198

Progress in metabolic engineering William R Farmer* and James C Liaot Major advances in metabolic engineering from the past two years include the production of ethanol from Zymomonas mobi/is grown on xylose at high yield, the production of aromatic amino acids from Escherichia co/i at theoretical yield, the over-production of the aspartate family of amino acids from Corynebacterium, and the successful production of polyhydroxybutyrate in Arabidopsis thaliana. Although insights into metabolic regulation in central metabolism have been gained, much remains to be investigated in the area of metabolic flux regulation and global responses.

Address *tDepartment of Chemical Engineering,Texas A&M University, College Station, Texas ?7840, USA; *e-mail: [email protected] te-mail: [email protected] Current Opinion in Biotechnology 1996, 7:198-204 © Current Biology Ltd ISSN 0958-1669 Abbreviations DAHP E4P PCK PDH PEP PHA PHB P(3HB-co-3HV) PPC PPS PTS TCA Vhb

3-deoxy-D-arabino-heptulosonate-7-phosphate erythrose 4-phosphate PEP carboxykinase pyruvate dehydrogenase phosphoenolpyruvate polyhydroxyalkanoate polyhydroxybutyrate poly-(3-hydroxybutyrate-co-3-hydroxyvalerate) PEP carboxylase PEP synthase phosphotransferase system tricarboxylic acid Vitreoscilla hemoglobin

Introduction Metabolic engineering is defined as the manipulation of the metabolic pathways of a living organism to achieve a desired goal. Different applications of metabolic engineering include the production of metabolites and biodegradable polymers, the biodegradation of toxic chemicals, and the modification of cell properties for biotechnological purposes. In addition, altering metabolic enzymes or pathways has become an important approach for investigating cell physiology.

T h e biosynthesis of chemicals is a safer and cleaner alternative to traditional chemical syntheses that generate environmental and safety hazards. Even so, biosynthetic processes are often financially uncompetitive. Thus, research has focused on modifying microorganisms to improve product yields and increase substrate ranges. One of the common goals in metabolic engineering is the design of an ideal pathway or microorganism for the desired product. Because of the complexity of metabolic regulation, however, it is difficult to predict the outcome

of pathway engineering. Increasing the concentration of an enzyme or decreasing the concentration of its competitors does not always lead to increased flux through the desired pathway. Many controlling factors need to be considered in an effort to modify metabolic fluxes. T h e difficulty in prediction calls for more research in cell physiology and metabolic regulation. As in other biological problems, the key to a successful investigation in this area is the proper design of control experiments. Without carefully designed controls, results may be interpreted wrongly. Because the field of metabolic engineering is broad and diverse, this review focuses on the production of metabolites and biopolymers from microorganisms and on related physiological investigations. Progress in the biosynthesis of ethanol, butanol, amino acids, biodegradable polymers, and other compounds is highlighted. Advances in cellular engineering to improve cell growth and recombinant protein over-expression are also discussed. In addition, we consider investigations into the central metabolism of Eschefichia coli. Because previous reviews [1,2] have established an excellent foundation for this field, we focus only on advances made from 1994 to the present. Research on the biosynthesis of secondary metabolites is not discussed here because of constraints on space. Investigations of central metabolism Central metabolism is of key importance to successful metabolic engineering because it involves the common pathways from which the metabolic fluxes are distributed to different branches. Understanding flux distribution and its regulation in central metabolism is fundamental to the progress of metabolic engineering as a whole. Although the pathways of central metabolism have been elucidated for a few decades, flux regulation and its roles in global physiology are still understood less well than most people perceive.

To determine the role of the tricarboxylic acid (TCA) cycle, Lee etal. [3°] have investigated the flux distribution in an E. coli mutant (gltA) lacking citrate synthase activity. Even though carbon flux from glycolysis was unable to enter the TCA cycle, the amount of ATP produced from glycolysis was comparable to that produced in the wild type. This result reconfirmed that the TCA cycle in E. coli has a primarily biosynthetic role and that ATP is produced from glucose mainly through glycolysis and the acetate pathway. Unexpectedly, the presence of 3% CO z in the feed gas causes the specific growth rate of the gltA strain to increase by almost 25% and the final growth yield to increase by 20%. In addition, production of formate and malate is drastically reduced in the presence of CO 2. None of these effects is observed in the wild type. It was speculated that the presence of

Progress in metabolic engineering Farmerand Liao 199

CO 2 activates the phosphoenolpyruvate (PEP) carboxylase (PPC)-catalyzed reaction so that the flux through formate formation is reduced. T h e reason for the reduction of malate production is, however, unknown. An important point of flux regulation is the junction between glycolysis and the T C A cycle, which includes enzymes interconverting PEP, pyruvate, and oxaloacetate. These enzymes are important for directing metabolic fluxes to various biosynthetic pathways. Our group [4] has systematically studied the flux regulation in this area. Surprisingly, we found that the over-expression of these enzymes caused metabolic imbalance and severe disturbance to global regulation in E. coll. These results suggest that some of the metabolites in central metabolism may be signals in the coordination of various global regulons, although detailed mechanisms remain to be elucidated. In an effort to investigate the flux distribution and the regulation of energy transduction, we [5 °*] have over-expressed two enzymes, PPC and PEP carboxykinase (PCK), simultaneously, and characterized the flux distribution. T h e simultaneous over-expression of these two enzymes could potentially cause futile cycling and results in ATP hydrolysis. Indeed, the over-expression of these two enzymes stimulated the respiration rate, glucose consumption, and the production of pyruvate and acetate. Over-expression of inactive PCK and PPC, or over-expression of either PCK or PPC alone, did not cause the same effects. This suggests that E. coli has a mechanism to sense or compensate for futile cycling. Again, the exact mechanisms underlying these effects are still under investigation. Another key step in central metabolism is catalyzed by the E. coli pyruvate dehydrogenase (PDH) complex, which converts pyruvate to acetyl co-enzyme A. This complex contains three components: pyruvate dehydrogenase (Elp), lipoate acetyhransferase (E2p), and lipoamide dehydrogenase (E3). T h e E2p chain (coded by aceF) in E. coli contains three highly homologous 80-residue lipoyl domains, whereas the E2p chains of Streptococcus faecalis and mammalian mitochondria contain two, and the E2p chains of Bacilli and yeast contain only one. The deletion of aceF subsegments (i.e. llip, which encodes one net lipoyl domain, or 2lip, which encodes two net lipoyl domains) has little effect on the specific activities of the purified complexes relative to the wild-type complex containing three lipoyls and PDH. Therefore, it was unclear why E. coli contains three lipoyl domains in the E2p chain. This question, although not directly related to practical metabolic engineering, is fundamental to the understanding of the role of P D H in central metabolic flux distribution. Dave et al. [6 °°] have constructed three isogenic E. coli strains that contain 1lip, 2lip, and 3lip (wild-type) segments of the aceF gene on the chromosome. T h e y found that the maximum growth rate of the 1lip

and 2lip strains is lower than that of the 3lip strain when they are grown on glucose, D/L-lactate, and succinate in batch cultures. When grown on pyruvate, however, the maximum growth rates for the mutants are similar to the wild type. T h e authors [6**] hypothesize that pyruvate lifts the repression of the transcription of the pdhR-aceEF-lpd operon, thus inducing higher levels of P D H complex expression. Presumably, this could compensate for the inefficiency of carbon flux to biomass or energy generation because of an insufficient number of lipoyl domains. In continuous cultures using a glucose minimal medium, the growth yield and the maintenance energy decreased systematically with the number of lipoyl domains in the E. coli strain: the wild-type (three domains) had the highest values, the two-domain mutant had intermediate values, and the one-domain mutant had the lowest values. These results show that three domains per E2p chain are optimal for balanced growth on carbon substrates requiring P D H activity. Thus, the role of the lipoyl domains can be demonstrated in vivo only using the altered enzyme complex. Alteration of cellular physiology Various investigators are interested in improving cell growth and recombinant protein over-expression by modifying cell physiology. T h e elevation of dissolved oxygen concentration in vivo and the reduction of acetate production are two areas that have been the focus for research.

Bailey and colleagues [7"] have reported that the expression of Vitreoscilla hemoglobin (Vhb) in E. coli increases the transcriptional activity of oxygen-regulated promoters and improves cell growth by increasing the dissolved oxygen concentration in vivo. Fnr, a regulator protein of various oxygen-regulated genes in E. coli, also activates the Vhb promoter. Intracellular expression of Vhb was shown to cause a 1.5-fold increase in the transcriptional activity of the promoter for cytochrome d, Pcyd, under microaerobic conditions. One explanation for this effect is that Vhb might raise the dissolved oxygen concentration to a level that triggers the Arc regulatory mechanism. ArcA, in conjunction with other regulators, is speculated to increase the transcriptional activity of Pcyd. In addition, evidence suggested that Pvhb competes with other oxygen-regulated promoters for transcriptional activators. The presence of Pvhb alone decreased the transcriptional activity of Pcyd by 3.5-fold. Further investigation is required to understand the interactive effects of Vhb with other cellular enzymes in E. coli. Various investigations have focused on the reduction of acetate production in E. coli to improve cell growth and recombinant protein over-expression. One method involves the over-expression of the gene for acetolactate synthase from Bacillus subtilis [8°,9]. Acetolactate synthase channels carbon away from acetate production to acetoin, which is 50-fold less toxic. Cell densities were reported

200 Biochemicalengineering

to increase by 35% compared with the control, and the activity of an over-expressed fusion protein, CadA-13galactosidase, increased by 60% in a fed-batch fermenter. Another approach involved modification of the glucose phosphotransferase system (PTS) by inactivation of the gene encoding enzyme IIGlc of the glucose PTS, ptsG [10"]. In complex media, growth rates of the mutant were unaffected, a cell density of >20gl -1 was achieved, and recombinant protein production increased by 1.5-fold to >l.6g1-1 in a batch bioreactor. Modifying the glucose uptake rate may have reduced the glycolytic flux, resulting in the apparent inhibitory effect on acetate production. Another investigation 'into acetate accumulation has focused on the reduction of the glycolytic flux in B. subtilis from another direction [11"]. Lee etal. [3"] have suggested that acid production is caused by unequal fluxes in glycolysis and the T C A cycle. To test this hypothesis, they added citrate to glucose minimal medium to reduce the glycolytic flux. As predicted, acetate production was suppressed. Other T C A intermediates were not observed to cause this suppression, which suggests that a specific mechanism involving citrate was involved. Goel et al. [11 °] speculate that the coupled transport of divalent metal ions with citrate may cause certain metabolic enzymes, such as phosphofructokinase and PDH, to have reduced activities.

Ethanol production Research in this traditional area has remained active and focuses on engineering an organism to utilize a broad range of substrates, thereby producing ethanol at high yields in a short period of time. On the basis of a number of different criteria, Zhang et al. [12] have selected two strains, Zymomonas mobilis and Lactobacillus, with the highest potential for the industrial production of ethanol. Some of the associated traits included high conversion yield, high ethanol tolerance, broad substrate range, and high fermentation selectivity. Z. mobilis was successfully engineered to ferment the pentose sugar xylose by introducing the xylose assimilatory pathway and the pentose phosphate pathway enzymes transketolase and transaldolase [13°°]. Ethanol production from xylose was 86% of the theoretical yield and a mixture of xylose and glucose was fermented to ethanol at 95% of the theoretical yield within 30 h. Lindsay et al. [14] have isolated E. coli mutants that are not able to ferment hexose sugars transported via the PTS. These mutants were still able to ferment non-PTS sugars and were used to ferment pentose sugars. Other mutants were able to ferment both hexose and pentose sugars to ethanol at higher yields than the parent, which preferentially fermented hexose sugars. Two of these mutants were able to produce 60g 1-1 ethanol from 120gl-I xylose. In addition, they produced >40gl-1 ethanol in 60 h from a sugar mixture consisting of 60 g 1-1 glucose and 30gl -1 xylose. Ethanol production from the sugar mixture approached 100% of theoretical yield.

One major problem with ethanol production from E. coli is ethanol tolerance. To overcome this, Lindsay et al. [14] suggested a two-stage fermentation process involving E. coli and yeasts to achieve higher ethanol concentrations. It remains to be seen, however, whether E. coli can be used successfully to produce bulk quantities of ethanol. In addition to their potential industrial relevance, investigations in this area have contributed to the general understanding of metabolic flux regulation.

Butanol production Expression of the cloned gene aad, which encodes aldehyde/alcohol dehydrogenase, in a mutant Clostridium acetobutylicum strain deficient in acetone and butanol production has recently been shown to restore butanol formation and butyraldehyde dehydrogenase activity [15"]. T h e final concentration of butanol in the medium in a batch fermentation was 84mM. Clearly, the results show that butyraldehyde dehydrogenase activity is primarily responsible for production of butanol, rather than ethanol or acetone. In an extensive physiological study of C. acetobutylicum in a phosphate-limited continuous culture, Girbal and Soucaille [16"] have reported that a high N A D H / N A D ratio inhibits glyceraldehyde-3-phosphate dehydrogenase and is associated with the induction of ferredoxin-NAD reductase and low-level expression of various enzymes, including NADH-ferredoxin reductase, ethanol dehydrogenase, butyraldehyde dehydrogenase, and butanol dehydrogenase. In addition, the in vitro activities of phospho-transbutyrylase and butyrate kinase decreased linearly with ATP concentration. These physiological studies in continuous cultures suggest that a low ATP level and a high N A D H / N A D ratio in vivo may lead to butanol production. This contradicts the hypothesis proposed by Papoutsakis [17] that holds the ATP level as the primary factor in acidogenesis/solventogenesis.

Aromatic compounds Aromatic compounds that can be produced from microbial processes include aromatic amino acids, indigo, catechol, and a variety of quinoid organics. The biosynthesis of these chemicals from renewable resources is a potential alternative to the traditional method of production from petroleum feedstocks. Frost and Lievense [18] discuss the advantages of aromatic biosynthesis and prospects for the future. These aromatic compounds are derived from the common aromatic pathway in bacteria, which uses PEP and erythrose 4-phosphate (E4P) as precursors to produce 3-deoxy-D-arabino-heptulosonate-7-phosphate (DAHP), the first metabolite committed toward aromatics production in the cell. Through a stoichiometric analysis, our group [19"',20"] has argued that the yield of aromatic metabolites from glucose is limited by the ability to recycle pyruvate to PEP, which is also consumed in the glucose transport process via the PTS. Guided by the stoichiometric analysis, we over-expressed PEP synthase (PPS), thereby attempting to recycle pyruvate to PEP. T h e initial effort was fruitless because of other limiting steps

Progress in metabolicengineeringFarmerand Liao 201

in the pathway. In the presence of over-expressed DAHP synthase (encoded by aroG) and transketolase (encoded by tktA), however, PPS was able to increase the amount of aromatic metabolites to near theoretical yield. T h e role of over-expressed transketolase was to increase the supply of E4P, another precursor for DAHP synthesis, and the over-expressed DAHP synthase removed the final bottleneck in DAHP production. These modulations resulted in a maximum yield of 86% (mol/mol) of aromatic amino acids from glucose. Because the PTS was found to affect production by converting PEP to pyruvate, an alternative suggestion was to use a non-PTS sugar as the substrate instead of over-expressing PPS [20°°]. Indeed, a comparable yield of 71% (mol/mol) was obtained on xylose. T h e s e results are encouraging because they suggest that predictions from stoichiometric analysis can be realized provided that a proper condition is found. Ikeda and Katsumata [21] suggested that the uptake rate of aromatic amino acids is an important consideration in amino acid over-production. By introducing the aromatic amino acid transport system into a tryptophan-producing Corynebacterium glutamicum mutant defective in the transport system, the production rate of tryptophan decreased tremendously. T h e presence of a transport system was hypothesized to increase the intracellular concentration of aromatic amino acids, which leads to feedback inhibition of the regulatory enzymes in the pathways. In another report, Dudley and Frost [22] reported the isolation of a Klebsiella oxytoca mutant that grows on arylsulfonates and produces the corresponding phenol products. T h e biosynthesis of phenols is an appealing potential alternative to the current process, which requires extreme temperatures, extreme pH levels, and unsafe process designs. Further work is required to determine the genes responsible for the metabolism of arylsulfonates in this mutant.

T h e a s p a r t a t e f a m i l y of a m i n o acids Aspartate type amino acids have an increasing range of applications, including use as supplements in animal feed, as nutritional supplements, and as therapeutic agents. Corynebacterium strains are the primary strains used to produce these amino acids. T h e physiology and genetics of these species have been reviewed elsewhere [23,24]. To engineer C. glutamicum for lysine over-production, Gubler et al. [25] have inactivated the pyruvate kinase gene, pyk. T h e y hypothesized that inactivation of pyk would create an equimolar ratio of PEP and oxaloacetate, thus optimizing lysine biosynthesis. Unexpectedly, lysine production in the pyk mutants was 40% lower than in the pyk ÷ strain, although the growth rates and glucose consumption rates were similar. In addition, derivatives of glycolytic intermediates were observed in the medium, possibly suggesting an accumulation of glycolytic inter-

mediates upstream of PEP. These observations point to complex regulation at the PEP branchpoint. Further investigations are required to understand the regulation in the metabolic pathways of C. glutamicum. For threonine accumulation, Colon et al. [26 °] have reported a strategy using established techniques to redirect carbon flux from the lysine biosynthetic pathway to threonine biosynthesis in a lysine-producing Corynebacterium lactofermentum strain. By abolishing feedback inhibition in the threonine pathway by isolating a feedback-insensitive allele (homdr) and over-expressing the operon homdr-thrB, threonine accumulated t o - l l . 8 g l - ! and the lysine concentration decreased to -0.8 gi -1. Furthermore, these authors [26 °] reported significant amounts of glycine and isoleucine, which were hypothesized to result from threonine degradation. In another study, Motoyama et al. [27] investigated the possibility of using Methylobacillus glycogenes to produce threonine from methanol. T h e y reported that over-expression of the hom-thrC genes resulted in threonine accumulation between 2.2 g l-I and 2.4gl -1. Glutamate production was still very high in this strain, and more work is needed to redirect the carbon flux from glutamate biosynthesis to that of the aspartate family amino acids. Colon et al. [28 °] have also reported the production of isoleucine from a threonine-producing mutant of C. lactofermentum. As expected, isoleucine production was achieved by redirecting carbon flux from threonine biosynthesis to isoleucine biosynthesis. 15gl-I of isoleucine were produced in a shake-flask fermentation. A carbon balance indicated that 80% of the carbon available for threonine had been converted to isoleucine. Colon et al. [28 °] observed that feedback inhibition of ilvA, which encodes threonine dehydratase, in the isoleucine pathway was abolished by simple over-expression.

O t h e r areas of m e t a b o l i t e production Hydromorphone and hydrocodone have wide clinical use as analgesics and antitussives, respectively. T h e y are produced in low yields from morphine using chemical processes requiring toxic chemicals. French et al. [29] have reported an alternative way to produce hydromorphone and hydrocodone using recombinant E. coli grown on a medium containing 25mM morphine. T h e genes encoding morphine dehydrogenase (MorA) and morphinone reductase (MorB) from Pseuclomonas putida M10 were introduced into E. coli. T h e authors [29] concluded that better yields could be achieved by both increasing the ratio of MorB activity to MorA activity and engineering MorA to change its co-factor specificity from NADP + to NAD ÷. Previous work on the biosynthesis of 1,3-propanediol had been conducted by Cameron and co-workers [30,31]. This area has gained increasing importance as 1,3-propanediol may be used as a monomer in the production of plastics.

202

Biochemical engineering

Biodegradable polyesters Research in this area is motivated by the environmental and waste-management problems caused by nonbiodegradable polymers. Interest in the production of polyhydroxyalkanoates (PHAs) in bacteria focuses on finding strains that either can grow on a variety of substrates or can incorporate different monomer precursors to produce co-polymers [32]. Zhang et al. [33"] have reported the production of 24g1-1 polyhydroxybutyrate (PHB) in a Klebsiella aerogenes strain growing on molasses within 32 h. In addition, they isolated afadR mutant strain of Klebsiella oxytoca that was able to produce the co-polymer poly-(3hydroxybutyrate-co-3-hydroxyvalerate) (P[3HB-co-3HV]). Incorporation of 3-hydroxyvalerate was -23% when 5 mM propionate was added to the medium. Rhie and Dennis [34] studied the mechanism of co-polymer biosynthesis in E. coil by investigating the mode of action of fadR and constitutive atoC mutations in the incorporation of 3-HV into P(3HB-co-3HV). T h e y rationalized that the fadAB gene is derepressed in the fadR mutants, thus allowing thefadAB gene product, 3-ketoacyl-CoA thiolase, to bind intermediates of P(3HB-co-3HV) and function in co-polymer synthesis. Furthermore, constitutive atoC mutants were hypothesized to increase propionate uptake from the medium by decreasing the specificity of the uptake system. Much interest focuses on the production of PHAs in genetically engineered crop plants because of the potentially large quantity of products possible at low production costs [35,36]. Nawrath etal. [37"'] have recently reported the successful targeting of the genes for PHB production from Alcaligines eutrophus into the plastids of Arabidopsis thaliana. Because fatty acid metabolism occurs in the plastids of A. thaliana, these authors reasoned that production of PHBs in the plastids would not cause any serious physiological defects. Previously, PHB accumulation in the cytoplasm of A. thaliana cells was observed to stunt growth [38]. Indeed, PHB production in the plastids resulted in PHB accumulation levels up to 14% of the dry weight of the plant.

Conclusions T h e past two years have seen major advances in our understanding of the metabolic pathways of microorganisms that have enabled the production of metabolites from a greater variety of substrates. Z. mobilis has been genetically engineered to grow on xylose in lignocellulose and produce ethanol at high yield. It is anticipated that similar modifications on the metabolism of Saccharomyces cerevisiae will allow the efficient conversion of pentose sugars to ethanol [39,40]. Even greater yields of ethanol are possible because of the high ethanol tolerance ofS. cerevisiae. Using pathway engineering techniques based on a knowledge of reaction stoichiometry, the production of aromatic amino acids from E. coli, and of the aspartate-family amino acids from CorynebacteHum, resulted in product levels

approaching the theoretical yield. Progress has also been made in the production of biodegradable co-polymers from bacteria and the production of PHAs from plants. It is expected that research into co-polymer production from plants will eventually make this the predominant method of biodegradable plastic production because of the potentially low production cost and high yields in plants. Finally, insights into metabolic regulation in E. coil have been obtained, although much remains to be elucidated. The ultimate goal in metabolic engineering would be to rationally design metabolic pathways to optimize metabolite production, cell growth, and recombinant protein over-expression. Consequently, continuing advances into deducing the regulation of metabolic pathways and global regulation are paramount to the progress of metabolic engineering as a whole.

References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: * ..

of special interest of outstanding interest

1.

Bailey JE: Toward a science of metabolic engineering. Science 1991, 252:1668-1875.

2.

Cameron DC, Tong IT: Cellular and metabolic engineering: an overview. Appl Biochem Biotechno/1993, 38:105-140.

Lee J, Goel A, Ataai MM, Domach MM: Flux adaptations of citrate synthase deficient Escherichia coll. Ann NYAcad Sci 1994, 745:35-50. This work attempts to determine the role of the TCA cycle in the metabolism of E. coil by knocking out the pathway connecting glycolysis to the TCA cycle. Glycolysis in the mutant is determined to provide an ATP yield comparable to the wild type. Modulation of acetate/formate production is hypothesized to be a key factor in ATP production in the mutant. Carbon dioxide is observed to increase the growth yield and the specific growth rate by some unknown mechanism. 3. •

4.

Liao JC, Chao Y-P, Patnaik R: Alteration of the biochemical valves in the central metabolism of Escherichia coil Ann NY Acad Sci 1994, 745:21-34.

5. Chao Y-P, Liao JC: Metabolic responses to substrate futile ,,• cycling in Escherichia coil. J Biol Chem 1994, 269:5122-5126. Simultaneous over-expression of PCK and PPC in E. coil stimulates respiration, glucose consumption, and the production of pyruvate and acetate. These observations are speculated to be cellular responses to substrate futile cycling; however, the detailed mechanisms involved are still unknown. Dave E, Guest JR, Attwood MM: Metabolic engineering in Escherichia coil: lowering the lipoyl domain content of the pyruvate dehydrogenase complex adversely affects the growth rate and yield. Microbiology 1995, 141:1839-1849. An E. coil mutant with PDH complex containing a lower lipoyl domain content exhibits a decrease in growth yield and maintenance energy on a variety of sugars. When pyruvate is the carbon source, the mutant growth yield is similar to the wild type. This leads to the interesting possibility that pyruvate lifts the repression of the transcription of the genes for the PDH complex. Thus, increased expression of the mutant PDH can compensate for its reduced activity. 6. ••

Tsai PS, Kallio PT, Bailey JE: Fnr, a global transcriptional regulator of Escherichia coil, activates the Vitreoscilla hemoglobin (Vhb) promoter and intracellular Vhb expression increases cytochrome d promoter activity. Biotechnol Prog 1995, 11:288-293. The expression of Vhb in vivo is shown to enhance the metabolism of E. coil Under microaerobic conditions, Vhb raises the intracellular dissolved oxygen concentration and increases the transcriptional activity of various oxygen-regulated promoters. The transcriptional activity of the promoter for cytochrome d is shown to increase 1.5-fold. In addition, Fnr, a global regulator in E. coil, activates the transcriptional activity of the heterologous promoter for Vhb. The use of Vhb may provide a method to investigate the aerobic physiology 7. •

Progress in metabolic engineering Farmer and Liao

of E. coil and other microorganisms, and it may be exploited to increase process productivity. 8. •

Aristidou AA, San K-Y, Bennett GN: Metabolic engineering of Escherichia coil to enhance recombinant protein production through acetate reduction. Biotechnol Prog 1995, 11:475-478. The investigators successfully demonstrate their hypothesis that acetate production can be reduced by drawing carbon away to the production of acetoin, which is less toxic to E. coil. The gene encoding acetolactate synthase from B. subtilis is introduced into E. coil. Cell densities are increased by 35% and recombinant protein production is increased by 60% compared with the wild type. 9.

Aristidou AA, San K-Y, Bennett GN: Modification of central metabolic pathway in Escherichia coil to reduce acetate accumulation by heterologous expression of the Bacillus subtilis acetolactate synthase gene. Biotechnol Bieeng 1994, 44:944-951.

10. •

Chou C-H, Bennett GN, San K-Y: Effect of modified glucose uptake using genetic engineering techniques on high-level recombinant protein production in Escherichia coil dense cultures. Biotechnol Bioeng 1994, 44:952-960. By genetic modification of the glucose PTS system in E. coil, these authors correlate a decrease in glucose uptake with a decrease in acetate excretion. Cell densities of 20 g 1-1 and recombinant protein production of 1.6 g 1-1 are achieved in complex media. This investigation provides additional evidence that controlling the glycolytic flux is an effective way to reduce acetate production. 11. •

Goel A, Lee J, Domach MM, Ataai MM: Suppressed acid formation by co-feeding of glucose and citrate in Bacillus cultures: emergence of pyruvate kinase as a potential metabolic engineering site. Biotechnol Prog 1995, 11:380-385. This investigation provides strong evidence that the unequal fluxes between glycolysis and the TCA cycle cause an increased production of acetate. The reduction of the glycolytic flux by the addition of citrate to the medium is shown to cause a suppression of acetate accumulation. Interestingly, the addition of other TCA intermediates does not cause the same effect. This points to the possibility of an unknown specific mechanism for citrate that decreases acetate accumulation. 12.

Zhang M, Franden MA, Newman M, McMillan J, Finkelstein M, Picataggio S: Promising athanologens for xylose fermentation. Appl Biochem Biotechnol 1995, 51/52:527-536.

13. ••

Zhang M, Eddy C, Deanda K, Finkelstein M, Picataggio S: Metabolic engineering of a pentose metabolism pathway in ethanologenic Zymomonas mobilis. Science 1995, 267:240-243. Z. mobilis is successfully engineered to utilize xylose for ethanol production by introducing a xylose assimilation pathway and two genes encoding transketolase and transaldolase. By constructing a complete pentose phosphate pathway, Z. mobilis is able to ferment xylose and produce ethanol at up to 86% of the theoretical yield. Fermentation of a mixture of xylose and glucose yields ethanol production at 95% of the theoretical yield. This investigation represents a significant advance toward sustainable ethanol production from renewable resources. 14.

Nair RV, Papoutsakis ET: Expression of plasmid-encoded aad in CIostridium acetobutylicum M5 restores vigorous butanol production. J Bacteriol 1994, 176:5843-5846. Genetic complementation is used to show that the primary function of Aad, the gene product of aad, is to form butanol. In addition, plasmid-encoded aad restores butyraldehyde dehydrogenase activity to a strain deficient in any solventogenic enzyme activity. Butanol accumulates in the medium to levels of 84 raM. Further investigations are required to map out the solventogenic pathways of C. acetobutylicum. Girbal L, Soucallle P: Regulation of Clostridium acetobutylicum metabolism, as revealed by mixed-substrate steady-state continuous cultures: role of NADH/NAD ratio and ATP pool. J Bacteriol 1994, 176:6433-6438. This extensive physiological study on C. acetobutyficum reveals further insight into the function of the ATP pool and the NADH/NAD ratio on product selectivity. 16. •

1 7.

18.

Patnaik R, Liao JC: Engineering of Escherichia coil central metabolism for aromatic metabollte production with near theoretical yield. App/Environ Microbio/1994, 60:3903-3908. Owing to the action of the PTS, PEP is the limiting factor in the production of aromatic amino acids in E. coil By recycling pyruvate to PEP and directing carbon flux to the production of PEP and E4P, production of aromatic amino acids approaches a maximum yield of 860/0 (mol/mol) from glucose. This work demonstrates that the PTS is a critical component that needs to be considered in designing metabolic pathways. 19. •.

20. •,

Patnaik R, Spitzer RG, Liao JC: Pathway engineering for production of aromatics in Escherichia coli: confirmation of stolchiometrlc analysis by independent modulation of AroG, TktA, and Pps activities. BiotechnoI Bioeng 1995, 46:361-3?0. This investigation examines the roles of AroG, TktA, and PPS in the production of aromatics in E. coll. Independent promoters are used to control the expression levels of these enzymes. Results confirm the prediction from stoichiometric analysis. Furthermore, xylose, a sugar transported into E. coil independently from the PTS, is tested as a substrate for the production of aromatic amino acids. Because the PTS is not involved with xylose assimilation, the yield of aromatic amino acids on xylose should be comparable to the yield on glucose, without having to recycle pyruvate back to PEP. Indeed, this hypothesis is verified and a yield of 710/0 (tool/tool) on xylose is obtained. 21.

Ikeda M, Katsumata R: Transport of aromaUc amino acids and its influence on overproduction of the amino acids in Cotynebacterium glutamicum. J Ferment Bioeng 1994, 78:420-425.

22.

Dudley MW, Frost JW: BiocatalyUc desulfurization of arylsulfonates. Bioorg Med Chem 1994, 2:681-690.

23.

SahrnH, Eggeling L, Eikmanns B, Kramer R: Metabolic design in amino acid producing bacterium Corynebacterium glutamicum. FEMS Microbiol Rev 1995, 16:243-252.

24.

Jetten MSM, Sinskey AJ: Recent advances in the physiology and genetics of amino acid-producing bacteria. Crit Rev Biotechnol 1995, 15:73-103.

25.

Gubler M, Jetten M, Lee SH, Sinskey AJ: Cloning of the pyruvate kinase gene (pyk) of Corynebacterium glutamicum and site-specific inactivation of pyk in a lysine-producing Corynebacterium lectofermentum strain. Appl Environ Microbiol 1994, 60:2494-2500.

26. •

Colon GE, Jetten MSM, Nguyen TT, Gubler ME, Follettie MT, Sinskey A.I, Stephanopoulos G: Effect of inducible thrB expression on amino acid production in Corynebacterium lactofermentum ATCC 21799. Appl Environ Microbiol 1995, 61:74-78. The metabolic pathway for lysine biosynthesis in C. lactofermentum is optimized for the production of threonine. By abolishing feedback regulation in the pathway through the isolation of a desensitized allele and by overexpressing the genes in the threonine biosynthetic pathway, threonine concentration is increased to 11.8 g i-1 and lysine concentration is decreased to -0.8 g I-1. 27.

LindsaySE, Bothast RJ, Ingrain LO: Improved strains of recombinant Escherichia coil for ethanol production from sugar mixtures. App/Microbio/Biotechno/1995, 43:70-?5.

15. •

Papoutsakis ET: A useful equation for fermentations of butyric acid bacteria. In The Acetone-Butanol Fermentation and Related Topics. Science and Technology Letters. Edited by Bu'Lock JD, Bu'Lock AJ. Kew; Chameleon Press Ltd; 1983:121-126. FrostJW, Lievense J: Prospects for the biocataly~dc synthesis of aromatics in the 21 st century. New J Chem 1994, 18:341-348.

203

Motoyama H, Yano H, Ishino S, Anazawa H: Effects of the amplification of the genes coding for the L-threonine biosynthetic enzymes on the L-threonine production from methanol by a Gram-negative obligate methylotroph, Methylobacillus glycogenes. Appl Microbio/ Biotechnol 1994, 42:6?-72.

28. •

Colon GE, Nguyen "I-1",Jetten MSM, Sinskey AJ, Stephanopoulos G: Production of isoleucine by over-expression of ilvA in a Corynebacterium lactofermentum threonine producer. Appl Microbiol Biotechnol 1995, 43:482-488. Over-production of isoleucine in a threonine-producing mutant of C./actofermentum (described in [25]) is achieved by over-expressing the gene for the first enzyme in the threonine to isoleucine pathway. 80o/0 of the carbon available for threonine is redirected to isoleucine biosynthesis. Isoleucine production up to concentrations of 15 g I-1 in a shake-flask fermentation is reported. 29.

FrenchCE, Hailes AM, Rathbone DA, Long MT, Willey DL, Bruce NC: Biological production of semisynthetic opiates using genetically engineered bacteria. Biotechnology 1995, 13:674-676.

30.

Tong IT, Cameron DC: Enhancement of 1,3-propanediol production by cofermentation in Escherichie coil expressing Klebsiella pneumoniae dha regulon genes. Appl Biochem Biotechnol 1992, 34/35:149-159.

31.

Tong IT, Liao HH, Cameron DC: 1,3-Propanediol production by Escherichia coil expressing genes from the Klebsiella pneumoniae dha regulon. Appl Environ Microbiol 1991, 57:3541-3546.

204

Biochemical engineering

32.

Jackson DE, Srienc F: Novel methods to synthesize polyhydroxyalkanoate¢ Ann NY Acad Sci 1994, 745:134-148.

33. •

Zhang H, Obias V, Gonyer K, Dennis D: Production of polyhydroxyalkanoates in sucrose-utilizing recombinant Escherichia coil and Klebsiella strains. Appl Environ Microbiol 1994, 60:1198-1205. Sucrose-utilizing E. coil and Klebsiella strains are modified to produce PHB from molasses, a cheap carbon source, by introducing the PHB pathway from A. eutrophus. The Klebsiel/a strains perform better, however, and K. aerogenes is able to produce 24gl -1 PHB within 32h. K. oxytoca is reported to produce the co-polymer P(3HB-co-3HV) when propionate is added to the medium. This investigation takes advantage of a major strength of polymer production in bacterial hosts by utilizing cheaper alternative carbon sources as substrates. 34.

35.

Rhie HG, Dennis D: Role of fadR and atoC(Con) mutations In poly(3-hydroxybutyrate-co-3-hydroxyvalerate)synthesis in recombinant pha + Eschenchie coil Appl Environ Microbiol 1995, 61:2487-2492. Poirier Y, Nawrath 'qC,Somerville C: Production of polyhydroxyalkanoates, a family of biodegradable plastics and elastomers, In bacteria and plants. Biotechnolog.y 1995, 13:142-150.

36.

Van der Leij FR, Witholt B: Strategies for the sustainable production of new biodegradable polyesters In plants: a review. Can J Microbio/1995, 41:222-238.

37 *,

Nawrath C, Poirier Y, Somerville C: Targeting of the polyhydroxybutyrate biosynthetic pathway to the plasUds of Arabidopsis thaliane results in high levels of polymer accumulation. Proc Nat/Acad Sci USA 1994, 91:12760-12764. PHB is produced in high amounts in A. the~lenaby targeting the PHB biosynthetic pathway from A. eutrophus to the plastids. The investigators successfully hypothesize that the localization of PHB production to a compartment with a high acetyI-CoA flux would not adversely affect the growth of the plant. PHB accumulates to levels up to 14O/oof the dry weight of the plant. 38.

Poirier Y, Dennis DE, Klomparens K, Somerville CR: Polyhydroxybutyrate, a biodegradable thermoplastic, produced in transgenic plants. Science 1992, 256:520-523.

39.

Jeffries TW, Kurtzmann CP: Strain selection, taxonomy, and genetics of xylose-fermentlng yeasts. Enzyme Microb Techno/ 1994, 16:922-932.

40.

Hahn-Hagerdal B, Jeppsson H, Skoog K, Prior BA: Biochemistry and physiology of xylose fermentation by yeasts. Enzyme Microb Techno/1994, 16:933-943.