Advances in Applied Microbiology, Volume 44

ADVANCES IN

Applied Microbiology VOLUME 44

This Page Intentionally Left Blank

ADVANCES IN

Applied Microbiology Edited by SAUL L. NEIDLEMAN Oakland, California

ALLEN I. LASKIN Somerset, New Jersey

VOLUME 44

Academic Press San Diego New York Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper. @ Copyright 0 1997 by ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at thebottom of the first page of a chapter in this book indicates the Publisher's consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per-copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923) for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-1997 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0065-2164197 $25.00

Academic Press a division of Harcourt Brace 6.Company 15 East 26'h Street, 15* floor, New York, New York 10010, USA http://www.apnet.com Academic Press Limited 24-28 Oval Road, London NW1 7DX, UK http://www.hbuk.co.uk/ap/ International Standard Serial Number: 0065-2164 International Standard Book Number: 0-12-002644-9

PRINTED IN THE UNITED STATES OF AMERICA 97 98 99 00 0 1 02 BB 9 8 7 6 5

4

3

2

1

CONTENTS

Biologically Active Fungal Metabolites

CEDIUCPEARCE 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. A Brief History of Fungal Products in Medicine. ......................

The Potential of Fungally Derived Chemical Diversity. . . . . . . . . . . . . . . . . . Approaches to Growth and Nutrition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mammalian Enzyme Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cholesterol Biosynthesis and Lipid Metabolism Inhibitors . . . . . . . . . . . . . . Receptor Binding Antagonists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antiinfective Agents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Antitumor and Cytotoxic Activity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . X. Miscellaneous Pharmaceutical Activity. ............................. XI. Agriculturally Active Compounds . . . . . . . . . . . XII. Summary and Conclusions. . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

111. IV. V. VI. VII. VIII. IX.

1 2 3 3 5 13

20 25 47 51 57 66 68

Old and New Synthetic Capacities of Baker’s Yeast

P. D’ARRIGO, G. PEDROCCHI-FANTONI, AND s. SERVI 11. Reducing Capa 111. The Formation

IV. Oxidations: Ge V. Hydrolytic Act VII. The Biogeneration of Aroma Compounds

IX. Conclusions. . . . ................. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

117 119

Investigation of the Carbon- and Sulfur-Oxidizing Capabilities of Microorganisms by Active-Site Modeling

HERBERT L. HOLLAND I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Models for Microbial Hydroxylations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Models for Sulfoxidation Reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. SummaryandPrognosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

V

125 133 150 158 159

vi

CONTENTS

Microbial Synthesis of o-Ribose: Metabolic Deregulation and Fermentation Process

P. DE WULF AND E. J.

VANDAMME

I. Introduction. . . . . . ................ 11. Natural Occurrence d Its Derivatives . . 111. Physicochemical Characteristics of D-Ribose. . . . . . . . IV. Detection and Identification of D-Ribose . . . . . . . . . . V. Applications of D-Ribose . . . . . . . . . . . . . . . . . . . . . . VI. Nonmicrobial Production of D-Ribose . . . . . . . . . . VII. Microbial Production of D-Ribose . . . . . . . . . . . . . . . . . . VIII. Pleiotropic Properties of D-Ribose-Producing Tran Bacillus Mutant spp.. . . . . . . . . . . . . . . . . . . . . . . . IX. o-Ribose Production by Fermentation with Bacillus X. Kinetics of D-Ribose Production by Bacillus spp.. . . XI. Conclusions and Future Perspectives . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . ...

...........

168 168 171 172 172 176 180 188 191 203 204 205

Production and Application of Tannin Acyl Hydrolase: State of the Art P. K. LEKHAAND B. K. LONSANE 11. Historical Highlights. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Tannin-Hydrolyzing Enzymes . . . . . . . . . IV. Source ofTannase . . . . . . . . . . . . . . . . . . .

VIII. Location of Tannase

.......................

X. Properties of Tannase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

................. XIV. Applications of Tannase

216 216 218 222 226 228 237 239 239 241 246 246 249 250 255 255

Ethanol Production from Agricultural Biomass Substrates

c.

RODNEYJ. BOTHASTAND BADAL SAHA I. Introduction. . . . ............................... ..... 11. Lignocellulosic Biomass . . . . . . . . . . . . . . . . . . . . . . . ................. 111. Pretreatment . . . . . . . . . . . . . . . . . . . . . .................

V. Fermentation. . . ............... VI. Technological Co p . . . . . . . . . . .. VII. Future Prospects . . . . . . . . . . . . . . . . . . . . . . . . . .

. . .. ....

261 263 265 267 2 74 279 282 282

CONTENTS

vii

Thermal Processing of Foods, A Retrospective, Part I: Uncertainties In Thermal Processing and Statistical Analysis

1. 11. 111. IV. V. VI.

M. N. RAMEsH, S. G. PRAPULLA, M. A. KLJMAR, AND M. MAHADEVAIAH ..................... ..................... .................... tion . . . . . . . . . . . . . . . . . .............. usions . . . . . . . . . .............. rences . . . . . . . . . .........

.. .. ..

..

.. ..

*.

288 289 305 308 310 311 312

Thermal Processing of Foods, A Retrospective, Part I I: On-Line Methods for Ensuring Commercial Sterility

M. N. M E S H , M. A. KUMAR,S. G. PRAPULLA, AND M. MAHADEVAIAH I. 11. 111. IV. V.

Introduction.. . . . . . . . FO Integrators. . . . . . . . ........ On-Line Monitoring Sy s . . .............................. Semiautomatic Retort Control Systems for Optimum Sterilization. . Computer-Aided Sterilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

..

VIII. Conclusions

.. ..

316 317 321 324 326 327 341 342 343

INDEX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

..

347

....

355

.. ..

.. ..

.. ..

CONTENTS OF PREVIOUS VOLUMES. ....................

This Page Intentionally Left Blank

Biologically Active Fungal Metabolites CEDRICPEARCE MYCOsearch, a subsidiary of Oncogene Science Inc. Durham, North Carolina 27707

I. Introduction 11. A Brief History of Fungal Products in Medicine 111. The Potential of Fungally Derived Chemical Diversity Iv. Approaches to Growth and Nutrition V. Mammalian Enzyme Inhibitors VI. Cholesterol Biosynthesis and Lipid Metabolism Inhibitors VII. Receptor Binding Antagonists VIII. Antiinfective Agents IX. Antitumor and Cytotoxic Activity X. Miscellaneous Pharmaceutical Activity XI. Agriculturally Active Compounds XII. Summary and Conclusions References

I. Introduction

The role of fungal metabolites in medicine and the unwanted biological effects of fungal metabolites have been observed and studied for many years. That being so, the intent of this chapter is to review more recent literature reporting significant biologically active compounds from fungi, drawing from those published up to December 1996. In many ways this is an exciting time for the discovery of bioactive metabolites from fungi, with new cultures being isolated from a wider variety of habitats and substrates and with the application of a greater understanding of fungal metabolism becoming more apparent by employing more sensitive and novel bioassays as screens, and by applying the recent advances in chemistry that have made the isolation and characterization of these compounds much more rapid. Analysis of the literature shows that proportionally more bioactive fungal products than actinomycete metabolites are being reported now than a decade ago. There are a number of excellent books and reviews of earlier work; for example, see Turner (1971), Turner and Aldridge (1983), Cole and Cox (1981),and Gloer (1995a, 1996). It is not the intention of this review to be encyclopedic nor to evaluate the relevance of any bioactivity to its translation into a mechanism of action for a new drug. Although 1 ADVANCES IN APPLIEO MICROBIOLOGY, VOLUME 44 Copyright 0 1997 by Academic Press, Inc. All rights of reproduction in any form reserved. 0065-2164/97$25.00

2

CEDRIC PEARCE

mycotoxins are strictly bioactive, they have not been included unless the ones described in the literature have demonstrated some more useful medicinal or agricultural attribute.

II. A Brief History of Fungal Products in Medicine

In the past 50 years, there have been a number of highly successful drugs based on fungal metabolites, and these have enjoyed a huge global market. Penicillin, for example, a potent antibiotic against sensitive Gram-positive bacteria was first observed many years ago being produced by a Penicillium notatum by Fleming (1946). Subsequent research showed that penicillins are also produced by Penicillium chrysogenum and a variety of other organisms (Lechevalier, 1975). Penicillin G, the original metabolite isolated, is still used clinically, and the utility of this and the very important semisynthetic derivatives as crucial antibacterial weapons cannot be overestimated, even in the face of antibiotic resistance. Studies on the biosynthesis and other biochemical, physiological, and genetic aspects of the production of these antibiotics is ongoing in many academic laboratories, and results from this work will undoubtedly provide a better understanding of how and why these organisms produce such compounds. The work of Baldwin and colleagues on p-lactam antibiotics is a high point in this area (Roach et al., 1995, and the references therein). The immunosuppressant cyclosporine was first discovered as an antifungal agent produced by a Tolypocladium inflatums (Beauvaria nivea) and Cylindrocarpon lucidum (Dreyfus et al., 1976). It was subsequently discovered to have excellent immunosuppressive activity and is used for treatment following organ transplant (Goodman Gilman et al., 1985). Cyclosporine A has been reported from a variety of different common organisms, including many strains of Tolypocladium inflatum and ?: geodes, as well as Acremonium, Beauvaria, Fusarium, Paecillomyces, and Verticillium species (Sanglier et al., 1990). Finally, among of the most successful drugs derived from a fungal product are the cholesterol biosynthesis inhibitors related to Lovastatin, which was initially reported by Merck from an Aspergillus terreus (Vagelos, 1991; see Section VI). There were reports of related compounds from Penicillium brevicompactum and Penicillium citrinum from groups at Beecham’s (Brown et al., 1976) and Sankyo (Endo et al., 1976). These compounds inhibit the rate-limiting step in cholesterol biosynthesis, HMG-CoA reductase, and they are used clinically to reduce cholesterol levels.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

3

Fungal natural products still provide important lead structures for drug development, a recent example being mycophenolic acid, mofetil. Mycophenolic acid was initially discovered from a Penicillium brevicompacturn in 1896 and subsequently found from a number of other penicillia. The structure was reported in 1952. Various bioactivities associated with this compound were discovered, including its immunosuppressive action (Wu, 1994). The mofetil derivative is a pro-drug that is broken down in the body to liberate mycophenolic acid, which inhibits the biosynthesis of certain precursors of nucleic acids and thereby inhibits proliferation of the cells involved in the immune response. Syntex/Roche has developed mycophenolic acid mofetil as an immunosuppressive drug, and this has been recently approved for use following kidney transplantation (Anonymous, 1995). This is a good example of the need to test the bioactivity of as many known natural products as possible in any new drug discovery screen, since a previously known compound can have the activity required and be the lead sought. Ill. The Potential of Fungally Derived Chemical Diversity

Hawksworth (1991) estimated that there are one and a half million fungi in existence, although at that time only about 69,000 had been described in the literature. Assuming that the actual rate of description of organisms by fungal taxonomists is approximately linear, that number may now be between 75,000 and 80,000. This is still small compared to the one and a half million thought to exist, and, apparently, the majority of fungi inhabiting the world have not been described. By implication, most of these fungi have not been screened in drug discovery programs. The potential chemical diversity of this vast untapped resource is surely one of the great driving forces behind today’s search for novel metabolites for use as drugs or leads to those drugs. IV. Approaches to Growth and Nutrition

Fungi require carbon, nitrogen, phosphorus, sulfur, minerals, vitamins, and other growth factors. They are sensitive to temperature and oxygen/carbon dioxide in their environment; pH is also a critical factor (Jennings, 1995). Nutrition to support growth is generally not difficult to satisfy, especially since many of cultures are isolated by growing out from substrates. Many of the nutritional requirements can simply be satisfied by using plant material. The challenge faced is to provide the organisms with conditions that allow expression of secondary metabo-

4

CEDlUC PEARCE

lism and accumulation of unusual metabolites (Martin et al., 1982; Bushell, 1989; Tanaka, 1992; Demain, 1992). There is probably no such thing as the best culture medium that will allow any fungus to express its secondary metabolic capability, and it is impossible to predict what a freshly isolated fungus will require for metabolite accumulation. In large screening programs, in order to manage the risk that the cultures as a group are being given a fair chance to express secondary metabolism, a number of media are generally employed. Initial evaluation of media is usually made on three levels: (1) suitability for fungal growth and metabolite accumulation; (2) compatibility with the screen being used; and (3) effect on isolation chemistry. Using HPLC in conjunction with a high-hit-rate bioassay, such as Bacillus subtilis, the suitability of a medium can be determined; this should probably be considered as a short-term fix that has to be confirmed by constantly monitoring the positive rate in the assays being employed for the drug discovery program. Selection of media is complex since the possible variations are so large. Simple media such as potato dextrose work very well as broth and agar, and this has been validated many times, with novel bioactive compounds being produced. Variations using potato dextrose media have been employed, including starting with fresh potato, although the most efficient is to simply employ commercial media, possibly supplementing with other nutrients. A more complex but commonly employed approach is to add a rapidly utilized sugar, for example, dextrose, and a more slowly utilized carbohydrate, for example, mannitol, together with a nitrogen source and minerals, in a medium. Fungi will frequently grow rapidly using the glucose and then enter a slower-growing phase favoring secondary metabolism. Many good ideas can be obtained from the literature (Jennings, 1995). Ammonia levels seem to control biosynthesis of many compounds. When fermentations were carried out using ammonia-trapping agents, up to tenfold increases in titers were observed (Tanaka, 1992). One approach to controlling this is to use either a complex slowly utilized nitrogen source such as soy meal or to use an ammonium salt of an organic acid, with the latter acting as the carbon source. If each part of the salt is utilized at the same rate, no accumulation of ammonium results and no dramatic change in pH is observed. Media containing an inorganic nitrogen source such as ammonium nitrate could be used providing the cultures are incubated long enough so that a stationary phase is entered well before the fermentation is processed. There is also a large amount of literature showing that phosphate levels control secondary metabolism (Tanaka, 1992) and that metabolite

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

5

accumulation is triggered once this has been depleted. Examples include the production of cephalosporins, a number of aminoglycosides, and chloramphenicol. This led to the development of media containing low phosphate and to media containing phosphate trapping agents. Use of these agents resulted in fivefold increases in antibiotic production in some cases. However, some groups deliberately used high-phosphate media and proceeded to discover novel compounds whose biosynthesis was not controlled by this mechanism. The implication is that by using such an approach they increased their chances of finding novel compounds; certainly, many new compounds were found. Some of these approaches have been used to grow actinomyces only, but would be worth investigating for fungal secondary metabolism. A variety of solid substrates can be usefully employed. These included nutrient grains such as rice or corn, or nonnutrient vermiculite containing a nutrient liquid medium. This type of environment has a number of advantages, most importantly that this reflects the natural state for a substrate encountered in the field and provides a suitable environment for partial or complete expression of the life cycle along with the associated chemical signaling agents. Further variations can be introduced by changing the incubation time and temperature; with a lean medium the incubation time can usually be shortened since the organisms become nutritionally stressed earlier. Temperatures either lower or higher than the traditional ZZOC are known to have a profound effect, but again there is no way of predetermining this. There are many other combinations of carbon source, nitrogen source, phosphate, sulfur, minerals, etc., which can be tried, and should be. Resins can be added to the media that bind certain metabolites and thus increase their production, or alter metabolism in a way that leads to novel compounds accumulating. This often results in dramatic changes in titers and types of end-products. V. Mammalian Enzyme Inhibitors

There have been many enzyme inhibitors isolated from microorganisms and fungi, and the mechanism of action for many bioactive compounds can be interpreted in terms of the inhibition of some key reaction. The search for an anticancer fungal metabolite is constant and is rejuvenated periodically by the identification of new targets. Protein kinases are a group of enzymes that phosphorylate proteins as a means of regulating their conformation and activity. Two such families of enzymes that play a vital role in cellular physiology are the protein kinase C group (PKC)and the protein tyrosine kinases (PTKs).The PKCs

6

CEDRIC PEARCE

(2)Emodin

(1 ) Balanol

0

H STRUCTURES 1, 2.

phosphorylate serine or threonine, while the PTKs are responsible for the phosphorylation of certain tyrosine residues. Both of these enzymes have been identified as playing crucial roles in cell proliferation and differentiation (Nishizuka, 1988; Chen et al., 1987). The PTK enzymes appear to be deregulated during tumor transformation to malignancy, as have PKCs in certain circumstances. PKC is the receptor for phorbol esters that promote tumor formation. The role of these enzymes in the development of cancer makes them an ideal target for screening natural product extracts. One of the more potent protein kinase C inhibitors ever to be reported is balanol (I), which was discovered from a fungus, Verticillium balanoides isolated by Barry Katz of MYCOsearch from a rhizomorph of a pine. Fermentations of this culture initially produced in MYCOsearch laboratories showed activity against PKC, the active compound (balanol) being isolated and characterized by the Sphinx Pharmaceuticals natural products group (Kulanthaivel et. al., 1993). Balanol was shown to inhibit PKC isozymes in the nanomolar range, having equal or better potency than the actinomyces product staurosporin, as well as showing some activity against PKA (Kulanthaivel, 1995). Three hundred analogues were subsequently prepared for structure activity studies. Unfortunately, the selectivity observed in the reported experiments was not sufficient to warrant pursuing this compound. A number of fungally derived anthraquinones related to emodin have been reported to inhibit protein tyrosine kinase. The anthraquinone emodin (2) was reported by Jayasuriya et al. (1992) from the roots of a Chinese medicinal plant, Polygonium caspiatum, and related

7

BIOLOGICALLY ACTIVE FUNGAL METABOLITES Paecilcquinones

OH

0

OH

HOO *H

OH HO

0

OH

(5)c

(9) Versiconol OH I

roH

0 STRUCTURES 3-9

bioactive compounds have been reported from fungi (Fredenhagen etal., 1995a,b). Some of the more potent of these were the paeciloquinones @-9), which were isolated from fermentations of Paecilornyces carneus, a culture isolated from a soil sample collected in a jungle region in Bolivia (Petersen et al., 1995). The production of paeciloquinones was closely dependent on the conditions of the fermentations. It was reported that shake-flask fermentations using a complex medium containing saccharose, mannitol, meat extract,

8

CEDRIC PEARCE

and ammonium nitrate and grown at 30°C yielded 50 mglliter of paeciloquinone A, but using similar conditions in a bioreactor the yield was very much reduced. Investigations showed this was probably due to insufficient oxygenation, and by carrying out the bioreactor experiment at a higher temperature, 33"C, comparable concentrations of the desired product were observed. Of the seven components produced by the fungus, the exact nature of products from a given fermentation could be changed by manipulating the medium. Thus, on the medium described above, paeciloquinones A, C, and D, and versiconol were produced, whereas on a medium containing sucrose, mannitol, peanut oil, meat extract, and ammonium nitrate inoculated and incubated in the same manner produced paeciloquinones B, E, and F. One explanation is that the peanut oil contains a six carbon-containing precursor for the side chain, but a more complex argument is needed to explain the origins of the four carbon fragment seen in components A, C, and D. This type of observation is commonly made with fermentations. However, without extensive biosynthetic and biochemical investigation, the cause is really speculative. These metabolites were inhibitors of v-abl, c-src, and EGF-R PTK, with IC50values in the micromolar range. Paeciloquinones A, C, and D were most potent against EGF-R with IC50values between 6.7 and 11 pM; A, C and F were most potent against v-abl with IC50values between 0.56 and 3.6 r-LM; A and C were most potent against c-src with IC50 values of 2 and 9 pM, respectively. None of the paeciloquinones displayed antimicrobial activity. During the invasive and metastatic phase of cancer, the tumor cells must penetrate the basement membranes of the tissue under attack. This degradative step is caused in part by protease and glycosidase. One of these enzymes is heparinase, which degrades the heparan sulfate of basement membranes. Heparinase has been identified in B16 melanoma, and its activity is associated with the metastatic capacity of malignancy. Trachyspic acid (lo) a, metabolite from Talaromyces trachyspermus that had been isolated from a Japanese soil sample, grown in a complex medium containing glycerol, potato, yeast extract, and malt extract, inhibits heparinase with an IC50of 36 pM (Shiozawa et al., 1995). Trachyspic acid showed no inhibitory activity towards bovine liver P-glucuronidase, thus demonstrating a degree of selectivity, although the activity in an invasive assay was not reported. Trachyspic acid is similar to three other reported polycarboxylic acids containing long alkyl side chains: (+)- and (-)-decylcitric (Gatenbeck and Mahlen, 1968; Brandage et al., 1976) and spiculisporic (Clutterbuck et al., 1931) acids, isolated fiom Penicillium spiculisporum, later reclassified as T

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

9

(10) Trachyspic acid

0

(11) Pachyrnic acid

II

(12) Dehydroturnulosicacid

II

STRUCTURES 10-12.

Trachysperrnus. The bioactivity of these previously reported compounds was not given, but, since they predate the discovery of the significance of the heparinase, it would be interesting to investigate their bioactivity in the assays employed here. Phospholipase A2 is responsible for producing fiee fatty acids and lysophospholipids. PLA2-I is present in secretions hom the pancreas, whereas PLA2-I1is involved in the inflammatory process. Inhibitors of PLA,-II might be useful antiinflammatory agents. PLA2 inhibitors have been isolated from a few fungi, including the Chinese fungus Poria cocos Wolf, a polyporaceae that is well known in traditional medicine for its diuretic, sedative, and tonic effects. The active compounds, pachymic acid (11)and dehydrotumulosic acid (12)were isolated and characterized (Cuellar et al., 1996) and were shown to inhibit snake venom PLA2 with ICs0 values of 2.9 and 0.84 pM, respectively.

10

CEDlUC PEARCE R1

I

OH

H3C'

.CH3

(16) MR 304A

STRUCTURES 13-16.

Acetylcholinesterase (AChE) is the enzyme responsible for the breakdown and subsequent inactivation of the neurotransmitter acetylcholine. From a screening program looking for microbial metabolites that inhibit AChE, Omura et al. (1995) found a Penicillium sp. FO-4259 isolated from a Japanese soil sample that produced such activity when grown in a medium containing saccharose, glucose, corn steep liquor, meat extract, agar (not enough to produce a solid medium) potassium was phosphate, and calcium carbonate. The compound, arisugacin (s), characterized and shown to be similar to two tremorgenic compounds, territrems B (14)and C (G),previously reported from Aspergillus terreus (Ling et al., 1979, 1984). Arisugacin had an of 1 nM against human erythrocyte AChE, approximately 200 times more potent than tacrine, an inhibitor that has potential for use in Alzheimer's disease (Summers et al., 1986). Arisugacin was 1500 less potent than tacrine against butyryl-cholinesterase and is thus far more selective than the latter. The territrems were also shown to be effective AChE inhibitors. Not all novel fungal metabolites are large complex structures with multiple stereochemical centers, as is shown by the compound known This compound was discovered from a program seekas MR304A (16). ing melanin biosynthesis inhibitors (Lee et al., 1995). The culture producing MR304A is Trichoderma harzianum isolated from a Korean soil sample. Bioactivity was determined using three assays: against mush-

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

11

(17 ) Melanoxadin

STRUCTURE 17.

room tyrosinase, melanin production in Streptomyces bikiniensis, and melanin formation in €316 melanoma cells. MR304A inhibited all three. It was shown to be a noncompetitive inhibitor of mushroom tyrosinase with an IC5,, of 7.5 pg/ml and at 1pg/ml inhibited melanogenesis in B16 melanoma without showing cytotoxicity. Other isonitrile fungal metabolites have been reported from Trichoderma spp. (Boyd et al., 1991; Brewer et al., 1979; Fujiwara et al., 1982); one of these, trichoviridin, was also shown to be a melanin biosynthesis inhibitor (Hashimoto et al., 1994). Using the larval haemolymph of the silkworm Bombyx mori as the basis for a melanin biosynthesis inhibitor, Hashimoto et al. (1995) When the haemolymphs are exposed, to discovered melanoxadin (17). air they change color from yellow to black, and this is due to the formation of melanin. These workers found a Trichoderma sp. that was isolated from a Japanese soil sample that produced the oxazole melanoxadin when grown in a complex medium. The IC5,, value in the haemolymph assay was 22.3 pg/ml compared to 13.1 pg/ml for trichoviridin. The IC5,, against mushroom tyrosinase was 98 pg/ml. The inhibition of aldose reductase is a new approach to the treatment of diabetes mellitus (Sarges, 1989). A number of microbial products have been isolated that inhibit this enzyme. The tricyclic salfredins 18-24 have been isolated from Crucibilum sp. RF-3817 by Matsumoto et al. (1995). The culture was isolated from a piece of rotten wood, and the active compounds produced in a complex medium containing starch, sucrose, and yeast extract. Seven related metabolites were produced. The most potent of these in a rat lens aldose reductase assay were salfredins A4 and C2, both of which contain a glycine moiety. When the producing organism was grown in a medium containing glycine, the yield of the A4 component was increased 25-fold and the C2 component 15-fold. By using a medium containing corn starch, glucose, corn steep liquor, peanut oil, dried yeast, and calcium carbon-

12

CEDRIC PEARCE

Salfredins

(18) A 3

R=-CH(COOH)CH2CH&OOH

(19) A 4

R=-CHSOOH

(20)A 7 R=-CH(CH&OOH

(21)Ci R=-H

(22)C2 R=-CH2COOH (23) C3 R=-CH(CH&OOH

STRUCTURES 18-24.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES (25)Compactin

13

(26)Lovastatin

STRUCTURES

25, 26.

ate, glycine supplementation at a rate of 1%gave the best yields: 2579 yg/ml A2 and 798 yg/ml C2. VI. Cholesterol Biosynthesis and Lipid Metabolism Inhibitors

Some of the most potent and clinically useful cholesterol metabolism effectors have been derived from fungal products, most notably the hydroxymethyl coenzyme A reductase (HMG-CoAreductase) inhibitors related to compactin (25) and produced by Penicillium brevicompactum (Brown et a]., 1976), and Lovastatin (26) from Aspergillus terreus (Alberts et al., 1980), which had also previously been reported from Monascus ruber (Endo, 1979). An interesting account of the discovery of Lovastatin is given in a review by Vagelos (1991). Subsequent to their initial discovery, these compounds have been shown to be common fungal metabolites but are uncommon from the otherwise prolific actinomycetes, and it is possible that in fungi that contain sterols as an integral part of their cell membranes the steroid biosynthesis inhibitors play some role in the control of the production of ergosterol and related metabolites. From the initial discovery of the HMG-CoA reductase inhibitors, there has been considerable attention paid to the fungi as a source of compounds for use in the control of other facets of mammalian steroid biochemistry. (The squalene synthase inhibitors squalestatins and zaragozic acids are reviewed in Section VIII). Acyl-CoA:cholesterol acyltransferase (ACAT) esterifies cholesterol and thereby has a central role in regulating intracellular free cholesterol levels in humans. In atherosclerosis, foam cells are formed that contain large quantities of cholesteryl esters, the presence of which is directly related to ACAT activity. It is thought that inhibitors of this enzyme will be potential antiatherosclerotic drugs (Sliskovic and White, 1991).

14

CEDlUC PEARCE

H

(30)Terpencble D

STRUCTURES 27-30.

A number of groups have reported inhibitors of ACAT from fungi, including compounds from a new genus. (For this review, as a general rule, only those most recently published are included, although some groups have published prolifically on this topic in the past few years.) Huang et al. (1995a,b) discovered in a new fungi, Albophoma yarnanashiensis, a coelomycete isolated from a soil sample collected in Yamanashi, Japan, which produced a series of metabolites, the terpendoles 27-30 when grown in a medium containing starch, glycerol, soybean meal, yeast, potassium phosphate, calcium carbonate, magne-

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

15

sium sulfate, and potassium chloride. Four of these products were novel, and two, paspaline (Fehr and Acklin, 1966) and emindole (Nozawa et al., 1987), had been reported previously from other fungi. In an ACAT assay using rat liver microsomes, terpendoles C and D were the most potent inhibitors, showing ICs0 values of 2.1 and 3.2 pM, respectively. In addition, in a bioassay designed to test the effect on both cholesteryl ester formation and cell viability in macrophage J774, terpendole D showed an IC50 of 0.048 pA4 and a CD50 (a measure of cytotoxicity) of >24.8 pA4. Terpendole C and D showed no acute toxicity in ddY mice at 100 mg/kg. It was postulated that the sterol-like side chain of the terpendoles might compete with the natural substrate of ACAT, cholesterol. When a different medium containing maltose, Ebios, yeast, potassium phosphate, magnesium sulfate, and potassium bromide was fermented with A. Yamanashiensis, eight new terpendoles were produced (Tomoda et al., 1995d); all other parameters appeared to be the same as the earlier fermentation, when some or all of the new products were produced, but at a level that made isolation and characterization impractical. None of the new metabolites approached the potency of terendole D, the closest being approximately one-tenth as active, and none of the terpendoles showed any antibacterial or antifungal activity. A second group of ACAT inhibitors, the pyripropenes 3 1 4 2 (Omura et al., 1993b; Tomoda et al., 1994; Kim et al., 1994; Tomoda et al., 1995a) and the GERI-BP001 series (4345)(Jeong et al., 1995), with a similar motif to the terpendoles, that is, the presence of a steroid-like moiety, were both reported from Aspergillusfumigatus. The most potent inhibitor in the whole pyripropene series, substance A, had an IC50 value of 0.16 pA4. While attempting to produce enough material for further biological work, the group exposed the fungus to N-methy1-N’nitro-N-nitrosoguanidine and isolated mutants, one of which was shown to produce 10 times the amount of pyripyropene than the parent did when grown in a medium containing starch, glycerol, soybean meal, yeast, potassium chloride, potassium phosphate, magnesium sulfate, calcium carbonate, and nicotinic acid (Tomoda et al., 1995a). Biosynthetic studies demonstrated that the nicotinic acid added to the fermentation medium was incorporated directly into the metabolites. This mutant strain also produced eight new compounds in the series. Pyripropene L was the most potent of these in the ACAT inhibition assay, with an value of 0.27 pM.None of these compounds showed antibacterial activity. Similar compounds were discovered in an independent investigation in Korea (Jeong et al., 1994, 1995). In this case, the producing organism, isolated from a soil sample collected from

16

CEDRIC PEARCE pyripropene

(43)GERI-BPWI M

R1&. R d H 3

(44) GERI-EP001 A

R1=OH, Rz=CH3

(45) GERl-BPOOl B Rt=H, R*=CH&H3 STRUCTURES

43-45.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

17

Mountain Dukyou, Korea, was cultured in a complex medium containing starch and soytone wherein it produced ACAT inhibitory activity, which leveled off at 6 days. The purified components, GERI-BP001 M, A and B, which are closely related to the pyripropenes, had IC50values against rat liver microsome enzyme of 42, 94, and 40 pM, respectively, with pyripyropene A, a known inhibitor included as a control, having an IC,, of 43 nM. These compounds showed no antibacterial or antiviral activity. The azaphilone group of ACAT-inhibiting metabolites, the isochromophilones 46-51 were discovered by the Kitasato Institute group (Arai et al., 1995). These compounds were produced by Penicillium rnulticolor FO-3216 isolated from a Japanese soil sample collected in Tokyo. Fermentation in a medium including starch, glycerol, soybean meal, yeast, potassium chloride, calcium carbonate, magnesium sulfate, and potassium phosphate for 3 days at 27°C in a fermenter led to the production of isochromophilones 3-5, with six being produced in the same medium but in shaken Erlenmeyer flasks. Isochromophilones 4 and 6 were the most potent of the group against ACAT, with IC50values of 50 pM. 4 also showed inhibitory activity against cholesteryl ester transfer protein, with an IC50 of 98 pM. 3, 5, and 6 had some weak antimicrobial activity as well as inhibiting the growth of B16 melanoma cells with IC50 values of 33, 36, and 30 pM, respectively. An alternative approach to decreasing circulating cholesterol levels by reducing its biosynthesis is to inhibit the absorption of dietary cholesterol in the intestine. Intestinal absorption involves the enzyme cholesterol esterase, which esterifies cholesterol prior to the generation of chylomicrons, the form in which lipids are transported around the body (Goodman Gilman et al., 1985). It has been shown that cholesterol absorption is markedly decreased in rats when they are depleted of cholesterol esterase (Gallo et al., 1984). Sakai and colleagues (1995) isolated a Stachybotrys sp. that produced a series of triprenylphenol cholesterol esterase inhibitors, two of which, K-76 (52)and stachybotryhad previously been reported from other strains of Stachydial (s), botrys complementi (Kaise et al., 1979) and Stachybotrys cylindrospora (Ayer and Miao, 1993), respectively. Ten of the remaining metabolites (54-63) were novel triprenylphenols that were produced when the Stachybotrys originally isolated fiom a soil sample collected in Japan was grown in a medium containing glucose, soybean meal, polypeptone, meat extract, yeast extract, potassium phosphate, and magnesium sulfate. The activity of the purified metabolites was determined using porcine pancreatic enzyme catalyzing the formation of cholesterol oleate. Of the 10 compounds isolated, stachybotrydial was the most potent,

lsochromophllones

OAc (49) IV

0 STRUCTURES 46-51.

(51) VI

19

BIOLOGICALLY ACTIVE FUNGAL METABOLITES Triprenylphenols

0

HO'"' HJC

"CH3 (52) 1 R>=OH, R&HO

(676)

(55) 3 Ri=OH. Rp=H, X=NH

(53)2 Rl=H, Rz=CHO (Stachybotrydial)

(56) 4 R,=OH. Rp=OCH,, X=NH

(54) 9 Ri=H, R&3i3

(57)5 Rl=OH. Rz=OCHs, X=NH (58) 7 Rl=OH, Rz=H, X=N(CHp)flH (59) 8

Rl=OH, Rz=H, X=N(CH,),COOCH,

(60)10 R+I,

Rz=H, X=(CHz)&DOCH3

(61) 12 Rl=H, Rz=H, X=O

(62)11

STRUCTURES 52-63

having an IC,, of 60 pM, with two other metabolites having IC,, values of 200 (K-76) and 270 (IX) yM. Stachybotrydial inhibited cholesterol esterase in a time-dependent manner and was irreversible. Stachybotrydial also inhibited the dietary absorption of cholesterol in rats and significantly reduced the serum levels of cholesterol in mice fed a diet containing 1%cholesterol. Since these compounds had no effect on ACAT, another enzyme postulated to control cholesterol absorption, their regulatory effect is presumed to work through control of the esterase. Other facets of lipid metabolism that have been targets for fungal metabolites include fat biosynthesis. The enzyme diacylglycerol acyltransferase catalyzes acylation from acyl-CoA to diacylglycerol, thus producing triacylglycerol. This is the last step in fat formation. Reports

20

CEDlUC PEARCE

of inhibitors of this pathway are rare. In their screening program to find metabolites that would inhibit this enzyme, Tomoda and co-workers (Tomoda et al., 1995b,c) found a Humicola sp. isolated from a Japanese soil sample in which a series of compounds, amidepsines A-D (64-67), was produced when grown in a liquid medium containing sucrose, glucose, corn steep liquor, meat extract, potassium phosphate, magnesium sulfate, ferrous sulfate, manganese chloride, zinc sulfate, copper sulfate, cobalt chloride, calcium carbonate, and agar. Amidepsines A-C consist of a tridepside and an amino acid, and as such are a new type of fungal metabolite, although tridepsides have been isolated from a variety of fungi and lichens and shown to have a number of bioactivities. Amidepsine D was previously reported from the lichen Parmelia damaziana (Elix et al., 1981). These compounds inhibited the target enzyme with IC5,, values of 10.2, 17.5, 19.2, and 51.6 pM (A-D, respectively). The effects of these compounds in Raji cells confirmed that they inhibit production of triacylglycerol formation with IC5,,values of 15.5, 3.35, 17.2, and 2.82 W(A-D, respectively). Compound B showed more selectivity since it had little inhibitory effect on the production of either phosphatidylcholine or phosphatidylethanolamine in Raji cells. While these compounds showed no cytotoxicity at the concentrations used in the assays, they were mildly antibacterial against Bacillus subtilis. In a second strain of Humicola sp. FO-5969 grown in the same medium, this same research group (Tomoda et al., 1996) discovered a methylated derivative (68) of amidepsine A, which was coproduced with B. This strain did not yield the other three amidepsines. The new metabolite, amidepsine E, is a weak inhibitor of the target enzyme, with an value of 124 pM. VII. Receptor Binding Antagonists

A number of fungal products have been shown to have neuropharmacological properties. Extreme examples include the well-known lysergic acid diethylamide (LSD) from Claviceps purpurea and psilocybin from, among others, Psylilocybe mexicana. Not surprisingly, other neuropharmacologically active compounds with various bioactivities have been reported. In a program screening fungal extracts for compounds that effect binding to the GABA-benzodiazepam receptor, Ainsworth et al. (1995) discovered xenovulene hom a strain of Acremonium strictum. This culture had been isolated from a foam sample collected from a tropical stream. When the fungus was cultured in a medium containing glucose, yeast extract, MES, Tween 80, Antifoam A,

BIOLOGICALLY ACTIVE FUNGAL METABOLITES OH

21

0

(64) Amidepsine A

(65) Amidepsine 6

-H

(66)Amidepsine C

-H

I

-H

0 (67)Amidepsine D

- CH3

(68)Amidepsine E

- CH3

- OH

-H

STRUCTURES 64-68

and carboxymethylcellulose, the humulene-related compound xenovulene (69)was produced. Humulene is a recognized plant product and is present in, for example, hops, but it has also been isolated from Coriolus consors (Nokoe et a]., 1976), and derivatives from other fungi (Turner and Aldridge, 1983), including the bioactive compounds pycnidione and upenifeldin (Harris et a]., 1993; Mayerl et al., 1993, 1994). Xenovulene inhibited the binding of flunitrazapam to the GABA-benzodiazepam receptor from ox brain with an ICs0 value of 40 nhl; the IC,, for the sodium salt was 10 nM.

22

CEDRIC PEARCE

Neuropeptide Y is a peptide regulator with a wide variety of biological effects that is distributed in catecholamine-containing neurones. It has been associated with the intake of food and water as well as regulation of pituitary function, vasoconstriction, and anxiolysis/sedation (Grundemar and Hakanson, 1994). Three classes of neuropeptide Y receptor have been described (Gehlert, 1994). Kodukula and co-workers (1995) isolated a tetracyclic compound, BMS-192548 (701, from Aspergillus niger that inhibits the binding of neuropeptide Y to NPYl and NPY2 receptors, with IC50values of 24 and 27 yM, respectively. BMS-192548 (Shu et al., 1995) is related to a number of known fungal metabolites and is a regio- and possibly a stereoisomer of the known compound TAN-1612 (71),a metabolite isolated from Penicilliurn claviforme (Ishimaru et al., 1994). TAN-1612 was shown to be a substance P inhibitor. These compounds are closely related to the actinomyces product tetracycline, although BMS-192548 showed no activity against a range of Gram-positive and -negative organisms at 100 yg/ml. Endothelins are a group of peptide vasoconstrictors of potent and long lasting effect. Three endothelin isopeptides are secreted by endothelial cells. Endothelins are thought to have a role in a variety of cardiovascular problems, including hypertension (Saito et al., 1989), congestive heart failure (Watanabe et al., 1990), and various related diseases. A number of streptomyces products are known to be antagonists of endothelin (Ihara et al., 1991; Miyata et al., 1992). In a screening program looking for endothelin antagonists from fungi, Nakamura et QI. (1995) discovered a series of novel metabolites, the stachybocins (72-74),from cultures of Stachybotrys sp. M6222. This culture was isolated from a Japanese soil sample, and, when grown in glucose, peptone, corn steep liquor, potassium phosphate, magnesium sulfate, and celite, produced the stachybocins. These compounds consist of two spirobezofurans, each attached to a decalin derivative with both units connected by a lysine (Ogawa et al., 1995b).The three metabolites inhibited the binding of radiolabeled endothelin-1 to human endothelin receptors type A with IC50values of 13, 1 2 , and 15 yM, respectively, for isomers A, B, and C, and with values for inhibiting binding to the type B receptor of 7.9, 9.5, and 9.4 yM, respectively. When these compounds were tested for relaxation activity in rabbit aortae treated with endothelin-1, contraction was relaxed 86, 70, and 68% for metabolites A, B, and C, respectively, at 30 yM, thus demonstrating that these compounds are endothelin antagonists. These compounds showed no antimicrobial effects. A second group of fungal metabolites, the azaphilones (Pairet et al., 1995), have been reported as metabolites from Penicillium

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

23

(69) Xenovulene A

0

OH

0

OH

CH30

0

OH (70) EMS-192518

OH

OH

CH3O

0

0

0

OH (71) TAN-1612

0

0

(72)A Rj=RpH (73) B Rj=OH, R2=H (74) C Rl=H. R2=OH

STRUCTURES 69-74.

sclerotiorurn, these also being antagonists of endothelin-A and endothelin-B. Two of these azaphilones (75 and 76) are novel sclerotiorin analogues and are similar to the isochromophilones, inhibitors of acyl-CoA:cholesterol acyltransferase, as discussed above. In binding studies with human ET, and ETb receptors using radiolabeled

24

CEDlUC PEARCE

Azaphih

(75)l R=CI

(76)2 R=H

(77) RES-12141 R*=OH, RFH, R,=OCH, (78) RES-1214-2

Rl=OH. R d I . R & C H 3

(79) Dihydroepiepofonin

I

OH

STRUCTURES 75-79.

endothelin-1, the IC,, values ranged between 56 and >250 pA4. In a third study, Ogawa and co-workers (1995) discovered two endothelin-1 antagonists from a Pestalotiopis sp. that had been isolated from a Japanese and -2 (781,which soil sample. These two compounds, RES-1214-1 (77) are simple aromatic bicyclic metabolites, showed IC5, values of 1.5 and 20 pA4 using ET, receptors and endothelin-1. In a screening program aimed at the discovery of interleukin-1 receptor binding antagonists, Kuo and colleagues (1995) found a PenicilIium patulum that produced bioactivity, and bioactivity-guided isolation led to the known compounds gentisyl alcohol and epiepoformin. Subsequent fermentations including HP-20 resin led to production of a novel compound, dihydroepiepoformin (79).The inclusion of resins in fermentations to improve titers and to enhance production of otherwise

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

25

very minor components is a previously reported approach used in a number of programs (e.g., Marshall et al., 1990). Mechanistically this approach probably works by removing feedback inhibition from metabolic pathways and by sequestering intermediates. The compounds described had IC50 values of 0.7 mM. VIII. Antiinfective Agents

There is a chronic unmet need for new antifungal agents to be used clinically along with amphotericin and the azoles. There have been a large number of reports of fungally produced antifungal agents; however, one of the biggest issues is to find agents that are selective against fungi without effecting the host. The structurally similar zaragozic acids and squalestatins, both inhibitors of squalene synthase and discovered by the groups at Merck and Glaxo independently, are some of the most potent and promising antifungal agents discovered. The squalestatins were discovered from fermentations of a Phoma sp. (Dawson et al., 1992; Sidebottom et al., 1992; Baxter et al., 1992) by investigators who had engaged in screening fungi for the production of inhibitors of squalene synthase. The zaragozic acids were initially discovered as metabolites from four fungi, Leptodontidium elatius, two strains of Sporormiella intermedia, and a sterile mycelia (Bergstrom et al., 1993).The L. elatius was isolated from a wood substrates collected in North Carolina, the S. intermedia were both isolated from dung collected in Arizona, and the Mycelia sterilia was isolated from a Spanish water sample. An excellent review of the discovery and structural, biosynthetic, and mechanistic studies on this group of compounds have been recently published by Bergstrom et al. (1995), and these exciting metabolites will not be discussed in depth herein. Zaragozic acids are a group of compounds with a similar basic structure with a considerable variety of minor modifications within these groups, and structures of A through F are given here (80-84). Zaragozic F is the first metabolite of this series to have a nonaromatic alkyl side chain attached to position 1 (Dufresne et d., 1996). The zaragozic acids are potent inhibitors of squalene synthase, with zaragozic acid A giving Kj values against rat liver microsoma1 enzyme of 78 pM (Bergstrom et al., 1993) to 1.6 nM (Hasumi et a]., 1993), and an IC5,, value of 12.5 nM (Baxter et al., 1992). At least 20 separate fungal isolates have been shown to produce members of the zaragozic acid/squalestatin group (Bills et al., 1994). An enlightening example of the importance of careful media preparation was given by

STRUCTURES 80-83.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

27

Connors et al. (1995),who demonstrated that, when a batch of medium was sterilized at a low heat input (R, = 33.4 min), the products from a subsequent fermentation differed from a batch prepared using a higher heat input (R, = 50.5 min), indicating significant thermal decomposition of some ingredient(s1. In the latter case, the yield of 4’-desacetoxy zaragozic acid C (85)was approximately 130 yg/ml compared to 60 yglml from normally sterilized media, although biosynthesis of the coproduced 4’-O-desacetyl zaragozic acid C (86)and zaragozic acid C remained the same. The different sterilization techniques also led to detectable changes in growth and metabolism of the fungus. The production medium contained lactose, glycerol, primatone, yeast extract, sodium citrate, and magnesium sulfate. The polyketide origins of zaragozic acid have been determined, and it was determined that the acetoxy group is derived from acetate (Byrne et a!., 1993). The 4‘-desacetoxy zaragozic acid C is probably a biosynthetic precursor to the other two compounds, and under the conditions established in this heat-altered medium, the production of this intermediate is in some way greater than the ability of the fungus to carry out the two-step conversion to zaragozic acid C. The related squalestatins (squalestatin 1= zaragozic acid A = 80)were first reported in 1992 (Dawson et al., 1992; Sidebottom et al., 1992; Baxter et al., 1992); these compounds are also very potent squalene synthase inhibitors. In addition to the naturally occurring compounds, to expand their analogue series the Glaxo group carried out microbial biotransformations on the squalestatins using actinomycetes and fungi and were able to prepare five novel compounds, all of which retained bioactivity (Middleton et al., 1995). Three thousand and five hundred cultures were screened to find these biotransforming microorganisms. Four of the fungi tested were able to deacetylate squalestatin (87),20% of the fungi examined were able to carry out the desacylation reaction (88),and a fusarium produced the methyl ester 89.The modified squalestatins were potent inhibitors of mammalian and fungal squalene synthase, with IC5,, values ranging from 3 to 30 nM. The echinocandins are also a very promising and potent group of antifungal agents, these lipopeptides being inhibitors of fungal glucan biosynthesis. Recent reviews have been published and should be consulted for in-depth information (Current et al., 1995; Debono, 1994, 1995; Tkacz, 1992). Briefly, echinocandin B is a product of Aspergillus nidulans and is a potent inhibitor of P-1,3-glucan biosynthesis, which results in it being also a potent inhibitor of Candida and Pneumocystis carinii (Buss and Waigh, 1995; Weinberg, 1996).

28 CEDRIC PEARCE

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

29

0

OH

STRUCTURES 87-89.

Other antifungal lipid-containing glucan biosynthesis inhibitors have been reported. The mechanistically related fusacandins 90 and 91 were isolated from Fusarium sambucinum (Jackson et al., 1995; Hochlowski et al., 1995). The producing fungi were isolated from a polypore fruit body collected in Illinois and was cultured in a production medium containing glucose, mannitol, glycine, dried lard water, soybean meal, sodium citrate, potassium phosphate, and cobalt chloride. The fusacandins, which are related to the papuolacandins, contain two galactose units, one glucose, and an aromatic moiety, with fusacandin A esterified with two long-chain fatty acids and fusacandin B with one. Fusacandin A had MIC values against a range of Candida and other fungi between

30

CEDNC PEARCE Fusacandins

STRUCTURES 90, 91.

3.12 and 6.25 pg/ml compared to papuolacandin B, which was generally twice as active as fusacandin B, which was much less active than the A component. These metabolites did not show any antimicrobial activity. Fusacandin A was also shown to be a potent inhibitor of glucan synthase, and the anti-Candida action was antagonized in the presence of the osmoprotectant sorbitol. When macromolecules were synthesized by Candid0 albicans in the presence of fusacandins, only glucan biosynthesis was significantly depressed, and it seems very probable that the fusacandins are inhibitors of (1,3)-P-glucansynthesis. Using a screen for identifying inducers of microbial differentiation in Phoma destructiva, Dornberger et al. (1995) discovered a strain of Apiocrea chrysosperma that produced the bioactive peptaibol metabolites chrysospermins 92-95 when incubated in a static medium contain-

W

CI

g

Chrysosperims and Trichorzins Chrysosperims and Trichorzins

93

AcPhe

94 - _ IAcPhe

Aib

Ser

Aib

I Aib I Ser I Aib

Aib Iva

Leu

Leu

Gln

Gly

I Gln I Gly

G r:

Aib

Aib

Ala

Aib

Aib

Pro

Iva

Aib

Aib

Gln

Trpol

Aib

Aib

Ala

Aib

Aib

Pro

Aib

Aib

hi Gln

Trwl Tpl

95

AcPhe

Aib

Ser

Aib

Iva

Leu

Gln

Gly

Aib

Aib

Ala

Aib

Aib

Pro

Iva

Aib

Aib

Gln

96

AcAib

Gly

Ala

Aib

A%

Gln

Aib

Val

fib

Gly

Leu

Aib

Pro

Leu

Aib

Aib

GIn

Leu01

STRUCTURES 92-97. STRUCTURES 92-97.

$D

F

32

CEDRIC PEAFCE

ing malt extract, glucose, yeast extract, and ammonium sulfate. The screening bioassay consisted of malt agar plates containing vegetative mycelia of r! destructiva. Samples were spotted on the surface of these plates, and cytodifferentiation was measured by the zone of pigmentation produced around the site of sample application. Chrysospermins were subsequently shown to be active against a variety of Gram-positive bacteria and yeast with MIC values from 5 pg/ml against Gram-positive organisms to 10-12.5 pg/ml against Sporobolomyces salmonicolor and Phoma destructiva. Nine somewhat related antifungal peptaibols, the trichorzins and 97,have been isolated from Trichoderma harzianum (Goulard et al., 1995; Hlimi et al., 1995). In this case, the metabolites contained 18 amino acid residues compared to 19 in the chrysospermins, so that belong to the recognized group of long-sequence peptaibols. Both groups contain similar residues at positions 4, 5, 9, 15, and 16. The trichorzins, which were isolated from two strains of T harzianum from soil collected in Uruguay and Malaysia, were produced in a defined medium containing glucose as the carbon and energy source and potassium nitrate as the nitrogen source, together with inorganic salts. Clearly, in a defined medium as the one described, the opportunity for directed biosynthesis of novel analogues exists, although these authors did not report on such experiments. The trichorzins were not active against Gram-negative bacteria, although they were active against Staphylococcus aureus. Since MIC values were not reported, an accurate determination of potency could not be acquired, although the zones of inhibition for the amount of material added did not seem great. Taking the approach into account, the results are not simple to interpret. These authors also performed experiments to determine the membrane effects of the various components isolated and by measuring the leakage of fluorescent dye from liposomes demonstrated a direct relationship between increased hydrophobicity and increased potency. Employing a physicochemical screening approach, Grafe et al. (1995) discovered the lipopeptide helioferins 98 and 99,which mediated the transfer of ions from an aqueous environment to an organic phase. The producing organism, Mycogone rosea, was isolated from the fruiting body of a Macrolepiota sp. and was grown as a static surface culture in a medium containing malt extract, glucose, yeast extract, and ammonium phosphate. Helioferin A and B are potent antibacterial and antifungal agents with an MIC value against C. albicans of 5 pg/ml. These compounds are toxic to chicken embryos at 0.5 mg/kg, cause hemolysis at > l o 0 pg/ml, and are cytotoxic against L I Z 1 0 leukemia and L929 mouse fibroblasts, with IC,, values of 0.01-0.04 pg/ml. The helioferins

I"

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

O = ? ( :I2

33

34

CEDRIC PEARCE

OH

OH

(105) (2-12371

R=H

(106) CJ-12372 R=OH

STRUCTURES 100-106.

are linear aminolipopeptides and are thought to work as protonophoric compounds disrupting ion flux and energy metabolism. A number of compounds containing bicyclic decalin and related structures with a variety of bioactivity, but commonly antifungal, have been reported recently. Thus, the diepoxins 100-103 (Schlingmann et

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

35

(107) Australifungin

OH (108) Australifunginol

STRUCTURES 107, 108.

al., 1993) were obtained from fermentations of a Mycelia sterilia initially isolated from a Panamanian tree sample at MYCOsearch. Of the four diepoxins reported, 101was inactive, but the remaining three had some antibacterial activity and inhibited C. albicans and Rhodotorula rubra. As well as producing the diepoxins, these fermentations contained the very active allenic polyacetylene, 3,4,5,6-tetrahydro-6-hydroxymycomycin (104)(Schlingmann et al., 1995). The MIC values for this compound were 2 4 pg/ml against Staphylococcus aureus and B. subtilis and 10 pg/ml against C. albicans. The allenic polyacetylene are very unstable compounds and would be difficult to develop further. Compounds structurally related to the diepoxins, CJ-12,371(105)and CJ-12,372 were isolated by Sakemi et al. (1995) as inhibitors of DNA gyrase. These compounds were also active against Gram-positive bacteria with MIC values ranging from 25 to 100 pg/ml. No other activity was reported. The diepoxins CJ-12,371/2 and Sch-50673/Sch-50676 (discussed in the Section IX) are all related to the preussomerins, which are antifungal agents isolated by Weber and Gloer (1991) from a coprophilous fungus. Biosynthetically, these compounds could be products from arrested or diverted melanin production and as such could be common metabolites from pigmented fungi. A more complex group containing the decalin-type motif, australiand australifunginol m), has been reported by Mandala fungin (107) and co-workers (1995) from fermentations of Sporomiella australis. This culture, isolated from moose dung, was initially cultured in a solid

(m),

36

CEDRIC PEARCE Fusarielins

HO

\

HO

\

HO

HO

H

+ '...,

...'

OH

H

0

(110) B

H

& 0...."

H

0

/

.. (112) D

(1ll)C

STRUCTURES 109-112.

medium containing cracked corn, ardamine, and inorganic salts, but, subsequent to the discovery of the activity, several liquid media were tested and one containing either mannitol or fructose, oat flour, fibco yeast, L-glutamic acid, and MES proved to support a four- to fivefold increase in metabolite. Australifungin was a very potent antifungal agent, giving MIC values of less than 1pg/ml against Candida pseudotropicalis, C. tropicalis, Cryptococcus neoformans, C. Albicans, and Aspergillus fumigatus. Australifunginol was a relatively weak antifungal agent.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

37

Certain antimitotic drugs cause deformation of Pyricularia oryzae, and this was used as the basis for a bioassay to screen for compounds interfering with microtubule function (Kobayashi et al., 1995). This approach led to the discovery of a series of antifungal agents from a Fusarium sp. originally isolated from a soil sample. When this fungus was cultured in a stationary potato dextrose broth, fusarielin A-D (109-112) were produced. Biosynthetic studies using I3C-labeled precursors demonstrated that these compounds are derived via a decaketide that undergoes five methionine-derived methylation reactions. Significantly, these results show that, while fusarielin contains the decalin motif, albeit somewhat disguised, it is not derived from a decalin precursor since the labeling pattern reported would be different if it were. Fusarielin A showed some antifungal activity but was not highly potent and had a narrow spectrum. This compound also did not inhibit microtubule assembly, thus showing the deformation assay employed probably has a more complex mechanism than speculated. The respiration inhibitors strobilurins and oudemansins are an exciting group of methoxyacrylic acids produced by a number of higher fungi, which have received considerable attention because of their antifungal, insecticidal, antiviral, and antitumor activities. These compounds were first reported in 1978, and work in this area has been reviewed elsewhere (Anke, 1995; Clough, 1993). New bioactive members of this class are still being discovered. A submerged culture of the basidiomycetes Favolaschia pustulosa was shown to produce 9-methoxystrobilurin L (113)and 9-methoxystrobilurin E (114) along with (Wood et al., 1996). The fungus had been isolated oudemansin L from fruit body tissue collected in a tropical forest. These compounds had some antibacterial and antifungal activity. 9-methoxystrobilurin L was shown to be cytotoxic, with an ICs0 against the human B lymphoblastoid cell line of 1.8 nM. 9-methoxystrobilurin A and K have been reported as metabolites from Favolaschia species (Zapf et al., 1995a). This latter group also isolated a sterol, favolon (m), from the same fungus (Anke et al., 1995). Favolon was produced in a medium containing yeast extract, malt extract, and glucose, and was present in the mycelia. Favolon displayed potent antifungal activity when tested against ascomycetes, basidiomycetes, oomycetes, and zygomycetes. The sterol was not active against bacteria, neither was it active against L1210 cells. This research group also isolated a new strobilurin, hydroxystrobilurin A (117),which was reported from basidiomycetes Pterula spp. (Engler et al., 1995). A mycelial culture obtained from the fruiting body of a fungus collected from a German forest was grown in a medium containing yeast extract, malt extract, and glucose in an aerated fermen-

m)

38

CEDRIC PEARCE

OCH3

(115)

STRUCTURES 113-115.

ter. The antifungal activity of hydroxystrobilurin A was similar to that reported for other strobilurins and oudemansins, with a wide spectrum of filamentous fungi and yeasts being inhibited, although, significantly, the additional hydroxyl group in hydroxystrobilurin results in decreased potency. New compounds have been discovered by investigating mutants of known producing strains in certain cases. Such was the case when

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

39

( 1 16) Favolon

CH3

STRUCTURE 116.

Sakuda et al. (1995) investigated a mutant of the patulin-producing Penicillium urticae, which was shown to produce a new azaphilone epoxide, patulodin (117a). This compound was shown to be weakly antifungal, having an MIC value of 50 p g h l against Pyricularia oryzae. As with the discovery of antiinsectan compounds (discussed in Section XI), Gloer and co-workers (1995a, 1996) have used an elegant discovery approach directed by nature and have sought antifungal compounds from fungi living in environments wherein the fungal inhabitants might be reasonably expected to compete for nutrients with each other. This area of research has been reviewed (Gloer, 1995a,b, 1996) and is only briefly discussed here. That there is interspecies chemical inhibition has been observed in many fungal communities. Gloer and his associates have been examining fungal isolates from dung that is sequentially colonized by different organisms. By looking for interspecies antagonism, a very large number of antifungal activities have been observed. Of approximately 250 coprophilous fungi examined, more than 60% were shown to produce antifungal metabolites (for more details of this work, see Gloer, 1996). The preussomerins previously noted as being the first of a new class of fungal metabolite, were isolated from Preussia isomera (Weber et a]., 1990; Weber and Gloer, 1991). Several of the preussomerins showed antifungal activity against fungi with which the producing culture may have to compete in nature, although activity against C. albicans was not observed. More recently, the antifungal agents cercophorins A-C (118m)were obtained from the coprophilus fungus Cercophora areolata in an isolate hom porcupine dung, along with the known trichothecene roridin E (Whyte et al., 1996). This fungus was used to ferment potato

m,

40

CEDRIC PEARCE

(1 17) Hydroxystrobilurin

(117a)Patulodin

(1 17b) Preussomerin A

OH

OH

OH STRUCTLJRES 3 17,11 7A,11 7B.

dextrose broth in shaken flasks at 25-28°C. The cercophorins were tested against dung from early successional fungi Sordaria fimicola and Ascobolus furfuraceus, and shown to inhibit the growth. These compounds were also active against some Gram-positive bacteria, with

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

41

(3343 (1 18) CerwphorinA

CHg (119) Cercophorin6

c33-43

(120) CerwphorinC

cercophorin A being the most potent, and had limited activity in the NCI tumor cell line. This appears to be a very promising and productive area of research. Mycophilic fungi associated with the fruiting bodies of either ascomycetes or basidiomycetes have been the subject of investigation (Zapf et al., 1995b), and this has led to the discovery of some interesting antifungal compounds. Sphaerellopsis filum, a widespread rust fungus mycoparasite, had been considered a biocontrol agent, although the chemistry of the toxins produced and the precise mechanism for toxicity was unknown. This culture produced antimicrobial bioactivity, which was isolated and shown to be due to the production of two xanthocillins, darlucins A (121) and B (122)(Zapf et al., 1995b). The darlucins were active against yeast and filamentous fungi, with the most potent MIC values being 2.5-5 yg/ml. These compounds also displayed antibacterial activity, with MIC values between 2.5 and 20 yg/ml against both Gram-positive and -negative bacteria, and showed some cytotoxicity against the BHK21, HeLaS3, L1210, and HL60 cell lines.

42

CEDRIC PEARCE Darlucins

OH

(121) A

NC

OH

(122) B

HO (1 23)Sch-57404

STRUCTURES 121-123.

In a recent preliminary report, an antifungal agent, Sch-57404 (123) containing the uncommon sodaricin nucleus was isolated from an unidentified fungus by Coval et al. (1995). This compound is only the second fungal metabolite having the unusual tricyclic nucleus shown. Sch-57404 was shown to be active against C. albicans, with MIC values of 16 pg/ml. As was demonstrated with the cyclosporines, fungally derived antifungal agents sometimes also demonstrate immunosuppressive effects with startling significance. Thus, cyclosporine A was initially identified

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

43

as an antifungal, and its effects on the immune system were recognized during further pharmacological investigation. The known ISP-1 (124) was isolated from Zsaria sinclairii by Fujita et al. (1994). This had previously been identified by two groups (Kluepfel et al., 1972; Aragozzini et al., 1972) as an antifungal agent, but Fujita and co-workers discovered this compound in a screen for immunosuppressive agents, and ISP-1 was reported to be 10 to 100 times more potent than cyclosporine A in suppressing lymphocyte proliferation in vitro and in other in vivo bioassays. This same research group also discovered the mycestericins 125-128 from a Mycelia sterilia (Sasaki et al., 1994; Fujita et al., 1996). These compounds were not as potent as ISP-1 on mouse allogenic mixed lymphocyte reaction, with mycestericins D and E being perhaps half as active, but this series of metabolites did allow for some speculative SAR conclusions. Fungi have been the source of many important antibacterial agents, including the penicillins and cephalosporins, both of which are used heavily 50 to 60 years after their discovery. However, from the reports in the literature there are not many novel antibacterial agents being isolated from fungi. Some of the antibacterial activity reported has been incidental to other biological activities, but these data demonstrate a level of interest in the area. With the emergence of antibiotic resistance, however, it would not be unexpected if more attention was paid to this area by drug discovery groups, The producing fungi may be in collections waiting to be tested. For example, a strain of the penicillin-producing fungus Penicillium chrysogenum was recently shown to produce a quinone, sorrentanone (m), that was active against Gram-positive and -negative bacteria, with MIC values of 16 pg/ml and higher (Miller and Huang, 1995). This is an area that may be worth revisiting simply because the targets have changed. There are a number of major viral diseases for which there is little or no treatment. Common viral diseases that affect huge populations include influenzas and colds. Progress has been made in some areas (e.g., human immunodeficiency virus and herpes), although none of the current usual chemotherapeutics are from natural sources. Of the few antiviral microbial products reported in the literature, the majority have been isolated from bacteria rather than fungi (Takeshima, 1992), and the rate of new antiviral fermentation products appearing in the literature seems to be increasing, which may be a reflection of better approaches to screening. In the HIV area, two reported screening strategies have led to the discovery of new products. During the HIV replication cycle there is a complex series of interactions between the virus and the host, including

44

CEDRIC PEARCE

0

HO

(124) ISP-1, Myriocin, Thermozymocld ' in

-

(125) Mycestericin D R= CH=CH-(CH2)6-Co-(CH2)5-CH3

0

0

0 (129)sonentanone

STRUCTURES 124-129.

specific binding of viral RNA with viral and host proteins. The viral protein, REV (for regulation of virion expression), regulates transport of viral RNA into cytoplasm, and the interaction of REV with the REV response element is essential for REV action. Using an approach based on the assumption that inhibition of this interaction could yield antivi-

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

45

(130) Harziphilone

0

(131) Fleephilone STRUCTURES 130, 131.

ral compounds, a productive screening program was established at Bristol-Myers Squibb. Thus, from a Trichoderma hQrZiQnZZmthat had been isolated from a Floridian soillplant sample, two active comwere obtained when pounds, harziphilone (130)and fleephilone the fungus was cultured in a medium containing glycerol, glucose, polypeptone, yeast extract, sodium chloride, and calcium carbonate (Qian-Cutrone et ~ l .1996a). , Interestingly, these metabolites show only slight structural similarity, although both inhibit binding of REV protein to REX response element, with ICs0 values of 2.0 pM for harziphilone and 7.6 pMfor fleephilone. In a secondary evaluation used to determine the effectiveness against virus in infected cells, in this case CEM-SS cells infected with HIV-1, neither compound had an effect using concentrations up to 760 pM harziphilone and 450 pM fleephilone. In this regard, it is interesting to note that an equally potent plant product, niruriside, which gave an ICs0value of 3.3 pMin the REV/REV response element assay, also failed to protect chronically infected cells at 260 pA4 (Qian-Cutrone et d., 1996b). In a second approach, inhibitors of gplZ0 binding to CD4 receptors were sought. In this case, it is speculated that, during the entry phase of virus to cell, the viral protein gplZ0 binds to the cellular surface CD4

(m),

46

CEDRIC PEARCE (132) lsochromophilone I

(133) lsochromophilone II

(135)Trypostatin B

( 134) TrypastatinA

STRUCTURES 132-135.

molecule, and compounds that inhibit this process should be antiviral (Johnson and Hoth, 1993). Omura et al. (1993a) and Matsuzaki et al. (1995a,b) reported the first hngally derived inhibitor of this process that also showed significant effects on viral replication. Fermentation of Penicillium multicolor isolated hom a Japanese soil sample produced together with the known inacisochromophilones I (132)and I1 tive metabolites sclerotiorin, ochrephilone, and rubrorotiorin. Isochro-

(m),

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

47

mophilones I and 11 inhibited gp120-CD4 binding with ICs0 values of 6.6 and 3.9 pA4, respectively. Isochromophilone I1 significantly inhibited HIV replication, as determined by measuring viral core protein p24 production in human lymphocytes. None of the isolated compounds had any anti-Gram-positive or -negative bacterial or antifungal activity. IX. Antitumor and Cytotoxic Activity

It is well known that natural products have antitumor activity and that some of the most useful drugs available are, either directly or via derivatization, from these sources (e.g., Vinca alkaloids). The podophyllotoxins etoposide and teniposide; taxol from the Pacific yew; actinomycin D, bleomycin, mithramycin, mitomycin, daunorubicin, and doxorubicin all from Streptomyces species (Goodman Gilman et a]., 1985). A variety of novel fungal products have been reported that add to the spectrum of potential approaches to the chemotherapy of cancer. Tryprostatins 134 and 135 are inhibitors of the mammalian cell cycle produced by a marine fungus (Cui et al., 1995). The producing culture, Aspergillus fumigatus, was isolated from a sediment sample collected at a depth of 760 m at the mouth of the Oi river in Japan. Fermentation of a medium containing glucose, starch, soybean meal, potassium phosphate, and magnesium sulfate produced the two active components that inhibited the cell cycle progression of the mouse tsFT210 cell line in the G2/M phase at concentrations of 50 pg/ml (A) and 12.5 pg/ml (B). In a screen to discover compounds that would inhibit metastasis, two novel compounds of preussomerin type were isolated from Nattrassia mangiferae (Chu et al., 1995a). This culture, also from the MYCOsearch collection, was isolated from dead leaves collected from an arid region of Guatemala. Fermentation of a medium containing neopeptone, cerelose, and calcium carbonate led to the production of Sch-50673 (136) and 50676 (137).These compounds are related to a number of recently discovered bioactive fungal metabolites, including the antimicrobials CJ-12371/12372 and the diepoxins discussed earlier, and were also isolated from fungi from the MYCOsearch collection. Sch-50673 and 50676 were active in a chamber invasion assay using HT1080 human fibroblast cells, with ICs0 values of 6.2 and 2.8 pM,respectively. Less than 10% toxicity was reported at 25 pM. Further evaluation of these compounds was not reported. Many tumors are either not susceptible to cytotoxic agents or develop resistance once exposed to the agent. A large amount of research has been carried out to overcome this latter problem. Nozawa et al. (1995)

48

CEDlUC PEARCE

(137) sch 50676

(136) sch 50673

+

CH3

(138) FD-211

CH3

(139)SCh 52900 R=CH(OH)CHa (140) SCh 52901 R=CH*CHs (141) Verticillin R Z H 3

STRUCTURES 136-141.

discovered a strain of Myceliophthora lutea isolated from a Chinese soil sample that produced a lactone compound, FD-211 which was active against adriamycin-resistant human promyelocytic leukemia (HL-60) cells. Thus, while the IC5,, values against a number a cell lines were about 8 to 10 times less potent than adriamycin, against the adriamycin-resistant cells FD-211 was 20 times more potent than adriamycin. FD-2 11inhibited incorporation of radiolabeled precursors into DNA, RNA, and protein in HeLa cells. It would be very instructive to learn more about the mechanism of this compound.

(m),

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

49

During the transition from quiescent to growing state, one of the earliest biochemical events is for the c-fos proto-oncogene to be induced, and this is triggered by a variety of mitogens. significantly, it has been shown that some oncogene products also stimulate production of c-fos,and in certain tumors c-fos is constitutively elevated. For this reason it has been postulated that inhibition of c-fos induction could be an important target for drug discovery. In the program at ScheringPlough, a Gliocladium sp. was shown to produce compounds Sch52900 (139)and Sch-52901 (140)that displayed such activity (Chu et al., 1995b). The culture, Gliocladium sp. SCF-1168, was isolated in the MYCOsearch program from fresh leaf litter collected in a Puerto Rican rain forest. When a medium consisting of neopeptone, cerelose, and calcium carbonate was fermented, the Gliocladium produced three bioactive compounds, two of which were novel, and a third, verticillin that had been reported previously from a Verticillium sp. (Katagiri et al., 1970). Using a validatedfos/lac Z reporter murine system, the two new metabolites were shown to be inhibitory, with ICS0values of 1.5 and 18 pM for 52900 and 52901, respectively, while verticillin had an ICS0of 0.5 pM. Activation of the Ras gene is associated with many human cancers. Biochemical studies have shown that Ras functions following localization within the membrane and binds with GTP prior to transforming cells. Protein farnesyltransferase is involved in farnesylation of Ras proteins, and inhibitors of this enzyme are predicted to change membrane localization and activation of Ras proteins, and will possibly have utility as anticancer agents (Gibbs, 1991). A number of fungal metabolites have been isolated that show this activity, including the wellknown gliotoxins (Van Der Pyl et al., 1992), the andrastins (Shiomi et al., 1996), kuraosins (Uchida et al., 1996a,b) and fusidienol (Singh et al., 1994a,b), for example. The kuraosins 142 and 143 (Uchida et al., 1996a,b)were produced by a Paecilomyces strain that had been isolated from a Japanese soil sample. When the fungus was grown in a medium containing starch, glycerol, soybean meal, fermipan, potassium chloride, calcium carbonate, magnesium sulfate, and potassium phosphate, karaosins A and B were produced. Structurally these compounds are relatively simple and were subsequently synthesized by the discovery group (Uchida et al., 1996b). The ICS0values against protein farnesyltransferase were 59 and 58.7 pM for the A and B isomers, respectively. In a second study, the French group at Rhone-Poulenc Rorer (Van Der Pyl et al., 1995) isolated a structurally more complex metabolite from Chrysosporium lobatum, also found in a soil sample. This compound, RPRl13228 (M), was produced in a liquid medium

(m),

CEDlUC PEARCE

50

HO

(1 42)Kurasoin A OH

(143)Kurasuin B

(144) RPR113228

(145) Fusidienol

&

HO

STRUCTURES 142-145.

containing malt extract and agar. Inhibition of farnesyl protein transferase using either lamin B terminal sequence peptide or recombinant p21H-ras protein was examined, and IC,, values of 0.83 and 2.1 pMwere obtained. In an investigation by Merck scientists, the tricyclic compound fusidienol (145) was discovered from a Fusidium griseum originally isolated at MYCOsearch (Singh et al., 1994a,b). E griseum produced fusidienol when cultured on a millet-based medium, and purified material was active against bovine and human enzyme, with IC5, values of 0 . 3 and 2.7 pM, respectively.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

51

(146) Taxol (147) 2a-hydroxydimeninol

HO.,,

STRUCTURES 146,147.

Supply is one of the inherent problems encountered while working towards bringing natural products to the clinic, and this is frequently the case with more exotic sources as well as with plant material. This is somewhat less of an issue with microbial fermentations since these are renewable, but they can never be ignored as a potential snag. This was highlighted during the development of the plant antitumor product taxol (146). With the discovery of an endophytic fungus, Taxomyces andreanae, associated with the Pacific yew Taxus brevifolia, which would produce taxol in culture (Stierle et al., 1993),certain expectations were raised about potential relatively facile production methods. This work has stimulated more investigations on the fungi associated with ?: brevifolia. In a report from Pulici et al. (1996),a Pestalotiopsis sp. isolated from Pacific yew leaves collected in Bozeman, Montana, was shown to produce the sesquiterpene 2-a-hydroxydimeninol (W), a compound related to a number of plant and fungal drimanes. This and related observations are very interesting in light of the speculation concerning the transfer of biosynthetic information between plants and fungi, although definitive proof that this occurs is lacking. In this report, 2-a-hydroxydimeninol was not shown to have any particular activity, but related compounds have been shown to be antitumor and antifeedants, and it would be interesting to test for these activities. X. Miscellaneous Pharmaceutical Activity

Platelet activating factor, l-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine (PAF),is a phospholipid involved in a number of physiological

52

CEDlUC PEARCE

(148) Phomactin E

(149) Phomactin F

(150)PhomactinG

(151) Phomactin B

STRUCTURES 148-151.

events, both normal and disease-associated. These responses include degranulation of a variety of blood cells, smooth muscle contraction and bronchoconstriction, vascular permeability and hypotension, and anaphylaxis, and are associated with inflammation and allergic reactions. It has been speculated for some time that PAF inhibitors would have a role in the control of disease, and a number of PAF antagonists have been discovered and tested clinically for effectiveness. A series of macrocyclic compounds have been reported from Phoma spp., one isolated from a marine environment and the other from a MYCOsearch terrestrial collection. The phomactins were initially reported in 1991 (Sugano et d.,1991) and further metabolites (148-151)in 1995 (Sugano eta]., 1995). These compounds seemed to be relatively potent, with IC50 values of 2.3, 3.9, and 3.2 pM for PAF-induced platelet aggregation by E, F, and G, respectively, and ICs0 values for inhibition of PAF binding to the receptor of 5.19, 35.9, and 0.38 pM for E, F, and G, respectively. Related metabolites Sch-47918, 49026, 49027, and 49028 (152-155) were isolated by Chu et al. (1993) from a fungus from the MYCOsearch collection. These compounds gave a range of ICsOvalues from 3 to 30 pM.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

53

(153) SCh 49026

,,+OH

@

0 H H

(154)Sch49027

OH

(156) ZG-1494a

STRUCTURES

152-156.

An alternative approach has been explored by West and colleagues screened for PAF biosynthetic inhibitors. PAF is biosynthesized via two systems. The first is the normal de novo pathway, and the second, the route used in disease response, is the remodeling pathway. The reason this group looked for biosynthetic inhibitors rather than antagonists is because PAF plays a role in normal metabolic function, and to block it may not be optimum for good health. From their screening program an inhibitor of PAF acetyltransferase, ZG-1494~~ (156) was discovered. The producing organism was Penicillium rubrum isolated from red pepper in Denmark. ZG-1494a selectively inhibited PAF acetyltransferase with an IC5,, of 40 pM. More(1996), who

54

CEDRIC PEARCE

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

55

over, the compound was shown to be sevenfold less inhibitory against the de novo system, with an ICs0 of 304 yM. In a further series of experiments to determine the antiinflammatory effect of ZG-l494a, this compound showed inhibition of PAF binding for PAF sites, pyrilamine binding to histamine HI receptors, and dexamethasone for glucocorticoid receptors, all with an ICs0 of 3 pM.These data suggest that ZG1494a may be a useful antiinflammatory agent. Blood coagulation has been implicated in a variety of conditions. Fungal metabolites that inhibit blood clotting and compounds that promote clot removal have been reported. Thus, a number of serine protease inhibiting asterriquinone (157-161) pigments were isolated from two Humicola fuscoatra, a Humicola grisea, a Botryotrichum sp., and an Aspergillus terreus, all of which had been isolated from soil samples collected in Mexico or Nevada (Mocek et al., 1996). The asterriquinones CT1-CT5 were specific inhibitors of certain serine proteases involved in blood coagulation. Factors VIIa, Xa/Va and Xa were inhibited, with ICs0 values between 3 and 135 yg/ml for those that were active, whereas thrombin and trypsin were very much less sensitive to these metabolites. Once blood clots have formed, the fibrinolytic system dissolves them over time. Plasmin, the primary degradation enzyme, is derived from plasminogen, which itself is activated by tissue-type activators, and this is accelerated on fibrin surfaces. Increasing the binding of plasminogen to fibrin is expected to enhance the whole system. A metabolite from Stachybotrys microspora has been isolated that increases the binding of plasminogen to fibrin (Shinohara et al., 1996). is a triprenylphenol shown to increase the binding of Staplabin (162) radiolabeled plasminogen to fibrin at concentrations between 0.3 and 0.6 mMto 110-270% of the control. The binding of labeled plasminogen to human U937 cells was increased approximately twofold by 0.37 nM staplabin. No further details of bioactivity were given in this preliminary communication. Platelet aggregation inhibitors have been found from extracts of Beauvaria bassiana (Kagamizono et al., 1995), the bioactive compound being a phenylalanine derivative, bassiatin (163). The producing culture was isolated hom a Chinese soil sample. Bassiatin inhibited rabbit platelet aggregation induced by ADP, collagen, and arachidonic acid, with ICs0 values of 0.19 mM, 0.38 mM, and 0.38 nM, respectively. Bassiatin did not inhibit ACAT at concentrations up to 10 yM. Cell adhesion and cell adhesion molecules have been implicated in various diseases, and this has created an interest in antiadhesion compounds. Hayashi et al. (1995) have isolated antiadhesant compounds from a Microsphaeropsis culture obtained from a soil sample. The two

56

CEDRIC PEARCE

OH

(163) Bassiatin

OH I

OH I

0

0

HO H3C

0U

C

H

s

(164) MacrospheliieA

(165) MacrosphelideB

0

H (166) Eporactaene

STRUCTURES 162-1 6 6.

macrosphelides (164 and 165)were tested in an adhesion assay using HL-60 cells that bind to human umbilical vein endothelial cells, and were shown to inhibit this process, with values of 3.5 and 36 pM for A and B, respectively. These compounds did not show cytotoxicity against a panel of human cell lines, and no acute toxicity was observed in mice at 200 mg/kg. Macrosphelide B was weakly antimicrobial, but the A component showed no activity against any of the bacteria or fungi used. The authors suggest it would be interesting to test these compounds against metastasis and tumor invasion. has been Finally, a very interesting compound, epolactaene reported from a Penicillium species isolated from a ocean sediment

(m),

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

57

sample (Kakeya et af., 1995). Epolactaene, which was produced in a complex medium containing glucose, starch, soybean meal, polypeptone, meat extract, yeast extract, potassium phosphate, magnesium sulfate, and 0.2% sodium chloride, causes changes in the morphology of SH-SY5Y cells, a neuroblastoma cell line. In control experiments, less than 5% of cells produced neurites from the cell bodies, but in the presence of 2.5-10 pg/ml of epolactaene 74% of the cells produced neurites. No further results were reported.

XI. Agriculturally Active Compounds

There is a long history of using natural products for agricultural and animal applications: for example, avermectins for antiparasitic compounds, and kasugamicin for rice blast disease. Natural products have been shown to be ideal agrichemicals in a number of ways, especially given their high potency and specificity. Arthrobotrys oligospora was first shown to trap nematodes in 1888 (Zopf, 1888), and subsequent study has demonstrated that this may be useful in the control of nematodes. Some chemical analyses of metabolites from Arthrobotrys oligospora have been reported, but only recently have nematocidal compounds been discovered. Thus, Anderson et a f . (1995) isolated a series of related compounds (167-172), including oligosporon, oligosporol A and B, 4’,5’-dihydro-oligosporon, hydroxyoligosporon, and lO’,ll’-epoxyoligosporon,from cultures of an Australian isolate of Arthrobotrys oligospora. Three of these compounds-oligosporon, oligosporol A, and oligosporol B-had been previously isolated (Stadler et af., 1993c) from another strain of Arthrobotrys ofigospora, although in this latter paper activity against Caenorhabditis elegans was not demonstrated. In the work by Anderson et al. (1995), the compounds were tested against an intestinal parasitic nematode, Haemonchus contortus, with oligosporon and 4’,5’-dihydro-oligosporon having LDS0 values of 25 and 50-100 &ml, respectively. The authors propose that in nature these compounds may play a role in the interactions of the fungus and its prey, as well as showing the potential for protection against infestations of nematodes. In a series of papers from the Kitasato Institute in Japan, a number of novel anticoccidial fungal and actinomycetes products have been reported. Fudecalone (173) was isolated from a Penicilfium sp. obtained from a soil sample collected in Shizuoka (Tabata et al., 1995b). This is a relatively simple terpene-type metabolite, structurally related to the plant products corymobtins and clerodane, and produced in a complex

58

CEDRIC PEARCE

0 (167) Oligosporin

0 (168) Oligosporol A

0 (169) Oligospord B

STRUCTURES 167-169.

medium containing starch, glycerol, soybean meal, yeast, potassium chloride, potassium phosphate, magnesium sulfate, and calcium carbonate. Fudecalone completely inhibited monensin-resistant Eimeria tenella at 16 pM, but no activity was observed against Gram-positive and -negative bacteria, or fungi. The bicyclic arohynapenes 174 (Masuma et al., 1994; Tabata et al., 1995a) were also isolated from a Penicillium sp. FO-2295 and shown to be active against Eimeria tenella. Of the four metabolites reported, arohynapene D was active against monensin-resistant organisms, with a minimum effective concentration of 0.51 pA4, and showed cytotoxicity against the BHK-21 cell line used at 1.0 pM. A screening program for ascomycete and basidiomycetes metabolites with activity against nematodes led to the discovery of a number of active compounds from Lachnum papyraceum. These include lachnumycorrhizin A chloromycorrhizin mon (175),lachnumol A A (m), and dechloromycorrhizin A (179) (Stadler et a]., 1993a,b).

(m),

(m),

59

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

(170) 4,5diydrc-oligospwon

(171) hydroxyoliispron

0 (172) 10,l I-epoxyoligospron

STRUCTURES 170-1 74.

Noting that these fermentation products contained chlorine atoms, these authors have investigated the effects of including other halogens in the production medium (Stadler et al., 1995a,b). In a medium in which the usual calcium chloride was replaced with calcium bromide, the fungus grew poorly and produced no significant secondary metabolites. However, when calcium chloride was used together with either calcium bromide or calcium fluoride, growth and appearance of the culture was similar to that of the usual fermentation. Although the fluoride-containing medium showed no difference in secondary metabolite profile, the inclusion of calcium bromide led to several significant

60

CEDRIC PEARCE 0

OH

CI

(175) Lachnumon

(176) Lachnumol A

0 (177) MycorrhizinA R,=H. R&l (178) Chloromyarrhizin A R j = R+I (179) dechlorormyarrhizinA Rl=R+-I

(169) Ghydroxymdlein R=H (181) 4-chloro-6,Edihydroxy-2-fnethylisochroman-l-one R=CI

OH

(182) 4-bromc-6,Edihydm~3-me~ylisochroman-l-one R=CI

CH30 R'

H

(183) Ehydroxy-Smethoxy-3-methylisochroman-1 one

R=H

(184) 4-chlorc-Ehydroxy-6-methoxy-3methylisochroman-l-one R=CI

(185) 4-chlorc-5.6,&trihydroxy-3-1nethyii~hroman-l -one

STRUCTURES 175-185.

changes. Thus, lachnumon, chloromycorrhizin, and dechloromycorrhizin were undetectable, while lachnumol A and mycorrhizin were produced as very minor components. Instead, five new metabolites were produced and a sixth, originally present as a very minor component, was produced at concentrations high enough for isolation and characterization. These compounds were the previously known 6-hydroxymellein (180),which had been isolated from a number of sources

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

61

including Gilmaniella humicola, and the following compounds previously unknown as natural products; 4-chloro-6-hydroxymellein 4-bromo-6-hydroxymellein (E), 6-methoxymellein 4-chloro-6(185). These methoxymellein W),and 4-chloro-6,7-dihydroxymellein compounds were weakly nematocidal, as well as showing some cytotoxicity, phytotoxicity, and antimicrobial activity. From fermentation and biosynthetic perspectives these results are very interesting. In the biosynthesis of mycorrhizins and similar compounds in Gilmaniella humicola (Chexal et al., 1979), it has been proposed that 6-hydroxymellein is an intermediate that is decarboxylated and prenylated. The lauchnumols and lachnumons are probably derived from a similar pathway, which in the presence of calcium bromide is inhibited. This leads to buildup of the intermediate 6-hydroxymellein, which is a substrate for more unusual halogenation, hydroxylation, and methylation reactions. Similar observations have been made on the effects of enzyme inhibitors on the metabolism and final products from actinomyces (Pearce, 1995; Pearce et al., 1991,1995, 1995), and this appears to be an area for further application by those interested in discovering novel metabolites from known organisms. Since calcium bromide appeared to inhibit the formation of derivatives of lachnumon or mycorrhizin, and because halogenation steps frequently occur towards the end of the biosynthetic route, it was reasoned that, if this was added after the end-products had started to appear, there may be a different effect on the final metabolites produced. When calcium bromide was added to 10-day-old fermentations, Lachnum papyrcreceurn produced eight novel metabolites, four of whichlachnumon B 1 (=), lachnumon B2 (=), mycorrhizin B 1 (=), and brominated (Stadler et al., 1995c,d,e).The mycorrhizin B2 (=)-were nonhalogenated compounds 1’Z)-dechloromycorrhizin A (190) and papyracons A, B, and C (191-193)are related structurally to mycorrhizins. All eight compounds had activity against Caenorhabditis elegans, a limited spectrum of fungi, were cytotoxic, and to varying degrees were antibacterial. Although more potent derivatives were not produced, this type of study is invaluable for understanding both the fermentation and biosynthetic processes, and for providing new compounds for additional structure-activity relationship studies. A number of interesting phytotoxins, some which are potential herbicides, are known to be produced by fungi, including tentoxin from. Alternaria alternata (Lax and Shepherd, 1988), various compounds from plant pathogenic fungi (Sakamura et al., 1988), cyclopenin and cyclopenol, 3,7-dimethyl-8-hydroxy-6-methoxyisochroman, and cinerain (Cutler et al., 1988), for example. The interest-

(m),

(m),

62

CEDRIC PEARCE 0

0

OH

II

OH

1

0 (192) papyrachonB Rj=OH, R+ (193) papyrachonC Rj=H. R@H

/

o

(194)Come*slin R=CH3 (195) Hydroxymmexislin R = C H P H

STRUCTURES 186-195.

ingly named cornexistin (194) was reported from Paecilomyces variotii Bainier (Nakajima et d., 1991). This compound had activity against grass and broadleaf species. Further work on the producing strain led to the isolation of a second metabolite, 14-hydroxycornexistin (195) (Fields et al., 1996),which was discovered in the fermentation broth of P variotii as a very active metabolite. Subsequent testing against a variety of broadleaf and grass weeds showed that cornexistin and hydroxycornexistin display similar herbicidal activity, with the latter compound being particularly good against broadleaf weeds. In a second report, a Chrysosporium sp. that had been isolated from a soil sample collected in Wisconsin was found to produce dihydrobisdechlorogeoa new compound, and the known bisdechlorogeodin (197) din

(m),

63

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

H

&

W

o

\

H

3

G

-

w0

\ COOCH3

OH

OH

(196) Dihydro-bisdechlorogeodin

0 COOCH3

(197) (-)-bisdechlorogeodin

STRUCTURES 196,197.

(Tanaka et al., 1996). These compounds were initially isolated because fermentations of this fungus in a medium containing starch, glycerol, soybean meal, yeast, potassium chloride, calcium carbonate, magnesium sulfate, and the phosphate trapping-agent allophane exhibited antibacterial activity. The antibacterial activity was shown to be due to bisdechlorogeodin, and this was also shown to be antagonized by the addition of casamino acids, or L-alanine, L-aspartic acid, or L-glutamic acid to the bioassay. This suggests that bisdechlorogeodin is an antimetabolite. In addition, both dihydrobisdechlorogeodin and bisdechlorogeodin showed herbicidal activity, although at what seems like high concentrations. It was further proposed that dihydrobisdechlorogeodin is transformed into bisdechlorogeodin intracellularly. A variety of approaches have been used to discover insecticidal fungal metabolites, including both random and more directed searches. As is pointed out in the fungicidal section, it is possible that fungi engage in chemical defense by producing compounds that exert an effect on either a predator or competitor. There is evidence that fungi may produce compounds that act as insect antifeedants, and this area has been reviewed extensively (Gloer, 199513). Briefly, it has been discovered that certain fungal survival structures, such as sclerotia, which may be exposed to insect predation, often contain compounds that are antiinsectans. For example, it has been observed that the beetle Carpophilus hemipterus avoids sclerotia of Aspergillus flavus while eating other parts of the same fungus. When A. flavus was grown on corn, analysis of the sclerotia led to the discovery of a series of novel indole Using a C. hemipterus diterpenoid compounds, called aflavines (198). feeding assay, it was shown that the most common aflavine completely inhibited feeding at concentrations in the diet far below those found in nature. The compounds were not present in other parts of the fungus,

64

CEDRIC PEARCE

H (199) Aflavarin

(198) Aflavine R=H m OH WH3

STRUCTURES 198, 199.

nor were they expressed in any liquid medium tested. This very interesting observation again focuses on the method used for cultivation as being critical for full metabolic expression from cultures being investigated. During the A. flavus study, Gloer’s group discovered a number of other novel compounds including aflavarin (199)and p-aflatrem (200). A wider search for metabolites from the sclerotia of fungi related to A. flavus led to the discovery of novel bioactive compounds from Aspergillus nomius, Aspergillus leporis, Aspergillus tubingensis, Aspergillus niger, Aspergillus carbonarius, Aspergillus sulphureus, Aspergillus rnelleus, and Aspergillus alliaceus (Gloer 1995b). Further studies by Gloer and his co-workers into the metabolites of other fungi led to investigation of antiinsectan compounds from the sclerotia and sclerotioid ascostromata of Penicillium and Eupenicillium. A number of novel metabolites were discovered including the sherinines 201-203 and isopentenylpaxillin 204 from Eupenicillium shearii (Belofsky et ul., 1995). These compounds were potent antiinsectans against H. zeu and C. hepipterus. The ascostromata of Eupenicillium crustaceurn were shown to contain 0.3% by weight of the aflavines previously discovered from Aspergillus tubingensis. This is a very interesting observation because both fungi, although quite different taxonomically, do inhabit similar areas and seem to have evolved or otherwise acquired similar biosynthetic routes. Other Eupenicillia were shown to contain quite different metabolites, emphasizing the distinct differences between the fungi and highlighting the value of testing various sources when looking for novel bioactive compounds.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

6

65

CEDRIC PEARCE

66

0

I (205)Destruxin A4

(206)Destruxin A5 STRUCTURES 205, 206.

From the white fly entomopathogen Aschersonia sp., two new dehave been isolated (Krasnoff et al., struxins, A4 (205) and A5 (m), 1996). In these new destruxins, methylisoleucine replaces the more common methylvaline residue found in other destruxins. Destruxin A4 and A5 showed insecticidal activity against Drosophila melanogaster, with LC50 values of 4 1 and 52 ppm, respectively. A completely different approach to that used by Gloer's group has been used by Morino et al. (1995),who discovered the insecticidal agent NK3 74200 (207). These workers screened for metabolites containing purine or pyrimidine. A Talaromyces sp. that had been isolated from a soil sample was shown to produce NK374200, and subsequent to the isolation and purification of this compound, by testing in various assays, it was shown to have antimosquito larval activity. This compound was shown to have low cytotoxicity against HeLa cells and in mice. No other activity was reported. XII. Summary and Conclusions

There is clearly a lot to be learned from the study of fungal metabolites, and clearly there is a lot we do not know about why and how these metabolites accumulate. Throughout the development of organic chemistry, there has been an interest in natural products, and fungal metabolites have received considerable attention. With the introduction of

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

67

(207)NK374200 STRUCTURE 207.

novel approaches to detecting biologically active compounds, there has been somewhat of a renaissance in the field, and, coupled with relatively nontraditional approaches to isolating fungi, new metabolites are being discovered. The importance of the new screening techniques is highlighted by the discovery of compactin and mevinolin and related compounds, and by the discovery of the zaragozic acid/squalestatin group. Both these families of metabolites were found independently by a number of groups and were subsequently shown to be quite common products hom not rare fungi. Recent discoveries of metabolites with a variety of activities are included in this chapter, but it is obvious that there are common themes. For example, there is a preponderance of antifungal metabolites, an interesting observation in light of Gloer’s ideas about antagonism between fungi in nature, and there have been a significant number of metabolites reported that affect cholesterol metabolism. Again this is perhaps of some significance in the metabolism of ergosterol and related compounds in the fungi. Both of these observations, however, could also be a reflection upon the fact that they comprise a number of complex enzymatically controlled events and as such provide a large number of different opportunities for inhibition. This is perhaps the reason why antiviral fungal metabolites are less common, because there are fewer significant unique biochemical events to be inhibited. Commonly, when a novel metabolite is detected, it is one of a family of metabolites produced by the culture being studied. This is apparent in the discoveries discussed herein, as well as in previous reports. This illustrates another powerful feature of drug discovery from fungal (and

68

CEDRIC PEARCE

other microorganism) sources. Analysis of the bioactivity associated with each metabolite allows for structure-activity relationships to be formulated, if in a preliminary state. The future for this area is promising. There is huge potential for novel chemistry from the untapped fungi yet to be isolated. The basic assumption is that biodiversity translates into chemidiversity, which, although true at the nucleic acid level, is not proven to be so in a practical sense. However, time will tell. One clear area for advances is in the methods used for culturing fungi in ways that engender production of these unusual metabolites so often associated with biological activity. Other issues to be dealt with include rapid characterization of products and swift yield improvement to obtain material for biological testing. In these trimmed down times of “lean and mean,” the challenges are there, but so are the sources of bioactive compounds. REFERENCES

Ainsworth, A. M., Chicarelli-Robinson, M. I., Copp, B. R., Fauth, U., Hylands, P. J., Holloway, J. A., Latif, M., O’Beirne, G. B., Porter, N., Renno, D. V., Richards, M., and Robinson, N. (1995). Xenovulene A, a novel GABA-benzodiazapine receptor binding compound produced by Acremonium strictum. J. Antibiot. 48, 568-573. Alberts, A. W., Chen, J., Kuron, G., Hunt, V., Huff, J., Hoffman, C., Rothcock, J., Lopez, M., Joshua, H., Harris, E., Patchett, A., Monaghan, R., Currie, S., Stapley, E., AlbersSchonberg, G., Hensens, O., Hirshfield, J., Hoogsteen, K., Liesch, J., and Springer, J. (1980). Mevinolin: A highly potent competitive inhibitor of hydroxymethylglutarycoenzyme A reductase and a cholesterol lowering agent. Proc. Natl. Acad. Sci. U.S.A. 77, 3957-3961.

Anderson, M. G., Jarman, T. B., and Rickards, R. W. (1995). Structures and absolute configurations of antibiotics of the oligosporon group hom the nematode-trapping fungus Arthrobotrys oligospora. J. Antibiot. 48, 391-398. Anke, T. (1995). The antifungal strobilurins and their possible ecological role. Can. J. Bot. 73,S940-S945.

Anke, T., Werle, A., Zapf, S.,Velten, R., and Steglich, W. (1995). Favolon, a new antifungal triterpenoid from a Favolaschia species. J. Antibiot. 48, 725-726. Anonymous (1995). First launch for Roche’s CellCept. Scrip 2036,19 (23 June). Aragozzini. F., Manachini, P. L., Craveri, R., Rindone, B., and Scolastico, C. (1972). Isolation and structure determination of a new antifungal a-hydroxymethyl-a-amino acid. Tetrahedron 28, 5493-5498. Arai, N., Shiomi, K., Tomoda, H., Tabata, N., Yang, D. J., Masuma, R., Kawakubo, T., and Omura, S. (1995). Isochromophilones 111-VI, inhibitors of acyl-CoA:cholesterol acyltransferase produced by Penicillium multicolor FO-3216. J. Antibiot. 48, 696-702. Ayer, W., and Miao, S. (1993). Secondary metabolites of the Aspen fungus Stacbybotrys cylindrospora. Can. J. Chem. 71,4 8 7 4 9 3 . Baxter, A., Fitzgerald, B. J., Hutson, J. L., McCarthy, A. D., Motteram, J. M., Ross, B. C., Sapra, M., Snowden, M. A., Watson, N. S., Williams, R. J., and Wright, C. (1992). Squalestatin 1, a potent inhibitor of squalene synthase, which lowers serum cholesterol in vivo. J. Biol. Chem. 276,11705-11708.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

69

Belofsky, G. N., Gloer, J. B., Wicklow, D. T., and Dowd, P. F. (1995). Antiinsectan alkaloids: Shearinines A-C and a new paxilline derivative from the ascostromata of Eupenicillium shearii. Tetrahedron 51, 3959-3968. Bergstrom, J. D., Kurtz, M. M., Rew, D. J., Amend, A. M., Karkas, J. D., Bostedao, R. G., Germenhausen, J. I., Bansal, V. S., Dufresne, C., VanMiddlesworth, F. L., Hensens, 0. D., Liesch, J. M., Zink, D. L., Wilson, K. E., Onishi, J., Milligen, J. A,, Bills, G., Bartizal, K. F., Rozdilsky, W., Abruzzo, G. K., Kaplan, L., Nallin, M., Jenkins, R. G., Huang, L., Meinz, M. S., Quinn, L., Burg, R. W., Kong, Y. L., Mochales, S., Mojena, M., Martin, I., Palaez, F., Diez, M., and Alberts, A. W. (1993). The zaragozic acids: A family of fungal metabolites that are picomolar competitive inhibitors of squalene synthase. Proc. Natl. Acad. Sci. U.S.A. 90,80-84. Bergstrom, J. D., Dufresne, C., Bills, G., Nallin-Omstead, M., and Byme, K. (1995). Discovery, biosynthesis, and mechanism of action of the zaragozic acids: Potent inhibitors of squalene synthase. Annu. Rev. Microbiol. 49,607-639. Bills, G., Pelaez, F., Polishook, J. D., Diez-Matas, M. T., Harris, G., Clapp, W. H., Dufresne, C., Byme, K. M., Nallin-Omstead, M., Jenkins, R. G., Mojena, M., Huang, L., and Bergstrom, J. D. (1994). Distribution of zaragozic acids (squalestatins) among filamentous ascomycetes. Mycol. Res. 98,733-739. Boyd, R. K., McAleess, A. J., Taylor, A., and Walker, J. A. (1991). Isolation of new isocyanide metabolites of Trichoderma hamatum as their (n-5-pentamthylcyclopentadieny1)- or (n-5-ethyltetramethylcyclopentadienyl)-di-~-thiocyana-to-rhodium complexes. J. Chem. Soc., Perkin Trans. 1, 1461-1465. Brandage, S., Josephson, S., Mahlen, A., Morch, L., and Vallen, S. (1976). (-)-Decylcitric acid and (+)-isocitric acid as metabolites from Penicillium spiculisporum-a correction. Acta. Chem. Scan. B30, 177-187. Brewer, D., Gabe, E. J., Hanson, A. E., and Taylor, A. (1979). Isonitrile acids from cultures of the fungus Trichoderma hamatum. J. Chem. SOC.,Chem. Commun., pp. 1061-1062. Brown, A. J., Smale, T. C., King, T. J., Hasenkamp, R., and Thompson, R. H. (1976). Crystal and molecular structure of compactin, a new antifungal metabolite from Penicilhm brevicompactum. J. Chem. Soc., Perkin Trans. 1, 1165-1170. Bushell, M. E. (1989). The process physiology of secondary metabolite production. In “Microbial Products: New Approaches” (S. Baumberg, I. S. Hunter, and P. M. &odes, eds.). Cambridge Univ. Press, Cambridge. Buss, A. D., and Waigh, R. D. (1995). Natural products as leads for new pharmaceuticals. In “Burger’s Medicinal Chemistry and Drug Discovery,” 5th ed., Vol. 1: “Principles and Practice” (M. E. Wolff, ed.), pp. 983-1034. Wiley, New York. Byme, K. M., Arison, B. H., Nallin-Omstead, M., and Kaplan, L. (1993). Biosynthesis of zaragozic acid. J. Org. Chem. 58,1019-1024. Chen, T.-B., and Weimer, D. D. (1991). Corymbotins A-I: Highly oxidized kolovane derivatives from Casearia corymbosa. J. Nat. Prod. 54,1612-1618. Chen, W. S., Lazar, C. S., Poenie, M., Tsien, R. Y., Gill, G. N., Rosenfeld, M. G. (1987). Requirement for intrinsic protein tyrosine kinase in the immediate and late actions of the EGF receptor. Nature 328, 820-823. Chexal, K. K., Tamm, Ch., Clardy, J., and Hirotsu, K. (1979). Gilmaniellin and dechlorogilmaniellin, two novel dimeric oxaphenalenones. Helv. Chim.Acta 62,1785-1803. Chu. M., Truumees, I., Gunnarsson, I., Bishop, W. R., Kreutner, W., Horan, A., Patel, M. G., Gullo, V. P., and Puar, M. S . (1993). A novel class of platelet activating factor antagonists from a Phoma sp. J. Antibiot. 46,554-563. Chu, M., Tmumees, I., Patel, M., Blood, C., Das, P. R., and Puar, M. S. (1995a). Sch 50673 and Sch 50676, two novel antitumor fungal metabolites. J. Antibiot. 48,329-331.

70

CEDRIC PEARCE

Chu, M., Truumees, I., Rothofsky, M. L., Patel, M., Gentile, f., Das, P. R., Puar, M. S., and Lin, S. L. (1995b). Inhibition of c-fos proto-oncogene induction by Sch 52900 and Sch 52901, novel diketopiperazines produced by Gliocladium sp. J. Antibiot. 48, 1440-1445. Clough, J. M. (1993). The strobilurins, oudemansins, and myxothiazols, fungicidal derivatives of b-methoxyacrylic acid. Nat. Prod. Rep. 10,565-574. Clutterbuck, P. W., Raistrick, H., and Rintoul, M.L. (1931). Studies in the biochemistry of micro-organisms, Part XVI: On the production from Glucose by Penicillium spiculisporum Lehman of a new polybasic fatty acid, C17H2806 (the lactone of y hydroxyPGdicarboxypentadecoic acid). Tmns. Roy. SOC.B220,301-330. Cole, R. J., and Cox, R. H. (1981). “Handbook of Toxic Fungal Metabolites.” Academic Press, New York. Connors, N., Prevomak, R., Brix, T., Seeley, A., Gbewonyo, R., Greasham, R., and Salmon, P. (1995). Effects of medium sterilization on the production of zaragozic acid by the fungus Leptodontidium elatius. J. Ind. Microbiol. 15,503-508. Coval, S. J., Puar, M. S., Phife, D. W., Terracciano, J. S., and Patel, M. (1995). SCH57404, an antifungal agent possessing the rare sodaricin skeleton and a tricyclic sugar moiety. J. Antibiot. 48,1171-1172. Cuellar, M. J., Giner, R. M., Recio, M. C., Just, M. J., Manez, S., and Rios, J. L. (1996). Two fungal lanostane derivatives as phospholipase A2 inhibitors. J. Nut. Prod. 59,977979. Cui, C.-B., Kakeya, H., Okada, G., Onose, R., Ubukata, M., Takahashi, I., Isono, K., and Osada, H. (1995). Tryprostatins A and B, novel mammalian cell cycle inhibitors produced by Aspergillus flaws. J. Antibiot. 48,1382-1384. Current, W. L., Tang. J., Boylan, C., Watson, P., Zeckner, D., Turner, W., Rodriguez, M., Dixon, C., Ma, D., and Radding, J. A. (1995). Glucan biosynthesis as a target for antifungals: The echinocandin class of antifungal agents. In “Antifungal Agents: Discovery and Mode of Action” (K. G. Dixon, L. G. Copping, and D. W. Holloman, eds.), pp. 143-160. BioScientific Publishers, Oxford, UK. Cutler, H. G., Ammerman, E., and Springer, J. P. (1988). Diverse but specific biological activities of four natural products from three fungi. In “Biologically Active Natural Products: Potential Use in Agriculture” (H. G. Cutler, ed.), pp. 79-90. American Chemical Society Symposium Series 380. American Chemical Society, Washington, DC. Dawson, M. J., Farthing, J. E., Marshall, P. S., Middleton, R. F., O’Neill, M. J., Shuttleworth, A., Stylli, C., Tait, R. M., Taylor, P. M., Wildman, H. G., Buss, A. D., Langley, D., and Hayes, M. V. (1992). The squalestatins, novel inhibitors of squalene synthase produced by a species of Phoma, I: Taxonomy, fermentation, isolation, physicochemical properties and biological activity. J. Antibiot. 45,639-647. Debono, M. (1994). The echinocandins: Fungicides targeted to the fungal cell wall. Expert Opin. Invest. Drugs 3, 65-82. Debono, M. (1995). Echinocandin lipopeptide antifungal agents: New agents and recent chemical modification studies. Expert Opin. Ther. Pat. 5, 771-786. Demain, A. L. (1992). Regulation of secondary metabolism. I n “Biotechnology of Filamentous Fungi: Technology and Products” (D. B. Finklestein and C. Ball, eds.). Butterworth-Heinemann, Boston. Dornberger, K., Ihn, W., Ritzau, M., Grafe, U., Schlegel, B., and Fleck, W. F. (1995). Chrysospermins, new peptaibol antibiotics from Apiocrea chrysospermin AplOl. J. Antibiot. 48,977-989.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

71

Dreyfus, M., Harri, E., Hofmann, H., Kobel, H., Pache, W., and Tscherter, H. (1976). Cyclosporine A and C, new metabolites from Trichoderma polysporum (Link et Pers.) Rifai. Eur. J. Appl. Microbial. 3, 125-133. Dufresne, C., Turner-Jones, E. T., Nallin-Omstead, M., Bergstrom, J. D., and Wilson, K. E. (1996). Novel zaragozic acids from Leptodontidiurn elatius. J. Nat. Prod. 59, 52-54, Elix, J. A., Jayanthi, V. K., and Leznoff, C. C. (1981). 2,4-di-O-methylgyrophoric acid and 2,4,5-tri-O-methylhiascic acid. New tridepsides from Parmelia damaziana. Aust. J. Chem. 34,1757-1761. Endo, A. (1979). Monacolin K, a new hypocholesterolemic agent produced by a Monascus species. J. Antibiot. 32, 852-854. Endo, A., Kuroda, M., and Tsujita, Y. (1976). ML-236A, ML-236B and ML-236C, new inhibitors of cholesterogenesis produced by Penicillium citrinum. J. Antibiot. 29, 1346-1 348. Engler, M., Anke, T., Klostermeyer, D., and Steglich, W. (1995). Hydroxystrobilurin A, a new antifungal E-P-methoxyacrylate from a Pterula species. J. Antibiot. 48, 884-885. Fehr, T., and Acklin, W. (1966). Die Isolierung zweier neuartiger Indol-derivate aus dem Mycel von Claviceps paspali Stevens et Hall. Helv. Chim. Acfa 49, 1907-1919. Fields, S. C., Mireles-Lo, L., and Gerwick, B. C. (1996). Hydroxycornexistin: A new phytotoxin from Paecilomyces variotii. J. Nut. Prod. 59, 698-700. Fleming, A. (1946). “History and Development of Penicillin,” pp. 1-23. Butterworth, London. Fredenhagen, A,, Hug, P., Sauter, H., and Peter, H. H. (1995a). Paeciloquinones A, B, C, D and F: New potent inhibitors of protein kinase produced by Paecilomyces carneus, 11: Characterization and structure determination. J. Antibiot. 48, 199-204. Fredenhagen, A., Mett, H., Meyer, T., Buchdunger, E., Regenass, U., Roggo, B., and Petersen, F. (1995h). Protein tyrosine kinase and protein kinase c inhibition by fungal anthraquinones related to emodin. J. Antibiot. 48, 1355-1358. Fujita, T., Inoue, K., Yamamoto, S., Ikumoto, T., Sasaki, S., Toyama, R., Chiba, K., Hoshino, Y., and Okumoto, T. (1994). Fungal metabolites, Part 11: A potent immunosuppressive activity found in Isaria sinclarii metabolite. J. Antibiot. 47, 208-215. Fujita, T., Hamamichi, N., Kiuchi, M., Masatoshi, K., Matsuzaki, T., Kitao, Y., Inoue, K., Hirose, R., Yoneta, M., Sasaki, S., and Chiha, K. (1996). Determination of absolute configuration and biological activity of new immunosuppressants, mycestericins D, E, F and G. I. Anfibiot.49, 846-853. Fujiwara, M., Okuda, T., Masuda, S., Shiomi, Y., Miyamoto, C., Sekine, Y., Takoe, M., and Fujiwara, M. (1982). Fermentation, isolation and characterization of isonitrile antibiotics. Agric. B i d . Chem. 46, 1803-1809. Gallo, L. L., Clark, S. B., Meyers, S., and Vahouny, G. V. (1984). Cholesterol absorption in rat intestine: Role of cholesterol esterase and acyl coenzyme A:cholesterol acyltransferase. J. Lipid Res. 25, 604-612. Gatenbeck, S., and Mahlen, A. (1968). A metabolic variation in Penicilliurn spiculisporum Lehman, I: Production of (+) and (-) decylcitric acids. Acta. Chem. Scan. 22, 26132616. Gehlert, D. R. (1994). Subtypes of receptors for neuropeptide Y Implication for targeting of therapeutics. Life Sci. 55, 551-562. Gibbs J. B. (1991). RasC-terminal processing enzymes: New drug targets? Cell 65, 1-4. Gloer, J. B. (1995a). The chemistry of fungal antagonism and defence. Can. J. Bot. 73, S1265-Sl274. Gloer, J. B. (1995b). Antiinsectan natural products from fungal sclerotia. Accounts Chem. Res. 28, 343-350.

72

CEDRIC PEARCE

Gloer, J. B. (1996). Applications of fungal ecology in the search for new bioactive natural products. The Mycota. In press. Goodman Gilman, A., Goodman, L. S., Rall, T. W., and Murad, F., eds. (1985). “Goodman and Gilman’s the Pharmacological Basis of Therapeutics,” 7th ed., pp. 1298-1299. Macmillan, New York. Goulard, C., Hlimi, S., Rebuffat, S., and Bodo, B. (1995). Trichorzins HA and MA, antibiotic peptides from Trichoderma harzianum, I: Fermentation, isolation and biological properties. J. Antibiot. 48,1248-1253. Grafe, U.,Ihn, W., Ritzau, M., Schade, W., Stengel, C., Schlegel, B., Fleck, W. F., Kunkel, W., Hartl, A., and Gutsche, W. (1995).Helioferins; Novel antifungal lipopeptides from Mycogone rosea:Screening, isolation, structures and biological properties. J. Antibiot. 48,126-133.

Grundemar, L., and Hakanson, R. (1994). Neuropeptide Y effector systems: Perspectives for drug development. Trends Pharm. Sci. 15, 155-159. Harris, G. H., Hoogsteen, K., Silverman, K. C., Raghoobar, S. L., Bills, G. F., Lingham, R. B., Smith, J. L., Dougherty, H. W., Cascales, C., and Pelaez, F. (1993). Isolation and structure determination of pycnidone, a novel bistropolone stromelysin inhibitor from a Phoma sp. Tetrahedron 49,2139-2144. Hashimoto, R., Takahashi, K., Hamano, K., Mori, T., and Nakagawa, A. (1994). New method for screening the melanin biosynthesis inhibitor by using the larval haemolymph of the silkworm, Bombyx mori, and the inhibitory activity of trichoviridin. Biosci. Biotech. Biochem. 58,1725-1 726. Hashimoto, R., Takehashi, S., Hamano, K., and Nakagawa, A. (1995). A new melanin biosynthesis inhibitor, melanoxadin from fungal metabolite by using the larval haemolymph of the silkworm, Bombyx mori. J. Antibiot. 48,1052-1054. Hasumi, K.,Tachikawa, K., Sakai, K., Murakawa, S., Yoshikawa, N., Kumazawa, S., and Endo, A. (1993). Competitive inhibition of squalene synthetase by squalestatin, 1. J. Antibiot. 46,689-691. Hawksworth, D. L. (1991).The fungal dimension of biodiversity: Magnitude, significance, and conservation. Mycol. Res. 95, 641-655. Hayashi, M., Kim, Y.-P., Hiraoka, H., Natori, M., Takamatsu, S., Kawakubo, T., Masuma, R., Komiyama, K., and Omura, S. (1995). Macrosphelide, a novel inhibitor of cell-cell adhesion molecules, I: Taxonomy, fermentation, isolation and biological activities. I. Antibiot. 48,1435-1439. Hlimi, S.,Rebuffat, S., Goulard, C., Duchamp, S., and Bodo, B. (1995). Trichorzins HA and MA, antibiotic peptides from Trichoderma harzianum, 11: Sequence determination. 1.Antibiot. 48,1254-1261. Hochlowski, J. E., Whittern, D. N., Buko, A., Alder, L., and McAlpine, J. (1995). Fusacandins A and B; Novel antifungal antibiotics of the papuolacandin class from Fusarium sambucinum, 11: Isolation and structural elucidation. J. Antibiot. 48,614618.

Huang, X.-H., Tomoda, H., Nishida, H., Masuma, R., and Omura, S. (1995a). Terpendoles, novel ACAT inhibitors produced by Albophoma yamanashiensis, I: Production, isolation and biological properties. J. Antibiot. 48,1-4. Huang, X.-H., Nishida, H., Tomoda, H., Tabata, N., Shiomi, K., Yang, D.-J., Takayanagi, H., and Omura, S. (1995b). Terpendoles, novel ACAT inhibitors produced by Albophoma yamanashiensis, 11: Structure elucidation of terpendoles A, B, C and D. J. Antibiot. 48,5-11.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

73

Ihara, M., Fukuroda, T., Saeki, T., Nishikibe, M., Kojiri, K., Suda, H., and Yano, M. (1991). An endothelin receptor (ETa) antagonist isolated from Streptomyces misakiensis. Biochem. Biophys. Res. Commun. 178,132-137. Ishimaru, T., Tsuboya, S., and Saijo, T. (1994). Tetracyclic compounds, their manufacture with Penicillium or Metarrizium, and inflammation inhibitors containing the tetracyclic compounds. Jpn. Pat. 0640995A. Jackson, M., Frost, D. J., Karwowski, J. P., Humphrey, P. E., Dahod, S. K., Choi, W. S., Brandt, K., Malmberg, L.-H., Rasmussen, R. R., Scherr, M. H., Flamm, R. K., Kadam, S., and McAlpine, J. B. (1995). Fusacandins A and B: Novel antifungal antibiotics of the papuolacandin class from Fusarium sambucinum, I: Identity of the producing organism, fermentation and biological activity. J. Antibiot. 48,608-613. Jayasuriya, H., Koonchanok, N. M., Geahlen, R. L., McLaughlin, J. L., and Chang, C.-J. (1992). Emodin, a protein tyrosine kinase inhibitor from Polygonum cuspidatum. J. Not. Prod. 55, 696-698. Jennings, D. H. (1995). “The Physiology of Fungal Nutrition.” Cambridge Univ. F’ress, Cambridge. Jeong, T. S., Kim, S. U., Kwon, B. M., Son, K. H., Kim, Y. K., Choi, M. U., and Bok, S . H. (1994). GEN-BP001, a new inhibitor of acyl-CoA:cholesterol acyltransferase produced by Aspergillus fumigatus F37. Tetrahedron Lett. 35, 33563570. Jeong, T.-S., Kim, S.-U., Son, K.-H., Kwon, B.-M., Kim, Y.-K., Choi, M.-U., and Bok, S.-H. (1995). GERI-BP001 compounds, new Inhibitors of acy1CoA:cholesterol acyltransferase from Aspergillusfumigatus F37, I: Production, isolation, and physico-chemical and biological properties. J. Antibiot. 48,751-756. Johnson, M. I., and Hoth, D. F. (1993). Present status and future prospects for HIV therapies. Science 260, 1286-1293. Kagamizono, T., Nishino, E., Matsumoto, K., Kawashima, A,, Kishimoto, M., Sakai, N., He, B.-M., Chen, Z.-X., Adachi, T., Morimoto, S., and Hanada, K. (1995). Bassiatin, a new platelet aggregation inhibitor produced by Beouveria bassiana K-717. J. Antibiot. 48,1407-1412. Kaise, H., Shinohara, M., Miyazaki, W., Izawa, T., Nakano, M., Sugawara, M., and Sugiura, K. (1979). Structure of K-76, a complement inhibitor produced by Stachybotrys complementi nov. Sp. K-76. J. Chem. SOC., Chem. Commun., pp. 726-727. Kakeya, H., Takahashi, I., Okada, G., Isono, K., and Osada, H. (1995). Epolactaene, a novel neuritogenic compound in human neuroblastoma cells, produced by a marine fungus. J. Antibiot. 48,733-735. Katagiri, K., Sato, K., Hayakawa, S., Matsushima, T., and Minato, H. (1970). Verticillin A, a new antibiotic from Verticillium sp. J. Antibiot. 23, 420-422. Kim, Y. K., Tomoda, H., Nishida, H., Sunazuka, T., Obata, R., and Omura, S. (1994). Pryripropenes, novel inhibitors of acyl-CoA:cholesterol acyltransferase produced by Aspergillus fumigatus, 11: Structure elucidation of pyripropenes A, B, C, and D. J. Antibiot. 47,154-162. Kluepfel, D.,Bagli, J., Baker, H., Charest, M.-P., Kudelski, A., Segal, S. N., and Vezina, C. (1972). Myriocin, a new antifungal antibiotic from Myriococcum albomyces. J. Antibid. 25, 109-115. Kobayashi, H., Sunaga, R., Furihata, K., Morisaki, N., and Iwasaki, S. (1995). Isolation and structure of an antifungal antibiotic, fusarielin A, and related compounds produced by a Fusarium sp. J. Antibiot. 48,42-52.

74

CEDRIC PEARCE

Kodukula, K., Arcuri, M., Qian-Cutrone, J., Hugill, R. M., Lowe, S. E., Pirnik, D. M., Shu, Y.-Z., Fernandes, P. B., and Seethala, R. (1995). BMS-192548, a tetracyclic binding inhibitor of neuropeptide Y Receptors, from Aspergillus niger WB2346, I: Taxonomy, fermentation, isolation and biological activity. J. Antibiot. 48, 1055-1059. Krasnoff, S. B., Gibson, D. M., Belofsky, G. N., Gloer, K. B., and Gloer, J. B. (1996). New destruxins from the entomopathogenic fungus Aschersonia sp. J. Nat. Prod. 59, 485-489.

Kulanthaivel, P. (1995). Naturally occurring protein kinase C inhibitors. In “Natural Products: Rapid Utilization of Sources for Drug Discovery and Development” (Nancy Mulford, ed.). Biomedical Library Series, International Business Communications, Southborough, MA. Kulanthaivel, P., Hallock, Y. F. Boros, C., Hamilton, S. M., Janzen, W. P., Ballas, M., Loomis, C. R., Jiang, J. B., Katz, B., Steiner, J. R., Clardy, J. (1993). Balanol: A novel and potent inhibitor of protein kinase C from the fungus Verticillium balanoides. J. Am. Chem. Soc. 115, 6452-6453. Kuo, M.-S., Yurek, D. A., Mizsak, S. A., Marshall, V. P., Liggett, W. F., Cialdella, J. I., Laborde, A. L., Shelly, J. A., and h e s d e l l , S. E. (1995). Production, isolation, and identification of dihydroepiepoformin as an IL-1 receptor antagonistic component in Penicilliurn patulum. J. Antibiot. 48, 888-890. Lax, A. R., and Shepherd, H. S. (1988). Tentoxin: A cyclic tetrapeptide having potential herbicidal usage. In “Biologically Active Natural Products: Potential Use in Agriculture’’ (H. G. Cutler, ed.), pp. 24-34. American Chemical Society Symposium Series 380. American Chemical Society, Washington, DC. Lechevalier, H. A. (1975). Production of the same antibiotics by members of different genera of microorganisms. In “Advances in Applied Microbiology” (D. Perlman, ed.), Vol. 19, pp. 2 5 4 5 . Academic Press, New York. Lee, C. H., Koshino, H., Chung, M. C., Lee, H. J., and Kho, Y. H. (1995). MR304A, a new melanin synthesis inhibitor produced by Trichoderma harzianum. J. Antibiot. 48, 1168-1 170.

Ling, K. H., Yang, C.-K., and Peng, F.-T. (1979). Territrems, tremorgenic mycotoxins of Aspergillus terreus. Appl. Environ. Microbiol. 37, 35 5-35 7. Ling, K. H., Liou, H.-H., Yang, C.-M., and Yang, C.-K. (1984).Isolation, chemical structure, acute toxicity, and some physicochemical properties of territrem C from Aspergillus terreus. Appl. Environ. Microbiol. 47, 98-100. Mandala, S. M., Thornton, R. A,, Frommer, B. R., Curotto, J. E., Rozdilsky, W., Kurtz, M. B., Giacobbe, R. A., Bills, G. F., Cabello, M. A., Martin, I., Pelaez, F., and Harris, G. H. (1995). The discovery of australifungin, a novel inhibitor of sphinganine N-acyltransferase from Sporomiella australis. Producing organism, fermentation, isolation, and biological activity. J. Antibiot. 48, 349-356. Marshall, V. P., McWethy, S. J., Sirotti, J. M., and Cialdella, J. I. (1990).The effect of neutral resins on the fermentation production of rubradirin. J. Ind. Microbiol. 5, 283-287. Martin, J. F., Revilla, G., Zanca, D. M., and Lopez-Nieto, M. J. (1982). Carbon catabolite regulation of penicillin and cephalosporin biosynthesis. In “Trends in Antibiotic Research: Genetics, Biosynthesis, Actions and New Substances” (H. Umezawa, A. L. Demain, T. Hata, and C. R. Hutchinson, eds.). Japan Antibiotics Research Association, Tokyo. Masuma, R., Tabata, N., Tomoda, H., Haneda, K., Iwai, Y., and Omura, S. (1994). Arohynapenes, new anticoccidial agents produced by Penicilliurn sp. J. Antibiot. 47,46-53.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

75

Matsumoto, K., Nagashima, K., Kamigauchi, T., Kawamura, Y., Yasuda, Y., Ishi, K., Uotani, N., Sato, T., Nakai, H., Terui, Y., Kikuchi, J., Ikenisi, Y., Yoshida, T., Kato, T., and Itazaki, H. (1995). Salfredins, new aldose reductase inhibitors produced by Crucibulurn sp. RF-3817, 1: Fermentation, isolation and structures of salfredins. J. Antibiot. 48,439446. Matsuzaki, K., Ikeda, H., Masuma, R., Tanaka, H., and Omura, S. (1995a). Isochromophilones I and 11, novel inhibitors against gp120-CD4 binding produced by Penicillium multicolor FO-2338, I: Screening, taxonomy, fermentation, isolation and biological activity. J. Antibiot. 48,703-707. Matsuzaki, K., Tanaka, H., and Omura, S. (1995b). Isochromophilones I and 11, novel inhibitors against gp120-CD4 binding produced by Penicillium multicolor FO-2338, 11: Structure elucidation. J. Antibiot. 48,708-713. Mayerl, F., Gao, Q., Huang, S., Klohr, S. E., Matson, J. A., Gustavson, D.,R., Pirnik, D. M., Berry. R. L., Fairchild, C., and Rose, W. C. (1993). Eupenifeldin, a novel cytotoxic bistropolone from Eupenicillium brefeldianum. J. Antibiot. 46,1082-1088. Mayerl, F., Huang, X., and Gao, Q. (1994). Fermentation manufacture of l,H-cycloundeca[1,2-P:5,6-P’]bis-cyclohepta(b)pyranderivatives as antitumor agents. U S . Pat. 5284866A. Middletown, R. F., Foster, G., Cannell, R. J. P., Sidebottom, P. J., Taylor, N. L., Noble, D., Todd, M., Dawson, M. J., and Lawrence, G. C. (1995). Novel squalestatins produced by biotransformation. J. Antibiot. 48,311-316. Miller, R. F., and Huang, S. (1995). Isolation and structure of sorrentanone: A new tetrasubstituted quinone from Penicillium chrysogenum. J. Antibiot. 48,520-521. Miyata, S.,Ohhata, N., Murai, H., Masui, Y., Ezaki, M., Takase, S., Nishikawa, M., Kiyota, S., Okuhara, M., and Kohsaka, M. (1992). WSOO9A and B, new endothelin receptor antagonists isolated from Streptomyces sp. no. 89009. J. Antibiot. 45,1029-1040. Mocek, U., Schultz, L., Buchan, T., Baek, C., Fretto, L., Nzerem, J., Sehl, L., and Sinha, U. (1996). Isolation and structure elucidation of five new asterriquinones from Aspergillus, Humicola and Botryotrichum species. J. Antibiot. 49,854-859. Morino, T., Nishimoto, M., Masuda, A., Fujita, S., Nishikiori, T., and Saito, S. (1995). NK374200, a novel insecticidal agent from Talaromyces, found by physico-chemical screening. J. Antibiot. 48,1509-1510. Nakajima, M.,Itoi, K., Takamatsu, Y., Sato, S., Furukawa, Y., Furuya, K., Honma, T., Kadotani, J., Kosaza, M., and Haneishi, T. (1991). Cornexistin: A new fungal metabolite with herbicidal activity. J. Antibiot. 44,1065-1072. Nakamura, M., Ito, Y., Ogawa, K., Michisuji, Y., Sato, S.-I., Takada, M., Hayashi, M., Yaginuma, S., and Yamamoto, S. (1995). Stachybocins, novel endothelin receptor antagonists, produced by Stachybotrys sp. M6222, I: Taxonomy, fermentation, isolation and characterization. J. Antibiot. 48,1389-1395. Nishizuka, Y. (1988). The molecular heterogeneity of protein kinase C and its implication for cellular regulation. Nature 334,661-665. Nokoe, S.,Furukawa, J., Sanakawa, U., and Shibata, S. (1976). Isolation, structure and synthesis of hirsutene, a precursor hydrocarbaon of coriolin biosynthesis. Tetrahedron Lett. 17,195-198. Nozawa, K., Nakajima, S., Kawai, K., Udagawa, S., Horie, Y., and Yamazaki, M. (1987). Novel indoloditerpenes, emindoles, and related compounds from Ernericelia spp. Chem. Nut. Prod., 29th Symp., Sapporo, Japan, pp. 637-643. Nozawa, O., Okazaki, T., Sakai, N., Komurasaki, T., Hanada, K., Morimoto, S., Chen, Z.-X., He, B.-M., and Mizoue, K. (1995). A novel bioactive S-Lactone FD-211: Taxonomy, isolation and characterization. J. Antibiot. 48,113-118.

76

CEDRIC PEARCE

Ogawa, T., Ando, K., Aotani, Y., Shinoda, K., Tanaka, T., Tsukuda, E., Yoshida, M., and Matsuda, Y. (1995a). RES-1214-1 and -2, novel non-peptide endothelin type a receptor antagonists produced by Pestalotiopsis sp. J. Antibiot. 48,1401-1406. Ogawa, K., Nakamura, M., Hayashi, M., Yaginuma, S., Yamamoto, S., Furihata, K., ShinYa, K., and Seto, H. (1995b). Stachybocins, novel endothelin receptor antagonists, produced by Stachybotys sp. M6222, II: Structure determination of stachybocins A, B and C. J. Antibiot. 48,1396-1400. Omura, S., Tanaka, K., Ikeda, H., and Masuma, R. (1993a). Isochromophilones I and 11, novel inhibitors against gp120-CD4 binding from Penicillium sp. J. Antibiot. 46, 1908-1 911. Omura, S., Tomoda, H., Kim, Y. K., and Nishida, H. (1993b). Pryripropenes, highly potent inhibitors of acyl-CoA:cholesterol acyltransferase produced by Aspergillusfumiga tus. 1. Antibiot. 46,1168-1169. Omura, S., Kuno, F., Otoguro, K., Sunazuka, T., Shiomi, K., Masuma, R., and Iwai, Y. (1995). Arisugarcin, a novel and selective inhibitor of acetylcholinesterase from Penicillium sp. FO-4259. J. Antibiot. 48,745-746. Pairet, L., Wrigley, S. K., Chetland, I., Reynolds, E. E., Hayes, M. A., Holloway, J., Ainsworth, A. M., Katzer, W., Cheng, X.-M., Hupe, D. J., Charlton, P., and Doherty, A. M. (1995). Azaphilones with endothelin receptor binding activity produced by Penicillium sclerotiorum: Taxonomy, fermentation, isolation, structure elucidation and biological activity. J. Antibiot. 48,913-923. Pearce, C. J. (1995). Discovering novel bioactive compounds from fungi. In “Natural Products: Rapid Utilization of Sources for Drug Discovery and Development” (Nancy Mulford, ed.), pp. 1.72-1.94. Biomedical Library Series, International Business Communications, Southborough, MA. Pearce, C. J., Carter, G. T., Neitsche, J., Borders, D. B., Greenstein, M., and Maiese, W. M. (1991). The effect of methylation inhibitors on citreamicin biosynthesis in Micromonospora citrea. 1.Antibiot. 44,1247-1251. Pearce, C. J., West, R. R., and Carter, G. T. (1995). The effect of sinefungin on the biosynthesis of ganefromycin. Structures of ganefromycins delta 1-4. Tetrahedron Lett. 36,1809-1812. Petersen, F., Fredenhagen, A., Mett, H., Lydon, N. B., Delmendo, R., Jenny, H.-B., and Peter, H. H. (1995). Paeciloquinones A, B, C, D, E and F: New potent inhibitors of protein tyrosine kinases produced by Paecilomyces carneus, I: Taxonomy, fermentation, isolation and biological activity. 1.Antibiot. 48,191-198. Pulici, M., Sugawara, F., Koshino, H., Uzawa, J., Yoshida, S., Lobkovsky, E., and Clardy, J. (1996). A new isodrimeninol from Pestalotiopsis sp. J. Nat. Prod. 59,47-48. Qian-Cutrone, J., Huang, S., Chang, L.-P., Pirnik, D. M., Klohr, S. E., Dalterio, R. A., Hugill, R., Lowe, S., Alam, M., and Kadow, K. (1996a). Harziphilone and fleephilone, two new HIV REV/RRE binding inhibitors produced by Tnchoderma har- zianum. J. Antibiot. 49,990-997. Qian-Cutrone, J., Huang, S., Trimble, J., Li, H., Lin, P.-F., Alam, M., Klohr, S. E., and Kadow, K. (1996b). Niruriside, a new HIV/RRE binding inhibitor from Phyllanthus niruri. J. Nat. Prod. 59,196-199. Roach, P. L., Clifton, I. J., Fulop, V., Harlos, K., Barton, G. J., Hajdu, J., Anderson, I., Schofield, C. J., and Baldwin, J. E. (1995). Crystal structure of isopenicillin synthase is the first from a new structural family of enzymes. Nature 375, 700-704. Saito, Y., Nakao, K., Mukoyama, M., and Imur, H. (1989). Increased plasma endothelin level in patients with essential hypertension. N. Engl. I. Med. 322,205.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

77

Sakai, K., Watanabe, K., Masuda, K., Tsuji, M., Hasumi, K., and Endo, A. (1995). Isolation, characterization and biological activities of novel triprenylphenols as pancreatic cholesterol esterase inhibitors produced by Stachybotrys sp. F-1839.1. Antibiot. 48, 447-456. Sakamura, S.,Ichihara, A., and Yoshihara, T. (1988). Toxins of phytopathogenic microorganisms: Structural diversity and physiological activity. In “Biologically Active Natural Products: Potential Use in Agriculture” (H. G. Cutler, ed.), pp. 57-64. American Chemical Society Symposium Series 380. American Chemical Society, Washington, DC. Sakemi, S., Inagaki, T., Kaneda, K., Hirai, H., Iwata, E., Sakakibara, T., Yamauchi, Y., Norcia, M., Wondrack, L. M., Sutcliffe, J. A., and Kojima, N. (1995). CJ-12,371 and c J - 1 ~72, 3 two novel DNA gyrase inhibitors. Fermentation, isolation, structural elucidation and biological activities. J. Antibiot. 48,134-142. Sakuda, S.,Otsubo, Y., and Yamada, Y. (1995). Structure of patulodin, a new azaphilone epoxide, produced by Penicillium urticae. J. Antibiot. 48,85-86. Sanglier, J. J., Baumann, G., Dreyfus, M., Fehr, T., Traber, R., and Scrierber, M. H. (1990). Immunosuppressants of microbial origin. In “Microbial Metabolites,” Developments in Industrial Microbiology Series, Vol. 32, pp. 1-28. Wm. C. Brown Publishers, Dubuque, IA. Sarges, R. (1989). Aldose reductase inhibitors: Structure-activity relationship and therapeutic potential. Adv. Drug Res. 18,139-175. Sasaki, S., Hashimoto, R., Kiuchi, M., Inoue, K., Ikumoto, T., Hirose, R., Chiba, K., Hoshino, Y., Okumoto, T., and Fujita, T. (1994). Fungal metabolites, Part 14: Novel potent immunosuppressants, mycestericins, produced by Mycelia Sterilia. J. Antibiot. 47,420-433. Schlingmann, G., West, R. R., Milne, L., Pearce, C. J., and Carter, G. T. (1993). Diepoxins, novel fungal metabolites with antibiotic activity. Tetrahedron Lett. 34,7225-7228. Schlingmann, G.,Milne, L., Pearce, C. J., Borders, D. B., Greenstein, M., Maiese, W. M., and Carter, G. T. (1995). Isoaltion, characterization and structure of a new allenic polyene antibiotic produced by fungus LG07F275.J. Antibiot. 48,375-379. Shinohara, C., Hasumi, K., Hatsumi, W., and Endo, A. (1996). Staplabin, a novel fungal triprenylphenol which stimulates the binding of plasminogen to fibrin and U937 cells. J. Antibiot. 49,961-966. Shiomi, K., Uchida, R., Inokoshi, J., Tanaka, H., Iwai, Y., and Omura, S. (1996). Adrastins A-C, new protein farnesyltransferase inhibitors, produced by Penicillium sp. FO3929. Tetrahedron Lett. 37, 1265-1268. Shiozawa, H., Takahashi, M., Takatsu, T., Kinoshita, T., Tanazawa, K., Hosoya, T., Furuya, K., Takahashi, S., Furihata, K., and Seto, H. (1995). Trachyspic acid, a new metabolite produced by Talaromyces trachyspennus, that inhibits tumor cell heparinase: Taxonomy of the producing strain, fermentation, isolation, structural elucidation and biological activity. J. Antibiot. 48,357-362. Shu, Y.-Z., Qian-Cutrone, J., Klohr, S. E., and Huang, S. (1995). BMS-192548,a tetracyclic binding inhibitor of neuropeptide Y receptors, from Aspergillus niger WB2346, 11: Physico-chemical properties and structural characterization. 1. Antibiot. 48,10601065.

Sidebottom, P. J., Highcock, R. M., Lane, S. J., Procopiou, P. A,, and Watson, N. S. (1992). The squalestatins, novel inhibitors of squalene synthase produced by a species of Phoma, 11: Structural elucidation. J. Antibiot. 45,648-658.

78

CEDRIC PEARCE

Singh, S. B., Jones, E. T., Goetz, M. A., Bills, G. F., Nallin-Omstead, M., Jenkins, R. G., Lingham, R. B., Silverman, K. C., and Gibbs, J. B. (1994a). Fusidienol: A novel inhibitor of ras farnesyl-protein transferase from Fusidium griseum. Tetrahedron Lett. 35,46934696. Singh, S. B., Katz, B. A., Lingham, R. B., Martin, I., and Silverman, K. C. (1994b). Inhibitors of farnesyl-protein transferase. U S . Pat. 5436263-A-950725. Sliskovic, D. R., and White, A. D. (1991). Therapeutic potential of ACAT inhibitors as lipid lowering and antiatherosclerotic agents. Trends Pharm. Sci. 12, 194-199. Stadler, M., Anke, H., Arendholz, W. R., Hansske, F., Anders, U., Sterner, O., and Berquist, K. E. (1993a). Lachnumon and lachnumol A, new metabolites with nematocidal and antimicrobial activities from the ascomycete Lachnum papyraceum (Karst.) Karst, I: Producing organism, fermentation, isolation and biological activities. J. Antibiot. 46, 961-967. Stadler, M., Anke, H., Sterner, O., and Berquist, K. E. (1993b). Lachnumon and lachnumol a, new metabolites with nematocidal and antimicrobial activities from the ascomycete Lachnum papyraceum (Karst.) Karst, 11: Structural elucidation. J. Antibiot. 46, 968-972. Stadler, M., Sterner, O., and Anke, H. (1993~).New biologically active compounds from the nematode-trapping fungus Arthrobotrys oligospora Fresen. 2.Naturforsch. C48, 843-850. Stadler, M., Anke, H., and Sterner, 0. (1995a). Metabolites with nematocidal and antimicrobial activities from the ascomycete Lachnum papyraceum (Karst.) Karst, ILI: Production of novel isocoumarin derivatives, isolation, and biological activities. J. Antibiot. 48,261-266. Stadler, M., Anke, H., and Sterner, 0. (1995b). New metabolites with nematocidal and antimicrobial activities from the ascomycete Lachnum papyraceum (Karst.) Karst, IV Structure determination of novel isocoumarin derivatives. J. Antibiot. 48,267-270. Stadler, M., Anke, H., and Sterner, 0. (1995~). Metabolites with nematocidal and antimicrobial activities from the ascomycete Lachnum papyraceum (Karst.) Karst, V: Production, isolation and biological activities of bromine-containing mycorrhizin and lachnumon derivatives and four additional new bioactive metabolites. J. Antibiot. 48, 149-1 53. Stadler, M., Anke, H., Shan, R., and Sterner, 0. (1995d). New metabolites with nematocidal and antimicrobial activities from the ascomycete Lochnum papyraceum (Karst.) Karst, VI:Structure determination of non-halogenated metabolites structurally related to mycorrhizin. J. Antibiot. 48,154-157. Stadler, M., Anke, H., and Sterner, 0. (1995e). New metabolites with nematocidal and antimicrobial activities from the ascomycete Lachnum papyraceum (Karst.) Karst, VII: Structure determination of brominated lachnumon and mycorrhizin A derivatives. 1.Antibiot. 48,158-161. Stierle, A.,Strobel, G., and Stierle, D. (1993).Tax01 and taxane production by Taxomyces andreanae, an endophytic fungus of Pacific Yew. Science 260, 214-216. Sugano, M., Sato, A., Iijima, Y., Oshima, T., Furuya, K., Kuwano, H., and Hanazawa, H. (1991). Phomactin A: A novel PAF antagonists from a marine fungus Phoma sp. J. Am. Chem. SOC.113, 5463-5464. Sugano, M., Sato, A., Iijima, Y., Furuya, K., Kuwano, H., and Hata, T. (1995). Phomactins E, F, and G: New phomactin-group PAF antagonists from a marine fungus Phoma sp. Antibiot. 48,1188-1190.

r.

BIOLOGICALLY ACTIVE FUNGAL METABOLITES

79

Summers, W. K., Majovski, L. V., Marsh, G. M., Tachiki, K., and Kling, A. (1986). Oral tetrahydroaminoacridine in long-term treatment of senile dementia, Alzheimer type. N . Engl. J. Med. 315,1241-1245. Tabata, N.. Tomoda, H., Iwai, Y., and Omura, S. (1995a). Arohynapene D, a new anticoccidial agent produced by Penicillium sp. FO-2295. I. Antibiot. 48,83-84. Tabata, N., Tomoda, H., Masuma, R., Iwai, Y., and Omura, S. (1995b). Fudecalone, a new anticoccidial agent produced by Penicillium sp. FO-2030. J. Antibiot. 48,53-58. Takeshima, H.(1992). Antiviral agents. In “The Search for Bioactive Compounds from Microorganisms” (S. Omura, ed.). Springer-Verlag, New York. Tanaka, Y. (1992). Fermentation processes in screening for new bioactive substances. In “The Search for Bioactive Compounds from Microorganisms” (S. Omura, ed.). Springer-Verlag. New York. Tanaka, Y., Matsuzaki, K., Zhong, C.-L., Yoshida, H., Kawkubo, T., Masuma, R., Tanaka, H., and Omura, S. (1996). Dechlorogeodin and its new dihydro derivatives, fungal metabolites with herbicidal activity. J. Antibiot. 49,1056-1059. TePaske, M. R. (1991). Isolation and structure determination of antiinsectan metabolites from the sclerotia of aspergillus species. Ph.D. Thesis, Department of Chemistry, University of Iowa. Tkacz, J. S. (1992). Glucan biosynthesis in fungi and its inhibition. In “Emerging Targets in Antibacterial and Antifungal Chemotherapy” (J. A. Sutcliffe and N. H. Georgopapadakou. eds.), pp. 495-523. Chapman and Hall, New York. Tomoda, H., Kim, Y. K., Nishida, H., Masuma, R., and Omura, S. (1994). Pyripropenes, novel inhibitors of acyl-CoA:cholesterol acyltransferase produced by Aspergillus fumigatus, I: Production, isolation and biological properties. J. Antibiot. 47,148-153. Tomoda, H.,Tabata, N., Yang, D.-J., Takayanagi, H., Nishida, H., and Omura, S. (1995a). Pyripropenes, novel ACAT inhibitors produced by Aspergillus fumigatus, 111: Structure elucidation of pyripyropenes E to L. J. Antibiot. 48,495-503. Tomoda, H., Ito, M., Tabata, N., Masuma, R., Yamaguchi, Y., and Omura, S. (1995b). Amidepsines, inhibitors of diacylglycerol acyltransferase produced by Humicola sp. FO-2942, I: Production, isolation and biological properties. J. Antibiot. 48,937-941. Tomoda, H., Tabata, N., Ito, M., and Omura, S. (1995~).Amidepsines, inhibitors of diacylglycerol acyltransferase produced by Humicola sp. FO-2942,II: Structure elucidation of amidepsines A, B and C. J. Antibiot. 48,942-947. Tomoda, H., Tabata, N., Yang, D.J., Takayanagi, H., and Omura, S. (1995d). Terpendoles, novel ACAT inhibitors produced by Albophoma yamanashiensis, 111: Production, isolation and structure elucidation of new components. J. Antibiot. 48,793-804. Tomoda, H.,Yamaguchi, Y., Tabata, N., Kobayashi, N., Masuma, R., Tanaka, T., and Omura, S. (1996). Amidepsine E, an inhibitor of diacylglycerol acyltransferase produced by Humicola sp. FO-5969. I. Antibiot. 49,929-931. Turner, W. B. (1971). “Fungal Metabolites.” Academic Press, New York. Turner, W. B., and Aldridge, D. C. (1983). “Fungal Metabolites, 11.” Academic Press, London. Uchida, R., Shiomi, K., Inokoshi, J., Masuma, R., Kawakubo, T., Tanaka, H., Iwai, Y., and Omura, S. (1996a). Kuraosins A and B, new protein farnesyltransferase inhibitors produced by Paecilomyces sp. FO-3684, I: Producing strain, fermentation, isolation, and biological activities. J. Antibiot. 49,932-934. Uchida, R.,Shiomi, K., Sunazuka, T., Inokoshi, J., Nishizawa, A,, Hirose, T., Tanaka, H., Iwai, Y., and Omura, S. (1996b). Kuraosins A and B, new protein farnesyltransferase inhibitors produced by Paecilomyces sp. FO-3684,II: Structure elucidation and total synthesis. J. Antibiot. 49,932-934.

80

CEDRIC PEARCE

Vagelos, P. R. (1991). Are prescription drug prices high? Science 252, 1080-1084. Van Der Pyl, D., Inokoshi, J., Shiomi, K., Yang, H., Takeshima, H., and Omura, S. (1992). Inhibition of farnesyl-protein transferase by gliotoxon and acetylgliotoxin. J. Antibiot. 45,1802-1805. Van Der Pyl, D., Cans, P., Debernard, J. J., Herman, F., Lelievre, Y., Tahraoui, L., Vuilhorgne, M., and Leboul, J. (1995). RPR113228, a novel farnesyl-protein transferase inhibitor produced by Chrysosporium lobatum. J. Antibiot. 48, 736-737, Watanabe, T., Suzuki, N., Shimamoto, N., Fujino, M., and Imada, A. (1990). Endothelin in myocardial infarction. Nature 344, 114. Weber, H. A., and Gloer, J. B. (1991). Preussomerins A-F: Novel antifungal metabolites from the coprophilous fungus Preussia isomera Cain. J. Org. Chem. 56, 4355-4360. Weber, H. A., Baenziger, N. C., and Gloer, J. B. (1990). Structure of preussomerin A: An unusual new antifungal metabolite from the coprophilous fungus Preussia isomera. J. Am. Chem. SOC. 112, 6718-6719. Weinberg, E. D. (1996). Antifungal agents. In “Burger’s Medicinal Chemistry and Drug Discovery,” 5th ed., Vol. 2: “Therapeutic Agents” (M. E. Wolff, ed.), pp. 637-652. Wiley, New York. West, R. R., Van Ness, J., Varming, A.-M., Rassing, B., Biggs, S., Gasper, S., McKernan, P. A., and Piggott, J. (1996). ZG-l494a, a novel platelet-activating factor acetyltransferase inhibitor from Penicillium rubrum: Isolation, structure elucidation and biological activity. J. Antibiot. 49, 967-973. Whyte, A. C., Gloer, J. B., Scott, J. A., and Malloch, D. (1996). Cercophorins A-C: Novel antifungal and cytotoxic metabolites from the coprophilus fungus Cercophora areolata. J. Nat. Prod. 59, 765-769. Wood, K. A., Kau, D. A., Wrigley, S. K., Beneyto, R., Renno, D. V., Ainsworth, A. M., Penn, J., Hill, D., Killacky, J., and Depledge, P. (1996). Novel P-methoxyacrylates of the 9-methoxystrobilurin and oudemansin classes produced by the basidiomycetes Favolaschia pustulosa. J. Nat. Prod. 59, 646-649. Wu, J. C. (1994). Mycophenolate mofetil: Molecular mechanism of action. Perspect. Drug Discovery Des. 2,185-204. [This review also contains a brief history, with references, of the discovery of mycophenolic acid.] Zapf, S., Werle, A., Anke, T., Klostermeyer, D., Steffan, B., and Steglich, W. (1995a). 9-Methoxystrobilurine-Bindeglieder zwischen Strobilurinen and Oudemansinen. Angew. Chem. 107, 255-257. Zapf, S., Hobfeld, M., Anke, H., Velten, R., and Steglich, W. (1995b). Darlucins A and B, new isocyanide antibiotics from Sphaerellopsis filum (Darluca filum). J. Antibiot. 48, 36-41.

Zopf, W. (1888). Zur Kenntnis der Infektions-Krankheiten neiderer Tiere und Pflanzen. Nova Acta Leopold Carol 52, 314-376.

Old and New Synthetic Capacities of Baker’s Yeast P. D’ARRIGO,

G. PEDROCCHI-FANTONI AND S. SERVI

CNR Centro per lo Studio delle Sostanze Organiche Naturali Dipartimento di Chimica Politecnico di Milano 20231 Milano, Italy

I. Introduction 11. Reducing Capacities A. Enzymes from the Fatty Acid Synthetase Complex B. Carbonyl Groups in Heterocyclic Compounds C. Nitrogen-Reducible Functional Groups D. Sulfur-Containing Compounds 111. The Formation of G--C Bonds n! Oxidations: Getting the Other Enantiomer V. Hydrolytic Activities: Phosphate Esters VI. Lyase Activity VII. The Biogeneration of Aroma Compounds VIII. Conflicting Reports IX. Conclusions References

I. Introduction

It is well documented in the literature of the last decade that biocatalysts (microorganisms and isolated enzymes) have found widespread application as selective reagents for the solution of synthetic organic chemistry problems. Although the extraordinary specificity of enzymatic catalysts has been known for a long time, they have only recently been applied in the selective transformation of unnatural substrates, mainly in the production of enantiomerically pure chiral compounds. It is now widely recognized that biologically active compounds used as pharmaceuticals should not be applied as racemates (Servi, 1996). Together with the absolute need to develop environmentally acceptable reagents and processes, these reasons are responsible for the observed growth of biocatalysis. The nature of enzymatic catalysis and the significance of stereochemistry in biological activity was already present in the work of Emil Fischer. The introduction to his article “Bedeutung der Stereochemie fur die Physiologie” (Fischer, 1889) could very well find a place in a 81 ADVANCES IN APPLIED MICROBIOLOGY, VOLUME 44 Copyright B 1997 by Academic Press, Inc. All rights of reproduction in any form reserved. 0065-2164/97 $25.00

82

P. D’ARRIGO et (11.

modern Introduction to Biocatalysis. Due to the growing number of papers in all sectors of chemistry, the wide range of interdisciplinary journals in which works in biocatalysis are published, and the rapid growth of this discipline, accurate awareness of the actual progress in this specific scientific domain requires continuous attention and the use of current awareness electronic bulletins and databases. While these tools are of fantastic help in information retrieval, the knowledge and citations of the work accomplished during or before the sixties are rapidly declining for the simple reason that they are not yet included in even the most comprehensive databases. Yet, the study of the archaeology of biocatalysis, to paraphrase the title of similar observations on enzymology (Neidleman, 1990), will not only be useful for the relevant scientific information found there, but also for the stimulating (sometimes frustrating) opportunity to compare the scientific novelty and intellectual effort present in today’s rapid communication with that met in similar studies of 80 years ago. Review articles would also help to fill this gap. These considerations apply equally well to the particular kind of biocatalysis represented by Baker’s yeast biotransformations. Baker’s yeast has had special merits in introducing the use of biocatalysts in the practice of organic synthesis, and hundreds of articles have been devoted to research in which yeast is applied in the biotransformation of unnatural substrates. At the end of the nineteenth century, microorganisms in general and yeast in particular were used for their production of hydrolytic enzymes that were applied in the transformation of polysaccharides. The specific action of yeast on different substrates was fundamental in the identification of new enzymatic activities and in the determination of their specificity. In “Bedeutung der Stereochemie fiir die Physiologie,” Fischer (1889) affirms that “bei der zuvor erwiihnten Versuchen hatte ich Gelegenheit, die Enzyme der hefe genauer kennen zu lernen” [following those experiments I had the opportunity to know with more precision the enzymes from yeast]. The process of submitting new substrates to yeast activity, allowing identification of new enzymatic activities, is of great interest in that genetic manipulation affords the possibility of enhancing synthesis of the enzyme with production of useful new catalysts. Baker’s yeast’s incredible ability to respond to new substrates with new synthetic activities gives the opportunity to uncover novel exploitable synthetic capacities. Comprehensive review articles (Czsuk and Gliinzer, 1991; Servi, 1990) and practical procedures (Roberts, 1992) for Baker’s yeast biotransformations have been reported. In this chapter some of the most interesting biocatalytic capacities expressed by yeast on old and new substrates will be critically considered.

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

83

I I . Reducing Capacities The reducing capacities of yeast have attracted the interest of modern organic chemists in the preparation of chiral secondary alcohols (Ward and Young, 1990). The observation that the addition of powdered sulfur in fermenting yeast resulted in the formation of H2S (Dumas, 1874) was probably the first published evidence of the reducing capacity of yeast. However, following the observation that proteins containing sulfhydryl groups produce hydrogen sulfide in the presence of sulfur, Heffter (1908) denied that yeast enzymes were at all involved in the reduction of sulfur. The attention given to discriminating between actual products from enzyme catalysis and artifacts was particularly intense. The reduction of carbonyl groups was first shown with aldehydes and then on ketones, nitro groups, and so on. Most of this work was reviewed by Neuberg (1949). Subsequently (Levene and Walti, 1943; Macleod et al., 1964; Guette and Spassky, 1972; Deol et al., 1976), the use of Baker’s yeast in carbonyl reduction has been considered a means of obtaining chiral secondary alcohols in enantiomerically pure form. This aspect was not considered previously because of the difficulty involved in assessing the absolute configuration and enantiomeric purity of chiral compounds. The method of reducing carbonyl groups for preparation of secondary alcohols has since become an established procedure. The stereochemistry of the reduction can be predicted since it follows the Prelog rule (Prelog, 1964). Microbial reduction with microorganisms is still preferred to the use of isolated oxidoreductases, since the regeneration of cofactor is still an unsolved practical problem. However, even optimized procedures are often not compatible with the strict economical requirements that must be met in industrial practice (Schmidt et al., 1992). Industrial preparation of trimegestone ( via l ) the Baker’s yeast catalyzed reduction of the corresponding 1,Z-diketone (2) (Fig. 1)has been reported (Crocq et al., 1997; Buendia et al., 1993). Using glycerol as the energy source, 240 g of yeast was used per g of substrate at a concentration of about 2 g/liter. The reaction was best performed at 4O0C in aerobic conditions. Despite the large fermentation volume, 99% pure hydroxyketone was recovered in 75% yield. The enantiomeric excess of the product was higher than 99%. Although other yeasts from culture collections proved to yield higher productivity, Baker’s yeast was preferred because of its lower price. The fact that yeast biomass is applied on an industrial scale in enantioselective reduction adds significance to this biocatalyst. Further applications of Baker’s yeast in industrial biotransformations will be mentioned.

P. D’ARRIGO et al.

84

2

1 FIG.1.

A. ENZVMES FROM THE FATTY ACIDSYNTHETASE COMPLEX

Some of the most interesting enzymatic capacities found in yeast are expressed by the fatty acid synthetase complex responsible for in vivo synthesis of fatty acids (Walsh, 1979a). The dehydrogenases catalyzing the reduction of the 3-oxothioester groups and of the C=C double bond conjugated with the carbonyl group can be exploited synthetically on structurally analogous substrates. Both the 3-oxoester and the C=C double bond reduction are considered by organic chemists to be of special interest for the production of chiral hydroxyesters, valuable chiral synthons, and of building blocks chiral for the presence of a methyl-bearing carbon, respectively. At least two different oxidoreductases are found in yeast that are able to reduce 3-oxoesters (3) with opposite stereochemical preferences (Heidlas et a]., 1988) (Fig. 2). The production of enantiopure 3-hydroxyesters (4) is an important achievement due to the value of these and related compounds as chiral synthons (Sheldon, 1993). A number of structurally related 3-hydroxyesters (4) of (S)-absolute configuration (L-series) can be obtained with yeast and other microorganisms with different conditions and efficiency and high to very high enantiomeric excesses. Simple structural variations allow for alteration of the stereoselectivity of the reduction. When R = CH3 the enantiomer of opposite configuration is available either by depolymerization of natural polyhydroxybutyrate (Seebach et a]., 1982) or by oxidation of butanoic acid with Candida rugosa (Hasegawa et a]., 1981). In practice, both enantiomers of 3-hydroxybutyrate and valerate are available in enantiomerically pure form through biotransformation or fermentation. The preparation of a-alkyl-P-hydroxyesters in the form of enantiomerically pure single diastereomers represents a more challenging synthetic problem and is not easily

85

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

2

RM o R R

3

OR

4 FIG.2.

L-enzyme 1 L-enzyme 2

l6

anti (2S.3S)

FIG.3.

accomplished by bioreduction of the corresponding racemic ketone. Also in this case, the carbonyl group is usually reduced with high enantiospecificity irrespective of the configuration of the adjacent chiral center. Enzymes with different stereochemical preferences for 3-0x0esters and 2-alkyl-3-oxoestershave been isolated (Heidlas et a]., 1988; Shieh et al., 1985; Shieh and Sih, 1993; Sih et al., 1983; Nakamura et al., 1991). They are NADPH-dependent enzymes and are able to catalyze the reduction of oxoesters of different type. Figure 3 shows that the two substrates 5 and 6 are in rapid equilibrium through the enol form (2).The L-enzyme (I)binds preferentially to the S-ketoester (5),whereas the L-enzyme (2) preferentially binds the R-ketoester (6). In both cases, the hydride equivalent is delivered to the Re face of the carbonyl group to give the respective products (Shieh and Sih, 1993). Enantiomerically pure anti and syn hydroxyesters can be obtained when the purified enzymes are used. In yeast the two enzymes have been estimated to be approximately 1:l. The kCat/Kmratio for the two

86

P. D'ARRIGO et a].

FIG.4.

FIG.5.

enzymes should account for the stereochemical outcome of the transformation. In the case where R, = ally1 and R2 = ethyl, the result is a 7 7 2 3 antisyn ratio. Creative variations of the use of yeast on 0-ketoesters have allowed kinetic resolution of racemic secondary alcohols and amines through Baker's yeast reduction of the corresponding acetoacetyl derivatives with low to medium enantiomeric excesses. The recognition of the sign of chirality on a carbon far removed from the actual center involved in biotransformation is remarkable (Hudlicky et a]., 1991, 1992) (Fig. 4). In a similar experiment, a-alkyl-P-ketoesters of a chiral alcohol were reduced but with a purified L-a-alkyl-P-ketoester reductase, and the resolution of the alcohol was complete, as shown in Fig. 5 (Kawai et al., 1995b). The following example refers to a remarkable case of diastereo- and enantioselectivity displayed on the same substrate, allowing establishment of two chiral centers in one operation (Fig. 6). The yields are very high (40 and 80% of maximum obtainable), as are the diastereoisomeric and enantiomeric excesses (Eh and Kalessem, 1995). The preparation of chiral compounds with a tertiary carbon atom chiral for the substitution pattern in Fig. 7 can be effected through the asymmetric reduction of a triply substituted C=C double bond.

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

87

40% yield, de>98%, ee=93%

FIG.6 . X

I

FIG.7.

This reaction, which has no general examples in asymmetric hydrogenation with chiral catalysts, can give very good results if the double bond is activated with strongly polarizing groups. The ability to reduce C=C double bonds has been attributed to the enoate reductases (EC 1.3.1.31) of the fatty acid synthetase (Cornforth, 1959; Dugan et al., 1970). Their activity has been found in several microorganisms (Holland, 1992), and their biocatalytic properties have been studied extensively in Clostridia (Thanos et al., 1987). While the reducing capacity of Baker’s yeast is quite accessible, microorganisms of the Clostridium type are anaerobic bacteria very sensitive to dioxygen and not as easy to grow (Kuno et al., 1985; Bader and Simon, 1980). However, they have extraordinary reducing capacities with very high productivity numbers (10-100 times higher than the corresponding reactions with yeasts), allowing reduction of large amounts of substrate with very little biomass. They have high reducing potentials, unique among microorganisms, allowing them to reduce carboxylic acids to primary alcohols, activity seldom reported in the literature (Fronza et al., 1995), or, as in this case, the C=C double bond of a$-unsaturated acids or esters. Moreover, in their use in biocatalysis with resting cells, the limited amount of cofactor present in the cells can be efficiently regenerated with hydrogen gas in the presence of small amounts of artificial mediators (viologen) or with an electrochemical regeneration system. The use of Clostridium tyrobutyricum DSM 1460 and of Clostridium kluyveri DSM 555 has been fully described (Thanos et al., 1987). These enzymatic systems have not become so popular among chemists due to the initiation required in manipulation of material and methods not familiar to organic chemists. The capacity of these systems are described in detail in the literature (Simon et al., 1985). The C=C reduc-

88

€? D’ARRIGO et a1

X = H , D, CH,, CI, Br, CF,

A = CHO. CH,OH, CH(OCH,),. COOR, NO,

FIG.8 .

ing capacity is assigned to an enzyme that has its best activity in a typical range of pH (5-8) and is usually employed at pH 6. It is NADHdependent, and a number of regenerating systems have been proposed. Ki for aliphatic enoates and their reduced products are rather high (500 mM), and they are therefore suitable for preparative applications. Stereochemistry of double bond reduction occurs with trans hydrogen delivery, and the chirality of the product obtained has been studied. The stereochemical preference is opposite for E and Z double bonds. Low enantiomeric excesses are obtained if the Z-E conversion is favored in the reaction conditions. Substrates accepted by the enzymatic systems are restricted to the ones having a small X, usually -CH3 in most examples, but also halogen or ethyl. Enoate reductase from yeast is a similar enzyme in many respects, but it is not able to reduce a$-unsaturated acids or esters, unless X is a halogen. The activating groups of the substrates are -CHO or NOz. Allylic alcohols are substrates since they are usually partially transformed with yeast cells to the corresponding aldehydes by the alcohol dehydrogenases present (Servi, 1990). Kinetic constants and operating parameters are not as well defined as for the parent enzyme from Clostridium that has been purified and used in the absence of other enzymes. All the examples reported in the literature for yeast enoate reductase make use of whole-cell biocatalysts. The range of substrates submitted to reduction is quite large, and the yields and enantiomeric excesses of products are remarkable. The two different types of triply substituted olefins in Fig. 8 can be the substrates for the reduction, and they give as products methyl alkanols (or corresponding products) of two different chirality senses.

89

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

/substrate/

[substrat4

AoMe W

C 0

OMe

H

T

0

Colt

O

i

+ O H .

?OH

C0,Et

b u g 1

8

O

H

9

voH A \

0

OH

OH

11

10

14

15

17

FIG.9.

The synthetic application of the compounds obtained in homochiral form using this method is found in the preparation of natural products. Compounds 8 , 9 (Leuenberger et al., 1979), and lo (Fuganti et al., 1988b) in Fig. 9 are equivalent bifunctional chiral synthons that have been employed in the synthesis of a-tocopherol (Fuganti and

P. D’ARRIGO et al.

90

5OHCHO

0

0

20 l9

\slow

-& fast

CHO

OH

0

22

21 FIG.10.

Grasselli, 1979, 1982). Compounds 11, 12, 13, and 14,deriving from olefins of type B, have the opposite configuration of the chiral methyl, Ohta et bearing carbon (Gramatica et al., 1987, 1988; Sat0 et ~ l .1988; al., 1989). Compounds 15 and show that the stereochemistry of the newly formed chiral center depends on the configuration of the double bond (Utaka et al., 1987). Compounds 17 and 18 are intermediates in the synthesis of zeaxanthin and of another precursor of a similar carotenoid. The products are prepared with yeast at a multi-kg scale (Leuenberger, 1985). The mechanism of the reduction of allylic alcohols has been studied in detail in the case outlined in Fig. 10. The aldehyde 19 is rapidly reduced to the allylic alcohol 20, and its level stays low and constant for the entire reaction time. The aldehyde present at equilibrium conditions is then transformed into the saturated aldehyde 21 and then into the final product 23. This behavior is confirmed by the fact that allylic alcohols are not reduced in the absence of an alcohol dehydrogenase, that is, with a purified enoate reductase. The reduction shown in Fig. 10 has been scaled to hundred of grams for the preparation of C5 chiral synthons (Fuganti et ~ l .1992). , B. CARBONYL GROUPSIN HETEROCYCLIC COMPOUNDS

Masking and unmasking carbonyl groups is a common technique in organic synthesis for the temporary protection of such a reactive func-

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

23

27

28

25

91

24

26

FIG.11.

tionality. One way of doing so is to build a heterocyclic ring across the functional group to be protected and subsequently release it under the mildest possible conditions. Among the possible heterocyclic rings employed to mask functionalities are isoxazoles, isoxazolines, thiazoles, furane derivatives, thiophenes, and dithianes. Acetyl pyridines, isoxazoles, isoxazolines, and thiazoles are effectively reduced by fermenting yeast to the corresponding secondary alcohols, usually with good control of enantiomeric purity (Fig. 11). Compounds 23 and 24 are obtained in good yield and high enantiomeric excess, although of unknown absolute configuration (Bianchi et al., 1984). From the acetyl isoxazolines, the two diastereoisomers 25 and S resulting from enantioselective carbonyl reduction are obtained. The enantiomeric excess is strongly enhanced by using 2-propanol as an additive (Ticozzi and Zanarotti, 1988).The thiazolyl ketone 22 is easily reduced to the carbinol of S-configuration. This compound is considered an equivalent of lactaldehyde (Dondoni et al., 1988).It has been reported that other isoxazoles fused in a six-membered ring are reduced in fermenting yeast with cleavage of the N-0 bond (Fig. 12). This kind of ring opening usually occurs with hydrogen gas in the presence of metal catalysis. Very functionalized molecules often do not withstand

P. D'ARRIGO et al.

92

X = CH,,

FIG.12.

FIG.13.

these conditions. However, this method seems deprived of practical , application since product yields are very low (Easton et ~ l .1994). Sulfur heterocycles can be easily transformed into saturated carbon chains via catalytic reduction. The reduction with Baker's yeast of a series of sulfur heterocycles acting as protecting groups is reported in Fig. 13. In reactions 1-4 the reduction takes advantage of the higher enantioselectivity usually displayed by cyclic substrates when com-

OLD AND NEW SYNTHETIC CAPACITIES OF BAKER'S YEAST

93

FIG.14.

pared to open-chain ones. In this way, a-substituted-P-hydroxyesters of high enantiomeric excesses are obtained (Hoffmann et al., 1981). In the case reported in reaction 5, the preparation of methyl alkanols via the thiophene intermediate was chosen because of the higher enantiomeric excess and efficiency observed in the production of the product. It was also shown that a family of 5-substituted compounds could be prepared by a general methodology and that all could be reduced with similar efficiency. In this way, a broad series of 2-methylalkanols were made accessible in enantiomerically pure form (Hogberg et al., 1992). C. NITROGEN-REDUCIBLE FUNCTIONAL GROUPS

The reduction of the nitro group to the corresponding amine is an important transformation, especially when homochiral amines become accessible. Neuberg studied the conversion of nitrobenzene to aniline in fermenting yeast (Neuberg and Welde, 1914b). The product was isolated in high yields. In order to elucidate the mechanism of such a reduction, and considering unlikely the direct transformation of the nitro group, a number of other possible intermediates were prepared and added to fermenting yeast. Figure 1 4 summarizes the positive results obtained. According to these findings (Neuberg and Welde, 1914a), it was deduced that nitrobenzene is initially reduced to nitrosobenzene and then to phenylhydroxylamine before this compound is further transformed into aniline. However, the yield of each individual step was lower than for the total transformation of nitrobenzene-aniline. This fact remains without an explanation.

94

P. D’ARRIGO et al.

FIG.15.

In a subsequent work (Neuberg and Reinfurth, 1923), m-dinitrobenzene was also submitted to fermenting yeast and the product of partial reduction was obtained (m-nitroaniline). The reduction of substituted m-dinitrobenzene has been reinvestigated and the regioselectivity of the reaction has been studied (Davey et al., 1994). The products obtained in regioselective reduction are the consequence of combined steric and electronic effects (Fig. 15). In the same work, o-nitrobenzonitrile (29) was reduced, but the corresponding aniline, expected from previous reports, was present in only minor amounts, the main reaction product being 2-aminobenzamide (30).This product might result from the initial reduction of the aniline to the hydroxylamine derivative 31, cyclized in turn to the benzoisoxazolidine 32 and then reduced again to the isolated product 30 (Fig. 16). This mechanism would reveal that the hydroxylamino compound is an intermediate in the reduction of nitro compounds, as previously inferred by Neuberg. A similar mechanism could also be invoked in the Baker’s yeast reduction of the nitroalkene bearing a cyan0 group in position 2 reported in Fig. 17, which leads to the formation of 5-amino isoxazole (34)(Navarro-Ocana et al., 1996). Since several grams of biomass per gram of substrate are usually employed for these biotransformations (24 in the last example), the possibility of a chemical reduction due to the presence of a stoichiometric amount of low-molecular-weight reducing compounds like carbohydrates or mercaptans should always be considered. In the latter case, for instance, it is known that unsaturated p-nitro nitriles are converted into amino-isoxazoles using mercaptans as reducing agents (Colau and Viel, 1980). In some cases the absence of enzymatic catalysis is evident (Baik et al., 1994) from the fact that operational conditions are employed (pH 14, 80°C) in which yeast does not retain any enzymatic activity. If nitro groups are presented in a different context, namely in the absence of an

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

29

95

30

!

I

31

32 FIG.16.

34

33 FIG.17.

aromatic ring, the group is not reduced. Thus, prochiral a$-unsaturated aliphatic nitro compounds are enantioselectively reduced to the chiral saturated compounds (Ohta et al., 1989), while y-nitro ketones are enantioselectively reduced to the corresponding (S)-4-nitroalcohols (Guarna et a].,1995). Other nitrogen-containing compounds have been transformed by other yeasts, including oximes and imines. It has been shown for instance that Torula yeasts are able to transform inorganic nitrates and nitrites into oximes and to further incorporate them into amino acids or their amides (Virtanen and Csaky, 1948). The reduction of aliphatic oximes has been reported. A mixture of E- and Z-stereoisomers of

96

P. D’AEWGO et 01.

L35

PhC ‘OOH

36

37 FIG.18.

38

39

R,, R, = H, Ph, Cyclohexyl R, = NHR, OH

FIG.19.

butanone oxime has been reduced to the corresponding amine of 58% ee. A similar result was observed by submitting to reduction the oxime ester 35 as a precursor of the oxime (Fig. 18). The oxime of benzoyl formic acid (36)of E-configuration was also reduced, although with low enantioselectivity. The (I?)-phenylglycine37 obtained was of only 20% ee. However, this reaction is of great potential interest. The involvement of a transaminase acting on the hydrolyzed oxime was excluded by an evaluation of the ratio of the compounds obtained (Fig. 19). From the aliphatic substrate, a small percentage of the hydrolysis product (Z-butanone) was isolated (Gibbs and Barnes, 1990). The authors considered the possibility that a specific hydrolase could be present in yeast. In fact, on other substrates and in different experimental conditions it was found that this hydrolytic activity is the prevalent or exclusive one. Indeed, a preparative method for the transformation of hydrazones and oximes of type 38 into the corresponding aldehydes and ketones 39 using Baker’s yeast has been proposed (Kamal and Reddy, 1992). The carbonyl compounds are obtained in yields higher

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

97

than 80%. 1 g of the hydrazone or the oxime was treated with 4 g of yeast in 70 ml of toluene and 6 ml of water. The reaction was complete within 15 h. The substantial differences between the two studies could be attributed either to substrate modification or to the experimental conditions (presence of solvent in the second case). No speculation about the nature of the catalysis was made. The N-oxides of pyridine, isoquinoline, morpholine, and others are reduced to the corresponding amino compounds in 40 to 70% yields in 5-6 days. 500 g of yeast per g of substrate were used (Takeshita and Yoshida, 1990). D. SULFUR-CONTAINING CO~OUNDS

Sulfur-containing compounds can undergo very different transformations by Baker's yeast, as shown from the following examples. Neuberg and Nord (1914) showed that thioacetaldehyde (40)in situ generated from the stable precursor 41can be reduced to ethylmercaptan ( 4 2 ) .Care was taken in showing that the reaction is not just an effect of the reducing capacities of carbohydrates. Higher homologues of thioaldehydes generated in situ by adding ammonia and hydrogen sulfide to the aldehyde were effectively reduced also using a cell-free enzymatic juice (Fig. 20). With other sulfur-containing substrates nonhomogeneous results have been observed. While diethyl disulfide was reported to be reduced to ethyl mercaptan (reaction 2) (Neuberg, 1949), some thiols, including benzyl mercaptans fermented under vigorously oxygenating conditions, were reported to be transformed into disulfides by yeast (reaction 3) (Rama Rao, 1992). However, other authors found that, under normal anaerobic fermenting conditions, benzyl mercaptan (43) was transformed in low yields into enanantiomerically pure (S)-benzyl thioglycerate (44). No trace of the disulfide mentioned in the preceding work (reaction 4) (Fronza et al., 1992a) was formed. In this example, selectivity was achieved by concomitant modulation of fermenting conditions and substrate modification. In the two equilibria between thiol and disulfide, it is likely that the catalyst is a thiol oxidase, which normally is active in the transformation of peptides and proteins containing sulfur functional groups. In the formation of disulfides, the concomitant chemical oxidation is occurring. The thioglycerate obtained in reaction 4 is probably formed in a mechanism in which benzyl mercaptans act as a nucleophile on an activated intermediate (a thioester, for instance) in the normal glycolytic pathway active in yeast metabolism. In fact, experiments with labeled glucose, fructose, and

P. D’ARRIGO et al.

98

I

H

41

2

42

40

vSH

\/s\s/\

3

4

___c

Ph-SH

OH

43

0

44

45

46

FIG.20.

mannose as nutrients have shown that the three carbon atoms found in (S)-benzylthioglycerate derive from the saccharides (Fronza et al., 199213). Moreover, experiments in D 2 0 show deuterium incorporation at various positions arising from the equilibrium between the three hexoses mentioned before and between dihydroxy acetone-P and glyceraldehyde 3-P. In an alternative mechanism, the thioglycerate could be formed by a reaction between benzyl mercaptan and hydroxypyruvaldehyde medi-

OLD AND NJZW SYNTHETIC CAPACITIES OF BAKERS YEAST

99

R = alkyl, aryl FIG.21.

ated by glyoxalase I in analogy with the behavior of glutathione with the same aldehyde. It is questionable, however, whether hydroxypyruvaldehyde is an intermediate in the alcoholic fermentation. Surprisingly, when racemic 2-methyl-3-phenylpropanethiol(45) (reaction 5) was used as the substrate under the same conditions as in reaction 4 of Fig. 20, (S)-2-methyl+phenylpropanethiol hemisuccinate (46)was obtained following partial kinetic resolution (40% ee) (Fuganti et al., 1991). In this case, the action of the nucleophilic thiol on an activated succinoyl unit probably mediated by succinate-thiokinase was invoked. The reactions of these sulfur derivatives are not particularly useful for preparative purposes. However, the results point out some exploitable enzymatic activities present in Baker’s yeast. It is worth noting, though, that performing reaction 4 of Fig. 20 in D 2 0 yielded isopropylidene glycerol selectively deuterated to a different extent at various positions depending on the monosaccharide used as the carbon source (glucose, fructose, or mannose) (Fronza et al., 1994). In none of the biotransformations described above involving mercaptans has the formation of products deriving from sulfur oxidation been observed. Specific strains of S. cerevisiae are reported as efficient biocatalysts in the oxidation of thio-analogues of fatty acids (Buist et al., 1990), of methyl-styryl sulfide (Fauve et al., 1991), and methyl tolyl sulfide (Beecher et al., 1995) (Fig. 21). Methyl thioethers of different structures have been oxidized to sulfoxides of R absolute configuration with the chloroperoxidase from Caldariomyces fumago (Colonna et al., 1990) and with the cyclohexanone monooxygenase from Acinetobacter calcoaceticus (Secundo et al., 1993). The Baker’s yeast reduction of the C=S bond in thioxoesters and thioketones has been the subject of investigation. The thiocarbonyl compounds analogues of P-ketoesters and ketones have been reduced. The thiols from the reduction of the thioxo group are accompanied by the alcohol from the reduction of the ketone formed by spontaneous

100

R

P. D'ARRIGO et al.

AO R

- - - C R

OR

OR'

FIG.22.

hydrolysis of the thio analogue. Comparison of the enantiomeric excess of the alcohol and of the thiol shows that the reduction of the thioxo group is less enantioselective than that of its keto analogue (Fig. 22). The effect of various additives used as inhibitors of oxidoreductases has been investigated. Increases in enantiomeric excesses are observed as in the case of the reduction of carbonyl groups. The ratio between substrate and yeast has been optimized with reference to the ee of products (Nielsen and Madsen, 1994). The conversion of thioureas and thiocarbamates to urea and carbamates has also been reported (Kamal et d.,1990).

I l l . The Formation of C-C

Bonds

The formation of carbon-carbon bonds is a fundamental operation in organic synthesis. If chiral centers are formed during C-C bond formation, stereochemical control is only observed when nonracemic chiral catalysts or reagents are involved. A valid complement to the asymmetric synthetic approach is represented by the use of biocatalysts, particularly aldolases and related enzymes, which find applications in the synthesis of rare carbohydrates (Wong et al., 1992). Although several aldolases, transketolases acyl-coA synthases, and other activities responsible for the formationhreaking of C--C bonds are present in yeast (Walsh, 1979b), they cannot generally be exploited in biocatalysis for biotransformation of unnatural substrates due to their narrow substrate specificity and the prevalent utilization of enzymes and natural substrates in the yeast metabolic cycles. Genetic engineering can provide a solution to the problem: fructose diphosphate (FruA) aldolase has been overexpressed in Saccharomyces cerevisiae and the whole cell organism utilized in biocatalysis using phenylacetaldehyde (Fig. 23) as the unnatural substrate (Compagno et al., 1993). Interestingly, the cells provide the cosubstrate (dihydroxyacetone phosphate, DHAP) and the phosphatase required to hydrolyze the initially formed phosphate ester (47).Although the product 9 is obtained in only low yields due to the concomitant utilization of the substrates by yeast, this system appears to be of great potential value. It can be compared with

OLD AND NEW SYNTHETIC CAPACITIES OF BAKER’S YEAST

/

101

Phosphatase

I

0

48 FIG.23.

multienzymatic systems where the same result is obtained by assembling the required enzymatic activities in a one-pot procedure, as in an artificial metabolism (Fessner, 1992). Other C-C bond-forming activities can be directly exploited in Baker’s yeast and constitute unique examples in biocatalysis of the utilization of whole-cell biocatalysts for this purpose. Although the biosynthesis of cholesterol and related compounds has been known for a long time, the efficient cyclization of polyenes to sterols has not been reported. The involvement of a single catalytic step from the open to the cyclized form has been postulated in lanosterol formation, but experiments were effected with minute amounts of protein from mammalian liver tissue. Oxido-sterol cyclase has been found in yeast. Efficient transformations of oxidosqualene (49) to lanosterol is possible if the biomass is sonicated in order to allow substrate diffusion probably by removing the obstructing outer cell membrane. Native Baker’s yeast can affect the transformation only marginally. Attention was specially devoted to the possibility of cyclizing unnatural substrates because of the relevant synthetic implication that this would, for instance, allow for preparation of remotely functionalized sterols like 50 in Fig. 24, a potent irreversible inhibitor of the 14-demethylase enzyme, with an important role in the biosynthesis of cholesterol (Kyler and Novak, 1992).

P. D’ARRIGO et al.

102

R = Methyl, Allyl, Ethinyl

52

51

FIG.24.

In this case, an efficient biomimetic cyclization was found before the equivalent enzymatic method was available (Johnson et a]., 1987). A similar enzymatic capacity has been observed from other sources (Abe et al., 1993; Xiao and Prestwich, 1991). The cyclization reaction would be of great interest in the cyclization of farnesyl pyrophosphate and its congeners to the bicyclic terpenes. The fact that such enzymes are not available prompted Kyler and Novak to design, on the basis of speculations of a hypothetical enzyme active site working model, intermediates that could eventually be cyclized by the yeast oxidosterol cyclase. In this way compound 51 was found to be a substrate for the enzyme to give a bicyclic product 52 convertible by desulfurization to the target bicyclic structure (Kyler and Novak, 1992). The success of this approach was only diminished by the low yields of the observed transformation (8%).

Pyruvate decarboxylase (PDC)is a thiamine pyrophosphate (TPP)-dependent enzyme catalyzing the decarboxylation of pyruvate to acetaldehyde or the transfer of a Cz unit of pyruvate onto acetaldehyde with formation of acetoin. The mechanism of the reaction is believed to occur through the initial formation of an adduct between TPP and pyruvate (Fig. 25). This adduct, depending on the organism and the environ-

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

\ +

103

/

\ N

HO OH

0-

I" TPP

+

OH

PDC R = CH,, X=OH. Y=H, R' = Ph BFD R = Ph, X=H, Y=OH, R = CH,

FIG.25.

mental conditions, can decarboxylate to give acetaldehyde (path a) or interact with another molecule of aldehyde (path b) to finally produce an acetoin that can be transformed by reducing enzymes to 2,3-butanediol. Neuberg found that added benzaldehyde in fermenting yeast entered path b as a second substrate giving rise to enantiomerically pure (R)-phenylacetyl carbinol (PAC) (Neuberg et al., 1923). The large-scale production of this compound has raised some interest, since L-ephedrine can be obtained by reductive amination. At present, PAC and L-ephedrine are manufactured by this method at BASF in Germany on a multi-ton scale. This process probably represents the largest industrial application of Baker's yeast on nonconventional substrates. The reaction efficiency is reduced by the concomitant reduction of benzaldehyde to benzyl alcohol. Various techniques have been devised to avoid the undesired side reaction. These include the slow addition of aldehyde to the fermenting mixture (Agarwal et aZ., 1987) or the use of alkyl pyridines as additives for the inhibition of the ADH responsible for aldehyde reduction (Gupta et ~ l .1979). , Why benzaldehyde is the preferred cosubstrate (after acetaldehyde) in the reaction with pyruvate is not known. Other aldehydes, including substituted benzaldehydes (Long et d.,1989; Ohta et QZ., 1989),furylacrolein (Fuganti et ~ l .1988a), , and cinnamaldehydes (Fuganti and Grasselli, 1977), are good substrates in parallel reactions. Higher reaction times usually allow production of enantiomerically pure vicinal diols (Fig. 26). The ones obtained from

104

P. D’ARRIGO et 01.

__c

OH

R = H. CH,

Epc synthesis

1

FIG.26.

cinnamaldehydes have found extensive application in the synthesis of enantiomerically pure compounds (EPCs) (Fuganti and Grasselli, 1985).

Incorporation of higher a-oxoacids has been explored, allowing for production of a higher homologue of the phenylacetyl carbinols, albeit with lower efficiency if compared to pyruvate (Fuganti et al., 1988b). Purified yeast pyruvate decarboxylase has been used with several aldehydes and pyruvate derivatives as C2-unitsas donors. The corresponding acyloins are usually obtained with lower yields if compared with the whole-cell system (Cardillo et al., 1991; Crout et al., 1991). Other decarboxylases with potential synthetic applications are found in other organisms. Benzoylformate decarboxylase (BFD) found in Pseudomonas putida and Acinetobacter calcoaceticus (Ward et al., 1992) operates with a mechanism similar to the one proposed for PDC. Benzoylformate is the first substrate instead of pyruvate, while the second substrate is acetaldehyde. The acyloin obtained in this case, IS)-2-hydroxypropiophenone, is isomeric with PAC obtained with PDC and benzaldehyde (Fig. 25). PDC may have application for production of acyloins containing a fixed acyl group with a variable aromatic group, while BFD may produce acyloins having a variable aliphatic group with a fixed aromatic component. A further C-C bond-forming reaction has been observed whose catalytic nature is not understood. Starting with a$-unsaturated ketones and esters in the presence of trifluoroethanol, fermenting yeast produces optically active fluorinated carbinols accompanied by the corre-

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

53

105

O

R = alkyl, OR'

J

R = OR'

54 FIG.27.

sponding allylic alcohols when a ketone is the substrate (Fig. 27). When the same transformation is applied to a#-unsaturated esters, the corresponding lactones are formed (Kitazume and Ishikawa, 1984). Enzymes catalyzing Michael-type addition to unsaturated ketones and esters with a broad substrate specificity would be of tremendous potential in biocatalysis, since there are apparently no general catalytic capacities of this kind. In another report on Baker's yeast applications, starting from cyanoacetone in fermenting yeast, the two stereoisomeric alkylated prodin Fig. 28 were obtained in high ee, and the results were ucts 55 and regarded by the authors as a consequence of the novel C-C bond-forming reaction catalyzed by Baker's yeast (Itoh et al., 1989). It was later shown that on similar active methylene compounds like 57 alkylation occurs through spontaneous aldol condensation between the acidic compound and the acetaldehyde presumably generated in situ by Baker's yeast oxidation of ethanol (Fuganti et al., 1990). The a$-unsaturated compound formed via dehydration of the aldol can as a then he reduced by yeast, affording the alkylated compound mixture of enantiomers, reflecting the stereochemistry of the alkene and the selectivity of the yeast reducing system. A similar mechanism was then proposed for the formation of the products obtained from cyanoacetone.

s

106

P. D’ARRIGO et a]. OH -

LcN - ‘y b.y.

OH

ACN

+

EtOH

55

56

cNx COOEt

CN

b.y.

v 57

EtOH

CN

/+

v

58

FIG.28.

IV. Oxidations: Getting the Other Enantiomer

Baker’s yeast is widely used as a reducing agent in organic synthesis for the preparation of enantiomerically pure chiral compounds. One enantiomer can often be prepared in a convenient way, while the other is usually not accessible with a similar procedure. This constitutes a serious drawback of the method, since in many applications of biocatalytic methods the goal is often to secure both enantiomers of a chiral drug for biological activity evaluation. When it is possible to obtain both enantiomers with the same biocatalyst, the advantage is obvious. The resolution of secondary alcohols with hydrolytic enzymes, mainly lipases, usually allows reaching that goal. Various techniques can be used to direct the enantioselectivity of carbonyl reduction by Baker’s yeast (the same concept is applicable to other microorganisms). Altering the surrounding of the carbonyl group (e.g., changing the substrate) is a strategy that can be followed and can be exemplified by the case of the reduction of P-ketoesters. The only variation that can be practical in this case is to modify the dimensions of the ester group. This technique has been successful in increasing the enantiomeric excess of alcohol obtained with the (S)-configuration.The rationale behind this behavior is well known: enzymes with opposite stereochemical requirements are

OLD AND NEW SYNTHETIC CAPACITIES OF BAKER’S YEAST

107

yeast

60 FIG.29.

present and acting at the same time on similar substrates. Increasing the differences in the groups flanking the carbonyl group, the object of the reduction, increases the differences in kinetics of the interaction of the two enzymes with the substrate, thus increasing selectivity. It has been shown that P-ketoesters reductases have very different k, and vma, thus explaining the shift in the enantiomeric excesses of the product when different concentration of substrate were used (Chen et al., 1983). If reductases of different stereochemical preferences are available, it is probably simpler to employ different microorganisms in order to obtain the opposite stereochemistry in carbonyl reduction. This is possible for this class of compounds since Baker’s yeast and Geotricum candidurn produce P-hydroxyesters of opposite chirality (Azerad and Buisson, 1992). In other cases, the use of selective inhibitors of enzymes with enantiomeric catalytic properties has been applied successfully. Figure 29 shows the application of this concept to the selective production of (Forni et the enantiomeric trifluoromethyl hydroxyketones 59 and

Ql.,

1994).

A further possibility for synthesizing the other enantiomer is to use the biocatalyst in the oxidation direction. Since enzymes catalyze reactions in both directions, it is conceivable that, while during the reduction of a prochiral carbonyl group 100% of the homochiral secondary alcohol can be obtained in theory, in a process that can be considered an asymmetric synthesis, the same enzymatic system, in catalyzing the oxidation in the reverse reaction, will act on the secondary alcohol with the same selectivity that gave the chiral alcohol in the reductive step. This means that the alcohol with the same chirality will be oxidized, leaving behind the other enantiomer in an enantiomerically enriched

P. D’ARRIGO et a].

108

61

62 0

OH

R = CH,, C,H,, n-C,H,,n-C,H,,

OH

Ph

FIG.30.

form. This process is a kinetic resolution and thus allows production of the other enantiomer, albeit at only 50% yield. Since the carbonyl compound thus obtained can be easily reduced to the racemic secondary alcohols by chemical means, in theory a cyclic succession of enzymatic oxidation and chemical reduction allows the entire transformation of the racemic ketone into the other enantiomer. This procedure was applied to the preparation of both enantiomeric forms of propylene glycol (PG) fiom 1-hydroxy-2-propanone (HP) (a) (Fig. 30). The Baker’s yeast reduction of HP gives (R)-PG in high yield and enantiomeric purity. This reaction was published in Organic Synthesis by Levene and Walti (1943). The same reduction is now effected on a large scale (Kometani et al., 1996). (R)-PG is actually used as a chiral building block for the preparation of (S)-ofloxacin (Kumobayashi et al., 1991). The R enantiomer in the racemic mixture is instead oxidized to HP, leaving behind (S)-PG (64). The ketone can then be reduced chemically, giving the racemic alcohol that can again be subjected to enantioselective oxidation. A series of higher homologues was prepared with the same methodology (Kometani et al., 1996). The oxidation was performed with fermenting yeast under aerobic conditions, and equilibrium was reached at 24 h. Kinetic studies of the reaction in the oxidative direction suggest that the activity responsible for the oxidation might be glycerol-DH. Lee and Whitesides (1986) used the GDH from bacterial sources for preparation via a similar oxidative resolution of (S)-1,2-butanediol. The Veschambre group reported the reduction of the diketone @ with Baker’s yeast, giving a mixture of the

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

66

65

109

67

bakefs yeast

FIG.31.

OH -

Ar

n

Ar = 2-furyl, 2-thieny1, phenyl

FIG.32.

hydroxyketone @ and the diol s7 (Besse et al., 1995). The diketone in Fig. 31 was shown to be the substrate for a yeast alcohol dehydrogenase (YADH), while the intermediate hydroxyketone, which is not the substrate for the same enzyme, was further reduced with GDH. From these results it was inferred that the two enzymes were actually involved in yeast reduction of the diketone (Besse et al., 1993). A similar oxidative method was proposed for the kinetic resolution of secondary alcohols of the type shown in Fig. 32. In this case, Baker’s yeast was employed as resting cells and the oxidation required several days to reach equilibrium conditions, in which the ketone/alcohol ratio varied according to the nature of the substrate. Very high enantiomeric excesses of the alcohols were obtained for conversions higher than 70%. Also in this case, of course, the alcohol with (I?)absolute configuration was obtained, opposite to the one obtained in the reduction (Fantin et al., 1993) In the example in Fig. 33, a mixture of ex0 and endo-bicycloheptenol (68)was oxidized by Baker’s yeast to give optically enriched bicycloheptenone (69) (fiom the oxidation of the endo alcohol) and the racemic exo-alcohol (70)unchanged (Dawson et d., 1983). The enantiomeric endo-alcohol was also recovered.

P. D’ARRIGO et ol.

110

68

69

70

FIG.33.

V. Hydrolytic Activities: Phosphate Esters

Proteases (Achstetter and Wolf, 1985), lipases (Schousboe, 1976), esterases (Parkkinen, 1980), phospholipases (Witt et al., 1984), amino acylases (Gliinzer et al., 1986, 1987a), and phosphatases (Trevelyan, 1966) have been recognized in yeast. Yeast glycosidases, in the hydrolysis of a-and P-glucosides, were first investigated by Emil Fischer, who developed the concept of lock-and-key enzyme-substrate interaction after these particular studies (Roberts et al., 1995). These enzymes are at work in many of the fermentative steps in which yeast is used in the food industry. Baker’s yeast has been used as an alternative to p-glucosidase in the synthesis of antirhine from secologanine, where the hydrolytic step is combined with a reductive one (Brown et al., 1991). The hydrolytic activities present, especially lipases or esterases, which can be undesirable, cause byproduct formation. In other cases, the unexpected hydrolytic activity is cooperating with a reductive step in the production of homochiral intermediates (Pedrocchi-Fantoni and Servi, 1991). Their use is justified In some specific applications, as in the resolution of racemic acetylenic alcohols-esters for which there are no efficient alternatives in the otherwise extremely versatile field of lipases (Gliinzer et al., 1987b). Peptide bond formation using immobilized viable Baker’s yeast in reversed micelles has been reported (Fadnavis et al., 1990). Good to excellent yields of peptide have been reported in the condensation reactions leading to a leucine enkephalin analogue (Fig. 34). Yeast was suspended in a reverse micellar medium consisting of aerosol/OT in iso-octane with a low water content. With controlled water activity, 80-90% peptide yields were obtained. Peptide bond formation competes with ester hydrolysis, which becomes important with long reaction times.

-

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST Ac-Phe-OMe + H-Leu-NH, Z-Ala-Phe-OMe + H-Leu-NH, Z-Gly-Gly-Phe-OMe + H-Leu-NH,

-

111

Ac-Phe-Leu-N H2 Z-Ala-Phe-Leu-N H, Z-Gly-Gly-Phe-Leu-NH,

FIG.34.

Enzymes responsible for the hydrolysis of phosphate esters are ubiquitous in living organisms. The number of specific phosphatases recognized in mammalian cells as well as in microorganisms probably exceeds that of any other enzymatic function. In Baker’s yeast the presence of phosphatases, associated with a number of metabolic functions, has been recognized. Many of these enzymes are linked with the metabolism of phospholipids and phosphopeptides. We have observed that some of the phosphate-hydrolyzing capacities present in Baker’s yeast can be useful in selective transformations in phospholipids. The action of fermenting Baker’s yeast transforms a mixture of phospholipids almost completely into glycerol, diglycerides, and inorganic phosphate. Toluene-induced autolysis of Baker’s yeast produces enzymes that catalyze extensive decomposition of the membrane phospholipids but leave substantial amounts of phosphatidylinositol (PI] unchanged. PI, which constitutes about 20% of all the phospholipids present in the yeast membrane, could then be recovered using a rather simplified procedure. Due to the separation difficulties usually encountered with phospholipid mixtures, this method was considered of interest for the preparation of PI (Trevelyan, 1966). The cell-free broth of the fermentation is actually deprived of phospholipase activity while still retaining acid and alkaline phosphatase activity. Therefore, in a crude mixture of phosphatidic acid (PA) containing soy phospholipids, there was selective hydrolysis of PA to inorganic phosphate and diglycerides, leaving the other components unchanged. This selectivity can be exploited in the purification of a mixture of phospholipids arising from a transphosphatidylation reaction on phosphatidylcholine (PC). Indeed, in the reaction of PC with a different alcoholic-bearing head group catalyzed by phospholipase D, a more or less significant portion of PA is formed due to a competing hydrolysis reaction (Fig. 35). Purification of the newly formed phosphatidic ester PX from PA is troublesome. Treatment with a phosphatidate phosphatase contained in the yeast broth allows selective hydrolysis of PA to diglycerides, easily removed by solvent partition (D’Arrigo, 1996).

112 P. D’ARRIGO et a].

OLD AND IWW SYNTHETIC CAPACITIES OF BAKERS YEAST OPO,H,

11 3

OPO,H,

H,O,PO

H,O,P 0

,..-

HO H0Q0p03H2 OH

OPO,H,

OPO,H,

72

71 FIG.36.

Whether the reaction is catalyzed by a specific phosphatase like phosphatidate phosphatase (Wu et al., 1993) or by other phosphatehydrolyzing enzymes has not been investigated. The broth exhibits phosphatase activity, with most phosphate esters submitted for hydrolysis, such as glucose 6-Pand p-nitrophenyl phosphate. Kinetic resolution of isopropylidene glycerol phosphate as either the barium or the calcium salt at alkaline pH with a cell-free fermentation broth at about 50% conversion gave products of low enantiomeric excess. A mammalian phosphatase had been used before in attempted resolution of glycerol in reverse hydrolysis conditions. However, phosphate ester obtained was isolated as a racemate (Pradines et al., 1986). Scollar et al. (1985)reported the resolution of phosphothreonine and phosphoserine using commercially available phosphatases from various sources. Yeast phosphatases have been exploited for hydrolysis of the natural (72) phytic acid 71 of Fig. 36 to ~-myo-inositol-1,2,6-tris-phosphate (Blum et a]., 1995). VI. Lyase Activity

Enzymes from the lyase group are not common in Baker’s yeast. However, some important reactions catalyzed by these enzymes from other microbial sources find important practical applications (Chibata et al., 1983, 1986).It has been found that addition of water onto the double bond of some a$-unsaturated aldehydes (73) generated optically active secondary alcohols (74) (Fronza et al., 1990) (Fig. 37). However, this reaction appears to be severely limited as far as variations in the substrate structure are concerned. Interestingly, it has been observed by other authors that a number of amines add efficiently onto cinnamic esters, giving S-amino acid esters

P. D'ARRIGO et al.

114

OH

0

R=H

ROWOH R

74

73 R' = H R" = COPh. CH,Ph

FIG.37.

of enantiomeric excesses ranging from 20 to 70% ee (Rama Rao et a]., 1990).

Other isolated cases of addition onto an activated double bond have been reported. VII. The Biogeneration of Aroma Compounds

There is current interest in the flavor industry in the generation of substantial quantities of products of relevant sensory properties, occurring in nature in trace amounts, to be used as food additives. In order to be considered as natural, these compounds must either be of extractive origin from natural sources or derive from natural products through manipulation while avoiding chemical reagents. Only biocatalytic or fermentative transformations are acceptable. Materials produced in this manner have enhanced commercial appeal and value since the label of natural products receives consumer acceptance (Fuganti et a]., 1993; Stofberg, 1986). Application of biocatalysis in this field is fairly active, and Baker's yeast has found industrial applications. Figure 38 shows two possible biocatalytic routes for the preparation of natural raspberry ketone, the impact flavor of the fruit. In the first preparation, betuloside, a glycoside from the bark of the birch Betula a h , is treated with a hydrolytic enzyme to give the secondary alcohol betuligenol. Oxidation catalyzed by the yeast C. boidinii allows production of the required ketone. In the second procedure, the unsaturated precursor obtained by basecatalyzed condensation of natural p-hydroxybenzaldehyde and acetone from fermentation is reduced to raspberry ketone, exploiting the wellknown selectivity of yeast in the reduction of a$-unsaturated ketones (Sakai et al., 1991; Kawai et al., 1995a). The ketone is produced industrially on a kilogram scale. (It might be of interest for the reader to know

OLD AND NEW SYNTHETIC CAPACITIES OF BAKER'S YEAST

115

116

P. D’ARRIGO et al.

75

76 FIG.39.

that the synthetic compound produced at about 20 tons per year has a price of 80 US$ per kg, while the biogenerated, but otherwise chemically identical material, is worth 4,000 US$ per kg.) A further example of the biogeneration of aroma compounds is given by the preparation of an important aroma found in fruits during ripen(I?)-&-decanolide0, ing. It can be prepared in its natural form by degradation of ricinoleic acid. Baker’s yeast was employed in the reduction of the corresponding 6-decenolide (75) (massoi lactone) isolated from the barks of the Massoi tree Cryptocaria massoia (Fronza et al., 1992b; van der Schaft et al., 1992) (Fig. 39).

VIII. Conflicting Reports

The idea of reproducibility in biocatalysis is not exactly the same as usually applied in organic synthesis, where a careful description of the experimental procedure must allow anyone to duplicate the experiments. The numerous parameters controlling the behavior of a microorganism are so complex that it is sometimes difficult to repeat the same performance even with the same strain. It is therefore to be anticipated with Baker’s yeast as the biocatalyst that there will be a certain range of uncertainty as far as isolated yields of products and enantiomeric excesses of chiral compounds obtained are concerned. In his pioneering work with yeast, Fischer affirmed that “Fiir die meisten Versuche mit Hefenenzyme ... kann man sich einer guten Brauereihefe bedienen, sicherer aber ist es, eine Reinkultur anzuwenden, wie sie heutzutage kauflich sind. Ich habe mich fiir alle entscheidenden Versuche einer Saccharomyces cerevisiae Typus Frohberg bedient” [For most of the experiments with enzymes from yeast one can use a good Brewer’s yeast, but it is more reliable to use a single-strain culture as they are available nowadays. For all my different experiments I have used a Saccharomyces cerevisiae type Frohbergl (Fischer, 1889). In order to give reproducible results it would suffice to

OLD AND NEW SYNTHETIC CAPACITIES OF BAKER’S YEAST

11 7

always describe a control experiment by using standard commercially available dried yeast (Sigma type 11). Experimental procedures for the preparation of various 3-hydroxyesters have been collected and the author’s procedures have been repeated by the editors using this kind of dried yeast. For all the reported transformations, good agreement between the controlled procedure and the one reported with local yeast has been observed (Roberts, 1992). Recently the NAD(P)H content in living cells in yeasts of different origin has been measured and a method for rapid evaluation of the reducing capacities of a Baker’s yeast brand has been proposed (Pereira, 1995). Sometimes, however, discrepancies in the results observed from different research groups on identical or very similar substrates are difficult to explain. A recent report concerns the cycloaddition of nitriloxide (77)to ethyl cinnamates (78) as catalyzed by Baker’s yeast (Fig. 40). In one article, the effect of added P-cyclodextrin on the regioselectivity of the cycloaddition was studied. It was concluded that the reaction was catalyzed by yeast and that the addition of P-cyclodextrin could completely reverse the regioselectivity of the reactions (Rama Rao et al., 1990, 1992). The isoxazolines 79 and 80 thus obtained were optically active. The authors stated that the reaction did not occur in the absence of yeast. The reaction of nitrile oxide with ethyl cinnamate was reinvestigated by other authors (Easton et al., 1995). They showed that yeast is not required for cycloaddition to take place and that the effect of P-cyclodextrin on regioselectivity was only marginal. That cycloaddition reaction had been reported before in the literature (Christ1and Huisgen, 1968). IX. Conclusions

Applications of Baker’s yeast to organic synthesis usually exploit old enzymes for new transformations. The continuous variation of the structural features of substrates presented to Baker’s yeast for biotransformation point to interesting new possible applications. In some cases, the real effect of Baker’s yeast catalysis is not understood. The nature of the studies published in organic chemistry journals indicates that Baker’s yeast is used for enantioselective reductions of carbonyl groups by researchers who only occasionally apply biocatalysis in organic synthesis. This confirms the tutorial function as well as the role as an asymmetric reducing agent that Baker’s yeast catalysis has had, particularly in the last 15 years. The attempted correlation of transformations

118 P. D'ARRIGO et a].

+

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

119

observed with known enzymatic activities widens the scope of the study of Baker’s yeast as a biocatalyst. The genetic modification of Baker’s yeast allows production of a series of microorganisms with a similar appealing quality as the wild-type organisms, but much enhanced in productivity and selectivity and of great potential for applications in organic synthesis.

REFERENCES Abe, I., Rohmer, M., and Prestwich, G. D. (1993). Chem. Rev. 93,2189. Achstetter, T.,and Wolf, D. H. (1985). Yeast 1,139. Agarwal, S. G., Basu, S. K., Vora, V. C., Mason, J. R., and Pirt, S. J. (1987). Biotechnol. Bioeng. 29, 783. Azerad, R., and Buisson D. (1992). In “Microbial Reagents in Organic Synthesis” (S. Servi, ed.), p. 421. NATO AS1 Series, Kluwer, Dordrecht. Bader, J., and Simon, H. (1980). Arch. Microbid. 127,279. Baik. W., Han, J. L., Lee, K. C., Lee, N. H., Kim, B. H., and Hahn, J.-T. (1994). Tetrahedron Lett. 35,3965. Beecher, J., Breckenridge, I., Roberts, S. M., Tang, J., and Willetts, A. J. (1995). J. Chem. Soc., Perkin Trans. 1 , p. 1641. Besse, P., Bolte, J., Fauve, A., and Veschambre, H. (1993). Bioorg. Chem. 21, 342. Besse, P., Bolte, J., and Veschambre, H. (1995). J. Chem. Edu. 72, 277. Bianchi, G., Comi, G., and Venturini, I. (1984). Gazz. Chim.It. 114,285. Blum, C., Karlsson, S., Schlewer, G., Spiess, B., and Rehnberg, N. (1995). Tetrahedron Lett. 36, 7239. Brown, R. T., Dauda, B., Santos E. N., and Cid, A. M. (1991). J. Chem. Soc., Chem. Commun., p. 825. Buendia, J., Crocq, V., Masson, C., Prat, D., and Vivat, M. (1993). Eur. Pat. Appl. 574,317. Buist, P. H., Marecak, D. M., Partington, E. T., and Skala, P. (1990). J. Org. Chem. 55,5667. Cardillo, R., Servi, S., and n n t i , C. (1991). AppJ. Microb. Biotech. 36, 300. Chen, C. S., Zhou, B. N., Girdaukas, G., Shieh, W. R., Van Middlesworth, F., Gupalan, A. C., and Sih, C. J. (1984). Bioorg. Chem. 12,98. Chibata, I., Tosa, T., and Takata, I. (1983). Trends Biotechnol. 1, 9. Christl, M., and Huisgen, R. (1968). Tetrahedron Lett. 50,5209. Colau, R., and Viel, C. (1980). Bull. Soc. Chim. Fr. 3-4, 163. Colonna, S.,Gaggero, N., Manfredi, A,, Casella, L., Gulotti, M., Carrea, G., and Pasta, P. (2990). Biochemistry 29, 10465. Compagno, C., Tura, A., Ranzi, B. M., and Martegani, E. (1993). Biotechnol. Bioeng. 42, 398. Cornforth, J. W. (1959). J. LipidRes. 1, 3. Crocq, V., Masson, C., Winter, J., Richard, C., Lemaitre, G., Lenay, J., Vivat, M., Buendia, J., and Prat D. (1997). Organic Process R6D. In press. Crout, D. H. G., Dalton, H., Hutchinson, D. W., and Miyagoshi, M. (1991). J. Chem. Soc., Perkin Trans. 1, p. 1329. Czsuk, R., and and Glkzer, B. (1991). Chem. Rev. 91,49. D’Arrigo, P. (1996) Ph.D. Dissertation, Dipartimento di Chimica, Politecnico di Milano, Italy.

120

P. D’ARRIGO et al.

Davey, C. L., Powell, L. W., Turner, N. J., and Wells, A. (1994). Tetrahedron Lett. 35,7867. Dawson, M. J., Lawrence, G. C., Lilley, G., Todd, M., Noble, D., Green, S. M., Roberts, S. M., Wallace, T. W., Newton, R. F., Careter, M. C., Hallett, P., Paton, J., Reynolds, D. P., and Young, S. J. (1983). I. Chem. Soc., Perkin Trans. 1, p. 2119. Deol, B. S., Ridley, D. D., and Simpson, G. W. (1976). Aust. J. Chem. 29,2459. Dondoni, A., Fantin, G., Fogagnolo, M., Mastellari, A,, Medici, A., Nefrini, E., and Pedrini, P. (1988). Gazz. Chim. It. 118, 211. Dugan, R. E., Slakey, L. L., and Porter, J. W. (1970). I, Biol. Chem. 245,6312. Dumas, J. B. (1874). Ann. Chim. Phys. 3,92. Easton, C. J., Merricc Hugues, C., Kirby, K. D., Savage, P., Simpson, G. W., and ‘Iiekink, E. R. T. (1994). J. Chem. Soc., Chem. Commun., p. 2035. Easton, C. J., Merricc Hugues, C., ‘Iiekink, E. R. T., Savage, I?, and Simpson, G. W. (1995). Tetrahedron Lett. 36,629. Eh, M., and Kalessem, M. (1995). Synlett. 837. Fadnavis, N. W., Deshpande, A., Chauhan, S., and Bhalerao, U. T. (1990). J. Chem. Soc., Chem. Commun., p. 1548. Fantin, G., Fogagnolo, M., Medici, A., Pedrini, P., and Poli, S. (1993). Tetrahedron Asymmetry 5,883. Fauve, A., Renard, M. F., Veschambre, H., Madesclaire, M., and Roche, D. (1991). Biocatalysis 4,265. Fessner, W.-D. (1992). In “Microbial Reagents in Organic Synthesis” (S. Servi, ed.), p. 43. NATO AS1 Series, Kluwer, Dordrecht. Fischer, E. (1889). Z. Physiol. Chemie 26,60. Forni, A., Moretti, I., Prati, F., and Torre, G. (1994). Tetrahedron Lett. 29, 11995. Fronza, G., Fuganti, C., Grasselli, P., and Barbeni, M. (1990). Tetrahedron Lett. 33, 6375. Fronza, G.,Fuganti, C., Mele, A., Pedrocchi-Fantoni, G., and Servi, S (1992a). J. Org. Chem. 57, 999. Fronza, G., Fuganti, C., Grasselli, P., Pedrocchi-Fantoni, G., and Servi, S. (1992b). Pure Appl. Chem. 64,1099. Fronza, G., Fuganti, C., Mele, A., Pedrocchi-Fantoni, G., and Servi, S (1994). Tetrahedron 50, 857

Fronza, G., Fuganti, C., Grasselli, P., Servi, S, Zucchi, G., Barbeni, M., and Villa, M. (1995). J. Chem. SOC.,Chem. Commun., p. 439. Fuganti C., and Grasselli, P. (1977). Chem. Ind. (London), p. 983. Fuganti C., and Grasselli, P. (1979). J. Chem. SOC.,Chem. Commun., p. 995. Fuganti C., and Grasselli, P. (1982). J. Chem. Soc., Chem. Commun., p. 205. Fuganti C., and Grasselli, P. (1985). In “Enzymes in Organic Synthesis: CIBA Symposium,” p. 112. Pitman, London. Fuganti C., Grasselli, P., Poli, G., Servi, S., and Zorzella, A. (1988a).J. Chem. Soc., Chem. Commun., p. 1619. Fuganti, C., Grasselli, P., Servi, S., and Hogberg, H. E. (1988b).J. Chem. Soc., Perkin Trans. 1, p. 3061. Fuganti, C., Pedrocchi-Fantoni, G., and Servi, S. (1990). Tetrahedron Lett. 31,4195. Fuganti, C., Pedrocchi-Fantoni, G., and Servi, S. (1991). Agric. Biol. Chem. 55,643. Fuganti, C., Grasselli, P., Servi, S., and Hogberg, H. E. (1992). In “Preparative Biotransformations: Whole Cells and Isolated Enzymes in Organic Synthesis” (S. M. Roberts, ed.). Wiley, New York. Fuganti, C., Servi, S., Barbeni, M., and Cabella, P. (1993). In “Studies in Natural Products Chemistry” (Atta Ur-Rahman ed.), Bioactive Natural Products (Part A), Vol. 13, p. 295. Elsevier, Amsterdam.

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

121

Gibbs, D. E., and Barnes, D. (1990). Tetrahedron Lett. 31, 5555. Gliinzer, B. I., Faber, K., and Griengl, H. (1986). Tetrahedron Lett. 27, 4293. Gliinzer, B. I., Faber, K., and Griengl, H. (1987a). Tetrahedron 43, 771. Glanzer, B. I., Faber, K., and Griengl, H. (1987b). Tetrahedron 43, 5791. Gramatica. P., Manitto, F., Monti, D., and Speranza, G. (1987). Tetrahedron 43, 4481. Gramatica, P., Manitto, F., Monti, D., and Speranza, G. (1988). Tetrahedron 44, 1299. Guarna, A., Occhiato, E. G., Spinetti, L. M., Vallecchi, M. E., and Scarpi, D. (1995). Tetrahedron 51, 1775. Guette, J. P., and Spassky, N. (1972). Bull. SOC.Chim.Fr. 11, 4217. Gupta, K. G., Singh, J., Sahni, G., and Dhawan, S. (1979). Biotechnol. Bioeng. 21, 1085. Hasogawa, J., Ogura, M., Hamaguchi, S., Shimazaki, M., and Watanabe, K. (1981). J. Ferment. Technol. 59, 257. Heffter, A. (1908). Mediz.-Naturwiss. Archiv 1, 81. Heidlas, J., Engel, K.-H., and Tressl, R. (1988). Eur. J. Biochem. 172, 633. Hogberg, H.-E., Hedenstroem, E., Faegerhag, J, and Servi, S. (1992).J. Org. Chem. 57,2052. Hoffmann, R. W., Ladner, W., Steinback, K., Massa, W., Schmidt, R., and Snatzke, G. (1981). Chem. Ber. 114, 2786. Hogberg, H. E., Hedenstrom E., Fagerhag, J., and Servi, S. (1992). J. Org. Chem. 57, 2052. Holland, H. L. (1992). In “Organic Synthesis with Oxidative Enzymes,” p. 393. VCH, New York Hudlicky, T., Tsunoda, T., Gadamasetti, K. G., Murray, J., and Keck, G. (1991). J. Org. Chem. 56, 3619. Hudlicky, T., Gillman, G., and Andersen, C. (1992). Tetrahedron Asymmetry 3, 281. Itoh, T., Takagi, Y., and Fujisawa, T. (1989). Tetrahedron Lett. 53, 6153. Johnson, W. S., Telfer, S. J., Cheng, S., and Schubert, U. (1987). J. Am. Chem. SOC. 109, 2517.

Kamal, A., and Reddy, B. S. P. (1992). Biotechnol. Lett. 14, 929. Kamal, A., Rao, M. V., and Rao, A. B. (1990). Chem. Lett., p. 655. Kawai, Y., Hida, K., Nakamura, K., and Ohno, A. (1995a). Tetrahedron Lett. 36, 591. Kawai, Y., Saitou, K., Hida, K., and Ohno, A. (1995b). Tetrahedron Asymmetry 6 , 2143. Kitazume, T., and Ishikawa, N. (1984). Chem. Lett., p. 1815. Kometani, T., Yoshii, H., and Matsuno, R. (1996). J. Mol. Cat. B 1, 45. Kumobayashi, H., Tachikawa, A., Okeda, Y., and Fujiwara, T. (1991). Jpn. Pat. Appl. 3-204873.

Kuno, S., Bacher, A,, and Simon, H. (1985). Hoppe Seyle’s Z. Physiol. Chem. 366, 463. Kyler, K. S., and Novak, M. J. (1992). In “Microbial Reagents in Organic Synthesis” (S. Servi, ed.), p. 3. NATO AS1 Series, Kluwer, Dordrecht. Lee, L. G., and Whitesides, G. M. (1986). J. Org. Chem. 51, 25. Leuenberger, H. G. W. (1985). In “Biocatalysts in Organic Synthesis” (J. Tramper, H. C. Van der Plas, and P. Linko, eds.), p. 99. Elsevier, Amsterdam. Leuenberger, H. G. W., Boguth, W., Barner, R., Schmid, M., and Zell, R. (1979). Helv. Chim.

Acta 62, 454. Levene, P. A., and Walti, A. (1943). Org. Synth. Collect. 2, 545. Long, A., and Ward, 0. F. (1989). J. Ind. Microbiol. 4, 49. Macleod, R., Prosser, H., Fikentschor, L., Lanyi, J., and Mosher, H. S. (1964). Biochemistry 3, 838.

Nakamura, K., Kawai, Y., Nakajima, N., and Ohno, A. (1991). J. Org. Chem. 56, 4778. Navarro-Ocana, A., Jimenez-Estrada, M., Gonzales-Paredes, M. B., and Barzana, E. (1996). Synlett, p. 695.

122

P. D’ARRIGO et al.

Neidleman, S. (1990). In “The Archaeology of Enzymology in Biocatalysis” (D. A. Abramowicz, ed.), p. 1. Van Nostrand Reinhold, New York. Neuberg, C. (1949). Adv. Carbohydr. Chem. 4,75. Neuberg, C., and Nord, F. F. (1914). Berichte 47,2264. Neuberg, C., and Reinfurth, E. (1923). Biochem. Z. 139,561. Neuberg, C., and Welde, E. (1914a). Biochem. Z. 60,474. Neuberg, C., and Welde, E. (1914b). Biochem. Z. 67,18. Nielsen, J. K., and Madsen, J. 0. (1994). Tetrahedron: Asymmetry 5,403. Ohta, H., Kobayashi, A., and Ozaki, K. (1989). J. Org. Chem. 54,1802. Parkkinen, E. (1980). Cell. Mol. Bid. 26, 147. Pedrocchi-Fantoni, G., and Servi, S. (1991). J. Chem. Soc., Perkin Trans. 1 , p. 1764. Pereira, R. S. (1995). Appl. Biochem. Biotechnol. 55,123. Pradines, A,, KlaBbB, A., PBriB, F., Paul, F., and Monsan, P. (1986). Tetrahedron 44,6373. Prelog, V. (1964). Pure Appl. Chem. 9,119. RamaRao, K. (1992). Pure Appl. Chem. 64,1141. Rama Rao, K., Bhanumathi, N., Srinivasan, T. N., and Sattur, P. B. (1990). Tetrahedron Lett. 31,899. Roberts, S. M., ed. (1992). In “Preparative Biotransformations: Whole Cells and Isolated Enzymes in Organic Synthesis” (S. M. Roberts, ed.). Wiley, New York. Roberts, S. M., Turner, N. J., Willetts, A. J., and Turner, M. K. (1995). “Introduction to Biocatalysis Using Enzymes and Microorganisms.” Cambridge Univ. Press, New York. Sakai, T., Matsumoto, S., Hidaka, S., Imajo, N., Tsuboi, S., and Utaka, M. (1991). Bull. Chem. Soc. Jpn. 64,3473. Sato, T., Hanayama, K., and Fujisawa, T. (1988). Tetrahedron Lett. 29,2197. Schmidt, E.,Blaser, H. U., Fauqueux, P. F., Sedelmeier, G., and Spindler, F. (1992). In “Microbial Reagents in Organic Synthesis” (S. Servi, ed.), p. 389. NATO AS1 Series, Kluwer, Dordrecht. Schousboe, I. (1976). Biochirn. Biophys. Acta 450,165. Scollar, M. P., Sigal, G., and Klibanov, A. M. (1985). Biotech. Bioeng. 27, 247. Secundo, F., Carrea, G., Dallavalle, S., and Franzosi, G. (1993). Tetrahedron: Asymmetry 4,1981. Seebach, D., and Zueger, M. (1982). Hefv. Chim.Acta 65,495. Servi, S. (1990). Synthesis 1. Servi, S. (1996). Chim.Ind. (Milan) 78,959. Sheldon, R. A. (1993). “Chirotechnology.” Dekker, New York. Shieh, W. R., and Sih, C. J. (1993). Tetrahedron: Asymmetry 4,1259. Shieh, W. R., Gopalan, A. S., and Sih, C. J. (1985). J. Am. Chem. Soc. 107,2993. Sih, C. J., Zhou, B., Gopalan, A. S., Shieh, W. R., and Van Middlesworth, F. (1983). In “Selectivity a Goal for Synthetic Efficiency” (W. Bartmann and B. M. Trost, eds.), p. 250. Proc. 14th Hoechst Workshop Conference. Verlag Chemie, Berlin. Simon, H., Bader., J., Guenther, H., Neumann, S., and Thanos, J. (1985). Angew. Chem. Int. Ed. Eng. 24,539. Stofberg, J. (1986). In “Biogeneration of Aromas” (T. H. Parliment and R. Croteau, eds.), p. 2. American Chemical Society, Washington, DC. Takeshita, M., and Yoshida, S. (1990). Heterocycles 3,871. Thanos, J., Bader., J., Guenther, H., Neumann, S., Krauss, F., and Simon, H. (1987). Meth. Enzymof. 136,302. Ticozzi, C., and Zanarotti, A. (1988). Tetrahedron Lett. 29,6167. Trevelyan, W. E. (1966). J. Lipid Res. 7,445.

OLD AND NEW SYNTHETIC CAPACITIES OF BAKERS YEAST

123

Utaka, M., Konishi, S., Okubo, T., Tsuboi, S., and Takeda, A. (1987). Tetrahedron Lett. 28, 1447.

van der Schaft, P. H., Van der Burg, N., Van den Bosch, S., and Cohen, A. M. (1992). AppJ. Microbiol. Biotechnol. 36, 712. Virtanen, A. I., and Csaky, T. Z. (1948). Nature 161,814. Walsh, C. (1979a). Chapter 2 7 in “Enzymatic Reaction Mechanisms.” Freeman, San Francisco. Walsh, C. (1979b). Section 5 in “Enzymatic Reaction Mechanisms,” Freeman, San Francisco. Ward, 0. P., and Young, C. S. (1990). Enzyme Microb. TechnoJ. 12,482. Ward, 0. P., Wilcocks, R., Prosen, E., Collins, S. Dewdney, N. J., and Hong, Y. (1992). In “Microbial Reagents in Organic Synthesis” (S. Semi, ed.), p. 67. NATO AS1 Series, Kluwer, Dordrecht. Witt, W., Schweingruber, M. E., and Mertshing, A. (1984). Biochim. Biophys. Acta 795, 108.

Wong, C. H., Ichikawa, Y., Kajimoto, T., Liu, K. K. C., Dumas, D. P., Lin, Y. C., and Look, G. C. (1992). In “Microbial Reagents in Organic Synthesis” (S. Semi, ed.), p. 35. NATO AS1 Series, Kluwer, Dordrecht. Wu, N. J., Lin, Y . P., Wang, E., Menyl, A. H., and Carman, G. M. (1993). J. Biol. Chem. 268,13830. Xiao. X.Y.. and Prestwich, G. D. (1991). Tetrahedron Lett. 32,6843.

This Page Intentionally Left Blank

Investigation of the Carbon- and Sulfur-Oxidizing Capabilities of Microorganisms by Active-Site Modeling HERBERT L. HOLLAND Department of Chemistry Brock University St. Catharines, Ontario L2S 3A2, Canada

I. Introduction A. Oxygenase Enzymes in Microbial Biotransformations B. Rationalization and Predictability of Oxygenase-Catalyzed Reactions C. Models for Microbial Biotransformations 11. Models for Microbial Hydroxylations A. “Active-Site” Models B. Relationship of Hydroxylase “Active-Site” Models to Enzyme Structure C. Hydroxylation and Phylogeny D. Hydroxylation and Biosynthesis 111. Models for Sulfoxidation Reactions A. “Active-Site” Models B. Sulfoxidation as a Model for Other Processes C. Relationship of Sulfoxidase “Active-Site” Models to Enzyme Structure IV. Summary and Prognosis References

I. Introduction

A. OXYGENASE ENZYMESIN MICROBIAL BIOTRANSFORMATIONS

Microbial biotransformation has a long and distinguished history as a tool for the selective manipulation of organic molecules for synthetic purposes, ranging from an extensive literature on the biotransformation of natural products such as steroids (Charney and Herzog, 1967; Smith, 1974; Mahato and Majumdar, 1993), alkaloids (Rosazza and Duffel, 1986), terpenes (Lamare and Furstoss, 1990), and antibiotics (Sebek, 1980),to applications for the production of low-molecular-weight chiral molecules useful as starting materials in organic synthesis (Davies et a]., 1989).

The range of reactions that can be carried out by microbial methodology is large, covering most of the standard reactions of organic chemistry (Drauz and Waldmann, 1995). Of these, however, the most dramatic transformations are those reactions such as hydroxylation at 125 ADVANCES IN APPLIED MICROBIOLOGY, VOLUME 44 Copyright 0 1997 by Academic Press, Inc. All rights of reproduction i n any form reserved. OO65-2164/97 $25.OO

HERBERT L. HOLLAND

126 H

H

H

R X R

R- X -C HRz

OH

R X R

A

RXH

+

RzCO

(X=N, 0,orS)

FIG.1. Oxygenase-catalyzed biotransformation reactions.

unactivated carbon for which no analogous chemical procedure is known. Such reactions are carried out by the oxygenases, enzymes that require the participation of molecular oxygen in their reaction cycle and that catalyze the introduction of oxygen atoms directly into the substrate. These enzymes are responsible for a wide range of transforma-

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

127

tions, summarized in Fig. 1,that comprise some of the more remarkable reactions of the biotransformation repertoire. For practical purposes, the reactions illustrated in Fig. 1 are most frequently carried out by whole-cell biotransformations using growing or resting cultures of fungi or bacteria. This protocol is dictated by the nature of the enzymes themselves, which for the most part are relatively intractable membrane-bound species that do not lend themselves to simple isolation. In addition, the oxygenase enzymes have extensive cofactor requirements that preclude their economic use in preparative applications. These limitations are particularly relevant in the case of the hydroxylase enzymes, which are generally assumed to be cytochrome P-450-dependent systems with all of the complex cofactor and electron transport requirements that that entails, and the bacterial dioxygenase systems that convert aromatic substrates to cis-dihydrodiol products (Holland, 1992). Those isolated microbial oxygenase preparations that have been described are limited in terms of their practical application to flavin-dependent monooxygenases, particularly BaeyerVilliger monooxygenases such as the cyclohexanone monooxygenase (CMO) of Acinetobacter (Ottolina et al., 1995) and the BVMO enzymes from Pseudomonas putida (Alphand et al., 1996). Many distinct oxygenase enzymes are capable of performing several of the reactions of Fig. 1. The cyt.P-450 monooxygenases are reported to be capable of hydroxylation of both aliphatic and aromatic carbon, epoxidation of olefins, heteroatom dealkylation (via hydroxylation a to oxygen, sulfur, or nitrogen), and oxidation at both sulfur and nitrogen (Holland, 1992). The flavin-dependent monooxygenases are capable of converting phenols to catechols, Baeyer-Villiger oxidation, and oxidation of sulfide to sulfoxide and of amines to N-oxides (Walsh, 1980), while the dioxygenases, once thought to be selective for dihydrodiol formation, are now known to carry out hydroxylation, desaturation, and sulfoxidation reactions in addition to their role in the direct oxidation of aromatic rings (Gibson et al., 1995; Lee et al., 1995). B, RATIONALIZATION AND PREDICTABILITYOF OXYGENASE-CATALYZED REACTIONS

This chapter will consider the various models that have been developed to interpret and predict the outcome of microbial hydroxylation and sulfoxidation reactions. Given the wide range of potentially useful reactions presented in Fig. 1 and the existence of many other useful whole-cell biotransformation reactions, it is not surprising that there

HERBERT L. HOLLAND

128

have been attempts to organize the wealth of empirical data available on microbial biotransformations into a predictively useful format. The rationale behind this type of analysis is twofold: first, the need to be able to predict the regio- and stereochemical outcome of biotransformation of a new substrate by a microorganism whose biocatalytic reactions on other substrates are known; and, second, the need to be able to predict the outcome of biotransformation of a given substrate by a hitherto uninvestigated microorganism. These disparate objectives are related in that their achievement lies ultimately in a knowledge of the natures and activities of the enzymes present in the relevant microorganisms. Considerable progress in this regard has been made in the areas of microbial hydrolytic and oxidoreductase enzyme activities, but in view of the difficulty in isolation and characterization of oxygenases of microbial origin, it is only comparatively recently that serious attempts have been made to explore the fundamental principles that lie behind the regio- and stereoselective nature of oxidative microbial biotransformations. This latter work has developed both as a result of an increasing amount of data having emerged on mammalian oxygenases and as a consequence of hard-won data finally becoming available on the structure and mechanism of a limited number of microbial oxygenase enzymes. As the fundamental approach to the problem of predictivity in wholecell biotransformations is common regardless of the nature of the enzymes concerned, Section 1.C will consider the models that have already been developed for microbial biotransformations other than the hydroxylation and sulfoxidation reactions that are the subject of this chapter.

c.

MODELSFOR MICROBULBIOTRANSFORMATIONS

1. Ester Hydrolysis

Microbial hydrolysis of esters, attributable to lipase activity, is a ubiquitous biotransformation. The reaction frequently proceeds with a degree of enantioselectivity in susceptible substrates, and examination of the hydrolysis of a large number racemic acetates by Rhizopus nigricans led Ziffer and co-workers (Kawai et al., 1981; Ziffer et al., 1983; Charton and Ziffer, 1987) to propose that the enantiomer shown in Fig. 2 is the more rapidly hydrolyzed. This model can also be applied to the hydrolysis of esters of cyclic carbinols (Kasai et al., 1984, 1985). The enantioselectivity of this reaction is highly dependent on the nature of the substituent groups and clearly has its basis in the substrate selectiv-

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

H

129

OCOCH3

SAL

S = small substituent, L = large substituent

FIG. 2. Structure of the enantiomer of acetate esters more rapidly hydrolyzed by Rhizopus nigricans.

S = small substituent, L = large substituent FIG.3. Prelog’s rule for the reduction of prochiral ketones.

ity of the lipases of R. nigricans. The availability of isolated lipase preparations has meant that enantioselective ester hydrolysis has relied less heavily on whole-cell biotransformations but has also reinforced the value of models such as that of Fig. 2 as their basis in enzyme structure and function becomes more apparent (Cygler et al., 1994). 2.

Carbonyl Reduction

Arising from a series of studies originally based on the fungus Curvularia falcata, Prelog (1984) proposed the model of Fig. 3 to account for the enantioface-selective reduction of simple acyclic and cyclic carbony1 compounds to chiral alcohols. Prelog recognized, however, that this rule did not apply to all systems, and that, as it was based on enzyme selectivity, could be complicated by the fact that many microorganisms contain several oxidoreductases that may operate on a common substrate with different selectivities. Although reduction of prochiral carbonyl compounds according to Prelog’s rule is observed for the majority of whole-cell-catalyzed examples, and many isolated microbial oxidoreductases do indeed operate according to the model of Fig. 3, there are several enzymes that express the opposite selectivity (Faber, 1992), and at least one whole-cell system (Yarrowia lipolytica) that reduces both acyclic and cyclic methyl ketones in an antiprelog sense (Fantin et al., 1996). An extension of this model to account for the yeast-catalyzed diastereoselective reduction of a-substituted-p-dicarbonyl compounds such as P-ketoesters and p-diketones was developed by Vanmiddlesworth

130

HERBERT L. HOLLAND

s

S = small substituent, L = large substituent FIG.4. Model for predicting diastereoselectivity in yeast reductions.

and Sih (1987). Their model, shown in Fig. 4, accounts for the regioand diastereoselectivity shown in a large number of such reactions by proposing a selective substrate binding based on the relative sizes of the a substituents, followed by diastereoface-selective delivery of a reducing equivalent from the enzyme’s cofactor. An extension of this model in which an internal hydrogen bond holds the ester group (“L”) away from the incoming reducing equivalent deals specifically with reductions of a-hydroxy-P-ketoesters (Sato et al., 1986). The more complex stereochemical possibilities offered by reductions of racemic bi- and polycyclic ketones have been dealt with by Nakazaki et al. (1980), who proposed a quadrant rule for reductions of such compounds by C u m l a r i a lunata and Rhodotorula rubra. Their analysis, illustrated in Fig. 5, is based on an optimal fit of substrate into four restrictive quadrants, with the relative ease of fit in the order UR > UL > LL > LR, followed by delivery of the reducing equivalent from the -z direction. Their analysis was able to account for the product distribution obtained hom microbial reduction of such ketones as (f)-norbornanone, (*)-twistanone, and (&)-4-protoadamantanone, although the relationship of the model to the selectivities of the various oxidoreductases present in C. lunata and R. rubra was not explored.

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

131

S = small substituent L = large substituent

+X

FIG.5. Quadrant rule for reductions of complex ketones by C. lunata and R. rubra.

3 . Epoxide Hydrolysis

Microbial hydrolysis of epoxides has recently been developed as a powerful tool for the preparation of both chiral diols and epoxides. The enantioselectivity of this reaction has been exploited in the preparation of a number of synthetic targets (e.g., Archelas et al., 19931, and its interpretation in microbial systems is soundly based on an understanding of the mechanism and selectivity of the corresponding mammalian microsomal epoxide hydrolase (Wistuba and Schurig, 1992; Wistuba et al., 1992; Lacourciere and Armstrong, 1993). Models have been proposed to account for the complementary enantioselectivities shown by Aspergillus niger and Beauveria sulfurescens in their conversion of substituted epoxides to vicinal diols (PedragosaMoreau et al., 1996a,b). These models, shown in Fig. 6, predict the preferred regiochemistry of attack by water, together with the predominant enantioselectivity (indicated on the substrate), the latter being determined by the relative sizes of two hydrophobic binding pockets, labeled RB (right back) and RF (right front), present in the respective enzymes. The rate of epoxide opening by B. sulfurescens, involving nucleophilic attack at the benzylic position, is further subject to influence by the conformation of the phenyl ring in the RF pocket (Pedragosa-Moreauet al., 1996a).

Baeyer-Villiger Oxidations The selectivity of Baeyer-Villiger oxidation of cyclic ketones to lactones by Acinetobacter NCIMB 9871 has been the subject of investigation by several groups. Alphand and Furstoss (1992) proposed the 4.

132

HERBERT L. HOLLAND

Epoxide opening by Aspergillus niger

Epoxide opening by Beauveria sulfurescens

FIG.6. Models for the regio- and stereochemistry of epoxide hydrolysis by Aspergillus niger and Beauveria sulfurescens.

Flavin

Flavin FIG.7. Preferred orientations of the enantiomers of a bicyclo[4.2.0]ketone undergoing enantioselective Baeyer-Villiger oxidations by Acinetobacter.

model shown in Fig. 7 based on the whole-cell oxidation of racemic bicyclic ketones. In this analysis, the preferred regio- and enantioselectivity of substrate oxidation was proposed to be determined by preferred binding of the enantiomeric forms of the reaction intermediate as shown, followed by a mechanistically dictated antiperiplanar arrangement of the migrating C-C bond and 0-0 peroxidic bonds of the intermediate, giving rise to the observed products.

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

133

6B FIG.8. Positions of testosterone a or vinylogous to carbonyl subject to hydroxylation under stereoelectronic control.

This mechanistic requirement was later shown to be a dominant feature in the enantioselectivity of the same reaction when catalyzed by the enzyme cyclohexanone monooxygenase (CMO) isolated from Acinetobacter (Kelly et a]., 1995). The latter authors also pointed out that, as the configuration of the tetrahedral peroxidic intermediate is controlled by the diastereofacial selectivity of the addition of a flavin peroxide to the substrate, it may be the latter process, and not interaction of the substrate with the enzyme’s active site, that indirectly controls the enantioselectivity of oxidation of small substrates. However, in an analysis of the oxidation of over 40 different substrates by this enzyme, Ottolina and co-workers (1996) suggested that enzyme-substrate interactions can indeed be a determinant of the selectivity of this reaction, and have proposed an active-site model for the Acinetobacter CMO based on restrictive-space descriptors.

II. Models for Microbial Hydroxylations

A. “ACTIVE-SITE” MODELS 1. Steroid Hydroxylation

As the single most frequently studied group of substrates for microbial hydroxylations, steroids have been the subject of several attempts to develop models that can be used to predict the outcome of these reactions (Holland, 1982). Hydroxylations that occur at positions a or vinylogous to carbonyl groups, such as those at C-2 and C-6 of A4-3-ketosteroids, illustrated in Fig. 8, are controlled by a stereoelectronic requirement for the axial addition of an electrophilic oxidizing species to the appropriate enolic intermediate. This process has been shown to

134

HERBERT L. HOLLAND

co

* OH

4

7.5 FIG.9. Relatiye positions of hydroxylation and binding sites in Calonectria decora (dimensions in A).

be operative for the C-6P-hydroxylation of A4-3-ketosteroidsby Rhizopus arrhizus (Holland, 1984), and for the C-20 hydroxylation of testosterone and related substrates by Gnomonia fructicola (Holland et al., 1988). The sites of hydroxylation at other, unactivated positions of the steroid nucleus may be directed by the position of existing oxygen substituents, which presumably act as binding sites for the substrate in the active site of the hydroxylase enzyme. A spatial relationship between such binding and hydroxylation sites is shown in Fig. 9. This relationship, first derived for hydroxylation of steroids by Calonectria decora, was also found to be generally applicable to hydroxylations carried out by Rhizopus nigricans, R. arrhizus, Wojnowickia graminis, Ophiobolus herpotrichus, Daedalea rufescens, and Leptosporus fissilis, but not to steroid hydroxylations carried out by some other microorganisms such as Aspergillus ochraceus (Jones, 1973). In this model, both carbonyl and hydroxyl groups are capable of acting as binding sites: monofunctional ketones are typically dihydroxylated at two sites as shown, but diketones are bound at two sites and hydroxylated at the third. Other oxygen substituents such as enol ethers and acetals can also exert a direct influence on the position of hydroxylation (Evans et al., 1975),but halogen atoms are ineffective in this role (Bird et al., 1980). Application of this model is, however, complicated by the fact that two-point binding of a steroid in an active site may occur in up to four different orientations, as shown in Fig. 10 (Brannon et al., 1967). A single substrate may thus give rise to up to four different hydroxylation products related by the binding patterns shown in Fig. 10. These spatial relationships account for such hydroxylations as that of 19-nortestoterone ( at l C-16 ) by the 2P-hydroxylator Gnomonia fructi-

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

135

FIG.10. Possible orientations of a steroid as a result of two-site binding [e) with equivalent hydroxylation sites at C-2K-16 and C-1/12 shown (+).

cola (Holland et al., 1988), of 5a-androstan-7,17-dione (2) at C-2 by the C-16 hydroxylating fungus L. fissilis (Denny et al., 1980), hydroxylation of testosterone (3)at both C - l a and C-12p by Penicillium species ATCC 12556 (Tweit et a]., 1962), hydroxylations of gP,lOa-retrosteroids (4J at C-9a by the Ila-hydroxylator Rhizopus arrhizus (Favero et a]., 1979), and the frequent cooccurrence of C-iiaIC-7p and C-llP/C-l4a hydroxylations of androstanes and pregnanes. Disruption of this binding pattern (e.g., by removal of one of the substrate’s oxygen atoms) leads to a shift in the balance of competitive binding, expressed as a change in the ratio of regioisomeric products (Zakelj-Mavric and Belic, 1987). Application of this model to nonsteroid substrates has met with some success. It may account for the hydroxylations of the bicyclic substrates

136

HERBERT L. HOLLAND

FIG.11. Three-point model for the hydroxylation of kauranones by R. nigricans.

(I)R = H (3) R=CH3

o--I’I’

0

4

(4) R = C(O)CH,, 0 or OH

(5) C-4(5) saturated (6) C4(5) unsaturated

STRUCTURES 1-6. 5 and 6 by C. decora and R. nigricans at positions 7-9 A away from the oxygen substituent as shown (Bailey et al., 1977), and has been used to rationalize the hydroxylations of gibberellins at C-15 (Fraga et al., 1993) by R. stolonifer and of 17-norkauran-16-one (2) and ent-17-norkauran16-one (8) by R. nigricans (McCrindle et ~ l .1975) , using the relationships illustrated in Fig. 11. -

2. Terpene Hydroxylation

.

The previously discussed successful application of the steroid hydroxylation model to some terpene hydroxylations suggests that similar predictive relationships may exist between binding and hydroxylation sites for other terpene substrates. The importance of binding sites in terpene hydroxylations is clearly demonstrated by the observations that,

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

137

"OH (17)

FIG.12. Hydroxylation of 1,4-and 1,8-cineole by Bacillus cereus.

whereas myrcene (9 and related hydrocarbons are generally poor substrates for microbial hydroxylations, the derived cyclic sulfones (e.g., lo)are typically hydroxylated in good yield by a variety of microorganisms (Abraham and Arfmann, 1992; Abraham et al., 1992). This phenomenon is also observed for hydroxylations of the phenylcarbamate by Beauveria sulfurescens (Hu et derivatives of dihydroartemisinin (TI) al., 1991) and of geraniol(l2) by A. niger and B. sulfurescens (Zhang et al., 1991), where hydroxylations of the parent molecules proceed in poor yield or not at all. When specific binding sites are absent, hydroxylation of terpenes typically leads to the formation of a mixture of isomeric products.

138

HERBERT L. HOLLAND

OCONHPh (11) STRUCTURES 9-12.

12

STRUCTURE 18.

Hydroxylation of 1,4-cineole (13)by Bacillus cereus yields a mixture of 2-ex0 and 2-end0 alcohols (14and l5),which may arise by hydroxylation of two bound forms of 13 in a single active site, as shown in Fig. 1 2 (Liu et d., 1988).Hydroxylation of 1,8-cineole (16) by B. cereus can be explained by the same model, the sole product (17)arising from preferred binding of the substrate in the orientation illustrated in Fig. 1 2 , left (Liu and Rosazza, 1990). Hydroxylations of the sesquiterpene cedrol 18 by a range of microorganisms occur predominantly in the region of the molecule around C-3 and C-12, and this has been attributed to the role of the C-8 alcohol group in anchoring the substrate in the hydroxylase active site and directing the site of hydroxylation (Abraham et d . , 1987; Lamare et a]., 1987; Fraga et d.,1996).

139

CARBON- AND SULFUR-OXIDIZING CAPABILITIES Hydroxylating site

Hydroxylating site

Binding site

Binding site

Flavonoids

koflavonoids

FIG.13. Model for the hydroxylation of flavonoids and isoflavonoids.

With the exception of the norkauranone substrates illustrated in Fig. 11,there have been no attempts to rationalize microbial hydroxylations

of the higher terpenes. 3 . Hydroxylation of Aromatic Compounds

The microbial oxidation of aromatic compounds to produce phenols is a common biotransformation, but few models dealing with this reaction have been proposed. Fig. 13 presents the proposal of Ibrahim and Abul-Hajj (1990a,b) for the hydroxylation of flavonoids by a range of microorganisms. This proposal accounts for the predominance of hydroxylations at the 3' and 4' carbons in both the flavone and isoflavone series of substrates by the assumption that binding of the substrate oxygen functionalities to polar groups of the enzyme is nonspecific. The importance of these oxygen atoms in substrate binding was confirmed by a study of the microbial hydroxylations of mono- and di-deoxyflavonoids, which demonstrated that at least one oxygen atom in the substrate was necessary for hydroxylation to occur (Abul-Hajj et al., 1991).

A model for the biotransformation of aromatic substrates by Streptomyces griseus (Fig. 14) that encompasses phenol formation together with 0- and N-demethylation has been proposed by Sariaslani and Rosazza (1984). This model suggests the existence of both a nonpolar binding region (the n-binding site) and a polar binding site (the electrophilic region), and accounts for the conversion by S. griseus of alkaloidal substrates such as 19 to phenolic products (R = H + OH), to 0-demethylated products (R = OCH, + OH), and to N-demethylated products (via inverted binding to bring N-CH, to the oxidation site).

140

HERBERT L. HOLLAND

0 = oxidation site E = electrophilic region n

(19) R = OCH, or H

= n-binding site

FIG.14. Model for hydroxylation and 0- and N-demethylation by S. griseus (dimensions in A).

4 . Benzylic Hydroxylation

Mortierella isabellina ATCC 42613 performs the benzylic hydroxylation of a range of aromatic compounds, and for the majority of these substrates the model shown in Fig. 15 can be used to predict the outcome of their biotransformation by this fungus. An earlier version of this model was first proposed to account for the absolute stereochemistry of the hydroxylation of ethylbenzene by M. isabellina (Holland et al., 1993),and the model was later extended to cover the biotransformation of phenyl-substituted olefins to vicinal diols, a reaction proposed to proceed via epoxidation according to the parameters of Fig. 15 (Holland et al., 1994a). The model was recently refined following an analysis of the biotransformation of over 60 substrates and potential substrates by M. isabellina (Holland et al., 1997a). The model proposes a binding region selective for an aromatic ring (A) and an aliphatic binding region (B), both of which are of fixed dimensions defined by molecular modeling of acceptable substrates, together with a polar binding region P located at the rear of B. Figure 15 illustrates its application for the hydroxylation of simple phenyl alkanes and benzylcycloalkanes. The most efficient conversions in terms of yield and enantioselectivity are achieved when R is either of the optimal size for pocket B (e.g., cyclobutyl) or contains a polar substituent capable of interacting with P (e.g., conversion of chroman, 20, to (R)-4-chromanol (a) in >98% enantiomeric purity (Holland et al., 1991a). The model also predicts the regioselectivity observed by Baciocchi et ~ l(1995) . during hydroxylation of 22 by M. isabellina at the site indicated. The ethyl group is best accommodated in pocket B, with the methyl substituent at the rear of A; the alternative binding mode with the ethyl group in A and methyl substituent in B is less favorable.

141

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

/ 6

4

fl

4

A: aromatic binding pocket B: aliphatic binding region P: polar binding site [O]:oxidking site dimensions in Angstrom

Benzyk hydroxyktion of phenyl alkanes (R = akyl) a d benzyl cycloalkanes (R = cycloakyl) by removal of H* to generate (R)alcohols FIG.15. Model for the benzylic hydroxylase of Mortierella isabellina.

5. Hydroxylation of Miscellaneous Compounds

Hydroxylation of 2,2-dimethylcyclohexanone(23) to give the 4-(S)-alcoho1 by E. coli containing the cloned genes of a cytochrome P-450 camphor hydroxylase from Pseudomonas putida has been interpreted in terms of the active-site structure of this enzyme (Yamamoto et al., 1990). The proposed model, shown in Fig. 16, assumes binding of the substrate in a boat configuration analogous to that of the natural substrate, camphor, with hydroxylation occurring exclusively from the direction indicated.

142

HERBERT L. HOLLAND

(20) R = H (21) R = O H

(22)

STRUCTURES 20-22.

S’

Hydrophobic interactions with VAL-295

[OH]+

H

H

*Binding

to lYR-96

Hydroxyhtbn of camphor

Hydrophobic

& interactions . with VAL-295 [OH]-

( 2 3 ) 4 + &

H

H

*Binding

to lYR-96

Hydroxylation of 2,2-dimethylcyclohexmne FIG.16. Hydroxylation of 2,2-dimethylcyclohexanoneby a cloned cyt.P-450 enzyme.

Hydroxylation of a series of substituted adamantanes (24) by various Absidia species leading predominantly to C-4=ial and C-3equatorial products has been explained by the model of Fig. 1 7 (Ridyard et al., 1996). The cooccurrence of hydroxylations at both these sites and the absence of C-4equatorialhydroxylated products requires an “end-on” approach (Fig. 17, right) of substrate to the heme-iron oxidizing center of the enzyme. The ratio of C-3 to C-4 hydroxylated products may then be determined by the nature of the interactions between the substituent R and the amino acid residues of the enzyme that lie perpendicular to the plane of the heme group.

143

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

JQ

H4eq

[OH]

(24)

&

H4ax

R

[OH] H3

H3

<*>

<=> "side-on"approach leading to C-4 equatorial and C-3 products

"end-on"approach leading to C-4 axial and C-3 products

FIG.17. Models for the hydroxylation of substituted adamantanes by Absidia species.

C = cyclic system E = electron rich binding group L = lipophilic substituent [O] = site of hydroxylation

FIG.18. Original model for hydroxylation of cyclic compounds by B. sulfurescens.

6. Hydroxylations by Beauveria Sulhrescens

Beauveria sulfurescens ATCC 7159 (previously classified as Sporotrichum sulfurescens and Beauveria bassiana) is one of the fungi most frequently used for microbial hydroxylations. It has been successfully utilized for the hydroxylation of a range of natural products, synthetic cyclic amides, substituted aromatic compounds, and hydrocarbon substrates, and its application for the hydroxylation of amides was the subject of the first systematic investigation of the parameters that may influence the site and stereochemistry of microbial hydroxylation (Fonken et al., 1967). Their model for the hydroxylation of cyclic alcohols and amides is shown in Fig. 18 and proposes hydroxylation at a methylene group that is part of a ring system, C, to which is attached an electron-rich binding group, E, at an optimum distance of 5.5 A from

144

HERBERT L. HOLLAND

[C] = cyclic structure [B] = binding site for arnide oxygen [O] = site of hydroxylation FIG.19. Later model for hydroxylation of cyclic compounds by B. sulfurescens.

FIG.20. Quadrant rule for hydroxylations of cyclic amides by B. sulfurescens.

C. Group E may also carry a lipophilic substituent, L, which may or may not be part of C. This model was later refined (Fig. 19) to take into account the absolute stereochemistry of the hydroxylation reaction and the experimental observation that hydroxylations could occur over a range of 4-7 A from the binding center (Archelas et al., 1984; Srairi and Maurey, 1987; Fourneron et al., 1989). Hydroxylations of cyclic substrates by B. sulfurescens have also been analyzed in terms of the octant system illustrated in Fig. 20 for the definition of substrate-product relationships (Johnson et al., 1968; Archelas et al., 1984). In this analysis, the newly introduced hydroxyl

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

145

group is placed along the z coordinate of an xyz coordinate system, with the carbon undergoing hydroxylation located at the origin. The bulk of the substrate then preferentially occupies the rear quadrants labeled UL, UR, LL, and LR, in the order UR > UL > LL and LR. The models of Figs. 18-20 were developed specifically to explain the hydroxylation of cyclic amides and related substrates. They fail to satisfactorily account for such other B. sulfurescens-catalyzed hydroxylations as the formation of phenols (Vigne et al., 1986) and hydroxylation of hydrocarbons (Johnson et al., 1973). A more complete analysis of the hydroxylations performed by B. sulfurescens suggests the existence in this microorganism of at least three distinct types of hydroxylase enzymes, one specific for amides and related substrates, one for benzylic hydroxylations, and one for conversion of arenes to phenols (Holland and Zabic, 1996). 7. Dioxygenase-CatalyzedReactions

The microbial conversion of arenes to cis-dihydrodiols has been extensively employed as a synthetic source of the latter. This reaction, performed exclusively by prokaryotic microorganisms, is most frequently carried out using mutants of various Pseudomonas species, and a model that predicts both the regiochemistry and stereochemistry of oxidation of mono- and disubstituted monocyclic arenes by Pseudom o m s putida UV4 has been proposed (Boyd et al., 1995). This model proposes that the outcome of dioxygenation is controlled by the size of substituent groups on the aromatic ring for substituents comprising hydrogen, the halogen group (F, Br, C1, I), and methyl. In all cases with the exception of fluorobenzene, products are formed in high (>99%) enantiomeric purity, as depicted in Fig. 21. B.

RELATIONSHIP OF HYDROXYLASE “ACTIVE-SITE” MODELS TO ENZYME STRUCTURE

The various models discussed above have largely been empirically derived by a correlation of substrate-product relationships. Although often described as “active-site models,” this description is a potential misnomer in the absence of any definite information to the effect that the hydroxylations of a range of substrates by a microorganism are indeed performed by a single enzyme, and in the absence of any complicating factors such as selective transport phenomena. The relevance of these models to the real topography of an enzyme’s active site is therefore speculative, but there are nevertheless several features of the

146

HERBERT L. HOLLAND

OH

major

minor

Q

OH

OH

FIG.21. Models for the oxidation of substituted arenes by Pseudornonas putida UV4.

models that suggest their relevance to specific enzyme-substrate interactions. 1. Significance of Binding Sites

The importance of relatively polar electron-rich binding sites in the substrate in determining both the efficiency and regiochemistry of microbial hydroxylation is a recurring theme. The beneficial effect on the hydroxylation of otherwise nonpolar molecules of introducing polar substituents such as the sulfone or carbamate residue was discussed in Section II.A.2, and the directing influence of oxygen substituents in steroid hydroxylations, discussed in Section 1I.A.1, is well established. The importance of this phenomenon has also been established at the isolated enzyme level, where, for example, the o-hydroxylation of lauric acid derivatives 25 by cyt.P-450 4A1 is dependent on the position of heteroatom substitution in the substrate (Bambal and Hanzlik, 1996).

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

147

(25) STRUCTURE 25.

0

0

'

FIG.22. Relationship between hydroxylations at C-Zcc (0) and C-7p (*) in cyt.P-450 2a-5.

The subtle binding effects that control enzyme-substrate interactions, and hence the regio- and stereochemistry of product formation, in hydroxylase enzymes are illustrated by the observation that a single-site mutation of Phe-209 in cyt.P-450 2a-5 is capable of altering the regiospecificity of hydroxylation of dehydroepiandrosterone (26)by this enzyme (Iwasaki et al., 1995). Substitution of Phe-209 by leucine resulted in an enzyme with 2a-hydroxylase activity, whereas the major site of hydroxylation was shifted to C-7p on replacement of Phe-209 by valine. This shift may be attributable to an inversion of the binding mode, illustrated in Fig. 22, with hydroxylations at C-2a (0) and C-7p (*) of two orientations of the substrate being performed at a single oxidizing site. 2.

Comparisons with cyt.P-450

With the exception of the dioxygenase-catalyzed reactions referred to in Section II.A.8, the bulk of the microbial hydroxylation reactions discussed above are presumed to be carried out by membrane-bound cytochrome P-450-dependent monooxygenases, although the number of examples for which this has been unequivocally demonstrated is small (Yamamoto et al., 1990; Ahmed et al., 1996; Ridyard et al., 1996). In

148

HERBERT L. HOLLAND

view of the lack of detailed structural information available for microbial cyt.P-450, the structure of the atypical soluble cyt.P-450cAM has been used as a starting point for analysis of microbial hydroxylation reactions. This enzyme, for which high-resolution X-ray crystallographic data are available (Poulos et al., 1985), possesses a narrow access channel through which the substrate can pass to a protected active site, and binds its substrate camphor by specific hydrophobic contact interactions with nonpolar active-site residues such as valine or phenylalanine, and by polar interactions though the carbonyl group hydrogen bonding to Tyr-96 (see Fig. 16). It therefore exemplifies the combined features of polar binding and nonpolar size-restricted interactions that appear to be dominant for the microbial hydroxylations discussed above. Indirect information concerning the active-site environment of microbial hydroxylases may therefore be gleaned from a comparison of the known features of these enzymes with similar properties of cyt.P4 5 0 ~ NMR ~ ~ relaxation . studies on Cyt.P-450BM3 from Bacillus megaterium indicate a relatively large active-site region (Modi et al., 1995), unlike the highly restricted environment of the cyt.P-450cAMactive site, and a similar picture may be gleaned from site-directed mutagenesis (Furuya et a]., 1989) and from partial sequence alignments of cyt. P-45ocAM and other membrane-bound cyt.P-450, based on the assumption that areas of common sequence have similar three-dimensional structures (Nelson and Strobel, 1988). This analysis is consistent with the observation that the membrane-bound microbial hydroxylases have a wider substrate specificity than does cyt.P-450cAM but that there are sufficient similarities for the specific substrate binding properties of the latter (the specific combination of polar and nonpolar interactions) to be used as a guide for the regio- and stereoselectivity of hydroxylation by the former.

c. HYDROXYLATION AND PHYLOGENY An alternative to the structure-based approach to rationalization of hydroxylation reactions is the phylogenic approach, the systematic study of the hydroxylation of a single substrate by a wide range of microorganisms in an attempt to relate the regio- and stereochemistry of hydroxylation to the taxonomic classification of the microorganisms concerned. This approach has been systematically applied to hydroxylation of several terpenes (Abraham, 1994; Abraham et al., 1996). Bio-

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

STRUCTURES

149

27-29.

transformation of (+)-isopinocampheol (27) can lead to hydroxylations at C-1, -2, -5, -7, -8 and -9, whereas the (-)-enantiomer 28 is hydroxylated only at C-1, -2, -4, or -9 (Abraham, 1994), and, whereas the by a single microorganism showed some hydroxylation of 22 or enantioselectivity with respect to product distribution, the pattern of hydroxylations across a range of microorganisms, particularly for 27, showed a pronounced phylogenic dependency. Hydroxylation at C-1 was confined almost entirely to bacteria, whereas fungi, particularly the Zygomycotina, showed a predilection for hydroxylations at C-5 and C-7. Hydroxylations at C-2 were carried out by both fungal and bacterial biocatalysts. A similar analysis of the biotransformations of aristolenepoxide (29) has determined the phylogenic frequency of hydroxylations at C-3, -5, -8, -9, -10, -15, and -16 (Abraham et al., 1996). D. HYDROXYLATION AND BIOSYNTHESIS

The eclectic substrate specificity of many of the enzymes of secondary metabolism has led to the biosynthetic pathways of fungal secondary metabolites being used as a guide both for the selection of substrates suitable for biotransformation by those fungi and for the type of reactions that those substrates may undergo (Alam and Hanson, 1990). This technique has been successfully applied to hydroxylations carried out by Cephalosporium aphidicola of analogues of the biosynthetic intermediates of the diterpene aphidicolin 30, a natural product of C. aphidicola. Hydroxylations of the aphidicolin analogues 31-35 by C. aphidicola occurred at the positions indicated in a series of reactions clearly linked to the biosynthesis of 30 (Gordon et al., 1988, 1992; Hanson and Jarvis, 1994), and the sites of hydroxylation of the structurally isomeric diterpene stemodin S also show some analogies to the reactions of the aphidicolin biosynthetic pathway (Hanson et al., 1994).

150

HERBERT L. HOLLAND

HO HO(31) R=CH3 (32) R=CzHS

+OH.*,

J

(33)

i“”,, J

0::’ HO-.‘

(35)

Ill. Models for Sulfoxidation Reactions

A. “ACTIVE-SITE” MODELS 1. Aspergillus niger

Aspergillus niger was among the earliest microorganisms used for the conversion of prochiral sulfides to chiral sulfoxides, and a model (Fig.

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

151

FIG.23. Model for chiral sulfoxidation of phenyl and benzyl sulfides by Aspergillus niger.

23) was proposed to account for the enantioselectivity of the oxidations of a range of substituted phenyl and benzyl sulfides by this fungus (Auret et al., 1968). This model, based solely on the steric differences between R, and R2, predicts that the preferred direction of oxidation is B when R2 is larger than R,. However, the model is not universal, even for A. niger, and predicts only a general trend in the enantioselectivity of oxidation. The model of Fig. 23 has, however, been used to interpret the oxidations of phenyl and benzyl sulfides carried out by an isolated liver microsomal cyt.P-450 enzyme (Takata et al., 1980), which lends credibility to this approach for the rationalization of sulfoxidase enantioselectivity. 2. Mortierella isabellina ATCC 42613

The simple model of Fig. 23 has also been applied to the oxidation of phenyl and benzyl sulfides by M. isabellina ATCC 42613 by redefinition of R1 as an alkyl and R, as a phenyl or benzyl group (Holland et d., 1985). These reactions are, however, more accurately described by the model of Fig. 15, which has recently been extended to cover the enantioselective oxidation of prochiral sulfides (Holland et al., 1997a). Inhibition and induction studies have indicated that the same enzyme of M. isabellina is responsible for both the benzylic hydroxylation and sulfoxidation reactions (Holland et al., 1987), and Fig. 15 can thus be used to predict the structures of substrates that will be efficiently converted to (R)-sulfoxides by M. isabellina. These include aryl alkyl sulfides (e.g., p-bromophenyl methyl sulfide (37,e.e. 100%) or 3,5-dimethylphenyl methyl sulfide (38, e.e. 84%)), in which the aromatic binding pocket (A) of Fig. 15 is optimally occupied, or substrates such as phenyl n-propyl sulfide (39, e.e. 100%) that fill the aliphatic binding region (B). 3. Helminthosporium Species NRRL 4671

The fungus Helminthosporium species NRRL 4671 converts a wide range of prochiral sulfides to the corresponding chiral sulfoxides, the

152

HERBERT L. HOLLAND

(37)

STRUCTURES 37-39.

Hs: s m l hydrophobic pocket HL: large hydrophobic pocket PHP: polar hydrophobic pocket P: polar binding site LPP: lone pair pocket 0: oxidationcentre optimumP to 0 distance, 8 -10 /

Top view, dimensions in 8,

FIG.24. Model for chiral sulfoxidation of Helminthosporium species NRRL 4671.

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

153

os'n~c

US'" H2CH3

(40) (S) sulfoxidation, e.e. 84%

(41) ( R )sulfoxidation, e.e. 25%

(42) (k) sulfoxidation, e.e. 0%

(43)

(s)sulfoxidation,e.e. 45%

STRUCTURES 40-43.

majority of which have (S)-configuration at sulfur. Analysis of over 90 such biotransformations has resulted in the development of a model based on restrictive space descriptors that has been used to rationalize these reactions and also as a predictor of the outcome of Helminthosporium-catalyzed sulfoxidations (Holland et d., 1997b). This model, presented in Fig. 24, was developed from energy-minimized structures of compounds divided into groups of acceptable (>lo% yield) and unacceptable (
s)

154

HERBERT L. HOLLAND

Xl-

R

(44) R = CN, (S)sulfoxidation,e.e. 98% (45) R = NH2, (S)sulfoxidation,e.e. 95%

(46) R = COCH3, (s)sulfoxidation, e.e. 92% (47) R = CN, (S) sulfoxidation,e.e. 80% (48) R = NCS, (S) sulfoxidation, e.e. 93% STRUCTURES 44-48.

group capable of interacting with this site. In all these cases (e.g., 44-48), moderate to good yields and high enantiomeric excesses of sulfoxide are obtained. The model also accounts for the enantioselectivity of Helminthosporium-catalyzed oxidations of nonaromatic sulfides such as 4648,in addition to the oxidations of disulfides such as dithianes and dithiolanes. B. SULFOXIDATION AS A MODEL FOR OTHER PROCESSES

Hydroxylation Reactions A comparison of the one-electron oxidation potential of sulfur in an organic sulfide with the energy required to break a carbon-hydrogen bond makes it clear that sulfoxidation is energetically preferred over hydroxylation (Holland and Carter, 1982; Holland et al., 1985).This has been experimentally verified for such substrates as 49-51, which present the possibility of an intramolecular competition between sulfoxidation and hydroxylation. Oxidative biotransformation of these and similar compounds results in exclusive sulfoxidation; no analogous hydroxylation products are observed (Holland et a]., 1991a,b, 1994a,b). Given the ability of oxygenases (both cyt.P-450- and flavin-dependent monooxygenases, and dioxygenases) to oxidize at sulfur in addition to their normal role of hydroxylation, and the close steric and electronegativity properties of a sulfur center and the methylene group, it 1.

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

(50)

155

(51) STRUCTURES 49-51.

becomes axiomatic that oxygenase-catalyzed oxidation of a substrate in which the normal hydroxylation site has been replaced b y sulfur will result in sulfoxidation. In addition to this site selectivity, the enantioselectivity of sulfoxidation should also correspond to the absolute stereochemistry of the hydroxylation reaction. This is indeed observed for the benzylic oxidations performed by M. isabellina according to the model of Fig. 15, and has also been shown for oxidations of thioanisole by cyt.P-450cAM(Freutel et al., 1994). The use of sulfur as a regio- or stereochemical probe for microbial hydroxylation reactions has not been extensively developed, but it has been applied to isolated enzyme studies. May and Phillips (1980) used the sulfide 52 as a mechanistic probe for dopamine P-hydroxylase; the sulfide was oxidized at a rate exceeding that by which the corresponding carbon analogue was hydroxylated, giving the (S)-sulfoxide corresponding in absolute stereochemistry to the normal product 53 of dopamine P-hydroxylase activity. Alterman and co-workers (1995) investigated the oxidation of the lauric acid analogue 54 (11-thialauric acid) by cyt.P-450 4A1, an enzyme that normally performs terminal oxidation of lauric acid. Both this enzyme and cyt.P-450 2B1 efficiently converted 54 to the corresponding sulfoxide, a process attributed to the greater length of the C(lO)-S bond in 54 relative to the corresponding C-C bond of lauric acid, bringing the sulfur atom within the oxidizing range of the enzyme’s heme iron center. 2. Baeyer- Villiger Reactions

The oxidation of sulfides by the Baeyer-Villiger oxidizing enzyme cyclohexanone monooxygenase from Acinetobacter was used in the

156

HERBERT L. HOLLAND

. . (53)

STRUCTURES 52-54.

M = main hydrophobic binding region HL= large hydrophobic pocket Hs = small hydrophobic pocket 0 = site of oxidation RF = flavin cofactor = 1.54 Angstrom

-

FIG.25. Active-site model for cyclohexanone monooxygenase based on oxidation of prochiral sulfides.

development of an active-site model for this enzyme based on cubic space descriptors (Fig. 25) (Ottolina et a]., 1995). This same model was later extended to the Baeyer-Villiger oxidation of cyclic ketones with little modification (Ottolina et al., 1996), clearly illustrating the value of sulfide oxidation as a stereochemical probe for oxygenase enzymes. 3. Desaturation Reactions

Replacement of a methylene group by a sulfur atom has been used by Buist and Marecak (1994) as a mechanistic probe for the Ag-desaturase

157

CARBON- AND SULFUR-OXIDIZING CAPABILITIES 0

H

SCoA yeast

(55)

\

0 9

SCoA

(56)

FIG.26. Use of sulfur oxidation as a mechanistic probe for fatty acid desaturation.

enzyme of yeast. Yeast containing this enzyme, which normally converts stearoyl CoA (=) to oleoyl CoA (=), as shown in Fig. 26, is capable of the stereoselective oxidation of the corresponding sulfides 57 and 58 with an enantioselectivity corresponding to that of hydrogen removal from 55. The oxidation of a range of 9-thiasubstituted substrates (59)was then used as a probe for the steric limits of the enzyme’s active site, and indicated a restriction in the region corresponding to C-9(10) of the substrate that was consistent with that necessary to impose a cis configuration on the product of desaturation.

c. RELATIONSHIP OF SULFOXIDASE “ACTIVE-SITE’’ MODELS TO ENZYME STRUCTURE

In the absence of any structural information on the sulfoxidase enzymes of A. niger, M. isabellina, or Helminthosporium, correlation of

HERBERT L. HOLLAND

158

R’

(59) R = phenyl, be@, I-naphthyl, 2-naphthyl STRUCTURE 59.

R.SW substrate

experimental

calculated

72:28

6535

48152

22:78

FIG.27. Experimental and calculated enantioselectivity by c y t . P - 4 5 0 ~ ~ .

the models presented in Section 1II.A with enzyme structure is speculative. There are indications, however (again based on analogies with cyt.P-450CAM), that the general features shown by these models, such as determination of enantioselectivity based on substrate fit into hydrophobic regions of defined size, in combination with polar binding where possible, are those that the substrate would experience in a sulfoxidase active site. The disparate stereochemistries of oxidation of thioanisole and p-methylthioanisole by cyt.P-450CAM(Fig. 2 7) are those predicted on the basis of their preferred energy-minimized binding orientation into the active site as determined by molecular mechanics 1994),an observation that lends credence to calculations (Freutel et d., the approach of using substrate-binding interactions to predict and interpret the stereochemistry of sulfoxidation reactions. IV. Summary and Prognosis

A recurring theme in the models discussed above is that both the regio- and stereochemistry of hydroxylation and sulfoxidation of a range

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

159

of substrates are controlled by a combination of two principal considerations, namely, general hydrophobic binding and polar interactions with the enzyme at specific sites. The substrate must clearly be sterically acceptable to the enzyme, both for binding in the active site and for access to that site (although these two different requirements are difficult to separate experimentally without detailed structural and mechanistic information). The view that most microbial oxygenases possess a relatively large active-site region accounts for their wide substrate specificity. Provided that the substrate does not exceed the available limits, it is acceptable to the enzyme, even if this involves a “loose” fit in the active site. However, if high regio- and stereoselectivity of oxidation are to follow, then the substrate should be a “tight” fit, and it is here that optimum size and polar binding of the substrate seem to be of paramount importance. Many of the models discussed in this chapter have had as their goal the definition of the substrate parameters necessary to achieve this “tight-fit’’situation. In view of the lack of isolated enzymes for the elucidation of structural information, the application of models for the regio- and stereochemistry of action of oxygenase enzymes has made remarkable progress. As the application of genetic methodology and other new techniques for the study of oxygenases progresses, more structural information on these enzymes will emerge, but at the present time the use of substrate-derived models provides the best method available for rationalization of microbial hydroxylation and sulfoxidation reactions. REFERENCES Abraham, W.-R. (1994). Phylogeny and biotransformation, Part 5: Biotransformation of isopinocampheol. Z. Natuzforsch. 49c, 553-560. Abraham, W.-R., and Arfmann, H.-A. (1992). Microbial hydroxylation of activated acyclic monoterpene hydrocarbons. Tetrahedron 48,6681-6688. Abraham, W-R., Washausen, P., and Kieslich, K. (1987). Microbial hydroxylation of cedrol and cedrene. Z. Naturforscb. 42c, 414-419. Abraham, W.-R., Arfmann, H.-A,, and Giersch, W. (1992). Microbial hydroxylation of precursors of sinensal. Z. Natuzforsch. 47c, 851-858. Abraham, W.-R., Riep, A,, and Hanssen, H.-P. (1996). Biotransformation and phylogeny, VI: Microbial oxidation of aristolenepoxide to phytotoxins. Bioorg. Chem. 24,19-28. Abul-Hajj, Y. J., Ghaffari, M. A,, and Mehrotra, S. (1991). Importance of oxygen functions in the biological hydroxylation of flavonoids by Absidia blackesleeana. Xenobiotica 21, 1171-1177. Ahmed, F., Williams, R. A. D., and Smith, K. E. (1996). Microbial transformations of steroids, Part X: Cytochromes P-450 1la-hydroxylase and C 1 7 4 2 0 lyase and A 1-ene dehydrogenase transform steroids in Nectria haematococca. 1. Steroid Biochern. Molec. Biol. 58, 337-349.

160

HERBERT L. HOLLAND

Alam, M., and Hanson, J. R. (1990). Biosynthetically directed microbiological transformations with the fungus Trichothecium roseum. Phytochemistry 29, 3801-3803. Alphand, V., and Furstoss, R. (1992). Microbiological transformations, Part 22: Microbiologically mediated Baeyer-Villiger reactions: A unique route to several bicyclic y-lactones in high enantiomeric purity. J. Org. Chem. 57, 1306-1309. Alphand, V., Furstoss, R., Pedragosa-Moreau, S., Roberts, S. M., and Willetts, A. J. (1996). Comparison of microbiologically and enzymatically mediated Baeyer-Villiger oxidations: Synthesis of optically active caprolactones. J. Chem. SOC., Perkin Trans. 1, pp. 1867-18 72. Alterman, M. A., Chaurasia, C. S., Lu, P., and Hanzlik, R. (1995). Heteroatom substitution shifts regioselectivity of lauric acid metabolism from o-hydroxylation to (w-1)-oxidation. Biochem. Biophys. Res. Commun. 214,1089-1094. Archelas, A., Furstoss, R., Waegell, B., Le Petit, J., and Deveze, L. (1984). Transformations microbiologiques, 3: Approche de la topologie du site hydroxylant de Beuuveriu sulfurescens. Tetrahedron 40,355-36 7. Archelas, A., Delbecque, J.-P., and Furstoss, R. (1993). Microbiological transformations, Part 30: Enantioselective hydrolysis of racemic epoxides: The synthesis of enantiopure juvenile hormone analogues (Bower’s compound). Tetrahedron: Asymmetry 4,2445-2446. Auret, B. J., Boyd., D. R., Henbest, H. B., and Ross, S. (1968). Stereoselectivity in the oxidation of thioethers to sulphoxides in the presence of Aspergillus niger. J. Chem. SOC.C, pp. 2371-2374. Baciocchi, E., d’Acunzo, F., and Galli, C. (1995). Steric effects and selectivity in the benzylic hydroxylation by metalloporphyrins and by the fungus Mortierella isabellina. Tetrahedron Lett. 36,315-318. Bailey, A. S.,Gilpin, M. L., and Jones, Sir Ewart R. H. (1977). Microbiological hydroxylation, Part 23: Bicyclic substrates for steroid-hydroxylating fungi. I. Chem. Soc., Perkin Trans 1, pp. 265-271. Bambal, R. B., and Hanzlik, R. P. (1996). Effects of steric bulk and conformational rigidity on fatty acid omega hydroxylation by a cytochrome P-450 4A1 fusion protein. Arch. Biochem. Biophys. 334,59-66. Bird, T. G. C., Fredericks, P. M., Jones, Sir Ewart R. H., and Meakins, G. D. (1980). Microbiological hydroxylation, Part 23: Hydroxylations of fluoro-5-a-androstanones by the fungi Calonectriu decora, Rhizopus nigricans, and Aspergillus ochraceus. J. Chem. SOC.,Perkin Trans. 1, pp. 750-755. Boyd, D. R., Shanna, N. D., and Dalton, H. (1995). Enzyme-catalyzed hydroxylations of aromatic substrates: Stereochemical and mechanistic aspects. In “Organic Chemistry and Reactivity” (B. T. Golding, R. J. Griffin, and H. Maskill, eds.), pp. 130-139. Royal Society of Chemistry, London. Brannon, D. R., Parrish, F. W., Wiley, B. J., and Long, J. (1967). Microbial transformation of a series of androgens with Aspergillus tamarii. J. Org. Chem. 32, 1521-1527. Buist, P. H., and Marecak, D. M. (1994). Use of aromatic thia fatty acids as active site mapping agents for a yeast Ag desaturase. Can. J. Chem. 72, 176-181. Charney, W., and Herzog, H. L. (1967). “Microbial Transformations of Steroids.” Academic Press, New York. Charton, M., and Ziffer, H. (1987). Contributions to steric, electrical, and polarizability effects in enantioselective hydrolyses with Rhizopus nigricans: A quantitative analysis. J. Org. Chem. 52, 2400-2403.

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

161

Cygler, M., Grochulski, P., Kazlauskas, R. J., Schrag, J. D., Bouthillier, F., Rubin, B., Serreqi, A. N., and Gupta, A. K. (1994). A structural basis for the chiral preferences of lipases. I. Amer. Chem. Soc. 116,3180-3186. Davies, H. G., Green, R. H., Kelly, D. R., and Roberts, S. M. (1989). “Biotransformations in Preparative Organic Chemistry.” Academic Press, London. Denny, W. A,, Fredericks, P. M., Ghilezan, I., Jones, Sir Ewart R. H., Meakins, G. D., and Miners, J. 0. (1980). Microbiological hydroxylation, Part XXIV: 16P,18-dihydroxylation of oxygenated 5-a-androstanes by the fungus Leptoporus fissilis. 1. Chem. Res. A4, pp. 345-360. Drauz, K., and Waldmann, H., eds. (1995). “Enzyme Catalysis in Organic Synthesis.” VCH, Weinheim. Evans, J. M., Jones, Sir Ewart R. H., Meakins, G. D., Miners, J. O., Pendlebury, A., and Wilkins, A. L. (1975). Microbiological hydroxylation, Part XVI: Incubation of derivatives (mainly acetals) of 5-a-androstane ketones with the fungi Calonectria decora, Aspergillus ochraceus, and Rhizopus nigricans. I. Chem. Soc., Perkin Trans. 1 , pp. 1355-1359. Faber, K. (1992). “Biotransformations in Organic Chemistry,” p. 142. Springer-Verlag, Berlin. Fantin, G., Fogagnolo, M., Giovannini, P. P., Medici, A,, Pedrini, P., Gardini, F., and Lanciotti, R. (1996). Anti-Prelog microbial reduction of prochiral carbonyl compounds. Tetrahedron 52,3547-3552. Favero, J., Ton That, T., and Winternitz, F. (1979). Bioconversion de 90,lOP steroides par Rhizopus arrhizus Fischer: Etude en RMN du 13C en serie r6tro-st6roide. Bull. Soc. Chem. Fr., pp. 11-56-11-60. Fonken, G. S., Herr, M. E., Murray, H. C., and Reineke, L. M. (1967). Microbiological hydroxylation of monocyclic alcohols. 1.Amer. Chem. Soc. 89, 672-675. Fourneron, J. D., Archelas, A., and Furstoss, R. (1989). Microbiological transformations, Part 10: Evidence for a carbon-radical intermediate in the biohydroxylations achieved by the fungus Beauveria sulfurescens. 1. Org. Chem. 54, 2478-2483. Fraga, B. M., Diaz Gomez, J. C., Ali, M. S., and Hanson, J. R. (1993). Biotransformation of gibberellin 20,lg-lactones by Rhizopus stolonifer. Phytochemistry 34, 693-696. Fraga, B. M., Guillermo, R., Hanson, J. R., and Truneh, A. (1996). Biotransformation of cedrol and related compounds by Mucor plumbeus. Phytochemistry 42,1583-1586. Freutel, J., Chang, Y.-T., Collins, J., Loew, G., and Ortiz de Montellano, P. R. (1994). l‘hioanisole sulfoxidation by cytochrome P45Ocm (CYP101): Experimental and calculated absolute stereochemistries. 1.Amer. Chem. Soc. 116,11643-11648. Furuya, H., Shimizu, T., Hirano, K., Hatano, M., Fujii-Kuriyama, Y., Raag, R., and Poulos. T. L. (1989). Site-directed mutagenesis of rat liver cytochrome P-450d: Catalytic activities toward benzphetamine and 7-ethoxycoumarin. Biochemistry 28, 6848-685 7. Gibson, D. T.,Resnick, S. M., Lee, K., Brand, J. M., Torok, D. S., Wackett, L. P., Schocken, M. J., and Haiglar, B. E. (1995). Desaturation, dioxygenation, and monooxygenation reactions catalyzed by naphthalene dioxygenase from Pseudomonas sp. strain 9816-4. J. Bacteriol. 177,2615-2621. Gordon, J. F., Hanson, J. R., and Ratcliffe, A. H. (1988). Active site mapping in a methyl group hydroxylation in aphidicolin biosynthesis. I. Chem. Soc., Chem. Commun., pp. 6-7. Gordon, J. F., Hanson, J. R., Jarvis, A. G., and Ratcliffe, A. H. (1992). Oxidation of aphidicolin and its conversion into 19-noraphidicolan-160-o1.J. Chem. Soc., Perkin Trans. 1 , pp. 3019-3022.

162

HERBERT L. HOLLAND

Hanson, J. R., and Jarvis, A. G. (1994). The biotransformation by Cephalosporium aphidicola of some aphidicolanes substituted on ring A. Phytochemistry 36, 1395-1398. Hanson, J. R., Reese, P. B., Takahashi, J. A., and Wilson, M. R. (1994). Biotransformation of some stemodane diterpenoids by Cephalosporium aphidicola. Phytochemistry 36, 1391-1393. Holland, H. L. (1982). The mechanism of the microbial hydroxylation of steroids. Chem. SOC.Rev. 11, 371-395. Holland, H. L. (1984). Biotransformations of A4-3-ketosteroids by the fungus Rhizopus arrhizus. Acc. Chem. Res. 17,398-402. Holland, H. L. (1992). “Organic Synthesis with Oxidative Enzymes,” pp. 5-17. VCH, New York. Holland, H. L., and Carter, I. M. (1982). The mechanism of sulphide oxidation by Mortierella isabellina NRRL 1757. Can. J. Chem. 60,2420-2425. Holland, H. L., and Zabic, M. (1996). Unpublished data. Holland, H. L., Popperl, H., Ninniss, R. W., and Chenchaiah, P. C. (1985). The oxidation of organic sulphides by Mortierella isabellina, Part 2: Effects of substituents on the stereochemistry of sulphoxide formation. Can. J. Chem. 63,1118-1120. Holland, H. L., Bergen, E. J., Chenchaiah, P. C., Khan, S. H., Munoz, B., Ninniss, R. W., and Richards, D. (1987). Side chain hydroxylation of aromatic compounds by fungi, Part 1: Products and stereochemistry. Can. J. Chem. 65,502-507. Holland, H.L., Brown, F. M., Chenchaiah, P. C., Chernishenko, M. J., Khan, S . H., and Rao, J. A. (1988). Microbial hydroxylation of steroids, Part 12: Hydroxylation of testosterone and related steroids by Gnomonia fructicola ATCC 11430. Can. J. Chem. 67,268-274. Holland, H. L., Manoharan, T. S., and Schweizer, F. (1991a). Preparation of homochiral chroman-4-01s and thiochroman-4-01s by microbial biotransformation. Tetrahedron: Asymmetry 2, 335-338. Holland, H. L., Rand, C. G., Viski, P., and Brown, F. M. (199lb). Microbial oxidation of benzyl sulfides and bibenzyl by Mortierella isabellina and Helminthosporium species. Can. J. Chem. 69, 1989-1993. Holland, H. L., Kindermann, M., Kumaresan, S., and Stefanac, T. (1993). Side chain hydroxylation of aromatic compounds by fungi, Part 5: Exploring the benzylic hydroxylase of Mortierella isabellina. Tetrahedron: Asymmetry 4,1353-1364. Holland, H. L., Destefano, D., and Ozog, J. (1994a). Side chain hydroxylation of aromatic compounds by fungi, Part 6: Biotransformation of olefins by Mortierella isabellina. Biocatalysis 10, 65-76. Holland, H. L., Brown, F. M., and Larsen, B. G. (1994b). Biotransformation of organic sulfides, Part 5: Formation of chiral para-alkyl benzyl methyl sulfoxides by Helminthosporium species NRRL 4671. Tetrahedron: Asymmetry 5, 1241-1248. Holland, H. L., Allen, L. J., Chernishenko, M. J., Diez, M., Kohl, A., Ozog, J., and Gu, J.-X. (1997a). Side chain hydroxylation of aromatic compounds by fungi, Part 7: A rationale for benzylic hydroxylation, dihydroxylation and sulfoxidation by Mortierella isabellina. J. Mol. Cat. B: Enzymatic. In press. Holland, H. L., Brown, F. M., Lakshmaiah, G., Larsen, B. G., and Patel, M. (1997b). Biotransformation of organic sulfides, Part 8: A predictive model for sulfoxidation by Helminthosporium species NRRL 4671. Tetrahedron: Asymmetry 8 , 683-697. Hu, Y., Highet, R. J., Marion, D., and Ziffer, H. (1991). Microbial hydroxylation of a dihydroartemisinin derivative. J. Chem. Soc., Chem. Commun., pp. 1176-1177. Ibrahim, A,-R., and Abul-Hajj, Y. J. (1990a). Microbiological transformations of (+)-flavanone and (+)-isoflavanone. J. Nat. Prod. 53,644-656.

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

163

Ibrahim, A.-R., and Abul-Hajj, Y.J. (1990b). Microbiological transformations of chromone, c:hromanone, and ring A hydroxyflavones. J. Nut. Prod. 53,1471-1478. Iwasaki, M., Davis, D. G., Darden, T. A., Pedersen, L. G., and Negishi, M. (1995). Multiple steroid-binding orientations: Alteration of regiospecificity of dehydroepiandrosterone 2- and 7-hydroxylase activities of cytochrome P-450 2a-5 by mutation of residue 209. Biochem. 1.306, 29-33. Johnson, R. A., Herr, M. E., Murray, H. C., and Fonken, G. S. (1968). Stereochemistry of microbiological hydroxylation. J. Org. Chem. 33,3217-3221. Johnson, R. A., Hall, C. M., Kreuger, W. C., and Murray, H. C. (1973). Microbial oxygenation of dialkylbenzenes. Bioorg. Chem. 2, ~ g - i i o . Jones, E. R. H. (1973). The microbiological hydroxylation of steroids and related compounds. Pure Appl. Chem. 33, 39-52. Kasai, M., Kawai, K., Imuta, M., and Ziffer, H. (1984). Enantioselective ester hydrolyses employing Rhizopus nigricans. A method of preparing and assigning the absolute stereochemistry of cyclic alcohols. 1.Org. Chem. 49, 675-679. Kasai, M., Ziffer, H., and Silverton, J. V. (1985). Enantioselective ester hydrolyses using Rhizopus nigricans: Stereoselective synthesis and absolute stereochemistry of (-)-cisand (-)-trans-l-hydroxy-4-methyl-1,2,3,4-tetr~ydronaphthalene.Can. 1.Chem. 63, 1287-1 291. Kawai, K., Imuta, M., and Ziffer, H. (1981). Microbially mediated enantioselective hydrolysis of racemic acetates. Tetrahedron Lett. 22,2527-2530. Kelly, D. R., Knowles, C. J., Mahdi, J. G., Taylor, I. N., and Wright, M. A. (1995). Mapping of the functional active site of Baeyer-Villigerases by substrate engineering. J. Chem. Soc., Chem. Commun. 729-730. Lacourciere, G. M., and Armstrong, R. N. (1993). The catalytic mechanism of microsomal epoxide hydrolase involves an ester intermediate. I. Amer. Chem. Soc. 115,1046610467. Lamare, V.. and Furstoss, R. (1990). Bioconversion of sesquiterpenes. Tetrahedron 46, 4109-4132. Lamare, V., Fourneron, J. D., Furstoss, R., Ehret, C., and Corbier, B. (1987). Microbiological transformations, Part 9: Biohydroxylation of a-cedrene and cedrol-synthesis of an odoriferous minor component of cedar wood essential oil. Tetrahedron Lett. 28, 6269-6272. Lee, K., Brand, J. M., and Gibson, D. T. (1995). Stereospecific sulfoxidation by toluene and naphthalene dioxygenases. Biochem. Biophys. Res. Commun. 212,9-15. Liu, W. G., and Rosazza, J. P. N. (1990). Stereospecific hydroxylation of 1,8-cineole using a microbial biocatalyst. Tetrahedron Lett. 31,2833-2836. Liu, W.-G., Goswami, A,, Steffek, P. P., Chapman, R. L., Sariaslani, F. S., Steffens, J. J., and Rosazza, J. P. N. (1988). Stereochemistry of microbiological hydroxylations of 1,4cineole. J. Org. Chem. 53, 5700-5704. Mahato, S. B., and Majumdar, I. (1993). Current trends in microbial steroid biotransformation. Phytochemistry 34,883-898. May, S.W., and Phillips, R. S. (1980). Asymmetric sulfoxidation by dopamine P-hydroxylase, an oxygenase considered specific for methylene hydroxylation. J. Amer. Chem. SOC. 102,5981-5983. McCrindle, R., Turnbull, J. K., and Anderson, A. B. (1975). Microbiological hydroxylation of 17-norkauran-16-one and ent-17-norkauran-16-one with the fungus Rhizopus nigricans. J. Chem. Soc., Perkin Trans. l , pp. 1202-1208.

164

HERBERT L. HOLLAND

Modi, S., Primrose, W. U., Boyle, J. M. B., Gibson, C. F., Lian, L.-Y., and Roberts, G. C. K. (1995). NMR studies of substrate binding to cytochrome P450BM3: comparisons to cytochrome P 4 5 0 ~Biochemistry . 34,8982-8988. Nakazaki, M., Chikamatsu, H., Naemura, K., and Aso, M. (1980). Microbial stereodifferentiating reduction of carbonyl compounds: Proposed quadrant rule. J. 0%.Cbem. 45,4432-4440. Nelson, D. R., and Strobel, H. W. (1988). On the membrane topology of vertebrate cytochrome P-450 proteins. J. Biol. Cbem. 263, 6038-6050. Ottolina, G., Pasta, P., Carrea, G., Colonna, S., Dallavalle, S., and Holland, H. L. (1995). A predictive active site model for the cyclohexanone monooxygenase catalyzed oxidation of sulfides to chiral sulfoxides. Tetrahedron: Asymmetry 6, 1375-1386. Ottolina, G., Carrea, G., Colonna, S., and Riickemann, A. (1996). A predictive active site model for cyclohexanone monooxygenase catalysed Baeyer-Villiger oxidations. Tetrahedron: Asymmetry 7, 1123-1136. Pedragosa-Moreau, S., Archelas, A., and Furstoss, R. (1996a). Microbial transformations, Part 32: Use of epoxide hydrolase mediated biohydrolysis as a way to enantiopure epoxides and vicinal diols: Application to substituted styrene oxide derivatives. Tetrahedron 52, 4593-4606. Pedragosa-Moreau, S., Morisseau, C., Zylber, J., Archelas, A., Baratti, J., and Furstoss, R. (1996b). Microbiological transformations, Part 33: Fungal epoxide hydrolases applied to the synthesis of enantiopure para-substituted styrene oxides. A mechanistic approach. J. Org. Chem. 61,7402-7407. Poulos, T. L., Finzel, B. C., Gunsalus, I. C., Wagner, G. C., and Kraut, J. (1985). The 2.6-A crystal structure of Pseudomonas pufida cytochrome P-450. J. Biol. Chem. 260, 16122-16130. Prelog, V. (1984). Specification of the stereospecificity of some oxido-reductases by diamond lattice sections. Pure Appl. Cbem. 9, 119-130. Ridyard, C. H., Whittaker, R. A., Higgins, S. D., Roberts, S. M., Willetts, A. J., Bailey, P. D., and Rosair, G. M. (1996). Site selective oxidation of tricycl0[3.3.1.1~.~]decane (adamantane) and some of its derivatives using fungi of the genus Absidia. J. Cbem. soc., Perkin Trans. 2, pp. 1811-1819. Rosazza, J. P. N., and Duffel, M. W. (1986). Metabolic transformations of alkaloids. In “The Alkaloids” (A. Brossi, ed.), Vol. 27, pp. 323-405. Academic Press, New York. Sariaslani, F. S., and Rosazza, J. P. N. (1984). Biocatalysis in natural products chemistry. Enzyme Microb. Tecbnol. 6, 242-253. Sato, T., Tsurumaki, M., and Fujisawa, T. (1986). Enantioselective synthesis of (ZS,3S)2,3-dihydroxyalkanoates by the Baker’s yeast reduction of 2-hydroxy-3-oxoalkanoates. Cbem. Lett., pp. 1367-1370. Sebek, 0. K. (1980). Microbial transformations of antibiotics. In “Microbial Enzymes and Bioconversions” (A. H. Rose, ed.), pp. 575-612. Academic Press, New York. Smith, L. L. (1974). Microbial transformations of steroids. In “Terpenes and Steroids” (K. H. Overton, ed.), Vol. 4, pp. 394-530. Royal Society of Chemistry, London. Srairi, D., and Maurey, G. (1987). Hydroxylations microbiologiques de pyrrolidinones-2. Bull. SOC. Cbim. Fr., pp. 297-301. Takata, T., Yamazaki, M., Fujimori, K., Kim, Y. H., Oae, S., and Iyanagi, T. (1980). Stereochemistry of sulfoxides by enzymatic oxygenation of sulfides with rabbit liver microsomal cytochrome P-450. Chem. Lett., pp. 1441-1444. Tweit, R. C., Dodson, R. M., and Muir, R. D. (1962). Microbiological transformations, Part XII: The substrate specificity of hydroxylations by a Penicillium sp., ATCC 12556. J. Org. Cbem. 27, 3654-3658.

CARBON- AND SULFUR-OXIDIZING CAPABILITIES

165

Vanmiddlesworth, F., and Sih, C. J. (1987). A model for predicting diastereoselectivity in yeast reductions. Biocatalysis 1, 117-127. Vigne, B., Archelas, A., Fourneron, J. D., and Furstoss, R. (1986). Microbial hydroxylations, Part 4: Regioselective para hydroxylation of aromatic rings by the fungus Beauveria sulfurescens. The metabolism of isopropyl N-phenyl carbamate (Propham). Tetrahedron 42,2451-2456. Walsh, C. (1980). Flavin coenzymes: At the crossroads of biological redox chemistry. Acc. Chem. Res. 13, 148-155. Wistuba, D., and Schurig, V. (1992). Enantio- and regioselectivity in the epoxide-hydrolase-catalyzed ring opening of simple aliphatic oxiranes, Part 1: Monoalkylsubstituted oxiranes. Chirality 4,178-184. Wistuba, D., Trager, O., and Schurig, V. (1992). Enantio- and regioselectivity in the epoxide-hydrolase-catalyzed ring opening of aliphatic oxiranes, Part 2: Dialkyl- and trialkylsubstituted oxiranes. Chirality 4,185-192. Yamamoto, H., Oritani, T., Koga, H., Horiuchi, T., and Yamashita, K. (1990). Enzymatic preparation of chiral 4-hydroxy-2,2-dimethyl-l-cyclohexanone as a chiral building block. Agric. Biol. Chem. 54, 1915-1921. Zakelj-Mavric, M., and Belic, I. (1987). Hydroxylation of steroids with 11-a-hydroxylase of Rhizopus nigricans. J. Steroid Biochem. 28,197-201. Zhang, X.-M., Archelas, A,, and Furstoss, R. (1991). Microbiological transformations, Part 19: Asymmetric dihydroxylation of the remote double bond of geraniol: A unique stereochemical control allowing easy access to both enantiomers of geraniol-6,7-diol. J. Org. Chem. 56,3814-3817. Ziffer, H., Kawai, K., Kasai, M., Imuta, M., and Froussios, C. (1983). Microbiologically mediated enantioselective ester hydrolyses utilizing Rhizopus nigricans: A new method of assigning the absolute stereochemistry of acyclic 1-arylalkanols. J. Org. Chem. 48,3017-3021.

This Page Intentionally Left Blank

Microbial Synthesis of D-Ribose: Metabolic Deregulation and Fermentation Process P.

DE WULF

Department of Microbiology and Molecular Genetics Harvard Medical School Boston, Massachusetts 021 15

E. J. VANDAMME Laboratory of Industrial Microbiology and Biocatalysis Department of Biochemical and Microbial Technology Faculty of Agricultural and Applied Biological Sciences University of Ghent B-9000 Ghent, Belgium

I. 11. 111. IV. V. VI.

VII.

VIII.

IX.

X. XI.

Introduction Natural Occurrence of u-Ribose and Its Derivatives Physicochemical Characteristics of u-Ribose Detection and Identification of u-Ribose Applications of o-Ribose Nonmicrobial Production of D-Ribose A. Chemical Hydrolysis of Yeast RNA B. Enzymatic Hydrolysis of Yeast RNA, Ribonucleosides, and Ribonucleotides C. Chemical Production of n-Ribose Microbial Production of u-Ribose A. ~-Ribose-5-Phosphate,an Intermediate of the Pentose Phosphate Pathway B. Goal-Oriented Creation of Transketolase-Deficient Microorganisms C. Random Isolation of D-Ribose Producing Microorganisms D. Creation of Industrially Applied Transketolase-Deficient D-RiboseProducing Strains E. Improvement of D-Ribose Productivity by Recombinant DNA Technology Pleiotropic Properties of u-Ribose-Producing Transketolase-Defective Bacillus Mutant spp. A. Defective Phosphoenolpyruvate-DependentPhosphotransferase System B. Deregulated Carbohydrate Catabolite Repression C. Altered Cell Membrane and Cell Wall Composition D-Ribose Production by Fermentation with Bacillus spp. A. Cultivation of the Inoculum B. Composition of the Fermentation Medium C. Fermentation Conditions D. Recovery of u-Ribose from the Fermentation Medium Kinetics of u-Ribose Production by Bacillus spp. Conclusions and Future Perspectives References 167 ADVANCES IN APPLIED MICROBIOLOGY, VOLUME 4 4 Copyright 0 1997 by Academic Press, Inc. All rights of reproduction in any form reserved. 0065-2164197 $25.00

168

P.

DE

WULF AND E. J. VANDAMME

I. Introduction

From the moment D-ribose was discovered by Kossel(1891),attention on this carbohydrate has shifted from its physicochemical characteristics to its role in cell metabolism (backbone of (deoxy)ribonucleic acids, ATP, and coenzymes), up to its use as a substrate to synthesize antiviral and anticancer medicines. This line of events can be correlated with the increasing knowledge on cell metabolism and the amount of D-ribose at hand, allowing a specific kind of research to be performed. Until the 1950s, only microscopic amounts of D-ribose were available, laboriously derived by hydrolyzing yeast RNA. Its subsequent organochemical synthesis from D-glucose (Sowden, 1950; Smith, 1955) made this carbohydrate accessible in quantitative amounts, allowing more fundamental research on this pentose, and on its role in cell physiology, thereby using isotopically labeled D-ribose. Excellent work performed by the Takeda Company (Japan)from the 1960s on led to the industrial synthesis of D-ribose with transketolase-defective Bacillus mutant spp. Optimized and alternative fermentation processes, as well as the introduction of recombinant DNA technology, subsequently enhanced production efficiencies (higher yields and decreased side-product formation). This chapter shall focus on the methods developed to produce D-ribose, and in particular on microbial D-ribose production. The mutant strains used, their altered biochemical characteristics, and the fermentation processes in which they are applied shall be considered in detail. This review could not be complete without an overview of the (potential) applications of D-ribose and suggestion of some perspectives on what the future may hold for this pentose. II. Natural Occurrence of D-Ribose and Its Derivatives

In 1891, Kossel found that an acid-based hydrolysis of what is now called ribonucleic acid (RNA)liberated a carbohydrate derivative. Only 3 years later, Hammarsten (1894)demonstrated that its sugar component was a pentose. Some 15 years elapsed before Levene and Jacobs (1908, 1909, 1911) succeeded in isolating the sugar in its crystalline form and in identifying it as D-ribose, which differed in its physical characteristics from the pentoses known at that time (L-arabinose, D-xylose, and D-lyxose). In addition to its occurrence as a backbone of ribonucleic acid and of nucleotides such as adenosine mono-, di-, and triphosphate, D-ribose has been found in combination with uric acid in the blood (Davis et al.,

MICROBIAL SYNTHESIS OF D-RIBOSE

169

1922) and united with 2-hydroxy-6-aminopurine in the croton (Croton tiglium) bean (Cherbuliez and Bernhard, 1932). Cobalamin (vitamin B,,), ribonucleosides, and coenzymes (NAD(P) and coenzyme A) all contain substituted D-ribose (Sasajima and Yoneda, 1989). Moreover, in all naturally occurring D-ribose derivatives the sugar is present in furanose form. The carbon atoms of D-ribose also serve in the synthesis of the pyrrole ring of the aromatic amino acids, the pyrrollopyrimidine nucleosides, and the pterine ring of folic acid (Suhadolnik, 1970). D-ribose is also found in the capsular polysaccharides of Escherichia coli strain 09:K138 (Hackland et al., 1991), E. coli strain 044:K74:H18 (Ahrens et al., 1988), Eubacterium sabeurum (Sasajima and Yoneda, 1989), Staphylococcus spp. (Kotilainen et al., 1990), Streptomyces spp. (Inoue et al., 1992), and the protozoans Blastocrithidia culicis (Merello et al., 1994) and Endotrypanum schaudinni (Merello et al., 1995). This may show that D-ribose is functionally associated with the adherence capacity of these strains and with the immunoresponse generated by the pathogenic species amongst them. It also occurs as a structural unit of the cell wall polysaccharides of Hafnia alvei (Jachymek et al., 1995), Salmonella arizonae (Shashkov et al., 1993), S . typhymurium (Sasajima and Yoneda, 1989), and the red alga Rhodella reticulata (Dubinsky et al., 1992). The immunologically active type-specific substance of Haemophilus influenzae type b consists of a polyribophosphate chain, as found in ribonucleic acid. The position of the purines and pyrimidines, though, is replaced by a second similar chain, linked to the first via 1,l’-glycosidic linkages (Zamenhof et al., 1953). The nucleoside nebularine, isolated from the mushroom Agaricus nebularis, gave purine and D-ribose on hydrolysis (Lofgren and Luning, 1953).The flavonoids of Phlebodium dictyocallis (Mett.)Gomes contain D-ribose as a component of the D-glycosidic moieties (Gomez and Wallace, 1986). Besides, D-ribose is found in the coccoliths of Emiliana huxleyi (Lohmann) Hay and Mohler as part of an acidic Ca2+-binding polysaccharide (Borman et al., 1987). That D-ribose is a highly reactive (aldehyde-contributing) substrate in nonenzymatic glycation (Maillard reaction) is well known. In fact, the D-ribose/arginine-lysine reaction model is used to study the kinetics of nonenzymatic browning (Alabed et al., 1995; Khalifah et al., 1996), the mutagenicity of the generated furfural (Cuzzoni et al., 1989), and the reaction conditions that favor/minimize its formation (Cuzzoni et al., 1988; Khalifah et al., 1996). D-ribose-based glycation of long-lived proteins may be a major contributor to the pathology of aging (Dyer et d.,

1993), diabetes (Bailey et al., 1995), cataracts (Liang and Rossi, 1990; Nagaraj and Monnier, 1992), and Alzheimer’s disease (Smith et al., 1995). The best-known phosphorylated D-ribose derivative, ~-ribose-5-phosphate, is a key intermediate in the pentose phosphate pathway (Wood, 1985), which amongst other functions (see later discussion) plays a central role in the synthesis of aromatic amino acids and vitamins. 5-phosphoribosyl-1-pyrophosphate(PRPP)and the 1’-phosphate ester of D-ribose are intermediates in purine and pyrimidine nucleotide synthesis (Rawn, 1989). ~-ribose-l,5-biphosphate,a coenzyme in the riboconvermutase-catalyzed ~-ribose-l-phosphate/~-ribose-5-phosphate sion during nucleoside synthesis (Leer and Hammer-Jespersen, 1975) has been isolated from human erythrocytes (Vanderheiden, 1970). 5-methylthioribose is a structural unit of 5-methylthioadenosine that plays a central role in the metabolism of adenosylthiomethionine (Cornell et al., 1996). It was first isolated by Baddiley (1955) from yeast and crude vitamin B1. Glutamyl-~-ribose-5-phosphatehas been identified as an abnormal storage product in the brain and kidneys of a patient suffering from renal failure and neurological deterioration (Williams et al., 1986). Contrary to D-ribose, 2-deoxy-~-riboseis the backbone pentose in deoxyribonucleic acid (DNA),which carries the genetic information of all living organisms, except RNA viruses (Rawn, 1989). 2-deoxy-D-ribose was first isolated from thymus nucleotides by Levene et al. (1930). 3-deoxy-~-riboseis a component of the antibiotic cordycepin, produced by Cordyceps militaris (Bentley et al., 1951).The D-ribose derivative 3-amino-3-deoxy-~-ribose was a hydrolytic product of puromycin, an antibiotic produced by Streptomyces alboniger (Baker and Schaub, 1953), and 3-amino-3-deoxyadenosine, produced by Helminthosporium spp. (Gerber and Lechevalier, 1962). Reduction products of D-ribose have also been found in nature. D-ribito1 was isolated from Adonis vernalis (Merck, 1893), identified in the roots of Bupleurum falcatum (Wessely and Wang, 1938), as well as in honeydews, secreted by the genus Ceroplastes (Hackman and Trikojus, 1952). This pentitol also occurs as a monomer in the capsular polysaccharide of pneumococci and in the cell walls of Proteus mirabilis and Vibrio parahaemolyticus (Sasajima and Yoneda, 1989). Besides, as a ribityl residue it is present in the cell wall-associated teichoic acid of most (Gram-positive)bacteria (Schlegel, 1984). The reduced ribityl residue forms part of riboflavin (vitamin B,) and the coenzyme flavine mononucleotide. Despite the absence of a ribitol ring structure in these components, the stereochemistry of this carbohydrate residue is of great

MICROBIAL SYNTHESIS OF D-RIBOSE

OH

171

OH

FIG.1. D-ribose (a-furanose form].

biological importance (Hirst, 1949). Analogous flavins that contain other sugar residues have different or no biological activity. Flavin adenine dinucleotide, a mixed diphosphate ester of riboflavin and adenosine, contains both ribofuranose and ribityl residues in its molecule (Sasajima and Yoneda, 1989). The oxidized forms of D-ribose, D-ribonic acid and D-riburonic acid, are produced by pseudomonads (Weimberg, 1961) and occur in the cell wall-associated lipopolysaccharide of Rhizobium melilotti (Amemura et a]., 1981), respectively. Ill. Physicochemical Characteristics of D-Ribose

D-ribose (C5HI0O5) has a molecular weight of 150.13 g/mol. Its structure is shown in Fig. 1. D-ribose is water-soluble but does not dissolve in ethanol, in which it forms colorless and hygroscopic needle-to-sheet-formed crystals (Sasajima and Yoneda, 1989). The melting point of D-ribose in an aqueous solution is between 84 and 87°C (Hiroshi et a]., 1985; Jeanlotz and Fletcher, 1951). The mutarotation of D-ribose in an aqueous solution ([a]20D: -23.7O) (Phelps et al., 1934) involves the reversible intramolecular attack of the sugar hydroxyl group upon the carbonyl carbon of the open-chain species, to form four ring anomers and one open-chain intermediate. The carbonyl carbon of the open-chain form can also be hydrated upon attack by water. Additional “pseudoacyclic” intermediates (Isbell et al., 1969) have been proposed for D-ribose (Maier, 1977), but there is no convincing experimental basis for their existence (Mega et al., 1990). In an aqueous D-ribose solution, an approximate furanose:pyranose equilibrium of 20:80 occurs at room temperature, while a 30:70 ratio is

found at 80°C (Carmona and Molina, 1990). At these temperatures, the a-and p-anomeric forms relate in 1:3 furanose and 1:2 pyranose ratios. Only small changes in the configuration of D-ribose occur between 2 1 and 48OC. The amount of a-pyranose and a-furanose forms slightly increases with temperature, at the expense of the P-configurations (Cortes et al., 1991). No significant differences in the conformational behavior of D-ribose appear in the pH range between 4.8 and 7.0 (Cortes et al., 1991). Lagrangian dynamics have been worked out with application to the pseudorotation dynamics of D-ribose (and 2-deoxy-~-ribose) (Rudnicki et al., 1994). NMR (Horton and Walaszek, 1982; Rudrum and Shaw, 1965) and infrared and Raman spectroscopy (Carmona and Molina, 1990) showed that crystalline D-ribose occurs in a P-pyranose configuration. A polarographic study of D-ribose was described by Cantor and Peniston (1940),while its crystallographic properties were reported by Bryant (1989). More information on the physicochemical properties of D-ribose and its derivatives can be found in classical review articles written by Jeanlotz and Fletcher (1951) and Overend and Stacey (1955). IV. Detection and Identification of D-Ribose

D-ribose can be identified using tests for analyzing reducing sugars and pentoses (e.g., orcinol reaction; Dische, 1962). For molecular characterization of D-ribose, substituted hydrazones (e.g., p-bromophenylhydrazone, benzylphenylhydrazone, and p-toluenesulfonylhydrazone) have been preferred (Jeanlotz and Fletcher, 1951). Such chromatographic methods as paper partition chromatography, thin-layer chromatography (assayed by densitometry), gas chromatography, high-pressure liquid chromatography, and ion chromatography have been used to identify and quantify D-ribose in mixtures of carbohydrates (Chaplin and Kennedy, 1986; De Wulf et al., 1996a). Determinations on crystalline samples of the melting point, specific rotation, optical crystallographic properties, and infrared, Raman, and 13C NMR spectroscopy may serve to differentiate D-ribose from other pentoses (Bryant, 1989; Carmona and Molina, 1990; Cortes et al., 1991). V. Applications of o-Ribose

On a large scale, D-ribose is used as a substrate to chemically synthesize riboflavin (vitamin B2), which is applied in pharmacy, medicine, cosmetics, (health) food, and feed (Sasajima and Yoneda, 1989).Vitamin

MICROBIAL SYNTHESIS OF D-RIBOSE

173

B2 is also considered a yellow biopigment (E101) (Artz and Jenkinson, 1994).

1)-ribose has been used to synthesize such 5’-nucleotide flavor enhancers as inosine monophosphate (IMP) and guanosine monophos, phate (GMP) (Atkinson and Mavituna, 1991; Miyagawa et ~ l .1992). D-ribose as such was tested to treat myocardial ischemia (e.g., thrombosis) (Zimmer, 1992). It enhanced the heart’s tolerance to ischemia, stabilized the cardiac membrane osmotically, and improved overall cardiac functionality (Zimmer, 1992; Zimmer et al., 1984). Besides, D-ribose infusions may facilitate thallium-201 redistribution in patients suffering from coronary artery disease and hence may improve identification of an ischemic myocardium (Perlmutter et d.,1991). However, placebo-controlled and double-blind experiments, could not reproduce these reports (Hendel et d.,1995). Exercise-induced muscular pain and/or stiffness due to myoadenylate deaminase deficiency (MAD disease) and muscle aches resulting from an intracellular shortage in phosphorylase (MacArdle disease) have been treated successfully by administering D-ribose (Gross et d., 1991; Zimmer, 1992). Double-blind and placebo-controlled crossover trials with MacArdle patients, however, showed no significant interpersonal differences in oxygen consumption or leg fatigue (Steele et al., 1996). More experiments need to clarify the disputes on this subject. Concerns about the safety of D-ribose therapy have been triggered by reports that D-ribose inhibits the proliferation of quiescent peripheral lymphocytes in vitro (Barbieri et d.,1994; Marini et d.,1985, Zunica et d . , 1986). This feature also remains controversial, since Pliml et al. (1993) showed that increased D-ribose plasma levels do not restrain human lymphocyte proliferation. Since therapeutic drugs need to interact with specific cell membrane receptors to be functionally active, medicines are often provided with a molecular tail to enhance cell-drug interaction. By ribosylating antibiotics, their therapeutic possibly may increase (Sasajima, 1976).On the other hand, drug activity can be inhibited by ribosylating the pharmaceutical compound, as was observed with Mycobacterium smegmatis DSM 43756, which inhibited the action of rifampicin by ribosylating the drug (Morisaki et al., 1995). Chemical alteration of the drug’s structure, to avoid ribosylation, may thus be a way to counteract resistance against leprostatic agents. Bicyclo[3.1.0]hexanes have been prepared from D-ribose by intramolecular cyclopropanation. These derivatives can be used to synthesize steroids, prostanoids, vitamin D analogues, terpenoids, and amino acids (Gallos et al., 1994b).

Starting with D-ribose, (1S,4S,5R)-N-acetyl-4-methoxymethyl-3-oxa6-azabicyclo[3.l.0]hexan-2-one, a cyclic analogue of aziridine 2-carboxylates, was synthesized by Dubois and Dodd (1993). From this analogue, modified a-and p-amino acids can be made by nucleophilic attack of the aziridine ring at C-2 or C-3, respectively. The large number of nucleophiles suitable for this reaction makes a great variety of optically pure substituted amino acids accessible, which might be used as antidiabetics (Horii et al., 1986). (2S,3R,4S)-2-amino-1-cyclohexyl-6-methylheptane-3,4-diol, a dihydroxyethylene dipeptide isostere of renin inhibitors, was regio- and stereospecifically synthesized from D-ribose in an overall yield of 22% (Chan and Hsiao, 1992). Renin inhibitors have attracted a great deal of interest due to their potential use as antihypertensive agents (Greenly, 1990). D-sedoheptulose, produced from D-ribose or D-ribose plus D-glucose by certain Bacillus subtilis strains (Yokota, 1993), can be used to synthesize chirally substituted furans and a-D-glucosidase-inhibitingpseudo amino sugars (e.g., N-substituted valioamine derivatives) that display antidiabetic activity (Horii et al., 1986; Yokota, 1993). Since the enzymes that bring about virus replication in a host cell differ remarkably kom their host cell counterparts, nucleoside analogues have been developed from (2-deoxy) D-ribose, which, due to the broadened substrate specificity of the virus-specified enzymes, are inhibitory only to viral nucleic acid synthesis (De Clercq and Torrence, 1978; D o h , 1985). In this view, such pharmaceuticals as ribavirin, pyrazofurin, 5-iodo-5'-amino-2',5'-dideoxy uridine, cytosine (adenine) arabinoside, and adenine (guanine) derivatives promise highly efficient (10 mg/kg body weight) and selective antiviral activity. Chemically originating from 2-deoxy-D-riboseand 2-deoxynucleosides, such antiviral 2,3-dideoxynucleosides as AZT (Zidovudine, 3'-azido-3'-deoxythymine) have been commercialized for some years now and are used in AIDS chemotherapy (treatment of AIDS) and AIDS chemoprophylaxis (prevention of AIDS) (Dueholm et d., 1992). Based on the substantial production of the above-mentioned dideoxynucleosides, it should be of interest to synthesize 2-deoxy-D-ribose from D-ribose in an efficient way. 2-deoxy-D-ribose can be produced chemically starting from D-glucose via an established methodology that results in an overall 55% conversion yield (Smith, 1955).Due to its high price and since the selective protection of the D-ribose hydroxyl groups is hard to perform, the chemical ~-ribose/2-deoxy-D-riboseconversion has not been worked out, at least to the best of our information (note of

MICROBIAL SYNTHESIS OF D-RIBOSE

175

the authors). As the efficiency of D-ribose synthesis via microorganisms has increased significantly over the last years, the D-ribose/Z-deoxy-Dribose conversion chemistry might be an interesting research target. D-ribose could, for example, be oxidatively converted into D-ribonolactone in the presence of bromine and barium carbonate. The latter could then be converted in a solution of hydrogen bromide and acetic acid Catalytic hydrogenolysis of into 2-bromo-2-deoxy-~-arabinolactone. this lactone derivative would yield the corresponding 2-deoxy-~-erythropentolactone, which can be reduced-in the presence of diisoamylborane-to 2-deoxy-D-ribose (Lundt, personal communication, 1993). Methyl inosine monophosphate, a new purine derivative prepared from D-ribose, is under preclinical development, as it attenuates HIV-induced immunosuppression (Hadden et al., 1992). Modified nucleoside antiviral and anticancer agents such as neplanocin-A (Hill et al., 1994), formycin and pyrazofurin (Rycroft et al., 1995), as well as some analogues of their key intermediates (Gallos et al., 1994a), have all been enantioselectively synthesized from D-ribose. The preparation of a homochiral pentacyclic molecule from D-ribose was reported by Cooper and Salomon (1990). This precursor was used to synthesize halochondrins, highly effective antitumor agents in vivo. Low doses of halochondrins (10 pg/kg body weight) provided a suppressive response to melanoma and leukemia in mice (Cooper and Salomon, 1990). The enantiomerically pure Z-vinyl iodide has been synthesized from D-ribose for conversion into taxol and its congeners (Paquette and Bailey, 1995), therapeutics effective against refractory ovarian cancer. The carba-sugars gabosine C (antibiotic, also isolated from Streptomyces) and gabosine E (inhibits cholesterol synthesis) both have been chemically prepared from D-ribose (Lygo et al., 1994). Due to the potent herbicidal and plant growth regulatory activity of hydantocidin, a compound secreted by Streptomyces hygroscopicus SANK 63584, considerable interest has grown in chemically derived hydantocidin and its stereoisomers (Fairbanks et al., 1993). Azidolactones, prepared from D-ribose, were transformed into l-epihydantocidin, which was more thermostable than the natural product (Fairbanks et al., 1993). The use of D-ribose as a chiral synthon has thus become a well-established tool to synthesize biologically active compounds. The search for molecules that possess therapeutic utility continues to present a challenging target for medicinal chemists. This may be stimulated by the increasingly efficient and economical (microbial) D-ribose production methods that are being developed.

176

F.'

DE

WULF AND E. J. VANDAMME

VI. Nonmicrobial Production of D-Ribose

A. CHEMICAL HYDROLYSIS OF YEASTRNA

Until the 1950s, yeast RNA was the best source of D-ribose, so that its preparation from yeast RNA has been investigated extensively. Levene and Clark (1921) developed a method that involved hydrolysis of ribonucleic acid with ammonia at high temperature and pressure. From the resulting nucleoside mixture, adenosine and guanosine were separated, purified, and subjected to further hydrolysis to liberate D-ribose. Levene (1935) described the preparation of D-ribose by an acidic hydrolysis of ash-free guanosine. It was considered essential by Levene to work with a colorless hydrolysate if pure D-ribose was to be obtained. To hydrolyze yeast ribonucleic acid, Phelps (1939)replaced ammonia by magnesium hydroxide. The mixture was heated at 145°C for 4 hours. The magnesium-containing phosphates were subsequently removed by filtration, whereas guanosine immediately precipitated, and adenosine was recovered from the mother liquor as a picrate. An acidic hydrolysis of the picrate yielded D-ribose. Enzymatic hydrolysis of yeast RNA thereby using enzyme preparations from sweet almonds, lucerne seeds, and germinated peas, followed by acidic hydrolysis of the liberated guanosine and adenosine, also yielded D-ribose (Bredereck and Rothe, 1938; Bredereck et al., 1940). Another way to isolate D-ribose from yeast RNA was patented by Laufer and Charney (1945a). Their method facilitated the separation of nucleosides from the ribonucleic acid hydrolysate. The addition of cuprous ions resulted in precipitation of purine nucleosides as cuprous salts. Acidic hydrolysis of these salts led to D-ribose, next to insoluble cuprous salts of adenine and guanine. Alternatively, yeast RNA was hydrolyzed to a mixture of nucleotides that were converted into their cuprous salts with copper sulfate and sodium bisulfite. The pyrimidine nucleotides remained in solution, while the purine nucleotides precipitated. After separation, these purine nucleotides were subjected to hydrolysis with sulfuric acid. Insoluble cuprous salts of adenine and guanine were removed by filtration, leaving a solution that contained D-ribose, sulfuric acid, and phosphoric acid. The addition of an alkaline-earth hydroxide to this mixture, followed by filtration, gave a solution of practically pure D-ribose (Laufer and Charney, 1945b). Due to the development of resin technology and the production of high-quality ion-exchange materials, nucleoside and nucleotide mixtures, obtained by acid hydrolysis of yeast RNA, subsequently were separated by ion-exchange chromatography (Cohn, 1949).

MICROBIAL SYNTHESIS OF D-RIBOSE

177

Due to the difficulty involved in obtaining large amounts of yeast RNA, combined with the laborious hydrolysis of RNA and the isolation of D-ribose from it, more convenient ways to prepare D-ribose were looked for.

B. ENZYMATIC HYDROLYSIS OF YEASTRNA, RIBONUCLEOSIDES, AND RIBONUCLEOTIDES 1)-ribose(and some of its derivatives) can be prepared from RNA, and its related compounds by hydrolytic enzymes obtained from microor, The disganisms (Bredereck and Rothe, 1938; Bredereck et ~ l .1940). covery that RNA and DNA bear life’s genetic information triggered investigations of the mechanisms of gene expression during which several enzymes were discovered that were involved in the metabolic conversion of nucleic acids. These included nucleoside hydrolases, nucleotide N-glycoside hydrolases, nucleoside phosphorylases, and nucleotide pyrophosphorylases. As illustrated below, D-ribose, D-ribose-5phosphate, D-ribose-1-phosphate, and-5-phosphoribosyl-1-pyrophosphate (PRPP) can be produced from ribonucleosides (1) and ribonucleotides (2) by hydrolysis with ribonucleoside hydrolases and ribonucleotide N-glycoside hydrolases, by phosphorolysis of ribonucleosides with ribonucleoside phosphorylases (3), or by pyrophosphorolysis of ribonucleotides with ribonucleotide pyrophosphorylases (4): ribonucleoside hydrolase

ribonucleoside

+ HzO *

t

base + D-ribose,

(1)

ribonucleotide N-glycoside hydrolase

ribonucleotide

+ Hz0

base + D-ribose-5-P, (2)

4

ribonucleoside phosphorylase

ribonucleoside

+ H3P04

. .

base + D-ribose-1-P, (3)

4

ribonucleotide pyrophosphorylase ribonucleotide + HzP030P03Hz

4

base + PFPP.

(4)

In addition, investigations into the production of the flavor enhancers IMP and GMP by enzymatic degradation of yeast RNA started in Japan in the 1950s (Ogata, 1976). Their successful production on an industrial

178

P. DE WULF AND E. J. VANDAMME

scale led to large amounts of hydrolytic ribonucleotide by-products (uridine monophosphate and cytidine monophosphate), which were subsequently used to enzymatically (N-glycoside hydrolase) prepare D-ribose. D-ribose production from 5-amino-4-imidazole-carboxamide-riboside (AICAR) (an intermediate in inosine monophosphate synthesis) by hydrolysis with a Bacillus thiaminolyticus nucleoside hydrolase was reported by Sano et al. (1977). The microbial (enzymatic) procedure resulted in a molar D-ribose conversion yield of 65% (Sano et al., 1977).

c. CHEMICAL PRODUCTION OF D-RIBOSE When D-ribose was discovered (Kossel, 1891), Fischer and Piloty (1891) had already succeeded in chemically synthesizing L-ribose,

which was obtained by epimerization of L-arabonic acid to L-ribonic acid, followed by a reduction of its lactone. Blanksma and Alberda Van Ekenstein (1908) repeated the work of Fischer and Piloty, and obtained a viscous product that was contaminated with L-ribitol. The crude L-ribose was purified through its p-bromophenyl hydrazone, and crystalline L-ribose was obtained for the first time (Alberda Van Ekenstein and Blanksma, 1909). The same workers repeated this route of synthesis for the D-series. In the presence of pyridine, D-arabonic acid was converted into D-ribonic acid, which, by reduction with sodium amalgam and purification through its p-bromophenylhydrazone, yielded crystalline D-ribose (Alberda Van Ekenstein and Blanksma, 1913). This D-ribose production method was reinvestigated by Steiger in 1936. D-arabinose, prepared from calcium D-ghconate, was electrolytically oxidized to D-arabonic acid. The latter was then epimerized in boiling aqueous pyridine. D-ribonic acid was isolated as its calcium salt and was converted to D-ribonolactone. A controlled reduction of this lactone with 2.5% sodium amalgam gave D-ribose in an overall yield of 17%.

D-ribose was obtained from D-arabinoseby Gehrke and Aichner (1927) using a method that was basically developed by Bergmann and Schotte (1921). D-arabinose was converted via its tetraacetate and 2,3,4-tri-Oacetyl-D-arabino-pyranosyl-bromide into 3,4-di-O-acetyl-D-arabinal. After deacetylation, D-arabinal was hydroxylated with perbenzoic acid, yielding D-arabinose and D-ribose, with the latter predominating. D-ribose was purified via its crystalline benzylphenylhydrazone. Sowden (1950) introduced a method wherein D-ribose was obtained from D-glucose. More specifically, 4,6-O-benzylidene-D-ghcose was catalytically reduced to 4,6-O-benzylidene-glucitol, which, on oxidation

MICROBIAL SYNTHESIS OF D-RIBOSE

179

with sodium metaperiodate, yielded 2,4-O-benzylidene-~-erythrose. Condensation of the latter with nitromethane gave 3,s-O-benzylidene1-deoxy-1-nitro-D-arabitol and 3,5-O-benzylidene-l-deoxy-l-nitro-~-ribitol. Both were separated, based on their different solubility in chloroform, in which the latter was very soluble. Hydrolysis of this D-ribitol derivative yielded amorphous 1-deoxy-1-nitro-D-ribitol,which was decomposed into D-ribose. This process was important, since it could be used to prepare 14C,-labeled D-ribose. D-glucose was converted in D-ribose by Smith (1955) via a two-step process. First, D-ghCOSe was esterified to 3-O-mesyl-D-glucose, which was oxidized with periodate into 2-0-mesyl-4-0-formyl-D-arabinose. The latter was treated with alkali at room temperature, yielding pure D-ribose, which was isolated as its toluene-p-sulfonylhydrazon in an overall yield of 24%. The methods described by Sowden (1950) and Smith (1955) were used for more than 20 years to produce D-ribose on a large scale. Because of the complexity of these processes, coupled with the use of mercury, a toxic and environmental pollutant, used to catalyze the electrolytic amalgam reduction step, several research groups started to work out alternative D-ribose production routes, thereby using cheap and renewable raw materials (Dondoni et al., 1986; Taniguchi et al., 1974; Zamojski and Grynkiewicz, 1984). Unfortunately, these processes were characterized by low production efficiencies and too many conversion steps. In 1992, Lacourt-Gadras and colleagues succeeded in generalizing methods that were developed by Barrett et al. (1977, 1979), Bessodes et al. (1981),Garegg and Samuelsson (1979),and Schaub and Weiss (1958) to prepare D-ribose from D-xylose, a pentose recovered from hydrolyzed plant waste biomass. By treating D-xylose with methanol, followed by tritylation with triphenylmethane, methyl-5-O-trityl-~-xylofuranoside was formed. The latter was converted into methyl-2,3-di-O-5-O-trityla,P-~-glycero-pent-2-endofuranoside with methanosulfonyl chloride and N,N-dimethyl formamide. By oxidizing this compound in an aqueous acetone solution, followed by triphenylamide treatment at high temperature, pure D-ribose was obtained in an overall yield of 38%. In 1985, Hiroshi and coworkers patented the preparation of D-ribose by epimerization of D-arabinose in an ammonium molybdate-borate mixture (78OC, 1 h). The final D-arabinose/D-riboseratio could be up to 6/94, depending on the reaction solvent used. After cooling the suspension and concentrating it by distillation, the molybdate and boric ions were removed with anion-exchange resins. After ethanol-based precipitation, crystalline D-ribose was obtained.

180

P. DE WULF AND E. J. VANDAMME

Kubala and Burdatsova (1991) improved the method developed by . by mixing an aqueous solution of D-arabinose with Hiroshi et ~ l(1985) an anion-exchange resin (amberlite) substituted with quaternary benzyltrimethylammonia groups and molybdate ions (95"C, 5 h). The aromatic amine formed was used to prepare crystalline N-phenyl-D-ribosylamine, from which D-ribose was released by steam hydrolysis. The much simpler, quicker, and more efficient production of D-ribose by microbial fermentation processes, mainly developed by Takeda Company in the mid-seventies, ruled out large-scale preparation of D-ribose via organic chemistry. Despite this, short, efficient, and nontoxic chemical procedures, aimed at recuperating or preparing D-ribose from (industrial) bulk waste always may be useful. VII. Microbial Production of o-Ribose Microbial D-ribose production occurs with mutant strains deficient in transketolase (TKT, EC 2.2.1.1) and/or ~-ribulose-5-phosphate-3-epimerase (EC 5.1.3.1). Grown on, for example, D-glucose, these mutants partly convert this carbon substrate via the oxidative pentose phosphate pathway, leading to intracellular accumulation of ~-ribose-5-phosphate at the transketolase conversion point. To overcome feedback inhibition effects exerted by this intermediate, ~-ribose-Ei-phosphateis dephosphorylated (alkaline phosphatase) and secreted into the medium. As an introduction, the biochemical background of D-ribose synthesis will be briefly overviewed. A. D-RIBOSE-5-PHOSPHATE, AN INTERMEDIATE OF THE PENTOSE PHOSPHATE PATHWAY

The active investigations of Warburg, Dickens, Horecker, Lipmann, and Racker from the 1930s to the 1950s led to the elucidation of a biochemical pathway (the Warburg-Dickens-Horecker pathway, the hexose monophosphate cycle, the phosphogluconate route, the pentose phosphate cycle (PPC)) that is uniquely situated between the glycolysis and various biosynthetic cascades (Fig. 2). This cycle of hexose breakdown can be split into an oxidative and a nonoxidative part. In the oxidative reactions, hexoses are converted into pentoses, thereby releasing COz. In the nonoxidative branch, the enzyme transketolase and transaldolase interconvert heptoses, hexoses, pentoses, tetroses, and trioses. The importance of the PPC partly lies in the carbohydrate interconversions that provide essential metabolic precursors necessary for cel-

\\\

D-gbcore-6-P

D-glucore 6-P dehydrogenare

m

D-gluconic acid ADP

6-P-D-gluconolaotone

H'p

laclonrrf

6-P-D-gluoononate

dehydrogenare D-rlbulore 5-P

nucldo rcldr cornrynno ATP

D-ribulooe 6-P

3-eplmerare

D-glyoeraldehyde 3-P

blomarr

----

D-lr D-glyoeraldehyde 3-P

-

-

L

O

PEP

D-fruotore 1.6-blP

9bGObBlr

A

t

I

aromallc amino acid. aromatlo vllamlnr

I I

I pyrwale

D-aedohoptulore 7-P

----

Krebr cycle

----+ respiration

FIG.2. The pentose phosphate pathway, the glycolysis, and some anabolic reactions (+ = o-ribulose-5-phosphate-3-epirnerase deficiency, = transketolase deficiency). See text for explanation.

*

182

P.

DE

WULF AND

E. J. VANDAMME

lular build-up. From D-ribulose-5-phosphate,riboflavin is synthesized; ~4bose-5-phosphateis the precursor of nucleotides, nucleic acids, ATP, and coenzymes, while D-erythrose 4-phosphate is an intermediate in aromatic amino acid (L-tryptophane, L-tyrosine, and L-phenylalanine) and vitamin (coenzyme Q, vitamin K2, and folic acid) synthesis. The pentose phosphate cycle has a hand in the biosynthesis of C7-carbohydrates that occur in the cell wall lipopolysaccharide layer of Gram-negative bacteria (Eidels and Osborn, 1971; Woisetschlager and Hogenauer, 1986),and in the formation of D-ribitol, a component of ribitol teichoic acid that configures the cell wall of Gram-positive bacteria. The reducing equivalents (NADPH + H+) liberated during the oxidative conversion steps are used in the synthesis of monomeric (e.g., amino acids, steroids, fatty acids, squalene) and polymeric cellular building blocks (e.g., proteins, polynucleotides, exopolysaccharides) (Schlegel, 1984).

B. GOAL-ORIENTED CREATION OF TRANSKETOLASEDEFICIENT MICROORGANISMS By exploiting the biochemical versatility of the PPC, Josephson and Fraenkel (1969) were the first to isolate transketolase-affected E. coli strains during research on the regulation of the pentose phosphate pathway. The principle for creating and isolating transketolase mutants was based on the idea that organisms that lack transketolase activity cannot grow on pentoses and cannot convert the intracellularly formed ~-ribose-5-phosphateinto aromatic biomolecules (amino acids, vitamins). Their method consisted of a chemical (ethylmethylsulfonate) mutation of E.coli strain K10, followed by an enrichment to transketolase-negative mutants in minimal medium, free of pentoses and aromatic amino acids, but provided with penicillin G. Induced transketolase-deficient mutant strains could not grow in the minimal medium, while tkt-positive species did. These were subsequently killed by the antibiotic. The nongrowing tkt mutants were isolated from the enrichment medium by transferring samples to agar-based minimal medium, supplemented with pentoses and/or aromatic amino acids. From the 1000 E. coli mutant colonies that were isolated, only 3 displayed pentose negativity and aromatic amino acid auxotrophy. Even these E. coli mutants were somewhat leaky for (i.e., not completely blocked in) the biosynthesis of aromatic intermediates of the shikimate pathway (Josephson and Fraenkel, 1974). This was explained by the fact that the requirement for TKT activity in aromatic amino acid synthesis is much less than for pentose degradation, for which transketolase is a major catabolic enzyme. In 1992,

MICROBIAL SYNTHESIS OF D-RIBOSE

183

Draths et al. determined the level of TKT activity of an E. coli transketolase mutant, isolated by Josephson and Fraenkel (1969), and showed it to be 15% of that measured in its parental strain. By employing the mutation and enrichment method developed by Josephson and Fraenkel (1969), Eidels and Osborn (1971) isolated transketolase mutants of Salmonella typhimurium. Of the approximately 5000 enriched colonies tested, only 2 were pentose-negative and required aromatic amino acids for growth. These mutants also had very low residual (0 to 4%) TKT activity. In addition, the transketolase mutants, sensitive to bile salts, were deficient in the synthesis of Lglycero-D-mannoheptose, a major constituent of the outer-membrane lipopolysaccharide of S. typhimuriurn. The defect in lipopolysaccharide formation could only be repaired by adding ~-sedoheptulose-7phosphate to the growth medium of the mutant strains.

c. RANDOM ISOLATION OF D-RIBOSE-PRODUCINGMICROORGANISMS The accumulation of small amounts of D-ribose was first observed in the culture medium of a Penicillium brevicompactum strain during a survey on microbial synthesis of benzene derivatives (Godin, 1953; Simonart and Godin, 1951). Since D-ribose was detected by paper chromatography, its concentration is unknown. Twelve years later, an unidentified bacterium secreted D-ribose or ~-ribose+phosphate and 5-phosphoribosyl-1-pyrophosphate in a culture medium with 100 g/liter D-glucose as the carbon source (Suzuki et al., 1963). When the medium contained Mn2+,Fez+,or Zn2+ions, D-ribose accumulated (5.2 g/liter); when these were absent, ~-ribose-5-phoswere secreted (up to 7 phate and 5-phosphoribosyl-l-pyrophosphate g/liter). Among the 1395 microbial strains isolated from soil, Saito and Sugiyama (1966) found nine species that secreted D-ribose. Next to D-ribose (0.72 g/liter from 50 g/liter D-glucose), the screened Pseudomonas reptilivora S-1104 produced D-fructose and an unidentified glucoside. Of the 2120 species isolated from sugar-rich substrates, De Wulf et (11. (1996a)found four strains that secreted D-ribose (between 0.10 and 0.36 g/liter, starting from 30 g/liter D-glucose). These included a Staphylococcus hominis strain, making it the second non-Bacillus bacterium described to secrete D-ribose (the first was I! reptilivora S-1104; Saito and Sugiyama, 1966), two Bacillus spp. that could not be identified at the species level, and the yeast Candida pelliculosa. Interestingly, the latter was the second yeast strain reported to secrete D-ribose (the first was a C. tropicalis strain; Kyowa Hakko Kogyo Co., Japan; Pat. 51-

184

P.

DE

WULF AND E. J. VANDAMME

057884, 1976). To increase the D-ribose titer, the screened C. pelliculosa strain was UV mutated and the transketolase-deficient mutants enriched in nystatin-based minimal medium free of pentoses and aromatic amino acidshitamins (Josephson and Fraenkel, 1969). Consequently, C. pelliculosa strain PDW 43 was isolated, which formed 1.2 g/liter D-ribose, starting from 30 g/liter D-glucose. Next to D-ribose, 7 g/liter Darabitol was secreted in the culture medium. Subsequent UV mutation and enrichment to D-arabitol nonconsuming mutants led to the isolation of C. pelliculosa strain PDW 43 ARA-, which produced 5 g/liter D-ribose, starting from 30 g/liter D-glucose. This strain probably was eliminated in the enzyme D-arabitol dehydrogenase, which catalyzes the 1996a). uptake and secretion of D-arabitol (De Wulf et d.,

D.

CREATION OF INDUSTRIALLY APPLIED TRANSKETOLASE-DEFICIENT

D-RIBOSE-PRODUCING STRAINS In order to create an inosine-producing Bacillus mutant strain, the spore-forming Bacillus strain no. 102, isolated from soil, was submitted to UV mutation. By selecting for adenine- and xanthine-negative phenotypes, an excellent inosine-forming mutant strain (named G3-46-226) was isolated. By submitting its spores to UV irradiation, coupled to an enrichment as used by Josephson and Fraenkel(1969),mutants were sought to be derived that would more efficiently convert the intracellularly accumulated ~4bose-5-phosphateinto inosine-5-phosphate. The inosine productivity of the isolated mutants did not improve, however. In fact, several strains lost their ability to produce this nucleotide (Sasajima et al., 1970). Unexpectedly, 14 strains accumulated D-ribose and/or D-ribulose in the fermentation medium and were shown to lack transketolase and/or ~-ribulose-5-phosphate-3-epimerase activity. Three strains in particular, named Shi5, Shi7 (shikimic acid-requiring strains), and Gluc34 (D-gluconatenonutilizing mutant) produced up to 35 g/liter D-ribose, whereas the epimerase mutants synthesized an unspecified amount of D-ribose and D-ribulose in a 2:l ratio (Sasajima and Yoneda, 1971). As it turned out, the intracellularly stacked D-ribose-5phosphate had feedback-inhibited its own conversion into inosine-5phosphate (Sasajima and Yoneda, 1974a,b).As true revertant transketolase-positive strains did not overproduce and secrete D-ribose, it was concluded that D-ribose formation was due to a lack of TKT activity (Sasajima and Yoneda, 1971). That these and all other transketolase mutants were never found to secrete D-XylOSe possibly may be explained by the fact that D-ribose-5phosphate has a more critical role in cell metabolism. In order to cope

MICROBIAL SYNTHESIS OF D-RIBOSE

PTS

D-glucose dehydrogenase

D-gluco8e

-

\ cytopl6sma

6-P

185

gluconokina> f 3

ATP

ADP

phosphatase

cell membrane

FIG.3. Possible conversion routes of D-glucose into D-ribose. See text for explanation.

with its transketolase deficiency status (synthesis of ATP, nucleotides, RNA (DNA)), a tkt mutant strain may preferably convert the intracellulady accumulated D-xylose-5-phosphate (by epimerization and subsequent isomerization) into ~4bose-5-phosphatefor use in, for example, energy synthesis (Fig. 2). As such, only ~4bose-5-phosphatemay become accumulated, such that only D-ribose may be secreted by tkt mutants (note of the authors). By transforming the D-ribose-secreting adenine- and xanthine-negative mutant strain Bacillus Shi7 (also named BG 1107) with chromosomal DNA from the D-ribose nonproducing and adenine- and xanthinepositive B. subtilis IF0 3026, a D-ribose superproducing adenine- and xanthine-positive transformant strain (named BG 1722) was selected, which secreted 60 g/liter D-ribose (Yokota and Sasajima, 1981; Yokota et al., 1979). Besides, this mutant constitutively synthesized D-glucose dehydrogenase, an enzyme that is normally induced during spore formation (Warren, 1968). Since the regulation of this enzyme seemed to be altered, this strain very efficiently converted D-glucose into D-ribose5-phosphate via two biochemical branches (Sasajima and Yoneda, 1989): first via the classical oxidative PPC, and second via a catalytic by-pass in which D-glucose dehydrogenase and D-gluconokinase are involved (Fig. 3).

186

P.

DE

WULF AND E. J. VANDAMME

Next to D-ribose, D-gluconic acid also accumulated in the fermentation medium. The secretion ratio of both metabolites depended on the partial oxygen pressure during the fermentation; oxygen depletion led to accumulation of D-ghconate rather than D-ribose (Sasajima and Yoneda, 1989). Under highly oxidative conditions, the strain secreted about 60 g/liter of D-ribose (Sasajima, 1976; Sasajima et al., 1972; Sasajima and Yoneda, 1984). Additional mutagenesis to induce asporogenous phenotypes of the tkt-affected mutants (Shi5, Shi7, and Gluc34) further enhanced the D-ribose yield to 70 g/liter (Sasajima, 1976; Sasajima et al., 1972; Sasajima and Yoneda, 1984). The basis of this causal relationship remained obscure until Rhaese and Groscurth (1976, 1979) showed that a high concentration of phosphorylated nucleotides initiates spore formation, even in the presence of excessive amounts of carbon, nitrogen, and phosphorous sources. Since D-ribose-leaking organisms are not able to build up a sufficient amount of phosphorylated nucleotides, the strains may remain asporogenous. The isolation of transketolase and/or ~4bulose-5-phosphate-3-epimerase mutants of such different bacterial strains as Bacillus subtilis, B. pumilus, Brevibacterium thiogenitalis, B. ammoniagenes, Arthrobacter globiformis, Aerobacter aerogenes, and Micrococcus denitrificans showed that only transketolase-deficient Bacillus spp. secrete D-ribose (Sasajima et al., 1985).In contrast, random screening led to the isolation of D-ribose-secreting non-Bacillus strains, such as Pseudomonas reptilivora (Saito and Sugiyama, 1966) and Staphylococcus hominis (De Wulf et al., 1996a),next to Penicillium brevicornpactum (Simonart and Godin, 1951), Candida tropicalis (Kyowa Hakko Kogyo Co., 19761, and C. pelliculosa (De Wulf et al., 1996a). Most D-ribose-overproducing Bacillus strains used in industry have been obtained via a two-step mutational approach based on irradiational or chemical treatment of a parent strain isolated from soil or obtained in any other way. The transketolase and/or D-ribulose-5-phosphate-3-epimerase-negative mutants isolated by selective enrichment were then mutated to induce an asporogenous phenotype. Moreover, during both postmutational selections, mutants with an enhanced 2-deoxy-D-glucose oxidase activity were sought. It is unclear why this characteristic guarantees high D-ribose productivity. Five D-ribose-overproducing B. subtilis strains-namely, IF0 13565, IF0 13586, IF0 13585, IF0 13621, and IF0 13573-were derived hom the D-ribose-nonproducing B. subtilis IF0 3026 by UV or nitrosoguanidine treatment via the two-step procedure mentioned above.

MICROBIAL SYNTHESIS OF IJ-RIBOSE

187

The D-ribose-producing bacteria used in industrial fermentation processes include: B. subtilis No. 429 (IF0 12603, ATCC 21359) *1 B. subtilis No. 483 (IF0 12604, ATCC 21360) *1 B. subtilis No. 608 (IF0 13323, FERM P-1490, ATCC 21952) *2 B. subtilis No. 957 (IF0 13565, FERM P-2259, ATCC 31096) *3 B. subtilis No. 941 (IF0 13573, FERM P-2360, ATCC 31097) *3 B. subtilis No. 1054 (IF0 13586, FERM P-2467, ATCC 31091) *3 B. subtilis No. 1067 (IF0 13588, FERM P-2468, ATCC 31092) *3 B. subtilis No. 1097 (IF0 13621, F E W P-2833, ATCC 31094) *3,4 B. subtilis TK 103 (IF0 15138, FERM BP-3290) *5 B. pumilus No. 716 (B. pumilus IF0 13322, B. subtilis FERM BP-812, B. subtilis ATCC 21951) *2 B. pumilus No. 503 (IF0 12600, ATCC 21356) *1 B. pumilus No. 537 (IF0 12601, ATCC 21357) *1 B. pumilus No. 558 (IF0 12602, ATCC 21358) *1 B. pumilus No. 911 (IF0 13566, FERM P-2260, ATCC 31095) *3 B. pumilus No. 1027 (IF0 13585, FERM P-2466, ATCC 31098) $3 B. pumilus No. 1083 (IF0 13620, FERM P-2832, ATCC 31093) *3,4 *1 = Jpn. Pat. *2 = Jpn. Pat. *3 = Jpn. Pat. *4 = Jpn. Pat. *5 = Jpn. Pat.

47-7948/1972, U.S. Pat. 3,607,648 50-1687811975, U.S. Pat. 3,919,046 51-7753/1976, Jpn. Pat. 52-1993/1977, U.S. Pat. 3,970,522 58-17591/1983 36129/1992, Eur. Pat. 0-501-765-A1

The industrially applied D-ribose-producingtkt-negative B. pumilus IF0 12600, B. pumilus IF0 12601, B. subtilis IF0 12603, and B. subtilis IF0 13323 were completely transketolase-deficient. Their parental strains had a transketolase activity of, respectively, 0.13 (B. pumilus) and 0.24 (B. subtilis) pmol/min mg transketolase (Sasajima and Yoneda, 1974~).

Since transketolase is only functional when its coenzyme thiamine pyrophosphate and Mg2+are coexistent (Wood, 1985), it was thought that thiamine-requiring mutants also would accumulate D-ribose. The amount of D-ribose produced by mutant strains isolated as such was much lower than that secreted by aromatic amino acid auxotrophic mutants (Sasajima and Yoneda, 1984; Sasajima et al., 1985). As several enzymes other than transketolase also use thiamine as a coenzyme, thiamine auxotrophy cannot be used to selectively isolate D-ribose-overproducing transketolase mutants.

188

P.

DE

WULF AND E. J. VANDAMME

E. IMPROVEMENT OF D-RIBOSE PRODUCTIVITY BY RECOMBINANTDNA TECHNOLOGY

Several D-ribose-producing bacilli secrete D-gluconate in a culture medium (Asai et al., 1978; Kintaka et al., 1986; Kishimoto et al., 1990; Miyagawa et al., 1992; Sasajima and Yoneda, 1989), thereby decreasing the pentose yield and disturbing D-ribose recovery. Based on the fact that D-ribose possibly may be formed via a gluconate bypass in tkt-affected Bacillus spp. (Yokota et al., 1979; Yokota and Sasajima, 1981) (Fig. 3), Miyagawa et al. (1992) developed a strategy to increase the expression level of the bacterial gluconate operon. This way, the cosecreted D-gluconic acid can be taken up and reconverted into D-ribose. The microbial gluconate (gnt) operon generally consists of four open reading frames: gntR, gntK, gntP, and gntZ (Fujita et a]., 1986). D-ghconate permease, encoded for by gntP, permeates D-gluconate into the cell, while D-gluconokinase, encoded for by gntK, catalyzes C6-phosphorylation of D-gluconate. D-gluconate-6-phosphate is subsequently oxidized and decarboxylated by ~-gluconate-6-phosphatedehydrogenase (gntZ gene product), thereby yielding ~-ribulose-5-phosphate, the direct precursor of ~-ribose-5-phosphate.The state of expression of both enzymes is negatively controlled by a repressor protein (encoded for by gntR) (Fujita and Fujita, 1987). Miyagawa et al. (1992) isolated the gnt operon from chromosomal DNA, extracted from a Bacillus sp. Before the gnt fragment was cloned in an expression plasmid, the gntR region was deleted and/or the promotor of this gluconate operon replaced with a highly active promotor sequence that triggers gnt expression in the accepting Bacillus strain (Miyagawa et al., 1992). The recombinant plasmid was transformed to the D-ribose- and D-gluconate-producingB. subtilis IF0 15138, resulting in a significantly enhanced D-ribose productivity of the strain (from 39 to 62 g/liter D-ribose, starting from 160 g/liter D-glucose; Miyagawa et a]., 1992). VIII. Pleiotropic Properties of D-Ribose-Producing Transketolase-Defective Bacillus Mutant spp.

As mentioned before, tkt mutants cannot utilize carbon sources that are catabolized via oxidative PPC. Moreover, transketolase-affected strains require an aromatic amino acid complement in the medium, since they are unable to autonomously synthesize these metabolites.

MICROBIAL SYNTHESIS OF D-RIBOSE

189

Interestingly, tkt mutants also have a pleiotropically changed physiology and morphology. A. DEFECTIVEPHOSPHOENOLPYRWATE-DEPENDENT PHOSPHOTRANSFERASE SYSTEM The inability to utilize D-glucose was noticed during a survey on the nutritional properties of tkt-affected bacteria (Sasajima and Kumada, 1979; Sasajima et al., 1977). The nonutilization O f D-glUCOSe by transketolase mutant B. subtilis BG 2607 was caused by a defective D-glucose transport function of the phosphoenolpyruvate (PEP)-dependent phosphotransferase system (PTS). In the parent strain, the uptake and phosphorylation of D-glucose and D-sorbitol was induced by the sugars themselves. In the tkt mutant, neither D-glucose nor D-sorbitol activated the transport function of the PTS, though both induced its phosphorylation activity (Sasajima and Kumada, 1979). On the other hand, D-mannitol and D-fructose, which are also phosphorylated and transported by the PEP-dependent PTS in B. subtilis (Marquet et al., 1970; Gay and Delobbe, 1977), were metabolized by the tkt mutant. Consequently, inactivation of the components of which the transport system consistsnamely, enzymes I, HPr, and PEP- could not be the cause of the transport deficiency. It appeared that a change had occurred in the membrane-bound enzyme I1 complex, which catalyzes the transportation and phosphorylation of carbohydrates. How tkt defectivity renders the PTS system to be inactive for D-glucose is unknown (Sasajima and Kumada, 1979). Since mutants were isolated that took up D-ghCOSe though they remained tkt-negative implies that this enzyme is indirectly related to the membrane function/structure through an unknown mechanism (Sasajima et al., 1977). B. DEREGULATED CAR~OHYDRATE CATABOLITE REPRESSION

The synthesis of enzymes that participate in o-mannitol catabolism, such as D-mannitol PEP-dependent PTS and D-mannitol-1-phosphate dehydrogenase, was hypersensitive to repression by D-glucose, D-gluconic acid, D-xylose, and L-arabinose in the tkt mutant strain B. subtilis BG 2607 (Sasajima and Kumada, 1981a). In contrast, the synthesis of enzymes that catalyze the conversion of D-sorbitol (D-sorbitol permease, D-sorbitol dehydrogenase) and glycerol (glycerol permease, glycerol kinase, and glycerol dehydrogenase) was insensitive to D-glucose repression (Sasajima and Kumada, 1981a). The changed regulation in enzyme synthesis seemed to be related to a defect in the cell surface

190

P.

DE

WULF AND E. J. V A N D M

structure of the tkt mutant, thereby generating several pleiotropic properties. Some component essential for the maintenance and functionality of the cell wall may not have been synthesized in the tkt-affected mutant strain. Hence, it is possible that the acquired tkt deficiency disturbed synthesis of the cell wall component ribitol teichoic acid, which is derived from pentose phosphates (Sasajima and Kumada, 1981b). When B. subtilis ATCC 21951 was grown in a medium containing D-glucose, D-fructose, or D-mannitol, plus D-gluconic acid, D-xylose, L-arabinose, D-xylitol, D-galactose, or D-glUCurOniC acid, both substrates were converted simultaneously (De Wulf, 1995; De Wulf et ~ l .1996b). , Carbon catabolite repression was significantly reduced (D-fructose and D-mannitol) or completely absent (D-glUCOSe).As illustrated by cultivating B. subtilis-type strain ATCC 6051 in the same double substrate , D-fructose-l,6-biphosmedia (De Wulf, 1995; De Wulf et ~ l .1996b), phate (glycolytically derived from D-glucose, D-fructose, or mannitol), normally triggers a reaction that ends in the formation of a phospho~erine~~-HPr-Ccpa protein complex. The latter then binds to a catabolite-responsive element localized inside the operon that encodes utili, Saier et ~ l .1996). , zation of the second carbon source (Fujita et ~ l .1995; Consequently, the cosubstrate is not utilized, and diauxic growth occurs. Since D-mannitol and D-frUCtOSe were more repressive than D-ghcose, and since both yielded lower D-ribose titers than D-glucose did when used as a single carbon source (De Wulf, 1995), it was proposed that a more active conversion O f D-glucose via the oxidative PPC had led to a lower intracellular amount of ~-fructose-l,6-biphosphate.Consequently, the critical threshold value needed to trigger catabolite repression may not have been reached when D-glucose was present in the double substrate medium (De Wulf et al., 1996b).

c. ALTERED CELL MEMBRANE AND CELL WALL COMPOSITION During exponential growth, tkt-affected bacilli appear as thick cells in a chainwise configuration, whereas the parent strains occur as slender and distinct rods (Sasajima and Kumada, 1981b). These changes suggested an altered cell wall synthesis. Chemical analysis of the cell envelope revealed that tkt-affected strains were defective in teichoic acid and teichuronic acid formation, or in the synthesis of lipoteichoic acids (Sasajima and Kumada, 1981b). The mutants were also very sensitive to phages SPOl and SPlO (Sasajima and Kumada, 1981b). Since tkt mutants undergo lysis at a slower rate, relative to their parent strains, and as they are nonmotile and deficient in flagellation (Sasajima and Kumada, 1981b, 1983a,b),an altered configuration of the

MICROBIAL SYNTHESIS OF D-RIBOSE

191

cytoplasmic membrane also was proposed (Sasajima et QL, 1985). Rhaese et d . (1977) reported the role of adenosine 3’(2’)-triphosphate5’-triphosphate in the initiation of sporogenesis. As these phosphorylated nucleotides are synthesized by membrane-bound enzymes (Rhaese and Groscurth, 1976; Rhaese et af., 1977), regulation of the sporulation process may be disturbed, possibly due to an altered cell membrane consistency. The isolation and morphological characterization of a truly revertant tkt-positive species that did not generate these phenotypic alterations confirmed the causal relationship of these phenomena with a mutation in the tkt locus (Sasajima and Kumada, 1979, 1983a,b). SDS-polyacrylamide gel electrophoretic patterns of the membrane proteins of a tkt mutant and its parent strain revealed that some new proteins had originated in the mutant. Moreover, the molecular weight of several membrane proteins differed from those in the wild-type strain , A thin-layer chromatographic examination of the (Sasajima et ~ l .1985). membrane lipids also showed an increased concentration of phosphatidylglycerol and lysylphosphatidylglycerol, whereas the quantity of phosphatidylethanolamine was lower in the tkt mutant (Sasajima and Yoneda, 1989). Tkt-altered strains of such Gram-negativebacteria as Escherichia coli and Salmonella typhimurium could not synthesize L-glycero-D-mannoheptose (Eidels and Osborn, 1971), a backbone component of the cell wall lipopolysaccharide layer (Smit et al., 1975; van Alphen et Qf., 1976). It is interesting that transketolase mutants of Gram-negative and Gram-positive bacteria are defective in cellular surface functions, although the basic mechanisms may differ. In conclusion, the enzyme transketolase is of fundamental physiological importance, not only for catalyzing carbohydrate interconversions and for supplying intermediate compounds in the synthesis of aromatic biomolecules, but also for regulating carbohydrate catabolism and for maintaining the bacterial membrane and cell wall structure. IX. D-Ribose Production by Fermentation with Bacillus spp.

Most information on microbial D-ribose production is found in the patent literature (Table I). The patents, all of Japanese origin (Takeda Co.), disclose batchwise D-ribose production and apply similar downstream processes to recover this pentose (Kintaka et d.,1986; Kishimoto et al., 1990; Miyagawa et al., 1992; Sasajima and Yoneda, 1974c; Sasajima et al., 1975; Yoneda and Sasajima, 1969).

192

P.

DE

WULF AND E. J. V A N D M TABLE I

OVERVIEW OF THE INDUSTRIALLY A p p L m D-RIBOSEPRODUCNG BACILLUS MUTANTS,THE OMPOSITION OF THE FERMENTATION MEIJIA (g/LJTER), THE FERMENTATION TIME(HOURS), AND THE OBTAINED D-RIBOSEYIELDSLITER) Sasalima & Yoneda 11971)

Yoneda and Sasajima (1969)

D-ghCOSe D-gluconic acid Soluble starch CSL Dried yeast (N&)zS04 MnS04 CaC03 Ca3(PO4I2 CaHPO, L-tryptophane L-phenylalanine L-tyrosine L-maleic acid m-leucine L-valine a-aminobutyric acid Shikimic acid Time of fermentation D-ribose

B. pumilus

B. pumilus

B. pumilus

ATCC 21356

ATCC 21357

ATCC 21358

125

125

B. subtilis ATCC 21359

8. subtilis

B. subtilis

ATCC 21360

ATCC 21951

100

100

15 15

20 18

25 15

10 5 5

10

10 5 5

125 100

25 15

25 15

25

10 5 5

10 5 5

10 5

15

5

5 5

0.1

0.15 0.05

66 29

55 31

80 29

72 27

68 30

144 35

The patents claim the use of a broad range of D-ribose producing mutant strains (Table I), developed by chemical or UV-induced mutagenesis (Kintaka et d., 1986; Kishimoto et al., 1990; Sasajima and Yoneda, 1974c; Sasajima et al., 1975; Yoneda and Sasajima, 1969), possibly improved by recombinant DNA technology (Miyagawa et al., 1992). Important fermentational parameters, such as the process temperature, the agitation speed, the aeration rate, the composition, and the pH of the medium, are only vaguely described in the patents. This and the fact that the most diverse mutant strains are used makes it hard to compare the different production methods (TableI) and to draw general conclusions from them.

193

MICROBIAL SYNTHESIS OF D-RIBOSE

TABLE I, continued Sasajima Sasajima and Yoneda (1974)

Bacillus strain

D-glucose o-gluconic acid Soluble starch CSL Dried yeast (Nb)’2S04 MnS04 CaC03 CaAPO4)2 CaHP04 L-tryptophane L-phenylalanine L-tyrosine L-maleic acid DL-leucine L-valine a-aminobutyric acid Shikimic acid Time of fermentation D-ribose

B. subtilis ATCC 21951

et a/.(1975)

B. subtilis

B. pumilus

B. subtilis

ATCC 21952

ATCC 31091

A T E 31092

150

150

150

150

150

150

10

10

5

5

12 3

20 5

10 5

10 5

20

20

20

20

20

20

0.10

0.05 0.05 0.05 0.50

0.05 0.10 0.05

0.05 0.05 0.05

0.05 0.05 0.05

60 66

60 67

72 45

0.01 90 60

0.75

a4 46

72 56

A. CULTIVATION OF THE INOCULUM

All patents describe a Bacillus inoculation medium that consists of D-sorbitol (20 g/liter), corn steep liquor (CSL, 20 g/liter), K2HP04 (3 g/liter), and KH2P04(10 g/liter) (Kintaka et al., 1986; Miyagawa et a]., 1992; Sasajima et al., 1975; Yoneda and Sasajima, 1969). Sasajima and Yoneda (1971) only once replaced D-sorbitol by soluble starch (10 g/liter). This inoculation medium could be supplemented with L-tryptophane, L-phenylalanine, and L-tyrosine (each 0.1 g/liter) (Kintaka et al., 1986; Sasajima and Yoneda, 1975). The inoculum is prepared in Erlenmeyer flasks (100- or 200-ml medium) or in small batch fermentors (1to 10 liter scale). The volumetric ratio of the inoculation and fermentation media varies between 1.7 and 10% (Asai et al., 1978; Kintaka et al., 1986; Kishimoto et al., 1990; Sasajima and Yoneda, 1971; Sasajima and Yoneda, 1974c; Sasajima et al., 1975; Yoneda and Sasajima, 1969).

194

P.

DE

WULF AND E. J. VANDAMME TABLE I, continued Sasajima et al. (1975)

Bacillus strain

D-glUCOSe D-gluconic acid Soluble starch CSL Dried yeast (Nh)zS04 MnS04 CaC03 CaAP0412 CaHPO4 L-tryptophane L-phenylalanine L-tyrosine L-maleic acid DL-leUCine L-valine a-aminobutyric acid Shikimic acid Time of fermentation o-ribose

B. subtilis

B. pumilus

ATCC 31093

ATCC 31094

150

150

150

150

150

150

10 5

10 5

10 5

10 5

10 5

10 5

20

20

20

20

20

20

0.05 0.05 0.05

0.05 0.05 0.05

0.05 0.05 0.05

0.05 0.05 0.05

0.05

0.05

0.05 0.05

0.05

60 72

60 71

60 64

60 62

60 65

60 63

B. subtilis

B. subtilis

ATCC 31095 ATCC 31096

B. pumilus

B.

subtilis

ATCC 31097 ATCC 31098

0.05

Although the process temperature (18 to 45°C) and the (initial) pH of the medium (4.5 to 9.0) are mentioned not to be critical, some patents suggest a temperature of 37°C and a pH between 5.5 and 8.0 (Kintaka et al., 1986; Miyagawa et al., 1992; Sasajima et al., 1974c, 1975; Yoneda and Sasajima, 1969). The agitation rate is not mentioned in any patent. Yet, this physical parameter determines the oxygen transfer to the cells, which may affect their D-ribose productivity. Interestingly, the D-ribose production capacity of the B. subtilis ATCC 21951 inoculum was maximal in the postexponential growth phase (2-3 x lo9 colony forming units (CFU)/ml) (De Wulf, 1995). Relative to that obtained with an exponentially grown inoculum (3 x lo8 to 2 x lo9 CFU/ml), the D-ribose productivity was three times as high (De Wulf, 1995). Noteworthy is that bacterial morphology increased during growth. More specifically, the cells occurred in a long-stretched (3-4 pm) chainwise configuration at the end of the

195

MICROBIAL SYNTHESIS OF D-RIBOSE

TABLE I, continued ~~

Asai et a / .

Kishimoto et al. (1990)

Kintaka et a / . (1986)

(1978)

Bacillus Bacillus strain

B. subtilk ATCC 21951

SP.

EMP-58

D-glucose D-ghlConiC acid Soluble starch CSL Dried yeast (N&)zS04

MnS04 CaC03 CadP0& CaHPO4 L-tryptophane L-phenylalanine L-tyrosine L-maleic acid DL-leucine L-valine a-aminobutyric acid Shikimic acid Time of fermentation D-ribose

140

B. subtilis ATCC 21951

180

200

180

200

220

20

22

22

26"

26"

5

5

5

7 0.05

20

16

5 0.50 16

16

20

5 0.05 20

0.10 0.10 0.10

0.05

0.05

0.05

0.03*

0.03'

0.10b

0.10b

55 92

91

10

0.50 0.25 0.25 55 64

72

72

72

80

91

78

72

active growth phase. By tracking the evolution in cell morphology of the inoculum culture, the best moment to start the fermentation could be determined easily (De Wulf, 1995). B . COMPOSITION OF THE FERMENTATION MEDIUM

Several carbon sources can be used to produce D-ribose. D-glUCOSe, D-fructose, D-mannitol, D-sorbitol, mannose, maltose, lactose, sucrose, glycerol, dextrine, soluble starch, hydrolyzed starch, molasses, acetic acid, D-ghconic acid, and ethanol have all been mentioned to be excellent substrates, ensuring high levels of D-ribose (Asai et d.,1978; De Wulf, 1995; De Wulf et a]., 1996b; Kintaka et d.,1986; Kishimoto et a]., 1990; Sasajima and Yoneda, 1971; Sasajima and Yoneda, 1974c; Sasajima et a]., 1975; Yoneda and Sasajima, 1969) (Table 1).

196

P.

DE

WULF AND E. J. V A N D A M

TABU I, continued Kishimoto et al.

Miyagawa et

(1990)

a/. (1992)

B. Bacillus strain

D-glucose D-glUcOniC acid Soluble starch CSL Dried yeast (NfhhS04 MnS04 CaC03 CadP0412 CaHPO4 L-tryptophane L-phenylalanine L-tyrosine L-maleic acid IJL-leUCine L-valine a-aminobutyric acid Shikimic acid Time of fermentation D-ribose

De Wulf et al. (1997)

subtilis

B. subfilis

subtilis

ATCC 21951

I F 0 15138

I F 0 15139

B.

200

160

160

200

100 100

100 50

26a

20

20

20

20

20

5 0.05 20

5

5

5

5

5

15

15

0.03'

0.05

0.05

0.25

0.25

72 39

72 62

156 40'

110 60d

84 45d

8.subtilis

ATCC 21951

0.10b

72 95

The processes are assumed to occur at free pH if not mentioned otherwise. OCSL free of L-tryptophane. ?he aromatic amino acid concentration is controlled at a constant value. "Process performed at controlled pH 6.0. dProcess performed at controlled pH 6.5. See text for explanations.

A fermentation medium composed of (NH,),SO,, CaHPO,, Ca3(P0,),, CaC03, and dried yeast (Sasajima and Yoneda, 1989; the component concentrations were not mentioned) and complemented with D-glUCOSe (125 or 150 g/liter), resulted in a D-ribose yield of 24 g/liter (Sasajima and Yoneda, 1989). With glycerol as the substrate, only 3.4 g/liter D-ribose was synthesized. Applying the glucose-based broth described above (Sasajima and Yoneda, 1989), but supplemented with yeast extract (15 g/liter) instead of dried yeast led to 30 g/liter D-ribose. When

MICROBIAL SYNTHESIS OF D-RIBOSE

197

CSL (25 glliter) was used, only 9.3 g/liter D-ribose was obtained (Sasajima and Yoneda, 1989). Nevertheless, Kintaka et al. (1986) and Kishimoto et (11. (1990) achieved very high D-ribose titers (90 glliter) with CSL as the main nitrogen source (Table I). The presence of other nutritional components, combined with optimized process conditions, may have stimulated the D-ribose productivity of the strain. It must be mentioned, though, that the data obtained by Kintaka et al. (1986) and Kishimoto et al. (1990) could not be reproduced. Even by growing the same mutant strain in the same medium under identical process conditions, D-ribose yields of maximally 20 g/liter were obtained (De Wulf, 1995). All kinds of nitrogen sources, including dried yeast, yeast extract, peptone, corn steep liquor, meat extract, fish meal, casein hydrolysate, amino acid mixtures, and ureum, and inorganic nitrogen sources, such as ammonia, ammonium sulfate, and ammonium nitrate, may be present in the fermentation broth (Asai et al., 1978; De Wulf, 1995; De Wulf et al., 1996b; Kintaka et al., 1986; Kishimoto et al., 1990; Sasajima and Yoneda, 1971; Sasajima and Yoneda, 1974c; Sasajima et al., 1975; Yoneda and Sasajima, 1969) (Table I). According to Sasajima (1976), the use of CSL or an anorganic nitrogen source such as (NH4)$04 requires the supplementation of shikimic acid (or its derivatives) or a mixture of aromatic amino acids (or their derivatives) to allow growth of the transketolase-negative D-ribose producing mutant strains (Table I). Suzuki et al. (1963) mentioned that Mn2+,Fez+,and Zn2+stimulated D-ribose production with an unidentified microorganism. D-ribose-5phosphate and 5-phosphoribosyl-1-pyrophosphate(PRF'P) were secreted in the absence of these cations. Although its mechanism is not known, this phenomenon remains interesting. As mentioned before, Asai et al. (1978), Kintaka et al. (1986), Kishimot0 et al. (19901, and Miyagawa et al. (1992) described the cosecretion of D-gluconic acid (up to 40 g/liter hom 140 g/liter D-glucose; Asai et al., 1978) during the fermentation process. Sasajima and Yoneda ( 1 9 7 4 ~ ) provided the medium with dicarboxylic organic acids (e.g., maleic acid, oxalic acid, glutaric acid) to limit the side formation of gluconic acid. Besides, the D-ribose yield increased with 40%. Kintaka et al. (1986) supplemented the fermentation medium with a compound that correR was H or a C1-C4 sponded to the molecular structure R-X-COOH. alkyl group substituted with 1 or 2 groups such as -OH, -SH, and -SCH3, while X could be C=O or CHNH,. The molecular configuration thus includes a broad range of organic acids and amino acids (Kintaka et al., 1986). The amount of D-gluconic acid formed was drastically

198

P.

DE

WULF AND E. J. VANDAMME

reduced in their presence (up to 96%) (Kintaka et al., 1986). Both observations (Kintaka et al., 1986; Sasajima and Yoneda, 1974c) were reinvestigated by De Wulf (1995) and turned out to be restricted to a very few organic acids. Sorbic acid (2 g/liter) enhanced D-ribose productivity by a factor of 1.5, while, in contrast, formic acid drastically decreased the D-ribose yield (De Wulf, 1995). Benzoic acid or glutaric acid did not influence bacterial metabolism (De Wulf, 1995). The pKa value of the acid possibly must be high enough for the acid to occur in the undissociated state to allow its diffusion into the cells. Also, the acid’s structural configuration and its (a)polar character may be important in this respect. How the acid stimulates D-ribose synthesis has been hypothesized based on the intracellular accumulation of acid-derived carboxylic protons. These may indirectly trigger ATP synthesis, since ATP is used to pump out the excessive carboxylic protons via the ATP-dependent ATPase. Since ATP is formed from ~-ribose-5-phosphate, D-glUCOSe conversion via the oxidative PPC may be stimulated, leading to higher ~-ribose-5-phosphateconcentrations at the TKT conversion point. As such, an increased amount of D-ribose may become secreted, relative to a blank (De Wulf, 1995). Noteworthy in this respect is that the use of D-ghconic acid as a carbon substrate (see later discussion), and the significant increase in D-ribose productivity obtained with it may be somewhat attributed to an acid-based effect, next to a much more efficient carbon flow through the oxidative PPC. CONDITIONS C. FERMENTATION

The industrial production of D-ribose is performed batchwise. The synthesis of D-ribose via continuous culture, fed-batch techniques, or immobilized cell systems has not been published. The physical characteristics of the fermentation processes (e.g., agitation and aeration rate) have been reported only vaguely. Since D-ribose is oxidatively formed from D-glucose, an aerobic (1w m + 1 atm; Asai et al., 1978), submerged cultivation of the mutant strain is preferred (Asai et al., 1978; Kintaka et al., 1986). This was examined by installing a series of varying aeration and agitation conditions at a 2-liter fermentor level (De Wulf, 1995). The more oxidative the process was performed, the faster the substrate (D-glucose) was utilized, the lower the amount of glycolytic side-products that were secreted, and the higher the final D-ribose yield (De Wulf, 1995). Using 200 g/liter D-glucose, an agitation of 700 rpm, and an aeration of 1.7 vvm at pH 6.0, 27 g/liter D-ribose was obtained with B. subtilis ATCC 21951. By increasing the

MICROBIAL SYNTHESIS OF D-RIBOSE

199

agitation speed to 1000 rpm and the aeration rate to 3 vvm, the D-ribose yield increased to 40 g/liter (De Wulf et al., 1997). The optimal temperature for microbial D-ribose synthesis lies between 32 (Yoneda and Sasajima, 1969) and 40°C (Kishimoto et al., 1990). Most authors mention the best temperature to be around 37°C (Asai et al., 1978; Kintaka et al., 1986; Kishimoto et a]., 1990; Sasajima and Yoneda, 1971; Sasajima and Yoneda, 1974c; Sasajima et al., 1975; Yoneda and Sasajima, 1969). The optimal pH of the fermentation medium is not defined in any patent. This parameter is stated to be of little importance, and a broad range of pH values (5.5 to 8.0) should lead to good D-ribose yields. Most fermentation processes probably are performed in a pH-free mode. Starting from pH 7.5 (Asai and Kono, 1984), pH 7.0 (Asai et al., 1978), or pH 6.0 (Sasajima and Yoneda, 1989), a minimal pH around 5.3 is obtained after approximately 24 hours of fermentation, after which the pH increases due to cell lysis (Asai et al., 1978; Sasajima and Yoneda, 1989). By performing a series of pH-controlled 2-liter batch fermentations with B. subtilis ATCC 21951, bacterial D-ribose productivity was shown to increase with decreasing medium pH (De Wulf et al., 1997). Neutral pH values actively stimulated biomass synthesis and glycolytic D-glucose conversion, inducing suboptimal D-ribose titers and a high concentration of unwanted glycolytic end-products (acetoin and 2,3-butanediol; De Wulf et al., 1997). A pH around 6.0 was most optimal, especially since lower pH values slowed down bacterial growth (De Wulf et a]., 1997). The fermentations take 50 to 1 2 0 hours, and the secretion of D-ribose starts after 1 2 to 24 hours of fermentation, both depending on the initial concentration of D-glucose (Asai et a]., 1978; Kintaka et a]., 1986; Kishimoto et al., 1990; Sasajima and Yoneda, 1971; Sasajima and Yoneda, 1974c; Sasajima et al., 1975; Yoneda and Sasajima, 1969). D-ribose is secreted during active growth and thus is a primary metabolite. The fact that transketolase deficiency induces a decreased D-glucose uptake rate may make D-ribose production a long process when high D-glucose concentrations are used (e.g., 200 g/liter, Fig. 4). By replacing D-glucose with another carbohydrate (e.g.,D-fructose, D-sorbitol,D-mannitol), this problem can be overcome, though the D-ribose yield significantly drops (De Wulf, 1995). A successful alternative consisted of partly replacing D-glucose with a PPC-converted cosubstrate, such as D-gluconic acid or D-xylose (De Wulf et al., 1996b). A significantly higher substrate conversion rate was obtained and the D-ribose productivity significantly increased, since overall cell metabolism was shifted

N

- I.2

1O O Q

200

0 0

180 1 .o

160

0 h c

-

m

140 120 100

0,

m

0 0 3

2

80

60 40

I I n

0

c

c

+-

-

n

a,

cn

40

2 0 U a,

m

cn

20

i

h

-

cn-

cn v

Y

80

.-C 0

g: n

2-

aJ

. L

20 15

b

10

20

0

7

30 0

->0

0.8

10 5

)O 6

0

0

' 0.0

Time (h) FIG.4. Typical o-glucose (200 glliterl-based batch fermentation profile obtained with the D-ribose-producing Bacillus s u b t i h strain ATCC 21951. The agitation speed was 1000 rpm and the aeration rate 3 vvm (De Wulf et a].,1997).

3

8

MICROBIAL SYNTHESIS OF D-RIBOSE

201

more significantly into the oxidative PPC (De Wulf et al., 1996b). Moreover, the concentration of glycolytic end-products drastically dropped (De Wulf et al., 1997), making D-ribose recovery easier and cheaper. In conclusion, an oxidatively cultured and postexponentially grown Bacillus inoculum should be used to start the fermentation. The composition of the inoculation medium seems not very important, and cell growth should only be sustained sufficiently (De Wulf, 1995). The fermentation process itself preferably should be performed as oxidatively as possible, this by applying overpressure using oxygen-enriched air or by controlling the dissolved oxygen concentration by a feedbackresponding agitation and/or aeration device (De Wulf et al., 1997). Regarding the composition of the fermentation medium, yeast extract shifts D-glucose into the glycolysis, leading to lower D-ribose yields and higher concentrations of glycolytic end-products and biomass. CSL plus (NHJ2SO4 was an excellent alternative in this respect (De Wulf, 1995). The concentration of phosphates in the medium may be important, since this may affect alkaline phosphatase activity, which probably dephosphorylates the intracellularly accumulated ~-ribose-5-phosphate (Fig. 2). So far, no data on this phenomenon have been reported. The pH of the medium preferably should be around 6.0. Higher values activate glycolytic catabolism and biomass synthesis, while lower values decrease biomass synthesis and the substrate conversion rate (De Wulf et al., 1997). The slow utilization of D-glucose by Bacillus tkt mutants is well known. Although D-glUCOSe is the most preferable substrate from an economical point of view, the use of an oxidative PPC-converted cosubstrate such as D-ghconic acid may be of interest, especially since transketolase mutants do not perform glucose-based catabolite repression on the utilization of this PPC-converted carbon source (De Wulf et al., 1997). D.

RECOVERY OF D-RIBOSEFROM THE

FERMENTATION MEDIUM

Many procedures have been developed to isolate D-ribose from the culture medium (Kintaka et al., 1986; Sasajima and Yoneda, 1971; Sasajima et al., 1985; Yoneda and Sasajima, 1969). A generalizing scheme is shown in Fig. 5. In summary, the cells are removed by centrifugation/filtration, and the cell-free supernatant liquidlfiltrate is chromatographically decolorized with activated charcoal (Yoneda and Sasajima, 1969). Subsequently, the broth is concentrated under reduced pressure to half (Sasajima et al., 1985) or one-tenth (Sasajima and Yoneda, 1971) its

202

P.

DE

WULF AND E. J. VANDAMh4E

fermentation medium

A

centrifugation/filtration

cells

SupernatanVfiltrate

1

decolorization

decolorized medium Iconcentration viscous solution removal of coexisting carbohydrates

1

centrifugation/filtration cells

SupernatanVfiltrate

1-

A

ethanol

centrifugation/filtration

cells

su pernat ant/f iltrate

1 1 1

concentration desalination concentration

syrupy solution 1-ethanol crystalline D-ribose FIG.5. Recovery of D-ribose from the fermentation medium.

volume. To prevent the isolation of D-ribose to be disturbed by coexisting carbohydrates (e.g., remaining substrate), the medium can be supplemented with an enzyme (e.g., D-glucose oxidase, rarely added) or with microbial cells (mostly Baker's yeast) that do not assimilate D-ribose. In this respect, two volumes of phosphate buffer (0.01 M , pH 6.0) and 10 to 30% (w/v) of Baker's yeast are added to the broth (37"C, 1h). The cells are subsequently removed by centrifugation/filtration and the supernatant/filtrate treated with ethanol (20%, v/v). After another centrifugation/filtration step, the supernatant/filtrate is concentrated under reduced pressure and desalted with ion exchange resins. More specifically, the solution is treated with 0.015 vol% of

203

MICROBIAL SYNTHESIS OF D-RIBOSE

amberlite IR120 (H+ form) to exclude cations and subsequently with 0.015 vol% of amberlite IRA 400 (borate form) to remove anions. The broth is next concentrated under reduced pressure, and three to four volumes of ethanol are added to the syrupy solution to induce the crystallization of D-ribose. The latter is then isolated by centrifugation/filtration in an (average) overall yield of 94% (Kintaka et d., 1986; Sasajima and Yoneda, 1971;Sasajima et al., 1975;Yoneda and Sasajima, 1969).

X. Kinetics of D-Ribose Production by Bacillus spp.

Kinetic models that define microbial growth and product formation sensu latiore have been presented by many researchers. Kono (1968) and Kono and Asai (1969)expressed the microbial growth rate and the production rate by the following equations:

dCx/dO = K, @ C,, d c p / d e = K~~@ c, + ~

(1) ~ - q)c,, ~ (

1

(2)

where C, and C,, respectively, represent the cell concentration and the D-ribose concentration, K, is a growth rate constant, and Kpland Kp2are production rate constants (Asai and Kono, 1984;Kono, 1968;Kono and Asai, 1969);0 stands for the time of fermentation, while 9 represents an apparent coefficient of growth activity, which depends on the bacterial growth phase, and is defined as (Asai and Kono, 1984) Induction phase @ = 0,

(3)

Transient growth phase @ = cp,

(4)

Exponential growth phase @ = 1,

(5)

Declining growth phase @

=

[CxC/(Cxm - C,ll

[(C,,

-

C,)/C,l,

(6)

where @ is a growth coefficient whose value (cp) increases from zero to unity in a transient growth phase. C, stands for the critical cell concentration, or the cell concentration at the boundary of the exponential and the declining growth phase. CXm is the maximal cell concentration, predicted by a theoretical procedure (Kono, 1968). In the declining growth stage, the value of @ decreases from unity to zero. When a constant growth phase is included in the time course of microbial growth, the value of @ is expressed as follows (Asai and Kono, 1984): Constant growth phase @ = Cx,/C,,

(7)

where Cxd represents the cell concentration at the boundary of the exponential growth phase and a constant growth phase, yielding

@ c x = Cx,growthj (1- @)cx=

Cx,nongrowth.

(8) (9)

The general equations that define the cellular growth rate and the microbial production rate hence can be rewritten as follows:

dcxldO = KxCx,growth,

(10)

dCp/d@= Kplcx,growth + KpzCx,nongrowth*

(11)

The latter equation, which describes the microbial production rate (dCp/dO), is based on the assumption that the microorganisms do not die throughout the fermentation process (Asai and Kono, 1984). Since cell lysis occurs during the fermentation, due to nutritional exhaustion, Eq. (11) could not fully explain the fermentation course of D-ribose production by Bacillus spp. (Asai et al., 1978; Asai and Kono, 1984). By introducing D, as the accumulated concentration of dead cells, the death rate can be expressed by (Asai and Kono, 1984) dDJd0 = KD(1 - @)C,,

(12)

in which KO is the death rate constant and C, is the theoretically accumulated concentration of cells. The growth rate of the living D-ribose-producing cells can be derived from Eqs. (1)and (12): dLxfdO = K, @ C, - K D ( ~ @)C,.

(13)

Finally, the production rate of D-ribose (dCp,JdO) and of the other microbial metabolites (e.g., D-gluconic acid [dCp,$dO]) can be expressed by Eq. (14), in which (1 - @)Cx- 0, represents the concentration of the nongrowing cells, some of which are dead and have no activity of product formation: dCp/de = Kpl @ C, + Kpz[(l - @)C, - Ox].

(14)

XI. Conclusions and Future Perspectives

During the last decades, microbial D-ribose synthesis has been commercially established. High production efficiencies have been achieved, notably due to the optimization of fermentation processes with mutant Bacillus strains that were derived by random mutagenesis. Even higher D-ribose yields may be obtained by introducing accurate genetic tech-

MICROBIAL SYNTHESIS OF D-RIBOSE

205

niques. By applying the ever-increasing amount of information available on the genetic background of, for example, Bacillus subtilis, mutant strains can be created with a genetically enhanced oxidative PPC activity, with a completely eliminated transketolase (possibly two isoenzymes in Bacillus spp., as in E. coli; Sprenger, 1995; and Saccharomyces cerevisae; Schaaf-Gerstenschlager and Zimmerman, 1993), and D-ribulose-5-phosphate-3-epimerase activity, and with an abolished secretion of unwanted side-products (e.g., D-gluconic acid, acetoin, and 2,3-butanediol). This approach, combined with the current process-technological knowledge may lead to shorter fermentation times, higher D-ribose yields, and simpler D-ribose recovery. Consequently, the D-ribose market price might drop, which may further stimulate applied biomedical research with this pentose. D-ribose has already captured the eye of medical carbohydrate chernists. Indeed, due to its (in)directrole in many aspects of cell metabolism, cleverly designed D-ribose-based drugs may be the pharmaceuticals of the future. Since many clinical trials have shown that D-ribose derivatives act as potent antiviral and anticancer therapeutics, the future of D-ribose may have never looked sweeter. REFERENCES Ahrens, R., Jann, B., Jann, K., and Brade, H. (1988). Structure of the K74 antigen from Escherichia coli 044:K74:H18, a capsular polysaccharide containing hranosidic pKDO residues. Carbohydr. Res. 179, 223-231. Alabed, Y., Ulrich, P., Kapurniotu, A,, Lolis, E., and Bucala, R. (1995). Model studies of the Maillard reaction of arg-lys with D-ribose. Bioorg Med. Chem. Lett. 5, 2929-2930. Alberda Van Ekenstein, W., and Blanksma, J. J. (1909). Uber de gekristalliseerde I-ribose. Chern. Weekbl. 6, 373-375. Alberda Van Ekenstein, W., and Blanksma, J. J. (1913). Uber u-ribose. Chern. Weekbl. 10, 664.

Amemura, A,, Hisamatsu, M., Ghai, S . K., and Harada, T. (1981). Structural studies on a new polysaccharide-containing u-riburonic acid, from Rhizobium melilotti IF0 13336. Carbohydr. Res. 91, 59-65. Artz, R., and Jenkinson, W. (1994). Natural food colours. Food Tech Europe 1, 204-208. Asai, T., and Kono, T. (1984). Industrial application of fermentation kinetics. In “Advances in Biotechnological Processes” (A. h4izrahi and A. L. van Wezel, eds.), Vol. 3, pp. 53-79. Alan R. Liss, New York. Asai, T., Doi, M., Kono, T., and Fukada, H. (1978). Kinetic study on the production of 11-ribose by Bacillus sp. J. Ferment. Technol. 56, 91-95. Atkinson, B., and Mavituna, F. (1991). “Biochemical Engineering and Biotechnology Handbook,” 2nd ed. M. Stockton Press, Macmillan, London. Baddiley, J. (1955). Roy. Inst. Chem., Lectures, Monographs Repts. 3,l-17. In “Chemistry of Ribose and Deoxyribose” (W. G. Overend and M. Stacey [1955]), “The Nucleic Acids” (E. Chargaff and J. N. Davidson, eds.), Vol. 1,pp. 9-80. Academic Press, New York.

206

P. DE WULF AND E. J. VANDAMME

Bailey, A. J., Sims, T. J., Avery, N. C., and Halligan, E. P. (1995).Non-enzymatic glycation of fibrous collagen: Reaction products of glucose and ribose. Biochem. J. 305, 385390.

Baker, B. R., and Schaub, R. E. (1953).Achromycin: Synthetic studies, 111: Synthesis of 3-amino-o-nbose, a hydrolytic product. J. Am. Chem. Soc. 75, 3864-3865. Barbieri, D., Grassilli, E., Monti, D., Salvioli, S., Franceschini, M. G., Franchini, A., Bellesia, E., Salomoni, P., Negro, P., Capri, M., Troiano, L., Cossarizza, A., and Franceschi, C. (1994).D-ribose and deoxy-D-ribose induce apotosis in human quiescent peripheral blood mononuclear cells. Biochem. Biophys. Res. Commun. 201, 1109-1 116.

Barrett, A. G. M., Barton, D. H. R., Bielski, R., and McCombie, S. W. (1977).Conversion of vicinal diols into olefins via the derived bisthiocarbonates. J. Chem. Soc., Chem. Commun., pp. 866-868. Barrett, A. G. M., Barton, D. H. R., and Bielski, R. (1979).Reactions of relevance to the chemistry of aminoglycoside antibiotics, Part 11:Preparation of olefins from vicinal diols. J. Chem. Soc., Perkin Trans. 1 pp. 2378-2381. Bentley, H. R., Cunningham, K. G., and Spring, F. S. (1951).Cordycepin, a metabolic product from cultures of Cordyceps militaris (Linn.). Link, Part 11: The structure of cordycepin. J. Chem. SOC.,pp. 2301-2305. Bergmann, M., and Schotte, H. (1921).h e r das Glucal-Problem. Ber. Dtsch. Chem. Ges. 54, 440-455. Bessodes, M., Abushanab, E., and Panzica, R. P. (1981).A new and versatile method for the preparation of unsaturated sugars. J. Chem. Soc., Chem. Commun., pp. 26-28. Blanksma, J. J., and Alberda Van Ekenstein, W. (1908).Bijdrage tot de kennis der suikers. Chem. Weekbl. 5, 777-782. Borman, A. H., de Jong, E. W., Thierry, R., Westbroek, P., Bosch, L., Gruter, M., and Kamerling, J. P. (1987).Coccolith-associated polysaccharides from cells of Emiliana huxleyi (Haptophyceae). J. Phycol. 23, 118-123. Bredereck, H., and Rothe, G. (1938).Nucleinsauren, VI: Mitteil: Darstellung der Nucleoside durch enzymatische Hydrolyse der Hefenucleinsaure; zugleich ein Beitrag zur Darstellung der d-ribose. Ber. Dtsch. Chem. Ges. 71, 408-411. Bredereck, H., Kothnig, M., and Berger, E. (1940).h e r die D-ribose (Darstellung einer kristallisierten Anhydroribose). Ber. Dtsch. Chem. Ges. 73,956-962. Bryant, W.M. D. (1989).Further optical crystallographic studies of biochemical substances. Microscope 37, 233-240. Cantor, S. M., and Peniston, Q. P. (1940).The reduction of aldoses at the dropping mercury cathode: estimation of the aldehyde structure in aqueous solutions. J. Am. Chem. Soc. 62, 2113-2121. Carmona, P., and Molina, M. (1990).Raman and infrared spectra of D-ribose and D-ribose 5-phosphate. J. Raman Spectrosc. 21, 395-400. Chan, M. F., and Hsiao, C.-N. (1992).A versatile and stereospecific synthesis of a dihydroxyethylene dipeptide isostere of renin inhibitors from D-ribose. Tetrahedron Lett. 33, 3567-3570. Chaplin, M. F., and Kennedy, J. F. (1986).“Carbohydrate analysis: A Practical Approach.” IRC Press, Oxford, Washington, DC. Cherbuliez, E., and Bernhard, K. (1932).Recherches sur la graine de croton, I: Sur le Helv. Chim. Acta 15, 464-471. crotonoside (2-hydroxy-6-aminopurine-~-riboside). Cohn, W.E. (1949).The separation of purine and pyrimidine bases and of nucleotides by ion exchange. Science 109, 377-378.

MICROBIAL SYNTHESIS OF D-RIBOSE

207

Cooper, A. J., and Salomon, R. G. (1990).Total synthesis of halochondrins: Enantioselective construction of a homochiral pentacyclic Cl-Cl5 intermediate from D-ribose. Tetrahedron Lett. 31, 3813-3816. Cornell, K. A.,Winter, R. W., Tower, P. A., and Riscoe, M. K. (1996).Affinity purification of 5-methylthioadenosine kinase and 5-methylthioribose/S-adenosylhomocysteine nucleosidase from Klebsiella pneumoniae. Biochem. J. 317,285-290. Cortes, S. J., Mega, T. L., and Van Etten, R. L. (1991).The l80isotope shift in 13C nuclear magnetic resonance spectroscopy, 14: Kinetics of oxygen exchange at the anomeric carbons of o-ribose and D-2-deoxyribose. 1.Org. Chem. 56,943-947. Cuzzoni, M. T., Stoppini, G., Gazzani, G., and Mazza, P. (1988).Influence of water activity and reaction temperature of ribose-lysine and glucose-lysine Maillard systems on mutagenicity, absorbance and content of furfurals. Food Chem. Toxicol. 26,815-825. Cuzzoni, M. T., Stoppini, G., Gazzani, G., and Mazza, P. (1989). Effect of storage on mutagenicity, absorbance and content of furfural in the ribose-lysine Maillard system. Food Chem. Toxicol. 27, 209-213. Davis, A. R., Newton, E. B., and Benedict, S. R. (1922).The combined uric acid in beef blood. 1. Biol. Chem. 54, 595-599. De Clercq, E., and Torrence, P. F. (1978).Nucleoside analogues with selective antiviral activity. J. Carbohydrates Nucleosides Nucleotides 5, 187-224. De Wulf, P. (1995). Microbial synthesis of o-ribose. Ph.D. thesis, University of Ghent, Belgium. De Wulf, P., Soetaert, W., Schwengers, D., and Vandamme, E. J. (1996a).Screening and mutational improvement of a D-ribose secreting Candida pelliculosa strain. J. Ferm. Bioeng. 82,1-7. De Wulf, P.,Soetaert, W., Schwengers, D., and Vandamme, E. J. (1996b).o-glucose does not catabolite repress a transketolase-deficient o-ribose producing Bacillus subtilis mutant strain. J. Ind. Microbiol. 17,104-109. De Wulf, P.,Soetaert, W., Schwengers, D., and Vandamme, E. J. (1997).Optimization of 1)-ribose production with a transketolase-affected Bacillus subtilis mutant strain in glucose and gluconic acid-based media. J. Appl. Microbiol. 82,1-6. Dische, Z.(1962).Color reaction of ketoses with carbazole and sulfuric acid. In “Methods in Carbohydrate Chemistry” (R. H. Whistler, M. L. Wolfrom, J. N. Bemiller, and F. Shafizadeh, eds.), Vol. 1,pp. 481-482. Academic Press, New York. Dondoni, A,, Fantin, G., Fogagnolo, M., and Medici, A. (1986).Synthesis of long-chain sugars by iterative, diastereoselective homologation of 2,3-O-isopropylidene-~-glyceraldehyde with 2-trimethylsilylthiazole. Angew. Chem., Int. Ed. Engl. 25,835-837. Dolin, R. (1985).Antiviral chemotherapy and chemoprophylaxis. Science 227, 12961302.

Draths, K. M., Pompliano, D. L., Conley, D. L., Frost, J. W., Berry, A., Disbrow, G. L., Staversky, R. J., and Lievense, J. C. (1992).Biocatalytic synthesis of aromatics from D-glucose: The role of transketolase. J. Am. Chem. SOC.114,3956-3962. Dubinsky, O.,Simon, B., Karamanos, Y., Geresh, S., Barak, Z., and Arad, S. (1992). Composition of the cell wall polysaccharide produced by the unicellular red alga Hhodella reticulata. Plant Phys. Biochem. 30,409-414. Dubois, L., and Dodd, R. H. (1993). Stereocontrolled synthesis of aziridine-2-lactones from D-ribose and D-lyXOSe. Tetrahedron 49,901-910. Dueholm, K. L., Motawia, M. S., Pedersen, E. B., Nielsen, C., and Lundt, I. (1992). Synthesis of 3’-alkylthio-~’,3’-dideoxynucleosides with potential anti-HIV activity from 2-deoxy-D-ribose, using a phosphorous pentoxide reagent. Arch. Pharm. 325, 597-601.

208

P.

DE

WULF AND E. J. VANDAM-

Dyer, D. G., Dunn., J. A,, Thorpe, S. R., Bailie, K. E., Lyons, T. Y., McCance, D. R., and Baynes, J. W. (1993).Accumulation of Maillard reaction products in skin collagen in diabetes and aging. J. Clin. Invest. 91,2463-2469. Eidels, L., and Osborn, M. J. (1971). Lipopolysaccharide and aldoheptose synthesis in transketolase mutants of Salmonella typhymurium. Proc. Natl. Acad. Sci. U.S.A. 68, 1673-1677.

Fairbanks, A. J., Ford, P. S., Watkin, D. J., and Fleet, G. W. J. (1993). Synthesis of 1-epihydantocidin from D-ribose. Tetrahedron Lett. 34, 3327-3330. Fischer, E., and Piloty, 0. (1891). iiber eine neue Pentonsaure und die zweite inactive Trioxyglutarsaure. Ber. Dtsch. Chem. Ges. 24,4214-4225. Fujita, Y.,and Fujita, T. (1987).The gluconate operon of Bacillus subtilis encodes its own transcriptional negative regulator. Proc. Natl. Acad. Sci. U.S.A. 84,4524-4528. Fujita, Y.,Fujita, T., Miwa, Y., Nihasi, J., and Aratani, Y. (1986). Organization and transcription of the gluconate operon, gnt, of Bacillus subtilis. J. Bid. Chem. 261, 13744-1 3753.

Fujita, Y., Miwa, Y., Galinier, A., and Deutscher, J. (1995). Specific recognition of the Bacillus subtilis gnt cis-acting catabolite responsive element by a protein complex formed between CcpA and seryl-phosphorylated HPr. Mol. Microbiol. 17,953-960. Gallos, J. K., Goga, E. G., and Koumbis, A.E. (1994a).Expeditious synthesis of aminocyclo-pentitols from D-ribose via intramolecular nitrone cycloaddition. J. Chem. Soc., Perkin Trans. I , pp. 613-614. Gallos, J. K., Koftis, T. V., and Koumbis, A. E. (1994b).Synthesis of enantiomerically pure bicyclo[3.l.Olhexanes from D-ribose by intramolecular cyclopropanation. J. Chem. SOC., Perkin Trans. 1, pp. 611-612. Garegg, P. J., and Samuelsson, B. (1979). Conversion of vicinal diols into olefins using triphenylphosphine and triiodoimidazole. Synthesis, pp. 813-814. Gay, P., and Delobbe, A. (1977).Fructose transport in Bacillus subtilis. Eur. J. Biochem. 79,363-373.

Gehrke, M., and Aichner, F. X. (1927). Uber das Arabinal. Ber. Dtsch. Chem. Ges. 6 0 , 918-92 2.

Gerber, N. N., and Lechevalier, H. A. (1962). 3-amino-3-deoxyadenosine, an antitumor agent fiom Helminthosporiurn spp. J. Org. Chem. 27,1731-1732. Godin, P. (1953). Etude du m6tabolisme ternaire de Penicillium brevicompactum, I: Analyse chromatographique du milieu de culture. Biochim. Biophys. Acta 11,114118.

Gomez, L. D., and Wallace, J. W. (1986).Flavonoids of Phlebodium. Biochem. Syst. Ecol. 14,407-408.

Greenly, W. J. (1990).Renin inhibitors. Med. Res. Rev. 10,173-236. Gross, M., Kormann, B., and Zollner, N. (1991).Ribose administration during exercise: Effects on substrates and products of energy metabolism in healthy subjects and a patient with myo-adenylate deaminase deficiency. Klin. Wochenschr. 69,151-155. Hackland, P. L., Parolis, H., and Parolis, L. A. (1991).Escherichia coli 09:K38capsular antigen: another ribofuranose-containing glycan. Carbohydr. Res. 219,193-201. Hackman, R. H., and Trikojus, V. M. (1952).The composition of the honeydew excreted by Australian coccoliths of the genus Ceroplastes. Biochem. J. 51, 653-656. Hadden, J. W., Ongradi, J., Specter, S., Nelson, R., Sosa, M., Monell, C., Strand, M., Giner-Sorolla, A., and Hadden, E. M. (1992).Methylinosine monophosphate: A potential immunotherapeutic for early human immunodeficiency virus (HIV) infection. Int. J. Immunopharmacol. 14,555-563. Hammarsten, 0. (1894).Zur Kenntniss der Nucleoprote'ide. 2. Physiol. Chem. 19,19-37.

MICROBIAL SYNTHESIS OF D-RIBOSE

209

Hendel, R. C., Vansant, J. P., Wilson, R. A,, Sonnemaker, R., Lally, R. M., and Hoff, J. V. (1995). The use of o-ribose in conjunction with TI-210 perfusion imaging for identification of reversible perfusion abnormalities: Results of a placebo-controlled, double blinded trial. Circulation 92, 1 4 4 7 . Hill, J. M., Hutchinson, E. J., Le Grand, D. M., Roberts, S. M., Thorpe, A. J., and Turner, N. J. (1994). Preparation of neplanocin-A from D-ribose and by a chemoenzymatic method. J. Chem. Soc., Perkin Trans. 1 , pp. 1483-1487. Hiroshi, N., Takeshi, T., and Masahiko, H. (1985). Herstellung von einer D-ribose-haltigen Lasung. Ger. Pat. DE-3437571-A1. Hirst, E. L. (1949). The occurrence and significance of the pentose sugars in nature, and their relationship to the hexoses. J. Chem. Soc. 1,522-533. Horii, S., Fukase, H., Matsuo, T., Kameda, Y., Asano, N., and Matsui, K. (1986). Synthesis of a-o-glucosidase inhibitory activity of N-substituted valioamine derivatives as potential oral antidiabetic agents. J. Med. Chem. 29, 1038-1046. Horton, D., and Walaszek, Z. (1982). Tautomeric equilibria of some sugars by partially relaxed 13Cpulsed Fourier-transform, nuclear magnetic resonance spectroscopy. Carbohydr. Res. 105,145-153 Inoue, T., Murakawa, S., and Endo, A. (1992). Production and partial properties of an extracellular polysaccharide from Streptomyces sp. A-1845. J. Ferment. Bioeng. 73, 440442.

Isbell, H. S., Frush, H. L., Wade, C. W. R., and Hunter, C. E. (1969). Transformation of sugars in alkaline solutions. Carbohydr. Res. 9, 163-175. Jachymek, W., Peterson, C., Helander, A,, Kenne, L., Lugowski, C., and Niedziela, T. (1995). Structural studies of the 0-specific chain and a core hexasaccharide of Hafnia alvei strain 1192. Carbohydr. Res. 269, 125-138. Jeanlotz, R. W., and Fletcher Jr., H. G., (1951).The chemistry of D-ribose. Adv. Carbohydr. Chem. Biochem. 6, 135-174. Josephson, B. L., and Fraenkel, D. G. (1969). Transketolase mutants of Escherichia coli. 1.Bacieriol. 100, 1289-1295. Josephson, B. L., and Fraenkel, D. G. (1974). Sugar metabolism in transketolase mutants of Escherichia coli.J. Bacteriol. 118,1082-1089. Khalifah, R. G., Todd, P., Booth, A. A., Yang, S. X., Mott, J. D., and Hudson, B. G. (1996). Kinetics of nonenzymatic glycation of ribonuclease A leading to advanced glycation end products. Paradoxical inhibition by ribose leads to facile isolation of protein intermediate for rapid post-Amadori studies. Biochemistry 35,4 6 4 5 4 6 5 4 . Kintaka, K., Sekitani, Y. H., and Yamaguchi, T. S. (1986). Verfahrung zur Erzeugung von n-Ribose. Ger. Pat. DE-3524549-A1. Kishimoto, K., Kintaka, K., and Ochiyama, N. (1990). Production of D-ribose. U.S. Pat. 4,904,587.

Kono, T. (1968). Kinetics of microbial cell growth. Biotech. Bioeng. 10, 105-131. Kono, T., and Asai, T. (1969). Kinetics of fermentation processes. Biotech. Bioeng. 11, 293-321.

Kossel, A. (1891). Uber die chemische Zusamensetzung der Zelle. Arch, Anat. Physiol., Physiol. Abt., pp. 181-186. Kotilainen, P., Oksman, P., Viljanen, M. K., Nikoskelainen, J., and Huovinen, P. (1990). Analysis of the relationship between bacterial adherence and extracellular production of mannose, galactose, glucose and ribose in Staphylococcus epidermidis and Staphylococcus hominis. Eur. J. Clin.Microbiol. Infect. Dis.9, 873-879. Kubala, J., and Burdatsova, A. (1991). Preparation of D-ribose. Czech Pat. CS-274863.

210

P.

DE

WULF AND E. J. VANDAMME

Lacourt-Gadras, B., Grignon-Dubois, M., and Rezzonico, B. (1992). Nouvelle voie d’accks au D-ribose et au D-lyxose. Carbohydr. Res. 235,281-288. Laufer, L., and Charney, J. (1945a). Production of D-ribose. U S . Pat. 2,379,913. Laufer, L., and Charney, J. (1945b). Production of D-ribose. U.S. Pat. 2,379,914. Leer, J. C., and Hammer-Jespersen, K. (1975). Multiple forms of phosphodeoxyribomutase from Escherichia coli. Physical and chemical characterization. Biochemistry 14,599607.

Levene, P. A. (1935). Note on the preparation of crystalline d-mannose and of crystalline d-ribose. J. B i d . Chem. 108, 4 1 9 4 2 0 . Levene, P. A., and Clark, E. P. (1921). D-ribohexaminic acids. I. Biol. Chem. 46, 19-33. Levene, P. A., and Jacobs, W. A. (1908). ijber die Inosinsaure. Ber. Dtsch. Chem. Ges. 41, 2703-2707.

Levene, P. A., and Jacobs, W. A. (1909). h e r die Inosinsaure. Ber. Dtsch. Chem. Ges. 42, 1198-1203.

Levene, P. A., and Jacobs, W. A. (1911). Uber die Inosinsaure. Vierte Mitteilung. Ber. Dtsch. Chem. Ges. 44, 746-753. Levene, P. A., Mikeska, L. A., and Mori, T. (1930). On the carbohydrate of thymonucleic acid. J. B i d . Chem. 85, 785-787. Liang, J. N., and Rossi, M. T. (1990). In vitro non-enzymatic glycation and formation of browning products in the bovine lens alpha-crystallin. Exp. Eye Res. 50,367-371. Lofgren, N., and Luning, B. (1953). On the structure of nebularine. Acta Chem. Scand. 7, 225.

Lundt, I. (1993). Personal communication. Lygo, B., Swiatyj, M., Trabsa, H., and Voyle, M. (1994). Synthesis of (+)-gabosines C and E from D-ribose. Tetrahedron Lett. 35,4197-4200. Maier, G. D. (1977). The study of the mutarotation of D-glucose, D-fructose and D-ribose by use of their circular dichroism. Carbohydr. Res. 53, 1-11. Marini, M., Zunica, G., and Franceschi, C. (1985). Inhibition of cell proliferation by D-ribose and deoxy-D-ribose. Proc. SOC.Exp. B i d . Med. 180, 246-257. Marquet, M., Wagner, M.-C., Delobbe, A., Gay, P., and Rapoport, G. (1970). Mise en Bvidence de systbmes de phosphotransferases dans le transport du glucose, du fructose et du saccharose chez Bacillus subtilis. C. R. Acad. Sci. 271, 449-452. Mega, T. L., Cortes, S. J., and Van Etten, R. L. (1990). The l 8 0 isotope shift in 13Cnuclear magnetic resonance spectroscopy, 13: Oxygen exchange at the anomeric carbon of D-glucose, D-mannose, and D-fructose. 1.Org. Chem. 55,522-528. Merck, E. (1893). Adonit, ein kristallisierenden Korper aus Adonis vernalis. Arch. Pharm. 231,129-131.

Merello, S., Xavier, M. T., and Parodi, A. J. (1994). Novel (rhamnosyl and ribosyl) and uncommon (xylosyl) monosaccharide residues are present in asparagine-linked oligosaccharides of the trypanosomatid Blastocrithidia culicis. J. B i d . Chem. 269, 20294-20298.

Merello, S., Xavier, M. T., and Parodi, A. J. (1995). The presence of galactofuranose and ribose units in asparagine-linked oligosaccharides of the digenetic trypanosomatic Endotrypanum schaudinni. Mol. Biochem. Parasitol. 69, 73-79. Miyagawa, K., Miyazaki, J., and Kanzaki, N. (1992). Method of producing D-ribose. Eur. Pat. 0-501-765-A1. Morisaki, N., Kobayashi, H., Iwasaki, S., Furihata, K., Dabbs, E. R., Yazawa, K., and Mikami, Y. (1995). Structure determination of ribosylated rifampicin and its derivative: New inactivated metabolites of rifampicin by mycobacterial strains. J. Antibiot. 48, 1299-1303.

MICROBIAL SYNTHESIS OF D-RIBOSE

211

Nagaraj, R. H., and Monnier, V. M. (1992).Isolation and characterization of a blue fluorophore from human eye lens crystallins: In vitro formation from Maillard reaction with ascorbate and ribose. Biochim. Biophys. Acta 1116,34-42. Ogata, K.(1976).History and prospects. In “Microbial Production of Nucleic Acid-Related Substances” (K. Ogata, S. Kinoshita, T. Tsunoda, and K. Aida, eds.), pp. xiii-xviii. Kodansha Ltd., Tokyo, and Wiley, New York. Overend, W. G., and Stacey, M. (1955).Chemistry of ribose and deoxyribose. In “The Nucleic Acids” (E. Chargaff and J. N. Davidson, eds.), Vol. 1, pp. 9-80, Academic Press, New York. Paquette, L. A., and Bailey, S. (1995).Evaluation of D-ribose as an enantiopure building block for construction of the C-ring of taxol and its congeners. J. Org. Chem. 60, 7849-7856.

Perlmutter, N. S., Wilson, R. A., Angello, D. A., Palac, R. T., Lin, J., and Brown, B. G. (1991).Ribose facilitates thallium-201 redistribution in patients with coronary artery disease. J. Nucl. Med. 32, 193-200. Phelps, F. P. (1939).Production of D-ribose. U.S. Pat. 2,152,662. Phelps, F. P., Isbel, H. S., and Pigman, W. (1934).Mutarotation of p-d-ribose and p-1-ribose. J. Am. Chem. SOC.56, 747-748. Pliml, W.,von Arnim, T., and Hammer, C. (1993).Effects of therapeutic ribose levels on human lymphocyte proliferation in vitro. Clin. Invest. 71, 770-773. Rawn, J. D. (1989).“Biochemistry.” Neil Patterson Publishers, Carolina Biological Supply Company, Burlington, NC. Rhaese, H. J., and Groscurth, R. (1976). Control of development: role of regulatory nucleotides synthesized by membranes of Bacillus subtiljs in initiation of sporulation. Proc. Natl. Acad. Sci. U.S.A. 73,331-335. Rhaese, H. J., and Groscurth, R. (1979).Apparent dependence of sporulation on synthesis of highly phosphorylated nucleotides in Bacillus subtilis. Proc. Natl. Acad. Sci. 1J.S.A. 76,842-846. Rhaese, H. J., Hoch, J. A,, and Groscurth, R. (1977).Studies on the control of development: Isolation of Bacillus subtilis mutants blocked early in sporulation and defective in synthesis of highly phosphorylated nucleotides. Proc. Natl. Acad. Sci. U.S.A. 74, 1125-1 129.

Rudnicki, W. R., Lesyng, B., and Harvey, S. C. (1994).Lagrangian molecular dynamics using selected conformational degrees of freedom, with application to the pseudorotation dynamics of furanose rings. Biopolymers 34, 383-392. Rudrum, M., and Shaw, D. F. (1965).The structure and conformation of some monosaccharides in solution. J. Chem. SOC.1, 52-57. Rycroft, A. D., Singh, G., and Wightman, R. H. (1995). Stereoselective synthesis of 3-P-~-ribofuranosylpyrazole from 2,3-O-isopropylidene-~-ribose; a new route to pyrazole C-nucleosides. J. Chem. SOC.,Perkin Trans. I, pp. 2667-2668. Saier Jr., M.H., Chauvaux, S., Cook, G. M., Deutscher, J., Paulsen, I. T., Reizer, J., and Ye, 7.-J. (1996). Catabolite repression and inducer control in Gram-positive bacteria. Microbiology 142,217-230. Saito. N., and Sugiyama, S.(1966).o-ribose formation by Pseudomonas reptilivora. Agric. B i d . Chem. 30,841-846. Sano, K., Yokozeki, K., and Mitsugi, K. (1977).Enzymatic production of 5-amino-4-imidazole-carboxamide from 5-amino-4-imidazole-carhoxamide-ribose by Bacillus thiaminolyticus. Optimal conditions for the enzyme formation and the enzymatic reaction. Agric. Biol. Chem. 41,2331-2334.

212

P.

DE

WULF AND E. J. VANDAMME

Sasajima, K. (1976). D-ribose. In “Microbial Production of Nucleic Acid-Related Substances” (K. Ogata, s. Kinoshita, T. Tsunoda, and K. Aida, eds.), pp. 199-204. Kodansha Ltd., Tokyo, and Wiley, New York. Sasajima, K., and Kumada, T. (1979). Deficiency of glucose transport in transketolase mutants Bacillus subtilis. Inst. Ferment. Res. Commun. 9, 17-26. Sasajima, K., and Kumada, T. (1981a). Change in the regulation of enzyme synthesis under catabolite repression in Bacillus subtilis pleiotropic mutant lacking transketolase. Agric. Biol. Chem. 45, 2005-2012. Sasajima, K., and Kumada, T. (1981b). Cell surface change of Bacillus subtilis pleiotropic mutant lacking transketolase. Inst. Ferment. Res. Commun. 10, 3-9. Sasajima, K., and Kumada, T. (1983a). Cell surface change of Bacillus subtilis pleiotropic mutant lacking transketolase: Properties of various revertant strains. Inst. Ferment. Res. Commun. 11, 3-9. Sasajima, K., and Kumada, T. (1983b). Deficiency of flagellation in a Bacillus subtilis pleiotropic mutant lacking transketolase. Agric. Biol. Chem. 47, 1375-1376. Sasajima, K., and Yoneda, M. (1971). Carbohydrate metabolism mutants of a Bacillus species, Part 11: o-ribose accumulation by pentose phosphate pathway mutants. Agric. Biol. Cbem. 35,509-517. Sasajima, K., and Yoneda, M. (1974a). Simple procedures for the preparation of D-ribose5-phosphate ketol-isomerase, ~-ribulose-5-phosphate-3-epimeraseand D-sedoheptulose-5-phosphate: o-glyceraldehyde-3-phosphateglycolaldehydetransferase. Agric. Biol. Cbem. 38,1297-1303. ~-glyceraldehyde-3Sasajima, K., and Yoneda, M. (1974b). D-sedoheptulose-7-phosphate: phosphate glycolaldehydetransferase and D-ribulose-5-phosphate 4-epimerase mutants of a Bacillus species. Agric. Biol. Chem. 38,1305-1310. Sasajima, K., and Yoneda, M. (1974~).Verfahrung zur Herstellung von o-ribose. Ger. Pat. DE-2330426-C2. Sasajima, K., and Yoneda, M. (1984). Production of pentoses by microorganisms. Biotecbno]. Genet. Engineer. Rev. 2 , 175-213. Sasajima, K., and Yoneda, M. (1989). Production of D-ribose by microorganisms. In “Biotechnology of Vitamins, Pigments and Growth Factors” (E. J. Vandamme, ed.), pp. 167-197. Elsevier Applied Biotechnology Series, Elsevier, New York. Sasajima, K., Nogami, I., and Yoneda, M. (1970). Carbohydrate metabolism mutants of a Bacillus species, Part I: Isolation of mutants and their inosine formation. Agric. B i d . Cbem. 34, 381-389. Sasajima, K., Fukuhara, T., Matsukura, A., Nakanishi, I., and Yoneda, M. (1972). o-ribose production by pentose phosphate pathway mutants of Bacillus. Paper presented at the Fourth International Fermentation Symposium, Kyoto, 19-25 March. Sasajima, K., Doi, M., Fukuharu, T., Yokota, A., Nakao, Y., and Yoneda, M. (1975). Verfahrung zur Herstellung von D-ribose. Patentschrift DE-2454931-C2. Sasajima, K., Kumada, T., and Yokota, A. (1977). A pleiotropy in carbohydrate metabolism of Bacillus subtilis mutant lacking transketolase. Inst. Ferment. Res. Commun. 8 , 69-77. Sasajima, K., Yokota, A., and Kumada, T. (1985). Alteration of the membrane composition of Bacillus subtilis pleiotropic mutant lacking transketolase. Inst. Ferment. Res. Commun. 12,5-18. Schaaf-Gerstenschlager, I., and Zimmerman, F. K. (1993). Pentose phosphate pathway in Saccbaromyces cerevisiae: Analysis of deletion mutants for transketolase. transaldolase, and glucose-6-phosphate dehydrogenase. Curr. Genet. 24, 373-376.

MICROBIAL SYNTHESIS OF D-RIBOSE

213

Schaub, R. E., and Weiss, M. J. (1958). The synthesis of the four possible methyl-3-amino3-deoxy-o-xylosides. A novel ring expansion of a furanoside to a pyranoside. J. Am. Chem. SOC.80, 4683-4692. Schlegel, H. G. (1984). “General Microbiology,” 6th ed. Cambridge Univ. Press, Cambridge. Shashkov, A. S., Vinogradov, E. V., Knirel, Y. A., Nifant’ev, N. E., Kochetov, N. K., Dabrowski, J., Kholodkova, E. V., and Stanislavsky, E. S. (1993). Structure of the 0-specific polysaccharide of Salmonello arizonae 045. Carhohydr. Res. 241,177-188. Simonart, P., and Godin, P. (1951). Production de pentoses, d’acide 2-cBto gluconique et d’acide glucuronique par Penicillium hrevicompactum. Bull. SOC.Chim. Belg. 60, 446448. Smit. J., Kamio, Y., and Nikaido, H. (1975). Outer membrane of Salmonella fyphymurium: Chemical analysis and freeze-fracture studies with lipopolysaccharide mutants. 1. Bacteriol. 124, 942-958. Smith, D. C. C. (1955). The preparation of D-ribose and 2-deoxy-o-ribose from glucose. Chem. Ind., pp. 92-93. Smith, M. A., Rudnicka-Nawrot, M., Richey, P. L., Praprotnik, D., Mulvihill, P., Miller, C. A., Sayre, L. M., and Perry, G. (1995). Carbonyl-related posttranslational modification of neurofilament protein in the neurofibrillary pathology of Alzheimer’s disease. J. Neurochem. 64, 2660-2666. Sowden, J. C. (1950). The condensation of nitromethane with o-erythrose and 2,4-benzylidene-D-erythrose. J. A m . Chem. SOC.72, 808-811. Sprenger, G. A. (1995). Genetics of pentose phosphate pathway enzymes of Escherichia coli K12. Arch. Microbiol. 164, 324-330. Steele, I. C., Patterson, V. H., and Nicholls, D. P. (1996). A double blind, placebo controlled, crossover trial of o-ribose in MacArdle’s disease. ], Neurol. Sci. 136, 174-177. Steiger, M. (1936). Zur herstellung von D-ribose. Helv. Chim. Acta 19, 189-195. Suhadolnik, R. J. (1970). “Nucleoside Antibiotics.” Wiley-Interscience, New York. Suzuki, T., Tanaka, N., Tomita, F., Mizuhara, K., and Kinoshita, S. (1963). Bacterial accumulation of ribose and ribose-5-phosphate. 1.Appl. Gen. Microbiol. 9, 457-458. Taniguchi, M., Koga, K., and Yamada, S. (1974). Stereoselective synthesis of D-ribose from L-glutamic acid. Tetrahedron 30, 3547-3552. van Alphen, W., Lugtenberg, B., and Berendsen, W. (1976). Heptose-deficient mutants of Escherichia coli K12 deficient in up to three major outer membrane proteins. Mol. Gen. Genet. 147, 263-269. Vanderheiden, B. S. (1970). Phosphate esters in human erythrocytes, VII: Further evidence for ribose-1,5-diphosphate as a natural metabolite. Biochim. Biophys. Acta 215, 242-248. Warren, S . C. (1968). Sporulation in Bacillus suhfilis. Biochemical changes. Biochem. J. 109, 811-818. Weimberg, R. (1961). Pentose oxidation by Pseudomonas fragi. J. Biol. Chem. 236, 629635. Wessely, F., and Wang, S. (1938). I h e r ein vorkommen von Adonit. Monatsh. Chem. 72, 168. Williams, J. C., Verani, R., Alcala, H., Butler, I. J., and Rosenberg, H. S. (1986). Glutamyl ~-ribose-5-phosphatestorage disease: Nephrotic syndrome and cerebral atrophy. Pediatr. Pathol. 5, 277-294. Woisetschlager, M., and Hogenauer, G. (1986). Cloning and characterization of the gene encoding 3-deoxy-~-manno-octu~osonate 8-phosphate synthase from Escherichia coli. J. Bacteriol. 168, 437-439.

2 14

P.

DE

WULF AND E. J. VANDAMME

Wood, T. (1985). “The Pentose Phosphate Pathway.” Academic Press, New York. Yokota, A. (1993). Production of sedoheptulose by Bacillus subtilis. J. Ferm. Bioeng. 75, 409413.

Yokota, A., and Sasajima, K. (1981). Derepressed synthesis of sporulation marker enzymes in a Bacillus species mutant. Agric. B i d . Chem. 45, 2417-2423. Yokota, A., Sasajima, K., and Yoneda, M. (1979). Reactivation of inactivated D-glucose dehydrogenase of a Bacillus species by pyridine and adenine nucleotides. Agric. B i d . Chem. 43, 271-278. Yoneda, M., and Sasajima, K. (1969). Biotechnisches Verfahren zur Herstellung von o-ribose. Patentschrift 19-04-265. Zamenhof, S., Leidy, G., Fitzgerald, P. F., Alexander, H. E., and Chargaff, E. (1953). Polyribophosphate, the type-specific substance of Hemophilus influenme, type b. J. Biol. Chem. 203,695-704. Zamojski, A., and Grynkiewicz, J. (1984). In “The Total Synthesis of Natural Products” (J. Simon, ed.), Vol. 6, pp. 141-232. Wiley-Interscience, New York. Zimmer, H.-G. (3992). The oxidative pentose phosphate pathway in the heart-regulation, physiological significance and clinical implications. Basic Res. Cardiol. 87, 303-316. Zimmer, H.-G., Ibel, H., Suchner, U., and Schad, H. (1984). Ribose intervention in the cardiac pentose phosphate pathway is not species specific. Science 223, 712-714. Zunica, G., Marini, M., Brunelli, M. A., Chiricolo, M., and Franceschi, C. (1986). D-ribose inhibits DNA repair synthesis in human lymphocytes. Biochem. Biophys. Res. Commun. 138, 673-678.

Production and Application of Tannin Acyl Hydrolase: State of the Art P. K. LEKHA AND B. K. LONSANE Fermentation Technology and Bioengineering Department Central Food Technological Research Institute Mysore 570 013, India I. Introduction 11. Historical Highlights 111. Tannin-Hydrolyzing EnzymesA. Enzymes That Act on Hydrolyzable Tannins B. Enzymes That Act on Condensed Tannins C. Significance of Tannins in Plants IV. Source of Tannase A. Microorganisms B. Plants and Animals C. Physiological Significance of Tannase in Plants, Animals, and Microorganisms V. Tannase Assay VI. Production of Tannase A. Tannase Production by Submerged Fermentation B. Tannase Production by Liquid-Surface Fermentation C. Tannase Production by Solid-state Fermentation VII. Regulation of Tannase Biosynthesis A. Regulation by Induction B. Regulation by Catabolite Repression C. Regulation by Feedback Inhibition VIII. Location of Tannase IX. Purification of Tannase X. Properties of Tannase A. pH Optimum and Stability B. Temperature Optimum and Stability C. Molecular Mass D. Enzyme Inhibition E. Kngof Tannase F. Carbohydrate Content XI. Isozymes of Tannase XI. Mode of Action of Tannase XIII. Immobilization of Tannase XIV. Applications of Tannase A. Instant Tea B. Beer Chillproofing C. Wine Making D. Production of Gallic Acid E. Animal Feed Additives F. Miscellaneous XV. Conclusions and Future Prospects References 215 ADVANCES IN APPLIED MICROBIOLOGY, VOLUME 44 Copyright 0 1997 by Academic Press, Inc. All rights of reproduction in any form reserved. 0065-2164197 $25.00

216

P. K. LEKHA AND B. K. LONSANE

I . Introduction Tannin acyl hydrolase (E.C. 3.1.1.20),commonly called tannase, catalyzes the hydrolysis of ester and depside bonds in such hydrolyzable tannins as tannic acid, thereby releasing glucose and gallic acid (Dykerhoff and Ambruster, 1933). Tannase acts only on the ester linkages present in hydrolyzable tannins (Dykerhoff and Ambruster, 1933) and does not act on condensed tannins (George and Sen, 1960). This enzyme finds widespread application in food and beverage processing. Tannase also finds extensive use in the manufacture of instant tea (Coggon et al., 1975). It is also used as a sensitive analytical probe for determining the structure of naturally occurring gallic acid esters (Haslam and Tanner, 1970). However, the practical use of this enzyme is at present limited by its high cost, as well as by a lack of sufficient knowledge about its properties, optimal production, and large-scale application. The available data on tannase are scattered, and there is no review that covers all of its aspects. The present chapter covers all the existing literature on tannase and presents a unified picture of the state of knowledge on the topic, with suggestions for new areas of research.

II. Historical Highlights

The series of events that led to the discovery and an understanding of tannase provides interesting insight into the economical and technological factors involved in the development of a new product. Teighem (1867) was the first to report the formation of gallic acid by two fungi, which occurs naturally when an aqueous solution of tannin or a filtered infusion of gall nut solution is exposed to air. These organisms, identified as Penicillium glaucum and Aspergillus niger, were able to grow with tannic acid as a the sole source of carbon, hydrolyzing it to gallic acid and glucose. Fernback (1901) showed that the hydrolysis of tannin is brought about with the help of a particular enzyme-tannase. This enzyme was isolated from A. niger, and the potential application of tannase in the manufacture of gallic acid using tannin-containing substances as raw materials was discovered. Pottevin (1901) studied the properties of tannase and reported that the enzyme is inducible. The first report on the isolation of tannase produced by A. oryzae in a wheat bran medium by solid-state fermentation was published by Kita

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

21 7

(1917). Further studies on the properties and reaction specificity of the enzyme were described by Dykerhoff and Ambruster (1933), Mikhlin and Gulkina (1932), and Thom and Raper (1945). Fang (1940) reported the use of tannase in the production of gallic acid by fermentation of tannic acid and a new method for cultivation of A. niger on a solid substrate. Toth (1944) showed that tannase is composed of two enzymes: gallic acid esterase (which acts on ester linkages) and depsidase (which acts on depside bonds). Madhavakrishna and Bose (1961) purified and crystallized tannase from a plant source: divi-divi (Caesalpinia coriaria) pods. Tannase from A. niger was purified and characterized by Dhar and Bose (1964).A new method for determining the activity of tannase was described by Iibuchi et al. (1967) based on the change in optical density of the substrate, tannic acid, at 310 nm. Tannase produced by A. flavus (Yamada et al., 19681 and A. oryzae (Iibuchi et al., 1968) was also purified and characterized. A patent was assigned to Tenco Brooke Bond Ltd. (1971) for enzymic solubilization of tea cream using tannase. Yamada and Tanaka (1972) reported the use of tannase in the treatment of grape juice and wine. Kimura et al. (1973) were issued a patent for precipitation of tannase using polyethylene glycol and reported on the characteristics of the enzyme immobilized on an inorganic support by covalent attachment. Coggon et al. (1975) described a continuous-column treatment for the production of cold water-soluble tea using tannase immobilized on glass beads. The presence of low levels of tannase in the rumen of cattle was reported by Begovic and Duzic (19761, who also purified bovine tannase from the mucosal membrane of the rumen and small intestine of cattle for the first time (Begovic and Duzic, 19771. A simple rapid method for detection of tannase on polyacrylamide gel was described by Aoki et al. (1979a). Jean et al. (1981) developed a gas-chromatographic method for assay of tannase activity. Katwa et al. (1981) described a procedure for the assay of immobilized tannase and studied the kinetic parameters of the enzymatic reaction. Deschamps et al. (1983) demonstrated the production of tannase by bacterial strains with chestnut bark as the sole source of carbon. Rajkumar and Nandy (19831 purified and characterized tannase produced by I! chrysogenurn. Weetal (1985a) described enzymatic gallic acid esterification with soluble and immobilized tannase using a wide range of alcohols and diols. Tsai (1985) described a process for the manufacture of instant tea by treating black tea leaves with tannase and cell wall lytic enzymes. Gathon et al. (1989) studied tannase entrapment in reverse micelles for the production of the antioxidant propyl gallate from tannic acid.

218

P. K. LEKHA AND B. K. LONSANE

Beverini and Metche (1990) purified the two isozymes of tannase, tannase I and tannase 11, from A. oryzae. Barthomeuf et al. (1994) reported that an in situ immobilized tannase from A. niger could be obtained by harvesting the mycelium after growth, followed by freezedrying, which could then be used in the hydrolysis of tannic acid to gallic acid. Ill. Tannin-Hydrolyzing Enzymes

The term “tannin” has been used in a wide sense in the botanical literature. Tannins are defined as water-soluble phenolic compounds with molecular weights ranging horn 500 to 3000 that have the property of combining with proteins, cellulose, gelatin, and pectin to form an insoluble complex (Swain and Bate-Smith, 1962). Many authors speak of tannin as if it were a single entity. Poor terminology can confuse readers, particularly that found in the older literature. Tannic acid has frequently been used incorrectly as a general term for tannins (Haslam, 1966). Tannins can be classified into two distinct groups: (i) hydrolyzable tannins and (ii) condensed tannins. Hydrolyzable tannins consist of a polyhydric alcohol esterified with gallic acid or derivatives of gallic acid (Nishira and Joslyn, 1968) (Fig. 1). The hydrolyzable tannins can be subdivided into two types: (a) gallotannins and (b) ellagitannins. Upon hydrolysis, gallotannin yields glucose and phenolic acids, gallic acid being predominant among the latter, for example, Chinese gallotannin (Rhus semilata) and sumac tannin (Rhus coriaria). When hydrolyzed, ellagitannins yield glucose and ellagic acid together with gallic acid, and frequently other acids structurally related to gallic acid, for example, myrobolan ( Terminalia chebula) tannin and divi-divi (Caesalpinia coriaria) tannin (Haslam et a]., 1961). Condensed tannins are made up of phenols of the flavone type and are often called flavolans because they are polymers of such flavan-3-01s as catechin or such flavan-3,4-diols as leucocyanidins. In contrast to the hydrolyzable tannins, they do not contain sugar residues (Goodwin and Mercer, 1983). A typical condensed tannin can be represented by the dimer procyanidin, to which molecules of flavan can be added as indicated (Fig. 2). No individual high-molecular-weight condensed tannin polymer has ever been isolated in the pure state (Grant, 1976). Examples of condensed tannins include wattle (Acacia mollisima) tannin and quebracho (Schinopsis lorentzii) tannin. Condensed tannins are less susceptible to microbial and chemical attack (Lewis and Starkey, 1962).

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE -

-19 219

5 I

OH

OH

H OHO -@:;b 0

I1

HO

/ /

OH

-C

HO

c=o

c=o

OH

OH

HO -4

OH

OH

HO

FIG.1. Hydrolyzable tannin (e.g., sumac tannin).

OH \

R = rnavan-3-01],,

OH

OH

/

OH

OH

FIG.2. Condensed tannin.

A. ENZYMESTHAT ACT ON HYDROLYZABLE TANNINS Tannase cleaves ester and depside linkages in such hydrolyzable tannins as tannic acid and chebulinic acid (Fig. 3). It also acts on the ester and depside linkages in methylgallate and m-digallic acid, respectively. Tannase hydrolyzes only those substrates that contain at least two phenolic OH groups in the acid component. The esterified COOH group must be on the oxidized benzene ring and must not be ortho to one of the OH groups (Dykerhoff and Ambruster, 1933). Chlorogenic acid, a depside of caffeic acid and quinic acid, is resistant to tannase, despite the fact that caffeic acid contains two phenolic OH groups. This is attributed to the presence of a double bond in the

220

P. K. LEKHA AND B. K. LONSANE

H~-o-R~ I

HC-O-Rz I RpO-CH

I

I

0

H~~-o-R~

Hk I

I

H2C- 0-RZ Tannic acid R, - Gallic acid R2- m-Digallic acid

m-Digallic acid

Chebulinic acid

Methyl gallate

FIG.3. Substrates of tannin.

side-chain carrying the esterified COOH group (Dykerhoff and Ambruster, 1933). However, Yamada and Tanaka (1972), in a patent for the use of tannase in wine making, reported that tannase hydrolyzes the chlorogenic acid present in grape juice to yield caffeic acid and quinic acid. Tannase is also reported to hydrolyze (-)-epicatechin gallate and (-)-epigallocatechin-3-gallate (Nierenstein, 1936; Bradfield and Penny, 1948), which are condensed tannins present in tea. B. ENZYMESTHAT ACTON CONDENSED TANNINS

Only a very limited number of microorganisms have been reported to degrade condensed tannins and catechins. The mechanism and the

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

221

enzymes involved in such degradation are unknown (Grant, 1976;Lewis and Starkey, 1962; Deschamps and Leulliette, 1984; GaliotouPanayatou and Macris, 1986).Galiotou-Panayatou et al. (1988)isolated a novel enzyme from Calvatea gigantea that hydrolyzes catechin, the building block of condensed tannin. This enzyme has been purified and characterized. The enzyme had an optimal pH of 8 and a temperature optimum of 35OC (Galiotou-Panayatou et al., 1988).Sambandam (1983) isolated catechin-2,3-dioxygenasefrom Chaetomium cupreurn, which cleaved catechin, releasing protocatechuic acid, catechol, and phloroglucinol carboxylic acid. The enzyme was active in the pH range of 2 to 8, with optimal pHs at 2.8 and 7.0.The enzyme was stable up to 50°C and had a molecular mass of 40,000daltons. Catechin oxygenase was a glycoprotein exhibiting two isoelectric points. Tannase does not act on condensed tannins (George and Sen, 1960). The biodegradation of condensed tannin by microorganisms remains an area that has not been understood or explored in detail. A lack of complete information on the enzymes involved in the hydrolysis of condensed tannin suggests that there is a need for further and continuing research in this area.

c.

SIGNIFICANCEOF TANNINS IN PLANTS

Tannins are believed to occur in the vacuoles of intact plant cells (Forsyth, 1964).Some plants accumulate tannin, particularly in the bark and heartwood. The exact function of tannins in plants is not clear. At the intracellular level of metabolism, tannins are of little value to the plant, although they have functions like wound healing or act structurally as pigments (White, 1957). Such accumulated tannins protect the vulnerable parts of the plants from microbial attack by inactivating viruses and invasive extracellular enzymes of microbes by direct tanning action (White, 1957).Enzymes secreted by attacking microorganisms are wholly or partially inactivated by complex formation with tannins (Goldstein and Swain, 1965),while such microbial substrates as polysaccharides and nonenzyme proteins, present in the plants, become highly resistant to microbial attack after binding to a tannin molecule (Betnoit et al., 1968).The plant, however, appears to protect itself from high concentrations of toxic phenols by the use of such specialized structures as the “glands” of gossypol (Singleton and Kratzer, 1969).The inhibitory action of tannins on the growth of bacteria (Sivaswamy, 1982; Henis et a]., 1964),fungi (Mur, 1953; Lewis and Papavizas, 1968),yeast (Jacob and Pignal, 1975),and viruses (Cadman, 1960)is well established.

222

P. K. LEKHA AND B. K. LONSANE

IV. Source of Tannase

A. MICROORGANISMS A number of microorganisms-including bacteria, fungi, and yeastshave been reported to produce tannase (Table I). Extensive screening studies have been conducted to select potent cultures for tannase production. Ganga et al. (1977, 1978) screened a number of fungi belonging to Penicillia and Aspergilli for production of tannase by submerged and liquid-surface fermentation. Among the different Aspergillus strains, A. oryzae, A. flavus, and A. japonicus were found to produce high titers of tannase compared to the other strains. Among the Penicillium spp., I? islandicum was reported to produce maximum tannase (Ganga et al., 1978). Tannase production by Penicillium strains was lower than that produced by Aspergillus species (Ganga et al., 1978). Yamada et al. (1967) conducted screening of about 80 fungal strains for tannase production and reported that two strains belonging to A. oryzae produced maximum tannase activity. The only commercial source of tannase available at present is produced by A. oryzae strain ATCC 9362. Reshetnikova et al. (1984) screened five species of Ascochyta (A. cucumeris, A. pisi, A. biochemica, A. boltshauseri, and A. viciae) for tannase production using a sucrose-mineral medium containing gallotannin. The highest tannase activity was produced by A. boltshauseri and A. viciae, whereas the lowest tannase activity was observed with A. pisi. Deschamps et al. (1983) isolated bacterial strains belonging to Bacillus, Corynebacterium, and Klebsiella that were able to degrade tannic acid and related compounds from decaying bark of pine (Pinus maritima) and oak (Quercus pedunculata). Among these strains, the best tannase production (0.064 U/ml) was observed in the case of Corynebacterium sp. Q 40 after 5 h of fermentation. These strains were able to produce tannase only in the presence of tannic acid in the medium (Deschamps et al., 1983). A similar strategy was adopted by Deschamps and Leulliette (1984) to isolate potent tannin-degrading yeasts. Aoki et al. (1976, 1979a,b) isolated several yeast-like strains from soil for tannase production. A strain identified as a Candida produced extracellular as well as intracellular tannase only in the presence of tannic acid. It is well known that product titers obtained in submerged fermentation need not be the same in solid-state fermentation. This has been observed in the cases of citric acid and a-amylase production by two different fermentation methods (Shankaranand et ul., 1992), thereby

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE TABLE I

MICROORGANISMS CAPABLE OF PRODUCING TANNASE Microorganism

Reference

Bacteria

Bacillus pumilis B. polymyxa Corynebacterium spp. Klebsiella pneumoniae

Deschamps et al., 1983 Deschamps et al., 1983 Deschamps et al., 1983 Deschamps et al., 1983

Yeast

Candida sp. K 1

Aoki et al., 1976, 1979a,h

Fungi

Ascochyta boltshauseri A. pisi

A . biochemica A . viciae Aspergill u s cam eus A. flaviceps A. f l a w s A. fumigatus A. japonicus A . leuchensis inui A. nidulans A . niger A. oryzae A. oryzae pseudoflaws A . parasiticus A. tamari A . terreus A. ustus Chaetomium lobosum Mucor pranii Myrothecium verrucaria Neurospora Rhizopus Trichothecium roseum Penicillium chrysogen u m I? fellutanum I? islandium I? notatum I? variable

Reshetnikova et al., 1984 Reshetnikova et al., 1984 Reshetnikova et al., 1984 Reshetnikova et al., 1984 Ganga eta]., 1977 Ganga etal., 1977 Chae and Yu, 1973 Lewis and Starkey, 1962 Ganga et a]., 1977 Rhind and Smith, 1922 Ganga et al., 1977 Nishira and Mugibayashi, 1960 Yamada et al., 1967 Seiji etal., 1973 Ganga et al., 1977 Vandamme et al., 1989 Ganga et al., 1977 Ganga et a]., 1977 Nishira and Mugibayashi, 1960 Nishira and Mugihayashi, 1960 Nishira and Mugibayashi, 1960 Nishira and Mugihayashi, 1960 Nishira and Mugibayashi, 1960 Nishira and Mugibayashi, 1960 Rajkumar and Nandy, 1983 Ganga et al., 1978 Ganga eta]., 1978 Ganga et al., 1978 Ganga eta]., 1978

223

224

P. K. LEKHA AND B. K. LONSANJZ

indicating a need for intensive screening to select a potent culture particularly suitable for a solid-state fermentation system. Eight strains of Penicillium, five strains of Aspergillus, three strains of Neurospora, and one strain each of Trichothecium roseum, Mucor pranii, Myrothecium verrucaria, and Chaetomium lobosum were screened for tannase production in a wheat bran medium containing 4% tannic acid as an inducer. Strains belonging to Aspergillus and Penicillium were found to be potent producers of tannase (Nishira and Mugibayashi, 1960). Lekha et al. (1993) also conducted extensive screening of fungal cultures for tannase production by solid-state fermentation and reported A. niger isolated by a baiting method to be the best tannase producer. Microorganisms used in the industrial production of food processing enzymes should be listed as GRAS (generally recognized as safe) (Fordham and Block, 1987). GRAS microorganisms are nonpathogenic, nontoxic, and should generally not produce antibiotics (Walsh and Headon, 1994). B. PLANTSAND ANIMALS

Tannase has been reported to be present in many tannin-rich plant materials, such as myrobolan ( Terminalia chebula) fruits, divi-divi (Caesalpinia coriaria) pods, dhawa (Anogeissuslatifolia) leaves and the bark of konnam (Cassia fistula), and babul (Acacia arabica) and avaram (Cassia auriculata) trees (Madhavakrishna et al., 1960). Its presence is also recorded in the rumen rnucosa of cattle (Begovic and Duzic, 1976). However, the levels reported are very low. Tannase was purified from bovine mucosa membrane of the rumen and small intestine (Begovic and Duzic, 1977). The gall larvae that undergo development in plant galls produce tannase to hydrolyze the tannic acid abundant in plant galls (Nierenstein, 1930). Although many enzymes are obtained from animal and plant sources (Godfrey, 1985), microorganisms are becoming the favored source for production of industrial enzymes because of their biochemical diversity and their technical and economic advantages (Underkofler, 1976). Microorganisms can be cultured in large quantities in a short time by established methods of fermentation. Thus, they can produce an abundant and regular supply of the desired enzyme. Microbial enzymes are more stable than analogous proteins obtained from plant or animal sources. Furthermore, microbes can be subjected to genetic manipulation more readily than animals and plants (Headon and Walsh, 1994; Walsh and Headon, 1994).

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

225

Genetic manipulation by mutation and selection has played a central role in increasing the yield of many enzymes produced by microorganisms (Walsh and Headon, 1994). Advances in the area of recombinant DNA technology have facilitated the development of an entirely new strategy to increase the levels of expression of specific enzymes (Walsh and Headon, 1994). For example, cellulase genes from Cellulomonas uda have been efficiently cloned and expressed in Zymomonas mobilis (Misawa et al., 1988). Large-scale fermentation of such recombinant organisms can yield appreciable quantities of any enzyme. Many commercially available enzymes are produced by this method (Cullen, 1987). However, no such studies have been conducted to improve tannase production.

c. PHYSIOLOGICAL SIGNIFICANCEOF TANNASE IN

PLANTS, ANIMALS, AND MICROORGANISMS

1. Plants

Tannase has been isolated from such hydrolyzable tannin-containing plant materials as myrobolan (Terminalia chebula) fruits, divi-divi (Caesalpinia coriaria) pods, and dhawa (Anogeissuslatifolia) leaves, as well as from condensed tannin-containing plant materials like avaram (CQSsia auriculata), babul (Acacia arabica), and konnam (Cassia fistula) bark (Madhavakrishna et al., 1960). Sometimes hydrolyzable and condensed tannin are produced in the same plant, but generally in separate tissues (Haslam, 1981; Haslam and Lilley, 1988). The physiological significance of tannase in plants has been discussed by Madhavakrishna et al. (1960). It was suggested that, together with large quantities of sugars, plants synthesize chebulinic acid, gallic acid, and hexahydroxyphenic acid during growth. As the fruit ripens, these acids may become esterified with glucose, with the help of tannase to form complex tannins. After abscission, the esterase hydrolyzes the tannins (Madhavakrishna et al., 1960). The structure of the condensed tannin molecule is such that it cannot be hydrolyzd by tannase (Dykerhoff and Ambruster, 1933). In the case of condensed tannins, tannase helps to synthesize, at one stage or another, some intermediates or precursors, which in turn undergo transformation into the complex tannin molecules (Madhavakrishna et al., 1960). 2. Animals

Tannins are present in a variety of plants utilized as feed (Salunkhe et al., 1982). The deleterious nutritional effects of dietary tannins have

226

P. K. LEKHA AND B. K. LONSANE

been reviewed (Price and Butler, 1980). They exert negative effects on protein, fat, and carbohydrate utilization of food (Salunkhe et al., 1982). Tannins complex with all the digestive enzymes and affect the digestion and availability of proteins. The secretions of the digestive tract contain mucoproteins (Buddecke, 1972), and tannins are known to react with these mucoproteins (Mitjavila et al., 1968). Any tannin that escapes this reaction can react with the proteins of the outer cellular layer of the gut and can reduce the passage of nutrients through the gut (Hand et al., 1966). The low levels of tannase reported to be present in the rumen mucosa of cattle (Begovic and Duzic, 1976) probably hydrolyze the tannic acid present in the diet to phenols and sugar, which can be readily absorbed (Glick and Joslyn, 1970), thereby alleviating the toxicity exerted by these compounds. 3. Microorganisms

Tannins are the fourth most abundant plant constituent, after cellulose, hemicellulose, and lignin (Swain, 1965). High tannin content in plant materials is associated with resistance to microbial attack, and the durability of certain long-lived trees and their wood has been ascribed to their high tannin content (White, 1957). Tannin accumulation in bark is particularly effective in preventing germination of spores of attacking fungi and penetration of fungal hyphae or such bacteria that rely on the action of extracellular enzymes to open up a pathway for the organism (White, 1957). Tannase is known to be produced by a number of microorganisms, including bacteria, yeasts, and fungi. Of these, the fungi are known to be the most potent enzyme producers (Lekha, 1996). Species of Ascochyta, phytopathogenic fungi, are also reported to produce tannase (Reshetnikova et al., 1984). Tannase produced by these microorganisms probably serves as a mode of invasion into the host plant by hydrolyzing these complex polyphenolic materials present in the bark of plants, which are known to confer protection against the attacking pathogenic fungi. Tannase present in soil microorganisms probably plays an active role in the decomposition and recycling of plant materials containing tannin. V. Tannase Assay

Tannase acts on the ester and depside linkages present in tannic acid, liberating glucose and gallic acid (Dykerhoff and Ambruster, 1933). Therefore, the activity of tannase can be measured by estimating the residual tannic acid (Deschamps et al., 1983) or gallic acid (Nicolson et

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

22 7

al., 1931) formed due to enzyme action. Most of the old methods were based on the titration of the gallic acid liberated by enzyme action (Freudenberg et al., 1927; Nishira, 1961; Haslam and Stangroom, 1966) and hence do not yield correct results due to the problem of determining the endpoint accurately (Madhavakrishna et al., 1960). Numerous photometric (Chen, 1969), colorimetric (Haslam and Tanner, 1970), and UV-spectrophotometric (Parmentier, 1970) methods have been described in the literature for the assay of tannase. The colorimetric and photometric methods are not specific (Jean et al., 1981). Madhavakrishna et al. (1960) described a procedure for quantitative estimation of tannase by determining the glucose liberated from tannic acid by the action of tannase. The reaction time used in this method is 24 h, so that it is not suitable for routine analysis of tannase. The spectrophotometric method developed by Iibuchi et al. (1967) has been used by most workers (Rajkumar and Nandy, 1983; Aoki et a]., 1976; Iibuchi et al., 1968). This method is based on the change in optical density of the substrate tannic acid at 310 nm. According to this method, one unit of enzyme activity is defined as the amount of enzyme that hydrolyzes 1 pmol of ester bond in tannic acid per minute. Yamada et d.(1968)used a modified version of this method using methyl gallate as the substrate. Haslam and Tanner (1970) were critical of the spectrophotometric method of Iibuchi et al. (1967), as it is based on an extremely narrow difference in UV absorption of gallic acid and its methyl ester or tannic acid, both of which have absorbance in the UV region. They developed a new spectrophotometric assay utilizing p-nitrophenyl esters of gallic acid as the substrate. However, this method did not find wide acceptance, probably due to the nonavailability of these substrates. Sanderson et al. (1974) described a method for tannase estimation that was based on a slight modification of the method of Iibuchi et al. (1967). This method is highly sensitive and is ideal for routine analysis of a large number of samples due to its short reaction time and continuous, direct monitoring of the reaction in the spectrophotometer. A major drawback of the method is that the enzyme activity cannot be expressed in SI units. Jean et al. (1981) developed a gas-chromatographic method for estimation of tannase activity by determining the gallic acid released as a result of enzymatic hydrolysis of methyl gallate. This method is said to be rapid, specific, reliable, and reproducible (Jean et al., 1981). Thomas and Murtaugh (1985) described a method for assay of tannase activity using an extract of black tea leaf as the substrate by a pH stat method.

228

P. K. LEKHA AND B.

K. LONSANE

VI. Production of Tannase

There are two main fermentation types that are generally used for production of commercial enzymes. These are submerged fermentation and solid-state fermentation (Frost and Moss, 1987). A literature survey indicates that tannase has been produced by liquid-surface, submerged, and solid-state fermentation, though production of tannase has been most extensively carried out in a submerged fermentation system. A. TANNASE PRODUCTION BY SUBMERGED FERMENTATION Submerged fermentation involves the growth of the microorganism as a suspension in a liquid medium in which various nutrients are either dissolved or suspended as particulate solids in many commercial media (Frost and Moss, 1987). Submerged fermentation is the preferred method for production of most of the commercially important enzymes, principally because sterilization and process control are easier to engineer in these systems (Aunstrup et al., 1979). Such details as media, fermentation time, temperature, and the location of tannase produced by different microorganisms in submerged fermentation are given in Table 11. No details on optimization of the processes described in Table I1 have been published, but it seems likely that fairly extensive optimization will be required to arrive at the medium and the conditions described. 1.

Media Composition

The fermentation medium must meet the nutritional requirements of the microorganism (Frost and Moss, 1987). It basically contains sources of carbon, nitrogen, minerals, and some growth factors, such as essential amino acids and vitamins (Volesky and Luong, 1985). Carbon Source. As tannase is an inducible enzyme, tannic acid itself was used as the sole carbon source as well as an inducer (Yamada et al., 1968; Aoki et a]., 1976). Additional carbon sources, like glucose (1”/0) and sucrose (3%), were used along with tannic acid for tannase production by A. oryzae (Fumihiko and Kiyoshi, 1975) and A. niger in submerged fermentation (Dhar and Bose, 1964), respectively. In fact, tannic acid concentration was found to be the crucial factor influencing the levels of enzyme. In submerged fermentation the concentration of tannic acid used ranged from 0.1 to 10% (Yamada et al., 1968; Nishira and Mugibayashi, 1956). In the case of A. oryzae, growth and tannase activity decreased when the tannic acid in the medium was increased from 0.5 to 1% (Ganga et al., 1977). A crude extract of gall nut powder

PRODUCTION AM) APPLICATION OF TANNIN ACYL HYDROLASE

229

TABLE I1 FERMENTATION CONDITIONS USEDFOR TANNASE PRODUCTION BY SUBMERGED FERMENTATION ~~

Microorganisms

Concentration

Time (h)

Temperature ("C)

Location of enzyme

Reference

Media

(Yo)

A. flavus

Tannic acid NaN03 MgSO4. 7Hz0 KC1 pH 6.0

0.1 0.2 0.05 0.005

96

30

A. niger

Tannic acid Sucrose NaN03

2 3 0.3

144

28

Dhar and Base 1964

70-120

30

Yamada et a/. 1967

48

30

Extracellular

Fumihiko and Kiyoshi, 1975

KzHP04 MgS04 KC1 FeS04 pH 6.5

A. o q m e

Tannic acid NHGI KHP04

MgSO4 7Hz0 AlC13 '6HzO pH 6.0

Extracellular

Yamada et a/., 1968

0.1

0.05 0.05 0.001 2

0.2 0.2 0.1 0.001

A. oryzae IAM 2636

Tannic acid Glucose NH4HzP04 KzHP04 MgS04. 7Hz0 pH 5.5

A. niger LCF.8

Tannin extract from gall nut powder 72.6% pH 4.5

29

33

Intracellular

Barthomeuf eta]., 1994

I! chrysogenum

Czapek-Dox medium + tannic acid 2%

120

28

Intracellular

Rajkumarand Nandy, 1983

1

6

40

Extracellular

Deschamps eta]., 1983

3 0.3

144

35

Extracellular

Aokiet al., 1976

NCIM 722

Bacillus SPP.

Candida sp. K 1

Commercial chestnut extract pH 6.8 Tannic acid NazHP0.1.12Hz0

KzHP04 MgS04. 7Hz0 Monosodium glutamate

2 1 1.4 0.2 0.05

0.3

0.05 1

230

P. K. LEKHA AND B. K. LONSANE

containing 72.6 g/liter of tannin, adjusted to pH 4.5, was also used as the culture medium for tannase production by A . niger in submerged fermentation (Barthomeuf et al., 1994). Nitrogen Source. The nitrogen sources used for tannase production include sodium nitrate (Yamada et al., 1968; Dhar and Bose, 1964), ammonium chloride (Vandamme et al., 1989), ammonium oxalate (Reshetnikova et al., 1984), and ammonium sulfate (Fumihiko and Kiyoshi, 1975). Reshetnikova et al. (1984) also reported that ammonium chloride was inhibitory for tannase production by Ascochyta spp. Organic nitrogen sources like monosodium glutamate, glutamic acid, and casein hydrolyzate have also been used for tannase production (Aoki et al., 1976; Lippitsch, 1961). Minerals. Traces of Fe, Zn, and Cu have been reported to be essential for the production of tannase by A. niger (Lippitsch, 1961), whereas Fe had no influence on tannase production in the case of Penicillium (Nishira, 1961).

pH Enzymes, being proteins, contain ionizable groups; consequently, the pH of the culture medium affects their structure and function (Frost and Moss, 1987). Most microbial extracellular enzymes are produced in greatest yield at a growth pH, somewhere near the pH for maximum enzyme activity (Volesky and Luong, 1985). In the case of tannase produced by submerged fermentation, the optimum initial pH used was in the acid range of 4.5-6.5 (Yamada et al., 1968; Fumihiko and Kiyoshi, 1975; Barthomeuf et al., 1994). However, most of the published work relates to shake-flask cultures run under uncontrolled conditions of pH, and the final pH values are rarely recorded. In the case of tannase production by A. niger in a 20-liter fermentor, pH was adjusted throughout fermentation by automatic regulation using aqueous ammonia or orthophosphoric acid (Pourrat et al., 1982). Maximum enzyme activity was obtained at pH 7. However, Barthomeuf et al. (1994) reported that, for tannase production by A. niger, at pH values below 3.5, the enzyme was unstable, and that substrate hydrolysis and diffusion of enzyme into the medium occurred at pH values above 5.5. 2.

3 . Sterilization

The most common method of sterilization for liquid, as well as for solid, media is by heat under pressure (Beckhorn et al., 1965). For liquid media in flasks and tubes, this is usually done at 120OC for 20 min. Tannic acid is thermolabile, so that the other medium components are usually dissolved in distilled water and sterilized at 120OC for 20 min.

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

231

Tannic acid dissolved in a minimal amount of water was filter sterilized and aseptically added to the rest of the autoclaved and cooled medium (Deschamps and Leulliette, 19841, or a tannic acid-containing medium was instantaneously sterilized separately at llO°C (Aoki et al., 1976).

Inoculum Preparation A highly developed production strain must be protected against the risk of degeneration, contamination, or loss of viability (Volesky and Luong, 1985). On a laboratory scale, spores are prepared by inoculating agar slants on suitable media by the microorganism, incubating the slants, and harvesting spores by adding 3-5 ml of water containing a non-ionic detergent (O.O1Yo],while aseptically scraping the spores from the agar surface (Mudgett, 1986). Tannase being an inducible enzyme, the maintenance medium for tannase-producing strains usually contains low levels of the inducer, tannic acid, along with other media components (Yamada et al., 1967; Aoki et al., 1976). It has been shown that the yield of tannase can be increased by growing the fungus for several successive asexual generations on gallotannin (Nicolson et al., 1931). Nishira (1961) also reported that the length of precultivation of Penicillium had an influence on the adaptation of the mold to the substrate tannin, when grown in a Czapek-Dox medium containing 0.5YO tannin. For tannase production by P chrysogenum, the inoculum was prepared by growing the culture on potato dextrose agar slants for 10 days (Rajkumar and Nandy, 1983). In the case of tannase production by Aspergillus and Penicillium strains in submerged fermentation, 2% inoculum was used (Ganga et al., 1977, 1978). A liquid medium was also used for inoculum production by Pourrat et al. (1982). Mycelia of A . niger weighing about 0.5 g were grown in 500 ml of Karrow medium (Prescott and Dunn, 1959) in which sucrose was replaced by 3% tannic acid and the pH was adjusted to 3.5. The medium was sterilized for 30 min at llO°C, incubated at 3OoC, and left undisturbed to form the mycelial mat. After 10 days, mycelium was harvested under sterile conditions, ground, and homogenized in 50 ml of sterile water. Two percent of this spore suspension, which had a spore density of 3 x lolo spores/ml, was used as inoculum for tannase production by submerged fermentation (Pourrat et al., 1982). A similar methodology was used for inoculum preparation by Barthomeuf et al. (1994) for tannase production by A. niger. 4.

5. Incubation Temperature

Enzyme fermentation is governed by the temperature, but the optimum for synthesis of a particular enzyme may differ from the optimum

232

P. K. LEKHA AND B. K. LONSANE

for growth (Frost and Moss, 1987). The optimum temperature for tannase production in most of the cases was found to be 30°C (Table 111, except in the case of bacteria and yeast, which required slightly higher temperatures of 40 and 35"C, respectively (Deschamps et al., 1983; Aoki eta]., 1976). 6. Aeration and Agitation

In laboratory-scale tannase production, shaken cultures in flasks (120 oscillations/min) were employed to ensure proper aeration and agitation (Yamada et al., 1968; Aoki et al., 1976). In the case of tannase production by A . niger in a 6-liter fermentor, the initial stirring rate was maintained at 300 rpm and then increased to 450 rpm after 24 h to offset an increase in the viscosity of the medium due to mycelial growth. The dissolved oxygen level was regulated at 3O-4O0h by means of an Ag-Pb electrode (Barthomeuf et al., 1994). They also reported that insufficient aeration impeded growth, while excessive aeration favored oxidation of tannins and had an inhibitory effect on the biosynthesis of tannase by A. niger. Small-jar fermentors provided with mechanical agitators (400 rpm) and air sparkers were used by Pourrat et al. (1982) for tannase production by A. oryzae. Constant airflow at 0.4 VVM was found to be better than dissolved oxygen at 30%, both in terms of growth and enzyme activity (Pourrat et al., 1982). In the cases of A. flavus and A. oryzae, a static condition enhanced tannase production, but stirred conditions favored enzyme production in the case of A. japonicus (Ganga et al., 1977). 7. Harvesting Time

Depending on the organism and the amount of tannase produced, fermentation may take 1 to 10 days. For tannase production by submerged fermentation, the fermentation time varied from 48 to 120 h in the case of A. oryzae (Okamura et al., 1988; Yamada et al., 1967; Fumihiko and Kiyoshi, 1975). In the case of Candida and A. niger, the fermentation was continued for 6 days (Aoki et al., 1976; Dhar and Bose, 1964). A shorter fermentation time of 29 h was also reported for tannase production by A. niger (Barthomeuf et al., 1994). Tannase produced by bacteria was released into the medium during the active growth phase, and maximum tannase was obtained after 5-6 h of fermentation (Deschamps et al., 1983). 8. Recovery of the Enzyme

In submerged fermentation, the next step after termination of fermentation is separation of microbial cells and suspended solids from the

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

233

liquid medium (Aunstrup, 1977). This is usually achieved by filtration in the case of mycelial organisms, or by centrifugation in the case of bacteria and yeast (Volesky and Luong, 1985). In the case of tannase from A. niger (Barthomeuf et a]., 1994) and P chrysogenum (Rajkumar and Nandy, 1983), the enzyme produced being intracellular, an additional step of sonication or grinding of the mycelium to release the enzyme is required. Barthomeuf et al. (1994) reported that, at maximum production (29 h), tannase was strongly bound to the mycelium, and not more than 5% of the enzyme could be released by chemical and physical methods (i.e., grinding the mycelium with sand and glass beads, polytrol grinding, osmotic shock, and sonication). Hence, 90-hour-old mycelium was used because, with the beginning of autolysis, the cell wall became fragile and recovery of enzyme was greatly increased. Tannase was obtained by physical disruption of the mycelium by congelation-decongelation and addition of concanavalin A into the medium. Addition of con A to the medium after decongelation increased tannase recovery from 22.5 to 34.5% by facilitating desorption of the enzyme from its binding site (Barthomeuf et al., 1994). Commercially, tannase is produced by submerged fermentation (Okamura and Yuasa, 1987). Table III gives a list of patents for tannase production using submerged fermentation. B. TANNASE PRODUCTION BY LIQUID-SURFACE FERMENTATION

Liquid-surface fermentation involves the growth of culture on the surface of a liquid medium at a shallow depth and held in a suitable container (e.g., stainless steel trays) (Mitchell and Lonsane, 1992). Only three preliminary reports (Pourrat et d., 1982; Ganga et al., 1977,1978) are available on the production of tannase by liquid-surface fermentation. In the case of tannase production by A. niger, the liquid medium containing all the necessary minerals along with carbon, nitrogen, and inducer was inoculated and left undisturbed so that a mycelial mat was formed on the surface. The enzyme was found to be intracellular, and tannase production was high when the initial pH of the medium was 7. The fermentation time was as long as 8 days (Pourrat et al., 1982). In the case of tannase production by different strains of Aspergillus, 25 ml of the medium was used in 100-ml Erlenmeyer flasks. Two percent tannic acid was used in the medium, and fermentation was continued for 7 days at 3OoC (Ganga et al., 1977). The inoculum level used was 2%.

234

P. K. LEKHA AND B. K. LONSANE TABLE 111 PATENTS FOR TANNASE PRODUCTION BY SUBMERGED FERMENTATION

1. Tannase produced by Aspergillus (Y. Fumihiko and M. Kiyoshi, Jpn. Pat. 7225,

786, 1975). 2.

Manufacture of tannase with Aspergillus (S. Okamura [Kikkoman Corp.] and K. Yuasa [Inabata and Co. Ltd.], Jpn. Pat. 62,272,973, 1987).

3. Fermentative manufacture of tannase (S. Okamura, K. Mizusawa, K. Takei, Y. Imai, and S. Ito [Kikkoman and Inabatal, Jpn. Pat. 63,304,981, 1988). 4.

Process for preparation of tannase intended for the production of gallic acid with Aspergillus (E. Vandamme, M. Jerome, A. Vermiera, and M. Maria, Eur. Pat. 339,011, 1989).

In the case of tannase produced by Penicillium spp. in liquid-surface fermentation, 0.5% tannic acid was used in the medium and fermentation was continued for 8 days (Ganga et al., 1978). More tannase was produced in liquid-surface fermentation as compared to submerged fermentation by A. flavus, A. oryzae, and I? islandicum (Ganga et al., 1977, 1978). However, the fermentation time required in all these cases was 7-10 days. These preliminary studies indicated that liquid-surface fermentation was not suitable for tannase production due to the longer fermentation time required and the intracellular nature of the enzyme. Other reasons cited for not adopting liquid-surface fermentation for production of other microbial metabolites include higher handling costs, the risk of infection, and the difficulty involved in applying modern methods of parameter and process control (Kumar and Lonsane, 1989).

c.

TANNASE PRODUCTION BY SOLID-STATE FERMENTATION

The essential feature of solid-substrate fermentation is the growth of microorganisms on an insoluble substrate without a free liquid phase (Mitchell and Lonsane, 1992). The moisture level in solid-substrate fermentation may be between 30 and 80%; for production of most enzymes it is typically in the region of 60% (Laukevics et al., 1984). The production of enzymes using solid-state fermentation has developed fkom the “koji process.” The traditional Japanese koji process involves growth of filamentous fungi (e.g., A. oryzae) on moist mixed substrates (e.g., rice, wheat, soybeans) to produce a mixture of extracellular amylolytic and proteolytic enzymes that were used in food pres-

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

235

ervation, flavoring, and texture modifications (Hesseltine and Wang, 1967; Yamada, 1977; Steinkraus, 1984). The literature on microbial production of tannase by solid-state fermentation is meager. Except for a few exploratory reports (Kita, 1917; Nishira, 1959a,b; Nishira and Mugibayashi, 1960) on production of tannase by solid-state fermentation using wheat bran, there are no data available on the effect of media parameters on tannase production by solid-state fermentation. The first report on tannase production by A. oryzae in a solid-state fermentation process was by Kita (1917). Tannase production increased with tannin content in the medium. Very high concentrations of tannin (20% of wheat bran) were found to inhibit the growth of the fungus (Kita, 1917). Nishira (1959a,b) reported the production of tannase by Penicillium spp. grown on wheat bran in the presence of gallotannin. The enzyme was produced only in the presence of tannic acid or gallic acid, and production was not growth-related. Gallotannin was completely degraded to gallic acid. Gallic acid decarboxylase was found to be co-produced along with tannase. Nishira and Mugibayashi (1960) studied the formation of tannase by different fungi in a wheat bran medium. Various strains of Penicillium, Aspergillus, Rhizopus, Neurospora, Trich0thecium rose um , Mucor pranii, Myrothecium verrucaria, and Chaetomium lobosum were cultured in a wheat bran medium containing 4% tannin and incubated at 2527°C for 3-10 days. Aspergillus and Penicillium species were found to produce higher titers of the enzyme, Tannase was leached from fermented bran using 1%sodium chloride, allowing a contact time of 24 h (Nishira and Mugibayashi, 1960). Work on tannase production by solid-state fermentation was initiated at the Central Food Technological Research Institute (CFTRI)in Mysore, India. This research yielded very promising results. Extensive screening of fungal cultures from a culture collection, as well as those isolated from soil and different tannin-rich plant materials, was carried out. A. niger PKL 104, a potent culture for tannase production, was isolated by a baiting method (Lekha et al., 1993). The physicochemical parameters for tannase production by this culture in solid-state fermentation was optimized by a response surface methodology (Lekha et al., 1994). Wheat bran was used as the solid substrate, and 6% tannic acid was added to the medium for maximum induction of tannase (Lekha, 1996). Addition of nitrogen and mineral salts did not have an effect on tannase production. This is because wheat bran provides mixed carbon, nitrogen, and other growth factors

236

P. K. LEKHA AND B. K. LONSANE

that can meet the nutritional requirements of the organism (Mudgett, 1986). The optimum pH and temperature for tannase production were 6.5 and 28OC, respectively. Moisure content is one of the critical parameters that affects enzyme production in solid-state fermentation. The optimum moisture content for tannase production was determined to be 62% (Lekha et a]., 1994). The inoculum level used was 5% (w/w), and the maximum amount of tannase was produced on day 3. Increased bed height was found to have a negative effect on tannase production, probably due to reduced aeration and increased heat buildup (Lekha, 1996).

Relevance of SSF for Tannase Production Solid-state fermentation offers a number of economic advantages over conventional submerged fermentation for enzyme production (Mudgett, 1986). The production medium is often simple, using agroindustrial by-products like wheat bran, rice bran, or wheat straw as the substrate (Mitchell and Lonsane, 1992). Because the moisture level is low, the volume of medium per unit weight of substrate is low; hence, enzyme activity is usually very high (Deschamps and Huet, 1985). Thus, to achieve a given enzyme productivity, fermentor volumes can be much smaller than in submerged fermentor systems (Mitchell and Lonsane, 1992). Additionally, effluent treatment requirements are also reduced (Moo-Young et al., 1983). Direct spore seeding is usually sufficient (Mitchell and Lonsane, 1992). Extraction of the enzyme from fermented bran with a small amount of water yields a relatively concentrated enzyme product (Frost and Moss, 1987).The inexpensive media, the high enzyme productivity, the easy enzyme recovery, and the reduced effluent treatment requirements involved in solid-state fermentation help in bringing down the cost of enzyme production (Ramakrishna et al., 1982). Solid-state fermentation has been largely neglected since World War Two in the West. Consequently, negligible research and developmental efforts have been put into the techniques involved in solid-state fermentation (Mitchell and Lonsane, 1992). No comparison of the economics was made between solid-state and submerged fermentation before submerged fermentation was selected for intensive development in the 1940s (Ralph, 1976). The technique of submerged fermentation in aerated stirred tanks, mainly developed for antibiotics, gained wide acceptance and was soon applied to the manufacture of industrial enzymes (Frost and Moss, 1987). It has been estimated that enzyme production by solid-state fermentation at present accounts for about only 5% of world total market enzyme sales (Frost, 1986).

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

237

Comparative studies of tannase production by A. niger using three different fermentation methods revealed that enzyme production with solid-state fermentation was 2.5 and 4.8 times higher compared to that for submerged and liquid-surface fermentation, respectively (Lekha and Lonsane, 1993). The fermentation time required to produce maximum enzyme was only 3 days, compared to 6 days for submerged and liquid-surface fermentation. The enzyme was completely extracellular in solid-state fermentation compared to the partly intracellular nature of the enzyme obtained from submerged and liquid-surface fermentation (Lekha and Lonsane, 1993). Tannase produced by solid-state fermentation was found to be more thermostable compared to the extraand intracellular tannase produced by submerged fermentation (Lekha and Lonsane, 1993). These studies clearly indicate the advantages of solid-state fermentation for tannase production. A number of commercial extracellular enzymes (e.g., protease, amylase, pectinase, amyloglucosidase, and rennin) are successfully produced by solid-state fermentation in Japan (Cannel and Moo-Young, 1980). A resurgence of interest has occurred in Western and European countries in response to the ever-rising demand for economy in the fermentation process (Steinkraus, 1984). Proper exploitation of this technique could lead to significant cost reductions in tannase production. The use of agroindustrial wastes as solid substrates will not only help to combat environmental pollution, but will also eliminate their treatment by cost-intensive waste-disposal techniques (Lonsane and Ghildyal, 1992). VII. Regulation of Tannase Biosynthesis

There are three modes of genetic regulatory mechanisms that affect the synthesis and secretion of an enzyme, viz., induction, catabolite regulation, and feedback regulation (Demain, 1971). A. REGULATION BY INDUCTION

In the case of inducible enzymes, an enzyme is synthesized only when the substrate or substrate analogue is present in the medium (Demain, 1971). Tannase is an inducible enzyme produced only in the presence of tannic acid or its end-product, that is, gallic acid (Knudson, 1913; Nishira and Mugibayashi, 1953). The minimum structural requirement for adaptive tannase formation is gallic acid (Nishira, 1959a). Kita (1917) reported that A. oryzae secretes tannase even when it is cultured on a substrate containing no tannin. However, in all the media

P. K. LEKHA AND B. K. LONSANE

238

reported for tannase production, tannic acid was present as an inducer. The minimum concentration of tannic acid that could stimulate formation of tannase was found to be 0.1% (Knudson, 1913). However, the mechanism of induction by tannic acid is not known. The large size and reactivity of tannic acid prevent uptake of the molecule through the cell membrane. The reaction of tannic acid with the cell wall has been reported to impair permeability (Herz and Kaplan, 1968). These facts suggest that tannic acid cannot be the inducing agent. The mechanism of induction is probably similar to that of cellulase (Singh and Hayashi, 1995),that is, the microorganism produces a basic level or a constitutive amount of tannase that hydrolyzes tannic acid to glucose and gallic acid, which can enter the microbial cell and function as an inducer. In fact, the rate of induction was faster when gallic acid was used as an inducer (Seiji et a]., 1973).

B.

REGULATION BY CATABOLITE REPRESSION

This involves the inhibition of the formation of certain enzymes by the catabolic products of the readily utilizable carbon source (Demain, 1971). The strongest repression is observed in a medium containing glucose (Eveleigh and Montenecourt, 1979). The repressed enzyme can be constitutive or inducible, but in most cases inducible enzymes are involved (Demain, 1971). However, in the case of tannase, no catabolite repression was observed. Knudson (1913) reported that the presence of 10% sucrose did not inhibit the secretion of tannase by A. niger, but the secretion of tannase was inhibited to a certain extent in the case of Penicillium spp. In addition, the presence of glucose did not repress tannase formation (Seiji et al., 1973). Another interesting observation was the shortening of the time lag for tannase production by glucose, ribose, and glycerol with tannic acid as the inducer. Transfer of the induced mycelium in a glucose-containing medium, without the inducer, caused an almost immediate cessation of tannase formation, indicating that the tannase mRNA of this strain is very unstable. Production of tannase resumed on addition of gallic acid (Seiji et d.,1973).

c. REGULATION BY FEEDBACK INHIBITION Feedback inhibition is the phenomenon by which the final metabolite of a pathway inhibits the synthesis of an enzyme, usually the first enzyme (Demain, 1971). No reports on feedback inhibition have been published in the case of tannase. In fact, gallic acid, which is the end-product, has been reported to induce tannase synthesis (Seiji et al.,

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

239

1973). Pyrogallol, which is also an end-product of tannin hydrolysis, does not stimulate enzyme formation (Knudson, 1913; Nishira, 1960). VIII. Location of Tannase

Tannase, when produced by submerged fermentation, has been reported to be intracellular in most fungi, for example, €? chrysogenum (Rajkumar and Nandy, 19831, A. flaws (Yamada et al., 1968), A. niger (Pourrat et al., 1982), and Aspergillus tamari (Vandamme et al., 1989). In the case of A. oryzae pseudoflavus, during growth in an inorganic salt medium containing an inducer, accumulation of both intra- and extracellular tannase was reported (Seiji et al., 1973). Okamura et al. (1988) reported that tannase produced by A. oryzae was extracellular. In yeasts and bacteria, it has been reported to be extracellular (Aoki et a]., 1976; Deschamps et al., 1983). Tannase produced by A. niger PKL 104 in submerged fermentation was completely intracellular when produced by submerged and liquidsurface fermentation during the initial 48 h of growth. Subsequently, the enzyme was secreted into the medium with progress of the fermentation (Lekha and Lonsane, 1993). However, tannase produced by the same culture in solid-state fermentation was completely extracellular throughout the course of fermentation (Lekha and Lonsane, 1993). IX. Purification of Tannase

Tannase has been purified from a variety of fungi, namely, A. flavlls (Yamada et al., 1968), A. oryzae (Iibuchi et al., 1968; Fumihiko and Kiyoshi, 1975), Candida spp. (Aoki et al., 1976), €? chrysogenum (Rajkumar and Nandy, 1983), and A. niger (Barthomeuf et al., 1994). The starting material was either culture filtrate (Fumihiko and Kiyoshi, 1975) or mycelial extract obtained by sonication of the mycelial cells (Yamada et al., 1968), depending on the localization of the enzyme. Tannase purification schemes have generally used standard column chromatographic techniques, mainly ion-exchange and gel filtration. The first step is usually ammonium sulfate (Yamada et al., 1968) or acetone precipitation (Beverini and Metche, 1990),which resulted in an initial concentration as well as purification. Aoki et al. (1976) reported the failure of ammonium sulfate to precipitate yeast tannase. Various other precipitating agents were tested, and it was found that rivanol (2-ethoxy-6,9-diaminoacridinium lactate) was suitable for precipitation of tannase from the culture broth. Precipitation of tannase using such

240

P. K. LEKHA AND B. K. LONSANE

polymers (1-90%) as polyethylene glycol, polyvinyl alcohol, and dextran has also been reported (Kazuo et al., 1973). The second step employed in most cases was ion-exchange chromatography (Rajkumar and Nandy, 1983; Yamada et al., 1968). Tannase is known to be an acidic protein (Adachi et al., 1968), and so an anion exchanger was used in all reported cases. DEAE-sephadex/cellulosewas used for purification of tannase from A. flavus, A . ozyzae, and l? chrysogenum (Yamada et al., 1968; Fumihiko and Kiyoshi, 1975; Rajkumar and Nandy, 1983). In the case of yeast tannase, the enzyme was adsorbed so strongly to DEAE-cellulose that, even with the use of a 0 . 9 4 4 phosphate buffer, the enzyme could not be eluted from the column. Therefore, ECTEOLA-cellulose,with a less exchangeable capacity, was used instead of DEAE-cellulose (Aoki et al., 1976). The enzyme was adsorbed to the ion-exchange column at an acidic pH (pH 5) and eluted from the column using gradient elution with increasing ionic strength of the buffer or salt (Iibuchi et al., 1968; Rajkumar and Nandy, 1983). The last step employed in tannase purification was gel-filtration chromatography (Rajkumar and Nandy, 1983; Yamada et al., 1968; Iibuchi et al., 1968). Since tannase is a high-molecular-weight protein (-200,000 daltons), sephadex G-200 was used by most workers (Yamada et al., 1968; Rajkumar and Nandy, 1983; Aoki et al., 1976). Iibuchi et al. (1968) used sephadex G-100 in the final tannase purification step. Affinity chromatography offers an extremely high-resolution method for purifying proteins and has been widely used in the laboratory (Volesky and Luong, 1985). It can significantly reduce the number of steps required to purify a protein due to its biospecificity (Cuatrecasas et ~ l .1968) , and has an excellent possibility of becoming a large-scale recovery technique (Clonis et al., 1986). However, this technique was not explored for tannase purification, except in one report on the separation of isozymes of tannase from A. oryzae. Tannase I, which has strong esterase activity, and tannase 11, which has strong depsidase activity, were separated on Con-A Ultrogel by elution with methyl-Dglucose at concentrations of 10 and 50 mM, respectively (Beverini and Metche, 1990). Barthomeuf et al. (1994) used a different protocol for purification of tannase from A. niger. Crude enzyme was obtained by physical disruption of the mycelium, employing congelation-decongelation and addition of concanavalin A into the medium. Insoluble materials were eliminated by centrifugation, and the supernatant was filtered through a 0.45-pm membrane. The filtrate was subjected to tangential ultrafiltration with a 200,000-dalton threshold membrane, followed by filtra-

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

241

TABLE IV SPECIFIC ACTMTESOF TANNASE BEFORE AND AFTER PURIFICATION ~~

Specific activity'

Source

Location

Crude extract

Purified enzyme

A. njger

Mycelium

68

1980

A. oryzae IAM 2636

Culture broth

6.8

77

A. oryzae

Culture broth

ND

A. j7avus

Mycelium

Final purification step High pressure size exclusion chromatography Sepbadex

Yield, Yo

15.1

4.0

G-100

Sephadex

78

lo3

135 x

lo3

Sephadex

F! chryso-

Mycelium

3.6

86

Condida spp.

Sephadex

48.2

Yamada et ol., 1968

18.5

Rajkumar & Nandy, 1983

7.3

Aoki e t a ] . ,

G-200

Culture broth

m

92

Sephadex G-200

Fumihiko & Kiyoshi, 1975 Iibuchi et a]., 1967

G-200

gen um

Barthomeuf et al., 1994

11.5

G-100 4.3 x

Reference

1976

'Specific activity in U/mg protein.

tion through a membrane with a molecular weight cutoff of 100,000 daltons. Addition of con A to the medium increased tannase recovery from 22.5 to 34.5%, probably by facilitating desportion of enzyme from its binding site (Barthomeuf et al., 1994). Table IV summarizes the recovery and specific activities of tannase purified from different microorganisms. X. Properties of Tannase

Tannases from many fungi have been extensively characterized. These results have led not only to an understanding of how these enzymes operate and are regulated, but also to an appreciation of their vastly different physicochemical properties.

242

P. K. LEKHA AND B. K. LONSANE

A. pH OPTIMUM AND STABILITY In general, tannase is an acidic protein with an optimum pH around 5.5 (Table V). In the case of A. niger and €? chrysogenum, the optimum pH was 6 (Barthomeuf et al., 1994; Rajkumar and Nandy, 1983). Tannases from yeast (Aoki et al., 1976) and A. oryzae (Iibuchi et a]., 1968) was stable in a broad pH range (3.5-8), whereas tannases produced by €? chrysogenum (Rajkumar and Nandy, 1983) and A. oryzae (Yamada et al., 1968) were stable in the narrow ranges of 4.5-6 and 5-5.5, respectively. B . TEMPERATURE OPTIMUM AND STABILITY

The temperature optimum reported for tannase activity was around 30°C in the case of A. oryzae (Iibuchi et d., 19681, I? chrysogenum (Rajkumar and Nandy, 1983), and A. niger (Barthomeuf et ~ l .1994), , except for 60 and 50°C in the cases of A. ~ ~ Q V U(Yamada S et al., 1968) and yeast tannase (Aoki et a]., 19761, respectively. Tannase from A. niger, A. oryzae, and €? chrysogenum were stable up to 30°C (Table V), and A. flavus tannase was stable up to 60°C for 10 min (Yamada et al., 1968). C. MOLECULAR MASS

Tannase is a high-molecular-weight protein whose molecular weight is reported to vary from 186,000 to 300,000 daltons, depending on the strain (Table V). The native enzyme consists of two different polypeptide chains (subunits) of similar molecular size (Adachi et al., 1968; Aoki et al., 1976; Rajkumar and Nandy, 1985). Amino acid analysis of tannase from Candida revealed that the enzyme consisted of 786 amino acid residues per molecule (Aoki et d., 1976). Alanine and arginine were identified as the N-terminal amino acids in the case of tannase from A. flavus (Yamada et al., 1968) and €? chrysogenum (Rajkumar and Nandy, 1985).The C-terminal amino acids lysine and glutamic acid were present only in €? chrysogenum tannase (Rajkumar and Nandy, 1985). The enzyme is a typical serine esterase and was completely inactivated by phenyl methyl sulfonyl fluoride (Rajkumar and Nandy, 1983). (D132P)tannase Radio-isotope studies using diisopr~pyl-~~P-phosphoryl revealed that the tannase molecule contains one essential serine (Adachi et al., 1971). The amino acid sequence around active serine was

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

similar to that of subtilisin, composed of -Thr-Ser-Methal., 1971).

243

(Adachi et

D. ENZYMEINHIBITION Tannase from A. niger was inactivated by 0-phenanthroline, phenyl methyl sulfonyl fluoride, ethylene diamine tetraacetic acid (EDTA), 2-mercaptoethanol, and sodium thioglycolate (Barthomeuf et al., 1994). Tannase fiom A. oryzae was also completely inactivated after dialysis against an EDTA solution (Iibuchi et al., 1968), whereas no inhibition was observed with such chelating agents as EDTA and O-phenanthroline and such SH agents as parachloromercuric chloride in the case of tannase from A. flavus, but it was completely inactivated by diisopropyl fluorophosphate (DFP) (Yamada et al., 1968). Yeast tannase was also not inhibited by EDTA (Aoki et al., 1976). Tannase from A . oryzae (Iibuchi et al., 1968) and I! chrysogenum (Rajkumar and Nandy, 1983) was most inhibited by Zn2+,Cu2+,and Fez+.Tannase from A. niger was strongly inhibited by copper and to a lesser extent by ferric and zinc ions at concentrations of 20 mM. This inhibition could be strongly attenuated by adding 20 mM EDTA into the reaction medium (Barthomeuf et al., 1994). Tannase from A. oryzae was inhibited competitively by substrate analogues like n-propyl gallate and isoamyl gallate, which have phenolic hydroxyl groups, except for 2,6-dihydroxy benzoic acid, which inhibits noncompetitively. Therefore, the binding site of tannase may be able to react with any phenolic hydroxyl group, although the substrate forming a true enzyme-substrate complex must be an ester compound of gallic acid (Iibuchi et al., 1972). Pyrogallol, gallic acid, gallaldehyde, and gallamide were reported to be competitive inhibitors for tannase produced by A. niger, with Kivalues of 217, 661, 1510, and 175 pM, respectively (Parmentier and Verbruggen, 1973). E. K, OF TANNASE The K, of tannase produced by I! chrysogenum was reported to be 0.48 x lo4 M when tannic acid was used as the substrate (Rajkumar and Nandy, 1983). The K,,, values of tannase produced by A. flavus were 0.5 x lo4 M for tannic acid, 1.4 x lo4 M for glucose-1-gallate, and 8.6 x lo4 M for methyl gallate (Yamada et al., 1968). The K,,, values for different substrates revealed that natural tannic acid is the best substrate for the enzyme and that methyl gallate is less reactive than tannic acid (Yamada et al., 1968).

P. K. LEKHA AND B. K. LONSANE

244

TABLE V

PROPERTIES OF Pmnm TANNASE Temperature, "C

PH

Microorganism

Properties

Opt.

Stab.

Opt.

Stab.

IEP

Mol. wt.

A. f l a w s IF0 5839

5.5

5-5.5

50-60

60

4

192,000

A. niger LCF 8

6.0

3.5-8

35

4-45

4.3

186,000

A . ozyzae

5.5

3-7.5

30-40

30

200,000

Asp. sp. AN 11

5.5

5-6.5

3040

30

200,000

Penicilliurn chzysogen urn

5-6

4.5-6

30-40

30

300,000

Candida sp. K 16

6

3.5-7.5

50

40

250,000

Opt = optimum: Stab = stability; IEP = isoelectric point.

F. CARBOHYDRATE CONTENT

All the fungal tannases reported thus far are glycoproteins. The carbohydrate content of tannase is relatively high, primarily consisting of such neutral sugars as mannose, galactose, and hexosamines (Aoki et al., 1976). l? chrysogenum tannase contained 66.2% neutral sugars and

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

245

TABLE V CONTINUED ~

Inhibitors (% inhition)

~

~

CarboInactivators

DFP (98%)

Protein content

80%

hY-

25.4%

hexose C U S O (68%) ~ ZnClz (39%) FeC13 (13%) DFP (83%)

O-phenanthroline PMSF, EDTA 2 mercaptoethanol sodium thioglycolate MgSO4, CaClz Mnc12, COClZ

Zn (70%) CU (70%)

EDTA

43%

sugar

Yamada et al., 1968 Adachi et al., 1971 Barthomeuf eta]., 1994

Iibuchi et al., 1968, 1972

Chae et al.,

cuc12 ZnClz CU (53%) Zn, Fe (45%) Mn (22%) Mg (17%)

Reference

drate

1983

PMSF

32%

66.2%

neutral sugar,

Rajkumar and Nandy, 1983,1985

1-9%

hexosamine 35%

61.9%

Aoki et al.

neutral sugars hexosamine

1976

2.2%

1.9% hexosamine (Rajkumar and Nandy, 1985). Yeast tannase contained 61.9% hexose and 2.2% hexosamine (Aoki et ~ l .19761, , whereas tannase from A. flavus had a comparatively smaller (25.4% hexose) carbo, hydrate content (Adachi et ~ l .1968).

The polypeptide moiety was relatively small and varied from strain S to strain, ranging from 12.5% (nitrogen) in the case of A. ~ ~ Q V U(Yamada

246

P. K. LEKHA AND B. K. LONSANE

et a]., 1968) to 38% (protein) in the case of yeast tannase (Aoki et al., 1976). The biological significance of such a high carbohydrate content is not known. It is well known that tannins associate strongly with proteins by hydrogen bond formation between the phenolic hydroxyl group of the tannin and the carboxyl group of the protein peptide bonds, forming insoluble precipitates (Canon, 1955). Fungal tannases not only escape the inhibitory effects of hydrolyzable tannins, they also render tannin innocuous by cleaving it into inert products like glucose and gallic acid (Strumeyer and Malin, 1970). It was suggested that the carbohydrate coating most probably protects the polypeptide backbone, which would then be less accessible to the tannin molecule. The carbohydrate coating on tannase not only protects the enzyme but also directs the substrate to the limited, but accessible, active-site region, where cleavage to non-tannin products can be accomplished (Strumeyer and Malin, 1970). XI. lsozymes of Tannase

Toth and Barsony (1943) reported that gallotannin-decomposing tannase contains two separate enzymes-an esterase and a depsidasewith specificities for ester linkage and m-digallic acid ester linkages, respectively (Fig. 4). Tannase is composed of a mixture of both (Toth, 1944). Among these enzymes, gallic acid esterase is predominant. The identities of these two enzymes were subsequently confirmed by Haslam et al. (1961).The esterase and depsidase were fractionated using column chromatography, and the two enzymes differed slightly with respect to their pH optimum and stability (Haslam and Stangroom, 1966). Beverini and Metche (1990) fractionated the two isozymes of tannase by affinity chromatography on Con-A Ultrogel. Tannase I has esterase activity, while tannase I1 had a strong affinity for m-digallic acid and pyrogallol derivatives containing depside groups. The values of K, for m-digallic acid and methyl gallate were 0.7 and 6.2 mM, respectively, for tannase 11, whereas the affinities of tannase I for ester and depside substrates were very similar (2.0 &for m-digallic acid and 1.7 mMfor methyl gallate). The two isozymes also differed in terms of their carbohydrate and polypeptide contents (Beverini and Metche, 1990). XII. Mode of Action of Tannase

An extensive study on the pathway of hydrolysis, substrate specificity, and inhibition of tannase produced by A. oryzae was carried out by

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

247

'esterase' HO OH

FIG.4. Esterase and depsidase activities of tannase.

Iibuchi et al. (1972). Tannic acid (Fig. 5(I)) was found to be completely hydrolyzed by the enzyme to gallic acid and glucose through 2,3,4,6tetragalloyl glucose (Fig. 5(III)) and two kinds of monogalloyl glucose (Fig. 5(IV)). The pathway of hydrolysis of tannic acid by the enzyme is discussed later in this chapter. The position of gallic acid in the two kinds of monogalloyl glucose has not been determined. Nishira (1962) reported the formation of an intermediate compound, that is, glucose-gallic acid, which on hydrolysis produced gallic acid and glucose. Deschamps et al. (1983) observed the formation of two intermediate products during degradation of tannic acid by Bacillus pumilis. These intermediate products were considered to be di- and trigallic acids, probably bonded to glucose. However, these intermediates were not found when other strains (e.g., Bacillus polymyxa, Klebsiella pneumoniae, and Corynebacterium spp.) were employed. Rajkumar and Nandy (1986) reported that l? chrysogenurn metabolizes tannic acid and chebulinic acid with the formation of an intermediate compound, which is subsequently converted to gallic acid. The intermediate product had an absorbance at 277 nm and produced gallic acid on hydrolysis with sulfuric acid. l? chrysogenum metabolizes gallic acid, producing oxalic acid as the final product (Rajkumar and Nandy, 1986). It was found that pyrogallol, an inhibitor of l? chrysogenum, was not metabolized by the organism, but tannic acid and gallic acid were metabolized after 3 and 4 days, respectively, in the presence of pyrogallol. Watanabe (1965) studied the decomposition of gallic acid by A. niger using fluoroacetate as the inhibitor. He reported that citric acid, cis-aconitic acid, and a-keto glutaric acid were the degradation products. Inhibition studies using 0.002 A4 fluoroacetate revealed that l? chrysogenum can metabolize gallic acid following a different metabolic path-

T

P Gallic acid -9

+

U

RI

Glucose

FIG.5. Hydrolyzing pathway of tannic acid by tannase.

!% m

F r 8 4

E n

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

249

way, where citric acid is the only product of metabolism (Rajkumar and Nandy, 1986). XIII. Immobilization of Tannase

With an understanding of the nature of industrial enzymes and their catalytic potential, the use of these enzymes has gradually been extended to a variety of fields, including food processing, brewing, pharmaceuticals, medicine, textiles, and detergents (Kennedy et al., 1987). Despite the advantages of using enzymes, their commercial use has been limited because (a) most of them are relatively unstable, (b) the cost of enzyme isolation and purification is still high, and (c) it is technically expensive to recover active enzyme from the reaction mixture after completion of the catalytic process (Kennedy et al., 1987). These problems can be solved by immobilizing the enzyme. Indeed, immobilized enzymes bypass many of the shortcomings that frequently make soluble enzymes commercially impractical (Trevan, 1980). Their superiority comes from cheaper cost due to reuse, improved stability, prevention of allergenic responses, or other undesirable interactions between the enzyme and specific components of the reaction mixture, including contamination of the final product by the enzyme (Trevan, 1980). The first report on the immobilization of tannase was by Weetal and Detar (1974). Tannase from a commercial source was covalently attached to alkylamine porous silica activated with glutaraldehyde. A comparative study on kinetic parameters, pH, temperature, and operational stability was carried out. The optimum pH of tannase was slightly shifted to the alkaline region and thermal stability was increased after immobilization (Weetal and Detar, 1974). Tannase from A. niger was immobilized according to this method and used for enzymatic gallic acid esterification in organic solvents (Weetal, 1985a). A comparison of the synthesis catalyzed by soluble and immobilized tannases indicated a difference in obtained yields. In the enzymatic synthesis of propyl gallate using immobilized tannase, the yield of propyl gallate was 41.4% with respect to gallic acid, compared to 3.5% when soluble tannase was used (Weetal, 1985b). In a soluble system, maximum synthesis was observed with methanol (40%), while in an immobilized system maximum synthesis (85%) was observed with amyl alcohol (Weetal, 1985a). In a patent on green tea conversion using tannase and natural tea enzymes, Sanderson et al. (1974) used immobilized tannase prepared by a technique employing diazo-coupling onto glass beads. Coggon et al. (1975) also described a process for production of instant tea using

250

P. K. LEKHA AND B. K. LONSANE

tannase immobilized on glass beads by diazo-coupling. The immobilized tannase was capable of solubilizing between 60 and 80% of the tea cream in tea extracts using either a batchwise or continuous-column treatment process. The immobilized tannase was shown to undergo little loss of enzyme activity on extended use. The enzyme can be readily separated from the reaction mixture in this technique, used again, and provides a means for complete removal of enzyme from the material undergoing treatment (Sanderson et al., 1974). Katwa et al. (1981)described a procedure for the assay of immobilized tannase with polyacrylamide, collagen, and Duolite as matrices, based on spectrophotometric determination of gallic acid formed in the enzymatic hydrolysis of tannic acid. The kinetic parameters of the enzymatic reaction have been studied and an assay procedure formulated. This method was found to be much more accurate than those reported previously. XIV. Applications of Tannase

Tannase is extensively applied in the food, feed, beverage, brewing, pharmaceutical, and chemical industries. A. INSTANTTEA This is the most promising application of tannase (Sanderson et a]., 1974). An important requirement of instant tea is cold-water solubility, as instant tea is mostly used to prepare iced tea (Coggon et al., 1975). The presence of tea cream, a cold water-insoluble precipitate that forms naturally in brewed tea beverages when allowed to stand for a few hours at or below 4"C, is therefore a major problem in instant tea manufacture (Sanderson, 1972). Tea cream is a hydrogen-bonded complex of the polymeric black tea polyphenols (i.e., thearubigins and theaflavins) with caffeine. The galloyl groups on the black tea phenols are involved in complexation with caffeine (Wickremansinghe, 1978). In the conventional process for preparing instant tea, the hot-water extract of tea is subjected low temperatures with agitation, followed by centrifugation of the tea cream (Nagalakshmi et al., 1985). The tea cream is usually discarded, which represents a considerable loss of the major flavor components (Coggon et a]., 1975). The chemical method for solubilizing tea cream involves treatment with sulfite and molecular oxygen together with an alkali (Sanderson, 1972). Instant tea powder produced by chemical methods, when recon-

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

(tannase + H,O)

OH

ROH

+

251

\ I

HO " > C O O H OH Gallic acid

FIG.6. Deesterification of tea polyphenols by tannase.

stituted as a hot beverage, reacts very badly on addition of milk by taking on a dull-blackish, unpleasant coloration (Coggon et a]., 1975). A patent assigned to Tenco Brooke Bond Ltd. (1971) describes the process for the preparation of instant tea from black tea. The enzymetreated sample at 5% was slightly hazy and had an undissolved solid content of 1.35%,while an untreated sample at 5°C was visually opaque and had an undissolved solid content of 7.5%. A process for the preparation of cold water-soluble instant tea from fresh green tea flush that uses tannase in a preconversion treatment was described by Sanderson et al. (1974). Black tea of high quality and good color that yields a good milk reaction and a substantially reduced level of cold water-insoluble solids was obtained by contacting green tea with tannase, followed by conversion of the green tea to black tea.

Action of Tannase on Tea Polyphenols Tannase catalyzes the hydrolysis of the ester linkages between galloyl groups and various compounds present in unconverted tea leaves (Sanderson et al., 1974). The reaction is a deesterification (Fig. 6), where R-OH can be epicatechin and epigallocatechin. This deesterification enhances the natural levels of gallic acid and epicatechin in nonconverted green tea leaf material. This favors the formation of large amounts of epitheaflavic acid during the tea conversion process on the tea leaf material, which has undergone preconversion tannase treatment. Epitheaflavic acid is responsible for the bright reddish-black tea-like color and has very good cold-water solubility (Nagalakshmi et al., 1985). Further deesterification of green tea leaf constituents prevents the formation of any gallated tea oxidation products by eliminating the precursors of these compounds, which are normally present in black tea infusions. Therefore, elimination of such poorly soluble compounds is probably important for producing instant tea with good color and solubility, and for obtaining a good yield when the green tea conversion process is

252

P. K. LEKHA AND B. K. LONSANE

carried out after preconversion tannase txeatment (Sanderson et al., 1974). A continuous process for solubilizing tea cream by passing a hot-water extract of black tea through a column containing immobilized tannase was also described by Coggon et al. (1975).A number of patents have been granted for manufacture of instant tea (Table VI). B. BEERCHILLPROOFING

Masschelein and Batum (1981) reported that tannase produced by a strain of A. f l a w s significantly reduced chill haze formation in beer. Discoloration and haze development during beer storage could be prevented by enzymic hydrolysis of wort phenolics with tannase and lactase (Rossi et al., 1988).

C. WINEMAKING Yamada and Tanaka (1972) described the use of tannase in wine making. The enzyme hydrolyzed chlorogenic acid to caffeic acid and quinic acid, which favorably influenced taste. Chae et al. (1983) explored the potential of tannase in the manufacture of acorn wine. Korean acorns (Quercus spp.) contain 6.5-7.5% tannic acid (Chae and Yu, 1973). Acorn wine was produced from a koji of rice powder and acorn powder (1:l)using an Aspergillus strain. The final wine had an ethanol content of lo%, a reducing sugar content of 7%, and a pH of 4.0 (Chae et al., 1983). Tannase was used along with lactase to treat grape juice and grape musts so as to remove phenolic substances for stabilization of the beverage (Cantarelli, 1986; Cantarelli et al., 1989). D. PRODUCTION OF GALLICACID

Tannase hydrolyzes tannic acid to glucose and gallic acid (Iibuchi et d . , 1972). As early as 1901, Fernback implicated the industrial application of the enzyme in the manufacture of gallic acid using tannin-containing substances as raw materials. Gallic acid is mainly used as a synthetic intermediate for the production of pyrogallol and gallic acid esters used in the food and pharmaceutical industries (Deschamps and Lebeault, 1984). In terms of pharmaceutical production, gallic acid is used in the synthesis of trimethoprim. At present, gallic acid is made industrially by chemical hydrolysis of naturally occurring gallotannins. Deschamps and Lebeault (1984) reported the production of gallic acid from tara tannin by using Klebsiella pneumoniae and Corynebacterium spp., where the yield of gallic acid

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

253

TABLE VI FOR TEACREAM SOLUBILIZATION USINGTANNASE PATENTS

1. Tenco Brooke Bond Ltd., U.K. Pat. 1,249,932, 1971. 2.

G. W. Sanderson, N. J. Englewood, P. Coggon, and N. Y. Orangeburg, Green tea conversion using tannase and natural tea leaves, US. Pat. 3,812,266, 1974.

3.

P. Coggon, H. N. Graham, and G. W. Sanderson, Cold water-soluble tea, U.K. Pat. 1,380,135, 1975.

4.

P. Coggon, H. N. Graham, A. C. Hoefler, and G. W. Sanderson (Unilever NV), Tea leaves extractable in cold water, Ger. Pat. 2,610,533, 1976.

5.

Y. Takino (Coca-Cola),Enzymic solubilization of tea cream, U.S. Pat. 3,959,497, 1976.

6. P. Coggon, H. N. Graham, A. C. Hoefler, and G. W. Sanderson, Tea, U.K. Pat. 1,546,508, 1979. 7. C. H. Tsai (Procter and Gamble), Enzymic treatment of black tea leaf, Eur. Pat. 135,222, 1985.

was 55%. Pourrat et al. (1985) described a method for production of gallic acid from tara tannin and sumac tannin using a strain of A. niger. The total yield of gallic acid was 30% at 45 h in the case of tara tannin and 9.7% in the case of sumac tannin, in terms of raw material weight (Pourrat et al., 1987). The low yield was due to biodegradation of the gallic acid produced by these strains. Use of mutant strains lacking the capacity to degrade gallic acid or use of immobilized tannase will result in high yields of gallic acid (Deschamps and Lebeault, 1984). Further studies in this direction are needed because, from an economical point of view, as the chemical process is not expensive, a biotechnological process has to be more selective and produce better yields. Gallic acid is also used in the enzymatic synthesis of propyl gallate, which is mainly used an an antioxidant in fats and oils, as well as in beverages (Weetal, 1985a). A patent was granted (Weetal, 198513) for the synthesis of propyl gallate using tannase immobilized on porous silica. E. ANIMALFEEDADDITIVES

The use of a number of enzymes in animal feed is gaining in importance (Berry and Paterson, 1990). The antinutritional effects of tannins are well known (Singleton and Kratzer, 1969). They are present in a variety of plant materials that are used as feed (Bate-Smith and Rasper,

254

P. K. LEKHA AND B. K. LONSANE

1969). The use of tannase in the pretreatment of tannin-containing feed may prove beneficial in removal of these undesirable compounds and also improved digestibility. Tannins form insoluble complexes with proteins (Swain, 1965) and are present in a wide variety of plant materials that are used as foods and feed (Bate-Smith and Rasper, 1969). The interaction of tannins with protein plays an important role in nonruminants (Tamir and Alumot, 1970). They exert their antinutritional effects by complexing with dietary and endogenous proteins, as well as with digestive enzymes, thereby interfering with normal digestion, leading to a drain of highquality proteins from the body (Salunkhe, et al., 1982). Tannins are also known to interfere with the absorption of iron (Lauren and Lee, 1988). Other deleterious effects of tannins include damage to the mucosal lining of the gastrointestinal tract, alteration in excretion of certain cations, and increased excretion of certain proteins and essential amino acids (Singleton and Kratzer, 1969). The use of tannase as an ingredient of animal feed would improve the digestibility of the feed. F. MISCELLANEOUS

Tannase may also find use in cosmetology to eliminate the turbidity of plant extracts and in the leather industry to homogenize tannin preparation for high-grade leather tannins (Barthomeuf et al., 1994). Tannase may also find potential application in the food industry for improving food quality by removing undesirable substances (Barthomeuf et al., 1994). Tannase could possibly find application in various fields, including animal nutrition and bioconversion of plant materials. XV. Conclusions and Future Prospects

Laboratory-scale studies on the production of tannase by solid-state fermentation showed several advantages over conventional submerged fermentation (Lekha et al., 1994). Proper exploitation of this technique would enable a significant reduction in the cost of enzyme production. Scale-up studies are essential to establish the feasibility of the process at an industrial scale. The advent of recombinant DNA technology has revolutionized research in the field of enzymology. Although tannase has been produced and purified from a number of microbial sources, isolation and characterization of tannase genes would be essential to understand the molecular biology of tannase biosynthesis. The role of the structural and

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

255

regulatory genes involved in tannase biosynthesis is also worth investigating in order to successfully clone and overproduce this enzyme in the desired host. All the tannases reported so far are inducible enzymes, and enzyme production was found to increase with increased tannic acid concentrations. Consequently, high concentrations of tannic acid have to be used for maximum enzyme induction. Genetic improvement of the strain to produce constitutive mutants that can synthesize tannase in the absence of tannic acid would be highly desirable. From an applications point of view, it would be desirable to immobilize the enzyme to enable its widespread use in food and beverage processing. An equally important aspect would be improvement in the thermostability of tannase. Tannase is known to have synthetic ability in nonaqueous media (Weetal, 1985a,b). This could be exploited to synthesize very specific gallic acid esters, which are used as antioxidants in the food processing industry. Some potential applications mentioned in the literature still require further research to emphasize their worth and economic feasibility so as to encourage the use of tannase on a commercial scale. REFERENCES Adachi, O., Watanabe, M., and Yamada, H. (1968). Agric. Biol. Chem. 32(9), 1079-1085. Adachi, O., Watanabe, M., and Yamada, H. (1971). Acta Biotechnol. 49(3), 230-234. Aoki, K., Shinke, R., and Nishira, H. (1976). Agric. Biol. Chem. 40(1), 79-85. Aoki, K., Tanaka, T., Shinke, R., and Nishira, H. (1979a). J. Chromatogr. 17,446-448. Aoki, K.,Kajiwara, S., Shinke, R., and Nishira, H. (1979b). Anal. Biochem. 95(2), 5 75-578. Aunstrup, K. (1977). In “Biotechnological Applications of Proteins and Enzymes” (Z. Bovak and N. Sharon, eds.), pp. 39-49. Academic Press, New York. Aunstrup, K., Anderson, O., Falch, E. A,, and Nielsen, T. K. (1979). In “Microbial Technology” (H. J. Peppler and D. Perlman, eds.), Vol. 1, pp. 281-310. Academic Press, New York. Barthomeuf, C., Regerat, F., and Pourrat, H. (1994). J. Ferment. Bioeng. 77(3), 320-323. Bate-Smith, E. C., and Rasper, V. (1969). 1.Food Sci. 34, 203-209. Beckhorn, E. J., Labbee, M. D., and Underkofler, L. A. (1965). 1.Agric. Food Chem. 13, 30-34. Begovic, S., and Duzic, E. (1976). Vetenaria 25,421-428. Begovic, S.,and Duzic, E. (1977). Vetenaria 26, 227-233. Berry, D. R., and Paterson, A. (1990). In “Enzyme Chemistry: Impact and Applications” (C. J. Suckling, ed.), pp. 306-351. Chapman and Hall, New York. Betnoit, R. E., Starkey, R. L., and Basaraba, J. (1968). Soil Sci. 105, 153-158. and Metche, M. (1990). Sci. Aliment. 10(4), 507-516. Beverini, M., Bradfield, A. E., and Penny, M. (1948). 1.Chem. SOC.,pp. 2249-2254.

256

P. K. LEKHA AND B. K. LONSANE

Buddecke, E. (1972). In “Glycoproteins: Their Composition, Structure and Function” (A. Gottschalk, ed.), 3rd ed., part A, p. 535. Elsevier, London. Cadman, C. H. (1960). In “Phenolics in Plant Health and Disease” (J. B. Pridham, ed.), pp. 101-105. Pergamon, New York. Cannel, E., and Moo-Young, M. (1980). Process Biochem. 15, 2-7. Canon, C. G. (1955). Mikrochim. Acta, 2, 555-588. Cantarelli, C. (1986). Vini. Ital. 28(3), 87-98. Cantarelli, C., Brenna, O., Giovanelli, G., and Rossi, M. (1989). Food Biotechnol. 3(2), 203-2 13.

Chae, S. K., and Yu, T. J. (1973). Hanguk Sipkum Kwahakhoechi 5(4), 258-267. Chae, S. K., Yu, T. J., and Kum, B. M. (1983). Hanguk Sipkum Kwahakhoechi 15(4), 333-334.

Chen, T. (1969). Ph.D. Thesis, University of California. Clonis, Y. D., Jones, K., and Lowe, C. R. (1986). J. Chromatogr. 363, 31-36. Coggon, P., Graham, N. H., and Sanderson, G. W. (1975). U.K. Pat. 1,380,135. Coggon, P., Graham, N. H., Hoefler, A. C., and Sanderson, G. W. (1976). Ger. Pat. 2,610,533. Coggon, P., Graham, N. H., Hoefler, A. C., and Sanderson, G. W. (1979). U.K. Pat. 1,546,509.

Cuatrecasas, P., Wilcheck, M., and Afinsen, C. B. (1968). Proc. Natl. Acad. Sci. U.S.A. 61, 636-643.

Cullen, D. (1987). Bio/Technology 5, 369-375. Demain, A. L. (1971). Adv. Biochem. Eng. Biotechnol. 1, 113-142. Deschamps, A. M., and Lebeault, J. M. (1984). Biotechnol. Lett. 6(4), 237-242. Deschamps, A. M., and Leulliette, L. (1984). Int. Biodeterioration 20(4), 237-241. Deschamps, A. M., Otuk, G., and Lebeault, J. M. (1983). J. Ferment. Technol. 61(1), 55-59. Deschamps, F., and Huet, M. C. (1985). Appl. Microbiol. Biotechnol. 22, 177-180. Dhar, S. C., and Bose, A. M. (1964). Leather Sci. 11, 27-38. Dykerhoff, H., and Ambruster, R. (1933). Z. Phys. Chem. 219, 38-56. Eveleigh, D. E., and Montenecourt, B. S. (1979). Adv. Appl. Microbiol. 25, 57-73. Fang, S. F. (1940). Hwang Hai 1(5), 5-8. Fernback, A. (1901). C.R. Acud. Sci. (Paris) 132, 1214-1215. Fordham, J. R., and Block, N. H. (1987). Dev. Ind. Microbiol. 28, 25-31. Forsyth, W. G. C. (1964). Annu. Rev. Plant Physiol. 15, 4 4 3 4 5 0 . Freudenberg, K., Blummel, F., and Frank, T. (1927). 2. Physiol. Chem., 164, 262-270. Frost, G. M. (1986). In “Development in Food Proteins” (B. J. F. Hudson, ed.), pp. 57-134. Elsevier Applied Science Publishers, London. Frost, G. M., and Moss, D. A. (1987). In “Biotechnology” (7. F. Kennedy, ed.), Vol. 7a, pp. 65-211. Verlag Chemie, Weinheim. Fumihiko, Y., and Kiyoshi, M. (1975). Jpn. Pat. 72,25,786. Galiotou-Panayatou, M., and Macris, B. J. (1986). Appl. Microbiol. Biotechnol. 23, 502-506.

Galiotou-Panayatou, M., Rodis, I?, Macris, B. J., and Stathakos, D. (1988). AppZ. Microbiol. Biotechnol. 28, 543-545. Ganga, P. S., Suseela, G., Nandy, S. C., and Santappa, M. (1977). Leather Sci. 24, 8-16. Ganga, P. S., Suseela, G., Nandy, S. C., and Santappa, M. (1978). Leather Sci. 25, 203-209. Gathon, A., Gross, Z., and Rozhanski, M. (1989). Enzyme Microb. Technol. 11, 604-609. George, E. C., and Sen, S. N. (1960). Bull. CLRI6, 279-280. Glick, Z., and Joslyn, M. A. (1970). J. Nutr. 100, 516-519.

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

257

Godfrey, A. (1985). In “Microbiology of Fermented Foods” (B. J. B. Wood, ed.), Vol. 1, pp. 345-371. Elsevier Applied Science Publishers, London. Goldstein, J. L., and Swain, T. (1965). Phytochemistry 4, 185-192. Goodwin, T. W., and Mercer, E. I. (1983). In “Introduction to Plant Biochemistry,” pp. 529-566. Pergamon, New York. Grant, W. D. (1976). Science 193, 1137-1139. Hand, D. W., Sanford, P. A., and Smyth, D. H. (1966). Nature 209, 618. Haslam, E. (1966). In “Chemistry of Vegetable Tannins.” Academic Press, New York. Haslam, E. (1981). In “Biochemistry of Plants” (E. Conn and P. K. Stumpf, eds.), Vol. 7, pp. 527-556. Academic Press, New York. Haslam, E., and Lilley, T. H. (1988). CRC Crit. Rev. Food Sci. Nutr. 27(1), 1-40. Haslam, E., and Stangroom, J. E. (1966). Biochem. J. 99, 28-31. Haslam, E., and Tanner, R. J. N. (1970). Phytochemistry 90, 2305-2309. Haslam, E., Haworth, R. D., Jones, K., and Rogers, H. J. (1961). J. Chem. SOC., pp. 1829-1859. Headon, D. R., and Walsh, G. (1994). Biotechnol. Adv. 12(4), 635-646. Henis, Y., Tagari, H., and Volcani, R. (1964). Appl. Microbiol. 12(3), 204-209. Herz, F., and Kaplan, E. (1968). Nature 217, 1258-1259. Hesseltine, C. W., and Wang, H. L. (1967). Biotechnol. Bioeng. 9, 275-288. Iibuchi, S., Minoda, Y . , and Yamada, K. (1967). Agric. Biol. Chem. 31(5), 513-518. Iibuchi, S., Minoda, Y., and Yamada, K. (1968). Agric. Biol. Chem. 32(7), 803-809. Iibuchi, S., Minoda, Y., and Yamada, K. (1972). Agric. B i d . Chem. 36(9), 1553-1562. Jacob, F. H., and Pignal, C. (1975). Mycopathologia 57, 139-148. Jean, D., Pourrat, H., Pourrat, A., and Carnat, A. (1981). Anal. Biochem. 110, 369-372. Katwa, L. C., Rarnakrishna, M., and Raghavendra, R. M. R. (1981). J. Biosci. 3(2), 135-142. Kazuo, K., Masaki, Y., and Yoshitake, T. (1973). Jpn. Pat. 7,335,478. Kennedy, J. F., Cabral, J. M. S., Sa-Correia, I., and White, C. A. (1987). In “Biotechnology” (J. F. Kennedy, ed.), Vol. 7a, pp. 340-404. Verlag Chemie, Weinheim. Kimura, K.. Yamamoto, M., and Tanaka, Y. (1973). Jpn. Pat. 7,335,478. Kita, J. (1917). J. Chem. Ind. (Tokyo] 20, 134-137. Knudson, L. (1913). J. Biol. Chem. 14(3),159-207. Kumar, P. K. R., and Lonsane, B. K. (1989). Adv. Appl. Microbiol. 34, 29-139. Laukevics, J. J., Aspite, A. F., Viesturs, H. E., and Tengerdy, R. P. (1984). Biotechnol. Bioeng. 26, 1465-1474. Lauren, S. J., and Lee, K. (1988). J. Food Sci. 53(1), 181-184. Lekha, P. K. (1996). Ph.D. Thesis, University of Mysore. Lekha, P. K., and Lonsane, B. K. (1993). Process Biochem. 29,497-503. Lekha, P. K., Ramakrishna, M., and Lonsane, B. K. (1993). Chem. Microbiol. Technol. Lebensm. 15(12), 5-10. Lekha, P. K., Nagin Chand, and Lonsane, B. K. (1994). Bioprocess Eng. 11, 7-15. Lewis, J. A,, and Papavizas, G. C. (1968). Phytopafh. Z. 63, 124-134. Lewis, J. A., and Starkey, R. L. (1962). Soil Sci. 707, 241-247. Lippitsch, M. (1961). Appl. Microbiol. 39, 209-220. Lonsane, B. K., and Ghildyal, N. P. (1992). In “Solid Substrate Cultivation” (H. W. Doelle, D. A. Mitchell, and C. E. Rolz, eds.), pp. 191-209. Elsevier Science Publishers, London. Madhavakrishna, W., and Bose, S. M. (1961). Bull. CLRI 8,153-165. Madhavakrishna, W., Bose, S. M., and Nayudamma, Y. (1960). Bull. CLRT 7, 1-11.

258

P. K. LEKHA AND B. K. LONSANE

Masschelein, C. A., and Batum, M. S. (1981). In “Proceedings of the Congress of European Brewers Convention,” pp. 359-370. Mikhlin, D. M., and Gulkina, G. L. (1932). Izv. Tzen. Nauch-Issled. Inst. Kozhevennoi Prom. 1, 8-11. Misawa, N., Okamota, T., and Kitamura, K. (1988). 1.Biotechnol. 4, 241-252. Mitchell, D. A., and Lonsane, B. K. (1992). In “Solid Substrate Cultivation” (H. W. Doelle, D. A. Mitchell, and C. E. Rolz, eds.), pp. 1-16. Elsevier Science Publishers, London. Mitjavila, S., Gaillard, D., and Derache, R. (1968). Biologie et Gastro-Enterologie 2, 183186.

Moo-Young, M., Moriera, A. R., and Tergendy, R. €? (1983). In “Filamentous Fungi” (J. E. Smith, D. R. Berry, and B. Kristiansen, eds.), Vol. 4, pp. 117-144. Edward Arnold, London. Mudgett, R. E. (1986). In “Manual of Industrial Microbiology and Biotechnology” (A. L. Demain and N. A. Solomon, eds.), pp. 66-83. American Society of Microbiology, Washington, DC. M u , A. (1953). Phytopathology 43, 32-38. Nagalakshmi, S., Jayalakshmi, R., and Seshadri, R. (1985). 1.Food Sci. Technol. 22(3), 198-201.

Nicolson, W. N., Nierenstein, M., and Prince, N. V. (1931). Biochem. 1.25, 752-757. Nierenstein, M. (1930). Nature 125, 348-349. Nierenstein, M. (1936). Analyst 61, 294-295. Nishira, H. (1959a). Halcko Kogaku Zasshi 37, 85-89. Nishira, H. (1959b). Kagaku 29, 42-43. Nishira, H. (1960). Kagaku 30, 644-645. Nishira, H. (1961). Hakko Kogaku Zasshi 39, 137-146. Nishira, H. (1962). Hyogo Noka Daigaku Henkyu Hokoku 5, 117-121. Nishira, H., and Joslyn, M. A. (1968). Phytochemistry 7, 2147-2156. Nishira, H., and Mugibayashi, N. (1953). Sci. Repts. Hyogo Univ. Agric., Ser. Agric. Chem. 1, 6-10.

Nishira, H., and Mugibayashi, N. (1956). Sci. Repts. Hyogo Univ. Agric., Ser. Agric. Chem. 2, 87-91.

Nishira, H., and Mugibayashi, N. (1960). Hyogo Noku Daigaku Henkyu Hokoku 4, 113116.

Okamura, S., and Yuasa, K. (1987). Jpn. Pat. 62,304,981. Okamura, S., Mizusawa, K., Takei, K., Imai, Y., and Ito, S. (1988). Jpn. Pat. 62,272,972. Parmentier, F. (1970). Arch. Int. Physiol. Biochem. 78(1), 131-133. Parmentier, F., and Verbruggen, C. (1973). Arch. Int. Physiol. Biochem. 81(2), 339-341. Pottevin, H. (1901). C.R. Acad. Sci. (Paris)131, 1215-1217. Pourrat, H., Regerat, F., Pourrat, A., and Jean, D. (1982). Biotechnol. Lett. 4(9), 583-588. Pourrat, H., Regerat, F., Pourrat, A., and Jean, D. (1985). J. Ferment. Technol. 63(4), 401-403.

Pourrat, H., Regerat, F., Morvan, P., and Pourrat, A. (1987). Biotechnol. Lett. 9(10), 731734.

Prescott, S. C., and Dunn, C. G. (1959). In “Industrial Microbiology” (B. M. Miller and W. Litsky, eds.), p. 548. McGraw-Hill, New York. Price, H. L., and Butler, L. G. (1980). Agric. Exp. Sta. Bull. 272, 37-39. Rajkumar, S., and Nandy, S. C. (1983). Appl. Environ. Microbiol. 46(2), 525-527. Rajkumar, S., and Nandy, S. C. (1985). Leather Sci. 32(11), 278-280. Rajkumar, S., and Nandy, S. C. (1986). Leather Sci. 33(8), 220-227. Ralph, B. J. (1976). Food Technol. Australia 28, 247-251.

PRODUCTION AND APPLICATION OF TANNIN ACYL HYDROLASE

259

Ramakrishna, S. V., Suseela, T. Ghildyal, N. P., Jaleel, S. A., Prema, P., Lonsane, B. K., and Ahmed, S. Y. (1982). Ind. J. Technol. 20, 476-480. Reshetnikova, I. A., Uspenskaya, G. D., Pazur, G. A., and Pobedinskaya, M. A. (1984). Mikol. Fito. Patol. 18, 483-488. Rhind, D., and Smith, F. E. (1922). Biochem. I. 16, 1-2. Rossi, M., Giovanelli, G., Cantarelli, C., and Brenna, 0. (1988). Bulletins Liaison Groupe Polyphenols 14, 85-88. Salunkhe, D. K., Jadhav, S. J., Kadam, S. S., and Chavan, J. K. (1982). CRC Grit. Rev. Food Sci. Nutr. 17(3), 277-305. Sambandan, T. (1983). Ph.D. Thesis, University of Madras. Sanderson, G. W. (1972). World Coffee ond Teo 12, 54-57. Sanderson, G. W., Englewood, N. J., Coggon, P., and Orangeburg, N. Y. (1974). U.S. Pat. 3,812,266. Seiji, D., Atsuhiko, S., Toshio, E., and Gyozo, T. (1973). HawCo Kogaku Zasshi 51(11), 768-774. Shankaranand, V. S., Ramesh, M. V., and Lonsane, B. K. (1992). Process Biochem. 27, 33-36. Singh, A,, and Hayashi, K. (1995). In “Advances in Applied Microhiology” ( S . Neidleman and A. I. Laskin, eds.), Vol. 40, pp. 1 4 4 . Academic Press, New York. Singleton, V. L., and Kratzer, F. H. (1969). J. Agric. Food Chem. 17,497-512. Sivaswamy, N. (1982). Ph.D. Thesis, University of Madras. Steinkraus, K. H. (1984). Acta Biotechnol. 4, 83-88. Strumeyer, D. H., and Malin, M. J. (1970). Biochem. J. 118, 989-991. Swain, T. (1965). In “Plant Biochemistry” (J. Bonner and J. E. Varner, eds.), p. 552. Academic Press, New York. Swain, T., and Bate-Smith, E. C. (1962). In “Comparative Biochemistry,” Vol. 3A, pp. 755-809. Academic Press, New York. Takino, Y. (1976). U.S. Pat. 3,959,497. Tamir, M., and Alumot, E. J. (1970). 1.Nutr. 100, 573-580. Teighem, V. (1867). C.R. Acad. Sci. (Paris)65, 1091-1094. Tenco Brooke Bond Ltd. (1971). U.K. Pat. 1,249,932. Thom, C., and Raper, K. B. (1945). In “A Manual of Aspergilli,” pp. 214-241. Williams and Wilkins, Baltimore. Thomas, R. L., and Murtaugh, K. (1985). 1.Food Sci. 50, 1126-1129. Toth, G. (1944). Magyar ‘rimor5, 1 4 . Toth, G., and Barsony, G. (1943). Enzymologia 11, 19-23. Trevan, M. D. (1980). In “Immobilized Enzymes: An Introduction and Application to Biotechnology,” pp. 56-103. Wiley, New York. Tsai, C. H. (1985). Eur. Pat. 135,222. Underkofler, L. A. (1976). In “Industrial Microbiology” (B. M. Miller and W. Litsky, eds.), pp. 128-164. McGraw-Hill, New York. Vandamme, E., Jerome, M., Vermoira, A., and Maria, M. (1989). Eur. Pat. 339,011. Volesky, B., and Luong, H. J. T. (1985). CRC Crit. Rev. BiotechnoL 2, 119-146. Walsh, G., and Headon, D. (1994). In “Protein Biotechnology,” pp. 10-38, Wiley, Chichester. Watanabe, A. (1965). Agric. B i d . Chern. 29, 20-26. Weetal, H. H. (1985a). Biotechnol. Bioeng. 27, 124-127. Weetal, H. H. (1985h). Eur. Pat. 137,601. Weetal, H. H., and Detar, C. C. (1974). Biotechnol. Bioeng. 16, 1091-1102.

260

P. K. LEKHA AND B. K. LONSANE

White, T. (1957). J. Sci. Food Agric. 8, 377-385. Wickremansinghe, R. L. (1978). Adv. Food Res. 24, 229-286. Yamada, H., Adachi, O., Watanabe, M., and Sato, N. (1968). Agric. Biol. Chem. 32(9), 1070-1 9 7 8.

Yamada, K. (1977). Biotechnol. Bioeng. 19, 1563-1621. Yamada, K., and Tanaka, T. (1972). Jpn. Pat. 2,224,100. Yamada, K., Iibuchi, S., and Minoda, Y.(1967). Hakko Kogaku Zasshi 45(3), 233-240.

Ethanol Production from Agricultural Biomass Substrates RODNEYJ. BOTHASTAND BADALC. S m Fermentation Biochemistry Research Unit National Center for Agricultural Utilization Research USDA* Agricultural Research Service Peoria, Illinois 61604

Introduction Lignocellulosic Biomass Pretreatment Enzymatic Conversion A. Cellulose B. Hemicellulose C. Lignin V. Fermentation A. Microorganisms B. Fermentation of Hydrolyzates VI. Technological Constraints to Scale-Up A. Enzymes B. Inhibitors C. Genetic Stability and Productivity D. Recovery of Dilute Ethanol VII. Future Prospects References I. 11. 111. IV.

I. Introduction

The increasing use of oxygenates as fuel additives provides an opportunity for large-scale expansion of the fuel ethanol industry. In 1994, about 1.3 billion gallons of fuel ethanol were produced in the United States, primarily from corn starch (House, 1995).However, the commercial viability of ethanol production from corn is dependent both on the price of corn and the income derived from the sale of corn oil, corn gluten meal, and fiber residues and stillage (corn gluten feed or distiller-dried grains, depending on whether the grain is processed by wet or dry milling) as an animal feed or co-product (Hohmann and Rendleman, *Names are necessary to report factually on available data; however, the USDA neither guarantees nor warrants the standard of the product, and the use of the name by USDA implies no approval of the product to the exclusion of others that may also be suitable. 261 ADVANCES IN APPLIED MICROBIOLOGY,VOLUME 44

262

RODNEY J. BOTHAST AND BADAL C. SAHA

Biomass

Enzymatic

I Fermentation I

c Ethanol I

Recovery

I

FIG.1. Production of ethanol from biomass by means of enzymatic hydrolysis.

1993). In 1996, corn prices rose above $5 per bushel, and, although the value of the co-products also increased, the net cost of corn as a feedstock for producing a gallon of ethanol exceeded the selling price of ethanol. Capital and operating costs have remained fairly stable. Overall, this scenario places the corn-to-ethanol industry in an economically challenged position and illustrates the need for new bioprocessing technologies that lower the cost of ethanol production and improve the competitiveness of ethanol as a fuel or fuel additive. Various lignocellulosic biomass sources, such as agricultural residues, food-processing wastes, wood, municipal solid wastes, and wastes from the pulp and paper industry have the potential to serve as low-cost and abundant feedstocks for production of fuel ethanol or chemicals (Fig. 1). The steps for production of fuels and chemicals from lignocellulosic biomass involve feedstock preparation, pretreatment, fractionation, enzyme production, hydrolysis, fermentation, product recovery, and waste treatment. The use of lignocellulosic biomass to produce fuel ethanol now faces significant technical and economic challenges. Its success depends largely on the development of environmentally friendly pretreatment procedures, highly effective enzyme systems for conversion of pretreated biomass to fermentable sugars, and efficient microorganisms to convert multiple sugars to ethanol. In this chapter, we review production of fuel alcohol horn various agricultural biomass substrates.

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

263

11. Lignocellulosic Biomass

Lignocellulosic biomass includes various agricultural residues (e.g., straws, hulls, stems, stalks, bagasse), fruit and vegetable wastes, deciduous and coniferous woods, municipal solid wastes (paper, cardboard, yard trash, wood products), waste from the pulp and paper industry, and herbaceous energy crops. The compositions of these materials vary. The major component is cellulose (35-50%), followed by hemicellulose (20-35%) and lignin (10-25%). Proteins, oils, and ash in widely varying ratios make up the remaining fraction of lignocellulosic biomass (Wyman, 1994a). The structure of these materials is very complex, and native biomass is resistant to enzymatic hydrolysis. In the current model of the structure of lignocellulose, cellulose fibers are embedded in a lignin-polysaccharide matrix. Xylan may play a significant role in the structural integrity of cell walls by both covalent and noncovalent associations (Thomson, 1993). Corn fiber represents a renewable resource that is available in sufficient quantities from the corn wet milling industry to serve as a low-cost feedstock for production of fuel alcohol. Corn fiber consists primarily of lignocellulosic biomass and starch (TableI) and is currently marketed in corn gluten feed. Several promising pretreatment and enzymatic processes have potential for conversion of corn fiber cellulose, hemicellulose, and the remaining starch to fermentable sugars. In all cases, these hydrolyzates are rich in pentoses (D-XylOSe and L-arabinose) and glucose. The production of ethanol from corn fiber has the potential to increase ethanol yields by a maximum of 0.3 gal/bushel (Table 11) in a wet milling process (Gulati et ~ l . 1997). , Improved technologies that integrate fiber conversion into a limited number of corn ethanol plants could have a significant impact, since 10 plants account for 75% of the fuel ethanol produced. On this basis alone the combination of fiber pretreatment, hydrolysis, and fermentation has the potential to increase ethanol production by more than 70 million gallons while providing test beds for new cellulose and hemicellulose conversion technologies. Gulati et al. (1997) have estimated that a wet milling facility that produces 100 million gal/year of ethanol from starch could generate an additional $4 to 8 million income annually if the fiber components were processed into ethanol. Hence, advances in fiber pretreatment, enzymatic hydrolysis, and pentose fermentation are likely to have a major impact on enhancing productivity of corn ethanol plants.

2 64

RODNEY J. BOTHAST AND BADAL C. SAHA TABLE 1 COMPOSnlON OF

CORN FIBER'

a. Carbohydrates Crude fiber

Starch

Glcb

Xylb

Arab

Galb

sugars

14.1'

19.68

37.19

17.58

11.25

3.59

69.6'

(0.97)d

(0.91)

(1.86)

(1.76)

(1.46)

(0.336)

(5.03)

TotalC

b. Other components Protein

Klason lignin

groups

Acetyl Ash

Crude fat

Unknown

10.98

7.78

1.71

0.6

2.53

6.79

(0.52)

(0.74)

(0.13)

(0.05)

(0.31)

'From Grohmann and Bothast (1997). bAverage concentration of each component is expressed in wt% of total dry solids. T h e sugar content is expressed on anhydrous basis. %umbers in parentheses refer to standard deviations.

TABLE I1

YIELD FOR WETMILLING PER BUSHEL OF CORN Product/ co-product

Starch Solubles Fiber Germ cake Gluten m e a l Crude oil Other Water Total

Pounds Per bushel 31.94 3.60 4.49 1.72 2.75 1.83 0.99 8.68 56

Theoretical yield (gal/bu)

Realistic yielda (gallbu)

2.76

2.48

unknown

-

0.30

0.22

-

-

3.06 3.06

2.71 2.70

Source: Gulati et 01. (1997)in part. "Realistic yield assumes 90% efficiency for hydrolysis and fermentation of starch and 74% efficiency for fiber.

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

265

Ill. Pretreatment

Pretreatment of lignocellulosic biomass is crucial to enzymatic hydrolysis. Dunning and Lathrop (1945) recognized the potential of such agricultural residues as corncobs, oat hulls, and flax sheaves as inexpensive sources of potentially fermentable sugars. In their classic paper, these authors presented clear-cut methods for low-temperature (10012OoC) dilute sulfuric acid extraction and hydrolysis of the hemicellulose components of biomass. Using ground corncobs as a model feedstock, 95% of the hemicellulose was removed from the biomass as product stream, consisting of about 86% xylose, 9% furfural, and 0.8% glucose. In the intervening years, these results were repeated many times by other researchers by applying the same or slightly modified dilute acid methods to a wide variety of biomass sources. The main advantages of the dilute acid treatment of biomass include the production of a soluble pentose stream that can be physically separated from the particulate residue. Second, a substantially increased rate of enzymatic hydrolysis of the residual cellulose portion results, presumably in large part due to acid-induced increased fiber porosity (Grethlein, 1985). On the other hand, acid treatment produces furfural, which is toxic to many microorganisms, and the residual acid must be neutralized. Dunning and Lathrop (1945) produced a clean fermentable 15% xylose stream by removing the furfural by vacuum distillation and removing the acid as a filterable calcium sulfate cake by way of addition of lime. However, these procedures are not commercially practical due to the considerable costs added to processing in ethanol fermentations. Grohmann and Bothast (1997) investigated a sequential saccharification of polysaccharides in corn fiber by treatment with dilute sulfuric acid at 100-160°C followed by partial neutralization and enzymatic hydrolysis with mixed cellulase and amyloglucosidase enzymes at 45°C. The sequential treatment achieved high (approximately 85%) conversion of all polysaccharides in the corn fiber. However, the formation of compounds inhibitory to fermentative microorganisms became readily apparent at all pretreatments tested at 140 and 160°C. A number of alternative methods have been proposed and used to overcome the toxicity and neutralization problems of dilute acid treatment. One of these is steam explosion, which is very efficient in fractionating wood biomass but results in large losses of hemicellulose sugars (Carrasco et al., 1994). A more gentle procedure is low-temperature (30-80°C) ammonia fiber explosion treatment (AFEX), which effectively disrupts nonwoody biomass such as coastal bermuda grass (De

266

RODNEY J. BOTHAST AND BADAL C. SAHA

La Rosa et al., 1994), but not softwoods. Moniruzzaman et al. (1997a) determined the best AFEX operating conditions for pretreatment of corn fiber. Approximate optimal pretreatment conditions for unground corn fiber containing 150% (dwb) moisture were found to be as follows: temperature = 90°C, ammonia:dry corn fiber mass ratio = 1:1, and residence time = 30 min (the average reactor pressure under these conditions was 200 psi). More than 85% of the theoretical sugar yield was obtained during enzymatic hydrolysis after pretreatment of corn fiber under these optimized conditions. However, unlike digestion of the starch-cellulose components of the corn fiber by commercial enzymes to monosaccharides, the hemicellulose or xylan component was only partially (25%) digested to monomeric sugars, leaving mixtures of unfermentable oligomeric sugars (Hespell et al., 1997). While this research indicates that the AFEX process does not produce toxic products, it does point to the need for better xylan-degrading enzymes. Various other pretreatment options are now available to fractionate, solubilize, hydrolyze, and separate cellulose, hemicellulose, and lignin components (Bungay, 1992; Wyman, 1994; Dale and Moreira, 1982; Weil et al., 1994). These include concentrated acid treatment (Goldstein and Easter, 1992), alkaline treatment (Koullas et al., 1993), treatment with SO2 (Clark and Mackie, 1987), treatment with hydrogen peroxide (Gould, 1984), and organic solvent treatments (Chum et al., 1988). In each option, the biomass is reduced in size and its physical structure opened. Pretreatment usually hydrolyzes hemicellulose to its sugars (xylose, L-arabinose, and other sugars) that are water-soluble (Bungay, 1992). The residue contains cellulose and often much of the lignin. The lignin can be extracted with such solvents as ethanol, butanol, or formic acid. Alternatively, hydrolysis of cellulose with lignin present produces water-soluble sugars and the insoluble residues that are lignin plus unreacted materials. The use of SO2 as a catalyst during steam pretreatment resulted in the enzymatic accessibility of cellulose and enhanced recovery of the hemicellulose-derived sugars (Brownell and Saddler, 1984). Steam pretreatment at 200-210°C with the addition of 1%SO2 (w/w) was superior to other forms of pretreatment of willow (Eklund and Zacchi, 1995). A glucose yield of 95% based on the glycan available in the raw material was achieved. A summary of various pretreatment options is given in Table III. Supercritical carbon dioxide explosion has been found to be very effective for pretreatment of cellulosic materials before enzymatic hydrolysis (Zheng et ~ l .1995). , Cao et al. (1996) reported a new pretreatment method that involves steeping of the lignocellulosic biomass (using corncob as a model feedstock) in dilute NH,OH at ambient temperature to remove

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

267

TABLE 111

METHODS FOR PRETREATMENT OF LIGNOCELLULOSIC BIOMASS Method

Example

Thermomechanical

Grinding, milling, shearing, extruder

Autohydrolysis

Steam pressure, steam explosion, supercritical carbon dioxide explosion

Acid treatment

Dilute acid (H2S04,HCl), concentrated acid (H2S04,HCl), acetic acid

Alkali treatment

Sodium hydroxide, ammonia, alkaline hydrogen peroxide

Organic solvent treatment

Methanol, ethanol, butanol, phenol

From Saha and Bothast (1997a).

lignin, acetate, and extractives. This is followed by dilute acid treatment that readily hydrolyzes the hemicellulose fraction to simple sugars, primarily xylose. The residual cellulose fraction of biomass can then be enzymatically hydrolyzed to glucose. Most pretreatments of lignocellulosic biomass are expensive with respect to costs and energy. IV. Enzymatic Conversion

A. CELLULOSE Cellulose is a linear polymer of 8,000-12,000 D-glucose units linked by 1,4-P-D-glucosidicbonds. The enzyme system for the conversion of cellulose to glucose comprises at least three enzyme types: an endo-1,4P-glucanase (EC 3.2.1.4), an exo-1,4-P-glucanase (EC 3.2.1.91), and a P-glucosidase (EC 3.2.1.21). The cellulolytic enzymes with p-glucosidase act sequentially and cooperatively to degrade crystalline cellulose to glucose. Endoglucanases act in a random fashion on the regions of low crystallinity of the cellulosic fiber, whereas exoglucanases remove cellobiose (P-1,4-glucosedimer) units from the nonreducing ends of the cellulose chains. Synergism between these two types of enzymes is attributed to the endo-exo form of cooperativity and has been studied extensively between cellulases in the degradation of cellulose by Trichoderma reesei (Henrissat et al., 1985). p-glucosidases hydrolyze cellobiose and in some cases cellooligosaccharides to glucose. This type of enzyme is generally responsible for kinetic regulation of the whole

268

RODNEY J. BOTHAST AND BADAL C. SAHA

cellulolytic process and is a rate-limiting factor during enzymatic hydrolysis of cellulose, as both endoglucanase and cellobiohydrolase activities are often inhibited by cellobiose (Woodward and Wiseman, 1982; Coughlan, 1985; Kadam and Demain, 1989). Thus, 0-glucosidase not only produces glucose from cellobiose but also reduces cellobiose inhibition, allowing the cellulolytic enzymes to function more efficiently. However, like P-glucanases, most P-glucosidases are subject to end-product (glucose) inhibition (Saha et al., 1995). The kinetics of the enzymatic hydrolysis of cellulose, including adsorption, inactivation, and inhibition of enzymes, have been studied extensively (Ladisch et al., 1983). For a complete hydrolysis of cellulose to glucose, the enzyme system must contain the three types of enzymes in proper proportions. Product inhibition, thermal inactivation, substrate inhibition, low product yield, and high enzyme costs are some barriers to commercial development of the enzymatic hydrolysis of cellulose. While many microorganisms are cellulolytic, only two genera of fungi (Trichoderma and Aspergillus) have been studied extensively for cellulase enzymes. There is an increasing demand for the development of thermostable, environmentally compatible, product and substrate-tolerant cellulases with increased specificity and activity for application in the conversion of cellulose to glucose in the fuel ethanol industry. Thermostable cellulases often have certain advantages, such as higher reaction rate, increased product formation, less microbial contamination, longer shelflife, easier purification, and better yield. Skory and Freer (1995) have isolated and cloned a Candida wickerhamii gene encoding a unique extracellular P-glucosidase that is highly resistant to D-glucose inhibition. The introduction of this gene into Saccharomyces cerevisiae has the potential to yield a cellodextrin-fermenting yeast with a cellulase system with much lower glucose inhibition and therefore ultimately improve ethanol productivity. In the continuing search for better yeasts and enzymes for biomass conversion, 48 yeast strains belonging to the genera Candida,Debaryomyces, Kluyveromyces, and Pichia were screened for production of extracellular glucose-tolerant and thermophilic P-glucosidase activity using p-nitrophenyl-P-D-glucoside as substrate (Saha and Bothast, 1996a). Glucose tolerance and thermoactivity found in the enzyme preparations from D. yamadae, K. marxianus, and C. chilensis are desired attributes of a 0-glucosidase suitable for industrial application for enzymatic hydrolysis of cellulose to glucose. Saha and Bothast (1996b)have purified and characterized a novel glucose- and cellobiosetolerant P-glucosidase from Candida peltata. The enzyme exhibited an

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

269

TABLE Tv BIOCHEMICAL CHRACTERISTICS OF P-GLUCOSIDASE FROM CANDIDA PELTATA NRRL Y-6888

Specific activity Molecular weight

108 U/mg protein

43,000 kDa

Optimum temperature

50°C

Optimum pH Specificity

5.o Hydrolyzes cellobiose and cello-oligosaccharide

Km value (mn/ll pNPPG (at pH 4.5, 75°C) Cellobiose (at pH 4.5, 75°C)

Metal ion requirement Substrate inhibition pNPPG (40 mM) Cellobiose (15%, w/v) Inhibition by glucose

2.3 mM 66 mM

None No inhibition No inhibition Competitive (Kj = 1.4 M]

Source: Saha and Bothast (1996b).

optimum activity at pH 5.0 and 50°C and was highly tolerant to glucose inhibition, with a Kivalue of 1.4 M (252 mg/ml). Substrate inhibition was not observed with 40 mMpNPPG or 15% cellobiose. Some properties of this enzyme are summarized in Table IV. Saha et al. (1994) have purified and characterized a thermostable P-glucosidase from a color variant strain of Aureobasidium pullulans. The enzyme displayed an optimum activity at 75OC and pH 4.5, but the enzyme was competitively inhibited by glucose, with a Kiof 5.65 mM. Cellulose hydrolysis to glucose is a significant component of the total production cost of ethanol from wood (Nguyen and Saddler, 1991). An overall economic process must include achieving a high glucose yield (>85% theoretical) at high substrate loading (>lo% w/v) over short residence times (4days). It has been shown that simultaneous saccharification (hydrolysis) of cellulose to glucose and fermentation of glucose to ethanol (SSF) improve the kinetics and economics of biomass conversion. This SSF reduces accumulation of hydrolysis products that are inhibitory to cellulase and P-glucosidase, reduces contamination risk because of the presence of ethanol, and reduces capital equipment requirements (Philippidis et al., 1993). An important drawback of a

2 70

RODNEY J. BOTHAST AND BADAL C. SAHA

yeast-based SSF is that the reaction has to operate at a compromised temperature of around 30°C instead of an optimum enzyme temperature in the range 45-5OOC. Enzyme recycling by ultrafiltration of the hydrolyzate can reduce the net enzyme requirement and thus lower costs (Tan et al., 1987). Hinman et al. (1992) reported that a preliminary estimate of the cost of ethanol production for SSF technology based on a woodto-ethanol process is $1.22/gal, of which the wood cost is $0.459/gal. Wright et al. (1986)evaluated a separate enzyme hydrolysis process and fermentation process for converting lignocellulose to ethanol. The cellulase enzyme was produced by the fungal mutant Trichoderma Rut C-30 (the first mutant with greatly increased 0-glucosidase activity) in a fed-batch production system. These findings showed that enzyme production is the single most expensive operation in the process. The efficient conversion of lignocellulosic biomass to fermentable sugars requires the use of complex enzyme mixtures tailored to the process. Enzyme recycling can increase the rates and yields of hydrolysis, reduce the net enzyme requirements, and thus lower costs (Lee et al., 1995). The first step in cellulose hydrolysis is considered the adsorption of cellulase onto a cellulosic substrate. As cellulose hydrolysis proceeds, the adsorbed enzymes (endo- and exo-glucanase components) are gradually released in the reaction mixture. Most P-glucosidases do not adsorb onto the substrate. All of these enzymes can be recovered and reused by contacting the hydrolyzate with fresh substrate. However, the amount of enzyme recovered is limited because some enzymes remain attached to the residual substrate and some enzymes become inactivated during hydrolysis. It has been shown that substrates containing a high proportion of lignin often result in poor recovery of cellulase (Tanaka et a]., 1988). Gusakov et al. (1992) found that cellolignin was completely converted to glucose by cellulase enzyme mixtures from ?: viride and A. foetidus. The cellolignin was an industrial residue obtained from the production of furfural from wood and corncobs when pretreated by dilute H2S04 at elevated temperature. The concentration of glucose in the hydrolyzate reached 4-5.5%, representing a cellulose conversion of about 80%. Kinetic analysis of cellolignin hydrolysis using a mathematical model of the process has shown that, with product inhibition, nonspecific adsorption of cellulase onto lignin and substrate-induced inactivation seem to negatively affect hydrolysis efficiency. Borchert and Buchholz (1987) investigated the enzymatic hydrolysis of different cellulosic materials (straw, potato pulp, sugar beet pulp) with respect to reactor design. The kinetic parameters studied included enzyme adsorption, inhibition, and inactivation. The results suggest the use of reactors with plug flow charac-

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

271

teristics to achieve high substrate and product concentrations and to avoid back-mixing to limit the effect of product inhibition. For efficient use of cellulases, a reactor with semipermeable hollow fiber or an ultrafilter membrane can be used. This allows cellulases to escape end-product inhibition (Henley et al., 1980; Klei et al., 1981; Ohlson et al., 1984; Kinoshita et al., 1986). Ghose and Ghosh (1992) reported a totally integrated bioprocess for conversion of rice straw into ethanol. The process includes: (i) ethanol refining of rice straw to segregate cellulose from pentose sugars and lignin, (ii) preparatior, of highly active mixed cellulase enzymes, (iii) a novel reactor system allowing rapid product formation involving enzymatic hydrolysis of cellulose to sugars followed by microbial conversion of the latter into ethanol and its simultaneous flash separation employing programmed recompression of ethanol vapors and condensation, and (iv) concentration of ethanol via alternative approaches. Direct microbial conversion of lignocellulosic biomass into ethanol could simplify the ethanol production process and reduce ethanol production costs. Lynd (1989) used Clostridium thermocellum, a thermoanaerobe, for enzyme production, hydrolysis, and glucose fermentation. Cofermentation with C. thermosaccharolyticum simultaneously converted hemicellulosic sugars to ethanol. However, formation of such by-products as acetic acid and low ethanol tolerance are two of the drawbacks of this system. Several reviews have dealt with the molecular biology of cellulose degradation, cellulolytic enzyme systems, and the structure and function of various domains found in the enzymes involved (Wilson, 1992; Beguin, 1990; Robson and Chambliss, 1989; Gilkes et al., 1988; Knowles et al., 1987). B . HEMICELLULOSE

Hemicelluloses are heterogeneous polymers of pentoses (xylose and L-arabinose), hexoses (mannose), and sugar acids. Xylans, major hemicelluloses of many plant materials, are heteropolysaccharides with a homopolymeric backbone chain of 1,4-linked p-D-xylopyranose units. Aside hom xylose, xylans may contain L-arabinose, D-glUCLUOniC acid or its 4-0-methyl ether, and acetic, p-coumaric, and ferulic acids. The total hydrolysis of xylan requires endo-P-1,4-xylanase(EC 3.2.1.8), p-xylosidase (EC 3.2.1.37), and several accessory enzyme activities, such as a-L-arabinofuranosidase (EC 3.2.1.55), a-glucuronidase (EC 3.2.1.1), acetyl xylan esterase (EC 3.2.1.6), feruloyl esterase, and p-coumaroyl esterase, which are necessary for hydrolyzing various substituted xy-

272

RODNEY J. BOTHAST AND BADAL C. SAHA

lans. The endo-xylanase randomly attacks the main chains of xylans, and P-xylosidase hydrolyzes xylooligosaccharides to xylose. The a-Larabinosidase and a-glucuronidase remove the arabinose and 4-0methyl glucuronic acid substituents, respectively, from the xylan backbone. The esterases hydrolyze the ester linkages between xylose units of the xylan and acetic acid (acetyl xylan esterase) or between arabinose side-chain residues and such phenolic acids as ferulic acid (feruloyl esterase) and p-coumaric acid (p-coumaroyl esterase). Synergistic action of depolymerizing and side group-cleaving enzymes has been proven using acetylated xylan as a substrate (Poutanen and Puls, 1989). Bachmann and McCarthy (1991) reported significant synergistic interactions between endo-xylanase, P-xylosidase, a-L-arabinofuranosidase, and acetyl xylan esterase enzymes of the thermophilic actinomycete Thermomonospora fusca. Many xylanases often do not cleave glycosidic bonds between xylose units that are substituted. The side chains must be cleaved before the xylan backbone can be completely hydrolyzed (Lee and Forsburg, 1987).On the other hand, several accessory enzymes only remove side chains from xylooligosaccharides. These enzymes require xylanases to hydrolyze hemicellulose partially before side chains can be cleaved (Poutanen et al., 1991). Although the structure of xylan is more complex than cellulose and requires several different enzymes with different specificities for complete hydrolysis, the polysaccharide does not form tightly packed crystalline structures and is thus more accessible to enzymatic hydrolysis (Gilbert and Hazlewood, 1993). The yeast-like fungus Aureobasidium is a promising source of xylanase (MW 20 kDa) with an exceptionally high specific activity (2100 U/mg protein) (Leathers, 1989). Xylanase represented nearly half the total extracellular protein, with a yield of up to 0.3 g of xylanase per liter (Leathers, 1986). A few reviews have dealt with the multiplicity, structure, and function of microbial xylanases, and the molecular biology of xylan degradation (Wong et d.,1988; Christov and Prior, 1993; Thomson, 1993). The utilization of hemicellulosic sugars is essential for efficient and economic conversion of lignocellulose to ethanol. The commercial exploitation of the pentose-fermenting yeasts for ethanol production from xylose is restricted mainly by their low ethanol tolerance, the slow rates of fermentation, the difficulty involved in controlling the rate of oxygen supply at the optimal level, and the sensitivity to microbial inhibitors, particularly those liberated during pretreatment and hydrolysis of lignocellulosic substrates (Du Preez, 1994; Hahn-Hagerdal et al., 1994a). Xylose can also be converted to xylulose using the enzyme xylose isomerase, and traditional yeasts can ferment xylulose to ethanol

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

273

(Gong et al., 1981; Hahn-Hagerdal et d., 1986). Xylose can be converted into xylitol, a potentially attractive sweetening agent, by a variety of microorganisms (yeasts, fungi and bacteria) (Nigam and Singh, 1995; Saha and Bothast, 1997b). C. LIGNIN

Lignin is a long-chain heterogeneous polymer composed largely of phenylpropane units most commonly linked by ether bonds. It effectively protects the woody plants against microbial attack, and only a few organisms including rot fungi and some bacteria can degrade it (Polvinen et al., 1991). The conversion of cellulose and hemicellulose to fuels and chemicals can generate lignin as a by-product that can be burned to provide heat and electricity, converted to low-molecular-weight chemicals, and used in the manufacture of various polymeric materials. As lignin makes up 15-25% of some lignocellulosic biomasses (especially woody ones), the selling price of lignin has a very large impact on ethanol price (Nguyen and Saddler, 1991). Removal of lignin from a lignin-carbohydrate complex (LCC) has received much attention because of potential application in the pulp and paper industry. The lignin barrier can be disrupted by a variety of pretreatments, rendering the cellulose and hemicellulose more susceptible to enzymatic attack (Chang et al., 1981).There are many papers about microbial breakdowns of lignin, the enzymes, and the pathways (Tien and Kirk, 1983; Tien, 1987; Umezawa and Higuchi, 1991; Fiechter, 1993). Degradation of lignin by the basidiomycete Phanerochaete chrysosporium is catalyzed by extracellular peroxidases (lignin peroxidase (Lip) and manganese peroxidase (MnP)) in an H,O,-dependent process (Shoemaker and Leisola, 1990; Zimmermann, 1990). However, due to the extreme complexity of lignin structures, a vast amount of research needs to be done to understand all the factors involved in the lignin biodegradation process (Vicuna, 1988). The cost of the enzymes for enzymatic hydrolysis of lignocellulosic biomass is clearly a critical parameter from an economic point of view. Most of the industrial enzymes are produced by organisms isolated from natural sources by labor-intensive and unpredictable classical screening, strain selection, medium optimization for overproduction, fermentation, and recovery process development. Screening of naturally occurring microorganisms (Fig. 2) may still be the best way to obtain new strains and/or enzymes for commercial applications (Cheetam, 1987). Recombinant DNA technology and protein engineering have also proven to be effective means of production of industrial enzymes

2 74

RODNEY J. BOTHAST AND BADAL C. SAHA

Screening for microorganism Culture selection Fermentation studies (preliminary) Isolation, purification and final characterization Evaluation Toxicology 0

Regulatory agency 0

Improvement of fermentation and recovery process development Product formulation Marketing

FIG.2. Strategies for commercial development of an enzyme.

(Arbige and Pitcher, 1989). The marketing of all enzymes is subject to a variety of federal laws and regulations. The “generally recognized as safe” (GRAS) status of an industrial enzyme depends on the source of its origin. Federal laws, regulations, and policies that have an impact on industrial enzymes have been reviewed by Fordham and Block (1987). V. Fermentation

A. MICROORGANISMS The potential increase in ethanol yield from fermenting D-xylose and L-arabinose from the fiber in corn grain is significant (Gulati et al., 1997). About 33% of the incremental ethanol can be derived from the D-xylose fiaction of the fiber and 24% from the L-arabinose fraction. The impact of D-xylose fermentation (to ethanol) for other lignocellulosic biomass materials is even greater since more than 25 % of the dry weight of lignocellulosic residues are xylans (Ladisch, 1989). The economic fermentation of D-xylose to ethanol is necessary for development of a fuel ethanol industry based on the use of cellulosic

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

Fungineast

Bacteria

Xylose

Xylose

Xylose reductase Xylitol dehydrogenase

-

275

Xylose isomerase

Xylito1

Xy lulose

XyIulose

-- *--

7

g

Pentose Phosphate Pathway

Acetyl-P

EMP Pathway Pyruvate

Pyruvate decarboxylase

Acetate

Acetaldehyde

Alcohol dehydrogenase

Ethanol

FIG.3. Xylose metabolism in fungi, yeast, and bacteria.

residues. The first step in xylose degradation is conversion to xylulose (Fig. 3). Bacteria generally accomplished this in one step with a xylose isomerase, while yeast uses a xylose reductase to reduce xylose to xylitol and a xylitol dehydrogenase to convert xylitol to xylulose (Smiley and Bolen, 1982). The traditional yeast Saccharomyces cerevisiae ferments glucose and sucrose and exhibits favorable ethanol production rates and tolerance. The bacterium Zymomonas mobilis also ferments glucose and sucrose at high rates. Other yeasts, such as Pichia stipitis, Candida shehatae, and Pachysolen tannophilus, have been discovered (Slininger et al., 1985; Jeffries and Sreenath, 1988; Neirinck et al., 1982; Schneider et al., 1983; Du Preez et al., 1986; Delgenes et al., 1988: Deverell, 1983; Hahn-Hagerdal et al., 1994b) that have the ability to ferment glucose and xylose but not L-arabinose, a major component (-11% of corn fiber). The yield and productivities of D-xylose fermentation by yeasts are lower than of hexose fermentation, with the major difference between hexose and pentose metabolism being that all pentoses have to be shuttled through the pentose phosphate pathway (Slininger et a]., 1987). In addition, these yeasts that ferment D-xylose directly must be improved with regard to ethanol tolerance and toler-

276

RODNEY J. BOTHAST AND BADAL C. SAHA

ance to potential inhibitors such as acetic acid present in hydrolyzates and less production of alcohol sugars as side-products. Consequently, there is significant impetus to continue to develop microorganisms that ferment D-glucose, D-xylose, and L-arabinose to ethanol. Much of the work to improve fermentation of multiple substrates has been in the development of genetically engineered microorganisms. Two fundamental molecular approaches have evolved for the development of superior ethanologenic microorganisms that can ferment pentose and hexose sugars. The first approachis to genetically engineer the ability to make ethanol into microbes that normally use multiple substrates. The second approach is to genetically engineer the ability to use multiple substrates into microbes that normally make ethanol (Bothast, 1994). In the first case, Dr. Lonnie Ingram and colleagues (Ingram et al., 1987, 1989; Ingram and Conway, 1988; Burckhardt and Ingram, 1992; Barbosa et al., 1992; Ohta et al., 1991; Doran and Ingram, 1983; Wood and Ingram, 1992) have developed a series of recombinant Escherichia coli and Klebsiella oxytoca strains for the fermentation of pentose and hexose sugars to ethanol. The alcohol dehydrogenase (adh B) and pyruvate decarboxylase (pdc) genes from Zymomonas mobilis have been introduced under the PET operon into these bacteria, and expression in the recombinant bacterium diverts its metabolism to high ethanol (Ohta et a]., 1991). Excellent conversion of sugars (glucose, xylose, and arabinose) to ethanol has been reported for model fermentation broths (Beall et al., 1991; Bothast et al., 1994) and for hemicellulose hydrolyzates (Lawford and Rousseau, 1991; Barbosa et al., 1992; Beall et al., 1992). The second approach is illustrated by the research of Zhang et al. (1995) of the National Renewable Energy Laboratory. They introduced the D-XylOSe assimilation and pentose phosphate pathway genes from E. coli into Z. mobilis. The yields thus far range from 0.15 to near the theoretical value of 0.51 g ethanol/g D-xylose. However, the sugar range (especially L-arabinose) and ethanol tolerance of this organism still needs further development. Advances in the genetics of engineered yeasts for pentose fermentation, particularly S. cerevisiae, are likely to have a more immediate impact on the corn processing industry because of their familiarity and experience with yeast fermentations and the potential robustness of the organisms. Most recently, Ho and Tsao (1995) genetically engineered a Saccharomyces strain to ferment xylose by expressing genes encoding a xylose reductase, a xylitol dehydrogenase, and a xylulose kinase. This recombinant was shown to efficiently ferment xylose, glucose, and galactose, and to produce ethanol with high yields (Moniruzzaman, 1997b).

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

277

B. FERMENTATION OF HYDROLYZATES

Ethanologenic E. coli have been developed for the fermentation of hemicellulosic syrups (Asghari et al., 1996). Final ethanol concentrations were limited to 40-50 g/liter, possibly due to product inhibition. In the study by Grohmann and Bothast (1997), dilute acid hydrolyzates of 15% corn fiber slurries were fermented with recombinant E. coli KO1 1 in 2-liter stirred bioreactors. Final ethanol concentrations exceeded 3% w/v in 3 days, and yields ranged from 19 to 62% of theoretical. The formation of inhibitory compounds was apparent for all pretreatments tested at 140 and 160OC. Acetic acid accumulated in the range of 0.06-0.30 wt% after treatment with acid and enzymes. Acetate was not responsible for the inhibition, because a separate set of experiments indicated that E. coli KO11 could tolerate up to 0.7-0.9% sodium acetate before inhibition of batch fermentations became evident. Also, inhibition also did not seem to be caused by accumulation of furfural or hydroxymethylfurfural because these compounds were not detected after pretreatment. However, the presence of trace (<500 ppm) amounts of these compounds cannot be ruled out due to the relatively low sensitivity of HPLC analyses. While the effects of inhibitors cannot be generalized for all fermentative microorganisms, their formation decreases the attractiveness of pretreatments performed at low acid loadings and high temperatures. Dien et al. (1997)used E. coli strains KO11 and SL40 to ferment dilute acid (1%v/v H,SO, at 121OC for 1h) hydrolyzates of corn fiber slurries in Fleaker fermentors (Beall et al., 1991). Both E. coli strains grew and performed well on the hydrolyzates (Table V). Maximum ethanol concentrations (3.2-3.5% w/v) were reached in approximately 48 h. Ethanol yields were 0.41 and 0.42 g/g (g of ethanol/g of sugar consumed), or approximately 80% of theoretical. Overall, strain KO11 tended to use sugars better than strain SL40. Glucose and arabinose were fermented rapidly and efficiently by both strains, but fermentation of xylose was slow and incomplete. Enzymatic digestion of the AFEX-pretreated corn fiber produced a hydrolyzate containing (g/liter): glucose, 30.2; xylose, 3.1; arabinose, 4.0; and galactose 2.0. Unfortunately, the commercial hemicellulase mixtures evaluated in this study left much of the xylose and arabinose in the form of short soluble oligomers and, in fact, proved almost as ineffective on various purified xylans (Hespell et al., 1997). Similarly, other pretreatment strategies currently being developed in which the hemicellulose is not completely hydrolyzed will face these same constraints. There is a continuing need to screen for microbial hemicellu-

2 78

RODNEY J. BOTHAST AND BADAL C. SAHA

TABLE V

FERMENTATION PERFORMANCE OF GENETICALLY ENGINEEREDMICROORGANISMS ON PRETREATED CORNFIBER HYDROLYZATES~

Pretreatment

Strain

Dilute acid

E. coli K O l l E. coli SL40 E. coli KOll E. coli SL40 Saccharomyces 1400

AFXX

Maximum ethanol (g/liter)

Ethanol yield (g/g)

(g/liter/h)

34.7

0.41

1.16

31.7

0.42

1.12

27.1

0.47

1.05

26.6

0.46

1.18

21.0

0.50

1.60

VPb

O a t a from Dien et al. (1997),Moniruzzaman et 01. (1996),and Moniruzzaman et al. (1997b). %laximum volumetric productivity.

lases in order to develop an effective system for complete digestion of AFEX-pretreated corn fiber. In order to compensate for incomplete digestion of the xylan in this study, the hydrolyzate was supplemented with additional xylose and arabinose, so that theoretical concentrations (g/liter) of xylose, 15.2, and arabinose, 10.5, were realized. Table V summarizes fermentation of the hydrolyzate by E. coli strains SL40 and KOll. Both strains utilized almost all of the sugars contained in the enzymatic hydrolyzate and produced maximum ethanol concentrations of 26.6 and 27.1 g/liter. Cell yields, maximum volumetric productivity, and ethanol yields are comparable to those reported (Barbosa et ~ l . , 1992; Beall et al., 1992; Dien eta]., 1997) for fermentation of dilute-acid hydrolyzates by K O l l . Fermentations were completed within 72 h by both strains. Furthermore, the absence of detectable inhibitors in enzymatic hydrolysis was confirmed by HPLC analysis. These results indicate good compatibility of AFEX pretreatment with subsequent fermentation by ethanologenic E. coli strains. Table V also summarizes fermentation of sugars produced by enzymatic digestion of AFEX-pretreated corn fiber by the recombinant SQCcharomyces strain 1400 (pLNH32). The hydrolyzate contained glucose (33.5 g/liter), xylose (7.5 g/liter), arabinose (5.0 g/liter), and galactose (1.0 g/liter). Fermentations were carried out in a pH 5.0 controlled Fleaker fermentor under anaerobic conditions with an inoculum level of ODGo0= 10. The recombinant strain efficiently utilized hydrolyzate sugars, glucose, xylose, and galactose and produced 21.0 g/liter ethanol

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

2 79

with yields of 0.5 g/g. Very little xylitol was detected in the fermentation medium, confirming the high theoretical yield. Further genetic engineering efforts will be needed so that this recombinant Saccharomyces can also ferment arabinose and maximize ethanol accumulation. VI. Technological Constraints to Scale-Up

A. ENZYMES

The industrial enzyme market approaches approximately $1 billion annually. Enzymes have already become commodity chemicals for such industrial applications as the production of various corn syrups and sweeteners, and fuel ethanol from starch. The volume market for enzymes involved in various lignocellulosic biomass conversion is limited and depends mainly on their use in the conversion of various lignocellulosic feedstocks to fermentable sugars for subsequent production of fuel alcohols and value-added chemicals. However, cellulolytic and xylanolytic enzymes are expensive and the hydrolysis rates often slow. The development of an environmentally compatible and highly efficient enzyme system free from product and substrate inhibitions for conversion of various pretreated agricultural residues to glucose, xylose, arabinose, galactose, etc., is very important for the use of these materials in the production of fuel alcohol. The market for these enzymes will expand rapidly if certain properties can be improved and if these enzymes are made available for biomass conversion at a competitive price, like starch-degrading enzymes. On the other hand, the development of a very efficient substrate pretreatment that increases the susceptibility of crystalline cellulose and hemicellulose to enzymatic hydrolysis significantly will lower the cost of producing ethanol from lignocellulosic biomass. B. INHIBITORS

The amount and nature of inhibiting compounds depends on the raw material, the prehydrolysis and hydrolysis procedures, and the extent of recirculation in the process. Knowledge about inhibitors and how to minimize their effects is of utmost importance for efficient fermentation of lignocellulosic hydrolyzates. Olsson and Hahn-Hagerdal (1996) divided inhibitors into five groups depending on their origin. First, substances released during prehydrolysis and hydrolysis include acetic acid (released when the hemicellulose structure is degraded) and extractives. The extractives comprise terpenes, alcohols, and aromatic

280

RODNEY J. BOTHAST AND BADAL C. SAHA

compounds such as tannins. Second, a group of inhibitors (furfural, 5-hydroxymethyl furfural, levulinic acid, formic acid, and humic substances) is produced as by-products in prehydrolysis and hydrolysis due to degradation of sugars. Third, lignin degradation products are also produced during prehydrolysis and hydrolysis. This group of inhibitors includes a wide range of aromatic and polyaromatic compounds with a variety of substituents. Fourth, products from the fermentation process, such as ethanol, acetic acid, glycerol, and lactic acid, inhibit the microorganisms. Finally, metals released from the equipment and additives such as SO2 will also inhibit fermentation. Several methods of detoxification have been reviewed by Olsson and Hahn-Hagerdal (1996) and include the addition of activated charcoal, extraction with organic solvents, ion exchange, and steam stripping. All these methods significantly increase the cost of the ethanol fermentation process. Hydrolysis processes that minimize the production of inhibitors are more desirable.

c. GENETICSTABILITY AND PRODUCTIVITY While genetically engineered bacteria and yeast hold tremendous potential for the fermentative conversion of multiple substrates to ethanol, questions still remain concerning the stability or hardiness of these organisms and their ability to perform in a large-scale industrial process. As noted earlier, Ingram et al. (1987) were the first to report a metabolically engineered E. coli for high alcohol production. The resultant recombinant strain produced over 4% wt/v ethanol from glucose in media containing ampicillin with positive selection pressure for the plasmid. A considerably more stable strain was developed by Ohta et al. (1991) by integrating the PET operon (gene cluster producing ethanol) and chloramphenicol (cm) resistance gene into the E. coli chromosome. The resultant E.coli KO11 strain did not require cm in the growth media for retention of the PET operon, but ethanol production in the absence of cm was lower, presumably due to reduced PET gene copy number. When mutants were selected for resistance to high levels (600 pg/ml) of cm, high ethanol production was restored. Hespell et al. (1996) have undertaken an alternative approach to the requirement for antibiotics. A lactate dehydrogenase (1dh)-pyruvate formate lyase (pfl) double mutant of E. coli was used as the cloning host. While capable of aerobic growth, this mutant strain is incapable of anaerobic growth because of its inability to regenerate oxidized pyridine nucleotides by reduction of pyruvate to lactate. Clones having recombinant plasmids containing an ldh gene can be isolated by complementing

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

281

for anaerobic growth by this strain. It was reasoned that an alternative complementation for anaerobic growth would be the PET operon containing plasmids, as expression of the pdc and adh genes would convert pyruvate to ethanol and regenerate oxidized pyridine nucleotides. If so, the resultant strains should be quite stable anaerobically for ethanol production. The strains should also not require antibiotics (ampicillin or tetracycline) in the growth media to maintain positive selective pressure for cells containing the PET operon plasmid because loss of the plasmid would be a conditionally lethal event under anaerobic growth conditions. E. coli strains FBRl and FBR2 were created by transforming E. coli FMJ39 with the PET operon plasmids pL01295 and pL01297, respectively. Both strains were capable of anaerobic growth and displayed no apparent PET plasmid losses after 60 generations in serially transferred (nine times) anaerobic batch cultures. In contrast, similar aerobic cultures rapidly lost plasmids. In high-cell-density batch fermentations, up to 4.4% wt/v ethanol was made from 10% glucose. Anaerobic glucoselimited continuous cultures of one strain grown for 20 days (51 generations: 23 with and then 28 after tetracycline removal) showed no loss of antibiotic resistance. Anaerobic serially transferred batch cultures and high-density fermentations were inoculated with cells taken after 57 generations from the previous continuous culture. Both cultures continued high ethanol production in the absence of tetracycline. The genetic stability conferred by selective pressure for PET-containing cells without requirement for antibiotics suggests potential commercial suitability for these E. coli strains. Also noted earlier, Ho and Tsao (1995)transformed a Saccharomyces yeast with a 2-pm replicating plasmid containing the I! stipitis xylose reductase (xyll) and xylitol dehydrogenase (xy12) genes and also the xylulose kinase gene from S. cerevisiae. A large amount of xylose was unused in low-inoculum fermentations, suggesting that the 2-pm plasmid harboring the xylose utilization genes may not be stably maintained when glucose is present in the medium (Moniruzzaman et a]., 199713). In the absence of selective pressure, such as the antimicrobial agent geneticin in the medium or xylose as a sole carbohydrate source, the 2-pm plasmid is most likely diluted or lost. When a high inoculum was used, this effect was minimized because of the smaller number of cell generations occurring in the nonselective medium. Efforts are currently directed at stabilizing the xylose-utilizing genes by integrating them into the genome. Additional genetic efforts will be necessary to allow for yeast fermentation of arabinose produced in the hydrolysis of corn fiber. Dien et al.

282

RODNEY J. BOTHAST AND BADAL C. SAHA

(1996) screened 116 different yeasts for the ability to ferment L-arabinose and have found the following species able to ferment the sugar: Candida a uringiensis, Candida succiphila , Ambrosiozyma monospora , and Candida sp. YB-2248. These yeasts produced low levels (up to 4.1 g/liter) of ethanol and are potential candidates for mutational or other genetic enhancements for increased ethanol production. D. RECOVERY OF DILUTEETHANOL

Regardless of the process, ethanol concentrations derived from biomass-derived sugar streams are likely to be on the order of 4 4 % . Consequently, energy costs for recovering dilute aqueous alcohol from such streams could be higher than that for the >lo% ethanol concentrations now processed in grain alcohol plants. The energy tradeoffs in recovering dilute alcohol can be moderated, however, by improved separations technology, for example, a process that combines distillation with adsorption, as discussed by Gulati et al. (1995), where a concentration of 5% ethanol to a dried fuel-grade product may be achievable at an energy input of about 20,000 Btu/gal. This approach is viable if an inexpensive adsorbent can be identified. VII. Future Prospects

Continuous pretreatments coupled with enzyme treatments are currently being optimized and scaled up in our pilot plant so that commercial hydrolyzates can be evaluated for ethanol productivity and yield. The most stable and hardy genetically engineered microorganisms available will be used for comparative fermentations. Application of these novel bioprocessing technologies has the potential to lower the cost of ethanol production and improve the competitiveness of ethanol as a fuel or fuel additive. REFERENCES Asghari, A., Bothast, R. J., Doran, J. B., and Ingram, L. 0. (1996). J. Ind. Microbiol. 16, 42-47.

Arbige, M. V., and Pitcher, W. H. (1989). Trends Biotechnol. 7 , 330-335. Bachmann, S. L., and McCarthy, A. J. (1991). Appl. Environ. Microbiol. 5 7 , 2121-2130. Barbosa, M. F. S., Beck, M. J., Fein, J. E., Potts, D., and Ingram, L. 0. (1992). Appl. Environ. Microbiol. 58, 1382-1384. Beall, D. S., Ohta, K., and Ingram, L. 0. (1991). Biotechnol. Bioeng. 38, 296-303. Beall, D. S., Ingram, L. O., Ben-Bassat, A., Doran, J. B., Fowler, D. E., Hall, R. G . , and Wood, B. E. (1992). Biotechnol. Lett. 14,857-862. Beguin, P. (1990). Annu. Rev. Microbiol. 44,219-248.

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

283

Bothast, R. J. (1994). In “Proc. Corn Utilization Conf. V.” National Corn Growers Association, St. Louis, MO. Bothast, R. J., Saha, B. C., Flosenzier, A. V., and Ingram, L. 0. (1994). Biotechnol. Lett. 16, 4 0 1 4 0 6 .

Borchert, A., and Buchholz, K. (1987). Process Biochem. 22,173-180. Brownell, H. H., and Saddler, J. N. (1984). Biotechnol. Bioeng. Syrnp. 14,55-68. Bungay, H. (1992). Enzyme Microb. Technol. 14,501-507. Burckhardt, G., and Ingram, L. 0. (1992). Appl. Environ. Microbiol. 58, 1128-1133. Cao, N.J., Krishnan, M. S., Du, J. X., Gong, C. S., Ho, N. W. Y., Chen, Z. D., and Tsao, G. T. (1996). Biotechnol. Lett. 18,1013-1018 Carrasco, J. E., Saiz, M. C., Navarro, A., Soriano, P., Saez, F., and Martinez, J. M. (1994). .4pp/. Biochem. Biotechnol. 45/46,23-34. Chang, M. M., Chou, T. Y.C., and Tsao, G. T. (1981). Adv. Biochern. Eng. 20, 15-42. Cheetham, P. J. (1987). Enzyme Microb. Technol. 9,194-212. Christov, L. P., and Prior, B. A. (1993). Enzyme Microb. Technol. 15,460-475. Chum, H. L., Johnson, D. K., Black, S., Baker, J., Grohmann, K. V., Sarkanen, K. V., Wallace, K., and Schroeder, H. A. (1988). Biotechnol. Bioeng. 31,643-649. Clark, T. A., and Mackie, K. L. (1987). J. Wood Chem. Technol. 7, 373-403. Coughlan, M. P. (1985). Biotechnol. Genet. Eng. Rev. 3, 39-109. Dale, B. E., and Moreira, M. J. (1982). Biotechnol. Bioeng. Symp. 12,31-43. De La Rosa, L. B., Reshamwala, S., Latimer, V. M., Shawky, B. T., Dale, B., and Stuart, E. D. (1994). Appl. Biochem. Biotechnol. 45/46,483-497. Delgenes, J. P., Moletta, R., and Navarro, J. M. (1988). Appl. Microbiol. Biotechnol. 29, 155-161.

Deverell, K. F. (1983). Biotechnol. Lett. 5,4 7 5 4 8 0 . Dien, B. S., Kurtzman, C. F!, Saha, B. C., and Bothast, R. J. (1996). Appl. Biochern. Biotechnol. 57/58,2 33-242. Dien, B. S., Hespell, R. B., Ingram, L. O., and Bothast, R. J. (1997). World J. Microbiol. Biotech. Accepted for publication 18 October 1996. Doran, J. D., and Ingram, L. 0. (1993). Biotechnol. Prog. 9, 533-538. Dunning, J. W., and Lathrop, E. C. (1945). Indust. Eng. Chern. 37, 24-29. Du Preez, J. C. (1994). Enzyme Microb. Technol. 16,944-956. Du Preez, J. C., Bosch, M., and Prior, B. A. (1986). Appl. Microbiol. Biotechnol. 23, 228-233.

Eklund, R., and Zacchi, G. (1995). Enzyme Microb. Technol. 17,255-259. Fiechter, A. (1993). J. Biotechnol. 30,49-55. Fordham, J. R., and Block, N. H. (1987). Dev. Indust. Microbiol. 28, 25-31. Gilkes, N. R., Henrisatt, B., Kilburn, D. G., Miller Jr., R. C., and Warren, R. A. J. (1988). Microbiol. Rev. 55, 303-315. Ghose, T. K., and Ghosh, P. (1992). In “Harnessing Biotechnology for the Zlst Century” (M. Ladisch and A. Bose, eds.), pp. 505-506. American Chemical Society, Washington, DC. Gilbert, H. J., and Hazlewood, G. P. (1993). J. Gen. Microbiol. 139,187-194. Goldstein, I. S., and Easter, J. M. (1992) Tappi J., pp. 135-140. Gong, C. S., Chen, L.-F., Flickinger, M. C., Chiang, L.-C., and Tsao, G. T. (1981). Appl. Environ. Microbiol. 41,4 3 0 4 3 6 . Gould, J. M. (1984). Biotechnol. Bioeng. 26,46-52. Grethlein, H. E. (1985). Biotechnology 3, 155-160. Grohmann, K., and Bothast, R. J. (1997). Process Biochem. Accepted for publication 16 October 1996.

284

RODNEY J. BOTHAST AND BADAL C. SAHA

Gulati, M., Westgate, P. J., Brewer, M., Hendrickson, R., and Ladisch, M. R. (1995). Appl. Biochem. Biotech. 57/58, 103-119. Gulati, M., Kohlmann, K., Ladisch, M. R., Hespell, R. B., and Bothast, R. J. (1997). Bioresour. Technol. 56, 253-264. Gusakov, A. V., Sinitsyn, A. P., Manenkova, J. A,, and Protas, 0. V. (1992).Appl. Biochem. Biotech. 34/35, 625-637. Hahn-Hagerdal, B., Berner, S., and Skoog, K. (1986). Appl. Microbiol. Biotechnol. 24, 28 7-2 9 3.

Hahn-Hagerdal, B., Halliborn, J., Jeppsson, H., Olsson, L., Skogg, K., and Walfridsson, M. (1993). In “Bioconversion of Forest and Agricultural Plant Residues” (J. N. Saddler, ed.), pp. 231-290. Oxon. Hahn-Hagerdal, B., Jeppsson, H., Skoog, K., and Prior, B. A. (1994a). Enzyme Microb. Technol. 16, 933-943. Hahn-Hagerdal, B., Jeppsson, H., Olsson, L., and Mohagheghi, A. (1994b). Appl. Microbid. Biotechnol. 41, 62-72. Henley, R. G., Young, R. Y. K., and Greenfield, P. F. (1980). Enzyme Microb. Technol. 2, 206-208.

Henrissat, B., Driguez, H., Viet, C., and Schulein, M. (1985). Bio/Technology 3, 722-726. Hespell, R. B., Wyckoff, H., Dien, B. S., and Bothast, R. J. (1996).Appl. Environ. Microbiol. 62,4594-4597.

Hespell, R. B., O’Bryan, P. J., Moniruzzaman, M., and Bothast, R. J. (1997).Appl. Biochem. Biotechnol. Accepted for publication 23 April 1996. Hinman, N. D., Schell, D. J., Rieley, C. J., Bergeron, P. W., and Walter, P. J. (1992). Appl. Biochem. Biotechnol. 34/35, 639-649. Ho, N. W. Y., and Tsao, G. T. (1995). PCT Pat. W095/13362. Hohmann, N., and Rendleman, C. M. (1993). Agricultural Information Bulletin No. 663. USDA, Washington, DC (January). House, C. (1995). Feedstufls 67, 21-22. Ingram, L. O., and Conway, T. (1988). Appl. Environ. Microbiol. 54, 397-404. Ingram, L. O., Conway, T., Clark, D. P., Sewell, G. W., and Preston, J. F. (1987). Appl. Environ. Microbiol. 53, 2420-2425. Ingram, L. O., Eddy, C. K., Mackenzie, K. F., Conway, T., and Alterthum, F. (1989). Dev. Indust. Microbiol. 30, 53-69. Jeffries, T. W., and Sreenath, H. K. (1988). Biotechnol. Bioeng. 31, 502-506. Kadam, S. K., and Demain, A. L. (1989). Biochem. Biophys. Res. Comrnun. 161, 706-711. Kinoshita, S., Chua, J. W., Kato, N., Yoshida, T., and Taguchi, T. (1986). Enzyme Microb. Technol. 8, 691-695. Klei, H. E., Sundstrom, D. W., Coughlin, K. W., and Ziolkowski, K. (1981). Biotechnol. Bioeng. Symp. 11, 593-601. Knowles, J., Lehtovaara, P., and Teeri, T. (1987). Trends Biotechnol. 5, 255-261. Koullas, D. P., Christakopoulos, P. E., Kekos, D., Koukios, E. G., and Macris, B. J. (1993). Biomass Bioenergy 4, 9-13. Ladisch, M. R. (1989). In “Biomass Handbook” (0.Kitani and C. W. Hall, eds.). Gordon and Breach, New York. Ladisch, M. R., Lin, K. W., Voloch, M., and Tsao, G. T. (1983). Enzyme Microb. Technol. 5, 82-102.

Lawford, H. G., and Rousseau, J. D. (1991). Biotechnol. Lett. 13, 191-196. Leathers, T. D. (1986). Appl. Environ. Microbiol. 52, 1026-1030. Leathers, T. D. (1989). 1. Ind. Microbiol. 4, 341-348. Lee, D., Yu, A. H. C., and Saddler, J. N. (1995). Biotechnol. Bioeng. 45, 328-336.

ETHANOL FROM AGRICULTURAL BIOMASS SUBSTRATES

285

Lee, S. F., and Forshurg, C. W. (1987). Can. J. Microbiol. 33, 1011-1016. Lynd, L. (1989). In “Adv. Biochem. Eng. Biotechnol” (A. Fiechter, ed.), Vol. 8, pp. 1-52. Springer-Verlag, Berlin. Moniruzzaman, M., Dien, B. S., Ferrer, B., Hespell, R. B., Dale, B. E., Ingram, L. O., and Bothast, R. J. (1996). Biotecbnol. Lett. 18,985-990. Moniruzzaman, M., Dale, B. E., Hespell, R. B., and Bothast, R. J. (1997a). Appl. Biocbern. Biotecbnol. Accepted for publication 11 April 1996. Moniruzzaman, M., Dien, B. S., Skory, C. D., Chen, Z. D., Hespell, R. B., Ho, N. W. Y., Dale, B. E., and Bothast, R. J. (1997b). World I. Microbiol. Biotech. Accepted for publication 23 September 1996. Nigam, P., and Singh, D. (1995). Process Biocbern. 30, 117-124. Neirinck, L., Maleszka, R., and Schneider, H. (1982). Biotecbnol. Bioeng. Symp. 12, 161-169.

Nguyen, Q. A., and Saddler, J. N. (1991). Bioresour. Technol. 35, 275-282. Ohta, K. D., Beall, D. S., Mejia, J. P., Shanmugam, K. T., and Ingram, L. 0. (1991). Appl. Environ. Microbiol. 57, 893-900. Ohlson, I., Trayardh, G., and Hahn-Hagerdal, B. (1984). Biotechnol. Bioeng. 26,647-653. Olsson, L.. and Hahn-Hagerdal, B. (1996). Enzyme Microb. Tecbnol. 18,312-331. Philippidis, G. P., Smith, T. K., and Wyman, C. E. (1993). Biotecbnol. Bioeng. 41, 846-853. Polvinen, K., Lehtonen, P., Leisola, M., and Visuri, K. (1991). In “Enzymes in Biomass Conversion” (G. F. Leatham and M. E. Himmel, eds.), pp. 225-235. American Chemical Society, Washington, DC. Poutanen, K., and Puls, J. (1989). In “Biogenesis and Biodegradation of Plant Cell Wall Polymers” (G. Lewis and M. Paice, eds.), pp. 630-640. American Chemical Society, Washington, DC. Poutanen, K., Tenkanen, M., Korte, H., and Puls, J. (1991). In “Enzymes in Biomass Conversion” (G. F. Leatham and M. E. Himmel, eds.), pp. 426-436. American Chemical Society, Washington, DC. Robson, L. M., and Chambliss, G. H. (1989). Enzyme Microb. Tecbnol. 11, 626-644. Saha, B. C., and Bothast, R. J. (1996a). Biotech. Lett. 18, 155-158. Saha, B. C., and Bothast, R. J. (1996b). Appl. Environ. Microbiol. 62, 3165-3170. Saha, B. C., and Bothast, R. J. (1997a).In “Fuels and Chemicals from Biomass” (B. C. Saha and J. Woodward, eds.), pp. 46-56. American Chemical Society, Washington, DC. Saha, B. C., and Bothast, R. J. (1997b). In “Fuels and Chemicals from Biomass” (B. C. Saha and J. Woodward, eds.), pp. 307-319. American Chemical Society, Washington, DC. Saha, B. C., Freer, S. N., and Bothast, R. J. (1994). Appl. Environ. Microbiol. 60,374-3780. Saha, B. C., Freer, S. N., and Bothast, R. J. (1995). In “Enzymatic degradation of Insoluble I’olysaccharides” (J. N. Saddler and M. H. Penner, eds.), pp. 197-207. American Chemical Society, Washington, DC. Schneider, H., Maleszka, R., Neirinck, L., Veliki, I. A., Wang, P. Y., and Cjan, Y. K. (1983). Adv. Biocbern. Eng. Biotecbnol. 27,57-71. Shoemaker, H.E., and Leisola, M. S. A. (1990). J. Biotechnol. 13,101-109. Skory, C. D., and S. N. Freer. (1995). Appl. Environ. Microbiol. 61,518-525. Slininger, P. J., Bothast, R. J., Okos, M. R., and Ladisch, M. R. (1985). Biotechnol. Lett. 7, 431436.

Slininger, P. J., Bolen, P. L., and Kurtzrnan, C. P. (1987). Enzyme Microb. Tecbnol. 9, 5-15. Smiley, K. L., and Bolen, P. L. (1982). Biotecbnol. Lett. 4, 607-610. Tan, L. U. L., Yu, E. K. C., Mayers, P., and Saddler, J. N. (1987). Appl. Microbiol. Biotechnol. 26, 21-27.

286

RODNEY J. BOTHAST AND BADAL C. SAHA

Tanaka, M., Fukui, M., and Matsuno, R. (1988). Biotechnol. Bioeng. 32, 897-902. Thomson, J. A. (1993). FEMS Microb. Revs. 104,65-82. Tien, M. (1987). CRC Crit. Rev. Microbiol. 15,141-168. Tien, M., and Kirk,T. K. (1983). Science 221, 661-663. Umezawa, T., and Higuchi, T. (1991).In “Enzymes in Biomass Conversion” (G. F. Leatham and M. E. Himmel, eds.),pp. 236-246. American Chemical Society, Washington, DC. Vicuna, R. (1988). Enzyme Microb. Technol. 10,646-655. Weil, J., Westgate, P., Kohlmann, K., and Ladisch, M. R. (1994). Enzyme Microb. Technol. 16,1002-1004.

Wilson, D. B. (1992). Crit. Rev. Biotechnol. 12,46-63. Wong, K. K. Y., Tan, L. U. L., and Saddler, J. N. (1988). Microbiol. Rev. 52,305-317. Wood, B. E., and Ingram, L. 0. (1992). Appl. Environ. Microbiol. 58,2103-2110. Woodward, J., and Wiseman, A. (1982). Enzyme Microbiol. Technol. 2, 73-79. Wright, J. D., Power, A. J., and Douglas, L. J. (1986). Biotechnol. Bioeng. Symp. 17, 285-302.

Wyman, C. E. (1994a). Appl. Biochem. Biotechnol. 45/46,897-915. Wyman, C. E. (1994b). Bioresour. Technol. 50, 3-16. Zhang, M., Eddy, C., Deanda, K., Finkelstein, M., and Picataggio, S. (1995). Science 267, 240-243.

Zheng, Y., Lin, H.-M., Wen, J., Cao, N., Yu, X., and Tsao, G. T. (1995). Biotech. Lett. 17, 845-850.

Zimmermann, W.J. (1990).J. Biotechnol. 13,119-130.

Thermal Processing of Foods, A Retrospective, Part I: Uncertainties In Thermal Processing and Statistical Analysis M. N. RAMESH Food Engineering Department Central Food Technological Research Institute Mysore 570 013, India

S. G. PRAPULLA Fermentation Technology and Bioengineering Department Central Food Technological Research Institute Mysore 570 023, India

M. A. KUMAR Central Instruments Facility Central Food Technological Research Institute Mysore 570 013, India

M. MAHADEVAIAH Food Packaging Technology Department Central Food Technological Research Institute Mysore 570 023, India

I. Introduction 11. Uncertainties In Thermal Processing A. Uncertainties During Thermal Death Time studies B. Uncertainties in Calculating the Value of D C. Uncertainties in Calculating the Value of z D. Uncertainties in Calculating the Value of F E. Uncertainties During Processing of Canned Products F. Inaccuracies in Evaluating Thermal Processing G. Uncertainties Due to pH H. Uncertainties Due to Temperature I. Uncertainties in the Safety Factor 111. Uncertainties in Aseptic Systems A. Uncertainties in the Aseptic Value of z B. Uncertainties in the Aseptic Value of F C. Uncertainties in Aseptic Processing IV. Statistical Analysis of Thermal Process Calculation V. Suggestions for Future Work VI. Conclusions References 287 ADVANCES IN APPLIED MICROBIOLOGY,VOLUME 44 Copyright 0 1997 by Academic Press, Inc. All rights of reproduction i n any form reserved. 0065-2164/97$25.00

288

M. N. RAMESH et al.

I. Introduction

The aim of thermal processing is to ensure adequate protection of a product by the destruction of potential spoilage organisms, based on reliable thermal death time (TDT) information and product heat-penetration data. Such various equipment as still retorts, rotary retorts, plate sterilizers, aseptic processing units, and ultrahigh-temperature (UHT) units have been developed for thermal processing. This was necessitated by the development of specialty and convenience foods, new container designs, new processing systems, quality considerations, energy savings, and processing regulations (Bee and Park, 1978). The basic procedure for calculating the time-temperature profile as the basis for ensuring commercial sterility as developed by Nicholas Appert has not changed much since the nineteenth century (Corcos, 1975). One of the major tools for calculating this time-temperature profile is Ball’s formula, introduced by Ball in 1923. Despite its limitations, the method is still widely used in the canning industry and stands as a classical tribute to the value of mathematics in food processing. However, improvements have been made in subsequent years (Merson eta]., 1978). A thermal process for foods should be very accurate, as it involves high temperatures and long durations, especially for low-acid foods. Underprocessing may result in spoilage of processed food products. Overprocessing can result in loss of valuable nutrients. The target in thermal processing is to ensure sterilization rather than nutrient retention, so as to avoid rejection and reprocessing of the processed foods. For this, a safety factor is needed, as the development of process schedules involves many uncertainties. Several publications have appeared on the ambiguity associated with parameters describing the temperature response of the inactivation of spores (Hicks, 1951,1961; Powers et al,, 1962; Herndon, 1971; Hurwicz and Tischer, 1956a;Perkins et al., 1975; Robertson and Miller, 1984; Bee and Park, 1978; Cleland and Robertson, 1985). Though a large number of publications have appeared on thermal process evaluation during the past 45 years, there has been little information on the magnitude of the uncertainty and error involved in thermal processing (Robertson and Miller, 1984). Thermal processing operations involve biological variability, which in turn affects commercial sterility. When some of the uncertainty related to the biological parameters cannot be avoided, the processors are forced to use safety factors. Instead of indiscriminately using the safety factors, several statistical techniques and computer methods that

THERMAL PROCESSING OF FOODS, PART I

289

were developed for applying safety factors can be adopted. Though this will not overcome the uncertainty, it will ensure optimum sterilization. The objective of this review is to identify the uncertainties with regard to Clostridium botulinum as the target microorganism and to summarize the statistical techniques available for evolving safety factors for sterilization within containers and aseptic processing. II. Uncertainties In Thermal Processing A. UNCERTAINTIES DURINGTHERMAL DEATHTIMESTUDIES

Thermal resistance of bacterial spores is dependent on water activity, pH, and the level of fats, proteins, carbohydrates, and salts present in , The specific effect of pH on the value of F foods (Hansen et ~ l .1963). has been worked out by Ito and Chen (1978). Hence, considering the effect of these parameters, it is always necessary to conduct TDT studies for the particular food product under investigation. Though there are some reports on TDT studies in particular foods, classical values derived from neutral phosphate buffer are generally used (Saikia and Ranganna, 1992). In the case of TDT studies for determination of the values of D and z, small glass tubes called ampules of wide-core diameter (7-13 mm) are used. Since these ampules have low thermal conductivity, the heating-up and cooling-down times have been considerable with lag time varying from 20 sec to 5 min (Shannon et al., 1970). To reduce the estimation error, a correction factor for the construction of TDT curves is added. Hence, the TDT data themselves would lead to approximation of the time-temperature data evolved. Pouches meant for TDT studies have been developed to considerably reduce error (Erdtsleck and Becimer, 1976). These do not require any safety factor. However, TDT ampules are still being used to evolve TDT data in the food processing industry. B . UNCERTAINTIES IN CALCULATING THE VALUE OF D

The uncertainties in derivation of D values from TDT curves (as indicated above) are multiples of the F values, as is evident from the following equation: F = D (log No - log Nf).

The value of D for a mesophilic spoilage organism is about five times greater than that for C. botulinum. The processor must decide what the

290

M. N. RAMESH et al.

target organism is and the acceptable level of spoilage and then calculate the required process Fo value (Stumbo et d.,1983). In practice, a process goal that provides reasonable assurance against spoilage is more important than that for commercial safety. The extent of safety margin varies for Fo values of 2 to 3 min (Alstrand et d.,1952) over and above that derived from the scientific studies. Xezones and Hutchings (1965) have studied the thermal resistance of C. botulinum spores as affected by certain food constituents and indicated that the D values of C. botulinum increase as the pH value increases. These increases in D values are statistically significant, as they are reflected in the calculation of F values and hence also in the duration of process. It appears that there are no conclusive data to establish the number of adequate trial runs and the number of containers to be used for determining a safe thermal process schedule (Robertson and Miller, 1984). The thermal death of bacteria is not truly logarithmic (Saikia and Ranganna, 1992). Hence, the D values determined from survivor curves are of limited validity, so that extrapolation of D values is inaccurate. Inoculated pack studies are carried out to confirm whether the calculated thermal process is adequate. . have developed proceHayakawa (1982) and Hayakawa et ~ l(1981) dures for mathematical analysis of logarithmic and nonlogarithmic survivor curves. The idea that thermal death of bacteria is exponential has motivated bacteriologists to determine the basis of thermal process calculation used for food processing (Ball and Olson, 1955; Schmidt, 1957; Stumbo, 1973). Vas and Proszt (1957) have postulated that 1. Populations of single strains of bacteria are homogeneous with re-

gard to heat resistance. 2. Thermal death of bacteria is unimolecular, that is to say, that death

occurs from inactivation of a single molecule. However, experimental evidence indicates that assumption 1 is seldom true and that assumption 2 is incompatible with evidence of sublethal injury (Ball and Olson, 1955; Moats et d.,1971). In view of this, there can be no theoretical basis for the common practice of assuming that logarithmic survivor curves and any deviation from linearity will lead to experimental errors. Still, the logarithmic curve forms the basis for all calculations used in thermal process evaluation, with generous safety factors added onto them. The impact of this assumption is reflected on the D values for thermal death of bacteria being

THERMAL PROCESSING OF FOODS, PART I

291

not truly logarithmic, so that, with these uncertainties, decimal reduction times ( D values) become meaningless (Hayakawa, 1982). One suggestion may be direct measurement of process times required for a given probability of kill at a given temperature. The direct method known as the spore count reduction technique involves inoculation of a representative number of cans with spores of a suitable test organism. In this operation, hundreds and sometimes thousands of cans packed with foods under study are inoculated with spores and subjected to different process types. The cans are then placed in incubation at a temperature near that for optimum growth of the test organism, and carefully observed for evidence of spoilage. The minimum required process corresponds to that giving the lowest processed lot showing no spoilage. However, this should not be confused with inoculation pack studies (Pilcher, 1949; Yawger, 1978; Berry and Bradshaw, 1986). This would give valid comparisons without introducing unwanted assumptions as to the logarithmicity of death rate (Moats et al., 1971). It has been demonstrated that, with this approach, F values are more reproducible and accurate than those derived from D values (Ott et d., 1961). Moats et al. (1971), from survivor curves carried through four or five log cycles, have also demonstrated that higher probabilities of kill could be seriously misleading. for C. botulinum, which is generally The suggested value of compatible with the 12D concept, is considerably lower than the value of suggested by Stumbo et d. (1983). If the initial spore numbers in the product are lo3 per unit and the probability of a final spore number is the spore logarithmic reduction will be iO-9/iO-3, that is, a 12D process would be required (Pflug and Odlaug, 1978). As suggested by Stumbo et al. (1975), a 13 to 15D process depending on the initial C. botulinum spore load is more compatible. An ideal thermal resistance determination method should be able to take into account all influencing factors. Many methods such as direct methods, indirect methods, particle methods, and mixing methods have been reported for thermal death determinations. Although most of these have individual advantages over previous methods, they are still rather laborious and leave important problems unsolved. These methods are discussed in detail by Brown and Ayres (1982). Being complex, direct methods do not allow working at low temperatures and using homogenized foods (Burton et d.,1977; Daudin and Cerf, 1977). In some methods, pH cannot be measured during heat treatment. The particle methods used for thermal death determination, aside from not allowing monitoring of pH and temperature, do not

292

M. N. RAMESH et al.

permit determinations in pure or homogenized foods (Brown et d., 1981).

The main advantage of the mixing method is elimination of lag phases and direct measurement of temperature in media. The literature provides procedures for determining heat resistance using survival curves without using food as the medium (Cerny, 1980; Brown et al., 1981). Determination of D values at ultrahigh temperatures (UHTs) with continuous pH measurement is not possible using the mixing method. Hence, most heat-resistant determinations are carried out using capillary tubes or employing some instrumental methods, like thermoresistometers (Cerny, 1980; Reichart, 1983; Mikolajacik and Rajkowsky, 1990; Brown et a]., 1981). A major drawback of all indirect methods is the inherent heating and cooling lag. In addition, although foods can be used as the medium, determinations at temperatures in the UHT range cannot be performed with high precision. The use of a capillary has drastically reduced lag time, but some still question whether the contents ever attain batch temperature (Brown and Ayres, 1982). C. UNCERTAINTIES IN CALCULATING THE VALUE OF Z The death rate kinetic factor z is assumed to be constant irrespective of a change in temperature. Actually, it follows a Arrhenius relationship and leads to the variable z. Gillepsey (1951) and Jonsson et d. (1977) suggested that straight calculations should be based on experimental results obtained at the actual temperature to be used during thermal processing and not by extrapolating data fiom other temperatures. D . UNCERTAINTIES IN CALCULATING THE VALUE OF F

In some cases, the 12D process adopted is inadequate when the initial spore count is high, as this prescribes an Fo value of 3 min assuming a defined initial and final number of spore counts. A better approach is to define an acceptable Nf value with a final spore reduction of and calculate the minimum safe process from the following equation (Cleland and Robertson, 1985): F = D[hg ( N f I N J ] .

Improperly cooled cans and faulty accounting of the cooling phase lead to erroneous Fo calculation. The steam can be shut off before total target F is reached, expecting that cooling phase lethality will increase the required overall Fo value. However, once the steam is off and cooling

THERMAL PROCESSING OF FOODS, PART I

293

commences, the lethal effect of the process is virtually fixed, and the contribution of the cooling phase is not properly evaluated and accounted for in terms of overall lethality. Hence, the standard practice is to attain the total a targeted F value in the heating phase and treat the cooling phase as a safety factor (Board et al., 1960). The weaknesses of this approach are that the process is longer, making it more expensive, and that destruction of heat-sensitive nutrients can be substantially greater than would occur if the minimum safe process were to be used (Cleland and Gesterkamp, 1983). F values determined in a small pilot plant retort are not directly applicable for the same process parameters in a commercial plant. Containers of different size processed for a particular F value do not have the same degree of commercial sterility. This is true because sterility is defined as the probability of survival of a single spore in the entire container at the cold point. Obviously, as the size and hence the volume increase, the containers should have a higher Fvalue to achieve the required final spore count (Cleland and Robertson, 1985). The F values are deduced by assuming pure conduction and pure convection heat transfer models. In practice, combined models usually exist. Particularly, the cooling phase time-temperature profile is highly irregular due to boiling, condensation, and agitation within the container (Cleland and Gesterkamp, 1983; Board et al., 1960). Therefore, there is inherent danger in expecting F to be identical for all cans unless the headspace within the container, the retort pressure during cooling, and the product temperature at the end of heating are closely monitored. Because of this uncertainty and complexity, the required lethality will be accomplished during the heating cycle, and that accomplished during cooling cycle will be a safety factor (Board et al., 1960). This may be safe, but will lead to overprocessing. Because of variations between individual containers, the rates of heat penetration may vary appreciably for the same product. Nevertheless, the same F, value is used though there are formulas to evolve the altered process schedules from one container size to another. Only rules of thumb are applied to get a revised process schedule, or the containers are simply processed for another 4-5 min. Therefore, it is always good practice to take at least three replicate containers for the experiment and perform sufficient experiments at each stage of product development to be sure that adequate processing is achieved under the most adverse conditions likely to be encountered (Shapton and Shapton, 1993). Food and Drug Administration (FDA) regulations demand that heat penetration data be taken from the regular production run. Despite this,

294

M. N. RAMESH et al.

a procedure commonly used in the canning industry to determine F values relies on data from a small number of cans processed in a pilot plant-scale retort. The implicit assumption is that the F value calculated for a few cans processed in a small retort will be the same as for production-run cans in a fully loaded commercial-scale retort (Robertson and Miller, 1984). The number of surviving spores is not easily determined because different parts of the container receive different thermal treatment, especially in the case of particulate foods. Thus, the assumption that the initial and final spore concentrations across a container are uniform will not be true. On the contrary, they will be position-dependent. With this uncertainty, calculation of the value of F using F = D (log No - log Nf) is also uncertain (Cleland and Robertson, 1985). This problem is minimized in pure convection because of mixing due to convection. This uncertainty can be considerably reduced by adopting numerical techniques with smaller increments. From time to time it is necessary to reestablish a process because changes in recipe, manufacturing method, filling method, or location can adversely alter the effectiveness of the sterilization process. Under these circumstances, heat penetration of other tests must be made to reestablish a satisfactory sterilization process. Process establishment using outside data is usually employed. When a new product is proposed, existing data are most often used in order to reduce the experimental work. In these circumstances, it is extremely important to ensure that the new product is essentially the same as the one for which the process was originally developed. Even a simple product, like a vegetable in brine, will have different process requirements depending on the size, shape, and amount of vegetable in the container (Shapton and Shapton, 1993).

E. UNCERTAINTIES DURINGPROCESSING OF CANNED PRODUCTS

The most common cause of C. botulinum outbreak is the failure of the operator to follow the specified process schedule (Pflug and Odlaug, 1978),which arises due to incorrect process calculation. All approaches applied for evaluating a thermal process schedule and for developing process lethality require the use of data that can only be determined experimentally. The usefulness of any calculated result is further limited by data uncertainties (Cleland and Robertson, 1985).

THERMAL PROCESSING OF FOODS, PART I

295

The National Canners’ Association (1968)has opined that one test run of 6 to 8 cans will provide sufficient data for homogeneous products. However, composition varies according to the processor used, the season, the preparation method, and many other factors. Hence, it is always safe to have data from several runs and then conclude that sufficient parameters, after ensuring the condition of slower heating, have been included in the study. With all of this, no reference has been made to the level of uncertainty associated with the calculated F value. According to Cleland and Robertson (1985), the reasons for underprocessing are: 1. Calculation of process parameters may not be sufficiently accurate, since the assumptions made to derive these parameters used are not valid for the circumstances under which they were derived. 2. The data used in calculations may not be precise. These data might be kinetic data for spore inactivation ( D and z values) or data used to determine temperature-time profiles. 3. The safety margin allowed is inadequate. Such process variations as initial product temperature, differences in steam supply to different parts of the retort, differences in retort venting and come-up time, and changes in cooling water temperature result in some of cans being processed improperly, thus yielding a lower F value. The scheduled thermal process has large and ill-defined unscientific safety margins (Board, 1977). In carrying out heat penetration studies, the process schedules have been standardized to decide on the number of test cans to be used for each run and the number of test runs to be adopted. The results reveal that the safety factor is quite considerable (Robertson and Miller, 1984). Though this important safety factor ensures a safe commercial process, it can lead to considerable nutritional loss. Product heating is dependent on both intrinsic factors (e.g.,the nature of product, filling weight, headspace, can size) and extrinsic factors (e.g., equipment, method of heating media, ambient conditions). Therefore, both of these factors must be taken into consideration when establishing a process because any factor that alters process lethality is a “critical factor” that must be specified in the schedule process. Extrinsic factors are related to the performance of filling and sterilization equipment. Filling is a factor critical not only for obtaining the correct solid-to-liquid ratio, but also in increasing the rate of heat transfer depending on headspace. An intrinsic factor like the cube size of diced vegetables can alter the rate of heat penetration and also the

296

M. N. RAMESH et al.

density of packing within the container, thereby altering the product heating data (Shapton and Shapton, 1993). Sterilizers of many different types (e.g., still retorts, agitating retorts, continuous retorts, plate sterilizers, aseptic units, and UHT units) are used in thermal processing of foods and use various heating media. The types and sizes of sterilization equipment have their own particular characteristics that must be known for both product heating rates and accuracy of sterilizing times and temperature. The standard practice in the canning industry is to follow the following assumptions, given by Stumbo et al. (1975): 1. Maximum 121.1”Cresistance of C. botulinum spores would be based on D = 0.20 min. 2. Prior to heat processing, all low-acid foods should be assumed to have a final population of the most resistant C. botulinum spores of one spore per gram (1cm3) of product. 3. The volume of any given container will be computed using the inside dimensions, with no correction for headspace. 4. The value of z, when characterizing the relative resistance of the most heat-resistant C. botulinum spores, should be taken as 14°C. 5. The probability of a C. botulinum spore surviving in any one heatprocessed container of low-acid food should be greater than With adoption of all of these assumptions, uncertainties in evaluating thermal process schedules still exist. F. INACCURACIES IN EVALUATING THERMAL PROCESSING

The formula method was introduced when many workers in the food industry were unable to appreciate its mathematical basis. Consequently, a “cookbook” approach to its use was developed. This approach persists, even when most food scientists have a better background in mathematics and physics. Ball’s formula method would be useful for food technologists and would lead to more intelligent use of process calculations. However, the formula method has its deficiencies, especially since it tends to yield inaccurate results. This may be due to: I. Calculations involving the value of fh and g found in published tables and graphs for these functions. fh is usually based on the eye-fit line or by linear regression.

THERMAL PROCESSING OF FOODS, PART I

297

2. Deviation from the assumed conditions in the original derivative

with respect to initial product temperature and retort temperature. In addition, g is calculated assuming uniform heating and cooling rates. 3. The difficulty in obtaining accurate values of required experimental parameters (e.g., f h , z). Despite these lacunae, Ball’s formula method has served the canning industry well since the 1930s. However, the method has undergone careful scrutiny, and a few errors have been found and possible solutions proposed. As more accurate computer methods based on numerical analysis become available for routine work, it is likely that the formula method will lose its prominence. Nonetheless, the development and application of this method will continue to stand as a milestone in food technology history (Merson et al., 1978). G. UNCERTAINTIES DUE TO pH

It is well known that the heat resistance of microorganisms is mostly determined by heat treatment parameters (time and temperature). However, many other microenvironmental factors, such as pH, water activity (aw),and medium composition, also exert influence. As the heat resistance of a microorganisms changes logarithmically with temperature, it is very important that the exact treatment temperature (with no undue fluctuations) be known. The effect of media pH on the heat resistance of microorganisms is very important. Some foods ( e g , macaroni, spaghetti, Spanish rice) have been reported to undergo changes in pH during heating (Brown and Thorpe, 1979; Reed et al., 1951, Xezones and Hutchings, 1965, Cerny, 1980; Montville and Sapers, 1981). Thus, ignorance of medium pH or of a change in pH during heat treatment can lead to misinterpretation of results (Brown and Thorpe, 1979; Reed et al., 1951). The composition of media used for TDT studies also exerts a significant effect on heat resistance (Pflug and Odlaug, 1978).Therefore, heat resistance should be determined in the foods themselves whenever possible (Stumbo, 1973; Brown and Ayres, 1982). The data in the literature on microbial resistance vary greatly. In addition to the natural variability in resistance of different spore crops of the same organism and in the age of spore suspensions, factors that contribute to data variability include the diversity of methods employed for TDT studies and differences in actual medium conditions that are

298

M. N. RAMESH et al.

not monitored during treatment (Schmidt, 1957; Brown and Ayres, 1982). Moreover, continuous monitoring of actual pH and temperature during heat treatment is not possible in TDT studies, and spores cannot be heated in highly viscous homogenized foods. Reichart (1983) has indicated that D values under 0.06 min for high-acid foods and at high temperatures cannot be determined by this method. Nevertheless, the capillary method has been used as the reference method for such applications (Davis, 1975; Burton et al., 1977; Mikolajacik and Rajkowsky, 1980).

It is a general assumption that the cutoff pH for low-acid foods is 4.6. However, the ability of pH 4.6 to inhibit growth of C. botulinum might be in doubt (Townsend et al., 1954; Thompson and Tanner, 1925; Baird-Parker and Freame, 1967; Odlaug and Pflug, 1979; Seeger, 1983). Rowley and Feeherry (1970) found that germination occurred at pH 6.5-7.5 and that no germination occurred below pH 4.8. Huhtanen et al. (1976) found that the minimum pH for outgrowth of spores with only a C. botulinum inoculum was pH 5.24. However, upon inoculation with mold and C. botulinum spores in tomato juice with an initial pH as low as 4.2, the medium became toxic. Upon taking the pH of the inoculated medium, they found a pH gradient ranging from the starting pH of 3.5 to as high as 8.2. This indicates that competitive microorganisms not normally present in properly prepared acid products could initially grow and change the environment to enhance conditions that support the growth of C. botulinum spores. Thus, growth in these acid foods is not due to the ability of C. botulinum to grow under the acidic conditions of the food (pH 4.6 or less), but rather to the pH rising above the inhibition level as a result of, for example, the action of other microorganisms. The above theory thus indicates that, although a pH of 4.6 will inhibit the growth of C. botulinum, the actual minimum pH at which growth is inhibited in a given food will be specific for that food and may be higher than pH 4.6 (Ito and Chen, 1978). However, no investigator appears to have critically controlled and quantitated anaerobiosis. Uncompromising exclusion of oxygen is known to encourage C. botulinum spore outgrowth under otherwise suboptimal conditions (Smith, 1975; Lund and Wyatt, 1984). C. botulinum is inhibited by oxygen because of its evident inability to synthesize catalase or superoxide dismutase. In the absence of these enzymes, which confer aerotolerance by dissimilating toxic oxides, oxygen and its products drain the anaerobe of cell reducing power. The essential reductants, for example, NAD(P) and NAD(H), are thus un-

THERMAL PROCESSING OF FOODS, PART I

299

available for yielding energy for cell metabolism. The ability of a microorganism to survive in an acidic medium may be related to the rate of proton migration into the cell relative to the proton-ejecting capacity of the cell (Corlett and Brown, 1980). In an oxygen-free substrate, cell energy that would otherwise be expended in scavenging oxygen and lowering the redox potential of the immediate environment, would be used to motivate expulsion of lethal hydrogen ions from the organism’s interior. Enhanced acid tolerance by C. botulinum in a relatively reduced system was first reported by Raatjees and Smelt (1979) and Smelt et al. (1982). C. botulinum was found to germinate, grow, and produce toxin in sterilized suspensions of soya protein (5.5%) acidified to pH 4 . 3 with hydrochloric acid and incubated in anaerobic jars. However, the results were not always reproducible, possibly because of varying oxygen tensions. The media were not prepared anaerobically, and the concentration of atmospheric oxygen in the jars was not measured. Nevertheless, these results suggest that a potential health hazard may exist in certain commercially produced foods that rely on acid to prevent C. botulinum growth. C. botulin um outgrowth and toxin formation in high-acid protein-rich environments is incompletely understood. The degree of anaerobiosis appears to be important, as does the extent of buffering. The protein may play several roles, such as a reducing agent, allowing the cell to overcome a limiting redox potential, a source of essential metabolites for growth and toxin production, and a buffer retarding acidification of the cell interior. Young-Perkins and Merson (1987) in their research have studied the effect of soya protein concentration on C. botulinurn spore germination, outgrowth, and toxin production in the presence of varying levels of an inorganic (hydrochloric) or organic (citric) acid under strict anaerobic conditions. They also investigated the interactions between pH, total acidity, and buffering capacity during the transition of C. botulinum from spore to active vegetative state at pH values less than 4.6; unequivocal evidence of C. botulinum spore germination, proliferation, mortality, and elaboration of neurotoxin under strict anaerobic conditions was firmly established. Although the pH of the medium and exposure temperature are known to influence the inactivation rates of bacteria, incorporation of these important environmental factors is not being followed in practical situations. Davey (1993) developed an extension of a generalized sterilization chart that combines temperature and pH. The procedure for its application is illustrated using inactivation of C. botulinum in a range of foods. This practical method should provide a convenient tool for

300

M. N. RAMESH et 01.

assessment of the necessary sterilization designs. It could be readily extended to a range of other foods and bacterial contaminants. Considering the various factors that affect thermal resistance, inoculum size, the number of strains, the medium, the incubation temperature, and the number of replicate samples used are very critical in determining the effect of pH of media during TDT studies (It0 and Chen, 1978). To optimize the probability of determining the minimum pH at which C. botulinum spores will grow in food, the following protocol may be considered (It0 and Chen, 1978): 1. The inoculum size that should present a natural contamination level of about lo4 spores per container would appear appropriate. 2. A reasonable number of types and strains should be utilized. Because of strain sensitivity to pH, different strains of each type should be utilized. 3. The product utilized should be the food to be tested rather than a bacteriological medium, or it should be food product supplemented with a bacteriological medium. 4. The product should be placed in 5-10 replicate tubes or other appropriate containers at a given pH level. 5 . The product should be incubated anaerobically at 30°C. If thioglycolate is utilized, care should be taken, as it may delay or prevent growth at the minimum pH level if used at too high a concentration 1966). (Segner et d., 6. The presence of toxin should be determined at the minimum pH level at which growth is observed. In addition, at least the next lower pH level should be examined for growth and presence of toxin. Raatjees and Smelt (1979) and Smelt et (11.(1982) have shown that the C. botulinum spore can grow at low pH values of 4.2. However, under specific conditions, the classified pH value for low-acid foods where C. botulinum is targeted is 4.6. Still, no attempt is made to redefine the low-acid food to include, foods with pH below 4.6. This leads to lower heat processing, as they are classified as high-acid foods. In spite of these reports, the knowledge available on the effects of reduced pH on bacterial spores is limited (Blocher and Busta, 1983). The most investigated species has been C. botulinum, and the evidence has thrown light on arriving at a pH range for this organism. However, there is a need for more data on the fundamental mechanisms of acid inhibition and studies on the practical implications of this information related to the safety of processed foods.

THERMAL PROCESSING OF FOODS, PART I

301

H. UNCERTAINTIES DUETO TEMPERATURE The cold point is defined as that point in a container that receives the least heat. This is assumed to be the slowest heating point in the container. Flambert and Deltour (1972) have shown that this point varies in the container over the whole heating and cooling cycle. They concluded that the location of the least-processed point lies on the vertical axis through the container for small Hlr, values and in a ringshaped region for large Hlr, values, where H is the height of the container and r, is radius of the container. Thermal processes are normally designed using time-temperature data measured in the slowest heating zone (the cold point). But for natural convection heating products, this cold point location varies during heating and cooling phases (Zechman and Pflug, 1989). The location of the cold point during natural convection heat transfer is determined by the fluid flow patterns of the product within the container. Zechman and Pflug (1989)concluded that the slowest heating zone along the vertical axis of vertically positioned metal containers with natural convection-heated liquids may be located near the bottom of the container, but usually not farther than 25% of the container height from its base. The size of the slowest heating zone is smallest, and therefore the exact location is more critical in the design of processes, for liquids of high initial viscosity in small containers. Hence, the use of one-point location for a specific container size for all natural convection-heated products is not appropriate. Since the cold point moves during various stages of heating, the location of the slowest heating zone may be different for the same product and container size when the process time or temperature is significantly changed. In conduction heating of products at the steam-off point, large temperature gradients can exist from the outside surface to the cold point. This leads to a kind of temperature abuse called overshooting during the initial stages of cooling. The F value of this period is appreciable and is on the order of 13% of the Fo value (Robertson and Miller, 1984). Robertson and Miller (1984) worked on some of the uncertainties between can-to-can variations and run variations and deduced that the variations may be due to the position of the thermocouple point, variations in retort operation, heat conduction along the thermocouple wire affecting temperature readings, differences in headspace, and the accuracy of the thermocouples. According to Navankasattusas and Lund (1978), a thermocouple with an accuracy range of fO.l-l°C will lead to a corresponding error in accomplishing lethality of 2.3

302

M. N.RAMESH et al.

to 26%. Cowell et al. (1959) reported that the variations in Fo values are largest during the early stages of heating and cooling. The FDA specifies that at least one mercury glass thermometer with divisions that are easily readable to 1°F be used for heat penetration studies. However, Pflug and Odlaug (1978) calculated that, if the accuracy of can center temperature measurement is +1"C, then the potential error in the value of Fo due to this temperature variation is about +12%. Hence, they have suggested that, for low-acid foods, Fvalues of 2-6 may be listed in steps of 0.5 min, 6-10 by 1min, 10-20 by 2 min, and greater than 20 by 5 min [Merson et a]., 1978). The method first introduced by Ball (1923) of considering 42% of the come-up time to be lethal (with the retort at the processing temperature) and the remaining 58% to be nonlethal is still widely used. Research to verify this procedure is necessary. The basic heat penetration data required for evolving time-temperature profiles are collected by measuring the temperature at regular intervals of time at the cold point of containers. Usually, T-type plug-in thermocouples are used for this purpose. However, Bee and Park (1978) listed 25 cases of unreliable data used in conducting heat penetration studies. Hence, care should be taken to avoid all these problems when collecting heat penetration data. Processors commonly conduct heat penetration tests in the laboratory and apply the data in establishing thermal processing requirements to other conditions. This is particularly true for the conditions of retort temperature (RT) and the initial temperature (IT) of the product when processing times are predicted by the Ball formula method (Stumbo, 1973).

The FDA reviews processes filed for low-acid canned foods for adequacy in protecting public health (Mulvaney eta].,1978). These reviews have revealed that the temperature difference at which the heat penetration data were taken and for which thermal processes were filed can range from a few degrees to as much as 50°C. Using products that exhibit broken-heating curves, Berry and Bush (1987) showed that the procedure of extrapolating heating data to other RTs and ITS can have a significant influence on predicting required processing times. Extrapolation of broken-heating heat penetration data to a higher value is a conservative practice. Predicted processing time determined from data taken at a lower RT will be greater than the processing time actually required at a higher RT. However, the converse is not true. Berry and Bush (1987) demonstrated that, by taking heat penetration data at higher RTs, higher than those intended for estab-

THERMAL PROCESSING OF FOODS, PART I

303

lishment of the thermal process, resulted in a calculated processing time as much as 10 min shorter than that demonstrated to be necessary. The effect of IT was even greater than that of RT for products exhibiting broken-heating curves. Their conclusion was that heat penetration data should be taken at the highest IT intended to make the process more safe. Extrapolation of heat penetration data taken at other RTs and ITS for product heating with straight-line heating curves as a result of significant induced convection can have an influence on calculated processing times. In general, this influence is less than that of product heating with broken-heating curves in a still retort. Taking heat penetration data at one RT and calculating processes at different RTs will have no effect on processing time for either rapid-heating or conduction-heating products. However, the effect will be significant for intermediate-viscosity products or large can size. The influence of extrapolation to different ITS is more significant with regard to determining required processing time than extrapolating data taken at different RTs, but the results will not be as consistent as those obtained from extrapolation of RT. The effect of product IT on heating depends on the product, and no compromise can be made as to the conservative direction of extrapolating data taken at different ITS, as it is product-dependent. Thus, extrapolating heating data to other conditions (RT or IT) can affect the establishment of thermal processes. The significance of this effect depends on the product, and serious underprocessing can occur with certain products and/or processing conditions. The conservative nature of the Ball formula method should not be expected to overcome this phenomenon, particularly for broken-heating products, where calculated processing times taken at other ITS will be deficient by as much as 50% (Berry and Bush, 1987; 1989). I. UNCERTAINTIES IN THE SAFETY FACTOR

It is felt that it is not feasible to apply valid statistical techniques because uncertainties in the data used to derive the processing parameters cannot be adequately assessed (Hicks, 1961). It is also argued that the safety factor is required for its own sake because the quantity or process being calculated must conform to the FDA safety standard, which cannot be specified very precisely in terms of the quantities calculated (Hicks, 1961). Hence, in the food processing industry it is probably necessary to think in terms of safety factors rather than precise confidence limits.

3 04

M. N. RAMESH et al.

Safety factors in process calculations are not always explicitly recognized. They are often introduced implicitly based on calculations of process schedules assuming initial spore population with specific properties (like D and z values). This hypothetical initial population is probably much harder to destroy than that likely to occur in commercial processed foods (Hicks, 1961). There is no specific point in the process calculation procedure at which to introduce a safety factor. That being the case, where and how does one compensate for the unknown? The canning industry must deal with a biological product that contains microorganisms. Product and processing conditions vary widely. Allowances should be made for unknown conditions, and there must be a way to compensate for realistic extremes. In this context, one should recognize the availability of scientific computing methods for determining the safety factors discussed in a later section. In the determination of process schedules, safety factors can be introduced in the response parameter, fh, and/or the lag factor, j . In earlier practices, a safety factor has been incorporated in the value of Fo itself. When carrying out heat penetration tests, the sterilization process engineer may introduce safety factors by using the slowest heating can in the design and overfill or make other manipulations to compensate for the unknown conditions. It has been stated as a rule of thumb that the Fo value for a continuous processed product should be double the Fo value used in the design of a still cook (Perkins, 1969). Perkins (1969) incorporated a similar safety factor in recommendations for the sterilization of hospital supplies based on his wide experience with sterilizers and sterilization conditions in hospitals. Lenz and Lund (1977a) and Lund (1978) have suggested that safety factors for canned foods may be calculated using statistical methods. At the same time, they have pointed out that there are not enough available background data to pursue these methods. Cues for Safety Factor Selection. Since the public health sterilization value requirement must include adequate safety, not only for microbiological variation, but also for process delivery variation, these factors should be considered in selecting the values of both F and z. Process delivery variation and the potential accompanying hazards are functions of the product, the container, and the processing system. In general, a greater safety factor is required as we go from metal to glass and other exotic packaging systems. The overall safety factor must be increased as we go from a conduction-heating product or a pure convection-heating product to products heated by both conduction and convection that may exhibit broken-heating curves. The safety factor must

THERMAL PROCESSING OF FOODS, PART I

305

increase as we go hom long-time processes at low temperatures to short-time processes at higher temperatures and as we go from still cooks to agitating cooks (Pflug and Odlaug, 1978). Ill. Uncertainties In Aseptic Systems

A. UNCERTAINTIES IN THE ASEPTIC VALUE OF z

The process assumes a constant z value irrespective of the applied temperature, as in TDT studies on a laboratory scale. Product sterilizers have confirmed that at temperatures above 120°C the value of z continuously increases, making extrapolation of TDT curves a suspect practice for high temperatures (Kaplan, 1984; Bockelmann, 1985).

B. UNCERTAINTIES IN THE ASEPTIC VALUE OF F Recommended Fo values have often been larger than necessary to destroy a given population of microorganism (Dignan et a]., 1989). This is to ensure that a variation in processing conditions will not impede achieving commercial sterility. It is a general practice to consider the lethality achieved during cooling as the safety factor. c.

UNCERTAINTIES IN

ASEPTIC PROCESSING

Having made a decision as to the amount of lethal heat required, it is not certain how the high temperatures and short times can be measured (Shapton and Shapton, 1993). Accuracy of measurement is of prime importance because of the way in which small changes in these values have a large numerical effect on evaluation of the process. Static temperatures are used to measure the dynamic nature of the aseptically processed food product. Even for a simple liquid-phase product, process evaluation is not straightforward. As prescribed by the FDA, only exposure in the holding section is being used. The contributions of the heating and cooling sections are not taken into account. Such a process schedule would result in significantly greater nutrient destruction than may be acceptable to consumers (Chandrana and Gavin, 1989). Hence, the very objective of aseptic processing to retain nutrients by high-temperature shorttime (HTST) processing is defeated. The type of flow in the heating and holding sections is a critical factor, since the type of flow determines the relationship of the minimum

306

M. N.RAMESH et a].

residence time to the average or mean residence time. There is much to be learned about the dynamics of a food particle carried by a viscous fluid through an aseptic processing system. Each system is unique, as is each section within each system. The establishment of a thermal process must be specific for a food product and processing system (Dignan et al., 1989). In modern HTST sterilization, application of temperature ranges from 130 to 150°C for a few seconds results in retention of valuable vitamins. From the experiments of Srimani et al. (1980),it is evident that it is not possible to destroy all microorganisms when they are heated to a high temperature and cooled down immediately within an average retention time of 60 sec. Thus, heat-resistant spores have to be destroyed by chemical or other means to ensure absolute sterilization in the HTST process. Process schedule evaluation for particulates is even more complex since they are usually carried or suspended in a liquid phase. The size, composition, and integrity of particles are critically important, together with such factors as the solid-to-liquid ratio. However, none of the above methods are considered in practice, and only rules of thumb for extrapolation of the F values of still processing are followed. There do exist several mathematical approaches to define F values, but they remain theories that have not been transferred to processing industries. Burton et al. (1977) found considerable discrepancies between data obtained for B. stearothermophilus in milk from capillary tubes and UHT sterilizers, and they urged caution in process evaluation. Srimani et al. (1980) found that the D value for B. stearothermophilus (without heat activation) and B. subtilis remained constant above 135 and 12O"C, respectively. Some researchers, including Manji and Van de Voort (1985), favored the Arrhenius model over the thermal death time (TDT) model for the temperature dependence of bacterial spore death, although recognizing that the differences may not be practically significant. Rodriguez et al. (1987) have developed a sophisticated model for spore death kinetics based on considerations of population dynamics and systems analysis. These and other studies suggest that the process designer must use caution in extrapolation of TDT data to aseptic temperatures, and that a suitable margin of safety should be employed in each situation. Process evaluation for large food particle suspensions is considerably more difficult than for liquids. The principal difficulty is in measurement of cold zone temperatures of individual food particles during continuous flow. When a suspension is heated within heat exchangers, the liquid heats up relatively rapidly, but particle thermal response may

THERMAL PROCESSING OF FOODS, PART I

307

be considerably slower. Further complications are introduced by the existence of residence time distributions in the process equipment (both heat exchangers and holding tubes) and an unknown value of the fluid-to-particle convective heat transfer coefficient. Consequently, process evaluation depends on development of accurate mathematical models for prediction of particle cold spot temperatures and reliable techniques for microbiological validation. Because of these limitations in aseptic processing, much effort is needed to establish the sterility of a newly installed plant. Examination after incubation of large numbers of samples taken from the filling line is necessary to ascertain whether or not the plant is operating in accordance with FDA standards (Hersom, 1985). A particular system used for filling one product is unique to that product, as the flow rate is dependent upon the viscosity and residence time of that particular food product. If any change in the product is envisaged, this calls for a major change in the equipment or the operating parameters. This is not economical for some food products. Different kinds of aseptic processing of such food forms as pulp, liquid, and particulate foods are in many ways distinctive. The sterilization methods used vary enormously depending on adaptation of direct, indirect, or electrical methods (Hersom, 1985). This indicates that it requires significantly more complex techniques than does the manufacture of canned foods in conventional retorts (Ahvenainen, 1988). Thermal death in the aseptic processing line is not the same as that with such small-scale laboratory methods as thermal shock, and the thermal death curve of a targeted microorganism cannot be efficiently simulated (Swartzel, 1985). This indicates that the best final result is obtained by testing sterilization effectiveness in the aseptic processing line itself, as there are no on-line methods for determining the sterility of a product. Only one final product evaluation for sterility (incubated packs) does not provide the necessary microbial safety. Other quality control measures (e.g., pH, pack integrity, seal efficiency) have to be used as complements (Brown and Ayres, 1982). Since aseptic processing is carried out at a high temperature of 12O-13O0C, even 1 min of overprocessing will lead to loss of product nutrients. Hence, the process schedule should be derived more accurately compared to canning (Lopez, 1987). Selection of the sterilization system that is most advantageous for a specific food product depends on many factors. The details of the system must be determined so that the process as a whole produces high-quality commercially sterile products. This particular setting is unique for that product. Since the product is first

308

M. N. RAMESH et al.

sterilized and then packed, utmost care should be taken to maintain system sterility. IV. Statistical Analysis of Thermal Process Calculations

When the uncertainties of thermal processing involving heat penetration studies and determination of thermal resistance data cannot be addressed, there are methods that can be used to evolve the safety factor so as to ensure the commercial sterility of a product with optimum sterilization. This also reduces losses of vitamins and other essential nutrients. Hicks (1961) was the first to provide data on statistical analysis of thermal process calculations. He studied the uncertainties due to physical measurement of temperature and those due to biological uncertainties in spore count. Different statistical distributions (e.g., normal, Poisson) have been applied to study the variation of thermal processing data. Massaguer et al. (1983) used a Poisson distribution to describe the error population of thermal process data, while Lund (1978) has used a normal distribution. As indicated by Hicks (1961), a sensible approach would be to define process schedules in terms of confidence levels or the coefficient of variations rather than adding safety factors implicitly. The coefficient of variation (CV) has been applied to define the probability of thermal process calculations in achieving commercial sterility. A CV can be defined for individual uncertain process parameters like the theoretical initial temperature (IT*), and g, D, f, and z values (Esselen et al., 1951; Tung and Britt, 1995). Herndon (1971) reported a CV of 25% for theoretical values in convection-heating foods. Hicks (1961) used a mean CV of 7%, a value obtained by considering the data of Hurwicz and Tischer (1956b) and Jackson and Olson (1940). Other statistical methods have been applied for evaluation of the D value of microorganisms. Lewis (1956) has extensively reviewed the application of Spellman-Karber and log-log methods for estimating the process schedule parameter and its standard deviations from experimental data. From his methods, the CV of the D value was approximately 10%. Stumbo et al. (1950) reported a CV in the z value of about 2%; whereas Kaplan et al. (1954) obtained a CV of 9% for PA3679, with a mean z value of 18.7"F. Hicks (1961) estimated that the effect of a CV of 2% for the z value would result in a CV of 5% in the calculated process time value; whereas a CV of 9% for the z value would result in

THERMAL PROCESSING OF FOODS. PART I

309

a CV of 23% in the calculated process time. Further, Hicks (1961) pointed out that the variances in D and z values are not independent. The dependent nature of the variances had not been studied until Lenz and Lund (1977b), who reported calculation results incorporating the dependent nature of the D and z values. Although the coefficient of variation is a valuable statistical parameter in that it provides information on the reliability or probability of having a given value of a parameter, its real significance is to incorporate uncertainty into the lethality calculation process. The usual method, as used by Hicks (1961), to incorporate statistical parameters into lethality calculations is to perform error analysis using the coefficient of variation. Although this is an entirely appropriate approach, there are other approaches that should be investigated. Recently, Lenz and Lund (1976,1977b) used a Monte Carlo procedure to determine the effect of parameters describing heating rate and microbial destruction on certainty of calculated lethality. In one study (Lenz and Lund, 1977b), lethality distributions were generated for conduction-heating foods, while in another study Lenz and Lund (1976) examined convection heating. Lund (1978) worked on a statistical analysis of thermal process calculations. In a Monte Carlo procedure, the confidence interval for calculated lethality is estimated by: 1. Establishing the population distribution of each variable under normal processing conditions by randomly choosing a value for each variable. 2. Calculating lethality using an appropriate calculation model. Lenz and Lund (197713) suggested that a knowledge of the standard deviation of lethality could be applied to establish a minimum safety factor that could be incorporated into the system. Finally, one should realize that the variability in calculated process schedules emanates from experimentally determined parameters that are subject to biological variations and from the accuracy range of measuring system parameters. The function of quantifying that with ability would be to use the information in assigning reasonable safety factors that would accomplish the objective of the process (inactivation of spores to ensure commercial sterility) while minimizing destruction of quality factors (Lund, 1978). Patino and Heil (1985) introduced a statistical approach to error analysis in thermal process calculations. Traditionally, D and z are felt to be separate parameters that arise from the two-stage procedure of first estimating the value of D and then estimating a z value from

310

M. N. RAMESH et a].

thermal resistance curves. But Patino and Heil(l985) showed that these two parameters are highly correlated. When the values of D and z are estimated simultaneously by using nonlinear least squares, the correlation coefficient was as high as 0.92-0.96. When these variables cannot be completely eliminated, one alternative is to develop scientific methods to estimate and accurately define the safety factors. Hayakawa et al. (1988)have developed a computerized procedure for estimating the variability of process lethality when there are variations in all independent thermal process parameters of conduction-heating food packages. They have applied a well-known statistical procedure of the Monte Car10 technique combined with a reliable mathematical method for thermal process evaluation. The coefficients of variation in sterilizing values (Fo) were estimated from heat penetration data collected by processing 2 1 1 x 300 and 307 x 409 cans of spaghetti in tomato sauce (60 cans each). Their results agreed well with those computed by the developed computerized procedure. V. Suggestions for Future Work

It is evident from this review that a great deal of attention has been focused on the effect of pH on sterility and the attempts to redefine lowand high-acid cutoff. Also, there is a need for more data on the fundamental mechanisms of acid inhibition and studies on the practical implications of this information related to the safety of processed foods. However, there have been few investigations on the effect of uncertainties in D and z values on process schedules. Studies on these issues would yield deeper insight into the concepts and would facilitate redefinition of this terminology. It is important that D and z values be evaluated simultaneously at the actual conditions in the food being studied. There is also a paucity of information about on-line methods for determining sterility during aseptic processing. The several limitations of aseptic processing that have been brought out in this review provide food for thought to direct future research. Studies on these topics would provide further insight into the roles of such alternative heating methods as ohmic heating and microwave heating, particularly in the case of particulate foods. It has been established that a loerror in measurement of the temperature of processed food will lead to a 13 to 15% difference in the sterility value. This is an important matter of concern that demands adoption of more accurate means of temperature measurement.

THERMAL PROCESSING OF FOODS, PART I

311

Hence, the future course of research should be mainly directed toward eliminating these uncertainties. There is a need to develop computer-aided methods for evaluating safety factors based on heat penetration data, to develop on-line control systems for achieving commercial sterility, and to develop direct methods for evaluating thermal processing by detecting living microbes, instead of D and z values. Improvements in aseptic processing should be made by devising on-line sterility evaluation techniques. Concerted efforts must be focused by academia and industry to develop methods and alternatives to eliminate the above-discussed uncertainties. Focusing on the above areas will provide a holistic approach for managing food safety in this era of streamlined processing and deregulation. Attempts should be made to improve the hitherto employed thermal processing techniques of the nineteenth century to lead to better thermal processing during the twenty-first century. VI. Conclusions

Despite these uncertainties, the long and continuing successes of commercial food processing are in large part due to the wide-ranging safety factors inherent in practice. Hence, it is necessary to develop alternative methods to minimize safety margins and thereby reduce nutrient losses and processing costs. One alternative is to develop either on- or off-line methods to ensure commercial sterility. Rapid microbial methods are being developed to evaluate thermally processed products for their sterility. Another area is development of on-line control systems. Though these approaches may not eliminate all uncertainties, they will at least ensure the development of minimum time-temperature profiles, with particular reference to nutrient retention, which will finally facilitate optimum sterilization. Subsequent parts of the series will deal with some of these techniques for optimum sterilization. It is important to prevent microbial contamination right from the point of harvesting foods through the preparation of recipes at the customer end. The hazard analysis and critical control point (HACCP) method and the good manufacturing practices (GMP) method are being developed for different food processing environments. These will be discussed in subsequent parts of this series. The research on eliminating the above-discussed uncertainties should reach the food processor as results become available. One possibility is to involve the food processor at the R&D stage of any new research in this direction. At the same time, the active participation of food industries will facilitate immediate implementation of the research outcome. The final result

312

M. N. RAMESH et al.

will be healthy, nutritious, and safe foods for customers, which is the basic objective of thermal processing. ACKNOWLEDGMENTS

The authors thank V. Prakash, Director of the Central Food Technological Research Institute, and A. Ramesh, N. G. Karanth, and R. Venkatakuppaiah, for their encouragement, and M. Asha, for her help in preparation of the manuscript. REFERENCES

Ahvenainen, R. (1988). Food Rev. Int. 4(1), 45. Alstrand, D. V., and Ecklund, 0. F. (1952). Food Technol. 6(5), 185. Baird-Parker, A. C., and Freame, B. (1967). J. Appl. Bacteriol. 30, 420. Ball, C. 0. (1923). National Research Council Bulletin, Vol. 7, Part 1, No. 37. Ball, C. O., and Olson, F. C. W. (1955). Food Res. 20, 666. Bee, G. R., and Park, D. K. (1978). Food Technol. 32, 56. Berry Jr., M. R., and Bradshaw, J. G. (1986). J. Food Sci. 51(2),477. Berry Jr., M. R., and Bush, R. C. (1987). J. Food Sci. 52,958. Berry Jr., M. R., and Bush, R. C. (1989). J. Food Sci. 54(4), 1040. Blocher, J. C., and Busta, F. F. (1983). Food Technol. 37(11), 87. Board, P. W. (1977). CSIRO Division of Food Research, Circular No. 7, Australian CSIRO. Board, P. W., Cowell, N. D., and Hicks, E. W. (1960). Food Res. 25,449. Bockelmann, I. V. (1985). Proc. IUFOST Symp. Aseptic Processing and Packaging of Foods, p. 134. Brown, K. L., and Ayres, C. A. (1982). In “Developments in Food Microbiology” (R. Davies, ed.), p. 119. Applied Science, England. Brown, K. L., and Thorpe, R. H. (1979).Tech. Memo No. 185, Campden Food Preservation Research Association, Chipping, England. Brown, K. L., Pitcher, Y. E., and Mottishaw, Y. (1981). Tech. Memo No. 281. Campden Food Preservation Research Association, Chipping, England. Burton, H., Perkins, A. G., Davies, F. L., and Underwood, H. H. (1977). J. Food Technol. 12,149.

Cerny, G. (1980). Zlebersm. Linters Forsch. 170, 180. Chandrana, D. I., and Gavin, A. (1989). J. Food Sci. 54(1), 198. Cleland, A. C., and Gesterkamp, M. F. (1983). J. Food Technol. 18, 411. Cleland, A. C., and Robertson, G. L. (1985). In “Developments in Food Preservation” (S. Thorne, ed.), Vol. 3, pp 1-43. Elsevier Applied Science Publishers, London. Corcos, A. (1975). Food Technol. 29(5), 114. Corlett Jr., D. A,, and Brown, M. H. (1980). In “International Commission on Microbiological Specifications for Foods,” Vol. 1, p. 92. Academic Press, New York. Cowell, N. D., Evans, H. L., Hicks, E. W., and Mellor, J. D. (1959). Food Technol. 13, 425. Daudin, J. D., and Cerf, 0. (1977). Lebens.-Wiss. U. Technol. 10, 203. Davey, K. R. (1993). Lebens.-Wiss. U. Technol. 26,476. Davis, F. L. (1975). J. Soc. Dairy Technol. 28(2), 69. Dignan, D. M., Berry, M. R., Pflug, I. J., and Gardine, T. D. (1989). Food Techno]. 3, 118. Erdtsleck, B., and Becimer, R. (1976). Lebens.-Wiss. U. Techno]. 9(3), 185.

THERMAL PROCESSING OF FOODS, PART I

313

Esselen, W. B., Anderson, E. E., Ruder, L. F., and Pflug, I. J. (1951). Food Technol. 5, 279. Flambert, F., and Deltour, J. (1972). Lebens.-Wiss. U. Technol. 5(1), 7. Gillepsey, T. G. (1951). J. Sci. Food Agric. 2, 107. Hansen, N. H., and Riemann, H. (1963). J. Appl. Ract. 26, 314 Hayakawa, K. (1982). Can. Inst. Food Sci. Technol. J. 15,116. Hayakawa, K., Matsuda, N., Komaki, K., and Matsunawa, K. (1981). Lebens.-Wiss. U. Technol. 14,70. Hayakawa, K., Massaguer, l? D., and Trout, R. J. (1988). J. Food Sci. 53(6), 1887. Herndon, D. H. (1971). J. Food Sci. 36, 299. Hersom, A. C. (1985). FoodRes. Int. 1(2), 215. Hicks, E. W. (1951). Food Technol. 5, 134. Hicks, E. W. (1961). FoodRes. 26, 218. Huhtanen, C. N., Nagahshi, J., Custer, C. S., and Russell, R. W. (1976). Appl. Environ. Microbiol. 32, 711. Hurwicz, H., and Tischer, R. G. (1956a). Food Res. 20, 377. Hurwicz, H., and Tischer, R. G. (1956b). Food Res. 21, 147. Ito, K. A., and Chen, J. K. (1978). Food Technol. 32, 71. Jackson, J. M., and Olson, F. C. W. (1940). Food Res. 5, 409. Jonsson, V., Snugg, B. G., Harnur, B. G., and Zachrisson, T. J. (1977). J. Food Sci. 42,1251. Kaplan, A. M., Reynolds, H., and Lichtenstein, H. (1954). Food Res. 19,173. Kaplan, S. R. (1984). Beverage World 103(1329), 60. Lenz, M. K., and Lund, D. B. (1976). Proc. Intl. Cong. Eng. Food, Zst, Boston Lenz, M. K., and Lund, D. B. (1977a). J. Food Sci. 42,989. Lenz, M. K., and Lund, D. B. (1977b). J. Food Sci. 42, 1002. Lewis, J. C. (1956). Appl. Microbiol. 4,211. Lopez, A. (1987). In “A Complete Course in Canning, Part 11” (A. Lopez, ed.). The Canning Trade, Baltimore. Lund, B. M., and Wyatt, G. M. (1984). FoodMicrobiol. 1,49. Lund, D.B. (1978). Food Technol. 32(3), 76. Manji, B., and Van de Voort, F. R. (1985). J. Food Prot. 48,359. Massaguer, P. R., Trout, R., and Kayakawa, K. (1983). Proc. Annu. Mtg. Food Technologists, 43rd, New Orleans, LA, Paper No. 432. Merson, R. L., Singh, R. P., and Carroad, P. A. (1978). Food Technol. 32, 66. Mikolajacik, E. M., and Rajkowsky, K. T. (1980). J. Food Prot. 43, 799. Moats, W. A,, Dabbah, R., and Edwards, V. M. (1971). J. Food Sci. 36, 523. Montville, T. J., and Sapers, G. M. (1981). J. Food Sci. 46, 1710. Mulvaney, T.R., Shaffner, R. M., Miller, R. A., and Johnson, M. R. (1978). Food Technol. 32(6), 73.

National Canners’ Association (1968). In “Laboratory Manual for Food Canners and Processors,” Vol. 1. AVI Publishing, Westport, CT. Nath, N., and Ranganna, S. (1983). J. Food Technol. 18,301. Navankasattusas, S.,and Lund, D. B. (1978). Food Technol. 32(3), 79. Odlaug, T. E., and Pflug, I. J. (1979). Appl. Environ. Microbiol. 37, 496. Ott, T. M., El-Bisi, H. M., and Esselen, W. B. (1961). J. Food Sci. 26, 1. Patino, H., and Heil, J. R. (1985). J. Food Sci. 50, 1110. Perkins, J. J. (1969). In “Principles and Methods of Sterilization.” Charles C. Thompson Co., Springfield, IL. Perkins, W. E., Ashton, D. H., and Evancho, G. M. (1975). J. Food Sci. 40, 1189. Pflug, I. J., and Odlaug, T. E. (1978). Food Technol. 32(6), 63.

3 14

M. N. RAMESH et a].

Pilcher, T. (1949). In “The Canned Food Reference Manual” (T. Pilcher, ed.), p. 320. American Can, New York. Powers, J. J., Pratt, D. E., Cannon, J. L., Somaotmadja, D., and Fortson, J. L. (1962). Food Techno]. 16, 80. Raatjees, G. J. M., and Smelt, J. P. P. M. (1979). Nature 281, 398. Reed, J. M., Bohrer, C. W., and Cameron, E. J. (1951). Food Res. 16, 383. Reichart, 0. (1983). Acta Alimentria 12(1), 35. Robertson, G. L., and Miller, S. L. (1984). J. Food Technol. 19, 631. Rodriguez, A. C., Teixeira, A. A., and Smerage, G. H. (1987) Proc. Annu. Mtg. Food Technologists, 47th, Las Vegas, NV. Rowley, D. B., and Feeherry, F. (1970). J. Bacteriol. 104, 1151. Saikia, L., and Ranganna, S. (1992). J. Food Sci. Tech. 29(4), 203. Schmidt, C. F. (1957). In “Antiseptics, Disinfectants, Fungicides, and Sterilization” (G. F. Reddissch, ed.). Lea and Febiger, Philadelphia. Seeger, M. L. (1983). NFPA Research Laboratories Annual Report, pp. 33-35. Segner, W. P., Schmidt, C. F., and Boltz, J. K. (1966). Appl. Microbial. 14, 49. Shannon, E. L., Reinbold, G. W., and Clark, W. S. (1970). J. Milk Food Technol. 33, 192. Shapton, D. A., and Shapton, N. F. (1993). In “Principles and Practices for the Safe Processing of Foods” (D. A. Shapton and N. F. Shapton, eds.), p. 334. Butterworth, London. Smelt, J. P. P. M., Raatjees, G. J. M., Crowther, J. S., and Verrips, C. T. (1982). J. Appl. Bacteriol. 52, 75. Smith, M. V. (1975). The effect of oxidation-reduction potential on the outgrowth and chemical inhibition of C. botulinum type E spores. Ph.D. thesis, Virginia Polytechnic Institute and State University, Blacksburg. Srimani, B., Reinhold, S., and Locin, M. (1980). Lebens.-Wiss. U. Techno]. 13, 186. Stumbo, C. R. (1973). In “Thermobacteriology in Food Processing” [C. R. Stumho, ed.). Academic Press, New York. Stumbo, C. R., Murphy, J. R., and Cochran, J. (1950). Food Techno]. 4, 321. Stumho, C. R., Purohit, K. S., and Ramakrishnan, T. V. (1975). J. Food Sci. 40, 1316. Stumbo, C. R., Prohit, K. S., Ramakrishnan, T. V., Erans, 0. H., and Francis, F. J. (1983). In “Handbook of Lethality Guides for Low-Acid Canned Foods,” Vol. 1. CRC Press, Boca Raton, FL. Swartzel, K. R. (1985). Proc. IUFOST Symp. Aseptic Processing and Packaging of Foods, p. 22.

Thompson, L., and Tanner, F. W. (1925). J. Infect. Disease 37, 344. Townsend, C. T., Yec, L., and Mercer, W. A. (1954). Food Res. 19, 536. Tung, M. A., and Britt, I. J. (1995). Food Res. Int. 28(1), 101. Vas, K., and Proszt, G. (1957). J. Appl. Bacteriol. 20, 431. Xezones, H., and Hutchings, I. J. (1965). Food Techno]. 19, 1003. Yawger, E. S. (1978). Food Technol. 32, 59. Young-Perkins, K. E., and Merson, R. L. (1987). J. Food Sci. 52(4), 1084. Zechman, L. G., and Pflug, I. J. (1989). J. Food. Sci. 54(1), 205.

Thermal Processing of Foods, A Retrospective, Part II: On-Line Methods for Ensuring Commercial Sterility M. N. RAMESH Food Engineering Department Central Food Technological Research Institute Mysore 570 013, India

M. A. KUMAR Central Instrum en ts Facility Central Food Technological Research Institute Mysore 570 013, India

S. G. PRAPULLA Fermentation Technology and Bioengineering Department Central Food Technological Research Institute Mysore 570 013, India

M. MAHADEVAIAH Food Packaging Technology Department Central Food Technological Research Institute Mysore 570 013, India

Introduction Fo Integrators On-Line Monitoring Systems Semiautomatic Retort Control Systems for Optimum Sterilization Computer-Aided Sterilization Process Indicators A. Bioindicators B. Color-Based Physical Indicators C. Chemical Markers D. Electronic Indicators VII. Suggestions for Future Work VIII. Conclusions References I. 11. 111. IV V VI.

315 ADVANCES IN APPLIED MICROBIOLOGY, VOLUME 44 Copyright 0 1997 by Academic Press, Inc. All rights of reproduction in any form reserved. on~,5-2164/97$25.00

316

M. N.RAMESH et al.

I. Introduction

The major cause for spoilage of thermally processed foods is the failure to follow a scientifically derived process schedule. The problem does not appear to lie in the process engineering area, where thermal process schedules are developed, but primarily in the manufacturing and production areas. The use of proper on-line methods involving trained personnel in canning plants is of utmost importance. The reason for the continuing spoilage and overprocessing of canned foods is that canning plants are often operated by untrained personnel who fail to administer the recommended sterilization processes. Though there is no way to remove the uncertainties in thermal processing, attempts have been made by food technologists and food engineers to critically evaluate the process, which have led to the development of control systems and instruments to ensure the commercial sterility of processed food products. These instruments indicate the sterility during processing, and the control system continuously monitors and controls the sterility value on-line and ensures commercial sterility without loss of nutrients. On-line methods are used to ensure minimum sterility, so as to avoid overprocessing and to facilitate nutrient retention. This indicates the actual time-temperature profile being experienced by the food. However, this method cannot eliminate the uncertainties involved in the development of time-temperature profiles. By this approach, a targeted sterility value can be accurately controlled and the effect of temperature abuse can be reduced to yield a better product, and one safe for immediate consumption. Sterilizing operations are used commercially in a number of industries, particularly in food and pharmaceuticals. The safety of the product being sterilized depends on correct operation of the sterilizing system, and thus a successful control procedure. The criterion for detection of adequate sterility is the sterilizing value Fo, which is determined by conducting heat-penetration tests by placing a temperature-measuring device at the point of slowest heating. During control operations, the temperature of the cold point is monitored so as to attain the prescribed Fo value. Simple thermocouples are used to determine the temperature distribution in commercial retorts. This chapter describes the on-line methods applied for within-container and aseptic processing. On-line methods for sterilization can be subclassified as:

THERMAL PROCESSING OF FOODS, PART I1

317

1. F,, integrators 2. On-line monitoring systems for determination of asepticity 3. Semiautomatic retort control systems 4. Computer-aided sterilization 5. Process indicators (time-temperature)

a. Bioindicators b. Color-based physical indicators c. Chemical markers d. Electronic indicators

II. Fo Integrators This is an instrument or a data acquisition system used to measure the accomplished lethality. It employs analog computing techniques to calculate process lethality or the F value of the process. The F value is defined as the duration of heat treatment at a suitable reference temperature that would be equivalent, with respect to destruction of a certain number of spoilage microorganisms, to the actual process. The instrument computes the F values continuously throughout a process from temperature signals derived from a sensor probe inserted at the cold point of the packed food. The instrument may be readily adapted for use with various types of temperature detection, including fine-wire thermocouples, or temperature telemetry used in case of agitating cookers. The temperature sensor generates an electrical signal proportional to the temperature at the sensing point. The electrical signal is converted to an equivalent Fvalue by suitable mathematical translation. As the F value is displayed rather than time-temperature profiles, it is easier to control the process when the targeted Fo value is reached. The principle of operation is depicted in Fig. 1. Skinner (1975) has developed an instrument that calculates the F value of a heat sterilization process using an analog technique. Holdsworth (1983) has listed a few of the commercially available F,, integrators indicated in Table I. Jairus and Shoemaker (1985) have developed a transducer to directly measure lethality rates during thermal processing. It consisted of a double-legged thermocouple, a thermocouple conditioner, and an antilog amplifier. The lethality rates measured in real time during thermal process experiments compared well with those calculated later from time-temperature profiles. A generalpurpose microcomputer data acquisition system was designed to evaluate this transducer. This system can also be used for data recording of similar routine and experimental processes.

M. N. RAMESH et al.

318

DispdY opening Menu

-

Initialize Lethality (L) = 0.00 Processing Time (t) 0.00 Time = 0.00 ScVlIntervd t = 0.00

-

I

Time

=

Time + t

-

Display / Print

FO E L

Processing Time(t)

FIG.1. Flowchart for the principle of operation of an Fo integrator.

319

THERMAL PROCESSING OF FOODS, PART I1 TABLE I CflMMERCIALLYAVAILABLE DATAACQUISTI~N UNITS WITH THEIR SPECIFICATIONS

Unit

Ball datatrace

Ultrakust Thermophil (PD30)

Redpost

No. of channels

Measuring range PC)

Resolution ("C)

Accuracy PC)

10 to 140

0.2 full range

10-40 : 0.5 40-120 : 0.25 120-140 : 0.5

1

Wessex Power Technology 189 Ashley Rd, Parkstone Poole, Dorset, BH14 9DL

-100 to 300

0.1

0.2%

3

Endamon Unit 6, Highfield Close Cove, Hampshire GUlW

f0.3

1 or 2

The Old Pumping Station Tolt Rd, Bourn Cambridge CB3 7TT

k0.5

8

full range

-5 to 100

0.1 (49-80)

Multitracker

190 to400

0.1

Address

Deanland House Cowley Rd, Cambridge CB4 4GU

Continuous data acquisition units (DAUs) for thermal process evaluation (May and Withers, 1993) are available that can withstand the extreme environment of the process while logging temperatures within the container. These DAU systems are battery-operated and employ platinum resistance thermometers, thermistors, or thermocouples for temperature measurement. May and Withers (1993) have also listed some of the commercially available data acquisition systems for thermal processing process evaluation. These DAUs and their features are given in Table 11. This is an open-loop system, and it only displays the accomplished lethality or Fo value. Since the actual Fo value or the accomplished lethality at the can center is displayed, it is convenient to control the process and the temperature of the retort. This leads to optimum sterilization without overcooking. The limitation of this approach is that it needs exba instrumentation and must be maintained properly for accurate readings. It also only displays the Fo value calculated based on an equation that is derived

M. N. RAMESH

320

et

al.

TABLE I1 DATAACQUISITION UNITSAND Fo INTEGRATORS Use

and

Temperature

Accuracy

"C

"C

range

Sensor facilities

Automatic FO value computer Ellab A/S Denmark

Static & rotary batch retorts

4/6 in product 1in retort also rpm Cu-Constantan

10-150

f0.5

Lethality meter NCFT Reading University

Static retorts

1thermistor probe

90-135

-

Sterimeter (SRA 711) Telemetric Instrument, Sweden

Static retorts

Platinum resistance or Cu/Constantan thermocouple

85-135

f0.2

Ursamat Berlin Institute of Control Tech., Germany

Static retorts

Copper/ Constantan thermocouples

50-130

-

Cyclometer counter FO

Autodata Acurex California, USA

Static retorts

1 , 2 Or 4 Platinum resistance

20-127.5

f0.5

Digital and chart, Fo

Q-DOS minicomputer Q-DOS-Kings

Static

1OX Copper/ Constantan thermocouples

70-130

f 0.3

Computer screen and FO

Minicomputer software development CFPRA Glos., UK Lynn, UK

Static

12X Copper/ Constantan thermocouples

70-130

f0.4

Computer screen and floppy disk & heat frame data + FO

OBSLED Czechoslovakia

Static

Copper/ Constantan thermocouples

90-135

-

Name

Display

Digital or chart Fo and rpm

Digital Fo Digital or chart alarm preset Fo

Digital and chart and Fo

using uncertainties and assumptions. However, this approach does assure minimum sterility as accomplished by the derived time-temperature profile.

THERMAL PROCESSING OF FOODS, PART II

321

Ill. On-Line Monitoring Systems

Because of the important emphasis that must be placed on public safety when it comes to canned foods, processors must operate in strict compliance with the canned food regulations (Teixeira et al., 1982). Particular emphasis is placed on product batches that experience an unscheduled process deviation, such as a drop in retort temperature, during the course of the process. In such cases, the entire batch must be reprocessed or set aside for reevaluation by a competent authority. Such practices are costly, so that processors tend to operate with higher retort temperatures and longer process times than those specified in their established scheduled processes, so as to minimize the frequency at which process deviations occur. At the same time, processors also recognize that higher retort temperatures and longer durations tend to have an adverse effect on product quality, and they would like to minimize such overprocessing. With the advent of low-cost microcomputers, the measured Fo value can be improved to control the process. It is no surprise, then, that they are becoming common (Datta et a]., 1986). Such controls yield uniform product quality while minimizing energy waste due to overprocessing. This leads to optimum sterilization. The temperature sensed at the cold point of the packed food is used to calculate the F value of the process at the initial temperature of the can and the retort temperature of the system at any given time. A microprocessor, which forms the heart of the control system, monitors the steam inlet. Once the calculated Fo is equal to the targeted Fo value, the steam inlet is closed using suitable valves. Thus, the temperature of a particular system is measured on-line and any deviation is suitably taken care of during F value calculation. By this method, whatever the temperature of the system may be, once the cold point of the packed food reaches the Fo value, heating stops. An accuracy of +O.l% produces little additional savings in terms of either process time or overprocessing consequences. For the high-temperature process, the gains from reduced error arise more in terms of reduction in overprocessing than from significant savings of process time. This is a closed-loop control system, along with display of the accomplished lethality, at the same time controlling the process, when the targeted lethality is accomplished. It also corrects the deviation by suitably calculating the required Fo value based on the such process parameters as initial temperature and retort temperature. Thus, it is on-line monitoring of the Fo value. Teixeira et al. (1982) have used a numerical computer program model that was developed to simulate the thermal processing of

322

M. N. RAMESH et al.

conduction-heated canned foods. The discussed control logic applies to batch-operated retorts in which the process time can be easily adjusted to compensate for an uncontrolled change in retort temperature. Datta et al. (1986) have developed and demonstrated the performance of an on-line computer-based retort system for processing canned foods that can ensure the desired level of sterilization automatically in real time despite arbitrary deviations in the heating medium temperature. They also considered the cooling part of the F,, value for accomplished lethality to terminate the process. Navankasattusas and Lund (1978) have discussed the basics of the development of the time-temperature profile and systems for on-line measurement of lethality. This also includes a lethality rate generator and a lethality rate integrator. They also discussed monitoring and control of continuous thermal processing by on-line measurement of accomplished lethality with a miniaturized lethality meter, without a display unit, in an insulated container. The accomplished lethality is read after the detector has traveled through the entire path of the cooker. Teixeira (1995) reviewed some intelligent on-line control systems for retorts that have been developed in Europe and North America. Various approaches based on real-time accumulation of the accomplished sterilization value of lethality are described. The lethality value is calculated from data taken from test containers using on-line data acquisition systems. The on-line control systems have been classified as: 0

Intelligent control via real-time data acquisition.

0

Control by estimated correction factor. Intelligent control with heat transfer models.

Intelligent retort control systems have the ability to calculate the accomplished sterilized value of a process (Fo)in response to the integrated time-temperature history experienced by the product at its cold point. This accumulating sterilization is compared with the target value to be reached at the end of heating as a control decision criterion to determine the end of heating for the process. These intelligent on-line retort control systems can be very effective and reliable and have been put into commercial practice. However, they require instrumented test containers with probes and lead wires into each and every retort batch. This approach is quite practical where the product value is very high, process times are very long, and considerable downtime is available between batches. However, it is considered cost-prohibitive, so that much attention is being focused on development of indirect means for

THERMAL PROCESSING OF FOODS, PART I1

323

accomplishing intelligent on-line retort control, either by using estimated correction factors or mathematical heat transfer models. Control by the estimated correction factor approach assumes that the lowest temperature of the retort during a deviation will remain sustained for the duration of the process and sets a new extended process time calculated at the reduced temperature. This approach requires that process times for a range of specified retort temperatures be calculated and stored in the memory of the control system, for quick reference whenever needed. The principal disadvantage of this approach is that, when a low-temperature extended process schedule has been adopted, it usually leads to overprocessing, since most process deviations would quickly recover, allowing the retort to resume its normal operating temperature. A revised method for calculating the correction factor avoids this unnecessary overprocessing by extending the process time only to the desired extent to compensate for the deviation. The correction factor is calculated statistically. Although this approach minimizes overprocessing, it becomes unwieldy in situations where multiple deviations occur and when deviations will not fully recover or might be difficult to analyze. This necessitates an accurate knowledge of the product center temperature history in response to any dynamic retort temperature variation without measuring internal can temperature with probes of any kind. Heat transfer models fill the void in this area and provide accurate process schedules for specific processing equipment and products. A heat transfer model for simulating thermal sterilization of canned food is a mathematical equation capable of predicting the internal temperature over time in response to a change in temperature applied at the surface. The solution of the equation is obtained by a finite-difference model employing numerical techniques. The actual retort temperature is read directly from sensors and is continuously updated with each iteration of the numerical solution. The model predicts the correct internal temperature response for any unexpected deviation of retort temperature. By programming the control logic, heating is continued until target lethality is accomplished, facilitating attainment of the desired level of sterilization regardless of any unscheduled process deviations. This approach shows the greatest promise for use with truly intelligent on-line retort control systems. The advantage of this system is that it controls the process and ensures optimum sterilization even with deviations from the set parameters. The limitation of the approach is that it only accepts the set parameters. Actual time-temperature profiles to be set are to be derived by using any of the approaches discussed earlier.

324

M. N. RAMESH et al.

IV. Semiautomatic Retort Control Systems for Optimum Sterilization

Ramesh et al. (1994) have developed a semiautomatic retort control system for sterilization of foods using a microprocessor. The control system has two modes: direct retort temperature monitoring and container cold-point monitoring. The importance of this control system is that the actual container center temperature is used to control the retort temperature and the entire process. This facilitates optimum cooking, leading to greater nutrient retention and savings in energy by avoiding overcooking. The cooling part of the lethality is also accounted for by suitably modifying the targeted Fo value to be achieved during the heating period. This system can be easily retrofitted onto existing horizontal and vertical manually controlled batch retorts. The system is economical compared to the more sophisticated large computer-controlled systems. It is mainly aimed at small- and medium-scale food processing industries that have manually operated batch retorts and do not necessarily involve huge investments. A block diagram of this system developed is shown in Fig. 2. Gill et al. (1985) have retrofitted two retorts with computerized data acquisition and control systems. Process control was carried out on the basis of up to 10 test cans equipped with thermocouples. The system provided excellent on-line feed-forward control for conductively heated products, with anticipatory correction for cool-down lethality. A separate algorithm was used for convectively heated products. The on-line monitoring models that have been developed require prior knowledge of thermal difhsivity for the heating and cooling phases of the process (Teixeira et al., 1982; Datta et al., 1986). Variations in the proportions of ingredients of heterogeneous mixed food from one processing batch to another can cause significant deviations in the thermal properties of foods, which may lead to underor oversterilization. Ryniecki and Jayas (1993) have developed a method for automatic determination of model parameters for computer control of the sterilization of heterogeneous mixed food with an unknown thermal diffusivity value. This can be done prior to processing by sterilization of some of the test cans. This model was employed for prediction of the cumulative lethality contributed by the cooling portion of the sterilization process. This lethality was used in heating turnoff decisions. The model is fast and accurate for this purpose. It was validated in the laboratory using 5-9% bentonite suspensions. The final monitored cumulative lethality was within 18% of the desired cumulative lethality for these

I I I I I I I I I I I I I I I I I I I I I I I I I

. 1 I

. 1 1

. 1 1

I

l

l

1 1 1 .I/

1 1 1 ‘I/

1 1 1 .I/

I

l

l

P 9

SOLENOID VALVES

AM?LIFIBR

I’

1 I I

’ /

KEY

BOARD

DslluT

/

FIG.2. Block diagram of the retort control system.

I

326

M. N. RAMESH et al.

suspensions. Lappo and Povey (1986) have developed a computer-based control system whose features include a modular design and the use of a high-level software language. The system is capable of following a prescribed time-temperature profile by altering its profile according to a computed Fo value. The Fo value is computed from any of 16 withincontainer temperature measurements or can be derived from retort temperatures. Control of pressure is also available during the cooling process in order to reduce stress on the container. Experimental data are presented to characterize the performance of the retort and instrumentation. The importance of instrumentation accuracy is discussed, and a computer calculation of heat penetration into a can is used to investigate the significance of errors in temperature screening with regard to process times. Completely automatic control of the retort heating profile is possible along with pressure control during cooling, so that on-line prediction and control of the Fo value is feasible. Also, a significant reduction in process time can be achieved by improving prediction of retort temperature measurement. Mihori et d.(19911 have developed an on-line control system that would achieve correct heat sterilization for conduction heating of food. The system collects a series of time-temperature data via a sensing probe during the early stages of the heating phase and analyzes the collected data for parameters of the conduction heating phase. During cooling, it integrates the lethality rate and determines the appropriate time to begin cooling to achieve a desired process lethality. This process has been validated by experiments. The advantage of this system is that it needs very little operator intervention and the process can be controlled independent of process deviations. Its limitations include the cost factor and maintenance requirements. The parameters to be controlled are to be predetermined. V. Computer-Aided Sterilization

M/s Steritech, a French company (Anon., 19951, in collaboration with the Agence Nationale de Valorisation de la Recherche, has developed a computer-controlled autoclave designed to conserve foods without much nutrient loss. Called Steritech, this system is equipped with a DOS software program that controls each phase of the sterilization process. The programming is flexible, and the process is interactive (to provide enhanced security). The computer and the autoclave maintain complete control over the sterilization process, which can be adapted to customer needs. As the temperature rises in the autoclave vessel, the quantity of steam injected is monitored and adjusted according to a

THERMAL PROCESSING OF FOODS, PART I1

327

number of criteria, including total product weight. The steam pressure inside the retort is also controlled. The computer continuously monitors and adjusts the opening of the proportional valve. At the start of the cooling phase, a powerful pump recycles the condensate at the bottom of the tank, resulting in savings in water usage. This patented system preheats the pressurized air to the temperature inside the autoclave before injecting it, which prevents thermal shock and ensures that the air will not expand. The calculated sterilization value (Fo)is integrated into the software, and sterilization automatically stops when the preset value is attained. Remote maintenance is performed by satellite by means of a modem equipped with software to access autoclaves anywhere in the world in real time. VI. Process Indicators

The determination of safe thermal process schedules is based on the time-temperature history at the cold point in packaged food. For still processing conditions, temperature monitoring can be easily established as it is static. However, for agitated forced convection-heated and particulate foods, identification of the cold point is rather difficult due to the practical difficulties involved in monitoring their temperature dynamic profiles. These difficulties have led investigators to assume still processing condition parameters with safety factors so as to compensate for uncertainty in particulate behavior (Pflug, 1987). Ronner (1990) has developed a bioindicator for monitoring and control of the sterilization process. Sastry et al. (1988) have developed bioindicators for verification and evaluation of thermal processing of particulate foods. Tobback et al. (1992) have described the application of immobilized enzymes as a time-temperature indicator system in thermal processing. Thus, the time-temperature indicator can be microbiological, enzyme-based, chemical, or physical systems and can be directly related to a change in any intrinsic food quality attribute. Much work has been targeted on the development of biological time-temperature indicators (TTIs) using immobilized enzymes. De Cordt et al. (1992) have developed one such TTI based on immobilized a-amylase on glass beads. This TTI can be employed in the range of 98-108°C and is only suitable for acid products below pH 4.6. The advantage of this device is that wireless measurements can be made. This is more vital for in-pack processing and for continuous aseptic processing of liquid foods. The feasibility of the use of TTIs with different carrier material has been discussed in detail by Maesmans et al. (1994a,b,c).

328

M. N. W S H et a].

In addition, temperature measurement of moving particles is a serious problem, and aseptic processing of particulate foods relies on microbiological/biological validation (Dignen et a]., 1989). Problems associated with the use of microorganisms as bioindicators have been detailed in several studies (Weng et al., 1991a,b; Pflug and Odlaug, 1978; Berry et a]., 1989; Sastry eta]., 1988). The methodology, though, is subject to some inherent disadvantages: 1. Any TTIs used will have to be carefully calibrated with particular reference to the z and D values of the TTI with a known standard

for accurate measurement. 2. The TTIs have to be located in a carrier material. Such thermal and physical properties as conductivity, heat capacity, and density need to be precisely known. 3. These devices are not applicable to all temperature ranges. For example, biological TTIs will be more sensitive to more lethal temperatures than to low temperatures. A distinction can be made between two directions toward which evaluation of the impact of food preservation processes is moving (Hendrickx et al., 1995). 1. With the entry of food commodities prepared by combined processes

using different unit operations and various steps of food processing (like mixing, soaking, and drying) onto the market, evaluation procedures have to be developed to monitor food quality properties for which temperature is not the only rate-determining factor. A complete understanding and description of all the factors that influence the kinetics is necessary for development of “product history integrators” as suitable monitoring systems that give relevant information on the complete history, safety, and quality of the product. 2. For such “new” heating techniques as scraped surface heat exchangers and ohmic heating, the existing approaches are inadequate for particulate foods. Because of the technological difficulties involved in developing small-sized wireless data transmitters, rapid development of new and accurate TTIs is required to prevent modern heating techniques from eluding their marketplace. In general, process indicators quantify the integrated time-temperature impact on foods. Depending on the monitoring mechanism, process indicators can be subdivided into: 1) bioindicators, 2) color-based physical indicators, 3) chemical markers, and 4) electronic indicators.

THERMAL PROCESSING OF FOODS, PART II

329

A. BIOINDICATORS

Food as such is nonhomogeneous, and the heating behavior of food is complex. It is practically impossible to determine the time-temperature curve and to identify the cold point in moving particulated foods, either due to agitation in the rotating retorts or in aseptic processing (Pflug and Odlaug, 1978). In such cases it is difficult to design, control, and verify the proper sterilization process. A bioindicator provides the required information and could be used as a tool to more accurately evaluate the impact of thermal processing on food (Weng et d.,1991a). Hendrickx et d. (1995) subdivided biological TTI systems into two categories: 1. Survivor-kill systems integrate the time-temperature history and indicate if a preset processing value is obtained or underprocessing has occurred depending on whether the microorganism systems are killed or have registered growth. By this method, if surviving organisms are present no conclusion can be drawn about the extent of heat treatment meted out to the product. 2. The count reduction approach, wherein the processing value can be

determined by the number of survivors. An inoculated pack is a practical example of the use of a biological indicator unit. The initial concentration of the microorganism spores (lo8) contained in the food is raised to a detectable level of lo-" after heat treatment. In this way, an endpoint defined as a probable number can also be detected by or even of surviving units (PNSUs) of multiplying the initial spore load (No) of the food product (under normal conditions) by spore reduction (= NIN,) in the inoculated pack (Pflug, 1987). This is based on the assumption that increasing the concentration of spores will not alter their thermostability. An inoculated pack study is classified as a TTI approach and not as an in situ approach, because calibrated spores with known kinetic properties have to be used. They are not identical to the actual bioburden of the food: they are used to mimic inactivation of the bioburden. Whereas the inoculated pack was at first used to verify process calculations and heat-penetration measurements for still retorts, since the 1950s it has been applied to design and monitor heat sterilization processes in machines where it is almost impossible to determine the Fo value by physical-mathematical methods (Pflug, 1988). Different tools have been designed to control the location of microorganisms or their spores at a specified site in the food product instead of

330

M. N. RAMESH et ol.

dispersing them over the entire contents of the can. Spore solutions have been enclosed in carrier units that largely simplify the recovery procedure after processing. A plastic rod filled with a calibrated B. stearothermophilus spore solution was suggested by Pflug (1976) that has been successfully used to validate sterilization processes for green beans and whole-kernel corn in a still retort and peas in brine in a Steritort. Sterilizing values determined from the physical-mathematical approach have been compared with their biological indicator units (BIUs) counterparts. These studies revealed that plastic rod BIUs can be used effectively to determine the sterilizing value given to cans processed in agitating retorts. According to Pflug et al. (1980), this method allows a routine determination of Fo with 15% accuracy. An aluminum biological indicator carrier has been proposed, as this material has the improved heat transfer characteristics and mechanical strength that is of particular importance in thermal processing of rapidly heating lowviscosity foods (Rodriguez and Teixeira, 1988; Smith et al., 1976, 1982). Paper strips inoculated with known amounts of spores are also employed for contamination-free test packaging. The problem with these biological thermocouples is that complete recovery of all heated organisms is not possible; in addition, the measurement is influenced by pH, the oxidation-reduction potential, the nutrient contents, etc. (Pflug et al., 1990). A bioindicator is encapsulated in a small vial and is placed at the cold point of the food package. This sensor is used to evaluate the lethality at that point during thermal processing. One of the bioindicators was commercially developed by the Swedish biotechnology company Diffchamb AB in Goteborg. This is an aqueous gel sphere in which specific microorganisms like the spores of B. stearothermophilus have been encapsulated. The sphere, which can be autoclaved, is surfacesterile and spore-tight. It possesses enough strength to allow passage through various types of process equipment. Mixed with food products, it can be treated during the sterilization or pasteurization process and later separated aseptically and cultivated in a nutrient solution. Should the heat treatment process prove to be insufficient, surviving microorganisms would bring about coloring of the gel sphere within 24 hours (Anon., 1990). Pflug et al. (1980) have carried out a series of experiments to evaluate the performance of plastic rod BIUs, which were used to measure the sterilization process to cans of food processed in a Steritort. The results indicated that BIUs can be used effectively to measure the Fo value

THERMAL PROCESSING OF FOODS, PART I1

331

delivered to containers of food heated in continuous agitating sterilization machines at an accuracy level that is essential on a par with that of time-temperature data. With the advent of aseptic processing, in the case of viscous liquid foods containing particulates, carrier systems for microbiological TTIs have been miniaturized to measure sterilization values under these conditions. Many small microbiological systems for determining the efficacy of in-pack sterilization have been studied. An overview of these systems is given in Table I11 (see Dignen et d., 1989). Various bioindicators used in different systems to evaluate thermal processing are indicated in Table IV. One advantage of using a microbiological TTI for monitoring the safety of a food product is that the temperature ranges in which the measuring device and target are sensitive to heat are equal, and in some cases the target and TTI may be identical. However, only calibrated microorganisms are to be used, and the heat inactivation kinetics of the microorganism have to be checked for the specific microorganism used in the TTIs. Also, using a microorganism to monitor the destruction of a quality attribute with specific kinetics (e.g., z = 10.C) is not sufficient, and not even tolerable. Without proper calibration in terms of D and z values in the actual heating, no conclusion can be drawn about the impact of heat treatment. Like any TTI, the biological TTI must have to comply with the basic requirement that z T T ~ = ztarget (Philipp and Sucker, 1990). Another advantage of the microbiological method is that heat penetration and thermal death time studies can be carried out simultaneously, thus eliminating some of the uncertainties involved in assembling the data obtained independently by these two types of studies. This method accounts for the effect of heat on the indicating microorganism at the scaled-up and as-applied states. This method is cost-effective and rapid. The results are easy to interpret and can be applied in both batch and continuous processes, that is, canning and aseptic processing (Anon., 1990). This can also serve as a complement to the instrument control system used for monitoring of sterilization of packaged containers. Only calibrated spores can be used to determine the killing power of the heat treatment. A biological indicator as such can never be a calibration standard from the methodology point of view. It will have to be validated first against a known physical standard (Pflug and Odlaug, 1986; Bruch, 1973, 1974).

332

M. N. W S H et al. TABLE 111 BIOINDICATORS EMPLOYING MICROSIZED MICROORGANISM-BASED TTIs ~

~

Carrier

Shape

Polymethylmethacrylate Alginate Glass Pea puree, meat puree Alginate Alginate + peach Polyacrylamide gel

Sphere Bead Bulb Cube Sphere Cube Sphere

Alginate + mushroom Alginate + potato puree Turkey Alginate + peach

Mushroom Cube Cube

~

~

Size (cm) 0.3 0.160.40

0.5 0.8-2.4

0.3 1.25

0.127

0.3-0.5

z value

Microorganism Bacillus anthracis B. stearothermophilus B. stearothermophilus B. stearothermophilus B. stearothermophilus B. stearothermophilus B. stearothermophilus B. subtilis E. polymyxa B. stearothennophilus Clostridium sporogenes C. sporogenes Z. bacilli

(“C) 60

8.5 10 11.4-1 1.8 9

N R ~ NR NR NR NR 12.5-12.7

8.5 NR

WR = not reported.

TABLE IV BIOINDICATORS USEDIN DIFFERENT SYSTEMS TO EVALUATE THERMAL PROCESSING OF FOODS Low-acids foods

B. stearothermophilus B. subtilis 5230 B. coagulans

Wet heat sterilization

C. sporogenes

Acid food pasteurization

Zygosaccharomyces bacilli

(60-65°C)

Pasteurization

Immobilized peroxidase (H,O,)

(95-100’C)

Continuous UHT processing

C. butyricum

B. COLOR-BASED PHYSICAL INDICATORS A color-based physical indicator is placed directly into the center of a can of food to determine heat penetration during processing. The indicator is held in a clear nylon pouch in a wire holder and placed in the test container before sealing. The nylon pouch permits penetration of the steam and yet provides protection for the indicator. The indicator

THERMAL PROCESSING OF FOODS, PART I1

333

is read after removal from the can. These indicators contain a chemically impregnated purple band that changes to green only if the process conditions are met. COOK-CHEX process indicators developed by Aseptic Thermo Co. are available for 14 different time-temperature conditions for all can sizes (Anon., 1972). It is a small and inexpensive device that indicates a time-temperature-dependent irreversible change that can be easily measured. This device mimics the change of a welldefined target quantity parameter of food undergoing the same variable temperature exposure (Taoukis and Labuza, 1989). Witonsky (1977) proposed a TTI that functions on a physical basis. A dry chemical in an embossed well of an aluminum plate is placed at the end of a paper wick. It is covered with a transparent film of known steam permeability. Steam permeating the film depresses the melting point of the colored chemical, and the molten chemical wicks up in the paper film, the distance of the color front from the base being a function of increasing temperature (Swartzel et al., 1991). The device could be calibrated in terms of Fo. The disadvantage of this device is that, as it is steam-activated, it cannot be used to monitor other types of heating media. Physical TTI systems are described as easy and accurate to prepare and calibrate, user-friendly in readout, and easy to recover. Further research should reveal physical phenomena with kinetic characteristics in the range of or equal to quality attribute inactivation in order to make such a system an interesting process evaluation tool. A retort temperature check card called TEMPILAQ has been developed by Williams (1969). It consists of a thick, colored bulb of known melting point. It is allowed to dry and is overlaid with thick porous paper (e.g.,Whatman no. 2 filter paper). As the temperature of the retort is raised, the color of the bulb is simultaneously developed. This only happens when the retort temperature reaches the set temperature. To prevent TEMPILAQ from soaking into the backing card, an aluminum foil interlayer can be used. TEMPILAQ lacquers with a wide range of melting points are available.

C. CHEMICAL MARKERS Selection of process parameters as well as the food particulate that will ensure commercial sterility at the cold point of the moving particulate is a challenging problem (Kim, 1994). The key processing parameters in a continuous thermal processing system are the temperature, flow rate, holding tube length, power setting, and conductivity of foods. A few alternatives like bioindicators and TTIs are available, as discussed earlier. However, instead of using external sources, it is better to use intrinsically formed compounds within the foods to demonstrate

334

M. N. RAMESH et al.

the sterility of thermally processed low-acid foods. Quantification of a microbiological TTI requires skills, and the analytical precision of the technique is rather low. The likelihood of contamination during microbiological sterility testing of foods has been recognized and documented (De Cordt et al., 1992). The inherent limitations of these methods for determining the efficacy of sterilization has prompted investigation of alternative approaches. These alternatives include chemical and physical TTIs and enzyme systems. For example, a covalently immobilized horseradish peroxidase together with an organic solvent was embedded in the food system and pasteurization efficiency aimed at destroying D-streptococci could be monitored in the temperature domain 65-90°C. These enzyme systems offer relatively easy readout and handling, which is a significant advantage over microbiological TTIs. The temperature domain wherein inactivation of enzyme occurs limits the monitoring of a sterilization process using these enzyme systems. Many methodologies have been attempted to assess the effect of heat treatment post facturn by monitoring the changing profile of innate biochemical markers. A commercial enzyme-monitoring system (Apizym, Analytlab Products) has been used to detect changes in the enzyme profile of meats before and after thermal processing (Townsend and Blankenship, 1987a,b; Brown, 1991). The enzyme fingerprint enables one to determine whether a specified temperature has been reached in the product. The result is dependent on the initial levels of enzyme, and the reproducibility of the results above 70°C is poor. Upper and lower limits of the temperature range are set by the biochemical markers, and conclusions on the degree of microbial destruction reached inside the product are hard to draw from these methodologies. Chemical TTIs detect a change in concentration of a chemical compound added to the food product to measure the efficacy of thermal processing. Thiamin mixed in beef puree and pea puree and added to peas-in-brine shows a reduction in its concentration after heat treatment. Even the thermal hydrolysis kinetics of disaccharides have been used for thermal evaluation. However, Wen Chin (1977), showed that the agreement was reasonable for processes with a negligible come-up time that were processed at a reference temperature of 121.l0C,whereas agreement was poor when the process had a considerable come-up time and thermal gradients throughout the system. A paper disk impregnated with reducing sugar and amino acid (Maillard's reaction) was employed by Favetto et al. (1988, 1989) to correlate the inactivation of foot-andmouth virus by monitoring the darkening of the disk. A disadvantage with such systems is the need to attach it to the packaging surface,

THERMAL PROCESSING OF FOODS, PART II

335

which makes cold-point evaluation of the product impossible. There is an effort underway to predict impact on food safety by monitoring the response of the innate chemical constituents of food products (Kim and Taub, 1993). Thus, chemical TTIs offer high hopes as promising tools for evaluation of thermal process. The only, yet crucial, deficiency is that no reactions have been identified in the literature on heat treatments of foods that feature the temperature dependency (the z value or activation energy) required to monitor food safety in the sterilization temperature range, and only a few are available that can be used to follow the deterioration of other quality attributes. A conceptual ground for validating thermal processing using intrinsic chemical markers (ICMs) was defined, and three potentially useful markers have been selected (Kim et al., 1992; Kim and Taub, 1993).

The Methodology of Chemical Markers The food sample flows through the heat exchanger into the holding tube. Measurement of marker concentration (Ml, M2, or M3, based on the food constituent) can be made at the end of the holding tube. Since carbohydrates are commonly present in most foods, the change in carbohydrate profile is monitored using anion exclusion chromatography (AEC) separation and photodiode array (PDA) detection (which is sensitive in the UV region and has scanning capability). The UV absorption spectra of compounds eluted from the chromatographic column are obtained every 6 sec with the PDA detector, stored in the computer, and then manipulated for display as a three-dimensional (3D) representation, a contour map, a spectrum at a specific retention time, or a chromatogram at a specific wavelength. The AEC-PDA system consists of a Wescan (Derf 111) anion exclusion column (sulfonated polystyrene/divinylbenzene, 7.8 x 100 mm) and a Waters (Milford Man) 990 PDA detector. The heated food sample is homogenized for 1 min with a tenfold excess of water using polytron. The extract is centrifuged, filtered through a 0.45-pm membrane filter, and injected into the chromatography system through a-ZO/yl injection loop. The eluent used is a 0.01 Nsulfuric acid solution with a 1-ml/min flow rate (Kim and Taub, 1993).

The thermal reactions leading to formation of these markers are intrinsic to the food, that is, the new compounds are formed without adding any compounds prior to heating. The markers are easily detectable because their absorption maxima occur when the unheated foods show no absorption.

336

M. N. RAMESH et al.

Two of the markers, M1 and M3, were identified as 2,3-dihydro-3,5dihydroxy-6-methyl-4(H)-pyran-4-one (MW = 144) and 5-hydroxymethyl-furfural (MW = 126). It has also been shown that M2 has a similar molecular weight and is produced from a water-soluble component and protein. M1 is formed upon heating almost any food, including fruits and vegetables, meats, and starch-based foods. M2 is formed almost exclusively in meats, while M3 is formed in fruits and vegetables (Kim, 1994). Relating the marker yield to lethality is a nontrivial problem. The relationship is straightforward under isothermal conditions as long as the rate constants and the activation energies of microbial destruction and marker formation are accurately known. However, as the center temperature of a food particulate increases in the holding tube by heat conduction, a changing time-temperature condition has to be considered. A piece of meat is a typical particulate in an aseptically processed low-acid food. Formation of two markers (M1 and M2) has been verified in all the meats tested (Kim, 1994). Therefore, the two-marker approach appears promising for validating aseptically processed meat particulates. Kim (1994) has described a two-marker approach that uses the ratio of the yields of two markers and its potential application in ohmic heating, and microwave sterilization as well as conventional aseptic processing. Several chemical indices have been developed to optimize the heating process. Mulley et al. (1975) used thiamin hydrochloride. Textural change in meat was used as a heating index by Tennigen and Olstad (1979). Berry et al. (1989) studied the destruction kinetics of methyl methionine sulfonium (MMS) in buffer solutions and found it to be suitable for indexing microbial lethality. They recommended MMS as a substitute for microorganisms in thermal process evaluation. The activation energy associated with microorganisms is higher than that for nutrients and enzymes. Hence, for high-temperature short-time (HTST) processing it is better to use enzymes as indicators rather than microbial spores and nutrients. For vegetable processing, peroxidase inactivation is often used as an indicator because of its reported heat stability (Schwartz, 1992). Along similar lines, trypsin-a family of enzymes that preferentially catalyzes hydrolysis of ester and peptide bonds-has been used as a chemical marker. Awuah et al. (1993) have evaluated the possibility of using trypsin as a bioindicator for aseptic processing of high- and low-acid foods in the range of 90 to 130°C in tris-HC1 and citrate buffer. The decimal reduction equivalent heating times (EHTs) at 1 3 O O C of trypsin was 30.7 min at pH 6.0, 98.3 min at pH 5.1, and 135 min at pH 3.8, compared to 2.24 min for B. stearothermophilus. These EHT values

THERMAL PROCESSING OF FOODS, PART I1

337

are at the reference temperature of 121.l0C, which should be differentiated from decimal reduction at a particular condition. This gives a more realistic comparison of the inactivation behavior of various enzymes at different temperatures because the process times are based on Fo values calculated using a z value of 1O"C, whereas enzyme inactivation is characterized by its z and Do values. At lower temperatures, B. stearothermophilus has a higher thermal resistance of 11.4 min of EHT as compared to 0.4-2.4 min of EHT for trypsin. This trend reverses as processing temperature exceeds 12OoC, as is indicated in Fig. 3. Hence, at the high temperatures employed for aseptic processing, enzymes serve as better indicators for verification of the process (Awuah et al., 1993). The effect of temperature and pH on EHT value is shown in Table V. As the means of chemical detection have improved, considerable interest has arisen concerning chemical markers, that is, substances are formed naturally in foods at known rates during processing. It is hoped that measurement of such compounds can generate useful information about the time-temperature history of the food in a container during processing. Chemical markers function similarly to time-temperature indicators (Wells and Singh, 1988; Hendrickx et al., 1992) and are possibly of value in application to continuous retort systems or heat exchangers, holding tubes, and suspended particulate behavior in ultrahigh-temperature (UHT) processing (Sapru et al., 1992; David and Merson, 1990). There is a need to demonstrate the kind of information that can be obtained from chemical-marker measurements and to define the properties that such markers ought to have. For example, if the properties of a putative marker are known, it would be possible to decide whether measuring its yield would give adequate and reliable information about microbial destruction or inactivation. Ross (1993) worked on the relationship of bacterial destruction to chemical marker formation during thermal processing. Ramaswamy et al. (1995) used a chemical marker to measure holding tube lethality. Meat beads containing a glucose processor of a chemical marker and alginate beads containing the spores of B. stearothermophilus were prepared and subjected to steam heating at 110°C for selected time intervals. Marker yields were related to spore survival data and lethality values obtained from timetemperature data to generate calibration curves, as indicated in Table VI. Meatballs fabricated with meat (marker processor) and alginate (microbial spores) beads placed at the center were heated in continuously flowing CMC solutions (0.5% w/w) at 110 + 0.5"C in an aseptic

W W

co 1so

0

100

A

3

c

+

SO

x 0

0

2 n

L-

8

100

1t o

110

-m

1so

I40

(c)

FIG.3. Equivalent heating time (EHT) for a decimal reduction in activity of trypsin and selected bioindicators at various temperatures.

339

THERMAL PROCESSING OF FOODS, PART I1 TABLE V TIME-TEMPERATURE INACTIVATION OF ENZYME [TRYPSIN) IN DILUTE HCl AND CITRATE BUFFER Temperature (“Cl

D value (min)

3.8

90 100 110 120 130

279.7 144.0 68.9 31.5 18.4

2.45 2.16 1.84 1.50 1.26

5.1

90 100 110 120 130

132.0 72.8 50.8 22.8 11.7

2.12 1.86 1.71 1.36 1.07

6.0

90 100 110 120 130

33.3 27.9 11.8 8.5 3.4

1.52 1.45 1.07 0.93 0.53

PH

log D (min)

processing holding tube simulator. All treated samples were analyzed for marker yield as well as spore survival. As indicated in Table VI, the lethalities and spore count reductions in terms of Fvalues calculated from marker yield data showed excellent correlations with experimental values. These results indicate that the chemical markers have a potential to provide data on accumulated lethality and spore count reductions in aseptic processing systems where direct temperature determination at the center of a particle is difficult. Hence, with proper corrections, marker yield data could provide valuable input on produced process lethality and microbial spore count reduction. D. ELECTRONIC INDICATORS It is not practical to include thermocouples located in containers in a production system, and much research has gone into evaluating the alternatives for determination of temperature history during processing (Holdsworth, 1983). As a result, possible systems based on electronic memory or radiotelemetry have been extensively studied.

340

M. N. RAMESH et a]. TABLE VI CALIBRATION DATAFOR MARKER (Ml) AND MICROBIAL SPORE KILL

Heating time (rnin)

FOa

M1

(min)

(aW

log N

2

0

0

5.76

0

0

10

0.45

0.004

5.58

0.18

0.14

20

1.23

0.19

4.96

0.80

0.63

log(No/N)

FOC (rnin)

30

2.0

0.042

3.98

1.78

1.40

40

2.78

0.068

2.32

3.44

2.71

50

3.56

0.10

1.0

4.76

3.76

‘Fo calculated using equation Fo = ji 10(T-nefl’zdt. h e a n of two measurements. “Fo calculated using DO(log(N0lN)).

A thermal memory cell (Swartzel et al., 1991) has been patented based on diffusion of ions in the insulator layer of a metal-insulator-semiconductor capacitor. The diffusion distance can be accurately read out by measuring the capacitance change of the cell before and after heat treatment. These authors claim that, by doping the insulator layer with at least two different mobile charge carriers of different activation energy and combining their readouts, conclusions can be drawn on the effect of the process on any food property, irrespective of their match with the kinetic characteristics (the E,, value) of ions used in the TTI. This is in contrast to the theoretical basis for proper TTI functioning. Advances in electrical circuit miniaturization have permitted the development of time-temperature memory devices for assessing thermal processing of packaged foods (Navankasattusas and Lund, 1978). Such a memory device consists of a temperature register module within an insulated container. This module is connected to a temperature transducer implanted inside a model food package. The module and model food package travel together through the continuous processing retorts, and the temperature history inside the model food package is registered at a constant preset time interval. As soon as the combined unit emerges from the retort, the module is connected to a temperature/accomplished lethality readout unit. Monitoring and control of a continuous processing unit by on-line measurement of accomplished lethality can also be made using a miniaturized lethality meter, without the display unit, in

THERMAL PROCESSING OF FOODS, PART I1

341

an insulated container. Wiring is eliminated by sending the lethality detector within an insulated container through the retort together with the model food package. The accomplished lethality is read out after the detector has traveled the entire path of the agitating retort, as with the system using the memory cell (Navankasattusas and Lund, 1978). The disadvantage of electronic memory is that it cannot be used for control, only for monitoring, whereas the radio telemetry-based system can be used for both. Sterility/lethality indicators that display the temperature in terms of lethality rates are available. They provide the advantage of direct display of lethality rates, thus eliminating transcription errors, but they have the disadvantage of not revealing the nature of heating or cooling. If the heating rate is uniform, this factor can be determined. However, for some types of products it is necessary to have more information because the slope of heating curve changes due to heating; that is, broken heating curves implying a change of heat transfer made from conduction to convection, or vice versa. A history of the time-temperature profile is very essential in calculating lethality rates. In addition, if time-temperature data are available, they can be used for calculation of process times for different-sized cans. Further, they provide the advantage of calculating process conditions at other temperatures (Holdsworth, 1983). VII. Suggestions for Future Work

Though on-line monitoring of sterility addresses the current needs of thermal processing to overcome process deviations, it does result in loss of products in the form of test cans. This poses a serious problem when highly valuable products are being processed, which can inhibit application of the approach. On-line control systems based on heat-transfer models provide solutions to overcome this limitation. However, these models require an accurate thermal diffusivity value for the food product. A better approach would be to combine nondestructive temperature sensing using heat-flux sensors mounted on the outside surface of the container with on-line determination of the thermal properties of the food during the initial stages of processing and application of the heat-transfer model. Though a few reports are available (e.g., Mihori et al., 1994; Watanabe et al., 1994), much needs to be done in this area. Although several researchers have pursued the legacy of the development of biological and chemical TTIs, a universal TTI has yet to have been developed. Calibrated microbiological TTIs have evoked much interest. With the advances in enzymatic and physical TTIs, renewed attention is being given to the possibilities and restrictions of these

342

M. N. RAh4ESH et al.

systems as TTIs in thermal processing of foods. There is need for better models to select a specific TTI for a food being processed, employing newer heating techniques like ohmic/microwave heating and scraped surface heat exchangers, and also for food with particulates, where existing approaches are inadequate. There is a need for the development of additional kinetic data and models, as determination of product temperature alone is not sufficient for ascertaining product quality. Such other factors as the physicalhhermal properties and the dynamic nature of the product should be considered in developing models for selecting TTIs. Though chemical markers offer an alternative for assessing the integrated time-temperature exposure of the food particulate, pertinent literary data are scarce and need to be developed. Since the intrinsic chemical-marker approach is safe, efforts should be made to commercialize the method in aseptic food processing. VIII. Conclusions

The production of sterilized foods requires that a product undergo a thermal process so as to render it microbiologically safe. Current practices in food industries for establishing process schedules require extensive heat-penetration experiments along with microbiological studies. Such a schedule is specific for each product-container-sterilizer combination. It is clear from the above discussions that on-line control systems facilitate optimum processing of foods and so ensure commercial sterility along with higher product quality and higher nutritional value as compared to conventional methods. On-line monitoring also allows processing times to be automatically adjusted to account for process deviations. On-line sterility monitoring ensures commercial sterility, which facilitates immediate release of processed products to consumers. On-line electronic control systems that employ thermocouples suffer the serious disadvantage of not being adaptable to aseptic processing due to the dynamic nature of the foods. Biological TTIs have the advantage of being able to be employed in HTST processing of foods. The applicability of TTIs for on-line monitoring of sterility in HTST processing indicates that the whole area of chemical indicators promises to be a fertile ground for researchers and patent hunters. The practical application of chemical TTIs is just the beginning. The use of a chemical index in sterilization processing has the potential of effecting a revolutionary change in the food and pharmaceutical industries (Mulley et al., 1975). A chemical marker-based TTI monitoring system can be encapsulated so that its kinetic charac-

THERMAL PROCESSING OF FOODS, PART I1

343

teristics can be determined independent of food composition, heating mode, or heating technology (Hendrickx et al., 1995). The concept of using intrinsic compounds developed during processing as internal time-temperature integrators is efficient and practical. These internally generated markers can be detected by spectrophotometry after their separation using anion-exclusion chromatography. Comparison of the relative marker concentration provides information on heat penetration within particulates. Also, the formation of these markers, which is associated with destruction of the bacterial population, facilitates computation of the sterility value of particulates subjected to thermal processing. Markers thus used could be applied to validate thermal processing of foods, particularly in aseptic food processing. Thus, these intrinsic markers are ideally suited to HTST profiles associated with conventional aseptic processing, ohmic-heated processing, and microwave-heated processing. When employed in food processing, intrinsic markers will assure the effectiveness of the process, so that the FDA should approve the process schedules developed using such TTIs. ACKNOWLEDGMENTS

The authors thank V. Prakash, Director of the Central Food Technological Research Institute, and A. Ramesh, N. G. Karanth, and R. Venkatakuppaiah, for their encouragement, and M. Asha, for her help in preparation of the manuscript. We also gratefully acknowledge the copyright permission granted by the different publishers for permitting us to utilize published material. REFERENCES Anon. (1972). Food Technol. 26, 18. Anon. (1990). Invention Intelligence 25, 155. Anon. (1995). Food Eng. Int. 20, 34. A m a h , G. B., Ramaswamy, H. S. Simpson, B. K., and Smith, J. P. (1993). 1.Food Processing Eng. 16,315. Berry, M. F., Singh, R. K., and Nelson, P. E. (1989). 1.Food Processing Preserv. 13,475. Brown, H. M. (1991). Tech. Memo No. 625, Campden Food and Drink Research Association, Chipping, UK. Bruch, C. W. (1973). Annu. 1.Pharm. Sci. NS2(1), 1. Bruch, C. W. (1974). Bull. Parenter. Drug Assoc. 28(3), 105. Datta, A. K., Teixeira, A. A., and Manson, J. E. (1986). J Food Sci. 51(2), 480. David, R. D., and Merson, R. L. (1990). 1.Food Sci. 55, 488. De Cordt, S. V., Van Hoof, K., Hu, J., Measmans, G., Hendrickx, M., and Tobback, P. (1992). Int. J. Food Sci. Technol. 27, 661.

344

M. N. RAMESH et 01.

Dignen, D. M., Berry, M. R., Pflug, I. J., and Gardine, T. D. (1989). Food Techno].43(3), 118.

Favetto, G. J., Chiriffe, J., Scorza, 0. C., and Hermida, C. A. (1988). J. Food Prot. 51(7), 542.

Favetto, G. J., Chiriffe, J., Scorza, 0. C., and Hermida, C. A. (1989). U.S. Pat. 4,834,017. Gill, T. A., Thompson, J. W., and Leblanc, B. (1985). Proc. Int. Congr. Eng. Food, 4th, Edmonton, Canada, p. 547. Hendrickx, M., Weng, Z., Maesmans, G., and Tobback, P. (1992). Int. J. Food Sci. Techno]. 27(1), 21.

Hendrickx, M., Maesmans, G., De Cordt, S., Noronaha, J., Vanloey, A., and Tobback, P. (1995). Crit. Rev. Food Sci. Nutr. 35(3), 231. Holdsworth, S. D. (1983). Process Biochem. 16(5), 24. Jairus, R. D., and Shoemaker, C. F. (1985). J. Food Sci. 50, 223. Kim, H. J. (1994). Activities of RbD Associates 46(1), 28. Kim, H. J., and Taub, I. A. (1993). Food Techno].47(1), 91. Kim, H. J., Taub, I. A., Richardson, M., Kustin, K., and Ross, E. (1992). Activities ofRbD Associates 44(1), 120. Lappo, B. P., and Povey, M. J. W. (1986). J. Food Eng. 5, 31. Maesmans, G. J., Hendrickx, M. E., De Cordt, S. V., and Tobback, P. (1994a). Food Res. Int. 27, 39. Maesmans, G. J., Hendrickx, M. E., De Cordt, S. V., Vanloey, A., Noronha, J., and Tobback, l? (1994b). Food Res. Int. 27, 413. Maesmans, G. J., Hendrickx, M. E., De Cordt, S. V., Vanloey, A., Noronha, J., and Tobback, P. ( 1 9 9 4 ~ )Food . Control 5(4), 249. May, N., and Withers, l? (1993). Food Techno].Int. Europe (1993 Annual), 97. Mihori, T., Watanbe, H., and Kaneko, S. (1991). J. Food Processing Preserv. 15, 135. Mihori, T., Xier, L., and Watanabe, H. (1994). Proc. ACoFOP Symp., 3rd, Paris. Mulley, A., Stumbo, C., and Hunting, W. (1975). J. Food Sci. 40, 993. Navankasattusas, S., and Lund, D. B. (1978). Food Techno].32(3), 79. Pflug, I. J. (1976). U.S. Pat. 3,960,670. Pflug, I. J. (1987). J. FoodProt. 50(4), 342. Pflug, I. J. (1988). Environmental Sterilization Laboratories, Minneapolis, MN. Pflug, I. J., and Odlaug, T. E. (1978). Food Techno].32(6), 63. Pflug, I. J., and Odlaug, T. E. (1986). J. Porenter. Sci. Technol. 40(5), 242. Pflug, I. J., Jones, A. T., and Blanchett, R. (1980). J. Food Sci. 45(4), 940. Pflug, I. J., Berry, M. R., and Dignen, B. (1990). J. Food Prot. 53(4), 312. Philipp, B., and Sucker, H. (1990). Phorm. Res. 7(12), 1273. Rarnaswamy, H. S., A m a h , G. B., Kim, H. J., and Choi, Y. M. (1995). Activities ofR@D Associates 47(3), 216. Ramesh, M. N., Kartik, V., Navneeth, L. V., and Bhanuprakash, K. N. (1994). Proc. Int. Con$ TEPEM 94,Madras, India. Rodriguez, A. C., and Teixeira, A. A. (1980). Trans. ASAE 31(4), 1233. Ronner, U. (1990). Food Techno].Int. Europe 90, 43. Ross, E. W. (1993). J. Food Processing Eng. 16, 247. Ryniecki, A., and Jayas, D. S. (1993). J. Food Eng. 19, 75. Sapru, V., Teixeira, A. A., Smerage, G. H., and Lindsay, J. A. (1992). J. Food Sci. 57, 1248. Sastry, S. K., Li, S. F., Patel, P., Konanayakam, M., Bafha, P., Doores, S., and Beelanan, R. B. (1988). J. Food Sci. 53(5), 1528. Schwartz, S. J. (1992). In “Advances in Aseptic Processing Technologies” (R. K. Singh and P. E. Nelson, eds.), p. 63. Elsevier Applied Science Publishers, New York.

THERMAL PROCESSING OF FOODS, PART I1

345

Skinner, R. (1975). Food Manuf. 50(10), 43. Smith, G., Pflug, I. J., and Chapman, P. (1976). Appl. Env. Microbiol. 32, 257. Smith, G., Kopelman, M., Jones, A., and Pflug, I. J. (1982). Appl. Env. Microbiol. 44, 12. Swartzel, K. R., Ganesan, S. G., Kuehn, R. T., Hamker, R. W., and Sadghi, F. (1991). U S . Pat. 5,021,981. Taoukis, P. S., and Labuza, T. P. (1989). J. Food Sci. 54(4), 783. Tennigen, A., and Olstad, S. (1979). In “Food Process Engineering,” Vol. 1: “Food Process Systems” (C. P. Linko, Y. Malkki, J. Olkku, and J. Larinkari, eds.), p. 146. Elsevier Applied Science Publications, London. Teixeira, A. A. (1995). Activities of R 6 D Associates 47(1), 205. Teixeira, A. A., and Manson, I. E. (1982). Food Technol. 36(4), 85. Tobback, P., Hendrickx, M. E., Weng, Z. M., Maesmans, G. J., and De Cordt, S. V. (1992). In “Advances in Food Engineering” (R. P. Singh and M. A. Wirakartakusumah, eds.), p. 561. CRC Press, Boca Raton, FL. Townsend, W. E., and Blankenship, L. C. (1987a). J. Food Sci. 52(2), 511. Townsend, W. E., and Blankenship, L. C. (1987b). 1.Food Sci. 52(62), 1445. Watanabe, H., Xier, L., and Mihori, T. (1994). Proc. ACoFOP Symp., 3rd, Paris, p. 375. Wells, J. H., and Singh, R. P. (1988). J. Food Sci. 53, 1866. Wen Chin, L. (1977). Ph.D. Dissertation, University of Massachusetts, Amherst. Weng, Z. M., Hendrickx, M., Measmans, G., and Tobback, P. (1991a). J. Food Sci.56, 567. Weng, Z. M., Hendrickx, M., Measmans, G., Gebruers, K., and Tobback, P. (1991b). 1.Food Sci. 56(2), 574. Witonsky, R. J. (1977). Bull. Parenter. Drug Assoc. 11(6), 274. Williams, M. L. B. (1969). Con. Inst. Food Technol. J. 2(4), 188.

This Page Intentionally Left Blank

INDEX

A

Absidia spp., 142 ACAT, see Acyl-CoA:cholesterol acyltransferase Acetylcholinesterase, 10 ACW, see Acetylcholinesterase Acremanium strictum, 20-21 Acyl-CoA:cholesterol acyltransferase cholesterol inhibition, 13-17 Adamantanes, 142 Additives, 253-254 AFEX, see Ammonia fiber explosion Affinity chromatography, 240 Albophoma yamanashiensis, 14-15 Alcohols, 90, 109 Aldose reductase, 11 l-O-Alkyl-2-acetyl-sn-glycero-3phosphocholine, 51-55 a-Alkyl-P-hydroxyesters, 84-85 Alternaria alternata, 61 Alzheimer’s disease, 10 Amidepsines A-D, 20 (2S,2R,4S)-2-Amino-l-cyclohexyl-6methylheptane-3,4-diol, 174 Ammonia, 4 Ammonia fiber explosion, 265-266, 277-278 Andrastins, 49 Apiocrea chrysosperma, 30-32 Arisugacin, 10 Arohynapene D, 58 Aromatic compounds, 139-140 Arthrobotrys oligospora, 57 Aschersonia spp., 66 Ascobolus furfuraceus, 4 0 4 1 Aspergillus f l a w s , 63 Aspergillus fumigatus, 15 Aspergillus nidulans, 27 Aspergillus niger, 22, 150-151 Aspergillus spp., 222-224, 268 Aspergillus tubingensis, 64 Aureobasidium pullulans, 269 Australifungin, 35-36 Azaphilones, 17, 22-24

B

B16 cells, 17 Bacillus cereus, 138 Bacillus spp. carbohydrate catabolite repression, 189-190

cell membrane, 190-191 o-glucose utilization, 189 inoculum, 193-195 inosine-producing, creation, 184-187 pleiotropic properties, 188-191 o-ribose fermentation with, 191-203 Bacillus subtilis, 20 Baeyer-Villiger oxidations, 131-133, 155-156 Baker’s yeast aroma compounds, 114-116 biocatalysis, reproducibility, 116-11 7 characteristics, 81-82 fatty acid synthetase complex, 84-90 heterocyclic compounds, carbonyl groups in, 90-93 history, 83 hydrolytic activities, 110-1 13 lyase activity, 113-114 nitrogen-reducible functional groups, 93-97 oxidations, 106-110 sulfur-containing compounds, 97-100 Beauvaria bassiana, 55 Beauveria sulfurescens, 143-145 Beer chillproofing, 252 Benzoylformate decarboxylase, 104 (S)-Benzylthioglycerate, 98 Betula alba, 114 BFD, see Benzoylformate decarboxylase BHK21 cells, 41 Bicyclo[3.1.Olhexanes, 173 Biogeneration, 114-116 Biomass composition, 263-264 347

348

INDEX

Biomass, ethanol conversion cellulose, 267-271 hemicellulose, 2 71-2 73 lignin, 273-274 pretreatment, 2 65-26 7 technological constraints, 279-282 Biosynthesis ammonia control, 4 melanin, 10 tannase, 237-239 vs. hydroxylation, 149 Biotransformation models, 128-133 oxidative, 154 oxygenase in, 125-127 Bisdechlorogeodin, 63-64 Blood coagulation, 55 C

Caenorhabdifis elegans, 63 Caesalpinia coriaria, 224 Cancer, 5-11 Candida albicans, 27, 42 Candida pelliculosa, 183-184 Candida spp., 29-30,222 Carbohydrate catabolite, 189-190 Carbonyl reduction, 129-131 Carpophilus hemipterus, 63 Cassia spp., 224 Catabolite repression, 238 Catechins, 220-221 CD4 receptors, 45-46 Cell adhesion, 55-56 Cellulose, 271 Cephalosporins, 43 Cercophora areolata, 39-40 Cercophorins, 39-41 c-fos gene, 49 Chlorogenic acid, 219-220 Chloromycorrhizin, 58, 60 Cholesterol biosynthesis, 13-17 Chrysospermins, 30-32 Chrysosporium loba f urn, 4 9-5 0 Clostridium bofulinum, thermal processing canned foods, 294 canned products, 296 pH factors, 298-301 value of D, 289-292

Closfridium spp, 87 Common viral diseases, 43 Continuous data acquisition units, 319 Corynebacterium spp, 222 Curvularia f a k a f a , 129-1 30 Cyanoacetone, 105 Cyclohexanone monooxygenase, 155-156 Cyclosporines, 42-43 Cytochrome P450,147-148

D Darlucins, 41 DAUs, see Continuous data acquisition units Dechloromycorrhizin, 58, 60 Dehydrogenases, 84 Desaturation, 156-157 Destruxins, 66 Diacylglycerol acyltransferase, 19-20 2,3-Dideoxynucleosides, 174 Diepoxins, 34-35 Dihydrobisdechlorogeodin, 63-64 Dihydroepiepoformin, 24 Dilute acid, 265 2,2-Dimethylcyclohexanone, 141 Dioxygenase, 145 DNA, recombinant, 188 o-Ribose applications, 172-1 77 characteristics, 171-172 detection, 171-1721 72 identification, 171-172172

E

Echinocandins, 27-29 Eimeria fenella, 58 Ellagitannins, 218 Emindole, 15 Enantiopure 3-hydroxyesters, 84 Endo-l,4-P-glucanase, 267-268 Endoglucanases, 267-268 Endothelins, 22-25 Enoate reductase, 88 Epimerization, 179-180 Epolactaene, 56-57 Epoxide hydrolysis, 131

INDEX

Escherichia coli ethanol conversion, 276-277 D-ribose conversion, 182-183, 191 Ester hydrolysis, 128-129 Ethanol biomass conversion to, 263-273 composition, 263-264 enzymatic, 267-274 pretreatment, 265-267 technological constraints, 2 79-282 dilute, recovery, 282 fermentation, 274-279 production cost, 261-262 Etoposide products, 47-57 Eupenicillium spp., 64 Exo-l,4-P-glucanase, 267-268

F Fatty acid synthetase complex, 84-90 Favolaschia pustulosa, 37 Favolon, 37 FDA, see Food and Drug Administration Feedback inhibition, 238-239 Fermentation hydrolyzates, 277-279 D-ribose, 191-203 D-XYlOSe, 274-275 tannase, 228-237 Fibroblasts, L929, 32 Flavalins, 218 Fleephilone, 45 Food and Drug Administration aseptic processing regulations, 305 canned food regulations, 293-294, 302-303 Fungal products agriculturally active, 57-66 antiinfective agents, 25-47 antitumor activity, 47-57 cholesterol biosynthesis, 13-20 enzyme inhibitors, 5-11, 13 growth, 3-5 history, 2-3 lipid metabolism inhibitors, 13-20 media, 4-5 neuropharmacological properties, 20-25 nutrition, 3-5 pharmaceutical activity, 51-57 potential, 3

349

Fusacandins, 29-30 Fusarielins, 37 Fusarium sambucinum, 29-30 Fusarium spp., 37 Fusidienol, 50 Fusidium griseum, 50

G Gallic acid, 253-254 Gallotannins, 218 Gilmaniella h umicola, 63 Gliocladium spp., 49 Gliotoxins, 49 P-Glucosidase, 267-268 Glutamyl-D-ribose, 170 Green tea, 249-252

H Haemonchus contortus, 57 Harziphilone, 45 Heat transfer model, 323 HeLa cells, 48 HeLaS3 cells, 41 Helioferins, 32-34 Helminthosporium spp., 151-1 54 Hemicellulose, 271-273 Heparinase, 8 HIV virus, 43-44, 175 HL-60 cells, 41, 48, 56 Humicola spp., 20, 53 Humulene, 2 1 Hydantocidin, 175 Hydrolyzates, 277-279 2-a-Hydroxydimeninol, 51 Hydroxyesters, 8 4 4 5 Hydroxylase, 145-147 Hydroxylation active-site models, 133-149 prognosis, 158-159 sulfoxidation model, 154-155 vs. biosynthesis, 149 vs. phylogeny, 148-149 6-Hydroxymellein, 60-61 Hydroxystrobilurin A, 3 7-38

INDEX

350 I

Immune system, 42-43,175 Intelligent retort control systems, 322-323 Isaria sinclairii, 43 Ischemia, myocardial, 173 Isochromophilones, 46-47 ISP-1.43

K P-Ketoesters, 86 Kuraosins, 49

L

L1210 cells, 32, 41 Lachnumol A, 58, 60 Lachnumons, 5 8 , 6 0 4 1 Lachnum papyraceum, 58-59 Lignins, 273-274 Lipid metabolism, 13-17 Lovastatin, 13 LSD, see Lysergic acid diethylamide Lyase, 113-114 Lysergic acid diethylamide, 20

M

MacArdle disease, 173 Macrolpiota spp., 32 Macrosphelides, 56 MAD disease, 173 Melanin, 10, 35 9-Methoxystrobilurins, 37 Methyl alkanols, 93 Methyl inosine monophosphate, 175 Methyl thioethers, 99 5-Methylthioribose, 170 Microtubles, 37 Mortierella isabellina, 140, 151 MR304A cells, 11 Mutarotation, 171 Mycelia sterilia, 35, 43 Myceliophthora lutea, 48 Mycestericins, 43 Mycogone rosea, 32

Myoadenylate deaminase deficiency, 173 Myocardial ischemia, 173

N National Canner’s Association, 295 Nattrassia mangiferae, 47 Nematodes, 57-58 Neuropeptide Y, 22 Nitrobenzene, 93 o-Nitrobenzonitrile, 94 P-Nitronitriles, 94 NK-374200, 66

0

Oudemansins, 37 Oxidation, Baeyer-Villiger, 131-133 Oxidoreductases, 84 Oxidosqualene, 101 a-Oxoacids, 104 Oxygenase, 125-128

P Paecilomyces carneus, 7 Paecilomyces spp., 49 Paecilomyces variotii, 62 Paeciloquinones, 7-8 PAF, see 1-0-Alkyl-2-acetyl-sn-glycero3-phosphocholine Paspaline, 15 Patulodin, 39 PDC, see Pyruvate decarboxylase Penicillins, 43 Penicillium brevicompactum, 183 Penicillium chzysogenum, 43 Penicillium claviforme, 22 Penicillium multicolor, 46-47 Penicillium patulurn, 2 4 Penicillium rubrum, 53-54 Penicillium sclerotiorum , 2 3 Penicillium spp. AChE inhibitor from, 10 antiinsectan compounds from, 64 epolactarne from, 56-57 tannase from, 222-224

351

INDEX

Penicillium urticae, 39 Pestalotiopis spp., 24 PET operon, 276,280-281 PG, see Propylene glycol (R)-Phenylglycine, 96 Phoma destructiva, 30-32 Phoma spp., 25 Phosphatases, 111 Phosphate esters, 110-113 Phosphatidylcholine, 20 Phosphatidylethanolamine, 20 Phosphoenolpyruvate, 189 Phospholipase Az, 9 5-Phosphoribosyl-l-pyrophosphate, 170 Phosphorylase, 173 Phosphotranferase, 189 Phylogeny, 148-149 Phytotxins, 61 Pin us maritima, 22 2-224 PKC, see Protein kinase C PLAz, see Phospholipase Az Platelet aggregation, 55 Pneumocystis carinii, 27 Polygonium caspiatum, 6-7 Poria COCOS, 9 Preussia isomera, 39 Process indicators chemical markers, 333-339 color-based physical, 332-333 electronic, 339-341 time-temperature, 32 7-332 Propylene glycol, 108 Protein kinase C, 5-9 Protein tyrosine kinases, 5-9 Pseudomonos putida, 141 Pseudomonas reptilivora, 183 PTK, see Protein tyrosine kinases Pyricularia oryzae, 39 Pyripropenes, 15 Pymvate decarboxylase, 102-103, 104

Q Quercus pedunculata, 222-224

R Raji cells, 20

Ras gene, 49 Recombinant DNA technology, 188 RES-1214-1, 24 Retort control systems, 324-326 Rhizopus nigricans, 128 Riboflavin, 172-173 o-Ribose fermentation, 191-203 composition, 195-198 conditions, 198-201 recovery, 210-203 natural occurrence, 168-171 production chemical, 178-180 microbial, 180-188 microorgamisms, 183-1 84 nonmicrobial, 176-180 recombinant DNA technology, 188 synthesis, 168 o-Ribose-5-phosphate, 170, 180-182 Ribosylation, 173

S

Saccharomyces cerevisiae ethanol conversion, 275-276, 281 oxidation catalysts, 99-100 Salmonella typhimurium, 183,191 Sch-50673, 47 Sch-50676,47 Sch-57404, 42 D-Sedoheptulose, 174 Sordaria fimicola, 40-41 Sphaerellopsis filum, 41 Sporobolomyces salmonicolor, 32 Sporomiella australis, 35-36 Sporormiella intermedia, 25 Squalestatins, 25-27 SSF, 269-270 Stachybocins, 22 Stachybotrydial, 17, 19 Stachybotrys microspora, 55 Stachybohys spp., 17,19-20, 22 Staphylococcus aureus, 32 Sterilization, see Thermal processing Steroid hydroxylation, 133-136 Stream explosion treatment, 265-266 Streptomyces griseus, 139 Streptomyces hygroscopicus, 175

352

INDEX

Sulfoxidase, 157-1 58 Sulfoxidation active-site model, 150-154 reaction model, 154-157 Sulfur heterocycles, 92 Sulfuric acid treatments, 265-267 Supercritical carbon dioxide explosion, 2 66-2 67

T Talaromyces trachyspermus, 8-9 Tannase applications, 250-255 biosynthesis, 237-239 fermentation liquid-surface, 233-234 solid-state, 2 34-2 37 submerged, 228-233 history, 216-218 immobilization, 249-250 isozymes, 246 location, 239 mode of action, 246-249 prognosis, 2 54-2 55 properties, 241-246 purification, 239-241 significance, 225-226 sources, 222-224 Tannin acyl hydrolase, see Tannase Tannins, 2 18-22 1 Taxol, 47, 51 Taxomyces andreanae, 51 Taxus brevifolia, 51 Teniposide, 47-57 Terminalia chebula, 224 Terpendoles, 1 5 Terpene hydroxylation, 1 3 6 1 3 9 Thermal processing aseptic system, 305-308 evaluation, 2 9 6 297 on-line methods computer-aided sterilization, 326-327 Fo integrators, 317, 319-320 monitoring systems, 321-323 overview, 316 process indicators, 327-341 prognosis, 341-343 retort control systems, 324-326

overview, 288-289 pH, 297-301 prognosis, 310-311 safety factors, 303-305 statistical analysis, 308-310 temperature, 301-303 uncertainties, 289 canned products, 294-296 value of D, 289-292 value of F, 292-294 value of Z, 292 Thiazolyl ketone, 91 Thioacetaldehyde, 97 Thioglycerate, 97-98 Thrombin, 55 a-Tocopherol, 89-90 Trachyspic acid, 8 Transketoloase-deficientmicroorganisms goal-oriented, 182-183 industrially applied, 184-187 pleiotropic properties, 188-191 Trichoderma harzianum antifungal peptaibols, 32 antivirals from, 45 MR304A from, 10-11 Trichoderma spp., 268 Trichorzins, 32 Trimegestone, 83 Triprenylphenol cholesterol esterase, 19-20 Tryprostatins, 47 Trypsin, 55

V Verticillin, 49 Verticillium balanoides, 6 Vitamin Bz, 172-173

W

Wine making, 252 X

Xenovulene, 20-21 X y h S , 271-273 D-Xylose, 274-275

INDEX

Y Yeast, see also specific types ethanol conversion, 275 RNA hydroxysis, 177-178

353 2

Zaragozic acids, 25-27 Z-vinyl iodide, 175 Zymomonas mobilis, 275-276

This Page Intentionally Left Blank

CONTENTS OF P R E V I O U S V O L U M E S

Volume 34

What’s in a Name?-Microbial Secondary Metabolism J. W Bennett and Ronald Bentley

Disclosure Requirements for Biological Materials in Patent Law Shung-Chang Jong and Jeannette M. Birmingham INDEX

Microbial Production of Gibberellins: State of the Art l? K . R. Kumar and B. K. Lansane Microbial Dehydrogenations of Monosaccharides MiloS Kulhdnek Antitumor and Antiviral Substances from Fungi Shung-Chang Jong and Richard Donovick Biotechnology-The V S. Malik

Golden Age

INDEX

Volume 35

Production of Bacterial Thermostable a-Amylase by Solid-state Fermentation: A Potential Tool for Achieving Economy in Enzyme Production and Starch Hydrolysis B. K. Lonsane and M. K Ramesh Methods for Studying Bacterial Gene Transfer in Soil by Conjugation and Transduction G. Stotzb, Monica A . Devanas, and Lawrence R. Zeph

Volume 36

Microbial Transformations of Herbicides and Pesticides Douglas J. Cork and James I! Krueger An Environmental Assessment of Biotechnological Processes M. S. Thakur, M. J. Kennedx and N . G. Karanth Fate of Recombinant Escherichia coli K-12 Strains in the Environment Gregg Bogosian and James E Kane Microbial Cytochromes P-450 and Xenobiotic Metabolism E Sima Sariaslani Foodborne Yeasts 7:Decik High-Resolution Electrophoretic Purification and Structural Microanalysis of Peptides and Proteins Erik I! Lillehoj and Vedpal S. Malik INDEX

Microbial Levan Youn W Han

Volume 37

Review and Evaluation of the Effects of Xenobiotic Chemicals on Microorganisms in Soil R. J. Hicks, G. Stotzkx and I? Van Voris

Microbial Degradation of the Nitroaromatic Compounds Frank K. Higson

355

356

CONTENTS OF PREVIOUS VOLUMES

An Evaluation of Bacterial Standards and Disinfection Practices Used for the Assessment and Treatment of Stormwater Marie L. O'Shea and Richard Field Haloperoxidases: Their Properties and Their Use in Organic Synthesis M. C. R. Franssen and H. C. van der Plas Medicinal Benefits of the Mushroom Ganoderma S. C. Jong and J. M. Birmingham Microbial Degradation of Biphenyl and Its Derivatives Frank K. Higson The Sensitivities of Biocatalysts to Hydrodynamic Shear Stress A l e s Prokop and Rakesh I(. Bajpai Biopotentialities of the Basidiomacromycetes Somasundoram Rajamthnam, Mysore Nanjarajurs Shashirekha, and Zakia Ban0

The New Antibody Technologies Erik I! Lillehoj and Vedpal S. Malik Anoxygenic Phototrophic Bacteria: Physiology and Advances in Hydrogen Production Technology K. Sasikala, Ch. V Ramana, r! Rahuveer Rao, and K. L. Kovacs INDEX

Volume 39

Asepsis in Bioreactors M. C. Sharma and A. K. Gurtu Lipids of n-Alkane-Utilizing Microorganisms and Their Application Potential Samir S. Radwan and Naser A. Sorkhoh Microbial Pentose Utilization Prashant Mishra and Ajay Singh

INDEX

Medicinal and Therapeutic Value of the Shiitake Mushroom S. C. Jong and J. M. Birmingham

Volume 38

Yeast Lipid Biotechnology Z . facob

Selected Methods for the Detection and Assessment of Ecological Effects Resulting from the Release of Genetically Engineered Microorganisms to the Terrestrial Environment G. S t o t z b , M. W Broder, J. D. Doyle, and R. A. Jones Biochemical Engineering Aspects of Solid-state Fermentation M. V Ramana Murthx N . G. Karanth, and K. S. M . S. Raghava Rao

Pectin, Pectinase, and Protopectinase: Production, Properties, and Applications Takuo Sakai, Tatsuji Sakamoto, Johan Hallaert, and Erick J. Vandamme Physicochemical and Biological Treatments for Enzymatic/Microbial Conversion of Lignocellulosic Biomass Purnendu Ghosh and Ajay Singh INDEX

CONTENTS OF PREVIOUS VOLUMES Volume 40

Microbial Cellulases: Protein Architecture, Molecular Properties, and Biosynthesis Ajay Singh and Kiyoshi Hayashi Factors Inhibiting and Stimulating Bacterial Growth in Milk: An Historical Perspective D. K. O’Toole Challenges in Commercial Biotechnology. Part I. Product, Process, and Market Discovery AleS Prokop Challenges in Commercial Biotechnology. Part 11. Product, Process, and Market Development AleS Prokop Effects of Genetically Engineered Microorganisms on Microbial Populations and Processes in Natural Habitats Jack D. Doyle, Guenther S t o t z b , Gwendolyn McCJung, and Charles W Hendricks Detection, Isolation, and Stability of Megaplasmic-Encoded Chloroaromatic Herbicide-Degrading Genes within Pseudomonas Species Douglas J. Cork and Amjad Khalil INDE:X

357

Improving Productivity of Heterologous Proteins in Recombinant Saccharomyces cerevisiae Fermentations Amit Vasavada Manipulations of Catabolic Genes for the Degradation and Detoxification of Xenobiotics Rup Lal, Sukanya Lal, l? S. Dhanaraj, and D. M. Saxena Aqueous Two-Phase Extraction for Downstream Processing of EnzymeslProteins K. S. M.S. Raghava Rao, N. K. Rastogi, M. K. Gowthaman, and N. G. Karanth Biotechnological Potentials of Anoxygenic Phototrophic Bacteria. Part I. Production of Single Cell Protein, Vitamins, Ubiquinones, Hormones, and Enzymes and Use in Waste Treatment Ch. Sasikala and Ch. V. Ramana Biotechnological Potentials of Anoxygenic Phototrophic Bacteria. Part 11. Biopolyesters, Biopesticide, Biofuel, and Biofertilizer Ch. Sasikala and Ch. V. Ramana INDEX

Volume 42

The Insecticidal Proteins of Bacillus th uringiensis l? Ananda Kumar, R. l? Sharma, and V S. Malik

Volume 41

Microbial Oxidation of Unsaturated Fatty Acids Ching 1: Hou

Microbiological Production of Lactic Acid John H. Litchfield Biodegradable Polyesters Ch. Sasikala

358

CONTENTS OF PREVIOUS VOLUMES

The Utility of Strains of Morphological Group I1 Bacillus Samuel Singer Phytase Rudy J. Wodzinski and A. H. J. Ullah INDEX

Volume 43

Production of Acetic Acid by Clostridium thermoaceticum Munir Cheryan, Sarad Parekh, Minish Shah, and Kusuma Witjitra Contact Lenses, Disinfectants, and Acanthamoeba Keratitis Donald G. Ahearn and Manal M. Gabriel

Marine Microorganisms as a Source of New Natural Products V S. Bernan, M. Greenstein, and N M. Maiese Stereoselective Biotransformations in Synthesis of Some Pharmaceutical Intermediates Ramesh N. Patel Microbial Xylanolytic Enzyme System: Properties and Applications Pratima Bajpai Oleaginous Microorganisms: An Assessment of the Potential Jacek Leman INDEX

This Page Intentionally Left Blank

ISBN 0-12-002644-9