Nitrogen Fixation: Methods and Protocols (Methods in Molecular Biology 766)

M E THODS IN M OLECULAR B IOLOGY Series Editor John M. Walker School of Life Sciences University of Hertfordshire Ha...

Author: Markus W. Ribbe

147 downloads 1057 Views 6MB 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

M E THODS

IN

M OLECULAR B IOLOGY

Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK

For further volumes: http://www.springer.com/series/7651

TM

Nitrogen Fixation Methods and Protocols

Edited by

Markus W. Ribbe University of California, Irvine, CA, USA

Editor Markus W. Ribbe Department of Molecular Biology & Biochemistry University of California Irvine, California, USA [email protected]

ISSN 1064-3745 e-ISSN 1940-6029 ISBN 978-1-61779-193-2 e-ISBN 978-1-61779-194-9 DOI 10.1007/978-1-61779-194-9 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2011931959 © Springer Science+Business Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Humana Press is part of Springer Science+Business Media (www.springer.com)

Preface Nitrogenase is a complex metalloenzyme that catalyzes one of the most remarkable chemical transformations in biological systems: the nucleotide-dependent reduction of atmospheric dinitrogen to bioavailable ammonia (designated biological nitrogen fixation). The fundamental significance of this process has prompted vigorous research on nitrogenase. However, few problems in protein biochemistry have proven to be as challenging and recalcitrant as the molecular description of nitrogenase. Although progress has been made toward deciphering the enzymatic and biosynthetic mechanisms of this enzyme system, further development is hampered by the complexity of nitrogenase that makes it impossible to study this enzyme by any singular method. To overcome this problem, the research area of nitrogen fixation has evolved into a highly interdisciplinary field that tackles the remaining questions of nitrogenase mechanism and biogenesis with a combination of methods. This volume attempts to provide an up-to-date, in-depth overview of the methods that have been applied to studying the nitrogenase at a molecular level. A large ensemble of approaches is covered in this volume, ranging from genetic, biochemical, spectroscopic, and chemical methods to theoretical calculations. In addition, techniques used to study an enzyme system that is homologous to nitrogenase are described in this book. A project of this scope requires the timely cooperation of many participants, and I greatly appreciate the willingness of all authors to face and meet such a demanding schedule. I hope that this volume, written by recognized experts in their particular areas, will be useful for anyone who is interested in nitrogenase research and who is willing to take charge of addressing the remaining mechanistic and biosynthetic questions of this fascinating enzyme system. Irvine, California, USA

Markus W. Ribbe

v

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

SECTION I

NITROGEN FIXATION

1.

Historic Overview of Nitrogenase Research . . . . . . . . . . . . . . . . . . . . Yilin Hu and Markus W. Ribbe

3

2.

Mechanism of Mo-Dependent Nitrogenase . . . . . . . . . . . . . . . . . . . . Zhi-Yong Yang, Karamatullah Danyal, and Lance C. Seefeldt

9

3.

Assembly of Nitrogenase MoFe Protein . . . . . . . . . . . . . . . . . . . . . . Chi-Chung Lee, Aaron W. Fay, Jared A. Wiig, Markus W. Ribbe, and Yilin Hu

31

4.

Genomic Analysis of Nitrogen Fixation . . . . . . . . . . . . . . . . . . . . . . Ina P. O’Carroll and Patricia C. Dos Santos

49

5.

Enzymatic Systems with Homology to Nitrogenase . . . . . . . . . . . . . . . . Jürgen Moser and Markus J. Bröcker

67

SECTION II

GENETIC AND BIOCHEMICAL METHODS

6.

Molecular Biology and Genetic Engineering in Nitrogen Fixation . . . . . . . . . Patricia C. Dos Santos

81

7.

Purification of Nitrogenase Proteins . . . . . . . . . . . . . . . . . . . . . . . . Jared A. Wiig, Chi-Chung Lee, Aaron W. Fay, Yilin Hu, and Markus W. Ribbe

93

8.

Assays of Nitrogenase Reaction Products . . . . . . . . . . . . . . . . . . . . . . 105 William E. Newton and Michael J. Dilworth

9.

Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Jürgen Moser and Markus J. Bröcker

SECTION III

SPECTROSCOPIC METHODS

10. X-Ray Crystallography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Lauren E. Roth and F. Akif Tezcan 11. X-Ray Absorption Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . 165 Serena DeBeer 12. Small Angle X-Ray Scattering Spectroscopy . . . . . . . . . . . . . . . . . . . . 177 David W. Mulder and John W. Peters

vii

viii

Contents

13. Electron Paramagnetic Resonance Spectroscopy . . . . . . . . . . . . . . . . . . 191 Karamatullah Danyal, Zhi-Yong Yang, and Lance C. Seefeldt 14. Magnetic Circular Dichroism Spectroscopy Brian J. Hales

. . . . . . . . . . . . . . . . . . . . 207

15. Mössbauer Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 Boi Hanh Huynh SECTION IV

CHEMICAL METHODS

16. Protocols for Cofactor Isolation of Nitrogenase . . . . . . . . . . . . . . . . . . 239 Aaron W. Fay, Chi-Chung Lee, Jared A. Wiig, Yilin Hu, and Markus W. Ribbe 17. Techniques for Functional and Structural Modeling of Nitrogenase . . . . . . . . 249 Patrick L. Holland SECTION V

THEORETICAL CALCULATIONS

18. Nitrogenase Structure and Function Relationships by Density Functional Theory . 267 Travis V. Harris and Robert K. Szilagyi 19. Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 Gregory M. Sandala and Louis Noodleman Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

Contributors MARKUS J. BRÖCKER • Institut für Mikrobiologie, Technische Universität Braunschweig, Braunschweig, Germany KARAMATULLAH DANYAL • Department of Chemistry and Biochemistry, Utah State University, Logan, UT, USA SERENA DEBEER • Department of Chemistry and Chemical Biology, Cornell University, Ithaca, NY, USA MICHAEL J. DILWORTH • School of Biological Sciences and Biotechnology, Center for Rhizobium Studies, Murdoch University, Murdoch, WA, Australia PATRICIA C. DOS SANTOS • Department of Chemistry, Wake Forest University, WistonSalem, NC, USA AARON W. FAY • Department of Molecular Biology and Biochemistry, University of California, Irvine, CA, USA BRIAN J. HALES • Department of Chemistry, Louisiana State University, Baton Rouge, LA, USA TRAVIS V. HARRIS • Department of Chemistry and Biochemistry, Astrobiology Biogeochemistry Research Center, Montana State University, Bozeman, MT, USA PATRICK L. HOLLAND • Department of Chemistry, University of Rochester, Rochester, NY, USA YILIN HU • Department of Molecular Biology and Biochemistry, University of California, Irvine, CA, USA BOI HANH HUYNH • Department of Physics, Emory University, Atlanta, GA, USA CHI-CHUNG LEE • Department of Molecular Biology and Biochemistry, University of California, Irvine, CA, USA JÜRGEN MOSER • Institut für Mikrobiologie, Technische Universität Braunschweig, Braunschweig, Germany DAVID W. MULDER • Department of Chemistry and Biochemistry, Astrobiology Biogeocatalysis Research Center, Montana State University, Bozeman, MT, USA WILLIAM E. NEWTON • Department of Biochemistry, Virginia Polytechnic Institute & State University, Blacksburg, VA, USA LOUIS NOODLEMAN • Department of Molecular Biology, The Scripps Research Institute, La Jolla, CA, USA INA P. O’CARROLL • HIV Drug Resistance Program, National Cancer Institute at Frederick, Frederick, MD, USA JOHN W. PETERS • Department of Chemistry and Biochemistry, Astrobiology Biogeocatalysis Research Center, Montana State University, Bozeman, MT, USA MARKUS W. RIBBE • Department of Molecular Biology and Biochemistry, University of California, Irvine, CA, USA LAUREN E. ROTH • Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, CA, USA GREGORY M. SANDALA • Department of Molecular Biology, The Scripps Research Institute, La Jolla, CA, USA

ix

x

Contributors

LANCE C. SEEFELDT • Department of Chemistry and Biochemistry, Utah State University, Logan, UT, USA ROBERT K. SZILAGYI • Department of Chemistry and Biochemistry, Astrobiology Biogeochemistry Research Center, Montana State University, Bozeman, MT, USA F. AKIF TEZCAN • Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, CA, USA JARED A. WIIG • Department of Molecular Biology and Biochemistry, University of California, Irvine, CA, USA ZHI-YONG YANG • Department of Chemistry and Biochemistry, Utah State University, Logan, UT, USA

Section I Nitrogen Fixation

Chapter 1 Historic Overview of Nitrogenase Research Yilin Hu and Markus W. Ribbe Abstract The history of nitrogenase research dates all the way back to the 1800s. This chapter provides a brief account of the advances in this particular research area over the past few hundred years, which include such events as the initial discovery of biological nitrogen fixation, the preparation of active cell-free extracts, the purification of nitrogenase enzyme, the proposal of the Thorneley–Lowe model, and the report of x-ray crystallographic structures of the component proteins of nitrogenase. Key words: Biological nitrogen fixation, nitrogenase, cell-free extracts, purification, x-ray crystallography.

1. The Discovery of Biological Nitrogen Fixation in 1888

The year 1888 represents one of the most important years in the history of biological nitrogen fixation. In this year, Hellriegel and Wilfarth demonstrated conclusively that the legumes could use atmospheric nitrogen (N2 ) for growth and that the fixation of nitrogen into the plant depended on bacteria that could be found in its root nodules (1). Despite the fact that there had been many claims before about the existence of biological nitrogen fixation, this report was the first in which this process was proven to exist without any doubt. In the same year, Beijerinck isolated the responsible bacteria from the root nodules (2) and the combination of these reports formed the foundation for textbooks on nitrogen fixation for decades to come. Several years later, the free-living nitrogen-fixing soil-bacteria Clostridium pasteurianum, Azotobacter chroococcum, and A. agilis were discovered (3, 4), and these discoveries resulted in attempts to establish

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_1, © Springer Science+Business Media, LLC 2011

3

4

Hu and Ribbe

research programs on this topic in many countries. On the other hand, although these organisms attracted a great deal of attention at this time, only few detailed studies on their physiology and metabolic reactions were reported. This was mainly due to the fact that techniques available at the time were not adequate to tackle the scientific problems in the field, ultimately forcing many scientists to leave this research area to perform easier experiments elsewhere. One major methodical barrier was the availability of a reliable method for the determination of nitrogen. The Kjeldahl assay was the only tool available to early scientists for nitrogen quantification; unfortunately, this method is not very sensitive, nor is it very accurate. It was apparent that the field was in a dire need for new techniques in order to revitalize the interest and ensure further progress in this scientific topic.

2. Major Advances Before 1960

3. The Successful Preparation of Nitrogen-Fixing Cell-Free Enzyme Extracts in 1960

A paper published in the year 1928 by Meyerhof and Burk (5) led to a revival of the interest in the chemistry and biology of nitrogen fixation. This was due to (i) the introduction of the use of a microrespirometer as a technique for the indirect measurement of fixed nitrogen and (ii) the development of concepts that allowed a detailed examination of the properties of a specific enzyme system in growing cells. Two years later, Bortels reported the dependency of the nitrogenase of A. chroococcum on molybdenum (6). In 1942, Burris presented the first conclusive evidence that ammonia—the compound terminating the inorganic phase of biological nitrogen fixation—might be the key intermediate of this process that was eventually incorporated into the cell proteins (7). In the meantime, despite the gradual recognition of certain fundamental biochemical properties of nitrogenase, further investigations of this enzyme system would still require a major methodological breakthrough.

The first cell-free enzyme preparation that had the consistent ability to fix N2 was reported in 1960 by Carnahan and his colleagues (8). This report marked the arguably most important breakthrough in the research field of nitrogen fixation and paved the way for meaningful investigations of the enzyme system

Historic Overview of Nitrogenase Research

5

required for nitrogen fixation on a molecular level. The secret to the success of Carnahan and his colleagues’ work was the exclusion of air from their preparations of active cell-free extracts from C. pasteurianum. Following this work, the preparation of active cell-free extracts from several other anaerobic organisms, such as Klebsiella pneumoniae, was reported within a month (9). However, the preparation of active extracts from an obligate, aerobic organism has remained unaccomplished for several more years. This task was eventually accomplished in 1964 by Bulen and colleagues, when they successfully prepared active cell-free extracts from A. vinelandii (10). The same group of researchers immediately followed their success by another discovery that sodium dithionite could be used as a nearly universal electron donor in studies of the nitrogenase enzyme system (11). Furthermore, they identified the last missing pieces of the puzzle for a highly active cell-free extract of nitrogenase: ATP and ATP-regenerating system (comprising creatine phosphate, and creatine kinase). Addition of these compounds served to (i) regenerate ATP from ADP, thus providing a constant supply of energy source for the enzymatic activity; and (ii) remove ADP from the reaction, which was an inhibitor for the enzyme system.

4. The Identification of the Binary Nature of Nitrogenase After 1960

The discovery of the key parameters for the reproducible preparation of nitrogen-fixing cell-free enzyme extracts laid the foundation for the subsequent success in the purification of nitrogenase enzyme. Between 1966 and 1967, Bulen, LeComte, Mortenson, and colleagues reported that nitrogenases of both A. vinelandii and C. pasteurianum were composed of two brown proteins and that both protein components were required for the enzymatic activity of nitrogenase (12, 13). The two nitrogenases from A. vinelandii and C. pasteurianum were very similar in properties: one component showed a relatively higher molecular mass (200,000–250,000) and contained molybdenum, iron, and acid-labile sulfur; the other was a smaller protein (50,000– 60,000) and contained only iron and acid-labile sulfur. Both component proteins were composed of multiple subunits and could be irreversibly destroyed by air. These nitrogenase proteins were given numerous different names in the follow-up publications. However, the most commonly used designations are molybdenum–iron (MoFe) protein or dinitrogenase for the larger component and iron (Fe) protein or dinitrogenase reductase for the smaller component.

6

Hu and Ribbe

5. Investigations of Nitrogenase on a Molecular Level Since the 1990s

The availability of nitrogenase in the purified state sets the stage for extensive investigations of this enzyme system on a molecular level. In 1985, Lowe and Thorneley proposed a model of nitrogenase mechanism, which has remained the best description of nitrogenase reaction to date (14). Between 1992 and 1993, Rees and coworkers solved the crystal structures of both component proteins of nitrogenase from A. vinelandii (15–17). These structural reports represented one of the biggest accomplishments in the field since the purification of nitrogenase. Subsequently, the structurally derived information prompted chemical synthetic efforts of the metal centers of nitrogenase (18–23) and, together with the genetically based knowledge (24; also see Chapter 4), it facilitated mechanistic (25–27; also see Chapter 2) and biosynthetic (28–30; also see Chapter 3) investigations of this unique enzyme system. Despite the significant advances in the field over the past decades, the complexity of nitrogenase and the difficulty in protein handling have made it clear to the community that it is impossible to obtain a detailed molecular description of nitrogenase on the basis of a singular approach. Thus, recent efforts toward addressing the remaining questions in the research area of nitrogenase often feature collaborations between various groups, taking full advantage of combined genetic (see Chapter 6), biochemical (see Chapters 7 and 8), spectroscopic (see Chapters 10, 11, 12, 13, 14, and 15), chemical (see Chapters 16 and 17), and theoretical (see Chapters 18 and 19) methods. Such an interdisciplinary approach may prove instrumental in solving the biological riddle of nitrogen fixation in the near future.

Acknowledgments The authors are supported by the National Institutes of Health grant GM 67626 (M.W.R.) and Herman Frasch Foundation grant 617-HF07 (M.W.R.). References 1. Hellriegel H, Wilfarth H (1888) Untersuchungen über die Stickstoffnährung der Gramineen und Leguminosen. Beil Z Ver dt ZuchInd, 1–234

2. Beijerinck MW (1888) Die Bakterien der Papilionaceen-Knölchen. Bot Ztg 46:725–735; 741–750; 757–771; 781–790; 797–804

Historic Overview of Nitrogenase Research 3. Winogradsky S (1893) Sur l’assimilation de l’azote gaseux de l’atmosphère par les microbes. C r hedb Séanc Acad Sci Paris 116:1385–1388 4. Beijerinck MW (1901) Über oligonitrophile Mikroben. ZentBl Bakt ParasitKde 7:561–582 5. Meyerhof O, Burk D (1928) On the fixation of air nitrogen through Azotobacter. Z Phys Chem 139:117–142 6. Bortels H (1930) Molybdän als Katalysator bei der biologischen Stickstoffbindung. Arch Mikrobiol 1:333–342 7. Burris H (1942) Distribution of isotopic nitrogen in Azotobacter vinelandii. J Biol Chem 143:509–517 8. Carnahan JE, Mortenson LE, Mower HF et al (1960) Nitrogen fixation in cellfree extracts of Clostridium pasteurianum. Biochim Biophys Acta 38:188–189 9. Burris RH (1969) Progress in the biochemistry of nitrogen fixation. Proc R Soc Lond B Biol Sci 172:339–354 10. Bulen WA, Burns RC, LeComte JR (1964) Nitrogen fixation: cell-free system with extracts of Azotobacter. Biochem Biophys Res Commun 17:265–271 11. Bulen WA, Burns RC, Lecomte JR (1965) Nitrogen fixation: hydrogensulfite as electron donor with cell-free preparations of Azotobacter vinelandii and Rhodospirillum rubrum. Proc Natl Acad Sci USA 53: 532–539 12. Bulen WA, LeComte JR (1966) The nitrogenase system from Azotobacter: twoenzyme requirement for N2 reduction, ATPdependent H2 evolution, and ATP hydrolysis. Proc Natl Acad Sci USA 56:979–986 13. Mortenson LE, Morris JA, Jeng DY (1967) Purification, metal composition and properties of molybdoferredoxin and azoferredoxin, two of the components of the nitrogenfixing system of Clostridium pasteurianum. Biochim Biophys Acta 141:516–522 14. Thorneley RNF, Lowe DJ (1985) Kinetics and mechanism of the nitrogenase enzyme system. In: Spiro, TG (ed) Molybdenum Enzymes, pp. 221–284. Wiley, New York, NY 15. Georgiadis MM, Komiya H, Chakrabarti P et al (1992) Crystallographic structure of the nitrogenase iron protein from Azotobacter vinelandii. Science 257:1653–1659 16. Kim J, Rees DC (1992) Structural models for the metal centers in the nitrogenase molybdenum-iron protein. Science 257:1677–1682 17. Chan MK, Kim J, Rees DC (1993) The nitrogenase FeMo-cofactor and P-cluster pair: 2.2 A resolution structures. Science 260:792–794

18. Lee SC, Holm RH (2003) Speculative synthetic chemistry and the nitrogenase problem. Proc Natl Acad Sci USA 100: 3595–3600 19. Lee SC, Holm RH (2004) The clusters of nitrogenase: synthetic methodology in the construction of weak-field clusters. Chem Rev 104:1135–1158 20. Groysman S, Holm RH (2009) Biomimetic chemistry of iron, nickel, molybdenum, and tungsten in sulfur-ligated protein sites. Biochemistry 48:2310–2320 21. Coucouvanis D (1994) Fe/S and Fe/Mo/S clusters as speculative models for the metal centers in uncommon Fe/S proteins and the Fe/Mo protein of the nitrogenases. Adv Inorg Biochem 9:75–122 22. Ohki Y, Sunada Y, Honda M, Katada M, Tatsumi K (2003) Synthesis of the P-cluster inorganic core of nitrogenases. J Am Chem Soc 125:4052–4053 23. Holland PL (2008) Electronic structure and reactivity of three-coordinate iron complexes. Acc Chem Res 41:905–914 24. Setubal JC, dos Santos P, Goldman BS, Ertesvåg H, Espin G, Rubio LM, Valla S, Almeida NF, Balasubramanian D, Cromes L, Curatti L, Du Z, Godsy E, Goodner B, Hellner-Burris K, Hernandez JA, Houmiel K, Imperial J, Kennedy C, Larson TJ, Latreille P, Ligon LS, Lu J, Maerk M, Miller NM, Norton S, O’Carroll IP, Paulsen I, Raulfs EC, Roemer R, Rosser J, Segura D, Slater S, Stricklin SL, Studholme DJ, Sun J, Viana CJ, Wallin E, Wang B, Wheeler C, Zhu H, Dean DR, Dixon R, Wood D (2009) Genome sequence of Azotobacter vinelandii, an obligate aerobe specialized to support diverse anaerobic metabolic processes. J Bacteriol 191:4534–4545 25. Seefeldt LC, Hoffman BM, Dean DR (2009) Mechanism of Mo-dependent nitrogenase. Annu Rev Biochem 78:701–722 26. Howard JB, Rees DC (2006) How many metals does it take to fix N2 ? A mechanistic overview of biological nitrogen fixation. Proc Natl Acad Sci USA 103: 17088–17093 27. Burgess BK, Lowe DJ (1996) Mechanism of molybdenum nitrogenase. Chem Rev 96:2983–3012 28. Hu Y, Fay AW, Lee CC, Yoshizawa J, Ribbe MW (2008) Assembly of nitrogenase MoFe protein. Biochemistry 47:3973–3981 29. Dos Santos PC, Dean DR, Hu Y, Ribbe MW (2004) Formation and insertion of the nitrogenase iron-molybdenum cofactor. Chem Rev 104:1159–1173 30. Schwarz G, Mendel RR, Ribbe MW (2009) Molybdenum cofactors, enzymes and pathways. Nature 460:839–847

7

Chapter 2 Mechanism of Mo-Dependent Nitrogenase Zhi-Yong Yang, Karamatullah Danyal, and Lance C. Seefeldt Abstract Nitrogenase is the enzyme responsible for biological reduction of dinitrogen (N2 ) to ammonia, a form usable for life. Playing a central role in the global biogeochemical nitrogen cycle, this enzyme has been the focus of intensive research for over 60 years. This chapter provides an overview of the features of nitrogenase as a background to the subsequent chapters of this volume that detail the many methods that have been applied in an attempt to gain a deeper understanding of this complex enzyme. Key words: Nitrogen fixation, Fe protein, MoFe protein, mechanism, metalloenzyme, MgATP.

1. Nitrogen Fixation Dinitrogen (N2 ) is the major constituent (79%) of the Earth’s atmosphere, representing the largest global pool of nitrogen (N). While nitrogen is essential to all life, the vast reservoir of dinitrogen in the atmosphere is unusable by most organisms (1, 2). This is largely a consequence of the high bond dissociation energy for the N2 triple bond (3), making the breaking of this bond and “fixation” of the nitrogen to forms usable to living organisms energetically challenging. Dinitrogen can be fixed with considerable energy input by addition of electrons and protons to yield two ammonia (NH3 ) molecules. In the industrial Haber–Bosch process for fixing dinitrogen, the reaction is carried out at high temperatures (∼450◦ C) and pressures (>200 atm) in the presence of an iron catalyst, with the electrons and protons coming from H2 (1, 4–6). This process is extremely energy demanding, utilizing approximately 1% of the total fossil fuel used globally (7).

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_2, © Springer Science+Business Media, LLC 2011

9

10

Yang, Danyal, and Seefeldt

The other major pathway for reduction of dinitrogen is through the action of select microorganisms (called diazotrophs) that carry out a process called biological nitrogen fixation (6, 8). The first step in the assimilation of N2 by these organisms is the reduction of N2 to two ammonia molecules catalyzed by a complex metalloenzyme called nitrogenase (9, 10). Nitrogenases occur across a wide range of microbes and sequencing of the nitrogenase genes (nif genes) reveals considerable sequence diversity among the enzymes (10–12). Despite this sequence diversity, the majority of nitrogenases share some common features. For example, most enzymes are composed of two component proteins: a large component having at least an α2 β2 subunit composition and a smaller component having a γ2 subunit composition. All known nitrogenases contain iron–sulfur clusters in both component proteins. The site of N2 binding and reduction is one of the iron–sulfur clusters contained in the larger component protein. This active site metal cluster can contain, in addition to Fe atoms, a heterometal atom (Mo or V) (9, 10, 13, 14). The best studied nitrogenase is the Mo-dependent enzyme, which appears to be the paradigm for nitrogenases (9, 13, 15–21). Given that most of the mechanistic information known about nitrogenases is for the Mo-based enzyme, this chapter will focus on this enzyme. Other nitrogenases, sometimes called alternative nitrogenases, are reviewed elsewhere (10, 12).

2. Mo-Dependent Nitrogenase: Overview

The two component proteins of the Mo-dependent nitrogenase are called the iron (Fe) protein (or dinitrogenase reductase) and the molybdenum–iron (MoFe) protein (or dinitrogenase) (Fig. 2.1). These two component proteins work together to catalyze the reduction of dinitrogen in a complex reaction with an ideal reaction stoichiometry shown as follows (22): N2 + 8e − + 16MgATP + 8H+ → 2NH3 + H2 + 16MgADP + 16Pi [1]

A breakthrough in understanding nitrogenase came from the X-ray crystal structures of the component proteins solved individually (23–42) and when bound together (43–46). The Fe protein, the only known reductant of the MoFe protein that can support substrate reduction, is a homodimer that contains two nucleotide (MgATP or MgADP) binding sites, one on each subunit, and a single [4Fe–4S] cluster that bridges the two subunits (Fig. 2.1) (25). The MoFe protein is an α2 β2

Mechanism of Mo-Dependent Nitrogenase

11

Fig. 2.1. Crystal structure of the Fe and MoFe protein components of Mo-dependent nitrogenase showing the nucleotides, metal clusters, and electron transfer pathways. (Left) Cartoon representation of MoFe protein (pdb code: 1M1N) with the α-subunits and the β-subunits and Fe protein (pdb code: 1FP6) with the γ-subunits. (Right) Structures of MgADP and the three metalloclusters of nitrogenase. The figure was generated using the computer program PyMol.

heterotetramer. Each αβ dimeric unit contains two unique metalloclusters: a P-cluster ([8Fe–7S]) and an FeMo cofactor ([7Fe9S-Mo-X-(R)-homocitrate]) (Fig. 2.1) (27, 28, 47). Each αβunit appears to function as a catalytic unit independent of the other αβ pair. During the catalytic cycle, an Fe protein binds transiently to one MoFe protein αβ unit. During this encounter, one electron is transferred from the [4Fe–4S] cluster of the Fe protein to the MoFe protein. This electron transfer step is coupled to the hydrolysis of a minimum of two MgATP molecules (16, 48). Following electron transfer and ATP hydrolysis, the Fe protein disengages from the MoFe protein and a new Fe protein binds in its place to repeat the cycle. Given that only one electron is transferred per cycle, a minimum of eight encounters must occur to reduce N2 (equation [1]). 2.1. Fe Protein

The Fe protein is a homodimer (coded for by the nifH gene) with a molecular mass of approximately 64,000 Da (25, 49). In addition to delivering electrons to the MoFe protein, the Fe protein also is known to function in the maturation of the MoFe

12

Yang, Danyal, and Seefeldt

protein and in the bioassembly of the active site metal cluster called FeMo cofactor (49, 50). This maturation role for the Fe protein does not appear to require electron transfer or the ATP hydrolysis function (9) and is covered in another chapter of this volume. 2.1.1. Redox Properties of the [4Fe–4S] Cluster

The Fe protein contains a single [4Fe–4S] cluster that serves as a carrier of electrons. The X-ray structure of the Fe protein revealed that this cluster is symmetrically ligated between the two Fe protein subunits, with each subunit contributing two cysteine ligands (Fig. 2.1). The [4Fe–4S] is known to access three redox states (9, 21): Em = −300mV

Em = −790 mV

⎯⎯⎯⎯⎯ →[4Fe−4S]1+←⎯⎯⎯⎯ ⎯⎯⎯⎯⎯ →[4Fe−4S]0 [4Fe−4S]2+←⎯⎯⎯⎯ ⎯ ⎯

[2]

The [4Fe–4S]2+/1+ redox couple is operational during substrate reduction supported by reductants such as dithionite or ferredoxin (9, 51, 52). A more reduced (0) oxidation state (termed the “all-ferrous” state) can be achieved in vitro by incubating the Fe protein with reductants with very negative potentials (e.g., Ticitrate) (53–57). It has been demonstrated that the all-ferrous state can participate in the delivery of electrons to the MoFe protein when strong reductants are used (58). However, the role of this all-ferrous state during catalysis in vivo remains unknown. The 1+ oxidation state (FeRed ) of the [4Fe–4S] cluster is the dominant state in the as-purified enzyme in the presence of the reductant dithionite (S2 O2− 4 ) (59–61). This oxidation state of the Fe protein is paramagnetic, with the four Fe atoms distributed as 3Fe2+ and 1Fe3+ . This state gives rise to an EPR spectrum at low temperatures (∼5 K) that has been assigned as a mixture of two spin states (S = 1/2 and S = 3/2 spin). This mixed spin state has been confirmed in the Mössbauer and MCD spectra. The ratio of the two spin states can be shifted by addition of other reagents such as urea or glycerol into the sample solution (61). The 1+ oxidation state of the [4Fe–4S] cluster can be reversibly oxidized by the removal of one electron, achieving the 2+ oxidation state (FeOx ), with the iron atoms distributed as 2Fe2+ and 2Fe3+ (54, 61, 62). This oxidation state of the [4Fe– 4S] cluster is diamagnetic and therefore is EPR silent. The oxidation of the [4Fe–4S] cluster from the 1+ to the 2+ oxidation state can be achieved by the treatment of the Fe protein with redox-active dyes of sufficiently positive potential (e.g., thionine, methylene blue, and indigo disulfonate) (54, 63), whereas reduction from the 2+ to the 1+ oxidation state can be achieved by the addition of a number of reductants (e.g., dithionite) (64). This reversibility allows the establishment of the midpoint reduction potential (Em ) for the [4Fe–4S]2+/1+ redox couple using

Mechanism of Mo-Dependent Nitrogenase

13

voltametric and coulometric methods (65–67). The values of Em are dependent on the organism from which the Fe protein is purified and the presence or absence of bound nucleotides (9). The Em for the [4Fe–4S]2+/1+ couple of the Fe protein from Azotobacter vinelandii is measured to be –300 mV in the absence of nucleotides (equation [2]) (62, 68). When MgATP is added to the protein, the Em value shifts more negative to –430 mV (68). MgADP shifts the Em to –440 mV (68). It is well established from kinetic and spectroscopic studies that the [4Fe–4S]2+/1+ redox couple of the [4Fe–4S] cluster in the Fe protein is functional during the catalytic cycle of nitrogenase (9, 16, 21). While the Fe protein in the 1+ oxidation state is bound to the MoFe protein, an electron is transferred from the Fe protein to the MoFe protein, resulting in oxidation of the [4Fe–4S] cluster to the 2+ oxidation state. The consensus model requires the oxidized Fe protein (2+) to dissociate from the MoFe protein and for the [4Fe–4S] cluster to be reduced back to the 1+ oxidation state by a reduced electron carrier protein (e.g., flavodoxin or ferredoxin) (51, 52), thereby readying the Fe protein for another round of electron transfer to the MoFe protein. 2.1.2. The Fe Protein Binds Nucleotides

Early work on nitrogenase revealed that the Fe protein could bind nucleotides and that the hydrolysis of nucleotides by the nitrogenase complex was critical to the transfer of an electron from the Fe protein to the MoFe protein (16, 69). The Fe protein binds two nucleotides, one to each subunit. The nucleotide binding sites on the Fe protein are on the opposite end of the Fe protein from the [4Fe–4S] cluster (30, 31, 37, 38, 41, 43–46). The dissociation constants (Kd ) for nucleotide binding to Fe protein have been determined by a number of techniques (70). Recent studies using isothermal titration calorimetry have supported earlier studies showing that the redox state of the [4Fe–4S] cluster impacts the affinity for nucleotide binding (71). The 2+ oxidation state binds MgATP with the highest affinity (Kd = 45 μM), while the 1+ oxidation state has a lower affinity (Kd = 500 μM) for this nucleotide. Further, these studies have revealed positive cooperativity in the binding of the two nucleotides to both the reduced and oxidized states of the Fe protein (72). For the reduced state (FeRed ) the Kd values for the binding of the first and second MgATP molecules are reported to be Kd1 = 500 μM and Kd2 = 170 μM, respectively. A divalent metal is required for the binding of nucleotides to the Fe protein. While a number of different metals will work, it is thought that Mg2+ is the physiologically relevant metal (73). The Fe protein can also bind other nucleotide triand di-phosphates (e.g., GTP and GDP) with reasonable affinity, although it is widely believed that ATP and ADP are the relevant nucleotides in vivo (74).

14

Yang, Danyal, and Seefeldt

While the Fe protein binds MgATP, it shows undetectable rates of hydrolysis in the absence of the MoFe protein (21). It is only after the Fe protein binds to the MoFe protein that hydrolysis is activated. This observation has been explained from examination of X-ray structures of the Fe protein with bound nucleotides as the movement of a likely catalytic base into place to activate hydrolysis following Fe protein binding to the MoFe protein. A detailed understanding of the specific interactions of nucleotides with the Fe protein has been achieved from the X-ray structures of Fe proteins in various nucleotide-bound states (43–46). 2.1.3. Nucleotide Binding Induces Protein Conformational Changes in the Fe Protein

There is ample evidence showing that the binding of nucleotides to the Fe protein induces conformational changes to the protein structure that impact many aspects of its function. For example, the binding of MgATP or MgADP to the Fe protein shifts the Em for the [4Fe–4S]2+/1+ redox couple to more negative potentials by about –120 mV (described above). Many other methods also reveal that nucleotides change the properties of the [4Fe–4S] cluster (16, 21). What is clear is that these changes are not the result of nucleotides binding directly to the [4Fe–4S] cluster, but rather a result of nucleotide-induced protein conformational changes impacting the cluster over a distance. The nucleotide binding sites are located approximately 15 Å away from the [4Fe–4S] cluster (30, 31, 37, 38, 41, 43–46). While a number of studies report changes in the electronic properties of the [4Fe–4S] cluster as a result of nucleotides binding to the Fe protein, it is more recent studies using both small angle X-ray scattering (SAXS) (75) and X-ray crystallography that are providing clearer pictures of the larger structural changes induced in the Fe protein upon nucleotide binding. These are reviewed in other chapters of this volume.

2.2. MoFe Protein

The MoFe protein is an α2 β2 heterotetramer (Mr ∼ 250,000 Da) with the α and β subunits encoded by the nifD and nifK genes, respectively (9, 49). Each αβ-dimeric catalytic unit contains one active site metallocluster, the FeMo cofactor ([7Fe-9S-MoX-(R)-homocitrate]) (76) and one electron carrier cluster, the P-cluster ([8Fe–7S]). The X-ray crystal structure of the MoFe protein reveals that the FeMo cofactor is embedded solely in the α-subunit, while the P-cluster is located at the interface between the α and β subunits (23, 24). Several structures of the complex of the Fe protein bound to the MoFe protein reveal the interfaces where the Fe protein and MoFe protein dock (43–46). These structures place the P-cluster directly in-line and between the Fe protein [4Fe–4S] cluster and the FeMo cofactor (Fig. 2.1). The distance between the [4Fe–4S] cluster and the P-cluster varies depending on the nucleotide-bound state of the Fe protein,

Mechanism of Mo-Dependent Nitrogenase

15

Fig. 2.2. Structure of P-cluster in oxidized and reduced states. The reduced state (PN ) of the P-cluster (a) and the oxidized state (POX ) of the P-cluster (b) are shown (3MIN.pdb and 2MIN.pdb, respectively).

leading to a model wherein one role of nucleotides is to alter this electron transfer distance and therefore the electron transfer rate. The arrangement of the three metalloclusters suggests an electron transfer chain from the [4Fe–4S] cluster to the P-cluster and finally to the FeMo-cofactor active site (46). 2.2.1. P-Clusters

Early Mössbauer spectroscopic studies of the MoFe protein revealed that the P-cluster was composed of eight ferrous Fe atoms in the resting state in the presence of dithionite (termed the PN state) (77). The X-ray structures revealed the nature of this unusual cluster (Fig. 2.2) as being composed of two cubic [4Fe–4S] subclusters sharing a common sulfide ligand at one corner (23, 24, 27, 33, 42). Each Fe atom is coordinated by two or three sulfide ligands and one terminal or bridging cysteinyl ligand from a cysteine residue in the α or β subunit. Upon oxidation of the MoFe protein, the P-cluster is oxidized by one or two electrons, which results in significant structural rearrangement (33, 42). Upon oxidation, one of the cubic units is opened up with two Fe–S bonds (Fe5–S1 and Fe6–S1) being cleaved and two novel Fe6–O and Fe5–N bonds being formed. Further, a serinate-O (β-188Ser ) and a backbone amide-N (α-88Cys ) become ligands to Fe atoms (Fig. 2.2) (42). From in vitro studies using dye oxidants, it has been shown that the resting state of the P-cluster (PN ) can be oxidized by up to three electrons (P1+ , P2+ , and P3+ ) (equation [3]) (78–81). The Em values measured for these redox couples are shown for the A. vinelandii MoFe protein as follows (62, 82, 83): ?

PRed

−309 −309

PN

P1+

+90

P2+

Em (mV)

P3+

[3]

The P1+ and P2+ oxidation states are often collectively referred to as the POx oxidation state, because both oxidation states are

16

Yang, Danyal, and Seefeldt

usually populated in oxidized states of the MoFe protein. The P3+ oxidation state is not reversible and so is not believed to be functioning during catalysis. More reduced states of the P-cluster from the PN state have not been observed and seem unlikely given that such a reduction would require reducing Fe atoms beyond the ferrous oxidation state. The Em for P2+/1+ redox couple is pH-dependent, shifting Em by –53 mV per increase of one pH unit (84). The relevance of this pH dependence is not known, but would be consistent with the coupling of proton and electron transfer reactions involving this cluster. 2.2.2. FeMo Cofactor—The Active Site of Nitrogenase

The FeMo cofactor (76), also called the M-cluster, is embedded in each α subunit of the MoFe protein. The structure of FeMo cofactor was resolved when the X-ray structure of the MoFe protein was solved (47). The structure revealed a heterometallocluster with a composition [7Fe-9S-Mo-X-(R)-homocitrate] (Fig. 2.3) (28). The cluster is ligated to the peptide matrix through one cysteine ligand (α-275Cys ) bound to the Fe atom at one end and through one histidine ligand (α-442His ) bound to the Mo atom at the other end. The six Fe atoms in the middle part are arranged as a prismatic structure with each Fe atom coordinated by three sulfide atoms. Recent high-resolution structures of the MoFe protein have revealed the presence of a light atom (C, N, or O) at the center of the Fe cage of unknown identity (designated as X) that is presumed to be bound to each of the central six Fe atoms (28). Homocitrate provides two oxygen atom (C1 carboxylate and C3 hydroxylate) ligands to the Mo (85). Thus, the overall structure of the FeMo cofactor can be viewed as one [4Fe-3S-X]

Fig. 2.3. Structure of FeMo cofactor and some of the amino acids surrounding it. Numbering of atoms uses the system in the original structure. Colors for atoms are Fe in rust, Mo in magenta, S in yellow, C in dark gray, N in blue, and O in red (online version only). The figure is generated from PDB file 1M1N.

Mechanism of Mo-Dependent Nitrogenase

17

cubane and one [Mo-3Fe-3S-X] cubane that are connected by three bridging sulfides with one shared μ6 -X atom at the center. Identifying the central atom X has proven difficult (69). ENDOR studies have suggested that it is not an exchangeable N atom, but have left open the possibility that it is a non-exchanging N atom (86–88). A vibrational spectroscopy technique (nuclear resonance vibrational spectroscopy or NRVS) supports the presence of a light atom at the center of the Fe cage, but does not resolve the identity of X (89). Likewise, a number of calculations support the presence of X, but do not provide a definitive assignment for X (90–93). To date, there remains no experimental evidence showing the involvement of X in the catalytic cycle of nitrogenase. Obviously, understanding X and its roles in catalysis remains a significant challenge for the field. FeMo cofactor can be reversibly oxidized or reduced from its resting state. The resting state of FeMo cofactor (MN ) occurs in the MoFe protein isolated in the presence of dithionite. This state is paramagnetic with a rhombic S = 3/2 spin EPR signal (94). Treatment of the MoFe protein with oxidizing dyes results in the one-electron oxidation to the MOx state (77, 78). The MOx state is diamagnetic (S = 0) and EPR silent. The Em for the MOx/N redox couple is about –40 mV (95). The MN state can be reduced. Incubation with the Fe protein in the presence of MgATP and dithionite results in the reduction of FeMo cofactor to an MR state with an integer spin (S > 1) state that is EPR silent (96, 97). The Em for the MN/R redox couple has not been measured, but has been estimated as –465 mV (98). The oxidation states of the Fe atoms and the Mo atom in the resting state of FeMo cofactor (MN ) have been examined by Mössbauer and ENDOR spectroscopies. The Mössbauer study suggested an assignment of [Mo4+ , 3Fe3+ , 4Fe2+ , 9S2– ] (97), which is supported by calculations using a model with the interstitial X atom (99). The 57 Fe ENDOR study suggested an assignment of [Mo4+ , 1Fe3+ , 6Fe2+ , 9S2– ] for the resting state FeMo cofactor (100). This later assignment is consistent with the result from calculations on the FeMo cofactor without an interstitial X atom (90). FeMo cofactor must accept multiple electrons (two or more) to complete the reduction of substrates. How these electrons are accumulated on FeMo cofactor is not known. Possible distribution between the P-clusters, FeMo cofactor, and bound intermediates remains to be established. FeMo cofactor can be extracted from the MoFe protein into organic solvents (76, 101). A number of studies have been conducted on such extracted FeMo cofactor. While some properties of the cofactor in solvent are similar to those for the cofactor in the protein (e.g., EPR spectrum), others are quite different. For example, the reactivity of FeMo cofactor in solvent is different from that of FeMo cofactor bound to the protein (e.g., substrate reduction ability) (102).

18

Yang, Danyal, and Seefeldt

A number of studies support FeMo cofactor as the site of substrate (and inhibitor) binding (19, 101), although the precise location of substrate binding is still being pursued (17–19). There are many possibilities for the substrate binding site: (1) the Mo atom; (2) one or more of the central Fe atom(s); and (3) some combination of Fe, S, and Mo atoms. Using spectroscopic methods such as ENDOR, there is growing evidence for binding of hydrides (103), alkynes (104, 105), and nitrogenous compounds (106–110) to one or more of the Fe atoms in the central portion of FeMo cofactor (Fig. 2.3). As yet, no experimental results have illustrated binding of any substrate or intermediate to Mo, although this possibility has not been ruled out. An important tool to gaining insights into substrate binding to FeMo cofactor has been substitution of amino acids in the MoFe protein using site-directed mutagenesis. An early study examined the roles of α-195His in nitrogenase catalysis (Fig. 2.3) (111–115). Substitution of the α-195His residue by glutamine resulted in a variant of the MoFe protein that cannot effectively reduce N2 or azide (N− 3 ), but which retained full rates of reduction of acetylene and protons. From these studies, it was concluded that α-195His might participate in delivery of protons during reduction of nitrogen-containing substrates. Slowing down proton delivery by substituting for α-195His has been exploited to trap presumed intermediate states during the reduction of a

Fig. 2.4. Some substrates for nitrogenase.

Mechanism of Mo-Dependent Nitrogenase

19

number of substrates including hydrazine (N2 H4 ) (106, 107), diazene (HN = NH) (109), and methyldiazene (MeN = NH) (107, 108). It is clear that α-195His is not the sole source of protons for substrate reduction as the rates of reduction of other substrates remain undisturbed when this residue is substituted (115). Recent work with MoFe proteins containing amino acid substitutions are providing solid evidence for the site of substrate binding on FeMo cofactor. These studies have recently been reviewed (13, 18, 19). While the physiological substrates for nitrogenase are N2 and protons, a number of other small, multiple bonded compounds have been demonstrated to be substrates. These have been extensively reviewed elsewhere (9, 19, 21, 116). Several of the substrates are shown in Fig. 2.4.

3. Nitrogenase Mechanism 3.1. Fe Protein–MoFe Protein Complex Formation

An essential step in the nitrogenase mechanism occurs when the Fe protein, with two bound MgATP molecules, associates with the MoFe protein. This associated complex is fleeting, existing for about 1 s during normal substrate reduction (9). Several events occur while the two proteins are associated, including the hydrolysis of the two MgATP molecules to two MgADP and two Pi molecules and the transfer of one electron from the Fe protein to the MoFe protein. The order of these two events has not been definitively established and is the subject of current studies (19, 20). The associated state of the nitrogenase complex has been trapped by a number of different approaches. Five different types of stable complexes that have been examined include the following: (1) A chemical cross-linked complex using a bifunctional chemical reagent 1-ethyl-3-[3-(dimethylamino) propyl]carbodiimide (EDC) (45, 117, 118). (2) A non-dissociating complex formed between the Fe protein from Clostridium pasteurianum and the MoFe protein from A. vinelandii (119–122). (3) A non-dissociating complex formed between an Fe protein with an amino acid deletion (127Leu ) and the MoFe protein that appears to mimic the ATP-bound state in the absence of ATP (44, 123). (4) A non-dissociating complex formed when ADP and AlF− 4 (or BeF− 3 ) are added to the Fe protein and MoFe protein. In this case, the AlF− 4 appears to be mimicking the departing phosphate following ATP hydrolysis (43, 124–127).

20

Yang, Danyal, and Seefeldt

(5) Relatively stable complexes of the MoFe protein and the Fe protein with MgADP or β,γ-methylene MgATP bound or without nucleotide bound (46). Analysis of these tight complexes by a number of different approaches, including X-ray crystallography, is providing insights into the nitrogenase complex such as the following: (1) AlF− 4 occupies the binding site where the γ-phosphate portion of MgATP was expected to be (43, 46). (2) The subunits of Fe protein are considerably reoriented during complex formation and nucleotide hydrolysis. The movement of two segments of amino acids in the Fe protein (called switches I and II) appears to connect to the [4Fe–4S] cluster, possibly controlling the nucleotideinduced changes in the properties of the [4Fe–4S] cluster (44, 46). (3) There are several distinct and mutually exclusive interaction sites on the MoFe protein surface that are selectively populated, depending on the Fe protein nucleotide state (46). (4) The distance between the Fe protein [4Fe–4S] cluster and the MoFe protein P-cluster changes by up to 5 Å depending on the nucleotide bound to the Fe protein (46). (5) The Em value of the [4Fe–4S] cluster and the P-cluster is shifted more negative when the two proteins are associated, favoring electron transfer from the Fe protein to the FeMo cofactor (123, 128). 3.2. Fe Protein Cycle

The Fe protein, being an ATP-dependent reductase of the MoFe protein, can be thought of as proceeding through a cycle during its function in the overall nitrogenase catalytic cycle (129, 130). This Fe protein cycle is summarized in Fig. 2.5. During nitrogenase catalysis, the reduced Fe protein, with the [4Fe–4S] cluster in its 1+ oxidation state, binds two MgATP molecules. The Fe protein in this state then associates with the MoFe protein (131). Within this complex, MgATP hydrolysis is activated and electron transfer occurs, followed by the dissociation of the two proteins. The spent Fe protein is reactivated by replacing MgADP with MgATP and reducing the 2+ oxidation state to the 1+ oxidation state (Fig. 2.5). While the general features of the Fe protein cycle are known, several important details remain to be resolved. For example, how does complex formation activate MgATP hydrolysis and electron transfer? Which comes first, electron transfer or nucleotide hydrolysis? How specifically is the energy from nucleotide hydrolysis used in the nitrogenase reaction? What specifically is the role of the P-cluster in brokering electrons between the Fe protein

Mechanism of Mo-Dependent Nitrogenase

21

Fig. 2.5. Fe–protein cycle showing the oxidation state changes and MgATP hydrolysis. Abbreviations used are FePOx for oxidized Fe protein, FePRed for reduced Fe protein, MoFeP for the oxidation state of MoFe protein before reduction, and MoFePRed for one more electron reduced state of MoFe protein.

and the active site FeMo cofactor? These and many other questions need to be addressed in the coming years. 3.3. MoFe–Protein Cycle

The MoFe protein must accumulate multiple electrons in order to achieve the reduction of bound substrates. The details of where and how these electrons are accumulated in the MoFe protein are not known (20). A simple notation to designate how many electrons have been transferred into the resting MoFe protein (designated as E0 ) is helpful (called the Lowe–Thornely model) (64, 132–134). This model does not differentiate electrons on the P-cluster from electrons on the FeMo cofactor (Fig. 2.6), but rather simply notes the number of electrons accumulated in the MoFe protein as E1 , E2 , etc., with the subscript indicating the number of electrons. The results of a number of kinetic studies have allowed construction of a MoFe-protein cycle as shown in Fig. 2.6. As

Fig. 2.6. Modified Lowe–Thorneley kinetic scheme for reduction of N2 . In this scheme, the En represents one functional αβ dimeric unit, which has been reduced by n electrons relative to the resting state E0 .

22

Yang, Danyal, and Seefeldt

noted in the cycle, there is good evidence indicating that different substrates bind to different reduction states (En ) of the MoFe protein. Dinitrogen is modeled to bind to E3 or E4 states (64), which is accompanied by the release of one equivalent of H2 (22, 135). In the absence of N2 , the less reduced E1 and E2 states are achieved, which are sufficient for proton binding and reduction to H2 . The non-physiological substrate acetylene is modeled to bind to the E2 state for reduction to ethylene. The binding of different substrates to different redox states of the MoFe protein can result in confusing inhibition patterns. For example, the inhibition of N2 reduction by acetylene appears to be non-competitive, while the inhibition of acetylene reduction by N2 appears to be competitive (136). This apparent contradiction can be explained by the fact that acetylene binds to the E2 state whereas N2 binds to more reduced states (64, 137). Thus, acetylene appears to be a non-competitive inhibitor of N2 reduction and N2 a competitive inhibitor of acetylene reduction. Insights into where and to what states of FeMo-cofactor substrates and inhibitors bind have come in recent years from the characterization of freeze-trapped MoFe protein during turnover with substrates or inhibitors. The inhibitor carbon monoxide (CO) has been extensively characterized trapped to FeMo cofactor (100, 138–142). Application of a variety of spectroscopic methods to this trapped state has revealed that at low CO concentrations, a single CO is likely bound bridging between two Fe atoms. At high CO concentrations, two CO molecules are proposed to be bound. Amino acid substitutions in the MoFe protein have been used, along with freeze trapping to capture a number of different substrates bound as intermediates to FeMo cofactor. The MoFe protein variant, the substrate, and the g values of the EPR spectrum observed are summarized in Table 2.1. How these states are trapped is discussed in more detail in another chapter of this volume. Some of the key findings from characterization of these trapped states include the following: (1) A hydride-trapped state is consistent with two hydrides bound to Fe atoms in FeMo cofactor (106). (2) The substrate propargyl alcohol and other alkyne substrates have been trapped bound side-on to one or more Fe atoms (104, 105, 143). (3) Intermediates have been trapped starting from the nitrogenous substrates hydrazine, diazene, methyldiazene, and N2 . These intermediates appear to be bound end-on to Fe atom(s) (106–110). (4) A specific Fe atom in the central portion of FeMo cofactor has been identified as the likely site of binding of several substrates tested so far (110, 143).

Mechanism of Mo-Dependent Nitrogenase

23

Table 2.1 Important variants of MoFe protein and EPR parameters of the resulting intermediates with different substrates in the turnover state Mutation

Substrate

EPR parameter

Ref.

Wild type

Dinitrogen (N2 )

S = 1/2, g = 2.08, 1.99, 1.97

(107, 110)

α-70Val→Ala

Propargyl alcohol (HC≡CCH2 OH)

S = 1.2, g = 2.12, 2.00, 1.99

(104, 105)

α-70Val→Ile

Proton (H+ )

S = 1/2, g = 2.14, 2.00, 1.96

(103)

α-195His→Gln

Methyldiazene (CH3 N=NH)

S = 1.2, g = 2.08, 2.02, 1.99

(107, 108)

α-70Val→Ala /α-195His→Gln

Diazene (HN=NH)

S = 1/2, g = 2.09, 2.01, 1.93

(109)

α-70Val→Ala /α-195His→Gln

Hydrazine (NH2 –NH2 )

S = 1/2, g = 2.09, 2.01, 1.93

(106, 107)

These studies have advanced our understanding of where and how substrates interact with the nitrogenase active site. Many other questions remain to be resolved, such as (1) Is the N–N bond broken in the trapped nitrogenous species characterized so far? (2) What is the level of reduction of the trapped states? (3) What is the level of proton addition to the trapped states? (4) Do intermediates migrate among the metals (Fe and Mo) during the course of substrate reduction? Answers to these and many other open questions will greatly advance our understanding of this complex enzyme.

4. Conclusions and Perspectives Nitrogenase is a complex enzyme that plays a central role in the global N cycle. Great strides have been achieved since its first discovery to understand many facets of this complex system. These advances are largely the result of the application of a wide range of methods, many of which are described in this volume. Clearly, much remains to be resolved about the mechanism of nitrogenase. Future advances will come from the application of the methods summarized in this volume, coupled with application of many new methods. While the prospects for advancing understanding of nitrogenase going forward are good, these advances are not likely to come easily.

24

Yang, Danyal, and Seefeldt

Acknowledgments The authors acknowledge the long collaboration with the Brian Hoffman and Dennis Dean laboratories in advancing understanding of nitrogenase. Work in the laboratory of the authors is supported by a generous grant from the National Institutes of Health (GM59087). References 1. Smil V (2001) Enriching the Earth: Fritz Haber, Carl Bosch, and the Transformation of World Food Production. MIT Press, Cambridge, MA 2. Ferguson SJ (1998) Nitrogen cycle enzymology. Curr Opin Chem Biol 2:182–193 3. Fryzuk MD, MacKay BA (2004) Dinitrogen coordination chemistry: On the biomimetic borderlands. Chem Rev 104:385–401 4. Haber F (1922) The production of ammonia from nitrogen and hydrogen. Naturwissenschaften 10:1041–1049 5. Haber F (1923) The history of the ammonia process. Naturwissenschaften 11:339–340 6. Cheng Q (2008) Perspectives in biological nitrogen fixation research. J Integr Plant Biol 50:786–798 7. Smith BE (2002) Nitrogen reveals its inner secrets. Science 297:1654–1655 8. Raymond J, Siefert JL, Staples CR et al (2004) The natural history of nitrogen fixation. Mol Biol Evol 21:541–554 9. Burgess BK, Lowe DJ (1996) The mechanism of molybdenum nitrogenase. Chem Rev 96:2983–3011 10. Eady RR (1996) Structure-function relationships of alternative nitrogenases. Chem Rev 96:3013–3030 11. Bishop PE, Joerger RD (1990) Genetics and molecular biology of alternative nitrogen fixation systems. Annu Rev Plant Physiol Plant Mol Biol 41:109–125 12. Masepohl B, Schneider K, Drepper T et al (2002) Alternative nitrogenases. In: Leigh GJ (ed) Nitrogenase Fixation at the Millennium, pp. 191–222. Elsevier, Amsterdam 13. Barney BM, Lee HI, Dos Santos PC et al (2006) Breaking the N2 triple bond: Insights into the nitrogenase mechanism. Dalton Trans 19:2277–2284 14. Ribbe M, Gadkari D, Meyer O (1997) N2 fixation by Streptomyces thermoautotrophicus involves a molybdenum- dinitrogenase and a manganese-superoxide oxidoreductase that couple N2 reduction to the oxida-

15. 16. 17. 18.

19. 20.

21.

22.

23.

24.

25.

tion of superoxide produced from O2 by a molybdenum-CO dehydrogenase. J Biol Chem 272:26627–26633 Rees DC, Howard JB (2000) Nitrogenase: Standing at the crossroads. Curr Opin Chem Biol 4:559–566 Seefeldt LC, Dean DR (1997) Role of nucleotides in nitrogenase catalysis. Acc Chem Res 30:260–266 Seefeldt LC, Dance I, Dean DR (2004) Substrate interactions with nitrogenase: Fe versus Mo. Biochemistry 43:1401–1409 Dos Santos PC, Igarashi RY, Lee HI et al (2005) Substrate interactions with the nitrogenase active site. Acc Chem Res 38: 208–214 Seefeldt LC, Hoffman BM, Dean DR (2009) Mechanism of Mo-dependent nitrogenase. Annu Rev Biochem 78:701–722 Hoffman BM, Dean DR, Seefeldt LC (2009) Climbing nitrogenase: Toward a mechanism of enzymatic nitrogen fixation. Acc Chem Res 42:609–619 Igarashi RY, Seefeldt LC (2003) Nitrogen fixation: The mechanism of the Modependent nitrogenase. Crit Rev Biochem Mol Biol 38:351–384 Simpson FB, Burris RH (1984) A nitrogen pressure of 50 atmospheres does not prevent evolution of hydrogen by nitrogenase. Science 224:1095–1097 Kim J, Rees DC (1992) Crystallographic structure and functional implications of the nitrogenase molybdenum iron protein from Azotobacter vinelandii. Nature 360: 553–560 Kim J, Rees DC (1992) Structural models for the metal centers in the nitrogenase molybdenum-iron protein. Science 257:1677–1682 Georgiadis MM, Komiya H, Chakrabarti P et al (1992) Crystallographic structure of the nitrogenase iron protein from Azotobacter vinelandii. Science 257: 1653–1659

Mechanism of Mo-Dependent Nitrogenase 26. Kim J, Woo D, Rees DC (1993) X-ray crystal structure of the nitrogenase molybdenumiron protein from Clostridium pasteurianum at 3.0-Å resolution. Biochemistry 32: 7104–7115 27. Chan MK, Kim J, Rees DC (1993) The nitrogenase FeMo-cofactor and P-cluster pair: 2.2 Å resolution structures. Science 260:792–794 28. Einsle O, Tezcan FA, Andrade SLA et al (2002) Nitrogenase MoFe-protein at 1.16 Å resolution: A central ligand in the FeMocofactor. Science 297:1696–1700 29. Sørlie M, Christiansen J, Lemon BJ et al (2001) Mechanistic features and structure of the nitrogenase α-Gln195 MoFe protein. Biochemistry 40:1540–1549 30. Jang SB, Seefeldt LC, Peters JW (2000) Insights into nucleotide signal transduction in nitrogenase: Structure of an iron protein with MgADP bound. Biochemistry 39:14745–14752 31. Jang SB, Jeong MS, Seefeldt LC et al (2004) Structural and biochemical implications of single amino acid substitutions in the nucleotide-dependent switch regions of the nitrogenase Fe protein from Azotobacter vinelandii. J Biol Inorg Chem 9:1028–1033 32. Jang SB, Seefeldt LC, Peters JW (2000) Modulating the midpoint potential of the [4Fe-4S] cluster of the nitrogenase Fe protein. Biochemistry 39:641–648 33. Mayer SM, Lawson DM, Gormal CA et al (1999) New insights into structure-function relationships in nitrogenase: A 1.6 Å resolution X-ray crystallographic study of Klebsiella pneumoniae MoFe-protein. J Mol Biol 292:871–891 34. Mayer SM, Gormal CA, Smith BE et al (2002) Crystallographic analysis of the MoFe protein of nitrogenase from a nifV mutant of Klebsiella pneumoniae identifies citrate as a ligand to the molybdenum of iron molybdenum cofactor (FeMoco). J Biol Chem 277:35263–35266 35. Bolin JT, Ronco AE, Mortenson LE et al (1990) Structure of the nitrogenase MoFe protein: Spatial distribution of the intrinsic metal atoms determined by X-ray anomalous scattering. In: Gresshoff PM, Roth LE, Stacey G, Newton WE (eds) Nitrogen Fixation: Achievements and Objectives, pp. 117–124. Chapman and Hall, New York, NY 36. Bolin JT, Campobasso N, Muchmore SW et al (1993) Structure and environment of metal clusters in the nitrogenase molybdenum-iron protein form Clostridium pasteurianum. In: Stiefel EI, Coucouvanis D,

37.

38.

39. 40.

41.

42.

43.

44.

45.

46.

47.

48. 49.

25

Newton WE (eds) Molybdenum Enzymes, Cofactors and Model Systems, pp. 186–195. ACS, Washington, DC Sen S, Krishnakumar A, McClead J et al (2006) Insights into the role of nucleotidedependent conformational change in nitrogenase catalysis: Structural characterization of the nitrogenase Fe protein Leu127 deletion variant with bound MgATP. J Inorg Biochem 100:1041–1052 Sen S, Igarashi R, Smith A et al (2004) A conformational mimic of the MgATP-bound “on state” of the nitrogenase iron protein. Biochemistry 43:1787–1797 Schmid B, Ribbe MW, Einsle O et al (2002) Structure of a cofactor-deficient nitrogenase MoFe protein. Science 296:352–356 Sarma R, Barney BM, Keable S et al (2009) Insights into substrate binding at FeMocofactor in nitrogenase from the structure of an α-70Ile MoFe protein variant. J Inorg Biochem 104:385–389 Jeong MS, Jang SB (2004) Structural basis for the changes in redox potential in the nitrogenase Phe135Trp Fe protein with MgADP bound. Mol Cells 18:374–382 Peters JW, Stowell MHB, Soltis SM et al (1997) Redox-dependent structural changes in the nitrogenase P-cluster. Biochemistry 36:1181–1187 Schindelin H, Kisker C, Schlessman JL et al (1997) Structure of ADP-AlF4 − stabilized nitrogenase complex and its implications for signal transduction. Nature 387:370–376 Chiu HJ, Peters JW, Lanzilotta WN et al (2001) MgATP-bound and nucleotide-free structures of a nitrogenase protein complex between the Leu 127@-Fe-protein and the MoFe-protein. Biochemistry 40: 641–650 Schmid B, Einsle O, Chiu HJ et al (2002) Biochemical and structural characterization of the cross-linked complex of nitrogenase: Comparison to the ADP-AlF4 – -stabilized structure. Biochemistry 41:15557–15565 Tezcan FA, Kaiser JT, Mustafi D et al (2005) Nitrogenase complexes: Multiple docking sites for a nucleotide switch protein. Science 309:1377–1380 Rees DC, Tezcan FA, Haynes CA et al (2005) Structural basis of biological nitrogen fixation. Philos Trans R Soc A 363: 971–984 Howard JB, Rees DC (1994) Nitrogenase: A nucleotide-dependent molecular switch. Annu Rev Biochem 63:235–264 Rubio LM, Ludden PW (2008) Biosynthesis of the Iron-Molybdenum cofactor of nitrogenase. Annu Rev Microbiol 62:93–111

26

Yang, Danyal, and Seefeldt

50. Hu Y, Fay AW, Lee CC et al (2008) Assembly of nitrogenase MoFe protein. Biochemistry 47:3973–3981 51. Martin AE, Burgess BK, Iismaa SE et al (1989) Construction and characterization of an Azotobacter vinelandii strain with mutations in the genes encoding flavodoxin and ferredoxin I. J Bacteriol 171: 3162–3167 52. Mortenson LE (1964) Ferredoxin requirement for nitrogen fixation by extracts of Clostridium pasteurianum. Biochim Biophys Acta 81:473–478 53. Angove HC, Yoo SJ, Burgess BK et al (1997) Mössbauer and EPR evidence for an allferrous Fe4 S4 cluster with S = 4 in the Fe protein of nitrogenase. J Am Chem Soc 119:8730–8731 54. Watt GD, Reddy KRN (1994) Formation of an all ferrous Fe4 S4 cluster in the iron protein component of Azotobacter vinelandii nitrogenase. J Inorg Biochem 53:281–294 55. Musgrave KB, Angove HC, Burgess BK et al (1998) All-ferrous titanium(III) citrate reduced Fe protein of nitrogenase: An XAS study of electronic and metrical structure. J Am Chem Soc 120:5325–5326 56. Strop P, Takahara PM, Chiu HJ et al (2001) Crystal structure of the all-ferrous [4Fe-4S]0 form of the nitrogenase iron protein from Azotobacter vinelandii. Biochemistry 40:651–656 57. Yoo SJ, Angove HC, Burgess BK et al (1998) Magnetic circular dichroism study of the allferrous [4Fe-4S] cluster of the Fe-protein of Azotobacter vinelandii nitrogenase. J Am Chem Soc 120:9704–9705 58. Nyborg AC, Johnson JL, Gunn A et al (2000) Evidence for a two-electron transfer using the all-ferrous Fe protein during nitrogenase catalysis. J Biol Chem 275: 39307–39312 59. Hagen WR, Dunham WR, Braaksma A et al (1985) On the prosthetic group(s) of component II from nitrogenase: EPR of the Fe-protein from Azotobacter vinelandii. FEBS Lett 187:146–150 60. Hagen WR, Eady RR, Dunham WR et al (1985) A novel S = 3/2 EPR signal associated with native Fe proteins of nitrogenase. FEBS Lett 189:250–254 61. Lindahl PA, Day EP, Kent TA et al (1985) Mössbauer, EPR, and magnetization studies of the Azotobacter vinelandii Fe protein. J Biol Chem 260:11160–11173 62. Thorneley RNF, Ashby GA (1989) Oxidation of nitrogenase iron protein by dioxygen without inactivation could contribute to high respiration rates of Azotobacter species

63.

64.

65

66.

67.

68.

69.

70.

71.

72.

73.

74.

and facilitate nitrogen fixation in other aerobic environments. Biochem J 261:181–187 Larsen C, Christensen S, Watt GD (1995) Reductant-independent ATP hydrolysis catalyzed by homologous nitrogenase proteins from Azotobacter vinelandii and heterologous crosses with Clostridium pasteurianum. Arch Biochem Biophys 323:215–222 Thorneley RNF, Lowe DJ (1985) Kinetics and mechanisms of the nitrogenase enzyme system. In: Spiro TG (ed) Molybdenum Enzymes, pp. 221–284. Wiley, New York, NY Ryle MJ, Lanzilotta WN, Mortenson LE et al (1995) Evidence for a central role of lysine 15 of Azotobacter vinelandii nitrogenase iron protein in nucleotide binding and protein conformational changes. J Biol Chem 270:13112–13117 Watt GD (1979) An electrochemical method for measuring redox potentials of low potential proteins by microcoulometry at controlled potentials. Anal Biochem 99:399–407 Watt GD, Wang ZC, Knotts RR (1986) Redox reactions of and nucleotide binding to the iron protein of Azotobacter vinelandii. Biochemistry 25:8156–8162 Lanzilotta WN, Ryle MJ, Seefeldt LC (1995) Nucleotide hydrolysis and protein conformational changes in Azotobacter vinelandii nitrogenase iron protein: Defining the function of aspartate 129. Biochemistry 34:10713–10723 Peters JW, Szilagyi RK (2006) Exploring new frontiers of nitrogenase structure and mechanism. Curr Opin Chem Biol 10: 101–108 Yates MG (1991) The enzymology of molybdenum-dependent nitrogen fixation. In: Stacey G, Burris RH, Evans HJ (eds) Biological Nitrogen Fixation, pp. 685–735. Chapman and Hall, New York, NY Lanzilotta WN, Parker VD, Seefeldt LC (1999) Thermodynamics of nucleotide interactions with the Azotobacter vinelandii nitrogenase iron protein. Biochim Biophys Acta 1429:411–421 Cordewener J, Haaker H, Van Ewijk P et al (1985) Properties of the MgATP and MgADP binding sites on the Fe protein of nitrogenase from Azotobacter vinelandii. Eur J Biochem 148:499–508 Weston MF, Kotake S, Davis LC (1983) Interaction of nitrogenase with nucleotide analogs of ATP and ADP and the effect of metal ions on ADP inhibition. Arch Biochem Biophys 225:809–817 Ryle MJ, Seefeldt LC (2000) Hydrolysis of nucleoside triphosphates other than

Mechanism of Mo-Dependent Nitrogenase

75.

76. 77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

ATP by nitrogenase. J Biol Chem 275: 6214–6219 Sarma R, Mulder DW, Brecht E et al (2007) Probing the MgATP-bound conformation of the nitrogenase Fe protein by solution small-angle X-ray scattering. Biochemistry 46:14058–14066 Shah VK, Brill WJ (1977) Isolation of an iron-molybdenum cofactor from nitrogenase. Proc Natl Acad Sci USA 74:3249–3253 Zimmermann R, Münck E, Brill WJ et al (1978) Nitrogenase X: Mössbauer and EPR studies on reversibly oxidized MoFe protein from Azotobacter vinelandii OP: Nature of the iron centers. Biochim Biophys Acta 537:185–207 Lindahl PA, Papaefthymiou V, OrmeJohnson WH et al (1988) Mössbauer studies of solid thionin-oxidized MoFe protein of nitrogenase. J Biol Chem 263: 19412–19418 Hagen WR, Wassink H, Eady RR et al (1987) Quantitative EPR of an S = 7/2 system in thionine-oxidized MoFe proteins of nitrogenase: A redefinition of the P-cluster concept. Eur J Biochem 169:457–465 Pierik AJ, Wassink H, Haaker H et al (1993) Redox properties and EPR spectroscopy of the P-clusters of Azotobacter vinelandii MoFe protein. Eur J Biochem 212:51–61 Tittsworth RC, Hales BJ (1993) Detection of EPR signals assigned to the 1-equivoxidized P-clusters of the nitrogenase MoFe protein from Azotobacter vinelandii. J Am Chem Soc 115:9763–9767 Lanzilotta WN, Fisher K, Seefeldt LC (1997) Evidence for electron transfer-dependent formation of a nitrogenase iron proteinmolybdenum-iron protein tight complex: The role of aspartate 39. J Biol Chem 272:4157–4165 Morgan TV, Mortenson LE, McDonald JW et al (1988) Comparison of redox and EPR properties of the molybdenum iron proteins of Clostridium pasteurianum and Azotobacter vinelandii nitrogenases. J Inorg Biochem 33:111–120 Lanzilotta WN, Christiansen J, Dean DR et al (1998) Evidence for coupled electron and proton transfer in the [8Fe-7S] cluster of nitrogenase. Biochemistry 37: 11376–11384 Hoover TR, Robertson AD, Cerny RL et al (1987) Identification of the V factor needed for the synthesis of the ironmolybdenum cofactor of nitrogenase as homocitrate. Nature 329:855–857 Lee HI, Benton PM, Laryukhin M et al (2003) The interstitial atom of the nitroge-

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

27

nase FeMo-cofactor: ENDOR and ESEEM show it is not an exchangeable nitrogen. J Am Chem Soc 125:5604–5605 Yang TC, Maeser NK, Laryukhin M et al (2005) The interstitial atom of the nitrogenase FeMo-cofactor: ENDOR and ESEEM evidence that it is not a nitrogen. J Am Chem Soc 127:12804–12805 Lukoyanov D, Pelmenschikov V, Maeser N et al (2007) Testing if the interstitial atom, X, of the nitrogenase molybdenum-iron cofactor is N or C: ENDOR, ESEEM, and DFT studies of the S = 3/2 resting state in multiple environments. Inorg Chem 46: 11437–11449 George SJ, Igarashi RY, Xiao Y et al (2008) Extended X-ray absorption fine structure and nuclear resonance vibrational spectroscopy reveal that NifB-co, a FeMo-co precursor, comprises a 6Fe core with an interstitial light atom. J Am Chem Soc 130: 5673–5680 Lovell T, Li J, Liu T et al (2001) FeMo cofactor of nitrogenase: A density functional study of states MN , MOX , MR , and MI . J Am Chem Soc 123:12392–12410 Pelmenshikov V, Case DA, Noodleman L (2008) Ligand-bound S = 1/2 FeMocofactor of nitrogenase: Hyperfine interaction analysis and implication for the central ligand X identity. Inorg Chem 47: 6162–6172 Dance I (2003) The consequences of an interstitial N atom in the FeMo cofactor of nitrogenase. Chem Commun 3: 324–325 Dance I (2007) The mechanistically significant coordination chemistry of dinitrogen at FeMo-co, the catalytic site of nitrogenase. J Am Chem Soc 129:1076–1088 Orme-Johnson WH, Hamilton WD, Jones TL et al (1972) Electron paramagnetic resonance of nitrogenase and nitrogenase components from Clostridium pasteurianum W5 and Azotobacter vinelandii OP. Proc Natl Acad Sci USA 69:3142–3145 Schultz FA, Gheller SF, Newton WE (1988) Iron molybdenum cofactor of nitrogenase: Electrochemical determination of the electron stoichiometry of the oxidized/semireduced couple. Biochem Biophys Res Commun 152:629–635 Huynh BH, Henzl MT, Christner JA et al (1980) Nitrogenase XII: Mössbauer studies of the MoFe protein from Clostridium pasteurianum W5. Biochim Biophys Acta 623:124–138 Yoo SJ, Angove HC, Papaefthymiou V et al (2000) Mössbauer study of the MoFe protein

28

98.

99.

100.

101. 102.

103.

104.

105.

106.

107.

108.

109.

Yang, Danyal, and Seefeldt of nitrogenase from Azotobacter vinelandii using selective 57 Fe enrichment of the M-centers. J Am Chem Soc 122:4926–4936 Watt GD, Burns A, Lough S et al (1980) Redox and spectroscopic properties of oxidized MoFe protein from Azotobacter vinelandii. Biochemistry 19:4926–4932 Lovell T, Liu T, Case DA et al (2003) Structural, spectroscopic, and redox consequences of a central ligand in the FeMoco of nitrogenase: A density functional theoretical study. J Am Chem Soc 125:8377–8383 Lee HI, Hales BJ, Hoffman BM (1997) Metal-ion valencies of the FeMo cofactor in CO-inhibited and resting state nitrogenase by 57 Fe Q-band ENDOR. J Am Chem Soc 119:11395–11400 Burgess BK (1990) The iron-molybdenum cofactor of nitrogenase. Chem Rev 90: 1377–1406 Smith BE, Durrant MC, Fairhurst SA et al (1999) Exploring the reactivity of the isolated iron-molybdenum cofactor of nitrogenase. Coord Chem Rev 185:669–687 Igarashi RY, Laryukhin M, Dos Santos PC et al (2005) Trapping H- bound to the nitrogenase FeMo-cofactor active site during H2 evolution: Characterization by ENDOR spectroscopy. J Am Chem Soc 127: 6231–6241 Benton PMC, Laryukhin M, Mayer SM et al (2003) Localization of a substrate binding site on FeMo-cofactor in nitrogenase: Trapping propargyl alcohol with an α70-substituted MoFe protein. Biochemistry 42:9102–9109 Lee HI, Igarashi RY, Laryukhin M et al (2004) An organometallic intermediate during alkyne reduction by nitrogenase. J Am Chem Soc 126:9563–9569 Barney BM, Laryukhin M, Igarashi RY et al (2005) Trapping a hydrazine reduction intermediate on the nitrogenase active site. Biochemistry 44:8030–8037 Barney BM, Yang TC, Igarashi RY et al (2005) Intermediates trapped during nitrogenase reduction of N2 , CH3 -N=NH, and H2 N-NH2 . J Am Chem Soc 127:14960– 14961 Barney BM, Lukoyanov D, Yang TC et al (2006) A methyldiazene (HN=NCH3 ) derived species bound to the nitrogenase active site FeMo-cofactor: Implications for mechanism. Proc Natl Acad Sci USA 103:17113–17118 Barney BM, McClead J, Lukoyanov D et al (2007) Diazene (HN=NH) is a substrate for nitrogenase: Insights into the pathway of N2 reduction. Biochemistry 46:6784–6794

110. Barney BM, Lukoyanov D, Igarashi RY et al (2009) Trapping an intermediate of dinitrogen (N2 ) reduction on nitrogenase. Biochemistry 48:9094–9102 111. Kim CH, Newton WE, Dean DR (1995) Role of the MoFe protein α-subunit histidine-195 residue in FeMo-cofactor binding and nitrogenase catalysis. Biochemistry 34:2798–2808 112. Thomann H, Bernardo M, Newton WE et al (1991) N coordination of MoFe cofactor requires His-195 of the MoFe protein α-subunit and is essential for biological nitrogen fixation. Proc Natl Acad Sci USA 88:6620–6623 113. Scott DJ, May HD, Newton WE et al (1990) Role for the nitrogenase MoFe protein α-subunit in FeMo-cofactor binding and catalysis. Nature 343:188–190 114. Dilworth MJ, Fisher K, Kim CH et al (1998) Effects on substrate reduction of substitution of histidine-195 by glutamine in the alpha-subunit of the MoFe protein of Azotobacter vinelandii nitrogenase. Biochemistry 37:17495–17505 115. Fisher K, Dilworth MJ, Newton WE (2000) Differential effects on N2 binding and reduction, HD formation, and azide reduction with α-195His - and α-191Gln substituted MoFe proteins of Azotobacter vinelandii nitrogenase. Biochemistry 39: 15570–15577 116. Burgess BK (1985) Substrate reactions of nitrogenase. In: Spiro TG (ed) Metal Ions in Biology: Molybdenum Enzymes, pp. 161– 219. Wiley, New York, NY 117. Willing AH, Georgiadis MM, Rees DC et al (1989) Cross-linking of nitrogenase components. J Biol Chem 264:8499–8503 118. Willing A, Howard JB (1990) Crosslinking site in Azotobacter vinelandii complex. J Biol Chem 265:6596–6599 119. Emerich DW, Burris RH (1976) Interactions of heterologous nitrogenase components that generate catalytically inactive complexes. Proc Natl Acad Sci USA 73: 4369–4373 120. Emerich DW, Burris RH (1978) Complementary functioning of the component proteins of nitrogenase from several bacteria. J Bacteriol 134:936–943 121. Emerich DW, Ljones T, Burris RH (1978) Nitrogenase: Properties of the catalytically inactive complex between the Azotobacter vinelandii MoFe protein and the Clostridium pasteurianum Fe protein. Biochim Biophys Acta 527:359–369 122. Chan JM, Ryle MJ, Seefeldt LC (1999) Evidence that MgATP accelerates primary

Mechanism of Mo-Dependent Nitrogenase

123.

124.

125.

126.

127.

128.

129.

130.

131.

132.

electron transfer in a Clostridium pasteurianum Fe protein-Azotobacter vinelandii MoFe protein nitrogenase tight complex. J Biol Chem 274:17593–17598 Lanzilotta WN, Seefeldt LC (1997) Changes in the midpoint potentials of the nitrogenase metal centers as a result of iron protein-molybdenum-iron protein complex formation. Biochemistry 36: 12976–12983 Clarke TA, Yousafzai FK, Eady RR (1999) Klebsiella pneumoniae nitrogenase: Formation and stability of putative beryllium fluoride-ADP transition state complexes. Biochemistry 38:9906–9913 Duyvis MG, Wassink H, Haaker H (1996) Formation and characterization of a transition state complex of Azotobacter vinelandii nitrogenase. FEBS Lett 380: 233–236 Renner KA, Howard JB (1996) Aluminum fluoride inhibition of nitrogenase: Stabilization of a nucleotide-Fe-proteinMoFe protein complex. Biochemistry 35: 5353–5358 Miller RW, Eady RR, Fairhurst SA et al (2001) Transition state complexes of the Klebsiella pneumoniae nitrogenase proteins: Spectroscopic properties of aluminum fluoride-stabilized and beryllium fluoridestabilized MgADP complexes reveal conformational differences of the Fe protein. Eur J Biochem 268:809–818 Spee JH, Arendsen AF, Wassink H et al (1998) Redox properties and electron paramagnetic resonance spectroscopy of the transition state complex of Azotobacter vinelandii nitrogenase. FEBS Lett 432: 55–58 Duyvais MG, Wassink H, Haaker H (1998) Nitrogenase of Azotobacter vinelandii: Kinetic analysis of the Fe protein redox cycle. Biochemistry 37:17345–17354 Christiansen J, Dean DR, Seefeldt LC (2001) Mechanistic features of the Mo-containing nitrogenase. Annu Rev Plant Physiol Plant Mol Biol 52:269–295 Hageman RV, Burris RH (1978) Nitrogenase and nitrogenase reductase associate and dissociate with each catalytic cycle. Proc Natl Acad Sci USA 75:2699–2702 Lowe DJ, Thorneley RNF (1984) The mechanism of Klebsiella pneumoniae nitrogenase action: The determination of rate constants required for the simulation of kinetics of

133.

134.

135.

136. 137.

138.

139.

140.

141. 142.

143.

29

N2 reduction and H2 evolution. Biochem J 224:895–901 Thorneley RNF, Lowe DJ (1984) The mechanism of Klebsiella pneumoniae nitrogenase action: Pre-steady-state kinetics of an enzyme-bound intermediate in N2 reduction and of NH3 formation. Biochem J 224: 887–894 Thorneley RNF, Lowe DJ (1984) The mechanism of Klebsiella pneumoniae nitrogenase action: Stimulation of the dependences of H2 -evolution rate on component-protein concentration and ratio and sodium dithionite concentration. Biochem J 224:903–909 Liang J, Burris RH (1988) Hydrogen burst associated with nitrogenase-catalyzed reactions. Proc Natl Acad Sci USA 85: 9446–9450 Rivera-Ortiz JM, Burris RH (1975) Interactions among substrates and inhibitors of nitrogenase. J Bacteriol 123:537–545 Davis LC, Shah VK, Brill WJ (1975) Nitrogenase: Effect of component ratio, ATP and H2 on the distribution of electrons to alternative substrates. Biochim Biophys Acta 403:67–78 Lee HI, Cameron LM, Hales BJ et al (1997) CO binding to the FeMo cofactor of COinhibited nitrogenase: 13 CO and 1 H Qband ENDOR investigation. J Am Chem Soc 119:10121–10126 Pollock CR, Lee HI, Cameron LM et al (1995) Investigation of CO bound to inhibited forms of nitrogenase MoFe protein by 13 C ENDOR. J Am Chem Soc 117: 8686–8687 Christie PD, Lee HI, Cameron LM et al (1996) Identification of the CO-binding cluster in nitrogenase MoFe protein by ENDOR of 57 Fe isotopomers. J Am Chem Soc 118:8707–8709 Maskos Z, Hales BJ (2003) Photo-lability of CO bound to Mo-nitrogenase from Azotobacter vinelandii. J Inorg Biochem 93:11–17 Maskos Z, Fisher K, Sørlie M et al (2005) Variant MoFe proteins of Azotobacter vinelandii: Effects of carbon monoxide on electron paramagnetic resonance spectra generated during enzyme turnover. J Biol Inorg Chem 10:394–406 Dos Santos PC, Mayer SM, Barney BM et al (2007) Alkyne substrate interaction within the nitrogenase MoFe protein. J Inorg Biochem 101:1642–1648

Chapter 3 Assembly of Nitrogenase MoFe Protein Chi-Chung Lee, Aaron W. Fay, Jared A. Wiig, Markus W. Ribbe, and Yilin Hu Abstract Biosynthesis of MoFe protein and, particularly, that of its associated P-cluster and FeMoco has raised a significant amount of interest because of the biological importance and chemical exclusiveness of these unique clusters. Following a brief introduction to the properties of Azotobacter vinelandii MoFe protein, this chapter will focus on the recent progress toward understanding the assembly mechanism of MoFe protein, with an emphasis on studies that provide important structural or spectroscopic insights into this process. Key words: Nitrogenase, MoFe protein, NifEN, FeMoco, P-cluster, assembly.

1. Introduction Nitrogenase is responsible for biological nitrogen fixation, an ATP-dependent process in which the atmospheric dinitrogen (N2 ) is converted to the bioavailable ammonia (NH3 ) (1–7). The ability of nitrogenase to break the triple bond of N2 —an extremely difficult chemical reaction—under ambient conditions makes it one of the most unique enzymes in nature. Through decades of research, it has been established that nitrogenase owes much of its amazing catalytic ability to its associated metalloclusters. The best-studied molybdenum (Mo) nitrogenase of Azotobacter vinelandii consists of two proteins, the iron (Fe) protein and the molybdenum–iron (MoFe) protein, each of which contains redox-active metal center(s). The α2 -homodimeric Fe protein is bridged by a single [Fe4 S4 ] cluster between the subunits and contains one ATP-binding site per subunit, whereas the M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_3, © Springer Science+Business Media, LLC 2011

31

32

Lee et al.

α2 β2 -heterotetrameric MoFe protein contains two unique metal clusters per αβ-dimer: the P-cluster ([Fe8 S7 ]), which is located at the α/β-subunit interface, and FeMoco ([MoFe7 S9 Xhomocitrate], X is considered to be N, C or O), which is buried within the α-subunit (8, 9). During substrate turnover, the Fe protein forms a transient complex with the MoFe protein and transfers electrons from its [Fe4 S4 ] cluster to the P-cluster of MoFe protein in an ATP-dependent manner. The electrons are subsequently transferred to FeMoco, where substrates are eventually reduced after the accumulation of certain amounts of electrons. Biosynthesis of MoFe protein and, particularly, that of its associated P-cluster and FeMoco has raised a significant amount of interest because of the biological importance and chemical exclusiveness of these unique clusters. Following a brief introduction about the properties of A. vinelandii MoFe protein, this chapter will focus on the recent progress toward understanding the assembly mechanism of MoFe protein, with an emphasis on studies that provide important structural or spectroscopic insights into this process.

2. Properties of MoFe Protein and Its Associated Clusters

The MoFe protein of A. vinelandii is an α2 β2 -tetramer of ∼230 kDa, and its α-subunit (∼55 kDa) and β-subunit (∼59 kDa) are encoded by the nifD and nifK genes, respectively. The P-cluster is found at the interface between the α- and β-subunits, 10 Å below the surface of the protein (8, 9). It contains eight Fe and seven S atoms, and it is normally covalently coordinated to the MoFe protein by six cysteinyl sulfur ligands: Cysα62 , Cysα88 , and Cysα154 from the α-subunit; and Cysβ70 , Cysβ95 , and Cysβ153 from the β-subunit (9–11). The P-cluster can assume two oxidation states: the dithionite-reduced PN state (12, 13) and the indigo disulfonate (IDS)-oxidized POX state (14). While the PN state is diamagnetic and believed to be all-ferrous, the POX state, which is two-electron-oxidized from PN , is paramagnetic and displays a characteristic g = 11.8 signal in the parallel mode EPR (13, 15) that is commonly used to identify a fully intact P-cluster species in assembly related studies (see Section 3.4 for detailed discussion). The PN and POX are not only different in electronic state but also different in structure. Upon oxidation, two Fe atoms lose their coordination to the center μ-sulfur atom and, in exchange, coordinate to the Oγ atom of Serβ188 and a backbone amide of Cysα88 , respectively (11). Thus, the P-cluster is structurally more “open” in the POX state than it is in the PN

Assembly of Nitrogenase MoFe Protein

33

state. It has been proposed that the P-cluster repeatedly undergoes the conversion between the “closed” PN state to the more “open” POX state during the catalytic cycle; however, the physiological relevance of the two oxidation states of the P-cluster has remained unclear at this point. FeMoco is located ∼14 Å away from the P-cluster and is buried in a cavity ∼10 Å below the surface of the MoFe protein α-subunit (see Section 3.2 for detailed discussion). It is attached to the protein by only two residues (Hisα442 and Cysα275 ) and, therefore, can be extracted as an intact entity into N-methylformamide; conversely, the isolated FeMoco can be readily inserted into the FeMoco-deficient MoFe protein, leading to the formation of a fully reconstituted MoFe protein (2, 16, 17). FeMoco has an overall stoichiometry of 1Mo:7Fe:9S:1homocitrate and can be viewed as an [Fe4 S3 ] and an [MoFe3 S3 ] cluster bridged by three inorganic sulfide atoms (8, 9, 17). In addition, a recent high-resolution (1.16 Å) x-ray structural analysis of A. vinelandii MoFe protein has revealed the presence of an extra, μ6 -interstitial, light atom (X) at the center of the Fe–S cage of FeMoco, although the identity of X is still a topic of ongoing research (8). The isolated FeMoco has been determined to be anionic (17) and its overall negative charge is attributed to homocitrate, an organic bidentate ligand of Mo that is −4 if the OH group is deprotonated. Both the isolated and the protein-bound forms of FeMoco have a well-characterized EPR signal, the spin state of which is determined to be S = 3/2 (1, 17).

3. Assembly of MoFe Protein The biosynthesis of MoFe protein is a highly complex process that involves the products of a number of genes, such as nifS, nifU, nifB, nifE, nifN, nifV, nifQ, nifZ, and nifH, within the nitrogen fixation (nif) gene cluster (18). Early genetic studies have provided valuable information regarding the factors required for the assembly of MoFe protein; however, the exact function of these protein factors in the assembly process have remained unclear. Fortunately, with the help of advanced techniques in molecular biology and biochemistry, several intermediates that are crucial for the assembly of MoFe protein have been obtained. Detailed characterization of these “snapshots” of the assembly process has led to the proposal of a much refined assembly mechanism of MoFe protein that entails (i) the “ex situ” assembly of FeMoco on NifS, NifU, NifB, and NifEN; (ii) the incorporation of FeMoco into MoFe protein; (iii) the “in situ” assembly of the P-cluster on

34

Lee et al.

Fig. 3.1. Assembly model of MoFe protein entailing the “ex situ” assembly of FeMoco (①), the insertion of FeMoco into the MoFe protein (②), and the stepwise “in situ” assembly of the two P-clusters (➂ and ➃).

MoFe protein; and (iv) the stepwise assembly of MoFe protein (Fig. 3.1). These four aspects of the MoFe protein assembly will be discussed in greater detail below. 3.1. The “Ex Situ” Assembly of FeMoco on NifS, NifU, NifB, and NifEN

The assembly of FeMoco occurs independently of the synthesis of the polypeptides of MoFe protein (hence the term “ex situ” assembly) (18). Based on genetic evidence, FeMoco assembly is likely initiated by the action of the NifUS protein complex (nifU and nifS gene products), during which process NifS acts as a pyridoxal phosphate-dependent cysteine desulfurase, which forms a protein-bound cysteine persulfide that is subsequently donated to NifU for the sequential formation of [2Fe–2S] and [4Fe–4S] clusters (19–23). These Fe–S fragments are subsequently transferred to NifB (nifB gene product), where

Assembly of Nitrogenase MoFe Protein

35

a FeMoco “core”—which likely contains all Fe and S necessary for the formation of FeMoco—is formed. The exact function of NifB in this process is unclear. Nevertheless, NifB is indispensable for FeMoco biosynthesis, as deletion of nifB leads to the formation of a cofactor-deficient form of MoFe protein. Sequence analysis shows that NifB contains a CXXXCXXC signature motif at the N-terminus, which is characteristic of a family of radical Sadenosyl-L-methionine (SAM)-dependent enzymes. Apart from this feature, there is an abundance of potential ligands in the NifB sequence for the coordination of the entire complement of Fe atoms of FeMoco. Therefore, generation of the Fe–S core on NifB could represent a novel synthetic route to bridged metal clusters that relies on radical chemistry at the SAM domain of NifB (24–27). It is not clear whether NifB assembles the FeMoco core alone or through its interactions with NifEN (nifE and nifN gene products); it is generally accepted that the synthesized FeMoco core is transferred to NifEN, the last “checkpoint” in FeMoco biosynthesis before its delivery to MoFe protein. The initial proposal that NifEN acts as a scaffold protein for the maturation of FeMoco “core” was based on the significant degree of sequence homology between NifEN and MoFe protein (nifD and nifK gene products). This observation has led to the hypothesis that NifEN contains cluster-binding regions that are analogous to the P-cluster and FeMoco sites in the MoFe protein (18, 28–30). While the P-cluster analog in NifEN was determined earlier to be an [Fe4 S4 ]-type metal cluster (29), the FeMoco analog in the protein has remained unidentified until the recent characterization of a NifEN complex isolated from a nifHDK-deficient background. The absence of the nifDKencoded MoFe protein (the terminal acceptor of FeMoco) and the nifH-encoded Fe protein [an essential factor for FeMoco maturation (see below for detailed discussion)] results in the accumulation of a precursor form of the FeMoco on NifEN. Biochemical analysis (31) reveals that the NifEN-bound precursor contains no Mo (and homocitrate) and exhibits a unique g = 1.92 EPR signal in the IDS-oxidized state (Fig. 3.2a, 1). More excitingly, Fe K-edge XAS/EXAFS analysis shows that the precursor is a Mo-free analog of FeMoco and not one of the more commonly suggested [Fe4 S4 ]-type metal clusters (32). Both the Fe8 model (Fig. 3.2b, 1) and the Fe7 model of the precursor resemble FeMoco with slightly elongated interatomic distances (32). This finding suggests that, instead of coupling a [Fe4 S3 ] and a [MoFe3 S3 ] subcluster, FeMoco is assembled by the formation of the Fe–S core prior to the insertion of the Mo atom. Furthermore, the identification of an Fe-only FeMoco precursor on NifEN suggests NifEN as the site for the incorporation of Mo and homocitrate into the FeMoco. Indeed, NifEN re-isolated after incubation with MoO2− 4 , homocitrate, Fe protein, MgATP,

36

Lee et al.

1.92

A

B 1

1

2

2

1000

2000

3000

4000

Magnetic Field (Gauss)

Fig. 3.2. EPR spectra (a) and EXAFS models (b) of FeMoco precursor (1) and FeMoco (2) of NifEN.

and dithionite (designated NifENFeMoco ) no longer exhibits the precursor-specific g = 1.92 feature in the IDS-oxidized state (Fig. 3.2a, 2) (33). Metal analysis confirms the presence of Mo in NifENFeMoco . Furthermore, NifENFeMoco is capable of activating the FeMoco-deficient nifB MoFe protein, suggesting the presence of a fully competent FeMoco in NifENFeMoco that can be directly transferred to the MoFe protein without other carrier proteins (33). Fe and Mo K-edge XAS/EXAFS studies of NifENFeMoco (prepared in 2 mM dithionite) show that the FeMoco on NifENFeMoco (Fig. 3.2b, 2) closely resembles the native cofactor in the MoFe protein; yet, there is a significant amount of disorder in the Mo environment of the cluster on NifENFeMoco (33). Recently, the preparation of NifENFeMoco has been improved by increasing the concentration of dithionite from 2 to 20 mM (34). EPR and Mo K-edge XAS/EXAFS analyses reveal that thusprepared NifENFeMoco (designated NifENFeMoco(20 mM) ), like the MoFe protein, contains a tightly coordinated Mo site; however, the coordination of the Mo atom in NifENFeMoco(20 mM) is somewhat asymmetric compared to that of the Mo atom in the MoFe protein, likely due to a different ligand environment at the Mo end of the FeMoco in NifEN (34). Consistent with the outcome of the spectroscopic studies, activity analysis shows that NifENFeMoco(20 mM) is capable of reconstituting the FeMocodeficient nifB MoFe protein at a level comparable to that by the isolated FeMoco (34). These findings indicate that redox chemistry plays an important role in the maturation of FeMoco on NifEN. The observation that the cluster conversion on NifEN is a redox-dependent process also underlines the significant role of Fe protein in this process, considering that Fe protein is the only known reductase in the maturation assay of NifEN-bound

Assembly of Nitrogenase MoFe Protein

37

precursor. It has been shown that Fe protein re-isolated after incubating with MgATP, MoO2− 4 , homocitrate, and NifEN (designated Fe proteinMo/homocitrate ) can directly serve as the Mo and homocitrate source for FeMoco formation on NifEN (35), suggesting that Fe proteinMo/homocitrate is “loaded” with Mo and homocitrate (Fig. 3.3). Comparative Mo K-edge XAS analysis of Fe proteinMo/homocitrate and free molybdate (MoO2− 4 ) shows an edge shift of ∼2.3 eV, indicating that the Mo atom associating with the Fe proteincomplete is more reduced compared to its supplied form (35). There is an additional ∼0.5 eV edge shift if homocitrate is omitted during the preparation of Fe proteinMo/homocitrate , suggesting an impact of the attachment of homocitrate to Mo on the Mo environment in Fe proteinMo/homocitrate (35). The loading of Mo and homocitrate on Fe protein is strictly dependent on ATP hydrolysis, as FeMoco maturation cannot occur when ATP is substituted with ADP or nonhydrolyzable ATP analogs, or when the wild-type Fe protein is replaced by Fe protein mutants defective in MgATP hydrolysis (33). Taken together, these observations imply that, like its role in catalysis, the Fe protein serves as an ATP-dependent

Fig. 3.3. Fe protein serves as the molybdenum and homocitrate insertase during the process of FeMoco assembly. HC, homocitrate; Mo, molybdenum.

38

Lee et al.

reductase in biosynthesis, although the ATP-hydrolysis-driven electron transfer by Fe protein is used for the mobilization of Mo and homocitrate for the maturation of NifEN-associated precursor in this case. 3.2. The Incorporation of FeMoco into the MoFe Protein

Upon the insertion of Mo and homocitrate into the NifENbound precursor, a fully matured FeMoco is generated on NifEN and ready to be transferred to the MoFe protein. It has been suggested earlier that the delivery of FeMoco to the MoFe protein requires carrier proteins, such as NifX (nifX gene product), NifY (nifY gene product), or NafY (nafY gene product). For example, NafY was proposed to interact with the α2 β2 -tetrameric MoFe protein of A. vinelandii in a manner similar to that of the additional γ-subunit of the α2 β2 γ2 -hexameric Klebsiella pneumoniae MoFe protein during the process of FeMoco assembly (18, 36–38). However, the absolute requirement of these FeMoco carriers was precluded by the previous observation that the unaltered nitrogen-fixing ability of hosts carrying deletions of the carrierencoding genes and the recent finding of the full reconstitution of the FeMoco-deficient nifB MoFe protein by NifENFeMoco upon direct protein–protein interactions (33, 34). As such, the previously hypothesized carrier proteins may instead serve as accessory factors that improve the in vivo efficiency of the FeMoco transfer. For instance, they can either act as chaperones that stabilize the protein complex or serve as post-translational modifiers that finetune the key residue(s) in the protein(s) for cluster formation. Recently, one such protein factor, NifX, has been reported to participate in FeMoco assembly in a chaperone-like function (39). In addition to the nif-encoded proteins, GroEL has also been shown to facilitate the incorporation of FeMoco into the MoFe protein, although the molecular details of how GroEL functions in this process have not been elucidated (40). The direct transfer of FeMoco between NifEN and MoFe protein is supported by the recent observation of complex formation between NifENFeMoco and the FeMoco-deficient nifB MoFe protein in native PAGE (41). Formation of such a complex cannot be observed between the precursor-bound NifEN and nifB MoFe protein (41), suggesting that, upon the maturation of precursor, NifEN undergoes a conformational rearrangement that enables it to dock on the MoFe protein for the subsequent cluster transfer. Sequence comparison between NifEN and MoFe protein reveals the presence of similar cluster-binding sites in the two proteins; however, several residues that either covalently ligate to or indirectly interact with the FeMoco in the MoFe protein are not conserved in the corresponding NifEN sequence. Thus, the docking of NifEN on the MoFe protein may bring their respective FeMoco-binding sites in close proximity, which facilitates the “diffusion” of FeMoco from its assembly site in NifEN

Assembly of Nitrogenase MoFe Protein

39

(low-affinity site) toward its target binding site in MoFe protein (high-affinity site). Once the FeMoco diffuses out of the NifEN and reaches the surface of the MoFe protein, it migrates through an insertion “funnel” toward its destined location within the MoFe protein. Identification of such a mechanism for FeMoco insertion was assisted by the comparative structural analysis of nifB and wild-type MoFe proteins, which reveals the presence of a positively charged funnel that extends all the way from the protein surface to the imbedded FeMoco-binding site, sufficiently sized to accommodate the insertion of the negatively charged FeMoco (42). The negative charge of the homocitrate moiety of the FeMoco, therefore, is essential for the incorporation of the cofactor into the MoFe protein. Indeed, a recent study shows that NifEN bound with a homocitrate-free form of FeMoco (designated NifENFeMo-cluster ) is unable to deliver FeMoco to the MoFe protein, although complex formation can be observed between NifENFeMo-cluster and the nifB MoFe protein in the native PAGE (41). Moreover, upon separation of NifENFeMo-cluster and the nifB MoFe protein, the FeMo cluster cannot be found in either NifEN or MoFe protein, suggesting that the cluster is lost upon such a treatment (41). Combined outcome of these studies suggests that (i) the insertion of Mo alone into the NifEN-associated precursor is sufficient to induce a conformational change of NifEN that allows it to form a complex with the nifB MoFe protein and (ii) the unsuccessful attempt of NifEN to deliver the FeMo cluster leaves the cluster “stuck” in between the two proteins in the complex, which is lost upon the separation of the two proteins. A closer examination of the insertion funnel identifies three distinct regions that are important for FeMoco insertion (42). One, designated the “lid loop,” consists of residues α353–α364 of the MoFe protein. Among these residues, Hisα362 , which is located at the tip of this flexible loop, could serve as the first docking point for FeMoco at the entrance of the insertion funnel (Fig. 3.4, 1). The second, termed the “His triad,” comprises Hisα274 , Hisα442 , and Hisα451 of the MoFe protein. Together, they could provide a histidine-rich, intermediary docking point for FeMoco halfway down the insertion funnel (Fig. 3.4, 2). The third, designated the “switch/lock,” is composed of Hisα442 and Trpα444 of the MoFe protein. Through a switch in positions at the bottom of the insertion funnel, Trpα444 could lock FeMoco in its binding site by its bulky side chain (Fig. 3.4, 3). The participation of these residues in the FeMoco insertion process was subsequently confirmed by mutational analyses (43–45), which demonstrate a specific and substantial reduction in the level of FeMoco incorporation upon the removal of the positive charge, ligand capacity, or steric effect at these positions.

40

Lee et al.

His362

“lid-loop”

His451

His274

“His-triad”

His442

“switch/lock”

Trp444

Fig. 3.4. Schematic presentation of the FeMoco insertion funnel in the nifB MoFe protein.

The incorporation of FeMoco into the MoFe protein is accompanied by an overall conformational rearrangement of the protein. Small angle x-ray scattering (SAXS) analysis shows that the Rg value of the wild-type MoFe protein (40.2 Å) is slightly, yet reproducibly, smaller than that of the nifB MoFe protein (42.4 Å) (46), suggesting the MoFe protein assumes a more compact conformation upon the insertion of FeMoco. Consistent with this finding, x-ray crystallographic analysis reveals a more “open” conformation of the nifB MoFe protein than that of the wild-type MoFe protein, particularly in the αIII domain, where the FeMoco insertion funnel resides. It is possible, therefore, that the insertion funnel is closed up upon the insertion of FeMoco and that this process triggers an overall conformational rearrangement of the MoFe protein conformation. The distance between the P-cluster and FeMoco could be optimized following the compacting of the MoFe protein, which enables an efficient transfer of electrons between the two clusters during the substrate turnover. 3.3. The “In Situ” Assembly of P-Cluster on MoFe Protein

Based on genetic evidence, the assembly of P-cluster, like that of FeMoco, is initiated by the action of NifUS (18). However, the post-NifUS biosynthetic steps of the P-cluster, unlike those of the FeMoco, occur within the MoFe protein (hence the term

Assembly of Nitrogenase MoFe Protein

41

“in situ” assembly). There is strong evidence that Fe protein is crucial for the maturation of P-cluster; however, the mechanistic details of the role of Fe protein in the process have remained unclear till the recent characterization of an MoFe protein isolated from the nifH-deletion background (designated nifH MoFe protein). In addition to the Fe protein, NifZ (gene product of nifZ), a small protein factor (∼19 kDa), has been shown to be essential for the assembly of P-cluster. The functions of Fe protein and NifZ in P-cluster assembly will be discussed in greater detail below. Earlier attempts at extracting the P-clusters from MoFe protein have resulted in the recovery of [Fe4 S4 ] clusters, the decomposed fragments of the P-cluster. This observation points to the possibility that the P-cluster is formed by the fusion of [Fe4 S4 ] fragments, an approach that has been employed by synthetic chemists to generate Fe/S clusters of high nuclearity (47). The biological evidence for such a hypothesis came from the characterization of the nifH MoFe protein (48). Like nifB MoFe protein, nifH MoFe protein is FeMoco deficient due to the absence of the nifH-encoded Fe protein, which is an essential factor for the maturation of FeMoco on NifEN (see Section 3.1 above). However, in contrast to the nifB MoFe protein, which contains the intact P-cluster, the nifH MoFe protein contains a variant form of the P-cluster, which displays an S = 1/2 EPR signal in the dithionite-reduced state (48). Fe K-edge XAS/EXAFS analysis (49) reveals that the P-cluster variant in nifH MoFe protein is composed of paired [Fe4 S4 ]-like clusters, either completely separated or bridged at the edges (Fig. 3.5), whereas VTVHMCD analysis provides further support for the model of separated [Fe4 S4 ] centers, showing that each [Fe4 S4 ] pair consists of one [Fe4 S4 ] cluster in the +1 oxidation state and one diamagnetic [Fe4 S4 ]-like cluster, which becomes paramagnetic upon IDS oxidation (50). A

Cysα154

B

C

Cysα62

S

Cysα62

D Cysα154

Cysα88 O Cysβ95

Fe

Cysβ153

Cysα88

Cysβ95 Serβ188 Cysβ70

Cysβ70

Cysβ153

Fig. 3.5. XAS/EXAFS-based structural models of P-cluster precursor in the nifH MoFe protein.

Lee et al.

The physiological relevance of the paired [Fe4 S4 ] clusters to P-cluster assembly is demonstrated by the fact that the nifH MoFe protein can be fully activated upon incubation with Fe protein, MgATP, dithionite, and isolated FeMoco in an in vitro assay (51). Conversion of the precursors to the Pclusters can be monitored by the decrease in the precursorspecific S = 1/2 signal and the concomitant appearance of the P-cluster (P2+ )-specific g = 11.8 signal (Fig. 3.6). It is interesting to note that the maturation of P-clusters in the nifH MoFe protein follows a biphasic pattern (Fig. 3.6), which suggests a sequential formation of the two P-clusters in this α2 β2 tetrameric protein (see Section 3.4 for detailed discussion). Furthermore, the formation of P-cluster, like that of FeMoco (see Section 3.2 above), is dependent on the concentration of the reductant, dithionite. An increase of the dithionite concentration from 2 to 20 mM results in a fourfold to fivefold increase in the level of P-cluster maturation, suggesting, once again, a key role of redox chemistry in reductively coupling the paired [Fe4 S4 ] subclusters into a mature [Fe8 S7 ] P-cluster (51). Given that nifH MoFe protein contains precursors in place of the fully assembled P-clusters, it represents an intermediate that occurs earlier than nifB MoFe protein along the biosynthetic pathway. Comparative SAXS analysis of the nifH, nifB, and wild-type MoFe proteins shows that the nifH MoFe protein exists in the most extended conformation (Rg = 45.7 Å), followed by the nifB MoFe protein (Rg = 42.4 Å), and then the wild-type MoFe protein (Rg = 40.2 Å) (46). The increase in the size of nifH MoFe protein is correlated to an increase in the solvent accessibility of the P-cluster precursor Fe atoms and can be best modeled by a 6 Å gap at the α/β-subunit interface that is absent from the structure of either nifB or wild-type MoFe protein (46). These results suggest that maturation of the P-cluster is

100

80 40 60 60 40 80

20

100

0 0

Relative Area of S = 1/2 Signal (%)

20

80 60 40 20

Average Specific Activity (%)

100

Relative Area of g = 11.8 signal (%)

42

0

20 40 60 80 100 120

Time (Minutes)

Fig. 3.6. Correlation among the precursor-specific S =1/2 signal (), the P-cluster (P2+ )-specific g = 11.8 signal (•), and the specific activity ( ) of the nifH MoFe protein during the process of P-cluster maturation.

Assembly of Nitrogenase MoFe Protein

43

likely accompanied by a conformational change of the MoFe protein that brings the α- and β-subunits together, which facilitates the subsequent fusion of the [Fe4 S4 ] subclusters. 3.4. The Stepwise Assembly of MoFe Protein

As we advance toward a better understanding of the biosynthesis of FeMoco and P-cluster, one important question emerges, that is, how does MoFe protein (a multimeric protein) coordinate the assembly of the multiple metal centers (two FeMoco and two P-clusters)? The capture of a FeMoco-deficient, yet P-clusterintact nifB MoFe protein (see above) seems to suggest the formation of P-cluster prior to the incorporation of FeMoco; yet, it is not clear how the MoFe protein orchestrates the assembly of its αβ-dimeric halves, each of which contains a set of FeMoco and P-cluster. The latter issue was addressed by a recent study, which reveals a biphasic pattern of P-cluster maturation in the nifH MoFe protein (51). The increase in the magnitude of the P-cluster (P2+ )-specific g = 11.8 EPR signal aligns well with the increase of activity of the matured protein, both showing two distinct phases that are separated by a “lag” period at ∼50% of their respective maximum levels. Such a biphasic pattern points to the stepwise formation of the two P-clusters, i.e., one at a time, in the two αβ-dimers of the MoFe protein. It should be noted that the disappearance of the precursor-specific signal occurs rather rapidly (within the first 5 min), while the appearance of the P-clusterspecific signal occurs much more slowly (over a time period of 2 h). This observation suggests that, following the initial coupling of the paired [Fe4 S4 ] clusters, the P-cluster is formed by a slow rearrangement of the cluster intermediate—a process that is likely facilitated by the conformational changes of the MoFe protein at the α/β-subunit interface. Interestingly, the half-matured nifH MoFe protein that appears transiently at the “lag” phase can be captured by isolating the MoFe protein from a nifZ/nifB-double deletion background (designated nifZnifB MoFe protein). Such an MoFe protein species is FeMoco deficient due to the absence of NifB. Moreover, like the half-matured nifH MoFe protein, the nifZnifB MoFe protein also contains one fully active P-cluster in one αβ-half and one paired [Fe4 S4 ]-like cluster in the other, indicating a specific role of NifZ in the formation of the “second” P-cluster (52–54). Consistent with its half P-cluster content, the nifZnifB MoFe protein can be reconstituted to ∼50% of the maximum activity upon the addition of the isolated FeMoco (54). Such a half-reconstituted nifZnifB MoFe protein can also be accumulated in a nifZ-deletion strain, which expressed an MoFe protein containing one P-cluster and one FeMoco in one αβ-dimer and a pair of [Fe4 S4 ]-like clusters and a vacant FeMoco site in

44

Lee et al.

the other (52). More importantly, following the incubation with NifZ and Fe protein/MgATP, the nifZnifB MoFe protein can be further activated to ∼86% of the maximum activity, suggesting a concerted action of NifZ and Fe protein in the maturation of the “second” P-cluster. EPR analysis provides further evidence of the maturation of P-clusters the nifZnifB MoFe protein, demonstrating a decrease of the precursor-associated S = 1/2 signal and a concurrent increase in the P-cluster (P2+ )-specific g = 11.8 signal during the maturation process (54). Curiously, the action of NifZ is required prior to that of the Fe protein/MgATP for the maturation of the “second” P-cluster in the nifZnifB MoFe protein (54). It can be speculated, therefore, that NifZ may induce a conformational change of the nifZnifB MoFe protein or covalently modifies certain residues of this protein, which facilitates the subsequent reductive coupling of the “second” pair of [Fe4 S4 ]-like subclusters by Fe protein.

4. Conclusion There has been major progress toward understanding the biosynthetic mechanism of MoFe protein and its associated metal clusters in the recent years. However, the job of piecing together this biological puzzle is far from finished. With regard to the assembly of FeMoco, the biosynthetic events that occur prior to NifEN and, in particular, those hosted by NifB remain largely unknown, whereas with regard to the assembly of P-cluster, the biosynthetic steps leading to the formation of the precursor, as well as the subsequent incorporation of these precursors into the MoFe protein, are poorly understood. Additionally, how assembly proteins interact with one another during the process of cluster transfer and what accessory factors are involved in which biosynthetic step have not been well explored. A better understanding of this complex process depends on future biochemical, spectroscopic, and structural studies that determine the mode of action of each biosynthetic component and establish the network of interactions among these components.

Acknowledgments The authors are supported by the National Institutes of Health grant GM 67626 (M.W.R.) and Herman Frasch Foundation grant 617-HF07 (M.W.R.).

Assembly of Nitrogenase MoFe Protein

45

References 1. Burgess BK, Lowe DJ (1996) Mechanism of molybdenum nitrogenase. Chem Rev 96:2983–3012 2. Howard JB, Rees DC (1996) Structural basis of biological nitrogen fixation. Chem Rev 96:2965–2982 3. Smith BE (1999) Structure, function, and biosynthesis of the metallosulfur clusters in nitrogenases. Adv Inorg Chem 47:159–218 4. Rees DC, Tezcan FA, Haynes CA et al (2005) Structural basis of biological nitrogen fixation. Philos Trans R Soc London, Ser A 363:971–984 5. Howard JB, Rees DC (2006) How many metals does it take to fix N2 ? A mechanistic overview of biological nitrogen fixation. Proc Natl Acad Sci USA 103:17088–17093 6. Schwarz G, Mendel RR, Ribbe MW (2009) Molybdenum cofactors, enzymes and pathways. Nature 460:839–847 7. Hu Y, Fay AW, Lee CC et al (2008) Assembly of nitrogenase MoFe protein. Biochemistry 47:3973–3981 8. Einsle O, Tezcan FA, Andrade SL et al (2002) Nitrogenase MoFe-protein at 1.16 Å resolution: A central ligand in the FeMocofactor. Science 297:1696–1700 9. Kim J, Rees DC (1992) Crystallographic structure and functional implications of the nitrogenase molybdenum iron protein from Azotobacter vinelandii. Nature 360:553–560 10. Chan MK, Kim J, Rees DC (1993) The nitrogenase FeMo-cofactor and P-cluster pair: 2.2 Å resolution structures. Science 260:792–794 11. Peters JW, Stowell MH, Soltis SM et al (1997) Redox-dependent structural changes in the nitrogenase P-cluster. Biochemistry 36:1181–1187 12. Zimmermann R, Orme-Johnson WH, Münck E et al (1978) Nitrogenase-X: Mössbauer and EPR studies on reversibly oxidized MoFe protein from Azotobacter vinelandii OP: Nature of iron centers. Biochim Biophys Acta 537:185–207 13. Surerus KK, Hendrich MP, Christie PD et al (1992) Mössbauer and integer-spin EPR of the oxidized P-clusters of nitrogenase: Pox is a non-Kramers system with a nearly degenerate ground doublet. J Am Chem Soc 114:8579–8590 14. Watt GD, Burns A, Lough S et al (1980) Redox and spectroscopic properties of oxidized MoFe protein from Azotobacter vinelandii. Biochemistry 19:4926–4932 15. Pierik AJ, Wassink H, Haaker H et al (1993) Redox properties and EPR spectroscopy

16. 17. 18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

of the P-clusters of Azotobacter vinelandii MoFe Protein. Eur J Biochem 212:51–61 Shah VK, Brill WJ (1977) Isolation of an iron-molybdenum cofactor from nitrogenase. Proc Natl Acad Sci USA 74:3249–3253 Burgess BK (1990) The iron molybdenum cofactor of nitrogenase. Chem Rev 90: 1377–1406 Dos Santos PC, Dean DR, Hu YL et al (2004) Formation and insertion of the nitrogenase iron-molybdenum cofactor. Chem Rev 104:1159–1173 Frazzon J, Dean DR (2002) Biosynthesis of the nitrogenase iron-molybdenum-cofactor from Azotobacter vinelandii. Met Ions Biol Syst 39:163–186 Yuvaniyama P, Agar JN, Cash VL et al (2000) NifS-directed assembly of a transient [2Fe2S] cluster within the NifU protein. Proc Natl Acad Sci USA 97:599–604 Agar JN, Yuvaniyama P, Jack RF et al (2000) Modular organization and identification of a mononuclear iron-binding site within the NifU protein. J Biol Inorg Chem 5: 167–177 Dos Santos PC, Smith AD, Frazzon J et al (2004) Iron-sulfur cluster assembly: NifUdirected activation of the nitrogenase Fe protein. J Biol Chem 279:19705–19711 Smith AD, Jameson GNL, Dos Santos PC et al (2005) NifS-mediated assembly of [4Fe-4S] clusters in the N- and C-terminal domains of the NifU scaffold protein. Biochemistry 44:12955–12969 Allen RM, Chatterjee R, Ludden PW et al (1995) Incorporation of iron and sulfur from NifB cofactor into the iron-molybdenum cofactor of dinitrogenase. J Biol Chem 270:26890–26896 Dean DR, Jacobson MR (1992) Biochemical genetics of nitrogenase. In: Stacey G, Burris RH, Evans HJ (eds) Biological Nitrogen Fixation, pp. 763–834. Chapman & Hall, New York, NY Curatti L, Hernandez JA, Igarashi RY et al (2007) In vitro synthesis of the ironmolybdenum cofactor of nitrogenase from iron, sulfur, molybdenum, and homocitrate using purified proteins. Proc Natl Acad Sci USA 104:17626–17631 Curatti L, Ludden PW, Rubio LM (2006) NifB-dependent in vitro synthesis of the ironmolybdenum cofactor of nitrogenase. Proc Natl Acad Sci USA 103:5297–5301 Roll JT, Shah VK, Dean DR et al (1995) Characteristics of NifNE in Azotobacter vinelandii strains: Implications for

46

29.

30.

31.

32.

33. 34.

35.

36.

37.

38.

39.

40.

Lee et al. the synthesis of the iron-molybdenum cofactor of dinitrogenase. J Biol Chem 270: 4432–4437 Goodwin PJ, Agar JN, Roll JT et al (1998) The Azotobacter vinelandii NifEN complex contains two identical [4Fe-4S] clusters. Biochemistry 37:10420–10428 Brigle KE, Weiss MC, Newton WE et al (1987) Products of the iron-molybdenum cofactor-specific biosynthetic genes, nifE and nifN, are structurally homologous to the products of the nitrogenase molybdenumiron protein genes, nifD and nifK. J Bacteriol 169:1547–1553 Hu Y, Fay AW, Ribbe MW (2005) Identification of a nitrogenase iron-molybdenum cofactor precursor on NifEN complex. Proc Natl Acad Sci USA 102:3236–3241 Corbett MC, Hu Y, Fay AW et al (2006) Structural insights into a protein-bound ironmolybdenum cofactor precursor. Proc Natl Acad Sci USA 103:1238–1243 Hu Y, Corbett MC, Fay AW et al (2006) FeMo cofactor maturation on NifEN. Proc Natl Acad Sci USA 103:17119–17124 Yoshizawa JM, Blank MA, Fay AW et al (2009) Optimization of FeMoco maturation on NifEN. J Am Chem Soc 131: 9321–9325 Hu Y, Corbett MC, Fay AW et al (2006) Nitrogenase Fe protein: A molybdate/homocitrate insertase. Proc Natl Acad Sci USA 103:17125–17130 Homer MJ, Dean DR, Roberts GP (1995) Characterization of the gamma-protein and its involvement in the metallocluster assembly and maturation of dinitrogenase from Azotobacter vinelandii. J Biol Chem 270:24745–24752 White TC, Harris GS, Orme-Johnson W (1992) Electrophoretic studies on the assembly of the nitrogenase molybdenum-iron protein from the Klebsiella pneumoniae nifD and nifK gene-products. J Biol Chem 267:24007–24016 Homer MJ, Paustian TD, Shah VK et al (1993) The NifY product of Klebsiella pneumoniae is associated with apodinitrogenase and dissociates upon activation with the iron-molybdenum cofactor. J Bacteriol 175: 4907–4910 Hernandez JA, Igarashi RY, Soboh B et al (2007) NifX and NifEN exchange NifB cofactor and the VK-cluster, a newly isolated intermediate of the iron-molybdenum cofactor biosynthetic pathway. Mol Microbiol 63:177–192 Ribbe MW, Burgess BK (2001) The chaperone GroEL is required for the final assem-

41.

42. 43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

bly of the molybdenum-iron protein of nitrogenase. Proc Natl Acad Sci USA 98: 5521–5525 Fay AW, Blank MA, Yoshizawa JM et al (2010) Formation of a homocitratefree iron-molybdenum cluster on NifEN: implications for the role of homocitrate in nitrogenase assembly. Dalton Trans 39:3124–3130 Schmid B, Ribbe MW, Einsle O et al (2002) Structure of a cofactor-deficient nitrogenase MoFe protein. Science 296:352–356 Hu Y, Fay AW, Schmid B et al (2006) Molecular insights into nitrogenase FeMoco insertion: Trp 444 of MoFe protein α subunit locks FeMoco in its binding site. J Biol Chem 281:30534–30541 Hu Y, Fay AW, Ribbe MW (2007) Molecular insights into nitrogenase FeMoco insertion: The role of His362 of MoFe protein α subunit in FeMoco incorporation. J Biol Inorg Chem 12:449–460 Fay AW, Hu Y, Schmid B et al (2007) Molecular insights into nitrogenase FeMoco insertion: The role of His 274 and His 451 of MoFe protein α subunit. J Inorg Biochem 101:1630–1641 Corbett MC, Hu Y, Fay AW et al (2007) Conformational differences between Azotobacter vinelandii nitrogenase MoFe proteins as studied by small angle X-ray scattering. Biochemistry 46:8066–8074 Lee SC, Holm RH (2004) The clusters of nitrogenase: Synthetic methodology in the construction of weak-field clusters. Chem Rev 104:1135–1158 Ribbe MW, Hu Y, Guo M et al (2002) The FeMoco-deficient MoFe protein produced by a nifH-deletion strain of Azotobacter vinelandii shows unusual P-cluster features. J Biol Chem 277:23469–23476 Corbett MC, Hu Y, Naderi F et al (2004) Comparison of iron-molybdenum cofactor deficient nitrogenase MoFe proteins by X-ray absorption spectroscopy: Implications for P-cluster biosynthesis. J Biol Chem 279:28276–28282 Broach RB, Rupnik K, Hu Y et al (2006) VTVH-MCD spectroscopic study of the metal clusters in the nifB and nifH MoFe proteins of nitrogenase from Azotobacter vinelandii. Biochemistry 45: 15039–15048 Lee CC, Blank MA, Fay AW et al (2010) Stepwise formation of P-cluster in nitrogenase MoFe protein. Proc Natl Acad Sci USA 106:18474–18478 Hu Y, Fay AW, Dos Santos PC et al (2004) Characterization of Azotobacter vinelandii

Assembly of Nitrogenase MoFe Protein nifZ-deletion strains: Indication of stepwise MoFe protein assembly. J Biol Chem 279:54963–54971 53. Cotton MS, Rupnik K, Broach RB et al (2009) VTVH-MCD study of the

47

nifBnifZ MoFe protein from Azotobacter vinelandii. J Am Chem Soc 131:4558–4559 54. Hu Y, Fay AW, Lee CC et al (2007) P-Cluster maturation on nitrogenase MoFe protein. Proc Natl Acad Sci USA 104:10424–10429

Chapter 4 Genomic Analysis of Nitrogen Fixation Ina P. O’Carroll and Patricia C. Dos Santos Abstract Advances in sequencing technology in the past decade have enabled the sequencing of genomes of thousands of organisms including diazotrophs. Genomics have enabled thorough analysis of the gene organization of nitrogen-fixing species, the identification of new genes involved in nitrogen fixation, and the identification of new diazotrophic species. This chapter reviews key characteristics of nitrogen-fixing genomes and methods to identify and analyze genomes of new diazotrophs using genome scanning. This chapter refers to Azotobacter vinelandii, a well-studied nitrogen-fixing organism, as a model for studying nitrogen-fixing genomes. We discuss the main nitrogen fixation genes as well as accessory genes that contribute to diazotrophy. We also review approaches that can be used to modify genomes in order to study nitrogen fixation at the genetic, biochemical, and biophysical level. Key words: Genome sequencing, genome scanning, genomics, Azotobacter vinelandii, nitrogen fixation.

1. N2 Fixation in the Post-genomic Era

Biological conversion of nitrogen gas (N2 ) to ammonia (NH3 ) is exclusive to archaeal and bacterial species. The identification of species capable of diazotrophy is not straightforward because nitrogen fixation is not limited to specific genera nor is it universally distributed among all species of the same genus. The challenge of experimental detection of biological nitrogen fixation further complicates the task. Diazotrophy is often a restricted process in prokaryotic systems and its activation through inducible expression only occurs when the appropriate physiological and nutritional conditions are met (1). For example, Klebsiella pneumoniae is a facultative anaerobe, but is only capable of diazotrophic growth under anaerobic conditions (2). In other

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_4, © Springer Science+Business Media, LLC 2011

49

50

O’Carroll and Dos Santos

species, such as the legume symbionts Rhizobium meliloti and Rhizobium leguminosarum, nitrogen fixation is fully restricted to symbiosis (3). Therefore, novel experimental demonstration of diazotrophy is complicated given that species-specific nutritional and physiological conditions have not been determined for many organisms. Those inherent challenges prevent the true assessment of the biodiversity and phylogenetic distribution of diazotrophic species. Nowadays, the rapid increase in the number of prokaryotic species with sequenced genomes enables the development of in silico predictions of biochemical pathways in microbes. Such assumptions, although very accurate, yield putative results and do not obviate the need for genetic and biochemical confirmation of gene function. During the post-genomic era, in silico searching tools are being used to guide scientists to identify complex pathways such as nitrogen fixation. This chapter will briefly summarize some of the current methods for the identification of nitrogenfixing species using genome scanning, and, as an example, will expand on genome analysis of the well-studied diazotrophic bacterium Azotobacter vinelandii. 1.1. Analyzing N2 -Fixing Genomes and Identifying New Diazotrophs

All diazotrophs studied so far use at least one of the four types of nitrogenases (4). The first three types of nitrogenase are structurally, phylogenetically, and mechanistically related and differ in the nature of the heterometal component to include either molybdenum or vanadium or contain a homometal cofactor that contains only iron (4). The fourth type of nitrogenase was identified from a single species, Streptomyces thermoautotrophicus, and is a superoxide-dependent, molybdopterin-containing nitrogenase (5). Neither the amino acid sequence of this fourth type of nitrogenase nor the sequence of the S. thermoautotrophicus genome has been determined yet. The inventory and distribution of species coding for the fourth type of nitrogenase is not known, a finding that underscores the limited number of sequenced genomes of diazotrophs.

1.1.1. Genome Scanning Using NifH as Query

Most, if not all, phylogenetic analyses in diazotrophs use as query the amino acid sequence of the well-studied Mo-dependent nitrogenase reductase also known as iron protein coded by nifH (6–10). Genes coding for NifH or proteins similar to NifH have been found in all known diazotrophs with sequenced genomes. The strict requirement of NifH in biological nitrogen fixation and its universal presence in diazotrophs has resulted in this protein serving as a sequence tag or barcode for the identification of nitrogen fixers. Such analysis has proved successful in the identification of species able to express not only Mo-dependent but also vanadium- and iron-only nitrogenases. Genomic analysis using the sequence of NifH as a query results in BLAST hits that include NifH, VnfH, and AnfH components of the Mo-, V-, and Fe-only

Genomic Analysis of Nitrogen Fixation

51

nitrogenases, respectively (10). Since the corresponding genes are located in distinct locations in the genome, proper functional assignment of their gene products must involve gene neighborhood scanning as well as phylogenetic analysis using sequences of studied proteins in orthologous systems. 1.1.2. Genome Scanning Using NifD as Query

Genome sequencing using NifD sequences as queries can also target the identification of novel nitrogen-fixing organisms, as well as the identification of the number and type of nitrogenases coded in these species (9–11). However, these analyses are not straightforward due to the existence of paralogous genes in these genomes. NifD, VnfD, and AnfD are paralogous sequences that code for the α-chain of the Mo-, V-, and Fe-only nitrogenases, respectively (12). In addition to Nif/Vnf/AnfD sequences, genomes of diazotrophs also code for additional proteins similar to NifD (Fig. 4.1) (10). The biosynthesis of the nitrogenase proteins and its cofactors involves a complex pathway (as described in

Fig. 4.1. Phylogenetic tree of NifD homologs sequences from A. vinelandii and P. stutzeri genomes. The tree was constructed as previously described (18).

52

O’Carroll and Dos Santos

Chapter 3) which includes, among several other proteins, the participation of the NifE–NifN and VnfE–VnfN complexes in Mo- and V-dependent synthesis of active site cofactors, respectively (13, 14). Sequence alignments and phylogenetic analyses show a significant degree of similarity among NifD/VnfD/AnfD and NifE/VnfE proteins (Fig. 4.1) (14). Furthermore, improper assignment of computational-search parameters such as matrix, filter, and threshold values can also reveal BLAST hits of Nif/Vnf/Anf-K and N sequences, which are similar to each other, but also show similarity to D- and E-type sequences. The presence of paralogous nitrogenase gene sequences in the genomes of lessstudied species often results in the misassignment of nitrogenase components during automated annotation. In some cases inspection of the genomic context can assist the functional assignment of nitrogenase coding sequences. For example, V-dependent and iron-only nitrogenases are composed of three gene products coded by vnfDGK and anfDGK. In these cases, the presence of G-type sequences can assist the diagnostic of vnf and anf sequences. Distinction between the different proteins on the basis of primary structure alone is challenging and needs to be accompanied by genetic and biochemical evidence. 1.1.3. Bioinformatic Analysis Needs to Be Accompanied by Manual Curation

Despite the sequence similarities among nitrogenase catalytic and biosynthetic proteins, all known diazotrophs with exception of S. thermoautotrophicus code for proteins whose sequences are similar to NifH and NifD (8). Therefore, genome sequence hits by independent BLAST searches using both these sequences as queries are strong indicators of the genetic makeup for nitrogen fixation. Exceptions are the non-diazotrophic species that use nitrogenase-like enzymes in the biosynthesis of chlorophylls and bacteriochlorophylls (15, 16). In these cases, careful bioinformatic analysis can assist in the proper assignment of such gene products to accurately predict their physiological roles. In silico pathway identification is a powerful tool for predicting diazotrophy in organisms with sequenced genomes. However, it does not avert the need to assay nitrogenase activity and establish that the organism can grow on dinitrogen as the sole nitrogen source. More than 1,000 fully sequenced and assembled genomes from archaeal and bacterial species have now been deposited in the NCBI database. BLAST analysis using NifH and NifD sequences identified over 150 genomes from 105 unique species in this database containing gene products with sequences similar to both NifH and NifD proteins (our unpublished results). This distribution suggests that nearly 10% of sequenced genomes are likely to be diazotrophs. Manual literature searches to validate computational predictions indicate that more than half of these species have not yet been confirmed as being diazotrophs. This observation may be used to initiate reverse genetics studies

Genomic Analysis of Nitrogen Fixation

53

to assess biological nitrogen fixation in these species. Complete genome sequences of prokaryotic diazotrophs also allow comparison of the genetic requirements for biological nitrogen fixation in distinct physiological niches such as sulfur-reducing, oxygenic, or hyperthermal environments. This type of analysis can reveal further clues about the immense biodiversity of nitrogen-fixing organisms.

2. A. vinelandii— A Model for Nitrogen Fixation

2.1. Genes Involved in Nitrogen Fixation

Among diazotrophs with sequenced genomes, A. vinelandii is one of the best studied organisms. A. vinelandii is a free-living bacterium that is able to fix nitrogen in an aerobic culture. For decades it has been used as a model for nitrogen fixation for numerous reasons, including (i) its amenability to genetic manipulation, (ii) its ability to fix nitrogen during aerobic growth, (iii) its nutritional flexibility—as evidenced by its capacity to fix nitrogen via three different pathways and its ability to adapt its metabolism to a variety of nutrients and media additives, and (iv) the fact that its genome has been fully sequenced and manually curated (17, 18). The genome of A. vinelandii is composed of a single circular chromosome of 5,365,318 bp predicted to code for 5051 proteins (18). The closest relative of A. vinelandii with a sequenced genome is the diazotroph Pseudomonas stutzeri (19). Although P. stutzeri is capable of aerobic growth, nitrogen fixation is restricted to microaerobic conditions (20, 21). On the other hand, A. vinelandii is able to catalyze nitrogen fixation, an oxygen-sensitive reaction, under ambient oxygen concentrations (∼20% O2 ) (22, 23). This unique ability to simultaneously perform two incompatible cellular processes, oxidative phosphorylation and nitrogen fixation, is possible thanks to its respiratory protection mechanism. During nitrogen fixation, A. vinelandii can adjust its respiration rate to maintain a low level of cytoplasmic oxygen (24). Genome sequencing revealed the presence of five terminal oxidases, in conjunction with extra NADH oxidoreductases and other respiratory complexes that supply electrons to terminal oxidases and thus increase oxygen consumption (18). Some of these genes, as in the case of cydAB I, are known to be involved in respiratory protection and play an essential role during aerobic nitrogen fixation (25, 26). The genome of A. vinelandii codes for three oxygen-sensitive nitrogenases (Mo-, V-, and Fe-only dependent enzymes), the structural genes of which are located in three distinct regions of the chromosome, but all are near the origin of replication

54

O’Carroll and Dos Santos

Fig. 4.2. Nitrogen-fixing gene regions in A. vinelandii. Diagrammatic representation of the nif, anf, and vnf gene regions and their genome locations. Predicted σ54 promoter regions are indicated by arrows. Genes required for a specific type of nitrogenase are shown in light gray, genes required for all three types of nitrogenase are shown in dark gray.

(Fig. 4.2) (18). Although the sequence of the nitrogenase genes has been known for more than two decades (27–30), the genomic location was only identified during the final assembly of the A. vinelandii genome (18). The two regions of nif genes are located equidistantly from the origin of replication. The proximity to the origin of replication has been suggested to increase gene dosage during active growth, an event that would contribute to the high level of expression of Mo-dependent nitrogenase (18). The genes coding for the Mo-dependent nitrogenase components (nifHDK) and their regulatory and assembly systems are located in two discrete regions. The major nif-region, spanning 25 kbp, is located at 1.5 min and includes nine genes that are essential for nitrogen fixation: nifH, nifDK, nifEN, nifUS, nifV, and nifM (27, 31). The minor nif-region of 7 kbp (at 57.8 min) includes three essential genes: nifA, nifB, and nifQ (32). The established or proposed functions of the 12 essential and other accessory nif genes are described in Table 4.1. The rnf1 genes, the products of which have been associated with accumulation of nitrogenase Fe protein (24), are located upstream of the minor nif gene region. The absence of molybdenum triggers expression of the vnf genes coding for the V-dependent nitrogenase components (vnfH, vnfDGK) (33) and their corresponding regulatory and assembly systems (29). The vnf genes are located in a single 20 kbp region of the chromosome, at 2.8 min (Fig. 4.2). The vnf locus includes genes coding for the transcription regulator VnfA

Genomic Analysis of Nitrogen Fixation

55

Table 4.1 A. vinelandii genes involved in Mo-dependent nitrogen fixation Gene

Protein name

Function in nitrogen fixation

nifH

Fe protein, nitrogenase reductase

Provides electrons to nitrogenase and ATP-derived energy for catalysis

nifD

α-chain of nitrogenase or MoFe protein

Catalyzes reduction of nitrogen to ammonia

nifK

β-chain of nitrogenase or MoFe protein

nifT

Unknown

nifY

Putative intermediate carrier of FeMo-cofactor

nifE

FeMo-cofactor assembly protein

nifN Putative intermediate carrier of FeMo-cofactor

nifX FeSII

Shethna protein

Provides respiratory protection

iscANif

IscANif

Involved in Fe–S cluster assembly protein

nifU

Fe–S cluster assembly scaffold protein

Assembles Fe–S clusters for nitrogen-fixing proteins

nifS

Cysteine desulfurase

Catalyzes transfer of sulfur from cysteine to NifU for Fe–S cluster formation

nifV

Homocitrate synthase

Required in biosynthesis of homocitrate moiety of FeMo-cofactor

cysE1

Serine O-acetyl transferase

Involved in biosynthesis of cysteine, which is used as a sulfur source of Fe–S cluster biosynthesis

nifW

Involved for biogenesis of functional MoFe protein

nifZ

Involved in biogenesis of functional MoFe protein

nifM

Peptidyl/prolyl cis/trans isomerase

Required for biogenesis of functional NifH

clpX

Clp protease

Subunit for ATP-dependent protease. Function in nitrogen fixation is unknown.

nifF

Flavodoxin

Involved in transfer of electrons to nitrogenase

nifL

Negative transcription regulatory element

nifA

Positive transcription regulatory element

nifB

Involved in biosynthesis of NifB-cofactor, a FeMo-cofactor precursor.

fdx

Ferredoxin

nifO

Thioredoxin-like protein. Function in nitrogen fixation is unknown.

nifQ

Molybdenum chaperone

rhdNif

Putative rhodanese

Function in nitrogen fixation is unknown

grx5Nif

Glutaredoxin

Function in nitrogen fixation is unknown

56

O’Carroll and Dos Santos

(34, 35) and for an incomplete assembly system (vnfEN, vnfX, vnfY, and vnfU) (14, 36, 37). In the absence of both molybdenum and vanadium, the anf genes are expressed leading to nitrogen fixation that uses the Fe-only nitrogenase. The anf region is the smallest of the three gene clusters spanning a 9 kbp fragment located at 55.5 min and codes for the Fe-only nitrogenase components (AnfH and AnfDGK) (30, 38), the transcription regulator (AnfA) (35, 39), the putative assembly protein (AnfU) (18), and two proteins of unknown function (AnfR and AnfO) (40). In A. vinelandii, the biosynthesis of active V- and Fe-only nitrogenases also requires expression of Nif components such as NifU, NifS, NifV, NifM, and NifB (28, 41, 42), which provide the basic building blocks for the Anf and Vnf nitrogenase cofactors (Fig. 4.2). Promoter analysis indicated that nif, vnf, and anf gene regions are expressed under the control of sigma 54-dependent promoters indicated with arrows in Fig. 4.2 (18). The nif genes are known to be regulated by NifA, a sigma 54-dependent transcriptional activator, and its associated regulatory protein NifL (43). In this system, NifL can sense availability of fixed nitrogen by interacting with GlnK. In the absence of fixed nitrogen the NifL–NifA complex dissociates and NifA acts as an activator of nif genes (44). NifL also regulates NifA in response to the redox status to prevent nif gene expression in the presence of damaging level of oxygen. Interestingly, genes coding for proteins with sequence similarity to NifA are found in the vnf and anf gene regions and were termed vnfA and anfA accordingly. They function in a way analogous to NifA in the activation of V- and Fe-only nitrogen fixation (45). In addition, genome scanning also identified gene regions containing sigma 54 promoters to include the aforementioned rnf1 genes, genes involved in the metabolism of molybdenum (mod1 and mod3 gene regions), and TonB-type siderophore receptors (18). Although not experimentally demonstrated, it is possible that these gene regions are also regulated by NifA or one of its paralogs. Identification of gene regions coregulated by sigma 54 and NifA facilitates the identification of additional genes involved in nitrogen fixation in A. vinelandii and other diazotrophs with sequenced genomes. 2.2. Accessory Genes Involved or Suspected to Be Involved in N2 Fixation

The two nif gene clusters as well as the vnf and anf gene sets comprise the main body of genes utilized during nitrogen fixation. In many nitrogen fixers, however, there are genes outside the main nif, anf, and vnf clusters that also contribute to nitrogen fixation. Many of these accessory genes are involved in Mo (46), Fe (47), and V transport. Others are involved in respiratory protection, i.e., keeping a microaerobic environment that is safe for the oxygen-sensitive nitrogenase enzymes. Recent sequencing and comprehensive annotation of the A. vinelandii genome

Genomic Analysis of Nitrogen Fixation

57

Table 4.2 A. vinelandii accessory genes involved or suspected to be involved in nitrogen fixation Gene/Operon

Genomic location (min)

Reference

rnfABCDGE (rnf1)

57.8

(24)

rnfABCDGE (rnf2)

21.5

(24)

rhdnif

57.9

(18)

grxnif

57.9

(18)

cydAB I

22.1

(26)

cydR

22.1

(25)

(nifD-like)

45.2

(18)

nifA2

30.3

(18)

vnfA2

38.3

(18)

vnfA3

53.5

(18)

modG, modEA1B1C1

57.5

(70)

1.4

(18)

(mod-like)

57.5

(18)

modA3

57.5

(18)

FeSII

1.7

(49)

Ndh

12.9

(71)

modA2B2C2

revealed several accessory genes that are involved or suspected to be involved in N2 fixation (Table 4.2) (18). Some of these were previously unknown and are located inside or outside the major and minor nif regions. The rnf operon encodes a set of membrane-bound proteins that facilitate electron transfer. There are two types of Rnf proteins, one of which, type I is found primarily in diazotrophs (24). In the case of the diazotroph Rhodobacter capsulatus, it has been suggested that electrons from the Rnf complex are transferred to a ferredoxin and eventually to nitrogenase (48). Type II Rnf proteins are found in non-diazotrophs, although in the case of A. vinelandii both types are present. Inactivation of A. vinelandii rnf1 has a significant impact on diazotrophic growth, whereas type II has a smaller but additive effect (24). Inactivation of the Rnf proteins was shown to result in the accumulation of Fe–S cluster-deficient (apo) NifH (24). The FeSII protein (also referred to as the Shethna protein) contributes to conformational protection of nitrogenase when damaging levels of oxygen are present. When oxidized this protein forms a complex with nitrogenase components that inactivates the enzyme but prevents its degradation (49). The rhdNif and grxNif genes are located in the

58

O’Carroll and Dos Santos

smaller nif gene cluster that codes for the regulatory genes nifL and nifA. RhdNif is a putative rhodanese enzyme based on its similarity to RhdA, a protein involved in sulfur transfer (18). GrxNif is a putative monothiol glutaredoxin and its inactivation has no detectable growth defect during diazotrophy but it does result in a twofold decrease in nitrogenase activity (18). The cydAB I genes code for the cytochrome bd oxidase, a protein that efficiently consumes oxygen to provide oxidative stress protection during nitrogen-fixing conditions (26). The genes are regulated by the CydR protein and their inactivation in A. vinelandii results in defective diazotrophic growth under aerobic conditions (25). Sequencing of the A. vinelandii genome revealed several previously unknown genes. A gene similar to nifD that is not part of the nif, vnf or anf clusters has been identified (Fig. 4.1). Genes similar to nifH or nifK were not found nearby and inactivation of the nifD-like gene did not affect diazotrophic growth in A. vinelandii (18). The regulatory NifA and VnfA proteins also seem to have paralogs of significant sequence similarity in A. vinelandii termed NifA2 and VnfA2 (18). The functions of these additional proteins have not yet been elucidated. Sequencing of the A. vinelandii genome also revealed an extraordinarily high number of genes involved in Fe transport. More than 30 genes similar to TonB-dependent siderophore receptor genes were identified as well as six TonB/ExbB/ExbD systems. This is in contrast to the two TonB-like systems present in Pseudomonas aeruginosa and one in Escherichia coli. During diazotrophy, nitrogenase is abundantly expressed, making up ∼10% of the total protein produced in A. vinelandii (17). Whether or not the observed multiple copies of Fe-transport genes serve to satisfy the high requirement for iron during nitrogen fixation remains to be determined.

3. Modifying Genomes to Study Nitrogen Fixation

The genetic malleability of A. vinelandii has made this organism an excellent model for genetic and in vivo biochemical studies. This bacterium can naturally take-up plasmid DNA when cells are starved for Mo and Fe, does not easily support autonomous replication of plasmids, and utilizes its efficient recombination machinery to alter its genome (17, 50). A DNA fragment of interest can be inserted directly into the A. vinelandii genome by double reciprocal recombination (51). Cells can be transformed with plasmids that contain the region of interest flanked by stretches that are identical to A. vinelandii genomic regions (52). These stretches need to be at least 200 bp long for effi-

Genomic Analysis of Nitrogen Fixation

59

cient recombination with an ideal length being 500–1,000 bp. When recombination occurs between both regions present on the plasmid and the genome, the region of interest gets incorporated into the genome. The presence of an antibiotic resistance cassette on the plasmid is used as a screening tool for double recombinants, i.e., the cassette is located outside the recombination region such that a single recombinant will be resistant to the antibiotic whereas a double recombinant will not. 3.1. Examples of Genomic Modification to Study Nitrogen Fixation

The double recombination method has been used successfully to introduce such genomic alterations as (i) gene inactivation by insertion of an antibiotic resistance cassette (53), (ii) gene deletion (54), (iii) amino acid substitution (52), (iv) gene fusion (55, 56), (v) replacement of genomic regions with self or foreign regions (57), and (vi) insertion of protein tags such as histidine tags (58–60). Such genetic manipulations have proved extremely valuable for genetic and biochemical studies that have elucidated the complex process of nitrogenase biosynthesis and metallocluster assembly. Nitrogenase metalloclusters require a consortium of proteins at various stages of their biosynthesis and for their incorporation into nitrogenase and other proteins (61, 62). These metallocluster biosynthetic proteins are uniquely present in nitrogen-fixing organisms and therefore, simple heterologous expression in E. coli, a non-diazotroph, cannot be used for the assembly of nitrogen fixation proteins that harbor metalloclusters. Additionally, metallocluster assembly and transfer also requires low cytoplasmic O2 levels to prevent their degradation, a situation that is achieved in A. vinelandii by a consortium of respiratory protection proteins (63). The genetic malleability of A. vinelandii and the relative abundance of nitrogen fixation proteins produced in the absence of a fixed nitrogen source have allowed the study of these proteins in their native form. The nitrogenase and nitrogenase reductase proteins (encoded by nifDK and nifH, respectively) are naturally produced in abundant levels during nitrogen-fixing conditions. Other proteins of interest, however, are not expressed at convenient levels. Goodwin et al. exploited the strength of the nitrogenase nifH promoter to express and purify the native NifEN proteins. By deleting the nifHDKTYorf3orf4 region, the nifEN genes were relocated such that they could be transcriptionally controlled by the strong nifH promoter resulting in higher protein yields (tenfold) (64). The same strain was used later to isolate a NifEN complex harboring a FeMo-cofactor precursor (65, 66). For direct comparison of proteins with and without metalloclusters, NifEN has also been purified from a strain in which NifB is deleted. NifB plays a role in an earlier stage of FeMo-cofactor biosynthesis and thus its inactivation results in a precursor-less NifEN complex. In another instance, elevated

60

O’Carroll and Dos Santos

expression of the NifB protein was accomplished by removing nifB from its endogenous locus and transferring it to the nifHDK locus such that it was co-expressed with the nifHDK genes at levels that were five times higher compared to endogenous nifB expression (67). 3.2. Using Polyhistidine Tags to Purify Proteins of Interest from Their Native Host

The use of a histidine tag has been favored for purification of many proteins because it reduces the number of purification steps, thereby increasing protein yields and decreasing the chance of losing unstable cofactors that might be bound to the protein. Insertion of the tag directly into the genome allows isolation of proteins from their native host and subsequently the biochemical characterization of the protein as produced by the cell in vivo. This is important for proteins that might undergo conformational rearrangements during the process that they are facilitating. The polyhistidine tag also allows purification of proteins that are produced at relatively low levels. In the case of nitrogen fixation, the biosynthesis of the FeMo-cofactor requires the participation of a number of protein complexes that synthesize and transfer immature and mature forms of the cofactor from one biosynthetic site to another. Comprehensive understanding of the mechanistic details of this biosynthetic pathway in vivo demands the isolation of protein intermediates from their natural environment. Insertion of the polyhistidine tag into the genomic locus of the protein enables biochemists and biophysicists to capture these intermediates and use them to generate or support models that describe FeMo-cofactor biogenesis. The polyhistidine tag is small enough that in most cases it does not interfere with protein activity. In the examples of NifEN and NifB, placement of the histidine tag at the N-terminus of NifE did not affect diazotrophic growth compared to the wildtype strain. However, in other cases, the presence of a histidine tag severely affects growth in A. vinelandii, for example when positioned at the N-terminus of NifU or its homolog IscU, a NifU homolog (60). Therefore, the location of the tag can affect protein activity in some cases and the effect of the tag on the function of the protein should be tested on a case by case basis before protein isolation. The number of histidines inserted at the ends of proteins seems to vary without a significant effect on protein activity in vivo. A range of 7–10 histidine residues has been employed to purify NifEN (64), apo-MoFe protein (58), and NifB (67).

3.3. Exploiting Remote Genomic Promoters to Control Protein Expression

Researchers have utilized remote genomic promoters that can be switched on and off in order to conditionally control gene expression. Such methods have allowed functional analysis of essential genes, increased expression of proteins of interest, as well as functional analysis of genes under various environmental conditions.

Genomic Analysis of Nitrogen Fixation

61

In A. vinelandii, one of the most genetically versatile nitrogenfixing organisms, controlled expression of nitrogen-fixing genes has been achieved in two different ways which are described briefly below and in more detail in reports published in the laboratory of Dennis Dean (55, 60, 68, 69). One way to control gene expression utilizes the genomic locus dedicated to sucrose metabolism. The genes involved in this pathway are clustered in an operon and are controlled by the regulatory protein ScrR encoded by an adjacent gene. Transcriptional and translational control occurs in a way similar to the lacI paradigm. In the presence of sucrose, ScrR binds its inducer and promotes expression of the genes involved in its catabolism. In the absence of sucrose, ScrR acts as a negative regulator of the sucrose catabolic genes in order to avoid futile production of their respective proteins (55). Since the scr gene region is not essential for sucrose metabolism in A. vinelandii, Johnson et al. took advantage of this system and replaced the sucrose catabolic gene scrX, which is negatively regulated by ScrR, with their genes of interest, namely the contiguous nifU and nifS genes which are involved in the biosynthesis of metalloclusters for nitrogenase and other nitrogen-fixing proteins (Fig. 4.3) (68). This allowed the expression of a duplicate copy of the nifUS genes in the presence of sucrose and turned off its expression in the presence of a different primary carbon source. This approach was used to provide proof that nifUS genes expressed in a remote location by non-endogenous regulatory elements can functionally replace the nifUS genes expressed from their natural transcriptional and translational elements. However, the nifUS genes controlled by scrR regulatory elements were not able to functionally replace their iscU and iscS counterparts, which are involved in the housekeeping pathway for iron–sulfur cluster biosynthesis, indicating lack of functional cross-talk between these two homologous processes (55).

Fig. 4.3. Diagrammatic representation of remote loci for incorporating genes or expression units into the A. vinelandii genome. (a) Schematic representation of homologous recombination into the scr region for sucrose-dependent expression of targeted genes as reported by Johnson et al. (55). (b) Schematic representation of homologous recombination into the acx gene region for arabinose-dependent expression of targeted genes as reported by Dos Santos et al. (69).

62

O’Carroll and Dos Santos

Another method for controlling gene expression using remote genomic loci has been described by Dos Santos and coworkers. In this case, nif genes were placed under control of the regulatory elements of arabinose catabolism. Arabinose is a 5-carbon sugar that binds to the regulatory protein AraC. The latter is a repressor protein that changes conformation upon binding arabinose, an event that prevents AraC from blocking gene expression resulting in activation of transcription. The arabinose catabolic genes (araBAD) and their expression and regulatory elements are not endogenous to A. vinelandii but are found in a number of other organisms. Using the convenient genetic recombination system of A. vinelandii, Dos Santos et al. placed the arabinose regulatory elements from E. coli in the acetone carboxylase locus in A. vinelandii (Fig. 4.3) (69). Inactivation of the acetone carboxylase region does not affect diazotrophic or non-diazotrophic growth of A. vinelandii, making this region an excellent candidate for the placement of foreign elements. Then, the nifUS genes were placed under control of the arabinose transcriptional and translational control elements, and it was shown that the NifUS proteins are expressed at elevated levels compared to NifUS either produced by endogenous expression or controlled by the sucrose regulatory elements (68). Elevated expression of the NifUS proteins allowed re-examination of the functional cross-talk between NifUS and IscUS, and it was shown that elevated NifU and NifS controlled by the arabinose promoter are able to functionally replace IscU but not IscS (69). A notable aspect of the two systems described above is their wide biotechnological applicability. The sucrose and arabinose regulatory elements can be used to control expression of any protein including foreign proteins from other organisms. These methods can be particularly useful for proteins that harbor oxygen-sensitive cofactors such as nitrogenase cofactors from other diazotrophs, Fe–S clusters, hydrogenase cofactors, and other Fe-containing cofactors. These regulatory systems can also be used for functional cross-talk studies that assess the ability of protein homologs from other organisms to functionally replace their A. vinelandii equivalents. References 1. Postgate JR (1982) The nitrogen cycle. Philos Trans R Soc Lond B 296:375–385 2. Hill S, Kennedy C, Kavanagh E et al (1981) Nitrogen fixation gene (nifL) involved in oxygen regulation of nitrogenase synthesis in Klebsiella pneumoniae. Nature 290:424–426 3. Weidner S, Puhler A, Kuster H (2003) Genomics insights into symbiotic nitrogen fixation. Curr Opin Biotechnol 14:200–205

4. Seefeldt LC, Hoffman BM, Dean DR (2009) Mechanism of Mo-dependent nitrogenase. Annu Rev Biochem 78:701–722 5. Ribbe M, Gadkari D, Meyer O (1997) N2 fixation by Streptomyces thermoautotrophicus involves a molybdenum- dinitrogenase and a manganese-superoxide oxidoreductase that couple N2 reduction to the oxidation of superoxide produced from O2 by

Genomic Analysis of Nitrogen Fixation

6.

7.

8.

9.

10. 11.

12.

13.

14.

15.

16.

17.

a molybdenum-CO dehydrogenase. J Biol Chem 272:26627–26633 Cummings SP, Gyaneshwar P, Vinuesa P et al (2009) Nodulation of Sesbania species by Rhizobium (Agrobacterium) strain IRBG74 and other rhizobia. Environ Microbiol 11:2510–2525 Mohamed NM, Colman AS, Tal Y et al (2008) Diversity and expression of nitrogen fixation genes in bacterial symbionts of marine sponges. Environ Microbiol 10:2910–2921 Zehr JP, Jenkins BD, Short SM et al (2003) Nitrogenase gene diversity and microbial community structure: a cross-system comparison. Environ Microbiol 5:539–554 Dedysh SN, Ricke P, Liesack W (2004) NifH and NifD phylogenies: an evolutionary basis for understanding nitrogen fixation capabilities of methanotrophic bacteria. Microbiology 150:1301–1313 Raymond J, Siefert JL, Staples CR et al (2004) The natural history of nitrogen fixation. Mol Biol Evol 21:541–554 Henson BJ, Watson LE, Barnum SR (2004) The evolutionary history of nitrogen fixation, as assessed by NifD. J Mol Evol 58: 390–399 Betancourt DA, Loveless TM, Brown JW et al (2008) Characterization of diazotrophs containing Mo-independent nitrogenases, isolated from diverse natural environments. Appl Environ Microbiol 74:3471–3480 Brigle KE, Weiss CM, Newton WE et al (1987) Products of the iron-molybdenum cofactor-specific biosynthetic genes, nifE and nifN, are structurally homologous to the products of the nitrogenase molybdenumiron protein genes, nifH and nifK. J Bacteriol 169:1547–1553 Wolfinger ED, Bishop PE (1991) Nucleotide sequence and mutational analysis of the vnfENX region of Azotobacter vinelandii. J Bacteriol 173:7565–7572 Watzlich D, Brocker MJ, Uliczka F et al (2009) Chimeric nitrogenase-like enzymes of (bacterio)chlorophyll biosynthesis. J Biol Chem 284:15530–15540 Yamamoto H, Kurumiya S, Ohashi R et al (2009) Oxygen sensitivity of a nitrogenaselike protochlorophyllide reductase from the cyanobacterium Leptolyngbya boryana. Plant Cell Physiol 50:1663–1673 Kennedy C, Rudnick P, MacDonald ML et al (2005) Genus III. Azotobacter Beijerinck 1901, 567al. In: Brenner DJ, Noel RK, Staley JT, Garrity GM (eds) Bergey’s Manual of Systematic Bacteriology—The Proteobacteria, pp. 384–402. Springer, New York, NY

63

18. Setubal JC, Dos Santos PC, Goldman BS et al (2009) Genome sequence of Azotobacter vinelandii, an obligate aerobe specialized to support diverse anaerobic metabolic processes. J Bacteriol 191:4534–4545 19. Yan Y, Yang J, Dou Y et al (2008) Nitrogen fixation island and rhizosphere competence traits in the genome of root-associated Pseudomonas stutzeri A1501. Proc Natl Acad Sci USA 105:7564–7569 20. Lalucat J, Bennasar A, Bosch R et al (2006) Biology of Pseudomonas stutzeri. Microbiol Mol Biol Rev 70:510–547 21. Desnoues N, Lin M, Guo X et al (2003) Nitrogen fixation genetics and regulation in a Pseudomonas stutzeri strain associated with rice. Microbiology 149:2251–2262 22. Dingler C, Kuhla J, Wassink H et al (1988) Levels and activities of nitrogenase proteins in Azotobacter vinelandii grown at different dissolved oxygen concentrations. J Bacteriol 170:2148–2152 23. Oelze J (2000) Respiratory protection of nitrogenase in Azotobacter species: is a widely held hypothesis unequivocally supported by experimental evidence? FEMS Microbiol Rev 24:321–333 24. Curatti L, Brown CS, Ludden PW et al (2005) Genes required for rapid expression of nitrogenase activity in Azotobacter vinelandii. Proc Natl Acad Sci USA 102:6291–6296 25. Wu G, Hill S, Kelly MJ et al (1997) The cydR gene product, required for regulation of cytochrome bd expression in the obligate aerobe Azotobacter vinelandii, is an Fnr- like protein. Microbiology 143:2197–2207 26. Kelly MJ, Poole RK, Yates MG et al (1990) Cloning and mutagenesis of genes encoding the cytochrome bd terminal oxidase complex in Azotobacter vinelandii: mutants deficient in the cytochrome d complex are unable to fix nitrogen in air. J Bacteriol 172:6010–6019 27. Jacobson MR, Brigle KE, Bennett LT et al (1989) Physical and genetic map of the major nif gene cluster from Azotobacter vinelandii. J Bacteriol 171:1017–1027 28. Joerger RD, Bishop PE (1988) Nucleotide sequence and genetic analysis of the nifBnifQ region from Azotobacter vinelandii. J Bacteriol 170:1475–1487 29. Joerger RD, Loveless TM, Pau RN et al (1990) Nucleotide-sequences and mutational analysis of the structural genes for nitrogenase-2 of Azotobacter vinelandii. J Bacteriol 172:3400–3408 30. Joerger RD, Jacobson MR, Premakumar R et al (1989) Nucleotide-sequence and mutational analysis of the structural genes

64

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

O’Carroll and Dos Santos (anfhdgk) for the 2nd alternative nitrogenase from Azotobacter vinelandii. J Bacteriol 171:1075–1086 Jacobson MR, Cash VL, Weiss MC et al (1989) Biochemical and genetic analysis of the nifUSVWZM cluster from Azotobacter vinelandii. Mol Gen Genet 219:49–57 Rodriguez-Quinones F, Bosch R, Imperial J (1993) Expression of the nifBfdxNnifOQ region of Azotobacter vinelandii and its role in nitrogenase activity. J Bacteriol 175: 2926–2935 Lee CC, Hu Y, Ribbe MW (2009) Unique features of the nitrogenase VFe protein from Azotobacter vinelandii. Proc Natl Acad Sci USA 106:9209–9214 Woodley P, Buck M, Kennedy C (1996) Identification of sequences important for recognition of vnf genes by the VnfA transcriptional activator in Azotobacter vinelandii. FEMS Microbiol Lett 135: 213–221 Walmsley J, Toukdarian A, Kennedy C (1994) The role of regulatory genes nifA, vnfA, anfA, nfrX, ntrC, and rpoN in expression of genes encoding the three nitrogenases of Azotobacter vinelandii. Arch Microbiol 162:422–429 Ruttimann-Johnson C, Rubio LM, Dean DR et al (2003) VnfY is required for full activity of the vanadium-containing dinitrogenase in Azotobacter vinelandii. J Bacteriol 185:2383–2386 Ruttimann-Johnson C, Staples CR, Rangaraj P et al (1999) A vanadium and iron cluster accumulates on VnfX during iron-vanadiumcofactor synthesis for the vanadium nitrogenase in Azotobacter vinelandii. J Biol Chem 274:18087–18092 Pau RN, Eldridge ME, Lowe DJ et al (1993) Molybdenum-independent nitrogenases of Azotobacter vinelandii: a functional species of alternative nitrogenase-3 isolated from a molybdenum-tolerant strain contains an iron-molybdenum cofactor. Biochemistry 293:101–107 Premakumar R, Loveless TM, Bishop PE (1994) Effect of amino acid substitutions in a potential metal-binding site of AnfA on expression from the anfH promoter in Azotobacter vinelandii. J Bacteriol 176: 6139–6142 Mylona PV, Premakumar R, Pau RN et al (1996) Characteristics of orf1 and orf2 in the anfHDGK genomic region encoding nitrogenase 3 of Azotobacter vinelandii. J Bacteriol 178:204–208 Kennedy C, Dean D (1992) The nifU, nifS and nifV gene products are required for

42.

43.

44.

45.

46.

47.

48.

49.

50. 51.

52.

53.

activity of all three nitrogenases of Azotobacter vinelandii. Mol Gen Genet 231:494–498 Lei S, Pulakat L, Gavini N (1999) Regulated expression of the nifM of Azotobacter vinelandii in response to molybdenum and vanadium supplements in Burk’s nitrogenfree growth medium. Biochem Biophys Res Commun 264:186–190 Dixon R (1998) The oxygen-responsive NIFL-NIFA complex: A novel twocomponent regulatory system controlling nitrogenase synthesis in gamma- proteobacteria. Arch Microbiol 169:371–380 Martinez-Argudo I, Little R, Shearer N et al (2004) The NifL-NifA system: A multidomain transcriptional regulatory complex that integrates environmental signals. J Bacteriol 186:601–610 Premakumar R, Jacobson MR, Loveless TM et al (1992) Characterization of transcripts expressed from nitrogenase-3 structural genes of Azotobacter vinelandii. Can J Microbiol 38:929–936 Luque F, Mitchenall LA, Chapman M et al (1993) Characterization of genes involved in molybdenum transport in Azotobacter vinelandii. Mol Microbiol 7:447–459 Knosp O, von Tigerstrom M, Page WJ (1984) Siderophore-mediated uptake of iron in Azotobacter vinelandii. J Bacteriol 159:341–347 Schmehl M, Jahn A, Meyer zu Vilsendorf A et al (1993) Identification of a new class of nitrogen fixation genes in Rhodobacter capsulatus: a putative membrane complex involved in electron transport to nitrogenase. Mol Gen Genet 241:602–615 Moshiri F, Kim JW, Fu C et al (1994) The FeSII protein of Azotobacter vinelandii is not essential for aerobic nitrogen fixation, but confers significant protection to oxygenmediated inactivation of nitrogenase in vitro and in vivo. Mol Microbiol 14:101–114 Page W, Von Tigerstrom M (1979) Optimal conditions for transformation of Azotobacter vinelandii. J Bacteriol 139:1058–1061 Bishop PE, Premakumar R, Dean DR et al (1986) Nitrogen fixation by Azotobacter vinelandii strains having deletions in structural genes for nitrogenase. Science 232: 92–94 Brigle KE, Setterquist RA, Dean DR (1987) Site-directed mutagenesis of the nitrogenase MoFe protein of Azotobacter vinelandii. Proc Natl Acad Sci USA 84:7066–7069 Morgan TV, Lundell DJ, Burgess BK (1988) Azotobacter vinelandii ferredoxin I: cloning, sequencing, and mutant analysis. J Biol Chem 263:1370–1375

Genomic Analysis of Nitrogen Fixation 54. Robinson AC, Burgess BK, Dean DR (1986) Activity, reconstitution, and accumulation of nitrogenase components in Azotobacter vinelandii mutant strains containing defined deletions within the nitrogenase structural gene-cluster. J Bacteriol 166:180–186 55. Johnson DC, Unciuleac MC, Dean DR (2006) Controlled expression and functional analysis of iron-sulfur cluster biosynthetic components within Azotobacter vinelandii. J Bacteriol 188:7551–7561 56. Suh MH, Pulakat L, Gavini N (2003) Functional expression of a fusion-dimeric MoFe protein of nitrogenase in Azotobacter vinelandii. J Biol Chem 278:5353–5360 57. Wang SZ, Dean DR, Chen JS et al (1991) The N-terminal and C-terminal portions of NifV are encoded by two different genes in Clostridium pasteurianum. J Bacteriol 173:3041–3046 58. Christiansen J, Goodwin PJ, Lanzilotta WN et al (1998) Catalytic and biophysical properties of a nitrogenase apo-MoFe protein produced by a nifB-deletion mutant of Azotobacter vinelandii. Biochemistry 37: 12611–12623 59. Hu Y, Corbett MC, Fay AW et al (2006) FeMo cofactor maturation on NifEN. Proc Natl Acad Sci USA 103:17119–17124 60. Raulfs EC, O‘Carroll IP, Dos Santos PC et al (2008) In vivo iron-sulfur cluster formation. Proc Natl Acad Sci USA 105:8591–8596 61. Rubio LM, Ludden PW (2008) Biosynthesis of the iron-molybdenum cofactor of nitrogenase. Annu Rev Microbiol 62:93–111 62. Hu Y, Fay AW, Lee CC et al (2008) Assembly of nitrogenase MoFe protein. Biochemistry 47:3973–3981 63. Jones CW, Brice JM, Wright V et al (1973) Respiratory protection of nitrogenase in Azotobacter vinelandii. FEBS Lett 29:77–81

65

64. Goodwin PJ, Agar JN, Roll JT et al (1998) The Azotobacter vinelandii NifEN complex contains two identical [4Fe-4S] clusters. Biochemistry 37:10420–10428 65. Hu Y, Fay AW, Ribbe MW (2005) Identification of a nitrogenase FeMo cofactor precursor on NifEN complex. Proc Natl Acad Sci USA 102:3236–3241 66. Soboh B, Igarashi RY, Hernandez JA et al (2006) Purification of a nifEN protein complex that contains bound Mo and a FeMo-co precursor from an Azotobacter vinelandii delta nifHDK strain. J Biol Chem 281:36701–36709 67. Curatti L, Ludden PW, Rubio LM (2006) NifB-dependent in vitro synthesis of the ironmolybdenum cofactor of nitrogenase. Proc Natl Acad Sci USA 103:5297–5301 68. Johnson DC, Dos Santos PC, Dean DR (2005) NifU and NifS are required for the maturation of nitrogenase and cannot replace the function of isc-gene products in Azotobacter vinelandii. Biochem Soc Trans 33: 90–93 69. Dos Santos PC, Johnson DC, Ragle BE et al (2007) Controlled expression of nif and isc iron-sulfur protein maturation components reveals target specificity and limited functional replacement between the two systems. J Bacteriol 189:2854–2862 70. Mouncey NJ, Mitchenall LA, Pau RN (1995) Mutational analysis of genes of the mod locus involved in molybdenum transport, homeostasis and processing in Azotobacter vinelandii. J Bacteriol 177:5294–5302 71. Bertsova YV, Bogachev AV, Skulachev VP (2001) Noncoupled NADH:ubiquinone oxidoreductase of Azotobacter vinelandii is required for diazotrophic growth at high oxygen concentrations. J Bacteriol 183:6869–6874

Chapter 5 Enzymatic Systems with Homology to Nitrogenase Jürgen Moser and Markus J. Bröcker Abstract Nitrogenase-like dark operative protochlorophyllide oxidoreductase (DPOR) is involved in the twoelectron reduction of protochlorophyllide to form chlorophyllide during chlorophyll biosynthesis. Formation of bacteriochlorophyll additionally requires a structurally related enzyme system which is termed chlorophyllide oxidoreductase (COR). During DPOR catalysis, the homodimeric subunit ChlL2 transfers electrons to the corresponding heterotetrameric catalytic subunit (ChlN/ChlB)2 . Analogously, subunit BchX2 of the COR enzymes delivers electrons to subunit (BchY/BchZ)2 . The ChlL2 protein is a dynamic switch protein triggering the ATP-dependent transfer of electrons via a [4Fe–4S] cluster onto a second [4Fe–4S] cluster located on subunit (ChlN/ChlB)2 . This initial electron transfer step of DPOR catalysis clearly resembles nitrogenase catalysis. However, the subsequent substrate reduction process was proposed to be unrelated since no molybdenum-containing cofactor or a P-cluster equivalent is employed. To investigate the transient interaction of both subcomplexes ChlL2 and (ChlN/ChlB)2 and the resulting electron transfer processes, the ternary DPOR enzyme holocomplex was trapped as an octameric (ChlN/ChlB)2 (ChlL2 )2 complex after incubation with non-hydrolyzable ATP analogs. Electron paramagnetic resonance spectroscopic experiments of various DPOR complexes in combination with circular dichroism spectroscopic experiments of the ChlL2 protein revealed a detailed redox catalytic cycle for nucleotide-dependent DPOR catalysis. Key words: Dark operative protochlorophyllide oxidoreductase (DPOR), chlorophyllide oxidoreductase (COR), chlorophyll biosynthesis, nitrogenase-like enzyme.

1. Introduction The sophisticated biochemistry used in nitrogen fixation is of relevance for a second fundamental process on earth, photosynthesis. The biosynthesis of chlorophylls and bacteriochlorophylls, a process yielding more than 6 billion tons of these abundant

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_5, © Springer Science+Business Media, LLC 2011

67

68

a

Moser and Bröcker

b

Fig. 5.1. Comparison of the three-subunit enzymes DPOR, COR, and nitrogenase. (a) During chlorophyll and bacteriochlorophyll biosynthesis ring D of protochlorophyllide (Pchlide) is stereospecifically reduced by the nitrogenase-like enzyme DPOR (composed of subunits ChlL, ChlN, and ChlB in chlorophyll synthesizing organisms and BchL, BchN, and BchB in bacteriochlorophyll synthesizing organisms) leading to chlorophyllide (Chlide). Synthesis of bacteriochlorophylls additionally requires the stereospecific B ring reduction by a second nitrogenase-like enzyme called COR (comprising subunits BchX, BchY, and BchZ) resulting in the formation of bacteriochlorophyllide (Bchlide). (b) The homologous nitrogenase complex is composed of subunits NifH, NifD, and NifK. The identical oligomeric architecture of the related three-subunit enzymes is indicated. Specific ring positions are designated according to the IUPAC nomenclature. R is either a vinyl or an ethyl moiety.

organic pigments annually, includes two enzyme systems sharing homology to nitrogenase (1). For the synthesis of chlorophylls and bacteriochlorophylls, the stereospecific reduction of the C17–C18 double bond of ring D of protochlorophyllide (Pchlide) catalyzed by dark operative protochlorophyllide oxidoreductase (DPOR) results in the formation of chlorophyllide (Chlide) (Fig. 5.1a, left) (2). A second reduction step at ring B (C7–C8) unique to the synthesis of bacteriochlorophylls converts Chlide into bacteriochlorophyllide (Bchlide) (Fig. 5.1a, right). Both of those two systems catalyze the chemically difficult two-electron reduction of the conjugated tetrapyrrole ring system. For this purpose a homodimeric subcomplex sharing a high degree of amino acid sequence identity to the Fe protein of nitrogenase (approximately 30%) is able to transfer electrons onto a second heterotetrameric subcomplex which only shares amino acid sequence identity values of approximately 15% when compared to the related subunits NifD and NifK of nitrogenase (3, 4). To mediate electron transfer and substrate reduction both subcomplexes of DPOR and COR carry redox-active [4Fe–4S] clusters. A comparison of the three-subunit enzymes DPOR, COR, and nitrogenase and the related substrates is given in Fig. 5.1a, b.

Enzymatic Systems with Homology to Nitrogenase

2. Dark Operative Protochlorophyllide Oxidoreductase (DPOR)

2.1. DPOR Subcomplex ChlL2 /BchL2 Mechanistically Resembles Nitrogenase Catalysis

69

Reduction of Pchlide is a central step in the biosynthesis of (bacterio)chlorophylls (5) and can be catalyzed by two evolutionarily unrelated enzymes. Monomeric, light-dependent Pchlide oxidoreductase (POR; NADPH Pchlide oxidoreductase) drives the NADPH-dependent reduction of Pchlide bound in the active site via absorption of light energy (6–8). This light dependency of POR catalysis prevents angiosperms from synthesizing chlorophyll in the dark (9, 10). The alternative nitrogenase-like Pchlidereducing system is termed the light-independent, dark operative Pchlide oxidoreductase (DPOR), which consumes ATP to drive the formation of Chlide (11–13). In chlorophyll synthesizing organisms, DPOR is encoded by the chlN, chlB, and chlL genes (3, 14, 15). The corresponding genes for bacteriochlorophyll synthesizing organisms are known as bchN, bchB, and bchL (2, 14). On the basis of a mutagenesis study in combination with iron determinations and activity assays, two cysteinyl ligands for a redox-active [4Fe–4S] center were proposed. These two residues can be found highly conserved in all BchL/ChlL sequences at an identical position as the respective cysteinyl ligands responsible for the formation of the [4Fe–4S] cluster located on NifH2 . Based on EPR studies it was concluded that residues Cys65 and Cys158 in Prochlorococcus marinus ChlL2 symmetrically coordinate an inter-subunit [4Fe–4S] cluster. Only recently these results were confirmed, when the three-dimensional structure of the BchL2 protein from Rhodobacter sphaeroides was found to share an overall structural similarity with the related nitrogenase protein (16). During DPOR catalysis the [4Fe–4S] cluster of BchL2 is reduced by a “plant type” [2Fe–2S] ferredoxin in vivo. However, for the in vitro investigation of DPOR, dithionite can be used as an artificial electron donor (2, 17) as in the standard nitrogenase activity assay (see Chapter 9). In analogy to nitrogenase, ChlL/BchL proteins contain a highly conserved ATP cofactor binding motif called P-loop Y8 GKGGIGK15 and the so-called switch II region L126 GDVVCGGF134 (numbering from Chlorobaculum tepidum, formerly denoted as Chlorobium tepidum). The switch II sequence motive of nitrogenase was shown to conformationally relay ATP binding to the [4Fe–4S] cluster inter alia by repositioning a single cluster ligand of NifH2 . Mutation of residues Lys10 and Leu126 of BchL from C. tepidum,

70

Moser and Bröcker

which corresponds to residues Lys15 (P-loop) and Leu127 (switch II motive) of nitrogenase (Azotobacter vinelandii numbering), abolished DPOR activity. ATP hydrolysis by ChlL/BchL facilitates electron transfer from the inter-subunit [4Fe–4S] cluster of dimeric subunit ChlL2 /BchL2 onto a second [4Fe–4S] cluster located on the heterotetrameric complex (ChlN/ChlB)2 /(BchN/BchB)2 . Therefore a transient interaction of both subcomplexes was demonstrated. Biotin label transfer experiments revealed the interaction of ChlL2 with both subunits ChlN and ChlB of the heterotetrameric subunit from P. marinus (18). In a related mutagenesis study, involvement of the surface exposed residues Leu70 , Val107 , and Lys109 of ChlN; Gly66 and Gln101 of ChlB; and Tyr127 of ChlL was proposed (18). Only recently the ternary DPOR enzyme holocomplex comprising subunits ChlN, ChlB, and ChlL was trapped as an octameric (ChlN/ChlB)2 (ChlL2 )2 complex after incubation with the nonhydrolyzable ATP analogs adenosine-5 (γ-thio)-triphosphate, adenosine-5 (βγ-imido)-triphosphate, or MgADP in combination with AlF− 4 , respectively. Additionally, a mutant ChlL2 protein, with a deleted Leu153 in the switch II region, allowed for the formation of a stable octameric complex (19) (Fig. 5.2). Furthermore, it was shown that complex formation is efficiently promoted in the presence of the substrate Pchlide. Electron paramagnetic resonance spectroscopy (EPR) of such ternary DPOR complexes after a dithionite reduction step revealed a reduced [4Fe–4S]+ cluster located on ChlL2 , whereas the [4Fe–4S] cluster located on (ChlN/ChlB)2 was found in a non-reduced, EPR silent state. As the employed ATP analog (MgADP-AlF− 4 ) is representing the transition state of ATP hydrolysis, it was concluded that complete ATP hydrolysis is a prerequisite for the inter-subunit electron transfer. Circular dichroism (CD) spectroscopic experiments indicated nucleotide-dependent conformational changes for ChlL2 after ATP binding. From these data a nucleotide-dependent switch mechanism for ChlL2 was deduced which triggers ternary DPOR complex formation and the related electron transfer (19). 2.2. Substrate Recognition and Reduction by DPOR is Unrelated to Nitrogenase Catalysis

The subsequent [Fe–S] cluster-dependent catalysis and the specific substrate recognition at the active site located on subunit (ChlN/ChlB)2 /(BchN/BchB)2 clearly differ from nitrogenase catalysis. The second [Fe–S] cluster of DPOR has no equivalent in nitrogenase which carries the [8Fe–7S] P-cluster and the [1Mo–7Fe–9S–1X–1homocitrate] (FeMoco) metallocenter instead (20, 21). Site-directed mutagenesis experiments for subunits BchN and BchB in combination with kinetic measurements

Enzymatic Systems with Homology to Nitrogenase

ATP

71

ATP ChlL

MgADP-AlF4–

ChlL MgADP-AlF4–

[4Fe-4S] [4Fe 4S]

ADP

ADP

ChlN

ChlB [4Fe-4S]

Pchlide

Pchlide ChlN

ChlB [4Fe-4S]

ATP MgADP-AlF4– ADP

ChlL

ChlL [4Fe-4S]

ATP MgADP-AlF4– ADP

Fig. 5.2. Ternary protein–protein interaction of DPOR subcomplexes. Dimeric ChlL2 is a dynamic switch protein triggering the ATP-dependent protein–protein interaction with the heterotetrameric (ChlN/ChlB)2 protein. ChlL2 and (ChlN/ChlB)2 both carry redoxactive [4Fe–4S] clusters responsible for the electron transfer onto the Pchlide substrate. The transient interaction of both subcomplexes was trapped in the presence of the ATP analog MgADP-AlF− 4 and Pchlide resulting in the formation of a stable octameric protein complex.

revealed the presence of four cysteine residues crucial for DPOR catalysis. Analysis of DPOR variants indicated a significant decrease of the overall iron and sulfur content for (BchN/BchB)2 modified in residues Cys21 , Cys46 , and Cys103 of BchN and residue Cys94 of BchB (C. tepidum numbering) (2). Subsequently, the postulated [4Fe–4S] cluster located on (ChlN/ChlB)2 /(BchN/BchB)2 was confirmed by EPR spectroscopy (22). Using spectroscopic and genetic approaches the involvement of an additional molybdenumcontaining cofactor or of a P-cluster equivalent was clearly ruled out and a direct electron transfer from the [4Fe–4S] cluster of (ChlN/ChlB)2 /(BchN/BchB)2 onto the substrate was proposed. Therefore the active site might be located in close proximity to the [4Fe–4S] cluster of (ChlN/ChlB)2 /(BchN/BchB)2 . The overall [Fe–S] cluster composition of the individual DPOR components in the octameric DPOR complex is summarized in Fig. 5.2. For the analysis of DPOR substrate recognition, 11 synthetic substrate derivatives with altered substituents on the

72

Moser and Bröcker

four pyrrole rings A–D and the isocyclic ring E plus eight chlorophyll biosynthetic intermediates were tested as DPOR substrates. Although DPOR tolerated minor modifications of the ring substituents on rings A–C, the catalytic target ring D was apparently found to be coordinated with high specificity. The 8-vinyl variant of Pchlide has been shown to be a biosynthetic intermediate when chlorophylls and bacteriochlorophylls are synthesized. In accordance to this, the [8-vinyl]-derivative of Pchlide was efficiently used as a DPOR substrate which is in agreement with the described, variable routes for (Baterio)chlorophyll biosynthesis. According to this both possible substituents at the C8 position of the substrate are indicated as R in Fig. 5.1a. 2.3. ATPase Activity of DPOR

For the understanding of the catalytic mechanism of DPOR, it is essential to characterize the ATPase properties of ChlL2 /BchL2 . For Pchlide, ATP, and the artificial reducing agent dithionite, Chlide formation followed Michaelis–Menten kinetics. Apparent Km values for Pchlide (6.1 μM), for ATP (13.5 μM), and for dithionite (52.7 μM) were determined for the C. tepidum enzyme (2). Taking into account the thermodynamic properties of Pchlide reduction the consumption of 4 mol of ATP per mol reduced Pchlide was proposed. This amount of ATP correlates with the dimeric composition of ChlL2 containing two ATP binding sites and a single inter-subunit [Fe–S] cluster (2). Only recently, new experiments allowed for the approximation of the stoichiometry of ATP hydrolysis versus Pchlide reduction in the employed in vitro test system. For this purpose free phosphate liberated during DPOR catalysis was detected as dark-green malachite metallophosphate complex. Interestingly, DPOR activity assays which were incubated in the absence of the Pchlide substrate also revealed significant ATPase activity (approximately 60%). It was concluded that this non-productive ATPase activity might not be relevant under in vivo conditions. However, this type of experiment indicated that the experimentally observed ratio of 14 mol of ATP hydrolyzed per mol of Chlide formation does not reflect optimal in vivo conditions. Analogously, for the related nitrogenase system an in vitro consumption of up to 36 ATP per reduced N2 versus an apparent minimal theoretical consumption of 16 ATP was observed (23–25). According to these observations, it was concluded that the minimal ATP consumption of DPOR catalysis might be in the range of 4 mol of ATP per mol of Chlide under in vivo conditions (19).

2.4. The Proposed Catalytic Redox Cycle of DPOR

The described experimental data suggested that ChlL2 containing two ATP binding sites is the functional electron donor leading to a catalytically active (ChlN/ChlB)2 protein. Furthermore, it was shown that the docking of ChlL2 to (ChlN/ChlB)2 is

Enzymatic Systems with Homology to Nitrogenase

73

a prerequisite for the observed ATPase activity. Obviously, this protein–protein interaction is important for the ATP hydrolyzing activity of ChlL2 . Additionally, this dynamic subunit interplay is strongly influenced by the presence of the substrate Pchlide. A high affinity of the substrate Pchlide for the (ChlN/ChlB)2 complex was experimentally demonstrated (22, 26). Since maximum DPOR ATPase activity was dependent on the presence of Pchlide and subcomplex (ChlN/ChlB)2 , we postulate that Pchlide binding is the initial step of DPOR catalysis. Subsequently, the substrate-bound complex (ChlN/ChlB)2 efficiently stimulates the ATPase activity of ChlL2 , which is required for substrate reduction. Such tight binding of the substrate is also beneficial in vivo to overcome the highly phototoxic nature of this hydrophobic molecule. Such harmful photodynamic properties of Pchlide have only recently been demonstrated (27). Various states of the DPOR catalytic cycle have been characterized with biochemical and biophysical methods (19). In the proposed redox cycle (Fig. 5.3) these individual states are highlighted (bold). The single electron reduction of ChlL2 is enabled in the absence of (ChlN/ChlB)2 (22). Therefore, it was concluded that during the initial step of DPOR catalysis the natural electron donor, which is a ferredoxin (2), transfers an electron onto the [4Fe–4S] cluster of the dimeric subunit ChlL2 (I). Since it was shown that the presence of ATP is not a prerequisite for effective dithionite reduction it was concluded that the (ChlN/ChlB)2[2+] Pchlide

ATP

ChlL2[1+]

ChlL2[1+] II

III

ATP

Fd red

[ChlL2[1+] (ChlN/ChlB)2[2+]] ATP

I Fd ox

Pchlide

[ChlL Leu1532[1+] (ChlN/ChlB)2[2+]] Pchlide

[ChlL2[2+]]

[ChlL2[1+] (ChlN/ChlB)2[2+]] MgADP-AlF4– Pchlide

[ChlL2[2+]

V

ADP

ADP ADP

IV

[ChlL2[2+](ChlN/ChlB)2[1+]]

(ChlN/ChlB)2[1+]

Pchlide Pi

Pchlide

(ChlN/ChlB)2[1+]

Fig. 5.3. Proposed redox cycle of DPOR catalysis. Schematic model for the electron transfer processes and dynamic subunit interaction during ATP-driven DPOR catalysis. Five intermediates were confirmed by EPR spectroscopy (bold), the individual redox state is indicated [1+] for reduced and [2+] for oxidized [4Fe–4S] clusters. According to this redox cycle two consecutive single electron reductions of (ChlN/ChlB)2 are required to provide the two electrons necessary for Pchlide reduction.

74

Moser and Bröcker

reduction of the [4Fe–4S] cluster of ChlL2 precedes nucleotide binding. Subsequently, binding of ATP molecules (II) results in slight conformational alterations, as indicated by CD spectroscopy data. However, these structural changes do not influence the [4Fe–4S]1+ cluster geometry of ChlL2 as described for the related nitrogenase protein. Binding of the substrate Pchlide on subunit (ChlN/ChlB)2 is then required for transient ternary complex formation (III). ATP hydrolysis by the ChlL2 subunits then facilitates the tight protein–protein interaction of ChlL2 and (ChlN/ChlB)2 and the simultaneous electron transfer from the [4Fe–4S] cluster of ChlL2 onto the [4Fe– 4S] cluster located on (ChlN/ChlB)2 . This transient complex has been characterized with two independent techniques. The octameric [(ChlL2 [1+] )2 (ChlN/ChlB)2 [2+] MgADP-AlF− 4 ] com[1+]

(ChlN/ChlB)2 [2+] plex and the octameric (ChlL2 Leu153 )2 complex represent a situation where the reduced component ChlL2 is close to the electron transfer onto (ChlN/ChlB)2 . EPR spectroscopy revealed that the transition state of ATP hydrolysis is not accompanied by a reduced [4Fe–4S] cluster in (ChlN/ChlB)2 or a partially reduced substrate. Therefore, complete ATP hydrolysis is required to obtain a [4Fe–4S]1+ cluster on (ChlN/ChlB)2 (IV) which then has the ability to transfer a single electron onto Pchlide. The presence of ADP then leads to the dissociation of the ternary complex, resulting in the liberation of ChlL2 in the [4Fe–4S]2+ state and (ChlN/ChlB)2 in the [4Fe–4S]1+ state (V). The dynamic switch protein ChlL2 has to trigger two rounds of this redox catalytic cycle in order to supply the two electrons necessary for the reduction of the substrate. These individual electron transfer processes are completed by the stereospecific addition of two protons on the C17 and C18 of Pchlide in the active site of the DPOR enzyme. Either protonated functional groups or precisely orientated water molecules might be the source for those two protons (19).

3. COR Catalysis Is Closely Related to DPOR

COR is the second nitrogenase-like enzyme which is responsible for the biosynthesis of all bacteriochlorophyll molecules. For this purpose COR catalyzes the reduction of ring B of Chlide to form Bchlide (Fig. 5.1a, right) (28). Therefore, accurate discrimination of the ring systems B and D is required. The individual subunits of COR share an overall amino acid sequence identity of 15– 22% for subunits BchY and BchZ and 31–35% for subunit BchX when compared to the corresponding DPOR subunits (BchN,

Enzymatic Systems with Homology to Nitrogenase

75

BchB, and BchL). In amino acid sequence alignments of BchX proteins with the closely related ChlL/BchL subunits of DPOR, both cysteinyl ligands responsible for the formation of the redoxactive inter-subunit [4Fe–4S] cluster are highly conserved (3). EPR experiments revealed a characteristic signal for a [4Fe–4S] cluster located on BchX2 (29). Furthermore, all residues involved in ATP binding and hydrolysis can be found at an identical position in all sequences of BchX proteins as for the related DPOR protein (18). These results argue for an identical catalytic mechanism for the BchX2 protein from COR. This hypothesis was recently confirmed experimentally when a chimeric DPOR enzyme was reconstituted. When the subunit ChlL/BchL in the standard DPOR assay was substituted with the BchX2 subunit of COR in the presence of Pchlide as a substrate, significant DPOR activity was observed. These results indicated that not only the ATP-driven electron transfer processes of BchX2 proteins but also the dynamic protein–protein interaction of subcomplexes BchX2 with (BchY/BchZ)2 resembles DPOR catalysis. Interestingly, the related nitrogenase subunit NifH2 does not have the ability to substitute for the ChlL/BchL subunit of DPOR, which might indicate that the chlorophyll biosynthetic reductases have evolved significantly from the nitrogenase system (18). Also, with respect to subcomplex (BchY/BchZ)2 it becomes evident that DPOR and COR have only evolved sparingly during the evolution of nitrogenase-like enzymes. Three cysteine ligands with relevance to the formation of the catalytic [4Fe–4S] cluster on (ChlN/ChlB)2 /(BchN/BchB)2 have a direct counterpart in the sequences of BchY of COR. However, theoretical analysis failed to give clear evidence for an additional ligand located on BchZ. Nevertheless, the postulated redox-active [4Fe–4S] cluster of (BchY/BchZ)2 was confirmed using EPR spectroscopic experiments (18). Based on these results one might conclude a conserved catalytic redox cycle for DPOR and COR. However, the orientation of the substrate bound in the active site of (BchY/BchZ)2 might differ substantially.

4. NflD and NflH: Crucial for Metabolism of Methanogens?

Genome analysis revealed the presence of NifH or NifD/NifK homologs sharing an atypical sequence of both nifH-like and nifD/nifK-like genes scattered among all methanogens and some phototrophs (30). These genes are distinct from nitrogenase homologs that allow nitrogen fixation in several methanogens. From phylogenetic analysis it was concluded that those nif-like sequences nflH and nflD (nfl stands for “nif-like”) lie basal in

76

Moser and Bröcker

the tree between the nitrogenase and the (bacterio)chlorophyll (DPOR/COR) clades (30). Amino acid sequence alignments of NflH with NifH and ChlL/BchL revealed a conserved P-loop ATP binding motive and two invariant cysteine residues suggesting an analogous inter-subunit [4Fe–4S] cluster, located at the interface of a dimeric reductase protein. In contrast to this, NflD was found to vary significantly from NifD and ChlN/BchN. None of the residues responsible for the ligation of the MoFecofactor (FeMoco) were found conserved. However, several amino acid residues responsible for the formation of the P-cluster in nitrogenase are conserved, which might indicate the presence of a less complex [Fe–S] cluster located on NflD. Methanocaldococcus jannaschii is a hyperthermophilic methanogen, a member of the Euryarchaeota, which does not perform photosynthesis or nitrogen fixation. Therefore, this organism is a good candidate for the elucidation of the function of nfl genes. It was shown that nflH and nflD are constitutively expressed independent of the availability of nitrogen. Furthermore, no involvement in the detoxification of compounds such as cyanide or azide was observed. With two-hybrid studies the interaction of NflH and NflD was shown. Based on theoretical considerations and the described initial experiments involvement of NflH/NflD in the biosynthesis of cofactor F430 was speculated. All known methanogenic archaea rely on methyl-coenzyme M reductase for their terminal step of methanogenesis. This enzymatic step is dependent on the nickel-containing heterocyclic cofactor F430 . In the present working hypothesis, NflD and NflH might be involved in one (or several) ring reduction step(s) responsible for cofactor F430 biosynthesis (31).

References 1. Field CB, Behrenfeld MJ, Randerson JT et al (1998) Primary production of the biosphere: integrating terrestrial and oceanic components. Science 281:237–240 2. Bröcker MJ, Virus S, Ganskow S et al (2008) ATP-driven reduction by darkoperative protochlorophyllide oxidoreductase from chlorobium tepidum mechanistically resembles nitrogenase catalysis. J Biol Chem 283:10559–10567 3. Burke DH, Hearst JE, Sidow A (1993) Early evolution of photosynthesis: clues from nitrogenase and chlorophyll iron proteins. Proc Natl Acad Sci USA 90:7134–7138 4. Fujita Y, Matsumoto H, Takahashi Y et al (1993) Identification of a nifDK-like gene (ORF467) involved in the biosynthesis of

5. 6.

7.

8.

chlorophyll in the cyanobacterium Plectonema boryanum. Plant Cell Physiol 34:305–314 Beale SI (1999) Enzymes of chlorophyll biosynthesis. Photosyn Res 60:43–73 Belyaeva OB, Griffiths WT, Kovalev JV et al (2001) Participation of free radicals in photoreduction of protochlorophyllide to chlorophyllide in an artificial pigmentprotein complex. Biochemistry (Moscow) 66:173–177 Heyes DJ, Hunter CN, van Stokkum IH et al (2003) Ultrafast enzymatic reaction dynamics in protochlorophyllide oxidoreductase. Nat Struct Biol 10:491–492 Heyes DJ, Ruban AV, Wilks HM et al (2002) Enzymology below 200 K: the kinetics

Enzymatic Systems with Homology to Nitrogenase

9.

10.

11.

12. 13. 14. 15.

16.

17.

18.

19.

and thermodynamics of the photochemistry catalyzed by protochlorophyllide oxidoreductase. Proc Natl Acad Sci USA 99: 11145–11150 Rüdiger W (2003) The last steps of chlorophyll biosynthesis. In: Kadish KM, Smith KM, Guilard R (eds) Porphyrin Handbook, Chlorophylls and Bilins: Biosynthesis, Synthesis, and degradation, pp. 71–108. Academic, New York, NY Masuda T, Takamiya K (2004) Novel insights into the enzymology, regulation and physiological functions of lightdependent protochlorophyllide oxidoreductase in angiosperms. Photosyn Res 81:1–29 Apel K (2001) Chlorophyll Biosynthesis – metabolism and strategies of higher plants to avoid photooxidative stress. In: Aro EM, Anderson B (eds) Regulation of Photosynthesis, pp. 235–252. Kluwer, Dordrecht Fujita Y (1996) Protochlorophyllide reduction: a key step in the greening of plants. Plant Cell Physiol 37:411–421 Schoefs B (2001) The protochlorophyllidechlorophyllide cycle. Photosyn Res 70: 257–271 Suzuki JY, Bollivar DW, Bauer CE (1997) Genetic analysis of chlorophyll biosynthesis. Annu Rev Genet 31:61–89 Bollivar DW, Suzuki JY, Beatty JT et al (1994) Directed mutational analysis of bacteriochlorophyll a biosynthesis in Rhodobacter capsulatus. J Mol Biol 237: 622–640 Sarma R, Barney BM, Hamilton TL et al (2008) Crystal structure of the L protein of Rhodobacter sphaeroides light-independent protochlorophyllide reductase with MgADP bound: a homologue of the nitrogenase Fe protein. Biochemistry 47:13004–13015 Fujita Y, Bauer CE (2000) Reconstitution of light-independent protochlorophyllide reductase from purified bchl and BchNBchB subunits. In vitro confirmation of nitrogenase-like features of a bacteriochlorophyll biosynthesis enzyme. J Biol Chem 275:23583–23588 Wätzlich D, Bröcker MJ, Uliczka F et al (2009) Chimeric nitrogenase-like enzymes of (bacterio)chlorophyll biosynthesis. J Biol Chem 284:15530–15540 Bröcker MJ, Waetzlich D, Saggu M et al (2010) Biosynthesis of (bacterio)chlorophylls: ATP-dependent transient subunit interaction and electron transfer of dark operative protochlorophyllide Oxidoreductase. J Biol Chem 285:8268–8277

77

20. Igarashi RY, Seefeldt LC (2003) Nitrogen fixation: the mechanism of the Modependent nitrogenase. Crit Rev Biochem Mol Biol 38:351–384 21. Howard JB, Rees DC (1994) Nitrogenase: a nucleotide-dependent molecular switch. Annu Rev Biochem 63:235–264 22. Bröcker MJ, Wätzlich D, Uliczka F et al (2008) Substrate recognition of nitrogenaselike dark operative protochlorophyllide oxidoreductase from Prochlorococcus marinus. J Biol Chem 283:29873–29881 23. Rainbird RM, Hitz WD, Hardy RW (1984) Experimental determination of the respiration associated with soybean/rhizobium nitrogenase function, nodule maintenance, and total nodule nitrogen fixation. Plant Physiol 75:49–53 24. Rees DC, Howard JB (2000) Nitrogenase: standing at the crossroads. Curr Opin Chem Biol 4:559–566 25. Erickson JA, Nyborg AC, Johnson JL et al (1999) Enhanced efficiency of ATP hydrolysis during nitrogenase catalysis utilizing reductants that form the all-ferrous redox state of the Fe protein. Biochemistry 38:14279–14285 26. Nomata J, Ogawa T, Kitashima M et al (2008) NB-protein (BchN-BchB) of dark operative protochlorophyllide reductase is the catalytic component containing oxygentolerant Fe-S clusters. FEBS Lett 582: 1346–1350 27. Walther J, Bröcker MJ, Wätzlich D et al (2009) Protochlorophyllide: a new photosensitizer for the photodynamic inactivation of Gram-positive and Gram-negative bacteria. FEMS Microbiol Lett 290:156–163 28. Nomata J, Kitashima M, Inoue K et al (2006) Nitrogenase Fe protein-like Fe-S cluster is conserved in L-protein (BchL) of dark-operative protochlorophyllide reductase from Rhodobacter capsulatus. FEBS Lett 580:6151–6154 29. Kim EJ, Kim JS, Lee IH et al (2008) Superoxide generation by chlorophyllide a reductase of Rhodobacter sphaeroides. J Biol Chem 283:3718–3730 30. Raymond J, Siefert JL, Staples CR et al (2004) The natural history of nitrogen fixation. Mol Biol Evol 21:541–554 31. Staples CR, Lahiri S, Raymond J et al (2007) Expression and association of group IV nitrogenase NifD and NifH homologs in the non-nitrogen-fixing Archaeon Methanocaldococcus jannaschii. J Bacteriol 189: 7392–7398

Section II Genetic and Biochemical Methods

Chapter 6 Molecular Biology and Genetic Engineering in Nitrogen Fixation Patricia C. Dos Santos Abstract Biological nitrogen fixation is a complex and tightly regulated process limited to a group of prokaryotic species known as diazotrophs. Among well-studied diazotrophs, Azotobacter vinelandii is the best studied for its convenience of aerobic growth, its high levels of nitrogenase expression, and its genetic tractability. This chapter includes protocols and strategies in the molecular biology and genetic engineering of A. vinelandii that have been used as valuable tools for advancing studies on the biosynthesis, mechanism, and regulation of nitrogen fixation. Key words: Competency, transformation, selection, rescue, congression, Azotobacter vinelandii.

1. Introduction Nitrogen fixation, the biological conversion of nitrogen gas to ammonia, is catalyzed by the complex metalloenzyme nitrogenase. The major and best studied catalyst in this process is the two-component Mo-dependent nitrogenase encoded by the genes nifH and nifDK (1, 2). The biosynthesis and activation of nitrogenase involves the participation of many components in a laborious and intricate pathway (2, 3). In diazotrophs, the nitrogenase genes are often located in genomic regions along with other genes, the products of which are involved in the synthesis of metalloclusters and the activation of nitrogenase (4). At least eight nif genes are strictly required for the synthesis of active nitrogenase: nifE, nifN, nifS, nifU, nifV, nifM, nifB, and nifQ (4–6). This complexity is one factor that prohibits the use of traditional molecular biology techniques and popular systems for M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_6, © Springer Science+Business Media, LLC 2011

81

82

Dos Santos

heterologous protein expression, such as the use of expression vectors in Escherichia coli. As a result, in order to conduct genetic and biochemical studies with active enzyme forms, nitrogenase and its biosynthetic enzymes need to be isolated from their native nitrogen-fixing host. Historically, the well-studied diazotrophs Klebsiella pneumoniae (facultative anaerobe) (7–10), Clostridium pasteurianum (anaerobe) (11), and Azotobacter vinelandii (strict aerobe) (1, 12, 13) have been used as sources of nitrogenase for studies of biosynthesis, structure, and mechanism. Among these three model systems, A. vinelandii offers special experimental advantages because of the convenience of its aerobic metabolism and ability to grow in the absence of special nutrients or growth conditions. Despite the fact that nitrogen fixation is an oxygensensitive process, high aeration rates lead to higher levels of cellular nitrogen fixation in A. vinelandii (14). In addition, the genetic tractability of this bacterium has allowed the development of many genetic and molecular biology tools for uncovering the biosynthetic scheme of nitrogenase-associated cofactors and for gaining insight into the mechanistic features of nitrogen reduction. This chapter provides a description of methods that combine genetics and molecular biology to study nitrogen fixation in A. vinelandii.

2. Materials 1. Burk’s medium (B medium)—prepared, autoclaved, and stored as two separate solutions: phosphate buffer (0.2 g KH2 PO4 , 0.8 g K2 HPO4 , and dH2 O up to 900 mL) and 10× salts (200 g of sucrose – table sugar, 2.0 g MgSO4 ·7H2 O, 0.9 g of CaCl2 ·2H2 O, 1 mL of 10 mM Na2 MoO4 ·H2 O, 50 mg of FeSO4 ·7H2 O, and dH2 O up to 1 L). Every liter of B medium is freshly prepared with 900 mL of phosphate buffer and 100 mL of 10× salts solution. When preparing B medium plates, 16 g of agar (Difco) is added to phosphate buffer prior sterilization. 2. Competent medium (B-Mo-Fe medium)—same as B medium but lacks FeSO4 ·7H2 O and Na2 MoO4 ·H2 O in its 10× salts solution. 3. Transformation buffer—filter-sterilized 20 mM 3-(Nmorpholino)propanesulfonic acid (MOPS) buffer, pH 7.2, with 20 mM MgCl2 . 4. Storage buffer—sterile phosphate buffer (0.2 g KH2 PO4 , 0.8 g K2 HPO4 , and dH2 O up to 900 mL) containing 1% v/v DMSO.

Molecular Biology and Genetic Engineering in Nitrogen Fixation

83

Table 6.1 Antibiotics commonly used in direct selection with A. vinelandii Antibiotic

Concentration

Reference

Ampicillin

50–100 μg/mL

(35, 46)

Carbenicillin

50 μg/mL

(47)

Gentamicin

0.05 μg/mL

(35)

Kanamycin

0.5–3 μg/mL

(48)

Rifampicin

5–10 μg/mL

(35, 46)

Spectinomycin

25 μg/mL

(48)

Streptomycin

0.1 μg/mL

(35)

Tetracycline

10 μg/mL

(49)

5. Antibiotic solutions—filter-sterilized 500–20,000× stock solutions are prepared and stored at –20◦ C, with exception of kanamycin that is stored in the fridge. Table 6.1 lists some antibiotics used as well as their optimal concentrations in growth media. Notice that for some antibiotics, the optimal concentration for selection in A. vinelandii is much lower than that used for other bacteria species.

3. Methods 3.1. Culturing A. vinelandii

Growth in liquid culture or on solid media is accomplished at an optimum temperature of 30◦ C using Burk’s medium (B medium) (15–19). Several variations of this recipe are found in the literature; one of these versions is provided above. The efficiency of nitrogen fixation and the bacterial growth rate vary with the carbon source (see Note 1). Interestingly, Chen and collaborators have reported increased nitrogenase activity in crude extracts of cells cultured in the presence of aromatic compounds (20). A. vinelandii grows with a doubling time of about 2 h when cultured in liquid B medium at 30◦ C/300 rpm, and colonies are visible after ~48 h when cultured on solid B medium (18). When sucrose is the sole carbon source, the growth rate remains unchanged in the presence or absence of a fixed nitrogen source (14, 21). Freezer stocks of A. vinelandii strains can be prepared by scraping cells off a 2-day-old B plate and mixing with 1 mL of storage buffer and stored at –80◦ C. Aliquots of a freezer stock can

84

Dos Santos

be obtained with a sterile disposable pipette tip from the freezer stock while it is frozen. In the presence of a fixed nitrogen source, the expression of the nif genes is repressed. When culturing cells under these nondiazotrophic conditions, a final concentration of either 10 mM urea (BU) or 10–30 mM ammonium acetate (BN) is used. Exhaustion or removal of this nutrient from the culture triggers the derepression of the nif genes and the accumulation of nitrogenase. The peak of protein expression is usually within 2–4 h after derepression (22). This physiological response is routinely used to yield high levels of nitrogenase from A. vinelandii cultures or when expressing nitrogen-fixing proteins from strains unable to sustain diazotrophic growth (23, 24). 3.2. A. vinelandii Competency

The exchange of genetic information in A. vinelandii was originally reported by Page and Tigerstrom in 1979 (25) and was first used to study the genetics of nitrogen fixation by Bishop and collaborators in 1985 (26). Most strains of A. vinelandii used in genetic engineering were derived from a non-gummy fluorescent strain resulting from spontaneous mutation, called strain OP by Bush and Winston (27). The OP strain was renamed UW (University of Wisconsin) and CA (North Carolina State University) and isolates derived from it were named MV (mutant vinelandii) and DJ (Deloriah Jacobs) by several investigators after their affiliated institutions or coworkers and these are considered “wildtype” strains by these laboratories (28). The completion of the genome sequencing of the A. vinelandii DJ strain showed a DNA fragment insertion within the coding sequence of the regulator AlgU (29), which explains the non-gummy phenotype of OP and its derivative strains (30). An efficient method for the transformation of A. vinelandii uses cells cultured under iron starvation (31). Under these conditions, A. vinelandii expresses and secretes a fluorescent green siderophore (32) (Fig. 6.1). Although there is experimental evidence linking siderophore production and the ability of the bacterium to take up DNA, the details of this relationship have not yet been identified. The preparation of competent A. vinelandii cells first involves at least one passage of the cells on agar plates containing BN-Mo medium. Each passage requires 2–3 days of incubation at 30◦ C. Competent cultures are then prepared in a 125 mL flasks containing 50 mL of BN-Mo-Fe medium. To inoculate the competent medium, about a quarter of a loop of cells (2 mm2 ) is scraped off a fresh BN-Mo Fe plate, added to the competent medium, and the flask is vortexed for a few seconds to help break up cell clumps. Both the cell inoculums and the dispersion of these cells into the medium are critical for the development of competency. The culture is then incubated at 170 rpm at 30◦ C for 18–20 h (see Note 2). For each

Molecular Biology and Genetic Engineering in Nitrogen Fixation

85

Fig. 6.1. A. vinelandii DJ strains. (a) A. vinelandii strains DJ (right) and DJ1418 (left) grown on a Burk’s medium agar plate containing X-gal after 3-day incubation at 30◦ C. The A. vinelandii DJ1418 {Johnson, 2006 #8878} has been genetically modified to contain the E. coli lacZ gene under the control of the scr-promoter. (b) Sample of A. vinelandii cultured in competent medium (BN-Fe-Mo).

transformation, the donor DNA is mixed with 200 μL of fresh competent cells and 200 μL of transformation buffer. This mixture is incubated for 20 min at room temperature before spreading on a selective solid medium or outgrowing in BN liquid medium for 16–20 h. 3.3. Transformation

The source of genetic material for transformation can vary from a DNA fragment, a circular plasmid, or genomic DNA (31, 33) (see Note 3). To prepare the donor DNA for a transformation experiment, entire genes or fragments of genes are cloned into recombinant plasmids and modified to contain a point mutation, fragment deletion or insertion, and/or antibiotic insertion. In designing such DNA constructs, there should be at least 150–500 bp of flanking sequences on either side of the genetic modification to facilitate recombination (19). About 1–10 μg of the donor DNA can be used in transformation experiments. During the transformation reaction, the competent cells take up the donor DNA and, via a double-reciprocal recombination event at the homologous flanking regions, the genetic modification is incorporated into the genome. The A. vinelandii RecA enzyme is a major contributor toward the high efficiency of recombination. Deletion of this gene greatly hampers the ability of A. vinelandii to modify its genomic information (34, 35).

3.4. Selection

Following recombination, modified cells need to be selected and segregated out of the transformation culture. In a typical transformation experiment only a small fraction of the competent cells will take up the donor DNA and undergo homologous recombination. Several rounds of cell division are necessary to isolate a progeny containing the desired modification. Therefore, identifying an effective selection method is critical for

86

Dos Santos

isolating transformants that contain the desired genetic modification. There are several strategies for the genetic manipulation and selection of modified strains. Commonly used techniques to study nitrogen fixation in A. vinelandii are listed below. 3.4.1. Direct Antibiotic Selection

Antibiotic resistance is used for selection in many transformation strategies in A. vinelandii. Specific genes can be inactivated by the insertion of an antibiotic resistance cartridge into their coding sequences. Genes can also be inactivated with antibiotic resistance cassettes via transposon mutagenesis (Tn5, Tn10, and Tn76) (36, 37), which has been used to generate knockout strains with various auxotrophic phenotypes (38). Direct antibiotic selection can be used as a screening tool for identifying gene modifications, such as point mutations or the insertion of sequences coding for affinity tags, when the site of modification is near the 3 end of the gene and the antibiotic cartridge is inserted between the end of the gene and its downstream region. Finally, direct antibiotic selection can also be used to distinguish between double-reciprocal recombination (when the two flanking regions serve as the sites for recombination) and single-reciprocal recombination (when only one flanking region serves as a site for recombination)—(Diagram 6.1). Double-reciprocal recombination results in the specific incorporation of the desired antibiotic resistance cassette only inside the gene of interest. Singlereciprocal recombination results in the incorporation of the entire plasmid into the chromosome—the original copy of the gene of interest will be intact and functional, a second copy will contain the desired antibiotic resistance cassette, and a second antibiotic resistance cassette will also be incorporated into the chromosome (Diagram 6.1). Several different antibiotic resistance cartridges have been successfully used in the direct selection of gene modifications in A. vinelandii. When using direct antibiotic selection, transformed cells are cultured for 16–20 h at 30◦ C/300 rpm in BN medium. Serial dilutions of this culture are then plated on

Diagram 6.1. Direct selection transformation strategies.

Molecular Biology and Genetic Engineering in Nitrogen Fixation

87

antibiotic-selective media. Isolated colonies are first streaked on single plates of the same selective media for at least two successive passages, then on plates without the antibiotic for two successive passages. Last, the isolates are screened on selective media again to ensure the stability of the newly constructed strain (see Note 4). 3.4.2. Rescue

In the study of nitrogen fixation focusing on the enzymes involved in both nitrogenase biosynthesis and catalysis, a convenient screening strategy exploits the ability of the desired strain to fix nitrogen. In these experiments, the parent strain contains a genetic modification, perhaps a point mutation, deletion, or an antibiotic resistance cassette insertion, that compromises the strain’s ability to grow on B medium. Cells are transformed with a donor DNA whose fragment contains flanking regions homologous to the sequences flanking both sides of the genetic modification of the parent strain genome (Diagram 6.2a). In rescue experiments, the recombination event restores the modified DNA region on the genome back to the original sequence or to an altered sequence that now contains new genetic information that enables nitrogen fixation to occur and growth on B medium can now be detected. This technique has been used in several reports to introduce point mutations or sequence tags that result in amino acid substitutions either in the catalytic nitrogenase components or in one of its essential biosynthetic enzymes (4, 39, 40).

3.4.3. Congression

Introduction of genetic modifications which lead to loss of nitrogen fixation can often be achieved through indirect selection. In this case, transformed cells are plated on medium that does not select for the targeted modification, because the desired phenotype is characterized by the lack of growth on B medium. A particular technical challenge met in this experimental approach

Diagram 6.2. Rescue and congression transformation strategies. Selection of transformed cells is based on the ability (Nif+ ) or disability (Nif– ) of the parent and desired strain to fix nitrogen.

88

Dos Santos

is that only a small fraction of competent cells will receive the donor DNA, which would result in the scoring of thousands of colonies for the identification of the desired strain. This experimental hurdle is circumvented by congression, defined as a coincidental transfer of two unrelated genetic markers. The procedure involves the co-transformation of competent cells with two different sources of DNA, one carrying the desired alteration and the other carrying a selectable marker, such as an antibiotic resistance cartridge. After transformation, cells are cultured in BN medium at 30◦ C/300 rpm for 16–20 h; serial dilutions of this culture are plated on a medium containing the antibiotic. Thus, transformed cells are plated on a medium that selects for cells containing the co-transforming DNA, not on a medium that selects for the desired modification—in many cases the screening phenotype for the desired modification is lack of growth (41). This indirect selection ensures that only cells that were actually competent and able to take up DNA survive. Since the fraction of actual competent cells in the “competent cells” mixture is small, congression significantly reduces the pool of cells that need to be scored for the modification of interest. For example, isolation of a targeted substitution that inactivates nitrogenase uses B medium (scoring medium), on which cells containing the modified genome will not grow and BN-antibiotic medium (selective medium), on which desired cells will grow (Diagram 6.2b) (see Note 5). 3.4.4. Blue and White Screening

Traditionally, the use of the lacZ gene as a reporter in gene expression has been used as a tool to understand the regulation of promoters involved in nitrogen fixation. As A. vinelandii does not encode for β-D-galactosidase, the E. coli lacZ gene can be inserted into the genome to report transcriptional or transcriptional/translational controls (42–44). In these experiments, levels of expression can be visualized on agar plates containing X-Gal (60 μg/mL) (Fig. 6.1) or β-D-galactosidase activity can also be quantified with the Miller assay using A. vinelandii cultures in the presence of o-nitrophenyl galactoside (ONPG) and 0.5% dimethyl formamide. Another use of lacZ in genetic manipulations is as an indicator for the successful exchange of genetic information during transformation. In this case, the parent strain contains a lacZ gene inserted between the sites of recombination, yielding blue colonies when the cells are plated on BN medium plus X-gal. During the exchange of genetic material through homologous recombination, the donor DNA is inserted into the genome replacing lacZ (35, 40). The use of X-gal in the screening of transformed cells allows the visual selection of white colonies (desired isolates) from a blue-colony population lacking the desired substitution (35, 40) (see Note 6).

Molecular Biology and Genetic Engineering in Nitrogen Fixation

3.4.5. Traditional Genetics

89

The sequencing of the A. vinelandii genome (30) and the development of molecular biology techniques for gene manipulation have allowed the development of several reverse genetic approaches, some of which are listed here. These methodologies for generating knockouts and introducing genetic modifications to alter protein structure and sequence have provided an avenue both for the identification of new gene products involved in nitrogen fixation and for the assignment of their specific functions. Nevertheless traditional genetics, in which the phenotype identification precedes the genotype, remains a valuable tool in the discovery of novel gene functions. In this process, the isolation of revertants—mutants coding altered proteins resistant to otherwise inhibitory conditions—provides a powerful approach for uncovering biochemical processes and/or revealing new players involved in these and other metabolic functions. Christiansen et al. isolated an altered protein resistant to acetylene, a potent noncompetitive inhibitor of nitrogen reduction (45). In their experimental approach, acetylene-hypersensitive strains of A. vinelandii were incubated on B medium plates in the presence of the inhibitory, but not explosive, concentrations of acetylene. After prolonged incubation, the appearance of revertant colonies was observed. The revertant genotype was predicted to be the result of a mutation within the nitrogenase coding sequence; this hypothesis was confirmed by sequencing the nitrogenase genes amplified from the genome of the revertant strain. Validation of the genotype associated with this phenotype was further confirmed when a DNA fragment containing the modification was transformed into the original parent strain generating an identical phenotype as that associated with the revertant strain. The isolation of spontaneous mutants does not always involve a predictable genomic target—in most cases, the identification of the genotype associated with the revertant phenotype is still the main obstacle in the functional assignment of genes. Today, with fast and affordable techniques for whole genome sequencing, it is possible to revisit classical molecular biology and genetic engineering techniques to study nitrogen fixation where modified genetic loci can now be rapidly identified by whole genome sequencing of isolated revertant strains.

4. Notes 1. Although sucrose is the most popular carbon source because of its low cost, A. vinelandii can also use a variety of carbon sources such as glucose, acetate, glycerol, and several aromatic and heterocyclic compounds (18).

90

Dos Santos

2. Routinely, when preparing competent cells, the fluorescence color intensity of the culture is used as a visual assessment of competence. 3. Typically, recombinant plasmids easily maintained in E. coli, such as pUC, M13, or the pET series, are not suitable for replication in A. vinelandii (18). 4. The proper selection of candidate strain isolates is an important aspect in isolating strains containing the desired substitution. A. vinelandii accumulates multiple copies of its chromosome during late exponential and stationary growth stages. Inactivation of essential genes through antibiotic cartridge insertion imposes a selective pressure for keeping two distinct genome copies: one containing the active gene of interest and the other containing the antibiotic cartridge. In targeting essential genes, the isolate will lose antibiotic resistance when selective pressure is relieved. For this reason, successive passages of candidate isolates under non-selective media helps segregate the chromosomes and gives a good indication of the homogeneity and genetic stability of the constructed strain. 5. In designing such experiments, it is important to consider that the DNA target containing the antibiotic marker insert should have flanking regions suitable for homologous recombination into the A. vinelandii genome; this insertion should not result in the same phenotype as the desired modification, i.e., when scoring for a desired modification that eliminates nitrogen fixation, the selectable marker should not interrupt any nif gene region essential to this process. 6. When combined with congression, blue–white selection is highly selective and greatly decreases the number of isolates that need to be manually screened. References 1. Seefeldt LC, Hoffman BM, Dean DR (2009) Mechanism of Mo-dependent nitrogenase. Annu Rev Biochem 78:701–722 2. Hu Y, Fay AW, Lee CC et al (2008) Assembly of nitrogenase MoFe protein. Biochemistry 47:3973–3981 3. Rubio LM, Ludden PW (2008) Biosynthesis of the iron-molybdenum cofactor of nitrogenase. Annu Rev Microbiol 62:93–111 4. Jacobson MR, Brigle KE, Bennett LT et al (1989) Physical and genetic map of the major nif gene cluster from Azotobacter vinelandii. J Bacteriol 171:1017–1027 5. Jacobson MR, Cash VL, Weiss MC et al (1989) Biochemical and genetic analysis of

the nifUSVWZM cluster from Azotobacter vinelandii. Mol Gen Genet 219:49–57 6. Imperial J, Ugalde RA, Shah VK et al (1984) Role of the nifQ gene product in the incorporation of molybdenum into nitrogenase in Klebsiella pneumoniae. J Bacteriol 158: 187–194 7. Mayer SM, Lawson DM, Gormal CA et al (1999) New insights into structure-function relationships in nitrogenase: a 1.6 A resolution X-ray crystallographic study of Klebsiella pneumoniae MoFe-protein. J Mol Biol 292:871–891 8. Brill WJ (1980) Biochemical genetics of nitrogen fixation. Microbiol Rev 44:449–467

Molecular Biology and Genetic Engineering in Nitrogen Fixation 9. Thorneley RNF, Eady RR (1973) Nitrogenase of Klebsiella pneumoniae: evidence for an adenosine triphosphate-induced association of the iron-sulphur protein. Biochem J 133:405–408. 10. Thorneley RNF, Lowe DJ (1983) Nitrogenase of Klebsiella pneumoniae – kinetics of the dissociation of oxidized iron protein from molybdenum iron protein – identification of the rate-limiting step for substrate reduction. Biochem J 215:393–403 11. Bolin JT, Ronco AE, Morgan TV et al (1993) The unusual metal clusters of nitrogenase: structural features revealed by x-ray anomalous diffraction studies of the MoFe protein from Clostridium pasteurianum. Proc Natl Acad Sci USA 90:1078–1082 12. Einsle O, Tezcan FA, Andrade SLA et al (2002) Nitrogenase MoFe-protein at 1.16 A resolution: a central ligand in the FeMocofactor. Science 297:1696–1700 13. Schmid B, Ribbe MW, Einsle O et al (2002) Structure of a cofactor-deficient nitrogenase MoFe protein. Science 296:352–356 14. Curatti L, Brown CS, Ludden PW et al (2005) Genes required for rapid expression of nitrogenase activity in Azotobacter vinelandii. Proc Natl Acad Sci USA 102:6291–6296 15. Esposito RG, Wilson PW (1958) Acetate as a calcium-sparing factor in nitrogen fixation by Azotobacter vinelandii. Proc Natl Acad Sci USA 44:472–476 16. Pena C, Campos N, Galindo E (1997) Changes in alginate molecular mass distributions, broth viscosity and morphology of Azotobacter vinelandii cultured in shake flasks. Appl Microbiol Biotechnol 48: 510–515 17. Fallik E, Hartel PG, Robson RL (1993) Presence of a vanadium nitrogenase in Azotobacter paspali. Appl Environ Microbiol 59:1883–1886 18. Kennedy C, Rudnick P, MacDonald ML et al (2005) Genus III. Azotobacter Beijerinck 1901, 567al. In: Brenner DJ, Noel RK, Staley JT, Garrity GM (eds) Bergey’s Manual of Systematic Bacteriology – The Proteobacteria, pp. 384–402. Springer, New York, NY 19. Mayer SM, Dos Santos PC, Seefeldt LC et al (2002) Use of short-chain alkynes to locate the nitrogenase catalytic site. In: Leigh GJ (ed) Nitrogen Fixation at the Millennium, pp. 137–154. Elsevier Science, Brighton, UK 20. Chen YP, Lopezdevictoria G, Lovell CR (1993) Utilization of aromatic-compounds as carbon and energy-sources during growth and N-2-fixation by free-living nitrogenfixing bacteria. Arch Microbiol 159:207–212

91

21. Upchurch RG, Mortenson LE (1980) In vivo energetics and control of nitrogen fixation: changes in the adenylate energy charge and adenosine 5 -diphosphate/adenosine 5 - triphosphate ratio of cells during growth on dinitrogen versus growth on ammonia. J Bacteriol 143:274–284 22. Shah VK, Davis LC, Brill WJ (1972) Nitrogenase. I. Repression and derepression of the iron-molybdenum and iron proteins of nitrogenase in Azotobacter vinelandii. Biochim Biophys Acta 256:498–511 23. Christiansen J, Goodwin PJ, Lanzilotta WN et al (1998) Catalytic and biophysical properties of a nitrogenase apo-MoFe protein produced by a nifB-deletion mutant of Azotobacter vinelandii. Biochemistry 37:12611– 12623 24. Goodwin PJ, Agar JN, Roll JT et al (1998) The Azotobacter vinelandii NifEN complex contains two identical [4Fe-4S] clusters. Biochemistry 37:10420–10428 25. Page WJ, von Tigerstrom M (1979) Optimal conditions for the transformation of Azotobacter vinelandii. J Bacteriol 139:1058–1061 26. Bishop PE, Premakumar R, Dean DR et al (1986) Nitrogen fixation by Azotobacter vinelandii strains having deletions in structural genes for nitrogenase. Science 232: 92–94 27. Bush JA, Wilson PW (1959) A nongummy chromogenic strain of Azotobacter vinelandii. Nature 184:381–384 28. Kennedy C, Bishop PE (2004) Genetics of nitrogen fixation and related aspects of metabolism in species of Azotobacter: history and current status. In: Klipp W, Masepohl B, Gallon JR, Newton WE (eds) Genetics and Regulation of Nitrogen Fixation in Free-Living Bacteria, pp. 27–44. Kluwer, Dordrecht 29. Martinez-Salazar JM, Moreno S, Najera R et al (1996) Characterization of the genes coding for the putative sigma factor AlgU and its regulators MucA, MucB, MucC, and MucD in Azotobacter vinelandii and evaluation of their roles in alginate biosynthesis. J Bacteriol 178:1800–1808 30. Setubal JC, Dos Santos P, Goldman BS et al (2009) Genome sequence of Azotobacter vinelandii, an obligate aerobe specialized to support diverse anaerobic metabolic processes. J Bacteriol 191:4534–4545 31. Page W, von Tigerstrom M (1979) Optimal conditions for transformation of Azotobacter vinelandii. J Bacteriol 139: 1058–1061 32. Page WJ, Grant GA (1987) Effect of mineral iron on the development of transformation

92

33. 34.

35.

36.

37.

38.

39.

40.

41.

Dos Santos competence in Azotobacter vinelandii. Fems Microbiol Lett 41:257–261 Doran JL, Page WJ (1983) Heat sensitivity of Azotobacter vinelandii genetic transformation. J Bacteriol 155:159–168 Venkatesh TV, Reddy MA, Das HK (1990) Cloning and characterization of the Azotobacter vinelandii recA gene and construction of a recA deletion mutant. Mol Gen Genet 224:482–486 Johnson DC, Unciuleac MC, Dean DR (2006) Controlled expression and functional analysis of iron-sulfur cluster biosynthetic components within Azotobacter vinelandii. J Bacteriol 188:7551–7561 Kennedy C, Gamal R, Humphrey R et al (1986) The nifH, nifM and nifN genes of Azotobcater vinelandii: characterisation by Tn5 mutagenesis and isolation from pLAFR1 gene banks. Mol Gen Genet 205: 318–325 Wu G, Hill S, Kelly MJ et al (1997) The cydR gene product, required for regulation of cytochrome bd expression in the obligate aerobe Azotobacter vinelandii, is an Fnr- like protein. Microbiology 143:2197–2207 Contreras A, Maldonado R, Casadesus J (1991) Tn5 mutagenesis and insertion replacement in Azotobacter vinelandii. Plasmid 25:76–80 Brigle KE, Setterquist RA, Dean DR et al (1987) Site-directed mutagenesis of the nitrogenase MoFe protein of Azotobacter vinelandii. Proc Natl Acad Sci USA 84:7066–7069 Dos Santos PC, Johnson DC, Ragle BE et al (2007) Controlled expression of nif and isc iron-sulfur protein maturation components reveals target specificity and limited functional replacement between the two systems. J Bacteriol 189:2854–2862 Robinson AC, Dean DR, Burgess BK (1987) Iron-molybdenum cofactor biosynthesis in Azotobacter vinelandii requires the iron

42.

43.

44. 45.

46.

47.

48.

49.

protein of nitrogenase. J Biol Chem 262: 14327–14332 Walmsley J, Toukdarian A, Kennedy C (1994) The role of regulatory genes nifA, vnfA, anfA, nfrX, ntrC, and rpoN in expression of genes encoding the three nitrogenases of Azotobacter vinelandii. Arch Microbiol 162:422–429 Premakumar R, Loveless TM, Bishop PE (1994) Effect of amino acid substitutions in a potential metal-binding site of AnfA on expression from the anfH promoter in Azotobacter vinelandii. J Bacteriol 176:6139–6142 Miller JH (1972) Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, New York, NY Christiansen J, Cash VL, Seefeldt LC et al (2000) Isolation and characterization of an acetylene-resistant nitrogenase. J Biol Chem 275:11459–11464 Bertsova YV, Bogachev AV, Skulachev VP (2001) Noncoupled NADH:ubiquinone oxidoreductase of Azotobacter vinelandii is required for diazotrophic growth at high oxygen concentrations. J Bacteriol 183:6869–6874 Fang FC, Helinski DR (1991) Broad-hostrange properties of plasmid RK2: importance of overlapping genes encoding the plasmid replication initiation protein TrfA. J Bacteriol 173: 5861–5868 Peralta-Gil M, Segura D, Guzman J et al (2002) Expression of the Azotobacter vinelandii poly-beta-hydroxybutyrate biosynthetic phbBAC operon is driven by two overlapping promoters and is dependent on the transcriptional activator PhbR. J Bacteriol 184:5672–5677 Kelly MJ, Poole RK, Yates MG et al (1990) Cloning and mutagenesis of genes encoding the cytochrome bd terminal oxidase complex in Azotobacter vinelandii: mutants deficient in the cytochrome d complex are unable to fix nitrogen in air. J Bacteriol 172:6010–6019

Chapter 7 Purification of Nitrogenase Proteins Jared A. Wiig, Chi-Chung Lee, Aaron W. Fay, Yilin Hu, and Markus W. Ribbe Abstract Nitrogenase is one of the most complex enzymes known to date. The extensively studied molybdenum nitrogenase consists of two protein components and three metal centers that are critical for nitrogenase activity. The inherent complexity of this enzyme system, which is further compounded by the sensitivity of the metal clusters toward oxygen, makes the large-scale purification of fully active nitrogenase proteins a formidable task. This chapter highlights several methods that have been developed for the purification of nitrogenase proteins over the past few decades. Techniques used include weak anion exchange chromatography, size exclusion chromatography, and immobilized metal affinity chromatography. These methods can be selectively applied to nitrogenase variants and other related proteins. Key words: Anaerobic protein purification, nitrogenase, MoFe protein, Fe protein, VFe protein, weak anion exchange chromatography (WAEC), gel filtration, immobilized metal affinity chromatography (IMAC).

1. Introduction Isolation of active nitrogenase proteins has been one of the major challenges for nitrogenase research (1, 2). Despite major advances in biochemical techniques over the last 50 years, the development of purification strategies of nitrogenase has been hampered by the difficulty in handling this fragile enzyme. Both component proteins of nitrogenase are bridged by metal centers at the subunit interface: the α2 -homodimeric iron (Fe) protein is ligated by a [Fe4 S4 ] cluster between the two subunits; whereas the α2 β2 -heterotetrameric molybdenum-iron (MoFe) protein is bridged by a [Fe7 S8 ] P-cluster between α and β subunits M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_7, © Springer Science+Business Media, LLC 2011

93

94

Wiig et al.

(3, 4). In addition, MoFe protein contains a FeMoco ([MoFe7 S9 X-homocitrate], where X = C, N, or O) in each α subunit. The integrity of all three metal clusters and their proper orientation within the proteins are crucial for the activity of nitrogenase, yet pose challenges for the handling of proteins during the purification process. Harsh treatment or extended purification procedures could lead to alterations of the quaternary structures of nitrogenase proteins and, subsequently, the disintegration of their associated metal clusters. Furthermore, the three metal clusters of nitrogenase are extremely vulnerable toward oxygen exposure and can be deactivated by being exposed to as low as 50 ppm of oxygen for a few minutes. Thus, proper anaerobic techniques have become a critical aspect of nitrogenase purification, which tend to complicate the purification procedures. Finally, most experiments in nitrogenase research demand large amounts of protein (5, 6). For example, a good FeMoco extraction attempt requires at least 500 mg of MoFe protein; an in vitro cofactor maturation experiment takes up to 200 mg of Fe protein; and even a routine activity assay or the preparation of one EPR sample consumes roughly 10–20 mg of MoFe protein or Fe protein (7–9). Clearly, an efficient method for large-scale nitrogenase purification is the prerequisite for the success of nitrogenase research (10). A general protocol has been established and refined for the purification of nitrogenase proteins over the last few decades (11, 12). This protocol primarily utilizes multiple salt gradient steps on a weak anion exchange column (WAEC) to first separate the MoFe protein and the Fe protein into different fractions and then purify each protein individually from these fractions. Such a procedure was originally designed for the purification inside a glove box, but it was soon modified to be performed on a Schlenk line outside the glove box, which then made the upscaling of the protocol possible. To date, this procedure has remained the most commonly used method for the preparation of non-tagged, wild-type MoFe protein and Fe protein. In addition, some of the MoFe protein variants such as the nifB MoFe protein and several point mutants of MoFe protein can also be purified by this procedure (13). Apart from the WAEC protocol, the recent advances in expressing affinity-tagged (i.e., His-tagged) nitrogenase proteins in its native host (i.e., Azotobacter vinelandii) via homologous recombination have facilitated the development of protein purification schemes using immobilized metal affinity chromatography (IMAC). This method greatly reduces the time and steps of purification, thereby significantly improving the stability and activity of the purified proteins (14). Such an approach is now applied to the preparation of the unstable MoFe protein variants or homologues, such as the nifH MoFe protein or the α2 β2 δ4 -octameric VFe protein of the alternative vanadium nitrogenase (15).

Purification of Nitrogenase Proteins

95

2. Materials Unless otherwise specified, all chemicals are obtained from Fisher Scientific (Pittsburgh, PA). 2.1. Cell Growth and Crude Extract Preparation

1. Burk’s nitrogen-free media: 4.6 mM K2 HPO4 , 1.5 mM KH2 PO4 , 0.8 mM MgSO4 ·7 H2 O, 3.4 mM NaCl, 0.3 mM CaSO4 ·2H2 O, 0.01 mM Na2 MoO4 ·2H2 O, 0.05 mM FeSO4 ·7H2 O, and 2% (w/v) sucrose. 2. Cell wash buffer: 0.05 M Tris, pH 7.4.

2.2. Purification of Non-tagged MoFe Protein

1. DEAE sepharose FF (GE, Piscataway, NJ) is suspended in 0.05 M Tris, pH 7.4. 2. Amicon microfiltration concentrators fitted with YM50 membranes (Danver, MA). 3. DEAE equilibration buffer I: 0.025 M Tris, pH 7.4, 0.1 M NaCl, and 1 mM Na2 S2 O4 . 4. Gradient buffers I—separation of MoFe protein and Fe protein: • Buffer A: 0.025 M Tris, pH 7.4, 0.1 M NaCl, and 1 mM Na2 S2 O4 in 1.2 L. • Buffer B: 0.025 M Tris, pH 7.4, 0.5 M NaCl, and 1 mM Na2 S2 O4 in 1.2 L. 5. Gradient buffers II—purification of MoFe protein: • Buffer A: 0.25 M Tris, pH 7.4, 0.1 M NaCl, and 1 mM Na2 S2 O4 in 1.2 L. • Buffer B: 0.25 M Tris, pH 7.4, 0.35 M NaCl, and 1 mM Na2 S2 O4 in 1.2 L. 6. Crystallization buffers: • Buffer A: 0.025 M Tris, pH 7.4, 1 mM Na2 S2 O4 in 1.0 L. • Buffer B: 0.025 M Tris, pH 7.4, 28 mM NaCl, and 1 mM Na2 S2 O4 in 0.05 L. • Buffer C: 0.025 M Tris, pH 7.4, 0.25 M NaCl, and 1 mM Na2 S2 O4 in 0.05 L.

2.3. Purification of Non-tagged Fe Protein

1. Sephacryl S-200 superfine (GE, Piscataway, NJ) is suspended in 0.05 M Tris, pH 7.4, 0.5 M NaCl, and 1 mM Na2 S2 O4 . 2. Sephacryl equilibration buffer: 0.05 M Tris, pH 7.4, 0.5 M NaCl, and 1 mM Na2 S2 O4 . 3. DEAE equilibration buffer II: 25 mM Tris, pH 7.4, 1 mM Na2 S2 O4 . 4. Gradient buffers III—purification of Fe protein:

96

Wiig et al.

• Buffer A: 50 mM Tris, pH 7.4, 0.1 M NaCl, and 1 mM Na2 S2 O4 in 0.35 L. • Buffer B: 50 mM Tris, pH 7.4, 0.35 M NaCl, and 1 mM Na2 S2 O4 in 0.35 L. 5. Elution buffer: 25 mM Tris, pH 7.4, 0.5 M NaCl, and 1 mM Na2 S2 O4 . 2.4. Purification of His-Tagged Nitrogenase Proteins

1. IMAC sepharose 6 FF (GE, Piscataway, NJ) is suspended in 0.025 M Tris, pH 7.9. 2. Zn(II) solution: 0.025 M Tris, pH 7.9, 100 mM ZnSO4 ·6H2 O. 3. Equilibration buffer: 0.025 M Tris, pH 7.9, 0.5 M NaCl. 4. Wash buffer: 0.025 M Tris, pH 7.9, 40 mM imidazole, 0.5 M NaCl. 5. Elution buffer: 0.025 M Tris, pH 7.9, 250 mM imidazole, and 0.5 M NaCl.

3. Methods The purification strategy of the non-tagged nitrogenase proteins is summarized in Fig. 7.1 (see Note 1). All steps following the rupture of the cells should be carried out anaerobically using the Schlenk system (see Note 2). The cell lysate is first fractionated by weak anion exchange chromatography into the “MoFe protein fraction” and the “Fe protein fraction.” The MoFe protein fraction is further purified using a salt gradient, followed by a quick crystallization step where the MoFe protein crystals are washed, desalted, re-dissolved, and concentrated; whereas the Fe protein fraction is further purified by a size exclusion chromatography step, followed by another WAEC procedure. Typically, ∼600 mg of the non-tagged, wild-type MoFe protein and ∼500 mg of the non-tagged, wild-type Fe protein can be purified from ∼800 g of cells. The specific activities of thus purified proteins in H2 formation are ∼3000 nmol/min/mg MoFe protein and ∼2100 nmol/min/mg Fe protein, respectively. The purification strategy for the His-tagged nitrogenase proteins is much simpler. The cell lysate is loaded on a Zn(II)charged IMAC column. Subsequently, the protein-bound column is washed to remove non-specifically associated impurities. Finally, the His-tagged protein is eluted with a buffer containing imidazole. It should be noted that the flow-through collected during the loading process of the crude extract can be saved for further purification of non-tagged nitrogenase components. Typically,

Purification of Nitrogenase Proteins

97

Crude Extract (850 g cells in 50 mM Tris, pH 8.0)

Supernatant

Sediment (discarded)

5’ at 56 ° C

Supernatant

Sediment (discarded)

DEAE-cellulose column (5 x 10 cm) Equilibration: 25 mM Tris, pH 7.4, 100 mM NaCl Gradient: 100–500 mM NaCl (1200 ml each, 400 ml/hours)

Fe protein fraction

MOFe protein fraction

DEAE-cellulose column (5 x 10 cm) DEAE-cellulose concentration column (2.5 x 7 cm) Equilibration: 25 mM Tris, pH 7.4, 100 mM NaCl Equilibration: 50 mM Tris, pH 8.0 Gradient: 100–350 mM NaCl (1200 ml each, 400 ml/hours) Elution: 500 mM NaCl (same buffer)

Sephacryl S-200 SF column (7.5 x 65 cm) Equilibration: 50 mM Tris, pH 8.0, 500 mM NaCl (800 ml/hr)

DEAE-sepharose column (2.5 x 30 cm) Equilibration: 50 mM Tris, pH 8.0, 100 mM NaCl Gradient: 150−400 mM NaCl (350 ml each, 200 ml/hr)

Pure Fe protein

Crystallization & wash

Re-dissolve crystals & concentrate

Pure MoFe protein

Fig. 7.1. Purification scheme of the non-tagged Fe protein and MoFe protein of nitrogenase.

∼900 mg of His-tagged, wild-type MoFe protein can be purified from ∼400 g of cells. The specific activity of thus purified protein is 2400 nmol/min/mg MoFe protein. 3.1. Cell Growth and Crude Extract Preparation

1. Small-scale cultures are grown in 5–10 1 L-Burk’s media supplemented with 0.2 mM ammonia acetate. 2. Large-scale cultures are grown in 100 L or 200 L fermentors, aerated at 40 ft3 /min. 3. Growth is followed by turbidity measurement at 436 nm. The wild-type strains can be harvested at mid-log phase or

98

Wiig et al.

at an OD436 of ∼1.0. The nif-deletion strains can be harvested after ammonia is exhausted or 3 h after the growth plateaus. Cells are harvested by centrifugation at 10,000 × g for 10 min. 4. Harvested cells are cleaned with the cell wash buffer, centrifuged at 10,000 × g for 10 min, and stored at –80◦ C till the time of use. 5. Frozen cells (∼800 g) are re-suspended in cell wash buffer in the ratio of 1 g of cells to 1.4 mL of buffer and degassed (see Note 3) for 1–2 h before they are ruptured in a Manton Gaulin homogenizer at 12,000 lb/in2 (psi). 6. The cell lysate is degassed for 1 h and centrifuged at 14,000 × g for 30 min at 4◦ C. The supernatant is collected and can either be stored at –20◦ C overnight in an air-tight container for later processing or used immediately in the following steps. 3.2. Purification of Non-tagged MoFe Protein

1. The cell lysate is incubated for 1 h at room temperature with 10 μg/mL deoxyribonuclease and then heated to 56◦ C for 5 min with constant stirring. This step precipitates up to 60% of proteins other than MoFe protein and Fe protein out of the cell lysate (see Note 4). 2. The cell lysate is then centrifuged at 70,000 × g for 45 min at 4◦ C to remove the precipitated proteins. 3. The supernatant is collected and loaded onto a DEAE column. The column is about 5 × 10 cm and should be preequilibrated overnight with anaerobic DEAE equilibration buffer I (see Note 5). 4. Following the loading of cell lysate, the column is washed with two to three column volumes of equilibration buffer. 5. Subsequently, the proteins are eluted from the column by a linear NaCl gradient, which is set up as follows: gradient buffer IB is pumped into gradient buffer IA at about 200 mL/h using a peristaltic pump; while gradient buffer IA is loaded onto the column at about 400 mL/h at the same time. Although the MoFe protein and Fe protein are expected to elute at approximately 0.12 M and 0.22 M NaCl in the gradient, the eluate should be monitored further at OD405 for protein content, and fractions containing the MoFe protein and Fe protein should be collected separately. 6. The Fe protein fraction is frozen in liquid nitrogen as protein pellets (see Note 6); whereas the MoFe protein fraction can also be frozen or used immediately in the following steps.

Purification of Nitrogenase Proteins

99

7. The MoFe protein fraction is diluted 1:1 with a buffer containing 0.025 M Tris, pH 7.4. The diluted sample is then loaded on a new 5 × 10 cm DEAE column, which is preequilibrated as described in Step 3. 8. The proteins bound to the column are subsequently eluted with the following NaCl gradient: gradient buffer IIB is pumped into gradient buffer IIA at about 250 mL/h using a peristaltic pump; while gradient buffer II A is loaded on the column at about 500 mL/h at the same time. The eluate is monitored at OD405 and the protein peak that corresponds to the partially purified MoFe protein is collected. 9. The partially pure MoFe protein is then transferred anaerobically to concentrators that are lined with YM50 membranes and pre-washed with crystallization buffer A. This concentration step is carried out with gentle stirring at ∼24 psi under a constant flow of argon, and the final volume of the protein solution should be approximately 5–10 mL. 10. The concentrated MoFe protein solution is then diluted with crystallization buffer A to a final concentration of 0.028 M NaCl (the dilution factor is approximately 4.4). 11. The MoFe protein crystals are subsequently formed upon incubation at 38◦ C for 1 h. 12. The solution containing the MoFe protein crystals is centrifuged at 20,200 × g for 15 min at 38◦ C. The centrifuge rotor should be equilibrated at 38◦ C beforehand. 13. The supernatant is discarded, and the sediment is washed by stirring it gently in pre-warmed (38◦ C) crystallization buffer B. 14. The solution containing protein crystals is centrifuged again at 20,200 × g for 15 min at 38◦ C. The supernatant is discarded and the pellet is dissolved in approximately 8 mL of crystallization buffer C which is pre-chilled to 4◦ C. 15. The solution is then equilibrated on ice for about 10 min to allow crystals to further dissolve in the solution. Following this step, the solution is centrifuged at 25,000 × g for 20 min at 4◦ C. 16. The supernatant, now containing the pure MoFe protein, is collected in a Schlenk flask. The protein is then frozen as pellets in liquid N2 . 3.3. Purification of Non-tagged Fe Protein

1. The Fe protein fraction (see Section 3.2) is diluted at least twofold with DEAE equilibration buffer II and then loaded on the pre-equilibrated 2.5 × 7 cm DEAE column. Subsequently, the protein-bound column is washed with two to three column volumes of the equilibration buffer, and eluted

100

Wiig et al.

with the elution buffer in the reverse direction of loading in order to concentrate this fraction down to ∼30 mL (see Note 7). 2. The concentrated Fe protein fraction is then loaded onto a 7.5 × 100 cm Sephacryl S-200 SF size exclusion chromatography column. This column is pre-equilibrated overnight with the Sephacryl equilibration buffer. 3. The Fe protein is eluted at 1.2 mL/min and monitored at OD405 . The collected Fe protein fraction is then diluted fourfold with DEAE equilibration buffer II and loaded on a 2.5 × 30 cm, pre-equilibrated DEAE column. 4. The column is briefly washed with equilibration buffer after loading, and the proteins bound to the column are then eluted in a linear NaCl gradient as follows: gradient buffer IIIB is pumped into gradient buffer IIIA at about 50 mL/h using a peristaltic pump; while gradient buffer II A is loaded on the column at about 100 mL/h at the same time. The elution of the purified Fe protein is monitored at OD405 . 5. The pure Fe protein collected from the NaCl gradient can be concentrated by another DEAE column step as described in Step 1. The protein is then frozen as pellets in liquid N2 . 3.4. Purification of His-Tagged Nitrogenase Proteins

1. For the purification of His-tagged nitrogenase proteins, only 400 g of cells are used to accommodate the capacity of the IMAC column. Cell lysate is prepared as described in Section 3.1. 2. The cell lysate is then incubated with 10 μg/ml deoxyribonuclease, 0.2 mM PMSF, and 0.5 M NaCl for 45 min (see Notes 8 and 9). 3. A 2.5 × 30 cm Zn(II)-IMAC column is packed and washed briefly with water. It is then charged with Zn(II) by passing three column volumes of Zn(II) sulfate solution through the column. The excess, unbound Zn(II) ions are removed by passing water through the column, which is followed by the equilibration of the column with five column volumes of the equilibration buffer. 4. The cell lysate is then loaded on the Zn(II)-charged, preequilibrated IMAC column at approximately 300 mL/h. The flow-through of the column is collected in an air-tight container for further purification of non-tagged nitrogenase proteins (see Section 3.3). 5. Subsequently, the column is washed with 5–10 column volumes of wash buffer. An extensive washing step here usually helps to remove most non-specifically bound contaminants from the column.

Purification of Nitrogenase Proteins

1

2

3

4

101

5

250 kDa 150 kDa 100 kDa 75 kDa 50 kDa

37 kDa

25 kDa 20 kDa

Fig. 7.2. SDS-PAGE for the IMAC purification of apo-MoFe protein (DJ1143). Coomassie brilliant blue: lanes 1, Mr standards; lane 2, sample of crude extract that was loaded onto the IMAC column; lane 3, IMAC column flow-through; lane 4, IMAC column eluate from 20 mM imidazole–HCl wash; and lane 5, protein collected from 250 mM imidazole–HCl elution.

6. The protein is then eluted in the elution buffer and collected by monitoring the elution profile at OD405 . The purified protein is frozen as pellets in liquid N2 . An example of the purified His-tagged MoFe protein is shown in Fig. 7.2.

4. Notes 1. The procedures described here are primarily designed for the purification of A. vinelandii nitrogenase. 2. The most commonly used Schlenk system in nitrogenase work is a vacuum–argon dual manifold. 3. All buffers (with the exception of the Zn(II) solution) are made anaerobic as follows: Buffers are first degassed for 1–2 h under vacuum (10–4 torr) with gentle stirring, followed by the addition of appropriate amounts of Na2 S2 O4 to the buffers. The buffers are then kept anaerobic under argon until use. If a buffer is anaerobic and capable of scavenging oxygen, a few drops of this buffer will turn 5 mL of methyl viologen solution (2 mM) from colorless to dark blue instantaneously. It is critical to always check buffers and columns by this method before applying them to protein solutions. In addition, the reductant, Na2 S2 O4 ,

102

Wiig et al.

tends to degrade and form reactive breakdown products over a longer period of time. Therefore, Na2 S2 O4 -containing buffers that are more than a day old should not be used. Buffers that need to be kept for an extended period of time are usually prepared without Na2 S2 O4 , and appropriate amounts of Na2 S2 O4 are added right before the purification procedure. 4. The purification procedure of non-tagged MoFe protein is typically performed over a period of 2–3 days. Preparation of materials, including the columns and cell lysate (see Section 3.1, Steps 5–6), are done on the first day and the cell lysate can be stored overnight as described above. The separation of Fe protein and MoFe protein (see Section 3.2, Steps 1–6) is usually performed on the second day and the further purification of the non-tagged MoFe protein (see Section 3.1, Steps 7–17) on the third day. 5. All resin materials should be suspended in the appropriate buffer at a 1:1 ratio, and degassed under vacuum for 1–2 h before being packed into a chromatography column. Typically, it takes about three column volumes of anaerobic buffer to fully equilibrate a column; however, the usage of up to 10 column volumes of equilibration buffer is not uncommon. 6. Nitrogenase proteins are best stored as frozen pellets in liquid N2 . Short-term storage of proteins in air-tight containers at –80◦ C is acceptable, but the proteins eventually lose their activities over an extended period of time. 7. To concentrate proteins, Q-sepharose can be used instead of DEAE sepharose. The same buffers and operational procedures can be used as described above. 8. The addition of deoxyribonuclease is not necessary, but it dramatically helps reduce cell lysate viscosity. 9. If a different protease inhibitor is used, make sure it does not chelate metals.

Acknowledgments The authors are supported by National Institutes of Health grant GM 67626 (M.W.R.) and Herman Frasch Foundation grant 617HF07 (M.W.R.).

Purification of Nitrogenase Proteins

103

References 1. Burgess BK, Wherland S, Stiefel EI et al (1980) HD formation by nitrogenase: a probe for N2 reduction intermediates. In: Newton WE, Otsuka S (eds) Molybdenum Chemistry of Biological Significance, pp. 73–84. Plenum Press, New York, NY 2. Stiefel EI, Burgess BK, Wherland S et al (1980) Azotobacter vinelandii Biochemistry: H2 (D2 ) relationships and some aspects of iron metabolism. In: Newton WE, OrmeJohnson WH (eds) Nitrogen Fixation, pp. 211–222. University Park Press, Baltimore, MD 3. Peters JW, Stowell MH, Soltis SM et al (1997) Redox-dependent structural changes in the nitrogenase P-cluster. Biochemistry 36:1181–1187 4. Georgiadis MM, Komiya H, Chakrabarti P et al (1992) Crystallographic structure of the nitrogenase iron protein from Azotobacter vinelandii. Science 257:1653–1659 5. Rawlings J, Shah VK, Chisnell JR et al (1978) Novel metal cluster in the iron-molybdenum cofactor of nitrogenase. Spectroscopic evidence. J Biol Chem 253:1001–1004 6. Cramer SP, Gillum WO, Hodgson KO et al (1978) The molybdenum site of nitrogenase. 2. A comparative study of molybdenum-iron proteins and the iron-molybdenum cofactor by x-ray absorption spectroscopy. J Am Chem Soc 100:3814–3819 7. Shah VK, Brill WJ (1977) Isolation of an iron-molybdenum cofactor from nitrogenase. Proc Natl Acad Sci USA 74:3249–3253 8. Hu Y, Fay AW, Ribbe MW (2005) Identification of a nitrogenase FeMo cofactor

9.

10. 11.

12.

13.

14.

15.

precursor on NifEN complex. Proc Natl Acad Sci USA 102:3236–3241 Kim CH, Newton WE, Dean DR (1995) Role of the MoFe protein alpha.-subunit histidine-195 residue in FeMo-cofactor binding and nitrogenase catalysis. Biochemistry 34:2798–2808 Burgess BK (1990) The iron-molybdenum cofactor of nitrogenase. Chem Rev 90: 1377–1406 Eady RR, Lowe DJ, Thorneley RNF (1978) Nitrogenase of Klebsiella pneumoniae: A pre-steady state burst of ATP hydrolysis is coupled to electron transfer between the component proteins. FEBS Lett 95: 211–213 Davis LC, Shah VK, Brill WJ (1975) Nitrogenase: VII. Effect of component ratio, ATP and H2 , on the distribution of electrons to alternative substrates. Biochim Biophys Acta 403:67–78 Burgess BK, Jacobs DB, Stiefel EI (1980) Large-scale purification of high activity Azotobacter vinelandii nitrogenase. Biochim Biophys Acta 614:196–209 Christiansen J, Goodwin PJ, Lanzilotta WN et al (1998) Catalytic and biophysical properties of a nitrogenase apo-MoFe protein produced by a nifB-deletion mutant of Azotobacter vinelandii. Biochemistry 37: 12611–12623 Lee CC, Hu Y, Ribbe MW (2009) Unique features of the nitrogenase VFe protein from Azotobacter vinelandii. Proc Natl Acad Sci USA 23:9209–9214

Chapter 8 Assays of Nitrogenase Reaction Products William E. Newton and Michael J. Dilworth Abstract Steady-state assays of nitrogenases share at least five requirements: an anaerobic environment, a consistent source of magnesium adenosine triphosphate (MgATP), a suitable source of reductant, a buffer system compatible with the product-quantification protocol to be used, and the desired substrate. The assay is initiated by injection of the component protein(s) of the enzyme or MgATP and terminated by injection of either acid or a solution of Na2 EDTA. The various nitrogenases catalyze the reduction of a wide variety of substrates. This chapter outlines the methods used to analyze the products of nitrogenasecatalyzed reactions involving nitrogen–nitrogen bonds, nitrogen–oxygen bonds, carbon–nitrogen bonds, carbon–carbon bonds, carbon–oxygen bonds, carbon–sulfur bonds, and hydrogen only. The usefulness of measurements of residual amounts of other components of nitrogenase assays is also discussed. Key words: Nitrogenase, substrate, catalysis, steady-state assays, electron balance.

1. Introduction Steady-state assays of the molybdenum-, vanadium-, and ironnitrogenases share at least five requirements. First, all require an anaerobic environment and O2 -free argon is the most commonly used. Assays are, therefore, conducted in rubber-sealed vials where the gas space can be conveniently changed either by flushing with flowing gas or by several cycles of evacuation and re-filling with the gas of choice. Second, all require a consistent source of magnesium adenosine triphosphate (MgATP), which is usually provided by the creatine phosphate/creatine phosphokinase system. Together, these components regenerate MgATP from its product, MgADP, and in doing so prolong the period of enzymatic activity, which would otherwise become inhibited by M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_8, © Springer Science+Business Media, LLC 2011

105

106

Newton and Dilworth

accumulating MgADP. Third, all require a source of reductant, usually provided by sodium dithionite, but either titanium(III) citrate or reduced flavodoxin can also serve this function for more specialized purposes. The reductant is added to sealed assay vials by injection after the gas phase has been rendered O2 -free. Fourth, the chosen buffer system must not only maintain the required pH for effective enzymatic activity but also be compatible with assay methods used for quantification of the reduction products from any particular substrate. Last, a source of the desired substrate must be added. If no other substrate is added, all nitrogenases catalyze the reduction of protons to produce dihydrogen (H2 ). In fact, even in the presence of added substrates, considerable amounts of H2 are often produced. High concentrations of some substrate sources, like sodium azide or sodium cyanide, may prove inhibitory to the activity of some nitrogenases under particular conditions. Once the above reaction conditions are established, the assay is initiated by one of several methods, most usually by injection of an appropriate mixture of the two component proteins of the enzyme, or of one of them, or of MgATP. Assays are terminated by injection of either acid, usually trichloroacetic or mineral acid, or a solution of Na2 EDTA at ca. pH 7.5. The use of EDTA avoids acid-catalyzed hydrolysis of pH-sensitive components and substrates (e.g., creatine phosphate). In addition to the rate of product formation, an assay can also indicate the extent of coupling between electrons transferred to substrate and the amount of MgATP hydrolyzed. To obtain this information, first the actual electron flux from reductant must be determined. This is performed by measuring either all products of the assay, each of which must account for specific number of electrons, or the actual consumption of reductant, i.e., the difference between the amount of, say, dithionite remaining and the initial amount added. Then, MgATP consumption must be calculated from the amount of either phosphate or creatine released from the creatine phosphate/creatine phosphokinase system (see below). These data give an index of coupling efficiency, which is usually expressed as moles of ATP hydrolyzed per electron pair transferred to substrates, the so-called ATP/2e– ratio. The various nitrogenases catalyze the reduction of a wide variety of substrates. The available data for the V- and Fe-enzymes are still quite limited. Frequently, product analysis has only ever been specified by one investigating group and the methods used have not, therefore, been widely tested. This chapter outlines the methods used but does not attempt to authenticate them. Moreover, the electron balance for the wide range of substrates, whose reductions are catalyzed by nitrogenase, is often incomplete and any difference between electrons extracted from dithionite (or other reductant) and electrons counted in products could indicate missing products.

Assays of Nitrogenase Reaction Products

107

2. Materials Assays are typically carried out in 9.25-mL assay vials at 30◦ C in 25 mM HEPES (or some other non-interfering buffer, like MOPS) at pH 7.4. A range of buffer types has been used with Mo-nitrogenase, which is optimally active within the pH range of 7–8 (1). A common reaction time is 8–10 min. Assays are often limited to 0.5–1.0 mg of total nitrogenase proteins in 1 mL to avoid kinetic complications arising from either highor low-protein concentrations. Each assay contains in the 1-mL final volume, 30 μmol creatine phosphate, 25 μmol HEPES buffer, pH 7.4, 20 μmol sodium dithionite, 5 μmol MgCl2 , 2.5 μmol ATP, and 0.125 mg creatine phosphokinase. Gaseous substrates and/or inhibitors (typically CO) are added by gas-tight syringe to a 101 kPa argon atmosphere and then vented to atmospheric pressure after temperature equilibration. Often, the MoFe (or VFe or FeFe) protein is added and, after a 3-min incubation at 30◦ C, the reaction is initiated by addition of Fe protein. Alternatively, an appropriate mixture of component proteins can be added. Reactions are terminated by injection of 0.25 mL 0.5 M EDTA-Na2 , pH 7.5. The creatine released from creatine phosphate as the ATP is recycled during the assay is usually measured spectrophotometrically by a method (2) modified from the original (3). Gaseous products are quantified by gas chromatography, using calibration gas mixtures, typically 1000 ppm C2 H4 in He, 1000 ppm C2 H6 in He, and 1% H2 in N2 (Scott Specialty Gases, Plumsteadville, PA), and appropriate detection methodology. Soluble products are assayed as described below.

3. Methods 3.1. Catalyzed Reactions Involving Nitrogen–Nitrogen Bonds 3.1.1. Dinitrogen

The physiological substrate for all nitrogenases is dinitrogen (N2 ). The major product of its catalyzed reduction is ammonia (NH3 ). It is the only nitrogenous product detected with Mo-nitrogenase (equation [1]), whereas a small but significant amount of the four-electron-reduced product, hydrazine (N2 H4 ), is detected with V-nitrogenase (equation [2]) (4). Hydrazine production is measured spectrophotometrically (at 458 nm) after reaction with acidic p-dimethylaminobenzaldehyde (5). N≡N + 8H+ + 8e− → 2NH3 + H2

[1]

N≡N + 4H+ + 4e− → H2 N−NH2

[2]

108

Newton and Dilworth

The standard colorimetric methods used to measure NH3 in solution involve either Nessler’s reagent, which produces an orange-colored mercuri-ammonia colloidal complex (measured at 490 nm), or the formation of indophenol blue by hypochlorite oxidation of alkaline phenol in the presence of NH3 and nitroprusside (measured at 625–630 nm). The Nessler reagent incorporates a protective colloid, usually either gum ghatti or gum acacia, to help stabilize the complex, but its instability limits the amount of ammonia that can be present in the assay. The indophenol method is much more sensitive than the Nessler method and gives a color that is stable over long times, but whose formation is subject to a variety of serious interferences. Even so, it is the most commonly used method. The reagents used and the way they are combined can cause some confusion and even bewilderment! When pre-combined reagents are used, rather than adding the alkaline phenol, nitroprusside, and hypochlorite separately (6), they have a very limited shelf life. Even when the phenol reagent is kept separate at 4–6◦ C, oxidation of the phenol causes the reagent to darken and produce an increased background absorbance in the absence of ammonia. The sensitivity of this assay can be increased by using more concentrated reagents (5, 7) than originally formulated (6, 8). NH3 can also be quantified using a reagent that contains 10 mM o-phthalaldehyde and 7 mM 2-mercaptoethanol to both terminate the nitrogenase assay and form an adduct with NH3 . This ammonia adduct is separated by HPLC using acetonitrile/0.01 M potassium phosphate (pH 7.3) and estimated fluorimetrically (9). Although rapid, requiring only a 2-min separation on a C18 guard column, this method has its drawbacks: (i) the o-phthalaldehyde reagent contains 2-mercaptoethanol, which can provoke allergic responses in sensitive individuals; (ii) the reactions with NH3 must be performed in the dark; and (iii) under some circumstances, the reaction produces more than the one peak for NH3 . This technique seems to have dropped out of routine use. The major difficulty in ammonia estimation is that sources of ammonia contamination are multitudinous and include filter papers, ammonia volatilized from ammonium hydroxide, contaminants in laboratory reagents, tobacco smoke (even when carried on clothing), and latex paints to name but a few, and accurate results depend on rigorous exclusion of such adventitious ammonia. The one-time use of disposable tubes is strongly recommended; these are infinitely cheaper than repeated experiments! These many problems explain why measurement of nitrogenase activity via its natural product, NH3 , is not the routine one. Furthermore, many of the chemicals in standard nitrogenase assays, e.g., dithionite, sulfite, creatine from creatine phosphate, N-containing buffers, interfere with ammonia estimation.

Assays of Nitrogenase Reaction Products

109

Consequently, most protocols for assaying ammonia depend on being able to separate it from these and other interfering compounds and use one of two broad approaches, either distillation or ion-exchange chromatography, to do so. (i) Microdistillation of Ammonia Microdistillation (see Note 1) involves diffusion of ammonia from an alkalinized reaction mixture into a drop of acid (usually H2 SO4 ) on an etched glass rod mounted in a rubber stopper inserted into the assay vial containing the mixture (10, 11). Alkalinization can be effected by adding either saturated potassium carbonate or various alkaline buffers, like sodium tetraborate (pH ca. 10). The latter is preferred if the assay mixture is strongly acidic because it avoids alkali splash accompanying CO2 evolution from K2 CO3 . The vials can either be left static overnight or rotated slowly at an angle, usually for at least 2 h, to continuously renew the surface and so promote diffusion. The ammonia collected in the acid drop on the glass rod is then diluted in a known volume of water and samples analyzed using either the Nessler or indophenol methods. An alternative distillation method uses dilute acid (either boric or sulfuric) in the central well of Conway diffusion dishes (12, 13), followed by either titration with a microburette or assay via the indophenol method. Microdistillation is preferred in nitrogenase assays driven by high concentrations of reductants, like titanium citrate (13, 14). (ii) Ion-Exchange Chromatography An alternative approach is to use ion-exchange chromatography to remove interfering substances (2). Passage through columns of Dowex-1 anion exchange resin (Cl– form) in glass wool-plugged Pasteur pipettes (2.5 cm long × 0.6 cm diameter) removes dithionite, sulfite, ATP, ADP, protein, unused creatine phosphate, phosphate, and perchlorate or trichloroacetate, when either acid is used to terminate the assays. When EDTA is used for assay termination, its removal requires a longer column (6 cm). (iii) Other Common Interferences Both dithionite and a number of N-containing buffers interfere with the indophenol method (15–17). Tris, tris(hydroxymethyl)aminomethane, is a relatively inexpensive buffer that is commonly used for nitrogenase purification; its use results in the greatest interference (17). In contrast, HEPES, N-(2-hydroxy-ethyl)piperazine-N -2ethanesulfonic acid, is fully compatible with nitrogenase assays and presents no such problems. Only relatively recently was creatine recognized as a major source of interference in measurements of ammonia.

110

Newton and Dilworth

A preliminary solution was to determine experimentally a correction factor from duplicate ammonia assays to which known quantities of creatine had been added to one set of replicates (2). This added element is not really a hardship because the amount of creatine produced from creatine phosphate in the assay (3) is often used as a measure of the amount of MgATP hydrolyzed, which is then used to estimate the ATP/2e– ratio. It, therefore, yields data for MgATP usage as well as for allowing a correction to be made for up to 1.5 μmol creatine per ammonia assay. This correction approach works well for wildtype nitrogenases with ATP/2e– ratios of ca. 4 because an ammonia assay of 50 nmol NH3 is accompanied by only ca. 400 nmol creatine (2). For variant nitrogenases, however, where the ATP/2e– ratio might be as high as 20 and where the enzyme’s affinity for N2 is low (18), the amounts of creatine produced can be so large that the ammonia produced is virtually invisible and the correction factor approach becomes unworkable. An answer to this problem is to capture the NH4 + in the eluate from the Dowex-1 column (see (ii) above) by adsorption onto a second (1.5 cm long × 6 mm diameter) column of a cationexchange resin (Dowex-50, Na+ form). The small fraction of the creatine that is also adsorbed can be easily removed by a 0.25 M NaCl wash. Elution with 2 M NaCl then removes the ammonia in a small volume for normal assay; the 2 M NaCl does not affect the color yield in the indophenol method (19). If the ATP/2e– ratio is not required, this two-column method obviates any need to measure creatine because no correction is required. 3.1.2. Hydrazine

The catalyzed reduction of hydrazine (N2 H4 ) to NH3 (equation [3]) (20) by Mo-nitrogenase was originally reported without details of either how ammonia was measured or how it was separated from the substrate. H2 N−NH2 + 2H+ + 2e− → 2 NH3

[3]

Hydrazine was confirmed as a substrate (21) with the NH3 product separated by alkaline diffusion before measurement by the indophenol method. No interference from N2 H4 was noted unless concentrations of >30 mM were used—presumably then N2 H4 distils into the trapping acid, but only at high concentrations does it interfere with indophenol color development. Freshly made N2 H4 solutions should be used to ensure satisfactory ammonia recovery (21). High concentrations of hydrazine are required to obtain any appreciable amounts of product because, even at pH 8 where the reduction appears to be maximal, the apparent Km (ca. 25 mM) is very high (21). Ammonia

Assays of Nitrogenase Reaction Products

111

derived from hydrazine reduction, after separation by distillation, has also been measured by the Chaykin procedure (8, 22). No column-based method for separating substrate and product has yet been developed. The form of the substrate used is important because the counter-ion to hydrazinium (N2 H5 + ) may inhibit enzyme activity due to high ionic strength (21) and sulfate can inhibit nitrogenase activity (K Fisher and M.J. Dilworth, unpublished data). Neutralization of hydrazine hydrate with acetic acid (21) is the procedure of choice. 3.1.3. Azide and Hydrazoic Acid

N2 and NH3 were originally reported as the only products (23, 24) from reduction of azide solutions by Mo-nitrogenase. However, this reaction forms multiple products (5, 25, 26), which are depicted in equations [4]–[7] (see Note 2). Azide (N− 3 ) is reduced by two electrons to yield N2 and ammonia (25) in a reaction with a relatively high Km (equation [4]). In contrast, hydrazoic acid (HN3 ) reduction yields N2 H4 and ammonia (equation [5]) with an extremely low Km . Even more interestingly, total ammonia production significantly exceeds the total of N2 plus N2 H4 produced. An additional ammonia-producing reaction must also be occurring (see equations [6] and [7]). One explanation is that the N2 produced (equation [4]) is subsequently further reduced to ammonia (equation [6]). This suggestion finds support by added H2 markedly suppressing this “excess” ammonia production. However, added H2 does not eliminate it completely (25). This result is somewhat surprising because the concentration of product N2 in the gas phase is very low (ca. 1 kPa) compared with the Km (of ca. 10 kPa) for N2 reduction by Mo-nitrogenase and so complete inhibition by H2 would be expected. Equation [7] offers the alternative explanation that “excess” ammonia is a product of a direct eight-electron reduction of azide to ammonia (5, 26). This alternative pathway is clearly supported by studies with V-nitrogenase (27). [N≡N−N]− + 3H+ + 2e− → N2 + NH3

[4]

N ≡ N−NH + 6H+ + 6e− → N2 H4 + NH3

[5]

N≡N + 6H+ + 6e− → 2NH3

[6]

[N ≡ N−N]− + 9H+ + 8e− → 3 NH3

[7]

Terminating assays with EDTA rather than acid avoids azide decomposition, and the use of an anion exchange column removes any excess azide. Hydrazine is easily measured spectrophotometrically, using acidic p-dimethylaminobenzaldehyde (5) and ammonia production as described for its formation from N2 (see Section 3.1.1). Although N2 H4 does interfere with the

112

Newton and Dilworth

color reaction for ammonia, it is a minor product in these reactions and unlikely to introduce serious error. In contrast, measurement of the N2 evolved presents a considerable problem: one that involves ensuring that atmospheric N2 does not contaminate either the assay vials or the gas-sampling syringes needed for gas chromatographic measurement. Atmospheric N2 contamination was probably a factor in the initial reports of azide reduction (23, 24). The contamination problem can be overcome by using either terminally 15 N-labeled azide (to give 50% 29 N2 and 50% 28 N2 ) or fully 15 N-labeled azide (to give 100% 30 N2 ). Terminally labeled KN3 is available commercially or can be synthesized from an excess of unlabeled hydrazine reacting with 15 N-labeled ethyl nitrite produced by treatment of Na15 NO2 with ethanol/HCl (28). Fully 15 N-labeled KN3 can be synthesized via the same route but using both fully labeled reagents (26). The gas phase in the assays is then analyzed mass spectrometrically and the amount of N2 produced calculated from either the 29 N2 or 30 N2 peaks by comparison with the 40 Ar peak (the inert atmosphere in the assays) with due allowance for the different ionization efficiencies in the mass spectrometer used. If the mass spectrum could be affected by other contributions to mass 29 (e.g., from 13 CO), fully 15 N-labeled substrate may have to be used. 3.2. Catalyzed Reactions Involving Nitrogen–Oxygen Bonds 3.2.1. Nitrous Oxide

Nitrous oxide (N2 O) was probably the first alternative substrate identified for Mo-nitrogenase. Its catalyzed reduction yields N2 and water initially (equation [8]), with the N2 subsequently reduced to NH3 (as in equation [6]) (29, 30). The ammonia can be assayed by any of the techniques described above (see Section 3.1.1), but quantification of the N2 evolved presents similar problems to those with azide (see Section 3.1.3). Again, use of labeled substrate (15 N2 O) with subsequent mass spectrometric measurement of 15 N2 is preferred, but now the 15 N2 O substrate must be removed by freezing assay mixtures in liquid N2 before entry into the mass spectrometer because it gives rise to peaks at m/e 28, 29, 30 (and 44 and 45) (31), which would confound the isotopic analysis. N≡N−O + 2H+ + 2e− → N≡N + H2 O

3.2.2. Nitrite

[8]

Nitrite (NO− 2 ) is a substrate for Mo-nitrogenase (equation [9]) and is reduced to ammonia (32), but its close relative nitrate is not. [O=N−O]− + 7H+ + 6e− → NH3 + 2H2 O

[9]

The ammonia produced was originally quantified by the o-phthalaldehyde method (9) but the ion-exchange method

Assays of Nitrogenase Reaction Products

113

(see Section 3.1.1) is also appropriate. Catalyzed nitrite reduction also irreversibly inactivates the Fe-protein component of Mo-nitrogenase (32). 3.3. Catalyzed Reactions Involving Carbon–Nitrogen Bonds 3.3.1. Cyanide

The actual substrate for Mo-nitrogenase appears to be HCN (33) and, like azide, its reduction results in multiple products. Four-electron reduction (equation [10]) produces methylamine (CH3 NH2 ), whereas six-electron reduction (equation [11]) leads to methane (CH4 ) and ammonia (NH3 ). At low NaCN concentrations, the amount of NH3 produced exceeds that of methane and this “excess” NH3 formation has been a source of debate for some time (see below). HC≡N + 4H+ + 4e− → CH3 −NH2

[10]

HC≡N + 6H+ + 6e− → CH4 + NH3

[11]

In addition, ethylene and ethane are produced (23, 34), but at very low concentrations compared to those of CH4 , NH3 , and CH3 NH2 ; for example, values of 0.08 and 0.07% of the amount of CH4 , respectively, have been reported (34). Measurement of the components of such complex product mixtures can be problematical. Some investigators have based their results on the observation of relatively constant ratios between the amounts of products formed (33) so that only one product (usually CH4 ) was quantified and the amount of the others calculated using those ratios. These methods have obvious shortcomings, which become acute when variant Mo-nitrogenases and V-nitrogenase are used and the ratios are different to those for wild type (27, 35). The hydrocarbon products can be quantified by gas chromatography with flame ionization detection (FID) (23, 33). To define the stoichiometry of the products of HCN reduction, both NH3 and CH3 NH2 were measured by HPLC separation of their Dabsyl [4-(dimethylamino)azobenzene-4 -sulfonyl] derivatives on a C18 column in acetonitrile–95% ethanol–water (1:1:1) by a method modified from the protocol by Lin and Lai (36). Some difficulty, however, was encountered with measuring small amounts of ammonia (33). Further, addition of KIO3 in HCl as the stop reagent was required to oxidize residual dithionite and so prevent it from reducing the Dabsyl reagent. Quantification of ammonia and methylamine in nitrogenase assay mixtures is complicated by their similar chemical properties. Methylamine reacts in the indophenol assay for NH3 , although the color yield is only about 7% of that for ammonia on a molar basis, making the accuracy of NH3 determinations questionable if the CH3 NH2 concentration is unknown. Furthermore, CH3 NH2 binds to columns of Dowex-50 (Na+ form) and is eluted by the 2 M NaCl used to elute NH4+ (19). Therefore,

114

Newton and Dilworth

a separation protocol, based on a combination of pH and ionic strength of eluant, was developed for sequentially eluting NH3 and CH3 NH2 from mini-columns of Dowex-50 (Na+ form) (see Note 3) (35). The NH3 can then be measured with the indophenol assay and the CH3 NH2 via its reaction with ninhydrin (37). This technique gives very similar ratios of products to those used by Li and coworkers (33) for wild-type Mo-nitrogenase. However, markedly different ratios of products are formed with some of the variant Mo-nitrogenases (35). The source of “excess” ammonia during HCN reduction was initially suggested to arise from hydrolysis of a two-electron reduced product, namely methyleneimine (CH2 =NH), that had “leaked” from the site. Its hydrolysis would result in formaldehyde (H2 C=O) and NH3 as products (equation [12]), but the presence of H2 C=O could not be confirmed (33). HC≡N + 2H+ + 2e− → [H2 C=NH] + H2 O → H2 C=O + NH3

[12]

However, added formaldehyde could be recovered from nitrogenase assay mixtures containing cyanide (27). Such assays are stopped with HCl, their pH adjusted to 4.5 with acetate buffer, a starch solution added, and the mixture titrated with iodine to the blue endpoint where both dithionite and sulfite have been oxidized. The oxidation of sulfite and dithionite is essential because both prevent subsequent color development in the reaction between formaldehyde and the acetylacetone/ammonium acetate reagent (38). Residual cyanide also interferes with color development and can be removed by precipitation with AgNO3 , followed by centrifugation. Added EDTA also inhibits background color development (apparently due to metals liberated from nitrogenase). Although quantities of 5–10 nmol of added formaldehyde could be measured with a recovery of about 90%, no trace of formaldehyde could be detected in assays for HCN reduction with wild-type Mo-nitrogenase from Azotobacter vinelandii. This lack of success led to HCN reduction assays with V-nitrogenase (27), which has an active site with a known tendency to “leak” reduction intermediates (4). Using all the modifications described above, significant amounts of H2 C=O were detected, so validating equation [12]. The potential product of a further twoelectron reduction of H2 C=O, namely H3 C–OH (methanol), was not detected (27). 3.3.2. Alkyl and Alkenyl Cyanides

Mo-nitrogenase catalyzes the reduction of alkyl cyanides (R–C≡N) up to and including n-butyronitrile with the corresponding alkane and ammonia as products (equation [13]) (39). The alkane is quantified by gas chromatography on either alumina

Assays of Nitrogenase Reaction Products

115

or Porapak columns, using FID. The corresponding alkylamine (R–CH2 –NH2 ) is a potential four-electron-reduced intermediate but it has neither been detected nor sought. Ammonia can be quantified as in Section 3.1.1. R−C≡N + 6H+ + 6e− → R−CH3 + NH3

[13]

Alkenyl cyanides are more reactive than their alkyl cyanide analogs toward Mo-nitrogenase-catalyzed reduction (39). Acrylonitrile (CH2 =CH–C≡N) is the most reactive (40) and yields NH3 , propene, and propane (equations [14] and [15]); propane is apparently not formed from subsequent propene reduction. The hydrocarbon products are quantified by gas chromatography on either alumina or Porapak-type columns with FID and NH3 by methods described in Section 3.1.1 using the indophenol assay. CH2=CH−C≡N + 6H+ + 6e− → CH2 =CH−CH3 + NH3 [14] CH2=CH−C≡N + 8H+ + 8e− → CH3 −CH2 −CH3 +NH3 [15] 3.3.3. Cyanamide

Both Mo- and V-nitrogenase catalyze the reduction of cyanamide (N≡C–NH2 ) by both a six-electron reduction (equation [16]), which results in CH3 NH2 plus NH3 , and an 8-electron reduction (equation [17]), yielding CH4 plus two NH3 (see Note 4) (41). N≡C−NH2 + 6H+ + 6e− → NH3 + H3 C−NH2

[16]

N≡C−NH2 + 8H+ + 8e− → CH4 + 2NH3

[17]

Product methane was measured by gas chromatography on Porapak N at 51◦ C using FID, whereas NH3 was measured with o-phthalaldehyde–mercaptoethanol (9). Methylamine was separated from the reaction mixture by microdistillation and estimated with 2,4-dinitrofluorobenzene using a modification of a published protocol (42). 3.3.4. Methyl Isocyanide

Reduction of methyl isocyanide (CH3 –N≡C) leads to a wide variety of products (34, 43). Reduction by four electrons produces dimethylamine, (CH3 )2 NH (equation [18]), whereas a sixelectron reduction (equation [19]) produces methane plus methylamine (44). The latter reaction produces “excess” methylamine when compared with the methane produced which may indicate hydrolysis of a two-electron reduced intermediate that escapes the active site and produces methylamine and an undetected

116

Newton and Dilworth

carbon-containing compound (possibly H2 C=O as observed for HCN reduction; see Section 3.3.1) on hydrolysis. CH3 −N≡C + 4H+ + 4e− → CH3 −N(H)−CH3

[18]

CH3 −N≡C + 6H+ + 6e− → CH3 −NH2 + CH4

[19]

In addition to CH4 , the reduction products include both C2 H4 and C2 H6 among other hydrocarbons (34, 44), which arise apparently from sequential reduction and insertion reactions. All hydrocarbons are measurable by gas chromatography on alumina, Porapak N, or Porapak R columns using FID. The nitrogenous products have only been fully quantified occasionally. Kelly (43) measured total amines by Conway diffusion into sulfuric acid and identification by paper chromatography. Rubinson and coworkers (44) measured CH3 NH2 by an HPLC fluorescence method, using the o-phthalaldehyde–mercaptoethanol reagent and reaction at 0◦ C, a C18 guard column, and methanol/0.05 M sodium citrate (pH 5) (55:45) as solvent. Dimethylamine was measured by reaction with 0.25 M copper nitrate in 0.5 M NH2 OH·HCl (pH 9.5), extraction with 2% CS2 in CHCl3 , and separation by HPLC using 436-nm absorption (44). 3.4. Catalyzed Reactions Involving Carbon–Carbon Bonds 3.4.1. Acetylene

The nitrogenase-catalyzed reduction of acetylene to ethylene (equation [20]) (45, 46) is the most commonly used assay in nitrogen fixation research (see Note 5). Although not normally a product of wild-type Mo-nitrogenase catalysis except at high temperature (47), ethane (equation [21]) is produced by the V-, Fe-, and some Mo-nitrogenases containing genetically altered MoFe proteins (48). HC≡CH + 2H+ + 2e− → H2 C=CH2

[20]

HC≡CH + 4H+ + 4e− → H3 C−CH3

[21]

Standard assays usually contain 5–20 kPa C2 H2 , but the acetylene concentration in the assay will depend on the Km of the nitrogenase under study (49). Acetylene and ethylene are readily separable on a variety of column packing materials, for example, alumina and Porapak N, R, or T, using either N2 or He as the carrier gas and FID. Porapak N is also capable of separating ethane from both acetylene and ethylene, but the separation can be marginal depending on the age and length of the column, carrier gas flow rate, etc. Activated alumina, although slower, gives better separation of these three gases (50). Dideuteroacetylene (C22 H2 , C2 D2 ) has been used to probe the stereospecificity of catalyzed acetylene reduction (see Note 5). When reduced in water (H2 O), using either Mo- or

Assays of Nitrogenase Reaction Products

117

V-nitrogenase, the product is >95% cis-C2 H2 D2 as determined by infrared spectroscopy of the gas phase (45, 51). This mechanistic probe has been used more recently with genetically altered nitrogenases. Their stereospecificity of C2 D2 reduction was examined by using a large-volume assay (10–20 times the standard nitrogenase assay). After stopping the reaction, the liquid contents are frozen, using solid CO2 /ethanol (not liquid N2 , which would condense the argon and the hydrocarbons), and a fraction of the gas phase is transferred through a direct connection to an evacuated 100-mL infrared gas cell (49). Smaller path-length gas cells can be used with more sensitive FTIR instruments. After recording the infrared spectrum, the heights of the peaks at 843 cm–1 (for cis-C2 H2 D2 ) and at 988 cm–1 (for trans-C2 H2 D2 ) are measured. Because the molar absorbance of cis-C2 H2 D2 at 843 cm–1 is about double that for trans-C2 H2 D2 at 988 cm–1 (52), the percentages of the two products can be readily calculated. 3.4.2. Propyne and Propargyl Alcohol

Like its homologue ethyne (acetylene), propyne (CH3 –C≡CH) is reduced by Mo-nitrogenase (equation [22]) by two electrons to propene (23). Originally, the two gases were separated by gas chromatography at 0◦ C on 20% ethyl-N ,N -dimethyloxalamide on 100–200-mesh firebrick with FID as developed for the acetylene–ethylene assay (53), but alumina columns are now used (51, 54). Propyne is only slowly reduced by wild-type Mo-nitrogenases (23, 54) but both propyne and the related propargyl alcohol (2-propyn-1-ol, HOCH2 –C≡CH) are much better substrates for the genetically altered MoFe protein with α-70Ala replacing α-70Val (54). Propyne is not reduced by the V-nitrogenase (51). Propargyl alcohol is reduced by Mo-nitrogenase (equation [23]) to give allyl alcohol (HOCH2 – CH=CH2 ), which is identified by a combination of gas chromatography and mass spectrometry (54). CH3 −C≡CH + 2H+ + 2e− → CH3 −CH=CH2

[22]

HOCH2 −C≡CH + 2H+ + 2e− → HOCH2 −CH=CH2 [23] 3.4.3. Butynes

Another large decrease in reactivity occurs for the next homologue in the series, butyne (39). Of the butynes, only 1-butyne (HC≡C–CH2 –CH3 ) is a substrate for nitrogenase. 1-Butyne and its two-electron reduced product, 1-butene, are also separable and quantifiable using gas chromatography on either alumina or Porapak columns with FID.

3.4.4. Ethylene

Ethylene (ethene) is reduced to ethane, but it is a very poor substrate. Its Km is usually greater than 100 kPa for both Mo- and V-nitrogenase. Therefore, a high concentration of ethylene must

118

Newton and Dilworth

be used in assays (equation [24]), which makes the ethane contamination always found in commercial ethylene samples a significant problem. Unless the ethylene used is extremely pure, the contaminating ethane must be removed (see Note 6). Ethylene and ethane are easily separated and quantified by gas chromatography using FID and an alumina column, which allow satisfactory separation of the small amount of ethane product from the large amount of ethylene substrate (50, 51). H2 C=CH2 + 2H+ + 2e− → H3 C−CH3 3.4.5. Allene

The catalyzed reduction of allene (CH2 =C=CH2 ; equation [25]) produces propene (CH2 =CH–CH3 ) as expected. However, because allene could isomerize to propyne (and give the same product on reduction) before it is reduced, this reaction has been studied in D2 O to determine the stereochemistry of proton addition. Catalyzed allene reduction in D2 O leads exclusively to 2,3-dideuteropropene (CH2 =CD–CH2 D) rather than 1,2-dideuteropropene (CHD=CD–CH3 ), indicating that isomerization to and reduction of propyne is not involved (55). The various gases are easily separated and quantified by gas chromatography (55) not only on the variety of column types initially used (53) but also on materials like the Porapak series, with FID. CH2 =C=CH2 + 2H+ + 2e− → CH2 =CH−CH3

3.4.6. Cyclopropene

[24]

[25]

Cyclopropene reduction is catalyzed by Mo-nitrogenase (equation [26]) and results in a mixture of cyclopropane and propene. Reduction in D2 O yields cis-1,2-D2 -cyclopropane and a mixture of dideuteropropenes (principally trans-1,3-D2 -propene, cis-1,3D2 -propene, and 2,3-D2 -propene). Cyclopropane and propene can be separated from cyclopropene by gas chromatography on AgNO3 -impregnated firebrick, which removes the cyclopropene, and Porapak N, which separates ethane, propene, and cyclopropane, and then quantified using FID (56, 57). The techniques for identifying the various deuterated products, which involve mass spectrometry, IR spectroscopy, and NMR, are described elsewhere (58). CH2

CH

CH

cyclopropene + 2 H+ + 2 e−

CH2 CH2

CH3−CH=CH2

CH2

cyclopropane

+

propene

[26]

Assays of Nitrogenase Reaction Products

3.5. Catalyzed Reactions Involving Carbon–Oxygen and Carbon–Sulfur Bonds 3.5.1. Carbon Dioxide

119

The Mo-nitrogenase catalyzes a two-electron reduction of CO2 (equation [27]) to CO plus H2 O (59). The product CO is quantified as the carbonyl–hemoglobin complex (60, 61). O=C=O + 2H+ + 2e− → H2 O + C≡O

[27]

An alternative two-electron reduction product (equation [28]) from CO2 (R.R. Eady and F.K. Yousafzai, personal communication) has been identified initially by 13 C-NMR spectrometry. This product, formate, may account for as much as 20% of electron flux. Formic acid was also steam distilled from assay mixtures and assayed enzymatically with formate dehydrogenase by measuring the rate of NADH formation at 340 nm. O=C=O + 2H+ + 2e− → O=(H)C−OH 3.5.2. Carbon Disulfide

Carbon disulfide (CS2 ) reduction catalyzed by Mo-nitrogenase results in H2 S (13), but the fate of the carbon of the substrate is unknown (equation [29]). H2 S is recovered from acidified assay mixtures by microdistillation into zinc acetate solution, where it is converted to methylene blue by addition of N,Ndimethylphenylenediamine in 5 M HCl and ferric chloride in 1.2 M HCl and quantified by its absorption at 670 nm (62). S=C=S + ?H+ + ?e− → ?H2 S + C??

3.5.3. Carbonyl Sulfide

[29]

Carbonyl sulfide reduction is catalyzed by Mo-nitrogenase (equation [30]) to produce CO and H2 S (59). The CO produced is quantified as described for carbon dioxide reduction (see Section 3.5.1) and the H2 S as described for carbon disulfide reduction (see Section 3.5.2). S=C=O + 2H+ + 2e− → H2 S + C≡O

3.5.4. Thiocyanate

[28]

[30]

The Mo-nitrogenase-catalyzed reduction of thiocyanate yields hydrogen sulfide (H2 S) and hydrogen cyanide (HCN) (equation [31]), along with methane and ammonia (13), which presumably arise from subsequent reduction of product HCN. Typically (see Section 3.3.1), the amount of ammonia produced is greater than the amount of methane. However, when thiocyanate reduction is run under a N2 atmosphere, CH4 formation from the putative further reduction of the HCN product is inhibited (13), which contrasts with the observation that CH4 formation is not inhibited by N2 when HCN is the only substrate available (33). HS−C≡N + 2H+ + 2e− → H2 S + HC≡N

[31]

120

Newton and Dilworth

H2 S is quantified as described for catalyzed carbon disulfide reduction (see Section 3.5.2). HCN is microdistilled from acidified reaction mixtures into 0.1 M NaOH, treated with Chloramine-T, and reacted with a mixture of 3-methyl-1phenyl-2-pyrazolin-5-one and 3,3 -dimethyl-1,1 -diphenyl-[4,4 bi-2-pyrazolin-5,5 -dione] in pyridine before spectrophotometric measurement at 632 nm (13, 63). 3.5.5. Cyanate

3.6. Catalyzed Reactions Involving Hydrogen Only 3.6.1. Proton Reduction

Apparently, the substrate for nitrogenase is HOCN rather than [OCN]– based on plots of activity versus calculated cyanate concentration over a range of pH values (13). The major products are CO and ammonia (equation [32]). In the presence of a CO-trapping reagent (hemoglobin), which removes CO as a nitrogenase inhibitor, small amounts of CH4 are also detected. The production of CH4 suggests that HOCN has alternative modes of reduction, namely either to (i) CO plus NH3 or to (ii) H2 O plus HCN (equation [33]). As suggested for HSCN above, the putative product HCN could then be further reduced to CH4 and other HCN reaction products (see Section 3.3.1). CO was measured spectrophotometrically using the known absorption of carbonyl-hemoglobin (59–61) and ammonia after Conway diffusion and estimation via the indophenol method. HO=C=N + 2H+ + 2e− → C≡O + NH3

[32]

HO−C≡N + 2H+ + 2e− → H2 O + HC≡N

[33]

A basic property of all nitrogenases is the ability to catalyze the reduction of protons to dihydrogen (H2 ) (equation [34]), even in the presence of other reducible substrates. H2 evolution, therefore, represents a major activity of these enzymes and so needs careful quantification. The amount of evolved H2 is normally determined by gas chromatography using argon as the carrier gas, thermal conductivity (TCD) as the detection system, and a variety of column types, principally either 5A or 13A molecular sieve (Supelco, Bellefonte, PA). Details of this technique, which has been described many times, can be found elsewhere (10). Typically, pre-standardized gas mixtures are used for both chromatograph calibration and quantification of the product formed. 2H+ + 2e− → H2

3.6.2. HD Formation

[34]

When Mo-nitrogenase is incubated under assay conditions in the presence of 2 H2 (D2 ), HD is produced but only in the presence of N2 (equation [35]). This reaction was first investigated using excised soybean root nodules (64). A very low rate of HD formation was observed under a D2 /argon atmosphere (65), but this is

Assays of Nitrogenase Reaction Products

121

almost certainly due to low-level N2 contamination of the argon (66–68). Surprisingly, HD is not formed by a simple exchange reaction but consumes electrons and, even more surprisingly, its stoichiometry is only one electron per molecule of HD formed (22, 66). +N2

−→2HD 2H+ + 2e− + D2 ——

[35]

The analysis of HD formation involves mass spectrometric measurement of masses 2, 3, and 4 in the gas phase. To do so, the relative ionization efficiencies for these three masses must be determined for the particular mass spectrometer and under the conditions being used. The amount of HD produced can then be estimated directly from the peak heights for masses 2, 3, and 4. Alternatively, the amount of HD can be determined by reference to the argon peak either directly, when argon is present as a diluent, or indirectly by introducing a known amount of argon into the assay vial. In either scenario, the assay vials must be frozen to minimize the entry of water vapor into the mass spectrometer (see Note 7). 3.7. Other Assay Components

Although not actually products of substrate reduction, measurements of residual amounts of other components of nitrogenase assays pose their own analytical problems and are included here for completeness.

3.7.1. Dithionite

The amount of the electron donor (usually dithionite) used is important in determining whether the quantities of detected products account for all the electrons abstracted from dithionite during the incubation. The actual electron donor to nitrogenase is the radical anion SO2·− formed by homolytic cleavage of [O2 S-SO2 ]2− in solution. Three quantification methods have been used: (i) the spectrophotometric determination of dithionite at 315 nm (69); (ii) iodine titration (32, 33); and (iii) polarography (70, 71). The first and last techniques provide a continuous monitor of dithionite concentration, whereas the second method accounts for total dithionite used for product formation. The spectrophotometric method monitors changes in dithionite concentration by absorbance changes at 315 nm (350 nm when higher concentrations of dithionite are used) in assays run in anaerobic cuvettes (72). The titration method (33) starts by terminating the anaerobic assay with excess formaldehyde, which complexes the product sulfite so that it is not titratable with I2 /KI. Then, without opening the reaction vial, a calculated amount of 2 M acetate buffer is injected to bring the pH to 4.0, followed by adding starch solution. The residual dithionite is titrated to the blue starch endpoint using a micrometer syringe for accuracy. Assays can also be terminated with acid

122

Newton and Dilworth

before addition of excess formaldehyde, provided that sufficient acetate buffer is subsequently added (50). A modified termination procedure was used for nitrite reduction by nitrogenase (32). To prevent oxidation of dithionite by HNO2 , the assay was frozen in an ice-ethanol bath before excess formaldehyde was added to complex both dithionite and sulfite. 3.7.2. Titanium(III) Citrate

As an alternative reductant for nitrogenase, it provides a continuous assay based on conversion of titanium(III) citrate to titanium(IV) citrate. This oxidation can be followed by spectrophotometry at 340 nm and quantified by using a molar absorbance change of 0.73 mM–1 cm–1 (73).

3.7.3. Creatine

Most often, the creatine phosphate/creatine phosphokinase system is used to regenerate ATP and so prevent ADP accumulation during nitrogenase assays. The amount of creatine generated provides a direct measurement of ATP consumption during the assay. The simplest approach is to stop the reaction with either EDTA or acid and then pass the supernatant through small (2.5 × 0.6 cm diameter) columns of Dowex-1 anion exchange resin (Cl– form) in glass wool-plugged Pasteur pipettes (2) as described in Section 3.1.1 (ii). The ion-exchange treatment removes unused creatine phosphate, thus minimizing any complicating and unwanted chemical (particularly acid-catalyzed) hydrolysis. Creatine can then be assayed in the eluate by spectrophotometric measurement at 530 nm following reaction with alkaline α-naphthol and diacetyl (3). Color development is normally complete by 40 min, after which the mixture is diluted to minimize any further slow color development. This creatine assay works well for most nitrogenase substrates. One exception is hydrazine, where the eluate from the Dowex-1 (Cl– ) column contains an appreciable amount of the substrate, which reacts with diacetyl and drastically slows color development. This problem can be largely overcome by increasing the amount of diacetyl added by two to threefold, but this increases the background absorbance and so decreases the sensitivity, i.e., the slope of the calibration curve.

3.7.4. Phosphate

An alternative method of estimating ATP hydrolysis during nitrogenase assays is to measure the phosphate released. When labile phosphates (such as creatine phosphate) are being used, care must be taken with respect to the acid used to terminate the assays because it may cause non-enzymatic phosphate release and so compromise the analysis. Moreover, most phosphate assays depend on formation of a blue color from reduction of acid phosphomolybdate (74), which can also hydrolyze sensitive phosphate esters (75). It is, therefore, clearly advisable to (i) terminate nitrogenase assays containing creatine phosphate with EDTA if at all possible and (ii) use the least acidic conditions for the

Assays of Nitrogenase Reaction Products

123

phosphate assay including a reductant, like Fe(II), which will work under mildly acid conditions (75). Other phosphate analysis methods used include the Fiske and Subbarow procedure (14, 74) and that described by Ottolenghi (76, 77). In general, if the creatine phosphate/creatine phosphokinase system is being used to regenerate ATP, creatine estimation is clearly preferred over the phosphate assay. For extremely sensitive and continuous measurements of phosphate release, as in pre-steady-state assays of ATP hydrolysis and its timing (78), specialist techniques have been developed. These methods depend on the phosphate-induced change in fluorescence emission at 464 nm of a fluorophore, N-[2-(1-maleimidyl)-ethyl]-7-(diethylamino)coumarin3-carboxamide, bound via a cysteine residue introduced by a directed mutation into the phosphate-binding protein from Escherichia coli (79). Phosphate binding results in a fivefold increase in the fluorescence emission at 464 nm, thus providing a continuous assay for phosphate release. To further increase the sensitivity of the assay, contaminating background phosphate can be removed by using a phosphate “mop” system of 1 mM 7-methylguanine and purine nucleotide phosphorylase (79). The level of phosphorylase added is adjusted so that it removes the background phosphate but does not interfere with the rapid increase in fluorescence resulting from binding of phosphate released from nitrogenase to the fluorophore/phosphate-binding protein complex. 3.8. Concluding Remarks

Our goal for this chapter was to emphasize the principles behind the assays used to measure the various products of nitrogenase action on a very wide variety of substrates rather than to reiterate the protocols themselves. We, of course, direct the reader to the original papers where the protocols are described in detail. As indicated at the beginning of this chapter, a number of the methods described were developed and used by a single researcher or research group and many have not been tested by others. We only guarantee our own developments!

4. Notes 1. As a technique, either mode of microdistillation has a number of drawbacks. It inserts a delay of at least 2 h between nitrogenase assay and ammonia determination. The glassrod technique requires considerable manual dexterity to avoid touching the acid-drop-laden rod onto the neck of the assay vial during removal. The Conway dish method

124

Newton and Dilworth

requires care in sealing the dishes and in preventing alkali “creep” into the central well. Further, the authors’ experience is that there is often a considerable spread in the data for replicate distillations, exceeding that for assays using other separation techniques, but why this is so remains unclear. 2. Not all equations [4]–[7] apply to all nitrogenases; product balances vary among them. For example, some variant Mo-nitrogenases catalyze the reduction of azide/HN3 (equations [4] and [5]) but are incapable of catalyzing the eight-electron reaction (equation [7]) because the NH3 produced is the equivalent to the sum of the N2 and the N2 H4 produced with no “excess” NH3 formed (26). Wild-type Mo-nitrogenases do not suffer serious inhibition by moderate concentrations of azide, but azide can inhibit electron transport through both variant Mo-nitrogenases (either without affecting the rate of ATP hydrolysis or by increasing it (18)) and V-nitrogenase (27). 3. The disadvantage of the mini-column protocol is that many fractions have to be measured for both NH3 and CH3 NH2 content for each HCN reduction assay performed. An alternative method, using 2,4-dinitrofluorobenzene (42), has been used for estimating methylamine produced from cyanamide (41), but apparently has not been used with HCN reduction assays. 4. Before use, contaminating ammonia and methylamine must be removed from cyanamide by passing it over a Dowex-50 (H+ form) column at pH 4.5. 5. Acetylene is usually generated by the action of water on calcium carbide and stored over water. C2 D2 can be prepared by injecting D2 O into an evacuated bottle containing a preweighed sample of calcium carbide (CaC2 ) (49). Due care must be taken to ensure that the bottle is of sufficient capacity to withstand the C2 D2 pressure generated! A preliminary experiment to determine the actual volume of gas liberated from the CaC2 is obviously desirable. 6. The purification protocol involves first exposing the ethylene to a solution of mercuric perchlorate with which it complexes, then pumping off any ethane contaminant, followed by regenerating highly purified ethylene by destroying the mercuric-ethylene complex using lithium chloride (50). 7. Importantly, if argon is being used as a reference, the assays must be frozen in dry ice/ethanol rather than in liquid nitrogen to prevent condensation of the argon (26).

Assays of Nitrogenase Reaction Products

125

References 1. Pham DN, Burgess BK (1993) Nitrogenase reactivity: Effects of pH on substrate reduction and CO inhibition. Biochemistry 32:13725–13731 2. Dilworth MJ, Eldridge ME, Eady RR (1992) Correction for creatine interference with the direct indophenol measurement of ammonia in steady-state nitrogenase assays. Anal Biochem 207:6–10 3. Ennor AH (1957) Determination and preparation of N-phosphates of biological origin. Methods Enzymol 3:850–856 4. Dilworth MJ, Eady RR (1991) Hydrazine is a product of dinitrogen reduction by the vanadium-nitrogenase from Azotobacter chroococcum. Biochem J 277: 465–468 5. Dilworth MJ, Thorneley RNF (1981) Nitrogenase of Klebsiella pneumoniae. Hydrazine is a product of azide reduction. Biochem J 193:971–983 6. Fawcett JK, Scott JE (1960) A rapid and precise method for the determination of urea. Clin Pathol 13:156–159 7. Thorneley RNF, Lowe DJ (1984) The mechanism of Klebsiella pneumoniae nitrogenase action. Pre-steady-state kinetics of an enzyme-bound intermediate in nitrogen reduction and of ammonia formation. Biochem J 224:887–894 8. Chaykin S (1969) Assay of nicotinamide deamidase. Determination of ammonia by the indophenol reaction. Anal Biochem 31:375–382 9. Corbin JL (1984) Liquid chromatographicfluorescence determination of ammonia from nitrogenase reactions: A 2-min assay. Appl Environ Microbiol 47:1027–1030 10. Burris RH (1972) Nitrogen fixation – assay methods and techniques. Methods Enzymol 24B:415–431 11. Liang J, Burris RH (1989) Nitrous oxide reduction and HD formation by nitrogenase from a nifV mutant of Klebsiella pneumoniae. J Bacteriol 171:3176–3180 12. Conway EJ (1960) Microdiffusion Analysis and Volumetric Error. Crosby, Lock and Sons, London 13. Rasche ME, Seefeldt LC (1997) Reduction of thiocyanate, cyanate, and carbon disulfide by nitrogenase: Kinetic characterization and EPR spectroscopic analysis. Biochemistry 36:8574–8585 14. Erickson JA, Nyborg AC, Johnson JL et al (1999) Enhanced efficiency of ATP hydrolysis during nitrogenase catalysis utilizing reductants that form the all-ferrous redox state of the Fe protein. Biochemistry 38:14279–14285

15. Beecher GR, Whitten GR (1970) Ammonia determination: Reagent modification and interfering compounds. Anal Biochem 36:243–246 16. O’Donovan DJ (1971) Inhibition of the indophenol reaction in the spectrophotometric determination of ammonia. Clin Chim Acta 32:59–61 17. Maryan PS, Vorley WT (1979) An improved spectrophotometric method for the determination of ammonia with particular relevance to in vitro nitrogenase activity. Lab Practice 28:251–252 18. Dilworth MJ, Fisher K, Kim C-H et al (1998) Effects on substrate reduction of substitution of histidine-195 by glutamine in the α-subunit of the MoFe protein of Azotobacter vinelandii nitrogenase. Biochemistry 37:17495–17505 19. Dilworth MJ, Fisher K (1998) Elimination of creatine interference with the indophenol measurement of NH3 produced during nitrogenase assays. Anal Biochem 256: 242–244 20. Bulen WA (1976) Nitrogenase from Azotobacter vinelandii and reactions affecting mechanistic interpretations. In: Newton WE, Nyman CJ (eds) Proceedings of the First International Symposium on Nitrogen Fixation, vol. 1, pp. 177–186. Washington State University Press, Pullman 21. Davis LC (1980) Hydrazine as a substrate and inhibitor of Azotobacter vinelandii nitrogenase. Arch Biochem Biophys 204:270–276 22. Wherland S, Burgess BK, Stiefel EI et al (1981) Nitrogenase reactivity: Effects of component ratio on electron flow and distribution during nitrogen fixation. Biochemistry 20:5132–5140 23. Hardy RWF, Knight E (1967) ATPdependent reduction of azide and hydrogen cyanide by nitrogen-fixing enzymes of Azotobacter vinelandii and Clostridium pasteurianum. Biochim Biophys Acta 139:69–90 24. Schöllhorn R, Burris RH (1967) Reduction of azide by the N2 -fixing enzyme system. Proc Natl Acad Sci USA 57:1317–1323 25. Rubinson JF, Burgess BK, Corbin JL et al (1985) Nitrogenase reactivity: Azide reduction. Biochemistry 24:273–283 26. Fisher K, Dilworth MJ, Newton WE (2000) Differential effects on N2 binding and reduction, HD formation, and azide reduction with α-195His - and α-191Gln -substituted MoFe proteins of Azotobacter vinelandii nitrogenase. Biochemistry 39:15570–15577 27. Fisher K, Dilworth MJ, Newton WE (2006) Azotobacter vinelandii vanadium

126

28.

29.

30. 31.

32. 33. 34.

35.

36.

37. 38. 39. 40. 41. 42.

Newton and Dilworth nitrogenase: Formaldehyde is a product of catalyzed HCN reduction and excess ammonia arises directly from catalyzed azide reduction. Biochemistry 45:4190–4198 Clusius K, Hürzeler H (1953) Reactions with nitrogen15 . X. Reduction and oxidation of hydrazoic acid. Helv Chim Acta 36: 1326–1332 Lockshin A, Burris RH (1965) Inhibitors of nitrogen fixation in extracts from Clostridium pasteurianum. Biochim Biophys Acta 111: 1–10 Jensen BB, Burris RH (1986) Nitrous oxide as a substrate and as a competitive inhibitor of nitrogenase. Biochemistry 25:1083–1088 Liang J, Burris RH (1988) Interactions among nitrogen, nitrous oxide and acetylene as substrates and inhibitors of nitrogenase from Azotobacter vinelandii. Biochemistry 27:6726–6732 Vaughn SA, Burgess BK (1989) Nitrite, a new substrate for nitrogenase. Biochemistry 28:419–424 Li J, Burgess BK, Corbin JL (1982) Nitrogenase reactivity: Cyanide as substrate and inhibitor. Biochemistry 21:4393–4402 Kelly M, Postgate JR, Richards RL (1967) Reduction of cyanide and isocyanide by nitrogenase of Azotobacter chroococcum. Biochem J 102:1C–3C Fisher K, Dilworth MJ, Kim C-H et al (2000) Azotobacter vinelandii nitrogenases with substitutions in the FeMo-cofactor environment of the MoFe protein: Effects of acetylene or ethylene on interactions with H+ , HCN, and CN– . Biochemistry 39:10855–10865 Lin J-K, Lai C-C (1980) High performance liquid chromatographic determination of naturally occurring primary and secondary amines with dabsyl chloride. Anal Chem 52:630–635 Spies JR (1957) Colorimetric methods for amino acids. Methods Enzymol 3: 467–477 Nash T (1953) The colorimetric estimation of formaldehyde by means of the Hantzsch reaction. Biochem J 55:416–421 Burns RC, Hardy RWF (1975) Nitrogen Fixation in Bacteria and Higher Plants. Springer, Berlin Fuchsman WH, Hardy RWF (1972) Nitrogenase-catalyzed acrylonitrile reductions. Bioinorg Chem 1:195–213 Miller RW, Eady RR (1988) Cyanamide: A new substrate for nitrogenase. Biochim Biophys Acta 952:290–296 Dubin PT (1960) Assay and characterization of amines by means of

43.

44.

45.

46. 47.

48.

49.

50.

51.

52.

53.

54.

2,4-dinitrofluorobenzene. J Biol Chem 235: 783–786 Kelly M (1968) The kinetics of the reduction of isocyanides, acetylenes and the cyanide ion by nitrogenase preparations from Azotobacter chroococcum and the effects of inhibitors. Biochem J 107:1–6 Rubinson JF, Corbin JL, Burgess BK (1983) Nitrogenase reactivity: Methyl isocyanide as substrate and inhibitor. Biochemistry 22:6260–6268 Dilworth MJ (1966) Acetylene reduction by nitrogen-fixing preparations from Clostridium pasteurianum. Biochim Biophys Acta 127:285–294 Schöllhorn R, Burris RH (1967) Acetylene as a competitive inhibitor of N2 fixation. Proc Natl Acad Sci USA 58:213–216 Dilworth MJ, Eldridge ME, Eady RR (1993) The molybdenum and vanadium nitrogenases of Azotobacter chroococcum: Effect of elevated temperature on nitrogen reduction. Biochem J 289:395–400 Scott DJ, Dean DR, Newton WE (1992) Nitrogenase-catalyzed ethane production and CO-sensitive hydrogen evolution from MoFe proteins having amino acid substitutions in an α-subunit FeMo cofactor binding domain. J Biol Chem 267:20002–20010 Fisher K, Dilworth MJ, Kim C-H et al (2000) Azotobacter vinelandii nitrogenases containing altered MoFe proteins with substitutions in the FeMo-cofactor environment: Effects on the catalyzed reduction of acetylene and ethylene. Biochemistry 39:2970– 2979 Ashby GA, Dilworth MJ, Thorneley RNF (1987) Klebsiella pneumoniae nitrogenase. Inhibition of hydrogen evolution by ethylene and the reduction of ethylene to ethane. Biochem J 247:547–554 Dilworth MJ, Eady RR, Eldridge ME (1988) The vanadium nitrogenase of Azotobacter chroococcum. Reduction of acetylene and ethylene to ethane. Biochem J 249: 745–751 Lin-Vien D, Fateley WG, Davis LC (1989) Estimation of nitrogenase activity in the presence of ethylene biosynthesis by use of deuterated acetylene as a substrate. Appl Environ Microbiol 55:354–359 Hardy RWF, Holsten RD, Jackson EK et al (1968) The acetylene-ethylene assay for N2 fixation: Laboratory and field evaluation. Plant Physiol 43:1185–1207 Mayer SM, Niehaus WG, Dean DR (2002) Reduction of short chain alkynes by a nitrogenase α-70Ala-substituted MoFe protein. J Chem Soc, Dalton Trans 5:802–807

Assays of Nitrogenase Reaction Products 55. Burns RC, Hardy RWF, Phillips WD (1975) Azotobacter nitrogenase: Mechanism and kinetics of allene reduction. In: Stewart WDP (ed) Nitrogen Fixation in Free-living Microorganisms, pp. 447–452. Cambridge University Press, Cambridge 56. McKenna CE, McKenna MC, Higa MT (1976) Chemical probes of nitrogenase. I. Cyclopropene. Nitrogenase-catalyzed reduction to propene and cyclopropane. J Am Chem Soc 98:4657–4659 57. Gemoets JP, Bravo M, McKenna CE et al (1989) Reduction of cyclopropene by NifV– and wild-type nitrogenases from Klebsiella pneumoniae. Biochem J 258:487–491 58. McKenna CE, McKenna MC, Huang CW (1979) Low stereoselectivity in methylacetylene and cyclopropene reductions by nitrogenase. Proc Natl Acad Sci USA 76: 4773–4777 59. Seefeldt LC, Rasche ME, Ensign SA (1995) Carbonyl sulfide and carbon dioxide as new substrates, and carbon disulfide as a new inhibitor, of nitrogenase. Biochemistry 34:5382–5389 60. Bonam D, Murrell SA, Ludden PW (1984) Carbon monoxide dehydrogenase from Rhodospirillum rubrum. J Bacteriol 159: 693–699 61. Kumar M, Lu WP, Ragsdale SW (1994) Binding of carbon disulfide to the site of acetyl-CoA synthesis by the nickel-iron-sulfur protein, carbon monoxide dehydrogenase, from Clostridium thermoaceticum. Biochemistry 33:9769–9777 62. Ensign SA (1995) Reaction of carbon monoxide dehydrogenase from Rhodospirillum rubrum with carbon dioxide, carbonyl sulfide and carbon disulfide. Biochemistry 34:5372–5381 63. Hargis LG (1978) Determination of carbon. In: Boltz DF, Howell JA (eds) Colorimetric Determination of Non-Metals, pp. 57–82. Wiley, New York, NY 64. Hoch GE, Schneider KC, Burris RH (1960) Hydrogen evolution and exchange, and conversion of N2 O to N2 by soybean root nodules. Biochim Biophys Acta 37: 273–279 65. Burgess BK, Wherland S, Newton WE et al (1981) Nitrogenase reactivity: Insight into the nitrogen-fixing process through hydrogen-inhibition and HD-forming reactions. Biochemistry 20:5140–5146 66. Guth JH, Burris RH (1983) Inhibition of nitrogenase-catalyzed ammonia formation by hydrogen. Biochemistry 22:5111–5122 67. Li JL, Burris RH (1983) Influence of pN2 and pD2 on HD formation by various nitrogenases. Biochemistry 22:4472–4480

127

68. Jensen BB, Burris RH (1985) Effect of high pN2 and high pD2 on ammonia production, hydrogen evolution, and hydrogen deuteride formation by nitrogenases. Biochemistry 24:1141–1147 69. Ljones T, Burris RH (1972) Continuous spectrophotometric assay for nitrogenase. Anal Biochem 45:448–452 70. Watt GD, Burns A (1977) Kinetics of dithionite ion utilization and ATP hydrolysis for reactions catalyzed by the nitrogenase complex from Azotobacter vinelandii. Biochemistry 16:264–270 71. Johnson JL, Tolley AM, Erickson JA et al (1996) Steady-state kinetics studies of dithionite utilization, component protein interaction, and the formation of an oxidized iron protein intermediate during Azotobacter vinelandii nitrogenase catalysis. Biochemistry 35:11336–11342 72. Shah VK, Davis LC, Brill WJ (1972) I. Repression and derepression of the ironmolybdenum and iron proteins of nitrogenase in Azotobacter vinelandii. Biochim Biophys Acta 256:498–511 73. Seefeldt LC, Ensign SA (1994) A continuous, spectrophotometric assay for nitrogenase using the reductant titanium(III) citrate. Anal Biochem 221:379–386 74. Fiske CH, Subbarow Y (1925) The colorimetric determination of phosphorus. J Biol Chem 66:375–400 75. Taussky HH, Shorr E, Kurzmann G (1953) A microcolorimetric method for the determination of inorganic phosphorus. J Biol Chem 202:675–685 76. Ottolenghi P (1975) The reversible delipidation of a solubilized sodium-plus-potassium ion-dependent adenosine triphosphatase from the salt gland of the spiny dogfish. Biochem J 151:61–66 77. Thorneley RNF, Ashby GA, Julius C et al (1991) Nitrogenase of Klebsiella pneumoniae. Reversibility of the reductantindependent magnesium-ATP-cleavage reaction is shown by magnesium-ADP-catalyzed phosphate/water oxygen exchange. Biochem J 277:735–741 78. Lowe DJ, Ashby GA, Brune M et al (1995) ATP hydrolysis and energy transduction by nitrogenase. In: Tikhonovich IA, Provorov NA, Romanov VI, Newton WE (eds) Nitrogen Fixation: Fundamentals and Applications, pp. 103–108. Kluwer, Dordrecht 79. Brune M, Hunter JL, Corrie JE et al (1994) Direct, real-time measurement of rapid inorganic phosphate release using a novel fluorescent probe and its application to actomyosin subfragment 1 ATPase. Biochemistry 33:8262–8271

Chapter 9 Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase Jürgen Moser and Markus J. Bröcker Abstract Nitrogenase-like dark operative protochlorophyllide oxidoreductase (DPOR) is involved in the biosynthesis of chlorophylls and bacteriochlorophylls in gymnosperms, ferns, algae, and photosynthetic bacteria. During protochlorophyllide (Pchlide) reduction, the homodimeric subunit ChlL2 of DPOR transfers electrons on the corresponding heterotetrameric catalytic subunit (ChlN/ChlB)2 . Although DPOR shares significant amino acid sequence homology to the nitrogenase system, only the initial catalytic steps of DPOR resemble nitrogenase catalysis. Investigation of the cyanobacterial DPOR from Prochlorococcus marinus indicated that subcomplex ChlL2 is functioning as an ATP-dependent switch protein, triggering the transient interaction of ChlL2 and (ChlN/ChlB)2 . This dynamic subunit interplay is responsible for the transfer of a single electron from the [4Fe–4S] cluster of ChlL2 onto a second [4Fe–4S] cluster located on (ChlN/ChlB)2 . However, the second part of DPOR catalysis is unrelated to nitrogenase catalysis, since no molybdenum-containing cofactor or a P-cluster equivalent is employed. Instead, two consecutive electron transfer steps are mediated via the [4Fe–4S] cluster of (ChlN/ChlB)2 , resulting in the reduction of the conjugated ring system of the substrate molecule Pchlide (Figs. 5.1a and 5.2). Key words: Dark operative protochlorophyllide oxidoreductase (DPOR), nitrogenase-like enzyme, chlorophyll biosynthesis, dynamic switch protein.

1. Introduction The heterologous overproduction of individual DPOR subunits in Escherichia coli was an important breakthrough for the subsequent biochemical and biophysical analysis of DPOR ([FeS] cluster composition, gel permeation chromatographic analysis, electron paramagnetic resonance spectroscopy). Therefore, the bicistronic overproduction of subunit ChlN as an N-terminal fusion protein with glutathione-S-transferase (GST) M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_9, © Springer Science+Business Media, LLC 2011

129

Moser and Bröcker

a

pGEX-PmNB Ptac

rbs

BamHI

gst

pGEX-PmL

rbs

chlN

b

NotI

SalI

Chlide 665 nm 0.05

chlB A

NotI

BamHI

Pchlide 626 nm

relative absorption

130

B

Ptac

rbs

C

chlL

gst

D E

PreScissionTM Protease cleavage site

600

d

c

620

640

ChlL

680

700 [nm]

(ChlN/ChlB)2

(ChlN/ChlB)2(ChlL2)2

116

660

ChlL

[4Fe−4S]

66.2

ChlB

ChlB

ChlN

absorption [280 nm]

[4Fe−4S]

45 33

1

2

3

[4Fe−4S] ChlN

0.05

ChlL

ChlN

[4Fe−4S] [4Fe−4S] ChlN

ChlB

ChlB

ChlL

[4Fe−4S]

0.02

4 75

125

175

225

[ml]

Fig. 9.1. Constructs for recombinant DPOR production, DPOR activity assay, analyses of ternary DPOR complexes by SDS-PAGE, and gel permeation chromatography. (a) Schematic representation of the employed constructs for the heterologous overproduction of subunits ChlN/ChlB (bicistronic) and ChlL. (b) DPOR activity assay using dithionite as an artificial reducing agent in the presence of an ATP-regenerating system. Trace A, DPOR assay containing 100 pmol purified (ChlN/ChlB)2 and 200 pmol ChlL2 from P. marinus; traces B and C, negative control without (ChlN/ChlB)2 (B) and without ChlL2 (C); trace D, negative control without ATP; trace E, negative control without dithionite. (c) Ternary DPOR complex formation in the presence of Pchlide. SDS-PAGE analyses of ternary DPOR complexes formed in the presence of MgADP153 (lane 4). Lanes 1, 2: purified subcomplexes ChlL (lane AlF− 2 4 (lane 3) or in the presence of mutant protein ChlLLeu 1) and (ChlN/ChlB)2 (lane 2). The relative molecular masses (× 1,000) of the employed marker proteins are indicated. (d) Preparative gel permeation chromatography of (ChlN/ChlB)2 [thin] and the ternary DPOR complex [bold]. 0.08 μmol of purified ternary DPOR complex and 0.26 μmol of the (ChlN/ChlB)2 subcomplex were analyzed on a Superdex 200 HR 26/60 column under anaerobic conditions (95% N2 , 5% H2 , <1 ppm O2 ) using a flow rate of 1.5 mL/min and monitoring the absorbance at 280 nm. The oligomeric architecture of the individual complexes is indicated.

in combination with ChlB based on a standard E. coli expression vector (pGEX-6P-1, GE Healthcare) was established. For this purpose it was of importance that the sequence of an E. colispecific ribosomal binding site was implemented during PCR amplification of the ChlB gene (1, 2). The chlN gene was then cloned into the BamHI and SalI sites of the vector. Subsequently, the chlB gene was then ligated into the SalI and NotI sites of the obtained vector (Fig. 9.1a). The resulting construct (pGEX-PmNB) then allowed for the protein production and copurification of the obtained GST-ChlN fusion protein in complex with stoichiometric amounts of the untagged ChlB protein. High amounts of the tight complex (ChlN/ChlB)2 can be easily purified using standard GST affinity chromatography followed

Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase

131

by on-column cleavage with PreScission protease according to the manufacturer’s instructions (GE Healthcare). The chlL gene was separately cloned in the identical pGEX-6P-1 vector. The obtained construct (pGEX-PmL) (Fig. 9.1a) allowed for the analogous purification of the GST-ChlL fusion protein, resulting in the elution of the ChlL2 complex after on-column cleavage (2). This type of affinity chromatography makes use of simple gravity flow columns, which is a good prerequisite for the efficient purification of the DPOR subcomplexes under strict anaerobic conditions in an anaerobic chamber. For the reconstitution of DPOR activity in vitro in the standard DPOR activity assay, dithionite is used as an artificial reducing agent in the presence of high concentrations of ATP and an ATP-regenerating system consisting of creatine phosphate and creatine phosphokinase. Despite using a chemical reducing agent instead of the natural electron donor (which is a ferredoxin), DPOR catalysis still is dependent on both subunits (ChlN/ChlB)2 and ChlL2 (Fig. 9.1b). To obtain optimal DPOR activity it is beneficial to evaluate the optimal ratio of ChlL2 versus (ChlN/ChlB)2 in preliminary pilot experiments. Therefore, the concentration of ChlL2 is titrated at a constant concentration of (ChlN/ChlB)2 . Only recently, non-productive ATPase activity for the reconstituted enzyme was shown in the absence of the substrate Pchlide, but also under conditions of the standard DPOR activity assay (3). This might indicate that the proper stoichiometry in the DPOR reconstitution assay is important to repress such non-productive ATPase activity in vitro. It was shown that the ‘trapping’ of the ternary DPOR protein–protein interaction is an efficient strategy to investigate the electron transfer processes of the DPOR catalytic redox cycle (3). Such experiments make use of analogs of the co-substrate ATP, representing the transition state of ATP hydrolysis. MgADP in combination with AlF− 4 has been shown to efficiently stabilize the transient DPOR protein–protein interaction. Furthermore, it was shown that this ternary interaction is additionally enforced in the presence of the Pchlide substrate. For the preparative formation and purification of ternary DPOR complexes, the following strategy was employed. Immobilization of ChlL2 was achieved by using the GST-ChlL fusion protein on glutathione-S-transferase column. Subsequently, ternary complex formation was initiated by the addition of an excess of the purified (ChlN/ChlB)2 protein in the presence of Pchlide. After an additional washing step, the purified DPOR complex (ChlN/ChlB)2 (ChlL2 )2 with an octameric stoichiometry was liberated from the column by PreScission protease treatment (Fig. 9.1c, d).

132

Moser and Bröcker

2. Materials 2.1. Protein Production

1. Luria-Bertani (LB) medium: 5.0 g/L yeast extract, 10.0 g/l tryptone, and 1 0.0 g/L NaCl. 2. Medium additives: (a) 100 mM Fe(III)citrate, autoclaved stock, stored at room temperature. (b) 100 mM L-cysteine, sterile filtered stock, stored at 4◦ C. (c) 100 mg/mL ampicillin stock, dissolved in water, stored at –20◦ C. (d) 34 mg/mL chloramphenicol stock, dissolved in ethanol (99.8%), stored at –20◦ C. (e) 1 M isopropyl-β-D-thiogalactopyranoside (IPTG) stock, dissolved in water, stored at –20◦ C. 3. Erlenmeyer flasks (250 mL, 1,000 mL) with chicanes. 4. Centrifugation bottles (1,000 mL) equipped with gas-tight cap assembly (O-rings) (Beckman Coulter, Palo Alto, CA).

2.2. Protein Purification

1. Standard Hepes buffer, anaerobic: (a) 100 mM Hepes/NaOH pH 7.5, from autoclaved 1 M stock. (b) 150 mM NaCl, from autoclaved 5 M stock. (c) 10 mM MgCl2 , from autoclaved 1 M stock. (stocks are stored at room temperature). 2. Distilled water, autoclaved, anaerobic. 3. 1 M sodium dithionite stock, dissolved in water, stored at –20◦ C. R pressure cell (40 kpsi, 35 mL) 4. Standard FRENCH (Thermo Scientific), 125 mL anaerobic septum bottles, silicone flexible tube, Luer fittings, and cannules.

5. Polycarbonate ultracentrifugation bottles (70 mL) equipped with gas-tight cap assembly (O-rings) (Beckman Coulter). 6. 25 mL Poly-Prep gravity flow column (Bio-Rad Laboratories Inc., Hercules, CA). R Glutathione Agarose 4B (Macherey–Nagel, Dül7. Protino men, Germany).

8. PreScissionTM Protease (GE Healthcare) stored at –20◦ C. 9. 100 mM reduced glutathione dissolved in anaerobic, standard Hepes buffer, freshly prepared. 10. Bradford reagent (Bio-Rad) stored at 4◦ C.

Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase

133

11. Anaerobic workstation: flexible glove chamber with forming gas atmosphere (95% N2 , 5% H2 ) equipped with a platinum catalyst (flexible vinyl chamber, COY Laboratory Products Inc., Grass Lake, MI). R stirred ultrafiltration cell (Millipore, Bedford 12. Amicon MA) equipped with 30,000 Da exclusion limit polyethersulfone membrane.

13. FPLC Äkta purifier, Superdex 200 26/60 column (GE Healthcare, Umea, Sweden). 14. Instrument for automated evaporation and nitrogen flushing of buffers (self-construction). 2.3. Activity Tests

1. 20 μM (ChlN/ChlB)2 (4.2 mg/ml) and 40 μM ChlL2 (2.6 mg/ml) in standard Hepes buffer; 500 μl aliquots stored in 1.5 ml HPLC vials in liquid nitrogen. 2. 2 mM Pchlide stock, dissolved in dimethyl sulfoxide (DMSO) (100%); Pchlide is produced with a Rhodobacter capsulatus mutant strain according to (4). 3. 1 M sodium dithionite stock dissolved in water. 4. 100 mM MgATP, 200 μl aliquots dissolved in anaerobic standard Hepes buffer. 5. 100 mM creatine phosphate, 200 μL aliquots dissolved in anaerobic standard Hepes buffer. 6. 20 units/3 μL creatine phosphokinase from rabbit muscle, 50 μL aliquots dissolved in anaerobic standard Hepes buffer. (Compounds 2–6 are stored in 1.5 mL HPLC vials at –20◦ C). 7. Acetone (100%). 8. Anaerobic box with forming gas atmosphere (95% N2 , 5% H2 ) equipped with platinum catalysts (MACS-MG-1000 anaerobic workstation, Don Whitley Scientific, West Yorkshire, UK). This system allows for sample manipulation in the absence of gloves. 9. Thermomixer compact (Eppendorf, Germany). R micro cuvettes (Hellma, 10. Sealed quartz SUPRASIL Müllheim, Germany).

11. V-550 UV/vis spectrophotometer Umstadt, Germany). 2.4. DPOR Complex Formation

(JASCO,

Groß-

1. 100 mM MgADP, 200 μL aliquots, stored at –20◦ C. 2. 50 mM AlCl3 , 15 mL, stored at 17◦ C. 3. 200 mM NaF, 50 mL, stored at 17◦ C. (Compounds are dissolved in anaerobic standard Hepes buffer).

134

Moser and Bröcker

3. Methods For the heterologous overproduction of the (ChlN/ChlB)2 complex (plasmid pGEX-PmNB) or the ChlL2 complex (plasmid pGEX-PmL) in E. coli (Fig. 9.1a), it is important to provide culture conditions that allow for efficient protein and subsequent [FeS] cluster biogenesis in vivo. For this purpose, the E. coli BL21 CodonPlusTM (DE3)-RIL strain additionally encodes for tRNA genes that most frequently limit translation of heterologous proteins in E. coli. Furthermore, cell growth was restricted by using a reduced growth temperature of 25◦ C. Additionally, the standard LB medium was supplemented by the addition of Fe(III)citrate and L-cysteine as a source of iron and sulfur. These two additives have significant influence on the entire assembly of the [4Fe–4S] clusters. The aerobic cultivation of the cell culture ensures protein yield in the milligram scale without hampering the activity of the purified protein. However, due to the highly oxygen-labile nature of the [FeS] clusters of DPOR, cell harvesting and all subsequent procedures have to be performed under strict anaerobic conditions. DPOR activity is reconstituted from purified (ChlN/ChlB)2 and ChlL2 subunits. Great care has to be taken to determine an optimal ratio of the two protein compounds. This ratio might differ in vitro from the optimal stoichiometry of 2 mol ChlL2 per mol (ChlN/ChlB)2 complex due to minor amounts of uncompleted [FeS] centers in the individual protein preparations. Based on the results of these pilot experiments, a robust assay system is obtained which allows for the variation of individual reaction parameters to characterize principles of DPOR catalysis (1). Approaches to generate ternary DPOR complexes comprising subunits (ChlN/ChlB)2 and ChlL2 make use of nucleotidedependent conformational rearrangements of ChlL2 that result in the trapping of the transient DPOR protein–protein interaction (3). MgADP-AlF− 4 is a compound that mimics the trigonal bipyramidal transition state of γ-phosphate hydrolysis of ATP. However, AlF− 4 is formed in situ from AlCl3 and NaF which also results in the formation of various hydroxide complexes. Therefore, the order of compound addition has significant effect on the equilibrium of the resulting aluminum compounds. The identical octameric (ChlN/ChlB)2 (ChlL2 )2 complex can also be analyzed by using a mutant protein of ChlL in the absence of any nucleotide cofactors. Deletion of Leucine153 in the switch-II sequence blocks the dynamic switch protein in a conformation with a high affinity for subcomplex (ChlN/ChlB)2 (3).

Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase

3.1. Protein Production

135

1. Starter cultures: Two 250 mL Erlenmeyer chicane flasks containing 150 mL LB medium supplemented with 100 μg/mL ampicillin and 34 μg/mL chloramphenicol are inoculated from glycerol stocks of E. coli BL21 CodonPlusTM (DE3)-RIL cells carrying either the plasmid pGEX-PmNB or pGEX-PmL. The cultures are incubated at 37◦ C and 150 rpm on a shaker overnight (approximately 17 h). 2. Production culture: Depending on the available centrifuge capacity up to 30 1,000 mL Erlenmeyer chicane flasks containing 500 mL LB medium supplemented with 1 mM L -cysteine, 1 mM Fe(III)citrate, 100 μg/mL ampicillin, and 34 μg/mL chloramphenicol are inoculated with 5 mL of the respective starter cultures to produce the (ChlN/ChlB)2 or the ChlL2 complexes, respectively. These cultures are incubated at 37◦ C and agitated at 150 rpm until they reach an optical density of OD578 ∼0.4, then they are shifted to 25◦ C and protein production is initiated by addition of 25 μM IPTG at an OD578 ∼0.5 (about 20–30 min later) for approximately 17 h. 3. Cell harvest: Starting with cell harvesting, all procedures have to be carried out under strictly anaerobic conditions to ensure the integrity of the oxygen-labile [FeS] clusters of DPOR. Anaerobic conditions of the employed buffer solutions have to be carefully adjusted by flushing with nitrogen gas. Repeated cycles of evacuation and nitrogen saturation automatically applied to the buffer solutions over a time range of 3 h are sufficient to ensure anaerobic conditions in 1,000 mL stock solutions of the required compounds. For cell harvesting, the production cultures are transferred to 1,000 mL centrifugation bottles and supplemented with 1.7 mM sodium dithionite. Subsequently these bottles are closed with a gas-tight cap assembly system and are incubated at 17◦ C for 3 h without agitation to allow the entire formation of [4Fe–4S] clusters (see Note 1). Following centrifugation for 12 min at 3,000 × g at 4◦ C the enclosed centrifuge bottles are transferred to a flexible vinyl chamber anaerobic workstation operated at 17◦ C. The supernatant is discarded before the sedimented E. coli cells are resuspended with anaerobic standard Hepes buffer (25 mL/pellet) and transferred to anaerobic septum bottles. These are tightly closed and allow storage of the resuspended cells at –20◦ C for at least 1 year without loosing DPOR activity.

3.2. Protein Purification

1. The suspended cell paste in the anaerobic septum bottles is thawed in the anaerobic workstation for 2 h. Then a

136

Moser and Bröcker

magnetic bar is added and the cell suspension is carefully homogenized on a magnetic stirrer for 5 min. R pressure cell 2. Cell disruption: In parallel, the FRENCH ◦ is cooled to 4 C and then transferred into the anaerobic workstation. This cell is then manually filled (at an inversed position) with up to 35 mL of the cell suspension and subsequently assembled according to the manufacturer’s instructions. The outlet valve is connected to another 125 mL anaerobic bottle via a flexible silicone tube fitted with a cannule. The resulting closed and air-tight system is then passed through the airlock and cell disruption is performed by a single passage through the pressure cell at approximately 10,000 psi (see Note 2).

3. Ultracentrifugation: The homogenized cells in the anaerobic septum bottle are transferred back to the anaerobic workstation and filled into 70 mL polycarbonate ultracentrifugation tubes. All tubes have to be completely filled (∼70 mL), otherwise the residual volume has to be filled up with standard Hepes buffer. The closed tubes are carefully balanced ( < 50 mg), centrifuged for 90 min at 95,000 × g at 4◦ C (Beckman 45-Ti Rotor) and subsequently returned to the anaerobic workstation. The resulting clear and brownish cytosolic extract is carefully decanted into 50 mL Falcon tubes. 4. Preparation of glutathione agarose-equipped gravity flow R Glutathione Agarose 4B is filled columns: 7 mL of Protino into 25 mL gravity flow columns. For equilibration, about 30 mL of distilled water is used to wash out the ethanol (20% v/v) of the storage solution, then the glutathione agarose is thoroughly washed with 70 mL anaerobic standard Hepes buffer before approximately 50 mL of the cell-free cytosolic extract is slowly applied per column. The resulting flow-through is then applied to the individual column for a second time. This results in an increased binding of recombinant proteins (see Note 3). 5. Affinity purification: Subsequently, the columns are washed extensively with at least 50 mL of standard Hepes buffer. DPOR subunits are brownish in color due to the presence of [4Fe–4S] clusters. Therefore, the glutathione agarose maintains a brownish color during all washing steps which is indicative for the successful purification of the individual DPOR subcomplexes. Due to the employed gravity flow technique about 3 h have to be scheduled for the overall affinity purification. 6. Elution of GST-tagged DPOR subunits: The employed pGEX-6P-1 vector allows for two alternative strategies for

Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase

137

the elution of recombinant proteins. The GST fusion proteins might be eluted from the affinity matrix by using a glutathione solution (100 mM in standard Hepes buffer) (see Note 4). Alternatively, the linker region between the GST tag and the N-terminus of the recombinant protein contains a specific cleavage site for PreScissionTM protease (Fig. 9.1a). While glutathione treatment allows for the rapid and quantitative elution of the fusion protein, several problems might arise from the remaining GST tag. The employed GST protein forms dimers in solution which might result in the oligomerization of the fusion protein and consequently impedes analyses concerning the oligomeric structure for the protein of interest. In contrast, PreScissionTM Protease treatment of the affinity resin allows for the elution of the recombinant protein in the absence of the protein tag. The employed protease is also fused to a GST tag; therefore this additional protein is efficiently retained on the affinity purification column. To allow quantitative cleavage 280 units of the protease were applied and carefully mixed with the 7 mL affinity matrix. Again, after incubation for 1 h the affinity matrix is manually agitated and subsequently incubated overnight at 17◦ C. The cleaved protein is then eluted by the stepwise addition of ∼20 mL standard Hepes buffer (see Note 4). Quality, identity, and yield of the purified protein are then analyzed by using the Bradford reagent according to the manufacturer’s instructions, followed by automated quantitative Edman degradation and/or SDS-Page analyses. 7. Gel permeation chromatography: As an additional purification step, DPOR subcomplexes were applied to preparative gel permeation chromatography using an Äkta purifier FPLC system equipped with a Superdex 200 26/60 column operating inside the anaerobic workstation at 17◦ C. 5 mL of purified (ChlN/ChlB)2 or ChlL2 at concentrations of approximately 6 mg/mL was run at a flow rate of 1.5 mL/min using standard Hepes buffer. Prior to injection, protein samples were centrifuged for 20 min at 12,100 × g in a minicentrifuge (Minispin Eppendorf) to sediment any traces of precipitated protein. A multi-wavelength UV cell is employed to detect protein elution at 280 nm and the presence of [4Fe–4S] clusters at 425 nm in parallel. Fractions containing DPOR subunits were collected at an elution volume ranging from 175 to 205 mL for tetrameric (ChlN/ChlB)2 and 150–185 mL for ternary DPOR complexes (Fig. 9.1d) (see Note 5). 8. Protein storage: Purified DPOR subcomplexes (ChlN/ ChlB)2 or ChlL2 are stored in anaerobic septum bottles at

138

Moser and Bröcker

4◦ C for up to 4 weeks or in 1.5 mL HPLC vials in liquid nitrogen for up to 1 year. DPOR complexes were stored at a concentration of 20 μM (4.2 mg/mL) for (ChlN/ChlB)2 and 40 μM (2.6 mg/mL) for ChlL2 , respectively, for later use in DPOR activity assays. For the concentration of proR stirred ultrafiltration cell equipped with a teins, an Amicon 30,000 Da size exclusion membrane was used. 3.3. Standard DPOR Activity Assay

1. For DPOR activity assays it is advantageous to use an anaerobic box which allows for the sample manipulation without wearing gloves. 2. Assays were carried out in a volume of 125 μl in 1.5 mL reaction tubes. 3. Ratio of (ChlN/ChlB)2 versus ChlL2 : To obtain the optimal specific activity of DPOR, the ratio of (ChlN/ChlB)2 versus ChL2 has to be determined in a pilot experiment. Therefore, 100 pmol of (ChlN/ChlB)2 (5 μl of a 20 μM stock) is combined with 150–400 pmol of ChlL2 (3.8–10 μl of a 40 μM stock) and supplemented with 84.6–78.4 μl standard Hepes buffer resulting in a volume of 93.4 μl, respectively. 4. Substrate and cofactors: For the standard assay a master mix containing 13 μM Pchlide, 2 mM sodium dithionite, 2 mM MgATP, 20 mM creatine phosphate, and 20 units creatine phosphokinase per assay was prepared, which is sufficient for 10 assays. Therefore, 8.1 μl of Pchlide (2 mM Stock), 2.5 μl dithionite (1 M stock), 25 μl ATP (100 mM stock), 250 μl creatine phosphate (100 mM stock), and 30 μl of the creatine phosphokinase (200 units) are mixed (see Note 6). 5. Initiation of DPOR catalysis: After the individual subunits (ChlN/ChlB)2 and ChlL2 as well as the standard Hepes buffer are placed in an 1.5 mL reaction tube, DPOR catalysis is started by addition of 31.6 μl of the master mix, resulting in a final assay volume of 125 μL. The assay mixtures are subsequently incubated in a thermomixer at 25◦ C and 500 rpm for 5–70 min. 6. Stopping of DPOR assays: DPOR catalysis is stopped by the addition of 500 μl acetone (100% v/v) to the reaction mixture. This results in the precipitation of all protein compounds, while Pchlide and the reaction product Chlide are dissolved in the resulting organic phase. Two successive centrifugation steps for 30 min at 12,100 × g result in the quantitative precipitation of all insoluble compounds. The obtained clear acetone phase (80% v/v) is then subjected to UV/vis analyses (see Note 7). 7. Quantification of DPOR activity: With the help of the extinction coefficients for Pchlide ε 626 = 30.4 mM–1 cm–1

Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase

139

and Chlide ε665 = 74.9 mM–1 cm–1 (in acetone 80% v/v), the amount of Chlide formed in the assay and the decline of the substrate Pchlide can easily be quantified (Fig. 9.1b). 8. Variation of reaction conditions: After determination of an optimal (ChlN/ChlB)2 versus ChlL2 ratio, the DPOR assay system allows for variations of reaction parameters. Temperature: Though P. marinus DPOR has a temperature optimum at 25◦ C, the reaction can be performed at temperatures ranging from 10 to 42◦ C for thermodynamic analysis of DPOR catalysis. However, the employed creatine phosphokinase (from rabbit muscle) tends to significantly loose activity at temperatures >42◦ C. Reaction time: Generally, reaction times of 2 min for the standard DPOR assay are sufficient for the quantification of Pchlide turnover. Nonetheless, extended reaction times for up to 70 min are beneficial for experiments with DPOR proteins showing reduced activity (e.g., mutant proteins). For the characterization of kinetic parameters great care has to be taken that DPOR catalysis remains in the linear range of the enzymatic conversion. Typical reaction times under Michaelis–Menten conditions are ranging up to 12 min. Substrate: In DPOR activity assays Pchlide might be substituted by a set of chemically altered derivatives based on Zn-protopheophorbide. Kinetics: For kinetic analysis of DPOR the concentration of one cofactor/substrate is varied, while the remaining cofactors are kept at a constant concentration. The artificial reductant dithionite might be replaced by a ferredoxin. This requires an additional ferredoxin regenerating system comprising 13 mM glucose-6-phosphate, 1.1 unit glucose-6-phosphate dehydrogenase from Torula yeast, 1.64 mM NADP+ , and 0.025 units ferredoxin NADP+ oxidoreductase from spinach. 3.4. DPOR Complex Formation

1. (ChlN/ChlB)2 is produced and purified according to Sections 3.1 and 3.2, eluted by PreScisionTM Protease cleavage and subsequently concentrated to 10 mg/mL (48 μM). 2. For DPOR complex formation using MgADP-AlF− 4 as an ATP analog, the GST-ChlL fusion protein is prepared from an overall of 5,000 mL production culture. A cell-free extract is prepared according to Sections 3.1 and 3.2 and the obtained fusion protein (∼42 mg) is coupled to 7 mL glutathione agarose and washed extensively with approximately 50 mL standard Hepes buffer. Ten μL of

140

Moser and Bröcker

this protein-bound affinity matrix is subjected to a Bradford test to quantify the amount of purified GST-ChlL. 3. Prepare a solution containing 50 mM NaF, 2 mM AlCl3 , 10 mM MgADP in standard Hepes buffer immediately before use. To yield 15 mL of this buffer 3.75 mL of NaF (from 200 mM stock), 0.6 mL of AlCl3 (from 50 mM stock), and 1.5 mL of MgADP (from 100 mM stock) are immediately mixed with 9.15 mL standard Hepes buffer. This solution is incubated for 30 min without agitation. In some cases small amounts of a visible precipitate might occur. If this is the case the supernatant of this mixture is used for DPOR complex formation. 8 mL of this MgADPAlF− 4 containing solution is then completed by the addition of 0.5 mol (ChlN/ChlB)2 per mol GST-ChlL bound to the glutathione agarose. Subsequently, a fivefold molar excess of Pchlide compared to the amount of (ChlN/ChlB)2 is added (see Note 8). 4. Ten mL of this bright green solution is then added to the GST affinity matrix carrying the immobilized GST-ChlL protein. When the employed resin is completely infiltrated (indicated by the green color of the resin) the gravity flow column is sealed. 5. The column is then incubated for at least 60 min to allow complex formation (see Note 9). 6. Next, the column is washed with 50 mL of standard Hepes buffer supplemented with 50 mM NaF (12.5 mL from 200 mM stock) and 2 mM AlCl3 (2 mL from 50 mM stock) to remove excessive nucleotide cofactor analogs. Initially, these wash fractions are brownish in color, whereas the late wash fractions appear colorless. Due to the binding of the Pchlide substrate as part of the ternary complex, the affinity matrix retains a bright green color (see Note 10). 7. For the elution of the ternary complex 40 units of the PreScissionTM Protease are applied to the affinity matrix. The resin is carefully agitated for 60 min to allow direct contact of the immobilized protease with the GST-ChlL fusion protein substrate. This reaction mixture is then incubated at 17◦ C for 16 h. 8. Bright green DPOR complexes are eluted with standard Hepes buffer (+ 50 mM NaF and 2 mM AlCl3 ) and subR sequently concentrated to 10 mg/mL with an Amicon stirred ultrafiltration cell (30,000 Da exclusion limit). 9. Alternatively, DPOR complexes can be formed with the help of a genetically modified ChlL protein. The ChlLLeu153 variant with a deleted Leucine153 in the switch-II region is produced and purified analogous as

Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase

141

the wild-type ChlL fusion protein (3). Following extensive washing with standard Hepes buffer in the absence of nucleotide analogs GST-ChlLLeu153 bound to the affinity matrix is quantified and then incubated with 0.5 mol (ChlN/ChlB)2 per mol GST-ChlLLeu153 in the presence of a fivefold molar excess of Pchlide. After incubation for at least 1 h, unbound (ChlN/ChlB)2 complex is washed from the affinity matrix with 20 mL standard Hepes buffer. After protease cleavage, the elution of the complex with standard Hepes buffer is then carried out as described for the DPOR complexes formed with MgADP-AlF− 4 (see Note 11). 10. In an additional purification step, these complexes are then subjected to a preparative gel permeation chromatography according to Section 3.2, step 7. Up to 5 mL of a protein sample at a concentration of 10 mg/mL is run in standard Hepes buffer supplemented with 25 mM NaF in the presence of 500 μM AlCl3 . UV detection is facilitated at 280 nm for protein compounds. In parallel, the presence of Pchlide is detected at 625 nm (see Note 12). 11. Purified DPOR complexes are analyzed by SDS-Page analyses (Fig. 9.1c).

4. Notes 1. Addition of 1.7 mM dithionite results in the rapid formation of anaerobic conditions in the centrifugation bottles. The subsequent anaerobic incubation of the cultures for 3 h prior to centrifugation ensured the accurate assembly of the [4Fe–4S] clusters of DPOR. The redox indicator resazurin might be added to buffer solutions according to the manufacturer’s instructions to indicate the required anaerobic conditions. 2. The outlet valve has to be carefully adjusted during operatR pressure cell. The employed pressure ing the FRENCH should not decline below 7,000 psi to ensure quantitative disruption of the cells with a single passage through the cell. We choose a valve gear fitted with a metal needle. This system allows a more precise control of the pressure when compared to the standard Teflon-fitted gear. 3. The yield of individual GST fusion proteins might differ drastically between glutathione affinity matrixes provided from different manufacturers. For the purification of DPOR, the highest amount of DPOR proteins R Glutathione Agarose 4B from was obtained with Protino Macherey–Nagel.

142

Moser and Bröcker

Following protein purification, the affinity matrix can be restored by washing with 10 mL of 100 mM glutathione, 50 mL distilled water, and finally 50 mL ethanol (20% v/v) before storage at 4◦ C. The affinity resin can be recycled for approximately 10 successive applications until the binding capacity begins to decline. 4. Yields of up to 7 mg GST fusion protein per mL glutathione agarose were obtained from a 1,000 mL production culture. If glutathione is used for the elution of DPOR from R Glutathione Agarose 4B significantly higher conProtino centrations than the proposed 10–20 mM glutathione are required for the quantitative protein elution. The use of PreScissionTM protease is the primary cost factor for DPOR purification. About one unit of protease per 100 μg protein to be cleaved is proposed by the manufacturer. Nonetheless, a more efficient protein cleavage is observed when higher amounts of the protease are used. 5. An additional purification step by gel permeation chromatography is not mandatory if DPOR subunits are used in activity assays. 6. Creatine phosphokinase tends to irreversibly precipitate after repeated cycles of freezing and thawing. Stock solutions of appropriate size should therefore be stored. Pchlide proved to be very stable when dissolved in DMSO. No degradation was observed when the pigment was exposed to light or when it was stored at room temperature. 7. Take care if acetone is transferred to the anaerobic workstation. Prevent contact of acetone with the vinyl or Perspex material of the workstation. DPOR assays stopped with acetone may be stored at –20◦ C without any effect on the pigment composition. 8. The 50 mM AlCl3 stock solution might be stirred for several hours until residual amounts have completely dissolved. Pchlide is applied in fivefold molar excess compared to (ChlN/ChlB)2 in complex formation experiments since a significant amount of the pigment is non-specifically adsorbed by the affinity matrix. Most plausible, this effect is caused by hydrophobic interaction of Pchlide with the employed linker of the glutathione agarose. The protein binding capacity of the affinity matrix is not affected by adsorption of Pchlide. No significant adsorption of the pigment to the Sepharose matrix of the Superdex 200 FPLC column is observed.

Methods for Nitrogenase-Like Dark Operative Protochlorophyllide Oxidoreductase

143

AlCl3 is combined with NaF to generate AlF− 4 . The exact amount of AlF− formed with this protocol is not deter4 mined. Several other aluminum compounds occur as side products, especially various aluminum hydroxides AlOHx that might cause a visible precipitate. These hydroxides do not interfere with proteins and/or the affinity matrix of the gravity flow column (despite slowing down the flow rate). This type of precipitate will sediment in freshly prepared MgADP-AlF− 4 solutions within 30 min. 9. Complex formation is also induced at significantly lower concentrations of the individual components and also after a shorter period of time. Nevertheless, incubation for 60 min in the presence of 50 mM NaF, 2 mM AlCl3 , and 10 mM MgADP ensures the formation of a stable complex with high yield. Obtained ternary DPOR complexes did not show any protein degradation after gel permeation chromatography. Complex formation was not reversible as indicated in subsequent DPOR activity assays. 10. Omission of MgADP in the buffer solution for the washing of column-bound complexes and in the gel permeation chromatography did not result in the dissociation of ternary complexes. 11. Deletion of Leucine153 in the switch-II region of ChlL has no influence on the recombinant production of the mutant protein. Furthermore, [FeS] cluster biogenesis was not affected when compared to the wild-type protein. 12. An additional purification step for ternary DPOR complexes via gel permeation chromatography is required to eliminate small traces of uncompleted DPOR complexes. The standard Hepes buffer employed for the gel permeation chromatography of ternary MgADP-AlF− 4 DPOR complexes contains only reduced amounts of NaF and AlCl3 to avoid the precipitation of AlOHx compounds on the column. References 1. Bröcker MJ, Virus S, Ganskow S et al (2008) ATP-driven reduction by darkoperative protochlorophyllide oxidoreductase from chlorobium tepidum mechanistically resembles nitrogenase catalysis. J Biol Chem 283:10559–10567 2. Bröcker MJ, Wätzlich D, Uliczka F et al (2008) Substrate recognition of nitrogenase-like dark operative protochlorophyllide oxidoreductase from Prochlorococcus marinus. J Biol Chem 283: 29873–29881

3. Brocker MJ, Waetzlich D, Saggu M et al (2010) Biosynthesis of (bacterio)chlorophylls: ATP-dependent transient subunit interaction and electron transfer of dark operative protochlorophyllide oxidoreductase. J Biol Chem 285:8268–8277 4. Walther J, Bröcker MJ, Wätzlich D et al (2009) Protochlorophyllide: a new photosensitizer for the photodynamic inactivation of Gram-positive and Gramnegative bacteria. FEMS Microbiol Lett 290: 156–163

Section III Spectroscopic Methods

Chapter 10 X-Ray Crystallography Lauren E. Roth and F. Akif Tezcan Abstract X-ray crystallography has been particularly important in the study of the enzyme nitrogenase, providing researchers with high-resolution structural models that have been essential to studying the enzyme’s unique metal clusters and nucleotide-binding modes and protein interactions. While several important nitrogenase structures have already been determined using X-ray crystallography, the technique still holds great potential for future significant discoveries involving reaction intermediates and redox states of the enzyme’s metal clusters. Thus, it is important to inform future nitrogenase researchers about the procedures for obtaining crystals of nitrogenase component proteins and their complexes and determining their structures. While nitrogenase component proteins from several bacteria have been crystallized, the majority of structures and those of highest resolution are of the nitrogenase proteins from the bacteria Azotobacter vinelandii. Therefore, the bulk of this chapter will focus on methods for crystallization and structure determination of nitrogenase component proteins from A. vinelandii. Key words: Protein crystallography, X-ray diffraction, nitrogenase, complexes.

1. Introduction 1.1. Background

Three-dimensional molecular structures of proteins determined by X-ray crystallography have significantly enhanced the scientific understanding of protein structure and function. Given the large sizes of its component proteins, which are prohibitive for NMR spectroscopy, and the complex spectroscopic signatures that its many iron–sulfur complexes exhibit, X-ray crystallography has been particularly important in the study of the enzyme nitrogenase. It is safe to say that the crystal structures of the nitrogenase components and complexes have been the single most important driving force and guiding light for researchers who have been

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_10, © Springer Science+Business Media, LLC 2011

147

148

Roth and Tezcan

trying to understand the many mechanistic intricacies of nitrogenase catalysis. Nitrogenase is a complex enzyme involving two component proteins, the molybdenum–iron protein (MoFeP), which holds the active site cluster, the iron–molybdenum cofactor (FeMoco), and the iron protein (FeP), an ATPase that is the only known reductant of MoFeP. These two proteins coordinate nucleotidedependent electron transfer to the FeMoco to catalyze the reduction of several substrates including dinitrogen (N2 ) (1). Prior to the first nitrogenase crystal structures, the compositions and structures of the redox-active centers in nitrogenase were studied using EPR, ENDOR, Mössbauer, and X-ray absorption spectroscopy. In addition, attempts were made to predict the protein ligands involved in metal cluster binding based on the chemical reactivity of both nitrogenase proteins and the isolated FeMoco cluster to various chelators (2). Despite extensive research, the exact structure of the nitrogenase metal centers remained unclear and several possible structures were proposed (3, 4). Rees and co-workers published the first crystal structures of the nitrogenase component proteins in 1992, a 2.9-Å structure of FeP and a 2.7-Å structure of MoFeP, both isolated from the nitrogen-fixing bacterium Azotobacter vinelandii (5, 6). Since then, the homologs of both proteins from Klebsiella pneumoniae and Clostridium pasteurianum have also been structurally characterized (7–10). The initial structures revealed that FeP is a homodimer of two γ-subunits, which are bridged by a 4Fe:4S cluster coordinated to two cysteines from each subunit (5). As expected form earlier characterizations, the iron–sulfur clusters of MoFeP proved to be much more intricate and unique. The P-cluster, proposed to function as an electron-relay center from the FeP cluster to the catalytic site FeMoco, was revealed to be an 8Fe:7S cluster sandwiched between the α- and β-subunits of MoFeP. In the as-isolated, dithionite-reduced state, the P-cluster essentially appears to be a fusion product of two 4Fe:4S clusters, which share one inorganic sulfide corner coordinated to two Fe’s from each half. Upon two-electron oxidation, the P-cluster “opens up,” whereby the bridging sulfide loses coordination to one of the four Fe’s (11, 12). The structure of FeMoco was much more unexpected. It was revealed to be a 7Fe:8S:1Mo:homocitrate cluster contained entirely within the α-subunit. Significantly, this resting-state, dithionite-reduced structure featured a coordinatively saturated Mo center, long thought to be a good candidate for N2 binding and activation, and six unusual central Fe’s, each coordinated to only three bridging sulfides (11). This discovery single-handedly brought nitrogenase back into the attention of a diverse group of inorganic, bioinorganic, biological, and computational chemists and

X-Ray Crystallography

149

reinvigorated research efforts in the area of biological nitrogen fixation. With these groundbreaking initial structures, the goal of finally understanding the catalytic mechanism of nitrogenase seemed very near. Yet, in the intervening two decades, nitrogenase has kept its reputation as an enzyme impervious to detailed characterization, where the most important aspects of the nitrogenase mechanism have remained largely unknown: How and where do substrates bind the active site? What are the catalytic intermediates? What is the role of ATP hydrolysis in nitrogen fixation? How is the transfer of multiple electrons and protons into FeMoco regulated? X-ray crystallography, in the meantime, also maintained its status as the primary driving force for nitrogenase research, and attempts to investigate these questions through crystallography resulted in several new insights. A 1.16-Å resolution structure of MoFeP from A. vinelandii revealed that a light atom (X = C, N, or O), not observed in lower resolution structures, was present in the center of the FeMoco (13). This new structure not only raised new questions about what the identity of X might be and how it might relate to catalysis, but also led synthetic and computational chemists to revise their strategies for constructing model complexes of FeMoco and for producing more accurate electronic descriptions of it. Structures of FeP have also been solved showing bound nucleotides (14, 15), as well as those of FeP mutants that are unable to hydrolyze or release nucleotides (16–18), and a version of FeP with an all-ferrous [4Fe:4S] cluster (19). X-ray crystallography led to another significant advancement in nitrogenase research with the determination of structures of FeP–MoFeP complexes. The initial 3.0-Å FeP–MoFeP complex structure—again from A. vinelandii—was obtained in the presence of Mg·ADP·AlF4 – , which is a transition-state analog for Mg·ATP hydrolysis and leads to the semi-irreversible formation of the complex. This structure revealed how the Fe:S clusters of FeP and MoFeP are arranged relative to one another for efficient electron transfer. It also showed the drastic conformational changes that need to occur during nucleotide binding in order to allow complex formation (20). These observations were repeated in the 2.2-Å structure of MoFeP complexed to the L127 mutant of FeP, which was proposed to mimic the ATP-bound state of the latter (21). The most recent set of complexed nitrogenase structures from the Rees Group revealed FeP bound to MoFeP in three distinct conformations and docking geometries depending on its nucleotide-binding state: nucleotide free, Mg·AMPPCP bound (a non-hydrolyzable ATP analog), and Mg·ADP bound. The protein conformations seen in each of the complexed structures may represent various enzyme transition states, perhaps necessary to coordinate electron and proton transfer, substrate

150

Roth and Tezcan

binding, and eventual reduction, and provide a starting point for experiments toward investigating the role of protein–protein interactions in nitrogenase catalysis (22). X-ray crystallography has undoubtedly proven to be one of the most important methods used to study nitrogenase. Crystal structures of the MoFeP and FeP have confirmed models of the enzyme’s unique metal clusters, provided additional data to aid in the interpretations of other nitrogenase experiments and raised interesting questions about the role of nucleotide binding and protein interactions in catalysis. Yet, it has also become clear in the last two decades that nitrogenase is a very dynamic system, and most of its interesting chemistry (which involves the processing of 8 electrons, 8 protons, and 16 ATP molecules) (1) happens beyond the resting-state structures and thus are not easily accessible by crystallography. For example, the pressurization of resting-state MoFeP crystals with very high pressures (>100 psi) of the substrates N2 and C2 H2 or inhibitor CO has shown no evidence for binding, clearly indicating the need for activated (i.e., reduced) states of FeMoco, which can only be fleetingly populated during enzyme turnover (F.A.T., O. Einsle and D.C. Rees, unpublished results). Nevertheless, given the unparalleled structural detail that X-ray crystallography can provide, this technique still holds great potential for future significant discoveries in nitrogenase research. For instance, crystal structures of reaction intermediates generated within nitrogenase crystals may reveal bound substrates or different redox states of the enzyme’s metal clusters. Moreover, the ever-growing number of nitrogenase mutants with altered reactivities and substrate-binding profiles may enable the capturing of structural states that resemble true catalytic intermediates (23). Thus, it is important to inform future nitrogenase researchers about the procedures for obtaining crystals of nitrogenase component proteins and their complexes and determining their structures. 1.2. Nitrogenase Crystallography

The process of using X-ray crystallography to determine a protein structure involves four main steps: protein preparation, crystal growth, diffraction data collection, and data processing. The single most important prerequisite to a successful crystallization experiment is protein homogeneity/purity. This is especially significant in the case of FeP and MoFeP, which pose certain challenges in isolation and purification, and are notoriously sensitive to oxygen damage. Methods for the proper isolation, purification, and handling of nitrogenase proteins are described elsewhere in this book. Determining solution conditions that will eventually result in crystal formation is typically the rate-limiting and the most challenging step in X-ray crystallography. Several conditions can be varied to affect protein solubility so that it will crystallize, including protein concentration, salt concentration, pH, the

X-Ray Crystallography

151

type and concentration of precipitating agents such as polyethylene glycol, and temperature. There are several methods available for setting up crystallization experiments. Among the published structures of nitrogenase proteins, high-quality crystals have been grown using sitting drop vapor diffusion, hanging drop vapor diffusion, capillary batch or microbatch, and liquid–liquid diffusion methods. While each method has its own advantages, the highest quality crystals of nitrogenase have thus far been grown using sitting drop vapor diffusion. This technique is also the least cumbersome for setting up crystallization experiments under anaerobic conditions and greatly facilitates crystal harvesting and handling. It will therefore be given the most attention in this chapter. Once nitrogenase crystals have formed and grown to maturation, they can be either directly harvested from the crystallization platform for diffraction experiments or further modified by soaking with small molecules such as nucleotides and substrates/inhibitors or pressurized with gaseous substrates. Crystals are then cryoprotected using a glassing agent and mounted into a suitable sized loop before being frozen in liquid nitrogen, where they can be stored until diffraction experiments. X-ray diffraction data collection, while not in itself challenging, requires sophisticated equipment and training and will not be addressed here. Once diffraction data have been collected, they can be processed using many freely available software programs to obtain protein structures on essentially a personal computer. In the case of nitrogenase components, several structures have already been solved and their coordinates deposited into the Research Collaboratory for Structural Bioinformatics (RCSB) Databank (see Tables 10.1 and 10.2). These coordinates can be used as starting models for molecular replacement methods, which readily lead to the structure determination of the desired MoFeP or FeP variant.

2. Materials 2.1. Protein Preparation

1. Anaerobic Tent (Coy Laboratory Products, Grass Lake, MI). 2. Econo-Pac 10DG desalting columns (Bio-Rad, Hercules, CA). 3. 10 kDa MW cutoff Micron centrifugal devices (Millipore, Billerica, MA).

2.2. Crystallization (See Note 1)

1. 0.2 μM syringe filters (Whatman, Florham Park, NJ). 2. 1 M Tris pH 7.4: Tris-HCl, OmniPur (EMD, Darmstadt, Germany), Tris Base, Molecular Biology Grade (Fisher Scientific, Pittsburg, PA).

PDB ID

1M1N

2MIN

3MIN

1FP4

3K1A

1L5H

1MIO

1MIO

1H1L

Protein

Av1 (13)

Av1, ox. P-cluster (12)

Av1, red. P-cluster (12)

Av1, αH195Q (24)

Av1, αV70I (23)

Av1, apo-FeMoco (25)

Cp1 (9, 29)

Cp1 (9, 29)

Kp1, apo-citrate (26)

1.9

3.0

3.0

2.3

2.3

2.5

2.0

2.0

1.16

Resolution (Å)

50 mM Tris, pH 7.4

80 mM Tris, pH 8.0

80 mM Tris, pH 8.0

0.1 M CHES, pH 9.5

0.1 M Tris, pH 8.0

0.1 M Tris, pH 8.0

0.1 M Tris, pH 8.5

0.1 M Tris, pH 8.5

0.1 M Tris, pH 8.0

Buffer

0.4–0.6 M MgCl2

0.21 M MgCl2 0.3 M CsCl

0.21 M MgCl2

0.17–0.19 M Na2 MoO4

0.2 M Na2 MoO4

0.2 M Na2 MoO4

0.2 M Na2 MoO4

–

Salt

Well solution conditions

7% PEG6000

18% PEG4000

15% PEG4000

20% PEG8000

30% PEG4000

30% PEG400

30% PEG8000

30% PEG8000

13% PEG8000

Precipitant

1 mM dithionite

Additive

Table 10.1 Nitrogenase structures deposited in the RCSD databank and reported crystallization conditions

25% ethylene glycol

20% MPD

Cryoprotectant

152 Roth and Tezcan

PDB ID

1QGU

1QH1

1QH8

2NIP

1G5P

1G1M

1XCP

1RW4

2C8V

1XD9

Protein

Kp1, reduced (8)

Kp1, oxidized (8)

Kp1, mixed oxidation state (8)

Av2 (10)

Av2 (19)

Av2, all ferrous (19)

Av2, F135W + MgADP (15)

Av2, L127 (17)

Av2, L127 + MgADP (18)

Av2, D19N + MgADP (16)

Table 10.1 (Continued)

2.8

2.5

2.5

3.2

2.25

2.2

2.2

1.6

1.6

1.6

Resolution (Å)

0.1 M Tris, pH 8.5, 0.2 M sodium acetate

Tris, pH 8.5

Tris, pH 8.5

0.1 M Tris, pH 8.5, 0.2 M sodium acetate

0.1 M Tris, pH 8.0

50 mM Tris

50 mM Tris, pH 8.0

50 mM Tris, pH 8.0

50 mM Tris, pH 8.0

Buffer

0.16 M MgCl2

0.16 M MgCl2

0.7–0.9 M NaCl

0.14–0.2 M Na2 MoO4

0.4–0.6 M MgCl2

0.4–0.6 M MgCl2

0.4–0.6 M MgCl2

Salt

Well solution conditions

30% PEG4000

24% PEG4000 20% glycerol

24% PEG4000 20% glycerol

30% PEG4000

9% PEG4000

25–30% PEG4000

7% PEG6000

7% PEG6000

7% PEG6000

Precipitant

1 mM dithionite 20 mM ATP 50 mM MgCl2

1 mM dithionite

10 mM MgADP

Additive

30% glycerol

30% glycerol

Cryoprotectant

X-Ray Crystallography 153

1.93

3.0

3.2

2.3

1DE0

1CP2

1N2C

1M1Y

1M34

Av2, F135W (27)

Cp2 (10)

Av1–Av2, ADP·AlF4 complex (20)

Av1–Av2, EDC crosslinked complex (28)

Av1–Av2, ADP·AlF4 complex (28)

2.4

2.8

1XDB

Av2, D129E (16)

Resolution (Å)

PDB ID

Protein

Table 10.1 (Continued)

0.1 M cacodylate, pH 6.5

75 mM Tris, pH 8.0

0.1 M cacodylate, pH 6.5

50 mM HEPES, pH 7.5

0.1 M Tris, pH 8.5, 0.2 M sodium acetate

0.1 M Tris, pH 8.5, 0.2 M sodium acetate

Buffer

30 mM MgCl2

20–50 mM MgCl2

0.23–0.27 M CaCl2

Salt

Well solution conditions

19% PEG8000

20.9% PEG6000

18.5–22.5% PEG4000, 5–15% glycerol 16–20% PEG8000

30% PEG4000

30% PEG4000

Precipitant

510 mM cadaverine, 2 mM dithionite

Additive

20% MPD

20% MPD

30% glycerol

Cryoprotectant

154 Roth and Tezcan

1G20

1G21

2AFH

2AFI

2AFK

Av1–Av2, delta L127 complex (21)

Av1–Av2, L127 + MgATP complex (21)

Av1–Av2, nucleotide free complex (22)

Av1–Av2, MgADP complex (22)

Av1–Av2, AMPPCP complex (22)

2.3

3.1

2.1

3.0

2.2

Resolution (Å)

0.1 M Tris, pH 8.5

0.1 M Tris, pH 8.5

0.1 M Tris, pH 8.5, 0.2 M sodium acetate

0.1 M Tris, pH 8.5, 0.2 M sodium acetate

Buffer

0–50 mM NaCl

0–50 mM NaCl

Salt

Well solution conditions

18–22% PEG8000

18–22% PEG8000

16–18% PEG10000

30% PEG4000

30% PEG4000

Precipitant

10 mM dithionite

10 mM dithionite

10 mM dithionite

1 mM dithionite, 50 mM MgATP soak

1 mM dithionite

Additive

Cryoprotectant

Av1, MoFeP from Azotobacter vinelandii; Av2, FeP from A. vinelandii; Cp1, MoFeP from Clostridium pasteurianum; Cp2, FeP from C. pasteurianum; Kp1, MoFeP from Klebsiella pneumoniae.

PDB ID

Protein

Table 10.1 (Continued)

X-Ray Crystallography 155

156

Roth and Tezcan

Table 10.2 Crystal symmetry and unit cell dimensions for high-resolution nitrogenase structures

Protein

PDB ID

Resolution (Å)

Av1 (13)

1M1N

1.16

Cp1 (9)

Kp1 (8)

Av2 (10)

Cp2 (10)

1MIO

1QGU

2NIP

1CP2

3.0

1.6

2.2

1.93

Space group P 21

P 21

C2

P 21

P 21

Av1–Av2 nucleotide free (22)

2AFH

2.1

P 2 1 21 21

Av1–Av2 + MgADP (22)

2AFI

3.1

P1

Av1–Av2 + MgAMPPCP (22)

2AFK

2.3

P 21 21 21

Unit cell dimensions Length (Å)

Angles (◦ )

a = 108.31

α = 90.00

b = 131.63

β = 108.37

c = 159.16 a = 69.96

γ = 90.00 α = 90.00

b = 151.30

β = 110.40

c = 121.90

= 90.00

a = 204.03

α = 90.00

b = 75.30

β = 122.61

c = 163.01 a = 56.80

γ = 90.00 α = 90.00

b = 92.90

β = 100.00

c = 63.60

γ = 90.00

a = 67.60

α = 90.00

b = 75.87

= 114.17

c = 53.55 a = 170.91 b = 75.87 c = 223.66

γ α β γ

= 90.00 = 90.00 = 90.00 = 90.00

a = 72.92 b = 141.43 c = 165.55 a = 110.53 b = 120.89 c = 264.83

α β γ α β γ

= 73.69 = 79.37 = 76.58 = 90.00 = 90.00 = 90.00

3. 1 M Tris pH 8.0: Tris-HCl, OmniPur (EMD), Tris Base, Molecular Biology Grade (Fisher Scientific). 4. 1 M Tris pH 8.5: Tris-HCl, OmniPur (EMD), Tris Base, Molecular Biology Grade (Fisher Scientific). 5. 1 M CHES pH 9.5: CHES (Fisher Scientific). 6. 1 M HEPES pH 7.5: HEPES, Molecular Biology Grade (Fisher Scientific). 7. 1 M cacodylate pH 6.5: sodium cacodylate trihydrate (EMS, Hatfield, PA). 8. 1 M sodium acetate: sodium acetate anhydrous, certified ACS (Mallinckrodt Baker, Phillipsburg, NJ).

X-Ray Crystallography

157

9. 1 M Na2 MoO4 : sodium molybdate dihydrate, certified ACS (EMD). 10. 4 M MgCl2 : magnesium chloride hexahydrate, extra pure (EMD). 11. 4 M NaCl: sodium chloride, certified ACS (Fisher Scientific). 12. 4 M calcium chloride: calcium chloride dihydrate, certified ACS (Fisher Scientific). 13. 50% polyethylene glycol 10,000: polyethylene glycol 10,000 (Hampton Research, Aliso Viejo, CA). 14. 50% polyethylene glycol 8000: polyethylene glycol 8000 (Fluka, Milwaukee, WI). 15. 50% polyethylene glycol 6000: polyethylene glycol 6000 (Hampton Research). 16. 50% polyethylene glycol 4000: polyethylene glycol 4000 (Fluka). 17. 100% polyethylene glycol 400: polyethylene glycol 400 (EMD). 18. 0.5 M sodium dithionite (Na2 S2 O4 ): sodium dithionite, 88% minimum (Mallinckrodt Baker) (see Note 2). 19. 24-well sitting drop crystallization plate (Hampton Research). 20. Clear packing tape (Hampton Research). 21. 24-well hanging drop crystallization plate (Hampton Research). 22. Cover slides, 22 mm × 0.22 mm (Hampton Research). 23. Vacuum Grease (Dow Corning, Midland, MI). 2.3. Crystal Harvesting and Cryoprotection

1. 100% glycerol: glycerol (EMD). 2. 100% polyethylene glycol 400: polyethylene glycol (EMD). 3. 100% 2-methyl-2,4-pentanediol: 2-methyl-2,4-pentanediol, >99% (MPD) (Sigma Aldrich, St. Louis MO). 4. 100% perfluoropolyether: perfluoropolyether (Hampton Research). 5. Sterile disposable scalpels (BD Medical, Franklin Lakes, NJ). 6. Magnetic Crystal Wand (Hampton Research). 7. Crystallography Micro-Tools (Hampton Research). 8. 3-Well glass crystal soaking plate (Corning Life Sciences, Union City, CA). 9. Various sized crystal loops (Hampton Research).

158

Roth and Tezcan

3. Methods Because of the oxygen-sensitive nature of nitrogenase, it is imperative that all crystallization experiments be done inside an anaerobic glovebox or tent. All reagents should be properly deaerated before use and all crystallization hardware, including desalting columns, pipet tips, and crystal trays, should be given adequate time to equilibrate with the oxygen-free environment. Typically all hardware is brought into the anaerobic environment at least the day before crystal trays are to be set up. The following procedures are based on our experiences with the crystallization of FeP and MoFeP from A. vinelandii. Below, they are referred to as Av2 (FeP) and Av1 (MoFeP). Crystallization conditions for Av1 and Av2, as well as their counterparts from other organisms, are listed in Table 10.1. 3.1. Preparation of Nitrogenase Proteins for Crystal Plates

1. Purity of nitrogenase component proteins can be estimated using SDS-PAGE. For optimal crystallization, both Av1 and Av2 should be at least 90–95% pure. 2. The protein should be exchanged into buffer conditions compatible with both protein stability and the proposed crystallization conditions. For example, if crystals will be grown in a well solution containing Tris, the protein should also be in a Tris solution. A small desalting column, such as the Econo-Pac 10DG (Bio-Rad), can be used to exchange Av1 or Av2 into an appropriate buffer. 3. Protein concentration is critical for obtaining a supersaturated solution that will produce crystals. Av2 stocks for crystallization are typically concentrated to 20–30 mg/mL but concentrations as high as 70 mg/mL have been reported. Av1 stock concentrations are usually 30–50 mg/mL. A high molecular weight cutoff centrifugal concentrator such as the Amicon Ultra-0.5 centrifugal filter unit (Millipore, Billerica, MA) used in a tabletop centrifuge in an anaerobic glovebox is a simple way to concentrate nitrogenase proteins.

3.2. Growing Nitrogenase Crystals—Sitting Drop Vapor Diffusion

Vapor diffusion crystallography creates a supersaturated protein solution by separating a larger liquid reservoir of reagent, or well solution, from a small drop of protein mixed with the same reagent. In a sealed environment, the protein drop will slowly lose water vapor to the larger reservoir, becoming increasingly supersaturated until protein nucleation and crystal growth can occur. Both hanging drop and sitting drop vapor diffusion techniques have been used to grow nitrogenase crystals; however, sitting drop crystallization has produced the highest resolution

X-Ray Crystallography

159

nitrogenase crystals. Typically these experiments are set up using 24-well plastic crystal plates though smaller plates, such as 96-well plates, can be used. The advantage of the 24-well plate is its larger size which is easier to handle in an anaerobic glovebox or tent. Crystal plates specifically for sitting drop vapor diffusion contain a stand in each well that is elevated a certain distance from the well bottom. Each stand has a small depression in the center for mixing protein and reagent drops. The following methods will assume use of a 24-well plate with 4 rows A-D and 6 columns, 1–6. 1. Mix the well solution for well A1. Well solutions can be made by mixing appropriate volumes of more concentrated stock solutions such as those listed in the materials section (see Note 3). 2. Pipet 250 μL of well solution A1 into well A1, then repeat for wells A2–B6. 3. Pipet 2 μL of protein onto the post of well A1. Pipet 2 μL of well solution from A1 directly on top of the protein drop. Repeat for wells A2–B6 (see Note 4). 4. Cut a piece of clear packing tape slightly larger than the crystal plate. Seal wells A1–B6 with the tape, making sure to avoid bubbles or gaps that will prevent a tight seal. 5. Repeat steps 1–3 for the second half of the plate, wells C1–D6. 6. Place the crystal plate in a temperature-controlled environment free of vibrations. 7. Use a standard microscope to check each well for crystal formation in the protein droplets, for example, each day for the first several days, and then only once a week. Nitrogenase crystals can take anywhere from 1 day to several months to form depending on the protein and well solution conditions. 3.3. Growing Nitrogenase Crystals—Formation of In Situ Nitrogenase Complex Crystals

1. Prepare the protein solutions. Both Av1 and Av2 should be exchanged into 100 mM Tris pH 7.75, 200 mM NaCl, and 5 mM dithionite and then concentrated to approximately 20 mg/mL for Av2 and 40 mg/mL for Av2. 2. Mix stocks of Av1 and Av2 to form the complex mixture. Diffraction quality crystals will form using ratios from 1:1.2 to 1:1.5 Av1:Av2 (v/v). 3. For AMPPCP or ADP complex crystals, prepare a 1 M MgCl2 stock and a 100 mM stock of either AMPPCP or ADP in 100 mM Tris pH 8.5. These solutions should be used immediately after mixing. Add an appropriate volume to the protein solution to reach a final concentration of 10 mM MgCl2 and 10 mM AMPPCP or ADP. The

160

Roth and Tezcan

nucleotide stock solutions should not be premixed with the MgCl2 solution, as this will lead to precipitation. 4. Mix the well solutions from more concentrated stock solutions. Crystals will form in well solutions containing 0–50 mM NaCl, 100 mM Tris pH 8.5, 10 mM dithionite, and 16–18% PEG10000 (w/v) or 18–22% PEG8000 (w/v) for nucleotide-free or ADP or AMPPCP complexes, respectively. Add 250 μL of solution corresponding to a specific well in a 24-well sitting drop plate. 5. Pipet 2 μL of the protein mixture onto the sitting drop stand. Pipet 2 μL of the well solution on top of the protein drop. Seal the plate with clear crystal tape. 6. Crystals should appear after 2–3 days and be fully grown after 1–2 weeks. 3.4. Growing Nitrogenase Crystals—Hanging Drop Vapor Diffusion

1. Apply a thin line of vacuum grease along the top edge of the plate wells A1–D6. Vacuum grease can be placed into a 3 mL syringe and then “injected” around the 24 reservoirs of a crystal plate. 2. Mix the well solution for well A1. Well solutions can be made by mixing appropriate volumes of more concentrated stock solutions such as those listed in the materials section (see Note 3). Pipet 250 μL of well solution into well A1. 3. Create a droplet of protein and reagent on a cover slide. Pipet 2.0 μL of Av1 or Av2 onto the glass slide first and then pipet 2.0 μL of well solution from A1 directly on top of the protein drop (see Note 4). For nitrogenase crystals, it is not necessary to mix the protein and reagent droplets together. 4. Grasp the edges of the cover slide either with forceps or fingers and carefully invert the slide so the drop is hanging upside down. Position the slide so the protein drop is centered over well A1. Gently press the cover slide onto the vacuum grease to create a tight seal around the well. 5. Repeat steps 2–4 for wells A2–D6. 6. Place the crystal plate in a temperature-controlled environment free of vibrations. 7. A standard microscope can be used to check each well for crystal formation in the protein droplets, for example, each day for the first several days, and then only once a week. Nitrogenase crystals can take anywhere from 1 day to several months to form depending on the protein and well solution conditions.

X-Ray Crystallography

3.5. Crystal Harvesting and Cryoprotection

161

1. Prepare solutions for cryoprotection. Cryoprotection typically involves soaking harvested crystals in solutions similar to the crystal growth conditions that additionally contain sufficient amounts of a glassing agent (e.g., 20% PEG400, 20% glycerol, 25% MPD). If, for example, PEG400 is to be used as a glassing agent, at least four cryoprotection solutions (∼1 mL) should be prepared that have identical buffer, salt, and precipitant concentrations but the first containing 5% PEG400, the second 10%, the third 15%, and the final 20% (see Note 5). 2. In a 3-well glass plate, pipet 40 μL of cryoprotection solution 1 into a well and set it aside until later. 3. Access the protein drop: a. In a sitting drop plate, use a scalpel to cut around the tape sealing a particular well. Be careful when removing the cut tape to avoid disturbing the protein drop. b. In a hanging drop plate, grasp the edge of the cover slide for a particular well with either forceps or fingertips and gently lift one side; be careful to avoid disturbing the protein drop. 4. Once the protein drop is accessible, the crystals must be removed from the well (see Note 6). Pipet 4 μL of the well solution onto the protein droplet to prevent the droplet from drying up during harvesting. Estimate the size of the crystals in the well and pick an appropriately sized crystal loop. Attach the crystal loop to a crystal wand and then use the loop to “fish” a single crystal from the protein droplet. 5. Transfer this crystal to the well containing 40 μL of cryoprotection solution 1. Allow the crystal to soak in this solution for 15 min. Pipet 40 μL of cryoprotection solution 2 into a separate well and then transfer the crystal to this well to soak for 15 min. Repeat this procedure for solutions 3 and 4. 6. Outside the anaerobic tent, prepare the set up for crystal freezing. Have a small bowl of liquid N2 ready for freezing. After the final cryoprotection soak, remove the tray with crystals from the anaerobic tent. Working quickly, pick the crystal out of the well using an appropriately sized crystal loop and plunge into liquid nitrogen. Cool the loop container in liquid nitrogen as well and, while still under liquid nitrogen, place the loop inside the container. 7. Keep crystal frozen in liquid nitrogen until data collection.

3.6. Data Processing

X-ray diffraction data collected on nitrogenase crystals can be processed using freely available software programs installed on

162

Roth and Tezcan

a personal computer. MOSFLM, which is available as part of the Collaborative Computational Project No. 4 (CCP4) program suite (http://www.ccp4.ac.uk/), can be used for data integration (download at http://www.mrc-lmb.cam.ac.uk/harry/mosflm). Data scaling, molecular replacement, and structure refinement can be accomplished with the SCALA, MOLREP, and REFMAC programs, which also are part of the CCP4 program suite (see Notes 7 and 8). Likewise, the freely available program COOT (download at http://www.biop.ox.ac.uk/coot/), which can be interfaced with REFMAC, can be used for manual refinement and placement of water molecules and metal clusters.

4. Notes 1. Diffraction quality nitrogenase crystals have been grown from a wide variety of solutions (see Table 10.1). All solutions used for crystallization should be filtered through a 0.2 μM filter and then thoroughly degassed with cycles of vacuum and argon before starting. More viscous solutions such as 100% glycerol or 50% PEG8000 may require longer degassing times to fully remove oxygen. 2. Dithionite solutions should be prepared freshly on the day crystal plates will be set up. Solid dithionite should be dissolved in a buffer compatible with the crystallization conditions. 3. Especially for solutions that contain high MW PEG’s (>2000) or glycerol and thus are highly viscous, it is recommended that they be premixed in 1.7 mL eppendorf tubes using a vortexer. Lower viscosity solutions can be directly mixed in the crystal plate reservoir. The volume of the well solution in each crystal well can be varied and will influence crystal growth. Typical volumes for well solutions for nitrogenase crystals are between 250 and 1000 μL. 4. The final volume of the protein–reagent drop and the ratio of protein to reagent can be varied. 5. If crystals were grown in the presence of a high concentration of glassing agent, for example, 20% glycerol, then additional cryoprotection is most likely unnecessary, and crystals can be flash frozen directly from the protein drop. 6. If PEGs or glycerol were used in the well solution, a film may develop over the protein droplet. Use a fine-pointed crystal tool and gently scrape the top of the protein drop to remove any film. Remove as much film as possible without disturbing any crystals present.

X-Ray Crystallography

163

7. Refinement with REFMAC requires library files to define the topologies and bond parameters for any non-proteinaceous ligands and cofactors such as metal clusters. The MoFePreduced P-cluster structure is located in the CCP4 library as “CLF” and the FeMoco as “CFM” (without atom X). The 4Fe:4S cluster in FeP is located in the library as “FS4.” If needed, these library files can be modified using the ccp4i program Sketcher. 8. Because of the internal twofold symmetry of both MoFeP and FeP, the noncrystallographic symmetry-restraint option can be used in REFMAC, particularly during refinement steps at low-resolution cutoffs (>2.0 Å).

References 1. Christiansen J, Dean DR, Seefeldt LC (2001) Mechanistic features of the Mo-containing nitrogenase. Annu Rev Plant Physiol Plant Mol Biol 52:269–295 2. Burgess BK (1990) The Iron-Molybdenum cofactor of nitrogenase. Chem Rev 90: 1377–1406 3. Madden MS, Krezel AM, Allen RM et al (1992) Plausible structure of the ironmolybdenum cofactor of nitrogenase. Proc Natl Acad Sci USA 89:6487–6491 4. Oliver ME, Hales BJ (1992) Structural information of the nitrogenase metal clusters deduced from paramagnetic interactions. J Am Chem Soc 114:10618–10623 5. Georgiadias MM, Chakrabarti KP, Woo D et al (1992) Crystallographic structure of the nitrogenase Iron protein from Azotobacter vinelandii. Science 257:1653–1659 6. Kim J, Rees DC (1992) Crystallographic structure and functional implications of the nitrogenase molybdenum-iron protein from Azotobacter vinelandii. Nature 360:553–560 7. Roe SM, Gormal C, Smith BE et al (1997) Crystallization and preliminary X-ray studies of nitrogenase component 1 (the MoFe protein) from Klebsiella pneumonia. Acta Cryst D 53:227–228 8. Mayer SM, Lawson DM, Gormal CA et al (1999) New insights into structure-function relationships in nitrogenase: a 1.6Å resolution X-ray crystallographic study of Klebsiella pneumoniae MoFe-protein. J Mol Biol 292:871–891 9. Kim J, Woo D, Rees DC (1993) X-ray crystal structure of the nitrogenase molybdenumiron protein from Clostridium pasteurianum at 3.0-A resolution. Biochemistry 32: 104–107

10. Schlessman JL, Woo D, Joshua-Tor L et al (1998) Conformational variability in structures of the nitrogenase iron proteins from Azotobacter vinelandii and Clostridium pasteurianum. J Mol Biol 280: 669–685 11. Chan MK, Kim J, Rees DC (1993) The nitrogenase FeMo-cofactor and P-cluster pair: 2.2 A resolution structures. Science 260:792–794 12. Peters JW, Stowell MHB, Soltis SM et al (1997) Redox-dependent structural changes in the nitrogenase P-cluster. Biochemistry 36:1181–1187 13. Einsle O, Tezcan FA, Andrade SLA et al (2002) Nitrogenase MoFe-protein at 1.16 Å resolution: a central ligand in the FeMocofactor. Science 297:1696–1700 14. Jang SB, Seefeldt LC, Peters JW (2000) Insights into nucleotide signal transduction in nitrogenase: structure of an iron protein with MgADP bound. Biochemistry 39:14745–14752 15. Jeong MS, Jang Se Bok (2004) Structural basis for the changes in redox potential in the nitrogenase Phe135Trp Fe protein with MgADP bound. Mol Cells 18: 374–382 16. Jang SB, Jeong MS, Seefeldt LC et al (2004) Structural and biochemical implications of single amino acid substitutions in the nucleotide-dependent switch regions of the nitrogenase Fe protein from Azotobacter vinelandii. J Biol Inorg Chem 9: 1028–1033 17. Sen S, Igarashi R, Smith A et al (2004) A conformational mimic of the MgATP-bound “on state” of the nitrogenase iron protein. Biochemistry 43:1787–1797

164

Roth and Tezcan

18. Sen S, Krishnakumar A, McClead J et al (2006) Insights into the role of nucleotidedependent conformational change in nitrogenase catalysis: structural characterization of the nitrogenase Fe protein Leu127 deletion variant with bound MgATP. J Inorg Biochem 100:1041–1052 19. Strop P, Takahara PM, Chiu H et al (2001) Crystal structure of the all-ferrous [4Fe4S]0 form of the nitrogenase iron protein from Azotobacter vinelandii. Biochemistry 40:651–656 20. Schindelin H, Kisker C, Schlessman JL et al (1997) Structure of ADPAlF4— stabilized nitrogenase complex and its implications for signal transduction. Nature 387: 370–376 21. Chiu H, Peters JW, Lanzilotta WN et al (2001) MgATP-bound and nucleotide-free structures of a nitrogenase protein complex between the Leu 127-Fe-protein and the MoFe-protein. Biochemistry 40: 641–650 22. Tezcan FA, Kaiser JT, Mustafi D et al (2005) Nitrogenase complexes: multiple docking sites for a nucleotide switch protein. Science 309:1377–1380 23. Sarma R, Barney BM, Keables S et al (2010) Insights into substrate binding at FeMocofactor in nitrogenase from the structure of

24.

25. 26.

27.

28.

29.

an alpha-70Ile MoFe protein variant. J Inorg Biochem 104:385–389 Sørlie M, Christiansen J, Lemon BJ et al (2001) Mechanistic features and structure of the nitrogenase α-Gln195 MoFe protein. Biochemistry 40:1540–1549 Schmid B, Ribbe MW, Einsle O et al (2002) Structure of a cofactor-deficient nitrogenase MoFe protein. Science 296:352–356 Mayer SM, Gormal CA, Smith BE et al (2002) Crystallographic analysis of the MoFe protein of nitrogenase from a nifV mutant of Klebsiella pneumoniae identifies citrate as a ligand to the molybdenum of iron molybdenum cofactor (FeMoco). J Biol Chem 277:35263–35266 Jang SB, Seefeldt LC, Peters JW (2000) Modulating the midpoint potential of the [4Fe-4S] cluster of the nitrogenase Fe protein. Biochemistry 39:641–648 Schmid B, Einsle O, Chiu H et al (2002) Biochemical and structural characterization of the cross-linked complex of nitrogenase: comparison to the ADPAlF4- -stabilized structure. Biochemistry 41: 15557–15565 Kim J, Rees DC (1992) Structural models for the metal centers in the nitrogenase molybdenum-iron protein. Science 257:1677–1682

Chapter 11 X-Ray Absorption Spectroscopy Serena DeBeer Abstract X-ray absorption spectroscopy (XAS) involves the excitation of core electrons to bound states localized on the photoabsorber and the eventual excitation of the photoelectron to the continuum. The resulting spectra are typically divided into two regions: (1) the edge region which provides electronic structure information and (2) the extended X-ray absorption fine structure (EXAFS) region, which provides information about the distance, number, and type of ligands. Here, a basic introduction to XAS theory, the information that can be obtained, and the experimental consideration are presented. The application of XAS to biological systems and the impact this has had on nitrogenase research are briefly highlighted. New experimental advances are described. Key words: XAS, EXAFS, nitrogenase, geometric structure, electronic structure.

1. Introduction 1.1. Background

X-ray absorption spectroscopy (XAS) provides a powerful probe of the local geometric and electronic structure surrounding a transition metal absorber and has had a particularly profound impact on our understanding of metal active sites in biological systems. In the last ∼30 years, the active sites of numerous metalloproteins have been characterized by extended X-ray absorption fine structure (EXAFS), often in the absence of any crystallographic data. EXAFS has been used to uniquely characterize many intermediates in enzymatic reactions, including the oxygen intermediates in the reaction cycles of methane monooxygenase (1), ribonucleotide reductase (2), P450 and chloroperoxidases (3), and multicopper oxidases (4), to name only a few. The current view of the Mn4 Ca cluster in photosystem II has been greatly influenced by the results of detailed extended X-ray

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_11, © Springer Science+Business Media, LLC 2011

165

166

DeBeer

absorption fine structure (EXAFS) studies (5, 6). Of particular note, in the context of this book, is that the first proposal of the structure of the FeMoco active site of nitrogenase was based on EXAFS studies, which preceded the crystal structure by 14 years (7). Though there were some differences between the EXAFS proposed structure and the crystallography, the similarities are striking, and testimony to the power of the technique even when interpreted with relatively simplistic models. In the years that have followed, the theoretical understanding of XAS (both in the XANES and EXAFS regions) has greatly matured and it is apparent that this is a method that is still poised to answer many key questions in bioinorganic chemistry. In this chapter, we briefly introduce the methods, the basic theory behind it, and the experimental considerations. Finally, new experimental advances, which enable increased selectivity (for oxidation state and spin state), are briefly described. For a more detailed review of XAS the reader is referred to (8, 9). An XAS edge results when a core electron absorbs energy equal to or greater than its binding energy. XAS edges are labeled according to the shell the electron originates from. A K-edge corresponds to a 1s core level; L-edges correspond to 2s and 2p levels; and M-edges to 3s, 3p, and 3d levels, as pictured in Fig. 11.1. For reference, the Fe K-edge occurs at 7.1 keV, the L-edges at ∼700 eV, and the M-edges between 50 and 100 eV. Due to the low energies, the Fe L- and M-edges require high vacuum conditions. For transition metals of biological interest, most experiments are performed at the metal K-edge. This is due in part to the greater ease of conducting experiments in the hard X-ray region (5–30 keV) and the reduction in beam-induced damage, due to the longer absorption path length. An XAS experiment involves monitoring the absorption coefficient, μ, as a function of energy. The resultant spectrum can be divided into two regions the “edge” or X-ray absorption near edge (XANES) region and the EXAFS, as shown in Fig. 11.2. The XAS edge is characterized by a sharp discontinuity (or increase in absorption), which results from the ionization of a core electron. The lower energy region of the spectrum,

Fig. 11.1. Atomic energy level diagram for an absorbing atom and the corresponding X-ray absorption edges.

X-Ray Absorption Spectroscopy

167

Fig. 11.2. The X-ray absorption near edge (XANES) and extended X-ray absorption fine structure (EXAFS) regions of an X-ray absorption spectrum.

which contains transitions to unoccupied orbitals localized on the photoabsorber is referred to as the pre-edge. After the electron has been ionized and has “escaped” to the continuum, it is able to interact with its neighboring atoms. As the photoelectron is deflected off the potential of the neighboring atoms, an interference pattern is generated known and the EXAFS region. This region primarily provides geometric structural information local to the photoabsorber. 1.2. The Edge or XANES Region

The edge region primarily provides electronic structure information. In the case of a K-edge, the intense edge feature results from as electric dipole allowed 1s to 4p transition (Fig. 11.3).

Fig. 11.3. The edge and pre-edge regions for Td (solid line) and Oh (dashed line) ferric complexes.

168

DeBeer

Fig. 11.4. Comparison of the normalized Fe K-edges for oxidized (solid line) and reduced (dashed line) FeS4 model complexes (a) and ferricyanide (solid line) and ferrocyanide (dashed line) (b).

Superimposed on the rising edge are weak pre-edge transitions, which correspond to 1s to 3d transitions. These transitions are formally electric dipole forbidden, but quadrupole allowed and can gain intensity through 3d-4p mixing in suitable symmetry. Hence pre-edges of centrosymmetric complexes (e.g. Oh ) are weak, while pre-edge of non-centrosymmetric complexes (e.g. Td ) have significantly higher intensities (Fig. 11.3). Metal K-edges are most frequently used to identify the oxidation state of an absorbing atom. In a simple picture, as a metal becomes more oxidized it takes more energy to ionize a core electron and the edge shifts up in energy. Figure 11.4a compares the Fe K-edge of ferrous and ferric FeS4 model complexes. For a change in one unit of oxidation state, the pre-edge shifts by ∼1 eV. However, even when the oxidation state is constant, changes in the local geometry can induce a pronounced change in the edge. This is clearly illustrated when comparing the data for ferro- and ferricyanide as shown in Fig. 11.4b. Edges are thus often used as a “fingerprint” of the local geometric and electronic structure. However, a more rigorous interpretation of the edge can provide more detailed insights. Recently significant progress has been made in the application of both multiple scattering and DFT based approaches to understanding the edge region. The reader is referred to (10–12) for a few representative examples. 1.3. The EXAFS Region

The EXAFS region primarily provides metrical information for the number and type of ligands surrounding a photoabsorber. When the photoelectron has been excited to the continuum it can be modeled as a wave, which propagates out from the photoabsorber and is backscattered by the electron density surrounding the neighboring atoms. The EXAFS region is defined in terms of χ(k), where

X-Ray Absorption Spectroscopy

χ(k) =

169

Nf (k) μ(k) − μ0 (k) 2 2 ; χ(k) = e −2k σ sin [2kR + δ(k)] 2 μ0 (k) kR s [1]

and k=

2me (E − E0 )

[2]

h¯ 2

In equation [1] μ(k) is the observed absorption coefficient and μ0 (k) is the atomic absorption (or the absorption coefficient that corresponds only to the absorbing atom, without any surrounding ligands). For convenience, EXAFS is expressed in terms of photoelectron wave vector space (k) rather than energy space. The relationship between energy (in eV) and k (in Å–1 ) is given in equation [2]. Here E0 defines the point at which the EXAFS start (i.e., k = 0) and represents the ionization threshold energy for the electron. In its simplest form EXAFS can be approximated as photoelectron scattering, where χ(k) =

e i2kR s

kR

kf (k)e iδ(k)

e i2kR kR

[3]

This equation can be expressed as a sine function, where χ(k) =

Nf (k) s

kR2

sin [2kR + δ(k)]

[4]

Here N refers to the number of scatterers of a given type, f(k) is the amplitude function for the backscattering atom, δ(k) is the phase shift for the absorber–backscatterer pair, and R is the absorber–scatterer (i.e., metal–ligand) distance. Averaging over all atoms and introducing the disorder parameter σ 2 then gives the EXAFS equation for a single absorber– scatterer pair, as follows: χ(k) =

Nf (k) s

kR2

e −2k

2σ 2

sin [2kR + δ(k)]

[5]

The contributions of each parameter to the EXAFS data are best illustrated through examples. Figure 11.5 shows the dependence of the EXAFS data on the metal–ligand distance (R) together with the Fourier transform of the data into R-space (i.e., Å). The Fourier transform allows one to visualize the radial distribution of electron density with respect to the central absorbing atom. As the Fe–O bond

170

DeBeer

Fig. 11.5. FEFF calculated FTs and EXAFS (inset) for an Fe–O interaction as it is elongated from 2 to 3 to 4 Å.

is elongated from 2 to 3 to 4 Å, the amplitude of the EXAFS decreases and the frequency increases (Fig. 11.5, inset). This is more readily visualized by examining the Fourier transform of the data, which shows the intensity of a scattering contribution at a given distance from the photoabsorber. It should be noted that the Fourier transforms are not phase shift corrected, and thus the distances cannot be read directly from the Fourier transform. In any case this plot clearly indicates the decrease in intensity as the Fe–O distance is increased. As illustrated by equation [5], the data have a 1/R2 dependence demonstrating that this is a local probe of metrical structure. As also indicated by equation [5], the EXAFS will have a linear dependence on coordination number. As the coordination number (N) of a given backscatterer increases the amplitude of the EXAFS signal and the intensity of the Fourier transform increase linearly. A similar effect is seen for a change in the disorder parameter (σ 2 ), where an increase in disorder decreases the amplitude of both the EXAFS and the FT, with an exponential dependence. The identity of the backscatterer will affect both the phase and amplitude of the EXAFS signal. Figure 11.6 shows the calculated FTs and EXAFS for Fe–O, Fe–S, and Fe–Fe vectors. For higher z backscatterers the FT intensity increases and the amplitude envelope of the EXAFS peaks at a higher k-value. This allows EXAFS to be used to determine the identity of the atoms ligated to an absorber. However, it should be noted that similar backscatterers will have similar phase and amplitude profiles, thus while EXAFS can distinguish oxygen from sulfur, it cannot distinguish oxygen from nitrogen. In general for first row atoms atomic numbers of Z ± 1 cannot be distinguished. Moving down the periodic table

X-Ray Absorption Spectroscopy

171

Fig. 11.6. FEFF calculated FTs and EXAFS (inset) for Fe–O, Fe–S, and Fe–Fe interactions all at a distance of 2.5 Å.

the required change in Z becomes larger, as the change in one electron becomes a smaller perturbation on the change in overall electron density surrounding the scatterer. In fitting EXAFS data the goal is to determine the sum of all the absorber–scatterer interactions that make up the total EXAFS signal. Using the EXAFS equation this may be described as a sum of damped sine waves in k-space. The radial distribution of backscatterers with respect to the absorber is most readily visualized by examination of the Fourier transform.

2. Materials 2.1. X-Ray Sources

EXAFS measurements on biological samples utilize intense tunable X-ray sources, provided by synchrotron facilities. A typical experimental setup is pictured in Fig. 11.7. The experimental setup uses a crystal monochromator for energy selection using Bragg reflection. The monochromator is scanned in energy from below the ionization energy of the absorber to well above the absorption edge. A standard EXAFS scan spans an energy range of ∼1000 eV. Longer EXAFS scans have the advantage of improved resolution (as discussed in the data analysis section), though collecting high k-data also can present significant challenges.

2.2. Experimental Setup

Typically, the incident beam intensity is measured using a gasfilled ionization chamber. For very concentrated samples (i.e., with a high mole ratio of absorbing atoms) a gas-filled ionization

172

DeBeer

Fig. 11.7. Experimental setup for a typical X-ray absorption experiment.

chamber after the sample can be used to measure the absorption of X-rays directly. For biological samples, however, this is not typically possible as the samples are far too dilute and the scattering dominates. For dilute biological samples, fluorescence detection is typically used. After the excitation of a core electron the core hole will be refilled by an electron from a higher energy shell and a fluorescent photon will be produced. When a 2p electron refills a 1s hole, the resultant fluorescence emission line is referred to as a K-alpha line. As the K-alpha emission line is the most probable event, the K-alpha fluorescence is most typically used for detection. This assumes that the number of emitted K-alpha photons is proportional to the number of photon absorbed. In most cases a reference foil is measured simultaneously with the sample for purposes of energy calibration, as pictured in Fig. 11.7. For most biological samples many scans will need to be taken and the results averaged in order to extract the EXAFS data.

3. Methods (See Notes for Advantages and Limitations of XAS) 3.1. Methodology

Once the raw XAS data have been obtained, the data will need to be processed to extract both XANES and EXAFS data. First, all scans are energy calibrated and averaged together. Then a pre-edge subtraction is performed in order to subtract the total absorption from the absorption edge of interest (i.e., to get removed of absorption due to lighter atoms/lower energy edges). At this stage, the post-edge can be normalized to one and the normalized edge or XANES data are available. In order to obtain the EXAFS a spline function is removed from the data. The spline models the atomic background or the background due to the photoabsorber in the absence of any ligands. As discussed above, EXAFS data are most typically expressed in k-space or photoelectron wave vector space rather than energy space. The resultant EXAFS data are typically k3 -weighted in order to enhance

X-Ray Absorption Spectroscopy

173

oscillation at high k, as the EXAFS signal dies out as the distance from the photoabsorber increases. When XAS data are fit they are typically modeled in k-space, though in principle one can also fit in R-space. The goal of fitting the data is to deconvolute the total signal into the component waves. Early EXAFS analyses used known model complexes to extract phase and amplitude parameters and then refined these parameters against an unknown. However, there are now many standard packages that calculate all the possible scattering paths associated with a given model. These include programs such as FEFF and GNXAS (13, 14). The calculated paths include both single scattering paths (i.e., when the photoelectron is scattered by the electron density of only one other atom before returning to the photoabsorber) and multiple scattering paths (i.e., paths where the photoelectron is scattered by more than one atom before returning to the photoabsorber). A detailed discussion of multiple scattering analyses is beyond the scope of this brief review. 3.2. Applications to Nitrogenase and New Directions

As mentioned in the introduction to this chapter, XAS and in particular EXAFS have had a profound impact on our understanding of nitrogenase. The first structural characterization of the FeMoco active site emerged from EXAFS. Initial Mo K-edge XAS studies by Cramer and Hodgson revealed the coordination site around the molybdenum as having primarily sulfur ligation and showed clear evidence for the presence of Mo = O interactions (15). This early study also shows that the Mo was part of a cluster containing Fe atoms. Later comparative model studies lead them to propose a structure based on the EXAFS, which was later shown by crystallography to be largely accurate (7). Considering the complexity of this cluster this was a remarkable achievement. In the years that have followed, XAS has been utilized in numerous studies, which have included to examination of the iron protein of nitrogenase in various redox states, examinations of nitrogenase biosynthesis, and addressing fundamental questions about the presence of an interstitial atom in FeMoco to name only a few (16–18). Singlecrystal XAS studies of nitrogenase have shown promise; however, they have not yet to our knowledge been fully explored in light of the high-resolution crystallography (19). Undoubtedly XAS studies of nitrogenase are in part hampered by the number of iron present in the active site (of both FeMoco and the iron protein) making the resultant average signal difficult to deconvolute. It is here that high-resolution detection methods utilizing the X-ray emission lines may provide new insights. By using the spin selectivity of the K-beta lines it has been shown that XAS and EXAFS data for Fe sites in different oxidation states may be separated (20). This could allow the redox states of the FeMoco factor to be explored in more detail. In addition, by applying oxidation

174

DeBeer

state-dependent diffraction, as employed by Rees and coworkers on ferredoxin, the redox sites could also be spatially resolved (21). Many key details of nitrogenase and particularly the FeMoco active site have yet to be revealed and XAS spectroscopy (in particular when combined with other methods) is poised to make continued contributions.

4. Notes XAS provides an element-specific probe of the local geometric and electronic structure about a photoabsorber. Essentially any metal of interest may be examined by these methods provided the K- or L-edge is accessible using a synchrotron X-ray source. For most biological XAS, the readily accessible range of experiments spaces the metal K-edges of Ca (4 keV) to Cd (27 keV). Samples may be examined in any form—solution, single crystal, or lyophilized proteins—though solution samples are most common. Typically concentrations on the millimolar level in the absorbing atom of interest are required, though micromolar and even nanomolar concentrations have been successfully measured. Analysis of the EXAFS data provides coordination numbers (within ± 25%), ligand identity (within Z ± 1 for first row scatterers, i.e., C, N, and O cannot be distinguished), and very accurate average metal–ligand bond distances (within ± 0.02 Å). The primary limitation is that the average of all photoabsorbers will contribute to the resultant signal. This means that for complex metalloprotein active sites, such as nitrogenase, there is no means to readily distinguish the contribution from different iron sites. In addition, in the characterization of protein intermediates the composition of the sample must be carefully characterized by other methods. For iron-containing metalloproteins Mössbauer is an ideal way to determine speciation of a sample, as it provides a superposition of all components, rather than an average. EPR and optical spectroscopy provide additional means for assessing sample composition. One of the primary limitations of XAS is photoreduction of biological active sites due to the X-ray beam (6, 22). This may be partially alleviated by running samples at liquid helium temperatures; however, for many biological active sites data collection is often challenging or even prohibitive. One must be extremely careful with the X-ray dose and exposure to know that the XAS-characterized active site corresponds to the site of interest. For this reason parallel spectroscopic measurements on the same sample both before and after synchrotron exposure are highly desirable. At some synchrotron facilities combined capabilities for XAS and optical measurements exist which can greatly aid in these

X-Ray Absorption Spectroscopy

175

assessments. Similarly great insights have come from combined XAS and XRD studies on photosystem II (5, 6).

Acknowledgments We thank Martha A. Beckwith for helpful comments on this chapter and Cornell University for generous financial support.

References 1. Shu LJ, Nesheim JC, Kauffmann K et al (1997) An (Fe2 O2 ) diamond core structure for the key intermediate Q of methane monooxygenase. Science 275:515–518 2. Burdi D, Willems JP, Riggs-Gelasco P et al (1998) The core structure of X generated in the assembly of the diiron cluster of ribonucleotide reductase: O-17(2) and (H2 O)-O-17 ENDOR. J Am Chem Soc 120: 12910–12919 3. Stone KL, Behan RK, Green MT (2005) X-ray absorption spectroscopy of chloroperoxidase compound I: Insight into the reactive intermediate of P450 chemistry. Proc Natl Acad Sci USA 102:16563–16565 4. Lee SK, DeBeer George S, Antholine WE et al (2002) Nature of the intermediate formed in the reduction of O2 to H2 O at the trinuclear copper cluster active site in native laccase. J Am Chem Soc 124:6180–6193 5. Yano J, Kern J, Sauer K et al (2006) Where water is oxidized to dioxygen: Structure of the photosynthetic Mn4Ca cluster. Science 314:821–825 6. Yano J, Kern J, Irrgang KD et al (2005) X-ray damage to the Mn4Ca complex in single crystals of photosystem II: A case study for metalloprotein crystallography. Proc Natl Acad Sci USA 102:12047–12052 7. Wolff TE, Berg JM, Warrick C et al (1978) Molybdenum-iron-sulfur cluster complex [Mo2 Fe6 S9 (SC2 H5 )8 ]3 - - Synthetic approach to molybdenum site in nitrogenase. J Am Chem Soc 100:4629–4632 8. Koningsberger DC, Prins R (1988) X-Ray Absorption, Principles, Applications Techniques of EXAFS, SEXAFS, and XANES. Wiley, New York, NY 9. Scott RA (2000) X-ray absorption spectroscopy. In: Que L Jr (ed) Physical Meth-

10.

11.

12.

13.

14.

15.

16.

17.

ods in Bioinorganic Chemistry, pp. 465–503. University Science Books, Sausalito, CA DeBeer George S, Neese F (2010) Calibration of scalar relativistic density functional theory for the calculation of sulfur K-edge X-ray absorption spectra. Inorg Chem 49:1849–1853 DeBeer George S, Petrenko T, Neese F (2008) Prediction of iron K-edge absorption spectra using time-dependent density functional theory. J Phys Chem A 112:12936– 12943 Ankudinov AL, Ravel B, Rehr JJ et al (1998) Real-space multiple-scattering calculation and interpretation of x-ray-absorption near-edge structure. Phys Rev B 58:7565– 7576 Ankudinov A, Conradson S, Rehr JJ (1997) Self-consistent calculations of XANES in Pu hydrates. Abstr Pap Am Chem S 214:52GEOC Filipponi A, Dicicco A, Tyson TA et al (1991) Ab initio modeling of X-ray absorption-spectra. Solid State Commun 78:265–268 Cramer SP, Hodgson KO, Gillum WO et al (1978) Molybdenum site of nitrogenase – preliminary structural evidence from x-ray absorption spectroscopy. J Am Chem Soc 100:3398–3407 Musgrave KB, Angove HC, Burgess BK et al (1998) All-ferrous titanium(III) citrate reduced Fe protein of nitrogenase: An XAS study of electronic and metrical structure. J Am Chem Soc 120:5325– 5326 Corbett MC, Hu YL, Fay AW et al (2006) Structural insights into a protein-bound ironmolybdenum cofactor precursor. Proc Natl Acad Sci USA 103:1238–1243

176

DeBeer

18. George SJ, Igarashi RY, Xiao Y et al (2008) Extended X-ray absorption fine structure and nuclear resonance vibrational spectroscopy reveal that NifB-co, a FeMo-co precursor, comprises a 6Fe core with an interstitial light atom. J Am Chem Soc 130:5673– 5680 19. Einsle O, Tezcan FA, Andrade SLA et al (2002) Nitrogenase MoFe-protein at 1.16 angstrom resolution: A central ligand in the FeMo-cofactor. Science 297:1696–1700 20. Glatzel P, Bergmann U (2005) High resolution 1s core hole X-ray spectroscopy in

3d transition metal complexes – electronic and structural information. Coord Chem Rev 249:65–95 21. Einsle O, Andrade SLA, Dobbek H et al (2007) Assignment of individual metal redox states in a metalloprotein by crystallographic refinement at multiple X-ray wavelengths. J Am Chem Soc 129:2210 22. Corbett MC, Latimer MJ, Poulos TL et al (2007) Photoreduction of the active site of the metalloprotein putidaredoxin by synchrotron radiation. Acta Crystallogr D 63:951–960

Chapter 12 Small Angle X-Ray Scattering Spectroscopy David W. Mulder and John W. Peters Abstract Conformational changes imposed upon the Fe protein during binding and hydrolysis of Mg·ATP are key to initiating the cycle of interactions within the nitrogenase complex that result in gated electron transfer and the eventual multiple electron reduction of dinitrogen to ammonia. Wonderful insights into how nitrogenase accomplishes this have been gleaned from a number of very nice crystal structures, but conformational changes can only be inferred in a very superficial manner, and a number of key conformations relevant to fully understanding conformational changes have eluded this approach. Alternatively, small angle x-ray scattering (SAXS) has proven to be a helpful method for complimenting x-ray crystallography and determining solution structures of various nucleotide-bound states. Key words: Nitrogenase, complex, structure, conformation, SAXS.

1. Introduction Biological nitrogen fixation is an involved process coupling protein association, nucleotide binding and hydrolysis, and unidirectional electron flow to a chemical reaction that is essential for making nitrogen readily available in a form that can be utilized by living systems. The molybdenum nitrogenase exists as a complex from two proteins, the Fe and MoFe proteins, which are encoded by the genes nifH and nifDK, respectively (1). Interactions between the Fe and MoFe proteins make possible the electron flow required for the reduction of dinitrogen to ammonia at the catalytic sites located in the MoFe protein (2). The Fe protein, a homodimer (α2 ), contains a [4Fe–4S] cluster at the dimer interface between the two subunits and a binding site for Mg·ATP. Conformational changes of the Fe protein are induced by binding M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_12, © Springer Science+Business Media, LLC 2011

177

178

Mulder and Peters

and hydrolysis of Mg·ATP (3). The MoFe protein, a heterotetramer (α2 β2 ), contains two types of complex FeS clusters: the P-clusters and FeMo-cofactors (FeMo-cos). The P-clusters ([Fe8 –S7 ]) function to mediate electron transfer from the [4Fe–4S] cluster at the Fe protein to the FeMo-cos ([Mo–7Fe– 9S–X–homocitrate], where X = C, O, N) at the MoFe protein, where N2 fixation occurs (4). Throughout catalysis by nitrogenase, the Fe and MoFe proteins repeatedly associate and dissociate with each other during electron transfer (2). This process, caused by the binding and hydrolysis of Mg·ATP at the Fe protein, creates gated and unidirectional electron flow from the Fe protein to the MoFe protein (3–7). Since eight electrons are required for the fixation process (N2 + 8 H+ + 8 e − + 16 ATP → 2 NH3 + H2 + 16 ADP + 16 Pi ), accumulation of electrons at the FeMo-cos requires multiple nucleotide binding and hydrolysis cycles at the Fe protein. Gated and unidirectional electron flow is essential to the process since only one electron can be transferred at a time in a sequential mechanism from the Fe protein to the MoFe protein (8– 11). Binding and hydrolysis of Mg·ATP coupled to Fe protein conformation change is the underlying factor for maintaining gated and unidirectional electron flow to the FeMo-cos. Interestingly, this function of coupling Mg·ATP binding and hydrolysis to electron flow is not common in biology and provides further interest toward the mechanism and role behind N2 fixation. Structurally, the Fe protein and complex formation between the Fe and MoFe proteins has been well studied by x-ray crystallography, and a number of different states in the presence and absence of nucleotides have been determined (12–19). The structure of the Mg·ATP-bound Fe protein state has remained elusive however, and this critical structure has not been able to be determined by x-ray crystallography. The structural features of this particular state potentially could expose key features behind the mechanism for gated and unidirectional electron transfer from the Fe protein to the MoFe protein during catalysis. Other approaches to x-ray crystallography are therefore desired to solve the mysteries of how nucleotide-dependent conformation changes interplay and relate to electron transfer in N2 fixation. We have implemented the technique of small angle x-ray scattering (SAXS) as a complimentary approach to x-ray crystallography to probe the unknowns behind nucleotide-dependent conformation change in nitrogenase (20). Much like x-ray crystallography, SAXS detects the scattering of x-rays by protein electrons. In contrast to x-ray crystallography, SAXS probes the solution state of proteins and detects scattering from groups of electrons from a protein in solution as opposed to single electrons from a protein in a static or crystalline state. Although SAXS does not yield high-resolution crystal structures as the technique of

Small Angle X-Ray Scattering Spectroscopy

179

x-ray crystallography, it does provide key information regarding the shape and size of proteins in solution. The results of this study show how different nitrogenase proteins and nucleotide-bound states, including the elusive Mg·ATP-bound Fe protein state, behave in regard to overall conformation in solution. This study highlights many experimental features of SAXS which makes it a valuable resource for answering broad research questions in structural biology and establishes general protocols for SAXS when utilizing for research which may involve different nitrogenase protein conformations, complexes, and nucleotide-bound states. Our particular study took advantage of previously solved x-ray crystal structures to approach probing the unknown Mg·ATP-bound Fe protein state (14, 17, 19). Central to this strategy was the crystal structure of a variant form (L127) of the Fe protein (17, 19). The L127 Fe protein contains a single deletion in the switch region between the Mg·ATP binding site and [4Fe–4S] cluster. In comparison to the native Fe protein structure, it is not compact and the two subunits are opened up in a rigid body motion to create a more elongated form of the Fe protein structure. This gave thought that the structure may be reflective of the Mg·ATP-bound state especially since both protein forms exhibit similar spectroscopic properties (21). We tested this hypothesis by SAXS and concluded that the solution Mg·ATPbound Fe protein state actually resembles the nucleotide-free Fe protein state. Importantly, this work set forth much groundwork for future studies implementing SAXS to compliment x-ray crystallography for the multitude of different protein states and conformations present during the catalysis of N2 fixation.

2. Materials 2.1. Initial Sample Preparation

1. Anaerobic sample buffer: 50 mM Tris pH 7.5, 0.5 M NaCl, 20% glycerol, 5 mM sodium dithionite (see Note 1). Store at room temperature in glove box (Coy Laboratory Products, Inc.). 2. Gel filtration resin: SephacrylTM S-300 (GE Healthcare). 3. Syringe: 5 mL Luer-LokTM (BD Syringe).

2.2. Sample Preparation at the Synchrotron Center

1. Nitrogen atmosphere glove box (Vacuum Atmospheres, Hawthorne, MA) operating in an environment less that 1 ppm level of oxygen. 2. Sealable glass sample vials: 3 mL (Wheaton), 13 mm aluminum seals and septa (Wheaton).

180

Mulder and Peters

Fig. 12.1. Overview of the SAXS Beamline 4-2 at the Stanford Synchrotron Radiation Laboratory (SSRL) and zoom into features of interest including the detector, sample area, and flow cell apparatus.

2.3. Beamline Preparation

1. Beamline 4-2 at the Stanford Synchrotron Radiation Laboratory (SSRL) is set up for SAXS data collection (22) (see Fig. 12.1 for overview of SAXS Beamline 4-2 at SSRL; see Note 2). 2. Detector: MarCCD 165 (MAR USA, Evanston, IL). 3. Cholesterol myristate and silver behanate (available at beamline 4-2 at SSRL). 4. Continuous flow cell: thin wall, oxygen impermeable x-ray capillary tubing connected to a computer controlled syringe dispenser (Hamilton 500 series, 250 μL syringe) (see Note 3). 5. Blu-ICE graphical interface (available at SSRL beamlines).

2.4. Initial Calibration

1. Reagent grade lysozyme (Fischer Scientific) at four different concentrations (2.5, 5, 10, and 20 mg/mL) in 10 mM ammonium acetate (pH 5.0) and 150 mM NaCl. 2. Program CRYSOL (23) (available for download at: http:// www.embl-hamburg.de/biosaxs/software.html).

2.5. Data Collection

1. Hamilton removable syringe needle (22 gauge, Cat. No: 7779-01). R (0.7 × 40 mm, Henke Sass 2. Syringe needle: Fine-Ject Wolf).

Small Angle X-Ray Scattering Spectroscopy

2.6. Nucleotide Addition and Data Collection

181

1. ADP and ATP (Sigma) stock solutions (100 mM) in 1 M Tris pH 8–8.5. Do not store and prepare only immediately prior to use (see Note 4). 2. 100 mM MgCl2 stock solution. 3. Gas-tight syringe (50 μL, Hamilton).

2.7. Data Processing, Analysis, and 3D Shape Reconstruction (Ab Initio)

1. Program MarParse (22) (available at Beamline 4-2 at SSRL). 2. Programs PRIMUS (24), CRYSOL (23), GNOM (25), DAMMIN (26), SUPCOMB (27), and DAMAVER (28) (available for download at: http://www.embl-hamburg.de/ biosaxs/software.html).

3. Methods 3.1. Initial Sample Preparation

1. Carry out expression and purification of the native Fe protein and L127 Fe protein as previously described (21). 2. After purification, exchange both proteins into anaerobic sample buffer using gel filtration chromatography (see Notes 1, 5–6). 3. Concentrate the purified proteins between 30 and 100 mg/mL using an Amicon cell apparatus under argon pressure (45 psi) (see Note 7). 4. Flash freeze the concentrated protein stocks at –80◦ C by dripping them in a periodic manner from a gas-tight syringe directly into liquid nitrogen. Make sure to wash syringe with anaerobic sample buffer prior to transfer from the concentration cell. The desired pellet size of the protein after freezing is c. 30–50 μL. 5. Store samples in liquid nitrogen using a small plastic container or Eppendorf tube (see Note 8). 6. Transfer frozen samples to the synchrotron center using a dry liquid nitrogen shipper.

3.2. Sample Preparation at the Synchrotron Center

1. Transfer protein stocks into nitrogen atmosphere glove box (see Note 9). 2. For initial data collection, prepare at least four different protein concentrations (2.5, 5, 10, and 20 mg/mL) by diluting the protein stock with the appropriate anaerobic sample buffer to 100 μL into sealable sample vials (see Notes 10 and 11). 3. Crimp and seal the vials and deliver them to the beamline just prior to data collection.

182

Mulder and Peters

3.3. Beamline Preparation

1. Set incident beam energy to 11 keV for initial experiments. 2. Position detector 1 m from the sample cell (see Note 12). 3. Calibrate the detector pixels to the momentum transfer Q (4π sin(θ)/λ, see Note 13) using the (100) and related higher order reflections from a cholesterol myristate or silver behanate powder sample. 4. Verify that the syringe dispenser and flow cell can be controlled remotely with the Blu-ICE graphical interface. 5. Set temperature controller for the flow cell sample area to 20◦ C.

3.4. Initial Calibration

1. Calibrate the experimental setup using lysozyme (29). 2. Measure the scattering curve of each sample in triplicate at the same beam energy to that used for protein sample data collection. 3. Compare the experimental scattering curve to the theoretical curve generated from CRYSOL (23) using the Protein Data Bank structure 193L (30).

3.5. Data Collection

The basic data collection strategy is to use an anaerobically maintained flow cell to minimize the negative effects of protein aggregation and x-ray radiation damage. For each protein sample, scattering data on the corresponding buffer is collected before and after to allow for proper background subtraction. 1. Prepare samples at the desired concentration only immediately before data collection is ready to take place. 2. Periodically wash and flush the flow cell with anaerobic water and anaerobic sample buffer to ensure the system is maintained at strict anaerobic conditions (see Note 14). 3. Attach sample vial containing protein buffer to the flow cell via Hamilton removable syringe needle. 4. Put vial under constant nitrogen over-pressure by attaching nitrogen input gas line with syringe needle. 5. Use a Hamilton removable syringe needle to release back pressure and to ensure that the flow cell does not pull vacuum on the system. 6. Load 100 μL of buffer from the flow cell using the BluICE interface. 7. Set up a continuous loop for the buffer passing it 30 μL in the forward and reverse direction (flow rate = 4 μL/sec) relative to the x-ray beam position. 8. Measure x-ray scattering with 20 exposures, each for duration of 3 s.

Small Angle X-Ray Scattering Spectroscopy

183

9. After the buffer scattering is measured, load a 10 μL N2 gap following the previously loaded buffer. 10. Load 100 μL of sample at 10 μL/sec following the N2 gap, and set up sample loop parameters identical to that used for protein buffer loop. Take extreme care to not introduce any bubbles into the flow cell. 11. Before collecting the data, lower the video scope to confirm that no N2 bubbles have been introduced into the sample loop (see Note 15). 12. Raise video scope. 13. Measure x-ray scattering with 20 exposures, each for duration of 3 s. 14. Load 10 μL N2 gap following sample and load 100 μL of corresponding protein buffer. 15. Collect data on buffer again using identical parameters for protein sample data collection. 3.6. Nucleotide Addition and Data Collection

1. Prepare nucleotide stock solutions in the glove box and transfer to the beamline in a sealed sample vial (see Note 4). 2. At the beamline, using a gas-tight syringe, add the nucleotide and MgCl2 stock solutions to the protein sample prepared in a separate sample vial so that the final concentration of ADP or ATP in the sample solution is 5 mm with 10 mM MgCl2 (see Note 16). 3. Allow to incubate for 5 min before immediately collecting data on the sample (see Note 17). 4. Collect scattering data of the nucleotide-containing buffer before and after data collection on the nucleotide-bound protein sample (see Note 18).

3.7. Data Processing, Analysis, and 3D Shape Reconstruction (Ab Initio)

1. After collecting data on each individual sample, use the program MarParse (22) to subtract background scattering (buffer scattering data) (see Notes 19 and 20, Fig. 12.2). 2. With the program PRIMUS (24), evaluate the Guinier plot to calculate the radius of gyration (Rg ) by using the first 20 intensity points past the beam stop (see Notes 21 and 22). 3. For structures that have been solved by x-ray crystallography, use the program CRYSOL to generate theoretical scattering curves for comparison (see Note 23). 4. Calculate the election pair distance distribution function, P(r), using the program GNOM (25) (see Note 24, Fig. 12.2).

184

Mulder and Peters

Fig. 12.2. Typical SAXS pattern (a), scattering curve (b), and electron pair distribution function plot (c) for the Fe protein of nitrogenase.

5. Use DAMMIN (26) to perform ab initio shape reconstruction calculations to obtain a 3D model from the distribution function (see Note 25). 6. Run at least 10 separate calculations and analyze them for spatial discrepancy using SUPCOMB (27) from DAMAVER (28). The final model is the model representing the most probable volume among all the separate models used in the calculation.

Small Angle X-Ray Scattering Spectroscopy

185

4. Notes 1. Give special attention to maintaining anaerobicity during the purification process. Degas all buffers using a Schlenk line with periodic exchange between nitrogen gas and vacuum pressure. Typically, 1 L of buffer should be degassed with constant exchanging for a minimum of 1 h. After degassing, transfer buffers to a glove box and crimp seal in 8 or 16 mL Wheaton sample vials. 2. SAXS experiments can also be carried out at the Advanced Photon Source Laboratory (Argonne National Laboratory) or the Lawrence Berkeley National Laboratory in the United States. All three laboratories (SSRL, APS, and LBNL) are operated by the U.S. Department of Energy. 3. The implementation of the flow cell into the beamline data collection strategy allows for x-ray exposure to be distributed evenly over the entire sample in different locations thus helping to eliminate the negative effects from x-ray radiation damage and protein aggregation. For the native and L127 Fe protein samples, implementation of a flow cell as opposed to a static cell eliminated x-ray radiation damage and protein aggregation. For samples where radiation damage and aggregation are not readily observed, a conventional flat-window cell made from thin mica may be used. 4. Prepare nucleotide stock solutions only immediately prior to time of data collection. Do not allow excessive time between making the solutions and data collection since nucleotide solutions may hydrolyze. 5. To help avoid aggregation at the beamline, purify the protein with a minimum of 5% glycerol present in the buffer. The presence of glycerol also helps counter radiation damage. If the protein is stable in glycerol, glycerol levels should be adjusted up to 20% to help avoid aggregation and radiation damage. Also, purification by gel filtration is desired and preparation of protein samples of high purity will also help avoid aggregation. 6. Both the native and the L127 proteins are stable with the 0.5 M NaCl present in the buffer. Although high levels of salt can increase background scattering noise, it is important to maintain a buffer concentration where the protein is stable to help avoid aggregation. 7. Both the native and the L127 proteins are stable at high concentrations and can be concentrated immediately

186

Mulder and Peters

after purification up to 100 mg/mL. For samples that may aggregate upon concentration, prepare at lower concentrations (1–5 mg/mL) and wait until just prior before data collection to concentrate the protein. 8. Since the protein was flash frozen in liquid nitrogen it is not important to seal the vials or plastic containers and because of safety it is desirable to leave them unsealed to prevent rapid explosion upon thawing at a later time. If storing in glass vials, do not submerge them into liquid nitrogen as liquid nitrogen may seep into the vial. For more safety details concerning safe handling of sealable glass vials at low temperatures, see http://ssrl.slac. stanford.edu/safety/advisory/useradvisory.html. 9. When transferring the samples into the glove box, allow for a small amount of liquid nitrogen to be present in the vials to ensure that the samples do not thaw until safely under the nitrogen atmosphere. 10. It is important for the early stages of the experiment to test a wide range of protein concentrations. Sample signal increases with sample concentration; however, high concentrated samples may not always be desired. Samples at high concentration my aggregate or oligomerize. Thus, it is necessary to consider a broad range of concentrations to find the correct balance between signal quality and intensity. Both the native and the L127 Fe protein are stable in solution at 5 mg/mL without aggregation and give strong signals at that concentration. 11. The final concentration of sodium dithionite in all anaerobic sample and buffer solutions should be 5 mM. It is important to maintain a high level of anaerobicity throughout experimentation. 12. Other sample-to-detector distances can be used, but for general experiments distances of 1 or 1.5 m are desirable. For collecting high angle data and focusing on interparticle interactions in the scattering curve, use shorter sample-to-detector distances (e.g., 0.25 m). 13. θ is one-half of the scattering angle and λ is the wavelength. 14. It is important to stay faithful in washing the flow cell with anaerobic water and anaerobic sample buffer. This ensures that the system is anaerobic and also prevents the flow cell from gumming up causing sample loading and flow issues which can be an issue due to the presence of glycerol in the sample buffer. 15. N2 bubbles in the sample loop will cause spikes in the scattering curve and make it not possible to process and average the collected data.

Small Angle X-Ray Scattering Spectroscopy

187

16. This concentration of nucleotides gives an excessive molar excess when the protein sample solution is 5 mg/mL and ensures that the protein will be in the nucleotide-bound form. Different concentrations can be used although the molar ratio of MgCl2 to ATP or ADP should always be 2:1. 17. Allowing the Mg·ATP solution to sit too long may result in hydrolysis to Mg·ADP and the addition immediately prior to data collection along with the short incubation time ensures that the protein sample will be in the Mg·ATPbound form. 18. It is always important to collect data on the exact buffer composition of that present for the corresponding sample and this should be done in a manner exact to data collection on the protein sample. 19. Process the data in parallel with the data collection process to monitor the quality of the data as the experiments progress and to ensure protein aggregation is not occurring. 20. MarParse processes the individual 2D data images and azimuthally integrates them, scaling them for the beam intensity, and averaging all of the frames resulting in the final scattering curve. 21. Use the Guinier plot to determine if the sample is aggregating. For aggregation, the Guinier plot will not be linear and sample data points will increase in intensity near I0 . 22. For calculating the Rg , use the Q range 0.023–0.052 Å–1 and 0.022–0.045 Å–1 for the native Fe protein and L127 Fe protein, respectively. 23. Theoretical scattering curves the native Fe protein and L127 protein can be calculated using CRYSOL from PDB files 2NIP (14) and 1RW4 (17), respectively. 24. Use the scattering range Q = 0.02–0.3 Å–1 and Dmax values, 70 and 80 Å for the native Fe protein and L127 Fe protein, respectively. Dmax values were calculated from the respective crystal structures and verified using GNOM. 25. For the native Fe protein and L127 protein, two-fold symmetry can be imposed on the calculation.

Acknowledgments Portions of the research carried out used to develop the protocols described in this chapter were conducted out at the Stanford Synchrotron Radiation Laboratory, a national user facility

188

Mulder and Peters

operated by Stanford University on behalf of the U.S. Department of Energy, Office of Basic Energy Sciences. The SSRL Structural Molecular Biology Program is supported by the Department of Energy, Office of Biological and Environmental Research and by the National Institutes of Health, National Center for Research Resources, Biomedical Technology Program, and the National Institute of General Medical Sciences.

References 1. Bulen WA, LeComte JR (1966) The nitrogenase system from Azotobacter: twoenzyme requirement for N2 reduction, ATPdependent H2 evolution, and ATP hydrolysis. Proc Natl Acad Sci USA 56:979–986 2. Hageman RV, Burris RH (1978) Nitrogenase and nitrogenase reductase associate and dissociate with each catalytic cycle. Proc Natl Acad Sci USA 75:2699–2702 3. Lanzilotta WN, Parker VD, Seefeldt LC (1998) Electron transfer in nitrogenase analyzed by Marcus theory: evidence for gating by MgATP. Biochemistry 37:399–407 4. Burgess BK, Lowe DJ (1996) Mechanism of molybdenum nitrogenase. Chem Rev 96:2983–3012 5. Howard JB, Rees DC (1994) Nitrogenase: a nucleotide-dependent molecular switch. Annu Rev Biochem 63:235–264 6. Seefeldt LC, Dean DR (1997) Role of nucleotides in nitrogenase catalysis. Acc Chem Res 30:260–266 7. Peters JW, Szilagyi RK (2006) Exploring new frontiers of nitrogenase structure and mechanism. Curr Opin Chem Biol 10:101–108 8. Lowe DJ, Thorneley RNF (1984) The mechanism of Klebsiella pneumoniae nitrogenase action. The determination of rate constants required for the simulation of the kinetics of N2 reduction and H2 evolution. Biochem J 224:895–901 9. Thorneley RNF, Lowe DJ (1984) The mechanism of Klebsiella pneumoniae nitrogenase action. Pre-steady-state kinetics of an enzyme-bound intermediate in N2 reduction and of NH3 formation. Biochem J 224: 887–894 10. Barney BM, Laryukhin M, Igarashi RY et al (2005) Trapping a hydrazine reduction intermediate on the nitrogenase active site. Biochemistry 44:8030–8037 11. Howard JB, Rees DC (2006) How many metals does it take to fix N2 ? A mechanistic overview of biological nitrogen fixation. Proc Natl Acad Sci USA 103:17088–17093

12. Georgiadis MM, Komiya H, Chakrabarti P et al (1992) Crystallographic structure of the nitrogenase iron protein from Azotobacter vinelandii. Science 257: 1653–1659 13. Schindelin H, Kisker C, Schlessman JL et al (1997) Structure of ADP x AIF4 (-)-stabilized nitrogenase complex and its implications for signal transduction. Nature 387:370–376 14. Schlessman JL, Woo D, Joshua-Tor L et al (1998) Conformational variability in structures of the nitrogenase iron proteins from Azotobacter vinelandii and Clostridium pasteurianum. J Mol Biol 280:669–685 15. Jang SB, Seefeldt LC, Peters JW (2000) Insights into nucleotide signal transduction in nitrogenase: structure of an iron protein with MgADP bound. Biochemistry 39:14745–14752 16. Schmid B, Einsle O, Chiu HJ et al (2002) Biochemical and structural characterization of the cross-linked complex of nitrogenase: comparison to the ADP-AlF4(-)-stabilized structure. Biochemistry 41:15557–15565 17. Sen S, Igarashi R, Smith A et al (2004) A conformational mimic of the MgATP-bound “on state” of the nitrogenase iron protein. Biochemistry 43:1787–1797 18. Tezcan FA, Kaiser JT, Mustafi D et al (2005) Nitrogenase complexes: multiple docking sites for a nucleotide switch protein. Science 309:1377–1380 19. Sen S, Krishnakumar A, McClead J et al (2006) Insights into the role of nucleotidedependent conformational change in nitrogenase catalysis: structural characterization of the nitrogenase Fe protein Leu127 deletion variant with bound MgATP. J Inorg Biochem 100:1041–1052 20. Sarma R, Mulder DW, Brecht E et al (2007) Probing the MgATP-bound conformation of the nitrogenase Fe protein by solution small-angle X-ray scattering. Biochemistry 46:14058–14066

Small Angle X-Ray Scattering Spectroscopy 21. Ryle MJ, Seefeldt LC (1996) Elucidation of a MgATP signal transduction pathway in the nitrogenase iron protein: formation of a conformation resembling the MgATP-bound state by protein engineering. Biochemistry 35:4766–4775 22. Smolsky IL, Liu P, Niebuhr M et al (2007) Biological small-angle X-ray scattering facility at the Stanford Synchrotron Radiation Laboratory. J Appl Cryst 40:s453–s458 23. Svergun D, Barberato C, Koch M (1995) CRYSOL – a program to evaluate x-ray solution scattering of biological macromolecules from atomic coordinates. J Appl Cryst 28:768–773 24. Konarev PV, Volkov VV, Sokolova AV et al (2003) PRIMUS: a Windows PC-based system for small-angle scattering data analysis. J Appl Cryst 36:1277–1282 25. Svergun D (1992) Determination of the regularization parameter in indirect-transform methods using perceptual criteria. J Appl Cryst 25:495–503

189

26. Svergun DI (1999) Restoring low resolution structure of biological macromolecules from solution scattering using simulated annealing. Biophys J 76:2879–2886 27. Kozin MB, Svergun DI (2001) Automated matching of high- and low-resolution structural models. J Appl Cryst 34:33–41 28. Volkov VV, Svergun DI (2003) Uniqueness of ab initio shape determination in smallangle scattering. J Appl Cryst 36:860–864 29. Krigbaum WR, Kugler FR (1970) Molecular conformation of egg-white lysozyme and bovine alpha-lactalbumin in solution. Biochemistry 9:1216–1223 30. Vaney MC, Maignan S, Ries-Kautt M et al (1996) High-resolution structure (1.33 A) of a HEW lysozyme tetragonal crystal grown in the APCF apparatus. Data and structural comparison with a crystal grown under microgravity from SpaceHab-01 mission. Acta Crystallogr D Biol Crystallogr 52: 505–517

Chapter 13 Electron Paramagnetic Resonance Spectroscopy Karamatullah Danyal, Zhi-Yong Yang, and Lance C. Seefeldt Abstract EPR spectroscopy has been an important tool in nitrogenase research for the last 50 years. The three metalloclusters in nitrogenase, the Fe protein [4Fe–4S] cluster, and the MoFe protein P-cluster, and FeMo-cofactor, all have EPR spectra when poised in the appropriate paramagnetic states. EPR spectroscopy can probe changes in the electronic properties of each metal cluster, such as when substrates bind, and can provide a definitive method for observing changes in the redox states of the clusters. In this chapter, the methods for analysis of the three metal clusters of nitrogenase by EPR spectroscopy are described, along with methods for trapping substrate-derived intermediates on the active site that are amenable to characterization by EPR and other magnetic resonance spectroscopy techniques. Key words: Nitrogenase, mechanism, intermediate, trapping substrate, electron paramagnetic resonance (EPR), electron spin resonance (ESR).

1. Introduction Electron paramagnetic resonance (EPR) spectroscopy has been a central tool in the investigation of nitrogenase mechanism since the 1970’s (1–4). EPR’s prominence in nitrogenase research results from the fact that nitrogenase contains three metal clusters that all can exist in paramagnetic states (3, 5–9) and from the power of EPR in detecting changes in redox states and electronic properties of paramagnetic metalloclusters at low temperature (1, 3, 10). For these reasons, EPR has proven to be useful in monitoring the redox states of the nitrogenase metal clusters and subtle changes to the electronic properties of the metal clusters, such as when substrates associate with the metal cluster.

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_13, © Springer Science+Business Media, LLC 2011

191

192

Danyal, Yang, and Seefeldt

EPR (sometimes called electron spin resonance or ESR) spectroscopy is widely used in the fields of inorganic and bioinorganic chemistry as a means to monitor paramagnetic metals (1). Only paramagnetic species (small number of unpaired electrons) are detected by this method, whereas diamagnetic (no unpaired electrons) and ferromagnetic (large number of unpaired electrons) species are not observable (11). For a paramagnetic species, the unpaired electrons align themselves with an external magnetic field. Each electron has both an intrinsic spin angular momentum and an orbital angular momentum. When the electron is placed within an external magnetic field (B0 ), it orients itself to this applied field. This interaction between the electron and the external magnetic field is known as the Zeeman interaction (Fig. 13.1) (12). To acquire an EPR spectrum, the applied magnetic field is varied at a fixed microwave frequency (Fig. 13.1). A paramagnetic species will resonate at a given magnetic field, resulting in an absorption of energy (11). The absorption (relative units) of energy is plotted on the Y-axis against the magnetic field (in gauss) on the X-axis. Most often, the first derivative of this absorption spectrum is reported, giving rise to the EPR spectrum that is typically reported (1). Each inflection in the absorption can be assigned a g value, with three g values for a given signal (gx , gy , and gz ). To assign a g value, the magnetic field (B, in gauss) and microwave frequency (ν, in GHz) are used in equation [1] (12). The g value for a given absorption is constant across multiple microwave frequency levels (e.g., X, Q, W bands), and thus the g value has universality across different types of EPR instruments. g=

714.48 × v B

[1]

Fig. 13.1. The Zeeman interaction. Resonance would be obtained when the energy of the magnetic field (increasing to the right) matches the energy difference (between the ms = −1/2 and ms = +1/2 states) at a fixed microwave frequency (arrow), resulting in the absorption of energy equal to gβB, where β is Bohr magneton and B is the magnetic field that satisfies the resonance conditions.

Electron Paramagnetic Resonance Spectroscopy

193

Fig. 13.2. Simulated derivative EPR spectra. (1) Isotropic EPR signal when gx = gy = gz (2), axial signal when gy = gz = gx (3), axial signal when gx = gy = gz , and (4) rhombic EPR signal when gx = gy = gz .

The interactions between the unpaired electrons in an atom and the other electrons in the same atom result in splitting of the signal into three separate transitions (gx , gy , and gz ). In the case where all g values have identical energy, the derivative EPR spectrum is called isotropic (Fig. 13.2, trace 1). If two of the three g values are identical then the resulting signal is axial (Fig. 13.2, traces 2 and 3). When all three g values are different, the resulting signal is called rhombic (Fig. 13.2, trace 4) (11). The EPR spectrum is also affected by hyperfine coupling, which is the coupling of the unpaired electron with nearby nuclei (1). A classic X-band EPR spectrometer (Fig. 13.3) consists of a sample cavity (capable of achieving low temperatures ∼4 K) located between two large magnets, accompanied by a microwave generator and a frequency reader (12). For metalloproteins such as nitrogenase, the fast relaxation time of the unpaired electrons requires low temperatures (2–20 K) (1). These low temperatures are achieved by flowing liquid He across the sample held in the cavity.

194

Danyal, Yang, and Seefeldt

Fig. 13.3. Schematic of an EPR instrument.

2. Materials Quartz EPR tubes are used to hold samples for EPR studies. Commonly used tubes are 250 mm long with an internal and external diameter of 3 and 4 mm and can be purchased from a number of sources (e.g., Wilmad Glass, Vineland, NJ). Since nitrogenase proteins are sensitive to oxygen, the EPR tubes must be flushed with O2 -free nitrogen or argon before the proteins are added. This can be accomplished by attaching a rubber septum to the top of the EPR tube and then either flushing the tube with argon through a needle inserted into the septum with a second needle attached for venting or evacuating the sealed tube through a needle attached to a vacuum pump. The evacuated tube is then refilled with argon prior to adding proteins. Alternatively, the EPR tubes can be filled inside an argon filled glove box, sealed with a septum, and removed from the box. The solution containing nitrogenase is added to the O2 -free tube with a gas-tight syringe by inserting through the rubber stopper. The sample is shaken to the bottom of the tube and then frozen. One method for freezing samples in EPR tubes is to slowly submerge the tube into a bath of liquid nitrogen. It is vital that the tube be immersed slowly to allow for expansion of the liquid as it is

Electron Paramagnetic Resonance Spectroscopy

195

freezing. Freezing by this method too rapidly can result in cracking of the EPR tube. An alternative method for freezing EPR tubes is to rapidly submerge them into an organic solvent (e.g., hexane) that has been frozen into a slurry with liquid nitrogen. Using this method, the samples can be immersed rapidly into the cryogen without risk of cracking the tubes. Once frozen, the EPR tubes are kept in a liquid nitrogen Dewar until the spectrum is taken.

3. Methods 3.1. Preparing EPR Samples of Nitrogenase

The simplest samples of nitrogenase proteins that are prepared for EPR analysis are of the individual component proteins (Fe protein and MoFe protein). Given the oxygen sensitivity of both proteins, the samples are made in buffers that are purged with O2 -free argon or nitrogen (see Notes 1–3). In addition, the samples usually contain 2–20 mM sodium dithionite (Na2 S2 O4 ) as a reductant and O2 scavenger (see Note 4). Fe protein concentrations typically used for EPR samples are between 30 and 50 mg/mL (180 and 300 μM), with a final volume of 400 μL. At the lower protein concentrations, a greater number of EPR scans are required to obtain acceptable signal to noise. The upper protein concentrations give the cleanest EPR spectra, but can become difficult to work with because of increased viscosity. For MoFe proteins, the typical protein concentration for EPR samples is 7–10 mg/mL (40–60 μM). The conditions for acquisition of the EPR spectrum of the Fe protein in the X-band are typically 9.44–9.51 GHz microwave frequency and 0.5–20 mW microwave power. Spectra are acquired at temperatures between 4 and 12 K and at power modulations of 0.8–1.26 mT (8.4–12.6 G). It is important to determine the optimal microwave power for a given sample to avoid signal saturation. In order to determine microwave power saturation, signals are obtained over a large range of microwave power (P) at a fixed temperature. The minimal modulation amplitude that is compatible with the low-power signal intensity should be chosen (8.4 G in the case of Fe protein). For each recorded spectrum, the measured signal height (S) is then divided by the square root of the microwave power. The log( √S ) is plotted P √ against the log( P). Those powers that result in a straight line are non-saturating, while those that result in a bent line are saturating (10). The conditions for acquisition of the EPR spectrum of the MoFe protein in the X-band are typically 9.44–9.51 GHz

196

Danyal, Yang, and Seefeldt

microwave frequency and 0.5–5 mW microwave power. Spectra are acquired at temperatures between 4 and 12 K. The power modulation is usually 8.4–12.6 G (0.84–1.26 mT). The optimal microwave power is determined as described above to avoid saturation of the signal. 3.2. EPR Spectra of Nitrogenase Proteins

All three metal clusters ([4Fe–4S], P-cluster, and FeMo-cofactor) contained in the two nitrogenase component proteins can be detected by low-temperature EPR spectroscopy if the clusters are poised in a paramagnetic state (3, 5–9, 13).

3.2.1. [4Fe–4S] Cluster

The Fe protein [4Fe–4S] cluster is paramagnetic in the as-isolated state in the presence of dithionite and is in the 1+ oxidation state with 3Fe2+ and 1Fe3+ (5, 14, 15). In this oxidation state, the [4Fe–4S] cluster exhibits a mixed spin state with an S = 1/2 and S = 3/2 (Fig. 13.4). This results in an EPR spectrum with features centered in the g = 2 region and g = 4–3 region. The g = 2 (S = 1/2) region signal is rhombic in line shape, with g values of 2.04, 1.94, and 1.89 while the S = 3/2 signal has inflections

Fig. 13.4. Derivative EPR spectra of Fe protein and MoFe protein. Trace 1 shows the EPR spectrum of Fe protein (70 μM) in the as-isolated state. Spectral conditions were 9.44 GHz microwave frequency, 1.26 modulation amplitude, and 15 mW microwave power at 12 K. Trace 2 shows the Fe protein in the presence of MgATP (7.5 mM) under the same spectra conditions as trace 1. Trace 3 shows the EPR spectrum of the MoFe protein (50 μM) in its as-isolated state. Spectral conditions were 9.65 GHz microwave frequency, 1.26 modulation amplitude, and 1.0 mW microwave power at 4.8 K.

Electron Paramagnetic Resonance Spectroscopy

197

around g ≈ 5 (5). The proportions of the two spin states can be changed by the addition of either glycerol or urea to the sample prior to freezing (5). Adding glycerol shifts the spin equilibrium in favor of S = 1/2, while adding urea shifts the spin equilibrium in favor of S = 3/2 (16). The addition of nucleotides (MgATP or MgADP) changes the line shape of the [4Fe–4S]1+ EPR spectrum. Adding MgADP results in subtle changes to the line shape, whereas adding MgATP results in a shift in the line shape from rhombic to largely axial (Fig. 13.4) (3, 6). The oxidized [4Fe–4S]2+ cluster is diamagnetic and therefore is silent in the EPR (17). The 1+ oxidation state can be reduced to an all ferrous state (0 oxidation state) by treating the Fe protein with low-potential electron mediators like Ti-citrate or reduced flavodoxin (18, 19). The [4Fe–4S]0 state has 4Fe2+ , and a spin S = 4 state that can be observed in the parallel mode of an EPR spectrometer. The parallel mode spectrum shows a single inflection centered at g = 16.4 (18, 19). 3.2.2. M- and P-Clusters

The MoFe protein has two metal clusters: the [8Fe–7S] P-cluster and the FeMo-cofactor (M-cluster) (20, 21). Both clusters can have EPR signals in the parallel mode or the perpendicular mode EPR (3, 7–9). In the as-isolated state, the M-cluster (MN ) is paramagnetic with S = 3/2 EPR signal (Fig. 13.4, trace 3) that is rhombic in line shape with g = 4.32, 3.64, and 2.00 (3). In some MoFe protein variants with amino acid substitutions near FeMo-cofactor, the inflections at g = 4.32 and 3.64 are shifted, indicating changes in the electronic properties of FeMo-cofactor (22–25). In the resting state, the P-cluster (PN ) is diamagnetic and EPR silent (9). When the MoFe protein is oxidized by treatment with dye mediators, the P-cluster and M-cluster can be oxidized (7–9). The oxidized M-cluster (Mox ) is diamagnetic and is EPR silent (26, 27). The P-cluster can be oxidized from the resting state to three additional states called P1+ , P2+ , and P3+ (8, 9, 26–28). The P+1 state is a mixed spin system of S = 1/2 and S = 5/2, with a rhombic signal with g = 2.06, 1.95, and 1.82 and minor inflections at g = 6.67 and 5.3 (9, 26). The P2+ oxidation state does not have a signal in perpendicular mode EPR. It has a parallel mode signal with g = 11.8 (8). This is due to a non-Kramer system integer spin S ≥ 3 (8). The P+3 oxidation state has a mixed spin S = 7/2 and S = 1/2 state with a signal with g = 2.0 and signals at g = 10.4 and 7.2 (7). The P-cluster in its as-isolated form is all ferrous and hence is thought not to be further reduced (27, 28). The FeMo-cofactor, on the other hand, can be reduced beyond the resting state when the Fe protein binds in the presence of ATP. The MN cluster is reduced to the MR state, which is an integer spin S ≥ 1 EPR

198

Danyal, Yang, and Seefeldt

silent, diamagnetic state (29). As the electron transfer from the Fe protein to the MoFe protein is not synchronized, it is likely that the MR state contains more than one oxidation state of the FeMo-cofactor (29). The P-cluster remains silent under turnover conditions, consistent with it remaining in the PN , EPR silent state (13). 3.3. Trapping Inhibitors and Substrates on Nitrogenase During Turnover

The EPR spectrum of the M-cluster in the resting state of the MoFe protein does not show any changes upon addition of substrates or inhibitors, indicating lack of changes to the redox state or electronic properties of the M-cluster or P-cluster which has been interpreted as a lack of binding of these compounds to the MN oxidation state (30). Under turnover conditions, with the M-cluster reduced to the MR state, it is possible to trap a number of different substrates and inhibitors on FeMo-cofactor, yielding a substrate–FeMo-cofactor complex that is EPR active (31, 32). Given that only paramagnetic states can be observed, these trapped states must arise from FeMo-cofactor in the MN oxidation state or even number of electron reduced states (33). Turnover conditions require both Fe protein and MoFe protein, ATP (10 mM), MgCl2 (15 mM), an ATP-regeneration system, and the substrate that is being trapped under anaerobic conditions (no O2 ). The ATP-regeneration system commonly used includes phosphocreatine (20 mM) and creatine phosphokinase (0.3 mg/mL). Following mixing of all of the components, the samples are frozen in liquid nitrogen as described below for each substrate.

3.3.1. Proton

Protons can be trapped bound to FeMo-cofactor in the wild-type MoFe protein, but the EPR signals are of low intensity (34). If the valine at position α-70 is substituted by an isoleucine, a much stronger EPR signal is observed (Fig. 13.5, trace 1) (23). The new rhombic signal arises from an S = 1/2 spin state of FeMo-cofactor, with g values of (2.14, 2.00, 1.96). This state is trapped for the MoFe protein under steady-state turnover about 20 s after initiation of the reaction. The sample is adjusted for pH (see Note 5) frozen rapidly using a liquid nitrogen frozen hexane slurry (described above). Use of 1 H and 2 H ENDOR spectroscopy reveals that the bound intermediate is associated with one or more Fe atoms and likely represents two bound hydrides (23). More recent analysis using 95 Mo has revealed that the intermediate is not bound to Mo, but rather to one or more Fe atoms (23, 35).

3.3.2. Alkynes

The non-physiological substrate propargyl alcohol (HC≡CCH2 OH) is reduced to allyl alcohol (H2 C=CH-CH2 OH) in the α-70Ala variant MoFe protein (36). When trapped

Electron Paramagnetic Resonance Spectroscopy

199

Fig. 13.5. EPR spectra of nitrogenase trapped during reduction of protons and propargyl alcohol. The top trace shows the EPR spectrum of the α-70Ile MoFe protein trapped during proton reduction. Turnover was created by mixing 100 μM MoFe protein with 50 μM Fe protein with MgATP and a MgATP-regeneration system (described in the text). Spectral conditions were 9.65 GHz microwave frequency, 1.26 modulation amplitude, and 2 mW microwave power at 8 K. The lower trace shows the α-70Ala MoFe protein trapped during turnover with propargyl alcohol. Spectral conditions were 9.64 GHz microwave frequency, 2.0 mW microwave power, and 1.26 mT modulation amplitude at 5 K.

during turnover with propargyl alcohol, a new S = 1/2 spin state spectrum is observed (Fig. 13.5, trace 2) (36). Other non-biological and alkyne substrates have also been trapped on FeMo-cofactor (e.g., CS2 and acetylene) (37, 38). 3.3.3. Dinitrogen

An intermediate derived from N2 can be trapped on the wild-type MoFe protein (39, 40). This is achieved by mixing Fe and MoFe protein (with an ATP-regeneration system) with a full atmosphere of N2 and allowing a steady state to be achieved (about 30 s) before rapidly freezing in a liquid N2 frozen slurry of hexane. The trapped state reveals a novel S = 1/2 EPR spectra with g values 2.08, 1.99, 1.97 (Fig. 13.6, trace 1) (39, 40). 14 N2 and 15 N have been used as substrates coupled with the appropriate 2 ENDOR spectroscopy to show that the N derived from N2 is bound to FeMo-cofactor in this trapped state. The presence of a second N atom has not been confirmed, but the current model is for this intermediate to represent an N2 bound end-on to one or more Fe atoms of FeMo-cofactor (39, 40).

200

Danyal, Yang, and Seefeldt

Fig. 13.6. EPR spectra of nitrogenase trapped during reduction of N2 , diazene, and hydrazine. Trace 3 shows the wild-type MoFe protein trapped during turnover with 1 atm of N2 . Spectral conditions were 9.65 GHz microwave frequency, 1.26 modulation amplitude, and 1 mW microwave power at 4.7 K. Trace 2 shows the wild-type MoFe protein trapped during turnover with diazene, while trace 1 shows the α-70Ala /α-195Gln MoFe protein trapped during turnover with diazene. Trace 4 shows the α-70Ala /α-195Gln MoFe trapped during turnover with hydrazine. In all cases, turnover conditions were achieved with 75 μM MoFe protein, 50 μM Fe protein, and MgATP including a regenerating system. Spectral conditions were 9.65 GHz microwave frequency, 1.26 modulation amplitude, and 20 mW microwave power at 8 K.

3.3.4. Diazenes

The stepwise reduction of a N2 bound to FeMo-cofactor is proposed to proceed through a partially reduced intermediate at the level of reduction of diazene (HN=NH) (41, 42). Diazene can be reduced by nitrogenase and can be trapped bound to FeMo-cofactor (43). Both the wild-type and α-195Gln MoFe protein can be trapped under turnover conditions with diazene or methyldiazene by rapidly freezing 30 s after initiating the reaction (43). In the wild-type protein, the signal is similar to the N2 wild-type bound state with g = 2.06 and 1.98 with a minor inflection at g = 4.25 (Fig. 13.6, trace 2). On the other hand, in the α-70Ala /α-195Gln variant of the MoFe protein, the diazene trapped state shows an S =1/2 spin state with g = 2.07 and 2.01 (Fig. 13.6, trace 3) (43, 44).

3.3.5. Hydrazine

Hydrazine (H2 N-NH2 ) is a substrate for nitrogenase, being reduced to two ammonia molecules (45). It is proposed that

Electron Paramagnetic Resonance Spectroscopy

201

during the reduction of N2 , an intermediate is bound to FeMo-cofactor that is at the level of reduction of hydrazine (46). It is possible to trap a hydrazine-derived intermediate on FeMocofactor (24). This is best achieved in a MoFe protein variant with the α-195His substituted by a glutamine. Trapping this variant MoFe protein during steady-state turnover with 50 mM hydrazine by rapidly freezing (see Note 6) results in a S = 1/2 spin state with EPR signals at g = 2.09, 2.01, and 1.93 (24). The intensity of the signals can be increased by also substituting the α-70Val by an alanine in the α-195Gln background (Fig. 13.6, trace 4) (24). 3.3.6. Carbon Monoxide

The inhibitor CO can also be trapped bound to FeMo-cofactor (30). Two known signals for the binding of CO to the FeMocofactor have been observed (47–49). One signal (Fig. 13.7, trace 1) is known as the hi-CO signal since it is achieved at 0.5 atm or higher partial pressure of CO, while the other signal is known as the lo-CO signal (Fig. 13.7, trace 2) since it is trapped under lower CO concentrations (0.08 atm partial pressure of CO) (48, 49). The hi-CO signal has been assigned to two CO molecules bound to FeMo-cofactor, while the lo-CO signal has been assigned to one CO molecule bound (30). The hi-CO

Fig. 13.7. EPR spectra of nitrogenase trapped during turnover with high and low CO. Trace 1 shows the Hi-CO EPR signal observed when MoFe protein (50 μM) and Fe protein (40 μM) are trapped during turnover under 0.5 atm CO. Trace 2 shows the Lo-CO EPR signal when nitrogenase is trapped under turnover with 0.08 atm CO. Analysis was conducted at 9.44 GHz microwave frequency, 1 mW microwave power, and 0.84 mT (8.4 G) modulation amplitude at 12.8 K for the Hi-CO, while the Lo-CO sample was obtained at 9.44 GHz microwave frequency, 20 mW microwave power, and 0.84 mT (8.4 G) modulation amplitude at 13 K.

202

Danyal, Yang, and Seefeldt

signal exhibits an axial EPR signal with g = 2.17, 2.06, while the lo-CO sample exhibits a rhombic EPR signal with g = 2.09, 1.97, 1.93 (49). 3.4. Summary and Prospects

EPR has been and will continue to be a critical tool in the quest to understand the mechanism of the complex enzyme nitrogenase. The use of EPR, coupled with the inclusion of new substrates and site-directed mutagenesis to change amino acids, continues to provide important insights into this complex enzyme.

4. Notes 1. Given the sensitivity of nitrogenase to oxygen, it is important that all buffers used to make EPR samples are made dioxygen free by exchanging with an inert gas like argon or dinitrogen. This can be achieved by repeated evacuation and refilling of the samples with these gases or by bubbling solutions with the exchanging gas. 2. The exchanging gas should be of high quality and scrubbed of oxygen by passage through an oxygen scrubbing system. 3. EPR tubes should be capped with rubber stoppers and evacuated and refilled with an inert gas. Alternatively, the tubes can be flushed with an inert gas using a long needle pushed to the bottom of the tube with a vent needle through the stopper. 4. All solutions should be transferred using gas-tight syringes that have been washed with argon or dinitrogen flushed solutions that contain sodium dithionite (5 mM). 5. When making buffers with D2 O instead of H2 O, the pH meter value should be adjusted to compensate for the error in the meter with D+ . The correction is pD = pH–0.4. So, for a H2 O solution with pH 7.0, an equivalent D2 O solution should read 6.6 on the pH meter. 6. Solutions containing hydrazine should have the pH adjusted after the hydrazine is added as the hydrazine shifts the pH.

Acknowledgments The authors acknowledge the long collaboration with the Brian Hoffman and Dennis Dean laboratories in advancing understanding of nitrogenase. Work in the laboratory of the

Electron Paramagnetic Resonance Spectroscopy

203

authors is supported by a generous grant from the National Institutes of Health (GM59087). The authors also acknowledge the help and guidance from Dr. Brian Bennett of the National Biomedical EPR Center for EPR simulations. References 1. Hales B (2007) Electron paramagnetic resonance (EPR) spectroscopy. In: Scott RA (ed) Applications of Physical Methods to Inorganic and Bioinorganic Chemistry, pp. 39–64. Wiley, Hoboken, NJ 2. Zumft WG, Mortenson LE, Palmer G (1974) Electron-paramagnetic resonance studies on nitrogenase. Eur J Biochem 46:525–535 3. Orme-Johnson WH, Hamilton WD, Jones TL et al (1972) Electron paramagnetic resonance of nitrogenase and nitrogenase components from Clostridium pasteurianum W5 and Azotobacter vinelandii OP. Proc Natl Acad Sci USA 69:3142–3145 4. Smith BE, Lowe DJ, Bray RC (1973) Studies by electron paramagnetic resonance on the catalytic mechanism of nitrogenase of Klebsiella pneumoniae. Biochem J 135:331–341 5. Lindahl PA, Day EP, Kent TA et al (1985) Mössbauer, EPR, and magnetization studies of the Azotobacter vinelandii Fe protein. J Biol Chem 260:11160–11173 6. Zumft WG, Palmer G, Mortenson LE (1973) Electron paramagnetic resonance studies on nitrogenase II. Interaction of adenosine 5-triphosphate with azoferredoxin. Biochim Biophys Acta 292:413–421 7. Hagen WR, Wassink H, Eady RR et al (1987) Quantitative EPR of an S = 7/2 system in thionine-oxidized MoFe proteins of nitrogenase: A redefinition of the P-cluster concept. Eur J Biochem 169:457–465 8. Pierik AJ, Wassink H, Haaker H et al (1993) Redox properties and EPR spectroscopy of the P-clusters of Azotobacter vinelandii MoFe protein. Eur J Biochem 212: 51–61 9. Tittsworth RC, Hales BJ (1993) Detection of EPR signals assigned to the 1-equivoxidized P-clusters of the nitrogenase MoFe protein from Azotobacter vinelandii. J Am Chem Soc 115:9763–9767 10. Beinert H, Orme-Johnson WH (1967) Electron spin relaxation as a probe for active centers of paramagnetic enzyme species. In: Ehrenberg A, Malmstrom BG, Vanngard T (eds) Magnetic Resonance in Biological Systems, pp. 221–248. Pergamon, New York, NY

11. Ragsdale SW (2008) Electron Paramagnetic Resonance (EPR) for the redox biochemist. In: Banerjee R (ed) Redox Biochemistry, pp. 237–247. Wiley, Hoboken, NJ 12. Weil JA, Bolton JR (2007) Electron Paramagnetic Resonance: Elementary Theory and Practical Applications. Wiley, Hoboken, NJ 13. Chan JM, Christiansen J, Dean DR et al (1999) Spectroscopic evidence for changes in the redox state of the nitrogenase P-cluster during turnover. Biochemistry 38:5779– 5785 14. Hagen WR, Dunham WR, Braaksma A et al (1985) On the prosthetic group(s) of component II from nitrogenase: EPR of the Feprotein from Azotobacter vinelandii. FEBS Lett 187:146–50 15. Hagen WR, Eady RR, Dunham WR et al (1985) A novel S = 3/2 EPR signal associated with native Fe proteins of nitrogenase. FEBS Lett 189:250–254 16. Lindahl PA, Gorelick NJ, Münck E et al (1987) EPR and Mössbauer studies of nucleotide-bound nitrogenase iron protein from Azotobacter vinelandii. J Biol Chem 262:14945–14953 17. Thorneley RNF, Ashby GA (1989) Oxidation of nitrogenase iron protein by dioxygen without inactivation could contribute to high respiration rates of Azotobacter species and facilitate nitrogen fixation in other aerobic environments. Biochem J 261:181–187 18. Angove HC, Yoo SJ, Burgess BK et al (1997) Mössbauer and EPR evidence for an allferrous Fe4 S4 cluster with S = 4 in the Fe protein of nitrogenase. J Am Chem Soc 119:8730–8731 19. Watt GD, Reddy KRN (1994) Formation of an all ferrous Fe4 S4 cluster in the iron protein component of Azotobacter vinelandii nitrogenase. J Inorg Biochem 53:281–294 20. Chan MK, Kim J, Rees DC (1993) The nitrogenase FeMo-cofactor and P-cluster pair: 2.2 Å resolution structures. Science 260:792–794 21. Kim J, Woo D, Rees DC (1993) X-ray crystal structure of the nitrogenase molybdenumiron protein from Clostridium pasteurianum at 3.0-Å resolution. Biochemistry 32: 7104–7115

204

Danyal, Yang, and Seefeldt

22. Thomann H, Bernardo M, Newton WE et al (1991) N coordination of MoFe cofactor requires His-195 of the MoFe protein α-subunit and is essential for biological nitrogen fixation. Proc Natl Acad Sci USA 88:6620–6623 23. Igarashi RY, Laryukhin M, Dos Santos PC et al (2005) Trapping H- bound to the nitrogenase FeMo-cofactor active site during H2 evolution: Characterization by ENDOR spectroscopy. J Am Chem Soc 127: 6231–6241 24. Barney BM, Laryukhin M, Igarashi RY et al (2005) Trapping a hydrazine reduction intermediate on the nitrogenase active site. Biochemistry 44:8030–8037 25. Maskos Z, Fisher K, Sørlie M et al (2005) Variant MoFe proteins of Azotobacter vinelandii: Effects of carbon monoxide on electron paramagnetic resonance spectra generated during enzyme turnover. J Biol Inorg Chem 10:394–406 26. Lindahl PA, Papaefthymiou V, OrmeJohnson WH et al (1988) Mössbauer studies of solid thionin-oxidized MoFe protein of nitrogenase. J Biol Chem 263: 19412–19418 27. Zimmermann R, Münck E, Brill WJ et al (1978) Nitrogenase X: Mössbauer and EPR studies on reversibly oxidized MoFe protein from Azotobacter vinelandii OP: Nature of the iron centers. Biochim Biophys Acta 537:185–207 28. Surerus KK, Hendrich MP, Christie PD et al (1992) Mossbauer and integer-spin EPR of the oxidized P-clusters of nitrogenase: Pox is a non-Kramers system with a nearly degenerate ground doublet. J Am Chem Soc 114:8579–8590 29. Huynh BH, Henzl MT, Christner JA et al (1980) Nitrogenase XII: Mössbauer studies of the MoFe protein from Clostridium pasteurianum W5. Biochim Biophys Acta 623:124–138 30. Pollock CR, Lee HI, Cameron LM et al (1995) Investigation of CO bound to inhibited forms of nitrogenase MoFe protein by 13 C ENDOR. J Am Chem Soc 117: 8686–8687 31. Barney BM, Lee HI, Dos Santos PC et al (2006) Breaking the N2 triple bond: Insights into the nitrogenase mechanism. Dalton Trans 2277–2284 32. Hoffman BM, Dean DR, Seefeldt LC (2009) Climbing nitrogenase: Toward a mechanism of enzymatic nitrogen fixation. Acc Chem Res 42:609–619 33. Thorneley RNF, Lowe DJ (1985) Kinetics and mechanisms of the nitrogenase enzyme

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

system. In: Spiro TG (ed) Molybdenum Enzymes, pp. 221–284. Wiley, New York, NY Lowe DJ, Fisher K, Thorneley RNF (1993) Klebsiella pneumoniae nitrogenase: Presteady state absorbance changes show that redox changes occur in the MoFe protein that depend on substrate and component protein ratio: A role for P-centres in reducing nitrogen? Biochem J 292:93–98 Lukoyanov D, Yang ZY, Dean DR et al (2010) Is Mo Involved in Hydride Binding by the Four-Electron Reduced (E4) intermediate of the Nitrogenase MoFe Protein? J Am Chem Soc 132:2526–2527 Benton PMC, Laryukhin M, Mayer SM et al (2003) Localization of a substrate binding site on FeMo-cofactor in nitrogenase: Trapping propargyl alcohol with an α70-substituted MoFe protein. Biochemistry 42:9102–9109 Lee HI, Sørlie M, Christiansen J et al (2000) Characterization of an intermediate in the reduction of acetylene by nitrogenase α-Gln195 MoFe protein by Q-band EPR and 13 C, 1 H ENDOR. J Am Chem Soc 122:5582–5587 Ryle MJ, Lee HI, Seefeldt LC et al (2000) Nitrogenase reduction of carbon disulfide: Freeze-quench EPR and ENDOR evidence for three sequential intermediates with cluster-bound carbon moieties. Biochemistry 39:1114–1119 Barney BM, Yang TC, Igarashi RY et al (2005) Intermediates trapped during nitrogenase reduction of N2 , CH3 -N=NH, and H2 N-NH2 . J Am Chem Soc 127: 14960–14961 Barney BM, Lukoyanov D, Igarashi RY et al (2009) Trapping an intermediate of dinitrogen (N2 ) reduction on nitrogenase. Biochemistry 48:9094–9102 Dance I (1996) Theoretical investigations of the mechanism of biological nitrogen fixation at the FeMo cluster site. J Biol Inorg Chem 1:581–586 Deng H, Hoffmann R (1993) How N2 might be activated by the FeMo-cofactor in nitrogenase. Angew Chem Int Edn 32: 1062–1065 Barney BM, McClead J, Lukoyanov D et al (2007) Diazene (HN=NH) is a substrate for nitrogenase: Insights into the pathway of N2 reduction. Biochemistry 46: 6784–6794 Barney BM, Lukoyanov D, Yang TC et al (2006) A methyldiazene (HN=NCH3 ) derived species bound to the nitrogenase active site FeMo-cofactor: Implica-

Electron Paramagnetic Resonance Spectroscopy tions for mechanism. Proc Natl Acad Sci USA 103:17113–17118 45. Davis LC (1980) Hydrazine as a substrate and inhibitor of Azotobacter vinelandii nitrogenase. Arch Biochem Biophys 204:270–276 46. Thorneley RNF, Eady RR, Lowe DJ (1978) Biological nitrogen fixation by way of an enzyme-bound dinitrogen-hydride intermediate. Nature 272:557–558 47. Yates MG, Lowe DJ (1976) Nitrogenase of Azotobacter chroococcum: A new electron paramagnetic resonance signal associ-

205

ated with a transient species of the MoFe protein during catalysis. FEBS Lett 72:121–126 48. Lowe DJ, Eady RR, Thorneley RNF (1978) Electron-paramagnetic resonance studies on nitrogenase of Klebsiella pneumoniae. Biochem J 173:277–290 49. Davis LC, Henzl MT, Burris RH et al (1979) Iron-sulfur clusters in the molybdenum-iron protein component of nitrogenase. Electron paramagnetic resonance of the carbon monoxide inhibited state. Biochemistry 18:4860–4869

Chapter 14 Magnetic Circular Dichroism Spectroscopy Brian J. Hales Abstract Being able to probe the structure and energy levels of metal ions in biological systems is an important goal of bioinorganic scientists. Several of the techniques used rely on the paramagnetic property of certain oxidation states of metal ions. MCD spectroscopy is one of those techniques and represents an effective way of obtaining structure/electronic information of paramagnetic metal ions. The basics of this technique are discussed along with examples of how MCD spectroscopy has been successfully used to elucidate the metal clusters of Nif proteins from nitrogen-fixing bacteria. Key words: MCD spectroscopy, EPR spectroscopy, spin–orbit coupling, integer spin states, paramagnetism, zero-field splitting, magnetization curves, FeMoco, P-cluster.

1. Introduction 1.1. Basic Concepts

The area of bioinorganic chemistry focuses on the role of metals in biological systems. Because of the unique spectroscopic properties of most metals, bioinorganic research tends to be highly technique oriented. In addition to structural techniques such as NMR spectroscopy, X-ray diffraction, and XAS/EXAFS spectroscopy, there are also techniques that utilize the paramagnetism of certain metal oxidation states. One of those techniques, magnetic circular dichroism (MCD) spectroscopy, is the focus of this chapter. When paired with EPR spectroscopy, MCD spectroscopy can yield important information regarding spectroscopic g-factors, zero-field splitting, rhombicity parameters (D and E/D, respectively, for S > 1/2 systems), cluster type (e.g., FeS clusters), and the characterization of electronic transitions for both half-integer and integer spin systems.

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_14, © Springer Science+Business Media, LLC 2011

207

208

Hales

Several excellent treatises (1–3) have been written on the theory of the interaction of a magnetic field with matter. Interested readers should consult these for a more detailed approach than will be presented here. To understand the origin of an MCD spectrum it is important to first understand the various ways a magnetic field can influence energy levels and electronic transitions. MCD spectroscopy is simply the detection of the circular dichroism (CD) spectrum induced by an external magnetic field. CD spectroscopy measures the differential absorption of left versus right circularly polarized light in regions where the material absorbs. Many metal centers possess natural optical activity, sometimes associated with specific chiral isomers or low symmetry environments. An external magnetic field can produce additional CD inflections by perturbing the electronic energy levels through a Zeeman interaction between the field and local magnetic moments. The resultant MCD dispersion can be categorized into three types of interactions expressed as A-, B-, and C-terms. The most basic theory of these interactions, called “rigid Shift,” assumes that transitions can be approximated using Born–Oppenheimer and Franck–Condon theories and that the magnetic field does not perturb the band shape. Using these assumptions, the MCD dispersion for the transition A → J can be written as A(A → J) = γ −A1

∂f ∂E

C0 + B0 + kT

f βHbl

[1]

In equation [1] γ is a collection of spectroscopic constants, k is the Boltzmann constant, and f is a normalized line shape function. Other terms are the transition energy (E), the Bohr magneton (β), the magnetic field strength (H), the sample concentration (b), and the path length (l). The three transition parameters (A1 , B0, and C0 ) depend on the electric dipole selection rules for circularly polarized light and represent the A-, B-, and C-terms. For oriented or isotropic molecules with light propagated parallel to the external magnetic field along the z-direction, these terms become 2 2 ˆ − Jλ − Aα m ˆ + Jλ A1 = Aα m [2] α,λ ˆ ˆ ˆ ˆ × Jλ Lz + 2Sz Jλ − Aα Lz + 2Sz Aα ⎧ ⎨ ˆ − Jλ KK m ˆ + Aα B0 = − d2A Re Aα m ⎩ α,λ KK =J

Jλ Lˆ z +2Sˆ z KK ˆ + Jλ KK m ˆ − Aα − Aα m EK −EJ + d1A

Magnetic Circular Dichroism Spectroscopy

+

KK =A

C0 = −

209

ˆ − Jλ Jλ m ˆ + KK − Aa m ˆ + Jλ Aα m

ˆ − KK Jλ m

KK Lˆ z +2Sˆ z Aα

[3]

EK −EA

1 2 2 ˆ ˆ − Jλ − Aα m ˆ + Jλ Aα m Aα Lz + 2Sˆ z Aα dA α,λ

[4]

In these equations Aα and Jλ are the ground and excited states, respectively, with components α and λ. KK represents any state that can mix with the ground or excited states in the presence of a magnetic field, and dA is the degeneracy of the ground state, Aα (see Note 1). Equations [2], [3], and [4] help us understand the origin of the A-, B-, and C-terms. In these equations, Lˆ z and Sˆ z represent the z-component of the orbitaland spin angular oper momentum ˆ ators. The terms Jα Lz + 2Sˆ z Jα and Aα Lˆ z + 2Sˆ z Aα represent nonzero matrix elements for degenerate excited and ground states, respectively (i.e., states with nonzero angular or spin magnetic dipole moments). Systems with only degenerate excited states have C0 = 0 while states with degenerate grounds states have nonzero A1 and C0 where the C0 term usually dominates. Finally, states with nondegenerate ground and excited states will only have B-term contributions. Zeeman (i.e., magnetic field) interaction splits degenerate states. Figure 14.1 shows an energy level diagram depicting a transition from a nondegenerate ground state (1 S) to an orbital degenerate excited state (1 P). In this figure the excited state has been split into its ML = +1, 0, and –1 levels by an external magnetic field. This situation represents a system where C0 = 0 and A1 = 0 and symbolizes a pure A-term contribution. Transitions to levels where ML > 0 are left circularly polarized while those to ML < 0 levels are right polarized. The overlap of both transitions results in a derivative-shaped spectrum. Transitions to the ML = 0 level are not circularly polarized. Systems (Fig. 14.1) with degenerate ground states (e.g., 1 P) and nondegenerate excited states (e.g., 1 S) represent C-terms (in this case, C0 = A1 ). Here the ground state is split into ML (and/or MS ) levels by the external field. Since the degenerate levels are split by the Zeeman energy E = βH (Lz + 2Sz ), which is much less than kT at most temperatures, all of the ground levels will be significantly, but differently, populated. The intensity of a transition from a given level is proportional to the relative population of that level (i.e., the Boltzmann term). Therefore, the

210

Hales

Fig. 14.1. Energy diagram depicting the transitions associated with the A- and C-terms. The A-term involves the transitions between nondegenerate ground states (1 S) and degenerate excited states (1 P in this example). In this case an external magnetic field splits the degeneracy of the excited state allowing the differential absorption of right circularly polarized (RCP) and left circularly polarized (LCP) light and yielding a derivativeshaped net absorption. The C-term reverses this order with the degenerate state as the ground state. In this situation the intensity of the lowest energy transition increases with decreasing temperature.

spectrum is dominated by the transition from lowest (i.e., most populated) M level and increases in intensity with decreasing temperature. This temperature dependency distinguishes the C-term from the other two terms in equation [1] and is a characteristic of paramagnetic systems. When neither the ground state nor the excited state is degenerate (i.e., A1 = C0 = 0), the MCD spectrum rises solely from the B-term in equation [3]. As implied by equation [2], the B-term is a second-order term produced by state mixing of the ground (Aα ) and/or excited (Jλ ) state with other close lying states (Kκ ). All systems will possess B-terms. The relative contribution of each of the three terms to the final intensity of the MCD spectrum is typically C > A > B. Because the C-term has an inverse temperature dependency, its contribution is the greatest and can be one- to two orders of magnitude larger than the other two terms at liquid He temperatures. This means that the C-term will dominate the MCD spectroscopy at low temperatures and is a major reason why this technique is helpful in the study of the ground state of a paramagnetic system. 1.2. Magnetization Curves

An example of the type of information gained with MCD spectroscopy can be seen by considering a simply isotropic S = 1/2 spin

Magnetic Circular Dichroism Spectroscopy

211

system. In the presence of an external magnetic field the energies of the ms + 1/2 and – 1/2 levels are split by E = gβH . At a constant temperature, the population of the upper state (Nα ) relative to that of the lower spin level (Nβ ) will be Nα = Nβ e −

gβH

/ kT

[5]

Since the relative intensity I (0 ≤ I = ε/K ≤ 1) of an MCD signal (C-term only) will be proportional to the relative population of the lower spin level to the upper level, it can be expressed as Nβ − Nα ε = K Nβ + Nα

[6]

Using equation [5], equation [6] becomes gβH gβH ε 1 − e − / kT e + / kT − 1 gβH = = gβH = tanh gβH K 2kT 1 + e − / kT e + / kT + 1

[7]

As H → ∞ and/or T → 0, the ground level becomes the dominant populated state and equation [7] goes to unity. This situation is called “saturation,” and a plot of equation [7] (called a “magnetization curve”) represents the fraction of magnetization of a spectrum. Because equation [7] contains the spectroscopic g-factor, the shape of the magnetization curve reflects the spin state of the system. This characteristic is true even for anisotropic state with S > 1/2. For a high spin (S > 1/2) metal system in an external magnetic field, there is an interaction among the electrons that occurs via spin–orbit coupling and is expressed by the Hamiltonian operator E ˆ 2 ˆ 2 2 ˆ ˆ ˆ H = go βH • S + D Sz − S(S + 1)/3 + S − Sy D x

[8]

The first term is the typical Zeeman interaction term involving ˆ and the external dot product of the total spin vector operator (S) magnetic field vector (H). The second term involves spin interaction where D and E are the zero-field axial and rhombic splitting, respectively. The ratio, E/D, is termed the rhombicity and has a range between 0 and 1/3 where 0 represents a purely axial system and E/D = 0 represents various degrees of rhombicity. In the absence of a magnetic field and zero rhombicity, the Sˆ z2 operator in equation [8] splits the spin state into |ms | doublets. For half-integer spin states these levels remain degenerate regardless of the rhombicity and are called Kramer’s doublets. The majority of metalloproteins have D greater than the microwave energy (gβH) at X-band (9.5 GHz) frequencies. Because of this,

212

Hales

the spectrometer’s microwaves do not have enough energy to induce transitions between different doublet levels and transitions are only observed within each isolated |ms | doublet. Therefore, each |ms | doublet can be viewed as an isolated S = 1/2 spin state with a unique set of g-factors that depend on the spin state and rhombicity. The g-factors of each doublet can be predicted using the “Rhombo” program designed by Hagen (4). In general, halfinteger spin states will have some inflections observable by EPR spectroscopy. Integer spin (non-Kramer’s) states present a different situation. In addition to |ms | doublets (when E = 0 and H = 0) there will also be an ms = 0 singlet. When rhombicity occurs, the doublet levels split into singles, even at zero magnetic field. As is the case with D, the magnitude of E is usually larger than the energy of the spectrometer’s microwaves, preventing any transitions and resulting in no EPR spectrum at X-band frequencies, i.e., integer spin systems are often “EPR silent.” Even though they may be “EPR silent,” integer spin system will still yield an MCD spectrum. This makes MCD spectroscopy more versatile than EPR spectroscopy (which often cannot detect integer spin systems) in the study of paramagnetic metal systems. However, it should be noted that because EPR inflections are not observed in an “EPR silent” system, neither D nor E is known or predictable and the absolute spin state cannot be determined using MCD spectroscopy (see Note 2). 1.3. Examples of Nif Proteins

There are numerous examples of the use of MCD spectroscopy in the characterization of the metal clusters of Nif proteins.

1.3.1. S = 1/2 States

To date, none of the metal clusters in the various Nif proteins exhibit a pure S = 1/2 spin state. All have mixed spin states. In fact, the nitrogenase Fe protein (NifH) was the first protein containing a [4Fe4S]+ cluster shown (5) to exhibit a mixed spin state (S = 1/2, 3/2). Since that early observation, mixed spin states have been detected in numerous metalloproteins. The relative amount of each spin state in NifH depends on the solvent composition with 50% ethylene glycol favoring the S = 1/2 state (∼90%) and 0.5 M urea favoring the S = 3/2 state (∼85%). In aqueous solution, the S = 1/2 to 3/2 spin state ratio is approximately 2:3. Using this solvent dependency, a study (6) was undertaken to investigate the influence of spin state on an MCD spectrum. A general outcome of that study is that a change in the spin state does not greatly perturb the energies of the electronic transitions. On the other hand, an increase in spin state induces a positive shift of the spectrum’s baseline.

1.3.2. S > 1/2 Half-Integer States

One of the most complex metal clusters studied by MCD spectroscopy is the FeMo cofactor (FeMoco) in the nitrogenase MoFe

Magnetic Circular Dichroism Spectroscopy

213

140 120

Δ ε (M–1 cm–1)

100 80 60 40 20 0 –20 400

500 600 Wavelength (nm)

700

800

Fig. 14.2. Low-temperature MCD spectrum of the FeMo cofactor (FeMoco, an Mo7Fe9SX-homocitrate cluster with an unknown center atom X) of the dithionitereduced state of the MoFe protein of nitrogenase. This cluster has an S = 3/2 spin state with D > 0.

protein. Initial EPR studies showed (7) that the cluster in the reduced protein exhibited an EPR spectrum with inflection g [4.3, 3.6, 2.0] associated with the ground ms = ±1/2 doublet of an S = 3/2 spin system with zero-field splitting D ∼6 cm–1 . The MCD spectrum (Fig. 14.2) of this cluster showed (8) inflections not previously seen in the spectra of classic FeS clusters and implying (prior to the determination of the MoFe protein structure (9) by X-ray diffraction) that FeMoco had a structure different from that of other known FeS clusters. Magnetization curves constructed from these spectra at different temperatures exhibit “nesting,” which is a flaring of the different temperaturedependent magnetization curves. Nesting occurs when a lowlying, paramagnetic excited level (with different magnetization parameters) becomes populated at high temperatures. Nesting is not observed in pure S = 1/2 states where there is only one ground-state doublet. Rather, it occurs when zero-field interactions split the levels of S > 1/2 states into pairs of |ms | components. For FeMoco in the MoFe protein, the low-lying level is the ms = ±1/2 doublet while the excited level is the ms = ±3/2 doublet, as predicted from the EPR spectrum. Simulation of the magnetization curves using the spectral parameters from the EPR spectrum illustrates the complimentary of the two techniques. 1.3.3. Integer States

The P-cluster of the nitrogenase MoFe protein provides an excellent example of an FeS cluster with integer spin states. The P-cluster is an [8Fe8S] cluster with three to four accessible oxidation states (10, 11). The lowest oxidation state (abbreviated as Pn or P0 ) is diamagnetic (S = 0) with all of eight Fe ions in their ferrous states (12). The MCD spectrum (13) of P0 (Fig. 14.3)

Hales 50

Δε (M–1 cm–1)

40

30

20

10

1.63 K 9.30 K 0 400

500 600 Wavelength (nm)

700

800

Fig. 14.3. Low-temperature MCD spectrum of the P-cluster (8Fe7S) of the dithionitereduced state of the apo-MoFe protein of nitrogenase. This cluster has an S = 0 spin state so that the spectrum arises from a B-term transition, meaning that its spectrum is temperature independent.

gives a good example of the spectrum of a diamagnetic system with a B-term in the MCD dispersion equation (equation [1]) meaning that the spectrum is temperature independent. It should also be noted that the intensity of this spectrum is much smaller

40

1.58 K 4.24 K

30

Δε (M–1 cm–1)

214

9.44 K 20

10

0 400

500 600 Wavelength (nm)

700

800

Fig. 14.4. Low-temperature MCD spectrum of the P-cluster (8Fe7S) of the thionineoxidized state of the apo-MoFe protein of nitrogenase. This cluster has an S = 3 or 4 spin state so that the spectrum arises from a C-term transition, whose intensity is temperature dependent.

Magnetic Circular Dichroism Spectroscopy

215

30

25

Δ ε ( M–1 cm–1)

4.24 K

9.44 K

1.58 K

20

15

10

5

0 0.0

0.2

0.4

0.6

0.8

1.0

1.2

β B /2 k T Fig. 14.5. Magnetization curves of the P-cluster (8Fe7S) of the thionine-oxidized state (S = 3 or 4) of the apo-MoFe protein of nitrogenase. The clustering of curves (i.e., nesting) at different temperatures arises from the thermal population of low-lying excited states of the spin system, each with different spectral parameters, as the temperature is raised.

than that of a typical paramagnetic sample (i.e., B-term << Cterm). Oxidizing the P-cluster (14) by two equivalents (P0 → P2+ ) generates an integer, paramagnetic state (S most likely 3 or 4). The MCD spectrum of this state (Fig. 14.4) is clearly that of a paramagnetic cluster with a high spin. This can be seen in the magnetization curve (Fig. 14.5) where the initial slope at low temperature (1.6 K) is very steep. Nesting can be clearly seen in these curves. Because of the uncertainty in the magnitudes of the zero-field parameters, D and E, the spin state cannot be definitively determined.

2. Materials 2.1. Sample

1. The majority of biological samples investigated using MCD spectroscopy are in a buffered aqueous solution. However, regardless of the solvent used a necessary criterion is that the sample forms an optical glass when frozen (see Note 3). This is because MCD spectroscopy, being an optical technique, requires the transmission of light through the sample for detection. For aqueous samples this is usually accomplished

216

Hales

by the addition of a cryoprotectant, typically glycerol or ethylene glycol (50% v/v). 2. The requirement of transmissibility applies to both the solvent and the solute. To obtain the most intense spectrum, the highest sample concentration should be used so long as the absorbance is below 1.0 in the wavelength range of investigation. Small additives such as dithionite, used for reduction, or organic dyes, used for oxidation/reduction control and mediation, also contribute to the total absorbance and should be used either at low concentrations or diluted by gel filtration. 2.2. Sample Cuvette

1. The typical sample cuvette has a path length of 1 mm, although a 2 mm path length can be used for samples of low concentration. Cuvettes with a path length longer than 2 mm have a difficulty in allowing the solution to form a glass upon freezing and have a chance of cracking during freezing and/or thawing. 2. The cuvette can be home-built using thin quartz plates cut to dimension and spaced with a rubber O-ring. A more convenient method is to use a commercial 1 mm optical cuvette and cut it to the appropriate size for easy insertion into the MCD sample holder.

3. Methods 1. To record an MCD spectrum, you first need a CD spectrometer. There are several excellent commercial spectrometers available. Be sure to choose one that has a detachable photomultiplier (to position on the other side of the magnet) and a wavelength range that goes up to at least 800 nm. 2. Place the CD spectrometer on a long table. Position a superconducting magnet next to the spectrometer on the detector side. To avoid magnetic field interference, the photomultiplier should be placed at least 4 ft from the magnet and connected to the magnet with a light tight tube. Like the spectrometer, there are many excellent super-conducting magnets available. The magnet is for optical detection, so it must have a sample chamber in the coil of the magnet with optical windows. The sample chamber is linked to the liquid He chamber of the magnet, which allows the sample temperature to be varied from room temperature to ∼1.6 K, with pumping. 3. The sample chamber and sample holder of the commercial magnet are usually equipped with thermocouples for temperature monitoring. The thermocouples roughly

Magnetic Circular Dichroism Spectroscopy

217

indicate the temperature of the chamber and sample. It is important to know the temperature of the sample to three significant figures. This is best accomplished by placing additional sensitive ceramic thermocouples directly above and below the sample. The average reading of the two monitors will closely approximate the temperature of the sample. 4. The initial setup of the MCD spectrometer involves cooling the magnet’s liquid He and liquid N2 chambers with liquid N2 . Following cool down and temperature equilibration (about 6 h) the liquid N2 is pumped out of the He chamber and the chamber is evacuated and flushed several times with He gas. Liquid He is then slowly transferred into the He chamber and the system is again allowed to stabilize for 6 h (see Note 4). 5. Prior to inserting the sample, the sample chamber must be partially filled with liquid He. This allows rapid freezing, necessary for good glass formation, when the sample is lowered into the chamber. The sample must be admitted into the chamber under a rapid flow of He gas to avoid the introduction of air into the sample chamber (air will freeze at liquid He temperatures and block light transmission) (see Note 5). 6. After the sample has been positioned in the sample chamber, the CD spectrum should be recorded. If an intense background CD spectrum (often caused by none ideal glass formation) is present, the sample may have to be thawed and refrozen. 7. The MCD spectrum is now ready to be recorded. Magnetization data are obtained by recording the signal intensity at an absorption maximum as the magnetic field is raised and lowered. After recording the MCD spectrum at a given temperature, the field is set back to zero, the polarity of the magnet is reversed, and the spectrum is again recorded. Since the sign of the MCD spectrum depends on the field direction while the CD spectrum is field independent, subtracting the spectra recorded in the two different field directions will yield the pure MCD spectrum without any CD background.

4. Notes 1. Equations [2], [3], and [4] contain matrix elements involving the orbital angular momentum. Therefore, in order for an MCD spectrum to be detectable, the ground state must have orbital angular momentum. This criterion is easily satisfied with metal ions, which have significant spin–orbit

218

Hales

coupling. However, this is often not the case with simple organic radicals, making them undetected in an MCD spectrometer. 2. Integer spin states are usually EPR silent. These samples are still detectable by MCD spectroscopy. However, because the spectral parameters (e.g., g-factors, axial and rhombic zerofield parameters, D and E, respectively) are unknown, the spin state cannot be determined from magnetization curves as it can be for half-integer states. 3. All samples must be able to form an optical glass when frozen. This is accomplished by adding various cryoprotectants to the solution. A sample that does not form a glass will be quickly recognized by its inability to transmit light when frozen. Always make sure that the cryoprotectant used does not adversely interact with the sample. 4. Safety precautions should always be used when transferring liquid N2 and liquid He into the magnet. The extreme low temperature of these two liquids can cause serious “burns.” 5. It is extremely important that air, and specifically oxygen, not be allowed to enter the liquid He dewar or sample chambers. At liquid He temperatures, air will form a solid and can cause serious blockages. Furthermore, oxygen is paramagnetic and solid oxygen is opaque. If solid oxygen is present in the sample chamber, the magnetic field will draw the solid into its light path, blocking the passage of light. This will be noted by a sudden decrease in transmission as the magnetic field is increased. If this happens, the sample can be slowly warmed to ∼150 K (DO NOT warm it above the freezing temperature of the sample!) and the sample chamber repeatedly evacuated and refilled with He gas to remove the oxygen. References 1. Piepho SB, Schatz PN (1983) Group Theory in Spectroscopy with Applications to Magnetic Circular Dichroism. Wiley, New York, NY 2. Thompson AJ, Cheesman MR, George SJ (1993) Variable-Temperature Magnetic Circular Dichroism. In: Riordan JF, Vallee BL (eds) Metallobiochemistry, Part C, pp. 199–231. Academic, San Diego, CA 3. Johnson MK (2000) CD and MCD Spectroscopy. In: Lawrence Que J (ed) Physical Methods in Bioinorganic Chemistry: Spectroscopy and Magnetism, pp. 233–286. University Science Books, Sausalito, CA

4. Hagen WR (1992) EPR Spectroscopy of Iron-Sulfur Proteins. In: Sykes AG, Cammack R (eds) Advances in Inorganic Chemistry: Iron-Sulfur Proteins, pp. 165–222. Academic, San Diego, CA 5. Lindahl PA, Day EP, Kent TA et al (1985) Mössbauer, EPR, and Magnetization Studies of the Azotobacter vinelandii Fe Protein. J Biol Chem 260:11160–11173 6. Onate YA, Finnegan MG, Hales BJ et al (1993) Variable Temperature Magnetic Circular Dichroism Studies of Reduced Nitrogenase Iron Proteins and [4Fe-4S]+ Synthetic Analog Clusters. Biochim Biophys Acta 1164:113–123

Magnetic Circular Dichroism Spectroscopy 7. Palmer G, Multani JS, Cretney WC et al (1972) Electron Paramagnetic Resonance Studies on Nitrogenase—I. The Properties of Molybdoferredoxin and Azoferredoxin. Arch Biochem Biophys 153:325–332 8. Johnson MK, Thomson AJ, Robinson AE et al (1981) Characterization of the Paramagnetic Centres of the Molybdenum-Iron Protein of Nitrogenase from Klebsiella pneumoniae Using Low-Temperature Magnetic Circular Dichroism Spectroscopy. Biochim Biophys Acta 671:61–70 9. Kim J, Rees DC (1992) Crystallographic Structure and Functional Implications of the Nitrogenase Molybdenum-Iron Protein from Azotobacter vinelandii. Nature 360: 553–560 10. Hagen WR, Wassink H, Eady RR et al (1987) Quantitative EPR of an S = 7/2 System in Thionine-Oxidized MoFe Proteins of Nitrogenase: A Redefinition on the P-Cluster Concept. Eur J Biochem 169:457–465

219

11. Surerus KK, Hendrich MP, Christie PD et al (1992) Mössbauer and Integer-Spin EPR of the Oxidized P-Clusters of Nitrogenase: Pox Is a Non-Kramers System with a Nearly Degenerate Ground Doublet. J Am Chem Soc 114:8579–8590 12. Münck E, Rhodes H, Orme-Johnson WH et al (1975) The MoFe Protein Component from Azotobacter vinelandii. Biochim Biophys Acta 400:32–53 13. Broach RB, Rupnik K, Hu Y et al (2004) VTVH-MCD spectroscopic Study of the Metal Clusters in the nifB and nifH MoFe Proteins of Nitrogenase from Azotobacter vinelandii. Biochemistry 45: 15039–15048 14. Lindahl PA, Papaefthymiou V, OrmeJohnson WH et al (1988) Mössbauer Studies of Solid Thionin-Oxidized MoFe Protein of Nitrogenase. J Biol Chem 263: 19412–19418

Chapter 15 Mössbauer Spectroscopy Boi Hanh Huynh Abstract Mössbauer spectroscopy has contributed significantly to the studies of Fe-containing proteins. Early applications yielded detailed electronic characterizations of hemeproteins, and thus enhanced our understanding of the chemical properties of this important class of proteins. The next stage of the applications was marked by major discoveries of several novel Fe clusters of complex structures, including the 8Fe7S P cluster and the mixed metal 1Mo7Fe M center in nitrogenase. Since early 1990 s, rapid kinetic techniques have been used to arrest enzymatic reactions for Mössbauer studies. A number of reaction intermediates were discovered and characterized, both spectroscopically and kinetically, providing unprecedented detailed molecular-level mechanistic information. This chapter gives a brief summary of the historical accounts and a concise description of some experimental and theoretical elements in Mössbauer spectroscopy that are essential for understanding Mössbauer spectra. Major biological applications are summarized at the end. Key words: Mössbauer spectroscopy, electronic information, nitrogenase, cofactors, Fe–S proteins.

1. Introduction 1.1. Overview

Mössbauer spectroscopy is a nuclear resonance technique that utilizes the recoilless nuclear γ-emission and absorption phenomena discovered by Rudolf L. Mössbauer in 1958 (1). Even though almost all elements with atomic number larger than 24 have at least one isotope that may be used for Mössbauer measurement, biological Mössbauer applications involve exclusively the isotope 57 Fe (2, 3). This is due to the fact that Mössbauer effect of the 57 Fe nucleus is relatively easy to detect and that Fe-containing proteins are abundant and perform vital biological functions. Development of biological Mössbauer applications may be described in three stages. In the early stage of the development

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_15, © Springer Science+Business Media, LLC 2011

221

222

Huynh

(1960 s), biological applications of Mössbauer spectroscopy were focused on the studies of hemeproteins (4), which contain a mononuclear Fe cofactor, heme. Through these studies, a solid foundation for extracting detailed electronic and magnetic properties of the mononuclear Fe site from its Mössbauer spectrum was laid down (5) and paved the way for later studies of proteins containing complex Fe cofactors. In the second stage of the development (1970–1990), the unique capability of Mössbauer spectroscopy to identify, quantify, and characterize multi-nuclear Fe cofactors was recognized. Major advances were made in the studies of Fe–S proteins, which contain cofactors composed of clusters of Fe and S atoms. In addition to providing electronic and magnetic information on the entire Fe–S clusters as whole entities, Mössbauer spectroscopy was shown to have the capability to (1) provide detailed electronic information on individual Fe sites within a cluster, (2) reveal the exchange coupling natures between the Fe atoms, and (3) determine the Fe nuclearity of the cluster. Furthermore, for the studies of multi-cluster containing proteins, the method has the capability to quantify and characterize the various types of Fe clusters in the molecule. This period was marked by several landmark studies, which firmly established the importance of Mössbauer spectroscopy in the study of Fe proteins. These studies include (1) characterizations of the antiferromagnetically coupled [2Fe–2S] and [4Fe–4S] clusters (6, 7), (2) the identification and characterization of the M and P cluster in the MoFe protein (8–10), (3) the discovery of the [3Fe–4S] cluster in ferredoxins (11, 12) and in aconitase (13), (4) the observation of facile interconversions between the [3Fe–4S] and [4Fe–4S] clusters, and its relevance to protein function (13, 14), and (5) the identification of a covalently linked [4Fe–4S]-heme active unit in sulfite reductases (15, 16). These studies have also stimulated a series of theoretical investigations on the exchange coupling nature of the Fe–S cofactors (17, 18). During the first two stages of development, biological Mössbauer studies generally are performed on stable states of proteins and enzymes. Information obtained is therefore static in nature. More recently (since the early 1990s), with the advances made in molecular biology, sufficient amounts of proteins and enzymes can be produced for kinetic Mössbauer investigations. By employing Mössbauer spectroscopy in conjunction with rapidmix rapid-freeze technique (19), which quenches biochemical reactions at desired reaction time points for spectroscopic investigation, a few groups of investigators have succeeded in characterizing, at a molecular level, the steps being taken during the reaction cycles of several oxygen-activating enzymes, including the R2 subunit of Escherichia coli ribonucleotide reductase (20–22), the soluble methane monooxygenase (23), frog M ferritin

Mössbauer Spectroscopy

223

(24), taurine:α-ketoglutarate dioxygenase (25, 26), myo-inositol oxygenase (27), toluene/o-xylene monooxygenase (28), and Chlamydia trachomatis ribonucleotide reductase (29). Novel reaction intermediates (29–33) as well as previously hypothesized intermediates (25, 26, 34, 35) were discovered and characterized, both structurally and kinetically, providing unprecedented insights into the reaction mechanisms of these Fe proteins. When combined with site-directed mutagenesis, the approach can also be used to reveal the functional roles of key residues (36, 37). 1.2. Applications to Nitrogenase MoFe Protein

Application of Mössbauer spectroscopy to the studies of the nitrogenase enzyme system begun in mid-1970s (38). By using an approach that combines Mössbauer, EPR, analytical, and biochemical techniques (8–10, 39, 40), Münck, Orme-Johnson, and coworkers were able to establish that the 30 Fe atoms in the MoFe protein are organized into two different types of clusters, namely the M (magnetic) and P (protein) cluster. The approach was then unprecedented, but, by now, it has become a standard methodology practiced by most biological Mössbauer spectroscopists. The presence of M and P cluster in MoFe protein was confirmed eventually by X-ray crystallographic studies (41–43) almost 15 years later. In the following a succinct description of the major Mössbauer/EPR findings is presented. The MoFe protein of the nitrogenase enzyme system is extremely sensitive to oxygen and has to be purified under strict anaerobic condition. Early spectroscopic investigations indicated that the anaerobically purified MoFe protein exhibited EPR resonances at g = 4.32, 3.65, and 2.01, arising from the ground Kramers doublet of a spin S = 3/2 state (D = 6 cm–1 , E/D = 0.055). Spin quantification of this S = 3/2 EPR signal yielded approximately 1 spin/Mo atom (8). No other EPR signals are detected. The 4.2 K Mössbauer spectrum of the as-purified MoFe protein recorded with a 36 mT magnetic field applied parallel or perpendicular to the γ radiation (Fig. 15.1) shows two types of spectral components: One spectral component (component I, indicated by solid lines) exhibits magnetic hyperfine split pattern with varying intensity depending on the orientation of the applied magnetic field. The other component (component II) showed no magnetic interaction and consisted of only quadrupole doublets. The magnetically split hyperfine pattern in weak applied field and the observed field-orientation dependence indicated that component I is arising from an EPR-active half-integer spin Fe center, while the absence of magnetic hyperfine structure for component II indicated that it is arising from a diamagnetic or an integer spin Fe center (see Section 3). Thus, the Mössbauer data indicate unambiguously that MoFe protein contains at least two types of Fe centers: an EPR-active half-integer spin and an EPR-silent integer spin Fe center.

224

Huynh

Fig. 15.1. Mössbauer spectra of as-purified MoFe protein from Azotobacter vinelandii. The experimental spectra (vertical bars) were taken from reference (8), which were recorded at 1.5 K with a 36 mT magnetic field applied parallel (a) or perpendicular (b) to the γ-beam. The solid lines are theoretical spectra of the m center simulated with the parameters of the six components given in reference (39). The 7th component was simulated by a quadrupole doublet described in reference (45). Line positions of the quadrupole doublets, Fe2+ , D, and S are indicated by brackets in (a).

Since the EPR data showed only one EPR-active species, component I must be associated with the S = 3/2 center. Because component I exhibited magnetic hyperfine pattern, the center was then termed the M center (8) (M for magnetic). In 1977, Shah and Brill isolated an iron- and molybdenum-containing complex (FeMo cofactor) from the MoFe protein (44). Spectroscopic investigations (9) of the isolated FeMo cofactor showed an S = 3/2 EPR signal and Mössbauer spectra similar to those of component I, establishing unambiguously that the M center is in fact the FeMo cofactor. Detailed analyses of the Mössbauer component I (39) indicated that the component can be decomposed into at least six sub-components of equal intensity. It was therefore proposed that the M center is a mixed metal cluster that contains 1 Mo and likely 6 Fe atoms (39). Later X-ray crystallographic

Mössbauer Spectroscopy

225

investigations (41, 42) showed that M center is indeed a mixed metal center containing 1 Mo and 7 Fe atoms. The presence of a 7th subcomponent for component I was eventually revealed by a Mössbauer study of MoFe protein with selectively 57 Fe-labeled M center (45). At a temperature higher than 30 K, the electronic relaxation of the M center becomes very fast, resulting in cancellation of the magnetic hyperfine interactions at the Fe nuclei and transforming component I into a quadrupole doublet (8). The observation of one single quadrupole doublet arising from a 7 Fe cluster indicates a valance delocalized system for M. The Mössbauer parameters, particularly the isomer shift of 0.41 mm/s, suggest an average oxidation state of about +2.4 for the Fe ions. Thus, it was proposed that the formal oxidation state of the M center may be considered as Mo4+ –4Fe2+ –3Fe3+ (45). Component II comprises three quadrupole doublets, labeled as the Fe2+ , D, and S doublet (8) (Fig. 15.1). They differ in absorption intensity and quadrupole splitting (EQ (Fe2+ ) = 3.02 mm/s, EQ (D) = 0.81 mm/s, and EQ (S) = 1.37 mm/s), but showed similar isomer shift of 0.64 mm/s (40). For Fe–S proteins, this value of isomer shift suggests strongly a high-spin ferrous state. Early Mössbauer investigations (10, 40) provided unequivocal evidence indicating that doublets Fe2+ and D are arising from a spin-coupled cluster: First, high-field Mössbauer studies demonstrated that both doublets are associated with a diamagnetic (S = 0) system, even though their parameters are indicative of high-spin S = 2 Fe2+ state (8). The least stunning explanation for these seemingly contradictive observations is that Fe2+ and D belong to an all-ferrous cluster, in which the high-spin Fe2+ ions are spin coupled to form a diamagnetic ground state. Second, oxidation titration of MoFe protein with thionine showed that the Fe2+ and D doublets converted in parallel into magnetic spectra. This observation is consistent with Fe2+ and D belonging to a same cluster. The cluster that is formed with the Fe2+ - and D-type iron atoms was then termed the P cluster (10) (P for protein). Subsequent Mössbauer investigations of MoFe protein with selectively 57 Fe-labeled P clusters showed that the S doublet is also associated with the P cluster (46). Correlating the results of metal analysis with Mössbauer absorption intensities of Fe2+ , D, and S suggested that approximately 16 Fe atoms per molecule of MoFe protein are associated with the P cluster. Thionine oxidation study indicated that it required the removal of four electrons to oxidize all the P clusters (10). Further, detailed analysis of the magnetic spectrum of the oxidized P cluster showed that it may be decomposed into eight components of equal intensity (10, 40). Taken together, these observations were initially interpreted to indicate that P cluster is a cubane-like 4Fe cluster, that each molecule of MoFe protein contains 4 P

226

Huynh

clusters and that one-electron oxidation of the diamagnetic P cluster converts the integer spin S = 0 state into a paramagnetic half-integer spin state. Integer spin EPR study, however, showed that the spin state of the oxidized P cluster is in fact integer (47). Thus, the initial interpretation was modified. The P cluster was proposed to comprise two bridged cubane-like 4Fe clusters, and each MoFe protein contains two P clusters (47). High-resolution X-ray crystallographic study revealed that the P cluster is an 8Fe–7S cluster composed of bridged [4Fe–4S] and [4Fe–3S] clusters (42).

2. Materials 2.1. Basic Mössbauer Spectrometer Setup

Conventional Mössbauer spectrometers for biological studies consist of four basic units: (1) a 57 Co γ-ray radiation source (10–100 mCi) for providing the 14.4 keV photons (see below), (2) a velocity transducer for driving the source at a constant acceleration motion, (3) a cryogenic dewar and its operation unit for controlling the sample temperature ranging from 1.5 to 220 K, and (4) a γ-ray counting system for recording the Mössbauer events as a function of the source velocity. For measurements in many nitrogenase related experiments (8, 48), as well as measurements of other Fe-containing proteins, a permanent magnet (24–60 mT) is typically used. If high-field measurements (1–10 T) are required, they can be performed with a cryogenic dewar equipped with a superconducting magnet system.

2.2. Source of Radiation

Common to any resonance spectroscopy, Mössbauer spectroscopy also requires a radiation source. For conventional Mössbauer studies of biological compounds, the source is 57 Co, a radioactive nucleus with a half-life of approximately 270 days. It captures an internal electron and decays to the 2nd excited state of 57 Fe (Fig. 15.2). Approximately 9% of the excited nuclei decay to the ground state and emit γ photons of 136 keV. The other 91% of the nuclei decay to the 1st excited state and emit γ photons of ∼122 keV. Subsequent decay of the 1st excited state to the ground state emits a γ photon of 14.4 keV. A significant fraction of these 14.4 keV γ photons will be emitted without recoil and thus will have a natural linewidth of approximately 5×10–9 eV. These recoilless 14.4 keV γ photons are the radiation source for promoting recoilless resonance transition from the ground state to the first excited state of the 57 Fe nuclei in the sample (Fig. 15.3). For a more complete description of Mössbauer effect, please refer to reference (5).

Mössbauer Spectroscopy

227

Fig. 15.2. Decay scheme of 57 Co.

Fig. 15.3. Energy level diagram of the ground and 1st excited states of 57 Fe nucleus under the influence of hyperfine interactions. Possible transitions are shown as arrows and the resulting Mössbauer spectra are depicted below the corresponding energy diagram. For illustrative purposes, the energy levels are not drawn to scale.

228

Huynh

2.3. Sample Preparation

For biological applications, the samples are generally frozen solutions. The optimal physical thickness for frozen solution samples is about 0.5 cm. The required volume obviously depends on the cross section of the sample. A well-designed sample holder should not require more than 500 μl of samples. In order to collect spectrum of good statistics within a manageable time (24–48 h), the concentration should be at least 1 mM in 57 Fe for each distinct Fe site.

3. Methods For biological studies, Mössbauer spectroscopy can be used to accomplish the following tasks (see Notes for principles of application). 3.1. Providing Detailed Description of the Electronic States of Iron Centers in Proteins

Mössbauer spectroscopy measures hyperfine interactions between the Mössbauer nucleus and its surrounding electrons. For biological applications, the nucleus is exclusively 57 Fe. In other words, the nuclear properties are constant for all measurements. Variations observed in the spectra are solely due to changes in the electronic state. Therefore, information extracted from the spectra is also pertaining only to the electronic states of the Fe ions.

3.2. Determining the Number and Electronic Properties of Fe Atoms in a Multi-iron Containing Center

In a multi-iron containing center, the electronic environment of each Fe sites may be differ due to differences in the intrinsic oxidation states and/or the effects of spin–spin coupling. Therefore, spectra originating from individual Fe sites may also be differ. Because the physical process involved in biological Mössbauer spectroscopy is the recoilless nuclear resonance between the ground and the first excited state of 57 Fe nucleus, the probability for resonance to occur is a constant irrespective of the electronic states of the Fe ions. Thus, different electronic environments may produce different Mössbauer spectra, but the absorption areas of the spectra are directly proportional to the concentrations of the Fe sites that generate the corresponding spectra. Consequently, from the relative absorption intensities and the number of different spectra observed for the center, the number of Fe contained in the center can be directly determined. The intrinsic oxidation states of the individual Fe sites can also be determined from the parameters obtained for the different spectra. Successful examples of such application are many (see above) including characterization of the MoFe protein (39) and discovery of the [3Fe–4S] cluster (11, 12).

Mössbauer Spectroscopy

229

3.3. Revealing the Organization of Fe Atoms in a Protein Containing Multiple Fe Centers

Since Mössbauer spectroscopy can be used to quantify relative concentrations of all Fe centers present in a sample and determine the nuclearity of each distinct Fe center, it is straightforward for Mössbauer spectroscopy to resolve the organization of Fe atoms in a complex protein molecule. For example, Mössbauer spectroscopy had been instrumental in establishing that the 30 Fe atoms in MoFe protein are organized into two types of Fe clusters, namely the 1Mo7Fe M center and the 8Fe P cluster (see Section 1.2).

3.4. Determining the Redox Potential of Fe Centers

Due to the fact that different redox states of a Fe center exhibit distinct Mössbauer spectra and that the relative concentration of each redox state can be directly measured from the absorption area of its corresponding spectrum, Mössbauer spectroscopy is ideal for monitoring redox titration of Fe proteins. This application is particularly useful for proteins containing multiple Fe centers, since all centers can be monitored simultaneously. For example, Mössbauer spectroscopy had been employed to determine the redox properties of all Fe centers in the Desulfovibrio gigas NiFe hydrogenase (49).

3.5. Identifying Substrate and Inhibitor Binding Sites

Binding of substrate or inhibitor to a particular Fe site alters its electronic properties and, therefore, its corresponding Mössbauer spectrum. This allows easy identification of the substrate/ inhibitor binding site within a multi-center containing protein. Likewise, Mössbauer spectroscopy has the capability of identifying binding site within a multi-Fe containing center. In addition, the electronic state of the substrate-bound Fe site/center can be characterized in detailed. Taken together, the information can provide molecular-level insights into enzyme mechanism. For example, Mössbauer studies of an S-adenosylmethionine (SAM)-radical enzyme, the pyruvate formate-lyase activating enzyme, showed that SAM binds to a unique Fe site of the [4Fe–4S] cluster at the 2+ oxidation state. This information indicates unambiguously that binding of the co-substrate SAM preceded the electron transfer and substrate cleavage steps (50).

3.6. Obtaining Kinetic and Spectroscopic Characterization of Reaction Intermediates

Mössbauer spectroscopy can also be used to characterize reaction intermediates trapped by rapid-mix rapid-freeze quench (RFQ) technique (51) or by annealing cryo-reduced enzymes (52). Since RFQ technique can quench reaction at well-defined reaction time points, and Mössbauer spectroscopy can provide quantitative measure on the accumulations of the intermediates and products, together, they can afford kinetic information on the temporal formations and decays of intermediates and products (21, 23, 28) and thus provide molecular-level mechanistic information.

230

Huynh

3.7. Investigating the Functional Role of Specific Residue

Combining site-directed mutagenesis with the above described kinetic Mössbauer approach, the functional role of specific residue can be addressed. Knowing the function of specific residue not only can enhance our understanding of how protein controls the reactivities of Fe cofactors, the information is also important for rational design of de novel enzymes. We have employed this method successfully in determining the functional roles of two key residues (W48 and D84) in the R2 subunit of E. coli ribonucleotide reductase (37, 53) and have demonstrated that proper substitution of these two residues converts the radical-generating R2 into a monooxygenase (36).

4. Notes 4.1. Energy Scale

The physical quantities measured by Mössbauer spectroscopy are the hyperfine interactions between the 57 Fe nucleus and its surrounding electrons. Hyperfine interactions are in the order of 10–7 –10–6 eV. Since the incoming recoilless photons have a linewidth that is two to three orders of magnitude smaller than the hyperfine interactions, they are ideal for measuring the hyperfine interactions. Conventional Mössbauer measurements employ the Doppler effect to vary the source photon energy by moving the source against a stationary sample. The change in the photon energy, E, is directly proportional to the velocity, v, of the source, E =

v × 14.4 keV c

[1]

where c is the velocity of light. Thus, the energy scale for 57 Fe Mössbauer spectra is generally expressed in units of velocity (mm/s) of the source. From equation [1], a source velocity of 1 mm/s is equivalent to a change of the source γ photon energy by ∼4.8×10–8 eV. 4.2. Hyperfine Interactions and Mössbauer Spectral Type

The 57 Fe nucleus is positively charged. Its ground state has a nuclear spin, I = 1/2, and possesses a magnetic moment, μ = gn βn I, where gn = 0.181 and β n is the nuclear magneton (0.05 × 10−27 J/T) . The 14.4 keV excited state has a spin I = 3/2 with gn = –0.103 and a quadrupole moment, Q , of about 1 × 10−23 m2 . The nucleus can therefore interact with its local electric field and magnetic field, generated by the surrounding electrons, through its charge (monopole interaction), quadrupole moment (quadrupole interaction) and magnetic moment (magnetic dipole interaction). Fig. 15.3 shows a schematic energy

Mössbauer Spectroscopy

231

diagram depicting the energy levels of the ground and 1st excited state of 57 Fe under the influence of hyperfine interactions. This figure also displays the possible nuclear transitions at different situations and the corresponding Mössbauer spectra. 4.2.1. Isomer Shift

The monopole interaction is an electrostatic interaction. It is proportional to the size of the nucleus and the electronic density at the nucleus, |ψ (0)|2 . For 57 Fe, the radius of the excited state, Re , is smaller than that of the ground state, Rg , resulting in an effect on the transition energy given by equation [2] (5), E =

2 2π 2 2 Z e (Re − Rg2 )ψ(0) 3

[2]

Since the 57 Fe nuclei in the source and sample have different electronic environments, the electronic densities at the source nucleus and at the sample nucleus are different. The transition energies between their ground and excited states are therefore also not the same. The net effect is shifting the entire Mössbauer spectrum away from the zero velocity of the source. The amount of the shift is termed the isomer shift, δ. For 57 Fe spectrum, δ is conventionally quoted in reference to the center of a room temperature spectrum of a metallic Fe foil. It is sensitive to the electronic environment surrounding the Fe nucleus and is therefore a good measure for the oxidation, spin, and ligation state of the Fe atom. The following lists the general trend observed for δ, • Oxidation dependence: δ(FeII ) > δ(FeIII ) > δ(FeIV ) • Spin dependence: δ(high spin) > δ(low spin) • Coordination dependence: δ(octathedral) > δ(tetrahedral) • Ligand dependence: δ(O) > δ(N) > δ(S) To illustrate the above dependences, Fig. 15.4 shows the range of δ generally observed for high-spin FeII and FeIII in various ligand environments. Of particular interests are the narrow ranges detected for tetrahedral sulfur coordinated FeII and FeIII in Fe–S proteins (Fig. 15.4). 4.2.2. Quadrupolar Interaction

The quadrupole moment of the excited state of the 57 Fe nucleus can interact with local electric field gradients (EFG) generated

Fig. 15.4. Values of isomer shift observed for high-spin FeII and FeIII centers with oxygenous/nitrogenous ligands (thick bars) or a tetrahedral sulfur coordination (arrows).

232

Huynh

by non-isotropic distributions of the surrounding electrons. This interaction splits the excited state into two levels (Fig. 15.3) with an energy separation given by equation [3]. eQ Vzz EQ = 2

η2 3

1+

[3]

where the asymmetry parameter η = (Vxx − Vyy )/Vzz and Vii s are the principal components of the EFG. The EFG has no effect on the energy of the ground state due to the absence of a quadrupole moment in the ground state. Thus, there are two possible transitions from the ground to the excited state, resulting in two equal intensity absorption lines in the Mössbauer spectrum (Fig. 15.3). Such a spectrum is called a quadrupole doublet, and the energy separation between the two absorption lines is equal to EQ (equation [3]), which is termed the quadrupole splitting. The value of EQ depends, to a large extent, on the distributions of the valence electrons and thus the oxidation and spin state of the Fe ion. To a lesser extent, it also depends on the symmetry of the ligand environment. The following lists the ranges and commonly observed values (in parentheses) of EQ for various spin and oxidation states of Fe. • High-spin FeII : 2.0–4.0 mm/s (∼3.0 mm/s) • Low-spin FeII : 0.3–2.0 mm/s (∼1.0 mm/s) • High-spin FeIII : 0.3–2.0 mm/s (∼0.5 mm/s) • Low-spin FeIII : 0.6–2.5 mm/s (∼2.0 mm/s) It can be seen that except for high-spin FeII , the ranges of EQ for other states overlap substantially. Consequently, by itself alone, EQ is not a reliable parameter for identifying spin and oxidation states of Fe complexes except for high-spin FeII . Judicial applications of ΔEQ in conjunction with other parameters and/or additional information are required. 4.2.3. Magnetic Hyperfine Interaction and Magnetic Spectrum

Due to the presence of the nuclear spin, I, both the ground and the excited state of 57 Fe can interact with local magnetic field at the nucleus, generated either by application of an external magnetic field, Bapp , or by the surrounding unpaired electrons, which is termed the internal field, Bint , or both. The energy Hamiltonian representing these magnetic interactions can be written as:

Hm = −gn βn [Bapp + Bint ] · I = −gn βn

S · A Bapp − · I [4] gn β n

where S is the expectation value of the electronic spin and A is the magnetic hyperfine tensor, which is a measure of the strength of interaction between the nuclear and the electronic spin. This magnetic interaction will split the ground state into two and the

Mössbauer Spectroscopy

233

excited state into four, resulting in, most likely, six allowed transitions and a magnetic spectrum consisting of six absorption lines of varying intensities (Fig. 15.3). For 57 Fe, the internal field per electronic spin, A/gn βn , is in the order of 20 T. Thus, in a weak applied field (e.g., 50 mT), the shape and overall spread of the magnetic Mössbauer spectrum is mainly determined by the magnetic hyperfine tensor, A, and the electronic spin expectation value S. The value of A is sensitive to ligand environments. For example, for high-spin FeIII , the value of A/gn β n ranges from 14 to 16 T for tetrahedral S coordination while for octahedral N/O ligations it varies from 20 to 22 T. The expectation value

S is determined by the electronic structure and relaxation. For example, at low temperature (4.2 K), the electronic relaxation rate is slow (<< 108 s−1 ). The expectation value S is equal to zero for a non-Kramers system (integer spin system) with significant zero-field splitting. On the other hand, S is not equal to zero for a Kramers system (half-integer spin system) (3). Consequently, integer spin systems generally exhibit quadrupole doublet at low temperature and weak field, while half-integer spin system shows magnetic Mössbauer spectrum. At temperatures higher than 77 K, electronic relaxation of Fe in proteins is generally fast. The thermal average of S is close to zero. Thus, high-temperature Mössbauer spectra of Fe proteins are generally quadrupole doublets.

References 1. Mössbauer RL (1958) Kernresonanzfluoreszenz von gammastrahlung in Ir-191. Zeitschrift Für Physik 151:124–143 2. Schunemann V, Winkler H (2000) Structure and dynamics of biomolecules studied by Mössbauer spectroscopy. Rep Prog Phys 63:263–353 3. Huynh BH, Kent TA (1984) Mössbauer studies of iron proteins. In: Eichhorn GL, Marzilli LG (eds) Advances in Inorganic Biochemistry, vol. 6, pp. 164–223. Elsevier, Amsterdam 4. Lang G, Marshall W (1966) Mössbauer effect in some haemoglobin compounds. Proc Phys Soc London 87:3–34 5. Debrunner PG (1983) Mössbauer spectroscopy of iron porphyrins. In: Lever ABP, Gray HB (eds) Iron Porphyrins, Part III, pp. 137–234. VCH Publishers, New York, NY 6. Sands RH, Dunham WR (1974) Spectroscopic studies on 2-iron ferredoxins. Quart Rev Biophys 7:443–504 7. Trautwein AX, Bill E, Bominaar EL et al (1991) Iron-containing proteins and related

8.

9.

10.

11.

analogs – complementary Mössbauer, EPR and magnetic-susceptibility studies. Struct Bond 78:1–95 Münck E, Rhodes H, Orme-Johnson WH et al (1975) Nitrogenase .8. Mössbauer and EPR spectroscopy – MoFe protein component from Azotobacter-vinelandii OP. Biochim Biophys Acta 400:32–53 Rawlings J, Shah VK, Chisnell JR et al (1978) Novel metal cluster in iron-molybdenum cofactor of nitrogenase – spectroscopic evidence. J Biol Chem 253:1001–1004 Zimmermann R, Orme-Johnson WH, Münck E et al (1978) Nitrogenase-X – Mössbauer and EPR studies on reversibly oxidized MoFe protein from Azotobacter-vinelandii OP – Nature of iron centers. Biochim Biophys Acta 537:185–207 Emptage MH, Kent TA, Huynh BH et al (1980) Nature of the iron-sulfur centers in a ferredoxin from Azotobacter vinelandii – Mössbauer studies and cluster displacement experiments. J Biol Chem 255: 1793–1796

234

Huynh

12. Huynh BH, Moura JJG, Moura I et al (1980) Evidence for a three iron center in a ferredoxin from Desulfovibrio gigas. J Biol Chem 255:3242–3244 13. Kent TA, Dreyer JL, Kennedy MC et al (1982) Mössbauer studies of beef-heart aconitase – Evidence for facile interconversions of iron-sulfur clusters. Proc Natl Acad Sci USA 79:1096–1100 14. Moura JJG, Moura I, Kent TA et al (1982) Interconversions of [3Fe-3S] and [4Fe-4S] clusters – Mössbauer and electron paramagnetic resonance studies of Desulfovibrio gigas ferredoxin II. J Biol Chem 257: 6259–6267 15. Christner JA, Münck E, Janick PA et al (1981) Mössbauer spectroscopic studies of Escherichia coli sulfite reductase – evidence for coupling between the siroheme and Fe4 S4 cluster prosthetic groups. J Biol Chem 256:2098–2101 16. Huynh BH, Kang L, Dervartanian DV et al (1984) Characterization of a sulfite reductase from Desulfovibrio vulgaris – Evidence for the presence of a low-spin siroheme and an exchange-coupled siroheme-[4Fe-4S] unit. J Biol Chem 259:5373–5376 17. Noodleman L, Case DA (1992) Densityfunctional theory of spin polarization and spin coupling in iron-sulfur clusters. Adv Inorg Chem 38:423–487 18. Blondin G, Bominaar EL, Girerd JJ et al (1995) Spin dependent electron delocalization, vibronic and antiferromagnetic couplings in iron sulfur clusters. In: LaMar GN (ed) Nuclear Magnetic Resonance of Paramagnetic Macromolecules, vol. 457, pp. 369–386. Kluwer, Dordrecht 19. Krebs C, Edmondson DE, Huynh BH (2002) Demonstration of peroxodiferric intermediate in M-ferritin ferroxidase reaction using rapid freeze-quench Mössbauer, resonance Raman, and XAS spectroscopies. Methods Enzymol 354:436–454 20. Bollinger JM, Edmondson DE, Huynh BH et al (1991) Mechanism of assembly of the tyrosyl radical dinuclear iron cluster cofactor of ribonucleotide reductase. Science 253:292–298 21. Bollinger JM, Tong WH, Ravi N et al (1994) Mechanism of assembly of the tyrosyl radical-diiron(III) cofactor of Escherichia coli ribonucleotide reductase. 2. Kinetics of the excess Fe2+ reaction by optical, EPR and Mössbauer spectroscopies. J Am Chem Soc 116:8015–8023 22. Ravi N, Bollinger JM, Huynh BH et al (1994) Mechanism of assembly of the tyrosyl radical-diiron(III) cofactor of Escherichia coli

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

ribonucleotide reductase. 1. Mössbauer characterization of the diferric radical precursor. J Am Chem Soc 116:8007–8014 Liu KE, Valentine AM, Wang DL et al (1995) Kinetic and spectroscopic characterization of intermediates and component interactions of methane monooxygenase from Methylococcus capsulatus (Bath). J Am Chem Soc 117:10174–10185 Pereira AS, Small W, Krebs C et al (1998) Direct spectroscopic and kinetic evidence for the involvement of a peroxodiferric intermediate during the ferroxidase reaction in fast ferritin mineralization. Biochemistry 37:9871–9876 Price JC, Barr EW, Glass TE et al (2003) Evidence for hydrogen abstraction from C1 of taurine by the high-spin Fe(IV) intermediate detected during oxygen activation by taurine: alpha-ketoglutarate dioxygenase (TauD). J Am Chem Soc 125:13008–13009 Price JC, Barr EW, Tirupati B et al (2003) The first direct characterization of a high-valent iron intermediate in the reaction of an alpha-ketoglutarate-dependent dioxygenase: A high-spin Fe(IV) complex in taurine/alpha-ketoglutarate dioxygenase (TauD) from Escherichia coli. Biochemistry 42:7497–7508 Xing G, Diao YH, Hoffart LM et al (2006) Evidence for C-H cleavage by an ironsuperoxide complex in the glycol cleavage reaction catalyzed by myo-inositol oxygenase. Proc Natl Acad Sci USA 103: 6130–6135 Murray LJ, Naik SG, Ortillo DO et al (2007) Characterization of the arene-oxidizing intermediate in ToMOH as a diiron(III) species. J Am Chem Soc 129:14500–14510 Jiang W, Yun D, Saleh L et al (2007) A manganese(IV)/iron(III) cofactor in Chlamydia trachomatis ribonucleotide reductase. Science 316:1188–1191 Bollinger JM, Stubbe J, Huynh BH et al (1991) Novel diferric radical intermediate responsible for tyrosyl radical formation in assembly of the cofactor of ribonucleotide reductase. J Am Chem Soc 113: 6289–6291 Murray LJ, Garcia-Serres R, Naik S et al (2006) Dioxygen activation at non-heme diiron centers: Characterization of intermediates in a mutant form of toluene/o-xylene monooxygenase hydroxylase. J Am Chem Soc 128:7458–7459 Hwang J, Krebs C, Huynh BH et al (2000) A short Fe-Fe distance in peroxodiferric ferritin: Control of Fe substrate versus cofactor decay? Science 287:122–125

Mössbauer Spectroscopy 33. Moenne-Loccoz P, Krebs C, Herlihy K et al (1999) The ferroxidase reaction of ferritin reveals a diferric mu-1,2 bridging peroxide intermediate in common with other O2 -activating non-heme diiron proteins. Biochemistry 38:5290–5295 34. Liu KE, Wang DL, Huynh BH et al (1994) Spectroscopic detection of intermediates in the reaction of dioxygen with the reduced methane monooxygenase hydroxylase from Methylococcus capsulatus (Bath). J Am Chem Soc 116:7465–7466 35. Lee SK, Fox BG, Froland WA et al (1993) A transient intermediate of the Methane monooxygenase catalytic cycle containing an Fe(IV)Fe(IV) cluster. J Am Chem Soc 115:6450–6451 36. Baldwin J, Voegtli WC, Khidekel N et al (2001) Rational reprogramming of the R2 subunit of Escherichia coli ribonucleotide reductase into a self-hydroxylating monooxygenase. J Am Chem Soc 123:7017–7030 37. Bollinger JM, Krebs C, Vicol A et al (1998) Engineering the diiron site of Escherichia coli ribonucleotide reductase protein R2 to accumulate an intermediate similar to H-peroxo, the putative peroxodiiron(III) complex from the methane monooxygenase catalytic cycle. J Am Chem Soc 120:1094–1095 38. Smith BE, Lang G (1974) Mössbauer spectroscopy of nitrogenase proteins from Klebsiella pneumoniae – Structural assignments and mechanistic conclusions. Biochem J 137:169–180 39. Huynh BH, Münck E, Orme-Johnson WH (1979) Nitrogenase-XI – Mössbauer studies of the cofactor centers of the MoFe protein from Azotobacter vinelandii-OP. Biochim Biophys Acta 576:192–203 40. Huynh BH, Henzl MT, Christner JA et al (1980) Nitrogenase.12. Mössbauer studies of the MoFe protein from Clostridium pasteurianum W5. Biochim Biophys Acta 623: 124–138 41. Chan MK, Kim JS, Rees DC (1993) The nitrogenase FeMo cofactor and P cluster pair – 2.2-angstrom resolution structures. Science 260:792–794 42. Peters JW, Stowell MHB, Soltis SM et al (1997) Redox-dependent structural changes in the nitrogenase P-cluster. Biochemistry 36:1181–1187 43. Einsle O, Tezcan FA, Andrade SLA et al (2002) Nitrogenase MoFe-protein at 1.16

44. 45.

46.

47.

48.

49.

50.

51.

52.

53.

235

angstrom resolution: A central ligand in the FeMo-cofactor. Science 297:1696–1700 Shah VK, Brill WJ (1977) Isolation of an iron-molybdenum cofactor from nitrogenase. Proc Natl Acad Sci USA 74:3249–3253 Yoo SJ, Angove HC, Papaefthymiou V et al (2000) Mössbauer study of the MoFe protein of nitrogenase from Azotobacter vinelandii using selective Fe-57 enrichment of the M-centers. J Am Chem Soc 122:4926–4936 Lindahl PA, Papaefthymiou V, OrmeJohnson WH et al (1988) Mössbauer studies of solid thionine-oxidized MoFe protein of nitrogenase. J Biol Chem 263:19412–19418 Surerus KK, Hendrich MP, Christie PD et al (1992) Mössbauer and integer-spin EPR of the oxidized P-cluster of nitrogenase – POX is a non-Kramers system with a nearly degenerate ground doublet. J Am Chem Soc 114:8579–8590 Emptage MH, Zimmermann R, Que L et al (1977) Mössbauer studies of cytochrome C ’ from Rhodospirillum rubrum. Biochim Biophys Acta 495:12–23 Huynh BH, Patil DS, Moura I et al (1987) On the active sites of the [NiFe] hydrogenase from Desulfovibrio gigas – Mössbauer and redox titration studies. J Biol Chem 262:795–800 Krebs C, Broderick WE, Henshaw TF et al (2002) Coordination of adenosylmethionine to a unique iron site of the [4Fe-4S] of pyruvate formate-lyase activating enzyme: A Mössbauer spectroscopic study. J Am Chem Soc 124:912–913 Krebs C, Price JC, Baldwin J et al (2005) Rapid freeze-quench Fe-57 Mössbauer spectroscopy: Monitoring changes of an ironcontaining active site during a biochemical reaction. Inorg Chem 44:742–757 Garcia-Serres R, Davydov RM, Matsui T et al (2007) Distinct reaction pathways followed upon reduction of oxy-heme oxygenase and oxy-myoglobin as characterized by Mössbauer spectroscopy. J Am Chem Soc 129:6662–6662 Krebs C, Chen SX, Baldwin J et al (2000) Mechanism of rapid electron transfer during oxygen activation in the R2 subunit of Escherichia coli ribonucleotide reductase. 2. Evidence for and consequences of blocked electron transfer in the W48F variant. J Am Chem Soc 122:12207–12219

Section IV Chemical Methods

Chapter 16 Protocols for Cofactor Isolation of Nitrogenase Aaron W. Fay, Chi-Chung Lee, Jared A. Wiig, Yilin Hu, and Markus W. Ribbe Abstract The iron-molybdenum cofactor (FeMoco) of the nitrogenase MoFe protein has remained a focal point in the field of bioinorganic chemistry for decades. This unique metal cluster has long been regarded as the actual site of dinitrogen reduction, and it is structurally complex and chemically unprecedented. A detailed characterization of the isolated FeMoco is crucial for elucidating the physiochemical properties of this biologically important cofactor. Such a study requires an effective technique to extract FeMoco intact, and in high yield, from the MoFe protein. A method involving the acid treatment of the MoFe protein and the subsequent extraction of FeMoco into an organic solvent was developed over 30 years ago and has been improved upon ever since. FeMoco isolated by this strategy is catalytically active and spectrally interesting, which provides a useful platform for future structure–function analyses of this unique cofactor. A general working protocol for FeMoco isolation is described in this chapter, along with some of the major modifications reported in the past years. Key words: Nitrogenase, FeMoco, NMF, DMF, extraction.

1. Introduction Molybdenum (Mo) nitrogenase is a complex metalloenzyme that utilizes the energy of ATP hydrolysis for the chemically challenging conversion of inert dinitrogen to bioavailable ammonia (1–4). This amazing feat cannot be achieved without the two unique metal clusters of its α2 β2 -tetrameric MoFe protein component: the P-cluster ([Fe8 S7 ]), which is located at the α/β subunit interface of the MoFe protein; and the iron-molybdenum cofactor, or FeMoco ([MoFe7 S9 X-homocitrate], where X = C, N or O), which is situated within the α subunit of the MoFe protein (5, 6). Both clusters are structurally complex and chemically M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_16, © Springer Science+Business Media, LLC 2011

239

240

Fay et al.

unprecedented, and both play critical roles in mediating the electron flow for nitrogenase catalysis. In particular, the FeMoco is believed to be the actual site of substrate reduction, which makes it an attractive target for investigations of mechanistic details of nitrogenase. The first successful attempt of FeMoco isolation was reported in 1977 (7). This work represented a major breakthrough in the field of nitrogenase research, as it provided an effective means for the biochemical and spectroscopic characterization of FeMoco without the interference of the P-cluster. Indeed, follow-up studies of the isolated FeMoco have provided valuable insights into the physiochemical properties, such as the net charges, optical properties, redox properties, chemical reactivity, and electronic structure, of this unique cluster (8–12). Furthermore, the ability of isolated FeMoco to reconstitute the FeMoco-deficient nifB MoFe protein not only confirms the physiological relevance of the extracted cluster species but also serves as an indispensable tool for investigations of the final step in the biosynthesis of MoFe protein, i.e., the insertion of FeMoco into its targeted location within the MoFe protein (13).

MoFe Protein

1) H2O 2) Citric Acid (or HCl or H3PO4) 3) Na2HPO4 (or NaOH or Na2CO3) Supernatant

Discarded

Supernatant

Pellet

1) Washed 2 times with DMF (or acidic NMF) 2) Centrifuged

Discarded Pellet

Supernatant

1) Washed 2 times with NMF (or Formamide) 2) Centrifuged

NMF solution of FeMoco

Pellet

Discarded

Fig. 16.1. Flow diagram of FeMoco extraction protocol.

Protocols for Cofactor Isolation of Nitrogenase

241

The major challenge for developing a strategy for FeMoco isolation lies in the fact that FeMoco is extremely labile toward O2 exposure, especially when it is not protected by the protein “cage.” Thus, the extraction of FeMoco must be performed under strictly anaerobic conditions, which significantly complicates the entire procedure. Moreover, it is difficult to extract FeMoco out of the protein environment with mild solvents and, upon isolation, FeMoco decomposes rapidly (at an approximate rate of 25% per hour) in aqueous solutions. As a result, certain “aggressive” organic solvents, such as N-methylformamide (NMF), need to be used for FeMoco extraction, which could impact the structural integrity and catalytic competence of the isolated cluster. Finally, a substantial amount of purified MoFe

Table 16.1 Modifications of the FeMoco isolation method Reagents used for Method

Unfolding and precipitation

Precipitate washing

FeMoco extraction

1a

Citrate/phosphate

0.1 M citrate 0.2 M Na2 HPO4

DMF

NMF (Na2 HPO4 )

1b

Citrate/phosphate, pH 2

Citrate (to pH 2) Na2 HPO4 (to pH 5)

DMF (Na2 HPO4 )

NMF

2

HCl/NaOH

0.1 M HCl 0.1 M NaOH

DMF

NMF (NaOH)

3

Acidic NMF

0.1 M citrate 0.2 M Na2 HPO4

Acidic NMF

NMF (Na2 HPO4 )

4

Formamide

0.1 M citrate 0.2 M Na2 HPO4

DMF

HCONH2 (Na2 HPO4 )

5

Phosphate

0.33 M H3 PO4 0.2 M Na2 HPO4

DMF

NMF (Na2 HPO4 )

6

S2 O24 free

DMF

7

DMF

0.1 M HCl 0.1 M NaOH 0.1 M citrate 0.2 M Na2 HPO4

NMF (NaOH) DMF ([Et4 N]2 [S2 O4 ]); ([Et4 N][OH]) (2 mM)

8

MeCN

0.1 M citrate 0.2 M Na2 HPO4

DMF

MeCN ([Et4 N]2 [S2 O4 ]); ([Et4 N][OH]) (50 mM)

9

DMSO

DMSO

DMF

NMF (Na2 HPO4 )

10

Pyrrolidinone

0.1 M citrate 0.2 M Na2 HPO4

DMF

Pyrrolidinone

−

DMF

242

Fay et al.

protein is required for the extraction of a limited amount of FeMoco, as the theoretically perfect yield of such a procedure is only 4 mg FeMoco per 500 mg MoFe protein (7). The original isolation of FeMoco (Fig. 16.1) involved acid treatment of purified MoFe protein (7, 14), followed by pH neutralization and extensive washing with dimethylformamide (DMF). The FeMoco is finally extracted in N-methylformamide (NMF) and concentrated to about 1–2 mM. This strategy is still the underlying framework of most FeMoco isolation protocols to date, although specific modifications of each experimental step have been made in these later protocols to optimize this procedure (Table 16.1) (14). Here, we describe the most commonly used and relatively simple protocol, which involves the precipitation of the MoFe protein by HCl–NaOH, the subsequent extraction of FeMoco by NMF, and the final concentration of the extracted FeMoco by vacuum distillation. We also describe an alternative method, which involves the citric acid–phosphate treatment of the MoFe protein and the pyrrolidinone-based extraction of FeMoco (7, 15) (see Note 1).

2. Materials Unless otherwise noted, all chemicals are purchased from SigmaAldrich (St Louis, MO). 2.1. Standard HCl–NaOH–NMF Extraction Method

1. 200 ml H2 O. 2. 200 ml 0.1 M HCl. 3. 200 ml 0.1 M NaOH. 4. 200 ml distilled DMF (for distillation technique see Section 3.1). 5. 200 ml distilled NMF (for distillation technique see Section 3.1). 6. 500 mg purified MoFe protein in a buffer containing 25 mM Tris, pH 7.4, 0.5 M NaCl. 7. Vacuum distillation concentrator apparatus: one 200 ml round-bottomed flask, a short glass tube with 24/40 joints, a three-necked 1 l round-bottomed flask with a glass stopcock adaptor to a vacuum tube.

2.2. Alternative Citrate–Phosphate– Pyrrolidinone Extraction Method

1. 200 ml H2 O. 2. 200 ml 1 M citric acid. 3. 200 ml 0.5 M sodium phosphate, dibasic (Na2 HPO4 ).

Protocols for Cofactor Isolation of Nitrogenase

243

4. 200 ml distilled DMF (for distillation technique see Section 3.1). 5. 200 ml re-crystallized 2-pyrrolidinone. 6. 500 mg purified MoFe protein in a buffer containing 25 mM Tris, pH 7.4, 0.5 M NaCl. 7. Vacuum distillation concentrator apparatus: one 200 ml round-bottomed flask, a short glass tube with 24/40 joints, a three-necked 1 l round-bottomed flask with a glass stopcock adaptor to a vacuum tube.

3. Methods 3.1. Distillation of NMF and DMF

1. NMF typically needs to be pretreated overnight with saturating amounts of sodium bicarbonate (9) (see Note 2). The excess salt in NMF should be filtered out using glass filters. 2. NMF is then transferred to the distilling flask with a thermometer-containing still head. Connect the distilling flask to the receiving flask with a condenser arm that is connected to a vacuum line. The entire apparatus should be placed inside a fume hood, and it should be examined carefully for any potential leak before usage. 3. Place the distilling flask on a heating mantle. Heat gently and carefully monitor the NMF solution to prevent it from boiling over. The head temperature should not significantly deviate from the heating temperature during this process. 4. NMF should condense between 50 and 60◦ C. Collect this fraction and store it at –20◦ C until use (see Note 3). 5. DMF can be distilled using the same method, following steps 2–4.

3.2. Standard HCl–NaOH–NMF Extraction Method

1. Prepare all solutions described in Section 2.1 (see Note 4). Degas 500 mg of purified MoFe protein in a three-necked vacuum flask, where a pH probe is inserted. 2. Place all solutions and the flask with protein on an ice bath. 3. Dilute MoFe protein to a final concentration of 5 mg/ml with the anaerobic H2 O. 4. Using a cannula, anaerobically add 0.1 M HCl to the protein solution. Titrate the solution to pH 2.2 (see Note 5). 5. Allow the protein solution to sit on ice for 2 min. Do not let the protein solution stay at low pH for more than 3 min since FeMoco will start to degrade at this point (see Note 6).

244

Fay et al.

6. Immediately titrate the solution to pH 5.5 with 0.1 M NaOH (see Note 7). The protein will precipitate at this pH and the solution will turn a cloudy, light gray-brown color. 7. Allow this solution to sit on ice for 25 min and then carefully transfer it to air-tight centrifugation bottles and centrifuge at 1000×g, 4◦ C for 10 min. The colorless supernatant is discarded afterward. 8. Add 10–20 ml of distilled DMF to each pellet and quickly vortex the mixture for 3 s to re-suspend the pellet. Insufficient (i.e., pellet is not re-suspended) or excess (i.e., more than 5 s) vortex lowers the yield of extraction significantly. 9. Centrifuge at 2000×g, 4◦ C for 10 min. The supernatant is discarded afterward (see Note 8). 10. Repeat steps 8 and 9 as above. The pellet should be much darker than the aqueous pellet formed right after the acidification step (step 6). 11. Add 10 ml of distilled NMF to the protein pellets (see Note 9). Vortex for 5 min, then centrifuge the mixture at 2000×g, 4◦ C for 10 min (see Note 10). 12. Carefully remove and save the supernatant. Add 8 ml of NMF to each pellet, vortex for 5 min, and centrifuge the mixture as described in step 9. 13. Carefully remove the supernatant and pool it with the supernatant from the previous batch of extraction. 14. If the pellet at this point is still dark-brown in color, an additional extraction step can be carried out as described in steps 12 and 13. 15. In a glovebox, centrifuge the pooled supernatant from the NMF extraction steps (i.e., a FeMoco-NMF solution) at 13,000×g for 5 min in an Eppendorf centrifuge. Carefully remove the supernatant and discard any residual protein precipitate. 16. Transfer the FeMoco–NMF solution from step 15 to a 200 ml round-bottomed flask and place a suitably sized stir bar in it. Connect this flask to a 1 l round-bottomed flask, which will serve as a liquid N2 vapor trap, with a short glass tube with 24/40 joints. The vapor trap (i.e., the 1 l flask) is then connected a strong vacuum line (7.6 × 10–5 mmHg) through a glass stopcock adaptor. To concentrate the FeMoco, the solution is slowly stirred at 30◦ C where the vapor trap is constantly chilled with liquid N2 and under vacuum suction. 17. Concentrate the FeMoco–NMF solution about 10-fold (see Note 11). The time required for this step varies between

Protocols for Cofactor Isolation of Nitrogenase

245

1 and 4 h. It is not recommended to dry off the NMF under vacuum, as it tends to lower the activity of FeMoco. 18. Store the concentrated FeMoco–NMF solution in sealed vials at –80◦ C. FeMoco typically stays active under this condition for 3–6 months. 3.3. Alternative Citrate–Phosphate– Pyrrolidinone Extraction Method

1. Prepare all solutions described in Section 2.2 (see Note 4). Degas 500 mg of purified MoFe protein in a three-necked vacuum flask, where a pH probe is inserted. 2. Place all solutions and the flask with protein on an ice bath except the sodium phosphate buffer (as the pKa of Na2 HPO4 varies with temperature) and the pyrrolidinone solution (as pyrrolidinone has a low melting point of 20.5◦ C). 3. Dilute the MoFe protein to a final concentration of 5 mg/ml with the anaerobic H2 O. 4. Using an air-tight syringe, add 1 M citric acid to the protein solution. Titrate to a final pH of 2.2 (see Note 5). It takes approximately 0.3 ml of 1 M citric acid for this pH adjustment. 5. Allow the protein solution to sit on ice for 2 min. Do not let the protein solution stay at low pH for more than 3 min since FeMoco will start to degrade after this point (see Note 6). 6. Immediately titrate the solution to pH 5.5 with 0.5 M Na2 HPO4 (see Note 7). The protein will precipitate at this pH and the solution will turn a cloudy, light gray-brown color. It takes approximately 0.5 ml of 0.5 M Na2 HPO4 for this pH adjustment. 7. Allow this solution to sit on ice for 25 min, and then carefully transfer it to air-tight centrifugation bottles and centrifuge at 1000×g, 4◦ C for 10 min. The colorless supernatant is discarded afterward. 8. Add 0.5–2 ml distilled DMF to each pellet. Quickly vortex the mixture for 3 s to re-suspend the pellet. Insufficient (i.e., pellet is not re-suspended) or excess (i.e., more than 5 s) vortex lowers the yield of extraction significantly. 9. Centrifuge at 2000×g, 4◦ C for 10 min. The supernatant is discarded afterward (see Note 8). 10. Repeat steps 8 and 9 as above. The pellet should be much darker than the aqueous pellet formed right after the acidification step (step 6). 11. Allow the pellet to sit at room temperature for 10 min. All steps from this point on should be carried out at room temperature.

246

Fay et al.

12. Add 0.5–2 ml of re-crystallized pyrrolidinone to each protein pellet. Vortex for 5 min, then centrifuge the mixture at 2000×g, room temperature for 10 min (see Note 10). 13. Carefully remove and save the supernatant. Add 8 ml of pyrrolidinone to each pellet, vortex for 5 min, and centrifuge the mixture at 2000×g, room temperature for 10 min. 14. Carefully remove the supernatant and pool it with the supernatant from the previous batch of pyrrolidinone extraction. 15. Repeat steps 12–14 unless the pellet is nearly colorless at this point. Typically, 3-pyrrolidinone extraction steps are required for a good yield of FeMoco isolation. 16. In a glovebox, centrifuge the pooled supernatant from the pyrrolidinone extraction steps (i.e., a FeMoco– pyrrolidinone solution) at 13,000×g for 5 min in an Eppendorf centrifuge. Carefully remove the supernatant and discard any residual protein precipitate. 17. Transfer the FeMoco–pyrrolidinone solution in a 200 ml round-bottomed flask and place a suitably sized stir bar in it. Connect this flask to a 1 l round-bottom flask, which will serve as a liquid N2 vapor trap, with a short glass tube with 24/40 joints. The vapor trap (i.e., the 1 l flask) is then connected to a strong vacuum line (7.6 × 10–5 mmHg) through a glass stopcock adaptor. To concentrate the FeMoco, the solution is slowly stirred at 30◦ C, where the vapor trap is constantly chilled with liquid N2 and under vacuum suction. 18. Concentrate the FeMoco–pyrrolidinone solution about 5–10 fold (see Note 11). The time required for this step is between 4 and 8 h. 19. Store the concentrated FeMoco–pyrrolidinone solution in sealed vials at –80◦ C.

4. Notes 1. It has been established that HCl and citric acid are equally effective in unfolding the MoFe protein. 2. It has been suggested that the basic form of NMF (Fig. 16.2) is much more efficient in FeMoco extraction (16). Thus, NMF is pretreated with base (i.e., sodium bicarbonate) before distillation (9, 14).

Protocols for Cofactor Isolation of Nitrogenase O

H C

H

O C

N CH3

247

H

N CH3

Fig. 16.2. NMF is ineffective under acidic conditions (left) and effective under basic conditions (right) for FeMoco isolation.

3. Condensation could start at temperatures between 25 and 35◦ C. However, the condensed liquid in this temperature range is likely the impurities from NMF and this fraction should be discarded. 4. Na2 S2 O4 degrades readily in acidic or basic solutions. Therefore, Na2 S2 O4 should be added to the solutions immediately before use. 5. Upon pH titration, the MoFe protein undergoes a series of color changes. The MoFe protein solution has a darkbrown color at pH 7.4. At pH 5.5, however, the MoFe protein starts to precipitate, and the solution turns into a cloudy, light brownish-gray color. The protein re-dissolves at pH 2.2, where the solution has a pale greenish-brown color (7, 14, 15). 6. The P-cluster is completely destroyed at low pH, whereas FeMoco is transiently stabilized by the unfolded polypeptide. Since FeMoco degrades with prolonged exposure to low pH, it is necessary to determine the conditions under which the protein is completely unfolded while the integrity of FeMoco is maintained. 7. Dissipation of the heat that is released from the acid–base neutralization step can significantly increase the yield and activity of the isolated FeMoco. 8. Although the DMF wash is typically colorless after the vortex and centrifugation steps, it occasionally carries a light brown color. This phenomenon can be minimized by centrifuging the DMF mixture at a low temperature (9, 14). 9. Even though the secondary amides appear to be the most effective type of solvent, it is not an absolute requirement for FeMoco extraction. There have been reports suggesting that FeMoco can be dissolved in a variety of solvents, including methanol, tetrahydrofuran, and acetonitrile (17), and stabilized by positively charged counterions, such as Na+ , K+ , or tetrabutylammonium (18). 10. Centrifugation of the FeMoco-containing solution at a high speed (e.g., 8000×g) has been reported to lower the yield of FeMoco extraction.

248

Fay et al.

11. Excessive concentration of FeMoco should be avoided to prevent a significant increase in the concentrations of contaminating salts and protein residues that cannot be completely removed from the FeMoco solution.

Acknowledgments The authors are supported by National Institutes of Health grant GM 67626 (M.W.R.) and Herman Frasch Foundation grant 617-HF07 (M.W.R.). References 1. Burgess BK, Lowe DJ (1996) Mechanism of molybdenum nitrogenase. Chem Rev 96:2983–3012 2. Smith BE (1999) Structure, function, and biosynthesis of the metallosulfur clusters in nitrogenases. Adv Inorg Chem 47:159–218 3. Rees DC, Tezcan FA, Haynes CA et al (2005) Structural basis of biological nitrogen fixation. Philos Trans R Soc Lond Ser A363:971–984 4. Howard JB, Rees DC (2006) How many metals does it take to fix N2 ? A mechanistic overview of biological nitrogen fixation. Proc Natl Acad Sci USA 103:17088–17093 5. Einsle O, Tezcan FA, Andrade SL et al (2002) Nitrogenase MoFe-protein at 1.16 Å resolution: A central ligand in the FeMocofactor. Science 297:1696–1700 6. Kim J, Rees DC (1992) Crystallographic structure and functional implications of the nitrogenase molybdenum iron protein from Azotobacter vinelandii. Nature 360:553–560 7. Shah VK, Brill WJ (1977) Isolation of an iron-molybdenum cofactor from nitrogenase. Proc Natl Acad Sci USA 74:3249–3253 8. Frank P, Ghellar SF, Newton WE et al (1990) Purification and spectroscopic characteristics in N-methylformamide of the Azotobacter vinelandii Fe-Mo cofactor. Biochem Biophys Res Commun 163:746–754 9. Schultz FA, Gheller SF, Burgess BK et al (1985) Electrochemical characterization of the iron-molybdenum cofactor from Azotobacter vinelandii nitrogenase. J Am Chem Soc 107:5364–5368 10. Rewlings J, Shah VK, Chisnell JR et al (1978) Novel metal cluster in the iron-molybdenum cofactor of nitrogenase. Spectroscopic evidence. J Biol Chem 253:1001–1004

11. Lindahl PA, Papaefthymiou V, OrmeJohnson WH et al (1988) Mössbauer studies of solid thionin-oxidized MoFe protein of nitrogenase. J Biol Chem 263: 19412–19418 12. Conradkn SD, Burgess BK, Newton WE et al (1987) Structural studies of the molybdenum site in the MoFe protein and its FeMo cofactor by EXAFS. J Am Chem Soc 109: 7507–7515 13. Hu Y, Fay AW, Lee CC et al (2008) Assembly of nitrogenase MoFe protein. Biochemistry 47:3973–3981 14. Burgess BK (1990) The iron-molybdenum cofactor of nitrogenase. Chem Rev 90: 1377–1406 15. Walters MA, Chapman SK, Orme-Johnson WH (1986) The nature of amide ligation to the metal sites of FeMoco. Polyhedron 5:561–565 16. Yang SS, Pan WH, Friesen GD et al (1982) Iron-molybdenum cofactor from nitrogenase. Modified extraction methods as probes for composition. J Biol Chem 257: 8042–8048 17. McLean PA, Wink DA, Chapman SK et al (1989) A new method for extraction of iron-molybdenum cofactor (FeMoco) from nitrogenase adsorbed to DEAE-cellulose. 1. Effects of anions, cations, and preextraction treatments. Biochemistry 28: 9402–9406 18. Wink DA, McLean PA, Hickman AB et al (1989) A new method for extraction of iron-molybdenum cofactor (FeMoco) from nitrogenase adsorbed to DEAE-cellulose. 2. Solubilization of FeMoco in a wide range of organic solvents. Biochemistry 28: 9407–9412

Chapter 17 Techniques for Functional and Structural Modeling of Nitrogenase Patrick L. Holland Abstract Synthetic compounds play an important role in developing our understanding of nitrogenase enzymes, and over the years, a multitude of new metal-containing compounds have been created using nitrogenase as an inspiration. The techniques for handling coordination compounds in organic solvents are different than those commonly encountered in a biochemistry or molecular biology laboratory. This chapter describes the precautions that are essential for successful synthesis of air- and moisture-sensitive synthetic compounds and gives details of the synthesis of some coordinatively unsaturated iron-dinitrogen and iron-sulfide compounds of interest in nitrogenase research. Key words: Synthesis, iron–sulfur cluster, dinitrogen, iron complexes, air sensitive.

1. Introduction The use of synthetic coordination complexes in bioinorganic chemistry has a long history, especially as a tool for gaining insight into the workings of the active sites of metalloenzymes. The synthetic complexes are often described as “model complexes” or “synthetic analogues” to highlight the strategic mimicry of the active site electronic structure of the metal, the geometry, the ligands, and/or the functional groups (1, 2). Of course, studies on any models have a disadvantage in that the enzyme itself is not under investigation, and therefore it is not possible to conclusively determine an enzymatic mechanism using synthetic analogues. However, the much greater flexibility of synthetic chemistry enables chemists to study the inherent behavior of a metal coordinated similarly to that in the enzyme, without the influence M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_17, © Springer Science+Business Media, LLC 2011

249

250

Holland

of “second-sphere” (not directly bound) effects like (a) hydrogen bonding to amino acid residues, the protein backbone, or water and (b) steric hindrance from the protein that usually creates a specifically shaped active site pocket. This elucidates how the protein modulates the reactivity of the metal and vice versa. The results of the synthetic studies often teach bioinorganic chemists how to design further protein mutants or synthetic complexes to accomplish catalytic reactions, to understand the determinants of interesting spectroscopic features, and to appreciate the manner in which proteins have evolved to control biologically important transformations. It is important to notice that the term “ligand” has a different use in coordination chemistry from the one most common in biochemistry. To inorganic chemists, a “ligand” is a distinct Lewis base group attached to a metal through one or more coordination bonds. The coordination chemistry terminology will be used below. In nitrogenase, there are three metal-containing cofactors of interest, each of which is an iron–sulfur cluster (3). The Fe protein contains a 4Fe–4S cluster that is unusual in that it can be reduced to an unprecedented state that is proposed to have all the iron atoms at the ferrous oxidation level (FeII 4 ) (4, 5). This idea inspired research that culminated in the first synthetic examples of all-ferrous synthetic 4Fe–4S clusters (6, 7). These novel compounds were emblematic in that they used abiological ligands that were strongly electron withdrawing (CN– ) (6) or very polarizable (N-heterocyclic carbene) (7). These complexes confirmed that all-ferrous four-iron clusters are possible and also enabled the study of the spectroscopic properties and reactivity of stable FeII 4 S4 species. The second cofactor of interest in nitrogenase is the P cluster, an 8Fe–7S cluster with an unusual shape that changes between the oxidized and reduced forms (8). Figure 17.1 shows the reduced form of the P cluster. Two research groups have synthesized iron– sulfur clusters with similar topology to the reduced P cluster (9, 10). Again, spectroscopic and structural comparisons between the synthetic 8Fe–7S clusters and the native P cluster were pursued as part of these studies. These examples also highlight

Fig. 17.1. The P cluster of molybdenum nitrogenases in the reduced form.

Techniques for Functional and Structural Modeling of Nitrogenase

251

another beneficial aspect of synthetic molecules: because of their smaller size, X-ray crystal structures have significantly higher precision and accuracy, enabling a look at the bond lengths and angles in greater detail than is possible in the natural protein (with the caveat that the environment is not exactly the same as in the enzyme). The most intriguing cluster in the nitrogenases is the M cluster, which is the location at which N2 is reduced to ammonia (11). Each of the three best-characterized nitrogenases has an eightmetal M cluster, but the metal constitution varies from MoFe7 in molybdenum-dependent nitrogenases, to VFe7 in vanadiumdependent nitrogenases, to Fe8 in iron-only nitrogenases (12). The EXAFS (extended X-ray absorption fine structure) spectra are very similar between the different M clusters, suggesting that they have similar shapes (13). The molybdenum-dependent nitrogenases have been characterized in detail, and X-ray crystallography of the enzyme has shown the topology of the M cluster in the molybdenum-dependent enzymes (14). The M cluster of Modependent nitrogenases is typically called the “iron–molybdenum cofactor” or FeMoco (Fig. 17.2). The shape and function of the FeMoco raise a multitude of interesting questions, many of which provide inspiration for synthetic studies. For example, there has been controversy regarding the identity of the central light atom (designated “X”) in the cluster, because C, N, and O are all consistent with the crystallographic data (14). No spectroscopic technique has yet shown clear evidence that would distinguish between the possible identities of X. However, recent NRVS (nuclear resonant vibrational spectroscopy) investigations coupled with theory somewhat favor X = C (15); computational chemistry experiments favor X = N (16); and ENDOR (electron nuclear double resonance) labeling studies favor X = O (17). Therefore, it is easy to imagine that synthesis of FeMoco clusters with purposeful incorporation of N, O, or C in the center would provide comparisons that could be used to deduce the identity of X in the biological system. However, it has not yet been possible to synthesize FeMoco-like clusters with light atoms in the center, so this is an area for continued research (18).

Fig. 17.2. The M cluster (“FeMoco”) of molybdenum nitrogenases.

252

Holland

Synthetic studies may also be able to answer other questions about the biosynthesis of the FeMoco. A series of particularly pertinent questions are raised by the research of Ribbe, which has shown that the precursor to the FeMoco is a seven-iron cluster with the overall shape of the FeMoco, and only after creation of this precursor does the cluster incorporate the molybdenum atom (19). The synthesis of Fe7 S9 clusters would enable chemists to study the mechanism of molybdenum incorporation in a simple model system. Again, the appropriate synthetic precursors are not yet available, although the ability to exchange metals in one corner of a site-differentiated 4Fe–4S cluster gives tentative evidence that the proposed metal addition may be reasonable (20). It is interesting that N2 (one of the weakest Lewis bases known) can somehow coordinate to the FeMoco, despite the fact that all of the metal atoms in the FeMoco are coordinatively saturated in the isolated form: the iron atoms have a tetrahedral geometry and the molybdenum atom has an octahedral geometry. However, during catalysis there must be some open site for N2 binding! This open site is apparently not present in nitrogenase’s crystallographically characterized form, and instead, substrates bind after reduction of the FeMoco by three or four electrons (21). Therefore, reduction of the FeMoco is likely to create an open site on a metal, enabling substrate binding. Mutations of iron–molybdenum nitrogenase most strongly suggest that binding (and hence this open coordination site) occurs on the iron “belt” of the cluster adjacent to Gly69α (22). Synthetic chemists have not yet produced an iron–sulfur cluster that binds N2 , and this is an outstanding challenge in the field. Synthetic chemists can contribute by synthesizing iron–sulfur clusters with open coordination sites, in order to comprehend the binding abilities, spectroscopy, and accessibility of potential FeMoco intermediates. On a more fundamental level, they can determine what types of iron environments enable iron to transfer electrons into N2 , enabling reduction of the N–N triple bond. Both goals require the formation of unsaturated complexes of iron. Synthetic complexes with biorelevant sulfur donors are preferred, but so little is known about unsaturated iron–sulfur clusters, and about Fe–N2 complexes with reactive N2 , that any unsaturated Fe–N2 and Fe–S complexes increase our comprehension significantly. The specific experimental procedures described in this chapter use bulky spectator ligands (unreactive ligands that modify the environment of the metal) to enforce the presence of open coordination sites on iron complexes, enabling the synthesis of unsaturated complexes of iron(II) (23). The specific complexes described here are abbreviated as [LMe Fe]2 S, [LMe Fe]2 N2 , and K2 [LMe Fe]2 N2 . In this terminology, LMe represents the anionic

Techniques for Functional and Structural Modeling of Nitrogenase

253

Fig. 17.3. The bulky β-diketiminate ligand used in the synthetic studies described here.

bidentate chelating ligand 2,4-bis(2,6-diisopropylphenylimido)3-pentyl (Fig. 17.3). These unsaturated complexes incorporate sulfide or partially reduced N2 , both key components of FeMoco activity. The syntheses described below have been reported previously in the peer-reviewed literature (24, 25). Nitrogenases are extremely air sensitive, and the FeMoco is sensitive to excess moisture when extracted from its protein scaffold (11, 26). Therefore, it is not surprising that small-molecule complexes containing similar unsaturated iron functionalities also decompose rapidly in the presence of air or moisture. Therefore, handling the compounds described below requires air-free techniques. Such techniques have been extensively developed in the context of organometallic chemistry, where there is a long history of synthesizing compounds that are highly oxygen and moisture sensitive. Excellent guidance on the basics of handling air-sensitive materials is available (27). The compounds described below are thermally stable in solution and in the solid state, if protected from O2 , moisture, or other reactive impurities. However, strict adherence to the guidelines is essential. Conditions used in the synthesis of these particular molecules are described below; they are similar to the methods used in numerous other laboratories that study unsaturated iron complexes, iron–N2 complexes, and iron–sulfur clusters (7, 8, 28, 29).

2. Materials 2.1. Air-Sensitive Techniques

1. All manipulations take place either in flasks attached to a glass double manifold (“Schlenk line” with parallel vacuum and inert gas manifolds) or in an inert atmosphere glove box. An atmosphere of purified N2 or argon is used; though argon is advantageous because it diffuses more slowly, the greater price of argon typically mandates the use of N2 atmosphere. In our work, the levels of O2 and H2 O in the atmosphere of the glove box are continuously monitored with

254

Holland

a Thermolyne 3190 O2 sensor and kept below 1 ppm O2 through the use of a recirculating system that passes the internal atmosphere through a “catalyst” consisting of solid pellets of water scavenger and O2 scavenger. Because of the moisture sensitivity of the compounds, wood, paper, or other objects that contain moisture are not allowed inside the glove box. 2. Common solvents for the synthetic iron compounds are alkanes, diethyl ether, tetrahydrofuran (THF), benzene, toluene, and acetonitrile (though the latter typically reacts with unsaturated complexes). Solvents and all other chemicals are purchased as HPLC grade and rigorously dried by treating with sodium metal, passing through activated alumina, or storing over 4 Å molecular sieves (30). Before use, solvents are always degassed (by bubbling N2 for 10 min, freezing and pumping three times, and/or distilling under N2 ) to remove any dissolved O2 . 3. Glassware (vials, Pasteur pipettes, flasks, and glass fritted filters) is dried by heating to at least 150◦ C overnight and brought into the dry glove box while still hot. 4. Celite (brand name of diatomaceous earth used as a filter aid) and graphite are heated to 200◦ C under vacuum overnight before use to remove traces of moisture. 5. Plasticware is subjected to vacuum for 15 min, then allowed to stand in the glove box for at least 1 day to off-gas any entrained air. 6. For especially sensitive compounds, such as K2 [LMe Fe]2 N2 described below, the most careful spectroscopic studies have used “silylated glass” which comes from treating glass with an aerosol of Me2 SiCl2 followed by baking at 150◦ C (31). This treatment deactivates any residual Si–OH groups on the surface of the glass that could react with the iron-containing compounds. 7. The vacuum pump attached to the manifold or the antechamber for the glove box should be able to reach below 100 mtorr, ideally 10 mtorr; this prevents trace O2 contamination. 8. Because the atmosphere inside the glove box is never completely devoid of trace O2 and H2 O, samples of the iron complexes (especially as solutions) should not be left standing at room temperature for longer than necessary. Solutions and solids are typically stored in screwcapped 20 mL scintillation vials at –40◦ C in internal freezers in the glove box. 9. Appropriate glove boxes are sold by M. Braun GmbH and by Vacuum Atmospheres Co.

Techniques for Functional and Structural Modeling of Nitrogenase

2.2. Purification and Characterization

255

1. Though organic molecules and biomolecules are often purified by chromatographic techniques, the organometallic molecules described here decompose rapidly on silica, alumina, or other stationary phases typically encountered in chromatography. Therefore, chromatography is generally ineffective. 2. The most common purification technique for metal-organic synthetic compounds is crystallization, either by cooling a concentrated solution or by vapor diffusion of a non-solvent (pentane, diethyl ether) into a concentrated solution. X-ray quality single crystals are not essential for purification, nor does crystallization necessarily guarantee purity (because of the potential co-crystallization of impurities). 3. Characterization of purity is often troublesome because biomimetic iron complexes are high spin and therefore paramagnetic. For paramagnetic molecules, nuclear magnetic resonance (NMR) spectra are often broad and highly shifted, and thus are not as conclusive as NMR spectra of diamagnetic (fully spin-paired) compounds like organic molecules. Therefore, it is important to use a range of complementary techniques that together support a proposed assignment. 4. The most important technique is X-ray crystallography, which in favorable cases gives both the constitution and the metrical details of a molecule (32). One should beware that X-ray crystal structures use only one crystal (typically 0.2 mm on a side) that may not be representative of the bulk sample. To verify bulk purity, the crystallographically determined structure should also be consistent with NMR, electron paramagnetic resonance (EPR or ESR), infrared (IR), and/or Raman spectra. 5. Airtight cells are necessary for spectroscopy. For IR spectra, KBr pellets are analyzed while inside an airtight container with NaCl windows. NMR, Raman, and ESR spectra use solutions prepared in the glove box and brought out inside glass or quartz tubes with Teflon caps. 6. In iron coordination compounds, another key technique is Mössbauer spectroscopy, which gives information on 57 Fe nuclei which are either present at natural-abundance levels of the 57 Fe isotope (about 2% of iron is iron-57) or at higher levels through purposeful isotopic labeling (33). Samples that are ∼1 mM in 57 Fe are routinely analyzed by labs having the appropriate equipment for measurement (cryostat and 57 Co γ-ray source). However, the air sensitivity of the synthetic compounds introduces significant challenges to spectroscopists, and these have been overcome by use of airtight sample cells. One particularly effective design uses

256

Holland

Delrin cups fitted with brass caps that shrink upon cooling to 77 K, providing an extremely tight seal (34). 7. Other useful techniques are microanalysis (giving percentage C, H, and N content) and the bulk magnetic moment (typically measured in solution using a method developed by Evans (35)). Excellent books giving details on these physical techniques are available (36). 2.3. Monitoring Reactions Using Synthetic Complexes

1. There exist several methods for evaluating the outcome of reactions using synthetic complexes. One of the most common is 1 H NMR spectroscopy, in which a solution of reactants in 0.4 mL of a deuterated solvent is monitored over time. It is most common to use NMR tubes with a Teflon screw cap (“J. Young” adapters), because the tubes are airtight yet can be reopened to add additional reagents or remove a product inside the glove box for further manipulation. 2. A common manipulation is to heat solutions in J. Young NMR tubes in a silicone oil bath or even heat them inside the NMR spectrometer. At a fixed temperature, one can record NMR spectra at fixed time intervals and integrate resonances in the starting material or product relative to an unreactive “internal standard.” These integrations are proportional to concentration, and thus can be plotted against the time course and fitted in order to extract rate constants. 3. Precautions must be taken with J. Young NMR tubes. The Teflon screw caps have a different thermal expansion coefficient than glass, and therefore the tubes leak if the screw cap is cooled. Another situation that arises is that many biorelevant synthetic compounds are paramagnetic and give broad and/or shifted NMR spectra. This is not a serious problem if the peaks do not significantly overlap, and if wide integration regions are used. In these cases, it is important to use an integration standard that is also paramagnetic (in our laboratory, we often use commercially available cobaltocene, Cp2 Co). To avoid reaction between the internal standard and the compounds under investigation, the internal standard is sealed within a 1–2 mm glass capillary. In unfavorable cases (especially S = 1/2 and S = 5/2 compounds), the NMR spectra are too broad to observe easily, and no distinct peaks may be seen (“NMR silent”). Because of the potential presence of NMR-silent species, it is wise to use a capillary integration standard in all reactions, in order to gauge whether the observed peaks truly account for all of the metal-containing complexes. 4. UV–visible spectroscopy can also be used for monitoring reactions; while this technique has the disadvantage of

Techniques for Functional and Structural Modeling of Nitrogenase

257

lower information content, it typically has a better signalto-noise ratio, and almost all biorelevant iron complexes have intense visible transitions. It is possible to purchase quartz cuvettes with tight screw caps (e.g., Spectrocell, Inc.) for use with air-sensitive samples. With both glass and quartz, careful drying in an oven (and possibly silylation) is important for avoiding reaction with traces of adsorbed moisture. 2.4. Specific Starting Materials Used Below

1. [LMe Li]x comes from air-stable LMe H, which is synthesized by the literature method (37), which follows. White solid LMe H is brought into the glove box, dissolved in pentane, cooled to 0 ◦ C, and 1.0 molar equivalent of n-butyllithium is added slowly to give [LMe Li]x . This deprotonation reaction is fairly exothermic and produces gas, so slow addition and cooling are important. The product [LMe Li]x should be crystallized by cooling a solution in pentane or hexanes and collecting the solid. 2. The iron starting material FeCl2 (THF)1.5 (THF = tetrahydrofuran) is synthesized by the method of Kern (38), which follows. This synthesis starts from anhydrous FeCl2 , which is commercially available, but is more pure when synthesized using a literature method (39). Using air-free techniques, FeCl2 is heated to reflux overnight in THF. Cooling the solution to room temperature and removing volatile materials in vacuo yields white (see Note 1) FeCl2 (THF)1.5 . 3. KC8 is prepared by heating a mixture of elemental potassium and graphite (no solvent) at 150◦ C under an argon atmosphere to give a bronze-colored powder. 4. Elemental sulfur (S8 ) is purified by cooling a concentrated solution in dry diethyl ether in the glove box, to give yellow crystals.

2.5. Abbreviations Used in Text Spectroscopic Descriptions

1. In the descriptions below, all peaks in the 1 H NMR spectra are referenced to residual protiated solvent. All peaks are singlets, because J coupling is masked by the line broadening from the paramagnetic metal center. In parentheses are listed peak integrations and assignments (for some complexes, not all signals have been assigned). 2. UV–vis spectra are listed with each peak given as its wavelength λ in nm (in parentheses is the extinction coefficient of the given band ε in mM–1 cm–1 ). 3. Elemental analyses (Elem. Anal.) were determined by Desert Analytics, Tucson, AZ. More recently, we have gotten significantly better results at the CENTC Elemental Analysis Facility in the Department of Chemistry of the University of Rochester.

258

Holland

3. Methods 3.1. [LMe Fe]2 Cl2 ·2LiCl

1. Within the glove box, a magnetic stir bar and [LMe Li]x (2.38 g, 5.60 mmol) are placed in a thick-walled tube with a resealable Teflon stopper (see Note 2). 2. Toluene (40 mL) is added to the flask at room temperature, and the solid dissolves with stirring to give a pale yellow solution. 3. White solid FeCl2 (THF)1.5 (see Note 1) (1.32 g, 5.72 mmol) is added to the solution in several portions. An immediate color change from pale yellow to orange is observed during the FeCl2 (THF)1.5 addition. 4. The flask is then closed, brought out of the glove box, and heated in an oil bath to 100◦ C with stirring. 5. After 18 h at 100◦ C, the red suspension is allowed to cool and is evacuated to dryness on the vacuum pump to give an orange solid (see Note 3). 6. The orange solid is dried under vacuum for an additional 30–60 min at 40◦ C. 7. The flask is then brought back into the glove box, and pentane (10 mL) is added. After stirring, the mixture is allowed to settle and the yellow supernatant is removed with a pipette. The pentane “wash” is repeated twice more (this step removes traces of organic products). 8. Toluene (20 mL) is added, the flask is closed, brought out of the glove box, stirred at 80◦ C overnight, and then pumped on vacuum. This completes the release of trace tetrahydrofuran from the LiCl byproduct. The tube is reintroduced to the glove box. 9. The orange powder obtained after drying this material, [LMe Fe]2 Cl2 ·2LiCl, is normally recovered in ≥95% yield. The LiCl is not separated from the iron compound at this stage, because later steps typically involve pentane extractions, and the LiCl can be filtered off easily at that time. 10. This yellow solid is very air sensitive and must be stored under nitrogen or argon. 11. Purity is determined by 1 H NMR spectroscopy (400 MHz, tetrahydrofuran-d8 ): δ 17.6 (4H), 4.9 (12H), 3.3 (4H), –6.8 (12H), –38.3 (2H), –67.9 (6H), –78.9 (1H) ppm. 12. Molecular weight including LiCl = 1103.

3.2. [LMe Fe]2 N2

1. Inside the glove box, a resealable tube (see Note 2), pentane (60 mL), and KC8 (see Note 4; especially notice the hazards

Techniques for Functional and Structural Modeling of Nitrogenase

259

Fig. 17.4. Structure of [LMe Fe]2 N2 . Ar represents the 2,6-diisopropylphenyl groups shown in Fig. 17.3.

of this material) (530 mg, 3.92 mmol) are added to yellow [LMe Fe]2 Cl2 ·2LiCl (1.80 g, 1.63 mmol). 2. The reaction mixture is stirred overnight at room temperature, and it slowly turns deep red with the formation of a black solid (graphite). 3. After settling for 2 h, the supernatant is filtered through Celite to give a deep red solution. 4. The solution is concentrated to 5–10 mL and stored at – 35◦ C. A very dark red solid is obtained (1.10 g, 69% yield). The structure of the product is shown in Fig. 17.4; single crystals were shown to contain one molecule of pentane per iron dimer. 5. Purity is determined by 1 H NMR spectroscopy (400 MHz, cyclohexane-d12 ): δ 53 (12H, CH3 ), –19 (24H, CH(CH3 )2 ), –20 (8H, m-H), –98 (4H, p-H), –112 (24H, CH(CH3 )2 ), –123 (8H, CH(CH3 )2 ), –280 (2H, backbone C-H) ppm. 6. UV–vis (pentane): 911 (9.1), 499 (20). 7. Elem. Anal. calculated for C58 H82 N6 Fe2 ·C5 H12 : C, 72.26%; H, 9.04%; N, 8.03%. Found C, 72.17%; H, 8.71%; N 8.07%. 8. Molecular weight = 975. 3.3. K2 [LMe Fe]2 N2

1. Inside the glove box, a resealable tube (see Note 2) is charged with [LMe Fe]2 Cl2 (2.50 g, 2.27 mmol) and pentane (25 mL). 2. While stirring the mixture, solid KC8 (see Note 4) (1.39 g, 10.3 mmol) is added to the yellow slurry, leading to the formation of a very dark blue mixture. 3. The reaction mixture is stirred overnight at room temperature and filtered through Celite to give a dark blue solution. (The filtrate contains some KC8 and should be handled very carefully; see Note 4.) 4. The solution is concentrated to 15 mL on the vacuum and cooled to –35◦ C to give dark blue crystals (1.21 g, 51% yield). The structure of the product is shown in Fig. 17.5.

260

Holland

Fig. 17.5. Structure of K2 [LMe Fe]2 N2 . Ar represents the 2,6-diisopropylphenyl groups shown in Fig. 17.3.

5. Purity is determined by 1 H NMR spectroscopy (400 MHz, benzene-d6 ): δ 25 (8H, m-H), 3 (12H, CH3 ), 1 (24H, CH(CH3 )2 ), –1 (24H, CH(CH3 )2 ), –38 (4H, p-H), –82 (2H, backbone C-H), –305 (8H, CH(CH3 )2 ) ppm. 6. UV–vis (pentane): 722 (11.4), 486 (7.7), 449 (8.5), 369 (17). 7. Elem. Anal. calculated for C58 H82 N6 Fe2 K2 : C, 66.14%; H, 7.85%; N, 7.98%. Found: C, 66.19%; H, 7.86%; N, 6.48%. 8. Molecular weight = 1053. 3.4. [LMe Fe]2 S

1. Inside the glovebox, [LMe Fe]2 N2 (600 mg, 0.615 mmol) is dissolved in diethyl ether (3 mL) at room temperature with stirring. 2. A solution of elemental sulfur (18.7 mg, 0.583 mmol) in tetrahydrofuran (2 mL) is added dropwise. 3. After 2 h of stirring at room temperature, the solution is filtered through Celite and concentrated to 2 mL under vacuum. 4. The solution is cooled to –40◦ C for 1 day to give deep red crystals (468 mg, 82% yield). The structure of the product is shown in Fig. 17.6. 5. Purity is determined by 1 H NMR spectroscopy (400 MHz, benzene-d6 ): δ 25.9 (2H, backbone C-H), 20.8 (12H, CH3 ), 4.7 (8H, m-H), –2.1 (24H, CH(CH3 )2 ), –10.4 (4H, p-H), –12.7 (8H, CH(CH3 )2 ), –21.5 (24H, CH(CH3 )2 ) ppm. 6. UV–vis (toluene): 516 (2.6). 7. Elem. Anal. calculated for C58 H82 N4 SFe2 : C, 71.15%; H, 8.44%; N, 5.72%. Found: C, 70.67%; H, 8.45%; N, 6.24%. 8. Molecular weight = 979.

Fig. 17.6. Structure of [LMe Fe]2 S. Ar represents the 2,6-diisopropylphenyl groups shown in Fig. 17.3.

Techniques for Functional and Structural Modeling of Nitrogenase

261

4. Notes 1. When synthesized from commercial FeCl2 , this material often has a beige tint, and may give less reproducible results. 2. Flasks of this type are very convenient for air-sensitive chemistry, and they consist of a glass tube roughly 5 cm in diameter closed at one side and having an 8 mm Teflon pin on the other side. The Teflon pin seals below a sidearm that goes to a 14/20 ground glass joint. The ground glass joint may be attached to a vacuum line for conveniently introducing gases or evacuating the flask. Appropriate glassware is sold by ChemGlass, catalog #AF-0523 (January 2010). 3. Sometimes the reaction becomes dark green at this point, which indicates iron(III) formation from traces of air. In favorable cases, it is still possible to obtain the product (in lower yield) by concentration of the solution and collection of the yellow precipitate, followed by crystallization. 4. Solid KC8 behaves like potassium metal, but because it is finely powdered, it reacts more rapidly than potassium chunks. Warning: this solid is highly pyrophoric and bursts into flame if exposed to air. It should be used only inside a glove box, and traces on flasks should be disposed with extreme care by mixing with hexanes, then adding acetone very slowly in a hood.

Acknowledgments Research on nitrogenase-relevant compounds in the Holland laboratory is funded by the National Institutes of Health (grant GM065313). Most of all, I thank the hardworking students and postdoctoral researchers at Rochester who have braved the difficulties of handling these sensitive yet interesting iron complexes. I also thank our spectroscopic collaborators, who have often struggled with issues of air and moisture sensitivity alongside us. In particular, Eckard Münck (and his students Hanspeter Andres and Sebastian Stoian) and Eckhard Bill (and his technician Bernd Mienert) designed and used cells that enabled Mössbauer analysis of these compounds. References 1. Ibers JA, Holm RH (1980) Modeling Coordination Sites in Metallobiomolecules. Science 209:223–235

2. Holm RH, Solomon EI (2004) Preface: Biomimetic Inorganic Chemistry. Chem Rev 104:347–348

262

Holland

3. Holland PL (2004) Nitrogen Fixation. In: McCleverty J, Meyer TJ (eds) Comprehensive Coordination Chemistry II, pp. 569– 599. Elsevier, Oxford 4. Watt GD, Reddy KRN (1994) Formation of an all Ferrous Fe4 S4 Cluster in the Iron Protein Component of Azotobacter vinelandii Nitrogenase. J Inorg Biochem 53:281–294 5. Yoo SJ, Angove HC, Burgess BK et al (1999) Mössbauer and Integer-Spin EPR Studies and Spin-Coupling Analysis of the [4Fe-4S]0 Cluster of the Fe Protein from Azotobacter vinelandii Nitrogenase. J Am Chem Soc 121:2534–2545 6. Scott TA, Berlinguette CP, Holm RH et al (2005) Initial Synthesis and Structure of an All-Ferrous Analogue of the Fully Reduced [Fe4 S4 ]0 Cluster of the Nitrogenase Iron Protein. Proc Natl Acad Sci USA 102: 9741–9744 7. Deng L, Holm RH (2008) Stabilization of Fully Reduced Iron-Sulfur Clusters by Carbene Ligation: The [Fen Sn ]0 Oxidation Levels (n = 4, 8). J Am Chem Soc 130: 9878–9886 8. Howard JB, Rees DC (1996) Structural Basis of Biological Nitrogen Fixation. Chem Rev 96:2965–2982 9. Zhang Y, Zuo JL, Zhou HC, Holm RH (2002) Rearrangement of Symmetrical Dicubane Clusters into Topological Analogues of the P Cluster of Nitrogenase: Nature’s Choice? J Am Chem Soc 124:14292–14293 10. Ohki Y, Sunada Y, Honda M et al (2003) Synthesis of the P-Cluster Inorganic Core of Nitrogenases. J Am Chem Soc 125: 4052–4053 11. Burgess BK, Lowe DJ (1996) Mechanism of Molybdenum Nitrogenase. Chem Rev 96:2983–3011 12. Eady RR (1996) Structure-Function Relationships of Alternative Nitrogenases. Chem Rev 96:3013–3030 13. Krahn E, Weiss BJR, Kröckel M et al (2002) The Fe-Only Nitrogenase from Rhodobacter capsulatus: Identification of the Cofactor, an Unusual, High-Nuclearity Iron-Sulfur Cluster, by Fe K-Edge EXAFS and Iron-57 Mössbauer Spectroscopy. J Biol Inorg Chem 7: 37–45 14. Einsle O, Tezcan FA, Andrade SLA et al (2002) Nitrogenase MoFe-Protein at 1.16 Å resolution: A Central Ligand in the FeMoCofactor. Science 297:1696–1700 15. Xiao Y, Fisher K, Smith MC et al (2006) How Nitrogenase Shakes - Initial Information About P-Cluster and FeMo-Cofactor Normal Modes from Nuclear Resonance

16.

17.

18.

19.

20.

21. 22. 23. 24.

25.

26. 27. 28.

29.

Vibrational Spectroscopy (NRVS). J Am Chem Soc 128:7608–7612 Lovell T, Liu T, Case DA et al (2003) Structural, Spectroscopic, and Redox Consequences of a Central Ligand in the FeMoco of Nitrogenase: A Density Functional Theoretical Study. J Am Chem Soc 125: 8377–8383 Lukoyanov D, Pelmenschikov V, Maeser N et al (2007) Testing if the Interstitial Atom, X, of the Nitrogenase Molybdenum-Iron Cofactor Is N or C: ENDOR, ESEEM, and DFT Studies of the S = 3/2 Resting State in Multiple Environments. Inorg Chem 46:11437–11449 Lee SC, Holm RH (2003) Speculative Synthetic Chemistry and the Nitrogenase Problem. Proc Natl Acad Sci USA 100: 3595–3600 Corbett MC, Hu Y, Fay AW et al (2006) Structural Insights into a Protein-Bound Iron-Molybdenum Cofactor Precursor. Proc Natl Acad Sci USA 103:1238–1243 Stack TDP, Holm RH (1988) SubsiteDifferentiated Analogs of Biological [4Fe-4S]2+ Clusters: Synthesis, Solution and Solid-State Structures, and Subsite-Specific Reactions. J Am Chem Soc 110:2484–2494 Thorneley RNF, Lowe DJ (1985) Kinetics and Mechanism of the Nitrogenase Enzyme System. Met Ions Biol 7:221–284 Seefeldt LC, Dance IG, Dean DR (2004) Substrate Interactions with Nitrogenase: Fe Versus Mo. Biochemistry 43:1401–1409 Holland PL (2008) Electronic Structure and Reactivity of Three-Coordinate Iron Complexes. Acc Chem Res 41:905–914 Smith JM, Sadique AR, Cundari TR et al (2006) Studies of Low-Coordinate Iron Dinitrogen Complexes. J Am Chem Soc 128:756–769 Vela J, Stoian S, Flaschenriem CJ et al (2004) A Sulfido-Bridged Diiron(II) Compound and Its Reactions with Nitrogenase-Relevant Substrates. J Am Chem Soc 126:4522–4523 Shah VK, Brill WJ (1977) Isolation of an Iron-Molybdenum Cofactor from Nitrogenase. Proc Natl Acad Sci USA 74:3249–3253 Shriver DF, Drezdzon MA (1986) The Manipulation of Air Sensitive Compounds. Wiley, New York, NY Peters JC, Mehn MP (2006) Bioorganometallic Approaches to Nitrogen Fixation Chemistry. In: Tolman WB (ed) Activation of Small Molecules, pp. 81–119. VCH, Weinheim Coucouvanis D (1991) Use of Preassembled Iron/Sulfur and Iron/Molybdenum/Sulfur Clusters in the Stepwise Synthesis of

Techniques for Functional and Structural Modeling of Nitrogenase

30.

31.

32. 33.

34.

Potential Analogs for the Fe/Mo/S Site in Nitrogenase. Acc Chem Res 24:1–8 Burfield DR, Lee K-H, Smithers RH (1977) Desiccant Efficiency in Solvent Drying. A Reappraisal by Application of a Novel Method for Solvent Water Assay. J Org Chem 42:3060–3065 Seed B (2000) Silanizing Glassware. Current Protocols in Cell Biology, Appendix 3E, A.3E.1–A.3E.2. See also http://www. currentprotocols.com/protocol/cba03e. Accessed 15 February 2010 Massa W (2000) Crystal Structure Determination, 2nd edn. Springer, New York, NY Münck E (2000) Aspects of Iron-57 Mossbauer Spectroscopy. In: Que L (ed) Physical Methods in Bioinorganic Chemistry, pp. 287–319. University Science Books, New York, NY Andres H, Bominaar E, Smith JM et al (2002) Planar Three-Coordinate High-Spin

35.

36. 37.

38. 39.

263

Fe(II) Complexes with Large Orbital Angular Momentum: Mössbauer, Electron Paramagnetic Resonance, and Electronic Structure Studies. J Am Chem Soc 124:3012–3025 Schubert EM (1992) Utilizing the Evans Method with a Superconducting NMR Spectrometer in the Undergraduate Laboratory. J Chem Educ 69:62 Drago RS (1992) Physical Methods for Chemists. Surfside, Gainesville, FL Feldman J, McLain SJ, Parthasarathy A et al (1997) Electrophilic Metal Precursors and a Beta-Diimine Ligand for Nickel(II)and Palladium(II)Catalyzed Ethylene Polymerizations. Organometallics 16: 1514–1516 Kern RJ (1962) Tetrahydrofuran Complexes of Transition Metal Chlorides. J Inorg Nucl Chem 24:1105–1109 Winter G (1973) Iron(II) Halides. Inorg Synth 14:101–104

Section V Theoretical Calculations

Chapter 18 Nitrogenase Structure and Function Relationships by Density Functional Theory Travis V. Harris and Robert K. Szilagyi Abstract Modern density functional theory has tremendous potential with matching popularity in metalloenzymology to reveal the unseen atomic and molecular details of structural data, spectroscopic measurements, and biochemical experiments by providing insights into unobservable structures and states, while also offering theoretical justifications for observed trends and differences. An often untapped potential of this theoretical approach is to bring together diverse experimental structural and reactivity information and allow for these to be critically evaluated at the same level. This is particularly applicable for the tantalizingly complex problem of the structure and molecular mechanism of biological nitrogen fixation. In this chapter we provide a review with extensive practical details of the compilation and evaluation of experimental data for an unbiased and systematic density functional theory analysis that can lead to remarkable new insights about the structure–function relationships of the iron–sulfur clusters of nitrogenase. Key words: Density functional theory, electronic structure, exchange functional, correlation functional, basis set saturation, spectroscopic calibration, broken symmetry calculation, ferromagnetic coupling, antiferromagnetic coupling, exchange interaction.

1. Introduction Density functional theory (DFT) builds on the premise of the Hohenberg–Kohn theorem (1) and Kohn–Sham equations (2) by providing a connection for a chemical system between its electron density distribution and its unique total energy with respect to combinations of nuclear and electron interactions. The electrostatic interactions can be described by the classical Coulomb law or by its Hartree expressions of J(ρ), where J is the Coulomb integral and ρ is the electron density (3). In practical application of DFT, the electrons are described by a set of small spatial M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_18, © Springer Science+Business Media, LLC 2011

267

268

Harris and Szilagyi

elements of the topologically complex electron density, which is obtained in a numerical integration procedure with a sufficiently fine electron density grid definition. Excluding the numerical nature of handling electron density, the Coulomb interactions are conceptually identical to those considered in wave functionbased molecular orbital (MO) methods that are often called ab initio MO methods. The dividing difference between DFT and ab initio MO methods comes from the treatment of the quantum phenomena of electron–electron exchange and correlation interactions. In ab initio MO methods, the exchange term is known exactly in the framework of orbitals and implemented by the Hartree–Fock approximation (Hartree–Fock exchange, HFX), while in DFT this is approximated by various mathematical expressions that are functions of spatially dependent electron density and termed as exchange functionals (density functional exchange, DFX). Improvement in DFX has been achieved by leading theoretical groups in going from early Xα (4) and Slater exchange (5) based local DFX functionals, through the modern Kohn–Sham first derivative-corrected semi-local DFX functionals termed generalized gradient approximations (GGA) (6) and second derivative-corrected functionals (meta-GGA) (6), to the non-local hyper-GGA (6) or random phase approximation (RPA) (6, 7) with exact exchange functionals. While the latter two are very attractive for transition metal systems, they suffer from prohibitive computational cost and from the fact that they are still only approximations. An alternative correction scheme introduced by Becke (8, 9) uses hybrid exchange functionals to mix exact non-local HFX and pure GGA or meta-GGA DFX functionals. From a coordination chemist’s point of view, this mixing is a useful tool for empirically tuning metal–ligand bonding in computational models and thus manipulation of the catalytically relevant redox-active orbitals at a modest computational cost. For example, GGA functionals, such as BP86 (10, 11) often lead to overly covalent descriptions of ground-state electronic structure, (12) and empirically derived hybrid functionals, such as B3LYP (9), can improve accuracy (see Note 1). However, no single functional is known to date that works for all transition metalcontaining systems. A method for choosing the most appropriate functional is described in Section 2.1. Due to the complex nature of the correlation interaction, there is no quick fix or exact analytical formula in either ab initio MO or DFT methods. However, the advantage of using ab initio MO methods is the availability of a theoretically converging series of post-HF or post-SCF methodologies that can systematically take into account more complete treatments of dynamic and static correlation effects by approaching the complete configuration interaction limit (full CI) and thus the exact solution of the Schrödinger equation. However, this comes with a prohibitive computational cost. Even considering

Nitrogenase Structure and Function Relationships by Density Functional Theory

269

multi-core compute servers equipped with graphical processing units and recent development in linear scaling methods, the practical applicability of higher correlation methods is limited to a few transition metal ions coordinated with highly truncated ligand models. By contrast, the density functional correlation (DFC) functionals share the acceptable cost of DFX for realistic systems. Therefore, DFT methods are currently the most practical approach for theoretical calculations involving large bioinorganic systems (13, 14) up to approximately 300 atoms. Local, semilocal (GGA and meta-GGA), and non-local (e.g., hyper-GGA and RPA) DFC functionals comprise an improving series of levels of theory; however, unlike with ab initio MO methods, a theoretically converging series for either DFC or DFX functionals is not yet known. The above-described features of ab initio MO and DFT methods are graphically illustrated in Fig. 18.1a and b, respectively. It is important to mention that there are exciting recent developments (15–17) in DFT for weak interactions, such as π–π stacking, H-bonding, van der Waals, and dispersion interactions; however, these address interactions that are at least an order of magnitude weaker than those governing the coordination chemistry of iron–sulfur clusters. A somewhat controversial issue, in particular for bioinorganic active centers with complex iron–sulfur architecture, is the treatment of magnetic interactions among iron centers with unpaired electrons. A general procedure known as the broken symmetry approximation was introduced and used to approximate the structure of these systems by Noodleman (18). This treatment is a quick fix for the limitation of HF ab initio MO or DFT calculations. The unrestricted nature of the single-reference wave

Fig. 18.1. (a) Schematic demonstration of theoretically converging series of ab initio MO methods and (b) relationship of hybrid and GGA semi-local DFT approximations.

270

Harris and Szilagyi

function or electron density is exploited by allowing for different occupation and mixing of the spin-up (α) and the spin-down (β) orbitals or densities, respectively. This will result in an electronic structure that is a solution of the HF or KS operator, but not the spin operator (19). Although Ms is well defined, the expectation value of the spin operator ( S2 ) deviates from the ideal S(S+1) as follows: Nα − Nβ N α − Nβ β αβ 2 ˆS 2 + 1 + Nβ − = niα ni Sii BS 2 2 i [1] where Nα and Nβ are the number of α and β electrons, respectively, ni α and ni β are spin-orbital populations, and Sii αβ is the overlap integral between the αi and βi spin-orbitals (20, 21). For example, the BS state of a system with two ferric iron centers (N α = N β = 5) bridged with sulfide ligands have S 2 BS ≈ 5 at the HF level, while S 2 = 0 for the true ground state. With the weak interactions between antiferromagnetically coupled Fe 2 centers, S αβ ii is nearly 0, and thus equation [1] reduces to the true ground state S 2 plus the number of unpaired β electrons. For a two-spin system, the true ground-state energy can be approximated using the total energy and the expectation values of the spin operator of a broken symmetry solution and the ferromagnetic spin state (22). Recently, a method has been proposed to use this projected ground state for geometry optimizations, which can result in significantly different structures by more than 0.1 Å deviations in bond lengths (23). For a polynuclear system such as FeMo-co, a more sophisticated approximation of the true ground-state based on the Heisenberg Hamiltonian can be used (24). Despite its conceptual limitations, the broken symmetry approach has already reliably described the electronic and geometric structures of [Fe–S] systems given a well-chosen functional and basis set (25). The size and quality of the basis set is an important input parameter for DFT calculations. The possible basis sets are Gaussian- (GTO) or Slater-type orbital (STO), numerical or plane wave. The latter is the preferred choice for periodic structures. Figure 18.1a illustrates on the horizontal axis of the ab initio MO chart a converging series of GTO basis sets as the splitting of the valence set and thus the overall basis set size increases. STO-based Kohn–Sham orbitals and other electronic structure descriptors are generally more chemically meaningful due to the hydrogenic radial distribution of STOs; however, they are more computationally expensive to integrate than GTOs. With respect to the quality of a given size basis set, there are some considerable variations especially when it comes to the description of catalytically important frontier molecular orbitals (FMO). For

Nitrogenase Structure and Function Relationships by Density Functional Theory

271

example, in our hands, the triple-ζ quality Ahlrich’s basis set (26, 27) with polarization functions appears to provide a more chemically reasonable description of unoccupied FMOs than the Pople’s 6–311G (28, 29) with both polarization and diffuse functions. In order to rigorously evaluate the basis set quality and how well the basis set saturation limit is approached, ab initio MO theory is of great practical help. Using a small computational model relevant to a given complex bioinorganic problem up to 20 atoms, this can be readily evaluated using a series of theoretically converging series of ab initio MO methods with increasing size of basis set to determine the smallest large basis that reasonably reproduces the available experimental observables. Using any larger basis set would only increase the computational cost without significantly improving the electronic structure description. In addition to the level of theory that includes the mathematical formalism for interactions and basis set considerations, there is another key component to electronic structure calculations and theoretical chemical modeling, which is the size and adequacy of the computational model. Truncation, simplifications, protein environmental effects, and explicit or implicit solvent interactions greatly affect computational cost and accuracy. In the following sections we will provide a detailed analysis of each of the above choices that need to be considered before a theoretical metalloenzymology study is initiated. We argue that these and only these allow for meaningful insights into the often remarkably unique electronic and geometric nature of the catalytic active component of a bioinorganic system that is useful for enzymologists, biochemists, and biomimetic chemists in generating experimentally testable hypotheses.

2. DFT Modeling of the Electronic and Geometric Structures of FeMo-co (See Notes) 2.1. Spectroscopic Calibration of DFT

Instead of looking at global electron density-based structural properties, such as molecular geometry, vibrational frequencies, or even UV–vis excitation data, Mössbauer shifts, or g-values for paramagnetic systems, we are a strong proponent of using experimental orbital compositions from the core-excited state spectroscopic technique of multi-edge X-ray absorption spectroscopy (XAS) (30, 31). Although as discussed in the introduction, there is no systematically converging series of density functionals, a fruitful method for customizing a functional is to use Becke’s (8, 9) hybrid approach that can adjust the calculated spin densities or frontier orbital compositions to those from X-ray absorption spectroscopy or EPR. This results in the most reasonable and economical mix of “pure” DFX and HFX for the studied system (12, 25). This method can be well demonstrated using

272

Harris and Szilagyi

S K-edge XAS data (32) and calculated S 3p orbital characters of metal–sulfide bonds in a prototypical model [Fe–S] cluster, [Fe4 S4 (SEt)4 ]2– . The theoretical electron donation from the sulfide and thiolate ligands can be derived from the difference between the S(3p) atomic populations of the covalent electronic structure and the theoretical ionic valence bond limit. The atomic populations can be obtained from the natural population analysis (NPA) (33) using a minimal valence set, where S 3d, Fe 4s, 4p, Mo 5p, and higher orbitals are treated as part of the Rydberg set. The combination of BP86 and SDD-VTZ(d) (34) with additional polarization functions for Fe (35), Mo (35), and S (36) gives S(3p) characters of 1.19 and 0.40 e – for the Fe-sulfide and Fe-thiolate bonds, respectively, which is a good approximation of the experimental 1.23 and 0.41 e – values. Employing a larger allelectron basis set (6–311+G(d)) and a separately experimentally validated (25) hybrid functional (B(5%HF)P86 (37)) only moderately improves this agreement, with S(3p) characters of 1.22 and 0.42 e – . Taking into account the little improvement in accuracy using B(5%HF)P86/6–311+G(d) at the expense of increase in computational time, BP86/SDD-VTZ(d) appears to be an ideal choice for performing a large number of calculations with an [Fe–S] cluster, such as FeMo-co. The effective core potential allows for a rapid structural optimization, and the electron density or total energy value for redox calculations can be improved by single point energy calculations at the more rigorously calibrated level of B(5%HF)P86/6–311+G(d). 2.2. Broken Symmetry Calculations

As discussed above, due to the presence of high-spin Fe centers embedded in weak-field ligand environments in FeMo-co, the method of treating ferro- and antiferromagnetic interactions is critical. Mössbauer (38) and electron nuclear double resonance (ENDOR) (39) measurements indicate four parallel spins and three antiparallel (4+3) to the majority spin (spin-up or α) direction. Noodleman defined 10 major 4+3 BS spin-coupling patterns (40), and including all alternative spin arrangements there are 35 non-equivalent St = 3/2 states. By constructing localized spin states with an ionic fragment approach (25) (see Notes 2 and 3), we can evaluate all 35 states along with functional and basis set dependence (see Notes 4). Figure 18.2 summarizes the relative energies of various spin-coupling schemes for a representative FeMo-co model with an oxidation state distribution of 4FeII – 3FeIII (see Note 5) and N3– (see Note 6) as the central ligand (41). The results in Fig. 18.2 display a clear trend of lower energy with increasing numbers of antiferromagnetically coupled pairs of Fe centers at four levels of theory, but there is a notable discrepancy in that the first excited state is functional dependent. With the pure DFT (GGA) functional, BP86, the first excited state is BS6, consistent with the findings of Noodleman and coworkers

Nitrogenase Structure and Function Relationships by Density Functional Theory

273

Fig. 18.2. Variation of the relative preference of broken symmetry states (as defined by Noodleman (40)) as a function of basis set and density functional calculated for the (CH3 S)[MoIV -4FeII -3FeIII -N3– ](glycolate)(imidazole) model of FeMo-co.

(42), who also used a GGA functional, PW91; however, we find that with the hybrid functional, B3LYP, the first excited state is BS10. Importantly, the ground state for each tested level of theory is BS7. This, in light of the overall trend, indicates that the primary factor affecting the stability of spin states is the number of antiferromagnetically coupled centers. A similar trend was also observed by Noodleman and coworkers using an early FeMo-co model without the interstitial atom, although BS6 was found to be the ground state (40). With X = N3– and a change in functional, they now find BS7 to be the ground state for X = N3– (42), in agreement with this work and with that of Blöchl and coworkers (43) who use a different theoretical approach. Additionally, ESEEM and ENDOR measurements indicate that if X = N or C the hyperfine coupling to X is weak, which, as shown by DFT, favors a spin state with symmetric spins in the central “6Fe belt,” such as BS7 (42). We emphasize that while the lowest energy BS state was the same regardless of the employed level of theory, there can be significant deviations in the excited BS states as a function of hybrid or GGA functional. This may lead to erroneous assignment and interpretation of experimental spectra in the case of ground states with degeneracy. 2.3. Computational Model

Assuming the chosen functional, basis set and method of handling magnetic interactions is adequate for describing [Fe–S] cluster chemistry, the results will still be inaccurate if the

274

Harris and Szilagyi

computational model is too small or has incorrect composition. This is particularly relevant for FeMo-co, since the presence of a central ligand was confirmed (44) about a decade after obtaining the first crystal structure (45). As an example of the consequences of having an unavoidably incomplete model, Lovell et al. initially found preference for the 6FeII –1FeIII oxidation state using an early FeMo-co model without the central ligand (40), while the 4FeII –3FeIII state was favored after inclusion of an anionic central ligand (46). An additional uncertainty in DFT calculations with FeMo-co is the unknown protonation state of the ligands. Truncation of ligands to save computational time can be reasonable, particularly when the covalent interactions in the cluster are preserved. For example, the uncoordinated carboxylates of homocitrate are not expected to significantly affect the Mo–O bonding, and thus homocitrate can be truncated to a glycolate moiety. Cys275 and His442 can be truncated to an alkyl thiolate and imidazole group, respectively, while critical covalent interactions will be lost with further truncation to a sulfhydryl and ammonia group. For the most accurate geometry optimizations (see Notes 7 and 8) and calculations of properties such as reduction potentials, truncations should be minimized, especially for charged groups, and the protein environment must also be considered (47–49). An efficient first-order approximation previously applied for FeMo-co (42, 50) is to mimic the dielectric environment around the cofactor with an implicit solvent implementation, such as the polarizable continuum model (PCM) (51–53). In order to introduce anisotropy, the surrounding protein environment can also be modeled by point charges (54) or by using explicit neighboring molecules treated with molecular mechanics (43, 55). However, if resources and time permit, expansion of the quantum mechanically treated portion of the model is the preferred way to increase accuracy as this is the only method that can fully capture simultaneously the environmental effects of H-bonding, dipole, electrostatic, and steric interactions of the inner and first-outer sphere residues in addition to the electronic interactions. We consider a computational model to be converged if additional outer sphere residues do not significantly affect the electronic interactions within the cluster (56). As demonstrated by Siegbahn (57), an alternative method for evaluating the adequacy of a model is to calculate a potential energy profile for a reaction using PCM with various dielectric constants. If the variations between the profiles are small, the model is large enough. This method can be generalized to use any molecular property as a reference, depending on the objectives of the study. As an example, Fig. 18.3 presents a 2D and corresponding 3D representation of the network of weak interactions around FeMoco that can affect the electronic structure of the cluster. This

Nitrogenase Structure and Function Relationships by Density Functional Theory

275

Fig. 18.3. Protein environment of FeMo-co. (a) Oval shapes represent second coordination sphere residues, circled numbers represent waters, and plain text represents further outer sphere charged residues. (b) The 3D representation includes the shaded residues from (a), truncated for use as a computational model. H atoms are omitted for clarity.

minimalist environmental model also includes residues shown experimentally to affect the FeMo-co structure and nitrogen fixation mechanism. The second sphere residues (oval shapes) and waters (circled numbers) sphere residues would need to be treated explicitly, while any other outer sphere charged residues can be approximated by point charges. The steric and electrostatic effects of the remaining protein environment can be modeled with molecular mechanics and PCM, respectively. 2.3.1. Experimental Summary of FeMo-co Structural Features

This section provides an example of the types of experimental facts and proposed interpretations that are to be used to validate a computational inquiry into the electronic and geometric structures of FeMo-co.

276

Harris and Szilagyi

2.3.1.1. Structure

FeMo-co has been identified as an [Mo-7Fe-9S-X] cluster from ´ crystal structure of the MoFe prothe atomic resolution (1.16 Å) tein, where X is an unknown interstitial atom with likely eight electrons (44). From a coordination chemistry perspective, it is remarkable that this 18-atom cluster is linked to the protein matrix by only the Cys275 and His442 residues. An additional non-protein ligand of homocitrate completes the hexacoordinate environment of the Mo center.

2.3.1.2. Charge State and Redox Forms

The overall charge of FeMo-co is an important factor in substrate binding, product release, and proton/electron transfer events, and it varies with the redox form, affecting interactions with substrates. The MoFe protein can be isolated with three forms of FeMo-co: resting or native (FeMo-coN ), one-electron reduced (FeMo-coRed ), and one-electron oxidized (FeMo-coOx ) (58) (see Note 9). The reduction potential for the FeMo-coOx /FeMocoN couple has been determined to be –40 mV, while the reduction step for FeMo-coN /FeMo-coRed has only been estimated to be –465 mV (58). Although the charge of protein-bound FeMo-co has not yet been directly measured, a cation-binding study indicated that the oxidized form of the extracted cofactor (exFeMo-coOx ) most likely has an overall charge of –2 (59). When FeMo-co is extracted into NMF, the homocitrate ligand remains coordinated to the Mo site, an NMF– anion replaces the deprotonated Cys275, and a neutral NMF molecule replaces nonprotonated His442 (60). Thus, if the protonation state of the homocitrate ligand remains unchanged, the charge is conserved, and a –2 charge for exFeMo-coOx corresponds to –2 for FeMocoOx and the –3 charged state for FeMo-coN .

2.3.1.3. Composition

To date, no direct experiments have been able to unambiguously distinguish between the carbido (C4– ), nitrido (N3– ), and oxo (O2– ) groups as the interstitial ligand in FeMo-co. ENDOR and ESEEM studies of 14/15 N and 12/13 C isotopomers of bound and extracted FeMo-co did not reveal signals for either possibility, indicating that if X = N or C, it must be magnetically decoupled from the FeMo-co spin system (42). Alternative electronically (8 e– ) and sterically (compact cations) reasonable assignments, such as B5– , F– , Na+ , or Mg2+ , can likely be excluded due to charge incompatibility or limited biochemical availability (B5– and F– ).

2.3.1.4. Oxidation States

XAS (61, 62) measurements suggest the Mo center is redox inactive with a formally +IV/d2 oxidation state in all three redox forms, and ENDOR (63) data support this assignment for the resting form. Interpretation of 57 Fe ENDOR experiments (64) led to the suggestion of the 6FeII –1FeIII assignment for CObound FeMo-co, while Mössbauer isomer shifts (38) were more

Nitrogenase Structure and Function Relationships by Density Functional Theory

277

consistent with the 4FeII –3FeIII state for the resting form. Independent electrochemical studies provided a link between these two assignments by suggesting that the CO-bound form is most likely 2e– reduced relative to FeMo-coN (60). 2.3.1.5. Spin States

FeMo-coN shows a characteristic total spin (St ) of 3/2 EPR signal, while the reduced form is EPR silent with integer spin St ≥ 1 (58). The oxidized form is diamagnetic (65–67) as observed at 4 K, but a low-lying excited St ≥ 1 state is populated between 240 and 320 K (68). An St = 1/2 state has been observed for a turnover intermediate in the presence or absence of N2 that has accumulated 4e– and presumably the same number of protons (69). Other unique St = 1/2 EPR signals have been observed for the CO-bound state (64), and for MoFe protein variants under turnover of protons, alkyne substrates, and nitrogenous substrates (70). The total spin values have important implications regarding the possible oxidation states of the Fe centers. Assuming weakfield ligand environments, the formally FeIII and FeII centers can only adopt ms = ±5/2 and ±2 states, respectively. The resting state St = 3/2 is a result of ferro- and antiferromagnetic coupling between the Fe centers (or [2Fe-2S] rhombs), of which four have parallel and three have antiparallel (4+3) spin orientations according to the interpretations of ENDOR (39) and Mössbauer (38) data. A formal electron counting analysis of FeMo-coN shows that the only possible oxidation states for iron are 6FeII –1FeIII , 4FeII – 3FeIII , and 2FeII –5FeIII with the above assumptions.

2.3.2. Coordination Chemical Computational Models

A variety of functionals, basis sets, and computational models have been used in studies aimed at identifying the unknown structure of FeMo-co. Vrajmasu et al. (71) used B3LYP with a valence triple-ζ GTO basis set plus polarization and diffuse functions to calculate Mössbauer isomer shifts for all possible Fe oxidation states and X = C4– , N3– , and O2– in a ferromagnetic [Fe8 -S9 -X-Cl2 ] model. They found that only the 4FeII – 3FeIII –N3– , 4FeII –3FeIII –C4– , and 2FeII –5FeIII –N3– oxidation state assignments are consistent with experiments. Huniar et al. (72) used a similar GTO basis set, but with BP86 and a larger model including the Mo center coordinated by imidazole and a deprotonated ethylene glycol; the 4FeII –3FeIII –N2– state was found to be slightly favored over the 4FeII –3FeIII –O2– based on geometric parameters, while C4– was not considered. Lovell et al. (46) used VWN+BP86 and a valence triple-ζ STO basis set plus polarization to study a large model including the full homocitrate with four H2 O molecules and the αGln191 side chain. Mössbauer isomer shifts favored the 4FeII –3FeIII oxidation state over 6FeII –1FeIII , and structural optimizations favored 4FeII –3FeIII with either N3– or C4– while disfavoring O2– . The calculated reduction potentials indicated that 4FeII –3FeIII –N3– was the most likely oxidation state and composition. Notably, the

278

Harris and Szilagyi

2FeII –5FeIII –N3– state was not evaluated. By comparing HOMO energies of FeMo-co and Fe–S clusters, Dance (73, 74) applied BLYP functionals and a numerical basis set to a medium-sized model with homocitrate truncated to glycolate and found preference for 4FeII –3FeIII –N3– , or 4FeII –3FeIII –C4– with a protonated bridging central sulfide or Fe. Comparison of the three lowest energy spin states showed that 2FeII –5FeIII –N3– had an St = 3/2 ground state, consistent with experiment, while 4FeII – 3FeIII –N3– had an St = 5/2 ground state with the St = 3/2 state about 20 kJ/mol higher. By contrast, using non-collinear spin calculations with plane waves, PBE functional, and a similar model, Schimpl et al. (43) reported St = 3/2 to be the ground state for 4FeII –3FeIII –N3– , while the ground states for 2FeII –5FeIII – N3– and 6FeII –1FeIII –N3– were St = 1/2 and St = 0.24, respectively. Geometry optimizations showed 4FeII –3FeIII –N3– giving the best agreement to experiment. In another plane wave basis set study with RPBE (75), Hinnemann and Nørskov calculated the binding energies of the interstitial atom possibilities using a minimal model with SH, NH3 , and 2(OH) for Cys, His, and homocitrate, respectively. In their models, H atoms were added to the central bridging sulfides according to the charge of the interstitial atom. They found the model with C4– to be unstable, while the structure with O2– was too expanded, thus favoring N3– . In summary, studies using different functionals, various basis sets, and a plethora of computational models do not always offer the same results, but most works identified the 4FeII –3FeIII –N3– state as one of the preferred states. Still, other states and compositions have not been unequivocally ruled out. In our recent work (24), we carried out a comprehensive evaluation of experimental and calculated geometric and electronic structural features, and we found that only the [MoIV –2FeII –5FeIII –9S2– –C4– ] composition with a hydroxyl-protonated homocitrate ligand satisfies all criteria considered.

3. Modeling of the Potential Energy Surface of Biological Nitrogen Fixation (See Notes) 3.1. Site-Directed Mutagenesis Studies

Studies of MoFe protein variants with substitutions in the vicinity of FeMo-co offer invaluable insights into the role of the protein environment. These perturbations are essential in validating the reliability of computational proposals for a given molecular mechanism. For example, substituting αVal70 with the smaller residues Ala or Gly allowed the binding and reduction of larger substrates such as propargyl alcohol (HOCH3 C≡CH), while the αIle70 variant was only capable of proton reduction (76). These results implicate the nearby 4Fe face composed of Fe2, 3, 6, and 7 as a likely binding site, which is sterically affected by the α70 side

Nitrogenase Structure and Function Relationships by Density Functional Theory

279

chain. Variants with a substitution of αArg96, also located near the same 4Fe face, bind acetylene and cyanide in the resting state, unlike the wild-type protein, indicating that αArg96 may have a gate-keeping role (77). Substitution of another residue in the same vicinity, αHis195, slows or eliminates N2 reduction, while allowing N2 to bind, implicating this residue in proton transfer and providing further evidence for the importance of this 4Fe face and the surrounding residues (78). Furthermore, a putative substrate channel that leads to the same face has also been identified (79). The binding site for alkyne substrates has been narrowed to Fe6 by site-directed mutagenesis studies of αHis195, αVal70, and αGln191 (70); however, locating the specific binding site of N2 has proven to be more difficult. Recently, rapid freeze-quenching study of a MoFe variant with a double mutation of αHis195 and αVal70 enabled 15 N ENDOR studies of trapped intermediates of the N2 reduction pathway. Diazene (N2 H2 ) and hydrazine (N2 H4 ) were shown to be trapped in either end-on or in a symmetric bridging coordination mode at the same reduced state of FeMo-co with St = 1/2 state (80, 81). These studies lend support to the “alternating” mechanism, in which alternate N atoms are protonated in successive steps (82). A less reduced intermediate has been trapped using N2 as a substrate for the wild-type protein, and the ENDOR data were consistent with N2 -bound end-on in a terminal or bridging mode to one or more metal ions with a small spin density, such as a low-spin FeII or FeIII , FeI , or even MoIV (83). The possibility of dinitrogen binding to Mo cannot be ruled out completely, and such mechanisms are closely linked to the role of homocitrate. A strong chemical argument for the Mo site is that to date the only molecular system capable of catalytic N2 reduction contains Mo (84, 85). MoFe protein variants with citrate instead of homocitrate lose the ability to reduce N2 , but acetylene reduction is unaffected (86). Citrate only differs from homocitrate in the length of the CH2 CH2 COO– arm by removal of a methylene group. The extracted FeMo-co [exFeMoco] with either citrate or homocitrate and NMF has an increasing rate of reaction with thiophenolate as [PhS– ]/[PhSH] ratio increases. However, when imidazole replaces NMF at the Mo site of exFeMo-co the rate decreases with [PhS– ]– /[PhSH], while it is unaffected by cyanide, azide, and the proton (87). This was interpreted as evidence for H-bonding between homocitrate and imidazole that is not possible with the shorter arm of citrate. However, even with the longer arm, homocitrate cannot reach the Mo-bound imidazole of His442 in the wild-type FeMo-co (87). Based on evidence from model complex chemistry (88), a mechanism was proposed that homocitrate becomes monodentate during the mechanism, which would now allow Hbonding to His442 (87). With this mechanism, N2 could bind

280

Harris and Szilagyi

to the vacant Mo site of the highly reduced FeMo-co, while acetylene reduction would take place away from Mo. An experiment designed to test this hypothesis showed that substitution of αLys426, a residue proposed to stabilize monodentate homocitrate, affected N2 reduction without affecting acetylene reduction (89). A somewhat forgotten support for an explanation of reactivity differences of C2 H2 and N2 comes from Coucouvanis’ model chemistry, in which a [4Fe–4S] cluster with chloride ligands was capable of acetylene reduction (90), while Mo is needed to cleave the N,N single and double bonds (91). 3.1.1. Coordination Chemist’s View of the Thorneley–Lowe Kinetic Model

The Thorneley–Lowe kinetic model provides a framework for our understanding of the N2 reduction mechanism at the molecular level by defining eight intermediate states for the MoFe protein, En , where n is the number of accumulated electrons and protons. To determine the structure of the active site in each intermediate state, we must consider the most likely binding sites of the protons/hydrides that can define preference among the Fe sites, as the reduction is generally delocalized over the entire FeMo-co cluster. For example, the E4 state was recently shown by ENDOR measurements to contain two hydrides, each bound in a bridging mode between Fe centers of FeMo-co (92), with the other two protons presumably bound to nearby residues (H+ (out) ). Due to lack of spectroscopic evidence, we assume that intermediates with protonated bridging sulfides are only short-lived species. This is a departure from earlier computational models, since we anticipate that protons can migrate from the sulfides (>SH– ) to the reduced Fe centers (FeII . . .FeII ) to form hydride ligands (FeIII –H– –FeIII and >S2– ) that can aid the accumulation of additional electrons on the cluster. Figure 18.4 shows an extended Thorneley–Lowe scheme for the first four electron and proton transfer steps that were constructed with the assumption that from a given En state there are two possible paths to reach the En+1 state: (1) a nearby residue is protonated (H+ (out) ) and the FeMo-co is reduced or (2) a hydride (initially H+ (in) ) forms and the metals are oxidized. Using these steps a few remarkable observations can be made. The EPR silent E1 state can be either reduced or oxidized by one electron relative to FeMo-coN depending on whether outer sphere protonation or hydride formation occurs. Initially, there are three possibilities for E2 , but FeMo-coSupRed would be short-lived with its high pKa , and the doubly hydride-coordinated FeMo-coSupOx is unlikely because it does not have a path for the H2 evolution observed during kinetics experiments (69). The hydride-bound FeMo-coN is the most attractive candidate for the E2 state, as it would likely display the observed (69) St = 3/2 signal and be ready for H2 evolution when the hydride becomes protonated from an outer sphere source. Consequently, the E3 state, prior to possible N2 binding, must have the same oxidation

Nitrogenase Structure and Function Relationships by Density Functional Theory

281

Fig. 18.4. Modified Thorneley–Lowe scheme showing possible speciation of electrons and protons/hydrides for each state, En , where n is the number of proton-coupled electron transfers to the MoFe protein. Protons that will bind or are already bound to nearby residues are colored blue, short-lived states are in brown with brackets, and less likely states are in red with braces (in the online version only).

state assignment as E1 and also be EPR silent. For E4 the trihydride state is unlikely (not shown in Fig. 18.4), and the highpKa FeMo-coSupRed H– state is expected to gain at least an additional hydride from the outer sphere, leaving FeMo-coN with two hydrides as the only possibility. The St = 1/2 spin state likely arises from structural perturbations by the two bridging hydrides that only have sufficient space within two faces of the 6Fe belt. This is attractive since it would leave the third face available for N2 binding. Similarly, the E3 state that most likely binds N2 (FeMo-coOx (H– )2 ) has two bridging hydrides. The alternate E3 state of hydride-bound FeMo-coRed would not be able to bind N2 because the resulting structure would be equivalent to that of the E1 state, which does not interact with N2 . Remarkably, in this extended kinetic scheme N2 binds to a formally equivalent oxidation state level to FeMo-coOx or FeMo-coN states in E3 or E4 , respectively. It is likely that a distorted structure relative to the crystallographically characterized resting form makes N2 binding possible. Due to the lack of an observed exchangeable hydride in the N2 -bound intermediate (83), the initial state with a single hydride and physisorbed N2 in E3 and E4 must be short-lived, with a hydride electron reducing N2 to N2 – . The leftover proton after reduction needs to bind to a bridging sulfide or nearby residue that will not be detectable in ENDOR. At this point we have arrived at only two N2 -bound structures: FeMo-coN H+ N2 – of E3 and FeMo-coRed H+ N2 – of E4 , which also have an additional outer proton. We expect that the E3 state is the experimentally observed N2 -bound intermediate with non-integer total spin and without experimentally observable exchangeable protons, because the E4 state is diamagnetic or an integer spin system.

282

Harris and Szilagyi

In both states we do not expect the inner proton to migrate to N2 – because the resulting states would be essentially the same as E0 and E1 , which do not bind N2 . From the E3 state, an additional outer- or inner-proton-coupled electron transfer will lead to the N2 – - or N2 H-bound E4 state, respectively, from which reduction to ammonia will occur. A final consideration for the actual structure of the reduced and bound N2 (chemisorbed) is that it is likely to be bridging to prevent the formation of an N2 Hx state at the E3 or E4 level. Terminal coordination would attract spontaneous protonation and thus manifest exchangeable protons attached to a paramagnetic center that have not been detected experimentally by ENDOR measurements. 3.2. Theoretical and Computational Insights into the Nitrogenase Function

There have already been countless computational attempts to assign molecular structures and relative energies to the abovediscussed steps of the Thorneley–Lowe scheme. A comprehensive and comparative review of these works is rather challenging due to the large variations in the computational model and level of theory employed. However, we propose that despite the differences in the sometimes contradicting conclusions, all mechanistic pathways have merit for understanding what in general molecular Fe– S clusters with heterometal substitution are capable of carrying out. In light of the discussion in Sections 3.1 and 3.2, we argue that the actual molecular steps of biological nitrogen fixation will only be revealed through studies that use a computational model encompassing all major inner and outer sphere coordination effects to FeMo-co, evaluate access to electron and proton transfer pathways, and consider reactant and product channels. These calculations are yet to be performed for realistic virtual chemical models of the nitrogenase active site. An earlier review (93) summarized the computationally proposed N2 coordination and conversion to NH3 in four major pathways: (a) substitution of a protonated μ2 -sulfide, (b) substitution of the interstitial ligand, (c) coordination to a central Fe site upon μ2 -sulfide protonation, and (d) activation via protonation at the Mo site. It is not unlikely that the actual molecular mechanism involves one or more of the above pathways or their combination. For example, the most difficult step of the initial N–N bond cleavage can take place at the central Fe sites with the most atypical Fe coordination geometry and the remaining bonds can be readily cleaved at the Mo site, as has been shown to occur in [Mo–3Fe–4S] clusters (91, 94). Furthermore, considering the ENDOR data and coordination chemical concepts, it is surprising that the roles of terminal or bridging hydride ligands are so underappreciated in past computational work. Instead a considerable focus is placed on protonated bridging sulfides to create coordinatively unsaturated Fe sites. While this may be relevant for higher reduced states of FeMo-co, due to the considerable S→Fe

Nitrogenase Structure and Function Relationships by Density Functional Theory

283

electron donation, the nucleophilicity of the Fe sites can be comparable, if not even greater than that of the S sites. Notably, the mechanistic relevance of H atom migration, as a simplified notation for coupled protonation and reduction, has recently been evaluated (95). It was proposed that the three H atoms in the E3 H3 state need to be strategically placed on the FeMo-co for favorable N2 binding leading to first N2 Hx and then NHx intermediates. Among the many possible N2 coordination modes it was found that the bridging coordination results in N2 dissociation or structural rearrangement to end-on coordination (96). The endo-η1 -N2 coordination to a specific Fe site of the central Fe belt appears to be most favored followed by H2 dissociation and formation of a bound N2 H intermediate. However, the full mechanism was developed for a side-on-coordinated N2 (97) due to its preference for hydrogenation relative to the former. The catalytic pathway was defined as “intramolecular” due to the requirement of the presence of protonated S ligands. The estimated activation barriers are 69–96 kJ/mol, similar to previous theoretical estimates; however, these can only be considered as qualitative since they do not include zero-point energy and entropy corrections. The neglect of hydride intermediates may contribute to finding mechanistic pathways with activation barriers of 66– 119 kJ/mol (98). The presence of an Fe-bound hydride is expected to allow instantaneous two-electron transfer to a coordinated N2 and concomitant protonation to stabilize the shortlived charge transfer state. In a more recent computational effort (99), in addition to the entire catalytic cycle, unproductive sidereactions were also identified commendably. These can be considered as control or “blank” computational experiments that are unfortunately often neglected from computational studies. The energy differences between the main branch and the unproductive side-branches were found to be about 20 kJ/mol, which is considerable; however, due to the simplified computational model with – SH, NH3 , and – OH ligands this cannot be taken as definite. Also one of the key pillars of this study is the controversial coordination flexibility of FeMo-co cofactor. Conversion of a central μ2 -sulfide ligand to a terminal sulfhydryl ligand via protonation is expected to seriously compromise the Fe–S cluster integrity and binding to the protein environment, in addition to causing the appearance of spectroscopic signatures that are greatly different from anything observed for the compact FeMo-co. Most critically, while the protonated bridging sulfides are expected to be ENDOR silent due to approximately equivalent electron donation to attached Fe sites from both α- and β-spin orbitals, the spin polarization would be eminent from terminal sulfhydryl groups. A computational work (100) with similar models uses a ferromagnetically coupled FeMo-co cluster assuming that the spin

284

Harris and Szilagyi

coupling only considerably affects the magnetic properties of the cluster and not its chemical properties. While the study is very thorough and can be even considered as a “mini-review” due to its extensive introductory section, this type of approximation should be avoided due to the gross underestimation of the role of Fe–S–Fe super-exchange and [2Fe–2S] . . . [2Fe–2S] directexchange interactions. These together strongly influence the electrophilicity of Fe sites and nucleophilicity of S sites, which are crucial in reduction/substrate binding and protonation, respectively.

4. Summary and Outlook In this chapter we overviewed the remarkable potential of DFTbased electronic structure calculations in probing the structural and mechanistic details of nitrogenase. Conclusive geometric or electronic structural data have been difficult to obtain from experiment, but DFT with an appropriate level of theory and computational model can rule out certain proposed structures and electronic states through comparison with experimental parameters. As shown by the Thorneley–Lowe scheme, the N2 reduction mechanism involves a large number of intermediate states, many of which are difficult to trap experimentally. DFT allows examination of every intermediate and assessment of the feasibility of proposed mechanistic pathways, including N2 -binding modes, protonation states, and the role of the protein environment. However, this must be done within the boundaries of a systematically and critically compiled experimental background and by employing a few clever computational “blank” experiments. Otherwise the calculated potential energy surface will only have a distant resemblance to the actual chemical process. As it stands now, only high-quality theoretical calculations have the potential to explain the electronic tuning by Mo, the role of weak interactions from protein environment, importance of magnetic coupling among the Fe centers, understanding the potential of the Mo center being the catalytically active site, and the complete mechanism at the level of detail of transition states. This combined information can provide an invaluable blueprint for rationalized design of biomimetic compounds. On the other hand, as FeMo-co is one of the most complex [Fe–S] clusters in biology it provides a challenging case study for theoretical method development and application. For example, treatments and approaches learned by studying the structure of FeMo-co and its chemical function are expected to be readily applicable to other bioinorganic systems, particularly those containing heterometal substituted [Fe–S] clusters. The use of

Nitrogenase Structure and Function Relationships by Density Functional Theory

285

larger computational models will be necessary for the most accurate calculations of structural, electronic, magnetic, and energetic properties of protein-bound and extracted FeMo-co. In addition, we have just begun understanding the intimate molecular details of the complex protein environmental effects. The numerous site-directed mutagenesis studies provide an excellent training set for the development of multi-layered computational models that include DFT, ab initio MO, and molecular mechanics-based regions. It will be invaluable to learn about the minimal outer sphere environment, similar to those presented in Fig. 18.3, in order to construct realistic computational models for capturing the effects of a network of H-bonding, electrostatic, and steric interactions, from which we can gain feedback for designing future site-directed substitutions and other biochemical experiments. With the continually increasing accessibility of larger computational resources and more accurate levels of theory for manyelectron systems, the already indispensable DFT techniques will soon have even greater ability to complement experiment and uncover the remaining details of biological nitrogen fixation. It is the mission of the computational bioinorganic chemistry community to maintain the highest quality and yet clarity of simulation results for the general metalloenzymology community to keep the conversation going and facilitate mutual learning.

5. Notes The notes below are described for calculations using the Gaussian 03 (101) implementation of density functional theory and related basis sets. These can be readily converted to be compatible with the latest versions of numerous other quantum chemical codes. 1. We do not expect to find large spin densities on atoms far removed from the paramagnetic centers, which can occur with functionals such as BP86 that provide highly covalent descriptions of bonding. In these cases introduction of even a limited amount of HFX (a few percent) can produce more reasonable spin densities. 2. DFT calculations for antiferromagnetically coupled systems, such as FeMo-co, cannot be carried out with the “black box” method of simply specifying the charge, spin, and atomic coordinates. The majority spin direction of each Fe center in the broken symmetry state must be explicitly defined. In the ionic fragment approach (25) the initial wave function is constructed by merging ionic fragment wave functions with the spins flipped for the desired Fe centers. The pattern of α and β majority Fe centers defines a unique wave function;

286

Harris and Szilagyi

however, the various specific assignments of FeIII and FeII that can achieve the desired total spin and charge all result in the same final converged state. Using the merged wave function as an initial guess for a DFT calculation will often result in convergence difficulties. We recommend beginning with loose convergence criteria (10–4 or “sleazy”) at HF level and gradually incorporating density functional from B(50%HF)P86 to B(25%HF)P86 and finally to BP86 with tight convergence (10–8 or “tight”). If convergence is still difficult to achieve, the calculation can be restarted from the last known good SCF cycle with the more computationally expensive quadratic convergence (QC) method. 3. The converged state can be evaluated by examining the absolute value of the Fe spin densities, which, in general, should not deviate from each other by more than 1 e– for [Fe–S] clusters. Large deviations were often observed for highly charged (more negative than –4 with homocitrate modeled as glycolate) FeMo-co models, but using a solvation model such as PCM alleviated this problem. 4. Using an electronic structure calculated with a different functional as an initial guess is a useful shortcut; however, we do not recommend switching basis sets in the same manner due to the misalignment of atomic orbitals and the orbital coefficients. 5. Oxidized and reduced states can be calculated by starting with the converged resting state; however, there may be convergence difficulties for certain spin states such as S = 0, 1, or 2. In these cases using an alternative spin state as a guess may help. 6. Converged electronic structures can serve as convenient initial guesses, when exploring alternative isoelectronic interstitial atoms. For example, N3– can be switched to C4– without having to remerge the initial guess. 7. When starting a geometry optimization from the crystallographic coordinates we save time by using the Gaussian 03 “Opt” keywords “Big, GDIIS.” We found that using loose optimization criteria is sufficient for finding stationary points without imaginary frequencies. 8. To avoid being trapped in a local minimum, we initially fix all atoms except the central ligand, X, and gradually release the rest of the cluster starting with the bridging sulfides, then the inner six Fe centers, the remaining sulfides, the terminal Fe and Mo, and finally all atoms. From this structure, X can be switched and the optimization can begin with all atoms free.

Nitrogenase Structure and Function Relationships by Density Functional Theory

287

9. We do not recommend switching between Fe oxidation states (e.g., 4FeII –3FeIII to 2FeII –5FeIII ); the new oxidation state distribution should be calculated/merged from scratch. 10. An exhaustive list of formatted checkpoint files containing pertinent electronic structure information such as optimized atomic position coordinates, charge and spin states, basis set, orbital coefficients, spin and total electron densities are available at http://computational.chemistry.montana.edu/ FeMo-co/. References 1. Hohenberg P, Kohn W (1964) Inhomogeneous electron gas. Phys Rev 136: B864–B871 2. Kohn W, Sham LJ (1965) Self-consistent equations including exchange and correlation effects. Phys Rev 140:A1133–A1138 3. Koch W, Holthausen MC (2000) A chemist’s guide to density functional theory, WileyVCH, Weinheim 4. Slater JC, Johnson KH (1972) Selfconsistent-field Xα cluster method for polyatomic molecules and solids. Phys Rev B 5:844–853 5. Slater JC (1951) A simplification of the Hartree-Fock method. Phys Rev 81: 385–390 6. Perdew JP, Schmidt K (2001) Jacob’s ladder of density functional approximations for the exchange-correlation energy. AIP Conf Proc 577:1–20 7. Ruzsinszky A, Perdew JP, Csonka GI (2010) The RPA atomization energy puzzle. J Chem Theory Comput 6:127–134 8. Becke AD (1993) A new mixing of HartreeFock and local density-functional theories. J Chem Phys 98:1372–1377 9. Becke AD (1993) Density-functional thermochemistry. III. The role of exact exchange. J Chem Phys 98:5648–5652 10. Perdew JP (1986) Density-functional approximation for the correlation energy of the inhomogeneous electron gas. Phys Rev B 33:8822–8824 11. Becke AD (1988) Density-functional exchange-energy approximation with correct asymptotic behavior. Phys Rev A 38: 3098–3100 12. Szilagyi RK, Metz M, Solomon EI (2002) Spectroscopic calibration of modern density functional methods using [CuCl4 ]2– . J Phys Chem A 106:2994–3007 13. Siegbahn PEM, Blomberg MRA (2000) Transition-metal systems in biochemistry studied by high-accuracy 0quantum

14.

15.

16.

17.

18.

19.

20. 21.

22. 23.

chemical methods. Chem Rev 100: 421–437 Noodleman L, Lovell T, Han WG et al (2004) Quantum chemical studies of intermediates and reaction pathways in selected enzymes and catalytic synthetic systems. Chem Rev 104:459–508 Grimme S (2006) Semiempirical GGA-type density functional constructed with a longrange dispersion correction. J Comput Chem 27:1787–1799 Chai JD, Head-Gordon M (2008) Longrange corrected hybrid density functionals with damped atom-atom dispersion corrections. Phys Chem Chem Phys 10:6615–6620 Zhao Y, Truhlar DG (2008) The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals. Theor Chem Acc 120:215–241 Noodleman L (1981) Valence bond description of anti-ferromagnetic coupling in transition-metal dimers. J Chem Phys 74:5737–5743 Neese F (2009) Prediction of molecular properties and molecular spectroscopy with density functional theory: from fundamental theory to exchange-coupling. Coord Chem Rev 253:526–563 Szabo A, Ostlund NS (1982) Modern theoretical chemistry, MacMillan, New York, NY Neese F (2004) Definition of corresponding orbitals and the diradical character in broken symmetry DFT calculations on spin coupled systems. J Phys Chem Solids 65: 781–785 Clark AE, Davidson ER (2001) Local spin. J Chem Phys 115:7382–7392 Kitagawa Y, Saito T, Ito M et al (2007) Approximately spin-projected geometry optimization method and its application to

288

24.

25.

26.

27.

28.

29.

30.

31.

32.

33. 34.

35.

36.

Harris and Szilagyi di-chromium systems. Chem Phys Lett 442:445–450 Harris TV, Szilagyi RK (2011) Comparative assessment of the composition and charge state of nitrogenase FeMo-cofactor. Inorg Chem in press, http://pubs.acs.org/doi/ abs/10.1021/ic102446n Szilagyi RK, Winslow MA (2006) On the accuracy of density functional theory for iron-sulfur clusters. J Comput Chem 27:1385–1397 Schaefer A, Horn H, Ahlrichs R (1992) Fully optimized contracted Gaussian basis sets for atoms Li to Kr. J Chem Phys 97: 2571–2577 Schaefer A, Huber C, Ahlrichs R (1994) Fully optimized contracted Gaussian basis sets of triple zeta valence quality for atoms Li to Kr. J Chem Phys 100: 5829–5835 Krishnan R, Binkley JS, Seeger R et al (1980) Self-consistent molecular orbital methods. XX. A basis set for correlated wave functions. J Chem Phys 72:650–654 McLean AD, Chandler GS (1980) Contracted Gaussian basis sets for molecular calculations. I. Second row atoms, Z = 11-18. J Chem Phys 72:5639–5648 Solomon EI, Hedman B, Hodgson KO et al (2005) Ligand K-edge X-ray absorption spectroscopy: covalency of ligand-metal bonds. Coord Chem Rev 249:97–129 Boysen RB, Szilagyi RK (2008) Development of palladium L-edge X-ray absorption spectroscopy and its application for chloropalladium complexes. Inorg Chim Acta 361:1047–1058 Dey A, Glaser T, Couture MMJ et al (2004) Ligand K-edge X-ray absorption spectroscopy of [Fe4 S4 ]1+,2+,3+ clusters: changes in bonding and electronic relaxation upon redox. J Am Chem Soc 126: 8320–8328 Reed AE, Weinhold F (1983) Natural bond orbital analysis of near-Hartree-Fock water dimer. J Chem Phys 78:4066–4073 Dolg M, Wedig U, Stoll H et al (1987) Energy-adjusted ab initio pseudopotentials for the first row transition elements. J Chem Phys 86:866–872 Ehlers AW, Bohme M, Dapprich S et al (1993) A set of f-polarization functions for pseudo-potential basis sets of the transition metals Sc-Cu, Y-Ag and La-Au. Chem Phys Lett 208:111–114 Hollwarth A, Bohme M, Dapprich S et al (1993) A set of d-polarization functions for pseudo-potential basis sets of the main group elements Al-Bi and f-type polarization func-

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

tions for Zn, Cd, Hg. Chem Phys Lett 208:237–240 Stephens PJ, Devlin FJ, Chabalowski CF et al (1994) Ab initio calculation of vibrational absorption and circular dichroism spectra using density functional force fields. J Phys Chem 98:11623–11627 Yoo SJ, Angove HC, Papaefthymiou V et al (2000) Mössbauer study of the MoFe protein of nitrogenase from Azotobacter vinelandii using selective 57 Fe enrichment of the Mcenters. J Am Chem Soc 122:4926–4936 True AE, Nelson MJ, Venters RA et al (1988) 57 Fe hyperfine coupling tensors of the FeMo cluster in Azotobacter vinelandii MoFe protein: determination by polycrystalline ENDOR spectroscopy. J Am Chem Soc 110:1935–1943 Lovell T, Li J, Liu TQ et al (2001) FeMo cofactor of nitrogenase: a density functional study of states MN , MOX , MR , and MI . J Am Chem Soc 123:12392–12410 Calculations involving other oxidation state distributions or compositions revealed that trends in the relative energies of spin coupling schemes are independent of the cluster composition and charge state Lukoyanov D, Pelmenschikov V, Maeser N et al (2007) Testing if the interstitial atom, X, of the nitrogenase molybdenum-iron cofactor is N or C: ENDOR, ESEEM, and DFT studies of the S = 3/2 resting state in multiple environments. Inorg Chem 46: 11437–11449 Schimpl J, Petrilli HM, Blöchl PE (2003) Nitrogen binding to the FeMo-cofactor of nitrogenase. J Am Chem Soc 125: 15772–15778 Einsle O, Tezcan FA, Andrade SLA et al (2002) Nitrogenase MoFe-protein at 1.16 Å resolution: a central ligand in the FeMocofactor. Science 297:1696–1700 Kim J, Rees DC (1992) Structural models for the metal centers in the nitrogenase molybdenum-iron protein. Science 257:1677–1682 Lovell T, Liu TQ, Case DA et al (2003) Structural, spectroscopic, and redox consequences of central ligand in the FeMoco of nitrogenase: a density functional theoretical study. J Am Chem Soc 125: 8377–8383 Li J, Nelson MR, Peng CY et al (1998) Incorporating protein environments in density functional theory: a self-consistent reaction field calculation of redox potentials of [2Fe2S] clusters in ferredoxin and phthalate dioxygenase reductase. J Phys Chem A 102:6311–6324

Nitrogenase Structure and Function Relationships by Density Functional Theory 48. Yang X, Niu SQ, Ichiye T et al (2004) Direct measurement of the hydrogen-bonding effect on the intrinsic redox potentials of [4Fe-4S] cubane complexes. J Am Chem Soc 126:15790–15794 49. Kitagawa Y, Shoji M, Saito T et al (2008) Theoretical studies on effects of hydrogen bonds attaching to cysteine ligands on 4Fe4S clusters. Int J Quantum Chem 108: 2881–2887 50. Pelmenschikov V, Case DA, Noodleman L (2008) Ligand-bound S = 1/2 FeMocofactor of nitrogenase: hyperfine interaction analysis and implication for the central ligand X identity. Inorg Chem 47:6162–6172 51. Tomasi J, Persico M (1994) Molecular interactions in solution: an overview of methods based on continuous distributions of the solvent. Chem Rev 94:2027–2094 52. Cossi M, Barone V, Cammi R et al (1996) Ab initio study of solvated molecules: a new implementation of the polarizable continuum model. Chem Phys Lett 255: 327–335 53. Barone V, Cossi M, Tomasi J (1998) Geometry optimization of molecular structures in solution by the polarizable continuum model. J Comput Chem 19:404–417 54. Lovell T, Li J, Case DA et al (2002) FeMo cofactor of nitrogenase: energetics and local interactions in the protein environment. J Biol Inorg Chem 7:735–749 55. Xie HJ, Wu RB, Zhou ZH et al (2008) Exploring the interstitial atom in the FeMo cofactor of nitrogenase: insights from QM and QM/MM calculations. J Phys Chem B 112:11435–11439 56. Rokhsana D, Dooley DM, Szilagyi RK (2008) Systematic development of computational models for the catalytic site in galactose oxidase: impact of outer-sphere residues on the geometric and electronic structures. J Biol Inorg Chem 13:371–383 57. Siegbahn PEM, Himo F (2009) Recent developments of the quantum chemical cluster approach for modeling enzyme reactions. J Biol Inorg Chem 14:643–651 58. Igarashi RY, Seefeldt LC (2003) Nitrogen fixation: the mechanism of the Modependent nitrogenase. Crit Rev Biochem Mol Biol 38:351–384 59. Huang HQ, Kofford M, Simpson FB et al (1993) Purification, composition, charge, and molecular weight of the FeMo cofactor from Azotobacter vinelandii nitrogenase. J Inorg Biochem 52:59–75 60. Pickett CJ, Vincent KA, Ibrahim SK et al (2003) Electron-transfer chemistry of the iron-molybdenum cofactor of nitrogenase:

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

289

delocalized and localized reduced states of FeMoco which allow binding of carbon monoxide to iron and molybdenum. Chem Eur J 9:76–87 Hedman B, Frank P, Gheller SF et al (1988) New structural insights into the iron-molybdenum cofactor from Azotobacter vinelandii nitrogenase through sulfur K and molybdenum L X-ray absorption edge studies. J Am Chem Soc 110:3798–3805 Corbett MC, Tezcan FA, Einsle O et al (2005) Mo K- and L-edge X-ray absorption spectroscopic study of the ADP·AlF4 - stabilized nitrogenase complex: comparison with MoFe protein in solution and single crystal. J Synchrotron Radiat 12:28–34 Venters RA, Nelson MJ, McLean PA et al (1986) ENDOR of the resting state of nitrogenase molybdenum iron proteins from Azotobacter vinelandii, Klebsiella pneumoniae, and Clostridium pasteurianum: 1 H, 57 Fe, 95 Mo, and 33 S studies. J Am Chem Soc 108:3487–3498 Lee HI, Hales BJ, Hoffman BM (1997) Metal-ion valencies of the FeMo cofactor in CO-inhibited and resting state nitrogenase by 57 Fe Q-band ENDOR. J Am Chem Soc 119:11395–11400 Huynh BH, Munck E, Orme-Johnson WH (1979) Nitrogenase XI: Mössbauer studies on the cofactor centers of the MoFe protein from Azotobacter vinelandii OP. Biochim Biophys Acta 576:192–203 Huynh BH, Henzl MT, Christner JA et al (1980) Nitrogenase XII: Mössbauer studies of the MoFe protein from Clostridium pasteurianum W5. Biochim Biophys Acta 623:124–138 Johnson MK, Thomson AJ, Robinson AE et al (1981) Characterization of the paramagnetic centers of the molybdenum-iron protein of nitrogenase from Klebsiella pneumoniae using low temperature magnetic circular dichroism spectroscopy. Biochim Biophys Acta 671:61–70 Conradson SD, Burgess BK, Holm RH (1988) Fluorine-19 chemical shifts as probes of the structure and reactivity of the ironmolybdenum cofactor of nitrogenase. J Biol Chem 263:13743–13749 Lukoyanov D, Barney BM, Dean DR et al (2007) Connecting nitrogenase intermediates with the kinetic scheme for N2 reduction by a relaxation protocol and identification of the N2 binding state. Proc Natl Acad Sci USA 104:1451–1455 Seefeldt LC, Hoffman BM, Dean DR (2009) Mechanism of Mo-dependent nitrogenase. Annu Rev Biochem 78:701–722

290

Harris and Szilagyi

71. Vrajmasu V, Munck E, Bominaar EL (2003) Density functional study of the electric hyperfine interactions and the redox-structural correlations in the cofactor of nitrogenase. Analysis of general trends in 57 Fe isomer shifts. Inorg Chem 42:5974–5988 72. Huniar U, Ahlrichs R, Coucouvanis D (2004) Density functional theory calculations and exploration of a possible mechanism of N2 reduction by nitrogenase. J Am Chem Soc 126:2588–2601 73. Dance I (2003) The consequences of an interstitial N atom in the FeMo cofactor of nitrogenase. Chem Commun (3):324–325. 74. Dance I (2006) The correlation of redox potential, HOMO energy, and oxidation state in metal sulfide clusters and its application to determine the redox level of the FeMo-co active-site cluster of nitrogenase. Inorg Chem 45:5084–5091 75. Hinnemann B, Nørskov JK (2003) Modeling a central ligand in the nitrogenase FeMo cofactor. J Am Chem Soc 125:1466–1467 76. Mayer SM, Niehaus WG, Dean DR (2002) Reduction of short chain alkynes by a nitrogenase α-70Ala -substituted MoFe protein. Dalton Trans 802–807 77. Benton PMC, Mayer SM, Shao JL et al (2001) Interaction of acetylene and cyanide with the resting state of nitrogenase alpha96-substituted MoFe proteins. Biochemistry 40:13816–13825 78. Fisher K, Dilworth MJ, Newton WE (2000) Differential effects on N2 binding and reduction, HD formation, and azide reduction with α-195His - and α191Gln substituted MoFe proteins of Azotobacter vinelandii nitrogenase. Biochemistry 39: 15570–15577 79. Barney BM, Yurth MG, Dos Santos PC et al (2009) A substrate channel in the nitrogenase MoFe protein. J Biol Inorg Chem 14:1015–1022 80. Barney BM, Lee HI, Dos Santos PC et al (2006) Breaking the N2 triple bond: insights into the nitrogenase mechanism. Dalton Trans 2277–2284 81. Barney BM, McClead J, Lukoyanov D et al (2007) Diazene (HN=NH) is a substrate for nitrogenase: insights into the pathway of N2 reduction. Biochemistry 46:6784–6794 82. Hoffman BM, Dean DR, Seefeldt LC (2009) Climbing nitrogenase: toward a mechanism of enzymatic nitrogen fixation. Acc Chem Res 42:609–619 83. Barney BM, Lukoyanov D, Igarashi RY et al (2009) Trapping an intermediate of dinitrogen (N2 ) reduction on nitrogenase. Biochemistry 48:9094–9102

84. Yandulov DV, Schrock RR (2003) Catalytic reduction of dinitrogen to ammonia at a single molybdenum center. Science 301:76–78 85. Schrock RR (2008) Catalytic reduction of dinitrogen to ammonia by molybdenum: theory versus experiment. Angew Chem Int Edn 47:5512–5522 86. McLean PA, Dixon RA (1981) Requirement of nifV gene for production of wildtype nitrogenase enzyme in Klebsiella pneumoniae. Nature 292:655–656 87. Smith BE, Durrant MC, Fairhurst SA et al (1999) Exploring the reactivity of the isolated iron-molybdenum cofactor of nitrogenase. Coord Chem Rev 185:669–687 88. Hughes DL, Ibrahim SK, Pickett CJ et al (1994) On carboxylate as a leaving group at the active-site of Mo nitrogenase – electrochemical reactions of some Mo and W carboxylates, formation of mono-hydrides, dihydrides and tri-hydrides and the detection of an MoH2 (N2 ) intermediate. Polyhedron 13:3341–3348 89. Durrant MC, Francis A, Lowe DJ et al (2006) Evidence for a dynamic role for homocitrate during nitrogen fixation: the effect of substitution at the α-Lys426 position in MoFe-protein of Azotobacter vinelandii. Biochem J 397:261–270 90. Coucouvanis D, Demadis KD, Malinak SM et al (1996) Catalytic and stoichiometric multielectron reduction of hydrazine to ammonia and acetylene to ethylene with clusters that contain the MFe3 S4 cores (M = Mo, V). Relevance to the function of nitrogenase. J Mol Catal A: Chem 107:123–135 91. Demadis KD, Malinak SM, Coucouvanis D (1996) Catalytic reduction of hydrazine to ammonia with MoFe3 S4 -polycarboxylate clusters. Possible relevance regarding the function of the molybdenum-coordinated homocitrate in nitrogenase. Inorg Chem 35:4038–4046 92. Lukoyanov D, Yang ZY, Dean DR et al (2010) Is Mo involved in hydride binding by the four-electron reduced (E4 ) intermediate of the nitrogenase MoFe protein? J Am Chem Soc 132:2526–2527 93. Peters JW, Szilagyi RK (2006) Exploring new frontiers of nitrogenase structure and mechanism. Curr Opin Chem Biol 10:101–108 94. Malinak SM, Simeonov AM, Mosier PE et al (1997) Catalytic reduction of cisdimethyldiazene by the [MoFe3 S4 ]3+ clusters. The four-electron reduction of a N=N bond by a nitrogenase-relevant cluster and implications for the function of nitrogenase. J Am Chem Soc 119: 1662–1667

Nitrogenase Structure and Function Relationships by Density Functional Theory

291

95. Dance I (2006) Mechanistic significance 98. Hinnemann B, Norskov JK (2006) Catalysis by enzymes: the biological ammonia syntheof the preparatory migration of hydrosis. Top Catal 37:55–70 gen atoms around the FeMo-co active site of nitrogenase. Biochemistry 45: 99. Kästner J, Blöchl PE (2007) Ammonia production at the FeMo cofactor of nitrogenase: 6328–6340 results from density functional theory. J Am 96. Dance I (2007) The mechanistically signifiChem Soc 129:2998–3006 cant coordination chemistry of dinitrogen at FeMo-co, the catalytic site of nitrogenase. J 100. McKee ML (2007) Modeling the nitrogenase FeMo cofactor with high-spin Fe8 S9 X+ Am Chem Soc 129:1076–1088 97. Dance I (2008) The chemical mecha(X=N, C) clusters. Is the first step for N2 nism of nitrogenase: hydrogen tunneling reduction to NH3 a concerted dihydrogen and further aspects of the intramolecutransfer? J Comput Chem 28:1342–1356 lar mechanism for hydrogenation of η2 -N2 101. Frisch MJ, Trucks GW, Schlegel HB et al (2006) Gaussian 03 Rev E.01, Gaussian Inc., on FeMo-co to NH3 . Dalton Trans (43): Wallingford, CT 5992–5998

Chapter 19 Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory Gregory M. Sandala and Louis Noodleman Abstract Density functional theory (DFT) represents a unified framework for gaining molecular level insight into molybdenum–iron (MoFe) nitrogenase. However, accurately describing the electronic structure of the spin-polarized and spin-coupled iron–molybdenum cofactor (FeMo-co) where N2 reduction occurs within MoFe nitrogenase is challenging. Therefore, the enhancement of DFT to include broken symmetry (BS-DFT) plus approximate spin projection has proven valuable because it provides a procedure to compute reliable geometries, energies, redox potentials, and quantities relevant to Mössbauer and ENDOR spectroscopies. After describing the theoretical tools necessary to obtain this information, we show by way of examples how BS-DFT is a very powerful partner to experiment. We expect that quantitative quantum chemical theory of this type will play an ever-increasing role in helping to decipher complex bioinorganic systems like those found in MoFe nitrogenase. Key words: Density functional theory (DFT), broken symmetry, spin coupling, spin polarization, spin projection techniques, redox potentials, Mössbauer parameters, ENDOR hyperfine parameters, FeMo cofactor, iron–sulfur clusters.

1. Introduction The ability of diazotrophs to reduce atmospheric nitrogen (N2 ) to ammonia (NH3 ) at ambient pressure and temperature represents one of nature’s most elegant and mysterious transformations. In molybdenum–iron (MoFe) nitrogenase the stoichiometry of the overall reaction can be written as follows: N2 +8H+ +8e − +16 MgATP → 2 NH3 +H2 +16MgADP+16Pi

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9_19, © Springer Science+Business Media, LLC 2011

293

294

Sandala and Noodleman

Fig. 19.1. Selected region of the active site of MoFe nitrogenase containing the redoxactive FeMo cofactor with its interstitial ligand X. The covalent ligands of FeMo-co include Cys275, His442, and homocitrate (HCA). Other potentially important residues are depicted in licorice representation.

The high cost to carry out this reaction (viz. the hydrolysis of 16 MgATP) reflects the difficulty in reducing the extremely strong and nonpolar triple bond of N2 . The efficient regulation of multiple 1e– reductions and proton transfers is notably challenging as well. In the MoFe enzyme, the unique iron–molybdenum cofactor (FeMo-co) is the site of this remarkable chemistry (1). Figure 19.1 depicts the FeMo cofactor and a selection from its first noncovalently bonded coordination sphere in the active site of MoFe nitrogenase. The first nitrogenase enzyme X-ray crystal structures from the early 1990s done by the Rees and Bolin groups showed that the MoFe7 S9 cluster took the form of a trigonal prism with six central Fe atoms and capped with an Fe and Mo at each end (2). The nine inorganic sulfides of the core are either μ2 S bridging around the central waist or μ3 S bridging within the two cuboidal [Fe4 ] and [Fe3 Mo] subunits. In contrast to most other known Fe–S clusters, the MoFe7 S9 cluster possesses only two covalent linkages to the protein, one involving Mo–N(δ)His442 and the other Fe–SCys275 . Homocitrate also coordinates to the Mo ion of the FeMo cofactor through its 2-hydroxy and 2-carboxyl groups. Its high negative charge (of –4) may help to chaperone the FeMo cofactor into position within MoFe nitrogenase via a positively charged channel (3). We note that in the resting state of the enzyme the Mo ion is formally Mo(IV), consistent with what is known about its physical state (4). Originally, the central site in the cluster appeared to be vacant, which would lead to each central Fe being only threecoordinate—a very unusual Fe environment for Fe–S clusters (5). Later, however, a high-resolution (1.16 Å) X-ray structure from Rees’s group showed that there is, in fact, a light atom at the center of the cluster (6). The identity of atom “X” could not be established unequivocally, as the electron density from the X-ray

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory

295

structure is consistent with X being either C4– , N3– , or O2– . The sulfide ion, S2– , is excluded as a candidate as atom X because it possesses a higher electron density than observed in the electron density map from the X-ray structure. The X-ray structure shows electron density maps for both the internal geometries of the FeMo cofactor and the P cluster, which is an associated Fe8 S7 cluster of MoFe nitrogenase that delivers electrons to FeMo-co (1), as well as for the overall protein structure. The resting state (called MN ) electronic spin of FeMo-co has been previously determined to possess a total spin of S = 3/2 (7), which is apparently stable in the X-ray beam. However, the data provided by the X-ray structure do not give direct information about the overall cluster oxidation state or those of the individual metal ions. Information like this is embedded in 57 Fe Mössbauer and ENDOR spectroscopic data, but this requires further analysis to gain an understanding of the metal sites and their spin coupling. Further, the interpretation of the appropriate oxidation state is connected to knowledge about the central atom. Clearly, this is a challenging task. Theoretical modeling, and in particular density functional theory (DFT), offers an attractive and convenient tool to address these issues, because it can examine different redox, protonation, and substrate binding states under which the enzyme may operate (8). Indeed, the electronic structure, relative energies, Mössbauer and ENDOR spectroscopic parameters, and redox potentials can be calculated and compared with experiment, offering wide scope to further our understanding of the nitrogenase system. We will demonstrate by way of examples how these theoretical tools have been applied successfully to a variety of issues. First, however, we will describe the theoretical framework for treating spin-coupled polynuclear transition metal complexes.

2. Theoretical Methodology 2.1. Broken Symmetry Density Functional Theory (BS-DFT)

We are concerned with using a type of DFT that is appropriate to investigate polynuclear mixed-valence metal systems with spin polarization and strong spin coupling like those found in the cofactor clusters of MoFe nitrogenase (9). This method is called broken symmetry plus approximate spin projection DFT, or simply BS-DFT (10). In this model, the electronic structure generates a spin-up (α) electron density that occupies a different spatial region from the spin-down (β) electron density, giving rise to orbitals that can overlap to give states of mixed spin symmetry and lower space symmetry (hence the term “broken symmetry”). The BS wave function is then a weighted average of pure spin states

296

Sandala and Noodleman

that are orthogonal and non-interacting with respect to the total Hamiltonian of the system. This property can be exploited to evaluate the energies and properties of the underlying pure spin states. In particular, hyperfine properties are extracted from the BS state by spin projection techniques based on the Wigner–Eckart theorem for vector operators (10). The need to apply BS-DFT to systems of this type is important because their electronic structures (and therefore properties) are heavily affected by the degree of electronic spin polarization, spin coupling, charge polarization (including metal–ligand (Fe–S) and metal–substrate covalency), and resonance delocalization (11). It is necessary to choose a balanced exchange-correlation functional that can describe spin-polarized and spin-coupled systems accurately. One issue that is of particular concern is that of spin crossover, which can arise when different spin states for one or more Fe sites (or other metal sites) lie close in energy. Swart et al. have assessed a series of exchange-correlation functionals using a set of seven Fe complexes that have been experimentally determined to possess low-, intermediate-, or high-spin ground states (12). Of the functionals examined, one group was found to disfavor high-spin states (e.g., LDA, BLYP, and PBE), while the other group, which includes hybrid and improved generalized gradient approximation (GGA) functionals and metaGGAs, tended to perform much better, particularly those using the OPTX exchange functional. In another study, Swart and co-workers examined the performance of the OPTX functional with the PBE, LYP, and Perdew correlation functionals to calculate thermochemical parameters, including atomization energies, reaction barriers, geometries, and vibrational frequencies (13). Again, it was found that functionals using the OPTX exchange functional perform better than conventional GGAs like PBE and BLYP. We have used a variety of functionals, including PW91 and BP86, with significant success for several types of Fe–S clusters and other Fe- and Mn-containing complexes (14). More recently, we have investigated the use of OPBE and OLYP in the reaction catalyzed by methane monooxygenase and found them to perform well (15). Several other functionals applied to the nitrogenase problem are in use, including RPBE, PBE, BLYP, and BP86 (see Section 3.3). 2.2. Spin Coupling, Spin Polarization, and Broken Symmetry (BS) States

In common with more typical and simpler Fe–S clusters, the Fe sites of the FeMo-co cluster are high spin and display antiferromagnetic coupling. There are eight metals within the cluster, each with the potential to carry unpaired electrons. For this reason, several valid sets of site spin vectors can be constructed that satisfy a given set of oxidation states and total spin. The challenge is to construct meaningful spin alignments for the FeMo-co spin

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory

297

Fig. 19.2. Collinear spin coupling alignments (BS states) for the seven iron sites of FeMo-co (reprinted with permission from (36). Copyright 2007 American Chemical Society).

state of interest. We have constructed such alignments based on a spin collinear (up ↑ or down ↓) approach for the high-spin ferric (Fe3+ , d5 , S = 5/2) and ferrous (Fe2+ , d6 , S = 2) sites. For the resting state of FeMo-co ([Mo4+ 3Fe3+ 4Fe2+ ], S = 3/2), 10 spin alignments of a 4↑:3↓ spin coupling pattern were constructed and analyzed (Fig. 19.2, denoted BS1–BS10). In addition to these BS states, a subsequent examination of the 2e– reduced ligandbound state of FeMo-co identified additional BS states with spin alignments of 3↑:4↓ and 5↑:2↓ (see Section 3.2). Broken symmetry states are obtained by first finding solutions to the ferromagnetic high-spin state (i.e., 7↑:0↓ with S = 31/2 for the [Mo4+ 3Fe3+ 4Fe2+ ] resting state and S = 29/2 for the [Mo4+ 1Fe3+ 6Fe2+ ] ligand-bound state), and then exchanging the α (↑) and β (↓) electron densities on the Fe atoms associated with the desired BS state. The calculation is then restarted and its wave function converged to this BS state, which can then be followed by geometry optimizations and/or property calculations. 2.3. Calculation of the Hyperfine Coupling Parameters for X and 57 Fe

Concerning the identity of the interstitial atom X (C, N, or O), it is necessary to project the calculated unrestricted broken symmetry isotropic contribution of the hyperfine coupling A tensor onto the total system spin. Quantitatively, this can be expressed as follows: Aiso (X) =

1 6

UBS UBS (|Ki /Si | St )Aiso (X) = PX Aiso (X)

[1]

i=2,3,...,7

where the sum runs over the six Fe atoms surrounding X and Aiso (X) is the calculated spin-coupled isotropic hyperfine parameter that can be compared directly to experiment. Ki is a

298

Sandala and Noodleman

spin projection coefficient that is defined as the projection of a given local Fe site spin vector Si onto the total FeMo-co spin vector St : Ki = Si · St / St · St = Si · St /St (St + 1)

[2]

These Ki coefficients are obtained with spin projection chain and sum rules. The chain rule expresses Ki as the product of a site spin projection coefficient onto a subunit spin, Ki q , and the subunit spin projection coefficient onto the total system spin, Kq t : q

Ki = Kit = Ki Kqt

[3]

If the total spin vector of the subsystems q and r can be written as Sq = Si + Sj and Sr = Sk + Sl , and so long as Si , Sj , Sq , and Sr are good spin quantum numbers, then the closed-form equations for K are as follows: q

Ki = [Sq (Sq + 1) + Si (Si + 1) − Sj (Sj + 1)]/2Sq (Sq + 1)

[4]

Kqt = [St (St + 1) + Sq (Sq + 1) − Sr (Sr + 1)]/2St (St + 1)

[5]

The chain rule can also be applied recursively for systems with multiple subsystems (33, 40). The sum rule states that the sum of the site spin projection coefficients, whether for the total system or for subsystems, will be unity (if all the spin centers are included in the system or subsystem):

Ki = 1

[6]

i=1,2,3,...,7

To calculate the transition metal hyperfine parameters we note that those calculated with DFT are prone to errors because of the difficulty in accurately calculating the Fermi contact term on Fe (and similarly on Mn) (16). Therefore, semi-empirical calculations of the 57 Fe site hyperfine coupling constants are evaluated with the following equation: Aicalc = Ki (P3d (F ei ) /2Si )aiionic = Ki dB (F ei )aiionic

[7]

3d Mulliken spin population of where P3d (Fei ) is the calculated the ith site, and P3d (Fei ) /2Si is the covalency factor, which relates the 3d spin populations to the 2Si maximum expected spin population in the valence-bond limit. At the valence-bond limit dB (Fei ) = 1 and decreases with increasing covalency. If dB (Fei ) ≥ 1 for a given BS state, then it is very likely that a chosen site spin Si is incompatible with the electronic structure of the system. The factor aionic is dependent on the Fe oxidation state and represents the intrinsic site hyperfine parameter for a purely ionic

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory

299

ion. Values developed for the P cluster of the MoFe protein and other iron–sulfur proteins are reported to be –34.0, –32.5, and – 31.0 MHz for the Fe2+ , Fe2.5+ , and Fe3+ ions, respectively (17).

2.4. Mössbauer Parameters 2.4.1. Isomer Shifts

Mössbauer isomer shifts are proportional to the s electron density at the nucleus and, in addition to quadrupole splittings, are key indicators of both the metal–ligand valency and the spin and oxidation states of the Mössbauer atoms. Important contributors to the magnitude of an isomer shift are electrostatic shielding of the 3s electrons by the Fe 3d electrons and valence contributions arising from covalency effects, changes in bond lengths, and changes in shielding by the 3d electrons (18). The isomer shift can be calculated by linear regression of the following equation: δ = α[ρ(0) − A] + C

[8]

where δ is the experimental isomer shift (mm s–1 ), ρ(0) is the calculated electron density at the nucleus (e a0 −3 ), α is the slope (e −1 a0 3 mm s−1 ), and C is the intercept (mm s–1 ). The value of A is chosen to be close to the nuclear density of Fe in its reference state, which is roughly 11,880. Hopmann and co-workers recently examined Mössbauer parameters of 21 non-heme iron complexes and found the OLYP functional to perform well against experimental isomer shift and quadrupole splitting values (19). Figure 19.3 shows the excellent

Fig. 19.3. Linear correlation between measured isomer shifts and calculated electron densities for 8 iron–NO complexes and 12 iron–sulfur complexes with 24 distinct Fe sites. Calculated at the OLYP/TZP level of theory with the COSMO solvation model using methanol as solvent (adapted with permission from (19). Copyright 2009 American Chemical Society).

300

Sandala and Noodleman

agreement obtained for 8 iron–NO complexes and 12 iron–sulfur complexes with 24 distinct Fe sites calculated at the OLYP/TZP level of theory with the COSMO solvation model using methanol as solvent. Parameters to determine δ were derived from calculations in the gas phase (α = −0.302, C = 0.416) and with the aforementioned COSMO model (α = −0.315, C = 0.432), where A = 11,877. Mean absolute errors for the associated 57 Fe isomer shifts were reported to be 0.040 and 0.039 mm s–1 , respectively. For calibration, experimental isomer shifts are often temperature corrected to 4.2 K (liquid He) by including second-order Doppler effects (19). At 300 K, the offset (δ4.2 K − δ300 K ) is 0.12 mm s–1 and is generally regarded to be linear with temperature (20). Relativistic effects will change the electron density at the Mössbauer nucleus. However, its inclusion is likely to change only the slope (and intercept) of the calibration curve due to a significant cancelation of errors. The good performance of nonrelativistic calculations to determine δ has been attributed to an essentially constant contribution from the 1s and 2s orbitals to the electron density at the nucleus (18). 2.4.2. Quadrupole Splittings

Quadrupole splittings arise due to the interaction of the quadrupole moment of the first nuclear excited state of 57 Fe (I = 3/2) with the asymmetric electronic charge distribution (i.e., electric field gradient) surrounding the iron nucleus. They give information on the populations of the 3d orbitals and their surrounding ligand environment and are thus valuable indicators in helping to understand the structures and properties of Fe–S clusters. The quadrupole interaction splits the 57 Fe (I = 3/2) nucleus into two doubly degenerate states (mI ± 1/2 and mI ± 3/2) whose energy difference is calculated with EQ =

1 eQ Vzz (1 + η2 /3)1/2 2

[9]

where e is the positive electric charge, Q is the nuclear quadrupole moment (in barns), and Vzz is the electric field gradient. η is parameter defined as (Vxx − Vyy )/Vzz with |Vzz | ≥ an asymmetry Vyy ≥ |Vxx |. The calculations of both δ and EQ in FeMo-co may well enable intimate knowledge of its electronic structure and the mode(s) with which the cofactor binds and manipulates substrates and intermediates. DFT calculations on a series of mixed-valence Fe compounds relevant to the reactions catalyzed by methane monooxygenase and ribonucleotide reductase were shown to yield results in very good agreement with experiment (21). It is thus fortunate that experimental Mössbauer parameters (δ and EQ ) have been recorded for the resting state (S = 3/2) of MoFe

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory

301

nitrogenase (22). Parameters for ligand-bound states (S = 1/2) would also be desirable. In both types of states, DFT calculations with evaluations of properties should prove valuable for developing our understanding of these systems. 2.5. Solvent and Environmental Effects

An accurate representation of the protein and solvent is important to describe the geometries, energetics, and spectroscopic properties of the FeMo cofactor. The high anionic charge of FeMoco (that includes the auxillary sulfide and homocitrate ligands) has a significant influence on both the hydrogen bonding and longer range electrostatic interactions from the protein and solvent. Second-shell, third-shell, and extended environmental interactions can therefore be important (23). Our approach to model such interactions beyond the active site quantum cluster has been to use Bashford’s Poisson–Boltzmann electrostatics code MEAD (macroscopic electrostatics with atomic detail) in conjunction with the ADF software package (24). With this approach we can evaluate the cluster–environment interaction and decompose this into the reaction field (Er ) and protein field (Ep ) energies. The former involves the response of the dielectric media of the protein and solvent interacting with the quantum cluster, while the latter is the screened dielectric interaction between the protein charges and the quantum cluster. The screening involves the three dielectric media for protein, solvent, and cluster with typical dielectric constants of εp = 4, εs = 80, εc = 1. For a simple onestep MEAD energy evaluation the total interaction energy can be expressed as Epr = Er + Ep =

1 ∗ qi Φreac (i) + qi Φprot (i) 2

[10]

∗ (i) is the electrostatic potential due to the polarization of Φreac the protein/solvent dielectric induced by the charge distribution on all atoms (qi ) in the active site cluster embedded in the protein/solvent environment. The qi are determined by electrostatic potential fitting of the potential outside of this active site region to an atom-centered point charge model. The boundary between the quantum cluster and the protein/solvent is defined by a van der Waals envelope. Fprot (i) is the electrostatic potential that arises from the charge distribution of the protein residues acting on the active site cluster. More elaborate methods lead to full SCRF (self-consistent reaction field) calculations for the quantum cluster in a protein plus solvent environment, which is characterized by the cluster being polarized by the environment (25). The size of the quantum cluster is adaptable, which is important for strong interactions, particularly those involving charge transfer. With the total cluster–environment interaction energy in hand, it is possible to calculate redox potentials of the FeMo cofactor with

302

Sandala and Noodleman 0 Eredox = IP + Epr − 4.43

[11]

where IP is the gas-phase ionization potential, Epr is the energy difference between the reduced and oxidized states, and ESHE = −4.43 is used to reference the potential to the standard hydrogen electrode (SHE). This is the value for the absolute hydrogen oxidation potential ESHE used in our iron–sulfur cluster redox calculations. We note that uncertainty of the Gibbs free energy of solvation of the proton may lead to different values of ESHE (26). The latest values are near –4.36 to –4.28 eV (27–30). In Fe4 S4 systems, we have included well-defined spin correction terms to account for energy differences between the BS state being investigated and its corresponding lowest energy pure spin state (24). 2.6. Structural Models

In order to describe the geometries and properties of the FeMoco cluster in a tractable manner, model structures are constructed based on appropriate X-ray structures (Fig. 19.1). These models in their simplest form consist of the [Mo-7Fe-9S-X] core and its covalent ligands Cys275, His442, and R-homocitrate (HCA) modeled as methylthiolate, imidazole, and glycolate, respectively. The interstitial atom (X = C4– , N3– , or O2– ) is also included into the interior of the core. To ascertain the function of other potentially important residues, the second coordination shell has to be considered as well. Further, some of our studies have included the fully deprotonated homocitrate molecule (with a charge of –4) plus four H2 O molecules. Enhancements like these are particularly important in the calculation of redox potentials since the negatively charged FeMo-cofactor must be screened in a physically realistic manner to obtain appropriate results for comparison with experiment.

3. The Reaction Pathway 3.1. The Resting State

Based on early 57 Fe Q-band ENDOR and EPR data, Hoffman’s group suggested oxidation states for the S = 3/2 resting state of FeMo-co to be 1Mo4+ , 6Fe2+ , and 1Fe3+ (alternatively 1Mo4+ , 5Fe2+ , and 2Fe2.5+ ) (31). However, subsequent Mössbauer experiments led Münck and co-workers to propose that the resting state is instead 1Mo4+ 4Fe2+ 3Fe3+ (32). Prior to the high-resolution X-ray structure that identified an interstitial atom within the Fe prismane (6), we carried out an investigation of the resting (and other redox) states of the FeMo cofactor (33). At the time we concluded that the most likely oxidation states of the Mo and Fe sites are Mo4+ 6Fe2+ 1Fe3+ . Indeed, this assignment produced metal hyperfine and Mössbauer isomer shifts that

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory

303

agreed well with experimental results. In addition, the calculated geometries displayed excellent agreement with the available X-ray structures. Calculations of an Mo4+ 4Fe2+ 3Fe3+ configuration were found to yield poorer agreement with experiment, and this was therefore proposed to be less likely. This study was extended to examine the energetics and redox potentials of the FeMo cofactor in the protein using a Poisson–Boltzmann representation of the protein/solvent environment (23). The calculated redox potential for the one-electron oxidized MOX form of FeMo-co was found to deviate from experiment by +0.82 eV, which was surprising given that calculations on closely related Fe–S clusters were generally found to be 0.2–0.3 eV lower than experiment (24, 34). It was therefore very satisfying from our perspective when an interstitial atom X in FeMo-co was uncovered in a high-resolution X-ray structure (6). A re-examination of the structure and energetics of FeMo-co with X = C4– , N3– , O2– concluded the oxidation states of the metals in FeMo-co to be Mo4+ 4Fe2+ 3Fe3+ (35), which is consistent with the assignment based on Mössbauer data (32). Structures of FeMo-co with X = C4– and N3– were found to yield geometries in good agreement with experiment. In contrast, poorer agreement was obtained with structures that possessed a vacant site or used O2– as the interstitial atom (Table 19.1). An analysis of the associated redox potentials for MOX + 1e − → MN proved very revealing (35). Figure 19.4 shows a correlation between the FeMo-co cluster charge (and active site composition) and its redox potential. Specifically, clusters with increasingly negative total charge (e.g., –5 to –7, see lower right quadrant of Fig. 19.4) tend to possess more negative redox potentials, while the

Table 19.1 Averaged bond distances (Å) of the FeMo cofactor core with X = C4– , N3– , O2– , and vacant (35) 6Fe2+ 1Fe3+

4Fe2+ 3Fe3+

Bond type

C

N

O

Vacancy

C

N

O

Vacancy

Expt

Mo–Fe

2.81

2.80

2.80

2.75

2.77

2.78

2.81

2.75

2.70

Fe–Fe

2.63

2.69

2.76

2.70

2.64

2.65

2.79

2.64

2.61

Fe–Xa

2.02

2.03

2.17

2.01

2.02

2.11

Fe–Fe

2.63

2.67

2.82

2.75

2.61

2.63

2.74

2.66

2.59

Fe –Fe

2.76

2.71

2.75

2.70

2.70

2.69

2.74

2.66

2.66

2.00

a Fe–X is the average of all six coordinated Fe–X and Fe –X distances. Fe atoms are those of the Fe cubane, while 4 Fe atoms correspond to the Fe3 Mo cubane (cf. Fig. 19.1)

304

Sandala and Noodleman

Fig. 19.4. Calculated DFT/Poisson–Boltzmann redox potentials (eV) for models of the FeMo cofactor with a vacant central site or with different central atoms X = C4– , N3– , O2– compared with the experimental value near 0.0 eV (dotted vertical line) (reprinted with permission from (35). Copyright 2003 American Chemical Society).

opposite is true for clusters of more positive total charge (e.g., –1 to –5, see upper left quadrant of Fig. 19.4). These charges of the active site model in state MN include the full homocitrate anion charge, which is –4. Fig. 19.4 also shows that the calculated redox potential for Mo4+ 4Fe2+ 3Fe3+ with C4– is too negative (at –1.31 eV) relative to experiment, while the redox potential of the Mo4+ 4Fe2+ 3Fe3+ structure incorporating O2– is found to be too positive (at +0.97 eV). The best agreement with the experimental redox potential (of +0.042 eV) is observed for the Mo4+ 4Fe2+ 3Fe3+ structure containing N3– (+0.19 eV), suggesting that this configuration is likely to be biologically relevant. To further help to determine the identity of the interstitial atom X we have carried out BS-DFT calculations on the FeMoco core in its resting state (36). Calculations of the lowest BS state (BS7, Fig. 19.2) suggest that any hyperfine signal arising from X (A iso (X), MHz) is likely to be quite small (14 N = 0.3, 13 C = 1.0, 17 O = 0.1). Extensive 14 N and 15 N ENDOR and ESEEM investigations of FeMo-co extracted from the resting state MoFe protein in N-methyl-formamide compared with intact FeMo-co in native MoFe protein and with FeMo-co incorporated into NifX protein (a chaperone during FeMo-co biosynthesis) show no positive evidence for a central N3– atom (36, 37). Further, earlier evidence that natural abundance (1%) 13 C observed by 35 GHz ENDOR might correspond to C4– as the central atom X site could be dismissed based on follow-up pulsed ENDOR with both natural and enriched (6%) samples (36, 37). Although the uncertainties associated with our calculations cannot allow us to unambiguously assign the identity of X, our findings suggest that the opposite alignments of the six prismatic Fe spin vectors for the

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory

305

lowest energy BS state may lead to very low spin density on X (36). Consequently, its spin signature would be difficult to characterize experimentally in the resting state. 3.2. The Ligand-Bound State

An exciting experimental development in recent years has been the characterization of ligand-bound substrates in MoFe nitrogenase (38). The binding of a ligand causes the total spin of the system to change from its resting state of S = 3/2 to S = 1/2, indicating that a 2e– reduction and (presumably) 2H+ protonation of the FeMo cofactor cluster has occurred. In addition to the ability of such studies to illuminate the mechanism of action of MoFe nitrogenase, these studies may also help to identify the interstitial atom X, since it is plausible that ligand-bound states will distort the otherwise symmetric resting state spin distribution of the FeMo-co core. Recall, we propose that the inability to detect N3– as the central ligand arises because of a small |(0)|2 electron spin density at the X nucleus on account of the overall 3↑:3↓ spin coupling arrangement of the six prismane irons (36). Recently, we examined the spin coupling patterns and X = N3– hyperfine signals arising from the S = 1/2 state of FeMoco with allyl alcohol (H2 C=CH–CH2 –OH) bound as an Fe6η2 (C=C) ferracycle product (39, 40). Initially, a 2e– reduced ligand-free state of FeMo-co was examined to determine spin alignments compatible with S = 1/2 using our newly developed combinatorial algorithm that generates BS states automatically. In addition to the ten BS states identified for the S = 3/2 resting state, which possess 4↑:3↓ spin coupling patterns (cf. Fig. 19.2), ten inverted 3↑:4↓ and six new 5↑:2↓ spin alignments were identified for the S = 1/2 state. The lowest energy BS states correspond to 4↑:3↓ spin alignments denoted as BS2, BS6, and BS7, with corresponding relative energies of 0.0, 3.3, and 0.2 kcal mol–1 , respectively. For our subsequent analysis of 14 N hyperfine parameters, we chose the BS2 state for our calculations. Although this is the lowest energy BS state in the 2e– reduced FeMo-co without a bound ligand, alternative spin couplings may be preferred when a ligand is bound. To examine this we have calculated the relative energies of the BS2, BS6, and BS7 states for three minima of allyl alcohol bound to FeMo-co (Fig. 19.5). It can be seen that BS2 remains the lowest energy BS state for two of the structures, while BS7 is only 1.6 kcal mol–1 lower than BS2 for the third structure. Therefore, the use of BS2 for our subsequent analysis of hyperfine parameters is reasonable. Our calculated hyperfine parameters for X = 14 N of the ligand-bound structures are larger than those determined previously for the resting state (40). Indeed, we find that the signal of the interstitial atom Aiso (X) is likely to increase by at least one order of magnitude (3.8 ≤ Aiso (X = 14 N) ≤ 14.7 MHz) when

306

Sandala and Noodleman

Fig. 19.5. Relative energies (kcal mol–1 ) of the three lowest broken symmetry (BS) states for each of the three identified allyl alcohol minima bound to FeMo-co with nitride, N3– , as the central ligand (40).

allyl alcohol is bound to Fe6 of FeMo-co (40). Figure 19.6 shows the hyperfine values for BS2, BS6, and BS7 for each ligand-bound structure. We find the largest Aiso values arise for structure 2, in which the coordination bond from Fe6 to Nx is broken. The smallest calculated Aiso values are found for structure 3 and appear to result from the combined electronic and structural symmetry perturbations that arise due to the broken Fe6–S2B bond. Intermediate Aiso values are observed for structure 1. Importantly, we anticipate these results to be fairly insensitive to the exact identity of X, which may prove valuable to future experimental studies. 3.3. The Reaction Pathway

Many theoretical groups have examined N2 reduction pathways involving the FeMo cofactor. Due to space limitations, we briefly outline only selected topics of these investigations, and the interested reader is referred to the original publications and other reviews for more details (41, 42). Nørskov and Hinnemann investigated how a central nitrogen atom could insert into FeMoco using periodic plane-wave DFT methods with the RPBE functional on a minimal model (43). The investigated pathways were calculated to possess barriers and reaction energies that are unlikely to be accessible at ambient conditions. Energy profiles of minima for the reduction of an end-on bound N2 were also calculated. The profiles differ in the number of (simultaneous)

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory

307

Fig. 19.6. Absolute hyperfine parameters (Aiso (X=14 N) MHz) of the three lowest broken symmetry (BS) states for each of the three identified allyl alcohol minima bound to FeMo-co with nitride, N3– , as the central ligand (40).

protons and electrons added to FeMo-co, and in all cases the most demanding step was found to be the initial reduction of N2 . Reduction was found to proceed in an alternating fashion giving rise to coordinated intermediates like –N2 H2 and –N2 H4 . In another approach, Blöchl and co-workers examined N2 binding to a model of FeMo-co using Car–Parrinello molecular dynamics with the PBE functional based on the projectoraugmented wave method (44). Axial and bridging modes of N2 binding to the prismane Fe sites were identified with concomitant cleavage of one of the adjacent protonated sulfur bridges. For the bridging structures, the bond from the central N ligand to Fe3 becomes broken. N2 adsorption at the molybdenum site was also investigated and found to be endothermic by ca. 8 kcal mol–1 . This latter result suggests that N2 binding and reduction does not take place at the Mo ion of FeMo-co, especially in light of the preferential adsorption of N2 at Fe (44). Of relevance here are the findings of Schrock and co-workers who have demonstrated that abiological reduction of N2 can occur under mild conditions with catalysts possessing a single Mo ion (45). For this reason, they have argued that a similar mechanism may be operative in nitrogenase. It is notable that Schrock’s triamidoamine Mo homogeneous catalyst has a

308

Sandala and Noodleman

very unusual trigonal planar amido coordination compared with the Mo coordination environment in the FeMo-co cluster (Fig. 19.1). Further, in the abiological catalyst, Mo is in its Mo(III) oxidation state for the first N2 binding step, and different Mo oxidation states are accessed up to Mo(VI) nitride during the catalytic cycle. By contrast, in the biological FeMo cofactor, the resting oxidation state contains diamagnetic Mo(IV), rather than paramagnetic Mo(III). There is no current experimental evidence favoring nitrogen-based substrate binding at the Mo site, or that Mo(III) is driving the catalytic cycle. However, there is EPR and ENDOR hyperfine evidence for nitrogen- and carbon-based substrate analogues at a prismatic iron–sulfur face upon reduction (38). Further investigations are necessary to determine the extent to which Mo is involved in the enzymatic mechanism. A subsequent analysis by Blöchl and Kästner examined the energetics of N2 reduction to find a full catalytic cycle involving N2 bridging two Fe sites (46). In addition to this “main branch” pathway, an early side branch involving N2 bound axially could not be ruled out. In any case, this two-step side branch connects to the main (bridging) pathway at an early stage of the mechanism. We note that these investigations involve N2 reduction occurring at Fe3 and Fe7 of the prismane (cf. Fig. 19.1). In mutagenesis studies of Val70 of MoFe nitrogenase, Fe6 appears to be the site at which analogues of the substrate bind (38). It is unclear whether a change in the principal iron coordination site will affect the aforementioned mechanistic proposal. A substantial body of work due to Dance has characterized several possible coordination sites and intermediates in the FeMoco catalyzed reduction of N2 using the BLYP functional (47). Beginning with an η2 -coordination of N2 at the endo position of Fe6, an intricate 21-step mechanism for N2 reduction has been developed (48). Noteworthy features of this mechanism are that (i) protons are supplied via a relay chain involving water molecule 679 to S3B, which then transfer serially to Fe2, S2B, Fe6, and N2 and its derivatives; (ii) hydrogenation occurs at each N atom sequentially from either S3B, Fe2, or S2B; (iii) hydrogenation occurs intramolecularly, i.e., without the participation from surrounding amino acid residues; and (iv) quantum-mechanical tunneling is likely to play a role, as the donor/acceptor distances in this intramolecular pathway are quite small. The final theoretical N2 reduction investigation we highlight is from Coucouvanis and co-workers, who have employed the BP86 functional in their calculations (49). The unique feature of this proposal is that the central nitrido ligand acts as an exchangeable entity. The mechanism is also characterized by large structural changes to FeMo-co, viz., S2B and the presence of an exchangeable water molecule at an Fe site. This proposal appears to be at odds with data from previous ENDOR/ESEEM

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory

309

experiments (37), but the authors (49) suggest that there remain open questions regarding the detectability of a central N3– ligand that possesses small spin density. Overall, this mechanism requires major structural changes within the MoFe7 S9 cofactor core, which may not be energetically feasible (43).

4. Summary and Outlook We have shown how broken symmetry density functional theory (BS-DFT) calculations can be an effective tool to increase our understanding of the nitrogenase system. This method is well suited to describe Fe–S clusters, like those found within MoFe nitrogenase, because it can accurately describe high-spin, anti-ferromagnetic spin-coupled systems. With appropriate care and effort, reliable geometries, energies, redox potentials, and Mössbauer and ENDOR parameters can be determined. The ultimate goal of these studies is to provide a comprehensive mechanistic landscape of nitrogenase that is consistent with the available experimental data. Although progress has been made, there remain several outstanding issues. For example, recent experiments have localized substrate binding to the Fe2–Fe3–Fe6–Fe7 face of FeMo-co (38, also see Fig. 19.1), but how do the reactions proceed? Does the Mo ion participate directly in the reduction process as abiological catalysts suggest or does it operate solely as a spectator? What is the role played by active site residues, in particular cations Arg96 and His195 (with variable protonation states), which lie very close to the proposed substrate binding site? In addition, how are the protons shuttled to FeMo cofactor and the substrate? Most of the theoretical studies relevant to nitrogenase have addressed aspects of these questions using various model systems. The advent of high-performance supercomputers makes possible the use of more elaborate models that may capture features of the reaction mechanism that might be otherwise unattainable, like longer range charge transfer, the role of second- or third-sphere coordination shells, or extended proton relays. Developments in the design of robust and widely applicable DFT functionals are also likely to broaden the scope of bioinorganic investigations. For example, the M06 suite (50) and TPSSh (51, 52) density functionals display promising performance when assessed against large thermochemical test sets, as do the OLYP and OPBE functionals (13, 19). All in all, BS-DFT in conjunction with such functionals may offer realistic prospects to elucidate the mechanism of action of MoFe nitrogenase. It is therefore an opportune time for collaborative efforts between theory and experiment.

310

Sandala and Noodleman

Acknowledgments The authors would like to acknowledge the other contributors to our nitrogenase work, whose names appear throughout the reference section. Most recently, Vladimir Pelmenschikov and David A. Case have made substantial contributions. We also gratefully acknowledge financial support by NIH grant GM039914.

References 1. Howard JB, Rees DC (2006) How many metals does it take to fix N2 ? A mechanistic overview of biological nitrogen fixation. Proc Natl Acad Sci USA 103:17088–17093 2. See, for example: Howard JB, Rees DC (1996) Structural basis of biological nitrogen fixation. Chem Rev 96:2965–2982, and references therein. 3. Dos Santos PC, Dean DR, Hu YL et al (2004) Formation and insertion of the nitrogenase iron–molybdenum cofactor. Chem Rev 104:1159–1173 4. Lukoyanov D, Yang ZY, Dean DR et al (2010) Is Mo involved in hydride binding by the four-electron reduced (E4 ) intermediate of the nitrogenase MoFe protein? J Am Chem Soc 132:2526–2527 5. Sanakis Y, Power PP, Stubna A et al (2002) Mössbauer study of the three-coordinate planar FeII thiolate complex [Fe(SR)3 ]– (R = C6 H2 -2,4,6-tBu3 ): Model for the trigonal iron sites of the MoFe7 S9 :homocitrate cofactor of nitrogenase. Inorg Chem 41: 2690–2696 6. Einsle O, Tezcan FA, Andrade SLA et al (2002) Nitrogenase MoFe-protein at 1.16 angstrom resolution: A central ligand in the FeMo-cofactor. Science 297: 1696–1700 7. Lee HI, Hales BJ, Hoffman BM (1997) Metal-ion valencies of the FeMo cofactor in CO-inhibited and resting state nitrogenase by 57 Fe Q-Band ENDOR. J Am Chem Soc 119:11395–11400 8. Noodleman L, Lovell T, Han WG et al (2004) Quantum chemical studies of intermediates and reaction pathways in selected enzymes and catalytic synthetic systems. Chem Rev 104:459–508 9. For an introductory text on the general principles of DFT see, for example: Koch W, Holthausen MC (2002) A chemist’s guide

10.

11.

12.

13.

14.

15.

16.

17.

to density functional theory, 2nd edn. WileyVCH, Weinheim Noodleman L (1981) Valence bond description of anti-ferromagnetic coupling in transition metal dimers. J Chem Phys 74: 5737–5743 For more on these issues see, for example: Noodleman L, Lovell T, Liu T et al (2002) Insights into properties and energetics of iron–sulfur proteins from simple clusters to nitrogenase. Curr Opin Chem Biol 6: 259–273 Swart M, Groenhof AR, Ehlers AW et al (2004) Validation of exchange–correlation functionals for spin states of iron complexes. J Phys Chem A 108:5479–5483 Swart M, Ehlers AW, Lammertsma K (2004) Performance of the OPBE exchangecorrelation functional. Mol Phys 102: 2467–2474 Noodleman L, Han WG (2006) Structure, redox, pKa , spin. A golden tetrad for understanding metalloenzyme energetics and reaction pathways. J Biol Inorg Chem 11: 674–694 Han WG, Noodleman L (2008) Structural model studies for the high-valent intermediate Q of methane monooxygenase from broken-symmetry density functional calculations. Inorg Chim Acta 361:973–986 Sinnecker S, Neese F, Noodleman L et al (2004) Calculating the electron paramagnetic resonance parameters of exchange coupled transition metal complexes using broken symmetry density functional theory: Application to a MnIII /MnIV model compound. J Am Chem Soc 126:2613–2622 Mouesca JM, Noodleman L, Case DA et al (1995) Spin-densities and spin coupling in iron–sulfur clusters - A new analysis of hyperfine coupling-constants. Inorg Chem 34:4347–4359

Modeling the MoFe Nitrogenase System with Broken Symmetry Density Functional Theory 18. Neese F (2002) Prediction and interpretation of the 57 Fe isomer shift in Mössbauer spectra by density functional theory. Inorg Chim Acta 337:181–192 19. Hopmann KH, Ghosh A, Noodleman L (2009) Density functional theory calculations on Mössbauer parameters of nonheme iron nitrosyls. Inorg Chem 48: 9155–9165 20. Fee JA, Findling KL, Yoshida T et al (1984) Purification and characterization of the rieske iron–sulfur protein from Thermus thermophilus - evidence for a [2Fe-2S] cluster having non-cysteine ligands. J Biol Chem 259:124–133 21. Han WG, Liu TQ, Lovell T et al (2006) DFT calculations of 57 Fe Mössbauer isomer shifts and quadrupole splittings for iron complexes in polar dielectric media: Applications to methane monooxygenase and ribonucleotide reductase. J Comput Chem 27:1292–1306 22. See, for example: Yoo SJ, Angove HC, Papaefthymiou V et al (2000) Mössbauer study of the MoFe protein of nitrogenase from Azotobacter vinelandii using selective 57 Fe enrichment of the M-centers. J Am Chem Soc 122:4926–4936, and references therein. 23. Lovell T, Li J, Case DA et al (2002) FeMo cofactor of nitrogenase: Energetics and local interactions in the protein environment. J Biol Inorg Chem 7:735–749 24. Torres RA, Lovell T, Noodleman L et al (2003) Density functional and reduction potential calculations of Fe4 S4 clusters. J Am Chem Soc 125:1923–1936 25. Li J, Nelson MR, Peng CY et al (1998) Incorporating protein environments in density functional theory: A self-consistent reaction field calculation of redox potentials of [2Fe2S] clusters in ferredoxin and phthalate dioxygenase reductase. J Phys Chem A 102:6311–6324 26. See, for example: Noodleman L, Case DA (2009) Broken symmetry states of iron– sulfur clusters. In: Solomon EI, Scott RA, King RB (eds) Computational Inorganic and Bioinorganic Chemistry, pp. 213–228. Wiley, New York, NY 27. Tissandier MD, Cowen KA, Feng WY et al (1998) The proton’s absolute aqueous enthalpy and Gibbs free energy of solvation from cluster-ion solvation data. J Phys Chem A 102:7787–7794 28. Lewis A, Bumpus JA, Truhlar DG et al (2004) Molecular modeling of environmentally important processes: Reduction potentials. J Chem Educ 81:596–604 29. Lewis A, Bumpus JA, Truhlar DG et al (2007) Molecular modeling of environmen-

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

311

tally important processes: Reduction potentials. J Chem Educ 84:934 Fee JA, Case DA, Noodleman L (2008) Toward a chemical mechanism of proton pumping by the B-type cytochrome c oxidases: Application of density functional theory to cytochrome ba(3) of Thermus thermophilus. J Am Chem Soc 130: 15002–15021 Lee HI, Hales BJ, Hoffman BM (1997) Metal-ion valencies of the FeMo cofactor in CO-inhibited and resting state nitrogenase by 57 Fe Q-band ENDOR. J Am Chem Soc 119:11395–11400 Yoo SJ, Angove HC, Papaefthymiou V et al (2000) Mössbauer study of the MoFe protein of nitrogenase from Azotobacter vinelandii using selective 57 Fe enrichment of the Mcenters. J Am Chem Soc 122:4926–4936 Lovell T, Li J, Liu TQ et al (2001) FeMo cofactor of nitrogenase: A density functional study of states MN , MOX , MR , and MI . J Am Chem Soc 123:12392–12410 Ullmann GM, Noodleman L, Case DA (2002) Density functional calculation of pKa values and redox potentials in the bovine rieske iron–sulfur protein. J Biol Inorg Chem 7:632–639 Lovell T, Liu TQ, Case DA et al (2003) Structural, spectroscopic, and redox consequences of central ligand in the FeMoco of nitrogenase: A density functional theoretical study. J Am Chem Soc 125:8377–8383 Lukoyanov D, Pelmenschikov V, Maeser N et al (2007) Testing if the interstitial atom, X, of the nitrogenase molybdenum–iron cofactor is N or C: ENDOR, ESEEM, and DFT studies of the S = 3/2 resting state in multiple environments. Inorg Chem 46: 11437–11449 Yang TC, Maeser NK, Laryukhin M et al (2005) The interstitial atom of the nitrogenase FeMo-cofactor: ENDOR and ESEEM evidence that it is not a nitrogen. J Am Chem Soc 127:12804–12805 See, for example: Seefeldt LC, Hoffman BM, Dean DR (2009) Mechanism of Modependent nitrogenase. Annu Rev Biochem 78:701–722 Igarashi RY, Dos Santos PC, Niehaus WG et al (2004) Localization of a catalytic intermediate bound to the FeMo-cofactor of nitrogenase. J Biol Chem 279:34770–34775 Pelmenschikov V, Case DA, Noodleman L (2008) Ligand-bound S = 1/2 FeMocofactor of nitrogenase: Hyperfine interaction analysis and implication for the central ligand X identity. Inorg Chem 47: 6162–6172

312

Sandala and Noodleman

41. Peters JW, Szilagyi RK (2006) Exploring new frontiers of nitrogenase structure and mechanism. Curr Opin Chem Biol 10: 101–108 42. Tuczek F (2009) Electronic structure calculations: Dinitrogen reduction in nitrogenase and synthetic model systems. In: Solomon, EI, Scott, RA, King, RB (eds) Computational Inorganic and Bioinorganic Chemistry, pp. 287–307. Wiley, New York, NY 43. Hinnemann B, Nørskov JK (2004) Chemical activity of the nitrogenase FeMo cofactor with a central nitrogen ligand: Density functional study. J Am Chem Soc 126: 3920–3927 44. Schimpl J, Petrilli HM, Blöchl PE (2003) Nitrogen binding to the FeMo-cofactor of nitrogenase. J Am Chem Soc 125: 15772–15778 45. Schrock RR (2008) Catalytic reduction of dinitrogen to ammonia by molybdenum: Theory versus experiment. Angew Chem Int Edn 47:5512–5522 46. Kästner J, Blöchl PE (2007) Ammonia production at the FeMo cofactor of nitrogenase: Results from density functional theory. J Am Chem Soc 129:2998–3006

47. Dance I (2007) The mechanistically significant coordination chemistry of dinitrogen at FeMo-co, the catalytic site of nitrogenase. J Am Chem Soc 129:1076–1088 48. Dance I (2008) The chemical mechanism of nitrogenase: Calculated details of the intramolecular mechanism for hydrogenation of η2 -N2 on FeMo-co to NH3 . Dalton Trans:5977–5991 49. Huniar U, Ahlrichs R, Coucouvanis D (2004) Density functional theory calculations and exploration of a possible mechanism of N2 reduction by nitrogenase. J Am Chem Soc 126:2588–2601 50. Zhao Y, Truhlar DG (2008) Exploring the limit of accuracy of the global hybrid meta density functional for main-group thermochemistry, kinetics, and noncovalent interactions. J Chem Theor Comput 4:1849–1868 51. Perdew JP, Tao JM, Staroverov VN et al (2004) Meta-generalized gradient approximation: Explanation of a realistic nonempirical density functional. J Chem Phys 120:6898–6911 52. Jensen KP (2008) Bioinorganic chemistry modeled with the TPSSh density functional. Inorg Chem 47:10357–10365

INDEX A A. agilis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Ab initio . . . . . . . . . . . . . . . . . . . . . 181, 183–184, 268–271, 285 MO methods . . . . . . . . . . . . . . . . . . . . . . . . 268–271, 285 Absorber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165, 169–171 Absorption coefficient . . . . . . . . . . . . . . . . . . . . . . . . 166, 169 Acetylene . . . . . . . . . . . . . . 18, 22, 89, 116–117, 124, 199, 279–280 Acid labile sulfur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Acid treatment, MoFe protein . . . . . . . . . . . . . . . . . . . . . 242 Adenosine diphosphate (ADP) . . . . . . . . 5, 13, 19, 37, 74, 109, 122, 159–160, 178, 181, 183, 187 Adenosine triphosphate (ATP) . . . . . . . . 5, 11–13, 19–20, 31–32, 37–38, 69–76, 105–107, 109–110, 122–124, 130–131, 134, 138–139, 149–150, 177–179, 181, 183, 187, 197–199, 239 Advanced Photon Source Laboratory (APSL) . . . . . . . 185 Aerobic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5, 53, 58, 82, 134 Affinity purification . . . . . . . . . . . . . . . . . . . . . . . . . . . 136–137 Affinity-tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Agar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82, 84–85, 88 Aggregation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182, 185–187 Air-free techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Air-sensitive materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Äkta purifier FPLC system . . . . . . . . . . . . . . . . . . . . . . . . . 137 AlF4 - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71, 149, 154 Alkyl cyanides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114–115 Alkynes . . . . . . . . . . . . . . . . . . . . 18, 22, 198–199, 277, 279 All-ferrous . . . . . . . . . . . . . . . . . . . . . . 12, 32, 149, 225, 250 Allyl alcohol. . . . . . . . . . . . . . . . . . . . . . . .117, 198, 305–306 Alternative nitrogenases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Aluminum seal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Amicon . . . . . . . . . . . . . . . . . . . . . . . 95, 133, 138, 140, 158, 181 Amino acid . . . . . . 16, 18–20, 22, 50, 59, 68, 74–76, 87, 197, 202, 250 Ammonia . . . . . . . . . . . . . . . . . . 4, 9–10, 31, 49, 81, 97–98, 107–114, 119–120, 123–124, 177, 200, 239, 251, 274, 282, 293 assays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Ammonium hydroxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Ampicillin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83, 132, 135 Anaerobic . . . . . . . . 5, 49, 94, 96, 98–99, 101, 102, 105, 121, 130–138, 141–142, 151, 158–159, 161, 179, 181–182, 185–186, 198, 223, 241, 243, 245 protein purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Anf genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Anisotropic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 Antibiotic resistance . . . . . . . . . . . . . . . . . . . . . 59, 86–88, 90

Antiferromagnetic . . . . . . . . . . . . 222, 270, 272, 273, 277, 285, 296 coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277, 296 Antiparallel spins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272, 277 Arabinose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61–62 Argon. . . . .99, 101, 105, 107, 117, 120–121, 124, 162, 181, 185, 194–195, 202, 253, 257–258 Aromatic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83, 89 Assays of nitrogenase reaction products assay initiation/termination, method. . . . . . . . . . . . 106 extent of coupling between electrons/amount of MgATP hydrolyzed . . . . . . . . . . . . . . . . . . . 106 product analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 buffer used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 quantification of gaseous products, detection methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 methods carbon–carbon bonds, catalyzed reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 116–118 carbon–nitrogen bonds, catalyzed reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 113–116 carbon–oxygen and carbon–sulfur bonds, catalyzed reactions . . . . . . . . . . . . . . . . . 119–120 hydrogen, catalyzed reactions . . . . . . . . . . 120–121 nitrogen–nitrogen bonds, catalyzed reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 107–112 nitrogen–oxygen bonds, catalyzed reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 112–113 other assay components . . . . . . . . . . . . . . . . 121–123 steady-state assays, requirements addition of desired substrate . . . . . . . . . . . . . . . . 106 anaerobic environment. . . . . . . . . . . . . . . . . . . . . .105 choice of buffer system. . . . . . . . . . . . . . . . . . . . . .106 consistent source of MgATP . . . . . . . . . . . . . . . . 105 suitable source of reductant . . . . . . . . . . . . 105–106 Assembly . . . . . . . . . . . . 31–44, 54, 56, 59, 132, 134–135, 141 See also Assembly of nitrogenase MoFe protein Assembly of nitrogenase MoFe protein assembly model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 biosynthesis of MoFe protein, nif genes in . . . . . . . . 33 “ex situ” assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . 34–38 Fe protein in FeMoco assembly process, role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36–38 NifB, sequence analysis . . . . . . . . . . . . . . . . . . . . . . . 35 NifEN, role in maturation of FeMoco “core,” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35–36 NifUS protein complex . . . . . . . . . . . . . . . . . . . . . . 34 XAS/EXAFS study of NifENFeMoco , . . . . . . . . . 36 incorporation of FeMoco into MoFe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38–40

M.W. Ribbe (ed.), Nitrogen Fixation, Methods in Molecular Biology 766, DOI 10.1007/978-1-61779-194-9, © Springer Science+Business Media, LLC 2011

313

NITROGEN FIXATION

314 Index

Assembly of nitrogenase MoFe protein (continued) conformational rearrangement of protein, SAXS analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . .40 docking of NifEN on MoFe protein . . . . . . . 38–39 migration of FeMoco into insertion funnel, outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39–40 role of carrier proteins. . . . . . . . . . . . . . . . . . . . . . . .38 “in situ” assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . 40–43 Fe protein and NifZ in P-cluster assembly formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41–43 maturation of P-cluster by nifH MoFe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42–43 XAS/EXAFS-based structural models of P-cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 properties and its associated clusters . . . . . . . . . . 32–33 FeMoco, isolation/protein bound forms of . . . 33 P-cluster, oxidation states . . . . . . . . . . . . . . . . . 32–33 stepwise assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43–44 biphasic pattern of P-cluster maturation . . . 43–44 ATP and ATP-regenerating system . . . . . . . . . . . . . . . . . . . . 5 ATPase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72–73, 131, 148 ATP hydrolysis . . . . . . . . . . 11–12, 19, 37–38, 70, 72, 74, 122–124, 131, 149, 239 Autoclave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82, 132 Auxotrophic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 A. vinelandii, model for nitrogen fixation . . . . . . . . 53–58 accessory genes involved in nitrogen fixation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56–58 cydAB I genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 FeSII protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 NifA2 and VnfA2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 RhdA protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 rnf proteins, types . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 genes involved in nitrogen fixation . . . . . . . . . . . 53–56 cydAB I in respiratory protection, role . . . . . . . . 53 Fe-only nitrogenase, genes/gene coding regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 gene regions with sigma 54 promoters, identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Mo-dependent nitrogenase, genes/gene coding regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54–55 nitrogen-fixing gene regions . . . . . . . . . . . . . . . . . . 54 promoter analysis of nif, vnf, and anf gene regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 V-dependent nitrogenase, genes/gene coding regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54–56 use in nitrogen fixation, reasons . . . . . . . . . . . . . . . . . . 53 Axial splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 Azotobacter chroococcum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Azotobacter vinelandii. . . . . .13, 31, 50, 70, 82, 94, 114, 148, 155, 224 competency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84–85 assessment by fluorescence color intensity of culture. . . . . . . . . . . . . . . . . . . . . . . . . . . . .84, 90n2 genetic transformation of A. vinelandii. . . .84–85 OP strains and its derivatives . . . . . . . . . . . . . . 84–85 culturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83–84 genetic transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 selection from transformation culture blue and white screening . . . . . . . . . . . . . . . . . . . . . 88 direct antibiotic selection . . . . . . . . . . . . . 83, 86–87 rescue and congression transformation strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87–88 traditional genetics . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 See also A. vinelandii, model for nitrogen fixation

B Backscatterer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169–171 Bacterial growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Bacteriochlorophyll . . . . . . . . . . . . . . . . . . 52, 67–69, 72, 74 Bacteriochlorophyllide (Bchlide) . . . . . . . . . . . . . . . . . . . . 68 Basis set saturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 β-D-galactosidase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 BchB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68–71, 75 BchL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68–70, 72, 74–76 BchN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68, 69–71, 74–76 B-diketiminate ligand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Beamline. . . . . . . . . . . . . . . . . . . . . . . . . . .180–181, 183, 185 Beijerinck, M.W. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Bioinorganic chemistry . . . . . . . . 166, 192, 207, 249, 287 Biological nitrogen fixation . . . . . . . . . . . . . . . . . . . . . . . . 3–4 See also Genomic analysis of nitrogen fixation; Molecular biology and genetic engineering in nitrogen fixation Biophysical analysis, DPOR . . . . . . . . . . . . . . . . . . . . 73, 129 Biosynthesis. . . . . .32–33, 35, 38, 43, 51–52, 56, 59–61, 67–69, 72, 74, 76, 81–82, 87, 173, 240, 252, 304 Biphasic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42–43 Blue-white selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 B3LYP, see Hybrid exchange functionals BLYP functional . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278, 308 Bohr magneton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192, 208 Born–Oppenheimer theory . . . . . . . . . . . . . . . . . . . . . . . . 208 Bortels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 BP86, see Generalized gradient approximation (GGA) Bradford . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132, 137, 140 Broken symmetry (BS) . . . . . . . . 269–270, 272–273, 287, 293–309, 295 calculation . . . . . . . . . . . . . . . . . . . . . . 269–270, 272–273 See also Modeling of MoFe nitrogenase system with BS-DFT states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296–297 BS-DFT method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295–296 exchange-correlation functionals OPBE and OLYP . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 OPTX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .296 hyperfine properties, evaluation . . . . . . . . . . . . . . . . . 296 spin crossover, issue . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Bulen, W.A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Bulk magnetic moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 Bulky spectator ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 Burk, D. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Burris, H. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

C Calcium carbide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Capillary batch/microbatch method . . . . . . . . . . . . . . . 151 Carbon disulfide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119–120 monoxide (CO) . . . . . . . . . . . . 22, 107, 119–120, 150, 201–202 -nitrogen bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . 113–116 -oxygen bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119–120 -sulfur bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119–120 Carbonyl sulfide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Carnahan, J.E. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–5 Car-Parrinello molecular dynamics . . . . . . . . . . . . . . . . . 307 Carrier proteins . . . . . . . . . . . . . . . 12–14, 36, 38, 116, 120

NITROGEN FIXATION Index 315 Catalysis. . . . . . . . . . . . . . . . .12, 16–18, 20, 37, 69–75, 87, 116, 131, 134, 138–139, 148–150, 178–179, 240, 252 Cavity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33, 193 CD, see Circular dichroism (CD) Cell disruption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Cell-free extracts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5, 139 Central ligand . . . . . . . . . . . . . . . . . 272, 274, 288, 305–307 Chaperone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38, 294 Chimeric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 ChlN . . . . . . . . 68–76, 130–131, 133–135, 137, 138–142 Chloramine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Chloramphenicol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132, 135 Chlorophyll . . . . . . . . . . . . . . . . . . . . . 52, 67–69, 72, 75, 76 biosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68, 72 Chlorophyllide (Chlide). . . . . . . . . . . . . . . . . . . . . . . . . . . . .68 Chlorophyllide oxidoreductase (COR) . . . . . . . 68, 74–76 Chromosome . . . . . . . . . . . . . . . . . . . . . . . . . . . 53–54, 86, 90 Circular dichroism (CD) . . . . . . . . . . . . . . . . . . 70, 207–218 Citrate . . . . . . . . 106, 109, 116, 122, 132, 134–135, 152, 197, 241, 242, 245–246, 279 Citric acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242, 245–246 -phosphate treatment . . . . . . . . . . . . . . . . . . . . . . . . . . 242 Cleavage . . . . . . . . . . . 121, 131, 137, 139, 141–142, 229, 282, 307 Clostridium pasteurianum . . . . . . . . . . 3, 19, 82, 148, 155 Cluster . . . 10–17, 20–21, 31–36, 38–44, 56–59, 61–62, 68–76, 93–94, 129, 134–137, 141, 143, 148–150, 152, 162–163, 165, 173, 177, 178–179, 181, 191, 196–198, 207, 212–215, 222–226, 228–229, 239–241, 247, 250–253, 269, 272–276, 278, 280, 282–286, 294–296, 299–305, 308, 309 Cofactor . . . . . . . . . 11–12, 14–22, 35–36, 39, 50–52, 56, 59–60, 62, 69, 71, 76, 82, 94, 134, 138–140, 148, 178, 196–201, 212–213, 222, 224, 230, 239–248, 250–251, 274, 276, 283, 294–295, 300–304, 305–306, 308–309 See also Iron-molybdenum cofactor (FeMoco) Collaborative Computational Project No. 4 (CCP4) program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 Competency, see Azotobacter vinelandii, competency Competitive inhibitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Complexes . . . . . . . . . . . 10–11, 13–14, 19–20, 23, 32–35, 38–39, 44, 50–51, 56–57, 59, 68, 70–74, 81, 108, 113, 119, 122–124, 130–131, 133–134, 139–143, 147–149, 154–155, 159, 174, 177–178, 198, 202, 212, 222, 224, 229, 239–240, 268–269, 271, 279, 284 Component proteins . . . . . . . . . . . . 5–6, 10, 93, 106–107, 147–148, 150, 158, 195–196 Computational model . . . . . . . . 268, 271, 273–278, 280, 282–284 Congression . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87–88, 88, 90 Continuum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167–168, 274 See also Polarizable continuum model (PCM) Cooperativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Coordination number . . . . . . . . . . . . . . . . . . . . . . . . 170, 174 Core electrons . . . . . . . . . . . . . . . . . . . . . . . . . . 166, 168, 172 Correlation functional . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 COSMO solvation model . . . . . . . . . . . . . . . . . . . . . . . . . . 300 Coulomb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267–268 Covalency factor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .298 Coy Laboratory . . . . . . . . . . . . . . . . . . . . . . . . . 133, 151, 179 Creatine kinase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

phosphate. . . . .5, 105–110, 122–123, 131, 133, 138 Cryoprotection . . . . . . . . . . . . . . . . . . . . . . . . . .157, 161–162 CRYSOL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180–183, 187 Crystal growth . . . . . . . . . . . . . . . . . . . . . . . . . 150, 158, 161–162 harvesting . . . . . . . . . . . . . . . . . . . . . . . . . . . 151, 157, 161 loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157, 161 monochromator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 structure . . . . . . . . . 6, 10–11, 14, 147–148, 150, 166, 178–179, 187, 251, 255, 274–275, 294 Crystallography . . . . . . . . . . . . 14, 20, 147–163, 166, 173, 178–179, 183, 251, 255 See also X-ray crystallography Cuvette . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121, 133, 216, 257 Cyanide . . . . . . . . . . . . . . . . . . . 76, 106, 113–115, 119, 279 Cyclopropene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 CydAB I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 CydR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57–58 Cysteine desulfurase. . . . . . . . . . . . . . . . . . . . . . . . . . . . .34, 55

D Dark operative protochlorophyllide oxidoreductase (DPOR) . . . . . . . . . 68–76, 129–131, 133–143 ATPase activity ATP hydrolysis versus Pchlide reduction . . . . . . . 72 Chlide formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 BchL/ChlL analogous to nitrogenase catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69–70 ATP hydrolysis by ChlL/BchL . . . . . . . . . . . . . . . 70 cysteinyl ligands in [4Fe–4S] cluster formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 DPOR subcomplex formation, EPR/CD study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70–71 catalytic redox cycle of DPOR . . . . . . . . . . . . . . . . 72–74 ATPase activity of ChlL2 . . . . . . . . . . . . . . . . . . . . . 73 electron transfer processes . . . . . . . . . . . . . . . . . . . . 74 Pchlide binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 ternary complex formation . . . . . . . . . . . . . . . . . . . 74 enzyme catalysis of chlorophyll biosynthesis genes involved . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 POR/DPOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 substrate recognition/reduction by . . . . . . . . . . . 70–72 See also Methods for nitrogenase-like DPOR Databank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151–152 Data collection . . . . . . . . . . 150–151, 161, 174, 180–183, 185–187 Data processing . . . . . . . . . . 150, 161–162, 181, 183–185 DEAE sepharose . . . . . . . . . . . . . . . . . . . . . . . . . . . 95, 97, 102 Dehydrogenase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119, 139 Deloriah Jacobs (DJ), see OP strains Density functional correlation (DFC) . . . . . . . . . . . . . . 269 Density functional exchange (DFX) . . . . . . . . . . . . . . . . 268 Density functional theory (DFT) . . . . 267–287, 293–309 ab initio MO methods, advantages . . . . . . . . . 268–269 for electronic structure calculations, parameters size and adequacy of computational model . . . 271 size and quality of basis sets (GTO/STO) . . . 270 electrostatic interactions in a chemical system . . . . 267 ab initio MO methods vs. DFT . . . . . . . . . 268–269 DFX functionals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 hybrid exchange functionals, tuning of metal–ligand bonding . . . . . . . . . . . . . . . . . . . 268 Hohenberg–Kohn theorem/Kohn–Sham equations, basis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 iron centers with unpaired electrons, treatment

NITROGEN FIXATION

316 Index

Density functional theory (DFT) (continued) broken symmetry approximation . . . . . . . 269–270 modeling of FeMoco, see DFT modeling of electronic/geometric structures of FeMoco for weak interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 Deprotonation . . . . . . . . . . . . . . . . . . . . . . 33, 276–277, 302 Derepression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Derivative EPR spectrums . . . . . . . . . . . . . . . . . . . . . . . . . 193 DFC, see Density functional correlation (DFC) DFT, see Density functional theory (DFT) DFT modeling of electronic/geometric structures of FeMoco broken symmetry calculations. . . . . . . . . . . . . .272–273 computational model . . . . . . . . . . . . . . . . . . . . . . 273–278 coordination chemical computational models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277–278 FeMo-co structural features . . . . . . . . . . . . 275–277 spectroscopic calibration of DFT . . . . . . . . . . . 271–272 DFX, see Density functional exchange (DFX) Diamagnetic . . . . . . . . . . . . 12, 17, 32, 41, 192, 197–198, 213–214, 223, 225–226, 255, 277, 281, 308 Diazene . . . . . . . . . . . . . . . . . . . . . . . . . . 19, 22, 23, 200, 279 Diazotrophs . . . . . 10, 49–53, 56–60, 62, 81–82, 84, 293 Diazotrophy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49–50 Diffraction . . . . . . . . . . . . . . . . . . . . 150–151, 159, 162, 174 See also X-ray diffraction Dihydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106, 120 Dimer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32, 43, 177, 259 Dimeric units . . . . . . . . . . . . . . . . 11, 14, 21, 43, 70–73, 76 Dimethylaminobenzaldehyde . . . . . . . . . . . . . . . . . 107, 111 Dimethylformamide (DMF) . . . . . . . . . 88, 241–245, 247 Dimethyl sulfoxide (DMSO) . . . . . . . . . 82, 133, 142, 241 Dinitrogen . . . . . . . . 9–10, 22–23, 31, 52, 107–110, 148, 177, 199–200, 202, 239, 279 fixation methods biological nitrogen fixation with diazotrophs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Haber–Bosch process. . . . . . . . . . . . . . . . . . . . . . . . . .9 Dinitrogenase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5, 10 Disorder parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . 169–170 Distillation . . . . . . . . . . . . . . . 109, 111, 123, 242–243, 246 Disulfide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119–120 Dithionite . . . . . . . . 5, 12, 15, 17, 32, 36, 41–42, 69–70, 72–73, 106–109, 113–114, 121–122, 130–133, 135, 138–139, 141, 148, 157, 159–160, 162, 179, 186, 195–196, 202, 213–214, 216 DMF, see Dimethylformamide (DMF) Doppler effect. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .230, 300 Double-reciprocal recombination . . . . . . . . . . . . . . . . 85–86 DPOR, see Dark operative protochlorophyllide oxidoreductase (DPOR) DPOR complex formation . . . . . . . . . . . . . . . 133, 139–141 See also Gel permeation chromatography Dynamic switch protein. . . . . . . . . . . . . . . . . . . . . . . . . . . .134

E Edge jump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 Edge region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167–168 Edman degradation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .137 Electric dipole . . . . . . . . . . . . . . . . . . . . . . . . . . . 167–168, 208 Electrochemical study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 Electron balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 donor . . . . . . . . . . . . . . . . . . . . . . 5, 69, 72–73, 121, 131

transfer . . . . . . . . . . . . . . 11–13, 15–16, 20, 38, 57, 68, 70–71, 73–75, 131, 148–149, 178, 198, 229, 276, 280–281, 283 Electronic information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 See also Mössbauer spectroscopy Electronic relaxation. . . . . . . . . . . . . . . . . . . . . . . . . .225, 233 Electronic states. . . . . . . . . .228, 284, 295–296, 298, 300 Electronic structure . . . . . 165, 167–168, 174, 240, 249, 268, 270–272, 274, 284, 288, 295–296, 298, 300 Electron nuclear double resonance (ENDOR) . . . 17–18, 133, 137, 148, 162, 181, 198–199, 244, 246, 251, 272–273, 276–277, 280–283, 295, 302, 304, 308–309 Electron paramagnetic resonance (EPR) . . . . . . . 70, 129, 191–203 silent . . . . . 12, 17, 70, 197–198, 212, 218, 223, 277, 280–281 spectra of nitrogenase proteins . . . . . . . . . . . . . 196–198 [4Fe–4S] cluster . . . . . . . . . . . . . . . . . . . . . . . 196–197 M- and P-clusters . . . . . . . . . . . . . . . . . . . . . . 197–198 spectroscopy, see EPR spectroscopy tube . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194–195, 202 Electron spin resonance (ESR) . . . . . . . . . . . . . . . . . . . . . 192 See also Electron paramagnetic resonance (EPR) Elemental analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 Energy calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 Enzymatic systems with homology to nitrogenase COR, see Chlorophyllide oxidoreductase (COR) DPOR, see Dark operative protochlorophyllide oxidoreductase (DPOR) NifH/NifD, metabolism of methanogens . . . . . 75–76 synthesis of chlorophylls and bacteriochlorophylls . . . . . . . . . . . . . . . . . . . . . . 68 comparison of DPOR, COR, and nitrogenase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68 enzyme systems and reduction steps . . . . . . . . . . 68 Enzyme . . . . . . . . 4–6, 10, 12, 23, 31, 35, 52–53, 56–57, 68–70, 72, 74–76, 81–82, 85, 87, 93, 106, 110–111, 120, 131, 147–150, 202, 222–223, 229–230, 239, 249, 251, 294–295 EPR, see Electron paramagnetic resonance (EPR) EPR spectroscopy detection of redox states of paramagnetic metalloclusters . . . . . . . . . . . . . . . . . . . . . . . . . . 191 EPR spectrum formation . . . . . . . . . . . . . . . . . . . . . . . 192 affected by hyperfine coupling. . . . . . . . . . . . . . .193 isotropic/axial/rhombic derivative EPR spectrums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 X-band EPR spectrometer . . . . . . . . . . . . . 193–194 materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194–195 freezing samples in EPR tubes . . . . . . . . . . 194–195 quartz EPR tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 methods EPR spectra of nitrogenase proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196–198 preparing EPR samples of nitrogenase . . 195–196 trapping substrates . . . . . . . . . . . . . . . . . . . . . 198–202 Zeeman interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Escherichia coli . . . . . . . . . . . . . . . . . . . 58, 82, 123, 129, 222 ESR, see Electron spin resonance (ESR) Ethanesulfonic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Ethanol . . . . . . . . . . . . . . . . . 112–113, 117, 122, 132, 136, 142 Ethylene . . . . . . . . 22, 113, 116–118, 124, 212, 216, 277 Ethylene glycol. . . . . . . . . . . . . . . . . . . . . . . . . .212, 216, 277

NITROGEN FIXATION Index 317 EXAFS, see Extended x-ray absorption fine structure (EXAFS) Exchange-correlation functional. . . . . . . . . . . . . . . . . . . .296 Exchange-coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 Exchange functionals, see Density functional exchange (DFX) Exchange interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 Excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172, 271 Excited state. . . . . . . .209–210, 215, 226–228, 230–233, 271–273, 300 Ex situ assembly. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33–38 Extended x-ray absorption fine structure (EXAFS) . . . . . . . . . . 35–36, 41, 165–174, 251 Extracted FeMo-cofactor . . . . . . . . . . . . . 17, 94, 240–241, 246–247, 279, 284 Extraction . . . . . . . . . . . . . . . . . . . . . . 94, 116, 240–247, 258

F FEFF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 FeMo-cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 FeMo cofactor (FeMoco) . . . . 11–12, 14–22, 55, 59–60, 178, 196–201, 212–213, 224, 294–295, 301–304, 306, 308–309 covalent ligands of . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 -deficient . . . . . . . . . . . . . . . . . . . 33, 36, 38, 41, 43, 240 extraction . . . . . . . . . . . . . . . . . . . 94, 240–241, 246–247 isolation . . . . . . . . . . . . . . . . . . . . . . . . 240–242, 246–247 acid treatment of purified MoFe protein with NMF/DMF . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 aid to nitrogenase research . . . . . . . . . . . . . . . . . . 240 FeMoco extraction protocol. . . . . . . . . . . . . . . . .240 modifications of . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 strategy development, challenges . . . . . . . 241–242 precursor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35–36 Fe–N2 complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 Fe protein . . . . . . . . . . . . . . . . 5, 10–14, 17, 19–21, 31–32, 35–38, 41–42, 44, 54–55, 68, 93–100, 102, 107, 113, 177–179, 181, 184–187, 195–201, 212, 250 binding to nucleotides . . . . . . . . . . . . . . . . . . . . . . . 13–14 binding to MgATP. . . . . . . . . . . . . . . . . . . . . . . . . . .14 determination of dissociation constants, techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 crystal structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20–21 [4Fe–4S] cluster, redox properties . . . . . . . . . . . . 12–13 all-ferrous state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1+ oxidation state . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2+ oxidation state . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 protein conformational changes on binding to nucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 SAXS/X-ray crystallography, study. . . . . . . . . . . .14 purification of . . . . . . . . . . . . . . . . . . . . . . 95–96, 99–100 Fe proteincomplete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Fe proteinMo/homocitrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Ferredoxin . . . . . . . 12–13, 55, 57, 69, 73, 131, 139, 174 Ferric chloride. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119 Ferromagnetic species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Fe–S cluster . . . . . 10, 55, 57, 61–62, 70–72, 76, 148, 222, 250, 252–253, 269, 272–273, 278, 282–286, 294, 296, 300, 302–303, 309 complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222, 225, 231 [Fe4 S4 ] cluster . . . . . . . . . . . . . . . . . . . . . . 31–32, 41–43, 93

FeSII protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Flavodoxin . . . . . . . . . . . . . . . . . . . . . . . . . . . 13, 55, 106, 197 Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . 90, 116, 123, 172 Fluorophore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 FMO, see Frontier molecular orbitals (FMO) Formic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Fourier transform. . . . . . . . . . . . . . . . . . . . . . . . . . . . .169–171 Franck–Condon theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 Free-living nitrogen-fixing soil-bacteria . . . . . . . . . . . . . . . 3 A. agilis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Azotobacter chroococcum. . . . . . . . . . . . . . . . . . . . . . . . . . .3 Clostridium pasteurianum . . . . . . . . . . . . . . . . . . . . . . . . 3 Freeze-trapped MoFe protein . . . . . . . . . . . . . . . . . . . . . . . 22 Frontier molecular orbitals (FMO) . . . . . . . . . . . . . . . . . 270 Frozen pellets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Funnel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39–40 Fusion protein . . . . . . . . . . . . . . . . . . . . . . . . . . 129–131, 137, 139–142

G Gas chromatography . . . . . . . . . . . . . . . 107, 113–118, 120 Gate-keeping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Gauss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36, 192 Gaussian-type orbital (GTO) . . . . . . . . . . . . . . . . . . . . . . . 270 GDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Gel filtration . . . . . . . . . . . . . . . . . . . . . . . 179, 181, 185, 216 Gel permeation chromatography . . . . . 129, 137, 143n12 Generalized gradient approximation (GGA) . . . 268, 296 Genetic manipulation. . . . . . . . . . . . . . . . . . . .53, 59, 86, 88 Genetic stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Genetic transformation of A. vinelandii OP strains, use of . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84–85 preparation of competent A. vinelandii cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84–85 siderophore production, culture medium used . . . . 84 Genome scanning . . . . . . . . . . . . . . . . . . . . . . . . . . . 50–52, 56 Genome sequencing . . . . . . . . . . . . . . . . . . . . . 51, 53, 84, 89 Genomic analysis of nitrogen fixation identification of nitrogen-fixing species genome scanning . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 in silico predictions . . . . . . . . . . . . . . . . . . . . . . . . . . 50 model, see A. vinelandii, model for nitrogen fixation modifying genomes to study nitrogen fixation 58–62 examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59–60 purification of proteins using polyhistidine tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 remote genomic promoters, control of protein expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60–62 N2 fixation in the post-genomic era . . . . . . . . . . 49–53 in diazotrophs, challenges . . . . . . . . . . . . . . . . 49–50 genomics and identification of new diazotrophs. . . . . . . . . . . . . . . . . . . . . . . . . . .50–53 Genomic modification. . . . . . . . . . . . . . . . . . . . . . . . . . . 58–62 examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59–60 double recombination method . . . . . . . . . . . . . . . 59 metallocluster assembly and transfer by respiratory protection proteins . . . . . . . . . . . . . . . . . . . . . . . 59 NifB protein expression . . . . . . . . . . . . . . . . . . 59–60 nifH promoter for NifEN protein expression . . 59 purification of proteins using polyhistidine tags . . . 60 remote genomic promoters, control of protein expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60–62 control of genomics of arabinose catabolism. . .62 control of regulatory elements of sucrose metabolism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61

NITROGEN FIXATION

318 Index

Genomic modification (continued) remote loci for incorporating genes into the A. vinelandii genome. . . . . . . . . . . . . . . . . . . . .61 Geometric structure . . . . . . . . . . . . . . . . . . . . . . . . . . 270–278 See also DFT modeling of electronic/geometric structures of FeMoco Geometry optimizations . . . . . . . 270, 274, 278, 288, 297 G-factor . . . . . . . . . . . . . . . . . . . . . . . . . . . 207, 211–212, 218 GGA, see Generalized gradient approximation (GGA) Glassing agent. . . . . . . . . . . . . . . . . . . . . . . . . . .151, 161–162 GlnK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Glove box . . . . . . . . . . . 94, 179, 181, 183, 185–186, 194, 253–259, 261 Glucose-6-phosphate dehydrogenase . . . . . . . . . . . . . . . 139 Glutathione-S-transferase (GST) protein . . . . . . 129–131, 136–137, 139–142 Glycerol . . . . 12, 89, 135, 153–154, 157, 161–162, 179, 185–186, 197, 216 Glycolate . . . . . . . . . . . . . . . . . . . . . 273–274, 278, 288, 302 GNXAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 GroEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Ground state . . . . . . . . . . . . 209–210, 213, 217, 225–226, 230–232, 268, 270, 273, 278, 296 GTO, see Gaussian-type orbital (GTO) GTP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

H Haber–Bosch process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Hamilton syringe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180–182 Hamilton, W.D. . . . . . . . . . . . . . . . . . . . . . 17, 191, 196–197 Hanging drop vapor diffusion. . . . . . . . . . . . . . . . .151, 160 Hartree–Fock approximation . . . . . . . . . . . . . . . . . . . . . . 268 Hartree–Fock exchange (HFX) . . . . . . . . . . . . . . . . . . . . 268 HCl . . . . . . 101, 112–114, 116, 119, 151, 156, 240–246 He . . . . . . . . . . . . . . . . . 107, 116, 193, 210, 216–218, 300 Hellriegel, H. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Heterologous. . . . . . . . . . . . . . . . . . . .59, 82, 129–130, 134 Heterometal . . . . . . . . . . . . . . . . . . . . . . 10, 16, 50, 282, 284 Heterotetramer . . . . . . . . . 11, 14, 32, 68, 70–71, 93, 178 Hexane slurry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 HFX, see Hartree–Fock exchange (HFX) Hi-CO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 High resolution structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 High spin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39–40 α-195His . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18–19, 23, 201 α-442His . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 His triad . . . . . . . . . . . 211, 215, 225, 231–233, 255, 272, 296–297, 309 1 H NMR spectroscopy . . . . . . . . . . . . . . . . . . . . . . . 256–257 Hohenberg–Kohn theorem . . . . . . . . . . . . . . . . . . . . . . . . 267 Homocitrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11, 14, 16, 32–33, 35, 37–39, 55, 70, 94, 148, 178, 213, 239, 276–280, 287, 294, 301–302, 304 Homodimer. . . . . . . . . . . . . . .10–11, 31, 68, 93, 148, 177 Homolytic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Hybrid exchange functionals . . . . . . . . . . . . . . . . . . . . . . . 268 Hydrazine . . . . . . . . . . . . . 19, 22–23, 107, 110–112, 122, 200–202, 279 Hydrazoic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111–112 Hydrocarbon. . . . . . . . . . . . . . . . . . . . . . . . . . . .113, 115–117 Hydrogen bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . 250, 301 Hydrogenic radial distribution . . . . . . . . . . . . . . . . . . . . . 270 Hydrolysis . . . . . . . 11–14, 19–21, 37–38, 70, 72, 74–75, 106, 114–116, 122–124, 131, 134, 149, 177–178, 187, 239, 294

Hydrophobic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73, 142 Hyperfine coupling . . . . . . . . . . . . . . . . . 193, 273, 297–299 Hyperthermal environments . . . . . . . . . . . . . . . . . . . . . . . . 53

I IMAC, see Immobilized metal affinity chromatography (IMAC) IMAC sepharose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Imidazole . . . . . . . . . . . . 96, 101, 273–274, 277, 279, 302 Immobilized metal affinity chromatography (IMAC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Incubation . . . . . . . . . . 17, 35, 42, 44, 70, 84–85, 89, 99, 107, 121, 137, 141, 143, 187 Indigo disulfonate (IDS) . . . . . . . . . . . . . 12, 32, 35–36, 41 Indirect selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87–88 Indophenol . . . . . . . . . . . . . . . . . . . 108–110, 113–115, 120 Insertase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 In situ assembly . . . . . . . . . . . . . . . 33–34, 40–43, 134, 159 Integer spin state . . . . . . . . . . . . . . . . . . . 211–213, 218, 226 Intermediate . . . . . . . . . . 4, 17–18, 22–23, 33, 42–43, 60, 72–73, 114–115, 149–150, 165, 174, 198–201, 223, 229, 252, 277, 279–284, 296, 300, 306–308 Interstitial . . . . . . . 17, 33, 173, 273, 276–278, 282, 288, 294, 297, 302–305 Intramolecular . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283, 308 In vitro . . . . . . . . . . . . . . . 12, 15, 42, 69, 72, 94, 131, 134 Ion-exchange chromatography . . . . . . . . . . . . . . . . . . . . . 109 Ionic . . . . . . . . . . . . . . . . . . . . . . . . . 111, 114, 272, 287, 298 Ionization chamber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 Ionization threshold energy . . . . . . . . . . . . . . . . . . . . . . . . 169 Iron (Fe) complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 See also Fe–N2 complexes; Fe–S; Unsaturated complexes of iron(II) -nitrogenase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 -NO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299–300 protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 -sulfur cluster . . . . . . . . 10, 55, 57, 61–62, 70–72, 76, 148, 222, 250, 252–253, 269, 272–273, 278, 282–287, 294, 296, 300, 302–303, 309 Iron-molybdenum cofactor (FeMoco) . . . . . . . . . . . . . . 251 air sensitive, see Air-free techniques biosynthesis of M cluster (“FeMoco”) of molybdenum nitrogenases . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 molybdenum incorporation into model system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 N2 binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 DFT modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271–278 See also DFT modeling of electronic/geometric structures of FeMoco structural features . . . . . . . . . . . . . . . . . . . . . . . . . 275–277 charge state and redox forms . . . . . . . . . . . . . . . . 276 composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 oxidation states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 spin states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 See also FeMo cofactor (FeMoco) Iron–sulfur (Fe–S) cluster . . . . 10, 55, 57, 61–62, 70–72, 76, 148, 222, 250, 252–253, 269, 272–273, 278, 282–287, 294, 296, 300, 302–303, 309 IscS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61–62 IscU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60–61 Isocyanide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115–116

NITROGEN FIXATION Index 319 Isoelectronic interstitial atoms. . . . . . . . . . . . . . . . . . . . . .288 Isolation . . . . . . . . . . . . . . . . . 60, 88–89, 93, 150, 239–248 Isomer shift . . . . . . . . . 225, 231, 276–277, 299–300, 303 Isotopic analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Isotopomers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .276 Isotropic . . . . . . . . . . . . . . . . . 193, 208, 210–211, 232, 297

J J. Young NMR tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

K K-edge . . . . . . . . . . . . 35–37, 41, 166–168, 173–174, 272 Kinetics . . . . . . . . . . . . 13, 21, 70, 72, 107, 139, 222–223, 229–230, 280–282 Kjeldahl assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Klebsiella pneumoniae . . . . . . . . . . 5, 38, 49, 82, 148, 155 K2 [LMe Fe]2 N2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259–260 Kohn-Sham equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 Kramers . . . . . . . . . . . . . . . . . . . . . . 197, 211–212, 223, 233 Kramer’s doublets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 K3 -weighted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

L LacZ gene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Lawrence Berkeley National Laboratory (LBNL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 L127 Fe protein . . . . . . . . . . . . . . . . . . 179, 181, 184–187 LeComte, J.R. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5, 177 L-edge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166, 174 Legumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 127Leu. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 L126 GDVVCGGF134 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Lid loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39–40 Ligand . . . . . 12, 15–16, 32–33, 35–36, 39, 69, 75, 148, 168–169, 172, 174, 231–233, 249–250, 252–253, 268–270, 272, 274, 276–278, 280, 282–283, 288, 294, 296–297, 299–302, 305–309 Light atom X . . . . . . . . . . . . . . . . 16–17, 33, 149, 251, 294 Liquid-liquid diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Liquid nitrogen . . . . . 98, 124, 133, 138, 151, 161, 181, 186, 194–195, 198 Lithium chloride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 [LMe Fe]2 Cl2 ·2LiCl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258 [LMe Fe]2 N2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258–259 [LMe Fe]2 S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 Lo-CO. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .201–202 Luria-Bertani . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 Lysozyme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180, 182

M Macroscopic electrostatics with atomic detail (MEAD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301 Magnesium adenosine triphosphate (MgATP) . . . 10–11, 13–14, 17, 105 Magnet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216–218, 226 Magnetic circular dichroism (MCD) . . . . . . . 12, 207–218 Magnetic field . . . . . . . . . . . . 36, 192, 208–212, 216–218, 223–224, 230, 232 Magnetic hyperfine interactions . . . . . . . . . . 225, 232–233 Magnetization curve . . . . . . . . . . . . . . . 210–213, 215, 218

Manton-Gaulin homogenizer . . . . . . . . . . . . . . . . . . . . . . . 98 MarParse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181, 183, 187 Maturation . . . . . . . . . . . . . 11–12, 35–38, 41–44, 94, 151 MCD, see Magnetic circular dichroism (MCD) MCD spectroscopy basic concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . .207–210 bioinorganic research . . . . . . . . . . . . . . . . . . . . . . . 207 CD spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 MCD coupled with EPR spectroscopy . . . . . . . 207 MCD dispersion, types of interactions (A/B/C) . . . . . . . . . . . . . . . . . . . . . . . . . . 208–209 nondegenerate/degenerate transitions, energy diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 “rigid Shift” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 Zeeman interaction, effects. . . . . . . . . . . . . . . . . .209 magnetization curves . . . . . . . . . . . . . . . . . . . . . . 210–212 isotropic S=1 /2 spin system, equations . . . . . . . . . . . . . . . . . . . . . . . . . . 210–211 Kramer’s doublets and rhombicity . . . . . . 211–212 Kramer’s doublets (half-integer spin states) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211–212 non-Kramer’s and rhombicity . . . . . . . . . . . . . . . 212 “saturation,” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 S>1 /2 spin system, equations . . . . . . . . . . . . . . . 211 materials sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215–216 sample cuvette . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216–217 Nif proteins, examples of . . . . . . . . . . . . . . . . . . 212–215 M-cluster. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16, 197–198 MEAD, see Macroscopic electrostatics with atomic detail (MEAD) M-edge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 Mercaptoethanol . . . . . . . . . . . . . . . . . . . . . . . . 108, 115–116 Mercuric perchlorate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Metal active sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 Metal cluster . . . . . . . . . . 10–12, 32, 35, 44, 94, 148, 150, 162–163, 191, 196–197, 212 Metallocluster . . . . . . . . . . 11, 14–16, 31, 59, 61, 81, 191 proteins . . . . . . . . . . . . . . . . . . . 165, 174, 193, 211–212 Metalloenzyme . . . . . . . . . . . . . . . . . . . . . . . 10, 81, 239, 249 Methanocaldococcus jannaschii. . . . . . . . . . . . . . . . . . . . . .76 Methanogens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75–76 Methods for nitrogenase-like DPOR constructs for recombinant DPOR production . . . . . . . . . . . . . . . . . . . . . . . . . 129–131 DPOR activity assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 materials activity tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 DPOR complex formation . . . . . . . . . . . . . . . . . . 133 protein production . . . . . . . . . . . . . . . . . . . . . . . . . 132 protein purification . . . . . . . . . . . . . . . . . . . . 132–133 methods DPOR complex formation . . . . . . . . . . . . . 139–141 protein production . . . . . . . . . . . . . . . . . . . . . . . . . 135 protein purification . . . . . . . . . . . . . . . . . . . . 135–138 standard DPOR activity assay . . . . . . . . . . 138–139 ternary DPOR complex formation . . . . . . . . . 131, 134 Methylamine . . . . . . . . . . . . . . . . . . . . . . . 113, 115–116, 124 Methyl-coenzyme M reductase . . . . . . . . . . . . . . . . . . . . . . 76 Methyldiazene . . . . . . . . . . . . . . . . . . . . . . . . . 19, 22–23, 200 Methylene blue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12, 119 Methylthiolate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302 Methyl viologen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Metrical information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .168 Meyerhof. O. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

NITROGEN FIXATION

320 Index

MgADP-AlF4 - . . . . . . 70–71, 73–74, 134, 139, 141, 143 MgATP, see Magnesium adenosine triphosphate (MgATP) Mg ATP hydrolysis . . . . . . . . . . . . . . . . . . . . . 20–21, 37, 110 Michaelis-Menten kinetics . . . . . . . . . . . . . . . . . . . . . 72, 139 Microanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256 Microbatch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Microdistillation . . . . . . . . . . . . . . . . . . . 109, 115, 119, 123 of ammonia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Conway dish method . . . . . . . . . . . . . 123n1–124n1 glass-rod technique . . . . . . . . . . . . . . . . . . . . . . 123n1 Microrespirometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Microscope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159–160 Microwave frequency . . . . . . . . . . 192, 195–196, 199–201 Microwave power . . . . . . . . . . . . . . . . . . 195–196, 199–201 Midpoint reduction potential. . . . . . . . . . . . . . . . . . . . . . . .12 Miller assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Mixed spin system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 MN (resting state of FeMoco) . . 17, 197–198, 296, 298, 303–304 Mod1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Model complexes . . . . . . . . . . . . . . . . . . 149, 168, 173, 249 Modeling (functional/structural) of nitrogenase, techniques bulky spectator ligands, synthetic study. . . . .252–253 β-diketiminate ligand . . . . . . . . . . . . . . . . . . . . . . . 253 iron(III) formation . . . . . . . . . . . . . . . . . . . . . . . . . . 251n3 iron–sulfur cluster (4Fe–4S) spectroscopic study and reactivity of FeII 4S4 species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 synthesis with open coordination sites for better binding ability . . . . . . . . . . . . . . . . . . . . . . . . . . 252 materials air-sensitive techniques . . . . . . . . . . . . . . . . . 253–254 monitoring reactions using synthetic complexes, see UV–visible spectroscopy purification and characterization techniques . . . . . . . . . . . . . . . . . . . . . . . . . 255–256 specific starting materials . . . . . . . . . . . . . . . . . . . . 257 spectroscopic descriptions, abbreviations used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 M cluster iron-only nitrogenases (Fe8 ) . . . . . . . . . . . . . . . . 251 molybdenum-dependent nitrogenases (MoFe7 ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 vanadium-dependent nitrogenases (VFe7 ) . . . 251 methods K2 [LMe Fe]2 N2 . . . . . . . . . . . . . . . . . . . . . . . 259–260 [LMe Fe]2 Cl2 ·2LiCl . . . . . . . . . . . . . . . . . . . . . . . . 258 [LMe Fe]2 N2 . . . . . . . . . . . . . . . . . . . . . . . . . . 258–259 [LMe Fe]2 S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 P cluster (8Fe–7S) of molybdenum nitrogenases . . . . . . . . . . . . . . . . 250 in synthetic molecules, benefits . . . . . . . . . . . . . . 251 synthetic complexes ligand, definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 “model complexes”/“synthetic analogues” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 model study, disadvantage. . . . . . . . . . . . . . . . . . .249 protein modulation of metal reactivity . . . . . . . . . . . . . . . . . . . . . . . . . . . 249–250 unsaturated iron complexes . . . . . . . . . . . . . . . . . . . . . 252 Modeling of MoFe nitrogenase system with BS-DFT MoFe nitrogenase reaction . . . . . . . . . . . . . . . . . . . . . 293 FeMoco, active site of. . . . . . . . . . . . . . . . . . . . . . .294

MoFe7 S9 cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 Mössbauer and ENDOR spectroscopic data, analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 resting state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 X-ray structure, data analysis . . . . . . . . . . . . . . . . 295 reaction pathway . . . . . . . . . . . . . . . . . . . . . . . . . . 306–309 the ligand-bound state . . . . . . . . . . . . . . . . . 305–306 the resting state. . . . . . . . . . . . . . . . . . . . . . . .302–305 theoretical methodology BS-DFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295–296 hyperfine coupling parameters for X and 59 Fe . . . . . . . . . . . . . . . . . . . . . . . . . . . 297–299 Mössbauer parameters . . . . . . . . . . . . . . . . . 299–301 solvent and environmental effects. . . . . . .301–302 spin coupling/polarization and BS states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296–297 structural models . . . . . . . . . . . . . . . . . . . . . . . . . . . 302 Modeling of potential energy surface of nitrogen fixation nitrogenase function, theoretical/computational insights . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282–284 site-directed mutagenesis studies . . . . . . . . . . . 278–282 Thorneley–Lowe kinetic model . . . . . . . . . 279–282 Mo-dependent nitrogenase mechanism Fe protein cycle . . . . . . . . . . . . . . . . . . . . . . . . . . 20–21 Fe protein–MoFe protein complex formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19–20 MoFe–protein cycle . . . . . . . . . . . . . . . . . . . . . . 21–23 nitrogen fixation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9–10 dinitrogen fixation, methods . . . . . . . . . . . . . . . 9–10 two component proteins crystal structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Fe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11–14 MoFe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14–19 Modulation amplitude . . . . . . . . . . . . . . 195–196, 199–201 MoFe protein . . . . . . . . . 5, 10–23, 14–19, 31–44, 93–99, 101–102, 117, 177–178, 195–201, 213–215, 222–226, 228–229, 239–240, 242–243, 245–247, 276–281, 299, 304 crystal structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 FeMo cofactor (M-cluster), active site of nitrogenase . . . . . . . . . . . . . . . . . . . . . . . . . . . 16–19 extraction from MoFe protein . . . . . . . . . . . . . . . . 17 α-195His in nitrogenase catalysis, role . . . . . 18–19 identification of central atom X, ENDOR studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 oxidation/reduction . . . . . . . . . . . . . . . . . . . . . . . . . 17 as substrate binding site . . . . . . . . . . . . . . . . . . . . . . 18 substrates for nitrogenase . . . . . . . . . . . . . . . . . . . . 18 X-ray structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 P-clusters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15–16 oxidation of MoFe protein, effects . . . . . . . . 15–16 X-ray structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 purification of . . . . . . . . . . . . . . . . . . . . 95, 98–99, 102n4 See also Assembly of nitrogenase MoFe protein MoFe7 S9 cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 Molecular biology and genetic engineering in nitrogen fixation materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82–83 antibiotics used in direct selection with A. vinelandii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 methods A. vinelandii competency . . . . . . . . . . . . . . . . 84–85 culturing A. vinelandii . . . . . . . . . . . . . . . . . . . 83–84 selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85–89 transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

NITROGEN FIXATION Index 321 nitrogenase synthesis model systems (diazotrophs), study . . . . . . . . . . . 82 nif genes required, complexity . . . . . . . . . . . . 81–82 Molecular-orbital . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268, 270 Molecular replacement methods . . . . . . . . . . . . . . . . . . . 151 Molybdenum . . . 4–5, 10, 31, 37, 50, 54–56, 62, 71, 93, 105, 148, 173, 177, 239, 250–252, 293–294, 307 Molybdenum–iron (MoFe) protein . . . . . . . . . . . . . . . . . . . 5 Mo-nitrogenase . . . . . . . . . . . . . . . . . . . . 107, 110–120, 124 Monodentate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Monomeric. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Monooxygenase . . . . . . . . . . . . . . 165, 222–223, 230, 296, 300 MoO42– . . . . . . . . . . . . . . . . .35, 37, 82, 95, 152–153, 157 Mortenson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Mössbauer . . . . . . . . 12, 15, 17, 148, 174, 221–233, 255, 271–272, 276–277, 295, 299–300, 302–303, 309 parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299–301 isomer shifts . . . . . . . . . . . . . . . . . . . . . . . . . . . 299–300 quadrupole splittings . . . . . . . . . . . . . . . . . . 300–301 spectrometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 spectroscopy, see Mössbauer spectroscopy Mössbauer spectroscopy applications to nitrogenase MoFe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223–226 Fe centers, types . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 M and P clusters, X-ray crystallographic studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 M center, EPR/Mössbauer/X-ray crystallographic study . . . . . . . . . . . . . . 224–225 Mössbauer spectra of MoFe protein from A. vinelandii . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 P clusters, high-field Mössbauer/X-ray crystallographic study . . . . . . . . . . . . . . 225–226 spectral components . . . . . . . . . . . . . . . . . . . . . . . . 223 development of biological applications, stages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221–222 in conjunction with rapid-mix rapid-freeze technique . . . . . . . . . . . . . . . . . . . . . . . . . . 222–223 energy scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 Doppler effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 hyperfine interactions . . . . . . . . . . . . . . . . . . . . . . . 230 hyperfine interactions and Mössbauer spectral type isomer shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 magnetic hyperfine interaction and magnetic spectrum. . . . . . . . . . . . . . . . . . . . . . . . . . .232–233 quadrupolar interaction . . . . . . . . . . . . . . . . 231–232 materials basic Mössbauer spectrometer setup . . . . . . . . . 226 radiation source . . . . . . . . . . . . . . . . . . . . . . . 226–227 sample preparation. . . . . . . . . . . . . . . . . . . . . . . . . .228 methods electronic properties/number of Fe atoms, determination . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 electronic states of Fe ions, information extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 functional role of specific residue, investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 identification of substrate/inhibitor binding site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 kinetic/spectroscopic characterization of reaction intermediates . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 organization of Fe atoms in multiple Fe centered proteins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .229

redox potential of Fe centers, determination . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 nuclear resonance technique . . . . . . . . . . . . . . . . . . . . 221 MOx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17, 197, 303 Multiple scattering paths . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Mutant vinelandii (MV), see OP strains Myo-inositol oxygenase . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

N NafY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Na2 S2 O4 . . . . . . . . . . . . . . . . . . 95–96, 101, 102, 195, 247 Nesting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 N2 -fixing genomes, analysis bioinformatic analysis . . . . . . . . . . . . . . . . . . . . . . . . 52–53 in silico pathway identification . . . . . . . . . . . . . . . . 52 genome scanning using NifD as query . . . . . . . . 51–52 biosynthesis of nitrogenase proteins, pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51–52 NifD homologs sequences from A. vinelandii/ P. stutzeri genomes . . . . . . . . . . . . . . . . . . . . . . . 51 paralogous genome sequences . . . . . . . . . . . . . . . . 51 genome scanning using NifH as query . . . . . . . . 50–51 BLAST hits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50–51 phylogenetic analysis in diazotrophs . . . . . . . 50–51 sequence tag/barcode, identification of nitrogen fixers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 NflH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75–76 Nickel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 NifA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54–56, 58 NifB . . . . . . . . 33–36, 38–44, 54–56, 59–60, 81, 94, 240 nifB MoFe protein. . . . . . . . . . . . . . . .36, 38–44, 94, 240 NifD . . . . . . . . . 14, 32, 35, 51–52, 55, 57–58, 68, 75–76 NiFe hydrogenase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 NifEN . . . . . . . . . . . . . . . . . . . . . . . 33–39, 41, 44, 54, 59–60 NifENFeMoco. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36, 38 Nif genes . . . . . . . . . . . . . . . . . . . . . . . . 10, 54, 56, 62, 81, 84 NifH. . . . .11, 33, 35, 41–43, 50, 52, 54–55, 57–59, 68, 75–76, 81, 94, 177, 212 nifH MoFe protein . . . . . . . . . . . . . . . . . . . . . . . . 41–43, 94 NifL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55–56, 58, 75–76 NifM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54–56, 81 Nif proteins, characterization by MCD spectroscopy . . . . . . . . . . . . . . . . . . . . . . . 212–215 half-integer states . . . . . . . . . . . . . . . . . . . . . . . . . 212–213 EPR spectrum of FeMoco . . . . . . . . . . . . . . . . . . 213 MCD spectrum of FeMoco . . . . . . . . . . . . . . . . . 213 nesting of magnetization curves . . . . . . . . . . . . . 213 integer states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213–215 magnetization curves of the P-cluster of thionite oxidized state . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 MCD spectrum of the P-cluster of dithionite reduced state . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 MCD spectrum of the P-cluster of thionite oxidized state . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 P-cluster, oxidation states . . . . . . . . . . . . . . . . . . . 213 S = 1 /2 states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 NifS . . . . . . . . . . . . . . . . . . . . . . . . . 33–34, 55–56, 61–62, 81 NifU. . . . . . . . . . . . . . . . . . . . .33–34, 40, 54–56, 60–62, 81 NifV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33, 54–56, 81 NifX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38, 55, 304 NifY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38, 55 NifZ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33, 41, 43–44, 55 Ninhydrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Nitrido ligand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 Nitrite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112–113, 122

NITROGEN FIXATION

322 Index

Nitrogenase . . 3–6, 9–24, 31–44, 50–59, 61–62, 67–76, 81–84, 87–89, 93–102, 105–124, 129–143, 147–152, 156, 158–162, 166, 173–174, 177–179, 184, 191, 193–196, 198–202, 212–215, 223–226, 239–247, 249–261, 267–287, 293–309 Nitrogenase-catalyzed reactions involving carbon–carbon bonds . . . . . . . . . . . . 116–118 acetylene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116–117 allene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 butynes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 cyclopropene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 ethylene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117–118 propyne and propargyl alcohol . . . . . . . . . . . . . . 117 involving carbon–nitrogen bonds . . . . . . . . . . 113–116 alkyl and alkenyl cyanides . . . . . . . . . . . . . . 114–115 cyanamide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 cyanide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113–114 methyl isocyanide . . . . . . . . . . . . . . . . . . . . . . 115–116 involving carbon–oxygen and carbon–sulfur bonds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119–120 carbon dioxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 carbon disulphide . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 carbonyl sulphide . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 cyanate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 thiocyanate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119–120 involving hydrogen . . . . . . . . . . . . . . . . . . . . . . . 120–121 HD formation . . . . . . . . . . . . . . . . . . . . . . . . . 120–121 proton reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 involving nitrogen–nitrogen bonds . . . . . . . . 107–112 azide and hydrazoic acid . . . . . . . . . . . . . . . 111–112 dinitrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107–110 hydrazine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110–111 involving nitrogen–oxygen bonds . . . . . . . . . . 112–113 nitrite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112–113 nitrous oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Nitrogenase-like enzyme, see Dark operative protochlorophyllide oxidoreductase (DPOR) Nitrogenase mechanism Fe protein cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20–21 MgATP hydrolysis and electron transfer . . . . . . . 20 Fe protein–MoFe protein complex formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19–20 complexes identified in the associated state, analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19–20 MoFe–protein cycle. . . . . . . . . . . . . . . . . . . . . . . . . .21–23 characterization of freeze-trapped MoFe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22–23 Lowe–Thorneley model . . . . . . . . . . . . . . . . . . . . . . 21 substrate binding to MoFe, inhibition patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 variants of MoFe protein and EPR parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Nitrogenase research, history biological nitrogen fixation, discovery (1888) . . . 3–4 free-living nitrogen-fixing soil-bacteria . . . . . . . . . 3 isolation of bacteria from root nodules . . . . . . . . . 3 Kjeldahl assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 in legumes and influence on its growth . . . . . . . . . 3 identification of binary nature . . . . . . . . . . . . . . . . . . . . . 5 investigations on a molecular level since 1990s . . . . . 6 crystal structures, reports . . . . . . . . . . . . . . . . . . . . . . 6 Thorneley–Lowe model . . . . . . . . . . . . . . . . . . . . . . . 6 major advances before 1960. . . . . . . . . . . . . . . . . . . . . . . 4 nitrogen-fixing cell-free enzyme preparation (1960) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–5

ATP and ATP-regenerating system . . . . . . . . . . . . . 5 from A. vinelandii (aerobic organism) . . . . . . . . . 5 from C. pasteurianum (Carnahan) . . . . . . . . . . . . . 5 from Klebsiella pneumoniae (anaerobic organism) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Nitrogenase structure/function relationships by DFT DFT modeling of electronic/geometric structures of FeMoco broken symmetry calculations . . . . . . . . . . 272–273 computational model . . . . . . . . . . . . . . . . . . 273–278 spectroscopic calibration of DFT . . . . . . . 271–272 modeling of potential energy surface of nitrogen fixation nitrogenase function, theoretical/computational insights . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282–284 site-directed mutagenesis studies . . . . . . . 278–280 Nitrogen fixation . . . . . 3–6, 9–24, 31–44, 49–62, 67–76, 81–90, 116, 149, 177, 278–284 Nitrous oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 N-methylformamide (NMF) . . . . . . . . 240–247, 276, 279 NMF, see N-methylformamide (NMF) NMR spectroscopy . . . . . . . . . . . . 147, 207, 256, 258–260 Nodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3, 120 NRVS, see Nuclear resonance vibrational spectroscopy (NRVS) Nuclearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41, 222, 229 Nuclear resonance vibrational spectroscopy (NRVS) . 17, 251 Nucleophilicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283–284 Nucleotide . . . . . . . . . . . 10, 13–14, 20, 70, 74, 123, 134, 140–141, 149–150, 156, 160, 177–179, 181, 183, 185, 187 binding . . . . . . . . . . . . . 13–14, 74, 149–150, 177–178

O O-nitrophenyl galactoside . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Open coordination sites . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 O-phthalaldehyde . . . . . . . . . . . . . . . . . . 108, 112, 115–116 OP strains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 OPTX exchange functional . . . . . . . . . . . . . . . . . . . . . . . . 294 Orbital angular momentum . . . . . . . . . . . . . . . . . . . 192, 217 Orthologous system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Oxidation . . . . . . 12–13, 15–17, 20–21, 32–33, 41, 108, 114, 122, 148, 153, 166, 168, 173, 196–198, 207, 213, 216, 225–226, 228–229, 231–232, 250, 272, 274, 276–277, 281, 287, 295–296, 298–299, 302, 308 state . . . . . . . . . . . . . . . . . 12–13, 15–17, 20–21, 32–33, 41, 153, 166, 168, 173, 196–198, 207, 213, 225, 228–229, 232, 272, 274, 276–277, 281, 295–296, 298–299, 302–303, 308 Oxidative phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Oxidoreductase . . . . . . . . . . . . . . . . . . . 53, 68–74, 129–143 Oxygen . . . . . . . . . . 16, 53, 56–58, 62, 94, 101, 112–113, 119–120, 134–135, 150, 158, 162, 165, 170, 179–180, 194–195, 202, 218, 222–223, 253 exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

P P1+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15, 197 P2+ . . . . . . . . . . . . . . . . . . . . . . . . . . . 15–16, 42–44, 197, 215 P3+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15–16, 197 Parallel mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32, 197 Paralogous gene sequences . . . . . . . . . . . . . . . . . . . . . . 51–52

NITROGEN FIXATION Index 323 Paramagnetic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12, 17, 32, 41, 70, 129, 191–203, 210, 212–213, 215, 218, 226, 255–257, 282, 285, 308 Path length . . . . . . . . . . . . . . . . . . . . . . . . 117, 166, 208, 216 PBE functional . . . . . . . . . . . . . . . . . . . . . 278, 306–307, 309 P-cluster . . . . . . . . . . . . . 11, 14–17, 20–21, 32–35, 40–44, 70–71, 76, 93, 148, 152, 163, 178, 196–198, 213–215, 239–240, 247 PCM, see Polarizable continuum model (PCM) PDB. . . . . . . . . . . . . . . . . . . . . . . . .11, 15–16, 152–156, 187 Periodic plane-wave DFT methods . . . . . . . . . . . . . . . . . 306 Perpendicular mode EPR . . . . . . . . . . . . . . . . . . . . . . . . . . 197 pH . . . 16, 82, 95–97, 99, 106–110, 114, 116, 120–121, 124, 132, 150–156, 159–160, 179–181, 198, 202, 241–245, 247 Phase shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169–170 Phenotypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Phosphate . . . 5, 19–20, 34, 72, 82, 105–110, 122–123, 131, 133–134, 138–139, 241–242, 245 Phosphokinase . . . . . . . . . . 105–107, 122–123, 131, 133, 138–139, 142, 198 Phosphomolybdate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Photoabsorber . . . . . . . . . . . . . . . . . . . . . 167–168, 172–174 Photoelectron . . . . . . . . . . . . . . . . . . . . . . 167–169, 172–173 wave vector space . . . . . . . . . . . . . . . . . . . . . . . . . 169, 172 Photomultiplier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 Phylogenetic analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51, 75 P-loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69–70, 76 P-loop Y8 GKGGIGK15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 PMSF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 PN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33, 197–198, 213 Point mutation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85–87 Polarizable continuum model (PCM) . . . . . . . . . . . . . . 274 Polyhistidine tag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Potassium phosphate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Precipitating agents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .151 Precursor . . . . . . . . . . . . . . . . . . . . . . . . 36–39, 41–44, 55, 59 Pre-edge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167–168, 172 Prismatic iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Propargyl alcohol . . . . . . . . . . . 22–23, 117, 198–199, 278 Propene. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115, 117–118 Protease . . . . . . . . . . . . . 55, 102, 131–132, 137, 139–142 Protein concentration . . . . . . . . 107, 150, 158, 181, 186, 195 conformation . . . . . . . . 14, 38–40, 42–44, 57, 60, 62, 69–70, 74, 134, 149, 177–179 expression . . . . . . . . . . . . . . . . . . . . . . . . . . . 60–62, 82, 84 interaction . . . 38, 60, 71, 73–75, 131, 134, 150, 174 ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148, 276 preparation . . . . . . . . . . . . . . . . . . . . . . . . . . 134, 150–157 Protochlorophyllide (Pchlide) . . . . . . . . . . . . . 68, 129–143 Protocols for cofactor isolation of nitrogenase materials alternative citrate–phosphate–pyrrolidinone extraction method . . . . . . . . . . . . . . . . . . 242–243 standard HCl–NaOH–NMF extraction method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 methods alternative citrate–phosphate–pyrrolidinone extraction method . . . . . . . . . . . . . . . . . . 245–246 distillation of NMF and DMF . . . . . . . . . . . . . . . 243 standard HCl–NaOH–NMF extraction method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243–245 nitrogenase catalysis, electron flow for FeMoco, role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240

Proton . . . 16, 18, 22–23, 118, 120, 149, 198–199, 276, 278–282, 294, 302, 309 Purification . . . . . . . . . . . . . . . . . . . . . 5–6, 60, 93–102, 109, 124, 130–133, 135–138, 141–143, 150, 181, 185–186, 255–256 Purification of nitrogenase proteins anaerobic techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 extended purification procedures, effects . . . . . . . . . 94 IMAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 materials cell growth and crude extract preparation . . . . . 95 purification of His-tagged nitrogenase protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 purification of non-tagged Fe protein . . . . . 95–96 purification of non-tagged MoFe protein. . . . . .95 metal clusters/its orientation activity of nitrogenase . . . . . . . . . . . . . . . . . . . . 93–94 methods cell growth and crude extract preparation . 97–98 non-tagged nitrogenase proteins, purification strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 purification of His-tagged nitrogenase protein . . . . . . . . . . . . . . . . . . . . . 96–97, 100–101 purification of non-tagged Fe protein . . . . 99–100 purification of non-tagged MoFe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98–99 Schlenk system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 WAEC protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Pyrrolidinone-based extraction . . . . . . . . . . . . . . . . . . . . . 242 Pyruvate formate-lyase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

Q Q-sepharose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102n7 Quadrupolar interaction . . . . . . . . . . . . . . . . . . . . . . 231–232 Quadrupole doublet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 Quadrupole splitting . . . . . . . . . . . . . . . 225, 232, 299–301 Quantum mechanical tunneling . . . . . . . . . . . . . . . . . . . . 308 Quantum phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 Quartz . . . . . . . . . . . . . . . . . . . . . . . 133, 194, 216, 255, 257

R Random phase approximation (RPA) . . . . . . . . . . . . . . . 268 Rapid-mix rapid-freeze quench (RFQ) technique . . . . . . . . . . . . . . . . . . . . . 222–223, 229 RCSB, see Research Collaboratory for Structural Bioinformatics (RCSB) Reaction intermediates . . . . . . . . . . . . . . . . . . 150, 223, 229 Reaction pathway . . . 10–11, 42, 50–53, 60–61, 81, 111, 279, 283, 302–309 Reaction stoichiometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Recoilless nuclear γ-emission . . . . . . . . . . . . . . . . . . . . . . . 221 Recombinant DPOR production, constructs for . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129–131 Reconstituted MoFe protein . . . . . . . . . . . . . . . . . . . . 33, 43 Redox cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72–73, 75, 131 potential . . . . . . . . . . . . . . . . . . 229, 295, 301–304, 309 state . . . 12–13, 22, 56, 73, 150, 173, 191, 198, 229, 302 Reductase . . . . . . . 5, 10, 20, 36, 38, 50, 55, 59, 76, 165, 222–223, 230, 300 Reduction intermediates . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 Rees, D.C. . . . . . . . . . . . . . . . . . . . . . . 6, 148–150, 174, 294 REFMAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

NITROGEN FIXATION

324 Index

Resazurin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Research Collaboratory for Structural Bioinformatics (RCSB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Residue . . . . . . . . . 15, 18–19, 38, 71, 123, 230, 278–281 Resistance cassettes . . . . . . . . . . . . . . . . . . . . . . . . . . 59, 86–87 Resonance delocalization . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Resting-state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148, 150 Rfn1 genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 RFQ technique, see Rapid-mix rapid-freeze quench (RFQ) technique RhdA protein. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Rhizobium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Rhombic. . .17, 193, 196–198, 202, 207, 211–212, 218 Rhombicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 “Rhombo” program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 Ribonucleotide reductase . . . . . . 165, 222–223, 230, 300 “Rigid Shift” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 Ring. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68, 72, 74, 216 Rnf proteins, types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 RPA, see Random phase approximation (RPA) RPBE functional . . . . . . . . . . . . . . . . . . . . . . . . 278, 296, 306 R-space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169, 173 Rubber septum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194

S S-adenosyl-l-methionine . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Salt concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Sample concentration. . . . . . . . . . . . . . . . . . . .186, 208, 216 SAXS, see Small angle X-ray scattering (SAXS) SAXS spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 177–187 catalytic mechanism by nitrogenase . . . . . . . . . . . . . 178 gated and unidirectional electron transfer . . . . 178 induction of conformational changes . . . 177–178 MgATP binding and hydrolysis coupled to Fe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 materials beamline preparation . . . . . . . . . . . . . . . . . . . . . . . 180 data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 data processing, analysis, and 3D shape reconstruction . . . . . . . . . . . . . . . . . . . . . . . . . . 181 initial calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 initial sample preparation . . . . . . . . . . . . . . . . . . . 179 nucleotide addition and data collection . . . . . . 181 sample preparation at the synchrotron center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 methods beamline preparation . . . . . . . . . . . . . . . . . . . . . . . 182 data collection . . . . . . . . . . . . . . . . . . . . . . . . . 182–183 data processing, analysis, and 3D shape reconstruction . . . . . . . . . . . . . . . . . . . . . 183–185 initial calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 initial sample preparation . . . . . . . . . . . . . . . . . . . 181 nucleotide addition and data collection . . . . . . 183 sample preparation at the synchrotron center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 MgATP-bound Fe protein state, study . . . . . . . . . . 179 molybdenum nitrogenase. . . . . . . . . . . . . . . . . . . . . . .177 vs. XAS, applications. . . . . . . . . . . . . . . . . . . . . . .178–179 Scaffold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35, 55, 253 Scatterer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169–171, 174 Scattering curve . . . . . . . . . . . . . . . . . . . . 182–184, 186–187 Schlenk system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96, 101n2 Schrödinger equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 Scr-promoter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 ScrR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

ScrX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Second sphere coordination shells . . . . . . . . . . . . . . . . . . 309 Serial dilutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86, 88 Siderophore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56, 58, 84 Single scattering paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Site directed mutagenesis . . . . . . . . 18, 70, 202, 223, 230, 279, 285 Sitting drop vapor diffusion . . . . . . . . . . . . . . 151, 158–159 Slater-type orbital (STO) . . . . . . . . . . . . . . . . . . . . . . . . . . 270 Small angle x-ray scattering (SAXS) . . . . . . . . . 14, 40, 42, 177–187 Sodium azide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 cyanide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Specific activity . . . . . . . . . . . . . . . . . . . . . . . . . 42, 96–97, 138 Spectroscopic calibration . . . . . . . . . . . . . . . . . . . . . . 271–272 Spin collinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 coupling. . . . . . . . . . . . . . . . . . .228, 272, 295–297, 305 dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 –orbit coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 polarization . . . . . . . . . . . . . . . . . . . . . . . . . . 283, 295–297 projection techniques . . . . . . . . . . . . . . . . . . . . . . . . . . 296 state . . . . . . . . . 12, 33, 166, 196–201, 211–215, 218, 226, 232, 270, 272–273, 277–278, 281, 286, 295–297, 302 Stanford Synchrotron Radiation Laboratory (SSRL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180, 187 Steady-state assays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Stepwise assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34, 43 Stereospecificity. . . . . . . . . . . . . . . . . . . . . . . . . . . .68, 74, 116 Steric . . . . . . . . . . . . . . . . . . . . . 39, 250, 274–276, 278, 285 STO, see Slater-type orbital (STO) Stoichiometry . . . . . . . . . . 10, 33, 72, 113, 121, 130–131, 134, 293 Streptomyces thermoautotrophicus. . . . . . . . . . . . . . . . . . . . .50 Structure refinement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .162 Substrate . . . . . . . . . . . . . . . . . . . . 10, 12, 17–19, 22–23, 32, 40, 68, 70–75, 106–107, 110–113, 117–122, 131, 138–140, 149–150, 198, 200, 229, 240, 252, 276, 278–279, 284, 295–296, 308–309 Subunit . . . . . . . . . . 10, 12–16, 31–38, 42–43, 55, 68–70, 72–76, 93–94, 129, 148, 222, 230, 239, 298 Sucrose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Sulfur . . . . . . . . . . . . . . . . 5, 10, 32, 34, 53, 55, 58, 61, 71, 109, 116, 119, 134, 147–148, 170, 173, 231, 250, 252–253, 257, 260, 269, 299–300, 302, 307–308 Superconducting magnet system . . . . . . . . . . . . . . . . . . . 226 Switch/lock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39–40 Symbiont . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Synchrotron . . . . . . . . . . . . . . . . . . 171, 174, 179–181, 187 Synthetic analogues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .249

T Techniques for functional/structural modeling of nitrogenase bulky spectator ligands, synthetic study. . . . .252–253 β-diketiminate ligand . . . . . . . . . . . . . . . . . . . . . . . 253 iron(III) formation . . . . . . . . . . . . . . . . . . . . . . . . . . 261n3 iron–sulfur cluster (4Fe–4S) spectroscopic study and reactivity of FeII 4 S4 species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

NITROGEN FIXATION Index 325 synthesis with open coordination sites for better binding ability . . . . . . . . . . . . . . . . . . . . . . . . . . 252 materials air-sensitive techniques . . . . . . . . . . . . . . . . . 253–254 monitoring reactions using synthetic complexes, see UV–visible spectroscopy purification and characterization techniques . . . . . . . . . . . . . . . . . . . . . . . . . 255–256 specific starting materials . . . . . . . . . . . . . . . . . . . . 257 spectroscopic descriptions, abbreviations used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 M cluster iron-only nitrogenases (Fe8 ) . . . . . . . . . . . . . . . . 251 molybdenum-dependent nitrogenases (MoFe7 ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 vanadium-dependent nitrogenases (VFe7 ) . . . 251 methods K2 [LMe Fe]2 N2 . . . . . . . . . . . . . . . . . . . . . . . 259–260 [LMe Fe]2 Cl2 ·2LiCl . . . . . . . . . . . . . . . . . . . . . . . . 258 [LMe Fe]2 N2 . . . . . . . . . . . . . . . . . . . . . . . . . . 258–259 [LMe Fe]2 S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 P cluster (8Fe–7S) of molybdenum nitrogenases . . . . . . . . . . . . . . . . 250 in synthetic molecules, benefits . . . . . . . . . . . . . . 251 synthetic complexes ligand, definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 “model complexes”/“synthetic analogues” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 model study, disadvantage. . . . . . . . . . . . . . . . . . .249 protein modulation of metal reactivity . . 249–250 unsaturated iron complexes . . . . . . . . . . . . . . . . . . . . . 252 Temperature controller . . . . . . . . . . . . . . . . . . 159–160, 182 Tetrahydrofuran . . . . . . . . . . . . . . . 247, 254, 257–258, 260 Tetramer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 Thionine oxidation study . . . . . . . . . . . . . . . . . . . . . . . . . . 225 Thiophenolate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Thorneley–Lowe model . . . . . . . . . . . 6, 21, 280–282, 284 Three-neck vacuum flask . . . . . . . . . . . . . . . . . 242–243, 245 Titanium(III) citrate. . . . . . . . . . . . . . . . . . . . . . . . . .106, 122 TonB-type siderophore receptors . . . . . . . . . . . . . . . . . . . . 56 Transformation . . . . . . . . . . . . . . . . . . . 82, 84–88, 250, 293 See also Genetic transformation of A. vinelandii Transition energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208, 231 metal . . . . . . . . . . . . . . . . . 165–166, 268–269, 295, 298 -state analog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Transposon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Trapping substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198–202 alkynes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198–199 carbon monoxide (CO) . . . . . . . . . . . . . . . . . . . 201–202 diazenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 dinitrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 hydrazine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200–201 protons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198–199 Trichloroacetate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106, 109 Truncation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271, 274 Turnover . . . . . . 22–23, 32, 40, 139, 150, 198–201, 277

U Ultracentrifugation . . . . . . . . . . . . . . . . . . . . . . . . . . . 132, 136 Unpaired electron . . . . . . . . . . . . . 192–193, 232, 269, 296 Unsaturated complexes of iron(II) K2 [LMe Fe]2 N2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .252 [LMe Fe]2 N2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

[LMe Fe]2 S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 UV–visible spectroscopy . . . . . . . . . . . . . . . . . . . . . . 256–257

V Vacuum distillation . . . . . . . . . . . . . . . . . . . . . . . . . . . 242–243 Vacuum pump . . . . . . . . . . . . . . . . . . . . . . . . . . 194, 254, 258 Vanadium . . . . . . . . . . . . . . . . . . . . . . . . 50, 56, 94, 105, 251 Van der Waals interactions . . . . . . . . . . . . . . . . . . . . 269, 301 Vectors . . . . 82, 130–131, 136, 169–170, 172, 211, 296, 298, 304 Velocity transducer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 VFe protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94, 107, 251 Vnf genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54–56 V-nitrogenase . . . . . . . . . . . . 107, 111, 113–115, 117, 124 VTVHMCD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

W WAEC, see Weak anion exchange chromatography (WAEC) Water . . . . . . . . . . 74, 100, 109, 112–113, 116, 121, 124, 132–133, 136, 142, 162, 182, 186, 250, 254, 308 Weak anion exchange chromatography (WAEC) . . . . . 94 Wheaton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179, 185 Wigner–Eckart theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Wilfarth. H. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

X XANES, see X-ray absorption near edge (XANES) X-band EPR spectrometer . . . . . . . . . . . . . . . . . . . . 193–194 X-gal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85, 88 X-ray absorption near edge (XANES). . . . .166–168, 172 X-ray absorption spectroscopy (XAS) . . . 35–37, 41, 148, 165–174, 207, 271–272, 276 absorbing atom and XAS edges atomic energy level diagram . . . . . . . . . . . . . . . . . 166 advantages/limitations . . . . . . . . . . . . . . . . . . . . 174–175 edge or XANES region . . . . . . . . . . . . . . . . . . . . 167–168 edge/pre-edge transitions . . . . . . . . . . . . . . 167–166 electronic structure information . . . . . . . . . . . . . 167 Fe K-edge of ferrous vs. ferric FeS4 model complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 multiple scattering and DFT based approaches168 EXAFS region . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168–171 characterization of intermediates in enzymatic reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 dependence on coordination number . . . . . . . . 170 fitting EXAFS data, goal . . . . . . . . . . . . . . . . . . . . 171 FTs and EXAFS for Fe–O, Fe–S, and Fe–Fe interactions . . . . . . . . . . . . . . . . . . . . . . . . 170–171 FTs and EXAFS for Fe-O interaction . . . 169–170 Mn4 Ca cluster in photosystem II . . . . . . . 165–166 photo-electron scattering, approximation . . . . 169 structural characterization of FeMoco . . . . . . . 166 materials experimental setup . . . . . . . . . . . . . . . . . . . . . 171–172 X-ray sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 methods applications to nitrogenase and new directions 173–174 scattering paths, evaluation . . . . . . . . . . . . . . . . . . 173 XANES/EXAFS regions, characterization . . 166–167

NITROGEN FIXATION

326 Index

X-ray crystallography . . . 14, 20, 40, 147–163, 178–179, 183, 223–224, 226, 251, 255 aid to future discoveries . . . . . . . . . . . . . . . . . . . . . . . . 150 catalytic mechanism of nitrogenase . . . . . . . . . . . . . . 149 determination of structures of FeP–MoFeP complexes . . . . . . . . . . . . . . . . . . . . . . . . . 149–150 materials crystal harvesting and cryoprotection . . . . . . . . 157 crystallization. . . . . . . . . . . . . . . . . . . . . . . . . .151–157 protein preparation . . . . . . . . . . . . . . . . . . . . . . . . . 151 methods crystal harvesting and cryoprotection . . . . . . . . 161 data processing, software programs used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161–162 formation of in situ nitrogenase complex crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159–160 hanging drop vapor diffusion. . . . . . . . . . . . . . . .160 protein preparation for crystal plates . . . . . . . . . 158 sitting vapor diffusion crystallography . . 158–159 nitrogenase crystallography . . . . . . . . . . . . . . . . 150–151 crystal harvesting and cryoprotection . . . . . . . . 151 determination of protein structure, steps . . . . . 150

high-resolution nitrogenase structures, crystal symmetry/dimensions . . . . . . . . . . . . . . . . . . . 156 nitrogenase structures/crystallization conditions in RCSD databank . . . . . . . . . . . . . . . . . 152–155 protein isolation/purification . . . . . . . . . . . . . . . 150 protein solubility for crystal formation, importance . . . . . . . . . . . . . . . . . . . . . . . . 150–151 setting up crystallization, methods. . . . . . . . . . .151 nitrogenase enzyme, components . . . . . . . . . . . . . . . 148 isolated from A. vinelandii, study of crystal structures . . . . . . . . . . . . . . . . . . . . . . . . . . 148–149 structure of nitrogenase metal centers, study . . . . 148 X-ray diffraction. . . . . . . . . . . . . . . . . . . .151, 161, 207, 213 X-ray emission lines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173 X-ray sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171, 174

Z Zeeman interaction. . . . . . . . . . . . . . . . . . . . . .192, 208, 211 Zero-field splitting . . . . . . . . . . . . . . . . . 207, 211, 213, 233 Zero rhombicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211