Trendsin Bioteehnology, VoL 2, No. 6, 1984
always hold true. In particular, it is now questionable whether thermophilic fermentations are indeed less prone to contamination. A summary of the expectations and realities of thermophilic fermentations will be given later in this review. In principle, the concept ofa thermophilic ethanol fermentation is a very simple one. That is - to perform a fermentation at high temperature (at Mark A. Payton least within the range of 60-70°C), Doubts about the viability of using classical yeast fermentations to reducing the need for power-consumproduce ethanol have led in recent years to a search for alternative ing cooling of reactor vessels and facilmicrobes capable of ethanol production. Among the front-runners for itating the constant removal of the consideration are the thermophilic bacteria and this article shows how volatile product, ethanol, by in situ close these extraordinary microbes have come to replacing yeast in evaporation/distillation. This could maintain a low ambient ethanol conethanol production. centration in the fermenter, circumIn recent years, the increased realiza- another mesophile which ranks high in venting the need for an organism tion of the finite nature of the world's ethanol-production terms, namely the resistant to high ethanol concentraoil supplies and vagaries in oil prices genus, Zymomona:. Though it is tions. In theory, therefore, one could have re-kindled interest in the pro- probably fair to say that there is less envisage a continuous system where the duction by fermentation of bulk chemi- known today about the process optim- substrate, be it cellulose, digests of cals, in particular potable and industrial ization of Zymomonas ethanologenic cellulose wastes, or sugar-cane juice, is ethanol. Such fermentation processes fermentations, there is a rapidly fed continuously into a fermenter conwould use as substrates renewable bio- expanding literature on the subject 3. taining an actively fermenting cukure mass sources such as cellulose, hemi- Also the recent development of suitable of a thermophile, while ethanol is concellulose and starch, and this has host-vector systems, facilitating tinuously evaporating out of the vessel focused attention on thermophilic broadening of the currently restrictive into a suitable condensation apparatus ethanologens ~ such as Clostridium substrate range of this organism 3, may allowing the constant collection of thermocellum (for cellulose utilization), make Zymomonas a close competitor to ethanol fractions. CI. thermohydrosulfuricum (for utilizing the thermophilic ethanol producers Which thermophile? cellulose degradation products) and which we shall consider here 4. Thermoanaerobium brockii (for Many thermophiles are well docufermentation of starch). The purpose of Why thermophiles? mented in the literature and a large Attention was focused on the use of number of these produce ethanolL In this review is to outline some of the organisms which display potential for thermophiles for ethanol production theory, at least, a thermophilic organuse in a thermophilic ethanologenic for the following reasons: ism suitable for ethanol production fermentation and in particular to could be any one of the types of organ1. Thermophiles have higher growth describe some of our work using ism in Table 1. and metabolic rates than mesophiles. the thermophilic bacillus, Bacillus I am aware of a number of attempts to 2. Thermophiles are themselves very stearothermophilus. approach the problem by option 3 in robust and contain very stable When considering ethanol producTable 1 which have not yet been fruitenzymes. tion by thermophilic fermentations, ful, and indeed one can appreciate that 3. Thermophiles generally have a low one must remember that there are this is a difficult task since it is likely cellular growth yield, hence more submesophilic organisms capable of that many enzymes would need to be strate carbon appears as ethanol. excellent ethanol production and that altered to achieve true thermophilyL It 4. Thermophilic fermentations may thermophiles must compare favourably remains then to compare the ethanol be less prone to sterility problems. production capacities of naturally with these. Probably the current 5. Growth at high temperature facilfavourite for large-scale production of occurring thermophiles. itates the removal of volatile products, ethanol is the yeast, Saccharomyces, In addition to the obvious requirenotably ethanol. since the fermentative capacity of this ments for the optimal ethanologenic organism has been honed to near perIt must be noted, however, that some thermophile, such as rapid fermentafection by extensive research and of these suggested advantages do not tion rates, strain stability and high development both on the host organism and on process optimization as a result Table I. Ways of obtaining the ideal thermophilic ethanologen of its historical importance in the brew- Option 1 Searchinglikelyhabitats for naturally occurringsuper ethanologenic ing industry. Before passing onto thermophiles. thermophiles I should also mention Option 2 Optimizingthe ethanol productionof a known thermophile using a combination of strain improvementby mutagenesis and process development. Adapting a mesophilicorganismwith excellent ethanol production MarkA. Payton is at Biogen S. A., 3, route Option 3 characteristics (e.g.S. cerevisiaeor Zymomonas)to thermophily. de Troinex, 1227 Geneva, Switzerland.
Production of ethanol by thermophilic bacteria
© 1984~ Elsevier Science Publishers B.V., Amsterdam 0166
2"rends in Biotechnology, VoL 2, No. 6, 1984
Table 2. A comparison of five thermophiles with potential for usein industrial ethanol production Organism
Wide substrate range especially Relativelylow ethanol cellulosics. production ( 1-1.2 tool ethanol (tool glucose)-1). Clostridium High ethanol yield (>1.5 moles Cannot use cellulosedirectly. thermohydrosulfuricum ethanol (tool glucose)-1). High temperature optimum (~70°C). Thermoanaerobium brockii High temperatureoptimum Cannot use cellulose directly. (65-70°C). Moderate ethanol production. Thermoanaerobacterethanolicus High ethanol yield. High Cannot use cellulose directly. temperature optimum (~70°C). Bacillus stearothermophilus Moderate ethanol yield. High Limited substrate range. metabolic rate. Clostridium thermocellum
temperature tolerance, one must also consider the substrate spectrum of the various thermophiles. This may, however, be of decreasing significance since several cellulose-degrading enzymes have already been cloned 7, and intergeneric cloning may, for prokaryotes at least, offer a method of broadening substrate utilization. The five thermophilic ethanologens I shall consider here are listed in Table 2. The possibility of combining more than one organism in a fermentation e.g. using the excellent cellulolytic capabilities of one organism and the high ethanol-production rate and capacity of a second organism has not been overlooked and will be considered later. To understand the advantages, disadvantages and potential for improvement of these five organisms better we should first consider the various pathways of ethanologenesis displayed by different microorganisms.
Microbial pathways of ethanologenesis Microorganisms possess three major types of pathway for the production of ethanol as a fermentative end product s. Two of these types of pathway differ in the way that pyruvate, the natural product of the glycolytic pathways the organisms possess, is subsequently metabolized. The third, relatively uncommon, pathway is found in organisms which possess a phosphoketolase pathway of glucose metabolism. The cleavage of xylose 5-phosphate by pentose phosphoketolase produces acetylphosphate and glyceraldehyde 3-phosphate. Further metabolism of these two products under fermentative conditions then produces ethanol and lactic acid. The major pathways of ethanologenesis, however, as mentioned, centre on the fate ofpyruvate, a critical point
not only in the divergence of different fermentation pathways but also in the divergence of fermentative and oxidative metabolic pathways in a single organism. The two alternative pathways are shown in Fig. 1. The pathway involving the direct decarboxylation of pyruvate to acetaldehyde via pyruvate decarboxylase has been termed the 'decarboxyclastic pathway' of ethanologenesis s. The enzyme pyruvate decarboxylase is found in only a few microorganisms 9 but is the unifying feature of the two most efficient mesophilic ethanologens, Zymomonas species and the yeast Saccharomyces cerevisiae. The key role of this enzyme in 'decarboxyclastic' ethanol production has been demonstrated by the inability ofS. cerevisiae mutants, lacking the enzyme, to grow fermentatively on glucose to produce ethanol 1°. Thermophilic ethanologens do not possess pyruvate decarboxylase but produce ethanol from pyruvate using acetyl-CoA as an intermediate. Variations in types of substrate and in fermentation conditions such as pH and temperature can drastically affect the ratios of the potential end-products of the fermentation. A number of the principles invotved in the manipulation of fermentative pathways and, in particular, the importance of the interrelationships of carbon balance and redox balance are demonstrated by our experience with B. stearotherrnophilus, which I shall consider later in some detail. In general, ethanol prodtiction by thermophilic ethanologens is governed by two factors. The first factor is the ethanol tolerance of the bacterium. Clearly, although ethanol may be constantly removed at, for example, 70°C, tolerance to ethanol concentrations which may be encountered during a fermentation is important and the
selection of tolerant strains is one approach to improving ethanol yields. Thermophiles such as Thermoanaerobiurn brockff have low ethanol tolerance H (1%) whereas some such as B. stearothermophilus have a naturally high ethanol resistance (4%) ~. The second factor influencing ethanologenesis is the pathway of carbon and electron flow present in the bacterium. For both Clostridium thermocellum and T. brockii it has been reported that one major obstacle to high ethanol production is the complication of alternative pathways producing lactate and acetate as alternative end-products ~3. All the thermophiles listed in Table 2 produce ethanol using essentially an identical pathway involving the EmbdenMeyerhoffpathway ofglycolysis. They all normally produce other fermentation products in addition to ethanol. It has been argued ~3 that the removal of potentially competing pathways by mutagenesis or optimization of the fermentation process by varying fermentation parameters such as pH
NADH NAD+~,~ I lactate CoA-~ C02~
l acetaldehyde ]
[ ethanol I phosphoroclastic
[ ethanoq decarboxyclastic
Fig. I. A comparison of the "phosphoroclastic" and 'decarboxyclastic' pathways of microbial ethanologenesis.
Trends in Biotechnology, VoL 2, No. 6, 1984
4 NAD ÷
t I~1111~ + / I 2 NAD "~- I I
CoA - "~,, I
[ ,ormate p,_
.,Will I 12acetylCoA]
i I, -
I.- ADP ~ATP / /
[ acetatep .....
I II I I
NADH . . . . . . . . . . .
NAD+.qt _ _ ..../II ] acetaldehyde ]
" " . . . . . I,~ 4 NAD +
dehydrogenase (PDH) which generates reducing equivalents (as NADH) in addition to those generated in glyco.~ 4 NAD + lysis. This additional generation of 'reducing potential' is, of course, not a problem under aerobic oxidative ~ 4NADH conditions since oxidatively growing organisms possess a high-efficiency 'energy-sink' in the form of the electron-transport or cytochrome + 4NAD chain. Under oxygen-limited fermentative conditions, however, this system is altered and B. stearothermophilus, in order to balance its redox state, reduces NADH production. This involves replacement of PDH by an alternative ~2' 4 N A D H enzyme system: pyruvate-formate lyase or PFL. The wild-type strain of B. stearothermophilus (NCA 1503) grows very rapidly on glucose or sucrose anaerobically. By conventional strain improvement techniques we isolated a faster-growing variant strain PSII which can grow anaerobically on complex media plus 0.2% (w/v) glucose at 65°C with doubling times of 20 minutes '2. We showed, as did others, that under such conditions L-lactate was the major fermentation product because NADqinked lactate dehydrogenase (probably in this context more 4 NADH properly termed pyruvate reductase) rapidly metabolized accumulated pyrnrate to L-lactate. In other words, the pathway used by B. stearothermophilus PSII could be represented as: 2glucose + 4NAD+---4~4 pyruvate + 4NADH
] 4 acetaldehyde ]
NADH . . . . .
. . . . .
~ \I t
:,. . . . . . . . . . . . . . . . .
2 lactate + I acetate + I ethanol + 4NAD +
We reasoned therefore that removal of lactate production as a potential NAD+~__ .../II i'.. . . . . . ~ - 4 N A D + 'sink' for the reducing equivalents by mutationally deleting NAD-linked L] 4 ethanol ] lactate dehydrogenase would make available both more carbon and more Fig. 2. The aerobk (-~) and fermentative (- ~ ) metabolism of glucose by Bacillus stearothermophilus. Enzymes: PDH = pyruvate dehydrogenase, PFL = pyruvate formate lyase, LDH = L- reducing equivalents for ethanololactate dehydrogenase, AcDH = acetaldehyde dehydrogenase, ADH = alcohol dehydrogenase, genesis according to the following PTA = phosphotransacetylase, ACK = acetate kinase. The putative anaerobic PDH-pathway is scheme: representedby . . . . ~2 glucose + 4NAD + - - - - ~ 4 pyruvate + 4 N A D H t
properties, despite its apparent original low ethanol production capacity, since we felt it offered a unique opportunity to employ a 'metabolic steering' approach to a novel organism for the production of a bulk chemical product. The pathways of aerobic and anaerobic glucose utilization displayed E t h a n o l p r o d u c t i o n by B. stearothermophilus by wild-type B. stearothermophilus are The thermophile B. stearothermo- shown in Fig. 2. Aerobically the fate of philus was chosen for its rapid growth pyruvate is metabolism via pyrnvate
might improve ethanol yields. As a general model for the optimization of an ethanologenic fermentation, I will describe some work performed in our laboratory using the thermophile B. stearothermophilus.
¢ 2 acetate + 2 ethanol + 4 NAD +
We selected mutants of PSII lacking lactate dehydrogenase by a novel method involving the use of resistance to a 'suicide substrate' the analogue fluoropyruvate (Payton and Hartley, unpublished results). Table 3 shows an analysis of fermentation products of such a mutant IId-15 relative to the
Trendsin Biotechnology, Vol.2, No. 6, 1984
wild-type, showing that the second Table3. A comparisonof the products of glucosefermentation by wild-type B. stearothermoequation above was stoichiometricaUy philus (PSII) and a mutant lacking NAD-linked t-lactate dehydrogenase01d- 15). correct, and that a single mutation Moles of product per mole of glucose used Strain Medium Lactate Acetate Ethanol increased ethanol yields by at least 100%. We believe the higher levels of PSII CG 1 2 0.5 acetate produced by PSII and the IIdI5 CG 0.04 3.38 1 PSII SG I 0.5 0.5 mutant IId-15, following growth on IIdI5 SG 0 1 1 complex medium, result from the use All fermentations were performed at 57°C. The media contained in addition to glucose, either 3% of amino acids in the tryptone and/or tryptone, 1% yeast extract (CG medium) or 0.1% tryptone, 0.05% yeast extract (SG medium). yeast extract. F e r m e n t a t i o n studies w i t h IId-15
increase ethanol yields from Iid-15. We reasoned that we now had an The results of such experiments ~14, organism with essentially no shortage suggested that ethanol yields could of carbon for ethanol production and in indeed be increased up to an overall which the obvious limiting factor, now, molar yield of 3 mol ethanol/hmol in improving ethanol yield from hexose sucrose. To date, we have no evidence or sucrose was the availability of of the cause of the 'switch' from the reducing equivalents in the form of theoretical maximum ethanol producNADH. From Fig. 2 it can be seen that tion allowed by the sole use of the 'PFLfor each mole of acetyl CoA metabol- pathway' to the higher levels of ethanol ized to ethanol via acetaldehyde there is production achieved in our fermentaa requirement for 2 moles of NADH tion. We have argued previously that since both acetaldehyde dehydrogenase though it is tempting to postulate a shift and alcohol dehydrogenase use towards our proposed PDH-pathway NAD(H) as co-factor. To achieve the of ethanologenesis, there are other maximal theoretical ethanol yield possible explanations for the observed allowed (i.e. 2 moles of ethanol per phenomena. For example, the same mole of glucose) we must produce 2 result would be obtained if the formate moles of acetyl CoA, and 4 moles of produced by PFL was used to generate NADH from one mole of glucose. We NADH via an NAD+-linked formate considered the possibility, suggested by dehydrogenase, thus providing suffiFig. 2, that under certain conditions we cient additional reducing equivalants to could manipulate the metabolism ofthe enable higher ethanol production. organism in such a way as to 'steer' Nevertheless, we have shown that using a combination of classical strain from a 'PFL-pathway', improvement techniques (the selection 1 glucose + 2NAD +- -"~21pyruvate + 2NADH of PSII) metabolic steering by removal of competing pathways (the isolation of Iid-15) and optimization of fermenta2 acetyl CoA+ 2 formate tion parameters can and did result in a I t , great improvement in ethanol yields by 1 acetate + 1 ethanol + 2NAD the thermophile B. stearothermophilus (productivity values of 1.3 g ethanol, g to a theoretical 'PDH-pathway', cells-', h-l). 1't This work, dearly, followed the path I glucose + 4NAD +----"~2 pyruvate + 2NADH [ +2NAD + outlined in Table 1 option 2, i.e. the I PDH improvement of an ethanologenic t thermophile to increase the ethanol 2 acetyl CoA + 2CO2 + 4NADH yields. It has served to introduce the I l importance of both carbon and redox l balance to the efficiency of ethanol pro2 etha~ol + 4NAD +. duction. These points have been As we have previously described ~4, a illustrated for other thermophiles by mutational approach to this problem other authors and, essentially, optimwas unsuccessful, largely owing to our ization of ethanologenesis in these lack of understanding of the normal bacteria has proceeded by a scheme control mechanisms governing the similar to that outlined above for B. switch from aerobic to anaerobic stearothermophilus. growth in this organism. We therefore At first sight, one major disadvantage concentrated on the variation of fer- of B. stearothermophilus might appear mentation parameters such as aeration, to be its inability to use, as carbon temperature and pH in an attempt to source, cellulosic substrates (Table 2);
particularly since we have seen that a number of thermophiles already possess the ability to digest cellulose wastes and other abundant (and cheap) energy sources, which is, of course, an advantage in the short term. However, a long term strategy may involve the use of recombinant DNA techniques to increase the number of potential feedstocks for a particular thermophile, and this need not be restricted to use in ethanologenic systems alone. The successful cloning and expression of cellulose genes has already been reportedL One must bear in mind, though, that the utilization of complex mixtures of cellulosic substances usually involves the simultaneous or sequential action of several enzymes, not just one. Thus in order to transfer the total celluiolytic capabilities of one organism to another, as many as four or five genes may need to be cloned. Also, to ensure control over when these enzymes are synthesized it may be necessary to express one, or all of the cloned genes from a host promoter. For example, to express a cellulose gene cloned from Trichoderma species in B. stearothermophilus the cellulose gene may have to be coupled to the promotor for B. stearothermophilus amylase. In addition, host-vector systems are being developed for thermophiles and notably for B. stearothermophilus 's. The cloning of B. stearothermophilus enzymes such as neutral proteasC 6 and a plasmid-born a-amylase~vinto other strains of B. stearothermophilus has been reported and it is likely to be only a matter of time before 'foreign' genes coding for cellulolytic activities could be expressed in organisms such as B. stearothermophilus. As mentioned previously an alternative approach to using a single organism with less than perfect properties is to use a co-culture combining the favourable features of two thermophiles in a single fermentation. A number of co-cultures have been studied in this respect, mostly
Trends in Biotechnology, VoL 2, No. 6, 1984
involving combinations including Cl. as one component because it shows efficient cellulose degradation. The products of cellulose degradation by C1. thermocellum are cellobiose and glucose and the aim has been to find a partner for this organism which will mop-up these products rapidly and convert them (more efficiently than Cl. thermocellum alone) to ethanol. Probably the best candidate as a fermentative partner to C1. thermocellum is C1. thermohydrosulfuricum, since it has high ethanol yields. Cocultures of these two organisms are stable at temperatures in the region of 60°C and their metabolic interrelationships have been extensively studied TM. Moreover, the combination of these two organisms is synergistic and allows ethanol production from sources previously untouched by either organism alone. The use of co-cultures appears highly favourable and development of such systems is being actively pursued in a number of laboratories. thermocellum
The economics of a thermophilic ethanol f e r m e n t a t i o n It is fair to say that arguments rage concerning the economics of thermophilic as opposed to mesophilic ethanol fermentations. The feed-stock of the fermentation is the major recurrent cost of such processes and therefore, not surprisingly, relative fluctuations in the prices of sugar-cane etc. result in widely differing assessments of the overall pro-
Glossary thermophile - an organism capable of growth between 45 and I O0°C, with a temperature optimum usually in excess of 5 0 - 6 0 ° C. mesophile - an organism capable of growth between 20 and 45°C, with a temperature optimum in the range 2837°C. ethanologenesis - used in this context to mean generation of ethanol by microorganisms. fermentation - used here to describe the process by which microorganisms degrade carbohydrate and use intermediates generated from the carbohydrates as electron acceptors producing reduced 'fermentation products'. suicide substrate - a compound which is metabolized by a particular enzyme, the 'target' enzyme to give a product toxic to the cell Hence resistance to a suicide substrate is often a result of a change in, or a loss of the target enzyme.
duction costs of ethanol. A theoretical consideration of the energetics of a thermophilic versus a mesophilic ethanologenic fermentation by an organism such as S. cerevisiae has been reported by us, in which we attempted a breakdown of predicted costs of a fermentation using B. stearothermophilus IId-15 ~4. We estimated that ethanologenic fermentations using this strain could offer savings over a fermentation using S. cerevisiae and published yeast fermentation technology. We also argued that further expected strain improvements could offer additional advantages and may result in savings of up to 25% on current production costs. The question of whether any process for the production of ethanol by fermentation is economical in absolute terms is of course a very complex one. In short, in areas where there is an excess of potential substrate, be it starch, cellulose or lactose, then it is economical to use these substrates for ethanol production. This would be particularly true when oil prices were high and the alternative means o f disposal of the potential fermentation substrates actually cost money, as is the case for lactose in whey permeate for example. The factors that determine the basic costs of ethanol fermentation and current views on the subject have been reviewed recently 4. How close then are thermophilic bacteria to replacing yeast in industrial ethanol production? At this time, one must accept that probably the only prokaryote which poses a serious threat to yeast is Zymomonas. Two major factors influence this: First, most thermophiles do not consistently produce sufficiently high levels of ethanol to be unanimously considered as economic. (Production o f and resistance to 6% ethanol is frequently considered almost economic). Second, there are several disadvantages associated with all thermophilic fermentations, not just those for ethanol production. For example, there is a problem of dissolution o f low-grade steel in current fermenters when operated at high temperatures. A thorough consideration of the pros and cons of such fermentations has been published recently 19. In conclusion, we have seen how the relatively simple concept underlying the development of a thermophilic ethanologenic fermentation can be achieved in practice. The process
whereby ethanol can be efficiently generated from a variety of potential substrates has already been developed. In the case ofB. stearothermophilus this currently involves only substrates such as cane-juice and other sucrosecontaining syrups currently used in ethanol production. However, for other systems, such as the CI. thermocellum/Cl, thermohydrosulfuricum coculture, cellulosic substrates could be used. The decisive factor in whether to substitute such processes for those based either on ethanol-production from oil or on conventional fermentation using yeast is an economic one. Not only is the efficiency of fermentation and product yields important but also factors such as geographical location and the availability of feedstocks. Trends in Biotechnology might be an appropriate forum for more detailed correspondence concerning these and other specific variables. If sufficient effort could be directed towards solving the biochemical, and engineering problems facing thermophilic ethanol production then the potential of some of the thermophilic ethanologens described here could be realized very soon. Acknowledgement I would like to thank Prof. Brian Hartley, Director of the Imperial College Centre for Biotechnology, London, in whose laboratories the work on B. stearothermophilus cited herein was conducted. References 1 Zeikus,J. G, Ben-Bassat, A, Ng, T. K. and Lamed, R. J. (1981) in Trends in the Biology of Fermentations for Fuels and Chemicals (HoUaender,A., ed.), pp. 441461, Plenum Press 2 Swings, J. and De Ley, J. (1977) Bacterial. Rev. 41, 1-46 3 Walia, S. K. and Ingram, L. O. (1984) A.S.M. 84th Annual Meeting, 1984, Abst. H169 4 Esser, K. and Karsch, T. (1984) Process Biochem. 19, 116-121 5 Ljungdahl,L. G., Bryant, F., Carrera, L. Saild, T. and Wiegel, J. (1981)in Trends in the Biology of Fermentationsfor Fuels and Chemicals (Hollaender, A., ed.)~ pp. 397-419, Plenum Press 6 Zuber, H. (1981)in Trends in theBiology of Fermentationsfor Fuels and Chemicah" (Hollaender, A., ed.), pp. 499-512, Plenum Press 7 Cornet et aL (1983) FEMS MicrobioL Lett. 16, 137-141 8 Zeikus,J. G. (1980)Ann. Rev. Microbiol. 34, 423-464
Trends in Biotechnology, VoL 2, No. 6, 1984
9 Wood, W. A. (1961) in The Bacteria 13 Lamed, R. and Zeikus, J. G. (1980)J. (Gunsalus, I. C. and Stanier, R. Y., eds) Bacteriol. 144, 569-578 11, 59-150, AcademicPress 14 Hartley, B. S., Payton, M. A., Pyle, 10 Lancashire, W. E., Payton, M. A., D. L., Mistry, P. and Shama, G. (1983) Webber, M. J. and Hartley, B. S. (1981) in Biotech 83, pp. 895-905, Online Mol. Gen. Genet. 81, 409-410 Publications Ltd., UK 11 Lamed, R. and Zeikus, J. G. (1980) J. Bacteriol. 141, 1251-1257 15 Imanaka, R., Fujii, M., Aramori, I. and 12 Hartley, B. S. and Payton, M. A. (1983) Aiba, S. (1982)J. Bacteriol. 149, 824Biochem. Soc. Syrup. 48, 133-146 830
16 Fujii, M., Takagi, M., Imanaka, R. and Aiba, S. (1983)ff. Bacteriol. 154, 831-837 17 Aiba, S., Kitai, K. and Imanaka, T. (1983) App. Env. MicrobioL 46, 1059-1065 18 Ng, T. K., Ben-Bassat, A. and Zeikus, J. G. (1981)App. Env. MicrobioL 41, 1337- 1343 19 Sonnleimer, B. and Fiechter, A. (1983) Trends Biotech. 1, 74-80
Amino acid biosynthetic enzymes as targets of herbicide action
typhimurium and the yeast Saccharomyces cerevisiae.
Case histories Glyphosate
R. A. LaRossa and S. C. Falco Several structurally-unrelated herbicides act by blocking a m i n o acid biosynthesis. Since a m i n o acid m e t a b o l i s m is s i m i l a r in plants and wellstudied, m a n i p u l a t a b l e m i c r o b i a l s y s t e m s , the opportunity exists for a particularly productive interaction b e t w e e n m i c r o b i a l and plant m o l e c u l a r biology. Such a s y m b i o s i s m a y lead to n e w m e t h o d s for the identification and design o f crop protection c h e m i c a l s . Selective toxicity The use of crop protection chemicals to reduce loss due to weeds, insects and disease has become an integral part of modern agriculture. About one half of the annual $15 billion expenditure for pesticides is used to purchase herbicides. These products have traditionally been found by random screen-; ing o f newly synthesized compounds. The screening tests are designed to identify chemicals that control weeds with as little toxicity to other organisms as possible, in other words compounds which inhibit metabolic pathways unique to the plant kingdom. The most obvious metabolic distinction between the plant and animal kingdoms is that plants obtain energy via photosynthesis and, therefore, it is not surprising that many herbicides interfere with this process. Many other metabolic differences between plants and animals exist, including synthesis of plant-specific hormones and the plant cell wall. No herbicides have been definitively shown to act on these processes, however. Plants and animals also differ in their ability to synthesize amino acids and vitamins. While plants are able to synthesize all these compounds, animals must obtain vitamins and the so-called essential amino acids from their diet. In this review we will focus on herbicides which act by R. A. LaRossa and S. C. Falco are at the Central Research and Development Department, E. I. du Pont de Nemours and Company, Wilmington, DE 19898, USA. © 1984, ElsevicrSciencePublishersB.V.,Amsterdara
inhibiting essential amino acid biosynthesis (Table 1). Since herbicides are of great agronomic importance, considerable effort has been devoted to physiological studies of their mode of action. It is only recently, however, that the molecular targets of some of them have been determined. The most extensively studied example is atrazine which blocks photosynthesis by binding to the 32 kDa protein ofphotosystem II, thus displacing a plastoquinone essential for electron transfer'. Several herbicides, e.g. glyphosate, chlorsulfuron and sulfometuron methyl, have recently been shown to act by inhibiting amino acid biosynthetic enzymes (Table 1 and Fig. 1). In these cases, determination of the molecular target was greatly aided by studying the action of the compounds on microorganisms such as the bacteria Escherichia coil and Salmonella ARGININE PROLINE GLUTAMINE
Glyphosate, sold by the Monsanto Company under the trade name Roundup ®, is widely used as a nonselective herbicide. Early studies of the action of glyphosate using the aquatic plant duckweed and the bacterium Rhizobium japonicum demonstrated that growth inhibition could be prevented by adding the aromatic amino acids phenylalanine, tyrosine and tryptophan to the culture medium 2. Analogous results were observed with E. coli, the unicellular alga Chlamydomonas and several higher-plant cell cultures 3'4.However, complete reversal in intact higher plants was generally not observed. Thus an indication that glyphosate inhibited an enzyme common to the biosynthesis of the three aromatic amino acids existed, but alternative explanations were also put forward 5. As indicated in Table 1 enzymological analyses suggest that glyphosate interferes with two steps of the common aromatic amino acids pathway 6,7. Molecular genetic techniques have shown that 5-enolpyruvyl-shikimate-3phosphoric acid (EPSP) synthase, an enzyme required for the synthesis of aromatic amino acids is an in vivo target
t .~'gPT GLUTAMATE t c~-ketoglutarate, pEP, erythrose-4-phosphate
~SERINE J t ; 3-phosphoglycerate
GP ~,l'~ TYROSINE- -
, PRPP, A TP oxaloacetate - ~ aspartate
~ AT HISTIDINE PHENYLALANINE
LYSINE~A~-~ I METHIONINE THREONINE [
ISOLEUClNE VALINE LEUCINE
Fig. I. An overview of amino acid biosynthesis. Metabolic intermediates are given in italics while the herbicides discussed in this article are boxed (AC, AC 243, 997; AT, aminotriazole; CS, chlorsulfuron; GP, glyphosate; PPT, phosphinothricin; SM, sulfometuron methyl). Amino acids essential to the mammalian diet are underlined; the metabolic pathways depicted below the broken line do not occur in mammals and are, therefore, potential targets, in the search far selectively toxic herbicides.