Proceedings of the International Biohydrometallurgy Symposium IBS 2003, Athens, Greece - Vol II

BIOHYDROMETALLURGY "A SUSTAINABLE TECHNOLOGY IN EVOLUTION" Proceedings of the International Biohydrometallurgy Symposi...

Author: Marios Tsezos | Artin Hatzikioseyian | Emmanouela Remoundaki (eds.)

115 downloads 1083 Views 11MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

BIOHYDROMETALLURGY "A SUSTAINABLE TECHNOLOGY IN EVOLUTION"

Proceedings of the International Biohydrometallurgy Symposium, IBS 2003, held in Athens, Hellas, September 14-19, 2003

Part II Biosorption, Microbiology Fundamentals, Molecular Biology and Taxonomy

Edited by Marios Tsezos Artin Hatzikioseyian Emmanouela Remoudaki

Associate Editors Pavlina Kousi Roza Vidali

NATIONAL TECHNICAL UNIVERSITY OF ATHENS School of Mining and Metallurgical Enginnering Laboratory of Environmental Science and Technology Heroon Polytechniou 9, 157 80 Zografou, Greece Tel: (+30) 2107722172, (+30) 2107722271, Fax: (+30) 2107722173 Contact: Professor Marios Tsezos, e-mail: [email protected]

© National Technical University of Athens. All rights reserved.

This work and the individual contributions contained in it are protected under the following terms and conditions: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Derivative works Permission of the publisher is required for all other derivative works, including compilations and translations. Electronic storage or usage Permission of the publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Contact the publisher at the address indicated. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without any prior written permission of the publisher. Address permission requests to: National Technical University of Athens, at the mail, fax and e-mail addresses noted above. Notice No responsibility is assumed of the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein.

First edition 2004

Electronic means of production – Cover design N. & J. ZOGRAPHAKIS Tel.: (+30) 2109731912 Fax: (+30) 2109707707

Printed in Greece. ISBN- (Set) : 960-88415-0-X ISBN- (Part II) : 960-88415-2-6

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Preface The present edition includes the proceedings of the 15th International Biohydrometallurgy Symposium (IBS 2003) held in Athens, Greece on September 14th19th, 2003. Continuing the effort on the understanding of the interactions between metals and microbial cells and on developing and applying biohydrometallurgical processes, International Biohydrometallurgy Symposia offer the opportunity of exchanging international experience on a wide range of topics from metal extraction to environmental remediation. During IBS 2003, the issues of sustainability and environmental remediation, worldwide priorities, were addressed from different points of view. Biohydrometallurgy is a sustainable innovative technology, which in many cases, during the last decade, has successfully replaced classical metal extraction processes, from minerals and rocks. Combining a competitive technology with minimum environmental impact is the challenge for optimization of technologies applied today and/or to be applied in the future. Recent advances towards the quantitative description of the interactions between metals and microbial cells as well as the identification of key parameters controlling these interactions, play an important role in metal extraction processes optimization and in the development of treatment technologies for liquid and solid metallurgical discharges. The 15th International Biohydrometallurgy Symposium opened by the invited plenary lecture: "Biohydrometallurgy: a sustainable technology in evolution" given by Professor Giovanni Rossi from the University of Cagliari, Italy. Professor Rossi honored the Symposium with his presence, reviewed the state of the art and pointed out to the future trends in different areas of biohydrometallurgy. The Symposium was organized along five sessions: Bioleaching Applications and Technology Developments. Bioremediation – Environmental Applications. Biosorption Fundamentals and Technology Developments. Microbiology Fundamentals. Molecular Biology and Taxonomy. All papers included in the present edition were previously reviewed by a minimum of two experts selected among the International Scientific Committee Members as well as among prestigious researchers in the biohydrometallurgy science and technology fields. Among the 160 papers included in this edition, the Organising Committee aimed at providing the opportunity to the Symposium participants to attend as many oral presentations as possible, according to originality and scientific merit. Sixty-five oral i

Preface

presentations were made during IBS 2003. The rest of the communications were presented in the poster session. A Closing Session, chaired by a panel of experts and pioneers in the corresponding areas, was organized to conclude the main topics of the Conference and to point out future trends in scientific areas of Biohydrometallurgy. From this position, we wish to acknowledge all the members of the International Scientific Committee: Antonio Ballester, Barrie Johnson, Bohumil Volesky, Borje Lindstrom, Carlos Jerez, Corale Brierley, David Holmes, Dominique Morin, Douglas Rawlings, Edgardo Donati, Eric Guibal, Giovanni Rossi, Gregory Karavaiko, Henry Erhlich, James Brierley, John Duncan, K. A. Natarajan, Kishore Paknikar, Klaus Bosecker, Marios Tsezos, Olli Tuovinen, Paul Norris, Piet Bos, Ralph Hackl, Ricardo Amils, Stoyan Groudev, Tomas Vargas, Tsuyoshi Sugio, Virginia Ciminelli, Wolfgang Sand, for participating in the reviewing and selection of the manuscripts submitted to IBS 2003. We also wish to express our appreciation to prestigious researchers non members of the International Scientific Committee for assisting the reviewing procedure: Anthimos Xenidis, Frantz Glombitza, Georgios Anastassakis, Konstantinos Komnitsas, Ludo Diels, Lynne Macaskie, Nymphodora Papassiopi, Styliani Agatzini-Leonardou. We also thank our colleagues at IBS 2003 from the National Organizing Committee: Anthimos Xenidis, Emmanouil Zevgolis, Georgios Anastassakis, Konstatina Tsaimou, Konstantinos Komnitsas, Nymphodora Papassiopi, Simos Simopoulos, Styliani AgatziniLeonardou, for their valuable assistance and support. Acknowledgements are also due to the many others who participated in the organization of the Symposium, the authors and the many participants who represented many countries around the world. Special thanks also to Mrs Pavlina Kousi and Mrs Roza Vidali, Ph.D candidate students of our Laboratory, for editing the final manuscripts for the preparation of the hardcopies of the IBS 2003 proceedings. Finally, we wish to thank the National Technical University of Athens (NTUA), The Ministry of Development: General Secretariat of Science and Technology, The Hellenic Ministry of Culture, The Hellenic Technical Chamber, The National Institute of Geology and Mineral Exploration for supporting the Symposium. Professor Marios Tsezos Dr. Emmanouela Remoudaki Dr. Artin Hatzikioseyian National Technical University of Athens, School of Mining and Metallurgical Engineering, Laboratory of Environmental Science and Engineering

ii

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Table of contents Preface .................................................................................................................................. i Table of contents ................................................................................................................ iii

PART I PLENARY LECTURE Biohydrometallurgy: a sustainable technology in evolution Giovanni Rossi ...................................................................................................................... 3

CHAPTER 1 BIOLEACHING APPLICATIONS A novel bio-leaching process to recover valuable metals from Indian Ocean nodules using a marine isolate Mukherjee A., Raichur A.M., Modak J.M., Natarajan K.A. ............................................... 25 A novel biotechnological process for germanium recovery from brown coal Xianwan Y., Yun Z., Yuxia G., Banghui G. ......................................................................... 35 Aerobic and anaerobic bacterial leaching of manganese Zafiratos J.G., Agatzini-Leonardou S. ................................................................................ 41 Anaerobic iron sulfides oxidation Schippers A. ........................................................................................................................ 55 Bacterial growth and propagation in chalcocite heap bioleach scenarios Petersen J., Dixon D.G. ...................................................................................................... 65 Bacterial leaching studies of a Portuguese flotation tailing Costa M.C., Carvalho N., Iglesias N., Palencia I. ............................................................. 75

iii

Table of contents

Bacterial tank leaching of zinc from flotation tailings Panin V.V., Adamov E.V., Krylova L.N., Pivovarova T.A., Voronin D.Yu., Karavaiko G.I. .................................................................................................................... 85 Behaviour of elemental sulfur in the biohydrometallurgical processing of refractory gold-sulfide concentrates of various mineral types Sedelnikova G.V., Savari E.E. ............................................................................................ 91 Beneficiation of phosphatic ores from Hirapur, India Agate A.D. ........................................................................................................................ 101 Biohydrometallurgy of antimony gold-bearing ores and concentrates Solozhenkin P.M., Nebera V.P. ........................................................................................ 107 Bioleaching of Argentinean sulfide ores using pure and mixed cultures Frizan V., Giaveno A., Chiacchiarini P., Donati E. ......................................................... 117 Bioleaching of complex gold-lead ores Ulberg Z., Podolska V., Yermolenko A., Yakubenko L., Pertsov N. ................................. 127 Bioleaching of electronic scrap material by Aspergillus niger Ten W.K., Ting Y.P. .......................................................................................................... 137 Bioleaching of metallic sulphide concentrate in continuous stirred reactors at industrial scale – Experience and lessons Morin D., d’Hugues P., Mugabi M. ................................................................................. 147 Bioleaching of natural zeolite – the processes of iron removal and chamfer of clinoptilolite grains Styriakova I., Kolousek D., Styriak I., Lengauer C., Tillmanns E. ................................... 157 Bioleaching of pyrite by defined mixed populations of moderately thermophilic acidophiles in pH-controlled bioreactors Okibe N., Johnson D.B. .................................................................................................... 165 Biolixiviation of Cu, Ni, Pb and Zn using organic acids produced by Aspergillus niger and Penicillium simplicissinum Galvez-Cloutier R., Mulligan C., Ouattara A. ................................................................. 175 Biooxidation of pyrite by Acidithiobacillus ferrooxidans in single- and multi-stage continuous reactors Canales C., Gentina J.C., Acevedo F. .............................................................................. 185 Chemical chalcopyrite leaching and biological ferric solvent production at pH below 1 Kinnunen P.H.-M., Salo V.L.A., Pehkonen S.O., Puhakka J.A. ....................................... 193

iv

Table of contents

Comparative study of the bioleaching of two concentrates of chalcopyrite using mesophilic microorganisms in the presence of Ag(I) Lopez-Juarez A., Rivera-Santillan R.E. ............................................................................ 203 Comparison of air-lift and stirred tank batch and semi continuous bioleaching of polymetallic bulk concentrate Tipre D.R., Vora S.B., Dave S.R. ...................................................................................... 211 Effect of pH and temperature on the biooxidation of a refractory gold concentrate by Sulfolobus metallicus Nancucheo I., Gentina J.C., Acevedo F. .......................................................................... 219 Effect of the pulp density and particle size on the biooxidation rate of a pyritic gold concentrate by Sulfolobus metallicus Valencia P., Gentina J.C., Acevedo F. ............................................................................. 227 Enhancement of chalcopyrite bioleaching capacity of an extremely thermophilic culture by addition of ferrous sulphate Rubio A., Garcia Frutos F.J. ............................................................................................ 235 Evaluation of microbial leaching of uranium from Sierra Pintada ore. Preliminary studies in laboratory scale Paulo P.S., Pivato D., Vigliocco A., Lopez J., Castillo A. ................................................ 243 Extraction of copper from mining residues and sediments by addition of rhamnolipids Mulligan C.N., Dahrazma B. ............................................................................................ 253 Improving of film coating bioleaching using biorotor process Shahverdi A.R., Oliazadeh M., Rohi R., Davodi M. ......................................................... 261 Isolation and evaluation of indigenous iron- and sulphur-oxidising bacteria for heavy metal removal from sewage sludge Matlakowska R., Sklodowska A. ....................................................................................... 265 Kinetics of ferrous iron oxidation with Sulfolobus metallicus at 70°C Meruane G., Carcamo C., Vargas T. ............................................................................... 277 Kinetics of sulphur oxidation: pH and temperature influence on bioleaching Patino E., Sandoval R., Frenay J. .................................................................................... 285 Leaching of iron from China clay with oxalic acid: effect of acid concentration, pH, temperature, solids concentration and shaking Mandal S.K., Banerjee P.C. ............................................................................................. 291 Mathematical modeling of the chemical and bacterial leaching of copper ores in stack Zeballos F., Filho O.B., de Carvalho R.J. ........................................................................ 301 v

Table of contents

Model for bacterial leaching of copper sulphides by forced aeration Sidborn M., Moreno L. ..................................................................................................... 311 Optimal oxygen and carbon dioxide concentrations for thermophilic bioleaching archaea de Kock S.H., Naldrett K., du Plessis C.A. ....................................................................... 319 Optimization study on bioleaching of municipal solid waste (MSW) incineration fly ash by Aspergillus niger Xu T.J., Ting Y.P. .............................................................................................................. 329 Production of an Acidithiobacillus ferrooxidans biomass using electrochemical regeneration of energetic substrate Morra C., Gondrexon N., Magnin J.-P., Deseure J., Ozil P. ........................................... 337 Removal of dibenzothiophene from fossil fuels with the action of iron(III)-ion generated by Thiobacillus ferrooxidans: Analytical aspects Beskoski V.P., Matic V., Spasic S., Vrvic M.M. ................................................................ 345 Solids loading in the bioleach slurry reactor: mechanisms through which particulate parameters influence slurry bioreactor performance Harrison S.T.L., Sissing A., Raja S., Pearce S.J.A., Lamaignere V., Nemati M. ............. 359 The development of a hybrid biological leaching-pressure oxidation process for auriferous arsenopyrite/pyrite feedstocks Dymov I., Ferron C.J., Phillips W. ................................................................................... 377 The development of the first commercial GEOCOAT® heap leach for refractory gold at the Agnes mine, Barberton South Africa Harvey T.J., Bath M. ........................................................................................................ 387 The electrochemistry of chalcopyrite bioleaching using bacteria modified powder micro-electrode Hongxu L., Dianzuo W., Yuehua H., Renman R. .............................................................. 399 The influence of crystal orientation on the bacterial dissolution of pyrite Ndlovu S., Monhemius A.J. ............................................................................................... 409 The influence of temperature and pH on the bioleaching of copper from a flotation concentrate of chalcopyrite Medrano-Roldan H., Salazar M.F.M., Pereyra-Alférez B., Solis-Soto A., Ramirez-Rodriguez D.G., Alvarez-Rosales E., Galan-Wong L.J. .................................... 419 The role of chemolitotrophic bacteria in the oxide copper ore heap leaching operation at Sarcheshmeh Copper Mine Seyed Baghery S.A., Shahverdi A.R., Oliazadeh M. ......................................................... 423

vi

Table of contents

Three-stage revolving drum biohydrometallurgical reactor for continuous operation Loi G., Trois P., Rossi G. ................................................................................................. 429 Use of biosurfactants for the mineral surfaces modification Sadowski Z., Maliszewska I., Polowczyk I. ....................................................................... 439

CHAPTER 2 BIOREMEDIATION ENVIRONMENTAL APPLICATIONS A new bench scale restoration method for a mercury-polluted soil with a mercury resistant Acidithiobacillus ferrooxidans strain SUG 2-2 Negishi A., Maeda T., Takeuchi F., Kamimura K., Sugio T. ............................................ 449 A novel type of microbial metal mobilization: cyanogenic bacteria and fungi solubilize metals as cyanide complexes Brandl H., Stagars M., Faramarzi M.A. ........................................................................... 457 An approach to cyanide degradation in wastewater of gold ore processing Podolska V., Ulberg Z., Pertsov N., Yakubenko L., Imanakunov B. ................................ 465 Available options for the bioremediation and restoration of abandoned pyritic dredge spoils causing the death of fringing mangroves in the Niger Delta Ohimain E. I. .................................................................................................................... 475 Bacterial reduction of TcO4- under the haloalkaline conditions Khijniak T., Medvedeva-Lyalikova N.N., Simonoff M. ..................................................... 483 Biodegradation of cyanides under saline conditions by a mixotrophic Pseudomonas putida Bipinraj N.K., Joshi N.R., Paknikar K.M. ........................................................................ 491 Bioleach of a fluvial tailings deposit material indicates long term potential for pollution Willscher S., Clark T.R., Cohen R.H., Ranville J.F., Smith K.S., Walton-Day K. ............ 497 Bioleaching of copper converter slag using A. ferrooxidans Seyed Baghery S.A., Oliazadeh M. ................................................................................... 507 Biooxidation of mine tailings using a mixed bacterial population Zahari M.A.K.M., Jaapar J., Bunyok M.A., Sohor S.H., Ahmad W.A. ............................. 513 Chromate reduction by immobilized cells of Desulfovibrio vulgaris using biologically produced hydrogen Humphries A.C., Penfold D.W., Forster C.F., Macaskie L.E. ......................................... 525 vii

Table of contents

Clean-up of mine waters from a uranium deposit by means of a constructed wetland Groudev S.N., Komnitsas K., Spasova I.I., Paspaliaris I. ................................................ 533 Degradation of tetracyanonickelate (II) by Cryptococcus humicolus in biofilm reactors Kwon H.K., Woo S.H., Sung J.Y., Park J.M. .................................................................... 541 Development of a bio-process using sulfate-reducing bacteria to remove metals from surface treatment effluents Battaglia-Brunet F., Foucher S., Denamur A., Chevard S., Morin D., Ignatiadis I. ....... 549 Effects of total-solids concentration on metal bioleaching from sewage sludge Villar L.D., Garcia O. Jr .................................................................................................. 559 Enhancement of electrodialytic soil remediation through biosorption Jensen P.E., Ottosen L.M., Ahring B.K. ........................................................................... 567 Fundamentals of the uranium separation in constructed wetlands Glombitza F., Karnatz F., Fischer H., Pinka J., Janneck E. ............................................ 575 Geomicrobiological risk assessment of abandoned mining sites Bosecker K., Mengel-Jung G., Schippers A. ..................................................................... 585 Immobilisation and growth of Acidithiobacillus ferrooxidans on refractory clay tiles Donati E., Martinez L., Curutchet G. ............................................................................... 595 Investigation of bioremediation techniques for cleaning-up arsenic contaminated soils Vaxevanidou K., Papassiopi N., Paspaliaris I. ................................................................ 603 Leaching characteristics of heavy metals from sewage sludge by Acidithiobacillus thiooxidans MET Cho K.S., Moon H.S., Yoo N.Y., Ryu H.W. ....................................................................... 613 Mercury removal by polymer-enhanced ultrafiltration using chitosan as the macroligand Kuncoro E.K., Lehtonen T., Roussy J., Guibal E. ............................................................ 621 Microbial recovery of copper from printed circuit boards of waste computer by Acidithiobacillus ferrooxidans Cho K.S., Choi M.S., Hong J.H., Kim D.S., Ryu H.W., Kim D.J., Sohn J.S., Park K.H. .. 631 Oxidation of iron, sulfur and arsenic in mine waters and mine wastes: an important role for novel Thiomonas spp Coupland K., Battaglia-Brunet F., Hallberg K.B., Dictor M.C., Garrido F., Johnson D.B. ................................................................................................................................... 639 viii

Table of contents

Oxidation of metallic copper by Acidothiobacillus Ferrooxidans Lilova K., Karamanev D. .................................................................................................. 647 Process monitoring of biodesulfurization of high sulfur coal in packed columns using molecular ecology methods Gómez F., Cara J., Carballo M.T., Moran A., Amils R., García Frutos F.J. .................. 653 Regeneration of hydrogen sulfide using sulfate reducing bacteria for photo catalytic hydrogen generation Takahashi Y., Suto K., Inoue C., Chida T. ........................................................................ 663 Remediation of sites contaminated by heavy metals: sustainable approach for unsaturated and saturated zones Diels L., Geets J., Vos J., Van Broekhoven K., Bastiaens L. ............................................ 671 Removal of chromium(VI) through a two-step process using sulphur-oxidising and sulphate-reducing bacteria Donati E., Viera M., Curutchet G. ................................................................................... 681 Removal of Mn(II) ions by manganese-oxidizing fungus at neutral pHs in the presence of carbon fiber Sasaki K., Endo M., Takano K., Konno H. ....................................................................... 689 Simultaneous removal of oil and heavy metals from waste waters by means of a permeable reactive barrier Groudeva V.I., Groudev S.N., Doycheva A.S. .................................................................. 697 The exploitation of sulphate-reducing bacteria for the reclamation of calcium sulphate sludges Luptakova A., Kusnierova M., Bezovska M., Fecko P. ..................................................... 703 The role of metal–organic complexes in the treatment of chromium containing effluents in biological reactors Remoudaki E., Hatzikioseyian A., Kaltsa F., Tsezos M. ................................................... 711 The selective precipitation of heavy metals by sulphate-reducing bacteria Luptakova A., Kusnierova M., Bezovska M., Fecko P. ..................................................... 719

APPENDIX Author index ................................................................................................................... A-3 Subject index ................................................................................................................. A-11

ix

Table of contents

PART II CHAPTER 3 BIOSORPTION A methodological approach to investigate the pH effect on biosorption process: experimental and modeling procedures Veglio F., Beolchini F., Pagnanelli F., Toro L. ............................................................... 731 A model for the copper biosorption in dried leaves de Carvalho R.P., De Sousa A-M.G., Freitas J.R., Rubinger C.P.L., Krambrock K. ...... 741 Agar-plate screening of effective metal biosorbents among year Podgorsky V.S., Lozovaya O.G., Kasatkina T.P., Fomina M.A. ...................................... 749 Bioremediation of chromium using Bacillus polymyxa Thyagarajan H., Subramanian S., Natarajan K.A. .......................................................... 759 Biosorption and bioaccumulation of heavy metals by bacteria isolated from contaminated sites of Karachi, Pakistan Nuzhat A., Uzma B., Fouad M. Qureshi, Fehmida F. ...................................................... 771 Biosorption equilibria with Spirogyra insignis Romera E., Fraguela P., Ballester A., Blazquez M.L., Munoz J.A., Gonzalez F. ............ 783 Biosorption of 226Ra and Ba by Sargassum sp. da Costa W.C., Garcia O. Jr., de Azevedo Gomes H. ...................................................... 793 Biosorption of arsenic and heavy metals on a ceramic-based biomass. Batch equilibrium experiments with Cu2+ model solutions Horak G., Willscher S., Werner P., Pompe W. ................................................................. 799 Biosorption of chromium (VI) by marine algal biomass Tan L.H., Chen J.P., Ting Y.P. ......................................................................................... 807 Biosorption of heavy metal ions from aqueous solutions by local seaweeds Sheng P.X., Chen J.P., Ting Y.P. ...................................................................................... 817 Biosorption of heavy metals onto an olive pomace: adsorbent characterisation and equilibrium modelling Pagnanelli F., Ubaldini S., Veglio F., Toro L. ................................................................. 825 Biosorption of Hg by vegetal biomasses Pimentel P.F., de Carvalho R.P., Santos M.H., Andrade M.C. ....................................... 835

x

Table of contents

Biosorption of lead in aquatic environment by Mucor rouxii biomass Som Majumdar S., Saha T., Bandhapadhyay T., Chatterjee S., Guha A.K. ..................... 843 Cadmium(II) biosorption by Aeromonas caviae: kinetic modeling Loukidou M.X., Karapantsios T.D., Zouboulis A.I., Matis K.A. ...................................... 849 Chromium uptake by pretreated cells of Aeromonas hydrophila isolated from textile effluents Zakaria Z.A., Ahmad W.A. ................................................................................................ 859 Copper ion adsorbed on chitosan beads: Physico-chemical characterization Chatterjee S., Som Majumdar S., Chatterjee B.P., Guha A.K. ......................................... 869 Development of a process for biosorptive removal of mercury from aqueous solutions Tupe S., Paknikar K. ......................................................................................................... 877 Effects of ionic strength, background electrolytes and heavy metals on the biosorption of hexavalent chromium by Ecklonia biomass Park D., Park J.M., Yun Y.-S. ........................................................................................... 883 Evaluation of silver recovery from photographic waste by Thiobacillus ferrooxidans and chitin Thiravetyan P., Nakbanpote W., Songkroah C. ................................................................ 891 Influence of the treatment of fungal biomass on sorption properties for lead and mercury uptake Spanelova M., Svecova L., Guibal E. ............................................................................... 899 Lanthanum and neodymium biosorption by different cellular systems Palmieri M., Garcia O. Jr. ............................................................................................... 911 Modeling of chromium biosorption by seaweed Sargassum sp. biomass in fixedbed column in series Cossich E.S., Silva E.A., Tavares C.R.G., Mesquita H.M., Eidan L.S. ............................ 919 Modelling and optimisation of copper ion uptake by Acidithiobacillus ferrooxidans Boyer A., Baillet F., Magnin J.-P., Ozil P. ....................................................................... 925 Platinum and palladium recovery from dilute acidic solutions using sulfate reducing bacteria and chitosan derivative materials Chassary P., de Vargas Parody I., Ruiz M., Macaskie L., Sastre A., Guibal E. .............. 935 Preliminary study of lead sorption by selected sorbents Ly Arrascue M., Bauer-Cuya J., Peirano Blondet F., Roussy J., Guibal E. .................... 947

xi

Table of contents

Regeneration of biomass after sorption of heavy metals Massacci P., Migliavacca E., Ferrini M. ......................................................................... 957 Structural modeling of arsenic biosorption using X-Ray spectroscopy (XAS) Teixeira M.C., Duarte G., Ciminelli V.S.T. ...................................................................... 965 Uranium and thorium removal by a Pseudomonas biomass: sorption equilibrium and mechanism of metal binding Sar P., Kazy S.K., D’Souza S. F. ...................................................................................... 975

CHAPTER 4 MICROBIOLOGY FUNDAMENTALS A model for iron uptake in Acidithiobacillus ferrooxidans based upon genome analysis Quatrini R., Veloso F., Jedlicki E., Holmes D.S. .............................................................. 989 Activity and occurrence of leaching bacteria in mine waste at Cartagena, Spain, in the years 1991 until 2000 Sand W., El Korchi-Buchert D., Rohwerder T. ................................................................ 997 An AFM-study on the adhesion of Acidithiobacillus ferrooxidans and Leptospirillum ferrooxidans to surfaces of pyrite Kinzler K., Sand W., Telegdi J., Kalman E. ................................................................... 1003 An X-ray photoelectron spectroscopy study of the mechanism of microbially assisted dissolution of chalcopyrite Parker A., Klauber C., Stott M., Watling H.R., Van Bronswijk W. ................................ 1011 Analysis of chalcopyrite (CuFeS2) electrodes utilizing galvanic current in the presence of Acidithiobacillus ferrooxidans Bevilaqua D., Benedetti A.V., Fugivara C.S., Garcia O. Jr. .......................................... 1023 Application of the bacterial weathering of silicate minerals in the improvement of quality of non-metallics Styriakova I., Styriak I. ................................................................................................... 1029 Assessment of acid production potential of sulphide minerals using Acidithiobacillus ferrooxidans and microbial sulphate reduction using Desulfotomaculum nigrificans Chockalingam E., Subramanian S., Natarajan K.A., Braun J.J. .................................... 1037 Comparative study on pit formation and interfacial chemistry induced by Leptospirillum and Acidothiobacillus ferrooxidans during FeS2 leaching Tributsch H., Rojas-Chapana J. ..................................................................................... 1047 xii

Table of contents

Composition of biofilm communities in acidic mine waters as revealed by combined cultivation and biomolecular approaches Kimura S., Coupland K., Hallberg K.B., Johnson D.B. ................................................. 1057 Computational fluid dynamics simulation of immobilized Acidothiobacillus ferrooxidans Metodiev B., Lilova L., Karamanev D. ........................................................................... 1067 Contribution to the quantification of the Acidithiobacillus ferrooxidans biomass concentration from the oxygen uptake rate Savic D.S., Veljkovic V.B., Lazic M.L. ............................................................................ 1077 Electrochemical and microbiological characterization of mercury in contact with mud Cruz F., Welzel A., Sampaio C., Englert G.E., Müller I.L. ............................................ 1085 Evaluating the growth of free and attached cells during the bioleaching of chalcopyrite with Sulfolobus metallicus Escobar B., Hevia M.J., Vargas T. ................................................................................. 1091 Experimental and modeling studies on inhibition effect of solution conditions on activity of Acidithiobacillus ferrooxidans during biooxidation of mixed sulphidic concentrates Chandraprabha M.N., Modak J.M., Natarajan K.A. ..................................................... 1099 Ferrous ion oxidation by an activated carbon cloth modified with Acidithiobacillus ferrooxidans de J. Cerino-Cordova F., Magnin J.P., Gondrexon N., Ozil P. ..................................... 1109 Heavy metal precipitation by off-gases from aerobic culture of Klebsiella pneumoniae M426 Essa A.M.M., Macaskie L.E., Brown N.L. ...................................................................... 1119 Influence of pH, Mg2+ and Mn2+ on the bioleaching of nickel laterite ore using the fungus Aspergillus niger O5 Coto O., Gutierrez D., Abin L., Marrero J., Bosecker K. .............................................. 1127 Mercury tolerance of thermophilic Bacillus sp. and Ureibacillus sp. Glendinning K.J., Brown N.L. ........................................................................................ 1137 Reduction of Pd(II) with Desulfovibrio fructosovorans, its [Fe]-only hydrogenase negative mutant and the activity of the obtained hybrid bioinorganic catalysts Mikheenko I.P., Baxter-Plant V.S., Rousset M., Dementin S., Adryanczyk-Perrier G., Macaskie L.E. ................................................................................................................. 1147 Removal of cobalt, strontium and caesium from aqueous solutions using native biofilm of Serratia sp. and biofilm pre-coated with hydrogen uranyl phosphate Paterson-Beedle M., Macaskie L.E. ............................................................................... 1155 xiii

Table of contents

Removal of soluble manganese from mine waters using a fixed bed column bioreactor Johnson D.B., Miller H., Ukermann S., Hallberg K.B. .................................................. 1163 Sulfane sulfur of persulfides is the actual substrate of the sulfur-oxidizing enzymes from Acidithiobacillus and Acidiphilium spp. Rohwerder T., Sand W. ................................................................................................... 1171 Sulfate reduction at low pH by mixed cultures of acidophilic bacteria Sen A.M., Kimura S., Hallberg K.B., Johnson D.B. ....................................................... 1179 Sulfur assimilation in Acidithiobacillus ferrooxidans Valdes J., Jedlicki E., Holmes D.S. ................................................................................ 1187 Survival of acidophilic bacteria under conditions of substrate depletion that occur during culture storage Johnson D.B., Bruhn D. F., Roberto F.F. ...................................................................... 1195 Synthesis of nanophase hydroxyapatite by Serratia sp. N14 Yong P., Sammons R.L., Marquis P.M., Lugg H., Macaskie L.E. .................................. 1205 The effect of maintenance on the ferrous-iron oxidation kinetics of Leptospirillum ferrooxidans Dempers C.J.N., Breed A.W., Hansford G.S. ................................................................. 1215 The kinetics of thermophilic ferrous-iron oxidation Searby G.E., Hansford G.S. ............................................................................................ 1227 The role of microorganisms in dispersion of thallium compounds in the environment Sklodowska A., Golan M., Matlakowska R. .................................................................... 1237

CHAPTER 5 MOLECULAR BIOLOGY AND TAXONOMY A promiscuous, broad-host range, IncQ-like plasmid isolated from an industrial strain of Acidithiobacillus caldus, its accessory DNA and potential to participate in the horizontal gene pool of biomining and other bacteria Goldschmidt G.K., Gardner M.N., van Zyl L.J., Deane S.M., Rawlings D.E. ............... 1249 Analysis of salt-induced outer membrane proteins in Acidithiobacillus ferrooxidans NASF-1 Kamimura K., Yamakado M., Shishikado T., Sugio T. ................................................... 1261

xiv

Table of contents

Bioinformatic analysis of biofilm formation in Acidithiobacillus ferrooxidans Barreto M., Rivas M., Holmes D.S., Jedlicki E. ............................................................. 1271 Diversity of Gram-negative bacteria at Malanjkhand copper mine, India Dave S.R., Tipre D.R. ..................................................................................................... 1279 Expression proteomics of Acidithiobacillus ferrooxidans grown in different metal sulfides: analysis of rhodanese-like proteins Ramirez P., Valenzuela L., Acosta M., Guiliani N., Jerez C.A. ..................................... 1287 Integration of metal-resistant determinants from the plasmid of an Acidocella strain into the chromosome of Escherichia coli DH5α Ghosh S., Mahapatra N.R., Nandi S., Banerjee P.C. ..................................................... 1297 Involvement of Fe2+-dependent mercury volatilization enzyme system in mercury resistance of Acidithiobacillus ferrooxidans strain MON-1 Sugio T., Fujii M., Takeuchi F., Negishi A., Maeda T., Kamimura K. ........................... 1305 Microbial diversity of various metal-sulphides bioleaching cultures grown under different operating conditions using 16S-rDNA analysis d’Hugues P., Battaglia-Brunet F., Clarens M., Morin D. .............................................. 1313 Molecular ecology of the Tinto River, an extreme acidic environment from the Iberian Prytic Belt González-Toril E., Llobet-Brossa E., Casamayor E.O., Amann R., Amils R. ................ 1325 Phenotypic characterization and copper induced stress resistance in the extremely acididophilic Archaeon Ferroplasma acidarmanus Baker-Austin C., Dopson M., Bowen A., Bond P. .......................................................... 1337 Pyrite oxidation by halotolerant acidophilic bacteria Norris P.R., Simmons S. ................................................................................................. 1347 Reversible loss of arsenopyrite oxidizing capabilities by Acidithiobacillus ferrooxidans is associated with swarming phenotype and presence of ISAfel Hurtado J.E. ................................................................................................................... 1353 Searching for physiological functions regulated by the quorum sensing autoinducer AI-1 promoted by afeI/afeR genes in Acidithiobacillus ferrooxidans Farah C., Banderas A., Jerez C.A., Guiliani N. ............................................................. 1361 Systematic analysis of our culture collection for "genospecies" of Acidithiobacillus ferrooxidans, Acidithiobacillus thiooxidans and Leptospirillum ferrooxidans Mitchell D., Harneit K., Meyer G., Sand W., Stackebrandt E. ....................................... 1369

xv

Table of contents

The strain genotypic heterogeneity of chemolithotrophic microorganisms Kondrateva T.F., Pivovarova T.A., Muntyan L.N., Ageeva S.N., Karavaiko G.I. .......... 1379

APPENDIX Author index ................................................................................................................... A-3 Subject index ................................................................................................................. A-11

xvi

C HAPTER 3 Biosorption

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

A methodological approach to investigate the pH effect on biosorption process: experimental and modeling procedures Francesco Vegliòa*, Francesca Beolchinia, Francesca Pagnanellib and Luigi Torob a

Dipartimento di Chimica, Ingegneria Chimica e Materiali, Universita’ degli Studi di L’Aquila, 67040 Monteluco di Roio, L’Aquila, Italy b Dipartimento di Chimica, Facoltà di S.M.F.N., Università degli Studi "La Sapienza", P.le A. Moro, 5, 00185 Roma, Italy Abstract A methodological approach to study and model equilibrium of heavy metals in biosorption processes has been proposed and discussed. Four cases of copper biosorption are here reported and discussed as examples of application: biosorption of copper onto Sphaerotilus natans, Rhizopus oligosporus, calcium alginate and olive mill residues (OMR) have been here described and discussed. Several empirical and semi-empirical models have been proposed and summarised, to consider the pH effect on the heavy metal up-take. The proposed models, originated from Langmuir isotherm, may be useful to fit experimental data avoiding pH control during biosorption tests and simply monitoring its equilibrium value. The adsorption isotherms were built considering experimental procedures at constant pH (in standard manner) and in pH edge conditions. Both empirical and semi-empirical models were able to fit these experimental results. The pH-edge experimental procedure coupled with the proposed pH-related models is proposed as useful tools to investigate and model biosorption processes with single heavy metals in solution. Keywords: biosorption, heavy metals, equilibrium, modelling, pH-effect 1.

INTRODUCTION Biosorption is an innovative technology aimed at the removal of toxic metals from polluted streams by using inactive and dead biomasses. Metals entrapment is due to chemico-physical interactions with active groups present on the cell wall: carboxylic, phosphate, sulfate, amino, amide and hydroxyl groups are the most commonly found, according to the biosorbent nature (Cox et al., 1999; Plette et al., 1995; Veglio’ and Beolchini, 1997). Considering its mechanism, biosorption is affected by several factors such as pH, simultaneous presence of other metals, kind of biosorbent material. In any case, the development of a model in agreement with experimental data is fundamental in

*

Corresponding author: [email protected]

731

Biosorption

order to simulate and predict biosorption courses. Equilibrium models are essential, considering that they represent the first step encountered in the development of kinetic models to be applied in unit operations typical of chemical engineering, such as fixed bed adsorption columns (Veglio’ et al., 1999) and membrane adsorbers (Veglio’ et al., 2000; Beolchini et al., 2002). In the case of sorption equilibrium data obtained at different pH, the effect of pH on heavy metal biosorption cannot be considered by simple empirical and physical equations such as Frendlich and Langmuir models (Esposito et al., 2001, 2002). These adsorption isotherms can be used to fit only experimental data obtained at constant equilibrium pH (Esposito et al., 2001). Consequently, more complex models have to be considered, with pH as a further independent variable. In the present work, some empirical and semi-empirical models are presented to describe copper sorption equilibrium by four different biosorbents: Sphaerotilus natans, Rhizopus oligosporus, calcium alginate and olive mill solid residues. 2.

MATERIALS AND METHODS

2.1 Biosorbents Sphaerotilus natans is a Gram-negative bacterium isolated from the waste streams of a water purification plant. Further details for biomass cultivation and separation can be found elsewhere (Esposito et al., 2001). Rhizopus oligosporus has been supplied by C.R.A.B. (Consorzio per le Ricerche Applicate alla Biotecnologia, Avezzano, Italy). Calcium alginate beads were prepared as described in Veglio’ et al. (2002). Olive mill residues (OMR) were the solid residues of oil production, provided by an olive mill in Abruzzo, Italy. Before their use, the olive mill solid wastes were ground and sieved at 400-1000 µm. Further details can be found elsewhere (Veglio’ et al., in press) 2.2 Equilibrium tests Equilibrium biosorption tests were realised with two experimental procedures: sorption test at constant pH (noted in the following as standard method – STD) and pH edge tests. In both cases a selected amount of lyophilised biomass (0.1 g) was placed in a shaken flask with a known volume of distilled water: the biomass was re-hydrated for 1 h. A selected volume of a copper solution (prepared by dissolving CuSO4 in distilled water, 1 g/L) was added to the shaken flask maintaining the liquid total volume at 100 mL. The shaken flasks were then placed in a shaker at 250 min-1 at room temperature (25°C) and heavy metal uptake was monitored sampling the solution and measuring the copper concentration by atomic absorption spectrophotometer (Varian Spectra 2000): the heavy metal uptake and the concentration in the solid phase q (mg of Cu2+ per g of biomass) was estimated by material balance. In the first series of adsorption tests (STD) the pH was continuously monitored and controlled by adding HCl 0.1 M or NaOH 0.1 M until the equilibrium conditions were reached (after 1 h): each isotherm (at constant pH) was obtained increasing the initial copper concentration in different shaken flasks. The second series of tests (pH-edge tests) were carried out as STD tests, but pH was changed from low to high values (from about pH 3 to 5, adding NaOH 0.1M or 1M) in each shaken flask and viceversa (from pH 5 to 3, adding HCl 0.1 N or 1 M): in this manner, the total amount of copper is constant for each test and the copper in solution is monitored after that each pH change has been induced. In both cases, particular care was paid to have negligible dilution by alkali or acid solutions during the pH control. After each pH change, an 732

Biosorption

equilibration time of 60 min was used before the collection of the liquid sample. The q values were calculated also considering the copper collected during the sampling procedure in order to avoid a propagation of this systematic error. Most of the biosorption tests were replicated twice and the c.v. values ranged from 2 to 5%. 2.3 Analytical determinations Copper concentration in the liquid phase was determined by atomic absorption spectrophotometer (Varian Spectra 2000). All samples were diluted with HNO3 at pH 2 and stored at 4°C before the analysis. 2.4 Mathematıcal models The proposed empirical and semi-empirical models used to fit equilibrium data are summarised in the following equations: (Veglio’ et al., 2002; Pagnanelli et al., in press): Model 1a:

(

) α Ceq + Ceq

q = α 1 ⋅ pH + α 2 ⋅

(1a)

3

Model 1b:

(

) α Ceq + Ceq

(1b)

) α Ceq + Ceq

(1c)

q = α 1 ⋅ pH α 2 ⋅

3

Model 1c:

(

q = α 1 ⋅ e −α 2 pH ⋅

3

Model 1d: Model 2a:

Model 2b:

Model 3a:

q=

α 1 ⋅ pH + α 2 Ceq ⋅ α 3 + pH (α 4 ⋅ pH + α 5 ) + Ceq

α 1 ⋅ eα ⋅ pH q= α 1 − 1 ⋅ (1 − eα α3 2

q=

α 1 ⋅ eα ⋅ pH α 1 − 1 ⋅ (1 − eα α3

α1

q=

10

− pH

10

− pH

1+

⋅

α2 α1

q=

⋅

α2 α1

q= 1+

Model 3c:

)

2

1+ Model 3b:

2 ⋅ pH

10

− pH

α2

⋅

2 ⋅ pH

)

(1d)

⋅

Ceq α 4 + Ceq

(2a)

⋅

Ceq (α 4 ⋅ pH + α 5 ) + Ceq

(2b)

Ceq α 3 + Ceq

(3a)

Ceq ⎛ 10 − pH α 3 ⋅ ⎜⎜1 + α2 ⎝

⎞ ⎟⎟ + Ceq ⎠

Ceq ⎛ 10 − pH α 3 ⋅ ⎜⎜1 + α4 ⎝

⎞ ⎟⎟ + Ceq ⎠

(3b)

(3c)

All these equations have been built considering Langmuir model as a reference model, introducing the pH effect in the two parameters qmax and Ks (Esposito et al., 2001): in fact at constant pH all the equations degenerate in the classical Langmuir model.

733

Biosorption

The empirical models reported in equations (1a), (1b), (1c) and (1d) were named Model 1a, 1b, 1c and 1d, respectively (Pagnanelli et al., in press; Veglio’ et al., 2002); in equations (2a) and (2b) two different versions of the logistic equation coupled with Langmuir model have been shown (Veglio’ et al., 2002): both models were named Model 2a and 2b; the equations (3a), (3b) and (3c) have been originated from non-competitive biosorption models (Esposito et al., 2002; Veglio’ et al., 2002) with some empirical changes introduced considering the obtained experimental data (these were named Model 3a, 3b and 3c respectively). A detailed description of the non-competitive mechanism between H+ ions and the metal can be found elsewhere (Esposito et al., 2002). 4.

EXPERIMENTAL RESULTS AND DISCUSSION Obviously each equilibrium model gives different results according to the experimental system; in fact, as well known, the empirical models are in general built on specific experimental results. Each model previously described was tested on different experimental systems. Some examples are shown in Figures 1 to 4. Figure 1 shows models 2a (empirical) and 3b (non competitive semi-empirical) fitting to copper sorption equilibrium data by Sphaerotilus natans. Figure 2 shows results obtained in the study of equilibrium sorption by Rhizopus oligosporus. In particular, Fig. 2a shows Langmuir model parameters dependence on pH, as estimated considering separately each test performed at constant pH, while Fig. 2b shows model 2a (empirical) fitting, performed on all tests. Figure 3 shows the semi-empirical model (equation 3b) application for copper biosorption by calcium alginate. Figure 4 reports the empirical fitting (model 1a) of copper sorption equilibrium by olive mill residues.

80 pH=3

qeq (mg/g)

60

pH=4 pH=5

40

pH=6 Mod. 2a

20

Mod. 3b

0 0

50 Ceq (mg/L)

Figure 1. Empirical (equation 2a) and non-competitive semi-empirical (equation 3b) models for copper biosorption by Sphaerotilus natans with 1 g/L biomass concentration (points represent experimental data, obtained in pH edge tests)

734

Biosorption

Figure 2. Copper biosorption by Rhizopus oligosporus. Langmuir parameter dependence on pH (a) and sorption isotherms (b). In the isotherms, points represent experimental data (pH edge tests), while lines have been calculated by Model 3a (equation 3a)

735

Biosorption

q (mg/g)

12

8

Model 3 b STD Test 1 Test 2

4

Test 3

0

0

1000

2000

3000

4000

5000

Ceq (mg/L)

Figure 3. Semi-empirical model (equation 3b) application for copper biosorption by calcium alginate at pH 4: STD (standard tests – 5 mL of beads; pH = 4.0; Temperature 22°C); Test 1 (pH-edge test - 5 mL of beads; pH = 3.9; Temperature 22°C); Test 2 (pH-edge test - 5 mL of beads; pH = 3.8; Temperature 22°C); Test 3 (pH-edge test - 10 mL of beads; pH = 3.8; Temperature 22°C) (Veglio’ et al., 2002) 4

3

q (mg/g)

pH 3 4 5

2

1

0 0

20

40

60

80

100

120

140

Ceq (mg/L)

Figure 4. Sorption isotherms in the case of copper sorption by olive mill residues (biosorbent 1 g/L, standard test) Continuous lines have been calculated by equation 3a 736

Biosorption

Models showed in Figures 1 to 4 are the ones characterised by the best agreement between experimental and calculated specific uptakes. The data fitting was performed by a non linear regression method, in order to evaluate the adjustable parameters of each model (αi; i =1, p) by minimizing the sum of the squared deviations of experimental from calculated values of q (Himmelblau, 1978). The model validity is indicated by the following statistical parameters (Himmelblau, 1978; Montgomery, 1991): i) the parameter standard error; ii) the model residual variance (s2res), calculated as n

2 S res =

∑ (q i =1

i exp

− qical ) 2

(4)

n− p

where n is the total number of experimental points, p is the number of estimated parameters. Table 1 shows as example the obtained results for Model 2a fitting to copper sorption by different biosorbents. Other results can be found elsewhere (Esposito et al., 2002; Veglio’ et al., in press; Pagnanelli et al., in press). The performances of the selected models were also compared by using an F-Test (not reported here) (Montgomery, 1991; Veglio’ et al., 2002). This statistical tool permits to evaluate if there is a difference in the accuracy of the investigated tested models. Considering the results of the F-tests and the other statistical parameters previously reported, it was possible to select the best model. For example, in the case of copper biosorption by calcium alginate the best models are N°3a and 3b because they have the lowest s2res values, they have a physical meaning and they are able to describe the experimental results with the lowest number of adjustable parameters (p = 3). An analogous discussion was performed for each biosorption system and the best models have been individuated and reported in Figures 1 to 4. Table 1. Empirical model (Model 2a) fitting in the case of copper sorption by different biosorbents: adjustable parameters (± standard error), degree of freedom (d.f.), regression coefficient (R2), model residual variance Biosorbent Sphaerotilus natans

Rhizopus oligosporus Calcium alginate

X (g/L)

Model parameter α3 α2 (mg/g) 2.6±0.5 100±10

R2

SR2 (mg/g)

α4 (mg/L) 1.8±0.3

d.f. 24

0.986 11.205

0.5

α1 (mg/g) 0.00037±0.00009

1.0 2.0 1

0.022±0.002 0.0045±0.0005 1.3 ± 0.8

1.6±0.2 1.9±0.2 1.1 ± 0.2

90±10 68±8 240 ± 70

5.5±0.7 3.9±0.7 125 ± 10

30 23 36

0.995 0.988 0.990

2.331 5.852 56.5

1

0.2±0.2

1.8±0.4

14.9±0.6

480±50

4

0.96

9.63

A further important aspect comes out from the analysis of Figure 3 (and from other results with other biosorbents, not reported here): a comparison among the different adsorption equilibrium data highlights that pH-edge tests give similar results obtained in standard adsorption tests. In this way, it is possible to conclude that pH edge tests can be applied in equilibrium studies with several advantages from the practical experimental point of view (saving biosorbent material and laboratory time to mantain pH constant during the building of an adsorption isotherm). 737

Biosorption

4.

CONCLUSIONS In this paper, equilibrium models for biosorption in single metal systems are described and applied for several biosorbent materials. pH was included as independent variable in the equilibrium models, in order to have more flexible models, suitable for data fitting also in the case of not constant pH. In this way, pH edge tests (Veglio’ et al., in press) coupled with the proposed pH-related models result to be an effective tool procedure for the study of biosorption equilibrium.

5.

REFERENCES

1. Benedetti, M.F., Milne, C.J., Kinniburgh, D.G., van Riemsdijl, W.H., & Koopal, L.K. (1995). Metal ion binding by humic substances: application of the non-ideal competitive adsorption model. Environmental Science & Technology, vol 29, pp 446457. 2. Beolchini, F., Pagnanelli, F., Veglio', F., (2001). Modeling of copper biosorption by Arthrobacter sp. in a UF/MF membrane reactor. Environmental Science & Technology, vol 35, pp 3048-3054. 3. Chong, K.H., Volesky, B. (1995) Description of Two-metal biosorption equilibria by Langmuir-type models. Biotechnology and Bioengineering, vol 47, no 4, pp 451-460. 4. Cox, J.S., Smith, D.S., Warren, L.A., Ferris, F.G. (1999). Characterizing heterogeneous bacterial surface functional groups using discrete affinity spectra for proton binding. Environmental Science & Technology, vol 33, pp 4514-4521. 5. Esposito, A., Pagnanelli, F., Lodi, A., Solisio, C., Vegliò, F. (2001). Biosorption of heavy metals by Sphaerotilus natans: an equilibrium study at different pH and biomass concentrations. Hydrometallurgy vol 60, pp 129-141. 6. Esposito, A., Pagnanelli, F., Veglio’, F. (2002). pH-related equilibria models for biosorption in single metal systems. Chemical Engineering Science, vol 57, pp 307313. 7. Himmelblau, D. M. (1978). Process Analysis by Statistical Methods, John Wiley & Sons, New York. 8. Montgomery, D.C. (1991). Design and analysis of experiments, Third Edition. John Wiley & Sons. 9. Pagnanelli, F., Esposito, A., Toro, L., Veglio’. F. Metal speciation and pH effect on Pb, Cu, Zn and Cd biosorption onto Sphaerotilus natans. Water Research, in press 10. Plette, C.C.A, van Riemsdijl, W.H., Benedetti, M.F., van der Wal, A. (1995). pH dependent charging behaviour of isolated cell walls of a Gram-positive soil bacterium. Journal of Colloid and Interface Science, vol 173, pp 354-363. 11. Vegliò, F., Beolchini, F. (1997). Removal of metals by biosorption: a review. Hydrometallurgy, vol 44, pp 301-316. 12. Veglio', F., Beolchini, F., Barba, D. (2000). Experimental study and simulation on the biosorption of copper (II) in membrane reactors: a preliminary study. Industrial Engineering Chemistry & Research, vol 39, no 7, pp 2480-2484. 13. Veglio’, F., Beolchini, F., Boaro, M., Lora, S., Corain, B., Toro, L. (1999) Polyhydroxoethyl-methacrylate resins as supports for copper (II) biosorption with Arthrobacter sp.: matrix nanomorphology and sorption performances. Process Biochemistry, vol 34, pp 367-373. 14. Veglio’, F., Esposito, A., Reverberi, A.P. (2002). Copper adsorption on calcium alginate beads: equilibrium pH-related models. Hydrometallurgy, vol 2048. 738

Biosorption

15. Warren, L.A., Ferris, F.G. (1998). Continuum between sorption and precipitation of Fe(III) on microbial surfaces. Environmental Science & Technology, vol 32, pp 23312337. 16. Yang, J., Volesky, B. (1999). Modeling uranium-proton ion exchange in biosorption. Environmental Science & Technology, vol 33, pp 4079-4085.

739

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

A model for the copper biosorption in dried leaves R.P. de Carvalho∗, A-M.G. de Sousa, J.R. Freitas, C.P.L. Rubinger and K. Krambrock Depto. de Física, ICEx, Universidade Federal de Minas Gerais, CP 702, 30123-970, Belo Horizonte, MG, Brazil Abstract We report on biosorption of copper ions by dried leaves. Sorption experiments with lettuce leaves (L. sativa), mainly composed of cellulose, and corn husks (Z. mays), mainly composed of lignin, showed that the sorptive performance of L. sativa is better than all other plant biomasses studied earlier and that Z. mays presents a very weak sorptive capacity. Electron Paramagnetic Resonance (EPR) spectra revealed that the copper ion occupies a site with axial symmetry inside the biomass. Fourier Transform Infrared (FTIR) absorption spectra indicated that the presence of copper affects CH2, CO and OH bonds of the biomass structure. These results allowed us to conclude that copper ions sorbed by dried leaves occupy the sites located between two glucose rings of cellulose fibers.

Keywords: biosorption, plant biomass, copper, sorption sites, cellulose, lignin, EPR, FTIR 1.

INTRODUCTION The biosorption, that is, the capacity of biomasses to retain metallic ions from solutions, is widely known [1, 2], but few studies propose sorption models that could explain this phenomenon. From our previous results of the sorption of copper by dried leaves using Atomic Absorption Spectroscopy (AAS), Electron Paramagnetic Resonance (EPR) and Fourier Transform Infrared Spectroscopy (FTIR) we concluded that: (i) the sorption sites are similar in all plant leaves studied [3]; (ii) after sorption, the Cu2+ ion is incorporated in the biomass in a site with axial symmetry [3]; (iii) the sorption is more efficient in the fibre residuals than in the complete leaves [3]; and (iv) the presence of copper inside the biomass affects C bonds in carbon rings but the Cu ion does not substitute any other ion of the biomass [4]. The aim of this work is to investigate the site responsible for the sorption in dried plant leaves in order to propose a model for the sorption sites. Since the fibres are more effective in sorption than the complete leaves, this site must be located near fibre macromolecules, such as lignin or cellulose.

∗

Corresponding author (R.P. de Carvalho): Tel: +55-31-34995633; Fax: +55-31-34995600; E-mail address: [email protected]

741

Biosorption

2.

MATERIALS PREPARATION AND EXPERIMENTAL METHODS The sorption of copper ions was studied by using CuSO4 solutions prepared with a Carlo Erba Atomic Absorption standard solution. The biomasses were prepared with dried leaves of lettuce (Lactuca sativa) and corn husks (Zea mays). Biomasses were sun dried during 4 weeks then dried at 35°C during 7 days. After grounding in a domestic blender to a size of about 1mm the biomasses were dried at 70°C during 24h and washed in a water solution with pH=4 using a drop of diluted HCl solution. Chemicals used were analytical grade Aldrich products. The sorption experiments were done as described in [5] (equilibrium experiments): 100mg biomass were contacted with 50mL CuSO4 solution, with Cu initial concentrations varying from 20µg/mL to 200µg/mL. The pH of the solution was controlled and adjusted to 4.0. After 1h of contact, the solution was filtered and the Cu final concentration was determined. The Cu concentrations of the initial and final solutions were determined by Atomic Absorption Spectrometry (AAS) in a CG-AA-7000 equipment, following traditional procedures [6]. The metal uptakes q were determined from the initial and final Cu concentration (Ci and Cf, respectively) in a solution of volume V, and with a mass M of biomass, as described in [7]: (1) q = V(Ci – Cf) / M For each biomass the uptake results were fitted using the Langmuir sorption model [8,9]: (2) q = qobCf / (1 + bCf) where qo and b are the characteristic parameters of the Langmuir isotherm: qo represents the saturation uptake for high equilibrium concentrations and b is related to the affinity of the metal ion with the biomass structure, which defines the inclination of the isotherm for low equilibrium concentrations. For EPR measurements copper-charged samples of L.sativa were dried and inserted in quartz tubes in a custom-build spectrometer for spectra recording. The powder-like spectrum of Cu2+ in the biomass was simulated for comparison with the spectrum obtained experimentally. Details of the sample preparation and of the EPR spectra recording and simulation are given in [4], as well as an overview of the EPR theory. For FTIR studies natural and copper-charged samples of L.sativa were prepared as pellets using KBr (Graseby Specac LTD.) as substratum. Spectra were recorded with a BOMEM-DA8 FTIR spectrometer and deconvoluted. Details of the sample preparation, FTIR measurements and analysis are given in [4], as well as an overview of the FTIR theory. 3.

RESULTS AND DISCUSSION

3.1 Sorption isotherms Figure 1 shows the sorption isotherms constructed by fitting Langmuir-type curves to the uptake values measured for L. sativa - lettuce (L) and Z. mays - corn husk (Z), compared with those reported for M.truncata fibres (F) and complete leaves (C) from [4]. Table 1 shows the adjusted values of qo and b for the biomasses studied, as well as the correlation coefficient χ2 for the adjusted curves. 742

Biosorption

Figure 1. Sorption isotherms for L.sativa (L- ), M.truncata fibres (F-1) [4], M.truncata complete leaves (C-X) [4] and Z.mays (Z- ) Table 1. Values for qo, b and χ2 for the biomasses showed in Figure 1 biomass L.sativa M.truncata (fibers) [4] M.truncata (complete) [4] Z.mays

qo (mg/g) 33 + 1 26 + 4 21 + 1 6+3

b (L/mg) 0.22 + 0.04 0.04 + 0.02 0.02 + 0.03 0.04 + 0,01

χ2 11 10 9 7

It can be seen that the sorption capacity of the Z.mays is much smaller than the sorption capacity of all other biomasses, and that L.sativa biomass presents the highest sorption uptake capacity. Since Z.mays is formed mainly by lignin while L.sativa is mainly composed of cellulose, we conclude that the cellulose is the macromolecule responsible for the metal sorption in dried leaves. 3.2 EPR spectra Figure 2 shows the EPR spectrum for copper-charged L.sativa and the simulated spectrum obtained by assuming that the Cu ion is located in a site of nearly axial symmetry. For the simulation, the following spin Hamiltonian H was used: (3) H = βSgB + SAI where the first term is the electron Zeeman and the second the hyperfine interaction. The symbols g and A denote the g-tensor and the hyperfine tensor, respectively. β is the Bohr Magneton and S and I are the electron and nuclear spin, repectively. For the analysis of the EPR spectrum of Cu2+, S = ½ and I = 3/2, with 100% natural abundance due to the two isotopes 63Cu and 65Cu, has been taken. The best simulation of the EPR spectrum for Cu2+ in L.sativa could be obtained assuming axial g and A tensors with values of g|| = 2.33 and g⊥ = 2.11 and A|| = 110 G and A⊥ = 50 G, respectively.

743

Biosorption

Intensity (a.u.)

2

B 0

A

-2 1500

2000

2500

3000

3500

4000

4500

5000

H (Gauss)

Figure 2. Εxperimental (A) and simulated (B) EPR spectra for copper-charged L. sativa

From these results, in particular the g and A tensors symmetry, we conclude that the Cu ion is incorporated in an axial site of the biomass structure. A similar result was found earlier for other biomasses [4]. 2+

Absorption (a.u.)

3.3 FTIR spectra Figure 3 shows the FTIR absorption spectra for natural (A) and copper-charged (B) L.sativa, normalized to account for thickness differences in the prepared samples. It can be seen that the presence of copper modifies the infrared absorption of the sample in the region between 800 cm-1 and 1800 cm-1. This region is amplified and deconvoluted in Figures 4A (natural L.sativa) and 4B (copper-charged L. sativa).

0

B A

1000

2000

3000

Wavenumber (cm

4000 –1

5000

)

Figure 3. FTIR absorption spectra of natural (A) and copper-charged (B) L. sativa

744

Biosorption

Figure 4. FTIR absorption spectra and deconvolution for natural (A) and coppercharged (B) L. sativa

The absorption peaks shown in Figures 4A and 4B can be assigned to vibration modes of molecular bonds of the glucose rings ramification of cellulose [10]. This assignment is shown in Table 2. Table 2. Assignment of FTIR absorption peaks (Figure 4) to cellulose molecular vibrations. Feature 1 2 3 4 5 6 7 8 9 10 11 12

Wavenumber (cm-1) 880 1010 1055 1120 1160 1245 1325 1385 1450 1525 1645 1745

Assignment C-H out-of-plane deformation C-O stretching coupled to ring modes C-O stretching coupled to ring modes C-O stretching coupled to ring modes C-O-C stretching C-O stretching ring breathing with C-O stretching CH bending CH2 symmetrical bending C=C in carbon rings C=C stretching or O-H bending C=O stretching

It can be seen that the presence of copper causes changes in the peaks 8 and 9 (CH2 vibrations); 2, 4 and 7 (CO modes); and 11 (OH vibration). No absorption peaks were created or extinguished by the presence of the copper in the biomass indicating that no molecular bonds were formed or destroyed after the sorption of the metal ion. It is worthy to note that our investigation was limited to the mid-IR region (500-2000 -1 cm ). In this region we can observe practically all-internal molecular vibrations and, thus, the characteristics of organic molecules. Nevertheless, absorption peaks originated from vibrational modes of ionic species like copper would be located at lower frequencies (below 500 cm-1) and cannot be observed in samples made with KBr substrata. Similar changes in the absorption peaks of dried plants charged with metallic ions were seen in other works that studied copper sorption in other biomasses [3,4] or other metals sorption in L. sativa [11]. 745

Biosorption

4.

DISCUSSION The results presented above allow us to conclude that cellulose is the macromolecule responsible for the copper sorption in biomasses prepared from dried plant leaves. Cellulose is a polymer composed of glucose monomers, joined by the sharing of an oxygen atom, which can rotate around the molecular long axis, creating hydrogen bonds between two monomers, depending on their relative positions. Figure 5 illustrates part of the cellulose chain. HO

O

CH

HO

O

HO

O

2

HO

2

HO

CH

O

HO

HO

HO

O HO

2

O

CH

2

O

HO

O

CH

HO

HO

Figure 5. Part of the cellulose chain showing glucose monomers and their bonds with neighbors

Our EPR spectra revealed that the copper ion is located in a site with strong axial symmetry and FTIR spectra showed that the presence of the copper ion affects CH2, CO and OH bonds of cellulose and that no molecular bonds were created or destroyed in the process. CH2 and OH are present in the ramification and CO in the glucose rings of cellulose. OH bonds can also come from H2O molecules present in the structure. To fit the information obtained, we propose the model shown in Figure 6, where the copper ion, hydrated with 4 H2O molecules, is located near two of the CH2-OH cellulose ramifications, from two different glucose chains or from the same chain, folded in this region. CO bonds of the glucose ring near the ramification are also affected by the presence of the metal ion. In the solution, the copper ions are hydrated with 6 H2O molecules, in an octahedral symmetry; after sorption in the biomass, the ion looses two of the H2O molecules, situated in an axial position, in order to accommodate inside the biomass structure. The four remaining hydration molecules are located in a plane orthogonal to the axis formed by copper and the CH2-OH ramifications, explaining the axial symmetry of the copper ion neighborhood found in the EPR studies. O

HO HO O

OH

CH2

O

HO

O CH2

O

OH OH

O

Figure 6. Model for the site of a copper ion (solid circle at the center) after biossoption by L. sativa

Brown and Kevan [12] report an EPR study in copper-containing clays where the spectra obtained are similar to ours and are explained with the presence of a 4-hydrated copper ion in an intermediate layer between the crystal layers.

746

Biosorption

In a recent work, Boutreau and col. [13] used computational methods to show that, in the minimum energy situation, the copper ion affects O-H and ramification bonds of glucose, in agreement with our infrared results. 5.

CONCLUSION Sorption isotherms, EPR spectra and FTIR absorption spectra of the biosorption of copper ions in L.sativa (lettuce leaves) and Z.mays (corn husk) biomasses led us to conclude that cellulose is responsible for biosorption in dried leaves, and that, after sorption, a hydrated copper ion is located near two glucose rings of the cellulose structure, in a site near the glucose ramification and with axial symmetry neighborhood. Dried lettuce leaves showed to be an efficient raw material in regards to metal biosorption and could be used in the treatment of metal-loaded disposal waters by means of, e.g., continuous-flow sorption columns [9]. Since there is a huge waste disposal of salad leaves in the food distribution centers of all big cities, the use of this biomass for metal sorption could help the depollution of contaminated waters and will be a way to alleviate the problems of unsold leaves disposal.

ACKNOWLEDGEMENTS We wish to thank Dr. Marilene Marinho Nogueira and Ms. Zabelê Dantas Moura, from the Depto. Fisiologia Vegetal (UFMG) for the discussions on the plant composition, Ms. Zenaide Souza Vasconcelos, from the Depto. Engenharia Química (UFMG), for the AAS determinations, and our colleague Dr. Roberto L. Moreira for his help with the infrared analysis. This work was supported by the Brazilian financing agencies CNPq and FAPEMIG. REFERENCES

1. A.Esposito, F.Pagnanelli, F.Beolchini, V.Dovì, F.Veglìo. In: Ciminelli, V.S.T., Garcia Jr., O. (Eds.), Biohydrometallurgy: Fundamentals, Technology and Sustainable Development vol. B, Elsevier, Amsterdam, p. 89-97 (2001) 2. M.L.Arrascue, H.M.Garcia, O.Horna, E.Guibal. In: Ciminelli, V.S.T., Garcia Jr., O. (Eds.), Biohydrometallurgy: Fundamentals, Technology and Sustainable Development vol. B, Elsevier, Amsterdam, p.119-128 (2001) 3. R.P. de Carvalho, K.J. Guedes, M.V.B. Pinheiro, K. Krambrock, Hydrom. 59 (2-3), 407-412 (2001) 4. R.P.de Carvalho, J.R.Freitas, A.-M. G. de Sousa, R. L. Moreira, M. V. B. Pinheiro and K. Krambrock, accepted for the IBS-2001 special issue of Hydrometallurgy 5. R.P. de Carvalho, K.- H. Chong, B. Volesky, Biotech. Lett. 16, 875-880(1994). 6. J. Ramirez-Munoz, Atomic Absorption Spectroscopy, Elsevier, Amsterdam, 1968. 7. R.P. de Carvalho, K.- H. Chong, B. Volesky, Biotech. Prog. 11, 39-44(1995). 8. I.Langmuir, J.Am.Chem.Soc. May 1915, pp 1139-1167 9. B.Volesky. In: Ciminelli, V.S.T., Garcia Jr., O. (Eds.), Biohydrometallurgy: Fundamentals, Technology and Sustainable Development vol. B, Elsevier, Amsterdam, p.69-80 (2001) 10. Michel, A.J. In: Schuerch, C. (Ed.), Proc. 10th Cellulose Conference, John Wiley and Sons, N.Y., 995 (1989). 11. M.H. Santos, M.C. Andrade, P.F.Pimentel, M.M.de Magalhães, R.L.Moreira, R.P.de Carvalho, accepted for presentation on IBS-2003 12. D.R.Brown, L.Kevan, J. Am. Chem. Soc 110 (9), 2743-2748 (1988) 747

Biosorption

13. L. Boutreau, J. Tortajada, A. Luna, M. Alcamí, O. Mó, M. Yáñez, Int. J. Quantum Chem. 86 (1), 138-144 (2002)

748

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Agar-plate screening of effective metal biosorbents among yeast Podgorsky V.S.a, Lozovaya O.G.a, Kasatkina T.P.a, Fomina M.A.a a

Institute of Microbiology and Virology of National Academy of Science of Ukraine, Zabolotnogo str. 154, Kiev 03134, Ukraine

Abstract The use of microbial cells as biosorbents for heavy metals offers a potential alternative to existing methods for decontamination or recovery of heavy metals from environment. Yeasts can be successfully used in metal sorption. An agar-plate screening method was developed for a rapid isolation of heavy metal-accumulating microorganisms and preliminary estimation of their biosorption capacity. In present investigation a variety of pink-coloured and pigment-less yeast cultures isolated from different habitats (67 strains) were screened for accumulation of zinc, copper, lead, chromium and cobalt ions using agar-plate method. According to the agar-plate screening data, the best copper and zinc accumulation capacity was found for pink-coloured yeast Rhodotorula mucilaginosa, Rhodotorula aurantiaca, Rhodotorula glutinis and pigmented-less yeast Candida krusei, Williopsis californica.

Keywords: metal, biosorption, yeast, agar-plate screening 1.

INTRODUCTION An increase of environmental heavy metals pollution over last years has led researchers to search for the efficient methods for the treatment of heavy metals using biosorbents. Among the most perspective groups of microorganisms which have the ability to sorb heavy metals such as copper, zinc, lead, cobalt, chromium are the yeasts [17]. Yeasts are easy to grow, produce high yields of biomass and at the same time can be manipulated genetically and morphologically [2, 7]. It is known, that the yeasts can remove heavy metals and radionuclides from aqueous solutions and soils in substantial quantities [4, 8]. The yeast is capable to grow at rather high concentration of heavy metals. The resistance of the yeast to heavy metals varied considerably with metal and yeast genera [9-11]. The aim of this study was an extended screening of the most widespread species of yeast for their metal sorption ability using agar-plate screening method based on the visualization and interpretation of the metal distribution between agar and colonies by chemical precipitation with hydrogen sulphide [12]. The best biosorbents found among tested yeasts were studied further as alive population in submerged culture for their resistance to zinc and copper and metal accumulation.

749

Biosorption

2.

MATERIALS AND METHODS

2.1 Microorganisms There was used a variety of yeast cultures from the collection of yeasts of the Industrial Microorganisms Physiology Department, Institute of Microbiology and Virology of National Academy of Sciences of Ukraine, Kiev, isolated from diverse habitats such as soil, water, plants, human gastrointestinal tract and from industrial processes. A total of 67 yeast cultures comprising species of genera Debaryomyces, Kluyveromyces, Saccharomyces, Williopsis, Candida, Cryptococcus, Rhodotorula, Sporobolomyces were screened for metal sorption property. 2.2 Biomass preparation The yeast biomass was grown in the medium of the following composition (g/l): (NH4)2SO4 – 3.0; K2HPO4- 0.1; KH2PO4 – 1.0; MgSO4 – 0.7; NaCl – 0.5; glucose – 10.0; extract of yeasts 1%; pH was 6.8. For inoculation, the yeast cultures were grown 24h in the above medium with agar (20 g/l). For agar-plate screening method we used the above medium containing 5mM Cu2+, Co2+, Cr6+ and 10mM Pb2+, Zn2+ and 10 g/l agar. Metals were used as salts Cu(NO3)2, CoCl2, K2Cr2O4, Pb(NO3)2, ZnCl2 that were added to the agar from a 100mM stock solutions after cooling the stock down to 55°C. 2.3 An agar-plate screening method for a rapid isolation of heavy metalaccumulating yeasts The agar-plates were inoculated by punctual inoculation with yeast strains (4-8 colonies in one plate). The plates were incubated for 48h at 28°C and 60% relative humidity until the biggest colonies reached about 4 mm in diameter in the dark to allow the yeasts to accumulate the metal. After this 48h they were exposed to gaseous H2S for 10 min in a desiccator [12]. H2S was generated by reacting 3g Na2S with the stoichiometric amount of 10% HCl. After incubation the metal became visualized. The main optical effects are the staining of colonies due to accumulated and precipitated metal and the formation of light haloes around the colonies within the uniformly darkened agar as a result of the diffusion of dissolved metal towards the organism. The main parameter, indicating metal accumulation ability of yeast, was diameter of a halo. The plates were inspected using light microscopy. Forming the light haloes around the yeast colonies was detected and yeasts of interest were taken from the plates for further investigation. 2.4 Determination of metal ion accumulation and metal resistance of the best-found biosorbents in submerged culture To estimate the metal resistance of selected isolates, the test tubes containing 9.9 ml of medium mentioned before and comprising Cu2+ or Zn2+ in concentration 5-500 mg/l were inoculated with 0.1 ml of cell suspension (109cells/ml) of the yeast cultures grown for 24h. The cultures in tubes were grown aerobically at 28°C on a shaker during four days. The criterion of resistance of yeast cultures was yield of biomass, measured as optical density at 540 nm and recalculated as biomass dry weight, at different initial metal concentrations compared to non-metal control variant of medium. Initial and final (after yeast growth) values of medium pH were estimated. Metal concentration was measured using AAS in liquid medium after yeast biomass separation.

750

Biosorption

Table 1. The ability of yeasts to form light haloes around the colonies. Mean values ±SEM, n=3 Yeast species Debaryomyces hansenii

D. occidentalis

D. polymorphus

D. vanriji

Candida krusei Kluyveromyces lactis

K. marxianus

S. cerevisiae S. unisporus

Williopsis californica Williopsis californica

Strains 148 143 1895 103 150 107 147 146 683 686 679 681 1521 153 156 155 189 173 188 179 61t 727 327 328 1891 1892 301 305 298 1996 295 962 556 666 675 2078 668 248 250 258

The light haloes, mm Cu2+ Cr6+ Pb2+ Co2+ Pigment-less: 0 0 0 1.2±0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.85±0.3 0 0 0 0 0 0 0 1.42±0.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.18±0.1 0.94±0.3 0 0 0 1.04±0.2 0 0 0 1.3±0.2 0 0 0 0 0 0 0.81±0.1 0.4±0.01 0 0 1.6±0.2 1.5±0.15 0 0 0 0 0 0 1.39±0.1 0.99±0.3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.1±0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 1.9±0.02 1.05±.003 1.3±0.02 0 0 1.3±0.02 1.45±0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.9±0.4 1.4±0.05 0 0 0 0 0 0 0 0

Zn2+ 0 0 0 0 0 0 0 0 1.4±0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0.95±0.1 0 0 0 0 0 0 0 0 1.15±0.1 1.15±0.1 0 1.1±0.3 0 0 1.6±0.1 0 0 751

Biosorption Yeast species

Strains 263

Cryptococcus albidus

Cryptococcus laurentii Rhodotorula aurantiaca

R. glutinis

R. minuta

R. mucilaginosa

Sporobolomyces roseus

S. cerevisiae Rhodotorula sp. Cryptococcus sp.

3

1001 1037 1034 1014 1091 1104 1195 1198 1202 1517 1321 1328 1335 1341 1342 1343 1775 1776 1803 1716 1447 1938 1940 1968 4 WT

The light haloes, mm Cu2+ Cr6+ Pb2+ Co2+ 0 0 0 0.5±0.05 Pigmented (pink): 0 0 0.7±0.03 0.7±0.03 0 0 0 0 0 0 0.6±0.1 0.8±0.02 0 0 0 0 0 0 0.8±0.02 0.6±0.1 0 0 0 0 0 0 1.85±0.05 1.2±0.1 0 0 0 1.7±0.05 0 0 0 0 0 0 0 1.9±0.2 0 0 0 0 0 0 0 0 0 0 1.2±0.03 1.2±0.03 0 0 0 1.15±0.1 0 0 1.68±0.1 1.3±0.1 0 0 0 0 0 0 1.45±0.1 1.3±0.02 0 0 1.8±0.05 1.95±0.01 0 0 1.5±0.1 1.8±0.1 0 1.45±0.1 1.6±0.1 1.15±0.1 0 0 0 2.1±0.1 0 0 0 0 0 0 0 1.1±0.3 Control strains [3]: 0 0 1.8±0.1 1.55±0.2 0 2±0.005 1.9±0.05 1,85±0.1 1.9±0.1 2±0.1 1.8±0.1 1.15±0.1

Zn2+ 0 0 0 0 0 0 0 11.6±0.2 1.5±0.2 0 0 0 0 0 0 1.6±0.1 0 1.15±0.1 0 0 0 2±0.1 0 0 0.6±0.01 2.1±0.05 1.6±0.1

RESULTS AND DISCUSSION A total of 67 different strains were used in the study of accumulation of metals (copper, lead, zinc, cobalt and chromium). Our agar-plate screening data demonstrated very high inter- and intraspecific variations in metal accumulating capacity evaluated as light haloes in agar for studied yeasts (Table 1). Different isolates of the same species showed remarkable differences in metal accumulation from total inactivity to median and high activity for at least three studied metals (e.g., K. lactis 383, 1891, 1892 versus K. lactis 327, or W. californica 250, 258 versus W. californica 248). Among all investigated strains the ability to sorb copper was found to be the best for genera of Rhodotorula and Debaryomyces, Williopsis, lead for genera of Rhodotorula, Cryptococcus, Kluyveromyces, Debaryomyces, zinc for genera of Rhodotorula and Cryptococcus, Williopsis, cobalt for genera of Rhodotorula and Debaryomyces, chromium for genera of Rhodotorula and Saccharomyces (Table 1). Toxic metal sorption capacity varied widely among different yeast strains. As control cultures, that have been studied previously [3], we used strains of 752

Biosorption

pigment-less Saccharomyces cerevisiae 1968 (industrial strain), pink Rhodotorula sp. 4 and black Cryptococcus sp. WT as examples of yeasts with different cell wall composition and with known metal sorption ability. Ability to accumulate heavy metals was observed for both pigmented and pigmented-less strains of yeast. The one-way ANOVA analysis of the data obtained (Table 1) showed that for all studied metal species the diameters of light haloes did not differ significantly for pigment-less yeast strains compared to pink ones (Pvalues>0.1 for all metal ions). However the best uptake capacity was shown by some pink-pigmented yeasts of genera Rhodotorula. Table 2. Growth of yeasts at different concentration of copper. *Metal-free control. Data are mean values ±SEM, n=3 Amount of biomass (g/l DW) Yeast strains Cu2+ (mg/l) 0* 5 10 50 100 Candida krusei 61T 1.8±0.1 1.8±0.04 1.6±0.02 0.45±0.6 0.4±0.1 Candida krusei 727 1.7±0.04 1.7±0.2 1.7±0.01 1.4±0.05 0.35±0.5 Cryptococcus sp. WT 2.1±0.2 1.6±0.1 1.2±0.05 0.95±0.1 0.21±0.01 R. aurantiaca 1195 2.2±0.03 1.8±0.3 1.8±0.3 1.8±0.05 1.7±0.5 R. aurantiaca 1198 1.75±0.02 1.7±0.2 1.7±0.2 1.65±0.3 0.6±0.1 R. minuta 1342 1.85±0.1 1.85±0.1 1.65±0.05 0.4±0.04 0.2±0.03 R. mucilaginosa 1776 2.1±0.03 2.1±0.25 1.4±0.03 1±0.05 0.3±0.01 R. mucilaginosa 1803 1.9±0.1 1.7±0.1 1.7±0.06 1.7±0.02 0.6±0.05 Rhodotorula sp. 4 2±0.2 2±0.2 2±0.02 1.3±0.1 0.6±0.01 S. cerevisiae 556 0.85±0.2 0.3±0.1 0.3±0.02 0.26±0.1 0.12±0.01 S. cerevisiae 962 0.85±0.4 0.85±0.2 0.85±0.6 0.78±0.1 0.5±0.02 S. cerevisiae 1968 1.25±0.1 1.2±0.4 1.15±0.1 1.1±0.2 0.2±0.02 W. californica 248 2±0.05 1.9±0.2 1.9±0.1 1.6±0.1 0.4±0.2

200 500 0.15±0.01 0 0.2±0.02 0 0.11±0.05 0 0.7±0.02 0.2±0.01 0.4±0.01 0 0.2±0.02 0 0.1±0.05 0 0.6±0.01 0 0.4±0.02 0 0.1±0.05 0 0.25±0.01 0 0.15±0.01 0 0.35±0.05 0

Table 3. Growth of yeasts at different concentration of zinc. *Metal-free control. Data are mean values ±SEM, n=3 Yeast strains 0* Candida krusei 61T 0.90±0.02 Candida krusei 727 0.46±0.03 Cryptococcus sp. WT 1±0.02 R. aurantiaca 1195 1.2±0.01 R. aurantiaca 1198 0.8±0.01 R. minuta 1342 1.7±0.01 R. mucilaginosa 1776 1.65±0.01 R. mucilaginosa 1803 1.6±0.02 Rhodotorula sp. 4 S. cerevisiae 556 S. cerevisiae 962 S. cerevisiae 1968 W. californica 248

5 0.6±0.02 0.16±0.02 0.9±0.02 1.2±0.02 0.8±0.01 0.9±0.02 1.2±0.2

Amount of biomass (g/l DW) Zn2+ (mg/l) 10 50 100 0 0.4±0.02 0.1±0.04 0 0.1±0.02 0.1±0.04 0.8±0.01 0.55±0.02 0.2±0.005 0.75±0.01 0.58±0.03 0.45±0.04 0.8±0.03 0.75±0.3 0.35±0.05 0.95±0.02 0.3±0.01 0.2±0.006 0.8±0.01 0.2±0.01 0.2±0.001

1.25±0.02 1.25±0.01 0.7±0.01

0.2±0.03

200 0 0 0.15±0.05 0.2±0.01 0.2±0.01 0.2±0.05 0.1±0.02

0.1±0.01 0 1.25±0.03 1.2±0.02 1.2±0.01 1±0.02 0.2±0.005 0 0.45±0.02 0.45±0.02 0.34±0.02 0.34±0.02 0.1±0.001 0 0 0.5±0.03 0.6±0.02 0.55±0.01 0.2±0.01 0.75±0.02 0.7±0.02 0.7±0.01 0.6±0.03 0.21±0.01 0.1±0.001 0 0 0.92±0.02 0.65±0.02 0.65±0.02 0.18±0.01

500 0 0 0 0 0 0 0 0 0 0 0 0 0 753

Biosorption

The sorption ability of metals depends on structure of cell wall of yeasts. The fungal (yeast) cell wall accounts for about 20-30% of cellular dry weight, is responsible for the shape of the cell, offers protection against mechanical damage and functions as a molecular sieve [13]. The best–studied from the Ascomycotina yeast in terms of cell wall composition and molecular architecture are currently S. cerevisiae and Candida albicans. The cell wall of the budding yeast S. cerevisiae contains four classes of components, namely chitin, β1,3-glucan, β1,6-glucan, and mannoproteins [13]. Many other yeasts cell walls contain α1,3–glucan. Other yeast’s cell wall carbohydrate polymers are proteinbound (galacto)mannan and acidic polysaccharides. Moreover, the cell walls of coloured yeasts contain different concentrations of pigments. The strains of pink-coloured yeasts (Rhodotorula) can produce very high quantity (630 µg/g DW biomass) of carotenoids pigments (beta-carotene, torulene, torularhodin) [14]. The black yeasts (e.g., Cryptococcus, Exophiala) blacken by the polymerization of dihydroxynaphthalene (DHN) into melanin (DHN-melanin) in their cell walls [15]. The main role in the processes of sorption by pigment-less yeasts is played by chitin and glucan-mannoprotein complex. Whereas for sorption by pigmented yeasts, the significant role is played by chelation properties of melanins and carotenoids. The major constituents of fungal cell walls such as chitin and melanin (for black yeasts) have significant metal binding abilities [2, 7, 9, 16]. The cell wall components of both main groups of tested yeast cultures, pigment-less Saccharomyces, Candida and Williopsis, containing chitin, glucans and mannan, and pink Rhodotorula and Sporobolomyces, containing chitin and carotenoid pigments and deprived of mannoproteins, could have active metal sorption sites [6, 7, 16]. Negative charge of yeast cell walls is formed mainly by carboxyl, hydroxyl, and phosphate groups. Composition, architecture and sorption properties of yeast cell wall can be altered by different conditions of growth, age and physiological state of cells [6, 10]. Metal sorption by yeast biomass has been studied for more than twenty years [1, 2, 4, 8, 9, 11, 17], but there is still a lack of information about a connection between sorption capacity of yeast cells and cell walls composition, as well as between sorption sites availability and physiological state of yeast cells. As a result of agar-plate screening, 13 cultures, that demonstrated the greatest values of light haloes on the plates with copper or zinc, were selected for further investigation of their zinc and copper toxicity and their ability to sorb heavy metals from medium in submerged culture (Table 2, 3, Fig. 1, 2). The pH values of medium after yeast growth decreased from initial 6.8 to 4.3-5.7 for copper-containing media, and to 3.9-4.8 for zinccontaining media. Resistance of yeast cultures to copper and zinc, measured as biomass yield in liquid medium, varied between species and strains (Table 2, 3). The most copperresistant yeasts were R. aurantiaca 1195 and R. mucilaginosa 1803 (Table 2). Rhodotorula strains showed also the highest resistance to zinc (Table 3). The most sensitive to zinc yeast cultures appeared to be pigment-less yeasts (Candida, Saccharomyces and Williopsis) (Table 3). In general, high metal resistance of pigmented yeasts compared to pigment-less suggests the protective role of yeast pigments (carotenoids, melanins) [17]. However, the ability of carotenoid pigments production by Rhodotorula cultures was altered by copper and zinc. It was found that pink cells of most Rhodotorula strains were gradually losing their colouring with increasing copper or zinc concentration and some of them completely lost the ability to form pigments at 100 mg/l (e.g. R. mucilaginosa 1803 (copper-medium) and 1776 (zinc-medium). Sorption screening experiments on the removal of metal ions by alive yeast biomass from liquid medium showed that the best copper sorbents were R. aurantiaca 1195 and C. 754

Biosorption

krusei 61T, regardless control strains S. cerevisiae 1968, Rhodotorula sp. 4 and Cryptococcus sp. WT (Fig. 1). There were found very similar zinc sorption data that did not differ statistically for the most of tested cultures. The high zinc sorption capacities were demonstrated by R. aurantiaca 1195, R. mucilaginosa 1776, R. minuta 1342 and W. californica 248 and were similar to the values, obtained for control strains S. cerevisiae 1968 and Cryptococcus sp. WT (Fig. 2). Such variation in metal accumulation by alive yeasts could be due to the differences in mechanisms of extra- and intracellular metal sequestration due to the differences in the cell wall and extracellular matrix structure, intracellular metal transport, compartmentation, efflux of metal cations, etc. [2, 7, 11, 16, 17]. It was also revealed some weak negative correlation (Zn: R=-0.57; Cu: R=-0.3) between biomass produced by yeasts, indicating their metal tolerance, and percent of metal removal from liquid medium at initial metal concentration 50 mg/l.

Metal sorption (%)

70 60 50 40 30 20

W. californica 248

S. cerevisiae 1968

S. cerevisiae 962

S. cerevisiae 556

Rhodotorula sp. 4

R. mucilaginosa 1803

R. mucilaginosa 1776

R. minuta 1342

R. aurantiaca 1198

R. aurantiaca 1195

Cryptococcus sp. WT

Candida krusei 61T

0

Candida krusei 727

10

Figure 1. Cu2+ sorption by yeasts at initial metal concentration 50 mg.l-1. Data are means derived from three replicated determinations, error bars are ±SEM

In terms of specific metal sorption capacity per gram of sorbing biomass, the highest specific copper sorption was found for R. aurantiaca 1195 (240 mg/g biomass) and R. mucilaginosa 1779 (255 mg/g biomass), being at least twice greater than values for all other tested cultures. The considerably higher values of specific zinc sorption (twice and threefold) were found for R. aurantiaca 1195 (182 mg/g biomass), and R. minuta 1342 (284 mg/g biomass) and Cryptococcus sp. WT (280 mg/g biomass). All mentioned above cultures were pigmented. According to the results of both agar-plates and aqueous solutions metal sorbent screening experiments, including analysis of specific sorption capacity per gram of 755

Biosorption

biomass, R. aurantiaca 1195 appeared to be the most efficient universal sorbent for both copper and zinc among all tested cultures. However, the analysis of the data of light haloes and percent of metal removal from liquid medium for zinc showed the absence of correlation between these two groups of data. Moreover, some strains (e.g., R. mucilaginosa, C. krusei (see Table 1)) that did not produce a light haloes around colonies in agar appeared to be efficient (40-50%) in zinc removal from liquid medium. Such difference between data from agar-plates and liquid medium screenings could be a consequence of abiotic and biotic factors. Abiotic factors could be different physico-chemical conditions in two methods affecting the changes in metal speciation and availability (and bioavailability). Biotic factors might be the differences between the growth conditions on solid and liquid medium altering metal toxicity and leading to the physiological and biochemical changes in yeast cells, including cell walls composition and sorption properties. Nevertheless, quite high positive correlation (R=0.75) between the data of light haloes and percent of metal removal from liquid medium was found for copper. Thus, the suitability of agar-plate screening for metal biosorbents depends on metal species as it has been already discussed [12].

70 Metal sorption (%)

60 50 40 30 20

W. californica 248

S. cerevisiae 1968

S. cerevisiae 962

S. cerevisiae 556

Rhodotorula sp. 4

R. mucilaginosa 1803

R. mucilaginosa 1776

R. minuta 1342

R. aurantiaca 1198

R. aurantiaca 1195

Cryptococcus sp. WT

Candida krusei 61T

0

Candida krusei 727

10

Figure 2. Zn2+ sorption by yeasts at initial metal concentration 50 mg.l-1. Data are means derived from three replicated determinations, error bars are ±SEM

756

Biosorption

4.

CONCLUSION It can be concluded that the agar-plate screening method is very useful for extended primary search for efficient metal biosorbents and it could be a great help for the initial qualification of the sorption properties of alive yeast biomass, being combined and verified with traditional methods of metal sorption from aqueous solutions. However, it should be taken into account that this method has some limitations, e.g., evaluates metal sorption capacity of biomass only indirectly and under specific conditions (surface growth on agar medium), and strongly depends on metal species and yeast strains. ACKNOWLEDGEMENTS We gratefully acknowledge Dr S.S. Nagornaya for the provision of yeast cultures from the collection of the Department of Industrial Microorganisms Physiology, Institute of Microbiology and Virology of NAS of Ukraine, Kiev. REFERENCES

1. Cervantes C., Campos-Garsia J., Devars S., Gutierrez-Corona F., FEMS Microbiology Reviews, 25 (2001) 335. 2. Gadd G.M., Current Opinion in Biotechnology, 11 (2000) 271. 3. Kasatkina T., Fomina M., Ignatova E., Nagornaya S. and Podgorsky V., International Biohydrometallurgy Symposium, IBS-2001, “Fundamentals, Technology and Sustainable Development”, Brazil, B (2001) 99. 4. Marques P.A.S.S., Rosa M.F. and Pinheiro H.M., Bioprocess Eng., 23 (2000) 135. 5. McEldowney S., Appl. Biochem. and Biotechnol., 26 (1990) 159. 6. Phaff H.J., Ann. Rev. Microbial., 17 (1963) 15. 7. Volesky B. and Holan Z.R., Biotechnol. Prog., 11 (1995) 235. 8. Wakatsuki T., Michiko I., Inahara H., J. Ferment. Technol., 66 (1988) 257. 9. Gadd G.M., Journal of Applied Bacteriology, 54 (1983) 57. 10. McEldowney S., Hardman D.J. and Waite S., Pollution: Ecology and Biotreatment, Longman Scientific and Technical, Singapore, 1993, 322. 11. Volesky B., May-Phillips H.A., Appl. Microbiol. Biotechnol., 42 (1995) 797. 12. Pumpel T., Pernfub B., Pinger B. et al., J. Indust. Microbiol., 14 (1995) 213. 13. De Nobel H., Sietsma J.H., Van Den Ende H. and Klis F.M., The Mycota VIII. Biology of the Fungal Cell, Springer-Verlag Berlin Heidelberg, 2001, 181. 14. Buzzini P. and Martini A., Bioresource Technology, 71 (2000) 41. 15. Huffnagle G.B., Chen G.H., Curtis J.L., McDonald R.A. et al., J. Immunol., 155 (1995) 3507. 16. Kapoor A. and Viraraghavan T., Bioresource Technology, 53 (1995) 195. 17. Mowll J.L. and Gadd G.M., J. Gen. Microbiol., 129 (1983) 3421.

757

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Bioremediation of chromium using Bacillus polymyxa Hemamalini Thyagarajan, S. Subramanian and K.A. Natarajan Department of Metallurgy, Indian Institute of Science, Bangalore – 560 012, India Abstract The use of a Gram positive, neutrophilic, facultative anaerobe, namely Bacillus polymyxa in the biosorption and the bioreduction of chromium species has been illustrated. Bioreduction of Cr (VI) and its biosorption has been monitored during bacterial growth using Cr (VI) adapted and unadapted strains. Bioremoval of both Cr (VI) and Cr (III) has been assessed with respect to time, pH, initial chromium concentration and biomass loading. The results indicate that bioreduction of Cr (VI) is feasible during growth of both adapted and unadapted strains. About 90% bioremoval of Cr (VI) could be achieved in 72 h using unadapted Bacillus polymyxa, while it took only 48 h using adapted bacteria for a similar amount of reduction. The bacterial metabolite is also found to be efficient in bioremoval of Cr (VI). Possible mechanisms of bioremoval of chromium species are discussed.

Keywords: bioremediation, chromium, Bacillus polymyxa, biosorption, bioreduction 1.

INTRODUCTION Chromium is a metal that exists in a variety of oxidation states, the most common being the +3 and the +6 forms. In the industrialized world, the use of chromium in industries like electroplating, textile, leather tanning, metallurgical metal finishing, photography, dye manufacturing, ink and pigments, power generation, and chemical manufacturing etc., is extensive, and hence it is not uncommon for the aqueous effluents from such industrial plants to have high amounts of chromium. Additionally, this can lead to the contamination of the soils or sediments that they contact. It has been estimated that more than 1,70,000 tonnes of chromium wastes are released annually in the United States of America, mainly due to industrial practices [1]. The consumption of basic chromium sulphate in the chrome tanning industries is reported to be about 50,000 tonnes per annum in India alone [2]. Hexavalent chromium (Cr (VI)) compounds are known to be toxic, mutagenic and carcinogenic, apart from being highly water-soluble. The Cr (III) form on the other hand, is innocuous and less soluble. Several conventional methods like precipitation and ion exchange are known for the elimination of toxic heavy metals from aqueous solution, but they are of high cost and low economic viability, especially at low and variable heavy metal concentration, and thus make these processes less advantageous. Accumulation of metals by microorganisms has been known for a few decades but has received more attention in recent because of its potential application in environmental protection or recovery of precious or strategic metals. Development of new biosorbent materials from microbial biomass is an emerging 759

Biosorption

area of significant interest. This biotechnological tool has great importance, as it is both cost effective and environmental friendly. Microbial reduction of toxic hexavalent chromium to the less soluble trivalent form represents a useful detoxification process that has been shown to be of practical importance for the removal of chromium from industrial wastewaters [3-4]. The unique ability of bacteria to reduce Cr (VI) may provide a possible means for cleaning up hazardous Cr (VI) wastes. The earliest reports pertained to a microorganism, Pseudomonas chromatophila, isolated from industrial sewage, which could use chromate or dichromate as a terminal electron acceptor during anaerobic respiration [5]. Subsequently, a number of microorganisms have been identified, which are capable of reducing hexavalent chromium, under different conditions [6-7]. It thus becomes imperative to devise suitable strategies to detoxify the Cr containing wastewater. Keeping these objectives in mind, the present investigation was taken up. The thrust of the present investigation has been directed towards the assessment of the potential of a Gram positive, heterotrophic, neutrophilic, facultative anaerobic soil bacterium, namely Bacillus polymyxa, for the bioremoval of Cr (VI) and Cr (III). 2.

EXPERIMENTAL

2.1 Bacterial strain A pure strain of Bacillus polymyxa NCIM 2539 was obtained from the National Collection of Industrial Microorganisms, National Chemical Laboratory, Pune, India and used for all the studies. The bacterium was cultured in the modified Bromfield medium [8]. The pH of the medium was adjusted to 7. 2.2 Reagents Analytical grade reagents namely potassium dichromate (K2Cr2O7) and chromium nitrate (Cr(NO3)3.9H2O) were procured from Qualigens, Mumbai, India and Loba Chemie Ltd., Mumbai, India, respectively and used as the hexavalent and trivalent forms of chromium respectively. Nitric acid and potassium hydroxide were used as pH modifiers. All the reagents used in this study were of analytical reagent grade and were made up in deionised double distilled water of conductivity < 1.5 µmho 2.3 Analytical estimation of Cr (VI) and Cr (III) A 0.25% w/v solution of diphenyl carbazide was prepared in 50% acetone. 15 ml each of the sample solution, containing various concentrations of Cr (VI) were pipetted out into 25ml standard flasks. To this 2ml of 3M H2SO4 was added followed by 1ml of diphenyl carbazide and the total volume was made upto 25 ml using deionised double distilled water such that the final concentrations were in the range 0.15 to 0.3 ppm. The intensity of the colour complex formed was measured using a Shimadzu model UV-260 uv-visible spectrophotometer. The absorbance was measured against a reagent blank at 540-nm wavelength maximum. A linear plot was obtained indicating adherence to the BeerLambert's law in the concentration range studied. The total chromium was analysed using a Thermo Jarrell Ash Video 11E atomic absorption spectrophotometer. The instrumental parameters were set as per the specifications provided [9]. The concentration of Cr (III) was obtained as the difference 760

Biosorption

between the values of the total chromium content and the amount of Cr (VI) estimated as per the procedures highlighted above. 2.4 Bacterial cell count The bacterial cell count was enumerated by microscopic counting using a PetroffHausser counter under a phase contrast Leitz microscope. 2.5 Bioremoval and bioreduction studies The bioremoval of Cr (VI) and Cr (III) was studied using 1. Metabolite devoid of cells 2. Cells The bioremoval of Cr (VI) and its subsequent reduction to Cr (III) was monitored 1. During growth studies with unadapted strain 2. During growth studies with the bacterial strain adapted to 2 ppm Cr (VI) The detailed procedures are described elsewhere [10]. 2.6 Biosorption studies Biosorption studies of both Cr (VI) and Cr (III) were studied, using the modified Bromfield medium grown cells, as a function of various parameters such as time, pH, amount of wet biomass and chromium concentration. 2.6.1 Kinetics of biosorption The cell pellet was dispersed in 100 ml of 2 ppm Cr (VI) or Cr(III). Ten such flasks were agitated in a Remi rotary shaker at 240 rpm at 30°C ± 2°C and at regular time intervals centrifugation was done and the supernatant was analysed for the residual chromium content. The final pH was also noted. A blank was maintained without adding the Cr (VI) or Cr(III) solution. Another experiment was performed using the cells, which were adapted to 2 ppm Cr (VI). In this, the adapted cells were dispersed in 100 ml of 2ppm of Cr (VI) and agitated as per the same conditions given above. At regular time intervals, the cells were separated by centrifugation and the supernatant was analysed for Cr (VI) content. All these experiments were carried out at natural pH of the respective solutions, namely 5.75 in the case of 2 ppm Cr (VI) and 5.45 in the case of 2 ppm Cr (III) 2.6.2 Effect of pH on the biosorption process The stability of 2 ppm Cr (VI) and 2 ppm Cr (III) in the pH range from 2 to 7 was studied. Nitric acid and potassium hydroxide were used as pH regulators. Around 0.3 g of the wet biomass was dispersed in 100ml of the solution containing 2 ppm Cr (VI) or Cr (III). Eight such flasks were maintained at different pH values ranging from 2 to 7. After an equilibration period of 48 h in the case of Cr (VI) and 25 h in the case of Cr (III), the solutions were centrifuged and the supernatant was analysed for the residual concentrations of the chromium ions. The final pH values have been plotted. 2.6.3 Effect of wet biomass loading A fully-grown culture was centrifuged and different weights of the biomass ranging from 0.1g to 1.5 g were dispersed in solutions containing the desired amount of Cr ion under consideration. The solutions were adjusted to the optimum pH, corresponding to 761

Biosorption

maximum biosorption of the Cr ion. These values were 2 for Cr (VI) and natural pH (5.45) for Cr (III). The flasks were equilibrated in a Remi rotary shaker at 240 rpm, the temperature being maintained as 30ºC ± 2°C, for the desired time period, namely, 48 h for Cr (VI) and 25 h for Cr (III). The solutions were later centrifuged and the metal ion concentrations were determined using the procedures described earlier. 2.6.4 Biosorption isotherms A known amount of the wet biomass (1.9g) was dispersed in a desired concentration ranging from 2 to 25 ppm of either Cr (III) or Cr (VI). In all these cases the initial pH was adjusted to that of the optimum value namely 2 for the Cr (VI) system and 5.45 for the Cr (III) system. The flasks were equilibrated for their respective time periods (48 h for Cr (VI) and 25 h for Cr (III)), at the end of which the residual concentrations were determined. 3.

RESULTS AND DISCUSSION

3.1 Bioremoval of Cr (VI) during the growth of Bacillus polymyxa (unadapted strain) The decrease in the concentration of Cr (VI) during the growth of the bacteria was monitored. The bioreduction of Cr (VI) to Cr (III) and the biosorption process contribute to the total bioremoval. The bioremoval of 2 ppm Cr (VI) was examined during the growth of Bacillus polymyxa and the results are depicted in Figure 1. In these experiments, both the Cr (VI) and Cr (III) contents were monitored as a function of time. The percentage removal of Cr (VI) steadily increases with increase in time upto 72 h, wherein nearly 90% removal is achieved. Beyond that period no further improvement is observed. The amount Cr (III) obtained by bioreduction, closely parallels the bioremoval trend of Cr (VI). Nearly 65% reduction is effected in 72 h and thereafter a saturation value is attained. The amount of Cr biosorbed marginally increases with time and reaches a maximum value of about 25% in 54 h. A further increase in the growth period does not improve the biosorption efficiency. These results highlight that significant bioremoval of Cr (VI) can be achieved during the growth of Bacillus polymyxa. Many bacteria have been reported to be able to detoxify Cr (VI) by reducing it to Cr (III). The prevention of the adverse effect of Cr (VI) on its growth is achieved by either reduction or accumulation inside the bacterium or adsorption of Cr (VI) on its surface [11 - 13]. The bioremoval of 5 ppm Cr (VI) is shown in Figure 2. It is worthy of observation that the kinetics of this experiment is much slower, when compared to that with 2 ppm Cr (VI). The reason for this could be due to the increase in toxicity. In 24 h, only 15% bioremoval is obtained, mainly contributed by means of reduction to Cr (III). The bioremoval efficiency steadily increases with increase in time upto about 160 h and thereafter attains a saturation value of about 85%. It is noteworthy that the bioreduction efficiency is about 73% during this period. The contribution due to biosorption is only about 10% with respect to Cr uptake. This trend is similar to that observed with 2 ppm Cr (VI) (Figure 1).

762

Biosorption

Figure 1. Bioremoval of 2 ppm Cr (VI) during the growth of unadapted Bacillus polymyxa

Figure 2. Bioremoval of 5 ppm Cr (VI) during the growth of unadapted Bacillus polymyxa

3.2 Bioremoval studies using cells adapted to Cr (VI) Figure 3 depicts the percentage bioremoval of 2 ppm Cr (VI) using the Bacillus polymyxa strain, which has been adapted to 2 ppm Cr (VI) concentration. The percentage bioremoval steeply increases with increase in time upto about 24 h and thereafter a slight decrease in the rate is observed. The percentage bioremoval is 62% in 24 h while the bioreduction is about 60%. The contribution of biosorption is only about 2% in 24 h and marginally increases to 8% between 48 h and 72 h. In 48 h complete bioremoval of Cr (VI) is effected, of which 92% is bioreduced to Cr (III) and 8% is biosorbed. A comparison of Figures 1 and 3 reveals that the adapted cell is more efficient for the bioremoval process. The time taken for 90% bioremoval is 72 h in the case of the unadapted strain, whereas the adapted one takes only around 48 h to achieve comparable results. Moreover, it is also evident that the adapted strain is more efficient in bringing about the reduction of Cr (VI) to Cr (III). For example, the percentage bioreduction with the adapted strain is about 1.5 times better than the unadapted one. It is also pertinent to observe that the biosorption efficiency of the adapted strain is much lower than that of the unadapted one. It is thus apparent that the bioreduction mechanism is a major contributing factor to the overall bioremoval process. Figure 4 shows the bioremoval of 5 ppm Cr (VI) as a function of time, using a strain initially adapted to 2 ppm Cr (VI). The percentage bioremoval increases with increase in time upto about 165 h and thereafter attains a saturation value. The maximum bioremoval obtained is around 90%. A similar trend is observed with respect to the bioreduction to Cr (III) and 70% reduction is achieved in about 165 h. On the other hand, only about 20% biosorption is achieved during this time period. As observed in the case of unadapted strain (Figure 2), bioreduction contributes a significant extent to the bioremoval process. Consequent to adaptation, there is a marginal improvement in the percentage bioremoval, though the percentage bioreduction achieved is very similar with and without adaptation. On the contrary, there is an improvement in the percentage biosorption from 10% to 20% after adaptation (Figures 2 and 4).

763

Biosorption

Figure 3. Bioremoval of 2 ppm Cr (VI) during the growth of Bacillus polymyxa adapted to 2 ppm Cr (VI)

Figure 4. Bioremoval of 5 ppm Cr (VI) during the growth of Bacillus polymyxa adapted to 2 ppm Cr (VI)

3.3 Bioremoval studies in the presence of metabolite The bioremoval of Cr (VI) was next investigated using the metabolite devoid of cells. From Figure 5 it is evident that there is a steep rise in the bioremoval efficiency upto 10 h and thereafter there is only a marginal improvement, for the two concentrations studied. On a comparative basis, the bioremoval efficiency is marginally better when a lower concentration of 2 ppm Cr (VI) is used. Over 80% bioremoval is achieved in about 10 h using 2 ppm Cr(VI), while it takes almost 48 h for a similar amount of removal to be effected using 5 ppm Cr(VI). Figure 6 depicts the bioremoval of Cr (VI) using the metabolite derived during the growth of the strain adapted to 2 ppm Cr (VI). The bioremoval efficiency is better than in the previous case (Figure 5) and complete removal is possible in 24 h. With respect to 5 ppm, nearly 36 h are required to bring about 90% removal. Metabolites contain organic acids, proteins and the exo-polysaccharides secreted by Bacillus polymyxa during its growth.

Figure 5. Bioremoval of Cr (VI) Figure 6. Bioremoval of Cr (VI) using using metabolite of unadapted metabolite of Bacillus polymyxa Bacillus polymyxa adapted to 2 ppm Cr (VI) The structure of the polysaccharides and the proteins are very complex and they are capable of binding to many metal ions. The formation of a complex of Cr (VI) with the 764

Biosorption

metabolic products can facilitate the removal of Cr (VI). These results corroborate the results of the earlier studies, wherein adaptation enhances the bioremoval efficiency. 3.4 Bioremoval studies on Cr (VI) and Cr (III) using cells 3.4.1 Kinetics of bioremoval The percentage biosorption of Cr (VI) and Cr (III) as a function of time is shown in Figure 7. In this experiment, the initial concentration was 2 ppm for both Cr (VI) and Cr (III). The pH of the system for Cr (VI) and Cr (III) was 5.7 and 5.5 respectively. The amount of wet biomass was around 0.3 g. It is evident that the kinetics of the biosorption process is very slow and the maximum amount of Cr (VI) biosorbed is only about 18% after 48 h. A further increase in the biosorption time upto 100 h does not improve the uptake. Hence the time of equilibration was fixed at 48 h in all further experiments using Cr (VI). It is evident that the amount of Cr (III) biosorbed steeply increases with increase in time upto 24 h and thereafter attains a saturation value of about 75%. It is noteworthy that there is a preferential biosorption of Cr (III) by the cells vis-à-vis Cr (VI). The equilibration time in all the subsequent experiments using Cr (III), was fixed at 24 h. Such slow kinetics can be associated with the intracellular uptake of Cr (VI). There are two possible reasons for the low uptake: 1. K2Cr2O7 around natural pH exists in the form of HCrO4- [14]. The bacterial cell wall is negatively charged and hence interaction with Cr (VI) is not favorable due to electrostatic forces of repulsion. 2. Cr (VI) is mutagenic and carcinogenic [15-16]. The bacteria are capable of sensing the toxicity of Cr (VI) and hence accumulate very less quantity of the metal. Reduction is possible only around very low pH values in the range of 2-3, and at natural pH, the bacteria are incapable of reduction, to overcome the adverse effect of the hexavalent salt. 3.4.2 Effect of pH The cells obtained by centrifuging were washed and later dispersed in solutions of 2 ppm Cr (VI) or Cr (III). The pH values range from 1.5 to 7 in the case of Cr (VI) and pH 2-6 for Cr (III). The time of equilibration was fixed at 48 h for Cr (VI) and 24 h for Cr (III), and the final pH values were noted. The percentage bioremoval of Cr (VI) as a function of pH is depicted in Figure 8. The Cr (VI) bioremoval shows a marginal increase from pH 1.5 to 2 and thereafter continuously decreases with increase of pH. The maximum bioremoval of about 75% is observed at pH 2, while at pH 7 it is significantly reduced to 15%. No precipitate of Cr (VI) was observed in the pH range studied. Cr (VI) is a strong oxidising agent and exists as monohydrogen chromate (HCrO4-) and chromate (CrO4-2) ion in solution and has a high solubility in water [17]. The following observations are noteworthy: 1. A variation in the pH values was observed between the initial and final readings. The shift was marginal in the pH range of 2 – 3 but was prominent at other pH values. 2. The bioreduction of Cr (VI) to Cr (III) is more pronounced in the acidic pH range and is significantly reduced at pH 7. 3. The percentage biosorption of Cr (VI) is about 22% and increases to about 30% in the pH range 2–3 and thereafter further increases, and at pH 7, it is enhanced to about 45%. 765

Biosorption

4.

A comparison of the bioreduction and biosorption processes reveals that bioreduction contributes to a greater extent to the overall bioremoval below pH 3, while biosorption is more dominant beyond that pH. 5. The decrease in the percentage bioremoval and biosorption below pH 2 may be attributed to the production of Cr (III) and also the effect of the hydronium ions, which will then compete for the binding sites. The effect of pH on the biosorption of Cr (III) is also shown in Figure 8. It is evident that the maximum uptake occurs at the natural pH (~6) of the system. The drop in the Cr (III) uptake at lower pH values can be attributed to the competition of H+ for the binding sites. As can be seen from the figure, there is hardly any significant uptake of Cr (III) upto pH 4. As the pH is increased, more negative sites are exposed and hence the biosorption increases. This can be visualized in the pH region of 4 to 6, wherein the biosorption increases from 12% to about 80%. After the maximum value, precipitation was observed beyond pH 6 and hence no experiment was conducted beyond pH 6. A comparison of the bioremoval results of Cr (VI) and Cr (III) reveals that Cr (VI) uptake is much higher than that of Cr (III) below pH 5. On the contrary, above pH 5, the trend is reversed with Cr (III) showing higher uptake. It is interesting to observe that Cr (VI) biosorption shows a maximum of about 75% at pH 2, while Cr (III) exhibits a maximum (79%) at pH 6. % of Cr (III) biosorbed) % of Cr (VI) biosorbed)

Total bioremoval of Cr (VI) Total bioremoval of Cr (III) Bioreduction of Cr (VI) Biosorption of Cr (VI)

100

80 80

% Bioremoval

% Biosorption

100

60 40

60

40

20 20

0

0

20

40

60

80

100

Time (h)

Figure 7. Kinetics of bioremoval of Cr (III) and Cr (VI)

0 1

2

3

4

5

6

7

8

pH

Figure 8. Effect of pH on the bioremoval of Cr (III) and Cr (VI)

3.4.3 Effect of biomass loading on Cr (VI) uptake The effect of biomass loading on the bioremoval of Cr (VI) is shown in Figure 9a. In these experiments, the pH was fixed at 2, initial Cr (VI) concentration at 2 ppm and the equilibration time at 48 h. In the same figure, the specific uptake expressed as mg chromium/g biosorbent is also portrayed. It is readily evident that the percentage bioremoval steadily increases with increase in the biomass loading upto 0.5 g and thereafter there is only a marginal increase. Over 90% Cr (VI) bioremoval is achieved with 0.5 g, while at 0.92 g almost 100% bioremoval takes place. As can be expected, the specific uptake of Cr (VI) continuously decreases with increase in the amount of biomass. For example, the specific uptake is the highest at 0.03 g biomass, namely 12 mg/g and steeply decreases and is negligible with 0.92 g biomass loading. The trends observed are in good agreement with those reported earlier [18-20].

766

Biosorption

Figure 9a. Effect of biomass loading on Figure 9b. Effect of biomass Cr (VI) uptake loading on bioremoval of Cr (VI) The contributions of the bioreduction and biosorption components to the bioremoval process have also been determined and these results are depicted in Figure 9b. The experimental conditions are the same as those pertaining to Figure 9a. The percentage bioreduction is found to increase with increase in the biomass loading. The percentage bioreduction is about 45% with 0.5 g biomass, while almost 100% bioreduction is achieved with 0.92g biomass. It may be recalled that the results of the growth studies carried out in the presence of Cr (VI) also revealed the significant contribution of bioreduction to the bioremoval process (Figure 1). The percentage biosorption increases with increase in the biomass loading upto 0.5 g and subsequently decreases (Figure 9b). It is thus apparent that the biosorption process is more effective at lower biomass loading, whereas the bioreduction process is more efficient when the amount of biomass is increased. The decrease in the biosorption capacity at higher biomass loading can be attributed to the "screening effect", wherein cell−cell interaction inhibits the metal uptake from the solution. Similar results have been reported by other workers [14, 18]. The effect of increasing the amount of the wet biomass on the uptake of Cr (III) with other conditions like the pH, time etc, being maintained as outlined earlier, is depicted in Figure 10. The specific uptake capacity of Cr (III) is also shown in this figure. The percentage uptake steeply increases to about 85% with increase in the amount of wet biomass upto 0.4 g and thereafter there is only marginal improvement to about 90% at 0.95 g biomass. The specific uptake continuously decreases with the increase in the amount of biomass, from about 5.25 mg Cr (III) per gram at 0.9 g biomass to about 1.8 mg/g at 0.85 g biomass. As explained earlier with respect to biosorption of Cr (VI), a higher amount of the biomass results in the "screening effect" and consequently Cr (III) uptake is diminished. 3.5 Biosorption isotherms

The biosorption isotherm for Cr (VI) at pH 2, carried out at a temperature of 28°C ± 3°C is shown in Figure11. In this experiment, the wet weight of biomass (substrate) used was 1 g and the equilibration time was fixed at 48 h. The adsorption density steadily increases with increase in equilibrium concentration upto about 7 ppm and thereafter attains a saturation value. The biosorption isotherm of Cr (III) is also depicted in Figure 11. The amount adsorbed steeply increases with increase in the equilibrium concentration of Cr (III) upto 15 ppm and subsequently, there is a decrease in the slope of the isotherm, tending to saturation coverage. Both the isotherms exhibit Langmuirian behaviour and resemble the type L2 of the Giles classification [21]. A comparison of the biosorption 767

Biosorption

isotherms of Cr (VI) and Cr (III) with respect to the bacterial cells highlights that saturation coverage is attained at a much lower concentration for Cr (VI) vis-à-vis Cr (III). This is understandable as Cr (VI) is toxic, while Cr (III) is not.

30

Cr (III) Cr (VI)

mg metal/g biosorbent

25

20

15

10

5

0 0

10

20

30

40

50

60

Equilibrium conc.

Figure 10. Effect of biomass loading on Cr (III) uptake

Figure 11. Biosorption isotherms for Cr (VI) and Cr (III)

3.6 Biosorption mechanisms Biosorption mechanisms involving living cells include both metabolism dependent and metabolism independent processes. Further, there is a potential for biologically altering the valency state of the metal through the bioreduction mechanism and in some cases biodegradation of organometallic complexes has been achieved. Clearly, metabolism independent metal binding to cell walls and external surfaces is the only mechanism present in the case of non-living biomass. Metabolism independent uptake essentially involves adsorption processes such as ionic, chemical and physical adsorption. A variety of ligands located on the fungal walls are known to be involved in metal chelation [22]. These include carboxyl, amine, hydroxyl, phosphate and sulfhydryl groups. Metal ions could be adsorbed by complexing with negatively charged reaction sites on the cell surfaces [23]. The relative importance of each functional group is often difficult to resolve. The microbial cell walls are rich in polysaccharides and glycoproteins such as glucans, chitin, chitosan, mannana and phosphomannans. These polymers are abundant sources of the above-mentioned metal binding ligands. The FTIR spectral results confirmed the presence of functional groups such as NH, NH2, CONH, OH, CO, PO, POC, CH2, CH3, and COO- groups [10]. It has been reported that different types of functional groups present on the cell wall such as carboxyl, amino, phosphate, hydroxyl are implicated in metal binding [24]. Cr ions could be adsorbed by complexing with the negatively charged reaction sites on the cell surfaces. Similar findings have been reported by other workers [23]. Furthermore, the chromium uptake is compounded by the complex solution chemistry of Cr (III) and Cr (VI) as a function of pH 4.

1.

768

CONCLUSIONS From the results of the present investigation, the following conclusions can be drawn: Both bioreduction and biosorption contribute towards the bioremoval of Cr (VI) during the growth of Bacillus polymyxa. Nearly 90% of 2ppm Cr (VI) is removed in 72 h, 65% by reduction to Cr (III) and 25% by biosorption.

Biosorption

2. 3.

4. 5.

Bioremoval studies using an adapted strain show a greater efficiency namely 90% bioremoval is effected in 48 h. Complete bioremoval of Cr (VI) was achieved at pH 2 using bacterial cells under optimum conditions. The bioreduction of Cr (VI) to Cr (III) was observed in the pH range of 1.5 to 4. About 90% uptake of Cr (III) could be obtained at a natural pH of 5.5. The biosorption isotherms for both Cr (III) and Cr (VI) follow Langmuirian behaviour.

REFERENCES

1. T.L. Marsh and M.J. Mc Inerney, Appl. Environ. Microbiol. 67 (2001) 1517. 2. T. Ramasami et al., Chem. Week., (1995) 155. 3. J.V. Bhide, P.K. Dhakephalkar and K.M. Paknikar, Biotechnol. Lett., 18 (1996) 667. 4. H. Ohtake and J.K. Hardoyo, Water Sci. Tech., 25 (1992) 395. 5. V.I. Romanenko and V.N. Korenkov, Mikrobiologiya, 46 (1977) 414. 6. P.C. DeLeo and H.L. Ehrlich, Appl. Microbiol. Biotechnol., 40 (1994) 756. 7. C. Cervantes, Antonie van Leeuwenhoek, 59 (1991) 229. 8. P. Anand, J.M. Modak and K.A. Natarajan, Int. J. Miner. Process., 48 (1996) 51. 9. Methods Manual, Thermo Jarrell Ash Corporation, U.S.A. 10. T. Hemamalini, Master of Science (Engg.) Thesis, Indian Institute of Science, Bangalore, July 2001. 11. P. Wang, T. Mori, K. Komori, M. Sasatsu, K. Toda and H. Ohtake, Appl. Environ. Microbiol., 55 (1989) 1665. 12. A.R. Shakoori, M. Makhdoom and R.U. Haq, Appl. Microbiol. Biotechnol, 53 (2000) 348. 13. J.W. Williams and S. Silver, Enzyme Microb. Technol., 6 (1984) 530 14. D. Kratochvil, P. Pimentel and B. Volesky, Environ. Sci. Technol., 32 (1998) 2693. 15. B.N. AMES, SCIENCE, 191 (1976) 241. 16. S. Venitt and L.S. Levy, Nature, 250 (1974) 493. 17. W. Stumm and J.J. Morgan, Aquatic Chemistry, 3rd Edition, Wiley Interscience Publication, New York (1996). 18. K. Chandra Sekhar, S. Subramanian, J.M. Modak and K.A. Natarajan, Int. J. Miner. Process., 53 (1998) 107. 19. J.M. Modak, K.A. Natarajan and B. Saha, Miner. Metall. Process.,13 (1996) 52. 20. M. Balakrishnan, J.M. Modak, K.A. Natarajan and J.S. Gururaj Naik, Miner. Metall. Process., 11 (1994) 197. 21. C.H. Giles, T.H. MacEwan, S.N. Nakhwa and D. Smith, J. Chem. Soc., (1960) 3973. 22. J.M. Modak and K.A. Natarajan, Miner. Metall. Process., 12 (1995) 189. 23. T.J. Beveridge and R.G.E. Murray, J. Bacteriol., 127 (1976) 1502. 24. J.M. Tobin, D.G. Cooper and R.J. Neufeld, Appl. Environ. Microbiol., 47 (1984) 821

769

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption and bioaccumulation of heavy metals by bacteria isolated from contaminated sites of Karachi, Pakistan Nuzhat Ahmed, Uzma Badar, Fouad M. Qureshi and Fehmida Fasim Centre for Molecular Genetics, University of Karachi, Karachi-75270, Pakistan Abstract Resistance to toxic heavy metals, their accumulation and biosorption by bacteria is a wide spread phenomenon that could be exploited for the improvement of the environment. This study describes investigation of metal resistant bacteria isolated from metal contaminated industrial sites in Karachi, Pakistan. Biosorption and bioaccumulation were studied with reference to the cadmium, copper, and chromium. The isolates were identified by 16s rRNA gene analysis and by API kits. The Pseudomonas species were more prevalent, showing multiple metal resistances to copper, chromium, cadmium, nickel, zinc and cobalt salts. Maximum accumulation and biosorption of cadmium, copper and chromium was found in CMG64, CMG462, CMG463, and CMG480 respectively. The biosorption and bioaccumulation were periodically monitored and the metal concentrations were estimated by atomic absorption spectrophotometer (AAS) and by enzyme assays. The localizations or deposition of heavy metal inside/outside of the cell surface were further confirmed by transmission electron microscopy (TEM) and by energy dispersive x-ray analysis (EDX). These microbes are good candidates for bioremediation purposes. 1.

INTRODUCTION Prevalence of heavy metals in effluent is a major cause of environmental damages. The most prevalent ones include barium, cadmium, chromium, copper, iron, lead, manganese, nickel and zinc. In bacteria toxic heavy metal ion resistance systems have been reviewed over the last decade (1-5). The resistance mechanisms against all these heavy metals are highly specific. There is no general mechanism for resistance to all heavy metals. Microorganisms can physically remove heavy metals from solution through either bioaccumulation or biosorption (6-8). Bioaccumulation plays an important role in the detoxification of hazardous heavy metals. The uptake of metal ions onto the cell surfaces and their subsequent translocation into the cell are well known natural processes but are highly specific (9). Microbial cells can intracellularly and extracellularly accumulate both essential and non-essential metals such as chromium, cadmium, copper, nickel, lead, iron, germanium, silver and zinc. Several species of bacteria have been reported for the accumulation/uptake of various metal e.g., Citrobacter species accumulated cadmium and uranium (10-12), Pseudomonas syringae accumulated copper (13), Pseudomonas stutzeri accumulated germanium (14) etc. Biosorption does not consume cellular energy. Positively charged metal ions are sequestered primarily through the adsorption of metals to the negative ionic groups on cell 771

Biosorption

surfaces (8), such as the polysaccharide coating found on most forms of bacteria, or other extracellular structures such as capsules or slime. Metallic cations are attracted to the negatively charged sites at the surface of the cell (15). Anionic ligands such as phosphoryl, carboxyl, sulfhydryl, and hydroxyl group of membrane proteins also involve in metal binding to the cell surface (16). These processes are applied to clean the effluents, contaminated ground waters and soils. For the development of this technology microorganism especially bacteria are of great importance. They have the ability to reduce the toxicity of metals; this ability of bacteria can be harnessed in biotechnological applications for the removal/control of excess metal in various environments such as industrial and other wastes. 2.

MATERIALS AND METHODS

2.1 Bacterial strains and growth media CMG64, CMG462, CMG463 and CMG480, local isolates, were used in this study. Nutrient broth (Oxoid) was used as a starter medium. Maximum tolerable concentrations of various heavy metals were estimated in tris-minimal medium (5). Reduction experiment was performed in acetate minimal medium (AMM) as described earlier (8). 2.2 Identification of bacterial strains The isolates (CMG462, CMG463 and CMG480) were identified using partial 16S rRNA gene analysis. A small colony of each was resuspended in 0.1 ml of sterile, deionised water (SDW), mixed and heated at 70°C (10 min) for cell lysis. Crude lysate (0.2 µl) was added to SDW (0.0198 ml) and used as a template in a polymerase chain reaction using the eubacterial 16S targeted PCR primers (pA and pH') as designed by Edwards (17). These are known to amplify a 1,536 base pair (approx. 1.5 kb) length of 16S rDNA. The reaction mixture for amplification was as published by Bruce et al, (1992). For PCR MJ thermalcycler (MJ Research Inc., USA) was used under tube temperature control and using 30 cycles of the following program: 94°C for 40 sec, 55°C for 1 min, 72°C for 2 min and a final 10 min extension at 72°C. PCR products were cleaned using Sephacryl S400 columns (Pharmacia, Sweden), and partially sequenced using 16S sequencing primer 943 reverse (Lane et al, 1985) by Alta Biosciences (University of Birmingham, UK). Sequences (up to 650 bp) were analysed using ADVANCED BLAST software to access the EMBL database, Heidelberg, Germany (Web ref: HTTP://www.ncbi.nlm.nih.gov/cgibin/Blast). Netscape browser interface was used. The isolate CMG64 was identified by API- kit. 2.3 Maximum tolerable concentration of heavy metal salts To determine the maximum tolerable concentration (MTC) of heavy metal salts such as CuSO4, NiCl2, Pb(CH3OO)2, ZnSO4, Cr2O7, CoCl2, and CdCl2, bacterial culture were streaked on tris mineral medium plates supplemented with variable salt concentrations. The plates were then incubated at 37°C and growth was observed after 24-48 h. 2.4 Accumulation of heavy metals The accumulation of heavy metals was assessed by growing bacterial cultures in tris minimal broth. The 50ml tris minimal broth in 250ml flask supplemented with variable concentrations of metal salts such as cadmium, copper and chromate were inoculated with 1 ml of overnight grown culture and incubated at 37°C in shaker incubator (100rpm). Each 772

Biosorption

day samples were collected for estimation of copper using copper assay and total cell protein content by protein assay using protein test kit (Sigma: TPRO 562). 2.5 Copper and chromate assay Copper was assayed by a method described by Macaskie (18) and modified by Qureshi et al (19) to increase the sensitivity of a 1 ml reaction by 20 folds. For determination of Cr(VI) the mixture contained 400 µl of 20 mM MOPS-NaOH buffer (pH7), 327 µl distilled water, 33 µl of 3M H2SO4, 40 µl of 0.25% diphenyl carbazide (DPC) solution and culture supernatant or standard solution (200 µl). The absorbance was determined immediately at 540 nm (1.5 ml cuvettes, 1 cm path length). DPC solution comprised 0.25% wt/vol diphenylcarbazide in 0.1M H2SO4 in AnalaR acetone (20). 2.6 Transmission Electron Microscopy and Energy Dispersive X-ray Analysis Bacterial pellets were harvested after 48 h, washed with distilled water and fixed by immediate resuspension in 2.5% vol/vol glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.2, 60 min). The cells were dehydrated in an ethanol series (70, 90, 100, 100, and 100% ethanol: 15 min each), twice with propylene oxide (15min) and then in a mixture of propylene oxide/epoxy resin (1:1; 45min). Samples were then embedded in epoxy resin under vacuum in plastic moulds (20 min) and left to polymerize at atmospheric pressure (24 h, 60°C). Sections (70 nm) were cut, collected on copper grids/aluminium grids and examined by transmission electron microscopy. For energy dispersive X-ray analysis (EDX) thicker sections (200-300 nm) were cut and examined by scanning transmission electron microscopy (JEOL JEM-100CXII) using a LINK ISIS X-ray analyzer to determine elemental distribution in/on and around the cells. 3.

RESULTS AND DISCUSSIONS Three isolates were homologous to strains of Pseudomonas (Table 1) CMG462 and CMG463 were identified as P. stutzeri. These pseudomonads were 99% similar to the matching sequences, while CMG64 was identified as Pseudomonas aeruginosa. The fourth one a tannery isolate gave a high homology to an isolate that was identified as "cucurbit yellow vine disease bacterium", an enterobacterial plant pathogen. This isolate was 97% similar to reference sequence on the EMBL database (Table 1). Table 1. Source of bacterial isolates Strain Code

Bacteria

Source

CMG462

Pseudomonas stutzeri

(Foundry Soil, Karachi Shipyard and Engineering works)

CMG463

Pseudomonas stutzeri

(Foundry Soil, Karachi Shipyard and Engineering work)

CMG64

Pseudomonas aeruginosa

Korangi Industrial Area, Sector 7-A, Karachi

CMG480

Cucurbit yellow vine disease bacterium

Korangi Tannery Air, Sector 7-A, Karachi

The metal resistance were studied in tris-based medium because the complexation with heavy metals is minimum therefore the shown metal concentration is approximately the free metal concentration (5).

773

Biosorption

All the isolates of this study were originated from various metal contaminated sites of Karachi, Pakistan and showed multiple metal resistances (Table 2). The isolates CMG462, CMG463 and CMG64 showed resistance against cadmium chloride up to 2 mM whereas CMG462 and CMG463 exhibited highest resistance against copper i.e., 8 mM and 10 mM respectively with respect to other tested heavy metal salts. The multiple metal resistance ability suggested the prior exposure of the isolates to these metals, which are present in the sampling sites; this phenomenon of multiple metal resistances has been reported by many workers (21-22). One of the potential metal resistance mechanisms is accumulation or uptake of metal by bacterial cell. Metal accumulation of cadmium by CMG64 and copper by CMG462 and CMG463 was studied. Results (Table 3) show 40% of cadmium accumulation by CMG64, whereas CMG462 and CMG463 accumulated copper 90.7% and 97.4% respectively. Accumulation might be due to the presence of extracellular components such as proteins or polysaccharide. Falla and Bloch (23) have reported that polysaccharide-producing strains are active metal accumulators. It is reported that the resistant strain have well developed mechanisms to prevent the shock, such as in Pseudomonas which accumulates metal in the periplasmic space that prevents the entrance of an excess amount of metal into the cytoplasm (24). Table 2. Maximum tolerable concentrations (MTC) of different metals Strain Code

Metal concentration (mM) ZnCl2 Cr2O4

CdCl2

CuSO4

CMG462

2

8

0

CMG463

2

10

CMG64

2

CMG480

0.5

CoCl2

NiCl2

1.5

0.5

0.5

0.5

1.5

0.5

0.5

1.5

2.5

-

2

2

0.5

0.5

1.0

-

0

Analysis of cell sections using transmission electron microscopy with energy dispersive X-ray analysis showed metal accumulation in these strains. Intracellular accumulation of cadmium in CMG64 was confirmed by TEM, which showed dark precipitates when cells were grown in the presence of 0.1 mM cadmium while no precipitates were observed in absence of cadmium. CMG462 and CMG463 showed accumulation/removal of copper from media while growing at 1 mM CuSO4. The concentration of copper removal was estimated by copper assay and calculated by copper standard curve. The intracellular/extracellular accumulation or the localization of copper was observed under electron microscope. The localization of copper was determined in stained and unstained cells (Figure 3). In CMG462 it was observed that the dark patches in side the cells i.e., in the cytoplasm and the darkly stained outer membrane was seen in the cells grown in presence of copper (Figure 3A, 3C). The EDX analysis also revealed the presence of copper inside the cell as well as at cell edges or the outer cell membrane (Figure 4). In contrast with CMG463, the copper was bound with the cells extracellularly which was clearly observed in unstained cells (Figure 3E). In EDX analysis the copper at the outer surface was extremely low or undetectable while it was detectable intracellularly (Figure 4). Accumulation of copper has been observed in several species of Pseudomonas as well as copper resistant Rhizobium loti (12, 25).

774

Biosorption

Table 3. Accumulation of metals in tris-minimal medium Strain Code CMG64 CMG462 CMG463

Metal salt CdCl2 CuSO4 CuSO4

Conc. (mM) 0.1 1.0 1.0

Metal Accumulation (%) 40.0 90.7 97.4

CMG462 and CMG463 P. stutzeri strains, were also evaluated for their ability to reduce and remove Cr(VI). Since Cr-stress can lead to resistance which may be associated with cellular exclusion of Cr(VI).The potential of the CMG462 and CMG463 strains were evaluated in parallel with another strain i.e., CMG480 which had been isolated from Korangi tannery air environment. All strains grew aerobically in 100 µM Cr(VI)supplemented minimal medium, the doubling time was generally 4-6 h but reduced little Cr(VI). These strains grew aerobically in 100 µM Cr(VI)-supplemented acetate medium the doubling time was generally 4-6 h but reduced little Cr(VI). Anaerobic growth at the expense of Cr(VI) as the electron acceptor was negligible, similar phenomenon has been reported earlier (26-27) but Cr(VI) was reduced and removed efficiently. These cultures were essentially resting cell suspensions. This was observed only after a lag of 31 h. Cr(VI) reduction aerobically was low in accordance with the provision of O2 as the primary electron acceptor and electron "sink". CMG462, CMG463 and CMG480 showed a comparable loss of Cr(VI) after 192 h. In contrast Cr(VI) reduction was apparent in these strains anaerobically; the best strains were CMG463 and CMG480 in terms of total Cr(VI) removed (88% and 76.08%, respectively) (Table 4). Table 4. Aerobic and anaerobic reduction rate observed in acetate medium Cell Incubation Aerobically

Anaerobically

Strain Code CMG462 CMG463 CMG480 CMG462 CMG463 CMG480

Residual Cr(VI) (µM) 87 + 1 85 + 1 91.18 + 4.86 42 + 3 12 + 3 23.92 + 7.18

Loss of Cr(VI) (µM) 14 15 8.82 58 88 76.08

Rate of reduction (nmol/mg protein/h) 0.1 0.4 0.13 3.0 2.6 4.29

This study suggested that there are two mechanisms of chromate resistance and biosorption. P. stutzeri CMG462 showed some electron opaque material at the outer cell surface but the localized concentration of this was below the sensitivity of the EDX technique. This suggested that the Cr trivalent precipitates dispersed in the medium more as compared with the cell surface adsorption. But these precipitates uniformly adsorbed at the cell surface (Figure 1), whereas in CMG480, the extracellular dark precipitates were attached with the cell surfaces when grown in presence of chromate (Figure 1E). It is suggested that bacteria have excellent nucleation sites for fine-grained mineral formation due to their high surface area to volume ratio (28) and the presence of electronegative surface functional groups for e.g., carboxyl group, phosphoryl and hydroxyl groups, the electron opaque material (Figure 1C) deposited by CMG480, identified as containing Cr and P by EDX. Similar results were found with B. pumulis strain which deposits Cr(VI) extracellularly (8). CMG463 showed occasional intracellular deposits of electron opaque material which gave a positive result for Cr and also for P by EDX. Analysis of the cell surface and background resin around the cells gave no detectable Cr (Figure 2). Detection of intracellular Cr suggests that anionic CrO42- or its reduced species Cr+3 enter the cell but for a waste remediation process a cell surface-localized deposit is preferable because it 775

Biosorption

may be possible to remove this and conserve the biomass for re-use. These results do suggest that these microbes are good candidates for exploitation of decontamination of cadmium, chromium and copper contaminated sites.

Figure 1. Transmission electron micrographs (TEM). Unstrained cells of CMG462 (A, B), CMG463 (C, D) and CMG480 (E, F) Bacterial strains grown in presence (A, C, E) and in absence (B, D, F) of Cr(VI) in medium

776

Biosorption

A

B

C

D

Figure 2. Energy disperssive X-ray analysis (EDX) of electron opaque areas. CMG462 (A), CMG463 (B), CMG480 (C) and background control (D)

777

Biosorption

Figure 3. Transmission electron micrographs (TEM) of bacterial strains grown in presence (A, C, E, G) and absence of copper (B, D, F, H) A, B, E, F: unstrained cells; C, D, G, H: stained cells; A, B, C, D: CMG462; E, F, G, H: CMG463

778

Biosorption

C

779

Biosorption

D

E

Figure 4. Energy Depressive X-ray Analysis (EDX) in presence of 1 mM CuSO4 of electron opaque material of CMG462 (A, B) and CMG463 (C, D); analysis inside the cells (A, C); analysis at cell surface (B, D) and control without metal (E) REFΕRENCES

1. N. L. Brown, Trends Biochem Sci., 10 (1985) 400. 2. T. J. Foster, CRC. Crit. Rev. Microbiol., 15 (1987) 117. 3. T. K. Misra, N. L. Brown, D. C. Fritzinger, R. D. Pridmore, W. M. Barnes, L. Habestroh and S. Silver, Proc. Natl. Acad. Sci. USA. 81 (1984) 5975. 4. S. Levy, R. Clowes and E. Koenig (eds), Molecular Biology, Pathogenicity, and Ecology of bacterial plasmids. Plenum New York. 1981. 5. M. Mergaey, D. Nies, H. G. Schlegel, J. Geriys, P. Charles, and F. Van Gijsegem, J. Bacteriol., 162 (1985) 328. 6. Gadd, G. M. Current Opinion in Biotechnology, 11 (2000) 271. 7. L. Xaing, L. C. Chan., and J. W. C. Wong. Chemosphere, 41 (2000) 283. 780

Biosorption

8. U. Badar, N. Ahmed, A. J. Beswick, P. Pattanapipitapaisal and L. E. Macaskie. Biotech. Lett., 22 (2000) 829. 9. M. N. Hughes and R.K. Poole. Metals and Microorganisms. Champan and Hall Ltd., London, New York, 1989. 10. L. E. Macaskie L. E. and A. C. R. Dean, J. Gen. Microbiol., 130 (1984) 53. 11. O. M. Neijessel, R. R. Vander Meer and K. Cu. A. A. Luyben (eds.), Proc. 4th Int. Conf. On Biotechnol., vol. 2. Elsevier, Amsterdam, 1987. 12. L. E. Macaskie, J. S. Blackmore and R. M. Empson, FEMS Microbiol. Lett., 55 (1988) 157. 13. D. A. Cooksey, and H. R. Azad, App. Environ. Microbiol. 58 (1992) 274. 14. M. I. V. Dyke, W. Parker, H. Lee, and J. T. Trevors, Appl. Microbiol. Biotechnol., 33:(1990) 716. 15. T. J. Beveridge, and R. G. E. Murray, J. Bacteriol., 141(1980) 876. 16. B. Volesky, (ed.), Biosorption of Heavy Metals. CRC press, F.L. Boca Raton. 1990. 17. U. Edwards, T. Rogall, H. Blocker, M. Emde, and E. C. Bottger, Nucleic Acids Res., 17 (1989) 7843. 18. L. E. Macaskie, Microbios., 84 (1995) 137. 19. F. M. Qureshi, U. Badar, and N. Ahmed, Appl. Environ. Microbiol., 67 (2001) 4390. 20. P. F. Urone, Anal. Chem. 27 (1955) 354. 21. G. S. Omenn and A. Hollaneder (eds.), Genetic Control of Environmental Pollutants. Plenum Publishing Corp. New York, 1984. 22. R. Qasim, S. N. Husnain, M. Ishaq, and A. Azhar (eds), Recent Advances in Biochemical Research in Pakistan, Karachi University Press, Karachi, 1994. 23. J. Falla and J.C. Baloch, FEMS Microbiol. Lett., 108 (1993) 347. 24. D. A. Cooksey, Appl. Environ. Microbiol., 56 (1990) 13. 25. N. Ahmed, F. M. Qureshi and O. Y. Khan (eds.), Industrial and Environmental Biotechnology, Horizon Scientific Press, Norfolk. U.K, 2001. 26. D. R. Lovely and J. D. Coats, Curr. Opin. Biotechnol., 8 (1997) 285. 27. H. Shen and Y. T. Wang, Appl. Environ. Microbiol., 59 (1993) 3771. 28. T. J. Beveridge, Can. J. Microbiol., 34 (1988) 363.

781

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption equilibria with Spirogyra insignis E. Romera, P. Fraguela, A. Ballester, M.L. Blázquez, J.A. Muñoz and F. González Dpto. de Ciencia de los Materiales e Ingeniería Metalúrgica, Facultad de Ciencias Químicas, Universidad Complutense. Ciudad Universitaria. E-28040 Madrid, Spain Tel.: 34 91 3944339; Fax: 34 91 3944357; E-mail: [email protected] Abstract The recovery of heavy metals from different dissolutions by algae has successfully been demonstrated. We have studied the biosorption process of cadmium, nickel, zinc, copper and lead using the filamentous green alga Spirogyra insignis as biomass. The most favourable sorption conditions using different pH and biomass concentration were set for monometallic solutions. This allowed to determine the sorption isotherms, as for monometallic as bimetallic systems. Both the sorption maximum capacity of the biomass and the equilibrium constant of the reaction between the biomass and the metal were determined by Lagmuir’s model. Finally, the biosorption process was simulated with a computer program, checking these results with those obtained experimentally.

Keywords: biosorption, algae, bimetallic systems, cadmium, nickel, zinc, copper, Spirogyra insignis 1.

INTRODUCTION Biosorption is a novel technique that uses dead biomass for the recovery of heavy metals from aqueous solutions and a clear alternative to the conventional methods for the treatment of contaminated effluents. Research and development of new biosorbent materials are especially focus on biomass made from algae. The fundamental reason is its high sorption capacity and its availability almost unlimited [1]. Nevertheless, most publications in biosorption in recent years have mainly been devoted to other biomass (fungi and bacteria) more than to algae. It is more and more accepted that algae form a homogeneous group within the vegetal kingdom. They are divided among different evolutionary via, completely independent each other. A huge simplification leads to the following classification: a "red branch" with the red algae (rhodophyta), a "grey branch" with the grey algae (phaeophyceans), among them, and a "green branch" that contains beside the green algae (chlorophyta), mosses, ferns and several plants. The wall cell of each one of these algae is different and, therefore, the material responsible for the sorption is substantially different. Research works in the literature have principally been orientated to studies with brown algae [2-7] whereas both green [8-10] and red [11] algae have been used in less extension. On this base, the aim of the present work was to determine the sorption 783

Biosorption

capacity of different heavy metals by Spirogyra insignis, a green alga used as biomass. Until now, there has not been mention of the sorption behaviour of this alga in the literature. Our study was carried out as with monometallic as with bimetallic systems, these last ones near to the actual situation in industrial effluents [12]. On the other hand, having in mind the idea of foreseeing the behaviour of a given biomass in the presence of a contaminated effluent, there has been an attempt to model mathematically the biosorption system tested by means of a computer program of chemical speciation. 2.

MATERIALS AND METHODS The biomass used came from sweet water, collected from the Valmayor reservoir (Madrid, Spain) and was constituted in 99% by Spirogyra insignis. This alga grows very easily in stagnant or slowly flowing waters. It forms filamentous colonies where each one of the cells conserves its identity. The filamentous, without branches, are constituted by the joint of numerous cylindrical cells, which keep united each other by a viscous pectin layer that covers them superficially. Each cell has only one nucleus and several chloroplasts with ribbon and S form, which adapt to the membrane as rolled spiral ribbons. The membrane is of cellulosic nature. More than 300 species of Spyrogiras have been described, in all cases from low pH waters [13]. The tests were initiated from the frozen alga, which proceeded from the biomass collected in the reservoir. After several washes with distilled water, the liquid supernatant was eliminated with a peristaltic Watson Marlow pump model 503S. The remaining pulp, enriched with the alga, was centrifuged at 5000 rpm for 30 minutes in an Eppendorf centrifuge model 5804. Then, the biomass was dried in an oven at 60ºC up to reach constant weight. Later, the biomass was ground to a particle size less than 0.5 mm. The experiments were performed with synthetic solutions of 1000 mg/L of Cd2+, Cu2+, Ni2+, Pb2+ and Zn2+, made from the corresponding sulphate salts, except for lead where nitrate was used, by dilution in distilled water. In all cases, chemically pure reagents were employed. The pH adjustment of the solutions was carried out with sulphuric or nitric acid, according with each case, 1% v/v diluted. 1 g/L NaOH was used as basic reagent. The pH values were tested between 2 and 6, depending on the metal, but always below the hydroxide precipitation pH. In all tests, the biomass concentration and the initial metal concentration were kept constant at 1 g/L and 50 mg/L, respectively Tests were carried out in 100 mL Erlenmeyer flasks set on a multiple stirring Selecta Multimatic plate model 5S. Periodically, aliquots were taken out of the mixture to follow the biosorption process. These samples were centrifuged at 5000 rpm and, on the supernatant liquid, the pH was measured and the metallic concentration was determined by atomic absorption spectrometry. 3.

RESULTS AND DISCUSSION

3.1 Influence of pH The pH influences the biosorption process in two ways [14]. On the one hand, on the total charge of the biosorbent, since protons have the chance to be either adsorbed or released. This behaviour would be affected by the functional groups on the cell wall of the alga, establishing the equilibrium state as a function of the pH value of the medium. On the other hand, the pH value affects the solubility of the metallic ion in solution. For 784

Biosorption

instance, at the metal concentration used, the hydroxide precipitation pH values are: 8.5 (Cd), 5.9 (Cu), 7.9 (Ni), 5.9 (Pb) and 7.5 (Zn). In order to study the effect of this variable and to determine its optimum value, for which the sorption capacity of the biomass is maximum (q, expressed in mg or mmol of metal/g of biomass), a series of tests were carried out by modifying the pH. The results obtained are shown in Figure 1.

q (mmol/g)

0,40

Cadmium Nickel Zinc Copper Lead

0,30 0,20 0,10 0,00 2

3

4

5

6

7

pH

Figure 1. Effect of pH on biosorption of metals by Spirogyra insignis

A general trend was observed: at pH 2 the amount of metal adsorbed was negligible, increasing biosorption with higher pH values. This suggests that, at low pH values, there is a preferential biosorption by protons versus metal ions on the active sites of the biomass. Thus, the sorption capacity values are kept constant above pH 4, or even decrease, as in the case of copper. Bearing in mind that the solubility of metal hydroxides is not only described by the precipitation pH, before reaching such pH value, hydroxo metal-ion complexes of the type [Me(OH)n]z-n can be present. This could justify the evolution of q versus pH. The highest sorption capacity was set between 4 and 6, depending on the metal (6 for Zn, Cd and Ni, 4 for Cu and 5 for Pb). These results agree with others given in the literature for another kinds of algae [15]. 3.2 Influence of biomass concentration The biomass concentration is another important variable that can affect the amount of metal retained. For a given equilibrium concentration, the biomass adsorbs more metallic ions at lower than at higher cell densities [16]. It has been suggested that electrostatic interactions between cells must be a significant factor on the biomass concentration dependence with respect to metal sorption. In this sense, by decreasing the biomass concentration in suspension, for a given metal concentration, an increase of both the metal/biosorbent ratio and the metal retention by sorbent unit is observed, as long as there is not saturation. High biomass concentrations can exert a shell effect protecting the active sites against the metal. As a result, the amount of metal adsorbed by biomass unit is lower. In these tests the initial metal concentration in solution (50 mg/L) and the pH value (the optimum for each metal, according with the previous section) were kept constant. The biomass concentrations tested were: 0.5, 1.0 and 2.0, and the results obtained are shown in Figure 2. As can be observed, the amount of metal retained by the biomass presents a maximum at 1.0 g/L, except for lead. This confirms the loss of efficiency for high concentrations of biosorbent. The same can be said for low concentrations, probably due to the weak electrostatic attraction forces with biomass concentrations so low. However, this is not 785

Biosorption

q (mmol/g)

observed for lead since, as it will later be described, is the metal with higher affinity for the biomass. Anyway, differences are small and could also be due to experimental errors typical of this kind of tests.

Cadmium Nickel Zinc Copper Lead

0,40 0,30 0,20 0,10 0,00 0

0,5

1

1,5

2

2,5

Algal dose (g/L.) Figure 2. Effect of biomass concentration on biosorption of metals by Spirogyra insignis 3.3 Sorption isotherms Monometallic systems The isotherm represents the apparent solute sorption versus its equilibrium concentration at a given temperature. At low concentrations, which is the case for many practical sorption situations, the isotherms present two main forms, classified by Giles et all. [17] in two types: type L, which is concave with respect to the concentration axis, and type S, or sigmoidal, which is firstly convex and then concave respect to the same axis. A type L isotherm, with a horizontal part well defined, is associated to the sorption of a monolayer of solute with a minimum competition for the solvent. With the aim of determining the sorption isotherms related with Spirogyra insignis, five series of tests were conducted with monometallic solutions varying the initial metal concentration between 10 and 150 mg/L. The pH values and biomass concentration were maintained at optimum, as indicated previously. In all cases, type L isotherms were recorded, with a rapid increase in the sorption values with low equilibrium concentrations, reaching a stable maximum value with higher equilibrium concentrations. Using the linear Lagmuir’s model, which supposes the adsorption of a monolayer on the sorbent, and is represented by the following equation: Ce C K = e + q q max q max

(1)

The results shown in Table 1 were obtained. qmax data correspond to the maximum sorption value for high equilibrium concentrations. b is the equilibrium constant for the interaction between the metal and the biomass, and K is its inverse value. Therefore, both parameters measure the biomass affinity for the metal cation. R2 is the fitting coefficient expressing the degree of linear fitting of the experimental results to equation (1); a value close to one reflects a good fitting of the data obtained to Lagmuir’s equation. From the comparison of qmax data, the maximum sorption values for the five metals are very similar; the lower K value for lead reveals a higher affinity of the biomass for this metal, followed by cadmium, copper and zinc. Nickel and zinc present the same affinity. 786

Biosorption

Table 1. Langmuir parameters for adsorption isotherms pH Cadmium Nickel Zinc Copper Lead

6 6 6 4 5

Biomass (g/L) 1 1 1 1 0.5

qmax (mmol/g) 0.20 0.30 0.32 0.25 0.27

K (mmol/L) 0.07 0.39 0.39 0.13 0.01

B (L/mmol) 13.6 2.57 2.58 7.86 73.13

R2 0.99 0.97 0.95 0.98 0.99

Bimetallic systems Industrial effluents use to have more than one metal ion in solution. The retention of a metal by dead biomass is significantly affected by the presence of other cations in solution [18]. This is due to the fact that many functional groups on the cell wall and membrane are non-specific and then there is a competition of the different cations for the binding sites. This leads to a lower retention of a given metal in mixed solutions than when it is alone. The decrease in the metal retention by the presence of other ones depends on the nature and concentration of the other cations, decreasing the metal retention when increasing other cations concentration. In this section the biomass sorption capacity from bimetallic solutions within the system: Zn-Cd-Ni was analyzed. The experiments were carried out with a biomass concentration of 1 g/L and a pH value of 6. The metal concentrations tested were: 0, 10, 25, 50, 100 and 150 mg/L, testing for each metal all the concentrations of the other metal. The most appropriated way of studying the sorption of two metals together is by fitting the experimental data to a mathematic model [19] able to generate several parameters that help to evaluate quantitatively the influence of the presence of one metal on the sorption capacity of the other. In this way, Langmuir’s equation of the binary type was used, which establishes the equilibrium between two metals (M1 and M2) and the species resulting from the sorption by the biomass (B), B- M1 and B-M2: [B][M 1 ] k B + M1 ⇔ B-M1 (2) K 1 = −1 = k1 [B − M 1 ] [B][M 2 ] k B + M2 ⇔ B-M2 (3) K 2 = −2 = k2 [B − M 2 ] Whose final expression is as follows: q max Ce (M 1 ) K1 q( M 1 ) = 1 1 1+ Ce (M 1 ) + Ce (M 2 ) K1 K2

(4)

A high value of K for M2 (K2) versus M1 (K1) means a higher affinity of the biosorbent for M1 than M2, since high values of K are associated to a high value of the ratio between desorption and sorption amount. These two parameters, together with qmax, allow quantifying the biosorption process. Several authors have worked during the last years with this type of binary systems although their number is relatively limited. Biosorption modelling considering two metals systems can be carried out using either empirical equations or chemico-physical mechanistic models [20]. Empirical models are more or less simple mathematical 787

Biosorption

equations with some adjustable parameters, which can be fitted to the experimental data. On the other hand, the mechanistic models provide a major understanding of the chemical and physical aspects involved in the biosorption of heavy metals. These models are not simple mathematical equations and are based on a series of hypothesised chemicophysical reactions among active sites of the biomass and ions in solution. Any case, mechanistic models require a wide and deep experimental investigation and its application to real systems is difficult. The experimental data obtained in the present work were fitted according to the empirical modelling approach. Table 2 collects all these parameters for the three bimetallic systems under study, which were obtained through a computing treatment of the experimental data by using the MATLAB program. It can be seen, similarly to monometallic systems, that the biomass affinity is higher towards Cd than to Zn and Ni. Of the three metals studied, nickel presents a higher K value and, therefore, less affinity for the biomass. With respect to the maximum sorption capacity, it is higher in those systems with Cd, which is in agreement with the higher affinity of the biomass for this element, within the trimetallic system used. Table 2. Langmuir parameters applied to bimetallic systems Ni - Cd Zn - Cd Zn - Ni

K1 (mmol/L) 0.23 0.12 0.08

K2 (mmol/L) 0.07 0.06 0.16

Qmax (mmol/g) 0.28 0.27 0.23

Using the same computer program, equation (4) could be represented by sorption isotherms surfaces in three dimensions, recording in X and Y axes the equilibrium concentrations of the two metals, and in Z-axis the sorption capacity either of one of the two metals or the sum of both. Figures 3, 4 and 5 are an example of the power of the program and show the equilibrium conditions for the systems Cd-Ni, Zn-Ni and Ni-Cd, respectively. These three cases have been choosing to show the influence of the metal with the less biomass affinity (Ni) on the sorption of the two others (Cd and Zn), and reciprocally the influence of the metal with the highest affinity (Cd) on the metal with the less one (Ni). It can be observed as Ni does not exert a clear influence on the sorption of Cd and Zn, since for high Ni concentrations there is not significant decrease in the q values of Cd and Zn (Figures 3 and 4). On the contrary, Cd provokes an outstanding decrease on the sorption levels of Ni (Figure 5).

Figure 3. Surface of the sorption isotherm of Cd in the presence of Ni

788

Biosorption

Figure 4. Surface of the sorption isotherm of Zn in the presence of Ni

Figure 5. Surface of the sorption isotherm of Ni in the presence of Cd

More information was obtained from the rest of 3-D figures (not shown here). For instance, in the case of Zn and Ni, in spite of its similar affinity for the biomass in monometallic systems, some differences were detected in bimetallic tests: with greater influence of Zn than Ni on the biosorption of Cd. 3.4 Simulation of the biosorption process The last stage of this study consisted in the simulation of the data corresponding to the equilibrium conditions of monometallic systems by means of a computer program of chemical speciation, PHREEQCI 6.2 [21]. Behind the comparison of the experimental data with those obtained from the model is the possibility of foreseeing the behaviour of the biomass theoretically. For this, the surface reactions between the biomass and each one of the metals were incorporated to the database of the program, with its corresponding equilibrium constant determined, as commented before, by the b constant of Langmuir’s model. Such reactions are: (5) B + Me+2 = BMe+2 Ce q (qmax - q) +2 (6) Spiroins + Cd = SpiroinsCd+2 log_k 4.1332 # (k = 13,590) +2 (7) Spiroins + Zn = SpiroinsZn+2 789

Biosorption

log_k 3.4121 # (k = 2,583) +2 (8) Spiroins + Ni = SpiroinsNi+2 log_k 3.4103 # (k = 2,572) +2 (9) Spiroins + Cu = SpiroinsCu+2 log_k 3.7414 # (k = 5,513) +2 (10) Spiroins + Pb = SpiroinsPb+2 log_k 5.0714 # (k = 117,870) To run the program it was necessary to specify the conditions of the dissolutions to deal with (pH, temperature, equilibrium metal concentration, ionic species, etc.), besides the own characteristics of the biomass (qmax, specific area and weight used). Table 3 collects both the experimental data and those obtained with the computer program. qmax-q values are related to the active sites of the biomass that remained unoccupied during the equilibrium, which does not necessarily mean the amount of metal in dissolution; this would only happen in the case of saturation of the biomass. Table 3. Comparison between experimental and program values

Lead

Copper

Zinc

Nickel

Cadmium

Ci Ce (mg/g) (mmol/L) 10 25 50 100 150 10 25 50 100 150 10 25 50 100 150 10 25 50 100 150 10 25 50 100 150

790

0.02 0.09 0.28 0.68 1.07 0.08 0.30 0.70 1.55 2.39 0.08 0.34 0.79 1.68 2.60 0.03 0.21 0.56 1.28 2.02 0.00 0.02 0.12 0.34 0.58

Experimental Program q qmax-q q qmax-q (mmol/g) (mmol/g) (mmol/g) (mmol/g) 0.06 0.14 0.04 0.16 0.11 0.09 0.11 0.09 0.14 0.06 0.15 0.05 0.18 0.02 0.18 0.02 0.20 0.00 0.18 0.02 0.09 0.21 0.05 0.25 0.13 0.17 0.12 0.18 0.17 0.13 0.18 0.12 0.22 0.08 0.23 0.07 0.27 0.03 0.24 0.05 0.11 0.22 0.05 0.27 0.16 0.17 0.14 0.18 0.19 0.13 0.20 0.12 0.23 0.09 0.25 0.07 0.31 0.02 0.27 0.05 0.12 0.18 0.04 0.26 0.18 0.13 0.15 0.15 0.20 0.11 0.22 0.09 0.24 0.07 0.26 0.05 0.30 0.01 0.27 0.03 0.08 0.17 0.01 0.24 0.18 0.07 0.17 0.08 0.22 0.03 0.23 0.018 0.24 0.01 0.24 0.01 0.25 0.00 0.25 0.00

Experimental qmax-q q (%) (%) 30.7 69.3 55.6 44.4 71.2 28.8 89 11 97.9 2.1 28.9 71.1 44.8 55.2 55.4 44.6 72.9 27.1 91.5 8.5 33.2 66.8 48 52 58.4 41.6 71 29 94.7 5.3 39.3 60.7 58 42 64.7 35.3 77.7 22.3 98.4 1.6 33.3 66.7 72.1 27.9 88 12 95.4 4.6 99.2 0.8

Program qmax-q q (%) (%) 19 81 54.4 45.6 77.1 22.9 88.7 11.4 92.2 7.8 17 83.1 40.9 59.1 60.5 39.5 75.8 24.2 82 18 17 83 43.3 56.7 63.1 36.9 77.2 22.8 83.2 16.8 14.2 85.8 49.4 50.6 71.7 28.3 84.4 15.6 89.1 10.9 5.3 94.7 68.8 31.2 92.7 7.3 97.3 2.7 98.3 1.7

Biosorption

It can be checked that, in general, there is a remarkable similarity between the experimental data and those obtained with the program. The value with the worse correlation is the one corresponding to the lowest initial metal concentration. This can be due to a higher competition of protons for the active sites on the cell wall of the alga, since the presence of the metallic species in dissolution was, in this case, little significant. By increasing the metal concentration, the difference between the experimental and calculated values is about 10%. This demonstrates the validity of the computer program used in order to simulate this type of processes. 4.

CONCLUSIONS In the sorption process of Cd, Ni, Zn, Cu and Pb by Spirogyra insignis the optimum sorption pH was set between 4 and 6 depending on the metal and the optimum biomass concentration was reached at 1 g/L, practically in all cases. The qmax values were very similar for all five metals, being lead the metal with higher affinity for the biomass. The fitting of the experimental data to Langmuir’s model was excellent. In multimetallic systems, Ni did not exert a remarkable influence on Cd and Zn sorption. On the contrary, Cd provoked a significant decrease on Ni sorption levels. Finally, the computer treatment of the data revealed an outstanding similarity between experimental data and those obtained through the computer program. Definitively, the biomass used in this study is a very efficient biosorbent for the recovery of metals from residual effluents. ACKNOWLEDGMENTS The authors wish to express sincere gratitude to Dirección General de Enseñanza Superior e Investigación Científica del Ministerio de Educación y Cultura (Spain) for funding this research. Also, we want to thank sincerely to Dr. E. Costas (UCM) for his help in the collection and supply of the biomass used in this study. REFERENCES

1. S. Klimmek, H.J. Stan, A. Wilke, G. Bunke and R. Buchholz, Environ. Sci. Technol., 35 (2001) 4283. 2. Z.R. Holan, B. Volesky and I. Prasetyo, Biotechnol. Bioeng., 41 (1993) 819. 3. K.H. Chong and B. Volesky, Biotechnol. Bioeng., 47 (1995) 451. 4. J.T. Matheickal and Q. Yu, Bioresource Technol., 69 (1999) 223. 5. J.T. Matheickal, Q. Yu and G.M. Woodburn, Water Res., 33, 2 (1999) 335. 6. Q. Yu, J.T. Matheickal, P. Yin and P. Kaewsarn, Water Res., 33, 6 (1999) 1534. 7. A.Leusch, Z.R. Holan and B. Volesky, J. Chem. Tech. Biotechnol., 62 (1995) 279. 8. G. Çetinkaya, Z. Aksu, A. Öztürk and T. Kutsal, Process Biochem., 34 (1999) 885. 9. Z. Aksu, Ü. Acikel and T. Kutsal, Separ. Sci. Technol., 34, 3 (1999) 501. 10. Z. Aksu, Ü. Acikel and T. Kutsal, J. Chem. Tech. Biotechnol., 70 (1997) 368. 11. Z.R. Holan and B. Volesky, Biotechnol. Bioeng., 43 (1994) 1001. 12. H.S. Lee and B. Volesky, Water Qual. Res. J. Canada, 34 (1999) 519. 13. Linda E. Graham and Lee W. Wilcox (eds.), Prentice Hall, USA, 2000. 14. E. Guibal, I. Saucedo, J. Roussy and P. Le Cloirec, React. Polym., 23 (1994) 147. 15. F. Veglio' and F. Beolchini, Hydrometallurgy, 44 (1997) 301. 16. S.K. Mehta and J.P. Gaur, Ecol. Eng., 18 (2001) 1. 17. C.H. Giles, T.H. Macewan, S.N. Nakhwa and D.J. Smith, J. Chem. Soc., 3 (1960) 3973. 18. K.H. Chong and B. Volesky, Biotechnol. Bioeng., 49 (1996) 629. 791

Biosorption

19. A. Sánchez, A. Ballester, M.L. Blázquez, F. González, J.A. Muñoz, A. Hammmaini, FEMS Microbiol. Rev., 23 (1999) 527. 20. F. Pagnanelli, A. Esposito, F. Veglio, Water Res., 36 (2002) 4095. 21. S.R. Charlton, C.L. Macklin and D.L. Parkhurst (eds.), Water-Resources Investigations. Report 97-4222, U.S. Geological Survey, Lakewood CO, 1997.

792

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption of 226Ra and Ba by Sargassum sp. Wilson Cervi da Costaa,b, Oswaldo Garcia Júniorb*, Heliana de Azevedo Gomesa a

Poços de Caldas Laboratory, Brazilian National Commission for Nuclear Energy Poços de Caldas, State of Minas Gerais, 37701-970, P.O. Box 913, Brazil b Department of Biochemistry and Chemical Technology, Institute of Chemistry, São Paulo State University, Araraquara, State of São Paulo, 14801-970, P.O. Box 355, Brazil Abstract The goal of the this work was to investigate the removal of 226Ra from a radioactive solution and to verify if barium can be an useful model for radium-226 biosorption since both elements are chemically similar. The non-living biomass utilized in this study was obtained from Sargassum sp. The biosorption kinetics experiments were carried out at pH 3.5 and the specific metal uptake (Ba) or radionuclide activity (226Ra) and efficiency of removal were the parameters evaluated. Sargassum sp. biomass showed an efficient removal of 226Ra from solution (~99%) and the behavior of barium biosorption was close to that of 226Ra, showing that it can be used as a model for this radionuclide in this process. 1.

INTRODUCTION Biosorption comprises binding of metals to the biomass by a process, which does not involve metabolic energy or transport, although this process may also occur simultaneously where live biomass is used [1]. Biosorption of heavy metals or radionuclides by microbial cells or biomass in general has been recognized as a potential alternative to existing technologies for removal of these contaminants from polluted waters [2]. This phenomenon is generally described as retention of ions from solution by microbial cells and this metal uptake is normally very efficient and frequently selective [3]. According to the International Atomic Energy Agency (IAEA), radioactive waste is any material that contains a concentration of radionuclides greater than those deemed safe by national authorities, and for which no use is foreseen. Based on this definition and on the anthropogenic radioecological effects, the percentage of radionuclides removed from effluents, waters deriving from nuclear installations and acid mine drainage must be maximized.

* Corresponding author: E-mail address: [email protected] OGJ acknowledges Conselho de Desenvolvimento Científico e Tecnológico - CNPq for research fellowship

793

Biosorption

Despite a great number of studies of biosorption applied to remove toxic metals from industrial effluents, mainly from mining activities, only few studies have been done focusing biosorption of radionuclides [4, 5]. Uranium and thorium are natural sources of radionuclides 226Ra e 228Ra, respectively. Mining activities and hydrometallurgical processing of ores and minerals of these elements can result in wastes and effluents containing both radionuclides. Effluents and/or residual waters from nuclear installations have ultra-diluted concentrations of radionuclides along with other diverse chemical species. Most of the time, the concentration of these diverse chemical species is much higher to the radionuclide concentrations. Therefore, the biosorbents must show elevated selectivity and removal efficiency (above 99%) of the present radioactive species. The main objective of this study was to investigate the biosorption of 226Ra and to compare its behavior with barium, a chemically similar element, utilizing Sargassum sp., a brown seaweed that has been utilized in several studies of biosorption [6, 7]. 2.

MATERIALS AND METHODS

2.1 Biomass The brown seaweed Sargassum sp. was collected at coast of São Sebastião – SP, Brazil. The biomass was washed four times with 500 ml of water purified to 18 MΩ in a 1 L beaker. The system was kept agitated and heated to 50°C in a magnetic agitator for a total of 12 hours. The changes in washing waters were performed every 3 hours, and the last operation of solid-liquid separation was accomplished through vacuum filtration. The washed biomass was sun dried and maintained at room temperature. For biosorption experiments the algae was shopped in pieces with size around 1 mm and 3 mm, washed for 2 hours in hydrochloric acid 0.1 mol.L-1 and after that it was separated in 2 fractions and washed in ultra-pure water up to pH value of 3.50 or 5.00. 2.2 Working solutions a) Ba2+ The reagent BaCl2.2H2O was used to prepare the stock standard solution at 1 g.L-1 of Ba. From this stock solution, were prepared by appropriated dilution working solutions in concentrations ranged from 0.1 mg to 10 mg L-1 of Ba. These working solutions were prepared at pH 3.50 and 5.00 using HCl for pH adjustments. b) Radioactive 226Ra The radionuclide 226Ra working solutions were prepared in the following nominal activity concentrations (Bq L-1): 900; 1,125; 1,350; 1,575; 1,800; 2,025; 2,250; 4,500; 22,500 e 45,000, adjusted to pH 3.00 or 5.00 with LiOH solution during the biosorption experiments. 2.3 Biosorption experiments Biosorption experiments were carried out by using 0.1 g of biomass in sealed 125 mL Erlenmeyers and 20 mL of working solutions of Ba or 226Ra, at initial concentrations (Ci) of 0.1 to 10 mg L-1 and 900 to 45,000 Bq L-1, respectively. The flasks were incubated in a shaker at 28°C, 250 rpm for 180 minutes (Ba2+) and 240 minutes (Ra2+) as final contact time. The pH values were adjusted to 3.50 or 5.00 every 30 minutes, whenever necessary. The experiments were performed in duplicate for each condition; water purified to 18 MΩ 794

Biosorption

at above mentioned pH values was used as a control. Due to adsorption of 226Ra on glass walls (physical adsorption, 13% at pH 3.50 and 36% at pH 5.00) the actual Ci and Cf (final concentration at equilibrium) values for 226Ra standard solutions were determined and considered in the final results. The following formula was used to calculate the accumulated Ba2+ mass or concentration of the accumulated activity for 226Ra2+ [8]: V × (C i − C f ) q= (1) b where: q is the mass (mg) of metal or activity (Bq) of radionuclide by 1g of dry biosorbent; V is the volume of the test solution (L); Ci and Cf are respectively, the initial and the equilibrium concentrations of the metal (mg L-1) or radionuclide (Bq L-1) in solution; b is the mass of dry biosorbent (g). 2.4 Chemical and radiometric determinations Barium was determined by ICP-AES, using a Jarrell-Ash model AtomComp 975. The radionuclide 226Ra activity was determined using a proportional alpha and beta meter, run with an ultra-low background gas, model ESM-EBERLINE FHT 770T. 3.

RESULTS AND DISCUSSION

3.1 Biosorption kinetics Figures 1 and 2 show, respectively, the kinetics of Ba and 226Ra biosorption by Sargassum sp. As it can be seen high affinities were observed between Ba2+ (a Class A metal) and the linkage sites at the Sargassum sp. cell walls. This fact agrees with the Class A metal ions preference to linkage sites that contain oxygen, which is the case of several oxygenated groups such as alginic acid, fucoidan, agars and carrageenans, present in algae cell walls, especially marine algae. This responds to the hydrophilic properties of the algae, making it susceptible to ionic exchange [9, 10, 11, 12]. 0.6

0.4

-1

q (mg g )

0.5

0.3 0.2 0.1 0.0 0

30

60

90

120

150

180

Time (min) Figure 1. Biosorption kinetics of Ba by Sargasssum sp. at pH 3.50 and initial concentration of 0.094 mg.L-1 795

Biosorption

160 140

-1

q (Bq g )

120 100 80 60 40 20 0 0

30

60

90

120

150

180

210

240

Time (min) Figure 2. Biosorption kinetics of 226Ra by Sargasssum sp. at pH 3.50 and initial radioactivity of 800 ± 40 Bq.L-1

Comparing both kinetics it is evident that Sargassum sp. showed a higher efficiency of biosorption of 226Ra than that observed for Ba. Group II elements, which belong to Class A ions, bind to radicals containing oxygen through ionic exchange mechanisms. The ionic exchange selectivity for these elements has the following qualitative prevalence: •

Selectivity increases with the increase of ion charge

•

Selectivity increases with the decrease of ion radius (hydrated)

•

Selectivity increases with the increase of ion polarization Therefore, the ranking of affinity of these elements for strong cationic resins containing sulfonic radicals (RSO3 -) are: 2+ 2+ 2+ 2+ 2+ Group II: Ra aq . > Ba aq. > Sr aq. > Ca aq. > Mg aq.

Besides, solubility product (Ksp) is another parameter that can be used to evaluate the affinity of radium and barium for these radicals. The Ksp values of barium sulfate and radium sulfate are 1.08 x 10-10 and 3.66 x 10-11, respectively [13], that is, radium has a higher affinity than barium for sulfate containing radicals, such as the fucoidan present in Sargassum sp. cell wall. However the difference in Ksp values is not much significant, which become barium a competitor ion with radium for these radicals; this is not true for calcium, for example, which has a much higher Ksp (3.14 x 10-5) [14]. So, barium can be used as a model for radium biosorption experiments. 3.2 Biosorption isotherms The isotherms of biosorption are curves that describe the equilibrium between the metal in solution and the biosorbent at a constant temperature. These curves are extensively used in the studies for comparison of biosorption performance of different biosorbents. Figures 3 and 4 show the biosorption equilibrium isotherm for Ba and 226Ra, respectively.

796

Biosorption

150 125

-1

q (mg g )

100 75 50 25 0 0

50

100

150

200

-1

250

300

350

Cf (mg L ) Figure 3. Ba biosorption equilibrium isotherms at pH = 5.00, for Sargasssum sp.

10,000

6,000

-1

q (Bq g )

8,000

4,000 2,000 0 0

200

400

600

800

-1

Cf (Bq L ) Figure 4. 226Ra biosorption equilibrium isotherms at pH = 5.00, for Sargasssum sp.

The results obtained from the Langmuir fitting showed a maximum biosorption coefficient (q max) of around 150 mg.g-1 for Ba and 10,000 Bq.g-1 for 226Ra. For others biosorbents tested in our laboratory, such as the microalgae Monoraphidium sp., Penicillium sp. (filamentous fungi) and the yeast Saccharomyces cerevisiae, these values (data not shown) were significantly lower than those obtained with Sargasssum sp. 4.

CONCLUSION Radium bioaccumulation has not been studied frequently, and with rare exceptions, few scientists tried to develop bioprocesses that employ some biological systems with potential for radium removal [4]. It has been demonstrated the superiority of biosorbents 797

Biosorption

compared with traditional adsorbents such as natural zeolites, manganese zeolites, zircon salts, Bio-rex exchange ions resin and activated carbon [15]. For example, radium sorption values ranged from 80 to 130 Bq.g-1 are reported for traditional sorbents whereas for biomass present in activated sludge the values ranged from 1,500 to 2,800 Bq.g-1 [15]. In the same reference [15] biomass of Penicillium chrysogenum showed capacities of radium biosorption ranged from 16,650 to 77,700 Bq.g1 depending on the initial concentration of radium and pH. In our work a maximum biosorption coefficient of 10,000 Bq.g-1 has been obtained for Sargassum sp., which means practically 100% of 226Ra removal at pH 3.50 as well as pH 5.00. This value is lower compared to the data reported before [15], but different conditions were used in that research, such as initial radium concentration and pH values which ranged from 7.00 to 10.00. In our work barium biosorption data indicated that it can be used as an “indicator” of the biosorption of 226Ra, since similar behavior was found in the assays carried out, despite some differences in the kinetics experiments in which 226Ra was more specific and more efficient for Sargassum sp. biosorption activity. The results obtained in this work are useful to define biosorption capacity of Sargassum sp. for 226Ra and will help to evaluate the potentiality of the process for its technological application. REFERENCES

1. 2. 3. 4. 5.

J.M. Tobin, C. White and G.M. Gadd, J. Ind. Microbiol., 1994; 13: 126–30. F. Veglio and F. Beolchini, Hydrometalurgy, 1997; 44: 301–16. A. Ozer, D. Ozer and H.I. Ekiz, Process Biochem., 1999; 34: 919–27. L.E Macaskie, Critical Reviews in Biotechnology, 1991; 11: 41-112. M.A. Fomina, V.M. Kadoshnikov and B.P. Zlobenko, In: R. Amils and A. Ballester (Eds) Biohidrometallurgy and The Enviroment Toward the Mining of the 21st Century, Elsevier Science, part B, (1999) 245-254. 6. M.C. Palmieri, B. Volesky and O. Garcia Jr., Hydrometallurgy, 67 (2002) 31-36. 7. M.M Figueira, B. Volesky, K. Azarian and V.S.T. Ciminelli, In R. Amils and A. Ballester (Eds) Biohidrometallurgy and The Enviroment Toward the Mining of the 21st Century, Elsevier Science, part B, (1999) 503-512. 8. B. Volesky, In: V.S.T. Ciminelli and O. Garcia Jr. (Eds.) Biohidrometallurgy: fundamentals, technology and sustainable development, Elsevier Science B.V., part B, (2001) 69-80. 9. W. Stumm and J.J Morgan, J.J. AQUATIC CHEMISTRY – Chemical Equilibra and Rates in Natural Waters. Third Edition ENVIRONMENTAL SCIENCE and TECHNOLOGY. Chap. 10 – Trace Metals: Cycling, Regulation, and Biological Role, pp 614-671. John Wiley & Sons, Inc., 1996. 10. L. Volterra and M.E. Conti, Int. J. Environment and Pollution, 2000; 13: 92-125. 11. H.L. Ehrlich, Biotechnol. Bioeng. Symp., 1986; 16: 227-238. 12. R.H. Crist, K. Oberholser, D. Schwartz, J. Marzoff, D. Ryder and D.R. Crist, Environ.Sci.Technol, 1988; 11: 755-760. 13. Handbook of Chemistry and Physics. David R. Lide, Editor-in-Chief, 83rd Edition, 2002-2003. 14. M. Tsezos, I. Baird, and L. Shemilt, “Adsorptive treatment with microbial biomass of 226 Ra-containing waste waters, “The Chemical Engineering Journal, 32: B29-B38 (1986). 15. M. Tsezos and D.M. Keller, Biotechnol. Bioeng., 1983; 25: 210-215.

798

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption of arsenic and heavy metals on a ceramic-based biomass. Batch equilibrium experiments with Cu2+ model solutions G. Horaka, S. Willschera*, P. Wernera, W. Pompeb a

TU Dresden, Faculty of Forest, Geo and Hydro Sciences, Institute of Waste and Site Management, Pratzschwitzer Str. 15, D - 01796 Pirna, Germany b TU Dresden, Institute of Material Sciences, Hallwachsstr. 3, D-01069 Dresden, Germany

Abstract The aim of this work was to study the suitability of a new ceramic-based biomass for the sorption of heavy metals from aqueous solutions. The biosorbens (Biocer) used here is consisting of a strain of Bacillus sphaericus immobilised in a ceramic material. In the first batch equilibrium experiments for the sorption of Cu2+ in model solutions was obtained a maximum sorption capacity of 6 mg Cu2+/g Biocer material (dry mass). A yield of up to 85% of the quantitative copper sorption was achieved after 60 minutes. Subsequently, the sorption rate of copper is relatively slow; it is fitting to a 1st order reaction. For practical use, a regeneration of the sorbens material is of major economical importance. Desorption experiments with 0.005 M citric acid resulted in an almost completely recovery of copper after a single regeneration. The Biocer material showed also a really good regeneration behaviour of recently 5 to 6 sorption/ desorption circles without any loss of its loading capacity or its stability. As a result, the material shows a good suitability not only for the decontamination of solutions with a low metal content, but also possibly for the recycling of economically valuable metals. 1.

INTRODUCTION Pit waters and seepage from mine tailings or dumps are containing a huge potential of harmful substances like arsenic and heavy metals. In the former uranium mining areas in Germany, especially uranium and arsenic are contaminants of enhanced environmental attention. The wastewaters from the flooded pits there as well as seepage waters from the large spoiled piles are containing a high future contamination potential. Many approaches were developed and tested to solve this problem; one of them is the use of biosorption for the treatment of mine drainage waters [1-5]. In the last years, several sorption experiments with different biomaterials were successfully implemented

* Author for correspondence: S. Willscher, e-mail [email protected]

799

Biosorption

[5-22]. Bacillus sphaericus, the Bacillus strain used here, is known for its excellent sorption capacity of uranium and other heavy metals [23,24]. In this context, a new biosorption material, called Biocer, was developed combining the good sorption properties of the bacteria with mechanical stability by means of the sol-gel technology [25]. The name Biocer means here the combination of a biological component with ceramic material. The vegetative cells are firmly bound in this ceramic material, whereas harmful metal ions can diffuse through the pores of the ceramics and then they become sorbed on the cell walls. With an optimised grain size, the Biocer material shall later be used in column reactors for the treatment of contaminated waters. Such reactors may be applied as small decentral constructions for especially low concentrations of harmful metal ions and arsenic. The Biocer material can also be used for the postprocessing of discharges of conventional treatment units in order to keep the limits of environmental contaminants into the surface water. In this work, first biosorption experiments with this Biocer material were carried out with synthetic solutions as model investigations in order to obtain fist equilibrium and kinetic data about the removal of metal ions and to get first results about the stability of the biosorbing material. 2.

MATERIALS AND METHODS

2.1 Material The Biocer material was supplied by "Kallies Feinchemie AG" Sebnitz (Germany). In these experiments was used a grain size of the biocer material of 90- 710 µm. Before starting the experiments, the material was conditioned with physiological sodium chloride solution or 0.1% sodium perchlorate solution. Therefore, 0.1 g of the material was four times washed for 15 min with 3 ml of the salt solution. After conditioning, the Biocer material was suspended in the metal containing solutions to perform the biosorption experiments. 2.2 Biosorption experiments The measurement of the sorption isotherms for copper(II) was carried out in batch experiments in closed test tubings. 1% of the Biocer material (w/v) was added to the copper solutions, and all was mixed in an overhead shaker for 2 hours. The concentration of the copper solutions used ranged from 0.1-250 mg/l and they were prepared by dilution of a stock solution (1 g/l Cu by dissolving of CuSO4 in deionised water). The pH of the experimental solutions was adjusted to pH 3 and pH 5 by adding HNO3 and NaOH as required. Finally, the metal content of the sample supernatants was analysed by atomic absorption spectroscopy (AAS, Perkin Elmer 4100). 2.3 Desorption experiments The desorption experiments of copper were carried out directly after the end of the sorption experiments. Therefore, the Biocer material was separated from the metal solution, and suspended in a 0.005 M citric acid adjusted to pH 3.2. The desorption experiments were carried out by mixing in an overhead shaker over a time of 20 h. For a new sorption cycle, the Biocer material was conditioned again 3-4 times with a physiological sodium chloride solution, as described in point 2.1, and further used as described above. 800

Biosorption

2.4 Kinetic sorption experiments Kinetic experiments of the biosorption of copper were carried out with 1 g Biocer in 100 ml copper solution at a pH of 5.3. Up to 10 samples were taken in different time intervals, each 5 ml. 3.

RESULTS AND DISCUSSION

3.1 Solution chemistry of Cu2+ For a better understanding of the sorption behaviour of the Cu2+ on the Biocer material, at first a few important items of the solution chemistry of CuSO4 shall be shortly outlined. The copper sulfate has a good solubility in aqueous solutions, and the Cu exists there in form of its dissociated bluish [Cu(H2O)4]2+ ions. The solutions react weakly acid, and at a concentration of 250 mg/l was measured a pH of 5.1. For this copper concentration, the solubility product is achieved at a pH of 5.8. This means, that at a pH higher than 5.8 the copper is already precipitated in different species. Additionally, in the literature is described the existence of undissolved copper species like Antlerite [Cu3SO4(OH)4] and Brochanite [Cu4SO4(OH)6] in the pH range of 4,5-6 [26]. By this reasons, a precipitation of different copper species in the biosorption experiments under the conditions described above should be avoided. 3.2 Sorption isotherms of Cu2+ The Biocer material is nearly completely neutralising weak acid solutions. After adding the Biocer material, an increase of the pH was measured in the copper solutions depending on the copper concentrations. The pH increased from the initially adjusted pH 3 to pH 5.2 in a 250 mg/l Cu solution, and to pH 6.3 in a solution of 1 mg/l Cu. In copper solutions initially adjusted to pH 5, the pH was increasing to 5.1 for 250 mg/l Cu and to pH 7.4 for 1 mg/l Cu. At the final pH of the solutions exists a high probability for the formation of precipitated copper species under these conditions.

loading mg Cu/g biosorbens

6

5

4

3

2

pH 5 1

pH 3

0 0

50

100

150

200

250

300

conc. in mg Cu/l

Figure 1. Sorption isotherms for Cu2+ in solutions of various initial pH

801

Biosorption

For this reason, additional samples were analysed parallel to the sorption experiments. A variance coefficient of the measurement data of 25% was calculated for the low concentrated copper solutions with an enhanced final pH of 6.3 and 7.4, resp., after a reaction time of 2 h. At a final pH of 5.5, a variance coefficient of less than 10% was calculated. In Fig. 1 are shown the sorption isotherms of copper at two different initial pH values of the solutions. An increase of the sorption capacity with an increasing pH is clearly recognisable (Fig. 1). The cell surfaces of the immobilised Bacillus strain contain a number of functional groups like hydroxy and carboxy groups, which are able to bind protons and metal ions coordinatively and to dissociate them in a sophisticated equilibrium. This is resulting in a strong pH dependency of the biosorption processes. The metal ions are easily replacing the protons at higher pH values. Furthermore, other binding mechanisms (complex or chelate bindings) are also involved in the biosorption process. Sorption model The Langmuir model was used to fit the experimental data (eq.(1)). The model is basing on the assumption of a confined number of binding places on the surface of the sorbens [27,28]. ceq 1 b = ⋅ c eq + (1) a q q max q max Under the supposition of a defined number of binding places on the sorbens surface, the model assumptions should be only of restricted validity for the biosorption experiments (because there is a superimposition of different binding mechanisms). Although, the experimental data are fitting satisfying with the Langmuir sorption model (Figs. 2 and 3) at the various pH values. 70

40 35

60

R2 = 0,9953 Ceq/ aq (g/l)

Ceq (g/l)

30 25 20 15

40 30

10

20

5

10

0

R2 = 0,9656

50

0

0,0

1,0

2,0

3,0

conc. (mmol Cu/l)

4,0

0,0

2,0

4,0

6,0

Ceq (mmol Cu/l)

Figure 2. Linearised Cu2+ isotherm, Figure 3. Linearised Cu2+ isotherm, initial pH 5 initial pH 3 As a result, the Langmuir equation can be used to fit the experimental biosorption data obtained with the Biocer material. So, the copper ions seem to be bound nearly exclusively on the surface groups of the biomass, and they are easily exchangeable. Under the given experimental conditions, precipitation processes as well as a longer lasting incorporation into the cells, resulting in a lower exchangeability of the copper ions, seem not to occur. Fig. 4 shows the sorption behaviour of the Biocer material at various Cu2+ concentrations of the solutions. Up to a concentration of 50 mg/l, the Cu2+ is sorbed in an average amount of 85% in the equilibrium experiments. At Cu2+ concentrations above, a 802

Biosorption

decrease of the yield of the sorbed Cu2+ was observed. In the equilibrium biosorption experiments was used a concentration of the Biocer material of 1% (Pt. 2.2.). This is resulting in a specific sorption capacity of this material of up to 6 mg Cu/g Biocer. 3.3 Desorption experiments For practical use, a regeneration of the sorbens material is of major economic importance. Experiments desorbing the copper(II) bound to the Biocer material were carried out with 0.1M NaOH and 0.5 M citric acid. Under these experimental conditions, a complete desorption, but also a partial destruction of the Biocer material was observed. Obviously, a pH of 3 is the lower limit of pH stability for this kind of material. Below this pH limit, a repeated sorption of copper is impossible. Probably, the Biocer material is then directly destructed. In further optimising experiments was found out, that the Cu2+ can be nearly quantitatively desorbed without destructing the Biocer material by using a 0.005 M citric acid at a pH of 3.2 (Fig.4).

sorption- desorption Cu in %

100

sorption

90 80

desorption

70 60 50 40 30 20 10 0 1

5

10

25

50

75

100

250

500

conc. (mg Cu/l)

Figure 4. Sorption and desorption behaviour of the Biocer material at various Cu2+ concentrations

Fig. 4 shows the amount of desorbed copper compared to the sorbed amount. At all loading concentrations used, a nearly quantitatively desorption of the bound Cu2+ was observed (Fig. 4). So, the optimised experimental conditions of the regeneration with 0.005 M citric acid at a pH of 3.2 seem to be really good suited for a nearly quantitatively removal of the sorbed Cu2+ ions from the Biocer material. 3.4 Regeneration behaviour of the Biocer material After the successful experiments of the desorption of Cu2+ from the Biocer material, the loading behaviour in various sorption/ desorption cycles was investigated under various concentrations of the Cu2+ solutions (Fig.5). In all sorption cycles was achieved a stable sorption/ desorption behaviour of the Biocer material, without a substantial loss of its sorption properties or the stability of the material (Fig. 5). So, a constant use of the Biocer material at least over 5 to 6 regeneration cycles seems to be feasible. The material seems to withstand several regeneration cycles without any loss of its loading capacity or its stability. The runoff of the metal concentrates in a regeneration cycle can be possibly used for further recovery processes. 803

Biosorption

loading mg Cu/g biosorbens

7 6 5

load 1 load 2

4

load 3 3

load 4 load 5

2 1 0 0

50

100

150

200

250

300

conc. mg Cu/l

Figure 5. Multiple regeneration cycles of the Biocer material at various Cu2+ concentrations (pH 5) 3.5 Sorption kinetics of the Biocer material In further experiments, the sorption kinetics of the Biocer material was investigated (Fig. 6). The equilibrium sorption capacity of 85% of the dissolved Cu2+ is achieved after 1 hour (Fig. 6). After that, the sorption process of the Cu2+ is only slightly increasing and approaching its maximum. 5.6 5.5

mg Cu/g sorbens

5.4 5.3 5.2 5.1 5 4.9 4.8 4.7 4.6 0

50

100

150

200

time in min

Figure 6. Sorption kinetics of the Biocer material with Cu2+ in aqueous solution

The results show, that the sorption process of the Cu2+ ions needs a time of around an hour for the diffusion through the porous ceramic material and until the Cu2+ is bound to the cell walls in a steady state equilibrium. In order to obtain closer predications about the kinetics, the reaction order of the biosorption process was graphically determined. The reaction rate is calculated by eq. (2), where aq* is the loading of the Biocer material, calculated from the difference of the maximum loading aqs and the loading aqt versus the time t. 804

Biosorption

By integration of eq. (2) was obtained the linearised eq. (3), with which the experimental data were fitted (Fig. 7). Fig. 7 shows, that the biosorption process of Cu2+ at the Biocer material obeys a reaction of 1st order. * da q * * * (2) ln a q = ln a q 0 − k ⋅ t (3) = k ⋅ aq dt 8 7 6

ln aq*

5 4 3 2 1 0 0

20

40

60

80

100

120

140

t in min

Figure 7. Biosorption of Cu2+ on the Biocer material as a 1st order reaction

Further experiments shall investigate the influence of mass transfer processes onto the sorption kinetics. 4.

CONCLUSIONS The Biocer material investigated here was good suited for the removal of copper from aqueous solutions. In batch experiments, first measurements have shown a good sorption capacity as well as an almost complete regeneration of the material with citric acid. The Biocer material seems to withstand several regeneration cycles without any loss of its loading capacity or its stability. The runoff of the metal concentrates in a regeneration cycle can be possibly used for further rewinning processes. So, the material is good suited not only for the decontamination of waters with a low metal content, but also possibly for the recycling of economically valuable metals. In future experiments shall be investigated in particular the mass transport behaviour of metal ions at the Biocer material, as well as the biosorption of ions like arsenate and arsenite because of their significance as contaminants in practice. Especially interesting will also be the performance of the Biocer material in column experiments to investigate its practical application. It is expected to obtain excellent sorption reslts in the column investigations.

ACKNOWLEDGEMENTS This work was supported as a part of a scientific-technical cooperation project of the PTJ/ BMBF (German Federal Ministry for Education and Technology, Proj. Nr. 03i 4004a). 805

Biosorption

REFERENCES

1. 2. 3. 4. 5.

B. Volesky, Biosorption of heavy metals, CRC Press, Boca Raton (1990) S.J. Allen and P.A. Brown, J. Chem. Tech. Biotechnol. 62 (1995), 17 G.M. Gadd and C. White, Trends Biotechnol. 11 (1993), 353 M. Tsezos and B. Volesky, Biotechnol. Bioeng. 23 (1981), 583 D.N. Edgington, S.A. Gorden, M.M. Thommes and L.R. Almodovar, Limnol. Ocean. 15 (1970), 945 6. A. Hakajama, T. Horikashi and T. Sakaguchi, Eur. J. Appl. Microbiol. Biotechnol. 16 (1982), 88 7. L.E. Macaskie, J.D. Blackmore and R.M. Empson, FEMS Microbiol. Lett. 55 (1988), 157 8. M. Tsezos, R.G.L. Mac Cready and J.P. Bell, Biotechnol. Bioeng. 34 (1989), 10 9. G. Naja, C. Peiffert, M. Cathelineau and C. Mustin, Process metallurgy 9B, Eds. R. Amils, A. Ballester, Elsevier, Amsterdam, NL (1999), 343 10. B. Volesky, and M. Tsezos, Separation of uranium by biosorption, U.S. Pat. No. 4, 320,093 (1981), Canad. Pat. No. 1, 143,007 (1983) 11. E. Guibal, C. Roulph and P. Le Cloirec, Water Res. 26 (1992), 1139 12. L.E. Macaskie, R.M. Empson, A.K. Cheetham, C.P. Grey and A.J. Skarnulis, Science 257 (1992), 782 13. M.Z.C. Hu, J.M. Norman, N.B. Faison and M. Reeves, Biotechnol. Bioeng. 51 (1996), 237 14. T. Horikoshi, A. Nakajima and T. Sakaguchi, Agric. Biol. Chem. 332 (1979), 617 15. J.J. Byerley, J.M. Scharer and A.M. Charles, Chem. Eng. J. 36 (1987), B 49 16. N. Kuyucak and B. Volesky, Biosorption by algal biomass, CRC Press, Boca Raton, FL, (1990) 17. J. Chen and S. Yiacoumi, Sep. Sci. Technol. 32 (1997), 51 18. E. Guibal, C. Milot and J.M. Tobin, Ind. Eng. Chem. Res. 37 (1998), 1454 19. A.E.Baes, T.Okuda, W.Nishijima, E. Soto and M.Okada, Wat. Sci. Technol. 35 (1997), 89 20. J.H. Min and J.G. Hering, Wat. Res. 32 (1998), 1544 21. G.L. Rorrer, T.Y. Hsien and J.D. Way, Ind. Eng. Chem. Res. 32 (1993), 2170 22. L. Dambies, A. Roze, J. Roussy and E. Guibal, Process metallurgy 9B, Eds. R. Amils, A. Ballester, Elsevier, Amsterdam, NL (1999), 277 23. S. Selenska-Pobell, P. Panak, V. Miteva, I. Boudakov, G. Bernhard and H. Nitsche, FEMS Microbiology Ecology 29 (1999) 59 24. P.J. Panak, J. Raff, S. Selenska-Pobell, G. Geipel, G. Bernhard and H. Nitsche, Radiochim. Acta88 (2000), 71 25. U. Soltmann, J. Raff and S. Selenska-Pobell, J. Sol-gel Sci. Technol. 26 (2002), in press 26. B. Volesky, J. Weber and R. Vieira, Process metallurgy 9B, R. Amils, A. Ballester (eds.), Elsevier, Amsterdam, NL (1999), 473 – 481 27. A. Wilke, Dissertation, Berlin 2001 28. R. Kümmel, Adsorption aus wässrigen Lösungen, Dt. Verlag für Grundstoffindustrie, 1990, 25-28

806

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption of chromium (VI) by marine algal biomass L.H. Tan, J.P. Chen and Y.P. Ting* Department of Chemical and Environmental Engineering, National University of Singapore, Kent Ridge Crescent, Singapore 119260 Abstract The ability of local brown seaweeds Sargassum sp. and Padina sp. to remove Cr(VI) anions from solution was examined. Batch studies showed a higher maximum uptake capacity for Padina sp (1.04 mmol/g) than Sargassum sp. (0.60 mmol/g). The biosorption of chromium was pH dependent; higher uptake was observed as the pH decreased, with the optimal uptake occurring at about pH 2. Over an initial Cr(VI) concentration of 0.38mmol/L to 10.6 mmol/L, the equilibrium adsorption data could be modeled according to the Freundlich and Langmuir adsorption isotherms. Both the algal biomass achieved about 90% of the maximum uptake capacity in 6-7 hours and attained equilibrium in 10 hours. Apart from the removal of Cr(VI) from solution by the biosorbent, a complete reduction of Cr(VI) to Cr(III) at low concentration occurred during the biosorption process. This offers an advantage over conventional physical/ chemical treatment methods where chromium reduction and removal are accomplished separately. 1.

INTRODUCTION The widespread use of heavy metals in various industries has created environmental problems due to their hazardous nature and non-biodegradability. Chromium in industrial effluent is of major concern; the heavy metal is primarily present in industrial effluent as anion in the hexavalent form as chromate (CrO42-) and dichromate (Cr2O72-) ions [1]. Compared with the relatively less harmful and mobile Cr(III), the hexavalent form of chromium is considered more toxic[2]. Conventionally, chemical precipitation, coagulation and flocculation, reverse osmosis and ion-exchange are some of the method commonly used in chromium treatment [3]. Chromium removal may involve several steps: the reduction of the Cr(VI) to trivalent form, the precipitation of Cr(III) as a metal hydroxide at high pH, followed by the settling and disposal of the dewatered sludge [4]. However, these processes have many drawbacks, including the generation of toxic sludge and high operational cost and incomplete reduction of Cr(VI). To overcome these shortcomings, biosorption appears to be an alternative. The use of cheap and abundant biomass (either living or dead) has attracted much attention in the past decades. The biosorption process does not generate toxic chemical sludge, yet it is cost

* Corrresponding author: [email protected]

807

Biosorption

effective especially for the treatment of low concentration metal-bearing wastewater. The use of algal biomass for the chromium-containing wastewater treatment has been reported. A maximum removal of about 14.7×103 mg metal/kg of dry weight biomass by filamentous algae Spirogyra species was reported by Gupta et. al. (2001). Chlorella vulgaris [5-6], Scenedesmus obliquus [6], Synechocystis sp. [6] and many other species of algal biomass have also been investigated for their biosorptive ability for chromium. Marine algal biomass, Sargassum sp. and Padina sp. are investigated in this study in view of their promising results for the biosorption of various heavy metals [7-10]. The biosorption of chromium by Padina sp. has not been previously reported. Factors affecting the biosorption efficiency were investigated in this study. These included a kinetic study on biosorption and reduction of chromate, as well as the effect of pH on the removal of chromium. Equilibrium studies using the Langmuir and Freundlich adsorption isotherms were also examined. 2.

MATERIALS AND METHODS

2.1 Preparation of the biomass The brown algal biomass Sargassum sp. and Padina sp. were harvested locally, and rinsed with copious quantity of water to remove all attached materials. The seaweed was sun-dried and ground into particles of various sizes; biomass of particle size 212-500µm was used in this study. The biomass was then washed with deionised distilled water and dried at 60°C overnight and stored in a dry cabinet before use. 2.2 Kinetic study A kinetic study was conducted in order to determine the time required for the biosorption process to reach equilibrium. The biomass (6.0g) was added to 2L chromate solution (1 mM) prepared using K2Cr2O7 (Merck Chemical). The pH of the solution was adjusted to pH 2, 4 or 6 initially using HCl or NaOH solutions, and was monitored continuously. A 4-mL sample was taken at predetermined time interval; the samples were filtered using a Whatman Autovial, with 0.45µm PTFE filter. The total chromium and Cr(VI) were analysed using ICP-ES and a colorimetric method 3500-Cr D respectively [11]. The Cr(III) concentration was obtained from the difference between these two analyses. 2.3 pH effect on chromium sorption 150 mg of the biomass was added to each conical flask containing 50 mL of chromate (Cr(VI)) solution and Cr(III) solution (prepared from Cr(NO3)3.9H2O), with initial concentration of 1 mM and pH over the range 1 to 8 and 1 to 5.5 (using NaOH or HCl). The mixtures were agitated on a rotary shaker at 150 rpm at ambient temperature (24 ± 1°C). After 12 hours of agitation, the supernatant of the solution was filtered and analysed for total chromium, Cr(VI), and Cr(III) concentration. The final pH of the solution was measured using ORION-420A. The total chromium uptake was calculated using the equation below: (2.1) Q = (Ci – Cf) V/ m where Ci and Cf are the initial and final total chromium concentration, V is the volume of the solution (L) and m is the amount of biomass (g).

808

Biosorption

2.4 Analysis of the adsorption isotherms 150 mg of biomass was contacted with 50 mL of chromate solution at an initial concentration ranging from 0.4-10.6 mmol/L, and adjusted initially to pH 2. The mixtures were agitated on a rotary shaker at 150 rpm at ambient temperature (24 ± 1°C). The pH was adjusted from time to time. The supernatant was filtered after 12 hours using a Whatman Autovial, with a 0.45µm PTFE filter, and analysed for total chromium concentration. Two commonly used adsorption models, the Langmuir and Freundlich isotherms were used to evaluate the experimental data. The Langmuir isotherm has the general form:

Q =

Q max bC e 1 + bC e

(2.2)

where Q = total chromium uptake; Qmax = maximum uptake of total chromium; b = affinity of sorbate for the sorbent and Ce = the equilibrium or final concentration of chromate solution. The Freundlich isotherm has the general form: (2.3) Q = KCe1/n where Q = total chromium uptake. K = biosorption equilibrium uptake and is indicative of the biosorptive uptake capacity, n = biosorption equilibrium constant, and Ce = the equilibrium or final concentration of chromate solution. 3.

RESULTS AND DISCUSSION

3.1 Kinetic study The kinetics of sorption describing the solute uptake rate which in turn governs the time of sorption reaction is one of the important characteristics defining the efficiency of sorption [12]. In order to determine the time required for the biosorption process to reach equilibrium, a kinetic study was carried out. From the experimental results (Figure 1a-b), both biomass achieved about 90% of the maximum uptake capacity in 6-7 hours. No further removal was observed after 10 hours. Similar results have also been reported by Luis et. al. [13]. Several studies on chromium uptake had assumed that an equilibrium is attained within 6 hours of biosorption [13, 14]. In our work, a contact time of 12 hours was used to ensure that the system has reached equilibrium. Kinetic studies at pH 4 and 6 were also conducted under the same condition. At higher pH, less time was needed for the sorption process to reach equilibrium (data not shown). An examination of the results showed that the removal of the hexavalent chromium is a coupled process of sorption and reduction at a different rate. Complete reduction of Cr(VI) to Cr(III) was observed during the biosorption process at pH 2. Padina sp. reduced chromate at a faster rate than Sargassum sp., where complete reduction was achieved within 8 hours. Indeed, chromate in the solution was reduced as soon as the seaweed was in contact with the solution (see Figure 2). It is clear that the instantaneous reduction is faster for Padina sp. than for Sargassum sp.; at 15 minutes, the ratio of Cr(VI):Cr(III) was 9.2:1 and 16.8:1 respectively. To reduce the ratio of Cr(VI):Cr(III) to 1.5, Padina sp. takes 3 hours whereas Sargassum sp. takes about 7 hours. Both the reduction processes occur until all the chromate was reduced to the trivalent form. This is in contrast with results reported by Hayashi et. al. (2001), who found that the Cr(VI)/Cr(III) ratio in the solution increased gradually from 1.5 to 2.5 over the time interval 30 to 150 minutes. The increase 809

Biosorption

in Cr(VI)/Cr(III) ratio was attributed to the possible sorption of Cr3+ species by the biomass. (b) (a) 1,20 Chromium Concentration (mmol)

Chromium Concentration (mmol)

1,20 1,00 0,80 0,60 0,40 0,20 0,00

1,00 0,80 0,60 0,40 0,20 0,00

0

5

10

15

20

25

30

0

5

10

Time (h) Cr(III)

Cr(VI)

15

20

25

30

Time (h) Cr(III)

Cr(Total)

Cr(VI)

Cr(total)

Figure 1: Progress of biosorption and reduction of (a) Padina sp. (b) Sargassum sp. (Biomass dosage = 3.0 g/L, Initial metal concentration = 1 mM, pH = 2, Chromium was present initially as Cr(VI))

Cr(VI):Cr(III) Ratio

20 15 10 5 0 0

5

10

15

20

25

Time (h) Padina sp.

Sargassum sp.

Figure 2. Reduction in Cr(VI):Cr(III) during biosorption by Padina sp. and Sargassum sp.. (Biomass dosage = 3.0 g/L, Initial metal concentration = 1 mM, pH = 2)

To examine the effect of pH on chromium uptake by the seaweed, Cr (VI) and Cr(III) (i.e., K2Cr2O7 and (Cr(NO3)3.9H2O) were used in order to compare the binding characteristic of both biomass to chromium with different oxidation state. Cr(III) uptake was favored at a higher pH (see Figure 3a-b). At pH 2, the uptake of Cr(III) was lower compared with that of the system with chromium present initially as Cr(VI). Thus, the gradually decrease in the Cr(VI):Cr(III) ratio (at pH 2) noted earlier maybe due to the less significant uptake of Cr(III). It is noteworthy that the Padina sp. showed comparable reduction ability with that of treated Sargassum biomass reported in Luis et al. [13]. 3.2 pH effects on the biosorption In the pH effects, the biomass was contacted with chromate solution with different pH. The uptake of chromium increases with a decrease in pH up to 2, and the uptake decrease at pH 1. The results indicate that pH 2 is the optimum for the biosorption of 810

Biosorption

chromium (Figure 3a-b), where the total chromium sorbed onto the biomass is about 62% and 50% for Padina sp. and Sargassum sp. respectively. Luis et al. [13], Kratochvil et al. [14], Gupta et al. [15] and Niu et al. [16] have earlier noted a similar optimum for chromium biosorption at pH 2. The similar trend in the effect of pH on the biosorption of chromium has been reported for a variety of biomass including Rhizopus nigricans [3], Scenedesmus obliquus, Synechocystis sp. [6], Chlorella vulgaris, Clodophara crispata, Rhizopus arrhizus, Saccharomymes cerevisiae, Zoogloea ramigera [17] and dried activated sludge [18]. It is well known that pH is an important parameter that influences the biosorption process [19]. Interactions between the metal ions and the functional groups of the biomass Sargassum sp. and Padina sp. depend not only on the nature of the biosorbent but also on the solution chemistry of the metals to be removed [3]. In this study, the solution chemistry of chromium is predicted using a chemical speciation prediction program, MINEQL (Figure 4). At pH 2, HCrO4- is the predominant species of chromium in solution. Depending on the extent of protonation on the seaweed, the chromate anions, HCrO4- and Cr2O72- that are likely to be present in the solution would be attracted to the positively charged functional group on the seaweed. A decrease in pH will result in a more positive charge on the surface of the seaweed.

100

80

80

60

60

40

40

20

20

0

0

100

100

80

80

60

60

40

40

20

20

0

0 1 2 3 4 5 6 7 8 9

0 0

1

2

3

pHf % Chromium Adsorbed

% Cr(III) Removal

100

% Chromium Adsorbed

(b)

% Cr(III) Removal

% Chromium Adsorbed

(a)

4

5

6

7

8

9

pHf % Chromium Adsorbed

% Cr(III) Removal

% Cr(III) Removal

Figure 3. pH effects on the removal of Cr(III) and Cr(VI). (a) Padina sp. (b) Sargassum sp. (Biomass dosage = 3.0 g/L, Initial metal concentration = 1 mM, pH = 2, Contact time = 12h) 100

HCrO4-

% species

80

CrO42-

60 40

H2CrO4

20

Cr2O72-

0 0

2

4

6

8

10

12

pH

Figure 4. Speciation of Cr(VI) predicted by MINEQL (Cr(VI) concentration = 1mM) 811

Biosorption

SOH + H+ → SOH2+ -

(at low pH)

+

(at high pH) SOH → SO + H The increase in the biosorption of Cr(VI) at the lower pH thus suggests that the negatively charged chromium species bind through electrostatic attraction to the positively charged functional groups on the biomass surface [3, 6,15,18]. The existence of the optimum pH was explained by the desorption of Cr(III) from the biomass at low pH and the effect of pH on the reduction potential of Cr(VI) in aqueous solutions [14]. Figure 5 shows that total removal of chromate by reduction occurred at lower pH, especially at pH 1-1.5. At these pH, the reduction of anionic Cr(VI) species to Cr(III) dominate the system. Further more, the sorption of Cr3+ is also not favored because positively charge hydrogen ions will compete with metal ions for the ligands on the cell wall of biomass This is also supported by the experimental results where the uptake of Cr(III) by the biomass at pH 1 is less than 20% of total chromium removal (Figure 3a-b). The Cr(VI) reduction during chromate biosorption has also been observed by the other workers. The reduction of chromate is greater at a lower pH. Figure 5 shows that the reduction of chromate is greatest at pH 1-2 in the Padina sp. system, and at pH 1-1.5 in the Sargassum sp. system. The observations that the reduction of chromate is lower in higher pH is also consistent with the calculations by Kratochvil et. al. [14] using the Nernst equation which showed that the redox potential of chromate was greater at lower pH.

% Cr(VI) Removal

100 80 60 40 20 0 1

2

3

4

5

6

7

8

pHf Padina sp.

Sargassum sp.

Figure 5. Cr(VI) removal by Padina sp. and Sargassum sp. as a function of equilibrium pH. (Biomass dosage = 3.0 g/L, Initial metal concentration = 1 mM, Contact time = 12h) 3.3 Langmuir and Freundlich adsorption isotherm The adsorption isotherms of the biomass are illustrated in Figures 6-7. The experimental results are plotted using linearized Langmuir and Freundlich adsorption isotherms over a concentration range of 0.4-10.6 mmol/L. As shown in Table 1, the experimental data are consistent with both the isotherms. Langmuir model is a theoretical model for monolayer adsorption while the Freundlich model allows for multilayer adsorption at heterogeneous surfaces. The conformity of the experimental results to both the models shows that the sorption of chromate onto the biomass is complex and may involve more than one mechanism [12]. A meaningful 812

Biosorption

physical interpretation thus may not be drawn; the sorption isotherms do not necessarily reflect the adsorption mechanisms involved [20-21]. Table 1. Freundlich and Langmuir model regression constants of the biomass Freundlich isotherm constants K n R2

Type of biomass Padina sp. Sargassum sp.

0.329 0.221

1.912 2.228

Langmuir isotherm constants Qmax b R2 (mmol/g) (L/mmol) 1.04 0.568 0.9878 0.60 0.679 0.9890

0.9673 0.9814

The Langmuir parameters, Qmax and b represent the maximum uptake and the affinity between the sorbate and sorbent respectively. A high value of b indicates a (desirable) steep beginning of the isotherm, which reflects the high affinity of the biosorbent for the sorbate [7]. In comparison with Sargassum sp., Padina sp. has a higher uptake capacity although its affinity constant, b is marginally lower than Sargassum sp. From the Freundlich Isotherm, the adsorption capacity, K of Padina sp. and Sargassum sp. is 0.329 and 0.221 respectively. Again, it is shown that the ability of Padina sp. to adsorb chromium ion is greater than Sargassum sp. In both cases, the intensity of adsorption, n, is greater than one, thus indicating that the adsorption of the chromium ion is favorable [3, 12]. 0 -1

0 0

1

2

-2

3

-1

0

-0,5

-0.5

-1

-1 ln Q

ln Q

-2

-1,5

-2

-2,5

-2.5

3

-3 ln Ce

Padina sp.

2

-1.5

-2

-3

1

Sargassum sp.

ln Ce

Padina sp.

Sargassum sp.

Figure 6. Freundlich isotherm over a Figure 7. Langmuir isotherm over a concentration range of 0.4-10.6 concentration range of 0.4-10.6 mmol/L (pH=2.0; Biomass dosage= mmol/L (pH=2.0; Biomass dosage= 3g/L) 3g/L) + Compared with H -protonated and Ca-treated Sargassum biomass [13-14], the uptake of chromium by both the biomass reported in this study is lower. This may due to the lower extent of protonation in the untreated biomass. As reported, the unique mixture of polysaccharides (mainly alginate and fucoidan) is largely responsible for the excellent metal sequestering ability of the brown algae. Carboxylate groups of alginate have been identified as the main metal binding site. Besides that, other negatively charged functional groups such as the sulphonate groups of fucoidan may also contribute to heavy metal complexation [7]. Even though the functional groups will be protonated at lower pH, it is believed that pre-treatment with acids will enhance the protonation, and thus increase the uptake of heavy metal. 813

Biosorption

4.

CONCLUSIONS The removal of chromium (VI) is a coupled process with biosorption and reduction. The process is pH dependent, and an optimum uptake occurs at pH 2. This is related to the speciation of chromium in the solution and the extent of protonation on the biomass. As most electroplating units discharge chromium in acidic solution [13, 22], minimal pH adjustment is required when biosorption is applied. Results from equilibrium studies showed that the Langmuir and Freundlich isotherms fitted the data well. Compared with Sargassum sp., Padina sp. was found to have higher adsorption capacity and reduction ability for chromium.

ACKNOWLEDGEMENTS This work was funded by the National University of Singapore research grant R-279000-123-112. REFERENCES

1. Handan, U., Y. K. Bayhan, Y. Kaya, A. Cakici and Ö. F. Algur. Biosorption of Chromium(VI) from Aqueous Solution by Cone Biomass of Pinus sylvestris, Bioresource Technology, 85, pp. 155-158. 2002. 2. Carlos, C., J. Campos-García, S. Devars, F. Gutiérrez-Corona, H. Loza-Tavera, J.C. Torres-Guzmán and R. Moreno-Sánchez. Interactions of Chromium with Microorganisms and Plants, FEMS Microbiology Reviews, 25, pp. 335-347. 2001. 3. Sudha, B.R. and T.E. Abraham. Biosorption of Cr (VI) from Aqueous Solution by Rhizopus nigricans, Bioresource Technology, 79, pp. 73-81. 2001. 4. Hayashi, A.M., W.B. Amorim, D.M. Pereira, P.F. Pimentel and M.G.C. da Silva. Biosorption of Cr(VI) In Algae Biomass: Kinetic Study IN Biohydrometallurgy: Fundamentals, Technology and Sustainable Development: Paper of the 14th International Biohydrometallurgy Symposium, 2001, New York: Elsevier, pp. 199206. 5. Aksu, Z. and Ü. Açikel. A Single-staged Bioseparation Process for Simultaneous Removal of Copper(II) and Chromium(VI) by Using C. vulgaris, Process Biochemistry, 34, pp. 589-599. 1999. 6. Dönmez, G.Ç., Z. Aksu, A. Öztürk and T. Kutsal. A Comparative Study on Heavy Metal Biosorption Characteristics of Some Algae, Process Biochemistry, 34, pp. 885892. 1999. 7. Davis, T.A.B., Volesky and R.H.S.F. Vieira. Sargassum Seaweed As Biosorbent for Heavy Metals, Water Research, 34, pp. 4270-4278. 2000. 8. Jalali, R., H. Ghafourian, Y. Asef, S. J. Davarpanah and S. Sepehr. Removal and Recovery of Lead Using Nonliving Biomass of Marine Algae, Journal of Hazardous Materials, 92, pp. 253-262. 2002. 9. Pairat, K. Biosorption of Copper(II) from Aqueous Solutions by Pre-treated Biomass of Marine Algae Padina sp., Chemosphere, 47, pp. 1081-1085. 2002. 10. Pairat, K and Q. Yu. Cadmium(II) Removal from Aqueous Solutions by Pre-treated Biomass of Marine Alga Padina sp., Environmental Pollution, 112, pp. 209-213. 2001. 11. Andrew, D. E., L. S. Clesceri and A. E. Greenberg. Standard methods for the examination of water and wastewater. pp. 3.59-3.60, Washington, D.C.: American Public Health Association, 1995.

814

Biosorption

12. Dönmez, G. and Z. Aksu. Removal of Chromium(VI) from Saline Wastewaters by Dunaliella species, Process Biochemistry, 38, pp. 751-762. 2002. 13. Luis, K. C., R.C. Agapay; J.L.L Rakels, M. Ottens, L.A.M. Van der Wielen. Potential of Biosorption for The Recovery of Chromate in Industrial Wastewaters, Industrial and Engineering Chemistry Research, 40, pp. 2302-2309. 2001. 14. Kratochvil, D., P. Pimentel and B. Volesky. Removal of Trivalent and Hexavalent Chromium by Seaweed Biosorbent, Environmental Science and Technology, 32, pp. 2693-2698. 1998. 15. Gupta, V. K., A. K. Shrivastava and N. Jain. Biosorption of Chromium(VI) from Aqueous Solutions by Green Algae Spirogyra species, Water Research, 35, pp. 40794085. 2001. 16. Niu, H. and B. Volesky. Biosorption of Anionic Metal Species, In: Biohydrometallurgy: Fundamentals, Technology and Sustainable Development: Paper of the 14th International Biohydrometallurgy Symposium, 2001, New York: Elsevier, pp. 189-197. 2001. 17. Nourbakhsh, M., Y. Sağ, D. Özer, Z. Aksu, T. Kutsal and A. Çağlar. A Comparative Study of Various Biosorbents for Removal of Chromium(VI) Ions from Industrial Waste Waters, Process Biochemistry, 29, pp. 1-5. 1994. 18. Aksu, Z., Ü. Aç1kel, E. Kabasakal and S. Tezer. Equilibrium Modelling of Individual and Simultaneous Biosorption of Chromium(VI) and Nickel(II) Onto Dried Activated Sludge, Water Research, 36, pp. 3063-3073. 2002. 19. Veglio, F. and F. Beolchini. Removal of Metals by Biosorption: A Review, Hydrometallurgy, 44, pp. 301-316. 1997. 20. Kratochvil, D. and B. Volesky. Advances in The Biosorption of Heavy Metals, Trends in Biotechnology, 16, pp. 291-300. 1998. 21. Volesky, B. and Z. R. Holan. Biosorption of Heavy Metals, Biotechnology Progress, 11, pp. 235-250. 1995. 22. Sharma, D.C. and C. F. Forster. Removal of Hexavalent Chromium Using Sghagnum Moss Peat, Water Research, 27, pp. 1201-1208. 1993.

815

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption of heavy metal ions from aqueous solutions by local seaweeds P.X. Sheng, J.P. Chen and Y.P. Ting* Department of Chemical & Environmental Engineering, National University of Singapore, Kent Ridge Crescent, Singapore 119260 Abstract The uptake capacities of the biomass of four seaweeds Sargassum sp., Padina sp., Ulva sp., and Gracillaria sp. (collected from Singapore coasts) for heavy metal ions (lead, copper, cadmium, nickel and zinc) were evaluated in this study. The metal ion removal rates were very rapid, with 90% of adsorption taking place within 10-60 min. The adsorption capacities of the biomass were strongly dependent on equilibrium solution pH; higher pH favored metal ion removal. Equilibrium experiments were carried out at pH 5.0 (for lead and copper) and pH 5.5 (for cadmium, zinc and nickel). Sargassum sp. and Padina sp. were found to have the highest potential in the removal of lead, copper, cadmium, nickel and zinc ions. The maximum uptake capacities ranged from 0.61 to 1.16mmol/g for Sargassum sp. and 0.63 to 1.25mmol/g for Padina sp. 1.

INTRODUCTION Heavy metal contamination is of worldwide environmental concern especially in developing countries. Conventional methods for removing heavy metals from industrial effluents (e.g. precipitation and sludge separation, chemical oxidation or reduction, ion exchange, reverse osmosis, membrane separation, electrochemical treatment and evaporation) are often ineffective and costly when applied to dilute effluents. A good sorbent to remove heavy metal should be both effective and inexpensive. Biosorption shows promise of fulfilling these requirements. Since biosorbents are essentially dead materials, no nutrition is needed to maintain the biomass. Problems associated with metal toxicity in living biomass and the need to provide suitable growth condition also do not arise. Indeed, many early studies have shown that nonliving biomass may be even more effective in sequestering metallic elements than living cells [1, 2]. The first major challenge for biosorption is the selection of the most promising types of biomass from an extremely large pool of readily available and inexpensive biomaterials. Many types of biomass in non-living form have been studied for their heavy metal uptake capacities and suitability for use as bases for biosorbent development. These include bacteria [3], fungi [4], yeast [5], fresh water algae [1] and marine algae [6, 7]. Several have focused on marine algae due to its easy availability and high uptake capacity

* Corresponding author: [email protected]

817

Biosorption

[8, 9]. The capacities of a few brown algae have been found to be much higher than those of other types of biomass, and even activated carbon, natural zeolite and synthetic ion exchange resins [4, 10]. For divalent heavy metal ions, Sargassum spp. [11, 12], Rhizopus spp. [12, 13], Durvillaea spp. [14, 15], Ecklonia spp. [14,15] and Lessonia spp. [15] showed very good biosorption performances among the marine algae studied. This study compares the metal biosorption performances of several local seaweeds which are abundant in Singapore seashores. This includes two brown seaweeds (Sargassum sp., Padina sp.), a green seaweed (Ulva sp) and a red seaweed (Gracillaria sp.). The choice of heavy metals (i.e. lead, copper, cadmium, zinc and nickel) was made with regard to their common industrial use and potential pollution impact. 2.

MATERIALS AND METHODS Preparation of biomass and chemicals The raw biomass of Sargassum sp., Padina sp., Ulva sp., Gracillaria sp. were collected from the Singapore coasts. The sun-dried biomass was ground to various particle sizes, from which particles of 500-800 µm were used in this work. The biomass were washed with DI water and dried at about 60°C overnight before use in the experiments. The stock metal solutions at various concentrations were prepared by dissolving lead nitrate, copper nitrate, cadmium sulfate, nickel nitrate and zinc nitrate respectively. All metal salts were of reagent grades. Batch adsorption experiments In the equilibrium experiments to study the pH effect, metal solutions at various initial pH were prepared. pH was adjusted using 0.1mol/l HNO3 or 0.1mol/l NaOH. The biomass was added into the conical flasks containing the metal solutions. The flasks were agitated at 200 rpm for 6 hours. The experiments were conducted at room temperature (22±1°C). The final solution pH was measured with an ORION 525A pH meter. The metal concentrations were analysed using an inductively coupled plasma emission spectrophotometer (ICP-ES) (Perkin-Elmer Optima 3000). In the kinetic experiments, the initial pH was adjusted to the selected optimum value before the biomass was added to the solutions while stirring. The pH was measured at 2030 minute intervals and adjusted accordingly. Samples were taken at periodic time intervals and the metal concentrations were analyzed. In the isotherm experiments, the solution pH was adjusted as in the kinetic experiments. The same amount of biomass was added to the solutions at various metal concentrations. All bottles were shaken at 200 rpm at room temperature. The initial and final metal concentrations were determined using the ICP-ES. Biosorption metal uptake (q) was calculated from the sorption system mass balance: q = V(C i − C f ) S

where V is the volume of the solution, S is the amount of biomass, and C i and C f are the initial and final metal concentrations respectively. The Langmuir adsorption isotherm was used to fit the experimental data:

(

q = (q max ⋅ C f ) b −1 + C f

)

q max and b are Langmuir constants, which reflect the maximum metal adsorption capacity and affinity between metal ion and biosorbent. 818

Biosorption

RESULTS AND DISCUSSION Effect of solution pH Figure 1 summarizes the results obtained using the chemical equilibrium program MINEQL [16] for the lead nitrate, copper nitrate, cadmium sulfate, nickel nitrate and zinc nitrate systems. All data sets were calculated considering the carbonate system naturally in equilibrium with atmospheric carbon dioxide (PCO2=10-3.5atm). For the lead nitrate system, Pb2+ is the dominant species present at lower pH. In this experimental system (maximum concentration of 2.0 mmol/l), Pb2+ remains the dominant species up to about pH 5.5. At pH higher than 5.5, solid lead hydroxide is thermodynamically the most stable phase. In addition, the effect of the influence of low concentrations of sodium and nitrate (present from pH adjustment during the sorption experiment) on the speciation was negligible. The same approach was applied to calculate the other four metal ion solution systems at a concentration of 2.0mmol/l. Cu2+ is the dominant species present up to pH 5.2 in the copper nitrate system. For the cadmium sulfate, zinc nitrate and nickel nitrate systems, the free metallic ions species are dominant at pH lower than 6.0. It is well documented that solution pH is an important parameter affecting biosorption of heavy metal ions [5, 17, 18]. Heavy metal ions (lead, copper, cadmium, nickel and zinc) adsorption by the seaweeds (Sargassum sp., Padina sp., Ulva sp., Gracillaria sp.) as a function of pH was studied. Only results for lead are shown here (Figure 2). On the whole, the uptake of metal ions increased sharply from pH 2 to 4.5; beyond pH 4.5, its increasing effect on uptake was reduced. These finding are in agreement with results reported earlier [17, 18]. 100

1.0

80

0.8

60

0.6

-1

q (mmol.g )

% of free metal ion species in solution

3.

40

Copper Nitrate Lead Nitrate Cadmium Nitrate Zinc Sulfate NIckel Nitrate

20

3

4

5

Padina sp. Sargassum sp. Ulva sp. Gracillaria sp.

0.2

0 2

0.4

6

7

8

pH

Figure 1. Free metal ion species in solution (Co = 2mmol/l, Calculated using MINEQL)

0.0 2

3

4

5

6

7

pH

Figure 2. pH effect on equilibrium of Lead uptake (m = 1.0g/l, Co = 1mmol/l)

The pH dependence of metal uptake is largely related to the surface functional groups (mainly carboxylic) [9] in the biomass cell wall and also on the metal chemistry in solution. Since the metals are present in their free ionic form (Figure 1) at pH less than 4.5, the sharp increase in metal adsorption from pH 2 to 4.5 cannot be explained by the change in metal speciation. This implies that the type and ionic state of the cell wall functional groups at these pHs determine the extent of sorption. As positively charged ions, hydrogen ions may compete with metal ions for the ligands on the cell wall. At lower pH, the concentration of hydrogen ions is higher, which leads to less ligands availability 819

Biosorption

on the biomass for metal ions sorption. As the pH is increased, more ligands are available for metal ions, thus resulting in an enhanced metal ion removal. In order to ensure that the metal ions were in their free ionic states during biosorption processes, the pH of the following kinetic and isotherm experiments were controlled at 5.0 for lead and copper, and at 5.5 for cadmium, zinc and nickel. Determination of equilibrium time Kinetic experiments were carried out to determine the equilibrium time required for the uptake of metal ions by different biomass. Only results for lead are shown here (Figure 3). In general, a two-stage kinetic behavior is seen: very rapid initial sorption for a few minutes, followed by a long period of a much slower uptake (especially for Padina-metal and Sargassum-metal systems). It is known that biomass cell walls are heterogeneous [9]. Various functional groups serve as adsorption sites that differ both with respect to the strength of the metal sorptive bond and the rate of adsorption on these sites. Hence this results in the classification of fast and slow uptake rates for the same metal ions. The results show that equilibrium times needed for the different metal-biomass systems ranged from 1 to 3 hours, with 90% of the total adsorbed metal ions occurring within 10-60 minutes. The contact time for the following isotherm experiments was set at 6 hours in order to assure the uptake equilibrium.

0.8

-1

q (mmol.g )

0.6

0.4 Padina sp. Sargassum sp. Ulva sp. Gracillaria sp.

0.2

0.0 0

100

200

300

400

500

time (min)

Figure 3. Kinetic experiments of Lead uptake (m = 1.0 g/l, Co = 1 mmol/l, pH = 5.0)

Adsorption Equilibrium Isotherm experimental results are shown in Figure 4a-4e. In all cases favorable isotherms are observed and the data could be modeled according to the Langmuir adsorption isotherm. Table 1 shows the maximum adsorption capacity (qmax) and affinity constant (b). Among the biomass screened, Sargassum sp. and Padina sp. are identified to be good biosorbents for removal of all the metal ions investigated. The equilibrium isotherms of Pb2+ on the studied biomass in their non-pretreated forms are shown in Figure 4a. The maximum adsorption capacities of Padina sp., Sargassum sp., Ulva sp., and Gracillaria sp. for lead were 1.16, 1.14, 0.83, 0.41 mmol/g respectively at the final concentration of about 1.0mmol/l. Table 1 shows that Padina sp. and Sargassum sp. exhibited the highest capacity and adsorption affinity (initial slope) with the maximum adsorption capacities at 1.25 and 1.16 mmol/g respectively. Ulva sp. sequestered more lead at higher residual concentrations than Gracillaria sp., while its affinity for lead is less than that of Gracillaria sp.

820

1.2

1.2

1.0

1.0

0.8

0.8

-1

q (mmol.g )

-1

q (mmol.g )

Biosorption

0.6

Padina sp. Sargassum sp.

0.4 Ulva sp. Gracillaria sp. Langmuir isotherm

0.2 0.0 0.0

0.3

0.6

0.9

1.2

1.5

0.6 Padina sp. Sargassum sp. Ulva sp. Gracillaria sp. Langmuir isotherm

0.4 0.2 0.0 0.0

1.8

0.3

0.6

-1

0.9

1.2

1.5

1.8

2.1

-1

Cf (mmol.L )

Cf (mmol.L )

Figure 4a. Equilibrium isotherm for Lead uptake (m=1.0g/l, pH=5.0)

Figure 4b. Equilibrium isotherm for Copper uptake (m=1.0g/l, pH=5.0) 0.8

0.8

0.6 -1

q (mmol.g )

Sargassum sp. Padina sp.

-1

q (mmol.g )

0.6

0.4

0.2

0.0 0.0

0.6

0.9

1.2

1.5

1.8

2.1

Padina sp. Sargassum sp. Ulva sp. Gracillaria sp. Langmuir isotherm

0.2

Ulva sp. Gracillaria sp. Langmuir isotherm

0.3

0.4

0.0 0.0

2.4

0.3

0.6

-1

0.9

1.2

1.5

1.8

2.1

2.4

-1

Cf (mmol.L )

Cf (mmol.L )

Figure 4c. Equilibrium isotherm for Cadmium uptake (m=1.0g/l, pH=5.5)

Figure 4d. Equilibrium isotherm for Zinc uptake (m=1.0g/l, pH=5.5)

-1

q (mmol.g )

0.6

0.4 Sargassum sp. Padina sp.

0.2

0.0 0.0

Ulva sp. Gracillaria sp. Langmuir isotherm

0.3

0.6

0.9

1.2

1.5

1.8

2.1

-1

Cf (mmol.L )

Figure 4e. Equilibrium isotherm for Nickel uptake (m=1.0g/l, pH=5.5) The sorption performance of the sorbent is manifested through the two Langmuir parameters: the maximum adsorption capacity qmax; and the affinity constant b. In biosorption, both a high qmax and b are desired. In the uptake of lead by Ulva sp., it can be 821

Biosorption

seen that although the seaweed possessed the highest qmax, unfortunately it has the lowest affinity constant amongst all the biosorbents (see Table 1). The biosorption potentials of the four seaweeds for the other four metal ions (Cu2+, 2+ Cd , Zn2+ and Ni2+) were further evaluated as shown in Figure 4b-4e and Table 1. The general shapes of isotherms were similar to that of lead. However, the adsorption capacities and affinities were significantly different. In general, the brown seaweeds (Padina sp. and Sargassum sp.) showed a better performance than the green seaweed (Ulva sp.) and red seaweed (Gracillaria sp.) for removal of all the metals investigated; for the same seaweed studied, the uptake capacity of lead was the highest, followed by copper, cadmium, zinc and nickel. Table 1. Parameters for Langmuir isotherms Pb

Cu 2

Padina sp. Sargassum sp. Ulva sp. Gracillaria sp.

qmax (mmol/g) 1.25 1.16 1.46 0.45

b (l/mM) 9.31 14.23 1.11 6.99 Cd

r 0.97 0.95 0.91 0.90

qmax (mmol/g) 1.14 0.99 0.75 0.59

b (l/mM) 8.39 8.78 4.14 10.25 Zn

r2 0.99 1.00 0.99 0.94

Padina sp. Sargassum sp. Ulva sp. Gracillaria sp.

qmax (mmol/g) 0.75 0.76 0.58 0.30

b (l/mM) 5.65 11.34 1.45 21.11 Ni

r2 1.00 0.94 0.98 0.81

qmax (mmol/g) 0.81 0.50 0.54 0.40

b (l/mM) 2.57 13.63 1.15 12.68

r2 0.99 0.95 0.97 0.83

Padina sp. Sargassum sp. Ulva sp. Gracillaria sp.

qmax (mmol/g) 0.63 0.61 0.29 0.28

b (l/mM) 1.98 4.69 1.58 9.71

r2 0.99 0.99 0.98 0.86

These uptake differences in isotherms may be related to the compositional differences among the algal biomass, as well as the binding mechanisms involved for different heavy metal ions. The biomass cell walls are known to be heterogeneous. For instance, the cell wall of brown algae contain algin, fucoidan and cellulose, while in most green algae, the outer part of the cell wall consists mainly of pectic substances and cellulose [9]. The mixture of polysaccharides, mainly alginate and fucoidan, is largely responsible for the excellent metal sequestering ability of the brown algae [1]. This may explain the reason that the biosorption performances of brown algae generally are superior to those of green algae. The alginate responsible for metal sorption is present in a gel form in the cell walls [6]. The cell walls of algae are often porous in their structure, which allows molecules and ions to pass freely through [19]. In addition to the porosity of the algal cell wall structure, the cell constituents can provide an array of ligands and functional groups, which bind metallic ions [9]. Numerous metal-binding mechanisms have been postulated in biosorption. These include chemisorption by ion exchange, complexation, coordination, chelation, physical adsorption and microprecipitation [20]. Oxidation/reduction reactions also may be involved [21]. Due to the complexity of the composition of the biomaterial, it is likely that some of these mechanisms are acting simultaneously to varying degrees. 822

Biosorption

Several studies have indicated a dominant role of ion exchange for metal biosorption [16, 22]. Further studies are needed in understanding the interaction behaviors between the biomass and heavy metal ions The binding strength of various divalent ions sequestered also is important in the uptake process [9]. The general affinity sequence for Padina sp. is Pb>Cu>Cd>Zn>Ni; and for Sargassum sp. is Pb>Zn>Cd>Cu>Ni. Both these (and especially the latter) differ from that for the alginates extracted from another brown seaweed (Laminaria digitata) Pb≥Cu≥Cd≥Ni≥Zn [23]. 4.

CONCLUSION In this study, kinetic and batch experiments were investigated for the biosorption of lead, copper, cadmium, zinc and nickel by four local seaweeds Sargassum sp., Padina sp., Ulva sp., and Gracillaria sp.. Metal removal was significantly dependent on pH. The metal ion uptake increased sharply from pH 2 to 4.5; beyond pH 4.5, its increasing effect was diminished. Results also showed that most of the adsorption capacities could be achieved in a very short time; 90% of the total adsorbed metal ions was achieved within 10-60 minutes. The general affinity sequence for Padina sp. is Pb>Cu>Cd>Zn>Ni; and for Sargassum sp. Pb>Zn>Cd>Cu>Ni. Overall, the brown seaweeds (Sargassum sp. and Padina sp.) showed better overall biosorption performance when compared with the green and red seaweeds (Ulva sp. and Gracillaria sp. respectively). In the biosorption studies with single metal ion, the maximum uptake capacities for lead, copper, cadmium, nickel and zinc were 1.16, 0.99, 0.76, 0.61 and 0.50 mmol/g respectively for Sargassum sp. and 1.25, 1.14, 0.75, 0.63 and 0.81 mmol/g respectively for Padina sp., which are higher than or comparable to those of most other types of biomass reported in the literature. The results suggest that these brown seaweeds could be used to develop high capacity biosorbent materials for the removal of heavy metal ions from aqueous solutions. Since this study served to identify local marine biomass that may be used in heavy metal removal, no pretreatment methods were applied. Pretreatment may be expected to enhance the biosorption behavior [9]. ACKNOWLEDGEMENTS This work was funded by the National University of Singapore research grant R-279000-123-112. REFERENCES

1. Crist, R.H., Oberholster, K., Shank, N. and Nguyen, M., Nature of binding between metallic ions and algal cell walls, Environ. Sci. Technol., 15 (1981) 1212. 2. Hassall, K.A., Uptake of Cu and its physiological effects on Chlorella vulgaris, Physiol. Plant, 16 (1963) 323. 3. Aksu, Z., Sag, Y. and Kutsal, T., The biosorption of copper (II) by C. vulgaris and Z. ramigera. Environ. Technol. 13 (1992): 579-586. 4. Matheickal, J.T. and Yu, Q., Biosorption of lead(II) from aqueous solutions by Phellinus badius. Miner. Eng. 10 (1997) 947-957. 5. Matheickal, J.T. and Yu, Q., Biosorption of lead from aqueous solutions by marine algae Ecklonia radiata. Water Sci. Technol. 34 (1996) 1-7.

823

Biosorption

6. Fourest E. and Volesky B. Alginate properties and heavy metal biosorption by marine algae. Appl. Biochem. Biotechnol. 67 (1997) 33-44. 7. Holan, Z.R.; Volesky, B.; Prasetyo, I., Biosorption of cadmium by biomass of marine algae, Biotechnol. and Bioeng., 41 (1993) 819-825 8. Schiewer, S. and Volesky B., (1999) Biosorption by marine algae. In Remediation, ed. J. J. Valdes, Kluwer, Dordrecht, The Netherlands. 9. Volesky, B., Biosorption of Heavy Metals, CRC Press. Inc., 1990. 10. Matheickal J.T., Feltham J. and Yu Q., Cu(II) binding by marine algae Ecklonia radita biomaterial. Environ. Technol. 18 (1997) 25-34. 11. Davis, T.A., Volesky, B., and Vieira, R.H.S.F., Sargassum seaweed as biosorbent for heavy metals, Water Research, 34 (2000) 4270-4278. 12. Volesky, B., Advances in biosorption of metals: selection of biomass types, FEMS Microbiology Reviews, 14 (1994) 291-302. 13. Yin, P., Yu, Q., Jin, B. and Ling, Z., Biosorption removal of cadmium from aqueous solution by using pretreated fungal biomass cultured from starch wastewater, Water Research, 33 (1999) 1960-1963. 14. Matheickal, J.T. and Yu Q. Biosorption of lead(II) and copper(II) from aqueous solutions by pre-treated biomass of Australian marine algae, Bioresource Technology, 69 (1999) 223-229. 15. Yu, Q., Matheickal, J.T., Yin, P., and Kaewsarn, P., Heavy metal uptake capacities of common marine macro algal biomass, Water Research, 33 (1999) 1534-1537. 16. Schecher W. D. (1991) MINEQL+: A Chemical Equilibrium Program for Personal Computers, Users Manual Version 2.22. Environmental Research Software, Inc., Hallowell, ME. 17. Kratochvil, D., and Volesky, B., Advances in the biosorption of heavy metals, TIBTECH, 16 (1998) 291-299. 18. Yang, J.B., and Volesky, B., Biosorption of uranium on Sargassum biomass, Water Research, 33 (1999) 3357-3363. 19. Hope, A.B. and Walker, N.A., The physiology of giant algal cells, Cambridge University Press, Cambridge, England, 1975. 20. Volesky, B., Detoxification of metal-bearing effluents: biosorption for the next century, Hydrometallurgy, 59 (2001) 203-216. 21. Kratochvil, D., Pimentel, P., and Volesky, B., Removal of trivalent and hexavalent chromium by seaweed biosorbent, Environ. Sci. Technol. 32 (1998) 2693–2698. 22. Sánchez, A., Ballester, B., Blázquez, M.L., González, F., Muñoz, J. and Hammaini, A. Biosorption of copper and zinc by Cymodocea nodosa, FEMS Microbiology Reviews, 23 (1999) 527-536. 23. Stokes, P.M., Maler T., and Riordan, J.R., A low molecular weight Cu-binding protein in a Cu-tolerant strain of Scenedesmus acutiformis, in Proc. 11th Annual Conf. on Trace Substances in environmental Health, University of Missouri, Columbia, 1977.

824

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption of heavy metals onto an olive pomace: adsorbent characterisation and equilibrium modelling Francesca Pagnanellia*, Stefano Ubaldinib, Francesco Vegliòc, Luigi Toroa a

Dipartimento di Chimica, Facoltà di S.M.F.N., Università degli Studi "La Sapienza", P.le A. Moro, 5, 00185 Roma, Italy (e-mail: [email protected] and [email protected]) b Instituto Geologia Ambientale e Geoingegneria (CNR), Via Bolognola 7, 00185 Roma (e-mail: [email protected]) c Dipartimento di Chimica, Ingegneria Chimica e Materiali, Universita’ degli Studi di L’Aquila, 67040 Monteluco di Roio, L’Aquila, Italy (e-mail: [email protected]) Abstract In this study an agricultural waste, an olive pomace, was used to remove heavy metals from aqueous solutions at different equilibrium pH. The adsorbent material was preliminary characterised from a physico-chemical point of view by size distribution analysis, SEM analysis, CEC, potentiometric titration and IR analyses. The acid-base properties of the functional groups on the adsorbent were then modelled by a model accounting for the functional heterogeneity. The modelling of the potentiometric data coupled with other experimental findings evidenced the presence of two main kinds of active sites: carboxylic and phenolic groups. Equilibrium tests carried out using different heavy metals at different equilibrium pH evidenced the positive effect of increasing pH on the specific adsorption capacity of the adsorbent. Moreover a general affinity series was observed in agreement with that reported in literature for other adsorbent materials (both inorganic and biological): Pb>Cu>Cd. This affinity series can be explained considering the HSAB theory of Pearson for complexes formation and it suggests a complexation reaction among metals and active sites. 1.

INTRODUCTION Heavy metal contamination is caused by different kinds of industrial, mining and military activities. Heavy metal ions are not biodegradable pollutants and tend to accumulate in the living organism, causing various diseases and disorders. Traditional treatment processes of contaminated wastewaters (ion exchange, precipitation and evaporation) present different technical and economical constraints about the fulfilling of law regulations. Biosorption is an alternative technology to remove heavy metal from dilute aqueous solutions based on the property of certain kinds of inactive and dead biomasses to bind

* Corresponding author: phone +39 06 49913333; fax +39 06 490631

825

Biosorption

and accumulate these pollutants. Biosorbents generally used for these purposes are wastes coming from agricultural and industrial activities or specially propagated biomasses of fungi, yeast and bacteria. The main advantage in biosorption is cost-effectiveness being based on the use of "low cost" materials. Olive pomace is one of these possible adsorbents that is abundant especially in countries of the Mediterranean area. In particular Italy is one of the major producing countries of olive oil; the manufacturing of olive oil yields an oily phase (30%), a solid residue (20%), and an aqueous phase (50%). The solid residue (pomace) is actually swallowed by means of controlled spreading on agricultural soil; a little amount is also used as natural fertilizer, combustible and added to animal food. An alternative use can be as low cost natural adsorbent (about 50 $/ton against 4500 $/ton of a granular actived carbon) [1]. At now the main enterprises concerning a wider application of biosorption technology in large-scale plants are related to product performance optimisation and continuous process development. Both these aspects are strictly connected to the understanding of the physico-chemical mechanisms involved in the interactions among active sites on solid adsorbent and ionic species in solution. Mathematical modelling is then a fundamental step in this paper both as understanding means and also as the base for dynamic process design. In particular considering pH as one of the most influencing operative condition in biosorption of heavy metals, the effect of this factor is studied both as acid-base behaviour of the adsorbent and as a factor affecting the equilibrium uptake of different heavy metals. The aim of this work is then twofold: modellistic and interpretative: characterizing the adsorbent and representing its acidbase behaviour technological and applicative: exploring the capacity of olive pomace in heavy metal removal from aqueous solutions in different operative conditions 2.

MATERIALS AND METHODS

2.1 Olive pomace Pomace was collected as pressed and sunny dried disks from an olive oil production plant in Italy. Pomace was grounded and particle size distribution was determined by an automatic sieve. Pomace samples were washed by distilled water: 20 g/L biomass suspensions were kept at 250 rpm in a shaker for 90 minutes at 25°C; solid-liquid separation was performed by centrifugation. After three successive washings solid samples were dried and used for titration and heavy metal biosorption [1]. 2.2 Potentiometric titration Pomace suspensions were preliminary acidified (6 mg in 20 mL of H2O acidificated with HCl 0.1 M, standard solution) and then potentiometrically titrated by successive additions of NaOH 0.1 M, standard solution. 2.3 Equilibrium tests Equilibrium biosorption of Pb(II), Cd(II), Cu(II) was determined by using 10 g/L pomace suspensions in which different initial metal concentrations were added. Samples were kept at constant pH and temperature under magnetic stirring until equilibrium conditions were reached (90 minutes). Sold-liquid separation was performed by centrifugation and the metal concentration was determined by an Inductively Coupled Plasma Spectrophotometer (ICP). For each sample, a blank test without biomass was 826

Biosorption

performed to determine the initial metal concentration by ICP and to avoid confusion between biosorption and possible metal precipitation. 3.

RESULTS AND DISCUSSION

3.1 Adsorbent characterisation Olive pomace is a very heterogeneous matrix both from a morphological and a functional point of view. The morphological heterogeneity was evidenced experimentally by SEM analyses (here not reported) which show the presence of particles with different shape and a wide range of dimensions. This is also evidenced by the distribution of particle size of a grounded sample (Fig. 1).

Figure 1. Size distribution of pomace particles expressed as % weight

A part from the different classes evidenced in the histogram two main types of particles come out from grinding: a type A with lower dimensions (<500µm) and a type B with larger dimensions (>500µm). This classification was chosen according to the different aspect of particles over and below 500µm. In particular type A is characterised by a dark brown colour and is preferentially made up of the cellulose, residual fats and polyphenolic substances of olive pulp, while type B fraction is dark yellow and composed of lignin fragments of olive seeds [2]. The functional heterogeneity associated to this morphological difference was put in evidence by the potentiometric titration data of olive pomace mix after grinding compared with those of a blank without biomass and of the two fractions (Fig. 2). The analysis of the experimental results shows that: acidic active site concentration is larger in the A fraction with lower dimensions then in the B one, so that morphological and functional heterogeneity results strictly related according to the different fractions composition already mentioned; in any case trends are rather flat testifying the heterogeneity of the matrix and the impossibility of distinguishing separated groups with well separated acid pK.

827

Biosorption

12 10

pH

8 6 Blank Pomace mix A fraction

4

B fraction

2 0 0

5

10

15

V NaOH (mL)

Figure 2. Potentiometric titration of the mixture obtained from the grounded pomace compared with two different fractions (A<500µm and B>500µm)

According to this last observation titration modelling of this material cannot be performed assuming the presence of few distinct sites. A continuous approach was necessary requiring the introduction of an affinity distribution for the logarithm of the protonation equilibrium constant of the active sites determining the acid-base behaviour of the system. This approach was developed for representing the interaction among ionic species in aqueous solutions with humic substances [3]. The continuous approach here considered relates the total fraction of protonated sites θT,H on the adsorbent given by the following general integral: θ T ,H =

[LH] = [L] + [LH] ∫

∆ log K

θ L,H (K H , H )f (log K H )d log K H

(1)

L

where L is the free active site with an affinity constant KH whose logarithm is distributed according to a certain affinity distribution f(logKH) over a specified range ∆logKH. This integral can be solved to obtain an analytical expression for θT,H choosing the proper local isotherm θL,H and f(logKH) distribution. In the literature there are different expressions obtained using different kinds of distributions for f(logKH) [4] and a Langmuir type model (eq. 2) as local isotherm θ L,H =

K H [H ] 1 + K H [H ]

(2)

where KH is the affinity equilibrium constant among active site L and protons KH =

[LH] [L][H]

(3)

Experimental data were firstly reported as negative charge for gram of adsorbent (QH mmol/g) as a function of pH by applying the charge balance in the system QH 828

V ([H ] + [Na ] − [OH ] − [Cl ])V = ∑ [L ] = m m i

− i

+

+

−

−

(4)

Biosorption

where V(L) is the suspension volume during titration, m (g) is the amount of biomass, [Na+] is the sodium concentration added in solution as sodium hydroxide (titrant) and [Cl-] is due to the initial addition of HCl to the biomass suspension to lower the initial pH (see section 2.2). The heterogeneity analysis of experimental data was then used to have preliminary information about the number of active site types and the shape of the logKH distribution [5]. Two main kinds of acidic sites were assumed to represent experimental data of QH versus pH considering a Sips distribution for each one f (log K H ) =

ln (10 )sin (mπ )

⎡⎛ K π ⎢⎜⎜ ~ H ⎢⎝ K H ⎣

⎞ ⎟ ⎟ ⎠

−m

⎛K + 2 cos(mπ ) + ⎜⎜ ~ H ⎝ KH

⎞ ⎟ ⎟ ⎠

m⎤

(5)

⎥ ⎥ ⎦

~ where K H is the median value of the distribution and m (0<m<1) is a parameter related to the shape of the distribution and in particular as m decreases the distribution becomes larger (for m=1 it is a Dirac distribution meaning a homogeneous type of site). By this way •

considering the relation among fraction of protonated sites θT,H, negative charge on the adsorbent QH and maximum negative charge QMax QH = Qmax (1 − θ H ,T )

•

(6)

extending the general Langmuir-Freundlich expression to two kinds of active sites, the following expression can be obtained relating the negative charge on the adsorbent QH with the hydrogen ion concentration in solution QH =

Qmax,1 Qmax,2 + m1 m2 ~ ~ 1 + K H ,1 [H ] 1 + K H , 2 [H ]

(

)

(

)

(7)

A non-linear regression of the experimental data was then performed to determine the adjustable parameters characteristic of each site, which define concentration Qmax,i ~ (mmol/g) and shape of the related affinity distribution ( K H and mi). In order to find significative values of the parameter, the range was restricted during non linear regression: ~ K H varied from 2 to 12, that is the range of pH studied during the titration, m varied from 0 to 1 in accordance with possible value of the Sips distribution and an estimate of Qmax was obtained by the graphic method of Gran [6]. Experimental data and model curve were reported in Figure 3 along with the affinity distribution for the two sites obtained putting the adjustable parameters (Table 1) in the Sips distribution to give an idea of the dispersion of the acid-base properties due to adsorbent heterogeneity. The goodness of the model evidenced by the agreement among experimental and predicted data is also enforced considering the obtained values of the adjustable parameters. In particular the total concentration of active sites is in agreement with the preliminary estimate made by Gran’s method (0.67 mmol/g), and the equilibrium acidic constants are realistic considering the component of olive pomace. On the base of the parameters values and from the literature analysis it is then reasonable to assume that these two kinds of functional groups are carboxylic and phenolic. Also the analysis of the IR spectrum on solid phase (here not reported) can be used to obtain some confirms about the nature of these sites. In fact even though this spectrum resembles the high heterogeneity of the pomace, some characteristic peaks can 829

Biosorption

be evidenced. In particular different peaks in the C=O stretching region (two peaks at 1704 and 1643 cm-1) are in agreement with the presence of the carboxylic groups and also confirm the heterogeneity of this functional group that is present both in fat and polyphenol compounds. The characteristic peaks of phenolic groups are those related to OH stretching which here cannot be distinguished by the other kinds of OH stretching typical of the cellulosic matrix (a large band around 3000 cm-1). Moreover a peak at 1507 cm-1 can be assigned to the stretching of C-C bound in aromatic ring present in polyphenols and lignin as confirmed by the comparison of the pomace spectrum with those of natural standard of lignin and cellulose which respectively present and do not present this specific peak (data here not reported).

Figure 3. Experimental data and model curve for potentiometric titration; affinity distribution calculated by Sips equation (eq. 5) using the adjustable parameters reported in Table 1 Table 1. Adjustable parameters for potentiometric titration modelling Site 1 Site 2

Qmax,i (mmol/g) 0.07 0.70

logKHi 4.7 9.0

mi 0.90 0.06

3.2 Equilibrium tests Experimental data for Pb(II), Cd(II), Cu(II) biosorption onto olive pomace at different equilibrium pH (3, 4 and 5) were reported in Figures 5-7. The analysis of these experimental data outlines some interesting results and in particular: the effect of equilibrium pH on adsorption performances; the effect of metal speciation onto adsorbent affinity. As for the effect of equilibrium pH onto adsorption performances it is evident that increasing the hydrogen ion concentration from pH 5 to 3 maximum metal uptake decreases. This is due to the competition among metals and hydrogen ions for active sites 830

Biosorption

on the adsorbent. In fact the acid-base properties evidenced by potentiometric titration data of a pomace suspension show the presence on the adsorbent of two main kinds of weakly acidic active sites probably carboxylic and phenolic ones. As pH increases the active sites dissociate themselves becoming negatively charged and then able to bind positive metallic species in solution.

Figure 5. Copper biosorption onto olive pomace at different equilibrium pH

Figure 6. Cadmium biosorption onto olive pomace at different equilibrium pH

831

Biosorption

Figure 7. Lead biosorption onto olive pomace at different equilibrium pH

As for the adsorbent affinity the comparison among metal specific maximum uptakes (the plateau of the isotherm) at the different equilibrium pH shows that lead removal is larger than copper that is larger than cadmium. This experimental result (affinity order: Pb>Cu>Cd) can be explained considering different interpretations. Considering the logarithmic values of the metal first hydrolysis constant (1) the order is LogKPb=-7.71 > LogKCu=-8.00 > LogKCd=-10.80 [7]. HO − H + Me 2 + → Me(OH ) + + H +

(8)

This correlation between metal acidic property and its uptake seems to be even more important than the specific functional groups present on the adsorbent surface; in other words, metal speciation predominates on adsorbent characteristics. This observation is strongly enforced considering that the same experimental behaviour was observed using different biological and inorganic matrices [8]. This experimental result can be explained considering the analogy between the reaction of metal hydrolysis (eq. 8) and the reaction between metal and active site (eq. 9): S − H + Me 2 + → SMe + + H +

(9)

where Me is the heavy metal and SH is the active site in the protonated form. In both reactions (8-9) the bond of hydrogen is broken (HO-H and S-H), a H+ ion is released and substitute by a metal (HO-Me and S-Me). From this point of view, it is logical that if a heavy metal is very acidic (because of its charge to mass ratio) then it will react more easily with a protonated site with respect to a weaker acidic heavy metal. The observed affinity order can be also explained considering the Hard Soft Acid Base theory of Pearson (HSAB) for complex formation which classifies different species as acid and bases arranged in a specific scale of hardness and says that hard bases react preferentially with hard acid and soft with soft ones [9]. Assuming the presence of hard bases as functional active groups on the adsorbent (such as carboxylic and phenolic ones) the hardness of the metals as acids reacting with the bases follows the order Pb > Cu > Cd. The existence of this particular affinity series can then help in the identification of the operating mechanism. 832

Biosorption

4.

CONCLUSIONS This paper evidenced the possibility of using olive pomace, an agriculatural waste, as adsorbent for heavy metal pollutants in aqueous solutions. In particular the effect of equilibrium pH was studied both regarding the acid-base behaviour of the active functional groups on the adsorbent and regarding the depression of the metal specific uptake for decreasing pH values.

REFERENCES

1. F. Pagnanelli, L. Toro and F. Vegliò, Waste Management 22 (2002a) 901. 2. L. Vlyssides J. Environm. Sci. Health 34 (1999) 737. 3. L.K. Koopal, W.H van Riemsdijk and D.G. Kinniburgh Pure Appl. Chem.73 (2001) 2005. 4. J.C.M. de Wit, W.H. van Riemsdijk and L.K. Koopal Environ. Sci. Technol. 27 (1993) 2015. 5. M.M. Nederlof, W.H. van Riemsdijk and L.K. Koopal, J.Colloid Interface Science 135 (1990) 410. 6. G. Gran, Acta Chem. Scand. 4 (1950) 559. 7. R. M. Smith and A.E. Martell 1976, Critical Stability constants. Plenum Press New York. 8. F. Pagnanelli, A. Esposito, L. Toro, and F. Vegliò Water Research, 37 (2003) 627. 9. R. G. Pearson Surv. Progr. Chem. 5 (1969) 1.

833

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption of Hg by vegetal biomasses P.F. Pimentela*, R.P. de Carvalhob, M.H. Santosa, M.C. Andradea a

CETEC, Biotechnology and Chemical Technology Department Av. José Cândido da Silveira 2.000, Horto CEP 31170-000, BH, MG, Brazil b Departamento de Física, ICEx – UFMG - CP 702 – CEP 30123/970, Belo Horizonte, Brazil Abstract The Hg sorption capacities as well as the evolution of these capacities during sorption/desorption cumulative cycles were tested for dried algae (Sargassum sp), rich in alginates, dried lettuce leaves (Lactuca sativa), rich in cellulose fibers, and corn husk (Zea mays), rich in lignin. The Sargassum biomass showed the more efficient uptake, and after three cycles of acidic desorption treatment the seaweed biomass still presented a very good sorptive capacity (20mg Hg/g(biomass)). The lettuce biomass presented sorptive capacity of 22 mg Hg/g(biomass) for saturation concentrations in the first sorption cycle. After 3 sorption/desorption cycles the biomass was degraded and showed no more the same sorption efficiency. The corn husk biomass presented a sorption capacity of 8,6mg Hg/g(biomass) in the first sorption cycle, a small value if compared to the other biomasses. FTIR spectra indicated that after sorption the metal is located in similar sites in all the biomasses studied, that is, near the ramifications of glucose rings of the alginate, cellulose or lignin fibers of seaweed, lettuce leaves or corn husk, respectively.

Keywords: biosorption, mercury, FTIR, Sargassum, L. sativa, Z. mays, sorption, desorption 1.

INTRODUCTION Mercury has been used in gold mining in Brazil since gold was first discovered in the Amazon basin in the 18th century. In the last ten years, between 1,000-2,000 tons of this highly poisonous liquid metal have been released into the environment. Concern over the contamination of the Amazon River and of local populations is escalating with hundreds of thousands of people living in the region thought to be at risk. Mercury pollution is one of the most serious environmental problems related to this activity [1]. Solutions for the problem have received little attention from researchers and governments. Clean-up technologies, which are capable of treating large volumes of soil, water or sediments contaminated with relatively low levels of mercury in a cost-effective way, are urgently needed and an integrated approach to mercury problem is necessary [2]. Bioremediation of toxic metals by biosorption as an alternative technology for the metal removal of industrial and mining waste has received much attention recently [3].

* E-mail: [email protected]

835

Biosorption

The first major challenge for the biosorption field is to select the most promising types of biomass from an extremely large pool of readily available and inexpensive biomaterials. The aim of this study is to evaluate the Hg sorption efficiency of inexpensive biomasses as well as the evolution of these efficiencies during sorption/desorption cumulative cycles. The Hg sorption capacity was tested for biomasses composed of dried sargassum algae (Sargassum sp), dried lettuce leaves (Lactuca sativa) and corn husk (Zea mays). At the same time, the modifications of the structure of biomasses caused by the presence of mercury ions were studied using Fourier Transform Infrared Spectroscopy (FTIR). 2.

MATERIALS AND METHODS

2.1 Biomass preparation For biosorption experiments the biomasses of Sargassum sp. and corn husk (Z. mays) were washed with distilled water and dried in the oven overnight at 60°C. Then, they were shopped in pieces of size around 0.3-0.6 cm, washed with 6 N HCl (Merck), rinsed three times with distilled water and oven dried for 24 hours at 60°.

Lettuce leaves (L. sativa) were sun dried during 4 weeks, then oven dried at 35°C for 7 days and at 70°C for 24h. After grounding in a domestic blender to a size of about 1 mm, the biomass was dried at 70°C during 24h and washed in an aqueous solution with pH=4 using a drop of diluted HCl solution (Aldrich). 2.2 Batch sorption experiments

Batch kinetics experiments were conducted at room temperature (25°C), in a rotary shaker, using 250 mL Erlenmeyer flasks. The Hg(II) solution was prepared in distilled water using HgCl2 (ACS-QM). Approximately 100 mg of dried biomass was combined with 50 mL of the metal solutions and the flasks were placed on the shaker for 6 h. The pH of the solutions before and during the sorption experiments was adjusted to 5.0 with NaOH and/or HNO3 solutions (Merck) in a pHmeter DMPH (Digimed). 2.3 Analysis of mercury The total concentration of mercury in the liquid samples was determined by Atomic Absorption Spectrometry (AAS) in the Perkin-Elmer Flow Injection Hydride Generation Atomic Absorption Spectrometers, models AA403 coupled to MHS-20 and AAS FIMS100. 2.4 FTIR spectra For Fourier Transform Infrared Spectroscopy (FTIR) studies, natural and Hg-charged samples of the biomasses were prepared as pellets using KBr (Graseby Specac LTD.) as a substratum. Spectra were recorded with a BOMEM-DA8 FTIR spectrometer and deconvoluted using Peakfit TM 4.00 software. Details of the sample preparation and of the spectra recording and deconvolution are given in [4], as well as an overview of the FTIR theory. 2.5 Data evaluation Uptakes of mercury were determined from the difference of metal concentrations in the initial and final solutions and the biosorption coefficient (q) was calculated as [5]: 836

Biosorption

q = (C i − C f )V / M

(1)

Where q is the amount of metal ion adsorbed onto the biosorbent (mg/g), Ci and Cf are the concentrations of the metal ions in the initial solution (mg/L) and after biosorption, respectively. V is the volume of the aqueous phase (L) and M is the amount of biosorbent (g). For each biomass the uptake results were fitted using the Langmuir sorption model [6]: q = q o bC f /(1 + bC f )

(2)

where qo and b are the characteristic parameters of the Langmuir isotherm: qo represents the saturation uptake for high equilibrium concentrations and b is related to the affinity of the metal ion with the biomass structure, which defines the inclination of the isotherm for low equilibrium concentrations. 2.6 Desorption experiments Desorption of Hg(II) was performed with 50 mL of HNO3 solution, 0.2M (Merck). The biomasses loaded with Hg(II) ions were placed in this desorption solution and stirred at 200 rpm for 3 hours at room temperature. Then the mercury concentration in the supernatants was analyzed using AAS. The sorbent samples regenerated were washed with distilled water and then reused for a new adsorption process. The loading and regeneration cycle was repeated three times. 3.

RESULTS AND DISCUSSION

3.1 Comparative sorption of different biomasses Figures 1 to 3 show the sorption isotherms for Sargassum sp., L. sativa and Z. mays, respectively, for 3 sorption/desorption cycles. The vertical axis scale (sorption uptake q) was made the same for the three curves, in order to facilitate the comparison of uptake values. During the three sorption cycles studied, Sargassum biomass showed the more efficient uptake. In the concentration range covered by the experiments it was not possible to attain the biomass saturation condition. For this reason the curves were not fitted using Langmuir equation (2), and the results are presented as smooth curves joining the average of the experimental points. Although the sorptive capacity of the biomass degrades after each sorption/desorption cycle, after three cycles the biomass still presents a very good sorption efficiency, with a q value of about 20 mg (Hg)/g (biomass) for the higher values of metal concentration studied. L. sativa presented also a good sorptive capacity, with qo = 22±2 mg (Hg)/g (biomass) in the first sorption cycle. For the second cycle few data were obtained and it was not possible to make a Langmuir fit. Due to a very low mechanical resistance, after 3 cycles the sorptive capacity is degraded and the maximum uptake is about half the original one. Z. mays presented qo = 8.6±0.8 mg(Hg)/g (biomass) in the first cycle, and the maximum uptake decreased to about half this value after 3 cycles. Data present a larger dispersion than for the other biomasses, due to the lower uptake values. Even if the sorptive capacity of corn husk is less effective than the one found for the other biomasses, it is worthy to note that this biomass has a bigger density than the others studied, and that this can compensate the low value of qo if the sorption is planned to be done in a column of fixed size. 837

Biosorption

1 s t c y c le 2 d c y c le 3 d c y c le

50

40

q (mg/g)

30

20

10

0

0

5

10

15

20

25

30

C f (m g /L )

Figure 1. Biosorption isotherms of dried sargassum algae (Sargassum sp) biomass in 3 cycles of Hg sorption/desorption 1 s t c y c le 2 d c y c le 3 d c y c le

50

40

q (mg/g)

30

20

10

0

0

20

40

60

80

100

C f (m g /L )

Figure 2. Biosorption isotherms of dried lettuce leaves (Lactuca sativa) biomass in 3 cycles of Hg sorption/desorption 1 s t c y c le 2 d c y c le 3 d c y c le

50

40

q (mg/g)

30

20

10

0 0

20

40

60

80

100

C f( m g / L )

Figure 3. Biosorption isotherms of corn husk (Zea mays) biomass in 3 cycles of Hg sorption/desorption 3.2 FTIR spectra Figure 4 shows the FTIR absorption spectra for natural (A) and Hg-charged (B) Sargassum, natural (C) and Hg-charged (D) L. sativa, and natural (E) and Hg-charged (F) Zea mays, in the region between 800 cm-1 and 1800 cm-1. The spectra were normalized to 838

Biosorption

-

COO CO

1000

Absorption (a.u.)

Absorption (a.u.)

A

1500

CO

1000

CO COC

1200

1250

Absorption (a.u.)

Absorption (a.u.)

Wavenumber (cm )

E

OH 1500 -1

Wavenumber (cm )

1750

CH

1600

Wavenumber (cm-1)

-1

1000

2000

D

800

1600

C-O-C

1500

Wavenumber (cm )

CH

1200

COO

-1

C

800

-

2000

Wavenumber (cm -1)

CO COC

B

Absorption (a.u.)

Absorption (a.u.)

take account for thickness differences in the prepared samples. It can be seen that the presence of mercury modifies the infrared absorption of all the samples in similar ways.

F

C-O-C 1000

1250

1750

1500 –1

Wavenumber (cm )

Figure 4. FTIR absorption spectra and deconvolution for natural (A) and Hgcharged (B) Sargassum, natural (C) and Hg-charged (D) L. sativa, and natural (E) and Hg-charged (F) Zea mays

a.

b.

on Sargassum samples, we can note a modification after Hg sorption on the peak with wavenumber 1520 cm-1, assigned to COO- bonds vibrations [7]: in spectrum A (natural) the height of this peak is equal to the height of its right and left neighbors, whereas in spectrum B (Hg-charged) the height of this peak is smaller than its neighbors height. We can also note modifications in the peak with wavenumber 1160 cm-1, assigned to CO bonds vibrations [7]: in spectrum A its height is smaller than its right neighbor height, whereas in spectrum B the heights are approximately the same. COO- is present in the ramification of alginate, main component of dried seaweed, and CO bonds are present in the alginate rings, near the ramification. on lettuce samples (L. sativa), peaks modified by the sorption of mercury have wavenumbers 1400 cm-1, assigned to CH bonds vibrations; 1160 cm-1, assigned to C839

Biosorption

O-C bonds vibrations, and 1115 cm-1, assigned to CO bonds vibrations [8]: in spectrum C (natural) the CH peak height is approximately 2/3 of the height of its left neighbor, while in spectrum D (Hg-charged) heights are almost equal; meanwhile, CO-C and CO peaks heights change from half the height of the left neighbor in spectrum C to about 1/3 the height of left neighbor in spectrum D. CH is present in the ramification of cellulose, the main component of dried lettuce leaves, and CO and C-O-C bonds are present in glucose rings of cellulose, near the ramification. c. on Z. mays samples, the sorption of mercury causes smaller modifications on infrared absorption peaks, due to the smaller quantity of metal retained by the biomass. Even though, we can note a small modification on the peak with wavenumber 1170cm-1, assigned to C-O-C bonds vibrations [8]: in spectrum F (Hg-charged) this peak is smaller as compared with its neighbors than it was in spectrum E (natural). C-O-C bonds are present in glucose rings of lignin, main component of corn husk. We note also the disappearance of a peak with wavenumber 1610 cm-1, which can be assigned to OH bonds of water present in some samples [9]. No absorption peaks that could be assigned to bonds of the biomass structure components were created or extinguished by the presence of the mercury in the biomass, indicating that no molecular bonds of the structure were formed or destroyed after the sorption of the metal ion. It is worthy to note that in our study, the absorption peaks have wavenumbers localized between 800 cm-1 and 2000 cm-1. These peaks can be associated to the bond vibrations of the organic molecules of the biomass. Absorption peaks associated to bond vibrations of mercury bonds would have a very low wavenumber, and would not be seen in our spectra. 4.

CONCLUSIONS The metals biosorption depends strongly on the nature of the biosorbent. It appears from the results of this study that the differences in the cell wall constituents in the sargassum biomass tested make the marked difference in the Hg(II) removal from solutions. Among the biomasses studied, Sargassum sp. far displays the best performance, followed by lettuce leaves (Lactuca sativa) and corn husk (Zea mays). The presence of the sorbed metal affects the bonds of ramifications of the structures, as well as the bonds of the carbon rings, near the ramifications. Similar features about the absorption peaks were seen in another work that studied copper sorption in L. sativa [10] where it was concluded that, after sorption, a hydrated copper ion is located near two glucose rings of the cellulose structure, in a site near the glucose ramification and with axial symmetry neighborhood. We thus conclude that the sorption mechanism of mercury is similar to the one for copper sorption. Also, our results agree with observations in [11] for Hg sorption on steam activated and sulphurised activated carbons prepared from bagasse pith. The authors proposed that in acidic medium Hg(OH)+ species present in the solution may be bond to COOH groups of the carbon materials. Finally, we conclude that, between the tested biomasses, Sargassum is the more efficient for mercury uptake and can be utilized mainly in coastal regions where it can be easily found. L. sativa also presented a good sorptive capacity but a very low mechanical resistance and this brings difficulties in its use on sorption columns. Corn husk (Z. mays) biomass is a less effective sorbent than the other biomasses. However, this biomass is very abundant in developing countries with mercury pollution problems, and its density is 840

Biosorption

higher than the density of the other biomasses studied. These factors can compensate the low value of Hg uptake, if the sorption is planned to be done in a column of fixed size. Biosorption studies have been conducted by several researchers using live or inactivated microbial systems and surface-modified adsorbents for removal of mercury from wastewater and synthetic solutions [12-15]. Such preparations offer advantages in terms of high biosorption capacities, mechanical strength and durability, handling and ease of scale up. Nevertheless, all these adsorbents are expensive and require several preparation steps. So, for scale up applications of mercury removing in remote areas, it is important to attain efforts on obtaining inexpensive and abundant biomasses with mechanical strength and durability, and good mercury sorption capacity. The biomasses proposed in this study answer to these requirements and present a novel option to large scale Hg decontamination. The structural information about sorption sites based on infrared absorption analyses, presented in this study, can also help in the understanding of the Hg sorption phenomenon and in the search of other convenient biomasses. ACKNOWLEDGEMENTS The author would like to thank CNPq and FNMA for the financial support and FAPEMIG for Ms. M. C. Andrade research fellowship. We are also grateful to Dr. Roberto L. Moreira (UFMG) for the discussions about FTIR, Mr. M.M de Magalhães (UFMG) and Ms. C.M.Pittella (CETEC) for laboratory support, and Ms. O.G. F. Rocha (CETEC) for AAS analyses. REFERENCES

1. M. Veiga, M. Introducing new technologies for abatement of global mercury pollution in Latin America. Rio de Janeiro: UNIDO/UBC/CETEM/CNPq, 94p (1997). 2. H. Von Canstein, Y. Li, K.N. Timmis, W-D. Deckwer, I. Wagner-Dobler. Removal of mercury from chloroalcaly electrolisys wastewater by a mercury resistant Pseudomonas putida strain. Applied and Environmental Microbiology 65 (1999) 5279-5284. 3. D. Kratochvil and B. Volesky, Advances in the biosorption of heavy metals, Trends Biotechnol., 16 (1998) 291-300. 4. R.P.de Carvalho, J.R.Freitas, A.-M. G. de Sousa, R. L. Moreira, M. V. B. Pinheiro and K. Krambrock, Biosorption of copper ions by dried leaves: Chemical bonds and site symmetry, accepted for the IBS-2001 special issue of Hydrometallurgy (to be published) (2003). 5. R.P.de Carvalho, K.H.Chong, B.Volesky. Effects of Leached Alginate on Metal Biosorption: Biotech. Lett. 16 (8) (1994) 875-880. a. Langmuir, Chemical Reactions at Low Pressures, J.Am.Chem.Soc. May (1915), 11391167. 6. C.Sartori, D.S.Finch, B.Ralph, K.Gilding, Determination of the cation content of alginate thin films by FT i.r. spectroscopy. Polymer 38 (1997) 43-51. 7. A.J. Michel. Second Derivative FTIR Spectra of Woods. In: Schuerch, C. (Ed.), Proc. 10th Cellulose Conference, John Wiley and Sons, N.Y., (1989), 995. 8. R.H. Wilson, A.C Smith, M. Kakuráková, P.K. Saunders, N. Wellner, K.W. Waldron. The mechanical propertie and molecular dynamics of plant cell wall polysaccharides studied by Fourier- transform infrared spectroscopy. Plant Physiol. 124 (2000) 397406. 841

Biosorption

9. R. P. de Carvalho, A-M. G. de Sousa, J. R. Freitas, C. P. L. Rubinger and K. Krambrock, A Model for the Copper Biosorption in Dried Leaves -accepted for presentation on IBS-2003 (2003). 10. K. Anoop Krishnan, T.S. Anirudhan. Removal of mercury (II) from aqueous solutions and chlor-alkali industry effluent by steam activated and sulphurised activated carbons prepared from bagasse pith: kinetics and equilibrium studies. Journal of Hazardous Materials B92 (2002) 161-183. 11. W. Bae, R.K. Mehra, A. Mulchandani and W. Chen. Genetic Engineering of Escherichia coli for Enhance Uptake and bioaccumulation of mercury. Appl. Environ. Microbiol 67 (2000) 5335-5338. 12. A. Özer, H. Ekiz, D. Özer, T. Kutsal, A. Caglar. A staged purification process to remove heavy metal ions from wastewater using Rhizopus arrhizus. Process Biochem. 32 (1997) 217-226. 13. A. Denizli, B. Salih, M.Y. Anka, K. Kesenci, V. Hasurci, E. Puskin. Cibacron F3GAincorporated macroporous poly(hydroxyethylmetacrylate) affinity membranes for heavy metal removal. J. Chromatogr. A 758 (1997) 217-226. 14. Y. Kaçar, Arpa, Ç, S. Tan, A. Denizli, Ö. Genç, M.Y. Arica. Biosorption of Hg(II) and Cd(II) from aqueous solutions: comparison of biosorptive capacity of alginate and immobilized live and heat inactivated Phanerochaete chrysosporium. Process Biochem. 37 (2002) 601-610.

842

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Biosorption of lead in aquatic environment by Mucor rouxii biomass Shrabani.Som Majumdar1, T. Saha1, T.S. Bandhapadhyay1, S. Chatterjee2 and A.K. Guha2* 1

Institute of Wetland Management and Ecological Design.B-04, Block LA, Sector-III, Saltlake, Kolkata 700 091, India 2 Department of Biological Chemistry, Indian Association for the Cultivation of science, Jadavpur, Kolkata-700032, India Abstract Kinetics and nature of biosorption process for the removal of lead ion from aqueous solution by fungal biomass Mucor rouxii were studied. Temperature, pH, residence time, metal ion and biomass concentration had been found to influence the biosorption process. At equilibrium sorption process followed Lagmuir isotherm model. Biosorption of lead ion by Mucor rouxii biomass is a fast process requiring less than 20 min to achieve more than 90% of adsorption. 1.

INTRODUCTION Physico-chemical processes usually remove heavy metals present in industrial wastewater before discharging into water system. Physico-chemical processes in use for heavy metal removal from wastewater include precipitation, coagulation, reduction, ion exchange, and membrane technology. All these processes are either costly or less effective. A search for low cost and easily available adsorbant has led to the investigation of materials of agricultural and biological origin including microorganism. Microorganism such as bacteria, fungi, yeast and algae [1] can remove heavy metal from aqueous solution in substantial quantities. Microbial biomass has been studied by several researchers [2-7]. Metal uptake by non-living biomass involves different types of adsorption processes. Biosorption is affected by various physical and chemical factors such as pH, temp, contact time, metal and biomass concentration etc [8]. The biosorption processes consist of two steps. (1) initial rapid process, followed by (2) slower second step [9]. In slower process the metal uptake can be due to number of mechanism including covalent bonding, surface precipitation, redox reaction [10,11]. During adsorption a rapid equilibrium is established between the adsorbed metal ion on the cell (Q) and unadsorbed metal ion in solution (Ce) and can be represented by either Freundlich or Langmuir adsorption isotherms. [12-14]. The present work aims to evaluate Pb biosorption by the fungus Mucor rouxii MTCC-386, study the adsorption isotherm and evaluate the kinetics of the reaction [15].

* Corresponding author: E mail: [email protected]

843

Biosorption

2.

MATERIALS AND METHODS

2.1 Materials 2.1.1 Microorganism Mucor rouxii (MTCC 386) was obtained from Institute of Microbial Technology Chandigarh, India and maintained on potato dextrose agar slants and subcultured at regular intervals of 30 days. 2.1.2 Chemicals Chemicals were obtained from E. Merck, Germany and agar powder was purchased from Hi- Media, India. 2.2 Methods 2.2.1 Production of Mucor rouxii biomass Mucor rouxii was grown in potato dextrose broth in 250-ml Erlenmeyer flasks containing 50-ml medium and inoculated with mycelia of M. rouxii grown previously. The flasks were incubated at 30°C in a rotary shaker with continuous shaking (120 rpm) for 80h. Mycelia were collected by filtration, washed with deionised water, dried at 80˚C, pulverized and kept in desiccators for adsoption studies. 2.2.2 Metal solution Stock solution of lead 1000 mg.L-1 was prepared by dissolving its nitrate salt into deionized water [16]. 2.2.3 Metal sorption studies A batch equilibrium method was used to determine the sorption of Pb ion by Mucor rouxii dried biomass. A set of 250-ml Erlenmeyer flasks each containing 50 ml of metal solution at varying concentrations (10-100 mg.L-1) were taken. Dried biomass (0.5 gm) was added in each of the flasks, which were incubated for 4 h at 30˚C on a rotary shaker at 120 rpm. pH was kept at 5.5 in each experiment. Biomass was separated by centrifugation and Pb2+ in the supernatant was analyzed by atomic absorption spectrometer (Varian Spectra AA55). Metal adsorbed by biomass was obtained from the following formula; (I) Q = V(Ci-Cf)/1000 M where Q is the specific metal uptake (mg metal/ g biosorbent), V is the volume of metal solution (mL), and Ci and Cf are the initial and final metal concentration in the solution (mg metal.L-1) and M is the dry weight of the biomass (gm). The metal sorption ability of the biomass was determined by the above-mentioned procedure, in all the experiments unless stated otherwise [17]. Data obtained with respect to the effect of initial metal concentration on metal biosorption was applied to the widely used Langmuir equation of adsorption isotherm (Fig. 2). 2.2.4 Reaction kinetics To examine metal biosorption kinetics, 0.5 gm of dry biomass was contacted with 50 ml of metal solution in 100-ml Erlenmeyer flasks. Flasks were incubated on a rotary 844

Biosorption

shaker at 30˚C. Samples of metal solutions were withdrawn from each flask at different time intervals (0-200 minutes) and metal content was analyzed by using the same formula (I). 3.

RESULTS AND DISCUSSION

3.1 Adsorption isotherm and biosorption characteristic In the biosorption studies of metal by M. rouxii, adsorption isotherm was studied under Langmuir and Freundlich adsorption isotherms. Plot of Log Ce/Q vs Log Ce gave the straight line with correlation value 0.96 (Fig.1), which indicates that the adsorption process obey the Langmuir equation. Ce/Q = Ce/Qmax + b.Qmax (II) where Ce be the equlibrium metal concentration at a fixed temperature and pH. Qmax be the maximum value of adsorption and b be the affinity of biomass for the metal. Extrapolation of log Ce/Q to Log Ce gave the value of Langmuir constant b. The values of b and Qmax were 81.23mg g-1 and 0.12 respectively. Qmax be the amount of adsorbate to form a complete monolayer on the adsorbate surface. Equilibrium sorption isotherm studies showed (Fig. 2) that metal uptake by M. rouxii was a chemically equilibrated and saturable mechanism. Thus there was an increase in metal uptake as long as binding sites were free [20, 21]. As the experimental data fit in the Langmuir model it indicates that biosorption of metals in the present study is characterized by a monolayer single type phenomenon with no interaction between sorbed metals [22]. Langmuir constant b is related to energy of adsorption through the Arrhenius equation, which also gives an indication of the affinity of the metal for binding sites on the biosorbent [23].

Figure 1. Adsorption isotherm 3.2 Kinetics study Kinetic experiments were necessary to determine the required time to reach the equilibrium condition [18]. Metal uptake data when plotted as a function of time (Fig. 2) at pH 5.5, showed that uptake was rapid in first 20 min of contact, and time required for attaining equilibrium was below 30 min. It could also be seen that the rate of uptake during the entire course of biosorption was independent of initial metal ion concentration 845

Biosorption

used. Thus, it is likely that kinetics of the process was influenced only by the step of metal transfer from solution to the binding sites [19].

Figure 2. Kinetics of metal adsorption 4.

CONCLUSION From the above study it can be stated that the Mucor rouxii biomass can adsorb lead ion from solution and this process is very fast. The equilibrium of the process reaches very quickly. It follows Langmuir adsorption isotherm. This biosorbent may provide a new technology of removal of lead ion from wastewater.

ACKNOWLEDGEMENT Authors are grateful to Dr. P.C.Banerjee, Deputy Director, Indian Institute of Chemical Biology, Kolkata-700032, India for his valuable suggestions. REFERENCES

1. B. Volesky, “Biosorption of heavy metals”, CRC Press, Inc. USA ISBN 0-8493-49176, (1990). 2. A. Esposito, F. Pagnanelli, F. Beolchini, V. Dovi “Cadmium and Copper biosorption on Sphaerotilus natans: influence of pH and biomass concentration on the biomass modeling” in: Biohydrometallurgy: Fundamentals, Technologyand sustainable development: V.S.T. Ciminelli and O. GarciaJr. Editors. Part-B, 2001, 89-97. 3. B.B. Nagar, S.P.Singh, “A comparative study of Pb(II) & Cr (II) ions removal by Solarthermal and Chemothermal activated Carbon” in ISSN, (5), 2002, 49-52. 4. V.G.S. Deniz, V.L. Da Silva, E.S. de Lima, C. M.A. de Aberu “ Lead biosorption in Arribadas Algal biomass” in: Biohydrometallurgy: Fundamentals, Technologyand sustainable development: V.S.T. Ciminelli and O. GarciaJr. Editors. Part-B, 2001,99117. 5. N. Aktar Md, S. Sastry, P. M. Mohan, “Biosorption of Ag2+ ions by processed Aspergillus Niger biomass” in Biotechnol. Letters Vol 17, (5) 1995, 551-556. 6. L. Mogollan, R. Rodriguez, W. Larrota, N. Ramirez, R. Torres, “Biosorption of Nickel using Filamentrous Fungi” in Applied Biochem. and Biotech.vol (70-72) 1998, 593601. 7. J.M. Lezcano, F. Gonazalez, I. Perez, M. Blazquez, J.A. Munez, A. Balleston, A. Hammaini “Use of waste biomass for decontamination of liquid effluents” in: 846

Biosorption

Biohydrometallurgy: Fundamentals, Technologyand sustainable development: V.S.T. Ciminelli and O. GarciaJr. Editors. Part-B, 2001,217-239. 8. A. Hammaini, J.M. Lezcano, F. Gonazalez, I. Perez, M. Blazquez, J.A. Munez, A. Balleston, “Activated sludge as biosorbent of heavy metals” in: biohydrometallurgy Proceedings of IBS 1999, 185-192. 9. E.W. Wilde, J. R. Benemann, Biotech. Adv., 11 1983 781-789. 10. D. Khummongol, G.S. Canterfordand C. Fryer, Biotech.Bioeng., 24. 11. J.M. Smith, Chemical Engneering Kinetics, 3rd Edn, Mc Graw-Hill, NY 1981. 12. J.C. Ballot and J.S. Condoret, Process Biochem., 28,1983 365. 13. W.Fritz and E.U. Schluender Chem Eng. Sci., 29, 1974,1279. 14. Y. Sag, T. Kutsal, “An overview of the studies about Heavy Metal Adsorption Processes by Microorganisms on the Lab Scale in Turkey.” in: biohydrometallurgy Proceedings of IBS 1999,307-316. 15. S.P. Mishra and G. Roy Chowdhury, “Kinetics of Cu Adsorption by Penicillum SP” Trans, Ins, Min Metal. (Sec C: Min Process. Extr. Metall.), 104,1995. 16. Puranik P.R. and Paknikar K.M. “ Biosorption of Lead Cadmium and Zinc by Citrobacter Strain MCM B- 181: Characterization study” Biotech Prog.15, 1999, 228237. 17. Y. Sag, and T. Kustal, Biotechnol. Letters, 11, 1989, 141. 18. Y. Sag, and T. Kustal, Process. Biochem., 32, 1997, 591. 19. Langmuir, I. The Adsorption of Gases on the Plane Surfaces of Glass Mica and Platinum. 20. Volesky, B. Holan. Z.R. “ Biosorption of Heavy Metals” Biotechnol. Prog. 11, 1995, 235-250. 21. Wong P.K. and C.M. So, “ Copper Accumulation by a strain of Pseudomonas Putida” Microbios 73: 1999 113-121. 22. F. Malekzadeh, “Accumulation of a Heavy Metals by a Bacterium isolated from Electroplating effluent”. Microbios. 69 1998 274-280. 23. Dodic, S.N. Popov S.D., Markov S.L. “ Insvestigation of Kinetics of Zinc Biosorption by Saccharomyces Cerevisiae cells” Nahrung. 45 (1): 2001, 59-61.

847

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Cadmium(II) biosorption by Aeromonas caviae: kinetic modeling M.X. Loukidou, T.D. Karapantsios, A.I. Zouboulis and K.A. Matis* Chemical Technology Division, School of Chemistry, Aristotle University, GR-54124 Thessaloniki Abstract The removal and depletion of cadmium from aqueous solutions by sorption on Aeromonas caviae particles was investigated in a well-stirred batch reactor. Kinetic experiments were performed at various initial bulk concentrations, biomass loads and temperatures, in order to investigate the probable mechanism of the process. It was seen that the sorption capacity is appreciably high for most experimental conditions, so Aeromonas caviae may be considered as a suitable sorbent for this application. A detailed analysis was conducted testing several chemical reaction and diffusion (external or intraparticle) kinetic models in order to identify a suitable kinetic model. In the present paper, the results obtained when using the so-called Ritchie 2nd order equation and also a pseudo 2nd order rate expression are given, with promising fitting of the experimental data.

Keywords: biosorption, modeling, cadmium, metal removal, wastewater 1.

INTRODUCTION The pollution of the environment with toxic heavy metals is spreading throughout the world along with industrial progress. Cadmium is one of the most toxic metals contaminating the environment, as it is widely used in rechargeable nickel-cadmium batteries, pigments, stabilisers, coatings, and alloys. The relevant E.U. Directive, as well as the U.S. E.P.A., have set the maximum contaminant level (MCL) for Cd(II) cations in domestic water supplies to be 5 µg.L-1 (Directive 98/36/EC). The commonly used treatment methods to remove Cd(II) ions from wastewaters include chemical precipitation, ion exchange, reverse osmosis and membrane processes. However, biosorption, the uptake of heavy metals by dead biomass, has gained credibility during recent years as it offers a technically feasible and economical approach [1,2]. Several biological materials investigated for heavy metals removal include bacteria, yeasts, algae and fungi [3,7]. The present study aims in examining biosorption by using biomass of Aeromonas caviae. Despite the fact that this microorganism is often present in groundwater and general in aquatic environments, little attention seems to have been given to its resistance to heavy metals [8]. The purpose of selecting this bacterium for studying biosorption was,

* To whom correspondence should be addressed. Fax: 00302310997794; E-mail: [email protected]

849

Biosorption

apart from its originality, to assess the possibility of utilizing a waste biomass for heavy metal removal. Equilibrium and kinetic analyses not only allow the estimation of sorption rates, but also lead to suitable rate expressions characteristic of possible reaction mechanisms [9,10]. Sorption kinetics may be controlled by several independent processes that can act in series or in parallel [11]. These processes fall in one of the following general categories: (i) bulk diffusion, (ii) external mass transfer (film diffusion), (iii) chemical reaction (chemisorption) and (iv) intraparticle diffusion. For sufficiently high agitation speed in the reaction vessel the bulk diffusion step can be safely ignored since then sorption onto sorbent particles is decoupled from mass transfer in the bulk mixture. Apart from that, it is quite common that more than one processes can contribute in the system performance at the same time. In this case, the extensive interrelationships between the various equations make the overall kinetic model exceedingly complicated to evaluate. A rather simplifying approach to circumvent this problem is to assume that each one of the co-current processes dominate over the others (i.e. the rate controlling step) at specific time regimes of the process and then study them independently [12] In order to identify the most appropriate mechanism for a process, several models apparently must be checked for suitability and consistency in a broad range of the system parameters. The model selection criteria proposed by Ho et al., [11] concerning sorption of pollutants in aqueous systems were used herein, as a general guideline. According to this, several reaction-based and diffusion-based models were tested in simulating our data. The finally chosen kinetic models are those, which not only fit closely the data, but also represent reasonable sorption mechanisms. 2.

MATERIALS AND METHODS

2.1 Biomass and grown condition Aeromonas caviae, a Gram-negative bacteria isolated from the raw water wells sample by enrichment culture technique. Culture units of microorganism were identified by Dr. J.M. Tobin (School of Biological Sciences, Dublin City University). The strain was grown at 29°C in a rotating shaker for 24 h in liquid medium containing: yeast extract (0.5% w/v), tryptone (1% w/v), NaCl (0.5% w/v) and FeSO4.7H2O (0.2 g.L-1). The produced biomass was separated by centrifugation at 3000 rpm, washed several times by a solution of NaCl (0.9% w/v), sterilized and stored. 2.2 Biosorption experiments Batch biosorption experiments were carried out at different: biomass feed (0.5, 1 and . -1 2 g L ), initial cadmium concentration (5 and 50 mg.L-1), temperature (20, 40 and 60°C). The experiments were performed in an Erlenmeyer flask at a 180-rpm agitation speed (Heidolph type, RZR 2102). This speed was selected to ensure that the effect of external film diffusion on biosorption rate is not significant and can be ignored in any analysis. The initial pH of the solution was adjusted to optimum value of 7, as it was determined during preliminary experiments. i.e. from the precipitation value, as metal hydroxide according to the aqueous speciation. For the equilibrium experiments, ample time (~2 h) was allowed after the beginning of adsorption to ensure that the experiments were concluded. For the kinetic study of cadmium adsorption, 2 mL samples were acquired at selected time intervals using a 10 mL syringe. Due to experimental constraints the sampling interval

850

Biosorption

was no less than 2 min. The residual concentration of cadmium in all samples was analyzed by atomic absorption spectrophotometry (AAS, Perkin-Elmer, model 2360). 3.

RESULTS AND DISCUSSION

3.1 Equilibrium experiments Experimental adsorption isotherms of cadmium ions obtained at several biomass loads and temperatures are presented in Figures 1a and 1b, respectively. For each isotherm the initial metal concentration was varied, while the biomass load and temperature was kept constant. 140

150

120 100

100

80 60 50

o

X:2 g L X:1 g L

o

T:40 C

-1

20

o

T:20 C

T:20 C

40

-1

X:0.5 g L

0

o

-1

T:60 C

X:1 g L

-1

0 0

50

100

150

200

250

0

50

100

-1

150

200

250

-1

Ceq (mg L )

Ceq (mg L)

Figure 1. Application of equilibrium Langmuir model to cadmium biosorption: (a) at different biomass concentrations; (b) at different temperatures

The results of fitting the Langmuir and the Freundlich models to these data are presented in Table 1. Table 1. Langmuir and Freundlich model regression constants Conditions T (°C) 20 20 20 40 60

Biomass load (g.L-1) 0.5 1 2 1 1

Langmuir constants

Freundlich constants

q bC q e = max e 1 + bC e

1n

qe = K F Ce

Kf (g.L-1)

N (-)

r2

qmax (mg.g-1)

b (L.mg-1)

r2

20.37 10.85 12.10 11.68 27.66

2.59 2.12 3.06 2.09 2.84

0.87 0.94 0.91 0.94 0.92

181.91 155.32 68.17 175.11 160.31

0.03 0.019 0.06 0.018 0.06

0.93 0.99 0.99 0.98 0.97

The higher values of the correlation coefficient (r2), show that the Langmuir equation is very suitable for describing the biosorption equilibrium of Cd(II) by A.caviae in the studied concentration range. The magnitude of the biosorption capacity, qmax, spans a range of values (68.17 to 175.11 mg.g-1) comparable to other types of sorbent reported earlier [13].

851

Biosorption

In addition, the large values of b clearly imply that A. caviae exhibits a high affinity for cadmium ions. The biosorption capacity, qmax, decreased with the increasing biomass load, indicating a poorer biomass utilisation (low efficiency). This may be attributed to a possible aggregation of solids occurring at higher biomass loads, capable of reducing the effective sorption area. The value of qmax (175.11 mg.g-1) obtained at 40°C appears to be higher in comparison with the uptakes obtained at the other temperatures. However, one should withhold judgement until experimental information for higher Ceq values is acquired. 3.2 Kinetic experiments Figure 2 presents the remaining concentration of cadmium in the bulk solution as a function of time at different experimental conditions. Unless differently stated, runs were performed at 20°C and with 1 g.L-1 biomass load. The key role of the initial metal concentration is apparent. That is, the adsorption capacity is markedly enhanced, at a higher initial concentration of Cd (II), since the initial concentration provides an important driving force to overcome mass transfer resistance of Cd (II) between the aqueous and solid phase. 60 X:0.5 g L

(a) 5

X:1 g L

-1

(b)

X:2 g L

X:0.5 g L

50

-1

X:1g L -1

X:2 g L

40

o

-1

o

T:40 C

3.5

-1

-1

T:40 C o

o

T:60 C

T:60 C

30

2

20 10

0.5 0

20

40

60

80

t (min)

100

120

140

0 0

20

40

60

80

100

120

140

t (min)

Figure 2. Cadmium biosorption kinetics by Aeromonas caviae obtained at initial cadmium concentrations of: (a) 5 mg.L-1; (b) 50 mg.L-1 (pH: 7, agitation speed: 180 rpm)

The qualitative resemblance among experimental data is considerable despite the different experimental conditions. Overall, there is a monotonous decreasing trend with time. The very steep descent at the beginning of sorption is succeeded by a less rapid decay down to about 20-30 minutes. From then on, the Cd (II) concentration gradually levels-off and remains almost constant till the end of the experiment (120 min). Thus, the major part of adsorption takes place within the first 30 minutes of the process. Moreover, the time required to reach the final equilibrium is practically the same for all experimental conditions. It is evident that adjusting the temperature or the biomass load in the examined range of values can lead to a comparable metal removal. In order to examine further whether the sorption of cadmium follows a mechanism of electrostatic or chemical nature, some biosorption experiments were performed by adding various concentrations of a nitrate salt. The impact of the presence of dissolved nitrate ions on the kinetics of Cd(II) biosorption is shown in Fig. 3. It appears that as the dosage of salt increases both the sorption capacity and sorption rate of Cd(II) ions decrease. This 852

Biosorption

effect is more intensive with the higher initial metal concentration (50 mg.L-1). The added ionic background alters both the equilibrium and the kinetic behavior of the sorbate/sorbent system. The effect of the ionic strength may be explained as the outcome of the competition between sodium and cadmium cations during electrostatical binding to the biomass. Co: 50 mg L

-1

Co: 50 mg L

-1

,0 M NaNO

3

,0.01 M NaNO

3

-1

Co: 50mg L ,0.5 M NaNO

3

Co: 5 mg L

-1

Co: 5 mg L

-1

Co: 5 mg L

-1

,0 M NaNO

,0.5 NaNO

50

10

40

8

30

6

20

4

10

2

0 0

20

40

60

80

100

120

3

,0.01 M NaNO

3

3

0 140

t (min)

Figure 3. Biosorption kinetic curves obtained for different concentrations of NaNO3 at 20°C 3.3 Kinetic modeling The relatively short duration of the present experiments, apart from the process advantages, is a first indication that adsorption of cadmium ions on A. caviae is a chemical reaction rather than a diffusion controlled process [11]. From surface titration experiments and analyzing IR spectrum, it was illustrated that the cell wall contains two or more main functional groups (carboxyl and phosphate) responsible for the uptake of heavy metals [14]. These types of groups one capable of removing metallic ions, usually cations, from aqueous solutions through the application of different mechanisms, such as cell surface sorption (complexation, surface precipitation etc) [5]. From the chemical reaction category (chemisorption), the best fit for the data sets of this study is achieved by 2nd order-type chemical reactions. The solution of the standard 2nd order reaction based on a constant stoichiometry of one metal ion per binding site, is [12]: Ct =

Co C 1 − o exp(− k 2 C e t ) Ce

(1)

where k2 is the reaction rate constant [L*(mg-1of metal)*min-1]. This adsorption model has been very effective in describing the kinetics of adsorption of gases on solids [12]. Figure 3 shows that equation (1) clearly fails to capture the steep concentration gradient of the early removal stage. This is a direct indication that adsorption on solids from a liquid phase is a different process than adsorption from a gas phase where traditionally the remaining bulk concentration dictates the kinetics [12]. 853

Biosorption

50

5 4.5

X:1 g L

4

eq. (1)

40

(a)

X:2 g L

3.5

X:1 g L eq. (1)

-1

-1

X:0.5 g L

2.5

-1

(b) -1

X:0.5 g L eq. (1)

30

eq (1)

3

X:2 g L eq. (1)

-1

-1

20

eq (1)

2

10

1.5 0

1 0

20

40

60 t (min)

80

100

120

140

0

20

40

60

80

100

120

140

t (min)

Figure 4. Cadmium biosorption kinetics obtained at initial cadmium concentrations of (a) 5 and (b) 50 mg.L-1

If the rate of sorption depends not on bulk concentration but on uptake by the sorbent this can be described by the so-called Ritchie 2nd order equation according to which one metal ion occupies two binding sites [15]: ⎧ ⎡ 1 ⎤⎫ qt = qe ⎨1 − ⎢ ⎥⎬ ⎩ ⎣1 + k 2 t ⎦ ⎭

(2)

where qt and qe are the amounts of adsorbed metal ions on the biosorbent at time t and at equilibrium, respectively (mg g-1) and k2 is the reaction rate constant [min-1]. When in the above treatment it is not necessarily qe that dictates the sorbate uptake then a pseudo 2nd order rate expression is more appropriate [11]: t 1 1 = + t qt k q 2 qm m m

(3)

where km is the reaction rate constant [g of biomass*(mg-1of metal)*min-1] and qm is a numerically determined parameter which under ideal 2nd order rate control corresponds to qe. However, equation (2) and (3) provide a quite suitable description of data for advancing time (Figures 4, 5). It is noteworthy that both models adequately capture the rapid rate of adsorption during the first minutes of the experiments. This already implies that the metal uptake by the sorbent is a satisfactory rate-controlling parameter under a 2nd order reaction mechanism. Table 2 displays the numerically best-fit values of the rate parameters of equations (2) and (3). The predicted equilibrium sorption capacities are quite close to the experimental values for both models. Nevertheless, the rate constant of the pseudo 2nd order model, km, is monotonously correlated with changes in the biomass load and in the bulk concentration, features that have been encountered in the past regarding biosorption [15]. On the contrary, the rate constant of the Ritchie 2nd order equation, k2, fluctuates beyond any physical reasoning. In addition, equation (3) exhibits better fitting statistics. Despite the goodness of fit for sorption at 40 and 60°C, the reaction rate constant of both models varies randomly with temperature. Preliminary calculations using the Arrhenius model between two temperatures every time gave activation energies always below 10 kJ/mol which is far less than what is expected for reaction controlled sorption processes [11]. The 854

Biosorption

morphological changes of the biomass surface at different temperatures and the dependence of sorption capacity on temperature may be blamed for this irregularity.

7 50

6 40

5 4

-1

X:1 g L

30

-1

o

X:2 g L

3

-1

X:0.5g L

2

T:60 C

20

-1

X:0.5 g L -1

o

T:40 C

(a)

1

o

T:60 C

X:2 g L

(b)

-1

10

X:1g L o

T:40 C

0

0 0

20

40

60

80

100

0

120 140

20

40

60

80

100

120

140

t (min)

t (min)

Figure 5. Comparison of experimental uptake curves against theoretical predictions based on the Ritchie 2nd order equation (equation 2) at initial cadmium concentrations of: a) 5 mg.L-1; b) 50 mg.L-1(pH:7)

70

6 X:1 g L

60

-1

X:0.5 g L

50

X:2 g L

X:1 g L

5

-1

X:0.5 g L

-1

(a)

X:2g L

4

-1

-1

(b)

o

o

T:40 C

T:40 C

40

-1

3

o

T:60 C

o

T:60 C

30 2

20 1

10

0

0 0

20

40

60

80

t (min)

100

120

140

0

20

40

60

80

100

120

t (min)

Figure 6. Comparison of experimental uptake curves against theoretical predictions based on the pseudo 2nd order equation (equation 3) at initial cadmium concentrations of: a) 5 mg.L-1; b) 50 mg.L-1(pH:7)

855

Biosorption

Table 2. Kinetically determined parameters and comparison with equilibrium sorption capacities Co (mg.L-1)

5

50

Conditions T (°C) 20 20 20 40 60 20 20 20 40 60

Biomass (mg.L-1) 0.5 1 2 1 1 0.5 1 2 1 1

Equilibrium qeq (mg.g-1) 6.65 3.53 1.9 3.44 4.33 45.11 32.32 21.74 36.20 45.87

Pseudo 2nd order eqn qeq (mg g-1) 6.22 3.66 1.94 3.98 4.33 56.18 43.48 25 41.67 43.47

km (mg.g-1.min-1) 0.133 0.192 0.373 0.118 0.132 0.015 0.034 0.051 0.034 0.132

Ritchie 2nd order eqn r2 0.995 0.991 0.989 0.993 0.995 0.986 0.990 0.987 0.999 0.999

qeq (mg.g-1) 6.02 3.50 1.85 3.98 4.36 48.73 37.25 21.79 40.73 43.21

k2 (min-1) 1.2 1.15 1.25 0.41 0.48 1.03 2.20 1.37 2.06 11.31

r2 0.981 0.975 0.974 0.992 0.985 0.993 0.998 0.997 0.992 0.999

4.

CONCLUSIONS Dead cells of Aeromonas caviae showed a high biosorption capacity for cadmium(II) ions comparing with other types of biosorbent. The present results demonstrate that temperature, initial metal concentration and biomass load highly affect the uptake capacity of the biosorbent.. The Freundlich and Langmuir adsorption models were tested for the mathematical description of the biosorption equilibrium of Cd(II) ions to A. caviae of various temperatures and biomass loads. The calculated isotherm constants were used to assess the biosorptive capacity of the biomass. The obtained results showed that the adsorption equilibrium data fitted fairly well to the Langmuir model in the examined concentration range. The suitability of the pseudo 2nd order chemical reaction for the sorption of Cd(II) ions onto biomass is also presented. This kinetic model may effectively describe the largest part of the process. The obtained kinetic expression is of a great practical value for technological applications, since kinetic modelling successfully replaces time- and material-consuming experiments, necessary for process equipment design. ACKNOWLEDGEMENTS Many thanks are due to Dr. John M. Tobin (School of Biological Sciences, Dublin City University) for his help with the microbiological identification of the microorganism used. REFERENCES

1. D. Kratochvil and B. Volesky, Trends Biotechnol, 16 (1998) 291. 2. T.J. Butter, L.M. Evison, I.C. Hancock, F.C. Holland, K.A. Matis, A. Philipson, A.I. Sheikh and A.I. Zouboulis, Wat Res., 32 (1998) 400. 3. A.I. Zouboulis, N.K. Lazaridis and K.A. Matis, J Chem Technol Biotechnol., 77 (2002) 958.

856

Biosorption

4. A.I. Zouboulis, E.G. Roussou, K.A. Matis and I.C. Hancock, J Chem Technol Biotechnol., 74(1999) 429. 5. F. Veglio and F. Beolcini, Hydrometallurgy, 44 (1997) 301. 6. B. Volesky, Hydrometallurgy, 59 (2001) 203. 7. A.I. Zouboulis, K.A. Matis, M.X. Loukidou and F. Sebesta, Colloids and Surfaces A: Physicochem. Eng. Aspects, 212 (2002) 185. 8. C.D. Miranda and G. Castillo, Sci Total Envir., 224 (1998) 167. 9. B. Volesky and Z.R. Holan, Biotechnol. Progr., 11 (1995) 230. 10. S. Yiacoumi and C. Tien, Kinetics of Metal Ion Adsorption from Aqueous Solutions. Kluwer Academic, Boston, 1995. 11. Y.S. Ho, J.C.Y. Ng and G. McKay, Sep Purific Methods, 2 (2000) 189. 12. E.H. Smith, Wat Res., 30 (1996) 2424. 13. A. Esposito, F. Pagnanelli, A. Lodi, C. Solisio and F. Veglio, Hydrometallurgy, 60 (2001) 129. 14. M.X. Loukidou, T. D. Karapantsios, A.I. Zouboulis and K.A. Matis, submitted for publication, J. Chem. Tech. Biotechnol. 15. C.W. Cheung, J.F Porter and G McKay, Wat. Res., 35 (2001) 605.

857

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Chromium uptake by pretreated cells of Aeromonas hydrophila isolated from textile effluents Zainul Akmar Zakaria and Wan Azlina Ahmad* Department of Chemistry, Faculty of Science, Universiti Teknologi Malaysia, 81310 Skudai, Johor Darul Takzim, Malaysia Abstract Effluent discharges from textile and dyestuff industries create serious risk of pollution to the environment. This is due to the presence of various types of pollutants including inorganic compounds and polymers, organic products and also heavy metals. Even though more attention was focused on colour, the presence of heavy metals should not be discounted due to its toxic effects through long-term accumulation. In this study, a locally isolated bacteria from wastewater discharge of a textile-based manufacturing industry identified as Aeromonas hydrophila was used as biosorbent to remove chromium species from simulated solution. Four types of pretreatment were used; autoclaving, 50%v/v acetone, 0.1N HCL and UV. The treated bacteria, plus the viable but non-culturable (VBNC) cells was then assessed for Cr (III) uptake ability using the shake-flask technique at pH 4.5. The application of Langmuir adsorption isotherm showed that the autoclavedacid treated cells gave the highest uptake capacity compared to others with a qmax value of 26.42 mg/g cell dry wt at around 1 hour of equilibration time. Pretreatment of the cells showed a slightly higher surface area for the autoclaved-acid treated and autoclavedacetone treated cells compared to the UV and untreated cells respectively.

Keywords: uptake, chromium, pretreatment, textile, Aeromonas 1.

INTRODUCTION The discharge of heavy metals into the environment especially from the industrial and agricultural source is of main concern lately. Heavy metal contamination exists in aqueous waste streams of many industries, such as metal plating facilities, textile industries and tanneries [1]. Simply, any industrial activity that utilizes metals in some part of its operation would have a metal disposal problem. This is because heavy metals tends to form toxic, carcinogenic or mutagenic compounds even in very low concentrations [2] which can accumulate through the food chain, leading to serious ecological and health problems [3,4]. Chromium (Cr) for instance is often present in the wastewaters of metal finishing parts, electroplating and textile industries, frequently at ppb level. In the textile industry, Cr compounds are contained in Cr complex dyes where it was introduced to the dyestuff

*

Corresponding author: [email protected]

859

Biosorption

molecules during the dyeing process i.e. chromate treatment in the textile industry. From all the manufacturing processes, Cr can seep into the environment directly via the effluents or indirectly via a slower release as solid wastes [2]. Chromium is found either as Cr(II), Cr(III) and Cr(VI) in water. The divalent state is unstable with respect to evolution of hydrogen, trivalent state has broad stability while hexavalent chromium occurs under strong oxidizing conditions [5] which accounts for it being more toxic to microorganisms such as bacteria [6] and fungi [7]. Cr(VI) is normally present in its chromate (CrO4-) or dichromate (Cr2O72-) forms. Biological membranes are impermeable to Cr(III) but permits the penetration of Cr(VI) into the cell cytoplasmic region where it can be reduced and precipitated there as Cr(III) which readily forms insoluble hydroxide compounds at pH 7.5 [5,8]. Amongst the proven methods for removing Cr from industrial wastes solution includes chemical precipitation, chemical oxidation or reduction, filtration, electrochemical treatment, membrane and evaporation technology [1,9-12]. The use of ion-exchange resins has its limitations in terms of low selectivity in metals recovered and is expensive. This makes the natural biological metal-microbe interactions such as biosorption, bioprecipitation, biodegradation and bioaccumulation an interesting yet feasible alternative treatment processes. Amongst the mechanisms suggested for the interactions are methylation, chelation, adsorption/ absorption, complexation and redox reaction [13]. A number of researchers have reported the use of fungus [14], seaweed [15,16], algae [17], cone biomass [18] and bacteria [19] for the biosorption of chromium from simulated or real industrial effluents. In this study, the biosorption of Cr by the Gram-negative bacterial cells of A. hydrophila, which was isolated from a Cr-laden textile effluent, was investigated. The effect of physical and chemical treatment on the cells Cr-uptake ability was elucidated using the simplest mode of Langmuir and Freundlich sorption isotherms i.e. without taking into considerations the effect of pH and the ion-exchange situation in the sorption process. Possible changes in the overall surface area of the bacterial cells due to the various treatments were also assessed. 2.

MATERIALS AND METHODS Unless otherwise stated, all experiments were performed in duplicates. Glasswares used were washed with 10%v/v HNO3 and rinsed with deionized water. All reagents used were of analytical grade. 2.1 Bacteria A. hydrophila used in this study was isolated from a local textile-based industrial effluent and was identified using the API20NE kit (Biomerieux, France). It was grown in nutrient broth (Merck, Germany) until early stationary phase at 200rpm in a Certomat-R, B.Braun orbital shaker at 30°C. 2.2 Pretreatment of the cells A. hydrophila cells grown to early stationary phase was harvested by centrifugation at 9000rpm, 10mins and 0°C (SIGMA 2K-15, B. Braun).The cell pellet obtained was resuspended in 5mL distilled deionized water (UHQII, Elgastat) and was then termed as VBNC (viable but non-culturable) cells.

The VBNC cells were also made non-living by autoclaving at 121°C, 101.325 kPa for 15 minutes (Fedegari, Italy) and by exposing to UV-irradiation. The autoclaved cells was 860

Biosorption

then subjected to treatment using 0.1N HCL and 50%v/v acetone respectively and termed as HA and HAt cells respectively. The cell pellet was washed twice using distilled deionised water before final suspension in the respective medium. 2.3 Specific surface area determination The effect of pretreatment was further investigated using the methylene blue adsorption method [20]. A stock solution (1mM, 0.01 to 0.14mL) of methylene blue (C16H18ClN3S.xH2O, 319.86gmol-1) was added to 20mg dry wt. of VBNC, UV-treated, HA and HAt cells suspension each in a 25mL polyethylene centrifuge tube. The volume was made up to 20mL using deionized water. Final methylene blue concentrations prepared ranged from 0.5 to 7µM. The mixture was then shaken in a rotary water bath shaker at 100rpm, 30°C for 4 hours. Centrifugation (SIGMA 2K-15, B.Braun) at 8225rpm, 10mins and 0°C was carried out to separate the cells from the mixture. The filtrate was analyzed spectrophotometrically (SPECTRONIC 21-D) at 661nm to determine the residual concentrations of methylene blue in solution. The amount of methylene blue adsorped by the cells was determined based on the difference between the initial and residual methylene blue concentrations. In determining the surface area of A. hydrophila cells, it is assumed that a complete monolayer of methylene blue was formed at the bacterial surface when the adsorption profile reached a plateau. To assess the reaction, the Langmuir adsorption isotherm was applied:

R=

Tm KCeq 1 + KCeq

(1)

where R is the amount of adsorbed methylene blue (µmol/g), Ceq is the concentration of methylene blue at equilibrium (µmol/ L), K is a constant related to the energy of adsorption (mg/L) and Tm is the amount of methylene blue needed to form a complete monolayer on the bacterial surface (µmol/g). The Tm value was then used to determine the bacterial surface area, S (µm2) according to the following equation: S = Tm N Aσ −1

(2)

where NA is the Avogadro no. (6.02 x 1023 molecules per mol) and σ is total area of methylene blue (0.55 x 10-18 m2) when a complete monolayer was formed [20]. 2.4 Chromium uptake experiment The chromium uptake experiment was carried out in 6, 250mL Erlenmeyer flasks. A. hydrophila, 20 mg cell dry wt. were mixed with 25mL of Cr(III) solution prepared by dissolving CrCl3.6H2O in distilled deionized water. The final concentrations of the Cr(III) solutions used ranged between 5-200 mg/L. The pH of the Cr solution was adjusted to 4.5 (WTW, Germany) using 0.1N HCL or 0.1N NaOH. The mixture was then shaken (Certomat-R, B.Braun) for 12 hours, 100rpm and 30°C. It was then centrifuged at 9000rpm, 5mins and 0°C and the filtrate collected was analyzed for Cr using AAS (Philips PU9100X). The amount of Cr removed was determined by the difference between initial and residual concentrations. A time-course study of the biosorption process was also investigated using living (VBNC) and non-living cells of A. hydrophila. The bacterial cell, 20 mg cell dry wt. were contacted with 25mL of Cr(III) solution in 6, 250mL Erlenmeyer flasks. Different 861

Biosorption

concentrations of Cr(III) were used i.e. 150mg/L and 50mg/L for the VBNC and HAt cells respectively. The cell suspension was shaken at 100rpm, room temperature with incubation times ranging between 1-12 hours. It was then centrifuged at 9000rpm, 5mins and 0°C and the filtrate collected was analyzed for Cr using AAS (Philips PU9100X). 3.

RESULTS AND DISCUSSIONS

3.1 Bacteria In this study, A. hydrophila was used as the biosorbent to remove Cr species from solution. It was originally isolated from a local textile-based (batek) manufacturing effluent using procedure suggested by Greenberg et al., 1985. Identification of the bacteria was made using the API20NE kit (Biomerieux, France) which gave between 99.7-99.9% positive results as A. hydrophila. This Gram-negative bacterium whose normal habitat include soil and water [21] is a member of the family Vibrionaceae. It is one of the frequently isolated species from drinking water due to its ability to withstand the chlorination procedure, thus recolonizing the water distribution networks [22]. 3.2 Specific surface area determination Methylene blue adsorption method was used to determine the specific surface area of pretreated cells of A. hydrophila namely VBNC, autoclaved, autoclaved-acid treated (HA) and autoclaved-acetone treated (HAt) cells. The basis of using this methodology for a Gram-negative bacterium has been reported elsewhere [23]. Table 1 shows profiles obtained from the Langmuir adsorption analysis: Table 1: Profile for the methylene blue adsorption based on Langmuir isotherm Cells Parameter Regression analysis, R2 R, Lµmol-1 (adsorbed MB) K, µM (energy of adsorption) S, m2mg-1 (bacterial surface area)

VBNC

UV-treated

HA

HAt

0.761 1.367 1.339 0.020

0.984 1.124 1.793 0.019

0.910 1.600 0.434 0.026

0.994 1.446 0.310 0.024

*VBNC-viable but non-culturable, HA- autoclaved and acid-treated, HAt- autoclaved and acetone treated

Based on table 1, it was noted that the surface area for HA and HAt cells were quite similar i.e. 0.026 and 0.024 m2.mg-1 respectively which might indicate the role of autoclaving in creating a larger surface area for the bacterial cell. UV-treatment did not impose any significant effect on the overall surface area based on values obtained for the VBNC cell. Zakaria, Z.A. [24] has reported the disappearance of an extracellular polymeric substance (EPS) from Thiobacillus ferrooxidans upon autoclaving, which has resulted in a 2-fold increase in the overall specific surface area measured as compared to the acidtreated cells only. Acidic treatment was not expected to cause significant physical modification on the cell, but rather towards the overall charges of the cell wall; by protonation of potential binding sites especially from the amino acid fraction of the bacteria [25] and to yield pure amino sugars such as D-glucosamine which could aid in the metal binding process [14]. Acetone treatment used did not cause significant changes to 862

Biosorption

the physical properties of the bacterial cell surface, which contradicts the findings of Bai and Abraham [14] who worked with the fungus, Rhizopus nigricans. This could be due to different cell wall properties for both fungus and bacteria respectively. The acetone treatment effect was more pronounced in Rhizopus nigricans due to the significant removal of lipids and proteins; these components are found in larger quantities in fungus as compared to bacteria. 3.3 Dynamics of chromium biosorption The time course experiment of chromium biosorption by A. hydrophila cells is shown in Figure 1. 160 140

q, uptake (mg/g)

120 100 80 60 40 20 0 0

2

4

6

8

10

12

14

hours Figure 1. Dynamics of chromium biosorption by A. hydrophila. (•) – VBNC, ( ) – HAt, pH - 4.5, [cell] - 20mg cell dry wt

Two types of A. hydrophila cells namely VBNC and HAt contacted with different concentrations of Cr (III) i.e. 150mg/L and 50mg/L respectively were used in the time course study. A higher concentration of Cr (III) was used for the VBNC cells based on the possibilities of a two-phase metal removal by these cells [26] hence, the ability to remove higher amounts of the metal. This was clearly indicated from the Cr uptake profiles shown in figure 1 where the VBNC cells showed much higher capacity to remove Cr from solution with a maximum uptake (qmax) value of 135.75mg/g cell dry wt. compared to 32.75mg/g cell dry wt. for the HAt cells. The HAt cells showed a favourable uptake profile where most of the metal were removed after 1 hour at a pH of around 4.5. This is comparable with the findings of other researchers; Ecklonia algae biomass – 12 hours, pH 4.0 [15] Sargassum sp. – 6 hours, pH 4.0 [16]. However, this was not the case for the VBNC cells where a saturation condition was not observed even after 12 hours. This could be attributed to the two-step metalbacteria interaction as suggested by Huang, et al [26]. The first step involving surface binding or extracellular association is rapid whilst the second step involving intracellular metal uptake is slow and might be the rate-limiting step. This is an energy consuming process and can be explained as follows. As there exist a difference in pH inside the 863

Biosorption

cytoplasmic membrane (pH 6.0) and the external environment (pH 4.0), a natural proton motive force exists which can play a role in ATP synthesis [27]. The ATP synthesized could well be the energy source required for the translocation of chromium. 3.4 Biosorption of chromium Figure 2 shows the Cr (III) biosorption profiles by pretreated cells of A. hydrophila. 30

qmax, uptake (mg/g)

25 20 15 10 5 0 0

50

100

150

200

Ceq (mg/L)

Figure 2: Profiles for the biosorption of Cr (III) by pretreated cells of A. hydrophila. (•) – HAt, ( ) – HA, (▲) – UV-treated, pH - 4.5, [cell] - 20mg cell dry wt

From figure 2, the non-living cells of A. hydrophila (HAt, HA and UV-treated) showed a good agreement to the Langmuir adsorption isotherm used as opposed to the living cells (VBNC). The inability to describe the uptake phenomenon by the VBNC cells might be attributed to the different types of metal deposition on the bacterial surface i.e. multilayer-type compared to the suggested monolayer-type of metal deposition. Based on figure 2, the HAt cells showed the highest capacity to remove Cr (III) from solution with qmax of 26.42mg/g cell dry wt. This was followed by the HA and UV-treated cells with qmax values of 20.50 and 18.60mg/g cell dry wt. respectively. This finding is in agreement with results from the specific surface area determination work which further supported the assumption that a larger surface area would lead to higher metal binding capacity [14, 28, 29]. The Langmuir parameters were analyzed using the nonlinear regression method and are summarized in Table 2: Table 2: Langmuir adsorption parameters for cells of A. hydrophila. Cells HAt HA UV-treated

qmax, (mg/g) 29.33 29.94 20.41

b, (mg/L) 0.055 0.013 0.062

R2 0.992 0.983 0.862

In elucidating the biosorption process of heavy metal ions, metal speciation especially its hydrolysis in water is an important factor to be considered. The metal speciation is closely related to the solution’s pH, hence, the overall valence of ionic charges which is 864

Biosorption

considered as two of the most crucial factor in biosorption together with biomass concentrations [15]. In this study, pH of the mixture was adjusted to 4.5 where most of the chromium is expected to exist as Cr3+ and CrOH2+ as shown in Equation 3 [15]. Cr3+ + H2O ↔ CrOH2+ + H+ (3) This might suggest the predominance of two-types of chromium removal mechanism i.e. ion-exchange and direct electrostatic interaction. For a bacterial species, most of the metal-binding properties can be attributed to the amino acid fraction of the bacterial cell wall emphasizing on potential binding sites such as carboxylate, amino and sulfhydryl groups. At pH 4.5, most of the carboxylic group would be deprotonated (pK, 1.8-2.9) resulting in negatively charged carboxylate ions which could serve as potential binding sites for the positively charged chromium ions. The involvement of amino groups was not taken into consideration due to the predicted repulsive action with the positively charged chromium ions while the contribution of sulfhydryl might not be significant due to its low distribution in a Gram-negative bacteria (Thiobacillus ferrooxidans, unpublished data). The possibilities of ion-exchange should not be discounted either especially between CrOH2+ and H+ from the carboxyl groups. This assumption was made based on the condition of CrOH2+ being the dominant species (~74%) at pH of around 4.0 [15]. Kratochvil et al. [16] have reported the equilibrium constant for equation 3 to be at pK of 3.82, approximating the presence of 50% of the total Cr in solution as CrOH2+. However, due to limitation on the scope of this study, further elucidation on the ion-exchange process between Cr and biomass binding sites was not carried out. 4.

CONCLUSION Based on this study, it can be concluded that A. hydrophila might serve as a useful biosorbent to remove Cr species from solution. Acetone treatment of the autoclaved cells has resulted in the highest Cr uptake capacity (26.42 mg/g cell dry wt.) with an overall surface area of 0.024m2/mg compared to the untreated bacterial cell. However, more work needs to be carried out in order to apply the A. hydrophila cells for the treatment of Cr from industrial effluent. ACKNOWLEDGEMENTS The authors would like to thank the Ministry of Science, Technology and Innovation for funding the project and the NSF scholarship award to Zainul Akmar Zakaria. REFERENCES

1. Bailey, S.E., Olin, T.J., Bricka, R.M. and Adrian. D.D. (1999) “A Review of Potentially Low-Cost Sorbents for Heavy Metals” Water Research. 33: 11. 2469-2479. 2. Merian, E. (1991) “Metals and Their Compounds in the Environment: Occurrence. Analysis and Biological Relevance” VCH Publishers Inc. New York. USA. 3. Mellado, R.P., Martinez, J.L. and Hernandez, A. (1999) “Metal Accumulation and Vanadium-Induced Multidrug Resistance by Environmental Isolates of Escherichia hermannii and Enterobacter cloacae” Applied and Environmental Microbiology. American Society for Microbiology. 65. 2: 489-498. 4. Volesky, B. (1987) “Biosorbents for Metal Recovery” Trends in Biotechnology. Elsevier Publications. Cambridge. Apr. 5: 96-101. 5. Baes, C.F. (1976) “The Hydrolysis of Cations” John Wiley and Sons, Inc. USA. 865

Biosorption

6. Petrilli, F.L. and DeFlora, S. (1977) “Toxicity and Mutagenicity of Hexavalent Chromium On Salmonella typhimurium” Applied and Environmental Microbiology. 33:805-809. 7. Babich, H., Schiffenbauer, M. and Stotsky, G. (1982) “Comparative Toxicity of Trivalent and Hexavalent Chromium to Fungi” Bulletin of Environmental Toxicology. 28:452-459. 8. Rapoport, A.I. and Mutter, O.A. (1995) “Biosorption of Hexavalent Chromium by Yeasts” Process Biochemistry. 30. 2: 145-149. 9. Volesky, B. (1990) “Biosorption of Heavy Metals” Boca Raton. Florida: CRC Press 10. Veglio, F. and Beolchini, F. (1997) “Removal of Metals by Biosorption: A Review” Hydrometallurgy. Elsevier Science Ltd.. 44: 301-316. 11. Volesky, B. and Holan, Z.R. (1995) “Biosorption of Heavy Metals” Biotechnology Progress. 11: 235-250. 12. Zouboulis, A., Rousor, E.G., Matis, K.A. and Hancock, I.C. (1999) “Removal of Toxic Metals from Aqueous Mixtures. Part 1: Biosorption” Journal of Chemical Technology and Biotechnology. 74: 429-436. 13. Mann, H. (1990) “Biosorption of Heavy Metals by Bacterial Biomass” in Volesky. B. (ed.). Biosorption of Heavy Metals. CRC Press. Boca Raton. Florida. USA. 14. Bai, R.S. and Abraham, T.E. (2002) “Studies on Enhancement of Cr (VI) Biosorption by Chemically Modified Biomass of Rhizopus nigricans”. Water Research. 36:12241236. 15. Yun, Y.S., Park, D.H., Park, J.M. and Volesky, B. (2001) “Biosorption of Trivalent Chromium On the Brown Seaweed Biomass”. Environmental Science and Technology. 35: 4353-4358. 16. Kratochvil, D, Pimentel, P. and Volesky, B. (1998) “Removal of Trivalent and Hexavalent Chromium by Seaweed Biosorbent”. Environmental Science and Technology. 32: 2693-2698. 17. Gupta, V.K., Shrivastava, A.K. and Neeraj Jain. (2001) “Biosorption of Chromium (VI) From Aqueous Solution by Green Algae Spirogyra Species”. Water Research. 35. 17: 4079-4085. 18. Ucun, H., Bayhan, Y.K., Kaya, Y., Cakici, A. and Algur F. (2002) “Biosorption of Chromium (VI) From Aqueous Solution by Cone Biomass of Pinus sylvetris”. Bioresource Technology. 19. Lodi, A., Solisio, C., Converti, A. and Borghi, D. (1998) “Cadmium, Zinc, Silver and Chromium (III) Removal From Wastewaters by Sphaerotilus natans”. Bioprocess Engineering. 19:197-203. 20. He, L.M. and Tebo, B.M. (1998) “Surface Charge Properties of and Cu(II) Adsorption by Spores of the Marine Bacillus sp. Strain SG-1” Applied and Environmental Microbiology. American Society for Microbiology. 64. 3: 1123-1129. 21. Greenberg, A.E., Trussell, R.R. and Clesceri, L.S. (eds.) (1985) “Standard Methods for the Examination of Water and Wastewater” American Public Health Association.16th ed. 22. McNicol, L.A., Aziz, K.M.S., Huq, I., Kaper, J.B., Lockman, H.A, Remmers, E.F., Spira, W.M., Voll, M.J. and Colwell, R.R. (1980) “Isolation of Drug-Resistant Aeromonas Hydrophila From Aquatic Environments” Antimicrobial Agents and Chemotherapy. 17. 3: 477-483. 23. Handfield, M., Simard, P. and Letarte, R. (1996) “Differential Media for Quantitative Recovery of Waterborne Aeromonas hydrophila” Applied and Environmental Microbiology. 62. 9:3544-3547.

866

Biosorption

24. Zakaria, Z.A. (2002) “A Study On the Removal of Gold by Thiobacillus ferrooxidans”. Universiti Teknologi Malaysia. MSc. Thesis. 25. Zakaria, Z.A. and Ahmad, W.A. (2001) “Gold Binding Study and Some Surface Characteristics of Thiobacillus ferrooxidans” in 6th Asean, Science and Technology Week, Brunei. In press. 26. Huang C.P., Huang C.P. and Morehart A.L. (1990) "The Removal of Cu(II) From Aqueous Solution by Saccharomyces cerevisiae" Water Research, 24, 4: 433-439. 27. Madigan, M.T., Martinko, J.M. and Parker, J. (2001) “Brock: Biology of Microorganisms” Prentice-Hall Int’l. USA. 9th ed., 52-62. 28. Vecchio, A., Finoli, C., Di Simine, D. and Andreoni, V. (1998) “Heavy Metal Biosorption by Bacterial Cell”, Fresenius Journal of Analytical Chemistry. SpringerVerlag. 361:338-342. 29. Kilbane II, J.J., Xie, J.Z. and Chang, H.L. (1996) “Removal and Recovery of Metal Ions from Wastewater Using Biosorbents and Chemically Modified Biosorbents” Bioresource Technology. 57. 127-136.

867

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Copper ion adsorbed on chitosan beads: Physico-chemical characterization S. Chatterjee1, S. Som Majumdar2, B.P. Chatterjee1 and A.K. Guha1* 1

Department of Biological Chemistry, Indian Association for the Cultivation of Science, Kolkata-700 032, India 2 Institute of Wetland Management and Ecological Design, B04, LA Block, Saltlake Kolkata-700 091, India

Abstract Beads prepared from chitosan having 89% of degree of deacetylation, polydispersity index around 1.2 and molecular weight 1×106 were used to study physico-chemical characterization of copper ion adsorption from aqueous solution. Kinetics of metal adsorption on chitosan beads had been found to follow first order rate law. More than 90% of the copper ion was adsorbed on the beads within 50 min. Temperature and pH of the solution influenced the rate of adsorption process. Desorption of copper ion from chitosan beads by changing pH to ~1 with sulfuric acid showed that this process also follows first order. 1.

INTRODUCTON Disposal of toxic heavy metals into the environment is increasing due to rapid industrialization. Copper ion is coming into wastewater due to its extensive use in electrical industries, antifouling paints and as fungicides. Present method for the removal of Cu (II) is to precipitate it as hydroxide by lime treatment. But this conventional procedure becomes less effective when metal ion concentration is in the range of 100 ppm [1]. This leads to research for the development of new technology for removal of toxic metals from wastewater [2] One of these technologies is the use of biopolymer for the removal of heavy metals from wastewater. Chitosan, polymer of β-1,4 glucosamine, due its wide availability, eco-friendly nature and capability of lowering transition metal ion concentration to ppb level through chelation [3] finds an alternative procedure for the treatment of wastewater. Metal binding capacity of chitosan depends on its physicochemical characteristics. Most of the researches [4-7] have been carried out with chitosan powder. In the present investigation we describe the adsorption and desorption kinetics of Cu (II) ions on chitosan in bead form. Influence of emperature and pH on adsorption process has also been reported.

*

Corresponding author: E mail: [email protected]; Fax: 91 33 2473 2805

869

Biosorption

2.

EXPERIMENTAL

2.1. Preparation of chitosan Chitin was isolated from prawn shells by the method described by Hackmann [8] and deacetylated to chitosan by treatment with 50% NaOH (Chitin:NaOH; 1:20) for 7 h at 120°C. Crude chitosan was purified by dissolving in 7% AcOH and precipitated out by adjusting pH to 8.5 with 1 (N) KOH. Chitosan thus obtained was washed first with water followed by ether. 2.2 Characterization of chitosan The degree of deacetylation and weight average molecular weight and chitosan were determined by first derivative of UV spectroscopy [9] and intrinsic viscosity [10] respectively. Polydispersity was determined by light scattering method. FTIR spectra of chitosan were taken in Shimadzu spectrophometer. Specific rotation was measured using Perkin Elmer Polarimeter (model 241 MC) 2.3 Preparation of chitosan beads Chitosan beads were prepared by drop wise addition of 3% (w/v) solution of chitosan in 7% (v/v) AcOH to alkali coagulating mixture [H2O:MeOH:NaOH; 4:5:1 (w/w)] [11]. Beads were collected by filtration; washed with water to neutral pH and then conditioned by treatment with 0.1 (M) (NH4)2SO4 [12]. 2.4 Preparation and estimation of Cu (II) solution 1-liter stock solution (1000 mg.L-1) of Cu (II) was prepared from analytical grade CuSO4.5H2O and diluted to working metal ion concentration. Initial pH was adjusted with dilute H2SO4. Cu (II) ion concentration was determined by atomic absorption spectrophotometer (Perkin Elmer model no 2380). 2.5 Adsorption of Cu (II) ion on chitosan beads The sorption experiments were carried out in batch process. 20 ml of CuSO4, 5 H2O solution containing 26.5 mg.L-1 Cu(II) were taken in different 250 ml Erlenmeyer flasks. 1 gm of chitosan beads having water content 96.4% was added to each flask. Flasks were agitated gently at 30˚C and 10 ml solution was taken out from duplicate flasks at different time intervals. Solution was filtered and diluted to its atomic absorption analysis. The experiments were performed at two different pHs (3.0 and 5.0). Experimental tests were also performed under different operating conditions: temperature 10-50°C and initial Cu (II) ion concentration (75-305 mg.L-1). 2.6 Desorption of Cu (II) ion from beads Desorption of Cu (II) ion from chitosan beads were performed immediately after the adsorption experiments. Beads were collected by filtration; washed with water and suspended in 20 ml of 0.1 (M) NH4SO4 and H2SO4 at pH 1.0 and gently agitated for 24 h. Beads were separated by filtration and Cu (II) ion concentration in the filtrate was estimated. Beads were reconditioned as described and reused.

870

Biosorption

3.

RESULTS AND DISCUSSION Properties of the chitosan used for the preparation of beads have been presented in Table I, which shows that chitosan was highly deacetylated and polydispersed.

Table 1.Characterization of chitosan Degree of deacetylation (%) 89.71

Molecular weight 10 × 105

Polydispersity index 1.205

Average molecular size (nm) 1448.6

Ash content (%) 0.6

Moisture content (%) 5.12

Specific rotation [αD] -21°

Co FTIR result (Fig. 1) shows that the chitosan used in this study was similar to that purchased from Sigma USA. Rate of adsorption of Cu (II) ion on chitosan beads at two different pHs is presented in the Fig 2a. The adsorption process during initial 1 min is presented in Fig. 2b. It appears from the figures that pH of the solution influenced the adsorption process and higher adsorption was noted with pH 5.0 than 3.0. Better adsorption of Cu (II) ion noted at pH 5.0 may be due to protonation of –NH2 groups of chitosan at lower pH inhibits chelation of metal ion by chitosan. Adsorption of Cu (II) ion on chitosan beads was very fast during the first minute, rate of adsorption gradually retreated down and 90% adsorption of Cu (II) ion was completed within 50 min. Initial high rate adsorption may be due to large availability of free amino groups of chitosan as well as high copper ion concentration. Fig. 3 represents the influence of temperature on sorption of Cu (II) by chitosan beads. The rate of adsorption increased with increase in temperature and maximum adsorption was noted at 30˚C and after that no further increase was observed. Order of adsorption of Cu (II) by chitosan beads was calculated by plotting log (residual Cu (II) ion mg.L-1 in solution) against time and presented in fig. 4.

Figure 1. Co FT IR spectra of chitosan [a ≡ chitosan prepared, b ≡ Sigma chitosan] 871

Biosorption

Cu(II)ads mg/g of chitosan

16 14 12 10 8 6 4

p H 5 .0 p H 3 .0

2 0 50

100

150

Figure 2a. Cu (II) adsorption Cu (II) ion ads/ mg of chitosan

200

8 6 4 2 0 0.0

0.2

0.4

0.6

250

300

1400

0.8

1.0

16 14 12 10 8 6 4 2 0

10

20

Time (min) Figure 2b. Cu (II) adsorption in the first min

30

o

40

50

Temp ( C )

Figure 3. Effect of temperature on adsorption

Figure 4. Adsorption kinetics of Cu (II) ion on chitosan beads

872

1450

Time (min)

Cuabs(mg)/g of chitosan

0

Biosorption

Since 50% of adsorption took place within first min, kinetics was studied only for that period. Straight-line curve obtained in this respect indicates that this adsorption process followed first order kinetics. There are different equations including Langmuir and Freundlich which described relationship between the amount adsorbed (X/m) and equlibrium concentration (Ce). Plot of log X/m vs. log Ce (Fig.5) gave a straight line indicating that it followed Freundlich’s adsorption isotherm. The desorption of Cu (II) from chitosan beads with time is presented in Fig. 6, which shows that this process was also very fast and almost 80% of desorption took place within first 60 min. This may be due to ready elution of copper ion from the outer surface of the beads. Removal of copper ion which goes inside of the porous beads due to capillary action takes some times to elute. The plot of Log [Cu (II) mg.L-1 in solution] with time shows linearity (Fig. 7), indicates desorption follows first order kinetics. -2 .0 -2 .2 -2 .4

log (x/m)

-2 .6 -2 .8 -3 .0 -3 .2 -3 .4 -3 .6 -3 .8 -4 .0 0 .0

-0 .5

-1 .0

-1 .5

-2 .0

-2 .5

lo g C

Figure 5. Adsorption isotherm

-3 .0

-3 .5

-4 .0

-4 .5

-5 .0

e

Cu(II)(des)( mg )/ gm

of chitosan

16 14 12

12

10

10

8

8 6

6

4 4

2 2

0

F ig 6

0 0

50

100

Time

150

0

200

10 250

20

30 300

40

50 1400

60 1450

(min)

Figure 6. Cu (II) desorption

873

Biosorption

Figure 7. Kinetics of Cu (II) desorption

Amount of Cu (II) adsorbed and desorbed per gm of chitosan are presented in the Table 2. Table 2. Adsorption and desorption of metal ion. Metal ion present initially in 20 ml of solution (mg)

Metal ion present finally in 20 ml of solution (mg)

(%) of metal ion removed

Chitosan content per gm of beads (g)

Amount of metal ion adsorbed by per gm of chitosan (mg)

Amount of metal ion desorbed per gram of chitosan (mg)

(%) of metal ion desorbed

0.53

0.016

96.98

0.0358

14.0000

13.7003

97.85

It is clear from the table that in this experiment around 97% of Cu (II) ion in the solution was removed by chitosan treatment and approximately 98% of could be desorbed from the beads. It was observed that beads could be reused five more times without affecting its efficacy. 4.

1. 2. 3. 4. 5. 6.

874

CONCLUSIONS It may be concluded from the present study Chitosan beads can remove more than 96% of Cu (II) ion from solution by sorption.

Optimum pH and temperature in this respect being 5.0 and 30°C respectively. Sorption process is very fast during 0 to 10 min. Cu (II) ion can be disorbed from the beads by changing pH and thus facilitates repetated use. Both sorption and desorption process follows first order rate law. Adsorption of Cu (II) on chitosan follows Freundlichs isotherm model.

Biosorption

REFERENCES

1.

A.Hammaini, A.Ballester, F. Gonzalez, M.L. Blazquez and J.A.Munaz, Biohydrometallargy and the environment toward the mining of the 21st centure R.Amils and A.Ballester (Eds) (1999) 185 2. J.R. Deans and B.G. Dixon, Water Res, 26 (1992) 469 3. W. Kaminski and Z. Modrzejewska, sep. Sci. Technol, 32 (1997) 2659 4. R.A.A. Muzzarelli and O. Tubertini, Talanta, 16 (1969) 1571 5. C.A. Eidn, C.A. Jewell and J.P. Wightman, J. Appl.Polymer Sci., 25 (1980) 1587 6. R. Maruca, B.J. Sdudar. And J.P. Wightman, J. Appl.Polymer Sci., 27 (1982) 4827 7. R.A.A. Muzzarelli and R. Rocchetti, Talanta, 21 (1969) 1571 8. R.H. Hcakman, Austral. J. Biol. Sci. 7 (1954) 168 9. Muzzarelli, R. A. A., Rocchetti, R., Stanic, V. and Weckx, M., “Methods for the determination of the degree of deacetylation of chitin and chitosan”, in R.A.A. Muzzarelli and M. G. Peter (eds.), Chitin Handbook, European Chitin Society, (1997) 109. 10. H. Chen and M.L. Tsaih, Int. J. Biol. Macromol, 23 (1998) 135 11. T. Mitani, N. Fukumuro, C.Yoshimo and H. Ishii, Agric. Biol. Chem., 55 (1991) 2419 12. R.A.A. Muzzarelli and R. Rocchetti, Talanta, 21 (1974) 1137

875

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Development of a process for biosorptive removal of mercury from aqueous solutions S.G. Tupe and K.M. Paknikar∗ Microbial Sciences Division, Agharkar Research Institute, G.G. Agarkar Road, Pune 411 004, India Abstract Mercury biosorption studies with fungal cultures revealed that optimum pH for biosorption was 8 and culture PK2 possessed highest efficiency. Metal loading capacity for PK2 was 76 mg/g. Effect of varying biomass concentration, time, and mercury concentration on biosorption by PK2 was checked. It was found that biosorbent concentration of 0.2 g, contact time of 60 min. and mercury concentration of 40 mg/L were optimal parameters. Pre-treatment with dimethyl sulfoxide (100%), hydrochloric acid (1M), sodium carbonate (1M) resulted in an increase in the biosorption efficiency, while ethanol (absolute), triton X-100 (1%), and ammonium sulphate treatment decreased the efficiency. It was observed that mercury uptake values could be fitted in the Freundlich and Langmuir isotherm models. The biosorbent beads of PK2 biomass were prepared by a proprietary process. The beads (1 g, pre-treated with dimethyl sulfoxide) were packed in glass columns (length 5.5 cm, internal diameter 1 cm) and a solution containing mercury (100 mg/L) was passed in an up-flow mode. It was observed that adsorption efficiency dropped below 60% and 50% after passing 10 and 35 bed volumes, respectively. However, a removal efficiency of 40% could be sustained beyond 100 bed volumes of mercury solution. 1.

INTRODUCTION Mercury is toxic in its metallic (including gaseous), ionic and organic (monomethyl, dimethyl and phenyl) forms, and has long been recognised as an environmental hazard. Its use in industry is widespread, especially in the production of chlorine, caustic soda, thermometers, certain pharmaceutical drugs, pesticides and agricultural products, electrical equipments such as batteries, metal switches, fluorescence lamps, etc. Although there has been a belated shift from the polluting mercury cell technology to the membrane cells, mercury pollution in India is still very high. India does not produce mercury and relies completely on imports. Between 1998-2001, the annual mercury imports for mercury cell plants stood at 170-190 tonnes, which is 10% of the global mercury consumption. Assuming that these mercury cell plants and other industries release only

∗ Corresponding author (E-mail: [email protected]) SGT thanks Council for Scientific and Industrial Research (CSIR, India) for research fellowship. 877

Biosorption

50% of the mercury, the total mercury pollution load generated per year will be about 125 tonnes of elemental mercury [1]. This is about five times more than the total mercury compounds dumped into the Minamata Bay in Japan in 36 years, equivalent to five Minamata disasters. Current technologies for mercury removal from wastewater are costly and ineffective to achieve desired permissible limits for discharge. Hence, there is a need for development of cost effective green technology. Biosorption provides a good alternative for the removal of metals from solution. Algae have been shown to accumulate [2-3] and volatilise mercury [4-5]. A process based on bioaccumulation of Hg [6] by genetically modified, mercury-resistant Pseudomonas putida, Aeromonas hydrophila and natural consortia has been developed in a bench-scale column. This paper describes studies on the uptake of mercury by fungal biomass. The study carried out will be helpful in developing full-scale biosorption process to clean up mercury pollution of water and wastewater. 2.

MATERIALS AND METHODS Biosorption studies were carried out with fungal cultures isolated in our laboratory through extensive screening for metal biosorbents. Fungal isolates Aspergillus niger (501), Alternaria alternata, PK1, PK2, Aspergillus niger (502), Aspergillus fumigatus, Fusarium oxysporum (NCIM-718), Absidia blakesleena (NCIM-889), Actinomucor sp. (NCIM1183) and Absidia corymbifera (NCIM-1233) were grown in bulk quantity in Sabourauds medium. After 5 days of incubation at room temperature (28±5°C) biomass was harvested by filtration. After washing the biomass with deionised water, it was dried in an oven at 60°C for 24 h. Dry biomass was powdered in a blender (particle size 0.1-0.2 mm) and then used in biosorption experiments. Standard protocol [7] was used for screening of cultures. The dried and finely powdered biomass obtained by above procedure was conditioned to pH 4, 5, 6, 7, and 8 by contacting with distilled water of respective pH. Metal solution was prepared by dissolving appropriate quantity of HgCl2 in water so as to get final concentration of 1 mM. The pH of the Hg solution was adjusted to 4, 5, 6, 7, and 8 separately. The dried biomass (0.2 g) conditioned to different pH was then contacted with 100 mL metal solution of same pH in a 250 mL Erlenmeyer flasks. The flasks were kept on shaker (120 rpm) for 30 min. A culture showing maximum mercury biosorption was selected for further studies. The mercury loading capacity or cumulative biosorption by the selected culture was determined by contacting the dried, pH conditioned (pH 8) biomass powder (1 g) several times with fresh batches of 100 mL mercury solution (200 mg/L, pH 8) till the biomass was saturated with mercury ions. To check the effect of biomass concentration, mercury solution (100 mL, 200 mg/L of Hg, pH 8) was contacted with varying amounts of dried, pH-conditioned biomass (pH 8) of the culture in the range of 0.1% to 1% (w/v). Rate of mercury uptake by the culture was studied by contacting the dried biomass (pH 8, 0.2 g) with 100 mL mercury solution (200 mL, pH 8), for various time intervals (10 - 80 min.). To check the effect of mercury concentration on biosorption, the culture biomass (0.2 g, and pH 8) was contacted with different concentrations (20, 40, 60, 80, 120, 160, 200 mg/L) of mercury solution (100 mL, pH 8) for 1 hour.

878

Biosorption

In order to check the effect of different pretreatments, freshly harvested biomass (10 g wet weight) was treated with 50 mL of the following solutions for 30 min.: - sodium carbonate (1 M), sodium hydroxide (1 M), hydrochloric acid (1 M), urea (1 M), ammonium sulphate (1 M), triton-X 100 (1%), dimethyl sulfoxide (100%), and ethanol (absolute). Biomass was separated by filtration and washed several times with deionised water to remove traces of adhering chemicals. It was then dried at 60°C, powdered and used in mercury biosorption experiments. Biosorbent beads were prepared by mixing the biomass with a polymeric matrix from waste poultry feathers by a proprietary process [15]. A solution containing 100 mg.L-1 mercury (pH 8) was passed in upflow mode through a glass column (length 5.5 cm, internal diameter 1 cm) containing 1 g biosorbent beads (pre-treated with dimethyl sulfoxide). The void volume of the column after packing the biosorbent was 2 mL. Solution flow rate was adjusted to 0.2 mL.min-1 using a programmable peristaltic pump (Ismatec, Switzerland, model MCP 552). In all the above experiments, after contacting with mercury solutions, the residual mercury content in solutions/filtrates were analysed for using an atomic absorption spectrometer (ATI-UNICAM, UK, model 929). The experiments were carried out in duplicates and appropriate experimental controls were run simultaneously. 3.

RESULTS AND DISCUSSION

3.1 Biosorption efficiency of the fungal cultures The data given in Table 1 show mercury sorbed (%) by the different fungal biomass at pH 8. Sorption of Hg2+ by the different fungal biomass was observed to vary according to the pH and found to occur over a pH range of 6-8. Mercury precipitation was not observed at pH 8. The optimum pH for sorption by all the cultures was 8. Biomass of culture PK2 was the most efficient in removing mercury from solutions; hence it was used at pH 8 in further experiments. The pH of metal solution influences metal biosorption by changing surface properties of biomass and metal speciation [8]. Most investigators have shown that a pH range of 4.0-8.0 is optimal for metal uptake [9]. At acidic pH metal uptake was less, which increased, as the pH was increased upto 8. As the pH level was increased, more negatively charged ligands would be available for binding positively charged metal ions. Table 1. Mercury sorbed (%) by the fungal biomass Culture A. niger (Culture no. 501) Alternaria alternata (NCIM-718) Culture PK1 Culture PK2 A. niger (Culture no. 502) Aspergillus fumigatus Fusarium oxysporum Absidia blakesleana (NCIM-889) Actinomucor sp. (NCIM-1183) Absidia corymbifera (NCIM-1233)

Mercury sorbed at pH 8 (%) 50 33 40 53 42 03 39 05 26 02

3.2 Metal loading capacity of PK2 biomass It was found that mercury-loading capacity of PK2 was 76 mg/g. The economics of metal biosorption process are more dependent upon metal loading capacity of a biosorbent 879

Biosorption

rather than its percent biosorption efficiency. Metal loading capacity gives more realistic figure as compared to percent metal uptake from solutions or the specific uptake values based on calculations [10]. 3.3 Effect of biomass concentration It was observed that amount of PK2 biosorbent required for maximum mercury biosorption was 0.2% (Figure 1). Further increase in the biosorbent concentration had not increased the biosorption efficiency. So 0.2% was considered as optimum biomass concentration for maximum Hg2+ uptake. 3.4 Rate of mercury uptake Mercury uptake increased gradually with time upto 60 min. of contact, following which there was no change in the sorption capacity (Figure 2).

Figure 1. Effect of biomass concentration on Hg sorption

Figure 2. Effect of time on Hg sorption

3.5 Effect of mercury concentration Initially there was increase in the efficiency of sorption as the concentration of mercury increased, but after 40 mg/L considerable change was not observed in biosorption efficiency, which remained almost stationary up to 200 mg/L. 3.6 Biomass pre-treatment Chemical pre-treatment to biomass result in unmasking or exposing the metal binding groups, adding new metal binding groups or modifying the existing ones and alterations in charge density on the surface. Such pre-treatments might be useful in improving the metal biosorption. Sodium carbonate, hydrochloric acid, and dimethyl sulfoxide treatment resulted in significant increase in biosorption efficiency of PK2. Maximum increase was due to dimethyl sulfoxide treatment (Table 2). The increase may be a result of change in charge density on the surface. Ethanol, ammonium sulphate, triton X-100 treatments resulted in reduced biosorption, probably as a result of denaturation, changes in surface charges or leaching of low molecular weight compounds [11-12]. Table 2. Effect of chemical pre-treatments to biomass Biomass pre-treatment None Sodium carbonate Ethanol Dimethyl sulfoxide Triton X-100 880

Mercury sorption (%) 44 62 24 78 18

Biosorption Biomass pre-treatment Sodium hydroxide Hydrochloric acid Ammonium sulfate Urea

Mercury sorption (%) 52 65 25 49

3.7 Adsorption isotherm Adsorption isotherm is an important tool that indicates relative affinities of biosorbents for a particular metal. It also helps in understanding the type of interaction that takes place between metal ions and microbial surfaces such as physical adsorption, nucleation or multilayer adsorption, etc. The metal uptake value (Q) was calculated using the equation: V (C i − C f ) Q= (1) 1000m where Q is the metal uptake (mg.g-1 biosorbent), V the volume of metal solution (mL), Ci the initial concentration of metal in solution (mg.L-1), Cf is final concentration of metal in solution (mg.L-1) and m is the mass of biosorbent (g). Figure 3 shows the isotherm for mercury biosorption. It could be seen that specific metal uptake (Q) increased with the increase in initial mercury concentration up to 160 mg.L-1. Above this concentration, the metal uptake remained constant indicating saturation of the binding sites. From the Q value obtained, adsorption isotherms were plotted according to Freundlich and Langmuir equations: ⎛1⎞ ln Q = ln k + ⎜ ⎟ ln C eq (Freundlich equation) (2) ⎝n⎠ C eq 1 C eq = Qmax + (Langmuir equation) (3) Q b Qmax where, Ceq is the liquid phase concentration of the metal (mg.L-1), b the Langmuir constant, Q the metal uptake (mg.g-1 biosorbent) and Qmax the maximum metal uptake (mg.g-1 biosorbent). The mercury uptake values could be fitted to the widely accepted Freundlich (Figure 4) and Langmuir model (Figure 5). Adsorption isotherm studies suggested that mercury biosorption by PK2 was a physical interaction characterised by monolayer adsorption onto heterogeneous surfaces at constant adsorption energy according to the basic assumptions of the Langmuir and Freundlich models [13-14]. 3.8 Breakthrough curve Based on the data breakthrough curve was plotted (Figure 6). It was observed that 10 and 35 bed volumes could be passed through the column before the mercury removal efficiency dropped below 60% and 50% respectively. The efficiency of the column remained above 40% even after passing more than 100 bed volumes of mercury solution. 4.

CONCLUSION The biomass of culture PK2 is a good biosorbent material for mercury removal. Although the residual mercury concentration in the column experiments performed using PK2 biosorbent beads was above the permissible limit, it is possible to improve the efficiency of the system further by employing a battery of columns.

881

50

2

40

1,5 Log Q

Q (mg.g-1)

Biosorption

30 20

R2 = 0,998

1 0,5

10

0

0 0

50

1

100 150 200 250 300

1,3

C/Co

-1

Ceq.Q

R2 = 0.9401

4 2 0 0

20

40

60 80 100 120 -1

Ceq (mg.L )

Figure 5. Langmuir isotherm for Hg biosorption

2,2

2,5

Figure 4. Freundlich isotherm for Hg sorption

8 6

1,9

Log Ci

Ci (mg.L-1)

Figure 3. Adsorption isotherm for mercury sorption

1,6

0,7 0,6 0,5 0,4 0,3 0,2 0,1 0 1

20

40

60

80

100

No. of bed volumes

Figure 6. Breakthrough curve for Hg sorption

REFERENCES 1. S. Narain, Down To Earth, 11 (2002) 27. 2. Y.J. Shieh and J. Barber, Planta, 109 (1973) 49. 3. D. Khummongol, G.S. Canterford and C. Fryer, Biotech. Bioeng,, 24 (1982) 2643. 4. D. Ben-Bassat and A.M. Mayer, Physiol. Plant., 40 (1977) 157. 5. S.C. Wilkinson, K.H. Goulding and P.K. Robinson, Biotech. Lett., 11 (1989) 861. 6. M. Brunke, D. Deckwer, A. Frischmuth, J.M. Horn, H. Lunsdorf, M. Rohricht, K.N. Timmis and P. Weppen, International Symposium on Biohydrometallurgy’91 (1991) 43. 7. K.M. Paknikar, P.R. Puranik and A.V. Pethkar, Biohydrometallurgy and the environment toward the mining of the 21st century, R. Amils and A. Ballester (eds), Process metallurgy 9B (1999) 363. 8. Z. Aksu and Kutsal, J. Chem.Technol. Biotechnol., 52 (1991) 109. 9. K.J. Blackwell, Singleton and J.M. Tobin, Appl. Microbiol. Biotechnol., 43 (1995) 579. 10. A.V. Pethkar, Uptake of heavy metals by filamentous fungi with special reference to biosorption and its newer applications, Ph.D. thesis, University of Pune, Pune, India, (1999). 11. P.R. Puranik and K.M. Paknikar, J. Biotechnol., 55 (1997) 113. 12. P.R. Puranik, N.S. Chabukswar and K.M. Paknikar, Appl. Microbiol. Biotechnol., 43 (1995) 1118. 13. H. Freundlich, Colloid and Capillary Chemistry, Methuen, London, (1926). 14. Langmuir, J. Am. Chem. Soc., 40 (1918) 1361. 15. K.M. Paknikar, J.V. Vernekar and A.V. Pethkar, A process for the preparation of a matrix from poultry waste and its use in the immobilisation of biomass, Indian Patent No. 186621 882

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Effects of ionic strength, background electrolytes and heavy metals on the biosorption of hexavalent chromium by Ecklonia biomass Donghee Parka, Jong Moon Parka∗ and Yeoung-Sang Yunb a

Department of Chemical Engineering, Pohang University of Science and Technology (POSTECH), San 31, Hyoja-dong, Nam-gu, Pohang, 790-784, Korea b Division of Environmental and Chemical Engineering, Chonbuk National University 664-14, 1 Ga, Duckjin-dong, Duckjin-ku, Jeonju, Chonbuk, 561-756, Korea

Abstract Brown seaweed, Ecklonia, biomass was used to remove Cr(VI) from wastewater. Previously, we found that the Cr(VI) was not really removed by adsorption to the biomass but reduced to Cr(III) on the surface of biomass. In this study, the effects of ionic strength and background electrolytes of the solution on the Cr(VI) reduction rates were evaluated. Although the increase of ionic strength slightly inhibited the Cr(VI) reduction, the rate equation proposed in our previous study was applicable over a wide range of ionic strengths with different background electrolytes. The presence of other heavy metals such as Cr(III), Ni(II), and Zn(II) did not nearly affect the Cr(VI) reduction, while it caused the slight reduction of the uptake of total Cr at equilibrium. In addition, Cr(VI)-Ni(II) binary system was analyzed using the competitive Langmuir model. 1.

INTRODUCTION Chromium in aquatic environments is known to be very toxic and classified by the US EPA in their list of human carcinogens (Group A) [1]. The major source of chromium is the wastewater from electroplating and metal-finishing industries [2]. Of its several oxidation states (e.g., di-, tri-, penta-, and hexa-), trivalent and hexavalent chromium are principal forms found in industrial effluents [3]. It is interesting that these two forms of chromium exhibit very different toxicity and mobility. Cr(III) is relatively insoluble over pH 5 in aqueous systems and exhibits little or no toxicity. In contrast, Cr(VI) usually exists as highly soluble and highly toxic chromate anions (HCrO4- or CrO42-), and is a suspected carcinogen and mutagen. The conventional treatment method of chromiumcontaining wastewater is based on precipitation of the hydroxide form of Cr(III). Initially chromium is present as Cr(VI) which is then converted to Cr(III) by reaction with reducing chemicals, followed by precipitation. However, chromium precipitation produces a large amount of chemical sludge which is a major disadvantage of this method in addition to costs associated with its chemical reduction. It is generally considered that ion

∗ Corresponding author E-mail: [email protected]

883

Biosorption

exchange method can minimize the sludge generation. However, due to the high costs of synthetic resins, its application for the wastewater treatment has been limited [4]. Recently, biomaterials such as microbial biomass (fermentation wastes) and seaweeds have been examined as an adsorbent alternative to synthetic ion exchange resin considering the economics [4-9]. Several types of seaweeds were found to successfully remove chromium species [4, 9]. However, the mechanisms involved in Cr removal and factors affecting the Cr removal efficiency have not been studied in detail. In our previous studies, we characterized the removal mechanism of Cr(VI) by Ecklonia biomass by using various experimental systems and techniques. It was found that Cr(VI) was removed through redox reaction by biomass. The rate equation was also established, which was applicable over a wide range of pH, [Cr(VI)]o, [B]o, and temperature. However, in general, industrial wastewater does not contain only chromium but also many other ions that may affect the removal efficiency of chromium by biomass. Therefore, it is necessary to examine the effects of these ions on the Cr(VI) reduction for the practical application to the actual wastewater. In this study, we examined the Cr(VI) reduction rates according to ionic strength, background electrolytes, and other heavy metals such as Cr(III), Ni(II) and Zn(II). In addition, Cr(VI)-Ni(II) binary system was analyzed using the competitive Langmuir model. 2.

MATERIALS AND METHODS

2.1 Biosorbent The brown macro-alga, Ecklonia was collected from the seashore of Pohang, South Korea and sun-dried. After cutting into approximately 5-mm size particles, the Ecklonia biomass was contacted with 1 M H2SO4 for 24 h, by which ions originally in the biomass were expected to be replaced with proton. The biomass was then washed several times with double-deionized water and dried at 80°C for 24 h. The resulting dried biomass was used in all experiments. 2.2 Biosorption experiments The sorption of chromium was investigated using batch kinetic and equilibrium experiments. All of experiments were carried out with 5 g.l-1 of biomass concentration. Flasks were mixed at 200 rpm on a shaker at room temperature. During the experiment, the solution pH was adjusted to the desired value by the incremental addition of concentrated H2SO4 or NaOH. Control experiment was conducted at 100 mg l-1 of initial Cr(VI) concentration, and pH 2. For kinetic experiments, samples were intermittently removed from the flasks in order to analyze the Cr(VI) concentration. For equilibrium experiments related to Cr(VI), all trials were conducted until Cr(VI) was completely removed from the solution. Depending on both the solution pH and initial concentration of Cr(VI), the contact time required for the complete removal of Cr(VI) from the solution was ranged from hours to weeks. While, equilibrium experiments with only Ni(II) were conducted for 12 hours which was sufficient to reach the equilibrium state. 2.3 Analytical methods The concentration of Cr(VI) in the liquid samples was determined colorimetrically by reaction with 1,5-diphenylcarbazide in the acid solution. The absorbance of the resulting red-violet sample was measured at 540 nm using a spectrophotometer (Spectronic 21, 884

Biosorption

Milton Roy Co., USA). To determine the total concentration of chromium, all chromium was converted into the hexavalent state by oxidation with potassium permanganate. Thereafter, the oxidized chromium was analyzed by the above-mentioned method for Cr(VI) analysis. Since chromium in solution is mostly in hexavalent or trivalent states, the concentration of Cr(III) can be obtained from the concentration difference between the total and hexavalent chromium. The analytical method of chromium is detailed elsewhere [10]. The concentration of Ni(II) was determined in atomic absorption spectrophotometer (Model SpectrAA.800, Varian) at the wavelength of 232 nm. 3.

RESULTS AND DISCUSSION

3.1 Effect of background electrolyte The dependence of reduction rates on the background electrolyte was examined by using three monovalent (Li+, Na+, K+) and two divalent (Mg2+, Ca2+) cations. The concentration of each cation was 0.1 M, and common anion was chloride. Fig. 1 shows the time courses of [Cr(VI)] change. There was almost no difference in the reduction rate of Cr(VI) according to the change of cations except Ca2+.

Cr(VI) concentration [mg l-1]

100 Control LiCl; 0.1 M NaCl; 0.1 M KCl; 0.1 M MgCl2; 0.1 M

80

60

CaCl2; 0.1 M

40

20

0 0

2

4

6

8

10

12

Time [h]

Figure 1. Effect of various cations on the reduction of Cr(VI) by Ecklonia biomass at pH 2

Effects of cations on the reduction of Cr(VI) was further evaluated by calculating the reduction rates (redox reaction between Cr(VI) and biomass). The high magnitude of correction factors reaffirmed that the removal of Cr(VI) by Ecklonia biomass followed the oxidation/reduction reaction. The reduction rate of control experiment was 0.0118mmol-1 l h-1. For monovalent cations, there was only a 0.0002 unit difference between the maximum and minimum value of reduction rates. This implies that the presence of monovalent cations did almost not affect the reduction of Cr(VI). However, divalent cations caused the decrease in the reduction rate of Cr(VI). Especially, Ca2+ significantly reduced the reduction rate of Cr(VI) (about 26%). It has been known that alkali-metal and alkaline-earth cations can bind to negatively charged groups of biomass primarily by electrostatic interaction. However, mono- or divalent light metals do not bind on the biomass at low pH, such as 2.0, and may not greatly affect the reduction of Cr(VI). It is also reasonably safe to presume that the conformational change of biomass did not occur at cationic concentration lower than those tested in our experiment. The decrease of the 885

Biosorption

reduction rate by Ca2+ can be explained as follows: Ca2+ forms insoluble complexes with OH- and SO42- on the surface of the biomass which might hinder the contact between Cr(VI) ions and the biomass. The effect of anions on the reduction rate of Cr(VI) was also examined by using various background electrolytes (Cl-, NO3-, CO32-, SO42-, HPO42-). The concentration of each anion was 0.1 M, and common cation was sodium ion. Figure 2 shows the time course of Cr(VI) change by the reduction.

Cr(VI) concentration [mg l-1]

100 Control NaCl; 0.1 M NaNO3; 0.1 M

80

Na2CO3; 0.1 M Na2SO4; 0.1 M

60

Na2HPO4; 0.1 M 40

20

0 0

2

4

6

8

10

12

Time [h]

Figure 2. Effect of various anions on the reduction of Cr(VI) by Ecklonia biomass at pH 2

The reduction rate constants varied up to 21% according to anions added. For the redox reaction between Cr(VI) and biomass, the intermediate of Cr(VI)-biomass has to be formed. However, the high concentration of anions inhibits the formation of Cr(VI)biomass intermediate, competitively. This inhibition is exaggerated by the increase of the electrostatic force of anions. The order of electrostatic force of anions added in this experiments is Cl- < NO3- < CO32-, SO42-, HPO42-, while the order of reduction rates measured in this experiment was Cl- > NO3- > CO32- > SO42-, HPO42-. 3.2 Effect of ionic strength Changes in ionic strength of the solution can cause conformational modification of the functional groups of biomass. It was suspected that these conformational changes would make the reactive functional groups either more or less accessible to Cr(VI), thereby altering the reduction rate. The Cr(VI) reduction by the biomass was studied in the ionic strength range 0.01-1M of NaCl. Time course of [Cr(VI)] change did not display a difference when the ionic strengths were below 0.1 M (Figure 3). However, significant differences were observed above 0.5 M of NaCl. The reduction rates at 0.5 M and 1.0 M of NaCl decreased to 29% and 41% of the control, respectively. These results suggest that conformational changes in the functional groups of biomass at ionic strength below 0.1M of NaCl did not greatly affect the reduction rate of Cr(VI). High ionic strength of the solution such as over 0.5 M definitely affected the reduction rate of Cr(VI) by Ecklonia biomass but this may be ignored in actual process since the ionic strength of usual wastewater would be lower than 0.1 M. In a separate experiment, it was found that many other factors such as pH and temperature gave much greater impacts (data not shown). 886

Biosorption

-1

Cr(VI) concentration [mg l ]

100 Control NaCl; 0.01 M NaCl; 0.05 M NaCl; 0.1 M NaCl; 0.5 M NaCl; 1.0 M

80

60

40

20

0 0

2

4

6

8

10

12

Time [h]

Figure 3. Effect of ionic strength on the reduction of Cr(VI) by Ecklonia biomass at pH 2 3.3 Effect of other heavy metals The reduction of Cr(VI) by the biomass ultimately results in the formation of Cr(III) which possibly affects the Cr(VI) reduction by binding functional groups on the surface of biomass. It is presumed that the binding of Cr(III) makes the biomass more resistant to oxidation by Cr(VI). In addition to Cr(III), other heavy metals coexisting in wastewater can also be bound on the surface of biomass, resulting in the decrease of reduction rate. The effects of other heavy metal on the reduction rate of Cr(VI) were estimated with a series of experiments added 500 mg l-1 of Cr(III), Ni(II), and Zn(II), respectively. As can be seen in Figure 4, there was almost no change in time course of [Cr(VI)] at each solution and, therefore, no significant difference in the reduction rate of Cr(VI). Similar to the ionic strength, other heavy metals possibly affect the reduction rate of Cr(VI) but the effects may be insignificant compared to other factors such as pH and temperature. It is likely that the functional groups responsible for the reduction of Cr(VI) are not the same groups responsible for cationic heavy metal binding. Therefore, the presence of other heavy metals may be ignored in the actual application of biosorption for Cr(VI) detoxification.

-1

Cr(VI) concentration [mg l ]

100 Control Cr(III); 500 mg l-1 Ni(II); 500 mg l-1 Zn(II); 500 mg l-1

80

60

40

20

0 0

2

4

6

8

10

12

Time [h]

Figure 4. Effect of other heavy metals on the reduction of Cr(VI) by Ecklonia biomass at pH 2 887

Biosorption

3.4 Cr(VI)-Ni(II) binary system At equilibrium state, since Cr(VI) completely reduced into Cr(III) by biomass, chromium exists only in the form of trivalent chromium in the solution. As mentioned above, other cationic heavy metals did not much affect the reduction rate of Cr(VI). However, at the equilibrium state, they can affect the binding efficiency of Cr(III). Ni(II) was chosen to determine the effect of co-existing heavy metal on the uptake of chromium. The maximum uptake of total Cr was about 8 times of the maximum uptake of Ni(II) in each single system. For each metal ion in single system, the individual adsorption data was fitted by mono-component Langmuir model, and high magnitude of correction factors guaranteed the validity of this model. The individual Langmuir constants, qmax and b, of total Cr and Ni(II) were determined as, 107.5 mg.g-1 and 0.0131 l.mg-1, 13.73 mg.g-1 and 0.0085 l.mg-1, respectively. Figures 5 and 6 show the adsorption isotherms of total Cr and Ni(II) in the binary system at pH 3, respectively.

-1

Utake of total Cr [mg g ]

80

60

40

[Ni(II)]0; 0 mg l-1 U [Ni(II)]0; 100 mg l-1 V [Ni(II)]0; 300 mg l-1

20

0 0

50

100

150

200 -1

Equilibrium concentration of total Cr [mg l ]

Figure 5. The isotherms of total Cr at various initial concentration of Ni(II). Lines were predicted by the comparative Langmuir model 15

[Cr(VI)]0; 0 mg l-1 U [Cr(VI)]0; 100 mg l-1 V [Cr(VI)]0; 300 mg l-1

-1

Utake of Ni(II) [mg g ]

12

9

6

3

0 0

100

200

300

400

500 -1

Equilibrium concentration of Ni(II) [mg l ]

Figure 6. The isotherms of Ni(II) at various initial concentration of Cr(VI). Lines were predicted by the comparative Langmuir model 888

Biosorption

To express the simultaneous biosorption phenomena of total Cr and Ni(II) in the binary system, the competitive Langmuir model was used with constants obtained from isotherms of the single metal system. In a two-metal system at an equilibrium, the uptake of one metal decreases as the other metal concentration increases. However, in the Cr(VI)Ni(II) system, the competitive Langmuir model showed poor performance in estimating the uptake of total Cr, while much better performance in estimating the uptake of Ni(II). The model gave us under-estimated values for the uptake of total Cr, which implies that the Ni(II) present in the solution did not properly compete with the total Cr on the biomass. On the contrary, the uptake of Ni(II) was completely inhibited by the competitive adsorption of Cr(III). The incomplete competition between Ni(II) and adsorbed total Cr suggested that the uptake of total Cr is slightly inhibited by Ni(II) because of the strong binding of chromium [especially Cr(III)] with the biomass, while the uptake of Ni(II) is completely inhibited by the competitive adsorption of reduced Cr(III). 4.

CONCLUSIONS In our previous report, we derived a rate equation applicable for the reduction of Cr(VI) by Ecklonia biomass over a wide range of pH, [Cr(VI)]o, [B]o, and temperature (data not shown). In this study, we found that this rate equation is applicable for over a range of ionic strength and different background electrolytes. Other heavy metals such as Cr(III), Ni(II), or Zn(II) did not nearly affect the reduction rate of Cr(VI), but caused the slight decrease of the uptake of total Cr on the biomass. Therefore, in actual wastewater treatment system, we may ignore the presence of other metals in the detoxification of Cr(IV). Although various parameters of actual wastewaters probably affect the reduction rate of Cr(VI) by Ecklonia biomass, these effects are relatively small compared to pH or temperature. In conclusion, the Ecklonia biomass is a good candidate for a biosorbent in the Cr(VI) detoxification process, and the scale up may be accomplished using the rate equation previously derived.

ACKNOWLEDGEMENTS This work was supported in part by Pohang Iron & Steel Co., Ltd. (POSCO). REFERENCES

1.

T.H. Shelton, Interpreting Drinking Water Quality Analysis: What Do the Numbers Mean? Rutgers Cooperative Extension, New Brunswick, 1989. 2. S.E. Mannahan, Environmental Chemistry, 5th edn, Lewis Publishers, Chelsea, 1991. 3. V.P. Evangelou, Environmental Soil and Water Chemistry, pp. 476-498, John Wiley & Sons, Inc., New York, 1998. 4. D. Kratochvil, P. Pimentel and B. Volesky, Environ. Sci. Technol., 32 (1998) 2693. 5. Z. Aksu, Y. Sag and T. Kutsal,. Environ. Sci. Technol., 11 (1990). 6. S. Alasheh and Z. Duvnjak, Biotechnol. Prog., 11 (1995) 638. 7. D.C. Sharma and C.F. Forster, Water Res., 27 (1993) 1201. 8. M. Pillichshammer, T. Pumpel, R. Poder, K. Eller, J. Klima and F. Schinner, Biometals 8 (1995) 117. 9. A.L. Rapoport and O.A. Muter, Process Biochem. 30 (1995) 145. 10. L.S. Clesceri, A.E. Greenberg and A.D. Eaton, Standard Methods for the Examination of Water and Wastewater, 20th edn, pp. 366-368, American Public Health Association, American Water Work Association and Water Environment Federation, Washington, D.C., 1998. 889

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Evaluation of silver recovery from photographic waste by Thiobacillus ferrooxidans and chitin P. Thiravetyana*, W. Nakbanpoteb and C. Songkroaha a

School of Bioresources and Technology, King Mongkut’s University of Technology Thonburi (Bangkhuntien), 83 Moo. 8 Thakham, Bangkhuntien, Bangkok, 10150, Thailand b Pilot Plant Development and Training Institute, King Mongkut’s University of Technology Thonburi (Bangkhuntien), 83 Moo. 8 Thakham, Bangkhuntien, Bangkok, 10150, Thailand Abstract Chitin was able to adsorb both silver ion and silver-thiosulphate complexes at their optimum pH of system 7.0-8.0 and 2.0, respectively. Silver ion was probably adsorbed by ion exchanging and/or co-ordination, whereas silver-thiosulphate complexes were adsorbed at only pH 2.0 by electrostatic interaction. Sodium acetate in the fixer did not affect to silver-thiosulphate adsorption due to its low ionic strength, but sodium thiosulphate let free thiosulphate to form strong silver-thiosulphate complexes, which could not be adsorbed by chitin. Adsorption isotherm showed that the maximum adsorption capacity (Qmax) and the affinity (b) of chitin to adsorb silver ion (Qmax = 4.67 mg Ag/g; b = 0.565) and silver-thiosulphate complexes (Qmax = 4.37 mg Ag/g; b = 0.518) were slightly different. These constants implied that silver ion and silver-thiosulphate complexes were adsorbed onto the same functional groups: acetylamino and amino groups. The biooxidation of silver-thiosulphate by Thiobacillus ferrooxidans was able to degrade thiosulphate. The biooxidation process also changed silver-thiosulphate complexes to silver ion as adsorption occurring at system pH of 7.0. The advantage of using T. ferrooxidans and chitin is the achievement of higher purity silver. Keywords: silver, photographic waste, biooxidation, chitin, adsorption 1.

INTRODUCTION Silver is a precious metal. It is also classified as a hazardous substance [1]. Film development causes photographic waste of spent fixer and rinse water containing 1,00010,000 mg Ag/L and 50-200 mg Ag/L, respectively, in the form of silver-thiosulphate complexes [2]. Therefore, it should be recovered completely for worth usefulness and environmental issues. Electrolysis, metallic replacement, and precipitation are used to recover the silver waste, but they are not suitable for recovery of silver at low concentrations (< 100 mg/L) [3]. Moreover, metallic replacement and precipitation produce impure sludge that requires refining through further treatment. The ion-exchange method can regain silver at low concentrations, but costs of the ion-exchanger and * Corresponding author: Fax: 662-4523455, E-mail:[email protected]

891

Biosorption

maintenance are high [4]. Therefore, the production of low-cost alternatives has been brought into focus. Chitin is a cheap biopolymer that is directly extracted in large quantities from crab and shrimp shells, and seafood wastes. The structure of chitin is similar to the structure of cellulose. It is a completely substituted polysaccharide carrying acetylamino and amino groups per glucose ring, which the portion of amino groups is less than 50%. Chitin is an alternative adsorbent due to its content of acetylamino and amino groups for chelating metal ions [5] and its stability in acid conditions [5,6]. Chitin has been studied for the adsorption of anionic complexes such as Cr(VI) [5-7], Mo(VI) [8], and of cationic ions such as Cd2+ [9,10] and Cu2+ [11]. The adsorption of silver in the forms of Ag+, Ag(NH3)2+, Ag(SCN)32- and Ag(S2O3)23- by fungi that contain chitin and chitosan in the cell wall has been studied [3,12], but the use of the biomass as an adsorbent has had problems because the content of these polymers is able to change during growth of mycelia, which can account for the variations in metal uptake capacity [3,12,13]. In addition, thiosulphate in the photographic waste causes impurity to silver and reduced silver adsorption. Therefore, thiosulphate should be degraded to sulphur and sulphite by Thiobacillus sp., according to Equation (1) [3,14,15]. Moreover, Thiobacillus sp. is tolerant of the toxicity of silver at high concentration [16]. Thiobacillus sp. 2-

S2O3 SO32-+S° (1) Therefore, the aim of this research is to study the adsorption of silver ion and silverthiosulphate complexes by chitin. The effect of fixer composition on silver adsorption was also studied. Adsorption isotherm of silver ion and silver-thiosulphate by chitin were then compared. Finally, biooxidation of silver-thiosulphate by T. ferrooxidans before silver adsorption by chitin was also examined. 2

MATERIALS AND METHODS

2.1 Chitin Seafresh Chitosan (Lab) Co., Ltd., Thailand provided chitin as powder. The particle size was 0.5-1.0 mm. The degree of deacetylation (% DD) was approximately 50%. In these experiments, chitin was used as received, without any further treatment. 2.2 Silver solutions and photographic waste Silver-thiosulphate solution was prepared by adding 10 ml of silver standard 1,000 mg Ag/L in 2% nitric acid (Scharlau Chemie) to a 500-ml volumetric flask and then adjusting the volume with 0.0020 M sodium thiosulphate (Na2S2O3) (Carlo Erba) to obtain 20 mg Ag/L of silver-thiosulphate complexes. The solution of silver ion (20 mg Ag/L) was also obtained by adjusting the volume with de-ionised water. The initial pHs of the solutions were adjusted to the desired value ± 0.05 pH units by nitric acid and sodium hydroxide. Fixer and rinse water was obtained from a KODAK photographic laboratory that uses Kodak 3000 as the fixer. The composition of Kodak 3000 is shown in Table 1.

892

Biosorption

Table 1. Composition of Kodak 3000 Composition Sodiumthiosulphate Sodiumacetate Sodiumbisulphite Boric acid Acetic acid

Amount in 1000 ml 40-45 g 40-45 g 1-5 g 1-5 ml 1-5 ml

2.3 Adsorption experiments Batch adsorption experiments were conducted in 15-ml dark bottles. Experiments were carried out by adding 0.2000 g of chitin to 10 ml of 20 mg Ag/L silver-thiosulphate solution, and then shaking at 150 rpm, 30°C for 2 hours without attempting to control the pH during the course of the experiment. Adsorption isotherms were studied by varying adsorbent dosage in the range of 0.25%-2.00% (w/v). A sample was collected by filtration with Whatman no. 4. The remaining silver and changing pH were detected by the Inductive Couple Plasma Spectroscopy (ICP) (JY-124, France) at 328 nm and a pH meter (Mettler Delta 340, USA), respectively. The thiosulphate concentration was examined by iodometric titration [3,17]. The percentage of silver adsorption was calculated according to Equation (2). The silver adsorption isotherm was analysed by the Langmuir models as in Equations (3) [18,19]. Silver Adsorption (%) =

Ci − C f Ci

×100

Langmuir model: Qe = where Ci Cf Ce Qe Qmax b

Qmax bCe Q C = max e 1 + bCe ( 1 + Ce ) b

is the initial concentration of silver in the solution (mg Ag/L), is the residual concentration of silver in the solution (mg Ag/L), is the residual concentration of silver in the solution at equilibrium (mg Ag/L), is the silver adsorption at equilibrium (mg Ag/g), is the maximum adsorption of silver (mg Ag/g), is the affinity coefficient (L/mg).

2.5 Thiosulphate biooxidation T. ferrooxidans DSM 583 was incubated in 9K medium [20] shaken at 150 rpm, 30°C for 15 hours. A 10% v/v of an inoculum was transferred to 250-ml Erlenmeyer flask containing 150 ml of silver-thiosulphate solution, 5 mg Ag/L within 0.0020 M thiosulphate, shaken at 150 rpm, 30°C for 5 days. The thiosulphate residue and pH chaining were determined by iodometric titration and the pH meter.

893

Biosorption

3.

RESULTS AND DISCUSSION

3.1 Effect of initial pH Under non-precipitation conditions pH < 10, the effect of initial pH to adsorb silver ion (positive charge) and silver-thiosulphate complexes (negative charge) by chitin was investigated to obtain the optimum pH and adsorption mechanism. Figure 1 (a) showed that 97% of silver ion was adsorbed since initial pH of 6-10, which the pHs of system became to approximate pH of 7.0-8.0 (Figure 1 (b)). The adsorption of silver ion was decreasing to 40% at the initial pH of 2.0. Whereas, silver-thiosulphate complexes was adsorbed to 98% and the adsorption was occurred at only pH 2.0. The system pH of both was changeless at the initial pH 2.0 as shown in Figure 1 (b). The change in solution pH from the initial value is still unclear, but the rise in pH has been attributed to the adsorption of protons from water by the amine group [4,21]. The drop in pH at a pH greater than 9 could be attributed to the reverse process, that is, the release of protons in the high pH conditions [4,21]. (a)

Silver adsorption (%)

100 80 60 40

Ag-ion silver adsorption (%)

20

Ag-thiosulphate silver adsorption (%)

0 0

2

4

6

8

10

Initial pH of solution

(b) pHsystem at equilibrium

10 8 6 4

pH system of silver ion Ag-ion adsorption pH system of Ag-thiosulphate il thi l h t d ti

2 0 0

2

4

6

8

10

Initial pH of solution

Figure 1. Effect of initial pH of solution for adsorption of silver ion and silverthiosulphate complexes by chitin 894

Biosorption

The adsorption of silver ion and silver-thiosulphate complexes was pH dependent, because the pka of chitin-chitosan is 6.0-6.5 [18,22]. The acetylamino and amino groups of chitin (Figure 2) were protonated at the low pH; it supported silver-thiosulphate adsorption by electrostatic interaction. The release of protons was occurred in the high pH conditions, it might support the adsorption of silver ion by ion exchanging [23] and/or coordination between lone pair electrons of nitrogen and oxygen and silver ion [9,24]. C H2O H H O

O

C H2O H H

O

H OH

H

H

NHCOCH 3

O

O H OH

H

H

N H2

(Acetylam ino group )

O

(Am ino group)

Figure 2. Acetylamino group and amino group of chitin structure Additionally, the results of thiosulphate residue (Figure 3) indicated that there are 10% of thiosulphate was decomposed at initial pH of 2.0, according to Equation 4. While other 40% of thiosulphate were disappeared due to silver was adsorbed in the form of silver-thiosulphate complexes. 2H+ + S2O32-

[HS2O3-]

HSO3- + S°

SO2 + S° + H2O

(4)

Thiosulphate residual (%)

100 80 60 40 Thiosulphate 0.002 M + Silver

20

Thiosulphate 0.002 M + Silver + Chitin

0 0

2

4

6

8

10

Initial pH of solution Figure 3. Thiosulphate residual at the initial pH of solution 3.2 Effect of sodium acetate and sodium thiosulphate Sodium acetate and sodium thiosulphate are the main composition of Kodak 3000 (Table 1). Therefore, the effect of silver adsorption by chitin was studied by adding an amount of sodium acetate or sodium thiosulphate into silver-thiosulphate solution. Figure 4 clearly shows that sodium acetate did not affect silver adsorption, because it cannot form any complexes with silver, and it has too little ionic strength to interfere the silver adsorption onto chitin surface. Whereas, sodium thiosulphate gave thiosulphate to form silver-thiosulphate complexes, which an increasing of thiosulphate caused more stable of 895

Biosorption

silver-thiosulphate complexes as shown in Table 2. The stable complexes were more difficult to be adsorbed by chitin as the adsorption of silver-thiosulphate from rinse water containing 0.089 M thiosulphate was only 76% as shown in Table 3. Table 2. Complex formation equilibria for Ag(I) and thiosulphate [2] pK (0.1 mol.L-1 ionic strength)

Equilibrium Ag+ + S2O32Ag+ + 2 S2O32Ag+ + 3 S2O322Ag+ + 4 S2O323Ag+ + 5 S2O326Ag+ + 8 S2O32-

⇔ ⇔ ⇔ ⇔ ⇔ ⇔

AgS2O3Ag(S2O3)23Ag(S2O3)35Ag2(S2O3)46Ag3(S2O3)57Ag6(S2O3)810-

6.93 12.72 14.78 28.23 42.58 85.23

Silver adsorption (%)

100 80 Sodium acetate

60

Sodium thiosulphate

40 20 0

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Concentraion (M)

Figure 4. Effect of sodium acetate and sodium thiosulphate to silver adsorption Table 3. Adsorption of silver-thiosulphate obtained from standard solution and rinse water Silver adsorption (%) Adsorbent Chitin powder Thiosulphate concentration

Standard solution (Ci = 20 mg Ag/L) 96.53 0.0020 M

Rinse water (Ci = 20.1 mg Ag/L) 75.85 0.0089 M

3.3 Adsorption isotherm The adsorption isotherm of silver ion and silver-thiosulphate complexes was studied at 30°C under their initial optimum pH of 6.0 and 2.0, respectively. The maximum adsorption capacity (Qmax) of chitin to adsorb silver ion and silver-thiosulphate complexes was 4.67 and 4.37 mg Ag/g, respectively. The b constants were 0.565 and 0.518, respectively (Figure 4). These indicated that chitin had the efficiency to adsorb silver ion a few higher than silver-thiosulphate complexes. This implied that silver ion and silverthiosulphate complexes were adsorbed onto the same functional groups of chitin: acetylamino and amino groups. Since the effect of pH (Figure 1) suggested that silver896

Biosorption

thiosulphate was adsorbed onto the protonated groups at low pH by electrostatic interaction, whereas silver ion was adsorbed by ion exchanging [23] and/or co-ordination [9,24].

5.0 Q e, Ag-ion = 4.67*C e /(1.77+C e )

Q e (mg Ag/g)

4.0

2

R = 0.989

3.0 Q e, Ag-thiosulphate = 4.37*C e /(1.93+C e ) 2

2.0

R = 0.999

Ag ion Ag-thiosulphate

1.0 0.0 0

2

4

6

8

10

12

C e (mg Ag/L) Figure 4. Adsorption isotherm of silver ion and silver-thiosulphate complexes by chitin 3.4 Biooxidation of thiosulphate Recovery of silver from photographic waste still had interference from high thiosulphate concentration in rinse water (Table 3). Therefore, biooxidation of silverthiosulphate by Thiobacillus ferrooxidans for degradation of thiosulphate was studied. The silver recovery from photographic waste by biooxidation with T. ferrooxidans and then adsorption by chitin was studied in this research. The result confirmed that T. ferrooxidans was able to degrade all of 0.0020 M thiosulphate in 5 days without silver precipitation. Silver-thiosulphate complexes were changed to silver ion as adsorption occurs better at pH 6 than pH 2.0. The advantage of biooxidation process is the reduction of the organic impurity from the solution. That will support silver purification processes. 4.

CONCLUSIONS Chitin was able to adsorb both silver ion and silver-thiosulphate complexes. The optimum pH of system to adsorb silver ion was 7.0-8.0, because it provided the silver ion adsorption by ion exchanging and/or co-ordination. Silver-thiosulphate complexes were adsorbed at only pH 2.0 because this system induced the chitin protonation that supported electrostatic interaction, and also reduced the stability of the silver complexes. Sodium acetate in the fixer did not affect silver-thiosulphate adsorption, but sodium thiosulphate could interfere the silver adsorption. Adsorption isotherm showed that the maximum adsorption capacity and the affinity of chitin to adsorb silver ion and silver-thiosulphate complexes were slightly different. This implied that silver ion and silver-thiosulphate complexes were adsorbed onto the same functional groups. The biooxidation of silverthiosulphate by T. ferrooxidans degraded thiosulphate and increased the purity of silver. In view of the usefulness, the cooperation between biooxidation and adsorption by chitin should be studied further for application to recovery silver from photographic waste.

897

Biosorption

ACKNOWLEDGEMENTS The authors would like to thank the Chin Sophonpanich Foundation and the Thailand Research Fund (TRF) for financial assistance. REFERENCES 1. A. Townshend (eds.), Encyclopedia of analytical science, Academic Press, London, 1995. 2. R.M. Smith and A.E. Martell, Inorganic complexes, Plenum, New York, 1976. 3. A.V. Pethkar and K.M. Paknikar, Process Biochem., 38 (2003) 855. 4. L. Lasko and M. P. Hurst, Environ. Sci. Technol., 33 (1999) 3622. 5. Y. Sag and Y. Aktay, Process Biochem., 36 (2000) 157. 6. Y. Sag and Y. Aktay, Process Biochem., 36 (2001) 1187. 7. S.Hoshi, K. Konuma, K. Sugawara, M. Uto, and K. Akatsuka, Talanta., 47 (1998) 659. 8. S.Hoshi, K. Konuma, K. Sugawara, M. Uto, and K. Akatsuka, Talanta., 44 (1997) 1473 9. B. Benguella, and H. Benaissa, Colloids and Surfaces A, 201 (2002) 143. 10. M. Bhanoori, and G.Venkateswerlu, Biochimica et Biophysica Acta (BBA) –General subjects., 1523 (2000) 21. 11. M. Tsezos, and S. MattarA, Talanta, 33 (1986) 225 12. A. Kapoor and T. Viraraghavan, Bioresource Technology, 53 (1995) 195. 13. P. Simmons and I.A. Singleton, Apply Microbiol Biotechnology, 45 (1996) 278. 14. Le Faou, B.S. Rajagopal, L. Daniels and G. Fauque, FEMS Microbiology Review, 75 (1990) 351. 15. R. Meulenberg, E.J. Scheer, J.T. Pronk, W. Hazeu, P. Bos and J.G. Kuenen, FEMS Microbiology Letters, 112 (1993) 167. 16. H. Sato, H. Nakazawa and Y. Kudo, Int. J. Miner. Process, 59 (2000) 17. 17. AOAC, Official methods of analysis of AOAC international, AOAC international, Virginia, 1995. 18. M. Ruthven, Principles of adsorption and adsorption process, John, Wiley & Sons, New York, 1984. 19. McKay, G. (eds.), Use of adsorbents for the removal of pollutants from wastewaters, CRC press, New York, 1995. 20. M. West, L. William and C. Hinde (eds.), Consolidation and growth, Metal & Minerals Annual Review, Mining Journal, London, 1996. 21. T. Tianwei, H. Xiaojing and D. Weixia, J. Chem. Tech. Biotech., 76 (2001) 191. 22. F.C. Wu, R.L Tseng and R.S. Juang, Ind. Eng. Chem. Res., 38 (1999) 270. 23. D. Kratochvil and B. Volesky, Tibtech, 16 (1998) 291. 24. F. Veglio and F. Beolchini, Hydrometllurgy, 44 (1997) 301.

898

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Influence of the treatment of fungal biomass on sorption properties for lead and mercury uptake M. Španělováa, L. Švecováa* and E. Guibalb** a

Institute of Chemical Technology, Department of Environmental Chemistry, Technická 5, 166 28 Prague 6, Czech Republic b Ecole des Mines d’Alès, Laboratoire Génie de l’Environnement Industriel, 6 avenue de Clavières, F-30319 Alès cedex, France

Abstract Aspergillus niger biomass, obtained as a by-product of fermentation processes, has been tested for lead and mercury sorption. Biomass was produced by two different ways of citric acid production: (a) surface growth, and (b) submerged growth. Their sorption properties have been compared to biomass subjected to an alkaline treatment. The influence of the pH has been studied to select the optimum pH for lead and mercury and pre-select the best sorbent samples. Optimum sorption performances for Pb were obtained with surface-grown biomass, which was submitted to alkaline treatment, while with Hg biomass produced by submerged culture was preferred. For lead the experimental pH was set to pH 3, while for mercury the experimental pH was fixed to pH 6. Experiments were continued on selected sorbents by determination of sorption isotherms and uptake kinetics. The influence of experimental parameters (particle size, sorbent dosage, metal concentration) has been checked on sorption kinetics to evaluate the best operating conditions. In the case of lead sorption properties the excellent sorption properties characterized by fast sorption kinetics and high sorption capacities (close to 600 mg Pb g1 ) may be explained by the release of a cell component or a by-product of culture procedure (residue of growing media or sub-product of the fermentation process) that induces lead precipitation at long contact time. Keywords: Aspergillus niger, waste biomass, sorption, lead, mercury 1.

INTRODUCTION The reinforcement of the regulations concerning industrial wastewaters has been the motive to an increasing research for developing new sorption processes. Indeed, conventional techniques such as precipitation and ion-exchange are sometimes noncompetitive or unable to reach the regulation levels. There is still a need for alternative processes that should be cost-effective, and easy to manage. Biosorption that consists in using biological materials or microorganisms for the sorption of target molecules (metals,

* L. Švecová thanks the French Embassy in Prague for the financial support during her stay in France. ** E. Guibal thanks the European Community for financial support under Growth Program (3SPM project, Contract G1RD-CT2000-00300) for attending the IBS’03 conference.

899

Biosorption

dyes, pesticides) has been widely studied since the early 80’s. A great diversity of biomass material has been investigated for the sorption of metal ions from bacteria, fungi, algae [1], to agriculture waste by-products or bio-industry wastes [2]. Using dead biomass makes easier the control of the process and using waste materials reduces the cost of the material. For these reasons, fungal biomass has frequently been considered as a suitable biosorbent [3-7] since it is produced in huge amounts in the fermentation processes (citric acid production) and this biomass is poorly valorizable. Fungal biomass may contain a wide range of functional groups including amine groups, carboxylic groups, and sulfur groups brought by cell constituents. For example fungal cell wall may be constituted by the assembling of chitin/chitosan layers (aminerich), proteins (amino-acids), glucans [5]. Fungal biomass, especially from Mucorales group, has been considered as an alternative source for chitosan [8-10]: an aminopolysaccharide, which is commercially produced from crustacean shells, very efficient for metal recovery [11-14]. However, chitin and chitosan are strongly bound to other cell wall constituents such as glucans and proteins. While the association of chitin to proteins may be positive for sorption purpose since the amino-acids have potential sorption activity, the glucans have very weak potential for metal sorption. The immobilization of amine functions by linkage with glucans can significantly reduce the sorption activity of the material. Chemical treatments should be considered for “purifying” the biomass from this "inert" material [8-10,15-17]. The composition of the membrane is also tributary of the harvesting time and the growing conditions [18]. A preliminary study has been launched on the comparison of the sorbing potential of Aspergillus niger prepared under surface and submerged fermentation conditions used (a) as produced, and (b) after alkaline treatment. The sorption properties are investigated for lead and mercury removal from dilute solutions. First the influence of the pH is studied and these results serve to select the best sorbents for the recovery of these metals. Then the sorption properties of selected biomass are studied through sorption isotherms and uptake kinetics. 2.

MATERIALS AND METHODS

2.1 Materials Aspergillus niger biomass was supplied as an industrial waste from fermentation processes (citric acid production) by a local company (Czech Republic). Two different samples have been tested: one being obtained by a surface-growth procedure while the second sample was obtained by submerged culture. After being collected the samples were abundantly rinsed in order to remove residual compounds present in the culture media. The samples were then dried or submitted to an alkaline treatment to prepare 5 different materials. The treated samples were finally submitted to a water-rinsing step and dried. The experimental conditions for their preparation are listed in Table 1. Typically, the treatment was performed with 100 kg (wet weight, 20-25% dry mass percentage) mixed with 500 L of sodium hydroxide solution at the appropriate concentration. After chemical treatment, the biomass was centrifuged, rinsed with water up to neutral pH and then dried in a fluidized bed dryer. The weight loss changed with the treatment. In the case of biomass B and D, which were selected for most of the experiments, the weight loss was 55% and 80%, respectively. The samples were ground and sieved before being used to prepare 4 different size fractions: G1 < 125 µm < G2 < 250 µm < G3 < 355 µm < G4 < 510 µm. 900

Biosorption

Lead chloride (Riedel-de-Haen, Germany) and mercury chloride (Fluka, Switzerland) have been used for the preparation of the solutions. Other reagents (hydrochloric acid, sodium hydroxide) were supplied by Carlo Erba (Italy). Table 1. Preparation characteristics of biosorbents Sample A B C D E

Culture conditions Surface Surface Submerged Submerged Submerged

Alkaline treatment (concentration) No Yes (1 M) No Yes (1 M) Yes (10 M)

Temperature

Time (h)

Room Room 107°C

24 24 6

2.2 Sorption procedures The study of the influence of pH on sorption performance was carried out by mixing sorbent with metal ion solution at selected pH (controlled with HCl and NaOH) for 3 days. The final pH was measured and the residual concentration in the solution (after filtration on membrane; pore size: 1.2 µm) was determined by ICP-AES (inductively coupled plasma atomic emission spectrometry). Sorbent dosage was fixed at 300 mg.L-1, while initial metal concentration was 50 mg Me L-1. In the case of lead, the initial pH was not raised at pH 7 to avoid lead hydroxide precipitation. For sorption isotherms, solutions were prepared at the appropriate pH (i.e. pH 3 for lead and pH 6 for mercury) controlled with HCl or NaOH, at different initial metal concentrations, Co (mg Me L-1), ranging between 10 and 200 mg Pb L-1 and between 10 and 100 mg Hg L-1. The sorbent dosage was varied using a variable mass of sorbent m: 10 and/or 20 mg in a fixed volume of solution, V, 0.1 L). After 3 days of contact the solutions were filtered and the residual concentration (Ceq, mg Me L-1) was determined by ICPAES. The mass balance equation was used for calculating the sorption capacity, g (mg Me g-1): q = V/m (Co – Ceq). For sorption kinetics, 1 L of solution (at given concentrations: 20, 50 or 100 mg Me -1 L ) at selected pH was mixed with selected sorbent (at given sorbent dosage: 100, 200, 300 or 500 mg L-1) for 3 to 4 days. Samples were regularly withdrawn and filtered on membranes and residual metal concentration was determined by ICP-AES. 3.

RESULTS AND DISCUSSION

3.1 Influence of pH on sorption properties and biomass selection A preliminary set of experiments was performed with standard conditions. This preliminary study was used to evaluate the efficiency of the different sorbents, to determine the optimum pH and to check the influence of the biomass on the pH of the solution. Results are summarized on Figures 1 and 2 for lead and mercury sorption, respectively. The contact of the biomass with metal ion solution significantly changed the pH of the solution, regardless of the type of metal. Significant differences were observed with the different biomass. With lead, biomass C is the only sorbent that decreases the pH of the solution at weakly acidic pH (i.e. pH 4-6). Biomass D hardly increased the pH (by less than 0.5 pH unit) over the whole pH range investigated. On the opposite hand, the other samples (biomass A, C, E) significantly increased the pH of the solution, especially at pH 901

Biosorption

4: in this case the pH can be increased up to 6.5-6.7 (biomass B). Above pH 5, pH variation was less significant: the biomass had a buffering effect around pH 6. Generally at acidic pH (below pH 4) the pH of the solution was not significantly changed (less than 0.3 pH unit): the alkaline charge of the sorbent and the influence of metal sorption (uptake of OH- in relation with metal sorption, proton release) were not sufficient to change the pH. For mercury sorption, the trends are generally similar, except for biomass C (for which the pH of the solutions was also decreased). The weak buffering effect of biomass was also observed at near-neutral pH. As expected the treatment of sodium hydroxide (regardless of NaOH concentration, temperature of the treatment and contact time) significantly impacted the pH of the solution. This may be explained by an insufficient washing/rinsing of the biomass after alkaline treatment or by the presence of citric acid released from the biomass (in the case of pH decrease). It is especially significant in the case of mercury sorption for which metal sorption was negligible, except in the case of biomass E (Figure 2 shows only sorption data for biomass E, for other biomass the sorption capacities were below 10 mg Me g-1). So the pH variation was only attributable to acid-base properties of the biomass.

8 7

B

6 pHf

Sorption Efficiency (%)

100 A C

5

D

4

E

3 2

A

80

B

60

D

C E

40 20 0

2

3

4 pHi

5

6

2

3

4 pHi

5

6

Figure 1. pH variation during lead sorption (left panel) and sorption efficiency (right panel) in function of initial pH (Co: 50 mg Pb L-1; Sorbent dosage: 300 mg.L-1) 100 A B C D E

8 pHf

Sorption Efficiency (%)

10

6 4 2

80 60 40 20 0

2

3

4

5 pHi

6

7

2

3

4

5

6

7

pHi

Figure 2. pH variation during mercury sorption (left panel) and sorption efficiency in function of pH with biomass E (Co: 50 mg Hg L-1; Sorbent dosage: 300 mg.L-1) 902

Biosorption

Lead sorption efficiency was only significant with biomass B, and to a lesser extent with biomass A (other biomass exhibited sorption efficiency below 20% under selected experimental conditions). Since the biomass strongly increased the pH above pHi = 4, and in order to avoid precipitation phenomena, the initial pH was selected at pH 3. The sorption efficiency was comparable at this pH to those observed at higher pH for biomass B. For these reasons biomass B was selected for further experiments using pH 3 as the optimum pH. It appears at this stage that biomass grown under surface fermentation conditions was more favorable to lead uptake, especially when submitted to alkaline treatment. In the case of mercury uptake, sorption efficiency was negligible in most cases, except with biomass E (submerged fermentation procedure with strongly alkaline treatment). The strong alkaline treatment may drastically change the structure of the biomass. Such a strong treatment is used for the deacetylation of chitin to prepare chitosan, which is characterized by a higher affinity for metals than the raw material. We could expect this treatment being able to remove inert material from cell walls and to transform weakly active sorption sites to strongly reactive functional groups. Less drastic treatment conditions (lower concentration of NaOH, lower contact time, lower temperature) were not able to reach the same reactivity of the biomass for mercury. In this case mercury sorption was mainly efficient at pH equal or greater to pH 4: below pH 4, the sorption efficiency drastically dropped. It is important to observe that the sorption efficiency is significantly lower than the levels reached with lead. For further experiments on mercury sorption, biomass E was selected and experiments were performed at pH 6 (in order to minimize pH variation during metal sorption procedure).

q (mg Pb/g)

3.2 Lead sorption isotherm Figure 3 shows lead sorption isotherm at pH 3 using biomass B. The sorption isotherm is very favorable (almost rectangular). At very low concentration (below 3-4 mg Pb L-1), sorption capacity was negligible. It may be explained by a possible complexation of lead ions by biological material released from the biomass; chelated lead is then less adsorbable on biosorbents. Above this limit concentration a sharp increase in the sorption capacity was observed up to a sorption capacity of about 550-600 mg Pb g-1, followed by a plateau (when the residual concentration exceeded 10 mg Pb L-1).

750 600 450 300

q m: 587.5 mg Pb/g b: 3.80 L/mg

150 0

2

R : 0.955

0

30

60 90 120 150 Ceq (mg Pb/L)

Figure 3. Lead sorption isotherm at pH 3 using sorbent B (symbols: experimental points, lines: Langmuir modeling (q = (qm b Ceq)/(1+ b Ceq)) after correction of the residual concentration to take into account the non-null residual concentration due to a possible complexation of lead by dissolved compounds) 903

Biosorption

Though under selected experimental conditions (pH and metal concentration) there is no precipitation of lead under hydroxide forms, the shape of the curve (almost rectangular) may be indicative of a mixed adsorption/precipitation phenomenon. This may proceed by reaction of lead with some chelating/precipitating agents released from the biomass (phosphate used in the fermentation media; by-products from citric acid production such as oxalate ions). 3.3 Influence of experimental parameters on lead sorption kinetics Figure 4 shows the influence of lead concentration at two different sorbent dosages (pH 3) using biomass B. Sorption capacities have been plotted in function of the square root of time in order to compare sorption kinetic rates: this type of curve is frequently used to evaluate the contribution of intraparticle diffusion resistance [19]. In a first section of the curve corresponding to a latency of about 1-2 hours, the sorption remained negligible, after this period the sorption capacity linearly increased with the square root of time and the slope was of the same order of magnitude for initial concentrations of 50 and 100 mg Pb L-1, significantly higher than those obtained at a metal concentration of 20 mg Pb L-1. The latency decreased with increasing sorbent dosage. 200 SD: 200 mg/L

400 300

Co: 100 mg/L Co: 50 mg/L Co: 20 mg/L

200 100 0

q (mg Pb/g)

q (mg Pb/g)

500

Co: 100 mg/L Co: 50 mg/L Co: 20 mg/L

150 100

SD: 500 mg/L

50 0

0

3 6 1/2 1/2 Time (h )

9

0

1

2 3 4 1/2 1/2 Time (h )

5

Figure 4. Influence of lead concentration on sorption kinetics at pH 3 at 2 sorbent dosages (SD) (Sorbent B; G1 particle size) Figure 5 (left) shows the influence of sorbent dosage on lead sorption kinetics (initial metal concentration: 100 mg Pb L-1). The same trends were observed as in Figure 4. Increasing sorbent dosage slightly decreased the slope of the plot of sorption capacities in function of the square root of time: increasing the sorbent dosage increases the excess of sorption sites (compared to metal ions). As a consequence the residual concentration strongly decreased in the initial stage of the sorption process (after the latency period), then the concentration gradient decreased and the slope of the curve also diminished. Figure 5 (right) shows the plot of sorption capacity as a function of square root of time for different particle sizes of biomass B at pH 3. The slopes of the curves were comparable (same order of magnitude) but the latency time increased with increasing the particle size of the sorbent. The shape of the kinetic curves is a bit unusual. It confirms that the mechanism involved in the removal of lead is not a simple sorption mechanism. Actually, at long contact time a trouble can be observed in the solution indicating the occurrence of precipitation phenomena (not due to hydroxide, but that could be explained by the presence of by-products as suggested above).

904

6

Biosorption

200

Co: 100 mg/L

400 300 200

SD: 200 mg/L

100

SD: 300 mg/L

q (mg Pb/g)

q (mg Pb/g)

500

150 100

SD: 500 mg/L

0 0

3 6 1/2 1/2 Time (h )

SD: 300 mg/L Co: 50 mg/L

50

G1 G2 G3 G4

0 0

9

3 6 1/2 1/2 Time (h )

9

Figure 5. Influence of sorbent dosage (left panel, G1 particle size) and sorbent particle size (right panel) on Pb sorption kinetics at pH 3 (Sorbent B; G1 particle size) To verify this hypothesis the biomass was submitted to a new washing treatment with water and acidic water (pH 3). After 2 days of contact with the “washing” solution, the solution was filtered. The biomass was used for kinetic experiments under comparable experimental conditions to those selected above (pH, metal concentration, sorbent dosage; to be able to compare). A concentrated lead solution (1 g Pb L-1) was added to the filtrate (to a final concentration of 50 mg Pb L-1) and the solution was mixed for 3 days in order to verify the occurrence of precipitation phenomena. Actually, the precipitation was observed in the filtrate obtained from acidic solution treatment of biomass B (pH 3) after a few minutes of contact, when the biomass was treated with demineralized water, even after 2-3 days of contact the precipitation was very low (a few percent of initial metal concentration). Using the washed biomass resulted in slower sorption kinetics (Figure 6). At equilibrium the sorption efficiency was only 50% (instead of 85%), taking into account the fraction of metal precipitated in the experiments performed with the acidic filtrate (about 35-40%, not shown): the mass balance is almost maintained.

1 SD: 300 mg/L Co: 50 mg Pb/L

C(t)/Co

0.8

RB WB

0.6 0.4 0.2 0 0

12

24 Time (h)

36

48

Figure 6. Influence of acid-washing treatment on sorption kinetics (RB: raw biomass, WB: washed biomass) 3.4 Mercury sorption isotherm Mercury sorption isotherms were performed at pH 6 using biomass E in presence of NaCl (0.1 M) and without salt addition (Figure 7). The addition of NaCl strongly reduced 905

Biosorption

mercury sorption and changed the type of sorption model that fitted better experimental data. Maximum sorption exceeded 250 mg Hg g-1; the Freundlich equation fitted better experimental data than the Langmuir equation: the curve can be characterized by an exponential trend and it was impossible to detect the plateau in the concentration range investigated in this study. It is important to notice that similarly to experiments on lead, a sample of biomass was submitted to a complementary washing treatment with water at the pH of sorption experiments, the filtrate was collected and completed with concentrated mercury solution (to a final concentration consistent with those selected for sorption studies). In this case, no precipitation was observed.

300

150

2

q (mg Hg/g)

q (mg Hg/g)

R : 0.950

200 2

100

R : 0.903 q = 21.15 Ceq

1/1.72

0

100 50

q m : 165.3 mg Pb/g b: 0.025 L/mg

0 0

20 40 60 Ceq (mg Hg/L)

80

0

20 40 60 Ceq (mg Hg/L)

80

Figure 7. Mercury sorption isotherm at pH 6 (left: without NaCl; right: with NaCl addition) using sorbent E (symbols: experimental points, lines: Langmuir or Freundlich modeling) When NaCl was added to mercury solution, the sorption capacity was significantly decreased (about two times). In this case the Langmuir equation better fitted experimental data than the Freundlich model. The theoretical maximum sorption capacity was close to 165 mg Hg g-1, higher than the maximum sorption capacity observed in the range of mercury concentration investigated in the present study (about 100 mg Hg g-1). The addition of sodium chloride influences the ionic strength of the solution. The influence of ionic strength is usually important in the case of ion-exchange mechanism. In the case of mercury sorption at pH close to neutrality metal is expected to be adsorbed by chelation on amine groups (less sensitive to ionic strength) or ion-exchange on carboxylic functions. The influence of sodium chloride may be interpreted by either the influence of ionic strength on the ion-exchange mechanism or by a change in the speciation of mercury (formation of chloro-complexes) that may affect its adsorbability. 3.5 Influence of experimental parameters on mercury sorption kinetics The influence of experimental parameters (mercury concentration, sorbent dosage) on sorption kinetics has been investigated (Figure 8). Kinetic decay curves are plotted versus time together with the plots of sorption capacities versus the square root of time. As expected, at decreasing sorbent dosage, the sorption efficiency decreased but with low sorbent dosage (100 mg L-1, not shown), at long contact time a partial release of mercury is observed, especially at high metal concentration and low sorbent dosage (SD: 300 mg L1 ; but also at SD: 100 mg L-1, not shown). This may explain some discrepancies between the results of sorption isotherms and equilibrium points of uptake kinetics. If the time of contact is not the same it can introduces some significant differences. 906

Biosorption

The initial section of the decay curves appeared to be independent of metal concentration. This first part corresponds to a step of the process that is controlled by external diffusion, while the second part of the curve is controlled by the resistance to intraparticle diffusion. Most significant differences were observed in the second stage of the process, indicating that sorbent dosage and metal concentration parameters mainly influenced mass transfer resistance to intraparticle diffusion. Even with a sorbent dosage as high as 500 mg L-1, the sorption efficiency did not exceed 75-80%.

1

150 SD: 500 mg/L

Co: 100 mg/L

Co: 50 mg/L

Co: 50 mg/L

0.6

q (mg Hg/g)

C(t)/Co

0.8

Co: 100 mg/L

Co: 20 mg/L

0.4

Co: 20 mg/L

100 50

0.2

SD: 500 mg/L

0

0 0

24

48 72 Time (h)

96

0

120

2

4 1/2

Time

6

8

1/2

(h )

Figure 8. Influence of mercury concentration on sorption kinetics at pH 6 at 2 sorbent dosages (SD) (Sorbent E; G1 particle size)

SD: 300 mg/L

C(t)/Co

0.8

Co: 100 mg/L Co: 50 mg/L Co: 20 mg/L

0.6 0.4

150 q (mg Hg/g)

1

SD: 300 mg/L

100

Co: 100 mg/L

50

Co: 50 mg/L

0.2

Co: 20 mg/L

0

0 0

24

48 Time (h)

72

96

0

2

4

6 1/2

Time

8

10

1/2

(h )

Figure 8. Influence of mercury concentration on sorption kinetics at pH 6 at 2 sorbent dosages (SD) (Sorbent E; G1 particle size) Particle size hardly influenced sorption kinetics (Figure 9). Regardless of the plot system (C(t)/Co versus time or q versus t0.5), the curves were very close. Decreasing the size of sorbent particles allowed increasing the slope of the kinetic curves but the differences were not very marked. At equilibrium the sorption capacities were very close: sorption did not occur only on the surface of sorbent particles (in this case the sorption capacity would be more sensitive to the external surface area of the sorbent) but also in the whole volume of the solution. But the sorption kinetics was not strictly controlled by the mass-transfer resistance for large particle. 907

Biosorption

100

1

C(t)/Co

0.6 0.4

q (mg Hg/g)

SD: 300 mg/L Co: 50 mg Pb/L

0.8

G1 G2 G3 G4

0.2 0

75 50

G1 G2 G3 G4

SD: 300 mg/L Co: 50 mg Pb/L

25 0

0

12

24 Time (h)

36

48

0

2

4 6 1/2 1/2 Time (h )

8

10

Figure 9. Influence of particle size on mercury sorption kinetics at pH 6 (SD: 300 mg.L-1; Co: 50 mg Pb L-1) 4.

CONCLUSIONS Waste materials (fungal biomass) from industrial fermentation processes have been successfully used for the sorption of lead and mercury after being chemically treated (alkaline treatment). Best sorbents for mercury and lead sorption were obtained after being treated with NaOH solutions with different concentration, contact time and reaction temperature. While best results were obtained at pH 3 for lead, in the case of mercury sorption may be performed at pH close to neutrality. High sorption capacities (between 1 and 2 mmol Me g-1) were observed in the case of lead uptake but it appeared that a part or metal removal was due to a precipitation phenomenon probably due to a release of some sub-products or by-products of the fermentation process (phosphate, organic or inorganic; or oxalate). This illustrates the necessity to use a carefully washed material (especially industrial waste products) in order to avoid artifacts; and the care to take in the interpretation of experimental results. Sorption capacities in the case of mercury were lower, especially in the presence of NaCl that significantly reduced sorption capacities. Sorption kinetics were performed at different sorbent dosages: though the initial sorption is quite fast compared to other systems, 24 hours of contact were necessary to reach the equilibrium (due to low sorbent dosage, below 500 mg L-1). Sorbent particle size has a negligible effect on equilibrium concentrations and kinetics. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

908

B. Volesky and Z. Holan, Biotechnol. Prog., 11 (1995) 235. S. Bosinco, J. Roussy, E. Guibal and P. Le Cloirec, Environ. Technol., 17 (1996) 55. M. Tsezos and B.Volesky, Biotechnol. Bioeng., 24 (1982) 385. E. Guibal, C. Roulph and P. Le Cloirec, Wat. Res., 26 (1992) 1139. Α. Kapoor, and T. Viraraghavan, Biores. Technol., 61 (1997) 221. W. Lo, H. Chua, K.-H. Lam and S.-P. Bi, Chemosphere, 39 (1999) 2723. W. Jianlong, Z. Xinmin, D. Decai and Z. Ding, J. Biotechnol., 87 (2001) 273. D.S. Wales and B.F. Sagard, J. Chem. Tech. Biotechnol., 49 (1990) 345. J. Synowiecki and N.A.Quawi Al-Khateeb, Food Chem., 60 (1997) 605. W.L. Teng, E. Khor, T.K. Tan, L.Y. Lim and S.C. Tan, Carbohydr. Res., 332 (2001) 305.

Biosorption

11. E. Guibal, C. Milot and J. Roussy, Sep. Sci. Technol., 35 (2000) 1021. 12. R. Bassi, S.O. Prasher and B.K. Simpson, Sep. Sci. Technol., 35 (2000) 547. 13. J. Guzman, I. Saucedo, R. Navarro, J. Revilla and E. Guibal, Langmuir, 18 (2002) 1567. 14. M. Jaworska, K. Kula, P. Chassary and E. Guibal, Polym. Int., in press (2003). 15. R.A.A Muzzarelli, F. Tanfani and G. Scarpini, Biotechnol. Bioeng., 22 (1980) 885. 16. Kapoor, and T. Viraraghavan, Biores. Technol., 63 (1998) 109. 17. B.J. Mcafee, W.D. Gould, J.C. Nadeau and A.C.A. da Costa, Sep. Sci. Technol., 36 (2001) 3207. 18. S. Arcidiacono and D.L. Kaplan, Biotechnol. Bioeng., 39 (1992) 281. 19. E. Guibal, I. Saucedo, J. Roussy, C. Roulph and P. Le Cloirec, Water S.A., 19 (1993) 119.

909

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Lanthanum and neodymium biosorption by different cellular systems M.C. Palmieria and O. Garcia Jr.† Department of Biochemistry, Institute of Chemistry, São Paulo State University, P.O. Box 355, Araraquara, SP 14.801-970, Brazil* Abstract Biosorption isotherms for lanthanum and neodymium were determined utilizing microalgae Monoraphidium sp., filamentous fungi Penicillium sp., algae Sargassum sp. and the yeast Saccharomyces sp. as biosorbents at pH from 2 to 5. The maximum biosorption coefficient for each isotherm was determined based on Langmuir Model fitting. A growth linear dependence with pH was observed for lanthanum biosorption (best condition achieved at pH 5), while neodymium biosorption presented little variation in the range studied. The highest biosorption coefficient was obtained using Monoraphidium sp. (1320 mg.g-1 for lanthanum and 1600 mg.g-1 for neodymium), while the Saccharomyces sp presented values around 280 and 460 mg.g-1 respectively. Sargassum sp. and Penicillium sp. presented very similar biosorption for both elements (60-80 mg.g-1). Desorption experiments using HCl 0.24 Mol.L-1 showed a total recovery for the metals initially adsorbed. 1.

INTRODUCTION Due to their unique physical and chemical properties, rare-earth (RE) metals are employed to manufacture advanced materials for high technology devices. The purity grade of these elements determines its level of technological applications; so far, numerous attempts are being made to develop efficient separation and concentration processes for rare-earth (RE) metals [1]. From the 15 RE elements, lanthanum and neodymium are the most common found in monazite sands at Brazilian shores, have been used in laser sources, hydrogen absorbents, magnetic substances and fluorescent materials [2]. Solvent extraction is well known as an effective technique used for separation and concentration of these RE metals at industrial scale. However, their very similar physical and chemical properties require a large number of stages in a series of mixer-settlers to obtain high purity products, increasing costs [3]. Biosorption of metals has been extensively studied as a low cost process for removal of ionic metals from industrial effluents and wastewater. Recently, the increasing development of biosorption process has attracted attention as an alternative for a †

Present address: Department of Chemistry – SCELISUL – CEP 11900-000, Registro – SP, Brazil. Corresponding author ([email protected]).

911

Biosorption

concentration of RE metals [4-8]. It is based on the immobilization of these metals by chemical and physical properties of cell components, such as carboxyl groups, sulphidril groups, amino groups, sulphate and phosphate groups. Due to the different chemical and physical constitution of each species of microorganism, the interaction with metallic ions occurs in varied forms, however it appears to be predominantly by ion-exchange mechanism if no-metabolic uptake is present [9]. Several biosorbents have been studied such as algae, microalgae, filamentous fungi, yeast and bacteria. Many metals have been investigated to be removed by biosorption processes, such as uranium, thorium, chromium, zinc, gold and many others [10-12]. Previous works have demonstrated the potential of different biomass in removing erbium and ytterbium from acid solutions [4]. The initial results obtained for lanthanum biosorption studies using Sargassum sp. [6] and neodymium on different kinds of biomass [5] also pointed out for the future use of this process as a low cost alternative method for the concentration of these metals 2.

MATERIALS AND METHODS

2.1 Organisms and culture conditions The fungi Penicillium sp. was kindly supplied by Dr. Rubens Monti from UNESP – Araraquara – SP, Brazil. Oat flour/agar plates were inoculated with fungi spores and incubated at room temperature (32ºC) for 72 h. Fungal mycelia were collected by scratching the culture surface and this biomass was washed with 100 mL boiling distilled water to remove agar residues. Microalgae Monoraphidium sp., was kindly donated by Dr. Armando Vieira from UFSCar – São Carlos - SP, Brazil. Flask containing 500 mL of WC medium [7] was inoculated with 20 mL pre-inoculum culture of algae and incubated at 25ºC for 60 days. The Saccharomyces sp. strain was obtained a single colony from commercial Baker’s yeast. Malt extract broth (10 mL) was inoculated with this culture and incubated at 32ºC for 16 h. For yeast biomass production, this previous culture was inoculated in 50 mL of malt extract broth and incubated on rotatory shaker at 150 rpm and 32ºC for 16 h. The biomass of the microorganisms used in this study was finally harvested by filtration on membrane (0,45 µm pore size) and washed twice with distilled water. The filtered biomass was then suspended in 20 mL of distilled water for further experiments. The brown seaweed Sargassum sp. was collected from the coast of São Sebastião – SP, Brazil. The biomass was washed and sun dried for stock at room temperature. For biosorption experiments the algae was shopped in pieces with size around 0.3-0.5 cm, washed twice in distilled water and twice in hydrochloric acid 0.1 Mol.L-1 and after that it was washed in distilled water up to close pH 4.0. 2.2 Lanthanum and neodymium stock solutions Lanthanum and neodymium stock solution were prepared by dissolution of the correspondent oxides (Aldrich, 99.9%) using concentrated HCl to a final concentration of 5.23 g.L-1 and 5.15 g.L-1, respectively, at pH 1.0. The metal content in solution was determined by inductively coupled plasma atomic spectroscopy (ICP-AS Thermo Jarel Ash, Trace Scan).

912

Biosorption

2.3 Biosorption assays Biosorption experiments for isotherm determination were carried out by using 0.1 g of selected biomass in 40 mL of lanthanum or neodymium solution in concentrations ranging from 0.01 to 2.5 g.L-1. The pulp was incubated at 30°C in a rotatory shaker and pH was adjusted every 1 h to keep it constant at 5.0 using 0.1 Mol.L-1 NaOH or HCl. The equilibrium was reached after 24 h and samples (5 mL) were taken out and filtered on membrane (pore size 0.45 µm) for metal determination in the filtrate. Each experiment was performed in triplicate. Metal biosorption coefficient (q) was calculated [8] according to equation (1): (1) q = (Ci-Cf)V/M . -1 where Ci is the initial metal concentration (mg L ), Cf is the final metal concentration after contact with biomass (mg.L-1), V is the solution volume (mL) and M is the sorbent weight in dry form (g). 2.4 Desorption assays Desorption experiments were carried out using 0.1 g of biomass previously loaded with metal (lanthanum or neodymium) at pH 5.0 as described above. The loaded biomass were drained to remove all solution remained and then mixed in 40 mL of chloridric acid 0.24 Mol.L-1 at 30°C in an orbitary shaker for 16 hours to reach the equilibrium. Samples from supernatant were taken out for metal and pH determination. Desorption coefficient were defined as: qd = Cf V/M (2) . -1 where: qd is the desorption coefficient (mg g ), V is the acid volume used (L) and M is the dry mass of biosorbent (g). 3.

RESULTS AND DISCUSSION

3.1 Effect of pH Considering biosorption systems similar to an ion-exchange system, where the components of the cell envelope proceed like binding sites for metal exchange, it is naturally expected that the pH has an important role in the interaction between the metal and the binding sites. The pH value affects the dissociation of chemical species in the biomass responsible for capturing the metal in solution. Indeed it can affect the speciation of the metal, changing its charge, producing complexes or associations that can favor or difficult the biosorption, once this interaction is basically electrostatic and depends on charges in both metal and binding site to be effective. Usually the increase in the pH value tends to increase the biosorption for a metallic cation. This effect can be observed for several metals and different kind of biomass. The Figure 1 shows the pH effects on the biosorption of lanthanum and neodymium by the biomass utilized. For lanthanum biosorption (fig. 1) it can be noted an evident linear increase in the qmax for Monoraphidium sp. and Saccharomyces sp. as the pH increases. For Sargassum sp. this effect was also detected (linear increase in the qmax from 10 to 80 mg.g-1) but it is not so evident in the graph, since the y scale was expanded to show Monoraphidium sp. curve. The biosorption of lanthanum by Penicillium sp. was not influenced by pH, since the qmax remains around 50 mg.g-1 in the pH values tested. As the pH increases the 913

Biosorption

dissociation of binding sites present in the biomass favor the ion-exchange. Similar results were observed in previous work for biosorption of erbium and ytterbium (other RE metal) in the pH range from 2 to 5 for the same kind of biomass utilized in this study [4]. The increase in pH value above 5 does not imply on increase in biosorption capacity, once this condition leads to the formation of insoluble metal hydroxide. 1500

Lanthanum

1250

1500

Neodymium

1250 -1

qmax (mg g )

-1

qmax (mg g )

1000 750 500

1000 750 500

250

250

0

0 2.0

3.0

4.0 pH

5.0

2.0

3.0

4.0

5.0

pH

Figure 1. Effect of pH on biosorption using Monoraphidium sp. ( ), Saccharomyces sp. ( ), Sargassum sp. ( ) and Penicillium sp. (O) The effect of pH value in neodymium biosorption by Saccharomyces sp., Sargassum sp. and Penicillium sp. was very similar to lanthanum biosorption (fig. 1). However, Monoraphidium sp. presents a distinct pattern, since its biosorption coefficients show, on the contrary, a little decrease as the pH increase from 2.0 to 5.0. The reasons for this behavior remain not clear until now, and new studies are being conducted to elucidate this interesting characteristic. Despite the high qmax at pH close to 2.0 for neodymium biosorption by Monoraphidium sp., when the pH is lowered below 1 all the metal can be desorbed as it can be seen in the results of desorption presented later. 3.2 Isotherms of biosorption The isotherms of biosorption are curves that describe the equilibrium between the metal in solution and the biosorbent at a constant temperature. These curves are extensively used for comparison of biosorption performance of different biosorbents. Figure 2 presents the biosorption isotherms for lanthanum and neodymium with the different biomass studied. For all systems the curves presented a hyperbolic shape, and after fitting using the Langmuir model, the qmax was determined, which indicates the maximum capacity of recovery of the metal for a biosorbent in these conditions. The results obtained from the Langmuir fitting showed a wide range of values reflecting the differences in chemical composition of each biomass studied. Monoraphidium sp. was the best biomass for biosorption of both metals (qmax 1300-1600 mg.g-1, values calculated from curve fitting), followed by the yeast Saccaromyces sp. with maximum biosorption coefficient of 280 and 460 mg.g-1 for lanthanum and neodymium respectively. Sargassum sp. and Penicillium sp. presented very similar biosorption performances (qmax around 60-80 mg.g-1) for both metals and lower when compared to Monoraphidium sp. and Saccharomyces sp. values. These results confirm preliminary studies of 914

Biosorption

biosorption with lanthanum and neodymium [5, 6] and pointed out that these two metals (light RE) have similar biosorption behavior than erbium and ytterbium (heavy RE) [4]. 1400 Lanthanum

1200

1000

1000

800

800

-1

q (mg.g )

-1

q (mg.g )

1200

1400

600 400

600 400

200

200

0

0 0,0

0,5

1,0

1,5 -1

Cf (g.L )

2,0

2,5

Neodymium

0,0

0,5

1,0

1,5

2,0

2,5

-1

Cf (g.L )

Figure 2. Biosorption isotherms using Monoraphidium sp. ( ), Saccharomyces sp. ( ), Sargassum sp. ( ) and Penicillium sp. (O) at 30°C. Continuous line represents a Langmuir curve fitting (Biosorbent concentration 0.1 g.L-1 and pH 5.0) 3.2 Desorption of lanthanum and neodymium Desorption consists in the shift of ion-exchange equilibrium in favor to release the metal previously loaded in the biomass. This shift can be achieved using several chemicals, but once binding groups in the biomass are known for their high selectivity for protons, small volumes of diluted mineral acids usually give complete desorption. As the desorption equilibrium is the reverse process of biosorption, it can also be determined a desorption isotherm (at constant temperature), that can be fitted according to Langmuir model, which allows to determine the maximum coefficient of desorption (qdmax). This coefficient is very similar to qmax, but while the first represents the amount of metal bound in the biomass, qdmax indicates the amount of metal released to solution. If there is complete desorption from the biomass, qmax should be equal to qdmax. In Table 1 are presented the maximum coefficients of biosorption (qmax) and the maximum coefficients of desorption (qdmax) for lanthanum and neodymium utilizing HCl 0.24 Mol L-1 for all biomass utilized. It is possible to observe that the coefficients are very close, confirming that both metals can be removed from the biomass efficiently. The differences observed in the values showed in table 1 are due to experimental deviances, once each value in qmax and qdmax are calculated from independent curve fitting, which explains some desorption values higher than adsorption. Despite the favorable biosorption behavior of neodymium in pH close to 2, it can be observed in table 1 that the ion exchange equilibrium can be shifted in lower pH in order to fully release the metal adsorbed. New studies are been conducted in order to establish a satisfactory explanation to the particular behavior of this RE metal, as stated above.

915

Biosorption

Table 1. Maximum coefficients of biosorption (qmax) and desorption (qdmax) for lanthanum and neodymium Biosorbents Monoraphidium sp. Saccharomyces sp. Sargassum sp. Penicillium sp.

Lanthanum . -1

Neodymium . -1

. -1

qmax (mg g )

qdmax (mg g )

qmax (mg g )

qdmax (mg.g-1)

1380.6 280.2 75.3 52.9

1368.4 273.5 80.8 55.1

1578.2 455.7 73.7 67.3

1580.9 448.6 75.2 66.1

4.

CONCLUSIONS Recent papers have pointed out the potential of RE biosorption as an alternative process for concentration of these metals. However, there is still a lack of information about biological, chemical, physics and engineering of RE biosorption process which has to be investigated to achieve a feasible industrial alternative. The results presented pointed out the similarities of biosorption behavior of two light RE metal (lanthanum and neodymium) with two heavy RE (erbium and ytterbium) studied in previous work. As the heavy element studied, lanthanum and neodymium are capable to be removed from solution by all the biosorbent utilized, with highest performance for microalgae Monoraphidium sp. However, the effect of pH presented some differences. While lanthanum biosorption showed in general an increase in the maximum biosorption coefficient with the increase of pH, as observed for erbium and ytterbium, it has practically no effect in the biosorption of neodymium for Monoraphidium sp. More studies have to be conducted on the speciation and interaction of neodymium with this cellular system to elucidate this point. The desorption of lanthanum and neodymium from the biomass was completely performed using hydrochloric acid and all the metal previously adsorbed was removed. This easily reversibility of the process using diluted mineral acid can be an interesting point for future industrial applications. The data obtained in this work and the advances in biosorption studies indicate a promising alternative for these metals concentration and separation process. ACKNOWLEDGEMENTS We acknowledge FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo) for a doctoral fellowship (M.C.P.) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) for research fellowship to O. G. Jr. We also acknowledge W.A.N. Meneses (IQ) and W.C. Costa (CNEN - MG) for technical assistance. REFERENCES 1. H. Hubicka and D. Drobek, Hydrometallurgy, 47 (1997) 127. 2. Roskil information service LTD. www.yesresources.com/reviews/industriaminerals/ milestone.asp - acessed 05/10/2001. 3. T. Kakoi, T. Nishyori, T. Oshima, F. Kubota, M. Goto, S. Shinkai and F. Nakashio, Journal of Membrane Science, 136 (1997) 261. 4. M.C. Palmieri and O. Garcia Jr. Biohidrometallurgy: fundamentals, technology and sustainable development, Elsevier Science B.V., part B, 2001. 916

Biosorption

5. M.C. Palmieri and O. Garcia Jr., Process Biochemistry, 36(5) (2000) 441. 6. M.C. Palmieri, B. Volesky, O. Garcia Jr., Hydrometallurgy, 67 (2002) 31. 7. T.R. Muraleedharan, L. Iyengar, L. Philip, C. Venkobachar (2001) -www.sph.umich. edu/eih/heavymetals/manuscripts/muraleedharantr.htm - accessed 05/10/2001. 8. A.C. Texier, Y. Andres, P. Cloirec, Environmental Science Technology, 33 (1999) 489. 9. H. Eccles, TIBTECH, 17 (1999) 462. 10. B. Volesky and Z.R. Holan, Biotechnology Progress, 11 (1995) 235. 11. M.G. Gadd and C. White, TIBTECH, August 11 (1993) 353. 12. F. Veglio and F. Beolchinni, Hydrometallurgy, 44 (1997) 301. 13. P. Ahuja, R. Gupta, R.K. Saxena, Process Biochemistry, 34 (1999) 77. 14. J.M. Modak, K.A. Natarajan, Minerals Metallurgical Processes, 12 (1995) 189.

917

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Modeling of chromium biosorption by seaweed Sargassum sp. biomass in fixed-bed column in series E.S. Cossicha, E.A. Silvab, C.R.G. Tavaresa, H.M. Mesquitaa and L.S. Eidana* a

Departamento de Engenharia Química, Universidade Estadual de Maringá - UEM Av. Colombo, 5790. CEP: 87020-900 - Maringá, PR - Brasil b Departamento de Engenharia Química, Universidade Estadual do Oeste do Paraná UNIOESTE – Toledo, PR. Abstract The biosorption of chromium by the marine alga Sargassum sp. were investigated in a system with two fixed-bed columns in a series configuration (temperature = 30°C; pH = 3.5). A model that describes the dynamics of chromium ion sorption in the columns was obtained from the mass balance in the fluid phase and the biosorbent. This model considers both the axial dispersion effect in the column and the mass transfer resistance in the biosorbent. The model performance was evaluated from experimental data (breakthrough curves) obtained at pH 3.5, flow rate of 6 mL/min, feed concentration of 0.97 mmol/litre and temperature of 30°C. The results showed that the model is appropriate to represent the chromium biosorption by seaweed Sargassum sp. biomass in fixed-bed columns in series. Keywords: chromium biosorption, Sargassum, fixed-bed columns, modeling 1. INTRODUCTION The increase in metal consumption in industrial scale represents an important environmental issue. Chromium is present in different types of industrial effluents, being responsible for environmental pollution. Traditionally, the chromium removal is made by chemical precipitation, a conventional method for removing metals. However, this method is not completely feasible to reduce the chromium concentration to levels as low as required by environmental legislation. Biosorption processes have been proposed as an alternative method for recovering and removing metals from industrial effluents with metal concentration in range from 1-100 mg/L [1]. Most separation and purification processes that employ sorption technology use continuous-flow columns. This operating mode ensures the highest possible concentration difference driving force and avoids a subsequent solid-liquid separation process. Starting at the inlet, the saturated solid sorbent zone gradually extends throughout the column; the sorbate eventually breaking through the column. The record of the breakthrough gives

* Corresponding author: [email protected]

919

Biosorption

usually a typical S-shaped breakthrough curve, whose shape and slope is the result of the equilibrium sorption isotherm relationships, the mass transfer to and throughout the sorbent in the column, and operation macroscopic fluid-flow parameters, such as axial mixing, affecting the deviation from the ideal plug-flow. The aim of the present work was to model chromium biosorption in a system with two fixed-bed columns in a series configuration by Sargassum sp., a brown seaweed present in abundance on the Brazilian coast. 2.

MATERIAL AND METHODS The biomass used was the brown seaweed Sargassum sp. It was washed in water, rinsed with distilled water and dried in an oven at 60°C during 24 hours. The columns were packed with biomass in natural size (with leafs and thallus). Dry weight of biomass was obtained after drying at 105°C for 24 hours. Chromium solution was prepared by dissolving CrK(SO4)2.12H2O (analytical grade) in deionized water. Continuous-flow sorption experiments were conducted in two steel columns with controlled temperature. The columns used had a height of 50 cm and a diameter of 2.8 cm. The bed length used in the experiments was 30.6 cm. A peristaltic pump fed the chromium solution (0.97 mmol/litre) to bottom of the first column with a flow rate of 6 mL/min, and the effluent of the first column fed the bottom of the second column. The pH of the solution in the feeding tank was maintained constant at 3.5. The temperature of stream feeding solution and of the column was controlled at 30oC through a thermostatic bath. Liquid samples of the concentration of chromium in the exit of each column were collected at pre-defined time intervals. The total concentration of chromium in liquid samples was determined by Atomic Absorption Spectroscopy (Varian SpectrAA-10 plus). When the system reaches equilibrium, the metal concentration in the fluid phase is constant along the column and equal to the feed concentration (C = C* = CF). The biosorption capacity of the chromium was calculated from the experimental breakthrough curves, using the following equation: t C F Q& q* = 1 − C z = L / C F dt (1) 1000 m s 0

∫(

)

The integral represented by Eq. (1) was solved analytically by means of the polynomial approach of the term 1 − C / C F .

(

)

The column void fraction, ε , was determined by the measure of the void volume (volume of distilled water required to fill the bed), as methodology proposed by Cossich [2]. At the end of each experiment, the solution present inside the columns was removed. The exhaustion of the solution was accomplished from the bottom of the column using a minimum period of 24 hours. Afterwards, a peristaltic pump (Cole Parmer) fed the columns from a reservoir that contained a defined volume of distilled water. The necessary volume of water to fill the bed was determined initially by the difference between the volume contained in the reservoir and the volume remaining after filling the bed. The column void fraction was calculated using the following equation: 920

Biosorption

ε=

VV Vb

(2)

where, VV is the bed void volume and Vb is the bed volume. 3.

MODEL DESCRIPTION In recent years, many mathematical models have been tested to represent the biosorption of different metals in fixed-bed columns [3-7]. The mathematical model for biosorption of a metal ion in a fixed bed column was obtained by means of the mass balance equations applied to an element of volume of the column in the liquid phase and in the solid phase (biosorbent). In the model development the following hypothesis were considered: Isobaric and isothermic process; Constant physical properties; Superficial adsorption; Negligible radial dispersion. The mass balance equation for the fluid phase is: ∂C 1 ∂q ∂C 1 ∂ 2C + ρb = −u + ∂τ ε b ∂τ ∂ξ Peb ∂ξ 2

(3)

with the following initial and boundaries conditions: C (ξ,0 ) = C 0

(4)

∂C = Peb (C (τ ,0) − C F ) ∂ξ

in ξ = 0

(5)

∂C =0 ∂ξ

in ξ = 1

(6)

To obtain the modeling of the copper adsorption rate in the biosorbent it is assumed that the driving force for the mass transfer is linear with the concentration for the solid phase (biosorbent) and the adsorption rate is represented by the following equation: ∂q = − Shm (q − q* ) ∂τ

(7)

with the following initial condition: q(ξ,0) = q 0

(8)

qm b C * 1 + b C*

(9)

The equilibrium concentration of chromium adsorption in the algae (q* ) were calculated by Langmuir isotherm model, described by following equation: q* =

The partial differential equation system of the model was solved by the Galerkin Method on finite elements [8-9]. The concentration of the chromium in the bulk fluid phase and the biosorbent were approximated in each element using quadratic polynomials Lagrange. 921

Biosorption

The resulting equation system of application of the method was solved using the DASSL subroutine [10], whose source code is in the FORTRAN computer language. The DASSL code was used to solves a system of algebraic /differential equations. The parameters of the model, that is, solid mass transfer ( K S ) and axial dispersion ( DL ) coefficients, had their estimated values obtained by minimizing an objective function using the Nelder and Mead method [11]. The minimized objective function was: np

F=

∑ (C

EXP out

MOD (K S , DL )) − Cout

2

(10)

i =1

where: EXP C out - Experimental concentration of the chromium in the outlet of the column; MOD C out - Concentration of the chromium determined by the solution of the model in the outlet of the column; np - Number of experimental data points;

The proposed model was able to represent the chromium biosorption by Sargassum sp. biomass in a fixed-bed column (Cossich, 2002; Cossich et al. 2002). Now, it was used to describe the sorption dynamic of chromium in two fixed-bed columns in series. 4.

RESULTS AND DICUSSION The results of the experimental and model simulated breakthrough curves, obtained are shown in Figure 1. The equilibrium data of chromium(III) biosorption system by seaweed Sargassum sp. (at pH 3.5, 30°C) and the Langmuir isotherm model curve obtained by fitting the experimental, obtained by Cossich [2], were used to perform the simulations. The Langmuir parameters used were qm = 1.07 mmol/g and b = 7.9 L/mol. The model parameters, mass transfer and axial dispersion coefficients, used to simulate the experimental breakthrough curves were KS = 1.502 x 10-3 min-1 and DL = 1.20 x 10-4 cm2/min. This values were estimated by Cossich [2] in sorption experiments with only one column (with same dimensions and at the same experimental conditions), at diferents flow rates (2, 4, 6 e 8 mL/min). As illustrated in Figure 1, the model proposed described adequately the chromium adsorption dynamics by seaweed Sargassum sp. in two columns in series. The first column presented a chromium(III) uptake capacity of q* = 0.75 mmol/g, while the second column presented a chromium(III) uptake capacity of q* = 1.28 mmol/g. As the biosorbent mass in the column were equal and they were in equilibrium, the chromium(III) uptake capacity should be the same for both columns. This fact could be caused by changes in the solution pH. The chromium solution leaves the first column and comes in the second column with a pH higher than 3.5. In this kind of system, the capacity of chromium biosorption by biomass increases with pH [2].

922

Biosorption

1.0

Cout / C

F

0.8

0.6 Experimental - Column 1 Experimental - Column 2 Model - Column 1 Model - Column 2

0.4

0.2

0.0 0

2000

4000

6000

8000

time (minutes) Figure 1. Experimental breakthrough curve and model calculated breakthrough curve 5.

CONCLUSIONS In this study, the chromium removal by using seaweed biomass as biosorbent in two fixed-bed columns in a series configuration was investigated. The experimental data showed that the capacity of chromium removal was different for each column. In the first column the capacity of chromium removal was about 0.75 mmol/g, while for the second one it was about 1.28 mmol/g. The mathematical model proposed, that assumes the overall sorption rate is controlled by the mass transfer resistance in the biosorbent, represented the experimental breakthroughs. NOMENCLATURE b - Langmuir isotherm constant (litre/mmol); C - Concentration of the copper(II) in the bulk fluid phase (mmol/litre); C * - Equilibrium concentration of the copper(II) in the bulk fluid phase (mmol/litre); C 0 - Initial concentration of the copper(II) in the bulk fluid phase (mmol/litre); C F - Concentration of the copper(II) in the inlet in the column (mmol/litre); D L - Axial dispersion coefficient (cm2/min); K S - Overall mass transfer coefficient in the biosorbent (min-1); L - Length of the bed (cm); m s - Dry weight of biomass (g); q - Concentration of copper(II) adsorption in the algae (mmol/g);

q* qm Q&

- Equilibrium concentration of copper(II) adsorption in the algae (mmol/g); - Langmuir isotherm parameter (mmol/g); - Volumetric flow rate (cm3/min); 923

Biosorption

t u z VV Vb

-

Time (min); Velocity (cm/min); Axial coordinate in the column (cm); Void volume (litre); Fixed bed volume (litre).

Dimensionless group

Peb Shm ε ρb τ ξ -

Peclet number for the bed (Lu / DL ) ; Modified Sherwood Number (K S L / u ) ; Column void fraction; Fixed bed density (g/litre); Dimensionless time coordinate (tu / L ) ; Dimensionless axial coordinate ( z / L ) ;

REFERENCES

1. B. Volesky (ed.), Biosorption of Heavy Metals, CRC Press: Boca Ranton, FL, 1990. 2. E.S Cossich, PhD thesis, Universidade Estadual de Campinas, Campinas, Brazil, 2000. 3. E.M. Trujillo, T.H Jeffers, C. Ferguson and H.Q Stevenson, Environ. Sci. Technol., 25 (1991) 1559. 4. B. Volesky and I. Prasetyo, Biotechnol. Bioeng., 43 (1994) 1010. 5. D. Kratochvil, B. Volesky and G. Demopoulos, Water Res., 31 (1997) 2327. 6. D. Kratochvil and B. Volesky, Water Res., 32 (1998) 2760. 7. E.S. Cossich, E.A. Silva, C.R.G. Tavares, L. Cardozo Filho and T.M.K. Ravagnani, Proceedings of III Brazilian Adsorption Symposium, (in Portuguese), Recife-PE, (2002) 231. 8. T.J. Chung, Finite element analysis in fluid dynamics, McGraw-Hill, Inc., New York, 1978. 9. B.A. Finlayson, Nonlinear analysis in chemical engineering, McGraw-Hill, Inc., New York, 1980. 10. L.R. Petzold, A description of DASSL: a differential/algebric equation system solver, STR, SAND82-8637, Livermore, 1982. 11. J.A. Nelder and R. Mead, The Computer Journal, 7 (1965) 308.

924

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Modelling and optimisation of copper ion uptake by Acidithiobacillus ferrooxidans A. Boyer, F. Baillet, J.-P. Magnin and P. Ozil Laboratoire d'Electrochimie et de Physico-Chimie des Matériaux et des Interfaces, UMR 5631 - CNRS - INPG – UJF, Equipe Génie des Procédés, BP 75, 38402 Saint Martin d'Hères, France Abstract The copper ion uptake capacities of a living Thiobacillus ferrooxidans biomass were modelled and optimised thanks to the Design Of Experiment (DOE) methodology. A composite centred in a cube design was used to find out an empirical quadratic model with respect of pH, agitation speed, temperature, protein concentration and bacterial physiological state, while metal concentration was fixed at 1 g.L-1. An optimal copper ion uptake capacity of 74.9 mg of copper / 100 mg bacterial dry weight was found with a middle exponential growth phase biomass, for a protein concentration of 45 mg.L-1, a pH of 5, a temperature of 30°C and an agitation speed of 50 rpm. The influence of each parameter on fixation was finally discussed.

Keywords: Acidithiobacillus ferrooxidans, DOE, modelling, copper fixation 1.

INTRODUCTION In the wastewater treatment field, the fixation of heavy metals on biomass has been more and more investigated [1]. Particularly, Acidithiobacillus ferrooxidans had shown great uptake capacities of diverse metal ions: Cu2+ [2], Cd2+ [3], Cr6+ [4], UO22- [5] or Ag+ [6]. Previous to consider a removal process using a biomass, its feasibility can be evaluated by studying the uptake capacities of the microorganisms. These capacities can be determined by modelling the metal ion uptake as a function of various parameters. Freundlich and Langmuir adsorption isotherms, traditionally used to describe solute-solid interaction as physi-sorption or chemi-sorption on inert surfaces, were often used to describe metal removal by died or dried biomass (microorganisms or exopolymers) [7-8]. However, for metal ion removal by a living biomass, other active processes could occur beside the simple surface adsorption (biosorption) and they should be taken into account to describe the metal removal. Uptake of metal by a micro-organism would then be a complex process including biosorption on the cell surface, bioaccumulation in the cytoplasm through non specific cation transport system or bioprecipitation of metal on the cell surface [1]. For such a complex process of fixation, Freundlich and Langmuir adsorption isotherms were no more adapted: these models were inadequate for some biosorption experiments on Chorella fusca [9] or on Echerischia coli [10]. Moreover, copper ion uptake by A. ferrooxidans could not be described satisfactory by such 925

Biosorption

isotherms [2]. Consequently another type of model should be used to characterise metal ion uptake by a living biomass. The classical well-known OVAT method (One Variable At Time) allows to study the influence of one parameter at a time, the value of the other being fixed. This method is rather long (243 trials for 5 parameters studied at 3 levels) and does not allow to take into account the interactions between the different parameters. On the other hand, Design Of Experiments (DOE) allows to simultaneously study from a minimal number of runs (32 runs for 5 parameters and 3 levels, for example) the influences of all the parameters by the way of a postulated polynomial response model [10-12]. However when a model has been determined from the experiments of a DOE, a careful analysis has to be performed for checking its adequacy and validity before using it for prediction or optimisation. This analysis, based on classical statistic tests such as t- and F- tests, is essential and too often neglected. Therefore, we describe each step of analysis through the modelling and the optimisation of copper ion uptake by a living biomass of A. ferrooxidans. The influence of the 5 chosen operating parameters is finally discussed. 2.

MATERIALS AND METHODS

2.1 Microorganisms and medium Acidithiobacillus ferrooxidans DSM583 was grown in 9K medium [13], consisted in FeSO4.7H2O (33 g.L-1), MgSO4.7H2O (0.4 g.L-1), K2HPO4 (0.4 g.L-1) and (NH4)2SO4 (0.4 g.L-1). pH was adjusted at 1.4. The bacterial cultures were performed in a 50 L-reactor with 9K medium at 30°C and under air flow agitation. Bacterial growth was controlled from measuring the amount of substrate, Fe2+ ions, remaining in solution. The ophenantroline colorimetric method [14] associated with Fe3+ reduction with hydroxylamine allowed to evaluate the ratio Fe3+/(Fe2++Fe3+), corresponding to the fraction of substrate still available. Once this ratio reached the desired value (65, 80 or 95%), bacterial biomass was separated by cross-flow filtration and concentrated by centrifugation at 5000 rpm. The protein concentration was then determined by the Lowry colorimetric method [15]. The bacterial dry weight concentration was 1.67 times the protein concentration [16]. Biomass was stored at 5°C for one night before utilisation. 2.2 Copper ion uptake experiments They were carried out in 100 ml-sterile polyethylene flask reactors filled with 40 ml of metal solution (1 g.L-1 Cu2+, CuSO4.5H2O Normapur, Prolabo) under rotary stirring. The solutions were previously heated in the incubator at 30°C or cooled with a waterglycol mixture circulation at 5 or 17.5°C, the selected temperatures. The required amount of biomass was centrifuged at 10000 rpm and the bacterial pellet was added to the metal solution at time zero. Then pH was adjusted by adding either concentrated sodium hydroxide or sulphuric acid depending on the selected pH. After a 90-minute contact, the metal solution was filtered on 2 µm Millipore filters to separate the bacterial biomass. The metal concentration of the solution free of bacteria was determined by atomic absorption spectrophotometry. Control experiments were performed without bacteria to check that copper ion concentration in solution was constant throughout the experimental procedure. 2.2 Experimental Design Method Acting parameters and response choice. First, some parameters were fixed at suitable values. The contact time was set at 90 minutes as a result of preliminary trials about the kinetics of metal fixation on the bacterium [2-4]. Actually, the equilibrium contact time 926

Biosorption

was found to be about 15 minutes for these previous experiments, results in good agreement with literature [8,17]. The metal concentration was fixed at 1 g.L-1 in order to obtain the maximum amount of metal the bacterium could uptake. The acting varying parameters have been determined from previous experiments [2-4] and their ranges were chosen as large as physically possible: * pH: pH, a well-known parameter for metal fixation on micro-organisms, has been previously shown to have a large effect on metal ion sorption. A large range from 1 to 5 was chosen. The lower limit was fixed at 1 to insure a pH avoiding any damage for the bacterium and the upper limit was fixed just under the pH corresponding to the solubility limit of copper ions. * ω: only one work [18] described the influence of stirring, but it seemed important to study this parameter because of its action on external diffusion which is one of the fixation steps. Agitation was varied by using a rotary shaker between the minimal efficient speed, 50 rpm and the upper limit of the device, 150 rpm. * T: temperature influence on fixation has often been studied, showing that the optimal temperature for adsorption is generally the growing temperature for most of the (living) microorganisms [17] but can also be 5°C for some of them [7]. The temperature ranged between 5 and 30°C. The lowest value corresponded to the limit allowing a reliable and accurate enough temperature control while the upper limit was set at 30°C, the growth temperature above which the bacterium could be damaged. * φ: the physiological state of the bacterium also acted on fixation as pointed out by some previous results on the influence of the age of culture on adsorption [3-4, 19]. The physiological state of the bacterium was correlated to the ferrous oxidation activity expressed by the ratio Fe3+/(Fe2++Fe3+) which varied from 65 to 95%. The lower limit was chosen in the middle of the exponential phase, which corresponded to 0.65, whereas the upper limit was fixed at the end of growth, in the stationary phase that corresponds to 0.95. * [prot]: the protein concentration, one of the most studied parameter, has been shown to have a great influence on adsorption [8, 20]. The protein concentration range was chosen as 45-320 mg.L-1 in connection with the ability of protein concentration determination, but also to ensure a significant difference in metal concentration for the analysis. The response under study, Y, was the quantity of fixed metal (in mg) by a 100 mg dry weight of bacteria (unity: mg / 100 mg dry weight). Design choice. The five chosen parameters were supposed to act on the response through a non-linear way and with possible interactions. Consequently a quadratic model with interactions was postulated: Y = bo + ∑ bi X i + ∑ bij X i X j + ∑ bii X i2 i

i, j≤ i

(1)

i

where Xi is the varying parameter and bi, bij the coefficients of the model. Among all the possible optimal designs, which are available for analysing a response with a quadratic model, one of the most classical and efficient DOEs is the so-called Central Composite Design [10]. Such a design can be inscribed in a cube or in a sphere. Although the central composite design in a sphere presents the advantage of respecting rotability, a central composite design in a cube can be a better choice because the experimental region of interest is a cube what avoids any problem of extrapolation out of a sphere. Therefore we selected here a central composite design in a cube for studying the 927

Biosorption

uptake of copper ions on A. ferrooxidans. The corresponding 32 trials are given in Table 1. The ECHIP® software has been used to build the DOE and analyse the experimental results Table 1. Experiments of the composite centred in a cube DOE and experimental response (Y) for copper fixation on A. ferrooxidans (mg Cu2+ / 100 mg bacterial dry weight) Exp. 1 2 3 4 5 6 7 8 9 10 11 121 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 1

φ = physiological state (% Fe3+/Fe2++Fe3+) 65 95 80 80 80 80 80 80 80 80 65 80 95 65 95 65 65 95 95 65 65 95 65 95 95 65 95

pH

ω = agitation rpm

3 3 1 5 3 3 3 3 3 3 5 3 5 1 5 1 5 1 5 1 5 1 5 1 5 1 1

100 100 100 100 100 100 50 150 100 100 150 100 150 150 50 50 50 50 150 150 150 150 50 50 50 50 150

T= temperature °C 17.5 17.5 17.5 17.5 5 30 17.5 17.5 17.5 17.5 30 17.5 5 5 30 30 5 5 30 30 5 5 30 30 5 5 30

[prot] = protein concentration (mg.L-1) 181.5 181.5 181.5 181.5 181.5 181.5 181.5 181.5 45 318 318 181.5 318 318 318 318 318 318 45 45 45 45 45 45 45 45 318

YCu mg / 100 mg dry weight 2.74 6.07 2.44 4.52 4.95 10.29 8.41 5.11 43.38 4.18 6.42 5.44 12.39 3.61 6.08 7.74 16.36 6.5 71.19 23.55 16.72 45.37 74.92 56.56 11.98 28.74 1.88

experiment repeated 5 times

3.

RESULTS The responses Y (fixed metal per dry weight, mg / 100 mg dry weight) are recapitulated in Table 1. The control experiments showed that copper ion concentration remained unchanged by the experimental procedure in absence of bacteria, meaning that the decrease in concentration was only due to the A. ferrooxidans biomass. As a first step, the coefficients of the postulated model are estimated by the way of a multi-linear regression, the so-called least-square method. The resulting model cannot be directly used for prediction but several steps of analysis have to be performed. First, the 928

Biosorption

model must be refined by suppressing the non-significant terms and then, the adequacy of the final model has to be checked from several statistical tests. Such an analysis prevents from wrong interpretations of experimental results. The model refinement and its checking are important enough to be described with some detail. 3.1 Pointing out the significance of effects and refining the response model The results of the multi-linear regression (calculation of the model coefficients) are given under either a graphic form (Fig. 1) or a table (not shown).

Figure 1. Pareto graph: visualisation of the absolute values of each effect sorted in descendant influence. The boundary of effect significance line (determined thanks to probability of significance P) shows the limit of effect significance Residual SD = 4.021626 Replicate SD = 3.714353 R squared = 0.974 Adjusted R squared = 0.959

The graph rapidly visualises the most important effect or the non-significant ones thanks to the boundary of effect significance. An effect is the change due to a given term in the response as the concerned variable(s) shift(s) from the low to the high limit(s). Significance is a probability assigned by a statistical test here used for checking if a coefficient has a zero value or not. A small probability means a significant result. To refine the model, all the non-significant effects (under the boundary in Fig. 1) must be suppressed. However, such an elimination cannot be performed by simultaneously suppressing all the terms having a significance greater than 0.05. It requires a stepwise procedure by suppressing the term having the highest significance, then recalculating the new model and the new significance of the remaining terms and so on. For the fixation of copper by A. ferrooxidans as a function of temperature, agitation, pH, protein concentration and physiological state, the model obtained after refining is given in next equation: 929

Biosorption

where the notation means the midrange. 3.2 Checking the adequacy and the validity of the model A first very simple information about the descriptive power was obtained by plotting the fitted response vs. the experimental response (Fig. 2a). 8

8

4

4

0

0

20 40 60 Experimental value

80

Residuals

a.

b.

0

-4

-4

-8

-8 0

10

20 Run order

30

Figure 2. Plots testing adequacy and validity of the model: fitted values vs. experimental values (a) and residuals vs. normal scale (b)

For the ideal case this graph should be the first bisector. A too large dispersion between predicted and experimental response values would be the sign of a poor fit due either to an unsatisfactory model or to an important experimental error. Moreover, two global statistics are available to check the description power of the model fitting. The first one is the classical R squared statistic which ranges from 0 to 1 and is nothing but the proportion of response variations explained by the model. This statistic is a global index of the descriptive power of the fitting but it is better to use a more realistic coefficient, the adjusted R squared, defined from mean squares instead of raw squares and connected to the R squared statistic by: 2 = 1 − (1 − R 2 ) R adj

n −1 n−p

(3)

where n is the number of experiment and p the number of coefficient. The adjusted R squared has a value greater than 0.9 for a rather good fit while it can be even negative when the fit is very poor. In the present work for copper ion uptake, a 2 =0.954 was obtained for the refined model. very satisfactory value R adj The second test to check model adequacy consisted in verifying that there was no lack of fit. It is based on a comparison of the standard deviation of the residuals (residual SD, Fig. 1) and of the standard deviation of the replicates (replicate SD, Fig 1) which must have the same magnitude. Residuals are, in fact, the differences between the experimental response and the response given by the model. The classical F- test uses a statistic based on the ratio r = [residual SD/replicate SD]2 which should be close enough to unity and in 930

Biosorption

practice lower than a critical value deduced by the F distribution. In this study the value of the replicated and residual standard deviation were 3.72 and 4.02 respectively and the Ftest allowed to conclude that there was no evidence of lack of fit at the usual significance level 5%. Last step before accepting the model is adequate and is an accurate analysis of the behaviour of the residuals. This last analysis is very important because it points out if the model was only descriptive or actually predictive and highlights the possible anomalies of the experimental investigation. The multi-linear regression which allows to determine the model coefficients requires the respect of a basic assumption: the experimental error must obey to a normal law with a mean equal to zero and a constant standard deviation. Consequently the normality of residuals, which would be nothing but the experimental errors if the postulated model was the right one, and the independence of the residuals, both on the value of the response and on the order of runs, must be checked. The normality of residuals is checked by using the classical normality plot consisting in plotting the observations under study (the residuals in this case) against a normal scale and expecting a straight line when the observations are normally distributed (Fig. 2b) [21]. When the graph is not linear, response transformations, such as logY, lnY or Y1/2, can be tried for stabilising the variance of the response. When some points appear to be far from the line, the corresponding runs are suspicious and should be replicated for concluding if there was an experimental mistake or if the model was too rigid to describe the phenomenon and has to be changed for a more complex one. For copper fixation on A. ferrooxidans, the curve is a straight line and confirms the normality of the residuals. The independence of residuals vs. the response value is checked by simply plotting the residuals vs. the observations (Fig. 3a). 80

b.

2.5

60 50

Residuals

Fitted response

5

a.

70

40

0

-2.5

30 20

-5

10 0

-7.5 0

10 20 30 40 50 60 70 80 Experimental response

-3

-2

-1

0

1

2

Normal

Figure 3. Plots for checking the independence of the residuals as regards the experimental value (a) and the run order (b)

The residuals are equally distributed around the zero value without any dependence on the response value. If not, all the zones of the region under study are not equivalent because of a varying experimental error or a variable descriptive power of the model. Finally the independence of the residuals vs. the run order of the experiments allows to verify that no evolution of the experimental system occurs in time (Fig. 3b). Here again the residuals are dispersed at random around zero without any dependence on the run order. 931

Biosorption

It is only after all these checking steps that the model can be accepted as correct for describing, predicting and optimising the response within the experimental region under study.

copper uptake (mg / 100 mg dry weight)

3.3 Optimisation Once the suitable response model has been accepted, useful visual information can be derived from it by using its graphical representation vs. the continuous factors: 2D contour-plots and 3D response surfaces (Fig. 4).

60

40

20

100 ein ot pr

150

n tio tra en nc co g/l) (m

200 300 50

100 d spee tio n agita rp m) (

Figure 4. Response surface as a function of protein concentration and agitation speed at pH = 5, T = 30°C, φ = 65%, optimal values

Moreover, it is possible to evaluate the optimal response in the studied field from the model and to locate it on the drawing in order to verify its robustness against possible variations of the controlled factors. A robust optimum is characterised by a rather smooth response surface around it and is always recommended for a process because of the ease to keep the response close to the optimum even if the controlled parameters vary a little. For the present example by using the refined model, the predicted maximal mass of copper fixed by 100 mg of dry weight of bacteria should be 74.94 mg and it is expected at pH 5 and 30°C, for a 50 rpm stirring speed, with a physiological state corresponding to the middle of the exponential phase (65%), and a 45 mg/L protein concentration (Fig. 4). Under these conditions, experiment leads to 74.91 mg per 100 mg dry weight, value which is in very good agreement with the prediction. This optimum is located on the border of the studied field. Generally, in such a case, the experimental region under study should be translated and enlarged so implying to perform a new DOE. However, for the present study, such a further investigation is physically unfeasible because the variation ranges of the parameters were as wide as possible (see Material and Methods). Analysing the response surface for the mass of copper fixed by 100 mg dry weight of bacteria provides useful information. First, a slight increase of the protein concentration leads to a sharp decrease of the response. This observation means that the protein concentration must be carefully controlled to keep up the response at its optimal value. The same conclusions can be derived for pH and temperature. On the other hand, stirring and physiological state appear to have a limited effect on fixation in the studied range so allowing a less severe control on these parameters. 932

Biosorption

4.

DISCUSSION The empirical model obtained thanks to the DOE allowed to know the influence of each of the 5 chosen parameters on copper ion uptake by a living biomass of A. ferrooxidans. Moreover, non-linearity or parameter interactions have been evaluated by this method, what is impossible with the common OVAT method or the Langmuir and Freundlich isotherms. Consequently, the interaction study is original compared to previous works where each parameter was analysed individually. The most influent parameter (seen on Pareto graph, Fig. 1) for copper ion uptake by A. ferrooxidans was the protein concentration. Moreover, the importance of the quadratic effect of protein concentration testified of the great non-linearity of its influence on copper ion uptake. The model showed that an increase in protein concentration led to a decrease in specific copper ions fixation. This phenomenon, already observed by Sakaguchi et al [22], was probably due to the formation of aggregates of bacteria at high concentration. These aggregates decreased the contact surface between the bacteria and the copper ions and, consequently, the metal fixation. Moreover, these aggregates have been observed for A. ferrooxidans [3] confirming this hypothesis. The second most influent parameter, the temperature, had different influences on metal fixation on living biomasses: either it increased with temperature with an optimal temperature close to the culture's one [17], or it decreased reaching the optimum of fixation at 5 °C [7]. Copper ion fixation on A. ferrooxidans increased with temperature with an optimal value at 30°C, growth temperature of the bacterium. This indicated that the metabolism had to be active for metal ion uptake. So copper ion fixation was a more complex process than simple surface biosorption, modelled by Langmuir and Freundlich isotherms. The third influent parameter was the pH. Most of the metal ion fixation on microorganisms was maximal for pH between 4 and 6 [8, 17, 20] more rarely close to 2. Copper ion uptake by A. ferrooxidans increased with pH towards an optimal value at 5. This value, different to pH growth medium (1.4), was close to pH copper precipitation. The high specific copper uptake capacity of the biomass (74.9 mg / 100 mg) had to be compared with previous results obtained for Cr6+ [4] and Cd2+ [3]: 50.9 mg and 31 mg / 100 mg. For these metal ions, a bioprecipitation has been observed on the surface (Baillet et al. 1998) of the bacteria. Such a mechanism could explain an optimal pH at 5 where precipitates dissolution was less rapid: although no pH variation was observed in the solution, an active metabolism of the bacteria could modified locally the pH on the bacterial surface causing copper ion bioprecipitation. Physiological state is also an influent parameter. The age of the culture was shown to be important in metal fixation [3,17,19]. The copper ion fixation was maximum in the middle of the exponential growth phase indicating that either the metabolism or the external surface were implied in copper ion fixation. The middle of exponential growth phase corresponded to the more intensive metabolism. Although the external surface of the bacteria was modified during growth (aspect, permeability or size...) which might modify the fixation process. Finally, the less influent parameter was the agitation speed. The fixation increased when the agitation speed decreased: the uptake was better without strong turbulence. This result was in agreement with the hypothesis of bioprecipitation: a too high agitation speed would be able to separate precipitates from bacteria. 933

Biosorption

The influence of each of the 5 chosen parameters was known thanks to the DOE modelling. This technique allowed us to understand the effect of physicochemical parameters (pH, agitation speed, temperature), but also the effect of biological ones (physiological state, protein concentration). The results revealed a more complex process of fixation than the simple biosorption: the copper ion uptake by A. ferrooxidans need an active metabolism to reach its optimal value: 74.9 mg / 100 mg dry weight. REFERENCES

1. G.M. Gadd, In: Rhem HJ, Reed G. (eds), Biotechnology, vol 6b, Verlag Chemie, Weinheim, 1988. 2. Α. Boyer, J.-P. Magnin, P. Ozil, Biotechnol. Lett., 20 (1998) 187. 3. F. Baillet, J.-P. Magnin, A. Cheruy, P. Ozil, Environ. Technol., 18 (1997) 631. 4. F. Baillet, J.-P. Magnin, A. Cheruy, P. Ozil, Biotechnol. Lett., 20 (1998) 99. 5. A.A. Dispirito, J.W. Talnagi, O.H. Tuovinen, Arch. Microbiol., 135 (1983) 250. 6. F.D. Pooley, Nature, 296 (1982) 642. 7. N.D. Mullen, D.C. Wolf, F.G. Ferris, T.J. Beveridge, C.A. Flemming, G.W. Bailey, Appl. Environ. Microbiol., 55 (1989) 3143. 8. Β. Mattuschka and G. Staube, J. Chem. Tech. Biotech., 58 (1993) 57. 9. Β. Wehrheim and M. Wettern, Appl. Microbiol. Biotechnol., 41 (1994) 725. 10. G.E.P. Box, W.G. Hunter, J.S. Hunter, Statistics for experimenters, John Wiley, New York, 1978. 11. G.E.P. Box and N. Draper, Empirical model-building and response surfaces, John Wiley, New York, 1987. 12. A.C. Atkinson and A.N. Donev, Optimum experimental designs, University Press, New York, Oxford, 1992. 13. M.P. Silverman and D.G. Lundgren, J. Bacteriol. 77 (1959) 642. 14. M.K. Muir and T.N. Anderson, Metall. Trans. B, 8 (1977) 517. 15. O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, J. Biol. Chem., 193 (1951) 267. 16. D.P. Kelly and C.A. Jones, In: Murr LE, Tama AE, Brierley JA (eds), Metallurgical application of bacterial leaching and related microbiological phenomena, Academic Press, New-York, San Francisco, London, 1978. 17. D.K. Sahoo, R.N. Kar, R.P. Das, Biores. Technol., 41 (1992) 177. 18. J.A. Scott and S.J. Palmer, Appl. Microbiol. Biotechnol., 33 (1990) 221. 19. D. Cotoras, M. Millar, P. Viedma, J. Pimentel, A. Mestrie, World J. Microbiol. Biotechnol., 8 (1992) 319. 20. Z. Aksu, T. Kutsal, S Gün, N. Haciosmanoglu, M. Gholaminejad, Environ. Technol., 12 (1991) 915. 21. D. Cuthbert, Technometrics, 1-4 (1959) 311. 22. T. Sakaguchi, T. Tsuji, A. Nakajima, Z. Horikoshi, Eur. J. Appl. Microbiol. Biotechnol., 8 (1979) 207.

934

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Platinum and palladium recovery from dilute acidic solutions using sulfate reducing bacteria and chitosan derivative materials P. Chassarya, I. de Vargas Parodyb, M. Ruizc, L. Macaskieb, A. Sastrec, E. Guibala∗ a

Ecole des Mines d’Alès, Laboratoire Génie de l’Environnement Industriel, 6 avenue de Clavières, F-30319 Alès cedex, France b School of Biosciences, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, United Kingdom c Universitat Politècnica de Catalunya, Department of Chemical Engineering, E.T.S.E.I.B., Diagonal 647, E-08028 Barcelona, Spain Abstract The present work focuses on the study of Pd and Pt sorption using different strains of sulfate reducing bacteria, SRB, (Desulfovibrio spp.) and the comparison of their sorption performance to that obtained using several chitosan derivatives. D. desulfuricans was the best strain: sorption capacities as high as 190 and 90 mg Metal (Me) g-1 for Pd and Pt respectively were obtained under optimum experimental conditions at pH 3. At pH 2, for Pd sorption the maximum sorption capacities with HNO3 and H2SO4 solutions were comparable (close to 120 mg Pd g-1) and higher than the levels reached with HCl solutions (less than 70 mg Pd g-1). Sorption kinetics are very fast: the equilibrium can be reached within the first 15-30 minutes of contact. For chitosan derivatives sorption capacities under selected experimental conditions (i.e. pH 2) appear to be significantly higher than with SRB between 200 and 600 mg Me g-1, depending on the metal and the derivative. However, compared to SRB systems the kinetics are significantly slower: Several hours of contact are necessary to reach the equilibrium.

Keywords: chitosan, sulfate reducing bacteria, platinum, palladium, isotherms, kinetics 1.

INTRODUCTION Platinum group metals (PGM) e.g. Pd, Pt and Rh are of interest due to their high value and catalytic properties. Demand for PGM is increasing due to their widespread and often obligatory utilization in automotive catalytic converters. Currently the main large-scale industrial recovery processes are hydro- or pyro-metallurgical but these conventional technologies also generate liquors containing residual precious metals [1-2]. The use of

∗ Authors would like to thank the European Community for financial support under Growth Program (3SPM project, Contract G1RD-CT2000-00300).

935

Biosorption

low-cost sorbents has been investigated for their potential to replace current costly methods of recovering heavy metals from solutions. Natural materials or waste products from industries can be economically employed. Thus research has evaluated a variety of sorbents e.g. lignin, tannin-rich materials, chitosan, microbial biomass, algae, alginate, seaweed or zeolites [3-4]. Metal sorbing biomasses can have a real potential as selective, competitive and cost-efficient sorbents for precious metal recovery [5-7]. An alternative approach is the use of enzymatic processes. Sulfate-reducing bacteria (SRB) are known to have a broad reducing metal activity coupled to hydrogenases or/and cytochromes [8-10]. This has been demonstrated in the case of reduction of Pd(II) to Pd(0) but the best reductive accumulation only occurred after a prior biosorption step which suggests a high metal uptake as a pre-requisite to enzymatic catalysis, especially in the case of Desulfovibrio desulfuricans [11-12]. The present work focuses on the optimization of sorption procedures and selection of the best strain for Pd and Pt recovery and the sorption properties will be compared to those obtained with chitosan derivatives. Indeed chitosan, which is an aminopolysaccharide produced from crustacean shell, is very efficient at removing metal ions from dilute solutions [13-17]. Sorption may involve different mechanisms including chelation (for metal cations in near-neutral solutions) or ion exchange (for metal anions in acidic solutions). The sorption properties may be enhanced by chemical modification, including amine grafting (Poly(ethyleneimine), PEI) or thiourea grafting. Sorption properties will be characterized for both SRB and chitosan derivative mediated processes by investigating the influence of the pH on sorption efficiency, and performing sorption isotherms and kinetics. 2.

MATERIALS AND METHODS

2.1 SRB, growth conditions and harvesting Desulfovibrio desulfuricans (NCIMB 8307, D.d.), Desulfovibrio vulgaris (NCIMB 8303, D.v.) and Desulfovibrio fructosivorans (DSM 2604, D.f.) were grown for 2 days anaerobically (inoculation 10% vol/vol from 24h pre-culture) in sealed bottles under a N2 atmosphere, as batch cultures in Postgate C medium at 30°C (D.d. and D.v.) or 37°C (D.f.) [18]. Harvesting was carried out by centrifugation and resting cells were obtained by centrifuging 1 L of culture and washing the cells 3 times in air with 20 mM MOPS-NaOH buffer (pH 7). Experiments were also performed with dead biomass (harvested cells were autoclaved at 120°C for 20 minutes); sorption capacities were comparable to those obtained with living biomass. Dry weight was determined by filtration and drying to constant weight. 2.2 Sorption procedures for SRB system Pd and Pt solutions were prepared in distilled water from Na2PdCl4 and Na2PtCl6 salts, respectively, using HCl, HNO3 and H2SO4 for pH control. For sorption isotherms, known volumes of metal solutions (V, usually 10 mL) at variable initial metal concentrations (Co, 5-50 mg.L-1) in sterilized and sealed bottles were put in contact with a fixed amount of sorbent (m, 1.5 mg, i.e. sorbent dosage: 150 mg.L-1) under agitation (150 rpm) at 30°C. The pH was maintained constant. Samples were collected by filtration after 4 days of contact and analyzed using the SnCl2/HCl spectrophotometric method for Pt/Pd determination (Ceq) [19]. The mass balance equation was used for determining the sorption capacity (q, metal concentration on the biomass, mg.g-1 or mmol.g-1): q = V/m (Co-Ceq). The influence of chloride concentration was tested on biosorption performance at pH 2, using D. desulfuricans with addition of increasing amounts of NaCl. Sorption kinetics 936

Biosorption

were carried out at a metal concentration of 50 mg metal L-1 under agitation (150 rpm) at 30°C and pH 2 (controlled with either HNO3 or HCl) in the absence or the presence of NaCl (1 M) using D. desulfuricans. Kinetic experiments were also carried out using the same procedure with solutions containing 25 and 10 mg metal L-1. Samples were regularly collected by filtration and analyzed. 2.3 Preparation of chitosan derivatives Chitosan was supplied by Aber Technologies (France) as a flaked material with a deacetylation percentage of about 87% and a molecular weight of 125,000 g.mol-1. Chitosan gel beads were prepared using a neutralization process consisting of a two-step procedure [20]: (a) the dissolving of chitosan (4-5% w/w) in acetic acid (excess of acetic acid to completely dissolve the polymer), (b) the distribution of the viscous solution (filtered and de-bubbled) through a thin nozzle into a neutralization bath (NaOH 2M). The beads were collected after 16 hours of contact with the neutralization bath and then rinsed with several demineralized water baths up to constant pH. The size of the beads was 2.5 mm ± 0.2 mm and the water content was approximately 95%. Chemical cross-linking of chitosan was performed by reacting 5 g of chitosan beads (dry mass) with 1.5 g of glutaraldehyde dissolved in 100 mL of demineralized water, for 24 h. For the synthesis of poly(ethyleneimine) grafted chitosan beads (PEI-GA), PEI was dissolved in dimethylacetamide (DMA). The beads were also mixed with DMA in order to exchange water with DMA and then favor the diffusion of PEI in the beads: 5 g of chitosan beads were added to 5 g (the amount was varied for some experiments using alternatively 3 g or 5g of PEI) of PEI dissolved in 100 mL of DMA. The mixture was reacted for 24 h. The chitosan beads were then rinsed with 30 mL of DMA, and reacted for 6 h with 1.5 g of glutaraldehyde dissolved in 50 mL of DMA: glutaraldehyde is thought to establish new linkages between the amine groups of the biopolymer and some amine groups of PEI. For the synthesis of thiourea derivative chitosan beads (PEI-ThGA), the standard procedure consists of the addition of (a) 5 g of PEI (b) 1.5 g of thiourea and (c) 5 g of chitosan to 100 mL of DMA. The components were pre-reacted for 24 h and rinsed with 30 mL of DMA before the addition of 1.5 g of glutaraldehyde dissolved in 50 mL of DMA. Following the treatments, the sorbents were extensively rinsed with demineralized water. Because of the instability of previous derivative containing thiourea, the grafting of sulfur groups was carried out according to the procedure described by Cardenas et al. (2001) with chitosan flakes. This technique uses epichlorohydrin as crosslinking agent. Chitosan (20 g) was contacted with 1 L of acetic acid/acetate buffer (pH 4.6); 40 mL of epichlorohydrin dissolved in 1 L of acetone were added, and the slurry mixed at 35°C for 36 h. Thiourea (20 g) was then added and the stirring continued for 6 h at 60°C. After that period, 46 g of thiourea were added and the system was stirred for 12 h at the same temperature. Then, 500 mL of NaOH (1 M) were added and the slurry was agitated for 4 h at 60°C. The solid product obtained (TDC) was filtered and successively washed with acetone, demineralized water and methanol. Finally, it was dried in an oven at 60°C. 2.4 Sorption procedures for chitosan derivatives system The pH of the solutions was controlled using HCl and NaOH concentrated solutions. For sorption isotherms, known volumes of metal ion solution (150 mL) at fixed concentrations were contacted with varying sorbent quantities (from 3 to 35 mg, wet mass) at room temperature (20°C). After 3 days of agitation, the solutions were filtered through 1.2 µm membranes and the filtrates were analyzed using ICP-AES (Jobin-Yvon JY36, France). For the study of sorption kinetics, one liter of metal ion solution at fixed 937

Biosorption

pH was mixed with a fixed amount (100 mg) of sorbent in a jar-test agitated system. Fivemilliliter samples were withdrawn at specified times, filtered through a 1.2 µm membrane and analyzed as previously specified. For the study of the influence of competitor anions, chloride ions were added in the form of NaCl. The isotherms were modeled using the Langmuir equation: q=

q m b C eq 1 + b C eq

where qm (mmol.g-1) and b (L.mmol-1) are the constants of the model. 3.

RESULTS AND DISCUSSION

3.1 Selection of SRB strain and influence of the acid used for pH control Figure 1 compares the sorption capacities at pH 2 for Pt and Pd for 3 different strains of Desulfovibrio using different acids for pH control. Maximum sorption capacities were obtained with D. desulfuricans, regardless of the acid used for pH control and metal. Comparing the sorption capacities for Pt and Pd on the basis of molar units, it appeared that the biomasses have a greater affinity for Pd than for Pt, regardless of the strain. Palladium sorption was higher in sulfuric and nitric acid solutions than in hydrochloric acid media. Bacterial strains have a greater affinity for Pt in nitric and sulfuric acid solutions. These differences may be explained by differences in the composition of cell walls or/and by the effect of metal speciation. Indeed, PGM are very sensitive to the composition of the solution with respect to the formation of complexes (especially chloro-complexes), which in turn may affect the sorption mechanism (chelation versus ion-exchange) and the affinity of metal species for sorption sites. Table 1 reports the parameters of the Langmuir equation that was used for the modeling of sorption isotherms in Figure 1. Table 1. Influence of the acid used for pH control at pH 2 on the Langmuir parameters for the modeling of Pd and Pt sorption isotherms using D. desulfuricans (D.d.), D. fructosivorans (D.f.) and D. vulgaris (D.v.) Acid HNO3 HNO3 HNO3 HCl HCl HCl H2SO4 H2SO4 H2SO4

938

Bacteria D.d. D.f. D.v. D.d. D.f. D.v. D.d. D.f. D.v.

qm 0.343 0.185 0.163 0.287 0.116 0.195 0.228 0.118 0.118

Pt b 68.5 78.0 88.5 1800 126.3 76.9 25.6 28.5 52.9

2

R 0.996 0.980 0.900 0.986 0.970 0.944 0.985 0.593 0.928

qm 1.24 1.59 1.09 0.682 0.583 0.562 1.27 1.17 1.77

Pd b 73.1 5.85 44.5 80.4 10.1 93.2 62.3 21.1 36.8

R2 0.997 0.816 0.998 0.999 0.940 0.998 0.998 0.972 0.995

Biosorption

D.d.

0.1

D.f.

0.05

H2 SO4

0 0

q (mmol Pt/g)

D.f. D.v.

0.2

D.d.

0.5 H2 SO4

0.1

0

0.1 0.2 Ceq (mmol Pt/L)

D.v.

0.1 0.2 0.3 Ceq (mmol Pd/L)

0.4

0.8 D.d.

0.4 HNO3

0 0

D.f.

1.2

0 0.3

0

0.1 0.2 0.3 Ceq (mmol Pd/L)

D.f. D.v.

0.4

0.8 D.d.

0.3

D.f.

0.2

D.v.

0.1

HCl

0 0

0.1 0.2 Ceq (mmol Pt/L)

q (mmol Pd/g)

0.4 q (mmol Pt/g)

1

0.3

D.d.

HNO3

1.5

0

0.1 0.2 Ceq (mmol Pt/L)

0.4 0.3

D.v.

q (mmol Pd/g)

0.15

q (mmol Pd/g)

q (mmol Pt/g)

0.2

0.6 0.4

D.d.

0.2

HCl

D.v.

0 0.3

D.f.

0

0.1 0.2 0.3 Ceq (mmol Pd/L)

0.4

Figure 1. Influence of the acid used for pH control and SRB strain on Pt and Pd sorption isotherms (at pH 2) (q: biomass total sorption capacity)

3.2 Comparison of chitosan derivatives for Pt and Pd sorption at pH 2 Figure 2 compares Pt and Pd sorption properties at pH 2 (controlled with HCl, which is the best medium for PGM sorption, as shown in previous work) for different chitosan derivatives. Experiments were performed at low metal concentration. It is interesting to observe that sorption capacities as high as 3 mmol Me g-1 were obtained and that sorption isotherms were characterized by a pseudo-rectangular shape: the saturation plateau was obtained at residual concentrations as low as 0.05 mmol Me L-1 (especially for PEI and TDC). The Langmuir equation failed to fit experimental data at low Pt concentration in the case of TDC.

939

Biosorption

4

3 2

GA PEI

1

TDC

q (mmol Pd/g)

q (mmol Pt/g)

4

0

3 2

GA PEI TDC

1 0

0

0.05

0.1

0.15

0.2

0

Ceq (mmol Pt/L)

0.05 0.1 0.15 Ceq (mmol Pd/L)

0.2

Figure 2. Pt and Pd sorption isotherms at pH 2 for GA, PEI and TDC chitosan derivatives

0.5 0.4 0.3 0.2 0.1 0

2

HNO3 HCl H2SO4

D.d.

q (mmol Pd/g)

q (mmol Pt/g)

3.3 Influence of pH on sorption performance Figure 3 shows the influence of pH on Pt and Pd sorption by D. desulfuricans (selected for its highest affinity for PGMs) with HNO3, HCl and H2SO4 solutions. As expected increasing the pH resulted in an increase in sorption capacities due to the weaker competitor effect of counter anions brought to the solution by the acid used for pH control. The differences between the different acidic solutions decreased at increasing the pH since the amount of acid used for pH control (and consequently the addition of counter anions) also decreased. Sorption capacities were again lower in HCl solutions. This result contrasts with other experiments performed with ion-exchange resins (chitosan sorbents) for which the control of the pH with HCl enhances the formation of chloro-complexes, readily adsorbable on protonated amine groups. It may indicate that the sorption mechanism is different to that observed with chitosan material. These results confirm the greater affinity of the sorbent for Pd than for Pt.

1.5 1 0.5

HNO3 HCl H2SO4 D.d.

0 0 0.5 1 1.5 2 2.5 3 3.5 pH

0 0.5 1 1.5 2 2.5 3 3.5 pH

Figure 3. Influence of pH and acid used for pH control on Pt (left figure) and Pd (right figure) sorption capacity (Co(Pd/pt): 50 mg Metal L-1; Sorbent dosage: 150 mg L-1)

The influence of the pH on the sorption of Pt and Pd was also investigated. The PEI and TDC (glutaraldehyde cross-linked material) was significantly less efficient at sorbing PGMs than the other derivatives, see above). Table 2 shows the parameters of the Langmuir equation that were used for the modeling of sorption isotherms in Figure 4. In most cases, except for TDC and platinum at low metal concentration, the Langmuir equation fitted well experimental data. 940

Biosorption

Table 2. Influence of the pH (controlled with HCl) on the Langmuir parameters for the modeling of Pd and Pt sorption isotherms using GA, PEI and DTC derivatives of chitosan Sorbent

1 2 3 4 1 2 3 2

qm 1.11 3.69 1.94 2.65 2.21 2.86 2.90 2.28

PEI PEI PEI PEI TDC TDC TDC GA

Pt b 27.4 33.25 2725 1130 196.9 161.6 100.8 52.85

q (mmol Pt/g)

4

pH 1

PEI

3

pH 2

2

pH 3

1

pH 4

0

R2 0.984 0.980 0.983 0.992 0.988 0.940 0.801 0.978

Pd b 37.85 408.8 114.7 74.3 295.5 146.8 194.4 9.48

pH 1

PEI

3

R2 0.991 0.995 0.992 0.955 0.997 0.975 0.996 0.871

pH 2

2

pH 3

1

pH 4

0

0

0.05 0.1 0.15 Ceq (mmol Pt/L)

0.2

0

0.05 0.1 0.15 Ceq (mmol Pd/L)

0.2

4

2 pH 1

1

TDC

pH 2 pH 3

0

q (mmol Pd/g)

3 q (mmol Pt/g)

qm 1.64 3.28 2.68 2.20 2.54 3.31 3.00 2.91

4 q (mmol Pd/g)

pH

3 2

pH 1

1

TDC

pH 2 pH 3

0 0

0.025 0.05 0.075 Ceq (mmol Pt/L)

0.1

0

0.05 0.1 0.15 Ceq (mmol Pd/L)

0.2

Figure 4. Influence of pH on Pt and Pd sorption isotherms using PEI and TDC

Maximum sorption capacities were obtained at pH 2 for the PEI derivative of chitosan, with a significant decrease at pH 1 and pH 3. At pH 4, the results should be considered with caution since precipitation may occur. On the other hand, with TDC the sorption isotherms were influenced negligibly by the pH (in the range pH 1-3): the difference in maximum sorption capacities did not exceed 10-20%. This may be explained by the different mechanisms involved in metal uptake for PEI and DTC. For PEI, sorption occurs through anion-exchange on protonated amine groups; this mechanism is very sensitive to the pH of the solution, and to the presence of competitor anions. For TDC, sorption may be the combination of anion-exchange (on protonated amine groups) and chelation (on the thio groups). Chelation on sulfur compounds is less sensitive to pH and to the presence of competitor anions. This may explain the high efficiency of TDC for Pt 941

Biosorption

and Pd sorption in a broader range of pH compared to the PEI derivative, although the sorption capacities were slightly higher for PEI than for TDC at the optimum pH (i.e. pH 2). The increase in the density of sorption sites (by grafting of supplementary amine functions) may explain this improvement in sorption properties.

1 0.9 0.8 0.7 0.6 0.5

1 C(t)/Co

C(t)/Co

3.4 Sorption kinetics The kinetics is also an important parameter in the design of a sorption process. Figure 5 shows Pt and Pd sorption kinetics using D. desulfuricans at pH 2 controlled with nitric acid and hydrochloric acid (with a superimposition of kinetic curve performed in HCl media in the presence of NaCl). Sorption kinetics was little influenced by the acid used for pH control. In the case of Pd sorption, a greater contact time was necessary to reach the equilibrium. The presence of NaCl strongly decreased the efficiency of sorption, especially in the case of Pd.

HNO3 HCl HCl/NaCl

Pt

0.8 HNO3 HCl HCl/NaCl

0.6 0.4

0

30

60 90 Time (min)

120

0

30

Pd

60 90 Time (min)

120

Figure 5. Influence of the acid used for pH control on Pt and Pd sorption kinetics at pH 2 using D. desulfuricans (Co(Pt/Pd): 50 mg Metal L-1; Sorbent dosage: 150 mg L-1)

1

1

0.8

0.8

0.6 0.4

Co: 50 mg/L Co: 25 mg/L

0.2

Co: 10 mg/L

Pt

0 0

C(t)/Co

C(t)/Co

In most cases, more than 90% of total sorption was achieved within the first 10 minutes of contact. Sorption kinetics was significantly faster with SRB than for chitosan derivatives (see below). The initial concentrations did not significantly influenced sorption kinetics, regardless of metal (Figure 6). The same contact time (about 10 minutes) was sufficient to achieve the same percentage of metal uptake (90%). The sorption was restricted to the external surface of the cells. The uptake was almost instantaneous (no effect of intraparticle diffusion resistance).

0.6 0.4 0.2

Pd

0 30

60 90 Time (min)

120

Co: 50 mg/L Co: 25 mg/L Co: 10 mg/L

0

30

60 90 Time (min)

120

Figure 6. Influence of metal concentration on Pt and Pd sorption kinetics using D. desulfuricans at pH 2 controlled with HCl (relative concentration decay curves: C(t)/Co= f(t))

942

Biosorption

1 0.8 0.6 0.4 0.2 0

Pt

PEI TDC

0

1440

2880 4320 Time (min)

5760

C(t)/Co

C(t)/Co

In the case of chitosan derivatives the reaction required a greater contact time to reach equilibrium: for PEI more than 24 hours of contact were necessary to achieve the complete removal of Pd and Pt. For TDC material, at the same sorbent dosage, the equilibrium was reached within 6 hours of contact. It is difficult to compare the data since sorbent dosage (100 mg.L-1) was sufficient to completely remove Pt and Pd. As a consequence it is possible that the sorbents were not fully saturated, or at least that the relative saturations were different for PEI and TDC. It is not possible to conclude on the predominance of the resistance to intraparticle diffusion on the kinetic control. Other experiments have shown using similar sorbents that sorption occurred in the whole mass of the sorbent. It is thus anticipated that the diffusion of metal ions to internal sites controls the time required to achieve complete recovery of PGMs. It is important to observe that PEI was prepared with chitosan gel beads of expanded structure but with greater particle size, while TDC was obtained by chemical modification of flakes with lower porous network with small particle size. 1 0.8 0.6 0.4 0.2 0

Pd

GA PEI TDC

0

1440 2880 4320 Time (min)

5760

Figure 7. Pt and Pd sorption kinetics on PEI and TDC at pH 2 (controlled with HCl, Co (Pt/Pd): 20-25 mg Metal L-1; Sorbent dosage: 100 mg L-1) 3.5 Influence of chloride concentration on sorption performance One purpose of this work was to develop a metal recovery technology applicable to acidic leachates obtained from solid scrap. The main characteristics of leachates from spent automotive catalysts and other wastewaters are a high concentration of chlorides and low pH, since aqua regia is necessary for effective PGM leaching. Hence, the influence of chloride on Pd and Pt sorption is of major concern for the evaluation of process applicability. Figure 8 compares relative sorption capacities (compared to reference performance, i.e. without NaCl addition) for the different systems with increasing concentrations of chloride. The addition of chloride strongly reduced Pt sorption capacities especially for SRB system and for PEI: when chloride concentration exceeded 0.25-0.5 M Pt sorption capacities were negligible for PEI and SRB. In the case of Pd, with PEI derivative and SRB there was also a significant decrease in sorption capacity, however it was necessary to drastically increase NaCl concentration: with SRB the sorption decrease was about 70% at a chloride concentration of 1 M; with PEI, at a 0.5 M chloride concentration the decrease did not exceed 50%. In the case of TDC, the addition of chloride decreased sorption capacities but even with a chloride concentration as a high as 2 M, the decrease of sorption capacity did not exceed 40%, regardless of metal. The grafting of supplementary amine groups (PEI) increased the number of sorption sites, and then decreases the competitor effect of chloride ions.

943

1 0.8 0.6 0.4 0.2 0

Pt

D. desulfuricans

0

2 4 [NaCl] (M)

Pd

q/qo

q/qo

Biosorption

6

1 0.8 0.6 0.4 0.2 0

TDC-Pt TDC-Pd PEI-Pt PEI-Pd

0

0.5

1 1.5 [NaCl] (M)

2

Figure 8. Influence of chloride concentration on Pt and Pd sorption at pH 2 (controlled with HCl; for D. desulfuricans: Co(Pd/Pt): 50 mg Metal L-1; Sorbent dosage: 150 mg L-1; for chitosan derivatives: Co(Pd/Pt): 20 mg Metal L-1; Sorbent dosage: 40 mg L-1)

In the case of thiourea grafting, new chelating groups less sensitive to the pH and the competitor effect of chloride anions, limited the influence of chloride concentration. The TDC appears to be perfectly designed for the treatment of industrial solutions containing high concentrations of chloride ions and low pH, despite the flaked conditioning that limits the use of sorbent particles in fixed-bed systems. 4.

CONCLUSIONS Sulfate-reducing bacteria (especially D. desulfuricans), and chitosan derivatives (more specifically thiourea derivative of chitosan) appear promising sorbents for the recovery of PGMs from dilute acidic solutions (in the range pH 1-3) containing chloride anions (below 0.5-1 M). Both sorbents have a preference for Pd over Pt, though more marked in the case of SRB. The SBR system is characterized by very fast sorption kinetics (equilibrium reached within 10 minutes of contact), while chitosan derivatives are characterized by slower kinetics but very high sorption capacities (up to 3 mmol Me g-1). Experiments are currently performed on the sorption of Pt and Pd in multi-component systems and in the presence of other metals (typical industrial solutions). REFERENCES

1. J.E. Hoffman, J. Met., 40 (1988) 40. 2. C.S. Brooks, Metal Recovery from Industrial Wastes, Lewis Publishers, Chelsea, MI, 1991. 3. S.E. Bailey, T.J. Olin, R.M. Bricka and D.D. Adrian, Wat. Res., 33 (1999) 2469. 4. B. Volesky and Z.R. Holan, Biotechnol. Prog., 11 (1995) 235. 5. E. Remoudaki, M. Tsezos, A. Hatzikoyian and V. Karakoussis, in: Process Metallurgy 9B, R. Amils, and A. Ballester, eds., Elsevier, Amsterdam, pp.449-462, 1999. 6. N. Kuyucak, and B. Volesky, Biotechnol. Lett., 10 (1988) 137. 7. A.V. Pethkar, S.K. Kulkarni and K.M. Paknikar, Bioresource Technol., 80 (2001) 211. 8. D.R. Lovley and E.J.P. Philips, Appl. Environ. Microbiol., 54 (1992) 1472. 9. J.R. Lloyd, P. Yong, L.E. Macaskie, Appl. Environ. Microbiol., 64 (1998) 4607. 10. J.R. Lloyd, J. Ridey, T. Khizniak, N.N. Lyalikova and L.E. Macaskie, Appl. Environ. Microbiol., 65 (1999) 269. 11. P. Yong, J.P.G. Farr, I.R. Harris and L.E. Macaskie, Biotechnol. Letters, 24 (2002) 205. 944

Biosorption

12. P. Yong, N.A. Rowson, J.P. Farr, I.R. Harris and L.E. Macaskie, J. Chem. Technol. Biotechnol., 77 (2002) 593. 13. K. Inoue, in: Recent Advances in Marine Biotechnology, Volume 2, Environmental Marine Biotechnology, M. Fingerman, R. Nagabhushanam and M.-F. Thompson, eds., Oxford & IBH Publishing PVT. Ltd, New Delhi, pp. 63-97, 1998. 14. E. Guibal, T. Vincent, A. Larkin, and J.M. Tobin, Ind. Eng. Chem. Res., 38 (1999) 4011. 15. M. Ruiz, A. Sastre, and E. Guibal, React. Funct. Polym., 45 (2000) 155. 16. R. Bassi, S.O. Prasher and B.K. Simpson, Sep. Sci. Technol., 35 (2000) 547. 17. E. Guibal, T. Vincent, T., and R. Navarro Mendoza, J. Appl. Polym. Sci., 75 (2001) 119. 18. J.R. Postgate, in: The Sulphate Reducing Bacteria, Cambridge University Press, Cambridge, pp. 30-50, 1979. 19. G. Charlot, Dosages Absorptiométriques des Eléments Minéraux, Masson, Paris, 1978. 20. E. Guibal, C. Milot and J.M. Tobin, Ind. Eng. Chem. Res., 37 (1998) 1454. 21. G. Cardenas, P. Orlando, and T. Edelio, Int. J. Biol. Macromol., 28 (2001) 167.

945

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Preliminary study of lead sorption by selected sorbents M.E. Ly Arrascuea, J.L. Bauer-Cuyaa, F. Peirano Blondeta,b, J. Roussyb and E. Guibalb∗ a

Universidad Peruana Cayetano Heredia, Departamento Academico de Quimica, Av. Honorio Delgado, 430 Urbanizacion Ingenieria, Lima 31, Peru b Ecole des Mines d’Alès, Laboratoire Génie de l’Environnement Industriel, 6 avenue de Clavières, F-30319 Alès cedex, France

Abstract Lead sorption was investigated using several biosorbents, including five fungi (Aspergillus niger, Mucor miehei, Penicillium chrysogenum, Rhizopus arrhizus and Rhizopus conhii), two algae (Ascophyllum nodosum, Lessonia trabeculata), one yeast (Saccharomyces cerevisiae), alginate and chitosan. Experiments were performed with non-living microorganisms. In the present work alginate was tested as a powder. Preliminary experiments have been performed investigating the influence of the pH on lead sorption (between pH 3 and pH 6). Complementary experiments were carried out using 3 of the most efficient sorbents: alginate, A. niger and A. nodosum. Sorption properties were determined through isotherms and kinetics at pH 4. Maximum sorption capacities were comparable for A. niger and A. nodosum, tending to 130-150 mg Pb g-1. Sorption kinetics was slightly faster for A. nodosum than for A. niger.

Keywords: lead, biosorption, fungal biomass, algal biomass, isotherms, kinetics. 1.

INTRODUCTION The toxic effects of heavy metals such as mercury, cadmium or lead require the treatment of industrial wastewater prior their discharge to the environment. Though precipitation processes are efficient at removing metal ions from solutions, these techniques are sometimes inappropriate for the treatment of dilute solutions for technical or economical reasons. Moreover, the production of toxic sludge as a by-product limits the interest of these processes: the precipitation results in a single pollution transfer. For these reasons, economical and efficient processes are still necessary to develop. Biosorption processes have been recognized as promising techniques, owing to the low cost and the great diversity of these materials. Since the early 80’s many studies have focused on the use of materials of biological origin for the recovery of metal ions [1]. Living as well as

∗ Authors thank the French Ministry of Foreign Office and the Concytec for the financial support under the Franco-Peruvian Program of Collaboration Raul Porras Barrenechea. E.G. thanks the European Community for financial support under Growth Program (3SPM project, Contract G1RD-CT2000-00300) for attending the IBS’03 conference.

947

Biosorption

non-living microorganisms have been used for metal recovery: algae [2-5], bacteria [6-7], fungi [8-12], and yeasts [13]. The identification of sorption sites on the cell walls of the microorganisms has been the motive for the increasing interest for metal uptake properties of the biopolymers entering in the composition of these cell walls: alginate [14], chitin and chitosan [15-16], or exopolysaccharides excreted by these microorganisms [17]. The present study focuses on a preliminary investigation of lead sorption using 5 fungi (Aspergillus niger, Mucor miehei, Penicillium chrysogenum, Rhizopus arrhizus and Rhizopus conhii), 2 algae (Ascophyllum nodosum and Lessonia trabeculata), 1 yeast (Saccharomyces cerevisiae) and 2 biopolymers (alginate and chitosan). The cell wall of selected fungi is characterized by the presence of chitin and/or chitosan material associated with proteins and other glucans. On the other hand, algal biomass may contain alginate (L. trabeculata) or fucoidan polymers (sulfate fucans, A. nodosum) [18-20]. Preliminary experiments focus on the study of the influence of the pH on 10 selected sorbents. Sorption isotherms and uptake kinetics are determined at pH 4 using alginate, A. niger and A. nodosum. 2.

MATERIALS AND METHODS

2.1 Materials The samples of Aspergillus niger, Penicillium chrysogenum, Rhizopus arrhizus, Rhizopus conhii and Mucor miehei were kindly donated by Gist Brocades (The Netherlands). They were supplied after being inactivated and dried. The particles were sieved and the fraction 125-250 µm was used for preliminary experiments. The brown alga Ascophyllum nodosum was collected on the Brittany coast while Lessonia trabeculata was collected on the Peru coast. The samples were dried and ground, the fraction 125-250 µm was used for experiments. The yeast (Saccharomyces cerevisiae) was obtained from local bakeries, as a commercial sample. Alginate was purchased from Janssen. Chitosan was supplied by ABER-Technologies (Brest-France) (Lot N°A17G28). Its characteristics were pKa = 6.2, number average molecular mass, MWn = 125,000 g.mol-1, weight average molecular mass, MWw = 191,000 g.mol-1, and deacetylation percentage = 87%. Lead solutions were prepared by dilution of an atomic absorption standard solution (1 g Pb L-1) supplied by Fluka (Germany). The metal was under the form of a nitrate-salt. The pH of the solutions was controlled with molar solutions of either sulfuric acid or sodium hydroxide. The ionic strength of the solution was adjusted by adding sodium nitrate at the final concentration 0.05 M. 2.2 Sorption procedure For the study of the influence of the pH, 200 mL of solution containing 50 mg Pb L-1 was mixed with 40 mg of sorbent for 72 hours. This contact time was considered sufficient for reaching equilibrium. The pH was not controlled during sorption but its final value was measured. Samples were filtered on a filter membrane (pore size: 1.2 µm). The residual lead concentration in the filtrate was analyzed using an ICP-AES equipment (Jobin Yvon 2000, Jobin-Yvon, France). Sorption isotherms were carried out on alginate, A. niger and A. nodosum. A fixed volume of solution (V: 200 mL), which pH was controlled with H2SO4 and NaOH solutions at different initial concentrations (Co, mg Pb L-1: 9, 22 and 46) was mixed for 72 hours with increasing sorbent masses m (10-40 mg). After filtration, the samples were analyzed using ICP-AES equipment for the determination of the residual metal concentration, Ceq (mg Pb L-1) and the mass balance equation was used for the determination of lead concentration in the sorbent, q, mg Pb g-1, 948

Biosorption

q = V/m (Co – Ceq). Sorption kinetics was determined using 1 L of solution at the concentration 25 mg Pb L-1 and 200 mg of sorbent. Samples were regularly collected, filtered and analyzed for lead content by ICP-AES. 3.

RESULTS AND DISCUSSION

3.1 Influence of pH on lead sorption and selection of sorbents Experiments were carried out at pH 3, 4, 5 and 6 with low metal concentrations to avoid precipitation artifacts. This part of the study served to select the best sorbents among the ten sorbents that were tested. Table 1 shows the change in the pH of the solution, the sorption efficiency and the sorption capacity. Table 1. Influence of pH on sorption efficiency (SE, %) and sorption capacity (SC, mg Pb g-1) for selected sorbents – Final pH (Co: 50 mg L-1; Sorbent dosage: 200 mg.L1 , NaNO3: 0.05 M) Sorbent

A. niger M. miehei P. chrysogenum R. arrhizus R. conhii A. nodosum L. trabeculata S. cerevisiae Alginate Chitosan

pHf 3.3 3.1 3.1 4.5 3.2 3.2 3.1 3.3 3.2 5.2

pH 3 SE SC 44.4 103 22.2 52 4.3 10 18.8 44 3.7 9 41.8 97 11.4 26 6.5 15 88.5 205 8.1 19

pHf 4.8 3.9 3.7 4.8 4.6 5.0 4.0 4.8 4.2 5.5

pH 4 SE SC 58.0 129 10.2 23 3.2 7 16.0 36 34.0 76 48.6 108 12.0 27 13.6 30 55.2 123 10.1 22

pHf 6.0 4.4 4.2 5.1 4.8 5.4 4.9 5.1 5.1 5.7

pH 5 SE 43.1 39.5 6.2 19.4 45.9 59.7 13.7 18.2 56.0 17.7

SC 107 95 15 48 113 149 34 45 140 44

pHf 6.1 4.5 5.0 5.4 5.0 5.6 5.2 5.2 5.9 6.0

pH 6 SE SC 54.6 102 34.5 74 11.1 21 31.4 59 44.3 83 68.4 127 23.8 45 27.5 52 31.5 59 21.5 40

Table 1 shows pH variation for selected sorbents. The sorbents can be classified in 3 groups: (a) whose that systematically decreased the pH of the solution (P. chrysogenum; M. miehei); (b) whose that systematically increased the pH of the solution (chitosan, A. niger); and (c) whose that increased the pH when the initial pH was below 5-5.5 and decreased the pH for pH above pH 6 (R. arrhizus, R. conhii, A. nodosum, L. trabeculata, S. cerevisiae, alginate). Even in the same kind of sorbent (fungal, algae) the pH variations were not homogeneous in value and in trend, despite similar structures and compositions. In the case of industrial fungal biomass it may be explained by the treatments used for the inactivation of the microorganisms (thermal treatment versus chemical treatment) or by the presence of impurities (residues of flocculating and filtrating material). The greatest pH variations were observed with chitosan, especially at low initial pH. The pH significantly influenced the sorption efficiency. However, the trends and the extent of the variation depended on the sorbent. In most cases, increasing the pH increased the sorption efficiency. In the case of A. niger and R. conhii, the sorption efficiency was almost independent of the pH, while in the case of alginate, the sorption efficiency tended to decrease with decreasing the initial pH. The best sorbents were A. niger, A. nodosum, R. conhii, M. miehei, and alginate. The optimum equilibrium pH was in most cases greater than pH 5, except for alginate. Indeed, in the case of alginate the sorption was much more efficient at acidic pH. Increasing the pH resulted in a decrease of the competition of protons with lead for sorption on the sorbents, while a mild pH was much more favorable to the chelation of lead by the ligands present on the biomass, and on amine functions of 949

Biosorption

chitosan. The greatest sorption capacities were obtained at around pH 5-5.5 for A. niger, A. nodosum, R. conhii, with sorption capacities as high as 100-150 mg Pb g-1 (0.5-0.75 mmol Pb g-1). In the case of alginate the sorption exceeded 200 mg Pb g-1 (1 mmol Pb g-1) at pH 3-3.5, but decreased at increasing the equilibrium pH. A. niger, A. nodosum and alginate have been selected for more complete sorption studies. 3.2 Sorption isotherms Sorption isotherms have been carried out at pH 4 as the initial pH. The pH was controlled during the sorption. Figure 1 show lead sorption isotherms using A. niger and A. nodosum. The curves were modeled using the Langmuir (solid lines) and the Freundlich (dashed lines) equations:

q=

Langmuir equation:

q m b C eq

(1)

1 + b C eq

where q, qm are the sorption capacity and the maximum sorption capacity at monolayer coverage, respectively (mg Pb g-1), b (L.mg-1) is the sorption affinity (proportional to the initial slope of the sorption isotherm curve). q = k F C1/n eq

Freundlich equation:

(2)

160

160

120

120

q (mg Pb/g)

q (mg Pb/g)

where kF and n are the parameters of the Freundlich model. For both A. niger and A. nodosum, the best fit of experimental data was obtained with the Langmuir equation (Figure 1, and Table 2). Though the maximum sorption capacity and the affinity coefficient varied reciprocally for A. niger and A. nodosum, and the term qm b, (i.e. the initial slope of the sorption isotherm curve) was comparable for A. niger and A. nodosum, (i.e. 97.0 and 97.8 L g-1, respectively). These sorbents showed comparable sorption efficacy on the basis of equilibrium performance.

80 A. niger

40 0

80 A. nodosum

40 0

0

10

20 30 40 Ceq (mg Pb/L)

50

0

10

20 30 40 Ceq (mg Pb/L)

50

Figure 1. Lead sorption isotherms on A. niger and A. nodosum at pH 4 (points: experimental data; dashed lines: modeling with the Freundlich equation; solid lines: modeling with the Langmuir equation)

In the case of alginate it was impossible to fit the experimental data with the models. Indeed, a great dispersion of sorption capacities was observed with changing the mass of the sorbent and the initial metal concentration (Figure 2). At using low initial metal concentration, the sorption capacity was very low (below 25 mg Pb g-1), while increasing the initial metal concentration, significantly increased the sorption capacity. However, for intermediary initial metal concentration (Co: 22 mg Pb L-1), unexpectedly, the sorption capacity decreased with increasing the residual concentration. At the highest initial metal 950

Biosorption

concentration, again, the sorption capacity did not vary continuously with increasing the residual concentration. Table 2. Lead sorption isotherms – Parameters of Langmuir and Freundlich models. Sorbent

A. nodosum A. niger

Langmuir Model qm b R2 125.4 0.78 0.979 161.7 0.60 0.984

Freundlich Model k n R2 53.85 3.73 0.788 70.19 3.77 0.820

q (mg Pb/g)

400 300 Co: 9 mg Pb/L Co: 22 mg Pb/L Co: 46 mg/L

200 100 0 0

10 20 Ceq (mg Pb/L)

30

Figure 2. Lead sorption isotherm on alginate at pH 4 (both initial concentration Co and sorbent dosage were varied to get the distribution of the metal between the two phases)

The experimental data were also worked on with plotting the sorption capacity and the sorption efficiency versus the sorbent dosage for each initial metal concentration (Figure 3, left and right panels, respectively). At low initial metal concentration, the sorption capacity and the sorption efficiency were not controlled by the sorbent dosage. At medium initial metal concentration, as expected increasing sorbent dosage resulted in a decrease in sorption capacity, however, increasing the mass of sorbent surprisingly decreased sorption efficiency. For the highest initial metal concentration, sorption capacity was almost constant at low sorbent dosage but seriously decreased at the highest sorbent dosage, while sorption efficiency reached a maximum when the sorbent dosage was 150 mg alginate L-1. A first attempt has been made to explain this surprising behavior of alginate in relation with the dissolving of the biopolymer. The ability of alginate to sorb metal ions is well documented but in most cases research has focused on the use of preformed alginate bead (gelled with calcium chloride, for example) [21], a few studies also dealt on the direct coagulation of alginate in metal ion solutions [22]. It is possible to suggest that lead sorption occurs by a dual mechanism involving the chelation of lead on functional groups of the biopolymer and by a gelation of the biopolymer, lead acting as a cross-linking agent. This ionic cross-linking / ionotropic gelation mechanism has been described by Dambies et al. in the case of chitosan gelation using molybdate as the crosslinking agent [23]. When there was an excess of polymer compared with lead, the amount of metal was not high enough to cross-link alginate chains. The biopolymer dissolved in acidic media and it was not able to form a stable network. Lead cations may be chelated to dissolved polymer chains and then they cannot be removed at filtration and the sorption efficiency decreased. When the metal concentration increased, the cross-linking with lead ions was more efficient and the biopolymer gelled with a simultaneous immobilization of metal ions. 951

Biosorption

Sorption Efficiency (%)

q (mg Pb/g)

400 300 200 100 0 50

100 150 SD (mg Alginate/L)

200

100 80 Co: 9 mg Pb/L Co: 22 mg Pb/L Co: 46 mg Pb/L

60 40 20 0 50

100 150 SD (mg Alginate/L)

200

Figure 3. Influence of sorbent dosage on uptake capacity and sorption efficiency at different initial metal concentrations (pH 4) 3.3 Sorption kinetics Though the maximum sorption capacity and affinity are important parameters in the determination of the performance of a biosorption system, it is also necessary to take into account the uptake kinetics in the selection of the optimum sorbent. Figure 4 shows lead sorption kinetics for A. niger and A. nodosum, respectively. Equilibrium was reached after 4 hours of contact in the case of A. nodosum, while in the case of A. niger, more than 90% of the total sorption was reached within the first 4 hours of contact, the sorbent continued to sorb small amounts of lead even after 24 hours of contact. It may be due to different sorption mechanisms, to diffusion limitations or to differences in the location of metal sorption. Tsezos and Volesky [24] showed in the case of uranium sorption on fungal biomass that the uptake involved different mechanisms including chelation, but also local precipitation of metal ions (in the membrane). The occurrence of successive different sorption steps could explain a longer time to be required for reaching the equilibrium. The sorption can be restricted to the external layer of the sorbent or can occur in the whole membrane: in this case the time required to penetrate, diffuse and be adsorbed can be significantly increased. Uptake kinetics may be controlled by several mechanisms including the intrinsic sorption rate, but also resistance to diffusion. There are different steps in the mass transfer of the solute from the solution to the sorption sites: (a) bulk diffusion; (b) film mass transfer resistance; and (c) intraparticle mass transfer resistance. Providing a sufficient agitation (stirrer speed, reactor geometry) allows neglecting the bulk diffusion as the limiting step. Film diffusion and intraparticle diffusion are actually the main controlling steps. Though the separation of these two steps has no physical significance, it is possible to assume, as a simplification, that the preliminary stage (first minutes of contact) is controlled by film diffusion resistance while the later stage is manly controlled by intraparticle diffusion resistance. In a first approximation, the sorption kinetics (within the first minutes) may be modeled by a first-order kinetic equation [25]; and the kinetic parameter ka (min-1) can be obtained by: ⎡ C(t) ⎤ Ln ⎢ ⎥ = ka t C o ⎣ ⎦

952

(3)

Biosorption

Puranik et al. [26-27] proposed a more sophisticated and appropriate model for the description of uptake kinetics under the following assumptions: (a) the particles are assumed to be spherical (uniformity in shape and size); (b) the bulk concentration of the solute is homogeneous in the reactor (correct mixing); (c) the intraparticle diffusion is negligible (sorption located at the surface of the sorbent); (d) instantaneous sorption at the surface of the particle (fast intrinsic sorption rate); (e) limited volume of solution; and (f) isothermal sorption mechanism described by either a Langmuir or a Freundlich equation. They used several equations, including the total mass balance equation: (4) Ws q(t) + C(t) = Co . -1 -1 where Ws is the sorbent dosage (g L ), q is the sorption capacity (mg Pb g ), C(t) is the bulk lead concentration (mg Pb L-1), and Co the initial metal concentration (mg Pb L-1). The change in the bulk concentration is proportional to the driving force for the sorption at the surface of the particle:

d C(t) = - K m a (C(t) - C p (t) ) (5) dt where Km is the external film mass transfer coefficient (m.min-1), a is the specific surface area of the sorbent particles per unit volume of reactor (m2.m-3) and Cp is the concentration of the metal at the liquid/sorbent interface (mg Pb L-1). Combining the mass transfer resistance equation and the mass balance equation, they established the following equation: ⎛ dq(t) ⎞ Ws ⎜ ⎟ = K m a (C(t) - C P (t) ) ⎝ dt ⎠

(6)

The differentiation of the Langmuir equation applied for lead concentration at the interface (Cp) gave after simplification: ⎤ qm b dq(t) dC P (t) ⎡ = ⎢ ⎥ dt dt ⎣ (1 + b C P (t) )2 ⎦

(7)

The combination of this equation with the preceding equation gave after simplification:

[

dC P (t) Km a (C(t) - C P (t) )(1 + b C P (t) )2 = dt W qm b

]

(8)

Using the dimensionless variables (C*(t) = C(t)/Co; and C*p(t) = Cp(t)/Co), the change of metal concentration in the bulk solution with time might satisfy the system of first ordinary differential equations: d C * (t) = - K m a (C * (t) - C p * (t) ) dt

[

dC P * (t) Km a (C * (t) - C P * (t) )(1 + b C P * (t) )2 = dt W qm b

(9)

]

(10)

with the boundary conditions: C*(t=0) = 1 and Cp*(t=0) = 0 (11) ® This system has been solved using Mathematica package for the modeling of kinetics using the Langmuir coefficients from Table 2. The sum of the square of residuals has been minimized in order to get the optimum Km a. Assuming that the sorbent particles were not porous, and that their (wet) density ρ (kg m-3) was approximately 1050, the 953

Biosorption

specific surface area of the sorbent particles per unit volume of reactor, a, can be calculated according to: a=

6 SD ≈ 18 m -1 dP ρ

where SD is the sorbent dosage (g L-1)

(12)

Figure 4 (left panel) shows the modeling of experimental data for A. nodosum and A. niger, the kinetic parameters are given in Table 3. The sorption kinetics is quite well described by the modeling, especially in the first part of the curve. In the final stage, the residual concentration was slightly overestimated for A. nodosum. It may be due to an underestimation of the sorption capacity deduced from the Langmuir parameters. In the case of A. niger the modeling was not so good: the final concentration was dramatically underestimated and the initial part of the curve was not perfectly described as it was with A. nodosum. The worst fit of experimental data by this model may be explained by inaccurate modeling hypotheses, and especially an overestimation of the contribution of intraparticle diffusion. Crank proposed a model whereby diffusion is controlled only by intraparticle mass transfer for well-stirred solutions of limited volume (V), assuming the solute concentration to be always uniform (initially Co), the sorbent sphere to be free from solute, and the external diffusion to be negligible [28]. Under these conditions, the fractional approach to equilibrium (FATE) that is the total amount of solute Mt (mg g-1) in a spherical particle after time t (min), expressed as a fraction of the corresponding quantity after infinite time (M∞) is given by: FATE =

∞ ⎡ Dq 2n t ⎤ 6 α (α + 1) M(t) = 1− ∑ Exp ⎢− 2 ⎥ M(∞) (9 + 9α + q 2nα 2 ) i =1 ⎣ dP ⎦

(14)

where D is the intraparticle diffusion coefficient (m2.min-1). The fractional approach to equilibrium, FATE, may be used to estimate the intraparticle diffusion coefficient D, when the external diffusion coefficient is neglected. α is the effective volume ratio, expressed as a function of the equilibrium partition coefficient (solid/liquid concentration ratio) and is obtained by the ratio Ceq/Co-Ceq. qn represent the non-zero solutions of the equation: tan q n =

3 qn 3 + α q 2n

(15)

The infinite terms are summed until the summation does not vary. This equation was used to determine the overall intraparticle diffusivity which best fit experimental data minimizing the sum of the square of the differences between experimental results and calculated data. Tien pointed out that this equation is only justified when the sorption isotherm can be approximated by a linear equation [29]. However this simplified model will be sufficient to approach coefficient D. Figure 4 (right panel) shows that the equation fitted well experimental data, especially for A. nodosum. The simplified equation does not take into account the sorption capacity of the sorbent, the best fit of kinetic curves in the later stage of the sorption process was thus expectable. Intraparticle diffusion coefficients were of the same order of magnitude for A. nodosum and A. niger, close to 4x10-12 m2.min1 . The Biot number was very high (between 4500 and 15000), indicating that film resistance is negligible for these sorbents.

954

Biosorption

1

0.8

A. niger

0.8

0.6

A. nodosum

0.6

0.4

FATE

C(t)/Co

1

0.2

0

0 5

10 15 20 Time (h)

25 30

A. nodosum

0.4

0.2 0

A. niger

Sorbent dosage: 200 mg/L Co: 25 mg Pb/L

0

5

10 15 20 Time (h)

25 30

Figure 4. Lead sorption kinetics on A. niger and A. nodosum at pH 4 (curves: modeling with external diffusion resistance model (left) and intraparticle diffusion resistance model (right)) Table 3. Lead sorption kinetics – Parameters of the kinetic models (Kma, min-1; Km, m.min-1; D, m2.min-1) and Biot number (dimensionless) Sorbent

A. nod. A. niger

First-order equation & Film Resistance 3 Kma* 10 Km * 103 SSR 22.35 1.22 0.081 5.09 0.278 0.287

Crank equation (Intraparticle Diffusion) D * 1012 SSR 4.94 0.038 3.67 0.173

Biot number Bi= Km dp/D 15400 4700

SSR: Sum of Square Residuals: Σi=1,n[(f(ti)exp.-f(ti)calc.)]2, n is the number of experimental points, f(ti)exp. and f(ti)calc. are transforms of lead concentrations at time ti on experimental data and modeled data, respectively.

4.

CONCLUSIONS A. niger and A. nodosum were selected for their high sorption for lead recovery from acidic solutions (around pH 4) with sorption capacities as high as 0.6-0.7 mmol Pb g-1. Simple models were used for modeling sorption kinetics and evaluate the order of magnitude for mass transfer coefficients for external and intraparticle diffusion. The Biot numbers show that film diffusion can be considered as negligible. Alginate was also characterized by high sorption capacities (about 1.6-1.7 mmol Pb g-1) when using high lead concentrations: it allows the gelification of the biopolymer. At low metal concentration, the amount of lead ions is not sufficient to maintain the stability of the polymer that dissolves in water and the sorption capacity strongly decreases. REFERENCES

1. 2. 3. 4. 5.

B. Volesky and Z.R. Holan, Biotechnol. Prog., 11 (1995) 235. D. Aderhold, C.J. Williams and R.G.J. Edyvean, Bioresource Technol., 58 (1996) 1. Q. Yu, J.T. Matheickal, P. Yin and P. Kaewsarn, Wat. Res., 33 (1999) 1534. K. Y.-H. Gin, Y.-Z. Tang and M.A. Aziz, Wat. Res., 36 (2002) 1313. R. Jalali, H. Ghafourian, Y. Asef, S.J. Davarpanah and S. Sepehr, J. Hazard. Mater., 2834 (2002) 1. 6. Y. Sağ and T. Kutsal, Chem. Eng. J., 60 (1995) 181. 7. F. Veglió, F. Beolchini and A. Gasbarro, Process Biochem., 32 (1997) 99. 8. E. Guibal, C. Roulph and P. Le Cloirec, Wat. Res., 26 (1992) 1139. 955

Biosorption

9. J.T. Matheickal and Q. Yu, Miner. Eng., 10 (1997) 847. 10. A. Kapoor, T. Viraraghavan and D.R. Cullimore, Bioresource Technol., 70 (1999) 95. 11. W. Jianlong, Z. Xinmin, D. Decai and Z. Ding, J. Biotechnol., 87 (2001) 273. 12. Z. Zulfadhly, M.D. Mashitah and S. Bhatia, Environ. Pollut., 112 (2001) 463. 13. P.A. Marques, H.M. Pinheiro, J.A. Teixeira and M.F. Rosa, Desalination, 124 (1999) 137. 14. V.V. Kobak, M.V. Sutkevich, I.G. Plashchina and E.E. Braudo, J. Inorg. Biochem., 61 (1996) 221. 15. E. Guibal, C. Roulph and P. Le Cloirec, Environ. Sci. Technol., 29 (1995) 2496. 16. Y. Tang, B. Chen and S. Mo, Talanta, 43 (1996) 761. 17. M. Loaëc, R. Olier and J. Guezennec, Wat. Res., 31 (1997) 1171. 18. N.P. Chandia, B. Matsuhiro and A.E. Vásquez, Carbohydr. Polym., 46 (2001) 81. 19. L. Chevelot, B. Mulloy, J. Ratiskol, A. Foucault and S. Colliec-Jouault, Carbohydr. Res., 330 (2001) 529. 20. M.F. Marais and J.-P. Joseleau, Carbohydr. Res., 336 (2001) 155. 21. J. Chen, F. Tendeyong and S. Yiacoumi, Environ. Sci. Technol., 31 (1997) 1433. 22. L.K. Jang, D. Nguyen, and G.G. Geesey, Wat. Res., 33 (1999) 2817. 23. L. Dambies, T. Vincent, A. Domard and E. Guibal, Biomacromolecules, 2 (2001) 1198. 24. M. Tsezos and B. Volesky, Biotechnol. Bioeng., 24 (1982) 385. 25. A. Findon, G. McKay and H.S. Blair, J. Environ. Sci. Health, A28 (1993) 173. 26. P.R. Puranik and K.M. Paknikar, Bioresource Technol., 70 (1999) 269. 27. P.R. Puranik, J.M. Modak and K.M. Paknikar, Hydrometallurgy, 52 (1999) 189. 28. J. Crank, The mathematics of diffusion. Clarendon Press, Oxford, U.K., 1975. 29. C. Tien, Adsorption Calculations and Modeling. Butterworth-Heinemann, Boston, 1994.

956

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Regeneration of biomass after sorption of heavy metals P. Massacci, E. Migliavacca and M. Ferrini Department of Chemical, Material, Raw Material and Metallurgical Engineering, University of Roma “La Sapienza”, Italy [email protected] Abstract The work takes into consideration caption of heavy metals carried by freshwaters contaminated by effluents coming from industrial plants. In the case of low metal concentration and high flow of water stream is very expensive to adopt decontamination methods based on chemical precipitation or ion exchanging. Heavy metal biosorption by non-living biomass appears to be an effective technology for caption of dissolved metals, especially in very diluted streams. Considering a processing approach, the aim of this study is to verify the sorption potential of non-living biomass coming from marine seaweed. Influence of pH on sorption kinetics has been assessed before testing the desorption behaviour of cached metals by the biomass. Batch experiments were carried out bringing into contact 1 g/l of non-living biomass with a metal solution bearing a set of heavy metals. A contact period of 90 minutes between the biomass and a solution, containing Cd, Zn, Pb, Cu and Ni (10 mg/l each) + Hg (1 mg/l Hg), has been assured. It has been ascertained that the biomass extracts on average 80% of total metal content. Since the concentration of sorbed Hg was found to be well correlated to that of the other sorbed ions (r2 = 0.75), mercury was assumed as an indicator of the process in sorption/desorption tests. Results show that fast sorption/de-sorption reactions occur between biomass and heavy metals solved in water. After 15 minutes contact with the metal solution (pH = 5÷7), the biomass extracted on average 70% of the initial Hg concentration. After a contact period of just 5 minutes (by washing with HCl solution at pH = 2) more than 50% of sorbed Hg has been recovered. This regeneration procedure shows to be effective and the biomass can be reused for new cycles of sorption. 1.

INTRODUCTION Industrial, mineral and metallurgical processes produce large amount of heavy-metalbearing effluents that are among the most dangerous contaminants, as they are persistent in the environment and bio-available. Mobilised metals tend to persist indefinitely in the environment, circulating and finally accumulating in the food chain, setting so a serious threat to flora, fauna and, eventually, humans. Interest in developing a system able to reduce toxic metal concentration in waters and soils is therefore increasing. In the case of low metal concentration and high flow of water stream is very expensive to adopt decontamination methods based on chemical precipitation or ion exchanging. 957

Biosorption

Biological treatment (based on living organisms) is a non-expensive option for effluent decontamination from low metal concentration, but it is highly affected by water temperature, pH, chemical and nutrient composition. The use of non-living biomass (biosorption) has been widely studied in the last years. It does not require nourishing and it does not suffer from toxicity problems [5-6, 12-15]. Since non-living biomass behaves as an ion-exchange resin ([8, 16]), heavy metal biosorption appears to be an effective technology for treatment of dissolved metals, especially in very diluted streams. It can be applied not only to remove toxic or radioactive metals, but also to recover precious metals. For instance, many kinds of seaweed and the microbial waste, coming from fermentation processes, are abundant and economic sources of biomass. The advantages of the bio-sorption are: −

use of low cost materials to extract metals from solution;

−

ability to purify very diluted effluents and to concentrate pollutants in a small quantity of contaminated biomass. The biomaterial can be disposed either as a toxic refuse or after incineration and eventually after metal extraction or after desorption to collect metals into a small volume liquor, that can be treated by traditional methods. Although heavy metal bio-sorption has been extensively studied in last few years, little attention has been paid to bio-sorption of multi-component systems. In fact, evaluation of experimental results is more difficult in presence of more than one metal in the system (although it is a condition more similar to real systems): the interactions between dissolved ions and biomass surface, which has a limited number of binding sites, plays a determinant role [15]. In this work the heavy-metals uptake properties of a vegetal biomass (green algae and leaves of marine macro-phytes), coming from a very polluted wetland, were investigated with a process viewpoint. The biomass was put in contact with a solution containing a set of metal nitrates (10 mg/l of Cd, Zn, Pb, Cu and Ni + 1 mg/l of Hg in distilled water) chosen because of their relevant presence in the wetland ad their accumulation in the living biomass [9]. 2.

EXPERIMENTAL

2.1 Seaweed collection The biomass was collected on the shore of the wetland of Boi Cerbus, in south-west Sardinia (Italy), where former studies [3] and [10] found heavy contamination of the environment coming from mining and metallurgical plants. Heavy metals contamination is mainly present in sediments due to precipitation induced by the high pH (around 9) of the wetland salt water. Collected biomass is made up of a green alga and, mainly, of leaves of marine seagrasses (Cymodocea, Zostera and Ruppia). Material was washed in tap and distilled water. All samples were dried for 15 days at room temperature (18÷22°C) and then oven dried at 50°C for 24 h. The final dry weight of the biomass was about 9% of the wet one. Dried samples were ground by a blade mill (RETSCH SM2000) equipped with 4 mm outlet screen.

958

Biosorption

2.2 Experimental procedure Sorption experiments (batch) were carried out using a solution containing metal nitrates (10 mg/l each of Cd, Zn, Pb, Cu and Ni + 1 mg/l of Hg) in distilled water. pH was adjusted using HNO3 and NaOH. This solution was brought into contact with 1 g/l of dry seaweed biomass and kept in agitation. At the end of the contact period, samples were filtered and the resulting solutions were analysed for metal content. Filtered biomass samples were preserved for the desorption step. De-sorption batch experiments were performed on the biomass coming from the sorption tests, using diluted HCl (pH adjusted by NaOH). The solution was kept agitated for a fixed contact period, and then filtered. The resulting solutions were analysed for metal content. 2.3 Metal analyses The analyses of the metal content of Zn, Pb, Cd, Ni and Cu in the water samples were performed by using Inductively Coupled Plasma-mass Spectrometry. The Hg content in all the experiments was analysed by AMA 254 model Atomic Absorption Spectrometer (λ = 253.65 nm). All the analyses were carried out on liquid samples. The difference between initial metal concentration and the remaining concentration after each step of the experiments was assumed to be bound to biomass. 2.4 Influence of pH on sorption and desorption To evaluate the influence of pH on sorption kinetics, a first set of experiments was performed at initial pH adjusted up to 5, 6, 7 and 9. For each pH value, 1,000 cm3 of metal solution were added to 1,000 mg of biomass into a magnetically stirred beaker. After 90 minutes, samples were filtered and the separated solutions were analysed for metal ions. Filtered biomass samples were conserved for the de-sorption step. In order to remove the bound metal ions from the biomaterial, filtered biomass samples were added to 1,000 cm3 of HCl solution at pH = 2 and magnetically stirred in a beaker. After 60 minutes, the samples were filtered and the separated solutions were analysed for metal content. In the sorption tests, the concentration of sorbed Hg was found to be well correlated to that of the other sorbed ions (r2 = 0.75). Therefore Hg content was assumed as an indicator of the sorption behaviour and in the next experiments the analyses were performed taking into consideration only the Hg content in the waters (notwithstanding the solution was contaminated by the set of heavy metals, as previous described). 2.6 Kinetics Sorption experiments were performed at pH = 6, adding 500 mg of biomass to 500 3 cm of metal solution in a magnetically stirred beaker. The solution was sampled (3 cm3) after 5, 10, 15, 20, 30, 45, 60, 75 and 90 minutes from starting. The samples were filtered and the resulting solutions were analysed for Hg content. Filtered biomass samples were preserved for the de-sorption step. In order to evaluate the de-sorption kinetic, a preliminary sorption step, under the previous described conditions for just 15 minutes contact, was carried out. 100 mg of the filtered biomass, coming from the sorption test, was added to 100 ml of HCl solution at

959

Biosorption

pH = 2: four samples (2 cm3) were taken after 5, 15, 30 and 60 minutes after starting the desorption experiment. Water samples were filtered and analysed for Hg content. 2.7 Sorption/desorption cycles In order to investigate the possibility to reuse the biomass for new sorption cycles, a set of experiments including 3 cycles of sorption and de-sorption was carried out. Every sorption step was performed adding 100 mg of biomass to 100 cm3 of the polluted solution initially adjusted at the pH 5, 7 or 9, as previously reported. After 15 minutes rocking the biomass was filtered and separated. Every de-sorption step was performed adding the filtered biomass to 100 cm3 of HCl solution at pH = 2. After 5 minutes rocking, samples were filtered and the separated biomass preserved for a new sorption step. All the filtered solutions were analysed for Hg content. Table 1. Overview of the batch experiments Experiment Sorption De-sorption Sorption kinetic De-sorption kinetic Sorption cycle De-sorption cycle

Sample

Initial pH

1 1 2 3 4,5,6 4,5,6

5, 6, 7, 9 2 6 2 5, 7, 9 2

Contact time (min) 90 60 90 60 15 + 15 + 15 5+5+5

Analysed metal ion Hg, Zn, Pb, Cd, Cu, Ni Hg, Zn, Pb, Cd, Cu, Ni Hg Hg Hg Hg

3.

RESULTS AND DISCUSSION Binding of Zn, Cd, Pb, Cu, Ni and Hg on dry seaweed is plotted in Figure 1 as a function of pH. Sorption occurs in a very short time as a confirmation of previous works ([4, 11]) reporting that the metal (Cd, Zn, Pb and Hg) absorption equilibrium on vegetable biomass was reached in less than 80 minutes.

It can be observed that sorption behaviour appears to be efficient at pH 5÷6. As a confrontation, sorption behaviour was performed also at pH = 9, which is the pH value of the contaminated seawater close to the site, where the seaweed were collected. At a pH between 5 and 7 the mean sorption capacity is the 80% of total metal content: from about 70% for Ni and Hg to almost 100% for Pb. De-sorption of the loaded biomass permits 70% recovery of bound metals. Collected seaweed shows special affinity with lead, which is the meaningful contaminant of the collected seaweed; furthermore lead is strongly bound to the biomass as resulting from sorption/de-sorption experiments shown in Table 2. The experiments carried out at pH = 9 showed a relevant reduction of metals into the solution, probably due to their precipitation, improving the extraction of metals by the biomass. At least 40% of the total metal content is bound onto the biomass, as it can be appreciated from balance of metals coming into the leaching (HCl) solution after desorption at pH = 2 for 60 minutes (see Fig. 2). At Hg concentration of 1 mg/l, in the range of pH 5÷9, activity-pH diagrams for Hg(II) species [7] show that mercury is present in solution as Hg(OH)2, available for sorption onto biomass. During the leaching step by HCl solution, as the pH value 960

Biosorption

decreases, all the Hg(II) species [HgCl-3, HgCl2-4, HgCl+, and Hg(OH)2] appear to be soluble. Table 2. Natural concentration, metal binding capacity (pH = 6) of the marine seaweed used for batch experiments. Metal recoveries (%) by HCl solution at pH=2 (from desorption experiments) are also shown Metals Cd Zn Pb Cu Ni Hg

Natural concentration into collected biomass (mg/g) < 20.0 263.0 1011.0 65.0 35.0 0.1

Binding capacity (mg/g) after 90 min at pH = 6 8.7 8.2 10.9 9.7 7.5 0.8

Metal recoveries (%) after 60 min desorption at pH = 2 78.0 72.2 44.7 78.2 68.9 51.4

Sorbedbound metal(%) (%) Metal

100 80 60 40 20

Cd

Zn

Pb

Cu

Ni

Hg

0 5

6

7 8 9 Initial pH Figure 1. Influence of pH on heavy metal sorption by seaweed after simultaneous feeding. Biomass (1 g/l) was stirred for 90 minutes with distilled water containing 10 mg/l of Pb, Zn, Cd, Cu and Ni + 1 mg/l of Hg as nitrates 100 Cd Pb Ni

Sorbed metal (%)

80

Zn Cu Hg

60 40 20 0 5

6

7 Sorption pH

8

9

Figure 2. Influence of sorption pH on de-sorption of heavy metals by seaweed. Biomass was stirred for 60 minutes with distilled water containing HCl (pH = 2) 961

Biosorption

9 8 7

0

6 0

15

30

45

60

Time (minutes)

75

S

Sorbed Hg (%)

1

0

90

0

D

15

30

45

60

Time (minutes)

75

90

Figure 4. Time dependency of desorption (initial pH = 2) of the metal loaded biomass

Figure 3. Sorbed Hg (mg/g of biomass) (o) and pH (Ð) as a function of the sorption time

200

Sorbed Hg (mg/g)

1

pH

Sorbed Hg (mg/g)

In such a kind of systems metal-speciation at different pHs is very complex [1]: nevertheless Hg content has been assumed as an indicator of the sorption behaviour of the metal set, since sorption tests showed an evident correlation between Hg content and the other sorbed ions, in every tested condition. From the statistical point of view metal speciation appears to be not relevant and the adoption of Hg as an indicator shorts the time required for analyses. For this reason, even if all batch experiments were performed using solutions bearing all the metals, as previously reported, only Hg analyses were carried out on the filtered solutions coming from the sorption/de-sorption tests. Kinetic batch experiments showed that 70% of the initial Hg content was sorbed after 15 minutes, covering 84% of the sorption capacity of the biomass achieved after 90 minutes of contact with the polluted solution (Figure 3). Metal de-sorption by HCl solution (pH = 2) for 5 minutes is enough to recover 90% of the previously bound Hg (see Figure 4). Results of sorption/de-sorption experiments are shown in Figure 5. Contact periods of 15 minutes and 5 minutes are adopted respectively for sorption and de-sorption experiments; the cycle was repeated three times. Results show the possibility to regenerate the biomass with HCl solution, even if the metal uptake capacity progressively decreases.

S

D

S

D

sorption pH=9 de-sorption pH=2 sorption pH=7 de-sorption pH=2 sorption pH=5 de-sorption pH=2

150 100 50 0

0

15

30

Time (minutes)

45

60

Figure 5. Hg (%) sorbed on biomass after 3 cycles of sorption (S) (15 minutes contact with the polluted solution) and de-sorption (D) (5 minutes contact with the HCl solution)

962

Biosorption

4.

CONCLUSIONS Results of batch experiments show that fast sorption/de-sorption reactions occur between biomass and heavy metals (Cd, Zn, Pb, Cu, Ni and Hg) solved in water. The total metal-uptake capacity of the biomass has been found up to 0.54 mMol/g dry weight. After a contact period of 90 minutes in a metal solution (10 mg/l of Cd, Zn, Pb, Cu and Ni + 1 mg/l of Hg) at pH between 5 and 9, the biomass extracts on average 80% of total metal content. Since the concentration of sorbed Hg was found to be well correlated to that of the other sorbed ions (r2 = 0.75) mercury was assumed as an indicator of the process in sorption/de-sorption tests. After 15 minutes contact with the metal solution (pH = 5÷7), the seaweed biomass extracts on average 70% of the initial Hg concentration. After a desorption time of just 5 minutes with HCl solution at pH = 2 it allows to recover more than 50% of the sorbed Hg. Regeneration of the biomass with HCl solution at pH = 2 shows to be effective and the biomass can be reused for more cycles. This study, carried out on a laboratory-batch scale, is the first step of an experimental series devoted to point out a continuous flow sorption/de-sorption system. Treating contaminated waters, intercepted before their influx into the water bodies, can prevent contamination of the recipient environment by heavy metals. REFERENCES

1. Brandon N.P., Francis P.A., Jeffrey J., Kelsall G.H., Yin Q., Thermodynamics and electrochemical behaviour of Hg-S-Cl-H2O systems. J. Electroan. Chem. 479 (2001): 18-32. 2. Catsiki V.A., Panayotidis P., Copper, Chromium and nickel in tissues of the mediterranean seagrasses Posidonia oceanica and Cymodocea nodosa (Potamogetonaceae) from Greek coastal areas. Chemosphere 26 (1993) 963-978. 3. De Giorgi F., Analisi e caratterizzazione dei sistemi ambientali degradati ai fini del loro risanamento: il caso della laguna di Boi Cerbus. (Tesi di laurea Dipartimento ICMMPM, Università di Roma “La Sapienza”, Roma, 1999). 4. Gardea-Torresdey J.L., Tiemann K.J., Gamez G., Dokken K., Effects of chemical competition for multi-metal binding by Medicago sativa (alfalfa). J. Hazard. Mater. B69 (1999) 41–51 5. Holan Z.R., Volesky B, Prasetyo I., Biosorption of cadmium by biomass of marine algae. Biotechnol. Bioeng. 41 (1993) 819-825. 6. Holan Z.R., Volesky B., Biosorption of lead and nichel by biomass of marine algae. Biotechnol. Bioeng. 43 (1994) 1001-1009. 7. Kragten J., Atlas of Metal-Ligand Equilibrium in Aqueous Solution (Wiley, New York, 1978) 8. Kratochvil D., Volesky B., Advances in the biosorption of heavy metals. TIBTECH 16 (July 1998) 291-300. 9. Migliavacca E., Cationi metallici dagli effluenti industriali all'ecosistema: assorbimento e decontaminazion.e (Tesi di laurea, Dipartimento ICMMPM, Università di Roma “La Sapienza”, Roma, 2002). 10. Provincia di Cagliari, Relazione sullo stato ambientale della laguna di Boi Cerbus (Portoscuso) e delle sue aree riparie. (Provincia di Cagliari - 1996). 963

Biosorption

11. Lacher C., Smith R.W. Sorption of Hg(II) by Potamogeton natans dead biomass. Min. Eng. 15 (2002) 187-191. 12. Rubio J., Schneider I.A.H., Treatment of effluents by sorption on freshwater macrophytes, SWEMP'96 Cagliari, Italy, Oct. 7-11, 1996. (Ciccu R., Cagliari, 1996), pp. 483-491. 13. Sanchez A., Ballester A., Blazquez M.L., Gonzalez F., Muñoz J., Hammaini A., Biosorption of copper and zinc by Cymodocea nodosa. FEMS Micro. Rev. 23 (1999) 527-536 14. Schiewer S., Volesky B., Modeling of the proton-metal ion exchange in biosorption. Environ. Sci. Technol. 29 (1995) 3049-3058. 15. Tsezos M., Volesky B., Holan Z.R., Biosorption of heavy metals. Biotechnol. Prog. 11 (1995) 235-250 16. Volesky B., Detoxification of metal-bearing effluents: biosorption for the next century. Hydrometallurgy 59 (2001) 203–216

964

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Structural modeling of arsenic biosorption using X-Ray spectroscopy (XAS) Mônica Cristina Teixeiraa, Graziele Duarte and Virgínia S.T. Ciminelli∗ a

Permanent address: Department of Pharmacy – Pharmacy School – Universidade Federal de Ouro Preto. Rua Costa Sena, 171. Ouro Preto - MG. CEP: 35400-000. Brazil Department of Metallurgical and Materials Engineering – Universidade Federal de Minas Gerais. Rua Espírito Santo, 35/206. Belo Horizonte-MG. CEP: 30160-030. Brazil Abstract The work describes a biological route for direct immobilization of aqueous As(III) species, which is the most toxic and mobile arsenic species found in soils. Based upon the biochemical mechanisms which explain arsenic toxicity, we propose that waste biomass with a high fibrous protein content can be used for As(III) immobilisation. Our investigations demonstrated that As(III) is specifically adsorbed on the biomass and, contrary to the behaviour observed with inorganic sorbents, the lower the pH the more effective the removal. Arsenic uptake reaches values of up to 20 mg As(III)/g of biomass. Analyses by Synchrotron light techniques, such as XANES, demonstrated that arsenic is immobilised in its trivalent state, an advantage over to the conventional techniques for As immobilisation, which usually require a previous oxidation stage. EXAFS analyses shows that each As atom is directly bound to three S atoms with an estimated distance of 2.26 Å. The uptake mechanism is explained in terms of the structural similarities between the As(III)-biomass complex structure and that of arsenite ions and Ars operon system encoded proteins and phytochelatins. The biological route presented here offers the perspective of a direct removal of arsenic in its reduced form.

Keywords: arsenite, sorption, bioremediation, EXAFS 1.

INTRODUCTION Arsenic and its compounds, the most usual being that of the water-soluble derived from arsenous (H3AsO3) and arsenic (H3AsO4) acids, are toxic and carcinogenic to all living organisms. The trivalent species is of great environmental concern in view of its considerably higher toxicity and mobility in soils. Recent disasters involving cases of poisoning have spawned a series of worldwide investigations towards As(III) immobilisation. Despite all work undertaken, there still remains a great lack of knowledge as regards arsenic sorption in general, especially with reference to the trivalent species.

∗

Corresponding author: [email protected]

965

Biosorption

Arsenic is found throughout the earth´s crust in small concentrations, arsenopyrite (FeAsS) being the most common arsenic mineral. In soils, As concentration may vary from 1.0 mg/Kg (apatite, fluorite, and calcite samples) to 77,000-126,000 mg/Kg (pyrite or arsenopyrite samples) [1]. Normal arsenic concentration found in soils is 6 mg/Kg [2]. Small amounts of arsenic and its compounds are used by the chemical and electronic industries to produce electronic components for laser equipment, wood preservatives, pesticides, and glasses, to name but a few of the numerous applications. Despite the many useful applications, there is a surplus of arsenic found in wastes, derived mainly from the mineral and metallurgical industries. The natural leaching of As-enriched soils and rocks for human water supplies, is yet another cause of human contamination, a good example being that of Bangladesh. Much attention has been paid to environmental contamination caused by heavy metals. Nevertheless, the understanding of the biochemical mechanisms, which are responsible for these elements’ toxicity, has not received due consideration. The toxicity of arsenic and its compounds is well established. Once ingested, arsenic provokes nausea and gastro-intestinal symptoms. The long-term exposure to this element causes skin problems like dermatitis, keratosis and cancer. From the toxicological point of view As(V) causes adverse effects to human and others living organisms due to its chemical similarity with phosphate [3, 4]. In this case, arsenic poisoning could be reverted by the administration of an excess of phosphate. The trivalent specie As(III) strongly binds to the sulphidryl (SH) active sites of some enzymes causing irreversible metabolic impairments and cellular mutagenesis [5, 6]. The wastewater produced by mining and metallurgical industries is a very important source of arsenic contamination. The conventional techniques used for As immobilisation usually require a previous oxidation stage in order to oxidize the more toxic and mobile trivalent arsenic species [7] to the pentavalent species. In the pentavalent state, arsenic acid (H3AsO4) species form stable complexes with soils constituents containing ferric, manganese or aluminium oxy-hydroxides, such as goethite, alumina, hematite, birnessite and gibbsite [1, 8-11]. This could explain its lower environmental mobility. On the other hand, the trivalent arsenous acid (H3AsO3) species are weakly bound to inorganic sorbents regardless of the pH uptake. To achieve immobilization, it is often necessary to oxidize As(III) ions to As(V). Biosorption is usually based on unspecific ion exchange mechanisms. For instance, positively charged chemical groups present in the biomass, like the amino group, are capable binding to negatively charged ions like arsenate, arsenite, chromate, sulphate or phosphate. The poor selectivity associated with unspecific ion exchange mechanism is probably one of the main limitations for applications to complex, multicomponent systems, which characterise real systems. Arsenic binding to sulphidryl groups available in some enzymes is irreversible and very specific, which explain arsenic toxicity. This strong affinity has led us to believe that waste biomasses with a high content of fibrous protein could be used for As immobilisation. This hypothesis was tested with a waste material produced by the poultry industry. Based on the results, we present, for the first time, a highly specific biosorbent for As (III) immobilisation. This specificity is explained by the molecular structure of the adsorbed complex determined by Syncrotron light-X-Ray Spectroscopy analyses. 2.

MATERIALS AND METHODS The biomass was prepared under laboratory conditions. White chicken feathers were rinsed exhaustively with warm tap water and dried at 45 ±5°C for 24 hours. The dried material was ground and sieved to obtain a size range below 0.037 mm (400 Mesh Tyler).

966

Biosorption

Biomass activation was accomplished by adding a basic ammonium thioglicolate solution to the final concentration of 0.78 g/L. Treatment did not imply any mass loss. After this activation step, powdered biomass was filtered and used in the adsorption tests. All chemicals used were of analytical grade. Water was first deionised and then ultrafiltered before being used to prepare the solutions. As(III) stock solutions of 10,000 mg/L were made from AsNaO2 (Fluka, 99.0%). The pH values were adjusted with 0.1 N HCl or NaOH solutions. Ionic strength (I) was controlled using 4 M NaCl or 0.01 M Na3PO4 electrolyte solutions. As(III) batch adsorption experiments were conducted at room temperature (28±3°C), by adding a known amount of biosorbent (1-10 mg/L) to each 250ml Erlenmeyer flask containing the metal solution (100 mL). Flasks were shaken (100 rpm) for 1 hour to achieve equilibrium. As(III) semi-continuous adsorption experiments were undertaken at constant temperature (25±0.2°C) using an apparatus similar to that described by Pagnanelli et al. [12], and the same proposed procedure called "Subsequent Additions Method" (SAM). The liquid volume in the reactor was 1,000 mL, biosorbent concentration was 2 g/L, and agitation and pH values were kept constant. One hour after each metal addition, the system reaches equilibrium and a 10 mL sample is collected and analysed. Reaction suspensions were filtered through a 0.45 µm cellulose membrane, and preserved with concentrated nitric acid (5 µL) for chemical analyses. Experiments were carried out in duplicate and results were averaged. X-Ray Absorption Near Edge Structure (XANES) and Extended X-Ray Absorption Fine Structure (EXAFS) analyses of humid biomass samples loaded with As(III) were performed using the synchrotron facilities at the Laboratório Nacional de Luz Synchrotron (LNLS), in Campinas, São Paulo. XANES and EXAFS data from the arsenic K edge (11,868 eV) were obtained at XAS workstation using beam currents of about 200 mA. All spectra were recorded at room temperature, using a Si(111) double crystal monochromator with an upstream vertical aperture of 0.6 mm. As K-edge X-ray absorption spectra were measured by monitoring the transmitted energy using a 15-element Ge detector (Canberra Industries). The energy resolution utilised were 0.8 eV at the XANES region located near the As K-edge (11855-11930 eV); 2 eV at the regions located between 11760 and 11855 eV, and 11930 and 12400 eV and 3 eV at 12400-13000 eV region. This procedure allowed us to obtain XANES and EXAFS spectra simultaneously. Counting times of 3 s were kept constant. XANES spectra were analysed using Origin 5.0 software and EXAFS collected data were analysed by using Winxas 97 software. EXAFS data fit was obtained using phase and amplitude parameters calculated with FEFF 6.01 software. 3.

RESULTS AND DISCUSSION Preliminary experimental results related to As(III) adsorption by the selected biomass are shown in Figure 1. Arsenite biosorption equilibria are achieved in less than 10 min for all experimental conditions. Results of tests performed with biomass prior to activation led to a negligible As uptake, a finding that supports the hypothesis that the reduced sulphidryl groups are responsible for As(III) adsorption. The milling process seems to negatively contribute to As uptake. Nevertheless, all the subsequent experimental tests were performed using ground material in order to favor biosorbent’s homogeneity. Loading capacities (8-13 mg As(III)/g of biomass) are promising and greater than those values obtained by using kaolinite and montmorillonite, 0.1 and 0.2 mg/L, respectively [13]; alumina, 0.2 mg/L [9]; goethite, 3.0 mg/L [8]. Ladeira et al. [11] reports significantly higher uptakes using thermally activated gibbsite, 25.4 mg/L. Driehaus et al. [10] and Meng et al. [14] also obtained higher arsenic adsorption capacities (20-40 mg/L) 967

Biosorption

but, in both cases, As(III) was previously oxidized to As(V). The very high value (69.2 mg of As(III) per each gram of Mo) reported by Dambies et al. [15], using a chitosan derivative biosorbent impregnated with molybdenum, can not be compared to the other results in view of the lack of information with regard to the quantities (mass) of biosorbent. 14

Adsorbed As(III)/g biomass

12 10 8 6 4

Whole Biomass (1g/L) Powdered Biomass (1g/L) Powdered Biomass (2g/L)

2 0

0

10

20

30

40

50

60

Time (min)

Figure 1. Milling and biomass concentration influence on As(III) uptake. Flask tests, As(III) initial concentration, 100mg/L; initial pH, 9.2; temperature, 28±3°C; pretreatment, 2h

The influence of pH on As(III) biosorption can be observed in Figure 2. The obtained results show that the lower the pH, the higher the uptake. This trend is just the opposite of that observed in As(III) uptake by inorganic sorbents, for which uptake increases with pH. During all the adsorption experiments, pH variation was less than 0.2 units. Taking into account that the pKa1 for arsenous acid is 9.2, experiments were performed at pH no higher than 8, i.e. 2.0, 5.0 and 8.0, in order to avoid or minimise the formation of anionic H3AsO3 species. Arsine formation during sorption experiments carried out at pH 2.0 was experimentally investigated and discarded (data not shown). A pH of 5.0 was chosen for all the subsequent experiments, as this pH value is consistent with conditions found in natural wastewater or industrial effluents. The results found in Figure 3 are most interesting. The great majority of As sorbents described in literature are active for both the pentavalent and the trivalent species, the main difference being the relatively higher mobility of the latter. The specificity of the proposed biosorbent towards As(III) is combined with a rejection of both phosphate species (Figure 3) and arsenate (not shown here) species. The well-described competition between arsenic and phosphate during sorptive experiments using biosorbents or resins [15-18] is not observed when the fibrous protein rich biomass is utilised. Both phosphate and arsenate molecules have the same tetraedric geometry which could explain their chemical similarity and their similar affinities for the same chemical ligands. Conversely, arsenite ion possesses a trigonal pyramidal geometry. It is possible that a steric hindrance may contribute to the rejection of the tetraedric arsenate and phosphate oxyanions.

968

Biosorption

12

12

10

Q (mg As/g )

8

Q (mg As/g biomass)

9

6 4 2

6

pH 5.0 pH 10.0

0 0

20

40

60

80

100

120 -1

3

As(III) - Concentração de Equilíbrio (mg.L )

pH 2.0 pH 5.0 pH 8.0

0 0

20

40

60

80

100 -1

As (III) - Equilibrium Concentration (mg.L )

Figure 2. Influence of pH on As(III) adsorption isotherms, SAM procedure (biomasss, 2g/L; temperature, 25±1°C) 20

Q (mg As/g biomass)

16

12

8

4

0.01M phosphate; 0.0M NaCl 0.0M phosphate; 0.08M NaCl

0 0

50

100

150

200

250

300 -1

As (III) - Equilibrium Concentration (mg.L )

Figure 3. Influence of phosphate ions on As(III) adsorption isotherms, SAM procedure (I = 0.1; pH=5; biomass, 2g/L; temperature, 25±1°C)

X-Ray Spectroscopy analyses, X-Ray Absorption Near Edge Structure (XANES) and Extended X-Ray Absorption Fine Structure (EXAFS), may be considered state of the art when investigating adsorption at molecular levels. These are modern and very precise tools, which provide information that could not be otherwise obtained through the traditional surface analyses techniques. Among the techniques described in the present work, XANES spectra offer electronic and structural information, such as oxidation state, with regard to adsorbed ion (photoabsorbing ion). EXAFS provides information, e.g. the 969

Biosorption

coordination number and interatomic distance, about the nature and position of the ligand atoms in the coordination shell of the photoabsorbing ion (scattering atom). The procedure involved in the analyses of the EXAFS data from arsenite-loaded biomass is illustrated in Figures 4 to 6. The As(III)-biomass EXAFS spectrum is presented in Figure 4. The averaged data from seven different spectra were converted to eV energy unit and had its background line extracted. The experimentally obtained K edge value (Eο) was 11,868 eV, the same obtained for the arsenite standard sample, confirming that arsenite was not oxidized by the biomass. XANES spectra (data not shown) confirmed this value and the trivalent state of arsenic atoms. The spectrum oscillations caused by all the atoms in the neighbouring coordination shells are also presented. This spectrum was submitted to Fourier Transform, thus allowing the identification of one amplified peak that corresponds to the Arsenic first coordination shell (Figure 5). 1.0

0.20

11.868 Kev K χ(K)

0.5

EXAFS-signal

2

0.15

0.0 -0.5 -1.0

0.10

-1.5 2

4

5

6

7 -1

KÅ 0.05

0.00

11.6

11.8

12.0

12.2

12.4

12.6

12.8

13.0

Energy (Kev)

Figure 4. EXAFS signal after background correction

0.10

First Coordination Shell

0.06

3

FTk χ(k)

0.08

0.04 0.02 0.00 0

2

4

R(Å)

Figure 5. Fourier Transform amplitude (K=3)

970

3

6

8

8

9

10

11

Biosorption

The signal obtained after submitting this data to another Fourier Transform treatment results in one spectrum that represents only the oscillations caused by the atoms in the As first coordination shell. At this point, it is possible to calculate the structural parameters such as interatomic distance between As and atoms in the first coordination shell, coordination number as well as to identify the "first neighbour" ligand. By adjusting the experimental data with the theoretical model [18] provided by the FEFF program (Figure 6) it was possible to confirm that arsenic is the scattering atom while sulphur is the retroscattering atom. It was also possible to affirm that each arsenic atom is bound to three sulphur atoms. The final structural parameters obtained in the analyses were coordination number (n) = 2.5±0.4 and interatomic distance (R) = 2.26±0.01 Å. 0.3

0.2

χ(k)

0.1

0.0

-0.1

experimental data theoretic prediction

-0.2

3

4

5

6

7

8

9

10

-1

k(Å )

Figure 6. Back Fourier Transform, first coordination shell. Experimental and theorethical curves (R=2.26±0.01 Å, n=2.5±0.4)

The structural parameters obtained during this work are quite different from those obtained by arsenic adsorption on inorganic matrices. As arsenic is adsorbed as an oxyanion, usually as a bidentate binuclear complex, the element found in the first coordination shell is always oxygen, with coordination numbers (n) varying from 3.6-4 and interatomic (R) distances in a range of 1.72 to 1.78 Å. The metal ligands (Fe or Al) are found in the second coordination shell with R values often greater than 3.0 Å [19-20]. The coordination number and interatomic distance obtained in the present study are, as expected, very similar to those reported in EXAFS analyses of biological As(III)/protein complexes, showing As atoms directly bound to S atoms in the first coordination shell. Each As atom is bound to the sulphur atoms coming from three different cysteine residues, R values vary from 2.20-2.25 Å [21-25]. The information provided by XAS analyses is consistent with the strong arsenic uptake reported here. The results explain that rather than adsorbed as a counter ion, or specifically adsorbed as arsenous species in the inner Helmoltz plane, As(III) undergoes a chemical reaction leading to dehydration of H3AsO3 molecule. Based on these findings we propose the following equation to describe arsenite immobilization by fibrous protein biomass: (1) H3AsO3 + 3B-CysSH → As(B-CysS)3 + 3H2O where B represents the biomass matrix. This adsorption mechanism is supported by the structural similarities between As(III)/biomass complex and those natural complexes formed between Arsenic atoms and 971

Biosorption

Ars Operon proteins [21, 22, 24-26], phytochelatins [23,27] or cysteine and glutathione [4, 24, 28, 29], previously identified. Finally, it may be stressed that the immobilization phenomenon studied here involves a chemical reaction between the dissolved arsenite ions and the sulphidryl groups of biomass, strong enough to dislocate oxygen atoms from the arsenic atom first coordination shell in arsenous acid molecules. The facts explain the specificity of this tested biosorbent by the trivalent species of arsenic as well as its minor affinity for phosphate and arsenate ions. 4.

CONCLUSIONS A waste biomass with a high content of fibrous protein was shown to specifically sorb arsenic in the trivalent state, therefore dispensing with the need of a previous oxidation treatment. Sulphidryl reduced groups are shown to be the active groups involved in the arsenic biosorption. Uptake increases as pH decreases; phosphate ions do not compete with arsenite ions for the biomass’ active sites. The uptake involves an inner sphere complexation phenomenon that takes place inside the arsenic first coordination shell, with the release of water and arsenic being directly bound to the sulphidryl group. XAS analyses indicated that each arsenic atom is bound directly to three sulphur atoms from the reduced cysteine aminoacids. The arsenic/sulphur interatomic distance was found to be 2.26±0.01 Å.

ACKNOWLEDGEMENTS Authors wish to acknowledge the National Synchrotron Light Laboratory (LNLS) in Campinas, São Paulo, for the use of XAS facilities. We also would like to express our gratitude to the Brazilian Scholarship Program–PICD from CAPES and to Dr. Maria do Carmo Alves for her important technical support during XAS data analyses. The support of the Millenium Science Initiative - Water a mineral approach and CNPq is also appreciated. REFERENCES

1. Smedley, P. L. and Kinniburgh, D. G. (2002). “A review of the source, behavior and distribution of arsenic in natural waters.” Apllied Geochemistry 17: 517-568. 2. Mandal, B. K. and. Suzuki, K. T (2002). “Arsenic Round the World: a Review.” Talanta 58: 201-235. 3. Nies, D. H. (1999). “Microbial heavy-metal resistance.” Applied Microbiology Biotechnology 51: 730-750. 4. Hughes, M. F. (2002). “Arsenic Toxicity and Potential Mechanisms of Action.” Toxicology Letters 133: 1-16. 5. Treagan, L. (1983). Metals and Immunity. Methods Involving Metal Ions and Complexes in Clinical Chemistry. H. Siegel. New York, Marcel Dekker, Inc. 16: 4783. 6. Flessel, P.; Furst, A. and Radding, S. B. (1980). A comparision of carcinogenic metals. Carcinogenicity and metal ions. H. Siegel. New York, Marcel Dekker, Inc. 10: 22-54. 7. Rawlins, B. G; Williams, T. M.; Breward, N.; Ferpozzi, L.; Figueiredo, B. and Borba, R. (1997). Preliminary in Investigation of Mining-related Arsenic Contamination in the Provinces of Mendoza and San Juan (Argentina) and Minas Gerais State (Brazil). Keyworth, Nottingham, British Geological Survey. 972

Biosorption

8. Deschamps, E., Ciminelli, V.; Weidler, P. G. and Ramos, A. Y. (2003). "Arsenic sorption onto soils enriched with manganes and iron minerals". Clays and Clay Minerals 51(2): 198-205. 9. Gupta, S. K. and Chen, K. C. (1978). “Arsenic removal by adsorption.” Journal of Water Pollution Control Federation 50(3): 493-506. 10. Driehaus, W.; R. Seith, R. and Jekel, M. (1995). “Oxidation of Arsenate(III) with Manganese Oxides in Water Treatment.” Water Research 29(1): 297-305. 11. Ladeira, A. C. Q.; Ciminelli, V. S. T.; Duarte, H. A.; Alves, M. C. M. and Ramos, A. Y. (2001). “Mechanism of Anion Retention from EXAFS and Density Functional Calculations: Arsenic (V) Adsorbed on Gibbsite.” Geochimica et Cosmochimica Acta 65(8): 1211-1217. 12. Pagnanelli, F.; Papini, M. P.; Trifoni, M.; Toro, L. and Veglio, F. (2000). “Biosorption of metal ions on Arthrobacter sp.: biomass characterization and biosorption modeling.” Environmental Science & Technology 34: 301-316. 13. Griffin, R. A.; Frost, R. R.; Au, A. K.; Robinson, G. D. and Shimp, N. F. (1977) "Attenuation of pollutants in municipal landfill leachate by clay minerals." Environmental Geology Notes 79:1-47. 14. Meng, X.; Korfiatis, G. P.; Bang, S. and Bang, K. W. (2002). “Combined Effects of Anions on Arsenic Removal by iron Hydroxides.” Toxicology Letters: 133 (1): 103111. 15. Dambies, L.; Vincent, T. e Guibal, E. (2002). “Treatment of Arsenic-Containing Solutions using Chitosan Derivatives: Uptake Mechanism and Sorption Perfomances.” Water Research: 36 (15): 3699-3710. 16. Dambies, L.; Roze, A.; Roussy, J. and Guibal, E. (1999). As (V) removal from dilute solutions using MICB (molybdate-impregnated chitosan beads). In: Amils, R. and Ballester, A.. (eds.) International Biohydrometallurgy Symposium - IBS99 Biohydrometallurgy and the environment toward the mining of the 21st century, Madrid, Elsevier, 277-288. 17. Korngold, E.; Belayev, N. and Aronov, L. (2001). “Removal of Arsenic from Drinking Water by Anion Exchangers.” Desalination 141: 81-84. 18. Alam, M. G. M.; Tokunaga, S. and Maekawa, T. (2001). “Extraction Arsenic in a Synthetic Arsenic-Contaminated Soil using Phosphate.” Chemosphere 43: 1035-1041. 19. Farquhar, M. L.; Charnock, J. M.; Livens, F. R. and Vaughan, D. J. (2002) "Mechanisms of Arsenic Uptake from Aqueous Solution by Interaction with Goethite, Lepidocrocite, Mackinawitte and Pyrite: Na X-Ray Absorption Spectroscopy Study." Environmental Science & Technology 36: 1757-1762. 20. Ladeira, A. C. Q. (1999). Utilizacao de Solos e Minerais para Imolizacao de Arsenio e Mecanismo de Adsorcao. Escola de Engenharia - Departamento de Engenharia Metalurgica e de Minas. Dr. thesis. Belo Horizonte, UFMG: 160. 21. Shi, W.; Dong, J.; Scott, R. A.; Kasenzenko, M. Y. and Rosen, B. (1996). “The Role of Arsenic-Thiol Interactions in Metalloregulation of the ars Operon.” The Journal of Biological Chemistry 271(16): 9291-9297. 22. Bhattacharjee, H. and Rosen, B. (1996). “Spatial Proximity of Cys 113, Cys172, and Cys 422 in the Metalloactivation Domain of the ArsA ATPase.” The Journal of Biological Chemistry 271(40): 24465-24470. 23. Pickering, I. J.; Price, R. C.; George, M. J.; Smith, R. D.; George, G. N. and Salt, D. E. (2000). “Reduction and Coordination of Arsenic in Indian Mustard.” Plant Physiology 122: 1171-1177.

973

Biosorption

24. Farrer, B. T.; McClure, C. P.; Penner-Hahn, J. E. e Pecoraro, V. L. (2000). “Arsenic (III)-Cysteine Interactions stabilize Three-Helix Bundles in Aqueous Solution.” Inorganic Chemistry 39(24): 5422-5423. 25. Martin, P.; DeMel, S.; Shi, J.; Gladysheva, T.; Gatti, D. L.; Rosen, B. P. and Edwards, B. F. P. (2001). “Insights into the structure, solvation and mechanism of ArsC arsenate reductase, a novel arsenic detoxification enzyme” Structure 9: 1071-1081. 26. Kaur, P. and Rosen, B. P. (1992). “Plasmid-Encoded Resistance to Arsenic and Antimony.” Plasmid 27: 29-40. 27. Schmoger, M. E. V.; Oven, M. and Grill, E. (2000). “Detoxification of Arsenic by Phytochelatins in Plants.” Plant Physiology 122: 793-801. 28. Knowles, F. C. and. Benson, A. A (1983). “The biochemistry of arsenic.” TIBS 8(5): 178-179. 29. Mukhopadhyay, R.; Rosen, B. P.; Phung, L. T. and Silver, S. (2002). “Microbial Arsenic: from Geocycles to Genes and Enzymes.” FEMS Microbiology 26: 311-325.

974

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Uranium and thorium removal by a Pseudomonas biomass: sorption equilibrium and mechanism of metal binding P. Sara*, S.K. Kazya and S.F. D’Souza α

Nuclear Agriculture and Biotechnolgy Division, Bhabha Atomic Research Centre, Mumbai 400 085, India

Abstract A Pseudomonas strain was characterized to develop biosorbent for removal of uranium and thorium from nuclear waste streams. The lyophilized bacterial biomass was found extremely good for U and Th uptake. 100% of added U and Th were removed by the biomass up to an initial concentration of 100 mg.l-1. Actinide biosorption was rapid, as well as a high affinity and efficient process, being optimum at pH 4-5, with a maximum loading capacity of 540 mg U g-1 and 430 mg Th g-1 at equilibrium. Bacterial cells grown in peptone rich medium or in minimal medium showed no significant difference in U accumulation at low U concentration (100 mg.l-1). However, at high concentration range (1000 mg.l-1) minimal medium grown cells showed a significantly high metal loading. Experimental sorption data showed good conformity to Langmuir model suggesting a monolayered metal binding process. Sorption in presence of several interfering cations and anions indicates a specific U and Th binding by the biomass with significant antagonism offered only by iron (III). Transmission electron microscopy (TEM) and Energy Dispersive X ray analyses (EDXA) of metal loaded biomass revealed an intracellular U and Th sequestration possibly via an ion exchange mechanism. Nuclear Magnetic Resonance (NMR) studies showed the role of cellular phosphoryl groups in radionuclide binding. Such observations were also confirmed from X ray diffraction (XRD) patterns of the metal loaded biomass that revealed the phosphide nature of sequestered U and Th. More than 90% of biomass bound radionuclide was recovered with Na- or Ca-carbonates. Bacterial biomass immobilized in a radiation polymerized polyacrylamide matrix showed a good uranium removal potential for continuous process application. Scanning electron microscopy of immobilized bio-beads revealed a highly porous nature of the matrix with bacterial cells embedded in pore walls. The overall data strongly indicate the future potential of the biosorbent in realistic application. 1.

INTRODUCTION Radionuclide and heavy metal pollution by nuclear and other industrial activities is of paramount environmental concern [1]. In view of increasingly strict legislative requirements for the discharge of large volumes of often-low activity contaminants, for a

Present address: Centre for Biotechnology, Biological Science Group, Birla Institute of Technology and Science, Pilani 333 031, Rajasthan, India. * Corresponding author. Email: [email protected]

975

Biosorption

which the conventional decontamination methods seems mostly ineffective or highly expensive, considerable recent interest has been generated in developing microbe-based remediation strategies [2, 3]. Among the several microbial processes that determine the environmental fate of metallic toxicants viz. reductive or enzymatic precipitation, solubilization, etc., biosorptive accumulation of uranium, and other radionuclides is of recent interest [4-6]. Compared to the conventional treatment methods, biomass based systems are more acceptable in being cost effective, with high efficiency of detoxification of even very dilute effluents, and minimizing the disposable sludge volume. It also offers the flexibility for developing non-destructive desorption techniques for biomass regeneration and/or quantitative metal recovery. The passive biosorptive microbial metal removal by purely physico-chemical way seems more appropriate for bioremediation that could be regulated by the characteristics of the microorganism, targeted metal and the microenvironment of the solution [7]. Since the chemical composition of the cell wall and other surface materials are responsible for cation sequestration, cell viability and other metabolic activities effectively have no impact on biosorption. For process application, one major requirement for microbe-based biosorption is the biosorbents mechanical stability and integrity during continuous operation. Immobilization or physical entrapment of biomass in a matrix is the most suitable way to enhance its mechanical strength which imparts operational flexibility and more effective biomass utilization by providing desired particle size, enhanced cell stability, easy solid-liquid separation, biomass regeneration for reuse and recovery of metals [8]. Although biosorptive uptake of several heavy metals is well documented, studies on radionuclide sorption are relatively less. Among the few reports on radionuclide sorption, fungal and bacterial biosorbents have been tested for uranium and thorium [5, 6, 9-11]. Previous studies on microbial metal sorption by our group have identified the strains of Pseudomonas as a potent accumulator of metals and radionuclides [5, 6, 12-15]. The present study was undertaken to evaluate the uranium (VI) and thorium (IV) biosorption capacity of a Pseudomonas soil isolate. Equilibrium sorption behavior of the lyophilized biomass was characterized. Localization of metal sequestration was ascertained employing transmission electron microscopy and X ray microanalysis. Development of an immobilized biomass system was emphasized. 2.

MATERIALS AND METHODS

2.1 Microorganism, growth medium and culture conditions Pseudomonas sp., isolated from a garden soil was grown and maintained in Trisminimal medium [13]. Mid exponential phase cells (culture O.D 0.6 at 600 nm) were collected by centrifugation (12000 × g, 30 min), washed thoroughly with distilled water, lyophilized and used for biosorption experiments. 2.2 Uranium and thorium biosorption experiments Except where otherwise described, for all biosorption experiments, 50 mg (dry wt.) of lyophilized biomass was contacted with 100 ml of a 100 mg uranium or thorium L-1 solution (as nitrate, UO2(NO3)2.6H2O or Th(NO3)2.5H2O, Merck, Germany). Experimental details are same as described earlier [5-6]. The biosorption equilibrium uptake (q, mg metal g-1 biomass dry wt.) for each sample was calculated according to the mass balance on metal ion expressed as: 976

Biosorption

q = V(Co - Ce) / M (1) . -1 where V is the sample volume (L), Co, the initial metal ion concentration (mg L ), Ce, the equilibrium or final metal concentration (mg.L-1), and M, the biomass dry weight (g). For sorption kinetic studies, samples were withdrawn at timed intervals from biomass-metal mixture, centrifuged, and finally dissolved U/Th was estimated. Role of pH was studied by adjusting the initial pH of the contact solution (100 mg U / Th L-1) over the pH range 2.0-8.0. Biosorption in simultaneous presence of other ions was tested in bimetallic combinations, by adding equimolar concentrations of uranium or thorium (430 µM Th or 420 µM U; equivalent to 100 mg U or Th L-1) and the test cation/anion. Experimental details are same as described earlier [5-6]. 2.3 Transmission electron microscopy (TEM) and X ray microanalysis For TEM studies, ultra-thin sections of radionuclide -free and -loaded bacterial cells were used. Experimental details are same as described previously [15]. Energy dispersive X ray analysis (EDXA) of samples was done using a Link Oxford ISIS EDX system. 2.4 Desorption of sorbed metal For desorption experiments, metal loaded biomass was contacted with respective desorbing agents (0.5 mg biomass mL-1). All other conditions were same as sorption experiment. 2.5 Immobilization of biomass Lyophilized Pseudomonas biomass was immobilized in radiation-polymerized polyacrylamide matrix as described earlier [16.]. The beads obtained were washed, resuspended in distilled water and stored at 0-4°C. 2.6 Uranium sorption by immobilized biomass Immobilized biomass (bio beads) were contacted with 100 ml uranyl nitrate (UO2(NO3)2,.6H2O) solution (100 mg U L-1) in a 250 ml conical flask (150 rpm, 30°C, 24 h). Following contact the beads were sieved out and the solution was filtered through a millipore filter (0.25 µm) and analyzed for uranium content. Initial pH of all U solutions was adjusted to 4.0 with the addition of 1M NaOH or 1M HNO3. In each set biomass free polyacrylamide beads were kept as control. Dissolved uranium was estimated by the methods described earlier. All data represents the mean of three independent experiments. Standard Deviations and error bars are indicated wherever necessary. All statistical analyses were done using Microcal Origin, Version 5.0. 3.

RESULTS AND DISCUSSION

3.1 Selection of radionuclide accumulating strain A number of metal tolerant bacterial strains were tested for their uranium accumulating capacity (Table 1). When exposed to an initial concentration of 100 mg U l1 , a copper resistant Pseudomonas sp.2 strain was found the best; accumulating a maximum of 63 mg U g-1 biomass dry wt followed by a cobalt resistant Bacillus coagulans strain with the loading capacity of 48 mg U g-1 dry wt. Based on this, the Pseudomonas sp. 2 strain was selected for further study on biosorption of uranium (VI) and thorium (IV). 977

Biosorption

Table 1. Selection of uranium accumulating bacterial strains U uptake (mg.g-1 dry wt.) 39 ± 1.4 45 ± 2.7 48 ± 2.9 63 ± 3.1

Bacterial strain

Pseudomonas sp. 1 P. aeruginosa Bacillus coagulans Pseudomonas sp. 2

3.2 Effect of growth medium, energy source and metabolic inhibitor on uranium biosorption To test Pseudomonas strain, cells were pre-grown in synthetic minimal- and enrichedmedium and tested for their U sorption capacity. At low uranium concentration (100 mg.L1 ), comparable metal sorption was observed by both enriched (63 mg.g-1 dry wt.) and minimal (60 mg.g-1 dry wt.) medium grown cells. However, improved (1.9-fold) uranium loading was found for minimal medium grown cells (245 mg.g-1 dry wt.) at higher uranium concentration (1000 mg.L-1). The present data corroborate very well with U sorption by P. aeruginosa CSU strain [11]. The energy dependency of U uptake by the biomass was tested by adding glucose (as carbon/energy source) or sodium azide (as metabolic inhibitor) in uranium uptake solution (100 mg.L-1). An insignificant difference in U uptake indicates the metabolic independency of the tested biomass in sequestering uranium. 3.3 Time course of uranium and thorium sorption The kinetics of uranium and thorium sorption by lyophilized biomass is shown in Fig. 1. For both radionuclides, the biomass exhibited a rapid cation uptake and more than 90% of equilibrium loading was reached within one (for Th) or ten (for U) minutes. In addition the process saturates after 2 (for U) or 4 (for Th) hours. This rapid binding of metal cations by microbial biomass is typical for radionuclide sorption as it was also shown earlier by P. aeruginosa [11], Rhizopus arrhizus [9, 10], Mycobacterium smegmatis [17]. The rapid cation uptake has been suggested as being essential for any good biosorbent as it allows short solution-sorbent contact time and would result in the use of much shallower contact beds of sorbent materials in column application [18]. 500

200

500

150

375

200

-1

mg metal g dry wt.

-1

mg metal g dry wt.

150

100

100

50

50

0

250

250

125

125

0

0

0 0

3

6

3

6

9 9

Time (h)

Figure 1. Time course of U (h) and Th (t) sorption (100 mg L-1) by Pseudomonas biomass 978

375

0

1 1

2 2

3 3

4 pH 4

5

6

7

5

6

7

Figure 2. Effect of pH on U (h) and Th (t) sorption (100 mg L-1) by Pseudomonas biomass

Biosorption

3.4 Effect of pH on uranium and thorium biosorption Initial solution pH plays a critical role in metal sorption by influencing both the bacterial surface chemistry as well as the chemistry of soluble metal ions. Uranium and thorium sorption by lyophilized Pseudomonas biomass was studied at a range of pH between 2 and 8 and 2 and 6, respectively (Fig. 2). It was observed that initial solution pH strongly affected the U and Th equilibrium loading capacity. Extreme acid condition (pH 2.0) did not favour sorption of both cations. As the pH increased, sorption of U and Th also increased and the maximum loading for both was attained at pH 4.0 and 5.0, respectively. Increase in pH beyond the optimum caused decline in sorption of respective cations. The reduced sorption at low pH could be attributed to: (i) the hydrolysis of biomass metal binding groups resulting an increased competition by H3O+, and (ii) the increased solubility and consequent reduced adsorptivity of thorium ions [19]. Furthermore, compared to the Th4+ and Th(OH)22+ ions formed at low pH that have been identified as a poor sorbate [10], the higher uptake at pH 4.0 could be correlated to the predominance of [Th2(OH)2]6+ and other polymerized species possessing a greater binding affinity and thus facilitating faster and enhanced metal sorption [20]. For uranium, the observed trend with regard to pH may be explained by an increasing binding affinity of monovalent uranyl species (UO2OH+, (UO2)3(OH)5+) formed at higher pH (pH 4.0-5.0) over the divalent (UO22+) at low pH (pH 2.0) [21]. 3.5 Biosorption isotherm The biosorptive U and Th uptake by Pseudomonas biomass was quantitatively evaluated by equilibrium sorption isotherms over a concentration range of 0-1200 mg.L-1 (Fig. 3). Representative isotherm curves for both cations exhibited very efficient metal binding, even at low residual concentration, and a high loading at equilibrium. The maximum sorption values obtained were 541 mg uranium g-1 dry wt. and 430 mg thorium g-1 dry wt. at an equilibrium concentration of 359 mg U L-1 and 885 mg Th L-1, respectively. Such impressive U and Th binding by the tested biomass significantly surpasses the economic threshold level (15% dry wt. basis) for practically usable biosorbents and also the previous values on Th [R. arrhizus (185 mg.g-1 dry wt.) [9] or P. chrysogenum (388 mg g-1 dry wt.) [20]] and U [R.. arrhizus and Penicillium chrysogenum (both 180 mg g-1) [9] P. aeruginosa CSU (110 mg.g-1) [11], and M. smegmatis (44.5 mg.g-1) [17]] uptake. 600 450 450

-1

mg metal g dry wt.

600

300 300

150 150

0

0 0

0

200 200

400 400

600 600

800 1000 800

1000

-1

Equilibrium concentration (mg l )

Figure 3. U (h) and Th (t) sorption isotherm for Pseudomonas biomass 979

Biosorption

Table 2. Freundlich and Langmuir constants for uranium and thorium sorption by Pseudomonas biomass

Freundlich

Langmuir

Uranium 199.00 0.206 0.931 555.5 0.0027 0.997

k 1/n r qmax b r

Thorium 159.20 0.176 0.973 476.19 0.0009 0.998

The relationship between equilibrium metal uptake capacity (q) and residual metal ion concentration (Ce) was further described using the model equation of Freundlich and Langmuir. Although linearized sorption isotherm for both the metals showed a reasonably good fit to both the models, the maximum correlation coefficient (r) was obtained with Langmuir equation. Values of respective sorption constants and correlation coefficients (r) are presented in Table 2. The better fitting of Langmuir model suggests a monolayerd U and Th binding on to the biomass [22]. The asymptotic maximum adsorption capacity as predicted by the Langmuir constant ‘qmax’ gives a very high value for U and Th, while a desirable high affinity of the biomass for test metals are evident from the low values of other constant ‘b’. 3.6 Effect of interfering ions on uranium and thorium biosorption Uranium and thorium sorption by Pseudomonas biomass was studied in the presence of equimolar amount of several competing ions (Table 3). Among the series of cations tested, a significant antagonism in U sorption was offered only by thorium (IV), iron (II and III), aluminium (III) and copper (II) while metals like cadmium (II), lead (II) silver (II), and anions like chloride (I), phosphate (II) and sulphate (II) had no effect. The order of inhibition to uranium binding by the cations was Fe3+>Th4+>Fe2+>Cu2+>Al2+. Iron (III), the cation considered as the most potent competitor of uranium for binding sites [11], also caused a severe decline (80%) in U loading. Noticeably, in case of thorium, except iron (III), no other tested cation showed an inhibition more than 20%. The order of inhibition by other cations was UO22+
Control Cations Na+ Ag2+ K+ Ca2+ Pb2+ Cd2+ 980

Percentage of sorption Uranium Thorium 100 100 100 98 98 98 98 96

100 84* 86* 100 100 100

Biosorption

Cu2+ Al3+ Fe2+ Fe3+ Th4+ UO22+ Anions ClPO42SO42CO32-

Percentage of sorption Uranium Thorium * 78 83* 82* 85* * 45 57* 20* 60* 37* 92 100 100 100 74*

100 100 100

a

Initial U/Th concentration 100 mg.L-1, pH 3.5, biomass 0.5 mg.mL-1. Indicates a significant difference (at 5% level) between control (U or Th alone) vs samples containing the competing ions along with U or Th. Significance is calculated using the absolute values (means) in statistical test

*

3.7 Transmission electron microscopy and X-ray microanalysis Transmission electron micrographs of both uranium and thorium loaded cells revealed a dark electron opaque region throughout the cytoplasm, which indicates the intracellular sequestration of biosorbed metals (Fig. 4). Conclusive identification of the deposited elements was achieved with energy dispersive X ray analysis (EDXA). Compared to the metal free control sample, presence of specific peaks for uranium and thorium in respective metal loaded samples confirmed the presence of accumulated radionuclides (Fig. 5). The present observations on cytosolic U and Th sequestration are in conformity with similar reports on P. aeruginosa [23] and M. smegmatis [24]. The intracellular U and Th deposition even by the lyophilized cells as observed in the present study could be attributed to the increased membrane permeability of these highly toxic radionuclides [25]. Once penetrates the cell membrane, these cations are often immobilized by binding to anionic sites in the cytoplasm or by precipitation due to change in solution chemistry (e.g. pH, phosphate concentration etc.). Further analyses of U and Th loaded biomass using infrared spectroscopy and 31P NMR techniques indicated the possible involvement of cellular phosphoryl groups in sequestering the biosorbed metals. X ray powder diffraction patterns of metal –free and –loaded biomass ascertained the sequestration of U and Th as phosphate compounds. 3.8 Desorption of bound uranium and thorium Recovery of sorbed metal is one of the most important aspects of any successful biosorption process development. In the present study, although mineral acids proved quite effective (>70% recovery), maximum amount of uranium or thorium could be recovered by carbonates of sodium (98%) and calcium (93%), respectively (Table 4). Noticeably, compared to the effective desorption of uranium or thorium (>90%) from other microbial biomass by only 10 mM carbonates (sodium or calcium), the high concentrations required (1M) to remobilize the sorbed metals in the present case indicates a relatively high intensity of bonding for these element with intracellular binding ligands. The observed effectiveness of sodium carbonate over mineral acids in recovering biosorbed uranium had also been reported earlier with different bacterial and fungal biomass [24, 26, 27]. As evident from the present data, use of carbonates as desorbent seems more useful in developing a cost effective 981

Biosorption

desorption technology as they will be less destructive to the biomass over mineral acids and therefore will favour the overall process economics.

Figure 4. Transmission electron micrographs of (a) metal-free (control), (b) U- and (c) Th- loaded Pseudomonas cells. Magnification x 45000

Figure 5. Energy dispersive X-ray analysis of (a) metal-free (control), (b) U- and (c) Th-loaded Pseudomonas cells

Table 5. Recovery of sorbed U and Th by various desorbents Desorbents HCl (1M) HNO3 (1M) H2SO4 (1M) CaCO3 (1M) Na2CO3 (1M) EDTA (0.01 M)

982

Amount of metal desorbed (percent of sorbed metal) Uranium Thorium 70 80 72 80 70 91 5 93 98 83 20 36

-1

U concentration in solution (mg l )

Biosorption

100

100

80

80

60

60

40

40

20

20

0

0 0

0

2

2

4

4

6

6

8 10 12 14 8

10

12

Time (h)

Figure 6. Uranium sorption kinetics by immobilized beads with (h) or without (t) Pseudomonas cells 3.9 Radionuclide sorption by immobilized Pseudomonas biomass In view of several disadvantages associated with the use of free microbial biomass in continuous effluent treatment processes [26], Pseudomonas biomass was immobilized in a radiation polymerized acrylamide matrix (bio-bead).

Figure 7. Scanning electron micrographs of radiation polymerized matrix. (a) surface view of control (biomass free)- and (c) with biomass bead. (b) cross section of control- and (d) with biomass-bead

The biosorption performance of such bio-beads was evaluated using a batch reactor system. As shown in Fig. 6, the uranium sorption kinetics of the immobilized biomass was much slower and till 12h of contact no steady equilibrium state was reached. When compared with its free counterpart (ref. fig. 1), such slow kinetics seems to be a 983

Biosorption

characteristic feature of immobilized biomass resulting from mass transfer limitation in the particle interior as also observed for radium biosorption by an entrapped fungal biomass [28]. Scanning electron microscope (SEM) (Fig. 7) elucidated a desirable highly porous nature of the bio-beads with bacterial cells entrapped over the pore walls. It was important to note that lyophilized bacterial cells retained their normal shape even in immobilized state. ACKNOWLEDGEMENTS Pinaki Sar and Sufia K. Kazy acknowledge the financial assistance made by Board of Research in Nuclear Sciences, Department of Atomic Energy, Govt. of India, and Council of Scientific and Industrial Research, Govt. of India, in the form of post doctoral fellowship. REFERENCES

1. T. Barkay and J. Schaefer, Curr. Op. Microbiol.., 4 (2001) 318. 2. J. R. Lloyd and L. E. Macaskie, In: Lovely DR, (eds.), Environmental Microbe-Metal Interactions, Washington DC, 2000. 3. M. Valls and V. de Lorenzo, FEMS Microbiol. Rev., 26 (2002) 327. 4. G. M. Gadd. Curr. Op. Biotechnol., 11 (2000) 271. 5. P. Sar and S. F. D'Souza, J. Chem. Tech. Biotechnol., 76 (2001) 1286. 6. P. Sar and S. F. D'Souza, Biotechnol. Lett., 24 (2002) 239. 7. M. Tsezos, E. Remoudaki and V. Angelatou, Int. Biodeterior. Biodeg., 32 (1995) 129. 8. M. B. Cassidy, H. Lee and J. T. Trevors, J. Ind. Microbiol., 16 (1996) 79. 9. M. Tsezos and B. Volesky, Biotechnol. Bioeng., 23 (1981) 583. 10. M. Tsezos and B. Volesky, Biotechnol. Bioeng., 24 (1982) 955. 11. M. Z. C. Hu, J. M. Norman, B. D. Faison and M. Reeves, Biotechnol. Bioeng., 51(1996) 237. 12. P. Sar, S. K. Kazy, R. K. Asthana and S. P. Singh, Curr. Microbiol., 37 (1998) 306. 13. S. K. Kazy, P. Sar, R. K. Asthana and S. P. Singh, W J. Microbiol. Biotechnol., 15(1999) 599. 14. P. Sar, S. K. Kazy, R. K. Asthana and S. P. Singh Sar, Int. Biodeterior. Biodeg., 44 (1999) 101. 15. P. Sar, S. K. Kazy and S. P. Singh, Lett. Appl. Microbiol., 32 (2001) 257. 16. A. Deshpande, S. F. D’Souza and G. B. Nadkani, J. Biosci., 11 (1987) 137 17. Y. Andres, H. J. MacCordick and J. C. Hubert, Appl. Microbiol. Biotechnol., 39 (1993) 413. 18. B. Volesky, Biosorption of heavy metals, CRC Press, Boca Raton, FL, 1990. 19. C.F. Beas Jr. and R. E. Mesmer, Hydrolysis of cations. Wiley, New York, 1976. 20. G. M. Gadd and C. White, Environ. Poll., 61 (1989) 187. 21. Y. E. Collins and G. Stotzky, Appl. Environ. Microbiol., 58 (1992) 1592. 22. I. Langmuir, J. Am. Chem. Soc., 40 (1918) 1361. 23. G. W. Strandberg, S. E. Shumate II and J.R. Parrot Jr., Appl. Environ. Microbiol., 41 (1981) 237. 24. M. Marques, X. Roca, S. M. Dolores, M. C. Fuste and F. Congregado, Appl. Microbiol. Biotechnol., 35 (1991) 406. 25. Y. Suzuki and J. F. Banfield,. In: Burns, P. C., Finch, R., (Eds.), Uranium-mineralogy, geochemistry and the environment, Mineralogical society of America, 2000. 26. M. Z. C. Hu and M. Reeves, Biotechnol. Prog., 13 (1997) 60. 27. A. Nakajima and T. Sakaguchi, Appl. Microbiol. Biotechnol., 38 (1993) 574. 984

Biosorption

28. M. Tsezos, M. H. I. Baird and L. W. Shemilt, Hydrometallurgy, 17 (1987) 357.

985

C HAPTER 4 Microbiology Fundamentals

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

A model for iron uptake in Acidithiobacillus ferrooxidans based upon genome analysis R. Quatrini1, F. Veloso1,3, E. Jedlicki2 and D.S. Holmes1,3* 1

Laboratory of Bioinformatics & Genome Biology, University of Santiago, Santiago, Chile 2 Program of Cellular and Molecular Biology, ICBM, University of Chile, Santiago, Chile 3 Millenium Institute of Fundamental and Applied Biology, Santiago, Chile Abstract During growth at acid pH, Acidithiobacillus ferrooxidans is confronted with an abundant supply of soluble iron. Therefore, it may have developed iron uptake mechanisms quite distinct from those of neutrophilic organisms who are faced with the problem of the uptake of insoluble forms of iron or of soluble iron at very low concentrations. An analysis of the nearly complete genome sequence of A. ferrooxidans reveals the presence of a number of potential genes and regulatory pathways involved in iron uptake typical of those found in other organisms. However, in addition, bioinformatic analysis suggests that A. ferrooxidans exhibits several unusual genetic features that may reflect special requirements for iron uptake and homeostasis required for life at acidic pH. Finally, an analysis of iron uptake mechanisms in A. ferrooxidans supports the growing contention that the organism can live at a variety pHs between pH1 and, at least, pH 5.5. Keywords: Acidithiobacillus ferrooxidans, iron uptake, iron regulation, FeoB, TonBdependent receptors, genome analysis 1.

INTRODUCTION One of the critical problems confronting all organisms is the acquisition and uptake of iron to fulfill metabolic requirements. Because iron is largely insoluble at neutral pH under aerobic conditions, neutrophilic organisms have evolved a variety of sophisticated iron uptake mechanisms. In addition, they have developed homeostatic controls to tightly balance iron intracellular levels, guarding cell integrity against the deleterious effects of excess iron (1,2). The regulation of iron metabolism and its coupling with the regulation of defences against oxidative stress must be strictly adjusted in response to environmental iron bioavailability. Judging by the plethora of genes and response mechanisms involved, iron

*

Corresponding author: David Holmes: [email protected]. Work supported by Fondecyt No. 1010623 and the Millenium Institute of Fundamental and Applied Biology, Santiago, Chile. We thank the Institute of Genome Research (TIGR) and Integrated Genomics, Inc. (IG) for the use of their partial sequence of the Acidithiobacillus ferrooxidans genome.

989

Microbiology Fundamentals

metabolism and homeostasis appear to be a major concern for life in the presence of oxygen. We are interested in understanding how the bacterium Acidithiobacillus ferrooxidans copes with iron uptake and homeostasis considering that it can grow aerobically at acidic pHs in environmental situations where soluble iron is abundant. In addition, a special feature of A. ferrooxidans, not found in the majority of organisms, is its use of iron as energy and electron sources. This utilization must be balanced with the need for iron in metabolism suggesting that A. ferrooxidans might exhibit novel iron regulatory mechanisms not observed in neutrophilic heterotrophic organisms. Contrary to what happens in most aerobic environments, where FeII rapidly autooxidizes to FeIII which precipitates as ferric hydroxides, in acidic conditions FeII is relatively stable and FeIII is much more soluble (>0.1M), exceeding by far the typical requirements of bacteria of 10-8M (3,4). Given that A. ferrooxidans is usually confronted with an abundant supply of both FeII and FeIII at acidic pHs, one can ask if it has evolved novel mechanisms for Fe uptake that are not present or are substantially different from those exhibited by neutrophilic organisms. One can also ask if this abundant supply of Fe has necessitated the development of special or different mechanisms in A. ferrooxidans for iron storage and for the avoidance of oxidative stress due to excess iron. Studies concerning the strategies used by A. ferrooxidans for iron assimilation, iron homeostasis and avoidance of the oxidative stress threat imposed by a unique iron-rich and O2-plentiful environment are lacking and prompted the present investigation. Knowledge of the physiology and metabolism of A. ferrooxidans has been impeded by the paucity of classical genetic and molecular biological tools that have been so successfully used to unravel the metabolism of other organisms. Due to its applied interest for biomining and to circumvent this limitation, A. ferrooxidans was the first biomining microorganism to have its genome almost completely sequenced (5) making it a candidate for bioinformatic gene predictions and subsequent functional and metabolic model building (6). In this paper, we present a partial bioinformatic survey of potential genes of A. ferrooxidans involved in iron uptake. Candidate genes and pathways were first identified by comparison to known iron uptake genes and pathways in other organisms. Additional candidate genes were then identified by linkage analysis to the previously known genes. Contributing to the elucidation of the physiological responses of A. ferrooxidans is of fundamental interest and might, eventually, have a biotechnological impact on the mining industry. 2.

METHODS Known metabolic pathways of iron acquisition were obtained from BIOCYC (http://biocyc.org:1555/META/server.html), KEGG (http://genome.ad.jp/kegg/kegg4. html) and ERGO (http://wit.integratedgenomics.com/WIT2/). Amino acid sequences derived from genes identified as being involved in iron acquisition were used as query sequences to search the partial genome sequence of A. ferrooxidans ATCC 23270 in the TIGR (http://www.tigr.org/) and ERGO data bases using TBlastN and BlastP respectively. When a prospective candidate gene was identified in TIGR or ERGO its predicted amino acid sequence was then used to formulate a BlastP (http://www.ncbi.nlm.nih.gov/BlastP/) search of the non-redundant database at NCBI. Only bidirectional best hits were accepted as evidence for putative homologs. Candidate genes and their translated proteins were further characterized employing the following bioinformatic tools available in the web: primary structure similarities (http://www.ebi.ac.uk/ClustalW/), secondary structure 990

Microbiology Fundamentals

predictions (HMM-based Protein Sequence Analysis http://www.cse.ucsc.edu/research/ compbio/HMM-apps/T99-query.html; JPred http://www.compbio.dundee.ac.uk/Software /JPred/jpred.html), transmembrane predictions (http://www.ch.embnet.org/software/ TMPRED_form.html), motif predictions (http://www.blocks.fhcrc.org/, http://www.ebi. ac.uk/interpro domain predictions (http://www.tolouse.inra.fr/prodom.htlm) and prediction of protein localization sites (http://psort.nibb.ac.jp/). 3.

RESULTS AND DISCUSSION Bioinformatic similarity searches for orthologous genes (evolutionarily homologous gene with the same function in different species) in A. ferrooxidans involved in iron uptake has revealed a number of candidate genes that can be assigned with a reasonable level of confidence, due to their similarity with experimentally proven iron uptake genes in other organisms (Table 1). Given the presence of such genes, metabolic models of iron uptake in A. ferrooxidans can be inferred based upon known models in other organisms. A summary of these results is shown in Figure 1. Additional work is in progress to identify genes and pathways involved in iron sensing, homeostasis, storage and detoxification. A. ferrooxidans encodes a putative FeII inner-membrane transport system termed FeoAB (Figure 1A). FeoAB is present in most of the sequenced genomes but has been observed to function only under anaerobic conditions where FeII is more stable and soluble (7-9). As in other microorganisms, the feoB gene in A. ferrooxidans is adjacent to a small open reading frame (ORF) termed feoA, of unknown function. Interestingly, in A. ferrooxidans feoA and feoB appear to constitute an operon together with an ORF encoding a putative porin (herein designated porA), which could facilitate FeII entrance in the periplasm. Upstream of this porin is a putative Fur-box, suggesting typical iron-dependent repression of FeII uptake mediated by a Fur-like repressor as has been demonstrated in Escherichia coli (10). Bioinformatic analysis of the putative FeoB reveals the presence of a highly conserved N-terminal GTPase motif and eight hydrophobic-rich transmembrane segments supporting the contention that it is a membrane transport protein. In addition, the second and third periplasmic loops have motifs similar to cytochrome CB5 and B6 which could represent iron recognition or iron interacting portions of the protein. Sequence comparison of the putative A. ferrooxidans FeoB ortholog and the other three FeoB proteins so far characterized (GenBank accession #s:P33650, P74884, D71909) reveals substantial amino acid sequence similarity except in the so-called connector loop and in the C-terminal half of the protein. FeoB is an inner membrane protein with several exposed loops facing the periplasm and it is these exposed loops that harbour the major differences between the FeoB of A. ferrooxidans and the other three orthologs of FeoB. We speculate that these differences evolved as a response to the need for function and structural integrity of the A. ferrooxidans FeoB in a low pH environment. In addition, FeoB proteins of A. ferrooxidans and Helicobacter pylori have predicted periplasmic loops with higher isoelectric points than E. coli. This could be a reflection of the functional adaptation of these proteins to acidophilic or acid tolerant conditions. It has recently been demonstrated that the FeoB of H. pylori plays an important role in iron acquisition in the low-pH and low-oxygen environment of the stomach (7). We propose that the FeoB of A. ferrooxidans may serve a similar function in iron uptake at low pH. Inspection of the A. ferrooxidans genome also reveals the presence of several putative outer membrane iron receptors (OMRs) (Figure 1B) with predicted affinities for either hydroxamate-type siderophores or for FeIII-dicitrate. The degree of primary structure 991

Microbiology Fundamentals

similarity and the general accordance of secondary structure predictions between the FhuA (FeIII-hydroxamate siderophores uptake) and FecA (FeIII-dicitrate uptake) of A. ferrooxidans and E. coli highlights the conserved organization of the proteins. These similarities include similar signal peptides, 2 to 4 N-terminal α-helices, 22 β-strands in the putative β-barrel domain and a TonB interaction C-terminal motif (11). Several of the genes encoding OMRs are next to genes encoding TonB energy-transducing systems (exbBDtonB) and some of them cluster together with all the necessary elements (OMR, TonB system, ABC transporter components) to transport FeIII iron into the cytoplasm (Figure 2) as observed in other organisms (12). The conservation of the siderophore receptor proteins and organization of their genes into putative operons together with relevant genes encoding other necessary components for iron uptake, allows us to propose that mechanisms of siderophore uptake exist in A. ferrooxidans that are similar to those found in other organisms.

Figure 1. Proposed model for iron uptake and storage A. ferrooxidans based upon a bioinformatics analysis of its genome. (A) FeII uptake via an outer membrane porin (black) and inner membrane FeoAB transporter complex (white). (B) FeIII uptake via an outer membrane TonB dependent receptor OMR (black) and a periplasmic siderophore binding and inner membrane uptake mechanism involving ABC transporter Fec-like proteins (white) and a TonB energy transducing system (dark gray). (C) Cytoplasmic FeII storage via a bacterioferritin-like protein Bfr. Interestingly, the several OMRs of A. ferrooxidans exhibit a wide range of isoelectric points (Table 1) that, by comparison to the OMRs of other organisms, would potentially allow them to import iron at different pHs ranging from pH 1 to at least pH 5.5. This supports the growing contention that A. ferrooxidans can live in various environments with marked differences in pH (13, 14). It is not known how A. ferrooxidans might regulate and coordinate the expression of these different OMRs to correspond to particular environmental requirements for iron uptake. An interesting aspect of the organization of the proposed FeIII-siderophore transport operon in A. ferrooxidans is the presence of a small ORF potentially encoding a hemoglobin-like protein (Figure 2). A bioinformatic analysis suggests that it belongs to a 992

Microbiology Fundamentals

family of genes known as truncated hemoglobins, identified in prokaryotes, protozoa, algae and plants. The role of this hemoglobin-like protein is still unclear, although it has been shown in plants to have low oxygen affinity and that the expression of its mRNA decreases in microaerophilic conditions (15). The hemoglobin-like gene is associated with other transporters annotated in genome database of ERGO in Burkholderia cepacia (downstream of a putative molybdenum transport), Ralstonia eutropha (downstream of an uncharacterized permease) and Sphingomonas aromaticivorans (JGI) (downstream of an OMR). In addition to transporting and storing O2 and facilitating its diffusion, several novel functions have emerged regarding hemoglobins, including control of nitric oxide levels in microorganisms, binding and transport of sulfide in endosymbiont-harboring species and protection against sulfide, scavenging of O2 in symbiotic leguminous plants, O2 sensing in bacteria and archaebacteria, and dehaloperoxidase activity useful in detoxification of chlorinated materials (16). In the case of A. ferrooxidans, its association with a siderophore operon might represent a strategy to diminish the risks of O2 toxicity caused by the entering iron until the metal has been targeted to its final destination. Table 1. Identification and properties of candidate genes in A. ferrooxidans involved in iron uptake. (A) name of genes involved in FeII uptake, FeIII-hydroxamate uptake and FeIII-dicitrate uptake. (B) The % similarity of the predicted proteins from the candidate genes to known proteins deposited in data banks. (C) lists the GenBank accession number of the known proteins and (D) the organisms from which they are derived. (E) lists the predicted location (loc) of the proteins where C: Cytoplasm, OM: Outer Membrane and IM: Inner Membrane. (F) : predicted isoelectric point (pI) of the mature proteins. (A) Gene

(B) %S

(C) Accesion #

(D) Organism

(E) Lo

(F) pI

FeII uptake

porA feoA feoB

42 55 53

ZP00053838 CAB49974 BAC09292

Magnetospirillum magnetotacticum Pyrococcus abyssi Thermosynechococcus elongatus

FeIII hydroxamate uptake

omr1 omr4 omr6 omr7 omr9 omr11

50 57 37 38 37 37

AAM72768 AAM43451 ZP00125990 ZP00003769 AAM35906 AAM42549

Chlorobium tepidum Xanthomonas campestris Pseudomonas syringae Nitrosomonas europaea Xanthomonas axonopodis Xanthomonas campestris

OM OM OM OM OM OM

6.98 8.18 9.02 9.15 6.53 8.63

FeIII dicitrate uptake

omr2 omr3 omr10

37 38 38

ZP00125403 ZP00125403 AAM35740

Pseudomonas syringae Pseudomonas syringae Xanthomonas axonopodis

OM OM OM

7.26 9.02 8.22

OM 5.85 C 11.14 IM 7.28

In accordance with what has been described for several siderophore transporting genes in other bacteria this putative operon could be iron regulated, since a well conserved Fur-box like sequence has been identified within the predicted promoter region of the globin-like gene (unpublished data). The existence in A. ferrooxidans of an gene with 79% similarity to the E. coli Fur ortholog (unpublished data) strengthens this interpretation and promulgates further intriguing questions about iron regulation in this organism.

993

Microbiology Fundamentals

Figure 2. Gene organization of the proposed FeIII-siderophore uptake region of the A. ferrooxidans genome. A: Globin-like gene; B: Outer Membrane TonB-dependent Receptor I; C: TonB energy transducing system (exbB-exbD-tonB); D: ABC transporter I Fhu-like proteins (perplasmic binding protein, permease, ATP-binding protein); E: Outer Membrane TonB-dependent Receptor II and F: ABC transporter II Fec-like proteins (perplasmic binding protein, permease, ATP-binding protein). Also shown is the putative 19 nt Fur–box. In the upper right corner is a scale bar in base pairs (bp) 4.

CONCLUSIONS A. ferrooxidans seems to have a complexity of iron uptake systems that rivals that found in neutrophilic organisms. This argues against a simplistic view that A. ferrooxidans has an easy time encountering and taking up readily available soluble iron in an acid environment. It is possible that it requires this complexity despite the readily available iron. However, a more reasonable argument is that it sometimes lives at higher pHs than those typically associated with its growth in acid ferrous sulfate and sulfur, in which it must scavenge less soluble forms of iron perhaps in strong competition with other organisms. Notable is the plethora of siderophore uptake systems of A. ferrooxidans characteristic of most organisms that have to scavenge and compete for poorly soluble FeIII at neutral pHs. Demonstrations of the capacity of A. ferrooxidans to live at higher pHs have appeared in the literature (13,14) and our findings, regarding iron acquisition systems, support this point of view. On the other hand, A. ferrooxidans iron uptake systems include components that seem to be adapted for function at low pH, most obvious are the higher isoelectric points of the periplasmic loops of membrane-associated iron uptake pumps. This characteristic is expected for the maintenance of structural integrity and possible function of proteins exposed to low pH. The presence of a hemoglobin-like protein in A. ferrooxidans possibly associated with a FeIII siderophore uptake operon is unusual, although descriptions of a similar protein are now appearing in the literature. The best guess for its function is that it protects cells from excess oxygen. Alternatively, and not mutually exclusive with this hypothesis, is the possibility that it may measure oxygen levels and regulate genes in response to the changes in oxygen concentration. In this paper, we have focused on iron uptake mechanisms. Future studies will address mechanisms to sense environmental levels of iron and the regulation of expression of iron uptake and storage genes and how these might be coupled to genes involved in the important capacity of A. ferrooxidans to oxidize iron for energy and electron gain. Preliminary evidence points to the presence of a complex network of Fur regulated operons and experimental work is underway to validate some of the bioinformatic models we have developed.

994

Microbiology Fundamentals

REFERENCES 1. J.M. De Freitas and R. Meneghini. Mutat. Res., 475 (2001) 153. 2. D. Touati, M. Jacques, B. Tardat, L. Bouchard and S. Despied. J. Bacteriol., 177 (1995) 2305. 3. K. Hantke and V. Braun. In: Bacterial Stress Responses. (Chapter 19) G. Storz & R. Hengge-Aronis. (Eds). ASM Press, Washington, D.C., 2000. 4. V. Braun and H. Killmann. Trends. Biochem. Sci., 24 (1999) 104. 5. E. Selkov, R. Overbeek, Y. Kogan, L. Chu, V. Vonstein, D.S. Holmes, S. Silver, R. Haelkorn & M. Fonstein. Proc. Natl. Acad. Sci. USA, 97 (2000) 3509. 6. D.S. Holmes, J. Valdéz, C. Dominguez, M. Barreto, C. Arriagada, S. Silver, S. Bueno and E. Jedlicki. International Biohydrometallurgy Symposium (2001) 237. 7. J. Velayudhan, N.J. Hughes, A.A. McColm, J. Bagshaw, C.L. Clayton, C.S. Andrews and D.J. Kelly. Mol. Microbiol., 37 (2000) 274. 8. R.M. Tsolis, A.J. Baumler, F. Heffron and I. Stojiljkovic. Infect. Immun., 64 (1996) 4549. 9. I. Stojiljkovic, M. Cobeljic and K. Hantke. FEMS Microbiol. Lett., 108 (1993) 111. 10. M. Kammler, C. Schön and K. Hantke. J. Bacteriol., 175 (1993) 6212. 11. R. Koebnik, K.P. Locher and P. van Gelder. Mol. Microbiol., 37 (2000) 239. 12. J.H. Crosa. Microbiol. Molec. Biol. Rev., 61(1997) 319. 13. N. Ohmura, K. Sasaki, N. Masumoto and H. Saiki. J. Bacteriol., 184 (2002) 2081. 14. M. Barreto, M. Rivas, E. Jedlicki and D.S. Holmes. International Biohydrometallurgy Symposium (2003) (submitted, the Proceedings of this Congress). 15. R.A. Watts, P.W. Hunt, A.N. Hvitved, M.S. Hargrove, W.L. Peacock and E.S. Dennis Proc. Natl. Acad. Sci. USA, 98 (2001) 10119. 16. R.E. Weber and S.N. Vinogradov. Physiol. Rev., 81(2001) 569.

995

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Activity and occurrence of leaching bacteria in mine waste at Cartagena, Spain, in the years 1991 until 2000 W. Sand, D. El Korchi-Buchert, T. Rohwerder* Department of Microbiology, Institute for General Botany, University of Hamburg, Ohnhorststr. 18, D-22609 Hamburg, Germany, Tel/Fax: +49 40 428 16 423, e-mail: [email protected] Abstract During the years 1991 until 2000 a total of 31 sampling campaigns have been performed in the abandoned mining area between Portman and La Union near the city of Cartagena in the province Murcia, Spain. Sampling was done at mine waste heaps near and in the open pit Brunita, in the open pit Gloria, and at a rock formation with an outcropping pyrite vein. More than 340 samples have been taken and analyzed for numbers of leaching bacteria, microbial activity, and occurrence of sulfur compounds. Besides, physicochemical factors such as pH, temperature, redox potential etc. were measured. The statistical evaluation of all these data yielded the following results: (i) The area is a typical habitat for acidophilic leaching bacteria. Consequently, a significant positive correlation between the proton concentration and the numbers of leaching bacteria, the concentrations of leaching products (iron ions, sulfate, and elemental sulfur), and the microbial activity was obtained. (ii) The temporal development of leaching activity indicates that in this climate biological leaching needs about 5 years to almost fully oxidize all metal sulfides exposed at the surface (0-30 cm). (iii) Furthermore, bioleaching is restricted to the humid season from November until April. Keywords: AMD, open pit mining, mining waste, monitoring, microcalorimetry 1.

INTRODUCTION Huge abandoned mining areas with waste heaps, open pits, and acidic lakes are the unpleasant but often tolerated side effect of the metal winning activities of modern times. Because these sites are often exposed to air and water and in most cases still hold significant amounts of metal sulfides they tend to pollute rivers and groundwater with acidic heavy-metal-containing solutions. This contamination, generally attributed to as acid mine or rock drainage (AMD, ARD), is caused by the activity of acidophilc iron(II)and sulfur-oxidizing bacteria [1]. Therefore, risk assessment studies and rehabilitation projects are only possible if information is provided on the parameters which determine the bioleaching process. In particular, data on leaching kinetics and the overall temporal development of the microbial flora in a special mining area have to be collected before * Present address: UFZ Center for Environmental Research Leipzig-Halle, Permoserstr. 15, D-04318 Leipzig, Germany

997

Microbiology Fundamentals

starting rehabilitation activities. To meet these prerequisites we have studied for nearly 10 years the changes and interdependencies of leaching-relevant parameters in the abandoned mining district between Portman and La Union in the southeast of Spain (province Murcia). Sampling campaigns were performed at waste heaps near and in the open pit mines Brunita and Gloria as well as at natural leaching sites (Fig. 1). The ore-rich region near the city of Cartagena has suffered from human mining activities for thousands of years. Thanks to the excellent natural harbor known as Portman (a contraction of the Latin Portus Magnus) at the Mediterranean coast the richest ore deposits of the hills surrounding this bay were already mined by the Romans for silver and lead. Intensive exploitation by surface extraction technologies started in the late 1950s by S. M. M. Penarroya. During the next 30 years waste materials from several open pit mines, ore processing, and smelters were deposited as heaps or dumped in the bay of Portman [2]. The disposal in the Mediterranean Sea was stopped in 1990, leaving the bay completely filled with 54 million tons of heavy-metal-containing mining waste [2,3]. Finely, metal-winning activities ceased with the closure of the open pit mine Los Blancos III in 1991 and the closure of the Pb smelter near Cartagena in March 1992 [4].

Figure 1. Mining area between La Union and Portman in the Cartagena mountain range (Autonomous Region of Murcia, southeast Spain). The dotted area marked the abandoned mining district with the arrowhead pointing at the position of the open pit mines Brunita and Gloria 2.

MATERIALS AND METHODS

2.1 Sample sites and sample analyses The investigated area between La Union and Portman (Fig. 1) belongs to a geological formation that contains Fe-, Pb-, Ag-, Cu-, Zn-, Mn-, and Sn-rich ore deposits [5]. Sampling campaigns were done at 4 sites: (a) A waste heap near La Esperanza to the north of the open pit mine Brunita, which was piled up with material out of the mine as well as 998

Microbiology Fundamentals

with flotation residues; (b) the abandoned open pit mine Brunita; (c) a natural leaching site at a rock formation with an outcropping pyrite vein, about 300 m to the south of Brunita; (d) the abandoned open pit mine Gloria. Analyses of leaching parameters were performed as described previously [6]. Briefly, sample humidity, temperature, pH, redox potential, calorimetric activity, and conductivity were recorded. Furthermore, concentrations of polythionates, sulfate, elemental sulfur, and iron ions were determined. For the assessment of the aerobic microbial flora acidophilic iron(II) oxidizers, acidophilic sulfur oxidizers, acidophilic chemoorganotrophs, neutrophilic sulfur oxidizers, and neutrophilic chemoorgano-trophs were counted by MPN enumeration. In order to correlate the data of sample analyses with the climatic situation in the province Murcia long-time mean values of temperature and precipitation for San Javier were applied [7]. On the basis of this data we were able to distinguish two seasons: “winter” covers the humid period from November to April and “summer” covers the dry period from May to October. 2.2 Statistical methods In a first statistical treatment the samples were compared with respect to all leaching relevant parameters after dividing them in groups that had been classified with respect to sample sites (see 2.1) and sample types (solid surface samples or from a depth of 20 to 30 cm, solid samples from the pyrite vein, lake sediment samples, and water samples). As the collected data showed no normal distribution only non-parametric (i. e. distribution-free) statistical tests could be applied. Multiple and bivariate group comparisons were performed with the Kruskal-Wallis and Mann-Whitney U-test, respectively. In order to elucidate the relationship between microbial activity and in situ conditions nonparametric, bivariate correlation analyses were performed (Kendall S-test). In this way obtained results were tested with partial correlation analyses. Changes in the leaching processes during the total campaign period were further tested by multiple linear regressions. In this case, climatic influences were also considered by using the season at which the samples were collected as a covariable. Generally, the significance level was set for 5% and was adjusted to the number of statistical tests performed (Bonferroni´s adjustment). The software SPSS 10.0.7 was used for the statistical analyses. 3.

RESULTS AND DISCUSSION

3.1 Microbial activity From previous studies it is already known that the microcalorimetrically measured activity is one of the most reliable parameter in order to characterize a certain leaching biotope [6,8]. A comparison of the microcalorimetric activity with the other parameters recorded for the samples from the La Union mining district confirms this finding, too. The mean activity of all samples was highest within the first 3 years of the campaigns (1991 to 1993) and then decreased continuously (Fig. 2). This temporal development of the leaching activity can be explained with the closure of most open pit mines in the late 1980s and early 1990s. Obviously, in this mining area it takes about 5 years to almost fully oxidize the metal sulfides exposed to the surface (0-30 cm). Samples which showed a correlation between calorimetric heat production and cell numbers of leaching bacteria had activity values strongly depending on the season at which the samples were collected. As expected, the calorimetric activity was high in the humid winter season and low in the dry summer time (Fig. 3A). In other words, significant biooxidation of metal sulfides at the surface of the leaching sites (0-30 cm) only occurred between November and April. In typical leaching biotopes the microbial processes are dominated by acidophilic bacteria. 999

Microbiology Fundamentals

Consequently, calorimetric values correlated negatively with the pH of the samples (Fig. 3B).

Figure 2. Temporal development of the microcalorimetrically measured activity of samples collected from 4 different leaching sites at the La Union mining district (sampling depth 0-30 cm)

Figure 3. Dependency of microbial activity of samples from the La Union mining district (sampling depth 0-30 cm) on season (A) and pH (B) 3.2 Group comparisons and correlation analyses of other parameters The interdependencies of the various leaching parameters were evaluated by group comparisons and bivariate correlation analyses (see 2.2). Generally, it could be confirmed that all sampling sites are typical acidophilic leaching biotopes (data not shown). Although the total number of samples was comparatively low and the sampling sites consisted of very heterogeneous materials a positive correlation was found between the proton concentration and numbers of acidophilic leaching bacteria (iron(II)- and FeS2-oxidizing bacteria). In contrast, but in conformity with the previous finding, the proton concentration correlated negatively with cell numbers of neutrophilic bacteria (thiosulfate-oxidizing and chemoorgano-heterotrophic microorganisms). As the leaching activity was highest in

1000

Microbiology Fundamentals

acidic samples (Fig. 3B) the amounts of leaching products (iron ions, sulfate, and elemental sulfur) showed a highly significant correlation with the proton concentration. 4.

CONCLUSIONS From the results of this study it can be concluded that rehabilitation projects should start concomitantly with or immediately after mining, if climatic conditions are comparable to the La Union district. In this region leaching bacteria oxidizes surfaceexposed metal sulfides (0-30 cm) almost completely within 3 to 5 years. Consequently, AMD/ARD can only be stopped if mitigation measures accompany the piling of waste heaps and the other depositing techniques. On the other hand, it is clearly shown that the simple closure of open pit mines is not an adequate solution for the problem. The consequences of this inactivity are acidic heavy-metal-contaminated soil and surface water. This contamination is build up in a few years but will leave an abandoned mining area unsuitable for many years with respect to farming and urban development as well as for tourism. ACKNOWLEDGMENTS The help of J.I. Manteca (Universidad Politécnica de Cartagena) in providing detailed information about the mining district La Union and the help of R. Buchert (University Hospital Eppendorf, Hamburg) with the statistical analyses are greatly acknowledged. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

H. L. Ehrich, Geomicrobiology, Marcel Dekker, New York, 2002. J. M. M. Orozco et al., Landscape Urban Plann., 23 (1993) 195. C. Auernheimer and S. Chinchon, Environ. Geol., 29 (1997) 78. S. Moreno-Grau et al., Atmos. Environ., 34 (2000) 5161. I. S. Oen et al., Econom. Geol., 70 (1990) 1259. A. Schippers et al., Appl. Environ. Microbiol., 61 (1995) 2930. http://www.inm.es/wwc/html/dclimat/SAN_JA.html W. Sand et al., Appl. Microbiol. Biotechnol., 40 (1993) 421.

1001

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

An AFM-study on the adhesion of Acidithiobacillus ferrooxidans and Leptospirillum ferrooxidans to surfaces of pyrite K. Kinzlera, W. Sanda*, J. Telegdib and E. Kalmanb a

Institut für Allgemeine Botanik, Mikrobiologie, Universität Hamburg, Ohnhorststrasse 18, D-22609 Hamburg, Germany b Hungarian Academy of Sciences, Chemical Research Center, Pusztazeri ut 59-67, H 1025 Budapest, Hungary

Abstract The attachment of Acidithiobacillus ferrooxidans and Leptospirillum ferrooxidans to pyrite surfaces has been analyzed using atomic force microscopy (AFM). Cells of both species attach preferentially to sites with visible surface defects. Although a part of the cells from the inoculum was still planktonical, a complete coverage of the pyrite surface never became detectable within 24 hours of attachment experiments. Only after the onset of growth, the surface became fully EPS-covered and, thus, turned into an "organic" surface. The extracellular polymeric substances (EPS) fill the space between the bacterial cell (wall) and the substratum surface. Furthermore, the EPS extend beyond the cell body. Especially the strain of L. ferrooxidans exhibited strong EPS-production and, consequently, an enormous surface coverage. Due to this coverage, the cells have an enlarged metabolic radius-of-action. Whereever EPS are in contact with the surface, the metal sulfide dissolution will be enhanced, due to the EPS-complexed iron(III)ions. 1.

INTRODUCTION Bioleaching is the bacterially mediated dissolution of metal sulfides to sulfuric acid (sulfate) and metal cations. This process is known for decades and has been shown to be of industrial importance as well as of environmental significance (acid rock drainage = ARD) [1,2,3,4]. Since its discovery there has been a debate about the mechanism, by which the microorganisms dissolve metal sulfides. Generally, a direct and an indirect mechanism are proposed. The argument pro and contra continue still nowadays [5,6,7]. Recent work of our group indicated that bioleaching is a process, which combines characteristics of both mechanisms [8,9,10]. This means that the dissolution of a metal sulfide is a chemical process caused by oxidative attack of iron(III)ions and/or protons. The iron(III)ions may be involved in any dissolution of metal sulfides, whereas protons are restricted to those metal sulfides, which are amenable to a hydrolytic dissolution. The group of metal sulfides, which obligately need iron(III)ions for a dissolution, comprise pyrite (FeS2), molybdenite (MoS2), and tungstenite (WS2). Their characteristic is the fact that the outer electron orbitals result from the metallic part of the compound, and not from the metal-sulfur bond. Thus, the oxidative attack by iron(III)ions is a requirement. The other metal sulfides have outer orbitals from the sulfur-metal bond and, thus, are attackable by protons and by iron(III)ions. In the presence of the latter the rate of 1003

Microbiology Fundamentals

dissolution may become considerably enhanced, which often causes erroneously the statement of a direct mechanism [11]. Furthermore, bioleaching is caused mainly by attached cells [8,10]. The planktonic population plays a negligible role under most conditions. Only in case of a high concentration of dissolved iron(III)ions planktonic cells could contribute. The iron(III)ions are provided by the leaching bacteria complexed in their extracellular, polymeric substances (EPS). The complex most likely consists of two moles of glucuronic acid and one mole of iron(III)ions. Both, At. ferrooxidans and Leptospirillum ferrooxidans, have these glucuronic acid - iron(III)ion complexes in their EPS. By an electrostatic interaction between the negatively charged pyrite surface (pyrite assumes a negative charge at pH below 3) and the net positive charge of the iron(III)ions-containing EPS, surface attachment of the cells becomes possible as the primary event in adhesion. Whether in later stages additional factors like pili etc. function, remains an open question. Some indications exist that such interactions may contribute also to the formation of robust surface attachment. Another open question is the site for attachment, whether it is a randomly or a deliberately chosen place. Some considerations seem to suggest that in case of bioleaching bacteria random attachment does not play the major role. Especially chemotaxis seems in this context to be a very important surface site recognizing system [12,13]. The present work was executed to visualize the attachment behavior of cells of two of the main bioleaching bacteria, At. ferrooxidans and L. ferrooxidans. The idea was to find criteria to judge the importance of EPS for attachment, to clarify the attachment mechanism, and to obtain information about the characteristics of the attachment site. For this purpose pure cultures of these species, which previously had been selected for good attachment properties, and coupons of selected pyrite cubes were incubated jointly. The coupons were analyzed for changes of the surface morphology in time-dependent adhesion and biofilm formation experiments using atomic force microscopy (AFM). AFM, a variety of tunneling scanning microscopy, allows to record the height profiles of plain surfaces by a surface scanning mechanism, and to put this together to produce a topographical map – AFM-images – of the investigated sample. Since this scanning action can be performed in air or water, living cells may be investigated without any interference by vacuum etc. Thus, the big advantages of this technique are the simple sample preparation and the measuring conditions at atmospheric pressure and room temperature. Consequently, for an investigation under natural conditions, AFM is the method of choice. The application of AFM for monitoring of microorganisms has been described in several papers [14,15,16]. 2.

MATERIALS AND METHODS

2.1 Strains For these investigations the strains At. ferrooxidans strain R7 and L. ferrooxidans strain R3 were selected. They are kept in our culture collection in nutrient solution at pH 1.9 [17]. 2.2 Pyrite sample preparation Pyrite was obtained from mineral shops as cubes of at least 1 cm³ volume. To obtain suitable samples for AFM investigations, the cubes were cut by a diamond saw into plates of about 2 mm thickness. Afterwards, the plates were washed with ethanol and hydrochloric acid to remove any organic contaminants and/or oxidation products like iron(II/III)ions, elemental sulfur, or polythionates. This procedure was followed by 1004

Microbiology Fundamentals

sterilization in 70% ethanol for 12h. These plates were introduced into the flasks with the bacterial cultures. 2.3 AFM experiments For experimental purposes, 500 mL cultures were grown, harvested by centrifugation after two-thirds of the iron(II)ions had been oxidized, and resuspended in 100 mL nutrient-free solution ([17], without iron(II)ions). After a period of 24h for substrate depletion (to allow the cells to oxidize any remaining substrate), pyrite samples were introduced into the cultures and incubated for periods of up to 96h with stirring. At different times pyrite samples were removed and analyzed using the AFM. The atomic force microscope (NanoScope III, Digital Instruments, Veeco, USA) was operated in air in contact mode. A silicium nitride tip (SiN4) was used for surface scanning. All figures contain images, which have been obtained without any processing except autoflattening. 3.

RESULTS AND DISCUSSION A sterile pyrite surface prior to any attachment experiment is shown in Figure 1. The various layers besides minor surface defects are clearly visible on the (typical) pyrite surface.

Figure. 1. AFM-image of a sterile pyrite surface with a typical layered surface structure and two inorganic crystals 2.1 Attachment of At. ferrooxidans strain R7 to pyrite Attachment experiments were performed with At. ferrooxidans and immersed pyrite plates. The following figures show the patterns of surface colonization by cells of At. ferrooxidans R7. For evaluating the question for preferential attachment, especially images in the time range of 2 to 8 h are useful. Such an image is given in Figure 2. Images obtained lateron yielded only equivocal information due to growth effects. It becomes obvious from Figure 2 and similar images (not shown) that the cells attached preferentially to sites with visible surface defects. More than 80% of all cells were attached to distortions of the pyrite surface (grooves, channels, edges of layers etc.) Images obtained at a later stage of the surface colonization process prove the growth of the attached bacteria (Fig. 3). Figure 4 demonstrates the existence of pits underneath bacteria. After removal of the cells, pits with the shape and size of cells become visible. Besides the bacteria footprints of cell-free EPS may remain on the surface (Fig. 5). Typical patterns of time-dependent surface colonization are presented in Figure 6A-D. Within 6 h several 1005

Microbiology Fundamentals

cells have attached to the pyrite surface. The amount of attached cells increases up to 24 h. Afterwards, until 48 h the amount of attached cells remains almost similar. One week later, the surface morphology exhibits considerable alteration. Only a few single cells remain visible. The majority is embedded in copious amounts of an amorphous mass, the EPS (measurements in the friction mode of the AFM proved that the amorphous mass is soft, whereas the free pyrite surface remained rough). The EPS cover a considerable part of the total surface. In Figure 7 a scheme of a surface-attached, bioleaching cell of At. ferrooxidans is shown.. It combines all our evidence collected from previous work [8] and from AFM images. Most important is the finding that the EPS extends over the limits of the cellbody and in this way considerably enlarge the metabolic radius-of-action.

Figure 2. AFM-image of a pyrite surface after 8 h of incubation with Acidithiobacillus ferrooxidans strain R7. Cells attached preferentially to sites with visible surface defects

Figure. 3. AFM-image of a pyrite surface after 24 h of incubation with Acidithiobacillus ferrooxidans strain R7. Pit formation is visible in the centre of the image

Figure 4. AFM-image of a cleaned pyrite surface after 24 h of incubation with cells of Acidithiobacillus ferrooxidans strain R7

Figure 5. AFM-image of a single cell of Acidithiobacillus ferrooxidans strain R7 attached to pyrite after 24 h of incubation. Around the cell cell-free exopolymers (footprints) are visible

1006

Microbiology Fundamentals

Figure 6A-D. Typical patterns of time dependent surface colonization by Acidithiobacillus ferrooxidans strain R7 over the course of 8 d

Cell A

cell of Acidithiobacillus ferrooxidans A

0

3+

S 2+ Fe S2O32-

attack

Figure 7. Schematical graph of a cell of Acidithiobacillus ferrooxidans attached to pyrite. Included is a model for the mechanism of the indirect or contact-mode leaching attack as catalyzed by A. ferrooxidans 1007

Microbiology Fundamentals

2.2 Attachment of Leptospirillum ferrooxidans strain R3 to pyrite The same experiments, as shown for At. ferrooxidans, were performed with cells of L. ferrooxidans strain R3. In Fig. 8 an image with a few cells of L. ferrooxidans attached to a pyrite surface is shown. In the middle of the image a dividing cell or two recently divided cells are visible, which are jointly embedded in their EPS. The cells are surrounded by copious amounts of EPS. The amount of EPS is considerably higher than the amount of the EPS of At. ferrooxidans strain R7 (Fig. 5). Furthermore, the EPS cover a considerably larger part of the surface than the naked cells would be able to. The surface coverage is in principle similar to the one shown in Figure 7 for At. ferrooxidans. Time-dependent adhesion is shown in Fig. 9A-C. In the course of the colonization a serious change of the surface morphology becomes visible. After 6 h of incubation (Fig. 9A) several cells have attached to the pyrite surface and excreted some EPS around the cells. After 24 h (Fig.9B) the morphology already has changed considerably. Single cells with their surrounding EPS are still identifiable, although the EPS starts to merge together. After 48 h (Fig. 9C) the surface of the pyrite is covered by an almost homogenous film of EPS (sea of EPS) with some embedded cells. Consequently, the structure of the pyrite is not visible anymore. The presence of the EPS again was proven by experiments using the frictionmode of the AFM (data not shown).

Figure 8. AFM-image of cells of Leptospirillum ferrooxidans attached to pyrite after 24 h of incubation

Figure 9A-C. AFM-image of cells of Leptospirillum ferrooxidans attached to pyrite after 6 (A), 24 (B) or 48 (C) h of incubation

1008

Microbiology Fundamentals

3.

CONCLUSIONS Summarizing, it is obvious that an attachment of these bioleaching bacteria takes place at selected sites. These sites have a surface defect, which most likely causes a high surface charge by electrical imbalance. This agrees well with the hypothesis of an electrostatic interaction mechanism between EPS and pyrite surface. It also becomes clear from these images that the EPS are always the compounds mediating the contact between a bacterial cell and pyrite. Once the cells are attached, they start to grow and, within a few days, may cover the surface with large amounts of EPS. Consequently, the previously inorganic surface turns to an organic one. It remains to be elucidated, what this means for further attachment. Preliminary data indicate that this organic surface film hinders other bacteria to attach, even those of the same strain. In any case, the EPS with their complexed iron(III)ions are able to oxidatively attack the pyrite. Thus, the cells radius-of-action becomes considerably larger than the dimensions of an EPS-free cell would indicate. Especially for L. ferrooxidans, whose only substrate are the low-energy iron(II)ions, this enlarged radiusof-action might confer an improved environmental competitiveness. REFERENCES 1. A. Schippers and K. Bosecker, Z. Angew. Geol., 48 (2002) 38. 2. D. Fortin, B. Davis and T.J. Beveridge, FEMS Microbiol. Ecol., 21 (1996) 11. 3. K.J. Edwards, T.M. Gihring and J.F. Banfield, Appl. Environ. Microbiol., 65 (1999) 3627. 4. N. Kuyucak, Min. Environ. Managem., 11 (2001) 12. 5. W.Sand, T. Gehrke, R. Hallmann and A. Schippers, Appl. Microbiol. Biotechnol., 43 (1995) 961. 6. M.P. Silverman, J. Bacteriol., 94 (1967) 1046. 7. F.K. Crundwell, In: V.S.T. Ciminelli and O. Jr. Garcia (eds.), Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, Elsevier, New York, 2001, Part A, p. 149. 8. W.Sand, T.Gehrke, P.-G.Jozsa and A. Schippers, Hydromet., 59 (2001) 159. 9. T. Rohwerder, P.-G. Jozsa, T. Gehrke and W. Sand, In: G. Bitton (ed.) Encyclopedia of Environmental Microbiology, Vol 2, Wiley, New York, 2002, p. 632. 10. D.E. Rawlings, Annu. Rev. Microbiol., 56 (2002) 65. 11. O.H. Tuovinen, In: H.L. Ehrlich and C.L. Brierley (eds.) Microbial Mineral Recovery, McGraw-Hill, New York, 1990 p. 55. 12. J. Acuna, J. Rojas, A.M. Amaro, H. Toledo and C.A. Jerez, FEMS Microbiol. Lett., 96 (1986) 37. 13. G.Meyer, T. Schneider-Merck, S. Böhme and W. Sand, Acta Biotechnol., 22 (2002) 391. 14. Y.F. Dufrene, Ch.J.P. Boonaert, H.C. van der Mei, H.J. Busscher and P.G. Rouxhet, Ultramicroscopy, 86 (2001) 113. 15. Y.F. Dufrene, Micron, 32 (2001) 153. 16. J. Telegdi, Zs. Keresztes, G. Pálinkás, E. Kalman and W. Sand, Applied Physics A, 66 (1998) S639. 17. M.E. Mackintosh, J. Gen. Microbiol., 105 (1978) 215.

1009

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

An X-ray photoelectron spectroscopy study of the mechanism of microbially assisted dissolution of chalcopyrite A. Parkera, C. Klauberb*, M. Stottb, H.R. Watlingb and W. van Bronswijka a

A.J. Parker Cooperative Research Centre for Hydrometallurgy, School of Applied Chemistry, Curtin University, P.O. Box U1987, Perth, WA 6845, Australia b A.J. Parker Cooperative Research Centre for Hydrometallurgy, CSIRO Minerals, P.O. Box 90, Bentley, WA 6982, Australia Abstract The acidic oxidative dissolution of samples of p-type chalcopyrite (Mt Isa Mines flotation concentrate and massive Mt Lyell) via the iron- and sulphur-oxidising microbes Sulfobacillus thermosulfidooxidans (50°C) and Sulfolobus metallicus (65°C) has been examined using X-ray photoelectron spectroscopy. This enabled surface speciation, encompassing sulphides, disulphides, elemental sulphur, sulphate and various jarosites, during the main dissolution stages. When compared to the case for the abiotic dissolution of chalcopyrite conducted under similar chemical conditions, both the similarities and differences can be highlighted. Principal differences between abiotic and microbially assisted dissolution arise from the microbial tendency to oxidise elemental sulphur and the effects of microbial nutrient media toward the type of jarosite formed. The role of the surface sulphide dimer S22- in the sulphide oxidation has also been clarified. 1.

INTRODUCTION The goal of complete copper recovery from chalcopyrite via low energy hydrometallurgical processing is critically dependent upon the control of "passivating" overlayer structures. The body of work previously undertaken by the authors (Klauber et al., 2001; Klauber, 2003; Parker et al., 2003), concentrating on the abiotic chemistry of Mt Isa Mines (MIM) and Mt. Lyell p-type chalcopyrites, for acidic sulphate oxidative dissolution, shows related behaviour over the thermophile temperature range and describes in detail a surface specific mechanism for abiotic chalcopyrite dissolution. The typical behavioural traits are oxidative dissolution to produce iron and copper in solution, with the sequential oxidation of the sulphide to disulphide, elemental sulphur and sulphate. The latter three all contribute to the initial surface reaction phase. Eventually the sulphate dominates, with a thin ferric sulphate precursor phase forming. With continued dissolution, the ferric solution concentration reaches a point where massive hydronium jarosite layers precipitate, hindering further chalcopyrite attack. Note that varied jarosites (KFe3(SO4)2(OH)6) are possible with cation substitution of the K+ by H3O+, NH4+ etc. The

* Corresponding author. PO Box 90, Bentley, WA 6982, Australia, e-mail: [email protected]

1011

Microbiology Fundamentals

initial ferric sulphate layer is observed to act as the crystallization precursor for jarosite precipitation. Iron control appears to be the key to chalcopyrite hydrometallurgy. Elemental sulphur plays a role but does not appear to be a significant candidate for hindered dissolution. Despite thorough investigation, no spectroscopic evidence was found for stable forms of the previously alluded to polysulphides (Parker et al., 2003). This work places the abiotic chemistry into context with bioleaching in the presence of two microbial strains. Particular aims were to examine the mechanism of sulphur oxidation further and jarosite composition as a function of microbially assisted leaching conditions. 2.

EXPERIMENTAL

2.1 Microbial culture Sulfolobus metallicus (DSM 6482T) and Sulfobacillus thermosulfidooxidans (DSM 9293T) were obtained from the German Collection for Microorganisms And Cell Cultures (DSMZ). Sulfolobus metallicus was cultured in a nutrient medium containing (NH4)2SO4 (1.3 g/L), KH2PO4 (0.28 g/L), MgSO4.7H2O (0.25 g/L) and S. thermosulfidooxidans in a nutrient medium containing (NH4)2SO4 (3.0 g/L), K2HPO4 (0.5 g/L), MgSO4.7H2O (0.5 g/L), KCl (0.1 g/L), Ca(NO3)2 (0.01 g/L). Both media were adjusted to pH 2.0 with concentrated H2SO4. The media were autoclaved at 121°C (2 atm) for 15 minutes. Sulfolobus metallicus runs were also conducted in potassium free media where K2HPO4 was replaced with (NH4)2HPO4 (0.27 g/L) and (NH4)2SO4 was reduced to 1.0 g/L. Yeast extract (0.02% (w/v)) was added aseptically to both media (via passage through a 0.22 µm pore size membrane). The chalcopyrite concentrate (CuFeS2 88%, FeS2 4% and gangue minerals, mainly quartz, 8%) was wet sieved to a size fraction of +20 µm –35 µm, and then dried in an oven at 30°C. Residual xanthates were then washed from the concentrate (Klauber et al., 2001). The cultures were grown at 65°C (S. metallicus) and 50°C (S. thermosulfidooxidans) in an environmental incubator shaker operated at 180 rpm. 2.2 Leaching method The microbially assisted leaches for the concentrate and the massive chalcopyrite were conducted in different experimental apparatus. The concentrate was leached in simple flask arrangements in an environmental incubator shaker. A 5 mL volume of solution containing S. metallicus or S. thermosulfidooxidans (late exponential phase, ~72 hours post-inoculation) was inoculated into a series of 250 mL Erlenmeyer flasks (one for each leach period), containing 95 mL of appropriate nutrient medium and 3.1 g concentrate. Samples of the massive chalcopyrite were silver dagged to sample stubs, dry polished with 2000 grit carbide paper and dipped onto an overflowing leachant meniscus; surface tension ensured contact only with the chalcopyrite surface. The leachant cell, containing 25 mL of inoculum and 0.03% (w/v) yeast extract in 500 mL of nutrient media, was run for 12 hours with 16.0 g concentrate prior to exposure to the massive chalcopyrite. 2.3 Solution sampling and analysis The pH and redox potential (Ag|AgCl) were measured periodically throughout the experiments. For the concentrate leaches, concentrations of ferrous ion, total sulphur, copper and iron in solution were determined after centrifugation to separate the solid material. Ferrous ion concentrations in solution were determined by spectrophotometry using the method of Wilson (1960). The total sulphur, iron and copper concentrations in 1012

Microbiology Fundamentals

solution were measured using Inductively Coupled Plasma - Atomic Emission Spectrometry (ICP-AES). 2.4 X-ray photoelectron spectroscopy At selected reaction intervals (12, 24, 36, 60 and 108 hours), the contents of a single flask were filtered (0.22 µm pore size membrane) and then washed with 10 mL of acidified MilliQ water (H2SO4 to pH 2) to remove soluble metal species. A portion of the solid material retained on the filter was then pressed into a sample stub and transferred into the photoelectron spectrometer and cooled to ~150K to minimise sulphur volatilisation. For the massive chalcopyrite, the surface was examined at shorter intervals (0.5, 2, 24, 48 and 72 hours). These pre-mounted samples went directly into the spectrometer after rinsing with acidified MilliQ water (H2SO4 to pH 2). XPS spectra were obtained with either Mg Kα or Al Kα radiation (300 W) from a conventional source in a VG ESCALAB Mk II fitted with 5-channeltron electron detection. Spectra were run at base pressure of ~1x10-10 mbar, using 6 mm analyser slits and a fixed analyser pass energy of 20 eV. The S 2p, Cu 2p, Fe 2p, O 1s, N 1s, K 2p, C 1s, K 2p, valence band, Fe 3s and full survey regions were acquired in sequence. Methods of data reduction employed such as X-ray line function removal (source function and satellites), spin-orbit splitting removal (in the case of S 2p) and comments on the peak fitting are covered in detail in earlier papers in the series (Klauber et al., 2001; Klauber et al., 2003; Parker et al., 2003). 3.

RESULTS AND DISCUSSION

3.1 Solution conditions Metal species solution concentration and redox potential for the concentrate cases are shown in figure 1 and are typical for the microbially assisted ferric acid leach conditions employed. As would be expected, Eh rises and solution ferrous concentration drops as the leachant activity and microbial numbers increase. Also note the low concentrations of soluble iron as compared to the leached copper (especially in the S. metallicus- assisted leach) indicating substantial jarosite formation. The pH remained constant at ~2 until extensive jarosite formation occurred. Note that a rate comparison to abiotic conditions is not simple. In abiotic dissolution the initially high ferric level contrasts to a low value in the microbial case and the microbial catalytic role of ferrous to ferric conversion is also absent. 3.2 Photoelectron spectra of the inorganic phases The matrix of four leach experiments (two microbes and two chalcopyrite types) over a number of sampling periods and the acquisition of multiple spectral lines allows only a few of the spectra to be shown. For clarity, the composite XPS spectra have arbitrary shifts in the ordinate position to avoid overlap. Spectra at increasing time intervals are arranged from top to bottom. For thin surface reaction layers, exhibiting negligible static charge shifts, the binding energy (BE) scales are referenced to the underlying chalcopyrite lattice sulphide species S2- at 161.15 eV (Klauber et al., 2001) to correct for spectrometer drift. This is important for the S 2p analysis. Figures 2 and 3 show the typical (unprocessed) S 2p data for the initial stages of the oxidation process for the concentrate. Note that once deposition of thick non-conducting overlayers has occurred there is an onset of static charging (typically 5-7 eV to higher apparent BE values). This has not been corrected for in the figures so as to illustrate that critical layer formation. For example, the raw sulphate peak in figure 2 can be seen to shift from 169.1 to 175.7 eV. 1013

Microbiology Fundamentals

Figure 1. Solution concentration of copper and iron and Eh variation during concentrate leaching

Figure 2. Sulphur 2p spectra from concentrate leached in the presence of S. thermosulfidooxidans at 50°C for periods varying from 12 to 108 hours

1014

Microbiology Fundamentals

Figure 3. Sulphur 2p spectra from concentrate leached in the presence of S. metallicus at 65°C for periods varying from 12 to 108 hours in nutrient media with and without K+ Due to the difficulty in reaction rate control for the microbially assisted leaches, caution needs to be taken in comparing the observed times for surface phase formations. Qualitatively, the same surface phase components form on both the MIM concentrate and the Mt. Lyell massive samples, but they do not necessarily correlate in terms of time. This is a consequence of differences in the experimental and sample set-ups (high surface area concentrate versus low surface area massive). Even within individual repeat leach runs exhibiting the same sequence of phases, the onset time for a given part of the sequence could be at variance by hours. This is despite care in reproducing both inoculum activity and the chalcopyrite sample. Other differences between the concentrate and massive relate solely to particle size, e.g. the deposition of jarosite related layers. For the concentrate, the static charge shifts are sudden as the particles are small and such layers reduce electrical conductivity between particles and the particles and earth (i.e. multiple insulating layers). By comparison the same relative jarosite levels on the massive material (in fixed contact with earth) resulted in negligible shifts. This does not indicate a difference in chemistry. As a diagnostic tool the Cu 2p spectrum is of limited value as all the spectra (e.g. figure 4) are typically representative of Cu[I] for the unoxidized chalcopyrite lattice, diminishing in intensity as the depositing overlayers attenuate the photoemission. The Fe 2p spectrum (figure 5) is more sensitive as iron species participate in the formation of coherent overlayers, starting with pre-cursor iron sulphate acting as a nucleation site for jarosite deposition. Analysis of the Fe 2p states via peak fitting is not practical, though a least squares principal component analysis (Parker et al., 2003) can reliably indicate the overlayer composition of iron containing phases.

1015

Microbiology Fundamentals

Figure 4. Copper 2p spectra from concentrate leached in the presence of S. metallicus at 65°C for periods varying from 12 to 108 hours in nutrient media with and without K+

Figure 5. Iron 2p spectra from concentrate leached in the presence of S. metallicus at 65°C for periods varying from 12 to 108 hours in nutrient media with and without K+ The S 2p peak remains the most useful means of understanding the oxidation process. Figure 6 shows a detailed sulphur analysis of the initial oxidation in terms of states largely already established (Klauber et al., 2001, Klauber 2003, Parker et al., 2003). Once the state of reaction has progressed to significant jarosite formation, the reduced emission 1016

Microbiology Fundamentals

intensity from the mixed states region below the sulphate makes their analysis less reliable. The component that evolved at a higher BE than the sulphate had not been previously observed. It appeared to be characteristic of the high initial sulphate levels for the concentrate, but not the massive. It is most likely to be sulphate emission from nucleating jarosite crystals. This may be an illustration of differential static charging on the surface or a genuine BE shift for conversion from a simple sulphate to one characteristic of a jarosite. Also noteworthy in figure 6, although the sulphide peak has been aligned to the lattice value of 161.15 eV, as the elemental sulphur evolves the measured sulphide feature actually shifts to a higher BE by ~0.13 eV whilst the sulphate BE remains approximately constant. At the same time the loss feature (Klauber, 2003) diminishes in intensity. Both this and the sulphide shift may be real and may point to a breakdown of the near surface chalcopyrite lattice. There are two factors influencing observed BE. Based on the Muliken charge population analysis (Hamajima et al., 1981) the sulphide donates charge toward the iron, so a lattice breakdown to fully ionic S2- should cause a shift to lower BE. Countering that, the simultaneous breakdown of the conduction band would mean less relaxation and a shift to higher BE.

Figure 6. Detail of the sulphur chemical state analysis for processed spectra showing the variation in states during the initial leaching. For figure clarity neither the final fits nor the residuals are shown

1017

Microbiology Fundamentals

Comparison of the S 2p at 60 hours with the S. thermosulfidooxidans to that for abiotic dissolution at 50°C after 2 hours (Parker et al., 2003) shows a very similar structure. Whilst the abiotic response appears "faster" (due to a higher initial ferric concentration), the microbially assisted creates more surface sulphate and appears to allow a more concerted attack on the sulphur dimer species by the ferric. The likely reason for this is the direct microbial attack on the elemental sulphur, a pathway absent in the abiotic case. 3.3 Photoelectron spectra of the entrained biomass Any entrained or adherent biomass will contribute to the observed photoelectron signal, principally for the C, N and O 1s peaks and to a lesser extent P 2p. In this study it exhibited mainly in the C 1s and N 1s peaks; the O 1s was overwhelmed by the inorganic phases. In terms of a microbial cell wall characterization, the use of XPS has generally been limited to a cluster analysis based on nett N/C, O/C and P/C ratios (Van der Mei et al., 2000). This is pragmatic as it circumvents applying detailed peak analysis that can be especially difficult for complex organic systems. The current microbial studies do produce some differences to the prior abiotic chemical investigations. This is due to the biomass and more especially its nutrient media, e.g. ammonium ion presence. In this work, some knowledge of the expected biomass XPS spectrum, especially the N 1s, is needed to compare the leaching regimes, although the role of XPS in directly increasing microbe knowledge is limited. Based on the generalized chemical compositions of cells (Neidhardt et al., 1996) the N 1s signal would be expected to primarily originate from proteins (~55%) with the RNA component also a major contributor (~21%). As the backbone structures for proteins are polypeptides, the basic N 1s signal for the protein should be similar to that observed for amide linkages R-CO.NH-. Purely from a photoelectron BE viewpoint (i.e. chemical environment), synthetic polymers such as nylon are polypeptide analogues. This provides a guide to expected BE values (Beamson and Briggs, 1992): polymer R C1s amide C 1s amide N 1s amide O 1s nylon 6,6 285.00 - 286.02 eV 288.02 eV 399.81 eV 531.37 eV nylon 6 285.00 - 286.00 eV 288.01 eV 399.77 eV 531.35 eV nylon 12 285.00 - 285.90 eV 287.97 eV 399.84 eV 531.33 eV Whilst the side groups on the polypeptide backbone define the protein type, it would be expected that the amide nitrogen atom would dominate the overall N 1s signal. Hence a BE of about 399.8 eV should be expected (confirmed by an examination of Escherichia coli biomass). Carrying the synthetic polymer analogy further, the aromatic nitrogen atoms in the heterocyclic RNA bases would be expected to have a BE value about 399.3 eV, i.e. 0.5 eV below that of the amide nitrogen atoms (Beamson and Briggs, 1992). Hence two nitrogen peaks with a small separation would be expected for the biomass. As can be seen from figure 7, two N 1s states are observed on the chalcopyrite sample. Although the ∆BE value of 1.9 eV is well above the expected amide-RNA separation, it perfectly matches the separation between that of ammonium jarosite standard and the biomass signature from E. coli. The jarosite was BE corrected for the sulphate component to match that of jarosite on massive chalcopyrite and for E. coli the C 1s was fixed to 285.0 eV. The higher BE component was also noted to vary disproportionately in relative intensity and found to dominate at long leach times with significant jarosite deposition. This additionally indicated a non-biomass origin. Based on the BE comparison, the low BE state can be identified as polypeptide amide and the 1018

Microbiology Fundamentals

higher BE state as the ammonium component in ammonium jarosite (the ammonium coming from the nutrient media). Curiously, the ammonium also is present much earlier in the oxidation. The signal-to-noise is insufficient to separate the protein-RNA components. Note also that Al Kα is preferred over Mg Kα radiation as the N 1s region is otherwise complicated by a series of copper Auger peaks.

Figure 7. Nitrogen 1s spectra from Mt. Lyell massive leached in the presence of S. thermosulfidooxidans at 50°C for 24 hours compared to E. coli and ammonium jarosite. Both raw and processed spectra are shown 3.4 Peak in elemental sulphur production Both the S. metallicus and S. thermosulfidooxidans assisted oxidative leaches exhibit what appear to be peaks in the concentration of surface adsorbed elemental sulphur in the 36-60 hour range (figure 6 compared, for example, to 108 hours in figure 3). They are both known to be elemental sulphur oxidisers (Huber and Stetter, 1991, Golovacheva and Karavaiko, 1978) and also capable of oxidising ferrous so the nett behaviour will be controlled by the availability of these species. As elemental sulphur forms, some of the population oxidise elemental sulphur to sulphate. The specifics of the microbially assisted and abiotic leaching mechanisms are too complex to reduce to consecutive, opposing and concurrent reactions and also fit the limited rate data. We can, however, simplistically view the paths involving elemental sulphur as two consecutive steps: ferric microbial S2- ⎯ ⎯⎯→ S° ⎯ ⎯ ⎯ ⎯⎯→ SO42For the classic case of irreversible first order consecutive reactions, the intermediate species would first increase to a maximum and then fall to zero. Neither of these steps are first order, but taking the concentrations of (solid) sulphide as approximately constant and that of ferric as slowly varying, it will mimic first order. Similarly, for approximately constant microbial activity, the oxidation of the elemental sulphur will mimic first order. This would predict a peak in elemental sulphur concentration. For the abiotic case sulphate is proposed to form via a thiosulphate intermediate (Parker et al., 2003). Thiosulphate is known to be unstable at temperature in the presence of sulphuric acid disproportionating to elemental sulphur and sulphate (Mizoguchi et al., 1019

Microbiology Fundamentals

1976). By comparison with the microbial case the abiotic case does not have a consumption route for elemental sulphur. 3.5 The disulphide intermediate in elemental sulphur production Analysis of the sulphur 2p states for both the 12-60 hour S. thermosulfidooxidans and 12-36 hour S. metallicus leaches (figure 6) shows an interesting pattern in the relative concentrations of the disulphide dimer S22- and elemental sulphur (assumed to be S8). We have previously proposed the role of the pyritic-like dimer as an intermediate in the chalcopyrite oxidative dissolution by ferric ions (Parker et al., 2003). From earlier work the dimer appears to be a consequence of the fracturing of a chalcopyrite lattice and thus systemic to the chalcopyrite system (Klauber, 2003). Interestingly, the detailed S 2p state analysis for S. metallicus and S. thermosulfidooxidans (repeated for S. metallicus) shows a clear interrelationship between S22- and S8. This is shown in figure 8. At the 36-hour point when Eh and ferric concentration rise, there is a substantial decrease in the S22concentration and a stoichiometrically commensurate increase in the S8, i.e. the combined sulphur content of the two states remains constant. It is the increased ferric activity that converts S22- to S8. The concentrate’s small pyrite content complicates this interpretation, but the changing levels of dimer and the subsequent quantity of elemental sulphur are well in excess of what could be expected from the pyrite content. Moreover, preliminary acidic ferric leaches on a pyrite concentrate showed no evidence for adsorbed elemental sulphur being formed. Though the pyrite is reactive, oxidation is straight to sulphate and, unlike the lower density dimer forming on chalcopyrite, appears not to be able to form elemental sulphur under these conditions.

Figure 8. Decrease in surface dimer concentration is mirrored by a commensurate increase in adsorbed elemental sulphur, indicating the dimer to be the intermediate 3.6 Cation role in jarosite formation Prior abiotic work, in the absence of alkali cations, has indicated that eventual deposition of hydronium jarosite, initiated by a ferric sulphate surface phase, hinders dissolution (Parker et al., 2003). For the microbially assisted cases, the presence of potassium enables potassium jarosite to form. Removal of the K+ content from the nutrient 1020

Microbiology Fundamentals

medium not only prevented potassium jarosite, but actually appeared to delay a jarosite onset. This then occurred as the alternate ammonium jarosite (Stott et al., 2001) and is qualitatively supported by the ammonium phase being observed in larger amounts, based on the higher BE N 1s component, in the absence of K+. Though only qualitatively examined, it would seem reasonable that the control of singly charged cations may offer alternate avenues for jarosite control in addition to employing ferric control. The relative sensitivity to cation type has not been explored, though any mechanism that alters the final jarosite may also alter its structural coherence and possibly encourage spontaneous nucleation rather than formation as an inhibiting layer. 4.

CONCLUSIONS In comparing earlier abiotic work on the acidic oxidative dissolution of samples of ptype chalcopyrite (Mt Isa Mines flotation concentrate and massive Mt Lyell) with that in the presence of oxidising microbes S. thermosulfidooxidans and S. metallicus, at 50°C and 65°C we have been able to affirm that the interface chemistry is driven by the chemical conditions. Ηowever, the microbial activity and environment can have an impact upon that chemistry. The microbial oxidation of the elemental sulphur is a key to the peak in elemental sulphur concentration observed and appears to allow a more concerted attack, by the ferric ion, on the surface sulphur dimer than in the abiotic case. Nutrient media composition is also important. Due to the high concentrations of ferric, sulphate and monovalent cations at elevated temperatures, microbial leach solutions strongly favour jarosite formation. The potassium free nutrient media delayed the onset of potassium jarosite, with the less stable ammonium jarosite forming at later reaction times. Controlling the formation of jarosite and jarosite type remains the key in eliminating the "passivation", or hindered dissolution, of chalcopyrite with respect to copper release, in both abiotic and microbially assisted systems. REFERENCES 1. Beamson, G., and Briggs, D., 1992. High Resolution XPS of Organic Polymers, The Scienta ESCA300 Database. John Wiley & Sons, Chichester. 2. Golovacheva, R.S. and Karavaiko, G.I., 1978. A new genus of thermophilic sporeforming bacteria, Sulfobacillus. Mikrobiologiya, 47, 658-665. 3. Hamajima, T., Kambara, T., Ken, I. and Oguchi, T., 1981. Self-consistent electronic structures of magnetic semiconductors by a discrete variational Xα calculation. III Chalcopyrite CuFeS2, Phys. Rev. B: Condens. Matter 24, 3349-3353. 4. Huber, G., and Stetter, O., 1991. Sulfolobus metallicus, sp. nov., a novel strictly chemolithoautotrophic thermophilic archaeal species of metal-mobilizers. System. Appl. Microbiol., 14, 372-378. 5. Klauber, C., Parker, A., van Bronswijk, W. and Watling, H., 2001. Sulphur speciation of leached chalcopyrite surfaces as determined by X-ray photoelectron spectroscopy. Int. J. Min. Process., 62, 65-94. 6. Klauber, C., 2003. Fracture induced reconstruction of a chalcopyrite CuFeS2 surface. Surf. Inter. Anal. (in press). 7. Mizoguchi, T., Takei, Y. and Okabe, T., 1976, The chemical behaviour of low valence sulfur compounds. X. Disproportionation of thiosulfate, trithionate, tetrathionate and sulfite under acidic conditions. Bull. Chem. Soc. Jpn., 49, 70-75.

1021

Microbiology Fundamentals

8. Neidhardt, F.C., Curtiss, III, R., Ingrahamn J.L., Lin, E.C.C., Low, K.B., Magasanik, B., Reznikoff, W., Riley, M., Schaechter, M. and Umbarger, H.E., 1996. Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, 2nd Ed. American Society of Microbiology, Washington D.C. 9. Parker, A., Klauber, C., Kougianos, A., Watling, H.R. and van Bronswijk, W., 2003. An X-ray photoelectron spectroscopy study of the mechanism of oxidative dissolution of chalcopyrite. Hydromet. (in press). 10. Stott, M.B., Watling, H.R., Franzmann, P.D. and Sutton, D.C., 2001. The effect of solution chemistry on jarosite deposition during the leaching of chalcopyrite by the thermophilic archaeon, Sulfolobus metallicus. In: Ciminelli, V.S.T. and Garcia Jr., O. (eds.), IBS’01: Biohydrometallurgy: Fundamentals, technology and sustainable development: Proceedings of the International Biohydrometallurgy Symposium IBS’01 (Ouro Preto, Brazil), Elsevier Science B.V. Amsterdam, A, pp. 207-216. 11. Van der Mei, H.C., de Vries, J. and Busscher, H.J., 2000. X-ray photoelectron spectroscopy for the study of microbial cell surfaces. Surf. Sci. Rep. 39, 1-24. 12. Wilson, A.D., 1960. The micro-determination of ferrous iron in silicate minerals by a volumetric and a colorimetric method. Analyst (Lond.), 85, 823-827.

1022

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Analysis of chalcopyrite (CuFeS2) electrodes utilizing galvanic current in the presence of Acidithiobacillus ferrooxidans D. Bevilaquaa*, A.V. Benedettib, C.S. Fugivarab, O. Garcia Jr.a† a

Department of Biochemistry and Chemical Technology, Institute of Chemistry, São Paulo State University, P.O. Box 355, Araraquara,SP-14.801-970, Brazil b Department of Physical Chemistry, Institute of Chemistry, São Paulo State University, P.O. Box 355, Araraquara, SP 14.801-970, Brazil Abstract The oxidative dissolution of chalcopyrite electrodes by Acidithiobacillus ferrooxidans was studied utilizing a bacteriology battery, consisting of an electrochemical cell with two identical compartments in two different conditions: (a) a solution in a natural condition without any air or gas forced flux through the cell (non-air saturated) and (b) the solution in the two compartments was saturated in air and a flux of air was passed over the solution during the experiment. One compartment was sterile and the other inoculated with A. ferrooxidans. During the experiment time course (around 10 days) the potential and galvanic current was measured in a zero resistance ammeter. Results showed that the forced aeration caused a progressive increasing in the values of potential and current due the continuous bacterial activity, while a mineral dissolution limited by depletion of oxygen in the solution was observed. 1.

INTRODUCTION The microbiological leaching of chalcopyrite (CuFeS2) is of great interest because of its potential application to many CuFeS2-rich ore materials. At present, copper bioleaching processes are more efficient to other copper sulfides such as covellite, bornite, etc., because their leaching rates are much more favorable as compared to those of CuFeS2. The slow dissolution rate reflects the recalcitrance of chalcopyrite to the chemical and microbial oxidation [1, 2]. The slow kinetics of chalcopyrite dissolution is the result of the intrinsic thermodynamic properties of this mineral. The electrochemical behavior of chalcopyrite and a number of other sulfide ores have been investigated [3, 4, 5]. However the biological mechanisms and the reactions that come into play are still poorly understood. The purpose of the present work was to study the oxidation of a research-grade chalcopyrite electrode by A. ferrooxidans monitoring changes in the rest potential and in the galvanic current during exposure of the sulfide to the bacterial attack. After the end of * Financial support from the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) for the project and for fellowship to D.B. is gratefully acknowledged. Acknowledgments are also due to the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) (O.G. J. and A. V. B.) † Corresponding author ([email protected])

1023

Microbiology Fundamentals

experiments, the electrodes were also analyzed by scanning electronic microscopy in order to verify the bacterial behavior regarding its colonization pattern on the sulfide. This study is a part of a project designed to elucidate the fundamental basis and underlying mechanisms involved in the microbiological leaching of sulfide minerals. 2.

MATERIAL AND METHODS

2.1 Bacterial strain and growth conditions Acidithiobacillus ferrooxidans strain LR was used in this work [6]. The culture was grown in a mineral salts medium [7] at pH 1.8 using ferrous sulfate as energy source. The cell suspension for electrochemical experiment was obtained after growth for 48 hours in a shaker (150 rpm and 30ºC) by successive washing and centrifugation (5000·g) to eliminate residual ferric ion from the medium [1]. The cells were then twice washed in Milli-Q water (18 MΩ.cm) and suspended in 10 mL of the mineral salts solution, medium T&K [7], without ferrous iron in Milli-Q water. The cell suspension collected aseptically was standardized by the modified Lowry protein determination method [8]. 2.2 Mineral sample and electrode preparation Research-grade chalcopyrite was obtained from Ward’s Natural Science Establishment (Rochester, N.Y.). Samples were cut in pieces of approximately 1 cm2 using a diamond saw. One face was polished using silicon carbide emery paper with different particle size and finally polished with an alumina suspension of 0.3 µm. To eliminate impurities the samples were rinsed with acetone, ethanol and Milli-Q water (15 min each one in an ultrasonic bath), and then dried with pure Argon and kept in desiccators before using. 2.3 Electrochemical measurements A modified Tait type electrochemical cell consisting of two compartments, each one containing a massive electrode of chalcopyrite similarly prepared as possible was used in the galvanic current and potential measurements. The working electrodes were fixed at the bottom of each compartment of the cell with a Viton O-ring and the reference electrode was fixed at the top of the cell. An area of 0.28 cm2 of each working electrode was exposed to the solution, which was maintained at 30°C. The electrodes and the electrochemical cell were sterilized by UV exposure for 2 hours before each experiment. The electrolytic solution (10 mL) composition was (g.L-1): 0.4 (NH4)2SO4, 0.4 K2HPO4, 0.4 MgSO4.7H20 at pH 1.8 adjusted with H2SO4 and sterilized by autoclaving (1 atm, 120°C for 20 min). After a certain immersion time, one compartment was inoculated with A. ferrooxidans-LR (≈ 8.5 109 cells.mL-1) and the other one was used as a control (non-inoculated). Both compartments were connected via a glass tube containing the electrolyte solution, and a cellophane membrane was fixed at the bottom of the glass tube to impede the contamination of the control compartment with bacteria. The galvanic current was measured at zero resistance ammeter by a short-circuiting the auxiliary and reference electrodes outputs of the potentiostat, and imposing a potential difference of 0 ± 0.3 mV between the two mineral samples used for recording the current fluctuations. Ag/AgCl/KClsat electrode connected to the solution through a Luggin capillary was used as reference. A quiescent solution was used in two conditions: (a) a solution in a natural condition without any air or gas forced flux through the cell (non-air saturated) and (b) the solution in the two compartments was saturated in air and a flux of air was passed over the solution during the experiment (air saturated). The experiments were conducted for at least 1024

Microbiology Fundamentals

10 days, and then the electrodes were rinsed thoroughly with Milli-Q water to remove any remaining unattached bacteria and examined using a scanning electron microscope (SEM) Topcon SM-300. 3.

RESULTS AND DISCUSSION The Figure 1 shows the potential (E) and the galvanic current (Ig) vs. time curves obtained for chalcopyrite electrode in the presence of A. ferrooxidans-LR in non-air saturated condition. A decrease of the potential probably related to the dissolution of oxides, naturally formed on the chalcopyrite surface was observed. After addition of the cell suspension the potential increased up to 250 mV as a consequence of biofilm formation and remained almost constant (around 250 mV) up to the fifth day (region 1 of the figure). Afterward the potential decreased until 150 mV, the minimum value, and increased again up to around 250 mV (region 2). The minimum observed in the E vs. t curve can be associated with different concurrent processes: the bacterial-assisted dissolution of iron from the mineral, and depletion of oxygen in the nonair saturated condition by the bacteria (cathodic reaction), leading to a decrease in the bacterial activity with a consequent increase in the potential (region 3). This situation can be illustrated by the global reaction of chalcopyrite oxidation by A. ferrooxidans, as shown bellow: 2CuFeS2 + 8½ O2 + H2SO4 → 2CuSO4 + Fe2(SO4)3 + H2O In a previous work, relevant iron dissolution was also detected only after 4 days of inoculation [1]. The galvanic current decreased initially and stabilized at around 1 nA up to the fifth day and then increased reaching a maximum (at about 10 nA) which corresponds to minimum potential (region 2), indicating an increasing in the mineral dissolution. Increasing the immersion time the current decreased and stabilized at around 2 nA from the eighth day to the end of the experiment. The same explanation given for the potentialtime curve is valid.

Figure 1. Potential and the galvanic current versus time curves obtained for chalcopyrite electrode in non-air saturated condition. Arrows indicate the time of bacterial addition into the electrochemical cell 1025

Microbiology Fundamentals

The experiment with air saturated solution (condition b), the E vs. t curve (Figure 2) showed different behavior when compared with non-air saturated condition (Figure1). The potential increased about 30 mV during the first hours of immersion and no significant changes were observed until the end of the experiment, stabilizing around 340 mV. This potential-time profile can be associated with the presence of oxidative atmosphere over the solution during all the time course of the experiment, being the oxygen supplied to the solution by a diffusion process. The minimum potential and the maximum current values observed for non-aerated condition (Figure1) were not detected in the aerated experiment. The oxygen diffusion controlled process was previously observed using electrochemical impedance spectroscopy [9]. Figure 2 also depicts the galvanic current-time curve and a slow and continuous increasing was observed, probably due to the progress of the bacterial activity, since there was no restriction of oxygen supplying.

Figure 2. Potential and the galvanic current versus time (Eoc vs. t and Igalv vs. t) curves obtained for chalcopyrite electrode in air-saturated condition. Arrows indicate the time of bacterial addition into the electrochemical cell Figure 3 (A and B) shows the adhesion of bacteria on chalcopyrite surface at the end of experiment in air-saturated solution. As it can be seen mainly in Figure 3A a great number of cells of A. ferrooxidans-LR attached on the chalcopyrite surface involved by a biofilm, probably formed by exopolymeric substances. In Figure 3B (a zoom of Fig. 3A) an unusual massive presence of diplobacilli can be observed, which is probably related with an intense metabolic activity. In the non-air saturated experiment a similar pattern of adhesion was observed (data not shown) despite of electrochemical differences detected between both conditions. The E and Ig curves of chalcopyrite electrodes in the presence of A. ferrooxidans-LR showed significant differences depending on the test condition. For air-saturated experiment the dissolved oxygen promoted a continuous activity of bacteria whereas in non-air saturated test this activity seems to be limited, as revealed by the current peak after around four days of incubation.

1026

Microbiology Fundamentals

(Α)

(Β)

Figure 3. Scanning electron microscopy of the A. ferrooxidans-LR on the surface of chalcopyrite electrode in the experiment with air saturated solution at the end of electrochemical measurements (12 days). The bars indicate the magnification REFERENCES 1. D. Bevilaqua, A. L. L. C. Leite, O. Garcia Jr, O. H. Tuovinen, Process Biochem., 38 (2002) 587. 2. C.L. Brierley, Crit. Rev. Microbiol., 6 (1978) 207. 3. T. Cabral and I. Ignatiadis, Int. J. Miner. Process, 62 (2001) 41. 4. P. Velásquez, D. Leinen, J. Pascual, J. R. Ramos-Barrados, R. Cordova, H. Gómez, R. Schrebler, J. Electroanalytical Chem., 494 (2000) 87. 5. P. Velásquez, H. Gómez, D. Leinen, J. R. Ramos-Barrados, Colloids and Surface A, 140 (1998) 177. 6. O.Garcia, Jr., Rev. Microbiol., 20 (1991) 1. 7. O.H. Tuovinen and D.P. Kelly, Arch. Mikrobiol., 88 (1973) 285. 8. E.F. Hartree, Anal. Biochem., 48 (1972) 422. 9. D. Bevilaqua, I. Diéz-Perez, C. S. Fugivara, F. Sanz, A. V. Benedetti, O. Garcia Jr. Proceedings of 15th International Corrosion Congress, Granada, Spain (2002) 063.

1027

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Application of the bacterial weathering of silicate minerals in the improvement of quality of non-metallics I. Štyriakováa, I. Štyriakb∗ a

Department of Biotechnology, Institute of Geotechnics of the Slovak Academy of Sciences, Watsonova 45, 05343 Košice, Slovakia, E-mail: [email protected] b Department of Microbiology, Institute of Animal Physiology of the Slovak Academy of Sciences, Šoltésovej 4-6, 040 01 Košice, Slovakia Abstract Silicate mineral weathering is one of the important biogeochemical processes in nature. Heterotrophic bacteria of Bacillus spp. are often relatively active in these processes because they are limited due to the lack of organic nutrients as well as other inhibitory factors in environment. However, bacterial activity can be significantly increased by forming of optimal conditions in bioleaching processes (pH, T, Eh, concentration of biogenic elements). These processes can be used in the improvement of quality of several non-metallics. Our results indicate that removal of iron from non-metallics is the first positive marker, which is visible after bioleaching. Except this, we observed the release of contaminants from intergranular spaces of silicate grains, where Fe bearing minerals are often impregnated. Moreover, the enrichment of non-metallics by fine-grained fraction is the result of destruction of silicate minerals (mica, feldspars) and rounding-off of their edges. These findings suggest important applications of Bacillus strains in biotechnology involving the quality improvement of kaolin, quartz sands, zeolite and feldspars. Keywords: non-metallics, deferritization, silicate minerals, Bacillus spp. 1.

INTRODUCTION Weathering of rocks is biogeochemical process occurring at earth's crust. While both laboratory and field studies have demonstrated that microorganisms can play a role in mineral weathering, the magnitude of this effect is still unknown. Recent surveys of deep subsurface aquifers (hundreds of meters) show relatively high microbial numbers (approximately 105 to 107 cells/cm3). Microbial abundance and diversity is variable but does not decrease systematically with depth [1].

∗

Acknowledgments: The authors are grateful to the Slovak Grant Agency for Science (Grant No. 2/2107/22).

1029

Microbiology Fundamentals

The identity of members of communities of microorganisms in natural environments, their distribution, and the combined effects of their metabolisms on mineral dissolution are currently poorly understood [2]. Weathering of minerals in soil releases major nutrients such as K, P, Fe, Ca, Mg, Si, as well as trace ions which are necessary for microbial and plant growth [3]. Weathering action may be the result of oxidative or reductive attack of appropriately reactive mineral constituents (oxyhydroxides of Fe, Mn, sulphides) of a rock or mineral [4]. Silicate bacteria are very active (at pH 7) in the release of impregnated sulphides from the silicate matrix and sulphide minerals are accessible to destruction (at pH 4) by another bacterial species of acidophilic sulfur-oxidizing bacteria [5]. In the case of metal sulphide ore deposits, these activities can be and often are industrially exploited for metal recovery without smelting in a process called biohydrometallurgy. Feldspars, mica and quartz are the most abundant silicate minerals at the earth's crust. While the geomicrobiology of silicate mineral weathering might seem an extremely arcane subject, consider the following short list of processes affected by biogeochemically mediated cycles: soil formation and plant nutrition, stability of architectural materials, geochemistry of groundwater and movement of contaminants, long term stability of geological nuclear repositories, effects of mineral weathering on climate on a geological time scale [2]. However, in the case of silicate minerals weathering, this activity can be also industrially used for improvement of quality of non-metallic raw materials and treatment of wastes. 2.

MATERIALS AND METHODS

2.1 Granitic eluvium Granitic eluvium (GE), derived from granitic rocks (Slovakia). It is composed of quartz (55-60%), feldspar (40-35%) and mica (10-20%). 2.2 Quartz sands Quartz sands (QS) from Vyšný Petrovec deposit composed of quartz (80-90%), mica (5-10%), kaolinite (3-8%), rutil (2-5%) and siderite (4-6%). 2.3 Kaolin sample The kaolin sample from Vyšný Petrovec deposit composed of quartz (4-7%), mica (815%), kaolinite (60-75%), rutil (1-5%). The kaolin sample from Horná Prievrana deposit composed of quartz (70-80%), mica (6-14%), kaolinite (10-20%). The iron-bearing minerals decrease the quality of these raw materials. 2.4 Bacteria and media Two bacterial strains of Bacillus spp. (Bacillus cereus and Bacillus pumilus) were isolated from a kaolin quarry in Horná Prievrana. They were identified by means of the Becton-Dickinson microbiology system (Becton Dickinson, USA). For the species identification, the strains were cultivated on Columbia agar plates according to recommendation of the system producer. This panel contains 29 enzymatic and biochemical substrates and a fluorescence control on tips of plastic prongs. The resulting pattern of the 29 reactions is converted into a ten-digit profile number that is used as the basis for identification of a wide variety of microorganisms. For experiment, the bacterial strains were grown in Nutrient broth No.2 (Imuna) at 37°C for 18 hours. Bacterial cells were 1030

Microbiology Fundamentals

subsequently centrifuged at 4000 rpm for 15 min, subsequently washed twice with saline solution (0.9% NaCl) and added in a concentration of 1010 cells per ml to modified Bromfield liquid medium. Bioleaching of the samples was carried out in 3000 ml Erlenmeyer flasks containing 2000 ml of modified Bromfield medium (NaH2PO4 – 0.5g/l, MgSO4.7H2O - 0.5g/l, (NH4)2SO4 – 1.0g/l, NaCl – 0.2g/l, molasses – 0.3g/kg) inoculated with a mixture of both Bacillus cereus and Bacillus pumilus strains. The flasks were incubated statically for 28-120 days at 28°C. The abiotic controls were cultivated under the same conditions. After incubation, the culture solutions were separated from the biomass by means of membrane filtration. The presence of vegetative bacterial cells in Erlenmeyer flasks and their morphology were regularly examined by light microscopy after Gram staining. 2.5 Chemical analyses Quantitative changes of samples (solid and liquid phases) investigated from view of element composition stability were evaluated by standard analytical method – atomic absorption spectrometry on a VARIAN spectrometer AA - 30 apparatus (Varian, Australia) after dissolution of the samples by standard procedure. 3.

RESULTS AND DISCUSSION The theses on weathering process as the most important geo-chemical process in the hypergene zone that causes material–structural changes of rocks and minerals, essentially come out of the numerous studies, but there are still a lot of unclear issues concerning the weathering stages, sequence of rock changes and participation of microorganisms in genesis of minerals, including clay ones. We have started the study of biogenous environmental transformation processes on the basis of similar considerations, which were also formulated by Ján Čurlík [6], which concerned the fact that so far there is not any precisely substantiating theory that could explain the genesis of eluvial processes and creation of mineral zones of weathering crusts. Microorganisms play an important role in the dissolution of silicate structure in the rock weathering process and in the genesis of clay minerals. The activity of heterotrophic bacteria and the release of organic compounds into circulating solutions in rock environment are displayed in nature by zonal formation of a deposit. In the porcelain clay deposit in Horná Prievrana, there is possible to observe the zones where the kaolin zone and the Fe oxidation zone with its typical mineralization are evident. The relicts of the bedrock of phyllites are also evident. The mire samples from Horná Prievrana kaolinite pit are characteristic also by a high number of bacteria, especially of Bacillus species. There are found bacteria of Bacillus genus with a wide range of species as B. cereus, B. pumilus a B. megaterium and B. mycoides, which coexist together. They are relatively active under aerobic and anaerobic conditions in environment with higher humidity [7]. The dissolution activity of selected Bacillus strains from these mire samples was investigated with non-metallics materials contained feldspars, quartz, kaolinite, mica and oxyhydroxides of iron. The chemical analyses as well as other methods indicated the ability of bacteria to destruct these silicates and to release iron and other elements. This fact suggests that these processes can be used in the removing of iron from many nonmetallics. 3.1 Granitic eluvium containing feldspars Experimental studies of the effect of microorganisms on mineral weathering of granitic eluvium with feldspars were performed in closed (batch) reaction vessels. In batch 1031

Microbiology Fundamentals

experiments, solutions were monitored over time, and solid phases were analysed in the end of experiment. Feldspars in granitic eluvium (GE) sample contain 1,44 wt.% iron. Iron was found especially in the form of Fe oxyhydroxides. The presence of iron minerals was identified also by visible brown-red colour of this material. The releasing of Fe was characteristic by the change of visible brown-red colour of these non-metallic materials to brown-white. The Bacillus strains used in this study markedly decreased Si content (from 73,3 to 53,5%) and partially decreased also content of potassium (from 2,48 to 2,15%) after 120 days’ bioleaching what resulted to a weak enrichment of Al content (from 11.51 to 11.73%) and destruction of feldspars. After 120 days of bioleaching, 31% Fe extraction was detected in GE. The release of Fe from GE suggests the removal of the cementing iron oxides from the silicate surface. This partial Si and Fe removal can also improve the access of organic acids to iron-bearing minerals and enable the increase of the feldspar raw material quality by subsequent leaching with less concentrated organic acids. That is why bacterial destruction of intergranular spaces of silicate grains, often impregnated with iron minerals, should be used as pretreatment of raw material before chemical treatment of the sample by oxalic acid. Bobos and Gomes [8] investigated K-feldspars weathering at the kaolin deposit of Sao Vicente de Pereira (Portugal). Their analyses indicated that the weathering products were typical by high content of Al and small contents of Si, Fe, and K. Moreover, altered Kfeldspar grains exhibit "porous" surfaces. This fact indicates that the K-feldspar has been rapidly dissolved or corroded by solutions with an acid pH, whereas no secondary mineral phases were produced. Mica and feldspars are the main constituents of rocks which were weathered into secondary minerals, according to the environmental conditions. Feldspars (especially Kfeldspars) are often considered as relatively stable, and phyllosilicates (mica or chlorite) are the main source of clay minerals, such as vermiculite or smectite. In humid tropical climates, both micas and feldspars are completely transformed into kaolinite and/or gibbsite, and iron oxides [9]. Hydrolysis is the process on which governs weathering. If intense leaching of silicates by water occurs, it may lead to dissolution of the structure, desilicification and precipitation of new secondary minerals (generally 1:1 clay minerals or oxides and hydroxides). This process is the main source of clays in tropical areas. Feldspars and micas experimentally weathered by water show pits in their structure in addition to formation of hydroxides of Al and Fe [10]. Weathering is inextricably bound to biological processes, for organisms inhabit a wide range of nichens in surface and subsurface environments and influence various mineral transformation reactions [11,12]. 3.2 Quartz sands Quartz sands are largely used as a raw material in the glass industry, however, are often associated with iron and titanium impurities that decrease their economic value and hinder their application. Iron impurities lower an important technological factor related to the degree of whiteness [13,14, 15]. In the first stage of bioleaching of the QS sample, there were removed visible Fe-bearing minerals coated over quartz grains. Moreover, the extraction of elements was observed due to the destruction of silicate minerals and intergranular spaces of brown polymineral grains with releasing of ferrihydrite sealing siderite nodules. Chemical analyses of the bioleached QS

1032

Microbiology Fundamentals

sample showed the slight increase of Fe2O3 content from 0.15 to 0.19% and FeO content decrease from 0.19 to 0.1% after 3 months bioleaching. Bacterial destruction of these polycomponent grains (fraction size 0.1-0.8 mm) caused the formation of a brown fine-grained fraction of mineral particles under 0.1 mm of size. This brown fine-grained fraction with a high Fe-bearing mineral content was removed from QS sample by simple elutriation process. Chemical analyses of QS show 47% removing of iron by bioleaching and elutriation. Upgrading of silica sand for the production of clear flat glass requires a partial removal of iron, typically present in the form of iron oxide as discrete particles or coatings on the surface of the sand particles. Traditional technology involves leaching with heated sulphuric acid and producing a large quantity of acidic leachate, which is costly to treat and discharge [16]. Other chemical methods consist of leaching with mineral acids, and treatment with reducers, such as sodium dithionite plus aluminium sulfate or sulphur dioxide plus aluminium powder. These bleaching methods are usually suitable for achieving a higher degree of iron removal but are more expensive, have complex operating conditions, and are environmentally hazardous [17]. Our experiments confirmed that heterotrophic bacteria of Bacillus spp. play active role at biodestruction as well as at deferritization of quartz sands. Therefore, their activity should be used in the improvement of quartz sands quality by a new progressive biotechnology. 3.3 Kaolin sample The release of Fe from mica or from oxyhydroxides was used as indicator of mineral dissolution and beneficiation of kaolin quality in our previous study [18]. Kaolin raw material composed of 0.92% - 1.43% iron, which was found especially in the form of Fe oxyhydroxides (free Fe). The presence of Fe oxyhydroxides was identified by visible brown-red colour of this material. However, a substantial part of iron in kaolin is bound in the aluminosilicate lattice. The Bacillus strains removed about 43% of free iron under anaerobic conditions from kaolin sample of brown-red colour within 28 days. The removal of structural Fe ions from kaolin of white colour under similar conditions was not so efficient because only about 15% of Fe was removed within 28 days. This fact can also suggest more difficult availability of iron for bacterial removal because of probable binding of iron in mica. Probable mechanisms involved in the biological removal of iron from elutriated kaolin were examined on kaolin sample from Vyšný Petrovec. The prolongation of bioleaching time from 1 month to 3 months and thus longer production of organic acids and metal-complexing substances by Bacillus strains caused the increase of iron removal from mica. There was observed after 3 months the 52% extraction of Fe atoms from octahedral position in mica when Al removal was only about 2%. Kaolin mined at Vyšný Petrovec deposit is economically interesting raw material but its properties could be improved by removal of unsuitable minerals (mica, oxyhydroxides). After such an improvement, it could be a suitable raw material for more extensive use in ceramic industry. Several strains of bacteria released by this way the cations from biotite (Si, Fe, Al), plagioclase, and feldspar (Si, Al) much more than abiotic procedures [2]. Microbial 1033

Microbiology Fundamentals

production of organics by fermentation, or reductive dissolution of Fe – Mn mineral phases can greatly accelerate weathering rates of aluminosilicate minerals [19, 20]. 4.

CONCLUSION The iron content of industrial minerals can be reduced by a number of physical, physicochemical and chemical methods. The appropriate method for the removal of iron from an industrial mineral depends on the mineralogical forms and the distribution of iron in the particular ore [21]. The red and yellow pigmentations noticed in many clay deposits are mainly due to the associated oxides, hydroxides and hydrated oxides of ferric iron such as hematite (red), maghemite (reddish brown), goethite (brownish yellow), lepidocrocite (orange), ferrihydrite (brownish red), etc. These oxides and hydroxides occur either as coatings on individual grains or as discrete fine particles throughout the clay mass. Quantities as low as 0.4% of ferric iron may be sufficient enough to impart colour to the deposit. Removal of these associated impurities improves the quality of the material. The beneficiated clay finds use in paper coating, paper filling and as extenders in paints and polymers [22]. Considerable efforts have been devoted to the problem of removing ferric contaminants by physical [23] and chemical means. High intensity magnetic separation is a standard method used, removing substantial quantities of iron and titanium as mineral impurities, consequently improving the brightness of silicates. The chemical methods consist of leaching with mineral acids, and treatment with reducers, such as sodium dithionite plus aluminium sulphate, sulphur dioxide plus aluminium powder, and sulphur dioxide plus zink powder. These bleaching methods are usually suitable for achieving a higher degree of iron removal but they are more expensive, have complex operating conditions, and are environmentally hazardous [17]. This paper also reports a possibility of microbial removal of iron from non – metallics and improvement of silicate properties via the action of bacterial destruction of Bacillus spp. Our results indicate the removal of visible iron coatings from non-metallics as well as iron contaminants from intergranular spaces of silicate grains. Moreover, the enrichment of non-metallics by fine-grained fraction is the result of destruction of silicate minerals (mica, feldspars) and rounding-off of their edges. It is also possible to accelerate bioleaching process by continual monitoring of sugars consumption from molasses, which was used as the energy source during samples bioleaching. REFERENCES 1. D.J. Balkwill, Geomicrobiology J., 7 (1989) 33. 2. W.W. Barker, S.A.Welch, S. Chu and J.F. Banfield, In: Banfield JF, Nealson KH (Eds) Geomicrobiology: Interactions between microbes and minerals, Reviews in Mineralogy, Vol. 35, Washington, D.C., USA, 1997 3. P. Hinsinger, B. Jaillard and J.E. Dufey, Soil Soc Am J, 56 (1992) 977. 4. H.L. Ehrlich, Chem Geol, 132 (1996) 1. 5. I. Štyriaková, I. Štyriak and M. Kušnierová, In: R. Amils, A. Ballester (eds) Process Metallurgy 9A, Elsevier, 1999. 6. Čurlík,J., 1988: Geochémia geologických procesov. Hypergénne procesy. Prírodovedecká fakulta Univerzity Komenského, Bratislava. 7. I. Štyriaková, I. Štyriak, Mineralia Slovaca, 34 (2002) 99. 8. I. Bobos and C. Gomes, Geologica Carpathica, 51 (2000) 49. 9. R. Romero, M. Robert, F. Elsass and C. Garcia, Clay Minerals 27, (1992) 21. 1034

Microbiology Fundamentals

10. M. Robert, D. Tessier, In: I.P. Martini and W. Chesworth (eds.): Weathering, Soils and Paleosols. ELSEVIER, 1992. 11. J.F. Banfield and R.A. Eggleton, Clays and Clay Minerals, 36 (1988) 47. 12. P.M. Huang and M. Schnityer, Soil Science Society of America Special Publication no.17 (1986) 606. 13. L.G. Yan, Y.J. Yu, S.S. Song, H.L Nan, Y.L. Cheng, X.M. Li, Q.W. Kong and Y.M Dai, Proc. 2nd World Congress on Non-Metallic Minerals, Beijing, China, 1987. 14. F. Veglio and L. Toro, International Journal of Mineral Processing, 39 (1993) 87. 15. F. Veglio, Hydrometallurgy, 45 (1997) 181. 16. I.I.Tarasova, A.W.L. Dudeney and S. Pilurzu, Minerals Engineering, 14 (2001) 639. 17. L.M.S. Mesquita, T. Rodrigues and S.S. Gomes, Minerals Engineering, 9 (1996) 965. 18. I. Štyriaková and I. Štyriak, Ceramics Silikáty, 44 (2000), 135. 19. S.A.Welch, W.J. Ullman, Geochim Cosmochim Acta 60 (1996) 2939. 20. P.C. Bennett, F.K. Hiebert and W.J. Choi, Chem Geol 132 (1996) 45. 21. M. Taxiarchou, D. Panias, I. Douni, I. Paspaliaris, and A. Kontopoulos, Hydrometallurgy, 46 (1997) 215. 22. V.R. Ambikadevi and M. Lalithambika, Applied Clay Science 16 (2000) 133. 23. S.J.F. Guimares, N. De Oliveira, F.L. De Salles, Trans Inst. Min. metall. 51 (1987) 13.

1035

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Assessment of acid production potential of sulphide minerals using Acidithiobacillus ferrooxidans and microbial sulphate reduction using Desulfotomaculum nigrificans Evvie Chockalingam, S. Subramanian, K.A. Natarajan and J.J. Brauna Department of Metallurgy and aIndo-French Centre for Water Sciences, Indian Institute of Science, Bangalore – 560 012, India Abstract The acid production potential of pyrite and chalcopyrite minerals has been assessed in the presence of Acidithiobacillus ferrooxidans in iron-free Silverman and Lundgren medium. The parameters such as pH, redox potential, bacterial cell number, ferrous, ferric and copper concentrations were monitored as a function of time. The inoculated samples show significant changes in the values of the parameters studied relative to the control samples. The pH values decrease to 1.6 after 100 days for pyrite and to 2.5 after 150 days for chalcopyrite respectively, from an initial value of about 7, after inoculation with Acidithiobacillus ferrooxidans cells. The ferrous ion concentration is found to be higher vis-à-vis that of ferric for both the mineral samples. The kinetics of copper dissolution from chalcopyrite is found to be slower compared to the dissolution of iron. Significant sulphate reduction and copper precipitation could be achieved using Desulfotomaculum nigrificans, both at pH 7 and 5.5. An increase in the cell number enhances the kinetics of sulphate reduction and copper precipitation. Keywords: acid production potential, Acidithiobacillus ferrooxidans, pyrite, chalcopyrite, Desulfotomaculum nigrificans 1.

INTRODUCTION A major environmental concern to the mining industry, both during the operational period and after closure is the generation of acidity, resulting from the oxidation of sulphide mineral wastes, catalyzed by acidophilic microorganisms. The dearth of basic or buffering materials to neutralise the acid, results in the leach water becoming acidic, leading to dissolution and build up of metal species. This phenomenon known as acid mine drainage, needs to be mitigated by suitable strategies. The technology employed for the treatment of acid mine drainage includes pH adjustment, chemical precipitation, aeration and settling. Several active and passive treatment methods, including the use of soil and water covers have been investigated to curtail the oxidation of sulphidic wastes [1]. Biotechnological approaches using sulphate-reducing bacteria (SRB) have also gained importance in recent times [2-4]. The treatment of metal-contaminated wastes by sulphatereducing bacteria has been investigated by several workers [5-7]. Specific technologies that have been successful in using SRB to remediate ground water contaminated with heavy metals include the Paques’ (THIOPAQ) system [8,9] and the NTBC bio-sulphide 1037

Microbiology Fundamentals

process [10]. In an ongoing research programme, environmental impact studies have been undertaken with reference to an abandoned copper mine. Keeping these objectives in mind, acid production potential studies on pyrite and chalcopyrite minerals using Acidithiobacillus ferrooxidans have been performed. Sulphate reduction studies have been carried out in modified Barr’s medium using Desulfotomaculum nigrificans, in the simultaneous presence of copper ions, as a function of pH and bacterial cell number. The variation in the sulphate and copper concentrations was monitored as a function of time. 2.

EXPERIMENTAL

2.1 Materials 2.1.1 Mineral samples Pure mineral samples of pyrite and chalcopyrite were obtained from Alminrock Indscer Fabriks, Bangalore, India. The samples were subjected to dry grinding in a porcelain ball mill and the < 74 µm size fraction was used for the experiments. 2.1.2 Bacterial strains The strain of Acidithiobacillus ferrooxidans was isolated from the ore samples of the Hutti gold mines, Karnataka, India, and used after purification. A pure strain of Desulfotomaculum nigrificans was obtained from the National Collection of Industrial Microorganisms, National Chemical Laboratory, Pune, India. 2.2 Methods 2.2.1 Medium and growth of Acidithiobacillus ferrooxidans The bacteria Acidithiobacillus ferrooxidans were grown in 9K medium [11] which has the following composition in grams per litre of solution: (NH4)2SO4, 3.0; MgSO4.7H2O, 0.5; K2HPO4, 0.5; KCl, 0.1; Ca(NO3)2, 0.1 and FeSO4.7H2O, 44.2. The pH was adjusted to 2 initially by the addition of sulphuric acid. The fully-grown culture of Acidithiobacillus ferrooxidans from the 9K medium was centrifuged to harvest the cells. The cells were then resuspended in pH 2 sulphuric acid to remove iron. This procedure was repeated twice to get iron free cell suspension. The iron-free Acidithiobacillus ferrooxidans cells thus obtained were used as inoculum for the acid production potential studies. 2.2.2 Medium and growth of Desulfotomaculum nigrificans The bacteria Desulfotomaculum nigrificans were grown in the modified Barr’s medium (MBM) [12] which contained 0.5 g K2HPO4, 1 g NH4Cl, 1 g CaSO4, 2 g MgSO4.7H2O, 5 g sodium lactate and 1 g yeast extract in 1000 mL distilled water. The pH of the Barr’s medium was adjusted to 7.5 and sterilized for 3 consecutive days. The pH was checked every time before sterilization and adjusted to the desired value. Before inoculation, for each 100 ml of the medium, 0.05 g of FeSO4(NH4)2SO4.6H2O and 0.01 g of sodium thioglycollate was added and filter sterilized. The media was stored in 250 ml serum bottles stoppered with a tight fitting rubber cork. To each 90 ml of the abovementioned media, 10 ml of an active culture was added through a sterilized syringe. Immediately after inoculation, nitrogen gas was flushed into the medium for 5 minutes through a syringe and sealed tightly with parafilm to maintain anaerobic conditions. The bottles were then incubated at 37°C in a rotary shaker at 160 rpm. The fully-grown culture 1038

Microbiology Fundamentals

was centrifuged and the supernatant was discarded. The cells were resuspended in distilled water. This process was repeated twice and the washed cells were used as inoculum for the sulphate reduction studies. 2.2.3 Assessment of acid production potential The acid production potential of pyrite and chalcopyrite minerals was evaluated adopting the following procedure: 5 g of the mineral sample (< 74 µm) was suspended in 100 ml of 9K- medium (without ferrous sulphate) containing ≈ 3 x 108 cells/ml under natural pH condition, in a 250 ml Erlenmeyer flask. The flask was agitated at 240 rpm in a Remi rotary shaker at 30 + 2°C. The parameters such as pH, ESCE, bacterial cell number, ferrous, ferric and copper concentrations were monitored periodically. A control sample was also maintained in which 10% alcoholic thymol was added as a bactericide. 2.2.4 Sulphate reduction and copper precipitation kinetics using Desulfotomaculum nigrificans In these experiments, the variation in the sulphate and copper concentrations was measured at pH ≈ 5.5 and 7 using different cell densities (≈ 107, 108 and 109 cells/ml), during growth of Desulfotomaculum nigrificans in the modified Barr’s medium. 10 ppm of copper was added as cupric sulphate pentahydrate to the MBM for these studies. 2.2.5 Analytical procedures The bacterial cell count was determined using a Petroff-Hauser counter attached to a Leitz phase contrast microbiological microscope (Laborlux K Wild Mps12). The ferrous and total iron concentrations were analysed using the orthophenanthroline method [13] in a Shimadzu model UV260 uv-visible spectrophotometer. The total iron concentration was determined by reducing the ferric iron in solution to ferrous by adding hydroxylammonium chloride. The ferric concentration was then estimated as the difference between total iron and ferrous concentrations. Copper was analysed using a Video 11E Thermo-Jarell Ash atomic absorption spectrophotometer. Sulphate concentration was determined by turbidimetric method using barium chloride at a wavelength of 420nm [14] using a Shimadzu model UV260 uv-visible spectrophotometer. 3.

RESULTS AND DISCUSSION

3.1 Acid production potential studies The acid production potential of pyrite and chalcopyrite in the absence and presence of Acidithiobacillus ferrooxidans was assessed by monitoring the variation in the pH, redox potential, bacterial cell number, ferrous, ferric and copper concentrations as a function of time. It must be emphasized that these experiments were carried out at a natural pH of 6-7, in order to ascertain the propensity for acid generation, with or without bacterial cells. 3.1.1 pH The variation of pH as a function of time for pyrite and chalcopyrite suspensions in 9K medium for the inoculated and control samples is shown in Figure 1. In the case of the pyrite sample, the pH steeply decreases from the initial value of ≈ 7 to ≈ 2 in about 50 days. The pH marginally decreases thereafter to ≈ 1.6 and remains more or less constant upto about 220 days. However, in the case of the control sample for pyrite, the pH marginally decreases to ≈ 6 in 50 days and is further lowered to about 4.5 in about 100 1039

Microbiology Fundamentals

days and remains at that value for the time period investigated. With respect to chalcopyrite, the pH steadily decreases from ≈ 6 to ≈ 2.5 in about 150 days and remains more or less unaltered thereafter. In the case of the control sample for chalcopyrite, the pH remains close to the initial value of 6 upto 90 days and is subsequently decreased to ≈ 4.6 in 150 days and remains at that value upto 220 days. These results highlight that acidity is generated in the case of both pyrite and chalcopyrite after inoculation with Acidithiobacillus ferrooxidans. On a comparative basis, the acid production potential of pyrite is higher than that of chalcopyrite after bacterial interaction. 3.1.2 Redox potential Figure 2 depicts the redox potential changes as a function of time for pyrite and chalcopyrite suspentions in 9K- medium. The ESCE value increases sharply from 100 mV to ≈ 400 mV in about 40 days for pyrite after interaction with the bacterial cells. The redox potential further increases to ≈ 425 mV in 50 days and remains unchanged upto about 220 days. In the case of the control sample for pyrite the ESCE increases from the initial value of 100 mV to ≈ 180 mV in 50 days. Beyond that period, the redox potential marginally increases and attains a value of 190 mV. The redox potential of the chalcopyrite sample after inoculation with Acidithiobacillus ferrooxidans steadily increases from 100 mV to ≈ 250 mV in about 90 days. A sharp increase is observed to a value of 500 mV in 130 days beyond which there is not much of a change. With respect to the control sample for chalcopyrite, the redox potential is more or less unaltered at 100 mV upto about 90 days. Thereafter, a steady increase is observed upto about 180 days and the redox potential attains a value of 270 mV. As can be expected, the redox potential values are higher in the case of the inoculated samples attesting to bacterial oxidation of the mineral samples. It is pertinent to mention that the redox potential values observed in these tests are lower than that commonly encountered in typical leaching systems. This is understandable as the initial pH of the system was close to neutral (pH 6-7), in contrast to a typical leaching system, wherein the pH is maintained close to 2-2.5. 900

-

Pyrite in 9K medium Pyrite in 9K medium + Af cells Chalcopyrite in 9K medium Chalcopyrite in 9K medium + Af cells

10

pH

6

4

2

600 450 300 150

0 0

50

100

150

200

Time (days)

Figure 1. pH as a function of time for pyrite and chalcopyrite mineral suspensions

1040

750

ESCE (mV)

8

-

Pyrite in 9K medium Pyrite in 9K medium + Af cells Chalcopyrite in 9K medium Chalcopyrite in 9K medium + Af cells

0 0

50

100

150

200

Time (days)

Figure 2. Redox potential as a function of time for pyrite and chalcopyrite mineral suspensions

Microbiology Fundamentals

3.1.3 Bacterial cell number The changes in the bacterial cell number after interaction with pyrite and chalcopyrite mineral samples as a function of time are portrayed in Figure 3. For both the samples, a decrease in the cell number is observed initially, presumably due to the attachment of the cells onto the mineral substrates. In the case of pyrite, the cell number steeply increases from 2 x 107 cells/ml to 8 x 108 cells/ml in about 60 days and thereafter remains unchanged. For chalcopyrite the cell number marginally increases from 7 x 107 to 2 x 108 cells/ml after 220 days. It is interesting to note that there is an increase in cell number after interaction with pyrite while after interaction with chalcopyrite there is a marginal decrease. The dissolution of copper ions from chalcopyrite as a function of time could plausibly lead to toxicity impeding cell growth. 10

Number of cells/ml

10

Pyrite Chalcopyrite 9

10

8

10

7

10

0

50

100

150

200

250

Time (days)

Figure 3. Bacterial cell number as a function of time for pyrite and chalcopyrite mineral suspensions 3.1.4 Dissolution kinetics of ferrous and ferric ions The changes in the ferrous and ferric concentrations as a function of time for pyrite are shown in Figure 4a. The ferrous concentration steeply increases initially upto about 50 days and attains a value of ≈ 6g/l in the case of the inoculated sample. Around 170 days, the ferrous concentration attains a value of 7g/l and remains at that value upto 220 days. The ferric concentration steadily increases to ≈ 3g/l in about 50 days and thereafter remains unchanged in the case of the inoculated sample. With respect to the control samples, the ferric concentration slightly increases to ≈ 0.5g/l in about 75 days and remains unaltered thereafter. The ferrous concentration in the control sample increases to about 1g/l in 75 days and further increases to about 1.5g/l between 160-220 days. It is evident that the ferrous concentration is higher in both the inoculated and control samples compared to the ferric concentrations. These results are in agreement with those reported earlier [15-17]. It is also pertinent to mention that the presence of the pyrite substrate facilitates the preferential oxidation of sulphur in the mineral, rather than the ferrous ions in solution, through the direct contact mechanism. The dissolution of iron species from chalcopyrite is portrayed in Figure 4b. Almost a similar trend as that observed for pyrite is revealed in this case also. In the case of the inoculated sample, the ferrous concentration shows a marginal increase upto about 60 days and subsequently a steep increase to 3.5 g/l is observed upto 160 days. Beyond that period, 1041

Microbiology Fundamentals

there is not much of a variation in the ferrous concentration. The ferric concentration gradually increases to about 1 g/l in 150 days and remains at that value for the inoculated sample. The ferrous concentration in the control sample marginally increases to about 0.75 g/l in 150 days and more or less remains unchanged thereafter. On the other hand, the ferric concentration in the control sample remains more or less constant at 0.25 g/l in the time frame investigated. A closer examination of Figures 4a and 4b highlights that the dissolution rate of iron species from pyrite is higher vis-à-vis that of chalcopyrite. The lower redox potential values observed in Figure 2 are consistent with the low ferric/ferrous ratios observed (Figs. 4a and 4b).

8 6

4

4

2

2

2+

2+

3+

6

0

0 0

50

100

150

2+

-

Fe in 9K medium 2+ Fe in 9K medium + Af cells 3+ Fe in 9K medium 3+ Fe in 9K medium + Af cells

5

6 5

4

4

3

3

2

2

1

1

3+

8

6

10

0

0

200

0

50

Time (days)

100

150

200

Time (days)

Figure 4a. Dissolution of ferrous and ferric from pyrite as a function of time

Figure 4b. Dissolution of ferrous and ferric from chalcopyrite as a function of time

3.1.5 Dissolution kinetics of copper ions Figure 5 depicts the kinetics of dissolution of copper from chalcopyrite in the case of the inoculated and control samples. The amount of copper dissolved is very low in the first 50 days for both the samples. The copper concentration steadily increases to about 3 g/l in the case of the inoculated sample after 220 days. However, the concentration of copper remains unchanged upto about 170 days at 0.2 g/l and thereafter shows a marginal increase to about 0.5 g/l in the case of the control sample. It becomes of interest to note from Figures 4b and 5 that the dissolution rate of ferrous is higher compared to that of copper from chalcopyrite. Similar results have been reported by other workers [15]. 5

Chalcopyrite

Cu conc. (g/l)

4

-

9K medium 9K medium + Af cells

3

2

1

0 0

50

100

150

200

Time (days)

Figure 5. Dissolution of copper from chalcopyrite as a function of time 1042

Fe conc. (g/l)

-

Fe conc. (g/l)

2+

7

7

Chalcopyrite

Fe in 9K medium 2+ Fe in 9K medium + Af cells 3+ Fe in 9K medium 3+ Fe in 9K medium + Af cells

10

Fe conc. (g/l)

12

Pyrite

Fe conc. (g/l)

12

Microbiology Fundamentals

3.2 Studies on sulphate reduction and copper precipitation using Desulfotomaculum nigrificans Studies have been carried out to assess the efficacy of Desulfotomaculum nigrificans to bring about the reduction of sulphate as well as the precipitation of copper. These studies are of relevance to the mitigation of acid mine drainage. The effect of pH and cell number on the reduction kinetics was assessed. The growth of the organism was carried out in the modified Barr’s medium [12] in the presence of copper ions (≈ 10 ppm). Figure 6 shows the change in sulphate concentration as a function of time and cell number. In these tests, the initial pH was maintained at 7. It is evident from the figure that the sulphate concentration decreases rapidly upto about 9 days in all cases and gradually thereafter. Further, the rate of sulphate reduction increases with increase in the cell number. For example, the sulphate reduction increases from 18.5% to 24.5% to 31.5% as the cell number is increased from ≈ 107 to ≈ 109 cells/ml. The change in copper concentration as a function of time and cell number was also monitored. The results are depicted in Figure 7. It is noteworthy that, complete precipitation of copper takes place by 3, 4 and 6 days respectively as the cell number is increased from ≈107 to 109 cells/ml. It was found that the pH increases from the initial value of 7 to about 7.8 after 4 days in all the samples and subsequently decreased back to 7. This may be attributed to the production of H2S by bacterial metabolism. Initial pH = 7 + 0.2

7

1500

1350

1200

1050

7

Initial pH = 7 + 0.2

12

1.4 x 10 cells/ml 8 1.1 x 10 cells/ml 9 1.6 x 10 cells/ml

1.4 x 10 cells/ml 8 1.1 x 10 cells/ml 9 1.6 x 10 cells/ml

10

Cu conc. (ppm)

Sulphate conc. (ppm)

1650

8 6 4 2

900

0 0

2

4

6

8

10

12

14

16

Time (days)

Figure 6. Bioreduction of sulphate as a function of time and bacterial cell number at pH 7

0

1

2

3

4

5

6

Time (days)

Figure 7. Bioremoval of copper as a function of time and bacterial cell number at pH 7

The feasibility of sulphate reduction at pH 5.5 was ascertained and the results are portrayed in Figure 8. It is apparent that, beyond 7 days the amount of sulphate reduced, increases with increase in cell number. A comparison of Figures 6 and 8 indicates that the rate of sulphate reduction is lower when the pH is reduced to 5.5 from 7. Prolonged experiments lasting for 57 days carried out at pH 7 and pH 5.5 respectively resulted in 95% and 91% sulphate reduction using ≈ 109 cells/ml. The variation of copper concentration as a function of time for experiments carried out at pH 5.5 at different cell densities is shown in Figure 9. It is evident that complete precipitation of copper takes place by 6, 8 and 15 days respectively as the cell number is increased from ≈107 to 109 cells/ml. As anticipated, the copper precipitation kinetics is 1043

Microbiology Fundamentals

slower at pH 5.5 when compared to pH 7 (Figures 7 and 9). The results presented in Figures 6 to 9 unequivocally demonstrate that Desulfotomaculum nigrificans is capable of significant sulphate reduction and complete precipitation of copper. The mechanisms of sulphate reduction and metal ion precipitation are well documented [18, 19]. 1.2 x 10 cells/ml 8 1.6 x 10 cells/ml 9 1.5 x 10 cells/ml

1650

1.2 x 10 cells/ml 8 1.6 x 10 cells/ml 9 1.5 x 10 cells/ml

12 10

1500

1350

1200

8 6 4 2

0

2

4

6

8

10

12

14

16

Time (days)

Figure 8. Bioreduction of sulphate as a function of time and bacterial cell number at pH 5.5 4.

7

Initial pH = 5.5 + 0.2

Cu conc. (ppm)

Sulphate conc. (ppm)

14

7

Initial pH = 5.5 + 0.2

0 0

2

4

6

8

10

12

14

16

Time (days)

Figure 9. Bioremoval of copper as a function of time and bacterial cell number at pH 5.5

CONCLUSIONS From the results of the present investigation, the following conclusions can be arrived

at: 1. Acid production potential studies reveal that the pH values of pyrite and chalcopyrite suspensions are decreased from ≈ 7 to about 1.6 and 2.5 respectively after inoculation with Acidithiobacillus ferrooxidans cells. 2. An increase in the bacterial cell number is observed in the case of pyrite while for chalcopyrite there is a marginal decrease in the cell number. 3. In the case of both pyrite and chalcopyrite minerals, the dissolution of ferrous ion is substantially higher compared to ferric ion. 4. The amount of copper dissolved from chalcopyrite is lower than that of ferrous. 5. About 95% and 91% sulphate reduction could be achieved at pH 7 and 5.5 respectively, using ≈ 109 cells/ml of Desulfotomaculum nigrificans. 6. Complete precipitation of copper could be achieved both at pH 5.5 and 7 using Desulfotomaculum nigrificans cells, though the kinetics of precipitation is slower at pH 5.5. 7. The kinetics of sulphate reduction and copper precipitation is found to increase with increase in the bacterial cell number. ACKNOWLEDGEMENTS The authors are grateful to the Indo-French Centre for Promotion of Advanced Research (CEFIPRA) and the Institute for Research and Development (IRD) France, for grant of research projects to carry out this investigation. 1044

Microbiology Fundamentals

REFERENCES 1. MEND, 1993, “2.11.2a: Literature Review Report: Possible Means of Evaluating the Biological Effects of Sub-Aqueous Disposal of Mine Tailings,” March 1993. 2. N. Kuyucak, Miner. Metall. Proc., No.2, 17 (2000) 85. 3. S. Foucher, F. Battaglia-Brunet, I. Ignatiadis and D. Morin, Chem. Eng. Sci., 56 (2001) 1639. 4. J.H. Tuttle, P.K. Dugan and C.I. Randles, Appl. Microbiol., No.2, 17 (1969) 297. 5. Kathy Jalali and S.A. Baldwin, Wat. Res., Vol. 34, 3 (2000) 797. 6. D.H. Dvorak, R.S. Hedin, H.M. Edenborn and P.E. McIntire, Biotechnol. Bioeng., 40 (1992) 609. 7. H. Kilty, J. Roussy, J.M. Tobin and J.R. Degorce-Dumas, Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, V.S.T. Ciminelli and O. Garcia Jr. (eds.), Elsevier Science B.V, Part B (2001) 357. 8. L.J. Barnes, F.J. Janssen, J. Sherren, J.H. Versteegh, R.O. Koch and P.J.H. Scheeren, Chem. Eng. Res. Des., 69 (1991) 184. 9. L.J. Barnes, F.J. Janssen, J. Sherren, J.H. Versteegh and R.O. Koch, Trans. Inst. Min. Metall., 101 (1992) C183-C199. 10. M.V. Rowley, D.D. Warkentin and V. Sicotte, Proceedings of the Fourth International Conference on Acid Rock Drainage, Vancouver, BC. Canada, 1997. 11. M.P. Silverman and D.G. Lundgren, J. Bacteriol., 77 (1959) 642. 12. E.S. Pankhurst, Isolation of Anaerobes, D.A. Shapton and R.G. Board (eds.), The Society for Appl. Bacteriol., Technical series No.5, Academic Press, 1971. 13. A.I. Vogel, Vogel’s Text Book for Quantitative Chemical Analysis, 5th edn, Longman, London, 1989. 14. APHA, AWWA and APCF, Standard Methods for the Examination of Water and Waste Water, 14th edn, 1975. 15. Shrihari, R. Kumar, K.S. Gandhi and K.A. Natarajan, Appl. Microbiol. Biotechnol., 36 (1991) 278. 16. M.G. Monroy Fernandez, C. Mustin, Philippe de Donato, O. Barres, P. Marion and J. Berthelin, Biotechnol. Bioeng., 46 (1995) 13. 17. K.J. Edwards, M.O. Schrenk, R. Hamers and J.F. Banfield, Am. Mineral., 83 (1998) 1444. 18. J.R. Postgate, The Sulphate Reducing Bacteria, Cambridge University Press, 2nd edn., 1984. 19. C. Garcia, D.A. Moreno, A. Ballester, M.L. Blazquez and F. Gonzalez, Miner. Eng., Vol. 14, 9 (2001) 997.

1045

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Comparative study on pit formation and interfacial chemistry induced by Leptospirillum and Acidothiobacillus ferrooxidans during FeS2 leaching Helmut Tributsch and José Rojas-Chapana Hahn-Meitner-Institut, Dept. Solare Energetik, 14109 Berlin, Germany Abstract After it had been assumed for a long time that the bacterium Leptospirillum ferrooxidans is only capable of oxidizing Fe2+ which then oxidizes the sulfide, highresolution electron microscopy has recently revealed that the bacterium can also selectively disintegrate a crystallized solid sulfide such as FeS2 in order to recover Fe2+. Characteristic properties are the mineralization of its surface including the interacting capsula, where sulfide nanoparticles can be identified, and localized pit formation, which can result in a pronounced collective effort to dissolve the crystalline environment. Acidothiobacillus ferrooxidans, on the other hand, profits from weakening and break up of chemical bonds mediated by the formation of cysteine-pyrite complexes. This might also be the case under natural conditions by the excretion of cysteine and cysteine-like metabolites. This contribution aims at a comparison of both Fe2+ oxidizing bacteria, to reveal their adaptation to the leaching of the FeS2 interface. It is shown how Leptospirillum has adapted electrochemical while Acidothiobacillus has adapted chemical mechanisms to obtain an optimum output of energy supplying Fe2+ and sulfur species respectively. 1.

INTRODUCTION Leaching of pyrite has received considerable attention especially due to its relevance for the hydrometallurgical recovery of gold but also because pyrite, which is a component in many sulfide ores, releases sulfuric acid during the oxidation process [1-12]. The bacteria typically found in the presence of pyrite under leaching conditions are Acidothiobacillus (formerly Thiobacillus) ferrooxidans (A.f.) and Leptospirillum ferrooxidans (L.f.). As their names indicates, both are able to oxidize Fe2+ to Fe3+ in the presence of oxygen, but while A.f. is also able to oxidize sulfide species, L.f. is apparently only able to oxidize Fe2+. Numerous studies have been performed to investigate the oxidation properties of these two bacterial species [13-14]. With respect to the bacteria’s ability for Fe2+-oxidation it has been documented that L.f. is able to utilize Fe2+ as an energy source at significantly, 200 mV, more positive oxidation potentials compared with A.f. [9]. Under such conditions, correspondingly less electrochemical energy is available for bacterial carbon dioxide fixation. In technical gold leaching installations with pyrite containing ores, L.f. are the bacteria which are doing the main job especially when 1047

Microbiology Fundamentals

the solutions contain high concentrations of Fe3+. This contribution aims at a comparative analysis of the two bacterial mechanisms for interacting with pyrite [13]. 2.

EXPERIMENTAL Bacterial strain and growth conditions: Leptospirillum ferrooxidans (isolated from commercial biooxidation tanks in South Africa) was kindly provided by G. Hansford (University of Cape Town, South Africa). L.f. was maintained on ferrous ions culture solution using the medium described by Tuovinen and Kelly for Thiobacillus ferrooxidans [15]. Bacterially mediated ferrous oxidation was followed by increasing coloration and gradual precipitation due to the low solubility of ferric iron at pH 1.6. For the culture of L. ferrooxidans on pyrite crystals (pyrite from Mogul-Turkey; particle size ca. 5 mm and at 10% pulp density) the same medium was used but without ferrous ions. Continuous agitation is required to maintain uniform pulp density. The cultivation of A.f. on synthetic pyrite layers was performed following the procedure described by Rojas-Chapana et al. [16-17]. Microscopy: High-resolution and analytical transmission electron microscopy (TEM) and scanning electron microscopy (SEM) were used to study the bacterial morphology, the stoichiometry and shape of the pyrite grains used as substrates as well as bacteria-pyrite assemblages formed during the bacterial life cycle. All samples for TEM-examinations were imaged in a Philips CM 12 electron microscope operating at 100 kV. High-resolution SEM-examinations were imaged in a Gemini SEM-microscope (LEO1530). The SEM microscope was operated in the 3-7 KeV range being configurated primarily for imaging (using secondary electrons) and compositional analysis (using energy-dispersive x-ray spectroscopy (EDS)). 2.1 Bacterial morphology in pyrite cultures Neither A.f. nor L.f. show peculiarities in their morphology when cultivated on Fe2+ as the only energy source. However, when cultivated on ground-up pyrite crystal material, they both show a significant adaptation, in which they develop a pronounced capsule of polysaccharides around their bacterial cells. Fig. 1a and 1b show A.f. bacteria surrounded by such capsules, which are dotted with nanoparticles. These nanoparticles turned out to be molecular sulfur [16] extracted from the pyrite interface. Fig. 1b shows a A.f. cell which is splitting-up into a mother- and daughter cell. It is seen, that also the capsule is separated during this process. The mechanism by which A.f. is interacting with the pyrite surface, via the organic material of the capsule has been discussed in some details in previous publications [11,16,17]. Since the main energy source for this bacterium is a sulfide species that liberates significantly more energy compared with Fe2+, the evolution of such a capsule is not surprising. The bacterium needs an organic matrix to circle the catalyst species, which are disrupting chemical bonds in the pyrite interface. Crystal-bound sulfur is extracted in the form of polysulfide to yield sulfur colloids as intermediate energy storing elements. The concentration, size and distribution of these colloids within the capsule may change during the life cycle of the bacterium and in dependence of the sulfide energy supply. It was a surprising discovery that L.f. are generating a capsule around their cells when interacting with a pyrite energy source. Fig. 1c shows three bacterial cells characterized by their spiral growth, partially overlapping each other in the transmission electron microscope picture. The two bacterial cells on the right are shown in a slightly higher magnification in Fig. 1d. The areas where the bacterial cells plus capsules overlap are 1048

Microbiology Fundamentals

clearly shown as areas of more intensive darkness. All over the capsule of L.f. dark spots of nanoparticles are seen. They are typically unevenly distributed and vary in a certain range with respect to their dimension. X-ray luminescence (EDAX) analysis has clearly shown that one is dealing with FeS2 particles. The cubic crystal structure of pyrite could be confirmed by electron diffraction analysis [18]. The question now arises: why does L.f., if it only oxidizes Fe2+, need a capsule? and, even more important; why do crystalline FeS2 particles accumulate in it? There is apparently a mechanism by which L.f. bacterial cells are interacting with solid pyrite interfaces to disrupt them and to recover nanoparticles for some metabolic purpose. In order to get a better insight into the interaction of L.f. with pyrite interfaces it became necessary to perform studies of crystallized pyrite interfaces in the presence of bacterial cells.

Figure 1a. Planktonic A.f. bacteria grown on synthetic pyrite film

Figure 1b. A.f. bacteria during cell division on synthetic pyrite film

Figure 1c. Planktonic L.f. bacteria grown on pyrite crystals

Figure 1d. Enlargement of 1c. L.f. bacteria covered with pyrite nanoparticles

2.2 Leaching patterns on crystallized pyrite interfaces Our research group has a long-term experience with the physical chemistry of pyrite interfaces and their electrochemical behavior. It was, therefore, quite easy to identify a very unusual type of leaching patterns which develop in presence of L.f.. In otherwise apparently untouched surfaces, pronounced 20 – 40 µm large, deep etching structures developed, which are shown in Fig. 2. They partially show a hexagonal pattern (which can be explained for pyrite interfaces [18]). The characteristic spiral-shaped cells of L.f. are rarely found on the exposed surface of pyrite. However, interestingly, when looking into the etching craters, they can be localized there in high concentrations. This is demonstrated with Fig. 2c. The bacterial 1049

Microbiology Fundamentals

concentration is so high, that the bacterial cells are touching each other. There appears to be a correlation between the pronounced etching pit formation and the high concentration of L.f. cells in them. This discovery raises new aspects in relation to L.f. activity. These bacteria have apparently developed a technique for actively dissolving pyrite. If they can do that, they presumably have an advantage from it regardless of the established notion that they are unable to oxidize sulfur species.

Figure 2a-c. SEM micrographs of a pyrite surface leached by L.f. (pitting). The bottom picture shows a high concentration of L.f. in the corrosion craters It is reasonable now to confront this observed leaching behavior with the leaching patterns observed with A.f. for pyrite. The pyrite leaching mechanism by these bacteria has been examined before in some detail, especially by in-situ observation of bacterial activity on 100 nm thin FeS2 synthetic layers [11,12,16]. A localized collective activity comparable to that observed to L.f. was never observed. The bacterial cells moved freely around on the surface interacting with it and leaching it away. Very homogeneous interfaces were that way leached away quite homogeneously. When the substrate, however, was inhomogeneous or when surface-active agents like Tween 80 were applied in very low concentrations the bacterial cells developed localized activities, which yielded pronounced bacterial etching pits. Such pits could be observed in transmitted light microscopic studies as shown in Fig. 3a. Such pictures have been published before [10,16,17], but they are shown here again to contrast the surface activity of L.f. The white spots are etching pits, which bacteria created, making the thin pyrite layer more transparent for light. Chains of such etching pits are clearly recognized and evidence a local multiplication of bacterial cells. Highresolution electron microscopical studies of such pits revealed that the nanocrystalline pyrite substrate has been chemically leached as shown in Fig. 3b. Here an etching pit is 1050

Microbiology Fundamentals

seen on the left hand side and the shadow of a bacterial cell on the right. As explained in preceding publications, such etching patterns arise when bacterial cells with their capsule interact with the pyrite interface. Fe3+, H+, or the addition of a complex forming thiol such as cysteine are cycled as catalysts to break-up interfacial chemical bonds of pyrite [19]. A.f. can use Fe3+ for oxidation of pyrite, however, this oxidation does not lead to an essential disruption of chemical bonds because the energy bands from which electrons are extracted have a non-bonding nature being d-states of iron. When, however, a complexing agent like cysteine is added, then the surface can be disrupted. This has been shown in complementary experiments in which pyrite layers were exposed to cysteine-containing solution. It could be shown that pyrite is indeed dissolved by this complex forming thiol.

Figure 3. Optical transmission photographs of 100 nm thin synthetic pyrite layers which are dotted with corrosion pits generated by bacterial activity. a) shows chains of corrosion pits indicating cell multiplication on the pyrite surface; b) shows a transmission electron micrograph of a corrosion pit with a bacterium nearby It can, therefore, be understood why A.f. is able to dissolve pyrite. This bacterium is specialized for the recovery of the sulfide species and is, therefore, able to disintegrate the crystallized solid material. How can, on the other hand, a bacterium such as L.f., which is only able to oxidize Fe2+, attack a pyrite interface where electron extraction does not primarily break chemical bonds due to the non-bonding character of the energy states involved (within 1 eV from the valence band edge)? 2.3 Comparative evaluation of leaching mechanisms Since electrons extracted from the pyrite valence band by Fe3+ do not generate broken bonds, a straightforward dissolution of pyrite by Fe3+ can, therefore, not be expected [10,19]. Only when the pyrite electrode is electrochemically polarized beyond the standard redox potential of Fe2+/3+ Eh = Eo −

RT ⎛ Fe 2 + ⎞ ⎟ ln⎜ zF ⎜⎝ Fe3+ ⎟⎠

(1)

an electrochemical dissolution of pyrite will occur. It is described by equation (2), which shows that water molecules are needed for this dissolution: FeS 2 + 8H 2O → 14e − + Fe 2+ + 2 SO42− + 16 H +

(2)

This has been verified by experiments in which small concentrations of water have been added to organic electrolytes. It is known from literature that L.f. can oxidize Fe2+ as an energy source at very high concentrations of Fe3+ corresponding to potentials, which are approx. 200 mV more positive compared to the iron oxidation abilities of A.f. [9-13]. L.f., which is able to oxidize Fe2+, however not sulfur containing species, may therefore be

1051

Microbiology Fundamentals

able to oxidize pyrite by locally inducing a very positive Fe2+/3+ redox potential as shown in Fig. 4a. B

A

FeS2

Fe2+ + 2x +2SO42-

biofilm

FeS2 Fe2+2+ Fe

S2O 32-

Fe

3+

Fe3+

Fe

2+

Acidothiobacillus ferrooxidans

Fe2+ 2+

Fe

anodic

Fe

2+

-

e e cathodic -

nanosized-FeS2

e

Fe2+ 3+

-

Fe

+

H+ H Fe2+

3+

Fe

e

-

FeS2 + 2x +4O2 (X= H+ +, Fe3+; Y (eg. cysteine))

dispersed sulfur nanoparticles

ee-

FeS2-mineralized Leptospirillum ferrooxidans

Figure 4. Schemes explaining the mechanisms for pyrite dissolution applied by L.f. (a) and of A.f. (b) The resulting local anodic dissolution of pyrite will liberate pyrite nanoparticles, which are indeed found in the capsule of this bacterial species (Fig.1). The thus generated 14 electrons per pyrite molecule may be driven to a cathodic site on the pyrite surface, where they may be engaged in the reduction of Fe3+ species according to equation (3).

14e − + 14 Fe3+ → 14 Fe 2+

(3)

When equation (2) and (3) are summed up, equation (4) results which corresponds to the equation proposed by Boone et al. [1-2] for ferric leaching of pyrite ores: FeS 2 + 8H 2O + 14 Fe3+ → 15Fe 2+ + 2 SO42− + 16 H +

(4)

While the same global mechanism results, the strategy, which had to be adopted by L.f. bacterial cells to accomplish this reaction turned out to be quite different from the mechanism originally assumed for ferric leaching. We can use standard thermodynamic considerations as applied to semiconductor electrochemistry [20] to understand the process. When the anodic dissolution process (2), which is proposed to be driven by bacterial activity is coupled with the reaction proceeding at the standard hydrogen electrode, 14 H + + 14e − ⇔ 7 H 2

(5)

the following global reaction results which for simplicity is assumed to be reversible. FeS 2 + 8 H 2O ⇔ FeSO4 + H 2 SO4 + 7 H 2

(6)

Knowing the ∆G° value (125,6 kcal/M) of this reaction, one can deduce the thermodynamic decomposition potential (against NHE). It is: ∆G 0 = +0.39V (7) + zF The decomposition may proceed if the redox potential of positive charges or holes is more positive than the decomposition potential (7), provided there are no kinetic barriers. Eo =

1052

Microbiology Fundamentals

Such kinetic barriers exist for pyrite. They arise because of the electronic d-character of its valence band. The position of the band edges varies if the electric charge on the surface varies, either by adsorbed ions, charged surface complexes or by high concentrations of holes in the surface. The increase of free energy of holes by electron extraction via L.f., which determine the oxidative dissolution of pyrite is described by ⎛ p + ∆p L ⎞ L ⎜⎜ ⎟⎟ = − ln E E kT p F V N V ⎝ ⎠

(8)

(EV = position of valence band edge, NV = concentration of states in the valence band, ∆PL = concentration of holes generated by electron extraction through L.f. beyond the existing concentration p). In this relation Fe3+ mediated electron extraction leading to an increase in ∆PL will pull the potential of holes towards positive values. Thus process will be accompanied by a positive shift of the valence band position EV due to positive charging of the pyrite surface. L.f. has obviously adapted to drive FeS2 beyond the decomposition potential and its kinetic barrier. A strategy was necessary which resembles the phenomenon of electrochemical corrosion in which anodic and cathodic sites develop on the same micro heterogeneous interface. Such a scenario is shown for L.f. activity and pyrite in Fig. 4a. Here high concentrations of bacterial cells are locally inducing anodic electrochemical dissolution processes while the thus liberated electrons are moving to cathodic sites where Fe3+ is reduced. The Fe2+ then diffuses to the bacteria where the oxidation of this energy source will occur. Experimental support for such a process are the pronounced large corrosion craters produced by L.f. and the high concentration of bacterial cells sitting in these craters. This special surface morphology created by bacterial activity may allow an intensive circulation of Fe2+ for the bacterial energy supply needed for CO2-fixation. While L.f. can sustain its metabolism by utilizing Fe2+ at very positive redox potentials, approx. 200 mV more positive compared to A.f., the latter species operates in a potential region where pyrite is typically not electrochemically dissolved. Since Fe3+ in small concentrations have only a very limited effect on disruption of chemical bonds in the pyrite interface, A.f. had to adopt a different strategy for energy recovery. This strategy is shown in a simplified way in Fig. 4b (for more detail see [19]. Basically this bacterium utilized protons, Fe3+, and complexing thiols (i.e extracellular polymeric substances (EPS) containing cysteine or cysteine-containing metabolites) for a catalytic disruption of pyrite interfaces. This works in such a way that protons interacting with the interface may partially break bonds forming interfacial SH- groups. Fe3+ may extract electrons from thus created interfacial states and cysteine may form complexes with the surface thus disrupting it and extracting iron sulfur aggregates bonded to cysteine. Since protons, Fe3+, species and cysteine are recycled this process may be considered catalytic. Viewed globally, the process developed by L.f. is also catalytic (since Fe3+ is recycled, Fig. 4a), however, the individual steps are electrochemical and require a separation of anodic and cathodic sites as known from electrochemical corrosion mechanisms. The effect of exogenous cysteine on the bacterial dissolution of pyrite has been demonstrated [12], but so far no proof has been given that cysteine is actually present in EPS secreted by A.f. in sufficient concentration. Though the EPS composition of some determined A.f. strains have been published, it cannot be extrapolated to bacteria directly interacting with the sulfide substrate. In-situ determinations of cysteine and cysteinecontaining metabolites in the presence of a biofilm on solid sulfides are required. In addition, cysteine and methionine have recently been found In EPS produced by sulfatereducing bacteria during corrosion of mild steel [21]. 1053

Microbiology Fundamentals

These distinct leaching strategies adopted by both pyrite oxidizing bacteria species show the evolutionary adaptability of bacteria to their inorganic energy source. Leptospirillum populations have apparently "discovered" that pyrite disintegrates under very positive Fe2+/3+ potentials and have improved their ability to recover reducing power for Fe3+ reduction by organizing themselves for cooperative leaching activities within corrosion craters which apparently improve their ability to recover Fe2+ species. Summarizing, the following conclusions can be drawn: While A.f. takes advantage of thiol complex chemistry to brake up sulfur species, their main energy source, from the sulfide interface, which are subsequently oxidized to sulfate for energy recovery, the survival interest of L.f. is aimed at recovering a maximum amount of reduced iron from the FeS2. Its interfacial activity aims at pushing the FeS2 towards electrochemical oxidation (iron sulfate formation) so that the reducing power of electrons can be harnessed for Fe3+ reduction elsewhere on the sulfide surface. L.f. bacteria apparently induce electrochemical corrosion by locally polarizing the FeS2 interface with Fe3+ immobilized in their capsule, thus oxidizing the sulfur species while generating microscopic cathodic currents for Fe3+ reduction. The specific leaching behavior of both species is not adapted for an efficient joint leaching activity. L.f. aims at a more positive redox potential and at sacrificing the energy rich sulfur species while utilizing electrochemically mobilized electrons for Fe3+ reduction to recover a maximum amount of Fe2+. Since very little energy is generated in the oxidation of ferrous to ferric ions, the bacteria must oxidize large amounts of iron in order to grow. A.f., on the other hand, is adapted to extract the sulfur species for its sulfur metabolism, which provides a high energy turnover and Fe2+ oxidation is just a parallel process which yields significantly less energy but supports the interfacial leaching process by providing Fe3+. ACKNOWLEDGEMENTS The authors would like to thank Ms U. Bloeck and Mr. M. Wilhelm for their valuable assistance in electron microscopy. REFERENCES

1. M. Boon and J. J. Heijnen, Biohydrometallurgical Technologies, A. E. Torma, J. E. Wey and V. L. Lakshmanan (eds.), The Minerals, Metals & Materials Society, Warrendale Pennsylvania, (1993) 217. 2. M. Boon and J. J. Heijnen, Hydrometallurgy 48 (1998) 27. 3. W. Sand, K. K. Rohde, B. Sobotke and C. Zenneck, Appl. Environ. Microbiol., 58 (1992) 85. 4. T. Rohwerder, A. Schippers and W. Sand, Proceedings of the XX International Mineral Processing Congress, Vol. 4, Solid-Liquid Separation, Hydro- und Biohydrometallurgy, H. Hoberg and H. von Blottnitz (eds.), Gesellschaft für Bergbau, Metallurgy, Rohstoff- und Umwelttechnik (GMDB), Clausthal Zellerfeld, (1997) 475. 5. T. Rohwerder, A. Schippers, W. Sand, Thermochimica Acta, 309 (1998) 79. 6. P. R. Norris and D. P. Kelly, The use of mixed microbial cultures in metal recovery. Microbial Interactions and Communities, A. T. Bull and J. H. Slater (eds.), Academic Press London, (1982) 443. 7. P. R. Norri, Recent Progress in Biohydrometallurgy, G. Rossi and A. E. Torma (eds.), Associazione Mieraria Sarda, Rom, Italy, (1983) 83.

1054

Microbiology Fundamentals

8. W. Sand, T. Gehrke, P.-G. Jozsa, A. Schippers, Proceedings of IBS 99. Biohydrometallurgy and the Environment Towards Mining of the 21st Century, Elsevier, (1999) 27. 9. G. S. Hansford and T. Vargas, Proceedings of IBS 99. Biohydrometallurgy and the Environment Towards Mining of the 21st Century, Elsevier, (1999) 13. 10. H. Tributsch, Proceedings of IBS 99. Biohydrometallurgy and the Environment Towards Mining of the 21st Century, Elsevier, (1999) 51. 11. J. Rojas-Chapana and H. Tributsch, Hydrometallurgy 59 (2001) 291-300. 12. J. A. Rojas-Chapana and H. Tributsch, Process Biochemistry 35 (8) (2000) 815. 13. D. Rawlings, H. Tributsch, G. Hansford, Microbiology 145 (1999) 5-13. 14. F. Battaglia-Brunet, P. d’Hughes, T. Cabral, P. Cezac, J. L. Garcia, D. Morin, Minerals Engineering 11 (1998) 195. 15. O. Tuovinen, D. Kelly, Arch Mikrobiol 88(1973) 285. 16. J. Rojas-Chapana, M. Giersig, H. Tributsch, Arch. Microbiol 163 (1995) 352. 17. H. Tributsch, J. Rojas-Chapana, C. Baertels, A. Ennaoui, W. Hofmann, Corrosion 54 (1998) 216. 18. J. A. Rojas-Chapana and H. Tributsch, submitted to FEMS Microbiol. 19. H. Tributsch and J. A. Rojas-Chapana, Electrochimica Acta 45 (2000) 4705. 20. H. Gerischer, J. Vac. Sci. Technol.15 (1978) 1422. 21. K. Chan, L.. Xu, H. Fang. Environ. Sci. Technol. 36 (2002) 1720

1055

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Composition of biofilm communities in acidic mine waters as revealed by combined cultivation and biomolecular approaches Sakurako Kimura, Kris Coupland, Kevin B. Hallberg and D. Barrie Johnson School of Biological Sciences, University of Wales, Bangor, LL57 2UW, U.K. Abstract Waters draining a pyritic outcrop and a disused underground copper mine in north Wales are both acidic and metal-rich. Both also contain copious amounts of microbial biomass growing as biofilms ("acid streamers"). We have studied the microbial communities in these streamers using a combination of cultivation-based and cultivationindependent molecular techniques. Isolation of acidophiles using a range of solid media indicated that the predominant microbes at one site were iron-oxidizing microbes whereas iron oxidizers and heterotrophic acidophiles were present in similar numbers in streamers from the other site. Streamer communities were also analyzed by fluorescent in situ hybridization (FISH) using a range of oligonucleotide probes of varying specificity. FISH revealed that the streamers from both mine sites were similar in overall gross structure, and were dominated (80-90% of the total bacterial counts) by β-Proteobacteria. The next most abundant group of microbes was γ-Proteobacteria, with Acidithiobacillus ferrooxidans accounting for about 25-90% of these. Archaea, low G+C Gram-positive bacteria and At. thiooxidans were not detected in streamers either by FISH and cultivation techniques, though "Ferrimicrobium"-like bacteria were detected using both approaches. These data show that the microbial communities of streamer biofilms from two different mine sites are remarkably similar, but differ substantially from those found in an abandoned pyrite mine at Iron Mountain, California.

Keywords: acid mine drainage; acid streamers; biodiversity; FISH 1.

INTRODUCTION The diversity of prokaryotic microorganisms that may live in highly acidic, metal-rich environments, such as acid mine drainage (AMD) and leachate liquors in biomining operations is known to be extensive [1]. The most obvious manifestations of microbial life in these environments are filamentous gelatinous growths, generally referred to as "acid streamers". The earliest report of acid streamers was by Lackey [2] who observed them in AMD streams in West Virginia. He originally believed that the long, colorless or light brown structures were fungal, but microscopic examination showed that they were "bacterial masses, presumably in a zoogloeal jelly". In contrast, Leathen [3] described acid streamers as consisting of fibrous masses of sheath-like structures, and that bacterial cells were absent. Temple and Koehler’s [4] description of acid streamers consisting of nonfilamentous, non-orientated bacteria embedded within a tough slime, was more similar to that of Lackey. They also noted that, inside mines, acid streamers were occasionally 1057

Microbiology Fundamentals

encrusted with orange-brown ferric precipitates and, in AMD streams outside mines, colonized by phototrophic Euglena spp., giving them a green appearance. Dugan et al. [5] isolated a neutrophilic Bacillus sp. from acid streamers, which produced copious amounts of extracellular slime in liquid media; they concluded that this bacterium was a significant member of the acid streamer microbial community. Johnson et al. [6] also isolated a variety of neutrophilic bacteria (several of which produced exopolysaccharides) from acid streamers growing extensively (estimated biovolume >100 m3) in an abandoned pyrite mine in North Wales. In later work [7, 8], acidophilic chemolithotrophic and heterotrophic bacteria were also isolated from the same site. These authors concluded that these acid streamers were a mixed community of ironand sulfur-oxidizing bacteria, which were the primary producers in this pyrite mine, and that the heterotrophic bacteria lived off exudates and lysates of the primary producers. Interestingly, one of the isolates (coded "CCH7") was a ferrous iron-oxidizing filamentous isolate that formed streamer-like growths in vitro. In contrast, Wakao et al. [9] considered that it was highly unlikely that heterotrophic bacteria could form acid streamers in AMD and concluded that such growths were composed of Acidithiobacillus ferrooxidans embedded in a gelatinous matrix. New insights into the microbial composition of acid streamer-like communities came from the work of Bond et al. [10] who studied a 1 cm-thick slime biofilm that developed on the surface of finely disseminated pyrite ore within an AMD stream at Iron Mountain, California. Phylogenetic analysis of 16S rRNA genes cloned from this microbial community indicated that the dominant sequences were of bacteria related to the ironoxidizer Leptospirillum. Archaeal (Thermoplasmales lineage), Acidimicrobium/ "Ferrimicrobium" and δ-proteobacterial gene sequences were also detected. On the basis of these (and other) results, a variety of gene probes were designed and the slime community further analyzed using fluorescent in situ hybridization (FISH). Whilst these findings contrasted greatly with earlier results that were based chiefly on cultivation and physiological criteria, the environment within Iron Mountain was also very different (temperature 30-50ºC; pH typically 0.5-1.0) from the cooler, less acidic (pH 2-3) AMD waters in which acid streamers have been reported elsewhere. In this paper, we describe the microbial diversity in acid streamers from AMD at two mine sites in north Wales, as revealed by using a combination of cultivation and molecular approaches. 2.

MATERIALS AND METHODS

2.1 Mine sites and water analysis Acid streamers were sampled at two mine sites in north Wales: the former Parys copper mine and Trefriw Spa, which is situated below the abandoned Cae Coch pyrite mine [11]. Small streamer biofilm samples (ca. 1 cm3) were removed and taken to the laboratory with one hour. On site analysis of pH, redox potential, temperature and conductivity used a Whatman Water Tester (Whatman, U.K.); dissolved oxygen (DO) was measured using a D400 DO meter (Whatman, U.K.). Ferrous iron was measured in filtered (0.2 µm pore-size) samples using the ferrozine assay [12]; other metals were analyzed by atomic absorption spectrometry, and sulfate was measured turbidometrically as barium sulfate (Hydrocheck, Cambridge, U.K.). Dissolved organic carbon was measured using a Protoc Analyzer (Pollution & Process Monitoring Ltd., U.K.).

1058

Microbiology Fundamentals

2.2 Isolation of bacteria Streamer samples were put into bottles containing basal salts solution [13], adjusted to pH 3.5, and vortexed for 5 min. The suspension was then centrifuged at low speed to remove undispersed streamer material, the supernatant was serially diluted, and samples plated onto a variety of solid media. The majority of these were overlay media, which facilitate the growth of a wide range of autotrophic and heterotrophic acidophiles. Full details of the media can be found elsewhere [13, 14]. R2 agar [15] was used to enumerate neutrophilic bacteria. Inoculated plates were incubated for up to 6 weeks at 20ºC under aerobic and anaerobic conditions (the latter using AnaeroGen system; Oxoid, UK). 2.3 Phylogenetic analysis of isolates The 16S rRNA genes of streamer isolates were amplified, using protocols described elsewhere [16]. Purified PCR products were sequenced using a BeckmanCoulter dye Terminator Cycle Sequencing kit and a CEQ8000 Genetic Analysis System (Beckman Coulter, U.K.). The resulting gene sequences were compared with those available in GenBank using BLAST [17]. 2.4 Total bacterial counts Bacteria in disrupted streamer suspensions (section 2.2) were harvested, washed with phosphate buffered saline (PBS) and the re-suspended cells fixed in 3% (v/v) paraformaldehyde in PBS. Fixed cells were washed with PBS, suspensions filtered through 25 mm black polycarbonate membranes (0.2 µm pore size) and stained with 4’, 6diamidino-2-phenylindole (DAPI; 10 ml of a 1 µg/ml solution). Membranes were placed onto glass slides and viewed using a Nikon ECLIPSE E600 microscope. A minimum number of 100 bacteria/membrane were counted. 2.5 Fluorescent in situ hybridization (FISH) The fixed streamer cells (section 2.4) were also utilized for FISH analysis, using a modified method of that described by Bond and Banfield [18]. The RNA probes used are listed in Table 1. Where the probe used targeted Gram-positive bacteria, fixed cells were first incubated (4°C; 10 min) with lysozyme (1027.5 units of activity/ml) prior to dehydration with ethanol. Hybridization was carried out using 25 ng of fluorescein-labeled eubacterial probe (EUB338Fl) and 25 ng of a more specific, Cy3-labeled probe at the same time. Mounting medium (70% glycerol in 100 mM sodium tetraborate, pH 9.2, containing 3 mg/ml N-propyll gallate) was applied to reduce fading of the probe signal. Various concentrations of formamide (0-50%, v/v) were tested to optimize specificity and maximize signal response for each probe, using pure cultures of target and related microorganisms (Table 1). To enumerate different groups of bacteria, Cy3-labeled cells were counted relative to those stained with the EUB338Fl probe. For the eubacterial and archaeal probes, counts of stained cells were made relative to those stained by DAPI.

1059

Table 1. List of 16S rRNA (except where noted) oligonucleotide probes used for FISH analysis and formamide concentrations required for optimum specificity. All probes, except for EUB338Fl, were labeled with the fluorochrome Cy3. Probe EUB338Fl ALF1B

Target Eubacteria α-Proteobacteria, some δProteobacteria and most spirochetes 23S rRNA of most β-Proteobacteria 23S rRNA of most γ-Proteobacteria Thiomonas Group 1 Thiomonas Group 2 Acidithiobacillus ferrooxidans Acidithiobacillus thiooxidans Acidimicrobium and “Ferrimicrobium” Acidimicrobium ferrooxidans Low G+C Gram-positive bacteria Leptospirillum groups 1, 2 and 3 Archaea

Probe sequence (5’f3’) GCTGCCTCCCGTAGGAGT CGTTCGYTCTGAGCCAG1

Formamide (%) 0 – 50 20

Reference [19] [20]

BET42a2 GCCTTCCCACTTCGTTT 35 [20] 2 GAM41a GCCTTCCCACATCGTTT 35 [20] TM1G138 GCAGTTATCCCCCATCAAT 40 This study TM2G138 GTAGTTATCCCCCATCACA 40 This study TF539 CAGACCTAACGTACCGCC 20 [21] ATT223 AGACGTAGGCTCCTCTTC 40 This study ACM732 GTACCGGCCCAGATCGCTG 35 [18] ACM995 CTCTGCGGCTTTTCCCTCCATG 10 P. Norris, unpublished LGC355 GGAAGATTCCCTACTGCTG 20 This study LF655 CGCTTCCCTCTCCCAGCCT 35 [18] ARCH915 GTGCTCCCCCGCCAATTCCT 40 [22] 1. Y = C or T 2. Hybridization using these two probes was carried out in the presence of a 10-fold excess of the other unlabeled oligonucleotide

Microbiology Fundamentals

3.

RESULTS

3.1 Mine water analysis Physico-chemical data from the two mine sites from where streamer biofilms were sampled are shown in Table 2. At both sites, conditions were reducing, with both dissolved oxygen and redox potentials being relatively low, and all of the soluble iron present was in the ferrous form. Table 2. Physicochemical data from the two AMD sites from which streamer biofilms were obtained. N.D. = not determined pH T (ºC) Eh (mV) Ec (µS/cm) DO2 (mg/L) DOC (mg/L) Fe2+ (mg/L) SO42- (mg/L) Al (mg/L) Cu (mg/L) Zn (mg/L)

Parys mine 2.70 8.8 +420 2994 0.55 6.0 473 1552 115 50 60

Trefriw spa 3.05 13.3 +374 1635 1.66 14.9 217 1093 N.D. N.D. N.D.

3.2 Plate and DAPI Counts Colonies that grew on overlay media were differentiated on the basis of whether or not they became encrusted with ferric iron (i.e. iron-oxidizing autotrophs and heterotrophs), and on other morphological characteristics (color, size, etc.). Iron-oxidizing isolates were further differentiated into moderate and extreme acidophiles, depending on whether they grew on ferrous iron/thiosulfate (pH ~4) or ferrous iron (pH ~2.5) overlay plates [14]. Enumeration of bacteria in dispersed biofilms determined by plate counts (on aerobically-incubated plates) is shown in Table 3. Plating efficiencies, determined relative to total (DAPI) cell counts were 1.8% for the Parys streamers and 0.96% of the Trefriw streamers. Numbers of isolates obtained on solid media incubated under anaerobic conditions were considerably, 102- to 103-fold, fewer than obtained with plates incubated aerobically. 3.3 Phylogenetic analysis of streamer isolates Identification of representative streamer isolates, differentiated on the basis of colony morphologies and identified from analysis of amplified 16S rRNA genes, is shown in Table 4. All of the iron-oxidizing autotrophic isolates examined in streamers at both sites were At. ferrooxidans rather than Leptospirillum spp. In addition, "Ferrimicrobium"-like isolates (Actinobacteria) were isolated from both streamers, as were Acidiphilium and Acidocella spp. (α-Proteobacteria). Parys streamer also included an isolate remotely related to Acidobacterium capsulatum, a Frateuria-like isolate (γ-Proteobacteria), and a Thiomonas-like isolate (β-Proteobacteria). Interestingly, the latter was initially classified as a heterotroph (being isolated on yeast extract solid medium) though it was later confirmed that it was capable of oxidizing thiosulfate [23]. 1061

Microbiology Fundamentals

Table 3. Plate counts (aerobic) and total microbes in the dispersed streamer samples from the two sites (CFU per g wet weight dispersed streamer) Parys streamers

Trefriw streamers

9.36 x 106 2.42 x 103 4.40 x 105 <102 9.80 x 106 5.44 x108

1.16 x 107 <102 4.80 x 106 <102 1.64 x 107 1.7 x109

Iron-oxidizing bacteria - extreme acidophiles - moderate acidophiles Heterotrophic acidophiles Neutrophiles Total plate counts Total counts (DAPI)

Table 4. Identification of acid streamer isolates based on analyses of amplified 16S rRNA genes

(i)

Parys streamers

Isolate

Physiological class

Nearest relative (% identity)

KP1

Iron-oxidizing heterotroph

Uncultured eubacterium TRA2-10 (99.9) “Ferrimicrobium acidophilum” (93.0)

PK11

Iron-oxidizing autotroph

At. ferrooxidans NO-37 (99.9)

PK51

Heterotroph

γ-proteobacterium WJ2 (100)

PK40

Heterotroph

Acidiphilium NO-17 (98.0)

PK46

Heterotroph

Acidiphilium organovorum (98.5)

PK48

Heterotroph

Acidiphilium acidophilum (97.5)

M21

Heterotroph

Acidocella NO-12 (99.4)

PK35

Heterotroph

Acidobacterium WJ7 (99.5)

KP3

Heterotroph

Uncultured eubacterium TRB82 (96.7) Acidobacterium capsulatum (91.8)

PK44

Heterotroph

Thiomonas CO2 (97.4)

ii)

Trefriw streamers

Isolate

Physiological class

Nearest relative (% identity)

CS11

Iron-oxidizing heterotroph

Uncultured eubacterium TRA2-10 (99.7) “Ferrimicrobium acidophilum” (93.0)

CCW10

Iron-oxidizing autotroph

At. ferrooxidans NO-37 (99.5)

CCW68

Iron-oxidizing autotroph

At. ferrooxidans ATCC 19859 (99.0)

CCW30

Heterotroph

Acidocella NO-12 (99.1)

CCP3

Heterotroph

Acidiphilium NO-17 (99.5)

3.4 FISH analysis of dispersed streamers Analysis of microorganisms in the disrupted streamers by FISH produced data that both confirmed and contrasted with results from cultivation techniques. The prokaryotes in the streamers were shown to be exclusively (>99.9%) bacteria; no cells were detected using the archaeal probe. Also in line with plating data, probes targeting Leptospirillum spp., At. thiooxidans and low G+C Gram-positive bacteria also failed to detect any cells. 1062

Microbiology Fundamentals

The Gram-positive bacteria that were detected by FISH were exclusively Actinobacteria detected by the Acidimicrobium/”Ferrimicrobium” probe but not by the Am. ferrooxidans-specific probe. The vast majority of cells in both streamer biofilms were Proteobacteria. Whilst α-, β- and γ-Proteobacteria were all detected by FISH, the dominant group in both Parys and Trefriw streamers were β-Proteobacteria (Figure 1). The only β-proteobacterium isolated was a Thiomonas species from the Parys streamer. By FISH analysis with the corresponding probe (TM1G138), however, this microbe was detected at less than 1 cell per thousand in the Parys streamer and was not detected at all in the Trefriw streamer. A second probe (TM2G138), targeting another group of Thiomonas species [23], also failed to detect cells in either streamer.

Figure 1. Counts of microbes in the streamer samples targeted by a range of Cy3labeled oligonucleotide probes, given as percentages of the fluorescein-labeled Eubacterial probe (EUB338Fl) 4.

DISCUSSION The application of molecular techniques is increasingly becoming important in microbial ecology [24, 25]. The power of these techniques lies in the ability to identify and enumerate microorganisms without the reliance on cultivation. Cultivation-based studies often lead to incorrect conclusions, such as the importance of Acidithiobacillus ferrooxidans in commercial bioleaching operations, which have been shown by a molecular approach to be dominated by Leptospirillum ferrooxidans and relatives [26]. In this study, we have taken the rarely used approach of combining cultivation with molecular studies to characterize the microbial populations of streamer biofilms that dominate mine water draining two metal mines in north Wales. Earlier cultivation-based studies were carried out on both sites, and the identities of the microbes present were determined solely on physiological characteristics. Since that time, improved solid media for the cultivation of acidophiles have been developed, and were employed in the present study along with 16S rRNA-based identification of the isolates. As would be expected from earlier studies, the dominant cultivatable isolates from both sites were iron-oxidizing microbes and the majority of these were found, by 16S rRNA gene sequence analysis, to be Acidithiobacillus ferrooxidans. 1063

Microbiology Fundamentals

The heterotrophs isolated from both streamers proved to be more diverse than previously thought, especially in the Parys streamer. The heterotrophs from both streamers included isolates from the well-known genera of acidophiles Acidiphilium and Acidocella, and amongst these were isolates that may represent new species of these genera. Some previously unrecognized microbes also inhabit the streamer from Parys Mountain. One isolate, PK51, is related to neutrophilic microbes of the genus Frateuria, PK51, and two other isolates are related to Acidobacterium capsulatum. The homology of these isolates to Ab. capsulatum and to each other is sufficiently low that they can be considered to belong to different genera. It is interesting to note that few microbes of the phylogenetic group that includes Ab. capsulatum have been cultivated, even though molecular studies show that they are the dominant microbes in many habitats [e.g. 27]. The latter two groups of microbes have also been isolated from a wetland constructed to treat acid mine drainage [14]. As with the cultivation-based study, the populations of both of the streamers appeared to be similar when analyzed with probes targeting ribosomal RNA. There was no indication of archaea present in either streamer community, as indicated by FISH and also by the absence of an amplification product following PCR with archaeal 16S rRNA genespecific primers (data not shown). In contrast to the cultivation study, both of the streamers were dominated by β-Proteobacteria and not γ-Proteobacteria, the probe for which detects At. ferrooxidans. The only isolate obtained from either of the streamers, PK44, that belongs to the β-Proteobacteria is related to the genus Thiomonas. Using probes to target the 16S rRNA gene of bacteria in this genus, however, it was shown that they are not the dominant microbes in either streamer, in keeping with the cultivationbased studies. Further molecular characterization of these streamers, including the generation of 16S rRNA gene libraries, is currently in progress and will reveal the identity of the dominant microbes in the streamer biofilms. Identification of the unknown microbe, or microbes, that dominate these streamers will be important to understanding the biogeochemistry of these microbial growths that are very prominent in AMD. Even though the streamers studied here were quite similar to each other, they differed substantially from similar microbial growths found in AMD at Iron Mountain, California. Analysis of 16S rRNA genes cloned from Iron Mountain streamer material showed that those streamers were dominated by Leptospirillum-like microbes [28], which have only been found there to date, and by archaea. Other microbes found were related to “Ferrimicrobium acidophilium” and Acidimicrobium ferrooxidans. The Parys and Trefriw isolates related to “Ferrimicrobium” were highly related (>99% 16S rRNA gene homology) to the Iron Mountain clone TRA2-10. FISH analysis [10] confirmed that the major microbes in the Iron Mountain streamers were the Leptospirillum-like microbes and archaea, the latter namely Ferroplasma. Neither the Acidimicrobium/”Ferrimicrobium” type bacteria, nor At. ferrooxidans, were detected by FISH, in contrast to the Parys and Trefriw streamers. While it is difficult to say what determines the differences in the microbial constitution of these various streamers, certainly the physico-chemical parameters play a central role, with temperature (~40°C at Iron Mountain vs ~10°C) and pH (<2 at Iron Mountain vs ~3) having significant effects. ACKNOWLEDGEMENTS Kris Coupland is grateful to the NERC (U.K.; ref. NER/S/C/2001/06450) and Rio Tinto Technologies for provision of a research studentship. K.B.H. and D.B.J. thank the BBSRC and DTI (U.K.; ref. 5/BRM18412) for partial support of this work. 1064

Microbiology Fundamentals

REFERENCES

1. K. B. Hallberg and D. B. Johnson, Adv. Appl. Microbiol. 49 (2001) 37. 2. J. B. Lackey, Public Health Rep. 53 (1938) 1499. 3. W. W. Leathen (1952) Special Summary Report of Microbiological Studies of Bituminous Coal MIne Drainage, Mellon Institute of Industrial Research, University of Pittsburg, Pittsburg, Pa. 4. K. L. Temple and W. A. Kohler (1954) In Research Bulletin No. 25, Engineering Exp. Station, University of West Virginia, Morgantown. 5. P. R. Dugan, C. B. Macmillan and R. M. Pfister, J. Bacteriol. 101 (1970) 982. 6. D. B. Johnson, W. I. Kelso and D. A. Jenkins, Environ. Pollut. 18 (1979) 107. 7. D. B. Johnson, M. A. Ghauri and M. F. Said, Appl. Environ. Microbiol. 58 (1992) 1423. 8. S. McGinness and D. B. Johnson, Sci. Total Environ. 132 (1993) 27. 9. N. Wakao, H. Tachibana, Y. Tanaka, Y. Sakurai and H. Shiota, J. Gen. Appl. Microbiol. 31 (1985) 17. 10. P. L. Bond, G. K. Druschel and J. F. Banfield, Appl. Environ. Microbiol. 66 (2000) 4962. 11. D. A. Jenkins and D. B. Johnson, J. Russell Soc. 5 (1993) 40. 12. D. R. Lovley and E. J. P. Phillips, Appl. Environ. Microbiol. 53 (1987) 1536. 13. D. B. Johnson, J. Microbiol. Meth. 23 (1995) 205. 14. K. B. Hallberg and D. B. Johnson, Hydrometallurgy (2003) in press. 15. D. J. Reasoner and E. E. Geldreich, Appl. Environ. Microbiol. 49 (1985) 1. 16. N. Okibe, M. Gericke, K. B. Hallberg and D. B. Johnson, Appl. Environ. Microbiol. 69 (2003) 1936. 17. S. F. Altschul, T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller and D. J. Lipman, Nucleic Acids Res. 25 (1997) 3389. 18. P. L. Bond and J. F. Banfield, Microbial Ecol. 41 (2001) 149. 19. R. I. Amann, L. Krumholz and D. A. Stahl, J. Bacteriol. 172 (1990) 762. 20. W. Manz, R. Amann, W. Ludwig, M. Wagner and K. H. Schleifer, Syst. Appl. Microbiol. 15 (1992) 593. 21. M. O. Schrenk, K. J. Edwards, R. M. Goodman, R. J. Hamers and J. F. Banfield, Science 279 (1998) 1519. 22. D. A. Stahl and R. Amann, Nucleic Acid Techniques in Bacterial Systematics, E. Stackebrandt and M. Goodfellow (eds.), Wiley, New York, (1991) 205. 23. K. Coupland, F. Battaglia-Brunet, K. B. Hallberg, M. C. Dictor, F. Garrido and D. B. Johnson, This Volume 24. I. Amann, W. Ludwig and K. H. Schleifer, Microbiol. Rev. 59 (1995) 143. 25. I. M. Head, J. R. Saunders and R. W. Pickup, Microbial Ecol. 35 (1998) 1. 26. N. J. Coram and D. E. Rawlings, Appl. Environ. Microbiol. 68 (2002) 838. 27. P. Hugenholtz, B. M. Goebel and N. R. Pace, J. Bacteriol. 180 (1998) 4765. 28. P. L. Bond, S. P. Smriga and J. F. Banfield, Appl. Environ. Microbiol. 66 (2000) 3842.

1065

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Computational fluid dynamics simulation of immobilized Acidothiobacillus ferrooxidans B. Metodieva, L. Lilovab and D. Karamanevb a

b

Blue EdgeValley, 103 Gerald Cr., London, Ontario, Canada N5Z 5A4 Department of Chemical and Biochemical Engineering, The University of Western Ontario, London, Ontario, Canada N6A 5B9

Abstract This paper presents the results of the simulation, using computational fluid dynamics, of the flow and ferric iron concentrations profiles around a single cell of A. ferrooxidans, immobilized on the surface of a sulfide crystal. It has been shown that there are significant concentration gradients of ferric iron concentration between the surface of the crystal and the bulk of liquid. The gradient can reach several hundred mg/L. These results can explain the difference between the chemical (sterile) and biological oxidation rates of metal sulfides in liquid media containing iron ions.

Keywords: computational fluid dynamics, Acidithiobacillus ferrooxidans, bioleaching 1.

INTRODUCTION The biological oxidation of metal sulfides by Acidothiobacillus ferrooxidans is of great importance for the hydrometallurgical production of base, heavy and noble metals such as copper, zinc, cobalt, uranium and gold. Because of the practical importance of these processes, the mechanism of sulfide biooxidation has been a subject of intensive studies in the recent decades. Two main mechanisms of the oxidation kinetics have been proposed [1]:

•

the direct oxidation is based on the oxidation of sulfide minerals by an immobilized on their surface microbial cell in the absence of intermediate oxidants such as ferric iron. It is based on the following reaction (in the case of pyrite oxidation): (1) 4FeS2 + 15O2 + 2H2O = 2Fe2(SO4)3 + 2H2SO4

•

the indirect oxidation is based on the ability of A. ferrooxidans and other ironoxidizing microorganisms to oxidize ferrous ions in aqueous solution: (2) 2FeSO4 + H2SO4 + ½O2 = Fe2(SO4)3 + H2O The ferric ions, produced as a result of this reaction, then can chemically oxidize metal sulfides, such as pyrite: 7Fe2(SO4)3 + FeS2 + 8H2O = 15FeSO4 + 8H2SO4 (3) The ferrous ions produced by reaction (3) are then oxidized by A. ferrooxidans to ferric ions according to reaction (2). The ferric sulfate, produced by reaction (2) is further 1067

Microbiology Fundamentals

reused for the chemical oxidation of pyrite (reaction 3), which closes the cycle of iron ions oxidation-reduction. After many years of research, it seems that a consensus has been reached that the indirect mechanism is if not the only [2-4], at least the predominant [5] mechanism in sulfide biooxidation by A. ferrooxidans. However, the rate of purely chemical (sterile) oxidation of pyrite by aqueous solutions of ferric sulfate is by an order of magnitude lower than that in the presence of iron oxidizing microorganisms [6]. Different explanations have been proposed to explain this discrepancy. Boon and Heijnen [6] explained the higher rate of the biological sulfide oxidation by the fact that the ferrous ions, produced as a result of the sulfide oxidation (3) are quickly reoxidized by microorganisms, which keeps the redox potential high; in the sterile oxidation of sulfides by ferric iron, the redox potential falls quickly because of the decrease in the ratio of ferric to ferrous ions concentration. Alternatively, Gehrke et al. [7] explained the acceleration of the bacterial process over the sterile one by the quick reoxidation of iron, which is complexed with microbial extracellular polymeric substances. The purely hydrodynamic retention of ferric ions near the point of contact of microorganisms with the sulfide surface has not been considered yet. In this work, we are trying to explain the higher rate of bacterial sulfide oxidation, compared to the sterile one, by the hydrodynamic retention of ferric ions. Computational fluid dynamics simulation is used to determine the ferric ions concentration profiles in the vicinity of a microorganism, attached physically to the surface of a sulfide crystal. 2.

NUMERICAL MODEL The concentration and velocity profiles around a single microbial cell, immobilized on the flat surface of a sulfide crystal were modeled numerically using the FLUENT computational fluid dynamics software. 2.1 Assumptions The following assumptions were made in order to develop the mathematical model: Geometrically, the A. ferrooxidans cell was approximated as an ellipsoid of rotation with a size along the axis of symmetry equal to 2 micrometers and a small axis, perpendicular to it, of 1 micron (Fig. 1). While the size and shape slightly differ in different strains, the one used here is a typical shape and size. The surface area of the microbial cell was 5.34x10-12 m2. The cell is a rigid solid particle, and therefore:

− It does not deform as a result of the acting hydrodynamic forces. − The cell does not deform as a result of immobilization on the solid surface, especially at the point of contact. The cell is immobilized on a flat surface (Fig. 1). The immobilization support is chemically inert. No extacellular polymeric substances (EPS) were assumed to exist between the cell and the sulfide surface. The available literature data shows that in some cases A. ferrooxidans forms EPS when immobilized on a sulfide surface, while in other they are absent. The ferric ions concentration in the bulk of the liquid is equal to zero. 1068

Microbiology Fundamentals

The ferric ions are produced evenly over the entire surface of the cell and are completely released into the liquid. The literature data show that the process of iron oxidation takes place outside of the cell membrane and the iron ions do not permeate the membrane [8]. The iron oxidation rate was assumed to be equal to 5.2x10-8 (kg Fe3+)/(s.(m2 of cell surface)). The approaching liquid flows in either laminar or turbulent mode, parallel to the solid surface (Fig. 2), with two bulk velocities: 0.1 and 1.0 m/s. The case for 1.0 m/s was calculated for two different cases – laminar and turbulent liquid flow. For the turbulent flow the turbulence intensity was 5% and the turbulent scale was chosen at 0.01 mm. The real geometry of the microbial cell has been scaled up by a factor of 100 in order to avoid round-off error of computing too small numbers (note that FLUENT computes only in SI units). This required scaling down the liquid velocity by the same factor in order to keep a constant Reynolds number. The computational mesh consisted of 0.83 mill tetrahedral cells. The surface mesh of the bacteria is shown in Fig. 3. Advance turbulent model, namely the FLUENT Shear-Stress Transport (SST) k-Ω model was applied to resolve the low-Reynolds number turbulence around the bacteria.

Figure 1. Geometry of the immobilized microbial cell

1069

Microbiology Fundamentals

Figure 2. Scheme of the modelling space

Figure 3. Bacteria surface mesh

2.2 Numerical results and discussion The velocity and concentration profiles in the plane x-z (side view) were obtained (Figs. 4-6) using the model, described above.

1070

Microbiology Fundamentals

The ferric iron concentration profiles in liquid, at the plane of contact with the sulfide crystal surface, are shown in Fig. 7. Actually, Fig. 7 shows the top view (x-y plane) of Figs. 4, 5 and 6, assuming that the microbial cell is "transparent".

Figure 4. Velocity and ferric ions concentration profiles around the immobilized cell. Liquid velocity: 0.1 m/s (laminar flow)

1071

Microbiology Fundamentals

Figure 5. Velocity and ferric ions concentration profiles around the immobilized cell. Liquid velocity: 1.0 m/s (laminar flow)

1072

Microbiology Fundamentals

Figure 6. Velocity and ferric ions concentration profiles around the immobilized cell. Liquid velocity: 1.0 m/s (turbulent flow)

1073

Microbiology Fundamentals

Figure 7. Top view of the ferric ions concentrations in liquid, contacting the sulfide crystal surface

The concentration profiles in Fig. 7 show that both the free stream velocity and flow structure (laminar vs. turbulent) of liquid affect the concentration distribution only slightly. This is due to the fact that the size of the microorganism is in the order of micrometers, thus the entire microorganism is within the laminar viscous layer of the liquid, regardless of the type of the flow – laminar or turbulent. Within the liquid viscous sublayer of 1 micrometer, the liquid velocity is very close to zero, which can be seen also from the velocity plots in Figure 4. Therefore, the upstream velocity magnitude, as well as 1074

Microbiology Fundamentals

the upstream turbulence does have a significant effect on the concentration distribution (Fig. 7). The tail of a relatively larger concentration downstream the bacteria is wider at smaller upstream velocity. At larger velocity the larger amount of the ferric ions are advected by the flow and the thinner tail of higher concentration is formed downstream of the microbial cell. 3.

CONCLUSIONS The results of the simulation show that there is significant spatial variation in ferric ions concentration in the close surrounding of the microbial cell (Fig. 7). While the concentration at the vicinity of the contact between the microbial cell and the crystal surface is well over 100 mg/L, it drops to zero at a distance of 1 micrometer from the cell surface (except for the tail behind the cell). Therefore, the concentration of ferric ions in the bulk of liquid, far away from the cell, is different - much lower - than that on the sulfide surface, surrounding the immobilized cell. The redox potential of the liquid, given by the Nernst equation: (4) E = Eo + (RT/F)ln(Fe3+/Fe2+) is a strong function of the ferric-to-ferrous ions concentration ratio. Therefore, since the concentration of ferric ions in the immediate cell surroundings is higher than the bulk concentration, the redox potential at the surface of the sulfide crystal will be higher that that in the bulk of liquid. Since the ferric iron concentration at the cell-sulfide interface is higher than the bulk concentration by approx. 150 mg/L, the redox potential at the cellsulfide interface will be significantly higher than the bulk potential mainly in the cases when the bulk ferric iron concentration is below approx. 1 g/L. ACKNOWLEDGEMENTS This work was supported financially by the Natural Sciences and Engineering Research Council of Canada and the Premier’s Research Excellence Award (Ontario). REFERENCES

1. 2. 3. 4. 5. 6. 7. 8.

Α.E. Torma, Adv. Biochem. Eng., 6 (1977) 1. W. Sand, T. Gehrke, P.-G. Jozsa and A. Schippers, Hydrometallurgy, 59 (2001)159. H. Tributsch, Hydrometallurgy, 59 (2001) 177. G.S. Hansford and T. Vargas, Proc. Metall., 9A (1999) 13. K. Lilova and D. Karamanev, (paper to appear in this volume) M. Boon and J. J. Heijnen, Hydrometallurgy, 48 (1998) 27. T. Gehrke, R. Hallmann, K. Kinzler and W. Sand, Water Sci. Tech., 43 (2001) 159. W. J. Ingledew, Biochim. Biophys. Acta, 683 (1982) 89.

1075

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Contribution to the quantification of the Acidithiobacillus ferrooxidans biomass concentration from the oxygen uptake rate D.S. Savić, V.B. Veljković, M.L. Lazić Faculty of Technology, 16000 Leskovac, Bulevar oslobodjenja 124 Abstract A simple, rapid and reproducible method for quantifying the A. ferrooxidans biomass concentration from measuring the oxygen uptake rate is described. The specific oxygen uptake rate, found to be 0.020 ± 0.001 mg O2/mgd.w./min, for both concentrated cell suspensions and during ferrous iron biooxidation, is used in the estimation of the viable biomass concentration. The ferrous iron concentration in samples of the culture medium, where the oxygen uptake is recorded, should be higher than 0.1 mg/L to avoid substrate limitation.

Keywords: Acidithiobacillus ferrrooxidans, qantification of viable biomass concentration 1.

INTRODUCTION The role of the chemolithotroph bacterium Acidithiobacillus ferrrooxidans in the leaching of metals from sulfide ores has been well recognized. The bacterium obtains energy for growth by the oxidation of ferrous to ferric iron using atmospheric oxygen, a desirable reaction from the hydrometallurgical point of view. As in other cases, when oxygen is closely related to the energetics of microbial growth under aerobic conditions [1], the oxygen consumption of A. ferrooxidans could be correlated to its growth rate and, therefore, used for indirect kinetic studies. A few earlier studies on oxidations by A. ferrooxidans, such as the ferrous iron oxidation rate by cell envelopes [2], the competitive inhibition of ferrous iron oxidation by cells and ferric iron [3, 4], and the relative contributions of adhered and free suspended cells to pyrite biooxidation [5], are based on the measurement of oxygen uptake. The oxygen uptake rate is also suggested for evaluating bacterial activity for metal leaching [6]. A high degree of synchronization among processes of oxygen uptake, biomass growth and ferrous iron utilization during ferrous iron oxidation by A. ferrooxidans in different bioreactors, independently of pH (1.8 or 2.0), has been observed [7]. Thus, the dissolved oxygen level could be used not only for studying the kinetics of the biooxidation process but also for determining the viable biomass concentration. Such a method, based on the measurement of the oxygen uptake rate, would be a simpler technique for quantifying the A. ferrooxidans growth during the biooxidation of ferrous iron than the standard ones, which have disadvantages in requiring considerable time, materials and manual labor (serial end-point dilution, enumeration of colonies) or refers to dead and living cells 1077

Microbiology Fundamentals

(direct microscope counting, the measurement of the total dry weight and protein content). The viable biomass concentration and the growth rate have been successfully determined with sufficient accuracy by measuring the oxygen uptake rate during aerobic fermentations, for instance by Saccharomyces cerevisiae, Streptomyces sp. and Thermoactinomyces sp. [8] as well as Clostridium boldinii [9]. In the present work, the growth and oxygen uptake of A. ferrooxidans cells growing on a ferrous iron medium were studied under different conditions with respect to the ferrous iron concentration, the biomass concentration and the ferric/ferrous ratios. The main aim was to reflect the relationship between microbial growth and oxygen uptake and to establish a quantitative base for determining the viable biomass concentration by measuring the oxygen uptake rate. 2.

MATERIAL AND METHODS

2.1 Microorganism and cultivation conditions A strain of A. ferrooxidans B5, isolated from the Bor Copper Mine (Serbia), was grown on the following mineral salt medium (in g/L): (NH4)2SO4 3; KH2PO4 0.5; MgSO4.7H2O 0.5; and KCl 0.1, the initial ferrous iron concentration being 9 g/L. The culture was grown in an Erlenmeyer flask (2 L) containing 1 L of the medium at 28°C on a rotary shaker at 200 rpm. 2.2 Biomass

The biomass was determined by measuring the dry weight - mgd.w (80°C, 24 h) of a membrane filter (0.45 µm pore diameter, Millipore, NVC 45218) after the sample (500 mL) was filtered under pressure (2 bar). 2.3 Preparation of a concentrated cell suspension The biomass was harvested at the end of the exponential growth phase; it was therefore assumed that the biomass contained a negligible number of dead cells. A volume of the growing culture (500 mL; 25 mgd.w./L) was centrifuged at low speed (1000 min-1, 10 min) to remove the insoluble ferric iron, filtered through a membrane filter (0.45 µm pore diameter, Millipore, NVC 45218), washed twice with 0.05 M H2SO4 solution (100 mL) and once with distilled water (100 mL), and concentrated by resuspending the biomass in the iron-free salt medium (25 mL). According to the ratio of the oxygen uptake rate of the original culture and that of the concentrated suspension, about tenfold concentration was achieved (238 mgd.w./L). 2.4 Oxygen uptake rate The oxygen uptake rate was calculated from the slope of the linear part of the relationship between the dissolved oxygen level and time. The dissolved oxygen level was measured by a Clarke-type oxygen probe (HANNA HI8043), and the oxygen solubility (6.7 mg O2/L) in the mineral salt medium was estimated [10]. The reaction mixture (5 mL) was prepared by mixing a certain volume of the concentrated cell suspension with different volumes of ferrous iron and ferric iron solutions in the mineral salt medium to achieve different ratios of ferric/ferrous iron. The oxygen uptake rate of the cell-free mineral medium was found to be negligible (0.007 mg O2/L/min).

1078

Microbiology Fundamentals

2.5 Analytical methods The ratio of ferrous and ferric ions was calculated from the redox potential measured using a redox probe (HANNA HI 3932). The pH was monitored with a pH meter (HANNA HI 9025). 3.

RESULTS AND DISCUSSION A summary of the effect of different biomass concentrations and ferric/ferrous ratios on the specific oxygen uptake rate of the concentrated bacterial cells is given in Tables 1 and 2. When the biomass concentration is held constant at a level of 8.0 mgd.w./L and the ferric/ferrous ratio is varied in the range from 1/10 to 1/0.1, which causes a change in the pH of the reaction mixture from 1.9 to 1.4, the specific oxygen uptake rate does not change. A mean value of 0.0185 ± 0.001 mg O2/mgd.w./min (Table 1) was obtained. Also, if the ferric/ferrous ratio of 1/10 is held constant and the biomass concentration in the reaction mixture is varied from 4.3 to 47.7 mgd.w./L, the specific oxygen uptake rate is also constant at a level of 0.020 ± 0.003 mg O2/mgd.w./min (Table 2). At the significance level of 0.05, as evaluated by the t-test (two populations), the two means are not significantly different. As both the biomass concentration and the ferric/ferrous iron ratio do not have a significant effect, the mean value of the specific oxygen uptake rate for all the experimental conditions applied was calculated: 0,020 ± 0.002 mg O2/mgd.w./min. Table 1. Effect of the ferric/ferrous iron ratio on the specific oxygen uptake rate (biomass concentration: 8.0 mgd.w./dm3) Fe(III)/Fe(II) 1/10 1/7 1/5 1/2 1/0.3 1/0.1 Mean value Standard deviation

qO2, mg O2/mgd.w./min 0.019 0.019 0.018 0.018 0.019 0.018 0.0185 ±0.001

Table 2. Effect of the biomass concentration on the specific oxygen uptake rate (ferric/ferrous iron ratio: 1/10) X, mgd.w./dm3 4.3 8.0 13.6 17.9 47.6 Mean value Standard deviation

qO2, mg O2/mgd.w./min 0.017 0.019 0.025 0.021 0.022 0.020 ± 0.003

The specific oxygen uptake rate found for the concentrated cells was practically the same as that determined in our studies on ferrous iron biooxidation in a bubble column without and with pH regulation at 1.8 and 2.0 and in a stirred tank with pH regulation at 2.0, as may be seen in Table 3. The mean value of the specific oxygen uptake rate was calculated from all our data: 0.020 ± 0.001 mg O2/mgd.w./min. This value agrees quite well 1079

Microbiology Fundamentals

with those measured for A. ferrooxidans grown in a batch reactor using the same type of oxygen probe [11] or the manometric technique [12] while other published values are twice [13, 14] or even tenfold [2] higher. The disagreement is probably due to different strains, measuring methods or cultural conditions applied in the different studies. Table 3. Specific oxygen uptake rate of A. ferrooxidans under different culture conditions* Culture condition Concentrated cells, pH 1.4-2.4 Bubble column, 9 gFe/L, 28°C without pH regulation (pH 2.4-2.7) Bubble column, 9 gFe/L, 28°C with pH regulation: 1.8 2.0 Stirred tank, 9 gFe/L, 28°C with pH regulation: 2.0 Fermentor, 9 gFe/L, 28°C, pH 3.2 Batch reactor, 10 gFe/dm3, 30°C, pH 1.6 Sample from uranium b leaching + 0.20 gFe, 30°C Not defined a

qO2, mgO2/mgd.w./min 0.020 ± 0.002 0.021 ± 0.001

0.018 ± 0.001 0.020 ± 0.001 0.021 ± 0.002 0.205 (230 mmol O2/mgpr./h)c 0.020 (630 nmol O2/mgd.w./min) 0.025 (700 mL O2/mgpr./h)c

Reference This work Unpublished data [5]

[5] [2] [11] [12]

0.045 [13] (20840 mL O2/mgN/h)c 0.042 [14] Shaken flasks, 0.56 gFe/dm3, 30°C, pH 1.6 (785 nmol O2/mgpr./min)c * In all cases, a Clarke oxygen electrode was used with the following exceptions: (a) not defined and (b) the manometric method c Assuming that the dry biomass of A. ferrooxidans contains about 60% of proteins [16, 17] and 12.4% of nitrogen [15].

The oxygen uptake rate of A. ferrooxidans, evaluated in experiments with the concentrated cell suspensions and during ferrous iron biooxidation is shown in Figure 1 as a function of the biomass concentration. All data fit a straight line, indicating that the oxygen uptake rate of the bacterial population, OUR (in mg O2/L/min), during ferrous iron biooxidation can be correlated to the biomass concentration, X (mgd.w./L), by the well-known equation:

OUR = qO2 X where qO2 is the specific oxygen uptake rate of the strain A. ferrooxidans B5 (0.020 mg O2/mgd.w./min). This correlation allows the determination of the viable biomass concentration of A. ferrooxidans from the oxygen uptake rate measured during the biooxidation process. The only limitation for applying the above equation is the ferrous iron concentration in the culture medium where the oxygen uptake rate is measured. From the stoichiometry the ferrous iron oxidation and cell synthesis by A. ferrooxidans [15], and taking the maximum oxygen rate and the usual duration time of the oxygen uptake rate measurement as 0.70 mg O2/L/min [7] and 20 minutes, respectively, the critical ferrous iron concentration in the medium, which limits the bacterial growth, was estimated to be about 0.1 g/L. This value was verified by measuring the oxygen uptake rates in samples 1080

Microbiology Fundamentals

taken from the stationary growth phase, where the ferrous iron concentration was below the critical value, and in the same samples after the addition of a ferrous iron solution to adjust the ferrous iron concentration above the critical level. A summary of the oxygen uptake rate of samples taken from the stationary growth phase with and without the addition of ferrous iron is presented in Table 4. The oxygen uptake rate in the latter is lower than that in the former, indicating the limitation of oxygen uptake by ferrous iron.

Figure 1. Oxygen uptake rate (OUR) as a function of dry biomass concentration (experiments with concentrated biomass suspension: ♦; the biooxidation of ferrous iron in a bubble column without pH regulation - U [unpublished data] and with pH regulation at 1.8 and 2.0 - ; and the biooxidation of ferrous iron in a stirred tank with pH regulation at 2.0: [7])

Table 4. The oxygen uptake rate of A. ferrooxidans in samples from the stationary growth phase in a bubble column* before and after the addition of ferrous iron (up to 9 g/L) Oxygen uptake rate, mg O2/L./min Before the addition of After the addition of ferrous iron ferrous iron 2.4-2.7 0.08 0.08 0.41 1.8 0.08 0.07 0.57 2.0 0.03 0.08 0.50 0.05 0.06 0.56 0.05 0.055 0.48 * Data from the experiments described in [7], but not shown there. pH

Ferrous iron concentration in the culture medium, g/L

1081

Microbiology Fundamentals

4.

CONCLUSIONS Quantification of the A. ferrooxidans biomass concentration using usual methods, such as serial end-point dilution and enumeration of colonies, is uncertain and timeconsuming. The measurement of dry biomass, although simpler and reproducible, can only be used for estimating the biomass concentration in large bioreactors since a volume of 500-750 mL [7] is needed as a sample. This work describes a simple method for quantifying the A. ferrooxidans biomass concentration indirectly by measuring the oxygen uptake rate by a cell population. The advantages of this method are that the test is rapid, small samples are needed, and the results are reproducible. The method could also be appropriate for enumerating aerobic microorganisms in bioleaching experiments, since many cells are attached to the ore and could not be detected by standard methods. The following assumptions were made when the indirect method was used. Firstly, all the oxygen consumed in the bioprocess was only used for ferrous iron oxidation, the oxygen, which became part of the biomass was derived from combined oxygen found in CO2, and the oxygen required for maintenance was negligible. Secondly, autooxidation of the ferrous iron did not occur, which was verified expermentally. REFERENCES

1. M.L. Shuler and F. Kargi, Bioprocess Engineering, Prentice Hall, Englewood Cliff, New Jersey, 1992. 2. C. Bodo and D.G. Lundgren, Can. J. Microbiol., 20 (1974) 1647 3. H.M. Lizama and I. Suzuki, Appl. Environ. Microbiol., 55 (1989) 2588 4. I. Suzuki, H.M. Lizama, P.D. Tackaberry, Appl. Environ. Microbiol., 55 (1989) 1117 5. D. Savić, V. Veljković, M. Lazić and M. Vrvić, in: R. Amils and A. Ballester (eds.), Biohidrometallurgy and the Environment toward the Mining of the 21st Century, Elsevier, Amsterdam, 1999, Part A, pp 625-630. 6. G.I. Karavaiko, G. Rossi, A.D. Agate, S.N. Groudev and Z.A. Avakyan, (eds.) Biogeotechnology of Metals, CIP GKNT, Moscow, 1988 7. V. Veljković, D. Savić, M. Lazić and M. Vrvić, in: R. Amilis and A. Ballester (eds.), Biohidrometallurgy and the Environment toward the Mining of the 21st Century, Elsevier, Amsterdam, 1999, Part A, pp 617-623. 8. D.W. Zabriskie and A.E. Humphrey, AIChE J., 24 (1978) 138 9. M. Reuß, R.P. Jefferis and J. Lehman, GBR Monogr., 3 (1976) 107 10. A. Schumpe, I. Adler and W. Deckwer, Biotechnol. Bioeng., 20 (1978) 145 11. J.T. Pronk, W.M. Meijer, W. Hazeu, J.P. vanDungen, P. Bos and G. Kuenen, Appl. Environ. Microbiol., 57 (1991), 2057 12. C.L. Brierley, in: L.E. Murr, A.E. Torma and J.A. Brierley (eds.) Metallurgical Applications of Bacterial Leaching and Related Microbiological Phenomena, Academic, New York, 1978, pp. 345-364 13. S.N. Groudev, F.N. Genchev and S.S. Gaidarjev, in: L.E. Murr, A.E. Torma and J.A. Brierley (eds.) Metallurgical Applications of Bacterial Leaching and Related Microbiological Phenomena, Academic, New York, 1978, pp. 253-272. 14. D.P. Kelly and C.A. Jones, in: L.E. Murr, A.E. Torma and J.A. Brierley (eds.) Metallurgical Applications of Bacterial Leaching and Related Microbiological Phenomena, Academic, New York, 1978, pp. 19-44. 15. J.R. Smith, R.G. Luthy and A.C. Middleton, J. WPCF, 60 (1988) 518 1082

Microbiology Fundamentals

16. C.A. Jones and D.P. Kelly, J. Chem. Tech. Biotechnol., 33B (1983) 241 17. A.H. Basaran and O.H. Tuovinen, Coal Prep., 5 (1987) 39

1083

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Electrochemical and microbiological characterization of mercury in contact with mud F. Cruza, A. Welzelb, C. Sampaioc, G.E. Englertc∗, I.L. Müllerc a

Material Department - UFRGS - Brazil b PPGEM - UFRGS - Brazil Metallurgical Department - UFRGS - Brazil Abstract HgCl2 poured onto liquid nutritive broth was used as culture media for aerobic bacteria, Bacillus sp. This media was used to evaluate the bacterial activity resistance to toxic metal and to activate the bacteria protective mechanism. Two interconnected flasks were projected to evaluate: i) growth of the bacteria and ii) the electrochemical behaviour of gaseous mercury trapped in acid solution. Bacteria were cultivated and maintained in plate count agar (PCA), a solid nutritive broth. Electrochemical experiments were done in a liquid culture medium using a Potentiostat PAR EG&G 273A. Potentiodynamic curves were obtained over a platinum surface during 10 days to follow the oxygen depletion by the bacteria. Electrochemical experiments were conducted in nutritive broth with and without mercuric chlorides and bacteria. At the end of the experiments bacteria remained viable despite the contact with toxic metal while the broth pH value changed of 7 through 10. 1.

INTRODUCTION It is well known that mercury is a toxic metal and any contaminated with it cannot be buried due to its hazardous effect. This kind of residues is catalogued as belonging to the Class I. Gaseous monatomic mercury Hg° may be produced by several natural sources and then it is delivered to the atmosphere. Almost 95% of atmospheric mercury is found in the Hg° state. However the fraction value that is free as a monatomic gas or adsorbed onto the solid particles, which are suspended in the air, may change considerably (1). On the other hand, in aqueous media, Hg can be oxidized to HgII as a result of the ozone activity (1). In this case bacteria with an active methylating mechanism seem to regulate environmental conditions that control the bioavailability of HgII. It was proposed in the literature that the accumulation rate of mercury in water depends on pH, selenium concentration, presence of sulfate, dissolved organic carbon in seepage lakes and decomposable organic matter (2). One of the ways of determining the rate of oxidation of mercury is by measuring the concentration of HgII. Some researchers have studied the detection of mercury through bio-indicators but they are not specific to a given form of HgII. It is also possible to detect the presence of mercury by electrochemical methods using platinum electrodes (3).

∗

Universidade Federal do Rio Grande do Sul - Av. Osvaldo Aranha 99/615-D.90035-090 Porto Alegre, RS_Brasil. [email protected] Phone numbers 55 51 3316 3344/3212 4547

1085

Microbiology Fundamentals

Many researchers have been looking for a technical solution to eliminate mercury from waters using microorganisms (1, 4, 5). Since microorganisms live in environments that can be reproduced artificially, it is possible to isolate them easily. Hence, it is possible to determine the effects that environmental factors play on the availability of inorganic mercury in the presence of bacteria in aquatic environments. Such knowledge is very important from an ecological point of view (3). 2.

EXPERIMENTAL PROCEDURE Bacterial strain. A portion of mud was collected from an industrial area in order to isolate one microorganism adapted to the environmental conditions. They were placed in contact with a liquid broth appropriated for aerobic bacteria. The liquid broth consisted of a solution of peptone, yeast extract and sodium chloride in water at pH 7. During one week the culture was kept at 30°C to produce a quick growth. After this stage, a liquid portion was poured in a dish plate in order to start to the bacterial isolation. The culture was then maintained at the same temperature indicated before. Bacterial identification was carried out by morphological aspects, temperature resistance (after the broth liquid with bacteria was heated up to 80°C for 15 minutes) and biochemical tests (Table 1). The temperature resistance indicates conditions for spore formation and the biochemical tests pointed out to Bacillus sp. Table 1. Biochemical tests to identify bacteria Motility Urease Citrate Gelatin Nitrate reduction Glucose Catalase Oxidase

+ + + + + + +

An aqueous solution with bacteria characterized previously was taken after 48 hours of incubation and poured into a fresh portion of modified liquid broth. The modified broth consisted of 1 mL of bacteria solution plus 99 mL of sterilized water. (In order to improve the bacterial growing, oxygen was continuously poured onto the flask). The resistance of the microorganisms to mercury was evaluated adding 1mL of 0.5M HgCl2 and HgCl2 which is a toxic solution to 0.2 L of broth liquid. The later solution pH was reduced from 7 to 4 by adding some drops of 0.1M nitric acid. The viability of the microorganisms was evaluated by counting them at Neusbauer camera. Some other tests were also carried out with the same solution mentioned before at pH 7, with and without mercury. Electrochemical parameters. Anodic and cathodic potentiodynamic curves, in broth culture with and without mercury, in the presence and in the absence of bacteria, were obtained using a potentiostat PAR G&G 273A.The saturated calomel electrode (SCE) was used as a reference and a graphite was an auxiliary electrode. The working electrode (where electrochemical reactions happened) consisted of a mini-electrode of platinum with an area of 0.00035 cm2. All the experiments started at -1.0V sce and went until +1.0V sce at a scan rate of 2 mV/s. All experiments were done at sterilized and quiescent conditions.

1086

Microbiology Fundamentals

3.

RESULTS AND DISCUSSION Biochemical tests to characterize the bacteria indicate Bacillus sp. as the isolated bacteria. They are considered resistant to mercury according to literature (1,4,5). In this work it was observed that those species cannot grow as well in modified media as they do in the original media, although both were contaminated by Hg. In the former media the organic concentration was low, since the solution was diluted so the bacteria were in closer contact with mercury. The morphological aspects after experimentation in modified media with Hg are shown in Figure 1, in which the stressed structure is clearly observed. Bacteria that survived in a complete media without Hg were not stressed as can be clearly seen in Figure 2. It was observed that bacteria species in contact with modified broth with Hg survived despite the oxygen concentration, which was reduced progressively after electrochemical experiments (Figure 3).

Figure 1. Bacteria after 30 days in modified media without Hg

Figure 2. Bacteria after 30 days in modified media with Hg

1087

Microbiology Fundamentals

1,5

st

th

1 day

after 5 day th

1,0

th

after 10 day

10 day

E (VSCE)

0,5 0,0 -0,5 -1,0 1E-7 1E-6 1E-5 1E-4 1E-3 0,01

0,1

1

10

100

2

i (A/cm )

Figure 3. Electrochemical curves in modified media with bacteria and without mercury

It was observed that, even reducing the oxygen concentration that was measured by the reduction of the current density, the bacteria were kept alive. The turbidity of the liquid (which remained for more than 10 days without any oxygen reposition) was an evidence that bacteria were still alive and was confirmed when bacteria grew in PCA. From electrochemical point of view it is possible to observe in Figure 3 that at the 5th day of the experimentation, there was a dislocation of the anodic curve to the right hand when, apparently, all oxygen had been removed by bacteria. The shape of the polarization curve is an indicative that the hydrogen cathodic reaction was the cathodic reactive, especially because pH was low. It is well known that the main cathodic reactions from the electrochemical point of view are: the reduction of oxygen dissolved in water and the reduction of the hydrogen from the water or from the acid media. The reintroduction of oxygen to the system brought the curve to the original behaviour with lower current density values. In order to compare the experiments with and without bacteria (data not shown), the same experiment was done with the liquid broth media alone and it was possible to observe that the cathodic current values did not suffer any change. This is a strong indicative that, despite oxygen is by far continuously removed by its electrochemical reduction, remotion is not enough to dislocate the curves as can be seen in Figure 3. Finally, a tiny anodic peak was observed and it may be an indicative that some change happened when comparing the experiments at the first day and the 15th day. As microorganisms grew, the production of their by-products is high enough to interact with the platinum surface. It is known that factors such as low concentration of organic carbon dissolved, the presence of sulphate, low alkalinity and low pH, may interfere in the mechanism by which bacteria retain mercury (2). However, a high concentration of total organic carbon may interfere in the mechanism that bacteria use to reduce mercury. From the anodic curve (data not shown) obtained in the broth liquid with mud, it was not possible to confirm the presence of Hg probably due to the interference to the high organic matter concentration. It is important to highlight that previous analysis through atomic absorption confirmed the presence of Hg in the mud in a concentration of a few ppm. The 1088

Microbiology Fundamentals

oxygen uptake is partially responsible to the high pH value observed at the end of all experiments since the initial value was 7 and the final one was 10. The participation of the by-products may also contribute to the high pH value observed. 4.

CONCLUSION Tested bacteria Bacillus sp. can survive without oxygen. They create self-mechanisms to produce a new by-product noticed by the changes in cathodic potentiodynamic curves and also in pH values. This observation was confirmed by pH 10 registered at the end of experimentation. The electrochemical tools may be jointly used with the microbiological ones to characterize the growth of bacteria during mercury remotion from contaminated liquids. ACKNOWLEDGEMENTS To FAPERGS and CNPq for the financial support in the form of scholarship as well to Prof. Washington Aliaga by his help to structure this work. REFERENCES

1. T. Smith, K. Pitts, J.A. McGarvey, A.O. Summers. Applied and environmental microbiology, Apr. 1998, vol. 64, No 4, p. 1328-1332. 2. T. Barkay, M. Gillman, R.R. Turner. Applied and environmental microbiology, Nov. 1997, vol. 63, No 11, p. 4267-4271. 3. Ionashiro, E.Y., Souza, G.R., Milaré, E., Bebedetti, A.V., Ionashiro, M. and Fertonani, F.L. XII Simpósio Brasileiro de Eletroquímica e eltroanalítica, April, 2001, CDrom, www.ufrgs.br/sibee. 4. A.O. Ogunseitan. Applied and environmental microbiology, Fev. 1998, vol. 64, No 2, p. 695-702. 5. M.D. Mullen, D.C. Wolf, F.G. Ferris. Applied and environmental microbiology, Dec. 1989, vol. 55, No 12, p. 3143-3149.

1089

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Evaluating the growth of free and attached cells during the bioleaching of chalcopyrite with Sulfolobus metallicus B. Escobar, M.J. Hevia and T. Vargas Centro de Estudios en Hidro/Electrometalurgia, Departamento de Ingeniería de Minas/Departamento de Ingeniería Química, Universidad de Chile, Tupper 2069, Santiago, Chile. Fax 56-2-6991084 Abstract Chalcopyrite can be dissolved at covenient rates in the presence of thermophilic microorganisms such as Sulfolobus metallicus, a chemolitothrophic archaea that can grow autotrophically at temperatures between 65 and 80°C. The catalytic influence of this microorganism is partially related to the microbial oxidation of ferrous iron and the oxidation of residual sulfur produced during chalcopyrite dissolution. Chalcopyrite bioleaching studies, however, normally report only the growth of the planktonic population of Sulfolobus metallicus which makes difficult to assess the pattern of utilization of these two subtrates by the microorganism during the process. The present work reports experimental studies on the bioleaching of a chalcopyrite concentrate with Sulfolobus metallicus in shake flasks at 70ºC and pH 1.8. The growth of the planktonic population was determined by cell counting, MPN and protein determination by Lowry. The population of microorganisms attached to the solids was evaluated by protein determination using a modified Lowry method. Results showed that about 68% of the Sulfolobus metallicus present in the flask grows attached to chalcopyrite particles, while the remaining 32% grows in the solution. The percentage associated to attached growth coincides with the fraction of microorganisms which initially attaches to chalcopyrite after addition of the inoculum. Cell growth during the first 72 hours of both attached and free microorganisms is mainly based on the oxidation of ferrous iron. Bacterial oxidation of sulfur, which is conducted only by attached microorganisms, becomes relevant only after 72 hours of leaching. After 120 hours of leaching there is a net reduction of both the planktonic and attached population which evidences an important process of cell death and breakage.

Keywords: S. metallicus, bioleaching, chalcopyrite, biomass estimation, attached cells 1.

INTRODUCTION Primary copper sulfides such as chalcopyrite has been demonstrated to be refractory to bacterial leaching with mesophiles such as At. ferrooxidans in ambient conditions (1). During the last years, however, several researchers have demonstrated that chalcopyrite can be dissolved at covenient rates in the presence of thermophilic microorganisms such

1091

Microbiology Fundamentals

as Sulfolobus metallicus, Acidianus brierley or Metallosphera sedula (2, 3). The use of Sulfolobus metallicus, a chemolitothrophic archaea that can grow autotrophically at temperatures between 65 and 80°C, has been largely investigated in the bioleaching of chalcopyrite (4-6). The catalytic influence of S. metallicus in this process is partially related to the microbial oxidation of ferrous iron, an activity which has been evidenced by several authors (1, 7). Other works have also indicated that S. metallicus presence have also a catalytic influence on the oxidation of residual elemental sulfur to sulfate (7,8). In chalcopyrite bioleaching studies only the growth of the planktonic S. metallicus is normally monitored, which makes difficult to assess the relative importance of these two substrates. The present work reports experimental studies on the bioleaching of a chalcopyrite concentrate with Sulfolobus metallicus in shake flasks at 70ºC and pH 1.8. An important methodological aspect, the growth of both the population of microorganisms attached to the sulfide particles and microorganisms free in solution were simultaneously monitored during these experiments. The relative importance of sulfur and ferrous iron as substrates for microbial growth was evaluated from relating bacterial growth to the evolution of dissolved metals and pH during the experiment. 2.

MATERIAL AND METHODS

2.1 Microorganism A strain of Sulfolobus metallicus was used in this study. The cells were grown in basal medium containing 1% chalcopyrite concentrate as the energy source. The composition of the basal medium was 0.4 (NH4)2SO4, 0.5 MgSO4.7H2O and 0.2 KH2PO4 g/l, adjusted to pH 1.6 with sulfuric acid. 2.2 Mineral A -325 +400 mesh (-0.045mm +0.038mm) sample of a high purity chalcopyrite sample prepared from an Andina Copper concentrate was used in the experiments. 2.3 Experimental procedure Attachment experiments were conducted in 125 ml Erlenmeyer flasks containing 50 ml basal medium. One g of chalcopyrite was contacted with an initial concentration of 1.1x108 cells of S. metallicus/ml. The flasks were incubated in an environmental shaker at 70°C and 100 rpm. The rate of cell attachment was determined by periodic sampling and determination of free cells by direct counting onto the microscope. Leaching experiments were conducted in 250 ml Erlenmeyer flasks containing 95 ml of the basal medium, 1% (w/w) chalcopyrite and 5 ml of active inoculum of Sulfolobus metallicus. The flasks, 8 in total, were incubated in an environmental shaker at 100 rpm and 70°C. Two flasks were removed at different leaching times to determine total iron, ferrous iron, copper and pH in solution, and bacterial growth in solution and solids. Total and ferrous iron in the solution were determined by o-phenanthroline colorimetric method (9); for solutions with high ferric:ferrous ratio ferrous iron was determined by the modified o-phenanthroline colorimetric method (10). The number of free cells in solution was determined by direct counting onto the microscope using a Petroff-Hausser chamber and by the MPN of iron oxidizing microorganisms, in which cells were grown at 70°C in basal medium containing 3.6 mM ferrous sulfate during 10 days. Also free cells were estimated by protein determination using the Lowry methodology (11). Attached 1092

Microbiology Fundamentals

microorganisms were also estimated by protein assay using the modified Lowry method proposed by Karan et al., (12). In order to correlate the amount of proteins determined on the solids to number of cells, a calibration curve obtained from measurements of planktonic cells with both methods, direct counting and modified Lowry was used (see Figure 1). 3.

RESULTS AND DISCUSSION

3.1 Adherence of S. metallicus to chalcopyrite Figure 2 shows the evolution of the population of free bateria in solution during attachment of Sulfolobus metallicus to chalcopyrite. The attachment of S. metallicus onto chalcopyrite concentrate was found to reach a stationary value within the first 100 minutes of contacting the cells with mineral, when 68% of the initial cells were attached to the mineral. 8.0E +09

Total Cell/50 ml

Cell/ml

6 .0 E + 0 8

4 .0 E + 0 8

2 .0 E + 0 8

0 .0 E + 0 0

6.0E +09 4.0E +09 2.0E +09 0.0E +00

0

10

20

30

P r o te in ( m g /m l)

Figure 1. Correlation between cell number of S. metallicus and measured protein

0

50

100

150

T im e (m in)

Figure 2. Attachment of S. metallicus onto 1 g chalcopyrite concentrate

3.2 Bioleaching of chalcopyrite Figure 3 shows the rate of copper and iron dissolution, determined during the bioleaching of chalcopyrite with Sulfolobus metallicus. The total amount of dissolved iron, 8.86 mg, was much smaller than the total amount of dissolved copper, 134 mg. Ferrous/ferric determinations indicated that iron released during chalcopyrite dissolution was quantitatively oxidized to ferric iron from the start of the experiment (results not included), which indicated an effective microbial catalytic activity. Very likely, most of the iron released during chalcopyrite dissolution was eventually precipitated, which showed in the presence of large amounts of jarosite-like compounds either in the bioleaching solution or adhered to the flasks wall. The amount of ferric iron precipitated at each time corresponds to the difference between the effective and measured iron concentration, the theoretical iron dissolution expected according to the stoichiometry:

(1) CuFeS2 + 4Fe+3 → Cu+2 + 5Fe+2 + 2S The steady increase of this difference indicates that ferric iron precipitation occurred continuously since the start of the experiment. The evolution of the number of cells and biomass of planktonik cells of S. metallicus during the experiment is shown in Figure 4. Figure 5 shows the evolution of 1093

Microbiology Fundamentals

microorganisms attached to the solids presents in the flasks, solids which includes residual chalcopyrite together with the ferric iron precipitates which were simultaneously separated from the solution by filtration. These results show that planktonic cells gives only account of 30% of cell growth occurring during the process and, consequently, about 70% of bacterial growth occurs in the cells attached to the solids. The relative importance of these two populations is similar to the distribution of cells obtained during the initial attachment stage, shown in Figure 2. 6.E+08

6.E+07

4.E+08

4.E+07

2.E+08

2.E+07

0.E+00

0.E+00

1400 1000 800 600

Protein/ml

1200

Cell/ml

Cu and Total Fe (mg/l)

1600

400 200 0 0

100

200

0

300

T im e (h o u rs )

200

300

Tim e (hours)

Figure 3. Copper ■, total iron measured ∆ and effective total iron ▲ disolution in bioleaching of chalcopyrite concentrate

Figure 4. Cells of S. metallicus in bioleaching solution determined by direct count ■, MPN ● and protein concentration ▲ 160

4.E+10

140 120

3.E+10

Sulfur (mg)

Total Cell and Adhered Cells

100

2.E+10 1.E+10

100 80 60 40 20

0.E+00 0

100

200

300

Time (hours) Figure 5. Total ■ and adhered cells ▲ of S. metallicus in the bioleaching chalcopyrite concentrate

0 0

100

200

300

Time (hours)

Figure 6. Sulfur generated ▲ (curve a) and oxidized ■ (curve b) estimated from copper and acidity of solution

The cell population of planktonic and attached cells initially increases reaching a maximum at 72 and 120 hours, respectively. After this initial stage there is a rapid decay in the population of both free and attached microorganisms. In the case of planktonic cells, the decrease in cellular activity, determined by MPN, is paralelled by a decrease of the number of cells, determined by counting. This indicates that the decrease in cellular activity is not due to some type of inhibition, but rather to a process of cell death and/or cell breakage. This phenomenon can be linked to the sensitivity which these microorganisms show to the presence of solids, as it is widely known (3,13). Elimination of cells by incorporation into ferric iron precipitates can be 1094

Microbiology Fundamentals

ruled out as a main cause of cell number reduction, as this would show in a simultaneous increase in the biomass of attached bacteria, which was not observed. It is interesting to outline that the decrease in cell number and activity of planktonic cells is paralelled by a decrease in protein concentration. This indicates that the constituent proteins of thermophilic microorganisms, once the cell breaks, are not stable at the temperature and acidity at which bioleaching solutions operate. This fact was confirmed in a complementary experiment in which Sulfolobus metallicus digested cell proteins were maintained in a shaker in an iron-free basal solution at pH 1.8 and at 70°C. The protein concentration in the solution after 72 hours in these conditions, determined by Lowry, decreased 28% with respect to the initial value. The decay in protein concentration observed on solids indicates that in the attached cell population also a stage is reached in which the rate of cell death and/or breakage predominates over the rate of net cell growth. 3.3 Bacterial utilization of ferrous iron and sulfur Ferric iron was not initially added in the reported experiments, hence the scarce ferric iron present in the solutions resulted from the microbial oxidation of ferrous iron released during chalcopyrite dissolution. Bioleaching solutions presented a high Eh (data not shown), indicating a high ferrous iron oxidation activity of S. metallicus. The efficiency of this oxidative activity was also evidenced from the fact that MNP cell number determinations, which were based on the reponse of microorganisms to ferrous iron oxidation, gave similar values to those determined by direct cell counting. Attached cells, however, are exposed to two substrates, ferrous iron and surface elemental sulfur formed after reaction (1). It is interesting to assess the pattern of utilization of these two substrates during bacterial growth and see how this relates to the chalcopyrite bioleaching process. The pH of the bioleaching solution is the result of dinamic balance linked to a series of chemical reactions which consume or produce protons. Protons are consumed during the oxidation of ferrous iron which occurrs according to:

(2) Fe+2 + 1/2O2 + 2H+ → Fe+3 + H2O On the other hand, protons are produced through: a) hydrolysis of the Fe(III) with jarosite formation according to the reaction: (3) M+ + 3Fe+3 + 6H2O + 2(SO4)-2 → MFe3(SO4)2(OH)6 + 6H+ + + + + Where M is either K , NH4 or H3O ; and through b) oxidation of sulfur to sulfate according to the reaction: (4) 2S° + 3O2 + 2H2O → 2H2SO4 In order to assess the amount of elemental sulfur oxidized by Sulfolobus metallicus to sulfate, the acidity resulting from this reaction was evaluated at different times during chalcopyrite bioleaching with the expression: ∆4 = Ct - Ci - ∆3 + ∆2

(5)

where ∆4 is the acid generated by reaction 4, Ct is the acid concentration determined in the solution from pH measurement, Ci is the initial acid concentration, ∆3 is the acid produced in reaction 3, and ∆2 the acid consumed in reaction 2. The amount of ferrous released at each time was calculated from the amount of copper dissolved from chalcopyrite, assuming that this follows the stoichiometry of reaction (1). Ferrous iron was assumed to be quantitatively converted into ferric according to reaction (2). The amount of ferric iron precipitated was calculated as the difference between the theoretical amount of ferrous 1095

Microbiology Fundamentals

iron dissolved and the iron experimentally determined in solution (Figure 3). The acid generation associated to ferric iron precipitation was calculated assuming formation of potassium jarosites which, according to XRD analysis, is the compound we have found that predominate in concentrates bioleached with thermophiles. Results of this calculation are shown in Figure 6. Curve a represents the amount of elemental sulfur generated during chalcopyrite dissolution according to reaction (1), and was calculated from the amount of copper dissolved at different leaching times. Curve b represents the amount of elemental sulfur oxidized to sulfate calculated from reaction (4) using the ∆4 values as generated acid. Comparison of data in Figure 6 with the evolution of cell growth in Figures 4 and 5 indicate that the fast cell growth observed during the initial 72 hours in both the planktonic and attached microorganisms, is mainly based on the oxidation of ferrous iron. The growth is faster on the solids as the attached initial population is larger. In addition, cells attached to chalcopyrite surface should have access to larger concentrations of ferrous ions which is released in that vicinity. The oxidation of elemental sulfur becomes significant only after the first 72 hours when already 50 mg of this compound have been accumulated on the chalcopyrite surface. At this stage the growth of the attached population continues fast for another 50 hours while the planktonic population started to decrease. It is very likely that the continued growth trend observed in the attached microorganisms is partially linked to triggering of sulfur oxidation, as these microorganisms are the only ones with can have direct access to that alternative solid substrate. On the other hand, the decrease in the planktonic population after the first 72 hours can be due to depletion of ferrous iron in the solution bulk. This can be related to the large population of attached microorganisms reached at this stage, which can consume most of the ferrous iron released from the sulfide. Studies done by several authors about the bioleaching of pyrite and chalcopyrite with thermophilic microorganisms show differences on the cells adsorption to the mineral; in some cases, in spite of the fact that the cellular attachment was not detected by microscope, they showed that pyrite leaching required the direct contact of Acidianus brierley with the mineral (14). Nemati and Harrison (15), working with S. metallicus and pyrite, showed a low cell adsorption (5-10%) and considered the concentration of the cells in the liquid phase as indicative of total microbial growth in the reactor. On the other hand, Konishi et al., (16) showed a high attachment of cells to chalcopyrite and found that the chalcopyrite leaching with A. brierley at 65°C takes place with a direct attack by adsorbed cells on the mineral surface, the ferric leaching being insignificant. The results presented in this work show that most of the population of S. metallicus present in a bioreactor remains attached to the solids, where they grow out of ferrous iron and sulfur oxidation. Any attempt to assess in depth the performance of a bioleaching reactor should include an adequate monitoring of this population. 4.

CONCLUSIONS Experimental results demonstrated that the growth of attached Sulfolobus metallicus can be efficiently monitored using the modified Lowry method of Karan et al.,(16). During bioleaching of chalcopyrite about 68% of the Sulfolobus metallicus present in the flask grows attached to chalcopyrite particles, while the remaining 32% grows in the solution. The percentage associated to attached growth coincides with the fraction of microorganisms which initially attaches to chalcopyrite after addition of the inoculum. Cell growth during the first 72 hours of both attached and free microorganisms is mainly based on the oxidation of ferrous iron. Bacterial oxidation of sulfur, which is 1096

Microbiology Fundamentals

conducted only by attached microorganisms, becomes relevant only after 72 hours of leaching. After 120 hours of leaching there is a a net reduction of both the planktonic and attached population which evidences an important process of cell death and breakage. ACKNOWLEDGEMENTS This work was funded by BHP Billiton and Conicyt Chile under Fondef Project D 00 I 1050. REFERENCES

1. P.R. Norris, Biomining: Theory, Microbes and Industrial Processes, D.E. Rawlings (ed), Landes, Bioscience, Austin, TX USA. (1997) 281. 2. P.R. Norris, L. Parrot and R.M. Marsh, Workshop on Biotechnology for the Mining, Metal-refining and Fossil Fuel Processing Industries, Biotechnol. Bioeng., Symposium Nª16, H.L. Ehrlich and D.S Holmes (eds), Wiley, New York (1986) 253. 3. D.A. Clark and P. Norris, Microbiology 142 (1996) 785. 4. P. d’Hugues, D. Morin and S. Foucher, Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, Part A, V.S.T. Ciminelli and O. Garcia (eds), Elsevier, Amsterdam (2001) 75. 5. M.L. Blázquez, A. Alvarez, A. Ballester, F. González and J.A. Muñoz, Biohydrometallurgy and the Environment toward the mining of the 21st Century, A. Ballester and R. Amils (eds.) Elsevier, Amsterdam (1999) 137. 6. Y. Rodriguez, A. Ballester, M. L. Blázquez, F. González and J.A. Muñoz, Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, Part A, V.S.T. Ciminelli and O. Garcia (eds), Elsevier, Amsterdam (2001) 125. 7. R. M. Marsh, P.R. Norris and N. W. Le Roux, Progress in Biohydrometallurgy, G. Rossi and A.E. Torma (eds.) Cagliari, Italy (1983) 71. 8. D.W. Shivvers and T. D. Brock, J. Bacteriol., 114 (1973) 706. 9. M. K. Muir and T. Anderson, Metall. Trans. 8B (1977) 517. 10. L. Herrera, P. Ruiz, J.C. Aguillón and A. Fehrmann, J. Chem. Tech. Biotechnol 44 (1989) 171. 11. O.H. Lowry, N.J. Rosenborough, A.L. Farr and R.J. Randall, J. Biol. Chem., 193 (1951) 265. 12. G. Karan, K.A. Natarajan and J.M. Modak, Hydrometallurgy 42 (1996) 169. 13. B. Escobar, J.M. Casas, J. Mamani and R. Badilla-Ohlbaum Biohydrometallurgical Technologies, A. E. Torma, J.E. Wey and V.I. Lakshmanan (eds.) The Minerals, Metals & Materials Society (1993) 195. 14. L. Larsson, G. Olsson, O. Holst and H. T. Karlsson, Biotechnol. Lett. 15 (1993) 1, 99. 15. M. Nemati and S.T.L. Harrison, Biohydrometallurgy and the Environment toward the mining of the 21st Century, Part A, A. Ballester and R. Amils (eds) Elsevier, Amsterdam (1999) 473. 16. Y. Konishi, M. Tokushige and S. Asai, Biohydrometallurgy and the Environment toward the mining of the 21st Century, Part A, A. Ballester and R. Amils (eds) Elsevier, Amsterdam (1999) 367.

1097

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Experimental and modeling studies on inhibition effect of solution conditions on activity of Acidithiobacillus ferrooxidans during biooxidation of mixed sulphidic concentrates M.N. Chandraprabhaa, Jayant M. Modakb and K.A. Natarajana a

b

Department of Metallurgy, Indian Institute of Science, India Department of Chemical Engineering, Indian Institute of Science, India

Abstract In this paper, we report experimental and modeling results aimed at characterizing the effect of solution conditions on the growth and ferrous iron oxidation ability of Acidithiobacillus ferrooxidans. The conditions we have chosen pertain to the bioleaching of mixed sulphidic (pyrite and arsenopyrite) gold ore concentrates resulting in the accumulation of ferric and arsenic ions in the solution. Apart from the dissolved metal species, the pH and CO2 concentrations in the solution are also important parameters in bioleaching processes. The iron oxidizing efficiency of the strain increases with increasing carbon dioxide supplementation of the sparged air upto 1% (V/V), beyond which the bacterial activity reduces. Maintaining lower growth pH (1.6) reduced the formation of precipitates to a large extent. Ferric and arsenic ions are found to be toxic to the bacterium and strains tolerant to higher concentrations of these metal ions could be obtained by adaptation. The growth kinetics of the bacterium has been modeled incorporating the ferric and arsenic inhibition. The adaptation of Acidithiobacillus ferrooxidans to higher metal concentrations is quantified by gradual reduction in inhibition constants. The simulation results were found to match well with the experimental data obtained. The applicability of the inhibition constants obtained to the batch bioleaching model of a mixed sulphidic concentrate has been analysed.

Keywords: Acidithiobacillus ferrooxidans, biooxidation, inhibition, adaptation, modeling 1.

INTRODUCTION Biooxidation rate depends to a large extent on the efficiency of the bacterial strain used. The regeneration of ferric iron by bacterial oxidation of ferrous sulphate in the solution phase is an essential step in the biooxidation process. The oxidation rate is, therefore, directly related to the activity of the bacteria in solution phase, which in turn is determined and driven by the solution conditions. Bioleaching experiments results in the formation of ferric and arsenic hydroxides that precipitate at pH 2. One way of minimizing the formation of these precipitates is by maintaining a low process pH. Studies by several workers (Jose et al, 1997; Kelly, 1978; 1099

Microbiology Fundamentals

Nagpal et al, 1992) have clearly indicated the reduction in precipitate formation by lowering pH. Since the bacterium employed has an optimum pH of 2, it has to be suitably adapted to the low pH required in order to retain its optimal activity. Studies done by Toxiarchou et al, (1994) on the extent of arsenopyrite and pyrite oxidation at various pH values showed that although the selective oxidation of arsenopyrite is enhanced at low pH, pH values lower than 1.6 adversely affects the overall oxidation kinetics since under these conditions the oxidation of pyrite is hindered. Hence a pH of 1.6 would be optimal for reducing the precipitate formation without hindering the oxidation kinetics. Oxygen and carbon dioxide are essential nutrients for bacterially catalyzed mineral oxidation and bacterial growth. Reduction in oxygen concentration in the aqueous phase below optimum levels leads to reduced growth rates. Similarly, reduction of aqueous phase carbon dioxide also sharply reduces the bacterial growth. Hence, the effect of CO2 enrichment on the growth of bacteria needs to be studied. Various toxic substances, often inherently present in the biooxidation of a sulphide mineral, have been found to inhibit the oxidation ability of bacteria (Bailey and Hansford, 1993). During biooxidation process, the concentration of the released metal ions in the solution continually increases and hence there will be build-up of ferric and arsenic ions. Both ferric and arsenic ions are toxic to the bacterial activity and hence the growth of bacteria in solution phase in presence of these metal ions needs to be examined. This detrimental effect can be overcome by adapting the bacteria to higher concentrations of these metal ions (Modak and Natarajan, 1995). The objective of our present work was to analyse the effect of solution conditions on the activity of the bacteria, model the growth kinetics of the bacterium incorporating the inhibition effects of the toxic metal ions present in the solution and to analyse the applicability of this to the batch bioleaching model of a mixed sulphidic concentrate. 2.

EXPERIMENTAL

2.1 Materials The bacterial culture used was a pure strain of Acidithiobacillus ferrooxidans that was isolated from Hutti Gold Mines (HGML) and is referred to as TfH6. The purity was ascertained by the procedure outlined by Karavaiko (1988). The bacteria were cultured in sterile 9K medium developed by Silverman and Lundgren (1959). The refractory sulphidic concentrate obtained from HGML was in finely ground form with an average particle diameter of 70 µm. The chemical composition was 31.95 wt% Fe, 29.25 wt% S and 22.5 g/MT gold. 2.2 Methods 2.2.1 Growth of bacteria The growth experiments were carried out by inoculating 10%(V/V) of active inoculum to sterilized 9K medium. The flasks were incubated at 30°C on a rotary shaker at 240 rpm. The ferrous and the total iron concentration were monitored at regular intervals by spectroscopic method. The solution cell number was determined by microscopic counting with Petroff-Hausser counter using a Leitz phase contrast microscope (Laborlux K Wild MPS12). pH changes were also monitored simultaneously using a Systronics digital pH meter.

1100

Microbiology Fundamentals

2.2.2 Effect of carbon dioxide on growth The effect of carbon dioxide availability on the activity of bacteria was monitored by measuring the iron oxidization rate of bacteria at different levels of carbon dioxide supplementation. These experiments were carried out in 17L bioreactors with the operating volume of 6L. The aeration rate and the carbon dioxide flow rate were monitored using suitably calibrated rotameters. 2.2.3 Growth inhibition and adaptation studies The inhibition effect of ferric iron and arsenic on the growth of bacteria was studied by adding ferric sulphate or sodium arsenite at different concentrations to the flasks containing 9K medium solution, prior to inoculation. The iron oxidation rate of these cultures were monitored and compared with that of the control without inhibition. Adaptation to lower pH was done by serial subculturing of the bacterial strain in 9K medium maintained at the required pH. Adaptation of the bacteria to the toxic ferric and arsenic ions was carried out by serial subculturing of bacteria in 9K medium in presence of externally added ferric sulphate and sodium arsenite as the sources of ferric and arsenic ions respectively. 2.2.4 Batch bioleaching experiments The bioleaching experiments were performed in batch mode in well-agitated and aerated 17L bioreactor with an operating volume of 6L. The reactor was aerated at the rate of 0.5 v/v/min with 1% carbon dioxide enrichment. The experiments were started with a low pulp density of 1%(W/V), which was then increased in stepwise with periodic addition of fresh concentrate when the concentrate in the reactor was completely oxidised. Complete oxidation was considered to be achieved when the total iron in the solution approached the theoretical stoichiometric value estimated. 3.

RESULTS AND DISCUSSION

3.1 Growth of Acidithiobacillus ferrooxidans The growth of a typical culture of Acidithiobacillus ferrooxidans (TfH6) on ferrous iron with time under optimum conditions of temperature (30°C) and pH (2.0) is shown in Figure 1a. No significant increase in cell density was observed during the initial lag period of 20hrs, beyond which the growth followed exponential kinetics. The growth remained in log phase till the 50th hr after which the cells entered the stationary phase. The maximum cell number was reached in 48-50 hrs. 3.2 Effect of pH Lowering of pH did not have any significant effect on the activity of the bacterial strain studied and just after one subculturing the activity of the adapted strain matched with that of the wild strain as seen from Figure 1b. Comparing the growth curves at pH 2 and pH 1.6 (Figures 1a and 1b), we find that at the end of exponential phase when all the ferrous iron is oxidized, the ferric iron concentration in the flask maintained at low pH is 8.9 g/L, which is higher than that observed at pH 2 (6.3 g/L). This clearly indicates that the precipitation of ferric hydroxide found at pH 2 is reduced to a large extent by lowering the pH.

1101

Microbiology Fundamentals

(a)

10

2+

Fe 3+ Fe

8

Iron concentration (g/L)

Iron concentration (g/L)

10

6 4 2 0

(b)

8 6

2+

Fe 3+ Fe

4 2 0

0

10

20

30

40

50

60

70

0

10

20

Time (h)

30

40

50

60

70

Time (h)

Figure 1. Growth of Acidithiobacillus ferrooxidans at (a) pH 2 and (b) pH 1.6 3.2 Effect of CO2 on growth of organism In order to study the effect of CO2 on the growth of bacteria, the growth experiments were conducted in 17L bucket bioreactor at different levels of CO2 supplementation (vol CO2/vol air). The results obtained are plotted in Figure 2. It is clear from the figure that the supplementation of CO2 to the sparged air resulted in enhancement of growth rate. In the absence of CO2 supplementation, the bacteria were able to oxidize the ferrous iron in 9K medium in 52 hours. With 0.5% CO2 supplementation, the time taken for complete growth was just 42 hours, which reduced to 38 hours at 1% CO2 supplementation. The cell yield also showed slight increase with increase in CO2 supplementation. However beyond 1% CO2 supplementation, there was no much benefit since the oxidation rate was less at 2% CO2 supplementation as compared to 1% CO2 supplementation. 25 (a)

8

Cell number X 10 (cells/L)

0% CO2 0.5% CO2 1% CO2 2% CO2

6 4

10

Ferrous iron concentration (g/L)

10

2 0 0

10

20

30

40

50

60

Time (h)

(a) Ferrous iron concentration

70

(b)

20 15 10

0% CO2 0.5% CO2 1% CO2 2% CO2

5 0 0

10

20

30

40

50

60

Time (h)

(b) Cell density

Figure 2: Effect of CO2 supplementation on growth of Acidithiobacillus ferrooxidans 1102

70

Microbiology Fundamentals

Nagpal et al., (1992) have shown that bacteria grown on a sulphide ore concentrate utilize oxygen and carbon dioxide in a molar ratio of approximately 20 mol oxygen per mole CO2. This suggests that air supplemented with 1% CO2 would provide the appropriate molar ratio of the two nutrients. During their tests at 16% solids, carbon dioxide concentrations in the liquor in excess of 10mg/L were found to be inhibitory to bacterial growth. The obtained results are thereby, in good agreement with that reported. 3.4 Effect of ferric and arsenic inhibition In order to study the effect of Fe3+ on ferrous oxidation, the growth of TfH6 was monitored with varying concentrations of Fe3+ added initially to the nutrient medium. The results of the experiments with 2, 5 and 8 g/L of initial Fe3+ is presented in Figure 3a. While 2 g/L initial Fe3+ had no significant effect on growth, an extended lag phase was observed with 5 g/L and 8 g/L initial Fe3+ concentration. This is accompanied by reduced rates of ferrous oxidation and cell growth. The arsenic ions coming to the solution due to solubilization of arsenopyrite is in the form of arsenite [As3+] and arsenate [As5+] ions. Since arsenate ions are less toxic compared to arsenite ions, the inhibition effect of only arsenite ions is considered in our studies. Figure 3b, shows the cell growth and ferrous iron oxidation by TfH6 in presence of different initial concentrations of arsenite ions. From the figure we can observe that the wild strain is naturally tolerant upto 400 ppm of arsenite concentration. Further increase in arsenite concentration resulted in extended lag phase and decrease in cell yield. The extended lag phase was significant when the initial arsenite concentration maintained was 1000 ppm and the wild strain took more than 58hrs to oxidize 9 g/L ferrous. It was observed that the total cell yield was less at arsenite concentrations of 500 and 1000 ppm. 10 (a)

8

3+

Fe concentration 0 g/L 2 g/L 5 g/L 8 g/L

6 4 2 0 0

10

20

30

40

Time (h)

50

60

70

Ferrous iron concentration (g/L)

Ferrous iron concentration (g/L)

10

(b)

8 6 3+

As concentration 4 0 ppm 100 ppm 200 ppm 2 500 ppm 1000 ppm

0

0

10

20

30

40

50

60

70

Time (h)

Figure 3: Effect of (a) ferric and (b) arsenite ions on growth of Acidithiobacillus ferrooxidans 3.5 Bacterial adaptation studies From Figures 3, it is clear that the growth of TfH6 is inhibited in presence of high initial ferric and arsenic concentration. However, repeated subculturing in presence of these toxic ions results in reduced lag period and increased growth rate due to its adaptation. The bacterial strain is said to be adapted when the cell growth and iron oxidation rate of the strain in presence of the externally added toxic metal ion is 1103

Microbiology Fundamentals

comparable with that of the control, which is the growth rate of the strain in the absence of any inhibition. The results obtained in presence of 8 g/L Fe3+ and 1000 ppm As3+ is shown in Figure 4. The growth rate of the wild bacterial strain in presence of the externally added toxic metal ion is represented as its natural tolerance. When grown in presence of 8g/L Fe3+ growth rate after two subcultures matched with that of the wild strain, thereby indicating that the strain is now adapted to 8 g/L Fe3+. Hence, with just two subcultures it was possible to obtain a strain tolerant to 8 g/L Fe3+. As evident from Figure 4b, it was possible to obtain a strain tolerant to 1000 ppm As3+ by subculturing the wild strain thrice in presence of 1000 ppm As3+. 25

25

(b)

Cell number X 10 cells/L

20 15

10

10

Cell number X 10 cells/L

(a)

10 5

Natural tolerance 1 subculture 2 subculture No inhibition (control)

0 0

10

20

30

40

Time (h)

50

60

70

20 15 10 Natural tolerance 1 subculture 2 subculture 3 subculture No inhibition

5 0 0

10

20

30

40

50

60

70

Time (h)

Figure 4. Adaptation of Acidithiobacillus ferrooxidans to (a) 8 g/L ferric; (b) 1000 ppm arsenite 3.6 Modeling of bacterial growth The model for the growth of bacteria on ferrous iron should be able to predict the rates at which ferrous is being oxidised under the changing conditions of ferrous and ferric concentrations and the cell densities. The growth rate of bacteria in liquid is assumed to be described adequately by Monod's model (Boon et al, 1999; Jensen and Webb, 1995) with ferrous iron as the substrate, that is, the rate of substrate utilization is given by:

− µ m Fe 2+ dFe 2+ = dt Y ( K S + Fe 2+ )

(1)

where µm is the maximum specific growth rate in hr-1 and KS is the saturation constant for ferrous iron in g/L respectively. Y is the yield coefficient given by, dX dFe 2+ = −Y (2) dt dt where X represents cell density in solution in cells/L. Both ferric and arsenic ions inhibit the growth of Acidithiobacillus ferrooxidans competitively without affecting the maximum specific growth rate. The rate is thus given by the below equation (Bailey and Olis, 1986), 1104

Microbiology Fundamentals

− µ m Fe 2+ dFe 2+ = dt Y ( K S (1 + K I Fe3+ + K a As 3+ ) + Fe2+ )

(3)

where KI is the inhibition constant for ferric iron and Ka for arsenic ion in L/g. Experimental data of growth in absence of any inhibition was used to fit the parameters of the above equation and the simulation result obtained is shown in Figure 5. The values of µm and KS obtained are 0.07 hr-1 and 1.6 g/L respectively.

(a) Ferrous iron oxidation

(b) Cell growth

Figure 5. Growth of Acidithiobacillus ferrooxidans at 30°C and pH 1.5

The experimental data obtained at different levels of CO2 supplementation were used in fitting the parameters of the Monod equation and it was found that the supplementation of CO2 affects the maximum specific growth rate. The µm obtained at different levels of supplementation are given in Table 1, from which it is clear that the optimum level of supplementation is 1% V/V of CO2. The growth rate data obtained in presence of 8 g/L Fe3+ and 1000 ppm As3+ was used to obtain the inhibition parameters in equation 3. Since both Fe3+ and As3+ ions inhibit competitively, µm remains unaffected. A comparison of the model parameters obtained with that reported in literature is given in Table 2. As see from the table, the values obtained are in close agreement with the reported values (Jensen and Webb, 1995; Lizama and Suzuki, 1989). Table 1. Variation in µm with CO2 supplementation

Table 2. Model parameters of bacterial growth

Percent CO2 supplementation 0 0.5 1 2

µm, hr-1

Parameter

0.07 0.0825 0.0925 0.0785

µm (hr-1) Ks (g/L) Y (cells/mol Fe2+) KI (L/g) Ka (L/g)

Estimated value 0.09 1.6 1.36x1012 0.14 1.15

Reported value 0.05 - 1.3 1-2 9x1013 - 5x1014 0.5 - 2.3 0.45 - 0.6

The comparison between model simulation using the parameters listed in Table 2 and the experimental data is shown in Figure 6. The bacteria, on continuous exposure, develop tolerance to ferric and arsenic ions as discussed earlier. The effect of bacterial adaptation 1105

Microbiology Fundamentals

to ferric and arsenic ions on the inhibition constants KI and Ka is shown in Figure 6. From the figure we observe that with repeated subculture, the value of the constants approaches zero.

Figure 6. Variation of inhibition constants with adaptation of Acidithiobacillus ferrooxidans to (a) 8 g/L ferric; (b) 1000 ppm arsenic 3.7 Application to batch leaching model A complete batch bioleaching model for a mixed sulphidic concentrate was developed earlier by considering the various subprocesses independently to obtain the model parameters (Chandraprabha et al, 2001). In this model growth of bacteria was described by equation 1, neglecting the inhibition effect of ferric and arsenic ions coming into the solution. Here the effect of the inhibition constants on the model prediction of the batch biooxidation of 1% pd pyrite-arsenopyrite concentrate was tested and the results obtained are shown in Figure 7. The bacterial growth in solution was described by equation 3 and the parameters listed in Table 2 were used for simulation. As seen from the figure, the model overpredicts the ferric iron concentration when the inhibition effect of ferric and arsenic ions is neglected (dotted lines). The presence of ferric and arsenic ions in the solution thereby inhibits the bacterial activity. After 5 days of leaching, when the concentrate was almost completely oxidized, the pulp density was increased from 1% to 2% by addition of fresh concentrate. Simulation done at 2% pd also showed similar behaviour (data not shown). The slurry was maintained at 2% pd for 5 days after which it was increased to 3% by addition of fresh concentrate. The results for 3% pulp density are shown in Figure 7b. During the course of biooxidation the bacteria are subjected to increasing concentrations of the toxic ferric and arsenic ions coming from the concentrate and grow in presence of these stressed conditions. When solid pulp density reached 3%, the bacteria were exposed to total of about 240 h. As revealed by our adaptation studies, the bacteria attain tolerance at 3 subcultures (approximately 200 h). Hence the later generations of bacteria are adapted strains that are tolerant to these toxic ions. This is clear from Figure 7b, which shows the simulation results obtained at 3% pd concentrate with KI and Ka values set to zero. Here we find that the prediction obtained by neglecting the inhibition effects of ferric and arsenic fits well with the experimental data points.

1106

Microbiology Fundamentals

Figure 7. Simulation of batch bioleaching at (a) 1% pd and (b) 3% pd concentrate. Solid lines represent model curves with ferric and arsenic inhibition and dashed lines without any inhibition. Symbols represent experimental data points 4.

CONCLUSIONS The effect of solution conditions on the activity of Acidithiobacillus ferrooxidans is reported and the following conclusions can be arrived at: Lowering of process pH resulted in reduction of the precipitate formation, thereby enhancing bacterial oxidation rates. Supply of air supplemented with CO2 enhanced the growth kinetics of the strain used. The iron oxidation rate increased with increase in CO2 supplementation upto 1% (V/V), beyond which the cell yield decreased. Presence of both ferric and arsenic ions inhibited the bacterial growth. Serial subculturing in presence of these metal ions resulted in adapted strains that were tolerant to higher concentrations of the metal ions. The growth kinetics of bacteria was modeled incorporating the inhibition constants obtained for ferric and arsenic inhibition and it was found that with repeated subculturing the value of the inhibition constant approached zero which clearly indicated the adaptation of bacteria to the toxic metal ions due to continuous exposure. The model developed for the batch bioleaching of pyrite-arsenopyrite concentrate was found to overpredict the ferric iron concentration in the solution at 1% pd when the effects of ferric and arsenic inhibition were neglected. This was overcome by incorporating the inhibition constants in the growth equation of bacteria. The predictions obtained at later stages at a higher pulp density were found to fit well with the experimental data even when the inhibition effects were neglected. This was due to the adaptation of the bacteria to the released metal ions in solution that was achieved due to their continuous exposure to these ions. REFERENCES

1. Bailey, J.E. and Hansford, G.S., 1993. Factors affecting bio-oxidation of sulphide minerals at high concentrations of solids: A Review. Biotechnol. Bioengg., 42, 11641174. 1107

Microbiology Fundamentals

2. Bailey, J.E. and Olis, D.F., 1986. Biochemical engineering fundamentals, second ed. McGraw Hill, Singapore. 3. Boon, M., Ras, C., Heijnen, J. J., 1999. The ferrous iron oxidation kinetics of Thiobacillus ferrooxidans in batch cultures. Appl. Microbiol. Biotechnol., 51, 813819. 4. Chandraprabha, M.N., Modak, J.M., Raichur, A.M. and Natarajan, K.A., 2001. Modeling of biooxidation of gold bearing pyrite-arsenopyrite concentrates by Acidithiobacilli ferrooxidans. In Ciminelli, V.S.T. and Garcia, O., (Eds.), Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, Process Metallurgy 11A, Elsevier, Amsterdam, pp.43-54. 5. Jensen, A.B. and Webb, C., 1995. Ferrous sulphate oxidation using Thiobacillus ferrooxidans: A review. Process Biochem., 30, 225-236. 6. Jose, M.G., Domingo, C. and Barrie, J., 1997. Comparison of the effects of temperature and pH on iron oxidation and survival of Thiobacillus ferrooxidans (type strain) and Leptospirillum ferrooxidans. Biotechnol. Bioengg., 42, 581-589. 7. Karavaiko, G.I., 1988 Methods of isolation, evaluation and studying of microorganisms. In Karavaiko, G.I., Rossi, G., Agate, A.D., Groudev, S.N., Avakyan, Z.A., (Eds.), Biotechnology of Metals – Manual, Center for International Projects GKNT, Moscow, pp.47-79. 8. Kelly, D.P. and Jones, C.A., 1978. Factors affecting metabolism and ferrous iron oxidation in suspensions and batch cultures of Thiobacillus ferrooxidans. In: Murr, L.E., Torma, A.E. and Brierly, J.A., (Eds), Metallurgical applications of bacterial leaching and related microbiological phenomenon, Academic, New York, pp. 19-44. 9. Lizama, H.M. and Suzuki, I., 1989. Synergitic competitive inhibition of ferrous iron oxidation by Thiobacillus ferrooxidans by increasing concentrations of ferric iron and cells. Appl. Environ. Microbiol., 55, 2588-2591. 10. Modak, J.M. and Natarajan, K.A., 1995. Development of special strains of Thiobacillus ferrooxidans for enhanced bioleaching of sulphide minerals. In: Jerez, C.A., Vargas, T., Toledo, H. and Wiertz, J.V., (Eds), Biohydrometallurgical processing, Vol I, University of Chile, Santiago, Chile, pp. 34-46. 11. Nagpal, S. and Dahlstrom, D., 1992. Effect of carbon dioxide concentration on the bioleaching of a pyrite-arsenopyrite ore concentrate. Biotechnol. Bioengg., 41, 459464. 12. Silverman, M.P. and Lundgren, D.G., 1959. Studies on chemoautotrophic iron bacterium Thiobacillus ferrooxidans. J. Bacteriol., 77, 642-646. 13. Taxiarchou, M., Adam, K. and Kontopoulos, A., 1994. Bacterial oxidation conditions for gold extraction from Olympias refractory arsenical pyrite concentrate. Hydrometallurgy 36, 169-185.

1108

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Ferrous ion oxidation by an activated carbon cloth modified with Acidithiobacillus ferrooxidans F. de J. Cerino-Cordova, J.P. Magnin, N. Gondrexon, P. Ozil Laboratoire d’Electrochimie et de Physico-chimie des Matériaux et des Interfaces, UMR 5631 INPG-CNRS- UJF, Equipe Génie des Procédés, ENSEEG BP75, 38402 St Martin d’Hères, France Abstract The formation of Acidithiobacillus ferrooxidans biofilm on an Activated Carbon Cloth (ACC) was performed during subcultures of a biologically modified ACC in sterile ferrous iron medium. Two ways of modification were investigated: the classical culture of free bacteria on ACC, an enhanced way using a preliminary sorption of concentrated bacteria on ACC surface. In the absence of bacteria, the ACC exhibited a low ferrous-iron oxidation capacity, 23 mg Fe2+ L-1/h. The presence of free (initial concentration 0.07 mg protein L-1) or immobilised bacteria on ACC WWP3 and WKL20 (4-8 mg protein/g carbon) increased the ferrous ion oxidation rate respectively up to 60-90 and 600 mg Fe2+ L-1/h. The preliminary sorption step allowed to decrease the time required for biofilm stabilisation of about 30 percent, depending on the initial concentration of bacteria immobilised on ACC. No difference was observed concerning the biofilm formation and activity for the two types of ACC, excepted the final fragility of bacterial modified KL20 ACC.

Keywords: Acidithiobacillus ferrooxidans, biofilm, activated carbon cloth 1.

INTRODUCTION Acidithiobacillus ferrooxidans is a chemoautotrophic acidophilic microorganism, which gets the energy indispensable for its growth from oxidation of ions ferrous, elemental sulphur and reduced sulphur compounds including metal sulphides [1]. This ability has been used from a long time in different processes such as bioleaching [2-6], H2S removal [7] and acid mine drainage treatment [8-9]. It is well known that the free growth of A. ferrooxidans is characterised by a low rate, thus requiring long residence times and consequently large reactor volumes resulting in high investment and operating costs [10]. Therefore many research works have been focused on increasing ferrous iron oxidation rate using immobilised biomass. Various configurations were investigated using different bioreactor types (packed bed [9-13], fluidised bed [13-15] trickle bed reactor [16] and rotating biological contactor [17,18,19]) with different non-metabolisable supports (glass beads [13-14]; ion exchange particles

1109

Microbiology Fundamentals

[13-14], diatomeous earth [20-21], polyurethane foam [22], silica gel [23], sand [8], nickel alloy fibre [24], and activated carbon [13-14, 23]). Activated carbon exists under various forms such as granule, fibre or cloth. Activated Carbon Clothes (ACC) offer faster kinetics than Granular Activated Carbon to treat micropolluants [26] and to sorb heavy metals [27]. Our previous studies were devoted to fixing A. ferrooxidans on Granular Activated Carbon in order to use this biological modified support as the biocathode of a bioelectroreactor [28]. Activated Carbon Cloth showed an improved biomass capacity for A. ferrooxidans immobilisation (respectively 25 mg protein /g Carbon for WWP3 and 14,9 mg protein /g Carbon for WKL20) [unpublished data]. Therefore the present paper focuses on developing a A. ferrooxidans biofilm on ACC and characterising it by its iron-oxidising activity and from Scanning Electron Microscopy observations. 2

MATERIALS AND METHODS

2.1 Culture and medium The bacterial strain Acidithiobacillis ferrooxidans DSM 583 was routinely maintained in 9K liquid medium (33 g.L-1 FeSO4, 0.4 g L-1 MgSO4.7H2O, 0.4 g.L-1 K2HPO4, 0.4 g.L-1 (NH4)2SO4) at pH=1.4 (H2SO4) After bacterial growth at 30°C in a 5-L bioreactor, the biomass was concentrated by cross-flow filtration before being harvested at 5000 g for 10 min. The resulting biomass was washed four times with H2SO4 (pH 1.4) and harvested at 1000 rpm during 10 min to remove jarosite precipitate. Finally, the biomass was suspended again in H2SO4 (pH 1.4) and kept at 5°C during 24 h before utilisation. 2.2 Activated Carbon Cloths The physical characteristics of ACC WWP3 and WKL20 (purchased from Actitex, France) are recapitulated in table 1. Table 1. Industrial characteristics of Activated Carbon Cloths WWP3 and WKL20 Characteristics Specific Surface (MEB) Weight Mechanical resistance Median Pore diameter Weaving Pore < 20 Å Pore < 500 Å Microporosity Thickness

WWP3 1000-1300 m2/g 110-130 g/m2 5.0 DaN/cm 8.4 Å Woven fabric 0.52 cc/g 0.65 cc/g >80% 0.5 m

WKL20 1000 m2/g 100 g/m2 7Å Knitted fabric 0.31 cc/g 0.33 cc/g >90% 0.5 m

2.3 Analytical procedures The concentration ratio of ferrous ion to total (ferrous + ferric) ions was measured by the o-phenantroline method [29]. The bacterial concentration in solution correlated with protein concentration was estimated by the Lowry method [30].

1110

Microbiology Fundamentals

2.4 Modification of ACC with bacterial biomass The ACC biological modification was performed via preliminary adsorption batch experiments at 30°C and 100 rpm. A piece of ACC (0.1345 g) was introduced in a 0.1 L flask containing 0.02 L H2SO4 (pH=1.4) and 0.2 g protein L-1. After 2h, this piece was rinsed with sterile demineralised water to eliminate non-attached bacteria. It was next used in the ferrous ion oxidation experiments. 2.5 Bacterial subcultures in the presence of ACC cloths The A. ferrooxidans bacterial cultures with modified or unmodified ACC (WWP3 or WKL20) were realised in 0.1 L flask containing 0.2 L of culture medium at 30°C under rotary shaker agitation (150 rpm). In the case of non-biologically modified ACC, biomass (0.07 mg prot. L-1) was introduced in the medium. During incubation, samples were regularly collected and analysed to determine both ferrous ion oxidation activity (ratio Fe2+/ (Fe2+ + Fe3+)) and protein concentration in the medium. When 90% of ferrous ion concentration were oxidised (corresponding to a time noted T90), the ACC samples were picked up under sterile condition, rinsed with sterile demineralised and introduced into a 20-ml sterile culture medium. This procedure was repeated several times until reaching constant T90 and bacterial concentrations in solution. 2.6 Scanning Electronic Microscopic ACC samples were rinsed with sterile demineralised water and put into a Nacacodylate solution (4% cacodilyc acid, sodium salt trihydrate 98%) dissolved in a 25% glutaraldehyde solution (pH 7). After 4-h incubation, the samples were transferred firstly into a water solution for 1h and then successively into 25%, 50%, 75% and 100% xylene/acetone solutions for 10 minutes each. Finally, the samples were air-dried and prepared before SEM observation (Jeol 6400). 3.

RESULTS AND DISCUSSION The WKL20 and WWP3 ACC under study exhibited a capacity to oxidise ferrous ion in solution as shown in Fig. 1. Without ACC and in an uninoculated medium, no ferrous iron was oxidised. In the absence of bacteria, 90% of ferrous ions were thus oxidised after 170 and 340 h respectively with WWP3 and WKL20. This Fe2+ oxidation capacity of ferrous ion oxidation was previously observed on activated carbon [11] and carbon fibre [31].

Figure 1. Ferrous oxidation by Activated Carbon Cloth WWP3 or WKL20 under different conditions: with and without bacteria 1111

Microbiology Fundamentals

The presence of bacteria in solution did not change the Fe2+ oxidation rate during the first 40h for of WKL20 and during 70 h for WWP3. Ferrous ions were rapidly oxidised either by ACC samples than by A. ferrooxidans biomass. After this period, bacterial Fe2+ oxidation was more important than that observed with ACC alone. 90% of ferrous ions were thus biologically oxidised within a short time (respectively 50 h and 90 h in the presence of WKL20 and WWP3 ACC). This oxidation activity was directly correlated to the free or immobilised biomass production. The protein concentrations (5 mg.L-1 with of WKL20 and 3 mg.L-1 with WWP3) were lower than that obtained without ACC (7 mg.L1 ). As a conclusion ACC supports present an inhibition effect for iron-oxidising activity such as observed for activated carbon [23], flowers of sulphur, fluorapatite, glass bead and pyrite [5,32] This bacterial diminution in solution was linked to the development of a biofilm onto the ACC surface associated with a jarosite precipitation (Fig. 2). No jarosite or biofilm structures were observed onto initial ACC samples (data not shown here). A. ferrooxidans modified ACC materials show a catalytic effect in ferrous ion oxidation as observed in the case of Biomass Support Particles with Activated Carbon coating [11].

Figure. 2. SEM observations of A. ferrooxidans biofilm obtained (a) onto WKL20 after 60h, (b) WWP3 after 90h

The development of a stable biofilm onto ACC was studied during subcultures of A. ferrooxidans modified ACC in Fe2+ growth medium (see Materials and Methods) (Fig. 3). The amount of biomass present within the biofilm was indirectly measured from the amount of biomass present in solution (via protein concentration) and associated to SEM observation of the biofilm. The total Fe2+ oxidation activity was estimated by determining the time required to oxidise 90% of ferrous ions (T90), resulting to the biological activity due to the biomass eventually present in solution or immobilised into the biofilm. The evolution of protein concentration or T90 appeared to be similar for both biologically modified ACC. The ferrous activity curves involved three distinct steps. During step I, the bacterial concentration in solution after oxidation of 90% Fe2+ was greater than that obtained during a batch culture without AC Cloths (5 mg.L-1). These freely dispersed bacteria resulted from two actions: either cells were divided in the biofilm, left the biofilm, and were suspended in medium, or / and the loss of some bacteria from the biofilm due to friction between the ACC support and glass flask. [15] In both cases, Fe2+ oxidation rate was increased as demonstrated by the fast T90 decreasing from 85h to 10-18 h depending on the ACC biofilm. The biofilm was more developed that that obtained after the first culture and consisted of jarosite precipitates and 1112

Microbiology Fundamentals

bacteria (Fig 4). Bacterial cells become embedded within iron precipitates and were adsorbed through complex multilayers patterns.

Figure 3. Biological characteristics (protein concentration and time necessary to oxidise 90% of ferrous ions, T90) of subcultures of ACC WKL20 (A) and WWP3 (B) in 9K medium. A. ferrooxidans biomass (0.07 mg.L-1) being only introduced during the first subculture). The vertical arrows with dotted lines indicate the different steps in the biofilm formation (see text)

During step II, T90 was increased from 10-18 h to 40 h, thus indicating a decrease of the overall activity for ferrous ions oxidation. Such diminution of biological activity was correlated to a large bacterial concentration in solution (6-11 mg protein L-1). The biomass present in solution exhibited a very low or a non-existent ferrous ion oxidation activity. However more jarosite deposits and bacteria were observed onto biofilm (Fig. 4). Step III led to the formation of a biofilm which is stable for its structure and its metabolic activity: no variation in the biomass concentration present in solution and in the time necessary to oxidise 90% of ferrous ions initially present in solution were observed. A weak bacterial concentration in solution (1-2 mg protein L-1) was reached after 13 subcultures. The final biofilm recovered the whole ACC surface and involved a significant jarosite deposit (Fig. 4). Finally, a stable and active A. ferrooxidans biofilm was obtained onto both ACC under study after about 300-350 h subcultures. This biofilm oxidised five times faster the substrate (ferrous ions) than a culture with free biomass. In order to decrease the time required to obtain a stable biofilm onto ACC, the clothes used latter were previously modified by A. ferrooxidans biomass. This modification resulted of an initial sorption of concentrated biomass onto ACC (see Materials and Methods). A 2 h contact period allowed to reach the equilibrium between non immobilised 1113

Microbiology Fundamentals

and immobilised cells. A better bacterial sorption capacity was obtained with WWP3 (7.8 mg protein L–1) than with WKL20 (4.6 mg protein L–1) as confirmed by SEM observations (Fig. 5).

Figure 4. SEM observations of A. ferrooxidans biofilm onto Activated Carbon Cloths WKL20 and WWP3 during subcultures in 9K medium (arrows =bacterial cell)

Figure 5. SEM observations of Activated Carbon Cloths modified with bacteria after 2h of bacterial fixation (pH 1.4, 30°C)

The development of A. ferrooxidans biofilm from these initially fixed biomasses was studied during subcultures of biologically modified ACC in 9K medium (Fig. 6A, B). No shape difference was observed on the evolution of biological characteristics whatever the ACC nature may be. However these curves were simpler than those observed when the cloth is directly introduced into the bacterial culture (Fig. 3): only two steps were observed instead of three.

1114

Microbiology Fundamentals

During the first culture, the ferrous ions were rapidly oxidised by bacteria as shown by the weak T90 value (12 h). This oxidation was correlated to a desorption process confirmed by measuring protein concentration in solution and by comparing SEM observations (Fig. 7 and Fig. 4 for WWP3).

Figure 6. Biological characteristics (protein concentration and T90) of subcultures for ACC WKL20 (A) and WWP3 (B) modified with A. ferrooxidans biomass in 9K medium. The vertical arrows with dotted lines indicate the different steps in the biofilm formation (see text)

Protein concentration in solution differed according to the AA cloth nature: 12 mg.L-1 for WKL20 and 4 mg.L-1 for WWP3. During the five first subcultures, the protein concentration in solution increased because of bacterial desorption. This desorption was correlated to an increase of the time required to biologically oxidise ferrous ions: 12h for WKL20 and 16 h for WWP3. However, the biofilm was developing as shown by SEM observations on Fig. 7. From the sixth subculture, protein concentration in solution continuously decreased down to 1-2 mg protein L-1. T90 simultaneously decreased indicating a high ferrous ion oxidising capacity. This capacity was mainly due to biofilm because of the very low bacteria concentration in solution.

1115

Microbiology Fundamentals

After 232-265h of total subculture, the biologically active biofilm oxidised the ferrous ion substrate within 10-12h, with a very little biomass desorption (1-2 mg protein L-1).

Figure 7. SEM observations of the formation of an A. ferrooxidans biofilm on Activated Carbon Cloth WKL20 and WWP3 4.

CONCLUSION Activated Carbon Cloth has been selected as a supporting material because of its large specific volume and high ability to adsorb A. ferrooxidans. The ACC here under study showed their capacity to oxidise ferrous ions. However, this capacity is improved by developing a A. ferrooxidans biofilm. As observed by SEM, the biofilm consisted in jarosite deposits wherein bacterial cells are able to develop. The development of a stable iron-oxidising biofilm may be explained by the model of "biofilm with adsorbed cells" previously described by Karamanev [15]. A very high iron-oxidising activity (600 mg Fe2+ oxidised L-1/h) was obtained for biologically modified ACC. A preliminary bacterial immobilisation onto ACC allowed an enhancement in developing and stabilising the biofilm. The formation of a stable biofilm and long time incubation in acidic solution led to a loss of mechanical resistance. Therefore ACC WKL20 rapidly breaks in contrast to ACC WWP3. ACKNOWLEDGEMENTS This work was carried out owing to a grant from "Consejo Nacional de Ciencia y Tecnologia" (CONACYT, Mexico) accorded to Mr Felipe Jesus Cerino Cordova.

1116

Microbiology Fundamentals

REFERENCES

1. A. Das and A.K. Mishra, App. Microbiol. Biotechnol., 45 (1996) 377. 2. J. Haddadin, C. Dagot, M. Fick, Enzyme Microbial Technol., 17 (1995) 290. 3. G. Mercier, M. Chartier, D. Couillard, J.F. Blais, Environ. Manag., 4 (1999) 517. 4. E. Donati, C. Pogliani, J.L. Boiardi, Appl. Microbiol. Biotechnol., 47 (1997) 636. 5. N. Wakao, M. Mishina, Y. Sakurai, H. Shiota, J. Gen. Appl. Microbiol., 30 (1984) 63. 6. T.A. Fowler, P.R Holmes, F.K. Crudndwell, Appl. Environ. Microbiol.,.7 (1999) 2987. 7. A. Pagella and D.M. De Faveri, Chem. Eng. Sci., 55 (1999) 2185. 8. T.A. Wood, K.R. Murray, J.G. Burgess, Appl. Microbiol. Biotechnol., 56 (2001) 560. 9. H.R. Diz and J.T. Novak, J. Environm. Eng., 2 (1999) 109. 10. M. Nemati and C. Webb, Appl. Microbiol. Biotechnol., 46 (1996) 250. 11. M. Nemati and C. Webb, J.Chem. Technol. Biotechnol., 74 (1999) 562. 12. A. B. Jensen and C. Webb, Process Bioch., 2 (1995) 225. 13. S. I. Grishi and O.H. Tuoniven, Appl. Environm. Microbiol., 12 (1988) 3092. 14. S. I. Grishi and O.H. Tuoniven, Appl. Environm. Microbiol., 31 (1989) 31. 15. D.G. Karamanev, J. Biotechnol., 20 (1991) 51. 16. A. B. Jensen and C. Webb, Biotechn. Techn., 2 (1994) 87. 17. L. Nikolov, D. Karamanev, V. Mamatarkova, D. Mehochev, D. Dimitrov, Biochem. Eng. J., 12 (2002) 43. 18. H. Olen and R.F. Unz, J. Water Pollut. Control Fed, 52 (1980) 257. 19. H. Olen and R.F. Unz, Biotechnol. Bioeng, 19 (1977) 1475. 20. T. Imaizumi, Biotechn. Bioeng Symp., 16 (1986) 363. 21. H. Magota, Y. Shiratori, C. Inoue, E. Yanagisawa, JP patent No.01215399 (1989). 22. H. Armentia and C. Webb, Appl. Microbiol. Biotechnol., 36 (1992) 697. 23. T. Kai, T. Takahashi, Y. Shirakawa, Y. Kawabata, Biotechnol. Bioeng., 36 (1990) 1105. 24. S. Sandhya and R.A. Pandey, J. Environ. Sci. Health, 2 (1992) 445. 25. J.M. Gomez, D. Cantero, C. Webb, Appl. Microbiol. Biotechnol., 54 (2000) 335 26. C. Brasquet, E. Subrenat, P. Le Cloirec, Wat. Sci. Tech. 39 (1999) 201. 27. K. Kadirvelu, C. Faur-Brasquet, P. Le Cloirec, Langmuir, 16 (2000) 8404. 28. A. Boyer, J.P. Magnin, P. Ozil, Int. Biohyd. Symp., Madrid, Spain, (1999) 211. 29. O.H. Tuoniven and D. P. Kelly, Arch. Microbiol., 88 (1973). 285. 30. M.K. Muir and T.N. Anderson, Metall. Trans. B., 88 (1977) 517. 31. M. Uchida, O. Shinohara, S. Ito, N. Kawasaki, T. Nakamura, S. Tanada, J. Colloid Interf. Sci., 224 (2000) 347. 32. A.A. Dispirito, P.R. Dungan, O.H. Tuoniven, Biotechnol. Bioeng., 23 (1981) 2761.

1117

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Heavy metal precipitation by off-gases from aerobic culture of Klebsiella pneumoniae M426 A.M.M. Essa, L.E. Macaskie∗ and N.L. Brown School of Biosciences, The University of Birmingham, Edgbaston, Birmingham B15 2TT UK Abstract In this study we have identified a new method for the removal of heavy metals from solution by uses the off-gas produced from aerobic cultures of Klebsiella pneumoniae M426. GC-Mass Spectroscopic analysis showed that the outlet gases contain dimethyldisulfide (DMDS), which precipitates some metals from their solutions. The outlet gases precipitated mercury, cadmium, palladium, and lead as metallo-thiol complexes but did not precipitate other metals such as nickel, cobalt, zinc and arsenite. Energy dispersive X-ray microanalysis showed that the metal precipitate consisted of the metal plus carbon, sulfur and oxygen. Formation of metal carbonates was discounted since precipitation was similar at pH 2.0; the precipitate did not have the physical characteristics of metal sulfides. Bio-mineralization of heavy metals by using DMDS is a novel approach because it can be used with concentrated or physiologically incompatible solutions, and since the metal precipitate is kept separate from the bacterial biomass, the biological and chemical end products can be managed independently. 1.

INTRODUCTION The discharge of heavy metals into the environment due to agricultural, industrial and military operations and the effect of this pollution on the ecosystem and human health are growing concerns. Since the natural mineralization of metals is a slow process, pollution by metals constitutes one of the most important environmental problems. Some metals seem to serve no biologically relevant function e.g. Cd, Pb, Sn, Hg instead they cause damage due to their avidity for the sulfhydryl groups of proteins which they block and inactivate [1-6]. Different procedures for the removal of toxic metal species from contaminated environments have been developed; most of them are based on ionexchange technologies and/or chemical precipitation of cations in an inert form. These methods are expensive and also require additional products for desorption of metals and for cleaning up of the inorganic matrix and lead to formation of concentrated secondary wastes. Remediation technologies using microorganisms are feasible alternatives to the physical cleaning of soils or the concentration of metals in polluted waters by physical or chemical means [7]. For example, sulphate-reducing bacteria have been used to precipitate metals as metal sulfides [8] but this requires control of emission of highly toxic H2S. As ∗ To whom correspondence should be addressed (Fax: +44(0)121-414-5925; e-mail: [email protected])

1119

Microbiology Fundamentals

an alternative S-based precipitation process organosulfur compounds have been considered, but these are too expensive for routine use [9]. Dimethyldisulfide (DMDS) has been reported to be produced from different organisms: Pseudomonas sp. [10], Streptomyces sp. [11], Aspergillus sp. [12], Clostridium sporogenes [13], Alteromonas putrefaciens, Pseudomonas fluorescens and Achromobacter sp. [14], Pseudomonas putida [15], Proteus sp. [16] and Pseudomonas fluorescens, Proteus vulgaris, and Serratia marcescens [17]. These bacteria were isolated from soil, wastewater, and fish. Some bacterial strains isolated from activated sludge [18] as Lactobacillus sp., Corynebacterium sp., Alcaligenes sp., and Pseudomonas sp. showed an ability to produce DMDS from activated sludge. The formation of methanethiol from methionine by cell extracts of Lactococcus lactis subsp. cremoris B78 was proposed by Engels et al. [19]. The first step is the transamination during which 4-methylthio-2-ketobutyric acid (KMBA) is produced. The intermediate KMBA is probably, converted to methanethiol, which under aerobic conditions is converted rapidly to dimethyldisulfide and/or dimethyltrisulfide. In this study we describe a novel approach for the removal of some heavy metals by using the outlet gases produced during the aerobic culture of Klebsiella pneumoniae M426. These gases contain a volatile sulfur compound dimethyldisulfide (DMDS), which is proposed to precipitate heavy metals with high efficiency. 2.

MATERIALS AND METHODS

2.1 Bacterial strains and bioreactor operation Klebsiella pneumoniae M426 strain was originally obtained from the NCTC (National Culture Type Collection) Colindale, London. Bacterial cells were grown in Luria Broth medium (LB; 20) in a shaker agitated at 200 rpm and incubated at 37°C. A batch bioreactor was constructed for the metal precipitation (Fig. 1) consisting of two chambers, one for the bacterial growth (about 1L in volume, maintained under aerobic conditions by pumping in a sterilised air, 37°C) and the other chamber (50ml) for the precipitation of metal via the pumping of the headspace gases from the culture through the metal solution. Cultures were inoculated (100 ml) using cells in an exponential growth phase (6hrs) and bacterial growth was monitored by measuring the optical density at 600nm. 2.2 Analytical methods 2.2.1 Analysis of the soluble metals Heavy metals (Hg2+, Cd2+, Pb2+, Pd2+) were prepared as aqueous solutions, and used as the following concentrations: mercuric chloride, 0.02-10.0 mg/ml; cadmium chloride, 0.7 mg/ml; lead nitrate, 0.6 mg/ml; sodium tetrachloropalladate, 0.9 mg/ml. The metal precipitates from the precipitation chamber were removed by centrifugation at 1600g for 15 minutes. Aliquots (10 ml) of the supernatant were filtered through 0.2 µm (Nalgene Syringe filters). The filtrates (Cd2+, Pb2+, Pd2+) were analysed using a 1999 Duo HR Iris advanced inductive coupled plasma (ICP) spectrophotometer [21]. Mercuric ion concentrations in the solutions were measured with a mercury analyser (Buck Scientific, model 400A) applying cold vapour atomic absorption spectrometry [22].

1120

Microbiology Fundamentals

Figure 1. A bioreactor used for metal precipitation via the culture off-gas consisting of two chambers one is the bacterial growth chamber and the second is the metal precipitation chamber 2.2.2 Analytical electron microscopy Samples from the precipitation chamber were centrifuged at 1600g for 15 minutes. The supernatant was removed and precipitates were washed in 20ml deionised water followed by centrifugation as before. This step was repeated 3 times. The precipitate was collected and dried at 30°C. The metal precipitates were examined with a JEOL JSM 5900 SEM. The elemental composition of the mineral phases was determined by energy dispersive X-ray microanalysis (EDX) using an Oxford Link ISIS System. 2.2.3 GC–mass spectral analysis The outlet gases produced from the bioreactor were collected in serum bottles (50ml) with a rubber caps. The samples were immediately analysed by GC–MS [23], using a Hewlett-Packard 5890/5970A system, with a HP1 column (50 mm x 0.2 mm fused silica capillary column, film thickness, 0.5 mm). GC oven initial temperature was 60°C and was programmed to 220°C at a rate of 2°C/min, and finally held at 220°C during 40 min. Operating conditions of the GC were as follows: helium was used as carrier gas (0.8 ml/min); the temperature of injector and detector was 250°C. The mass spectra were determined at 70 eV in the mass range of 35-400.

1121

Microbiology Fundamentals

3.

RESULTS AND DISCUSSION

3.1 Precipitation of mercury as HgS by growing cultures During the aerobic growth of K. pneumoniae M426 a black precipitate was visible in the extracellular matrix (Fig. 2a). EDX analysis data showed that the precipitate contained mercury and sulfur (Fig. 2b). Accordingly, it was assumed that K. pneumoniae M426 has the ability to produce hydrogen sulfide (H2S) under aerobic conditions, which can precipitate Hg2+ as insoluble HgS [24]. These results were confirmed when a suspended filter paper saturated with HgCl2 turned black after exposure to the culture outlet gases. This phenomenon was recognised previously by Aiking et al. [25], who recorded an aerobic mechanism for mercury bioremediation, in which Klebsiella aerogenes formed HgS when grown in continuous aerobic culture in the presence of HgCl2 and at the same time there was no bacterial mercury volatilization. This strain showed mercuric ion sensitivity under sulfate-limited conditions and elevated levels of total cellular sulfide were detected in cells grown in the presence of mercuric ions. It was suggested that this was due to the formation of HgS. A similar phenomenon was recorded by Wang et al. [26] who showed the ability of Pseudomonas aeruginosa to remove more than 99% of the cadmium from 5mM cadmium solution by precipitation on the cell wall as cadmium sulfide under aerobic conditions.

Figure 2. (a) Electron micrograph of aerobic culture of K. pneumoniae M426 showing an extracellular dark precipitate, (b) EDX spectrum of the dark precipitate in (a) 3.2 Mercury and cadmium bio-mineralization using culture off-gases A bioreactor for mercury bioremediation was designed according to the ability of K. pneumoniae M426 to precipitate the soluble Hg2+ as HgS. The bacterial cells were used as a gas generator in a separate chamber and the headspace gases produced during aerobic growth were pumped into the metal solutions in the precipitation chamber through a bacterial membrane filter to avoid any bacterial contamination. A black HgS precipitate was anticipated, but more than 99% of the soluble Hg2+ was recovered as a yellowish white precipitate within 2h (Fig. 3). A similar result was obtained 1122

Microbiology Fundamentals

with cadmium; within 1h there was more than 99% of Cd2+ removal from CdCl2 solution. EDX analysis showed that the precipitates consisted of mercury or cadmium plus carbon, sulfur and oxygen and there was no phosphorus in the precipitate, which confirms that there was no bacterial contamination (Fig. 4a & 4b, 5a & 5b). 1

900 Bacterial growth

0.9

Residual soluble mercury 800

0.8

700

0.7

600

0.6 500

Bacterial growth

0.5 400 0.4

(O.D600)

300

0.3

Residual soluble Hg2+ (mg/l)

200

0.2

100

0.1 0

0 0

20

40

60 Time (min)

80

100

120

Figure 3. A time course for bacterial growth and mercury bio-precipitation by using the outlet gases produced from the aerobic culture of K. pneumoniae M426. The data shown are for an initial Hg2+ concentration of 0.8 mg/ml, similar results were obtained using Hg2+ up to 10.0 mg/ml (Error bars represent means ± standard errors)

Figure 4. Mercury precipitation by using the volatile thiols produced by K. pneumoniae M426, (a) an electron micrograph of the mercury precipitate, (b) EDX spectrum of the mercury precipitate

1123

Microbiology Fundamentals

Figure 5. Cadmium precipitation by using the volatile thiols produced by K. pneumoniae M426, (a) an electron micrograph of the cadmium precipitate, (b) EDX spectrum of the cadmium precipitate

The EDX spectrum for the Hg precipitate cannot easily distinguish between Hg and sulfur due to peak overlap, but the peaks were clearly differentiated in the corresponding Cd-precipitate. It could be suggested that metal carbonates were precipitated. However similar results were obtained at pH 2.0, at which pH little HCO3- would be present in the solution and no CO32- [27]. GC-Mass Spectroscopic analysis showed that the outlet gases from the bioreactor contained volatile dimethyldisulfide (DMDS). According to these results, we suggest a novel mechanism for mercury bioremediation by the production of DMDS as a metabolic by-product during the aerobic growth of K. pneumoniae M426. This volatile species may form insoluble mercurithiol-compounds. There was more than 99% mercury precipitation by this mechanism with different concentrations of HgCl2 (from 20 µg/ml to 10 mg/ml) within 2 hrs. This bioprocess also gave high metal removal efficiency over a wide pH range (acidic, neutral or alkaline) and also at salinity levels up to 1M NaCl. DMDS for bio-mineralization of heavy metals The head gas produced by aerobic culture of K. pneumoniae M426 did not precipitate some metals e.g. nickel, cobalt, zinc and arsenite but showed a high efficiency of precipitation of cadmium, lead and palladium from solution with about 99% metal removal efficiency within 1h. EDX microanalysis showed that the metal precipitates consisted of metal plus carbon, sulfur and oxygen, while GC-Mass Spectrum analysis showed the presence of dimethyldisulfide (DMDS) in the culture off-gas. We attributed the precipitation to the high affinity between DMDS and some metal ions to produce metallo-thiol complexes. The formation of metal sulfides seems unlikely since mercury and cadmium deposits were white instead of black and yellow respectively and the characteristic X-ray powder diffraction spectra of metal sulfides could not be obtained (data not shown). There are some reports of the use of organosulfur compounds [2,3Dimercaptopropanol, meso-2,3-Dimercaptosuccinic acid (DMSA), Sodium 2,3dimercapto-1-propanesulfonate (DMPS)] as therapeutic chelating agents for the removal of some heavy metals (arsenic, cadmium, lead and mercury) from living cells. The formation of complexes with metal ions can be attributed to the presence of sulfhydryl group in these chelators, which show a great affinity for metal ions producing the metal 1124

Microbiology Fundamentals

complexes [28-35]. According to our knowledge this is the first study using a naturally produced organo-volatile sulfur gases as a chelating agent to precipitate heavy metals with such high metal removal efficiency under aerobic conditions. Current studies are aiming to quantify DMDS production and its contribution to metal precipitation. 4.

CONCLUSIONS There have been attempts to engineer aerobic sulfate or thiosulfate reducing systems for metal removal to produce more sulfide [36, 37]. This study describes a novel natural process not only for mercury bioremediation but also for some other heavy metals e.g. cadmium, lead and palladium via aerobic bio-precipitation of metal ions as a metal-thiol complex. This mechanism provides an excellent approach for bioremediation of some heavy metal and has some advantages. Firstly, since culture off-gas is used, it is applicable for concentrated or non-physiologically compatible solutions under a wide range of environmental factors such as pH and salinity. Secondly, with this bioprocess one organism, K. pneumoniae, was used to precipitate different metals from their solutions and could be used with mixed metal solution. Thirdly, and an important advantage, during metal bio-mineralization there is no direct contact between the bacterial biomass and the metal precipitate; thus keeping the biomass away from the original contaminated waste water, ensuring that there is no cross-contamination between the biomass and the final metal precipitate. REFERENCES

1. G. M. Gadd, Accumulation of metals by microorganisms and algae. In: Biotechnology, Ed. H. J. Rehm. VCHV, Weinheim, 1988. 2. G. M. Gadd, Microbial control of heavy metal pollution. In: Microbial Control of Pollution, Eds J. C. Fry, G. M. Gadd, R. A. Herbert, C. W. Jones and I. Watson-Craik. Cambridge University Press, Cambridge, 1992. 3. G. M. Gadd, FEMS Microbiol. Rev. 11 (1993) 297. 4. M. D.Mullen, D. C. Wolf, F. G. Ferris, T. J. Beveridge, C. A. Flemming and G. W. Bailey, Appl. Environ. Microbiol. 55 (1989) 3143. 5. S. G.Walker, C. A. Flemming, F. G. Ferris, T. J. Beveridge and G. W. Bailey, Appl. Microbiol. Biotechnol. 55 (1989) 2967. 6. T. J. Beveridge, Rev. Microbiol. 43 (1989) 147. 7. M. M. Valls and V. de Lorenzo, FEMS Microbiol. Rev. 26 (2002) 327. 8. G. M. Gadd and C. White, Trends Biotechnol. 11 (1993) 353. 9. R. M. Austrain, Ph.D Thesis, Institut für Thermische Verfahrenstechnik und Umwelttechnik, Germany, 1994. 10. R. E. Kallio and A. D. Larson, Methionine degradation by a species of Pseudomonas. In W. D. McElroy and H. B. Glass (ed.), A symposium on amino acid metabolism. Johns Hopkins University Press, Baltimore, 1955. 11. W. Segal and R. L. Starkey, J. Bacteriol. 98 (1969) 908. 12. J. Ruiz-Herrera, and R. L. Starkey, J. Bacteriol. 99 (1969) 544. 13. W. Kreis and C. Hession, Cancer Res. 33 (1973) 1862. 14. A. Miller, R. A. Scanlan, J. S. Lee and L. M. Libbey, Appl. Microbiol. 26 (1971)18. 15. S. Ito, T. Nakamura and Y. Eguchi, J. Biochem. 79 (1976) 1263 16. N. J. Hayward, T. H. Jeavons, A. J. C. Nicholson and A. G. Thornton, J. Clin. Microbiol. 6 (1977) 187. 17. M. Pohl, E. Bock, M. Rinken, M. Avvdin and W. A. Konig. Z, Naturforsch. Teil C. 39 (1984) 240. 1125

Microbiology Fundamentals

18. B. Tomita, N. Hmamura and Y. Ose. Eisei Kagaku, 27 (1981) 98. 19. W. J. M. Engels, A. C. Alting, M. M. T. G. Arntz, H. Gruppen, A. G. J.Voragen, G. Smit and S. Visser, Int. Dairy J. 10 (2000) 443. 20. J. Sambrook, E. F. Fritsch and T. Maniatis, Molecular Cloning: a Laboratory Manual, 2nd ed., vol. 3. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 1989. 21. M. M. Matlock, B. S. Howerton and D. A. Atwood. J. Hazardous Materials 84 (2001) 73. 22. J. S. Chang, J. Hong, Q. A. Ogunseitan and B. H. Olson, Biotechnol. Prog. 9 (1993) 526. 23. M. Mestres, O. Busto and J. Guasch, J. Chromatography 945 (2002) 211. 24. A. M. M. Essa, L. E. Macaskie and N. L. Brown, Biochem. Soc. Trans. 4 (2002) 672. 25. H. Aiking, H. Govers and J. 't Riet, Appl. Environ. Microbiol. 50 (1985) 1262. 26. C. L. Wang, P. C. Michels, S. C. Dawson, S. Kitisakkul, J. A. Bross, J. D. Keasling and D. S. Clark, Appl. Environ. Microbiol. 63 (1997) 4075. 27. C. N. Mulligan, R. N. Yong and B. F. Gibbs, J. Hazardous Materials, 85 (2001) 111. 28. H. V. Aposhian and M. M. Aposhian, Ann. Rev. Pharmacol. Toxicol. 30 (1990) 279. 29. H. V. Aposhian, D. C. Bruce, W. Alter, R. C. Dart, K. M. Hurbult and M. M. Aposhian, FASEB J. 6 (1992) 2472. 30. M. M. Aposhian, R. M. Maiorino, Z. Xu and H. V. Aposhian, Toxicol. 109 (1996) 49. 31. P. Asiedu, T. Moulton, C. B. Blum, E. Roldan, N. J. Lolacono and J. H. Graziano, Environ. Health Perspect. 103 (1995) 734. 32. J. Chisolm, J. Clin. Toxicol. 30 (1992) 493. 33. J. H. Graziano, N. J. Lolacono, T. Moulton, M. E. Mitchell, V. Slavkovich and C. Zarate, J. Pediatr. 120 (1992) 133. 34. M. M. Jones, Crit. Rev. Toxicol. 21 (1991) 209. 35. R. A. Peters, L. A. Stocken and R. H. S. Thomson, Nature. 156 (1945) 616. 36. S. W. Bang, D. S. Clark and J. D. Keasling, Appl. Environ. Microbiol. 66 (2000) 3939. 37. C. L.Wang, P. D. Maratukulam, A. M. Lum, D. S. Clark and J. D. Keasling, Appl. Environ. Microbiol. 66 (2000) 4497.

1126

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Influence of pH, Mg2+ and Mn2+ on the bioleaching of nickel laterite ore using the fungus Aspergillus niger O5 O. Cotoa, D. Gutierreza, L. Abína J. Marreroa and K. Boseckerb a

Departamento de Microbiología. Facultad de Biología. Universidad de la Habana. Calle 25 entre J e I. Plaza. Vedado. Ciudad Habana. Cuba b Federal Institute for Geosciences and Natural Resources (BGR) Stilleweg 2, D 30655 Hannover, Germany

Abstract The main global nickel deposits are laterite deposits. Their nickel contents vary from 0.3-1.5% and are associated with a non-soluble phase that cannot be processed completely by conventional methods. The Cuban nickel deposits are recognized as one of the largest deposits in the world. Besides nickel they contain other valuable metals, such as cobalt and manganese, which in spite of low concentrations of valuable metals, are of important commercial interest. The aim of this investigation was to study the influence of pH and Mg2+ and Mn2+ ion concentrations on the dissolution of Ni, Co, Mn, and Mg from serpentinite in the Punta Gorda deposit in Moa (Cuba). For leaching, Aspergillus niger O5 was chosen because of its ability to dissolve a high percentage of Co, Ni and Mn from the mineral substrate and for its production of citric acid. The strain produced higher concentration of citric acid (25-27 g/L) when grown in the presence of the ore than in the absence of ore (14 g/L). The production of organic acid by the fungus varied with the conditions of incubation. The initial pH of the medium was a very important variable in this process; its change during the course of the time affected the bioleaching efficiency and gave rise to biosorption processes.

Keywords: bioleaching, heterotrophic microorganisms, Aspergillus niger, Penicillium simplicissimum, citric acid, organic acids, laterite ore, oxide ore 1.

INTRODUCTION Cuba has an estimated world reserve of Ni of 37% and ranks first in the world reserves of Ni and second in world reserves of Co. The main nickel deposits in Cuba are on the east side of the island in the Moa-Nicaro area of Holguín province. In this area there are three commercial plants where Ni+Co oxides and Ni+Co sulfides are obtained by hydrometallurgy and pyrometallurgy, respectively. Both technologies require high temperature, special equipment and constitute a high environmental risk. For this reason, bioleaching with organic acid produced by fungi (1-9) might be an attractive alternative for the Cuban nickel industry in order to develop a sustainable technology. The number of scientific papers and patents on bioleaching of sulfide ores is quite large and exceeds several times the number of papers related to bioleaching of non-sulfide 1127

Microbiology Fundamentals

ores. Bioleaching of non-sulfide ores is based on the ability of fungi to produce organic acids such as citric, oxalic, and fumaric acid, which act as leaching agents (1–9). They are excreted into the environment and dissolve heavy metals by direct displacement of metal ions from the ore matrix by hydrogen ions and by the formation of soluble metal complexes and chelates (2-3). Fungi most active in leaching are Penicillium simplicissimum and Aspergillus niger. Studies of nickel leaching have revealed that the best recovery has been obtained with citric acid (2, 3, 6, 8). The main objective of this paper was to study the influence of pH and Mg2+ and Mn2+ ion concentrations on the dissolution of Ni, Co, Mn and Mg present in the serpentinite from the Punta Gorda deposit at Moa, Cuba. 2.

MATERIALS AND METHODS

2.1 Laterite ore Nickeliferous ore from the Punta Gorda, Moa (Cuba) deposit was used in all experiments. Table 1. Chemical and mineralogical composition of Cuban laterite ore

Concentration (%) Phases %

Ni 1.54 Al2O3 1.64

Co 0.27 Fe2O3 17.53

Fe 14.8 MnO 0.253

Element Mg 19.86 MgO 29.56

Mn 0.49 SiO2 35.87

Al 1.95

Cr 1.02

2.2 Screening of microorganisms for bioleaching Sixteen strains of fungi were used in this work, 15 of them identified as Penicillium simplicissimum (strains 2, 3, 14, 16, 22, 23, 25, 26, 42, 43, 47, 50, 56, 60, 61). They were obtained from the Federal Institute for Geosciences and Natural Resources, Hannover, Germany. The other fungus was Aspergillus niger strain O5 (6). All strains were grown on solid Saboraud medium and incubated for one week at 30°C. Culture medium 1a (100 mL) was inoculated with 2 mL containing 107-108 spores/mL in a 500 mL Erlenmeyer flask. Culture medium 1a was prepared as follows: 140 g/L sucrose; 0.3 g/L yeast extract; 0.5 g/L KH2PO4; 0.250 g/L MgSO4 and 0.125 g/L MnSO4. The pH was adjusted to 6.5. The spores were counted using a Thoma counting chamber. All leaching experiments were carried out at 5% pulp density of laterite ore. The ore was milled to a mean particle size of 0.250 mm. The culture medium was incubated at 100 rpm at 30°C. Each time a sample was taken from the solution the same volume was replaced with sterile culture medium. One Erlenmeyer flask without inoculum was used as a control. 2.3 Culture conditions for bioleaching To study the effect of oxygen on bioleaching, one Erlenmeyer flask was placed on a shaker for the duration of the experiment, a second one was not shaken. The influence of the culture medium was investigated using seven variants of the growth medium. Culture medium 3a was used only for the Ni leaching experiment.

1128

Microbiology Fundamentals

Table 2. Variants of culture media

Mg2+ Mn2+ pH

Culture media (with 140 g/L sucrose and 0.3 g/L yeast extract) 1 1a 2 2a 3 3a 3b + + + + + + + + 6.5 6.5 5.5 3.0 5.5 5.0 3.0

2.4 Bioleaching in two-stages In the first step, culture medium 3a (100 mL in a 500 mL Erlenmeyer flask) was inoculated with A. niger O5 and incubated at 150 rpm/min for 8 days at 30°C. In the second step, the biomass was removed by filtration; the filtrate was autoclaved and used to leach Ni ore with a 5% pulp density. 2.5 Analytical methods At regular intervals, samples were withdrawn for pH measurement and element analysis by ICP-AES. The type and concentration of the organic acids produced were determined using a densitometer after separation by thin layer chromatography (TLC). The spots became apparent under ultraviolet light (366 nm) and after staining at visible light. The mineralogical composition of the Cuban laterite ore and some solid residues from bioleaching was determined by X-ray diffraction (XRD). This was done using a Philips X'pert system PW1730 equipped with a Cu tube, primary and secondary Soller slits and a secondary monochromator. The main phases are chrysotile, magnesioferrite and goethite. 3.

RESULTS AND DISCUSSION

3.1 Screening for microbial leaching In a preliminary experiment, we tested the ability of 16 strains of fungi to leach nickel and other metals from Cuban laterite ore. Most of of the strains were able to dissolve Ni, Co, Mn, Al, Fe, and Cr to some extent, whereas strains 42, 56, 60 and 61 dissolved less than 10% (data not shown). The amount of metal leached by four strains of the 16 strains studied is shown in Figure 1 as a function of time. The trend is rather similar with all four strains. In the majority of strains, the initial pH of 6.2 decreased, indicating production of organic acid by the fungus, and the most of the change in pH occurred within the first seven days. We took Ni solubilization as the main criterion for screening programs, because Ni is the metal of the most economic importance in Cuban laterite ore. Table 3 shows the percentage of nickel recovered after one and two weeks of bioleaching, indicating that the strains can be separated into three groups: the first group with the strain 23 and O5, recovered up to 60% Ni within 14 days. Strains 2, 3, 14, 22, 25, 26 and 50 form the second group, dissolving more than 20% and less than 50% Ni. The fungi of the third group, including strains 16, 42, 43, 47, 56, 60 and 61, extracted less than 15% Ni. The best strain was A. niger O5.

1129

Microbiology Fundamentals

Figure 1. Laterite bioleaching in medium 1a: A) strain 50, B) strain 3, C) strain 2 and D) strain 42. Pulp density: 5%; carbon source: sucrose 140 g/L; temperature: 30°C; reciprocal agitation: 100 rpm.min-1. (•: Mn, ○: Mg, ∆: pH, ■: Co, □: Ni, ▲: Cr, ♦: Al and ◊: Fe) Table 3. Solubilization of Ni laterite ore by different strains Fungal strain P. simplicissimum 2 3 14 16 22 23 25 26 42 43 47 50 56 60 61 A. niger O5 1130

% recovery after 7 days 14 days 0.27 14.73 12.64 9.56 15.61 30.15 16.10 21.51 4.93 6.92 11.42 17.52 1.60 10.60 9.36 34.15

34.30 25.89 21.02 14.91 24.42 56.23 24.92 35.43 6.57 11.18 13.72 31.71 2.09 12.35 10.01 57.8

Microbiology Fundamentals

The selection of the best strain is a very important factor in the bioleaching of Ni laterite ore and the leaching efficiency depends on the amount of acid produced by the fungus (1, 2, 6, 8). 3.2 Effect of incubation conditions on bioleaching of metals The results of bioleaching and change of pH with strain O5 under shaking and static conditions are illustrated in Figure 2. Under both conditions the pH decreased to 3.8 within 7 days and remained below 4.0 thereafter. Manganese was leached best and within 10 days about 70% Mn was dissolved. However, the pattern of bioleaching changed in the course of time depending on the incubation conditions. Aspergillus niger O5 was the only strain that showed loss of Ni, Mn and Co from dissolution when cultivated under shaking conditions (Figures 1 and 2). This was observed after 14 days of experiment (Figure 2A) and was due to the uptake of Ni, Co, and Mn by Aspergillus niger, which is known for nickel accumulation in the biomass (2, 10). Some abiotic factors, such as pH, allow the biosorption of heavy metals by fungi (10); precipitation of metals in the presence of oxalic acid (1) or electrosorption (11) may occur as well. Besides manganese and nickel, up to 40% of cobalt and magnesium was removed from the laterite ore (Figures 1 and 2). Previous mineralogical studies of this Cuban laterite ore revealed that nickel and cobalt are associated with goethite and manganese mineral phases, respectively. In general, there are two phases in the bioleaching of nickel laterite ore, similar to the bioleaching of chalcopyrite (12). The first phase is characterized by an increasingly high rate of metal dissolution, due to a strong decrease in the pH, which reflects a high metabolic activity of the microorganisms (7). In the second phase, the rate of metal dissolution decreases or remains almost constant.

Figure 2. Laterite bioleaching with A. niger O5 in medium 1a. A) shaking conditions (reciprocal agitation: 100 rpm.min-1) and B) static conditions. Pulp density: 5%, carbon source: sucrose 140g/L, temperature: 30°C. (•: Mn, ○: Mg, ∆: pH, ■: Co, □: Ni, ▲: Cr, ♦: Al and ◊: Fe) 3.3 Influence of pH and Mg2+ and Mn2+ ion concentrations on the bioleaching Considering that some cations, such as Mn2+, Mg2+, Fe2+, and Fe3+, can affect cell morphology, enzymatic activity and citric acid production by Aspergillus, they are essential micronutrients in the metabolism. We used some variations of the culture 1131

Microbiology Fundamentals

medium to study the effect of these cations on the release of Ni from Cuban laterite and on the type and concentration of organic acid produced by the fungus. The composition of the culture medium considerably influenced metal removal as well as type and amount of organic acids produced by strain O5 (Figures 3 and 4). The highest amount of Ni was leached using medium 3a. After 18 days of incubation, 60% Ni was dissolved, and 86.6% Ni was extracted at the end of experiment (Figure 3C). Medium 3b was the worst for A. niger O5. A different behaviour was observed with Penicillium simplicissimum strain 2 (data not shown). Regardless of the initial pH in the growth medium, all pH values were around 5 after 7 days of bioleaching, after 14 days pH values of 3.8-4.8 were measured. The pH behaviour in medium 3a was different from that of the other media. The concentration of dissolved Mn, Ni and Co increased when media 1a, 2, 2a and 3 were used. Later, a loss of dissolved metals was observed (Figure 3). A)

B) Ni Dissolved (%)

Mn Dissolved (%)

120 100 80 60 40 20

100 80 60 40 20 0

0 0

5

10

15 20 Time (days)

25

0

30

10

20

30

Time (days)

D)

30 25 20 15 10 5 0 0

10

20

Mg Dissolved (%)

Co Dissolved (%)

C)

30

20 15 10 5 0 0

Time (days)

10 20 Time (days)

30

E) 8

pH

6 4 2 0 0

10 20 Time (days)

30

Figure 3. Influence of initial pH, Mg and Mn on bioleaching of Cuban laterite ore by A. niger O5 (A, B, C and D) and change of pH (E). Pulp density: 5%, carbon source: 140g/L sucrose, temperature: 30°C, reciprocal agitation: 100 rpm.min-1. Media with Mg and Mn: •: 1a, □: 3, ∆: 3a, ■: 3b and media without Mg and Mn: ▲: 2, ♦: 2a, x: 1 1132

Microbiology Fundamentals

Oxalic acid (g/L)

60 50 40 30 20 10

Citric acid (g/L)

As shown in Figure 4, the initial pH of the media and the concentrations of Mn2+ and Mg2+ cations influenced the amount and type of organic acid produced by fungal activity. When the strain was cultivated in medium 3a (pH 5.0), the yield of citric acid was considerably enhanced and oxalic acid production was inhibited. This is of considerable advantage, since oxalic acid inhibits Ni laterite bioleaching (1). This strain produced a higher concentration of citric acid (25-27 g/L) when grown in the presence of the ore than in the absence of ore (14 g/L) (data not shown). 30 25 20 15 10 5 0

2a 3b 3 1 1a 2 3a

0

Figure 4. Influence of different culture media on oxalic acid (A) and citric acid (B) production by A. niger strain O5 to10 days of Cuban laterite bioleaching. Pulp density: 5%, carbon source: sucrose 140 g/L, temperature: 30°C, reciprocal agitation: 100 rpm.min-1. Medium: 1, 2, and 2a (without Mg and Mn) and 1a, 3, 3a and 3b (with Mg and Mn)

Some released metal cations inhibit citric acid synthesis by A. niger O5. Mg2+ is a cofactor of the enzyme isocitrate dehydrogenase, and some concentrations of Mn2+ act in a negative way on the citric acid cycle, and citric acid is the best leaching agent for Cuban laterite nickel ore (6). Therefore, we need to define the optimum concentration of metal ions for maximum yield of citric acid. The X-ray diffraction analysis of some residue samples from this set of experiment demonstrated a new phase in the solid residue after bioleaching. This was identified as magnesium oxalate hydrate (Figure 5).

A

B

Figure 5. X-ray diffraction: A) Cuban laterite ore before bioleaching and B) solid residues from bioleaching with A. niger O5 in medium 3 (pH 5.5) 1133

Microbiology Fundamentals

3.4 Two-stage bioleaching The leaching agent produced in the first step was used for the leaching of the ore in the second step. The results of the leaching of laterite ore are shown in Figure 6. The only organic acid metabolite accumulated in the filtrate was citric acid (data not shown). The percentage of Ni recovered using this method (Figure 6A) was similar to that obtained using a single stage method (Figure 3B). pH during the leaching phase is shown in Figure 6B as a function of time. Note that the increase in pH with time (Figure 6B). This behavior is due to the release of 30% Mg, which is a neutralizing agent. In the single-stage bioleaching method, the pH decreased with time (Figure 3E) as consequence of metabolite activity of the fungus.

Figure 6. (A) Chemical leaching of Cuban laterite ore for 12 days by metabolic products of A. niger O5 cultivated in medium 3a (pH 5) and (B) pH without ore. Pulp density: 5%, temperature: 30°C and reciprocal agitation: 100 rpm.min-1 4.

CONCLUSIONS Cuban laterite ore was more amenable to bioleaching with Aspergillus niger O5 than with Penicillium simplicissimum. A. niger is capable of producing oxalic and citric acid or only citric acid depending on the culture conditions. The best leaching agent for Cuban laterite nickel ore was citric acid. The initial pH of culture medium and the composition of growth medium is an important variable for the bioleaching of non-sulfide ore. Cuban sugar cane can be a substitute for sucrose in the culture medium. There was a high solubilization of nickel (89% after 28 days of bioleaching) in medium 3a with an initial pH of 5.0. The single-stage and two-stage bioleaching methods yielded similar results. On the basis of these results, we recommend that a preliminary engineering and economic study be made. REFERENCES

1. K. Alibhai, D. Leak, A.W.L. Dudeney, S. Agatzini and P. Tzeferis, in: R.W. Smith and M. Misra (eds.), Mineral Bioprocessing, The Minerals, Metals and Materials Society, TMS (1991) 191. 2. K. Bosecker, in: R.L. Lawrence, R.M.R. Branion and H.H Ebner (eds.), Fundamental and Applied Biohydrometallurgy, Elsevier, Amsterdam (1986) 367. 3. K. Bosecker, in: J. Salley, R.G.L. McCready, and P.L. Wichlaz (eds.), Biohydrometallurgy, CANMET SP89-10 (1989) 15. 1134

Microbiology Fundamentals

4. W. Burgstaller and F. Schinner, in: A.E. Torma, J.E. Wey and V.I. Lakshmanan (eds.), Biohydrometallurgical Technologies, Vol. 1, The Minerals, Metals & Materials Society, TMS (1993) 325. 5. W. Burgstaller and F. Schinner, J. Biotechnol. 26 (1993), 340. 6. O. Coto, N. Bruguera, L. Abín, J. Gamboa and Y. Gómez, in: V.S.T Ciminelli and O. García Jr. (eds.), Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, Part A, Elsevier, Amsterdam (2001) 175. 7. L. Abin, O. Coto, B. Marrero and J. Marrero, Ciencias Biológicas CENIC No 3 (2002) In press. 8. P. Tzeferis, Int. Journal of Mineral Processing, 42 (1994) 267. 9. L.B. Sukla and V.V. Panchanadikar, in: A.E. Torma, J.E. Wey and V.I. Lakshmanan (eds.), Biohydrometallurgical Technologies, Vol. 1, The Minerals, Metals & Materials Society, TMS (1993) 373. 10. B. Volesky, in: R. Amils and A. Ballester (eds.), Biohydrometallurgy and the Environment towards the Mining of the 21st Century, Part B, Elsevier, Amsterdam (1999) 161. 11. M. Valix, J.Y. Tang and W.H. Cheung, Minerals Engineering., 14 (12) (2001) 1629. 12. O. Coto, A. Ballester, M.L. Blázquez and F. González, Biorecovery, 2 (3) (1993) 121.

1135

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Mercury tolerance of thermophilic Bacillus sp. and Ureibacillus sp. K.J. Glendinning∗ and N.L. Brown School of Biosciences, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, U.K. Abstract Although resistance of microorganisms to Hg(II) salts has been widely investigated and resistant strains have been reported from many eubacterial genera, there are few reports of mercuric ion resistance in extremophilic microorganisms. Thermophilic mercury resistant bacteria were selected by growth at 62°C on Luria agar containing HgCl2. Sequence analysis of 16S rRNA genes of two isolates showed the closest matches to be with Bacillus pallidus and Ureibacillus thermosphaericus. Minimum inhibitory concentration (MIC) values for HgCl2 were found to be 30 µg/ml and 80µg/ml for these isolates respectively, compared to 10 µg/ml for B. pallidus H12 DSM3670, a mercurysensitive control. The best-characterised mercury-resistant Bacillus strain, B. cereus RC607, had an MIC of 60 µg/ml. Growth of the thermophilic isolates in media containing HgCl2 resulted in the formation of a black precipitate. X-Ray Diffraction (XRD) analysis of this precipitate showed it to be HgS. Growth in the presence of radioactive 203HgCl2 at 45°C or 62°C showed progressive removal of mercury from the medium, but the mechanism by which this occurs is uncertain. Biochemical assays for mercuric reductase, which reduces 203Hg2+ to volatile 203Hg°, did not detect this enzyme. It is possible that volatilisation of mercury is occurring as a by-product of normal metabolism, through complexation into volatile compounds.

Keywords: Ureibacillus sp., Bacillus sp., thermophilic, mercury resistance 1.

INTRODUCTION Compounds of mercury, such as mercuric chloride and organomercurials, have long been known to be toxic to both eukaryotic and prokaryotic cells. These compounds can pass through biological membranes [1] and bind with high affinity to thiol (SH) groups in proteins [2], thus causing damage to membranes and inactivating enzymes. Mercury is also genotoxic; inorganic Hg(II) is capable of strong reversible interactions with the nitrogens in purines and pyrimidines, and organic mercury compounds, e.g. methylmercury, also produce irreversible damage to nucleic acids [3]. Environmental contamination with mercury compounds can have devastating effects, as mercury toxicity

∗ Corresponding author. Current address: Birmingham Women's Health Care, Metchley Park Road, Birmingham B15 2TG, UK. Email: [email protected]; Fax: +44-121-627-2711.

1137

Microbiology Fundamentals

is cumulative, with the highest levels of mercury compounds being found in consumers at the top of the food chain. A number of microorganisms have evolved resistance mechanisms to deal with mercury compounds. Mercury resistance was first reported in Staphylococcus aureus [4] and since then has been described in a number of bacterial species. One of the best-defined mercury resistance determinants is the mer operon encoded by transposon Tn501, found in Gram-negative bacteria. The functions of the minimal number of proteins required to confer full resistance are as follows [5]: MerR is the mer regulatory protein which controls the expression of all the other proteins in the operon in response to the presence of Hg(II) salts. MerP is a periplasmic protein which binds mercuric ions entering the periplasm through the porins OmpC and OmpF, and transfers Hg(II) to the MerT transport protein located in the cytoplasmic membrane. The mercuric ions are then passed from cysteine residues in MerT which are exposed to the periplasmic compartment to cysteine residues on the cytoplasmic face of the membrane, and thence to the cytoplasmic mercuric reductase (MR, the merA gene product). Hg(II) is complexed with the active site cysteine residues of MR and reduced from Hg(II) to Hg(0) using NADPH as the reductant. Hg(0) is then lost from the cell in the gas phase. Mercuric ion resistance has also been characterised in Gram-positive genera. The resistance determinants from Bacillus sp. RC607 [6], Streptomyces lividans [7] and Staphylococcus aureus [8] have been characterised. The mechanisms of induction, mercury transport and mercury reduction in Gram-positive genera are similar to those of the Gram-negative systems, but the transport functions differ, as might be expected from the differences in the surfaces of Gram-positive and Gram-negative cells [5]. Although resistance to Hg(II) has been widely investigated, there are few reports of mercuric ion resistance in extremophilic microorganisms [9, 10]. Given the wide spread of mercuric ion resistance/tolerance throughout the bacterial kingdom, it seems likely that many extremophilic microorganisms will encode mercury resistance. This paper describes the investigation of such resistances. Here we report the isolation of thermophilic Bacillus sp. [11] and Ureibacillus sp. [12] and characterisation of their observed mercuric ion resistance. 2.

MATERIALS AND METHODS

2.1 Isolation and identification of organisms from compost Suspensions of garden compost were plated onto LB agar plates supplemented with 0, 15 and 20 µg/ml HgCl2, and the plates were incubated at 62°C for 48 hours. Colonies were picked from the LB agar plates containing mercury and repeatedly subcultured on LB agar containing 15 µg/ml HgCl2 to obtain two axenic cultures. Approximately 700bp of 16S rDNA gene was amplified from each organism using primers pA [13] and 530r [14]. The PCR products obtained were sequenced and subjected to BLAST analysis. 2.2 HgCl2 resistance testing Minimum Inhibitory Concentration (MIC) assays for HgCl2 were performed on Ureibacillus sp. and Bacillus sp. The controls used were a mercuric ion sensitive strain, Bacillus pallidus H12 DSM3670, and a known mer operon-containing mercury resistant Bacillus cereus RC6067 [6]. MIC assays were performed on LB agar plates containing 1138

Microbiology Fundamentals

HgCl2 in 10 µg/ml increments from 0µg/ml to 100µg/ml. Plates were incubated at 37°C and 62°C. 2.3 Identification of culture precipitate It was noted that when Bacillus sp. and Ureibacillus sp. were grown in culture with HgCl2 a black precipitate was formed. Bacillus sp. cells were grown with and without HgCl2 followed by transmission electron microscopy (TEM) to localise the precipitate. In order to test whether the precipitate could be HgS formed by the reaction between Hg(II) and microbially produced H2S [15], [16], H2S production was determined. The positive control used was Proteus vulgaris, a known aerobic H2S producer, with Escherichia coli as the negative control. P. vulgaris, E. coli and B. cereus RC607 were grown overnight at 37°C in LB broth, and B. pallidus H12 DSM 3670, Ureibacillus sp. and Bacillus sp. were grown overnight at 62°C in LB broth. Overnight cultures were diluted 1:100 into sterile glass universal bottles containing 5ml volumes of fresh LB broth with 0µg/ml and 10 µg/ml HgCl2 (only used for mercury resistant strains). Lead acetate strips (BDH) were folded over the neck of the bottles (care was taken not to wet the strips) and the caps were screwed on. These were then incubated at 37°C or 62°C (as above) and examined after 24h and 48h. In order to identify the precipitate, Bacillus sp. cells were grown in the presence of 50 µg/ml HgCl2 for 48 hours. Cells and precipitate were harvested by centrifugation and separated by careful washing with sterile distilled water. The precipitate was dried and examined by X-ray diffraction (XRD) analysis. 2.4 Mercury volatilisation assays Volatilisation assays were performed essentially by the method of [17] with the exception that LB broth was used instead of M9 medium and induction was performed with 2.5 µg/ml HgCl2 for mercury resistant isolates and 0.25 µg/ml for sensitive isolates. Cultures were grown to OD600 = 0.4-0.5, then preinduced with HgCl2 for one hour. Cultures were adjusted to precisely OD600 = 0.4 followed by washing and resuspension of cells in fresh LB broth. Three 1ml aliquots of the resuspended cells were dispensed into separate 50ml conical flasks. Three control flasks each containing 1ml uninoculated LB broth were prepared. Flasks were placed in a shaking waterbath set to the desired temperature and to each flask was added 4ml pre-warmed LB broth supplemented with 1.25µM HgCl2 and 1:2000 diluted 203HgCl2 (at a concentration of 2.55mg/ml Hg; Amersham Pharmacia plc). Timing commenced upon the removal of the first sample. 100 µl samples were taken from each flask at 0, 20, 40, 60, 120, 180 and 240 minutes. These samples were immediately added to 5ml volumes of scintillation cocktail (OptiPhase "HiSafe" 2) and counted in a Packard TR1-CARB 2700 TR liquid scintillation analyser with the window set at 0.0-220 MeV. 2.4.1 Mercuric reductase assays Mercuric reductase assays were performed by the method of Lund and Brown [18]. Cells were grown overnight, then inoculated 1:25 into fresh LB broth without selection. Cells were grown at the appropriate temperature (45°C or 62°C) with shaking to OD600 = 0.5-0.6, induced with 10 µg/ml HgCl2 for one hour, then a further 10 µg/ml HgCl2 for a further hour (mercury sensitive B. pallidus H12 DSM 3670 cells were induced with 1 µg/ml HgCl2 for one hour followed by a further 1 µg/ml HgCl2 for a further hour). Cells were harvested by centrifugation, then washed in 20 ml wash buffer (30mM Tris-HCl 1139

Microbiology Fundamentals

pH7.6, 20% sucrose, 1mM DTT, 1mM EDTA). The cell suspension was re-centrifuged and the supernatant discarded. Cell pellets were washed twice with 20ml resuspension buffer (100mM sodium phosphate pH7.0, 1mM β-mercaptoethanol, 0.5mM EDTA, 0.05mM PMSF, 0.1mM benzamidine) followed by resuspension in 5ml fresh buffer. The cells were lysed by sonication and cell debris was pelleted by centifugation. Supernatants were retained and the pellets discarded. The supernatant was clarified by two further rounds of centrifugation in order to remove all traces of cellular debris. Lysates were stored at 0°C overnight. The following day, 7ml aliquots of assay buffer (0.1M sodium phosphate buffer pH 7.0, 1mM β-mercaptoethanol, 0.5mM EDTA, 0.1mM NADPH) were dispensed into sterile glass universal bottles and stored on ice until use. For the assay, 70µl of cell extract was added to 7ml assay buffer, mixed well and the mixture was preincubated at the desired temperature for 5 minutes. 1ml was removed and an A340 reading was taken. This was immediately followed by addition of 60µl of 10mM HgCl2 stock, which gave a final concentration in the assay mix of 100µM HgCl2; 60µl sterile distilled water was added to the control assay mix. Samples were mixed well and incubated at the desired temperature for 1 minute followed by an A340 reading. Further readings were taken at 3 and 5 minutes after addition of HgCl2/sterile distilled water. 3.

RESULTS AND DISCUSSION

3.1 Identification of microorganisms The 16S rDNA PCR products obtained from the two cultures were sequenced and BLAST analysis of the sequences obtained revealed 97% homology with Bacillus pallidus H12 DSM3670 [11] and >99% homology with Ureibacillus thermosphaericus [12]. 3.2 HgCl2 resistance Table 1. HgCl2 MIC values (µg/ml) for Bacillus sp. and Ureibacillus sp. Organism

HgCl2MIC (mg/ml) at 37°C

HgCl2MIC (mg/ml) at 62°C

Bacillus cereus RC607 Bacillus pallidus DSM 3670 Bacillus sp. Ureibacillus sp.

60 10 60 30

10 80 30

No data are shown for Bacillus cereus at 62°C as this temperature was outside the growth range for this organism. The results obtained from MIC testing showed that mercury resistance of the thermophilic isolates was not temperature dependent. The growth of Bacillus sp. was reduced on normal LB agar at 37°C; this temperature is near to the bottom end of the organism's growth range, which may help explain why mercury resistance was reduced, but not abolished. 3.3 TEM of Bacillus sp. The results of Transmission Electron Microscopy (TEM) (Figure 1) show a clear difference in the appearance of bacterial cultures grown with and without mercury. There is a large amount of black precipitate visible when Hg(II) is present, which was thought to be HgS. The appearance of the cells in the TEM suggests that the precipitate may be forming within the cells, possibly causing cells to rupture. 1140

Microbiology Fundamentals

Figure 1(a). TEM (10,000 X magnification) of Bacillus sp. grown with 50µg/ml HgCl2

Figure 1(b): TEM (10,000X magnification) of Bacillus sp. grown without HgCl2

3.4 Identification of the precipitate XRD analysis was performed upon a sample of dried crystalline precipitate collected from a culture of Bacillus sp. grown with HgCl2. The thick black lines on the XRD profile (Figure 2) show the peaks obtained for the experimental sample, and the fainter, vertical lines show the peaks obtained from a database for an ultra-pure reference sample of βHgS. There was sufficiently good correlation between the two sets of information to make a positive identification of the experimental sample as β-HgS. The profile was very similar to one obtained for HgS precipitated by Desulfovibrio desulfuricans [19].

1141

Microbiology Fundamentals

Figure 2. X-Ray Dispersion Analysis of the precipitate from Bacillus sp. grown in the presence of Hg(II) 3.5 Mercury volatilisation assays The results of mercury volatilisation experiments (Figure 3) showed that Ureibacillus sp. removed Hg(II) from the media more efficiently than Bacillus sp. at 45°C than at 62°C. This may reflect the growth temperature optima for these organisms. From 16S rRNA data, the most closely related species to Bacillus sp. was Bacillus pallidus, which is reported as having a maximum growth temperature of 70°C with a temperature optimum of 60-65°C. The closest species to Ureibacillus sp. was Ureibacillus thermosphaericus which has a maximum growth temperature of 64°C. Mercury Volatilisation at 45oC

Mercury Volatilisation at 62oC

100

100 %Hg remaining

120

%Hg remaining

120

80 60 40

B.cereus RC607

80

Bacillus sp. Ureibacillus sp.

60

Bacillus pallidus H12 DSM 3670 Luria broth

40 20

20

0

0 0

50

100 150 Time (min)

200

0

50

100 150 Time (min)

200

Figure 3: Results of volatilisation assays on Bacillus spp. and Ureibacillus sp. at 62°C and 45°C. Results are shown as the means of triplicate readings with error bars shown as sample standard deviations (σn-1)

This suggests that Ureibacillus sp. probably has a lower optimum growth temperature than Bacillus sp., hence Ureibacillus sp. may remove 203Hg(II) more efficiently at 45°C. At 45°C, 203Hg(II) was being removed from the media at a high rate by B. cereus RC607. 1142

Microbiology Fundamentals

This rate was reduced at 62°C, probably due to the fact that 62°C was outside the growth range of this organism. There appeared to be no appreciable difference in the rates of removal in LB broth alone and in the presence of B. pallidus H12 DSM3670 at 45°C. There did appear to be some removal of 203Hg(II) from the media by B. pallidus H12 DSM3670 at 62°C, the reason for which is not known. 3.6 Mercuric reductase assays The results of mercuric reductase assays are shown in Figure 4. Mercuric reductase activity was observed in B. cereus RC607 at 37°, 45° and 62°C. There appeared to be no specific NADPH-dependent mercuric reductase activity in B. pallidus H12 DSM3670, Ureibacillus sp. or Bacillus sp. This result was surprising in view of the mercuric ion resistance and results from volatilisation assays, which showed Ureibacillus sp. and Bacillus sp. as capable of removing 203Hg(II) from the medium. An alternative, non-mer mechanism of mercury detoxification and volatilisation may operate within Ureibacillus sp. and Bacillus sp., possibly as a by-product of normal metabolism.

90 o

37 C with HgCl2

nmoles NADPH oxidised/min/mg of protein

80

o

45 C with HgCl2 o

62 C with HgCl2

70

o

37 C without HgCl2 o

45 C without HgCl2

60

o

62 C without HgCl2 50 40 30 20 10 0 B.cereus RC607

Bacillus sp.

Ureibacillus sp.

B.pallidus H12 DSM3670

Figure 4. Mercuric reductase activity of Bacillus spp. and Ureibacillus sp. at 37°C, 45°C and 62°C. Errors bars show one standard deviation (σn-1) of triplicate readings 3.7 Testing for H2S production The precipitation of HgS by microorganisms is often due to a reaction of Hg(II) with H2S gas. The strains in this study were tested for H2S production as described in section 2.3. Growth occurred in all bottles; however, blackening of the lead acetate strip only occurred in the bottle containing P. vulgaris, indicating that only this isolate produced detectable amounts of H2S. It seems likely, therefore, that any production of HgS by Ureibacillus sp. and Bacillus sp. proceeded due to complexation of Hg(II) with cellular sulphides and not with H2S. 1143

Microbiology Fundamentals

4.

CONCLUSIONS Mercuric ion tolerant isolates of thermophilic Bacillus sp. and Ureibacillus sp. were isolated from compost. Mercury tolerance in these isolates was apparently not due to classical mer operon-mediated mercury reduction, but by some other mechanism. Mercury tolerance was not temperature dependent, as both isolates showed elevated tolerance to mercuric ions at 37°C and 62°C. When considering features of mercuric ion resistance in these isolates it is important to make the distinction between mercury resistance and mercury tolerance. Mercuric ion resistance can be thought of as a genetically encoded detoxification mechanism, which is specifically induced in response to mercurials. Mercury tolerance can be thought of as a detoxification mechanism, which is a by-product of normal metabolism and is not specifically induced [19]. From the results obtained it seems possible that Ureibacillus sp. and Bacillus sp. are exhibiting mercury tolerance mechanisms rather than a specific mer operon-encoded mercury resistance. The evidence to support this is the lack of detectable NADPH-dependent mercuric reductase activity, and precipitation of HgS in the medium. Several organisms have been reported as being mercury tolerant due to HgS precipitation [15, 16, 20, 21]. This mechanism has not been shown to be specifically induced in response to mercurials, suggesting that it is indeed a by-product of normal metabolism. HgS can be formed by various mechanisms in microbial mercury cycling. The growth of Desulfovibrio desulfuricans under sulphate-reducing conditions resulted in the precipitation of HgS when the medium was spiked with Hg(II) [15, 19]. Hg(II) reacts directly with H2S produced by D. desulfuricans [15] and Clostridium cochlearium [16] to form HgS. This is thought to provide a means of mercury tolerance in these organisms. The formation of HgS from monomethylmercury (MMHg) by D. desulfuricans may occur via the reaction of MMHg with H2S to produce unstable dimethylmercurysulphide (DMHgS), which breaks down to HgS and volatile dimethylmercury (DMHg) [15]. Precipitation of HgS by aerobic microorganisms has also been suggested as a tolerance mechanism. Klebsiella aerogenes NCTC418 was thought to form HgS when grown in continuous (aerobic) culture with the addition of 2ppm (equivalent to 2µg/ml) HgCl2. No bacterial mercury volatilisation could be detected, and it was claimed mercuric ion sensitivity was increased under sulphate-limited conditions [20]. K. aerogenes is unable to produce H2S gas, therefore any precipitation of HgS presumably occurred via an alternative mechanism. There are few reports of precipitation of HgS by bacterial cultures, which makes it of unknown importance in microbial mercury detoxification. At least part of the mercury tolerance shown by Ureibacillus sp. and Bacillus sp. may be due to precipitation of HgS. This probably occurs by reaction of Hg(II) with cellular sulphides, as H2S was not produced by these microorganisms when grown with or without HgCl2. The results of mercuric reductase assays showed that no NADPH-dependent mercuric reductase activity was present in Ureibacillus sp. and Bacillus sp., therefore mercury removal was not due to reduction of Hg(II) to Hg(0) by mercuric reductase, unless it uniquely uses another cofacter. It is possible that Ureibacillus sp. and Bacillus sp. may produce high levels of non-specific reductases which are capable of reducing Hg(II) to Hg(0). Alternatively, there may be a different, as yet unknown, mechanism for mercury removal operating in these organisms. A possible explanation for removal of mercury from the media by Ureibacillus sp. and Bacillus sp. may be the methylation of mercury to either MMHg or DMHg. MMHg 1144

Microbiology Fundamentals

and DMHg are volatile, and could be lost from the medium in a similar fashion to Hg(0) The formation of DMHg by microorganisms may represent a mercury detoxification method due to the volatility of the compound [19]. Methylation of mercury by anaerobic microorganisms has been reported [22]. At present there are no reliable reports of mercury methylation by aerobic cultures. The mechanisms of mercury tolerance in Ureibacillus sp. and Bacillus sp. are not fully understood. Whilst precipitation of HgS in culture is one tolerance mechanism, the means by which mercury volatilisation occurs in the absence of specific mercuric reductase is unclear. ACKNOWLEDGEMENTS This work was supported by a BBSRC CASE studentship to KJG. Thanks are due to Dr Jon L. Hobman and Dr Ching Leang for discussion and advice in setting up assay systems. REFERENCES

1. 2. 3. 4. 5. 6.

Gutknecht, J. J. Memb. Biol. (1981). 61:61. Leach, S.J. J. Aust. Chem. (1960). 13:520. Sletten, E. and Nerdal, W. Metal Ions In Biological Systems. (1997). 34:479. Moore, B. Lancet. (1960). II:453. Hobman, J.L. and Brown, N.L. Metal Ions In Biological Systems. (1997). 34:527. Wang, Y., Moore, M.,Levinson, H.S., Silver, S., Walsh, C. and Mahler, I. J. Bacteriol. (1989). 171:83. 7. Sedlmeier, R. and Altenbuchner, J. Mol. Gen. Genet. (1992). 236:76. 8. Laddaga, R.A., Chu, L., Misra, T.K. and Silver, S. Proc. Natl. Acad. Sci. (1987). 84:5106. 9. Olson, G.J., Porter, F.D., Rubinstein, J. and Silver, S. J. Bacteriol. (1982). 151:1230. 10. Bogdanova, E.S., Mindlin, S.Z., Kalyaeva, E.S. and Nikiforov, V.G. FEBS Lett. (1988). 234:280. 11. Scholz, T., Demharter, W., Hensell, R. and Kandler, O. Syst. Appl. Microbiol. (1987). 9:91. 12. Fortina, M.G., Rüdiger, P., Schumann, P., Mora, D., Parini, C., Manachini, L.P. and Stackebrandt, E. Int. J. Syst. Evol. Microbiol. (2001). 51:447 13. Edwards, U., Rogall, T., Blocker, H., Emde, M. and Bottger, E.C. Nucleic Acids Res. (1989). 17:7843. 14. Lane, D.J., Pace, B., Olsen, G.J., Stahl, D.A., Sogin, M.L. and Pace, N.R. Proc. Natl.Acad. Sci. (1985). 82:6955. 15. Baldi, F., Pepi, M. and Filippelli, M. Appl. Environ. Microbiol. (1993). 59:2479. 16. Pan-Hou, H. and Imura, N. Arch. Microbiol. (1981). 129:49. 17. Leang, C., PhD thesis (1999), University of Birmingham: Birmingham, UK. 18. Lund, P.A. and Brown, N.L. Gene. (1987). 52:207. 19. Baldi, F. Metal Ions In Biological Systems. (1997). 34:213. 20. Aiking, H., Govers, H. and van 't Riet, J. Appl. Environ. Microbiol. (1985). 50:1262. 21. Belliveau, B.H., Starodub, M.E. and Trevors, J.T. Can. J. Microbiol. (1991). 37:513. 22. Choi, S.C. and Bartha, R. Appl. Environ. Microbiol. (1993). 59:290.

1145

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Reduction of Pd(II) with Desulfovibrio fructosovorans, its [Fe]-only hydrogenase negative mutant and the activity of the obtained hybrid bioinorganic catalysts I.P. Mikheenkoa, V.S. Baxter-Planta, M. Roussetb, S. Dementinb, G. Adryanczyk-Perrierb and L.E. Macaskiea a

School of Biosciences, the University of Birmingham, Edgbaston, Birmingham, B15 2TT, England, United Kingdom b Bioenérgétique et Ingénierie des Protéines, CNRS, 31 Chemin Cedex 20, Joseph Aiguier, 13402 Marseille France Abstract A novel biocatalyst was obtained via reduction of Pd(II) to Pd(0), at the expense of hydrogen, onto the cell surface of Desulfovibrio fructosovorans and a mutant of D. fructosovorans with deactivated [Fe]-only hydrogenase. The ability of the palladiumcoated biomass to reduce chromium (Cr+6 to Cr+3), to release hydrogen from sodium hypophosphite and reductively dehalogenate chlorophenol and polychlorinated biphenyl congeners was demonstrated. Dried, palladised cells of D. fructosovorans wild type and its mutant were more effective catalysts than Pd(II) reduced chemically under hydrogen or commercially available sub-micron size Pd(0) powder. Differences were observed in the catalytic activity of the wild type and the mutant strain of D. fructosovorans when compared with each other. Negligible chloride release occurred from the chlorophenol and polychlorinated biphenyl species using biomass alone. The structure of the bioinorganic catalyst was investigated using transmission electron microscopy. 1.

INTRODUCTION Palladium is one of the world’s most expensive metals. Catalyst systems based on palladium are widely used in chemical manufacturing and processing (1). The highest palladium consumer is the auto motive industry, where this metal, together with other platinum group metals, is used in automobile catalytic converters to reduce the toxicity of exhaust gases. Palladium is also extensively used in the electronic industry. Since palladium is a highly valuable metal with only limited world resources (2), developing new methods of Pd recovery and reprocessing of scrap is necessary since existing hydroand pyrometallurgical routes are energy demanding and/or not environmentally friendly. In the work of Lloyd et al. (3) it was shown that palladium could be effectively recovered from solution by resting cells of the sulphate-reducing bacterium Desulfovibrio desulfuricans at ambient temperature. It was concluded that Pd(II) was reduced by the activity of hydrogenase and possibly cytochrome c3. Yong et al. (4) demonstrated that palladised D. desulfuricans biomass could be used as a catalyst without any further processing. It has also been reported (5, 6) that in the 1147

Microbiology Fundamentals

process of biorecovery metallic Pd with unusual magnetic properties could be obtained: ferromagnetic Pd nanocrystals were formed on the biomass. Subsequent studies showed (7) that biologically reduced palladium (Bio-Pd) can be used as a catalyst for partial dehalogenation of polychlorinated biphenyls, which are very stable recalcitrant xenobiotics which are not generally susceptible to aerobic or anaerobic microbial metabolization. The catalytic properties of biomineralized palladium (Bio-Pd) depend on the size and the distribution pattern of metal deposits on the cell surface. To obtain a Bio-Pd catalyst with certain catalytic properties it is important to understand the process of Pd(II) reduction and Pd(0) grain formation which take place within the periplasm of sulphate reducing bacteria (5). According to Lloyd et al. (3) hydrogenase is one of the important components of the Pd(II) reducing system. Sulphate-reducers have more than one hydrogenase system, with Desulfovibrio sp. possessing up to four different hydrogenases, which differ by structure (enzymes with Fe, Fe-Ni or Fe-Ni(Se) in their active centres) and localisation within the cell (periplasmic and cytoplasmic, membrane bound and nonbound) (8). Several hydrogenase mutants of Desulfovibrio sp. have been constructed lacking different hydrogenases (9, 10). The aim of this work is to investigate the catalytic properties of Bio-Pd obtained via Pd(II) reduction by the sulphate reducer D. fructosovorans and its [Fe]-only hydrogenase mutant. The hydrogenase and cytochrome complex of the Desulfovibrio sp. has been intensively investigated (8, 11, 12 ). 2.

EXPERIMENTAL

2.1 Preparation of cells D. fructosovorans and its mutant (10) were maintained anaerobically (13). Growth medium was inoculated with 10% of culture and cells were incubated at 37 °C for three days before harvesting. The medium for D.fructosovorans culture was as in (13). For Bio-Pd production the cells were harvested by centrifugation at 6000 rpm for 15 min at 4°C, washed three times in degassed 20 mM MOPS buffer at pH 7.0. The cells were stored refrigerated. 2.2 Preparation of Bio-Pd Aliquots of the washed cells were resuspended as a concentrated suspension (diluted to 4-6 mg dry wt/ml for use) in MOPS buffer and transferred under N2 to 100 ml of Pd(II) (2 mM sodium tetrachloropalladate(II) Na2PdCl4 (Aldrich Chemical Company) in 0.01 M HNO3, pH 2.3) solution in 100 ml serum bottles sealed with butyl rubber stoppers and preequilibrated with N2. The suspension was allowed to stand for 1 h at 30°C to form nucleation sites on the biomass. The electron donor used was H2 gas, which was bubbled through the solution for 20 min. After the black precipitate had formed, the solution was centrifuged at 6000 rpm for 10 min and the precipitate was washed three times with distilled water, dried in air and then washed with acetone. Chemical palladium was prepared identically except that no biomass was added and reduction took approximately 60 min under H2 in the absence of cells. 2.3 Pd(II) and Cr(VI) assay The kinetics of Pd(II) reduction were studied by recording the A420 of the residual Na2PdCl4 solution, withdrawn (1.5 ml) from the reaction mixture each 20 min, (Unicam 5600 UV-VIS spectrophotometer: Cambridge, UK). The maximum height of the specific 1148

Microbiology Fundamentals

absorption peak of [PdCl4]2- at 420 nm depends linearly on the palladium salt concentration. The concentration of Pd(II) in solution was also determined spectrophotometrically with Sn(II) (14). For the Pd(II) assay 200 µl of sample was added to 800 µl of SnCl2 solution. Readings were taken after 60 min at a wavelength of 463 nm. The residual concentration of Cr(VI) (where Pd(0) was used as a catalyst for the reduction of Cr(VI) (below)) was measured using the diphenylcarbazide method (5). Test solution (100 µl) was added to 1.4 ml of the diphenylcarbazide solution, mixed well, left to stabilise for 5 min and read at 540 nm. 2.4 Examination of Pd-loaded cells by electron microscopy Pellets of Pd-loaded bacteria were harvested using a Heraeus Sepatech Biofuge 13 microfuge (13000 rpm, 5 min) and fixed in 2.5% (wt/vol) glutaraldehyde in 0.1 M Nacacodylate buffer, pH 6.8. After centrifugation the pellet was resuspended in 1.5 ml of 0.1 M Na-cacodylate buffer, pH 6.8. The preparation was stained in 1% osmium tetroxide in 0.1 M phosphate buffer, pH 7.0, dehydrated and embedded in epoxy resin. Sections (100 150 nm thick) were cut from the resin block and placed onto a copper grid prior to analyses. Sections were viewed with a JEOL 120CX2 transmission electron microscope. 2.5 Catalytic activity measurement Three test reactions were used to characterise the catalytic activity of the Bio-Pd: hydrogen release from sodium hypophosphite, Cr(VI) reduction and reductive dehalogenation (RD) of polychlorinated aromatic compounds. 2.5.1 Hydrogen release from sodium hypophosphite The reaction was initiated via the addition of 4 mg of Bio-Pd (Pd:dry wt biomass 1:3) to 20 ml of a 2% solution of hypophosphite (NaH2PO2) at 30°C. The volume of liberated hydrogen was recorded at 5 min intervals starting 20 mins after the initiation of the reaction (reaction onset time). The reaction rate was calculated as the rate of H2 release from 5 replicate experiments. Chemical palladium (4 mg) was used as a control test reaction. 2.5.2 Cr(VI) reduction Tests were performed under oxygen free nitrogen (OFN) in 12 ml sealed serum bottles. Bio-Pd (2 mg) was added to 5 ml of 7 mM Na2CrO4 in 20 mM MOPS-NaOH buffer at pH 7.0. The solution was equilibrated at 30°C for 10 min and the reaction was initiated by addition of sodium formate (electron donor) to a final concentration of 25 mM. Samples were taken from the bottle via a rubber septum and centrifuged for 5 min at 13000 rpm to remove Pd. The supernatant was analysed for residual Cr (VI). The catalytic activity was expressed as the percent of reduced Cr(VI). Chemical palladium (2 mg) was used as a control test reaction. 2.5.3 Reductive dehalogenation of chlorinated aromatic compounds The Bio-Pd (2 mg, 1:3 Pd:dry wt biomass) was placed in 12 ml serum bottles and the chlorinated aromatic compound added to the required concentration in a carrier of 20 mM MOPS NaOH buffer pH 7.0. Experiments were initiated via the addition of formate (10 mM sodium formate, pH 7.3). The final volume was 10ml. Dehalogenation of the chlorinated aromatic substance was estimated in periodically withdrawn samples as release of chloride ion assayed spectrophotometrically by the mercury (II) thiocyanate method (15) versus sodium chloride as standard. For these tests the washing procedures 1149

Microbiology Fundamentals

were carried out during preparation of the Bio-Pd with omission of Cl- from all processing steps to ensure a low background level of Cl-, against which the Cl- release was measured. The limit of sensitivity was in the range 0.5-100 µg/ml Cl-. 2.5.4 Chlorinated aromatic compounds Chlorophenols were used as aqueous solutions diluted to the required concentration (usually 5 mM) in the media. Polychlorinated biphenyls (PCBs) were dissolved in hexane and diluted to the required nominal concentrations (80 µg/ml as chloride; shaken to mix) as hexane suspensions in the media. 3.

RESULTS AND DISCUSSION The bioreduction of Pd(II) by D. fructosovorans and its [Fe]-only hydrogenase mutant was compared. Solutions (2 mM) of sodium tetrachloropalladate (Na2PdCl4) in 0.01 M HNO3 or tetraamminepalladium chloride ([Pd(NH3)4]Cl2) in 20 mM MOPS buffer, pH 7.0 were poured into 10 ml sealed vials and degassed under OFN. The resting cell suspensions of each strain were added to the Pd(II) solution (3 parts dry weight of biomass to 1 one part of Pd(II) and left for biosorption at 30°C. After 1 h samples of supernatant were analysed for remaining Pd(II). This concentration of Pd(II) was regarded as the time-zero concentration. Following the one hour of biosorption the Pd(II) bioreduction was initiated via the addition of electron donor (20 mM sodium formate solution, pH 7.0). The reaction was carried out anaerobically at 30°C. The residual Pd concentration in the supernatant is shown in Fig. 1.

Figure 1. Reduction of Pd(II) by D. fructosovorans, wild type and [Fe]-only hydrogenase negative mutant. Mass ratio Pd(II) dry biomass is was 1:3. Solution: 2 mM Na2PdCl4, in 0.1 M HNO3. Electron donor 0.25 mM sodium formate

The rate of Pd(II) reduction was monitored via the decreasing concentration of the coloured [PdCl4]-2 complex ion (A420), and confirmed using SnCl2. Both methods gave identical results. The reduction of Pd by sodium formate alone was used as a control. Dead cells (killed by autoclaving at 121˚C for 15 mins) did not reduce Pd (not shown). Fig 1 shows that both wild type and mutant reduce Pd(II) at approximately the same rate hence the absence of [Fe]-only hydrogenase did not significantly effect the Pd(II) reduction process. It should be noted that [Fe]-only periplasmic hydrogenase is less abundant 1150

Microbiology Fundamentals

mM reduced Cr(VI)/mg Pd.h

mM reduced Cr(VI)/mg Pd.h

compared to the [Fe-Ni] enzyme, which constitutes a larger part of hydrogenase activity in the periplasm (8). The catalytic activity of Bio-Pd samples obtained using D. fructosovorans and its mutant with respect to Cr(VI) reduction was substantially higher when compared to palladium powder of 39 µm particles size (Fig. 2). The Bio-Pd obtained under acidic conditions (pH 2.3) using Na2PdCl4 salt (Fig.2a) proved to be a better catalyst when compared to the catalyst obtained at pH 7.0 using [Pd(NH3)4]Cl2 salt (Fig.2b). Interestingly, Bio-Pd obtained with [Fe]-only hydrogenase negative mutant was significantly more catalytically active when compared to Bio-Pd obtained from the wild type regardless of which palladium salt or pH was used in the preparation of Bio-Pd (Fig. 2). 2,5 2,0

a

1,5 1,0 0,5 0,0

1

2

0,6 0,5

b

0,4 0,3 0,2 0,1 0,0

3

1

2

3

1 - Control: commercial Pd powder ; Bio-Pd: 2 - wild type; 3 - [Fe]-only hydrogenase negative mutant.

Figure 2. The reduction of Cr(VI) to Cr(III) using D. fructosovorans and its [Fe]-only hydrogenase negative mutant Bio-Pd produced with Na2PdCl4 salt, pH 2.3 (a); and [Pd(NH3)4]Cl2 salt, pH 7.0 (b)

The Bio-Pd was also a more active catalyst than its chemical counterpart in the release of hydrogen from sodium hypophosphite (Fig. 3). The rate of hydrogen release was higher in the presence of mutant-based Bio-Pd catalyst obtained with Na2PdCl4 salt (made at pH 2.3) (Fig. 3a) when compared to wild type-based Bio-Pd. However, in the case of Bio-Pd obtained using [Pd(NH3)4]Cl2 salt made at pH 7.0 (Fig. 3b) the difference between the two types of biocatalyst was not significant. 0,14

a

0,12

ml H2/mg Pd.min

. ml H2/mg Pd min

0,30 0,25 0,20 0,15 0,10 0,05

b

0,10 0,08 0,06 0,04 0,02 0,00

0,00

1

2

3

1

2

3

1 - Control: commercial Pd powder ; Bio-Pd: 2 - wild type; 3 - [Fe]-only hydrogenase negative mutant.

Figure 3. The release of hydrogen from sodium hypophosphite using D. fructosovorans and its [Fe]-only hydrogenae negative mutants produced Bio-Pd with Na2PdCl4 salt, pH 2.3 (a); and [Pd(NH3)4]Cl2 salt, pH 7.0 (b) bars. Bars indicate standard error 1151

Microbiology Fundamentals

The reductive dehalogenation of chlorinated aromatic compounds was investigated. Chloride release from 4-chlorobiphenyl, 2,3,4,5-tetrachlorobiphenyl and 2,2’,4,4’,6,6’hexachlorobiphenyl using Desulfovibrio fructosovorans and its hydrogenase negative mutant Bio-Pd produced with Na2PdCl4 salt (pH 2.3) was tested. Throughout, no Clliberation was promoted by non-palladised bacteria, chemically prepared Pd(0) or commercially available finely divided Pd(0) (not shown) although the Bio-Pd promoted Cl- release from all the PCBs (Figs. 4a-c). However only small differences were seen between the parent and the mutant Bio-Pds in each case. 16

b

a Chloride, µg/ml

Chloride, µg/ml

16

12 8 monoPCB wild type monoPCB [Fe]-negative hydrogenase mutant

4

1

2

3

4

12

8 TetraBCB wild type TetraPCB [Fe]-negative hydrogenase mutant

4

1

5

2

Time, hours

20

4

5

16

c

12 8 HexaPCB wild type HexaPCB [Fe]-negative hydrogenase mutant

4 1

2

3

Time, hours

4

5

Chloride, µg/ml

d

16

Chloride, µg/ml

3

Time, hours

12

8 2-chlorophenol wild type 2-chlorophenol [Fe]-negative

4

hydrogenase mutant

1

2

3

4

5

Time, hours

Figure 4. Reductive dehalogenation of 4-chlorobiphenyl (a) and 2,3,4,5tetrachlorobiphenyl (b) of 2,2’,4,4’,6,6-hexachlorobiphenyl (c) and 2-chlorophenol (d) using D.fructosovorans and its hydrogenase negative mutants Bio-Pd

The rate of chloride release from 5 mM 2-chlorophenol was also studied under similar conditions. In this case the rate of Cl- release for the [Fe]-only hydrogenase mutant was lower than for the wild type (Fig. 4d). The controls included resting cells of the bacteria without bound Pd(0), chemically prepared Pd(0) and commercially available finely divided Pd(0). In confirmation of a previous study the chemically prepared Pd(0) and commercially available finely divided Pd(0) gave a similar rate of Cl- release to D. fructosovorans, whereas no Cl- liberation was promoted by the non-palladised bacteria (not shown). Thus, the release of Cl- from 2-chlorophenol was achieved using both Bio-Pd preparations, with no advantage over "chemical" Pd(0). However, the reductive dehalogenation of polychlorinated biphenyls was achieved using the Bio-Pd obtained from D. fructosovorans and its hydrogenase negative mutant under conditions where the chemically produced Pd counterpart is largely ineffective, but no advantage was demonstrated using the mutant strain. The bioreduction tests and studies using X-ray diffraction analysis (not shown) showed that D. fructosovorans and its [Fe]-only hydrogenase negative mutant reduced 1152

Microbiology Fundamentals

palladium to its metallic state on their surface. Electron microscopic investigation of BioPd in this study showed that for D. fructosovorans and its mutant the initial reduction of Pd takes place in the same way as that previously described for in D. desulfuricans (5). It begins within the periplasmic space (the site of localisation of hydrogenases and cytochromes) and then the growing crystals protrude through the outer cell membrane and form larger crystals on the cell surface (Fig. 5). The pattern of Pd crystal deposition was very similar in both the wild type and mutant stain. The catalytic activity of Bio-Pd obtained using the [Fe]-only hydrogenase negative mutant was higher when compared to the Bio-Pd obtained using wild type strain in the reduction of Cr(VI) and in the other tests the difference varied according to the Pd(II) salt used and pH of its preparation (hydrogen release tests) or chlorinated aromatic substrate (RD tests). We assume that the reason for differences in catalytic activity is the size and, consequently, the surface area of palladium crystals formed in the process of Pd(II) reduction. It may be possible that the lack of one of the hydrogenases leads to formation of a larger quantity of Pd(0) nanoclusters via the activity of other redox-enzymes within the periplasm. These nanoparticles are not visible on the electron micrographs but the presence within the preparation can be detected through their enhanced catalytic activity. The presence of such Pd nanoclusters was shown in the Bio-Pd preparations of D. desulfuricans via the measurement of the magnetic moment of Bio-Pd powder in a varying external magnetic field (5, 6); magnetic measurements to determine the Pd-cluster size on the wild type and mutant D. fructosovorans are in progress.

Figure 5. Electron micrographs of Pd deposition on the surface of D. fructosovorans (a) and its mutant (b) cell surface. Scale bars indicate 500 nm 4

CONCLUSIONS The catalytic activity of Bio-Pd obtained under different conditions using D. fructosovorans and its hydrogenase negative mutant vary in absolute values although shares a general trend. It was significantly higher in all tests (except RD of chlorophenol) when compared to chemical palladium. The [Fe]-only hydrogenase negative mutant showed better catalytic activity compared to the wild type in most reactions. This provides evidence that, firstly, hydrogenase is not the only enzyme involved in Pd(II) reduction, and, secondly, it shows that the mutation has an additional effect, which causes variations in Bio-Pd catalytic activity. AKNOWLEDGEMENTS We acknowledge the financial support of the EU BIO-CAT (Grant number GRD12001-40424) and the BBSRC (Grant number E13817) and EPSRC (Grant NoGR/NO8445). 1153

Microbiology Fundamentals

REFERENCES

1. 2. 3. 4.

Platinum 2002 Interim Review. Johnson Matthey. November (2002) 28. P. Crowson. Minerals Handbook. Macmillan Ltd., Surrey, 1982. J.R. Lloyd, P. Yong, L.E. Macaskie. Appl. Environ. Microbiol., 64 (1998) 4608. P.Yong, N.A. Rowson, J.P. Farr, I.R. Harris, L.E. Macaskie. Biotechnol. Bioeng., 80 (2002) 369. 5. I. Mikheenko, P. Mikheenko, C.N.W. Darlington, C.M. Muirhead, L.E. Macaskie. In Biohydrometallurgy: Fundamentals, Technology and Sustainable Development Eds V.S.T. Ciminelli & O Garcia, Elsevier (2001) 525. 6. I.P. Mikheenko, L.E. Macaskie, P.M. Mikheenko, C.M. Muirhead. Pd. 19th General Conference of the EPS Condensed Matter Division held jointly with CMMP 2002 – Condensed Matter and Materials Physics. Brighton. Europhysics Conference Abstracts. 26A (2002) 93. 7. V.S Baxter-Plant, I.P. Mikheenko, L.E. Macaskie, Biodegradation (2002) in press. 8. R.Cammack, M. Frey, R. Robson (eds.), Hydrogen as a Fuel. Learning from Nature. Taylor & Francis. London, New York, 2001. 9. L. Casalot, De Luca G, Dermoun Z, Rousset M, de Philip P. J. Bacteriol.,184 (2002) 853. 10. B.K. Pohorelic, J.K. Voordouw, E. Lojou, A. Dolla, J. Harder, G. Voordouw. J. Bacteriol., 184 (2002) 679. 11. C. Aubert, M. Brugna, A. Dolla, M. Bruschi, M-T. Giudici-Orticoni. Biochim. Biophys. Acta - Protein Structure & Molecular Enzymology, 1476 (2000) 85. 12. C. Wawer, G Muyzer. Appl. Environ. Microbiol., 61 (1995) 2203. 13. M. Rousset, L. Casalot, B. J. Rapp-Giles, Z. Dermoun, P. de Philip, J. P. Belaich, J. D.Wall. Plasmid, 39 (1998) 114. 14. G.C. Dasages. Absorptiometriques des elements mineraux. Ed. Masson, Paris, 1978. 15. G.H. Jeffrey, J. Bassett, J. Mendham, R.C. Denny. Vogels textbook of quantitative chemical analysis, Fifth Edition, Bath Press, Avon, UK, 1989.

1154

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Removal of cobalt, strontium and caesium from aqueous solutions using native biofilm of Serratia sp. and biofilm pre-coated with hydrogen uranyl phosphate M. Paterson-Beedle and L.E. Macaskie School of Biosciences, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK Abstract Heavy metals are removed by Citrobacter sp. (NCIMB 40259) now identified as a Serratia sp. via the activity of an acid-type phosphatase enzyme, which liberates HPO42from a suitable organic phosphate donor with the stoichiometric precipitation of heavy metal cations (M2+) as insoluble MHPO4 at the cell surface. Previous studies have shown that it is possible to remove Ni2+ effectively via its intercalation into pre-formed or growing crystals of hydrogen uranyl phosphate (HUP), with the "guest" metal species intercalating within the HUP matrix. Also, the effective biomineralisation of HUP from uranium mine waters has been demonstrated. Nuclear wastes contain not only uranium but also fission products like 90Sr, 137Cs, and 60Co. We have shown that in the presence of uranyl ion and glycerol 2-phosphate the deposited HUP is able to remove "cold" surrogates of the fission products via intercalation using either continuous co-crystal growth or by using cells pre-coated with HUP as inorganic ion-exchangers. Using Serratia sp. biofilm immobilised onto polyurethane reticulated foam continuous removal of fission products surrogates was obtained by intercalation techniques, but not using the phosphatase-biomineralisation route alone, in the absence of the HUP host crystal. 1. INTRODUCTION Fission of 235U yields radioactive fission products such as the isotopes 60Co, 137Cs and 90 Sr. Urgency is prompted by the nuclear industry to treat radionuclide-loaded liquid wastes. Biological methods for removal of metals from nuclear wastes can succeed where traditional physico-chemical treatments fail [1, 2]. A model for phosphatase-mediated uranyl ion accumulation, with exocellular deposition of polycrystalline (i.e., multitude of small crystals [3]) hydrogen uranyl phosphate (HUO2PO4.4H2O, HUP) that is identical to the structure of HUP prepared by inorganic crystal growth has been well published [4]. "Chemical" HUP is an established ion-exchange material, with the intralamellar mobile protons freely exchangeable for other ions like Na+, Ni2+ [5] and Co2+ [6, 7]. Previous studies have demonstrated two approaches for the bioremediation of Ni2+ from dilute aqueous flows using Serratia sp. cells [8-12]. The first approach was based on a two step process: (i) enzymatically-mediated generation of a polycrystalline ‘host lattice’ (priming layer), i.e. HUP bound to the surface of the Serratia sp. cells and (ii) cation-exchange intercalation of Ni2+ into the interlamellar spaces of HUP [8-12]. The second approach [8] was to co-challenge immobilised Serratia cells in a packed-bed reactor with a Ni2+ and 1155

Microbiology Fundamentals

UO22+ solution in the presence of glycerol-2-phosphate (phosphate donor for phosphate release and metal precipitation) giving a sustained removal of both metals. Since Co and Ni are in the same group of the periodic table of the elements they should behave similarly. Therefore, the purpose of this study was to evaluate the methods described above using Serratia sp. biofilm immobilised onto polyurethane foam for the removal of Co2+ from solution. Previous studies showed the uranyl phosphate precipitate to be, more likely, NaUO2PO4.4H2O [13], therefore, Cs+ should be similarly removed as CsUO2PO4 and, in the presence of both fission products surrogates, bio-driven formation of mixed metal uranyl phosphate is distinctly feasible. Previous studies showed removal of Sr2+ as the phosphate precipitate [14] and removal of Sr2+ as a co-crystal was also explored. 2.

MATERIALS AND METHODS

2.1 Support, organism and biofilm production Citrobacter sp. (NCIMB 40259) now identified as Serratia sp. [15] was grown as biofilm onto polyurethane reticulated foam Filtren TM30 (supplied by Recticel, Belgium) in an air-lift fermenter [16] and the phosphatase specific activity of the cells from the fermenter outflow was determined as described previously [16]; the steady state specific activity was ~ 3000 units (nmol p-nitrophenol released from p-nitrophenyl phosphate min1 mg protein-1). 2.2 Preparation of packed-bed reactor systems for metal bioaccumulation Cubes of polyurethane foam (88, 125 mm3) coated with Serratia sp. biofilm were packed in a cylindrical glass column (length 9.0 cm and internal diameter 1.5 cm) of working volume of ca. 13 ml. The total amount of protein immobilised onto the foam was ca. 95 mg per reactor, measured by the bicinchoninic acid/CuSO4 method (Sigma protein test kit TPRO-562) using bovine serum albumin as standard [17]. Considering that the amount of protein is ca. 50% of the dry weight of biomass [unpublished], the total biomass per reactor was, therefore, ca. 190 mg. The reactors were challenged with an upward flow (ca. 10 ml h-1) via an external peristaltic pump (Watson-Marlow). All tests were done at ambient temperature. Challenge solutions comprised sodium citrate buffer (2 mM), pH 6.0 and sodium glycerol-2-phosphate (5 mM) supplemented with Co(NO3)2.6H2O (1 mM), CsCl (1 mM), Sr(NO3)2 (1 mM) or UO2(NO3)2.6H2O (1 mM) as specified in individual experiments. 2.3 Metal biosorption by packed-bed reactors Reactor systems were prepared, similar to those described in section 2.2, and challenged with metal (1 mM) in the presence of citrate buffer (2 mM), pH 6.0, but in the absence of glycerol-2-phosphate. 2.4 Co-crystallisation of metals by packed-bed reactors not previously primed with HUP Reactor systems were prepared, similar to those described in section 2.2, and challenged with solutions comprising UO2(NO3)2.6H2O (1 mM), sodium citrate buffer (2 mM), pH 6.0 and sodium glycerol-2-phosphate (5 mM) supplemented with Co(NO3)2.6H2O (1 mM), CsCl (1 mM) or Sr(NO3)2 (1 mM) as specified in individual experiments.

1156

Microbiology Fundamentals

2.5 Use of reactor packed-bed systems previously primed with HUP Reactor systems were "primed" (deposition of HUP onto the biomass) using a solution of UO2(NO3)2.6H2O (1 mM), sodium glycerol-2-phosphate (5 mM) and sodium citrate buffer (2 mM), pH 6.0. First step: four reactors were primed for 24.3 h at a flow rate of 24 ml.h-1 giving a removal of ca. 65% of the input uranyl ion (total uranium loaded was ca 139 mg). Second step: two reactors were challenged with a solution comprising Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM), pH 6.0 and the other two were challenged with a solution comprising Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM) and sodium citrate buffer (2 mM), pH 6.0. 2.6 Spectrophotometric analysis of Co2+ The Co2+ contents of reactor outflows were estimated using the method described by Onishi [18], with a slight modification. The sample or standard containing Co2+ (30 µl, 110 µg) was transferred to a test tube, citric acid solution (0.2 M, 250 µl) and phosphateboric acid buffer (6.2 g of boric acid, 35.6 g of disodium phosphate dehydrated, and 500 ml of 1 M sodium hydroxide in a total volume of 1 l, 300 µl) were added. The pH of the solution should be close to 8.0. Nitroso-R salt (supplied by Fluka, UK) solution (0.2%, 125 µl) was added while stirring. The test tubes containing the samples were covered, transferred to a water bath at 100°C and left for 1 min. Concentrated HNO3 (250 µl) was added and the samples left for a further 1 min. Samples were cooled in the dark and then deionised water was added (325 µl). The transmittance of the solution was measured at 420 nm. 2.7 Spectrophotometric analyses of Sr2+ and/or UO22+ A method was developed, similar to that described by Yong et al. [19], for the simultaneous determination of uranium and strontium in mixed solutions using arsenazo III. Michaylova and Kouleva [20] showed that the complex formation of strontium with arsenazo III is maximum at pH 9-10. To each test tube containing 30 µl of a target metal (or metal mixture) was added 1.97 ml of one of the two solutions, as appropriate to the metal under test (0.1 M HCl for uranium and 0.1 M borate buffer, pH 9.0 for strontium). Metal was visualised by the addition of 0.1 ml of 0.15% (w/v) arsenazo III, with estimation of the blue-violet complex at 652 nm and 649 nm for uranium and strontium, respectively. The blue-violet complex develops in 25 min with strontium [20]. 2.8 Spectrophotometric analysis of Cs+ The Cs+ contents of reactor outflows were estimated using the method described by Huey and Hargis [21] with modifications. Sample or standard containing Cs+ (60 µl, 1-50 µg) was transferred to an Eppendorf tube (1.5 ml), perchloric acid solution (6 M, 84 µl) was added and diluted with deionised water (192 µl). Phosphomolybdic acid (supplied by Riedel-deHaën) solution (14%, w/v, 60 µl) was added, mixed and allowed to stand for 30 min. The suspension was centrifuged at 16,060 g for 30 min, the supernatant discarded using a glass Pasteur pipette and the precipitate washed with perchloric acid (1.2 M, 300 µl). The suspension was centrifuged at 16,060 g for 30 min, the supernatant discarded and the precipitate was dissolved in borate buffer pH 9.0 (0.05 M, 360 µl) during 30 min. Borate buffer (840 µl) was added to give a final volume of 1.2 ml. The absorbance was measured at 226 nm.

1157

Microbiology Fundamentals

3.

RESULTS AND DISCUSSION

3.1 Biosorption and bioaccumulation of Co2+ by Serratia sp. biofilm The first experiment was designed to test the background adsorption/biosorption of Co2+ from aqueous flows by polyurethane foam coated with Serratia sp. biofilm (Fig. 1). Approximately 11% of the Co2+ was removed initially (up to 21 ml, or ca. 2 column volumes), decreasing rapidly thereafter. Biosorption of Co2+ by the biomass was negligible (less than 1% of the biomass dry weight). Bioaccumulation of Co2+ by Serratia sp. in the presence of glycerol 2-phosphate was also negligible (Fig. 1) even though excess phosphate was produced (not shown). These results are similar to those obtained using immobilised Serratia sp. cells for the removal of Ni2+ [8, 11], and show that phosphatasemediated metal bioprecipitation is not appropriate for these metals.

Figure 1. Co2+ removal by reactors containing Serratia sp. biofilm immobilised onto polyurethane foam and challenged with: ○, Co(NO3)2.6H2O (1 mM) in the presence of citrate buffer (2 mM), pH 6.0 (biosorption); ●, Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM) and citrate buffer (2 mM), pH 6.0 (bioaccumulation) 3.2 Co-crystallisation of Co2+ and HUP Previous experiments [8] showed that Serratia sp. cells, when challenged with a solution of uranyl nitrate (0.5 mM) and nickel nitrate (0.5 mM) in the presence of glycerol-2-phosphate and citrate buffer, accumulated both metals. The flow rate was set to give a removal of UO22+ of 56.3% and the corresponding removal of Ni2+ from solution was maintained at 27.3% (during 42 h, i.e. 25 column volumes). The proportion of metal removed suggest the formation of Ni(UO2PO4)2.7H2O; the identity of the material accumulated by the cells was confirmed by X-ray powder diffraction [8]. We tested cells immobilised onto polyurethane foam challenged with a solution of UO2(NO3)2.6H2O (1 mM) and Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate with a flow rate set at 10 ml.h-1. The removal of UO22+ was ca. 99% but the corresponding removal of Co2+ was less than 10% (Fig. 2). 3.3 Uptake of Co2+ onto/into HUP previously accumulated onto the biomass In this study the reactors were primed with HUP as described in section 2.5 which resulted in ca. 90 mg uranium deposited per reactor (ca. 47% of bacterial dry weight). The initial challenge with Co2+ (Fig. 3) resulted in a slowly decreasing removal of Co2+ over 226 ml, with no sharp breakthrough after saturation of the HUP host. This pattern is similar to that obtained using Serratia biofilm immobilised onto ceramic support, primed 1158

Microbiology Fundamentals

with HUP, and initially challenged with Ni2+ solution [11]. The theoretical capacity of a reactor containing 90 mg uranium (in the form of HUP) would be 11 mg of cobalt at a molar ratio of Co:U of 1:2, i.e. for the formation of Co(UO2PO4)2. A loss of Co-removing capacity after 226 ml corresponds to ca. 10 mg of deposited cobalt which, therefore, is attributable to column saturation. As expected the presence of citrate in the challenge solution reduced the amount of Co2+ deposited (Fig. 3), since the metal would be presented to the cells as the citrate complex; citrate was shown previously to remove Ni2+ from its position within the HUP host crystal. [12].

Figure 2. Metal removal by reactors containing Serratia sp. biofilm immobilised onto polyurethane foam and co-challenged with a solution of Co(NO3)2.6H2O (1 mM) and UO2(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM) and citrate buffer (2 mM), pH 6.0. ■, UO22+ removed. ○, Co2+ removed

Figure. 3. Co2+ removal by reactors containing Serratia sp. biofilm immobilised onto polyurethane foam, primed with HUP and challenged with: ○, a solution of Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM), pH 6.0; ●, a solution of Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM) and citrate buffer (2 mM), pH 6.0. 3.4 Co-crystallisation of Cs+ and HUP Serratia sp. cells immobilised onto polyurethane foam were co-challenged with UO2(NO3)2.6H2O (1 mM) and CsCl (1 mM) in the presence of glycerol-2-phosphate at a flow rate of ca. 10 ml.h-1. The removal of UO22+ was ca. 96% and the corresponding removal of Cs+ was ca. 58% (Fig. 4). The removal of Cs+ using the phosphatasebiomineralisation route alone, in the absence of the HUP host crystal, was less than 20%. Thus it was concluded that although co-crystallization did not promote removal of Co2+ (above) this approach has potential for the removal of Cs+. It should be noted that since glycerol 2-phosphate was provided as the sodium salt (5 mM) Na+ would be present to a five-fold excess over Cs+ and in this respect it can be suggested that co-crystal formation 1159

Microbiology Fundamentals

favours formation of the CsUO2PO4 precipitate. However, it is not easy to distinguish between HUO2PO4 and NaUO2PO4 by X-ray powder diffraction because they both have very similar crystal lattice [13].

Figure 4. Metal removal by reactors containing Serratia sp. biofilm immobilised onto polyurethane foam and co-challenged with a solution of CsCl (1 mM) and UO2(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM) and citrate buffer (2 mM), pH 6.0. ■, UO22+ removed. ○, Cs+ removed. 3.5 Co-crystallisation of Sr2+ and HUP Serratia sp. cells immobilised onto polyurethane foam were co-challenged with UO2(NO3)2.6H2O (1 mM) and Sr(NO3)2 (1 mM) in the presence of glycerol-2-phosphate at a flow rate of ca. 10 ml.h-1. The steady-state removal of UO22+ was more than 92% and the corresponding removal of Sr2+ was in the range of 50-56% (Fig. 5). The removal of Sr2+ using the phosphatase-biomineralisation route alone, in the absence of the HUP host crystal, was negligible. The molar ratio accumulated was in accordance with the deposition of Sr(UO2PO4)2.

Figure 5. Metal removal by reactors containing Serratia sp. biofilm immobilised onto polyurethane foam and co-challenged with a solution of Sr(NO3)2 (1 mM) and UO2(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM) and citrate buffer (2 mM), pH 6.0. ■, UO22+ removed. ○, Sr2+ removed 4.

CONCLUSIONS We have shown that using the continuous co-challenge system, i.e. in the presence of uranyl ion and glycerol 2-phosphate, the deposited HUP is able to promote removal of ca. 58% of Cs+ and 50-56% of Sr2+ (during 53 column volumes) but was not able to remove the Co2+ (less than 10%). However, using cells pre-coated with HUP as inorganic ionexchangers it was possible to remove the Co2+ until the column reached saturation corresponding to a molar ratio of U:Co of ~2:1 in accordance with a crystal of 1160

Microbiology Fundamentals

Co(UO2PO4)2. Further work to be reported will confirm the identity of the crystals formed, using X ray powder diffraction analysis and EXAFS. However, the preliminary data reported here show that removal of Co2+, Sr2+ and Cs+ as, respectively, Co(UO2PO4)2, Sr(UO2PO4)2 and CsUO2PO4 is feasible since no removal of the surrogate fission products occurred in the absence of the uranyl ion. Such uranyl phosphate-driven co-deposited or intercalation is entirely feasible for the treatment of real nuclear wastes since the fission products residues often occur with an excess of residual uranium [1, 2] and, furthermore, the metal precipitating phosphatase is highly radioresistent [22]. ACKNOWLEDGMENTS This work is supported by the Biotechnology and Biological Sciences Research Council. The authors thank Recticel (Belgium) for the polyurethane reticulated foam. REFERENCES

1. L. E. Macaskie, CRC Crit. Rev. Biotechnol. 11 (1991) 41-112. 2. J. R. Lloyd and L. E. Macaskie, In: Environmental Microbe-Metal Interactions, (D. R. Lovely, ed.), ASM Press, Washington, (2000) 277-327. 3. B. K. Vainshtein, Fundamentals of Crystals, Modern Crystallography, Vol. 1, 2nd edn., Springer, Berlin, 1994. 4. L. E. Macaskie, R. M. Empson, A. K. Cheetham, C. P. Grey and J. Skarnulis, Science 257 (1992) 782-784. 5. A. Clearfield, Chem. Rev. 88 (1988) 125-148. 6. R. Pozas-Tormo, L. Moreno-Real, M. Martinez-Lara and S. Bruque-Gamez, Can. J. Chem. 64 (1986) 30-34. 7. R. Pozas-Tormo, S. Bruque-Gamez, M. Martinez-Lara and L. Moreno-Real, Can. J. Chem. 66 (1988) 2849-2854. 8. K. M. Bonthrone, G. Basnakova, F. Lin and L. E. Macaskie, Nature Biotechnol. 14 (1996) 635-638. 9. G. Basnakova and L. E. Macaskie, Biotechnol. Bioeng. 54 (1997) 319-328. 10. G. Basnakova, A. J. Spencer, E. Palsgard, G. Grime and L. E. Macaskie, Environ. Sci. Technol., 32 (1998) 760-765. 11. G. Basnakova, J. A. Finlay and L. E. Macaskie, Biotechnol. Lett., 20 (1998) 949-952. 12. G. Basnakova and L. E. Macaskie, Biotechnol. Lett., 23 (2001) 67-70. 13. P. Yong and L. Macaskie, J. Chem. Tech. Biotechnol. 63 (1995) 101-108. 14. L. E. Macaskie and A. C. R. Dean, Biotechnol. Lett., 7 (1985) 627-630. 15. P. Pattanapipitpaisal, A. N. Mabbett, J. A. Finlay, A. J. Beswick, M. Paterson-Beedle, A. Essa, J. Wright, M. R. Tolley, U. Badar, N. Ahmed, J. L. Hobman, N. L. Brown and L. E. Macaskie, Environ. Technol., 23(7) (2002) 731-45. 16. J. A. Finlay, V. J. M. Allan, A. Conner, M.E. Callow, G. Basnakova and L. E. Macaskie, Biotechnol. Bioeng., 63 (1999) 87-97. 17. K. P. Nott, M. Paterson-Beedle, L. E. Macaskie and L. D. Hall, Biotechnol. Lett., 23 (2001) 1749-1757. 18. H. Onishi, Photometric Determination of Traces of Metals, Part IIA: Individual Metals, Aluminum to Lithium, 4th edn., John Wiley, New York (1986) 454-459. 19. P. Yong, H. Eccles and L. E. Macaskie, Anal. Chim. Acta, 329 (1996) 173-179. 20. V. Michaylova and N. Kouleva, Talanta, 21 (1974) 523-532. 21. F. Huey and L. G. Hargis, Anal. Chem., 39(1) (1967) 125-127. 22. B. C. Jeong, Studies on the atypical phosphatase of a heavy metal accumulating Citrobacter sp., D. Phil. Thesis, University of Oxford, UK (1992). 1161

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Removal of soluble manganese from mine waters using a fixed-bed column bioreactor D. Barrie Johnson, Helen Miller, Sandra Ukermann and Kevin B. Hallberg School of Biological Sciences, University of Wales, Bangor, LL57 2UW, U.K. Abstract Acid mine drainage waters vary considerably in the range and concentration of heavy metals they contain. Besides iron, aluminum and manganese are frequently present at elevated concentrations in waters draining both coal and metal mines. Passive treatment systems (aerobic wetlands and compost bioreactors) are designed to remove iron by biologically induced oxidation/precipitation, and aluminum by precipitation as Al(OH)3 as a result of biologically-generated alkalinity. Manganese, however, is problematic as it does not readily form sulfidic minerals (in compost bioreactors) and requires elevated pH (>8) for oxidation of Mn(II) to insoluble Mn(IV). As a result, manganese removal is often less effective than iron and aluminum removal, such as at the pilot passive plant treating water draining the former Wheal Jane tin mine in Cornwall, U.K. We have sought to devise a novel microbiological approach for effective removal of manganese from mine waters at pH 5-7. Ferromanganese nodules (about 2 cm diameter) were collected from an abandoned mine adit in the Gwydyr forest, north Wales, and used to inoculate a fixed bed bioreactor (working volume ca. 700 ml). Pumice-like porous beads, made from inert recycled glass, were used as carrier materials for microbial biofilms. Following colonization of the beads, the aerated reactor was tested for removal of soluble manganese in synthetic and actual mine waters, using a continuous plug-flow mode of operation. Data from preliminary experiments show that the bioreactor is highly efficient at catalyzing the removal of manganese from the mine waters via oxidation of soluble Mn (II) and precipitation of the resultant Mn (IV) in the bioreactor. Such an approach appears to be a suitable alternative to current bioremediation strategies employed for manganese removal from mine waters.

Keywords: acid mine drainage; bioreactors; bioremediation; manganese 1.

INTRODUCTION Soluble manganese (Mn (II)) is often found in considerably greater amounts in AMD than in unpolluted streams and groundwater [1]. Even though there are uncertainties regarding the toxicity of manganese, the removal of this metal from surface- and groundwaters is desirable for several reasons. As with iron and aluminum, manganese also contributes to the total mineral acidity of mine waters when it is oxidized. Additionally, the presence of manganese in water sources for human consumption is undesirable because it can impart a metallic taste to water, it will stain laundry and water fixtures and, as it precipitates readily as Mn (IV), manganese tends to block water distribution 1163

Microbiology Fundamentals

networks. Due to these concerns, the U.S. Environmental Protection Agency (EPA) has set a secondary maximum contaminant level for Mn of 0.05 mg/l. The EPA has also established guidelines limiting the concentration of Mn in acidic waters discharging from mines at maximum of 4 mg/l, as long as average discharges for a 30-day period to not exceed 2 mg/l. To address such concerns, and to meet the potentially more stringent limits that are being discussed in Europe and in the U.S., effective means of removal of Mn from mine waters are required. Although manganese readily precipitates as Mn(IV), little oxidation of Mn(II) occurs in solutions below pH 8, in contrast to iron, and the kinetics for manganese oxidation is slow, relative to that of ferrous iron. In addition, biological oxidation of Mn(II) does not proceed rapidly in the presence of Fe(II), and thus it is not removed significantly in wetlands where the concentration of iron exceeds 1 mg/l [2]. The lack of affinity of manganese for sulfide prohibits any significant removal as a sulfide in compost wetlands and bioreactors. Therefore, alternative approaches are needed for the removal of manganese from mine drainage waters. The treatment systems currently used are typically applied at the end of any treatment process scheme, to ensure the all iron has been removed first. An abiotic Mn-removal system consisting of columns filled with limestone (to significantly raise the pH for Mn oxidation and subsequent precipitation) has been described [3]. Biological approaches that have been proposed for removing manganese from solution, by causing solution pH to rise above 8 via oxygenic photosynthesis, include columns of immobilized cyanobacteria [4] and small pools filled with rocks that have been colonized by algae [5]. The latter approach has been applied at the Wheal Jane passive treatment plant. This study site was built in 1993 following a catastrophic spill of acid mine drainage from the Wheal Jane tin mine in Cornwall, England. However, the algal ponds have proved highly ineffective at removing soluble manganese (as described below) and an alternative approach, using a fixed-bed column bioreactor, has been initiated as an alternative strategy, which would be equally applicable for treating other mine waters. 2.

MANGANESE REMOVAL AT THE WHEAL JANE PASSIVE TREATMENT PLANT A composite, pilot-scale passive treatment plant ("PPTP") was established at the former Wheal Jane tin mine in Cornwall, England in 1995, as a large-scale experimental facility to examine the efficacy of using constructed "wetlands" to remediate acidic, metalrich minewaters. A more complete description of the design and operation of the treatment system can be found elsewhere [6]. In brief, a small fraction (<5%) of the acid (pH ~ 3.5) mine drainage (AMD) water flows into the pilot passive treatment plant, where it is fed into three similar composite passive systems (Figure 1). The main variations are the pretreatment that the AMD is subjected to ahead of entering the "wetlands". These are: (i) addition of lime (CaO) to adjust the AMD to a pre-determined pH value (the "lime-dosed" system); (ii) passage through an anaerobic (manure-based) cell to remove dissolved oxygen and then through an anoxic limestone drain (to passively add alkalinity; the "ALD" system); (iii) no pretreatment (the "lime-free" system). In each system, AMD then flows (by gravity) through a series of five “aerobic cells” (which are considerably larger in the "lime-free" system), then through a single "anaerobic cell" (containing sawdust and cow manure; the "compost bioreactor") and finally through a series of ten smaller "rock filters" (algal ponds). The key roles of the components in the passive systems are: (a) aerobic cells: oxidation and precipitation of iron, and co-precipitation of arsenic; 1164

Microbiology Fundamentals

(b) anaerobic cell: sulfidogenesis and alkalinity genesis, causing precipitation of chalcophilic metals as sulfides (e.g. copper, zinc and cadmium) and other metals as hydroxides (e.g. aluminum); (c) rock filters: (i) oxygenic photosynthesis causing aeration and further increase in pH and precipitation of manganese (as carbonate and/or oxide) and (ii) polishing of processed AMD (e.g. removal of organic carbon compounds draining the compost bioreactors). Fe2+ oxidation and Fe/As removal

MeS Mn removal removal

Rock Filters (10)

Compost Bioreactor

Aerobic Cell 5

Aerobic Cell 3

Aerobic Cell 1

Inflowing AMD

Outflow Figure 1. Schematic of the pilot passive treatment plant for remediating AMD at the former Wheal Jane tin mine, Cornwall, England The performance of the PPTP has been monitored by a group of geochemists and microbiologists since 1998. In general, the aerobic cells have proved highly effective in fulfilling their roles, but in only one of the three systems (the "lime-free" system) has the performance of the compost bioreactors and algal ponds been satisfactory [7]. Figure 2 shows the concentrations of manganese in waters sampled throughout the three systems in the PPTP in April 2002, and is representative of the general trends that have been observed. Data in Figure 2 show that changes in soluble manganese do occur within the aerobic cells (which are mostly highly acidic, pH <3, due to ferric iron hydrolysis); this is, in part, due to dilution of the AMD by rainwater. However, significant concentrations (~25 mg/l) are present in the water draining the compost bioreactors and, as shown in Figure 2, is removed effectively only in the "lime-free" system. Indeed there was evidence of remobilization of manganese in the rock filters of the "lime-dosed" system on this sampling occasion. A major reason for the contrasting trends observed in Mn geochemistry found within the algal ponds is that the necessary elevation in pH resulting from oxygenic photosynthesis only occurred in the "lime free" system, as illustrated in Figure 3. The discrepancy in Mn removal is due, in the main, to the variation in the performance of the compost bioreactors in each of the three systems. Water draining these bioreactors in the "ALD" and "lime-free" systems contains higher concentrations of both ferrous iron and sulfide. Microbiological (and abiotic) oxidation of these species in the algal ponds causes the pH of the water to decline, which is counter-productive to the overall objective [7]. As the treatment plant is built in a temperate climate that receives considerable rainfall, photosynthetic activity appears to be limited such that the rock filters are unable to overcome the additional acid burden from the poorly performing compost bioreactors. 1165

Microbiology Fundamentals

Figure 2. Manganese concentrations in various components of the pilot passive treatment plant at the former Wheal Jane tin mine; sampling date April 2002

Figure 3. Manganese (bars) and pH (lines) in the compost bioreactor influent (CB in) and effluent (CB out) and in each of the 10 rock filter pools (RF1-RF10). The samples were taken in April, 2002, from the ALD (A) and LF (B) systems. The straight lines at the bottom of each graph represent the lower limit of manganese detection, which was 0.25 mg/l

1166

Microbiology Fundamentals

3.

REMOVAL OF SOLUBLE MANGANESE BY FRESHWATER FE-MN NODULES To overcome the limitations of the rock filters, we have proposed an alternative treatment strategy based on bioreactors containing immobilized manganese-oxidizing bacteria. Ferromanganese nodules were collected from a trial mine adit located in the Gwydyr forest, north Wales. The irregular, spherical nodules, ca. 1-2 cm in diameter, were found in a pool within the adit (pH 6.9, soluble [Mn] 30 mg/l, conductivity 149 µS/cm). These were transferred to the laboratory and stored at 4°C. Two of the nodules were placed into replicate 250 ml conical flasks containing 100 ml of filter-sterilized "MMS" medium [8], which had been developed to promote the growth of manganese oxidizing/depositing bacteria. MMS medium contains sodium pyruvate (2.4 g/l) and basal salts (pH 7); 50 µl of a 100 mM manganese sulfate solution (filter-sterilized) was added just prior to inoculation with the nodules, and the flasks were incubated, shaken (100 rpm) at 20°C. Samples were withdrawn at regular intervals and concentrations of soluble Mn analyzed colorimetrically by complexing with formaldoxime (Hydrocheck, Cambridge, U.K.). Results, given in Figure 4, show that concentrations of soluble manganese declined steadily during incubation; this corresponded with the appearance of brown-black precipitates (and associated bacteria) within cultures.

Figure 4. Removal of soluble manganese in shake flask cultures containing freshwater ferro-manganese nodules. The error bars (occluded in most cases by symbols) show the range of soluble [Mn] at each sampling occasion 4.

COMMISIONING AND PERFORMANCE OF A FIXED-BED BIOREACTOR A fixed-bed bioreactor for removing soluble manganese from mine waters was set up by mixing freshwater ferromanganese nodules (ca. 10%, v/v) with porous, acid-washed glass beads (8-16 mm) manufactured by Poraver GmbH, Germany [9]. The nodules/beads mixture was packed (to a depth of 14.5 cm) into a 25 by 10 cm perspex column, between 3.5 cm layers of acid-washed pea gravel (Figure 5a). The base and cap of the column contained ports for aeration (applied at ca. 2 l/min), medium circulation and sampling. The column was filled to the top of the upper gravel layer surface with MMS medium, which was then re-circulated through the column using a peristaltic pump until the soluble 1167

Microbiology Fundamentals

manganese had declined to <10% of the initial concentration, at which point the column was drained and fresh medium introduced. This sequence was repeated several times, and the column was observed to accumulate increasing amounts of brown-black (Mn (IV)) precipitates (Figure 5b). The external plumbing on the bioreactor was then re-configured to allow the system to be tested in continuous plug-flow mode. Representative data showing removal of soluble manganese from MMS medium with the fixed-bed bioreactor run in recirculating and plug-flow modes are shown in Figure 6. In both cases, >90% of soluble Mn was removed within 4 hours. When run in plug-flow mode, 50% of soluble manganese was removed with a flow rate of 400 ml/h, and 68% when the flow rate was lowered to 195 ml/h.

Figure 5. Fixed-bed bioreactor for removing soluble Mn from solution: (a) after initial construction; (b) several weeks after continuous operation

Figure 6. Performance of the fixed-bed bioreactor in removing soluble manganese from MMS medium when run in recirculating ( ) and plug-flow ( ) modes 1168

Microbiology Fundamentals

5.

DISCUSSION Manganese is the second most abundant heavy metal in the lithosphere, occurring at between 0.002 and 10% (w/w) in terrestrial ecosystems, ca. 0.2 µg/l in marine waters, and an average of 8 µg/l in fresh waters [10]. As with many other metals, greatly enhanced concentrations of soluble manganese can occur in waters draining coal and metal mines. The major oxidation states of manganese are +2 and +4: Mn (II) can occur as free ions in solution (Mn2+, and Mn(OH)+ in circum-neutral pH waters) or complexed with a variety of organic and inorganic ligands. In contrast, Mn (IV) forms very insoluble oxide phases; MnO2 has a point of zero charge at pH 2.8 which, together with its large specific surface area, gives rise to a marked capacity to adsorb metal cations, including Mn (II). Thermodynamically, Mn (II) should oxidize spontaneously to Mn (IV) in aerated, neutral pH waters, but the activation energy required is relatively high, and this greatly slows down the process, causing Mn (II) to be far more stable in non-acidic waters than is ferrous iron. In contrast, the adsorption of Mn (II) onto MnO2 is much more rapid. The "activation energy barrier" for Mn oxidation can be overcome biologically. Many microorganisms are known to be able to catalyze the oxidation of Mn (II), and at least some of these are known utilize the energy available from the reaction [10]. Manganese (II) may also form a highly insoluble carbonate (MnCO3; rhodocrosite). Since, however, the concentration of CO32- in solution is dictated by pH, formation of this mineral is only significant in neutral-alkaline water bodies. The existence of marine "manganese nodules" has been known for some time, and these are mined as an ore of the metal. Iron/manganese-rich nodules may also occur in freshwaters, though these tend to be much smaller than their marine equivalents. Stein et al. [11] analyzed prokaryotic populations associated with freshwater ferromanganese micronodules (mm scale) and sediments by extracting nucleic acids, amplification of 16S rRNA genes and phylogenetic analysis of cloned products. Two groups of known Mn (II)oxidizing genera were identified, Leptothrix and Hyphomicrobium, as well as a group related to metal-reducing bacteria. To date, the microorganisms associated with the freshwater ferromanganese nodules used in the present study have not been subjected to rigorous analysis, though Mn(IV)depositing bacteria have been isolated on solid media inoculated with crushed nodules and bioreactor liquor. The increasingly-efficient performance of the bioreactor (in terms of removing soluble Mn(II)) as Mn(IV) deposits accumulate may be due both to the activities of planktonic-phase Mn(II)-oxidizers and to the abiotic uptake of Mn(II) onto MnO2 precipitates, and subsequent oxidation of the metal by sessile bacteria. What is clear is that a simple fixed-bed bioreactor containing Mn(II)-oxidizing/depositing microorganisms can be highly effective in catalyzing the removal of the metal from a synthetic, metal-rich waste water. So far, limited tests of the technology for bioremediating actual Mn-rich mine waters has shown considerable promise (data not shown). Fixed-bed bioreactors for oxidizing and precipitating ferrous iron from AMD have been described by a number of research groups [e.g. 12]. A major advantage offered by metal-removing bioreactors is that they occupy a far smaller "footprint" than, for example, the rock filters/algal ponds at the Wheal Jane passive treatment plant. Land availability frequently restricts the application of passive mine water treatment systems, and low-maintenance biological systems are attractive alternatives in such circumstances.

1169

Microbiology Fundamentals

ACKNOWLEDGEMENTS We wish to thank DTI (ref. # BTL/20/71), MIRO (U.K.) and the Environment Agency (U.K.) for financial support. We also acknowledge the contributions made by our colleagues at the University of Reading, Imperial College, London, the Centre for Ecology and Hydrology, the Camborne School of Mines and the Knight Piésold Consulting Group (U.K.) to the "Wheal Jane Project". REFERENCES 1. D. Banks, P. L. Younger, R.-T. Arnesen, E. R. Iversen and S. B. Banks, Environ. Geol. 32 (1997) 157. 2. B. Nairn and R. S. Hedin, Constructed wetlands for water quality improvement, G. A. Moshiri (eds.), Lewis Publishers, Boca Raton, Fl, (1993) 187. 3. F. C. Thornton, Ecol. Eng. 4 (1995) 11. 4. J. Bender, J. P. Gould, Y. Vatcharapijarn, J. S. Young and P. Phillips, Water Environ. Res. 66 (1994) 679. 5. P. Phillips, J. Bender, R. Simms, S. Rodriguezeaton and C. Britt, Water Sci. Technol. 31 (1995) 161. 6. Q. U. I. Hamilton, H. M. Lamb, C. Hallett and J. A. Proctor, J. Chart. Inst. Water. Environ. Manage. 13 (1999) 93. 7. K. B. Hallberg and D. B. Johnson, Mine Water Treatment:A Decade of Progress, C. A. Nuttall (eds.), Newcastle upon Tyne, U.K., (2002) 143. 8. Y. M. Nelson, L. W. Lion, W. C. Ghiorse and M. L. Shuler, Appl. Environ. Microbiol. 65 (1999) 175. 9. Å. Kolmert and D. B. Johnson, J. Chem. Technol. Biotechnol. 76 (2001) 836. 10. H. L. Ehrlich, Geomicrobiology, 3rd edition. Marcel Dekker, New York (1996). 11. L. Y. Stein, M. T. La Duc, T. J. Grundl and K. H. Nealson, Environ. Microbiol. 3 (2001) 10. 12. A. Mazuelos, F. Carranza, I. Palencia and R. Romero, Hydrometallurgy 58 (2000) 269.

1170

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Sulfane sulfur of persulfides is the actual substrate of the sulfur-oxidizing enzymes from Acidithiobacillus and Acidiphilium spp. T. Rohwerder∗ and W. Sand Department of Microbiology, Institute for General Botany, University of Hamburg, Ohnhorststr. 18, D-22609 Hamburg, Germany Tel/Fax: +49 40 428 16 423, e-mail: [email protected] Abstract The sulfur chemistry of sulfur dioxygenase (EC 1.13.11.18) was analyzed in various strains of the genera Acidithiobacillus and Acidiphilium. Cell-free extracts oxidized elemental sulfur well, however only in the presence of glutathione (GSH). Sulfur had to be merged with GSH to form the organic persulfide or higher homologues (GSnH, n > 1) prior to the enzymic oxidation to sulfite. The latter reacted non-enzymically further either to sulfate, thiosulfate, or glutathione S-sulfonate (GSSO3-). In contrast, oxidation of hydrogen sulfide by the dioxygenase system did not require the reduced but the oxidized form of glutathione (GSSG). The latter formed with sulfide non-enzymically GSSH prior to enzymic oxidation. Persulfide species as sulfur donors could not be replaced by other sulfane-sulfur-containing compounds such as thiosulfate, polythionates, aryl thiosulfonates, or bisorganyl-polysulfanes. In conclusion, persulfides and higher monoorganylpolysulfanes are exclusively the substrates for the sulfur dioxygenases of meso-acidophilic sulfur bacteria. On the basis of these findings a biochemical model for both elemental sulfur and sulfide oxidation in Acidithiobacillus and Acidiphilium spp. is proposed.

Keywords: sulfur dioxygenase, sulfur oxidation, sulfide oxidation, Acidithiobacillus, Acidiphilium 1.

INTRODUCTION For the oxidation of elemental sulfur a conclusive biochemical pathway has not yet been identified. Diverse enzymic systems like a reverse sulfite reductase [1,2], a thiosulfate-oxidizing enzyme complex [3,4], and dioxygenases or oxygenase-reductases [5,6,7,8] are considered to be involved. The meso- and moderately thermophilic leaching bacteria seem to have a pathway depending on low-molecular thiols. For in vitro assays glutathione (GSH) is the preferentially used thiol compound (reviewed in [9]). Due to its capability to incorporate molecular oxygen into the sulfur substrate and to produce sulfite as the primary reaction product the system has been named sulfur dioxygenase (EC ∗ Present address: UFZ Center for Environmental Research Leipzig-Halle, Permoserstr. 15, D04318 Leipzig, Germany

1171

Microbiology Fundamentals

1.13.11.18; [5]). This GSH-dependent activity has been detected in Acidithiobacillus thiooxidans and Acidithiobacillus ferrooxidans [10,11,12] as well as in Sulfobacillus thermosulfidooxidans [13]. Although the known sulfur dioxygenases are very similar in molecular size and cofactor content, the role of GSH is under dispute. The work of Suzuki [5,10] and Silver & Lundgren [11] suggest a catalytic role for GSH, where organic polysulfanes are formed and sulfur is oxidized at the zero valence state to sulfite (equation 1). In contrast, Sugio and coworkers [12,14] described an enzyme system that probably used dissolved sulfide as substrate. Consequently, GSH would be consumed stoichiometrically in the course of sulfur oxidation due to the preceding non-enzymic reduction of elemental sulfur to sulfide. GSnH + O2 + H2O GSn-1H + SO32- + 2 H+ n>1 (1) So far, the sulfur chemistry and especially the fate of GSH has not been analyzed leaving the actual activation mechanism for elemental sulfur unresolved. Therefore, we reinvestigated the GSH-dependent sulfur-oxidizing system. Special analytical attention was given to the organic sulfur species that occurred in the in vitro assay. Besides elemental sulfur other sulfane compounds and hydrogen sulfide were tested to elucidate the actual substrate used by the sulfur dioxygenase. For this purpose the enzymic activity of cell-free extracts of At. thiooxidans, At. ferrooxidans, Acidiphilium acidophilum, and Acidiphilium cryptum was investigated. 2.

MATERIALS AND METHODS

2.1 Bacterial strains, growth conditions, and cell-free preparations The bacterial strains of this study are listed in Table 1. Growth on elemental sulfur powder (5 g/L) was performed in a salt solution modified from Mackintosh [15] with 2 mM NH4Cl and an initial pH of 3.0 adjusted with HCl. In addition, a lithotrophic iron medium [15] with 5 g/L iron(II) ions and a glucose-based medium [16] were applied for the growth of iron-oxidizers and facultative or obligate heterotrophs, respectively. All strains were cultivated aerobically at 28°C. Cell disruption was performed at ≤4°C under anaerobic conditions by treatment with glass beads [17]. Suspensions of disrupted cells were centrifuged (20 min, 25,000 x g, twice) to remove intact cells and cell residues. The supernatant, referred to as crude or cell-free extract, usually contained 1 to 2 g/L protein and was used either directly for activity assays or was stored under an anaerobic atmosphere at –25°C (less than 3 months). 2.2 Enzyme assays All assays were performed aerobically at 30°C in phosphate buffer (10 mM, pH 6.5) with stirring at 300 rpm. If needed, the pH was maintained by titration with 50 mM KOH or 50 mM HCl. In order to determine non-enzymic reactions, assays without protein, with 0.2 g/L BSA, or with heat-inactivated crude extracts (90°C for 30 min) were used. For measuring elemental sulfur oxidation a system of dispersed sulfur in water was developed. To 50 mL of deionized water an equal volume of a saturated acetonic sulfur solution was added and dialyzed against 5 L deionized water to remove the acetone. The whitish dialysis product contained sulfur droplets of 2 to 10 µm in diameter and was used at a concentration of 4 mM. When sulfide, thiosulfate, tetrathionate, and p-toluolthiosulfonate were tested stock solutions of the respective potassium or sodium salts in deionized water were prepared. A mixture of GSSG and higher homologues was obtained by an incubation of 500 mM elemental sulfur (powder) with 100 mM GSH at pH 7.5 (adjusted with KOH) under stirring and anaerobic conditions until the solution got lemon-colored. At this point 1172

Microbiology Fundamentals

the pH was lowered to 5.0 by addition of HCl. The resulting hydrogen sulfide was removed by evacuation of the gas phase. Finally, the pH was adjusted to 6.5. 2.3 Analyses of sulfur compounds Thiosulfate, polythionates, GSH, GSH-derivatives, and p-toluolthiosulfonate were analyzed by ion pair chromatography using tetrabutylammonium as counter ion [17]. A HPLC system from Kontron/BIO-TEK Instruments was applied. Chromatograms were recorded at 205, 215, 265, and 300 nm concomitantly with spectra from 190 to 320 nm. Elemental sulfur was analyzed by reversed-phase chromatography followed by UVdetection [18]. Sulfite and sulfate were quantified by ion exchange chromatography and conductivity detection [18] applying a Dionex DX 500 system. Sulfite was fixed with methanal prior to analysis [19]. Dissolved sulfide was determined by the methylene blue method [20]. With the exception of elemental sulfur quantification, all samples had been filtered (nylon filter, 0.2 µm) prior to analysis in order to remove elemental sulfur. 3.

RESULTS AND DISCUSSION

3.1 Oxidation of elemental sulfur via glutathione persulfide The enzymic oxidation of elemental sulfur by cell-free extracts from various mesoacidophilic sulfur-oxidizing bacteria was performed mainly according to the sulfur dioxygenase assay of Suzuki [5,10]. As has been demonstrated previously for similar enzyme preparations (reviewed in [9]) enzymic oxidation of elemental sulfur could not be observed in the absence of GSH. With GSH, however, elemental sulfur was readily oxidized by crude extracts derived from sulfur oxidizers. In contrast to data of a previous study [21], preparations from the obligate iron(II) oxidizer Leptospirillum ferrooxidans did not show any sulfur oxidation activity (Table 1). The main reaction product of positive assays was thiosulfate (99%). Besides, traces of sulfite, sulfate, and glutathione Ssulfonate were detectable (data not shown). Obviously, sulfite as the first oxidation product rapidly reacted further with the finely dispersed sulfur to thiosulfate (equation 2). Only trace amounts were consecutively oxidized to sulfate or reacted with GSSG to glutathione S-sulfonate (equation 3). SO32- + 1/8 S8 S2O32(2) GSSG + SO32- + H+ GSSO3- + GSH (3) GSSG was regularly detected in the assay solutions. It was formed in the course of the non-enzymic reduction of elemental sulfur with GSH (equation 4 and 5). This reaction occurred in parallel to the enzymic activity and, besides GSSG, GSnG species with n ranging from 3 to 5 and hydrogen sulfide were formed (data not shown). S8 + GSH (GS9H) GSnH + (9-n)/8 S8 (4) x ≥ 2, y ≥ 1 (5) GSxH + GSyH GSx+y-1G + H2S Comparing sulfur oxidation and reduction activities, it turned out that GSH was not consumed stoichiometrically in the course of elemental sulfur oxidation. In other words, free sulfide was not a substrate for the sulfur-oxidizing enzyme under the experimental conditions. Interestingly, the non-enzymic sulfur reduction by GSH (equation 4 plus 5) and the GSH-dependent enzymic sulfur oxidation were connected in another way: the higher the rates of the enzymic sulfur oxidation were, the lower were the GSH oxidation rates (Fig. 1). Furthermore, for a certain GSH concentration a maximal sulfur dioxygenase activity was obtainable, irrespective of the amount of enzyme, at which no further GSH oxidation occurred. For example, with 100 µM GSH an upper limit of about 350 µM/h 1173

Microbiology Fundamentals

was achieved (Fig. 1). The explanation for this phenomenon can be deduced from equations 1 and 5. Both reactions compete for the highly unstable monoorganylpolysulfane derivatives (GSnH, n > 1) of GSH. Summarizing, from the analyses of all relevant sulfur compounds in these sulfur dioxygenase assays we can fully confirm Suzuki´s proposal [5] of persulfides and their higher homologues being the actual substrates for the sulfur-oxidizing enzyme system of meso-acidophilic sulfur oxidizers (equation 1). This finding is valid for the genera Acidithiobacillus and Acidiphilium. Table 1. Specific activities of sulphur dioxygenase in crude extracts of different strains of meso-acidophilic bacteria strain

substrate for cell growth*

activity† µmol.h-1.mg-1

sulfur 3.8 ± 1.4 Acidithiobacillus ferrooxidans R1 iron(II) 3.5 ± 0.6 Acidithiobacillus thiooxidans DSM 504 sulfur 2.4 ± 0.8 Acidithiobacillus thiooxidans K6 sulfur 1.5 ± 0.5 sulfur 22.4 ± 5.4 Acidiphilium acidophilum DSM 700 glucose 7.8 ± 2.5 Acidiphilium cryptum DSM 2389 glucose 0.3 ± 0.1 Leptospirillum ferrooxidans DSM 2705 iron(II) ND‡ *Strains were subcultured at least for 2 years on the indicated substrates. †Specific activity is expressed as the amount of sulfur atoms oxidized to the valence state of sulfite within 1 hour by 1 mg protein. ‡ND not detectable.

Figure 1. Sulfur dioxygenase assay with 100 µM GSH. Relationship between initial GSH consumption rate Vο and enzymic activity

1174

Microbiology Fundamentals

3.2 Oxidation of hydrogen sulfide Although the free sulfide formed during the incubation of elemental sulfur with GSH was not oxidized enzymically, hydrogen sulfide was tested separately for enzymic oxidation. As expected, cell-free preparations incubated with or without GSH did not oxidize the sulfide. Surprisingly, low but significant enzymic activity was observed after GSSG addition to the assays (at 1 mM). In this case, the formation of thiosulfate and glutathione S-sulfonate was observed (Fig. 2). The latter was the main oxidation product. Furthermore, GSSG was reduced to GSH. From the stoichiometry of all these compounds (Fig. 2) the following reaction mechanism can be deduced. GSSG reacted with hydrogen sulfide to GSH and GSSH (equation 6). The persulfide was consecutively oxidized by the sulfur dioxygenase to sulfite (equation 1) and then formed with excess GSSG the glutathione S-sulfonate (equation 3). GSSG + H2S GSSH + GSH (6) + 2 GSSG + H2S + O2 + H2O 3 GSH + GSSO3 + H (7) Summing up, GSH and glutathione S-sulfonate should be produced in a ratio of 3:1 (equation 7) which agrees quite well with the observed values (Fig. 2). Consequently, sulfide was not oxidized directly, as proposed by Sugio and coworkers [14], but via the formation of persulfide species.

concentration (µM)

3.5

15

3.0

10

2.5 2.0

5

1.5

0

ratio GSH/GSSO3-

4.0

20

1.0 0.0

GSH thiosulfate

0.4

0.8 time (h)

1.2

GSSO3ratio GSH/GSSO3-

Figure 2. Oxidation of sulfide by cell-free extracts in the presence of 1 mM GSSG 3.3 Oxidation of other sulfane compounds From the results of the oxidation tests performed with elemental sulfur and sulfide, it can be clearly deduced that the active enzyme preparations contained no enzymic activity for the direct oxidation of these sulfur species. Only the sulfane sulfur of persulfides was oxidizable. In additional experiments the specificity of this enzymic activity was tested with various inorganic and organic sulfane compounds (see 2.2.). Briefly, with none of these compounds an oxidation activity was observed. However, when a mixture of GSSG and higher GSnG species (2
Microbiology Fundamentals

Consequently, it was highly likely formed from sulfite and GSSG because this reaction produced equimolar amounts of GSSO3- and GSH (equation 3).

500 300 200

50

100 0

0 0.0

GSnG (mAbs min)

A

B

-

400 100

0.5

1.0 1.5 time (h)

thiosulfate

2.0

GSH

GSSO3-

150

1100

100

1000

50

900

0

800 0.0

0.5

1.0

1.5

GSSG (µM)

thiosulfate (µM)

150

GSH, GSSO3 (µM)

(8) GSnG + GSH GSxH + GSyGn ≥ 3, x ≥ 2, y ≥ 2 Summarizing, only the sulfane sulfur of persulfide compounds can be oxidized by the sulfur dioxygenase of meso-acidophilic bacteria. Neither polythionates nor bisorganylpolysulfanes are oxidized directly.

2.0

time (h) GS3 G

GS4 G

GS5 G

G SSG

Figure 3 A+B. Oxidation of GSnG species (2 < n < 6) by cell-free extracts in the presence of 200 µM GSH. Amounts of GSnG are expressed as peak areas recorded at 215 nm 3.4 Proposal of a biochemical model for sulfur and sulfide oxidation Compiling all data available, a model for the oxidation of sulfide and elemental sulfur can now be proposed (Fig. 4). Extracellular elemental sulfur (S8) reacts with thiol groups of special outer-membrane proteins (OMP-SH) and, thus, can be transported as persulfide sulfur (OMP-SSH) into the periplasmic space. Here sulfur dioxygenase (SDO) takes over the sulfane sulfur and oxidizes it to sulfite. Sulfite is oxidized further to sulfate by a sulfite:acceptor oxidoreductase (SOR). This enzyme most probably uses cytochromes (Cyt) as electron acceptors [22,23,24]. Free sulfide is oxidized by a separate sulfide:quinone oxidoreductase (SQR), which is located at the periplasmic site of the cytoplasmic membrane [25,26]. For this reaction polysulfides have recently been proposed to be the oxidation products [27]. However, at least for acidophilic bacteria we believe that membrane-bound persulfides (OMP-SSH) as products of SQR activity are more likely (Fig. 4). In fact, the membrane binding of sulfur at zero valence state was demonstrated in sulfide oxidation experiments with cell-free extracts from strains of At. thiooxidans [25,28]. Candidates for the not yet identified thiol-bearing OMP are the sulfide-binding 1176

Microbiology Fundamentals

protein isolated from At. ferrooxidans AP 19-3 [29] and several outer membrane proteins that have been associated with sulfur oxidation in strains of At. ferrooxidans [30,31,32].

Figure 4. Proposed biochemical pathway for sulfide and elemental sulfur oxidation in leaching bacteria of the Gram-negative genera Acidithiobacillus and Acidiphilium (see text). ACKNOWLEDGEMENTS This work was supported by grants to W. S. from BMBF via UBA (1490954), DBU (05333), and the Max-Buchner-Forschungsstiftung (DECHEMA e.V.). REFERENCES 1. M. Schedel and H. G. Trüper, Biochim. Biophys. Acta, 568 (1979) 454. 2. A. S. Pott and C. Dahl, Microbiology, 144 (1998) 1881. 3. C. G. Friedrich et al., Appl. Environ. Microbiol., 67 (2001) 2873. 4. D. Rother et al., J. Bacteriol., 183 (2001) 4499. 5. I. Suzuki, Biochim. Biophys. Acta, 110 (1965) 97. 6. T. Emmel et al., J. Gen. Microbiol., 132 (1986) 3415. 7. A. Kletzin, J. Bacteriol., 171 (1989) 1638. 8. A. Kletzin, J. Bacteriol., 174 (1992) 5854. 9. I. Suzuki, Methods Enzymol., 243 (1994) 455. 10. I. Suzuki, Biochim. Biophys. Acta, 104 (1965) 359. 11. M. Silver and D. G. Lundgren, Can. J. Biochem., 46 (1968) 457. 12. T. Sugio et al., J. Bacteriol., 169 (1987) 4916. 13. E. N. Krasil`nikova et al., Mikrobiologiya, 67 (1998) 156. 14. T. Sugio et al., Biochim. Biophys. Acta, 973 (1989) 250. 15. M. E. Mackintosh, J. Gen. Microbiol., 105 (1978) 215. 1177

Microbiology Fundamentals

16. A. P. Harrison Jr., Int. J. Syst. Bacteriol., 31 (1981) 327. 17. T. Rohwerder, Untersuchungen zur enzymatischen Oxidation von Elementarschwefel bei acidophilen Laugungsbakterien, PhD thesis, University of Hamburg, 2002 (www.sub.uni-hamburg.de/disse/849/dissertation.pdf). 18. A. Schippers and B. B. Jørgensen, Geochim. Cosmochim. Acta, 65 (2001) 915. 19. J. Weiß, Ionenchromatographie, VCH, Weinheim, 1991. 20. Anonymous, Deutsche Einheitsverfahren zur Wasser-, Abwasser- und Schlammuntersuchung (DEV), VCH, Weinheim, 1984. 21. T. Sugio et al., Appl. Environ. Microbiol., 58 (1992) 431. 22. J. R. Vestal and D. G. Lundgren, Can. J. Biochem., 49 (1971) 1125. 23. K. Nakamura et al., Biosci. Biotech. Biochem., 59 (1995) 11. 24. G. A. H. de Jong et al., J. Mol. Catal. B, 8 (2000) 61. 25. D. J. W. Moriarty and D. J. D. Nicholas, Biochim. Biophys. Acta, 184 (1969) 114. 26. T. Nübel et al., Arch. Microbiol., 173 (2000) 233. 27. C. Griesbeck et al., Biochemistry, 41 (2002) 11552. 28. D. J. W. Moriarty and D. J. D. Nicholas, Biochim. Biophys. Acta, 197 (1970) 143. 29. T. Sugio et al., Agric. Biol. Chem., 55 (1991) 2091. 30. V. Buonfiglio et al., FEMS Microbiol. Rev., 11 (1993) 43. 31. V. Buonfiglio et al., J. Biotechnol., 72 (1999) 85. 32. N. Ohmura et al., J. Bacteriol., 178 (1996) 5776.

1178

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Sulfate reduction at low pH by mixed cultures of acidophilic bacteria Anna M. Sen, Sakurako Kimura, Kevin B. Hallberg and D. Barrie Johnson School of Biological Sciences, University of Wales, Bangor, LL57 2UW, U.K. Abstract Mixed cultures of acid-tolerant sulfate reducing bacteria (SRB) and other acidophilic bacteria were shown to form stable microbial communities in sulfidogenic bioreactors. Cultures were set up in pH-controlled (pH 2.5-5.5) bioreactors, and run in batch and continuous mode using glycerol or ethanol as electron donor. Reduction of sulfate (and corresponding formation of hydrogen sulfide), acid-consumption and changes in the indigenous microflora (SRB and other bacteria) were monitored. Sparging of the bioreactors with nitrogen effectively removed the H2S, which was used to precipitate copper as CuS (off-line) or zinc as ZnS (within the bioreactor). Following earlier experiments with undefined mixed populations of anaerobic acidophiles, we studied sulfidogenesis at low pH by mixed cultures of Desulfosporosinus spp. (SRB) and an Acidocella-like acidophilic heterotroph. In these cultures, accumulation of acetic acid (the product of incomplete glycerol oxidation by the SRB) was much less than in the undefined mixed cultures, leading to more efficient metal (zinc) immobilization in. Syntrophy between acid-tolerant SRB and acidophilic heterotrophic bacteria appears to be important in mediating sulfidogenesis in acidic liquors.

Keywords: acid mine drainage; acidophiles; bioremediation; metal recovery; sulfate reducing bacteria 1.

INTRODUCTION Studies of the microbiology in extremely acidic liquors, such as acid mine drainage (AMD) and mineral leachates, have focused primarily on aerobic processes, chiefly because the major reactions responsible for mineral dissolution are oxidative. Some acidophiles are able to grow in anoxic environments [1]. Iron respiration (where ferric iron acts as electron acceptor) seems particularly widespread amongst acidophiles. Ironoxidizing moderate thermophiles and obligately heterotrophic Acidiphilium spp. can couple the oxidation of organic substrates to the reduction of ferric iron [2-4]. Acidithiobacillus ferrooxidans can also grow anaerobically by iron respiration, using either reduced sulfur or hydrogen as electron donor [5, 6]. Sulfate is found in mineral leaching environments often at concentrations of many grams per liter. Dissimilatory sulfate reduction might therefore be anticipated to be widespread in acidic anaerobic zones (sediments etc.), and there are a number of reports in the literature documenting sulfidogenesis in these situations [e.g. 7, 8, 9]. However, attempts to isolate and cultivate acid-tolerant/acidophilic sulfate reducing bacteria (aSRB) 1179

Microbiology Fundamentals

from mine waters have mostly been unsuccessful. Most mine water isolates appear to be neutrophilic, and are not active below pH ~ 5 [e.g. 7, 10]. There are a number of reasons for this, including the use of inappropriate media (substrates, such as lactate, used routinely to enrich neutrophilic species may be toxic to acidophiles). Sen and Johnson [11] obtained enrichment cultures of aSRB from sediment samples from a copper mine and from geothermal areas on Montserrat. Sulfidogenesis by these mixed cultures occurred at pH <3, but was more rapid between pH 3 and 5. This paper describes sulfidogenesis by undefined and defined cultures of aSRB, and the immobilization of heavy metals at low pH due to biogenic sulfide. Syntrophy between aSRB and a non-sulfidogenic heterotrophic acidophile is discussed. 2.

MATERIALS AND METHODS

2.1 Establishment and maintenance of enrichment cultures Enrichment cultures containing acidophilic/acid-tolerant strains of SRB were obtained from sediments from the Afon Goch, Parys copper mine (north Wales) and Montserrat, as described elsewhere [11]. These were subcultured in ~ 500 ml of ethanol-containing liquid medium (below) containing ~ 300 ml of porous glass beads (2-4 mm diameter; [12]). The beads were included to act as a surface for bacterial colonization. These cultures were used to inoculate 2 l bench-scale bioreactors (P350; Electrolab, U.K.), which were run in either batch or continuous mode. A schematic diagram of the continuous mode set up is shown in Figure 1. Cultures were maintained at 30°C, and pH control was maintained via addition of either 0.1 M HCl or KOH. Impeller Sample point (input)

MEDIUM PUMPED IN

ORP/pH electrodes

Acid/ alkali

Filter

OFN

in WASTE PUMPED OUT (to sample point output and waste vessel)

2L bioreactor vessel

Biofilm of aSRB on Poraver beads

Gas collection bottles each containing 500 ml copper acetate (10 mM)

Figure 1. Bioreactor system set-up for continuous mode experiments. OFN = oxygenfree nitrogen 1180

Microbiology Fundamentals

2.2 Batch and continuous cultures of mixed (undefined) SRB cultures The growth medium for batch culture experiments contained 10 mM ethanol, 0.002% yeast extract, 10 mM FeSO4, together with basal salts and trace elements [13]. The medium was heat-sterilized prior to addition of filter-sterilized ethanol and FeSO4. The cultures (working volumes, 1.8 l) were sparged daily for 5-10 minutes with oxygen-free nitrogen (OFN) to remove H2S, which might otherwise inhibit the SRB. Samples were removed at regular intervals and analyzed for sulfate, ferrous iron, ethanol, acetate and protein. Bacteria were isolated and enumerated on overlay solid media, as described elsewhere [11]. The bioreactor was run in continuous mode for the Parys mixed culture only. In this case, a modified growth medium was used, with glycerol (at 5 mM) replacing ethanol as electron donor. In addition, the concentration of FeSO4 was lowered to 1 mM, with sulfate concentrations being maintained as earlier by the inclusion of 9 mM K2SO4. The sterile feed medium was adjusted to pH 3.0 with H2SO4, and was added to the culture vessel at 25 ml/h, while that in the bioreactor was maintained at pH 2.5, 3.0, 3.5 or 4.0. In addition, inflow rates of 50 and 100 ml/h were applied in the pH 4.0-maintained culture. Throughout the working day, the reactor vessel was sparged for about 5 min/h with OFN and the off-gas bubbled though gas bottles containing 10 mM copper acetate, causing precipitation of CuS. Concentrations of glycerol and acetate in the growth vessel were also measured in continuous mode experiments. 2.3 Batch cultures of defined mixed SRB cultures During the course of purifying aSRB on solid media, a novel uncolored colony morphology was observed on several overlay plates. Physiological and phylogenetic work confirmed that this was an Acidocella-like isolate (unpublished); bacteria of this type had not previously been found in anaerobic bioreactor cultures. A mixed culture of this isolate, PFBC, and Desulfosporosinus isolates from the Parys culture (isolate P1) and from the Montserrat culture (isolate M1) was grown in batch mode in an anaerobic bioreactor. A modified medium was used in this case, containing 5 mM glycerol, 5 mM ZnSO4 5 mM K2SO4 and 1 mM FeSO4. Yeast extract was replaced by the vitamin mixture proposed by Widdel and Pfennig [14] for SRB, added at 1 ml vitamin solution/l medium. 2.4 Analytical techniques Samples were removed from the bioreactor vessels and filtered through 0.2 µm cellulose nitrate membranes (Whatman, U.K.). Concentrations of ethanol, glycerol and acetate were measured enzymatically (R-Biopharm GmbH, Germany), and protein using the Bradford assay [15]. Sulfate was measured turbidometrically as barium sulfate (Hydrocheck, Cambridge, U.K.), and ferrous iron using the ferrozine reagent [16]. Sulfide production in continuous cultures was estimated by analysis of copper concentrations in the gas bottles by atomic absorption spectrophotometry. Proton consumption by batch and continuous cultures was determined from the amounts of alkali that were required to maintain the cultures at pre-set pH levels. 2.5 Calculations Sulfate reduction rates (SRR) were calculated in mmol SO42- reduced l-1·h-1. Hydraulic residence time (HRT), i.e. the average time that the culture medium was retained in the bioreactor, was calculated by dividing the working volume by the flow rate of feed medium. 1181

Microbiology Fundamentals

3.

RESULTS Mixed cultures of aSRB and other anaerobic bacteria were established readily in enrichment cultures, as described previously [11]. Both the Parys and Montserrat cultures contained aSRB and other bacteria (as evidenced by the formation of black and non-black colonies, respectively, on overlay plates). Since these mixed cultures appeared to be robust (i.e. they were maintained through many sub-cultures) it was decided to carry out initial batch- and continuous-mode experiments with them. Data from a batch culture run (at pH 4.0), using the Parys culture, is shown in Figure 2. Sulfate reduction in this culture corresponded with proton consumption (alkali generation) that was due, at least in part, to production of H2S from sulfuric acid. Ferrous sulfide was not formed in significant amounts (as would be predicted, given the pH of the culture). Interestingly, ethanol oxidation was only about 3 mM, which was much less than the sulfate reduced (~ 10 mM). Similar data were obtained with the Montserrat culture (not shown). In both cultures, there were general, though sporadic, increases in protein concentrations during culture runs, indicating that the cultures were actively growing under these conditions (data not shown).

Figure 2. Sulfate reduction by the undefined Parys consortium, run in batch mode at pH 4.0 Key: , sulfate; S, ethanol; , ferrous iron; , protons consumed Sulfate reduction was observed with both cultures over the pH range 2.5-5.5. Rates of sulfate reduction were, however, affected by culture pH, and the Parys and Montserrat cultures showed difference responses (Figure 3). The Parys culture displayed a pH optimum for sulfate reduction at 4.0. In contrast, rates of reduction were similar with the Montserrat culture between pH 3.0 and 4.0 (and greater than at pH 2.5 and 4.5) and the greatest rate of sulfate reduction was observed at pH 5.5. The shape of the graph obtained with the Montserrat culture indicated that a "second" maximum in sulfate-reduction rate may occur above this pH (Figure 3). Data from the Parys culture run in continuous mode (pH 4.0; hydraulic retention time 16.5 h) is shown in Figure 4. Sulfate concentrations in the outflow liquor (which were very similar to those in the bioreactor vessel) were 4-5 mM lower than in the inflowing liquor, as a result of sulfate reduction. Similarly, glycerol concentrations were 3-4 mM 1182

Microbiology Fundamentals

less in the outflow than in the feed liquor, indicating that it was being utilized as the electron donor for sulfate reduction. Acetic acid was present at about 2 mM in the bioreactor vessel and outflow liquor, presumably as a result of incomplete oxidation of glycerol. In cultures maintained at lower pH, glycerol oxidation again appeared to be coupled to sulfate reduction, though concentrations of inflow/outflow glycerol were more similar at pH 3.0 and 2.5 (indicating less active sulfate reduction) than at pH 3.5 and 4.0. Data from the Parys culture run in continuous mode, showing the effects of culture pH and hydraulic retention times, are shown in Table 1.

Figure 3. Sulfate reduction rates (SRR) for batch cultures of ( ) the Parys and ( ) the Montserrat undefined mixed cultures

Figure 4. Sulfidogenesis by the undefined Parys consortium run in continuous mode (pH 4.0; HRT 16.5 h). Key: ( ) sulfate (in); ( ) sulfate (out); (S) glycerol (in); (U) glycerol (out); (V) acetate (out)

1183

Microbiology Fundamentals

Table 1. Sulfate reduction rates (SRR) in continuous culture experiments with the Parys undefined mixed culture, as affected by culture pH and hydraulic retention times (HRT) pH 2.5 3.0 3.5 4.0 4.0 4.0

HRT (h) 66 66 66 66 33 16.5

SRR (µmol SO42- reduced/h) 40 82 166 134 340 460

Data from batch growth of the defined SRB-Acidocella culture are shown in Figure 5. The inclusion of zinc sulfate in this instance led to the formation of zinc sulfide, thereby reducing the amount of acid consumption (to maintain a low pH) by the culture, due to the reaction: Zn2+ + H2S → ZnS + 2H+ Again, glycerol oxidation was coupled to sulfate reduction, which in turn resulted in the removal from solution of stoichiometric amounts of zinc. It was also interesting that, although some acetic acid was produced during days 1-5, this was rapidly consumed during day 6, with concentrations falling from about 2 to <0.1 mM (Figure 5).

Figure 5. Sulfidogenesis coupled to precipitation of zinc in a defined culture of Desulfosporosinus and Acidocella PFBC, grown in batch mode. Key: ( ) sulfate reduced; ( ) soluble zinc; (S) glycerol; (V) acetate 4.

DISCUSSION Dissimilatory reduction of sulfate to sulfide is of significance in two major respects, both of which are important for the bioremediation of acidic, metal-rich waters, such as AMD. First is the production of net alkalinity, resulting from the conversion of a strong acid (sulfuric) to a weak acid (hydrogen sulfide), and from the generation of bicarbonate. Secondly, sulfate reduction leads to the precipitation of a number of potentially toxic

1184

Microbiology Fundamentals

heavy metals (cadmium, copper, zinc etc.) as highly insoluble sulfides. Both of these key features were illustrated in the present study, where mixed cultures of aSRB and other acidophiles required addition of protons to bioreactor cultures in order to maintain the preset pH levels, and where copper and zinc were both removed from solution (copper offline by sparging the sulfide from the culture vessel with nitrogen, and zinc in situ). Sulfidogenesis is used commercially to remediate acidic waters and to recover metals, such as at the Budelco zinc refinery in the Netherlands [17]. However, these systems use neutrophilic strains of SRB, and it is necessary to preclude contact between the SRB cultures and the acidic wastewaters, which is usually achieved by stripping H2S/HS- from the SRB reactors and allowing the sulfide to contact the waste stream in separate tanks. Considerable savings in costs and engineering complexity could be achieved by using single SRB/waste stream tanks, maintained at low pH to allow selective precipitation of metals (e.g. zinc and copper, but not iron). This work has shown that this can be achieved using acid-tolerant/acidophilic strains of SRB, grown in mixed cultures with other acidophiles. Even though ferrous iron was present in all of the culture media used, FeS was not formed in the pH 4.0 reactors, since its solubility product (4.0 x 10-19 M2) is considerably greater than that of other metal sulfides such as zinc sulfide (1.0 x 10-23 M2), which does precipitate at pH 4.0. Other studies have indicted that, although it is possible to purify aSRB from the consortia, the mixed cultures are inherently more robust than pure cultures (unpublished). One reason for this is that the Desulfosporosinus isolates studied here are incomplete oxidizers, i.e. they produce (and accumulate) acetate from ethanol and glycerol. This is a common trait amongst SRB, but acidophiles have the added problem that organic acids (which occur primarily as undissociated acids in low pH liquors) are toxic at relatively low concentrations [18]. The presence of other bacteria in the anaerobic cultures might therefore serve to lower concentrations of acetic acid to the point at which they are nontoxic. Support for this hypothesis came from the work with the defined population of Desulfosporosinus sp. M1 and Acidocella-like acidophiles. The latter is most closely related to Acidocella sp. WJB3, which is known to be unusually adept (for an acidophile) at metabolizing acetic acid [19]. However, Acidocella spp. are all supposedly obligate aerobes. Attempts to grow pure cultures of strain PFBC anaerobically on acetic acid (or any other substrate) have all failed, yet data from molecular analysis of mixed aSRB/PFBC cultures have shown that the anaerobic aSRB and “aerobic” PFBC grow concurrently in anoxic media (data not shown). Work is currently underway to resolve this apparent conundrum. What was apparent is that acetate concentrations in mixed cultures of Desulfosporosinus M1/Acidocella PFBC were considerably lower than in the "undefined" aSRB consortia (where Bacillus-, Cellulomonas- and Luteococcus-like isolates were detected [13] and Acidocella-like bacteria were not detected). This work has confirmed that SRB may be metabolically active in acidic liquors, and that their potential for remediation of AMD and related wastewaters is considerable. ACKNOWLEDGEMENTS A.M.S. is grateful to the BBSRC (U.K.) and British Nuclear Fuels Limited (U.K.) for provision of a research studentship. K.B.H. and D.B.J. thank the BBSRC (ref. 5/T06496) and the DTI (U.K.; ref. BTL/20/71) for partial support of this work. REFERENCES 1. D. B. Johnson, FEMS Microbiol. Ecol. 27 (1998) 307. 1185

Microbiology Fundamentals

2. T. A. M. Bridge and D. B. Johnson, Appl. Environ. Microbiol. 64 (1998) 2181. 3. T. A. M. Bridge and D. B. Johnson, Geomicrobiol. J. 17 (2000) 193. 4. K. Küsel, T. Dorsch, G. Acker and E. Stackebrandt, Appl. Environ. Microbiol. 65 (1999) 3633. 5. J. T. Pronk, J. C. de Bruyn, P. Bos and J. G. Kuenen, Appl. Environ. Microbiol. 58 (1992) 2227. 6. N. Ohmura, K. Sasaki, N. Matsumoto and H. Saiki, J. Bacteriol. 184 (2002) 2081. 7. J. H. Tuttle, P. R. Dugan, C. B. Macmillan and C. I. Randles, J. Bacteriol. 97 (1969) 594. 8. R. A. Gyure, A. Konopka, A. Brooks and W. Doemel, FEMS Microbiol. Ecol. 73 (1990) 193. 9. D. Fortin, M. Roy, J. P. Rioux and P. J. Thibault, FEMS Microbiol. Ecol. 33 (2000) 197. 10. K. A. Küsel, U. Roth, T. Trinkwalter and S. Peiffer, Environ. Exp. Bot. 46 (2001) 213. 11. A. M. Sen and D. B. Johnson, Biohydrometallurgy and the Environment Toward the Mining of the 21st Century, R. Amils and A. Ballester (eds.), Elsevier, Amsterdam, (1999) 709. 12. Å. Kolmert and D. B. Johnson, J. Chem. Technol. Biotechnol. 76 (2001) 836. 13. A. M. Sen, Ph.D. Thesis, University of Wales, Bangor, (2001). 14. F. Widdel and N. Pfennig, Arch. Microbiol. 129 (1981) 395. 15. M. M. Bradford, Anal. Biochem. 72 (1976) 248. 16. D. R. Lovley and E. J. P. Phillips, Appl. Environ. Microbiol. 53 (1987) 1536. 17. L. J. Barnes, P. J. M. Scheeren and C. J. N. Buisman, Emerging Technology for Bioremediation of Metals, J. L. Means and R. E. Hinchee (eds.), Lewis Publishers, Boca Raton, Fl., (1994) 38. 18. B. Alexander, S. Leach and W. J. Ingledew, J. Gen. Microbiol. 133 (1987) 1171. 19. K. B. Hallberg, Å. K. Kolmert, D. B. Johnson and P. A. Williams, Biohydrometallurgy and the Environment Toward the Mining of the 21st Century, R. Amils and A. Ballester (eds.), Elsevier, Amsterdam, (1999) 719.

1186

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Sulfur assimilation in Acidithiobacillus ferrooxidans J. Valdés1, E. Jedlicki2 and D.S. Holmes1,3 1

*

Laboratory of Bioinformatics & Genome Biology, University of Santiago, Santiago, Chile, 2 Program of Cellular and Molecular Biology, University of Chile, Santiago, Chile 3 Millenium Institute of Fundamental and Applied Biology, Santiago, Chile

Abstract A bioinformatic analysis of the nearly complete genome sequence of Acidithiobacillus ferrooxidans, available from The Institute for Genome Research (TIGR) and Integrated Genomics (IG), reveals the presence of a number of genes potentially involved in sulfur uptake and metabolism. The coding potential of these genes is compared to similar genes for known sulfur metabolic pathways in other microorganisms. In addition, the proposed sulfur assimilation genes of A. ferrooxidans are shown to be present in operons in an organization similar to that found in some other organisms. Genes potentially involved in the regulation of sulfur uptake and assimilation have also been detected. Based upon these observations, we present a metabolic model of sulfur uptake and assimilation in A. ferrooxidans that is substantially similar to the genetic organization and metabolic pathway structure that has been experimentally demonstrated to function in other organisms. An interesting exception to this similarity is the observation that A. ferrooxidans appears to have the genetic potential to encode for two different pathways of sulfur activation whereas many organism only have one such pathway. In addition, an interesting gene fusion event seems to have taken place in A. ferrooxidans involving the joining of a cysN gene, encoding a sulfurylase enzyme subunit B, with a cysC-like gene encoding an APS activity.

Keywords: sulfur assimilation, sulfur regulation, cysteine biosynthesis, genome analysis, Acidithiobacillus ferrooxidans 1.

INTRODUCTION The assimilation of sulfur is an essential process in all cells being required for the formation of the amino acids, methionine and cysteine. Sulfur is also used for the construction of FeS centers in proteins involved in electron transfer processes (1) and is found in certain membrane components such as the nod factor of the plant symbiont Rhizobium sp. (2) and in various sulfolipids of Mycobacterium sp. that are thought to bevirulence factors (3). * Corresponding author: David Holmes: [email protected]. Work supported by Fondecyt No. 1010623 and the Millenium Institute for Fundamental and Applied Biology, Santiago, Chile. We thank the Institute of Genome Research (TIGR) and Integrated Genomics, Inc. (IG) for the use of their partial sequence of the Acidithiobacillus ferrooxidans genome.

1187

Microbiology Fundamentals

The process of sulfur assimilation, in all organisms begins with the uptake of sulfate ions from the environment through the cellular membrane via an ABC-type membrane pump mechanism. Sulfate ions are generally available in most environments and are abundantly available in mining environments rich in sulfuric acid. After entry into the cell, the sulfate is made biologically available by activation, a process that involves the reaction between sulfate and ATP, catalyzed by the enzyme ATP sulfurylase to yield adenosine 5´-phosphosulfate (APS) and PPi. A second enzyme, APS kinase, can further phosphorylate APS with another ATP to give 3´-phosphoadenosine 5´phosphosulfate (PAPS)(1). In plants and phototrophic bacteria the intermediate of assimilative reduction is APS, but in fungi and some chemotrophic bacteria it is PAPS. Depending on the organism, these intermediates are then further reduced and fed into pathways that are shared in all organisms leading to the formation of the amino acids cysteine and methionine. The reason for the evolution of two mechanisms for sulfur assimilation by different organisms is unclear. The enzymes involved in the formation of APS or PAPS show significant sequence similarity and may evolve from a common ancestor. We have carried out a bioinformatic analysis of the almost complete genome sequence of A. ferrooxidans in an attempt to identify candidate genes potentially involved in sulfur assimilation. Using this information, we suggest possible biochemical pathways for sulfur assimilation and for the regulation of sulfur uptake. It is beyond the scope of this paper to connect the proposed sulfur assimilation pathways with proposed sulfur oxidation genes and pathways in A. ferrooxidans. However, the present paper will aid future studies designed not only to make these connections but ultimately to deduce the interconnected regulatory mechanism that balance sulfur assimilation with sulfur oxidation in a variety of environmental situations. 2.

METHODS Known metabolic pathways of sulfur assimilation were obtained from BIOCYC (http://biocyc.org:1555/META/server.html), KEGG (http://genome.ad.jp/kegg/kegg4. html) and ERGO (http://wit.integratedgenomincs.com/WIT2/CGI). Amino acid sequences derived from genes identified as being involved in sulfate acquisition were used as query sequences to search the partial genome sequence of A. ferrooxidans ATCC 23270 in the TIGR (http:// www.tigr.org/) and ERGO data bases using TBlastN and BlastP respectively. When a prospective candidate gene was identified in TIGR or ERGO its predicted amino acid sequence was then used to formulate a BlastP (http://www.ncbi.nlm.nhi.gov/BlastP/) search of the non-redundant database at NCBI. Only bidirectional best hits were accepted as evidence for putative homologs. Candidate genes and their translated proteins were further characterized employing the following bioinformatic tools available in the web: primary structure similarity relations (http://www.ebi.c.uk/ClustalW/), secondary structure predictions JPred http://www. compbio.dundee.ac.uk/Software/JPred/jpred.html), transmembrane predictions (http://www.ch.embnet.org/software/TMPRED_form.html), motif predictions (http:// www.blocks.fhcrc.org/, http://www.ebi.ac.uk/interpro/, and domain predictions (http:// www.tolouse.inra.fr/prodom.htlm), and prediction of protein localization sites (http://psort.nibb.ac.jp/).

1188

Microbiology Fundamentals

3.

RESULTS AND DISCUSSION Sulfur assimilation, in most organisms, typically starts with sulfate uptake from the environment by a membrane complex that includes an ABC-type membrane pump (Figure 1). A bioinformatic analysis of the predicted genes of A. ferrooxidans reveals the presence of candidate genes that could encode the proteins for such a complex. One of these genes encodes a possible sulfate-permease protein and lies adjacent to a gene potentially encoding an ABC-type ATP binding membrane pump and a protein of unknown function. The three genes appear to be organized as an operon (data not shown) lending support to the idea that they encode proteins involved in sulfate uptake. On the other hand a positive identification of genes and proteins involved in sulfate uptake is difficult because ABCtype membrane pump proteins belong to a large family of paralogous genes that encode proteins with similar sequences and structures but that differ in the substrate that they transport. Thus bioinformatic identification of an ABC transporter for sulfate uptake by sequence similarity is insufficient to pinpoint its function and experimental validation will be essential. On the other hand, genes involved in the subsequent activation and utilization of sulfur do not appear to belong to extensive paralogous families. In addition, the genes and pathways of sulfur assimilation are quite conserved among organisms. These considerations make it possible not only to identify candidate orthologous genes in A. ferrooxidans by bioinformatic analysis but also to have a reasonable degree of confidence in such assignations. It also allows us to propose a biochemical model of sulfur assimilation for A. ferrooxidans. It should be stressed that such a model requires experimental validation before it can be accepted. Such experiments are now underway and will be reported in the future. However, it is useful to note that, in the absence of well developed methods of gene transfer in A. ferrooxidans and the concomitant paucity of genetic analyses, metabolic models derived by bioinformatics provide an important way to investigate the potential biochemistry and genetics of the organism. Such studies have already revealed valuable information regarding the amino acid biosynthesis (4), nitrogen metabolism, metal fluxes and other characteristics of A. ferrooxidans. (5). Table 1. Identification and some properties of the genes predicted to be involved in sulfur assimilation in A. ferrooxidans. (A) Proposed gene name. (B) Putative function assigned to each gene. (C) The best BlastP hit of each candidate gene to orthologs in GenBank and the organism in which the ortholog is found and (D) the percent similarity of the amino acid sequence of the candidate gene with its ortholog from GenBank A. Gene CysD1 CysD2 CysN1 CysN2 CysN3 CysH CysI CysJ CysB1 CysB2

B. Putative function assigned ATP sulfurylase, catalitic subunit ATP sulfurylase, catalitic subunit ATP sulfurylase, GTP-binding subunit ATP sulfurylase/Adenylylsulfate kinase ATP sulfurylase/Adenylylsulfate kinase APS reductase Sulfite reductase hemoprotein Sulfite reductase flavoprotein Transcriptional regulator Transcriptional regulator

C. Best Blastp hit and organism Neisseria meningitidis Mesorhizobium loti Neisseria meningitidis Mesorhizobium loti Coxiella burnetii Allochromatium vinosum Bacillus subtilis Bacillus subtilis Pseudomonas aeruginosa Pseudomonas aeruginosa

D. % Similarity 71 78 69 72 61 74 67 59 69 52 1189

Microbiology Fundamentals A. Gene CysQ CysA CysW CysM

B. Putative function assigned PAPS catabolism Sulfate /thiosulfate transport Sulfate/ thiosulfate transport O-acetyl serine (thiol)-lyase-B

C. Best Blastp hit and organism Aquifex aeolicus Mesorhizobium loti Halobacterium sp. Pseudomonas putida

D. % Similarity 72 53 54 70

In contrast to many organisms, where sulfur activation and subsequent reduction occurs either via APS or PAPS, candidate genes have been identified A. ferrooxidans that are predicted to encode enzymes involved in both APS and PAPS formation (Figure 1 and Table 1). In addition, genes have been identified for proteins involved in the subsequent utilization of APS and PAPS and also in the regulation of the uptake and incorporation of sulfur (Figure 1 and Table 1).

Figure 1. Proposed biochemical pathway and genes for the uptake of sulfate and the assimilation of sulfur in A. ferrooxidans based upon bioinformatics analysis of its genome. Potential genes are shown in parentheses. The reason for the dotted line that connects PAPS and sulfite is discussed in the text. Also shown is the feedback inhibition of serine acetyl transferase by cysteine and the regulation of the sulfur uptake and assimilation genes by N-acetyl-L-serine which is discussed in the text The levels of similarity (Table 1) of the predicted A ferrooxidans amino acid sequences to known proteins involved in sulfur activation are sufficient to allow a reasonable degree of confidence in the assignments made. In addition, many of the candidate sulfur assimilation genes appear to have an operon-like organization similar to those found in other organisms (Figures 2 and 3). Operon-like organization is a strong predictor of similar function for the genes within the operon. Thus, the finding that many 1190

Microbiology Fundamentals

of the candidate sulfur assimilation genes of A. ferrooxidans are display such organization strengthens the likelihood of the assignments of gene function which we have made using bioinformatics predictions based on comparative protein sequence similarity. The formation of APS in A. ferrooxidans could be accomplished by the protein products of cysD1 and cysN1, which, in E. coli, encode the heterodimeric enzyme ATP sulfurylase (6). In both A. ferrooxidans (Figure 2) and E. coli cysD1 and cysN1 appear to be organized in an operon. However, the E. coli operon also includes the gene cysC that catalyzes the phosphorylation of APS to give PAPS. This gene appears to be absent in the equivalent A. ferrooxidans operon. However, the A. ferrooxidans genome has an additional two copies of cysN1, which we term cysN2 and cysN3 (Figures 1,2 and Table 1). Interestingly, a detailed examination of the predicted protein sequences encoded by cysN2 and cysN3 reveals the presence of an additional domain in their C terminal regions that exhibit significant sequence similarity to the E. coli cysC (Figure 3). This raises the possibility that the A. ferrooxidans cysN2 and cysN3 encode bi-functional proteins formed by a gene fusion event between cysN1 and cysC and that the resulting fused protein can assume the function of the missing cysC catalyzing the formation of PAPS. A similar bifunctional protein has also been observed in Rhizobium sp. (7). Additional support for this conjecture comes from the observation that a cloned segment of A. ferrooxidans DNA was able to complement an E. coli cysC mutant, demonstrating that the organism has the capacity to encode a cysC-like function (8). It appears that a transpose-encoding element (Tpa, Figure 2) has inserted between cysN2 and cysQ that may interrupt the coordinated regulation of the cysDNQ operon. The significance of this observation is not known. Upstream of cysD1 and cysN1 in the A. ferrooxidans genome there are candidate genes for cysH, cysI and cysJ organized in an operon-like structure (Figure 2). Many microorganisms, such as E. coli and S. typhimurium, have a similar gene organization (1). cysI and cysJ encode a heterodimeric sulfite reductase with the form α4β8. This catalyzes the reduction of sulfite to sulfide which s the final reduction state of sulfur for its assimilation and the putative cysI and cysJ in A. ferrooxidans can be assigned with reasonable confidence. The role of cysH in A. ferrooxidans is more enigmatic. In E. coli, cysH encodes a PAPS reductase without APS reductase activity, but in the plant A. thaliana the cysH product is exclusively an APS reductase. In the bacteria Pseudomonas putida and Mycobacterium spp., cysH exhibits both APS and PAPS reductase activity, but with differing levels of activity (9). The A. ferrooxidans cysH ortholog exhibits a motif characteristic of APS reductase (data not shown) (10), suggesting that it is has at least an APS reductase activity but leaves open, for the present, the possibility that it also has a PAPS reductase activity. For this reason we have drawn in Figure 1 a solid line connecting APS reduction with cysH, signifying a reasonable level of confidence in the assignment, and we have placed a dotted line connecting PAPS reduction with cysH to emphasize that some ambiguity remains regarding cysH function. The final step in sulfur assimilation is the synthesis of L-cysteine from O-acetyl-Lserine and sulfide. This reaction is catalyzed by one or both of two O-acetyl-L-serine (thiol)-lyase isozymes, designated A and B and encoded by cysK and cysM genes in S. typhimurium and E. coli (1). A. ferrooxidans has a potential cysM ortholog that is predicted to accomplish this final step in sulfur assimilation (Table 1, Figures 1 and 2).

1191

Microbiology Fundamentals

Figure 2. Examples of two operon-like organizations of the candidate genes of A. ferrooxidans postulated to be involved in sulfate uptake and sulfur assimilation. The proposed function of the various genes designated with prefix "cys" can be found in Figure 1. tpa = probable transposase. hyp = hypothetical gene of unknown function

Figure 3. Comparison between the organization and function of the genes found in the cysDNC operon of E. coli with similar genes of A. ferrooxidans. In two instances in A. ferrooxidans a cysN gene (N2 and N3), potentially encoding a sulfurylase subunit, appears to be fused with a gene with significant similarity to the cysC gene of E. coli encoding an APS activity In addition to exhibiting potential orthologs of genes involved in sulfur assimilation, A. ferrooxidans also appears to have characteristic mechanisms for the regulation of the relevant genes and enzymes. In plants and enteric bacteria the activity of the sulfur uptake and assimilation genes is regulated by a combination of feedback inhibition of serine transacetylase by cysteine and a gene regulatory system called the cysteine regulon (1). Excess cysteine can bind to acetylserine transferase reducing the activity of this enzyme, which, in turn, limits the production of O-acetyl-L serine. Since this latter compound is a precursor of cysteine its reduction causes a decrease of the synthesis of cysteine. Since A. ferrooxidans appears to have the gene encoding acetylserine transferase it can be assumed that it is subject to a similar feedback inhibition mechanism. On the other hand, a reduction in the availability of sulfide leads to the accumulation of O-acetylserine, which spontaneously isomerizes to N-acetylserine. This latter compound binds to cysB and the resulting complex serves as an activator of several genes involved in sulfate uptake and sulfur assimilation (11, 12). cysB is a transcriptional regulator that belongs to the lysR family of transcriptional activators. A ferrooxidans has 1192

Microbiology Fundamentals

two copies of this gene both of which display the characteristic helix-turn-helix motif. of this family. Binding of N-acetylserine to cysB promotes the activation of genes involved in sulfate uptake and sulfur assimilation (1). A third mechanism of regulation involves the control of levels of PAPS carried out by the cysQ gene product. Reports indicates that accumulation of PAPS is toxic for the cell (13). In A. ferrooxidans we found an ortholog of cysQ associated in an operon-like organization with cysD2 and cysN2 genes (Figure 2). A similar organization can be detected in Brucella abortus by inspection of the Integrated Genomics Inc. database of genome sequences (http://wit.integratedgenomics.com/ WIT2/CGI). 3.

CONCLUSIONS Genes and enzymes predicted to be responsible for sulfur uptake and assimilation in A. ferrooxidans exhibit significant similarity, including conservation of key protein motifs, to those found in a number of other microorganisms. This permits a metabolic model to be constructed that suggests that sulfur assimilation in A. ferrooxidans exhibits substantial similarity to other organisms. The operon-like organization of these genes in A. ferrooxidans is also similar to that observed in other microorganisms. In addition, A. ferrooxidans has a repertoire of genes predicted to be involved in the regulation of sulfur assimilation and this characteristic, together with the operonic organization of the relevant genes, implies that key aspects of the genetic and biochemical regulation of sulfur assimilation are shared between A. ferrooxidans and other organisms. However, two important differences between A. ferrooxidans and other organisms stand out. First, it appears that A. ferrooxidans may have encoded enzymes that permit sulfur to be activated by both APS and PAPS intermediates, a characteristic that is shared by some but not all organisms. Secondly, compared to most organisms, A. ferrooxidans exhibits a probable gene fusion event in which an ATP sulfurylase-encoding gene has been joined to a gene encoding an APS kinase activity. The recognition and analysis of gene fusion events are becoming increasingly important for building models of gene evolution. In addition, gene fusions, known as Rosetta stone sequences, serve as powerful predictors of gene and protein function in bionformatic analyses (14). The model of sulfur assimilation shown in Figure 1 was inferred by a range of bioinformatic inferences, including gene similarity comparisons, detection of possible operons and identification of putative genes involved in the regulation of sulfur assimilation. However, the model requires experimental validation, some of which is now underway in our laboratory. In addition to sulfur assimilation, A. ferrooxidans uses sulfur as an energy and electron source. Several laboratories are actively engaged in deducing the transduction of sulfur to energy in A. ferrooxidans and a comprehensive model of sulfur utilization this organism will require an understanding of the metabolic and regulatory connections between the two pathways. REFERENCES 1. Kredich, N. M. Biosynthesis of cysteine, Escherichia coli and Salmonella typhimurium: cellular and molecular biology, In F. C. Neidhardt, J. L. Ingraham, K.B. Low, B. Magasanic, C. M. Schaechter, and H. E. Umbarger (ed.). vol. 1, pgs. 419-428. (1987). 2. Dénarié, J., Debellé, F, and Rosemberg, C. Annu. Rev. Microbiol. 46, pgs. 497-531. (1992). 1193

Microbiology Fundamentals

3. Williams, S.J., Senaratne, R. H., Mougous, J. D., Riley, L. W., Bertozzi, C. R. J. Bacteriol. 277, pgs.32606-32615.(2002) 4. Selkov, E., Overbeek, R., Kogan, Y., Chu, L., Vonstein, V., Holmes, D., Silver, S., Haselkorn, R., Fonstein, M. PNAS. 7, pgs. 3509-3514. (2000). 5. Barreto, M., R. Quatrini, S. Bueno, C. Arriagada, J. Valdes, S. Silver, E. Jedlicki and D. S. Holmes. (In press, Hydrometallurgy, 2003). 6. Leyh, T. S., Taylor, J. C., Markman, G. D. J. Biol. Chem. 263, pgs. 2409-2416. (1988) 7. Schwedock, J. S., Liu, C., Leyh, T. S., Long, S. R. J. Bacteriol. 176, pgs. 7055-7064. (1994). 8. Fry,I. J., and E. Garcia. In J. Salley, R. G. L. McCready, and P. L. Wichlacz (ed.), Biohydrometallurgy. pgs. 171-185. (1989). 9. Bick, J. A., Dennis, J. J., Zylstra, G. J., Nowack J, Leustek T. J. Bacteriol. 182, pgs.135-142. (2000). 10. Kopriva, S., Buchert, T., Fritz, G., Suter, M., Benda, R., Schunemann, V., Koprivova, A., Schurmann, P., Trautwein, AX., Kroneck, PM., Brunold, C. J. Biol. Chem. 277. pgs. 21786-21791. (2002). 11. Hryniewicz, M. M., and Kredich, N. M. J. Bacteriol. 173, 5876-5886. (1991). 12. Monroe, R.S., Ostrwski, J., Hryniewicz, M. M., Kredich, N. M. 172, pgs.6919-6929. (1990). 13. Neuwald, A. F., Krishnan, B. R., Brikum, I., Kulakauskas, S., Suziedelis, K., Tomcsanyi, T., Leyh, T. S., Berg, D.E. J. Bacteriol. 174, pgs. 415-425. (1992). 14. Marcotte, E. M., Pellegrini, M., Ng HL., Rice, D. W., Yeates, T. O., Eisenberg, D. Science.285, pgs.751-3. (1999).

1194

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Survival of acidophilic bacteria under conditions of substrate depletion that occur during culture storage D. Barrie Johnsona, Debby F. Bruhnb and Francisco F. Robertob a

b

School of Biological Sciences, University of Wales, Bangor, LL57 2UW, U.K. Lockheed Idaho Technologies Co., Idaho National Engineering and Environmental Laboratory, P.O.Box 1625, Idaho Falls, Idaho, 83415-22032

Abstract Survival of four mesophilic acidophilic bacteria (two chemolithotrophs, Acidithiobacillus ferrooxidans and Leptospirillum ferrooxidans, and two heterotrophs, "Ferrimicrobium acidiphilum" and Acidiphilium SJH) was investigated during 500 days of starvation stress. Counts of plateable At. ferrooxidans remained fairly stable for 25 days following completion of ferrous iron oxidation, after which they declined; however, 2 x 104 viable cells/ml stored culture were still present after 500 days of starvation stress. In contrast, numbers of viable L. ferrooxidans went into immediate decline following iron oxidation, and no viable planktonic cells were detected 45 days after iron oxidation. Pure cultures of the two heterotrophic acidophiles displayed similar trends to At. ferrooxidans. When mixed populations of acidophilic mesophiles were subjected to long-term starvation stress, mortality rates of At. ferrooxidans and Acidiphilium SJH were greater, and those of L. ferrooxidans and "F. acidiphilum" less than in corresponding pure cultures. The significance of metabolic diversity, sensitivities to organic compounds and scavenging potential on the survival of acidophilic bacteria is discussed.

Keywords: acidophilic bacteria; mixed cultures; pH stress; starvation stress 1.

INTRODUCTION Bacteria that inhabit acidic, metal-rich environments are physiologically diverse, in their carbon assimilation, temperature-response characteristics and abilities to catalyze the dissimilatory transformations of iron and sulfur [1]. Interest in this group stems primarily from their role in the genesis of acid mine drainage [2] and in their exploitation in mineral processing operations [3]. Besides being characterized by low pH (generally 2-4), environments populated by acidophilic microorganisms generally contain large concentrations of sulfate, iron, other (heavy) metals and metalloids (depending on the geochemistry of the local environment), and small concentrations of dissolved organic carbon. Bacteria such as the iron-oxidizing chemolithotrophs Acidithiobacillus ferrooxidans and Leptospirillum spp. are obligate acidophiles, and are inactive in neutral and mildly acidic environments. A variety of interactions between acidophilic microorganisms have been described [4]. For example, mutualistic associations occur between autotrophic and heterotrophic acidophiles, whereby utilization of organic materials (e.g. leakage products from autotrophic cells) by the latter "detoxify" the 1195

Microbiology Fundamentals

environment for acidophilic autotrophs, which are often highly sensitive to small molecular weight organic compounds. The effects of nutrient stress and substrate depletion on the survival of bacteria are of fundamental importance in microbial ecology. Long-term survival in environments that are depleted in nutrients, and which also may be stressful in other ways (e.g. pH, temperature), may result from bacteria having specific survival strategies. Although acidophilic bacteria are important in ore leaching and genesis of acid mine drainage pollution, relatively little has been published on their abilities to survive starvation stress. Zychlinsky and Matin [5] found that survival of the acidophilic mixotroph Acidiphilium acidophilum during "long-term starvation" (up to 22 days) was similar to that reported for neutrophilic bacteria. In this paper, we describe the responses of four mesophilic acidophiles to starvation stress for periods of up to 500 days. 2.

MATERIALS AND METHODS

2.1 Bacteria Survival of four mesophilic acidophiles was investigated: (i) At. ferrooxidans (ATCC 23270; the type strain); L. ferrooxidans (strain CF12, isolated from the Noranda Blackbird-cobalt mine, Cobalt, Idaho); (iii) "Ferrimicrobium acidiphilum" (strain T-23 [1]); (iv) Acidiphilium SJH [1]. Two of these bacteria are autotrophs (At. ferrooxidans and L. ferrooxidans) and two are heterotrophs ("F. acidiphilum" and Acidiphilium SJH). Three species (At. ferrooxidans, L. ferrooxidans and "F. acidiphilum") oxidize ferrous iron to ferric, and three (Acidiphilium SJH, F. acidophilum and At. ferrooxidans) can reduce ferric iron to ferrous in oxygen-limited environments. 2.2 Experimental protocols Pure cultures of each acidophile, and a mixed culture containing all four, were grown in 20 mM ferrous sulfate liquid medium, poised initially at pH 1.8 (to avoid precipitation of ferric compounds in oxidized cultures). Cultures (250 ml in 500 ml conical flasks) were incubated (shaken) for three days at 30°C, and subsequently stored at room temperature (23 ± 2°C) in the dark. Flasks were loosely covered with aluminum foil to reduce evaporation losses. A second experiment was set up to monitor the effect of pH on the survival of mesophilic acidophiles under conditions of nutrient depletion. Individual cultures of the four bacteria were grown in appropriate liquid media (20 mM ferrous sulfate for At. ferrooxidans and L. ferrooxidans; 20 mM ferrous sulfate/0.02% (wt/vol) yeast extract for "F. acidiphilum"; 0.002% (wt/vol) yeast extract for Acidiphilium SJH). When cell numbers had reached approximately 107/ml, bacteria were harvested aseptically by centrifugation, washed, and re-suspended in acidified (pH 2.5) basal salts. Aliquots (0.5 ml) of cell suspensions were added, separately, to basal salts solutions (200 ml in 500 ml flasks) adjusted at either pH 2.0, 5.0 or 8.0, using dilute H2SO4 or NaOH. Cultures were incubated, unshaken, at room temperature in the dark. All experiments used duplicate cultures. 2.3 Analytical techniques Total numbers of bacteria were assessed by acridine orange direct counts (AODCs) using a modification of the method of Hobbie et al. [6]. This involved filtering culture aliquots through black polycarbonate membranes (0.2 µm pore size; Nuclepore, U.S.A.), 1196

Microbiology Fundamentals

rinsing first with acidified (pH 1.0; H2SO4) distilled water and then with alkaline (pH 11.0; NaOH) water, fixing cells with 4% (v/v) glutaraldehyde (10 mins.) followed by a second pH 11.0 water rinse, staining with 0.01% (w/v) acridine orange solution (pH 11.0; 6 mins.) and a final rinse in distilled water. Counts of viable bacteria were made by plating diluted culture aliquots onto selective solid media [7]. These were ferrous iron overlay medium (Feo) for enumerating At. ferrooxidans, L. ferrooxidans and "F. acidiphilum", and yeast extract medium for Acidiphilium SJH. Identification of bacteria was made on the basis of differences in colony morphologies, combined with microscopic examination, as described previously [7]. To prepare samples for scanning electron microscopy (SEM), culture aliquots (containing 106-107 cells/ml) were filtered through Nuclepore membranes (0.2 µm pore size) and rinsed, sequentially, with 0.1 M H2SO4, acidified (pH 2.0) basal salts and "nanopure" water. After fixing with 5% (v/v) glutaraldehyde, cells were dehydrated by flushing with solutions of ethanol (30%, 60%, 90%, 100% (x2); all vol/vol) and critical point dried. Bacteria were viewed in an Amray Model 1830 SEM; cell dimensions were recorded and biovolumes calculated. Other analyses included determination of ferrous iron concentrations (by titration with 1 mM KMnO4 in dilute H2SO4), dissolved organic carbon (samples pre-filtered through 0.2 µm membranes and DOC estimated using an OI Corporation Model 700 DOC analyzer), and pH using a glass electrode. 3.

RESULTS

3.1 Survival trends in pure cultures of acidophiles The acidophilic bacteria used in these experiments varied widely in their abilities to survive under conditions of substrate depletion. Their survival was affected by the presence of other bacteria (i.e. mixed versus pure cultures) and by environmental factors. Figure 1 shows how total counts of all four acidophiles remained fairly constant during the first 100 days of substrate depletion while, in contrast, counts of cultivatable bacteria differed markedly between isolates. Counts of plateable At. ferrooxidans remained fairly static for ca. 25 days following completion of iron oxidation (which took 3 days) after which time they declined, though significant numbers of cultivatable cells (ca. 2 x 104/ml) were still present some 500 days after the start of the experiment. In contrast, plate counts of the other iron-oxidizing chemolithotroph, L. ferrooxidans, declined dramatically following iron oxidation, and no cultivatable bacteria were recovered on day 45. However, cultures plated after 87 and 100 days did produce colonies (confirmed as L. ferrooxidans), though none was detected beyond these times. Viable counts of pure cultures of the two acidophilic heterotrophic isolates also displayed interesting differences. Plate counts of "F. acidiphilum" showed an initial decline in contrast to Acidiphilium SJH whose numbers increased during this time, though these stabilized from day 28 and, by the end of the experiment, viable counts of "F. acidiphilum" were similar to those of both At. ferrooxidans and Acidiphilium SJH. One difference between "F. acidiphilum" and the other iron-oxidizing acidophiles is the former's more protracted ferrous iron oxidation when cultured in "inorganic" media. Whereas iron oxidation was complete in cultures of At. ferrooxidans and L. ferrooxidans by the time that cultures were put into storage (3 days after inoculation), residual ferrous iron concentrations in cultures of "F. acidophilum" were 7 mM (after 9 days) and 3 mM (after 17 days). Slow ferrous iron oxidation is a feature of iron-oxidizing heterotrophs when grown in the absence of organic carbon (D.B. Johnson, unpublished data).

1197

Microbiology Fundamentals

Figure 1. Effect of starvation-stress on, (a) total bacterial numbers (AODCs) and (b) culturable bacteria. Symbols: (S) At. ferrooxidans; () L. ferrooxidans; (z) F. acidophilum; (T) Acidiphilium SJH. Error bars represent standard deviations Table 1. Changes in cell dimensions* of acidophilic bacteria stored under conditions of substrate limitation

Three of the acidophiles displayed net decreases in cell biovolumes after 37 days compared with those recorded at day 3 (Table 1). The single exception was Acidiphilium SJH, whose cells were longer and thinner at day 37 but which showed no net change in biovolume. Decrease in mean cell biovolume was smallest for At. ferrooxidans among the iron-oxidizing acidophiles. Net biovolume reductions were similar for L. ferrooxidans and "F. acidophilum" during this time; however, while L. ferrooxidans cells had shrunk by similar proportions in both length and width, those of "F. acidophilum" were much shorter than at day 3, though of similar width. 1198

Microbiology Fundamentals

3.2 Survival trends in mixed cultures of acidophiles Patterns of bacterial survival were subtly different when the acidophiles were grown in mixed culture (Fig. 2). In particular, numbers of plateable L. ferrooxidans declined at a slower rate than was observed in pure cultures, and plate counts of "F.acidophilum" increased (rather than decreased) during the first 25-30 days of culture storage. In contrast, numbers of viable At. ferrooxidans and Acidiphilium SJH went into more immediate decline than in pure cultures of these acidophiles. As with pure cultures, after 500 days incubation viable cells of all acidophiles (except L. ferrooxidans) were detected in mixed cultures, with "F. acidiphilum" being the most abundant viable acidophile.

Figure 2. Comparison of plate counts of acidophiles, grown in pure and mixed cultures. Data shown are for changes in populations that occurred during the first 45 days of starvation-stress. Error bars represent standard deviations. (a) At. ferrooxidans; (b) L. ferrooxidans; (c) `F. acidiphilum’; (d) Acidiphilium SJH 3.3 Survival trends of acidophiles at different pH The pH of the substrate-depleted media in which cultures of bacteria were maintained had a major influence on the survival rates of the mesophilic acidophiles. Because of the poor buffering capacity of the basal salts solutions, the addition of the bacterial suspensions caused a shift in pH in all cases, resulting in pH values of 2.15, 4.10 and 6.20 for media adjusted initially to pH 2.0, 5.0 and 8.0, respectively. Subsequent changes in pH in these cultures were minor (<0.1 pH unit). Concentrations of dissolved organic carbon (ca. 4-5 mg/L at the start of the experiment) were about 50% of levels recorded in the first experiment since, in the latter, the culture liquor was spent bacterial medium (which would have accumulated cell leakage products etc.) rather than basal salts solution. Changes in viable bacterial counts recorded in this experiment are shown in Fig. 3. No colonies of L. ferrooxidans were recovered from any culture in this experiment, indicating either that the methodology used in harvesting cells of this acidophile proved lethal, or 1199

Microbiology Fundamentals

that L. ferrooxidans lost viability very rapidly in the basal salts solutions. At pH 2.15, the overall trends (Fig. 3a) were similar to those found in experiment 1. Viable counts of "F. acidiphilum" exceeded those of other bacteria from day 14 onwards, and plate counts of this heterotrophic iron-oxidizer remained fairly stable throughout the 100 days of culture storage. Viability of At. ferrooxidans fell sharply during the first 40 days of culture storage, but numbers stabilized at about 104 cells/ml by day 100. Counts of viable Acidiphilium SJH, in contrast, were fairly stable for the first 40 days of storage, but then went into severe decline, and no viable cells were detected at day 100. At pH 4.1, patterns of survival of the three acidophiles were quite different, with both At. ferrooxidans and Acidiphilium SJH displaying more prolonged cell viability than at the lower pH; in contrast, plate counts of "F. acidiphilum" declined rapidly and from day 72 onwards no viable cells of this acidophile were detected (Fig. 3b). At the highest pH (6.2), the mortality rate of "F. acidiphilum" was even more pronounced. At. ferrooxidans also died out more rapidly at this pH than at either pH 2.15 or 4.1, but Acidiphilium SJH survived in reasonable numbers (ca. 103-104/ml) throughout the period of culture storage (Fig. 3c). 4.

DISCUSSION Microorganisms have developed a variety of strategies for surviving times of depletion of substrate or of a particular essential nutrient [8]. This may involve the possession and expression of "survival genes", cell proliferation, and reduction in cell size (producing "ultramicrocells" of <0.3 µm diameter). Various methods have been used to differentiate between viable and non-viable cells in a population [9]. The use of plate counts to enumerate viable cells has been criticized, since bacteria may be able to take up and metabolize nutrients without resuming growth and division under the imposed conditions of plating. Oliver [10] described "viable but not culturable" cells as those that can be demonstrated to be metabolically active but which are incapable of sustaining cellular division in media that normally support growth. The solid media and experimental techniques used in the present study have been developed over several years specifically to produce high plating efficiencies of acidophilic bacteria from environmental materials (as well as to facilitate identification of isolates). Numerical estimations of iron-oxidizing acidophiles in laboratory cultures and environmental samples (where bacteria are frequently in a state of substrate limitation) as determined using "overlay" plates have been shown to be similar or greater than the "Most Probable Number" approach [7]; the latter uses liquid media that might be considered to be more conducive to the growth of stressed cells. It is worth noting that the stored cultures that produced zero plate counts of L. ferrooxidans also proved negative (i.e. no iron oxidation) when samples were subcultured in liquid media. Oliver [10] found that "starved" cells responded quickly to the reversal of nutrient limitation whereas the response in viable non-culturable cells was slow; zero response, as described for L. ferrooxidans, indicates that cells were indeed nonviable. Trends of starvation-survival displayed by the four species of mesophilic acidophiles used in the present work were often very different. The chemolithotrophic iron-oxidizer, L. ferrooxidans displayed much more limited longevity than At. ferrooxidans. This may be due either to an inherently greater capacity of At. ferrooxidans to survive in the absence of extraneous substrates (using storage reserves etc.) or to its greater metabolic versatility. Ferrous iron is the only substrate known to be utilized by L. ferrooxidans; whilst it is possible that some ferrous iron might have been regenerated during culture storage, this is highly unlikely. Even though both Acidiphilium SJH and "F. acidiphilum" can couple the oxidation of organic substrates to the reduction of ferric iron [4] this is generally only 1200

Microbiology Fundamentals

Figure 3. Effect of pH on long-term survival of mesophilic acidophiles. Symbols: (U) At. ferrooxidans; ({) "F. acidiphilum"; (V) Acidiphilium SJH. Error bars represent standard deviations observed under conditions of limited dissolved oxygen concentrations, for example, in cultures containing extraneous organic materials. Since no reduced sulfur compounds 1201

Microbiology Fundamentals

were present in the spent media or the basal salts suspension, the only other alternative potential substrates would have been organic. These would have arisen from leakage from active cells, from lysis from dead and dying cells, and as atmospheric inputs. While At. ferrooxidans is often considered to be an obligate autotroph, its ability to incorporate methionine [11] indicates that it has at least some capacity to utilize organic carbon compounds. Further evidence that the survival of At. ferrooxidans was at least in part due to heterotrophic metabolism came from comparing its longevity in pure and mixed cultures. More limited survival in mixed cultures may have been due to reduced availability of organic substrates, resulting from competition with the two acidophilic heterotrophs present. In contrast, the greater sensitivity of L. ferrooxidans to dissolved carbon compounds [12] combined with its inability to utilize organic materials would imply that heterotrophic activity would have a net beneficial effect on survival of this iron-oxidizer, as was observed. The reason for the "re-appearance" of colonies of L. ferrooxidans at days 87 and 100 is unclear, but may relate to the break-up of cell aggregates. L. ferrooxidans is known to form flocs consisting of cells enmeshed in extracellular polymers [13] in liquid media. These would have not been enumerated by direct or plate counts, under the protocol used. It is possible that these entrapped cells have greater longevity that free-swimming cells; disintegration of flocs during protracted starvation stress may have resulted in the release of viable planktonic cells, which again tended to die out relatively rapidly. The relative survival of the two heterotrophs, Acidiphilium SJH and "F. acidiphilum", was also affected by competition for the limited organic resources available. Acidiphilium SJH is a particularly adept scavenger in pure cultures, but was out-competed by the ironoxidizing heterotroph "F. acidiphilum" in highly acidic mixed cultures. Indeed, the greater capacity of "F. acidiphilum" to survive starvation stress than the other mesophilic acidophiles tested suggests that these bacteria may be better adapted to situations where substrate depletion is a common phenomenon, and may also be a reason for the dominance of such bacteria in some environments. However, mortality rates of "F. acidiphilum" in moderately acidic and near-neutral solutions were much greater than in media poised at about pH 2. The resulting reduction in competition for available organic materials was considered to be a major reason for the enhanced survival of both At. ferrooxidans and Acidiphilium SJH in mixed cultures at pH 4.10 than in those poised at pH 2.15. However, in solutions of pH 6.10, the combined stresses of starvation and inhospitable pH resulted in more limited survival of all acidophilic bacteria than in more acidic cultures. It is surprising that near-neutral external pH should have lessened the survival of acidophiles, as the internal pH of these bacteria is close to neutral in active cells, and 2-3 pH units above external pH in starved cells [14]. It is known that the maintenance of ∆pH values in resting acidophiles is passive rather than active. The trends identified in this study have highlighted the ability of acidophilic bacteria to survive long-term starvation stress, though surviving populations may be very different in composition to those present at the onset of nutrient depletion. Environmental factors appear to have an important influence on survival of these bacteria. Metabolic versatility appears to be one reason for the superior survival of some species of acidophiles. Evidence for survival strategy based on the formation of viable ultramicrocells was limited to the heterotrophic iron-oxidizer "F. acidiphilum". The results show that At. ferrooxidans, Acidiphilium SJH and "F. acidiphilum" can survive in significant numbers for up to at least 500 days in substrate-depleted situations, though L. ferrooxidans has a far lesser capacity to survive under the conditions imposed in our experiments. A limitation of this work was that only one strain of each acidophile was used in experimental work. It is 1202

Microbiology Fundamentals

conceivable that considerable variation also exists between strains of a particular acidophilic species, and between species in the case of Acidiphilium. ACKNOWLEDGEMENTS This work was supported by funding provided by the U.S. Department of Energy, Office of Science, to the Idaho National Engineering and Environmental Laboratory, operated by Bechtel BWXT Idaho, LLC, under contract DE-AC07099-ID13727. In addition, DBJ was supported during the course of this work by a Faculty Fellowship program sponsored by the U.S. Department of Energy, Office of University and Science Education Programs. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9.

K.B. Hallberg and D.B. Johnson. Adv. Appl. Microbiol. 49 (2001) 37. D.B. Johnson. Water, Air & Soil Pollut. Foc. 3 (2003) 47. D.E. Rawlings. Annu. Rev. Microbiol. 56 (2002) 65. D. B. Johnson. FEMS Microbiol. Ecol. 27 (1998) 307. E. Zychlinsky and A. Matin. J Bacteriol. 153 (1983) 371. A.T. Hobbie, J.R. Daley and S. Jaspar. Appl. Environ. Microbiol. 33 (1977) 1225. D.B. Johnson. J. Microbiol. Meth. 23 (1995) 205. S. Kjelleberg. (ed.) Starvation in bacteria. Plenum Press, New York, 1993. D.A. Siegele, M. Almiron and R. Kolter. p.151. In: S. Kjelleberg (ed.), Starvationin bacteria. Plenum Press, New York, 1993. 10. D.J. Oliver. p.239. In: S. Kjelleberg (ed.), Starvation in bacteria. Plenum Press, NewYork, 1993. 11. D.J. Oliver and J.K. VanSlyke. Arch. Biochem. Biophys. 263 (1988) 369. 12. D.B. Johnson, M.F. Said, M.A. Ghauri and S. McGinness. p. 403. In: J. Salley, R.G.L. McCready and P.L. Wichlacz (eds.), Biohydrometallurgy 1989. Canmet, Ottawa, 1990. 13. R. Hallmann, A. Friedrich, H-P. Koops, A. Pommerening-Roser, K. Rohde, C. Zenneck and W. Sand. Geomicrobiol. J. 10 (1992) 193. 14. P.R. Norris and W.J. Ingledew. p. 115. In R.A. Herbert and R.J. Sharp (eds.), Molecular biology and biotechnology of extremophiles. Blackie, Glasgow, 1992.

1203

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Synthesis of nanophase hydroxyapatite by Serratia sp. N14 P. Yonga, R.L. Sammonsb, P.M. Marquisb, H. Luggb and L.E. Macaskiea* a

Schools of Biosciences, The University of Birmingham, Birmingham B15 2TT, UK b School of Dentistry, The University of Birmingham, Birmingham B15 2TT, UK

Abstract A species of Serratia isolated from heavy metal contaminated soil and previously identified as Citrobacter sp. N14 expresses an atypically high level of acid phosphatase. In the presence of CaCl2 and glycerol-2-phosphate (G2P), biosynthesis of nano-crystalline hydroxyapatite (HA) was achieved by use of immobilised Serratia cells. Calcium ions accumulated as cell-bound calcium phosphates utilizing phosphate released by the enzymatic cleavage of the substrate (G2P). Addition of citrate (to restrict the free Ca2+ concentration) and/or reducing the substrate concentration reduced the size of the HA crystals and promote formation of the nano-phase. Hydroxyapatite crystals were located using Scanning Electron Microscopy (SEM) and identified by X-ray diffraction (XRD). The size of the crystals was calculated to be between 15 and 25 nm, a size more suitable for biomedical applications than that obtained by chemical synthesis.

Keywords: biomineralization, calcium phosphate, hydroxyapatite, Serratia 1.

INTRODUCTION Hydroxyapatite (HA) is a well-established bone substitute biomaterial, which has been successfully used for replacement and augmentation of diseased or traumatised bone [1]. The biocompatibility of this material is related to its close similarity to the composition of the mineral component of natural bone. Hydroxyapatite can also be used to coat metal prostheses, and for orthopaedic and dental applications, using techniques such as plasma spraying. Unfortunately the coating methods currently employed lead to coatings containing a variety of phosphate phases, both amorphous and crystalline, which can be susceptible to dissolution or delamination in clinical use [2,3]. Equivalent problems are also encountered in the fabrication of dense hydroxyapatite components, where difficulties in densification inevitably lead to implants with poor mechanical properties, notably strength and fracture toughness, which cannot be used in load bearing applications. It is well known that as the grain size of materials reduces towards the nanoscale (<200 nm, ideally below 50 nm), properties such as strength are dramatically improved [4]. To achieve this goal requires precursor materials of this nanoscale dimension, which can be subsequently manipulated and fabricated without loss of the fine grain size. Conventional chemical synthesis approaches do not produce particles of the

* To whom correspondence should be addressed (Fax: +44(0)121-414-5925; e-mail: [email protected])

1205

Microbiology Fundamentals

required dimensions and also invariably lead to clusters, or agglomerates of particles, which lead to inherent weakness in the final implant. Microbially mediated biomineralization has a high potential for the production of metallic nano-particles. Klaus [5] found that Pseudomonas stutzeri extracted silver from ores and deposited pure metallic silver particles of up to 200 nm diameter inside the bacteria. Palladium (II) can be reduced enzymatically to Pd(0) at the surface of Desulfovibrio desulfuricans [6,7]; the mean crystal size of biologically-manufactured Pd(0) was only half of that obtained by reduction of soluble Pd(II) under H2 without cells, but it is not clear which biological features contributed to this phenomenon [7,8]. In biomineralization, the components of microbial cell walls and their associated structural polymers may be used as organic "scaffolds" for the patterning of extracellular vesicles and associated inorganic minerals [9]. The compartmentalisation of the biological space on and around cells could be utilised to promote spatial localisation for crystal growth, and nucleation could be mediated by organic polymeric substrates in or on the spatial boundary. The inorganic mineral crystals could be in membrane-bound vesicles, in the mucilaginous layers of cell walls or impregnated in biopolymers in the extra-cellular space [9]. A biomineralization method was developed for the accumulation of metal ions from aqueous solution utilizing cells of a Serratia sp., which was originally classified as Citrobacter [10,11]. It was recently reassigned as Serratia on the basis of biochemical and physiological criteria and 16S rRNA homology [12]. In this bioaccumulation process, metal uptake is mediated by the activity of phosphatase enzyme that is localised both periplasmically or within extracellular polymeric materials [13]. This enzyme continues to function in resting cells to liberate phosphate ions via the hydrolytic cleavage of an organic phosphate molecule (e.g. glycerol 2-phosphate (G2P)) and promotes stoichiometric deposition of cell-bound metal phosphate [13]. Early studies showed that under appropriate conditions (high pH) strontium could be desolubilized using this method [14]. The biomineralization of calcium phosphate should occur in similar manner and the purpose of this study was to evaluate the potential of this approach for the microbially mediated synthesis of calcium phosphate. The investigation focused on the biosynthesis of nanocrystalline hydroxyapatite (HA) with the objective of producing a better HA biomineral for biomedical application. The effect of the challenge solution on metal phosphate deposition, the chemical structure and the size of crystals, and their arrangement on the biomass, was investigated. 2.

EXPERIMENTAL

2.1 Preparation of biomass Biomass was grown under carbon (lactose)-limited continuous culture at 30°C in an airlift fermenter (Fig. 1) as described previously [15]. Minimal medium (2.5 l) [15] was added to the fermenter containing up to 200 polyurethane reticulated foam cubes (TM30 supplied by Recticel, Belgium, washed with distilled water three times) threaded on cotton strings, and the culture was started by the addition of 50 ml inoculum which had been pregrown in similar medium. The culture was allowed to grow batch-wise and switched to continuous mode after 24 h (OD600 ~ 0.6), maintained at steady-state for six days [15], and the foam cubes (heavily loaded with biofilm) were withdrawn and kept in air in a sealed vessel over isotonic saline (8.5 g/l) at 4°C until use. The culture was monitored by measurement of phosphatase activities and cell densities of free cells in the culture outflow [15]. 1206

Microbiology Fundamentals

2.2 Phosphatase activity assay Phosphatase activity was assayed by the release of p-nitrophenol (PNP) from pnitrophenyl phosphate (PNPP) [14,15]. Phosphatase activity (units) is defined as nmol of PNP liberated per min per mg bacterial protein, with OD600 (cell optical density at 600 nm) of the cells converted to bacterial protein using a conversion factor obtained by the assay method of Lowry (0.278 mg protein/ml at OD600 =1; path length = 1 cm) [16]. 2.3 Bioaccumulation of calcium phosphate by Serratia sp. N14 Ten biofilm-loaded foam cubes (~4.5 mg dry cell mass per cube, washed by dipping into isotonic saline) were challenged in 50 ml of 50 mM TAPSO/NaOH buffer pH 9.2 with CaCl2 (1 mM or 5 mM) and sodium glycerol 2-phosphate (G2P: 1 mM or 5 mM) with or without citrate (2 mM). The flask was shaken at room temperature. Additional CaCl2, G2P and citrate were dosed into the flask daily (additional final concentrations of 1 mM, 5 mM and 2 mM as appropriate, after allowing for volume reduction after sampling). The dosing regime is summarised in Table 1. Samples were taken daily before dosing for monitoring of residual Ca(II) and phosphate concentrations. Calcium phosphate-loaded cells harvested by centrifugation after eight days of dosing (or as otherwise specified) were treated and analysed by X-ray powder diffraction and electron microscopy as described. Representative data are shown throughout. 2.4 Assay of phosphate and calcium Samples (up to 5 mM Pi) were diluted 30-fold. Sodium molybdate (0.6 ml; 2.5% (w/v) in 1.67 M H2SO4 was added into 1 ml assay solution and phosphate was visualized by adding 0.4 ml SnCl2. (Fresh solution of SnCl2 was made from 0.25 ml stock solution diluted into 100 ml of 1 M HCl, and the stock was made monthly by dissolving 1.5 g SnCl2 in 2.5 ml concentrated HCl.) The quantity of resultant blue complex was measured at 720 nm (Perkin-Elmer spectrophotometer) and the phosphate concentration estimated by reference to a standard curve, similarly prepared. For assay of Ca(II) 0.1 ml of sample (up to 1 mM Ca(II)) was added to 1.9 ml 20 mM MOPS/NaOH buffer pH 7, and mixed, and Ca(II) was visualized by the addition of 0.1 ml of 0.15% (w/v; aq.) Arsenazo III, with estimation of the pink complex at A631 nm (Perkin-Elmer spectrophotometer) versus a standard solution of CaCl2. 2.5 Localization of cell-bound deposits using scanning electron microscopy Calcium phosphate loaded cells (squeezed from one foam cube) were centrifuged, fixed with 2.5% glutaradehyde fixative (in 0.1 M Na cacodylate/HCl buffer pH 5.2) at 4°C for one hour, dehydrated at room temperature, treated and examined under a scanning electron microscope as described previously [6]. 2.6 Identification of cell bond material and crystal size determination by X-ray powder diffraction (XRD) analysis The Ca-loaded cells (squeezed from 10 foam cubes) were centrifuged after eight days of challenge with the same daily addition, washed with distilled water and acetone, and then dried at room temperature. The samples were laid on a silicon crystal plate and analysed using a high precision X-ray powder diffractometer (School of Physics, The University of Birmingham) for data collection as described previously [7,11]. The powder diffraction patterns were recorded from 5° to 70° (2θ) with a step length of 0.05° (2θ). The crystal size of cell-bound calcium phosphate/HA was calculated for samples obtained 1207

Microbiology Fundamentals

from loaded dried biomass or from commercial reference material HA (Capital 60 hydroxylapatite: Plasma Biotal Ltd., Tideswell, Derbyshire, UK) by computing the diffraction pattern using the Langford programme [17]. 3.

RESULTS AND DISCUSSION When the cells are challenged with calcium and a phosphatase substrate i.e. G2P, there is a biocatalytic (enzymically-mediated) and an inorganic precipitation combined process, in which the component reactions are: Glycerol-2-phosphate Phosphatase> Glycerol + Phosphate (1) ____ Phosphate + Calcium > Calcium phosphate (2) The synthesis of calcium phosphate would occur where the Ca(II) ions interact with the liberated precipitant, phosphate ions. Calcium phosphate can be synthesised chemically, and many authors have studied the factors of affecting this synthesis [1,2,18]. The type of calcium phosphate and the crystal size are largely dependent on the conditions of synthesis, i.e. the setting time, solution pH, ions concentrations, temperature, etc. Indeed, the bioprecipitation of calcium phosphate by Serratia cells would be affected similarly by these physicochemical factors. The form of phosphate ions released by reaction (1) is pH-dependent. Phosphate is a tri-basic ion (K1 = 7.5 x 10-3, K2 = 6.2 x 10-8, K3 = 5.0 x 10-13) [19]. The major forms of which in aqueous solution would be H3PO4 and H2PO4- below pH 4.6, H2PO4- and HPO42- from pH 4.6 to 9.7 and HPO42- and PO43- for pH values above 9.7. In addition, the rate of phosphate release in reaction (1) is dependent on the solution pH since phosphatase activity is pH dependent. Phosphatase activity of the cells was reduced by 70% when the solution pH was changed to 9.0 from 5.5 (Fig. 1a). However, after 24 h the inorganic phosphate concentrations were not significantly different in solutions at pH 7 to 9.2 although the phosphate released was lower in the solution at pH 10.0 (Fig. 1b) attributable to the lower reaction rate. The observation that phosphate release was >95% complete at 24 h (Fig.1b) justified the choice of daily intervals for the dosing regime. The requirement for alkaline pH was shown previously in the case of Sr(II) removal [14] where a compromise was required between the requirement for high pH for precipitation of the alkaline earth metal phosphate and retention of high cellular phosphatase activity. Many authors have studied the solubility of calcium phosphate salts [1,20,21]. The type of calcium phosphate precipitated would be dependent on the pH value of the challenge solution [20-22]. It has been reported that hydroxyapatite (Ca5(PO4)3(OH)) is the most stable calcium phosphate salt at pH>4, but it precipitates directly only at above pH 8 [20,21]. The current investigation was carried out to define the conditions for the biosynthesis of HA for potential medical applications; the quality of the calcium phosphate crystals is much more important than the overall yield. Calcium bioprecipitation by Serratia was tested in a solution of 50 mM TAPSO/NaOH buffer (pH 9.2) and 1 mM CaCl2, with phosphatase substrate G2P at saturating [23] (5 mM) or limiting (1 mM) concentrations using biofilm-immobilized cells. The biomineralization process (controlled by enzyme-mediated inorganic phosphate (Pi) concentration), relies on the cellular microenvironment becoming supersaturated with inorganic phosphate released from G2P by the cells. However, nucleation represents an activation energy barrier to the spontaneous formation of a solid phase from a supersaturated solution. Increasing the degree of super-saturation and lowering the interfacial energy can reduce the activation energy for nucleation [24]. Indeed, for 1 mM free Ca2+ in the bulk solution without complexing reagent, it was calculated that 1208

Microbiology Fundamentals

precipitation should occur for CaHPO4 at 0.1 mM HPO4- (Ksp=1x10-7 for CaHPO4) [25], and for Ca3(PO4)2 at 1.4x10-7 mM PO4- (Ksp= 2.0x10-29 for Ca3(PO4)2) [25]. However, biosorption of calcium was observed only after 24 h even in the solution with 5 mM G2P and a released phosphate concentration of higher than 4 mM (Fig.2a.). In the solution with 1 mM substrate the initiation of precipitation was retarded even further (Fig.2b). When the cells were challenged in the presence of citrate, the concentration of free Ca(II) ions was reduced significantly because citrate is a calcium complexing reagent (K = 4.79x104) [25]. By calculation, there would be only 0.02 mM free ions for Ca(II) at a input concentration of 1mM in the solution with 2 mM citrate, which, in turn, would need a phosphate concentration above 5 mM to form calcium phosphate precipitate. It was shown that the initial precipitation of calcium phosphate from the solution with 2 mM citrate was the slowest (Fig. 2c) compared to that from the other two solutions. However once the precipitation started, the rate of calcium uptake was the same in 1mM as in 5 mM G2P solution with or without citrate, although there was a much lower phosphate or free calcium ion concentration in the latter two solutions (Fig.2b, 2c). Therefore, for the formation of calcium phosphate, nucleation was the controlling stage initially. Once the barrier of nucleation was overcome, continued precipitation of calcium phosphate became feasible. Even in a solution with 1 mM CaCl2, 2 mM citrate and 1 mM G2P (added daily), the precipitation was completed within eight days (not shown).

6

140

Pi concentration (mM)

Relative activity (%)

160 120 100 80 60 40 20

5 4 3 2 1 0

0 4

5

6

7

8

9

10

11

pH

Figure 1a. The effect of pH on phosphatase activity of Serratia cells. Cells were squeezed from the foam cubes and resuspended in 200 mM buffer (pH 5-6: acetate buffer, pH 6-7: MOPS/NaOH buffer, pH 8-9: TAPSO buffer, and pH 10: AMPSO buffer), and phosphatase activities were assayed as described in Experimental. The activity of the cells at pH 7 (expressed as 100%) was 2384 units

7

8

8.6

9.2

10

pH

Figure 1b. The effect of pH on inorganic phosphate (Pi) release by immobilized Serratia cells. Five foam cubes were challenged in 25 ml 50 mM buffer (pH 7-9.2: TAPSO, and pH 10 AMPSO) 5 mM G2P and 2 mM citrate. Samples were taken after 24 hours and centrifuged and the phosphate content of supernatants assayed as described in Experimental

1209

5 0 0 1 2 3 4 5 6 Time (Days)

20 15 10 5 0

25

8 7 6 5 4 3 2 1 0

20 15 10 5 0

0 1 2 3 4 5 6

0 1 2 3 4 5 6

Time (Days)

Time (days)

Ca precipitated (mM)

10

8 7 6 5 4 3 2 1 0

Pi concentration (mM)

15

25

Ca precipitated (mM)

20

Pi concentration (mM)

8 7 6 5 4 3 2 1 0

25

Ca precipitated (mM)

Pi concentration(mM)

Microbiology Fundamentals

a. b. c. Figure 2. Phosphate release and calcium accumulation by immobilized Serratia cells Ten biofilm-foam cubes (~4.5 mg dry cell mass per cube; phosphatase activity was 2834 units at harvest) were challenged in 50 ml of 50 mM TAPSO/NaOH buffer (pH 9.2), and: a) 1 mM CaCl2 and 5 mM G2P with same addition daily; b) 1 mM CaCl2 and 1 mM G2P with same addition daily; c) 1 mM CaCl2, 5 mM G2P and 2 mM citrate with same addition daily. ¡, inorganic phosphate concentration in the challenge solution (mM), , calcium removed by Serratia cells (mM) A scanning electron microscope study (SEM) was performed for samples harvested after eight days using cells challenged with CaCl2 (1 mM), and citrate (0 mM or 2 mM) without (Fig. 3a) or in the presence of 5 mM G2P (Fig. 3b) added daily. Calcium phosphate crystals were visible heavily coating the cell surface but with no apparent differences between the precipitate made from solutions with different concentration of calcium chloride, G2P and citrate. HA crystals were identified by XRD by comparing the pattern with that of the commercial HA and also the XRD standard data file (Fig. 4). The size of HA crystals was calculated by computing the diffraction pattern using the Langford programme [17]; the HA crystals on the cells were only 15-25 nm, substantially below a typical commercial product (Table 1). The largest size (25 nm) of HA crystal was from solution with addition of substrate (G2P) at its saturating concentration (5 mM) in the absence of citrate. The crystal size was considerably reduced when either the concentration of G2P was lowed to 1 mM or citrate (2 mM) was introduced into the solution. When the cells were challenged with 1 mM G2P and 2 mM citrate, Pi release was at the slowest rate and the concentration of free calcium ions was also very low due to citrate complexing. Hence Pi had to remove Ca from the complex to form calcium phosphate precipitate. Indeed, the smallest crystals were from solution with daily addition of 1 mM G2P and 2 mM citrate (15-16 nm). Therefore, the crystals formed at a size governed by the availability of free calcium ions and the rate of phosphate release using the bacterial cell surface as the nucleation template. 4.

CONCLUSION The synthesis of nano-particulate calcium phosphate was achieved by challenging Serratia cells in the solution of 50 mM TAPSO buffer (pH 9.2), 1 mM calcium chloride, 2 mM (or 0 mM) citrate and 5 mM (or 1mM) G2P with the same daily doses. The biosynthesised crystals were located by scanning electron microscopy, and identified by X-ray powder diffraction. The sizes of crystals were 15-25 nm, substantially below a typical commercial product. However, a more detailed study is needed for this biomineralization system to produce the best HA biomineral for commercial exploitation. The mechanical strength and potential biomedical applications of the biosynthesised HA will also need further investigation. 1210

Microbiology Fundamentals

a.

b.

Figure 3. Calcium accumulation by Serratia cells viewed under SEM. Five biofilmfoam cubes were challenged in 25 ml of 50 mM TAPSO/NaOH buffer (pH 9.2) solution with 1 mM CaCl2 only (a.) or with 1 mM CaCl2 and 5 mM G2P with or without 2 mM citrate (b.) with the same daily addition (for eight days). Cells were harvested and treated as described in Experimental.

Figure 4. The X-ray powder diffraction pattern of the cell-bound material and hydroxyapatite. Upper: pattern for the sample from 50 mM TAPSO buffer (pH 9.2) solution with 1 mM calcium chloride, and 5 mM G2P with the same addition daily for eight days. Lower: pattern for the commercial HA crystals. Vertical lines: superimposition of the obtained pattern upon the standard database files for HA [26]

1211

Microbiology Fundamentals

Table.1. The size of crystals for HA precipitated from different solutions with immobilized cells and the commercialised HA. Samples were centrifuged after eight days, treated and the calcium phosphate crystals were identified by XRD as HA as described in Experimental. The crystal size was calculated by computing the diffraction pattern using the Langford programme [17] Samples First day addition: 50 mM TAPSO buffer, pH 9.2 Daily addition to final concentration for seven days HA size (nm)

CaCl2

1 1 mM

2 1 mM

3 1 mM

4 1 mM

5* 5 mM

6* 5 mM

HA** N/A

G2P

5 mM

5 mM

1 mM

1 mM

1 mM

1 mM

N/A

Citrate CaCl2

2 mM 1 mM

0 mM 1 mM

2 mM 1 mM

0 mM 1 mM

2 mM 1 mM

0 mM 1 mM

N/A N/A

G2P

5 mM

5 mM

1 mM

1 mM

1 mM

1 mM

N/A

Citrate

2 mM 21

0 mM 25

2 mM 15

0 mM 20

2 mM 16

0 mM 20

N/A 167

*Note that following the initial dose of 5 mM Ca2+ 1 mM Ca2+ was dosed subsequently in samples 5 and 6. **Commercial HA: Capital 60 hydroxylapatite, Plasma Biotal Ltd., Tideswell, Derbyshire, UK.

ACKNOWLEDGEMENTS This project was supported by the BBSRC (Grant No. 6/E11940). The authors wish to thank Dr J. I. Langford (School of Physics & Space Research, University of Birmingham) for assistance with the XRD technique and Mrs L. Tomkins and Mr P. Whittle for help with electron microscopy. The authors also thank Recticel (Belgium) for the gift of reticulated foam for the growth of biofilm. REFERENCES 1. M. Jarcho, Clin. Orthopaed. Rel. Res., 207(1981)290. 2. S. D. Cook, J. F. Kay, K. A. Thomas and M. Jarcho, Int. J. Oral Maxillofacial Impl., 2(1987)1. 3. A. N. Cranin, M. Mehrali and M. Baraoidan, J. Implantol., 20(1994)235. 4. R. W. Siegel, Mechanical properties of nanophase materials. In Synthesis and Properties of Mechanically Alloyed and Nanocrystalline Materials, Parts 1 & 2, Materials Science Forum, 235 (1997) 851. 5. T. Klaus, R. Joerger, and C. Granqvist. Proc. Nat. Acad. Sci., 96 (1999) 13611. 6. P. Yong, N. A. Rowson, J. P. G. Farr, I. R. Harris and L.E. Macaskie. J. Chem. Tech. Biotechnol. 77 (2002) 593. 7. P. Yong, N. A. Rowson, J. P. G. Farr, I. R. Harris and L. E. Macaskie. Biotechnol. Bioeng., 80(2002)369. 8. P. Yong, N. A. Rowson, J. P. G. Farr, I. R. Harris and L. E. Macaskie. Environ. Technol., In press (2003). 9. S. Mann, J. Chem. Soc Dalton Trans., 21 (1997) 3953. 10. L. E. Macaskie, R. M. Empson, A. K. Cheetham, C. P. Grey, and A. J. Skarnulis, Science, 257 (1992) 782. 11. P. Yong and L.E. Macaskie J. Chem. Tech. Biotechnol., 63(1995)101. 12. P. Pattanapipitpaisal, A. N. Mabbett, J. A. Finlay, A. J. Beswick, M. Paterson-Beedle, A. Essa, J. Wright, M. R. Tolley, U. Badar, N.A. Ahmed, J. L. Hobman, N. L. Brown and L. E. Macaskie, Environ. Technol., 23 (2002) 731. 1212

Microbiology Fundamentals

13. L. E. Macaskie, K. M. Bonthrone, P. Yong, and D. T. Goddard, Microbiology, 146 (2000) 1855. 14. L.E. Macaskie and A.C.R. Dean. Biotechnol. Lett., 7 (1985) 627. 15. J. A. Finlay, V. J. M. Allan, A. Conner, M. E. Callow, G. Basnakova and L. E. Macaskie, Biotechnol. Bioeng., 63(1999)8. 16. B. C. Jeong and L. E. Macaskie, Enz. Mic. Technol., 24 (1999) 218. 17. J. I. Langford. National Institute of Standard and Technology Special Publication 846. Proceedings of the International Conference. Accuracy in Powder Diffraction II, NIST, Gaitherburg, MD, May 26-29, (1992) 110. 18. C. Durucan and P. Brown. J. Mat. Sci.: Materials in Medicine, 11 (2000) 365. 19. G. H. Jeffery, J. Bassett, J. Mendham, R. C. Denney (eds). Vogel’s Textbook of Quantitative Chemical Analysis, Fifth Edition. Longman Scientific Technical, John Wiley And Sons Inc., New York, USA (1989). 20. P. Boudeville, S. Serraj, J.M. Leloup, J. Margerit, B. Pauvert, and A. Terol. J. Mat. Sci.: Materials in Medicine, 10 (1999) 99. 21. S. Matsuya, S. Takagi, and L.C. Chow. J. Mat. Sci.: Materials in Medicine, 11 (2000) 305. 22. E. Fernandez, F.J. Gil, M.P. Ginebra, F.C.M. Driessens, J.A. Planell and S.M. Best. J. Mat. Sci.: Materials in Medicine, 10 (1999) 223. 23. P. Yong and L.E. Macaskie. Biotechnol. Bioeng., 55(1997)821. 24. S. Mann. Nature, 332 (1988) 119. 25. J.A. Dean (ed). Lange's Handbook of Chemistry, McGraw-Hill, Inc. New York, 1992. 26. Anon. Powder Diffrection File. Card No. 09-0432, JCPDS, Swarthmore, PA, USA (1992).

1213

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

The effect of maintenance on the ferrous-iron oxidation kinetics of Leptospirillum ferrooxidans C.J.N. Dempers1, A.W. Breed2 and G.S. Hansford Gold Fields Minerals Bioprocessing Laboratory, Department of Chemical Engineering, University of Cape Town, Rondebosch, 7701, South Africa Abstract The results of batch experiments carried out using a predominantly Leptospirillum ferrooxidans culture indicated that the maintenance requirement of energy sufficient cultures is a combination of the variable and constant maintenance energy requirements. For this reason, modelling the bioenergetics of energy sufficient cultures assuming a constant maintenance energy requirement results in a significant overestimation of the maximum bacterial yield. However, the variable maintenance equation proposed by Pirt (Pirt, 1982) may be used to determine the maximum bacterial yields and the maintenance coefficients from both substrate lean and substrate rich experiments. This is because this model is able to account for variations in the growth rate as a result of changes in the maintenance requirement due to the microorganisms being limited by more than one factor.

Keywords: batch processing, bioreactors, Chemoautotrophs, kinetic parameters, maintenance requirements, modelling 1.

INTRODUCTION Recent work has provided strong evidence that the bioleaching of sulfide minerals occurs via a multiple sub-process mechanism [1-5]. According to Sand et al. [2], the disulfides pyrite (FeS2), molybdenite (MoS2), and tungstenite (WS2) are chemically oxidised by the ferric-iron present in the bioleaching medium via the main intermediate thiosulfate:

FeS2 + 6 Fe3+ + 3 H 2O → S2O32 − + 7 Fe2 + + 6 H +

(1)

The thiosulfate is subsequently, degraded in a cyclic process to sulfate, with elemental sulfur being a side product [2]: S2O32 − + 8 Fe3+ + 5 H 2O → 2SO 24 − + 8 Fe2 + + 10 H +

(2)

The above thus explains why ferrous-iron oxidizing bacteria alone, e.g. Leptospirillum ferrooxidans, are able to oxidise these metal sulfides.3 1

MRT Africa, Park Place East, 104 North Rand Road, Hughes, Boksburg, 1460, South Africa Department of Chemical Engineering, University of Sydney, NSW 2006, Australia 3 The genus Leptospirillum has recently been shown to have two species, viz. Leptospirillum ferrooxidans and Leptospirillum ferriphilum [6]. This differentiation is not however considered in this work. 2

1215

Microbiology Fundamentals

In contrast to the above, metal sulfides such as galena (PbS), sphalerite (ZnS), chalcopyrite (CuFeS2), hauerite (MnS2), orpiment (As2S3), and realgar (As4S4) are oxidised by both the ferric-iron and protons present in the bioleaching solution; the main intermediates of these reactions are polysulfides and elemental sulfur [2]: MS + Fe3+ + H + → M 2 + + 0.5 H 2Sn + Fe2 + (n > 2)

(3)

0.5 H 2Sn + Fe3+ → 0.125 S8 + Fe2 + + H +

(4)

0.125 S8 + 1.5 O 2 + H 2O + → SO 24 − + 2 H +

(5)

These metal sulfides are thus degradable by bacteria that are able to oxidize sulfur compounds, e.g. Acidithiobacillus thiooxidans.4 If ferrous-iron oxidising bacteria are present within the system, then the ferrous-iron produced by Reactions 1 to 4 is subsequently oxidised to the ferric form by these bacteria: 4 Fe 2+ + O 2 + 4 H + → 4 Fe 3+ + 2 H 2 O

(6)

A multiple sub-process mechanism, such as the one described above, thus suggests that the overall process can be expressed as a number of interconnected chemical and bacterial sub-processes, the kinetics of which may be studied separately, and the results obtained used to predict the performance of bioleach reactors for a variety of different minerals, micro-organisms and operating conditions. To date a number of kinetic models for bacterial ferrous-iron oxidation have been proposed [8]. These models can be broadly classified as either empirical or MichaelisMenten/Monod based. In the proposed models, the bacterial specific growth rate, µ, is usually related to the bacterial specific ferrous-iron utilisation rate, q Fe2 + , via the , and a constant maintenance coefficient maximum bacterial yield on ferrous-iron, YFemax 2+ X on ferrous-iron, m Fe2 + ; viz. using the Pirt Equation [9]: q Fe 2 + =

µ + m Fe2 + YFemax 2+ X

(7)

The bacterial specific ferrous-iron utilisation rate, q Fe2 + , is defined as follows: q Fe2 + =

− rFe2 + cX

(8)

where: rFe2 + is the ferrous-iron production rate, and c X is the bacterial concentration expressed as mol C.l-1.

From Equation 7 it is thus apparent that the Pirt Equation [9] is based on the assumption that the biomass specific maintenance requirement is constant and independent of the growth rate; the maintenance energy requirement is the minimum amount of substrate, per unit of biomass, which is required to maintain the vital functions of the microorganisms. Although the Pirt Equation [9] has been widely used to model the ferrous-iron oxidation kinetics of the bacteria encountered in bioleaching, a dependence of the experimentally determined yield and maintenance coefficients on the growth conditions 4

Acidithiobacillus thiooxidans was previously named Thiobacillus thiooxidans [7].

1216

Microbiology Fundamentals

has been reported [5]. Boon [5] reported that it was necessary to fit the Pirt Equation [9] to the initial and late batch data separately, i.e. the calculated values of YFemax and m Fe2 + 2+ X depended on whether the growth was energy sufficient or energy limited. Similar dependencies on the growth conditions have been observed during experiments performed using other microorganisms [10-13]. Neijssel and Tempest [14] suggested that these deviations could be attributed to the growth of the microorganisms being limited by factors other than the energy source. This led these workers to propose that the maintenance requirement was not constant, but varied with changes in the bacterial growth rate. Pirt [15] modified the relationship developed by Neijssel and Tempest [14] by assuming that the energy required for maintenance included a term that decreased with an increase in the specific growth rate. Although the variable maintenance approach has been successfully used to describe the kinetics of non-bioleaching microorganisms, it has yet to be used in the modelling of the ferrous-iron oxidation kinetics of the bacteria used in bioleaching operations. The primary objective of the work presented below was to determine whether or not the variable maintenance equation proposed by Pirt [15] could be used to describe the batch growth data of a predominantly L. ferrooxidans culture during both energy sufficient and energy-limited growth. A further objective of the work was to determine whether or not the kinetic parameters determined in this way were similar to those determined previously during continuous ferrous-iron oxidation experiments performed using the same bacterial culture and similar growth conditions.5 2.

THEORETICAL ASPECTS If bacteria are grown in continuous culture and under conditions in which the carbon source is limited to carbon dioxide, then the carbon dioxide production rate, rCO 2 , can be

used to estimate the biomass production rate growth rate, rX , [5]: rX = − rCO 2

(9)

However, if the bacteria are grown in batch culture and under conditions in which the carbon source is limited to carbon dioxide, then the bacterial concentration at time, t, can be estimated from the carbon dioxide production rate, rCO 2 , and the initial bacterial concentration, c X 0 : t =t

c X = c X0 +

∫−r

CO 2

dt

(10)

t =0

Furthermore if the stoichiometric formula of bacteria is assumed to be CH1.8O0.5N0.2 [18, 19] and energy for bacterial growth and maintenance is obtained from the oxidation of ferrous-iron, performing mass and charge balances and solving in terms of the carbon dioxide and oxygen production rates, rCO 2 and rO 2 , respectively, yields the degree-ofreduction balance: − rFe2 + = − 4rO 2 − 4.2 rCO 2

(11)

5

The continuous experiments are reported elsewhere [16, 17] whereas the batch experiments were performed during the course of the current study.

1217

Microbiology Fundamentals

If the maintenance energy requirement is assumed to be constant and independent of the growth rate and the theoretical maximum growth yield, YFemax , occurs when the 2+ X maintenance energy, m Fe2 + , is zero, then the relationship between the amount of substrate consumed by the biomass for bacterial growth and maintenance can be described using Equation 7, i.e. the Pirt Equation [9]. However, if the maintenance requirement is assumed to vary with changes in the bacterial growth rate and if the maintenance energy term is assumed to include a portion that decreases with an increase in the specific rate of growth, then the relationship between the bacterial specific substrate utilisation rate, the maximum substrate specific yield and the and the growth rate must be described using [15]: q Fe 2 + =

µ Y

max

v + m Fe 2 + + m Fe 2 + (1 − k µ )

(12)

Fe 2 + X

In Equation 12, m Fe2 + is the constant maintenance coefficient on ferrous iron, whereas v m Fe 2 + and k are growth dependent maintenance coefficients. Pirt [15] also postulated that

the growth rate dependent maintenance energy requirement decreases to zero as the specific growth rate, µ, approaches the maximum bacterial specific growth rate, µ max . This in turn implies that: k=

1 µ

(13)

max

Combining Equations 12 and 13 and rearranging yields: q Fe 2 +

v ⎛ 1 m Fe 2+ ⎜ = − max max ⎜ Y 2+ µ ⎝ Fe X

⎞ ⎟ µ + m 2+ + m v 2+ Fe Fe ⎟ ⎠

(14)

A graphical representation of Equation 14 is shown in Figure 1; the two lines, (a) and (b), represent experimental data obtained under different energetic conditions. From Figure 1 and the definition of the constant maintenance requirement, it is apparent that, under conditions in which the growth rate is limited by the concentration of the energy source, i.e. during energy limited growth, the maintenance energy requirement of the microorganisms is constant and lower than the maintenance energy requirement when the growth is limited by another factor, e.g. the concentration of a specific trace metal. Under these conditions the constant maintenance coefficient can thus be obtained from the intercept of the solid line and the y-axis, i.e. from line (a). Furthermore, because the variable maintenance requirement decreases to zero under energy-limited conditions, the maximum growth yield can also be determined from the slope of line (a). In contrast to the above, the maintenance energy requirement of energy sufficient cultures is a combination of the variable and constant maintenance energy requirements (see line (b) in Figure 1). The variable maintenance energy requirement can therefore be determined from the difference between the values of the y-intercepts determined during energy sufficient and energy limited growth experiments; the value of the variable maintenance coefficient will thus depend on the limiting factor. These factors include the concentrations of phosphate, sulfate and ammonium ions, the concentration of heavy metals and the bacterial growth rate.

1218

Microbiology Fundamentals

b

2+

-1

-1

qFe2+ (mol Fe .(mol C) .h )

a qFe2+max

mFe2+ + mFe2+v 1/YFe2+max mFe2+ 0 0

max

µ

-1

µ (h )

Figure 1. Relationship between the bacterial specific growth rate, µ, and the bacterial specific ferrous-iron utilisation rate, q Fe2 + , assuming a) constant maintenance and b) variable maintenance In addition to the above, it can be seen that the maximum bacterial specific ferrousiron utilisation rate, q max , and the maximum bacterial specific growth rate are defined by Fe 2 + the intersection of the solid and the broken lines. Equation 14 can also be written in terms of the bacterial specific oxygen utilisation rate, q O 2 , hence the maximum bacterial yield on oxygen, YOmax , the constant maintenance 2X coefficient on oxygen, m O 2 and the growth dependent maintenance coefficient on oxygen, m Ov 2 , may be calculated in the same way as the ferrous-iron based parameters. The

validity of the yield and maintenance coefficients calculated determined in this manner can in turn be can be checked using the degree of reduction balance, Equation 10; i.e. using: YOmax = 2X m O2 =

4 YFemax 2+ X 1 − 4.2 YFemax 2+ X

(15)

m Fe2 +

(16) 4 It is thus also apparent from Figure 1 and the preceding discussion that the values of the kinetic parameters determined from experimental data can be highly dependent on the conditions used.

3.

MATERIALS AND METHODS The batch ferrous-iron oxidation experiments performed during the course of this study were carried out at pH 1.70 and 30, 35 and 40°C and at 40°C and pH 1.10, 1.30, 1.50 and 1.70, using a predominantly L. ferrooxidans culture. The bacterial culture used was initially obtained from a vat-type two-stage (2×20 l) continuous bioleaching miniplant treating an arsenopyrite/pyrite concentrate from Fairview Gold Mine in Barberton, South Africa [20]. The inoculum for each of the batch experiments was obtained from steady state continuous cultures of microorganisms from the bioleaching mini-plant grown 1219

Microbiology Fundamentals

on ferrous-iron medium at a dilution rate of 0.04 h-1 and the temperature and pH to be used in the batch experiment. The batch experiments themselves were carried out in 2l air-sparged, agitated bioreactors. The bioreactors had a H/D of 1.32 and a working volume of 1l. Circulating water from constant temperature baths through the bioreactor jackets controlled the temperature in the bioreactors. The pH of the solution in the bioreactors was monitored continuously and maintained at the required pH by the addition of concentrated sulphuric acid (98%). The redox potential of the bioleaching solution in the bioreactors was measured continuously using Metrohm redox electrodes (Pt-Ag/AgCl) and logged by computer. The ferrous iron media used consisted of 12 g.l-1 FeSO4.7H2O, 1.11 g.l-1 K2SO4, 0.53 g.l-1 (NH4)2HPO4, 1.83 g.l-1 (NH4)2SO4 and 10 ml.l-1 trace element solution [21] adjusted to between pH 0.95 and pH 1.30 using H2SO4conc. Air, at a flow rate of 1 l.min-1, was supplied to the bioreactors using Brooks mass flow controllers. The off-gas from the bioreactors was dried prior to passing through CO2 and O2 gas analysers. This enabled the oxygen utilisation rate, -rO2, carbon dioxide utilisation rate, -rCO2, and biomass concentration, cX, to be determined [4]. The experimental equipment used is described in greater detail elsewhere [3]. The total iron concentration in solution was determined by both atomic adsorption spectroscopy (AAS) and by titration with potassium dichromate [22]. This enabled the ferric/ferrous-iron ratio and the ferrous- and ferric-iron concentrations to be determined using a calibration curve for the specific electrode and the Nernst equation [5]. Changes in the ferrous- and ferric-iron and total iron concentrations with time allowed the ferrous-iron production rate, rFe2 + , to be calculated.

4.

RESULTS AND DISCUSSION In the first instance an attempt was made to determine the kinetic parameters, YFemax2 + X ,

and m Fe 2 + using the Pirt Equation [9] i.e. by plotting µ vs. q Fe 2 + . A typical plot of µ vs. q Fe 2 + for the experiment performed at 40°C and pH 1.10 is shown in Figure 2. From Figure 2 it is evident that there are three distinct linear regions, hence it was not possible to determine the values of YFemax2 + X and m Fe 2 + in this way. However, the similarity between Figs. 1 and 2 suggests that the micro-organisms used were in fact limited by different factors during the period over which the experiment was performed and should therefore be modelled using the variable maintenance equation, viz. Equation 14. The values of the maximum bacterial yield on ferrous-iron, YFemax2 + X , and the constant maintenance coefficient on ferrous-iron, m Fe 2 + , were therefore calculated by linear regression of the data obtained during the latter stages of the batch experiments (region cd); it was assumed that there was no variable maintenance requirement in this region. v This allowed the variable maintenance coefficient on ferrous-iron, m Fe 2 + , to be estimated from a regression line drawn through the data obtained during the initial stages of the batch experiment (region ab). Finally, the maximum bacterial specific growth rate, µ max , , were determined and the maximum bacterial specific ferrous-iron utilisation rate, q max Fe 2 + 1220

Microbiology Fundamentals

from the intercept of the regression lines drawn through the data obtained during the initial and latter stages of the batch experiments, as demonstrated in Figure 1. The average ferrous-iron and oxygen-based parameters calculated in this manner are listed in Table 1 and Table 2. The validity of the yield and maintenance coefficients listed in Table 1 and Table 2 was checked by comparing the experimental (calculated) values with the values predicted by the "degree of reduction" balance i.e. Equations 15 and 16. These comparisons are displayed in Figure 3, from which it is apparent that the correlation is good (average correlation coefficient R2 = 0.9596).

a

-1

12

2+

-1

qFe2+ (mol Fe .(mol C) .h )

16

8

b c

4

0 0.00

d 0.02

0.04

0.06

0.08

-1

µ (h )

Figure 2. Variation in the experimentally observed bacterial specific ferrous-iron utilisation rate, q Fe2 + , with changes in the bacterial specific growth rate, µ Table 1. Average ferrous-iron based bioenergetic parameters Experimental conditions 30°C

pH 1.70 35°C

40°C

14.65

22.31

(mol Fe2+.(mol C)-1.h-1)

0.0056

m Fe2 + (mol C.(mol Fe2+)-1.h-1) v 2+ -1 -1 m Fe 2 + (mol Fe .(mol C) .h )

q

max Fe 2 +

-1

-1

(mol Fe .(mol C) .h )

max Fe 2 + X

Y

2+

µ max (h-1)

pH 1.50

40°C pH 1.30

pH 1.10

10.65

14.32

18.30

16.36

0.0055

0.0036

0.0058

0.0054

0.0058

0.540

0.855

0.002

0.95

2.26

1.00

7.85

6.50

11.32

6.06

10.08

5.95

0.079

0.119

0.038

0.077

0.087

0.089

1221

Microbiology Fundamentals

Table 2. Average oxygen based bioenergetic parameters Experimental conditions 30°C

pH 1.70 35°C

40°C

q Omax (mol Fe2+.(mol C)-1.h-1) 2

3.37

5.63

YOmax (mol Fe2+.(mol C)-1.h-1) 2X

0.0228

m O 2 (mol C.(mol Fe ) .h )

pH 1.50

40°C pH 1.30

pH 1.10

2.61

3.50

4.49

4.00

0.0227

0.0158

0.0236

0.0221

0.0239

0.131

0.203

0.0917

0.2312

0.5638

0.2423

m Ov 2 (mol Fe2+.(mol C)-1.h-1)

2.32

1.59

2.74

1.52

2.51

1.49

µ max (h-1)

0.074

0.123

0.040

0.077

0.087

0.090

2+ -1

-1

0.032

0.60 Figure 3(b)

0.024

-1

0.016

2+

4YFe2+/(1-4.2YFe2+)

-1

(mFe2+)/4 (mol Fe .(mol C) .h )

Figure 3(a)

0.008

0.000 0.000

0.008

0.016

0.024

0.45

0.30

0.15

0.00 0.00

0.032

0.15

-1

YO (mol C.(mol O2) )

0.30

0.45 -1

0.60

-1

mO (mol O2.(mol C) .h ) 2

2

Figure 3(c)

v

2+

-1

-1

(mFe2+ )/4 (mol Fe .(mol C) .h )

3.2

2.4

1.6

0.8

0.0 0.0

0.8 v

1.6

2.4 -1

3.2 -1

mO (mol O2.(mol C) .h ) 2

Figure 3. Comparison of the predicted and experimental relationships for (a) the maximum bacterial yield, (b) the constant maintenance coefficient and (c) the variable maintenance coefficient

1222

Microbiology Fundamentals

In addition to the above, it is also apparent from the results listed in Table 1 and Table 2 that the maximum bacterial yields on ferrous-iron and oxygen and their respective constant and variable maintenance coefficients, did not vary significantly with either temperature or pH. For this reason, average maximum bacterial yield and constant maintenance coefficients on ferrous-iron and oxygen were calculated by linear regression using the data from the latter stages of all the batch experiments performed during this investigation. These values are listed in Table 3, together with previously reported values of these parameters. From the results presented in Table 3 it is apparent that the average maximum bacterial yields and constant maintenance coefficients determined during the course of this investigation are similar to previously published values.

Table 3. Average values of the maximum bacterial yield and constant maintenance coefficient on ferrous-iron and the maximum bacterial yield and maintenance coefficient on oxygen together with the values reported previously Breed [3]

Van Scherpenzeel et al. [23]

Current work

YFemax (mol C.(mol Fe2+)-1) 2+ X

0.0075

0.011

0.0060

m Fe 2 + (mol Fe2+.(mol C)-1.h-1)

1.196

0.444

1.261

YOmax (mol C.(mol O2)-1) 2X

0.0279

0.046

0.0247

m O 2 (mol O2.(mol C)-1.h-1)

0.2376

0.0425

0.3096

The maximum bacterial specific ferrous-iron utilisation rates, q max , and the maximum Fe 2+

bacterial specific growth rates, µ max , calculated using the values listed in Table 3 and the variable maintenance equation, Equation 14, are listed in Table 4, together with previously published values of these parameters.

Table 4. Comparison of the maximum bacterial specific growth rates and ferrousiron utilisation rates calculated using the variable maintenance equation and those published by Breed [3] Experimental conditions 30°C; pH 1.70 35°C; pH 1.70 40°C; pH 1.70 40°C; pH 1.50 40°C; pH 1.30 40°C; pH 1.10

Breed [3]

µ

max

(h-1) 0.0397 0.0638 0.0862 0.1238 0.1077 0.1027

q

Current work max Fe 2 +

(mol Fe2+.(mol C)-1.h-1) 8.65 11.01 13.62 19.02 15.57 15.26

µ

max

(h-1) 0.074 0.123 0.040 0.077 0.087 0.090

q max Fe 2 +

(mol O2.(mol C)-1.h-1) 14.65 22.31 10.65 14.32 18.30 16.36

From these results it is apparent that the values of q max and µ max calculated using the Fe variable maintenance model are also similar to previously published values of these parameters. The results presented thus suggest that during the initial stages of a batch experiment (region ab in Figure 2) the microorganisms are not substrate-(energy-)limited, whereas in the latter stages of the experiment (region cd in Figure 2) they are substrate (energy) limited. The points, (b) and (c) in Figure 2 represent points at which a change in the limiting substrate occurs. In fact, within the transition region (region bc in Figure 2) the limiting substrate is continually changing, hence a line with a negative slope is 2+

1223

Microbiology Fundamentals

obtained. In other words, during batch experiments the growth of the microorganisms are limited by more than one factor separately, hence the data cannot be modelled using the constant maintenance equation. However, it can be modelled using the variable maintenance equation.

5.

CONCLUSIONS The results of the batch experiments performed during the course of this study indicate that the maintenance requirement of microorganisms is actually a combination of the variable and constant maintenance energy requirements. The variable maintenance requirement depends on the limiting substrate, whereas the constant maintenance requirement is dependent on the energy source. Where the bacterial growth rate is limited solely by the energy source, the variable maintenance requirement decreases to zero. Modelling the bioenergetics of energy sufficient cultures assuming a constant maintenance energy requirement thus results in a significant overestimation of the maximum bacterial yield. However, the variable maintenance equation proposed by Pirt [15] may be used to determine the maximum bacterial yields and the maintenance coefficients from both substrate lean and substrate rich experiments. This is because this model is able to account for variations in the growth rate as a result of changes in the maintenance requirement due to the micro-organisms being limited by more than one factor, e.g. ferrous-iron, ammonia, phosphate, sulfate, etc. It is thus suggested that the variable maintenance equation be used instead of the constant maintenance equation when quantifying the bioenergetics of microorganisms, including those encountered during the bioleaching, unless the energy source is known to be the limiting substrate. The above is especially important for the case of bioleaching using ferrous-iron oxidising microorganisms at low redox potentials (high ferrous-iron concentrations), e.g. during mesophilic chalcopyrite bioleaching and during thermophilic bioleaching. REFERENCES 1. G.S. Hansford and T. Vargas, Hydrometallurgy, 59 (2001) 135. 2. W. Sand, T. Gehrke, P.-G. Jozsa and A. Schippers, Hydrometallurgy, 59 (2001) 159. 3. A.W. Breed, Studies on the mechanism and kinetics of bioleaching with special reference to the bioleaching of refractory gold-bearing arsenopyrite/pyrite concentrates, PhD Thesis, University of Cape Town, Cape Town, South Africa, 2000. 4. M. Boon, G.S. Hansford and J.J. Heijnen, Biohydrometallurgical Processing I, T. Vargas, C.A. Jerez, J.V. Wiertz and H. Toledo (eds.), University of Chile, Santiago, Chile, 1995. 5. M. Boon, Theoretical and experimental methods in the modelling of biooxidation Kinetics of sulphide minerals, PhD Thesis, Technische Universiteit Delft, The Netherlands, 1996. 6. N.J. Coram and D.E. Rawlings, Appl. Environ. Microbiology, 68 (2002) 838. 7. D.P. Kelly and A.P. Wood, Int. Jour. Systematic and Evolutionary Microbiol., 50 (2000) 511. 8. M. Nemati, S.T.L Harrison, C. Webb and G.S. Hansford, Biochem. Eng. J., 1 (1998) 171. 9. S.J. Pirt, Proceedings of the Royal Society of London. Series B: Biological Sciences, 163 (1965) 224.

1224

Microbiology Fundamentals

10. J.G. Kuenen, Growth yields and “maintenance energy requirement” in Thiobacillus species under energy limitation, Arch. Microbiol., 122 (1979) 183. 11. A.J. Downs and C.W. Jones, Arch. Microbiol., 105 (1975) 159. 12. O.M. Neijssel and D.W. Tempest, Arch. Microbiol., 106 (1975) 251. 13. A.H. Stouthamer and C.W. Bettenhausen, Arch. Microbiol., 102 (1975) 187. 14. O.M. Neijssel and D.W. Tempest, Arch. Microbiol., 107 (1976) 215. 15. S.J. Pirt, Arch. Microbiol., 133 (1982) 300. 16. A.W. Breed, C.J.N. Dempers, G.E. Searby, M.N. Gardner, D.E. Rawlings and G.S. Hansford, Biotechnol. Bioeng., 65 (1999) 44. 17. A.W. Breed and G.S. Hansford, Biochem. Eng. J., 3 (1999) 193. 18. C.A. Jones and D.P. Kelly, J. Chem. Technol. Biotechnol., 33B (1983) 241. 19. J.A. Roels, Energetics and kinetics in biotechnology, Elsevier Biomedical Press, Amsterdam, 1983. 20. A.W. Breed, S.T.L. Harrison and G.S. Hansford, IBS-BIOMINE 97, Australian Mineral Foundation, Glenside, Australia, 1997. 21. W. Vishniac and M. Santer, Bacteriol. Revs., 21 (1957) 195. 22. G.H. Jeffrey, J. Bassett, J. Mendham and R.C. Denney (eds.), Vogel’s Textbook of Quantitative Chemical Analysis, 5th edition, Longman Scientific & Technical, New York, 1989. 23. D.A. van Scherpenzeel, M. Boon, C. Ras, G.S. Hansford and J.J. Heijnen, Biotechnol. Prog., 14 (1998) 425.

1225

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

The kinetics of thermophilic ferrous-iron oxidation G.E. Searby and G.S. Hansford Gold Fields Mineral Bioprocessing Laboratory, Department of Chemical Engineering, University of Cape Town, Private Bag, Rondebosch, 7701, South Africa Fax: +27-21-689 7579. E-mail: [email protected] Abstract The kinetics of ferrous iron oxidation by a thermophilic archaeal culture was investigated in continuous culture, at temperatures ranging from 65 to 75ºC, and a pH of 1.5. The reaction kinetics were followed by on-line monitoring of the solution redox potential and of the concentrations of oxygen and carbon dioxide in the off gas stream. The specific rate of ferrous iron utilisation was calculated from the rates of oxygen and carbon dioxide utilisation via a degree of reduction balance, and examined as a function of the ferric/ferrous iron ratio, determined from the redox potential and checked by measuring dissolved iron concentrations. The results were modelled using a kinetic model developed for mesophilic oxidation based on Michaelis-Menten enzyme kinetics and proportional to the ferric/ferrous iron ratio. The effect of temperature was accounted for by the incorporation of an Arrhenius term. The thermophiles achieved specific iron utilisation rates similar to those reported for L. ferrooxidans, but was active at significantly lower redox potentials, and less active at high redox, more comparable to At. ferrooxidans.

Keywords: thermophiles, bioleaching, kinetics, iron oxidation, continuous culture 1.

INTRODUCTION The recalcitrance of chalcopyrite to bioleaching by mesophilic microorganisms has led to interest in high temperature leaching. Thermophilic archaea have been shown to be capable of effective leaching. (Le Roux, 1988, and Dew et al, 1999). Like mesophilic systems, thermophilic mixed cultures often contain species that are predominantly iron oxidisers and others that are predominantly sulfur oxidisers. This indicates the possibility that the leaching mechanism is similar to that generally accepted for conventional bioleaching microbes, i.e. a chemical ferric leach followed by microbially mediated ferrous iron oxidation reforming the primary ferric leach reagent, and microbial oxidation of sulphur compounds released. An understanding of the ferrous iron oxidation kinetics of these archaea at their operating temperature is therefore desirable so that their bioleaching capabilities may be fully exploited. 1227

Microbiology Fundamentals

Mesophilic ferrous iron oxidation (mostly focussed on Acidithiobacillus ferrooxidans) has been studied extensively. A number of kinetic models have been proposed, starting with empirical systems such as the logistic equation and simple Monod models, followed by more developed models. These were often based on Michaelis-Menten enzyme kinetics and adapted to show substrate and/or product inhibition and effects of various parameters such as dissolved oxygen concentration, pH, and temperature. Ingledew (1982) proposed a chemiosmotic theory for ferrous iron oxidation in At. ferrooxidans, involving the generation of a transmembrane potential and a proton-motive force by splitting the two half reactions in the overall chemical reaction of ferrous iron oxidation across the cell membrane. This model highlighted the influence of the solution redox potential on the microbial growth and oxidation kinetics and has led to the generation of models based on redox potential (Huberts, 1994; Boon, 1995; and Meruane et al, 2002). It is proposed that the oxidation of ferrous iron by thermophiles may be controlled by a similar mechanism, and hence that the kinetics can be described using a similar model.

2.

MATERIALS AND METHODS Experiments were performed in three 1.7L, air-sparged glass vessels, each with an H/D ratio of 1.32, using a working volume of 1L. Temperature in the reactors was maintained at 65, 70 and 75ºC by a heated water jacket. pH was maintained at 1.5 by manipulation of the feed pH and by addition of concentrated sulphuric acid. The reactors were kept airtight, allowing the measurement of the change in oxygen and carbon concentration in the entering and exiting gas streams. Off-gas analysis and the measurement of the change in the solution potential have been shown to be an effective system for the investigation of ferrous-iron oxidation (Boon, 1996). Compressed air was passed through a gas-chiller and a series of filters to obtain bonedry medical quality air and fed into the reactor at a constant rate via a Brooks 5850S mass flow control valve, and sparged below the impeller. Agitation and gas mixing was achieved by a Lightnin 315 impeller running at 400 rpm. Air leaving the reactor passed through a reflux condenser, which served the dual purpose of drying the gas before it enters the gas analysis system, and avoiding evaporative water loss in the reactor. The offgas then passed through a cloth filter and a Hartmann and Braun CGEK sample gas conditioner before entering the analysers. Oxygen concentration was measured by a Hartmann and Braun Magnos 6G paramagnetic oxygen analyser and carbon dioxide using a Hartmann and Braun Uras 4 infrared photometer. The culture used was a mixed culture of extreme Sulfolobus-like thermophilic archaea grown at pH 1.5 and at a temperature of 70ºC. The culture was taken initially from a semicontinuous system grown on a chalcopyrite concentrate and has been maintained in continuous culture on ferrous iron for more than 2 years. The culture was grown in a basal salt medium containing 0.5 g/L MgSO4.7H2O, 0.4 g/L (NH4)2SO4 0.02 g/L K2HPO4.3H2O and 0.1 g/L KCl (Clark and Norris, 1996). The energy source was 12g/L FeSO4. Reduced sulphur was added as 0.225g/L K2S4O6 as the thermophiles are unable to assimilate sulphate. Experiments were performed in continuous culture at dilution rates ranging from 0.015 to 0.9 hr-1. Continuous flow was obtained by feeding the fresh nutrient medium via a variable speed peristaltic pump, and by removing the resultant reactor liquor at a fixed liquid level by a fixed speed peristaltic pump, maintaining a constant volume. The reactors 1228

Microbiology Fundamentals

were maintained at each dilution rate for 5 residence times before steady state was assumed. Steady state was verified by the generation of steady data for a further residence time. Wall growth was eliminated by daily scrubbing of all reactor surfaces. The reaction kinetics was followed by off-gas analysis. In an autotrophic system, the only carbon source is atmospheric carbon dioxide, and thus the carbon dioxide utilisation rate can be used to follow the growth kinetics. rX = − rCO 2

(1)

This provides a non-invasive on-line measurement of the cell concentration, in terms of moles of carbon fixed. cx =

− rCO 2

D Autotrophic microbes oxidise iron as their energy source,

4Fe2 + + 4H + + O 2

⎯⎯ ⎯⎯→ 4Fe3+ + 2H 2O microbes

(2)

(3)

and utilise the energy produced for cell synthesis (typical cell stoichiometry CH1.8O0.5N0.2) CO2, H2O, NH +4

⎯ ⎯→ CH1.8O0.5 N 0.2

(4)

Simultaneous solution of the species and charge balances for the species in Reactions 3 and 4, or the degree of reduction balance over the same set of species (Roels, 1983) (5) ∑ γ i ri = 0 i

yields the rate of microbial ferrous iron oxidation in terms of the measured off-gas parameters. − rFe2 + = −4.2rCO 2 − 4rO 2

(6)

The solution redox potential was measured using a Pt Ag/AgCl2 redox electrode. This was related to the ferric/ferrous-iron ratio in the reactor via the Nernst equation, approximating activities as concentrations. E

=

E '0

RT ⎛ [Fe3+ ] ⎞ ⎟ + ln⎜ zF ⎜⎝ [Fe2 + ] ⎟⎠

(7)

The ferrous-iron and total iron concentrations were determined by titration with K2Cr2O7 (Jeffrey et al, 1989). Ferric iron concentrations can then be calculated by the difference between the ferrous and the total iron concentrations. This provided a check for the ferric/ferrous-iron ratio as determined by redox measurements. The specific iron utilisation rate was modelled using a Michaelis-Menten-form model dependent on the ferric/ferrous ratio, previously used to model L. ferrooxidans (van Scherpenzeel, 1998, Breed et al, 1999) and At. ferrooxidans (Boon, 1996) q Fe 2 + =

q max Fe 2 + 1 + K Fe 2 +

[ Fe 3+ ] [ Fe 2 + ]

(8)

Breed et al (1999) described the effect of temperature on Leptospirillum ferrooxidans as an Arrhenius function of the qmax constant

1229

Microbiology Fundamentals

q

max Fe 2 +

= k 0e

−

Ea RT

(9)

and a linear function of the KFe2+ K Fe 2 + = 0.0002T − 0.0453

(10)

Nemati and Webb described the effect of temperature on At. ferrooxidans simply as an Arrhenius function of qFe2+max.

RESULTS AND DISCUSSION Continuous reactors were run at temperatures of 65, 70, and 75ºC, pH 1.5, 12gFe2+.L-1 and steady state data was obtained at dilution rates of 0.02, 0.033, 0.04, 0.05, 0.059, 0.071 and 0.08 h-1. Washout occurred at dilution rates between 0.09 and 0.11 hr-1. All three systems are characterised by a maximum cell concentration at intermediate dilution rates decreasing at both high and low dilution rates (Figure 1). At low dilution rates the system is energy-limited increasing the cell maintenance and reducing the cell concentration, whilst at high dilution rates, the high growth rates stress the cells, which also leads to increased maintenance requirements. The rates of ferrous iron oxidation determined from the rates of oxygen and carbon dioxide utilisation via Equation 6 were compared in Figure 2 to those determined from the difference between feed and reactor contents iron concentrations. The results confirm the applicability of the degree of reduction balance in thermophilic systems. 15

2

-1

Cell concentration (mmolC.l)

3.

1.5

-rFe2+

10

1

5 0.5 0

0 0

0.05

0.1

-1

Dilution Rate (hr )

0.15

Figure 1. Steady state cell concentrations for each dilution rate at 65(∆), 70 (●) and 75ºC (□)

0

5

10

15

-4rO2 - 4.2 rCO2 Figure 2. Confirmation of degree of reduction balance at 65(∆), 70 (●) and 75ºC (□)

The dependence of the system on the redox potential via the ferric/ferrous iron ratio, as predicted by Equation 8, was demonstrated by plotting the specific iron utilisation rate obtained at each steady state against the ferric/ferrous iron ratio measured (Figure 3).

1230

Microbiology Fundamentals

(molFe .molC .hr )

Specific iron utilisation rate

20

15

10

5

0 0.1

1

10 3+

100

1000

2+

[Fe ]/[Fe ]

Figure 3. Specific iron utilisation rates obtained at 65(∆), 70 (●) and 75ºC (□) as a function of the ferric/ferrous iron ratio and modelled by Equation 8

The specific oxidation rates increase with an increase in temperature, this would indicate that the experiments were run within the microbes' preferred operating range. Above this range, thermal deactivation is expected, and a consequent drop in rate. The results show a similar dependence on the ferric/ferrous iron ratio as was observed in mesophilic systems. The specific utilisation rate reaches a maximum at the lowest ferric/ferrous iron ratio corresponding to the fastest dilution rate before washout. The rate decreases with increasing redox potential until the ferrous iron concentration reaches levels below the minimum required for further growth. This has been described as the threshold substrate concentration (Braddock et al, 1984). The data was modelled using Equation 8, obtaining the kinetic constants q max and Fe 2+ K Fe2 + by linear regression. The results show good agreement between data and prediction over the range of potentials measured, suggesting that thermophilic ferrous iron oxidation is proportional to the ferric/ferrous iron ratio and in turn the redox potential. Table 2. Kinetic constants qmax and K Fe2+ derived from the data and compared to Fe 2+

literature values

Present work L. ferrooxidans (Breed et al, 1999) At. ferrooxidans (Boon, 1995)

Temperature

q max Fe 2+

K Fe2 +

65 70 75

11.08 12.32 19.36

0.155 0.098 0.227

40

13.58

0.0033

30

8.8

0.05

The K Fe2 + values obtained show more commonality with those of Acidithiobacillus ferrooxidans, than with those of Leptospirillum ferrooxidans. This has been described as the affinity of the microbe for ferrous iron, with L. ferrooxidans having the capacity to oxidise much lower concentrations of iron and hence operate at much higher redox potentials. Higher K Fe2 + values indicate that the ferric/ferrous iron ratio or redox potential 1231

Microbiology Fundamentals

at which the oxidation and growth rates begin to drop away is lower. This may be of interest in the investigation of why thermophiles are more able to bioleach chalcopyrite, where better leaching may be achieved at lower solution redox potentials.

(molFe .molC .h )

Specific iron utilisation rate

25 20 15 10 5 0 0.1

1

10

100 3+

1000

10000

2+

[Fe ]/[Fe ]

Figure 4. A comparison of the relationship between q Fe2+ and the ferric/ferrous iron ratio for thermophiles (70ºC (—), 75ºC (—)), and mesophiles At. ferrooxidans (----), and L. ferrooxidans(----)

The specific growth rate; this was compared to the specific iron utilisation rate to determine the yield and maintenance coefficients from the Pirt equation (Pirt, 1965) q Fe 2 + =

µ Y max 2+ Fe

+ m Fe 2 +

(11)

X

Straight lines fitting the constant maintenance Pirt equation (Pirt, 1965) are only expected for substrate limited growth. Deviations from linearity at high redox potential and high growth stress may be attributed to increased maintenance requirements and may be better described using the variable maintenance equation (Pirt, 1982). max Table 3. Bioenergetic parameters Yield ( YFe ) and the maintenance coefficient 2+ X

(mFe2+) Temperature (ºC) 65 70 75

YFemax 2+ X .

m Fe2 +

2+ -1

(mol C (mol Fe ) ) 0.0082 0.0082 0.0071

2+.

(mol Fe mol C-1.h-1) 0.394 1.265 1.308

The maximum yield is a very weak function of temperature, showing no significant change within the temperature range measured, and not differing greatly from values measured at mesophilic temperatures, lying between two values obtained for Leptospirillum-like systems — 0.0059 and 0.011 molC.(molFe2+)-1 obtained by Breed et al (1999) and van Scherpenzeel (1998) respectively. The maintenance coefficient is a much stronger function of temperature, indicating increased stress with increased temperature. (Figure 5) can be modelled by the Arrhenius The effect of temperature on q max Fe 2+ equation. The data can thus be modelled as 1232

Microbiology Fundamentals −

54 .437 RT

2 .70 × 10 e [ Fe 3 + ] 1 + K Fe 2 + [ Fe 2 + ]

4

15

3

10

ln(qmax)

-1

qFe2+ (molFe .molC .h )

20 -1

(12)

2+

q Fe 2 + =

9

2

5

1

0

0 0.00285

0

0.05

-1

0.1

y = -6547.6x + 21.718 R2 = 0.8814

0.0029

0.00295 -1

0.003

1/T (degC )

Specific Growth Rate µ (h )

Figure 5. Determination of the max ) energetic parameters, yield ( YFe 2+ X

Figure 6. Effect of temperature on qmax , showing fit of Arrhenius Fe 2+

and the maintenance coefficient ( m Fe2+ ) at 65(∆), 70 (●) and 75ºC (□)

equation

(molFe .molC .h )

-1

15.00

-1 2+

Specific iron utilisation rate

20.00

10.00

5.00

0.00 0.1

1

10 3+

100

1000

2+

[Fe ]/[Fe ]

Figure 7. Specific iron utilisation rates obtained at 65(∆), 70 (●) and 75ºC (□) were compared to values predicted using Equation 8 (—) and Equation 12 (---) 4.

CONCLUSIONS The kinetics of ferrous iron oxidation using thermophilic archaea was investigated in continuous culture. The rate of iron consumption calculated from measured iron concentrations correlated well with that calculated from off gas analysis via the degree of reduction balance. This indicates that the overall reaction stoichiometry for microbial ferrous iron oxidation remains the same as at mesophilic temperatures. 1233

Microbiology Fundamentals

The results were successfully modelled using a Michaelis-Menten form kinetic model, developed by Boon (1996) for mesophilic bioleaching, in which the specific rate of iron oxidation is proportional to the ferric/ferrous iron ratio. The effect of temperature was accounted for by the incorporation of an Arrhenius term. The maximum specific iron utilisation rates determined for the thermophilic systems were not greatly different to those achievable in mesophilic systems, so simple rate of iron oxidation cannot explain the advantage of bioleaching at elevated temperatures. However, the lower tolerance of high redox potential can lead to a bioleach system that self-regulates its redox to a lower level where the rate of ferric leaching is enhanced. NOMENCLATURE cX

concentration of bacteria

mmol C.L-1

D

dilution rate

h-1

E

redox potential of the solution (Pt-Ag/AgCl)

mV

E '0

equilibrium redox potential for Ag/AgCl electrode

mV

Ea

activation energy

kJ.mol-1

F

Faraday constant

coulombs.mol-1

[Fe2+]

concentration of ferrous-iron

mmol Fe2+.L-1

[Fe3+]

concentration of ferric-iron

mmol Fe3+.L-1

Ko

Arrhenius constant

mol Fe2+.(mol C)-1.h-1

KFe2+

ferrous-iron based kinetic constant in bacterial ferrous-iron oxidation

dimensionless

mFe2+

maintenance coefficient on ferrous-iron

mol Fe2+.(mol C)-1.h-1

qFe2+

bacterial specific ferrous-iron utilisation rate

mol Fe2+.(mol C)-1.h-1

q max Fe 2 +

maximum bacterial specific ferrous-iron utilisation rate

mol Fe2+.(mol C)-1.h-1

R

Universal gas constant

kJ.K-1.mol-1

-rFe2+

ferrous-iron utilisation rate

mmol Fe2+.L-1.h-1

-rCO2

carbon dioxide utilisation rate

mmol CO2.L-1.h-1

-rO2

oxygen utilisation rate

mmol O2.L-1.h-1

rX

biomass production rate

mmol C.L-1.h-1

max YFe 2+ X

maximum bacterial yield on ferrous-iron

mol C.( mol Fe2+)-1

z

number of electrons involved in a reaction

dimensionless

γi

Degree of reduction of species i

dimensionless

µ

bacterial specific growth rate

h-1

REFERENCES 1. Boon M., 1996, "Theoretical and Experimental Methods in the Modelling of Biooxidation kinetics of Sulphide Minerals", PhD Thesis, Technische Universiteit Delft, The Netherlands. 2. Boon M., Hansford G.S. and Heijnen J.J., 1995, "The Role of Bacterial Ferrous Iron Oxidation in the Bio-Oxidation of Pyrite", Vargas T., Jerez C.A., Wiertz J.V. and 1234

Microbiology Fundamentals

Toledo H. (eds.), Biohydrometallurgical Processing, 1, Santiago: University of Chile, 153-163. 3. Braddock J.F., Luong H.V. and Brown E.J., 1984, "Growth kinetics of Thiobacillus ferrooxidans isolated from arsenic mine drainage", Applied and Environmental Microbiology 48, (1), 48-55. 4. Breed A.W., Dempers C.J.N., Searby G.E., Gardner M.N., Rawlings D.E. and Hansford G.S., 1999, "The Effect of Temperature on the Continuous Ferrous-iron Oxidation Kinetics of a Predominantly Leptospirillum ferrooxidans Culture", Biotechnology and Bioengineering, 65, 44-53. 5. Clark, D.A., Norris, P.R., 1996, “Oxidation of Mineral Sulphides by Thermophilic Micro-organisms” Minerals Engineering vol. 9 no. 11, pp 1119-1125 6. Crundwell, F. K., 1997, “The kinetics of the chemiosmotic proton circuit of the ironoxidising bacterium Thiobacillus ferrooxidans”, Bioelectrochemistry and Bioenergetics, 43, 115-122 7. Dew, D.W., van Buuren, C., McEwan, K., Bowker, C., 1999, “Bioleaching of Base Metal Sulphide Concentrates: A Comparison of Mesophile and Thermophile Bacterial Cultures”, In: Proceedings of International Biohydrometallurgy Symposium IBS ‘99, Ed: by R. Amils, A. Ballester, Madrid, Elsevier, pp. 229-238 8. Huberts R., 1994, "Modelling of ferrous sulfate oxidation by iron oxidising bacteria - a chemiosmotic and electrochemical approach", PhD Thesis, University of the Witwatersrand, Johannesburg, South Africa. 9. Ingledew W.J., 1982, “Thiobacillus ferrooxidans: The bioenergetics of an acidophilic chemolithotroph”, Biochimica et Biophyisca Acta, 683, 89-117 10. Jeffery, G.H., Basset, J., Mendham, J. and Denney, R.C., 1989, “Vogel’s textbook of quantitative chemical analysis”, 5th edition, Longman Scientific and Technical, New York. 11. Le Roux, N.W., Wakerley, D.S., 1988, “Leaching of chalcopyrite (CuFeS2) at 70 ºC using Sulfolobus”, In: Biohydrometallurgy: Proceedings of the International Symposium Warwick 1987, Ed. by P.R. Norris and D.P Kelly, University of Warwick, 305-319. 12. Meruane, G., Salhe, C., Wiertz, J., Vargas, T., 2002, A novel electro-chemicalenzymatic model which quantifies the effect of the solution Eh on the kinetics of ferrous iron oxidation of Acidithiobacillus ferrooxidans.”, 80, (3), 280-288 13. Nemati M. and Webb C., 1997, "A kinetic model for biological oxidation of ferrousiron by Thiobacillus ferrooxidans", Biotechnology and Bioengineering, 53, (5). 478486. 14. Pirt S.J., 1965, "The maintenance energy of bacteria in growing cultures", Proceedings of the Royal Society B, 163, 224-231. 15. Pirt S.J., 1982, "Maintenance energy: a general model for energy-limited and energysufficient growth", Archives of Microbiology, 133, 300-302. 16. Roels J.A., 1983, “Energetics and kinetics in biotechnology”, Elsevier Biomedical Press, Amsterdam, 20-31 17. van Scherpenzeel D.A., Boon M., Ras C., Hansford G.S. and Heijnen J.J., 1998, "Kinetics of ferrous-iron oxidation by Leptospirillum bacteria in continuous cultures", Biotechnology Progress, 14, 425-433.

1235

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

The role of microorganisms in dispersion of thallium compounds in the environment A. Sklodowska, M. Golan and R. Matlakowska Warsaw University, Faculty of Biology, Laboratory of Environmental Pollution Analysis, CEMERA – Centre of Excellence, Miecznikowa 1, 02-096 Warsaw, Poland E-mail: [email protected] Abstract Thallium is a highly toxic element and very rarely studied in the context of environmental hazards connected with zinc and non-ferrous metal industry. Microorganisms naturally existing in post-flotation and smelt wastes can participate in thallium release from waste deposits and can contribute to its dispersion in the environment. Twenty-one isolates were obtained from wastes of a non-ferrous smelter in Southern Poland characterised by high heavy metal contamination. Ten isolates showed high activity in thallium leaching from wastes (post-flotation and smelt wastes) as well as from pure thallous sulphide. Additionally, cadmium and lead were bioleached from wastes. The isolated bacteria indicated thallium resistance at a concentration up to 100 mg/l and some of them were able to survive in good condition at a concentration of up to 4 g/l. The same bacteria were isolated from rivers and wastewater in this region. A preliminary characterisation of isolates was performed. It was shown that some petroleum products i.e. asphalt-base crude occasionally used for waste immobilisation at the edge of pond or flotation surfactants partially stopped the activity of sulphide oxidising bacteria. 1.

INTRODUCTION Microorganisms naturally existing in post-flotation and smelt wastes can participate in thallium release from waste deposits and can contribute to its dispersion in the environment. The role of microbes in the dispersion of inorganic metal salts (especially sulfides) has been known for years. Oxidation of these compounds is the way of gaining energy, needed in many biochemical processes such as CO2 fixation etc. Metal ions are unused products of reactions, which can penetrate into the environment. Since the 70’s it has been known, that thiobacilli (today genus Thiobacillus is divided into a few genera, belonging to another subclasses of Proteobacteria [6]) can divide thallous sulfide, to obtain sulfide ions – the energy source [5]. Free thallous ions can penetrate into the environment, and take part in many biogeochemical cycles. Microbes can methylate thallium ions, producing dimethylthallium – Me2Tl+. A method of estimating the concentration of this compound in environmental samples was worked out by Schedlbauer and Heumann [8]. The mechanism of this process is still unknown. Probably methylated cobalamine is needed, as in the mercury methylation [4]. 1237

Microbiology Fundamentals

The role of this process in the environment is unknown. Me2Tl+ is very stable [8]. This can suggest, that it is the way of detoxification of the microbe’s environment, similarly to the methyllation of mercury [4]. For the majority of microbes thallium is highly toxic. By disturbing metabolic processes, it stops cell division. For Anacystis nidulans 15 ppm of thallium in medium completely inhibit all metabolic processes, for Chlamydomanas reinhardtii only 5 ppm [7]. Thallium is a highly toxic element and very rarely studied in the context of environmental hazards connected with zinc and non-ferrous metal industry. The research carried out in Southern Poland enabled the identification of several regions, which are seriously threatened by thallium as well as to indicate direct sources of pollution. Polluted regions included mainly the surroundings of the zinc smelter and post-flotation waste ponds.

2

MATERIALS AND METHODS

2.1 Collecting and storing of samples Samples of wastes of a non-ferrous smelter in Southern Poland were collected in October 2001. The material was stored in single-use, sterile plastic tubes. Microbiological analysis was carried out within 24 hours. The remaining material was stored at -20°C. 2.2 Bacterial strains Ten strains were isolated from wastes. Non-modified strains of Halothiobacillus neapolitanus and Paracoccus versutus received from The Department of Bacterial Genetics, Warsaw University were used as a reference due to their potential similarity to isolates from wastes. 2.3 Isolation of strains 5 g of fresh wastes were added to 50 ml of sterile Beijerinck’s medium. The mixture was incubated on rotary shaker, at room temperature. After 24 hours, the culture was transferred to solid medium and cultivated at 28°C for 4 days. Single colonies were inoculated on fresh solid medium every 5 days. 2.4 Media 1. Beijerinck’s medium [2] with 10 g of Na2S2O3.5H2O, as the only energy source and 2 ml of Tuovinen’s salts [10]. Solid medium contained 25 ml of 3% phenol red’s solution, pH = 7,5. 2. Modified LB medium [9] with 20 g of NaCl, pH = 7,5 3. Modified LB medium with thiosulphate [9] with addition of 20 g Na2S2O3.5H2O, pH = 7,5. 4. Davis medium [3] with glucose, as carbon source. 5. Modified Beijerinck’s medium I with 10g of Tl2S instead of thiosulphate. 6. Modified Beijerinck’s medium II: 60 g of wastes, dried to dry weight in 105°C instead thiosulphate, and without Tuovinen’s salts.

1238

Microbiology Fundamentals

2.5 Estimation of bacterial cell number in wastes 5 g of fresh wastes were added to 50 ml of sterile NaCl solution (0.9%) and stored on laboratory rotary shaker for ½ hour. Solutions of this culture were then inoculated on solid media: Beijerinck, modified LB with NaCl and Davis. Plates were stored at 28°C for 7 days. After the incubation the colonies on every plate were counted. 2.6 Characterisation of isolates Isolated strains were inoculated on solid media to test their ability to grow in different conditions. Beijerinck’s, modified LB, modified LB with Na2S2O3 and Davis media were used. All isolates were stained using Gram method. The resistance of strains to thallous ions was tested. Bacterial strains were inoculated on solid modified LB medium with Na2S2O3, containing Tl+ (as TlNO3) in concentrations: 25, 50, 75, 100 ppm, or solid Beijerinck’s medium with analogous ratios of thallim. Paracoccus versutus and Halothiobacillus neapolitanus were inoculated on the same media. 2.7 Preparation of a mixture of strains Fresh cultures of all isolated strains on Beijerinck’s medium were prepared. After 5 days of incubation at room temperature on a laboratory shaker 1 ml of each culture was added to sterile medium without thiosulphate. This mixture was used as an inoculum in next experiments. 2.8 Experiments 2.8.1 Thallium bioleaching from thallous sulphide 100 ml of Beijerinck’s medium with 10 g of Tl2S was inoculated with mixture of bacterial strains. Experiment was conducted in Erlenmayers’ flasks (300 ml) on a laboratory shaker at room temperature. Non-inoculated medium, cultivated under the same conditions, as cultures was control for this experiment. Two series of culture (designated 1st culture and 2nd culture) were prepared. Every day pH, Tl+ concentration and the number of bacterial cell were measured. 2.8.2 Thallium bioleaching from wastes 200 ml of Beijerinck’s medium with 60 g of wastes (dried at 105°C) was inoculated with mixture of strains. Incubation was conducted in Erlenmayer flasks (500 ml), on a shaker at room temperature. Non-inoculated medium served as a control for this experiment. Two series of culture (designated 1st culture and 2nd culture) were prepared. The concentration of thallium, cadmium and lead was estimated. Additionally, the pH and the number of bacterial cell were assessed. The number of bacterial cells was estimated after staining with DAPI and counting on filters under the epifluorescence microscope.

2.9 Chemical analysis The concentration of metals in acidified supernatant and mineralised wastes was measured using Flame Atomic Absorption Spectrometer SOLAAR M6. Before the analysis wastes were dried at 105°C and mineralised in Millestone Laboratory Microwave System with 65% HNO3 and 36% H2O2 (9:1). 1239

Microbiology Fundamentals

The analysis of thallium in soils was carried out with diisopropylether extraction, according to Asami et al. [1]. Mineralised wastes were filtered and distilled water was added, giving the total volume of 100 ml in the flask 50 ml were moved to Erlenmayer flask, 6 ml of HBr (36%) and 1 ml of 0,1% solution of CeSO4.4H2O was added. After 15 minutes the total volume was placed into a separator (200 ml), 5 ml of ether was added and then shaken for 5 minutes. Organic phase was collected and evaporated. The sample was then dissolved in 5 ml of 3% HNO3

3.

RESULTS AND DISCUSSION Concentration of heavy metals and pH of wastes of a non-ferrous smelter were measured and presented in Table 1.

Table 1. Heavy metals concentration and pH of wastes pH 7.00 – 7.10

Tl 40 – 50

Concentration of heavy metals [mg/kg d.w.] Cd Pb Zn 120 – 130 18000 – 21000 4500 – 5000

The bacterial cell number isolated from wastes able to grow on different media was estimated. For all media (Beijerink’s, modified LB and Davis) similar results were observed: 104-105 cells per mg of wastes (wet weight). Twenty-one bacterial strains were isolated from wastes on Beijerinck’s medium. From all isolates, 10 were chosen on the basis of their growth ability. All isolates were Gram-negative, or Gram-variable (young cells were negative, and after 10 days of incubation, positive). Eight of strains were rods, and two of them were too small to identify their morphology under a light microscope.

Table 2. Characteristics of isolates Growth on medium: Strain

Gram

1. negative 2. negative 3. negative 4. variable 5. variable 6. variable 7. variable 8. negative 9. negative 10. negative n/e – not estimated

Morphology rod rod rod rod rod rod n/e rod rod n/e

Beijerinck Modified LB + + + + + + + + + +

+ +

Davis + + +

Modified LB with Na2S2O3 + + + + + + +

The isolated bacteria indicated thallium resistance at a concentration up to 100 mg/l and some of them were able to survive in good condition at a concentration of up to 4 g/l (data not shown).

1240

Microbiology Fundamentals

Table 3. Resistance of freshly isolated bacterial strains to different concentration of thallium ions Strain 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. H.neapolitanus P. versutus

Growth on medium with thallium in concentration: 25 ppm 50 ppm 75 ppm 100 ppm + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + -

The ability of isolates to thallium leaching from pure Tl2S was tested in the first experiment. Sulphide ion was the only energy source for microorganisms. The concentration of Tl+ in a supernatant indicated the rate of the leaching process. The highest concentration of Tl+ (1624 ppm) on the 9th day of cultivation was obtained (Fig. 1). In the 2nd culture the highest concentration was observed at the beginning of the experiment – 250 ppm. From the 5th day it droped to 110 ppm on the last day of cultivation. Throughout the experiment the biggest concentration was noticed in the 1st culture. For the first five days the concentration was stable 210-320 ppm, then it started to increase, reaching 1624 ppm on the 9th day in the 1st culture. This phenomenon may be explained by the irregular structure of the crystal, or some impurities in crystal net. In both cultures brown, stable sediment, localised on flask walls, under the liquid line was observed.

Figure 1. Bioleaching of thallium from Tl2S – concentration of Tl ions in leaching solution Throughout the experiment bacterial cell number systematically increased in both cultures. For both series of cultures similar results were obtained – from about 1 x 104/ml to about 7 x 105/ml (Fig. 2).

1241

cell number [104/ml]

Microbiology Fundamentals

80 60 40 20 0 1

2

3

4

5

6

7

8

9

10

time (days of experiment) first culture

second culture

Figure 2. Bacterial cell number per ml of culture during bioleaching of Tl2S

Figure 3 Bacterial cells stained with DAPI adhered to particle of wastes during bioleaching process Figure 3 presents mixture of isolates cultivated in mineral medium containing wastes. The bacterial cells stained with DAPI attached to particle of wastes are visible. In both cultures pH decrease of about 1 unit was observed. In the first culture the decrease from 7.40 to 6.50 and in the 2nd one from 7.00 to 6.20 was observed while in the control from 7.29 to 7.00. In both cultures pH was stable for the first 5 days, then it decreased (Fig.4.). To check the ability of isolates to leach heavy metals (esp. thallium) from wastes, the experiment using modified Beijerinck’s medium II was carried out. Thallium concentration in wastes (mg/kg d.w.) was measured and its decrease showed the process efficiency. Additional parameters of process were: bacterial cell number, pH changes. The rate of cadmium and lead bioleaching were also measured. Apart from the fouling of culture medium, no other differences between cultures and control images were obsewved during the experiment. In the first culture 40% of thallium was leached. This was more than in the second culture, where it was only 25%. In the control less than 10% of thallium was leached. At 1242

Microbiology Fundamentals

the beginning of the experiment the concentration of thallium was different for both cultures and the control. Probably because the wastes are an unhomogenic blend with unspecified ratio of different compounds (Fig. 5). Cadmium was bioleached in 100% in the 1st culture and more than in 70% in the 2nd one. In the control flask 17% decrease of the concentration was noticed (Fig. 6).

Tl [ppm]

Figure 4. Changes of pH of supernatant during bioleaching of Tl2S 50 40 30 20 10 0 1

4

7

10

13

time (days of experiment) first culture

second culture

control

Figure 5. Thalium concetration in wastes during bioleaching experiment

Cd [ppm]

150 100 50 0 1

4

7

10

13

time (days of experiment) first culture

second culture

control

Figure 6. Cadmium concentration in wastes during bioleaching process For lead 70% decrease of the concentration for the 1st culture and more than 50% for the 2nd one was obtained at the end of the experiment, while in the control about 25%. For 1243

Microbiology Fundamentals

lead, differences in the concentration at the beginning of the experiment, were smaller than for thallium and cadmium (Fig. 7). 25000 Pb [ppm]

20000 15000 10000 5000 0 1

4

7

10

13

time (days of experiment) first culture

second culture

control

Figure 7. Lead concentration in wastes during bioleaching process

cell number [106 /ml]

Bacterial cell number during the experiment was similar for both cultures: from 3.20x106 to 8.05x107 per ml in the 1st one and 5.40x106 – 5.00x107 for the 2nd one. Lagphase was longer in the 2nd culture, than in the 1st one (Fig. 8). 100 80 60 40 20 0 1

2

3

4

5

6

7

8

9 10 13

time (days of experiment) first culture

second culture

Figure 8. Bacteria cell number during bioleaching of wastes Throughout the experiment the pH was stable due to probable buffer capacity of wastes that includes a significant amount of carbonate. The maximal ratio of changes was 0,8 unit in the 1st culture, 1 unit in the 2nd one and only 0,1 unit in the control (Fig. 9). Isolated strains are able to leach thallium from both: pure salts and ores. Isolates are very interesting due to their ability to grow in high concentrations of thallium and other heavy metals, which are thought to be highly toxic. It was shown that microbes' activity might be one of the reasons of thallium contamination near ores treatment plants and near postflotation wastes deposits. This problem raises many new questions concerning the physiology of their resistance to metal ions. Additionally, isolated microbes might be used in the biohydrometallurgical methods of this metal extraction. This research will be continued to understand the problem of the taxonomy and physiology of these isolates, and find possibilities of using them into biotechnological processes. 1244

Microbiology Fundamentals

pH

Another issue which arises is how to inhibit microbes activity in the deposits, or how to retain dissolved metal ions within the deposit. 7,6 7,2 6,8 6,4 6 1

2

3

4

5

6

7

8

9 10

time (days of experiment) first culture control

second culture

Figure 9. Changes of pH of supernatant during bioleaching of wastes REFERENCES 1. T. Asami, C. Mizui, T.Shimada, M. Kubota, Fresenius J. Anal. Chem. 356 (1996) 348. 2. M.W. Beijerinck, Zentralbl. Bakteriol. Parasitenkd. Infenktionskr. Hyg. Abt. II , 11 (1904) 593. 3. D.H. Davis, M. Doudoroff, R.Y. Stanier, M. Mandel, Int. J. Syst.Bacteriol., 19 (1969) 375. 4. G.M. Gadd, Encyclopedia of Microbiology, Vol. 2, Academic Press, Inc., 1992. 5. M. Galizzi, E. Ferrari, Appl. Environ. Microbiol., 32 (1976) 433. 6. D.P. Kelly, A.P. Wood, Int. J. Syst. Evolution. Microbiol., 50 (2000) 511. 7. B. Lustigman, L.H. Lee, J. Morate, F. Khan, Bull. Environ. Contam. Toxicol., 64 (2000) 565. 8. O.F. Schedlbauer, K.G. Heumann, Appl. Organometal. Chem., 14 (2000) 330. 9. A. Sklodowska, R. Matlakowska, W. Ludwig, Acta Microbiol. Polon., Vol. 45 No 2 (1996) 131. 10. O.H. Tuovinen, D.P. Kelly, Arch. Microbiol., 111 (1973) 257.

1245

C HAPTER 5 Molecular Biology and Taxonomy

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

A promiscuous, broad-host range, IncQ-like plasmid isolated from an industrial strain of Acidithiobacillus caldus, its accessory DNA and potential to participate in the horizontal gene pool of biomining and other bacteria Gunther K. Goldschmidt, Murray N. Gardner, Leonardo J. van Zyl, Shelly M. Deane and Douglas E. Rawlings Department of Microbiology, University of Stellenbosch, Private Bag X1, Matieland 7602, South Africa Abstract A consortium of bacteria consisting primarily of the iron-oxidizing, Leptospirillum ferriphilum and the sulfur-oxidizing, Acidithiobacillus caldus were found to dominate the population of organisms in industrial continuous-flow tank reactors used to oxidize arsenopyrite concentrate. A 14.15 kb plasmid was isolated from At. caldus strain f which was present in the consortium of cells. The plasmid, pTC-F14, was found to belong to the IncQ-like group of highly promiscuous, mobilizable, broad host-range plasmids. Plasmid pTC-F14 has a replicon and mobilization region closely related to pTF-FC2, a 12.2 kb plasmid isolated from Acidithiobacillus ferrooxidans about 15 years previously. Surprisingly, the replication and mobilization proteins of another broad host-range IncQlike plasmid, pRAS3.2 (isolated from the fish pathogen, Aeromonas salmonocida in Norway), are even more closely related to pTF-FC2 than plasmids pTC-F14 and pTF-FC2 are to each other. This suggests that these highly promiscuous IncQ-like plasmids are potential vehicles for the horizontal transfer of DNA between bacteria from very different environments. The sequence of plasmid pTC-F14 has been completed and the region that contains the accessory genes has been analysed. Present within this region is an insertion sequence ISAtc1, that is most closely related (92% nucleotide identity) to the mobile element, ISAfe1, previously identified in many isolates of At. ferrooxidans and At. thiooxidans. ISAtc1 is present in three At. caldus strains isolated from South Africa but not present in three At. caldus strains from Europe or Australia. The presence of insertion sequences on both a plasmid and the chromosome allows plasmids to integrate into the chromosome and provides an enhanced level of genome plasticity. Plasmids pTC-F14 and pTF-FC2 and the accessory genes that they contain are analysed and compared. Keywords: Acidithiobacillus caldus, plasmids, accessory genes, horizontal gene pool

1249

Molecular Biology and Taxonomy

1.

INTRODUCTION Plasmids are pieces of extrachromosomal DNA that are replicated independently of the host chromosome and may contain genes that, while not essential for host survival under some conditions, may enhance survival of a host cell under other circumstances (e.g. antibiotic or metal ion resistance genes). Some plasmids can be transferred between bacterial hosts by a mating process called conjugation [1]. Self-transmissible plasmids contain all of the genes required for conjugation, while mobilizable plasmids encode for a subset of genes required for DNA processing only and require the presence a selftransmissible plasmid for conjugation. Self-transmissible plasmids are generally larger in size (>30 kb) than mobilizable plasmids (typically 6 to 20 kb). Plasmids are widespread amongst bacteria and have been reported to contribute from 1 to >10% of the total genome of many bacterial species [2]. As conjugation is not restricted to members of the same species, but also takes place between species, self-transmissible and mobilizable plasmids play an important role in the horizontal gene pool that is shared between many organisms. Although most plasmids are narrow host-range and can replicate only in closely related species, other plasmids are capable of replication in many types of bacteria. IncQ or IncQ-like plasmids are relatively small in size (5 to 15 kb) and capable of replication in wide variety of Gram-negative and Gram-positive bacteria [3]. Furthermore, these plasmids are mobilized by a family of broad host-range plasmids known as IncP plasmids (as well as the Ti-plasmids of Agrobacteria). As a result, IncQ and IncQ-like plasmids are highly promiscuous. We investigated plasmids from biomining bacteria to discover what types of genes are present within the mobile gene pool of bacteria growing in low pH inorganic mineral environments and whether the replication and mobilization genes of plasmids from these bacteria are related to those of other bacteria. This research should help address the question of whether plasmids from acidiphilic, chemolithotrophic bacteria are part of an isolated gene pool or whether they are active participants in the horizontal gene pool shared by other bacteria. We report on the analysis of an IncQ-like plasmid from a strain of the sulfur-oxidizing, moderately thermophilic bacterium (optimum 45-50°C), Acidithiobacillus caldus [4]. 2.

MATERIALS AND METHODS Media and growth. At. caldus strains were grown at 37°C (rather than the 45-50°C optimum as aeration facilities were better) in tetrathionate medium (3 mM), sterilised and adjusted to pH 2.5 as reported previously [5]. At. caldus cultures were purified using solid FeSo overlay medium that incorporates the acidophilic heterotroph Acidiphilium SJH into the lower layer [6]. Bacteria and plasmids are shown in Table 1. Southern hybridization. Labelling of probes, hybridization and detection was performed by using a digoxigenin-dUTP non-radioactive DNA labelling and detection kit (Roche). Hybridization was at 40°C in Easy Hyb (Roche) followed by two non-stringent washes at 25°C (in 2 X SSC, 0.1% SDS) and two stringent washes at 65°C (0.1 X SSC, 0.1% SDS). DNA sequencing and bioinformatics. The isolation and cloning of plasmid pTC-F14 was described previously [7]. DNA sequencing was by the dideoxy chain termination method, using an ABI PRISMTM 377 automated DNA sequencer and the sequence was analysed using a variety of software programmes but mainly the PC based DNAMAN (version 4.1) package from Lynnon BioSoft. Comparison searches were performed using 1250

Molecular Biology and Taxonomy

the gapped-BLAST program at the National Center for Biotechnology Information. The phylogenetic trees were constructed using the ClustalW-based multiple sequence alignment tool in DNAMAN. Table 1. Details of bacterial strains and plasmids used in this study Bacterial strain At. caldus "f" #6 MNG C-SH12 BC13 KU DSM8584 Plasmids pTF-FC2

pTC-F14

3.

Geographical origin

Source or reference

Nickel pilot plant, Billiton, Randburg, South Africa Fairview mine, Barberton, South Africa Arsenopyrite pilot plant, UCT Continuous bioreactor, Brisbane, Australia Birch Coppice, Warwickshire, UK Kingsbury coal spoil, UK

Own laboratory Own laboratory Murray Gardner Kevin Hallberg [4] [4]

From At. ferrooxidans from a mixed culture used to biooxidize an arsenopyrite concentrate from the Fairview mine, Barberton, South Africa From At. caldus strain ‘f’ above

[8]

[7]

RESULTS

3.1 Comparison of plasmid ‘backbone’ genes At. caldus strain f, contains at least two plasmids, one of approximately 45 kb, which has not yet been cloned and a smaller plasmid called pTC-F14 [7], the DNA sequence of which has recently been determined (14,149 bp, unpublished). Plasmid pTC-F14 is closely related to pTC-FC2 (Figure 1, Table 2) previously isolated from At. ferrooxidans [8]. Both belong to the family of IncQ-like plasmids, and are therefore broad host-range, mobilizable, highly promiscuous plasmids. The plasmid ‘backbone’ consists of those genes and sites associated with aspects of plasmid biology and includes functions such as replication, conjugation (mobilization) and stability [3]. pTF-FC2 0

12180

orf18.9 mobD oriT mobA/repB mobE mobC mobB mobA repB

repA pasABC

oriV repC

38bp grx orf43.4 res 38bp merR-like tnpR/tnpA* Tn5467

pTC-F14 0

14149

mobD oriT mobA/repB mobE mobC mobB mobA repB

repA pasAB

oriV repC

orf13 orf20.8 orf17.4

orf33.2

26bp tnp 26bp ISAtc1 orf9.5

Figure 1. Comparison of genes, open reading frames and sites of plasmids pTF-FC2 and pTC-F14 A comparison of the proteins of pTF-FC2 and pTC-F14 involved in plasmid replication, mobilization and the toxin-antitoxin stability systems is shown in Table 2. All three replication proteins (RepA, RepB and RepC), two of the plasmid addiction system proteins (PasA and PasB), as well as two of the five mobilization proteins (MobA and MobB) are highly conserved with amino acid sequence identities of between 72 and 81%. 1251

Molecular Biology and Taxonomy

Likewise the location of these proteins with respect to each other on the two plasmids is highly similar (Figure 1). This suggests that the two plasmids had a common plasmid ancestor. Table 2. Comparison of the replication, addiction and mobilization proteins of plasmids pTCF14 and pTF-FC2 Protein RepA RepB RepC PasA PasB PasC MobA-RepB MobB MobC MobD MobE

Function replication specific helicase plasmid specific DNA primase iteron-specific binding protein antitoxin of plasmid addiction system toxin of plasmid addiction system toxin-antitoxin accessory protein oriT-specific relaxase oriT-processing accessory protein DNA-binding accessory protein mobilization protein of unknown function mobilization protein of unknown function

Amino acids 291 352 303 74 90 833 103 131 226 220

pTC-F14 Mol mass (Da) 31289 40623 33712 8523 10483 95792 11198 13969 24698 23811

pI 5.92 9.73 9.28 4.46 10.36 9.50 9.72 10.03 6.60 5.53

Amino acids 290 352 299 74 90 71 831 106 118 227 213

pTF-FC2 Mol mass (Da) 31227 40111 33740 8453 10307 7676 94854 11605 12941 25274 23093

pI 6.21 9.77 8.99 4.71 10.4 3.76 9.59 9.79 10.01 5.25 8.19

% amino acid identity 81.0 78.4 74.2 81.1 72.2 75.0 77.4 22.7 39.4 19.8

Recently, another IncQ-like plasmid has been isolated from the fish pathogen, Aeromonas salmonicida in Norway (L'Abée-Lund, NCBI accession number, AY043298). This plasmid, called pRAS3, carries a tetracycline resistance gene and regulator. Surprisingly, two of the three replication proteins and all five of the mobilization proteins are substantially more closely related to pTF-FC2, than pTF-FC2 is to pTC-F14 (Figure 2A and B). The most likely interpretation of this observation is that pRAS3 and pTF-FC2 diverged from a common ancestor more recently than pTF-FC2 and pTC-F14, even though the latter two plasmids were isolated from At. ferrooxidans and At. caldus, bacteria that share the same ecological niche. The observation that pTF-FC2 and pRAS3 are closer relatives than pTF-FC2 and pTF-F14 is supporting evidence of how promiscuous the IncQ-like plasmids may be. These IncQ-like plasmids are therefore potentially important vehicles in the horizontal distribution of the genes they carry between amongst a broad bacterial community. 3.2 Accessory genes Accessory DNA contains ‘passenger’ genes and functions that are not directly involved with plasmid biology but which may be either parasitic or increase the fitness of the host in which the plasmid resides. Those accessory genes that improve host fitness are expected to be preferentially selected, as these should help to counter the additional metabolic burden that replication and maintenance of the plasmid places on the host. It is therefore of great interest to examine the accessory DNA in the hope of identifying what types of genes the plasmid has acquired. A list of the accessory genes found on plasmids pTF-FC2 and pTC-F14 is shown in Table 3.

1252

Molecular Biology and Taxonomy

pRAS3

pRAS3

94%

pTF-FC2 pTC-F14 RSF1010

78%

pTC-F14 44%

94%

pIE1130

pIE1107/ pDN1

15%

98% 91%

pIE1115

92%

74%

pIE1130

pIE1107 pDN1

A

97%

pTF-FC2

81%

99%

RepB

RF1010

RepA

99%

pIE1115

pRAS3 pTF-FC2

74%

35%

pTC-F14 62%

pIE1107/pDN1 /pIE1115 pIE1130

93% 90%

RepC

RSF1010

25%

pTC-F14

pTC-F14

22%

91%

27%

77%

pTF-FC2

72%

pTF-FC2

58%

R751

67%

R751

pRAS3

B

RP4

RP4

92%

pRAS3

37%

pRA2

pRA2

MobB/TraJ

MobA/TraI 400 aa RP4

53%

R751 pRAS3

32%

pTF-FC2

23%

pTC-F14

24%

pRA2

RP4 R751

MobD/TraL

18%

pTF-FC2

pTF-FC2

pRA2

69%

R751

17%

pTC-F14

pRAS3

MobE/TraM

RP4

85%

MobC/TraK

96%

pRAS3

39% 98%

10%

pTC-F14

92% 23%

10%

pRA2

Figure 2, A and B. Phylogenetic relationships between the replication (A) and mobilization proteins (B) of the IncQ plasmid family. Percentages are amino acid sequence identities 1253

Molecular Biology and Taxonomy

3.2.1 Accessory DNA from pTF-FC2 isolated from At. ferrooxidans Most of the accessory DNA of plasmid pTF-FC2 consists of a defective Tn21-like transposon, Tn5467, which has been reported previously [9]. Although Tn5467 was not able to transpose or resolve on its own, its 38 bp terminal repeats and res sites were functional, as Tn5467 was able to transpose and resolve if the genes for Tn21 transposase and resolvase were provided in trans. Tn5467 contains three open reading frames (ORFs) which are potentially functional. One of these, encoded by a gene now called grx, encodes a glutaredoxin-like protein that was shown to functionally complement thioredoxin deficient mutants of the bacterium, Escherichia coli for the ability to grow on minimal medium lacking glutathione. Thioredoxin is also essential for the arsenate resistance activity of a family of arsenate reductases (product of arsC gene) [10]. These enzymes use thioredoxin to reduce arsenate to arsenite prior to its removal from a cell by an arsenite efflux pump (product of arsB gene). We have recently shown (B. Butcher, unpublished), that product of the glx gene present on Tn5467 is able to substitute for thioredoxin and allows the cloned At. ferrooxidans arsC gene product to reduce arsenate to arsenite, thereby conferring additional arsenate resistance to an E. coli thioredoxin (trxA) mutant. Table 3. Accessory proteins of plasmids pTF-FC2 and pTC-F14 Size amino Putative acids and protein or (Da) ORF Plasmid pTF-FC2 ORF18.9 170 (18925) ORF8 85 grx (9042) MerR-like 137 regulator (15097) ORF43

406 (43416)

Plasmid pTC-F14 ORF13 124 (13008) ORF20.8 189 (20795) ORF17.4 153 (17405) ORF33 transposase ORF9.5

286 (33169) 404 (46188) 86 (9485)

Putatative RBS GAGGG GGAGAA GGAGGA GGAGAA

AGGAGA AGGCGA AGGAG

AGGAG AGGAG AGGAG

Most related protein and proposed function and predicted size

% identity/ similarity (part of protein)a

BLAST E value

Reference NCBI accession no.

No meaningful BLAST hits

NA

NA

NA

thioredoxin from 186 kB plasmid, named beta and present in Nostoc sp. PCC7120, 83 aa copper efflux transcriptional regulator, hmrR, preventing low pH copper toxicity in Sinorhizobium meliloti, 147 aa 12 transmembrane segment, multidrug resistance-like protein in genome of Synechocystis sp. PCC 6803, 418 aa no meaningful BLAST hits invertase/recombinase protein, Xanthomonas axonopodis pv. citri str. 306, 209 aa conserved hypothetical protein, Geobacter metallireducens, 110 aa Pseudomonas fluorescens, 146 aa Ralstonia metallidurans, 145 aa aminotransferase, Bacillus halodurans, 397 aa transposase from ISAfe1, Acidithiobacillus ferrooxidans, 404 aa no meaningful BLAST hits

-9

42/67 (63)

3e

41/57 (128)

4 e -19

Q9X5X4

36/48 (323)

2 e -54

NP_442543

NA

NA

NA

61/71 (186)

-54

5e

AP003602

NP_644692

65/88 (110) 53/70 (141) 51/74 (145) 27/48 (133) 92/95 (404)

1e-35 2e-35 3e-35 6.8e-2

ZP_00080185 ZP_00082921 ZP_00025386 NP_244179

0.0

AAB07489

NA

NA

NA

a

Part of protein, is the number of amino acids over which the similarity/identity to the highest match in the NCBI data base was determined. NA, not applicable

New closest matches to proteins in the database have been obtained using the BLAST program. Interestingly, the closest match to what was previously called the MerR-like family regulator is to a transcriptional regulator of a copper efflux mechanism that reduces the toxicity of copper at low pH in Sinorhizobium meliloti [11]. While the closest match to Tn5467 ORF43 is to a twelve, transmembrane-spanning protein that is related to the family of multidrug exporters. It is likely that the MerR-like family transcriptional regulator regulates expression of ORF43, however, attempts to detect increased resistance 1254

Molecular Biology and Taxonomy

to metal ions or antibiotics in E. coli due to the presence of these two ORFs have so far been unsuccessful (unpublished). The amino acid sequence of ORF18.9, which falls outside of the Tn5467 inverted repeat sequences did not give any meaningful similarity hits using the BLAST program. 3.2.2 Accessory DNA of pTC-F14 isolated from At. caldus Six open reading frames that gave putative translation products of 9 kDa or larger and that were preceded by putative ribosome binding sites were detected in the accessory DNA of pTC-F14 (Table 3). Two of these, ORFs 13 and 9.5 gave no meaningful similarity hits using the BLAST program. ORF33, gave relatively weak similarity and identity to approximately one third of the amino acid sequence of an aminotransferase. However, this level of similarity was considered to be insufficient to assign this as a likely function of the putative protein. The remaining three ORFs had strong amino acid sequence relationships to ORFs already present in the NCBI database. ORF20.8 had the highest sequence match to an invertase or recombinase, previously found in Xanthomonas axonopodis pv citri. The function of these enzymes is the rearrangement of DNA within a sequence or the exchange of DNA between sequences, but this property is not specifically associated with the environment of At caldus. Similarly, ORF17.4 is related to a hypothetical protein that is highly conserved in a wide variety of bacteria of which only the three highest matches are shown in Table 3. In spite of its wide distribution, the function of this hypothetical protein is unknown. The remaining ORF of 404 amino acids had very high sequence identity to the transposase of an insertion sequence, ISAfe1 [12], previously identified in At. ferrooxidans a relative of At. caldus. 3.3 ISAtc1 and its comparison with ISAfe1 ISAfe1 (previously called IST1), is one of two types of insertion sequences found in the genome of a several, but not all, strains of At. ferrooxidans and At. thiooxidans [12, 13]. The putative transposase found on At. caldus plasmid pTC-F14, is clearly a close relative of ISAfe1 and we have named it ISAtc1. The two transposases are both 404 amino acids in length and sequence alignment indicated that the proteins are 92% identical. Both IS elements are 1303 bp in size. Like ISAfe1, ISAtc1 is flanked by two sets of imperfectly conserved 26 bp inverted repeat sequences which are strongly conserved, with the two left and two right termini having 1 and 3 bp sequence variations between the two IS elements respectively (underlined Table 4). Table 4. Alignment of the 5` and the 3` terminal inverted repeats of ISAfe1 with ISAtc1 Insertion sequence

Left and right terminal IR sequences

ISAfe1

5`-GGCTCTTCGTCGGATTGAGTGGGTAG 3`-GGCTCTTCGTCATTTCAAGTGGGTAG 5`-GGCTCTTCGTCAGATTGAGTGGGTAG 3`-GGCTCTTCGACGTTTCATGTGGGTAG

ISAtc1

The complementary strand of the right hand IR is shown to facilitate comparison. We wished to determine whether ISAtc1 elements are as widespread amongst At. caldus strains as ISAfe1 is amongst At. ferrooxidans strains [12, 13]. Genomic DNA was prepared from six At. caldus strains, two of which originated from Europe, one from Australia and three from South Africa. A Southern hybridization experiment was carried 1255

Molecular Biology and Taxonomy

out in which the genomic DNA was hybridised to a labelled probe prepared from ISAtc1 and the result is shown in Figure 3. Positive hybridization signals were obtained for all three of the At. caldus isolates from South Africa but not from any of the others. Approximately 11-14 bands were obtained for each of the South African isolates, which indicated that multiple copies of ISAtc1 were present. Some of the bands appeared to be of a similar in size, but there were also clear differences in banding pattern between isolates.

Figure 3. DNA hybridization experiment showing that three strains have this insertion sequence on their chromosomal DNA. Lane 1, MNG; lane 2, “f”; lane 3, #6; lane 4, CSH12; lane5, BC13; lane 6, KU and lane 7, a subclone of pTC-F14 containing only accessory DNA and used as positive control. The probe was made using an internal fragment of the tnp gene of ISAtc1 4.

DISCUSSION Free-living bacteria typically have in excess of one thousand genes, the majority of which encode for essential cell functions that are required for cell viability. These are the so-called, ‘house-keeping genes’. Bacteria also have access to a pool of genes that can be acquired from other bacteria or even non-bacteria, known as the "horizontal" gene pool [2]. Although some house-keeping genes may become caught up in this horizontal gene pool, the horizontal gene pool is thought to consist mostly of genes that are not essential to host survival under some circumstances. However, genes that may increase host cell fitness under certain circumstances can be recruited from the horizontal gene pool and then lost again when no longer advantageous [14]. Several types of genetic elements can play a role in moving DNA between bacteria, and sometimes integrating them into the chromosome, including plasmids, bacterial phages, transposons and insertion sequences. From this and previous studies, the environmental niche in which the IncQ-like plasmids are found clearly includes the highly acidic, mineral rich, inorganic environments in which the acidithiobacilli grow. These plasmids can therefore presumably move between bacteria that grow within this ecological niche. Although pTF-FC2 and 1256

Molecular Biology and Taxonomy

pTC-F14 are closely related, they have relicons that are compatible with each other and can be coresident in the same cell for extended periods in the absence of external selection [7]. They therefore appear to have adapted to living in the same environment as each other. The discovery of a close relative of pTF-FC2 in a Norwegian salmon pathogen further supports the view that these plasmids are capable of moving between bacteria from very different ecological niches. IncQ-like plasmids are therefore potential role players in the movement of genes between many different types of bacteria. Most of the IncQ-like plasmids already discovered carry genes for antibiotic resistance [3]. This is not unexpected as antibiotic resistance is a particularly easy property to screen for. Only one IncQ-like plasmid discovered to date, carries no recognizable accessory DNA (plasmid pDN1, isolated from an Australian strain of the sheep foot-rot causing bacterium, Dichelobacter nodosus [15]). Analysis of the accessory DNA present on pTC-F14 was disappointing as no genes that confer a property that is known to provide a selective advantage to the host were detected. The products of open reading frames, ORF13, ORF33.2 and ORF17.4 may be advantageous to a host cell, but the functions of these putative proteins are unknown. Of these three ORFs, ORF17.4 is particularly interesting as highly related ORFs are present among a wide range of bacteria, which improves the likelihood that its function will be discovered. The insertion sequence ISAtc1 is clearly closely related, though not identical to ISAfe1 of At. ferrooxidans ATCC19859 and related bacteria. It has been proposed that movement of ISAfe1 is responsible for the phenomenon of phenotypic switching in At. ferrooxidans between a wild-type state in which both ferrous iron and reduced sulfur compounds can be oxidized and a mutant state during which the ability to oxidize ferrous iron is lost but the ability to oxidize reduced sulfur is retained [16]. ISAfe1 has been shown to insert within the resB gene which encodes for a putative cytochrome c-type biogenesis protein. It is proposed that this insertion results in loss of activity of this c-type cytochrome which is thought to be required for ferrous iron but not sulfur oxidation. Here we have shown that a copy of ISAtc1 exists on a plasmid as well as several copies in the chromosome of At. caldus. The presence of IS elements on both plasmid and chromosome provides several potential sites for the plasmid to integrate into the chromosome [17, 18 and references therein]. A small number of integrated plasmids may excise from the chromosome carrying pieces of chromosomal DNA (prime plasmid formation), which then may be transferred to new cells by conjugation. This is a mechanism by which chromosomal genes can enter the horizontal gene pool and IS elements have been associated with the assembly of sets of accessory genes [18]. IS elements can also confer an increased level of plasticity to a chromosome, serving as sites of chromosome rearrangments such as DNA recombination, inversion, integration and deletion. Although an analysis of the accessory DNA from pTF-FC2 has been reported in 1995, we have reanalysed this DNA in the light of the greatly expanded database now available. The highest BLAST hits to the MerR-like regulator and ORF43.4 are particularly interesting. The MerR-like regulator has the closest match to the heavy metal response regulator (HmrR) from S. melitoli and that has been shown to regulate the protein AtcP a copper transporting ATPase [11]. Next highest matches are to MerR-family regulators, mostly of unknown function, but including a zinc-responsive transcriptional regulator (ZntR) of 141 amino acids in length from Salmonella typhimurium (NCBI accession number NP_462316, similarity 62%, identity 35% over 129 amino acids, e-16). ORF43.4 is clearly related to a large family of membrane located multidrug efflux proteins, and some of the multidrug efflux family of proteins are capable of conferring resistance to metal ions such as cobalt, nickel, cadmium or zinc [19]. Furthermore, many of the multidruglike transporters are members of the MerR-like family of regulators [20]. There is 1257

Molecular Biology and Taxonomy

therefore a possibility that the MerR-like regulator and associated 12 TMS multidrug resistance family-like ORF represent a metal ion transport mechanism. The reason for association of the functional glutaredoxin-like encoding gene (grx) with these two putative genes is uncertain, but its ability to substitute for thioredoxin as an electron donor in the reduction of arsenate, suggests that it could also be used as an electron donor for the reduction of other metals. ORF43.4 may not be functional in E. coli since it is a membrane protein and the pH gradient across the membrane is very different in E. coli compared with At. ferrooxidans. However, we intend to renew efforts to determine whether ORF43.4 and the MerR-like regulator are functional and to identify their target (s). ACKNOWLEDGEMENTS This work was supported by grants from the National Research Foundation, The Human Resource for Industry Programme (Pretoria, South Africa), BHP-Billiton as well as the University of Stellenbosch. REFERENCES 1. Clewell, D.B. (ed) Bacterial conjugation. Plenum Press, New York and London. (1993). 2. Thomas, C.M. (ed) The Horizontal Gene Pool. Harwood Academic Publishers, Amsterdam. (2000). 3. Rawlings D.E. and E. Tietze. Microbiol. Mol. Biol. Rev. 65 (2001) 481-496. 4. Hallberg, K.B., and E.B. Lindström. Microbiol. 140, (1994) 3451-3456. 5. Butcher, B.G., S.M. Deane and D.E. Rawlings. Appl. Environ. Microbiol. 66 (2000) 1826-1833. 6. Johnson, D.B. Microbiol. Methods 23 (1995) 205-218. 7. Gardner, M.N., S.M. Deane and D.E. Rawlings. 2001. J. Bacteriol. 183 (2001) 33033309. 8. Rawlings, D.E., I-M. Pretorius and D.R. Woods. J. Bacteriol. 159 (1984) 737-738. 9. Clennel, A-M., B. Johnston and D.E. Rawlings. Appl. Environ. Microbiol. 61 (1995) 4223-4229. 10. Mukhopadhyay, R., B.P. Rosen, L.T. Phung and S. Silver. FEMS Microbiology Reviews 26 (2002) 311-325. 11. Reeve, W.G., R.P. Tiwari, N.B. Kale, M.J. Dilworth and A.R. Glenn. Mol. Microbiol. 43 (2002) 981-991’ 12. Holmes D.S., H-L. Zhao, G. Levican, J. Ratouchniak, V. Bonnefoy, P. Varela and E. Jedlicki. J. Bacteriol. 183 (2001) 4323-4329. 13. Holmes, D.S. and R. Ul Haq. Biohydrometallurgy 1989. Salley, J., McCready, R.G.L. Wichlacz, P.L. (eds) Canmet, Ottawa, Canada. (1989) pp 115-127. 14. Ochman, H., J.G. Lawrence and E.A. Groisman. Nature 405 (2000) 299-304. 15. Whittle, G., M.E. Katz, E.H. Clayton and B.F. Cheetham. Plasmid 43 (2000) 230-234. 16. Cabrejos, M-E., Zhao, M. Guacucano, S. Bueno, G. Levican, E. Garcia, E. Jedlicki and D.S. Holmes. FEMS Microbiol. Lett 175 (1999) 223-229. 17. Neidhardt, F.C., R. Curtiss III, J.L. Ingraham, E.C.C. Lin, K.B. Low, B. Magasanik, W.S. Reznikoff, M. Riley, M. Schaecter, H.E. Umbarger (eds) Escherichia coli and 1258

Molecular Biology and Taxonomy

Salmonella: cellular and molecular biology. 2nd ed. ASM Press, Washington, D.C. (1996) 18. Mahillon, J. and M. Chandler. Microbiol. Mol. Biol. Rev. 62, (1998) 725-774 19. Paulsen, I.T., M.H. Brown and R.A. Skurray. Microbiol. Mol. Biol. Rev. 60, (1996) 481-496. 20. Putman, M., H. W. van Veen and W.N. Konings. Microbiol. Mol. Biol. Rev., 64 (2000) 672-693.

1259

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Analysis of salt-induced outer membrane proteins in Acidithiobacillus ferrooxidans NASF-1 K. Kamimura, M. Yamakado, T. Shishikado and T. Sugio Division of Science and Technology for Energy Conversion, Graduate School of Natural Science and Technology, Okayama University, 3-1-1 Tsushima-Naka 700-8530, Japan Abstract Acidithiobacillus ferrooxidans is an acidophilic chemolithotrophic bacterium capable of oxidizing ferrous ion or reduced inorganic sulfur compounds. Outer membrane proteins of this bacterium are probably involved in response to environmental changes. The clarification of molecular mechanisms involved in environmental adaptation is very important to understand the physiology of this acidophilic bacterium. Effects of salt on the composition of outer membrane proteins in At. ferrooxidans strain NASF-1 were examined by polyacrylamide gel electrophoresis. The amount of two proteins with apparent molecular masses of 30 kDa and 40 kDa increased in the outer membrane prepared from cells grown in Fe2+-medium supplemented with NaCl. The N-terminal amino acid sequence of 40 kDa protein had almost same sequence as that of Omp40 previously detected in phosphate-starved At. ferrooxidans strain ATCC 18959. Northern blot hybridization analyses revealed that the expression of omp40 gene was stimulated in cells incubated in Fe2+-medium supplemented with NaCl, but not in cells incubated in Fe2+-medium with KCl or Na2SO4. A search using the N-terminal sequence of the 30 kDa protein in the TIGR pre-released genomic data of At. ferrooxidans ATCC 23270 using the Blast algorithm revealed the presence of one open reading frame having the same Nterminal amino acid sequence as that of 30 kDa protein. The gene encodes a protein of 217 amino acids with a predicted molecular mass of about 20 kDa. The first 27 amino acids were not present in the mature protein and probably represent a signal peptide. Homology search in databases using the Blast algorithm revealed no protein with sequence similarity to the N-terminal part of the protein. The C-terminal part of the protein had strong sequence similarity with proteins of the OmpA family. 1.

INTRODUCTION An outer membrane of Gram-negative bacterium is a structure exposed directly to environmental changes by external stimuli. The outer membrane acts as a molecular sieve that allows the passage of ions and small hydrophilic organic molecules. This property is due to the presence of a major group of proteins, porins, that form diffusion pore [1-5]. Porins have been well characterized in Escherichia coli. Their primary and secondary structures are known especially for general diffusion porins, OmpC and OmpF, which had been well studied in response to osmotic pressure. OmpC and OmpF are similar in amino 1261

Molecular Biology and Taxonomy

acid sequence [6], immunological reactivity [7] and ion selectivity [8], yet different in pore size [9], phage selectivity [10] and regulation [11]. The mechanism for response to osmotic pressure in E. coli involves two-component system, resulting in the qualitative and quantitative changes of outer membrane proteins. Cells in high-osmolarity medium have high levels of OmpC and low levels of OmpF, while the opposite is true in lowosmolarity medium. Two components, EnvZ and OmpR, are involved in the regulatory system for the expression of OmpC and OmpF proteins in E. coli. EnvZ is an innermembrane protein responsible for sensing the external osmolarity and OmpR acts as a transcriptional regulator. As EnvZ senses high levels of osmolarity, it phosphorylates OmpR [12, 13]. The phosphorylated form of OmpR binds the ompF and ompC regulatory regions and regulates transcription [14]. Acidithiobacillus ferrooxidans is a Gram-negative, acidophilic chemolithotrophic bacterium capable of oxidizing ferrous ion or reduced inorganic sulfur compounds, and is involved in bacterial leaching of metals from sulfide ores [15-19]. It has been reported that the relative synthesis of proteins of At. ferrooxidans were influenced by environmental factors, such as pH [20], substrate [21-23] and phosphate source [24,25]. In At. ferrooxidans ATCC 18959, a major outer membrane protein having an apparent molecular mass of 40 kDa (Omp40) has been studied [24,26], and a possible role for the protein in forming small pores has been reported [27]. The studies on Omp40 protein from At. ferrooxidans ATCC 18959 have indicated that the protein was organized in a trimeric structure and formed a small ionic channel [27,28]. The degree of identity of amino acid sequence of Omp40 protein to porins from enterobacteria was only 22%. Nevertheless, multiple alignments of this sequence with OmpC porin from E. coli has shown several important features conserved in the At. ferrooxidans surface protein [28]. These results have strongly supported its role as a porin in the chemolithotrophic acidophilic bacterium [28]. However, little detailed information is available about the molecular mechanism by which At. ferrooxidans responds and adapts to external environmental changes. In this report, we examined effects of NaCl concentrations on the composition of outer membrane proteins, and found that the amount of two proteins increased in response to increasing concentration of NaCl. The expression of one of the proteins was examined by Northern blot hybridization analysis. 2.

MATERIALS AND METHODS

2.1 Bacterial strain and growth conditions The iron-oxidizing bacterium used in this study was At. ferrooxidans strain NASF-1. Cells were grown at 30°C under aerobic condition in Fe2+-medium as described previously [29]. 2.2 Preparation of outer membrane proteins Outer membrane proteins from NASF-1 cells were prepared according to the method of Silva et al. [27], although a slight modification was done. The cells were harvested in the mid- to late-exponential phase by centrifugation (15,000×g for 15 min at 4°C). The cell pellet was washed three times with 0.1M β-alanine-SO42- buffer (pH 3.0), two times with 20 mM 2-[4-(2-Hydroxyethyl)-1-piperazinyl] ethanesulfonic acid (HEPES) buffer (pH 8.0), and suspended in 20 mM HEPES buffer (pH 8.0). The cell suspension was sonicated (three times for 1 min). The lysate was centrifuged at 15,000×g for 10 min to remove cellular debris. The supernatant was centrifuged at 105,000×g for 1 h at 4°C. The precipitate was washed with 20 mM HEPES buffer (pH 8.0), resuspended in 20 mM 1262

Molecular Biology and Taxonomy

HEPES buffer containing 1% (w/v) sodium N-laurylsarcosine (Sarkosyl), and incubated for 1 h at 30°C. The suspension was centrifuged at 105,000×g for 1 h at 4°C to pellet the detergent-insoluble outer membrane fraction. The precipitate was washed with 20 mM HEPES buffer (pH 8.0) and used as an outer membrane fraction. 2.3 Protein analysis and N-terminal amino acid sequencing Protein concentrations were determined by Lowry method with crystalline bovine serum albumin as a reference protein [30]. Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulphate (SDS-PAGE) was performed in 12.5% (v/v) polyacrylamide slab gel with the Tris-glysine buffer. Outer membrane proteins separated by SDS-PAGE were electroblotted to a PVDF membrane (Hybond-P, Amersham Biosciences) using a blotting apparatus (Trans-Blot Cell system, Bio-Rad, U.S.A) according to the manufacture’s recommendations. N-terminal amino acid sequences of outer membrane proteins were determined by Edman analysis using an automatic protein sequencer (Model 610A NH2-terminal sequencer, Perkin-Elmer Corporation, U.S.A). 2.4 DNA manipulations The genomic DNA (gDNA) from NASF-1 cells was prepared by phenol/chloroform/ isoamylalcohol after lysis by a solution containing 20 mM Tris-HCl (pH 8.0), 20 mM EDTA and 0.4% sodium dodecyl sulfate. The DNA was used as a template for PCR reaction to amplify the Omp40 gene, that is an outer membrane protein previously detected in At. ferrooxidans [28]. Primers used for a PCR-amplification of Omp40 gene were constructed by using a sequence reported by Guiliani and Jerez [28]. Taq polymerase from Takara was used according to the manufacture’s recommendations. The PCR reaction was as follows: 3 min at 95°C, followed by 25 cycles at 95°C for 25s, 55°C for 30s, and 72°C for 45s, and then 3 min at 72°C. After the electrophoresis of PCR-amplified DNA fragments, the DNA was purified with Geneclean Kit (Q BIOgene) and directly sequenced with Thermo Sequenase Fluorescent Labelled Primer Cycle Sequencing Kit (Amersham Biosciences) and an automated sequence analyzer (Model DSQ-1000L; Shumadzu Co.). The PCR product purified from gel with Geneclean Kit was labeled with digoxigenin by using DNA Labeling and Detection Kit (Roche) according to the manufacture’s recommendations, and used as a probe in Southern and Northern blot hybridization experiments. Restriction enzyme digestions were performed according to the manufacture’s recommendations. Southern blotting was performed with total DNA digested with different restriction enzymes. After an electrophoresis, the DNA was denatured and transferred to a positively charged nylon membrane (Zeta-Probe, Bio-Rad) using a TransBlot Cell system. Prehybridization and hybridization with a DIG-labeled prove were performed under stringent conditions according to the manufacture’s recommendations (Roche). DNA was detected with the colorimetric reactions by using DNA labeling and Detection Kit (Roche) according to the manufacture’s recommendations. 2.5 RNA manipulations Total RNA of strain NASF-1 cells was extracted by using RNeasy Mini Kit (Qiagen) according to the manufacture’s recommendations. After the electrophoresis of RNA on formaldehyde agarose gel, RNA was transferred to a positively charged nylon membrane (Hybond-N+, Amersham Biosciences) using a Trans-Blot Cell system. The DIG-labeled probe described above was used for the detection of specific mRNA. Prehybridizition and hybridization with DIG-labeled probe were performed under stringent conditions 1263

Molecular Biology and Taxonomy

according to the manufacture’s recommendations (Roche). RNA hybridized with the probe was detected as described above in Southern blot hybridization experiment. 2.6 Database analysis Preliminary sequence data for genes of 30kDa and 40kDa proteins detected in this report was obtained from The Institute for Genomic Research website at http:// www.tigr.org. 3.

RESULTS

3.1 Effect of salts on the composition of outer membrane protein Strain NASF-1 cells were grown in Fe2+-media supplemented with different concentrations of NaCl or Na2SO4. The growth was observed in Fe2+-media supplemented with NaCl at concentrations up to 0.3 M. Strain NASF-1 cells could grow in Fe2+-media supplemented with Na2SO4 at concentrations up to 0.5 M. Outer membrane fractions were prepared from NASF-1 cells grown in Fe2+-media supplemented with NaCl or Na2SO4 at concentration up to 0.3 M and analyzed by SDS-PAGE. Outer membrane proteins with apparent molecular masses of 30 kDa and 40 kDa increased in cells grown in Fe2+medium supplemented with NaCl (Fig. 1A). The increases were not observed in the outer membrane fraction prepared from cells grown in Fe2+-medium supplemented with Na2SO4 (Fig. 1B). Proteins with molecular mass of 30 kDa and 40 kDa were designated as FopA (Acidithiobacillus ferrooxidans outer membrane protein A) and Fop40, respectively, in this report. The composition of outer membrane proteins may be influenced by growth phases. Therefore, outer membrane fractions prepared from cell grown in Fe2+-medium in different growth phases were analyzed by SDS-PAGE. The compositions of outer membrane proteins did not change in cells grown in log phase (4 days-culture), stationary phase (7 days-culture), and late stationary phase (14 days-culture) (data not shown). These results indicated that the increases of Fop40 and FopA proteins were due to response of cells to the increasing concentration of NaCl.

Figure 1. Composition of outer membrane proteins prepared from At. ferrooxidans NASF-1 cells grown in various concentration of salts. A; Outer membrane fractions were prepared from cells grown in Fe2+-medium supplemented with 0 M (lane 1), 0.1 M (lane 2), 0.2 M (lane 3) NaCl, and analyzed by SDS-PAGE. Lane M corresponds to MW marker proteins. Numbers to the left indicate molecular masses in kDa. The gel was stained with Coomassie blue. B; Outer membrane fractions were prepared from cells grown in Fe2+-medium supplemented with 0 M (lane1), 0.1 M (lane 2), 0.2 M (lane 3), 0.3M (lane 4) Na2SO4. The gel was stained with silver 1264

Molecular Biology and Taxonomy

3.2 N-terminal amino acid sequences of FopA and Fop40 protein The N-terminal amino acid sequences of FopA and Fop40 proteins were determined to be DGGYVGYAVNHGAKPVVTSR and ADTSNANTGPVVFGYAQI, respectively. Although the expressions of FopA and Fop40 proteins were stimulated in response to the increasing concentration of NaCl in medium, no homology was observed in the Nterminal amino acid sequences between FopA and Fop40. Computer searches of available databases revealed that N-terminal amino acid sequence of FopA protein had no significant homology with any other known prokaryotic protein. The N-terminal amino acid sequence of Fop40 protein was almost the same as the outer membrane protein (Omp40) reported previously as an outer membrane protein influenced in At. ferrooxidans ATCC 18959 under phosphate starvation [24], except in one amino acid. The nucleotide sequence had been already analyzed by Guiliani and Jerez [28]. The determination of whole genome sequence of At. ferrooxidans ATCC 23270 is in progress now, and the sequence data can be available (http://www.tigr.org). A search using the N-terminal amino acid sequence of the Fop40 protein in TIGR pre-released genomic data using the Blast algorithm revealed only one reading frame encoding Omp40 protein. Therefore Fop40 protein was the product of the gene encoded Omp40 protein. On the other hand, a homology search to the N-terminal amino acid sequence of the FopA protein in TIGR prereleased genomic data also revealed only one open reading frame (651 bp) encoding the same N-terminal amino acid sequence. As expected for an outer membrane protein, the gene contained a signal peptide sequence corresponding to 27 amino acids as shown in Fig. 2. The deduced mature protein had 190 amino acids and molecular mass of 20,210 Da. The molecular mass deduced from the gene was smaller than the apparent molecular mass of FopA protein determined by SDS-PAGE. Our BLASTP search of the SwissProt database at the National Cancer for Biotechnology Information Web site identified 30 proteins homologous to FopA protein with scores exceeding 70 and E value of < 2e-11. Members of the OmpA family belong to this cluster. A protein with the highest score was the outer membrane protein of Fusobacterium nucleatum [31]. However, the N-terminal region of the putative FopA gene product was shorter than that of typical proteins of the OmpA family, such as OmpA of E. coli, OprF of Pseudomonas spp., and MopB of Methylococcus [32]. Peptidoglican-associated lipoprotein (Pal) also showed a low homology to FopA protein. 3.3 Amplification of Fop40 gene by PCR The N-terminal amino acid sequence of Fop40 protein from NASF-1 cell had almost the same sequence as the Omp40 protein previously reported in response to phosphate starvation in At. ferrooxidans ATCC 18959 [24]. As the gene encoding Omp40 protein has already been sequenced [28], primers were designed to amplify Fop40 gene. The PCRamplified product had an expected length (447 bp) of Omp40 gene on agarose gel. Therefore, the PCR-amplified product was purified, labeled with DIG, and used as a probe for hybridization experiments. 3.4 Specificity of Fop40-probe by Southern hybridization analysis A southern hybridization analysis was carried out to examine the specificity of the DIG-labeled probe. The hybridizations were carried out with the genomic DNA (gDNA) of strain NASF-1, the PCR product of Fop40 gene and the gDNA fragments digested with Sal I, Sac I, or Sma I. Only one hybridization signal was observed in each gDNA digested with the different endonuclease. These results indicated that only one Fop40 gene was present in the genome of NASF-1 cell. 1265

Molecular Biology and Taxonomy

Figure 2. Nucleotide sequence and derived amino acid sequence of FopA gene detected in TIGR pre-released database of At. ferrooxidans ATCC 23270 genome. The possible –10 and –35 regions are underlined. The putative ribosome-binding site is indicated in boldface. The signal sequence is underlined in bold. A putative transcription terminator (underlined with arrow heads) is shown after the coding sequence 3.5 Expression of mRNA of Fop40 protein in NASF-1 cells The expression of Fop40 protein was stimulated in response to the increasing concentration of NaCl. The expression of mRNA was examined to clarify whether the increase of Fop40 protein was due to transcriptional activation or translational activation. The analysis of Fop40 gene obtained from TIGR pre-released genomic data of At. ferrooxidans ATCC 23270 revealed that the open reading frame was preceded by a plausible ribosome-binding site with a AGGA sequence and –10 and –35 promoter sequences. A stem-loop structure followed by a T-rich sequence was found downstream from the stopping UAA codon, representing an independent transcriptional terminator (data not shown). Therefore, the inferred length of transcribed mRNA is though to be about 1.3 kb. One hybridization signal having the expected length was observed by Northern blot hybridization. The expression of Fop40-mRNA was stimulated in cells grown in Fe2+-medium supplemented with 0.2 M NaCl as shown in Fig. 3A. To determine the induction period for the expression of Fop40-mRNA, RNAs were prepared from cells incubated in Fe2+-medium supplemented with 0.2 M NaCl for 0, 1, 3 or 5 h and analyzed by Northern blot hybridization. A relative strong hybridization signal was observed after the incubation for 5 hours as shown in Fig. 3B. The expressions of proteins of At. ferrooxidans have been reported to be influenced in the external medium pH [20]. Therefore, the effect of pH on the expression of Fop40-mRNA was examined. RNAs were prepared from cells incubated for 5 h in Fe2+-medium adjusted at pH 1.5, 2.5, 3.5, or 4.5, and analyzed. The expression of Fop40-mRNA was stimulated when cells were incubated at pH higher than 2.5 as shown in Fig. 3C. Although the expression of Fop40-mRNA was stimulated in cells grown in Fe2+-medium supplemented with NaCl, SDS-PAGE analysis revealed that the stimulation did not occur in cells grown in Fe2+-medium supplemented with Na2SO4. Therefore, the effect of salts on the expression of Fop40-mRNA was examined. mRNAs were prepared from cells incubated for 5 h in Fe2+-medium supplemented with 0.2 M NaCl, 0.2 M KCl or 0.1 M Na2SO4, and analyzed by Northern blot hybridization. A strong hybridization signal was observed only with mRNA prepared from cells grown in Fe2+-medium supplemented with 0.2 M NaCl as shown in Fig. 3D.

1266

Molecular Biology and Taxonomy

Figure 3. Effects of NaCl (A), incubation periods (B), pH (C) and salts (D) on Fop40 gene expression analyzed by northern hybridization. A; RNAs were prepared from cells grown without (lane 1) or with (lane 2) 0.2 M NaCl. B; RNAs were prepared from cells incubated in Fe2+-medium supplemented with 0.2M NaCl for 0 (lane 1), 1 (lane 2), 3 (lane 3), or 5 h (lane 4). C; RNAs were prepared from cells incubated for 5 h in Fe2+-medium adjusted at pH 4.5 (lane 1), 3.5 (lane 2), 2.5 (lane 3), or 1.5 (lane 4). D; RNAs were prepared from cells incubated in Fe2+-medium supplemented with 0.2 M KCl (lane 1), 0.1 M Na2SO4 (lane 2), 0 M NaCl (lane 3), or 0.2 M NaCl (lane 4). Northern hybridizations were carried out with DIG-labeled Fop40 probe. Lower photograph is an ethidium bromide-stained gel, indicating equal loadings of rRNA 4.

DISCUSSION The results obtained by SDS-PAGE analysis of outer membrane fractions prepared from cells grown in Fe2+-medium supplemented with NaCl revealed the increase of two specific proteins, FopA and Fop40 proteins. The N-terminal amino acid sequence of Fop40 protein was almost the same as the membrane protein previously detected in At. ferrooxidans ATCC 18959 under phosphate starvation. Northern blot hybridization analyses using DIG-labeled PCR-product of Fop40 gene as a probe revealed that the expression of mRNA of Fop40 protein was stimulated when cells were exposed to NaCl, or pH 3-4. Although the synthesis of a protein having an apparent molecular mass of 36 kDa has been reported to increase when At. ferrooxidans cells grown at pH 1.5 were shifted to pH 3.5, the synthesis of a major membrane protein with a molecular mass of 40 kDa (Omp40) has not been significantly influenced with pH shift [20]. The result is inconsistent with the data obtained with NASF-1 cells. The reason for this contradiction is unclear. When the cells were incubated in Fe2+-medium supplemented with NaCl, the expression of Fop40-mRNA was stimulated after the incubation for 5 h. This long stimulation period for Fop40-mRNA transcription may be due to the long generation time (about 8 h) of this bacterium. The stimulation did not occur in Fe2+-medium supplemented with KCl or Na2SO4. The results was consistent with the data obtained by SDS-PAGE analysis shown in Fig. 2B, and may indicate that the increase of Fop40 is not due to an osmotic change in medium.

1267

Molecular Biology and Taxonomy

Figure 4. Alignment of the amino acid sequence derived from the putative FopA gene from At. ferrooxidans ATCC 23270 with sequences of the homologous outer membrane proteins. OrpF; OmpA-like protein of P. fluorescens (AF117969), OmpA; outer membrame protein of E. coli (V00307), Fomp; outer membrane protein of F. nucleatum (N_003454), Pal; peptidoglycan-associated lipoprotein of E. coli (X05123). Residues asterisked under the sequences are conserved in all sequence. Residues dotted are conserved in OmpA-related proteins. The linker region between Nterminal and C-terminal region of OmpA is in italic. An underlined part indicated the peptidoglycan-binding domain of OmpA The stimulation of Fop40 expression in At. ferrooxidans NASF-1 cells occurred with different stimuli, such as NaCl concentration and pH shift. Although it has been reported that Omp40 protein is a porin and has a pore-forming activity [28, 27], homologous proteins to Omp40 protein from At. ferrooxidans have not been detected in databases. In E. coli, different porins, OmpC or OmpF, function when cells are exposed to osmotic change, and the pH dependence of OmpC and OmpF expression is also well known [33, 34]. E. coli involves EnvZ and OmpR functioning as a sensor of osmotic change and as regulator, respectively. The homologous gene to EnvZ or OmpR could not be detected in the pre-released database of At. ferrooxidans ATCC 23270 genome. Therefore, the regulatory mechanism of Fop40 expression may be different from that of OmpC or OmpF expression in E. coli. The investigation of regulatory mechanism for the expression of Fop40 protein is very important to understand the mechanism of environmental adaptation of this acidophilic bacterium, as pointed previously [28]. On the other hand, the expression of FopA protein was also stimulated in NASF-1 cells grown in Fe2+-medium supplemented with NaCl. Although the deduced molecular mass of the gene product detected in the pre-released database of At. ferrooxidans ATCC 23270 genome was smaller than the apparent molecular mass of FopA protein estimated by SDS-PAGE, we could not find out any other homologous genes in the database. Therefore, the open reading frame detected in the database seems to be the gene encoding the FopA protein. The C–terminal region of FopA had strong sequence similarity with 1268

Molecular Biology and Taxonomy

proteins of the OmpA family, although no homologous protein with the N-terminal region of FopA was observed in the database. The N-terminal region of FopA was shorter than that of other typical proteins in the OmpA family, such as OmpA, and OprF (Fig. 4). OmpA-related proteins from other bacteria have a function needed to maintain the structure of cell by interacting with peptideglycan [35-37]. FopA also seems to associate with peptideglycan. N-terminal domains of OmpA-related proteins have shown to cross the membrane eight times in antiparallel β-strand [38]. We cannot find at least 3 sequences capable to form β-strand in FopA. FopA contained two hydrophobic parts in the Nterminal region, although Pal does not contain any hydrophobic parts in the N-terminal region. The linker region observed in OmpA of E. coli is also conserved in FopA. Therefore, we concluded that FopA is a new OmpA-like protein associating with peptidoglycan. We can find many proteins having similar structure as FopA protein in databases. The functions of these OmpA-like proteins having a short N-terminal region have not been examined in detail, yet. Some OmpA-related proteins have been known as a heat-modifiable proteins, which changes the mobility on SDS-PAGE due to the heat-induced conformational change. The difference between the apparent molecular mass estimated by SDS-PAGE and the molecular mass deduced from the putative gene of FopA may be due to the heat modifiability of FopA protein. The purification and analysis of FopA protein is now in progress to characterize the properties of FopA protein in detail. ACKNOWLEDGMENTS We thank Hidenori Yamada (Graduate School of Natural Science and Technology, Okayama University) for the N-terminal sequencing of FopA and Fop40 proteins. Preliminary sequence data for the At. ferrooxidans strain 23270 was obtained from The Institute for Genomic Research (http://www.tigr.org). This work was supported in part by a grant (No.12876022) from The Ministry of Education, Culture, Sports, Science and Technology. REFERENCES 1. W. Achouak, T. Heulin, and J. M. Pages, FEMS Microbiol. Lett., 199 (2001)1. 2. R. Benz, and K. Bauer, Eur. J. Biochem., 176 (1988) 1. 3. B. K. Jap, and P. J. Walian, Rev. Biophys., 23 (1990) 367. 4. B. K. Jap, and P. J. Walian, Physiol. Rev., 76 (1996) 1073. 5. H. Nikaido, and M. Vaara, Microbiol. Rev., 49 (1985) 1. 6. T. Mizuno, M. –Y. Chou, and M. Inouye, J. Biol. Chem., 258 (1983) 6932. 7. H. Hofstra, and J. Dankert, J. Gen. Microbiol., 119 (1980) 123. 8. R. Benz, A. Schmid, and R. E. W. Hancock, J. Bacteriol., 162 (1985) 722. 9. H. Nikaido, and E. Y. Rosenberg, J. Bacteriol., 153 (1983) 241. 10. D. B. Datta, B. Arden, and U. Henning, J. Bacteriol., 131(1977) 821. 11. L. A. Pratt, W. Hsing, K. E. Gibson, and T. J. Silhavy, Mol. Microbial., 20 (1996) 911. 12. H. Aiba, T. Mizuno, and S. Mizushima, J. Biol. Chem., 264 (1989) 8563. 13. S. Forst, J. Delgado, and M. Inouye, Proc. Natl. Acad. Sci. USA., 86 (1989) 6052. 14. H. Aiba, H., F. Nakasai, S. Mizushima, and T. Mizuno, J. Bacteriol., 106 (1989) 5. 15. C. L. Brierley, Crit. Rev. Microbiol., 6 (1978) 207. 16. J. G. Cobley, and J. C. Cox, Microbial. Rev., 47 (1983) 579. 17. A. P. Harrison, Jr. Annu. Rev. Microbiol., 38 (1984) 265. 18. W. J. Ingledew, Biochem. Biophys. Acta., 683 (1982) 89. 1269

Molecular Biology and Taxonomy

19. O. H. Tuovinen and D. P. Kelly. Z. Allg. Mikrobial. 12 (1972) 311. 20. A. M. Amaro, C. Chamorro, R. Arredondo, I. Peirano, and C. A. Jerez, J. Bacteriol., 173 (1991) 910. 21. V. Buonfiglio, M. Polidoro, L. Flora, G. Citro, P. Valenti, and N. Orsi, FEMS Microbiol. Rev., 11 (1993) 43. 22. V. Buonfiglio, V., M. Polodoro, F. Soyer, P. Valenti, and J. Shively, J. Biotechnol., 72 (1999) 85. 23. N. Ohomura, K. Tsugita, J. –I. Koizumi, and H. Saiki, J. Bacteriol., 178 (1996) 5776. 24. A. C. Jerez, M. Seeger, and A. M. Amaro, FEMS Microbiol. Lett., 98 (1992) 29. 25. M. Seeger, and C. A. Jerez, FEMS Microbiol. Lett., 108 (1993) 35. 26. M. Rodriguz, S. Campos, and B. Gomz-Silva, Appl. Biochem., 8 (1986) 292. 27. M. Silva, A. Ferreira, M. Rodriguez, and D. Wolff. FEBS Lett., 296 (1992) 169. 28. N. Guiliani, C. A. Jerez, Appl. Environ. Microbiol., 66 (2000) 2318. 29. K. Kamimura, S. Fujii and T. Sugio, Biosci. Biotechnol. Biochem., 65 (2001) 63. 30. O. H. Lowry, N. J. Rosenbrough, A. L. Farr, and R. J. Randall, J. Biol. Chem., 193 (1951) 265. 31. V. Kapatral, I. Anderson, N. Ivanova, G. Reznik, T. Los, A. Lykidis, A. Bhattacharyya, A. Bartman, W. Gardner, G. Grechkin, L. Zhu, O. Vasieva, L. Chu, Y. Kogan, E. Goltsman, A. Bernal, N. Larsen, M. D’Souza, T. Walunas, G. Pusch, R. Haselkorn, M. Fonstein, N. Kyrpides, and R. Overbeek, J. Bacteriol., 184 (2002) 20052018. 32. A. Fjellbirkeland, V. Bemanian, I. R. McDonald, J. C. Murrell, and H. B. Jensen, Arch. Microbiol., 173 (2000) 346. 33. M. Heyde, and R. Portalier, Mol. Gen. Genet., 208 (1987) 511-517. 34. M. Sato, K. Machida, E. Arikado, H. Saito, T. Kakegawa, and H. Kobayashi, Appl. Environ. Microbiol., 66 (2000) 943-947. 35. R. Domot, and J. Vanderleyden, Mol. Microbiol., 12 (1994) 333. 36. L. Sonntag, H. Schwartz, Y. Hirota, and U. Henning, J. Bacteriol., 136 (1978) 280. 37. E. Sugawara, and H. Nikaido, J. Biol. Chem., 269 (1994) 17981. 38. A.Pautsch, G.E. Schulz, Nat. Struct. Biol. 5 (1998) 1013-1017.

1270

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Bioinformatic analysis of biofilm formation in Acidithiobacillus ferrooxidans M. Barreto1, M. Rivas2, D.S. Holmes1,3 and E. Jedlicki2* 1

Laboratory of Bioinformatics and Genome Biology, University of Santiago (USACH), Santiago, Chile 2 Program of Cellular & Molecular Biology, ICBM, University of Chile, Santiago, Chile 3 Millenium Institute of Fundamental and Applied Biology, Santiago, Chile Abstract The role of biofilm formation in the growth of Acidithiobacillus ferrooxidans in natural environments and on simulated laboratory mineral surfaces has been well documented. However, despite the fundamental and industrial interest of such biofilm formation, little has been done to investigate its underlying genetic and physiological basis in A. ferrooxidans. Using the almost complete genome sequence of A. ferrooxidans made available by The Institute for Genome Research (TIGR) and Integrated Genomics Inc. (IG), we have undertaken a preliminary evaluation of possible genes and pathways potentially involved in biofilm formation. A. ferrooxidans appears to have a substantial repertoire of genes necessary to synthesize the polysaccharide building blocks of biofilms. It also has genes to polymerize these building blocks into complex polysaccharides on a membrane associated lipid anchor. In addition, it has genes to form this lipid anchor and also genes to export and mature the extracellular polysaccharides that are the major constituent of most biofilms. Using this information, a model is proposed for the biofilm formation in A. ferrooxidans. Future studies will seek to provide experimental evidence for the model. Keywords: biofilm formation, Acidithiobacillus ferrooxidans, genome analysis, extracellular polysaccharides, galactose 1.

INTRODUCTION The formation of biofilms on mineral surfaces and their probable role in mineral dissolution has been an area of study not only for fundamental interest but also because of its relevance to the industrial activity of this microorganism. However, little has been established regarding the underlying genetics and physiology of biofilm formation.

*

Corresponding author: Eugenia Jedlicki, [email protected]. Work supported by Fondecyt No. 1010623 and the Millenium Institute of Fundamental and Applied Biology, Santiago, Chile. We thank the Institute of Genome Research (TIGR) and Integrated Genomics, Inc. (IG) for the use of their partial sequence data of the Acidithiobacillus genome.

1271

Molecular Biology and Taxonomy

A biofilm is a highly structured community of organisms typically enclosed in an extra-cellular matrix and separated from neighboring communities by water channels (1). A structured community can include regional differentiation of function, for example aerobically respiring bacteria near the outside with anaerobes inside, or bacteria with different tolerances to pH distributed in a gradient from the outside to the inside of the biofilm. Usually, regional differentiation utilizes complex inter-cellular signaling for its development and maintenance. A typical biofilm is illustrated in figure 1. This figure is a composite, constructed from an analysis of a compilation of confocal microscope images obtained using totally hydrated biofilms derived from a number of different locations such as mountain rivers and acid mine drainage (2). One of the first and best characterized step in the formation of a biofilm is the event in which bacteria pass from a reversibly attached stage to one in which they are irreversible bound to their substrate. Reversible attachment includes substrate identification that sometimes, but not always, involves chemotaxis, followed by electrostatic interactions between the bacterial cell wall and the substrate. The switchover to irreversible attachment involves the production and excretion of extracellular polysaccharides or, more accurately, extracellular polymorphic substance (EPSs) (3). EPS is a term that refers to a diverse set of biopolymers that can contain substituted or non-substituted polysaccharides and substituted or non-substituted proteins and may include nucleic acids and phospholipids (4). Several studies have shown that attachment and adherence of A. ferrooxidans to mineral surfaces can occur and that the latter process is accompanied by the production of EPS (5-10). EPS production has an important role in the bacterial-substratum interactions and subsequent biofilm formation (8). In the environment, it is most likely that A. ferrooxidans forms a part of natural biofilms that cover exposed rock and mineral surfaces.

Figure 1. Formation and maturation of a typical bacterial biofilm. (A) Initial adhesion of a cell to a charged (often positively charged) abiotic or biological surface. (B) Formation of a monolayer of cells. (C) Development of strong inter-cellular contacts and formation of microcolonies. (D) Differentiation of a mature biofilm within a matrix of exopolisaccharides (EPS) separated by water channels A. ferrooxidans can also exist in the planktonic state and probable colonizes new substrates while in this state. In the process of bioleaching, especially in the case of dump leaching, solubilization of the mineral involves attachment of various bacteria, including A. ferrooxidans, to the mineral substrate followed by biofilm formation. Attachment of bacteria to the mineral substrate probably also occurs during heap leaching, but the extent of attachment and subsequent biofilm formation would depend on the length of time during which the heap bioleaching process is allowed to occur. 1272

Molecular Biology and Taxonomy

Whereas the capacity of A. ferrooxidans to form biofilms has been well established, little is known about the underlying physiology, genetics and regulation of biofilm formation in this microorganism. The long-term objective of our work is to address this deficiency and in this paper we present preliminary evidence for the presence of genes potentially involved in the formation of EPS and we propose a working model for EPS formation. This information provides a first step for understanding the physiology of biofilm formation in A. ferrooxidans. 2.

MATERIALS AND METHODS. Known metabolic pathways involved in the formation of galactosides were obtained from BIOCYC (http://biocyc.org:1555/META/server.html), KEGG (http://genome.ad.jp/ kegg/kegg4.html) and ERGO (http://wit.integratedgenomics.com/WIT2/CGI). Amino acid sequences derived from genes identified as being involved in galactose utilization were used as query sequences to search the partial genome sequence of A. ferrooxidans ATCC 23270 in the TIGR (http:// www.tigr.org/) and ERGO data bases using TBlastN and BlastP respectively. When a prospective candidate gene was identified in TIGR or ERGO its predicted amino acid sequence was then used to formulate a BlastP (http://www.ncbi.nlm.nhi.gov/BlastP/) search of the non-redundant database at NCBI. Only bidirectional best hits were accepted as evidence for putative homologs. Candidate genes and their translated proteins were further characterized employing the following bioinformatic tools available in the web: primary structure similarity relations (http://www.ebi.c.uk/ClustalW/), secondary structure predictions (HMM-based Protein Sequence Analysis http://www.cse.ucsc.edu/research/ compbio/HMM-apps/T99query.html; JPred http://www.compbio.dundee.ac.uk/Software/ JPred/jpred.html), transmembrane predictions (http://www.ch.embnet.org/software/ TMPRED_form.html), motif predictions (http://www.blocks.fhcrc.org/, http://www.ebi.ac.uk/interpro/, http://www.biochem.ucl.ac.uk/bsm/dbbrowser/PRINTS/printscontents.html/, http://www. sanger.ac.uk/Software/Pfam/) and prediction of protein localization sites (http://psort.nibb. ac.jp/). 3.

RESULTS AND DISCUSSION Our search for possible genes in A. ferrooxidans, involved in EPS, started with the assumption that such genes would be recognizable orthologs of genes in other organisms known to be involved in biofilm formation. This assumption is justified because of the known conservation of EPS formation genes among various bacterial species (11), although subsequent metabolic routes to mature biofilm formation are varied and not well understood (12). The basic building blocks of the EPS are typically the two galactosides, UDP-glucose and UDP-galactose (13). The enzymes involved in their production are UDP-glucosepyrophosphorylase, encoded by the gene galU and UDP-glucose-4-epimerase, encoded by galE. In addition, in lactic acid bacteria, the enzyme phosphoglucomutase, encoded by the gene pgm, is also involved (14). Recognizable orthologs of these genes were found in A. ferrooxidans (Figure 2A, Figure 3 and Table1). Having established that A. ferrooxidans has the necessary genetic capacity to synthesize the galactoside building blocks of EPS, additional bioinformatic analysis revealed the presence of a suite of genes potentially involved in the polymerization of the building blocks into EPS and its resulting processing and exportation through the outer membrane (Figure 2B,C and D, Figure 3 and Table 1). These include genes for the 1273

Molecular Biology and Taxonomy

synthesis of the glycosyltransferases (GTFs) that are responsible for the polymerization of the galactosides into EPS on a membrane associated lipid anchor (Figure 2B, Figure 3), the export of the EPS through the membrane (Figure 2C, Figure 3) and its attachment to the outer surface of the outer membrane (Figure 2D, Figure 3). In addition, genes have been identified that potentially encode functions related to the biosynthesis of the membrane associated lipid anchor transporter, the modification of EPS and the construction of the outer membrane exporter of the EPS (Table 1, Figure 3). The identification of a suite of candidate genes in A. ferrooxidans that potentially encode functions related to EPS formation was based upon bioinformatic amino acid sequence similarity comparisons made with genes experimentally implicated in EPSformation. Corroborating evidence for these functional assignments comes from additional bioinformatic analyses that reveal structural and functional motifs and domains in a number of these candidate proteins characteristic of genes involved in EPS formation (data not shown). An analysis of the genomic locations of the candidate genes in A. ferrooxidans suggests that a number of them are located in operon-like organizations typical of those found in other microorganisms involved in EPS formation. Three examples of such putative operons are illustrated in Figure 3. The case illustrated in Figure 3A presents the organization of a proposed operon that includes genes potentially encoding galE (synthesis of galactosides) wza, ywqE, mir, exoT and alr (polymerization of galactosides) and exoP (export EPS). This proposed operon has similarities in gene content to that involved in EPS formation in the nitrogen fixing bacterium Rhizobium melliloti (15). The proposed operon shown in Figure 3B includes lpxB (polymerization of galactosides) and nmb, cdsA, dxr, lpxD, fabZ and lpxA (modification of EPS). It also includes three predicted membrane proteins of unknown function. This proposed operon has similarities of gene content to operons found in Lactococcus lactis, Streptococcus thermophilus and E. coli (14, 16 and 17).

Figure 2. Proposed model for the biosynthesis and excretion of exopolysaccharides in A. ferrooxidans potentially capable of leading to the formation of a biofilm. This model is based on a similar model for the formation of EPS experimentally established in a wide variety of microorganisms (Boels et al 2001). An explanation of the steps A-D in EPS formation is provided in the text

1274

Molecular Biology and Taxonomy

Table 1. Candidate genes involved in EPS biosynthesis in A. ferrooxidans. (A) Proposed function ordered in a way consistent with the pathway shown in Figure 2, (B) gene name, (C) proposed enzyme activity, (D) Organism with the best BlastP hit to the candidate gene and (E) the % similarity of candidate gene to that found in the organism listed in column D A. Proposed function Galactoside synthesis

B. Gene galE pgm galK galU

Polymerization of EPS

exoT ywqE mlr alr lpxB wza

Maturation and modification of EPS

nmb lpxD

fabZ lpxA cdsA

Exportation

dxr exoP

C. Proposed enzyme activity

D. Best Blastp hit

E. % Sim.

UDP-glucose epimerase Phosphoglucomutase Galactokinase UDP-glucose pyrophosphorylase Polysaccharide biosynthesis protein Capsular polisacharide biosynthesis protein UDP-glucose/GDP-mannose dehydrogenase Glycosyltransferase (GTF) Lipid A-disaccharide synthase capsular polysaccharide transport Lipid carrier synthetase Probable UDP-3-o-3hydroxymyristoyl glucosamine n-acyltransferase protein Probable 3R-hydroxymyristoylacyl carrier protein dehydratase UDP-N-acetylglucosamine acetyltransferase Phosphatidate cytidilyltransferase Xylulose 5-phosphatase Exportation of EPS

M. thermautotrophicus B. melitensis S. coelicolor B. pseudomallei

71% 76% 48% 74%

G. xylinus

56%

B. subtilis

44%

P. aeruginosa

73%

Nostoc sp. P. aeruginosa V. vulnificus

49% 50% 48%

N. meningitidis R. solanacearum

64% 60%

R. solanacearum

58%

E. coli K12

69%

P.aeruginosa PA01

53%

R. solanacearum E. coli O157

68% 79%

The marked similarity of the organization of the proposed EPS formation operons of A. ferrooxidans with those found in a variety of microorganisms provides additional supporting evidence for the assignations of gene functions listed in Table 1. It also suggests that there may be an underlying commonality of gene regulatory networks controlling the expression of genes involved in EPS formation. To substantiate this conjecture, we tried to compare, by bioinformatic analysis, known DNA regulatory components of established EPS operons with the putative operons of A. ferrooxidans. Unfortunately, at present, such bioinformatic approaches are generally hampered by the typical shortness of regulatory DNA sequences. However, using methods that have revealed potential Fur binding sites in A. ferrooxidans (18), a possible catabolite activator (CAP) binding site was detected upstream of the proposed gal operon (Figure 3C). A CAP binding site has also been mapped in the galactose operon of E. coli where the CAP protein has been shown to serve as a transcriptional activator of the operon (19). The similarity of the 1275

Molecular Biology and Taxonomy

organization of the proposed A. ferrooxidans and E. coli gal operons suggests that they may perform a similar function with, possibly, a closely related mechanism of regulation. The model that we have proposed for the formation of EPS in A. ferrooxidans can be considered a first step in understanding the genetics, physiology and regulation of biofilm formation in these microorganisms and now it is important to validate experimentally the model. In addition, future work will investigate the relationship of the proposed excreted EPS to later stages of biofilm formation.

Figure 3. (A), (B) and (C) examples of operon-like organizations of A. ferrooxidans genes proposed to be involved in EPS formation. In addition, in (C), the E. coli galactose operon and separate pgm gene is shown for comparison with the two proposed equivalent A. ferrooxidans gal operons Scale is shown in kb (kilobases). Each gene is coded according to its proposed function (see Key) consistent with the functions shown in Figure 2 and Table 1 REFERENCES 1. J. Costerton and P. Stewart, Scientific American., 285 (2001) 75 2. P. Stoodley, K. Sauer, D. Davies and J. Costerton, Annu. Rev. Microbiol., 56 (2002) 187 3. W. Characklis, in Biofilms, ed. W. Characklis, K. Marshall, New York: Wiley, (1990) 195. 4. J. Windenger, T. Neu and H. Flemming, ed. J. Windenger, T; Neu, T, Flemming, H. Berlin: Springer. (1999) 93 5. D. Karamanev, J. Biotechnol., 20 (1991) 51 6. N. Wakao, K. Endo, K. Mino, Y. Sakurai, H. Shiota, J. Gen. Appl. Microbiol., 40 (1994) 349 7. A. Schippers, T. Rohwerder and W. Sand, Appl. Microbiol. Biotechnol., 52 (1999) 104 8. T. Gehrke, J. Telegdi, D. Thierry and W. Sand, Appl. Environ. Microbiol., 64 (1998) 2743 1276

Molecular Biology and Taxonomy

9. C. Pogliani and E. Donati, Process Biochemistry, 35 (2000) 997 10. K. Kinzler, T. Gehrke and W. Sand, in Biohydrometallurgy: Fundamentals, Technology and Sustanable Development, V.S.T. Ciminelli, O. Garcia (eds.), Ouro Preto, Minas Gerais, Brazil (2001) 191. 11. C. Ingeborg, A. Ramos, M. Kleerebezem and W. De Vos, Appl. Environ. Microbiol., 67 (2001) 3033 12. J-M Ghigo, Research in Microbiology, (2003, in press) 13. K. Bettenbrock and C-A. Alpert, Appl. Environ. Microbiol., 64 (1998) 2013 14. B. Degeest and L. Vuyst, Appl. Environ. Microbiol., 66 (2000) 3519 15. M. Glucksmann, T. Reuber and G. Walker, J. Bacteriol., 175 (1993) 7043 16. V. Stout, J. Bacteriol., 178 (1996) 4273 17. P. Looijesteijn, I. Boels, M. Kleerebezem and J. Hugenholtz, Appl. Environ. Microbiol., 65 (1999) 5003 18. R. Quatrini, F. Veloso, E. Jedlicki, and D.S. Holmes, International Biohydrometallurgy Symposium, Greece (2003). This volume. 19. M. Weickert and S. Adhya, Mol. Microbiol., 10 (1993) 245

1277

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Diversity of Gram-negative bacteria at Malanjkhand copper mine, India S.R. Dave* and D.R. Tipre Department of Microbiology, School of Sciences, Gujarat University, Ahmedabad 380 009, Gujarat, India Abstract Samples collected from Malanjkhand open-pit copper mine, India showed major variation in pH from 3.3 to 8.0, redox potential from 250 to 505 mV and soluble copper from 0.13 to 0.63 g/l. Shannon-Wiener diversity indices (H') of heterotrophic cultivable bacterial species was in the range of 1.19 to 2.17 corresponding to evenness of 0.57 to 0.91 respectively. The major heterotrophic Gram-negative bacteria, which grew on sodium thiosulphate medium supplemented with glucose or yeast extract, were identified as Pseudomonas stutzeri, Pseudomonas aeruginosa, Brevundimonas diminuta, Stenotrophomonas maltophilia and Alcaligenes species. Collected samples also showed the presence of acid tolerant heterotrophs capable of growing on reduced sulphur compounds. Autotrophic sulphur and iron oxidizers were successfully cultivated with sulphur, tetrathionate, thiosulphate, metal sulphide and ferrous in liquid as well as solid media. They were identified on the basis of morphological, biochemical and physiological characteristics as Acidithiobacillus ferrooxidans, Acidithiobacillus thiooxidans, Thiobacillus thioparus, Thiobacillus versutus and Thiobacillus intermedius. The Malanjkhand open-pit copper mine habitats showed considerable diversity among mesophilic Gram-negative bacterial species. Presence of both heterotrophic and autotrophic iron oxidizing acidophilic bacteria was noted in various proportions at different sites of the mine. Addition of yeast extract in the medium proved to be the choice of organic material for overall ferrous oxidation by the enriched cultures. Keywords: mesophilic bacteria, Gram-negative bacteria, diversity, copper mine 1.

INTRODUCTION Microorganisms occupy important niches in all ecosystems and are responsible for the cycling of elements, degradation and formation of minerals even in the extreme environments [1-3]. The information about the microbial community structure at mining environment is necessary in order to gain a thorough understanding of the functioning of these ecosystems and their impact on the surrounding environment [4, 5]. The mining ecosystems are dominated by acidophilic sulphur and iron oxidising organisms. But due to sharp physical and chemical gradients, microbial ecosystems offer a variety of habitats and microniches, which can potentially be inhabited by metabolically diversed *

Corresponding author: S.R. Dave, E-mail: [email protected]

1279

Molecular Biology and Taxonomy

microorganisms. However, little is known about the distribution of such microbial population thriving in these environments [6]. Therefore, the interest in the biodiversity of the microorganisms, which inhabit such extreme environments, has increased significantly over the past few decades. The most familiar and well-studied microbes of acidic mineral leaching environments are Acidithiobacillus ferrooxidans, Acidithiobacillus thiooxidans and Leptospirillum ferrooxidans [5]. But various fungi and heterotrophic organisms can also accomplish metal dissolution from oxidised and sulphide minerals even at weakly acidic or alkaline pH [5, 7-10]. Therefore, the study regarding such microbial community at mining environments is essential. The cultivation based methods are not well suited for the fastidious microbial community study at the same time molecular studies also suffer from some drawbacks and it cannot provide the information regarding the physiological information of microorganisms. Thus, cultivation and molecular methods compliment each other [6]. In this context, the present work was undertaken to study the microbial diversity of mesophilic Gram negative bacteria of Malanjkhand Copper Mine, as India has numerous base metal ore reserves among which Malanjkhand Copper Mine is an important reserve situated in Madhya Pradesh. 2.

MATERIALS AND METHODS

2.1 Sample collection Seven soil and water samples were collected from Malanjkhand Copper Mine. This mine is located at 80° 43' longitude and 22° 2' latitude at Malanjkhand, Madhya Pradesh, India. It is an open pit operation producing 2 million tonnes per year of ore containing 1.2% grade of copper. The deposit is of Proterozoic age and consists of a large body of primary copper ore (chalcopyrite) in quartz veins and granite rocks. The secondary sulphides formed are covellite and chalcocite. The main ore mineral is chalcopyrite with minor sulphide minerals viz. pyrite, sphalerite, molybdenite, chalcocite and bornite. All the samples were collected in the month of October 2000 when the mean day temperature was 30±2°C. Samples were collected in sterilised containers and polythene bags with the help of sterile sampler. During sample collection, pH and oxidation-reduction potential at sites were recorded with portable instruments (model-Eutech Cybernetics). Samples were brought immediately to the laboratory and stored at 4°C temperature till analysed. Dissolved copper was estimated by standard methods [11]. 2.2 Isolation and enumeration Neutrophilic heterotrophic bacterial diversity study was carried out using High Plate Count medium [12]. For the isolation and enumeration of autotrophic sulphur oxidisers, inorganic Starkey's basal salt medium [13] containing 2.5% (w/v) sodium thiosulphate / potassium tetrathionate / sulphur powder as energy source was used. For sulphur oxidising mixotrophs the above media were supplemented with 0.02% (w/v) organic substrate viz. glucose/glycerol/yeast extract. For isolation and purification of isolates all the above solid media were prepared with 1% (w/v) washed agar-agar powder as solidifying agent [13]. In case of iron oxidisers, basal salt medium was prepared as described by Johnson [14] with ferrous sulphate as energy source. For heterotrophic iron oxidisers 0.02% (w/v) yeast extract was supplemented as organic substrate in the medium and for the solid medium 0.8% (w/v) washed agar-agar powder was used as solidifying agent. Cell count was 1280

Molecular Biology and Taxonomy

carried out by standard 3-tube Most Probable Number (MPN) technique and results were recorded after 7, 14 and 21 days of incubation [15]. For the ferrous oxidation rate of isolated cultures, 100 ml system with 2% (w/v) ferrous sulphate was taken. For the identification of the heterotrophic sulphur oxidisers, the cultures were transferred on nutrient agar plates and various biochemical tests were performed as per the Bergey’s manual [16, 17] and identification of these cultures were done by Biolog® GN-2 identification microplates (Biolog Inc., USA). All the experiments were carried out in triplicates. Plates and tubes were incubated at 30±2°C. 2.3 Analysis Similarity index, Shannon Weiner diversity, richness and evenness was calculated by the standard formula [18]. Soluble ferrous, sulphate, acidity-alkalinity and thiosulphate were determined by the standard titrimetric and spectrophotometric method [11]. 3.

RESULTS AND DISCUSSION Soil and water samples collected from Malanjkhand Copper Mine showed pH variations between 3.3 to 8.0 and redox potential of 505 to 250 mV. In terms of copper concentration also, they showed nearly 10 fold variations in terms of minimum and maximum copper present (Table 1). The detail description of appearance, conductivity and sulphate concentration is reported elsewhere [18]. The observed variation in the various parameters was obviously due to the selective representative sites from different ecological niches. The sample MJ-5 showed the highest copper as it was water oozing out from the chalcopyrite rock where the upper layer of the rock was of blue colour indicating the formation of secondary copper minerals like covellite and chalcocite. Table 1. Characteristics of sample collection sites at Malanjkhand Copper Mine Sample No. MJ-1 MJ-2 MJ-3 MJ-4 MJ-5 MJ-6 MJ-7

7.4 8.0

Redox potential (mV) 250 280

Soluble copper (g/l) 0.21 0.46

6.6

290

0.13

7.0 4.9 3.3

300 355 505

0.06 0.63 0.42

4.5

325

0.13

Collection site and sample appearance

pH

Clear flowing water at base of open pit mine Brown sediments from the base of open pit mine Clear water from water pond at base of open pit mine Reddish brown mud from the mine pond Light blue water oozing from chalcopyrite rock Greenish blue turbid leachate of acid heap leaching Yellowish brown dry sediments of heap leaching from collection pond

Qualitative and quantitative determination of mesophilic neutrophilic heterotrophic bacteria from the various sites of the Malanjkhand Copper Mine were carried out and results are shown in terms of variety and diversity indices in Table 2. The highest varieties of heterotrophic bacteria were isolated from sample MJ-1 that was having slightly alkaline pH. On the other hand sample MJ-5 having pH 4.9 also showed as high as ten varieties of neutrophilic bacteria indicating the acid tolerant nature of these isolates. The lowest variety observed was 6, even when the sample pH was as low as 3.3. This indicates quite 1281

Molecular Biology and Taxonomy

substantial number of acid tolerant population in Malanjkhand Copper Mine. When the richness, evenness and diversity were considered, the sample MJ-1 showed the highest figure. This could be due to the reason that the sample was flowing water passing through various locations of the mine, thus having population from various sites. Similarly even sample MJ-5 (oozing water) showed second highest richness. When the overall data is consider acidic water showed higher indices compared to alkaline samples. The diversity of the sample is also clear from the similarity matrix shown in Table 3. It can be seen from the data the highest similarity observed was just 42% while it was as low as 0% in three samples. This determination indicates the selection of diverse sites at a mine, which provides maximum possible microbial varieties. Table 2. Diversity indices of neutrophilic heterotrophic bacteria at Malanjkhand Copper Mine Site MJ-1 MJ-2 MJ-3 MJ-4 MJ-5 MJ-6 MJ-7

Number of varieties 11 8 9 7 10 6 8

Richness RMargalef 1.40 0.85 0.98 0.85 1.30 0.74 1.07

Shannon Weiner diversity (H') 2.17 1.19 1.49 1.60 1.51 1.20 1.91

Evenness EPielou 0.92 0.57 0.68 0.82 0.66 0.67 0.92

Table 3. Similarity matrix of the collection sites for heterotrophic bacterial isolates Site MJ-1 MJ-2 MJ-3 MJ-4 MJ-5 MJ-6 MJ-7

MJ-1 1

MJ-2 0.42 1

MJ-3 0 0.18 1

MJ-4 0 0.13 0.19 1

MJ-5 0.05 0.11 0.11 0.06 1

MJ-6 0 0.07 0.20 0.23 0.19 1

MJ-7 0.05 0.06 0.18 0.20 0.17 0.21 1

Table 4. MPN count of iron and sulphur oxidising bacteria in various media (21 days incubation) Substrate ST Tetrathionate Sulphur ST + Y.E. ST + Glucose ST + Glycerol Ferrous

Group of organism Autotrophic 'S' oxidiser Mixotrophic 'S' oxidiser

MJ-1 2.8x103 0 0 2.3x103 9.0x102 4.6x104

Autotrophic 2.9x102 Fe2+ oxidiser Ferrous + Y.E. Mixotrophic 3.0x101 Fe2+ oxidiser Y.E.:yeast extract, ST: sodium thiosulphate 1282

Counts per ml Sampling site MJ-3 MJ-5 7x102 2.4x104 0 4.3x103 0 2.3x103 1.4x103 4.3x103 4.3x103 2.3x103 1.1x105 1.5x104

MJ-6 4.3x103 2.3x103 2.3x103 2.3x103 9.0x102 1.5x104

MJ-7 0 0 0 0 2.8x103 4.6x104

5.5x104

1.1x104

4.3x102

9.3x103

4.3x103

7.5x102

3.0x101

2.3x103

Molecular Biology and Taxonomy

The quantitative distributions of auto- and heterotrophic sulphur and iron oxidisers at five different sites using most probable number technique is shown in Table 4. Presence of both, auto- and heterotrophic iron oxidisers as well as heterotrophic sulphur oxidisers were recorded from all the sites. The autotrophic sulphur oxidisers utilising thiosulphate were more prevalent as compared to those, which were utilising tetrathionate or sulphur as energy source. In spite of the diverse pH of the samples, all the sites showed presence of acidophilic iron oxidising organisms while the acidophilic sulphur oxidiser were more prevalent in sample MJ-5 and MJ-6, obviously due to the acidic pH of these samples. The MPN count analysis also indicate the presence of both autotrophic and mixotrophic iron and sulphur oxidisers. On the basis of studied ratio of inorganic to organic compounds at very low concentration such as thiosulphate / ferrous and glucose / yeast extract / acetate some isolates could be grouped as mixotrophs or facultative chemolithotrophs. There were seven varieties of mesophilic Gram negative bacteria observed on Starkey's basal medium containing sodium thiosulphate supplemented with either glucose or yeast extract. Out of these seven isolates, five were grown on Nutrient agar medium and were differentiated depending on the basic characteristics as depicted in Table 5. When these isolates were further characterised by Biolog® GN-2 plates, they were identified as Pseudomonas stutzeri, Pseudomonas aeruginosa, Brevundimonas diminuta, Stenotrophomonas maltophilia and Alcaligenes spp. All the isolates could be grouped in the genus Pseudomonas except Alcaligenes, when they were examined for their classification in Bergey's manual of determinative bacteriology [16] and Bergey's manual of systematic bacteriology [17]. Table 5. Characteristic of identified heterotrophic thiosulphate oxidisers Test Medium Motility Fluorescence Pigment Growth at 41°C Oxidase Gelatine Starch Glucose Fructose T.S.I. Isolated from sample Identification

1 ST+Glu. + − Yellow ± + − + + + Ak/Ak MJ-3, MJ-5 Ps. stutzeri

2 ST+Y.E. + + Bluish green + + + − + + Ak/Ak MJ-1, MJ-5 Ps. aeruginosa

Isolate number 3 4 ST+Y.E. ST+Y.E. + + − − − − n.d. ± + + + − − − + − + − Ak/Ak Ak/Ak MJ-1, MJ-3, MJ-1 MJ-5 B. diminuta S. maltophilia

5 ST+Glu. + − − n.d. + − − − + Ak/Ak MJ-3, MJ-5 Alcaligenes spp.

ST: sodium thiosulphate; Y.E: yeast extract; Glu: glucose; Ak: alkaline; n.d: not determined

All this isolates grew on Starkey's medium indicating their role in thiosulphate oxidation. When they were grown in broth medium with various pH for thiosulphate oxidation, the maximum oxidation observed was in the range of 70 to 93% at 4.5 initial pH of the medium. They also oxidised thiosulphate upto pH 8.0. During the thiosulphate oxidation, first the pH of the medium increased and reached to as high as 8.5 to 9.0 and thereafter for two isolates, it decreased to acidic side as low as pH 4.0 (data not shown). 1283

Molecular Biology and Taxonomy

In case of ferrous oxidation, direct growth on solid medium from the sample was not achieved. Therefore, positive tubes of the MPN count were selected for the isolation of iron oxidisers on solid media. Twelve different varieties of colonies were observed on ferrous and ferrous supplemented with yeast extract medium. Out of which 10 varieties were successfully purified and their ferrous oxidation patterns were studied between pH 1.0 and 3.0. The mg/l/h ferrous oxidation of these isolates at various pH are shown in Table 6. As can be seen from the data, all the isolates showed almost similar ferrous oxidation activity both at pH 2.3 and 3.1. The ferrous oxidation activity reduced by 5 to 20 folds at pH 1.2 except for the isolates 1 and 3, which did not show any activity at this pH. When the activity was compared between auto- and mixotrophs, the mixotrophs showed more activity irrespective of the pH studied. This finding once again shows the importance of mixotrophic organisms in mining activity. Various authors have reported the importance of heterotrophs in biomining [7, 19, 20]. Table 6. Iron oxidation rates of autotrophic and mixotrophic iron oxidising isolates Isolate no. 1 2 3 6 8 9 10 11 12 15

Growth substrate in medium Ferrous

Ferrous + yeast extract

1.2 0 6.3 0 12.2 21.9 22.6 17.4 20.7 26.0 29.3

Ferrous oxidation rate (mg/l/hr) pH 2.3 109.6 116.6 120.0 170.0 168.9 160.4 156.0 167.0 175.6 156.3

3.1 107.4 124.8 141.1 173.0 176.0 173.7 170.0 170.0 183.0 153.7

The attempts were made for the identification of the mesophilic auto- and mixotrophic iron and sulphur oxidizers. On the basis of their growth pattern and some of the biological tests, they were identified as Acidithiobacillus ferrooxidans, Acidithiobacillus thiooxidans, Thiobacillus thioparus, Thiobacillus versutus and Thiobacillus intermedius. Other mixotrophic iron oxidisers growing in the presence of yeast extract were not identified due to the limitation of the facility available. 4.

CONCLUSION Malanjkhand Copper Mine represents quite diversed physico-chemical ecosystem. Inspite of mining environment, the ecosystem showed considerable richness, evenness and diversity for neutrophilic heterotrophic bacteria. None of the selected site has more than 42% similarity. Almost all the sites showed considerable number of auto- and heterotrophic iron and sulphur oxidisers. Thiosulphate was proved to be better substrate for the study of sulphur oxidisers. The mining ecosystem showed Pseudomonas as a keystone genus among the heterotrophic sulphur oxidisers. In case of cultivable iron oxidisers on solid media, the mixotrophic group was found to be widespread and dominant as compared to autotrophs. The overall finding of mesophiles reveal that autotrophs, mixotrophs and heterotrophs may play equally important role in mining activity. These findings are very much encouraging and suggest that the mixotrophic and heterotrophic Gram-negative bacteria can be used for biohydrometallurgical processes along with the autotrophs. 1284

Molecular Biology and Taxonomy

ACKNOWLEDMENTS We are thankful to Department of Biotechnology, New Delhi, India for the project grant and Research Associateship to D. R. Tipre. We are grateful to Hindustan Copper Ltd., Malanjkhand for helping in sample collection. We also acknowledge the assistance provided by K.P. Ladhawala and V.V. Gajjar. REFERENCES 1. D. B. Johnson, In: Biohydrometallurgy and the Environment toward the Mining of the 21st Century, Part A, R. Amils and A. Ballester (eds.), Proc. Intl. Biohydrometallurgy Symp., Elsevier, Amsterdam, (1999) 645. 2. Pandey and L. M. S. Palni, J. Sci. Indus. Res., 57 (1998) 668. 3. K. A. Natarajan, In: Microbes, Minerals and Environment, Geological Survey of India, (1998). 4. S. R. Dave and K. A. Natarajan, Trans. I.I.M., 40 (4) (1987) 315. 5. D. B. Johnson and F. F. Roberto, In: Biomining: Theory, Microbes and Industrial Processes, D. E. Rawlings (ed.), Landes Bioscience, USA, (1997) 302. 6. S. M. Sievert, T. Brinkhoff, G. Muyzer, W. Ziebis and J. Kuever, Appl. Environ. Microbiol., 65 (9) (1999) 3834. 7. D. B. Johnson, FEMS Microbiol. Ecol. 27 (1998) 307. 8. S. R. Dave and K. A. Natarajan, Hydrometallurgy, 7 (1981) 235. 9. M. Hahn S. Willscher and G. Straube, In: Biohydrometallurgical Technologies I, A. E. Torma, M. L. Apel and C. L. Brierley (eds.), IBS, TMS, USA, (1993) 99. 10. H. L. Ehrlich, In: Biohydrometallurgy and the Environment toward the Mining of the 21st Century, Part A, R. Amils and A. Ballester (eds.), Proc. Intl. Biohydrometallurgy Symp., Elsevier, Amsterdam, (1999) 3. 11. D. Eaton, L. S. Clesceri and A. E. Greenberg (eds.), Standard methods for the examination of water and waste water, 19th ed, APHA, USA, (1999). 12. The Himedia Manual for Microbiology Laboratory Practice, Himedia Laboratories Pvt. Ltd., Mumbai, India, (1998). 13. S. R. Dave, Ph. D. Thesis, The University of Mysore, Mysore, India, (1980). 14. D. B. Johnson, J. H. M. Macvicar and S. Rolfe, J. Microbio. Methods, 7 (1987) 9. 15. B. Escobar and I. Godoy, In: Biohydrometallurgy and the Environment toward the Mining of the 21st Century, Part A, R. Amils and A. Ballester (eds.), Proc. Intl. Biohydrometallurgy Symp., Elsevier, Amsterdam, (1999) 681. 16. Bergey's Manual of Determinative Bacteriology, In: J. G. Holt, N. R. Krieg, P. H. A. Sneath, J. T. Staley and S. T. Williams (eds.), 9th ed., Lippincott Williams and Wilkins, USA, (1994). 17. Bergey's Manual of Systematic Bacteriology, In: J. T. Staley, M. P. Bryant, N. Pfennig and J. G. Holt (eds.), Vol 3, 1st ed., Williams and Wilkins, USA, (1989). 18. S. R. Dave, D. R. Tipre and V. V. Gajjar, Asian J Microbiol. Biotech. Env. Sc., 4 (3) (2002) 367. 19. Schippers, R. Hallmann, S. Wentzien and W. Sand, Appl. Environ. Microbiol., 61 (8) (1995) 2930. 20. M. A. Ghauri and D. B. Johnson, FEMS Microbiol. Ecology, 85 (1991) 327.

1285

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Expression proteomics of Acidithiobacillus ferrooxidans grown in different metal sulfides: analysis of rhodanese-like proteins P. Ramírez, L. Valenzuela, M. Acosta, N. Guiliani and C.A. Jerez Laboratory of Molecular Microbiology and Biotechnology, Department of Biology, Faculty of Sciences, and Millennium Institute for Advanced Studies in Cell Biology and Biotechnology, University of Chile, Santiago, Chile. [email protected] Abstract By expression proteomics of Acidithiobacillus ferrooxidans ATCC 19859 we characterized a set of proteins changing their levels of expression during growth of the microorganism in metal sulfides and elemental sulfur compared with growth in ferrous iron. By determination of the N-terminal amino acid sequences of these proteins present in proteomic arrays obtained after two-dimensional polyacrylamide gel electrophoresis and by using the available preliminary genomic sequence of A. ferrooxidans ATCC 23270 we identified several of them. The genomic context around these protein genes suggests their involvement in the sulfur metabolism of A. ferrooxidans. Amongst some of the proteins highly upregulated by growth in sulfur compounds (and downregulated by growth in ferrous iron) we found an outer membrane protein, an exported putative thiosulfate sulfur transferase (rhodanese) protein and a 33 kDa putative thiosulfate/sulfate binding protein amongst others. In the present work, we further analyzed the genome sequence from A. ferrooxidans and found two other rhodanase-like proteins: P14 and P16 whose genes did not contain signal peptides. The predicted tertiary structures of P21, P16 and P14 were very similar, especially in their putative active site for thiosulfate binding. The genomic context of the genes for these proteins was annotated in an attempt to suggest their possible roles. We further isolated from the DNA of A. ferrooxidans the gene coding for P14 and cloned and expressed the protein in E. coli, detecting a functional rhodanase activity for P14. This family of rhodanese-like proteins may be important in the sulfur metabolism of A. ferrooxidans. Keywords: rhodanese, Acidithiobacillus ferrooxidans, sulfur metabolism, proteomics 1.

INTRODUCTION Acidithiobacillus ferrooxidans is a chemolithoauthotrophic bacterium that obtains its energy from the oxidation of ferrous iron, elemental sulfur, or partially oxidized sulfur compounds (1, 2, 3). The ability of these and other microorganisms present in their habitat to solubilize metal sulfides is succesfully applied in biomining operations (2). Recently, it has been proposed that pyrite (FeS2) and other metal sulfides are degraded by an indirect mechanism generating thiosulfate as the main intermediate (4). Iron (III) ions are exclusively the oxidizing agents for the dissolution. Thiosulfate would be consequently degraded in a cyclic process to sulfate, with elemental sulfur being a side product. This explains why only Fe(II) ion-oxidizing bacteria are capable of oxidizing these metal 1287

Molecular Biology and Taxonomy

sulfides (4). In addition, enzymes for thiosulfate or sulfite oxidation from A. ferrooxidans or A. thiooxidans may succesfully compete with the chemical reactions with iron (III) ions as an oxidizing agent (4). A rhodanese activity has been previously described in A. ferrooxidans (5). This enzyme is a thiosulfate sulfur-transferase, which breaks the S-S bond present in thiosulfate, generating sulfur and sulfite. Other enzymes may also participate in the thiosulfate mechanism, such as the thiosulfate-oxidizing enzyme of A. ferrooxidans (6). By proteomic analysis, we have previously studied the global changes in gene expression of A. ferrooxidans when the microorganism was grown under different conditions and have identified an exported rhodanese-like protein (P21) which is induced when A. ferrooxidans is grown in metal sulfides and different sulfur compounds but is almost entirely repressed by growth in ferrous iron (7). Unlike cytoplasmic rhodaneses, P21 was located in the periphery of A. ferrooxidans cells and was regulated depending on the oxidizable substrate. By using the available preliminary genomic sequence of A. ferrooxidans ATCC 23270, the genomic context around gene p21 showed the presence of other ORFs corresponding to proteins such as thioredoxins and sulfate-thiosulfate binding proteins, clearly suggesting the involvement of P21 in inorganic sulfur metabolism in A. ferrooxidans (7). Here, we extend our genomic analysis, and define two new rhodaneselike proteins, P14 and P16. The gene coding for P14 was isolated and after its cloning and expression in E. coli, its functional rhodanese activity was demonstrated. 2.

MATERIALS AND METHODS

2.1 Bacterial strains and growth conditions A. ferrooxidans strain ATCC 19859 was grown in ferrous iron-containing modified 9K medium or in sulfur or pyrite as described before (7). E. coli strain BL21(DE3) containing plasmid pGZ105 with the glpE insert coding for the E. coli rhodanese (8) was a kind gift of T. Larson. E. coli strains BL21(DE3) and derivatives were grown in LB medium (9). 2.2 2-D NEPHGE and SDS-PAGE Total cell proteins were separated by SDS-PAGE or two-dimensional non-equilibrium pH polyacrylamide gel electrophoresis (2-D NEPHGE) as described before for A. ferrooxidans (7). 2.3 Primers and PCR conditions The oligonucleotide primers were purchased from Genset Corporation. Taq and Pwo polymerases were from Promega and Roche, respectively, and were used according to the manufacturer's recommendations. The oligonucleotide primer sequences were deduced from the ORFs found in the available almost finished DNA genomic sequence of A. ferrooxidans strain ATCC 23270 (http://www.tigr.org). These primers were P14NTER-NdeI (5´-gTTTTTAgTCATATggggAAggTCATgg-3´) and P14CTER-XhoIHT (5´-TAggCTCCggCTCTCgAgggAAACgAC-3´). To amplify the p14 gene we used a two-step HotPCR protocol: 3 min at 95°C followed by 20 cycles at 95°C for 30s, 62°C for 30s and 45s at 72°C and finally 72ºC for 3 min.

1288

Molecular Biology and Taxonomy

2.4 DNA manipulations Restriction enzyme digestions, T4 DNA ligase and recombinant DNA techniques were carried out according to standard laboratory procedures (9). The dideoxy chain termination method was employed to sequence DNA using [γ−33P]ATP and the dsDNA Cycle Sequencing System from Gibco BRL. The DNA sequences were compiled and analysed with the University of Wisconsin GCG Package (Version 9.1, Genetics Computer Group, Madison, Wis.). 2.5 P14 gene cloning and expression We used pGEM-T (Promega) and the pET System (Novagen). The p14 gene was obtained by PCR using P14NTER-NdeI and P14CTER-XhoIHT primers corresponding to the N-terminal and C-terminal end sequences of P14 and containing NdeI and XhoI restriction sites, respectively. The DNA fragments separated by electrophoresis in 1% agarose gels were recovered, purified with Wizard PCR Prep (Promega) and ligated to pGEM-T vector (Promega). The ligation products were used to transform E. coli JM109. The positive clones were analyzed by using colony PCR and the corresponding plasmids with inserts were purified. The DNA fragment of interest was ligated to pET21b(+) vector (Novagen), previously digested with NdeI and XhoI. The ligation product (p14H vector) was used to transform E. coli strain BL21(DE3). The recombinant clones were selected on LB solid medium supplemented with ampicillin (100 µg/ml). The induction/expression analysis was done in the presence or absence of 1 mM IPTG, added when the cultures reached an OD600 of 0.6. Expression of the recombinant P14 (rP14) was analyzed by SDSPAGE of total cell extracts. 2.6 Determination of rhodanese activity Rhodanese (thiosulfate:cyanide sulfurtransferase; EC 2.8.1.1) activity was assayed in crude enzyme extracts or with the purified recombinant protein rP21. As a control, we used the recombinant rhodanese GlpE from E. coli (8). The assay was done at pH 7.5 - 8.5 essentially as described before by Singleton and Smith (10), and Gardner and Rawlings (11). 2.7 Sequence analysis Identity/similarity searching in databases was done by using the BlastP program (12) from NCBI (http://www.ncbi.nlm.nih.gov) and from the unfinished A. ferrooxidans ATCC 23270 genome site (http://www.tigr.org). Multiple alignments, molecular masses and isoelectric points of ORFs, the presence of transmembrane domains in the analyzed ORFs and the putative functions and predicted subcellular locations of the proteins coded by the different ORFs were analyzed as described before (7). 3.

RESULTS AND DISCUSSION

3.1 Determination of rhodanese activity in A. ferrooxidans We could not find in vitro rhodanese activity in the rhodanese-like P21 protein that we recently described (7). This lack of activity of the recombinant P21 could be to a number of reasons. However, one possibility is that P21 does not correspond to the previously described rhodanese activity in A. ferrooxidans (5). We therefore measured rhodanese activity in crude cell-free extracts from A. ferrooxidans ATCC 19859 grown under different conditions. Figure 1 shows that rhodanese activity in crude extracts from 1289

Molecular Biology and Taxonomy

µMol SCN-/mg prot/min uMol SCN-/mg de prot/ min

cells grown in sulfur, ferrous iron or pyrite, gave values in the range of those reported by Tabita et al. (5) (0.250 µMoles SCN-/min/mg of protein), the differences between ironand sulfur-grown cells being only about 25%. On the other hand, the synthesis of P21 appeared to be regulated by the presence of iron in the growth medium of sulfur grown cells since P21 levels are decreased by more than 30 fold in the presence of 10 mM Fe (II) (7). This result is clearly different from that expected if the levels of P21 synthesized in ferrous iron- or sulfur-grown cells were responsible of the observed rhodanese activity. Recently, Gardner and Rawlings (11) detected thiosulfate-sulfur transferase activity in whole cells and crude extracts from A. ferrooxidans, A. thiooxidans, and A. caldus whereas this activity was absent from Leptospirillum ferrooxidans, since this microorganism is only capable of oxidizing ferrous iron or the iron contained in pyrite, but not its sulfur moiety (4). These results support the idea of rhodanese being involved during in vivo sulfur oxidation. 0.3 0.25 0.2 0.15 0.1 0.05 0

So

Fe 2+

FeS2

Figure 1. Rhodanese activity of crude cell extracts from A. ferrooxidans. Cells of A. ferrooxidans ATCC 19859 were grown as indicated in elemental sulfur, ferrous iron or pyrite and the corresponding cell-free extracts were prepared to determine rhodanese activity In the studies of Gardner and Rawlings (11), the rhodanese activity levels in A. ferrooxidans were also about the same when cells were grown either in ferrous iron or in sulfur. Since rhodaneses have been reported as constitutive (13), the activity measured by Gardner and Rawlings most likely corresponded to the cytoplasmic rhodanese. The lack of rhodanase activity of P21 may be due to the need of additional polypeptides required for it to be active, as it occurs with the thiosulfate oxidizing complex from Paracoccus versutus (13). Alternatively, the regulated exported P21 may have a different role during sulfur metabolism. 3.2 Search in the genomic sequence of A. ferrooxidans of putative rhodanase genes and their genetic context In the unfinished A. ferrooxidans genome sequence we found at least two other small sequences with rhodanese-like similarities: P14 and P16 (7). These two putative ORFs did not present signal peptides and the corresponding putative proteins had isoelectric points of 4.8 (P14) and 9.3 (P16). Figure 2 shows the genomic analysis of the regions surrounding genes p21, p14 and p16. Several putative ORFs related to sulfur metabolism were deduced in the context of p21 (7): upstream of p21 a terminal oxidase subunit (tox1) and a sulfate/thiosulfate binding protein (sbp1) were located. These putative genes together with a putative C4-dicarboxilate transporter (cdt), an unknown ORF and a hypothetical protein ORF apparently form a cluster with the same orientation. On the other hand, a putative gene with high similarity to a periplasmic thioredoxin (trx), together with a terminal oxidase subunit (tox2), a sulfate/molybdate 1290

Molecular Biology and Taxonomy

binding protein (sbp2) and an unknown ORF were oriented in a divergent way from the p21 cluster. The existence in A. ferrooxidans ATCC 19859 of an exported protein P21 similar to a thiosulfate-sulfur-transferase and which is regulated depending on the oxidizable substrate is very interesting considering the proposal that the oxidation of pyrite generates thiosulfate as one of the main intermediates (4). A

cdt

sbp1 (p33)

tox1

p21

sams

sk

trx

tox2

sbp2

B

prk

5,10 mthfr

pm

amid

sahch

p14

gf

2,3 bfgk

C

trx gtx p16

acr1

acr2

acr3

tr

Figure 2. Schematic map of the contig regions containing the putative gene cluster contexts around genes p21(A), p14 (B) and p16 (C) from A. ferrooxidans. ORFs possibly related with sulfur metabolism are in gray. Coding regions containing putative signal peptides for the Sec system are indicated with black vertical rectangles. The names of the genes coding for P14, P16 and P21 are indicated in bold. The putative ORFs present in these regions are described in the text On the other hand, it has been shown that A. ferrooxidans generates thiosulfate when grown in a medium containing elemental sulfur (15). This could explain why P21 is induced when cells are grown in elemental sulfur. If P21 is involved in thiosulfate metabolism, one should expect an increased expression of the protein when the cells are grown in pyrite, thiosulfate or sulfur, as we have observed (7). The lack of repression of P21 synthesis by growth in pyrite when compared with that obtained by growth in ferrous iron was unexpected. However, during pyrite attack, much smaller amounts of free ferrous iron are probably present, and as we have shown, the levels of P21 drastically decreased at higher concentration of ferrous iron in the growth medium. The studies on a small rhodanese-like protein from Wolinella succinogenes showed that it acts as a periplasmic sulfide dehydrogenase and uses the same catalytic cysteine involved in anion transferase and hydrolase activity (16). This suggests a possible redox function for rhodanese-like proteins similar to that of the thioredoxin proteins. This is supported by the presence on the C-terminal end of P21, and not in P14 or P16, a cysteine motif Cys-XX-Trp-XX-Cys known to bind iron-sulfur clusters in electron transport complexes (14). It is also possible that P21 from A. ferrooxidans has a dithiol-disulfide redox activity analogous to the one in W. succinogenes. The ORFs coding for P14 and P16 were located in different contigs of the incomplete genome from A. ferrooxidans ATCC 23270 (contigs 7920 and 7913 respectively), and with entirely different neighbouring putative genes. Upstream of p14 a sugar kinase (sk) is located in a divergent direction and followed by three ORFs apparently forming a cluster 1291

Molecular Biology and Taxonomy

with the same orientation: S-adenosylmethionine synthetase (sams), S-adenosyl-Lhomocysteine hydrolase (sahch) and a 5, 10-methylene tetrahydrofolate reductase (5-10 mthfr). Downstream of p14 we found an unknown ORF with the same orientation followed in opposite direction by a putative glycogen phosphorylase (gf) and a 2,3biphosphoglycerate kinase (2,3-bfgk). Upstream of p16 and with the same orientation we found ORFs coding for putatives glutarhedoxin (gtx), a thioredoxin (trx) and an Nacetylmuramoyl-L-alanine amidase (amid). Downstream of the putative rhodanese-like p16 we found with the same orientation, three ORFs coding for putative genes for acryflavin resistance (acr1, acr2 and acr3) and a possible transcriptional regulator (tr). 3.3 Structural comparison of the putative rhodanase-like proteins P14, P16 and P21 with the rhodanese GlpE from E. coli The comparative analysis of the amino acid sequences of P14, P16 and P21 with several known thiosulfate-sulfur transferases, which activity has been demonstrated in vitro, showed a significant similarity (average 40%). The three proteins also contained the highly conserved structural domains CH2A, CH2B and a catalytic site with a Cys, typical of thiosulfate-sulfur transferases (Fig. 3).

Figure 3. Alignment of the amino acid sequences of the rhodanese-like proteins P14, P16 and P21 from A. ferrooxidans with GlpE, the rhodanese from E. coli. The active site and the structural domains CH2A and CH2B, which are conserved in all rhodaneses are enclosed by rectangles. The secondary structure elements are indicated above the alignment (black arrows, for β-strands and grey rectangles for the α-helices) Recently, Spallarossa et al., (17) compared the crystalline structure of the GlpE protein or rhodanese from E. coli with that of other rhodaneses and described that the catalytic site for the thiosulfate sulfur transferase activity is formed by six amino acids containing a Cys-65. The loops forming part of the site contain the βD sheet and the D αhelix previously described (18). When we aligned the structure of GlpE present in the data 1292

Molecular Biology and Taxonomy

banks (MMDB Id: 18023 PDB Id:1GN0) with those putative structures of P14, P16 and P21 by means of the program Cn3D (19), we found similar overall three-dimensional structures with a great conservation in the active sites where thiosulfate would be bound, as shown in Figure 4. In spite of the observed amino acid variability, the active site loop conformation is similar in all the proteins compared in Fig. 4. At least three polar (often charged) residues are observed at these six sites in rhodanese enzymes, contributing to the buildup of a positive electrostatic field, expected to lower the pKa of the catalytic Cys residue. An abundance of potentially functional rhodanese-like proteins has been observed in several genomes (20) and we have observed so far three such proteins in A. ferrooxidans. As pointed out by Spallarossa et al. (17), the amino acid variability observed for the putative active-site loops in all the identified homologs suggests a diversification of substrate specificity, while keeping the enzymatic activities related to the formation, interconversion and transport of compounds containing sulfane sulfur atoms. The three rhodanese-like proteins that we describe here belong to the ubiquitous rhodanese protein superfamily, and may have important roles in sulfur metabolism and or acquisition by A. ferrooxidans. Nevertheless, it is known that rhodanese-like proteins could show several alternative catalytic activities, amongst them, detoxification of toxic compounds such as arsenate and cyanide by either transferring anions or reducing them, and a chaperone activity to allow efficient assembly of iron-sulfur complex-containing proteins (20). GlpE

P21

Cys147

Cys65 Active site P14 P16 GlpE P21 p14 P14

P16

Cys101 Cys92

Figure 4. Three-dimensional structural comparisons of the regions containing the active site of rhodaneses. Taking the crystaline structure of GlpE as a model (17), the corresponding homolog active sites of P14, P16 and P21 are presented. The amino acid sequences of the compared structures are also indicated. Identical amino acids are in black and similar amino acids are in grey

1293

Molecular Biology and Taxonomy

µMolSCN-/mg prot/min

3.4 Isolation, cloning and expression of the p14 gene from A. ferrooxidans in E. coli Based on the greater similarities between P14 and GlpE, we decided to isolate and clone the p14 gene from A. ferrooxidans in E. coli. By using PCR and the appropriate primers, we isolated the DNA fragment of A. ferrooxidans containing the p14 gene and cloned it in an expression vector to transform E. coli. Fig. 5A shows the in vivo overexpression of a protein band with the molecular mass expected for P14. To find out if this E. coli transformant showed an increased rhodanese activity, we determined the capacity of the crude extract overexpressing P14 to transfer the sulfane sulfur from thiosulfate to cyanide (Fig. 5B).

Figure 5. In vivo overexpression of A. ferrooxidans p14 gene in E. coli and determination of rhodanese activity in the crude E. coli cell extracts. A. The pET21b(+) plasmid containing the p14 gene insert (lanes a, b) or the vector without the insert (lanes c, d) were used to transform E. coli strain BL21(DE3). All of the strains were grown for 2 h in the presence (lanes b, d) or in the absence (lanes a, c) of 1 mM IPTG added at the half-logarithmic phase of growth. Total cell proteins were separated by SDS-PAGE and stained with Coomassie blue. The arrow head indicates the migrating position of protein P14. B. Rhodanase activity in crude cell-free extracts. Column a, activity of extracts from cells containing the plasmid vector carrying gene p14. Column b, activity of cell extracts from bacteria carrying only the vector. Both strains were grown up to the exponential phase and were then induced for 2 h with 1 mM IPTG The transformed E. coli strain showed twice the activity of the control strain transformed with the plasmid without the p14 gene. Although the rhodanese activity observed was smaller than that seen in E. coli overexpressing GlpE from a plasmid (8), these results clearly indicate that gene p14 codes for a functional protein with rhodanese activity. Most likely, we conclude that P14 could be responsible for the rhodanese activity we detected in the crude extracts from A. ferrooxidans grown in sulfur, ferrous iron or pyrite (Fig. 1). A working model which summarizes some of our previous findings and those presented here is shown in figure 6. It is possible that P21 is not a periplasmic rhodanase enzyme but it is rather part of a possible complex in charge of thiosulfate oxidation. This putative complex could be different from the Sox model proposed for sulfur oxidation in many bacteria (13) since so far, we have not found any Sox-like genes in the available unfinished genome of A. ferrooxidans.

1294

Molecular Biology and Taxonomy

Sº + H2 SO3 FeS2

S2 03

2-

Fe 3+ Fe 2+ OM

PS

S-SO3 2-

TD

- O S-S-S-SO - + 2e3 3 TH - O S-S-S- O S-S-S-SO - + H O 3 3 2 3

S-SO3 2- + 2H2 O SBP (P33)

SO4 2-

+ SO4 2- +2H+

Sº + H2 SO3 + 2OH- SOR H2 SO3 + H2 O SO4 2- + 2e - + 4H+ TST

(P21)

IM

Cytosol

SO4 2-

S-SO32-

Cysteine biosynthesis

Synthesis of sulfur oxidation enzymes (TST, TD, TRX, GTX, CYT, etc.)

P14 ?

Figure 6. Working model for thiosulfate metabolism in A. ferrooxidans. The thiosulfate generated from pyrite or by chemical reaction between elemental sulfur and sulfite would be oxidized in the periplasm by means of two possible pathways: by thiosulfate dehydrogenase (TD) to give tetrathionate (3, 6, 13) or by means of a thiosulfate sulfur transferase (TST) (3) or (P21) (3). The sulfite generated would be then oxidized by a sulfite oxido reductase (SOR) (3). Sulfate or thiosulfate would be transported to the cytoplasm by a sulfate/thiosulfate binding protein (SBP or P33) (7). OM, outer membrane; PS, periplasmic space; IM, inner membrane 4. CONCLUSIONS 1. We have found three genes coding for rhodanase-like proteins in A. ferrooxidans. Their genomic context strongly suggests the involvement of these proteins in sulfur metabolim in this bacterium. 2. Protein P14 most likely is a rhodanese. However, due to the lack of an appropriate workable genetic system to perform functional genomics in A. ferrooxidans, at this point it is not possible to assign definitive roles to P14, P16 and P21 in sulfur metabolism in this bacterium.

1295

Molecular Biology and Taxonomy

AKNOWLEDGEMENTS This work was supported in part by grants from FONDECYT (projects 1000967 and 1030767), and ICM- P99-031-F. P.R. was the recipient of a DAAD Ph.D. scholarship. Preliminary sequence data for the A. ferrooxidans strain 23270 was obtained from The Institute for Genomic Research website at http://www.tigr.org. REFERENCES 1. 2. 3. 4. 5. 6. 7.

A.P. Harrison, Annu. Rev. Microbiol., 38 (1984) 265. D.G. Lundgren, Annu. Rev. Microbiol., 34 (1980) 263. I. Suzuki, Can. J. Microbiol., 45 (1999) 97. A. Schippers and W. Sand, Appl. Environ. Microbiol., 65 (1999) 319. R. Tabita, M. Silver and D.G. Lundgren, Can. J. Biochem., 47 (1969) 1141. M. Silver and D.G. Lundgren, Can. J. Biochem., 46 (1968) 1215. P. Ramírez, H. Toledo, N. Guiliani and C.A. Jerez., Appl. Environ. Microbiol., 68 (2002) 1837. 8. W.K. Ray, G. Zeng, M.B. Potters, A.M. Mansuri and T.J. Larson, J. Bacteriol., 182 (2000) 2277. 9. J. Sambrook and D.W. Russell (eds.), Molecular cloning, A laboratory manual. Cold Spring Harbor Laboratory Press, New York, 2001. 10. D. R. Singleton and D. W. Smith, Appl. Environ. Microbiol., 54 (1988) 2866. 11. M.N. Gardner and D.E. Rawlings, J. Appl. Microbiol., 89 (2000) 185. 12. S.F. Altschul, T.L. Madden, A.A. Schäffer, J. Zhang, Z. Zhang, W. Miller and D.L. Lipman, Nucleic Acids Res., 25 (1997) 3389. 13. C.G. Friedrich, Adv. Microb. Physiol., 39 (1998) 235. 14. B.C. Berks, S.J. Ferguson, J.W.B. Moir and D.J. Richardson, Biochim. Biophys. Acta., 1232 (1995) 97. 15. Shrihari, S.R. Bhavaraju, J.M. Modak, R. Kumar and K.S. Gandhi, Biotechnol. Bioeng., 41 (1993) 612. 16. V. Kreis-Kleinschmidt, F. Fahrenholz, E. Kojro, and A. Kröger, Eur. J. Biochem., 227 (1995) 137. 17. A. Spallarossa, J.L. Donahue, T.J. Larson, M. Bolognesi and D. Bordo, Structure, 9 (2001) 1117. 18. E.B. Fauman, J.P. Cogswell, B. Lovejoy, W.J. Rocque, W. Holmes, V.G. Montana, H. Piwnica-Worms, M.J. Rink and M.A. Saper, Cell 93 (1998) 617. 19. C.W.V. Hogue, Trends Biochem. Sci., 22 (1997) 314. 20. G. Storz and R. Hengge-Aronis (eds.), Bacterial stress responses, ASM Press, Washington, D.C., 2000.

1296

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Integration of metal-resistant determinants from the plasmid of an Acidocella strain into the chromosome of Escherichia coli DH5α S. Ghosh*, N.R. Mahapatra*, S. Nandi+ and P.C. Banerjee++ Indian Institute of Chemical Biology, 4 Raja S. C. Mullick Road, Kolkata 700032, India Abstract Integration of plasmid DNA into the chromosome of Escherichia coli DH5α conferring Zn2+- and Cd2+-resistance was suggested when the strain was transformed with plasmid DNA preparation from Acidocella sp. strain GS19h [1]. As evidence, pulsed-field gel electrophoresis (PFGE) pattern of genomic DNA of the transformants was observed to differ markedly with that of the untransformed DH5α strain. Further, the recombinant plasmids constructed with plasmid DNA pieces of strain GS19h at BamHI site of pBluescript-II KS+/- when used to transform E. coli DH5α strain, no plasmid DNA was detected in some of the lactose-negative, ampicillin- and zinc-resistant clones. The PFGE pattern of a transformed clone differed from that of the parent strain suggesting chromosomal integration of the recombinant plasmid(s). That the recombinant plasmid DNA(s) containing both the resistant genetic markers was integrated into chromosome of the transformed E. coli strain was reflected from hybridization of chromosomal DNA with the probes made from the plasmid DNA of strain GS19h and the vector DNA. Keywords: Acidocella strain, plasmid, chromosomal integration, metal resistance, E. coli 1.

INTRODUCTION Metal resistance is a plasmid-borne property in many bacterial species [2-4]. This property of the highly metal-resistant acidophilic heterotrophic bacterium Acidocella sp. strain GS19h was expressed in heterologous Escherichia coli system through transformation with the plasmid preparation from this strain [1]. Since existence of any plasmid in the transformed E. coli cells could not be detected, it was suggested that the metal-resistance conferring plasmid or a part of it was integrated into the E. coli chromosome rendering the transformants metal-resistance phenotype [1]. Integration of self and foreign plasmid DNA in both prokaryotic and eukaryotic microorganisma, viz. E. coli [5], Myxococcus xanthus [6], Streptomyces griseofulvus [7], Vibrio cholarae [8], Saccharomyces cerevisiae [9] and others [10-12] was reported earlier. Moreover, Inagaki

+

Department of Biochemistry, Molecular Biology & Cell Biology, Northwestern University, Evanston, IL 60208, USA ++ Corresponding author - Fax: 91-33-24735197/24730284; Email: [email protected] Present address: *Department of Medicine, University of California at San Diego, San Diego, CA 92161, USA

1297

Molecular Biology and Taxonomy

et al [13] observed gradual disappearance of a 8.8 kb chimeric plasmid that was introduced by electroporation into an Acidocella facilis strain which retained donor’s property indicating chromosomal integration of the recombinant plasmid. However, further experimentation supporting this integration event was not conducted. In this report, we present evidences for such integration of plasmid(s) of Acidocella sp. GS19h strain into the chromosome of E. coli DH5α employing pulsed field gel electrophoresis (PFGE) of the E. coli chromosomal DNA before and after transformation with (i) the plasmids of strain GS19h [1] and (ii) the recombinant DNA molecules containing plasmid DNA fragments of strain GS19h [14]. 2.

MATERIALS AND METHODS

2.1 Bacterial strains and culture conditions Acidocella sp. strain GS19h, E. coli DH5α and two of its plasmidless, Zn2+-resistant derivatives were used in this study. One of the Zn2+-resistant derivatives was obtained via transformation of the E. coli strain with plasmids of the Acidocella strain [1]. The other derivative (also resistant to ampicillin) evolved during cloning of putative plasmidmediated metal-resistant genes of Acidocella in this E. coli strain [14]. Medium and growth conditions of Acidocella sp. strain GS19h have been described [1]. The E. coli strain and its plasmidless derivatives were grown at 37°C on a rotary shaker in LB medium and the same containing either 12-16 mM Zn(SO)4 or 100 µg ml-1 ampicillin, respectively, as indicated in the text. 2.2 Plasmid purification Isolation and purification of plasmid DNA from the Acidocella sp. strain GS19h has been described [1,15]. The pBluescript-II KS+/- (Strategene) plasmid was either purchased or purified from an E. coli transformant of the same following standard procedure [15]. Electrophoresis of DNA samples was carried out in 0.5-0.8% (w/v) agarose gels. DNA bands were detected by ethidium bromide staining, as usual [16]. 2.3 Preparation of chromosomal DNA Chromosomal DNA from bacterial cells was prepared by CTAB method [17]. For pulsed field gel electrophoresis (PFGE), genomic DNA was prepared in situ in 0.7% (w/v) low melting agarose gel plugs. Cells of logarithmic phase were suspended in 10 mM TrisHCl (pH 7.6) and 1M NaCl. The suspension was mixed with equal volume of molten 1.4% (w/v) low melting agarose and allowed to set into plugs. The plugs were incubated at 37°C for 16 hr with gentle shaking in lysis solution [18]. The lysis solution was replaced by a solution containing 0.5 M EDTA (pH 9.0), sarkosyl (1%, w/v) and proteinase K (1 mg ml-1), and the incubation was continued at 50°C for 2 days or more until the blocks became transparent [19]. Agarose blocks were then treated with 1mM phenylmethylsulfonylfluoride (PMSF) and washed with TE (10 mM Tris, pH 8.0 and 1 mM EDTA) buffer. The blocks were digested with restriction enzyme for 4-12 hr. Enzyme- digested DNA was separated in 1% (w/v) PFGE-grade agarose gel using 0.5xTAE (20 mM Tris-acetate, pH 8.3 + 0.5 mM EDTA) as the running buffer at 10 Volt/cm and pulses ramping from 5-25 sec for 22 hr using Pulsaphor Plus System with a hexagonal electrode array (Pharmacia).

1298

Molecular Biology and Taxonomy

2.4 Selection of metal-resistant clones from the miniplasmid library Construction of miniplasmid library and selection of lactose-negative, ampicillin- and zinc-resistant (AmprZnr) E. coli DH5α clones have been described [14]. While some clones carried recombinant plasmid, others did not contain any plasmid indicating chromosomal insertion of resistant determinants. One of the plasmidless clones was selected for this study. 2.5 Hybridization Purified plasmid and chromosomal DNA samples were spotted on nylon membrane, the latter after digestion with BglII, by dot blotting [20]. The plasmid DNAs were radiolabelled with [α-32P] dATP by nick translation or random priming. Hybridization was carried out following standard methods [16]. 3.

RESULTS AND DISCUSSION It was previously suggested that the metal-resistant determinants from the plasmids of Acidocella sp. strain GS19h were integrated into the chromosome of E. coli DH5α imparting stable metal-resistance characteristics to it [1]. In this event a change in PFGE pattern of the transformed compared to that of the parent E. coli strain should be observed. Figure 1 shows that a lot of changes were introduced into the E. coli DH5α chromosome after transformation supporting the view of plasmid integration into the chromosome of the strain.

Figure 1. Pulsed-field gel electrophoretogram of genomic DNA digested with SfiI. Lane 1, E. coli DH5α; lane 2, Acidocella sp. strain GS19h; lane 3, transformed E. coli DH5α; lane 4, concatameric DNA molecular weight marker It is also evident from Figure 1 that many new sites for the sequence GGCC(N)5GGCC (for SfiI) were available in transformed E. coli chromosome after partial or full integration of plasmid(s) that conferred metal-resistance in E. coli DH5α. Major changes in the PFGE pattern were also observed (figures not shown) after digestion of the 1299

Molecular Biology and Taxonomy

transformed chromosome with NotI (GC↑GGCCGC), AseI (AT↑TAAT) and XbaI (T↑CTAGA). Such drastic changes in PFGE pattern may occur for two reasons. First, if the metal resistance conferring genes are contained in a composite transposon, and the insertion element components of the same are capable of multiple insertions in the E. coli chromosome; second, due to mutations at multiple locations producing new sites for restriction enzymes and conferring metal-resistance characteristics. Further, when a plasmidless, Zn2+- and ampicillin-resistant (ZnrAmpr) clone of E. coli DH5α was subjected to PFGE, a lot of difference was detected between the PFGE pattern of the clone and the parent strain (Figure 2). Several new bands appeared in the clone DNA while some original bands were missing in the same. The PFGE patterns also differed when the DNA samples were digested with AseI, NotI and SfiI individually (data not shown). These observations again suggest that plasmid-borne metal resistant determinants of the Acidocella strain can integrate into the chromosome of E. coli DH5α by a RecA independent recombination method.

Figure 2. PFGE profile of genomic DNA digested with XbaI. Lane 1, E. coli DH5α; lane 2, AmprZnr E. coli DH5α clone. Symbols denote some of the specific bands which are present in the respective lanes but are absent in the other lane The ZnrAmpr E. coli DH5α clone might had acquired its ampicillin-resistance marker from the vector pBluescript-II KS+/-, while the other resistance determinant, i.e. for ZnSO4, might had come from the plasmids of the Acidocella strain. Hybridization of BglII-digested chromosomal DNA of the ZnrAmpr E. coli clone with the plasmid DNA preparation of the Acidocella strain (Figure 3) and pBluescript-II KS+/- (Figure 4) supported the view that partial or full integration of Acidocella plasmid(s) and pBluescript-II KS+/- had occurred into the chromosome of E. coli DH5α during transformation and cloning.

1300

Molecular Biology and Taxonomy

Figure 3. Dot-blot hybridization of DNA samples using plasmid preparation from strain GS19h as probe. Lane 1, plasmid preparation from strain GS19h; lane 2, BglII digested chromosomal DNA of AmprZnr E. coli DH5α clone; lane 3, BglII digested chromosomal DNA of E. coli DH5α. Lanes a and b contained different amount of the same DNA sample

Figure 4. Dot-blot hybridization of DNA samples using pBluescriptII KS+/-. Lane 1, pBluescriptII KS +/- DNA; lane 2, BglII digested chromosomal DNA of E. coli DH5α; lane 3, BglII digested chromosomal DNA of AmprZnr E. coli DH5α clone. Lanes a and b contained different amount of the same DNA sample This study confirms that plasmids of Acidocella sp. GS19h strain can integrate into the chromosome of E.coli, a bacterium distinctly unrelated to Acidocella in respect of physiology and habitat. This observation leads to interpret that bacteria of this and related genera having similar physiological properties, such as Acidiphilium [21], donate genetic elements to other microorganisms inhabiting the same acidic mine environment [22] enriching their genetic repository to combat metal stress. The chromosomally inherited metal resistance conferring genes of Acidithiobacillus ferrooxidans [23,24], the most widely studied biomining bacterium [25], might probably had come from such extraneous genetic sources like plasmids of Acidocella (or Acidiphilium?) that can integrate even into habitually quite unknown bacterium like E. coli (this article). Although strongly speculative, this wild assumption is supported by the very fact that none of the so many native plasmids of Acidithiobacillus ferrooxidans harbours metal resistant determinants [24,26] although most of its plasmidless or plasmid-bearing strains exhibit metal resistance to different extent depending on the growth conditions [25,27]. Further study on metal resistance conferring genetic elements of these acidophilic heterotrophs may unveil many interesting aspects of bacterial metal resistance and consequent application of these genes in research and biotechnology [28-30].

1301

Molecular Biology and Taxonomy

ACKNOWLEDGEMENTS We are thankful to Late Dr. J. Das, Ex-Director, Dr. R. K. Ghosh and Dr. Rupak Bhadra, Scientists, Mr. Partha Sarkar, Mr. Chirojyoti Deb and Mr. Amit Chakraborty, Research Fellows, and other staff members of Indian Institute of Chemical Biology for helping us in various ways. Fellowships provided by Council of Scientific and Industrial Research, New Delhi, to S.G., N.R.M. and S.N. are gratefully acknowledged. REFERENCES 1. S. Ghosh, N. R. Mahapatra and P. C. Banerjee, Appl. Environ. Microbiol., 63 (1997) 4523. 2. J. T. Trevors, K. M. Oddie and B H Belliveau, FEMS Microbiol. Rev., 32 (1985) 39. 3. S. Silver and L. T. Phung, Annu. Rev. Microbiol., 50 (1996) 753. 4. K Suzuki, N Wakao, T Kimura, K Sakka and K. Ohmiya, Appl. Environ. Microbiol., 64 (1998) 411. 5. R. C. Deonier and N. Davidson, J. Mol. Biol., 107 (1976) 207. 6. A M Breton, S Jaoua and J Guespin-Michel, J. Bacteriol., 161 (1985) 523. 7. J. L. Larson and C. L. Hershberger, Plasmid. 23 (1990) 252. 8. S. Kar, R. K. Ghosh, A. N. Ghosh and A. Ghosh, FEMS Microbiol. Lett., 145 (1996) 17. 9. Α. Sakai, Y. Shimizu and F. Hishinuma, Appl. Microbiol. Biotechnol., 33 (1990) 302. 10. J. Casey, C. Daly and G. F. Fitzgerald, Appl. Environ. Microbiol., 57 (1991) 2677. 11. D. K. Tang, S. Y. Qiao and M. Wu, Biochem. Mol. Biol. Int., 36 (1995) 1025. 12. D. K. Mercer, K. P. Scott, C. M. Melville, L. A. Glover and H. J. Flint, FEMS Microbiol. Lett., 200 (2001) 163. 13. K. Inagaki, J. Tomono, N. Kishimoto, T. Tano and H. Tanaka, Biosci. Biotechnol. Biochem., 57 (1993) 1770. 14. S. Ghosh, N. R. Mahapatra and P. C. Banerjee. In: R. Amils and A. Ballester (eds.), Process Metallurgy 9B: Biohydrometallurgy and the environment toward the mining of the 21st century, Elsevier, Amsterdam, Part B, 1999, p. 21. 15. H. C. Birnboim and J. Doly, Nucleic Acids Res., 7 (1979) 1513. 16. J. Sambrook, E. F. Fritsch and T. Maniatis, Molecular cloning: a laboratory manual, 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbour, N.Y., 1989. 17. F. M. Ansubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith and K. Struhl, Current protocols in molecular biology, vol. 1, John Wiley and Sons, New York, 1995. 18. S. Nandi, G. Khetawat, S. Sengupta, R. Majumder, S. Kar, R. K. Bhadra, S. Roychoudhury and J. Das, Int. J. Syst. Bacteriol. 47 (1997) 858. 19. R. Majumdar, S. Sengupta, G. Khetawat, R. K. Bhadra, S. Roychoudhury and J. Das, J. Bacteriol., 178 (1996) 1105. 20. N. J. Dyson. In: T. A. Brown (ed.), Essential molecular biology: a practical approach, vol. II, IRL Press, Oxford, 1993, p.111. 21. N. Kishimoto, Y. Kosako, N. Wakao, T. Tano and A. Hiraishi, Syst. Appl. Microbiol., 18 (1995) 85. 22. B. M. Goebel and E. Stackebrandt, Appl. Environ. Microbiol., 60 (1994) 1614. 23. B. G. Butcher, S. M. Deane and D. E. Rawlings, Appl. Environ. Microbiol.,66 (2000) 1826. 24. D. E. Rawlings and T. Kusano, Microbiol. Rev., 58 (1994) 39. 25. Α.E. Torma, Adv. Biochem. Eng. 6 (1977) 1. 1302

Molecular Biology and Taxonomy

26. T. Shiratori, C. Inoue, M. Numata and T. Kusano, Curr. Microbiol., 23 (1991) 321. 27. S. Bhattacharyya, A. Das, B. K. Chakrabarti and P. C. Banerjee, Folia Microbiol. 37 (1992) 33. 28. D. E. Rawlings and S. Silver, Bio/Technol, 13 (1995) 773. 29. T. Barkay and J. Schaefer, Curr. Opinion Microbiol., 4 (2001) 318. 30. O. P. Dhankher, Y. Li, B. P. Rosen, J. Shi, D. Salt, J. F. Senecoff, N. A. Sashti and R. B. Meagher, Nature Biotechnol., 20 (2002), 140.

1303

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Involvement of Fe2+-dependent mercury volatilization enzyme system in mercury resistance of Acidithiobacillus Ferrooxidans strain MON-1 Tsuyoshi Sugio1, Mitsuko Fujii1, Fumiaki Takeuchi2, Atsunori Negishi3, Terunobu Maeda4 and Kazuo Kamimura1 1

Graduate School of Natural Science and Technology, Science and Technology for Energy Conversion, Okayama University, Tsushima Naka 1-1-1, Okayama 700-8530 2 Administration Center for Environmental Science and Technology, Okayama University, Tsushima Naka 1-1-1, Okayama 700-8530, 3 Technical Research Institute, Hazama Corporation, 515-1 Nishimukai, Karima, Tsukuba 305-0822, 4 Civil Chemical Engineering Corporation, 3411 Sanuki-machi Ryugasaki 301-0033 Japan Abstract The mechanism of mercury resistance in the highly mercury resistant strain Acidithiobacillus ferrooxidans MON-1 was studied by comparing with the moderately mercury resistant A. ferrooxidans strain SUG 2-2. Strain SUG 2-2 grew in a Fe2+ medium containing 6µM Hg2+ with a lag time of 22 days. Eight times successive cultivation of SUG 2-2 in the medium markedly shortened the lag time to 4 days. From this adapted culture, strain SUP-1 was newly isolated as a single colony on a 1.0% gellan gum plate containing ferrous iron. Strain SUP-1 could grow in a Fe2+ medium containing 20µM Hg2+ with the lag time of 10 days. Five times successive cultivation of SUP-1 in a Fe2+ medium containing 20µM Hg2+ shortened the lag time to 4 days. From this adapted culture, strain MON-1 was isolated as a single colony. Strain MON-1 could grow in a Fe2+ medium containing 20µM Hg2+ with the lag time of 2 days. The ability of strain MON-1 to grow rapidly in a Fe2+ medium containing 20µM Hg2+ maintained stably after the strain was cultured many times in a Fe2+ medium without Hg2+. The activities of the enzymes involved mainly in mercury detoxification were compared between strains SUG 2-2 and MON-1. Similar levels of NADPH-dependent mercury reductase activity were observed in cell extracts from strains SUG 2-2 and MON-1. Fe2+-dependent mercury volatilization activity was measured in 20 ml of water acidified with sulfuric acid (pH 2.5) containing resting cells (1.0 mg) and 140 nmol HgCl2 by incubating for 3 hours at 30°C. The amounts of mercury volatilized from the reaction mixture were 4.7 nmol for strain SUG 2-2 and 27 nmol for strain MON-1, respectively. Addition of 100 µmol of ferrous sulfate to the reaction mixture containing MON-1 cells enhanced the level of mercury volatilization activity 2.5 fold. In contrast, a 1.2 fold enhancement was observed in the case of SUG 2-2. These results indicate that a marked enhancement of the Fe2+-dependent mercury volatilization enzyme system conferred strain MON-1 the ability to grow rapidly in a Fe2+ medium containing 20µM Hg2+. 1305

Molecular Biology and Taxonomy

1.

INTRODUCTION The iron-oxidizing bacterium, Acidithiobacillus ferrooxidans is an acidophilic chemolithotrophic bacterium that can use both ferrous iron (Fe2+) and reduced inorganic sulfur compounds as energy sources. The bacterium is one of the most important bacteria for bioleaching of sulfide ores (1-3). A. ferrooxidans strains which possess a high ironoxidizing activity in an environment with many kinds of and high concentrations of heavy metals are required for microbiological leaching of low grade ores. It has been known that A. ferrooxidans cells are in general resistant to many heavy metals including iron, copper, zinc and nickel, but sensitive to mercury, silver, molybdenum and tungsten (4-10). We have reported the growth inhibition of A. ferrooxidans cells by mercury, silver, molybdenum and tungsten and clarified inhibition sites for these toxic metals (7-11). Toxic metals such as mercury are highly toxic for almost all organisms because they have a strong affinity for thiol groups in proteins (12, 13). The bacteria that are resistant to Hg2+ and/or organomercurial compounds have the ability to volatilize metal mercury (Hg°) from inorganic and organic mercurial compounds (12, 14-16). A wide range of Gramnegative and Gram-positive bacteria has mercury reductases that reduce Hg2+ with NADPH as an electron donor (13, 17-20). A. ferrooxidans cells have mercury reductase activity (9, 21-23) and the genes involved in the volatilization of mercury have been cloned and characterized in detail (21, 24-28). A. ferrooxidans SUG 2-2 was isolated as a mercury resistant strain among one hundred A. ferrooxidans strains isolated from natural environments (10). Strain SUG 2-2 has an ability to volatilize metal mercury from mercury-polluted wastewater and soil under acidic conditions in the presence of ferrous iron (29-31). We recently showed that A. ferrooxidans SUG 2-2 has not only NADPH-dependent mercury reductase activity but also a Fe2+-dependent mercury reductase activity in the cells (10) and cytochrome c oxidase purified from strain SUG 2-2 volatilizes mercury in the presence of Fe2+ (29). In this study, A. ferrooxidans strain MON-1 that is more resistant to mercuric chloride than strain SUG2-2 was isolated from a culture of SUG 2-2 to study the involvement of the iron oxidation enzyme system in mercury reduction of A. ferrooxidans. 2.

MATERIALS AND METHODS

2.1 Microorganisms, medium and growth conditions The iron-oxidizing bacteria used in this study were A. ferrooxidans AP19-3 (32) and A. ferrooxidans SUG 2-2 (10). Each strain was cultivated at 30°C under aerobic conditions in a Fe2+ medium (pH 2.5) containing 30 g of FeSO4.7H2O, 3 g of (NH4)2SO4, 0.5 g of K2HPO4, 0.5 g of MgSO4.7H2O, 0.1 g of KCl and 0.01 g of Ca(NO3)2 per liter The resting cells were prepared as follows. Each strain of iron-oxidizing bacteria was grown in 70 l of Fe2+ medium under aeration for one week. The culture medium was filtered with a Toyo no.2 filter paper to remove the bulk of the ferric precipitates and then centrifuged with a Hitachi 18pR-52 continuous-flow rotor at 15,000 × g and a flow rate of 200 ml/min. Harvested cells were washed three times with 0.1 M β–alanine-SO42- buffer (pH 3.0) before use. 2.2 Analysis of mercury volatilized from the culture medium of A. ferrooxidans A 50-ml culture flask with a screw cap contained 19 ml of Fe2+ medium (pH 2.5) supplemented with 1.0 or 5.0µM Hg2+ and 1 ml of an active seed culture of A. ferrooxidans. A small test tube containing 2 ml of a KMnO4 solution was inserted in the 1306

Molecular Biology and Taxonomy

50-ml culture flask to trap the Hg2+ volatilized form the culture medium. The KMnO4 solution used (100 ml) was composed of a 10-ml solution containing 0.6 g of KMnO4, 5 ml of concentrated H2SO4, and 85 ml of deionized water. After the culture medium was aerated by shaking at 30°C and 100 rpm, the concentration of Hg° trapped in the KMnO4 solution was measured by cold-vapor atomic absorption spectroscopy. 2.3 Mercury reductase activity The reaction mixture (2.5 ml) contained 50 mM sodium phosphate buffer (pH 7.0), 0.5 mM EDTA, 0.2 mM MgSO4.7H2O, 1 mM β-mercaptoethanol, 0.2 mM NADPH, 1.5 mg bovine serum albumin, 0.1 mM HgCl2, and the cytosol prepared from A. ferrooxidans strain SUG 2-2 and MON-1. The cytosol was prepared by centrifugation of a cell extract at 105,000×g for 1 h. After the reaction mixture was incubated at 37°C for 60 min, the reaction was started by the addition of NADPH. The activity was measured by the rate of oxidation of NADPH by monitoring the decrease of absorbance at 340 nm. 2.4 Fe2+-dependent mercury volatilization activity Each of several 50-ml flasks with screw caps contained a reaction mixture plus 2 ml of a KMnO4 solution described above. The gas phase was air, and the reaction mixture was shaken at 100 rpm at 30°. The reaction mixture used for the measurement of Fe2+mercury volatilization activity was composed of water acidified with sulfuric acid (20 ml), resting cells of A. ferrooxidans (1 mg of protein), mercuric chloride (0.7-7 µM), and ferrous sulfate (100 µmol). After the reaction mixture was aerated by shaking at 30°C and 100 rpm, the concentration of Hg° trapped in the KMnO4 solution was measured by coldvapor atomic absorption spectroscopy. 2.5 Protein content Protein content was determined by the method of Lowry et al. (33) with crystalline bovine serum albumin as the standard. 3.

RESULTS AND DISCUSSION

3.1 Isolation of strain MON-1 from the culture medium of A. ferrooxidans SUG 2-2 The processes to isolate strain MON-1 which is more resistant to mercuric chloride than the moderately mercury resistant A. ferrooxidans strain SUG 2-2 are shown in Table 1. Growth characteristics of mercury sensitive strain A. ferrooxidans AP19-3 (32) in a Fe2+ medium containing Hg2+ are also shown in the table as a reference. Mercury-sensitive strain AP19-3 cannot grow in a Fe2+ medium containing 0.7 µM Hg2+, but it can grow in Fe2+ medium containing 0.6 µM Hg2+ after a lag time of 24 days. In contrast, moderately mercury resistant strain SUG 2-2 grew in Fe2+ media containing 0.7 and 6µM Hg2+ with a lag time of 1 and 22 days, respectively. Eight times successive cultivations of SUG 2-2 in the medium markedly shortened the lag time to 4 days. From this adapted culture, strain SUP-1 was newly isolated as a single colony on a 1.0% gellan gum plate. Strain SUG 2-2 could not grow in a Fe2+ medium containing 10µM Hg2+. However, strain SUP-1 could grow in a Fe2+ medium containing 20µM Hg2+ with a lag time of 10 days. Five times successive cultivation of SUP-1 in a Fe2+ medium containing 20 µM Hg2+ shortened the lag time to 4 days. From this adapted culture, strain MON-1 was isolated as a single colony on a 1.0% gellan gum plate. Strain MON-1 could grow in a Fe2+ medium containing 20µM with a lag time of 2 days. Strain MON-1 slightly grew in a Fe2+ medium containing 40µM Hg2+, but could not grow in a Fe2+ medium containing 80 µM Hg2+. The 1307

Molecular Biology and Taxonomy

ability of strain MON-1 to grow rapidly in a Fe2+ medium containing 20µM Hg2+ was stably maintained after the strain was cultured successively in a Fe2+ medium without Hg2+. Table 1. Growth of A. ferrooxidans strains in Fe2+-medium containing mercuric ion Number of cultivation times

Con. of Hg2+ (µM)

2

4

6

8

10

12

14

16

18

20

22

24

26

28

AP19-3

1

0.6

-

-

-

-

-

-

-

-

-

-

-

+

++

+++

AP19-3

1

0.7

-

-

-

-

-

-

-

-

-

-

-

-

-

-

SUG 2-2

1

0.7

++

+++

+++

SUG 2-2 SUG 2-2

1 8

6.0 6.0

-

++

+++

-

-

-

-

-

-

--

+

++

+++

SUG 2-2

1

10

-

-

-

-

-

-

-

-

-

-

-

-

-

-

SUP-1

1

20

-

-

-

-

+

++

+++

SUP-1

5

20

-

+++

+++

MON-1

1

20

-

+++

+++

MON-1 MON-1 MON-1

1 5 1

40 40 80

-

+ + -

++ + -

+++ ++ -

-

-

-

-

-

-

-

-

-

-

Strain

Cultivation time (days)

3.2 Volatilization of metal mercury from a Fe2+ medium containing 1.0 or 5.0 µM Hg2+ by A. ferrooxidans cells The amounts of mercury volatilized from 20 ml of Fe2+ medium containing 1.0 or 5.0 µM of mercuric chloride were measured (Fig. 1). When cells of strains SUG 2-2 and MON-1 were cultured in a Fe2+ medium containing 1.0 µM of mercuric chloride for 2 days at 30°C, 10 and 19 nmol of mercury were volatilized from the culture medium. In contrast, no mercury was volatilized from the medium containing strain AP19-3 cells. When strains AP19-3, SUG 2-2 and MON-1 cells were cultured in a Fe2+ medium containing 5.0 µM of mercuric chloride for 2 days at 30°C, 0, 5 and 92 nmol of mercury were volatilized from the culture medium, indicating that mercury volatilization activity of strain MON-1 is much higher than that of SUG 2-2. 3.3 NADPH-dependent mercury reductase activity The activities of enzymes that are involved in detoxification of mercuric chloride were compared between A. ferrooxidans SUG 2-2 and MON-1 cells using cell extracts prepared from the two strains. At first, NADPH-dependent mercury reductase activity was measured by the rate of Hg2+-dependent oxidation of NADPH and NADPH-dependent mercury volatilization activity. Mercury reductase activity measured with oxidation of NADPH increased in proportion to the concentration of cell extracts from strains SUG 2-2 nd MON-1. Similar levels of activities were observed in both strains (Fig. 3A). The activities of NADPH-dependent volatilization of mercury were nearly the same in both strains (Fig. 3B). These results indicate that strains SUG 2-2 and MON-1 have a similar level of NADPH-dependent mercury reductase activity in spite of the difference in mercury resistance.

1308

Molecular Biology and Taxonomy

Hg2+ = 5.0 µ M

Hg0 volatilized (nmol)

Hg0 volatilized (nmol)

Hg2+ = 1.0 µ M

Time ( d )

Time ( d )

Figure 1. Volatilization of metal mercury from a Fe2+ medium containing 1.0 or 5.0 µM Hg2+ by A. ferrooxidans strains AP19-3, SUG 2-2 and MON-1 3.4 Fe2+-dependent mercury volatilization activity We recently showed that A. ferrooxidans SUG 2-2 cells contain has not only NADPH-dependent mercury reductase activity but also Fe2+-dependent mercury reductase activity and that the latter enzyme system markedly contributes to the mercury volatilization activity of strain SUG 2-2 (10). Fe2+-dependent mercury volatilization activity was measured in 20 ml of water acidified with sulfuric acid (pH 2.5) containing resting cells (1.0 mg) and 140 nmol HgCl2 by incubating for 3 h at 30°. The concentration of Hg2+ in the reaction mixture was 7 µM. The amounts of mercury volatilized from the reaction mixture were 4.7 nmol for strain SUG 2-2 and 27 nmol for strain MON-1, respectively (Fig. 4). Addition of 100 µmol of ferrous sulfate to the reaction mixture containing MON-1 cells enhanced the level of mercury volatilization activity by 2.5 fold. In contrast, an 1.2 fold of enhancement was observed in the case of SUG 2-2. Fe2+dependent volatilization of mercury was also found in SUG 2-2 cells mg) when incubated in 20 ml of acidic water containing 14 or 100 nmol HgCl2. However, in contrast to strain MON-1, the amount of mercury volatilized from the reaction mixture in the presence of 100 µmol Fe2+ markedly decreased in the presence of 7 µM Hg2+, suggesting that Fe2+dependent volatilization enzyme system of SUG 2-2 was inhibited by a high concentration of Hg2+. No remarkable enhancement of mercury volatilization was observed between strains SUG 2-2 and MON-1 when 100 µmol of ferrous sulfate was added to the reaction mixture containing 0.7 µM of Hg2+. The results obtained in this work strongly suggest that marked enhancement of Fe2+dependent mercury volatilization activity in strain MON-1 cells conferred on the strain the ability to grow rapidly in a Fe2+ medium containing 20µM Hg2+. We recently showed that cytochrome c oxidase purified from A. ferrooxidans strain SUG 2-2 can volatilize mercury using Fe2+ as an electron donor (29) and suggest that cytochrome c oxidase is involved in Fe2+-dependent mercury volatilization in A. ferrooxidans SUG 2-2 cells. Therefore, to further study the Fe2+-dependent mercury volatilization system of A. ferrooxidans strain MON-1 in relation to cell’s mercury resistance, purification and characterization of cytochrome c oxidase from strain MON-1 is now underway. 1309

Molecular Biology and Taxonomy

40

0.003

(B)

0.002 SUG 2-2 MON-1

0.001

Hg0 volatilized (nmol)

∆ A340 /min

(A)

30

0.05

SUG 2-2

10

0

0 0

MON-1

20

0

0.1

Protein (mg)

30

60

90

120

Time (min)

Figure 3. Mercury reductase activity measured in cell extracts of A. ferrooxidans strains SUG 2-2 and MON-1

Fe 2+= 5.0 mM

80

Hg 0 volatilized (nmol/3h)

Hg 0 volatilized (nmol/3h)

Fe 2+= 0 mM , Strain SUG 2-2

60 40 20 0

0.7

5

7

Conc. of Hg2+ (∆ ╩ M)

80

Strain MON-1

60 40 20 0

0.7

5

7

M) Conc. of Hg2+ (∆ ╩

Figure 4. Fe2+-dependent mercury volatilization by resting cells of A. ferrooxidans SUG2-2 and MON-1 REFERENCES 1. E. Torma. Adv. Biochem. Eng. 6 (1977) 1. 2. D. G. Lundgren and M. Silver. Ann Rev. Microbiol. 34 (1980) 263. 3. S R. Hutchins, M. S. Davidson, J. A. Brierley and C. L. Brierley. Ann. Rev. microbial. 40 (1986) 311. 4. O. H. Tuovinen, S. I. Niemela and H. G. Gyllenberg. Antonie van Leeuwenhoek 37 (1971) 489. 5. K. Imai, T. Sugio, T. Tsuchida and T. Tano. Agric. Biol. Chem. 39 (1975) 1349. 6. G. J. Olson, W. P. Iverson and F. E. Brinckman. Current Microbiol. 5 (1981) 115. 7. T. Sugio, T. Tano and K. Imai. Agric. Biol. Chem. 45 (1981) 2037. 8. K. Y. Ng, M. Oshima, R. C. Blake II and T. Sugio. Biosci. Biotechnol. Biochem. 61 (1997) 1523. 9. F. Takeuchi, K. Iwahori, K. Kamimura and T. Sugio. J. Biosci. Bioeng. 88 (1999) 387. 10. K. Iwahori, F. Takeuchi, K. Kamimura and T. Sugio. Appl. Environ. Microbiol. 66 (2000) 3823. 1310

Molecular Biology and Taxonomy

11. T. Sugio, H. Kuwano, A. Negishi, T. Maeda, F. Takeuchi and K. Kamimura. Biosci. Biotechnol. Biochem. 65 (2001) 555. 12. J. B. Robinson and O. H. Tuovinen. Microbiol. Rev. 48 (1984) 95. 13. A. Velasco, P. Acebo and N. Flores. Extremophiles. 3 (1999) 35. 14. O. Summer and S. Silver. Ann Rev. Microbiol. 32 (1978) 637. 15. S. Silver and T. K. Misra. Ann Rev. Microbiol. 42 (1988) 717. 16. S. Silver and M. Walderhaug. Microbiol. Rev. 56 (1992) 195. 17. S. Silver and L. T. Pheng. Ann Rev. Microbiol. 50 (1996) 753. 18. J. L. Schottel, A. Mandal, D. Clark, S. Silver and R. W. Hedges. Nature 251 (1974) 335. 19. J. L. Schottel. J. Biol. Chem. 253 (1978) 4341. 20. K. Babich, M. Engle, J. S. Skinner and R. A. Laddaga. Can. J. Microbiol. 37 (1991) 624. 21. G. J. Olson, W. P. Iverson and F. E. Brinckman. Current Microbiol. 5 (1981) 115. 22. G. J. Olson, and F. D. Porter. J. Bacteriol. 151 (1982) 1230. 23. J. E. Booth and J. W. Williams. J. Gen. Microbiol. 130 (1984) 725. 24. D. G. Rawlings and T. Kusano. Microbiol. Rev. 58 (1994)39. 25. T. Kusano, G. Ji, C. Inoue and S. Silver. J. Bacteriol. 172 (1990) 2688. 26. Inoue, K. Sugawara, T. Shiratori, T. Kusano and Y. Kitagawa. Gene 84 (1989) 47. 27. Inoue, K Sugawara and T. Kusano. Gene 96 (1990) 115. 28. Inoue, T. Kusano and M. Silver. Biosci. Biotech. Biochem. 60 (1996) 1289. 29. T. Sugio, K. Iwahori, F. Takeuchi, A. Negishi, T. Maeda and K. Kamimura. J. Biosci. Bioeng. 92 (2001) 44. 30. Takeuchi, K. Iwahori, K. Kamimura, A. Negishi, T. Maeda and T. Sugio. Biosci. Biotechnol. Biochem. 65 (2001) 1981. 31. Takeuchi, A. Negishi, T. Maeda, K. Kamimura and T. Sugio. J. Biosci. Bioeng. (2003) in press 32. S. Sugio, W. Mizunashi, K. Inagaki, and T. Tano. J. Bacteriol. 169 (1987) 4916. 33. O. H. Lowry, N. J. Rosebrough, A.L. Farr and R. J. Randall. J. Biol. Chem. 193 (1951) 265.

1311

1312

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Microbial diversity of various metal-sulphides bioleaching cultures grown under different operating conditions using 16SrDNA analysis P. d’Hugues, F. Battaglia-Brunet, M. Clarens and D. Morin BRGM, 3 Av. Claude Guillemin, BP 6009 45060 Orléans Cedex 2, France Abstract The microbial diversity of various metal sulphides bioleaching cultures was studied using the Single-Strand Conformation Polymorphism (SSCP) technique. Two sets of SSCP analyses were carried out on microbial populations subcultured at laboratory scale on five sulphidic substrates. The SSCP technique was also used to study a population grown on a cobaltiferous pyrite in different operating conditions (laboratory, pilot and industrial scales, batch and continuous modes, air-lift reactor and mechanically-agitated reactors). The 16S rDNA sequencing of the predominant organisms (seven strains out of eleven) revealed the presence of organisms, respectively affiliated to Leptospirillum ferrooxidans (two strains), Acidithiobacillus caldus, Acidithiobacillus thiooxidans, Acidithiobacillus ferrooxidans, Sulfobacillus thermosulfidooxidans and Sulfobacillus montserratensis. Whichever sulphide substrate used, organisms related to L. ferrooxidans and A. thiooxidans were always present amongst a microbial diversity of 2 to 7 bacterial strains. Depending on culture conditions or mineral characteristics, the occurrences of A. ferrooxidans, A. caldus, S. thermosulfidooxidans and S. montserratensis were more variable. In laboratory batch tests with pyrite, A. thiooxidans was significantly present at the beginning of the tests. Nevertheless, L. ferrooxidans-like organisms always appeared as the major contributor to the bioleaching efficiency, especially at industrial scale. The analysis of the biodiversity showed that the industrial culture contained strains that were also present in the cultures used for process development study. This work demonstrated that SSCP technique is a very convenient and reliable technique to monitor bioleaching populations. Keywords: bioleaching, biodiversity, SSCP, 16S rDNA 1.

INTRODUCTION Although bioleaching and biooxidation processes are an industrial reality, considerable work remains to be carried out on microbial ecology of these systems. Characterisation of mixed bacterial populations by classical microbiological methods has been limited due to the difficulties in plating, isolating and enumerating individual species and strains [1].

1313

Molecular Biology and Taxonomy

During the last decade, breakthrough in investigating microbial ecology was mainly achieved thanks to advances in molecular biology and phylogeny techniques [2,3]. Molecular tools developed to study microbial communities are essentially based on detection and analysis of 16S rDNA molecules [4,5]. Some of these molecular biology approaches were implemented in order to study iron- and sulphur- oxidisers from acidic mineral leaching environment. Goebel and Stackebrandt [1] published 16S rRNA sequence analysis of 33 strains of acidophilic bacteria obtained from both, acidic runoffs and laboratory scale bioreactors (batch and continuous-flow). The 16S-23S rDNA intergenic spacing method was used on a copper heap leaching system [6]. A PCR-based technique, using selected primers based on published 16S rRNA sequences, was implemented on a microbial population obtained from a silver-catalysed bioleaching column for chalcopyrite [7]. The bacteria present in commercial-scale biooxidation tanks, running to pre-treat gold-bearing arsenopyrite concentrate, were determined by implementing restriction enzyme patterns analysis [8]. Microbial populations, involved in the generation of acid mine drainage at Iron Mountain (California) were extensively studied by researchers of the University of Wisconsin using both clone-library generation using PCR and taxon specific hybridisation probes [9]. In 2002, BRGM published work for the monitoring of bioleaching operations using a recently developed PCR-based method, the Single-Strand Conformation Polymorphism (SSCP) technique [10]. It was applied on bioleached pulps sampled from two types of bioreactors, a bubble column and a mechanically stirred reactor. These various studies have provided new insights into microbial ecology of acidic mineral leaching environment. It was demonstrated that Acidithiobacillus ferrooxidans (previously Thiobacillus ferrooxidans) was not the main catalyst in biomining processes and AMD production. Other organisms involved in sulphide oxidation were gradually identified and studied: Leptospirillum ferrooxidans, Acithiobacillus thiooxidans, Acithiobacillus caldus, Sulfobacillus thermosulfidooxidans and Acidiphilium cryptum. The objective of the work presented in this paper was to investigate and compare the microbial diversity of various bioleaching cultures: (i) on 5 different mineral sulphide concentrates, and (ii) on a cobaltiferous pyrite in function of operating conditions (from laboratory scale up to industrial scale – Kasese Cobalt Company, KCC industrial plant located in Uganda). The originality of BRGM's studies lies on the implementation, for bioleaching samples, of the recently developed SSCP technique. The originality is also linked on the specific monitoring of identified bacterial strains in function of changes in operating conditions and on the first bioleaching industrial plant for base-metal recovery. 2.

MATERIAL AND METHODS

2.1 Bacterial inocula description The bacterial population was originally obtained from an enrichment culture of mine waters sampled by BRGM on a mining site, 15 years ago. Collinet-Latil [11] isolated strains of Thiobacillus ferrooxidans and Thiobacillus thiooxidans from this population. This first culture was then used as an inoculum for various studies carried out at BRGM on sulphide minerals (1 pyrite, 2 types of arsenopyrite, 1 chalcopyrite and 1 sphalerite). When these studies were achieved, the culture was maintained active by subculturing on the corresponding substrate. The same inoculum was used at the beginning of the "KCC Project". For this project, bioleaching of a cobaltiferous pyrite was studied in batch tests at laboratory scale and in 1314

Molecular Biology and Taxonomy

continuous operations with agitated tank reactors from 80 litres to 65m3 [12,13]. During continuous-bioleaching experiments, Leptospirillum-like bacteria associated with the initial rod-shape bacterial population were identified [14,15]. A culture originating from BRGM was used as the inoculum of the industrial operation that started on site in Uganda in 1998. 2.2 Description of the cultures used for SSCP studies Batch experiments were performed in (i) air-lift tubes of 200-ml effective capacity [16] (ii) a 22-litre mechanically stirred reactor [13], (iii) a 17-litre suspended solid bubble column [17]. The industrial samples were collected on site in 1,350 m3 stirred reactors running in continuous mode. The description of the various tests carried out is presented in Table 1. For the bioleaching experiment carried out at BRGM, 0Km [12,13] media was used and contained in g.l-1 (NH4)2SO4 (3.7), H3PO4 (0.8), MgSO4,7H2O (0.52), KOH (0.48). The composition of the industrial medium was very similar in N, P, and K concentrations but composed of fertilisers at industrial grades. The initial pH of batch culture was 2, and maintained above 1.1 in 20-litre reactors. It ranged from approximately 1.5 to 1.7 in industrial scale reactors. The temperature was maintained at 35°C for BRGM experiments and 40°C in the industrial reactors. The solids concentration (w/w) was 10% for the batch test in air-lift tubes and 20% for the other tests at larger scale. Table 1. Description of the bioleaching cultures used for SSCP analysis Culture identification

Operating system

Sulphide substrate

CHES-Cu

Air-lift tube (200 ml)

Chalcopyrite concentrate (Cu 28%)

CHES-Zn

Air-lift tube (200 ml)

Sphalerite concentrate (Zn 59%)

SALS

Air-lift tube (200 ml)

Arsenopyrite (55%), Pyrite (20%)

NIEJDA

Air-lift tube (200 ml)

Arsenopyrite (27%), Pyrite (26%)

KCC - AL

Air-lift tube (200 ml)

Pyrite (80%)

KCC - SSBC

Suspended-solids bubble column (25 l)

Pyrite (80%)

KCC - MAR

Mechanically agitated reactor

Pyrite (80%)

KCC - Bioco

Industrial stirred reactor (1350 m3)

Pyrite (80%)

Test identification Batch No1 Batch No2 Batch No1 Batch No2 Batch No1 Batch No2 Batch No1 Batch No2 Batch No1 Batch No2 Batch No1 Batch No2 Batch No3 Inoculum (2 L) BatchNo 2 (20 L) Continuous mode 3 Primary tanks and 1 secondary tank

Two sets of SSCP analyses were carried out on microbial populations subcultured at laboratory scale on five sulphide concentrates. The second SSCP analysis was carried out after a six-month subculturing period on each respective substrate. In all cases, the sampling was carried out at the end of the culture after 11 days. The SSCP technique was also used to study a population grown on a cobaltiferous pyrite in different operating conditions (laboratory, pilot and industrial scales, batch and continuous modes, air-lift reactor and mechanically agitated reactors). 1315

Molecular Biology and Taxonomy

2.3 Microbial communities analysis The strategy of the microbial diversity analysis comprised two parts with a first and common step being the extraction and purification of total genomic DNA of each sample tested. The details concerning the various experimental protocols used in this study were published previously [10]. They can be summarised as follows. The first part was based on the SSCP fingerprinting method and relies on the different electrophoretic mobility (in non-denaturing gel) of single strand DNA molecules according to the difference in their secondary structure. Thus, DNA fragments of the same size but with a single base modification can be separated. The SSCP-PCR amplification and SSCP electrophoresis of amplified target DNA (variable region V3 of 16S rRNA gene) was carried out. Proportions of each amplified V3 region (peak) on sample SSCP profiles varied during SSCP monitoring. The relative height of the peaks obtained was likened to the proportion of DNA from each species present in the PCR product. As it is known that potential biases can arise from DNA extraction, PCR amplification and 16S rDNA gene copy number, several DNA extractions and SSCP analyses were performed on the same bioleaching sample. Those showed good reproducibility in the ratio between specific peak heights. The second part was based on the 16S rDNA inventory of the population after screening, sequencing and identification of clones. The screening of clones to be used for total 16S rDNA sequence analysis was carried out using both restriction fragment length polymorphism technique (RFLP) and individual SSCP pattern. In order to assign peaks on SSCP patterns of a bacterial community, the V3 region of the different OTUs (Operational Taxonomic Unit) of the selected clones was also analysed by SSCP. The identification of SSCP peaks was realised thanks to the sequencing of the corresponding clones. The total sequence of 16S rDNA was used for sequence analysis. Sequences were compared with sequences available in databases (GenBank and RDP). The nucleotide sequence data reported in this work will appear in the GenBank nucleotide database under accession numbers AF460981 to AF460987. 3.

RESULTS AND DISCUSSION

3.1 Inventory of the bioleaching populations The first inventory of a BRGM bioleaching culture was performed using a sample from an inoculum (prepared in a 2-litre mechanically agitated reactor)[10]. The other inventories were performed on the following bioleaching cultures when SSCP patterns revealed the presence of an unidentified representative peak. The results of the 4 inventory tests carried out are presented in Table 2. The different clones identified are classified in function of their electrophoretic mobility. Thanks to these inventory studies, 7 peaks out of the 11 peaks observed on the various SSCP tests could be assigned to 1 OTU (representing 1 bacterial strain represented by 1 clone). Two peaks with different electrophoretic mobilities were found in SSCP patterns of the first sample tested by SSCP. A total of 53 clones were then screened by restriction fragment length polymorphism technique (RFLP) using HaeIII. Two RFLP patterns were found, 47 clones corresponded to one pattern and 6 to the other. Two clones, one for each pattern, were tested by individual SSCP. The two different OTUs represented by clones K01 and K13 showed an electrophoretic mobility that could be assigned to the two SSCP peaks observed in the sample. 16S rDNA sequences of K01 and K13 were compared to referenced sequences from Genbank and RDP. For K01, the closest 16S rDNA sequences were uncultivated bacteria from natural acidic environments (clone OS7) and L. 1316

Molecular Biology and Taxonomy

ferrooxidans strain C-Lf30A. The closest sequences for K13 were the uncultured Acidithiobacillus sp. V1 and A. thiooxidans KCTC 8929P. A third strain, not detected in the initial inoculum, was revealed by SSCP and identified from DNA extracted in the last sample of test KCC-SSCB (Batch No1). Clone K55, with an SSCP mobility corresponding to the unidentified bacterium, had a 16S rDNA sequence closely related to the uncultured bacterium detected in samples (BU138) from acid mine environments. The most similar 16S rDNA sequence of isolated organism diverged from more than 3% from clone K55 and was identified as S. thermosulfidooxidans str. AT-1 (DSM 9293). When applied on industrial samples, SSCP patterns revealed the presence of 3 unidentified peaks. According to the semi-quantitative approach, one peak present in all samples probably corresponded to the major organism of the industrial samples. The SSCP mobility of the corresponding selected clone (KCC-IND) was very closely related to the one of K01 (1% divergence), identified as closely affiliated to L. ferrooxidans. The closest related 16S rDNA sequence was the one of L. ferrooxidans strain C-Lf30A. The SSCP tests carried out on the laboratory scale cultures grown on various substrates showed SSCP patterns with 3 peaks sometimes observed as related to major organisms in the samples but that could not be assigned to any of the already identified clones. Clones corresponding to these unidentified 16S rDNA sequences were obtained from DNA extracted in the more appropriate cultures. Total 16S rDNA sequences of 3 representative clones, P3-5, P5-10 and P6-2, showed that there were respectively closely related to S. montserratensis, A. caldus and A. ferrooxidans. Table 2. SSCP peaks identification SSCP Peak identification

Culture and inventory identification

Peak no 3 (S.t)

KCC-SSBC (inventory n°2)

Peak No 4 (Sm) Peak No 5 (Lf-1)

NIEJDA (inventory n°4) KCC-IND (inventory n°3)

Clone for Genbank Closest related clone and/or identified SSCP-Peak Accession species assignation Number (% divergence) Uncultured bacterium BU138 (0.2%) AF46098 K55 4 S. thermosulfidooxidans str. AT-1 (DSM 9293) - (3.3%) P3-5

AF46098 5

S. montserratensis L15 - (0.9%)

KCC-IND

None

L. ferrooxidans str. C-Lf30A (DSM 9468) - (0.4%)

Peak No 6 (Lf-2)

KCC-MAR (inventory n°1)

K01

AF46098 1

Peak No 8 (A.c)

CHES-Zn (inventory n°4)

P5-10

AF46098 6

Peak No 9 (At)

KCC-MAR (inventory n°1)

K13

AF46098 3

Peak No 10 (A.f)

SALS (inventory n°4)

P6-2

AF46098 7

Clone OS7 - (0.2%) L. ferrooxidans str. C-Lf30A (DSM 9468) - (0.4%) A. caldus KU DSM8584 - (0.2%) Uncultured Acidithiobacillus sp. V1 (2.6%) A. thiooxidans KCTC 8929P - (2.9%) A. ferrooxidans NFe4 (0.3%)

1317

Molecular Biology and Taxonomy

3.2 Influence of sulphide substrates on microbial biodiversity The SSCP patterns observed on the various bioleaching cultures revealed the presence of 9 different peaks that can be associated to the presence of 9 different bacterial strains (Table 3). The assignment of SSCP peaks to their respective 16S rDNA sequences was carried out for 5 of them. All SSCP patterns carried out on the first set of tests and some of them for the second set of tests were doubled (from the extraction step) and showed a good reproducibility in the ratio between specific peak heights (data not shown). The identification of SSCP peaks was undertaken if they were present in large proportion in any of the samples tested or if they were widely present in the samples. Table 3. SSCP patterns for cultures carried out on 5 different sulphide concentrates Culture

Test

pH

CHES-Cu

AL1

1.28

4

AL2

1.34

4

AL1

1.48

AL2

CHES-Zn SALS NIEJDA KCC-AL

1

2

3 S.t

SSCP peaks identification (%) 4 5 6 7 8 S.m Lf.1 Lf.2 A.c 1

9 A.t

29

25

40

37

16

43

4

34

57

1.53

25

31

44

AL1

1.75

1

AL2

1.93

1

AL1

1.15

4

AL2

0.9

2

AL1

1.10

12

AL2

0.8

12

4

1

22

2

33

9

1

45 79

11 1

1 76

75 1

81 2

22

10 A.f

52 17

2

38

1

9

The SSCP peaks No 5 and No 9 were present in all cultures and corresponded respectively to clones KCC-IND (closely related to L. ferrooxidans) and K13 (closely related to A. thiooxidans). The proportion of clone P3-5 (peak No 4, closely related to S. montserratensis) was important on the first batch test with concentrate Niejda (pyritearsenopyrite), but disappeared from SSCP pattern after 6 months of subculturing. It seems that it was essentially replaced by an important development of clone KCC-IND. Clone P5-10 (peak No 8, closely related to A. caldus) was only present on chalcopyrite and sphalerite concentrates and remained present after 6 months of subculturing. Clone P6-2 (Peak No 10, closely related to A. ferrooxidans) was one of the major organisms of the culture on arsenopyrite (Sals), but disappeared from SSCP pattern after 6 months of subculturing. When looking at the 10 SSCP patterns carried out on the 5 substrates, the presence of a maximum of 6 strains and of a minimum of 3 strains can be observed. According to the semi-quantitative information given by the peak height, 2 or 3 bacterial strains seemed predominant in the culture, while the others occuring in very small proportions. The 2 SSCP patterns carried out on both on chalcopyrite and sphalerite demonstrated a good stability of the culture composition for the predominant organisms. The composition of cultures grown on the 2 arsenopyrite concentrates was different. For the same concentrate, they were also unstable when the SSCP patterns were compared after 6 months of subculturing. The difference of composition between the 2 arsenopyrite concentrates can 1318

Molecular Biology and Taxonomy

be related to the presence of carbonate in one of them with a direct effect on pH trend. With a pH maintained above 1.7, the presence of an organism affiliated to A. ferrooxidans could be observed. The presence of this organism was not stable throughout the test. It seemed to be more or less replaced by an organism related to S. montserratensis. With the second arsenopyrite concentrate, a significant decrease of pH was always observed. The organisms related to S. montserratensis disappeared from SSCP patterns between the two batch tests. It was replaced in the second test, where the pH dropped below 1, by the emergence of an organism related to L. ferrooxidans. Below pH 1, the organism related to L. ferrooxidans seemed also to become predominant to the detriment of the organism related to A. thiooxidans. This result is particularly marked on the SSCP patterns carried out on KCC pyrite concentrate. 3.3 Influence of operating conditions on microbial biodiversity SSCP technique was used to study the changes in composition of a bacterial mixed culture grown on the KCC sulphide concentrate under different operating conditions (Table 4). The results obtained in small-scale air-lift tubes (presented in the previous chapter) were obtained from samples collected at the end of the culture. In order to study the evolution of the bacterial consortium during batch oxidation, the same approach was implemented on pulp samples collected during the whole length of batch cultures. Table 4. SSCP patterns on bioleaching cultures carried out on KCC concentrate Test description SSBC Batch No1 Inoculum 8 days 15 days 21 days 23 days (inoculum batch 2) SSBC Batch No2 7 days 17 days MAR Batch No2 7 days 17 days 20 days (inoculum SSBC SSBC Batch No3 12 days 19 days KCC Industrial reactors Bioco primary reactors Sampling Campaign 1

1

2

1

2

1

2

1

2

3

1

2

3 4 6 4 3

4 5

2 5

Bioco primary reactors Sampling Campaign 2 Bioco primary reactors Sampling Campaign 3 Bioco secondary reactor Sampling Campaign 1

3 0 15 20 18 22 3 2 37 3 1 9

8 6

SSCP peak proportions (%) 4 5 6 7 8 100 49 61 74 74 4 5 6 7 8 35 49 7 4 5 6 8 47 86 100 7 4 5 6 8 75 86 4 5 6 7 8 96 94 93 94 94 90 90 86 6

9 0 36 19 8 4 9 63 14 9 52 5 0 9 25 14 9 0 0 3 3

10

11

10

11

10

11

10

11

10

11

3 2

1319

Molecular Biology and Taxonomy

When compared with results obtained in air-lift tubes, the SSCP patterns obtained from cultures carried out in larger devices (20 litres - 25 litres) showed 1 common SSCP peak and 2 different ones. Peak No 9, associated to the K13 clone (A. thiooxidans-like organism) is still widely represented. In contrast, peaks No 6 and No 3 were not present in SSCP patterns carried out on KCC cultures in air-lift tubes. The main differences between the two types of tests were solids concentration and agitation-aeration system. Both pH trend evolution and temperature were comparable. The study of the bioleaching community dynamics showed that whereas L. ferrooxidans-like organism (K01) and A. thiooxidans-like organism (K13) were both almost equally represented in the first days of the culture, L. ferrooxidans-like organism became the predominant organism towards the end of the bioleaching process. In contrast to K01 and K13, the behaviour of K55 (Sulfobacillus-related organism) was significantly different from one test to another. Thus, K55 was not detected during the last batch test using the SSCB. However, SSCP monitoring revealed that the development of K55 was favoured in the column bioreactor as compared to the mechanically stirred reactor (batch 2 carried out in parallel with the same inoculum). The proportion of K55 usually increased at the end of the test. Using the same PCR-SSCP approach, a more detailed analysis was carried out on the same samples to determine the proportions of each organism attached on the solid particles or freely suspended in the medium [10]. In the liquid, the A. thiooxidans-related bacteria were dominant during the early phase of the batch, then supplanted by the L. ferrooxidans related bacteria. L. ferrooxidans related organisms were always in the majority on the solids. S. thermosulfidooxidans-related bacteria generally occurred more on the solids than in free suspension in the liquid phase. Whatever the changes in population composition and the changes in operating condition (reactor type, air flow-rate and inoculum) it is important to note that the pyrite oxidation kinetics were not significantly affected [17]. For this study, some samples had been also collected in the industrial KCC bioreactors. The temperature was 40°C on average and the pH maintained between 1.5 and 1.9 by limestone addition. The SSCP patterns showed the presence of 6 distinctive peaks. They were all present in BRGM laboratory scale studies, and not only on KCC pyrite cultures. L. ferrooxidans-like organism corresponding to peak No 5, seemed to be always the largely dominant organism of the industrial culture, both in the primary and secondary stages. Surprisingly, this peak No 5 was detected in all BRGM air-lift tests, including the one on KCC pyrite but absent from the batch tests carried out both in mechanically agitated reactor and column reactor. All the other organisms present in the industrial culture represented a very small proportion of the total population. Organisms corresponding to peak No 1 and peak No 2 could not be identified by a corresponding clone sequence analysis. Peak No 3 corresponding to clone K55 (Sulfobacillus - like organism) was also detected on KCC cultures carried out both in mechanically agitated reactor and column reactor. Peak No 8 assigned to an OTU corresponding to an A. calduslike organism was largely present in chalcopyrite and sphalerite samples. Peak No 9 corresponding to an A. thiooxidans-like organism was present in all samples tested by SSCP. 3.4 Discussion on microbial diversity in bioleaching The 16S rDNA sequencing of the predominant organisms (seven strains out of eleven) detected by the SSCP analysis, revealed the presence of organisms closely related to L. ferrooxidans (two strains), A. caldus, A. thiooxidans, A. ferrooxidans, S. thermosulfidooxidans and S. montserratensis. Considering that these tests were undertaken 1320

Molecular Biology and Taxonomy

on 5 different sulphide concentrates, the important biodiversity of BRGM cultures was not really surprising. On the other hand, the presence of the same representative strains was more surprising. It can be assumed that through out the different subculturing and tests carried out, the composition of the original inoculum changed by crossed contamination and selection of the more appropriate strains which where naturally present in the original sulphide substrates tested. This result is important for the question highlighted in a recent review on the importance of microbial ecology in the development of new mineral technologies [2]. Is the acidophilic population responsible for mineral oxidation in industrial bioleaching operation the one, which are originally associated with the mineral deposit itself? From the test carried out on KCC industrial operation, it was shown that the industrial population was composed of microorganisms that were also present in laboratory scale cultures on different concentrates. As biodiversity assessment on the KCC deposit itself would required extensive studies, it was not possible at that stage to determine whether some of them were initially present on KCC concentrate and established themselves on other concentrates or conversely. The presence of similar strains in industrial and laboratory scale cultures gives sense to any extended process optimisation studies carried out prior to the implementation by the extractive mineral industry. Nevertheless, when the same substrate is treated in similar conditions by batch or continuous culture, the differences in terms of maximum oxidation rate are always significant. On KCC project, the gain by running continuous bioleaching was evaluated to a minimum of 30%. It shows the importance of running continuous bioleaching tests when the objective is to evaluate performances in view of an application to real scale. The difference in oxidation rates between batch and continuous cultures might be related to the difference of composition and dynamics of the bioleaching populations. SSCP tests carried out in this study on batch cultures showed that both Acidithiobacillus-like organisms and Leptospirillum-like organisms coexisted during the sulphides oxidation process. When looking at their presence as pyrite oxidation progressed, an increasing contribution of Leptospirillum-like organisms was observed. In industrial continuous culture, Leptospirillum-like organisms were the dominant organism. Other authors already observed this phenomenon on industrial units for arsenopyrite concentrate biooxidation [3]. The predominance of Leptospirillum-like organisms over both A. thiooxidans and A. caldus in industrial cultures or at the end of the bioleaching process is surprising as they apparently do not compete for the same substrate. L. ferrooxidans-like organism was also found as the major solid coloniser on pyrite, where as sulphur-oxidizing Acidithiobacillus-like organisms are less represented on the solids fraction [10]. From these various observations on microbial community, it could be assumed that pyrite oxidation mainly results from an indirect oxidation by ferric iron, located at the interface between the pyrite surface and the attached Leptospirillum-like organisms. The role of Leptospirillum-like organism would be the subsequent re-oxidation of ferrous iron produced by pyrite oxidation. The development of A. thiooxidans or A. caldus would be of less (if not any) importance on the bioleaching efficiency. These two organisms would grow thanks to the elemental sulphur and the reduced sulphur compounds produced by the Leptospirillum sp. driven pyrite oxidation. Then, the question of whether or not their presence is of any importance on sulphide oxidation can be asked, especially when looking at industrial scale SSCP patterns where they were sometime not even detectable. In batch tests, the relatively smooth selective pressure would have, as a consequence, the development of both Leptospirillum-like organisms, the indirect but only pyrite oxidisers and of organisms able to use the products of the pyrite oxidation. At the end of a batch test or in continuous culture condition, with higher selective pressures, Leptospirillum-like organisms, the only adapted to the main substrate (ferrous iron from 1321

Molecular Biology and Taxonomy

pyrite oxidation) would then outgrow the sulphur oxidisers. If this statement was confirmed, then the question of another acid production mechanism and elemental sulphur transformation (no accumulation of S° was observed in industrial scale reactors) would have to be considered. The idea of Leptospirillum-like organism as the only pyrite oxidiser by indirect mechanism makes sense when looking at pyrite-bearing sulphides. The oxidation mechanisms of chalcopyrite and sphalerite type of concentrates might be comparable. The development of A. thiooxidans or A. caldus would be a consequence of a prior oxidation process by an iron oxidiser. The presence of members of the genus Sulfobacillus in the various cultures appeared as more uncertain. As they can oxidise either reduced sulphur compounds or ferrous, S. thermosulfidooxidans and S. montserratensis could be competing with both, A. thiooxidans and A. caldus, for sulphur and with L. ferrooxidans and A. ferrooxidans, for ferrous iron. However, experimental data on both pyrite and arsenopyrite suggested that Sulfobacillus-like organisms were more in competition with iron oxidisers than with sulphur oxidisers. On pyrite, the development of Leptospirillum-like organism to the detriment of S. thermosulfidiooxidans-like bacteria seemed to be related to a better resistance to more constraining agitation-aeration operating conditions. At laboratory scale, in air-lift tube, the development of Leptospirillum-like organism to the detriment of S. thermosulfidiooxidans-related bacteria seemed to be related to its ability to grow in very acidic (pH below 1) environment. 4.

CONCLUSION The use of molecular biology tools, such as SSCP, are beneficial for both ecological and industry-focused research into acidophilic microbiology, i.e.: (i) academic work on sulphide oxidation mechanisms and microbial interactions; and (ii) bioleaching processes studies from laboratory scale development up to monitoring of industrial operation. A more specific monitoring of iron and sulphur oxidising populations helps for the debate still open on the respective importance of direct and indirect mechanisms in the oxidation of sulphide minerals. The molecular techniques provide new insights on microbial interaction phenomena occurring within a consortium. More information will be then available to understand whether mixed populations are really more efficient than corresponding pure cultures. It will be also possible to determine whether microorganisms responsible for mineral oxidation in industrial bioleaching operations are those associated with the mineral deposit (endemic strains) or/and those used at laboratory-scale for process development steps. This phenomenon, like the difference observed on population dynamics between batch or continuous cultures is of great importance when implementing process development studies. In the future, a possible design of specific microbial consortia could be envisaged thanks to a better knowledge of acidophilic environments. Even though the first tests using different molecular biology approaches led to quite promising, reproducible and coherent results, the work to be carried out on the ecology of iron- and sulphur- oxidisers remains considerable. First of all, because whatever the technique chosen, they always generate a panel of bias [4,5,9]. Studies of population sampled in similar environments have to be crosschecked with different techniques in order to establish a reliable picture of the reality. Furthermore, whereas numerous studies on the ecology of these systems were focused on the identification of the various strains present, only a few of them took into account the microbial population dynamics.

1322

Molecular Biology and Taxonomy

ACKNOWLEDGEMENTS This paper is published with the permission of BRGM as scientific contribution No.02499. REFERENCES 1. 2. 3. 4. 5.

Goebel, M.B. and Stackebrandt, E., Appl. Environ. Microbiol., 60, (1994), 1614. Johnson D.B., Hydrometallurgy, 59, (1999), 147. Rawling, D.E., Tributsh, H. and Hansford G.S., Microbiology, 145, (1999), 5. Head, I.M., Saunders, J.R. and Pickup, R.W., Microbial Ecology, 35, (1998), 1. Dabert, P., Delgenès, J.P., Moletta, R. and Godon J.J., Re/Views in Environmental Science & Technology 1, (2002), 39. 6. Pizzaro J, Jedlicki E, Orellana O, Romero J, Espejo RT, Appl Environ Microbiol 62, (1996), 1323. 7. De Wulf-Durand P, Bryant LJ, Sly L.I., Appl. Environ. Microbiol. 63, (1997) 2944. 8. Rawling, D.E., Biohydrometallurgical Processing, Volume II, University of Chili, (1995), 9. 9. Edwards, K.J., Goebel, B.M., Rodgers, T.R., Schrenk, M.O., Gihring, T.M., Cardona, M.C., Hu, B., McGuire, M.M., Hamers, R.J., Pace, N.R., Banfield, J.F, Geomicrobiology Journal, (1999), 16, 155. 10. Battaglia-Brunet, F., Clarens, M., d’Hugues, P., Godon, J.J., Foucher, S., and Morin, D., Appl. Microbiol. Biotechnol., 60, (2002), 206. 11. Collinet-Latil M.N., Morin D., Antonie van Leeuwenhoek 57, (1990), 237. 12. Morin, D., Ollivier, P., and Hau, J.M., Waste Processing and Recycling in Mineral and Metallurgical Industries, II, The Canadian Institute of Mining, Metallurgy and Petroleum (1995), 23. 13. d'Hugues P., Cézac P., Cabral T., Battaglia F., Truong-Meyer X.M. and Morin D., Minerals Engineering, Vol. 10, n° 5 (1997), 507. 14. Battaglia, F., Morin, D., Garcia, J.L., and Ollivier, P., Antonie Van Leeuwenhoek 66, (1994) 295. 15. Battaglia-Brunet F., d'Hugues P., Cabral T., Cézac P., Garcia J.L. and Morin D., Minerals Engineering, Vol. 11, n°2, (1998) 195. 16. Battaglia, F., Morin, D., and Ollivier, P., J. Biotechnol. 32, (1994) 11. 17. Foucher, S., Battaglia-Brunet, F., d'Hugues, P., Clarens, M., Godon, J.J., and Morin, D., Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, Part A, Elsevier, (2001), 3.

1323

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Molecular ecology of the Tinto River, an extreme acidic environment from the Iberian Prytic Belt E. González-Torila, E. Llobet-Brossab, E.O. Casamayorb, R. Amannb, R. Amilsa,c a

Centro de Biología Molecular (CSIC-UAM), Cantoblanco, Madrid 28049, Spain Max Planck Institut for Marine Microbiology, Celsiusstraβe 1, D-28359 Bremen, Germany c Centro de Astrobiología (CSIC-INTA), Torrejón de Ardoz, Madrid 28850, Spain

b

Abstract Complementary molecular ecology techniques have been used to characterize the Tinto River, an extreme environment with a rather constant acidic pH all along its course (mean pH 2.3) and high concentration of heavy metals (Fe, Cu, Zn, As and Cr). Comparative sequence analysis of amplified 16S rRNAs and 16S rRNA genes resolved by denaturing gradient gel electrophoresis (DGGE) allowed members of four bacterial genera: Acidithiobacillus, Leptospirillum, Acidiphilium and Ferrimicrobium, and two archaeal genera: Ferroplasma and Thermoplasma to be identified at different sampling stations along the river. The quantitative evaluation of the prokaryotic diversity using in situ hybridization with fluorescence labeled rRNA-targeted oligonucleotides (FISH) showed that the bulk prokaryotic biomass of the water column, up to 80%, corresponded to members of three bacterial species: Acidithiobacillus ferrooxidans, Leptospirillum ferrooxidans and Acidiphilium spp., all of them related to iron metabolism. Taking into consideration the characteristics of the habitat, the physiological properties and spatial distribution of the identified microorganisms, a model for the Tinto ecosystem based on the iron cycle is advanced and its biohydrometallurgical implications discussed. 1.

INTRODUCTION The Tinto River, a 100 km-long river in Southwestern Spain, is an unusual ecosystem due to its acidity and high concentration of metallic cations in solution (1, 2). The river springs up in the core of the Iberian Pyritic Belt (IPB) at Peña de Hierro (Iron Mountain), and flows into the Atlantic Ocean at Huelva (Fig. 1). The extreme conditions of the Tinto ecosystem are the product of the very active chemolithotrophic metabolism of microorganisms growing on the rich sulfidic mineral ores of the IPB and not, as formerly believed, the result of industrial mining contamination (3, 4, 5, 6). The existence of massive laminated iron bioformations (iron stromatolites) corresponding to old terraces of the river, predating the oldest mining activity reported in the area and similar to the laminar structures that are being currently formed in the river, is considered a strong argument in favour of a natural origin of the river (7). In spite of the extreme conditions of acidity and heavy metal content, the Tinto ecosystem holds a high level of unexpected eukaryotic diversity (2, 8). 1325

Molecular Biology and Taxonomy

It was not surprising that the results obtained using conventional microbiological methods, isolation from enrichment cultures and phenotypic characterization, showed the presence in the Tinto ecosystem of sulfur and iron-oxidizing microorganisms, such as Acidithiobacillus ferrooxidans (formerly Thiobacillus ferrooxidans), in rather high numbers due to the acidic characteristics of the habitat (1, 2). Since its isolation in the early fifties At. ferrooxidans has been considered the principle agent of acid mine drainage (AMD), a problem of environmental concern in metal mining (9). The development of industrial bioleaching, facilitating the metabolism of acidic chemolithotrophic microorganisms, focussed attention on this microorganism. As a result a number of rather contradictory reports have been produced during the search for the basic mechanisms involved in the process (10). Recently it has been proposed that the properties of the substrate (sulfide minerals) rather than the microorganisms can explain most of the contradictory information in the field, giving to ferric iron a key role in the oxidation mechanism of sulfidic ores (11). Accordingly, molecular ecology studies have challenged the role of At ferrooxidans in bioleaching processes showing that strict iron-oxidizing microorganisms, like Ferroplasma spp. and Leptospirillum spp., were mainly responsible for AMDs generation and the main microbial populations in different bioleaching processes (12, 13, 14). The recent introduction of molecular biology techniques into the field of microbial ecology has produced a significant advance (15, 16), especially in poorly characterized environments, such as extreme ecosystems. We present in this work a succinct report of the prokaryotic diversity of the Tinto River, a chemolithotrophically sustained ecosystem, using molecular ecology techniques (for a full report see reference 17). The combination of molecular ecology with the physiological characterization of isolated microorganisms has provided sufficient information to generate a geomicrobiological model of this unusual ecosystem of interest in different fields (18), particularly in biohydrometallurgy (19). 2.

MATERIALS AND METHODS

2.1 Sampling and analysis of physico-chemical parameters Samples were collected in triplicate from different sampling stations along the river (Fig. 1) in June and October of 1999 and May of 2000. Total content of metals was measured by TXRF and ICP-MS. Sulfate concentrations were determined by a turbidimetric method, (2) and ferrous iron by a colorimetric method (2). Conductivity, pH and redox potential were measured in situ using specific electrodes. A Crison 506 pH/EHmeter was used to measure redox potential and pH, and an Orion-122 conductivity-meter for conductivity. 2.2 Nucleic acid extraction and cell fixation Samples for DNA extraction were collected into one liter bottles and kept on ice until filtered through nitrocellulose Millipore filters (0.22 µm). Filters were stored at -20ºC until processed. Nucleic acid extraction was performed as described in (17). Samples for FISH were immediately fixed with 4% formaldehyde-minimal Mackintosh media and processed as described in (17).

1326

Molecular Biology and Taxonomy

Figure 1. Geographical location of the Tinto River and the different sampling sites 2.3 PCR amplification, DGGE, sequencing and phylogenetic analysis PCR amplification of 16S rRNA gene fragments between E. coli positions 341 and 907 for the domain Archaea (20, 21) and between E. coli position 344 and 907 for the domain Bacteria (22), reverse transcription of 16S rRNA and amplification of the 16S rRNA gene, denaturing gradient gel electrophoresis (DGGE), excision of bands, and reamplification were performed as previously described (22). Taq Dyedeoxy Terminator Cycle Sequencing kit (Applied Biosystems, Forster City, USA) was used to sequence the 16S rRNA gene fragments. Sequencing reactions were run on an Applied Biosystems 373S DNA sequencer. New partial sequences were added to an alignment of about 8,500 homologous 16S rRNA primary structures (23) by using the aligning tool of the ARB package (24). Aligned sequences were inserted within a stable tree using the parsimony tool ARB that enables reliable positioning of new sequences without alignment (25). Sequences of DGGE bands and 16S rRNA gene clones were initially compared with references sequences contained in the EMBL Nucleotide Sequences Database using the BLAST program and subsequently aligned with 16S rRNA reference sequences in the ARB package (http://www.mikro.biologie.tu-muenchen.de) (24). DGGE partial sequence dendrograms were obtained using the parsimony tool DNAPARS included in the ARB software.

1327

Molecular Biology and Taxonomy

2.4 Cell counts and FISH Hybridization and microscopy counting of hybridized and 4',6'-diamidino-2phenylindole (DAPI)-stained cells were performed as described previously (15). Mean values were calculated by using ten to twenty randomly chosen fields for each filter section, which corresponded to 800 to 1,000 DAPI-stained cells. Counting results were corrected by subtracting signals observed with the control probe NON338. Probes used in this work are listed in Table 1. Hybridization conditions are described in (17). Cy3 labeled probes were synthesized by Interactiva (Ulm, Germany) and by Quiagen (Barcelona, Spain). Table 1. Fluorescence labeled oligonucleotide probes used for in situ hybridization experiments Probe EUB338 ALF968 ACD638 BET42a GAM42a THIO1 ACT465a ACT465b ACT465c ACT465d NTR712b LEP154 LEP634 LEP636 SRB385 ACM1160 ARCH915 FER656 TMP654 NON338

3.

Target 16S 16S 16S 23S 23S 16S 16S 16S 16S 16S 16S 16S 16S 16S 16S 16S 16S 16S 16S -----

Sequence (5’ to 3’) GCT GCC TCC CGT AGG AGT GGT AAG GTT CTG CGC GTT CTC AAG ACA ACA CGT CTC GCC TTC CCA CTT CGT TT GCC TTC CCA CAT CGT TT GCG CTT TCT GGG GTC TGC GTC AAC AGC AGC TCG TAT GTC AAC AGC AGA TCG TAT GTC AAC AGC AGA TTG TAT GTC AAT AGC AGA TTG TAT CGC CTT CGC CAC CGG CCT TCC TTG CCC CCC CTT TCG GAG AGT CTC CCA GTC TCC TTG CCA GCC TGC CAG TCT CTT CGG CGT CGC TGC GTC AGG CCT CCG AAT TAA CTC CGG GTG CTC CCC CGC CAA TTC CT CGT TTA ACC TCA CCC GAT C TTC AAC CTC ATT TGG TCC ACT CCT ACG GGA GGC AGC

Specificity Bacteria domain α Proteobacteria Acidiphillium spp. β Proteobacteria γ Proteobacteria Acidithiobacillus spp. Group a Acidithiobacillus spp. Group b Acidithiobacillus spp. Group c Acidithiobacillus spp. Group d Acidithiobacillus spp. Nitrospira group Group b L. ferriphilum Group a Leptospirillum spp. Group c L. ferrooxidans δ Proteobacteria Acidimicrobium spp. Archaea domain Ferroplasma spp. Thermoplasma spp., Picrophilus spp. Negative control

RESULTS AND DISCUSSION

3.1 Physico-chemical characterization of the samples An important characteristic of the Tinto River is its constant acidic pH, a direct consequence of the strong buffer capacity of ferric iron (2, 26). The mean pH value measured in the different samples used in this work was 2.4, with sampling site RT3 as the only exception, where the pH was always higher (mean 4.7), probably as a consequence of the lack of iron in solution to buffer the stream. The mean iron concentration for the different sampling sites was 4.9 g/l, although its concentration at the origin could be as high as 20 g/l. The concentration of iron decreases along the river as a consequence of its precipitation due to the dilution effect produced by tributaries or rain (19). A mean concentration of 9.2 g/l was found for total sulfur, most of it corresponding to sulfate (2). Redox potentials were high and relatively constant along the river, between +421 and +608 mV, except for station RT3 in which the values were always much lower. The dissolved oxygen varied in the different sampling stations depending on the hydrological regime of the river (7). Interestingly enough, anoxic conditions were found at the bottom 1328

Molecular Biology and Taxonomy

of different sampling stations (e.g. RT7 and Berrocal). For complementary information concerning physico-chemical parameters measured in the Tinto ecosystem see references (2,17). 3.2 DGGE analyses Comparative DGGE analysis was performed to evaluate the level of microbial diversity along the river (Fig. 1) in different seasons. Using specific primers for Bacteria and Archaea partial 16S rRNAs and 16S rRNA genes were amplified. The amplifications resulted in reproducible DGGE fingerprints with a small number of bands (Fig. 2). Out of fifty-seven sequenced bands, thirty-eight showed over 96% similarity with members of six genera, four from the bacterial domain: Acidithiobacillus, Leptospirillum, Acidiphilium and Ferrimicrobium, and two from the archaeal domain: Ferroplasma and Thermoplasma, all of them involved in the iron cycle. All the identified prokaryotes have been detected in previous studies in the Tinto River (2) and in different AMD systems (27). The comparative analysis of all the Acidithiobacillus sequences retrieved from the Tinto ecosystem showed that they cluster with each of the four phylogenetic groups obtained for At. ferrooxidans (17). Accordingly, specific probes for each of the four groups were designed to follow their distribution along the river. All sequenced Leptospirillum clustered into three groups, a, b and c (17), which agrees with previously reported results (28, 29), although the retrieved Tinto’s leptospirilli sequences belong to a different group than the ones reported for Iron Mountain (28) and those found in industrial bioleaching processes (29). Specific probes have been designed to distinguish between these three groups (17). Most of the Acidiphilium sequences showed high homology with the group represented by Acidophilium organovorum, Acidiphilium cryptum and Acidiphilium multivorum (17). The number of sequences corresponding to the gram-positive Ferrimicrobium/ Acidimicrobium group was rather low and they appeared mainly at the origin of the river. A similar situation was observed with the sequences homologous to the archaeal Ferroplasma/Thermoplasma group.

Figure 2. DGGE fingerprint of 16S rRNA gene and reverse transcribed 16S rRNA (*) using universal primers for members of domain Bacteria in different samples from October 1999. Numbers correspond to sampling sites of Fig. 1. Arrows label sequenced bands 1329

Molecular Biology and Taxonomy

3.3 FISH analysis using group- and species-specific probes Total cell counts ranging between 105 and 107 cells/ml were found, regardless of the sampling station and season. These values were similar to those reported previously by López-Archilla using the Most Probable Number (2). The majority of the DAPI-stained cells hybridized with the bacterial probe EUB338 (mean value of 78%). Very few Archaea were found throughout the river, which agrees with the DGGE results. Around 68% of the cells detected with the universal bacterial probe EUB338 could be assigned using groupspecific probes (Fig. 3). Variable percentages of hybridization, up to 69%, were obtained with the specific probe for γ-Proteobacteria (GAM42a), depending on the sampling station (Fig. 3). A specific probe for Acidithiobacillus (THIO1) was also used to detect members of this genus in the Tinto ecosystem. The yields for THIO1 and GAM42a were very similar in most of the samples (17), meaning that most of the γ-Proteobacteria found in the Tinto belong to the genus Acidithiobacillus, which agrees with the DGGE analysis. The use of specific probes designed for each of the four At. ferrooxidans groups gave positive hybridization results with different samples. Interestingly enough, variable populations of At. ferrooxidans, were found along the river (Fig. 4), suggesting that diverse groups of At. ferrooxidans have adapted to the different existing conditions (iron concentration, toxic heavy metals, oxygen concentration, etc.). Values of hybridization up to 65% were obtained with probe NTR712, targeted to members of the Nitrospira phylum (Fig. 3). Former studies suggested that Leptospirillum spp. were the most likely genus of this phylum to be present in the Tinto River (2, 28, 29, 30, 31, 32). Using the specific probes designed for Leptospirillum spp. (17) it was found that the most representative group of leptospirilli along the river was group c (17), which agrees with the DGGE analysis and underlines the difference of the Tinto system from other reported bioleaching systems (28, 29). According to our clustering results, the leptospirilli identified and isolated from the Tinto River correspond to strains of L. ferrooxidans because they cluster with the type strain defined for this species (17). The total cell count detected with L. ferrooxidans specific probes was lower than that detected with NTR712 probe, meaning that some members of this group escaped detection. ALF968 probe was used to detect α-Proteobacteria in the Tinto ecosystem. The yields relative to DAPI-stain reached values of 51%, corresponding to the third most representative group of bacteria in the Tinto River (Fig. 3). Previous studies have described Acidiphilium as the most probable α-Proteobacteria for this type of environment (27). This member of the α-Proteobacteria seems to be associated to At. ferrooxidans and L. ferrooxidans (33). Considering the results obtained with DGGE, a specific probe (ACD638) designed to identify members of the phylogenetically related species A. organovorum, A.multivorum and A. crytum was used to quantify this group of Acidiphilium spp. The percentage of positive hybridizations obtained with this probe was similar in most samples to the values obtained using the general ALF968 probe (17). β- Proteobacteria were detected with probe BET42a. This group of bacteria was a minority in all samples from all stations with the exception of RT3, where a mean value of 73% of total cell was obtained (Fig. 3). Due to the high pH of this sampling site no further characterization of this group of bacteria has been pursued. To quantify Ferrimicrobium and the related genus Acidimicrobium, both characteristic bacteria in AMD systems, probe ACM1160 was designed. These Actinobacteria were found at rather low percentages in the different Tinto River samples (17). 1330

Molecular Biology and Taxonomy

Probes FER656 and TMP654, specific for Ferroplasma and Thermoplasma genera respectively, were used for the detection of these iron-oxidizing Archaea associated to AMD systems (12, 13). In our case, both Ferroplasma and Thermoplasma cells were detected although in a rather small percentage, less than 3%.

Figure 3. Fraction of total cells detected with FISH using bacteria and group specific probes in different sampling sites

Figure 4. Variation of At. ferrooxidans populations along the Tinto River 3.4 Microbial ecology model of the Tinto ecosystem The identification and quantification of the main prokaryotic microorganisms thriving in the Tinto River complements the physiological characterization of the respective isolates (1, 2, 34). This allows a geomicrobiological model for the Tinto River based on the physico-chemical characteristics of the system as well as on the physiological properties of the major species present in the river: At. ferrooxidans, L. ferrooxidans and 1331

Molecular Biology and Taxonomy

Acidiphilium spp (Fig. 5) to be proposed. The Tinto ecosystem is under the control of an active iron cycle. Reduced iron, from mineral ores and solution, is the energy source for both, L. ferrooxidans and At. ferrooxidans, resulting in the production of ferric iron. Oxidized iron is responsible for maintaining a constant acidic pH along the river. On the other hand, At. ferrooxidans can grow under anaerobic conditions using reduced sulfur compounds as electron donors, such as those generated by the polysulfide oxidation mechanism of metal sulfides (11, 19), and ferric ion as an electron acceptor. Acidiphilium spp. have the capacity to respire reduced carbon compounds anaerobically using ferric iron as an electron acceptor, even at dissolved oxygen concentrations of 60% (35). Therefore, this member of the α-Proteobacteria may be, together with At. ferrooxidans, an important element for the reduction of ferric iron under anaerobic or microaerobic conditions, just the conditions found at several locations along the river (RT7, Berrocal). The iron cycle would be completed with these three species and the constant acidic pH would be also explained. Interestingly enough, preliminary results showed that some L. ferrooxidans isolates from the Tinto River are able to anaerobically oxidize (respire) iron using reduced metals as electron acceptors, which suggests that a complete anaerobic iron cycle is also operative in the Tinto ecosystem, with obvious biohydrometallurgical implications (Garcia-Moyano et al, personal communication). Concerning the sulfur cycle, only At. ferrooxidans, able to oxidize sulfur aerobically and anaerobically, has been detected in important numbers using both conventional and molecular ecology techniques. Although isolation of At. thiooxidans from the Tinto River has been reported previously (1, 2), none of the molecular techniques used in this work (DGGE, FISH using specific probes) have been able to detect them. Given the close phylogenetic relationship between At. thiooxidans and At. ferrooxidans (17) further investigation is needed before a final conclusion concerning the status of this species in the river can be reached. As to sulfur reduction, this type of activity has been detected by in situ hybridization at several sampling sites of the Tinto ecosystem, although isolation and physiological characterization will be required to confirm the existence of this important activity in the acidic waters of the Tinto River (17). Different reports have described sulfate reducing activity in other acidic environments, thus it is reasonable to conceive that this activity is also occurring in the Tinto ecosystem, although at a rather low level, probably as a consequence of the high concentration of ferric iron present in the system (36). The microbial ecology of the Tinto River corresponds to what can be expected from the coincidence of two important aspects: the role that iron seems to play in the oxidation of metal sulfides (11) and the mineral composition of the Iberian Pyritic Belt, in which pyrite is the dominant mineral. One noticeable difference between the Tinto ecosystem and other reported acidic systems is the relatively high number of active At. ferroxidans found in the sampling stations along the river, in contrast to the low numbers reported for this microorganism in other bioleaching systems. One possible explanation could be that in industrial bioleaching, aerobic oxidation of iron is favored, so the complete microbial iron cycle is not operative, while in the Tinto ecosystem the microbial oxidation and reduction of iron is performed along the length of the entire river. Probably the main role of At. ferrooxidans is to reduce iron rather than to oxidize it. In this context an interesting correlation has been found between the high concentration of reduced iron in the anoxic parts of the river and the relatively high number of active At. ferrooxidans, suggesting that the role of this microorganism is more closely related to iron reduction rather than to its oxidation (Malki et al, personal communication). Also, the population change of this bacteria along the river might be linked to its adaptation to different metabolic conditions. 1332

Molecular Biology and Taxonomy

Further research has to be done in this line to verify these observations and to evaluate its possible implications in industrial bioleaching operations.

Figure 5. Suggested geomicrobiological model of the Tinto ecosystem. Only representative microorganisms are shown associated to their physiological role in the correspondent iron and sulfur cycles Obviously many questions still remain unanswered (e.g., possible operation of an anaerobic iron cycle, the physiological and metabolic properties of the Leptospirillum spp. isolated in different acidic ecosystems, the low acidity of iron-oxidizing archaea in the Tinto River, etc.), but the tools designed for this study can be used to further explore the microbial ecology of the Tinto River and to compare it with other acidic environments. This information could then be used to generate DNA arrays to monitor the microbial population of industrial bioleaching processes, facilitating their control and detecting anomalous populations that could need correction. Currently, this methodology is being used to control microbial populations involved in coal biodesulfurization processes (see this volume). DNA micro-arrays using genomic libraries of key microorganisms could generate snapshots of gene expression during the process, helping to monitor its progress or to alert about possible malfunctions, facilitating their optimal performance. All these applications are in progress in our lab in collaboration with the Centro de Astrobiologia (Parro et al., personal communication). It is obvious that the application of molecular ecology techniques to biohydrometallurgy has been a watershed. We hope that these tools will help to expand the use of this environmentally friendly biotechnology in a near future. 4.

CONCLUSIONS The use of molecular and conventional microbiological techniques allowed the prokaryotic diversity and the relative concentration of the key microorganisms of an acidic ecosystem, the Tinto River to be studied. In spite of the high level of eukaryotic diversity found in this peculiar habitat, only three bacterial species seem to play an important role in the generation and maintenance of the extreme conditions of the habitat: At. ferrooxidans, L. ferrooxidans and Acidiphilium spp., all of them conspicuous members of the iron cycle. 1333

Molecular Biology and Taxonomy

Other detected microorganisms may also be involved in this system, although they are present in lower numbers. This would be the case of Ferrimicrobium acidophilum and Ferroplasma acidiphilum, whose metabolism is very similar to L. ferrooxidans. Due to the characteristics of the habitat and given what is known about iron’s geomicrobiology, we postulate that the Tinto ecosystem is under the control of iron, a model with important biohydrometallurgical implications. ACKNOWLEDGMENTS This work was supported by grants BIO99-0184 and BX2000-1385 from the Ministerio de Educación y Cultura and 07M/0023/199 from the Comunidad Autónoma de Madrid, an institutional grant from the Fundación Areces to the CBM, and by the Max Planck Society. REFERENCES 1. López-Archilla, A. I., I. Marín, and R. Amils. 1993. Geomicrobiology Journal, 11:223233. 2. López-Archilla, A. I., I. Marín, and R. Amils. 2001. Microbial Ecology, 41:20-35. 3. Davis Jr. R.A., A.T. Welty, J. Borrego, J. A. Morales, J.G. Pendon and J.G. Ryan. 2000. Environmen. Geol., 39:1107-1116. 4. Elbaz-Poulichet, F., C. Braungrardt, E. Achterberg, N. Morley, D. Cossa, J.M. Beckers, P. Nomérange, A. Cruzado and M. Leblanc. 2001. Continental Shelf Research, 21:1961-1973. 5. Geen A. Van, J.F. Adkins, E.A. Boyle, C.H. Nelson and A. Palanques. 1997. Geology, 25:291-294. 6. Leblanc, M, J.A. Morales, J. Borrego and F. Elbaz-Poulichet. 2000. Econom. Geology, 95:655-662. 7. Fernández-Remolar, D.C., N. Rodríguez, F.Gómez and R. Amils. 2003. Journal of Geophysical Research, in press. 8. Amaral-Zettler, L.A., Gómez, F., Zettler, E., Keenan, B.G., Amils, R., Sogin, M.L. 2002. Nature, 417: 137 9. Colmer A.R., K. L. Temple, H.E. Hinkle. 1950. Journal of Bacteriology, 59:317-328. 10. Ehrlich H.L. 2001. Geomicrobiology, fourth edition, Marcel Dekker, Imc., New York. 11. Sand, W., T. Gehrke, P.G. Jozsa, and A. Schippers. 2001. Appl. Hydrometallurgy, 59:159-175. 12. Edwards, K. J., P. l. Bond, T. M. Gihrin, and J. F. Banfield. 2000. Science, 287:17961798. 13. Golyshina, O. V., T. A. Pivovarova, G. I. Karavaiko, T. F. Kondrateva, E. R. Moore, W. R. Abraham, H. Lunsdorf, K. N. Timmis, M. M. Yakimov, and P. N. Golyshin. 2000. International Journal of Systematic Bacteriology, 50:997-1006. 14. Rawlings, D. E., H. Tributsch, and G. S. Hansford. 1999. Microbiology, 145:5-13. 15. Amann, R. I., B.J. Binder, R.J. Olson, S.W. Chisholm, R. Devereux and D.A. Stahl.. 1990.Applied and Environmental Microbiology, 56(6):1919-1925. 16. Amann, R. I., W. Ludwig, and K. H. Schleifer. 1995. Microbiological Reviews, 59(1):143-169. 17. González-Toril, E., Llobet-Brossa, Casamayor E.O., Amann R., Amils R. 2003. Applied and Environmental Microbiology, in press. 18. Amils R., E. González-Toril, D. Fernández-Remolar, F. Gómez, N. Rodríguez and C. Durán. 2003. Reviews in Environmental Science and Technology, in Press. 1334

Molecular Biology and Taxonomy

19. González-Toril, E., F. Gómez, N. Rodríguez, D. Fernández, J. Zuluaga, I. Marín, and R. Amils. 2002. Hydrometallurgy, in press. 20. Raskin, L., J. M. Stromley, B. E. Rittmann, and D. A. Stahl. 1994. Applied and Environmental Microbiology, 60:1232-1240. 21. Stahl, D. A., and R. Amann. 1991, pp. 205-248. In E. Stackebrandt and M. Goodfellow (ed.), Nucleic Acid Techniques in Bacterial Systematics. John Wiley & Sons Ltd., Chichester, UK. 22. Muyzer, G., S. Hottenträger, A. Teske, and C. Wawer. 1996, pp. 3.4.4.1-3.4.4.22. In A. D. L. Akkermans, J. D. van Elsas, and F. J. de Bruijn (ed.), Molecular Microbial Ecology Manual, 2nd ed. Kluwer Academic Publishers, Dordrecht, The Netherlands. 23. Maidak, B. L., J. R. Cole, T. G. Lilburn, C. T. Parker, P. R. Saxman, J. M. Stredwick, G. M. Garrity, B. Li, G. J. Olsen, S. Pramanik, T. M. Schmidt, and J. M. Tiedje. 2000. Nucleic Acids Research, 28(1):173-174. 24. Strunk, O., O. Gross, B. Reichel, M. May, S. Hermann, N. Stuckmann, B. Nonhoff, T. Ginhart, A. Vilbig, M. Lenke, T. Ludwig, A. Bode, K.-H. Schleifer, and W. Ludwig. 1998-2000. http://www.mikro.biologie.tu-muenchen.de. 25. Ludwig, W., O. Strunk, S. Klugbauer, N. Klugbauer, M. Weizenegger, J. Neumaier, M. Bachleitner, and K. H. 16. Schleifer. 1998. Electrophoresis, 19(4):554-568. 26. González-Toril, E., F. Gómez, N. Rodríguez, D. Fernández, J. Zuluaga, I. Marín, and R. Amils. 2001. Part B: 639-650. In V.S.T. Cuminielly and O. García (ed.), Biohydrometallurgy: fundamentals, technology and sustainable development. Elsevier, Amsterdam. 27. Hallberg K.B. and D.B. Johnson. 2001. Advances in applied microbiology, 49:37-84. 28. Bond, P. L., S. P. Smriga, and J. F. Banfield. 2000. Applied and Environmental Microbiology, 66:3842-3849. 29. Coram N.J. and Rawling D.E. 2002. Applied and Environmental Microbiology, 68:838-845. 30. Ehrich, S., D. Behrens, E. Lebedeva, W. Ludwig, and E. Bock. 1995. Archives of Microbiology, 164:16-23. 31. Hippe, H. 2000. International Journal of Systematic and Evolutionary Microbiology, 50:501-503. 32. Edwards, K. J., T. M. Gihring, and Banfield. J.F. 1999. Applied and Enviromental Microbiology, 65:3627-3632. 33. Harrison, J. A. P. 1985. Annual Review of Microbiology, 38:265-292. 34. Irazabal N., I. Marín and R. Amils. 1999. Journal of Bacteriology, 179:1946-50. 35. Johnson D.B. and Brige T.A.M. 2002. Journal of Applied Microbiology, 92:315-321. 36. Lovely D.R. 2000. Edited by Lovely. AMS Press. Washington D.C. Chapter 1, pp:330.

1335

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Phenotypic characterization and copper induced stress resistance in the extremely acididophilic Archaeon Ferroplasma acidarmanus Craig Baker-Austin1*, Mark Dopson1, Andrew Bowen1 and Philip Bond1,2 School of Biological Sciences1 and School of Environmental Sciences2, University of East Anglia, Norwich, England Abstract The Ferroplasmales are a group of extremely acidophilic archaea believed to play a significant role in the oxidization of sulfide minerals associated with the production of acid mine drainage. Here we present a characterization of the isolate Ferroplasma acidarmanus Fer1, and using a range of culture-based and proteomic techniques examine copper resistance and biofilm induction mechanisms adopted by this archaeon. Fer1 grows chemoorganotrophically utilizing yeast extract or sugars as a carbon and energy source, but grows optimally chemomixotrophically utilizing ferrous iron and yeast extract or sugars. Fer1 has temperature and pH optimums of 42°C and 1.2 respectively, but is capable of growing at near pH 0, which represents one of the most extreme examples of acidophily reported to date. Fer1 exhibits remarkably high tolerance to copper ions when adapted to growth at higher concentrations by multiple-step culturing in the presence of 1 g/l (0.0157 mol/l) copper. Exposure to sub-toxic concentrations of copper results in production of exopolysaccharides and the over and under-expression of cellular proteins as detected by 2-dimensional polyacrylamide gel electrophoresis. These results suggest that Fer1 possess a highly efficient Cu2+ homeostasis mechanism to deal with chronic metal stresses. Keywords: Archaea, acidophile, proteomic, exopolysaccharide, metal resistance 1.

INTRODUCTION Ferroplasma acidarmanus is a mesophilic, iron-oxidizing extreme acidophile of the archaeal family Ferroplasmaceae. All of the characterized species within this family are cell-wall lacking obligate acidophiles that grow optimally around 40°C, in highly acidic conditions (~ pH 1-2) utilizing ferrous iron and yeast extract via a chemomixotrophic mode of growth [1, 2]. Ferroplasma acidarmanus strain Fer1 was isolated from the Iron Mountain superfund site in Northern California, where it was shown to constitute 85 ± 7% of a biofilm community by fluorescent in situ hybridization (FISH) [1]. The relative numerical dominance of this archaeon suggests that these organisms play a significant role in the production of acid mine drainage, a process previously thought to be dominated by bacterial iron-oxidizing species such as Acidithiobacillus ferrooxidans and Leptospirillum *

[email protected]

1337

Molecular Biology and Taxonomy

ferrooxidans. Geochemical data indicates pH values as low as -3.6, total dissolved metal concentrations as high as 200 g/l and sulfate values as high as 760 g/l having been recorded in the Iron Mountain superfund site [3]. The presence of biofilm-bound Fer1 communities in this environment implies it is extremely pH and metal tolerant as it is capable of thriving in one of the most acidic and metal-rich natural ecosystems reported to date [3, 4]. Cu2+ is an essential trace element involved in a number of fundamental cellular roles, acting as a cofactor for enzymes as diverse as cytochrome c oxidases, lysyl oxidases or tyrosinases, but can cause serious cell damage through radical formation [5]. Coping with this duality requires regulated pathways to control intracellular copper availability [6]. The levels of copper in the Iron Mountain site vary significantly depending on micro-scale seasonal and hydrological events, but ambient concentrations have been measured varying from 290-9800 mg/l [3]. Strain Fer1 must be tolerant to concentrations of copper within this range; however the underlying genetic and biochemical mechanisms of this resistance are unknown. The partially annotated Fer1 genome sequence (http://www.jgi.doe.gov/ index.html) contains a number of putative open reading frames involved in metal resistance [7], including a copper-transporting ATPase. Here we report initial investigations of copper-induced stress responses of Fer1 by examination of protein expression and exopolysaccharide (EPS) production. The findings have a direct relevance to understanding the characteristics of an extreme acidophilic archaeon involved in geochemical iron and sulfur cycling in bioleaching environments. 2.

MATERIALS AND METHODS

2.1 Archaeal strains and growth conditions F. acidarmanus strain Fer1 was used for all experiments. This strain was isolated from the Iron Mountain superfund site via a series of enrichment cultures from a water and sediment sample obtained in July 1997 [1]. The inoculum for all batch culture experiments (unless otherwise stated) were steady state cells obtained from a chemostat while growing mixotrophically on ferrous iron and yeast extract. Details of chemostat growth conditions are described elsewhere [8]. Fluorescent in situ hybridization (FISH) using the Ferroplasma-specific genus probe Fer-656 (5’-CGTTTAACCTCACCCGATC-3’) was applied to samples originating from the chemostat to ensure the inoculum was a pure and uncontaminated Fer1 source [9]. Unless otherwise stated, batch cultures were carried out in 100 ml mineral salts medium (MSM) and trace elements. The MSM consisted of the basal salts (g l-1) (NH4)2SO4 (3.0), Na2SO4.10H2O (3.2), KCl (0.1), K2HPO4 (0.05), MgSO4.7H2O (0.5), Ca(NO3)2 (0.01) and filter sterilized trace elements (mg l-1): FeCl3.6H2O (11.0), CuSO4.5H2O (0.5), HBO3 (2.0) and MnSO4.H2O (2.0), Na2MoO4.2H2O (0.8), CoCl2.6H2O (0.6), ZnSO4.7H2O (0.9) and Na2SeO4 (0.1) and 0.02% (w/v) yeast extract. The medium was altered to pH 1.2 and supplemented with 70 g/l ferrous iron. All batch cultures were incubated on a rotary shaker at 150 r.p.m at 37°C for 63 hours unless otherwise stated. Growth was monitored by either cell counts with a hemocytometer (Hawksley) on an Olympus BX50 phase-contrast microscope or via protein concentration (Bio-Rad protein assay kit) [10]. All batch experiments were conducted in triplicate unless stated otherwise, with mean and ± standard deviation (SD) presented. 2.2 Batch toxicity and resistance induction experiments Cu2+ toxicity was tested by growth of Fer1 in MSM with the indicated Cu2+ concentrations. Inoculated batch cultures (10 µg of protein) were incubated for 72 hours at 1338

Molecular Biology and Taxonomy

37°C and growth was measured as protein concentration (as above). To test for induction of Cu2+ resistance/adaptation 100 ml shake flasks were inoculated with 10 µg of Fer1 protein previously grown at a sub-toxic concentration of Cu2+ (1 g/l) into medium containing higher Cu2+ concentrations. These cultures were then incubated for a further 72 hours and growth was measured (as above). 2.3 Congo-Red quantitative exopolysaccharide determination assay Production of exopolysaccharide (EPS) during growth and biofilm formation was monitored with Congo red [11]. Fer1 was grown mixotrophically in batch 1 l cultures (as described above) for 72 hours, then exposed to various levels of Cu2+ and incubated for a further 24 or 72 hours. A protein assay was performed and cell samples were centrifuged at 12 000 g for 30 min at 4°C and the supernatants decanted. Cell slurries containing 1 mg of protein were resuspended in 100 µl of 1 M Tris HCl (pH 6.9) and added to 50 µl of a saturated solution of Congo red in 75% ethanol. Mixtures were vortexed for 10 sec and incubated for 3.5 h at 4°C. The volume of each sample was adjusted to 1 ml with ultrapure H2O and centrifuged at 13 000 g for 15 min at 25°C to remove cells, biofilm, and bound Congo red. Supernatants were diluted 1:1 with distilled ultrapure H2O, and transmittance at 500 nm was determined on a Phillips PU8730 Spectrophotometer. All experiments were conducted in triplicate unless stated otherwise, with mean and ± SD presented. 2.4 2-Dimensional polyacrylamide gel electrophoresis Fer1 batch cultures (1 l) inoculated with 100 µg of protein were grown in the absence and presence of 2 g/l Cu2+ for 72 hours. Following growth, cells were centrifuged (as above) and total cellular protein extracted using an urea/thiourea cell lysis buffer [12]. Cell extracts were sonicated and 200 µg protein loaded onto pH 4-7 immobilized pH gradient (IPG) strips (Amersham Pharmacia). The strips were focused for the first dimension separation using a Investigator 5000 focusing unit (Genomic Solutions) with a 24 hour programme that consisted of a maximum voltage of 5000 V, and a volt-hours setting of 80 µA/gel. Standard second-dimensional gel electrophoresis was performed as previously described [13]. 2D gels were either stained with EZ brilliant blue colloidal coomasie G250 (Sigma) or silver stained as previously described [14]. Gels were scanned using the Proteomic Imaging System (Perkin Elmer Life Sciences) and analyzed using Proteomweaver version 1.3 (Definiens). 3.

RESULTS

3.1 Phenotypic characterization experiments Fer1 was capable of growing chemomixotrophically on ferrous iron in the presence of yeast extract, and chemoorganotrophically on yeast extract alone (Table 1). Fer1 grew to a higher cell density chemomixotrophically utilizing ferrous iron and yeast extract (2.74 ± 0.18 mg/l protein), as opposed to chemoorganotrophically on yeast extract alone (1.34 ± 0.38 mg/l protein). The difference in final protein concentration strongly suggests that Fer1 cells actively utilize ferrous iron as a primary energy source, and use yeast extract as a carbon source, but are capable of sub-optimally utilizing yeast for a chemoorganotrophic mode of growth. Fer1 was also capable of growth within the pH ranges 0.2-2.5, with an optimum at around 1.2, and grows between 32 to 51°C, with an optimum growth rate observed at 42°C (Table 1).

1339

Molecular Biology and Taxonomy

Table 1. Optimal growth conditions of Fer1 compared to a range of other extremely acidophilic archaeal species Species Fer1 Fer1 F. acidophilum P. oshimae T. acidophilum

µg protein ml-1 1.34 ± 0.38 2.74 ± 0.18 1.18 ± 0.48 -

Growth Ae (Het) Ae (Mixo) Ae/An Ae Ae/An

Temperature 42°C (23-46) 42°C (23-46) 35°C (15-45) 40°C (45-65) 59°C

pH 1.2 (0.2-1.5) 1.2 (0.2-1.5) 1.7 (1.3-2.2) 0.7 (0-3.5) 1.8-2

Values for Fer1 are from this study; those for Ferroplasma acidophilum [2, and unpublished data], Picrophilus oshimae [15], and Thermoplasma acidophilum [16]. Abbreviations: Ae, aerobic, Ae/An, facultatively anaerobic. Brackets denote the range of temperature and pH where growth occurs. 3.2 Copper toxicity experiments In the absence of Cu2+ Fer1 grew to just under 8 µg/ml, but was strongly inhibited at concentrations of Cu2+ of over 500 mg/l, indicating that Cu2+ detrimentally affected cell growth (Fig. 1A). Further increases of Cu2+ only marginally decreased the protein concentration (Fig. 1A) or cell counts (data not shown). At the highest Cu2+ concentration used (2400 mg/l) the protein concentration was approximately 1.8 µg/ml, and cells were still viable by subsequent inoculation into fresh MSM in the absence of Cu2+. Fer1 cells previously exposed to 1 g/l Cu2+ were capable of growth in MSM containing higher concentrations of Cu2+ (Fig. 1B). The results show a clear reduction in Fer1 protein concentration and cell counts (data not shown) when exposed to higher Cu2+ concentrations. The protein concentration of Fer1 cells grown for 72 hours at 12.8 g/l was 0.8 µg/ml, but these cells were unable to propagate in fresh MSM in the absence of Cu2+ (data not shown). 8

8

A

B

7

Protein (µg/ml)

Protein (µg/ml)

7 6 5 4 3 2

6 5 4 3 2 1

1

0

0 0

0.5

1

1.5

2

Copper (g/l)

2.5

0

2

4

6

8 10 12 14

Copper (g/l)

Figure 1. Growth of Fer1 cells in increasing concentrations of copper (A) and after a previous exposure to a sub toxic concentration of Cu2+ (B). Values in (A) are means ± SD (n = 3) and representative values in (B) (n = 1)

1340

Molecular Biology and Taxonomy

Trans (500 nm)

0.100 0.050 0.000

1 Day 3 Days

-0.050 -0.100 -0.150 0.001

0.01

0.1

1

5

Copper (g/l)

Figure 2. EPS production in Fer1 cultures exposed to increasing Cu2+ stress. The data are means ± SD (n = 3) 3.3 Congo-Red quantitative EPS determination assay The Congo red assay detects EPS production. Cultures producing EPS show an increase in transmittance because polysaccharide in biofilm binds Congo red, removing it from solution. Fer1 cells produce detectable amounts of EPS in response to Cu2+ stress, but no EPS was detected in the absence of Cu2+. Cultures inoculated with increasing concentrations of Cu2+, and cultures stressed for longer periods (3 days) showed marked increases in EPS production compared to control cultures grown for the same period of time (Fig. 2). 3.4 Proteomic analysis of Fer1 Cu2+exposed cells Total protein from Fer1 cells from the different growth cultures were separated using 2-dimensional polyacrylamide gel electrophoresis (2D PAGE). Approximately 600 protein spots are resolved from Fer1 cells cultured with 2 g/l Cu2+ (Fig. 3A). A portion of 2D gel electrophoresis separations from control cells (no Cu2+ exposure) and exposure to 2 g/l Cu2+ gel have been magnified to show examples of up and down regulated proteins in the absence (Fig. 4B) and presence of Cu2+ (Fig. 4C). By employing 2D PAGE we identified approximately 12 proteins up-regulated and 4 proteins down regulated compared to controls. 4.

DISCUSSION The isolate Fer1 represents a species of mesophilic, chemohetero- and chemomixotrophic archaea that grows under extremely acidic conditions. A draft annotation of the genome data (97% complete) has identified a range of incomplete amino acid biosynthetic pathways, an array of intra and extracellular proteases and amino acid uptake pumps. Those results are supported by our findings, which show that Fer1 grows optimally through the chemomixotrophic utilization of ferrous iron and yeast extract (Table 1). Fer1 is shown to grow optimally at pH 1.2, and is capable of growth at close to pH 0. Only two other archaea [15], a few eukaryotic fungi and an algal species have been reported to be capable of growth at pH values around 0 (16, 17]. 1341

Molecular Biology and Taxonomy

Figure 3. Representative silver stained 2D PAGE gel of Fer1 in the presence of 2 g/l Cu2+ (A). The inset area shows the approximate section magnified for Fer1 cultured in the absence (B) and presence of 2g/l Cu2+ (C). Relative isoelectric point and molecular masses shown. Arrows indicate putative up-regulated proteins Microorganisms isolated from highly acidic environments such as acid mine drainage sites polluted with high concentrations of metals exhibit considerable tolerance to these elements. This tolerance may be due to abiotic factors (pH, temperature, nutrients in the environment or growth media) or to the physiological and genetic adaptations of these organisms [18]. Based on the results of batch toxicity and proteomic analysis Ferroplasma 1342

Molecular Biology and Taxonomy

acidarmanus strain Fer1 appears to have developed an inducible and highly efficient Cu2+ resistance system. Fer1 cells grow efficiently at relatively low concentrations of dissolved Cu2+ (<500 mg/l) but growth appears to be inhibited at concentrations exceeding 1 g/l Cu2+. The apparent reduction in cell densities at higher Cu2+ concentrations suggests that Cu2+ exerts a significant toxicological effect. However, all of the cells harvested within these experiments, except those at 12.8 g/l, were shown to be viable via reinnoculation of cells into fresh, unspiked media. This suggests that relatively low-level exposure of Cu2+ to Fer1 cells does not exert a permanent toxicological effect, and that Fer1 cells are capable of withstanding acute, and relatively low-level exposure to Cu2+ ions with no long-term effect on viability. This resistance phenotype may have significant advantages for an organism that grows in a habitat characterized by significant Cu2+ fluxes [2]. Cells previously exposed to copper ions were significantly better adapted to growth in higher concentrations of this metal (Fig. 2), and for surviving for longer periods than that used during the toxicity experiments (2 weeks, data not shown). These data suggest that the Fer1 resistance phenotype is up-regulated or an adaptive response to exposure to copper. Results from the initial phase in our proteomic analysis of Cu2+ exposed Fer1 show significant variations in proteome expression compared to control cells cultured without Cu2+ (Fig. 3B and C). The up- and down regulation of numerous proteins in response to Cu2+ is detected. Fer1 produced significant amounts of EPS when stressed with Cu2+ (Figure 3). The production of EPS appears to be increased when toxic concentrations of Cu2+ are applied to the culture media, and when the stress period was increased to 3 days. These results suggest Fer1 may produce EPS as a stress response to acute Cu2+ exposure, although the actual mechanisms and rationale for EPS production in this archaeon are unknown. Biofilm production has been observed in a wide variety of archaea from functionally and phylogenetically distinct taxonomic groups, and these observations point to the possibility that this is a widespread stress response among the archaea [10]. One explanation is that EPS production may be a population/community level mechanism of heavy metal resistance. Archaeal cultures stressed by high concentrations of metals are often able to incorporate metals into an insoluble matrix [10, 19]. Although the mechanism of sequestration is unknown, polypeptides found in biofilm may act like phytochelatins to trap metals [19], presumably reducing their bioavailability. Recently, there has been a broad recognition that microbial biofilms are an important mode of growth in natural environments [19, 21]. This is the preferred mode of growth for bacteria in many ecosystems because attached growth confers distinct advantages to the biofilm bacteria, such as enhanced nutrient gathering capacity and several-fold resistance to antimicrobials and biocides [22, 23]. This phenomenon has been the subject of intense research in medically important and pathogenic bacteria that produce biofilms, particularly Klebsiella pneumoniae and Pseudomonas aeruginosa (reviewed in [20]). Interestingly, little research has focussed upon biofilm formation within the archaea, or within extremophilic microorganisms, and this area requires further investigation. An understanding of acidophilic biofilm formation may provide insights into the structural stability of these organisms within these highly inhospitable and stochastic environments, and may increase our understanding of the population and community-level implications of surface attached growth, particularly with regards to metal resistance. We are currently analyzing protein spots by trypsin digestion coupled with MALDITOF mass spectrometry. By comparison of protein mass-fingerprinting results to genome sequence data we will identify proteins and genes involved in EPS production and Cu2+ resistance. The recent sequencing and partial annotation of the Fer1 genome has facilitated a range of post-genomic techniques to be used to investigate this archaeon in considerable 1343

Molecular Biology and Taxonomy

depth. We have also been using 2D PAGE to select and excise proteins of interest isolated from cells grown under a wide variety of culture conditions, including anaerobisis, chemoheterotrophy and variations in physio-chemical growth conditions such as temperature and pH. The sequences of these differentially expressed proteins are being determined, and a heterologous prokaryote system is currently being prepared for gene cloning experiments based on Fer1 transcripts. This approach should provide fascinating insights into the genetics, biochemistry and physiology of an extremely acidic archaeal lifestyle. 4.

CONCLUSIONS This study demonstrates that the extremely acidophilic archaeal species Ferroplasma acidarmanus isolate Fer1 grows optimally utilizing ferrous iron and yeast extract as growth substrates at 42°C and pH 1.2. Experimental data show Fer1 to be extremely resistant to Cu2+ and to produce significant and detectable amounts of EPS when exposed to Cu2+ ions. EPS production was also increased with longer stress periods in Fer1 cultures. Both biofilm and batch Cu2+ exposed cultures of Fer1 cells analyzed using 2D PAGE showed marked up- and down regulation of proteins compared to control cultures, suggesting significant genetic and biochemical adaptation to changing culture conditions. ACKNOWLEDGEMENTS This work was supported by the BBSRC. We would also like to thank Francis Mulholland for useful discussions, and the Institute of Food Research (Norwich, England) for use of 2D PAGE scanning equipment. REFERENCES 1. K.J. Edwards, P.L. Bond, T.M. Gihring, and J.F. Banfield, Science, 287 (2000) 17961799. 2. O.V. Golyshina, T.A. Pivovarova, G.I. Karavaiko, T.F. Kondrat'eva, E.R.B. Moore, W.R. Abraham, H. Lünsdorf, T.M. Timmis, M.M. Yakimov, P.M. Golyshin, Int. J. Syst. Evol. Microbiol., No. 50 (2000) 997-1006. 3. D.K. Nordstrom, C.N. Alpers, C.J. Ptacek, D. Blowes, Environ. Sci. Technol., No. 34 (1999) 254-258. 4. D.K. Nordstrom., Ph.D. Dissertation. Stanford Univ., Stanford, CA (1977). 5. D. Strausak, and M. Salioz, J. Biol. Chem., No 272 14 (1997) 8932-8936. 6. F.W. Outten, D.L. Huffman, J.A. Hale, and T.V. O'Halloran, J. Biol. Chem., No. 276 (2001) 30670-30677. 7. T.M. Gihring, P.L. Bond, S.C. Peters, and J.F Banfield, Extremophiles, (2003) (in press). 8. Dopson, M., Baker-Austin, C.M., Hind, A. & Bond, P.L., (unpublished data, 2003). 9. P.L. Bond, and J.F. Banfield, Micro. Ecol., No. 41 (2001) 149-161. 10. M. Bradford, Anal. Biochem., 72 (1976) 248. 11. C. LaPaglia, and P. Hartzell, Appl. Environ. Microbiol., No. 63 (1997) 3158-3163. 12. V. Santoni, M. Molloy and T. Rabilloud, Electrophoresis, No 21 (2000) 1054-1070. 13. A. Hesketh, D. Fink, and B Gust, Mol. Micro., No. 46 2 (2002) 319-330. 14. H. Blum, H. Beier, and H.J.Gross, Electrophoresis, 8 (1987) 93-99. 15. C. Schleper, G. Pühler, I. Holz, I. Gambacorta, D. Janekovic, U. Santarius, P.H. Klenk, and W.J. Zillig, Bacteriol., No 177 (1995) 7050-7059.

1344

Molecular Biology and Taxonomy

16. T.D. Brock, Thermophilic microorganisms and life at high temperatures. SpringerVerlag, New York (1978). 17. O. Sletten, and C. E. Skinner, J. Bacteriol., No 56 (1948) 679-681. 18. M.C. Romero, E.M Gatti, and D.E. Bruno, W. J. Micro. & Biotech., No 15 (1999) 179184. 19. S. Karthikeyan, and T.J., Beveridge, Env. Micro., No. 4 (2002) 667-675. 20. P.S. Stewart, and J. Costerton, Lancet, No. 358 (2001) 135-138. 21. J.W. Costerton, Z. Lewandowski, D. de Beer, D. Caldwell, D.R. Korber, and H.L Lappin-Scott, Annu. Rev. Microbiol., No. 49 (1995) 711–745. 22. G.M. Wolfaardt, J.R. Lawrence, R.D. Robarts, and D.E. Caldwell, Can J. Microbiol., No 40 (1994) 331–340. 23. D.R. Korber, A. Choi, G.M. Wolfaardt, and D.E. Caldwell, Appl. Environ. Microbiol., No. 62 (1996) 3939–3947.

1345

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Pyrite oxidation by halotolerant, acidophilic bacteria P.R. Norris and S. Simmons Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, UK Abstract A pyrite enrichment culture of mesophilic Thiobacillus prosperus-like bacteria and salt-tolerant Acidithiobacillus species was maintained through serial cultures over many months with up to 6% w/v NaCl. The dominant bacteria in the culture were indicated by analysis of ribosomal RNA genes that were amplified by PCR with DNA extracted from laboratory bioreactors. The potential of such cultures for mineral processing where freshwater is not available was confirmed by demonstration of pyrite oxidation rates equivalent to those obtainable with Acidithiobacillus ferrooxidans in the absence of salt. Keywords: acidophiles, salt tolerance, Acidihalobacter 1.

INTRODUCTION The inhibition of mineral sulfide-oxidizing bacteria by salt could preclude biohydrometallurgical mineral processing where only saline water is available. Acidophilic bacteria are generally inhibited by anions (except sulfate) because they accumulate intracellularly in response to positive membrane potentials, the consequent reduction of which leads to denaturing acidification of the cytoplasm [1]. Well-studied mineral sulfide-oxidizing, acidophilic bacteria, such as Acidithiobacillus ferrooxidans, do not, therefore, grow in seawater. However, several strains of mineral sulfide-oxidizing bacteria that could grow with up to 3.5% w/v NaCl have been isolated from marine, geothermal sediments of Vulcano, one of the Aeolian Islands [2,3]. One strain was characterized and named Thiobacillus prosperus [3]. The phylogenetic placement of T. prosperus [4] some distance from the genus Acidithiobacillus which was created for acidophilic thiobacilli [5] has indicated the requirement for its transfer to a novel genus. Solubilization of metals from an ore mixture by T. prosperus was demonstrated and described as similar to that by At. ferrooxidans, but little mineral leaching data was presented [3]. Thiobacillus prosperus-like bacteria have been re-isolated from ferrous iron enrichment cultures of acidic, saline sediment samples from Vulcano [6]. These were divided into two types, initially by comparison of whole cell electrophoretic protein profiles and subsequently by sequencing of their 16S rRNA genes (Norris, unpublished work). One type (strain V6) appeared to be very similar to T. prosperus. However, it differed from the described behaviour of the type strain in that its growth rate on ferrous iron was very similar to that of Acidithiobacillus ferrooxidans, whereas growth of the type strain with ferrous sulfate has been described as poor [3]. The second type (strain V8)

1347

Molecular Biology and Taxonomy

represents a closely related but new species of the genus, for which the name Acidihalobacter ferrooxidans will be proposed in a full description to be published elsewhere. Two Acidithiobacillus species were also isolated from tetrathionate enrichment cultures of the sediment and water samples. Strain V1 was a novel, sulfur- (but not iron-) oxidizing Acidithiobacillus species and strain V2 was a strain of Acidithiobacillus thiooxidans [6]. The experiments described in this paper used a 35°C pyrite enrichment culture that was established with the sample that also gave rise to the cultures from which these Acidithiobacillus species and strains V6 and V8 were isolated. 2.

MATERIALS AND METHODS

2.1 Culture conditions A pyrite enrichment culture was established at 35°C with water and sediment from an acidic, geothermal site on Vulcano Island [4]. The medium used for enrichment, for maintenance of the culture through serial culturing, and for the growth experiments contained (g l-1) (NH4)2SO4 (0.4), MgSO4.7H2O (0.4), K2HPO4 (0.2) and NaCl (30), adjusted to pH 2 with H2SO4. The NaCl concentration was varied as indicated in the text. The pyrite used was a finely ground concentrate (minus 75 µm) with approximately 40% w/v iron. Solubilization of iron from pyrite was followed in air-lift reactors containing a central draft tube, 440 ml medium and 2% w/v pyrite. Iron in supernatants of centrifuged samples was determined by atomic absorption spectrophotometry. 2.2 DNA extraction and analysis DNA extraction, polymerase chain reaction (PCR) amplification of 16S rRNA genes, sorting of cloned genes on the basis of restriction fragment length polymorphisms (RFLPs) obtained from EcoRI/RsaI and EcoRI/Sau3AI double digests, and sequencing were carried out as described previously [6]. Clone libraries of rRNA genes were constructed using the TOPO TA Cloning Kit (Version J) vector (pCR 2.1-TOPO) and host Escherichia coli strain (Invitrogen). Clone libraries of rDNA from cultures in bioreactors were also used in dot blot hybridizations (using the digoxigenin labelling system (Boehringer) with oligonucleotide probes that allowed differentiation of bacterial strains V1, V2A, V6 and V8 (see later). Target sequences for the probes (with the Escherichia coli 16S rRNA gene numbering region equivalent in brackets) were 5'-GTGCTTGCACCTGGTGG-3' (78-99), 5'GGGTGCTAATATCGCTGCTG-3' (459-479), 5'-GCGTGCGTAGGTGGTTGGGT-3' (576-595) and 5'-GCGTGTGTAGGCGGTTTAGT-5' (576-595) for strains V1, V2A, V6 and V8 respectively. 3.

RESULTS AND DISCUSSION

3.1 Effect of salt and temperature on pyrite dissolution by halotolerant acidophiles Growth of T. prosperus on mineral sulfides has been described as fastest in a mineral medium without additional salt, with the medium already containing about 0.25% w/v chloride from ammonium, magnesium, calcium, potassium and sodium salts [2]. This contrasts with the observation that growth-associated pyrite oxidation by similar bacteria was enhanced when the medium contained 2 or 3% w/v NaCl compared to growth with 1% w/v salt (Fig. 1a). Although growth was clearly inhibited by 5% w/v NaCl (Fig. 1b), growth and pyrite oxidation was still maintained over many months with weekly serial 1348

Molecular Biology and Taxonomy

culture in medium containing 6% w/v NaCl. Strain V8 showed slightly greater salt tolerance than strain V6 in further work with ferrous iron as substrate (data not shown). These results confirm a tolerance of salt by these strains far greater than that by At. ferrooxidans (< 1% w/v NaCl) and similar to that by some iron-oxidizing Alicyclobacillus-like bacteria [7].

Figure 1. The effect of salt on the solubilization of iron during growth of pyrite enrichment cultures at 35°C in a bioreactor with 2% w/v pyrite

Figure 2. Solubilization of iron during growth of a pyrite enrichment culture at different temperatures with 2.5% w/v NaCl and 2% w/v pyrite The halotolerant enrichment culture was particularly active over a temperature range of 25 to 40°C with maximum iron solubilization from pyrite at 40°C (Fig. 2). This is close to the optimum temperature for growth of strains V6 and V8 on ferrous iron, which has been determined as about 37°C (data not shown). 1349

Molecular Biology and Taxonomy

3.2. Composition of the pyrite-oxidizing population The halotolerant pyrite enrichment culture was maintained in air-lift bioreactors for many months through serial batch culture, with inoculation (10% v/v) into new medium every one to two weeks. The initial pyrite concentration was 2% w/v for each culture. One culture was maintained in this way for 19 months with a NaCl concentration of 3% w/v. After 11 months, an inoculum was taken to establish a second culture that was then maintained with 6% w/v NaCl for a further 8 months. DNA was extracted directly from these cultures, rRNA genes were PCR-amplified and analysed as described in the Materials and Methods. Representative clones with different RFLPs were sequenced to reveal the corresponding source strains (Fig. 3).

Figure 3. Clone types in libraries constructed from PCR-amplified rRNA genes extracted from a pyrite enrichment culture grown with 3% w/v NaCl. The time scale indicates time elapsed since inoculation of the first of the serial batch cultures. The key indicates the strains to which cloned rRNA sequences corresponded. The number of clones analysed by RFLP comparisons was 90 (after 5 months serial culturing), 60 (after 12 months) and 62 (after 15 months). 79 clones were analysed by dot-blotting with specific probes after 19 months Only three strains were represented in the clone libraries, but others could have been present in the culture in relatively low numbers. Strain V8 was indicated as the principal (and possibly the only) ferrous iron-oxidizing bacterium in the culture. Two Acidithiobacillus strains were indicated, strain V1 and a strain (referred to as strain 2A) whose rRNA sequence was almost identical to that of Acidithiobacillus thiooxidans and to that of strain V2, previously isolated from Vulcano samples [6]. The culture grown with 6% w/v NaCl was analysed by dot blotting of plasmid DNA from clone libraries constructed after 5 months (76 clones) and 8 months (64 clones) of serial culture. The clone libraries were dominated by sequences corresponding to strain V8, with no clones corresponding to strain V1 and only 1% (after 5 months) and 8% (after 8 months) of the clones corresponding to strain V2A. Sulfur-oxidizing Acidithiobacillus species, at least one of which is probably a strain of At. thiooxidans, were present in the mixed cultures with 3 or 6% w/v NaCl, although the 1350

Molecular Biology and Taxonomy

type strain of At. thiooxidans does not possess such salt tolerance. The closely related At. ferrooxidans was not observed in any experiments with NaCl-supplemented medium. 4.

CONCLUSIONS At the time of these analyses, rRNA sequences were available for three iron-oxidizing T. prosperus-like bacteria: T. prosperus itself, strain V6 and strain V8. A fourth, related sequence (V3) was cloned from DNA extracted from a Vulcano sample but no source strain has yet been isolated [6]. This extra sequence, and a strain (VC15) that showed little DNA:DNA sequence hybridization with a T. prosperus isolate in the original description of such bacteria [3], indicates several species within this group of halotolerant acidophiles. Therefore, the apparently specific RFLP patterns and oligonucleotide probes used to assess the mixed population in bioreactors (Fig. 3) might not have revealed any further diversity if this was present as uncharacterized strains with similar 16S rRNA sequences. Further work is also required to establish whether strain V8 has any competitive advantage over strain V6 during growth on pyrite, as might be indicated from its dominance in the bioreactor population. Strains V6 and V8 appear to grow equally well on pyrite in pure culture. There appears to have been no reports of mineral oxidation by such bacteria since the description of T. prosperus fourteen years ago, but this paper confirms activity worthy of further consideration for mineral processing under particular (i.e. saline) conditions. ACKNOWLEDGMENTS We gratefully acknowledge the support of a Biological and Biotechnology Research Council studentship. REFERENCES 1. B. Alexander, S. Leach and W.J. Ingledew, J. Gen. Microbiol., 133 (1987) 1171. 2. G. Huber, H. Huber and K.O. Stetter, pp. 239-251 in H.L. Ehrlich and D.S. Holmes (eds.), Workshop on Biotechnology for the Mining, Metal-Refining and Fossil Fuel Processing Industries, Wiley, NewYork, 1985. 3. H. Huber and K.O. Stetter, Arch. Microbiol., 151 (1989) 479. 4. B.M. Goebel, P.R. Norris and N.P. Burton, pp. 293-314 in F.G. Priest and M. Goodfellow (eds.), Applied Microbial Systematics, Kluwer, Dordrecht, 2000. 5. D.P. Kelly and A.P.Wood, Int. J. Syst. Evol. Microbiol., 50 (2000) 511. 6. S. Simmons and P.R. Norris, Extremophiles, 6 (2002) 201. 7. P.J. Holden, L.J. Foster, B.A. Neilan, G Berra, and Q.M. Vu, pp. 283-290 in V.S.T. Ciminelli and O. Garcia Jr. (eds.), Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, Part A, Elsevier, Amsterdam, 2001.

1351

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Reversible loss of arsenopyrite oxidizing capabilities by Acidiothiobacillus ferrooxidans is associated with swarming phenotype and presence of ISAfe1 Jasmin E. Hurtado Laboratorio Biotecnología Ambiental, Departamento de Microbiología, Universidad Peruana Cayetano Heredia, A.P. 4314, Lima-Peru E-mail: [email protected] Abstract There is a direct relationship between swarming phenotype and loss of arsenopyrite oxidizing capability in isolates of Acidiothiobacillus ferrooxidans. After several years of culturing A. ferrooxidans on arsenopyrite concentrates it was possible to isolate strains that show periodical decay on the level of arsenopyrite oxidation. This low activity is correlated with phenotypic changes in growth in solid media of the swarming phenotype and with the loss of the ability to use oxidation of ferrous to ferric iron as an energy source. Previous work has shown that oxidation of ferrous iron and the swarming phenotype are associated with the presence of the insertion sequence ISAfe1. We hypothesize that the presence of this insertion sequence in our cultures may account for the phenotypic changes as well as the loss of ability to oxidize arsenopyrite. Keywords: Acidiothiobacillus, arsenopyrite, swarming, ISAfe1 1.

INTRODUCTION Acidiothiobacillus ferrooxidans is an acidophilic, obligately chemolithoautothrophic Gram negative rod that oxidizes either ferrous to ferric iron or reduced sulfur compounds to sulfuric acid for energy generation. Initially isolated by Temple and Colmer in 1951 from water mine, its presence has been demonstrated in almost all acidic mining environments. Also, A. ferrooxidans possesses high resistance to toxic metal ions. Acidiothiobacillus ferrooxidans exhibits considerable genetic variation [1, 2] and can rapidly adapt to a wide range to growth conditions [3, 4]. In 1987, the presence of two families (family 1 and family 2) of repetitive DNA sequences in the genome of A. ferrooxidans [5] was described. More recently, these were identified as insertion sequences IST1 and IST2 [6]. A phenotypic switching that includes loss of ability to oxidize ferrous iron and a swarming phenotype is associated with the transposition of IST1 [7]. In 1999, it was proposed that loss of Fe(II) oxidation is due to reversible transposition of the insertion sequence IST1(now named ISAfe1 [8]) into resB gene [9]. In 1995, the structure, sequence and biological activity related to a glutaredoxin of Tn5467, a Tn21-like Transposon located on the A. ferrooxidans plasmid pTF-FC2 were 1353

Molecular Biology and Taxonomy

described [10]. In 1997 an insertion sequence, IST3091 in a T. ferrooxidans plasmid PTF91 was demonstrated [11]. The presence of a Tn7-like Transposon present in the glmUS region of A. ferrooxidans [11] has been reported. Finally, location of the pst operon, which encodes a phosphate-specific transport system, was reported to be downstream of the glmUS genes as in E. coli [12]. In recent years the most rapidly developing area of the byohidrometallurgy has been associated with the extraction of gold from refractory ores and concentrates. A number of commercial and semicommercial operations using arsenopyrite biooxidation tanks and piles are cited in literature [13,14]. During arsenopyrite leaching, high arsenic levels are present in the pregnant solution. In E. coli and other bacteria, arsenate is taken up by phosphate transport systems such as the ATP-coupled Pst pump, a member of the ArsABC superfamily of transport ATPases [15]. Recently, homologous genes of four Ars (arsenic resistance genes) have been identified in A. ferrooxidans [16]. In a previous work [17], we found that during two years of transfers in arsenopyrite, most strains of A. ferrooxidans showed changes in the levels of arsenopyrite leaching, and it was confirmed that degree of arsenopyrite leaching activity was directly related to iron oxidation. The aim of the present work was to examine the relationship between phenotypes associated with IST1 presence and the reversible loss of arsenopyrite oxidizing capabilities of A. ferrooxidans. 2.

MATERIAL AND METHODS

2.1 Bacterial strains Six different strains of A. ferrooxidans were included in this study, namely (1) JH6-L, (2)JH1-DP and (3) JH1-DA isolated from Coricancha gold mine, and (4) JH8A-CE, (5) JH8B-CE and (6) JH9-CE isolated from Centraminas Gold mine. Strains 1-3 and 6 have been previously reported [17]. Strains 4 and 5, variants isolated from JH8-CE, also have been reported [17]. 2.2 Media used Liquid culture media for growth and leaching experiments were prepared according Tuovinen and Kelly [18]. When arsenopyrite was used, ferrous iron was omitted in the preparation. Solid iron media were prepared as previously described [17]. Solid 10:10 medium and 100:10 medium (Medium 100:10 contains 100% of the thiosulfate concentration found in standard solid medium and 10% of the ferrous iron concentration found in standard 9K medium) were made as described by Schrader and Holmes [7] with the following modification: It was used 0.8% Purified Agar (L28, Oxoid Limited, Basingstoke, Hampshire, England.) as a solidifying agent. This purified agar was previously washed twice with distilled water, three times with alcohol and finally twice with distilled water, it was then autoclaved in order to eliminate additional impurities. Liquid 10:10 medium and 100:10 medium were prepared as described previously by Schrader and Holmes [7]. A modified 10:10 soft agar was prepared by using 0.5% Purified Agar washed as described before and plated onto the surface of a modified 10:10 agar. 2.3 Oxidation of iron and thiosulfate Oxidation of iron and thiosulphate was assayed as previously described [17]. 1354

Molecular Biology and Taxonomy

2.4 Arsenopyrite oxidation The procedure used for the arsenopyrite growth measurements has been previously described [17]. In these experiments JH6-L, JH1-DA and JH1-DP were grown in arsenopyrite concentrates containing 21% arsenic and 0.5 Oz/ST gold. Strains JH8A-CE, JH8B-CE and JH9-CE were grown in arsenopyrites concentrates containing 30.1% arsenic, 0.6 Oz/ST gold and 0.7 0z/ST silver. During 7 years, these strains were periodically transferred to arsenopyrite containing medium. 2.5 DNA manipulation DNA isolation were carried out using standard procedures [19] from strains JH8-CE and JH9-CE, which have been grown in liquid media containing iron and thiosulfate as energy sources. PCR amplification of ISAfe1 was done using the following primers derived from ISAfe1(NCBI U66246): A(5’TGCCCCGTCTGTGGTGAGGATG3’) and B(5’TGCCGGGCGTAGCGAACAAGAGT 3’). PCR amplification was carried out as follows: 2 min and 30s at 94° C followed by 30 cycles at 94°C for 30 s, 66°C for 30s, 72°C for 30s, and then 2 min and 30s at 72°C. 3.

RESULTS

3.1 Arsenopyrite oxidation Two levels of arsenopyrite oxidation were exhibited over the years:(1) a low level of arsenopyrite oxidation which had very low concentration of arsenic in solution (0.2-1 g/L), and (2) a high level of arsenopyrite oxidation with more than 12 g/L of arsenic in solution. In a previous work [17] we showed that JH8-CE maintained a high level of ability to oxidize arsenopyrite over a period of two years. To maintain the strain, we streaked JH8CE in a solid iron media. After a week, twenty typical A. ferrooxidans colonies were transferred to equal number of Erlemeyer flasks with arsenopyrite liquid media. After five years of transferring these 20 cultures, only five remain viable. Two of these, JH8A-CE and JH8B-CE were used in this study. While JH8A-CE during these years always has shown a high level of oxidation, JH8B-CE has had three periods of low-level oxidation. One of these low-oxidation periods was eliminated by subculturing in thiosulfate medium. JH1-DA, JH1-DP, JH6-L and JH9-CE have demonstrated continuous cycling between high and low levels of oxidation. We found that change from low to high level of oxidation which might occur spontaneously (sometimes after two to six months) could be induced by growing in thiosulfate. 3.2 Development on solid media and oxidation of iron and thiosulfate Fig. 1 shows the different bacterial colonies observed in this work. The solid media revealed for types of colonies in the media: Wild-type, white, large spreading and iron-oxidizing slimy colonies. The wild type colonies are generally circular, 1 mm diameter, with entire or lobated margin, with a pailla-like protusion in the center dark brown with a yellow-orange band just outside the brown colony. The white colonies were circular, transparent, 1 mm colonies.

1355

Molecular Biology and Taxonomy

The LSC colonies were flat and 1-3 cm in diameter. Growth seemed to be limited for the other colonies. They had a reddish- brown center, 1-2 diameter, surrounded by a 2-5 mm white yellow halo. The rest of the colony was white, with lobated borders (Fig. 1A and 1B). The IOS colonies were convex, reddish brown to yellow, significantly larger than the wild type (JH6-L showed a 3 cm colony of this type), with a mucoid consistency (Fig. 1A and 1C). In modified solid medium 100:10 JH1-DP formed white colonies. Two isolated colonies were picked and transferred to tubes containing 5 ml iron liquid media. In one of them, was obtained 20% iron oxidation whereas in the other the iron oxidation was total. Triplicate samples of one microliter of these cultures were poured on two plates of soft modified 10:10 agar. After 4 days, three of the plates from tubes that obtained 20% iron oxidation showed LSC (Large Spreading Colony) phenotype. Plates from 100% iron oxidation showed IOS (Iron- Oxidizing Slimy) variants. JH1-DA and JH6-L in modified 10:10 soft agar showed LSC colonies and IOS colonies. JH8B-CE showed wild type colonies, iron oxidation slimy variants, and, in some cases circular clearings visible against a lawn of wild type colonies (Fig. 2). 3.3 Presence of ISAfe1 The presence of ISAfe1 in JH8-CE and JH9-CE was evaluated by PCR amplification using ISAfe1 internal primers. The expected PCR product for ISAfe1 is 407 pb, which was confirmed by agarose gel electrophoresis.

Figure 1. Colonies of A. ferroxidans. Upper left: LSC limited by IOS variants, upper right: LSC and wild type colonies (WT), lower left: IOS and lower right: wild type colonies. JH8A-CE always showed typically A. ferrooxidans colonies: flat, 1-2 mm, reddish brown

1356

Molecular Biology and Taxonomy

Figure 2. Colonies of A. ferroxidans in modified 10:10 iron media. Arrow points to circular clearings observed in this media

Figure 3. Characterization of PCR amplified products. Lane 1 shows molecular weight marker DNA (400 pb), lane 2 and 3 show JH9-CE and lanes 4 and 5 show JH8-CE 4.

DISCUSSION A. ferrooxidans is a microorganism which, under toxic conditions, grows resistant forms and is able to survive. Nature has helped it. To date three insertion sequences and two transposes in different strains of A. ferrooxidans, have been reported [6, 10, 11, 12]. However, only few phenotypes may be related to the presence of these elements. One of them ISAfe1 is related to a swarming phenotype. ISAfe1 also is associated with loss of Fe(II) oxidation due to the reversible transposition of the insertion into resB, encoding a putative cytochrome c-type biogenesis protein [9]. As JH1-DP, JH1-DA, JH6-L and JH9 have showed the LSC phenotype directly related to loss of iron oxidation and loss of arsenopyrite oxidation ability, it is very probable that these strains also have the insertion sequence ISAfe1in the resB gene. Phenotypes related to movement of ISAfe1 have been shown by these strains with very high frequency Additional studies are needed in order to find factors inducing these insertion movements Several species are able to form expanding colonies on the surface of semisolid growth media by means of swarming motility. Development of a swarming phenotype requires a series of environmental and intracellular signals [20], including polysaccharide formation [21]. Large Spreading Colonies (LSC) are directly associated to ISAfe1. However, the mucoid consistency of the iron-oxidation slimy (IOS) variants might be related to special insertion zones ISAfe1 or other IS. Our results have shown that IOS variants are related to the formation of LSC. 1357

Molecular Biology and Taxonomy

A Tn5468 transposon which has been reported in A. ferrooxidans [12], can be located in the glmU gene. This gene lies upstream to the pst operon, which encodes a phosphatespecific transport system. Phosphate is very important in the arsenopyrite leaching process, because is one of the few nutrients added in industrial operations [21]. In our strains, transfer of one strain with low activity to a fresh medium where more phosphate is available does not always result in a change of activity; this will suggest that activity might not be related to the Tn5468. All the strains reported here showed very high level of arsenic resistance. The oxyanions of arsenic are quite toxic. In E. coli and other bacteria, bacterial operons encoding resistance to arsenicals and antimonials can be found on transmissible plasmids and in chromosomes. These operons usually have either three (arsRBC) or five (arsRDABC) genes [15]. In A. ferrooxidans chromosomal arsenic resistance genes arsBCH as well as a putative ars R gene has been demonstrated [16]. The present work also reports the presence of ISAfe1 in JH8-CE, who shows a very stable arsenopyrite oxidation during these years. Insert of ISAfe1 does not ocurr into genes related to iron oxidation. One of our strains presented clearing zones in A. ferrooxidans solid cultures. This strain showed wild type phenotype clearing zones resembling plaques from bacteriophages or bdellovibrios. Schrader and Holmes [7] described plaques resembling bacteriophages in LSC colonies. Five years ago, we inoculated 20 flasks of arsenopyrite cultures from JH8-CE and one flasks of each JH1-DA, JH1-DP, JH9-CE and JH6-L. During these years, all 24 cultures were transferred periodically and were provided with identical care and environmental conditions. However, 25% of JH8-CE cultures, whereas 100% survival was observed for all other cultures. JH1-DA, JH1-DP and JH6-L have phenotypes related to movement of ISAfe1. ACKNOWLEDGMENTS This work represents part of the research for a PhD in the Universidad Peruana Cayetano Heredia (J. Hurtado). I thank BH Consultores for chemical assays. REFERENCES 1. A.P. Harrison, Arch. Microbiol.,131 (1982), 3451. 2. D.P. Kelly and A.P. Wood, Int. J. Syst. Evol. Microbiol. 50 (2000) 511. 3. D.S. Holmes and R.U. Haq. In: J. Salley, R.G.L. McCready and P.L. Wichlacz (eds.) Biohydrometallurgy, Otawa: Canadian Centre for Mineral and Energy Technology, 1989. 4. T. F. Kondratyeva, L.N.Muntyan and G.I. Karavaiko, Microbiol., 141 (1995) 1157 5. Yates, J.R. and D.S. Holmes, J. Bacteriol. 169 (1987) 1861 6. D. S. Holmes, J.R. Yates and J. Schrader. In: P. R. Norris and D.P. Kelly (eds.) Biohydrometallurgy. Science and Technology Letters. 1988 7. J. A. Schrader and D.S. Holmes, J. Bacteriol., 170 (1988) 3915. 8. D.S. Holmes, H. Zhao, G. Levican, J, Ratouchnick, J. Bonnefoy, P. Varela and E. Jedlicki. J. bacteriol., 183 (2001) 4323. 9. M.E. Cabrejos, H. Zhao, M. Guacucano, S. Bueno, G. Levican, E. García, E. Jedlicki and D. S. Holmes, FEMS Microbiol. Let. 175 (1999) 223 10. Α. Clennel, B. Johnston and D. E. Rawlings, App. Environ. Microbiol., 61(1995) 4223. 1358

Molecular Biology and Taxonomy

11. L. Chakravarty, J.D. Kittle and O.H. Tuovinen, Can. J. Microbiol., 43 (1997) 503 12. J. C. Oppon, R.J. Sarnovsky, Nancy L. Craig and D. E. Rawlings, J. Bacteriol., 180 (1998) 3007. 13. S. A. Shuey, Eng. & Min. J., pp 16. May 1999. 14. J.E. Hurtado, J.L. Bauer, H. Maldonado, E. Cruz and E. Lazcano, In: G.I.Karavaiko, G. Rossi and Z.A. Avakyan (eds.) International Seminar on Dump and underground bacterial leaching metals from ores. UNEP/USSR Centre for International Projects. 1990. 15. C. Rensing, M. Ghosh and B.P. Rosen, J. Bacteriol., 181 (1999) 5891. 16. B. G. Butcher, S.M. Deane and D.E. Rawlings, App. Environ. Microbiol. 66 (2000) 1826. 17. J.E. Hurtado and A. Berastain, In: Biohidrometallurgical Processing, T. Vargas, C.A. Jerez and H. Toledo (eds.) University of Chile, 1995. 18. 0.H. Tuovinen and D.P. Kelly, Arch. Microbiol., 88 (1973) 285. 19. Ausubel, F.M., R. Brent, R.E. Kingston, D.D. Moore, J.G. Seidman, J.A. Smith and K. Strhul.1992, Current protocols in Molecular biology, Greene Publishing, N.Y.. 20. P.W. Lindum, U. Anthoni, C: Christophersen, l. Eberl, S. Molin and M. Givskov, J. Bacteriol., 180 (1998) 6384. 21. A. Toguchi, M. Siano, M. Burkhart and R.M. Harshey, J. Bacteriol., 182 (2000) 6308. 22. D.E. Rawlings and S. Silver, Bio/Technology, 13 (1995) 773.

1359

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Searching for physiological functions regulated by the quorum sensing autoinducer AI-1 promoted by afeI/afeR genes in Acidithiobacillus ferrooxidans C. Farah, A. Banderas, C.A. Jerez and N. Guiliani* Millenium Institute for Advanced Studies in Cell Biology and Biotechnology, Faculty of Sciences, University of Chile, Santiago, Chile Abstract Acidithiobacillus ferrooxidans (former Thiobacillus ferrooxidans) is one of the main acidophilic chemolithotrophic bacteria involved in the bioleaching of metal sulfide ores. Different reports revealed the capacity of A. ferrooxidans to develop biofilms and the use of immobilized A. ferrooxidans cells seems to be a promising method especially to improve the rate of ferrous iron oxidation. Gene regulation mediated by acyl homoserine lactone (acyl-HSL) (also known as AI1) has been shown to influence many physiological functions in Gram-negative and Grampositive bacteria. Biofilm formation is one of them, and interest in acyl-HSL-mediated quorum sensing and bacterial biofilm formation has increased in recent years. Biofilm maturation requires the synthesis of AI-1 by a member of the LuxI protein family and a member of the transcriptional regulator LuxR protein family which senses AI-1 in a celldensity-dependent manner. The purpose of our study was to identify the physiological functions of A. ferrooxidans regulated by AI-1 quorum sensing with a specifical interest in biofilm formation. The bioinformatic analysis of the incomplete genomic sequence of A. ferrooxidans ATCC 23270 allowed us to identify two open reading frames encoding for proteins with high similarities to LuxI and LuxR family members. We called these ORFs afeR and afeI. They are both located in the same genetic locus but have opposite orientations, as it has been described for Aeromonas hydrophila. Looking for a sequence similar to a lux box, we identified an 18 bp inverse repeat sequence upstream of the afeR gene which could correspond to the putative DNA binding site for a transcriptional regulator. We have successfully isolated the afeR and afeI genes by PCR procedures. Both genes and subsequences encoding different parts of the AfeR carboxyl domain have been cloned independently in overexpression vectors. We are currently studying the regulation of the expression of these genes and the functionality of AfeR in A. ferrooxidans. Keywords: quorum sensing, Acidithiobacillus ferrooxidans, biofilm

* Corresponding author: [email protected]

1361

Molecular Biology and Taxonomy

1.

INTRODUCTION Acidithiobacillus ferrooxidans is a chemolithoautotrophic acidophilic Gram-negative bacterium that belongs to the γ subdivision of the proteobacteria group (1) and obtains its energy from the oxidation of ferrous iron or reduced sulfur compounds. From the different studies of A. ferrooxidans, it is possible to underline two important points: the role of the interaction bacteria/ore in bioleaching process (2) and the second one, is the lack of an efficient and readily available technique to transfer genetic material with a high frequency (3). A. ferrooxidans develops biofilm structures and presents morphological modifications during the cellular adhesion process (4-6). Moreover, A. ferrooxidans active biofilms have been described during pyrite oxidation (7, 8) and it has been demonstrated that A. ferrooxidans biofilms are resistant to various acid washes (9). An analysis of the unfinished bacterial genomes database has revealed the presence of different open reading frames encoding for proteins involved in type IV secretion system in A. ferrooxidans (10). Independently, it is accepted that A. ferrooxidans has a flagellum apparatus (also known as type III secretion system) which seems to be involved in adhesion mechanisms (11, 12). In many Gram-negative bacteria, development of biofilm structure and the secretion system are controlled by a quorum sensing response (13, 14). Initially described in the marine bacterium Vibrio fischeri (15), it is now accepted that quorum sensing is a widespread phenomenon in bacteria. Quorum sensing is a cell density-dependent process that enables bacterial cells to establish a cell-cell communication and to regulate the expression of specific genes in response to local changes in cell-population density. This provides bacteria the means to coordinate their activities in order to function as a multicellular unit (16, 17). Among Gram-negative bacteria and depending of the autoinductor (AI) molecule, two quorum sensing processes have been described: AI-1 which is involved in intraspecies communication and AI-2, related to interspecies communications (17, 18). In V. fischeri, the quorum sensing AI-1 regulatory system is composed of the activator LuxR, a cis-acting DNA inverted repeat called the lux box, and the AHL signal molecule which is synthetized by the bacterial LuxI protein. LuxR and LuxI protein families have been initially defined based on the different studies on the lux operon from Vibrio sp. (19, 20) and confirmed by a phylogenetic study of 76 individual LuxI and LuxR homologues which have demonstrated an early origin during the evolution of Proteobacteria and assumed coevolution for the LuxI/LuxR regulatory cassettes (21). The signal diffuses freely between the cellular and external environments and complexes with LuxR protein only at high population density (22). The AHL-LuxR complex then binds to the lux box: 20-nucleotide inverted repeat centered 44 nucleotides upstream of the start site of the luminescence operon. In Agrobacterium tumefaciens similar 18-bp boxes, tra boxes, are found upstream the TraR-regulated promoters (23). Our working hypothesis is that A. ferrooxidans, like many Gram-negative bacteria, possesses a quorum sensing mechanism that could regulate some important physiological functions such as biofilm formation and secretion apparatus synthesis which are important in bioleaching process. We have just started research to determine if A. ferrooxidans possesses a quorum sensing system and the kind of physiological functions it would control. 1362

Molecular Biology and Taxonomy

2.

MATERIALS AND METHODS

2.1 Bacterial strains, growth conditions and cloning vectors The A. ferrooxidans ATCC 19859 and ATCC 23270 strain were used in this work. Growth in ferrous iron was done at 30°C in modified 9K medium as described before (24, 25). The E. coli strains TOP10F’ and BL21(DE3) were grown in Luria-Bertani (LB) medium (26) at 37°C. We used the pCRT7/NT (Invitrogen®) expression vector to clone the different amplified DNA fragments. 2.2 Genomic analysis Genomic analysis was done by using ORF finders, Blast and Structural Domain Database from NCBI site (http://www.ncbi.nlm.nih.gov/) and the GCG software package (27) to walk on the genome up- and downstream to the afe locus and to analyse the intergenic sequence. 2.3 DNA manipulations The different recombinant clones were selected by PCR in colony. The plasmids from positive recombinant clones were purified by Wizard Miniprep Kit (Promega®) and by QUIAGEN® plasmid Kit according to the manufacturer’s recommendations. The different recombinant plasmids used in expression experiments were sequenced using the Applied Biosystems Hitachi technologies in the laboratory of Genetics, Faculty of Sciences, University of Chile and transformed in E. coli BL21(DE3) strain. 2.4 Primers and PCR conditions The oligonucleotide primers were designed on the basis of the genome sequence and purchased from PROLIGO Corporation®. Taq (Promega®) and Pwo (Roche®) polymerases were used according to the manufacturer’s recommendations. The general PCR conditions were as follows: 3 min at 95°C, followed by 20 cycles at 95°C for 30s, X°C for 30s, 72°C for Ys, and then 3 min at 72°C. For each fragment we used a specific couple [X,Y] depending of the primer Tm and fragment size: afeI gene [65ºC, 30 s], afeR gene [72ºC, 40 s] and for the afeR carboxy-subdomain [69ºC, 40 s]. 2.5 Western immunoblotting The total cell proteins separated by SDS-PAGE were electroblotted onto a polyvinylidene difluoride (PVDF) Immobilon P (Millipore®) membrane as described by Towbin (28), by employing the Trans-Blot Cell System (BioRad®). For the antigenantibody reaction, the membrane was treated with mouse antibody against His-Tag (Amersham®) serum, as the primary antibody (1:2,500 dilution), and monoclonal antimouse antibodies conjugated with peroxidase (Amersham®) as the secondary antibody (1:3,000 dilution). 3.

RESULTS AND DISCUSSION

3.1 Genome analysis By analyzing the incomplete genome sequence from A. ferrooxidans ATCC 23270, we found two open reading frames located in the same locus but in opposite orientations 1363

Molecular Biology and Taxonomy

(Fig. 1A). We called them afeI and afeR. Upstream afeR coding region, there were a Shine and Delgarno sequence and a promoter box. More interestingly, an 18-nucleotide inverted repeat centered 52 nucleotides upstream of putative afeR transcription site was seen (Fig. 1B). We called this sequence r-box since it could play a regulatory role similar to that of lux- and tra boxes. In a similar way, AfeR could form a dimeric asymmetric structure to be functional, as it occurs with LuxR and TraR (29). Both afeI and afeR ORFs encode for two putative proteins which present 60% and 64% similarity with proteins BveR and BveI from Burkholderia cepacia respectively, both members of the LuxI and LuxR proteins family (Figs. 2 and 3). The deduced proteins AfeI and AfeR have 183 and 215 amino acids, molecular masses of 19.9 kDa and 23.7 kDa and theorical isoelectric points of 5.77 and 8.96, respectively.

Figure 1. Identification of the A. ferrooxidans putative genes afeR and afeI and a putative afe box. A, The bioinformatic analysis of the A. ferrooxidans ATCC 23270 incomplete genome revealed the presence of two ORFs, afeR and afeI. B, the sequence analysis of the intergenic region revealed the presence of a putative afe box (boxed bold letters) which is typically centered 52 nucleotides upstream of a putative +1 transcription start site (arrow). Underlined bold nucleotides represent a putative –10 box. SD underlined nucleotides represent a putative Shine and Delgarno site located 8 bp upstream the afeR initiation codon (bold ATG)

Figure 2. Multiple protein sequence alignments of N-acyl-L-homoserine lactone synthase (AHLS). The putative AHLS from A. ferrooxidans (AfeI) was aligned using the NCBI conserved domain database with consensus sequence (Cons) which was generated by multiple alignments of 12 protein sequences belonging to the LuxI protein family. Identical and conserved residues are denoted by * and º respectively 1364

Molecular Biology and Taxonomy

3.2 Structural Analysis of The Transcriptional Regulator AfeR Like the LuxR-type proteins, AfeR appeared to be composed of two functional domains (Fig. 3): the AHL binding domain located in the 2/3 amino terminal part and the DNA binding domain in the carboxyl terminal third (30, 31). Based on this result, we assume that the structure of AfeR could be similar to the TraR structure (Fig. 4) and deletion of the N-terminal domain of AfeR could result in a constitutively active transcriptional regulator without autoinducer (29, 32, 33). The crystal structure presents an asymmetric homodimer (Fig. 4A) whereas the Agrobacterium auto-inductor (AAI) is completely embedded in an enclosed cavity of 200 Å between five-stranded antiparallel βSheets and three helices in the amino-terminal domain (Fig. 4B). While the DNA binding domain is conformed by four α-helices where the two central helices conform the HTH motif.

Figure 3. Characterization and location of the two putative AfeR functional subdomains. The two putative AfeR functional subdomains have been aligned with the consensus sequences (Cons) for the Autoinducer Binding Domain (A) and the HTH-LuxR Domain (B) by the NCBI conserved domain database. Both subdomain consensus sequences were generated by multiple alignments of protein sequences belonging to the LuxR protein family. Identical and conserved residues are denoted by * and º, respectively. Boxes denote the putative α-helices involved in the HTH motif. C, protein sequence location of the autoinducer and HTH subdomains

1365

Molecular Biology and Taxonomy

Figure 4. AAI-TraR complex structure. Structural representations of the AAI-TraR asymmetric dimeric complex (A) and AAI binding domain (B). The structures were obtained by using the Swiss-PdbViewer program (33) and the data file 1H0M (28) from the Protein Data Base (PDB). The structure of the autoinducer domain enclosed in the broken line square (A) was amplified in B 3.3 In vivo expression of the afeI and afeR genes in E. coli The PCR-amplified DNA fragments corresponding to the complete afeI and afeR genes and a partial afeR sequence encoding for 141 amino acids of the AfeR carboxyl subdomain were cloned in the pCRT7/NT vector and used to overproduce in E. coli BL21(DE3) the A. ferrooxidans recombinant proteins including His-Tag in their Nterminal ends. All the recombinant plasmids were analysed by sequencing to confirm the constructions (results not shown). As Fig. 5A shows, there was an increased level of synthesis of three protein bands (arrows) with the expected molecular weights for the different recombinant polypeptides AfeI, AfeR and AfeR141. For an unknown reason, these products were not strictly under the control of the lac promoter and they were expressed with or without 1 mM IPTG. To confirm that these different protein bands corresponded to the expected chimeric polypeptides, the total proteins of the corresponding cells were subjected to Western Blotting by using an antibody against HisTag (Fig. 5B). All the protein bands overexpressed clearly gave a positive result with the antibody against His-Tag. 4.

CONCLUSION By genome sequence analysis and PCR techniques, we have isolated and cloned in an overexpression vector the afeI and afeR genes and DNA fragments from A. ferrooxidans encoding for 78 (result not shown) and 141 amino acids of AfeR carboxyl-terminal subdomain. All the polypeptides were overexpressed in E. coli. This is the first report about quorum sensing studies in A. ferrooxidans. Obviously, further studies will be necessary to understand the regulation of the expression of the afeRI genes and the role of this quorum sensing system in the regulation of other important physiological functions in A. ferrooxidans.

1366

Molecular Biology and Taxonomy

Figure 5. Expression of AfeI, AfeR and AfeR141 proteins from A. ferrooxidans in E. coli BL21(DE3). The afeI (1) and afeR (2) genes and an afeR141 DNA subfragment (3) were amplified by PCR using Pwo Polymerase and a low number of cycles. The amplified fragments were cloned in the pCRT7/NT expression vector which were employed to transform E. coli strain BL21(DE3). All the recombinant strains were grown until D.O = 0.4-0.5 and induced (+) or not (-) with 1 mM IPTG for 6 h. The total cell proteins from each recombinant bacterial strains were separated by SDSPAGE and stained with Coomassie Blue (A) or were subjected to Western Blotting employing antibodies against His-Tag (B). The black symbols (X, •, g) indicate the different overproduced recombinant proteins ACKNOWLEDGMENTS This research was supported by grants DID I-02/4-2 from the Universidad de Chile and ICM P99-031-F. REFERENCES 1. D. Kelly and A. Wood Int. J. Syst. Evol. Microbiol., 50 (2000) 511. 2. R. Arredondo, A. Garcia and C.A. Jerez Appl. Environ. Microbiol., 60 (1994) 2846.

1367

Molecular Biology and Taxonomy

3. Z. Liu, N. Guiliani, C. Appia-Ayme, F. Borne, J. Ratouchniak and V. Bonnefoy J. Bacteriol., 182 (2000) 2276. 4. D. Karamanev J. Biotechnol., 20 (1991) 51. 5. J. Blais, R. Tyagi, N. Meunier and J. Auclair Process Biochem., 29 (1994) 475. 6. W. Sand, T. Gehrke, R. Hallmann and A. Schippers Appl. Microbiol. Biotechnol., 43 (1995) 961. 7. J. Rojas, M. Giersig and H. Tributsch Arch. Microbiol. 163 (1995), 352. 8. C. Pogliani and E. Donati, Process. Biochem. 35 (2000) 997. 9. M. Oprime and O. Garcia Jr. In International Biohydrometallurgy Symphosium, Part A. Ouro Preto, Minas Gerais. Brazil (2001), 369. 10. P. Christie and J. Vogel Trends Microbiol., 8 (2000) 354. 11. A. DiSpirito, M. Silver, L. Voss and O. Tuovinen Appl. Environ. Microbiol., 43 (1982) 1196. 12. N. Ohmura, K. Tsugita, J. Koizumi and H. Saiki J. Bacteriol., 178 (1996) 5776. 13. V. Sperandio, M. Jay, W. Nguyen, S. Shin and J. B. Kaper Proc. Natl. Acad. Sci. U.S.A., 96 (1999) 15196. 14. K. Sauer, A. Camper, G. Ehrlich, J. W. Costerton and D. J. Davies Bacteriol., 184 (2002) 1140. 15. K. Nealson and J. Hastings Microbiol. Rev., 43 (1979) 496. 16. N. Whitehead, A. Barnard, H. Slater, N. Simpson and G. Salmond FEMS Microbiol. Rev., 25 (2001) 365. 17. S. Winans and B. Bassler J. Bacteriol., 184 (2002) 873. 18. B. Bassler Cell., 109 (2002) 421. 19. J. Engebrecht and M. Silverman Nucleic Acids Res. 15 (1987), 10455. 20. B. Bassler, M. Wright, R. Showalter and M. Silverman Mol. Microbiol., 9 (1993) 773. 21. K. Gray and J. Garey Microbiology., 147 (2001)2379. 22. C.Fuqua, S. Winans and E. Greenberg J. Bacteriol., 176 (1994) 269. 23. C. Fuqua and S. Winans J. Bacteriol., 178 (1996) 435. 24. P. Ramírez, H. Toledo, N. Guiliani and C. A. Jerez Appl. Environ. Microbiol., 68 (2002) 1837. 25. N. Guiliani and C. A. Jerez, Appl. Environ. Microbiol., 66 (2000) 2318. 26. J. Sambrook, E. F. Fritsch and T. Maniatis (eds.), Molecular cloning; a laboratory manual, Cold Spring Harbor, N. Y. (1989). 27. Wisconsin Package Version 10.0, Genetics Computer Group (GCG), Madison, Wisc. 28. H. Towbin, T. Staehelin and J. Gordon, Proc. Natl. Acad. Sci. U.S.A., 76 (1979) 4350. 29. A.Vannini, C. Volpari, C. Gargioli, E. Muraglia, R. Cortese, R. De Francesco, P. Neddermann and S. Marco EMBO J., 21 (2002) 4393. 30. S. Choi and E. Greenberg J. Bacteriol., 174 (1992) 4064. 31. Z. Luo and S. Farrand Proc. Natl. Acad. Sci., 96 (1999) 9009. 32. R. Zhang, T. Pappas, J. Brace, P. Miller, T. Oulmassov, J. Molyneaux, J. Anderson, J.Bashkin, S. Winans and A. Joachimiak Nature., 417 (2002) 971. 33. S. Choi and E. Greenberg Proc. Natl. Acad. Sci. U.S.A., 88 (1991) 11115. 34. N. Guex and M.C. Peitsch Electrophoresis 18 (1997) 2714.

1368

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Systematic analysis of our culture collection for "genospecies" of Acidithiobacillus ferrooxidans, Acidithiobacillus thiooxidans and Leptospirillum ferrooxidans D. Mitchell1, K. Harneit1, G. Meyer1, W. Sand1*, E. Stackebrandt2 1

Universität Hamburg, Institut für Allgemeine Botanik, Abteilung Mikrobiologie, Ohnhorststr. 18, 22609 Hamburg, Germany 2 Deutsche Sammlung von Mikroorganismen und Zellkulturen, DSMZ, Mascheroder Weg 1b, 38124 Braunschweig, Germany

Abstract A strain collection of leaching bacteria containing 42 strains of Acidithiobacillus (At.) ferrooxidans, 17 strains of At. thiooxidans, and 26 strains of Leptospirillum ferrooxidans was investigated by genomic methods. The strains originate from mining sites all over the world. In previous studies, some of these strains have been characterised physiologically and partly genetically resulting in the notion that these species need to be subdivided on the basis of genomic differences. Following extraction of genomic DNA, a 566 bp fragment of the 16S rRNA gene was amplified by the polymerase chain reaction (PCR) and the resulting fragments analysed by denaturing gel electrophoresis (DGGE) in a 40 to 70% urea-formamide gradient. In case of At. ferrooxidans and L. ferrooxidans, 5 different subgroups emerged, while 3 subgroups were found among strains of At. thiooxidans. For sequence analyses of the 16S rRNA genes, one strain of each ermerging subgroup was chosen. The findings confirmed the results of the DGGE analyses. The subgroups 1 to 4 of At. ferrooxidans showed a similarity between 98.2% and 99% with each other. The representative of subgroup 5 is only distantly related to the type strain of At. ferrooxidans ATCC 23270T, but comparably close to strain DSM 2392, which in turn is related to the genus Nevskia (although lithotrophic growth at pH 2 with ferrous iron occurs). The 3 subgroups of At. thiooxidans were only distantly related among each other. In case of L. ferrooxidans subgroups 1 and 3 had a rDNA identity of 99.5% and 98.8% with the type strain DSM 2705T. The only strain of subgroup 2 had a value of 98.5% to the recently defined type strain L. ferriphilum (ATCC 49881T). The remaining 2 subgroups 4 and 5 were not included in this analysis. Keywords: DGGE, 16S rRNA, At. ferrooxidans, At. thiooxidans, L. ferrooxidans

*

Corresponding author. Mailing address: Universität Hamburg, Institut für Allgemeine Botanik, Abteilung Mikrobiologie, Ohnhorststr. 18, 22609 Hamburg, Germany. Phone: + 49 40 42816 423, Fax: + 49 40 42816 423. Email: [email protected]

1369

Molecular Biology and Taxonomy

1.

INTRODUCTION The interest in biodiversity of prokaryotic microorganisms from extreme habitats has increased constantly during the past decades. One reason is the understanding that, once isolated from these locations, these microorganisms may represent life in the early periods of earth. Another reason is the potential application of these extremophilic bacteria in various industrial and biotechnological processes, e.g. as suppliers for enzymes or as biocatalysts. Among these, acidophilic prokaryotes from habitats with extreme low pH are of great interest [1] for the industrial exploitation of metals or coal. Though more gentle than chemical leaching, bioleaching may result in environmental pollution of the surroundings of piles and mines, termed "acid mine drainage" [2-4]. The bacteria involved in this processes belong to the genus Thiobacillus, Acidithiobacillus and Leptospirillum [5-10]. Since the first description by Beijerinck [11], taxonomists were aware about the heterogeneity of the genus Thiobacillus. However, reclassification into different species or genera was not possible due to insufficient information about genetic and physiological differences [12]. Only when DNA-DNA hybridisation and sequence analysis of 16S rRNA genes were introduced into systematics, it became possible to analyse the degree of interrelationship within the genus Thiobacillus. As a result, it became obvious that the members of the genus Thiobacillus are distributed in the α-, β- and γ-subclasses of the proteobacteria [13-15]. As a consequence, Thiobacillus versutus was transferred into the genus Paracoccus [16,17], and Thiobacillus acidophilus was moved to the genus Acidiphilium [18]. Thiobacillus intermedius, Thiobacillus perometabolis, Thiobacillus thermosulfatus and Thiobacillus cuprinus were reclassified by Moreira and Amils [19] as species of the new genus Thiomonas. Kelly and Wood [20] established three new genera, embracing eight Thiobacillus species: Halothiobacillus was described for Thiobacillus neapolitanus, Thiobacillus halophilus, and Thiobacillus hydrothermalis; Thermithiobacillus contained the former Thiobacillus tepidarius; Acidithiobacillus was established for the phylogenetically very closely related species Thiobacillus ferrooxidans, Thiobacillus thiooxidans, Thiobacillus caldus, and Thiobacillus albertis. The genera Acidithiobacillus, Thermithiobacillus, and Halothiobacillus belong in the γ-subclass of the proteobacteria [20]. Further analysis indicated that even the species At. ferrooxidans and At. thiooxidans were phylogenetically heterogeneous [21-23]. The polyphyletic genus Leptospirillum was suggested by Lane et al. [13], De WulfDurand et al. [24], and Hippe [25] to constitute an independent class of the proteobacteria. Also L. ferrooxidans consists of genetically different strains [26]. Using denaturing gradient gel electrophoresis (DGGE) and sequence analysis of 16S rRNA genes it was the goal of this study to analyse the degree of genomic heterogeneity of At. ferrooxidans, At. thiooxidans and L. ferrooxidans. 2.

MATERIALS AND METHODS

2.1 Bacterial strains, media and growth Cultures of At. ferrooxidans, At. thiooxidans, and L. ferrooxidans are listed in Tables 1, 2 and 3, respectively. Growth of iron-II-ion oxidizing bacteria was done in a medium as described by Mackintosh [27] at pH 2.0 and 28°C. Sulfur-oxidizing bacteria were cultivated according to Hutchinson [28] at pH 4.5 and 28°C. Acidiphilium acidophilum

1370

Molecular Biology and Taxonomy

was grown according to Harrison [29] and Ferromicrobium acidophilus was kept according to Johnson [30]. Table 1. Designation and origin for strains of At. ferrooxidans Strain R1 R2 R4 R5 R6 R7 R8 R9 R10 WVa WR1 WR2 WR3 A1 A2 A-4 A-6 C-52 D-26 AS1 AS2 2Y Yellow3 Yellow4 Yellow7 Yellow8 SPIII/3 SPIII/7 F221 Chile 1 Van IB DECp LP 2B DSM 583 ATCC 19859 ATCC 23270* 13598 p BKM-b-458 F 427 RAM 6F PH

Origin Ilba mine, Romania Ilba mine, Romania Ilba mine, Romania Ilba mine, Romania Ilba mine, Romania Ilba mine, Romania Ilba mine, Romania Ilba mine, Romania Ilba mine, Romania Coal mine, Preston County, West Virginia, USA Uranium waste heap, Ronneburg, Germany Uranium waste heap, Ronneburg, Germany Uranium waste heap, Ronneburg, Germany Ore from Rammelsberg, Goslar, Germany Ore from Rammelsberg, Goslar, Germany Copper mine, Rincón Andacollo, Chile Copper mine, Vista Hermosa, Andacollo, Chile Copper mine, St.Rita Combarbala, Chile Copper mine, Disputa de las Condos, Chile Hamburg, Germany Garbage dump, Hamburg, Germany Yellowstone, USA Yellowstone, USA Yellowstone, USA Yellowstone, USA Yellowstone, USA Cartagena, Spain Cartagena, Spain Uranium mine, Forstau, Austria unknown Ilba, Valea Ardeleana, Romania Sulfur spring, New Mexico, USA Coal mine, West Pennsylvania, USA Copper mine, Rammelsberg, Goslar, Germany Coal mine, Pennsylvania, USA Copper mine drainage, British Columbia, Canada Copper mine drainage, USA Copper mine, Bingham Canyon, Utah, USA Coal pile, Moscow, Russia Ore heap, Düsseldorf, Germany Rammelsberg, Goslar, Germany Coal seam, Randolph county, Missouri, USA

1371

Molecular Biology and Taxonomy

Table 2. Designation and origin for the strains of At. thiooxidans Strain WRA WRB WRC DSM 504 DSM 622 DSM 9463 ATCC 19377* RAM 8T K6 K16 365 R20 B BKM-b-460 Van IB IY-3 AS3

Origin Uranium waste heap, Ronneburg, Germany Uranium waste heap, Ronneburg, Germany Uranium waste heap, Ronneburg, Germany Soil, New Yersey, USA Lake, Göttingen, Germany Bioreactor, Australia Sulfur lake, Libya Copper mine drainage, Goslar, Germany Sewage pipeline, Hamburg, Germany Sewage pipeline, Hamburg, Germany Minzenschwand, Germany Ilba mine, Romania Ore from Rammelsberg, Goslar, Germany Sulfur deposit, Razdol, Russia Yellowstone, USA Coal mine, Pennsylvania, USA Yellowstone, USA

Table 3. Designation and origin for the strains of L. ferrooxidans Strain R3 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 R40 AS1 US1 BKM-b-1339 Van IB Robin DSM 2705* DSM 2391 ATCC 49879 ATCC 49880 ATCC 49881 P3A Lf ag 29 S53

1372

Origin Ilba mine, Romania (ATCC 49879) Ilba mine, Romania (ATCC 49880) Ilba mine, Romania Ilba mine, Romania Ilba mine, Romania Leach heap, Rosia Poieni, Romania Leach heap, Rosia Poieni, Romania Leach heap, Rosia Poieni, Romania Leach heap drainage, Roisa Poieni, Romania Baia Sprie mine, Romania Baia Sprie mine, Romania Baia Sprie mine, Romania Yellowstone, USA unknown Copper deposit, Lagerdy, Armenia Ilba, Valea Ardeleana noua, Romania Stockpile, Romania Copper deposit, Armenia Copper mine drainage, Gramatikovo, Bulgaria Ilba mine, Romania(=R3) Ilba mine, Romania(=R30) Isolated from mixed culture (=P3A), origin Peru Isolated from mixed culture (ATCC 49881), origin Peru unknown Opencast mining, Ronneburg, Germany Rammelsberg, Goslar, Germany

Molecular Biology and Taxonomy

2.2 DNA extraction Bacterial DNA was extracted according to Jayarao et al. [31]. Enzymatic lysis of bacterial cells was followed by phenolic extraction and alcoholic precipitation of DNA. Yield and quantity was checked by agarose gel electrophoresis. The DNA extract was resuspended in 15 µl TAE buffer and stored until processing at -20°C. 2.3 PCR amplification PCR amplifications were carried out to synthesize 566 bp long 16S rRNA gene fragments, using the primer pair 341 F GC (5'-CGCCCGCCGCGCGCGGCGG GCGGGGCGGGGGCACGGGGGGCCTACGGGAGGCAGCAG-3') and 907 R (CCGT CAATTC(AC)TTTGAGTT-3'). 2 µl of extracted chromosomal DNA was mixed in a 0,5 ml PCR-tube with 50 pmol oligonucleotide primers each, 200 µmol of dNTPs each, 1 mmol of MgCl2, 5µl of 10x PCR-Buffer (Peqlab), and 2,5 units of TaqPolymerase (Peqlab). The reaction mixture was filled up to a total volume of 50 µl with sterile PCRwater and coated with two drops of mineral oil as evaporation protection. The PCR reaction was performed in a thermocycler (Minicycler PTC 150, Biozym) as a touchdown program [32]. Amplification steps were as follows: Initial denaturation of genomic DNA at 92°C for 180s followed by primer annealing at 68°C and elongation at 72°C for 60s. Then, denaturation was performed during cycles 2 to16 at 92°C for 45 s each, elongation at 72°C for 60s each, while simultaneously stepwise the annealing temperatur was decreased down to 57°C. The next 13 cycles were run at 92°C for 45s, 57°C for 60s und 72°C for 60s. A terminal elongation step was at 72°C for 300s. 2.4 Denaturing gradient gel electrophoresis DGGE analysis was carried out in a vertical gel chamber (Protean II, Biorad) using 16x18 cm glas sheets and 1.5 mm spacers as well as 25 well loading combs. Further details relating to experimental construction are described in [33-37]. The gradient used for DGGE analysis was in the range between 40% to 70% urea-formamide. Each lane was loaded with approximately 100 ng of a PCR product. Gels were run in 1x TAE buffer (40 mmol Tris, 20 mmol Na-acetate, and 1 mmol Na-EDTA at pH 7.4) at 60°C. Initially, 200 volts for 10 minutes were applied to allow the DNA migrate into the gel, followed by 100 volts for 18 hours. DNA patterns were visualised by silver staining. 2.5 DNA sequencing Sequencing of PCR products was carried out as described by Rainey et al. [38]. Purified PCR products were sequenced directly using the Taq DyeDeoxy Terminator Cycle sequencing kit (Applied Biosystems). An Applied Biosystems 373A DNA genetic analyser was used for the electrophoresis of the sequence reaction products. The ae2 editor [39] was used to align the 16S rDNA sequences of the strains analyzed against the sequences available in public databases. Pairwise evolutionary distances were computed using the correction of Jukes and Cantor [40]. 3.

RESULTS The 42 strains of At. ferrooxidans could be divided into 5 subgroups, due to the different melting points of their 16S rDNA gene fragments. Figure 1 shows a DGGE profile of subgroup 2 plus the type strain ATCC 23270T, which is a member of subgroup 1. Each DGGE gel contained a G+C ladder in lane 1 to facilitate orientation and comparability between gels. This ladder was made from 16S rDNA gene fragments of the 1373

Molecular Biology and Taxonomy

type strains of: 1. At. ferrooxidans ATCC 23270 (G+C content of 56,7 mol%); 2. At. thiooxidans DSM 9463 (57,2 mol%); 3. L. ferrooxidans DSM 2705 (57,4 mol%); 4. Acidiphilium acidophilum DSM 700 (57,6 mol%); and 5. Ferromicrobium acidophilus T2 (60,4 mol%).

Figure 1. DGGE-pattern of 19 strains of At. ferrooxidans. The PCR-amplified 16S rDNA fragment is about 566 bp long and correlates with position 341 to 907 in the 16S rRNA gene of E. coli. Strain designations and origin see Table 1. DGGE-gel with urea-formamide gradient (40 to 70%), 18h at 100V. Lane 1 contains G+C ladder from 1: At. ferrooxidans ATCC 23270, 2: At. thiooxidans DSM 9463, 3: L. ferrooxidans DSM 2705, 4: Acidiphilium acidophilum DSM 700, 5: Ferromicrobium acidophilus T-23 An overview of the distribution of 42 strains of At. ferrooxidans in five subgroups is shown in Table 4. Table 4. Distribution of 42 strains of At. ferrooxidans into 5 subgroups. Strains selected for 16S rRNA gene sequence analyses are in bold Subgroup 1: Subgroup 2: Subgroup 3: Subgroup 4: Subgroup 5:

13598 p, A1, A2, AS2, ATCC 19859, ATCC 23270, BKM-b-458, DSM 583, F 427, R6, RAM 6F, WR1 2B, A-4, C-52, Chile 1, DECp, F221, LP, R1, R2, R4, R5, R7, R8, R9, R10, Van IB, Wva, WR3 2Y, A-6, AS1, D-26, PH, Yellow 3, Yellow 4, Yellow 7, Yellow 8 SPIII/7, WR2 SPIII/3

For subsequent 16S rDNA sequence analysis one strain of each of the 5 subgroups were selected (Table 4). Representatives of subgroups 1 to 4 exhibited between 98.2% to 99.0% sequence similarity among themselves. Strain SPIII/3, however, shows 99.6% 1374

Molecular Biology and Taxonomy

sequence identity to strain DSM 2392, which belongs to a separate line within the γproteobacteria, comprising members of Gallionella and Nevskia. The 17 strains of At. thiooxidans were subdivided into three subgroups (Table 5) (DGGE-gels not shown). Table 5. Distribution of 17 strains of At. thiooxidans into 3 subgroups. Strains selected for 16S rRNA gene sequence analyses are in bold Subgroup 1: Subgroup 2: Subgroup 3:

365, AS3, ATCC 19377, BKM-b-460, DSM 504, DSM 622, DSM 9463, K6, K16, R20, RAM 8T, Van IB, WRA, WRB, WRC B 1Y-3

For further sequence analysis the strains R20, B, and 1Y-3 were selected. The data (not shown) indicate that these strains are only quite distantly related to each other. It may thus be speculated that strain B and strain 1Y-3 are probably not related with the species At. thiooxidans, but represent two as yet undescribed species. DGGE analyses of the 26 strains of L. ferrooxidans yielded 5 subgroups (Table 6) (gels not shown). Table 6. Distribution of 26 strains of L. ferrooxidans into 5 subgroups. Strains selected for 16S rRNA gene sequence analyses are in bold Subgroup 1: Subgroup 2: Subgroup 3: Subgroup 4: Subgroup 5:

AS1, ATCC 49879, ATCC 49880, BKM-b-1339, DSM 2705, R3, R30, R32, R33, R34, R35, R36, R37, Robin, US1, Van IB 29, ATCC 49881, LF ag, P3A, R38, R39, R40 R31 S53 DSM 2391

The subgroups 4 and 5 could not be analysed due to technical problems. R30 and R31 have 99.5% and 98.8% sequence identity, respectively, to the type strain DSM 2705T of L. ferrooxidans. Strain R38 shows a 98.5% sequence similarity to the recently described second species within the genus Leptospirillum, L. ferriphilum [26]. 4.

DISCUSSION Research performed within the past decades aimed at characterizing the physiological and genomical variability of strains of At. ferrooxidans, At. thiooxidans, and L. ferrooxidans. The findings revealed a considerable heterogeneity of strains contained under these species [21-26]. Having this in mind, the strain collection of these three species were characterized by an molecular approach. The results clearly indicate that the 42 strains of the "species" At. ferrooxidans fall into one of 5 different subgroups. While strains of subgroups 1 to 4 seem to be closely related, strain SPIII/3 represents an own line within the γ-proteobacteria, branching adjacent to members of the iron-bacterium Gallionella and to Nevskia. The 17 strains of At. thiooxidans were subdivided into 3 groups. Following 16S rRNA gene sequence analysis data we postulate that the strains B and 1Y-3 belong to two new species. Strains of the species L. ferrooxidans are also not genomically homogeneous as at least five subgroups were determined. Strains of subgroup 2 belong to the species L. ferriphilum, recently described by Coram and Rawlings [26]. In conclusion, the assignment of strains included in this study to DGGE-patterndefined subgroups, is in total agreement with the cluster analysis based on 16S rDNA gene 1375

Molecular Biology and Taxonomy

sequences. Thus the "species" At. ferrooxidans, At. thiooxidans and L. ferrooxidans embrace several "genospecies" that need to be subjected to thorough phenotypic testing in order to decide whether they can be described as individual species or subspecies. REFERENCES 1. Johnson, D. B. 1998. Biodiversity and ecology of acidophilic microorganisms. FEMS Microbiol. Ecol. 27: 307-317. 2. Colmer, A. R., and Hinkle, H. E. 1947. The role of microorganisms in acid mine drainage: a preliminary report. Science 106: 253-256. 3. Lowell, H. L. 1983. Coal mine drainage in the United States - an overview. Water. Sci. Techn. 15: 1-25. 4. Schippers, A., Hallmann, R., Wentzien, S. and Sand, W. 1995. Microbial diversity in uranium mine waste heaps. Appl. Environ. Microbiol. 61: 2930-2935. 5. Touvinen, O. H. and Kelly, D.P. 1972. Biology of Thiobacillus ferrooxidans in relation to the microbiological leaching of sulfide ores. Z. Allg. Microbiol. 12: 311346. 6. Sand, W., Rohde, K., Sobotke, B. and Zenneck, C. 1992. Evaluation of Leptospirillum ferrooxidans for leaching. Appl. Environ. Microbiol. 58: 85-92. 7. Hallmann, R., Friedrich, A., Koops, H. P., Pommering-Röser, A., Rohde, K., Zenneck, C. and Sand, W. 1992. Physiological characteristics of Thiobacillus ferrooxidans and Leptospirillum ferrooxidans and physicochemical factors influence microbial metal leaching. Geomicrobiol. J. 10: 193-206. 8. Rawlings, D. E. and Kusano, T. 1994. Molecular Genetics of Thiobacillus ferrooxidans. Microbiol. Rev. 58: 39-55. 9. Schrenk, M. O., Edwards, K. J., Goodman, R. M., Hamers, R. J. and Banfield, J. F. 1998. Distribution of Thiobacillus ferrooxidans and Leptospirillum ferrooxidans: Implications for Generation of Acid Mine Drainage. Science 279: 1519- 1522. 10. Rawlings, D. E., Tributsch, H. and Hansford, G.S. 1999. Reasons why `Leptospirillum´-like species rather than Thiobacillus ferrooxidans are the dominant iron-oxidizing bacteria in many commercial processes for the biooxidation of pyrite and related ores. Microbiology 145: 5-13. 11. Beijerinck, M. W. 1904. Über die Bakterien, welche sich im Dunkeln mit Kohlensäure als Kohlenstoffquelle ernähren können. Centralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt. II 11: 593-599 12. Kelly, D. P. and Harrison, A. P., JR.. 1989. Genus Thiobacillus. In Bergey’s Manual of Systematic Bacteriology, 1st edn, vol.3, pp 1842-1458. Edited by J. T. Staley, M. P. Bryant, N. Pfennig and J. G. Holt. Baltimore: Williams and Wilkins. 13. Lane, D. J., Harrison, A. P., JR., Stahl, D., Pace, B., Giovannoni, S. J., Olsen, G. J. and Pace, N. R. 1992. Evolutionary Relationships among Sulfur- and Iron-Oxidizing Eubacteria. J. Bacteriol. 174: 269-278. 14. Donald, I. R., Kelly, P. D., Murrell, J. C. and Wood, A. P. 1997. Taxonomic relationship of Thiobacillus halophilus, T. aquaesulis, and other species of Thiobacillus, as determined using 16S rDNA sequencing. Arch. Microbiol. 166:394398. 15. Goebel, B. M., Norris, P. R. and Burton, N.P. 1999. Acidophiles in biomining. Appl. Microbial Sys. Edited by F. G. Priest and M. Goodfellow. Dordrecht: Kluwer. 16. Katayama, Y., Hiraishi, A. and Kuraishi, H. 1995. Paracoccus thiocyanatus sp. nov., a new species of thiocyanate-utilizing facultative chemolitotroph, and transfer of

1376

Molecular Biology and Taxonomy

Thiobacillus versutus to the genus Paracoccus as Paracoccus versutus comb. nov. with emendation of the genus. Microbiology 141: 1469-1477. 17. Rainey, F. A., Kelly, D. P., Stackebrandt, E., Burghardt, J., Hiraishi, A., Katayama, Y. and Wood, A. P. 1999. A re-evaluation of the taxonomy of Paracoccus denitrificans and a proposal for the creation of Paracoccus pantotrophus comb. nov. INT. J. Syst. Bacteriol. 49: 645-651. 18. Hiraishi, A., Nagashima, K. V. and Katayama, Y. 1998. Phylogeny and photosynthetic features of Thiobacillus acidophilus and related acidophilic bacteria: its transfer to the genus Acidiphilium as Acidiphilium acidophilum comb. nov. Int. J. Syst. Bacteriol. 48: 1389-1398. 19. Moreira, D. and Amils, R. 1997. Phylogeny of Thiobacillus cuprinus and other mixotrophic thiobacilli: proposal for Thiomonas gen. nov.. Int. J. Syst. Bacteriol. 47: 522-528. 20. Kelly, D. P. and Wood, A. P. 2000. Reclassification of some species of Thiobacillus to the newly designated genera Acidithiobacillus gen. nov., Halothiobacillus gen. nov. and Thermithiobacillus gen. nov.. Int. J. Sys. Evol. Microbiol. 50: 511-516. 21. Harrison, A.P., JR. 1982. Genomic and Physiological Diversity Amongst Strains of Thiobacillus ferrooxidans, and Genomic Comparison with Thiobacillus thiooxidans. Arch. Microbiol. 131: 68-76. 22. Pizarro, J., Jedlicki, E., Orellana, O., Romero, J and Espejo, R. T. 1996. Bacterial Populations in Samples of Bioleached Copper Ore as Revealed by Analysis of DNA Obtained before and after Cultivation. Appl. Environ. Microbiol. 62: 1323-1328. 23. Paulino, L. C., Bergamo, R. F., De Mello, M. P., Garcia, O., Manfio, G. P. and Ottoboni, L. M. 2001. Molecular characterization of Acidithiobacillus ferrooxidans and A. thiooxidans strains isolated from mine wastes in Brazil. Antonie van Leeuwenhoek 80: 65-75. 24. De Wulf-Durand, P., Bryant, L. J. and Sly, L. I. 1997. PCR-Mediated Detection of Acidophilic, Bioleaching-Associated Bacteria. Appl. Environ. Microbiol. 63: 29442948. 25. Hippe, H. 2000. Leptospirillum gen. nov. (ex Markosyan 1972), nom. rev., including Leptospirillum ferrooxidans sp. nov. (ex. Markosyan 1972), nom. rev. and Leptospirillum thermoferrooxidans sp. nov. (Golovacheva et al. 1992). Int. J. Sys. Evol. Microbiol. 50: 501-503. 26. Coram, N. J. and Rawlings, D. E. 2002. Molecular Relationship between Two Groups of the Genus Leptospirillum and the Finding that Leptospirillum ferriphilum sp. nov. Dominates South African Commercial Biooxidation Tanks That Operate at 40°C. Appl. Environ. Microbiol. 68: 838-845. 27. Mackintosh, M.E. 1978. Nitrogen fixation by Thiobacillus ferrooxidans. J. Microbiol. 105: 215-218. 28. Hutchinson, D. N. and Thompson, K. M. 1965. The taxonomy of certain thiobacilli. J. Gen. Microbiol. 3: 357-366. 29. Harrison, A. P., JR. 1981. Acidiphilium cryptum gen. nov., sp. nov., heterotrophic bacterium from acidic environments. Int. J. Syst. Bacteriol. 31: 327-332. 30. Johnson, D. B. 2002 (pers. comm.) 31. Jayarao, B. M., Dore, J. J., Baumbach, G. A., Matthews, K. R. and Oliver, S. P. 1991. Differentiation of Streptococcus uberis from Streptococcus parauberis by polymerase chain reaction and restriction fragment length polymorphism of 16S ribosomal DNA. J. Clin. Microbiol. 12: 2774-2778.

1377

Molecular Biology and Taxonomy

32. Don, R. H., Cox, P. T., Wainwright, B. J., Baker and K., Mattick, J. S. 1991. Touchdown PCR to circumvent spurious priming during gene amplification. Nucl. Acids Res. 19: 4008. 33. Muyzer, G. 1999. DGGE / TGGE a method for identifying genes from natural ecosystems. Current Opinion in Microbiology 2: 317-322. 34. Muyzer, G., Brinkhoff, T., Nübel, U., Santegoeds, C., Schäfer, H. and Wawer, C. 1998. Denaturing gradient gel electrophoresis (DGGE) in microbial ecology. Mol. Microbial Ecol. Manual 3.4.4: 1-27. 35. Muyzer, G., De Waal, E. C. and Uitterlinden, A. G. 1993. Profiling of Complex Microbial Populations by Denaturing Gradient Gel Electrophoresis Analysis of Polymerase Chain Reaction-Amplified Genes Coding for 16S rRNA. Appl. Environ. Microbiol. 59: 695-700. 36. Muyzer, G., Ramsing, N. B. 1995. Molecular Methods to Study the Organization of Microbial Communities. Wat. Sci. Tech. 32: 1-9. 37. Muyzer, G. and Smalla, K. 1998. Application of denaturing gel electrophoresis (DGGE) and temperature gel electrophoresis (TGGE) in microbial ecology. Antonie van Leeuwenhoek 73: 127-141. 38. Rainey, F.A., Ward-Rainey, N., Kroppenstedt, R.M. and Stackebrandt, E. 1996. The genus Nocardiopsis represents a phylogenetically coherent taxon and a distinct actinomycete lineage: proposal of Nocardiopsaceae fam. nov. Int J Syst Bacteriol 46, 1088-1092. 39. Maidak, B.L., Cole, J.R., Parker, C.T., JR and 11 other authors. 1999. A new version of the RDP (Ribosomal Database Project). Nucleic Acids Res 27, 171-173. 40. Jukes, T.H. and Cantor, C.R. 1969. Evolution of protein molecules. In Mammalian Protein Metabolism, vol. 3, pp21-132. Edited by H.N. Munro. New York: Academic Press.

1378

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

The strain genotypic heterogeneity of chemolithotrophic microorganisms T.F. Kondrat’eva, T.A. Pivovarova, L.N. Muntyan, S.N. Ageeva and G.I. Karavaiko Institute of Microbiology, Russian Academy of Sciences, Pr. 60-letiya Oktyabrya 7, k. 2, Moscow, 117312 Russia* Abstract The study on the role of bacterial strain diversity and variability in biohydrometallurgical processes is one of the key tasks on a way to increase the rate and efficiency of leaching of valuable metals from sulfide ores and concentrates. More than 100 strains of Acidithiobacillus ferrooxidans and other bacteria (A. thiooxidans, Sulfobacillus thermosulfidooxidans, S. sibiricus, Leptospirillum ferrooxidans) and archae (Acidianus brierleyi, Ferroplasma acidiphilum) were isolated from various econiches and dense pulps of the concentrates used in bacterial-chemical technologies under mesophilic conditions. The strain heterogeneity of chemolithotrophic bacteria A. ferrooxidans was evidenced from the genotype characteristics, such as: (i) the G+C content in DNA (5658%); (ii) genome size (2.2⋅109 – 2.8⋅109 Da); (iii) DNA-DNA homology values (4 genomovars and 6 strains with the relatedness level of 13 to 43% were distinguished between); (iv) 16S rDNA sequence (all the studied strains were subdivided into 3 phylogenetic groups lying in just the same cluster); (v) chromosomal DNA structure (each strain is characterized by a unique chromosomal DNA structure analyzed by pulsed-field gel electrophoresis of native DNA digested by the same restriction endonuclease); (vi) plasmid profile (the strain were shown to differ in the number and size of plasmids). The main environmental factors that cause the genotype diversity of chemolithotrophic microorganisms are energy substrate and metal ions. Mutation and insertion processes result in the intraspecies variability and microevolution of bacteria. Keywords: chemolithotrophic microorganisms, strain polymorphism, chromosomal DNA, plasmids, adaptation 1.

INTRODUCTION As is well known, the species diversity of acidophilic chemolithotrophic microorganisms, capable to obtain energy from oxidation of Fe2+, S2-/S° and sulfide minerals, is relatively minor in mesophilic communities. Thus, our studies have demonstrated that the Gram-negative bacteria Acidithiobacillus ferrooxidans hold the *

This work was supported by grants nos. 00-15-97765 and 01-04-48270 from the Russian Foundation for Basic Research and by grant within the scope of the Federal Central Scientific and Technical Program "Technology of the Living Systems".

1379

Molecular Biology and Taxonomy

dominating position among the members of microbial communities isolated from natural habitats or dense pulps of sulfide ores and concentrates used in bacterial-chemical processes of gold recovery and non-ferrous metal leaching. The second place in terms of the number is occupied usually by A. thiooxidans. Also, representatives of the Grampositive bacteria of the genus Sulfobacillus are almost always found in these communities, along with Leptospirillum ferrooxidans and the archae belonging to Acidianus brierleyi and Ferroplasma acidiphilum species. The quantitative proportions of these microorganisms are changed depending upon the surrounding conditions in natural habitats. In general, not species but the strains, possessing the maximum growth rate and the highest activity of oxidizing an energy source in given conditions in comparison with the other strains, participate in biohydrometallurgical processes, as well as in other microbial technologies. This is why selection of highly efficient strains that are resistant to extreme factors under permanently changing surrounding conditions and have great adaptive potentialities is a very important task. Yet, the study on a role of bacterial strain diversity and variability in biohydrometallurgical processes has not attracted a proper attention, albeit, in our view, is one of the key issues in solving the problems related to increasing the efficiency of valuable metals recovery from sulfide ores and concentrates. The aim of this work was to study the genotypic diversity of acidophilic chemolithotrophic bacterial strains with a focus on revealing such characteristics as the G+C content in DNA, the genome size, the DNA-DNA homology level, nucleotide sequences of 16S rRNA genes, the chromosomal DNA structure, and plasmid profiles and to elucidate a role of energy substrates and content of metal ions in microevolution of these bacteria. 2.

MATERIALS AND METHODS More than 100 strains of different species of microorganisms: Acidithiobacillus ferrooxidans, A. thiooxidans, Sulfobacillus thermosulfidooxidans, S. sibiricus, Leptospirillum ferrooxidans, Acidianus brierleyi, and Ferroplasma acidiphilum were isolated from various environments (mine waters, ore deposits and ore concentrates, dens pulps obtained by processing complex sulfide concentrates). Batch cultivation of these microorganisms was performed on a shaker (180 rpm) in 250-ml Erlenmeyer flasks, containing 100 ml medium or in 5 l flasks containing 3 l medium at temperature: 28°C for A. ferrooxidans, A. thiooxidans, L. ferrooxidans; (42-48)°C for S. thermosulfidooxidans; 55°C for S. sibiricus; 65°C for A. brierleyi and (35-42)°C for F. acidiphilum. The composition of media used for cultivation of A. ferrooxidans, A. thiooxidans and L. ferrooxidans was described in [1]; for S. thermosulfidooxidans in [2], for A. brierleyi in [3], for F. acidiphilum in [4]. The initial pH of the medium was from 1.8-2.0 to 4.5 depending on the microorganism. Inoculum was introduced into the growth medium in amounts 10% v/v. Strain polymorphism in chromosomal DNA was analyzed by pulsed-field gel electrophoresis [5]. The G+C content of DNA, genome size, analysis of the total genome by DNA-DNA hybridization, nucleotide sequences of the 16S rDNA were studied as in [6]; interaction of chromosomal and plasmid DNA - as in [7].

1380

Molecular Biology and Taxonomy

3.

RESULTS

3.1 Strain polymorphism in chromosomal DNA from chemolithotrophic microorganisms 3.1.1 Chromosomal DNA structure Most clearly, the genotypic polymorphism of chemolithotrophic microbial strains manifested itself in the structure of their chromosomal DNA analyzed by pulsed-field gel electrophoresis (PFGE) of fragments formed upon cleavage of native DNA by restriction endonucleases. Earlier, the strain polymorphism of chromosomal DNA structure was noted in strains of the Gram-negative bacteria A.ferrooxidans [5], of the Gram-positive bacteria S thermosulfidooxidans [2], and in strains of archae A.brierley [3]. As seen from the restriction patterns (fig. 1), each strain of A. thiooxidans, S. sibiricus, and F. acidiphilum were characterized by the specific profile - the number and size of chromosomal DNA fragments. The PFGE analysis of chromosomal DNA can successfully be used to identify new strains, to detect the known strains in biotechnological processes and experiments, and to gain more insight into factors and mechanisms responsible for the strain variability.

Figure 1. Restriction patterns of chromosomal DNA from different strains of acidophilic chemolithoautotrophic microorganisms: A – A. thiooxidans: (1) TTA, (2) TTMt, (3) TTN-d, (4) size marker. B – S. sibiricus: (1) N1, (2) size marker, (3) B2, (4) OSSO. C – F. acidiphilum: (1) Y-1, (2) size marker, (3) Y-2, (4) Y-3, (5) Y-4. Restriction endonucleases: (A) XbaI, (B) NotI, (C) XhoI. PFGE conditions were 13 V/cm, 10-s pulse, 68 h run at 12-14°C. At the right side is indicated a size of DNA fragments, Kbp 1381

Molecular Biology and Taxonomy

3.1.2 The G+C content of DNA and genome size The studied A. ferrooxidans strains differed in the G+C content in DNA which varied from 56.1 to 58.1 mol % and in the genome size (2.2⋅109 - 2.8⋅109Da). The most frequent genome size was (2.3 – 2.4)⋅ 109Da, and the most frequent G+C content was about 57 mol%. 3.1.3 Similarity analysis of the total genomes by DNA-DNA hybridization The DNA-DNA hybridization analysis of twenty-three A. ferrooxidans strains showed their considerable genetic heterogeneity (Fig. 2). The strains were divided into four groups characterized by a high degree of genome similarity. Such genome groups are presently called as genomovars. The studied A. ferrooxidans strains were divided into four genomovars, and six strains with the relatedness level of 13 to 43% could be referred to none of the genomovars. The data of DNA-DNA hybridization indicated a high genotypic divergence of A. ferrooxidans strains.

Figure 2. The approximate relatedness dendrogram of A. ferrooxidans strains derived from the DNA-DNA hybridization data. The scale bar corresponds to the level of genome similarity 3.1.4 Nucleotide sequences of the 16S rDNA We determined nearly complete nucleotide sequences (about 1450 nucleotides) of the 16S rRNA genes of five A. ferrooxidans strains: the type strain ATCC 23270T (genomovar 1), strain TFN-d (genomovar 2), strain TFY (genomovar 3), and strains TFI and TFD (genomovar 4). According to the results of this analysis, most of A. ferrooxidans strains fell into three phylogenetic groups [6]. All newly studied A. ferrooxidans strains belonged to the major phylogenetic cluster, which included also A. thiooxidans strains. The type 1382

Molecular Biology and Taxonomy

strain A. ferrooxidans fell within phylogenetic group I, strain TFN-d fell within group II, and strains TFI, TFD, and TFY fell within phylogenetic group III. Strains of the species A. ferrooxidans isolated from different habitats exhibited a considerable genotypic diversity, which manifested itself in a noticeable divergence of their 16S rRNA gene sequences, a low similarity level of the total DNA, and significant differences in the structure of chromosomal DNA revealed by PFGE. 3.2 Polymorphism of A. ferrooxidans plasmid profiles We studied plasmid profiles of twenty-seven A. ferrooxidans strains, which were isolated from various geographic sites and substrates with different composition of main sulfide minerals or were selected in experiments and possessed an elevated resistance to heavy metal ions [8]. Of twenty strains isolated from various substrates, cells of sixteen A. ferrooxidans strains harbored one to four plasmids of different size. Plasmids were found in all the strains from gold-containing arsenopyrite ores and concentrates, in nine of eleven strains isolated from non-ferrous metal ores and concentrates, and only in half the strains from mine waters, pyrite-containing coals, and active sludge. A proportion of plasmid-free strains was as greater as the simpler was composition of substrates. It is not excluded that plasmids participate in regulation of energy substrate oxidation. 3.3 Changes in chromosomal DNA structure of chemolithotrophic microorganisms during adaptation to new energy substrates or increased concentrations of metal ions Earlier, the changes in chromosomal DNA structure in the result of an adaptation to new energy substrates were noted in strains of the Gram-negative bacteria A. ferrooxidans [9, 10] and the Gram-positive bacteria S. thermosulfidooxidans [2]. In this work, changes in the structure of chromosomal DNA at a metabolic switch from Fe2+ oxidation for S° oxidation were demonstrated also for representatives of the other kingdom - Archae. Thus, an additional 75⋅103 -bp fragment appeared in a XhoI-restriction pattern of chromosomal DNA from the archae F.acidiphilum strain Y-4 (fig. 3). As to several A.ferrooxidans strains, we studied a correlation of experimentally acquired resistance to copper, arsenic, nickel, mercury, zinc, and ferric oxide ions to the chromosomal DNA structure. Each strain was characterized the specific threshold concentration of metal ions at which it could grow and oxidize ferrous iron. Changes in the chromosomal DNA structure were revealed only for the TFI-Fe strain with a greater experimentally acquired resistance to 50 g Fe3+ /l [11]. These changes did not disappear after multiple passages on a medium with 9 g/l of ferrous iron. In such a way, a new strain with changed chromosomal DNA structure (in comparison to the parent strain) was obtained under experimental conditions using the medium with a great content of Fe3+ ions. Evidently, the formation of new microbial strains under an influence of varying metal ion concentrations may occur in natural conditions.

1383

Molecular Biology and Taxonomy

Figure 3. XhoI-restriction patterns of chromosomal DNA from strains of F. acidiphilum Y-4: (1) size marker, (2) on the medium with S°, (3) on the medium with Fe2+. PFGE conditions are as in legend to fig. 1 3.4 Changes in plasmid profiles of A. ferrooxidans strains adapted to new energy substrates In a next series of experiments, we studied the plasmid profiles of A. ferrooxidans strains that were grown on the medium with ferrous iron and then adapted to various oxidation substrates: S°, FeS2, or sulfide concentrate. Cells of the A. ferrooxidans strain TFL-2 grown on Fe2+-containing medium harbored one plasmid, whereas two plasmids occurred in the strains TFO, TFBk, and TFV-1 and three plasmids, in the strain TFN-d. Several strains, i.e. TFL-2, TFO, and TFBk, displayed the unchanged plasmid copy number depending upon the oxidized substrate. While, adaptation of the strain TFN-d to sulfide concentrate or adaptation of the strain TFV-1 to sulfur, pyrite, or sulfide concentrate was followed by changes in the number of plasmids. Evidently, there are no common patterns in response of different strains to changing oxidation substrates, but a metabolic switch for the oxidation of a new substrate sometimes was accompanied by changes in the plasmid profile. Albeit plasmids of this species are considered as cryptic, our data suggest their possible participation in adaptation to changing environmental factors and in exchange of genetic information between plasmid and chromosomal DNA. This exchange manifested itself, for example, in localization of expressed genes responsible for Hg-resistance on chromosomal DNA and localization of single defect operone genes on a plasmid pTF-FC2 of A. ferrooxidans [12]. 3.5 Interaction of chromosomal and plasmid DNA in cells of A. ferrooxidans strains adapted to new energy substrates Plasmids of the strain TFBk served as probes for Southern hybridization with blots of chromosomal DNA fragments from several A. ferrooxidans strains adapted to various energy substrates. In particular, the pattern of hybridization of chromosomal DNA fragments from the strain TFBk, adapted to S°, FeS2, or sulfide concentrate, to pTFK2 isolated from this strain is presented in Figure 4. Many fragments of chromosomal DNA from all the studied strains showed weak hybridization signals. In some cases, we 1384

Molecular Biology and Taxonomy

observed the changes in localization of weakly signaling hybridization bands (indicated with double arrows) if A.ferrooxidans cells were pre-adapted to new substrates. Chromosomal DNA of all the studied strains TFBk, TFN-d, TFL-2, TFO, TFV-1 contained multiple nucleotide sequences complementary to those in the plasmid of the strain TFBk. Due to the relatively small size as judged by a low intensity of hybridization signals, multiplicity, and localization on chromosomal and plasmid DNA, these sequences can be referred to IS-elements that previously were found in the A. ferrooxidans genome by Holmes et al. [13]. The changes in localization of IS-elements in the result of an adaptation to new energy substrates led us to assume the participation of these elements in such an adaptive mechanism. Earlier, it was assumed by Holmes et al. [13] that these transposable elements of the genome are involved in the mechanism of adaptation to changing environmental conditions, in evolution of gene structure and functions, in regulating the gene expression, and in genetic information exchange between different strains. More intense signals of DNA hybridization to plasmid pTFK2 were noted (fig. 5, indicated by arrows) in comparison to multiple low-intensity signals as in the case of pyrite-adapted strains TFO and TFN-d. Possibly, cells of the above listed strains contained plasmids with long nucleotide sequences, homologous to plasmid pTFK2 of the strain TFBk. Intense hybridization signals may arise due to an insertion of plasmid DNA into the chromosomal DNA at the metabolic switch for oxidation of a new energy substrate. Thus, our studies demonstrated the possibility of chromosomal and plasmid DNA interaction in A. ferrooxidans strains during adaptation to new energy sources. This mechanism may be invoked in intraspecies variability that underlies the microevolution processes, resulting in the appearance of strain genotypic polymorphism.

Figure 4. А. XbaI-restriction patterns of chromosomal DNA from A. ferrooxidans strain TFBk adapted to new energy substrates: (1) size marker, (2) plasmid DNA of the strain TFBk, (3) Fe2+, (4) S°, (5) FeS2, (6) sulfide concentrate. PFGE conditions were 12 V/cm, 25-s pulse, 44 h run at 13°C (here and on Fig. 5). B. Radioautograph of XbaI-fragments of chromosomal DNA from A. ferrooxidans strain TFBk hybridized with pTFK2 DNA of this strain

1385

Molecular Biology and Taxonomy

Figure 5. A. XbaI-restriction patterns of chromosomal DNA from A. ferrooxidans strains TFO (1) and TFN-d (2), adapted to FeS2 В. Radioautograph of XbaIfragments of chromosomal DNA from the strains hybridized with pTFK2 DNA of TFBk strain 4.

CONCLUSION Taken together, the obtained data support that the strains of acidophilic chemolithotrophic microorganisms possess the high adaptive potentiality and widely ranging reaction norms due to their evolution in extreme and frequently changing environmental conditions in nature. Furthermore, microevolution of these microorganisms is accelerated in biohydrometallurgical processes in which highly dense pulps, great concentrations of metal ions, and ultra-low pH values are rather common. For this reason, the genotypic strain polymorphism of acidophilic chemolithotrophic microorganisms should be considered in biotechnology. As demonstrated in our numerous studies, the introduced strains, if they were not isolated from an oxidized substrate, are replaced by the natural aborigine strain, which is better adapted to a given substrate and a complex of conditions created in pulps. Undoubtedly, an intimate "substrate-strain" relation must be considered when trying to achieve the maximum rate and efficiency of biohydrometallurgical processes. ACKNOWLEDGEMENTS Authors thanks T.I. Bogdanova, V.S. Melamud and T.A. Tsaplina for supplying Sulfobacillus sibiricus strains. REFERENCES 1. G.I. Karavaiko, G. Rossi, A.D. Agate, S.N. Groudev, and Z.A. Avakyan (eds.), Methods of isolation, evaluation and studying of microorganisms, Centre for international projects GKNT, Moscow, 1988. 2. T.F. Kondrat'eva, V.S. Melamud, I.A. Tsaplina, T.I. Bogdanova, A.A. Senyushkin. T.A. Pivovarova and G.I. Karavaiko, Microbiology, 67 (1998) 13. 3. T.F. Kondrat'eva, T.A. Pivovarova and G.I. Karavaiko, Microbiology, 68 (1999) 443.

1386

Molecular Biology and Taxonomy

4. O.V. Golyshina, T.A. Pivovarova, G.I. Karavaiko, T.F. Kondrat,eva, E.R.B. Moore, W.-R. Abraham, H. Lunsdorf, K.N. Timmis, M.M. Yakimov and P.N. Golyshin, Int. J. Syst. Evol. Microbiol., 50 (2000) 997. 5. T.F. Kondrat'eva, L.N. Muntyan and G.I. Karavaiko, Microbiology UK, 141 (1995) 1157. 6. G.I. Karavaiko, T.P. Turova, T.F. Kondrat'eva, A.M. Lysenko, T.V. Kolganova, S.N. Ageeva, L.N. Muntyan and G.I. Karavaiko, Int. J. Syst. Evol. Microbiol, 53 (2003) 113. 7. T.F. Kondrat'eva, V.N. Danilevich, S.N. Ageeva and G.I. Karavaiko, Microbiology, 73 (2004) 308. 8. T.F. Kondrat'eva, S.N. Ageeva, L.N. Muntyan, T.A. Pivovarova and G.I. Karavaiko, Microbiology, 71 (2002) 319. 9. T.F. Kondrat'eva, T.A. Pivovarova, L.N. Muntyan and G.I. Karavaiko, Microbiology, 65 (1996) 59. 10. T.F. Kondrat'eva, T.A. Pivovarova and G.I. Karavaiko, Microbiology, 65 (1966) 675. 11. T.F. Kondrat'eva, T.A. Pivovarova, L.N. Muntyan and G.I. Karavaiko, Biohydrometallurgy and the environment toward the mining of the 21st century, R. Amils, A. Ballester (eds.), Elsevier, Amsterdam, Part B (1999) 89. 12. D.E. Rawlings and T. Kusano, Microbiological Reviews, 58 (1994) 39. 13. D.S. Holmes, J.R. Yates and J. Schrader, Biohydrometallurgy, P.R.Norris and D.P.Kelly (eds.), Antony Rowe Ltd., Chippenham, Wiltshire, Great Britan (1988) 153.

1387

A PPENDIX

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Author index

A Abin L., 1127 Acevedo F., 185, 219, 227 Acosta M., 1287 Adamov E.V., 85 Adryanczyk-Perrier G., 1147 Agate A.D., 101 Agatzini-Leonardou S., 41 Ageeva S.N., 1379 Ahmad W.A., 513, 859 Ahring B.K., 567 Alvarez-Rosales E., 419 Amann R., 1325 Amils R., 653, 1325 Andrade M.C., 835

Beskoski V.P., 345 Bevilaqua D., 1023 Bezovska M., 703, 719 Bipinraj N.K., 491 Blazquez M.L., 783 Bond P., 1337 Bosecker K., 585, 1127 Bowen A., 1337 Boyer A., 925 Brandl H., 457 Braun J.J., 1037 Breed A.W., 1215 Brown N.L., 1119, 1137 Bruhn D.F., 1195 Bunyok M.A., 513

B Baillet F., 925 Baker-Austin C., 1337 Ballester A., 783 Banderas A., 1361 Bandhapadhyay T., 843 Banerjee P.C., 291, 1297 Banghui G., 35 Barreto M., 1271 Bastiaens L., 671 Bath M., 387 Battaglia-Brunet F., 549, 639, 1313 Bauer-Cuya J., 947 Baxter-Plant V.S., 1147 Benedetti A.V., 1023 Beolchini F., 731

C Canales C., 185 Cara J., 653 Carballo M.T., 653 Carcamo C., 277 Carvalho N., 75 Casamayor E.O., 1325 Castillo A., 243 Chandraprabha M.N., 1099 Chassary P., 935 Chatterjee B.P., 869 Chatterjee S., 843, 869 Chen J.P., 807, 817 Chevard S., 549 Chiacchiarini P., 117 Chida T., 663 A-3

Author index

Cho K.S., 613, 631 Chockalingam E., 1037 Choi M. S., 631 Ciminelli V.S.T., 965 Clarens M., 1313 Clark T.R., 497 Cohen R.H., 497 Cossich E.S, 919 Costa M.C., 75 Coto O., 1127 Coupland K., 639, 1057 Cruz F., 1085 Curutchet G., 595, 681 D d’Hugues P., 147, 1313 d’Souza S.F., 975 da Costa W.C., 793 Dahrazma B., 253 Dave S.R., 211, 1279 Davodi M., 261 de Azevedo Gomes H., 793 de Carvalho R.J., 301 de Carvalho R.P., 741, 835 de J. Cerino-Cordova F., 1109 de Kock S.H., 319 de Sousa A-M.G., 741 de Vargas Parody I., 935 Deane S.M., 1249 Dementin S., 1147 Dempers C.J.N., 1215 Denamur A., 549 Deseure J., 337 Dianzuo W., 399 Dictor M.C., 639 Diels L., 671 Dixon D.G., 65 Donati E., 117, 595, 681 Dopson M., 1337 A-4

Doycheva A.S., 697 du Plessis C.A., 319 Duarte G., 965 Dymov I., 377 E Eidan L.S., 919 El Korchi-Buchert D., 997 Endo M., 689 Englert G.E., 1085 Escobar B., 1091 Essa A.M.M., 1119 F Farah C., 1361 Faramarzi M.A., 457 Fecko P., 703, 719 Fehmida F., 771 Ferrini M., 957 Ferron C.J., 377 Filho O.B., 301 Fischer H., 575 Fomina M.A., 749 Forster C.F., 525 Fouad M.Q., 771 Foucher S., 549 Fraguela P., 783 Freitas J.R., 741 Frenay J., 285 Frizan V., 117 Fugivara C.S., 1023 Fujii M., 1305 G Galan-Wong L.J., 419 Galvez-Cloutier R., 175 Garcia Frutos F.J., 235, 653 Garcia Jr. O., 559, 793, 911, 1023 Gardner M.N., 1249 Garrido F., 639

Author index

Geets J., 671 Gentina J.C., 185, 219, 227 Ghosh S., 1297 Giaveno A., 117 Glendinning K.J., 1137 Glombitza F., 575 Golan M., 1237 Goldschmidt G.K., 1249 Gomez F., 653 Gondrexon N., 337, 1109 Gonzalez F., 783 Gonzalez-Toril E., 1325 Groudev S.N., 533, 697 Groudeva V.I., 697 Guha A.K., 843, 869 Guibal E., 621, 899, 935, 947 Guiliani N., 1287, 1361 Gutierrez D., 1127 H Hallberg K.B., 639, 1057, 1163, 1179 Hansford G.S., 1215, 1227 Harneit K., 1369 Harrison S.T.L., 359 Harvey T.J., 387 Hatzikoseyan A., 711 Hevia M.J., 1091 Holmes D.S., 989, 1187, 1271 Hong J. H., 631 Hongxu L., 399 Horak G., 799 Humphries A.C., 525 Hurtado J.E., 1353 I Iglesias N., 75 Ignatiadis I., 549 Imanakunov B., 465 Inoue C., 663

J Jaapar J., 513 Janneck E., 575 Jedlicki E., 989, 1187, 1271 Jensen P.E., 567 Jerez C.A., 1287, 1361 Johnson D.B., 165, 639, 1057,1163, 1179, 1195 Joshi N.R., 491 K Kalman E., 1003 Kaltsa F., 711 Kamimura K., 449, 1261, 1305 Karamanev D., 647, 1067 Karapantsios T.D., 849 Karavaiko G.I., 85, 1389 Karnatz F., 575 Kasatkina T.P., 749 Kazy S.K., 975 Khijniak T., 483 Kim D.J., 631 Kim D.S., 631 Kimura S., 1057, 1179 Kinnunen P.H.-M., 193 Kinzler K., 1003 Klauber C., 1011 Kolousek D., 157 Komnitsas K., 533 Kondrateva T.F., 1379 Konno H., 689 Krambrock K., 741 Krylova N., 85 Kuncoro E.K., 621 Kusnierova M., 703, 719 Kwon H.K., 541 L Lamaignere V., 359 Lazic M.L., 1077 A-5

Author index

Lehtonen T., 621 Lengauer C., 157 Lilova K., 647, 1067 Llobet-Brossa E., 1325 Loi G., 429 Lopez J., 243 Lopez-Juarez A., 203 Loukidou M.X., 849 Lozovaya O.G., 749 Lugg H., 1205 Luptakova A., 703, 719 Ly Arrascue M., 947

Monhemius A.J., 409 Moon H.S., 613 Moran A., 653 Moreno L., 311 Morin D., 147, 549, 1313 Morra C., 337 Mugabi M., 147 Mukherjee A., 25 Müller I.L., 1085 Mulligan C.N., 175, 253 Munoz J.A., 783 Muntyan L.N., 1379

M Macaskie L.E., 525, 935, 1119, 1147, 1155, 1205 Maeda T., 449, 1305 Magnin J.P., 337, 925, 1109 Mahapatra N.R., 1297 Maliszewska I., 439 Mandal S.K., 291 Marquis P.M., 1205 Marrero J., 1127 Martinez L., 595 Massacci P., 957 Matic V., 345 Matis K.A., 849 Matlakowska R., 265, 1237 Medrano-Roldan H., 419 Medvedeva-Lyalikova N.N., 483 Mengel-Jung G., 585 Meruane G., 277 Mesquita H.M., 919 Metodiev B., 1067 Meyer G., 1369 Migliavacca E., 957 Mikheenko I.P., 1147 Miller H., 1163 Mitchell D., 1369 Modak J.M., 25, 1099

N Nakbanpote W., 891 Naldrett K., 319 Nancucheo I., 219 Nandi S., 1297 Natarajan K.A., 25, 759, 1037, 1099 Ndlovu S., 409 Nebera V.P., 107 Negishi A., 449, 1305 Nemati M., 359 Norris P.R., 1347 Nuzhat A., 771

A-6

O Ohimain E.I., 475 Okibe N., 165 Oliazadeh M., 261, 423, 507 Ottosen L.M., 567 Ouattara A., 175 Ozil P., 337, 925, 1109 P Pagnanelli F., 731, 825 Paknikar K.M., 491, 877 Palencia I., 75 Palmieri M., 911 Panin V.V., 85 Papassiopi N., 603

Author index

Park D., 883 Park J.M., 541, 883 Park K.H., 631 Parker A., 1011 Paspaliaris I., 533, 603 Paterson-Beedle M., 1155 Patino E., 285 Paulo P.S., 243 Pearce S.J.A., 359 Pehkonen S.O., 193 Peirano Blondet F., 947 Penfold D.W., 525 Pereyra-Alferez B., 419 Pertsov N., 127, 465 Petersen J., 65 Phillips W., 377 Pimentel P.F., 835 Pinka J., 575 Pivato D., 243 Pivovarova T.A., 85, 1379 Podgorsky V.S., 749 Podolska V., 127, 465 Polowczyk I., 439 Pompe W., 799 Puhakka J.A., 193 Q Quatrini R., 989 R Raichur A.M., 25 Raja S., 359 Ramirez P., 1287 Ramirez-Rodriguez D.G., 419 Ranville J.F., 497 Rawlings D.E., 1249 Remoudaki E., 711 Renman R., 399 Rivas M., 1271 Rivera-Santillán R.E., 203

Roberto F.F., 1195 Rohi R., 261 Rohwerder T., 997, 1171 Rojas-Chapana J., 1047 Romera E., 783 Rossi G., 3, 429 Rousset M., 1147 Roussy J., 621, 947 Rubinger C.P.L., 741 Rubio A., 235 Ruiz M., 935 Ryu H.W., 613, 631 S Sadowski Z., 439 Saha T., 843 Salazar M.F.M., 419 Salo V.L.A., 193 Sammons R.L., 1205 Sampaio C., 1085 Sand W., 997, 1003, 1171, 1369 Sandoval R., 285 Santos M.H., 835 Sar P., 975 Sasaki K., 689 Sastre A., 935 Savari E.Ε., 91 Savic D.S., 1077 Sayed Baghery S.A., 423, 507 Schippers A., 55, 585 Searby G.E., 1227 Sedelnikova G.V., 91 Sen A.M., 1179 Shahverdi A.R., 261, 423 Sheng P.X., 817 Shishikado T., 1261 Sidborn M., 311 Silva E.A., 919 Simmons S., 1347 Simonoff M., 483 A-7

Author index

Sissing A., 359 Sklodowska A., 265, 1237 Smith K.S., 497 Sohn J.S., 631 Sohor S.H., 513 Solis-Soto A., 419 Solozhenkin P.M., 107 Som Majumdar S., 843, 869 Songkroah C., 891 Spanelova M., 899 Spasic S., 345 Spasova I.I., 533 Stackebrandt E., 1369 Stagars M., 457 Stott M., 1011 Styriak I., 157, 1029 Styriakova I., 157, 1029 Subramanian S., 759, 1037 Sugio T., 449, 1261, 1305 Sung J.Y., 541 Suto K., 663 Svecova L., 899 T Takahashi Y., 663 Takano K., 689 Takeuchi F., 449, 1305 Tan L.H., 807 Tavares C.R.G., 919 Teixeira M.C., 965 Teledgi J., 1003 Ten W.K., 137 Thiravetyan P., 891 Thyagarajan H., 759 Tillmanns E., 157 Ting Y.P., 137, 329, 807, 817 Tipre D.R., 211, 1279 Toro L., 731, 825 Tributsch H., 1047 Trois P., 429 A-8

Tsezos M., 711 Tupe S., 877 U Ubaldini S., 825 Ukermann S., 1163 Ulberg Z., 127, 465 Uzma B., 771 V Valdes J., 1187 Valencia P., 227 Valenzuela L., 1287 van Broekhoven K., 671 van Bronswijk W., 1011 van Zyl L.J., 1249 Vargas T., 277, 1091 Vaxevanidou K., 603 Veglio F., 731, 825 Veljkovic V.B., 1077 Veloso F., 989 Viera M., 681 Vigliocco A., 243 Villar L.D., 559 Vora S.B., 211 Voronin D.Y., 85 Vos J., 671 Vrvic M.M., 345 W Walton-Day K., 497 Watling H.R., 1011 Welzel A., 1085 Werner P., 799 Willscher S., 497, 799 Woo S.H., 541 X Xianwan Y., 35 Xu T.J., 329

Author index

Y Yakubenko L., 127, 465 Yamakado M., 1261 Yermolenko A., 127. Yong P., 1205 Yoo N.Y., 613 Yuehua H., 399 Yun Y.-S., 883 Yun Z., 35

Yuxia G., 35 Z Zafiratos J.G., 41 Zahari M.A.K.M., 513 Zakaria Z.A., 859 Zeballos F., 301 Zouboulis A.I., 849

A-9

15th International Biohydrometallurgy Symposium (IBS 2003) September 14-19, Athens, Hellas "Biohydrometallurgy: a sustainable technology in evolution"

Subject index 16S rDNA, 1313 16S rRNA, 1325, 1369 A Accessory genes, 1249 Acid Mine Drainage (AMD), 497, 585, 639, 997, 1057, 1163, 1179 Acid production potential, 1037 streamers, 1057 sulphate soils, 475 Acidification, 475 Acidihalobacter, 1347 Acidiphilium, 1171 Acidithiobacillus, 85, 117, 423, 1171, 1353 caldus, 1249 ferrooxidans, 337, 429, 449, 595, 631, 647, 925, 989, 1003, 1023, 1037, 1047, 1067, 1077, 1099, 1109, 1187, 1261, 1271, 1287, 1305, 1361, 1369 thiooxidans, 613, 681, 1369 Acidocella strain, 1297 Acidophile heterotrophs, 497, 585 Acidophiles, 1179, 1337, 1347 Acidophilic bacteria, 1195 Activated carbon cloth, 1109 Activation energy, 291 Adaptation, 1099, 1379 Adsorption, 439, 891 Aeromonas, 859 Agar-plate screening, 749 Air-lift, 211

Algae, 783 Algal biomass, 947 Analytical aspects, 345 Antimony, 107 Archaea, 219, 319, 1337 Arsenic, 603, 639, 799 precipitation, 377 Arsenite, 965 Arsenopyrite / Pyrite, 377 Arsenopyrite, 377, 1353 Aspergillus niger, 137, 291, 329, 899, 1127 Atomic Force Microscopy (AFM), 1003 Attached cells, 1091 B Bacillus polymyxa, 759 sp., 157, 1029, 1137 Bacteria modified, 399 Bacterial activity, 277 attachment, 409 leaching, 75, 311 Barium, 793 Batch processing, 1215 Bench scale, 449 Beneficiate, 101 Bimetallic systems, 783 Bioaccumulation, 771 Biodegradation, 491, 541 Biodiversity, 1057, 1313 Biofilm, 541, 1109, 1361 formation, 1271 A-11

Subject Index

Biohydrometallurgy, 91, 107 Bioleaching, 25, 41, 85, 117, 127, 137, 147, 157, 203, 219, 235, 277, 285, 301, 319, 329, 359, 399, 429, 507, 559, 613, 1067, 1091, 1127, 1227, 1313 Biolixiviation, 175 Biological leaching, 377 Biomass estimation, 1091 Biomineralization, 1205 Biooxidation, 65, 219, 387, 513, 891, 1099 Bioreactors, 541, 1163, 1215 Bioreduction, 759 Bioremediation, 759, 965, 1163, 1179 Biosorption, 567, 731, 741, 749, 759, 771, 783, 793, 799, 807, 817, 825, 835, 843, 849, 859, 877, 883, 911, 919, 947 Biosurfactant, 439 Biotechnology, 35 Brown coal, 35 C Cadmium, 783, 849 Caesium, 1155 Calcium phosphate, 1205 sulphate, 703 Carbon dioxide, 319 fiber, 689 Catalytic effect, 203 Cellulose, 741 Chalcocite, 203 Chalcopyrite, 193, 203, 235, 399, 419, 1011, 1023, 1037, 1091 Chalk, 439 Chelation, 621 Chemoautotrophs, 1215 Chemolithotrophic microorganisms, 1379 A-12

Chitin, 891 Chitosan, 621, 869, 935 Chromate reduction, 525 Chromium, 711, 759, 807, 859, 883, 919 biosorption, 859, 919 removal, 681 Chromobacterium, 457 Chromosomal DNA, 1379 integration, 1297 Citric acid, 1127 Clay tiles, 595 Coal biodesulphurization, 653 Cobalt, 1155 Complex sulfides concentrates, 429 Complexing organic matter, 497 Computational fluid dynamics, 1067 Concentrates, 107 Continuous culture, 1227 Continuous electrochemical regeneration, 337 Converter slag, 507 Copper, 75, 253, 507, 631, 647, 741, 783, 799, 869 bioleaching, 311, 419 fixation, 925 heap, 423 mine, 1279 sulphides, 301 Corrosion patterns, 409 Crystal orientation, 409 CSTR, 185 Cyanide, 491, 541 Cyanogenic microrganisms, 457 Cysteine biosynthesis, 1187 D Deferritization, 1029 Design Of Experiment (DOE), 925 Desorption, 835 Desulfotomaculum nigrificans, 1037

Subject Index

Desulfovibrio, 681 fructosovorans, 1147 vulgaris, 525 DGGE, 1369 Dibenzothiophene, 345 Diversity, 1279 Dolomite, 439 Dredging, 475 E E. coli, 1297 EDTA, 603 Electric field, 465 Electrochemistry, 399 Electron paramagnetic resonance (EPR), 741 Electronic scrap materials, 137 Elemental sulfur, 91 Enargite, 219 Environmental risk potential, 585 Equilibrium, 731, 799, 825, 975 EXAFS, 965 Exopolysaccharide, 1337 Extracellular polysaccharides, 1271 Extremely thermophilic culture, 235 F FeoB, 989 Ferric leaching, 193 Ferrooxidans, 301 Ferrous iron catalysis, 235 oxidation, 277 Fixed-bed columns, 919 Flotation tailings, 85 Fluorescence in situ hybridization, 653 Fluorescent in situ hybridization (FISH), 1057 Fluvial tailings deposit, 497 Fly ash, 329 Forced aeration, 311

Fossil fuels, 345 FTIR, 741, 835 Fungal biomass, 947 G Galactose, 1271 Galena, 127 Galvanic couple, 127 Ge, 35 Genome analysis, 989, 1187, 1271 GeoBiotics process, 261 Geochemistry, 175 GEOCOAT®, 387 Gold recovery, 377 Goldbearing ores, 107 sulfide concentrates, 91 Gram-negative bacteria, 1279 Growth inhibition, 65 kinetics, 65 H Haloalkaliphilic bacteria, 483 Halomonas, 483 HCN, 457 Heap leaching, 65 Heavy metals, 175, 613, 671, 731, 771, 799, 817, 825, 883, 957, 1119 Hematite, 439 Heterotrophic microorganisms, 1127 High-cell density, 337 Horizontal gene pool, 1249 Hydrogen generation, 525, 663 peroxide, 689 sulphide, 663 uranyl phosphate, 1155 Hydroxyapatite, 1205 I Immobilization, 595 A-13

Subject Index

Immobilized cells, 525 Indigenous thiobacilli, 559 Indirect action, 277 Inhibition, 1099 Iron leaching, 291 oxidation, 193, 653, 1227 oxide, 55 regulation, 989 solubilisation, 227 uptake, 989 Ironoxidising bacteria, 265 oxidizing acidophiles, 585 reducing bacteria, 603 ISAfe1, 1353 Isotherms, 935, 947 K Kaolin, 439 Kinetic parameters, 1215 Kinetics, 935, 947, 1227 Klebsiella pneumoniae, 1119 L Lactuca Sativa, 835 Lanthanum, 911 Laterite ore, 1127 Lead, 843, 899, 947 Leptospirillum, 165 ferrooxidans, 429, 1003, 1047, 1369 Lignin, 741 Low-grade ores, 41 M Magnetite, 127 Maintenance requirements, 1215 Manganese, 41, 1163 oxide, 55 Mangrove, 475 Marine A-14

algal, 807 isolate, 25 Mechanism, 399 Membrane proteins, 1261 Mercury resistance, 1137 Mercury, 449, 621, 835, 877, 899, 1085, 1305 Mesophilic, 203 bacteria, 1279 Metabolic activity, 359 Metal, 749 binding, 975 complexed cyanide, 465 cyanides, 457, 541 recovery, 1179 removal, 849 resistance, 1297, 1337 sorption, 697 sulfide oxidation, 55 Metal-organic complexes, 711 Metals, 559 Microbial cell damage, 359 degradation, 465 leaching, 243 monitoring, 653 recovery, 631 sulphate reduction, 697 Microcalorimetry, 997 Mine drainage, 533 Mineral suspension, 439 Mining tailings, 253, 513 waste, 585, 997 Mixed cultures, 165, 513, 1195 Mn-oxidizing fungus, 689 Modelling, 301, 311, 731, 825, 849, 919, 925, 1099, 1215 Moderate acidophiles, 639 Molecular ecology, 1325 Monitoring, 997

Subject Index

Mucor rouxii, 843 Mud, 1085 N Neodymium, 911 Nickel, 783 Niger delta, 475 Non-metallics, 1029 Numerical modelling, 311 O Oil, 697 Open pit mining, 997 Optimal conditions, 227 Optimization, 329 Organic acids, 175, 1127 carbon source, 689 Oxalic acid, 291 Oxidation, 399, 647 Oxide ore, 1127 Oxygen, 319 P Packed bed reactor, 653 Palladium, 935, 1147 Passivation, 193 Penicillium simplicissimum, 1127 Permeable reactive barrier, 697 Pertechnetate reduction, 483 pH, 419, 483 stress, 1195 effect, 731 Phosphatic ores, 101 Photographic waste, 891 Piloting, 377 Plant biomass, 741 Plasmids, 1249, 1297, 1379 Platinum, 935 Polymetallic concentrate, 211 Polysulfide, 663

mechanism, 55 Powder microelectrode, 399 Precipitation of heavy metals, 719, 1119 Pressure drop, 621 oxidation, 377 Printed circuit board, 631 Process control, 147 Process flowsheet, 85 Proteomics, 1287, 1337 Pseudomonas aeruginosa, 439 putida, 491 sp., 975 Pulp density, 85 Pyrite, 147, 165, 203, 261, 387, 409, 475, 1003, 1037,1047 biooxidation, 185 oxidation, 55 Q Quantitative ecology, 585 Quorum sensing, 1361 R Radium, 793 Reaction rate, 291 Reactor configuration, 185 Refractory gold concentrate, 185 gold, 377 ore, 387 Regeneration, 663, 957 Rehabilitation, 475 Remediation, 567, 671 Removal, 345, 1155 Response surface, 227 Restoration, 475 Retention efficiency, 621 Revolving drum bioreactor, 429 Rhamnolipids, 253 A-15

Subject Index

Rhodanese, 1287 S Salt tolerance, 1347 Sargassum sp., 793, 835, 919 Seaweeds, 817 Sediments, 253 SEM, 265 Semi-continuous, 211 Serratia sp., 1155, 1205 Settling, 439 Sewage sludge, 265, 559, 613 Shrinking-core, 301 Silicate bacteria, 157 minerals, 1029 Silver, 891 Soil remediation, 603 Solids loading, 359 Solubility curves, 711 Sorption, 835, 899, 957, 965, 975 sites, 741 Spectroscopy, 1011 Sphalerite, 85 Spirogyra insignis, 783 SSCP, 1313 Stability, 439 Starvation stress, 1195 Static mixers, 337 Stirred tank, 211 Strain polymorphism, 1379 Strontium, 1155 Sulfate Reducing Bacteria (SRB), 549, 663, 703, 719, 935, 1179 Sulfide, 387 concentrate, 147 ores, 117, 423 oxidation, 1171 Sulfidic dredge spoils / sediment, 475 Sulfolobus metallicus, 277, 319, 1091 A-16

Sulfur assimilation, 1187 cycle, 663 dioxygenase, 1171 metabolism, 1287 oxidation, 285, 1171 regulation, 1187 Sulfuroxidizing acidophiles, 585 oxidising bacteria, 265 oxidizing neutrophiles, 585 Surface treatment effluents, 549 Swarming, 1353 Synergism, 165 T Tailing, 75 Tc(III), Tc(IV), Tc(VII), 483 TDS, 65 TEM, 265 Temperature, 419 Textile, 859 Thallium, 1237 Thermophiles, 165, 219, 1227 Thermophilic, 319, 1137 archaeon, 227 Thiobacillus, 41 ferrooxidans, 101, 261, 345, 399 Thiomonas sp., 639 Thiosulfate mechanism, 55 Thorium, 975 TonB dependent receptors, 989 Total solids, 559 U Ultrafiltration, 621 Uranium, 243, 533, 575, 975 Ureibacillus sp., 1137 W Waste biomass, 899

Subject Index

Wastewater, 465, 697, 849 Wetlands, 533, 575 X X-Ray photoelectron spectroscopy, 1011 Y Yeast, 749

Yield, 65 Z Zea Mays, 835 Zeolite, 157 Zeta-potential, 127, 439 Zinc, 75, 117, 783

A-17

Proceedings of the International Biohydrometallurgy Symposium IBS 2003, Athens, Greece - Vol I

Configurational Mechanics: Proceedings of the Symposium on Configurational Mechanics, Thessaloniki, Greece, 17-22 August 2003

Proceedings of the International Biohydrometallurgy Symposium IBS 2003, Athens, Greece - Vol II