Synthesis of optically active amino acids from alpha-keto acids with Escherichia coli cells expressing heterologous genes.

We describe a simple method for enzymatic synthesis of L and D amino acids from alpha-keto acids with Escherichia coli cells which express heterologous genes. L-amino acids were produced with thermostable L-amino acid dehydrogenase and formate dehydrogenase (FDH) from alpha-keto acids and ammonium formate with only an intracellular pool of NAD+ for the regeneration of NADH. We constructed plasmids containing, in addition to the FDH gene, the genes for amino acid dehydrogenases, including i.e., leucine dehydrogenase, alanine dehydrogenase, and phenylalanine dehydrogenase. L-Leucine, L-valine, L-norvaline, L-methionine, L-phenylalanine, and L-tyrosine were synthesized with the recombinant E. coli cells with high chemical yields (> 80%) and high optical yields (up to 100% enantiomeric excess). Stereospecific conversion of various alpha-keto acids to D amino acids was also examined with recombinant E. coli cells containing a plasmid coding for the four heterologous genes of the thermostable enzymes D-amino acid aminotransferase, alanine racemase, L-alanine dehydrogenase, and FDH. Optically pure D enantiomers of glutamate and leucine were obtained.     

Production of aromatic d-amino acids from α-keto acids and ammonia by coupling of four enzyme reactions

A multi-enzyme system composed of glutamate racemase, thermostable d-amino acid aminotransferase, glutamate dehydrogenase and formate dehydrogenase was employed for the production of aromatic d-amino acids, d-phenylalanine and d-tyrosine, from the corresponding α-keto acids, phenylpyruvate and hydroxyphenylpyruvate, respectively. The optimal concentration of ammonium formate for the production of these d-amino acids was found in the range of 0.25–1.0 M. The optimal concentration of α-keto acid was determined to be 50 mM, above which the productivity greatly decreased. To keep the concentration of α-keto acid around this concentration, α-keto acid was intermittently fed into the multi-enzyme system during the production period. By running the multi-enzyme system for 35 h, 48 g l⁻¹ of d-phenylalanine and 60 g l⁻¹ of d-tyrosine were produced with 100% of optical purity from the equimolar amounts of phenylpyruvate and hydroxyphenylpyruvate, respectively. The production levels of both aromatic d-amino acids were demonstrated to be dependent on the stability of glutamate racemase.     

Occurrence of D-amino acid aminotransferase in pea seedlings

Aminotransferase reacting between D-amino acids and α-keto acids was found in the germinating pea seedlings. With partial purification, it was evident that the enzyme preparation contained no racemase and L-amino acid aminotransferase and was only specific for D-amino acid transamination. Large amounts of D-alanine found in the germinating pea seedlings were assumed to be formed by this enzyme action.     

Interaction between glutamate dehydrogenase (GDH) and L-leucine catabolic enzymes: intersecting metabolic pathways

Branched-chain amino acids (BCAAs) catabolism follows sequential reactions and their metabolites intersect with other metabolic pathways. The initial enzymes in BCAA metabolism, the mitochondrial branched-chain aminotransferase (BCATm), which deaminates the BCAAs to branched-chain α-keto acids (BCKAs); and the branched-chain α-keto acid dehydrogenase enzyme complex (BCKDC), which oxidatively decarboxylates the BCKAs, are organized in a supramolecular complex termed metabolon. Glutamate dehydrogenase (GDH1) is found in the metabolon in rat tissues. Bovine GDH1 binds to the pyridoxamine 5′-phosphate (PMP)-form of human BCATm (PMP-BCATm) but not to pyridoxal 5′-phosphate (PLP)-BCATm in vitro. This protein interaction facilitates reamination of the α-ketoglutarate (αKG) product of the GDH1 oxidative deamination reaction. Human GDH1 appears to act like bovine GDH1 but human GDH2 does not show the same enhancement of BCKDC enzyme activities. Another metabolic enzyme is also found in the metabolon is pyruvate carboxylase (PC). Kinetic results suggest that PC binds to the E1 decarboxylase of BCKDC but does not effect BCAA catabolism. The protein interaction of BCATm and GDH1 promotes regeneration of PLP-BCATm which then binds to BCKDC resulting in channeling of the BCKA products from BCATm first half reaction to E1 and promoting BCAA oxidation and net nitrogen transfer from BCAAs. The cycling of nitrogen through glutamate via the actions of BCATm and GDH1 releases free ammonia. Formation of ammonia may be important for astrocyte glutamine synthesis in the central nervous system. In peripheral tissue association of BCATm and GDH1 would promote BCAA oxidation at physiologically relevant BCAA concentrations.     

The formation of ornithine from proline in animal tissues

1. Homogenates of liver or kidney from rat, mouse, dog and guinea pig formed ornithine from proline but not from glutamate. Rat kidney was most active in this reaction and was used for further studies. 2. The overall reaction was found to be catalysed by proline oxidase to yield glutamic gamma-semialdehyde, followed by transamination of this product with glutamate as catalysed by ornithine-keto acid aminotransferase. 3. The unfavourable equilibrium of the ornithine-keto acid aminotransferase reaction was overcome chiefly by glutamate dehydrogenase in the tissue, which removed the alpha-oxoglutarate produced, by reduction with endogenous ammonia and NADH. 4. Aspartate aminotransferase in these preparations also aided in the removal of alpha-oxoglutarate. In this case the overall reaction was driven also by the rapid decarboxylation of oxaloacetate. 5. No evidence could be found for a pathway of ornithine synthesis involving acylated intermediates as has been observed in some micro-organisms. 6. The rate of ornithine synthesis in homogenates of several rat tissues paralleled the activity of ornithine-keto acid aminotransferase in these tissues, indicating that this enzyme was rate-determining for the synthesis. 7. The possible influence of these reactions on urea synthesis is discussed.     

Inhibition of glycine synthase by branched-chain α-keto acids: A possible mechanism for abnormal glycine metabolism in ketotic hyperglycinemia

The effect of various amino acid metabolites on glycine oxidation by rat liver homogenate was investigated. Three compounds, α-ketoisovaleric acid, α-ketoisocaproic acid, and α-keto-β-methylvaleric acid, were found to inhibit glycine oxidation by 40–60%. In addition, these compounds also inhibited the glycine-CO₂ exchange reaction, a partial reaction of glycine synthase. The reverse reaction, glycine synthesis, was stimulated 4-fold by these α-keto acids. Pyruvate and α-ketoglutarate had no effect on any of these reactions. The parent amino acids, valine, isoleucine, and leucine, also had no effect on the reactions nor did any of their other metabolites with the exception of the branched-chain α-keto acids. The concentration dependence of the inhibition of glycine oxidation and stimulation of glycine synthesis by these branched-chain α-keto acids suggested that the inhibition of glycine oxidation by these compounds was the result of their further oxidation by branched-chain α-keto acid dehydrogenase. However, the products of the branched-chain α-keto acid dehydrogenase, isobutyryl CoA, isovaleryl CoA, or α-methylbutyryl CoA had no effect on glycine oxidation. Thus, it appeared that either the branched-chain α-keto acids altered glycine oxidation by direct binding to glycine synthase or that electrons derived from the oxidation of branched-chain α-keto acids were transferred to the glycine synthase system. It is proposed that glycine synthase and branched-chain α-keto acid dehydrogenase either share a common subunit, possibly lipoamide dehydrogenase, or are so arranged on the mitochondrial membrane that electron transfer between these two enzymes occurs.     

Desilicification from the High Silica Containing Kraft Black Liquor by the Improved Carbon Dioxide Method

Addition of NADH to crude but not to pure branched-chain α-keto acid decarboxylase decreased the CO₂ production from α-keto-β-methylvalerate (KMV) suggesting the presence of an NADH dependent inhibitor in the crude enzyme from Bacillus subtilis. This NADH-dependent decarboxylase inhibitor was purified to homogeneity by a fast protein liquid chromatography system.

The purified inhibitor was identical with leucine dehydrogenase as to N-terminal amino acid squence (35 residues) and molecular weight, and catalyzed the oxidative deamination of three branched chain amino acids (BCAAs), valine, leucine, and isoleucine. The decarboxylase inhibitor was therefore identified as leucine dehydrogenase. A decreased substrate availability caused by leucine dehydrogenase thus reasonably accounted for the NADH dependent inhibition of the decarboxylation. In turn, the observation that leucine dehydrogenase competes with the decarboxylase for branched-chain α-keto acid (BCKA) suggested an involvement of this enzyme in the branched chain fatty acid (BCFA) biosynthesis. This view was supported by the observation that addition of NAD to crude fatty acid synthetase increased the incorporation of isoleucine into BCFAs. Pyridoxal-5′-phosphate and α-ketoglutarate, cofactors for BCAA transaminase, modulated BCFA biosynthesis from isoleucine in vitro, suggesting also the involvement of transaminase reaction in BCFA biosynthesis.     

Biotransformation of D-methionine into L-methionine in the cascade of four enzymes

D-Methionine was converted to L-methionine in a reaction system where four enzymes were used. D-amino acid oxidase (D-AAO) from Arthrobacter protophormiae was used for the complete conversion of D-methionine to 2-oxo-4-methylthiobutyric acid. Catalase was added to prevent 2-oxo-4-methylthiobutyric acid decarboxylation. In the second reaction step, L-phenylalanine dehydrogenase (L-PheDH) from Rhodococcus sp. was used to convert 2- oxo-4-methylthiobutyric acid to L-methionine, and formate dehydrogenase (FDH) from Candida boidinii was added for NADH regeneration. Enzyme kinetics of all enzymes was analyzed in detail. Mathematical models for separate reactions steps, as well as for the complete system were developed and validated in the batch reactor experiments. Complete conversion of D-methionine to L-methionine was achieved. Considering that both enzymes act on different substrates, such a system could be easily employed for the synthesis of other amino acids from D-isomer, as well as from the racemate of a certain amino acid (DL-amino acid).     

Photorespiratory nitrogen cycle – A critical evaluation

The photorespiratory nitrogen cycle proposed by Keys et al. (Nature 275: 741–743, 1978) involved formation of glycine by transamination of glyoxylate in the peroxisomes utilizing glutamate. Subsequently, glycine is oxidized to ammonia, serine and CO₂ in the mitochondria. The ammonia is reassimilated via the GS/GOGAT pathway generating glutamate. In this article, experimental evidence which suggests the occurrence of alternative mechanisms of glycolate and serine synthesis as well as of CO₂ and ammonia evolution is discussed. The problem of utilization of NADH coupled to ATP synthesis during photosynthesis is still unresolved, which complicates the glycine oxidation reaction in light. Further, factors are presented that determine the availability of amino donors in the peroxisomes and of amino acids viz., glycine, serine and glutamate for the operation of the photorespiratory N cycle. Recent evidence regarding the role of formate arising out of the reaction of glyoxylate with H₂O₂ in the regulation of photosynthetic electron flow in the Hill reaction, as well as of photorespiratory substrates functioning as carbon sources for the citric acid cycle in the light or for export to the growing tissues, suggests that the role of photo-respiration in plant metabolism needs to be reexamined.     

D-Amino acid metabolism in mammals: biosynthesis, degradation and analytical aspects of the metabolic study

It was believed for long time that d-amino acids are not present in mammals. However, current technological advances and improvements in analytical instruments have enabled studies that now indicate that significant amounts of D-amino acids are present in mammals. The most abundant D-amino acids are D-serine and D-aspartate. D-Serine, which is synthesized by serine racemase and is degraded by D-amino-acid oxidase, is present in the brain and modulates neurotransmission. D-Aspartate, which is synthesized by aspartate racemase and degraded by D-aspartate oxidase, is present in the neuroendocrine and endocrine tissues and testis. It regulates the synthesis and secretion of hormones and spermatogenesis. D-Serine and D-aspartate bind to the N-methyl-D-aspartate (NMDA) subtype of glutamate receptors and function as a coagonist and agonist, respectively. The enzymes that are involved in the synthesis and degradation of these D-amino acids are associated with neural diseases where the NMDA receptors are involved. Knockout mice for serine racemase and D-aspartate oxidase have been generated, and natural mutations in the d-amino-acid oxidase gene are present in mice and rats. These mutant animals display altered behaviors caused by enhanced or decreased NMDA receptor activity. In this article, we review currently available studies on D-amino acid metabolism in mammals and discuss analytical methods used to assay activity of amino acid racemases and D-amino-acid oxidases.     

The Role of Glutamine Synthetase and Glutamate Dehydrogenase in Cerebral Ammonia Homeostasis

In the brain, glutamine synthetase (GS), which is located predominantly in astrocytes, is largely responsible for the removal of both blood-derived and metabolically generated ammonia. Thus, studies with [¹³N]ammonia have shown that about 25?% of blood-derived ammonia is removed in a single pass through the rat brain and that this ammonia is incorporated primarily into glutamine (amide) in astrocytes. Major pathways for cerebral ammonia generation include the glutaminase reaction and the glutamate dehydrogenase (GDH) reaction. The equilibrium position of the GDH-catalyzed reaction in vitro favors reductive amination of α-ketoglutarate at pH 7.4. Nevertheless, only a small amount of label derived from [¹³N]ammonia in rat brain is incorporated into glutamate and the α-amine of glutamine in vivo. Most likely the cerebral GDH reaction is drawn normally in the direction of glutamate oxidation (ammonia production) by rapid removal of ammonia as glutamine. Linkage of glutamate/α-ketoglutarate-utilizing aminotransferases with the GDH reaction channels excess amino acid nitrogen toward ammonia for glutamine synthesis. At high ammonia levels and/or when GS is inhibited the GDH reaction coupled with glutamate/α-ketoglutarate-linked aminotransferases may, however, promote the flow of ammonia nitrogen toward synthesis of amino acids. Preliminary evidence suggests an important role for the purine nucleotide cycle (PNC) as an additional source of ammonia in neurons (Net reaction: l-Aspartate?+?GTP?+?H₂O?→?Fumarate?+?GDP?+?P_i?+?NH₃) and in the beat cycle of ependyma cilia. The link of the PNC to aminotransferases and GDH/GS and its role in cerebral nitrogen metabolism under both normal and pathological (e.g. hyperammonemic encephalopathy) conditions should be a productive area for future research.     

Enzymatic in situ determination of stereospecificity of NAD-dependent dehydrogenases

Amino acid racemases inherently catalyze the exchange of alpha-hydrogen of amino acids with deuterium during racemization in 2H2O. When the reactions catalyzed by alanine racemase (EC 5.1.1.1) and L-alanine dehydrogenase (EC 1.4.1.1), which is pro-R specific for the C-4 hydrogen transfer of NADH, are coupled in 2H2O, [4R-2H]NADH is exclusively produced. Similarly, [4S-2H]NADH is made in 2H2O with amino-acid racemase with low substrate specificity (EC 5.1.1.10) and L-leucine dehydrogenase (EC 1.4.1.9), which is pro-S specific. We have established a simple procedure for the in situ analysis of stereospecificity of C-4 hydrogen transfer of NADH by an NAD-dependent dehydrogenase by combination with either of the above two couples of enzymes in the same reaction mixture. When the C-4 hydrogen of NAD+ is fully retained after sufficient incubation, the stereospecificity of hydrogen transfer by a dehydrogenase is the same as that of alanine dehydrogenase or leucine dehydrogenase. However, when the C-4 hydrogen of NAD+ is exchanged with deuterium, the enzyme to be examined shows the different stereospecificity from alanine dehydrogenase or leucine dehydrogenase. Thus, we can readily determine the stereospecificity by 1H NMR measurement without isolation of the coenzymes and products.     

Role of Branched-Chain Aminotransferase Isoenzymes and Gabapentin in Neurotransmitter Metabolism

Abstract: Because it is well known that excess branched-chain amino acids (BCAAs) have a profound influence on neurological function, studies were conducted to determine the impact of BCAAs on neuronal and astrocytic metabolism and on trafficking between neurons and astrocytes. The first step in the metabolism of BCAAs is transamination with α-ketoglutarate to form the branched-chain α-keto acids (BCKAs). The brain is unique in that it expresses two separate branched-chain aminotransferase (BCAT) isoenzymes. One is the common peripheral form [mitochondrial (BCATm)], and the other [cytosolic (BCATc)] is unique to cerebral tissue, placenta, and ovaries. Therefore, attempts were made to define the isoenzymes' spatial distribution and whether they might play separate metabolic roles. Studies were conducted on primary rat brain cell cultures enriched in either astroglia or neurons. The data show that over time BCATm becomes the predominant isoenzyme in astrocyte cultures and that BCATc is prominent in early neuronal cultures. The data also show that gabapentin, a structural analogue of leucine with anticonvulsant properties, is a competitive inhibitor of BCATc but that it does not inhibit BCATm. Metabolic studies indicated that BCAAs promote the efflux of glutamine from astrocytes and that gabapentin can replace leucine as an exchange substrate. Studying astrocyte-enriched cultures in the presence of [U-¹⁴C]glutamate we found that BCKAs, but not BCAAs, stimulate glutamate transamination to α-ketoglutarate and thus irreversible decarboxylation of glutamate to pyruvate and lactate, thereby promoting glutamate oxidative breakdown. Oxidation of glutamate appeared to be largely dependent on the presence of an α-keto acid acceptor for transamination in astrocyte cultures and independent of astrocytic glutamate dehydrogenase activity. The data are discussed in terms of a putative BCAA/BCKA shuttle, where BCATs and BCAAs provide the amino group for glutamate synthesis from α-ketoglutarate via BCATm in astrocytes and thereby promote glutamine transfer to neurons, whereas BCATc reaminates the amino acids in neurons for another cycle.     

Oxidation of NADH in Glyoxysomes by a Malate-Aspartate Shuttle

Glyoxysomes isolated from germinating castor bean endosperm accumulate NADH by β-oxidation of fatty acids. By utilizing the glutamate: oxaloacetate aminotransferase and malate dehydrogenase present in glyoxysomes and mitochondria, reducing equivalents could be transferred between the organelles by a malate-aspartate shuttle. The addition of aspartate plus α-ketoglutarate to purified glyoxysomes brought about a rapid oxidation of accumulated NADH, and the oxidation was prevented by aminooxyacetate, an inhibitor of aminotransferase activity. Citrate synthetase activity in purified glyoxysomes could be coupled readily to glutamate: oxaloacetate aminotransferase activity as a source of oxaloacetate, but coupling to malate dehydrogenase and malate resulted in low rates of citrate formation. Glyoxysomes purified in sucrose or Percoll gradients were permeable to low molecular weight compounds. No evidence was obtained for specific transport mechanisms for the proposed shuttle intermediates. The results support a revised model of gluconeogenic metabolism incorporating a malate-aspartate shuttle in the glyoxysomal pathway.     

The potential use of mechanism-based enzyme inactivators in medicine

Mechanism-based enzyme inactivator, alanine racemase, S-adenosylhomocysteine hydrolase, D-amino acid aminotransferase, gamma-aminobutyric acid aminotransferase, arginine decarboxylase, aromatase, L-aromatic amino acid decarboxylase, dihydrofolate reductase, dihydroorotate dehydrogenase DNA polymerase I, dopamine beta-hydroxylase, histidine decarboxylase, beta-lactamase, monoamine oxidase, ornithine decarboxylase, serine proteases, testosterone 5 alpha-reductase, thymidylate synthetase, xanthine oxidase.     

Continuous production of 3-fluoro-L-alanine with alanine dehydrogenase

Alanine dehydrogenase catalyzed the conversion of 3-fluoropyruvate into 3-fluoro-L-alanine in the presence of NADH and ammonia. The optimum pH of the reaction was 7.8. The K(m) values of the enzyme for 3-fluoropyruvate, polyethylene glycol-bound NADH, and ammonia were 2.94, 0.56, and 105mM, respectively. 3-Fluoro-L-alanine was selectively and continuously produced from 3-fluoropyruvate and ammonium formate in an enzyme membrane reactor by the multienzyme reaction system of alanine dehydrogenase and formate dehydrogenase with a simultaneous coenzyme regeneration. The average conversion and the space-time yield were 73% and 75 g/L day, respectively, with operation of the reactor for 4 days. Alanine dehydrogenase and formate dehydrogenase consumed were 11, 370 and 22, 950 units/kg 3-fluoro-L-alanine, respectively. The cycle number was 3150 mol/mol NAD.     

Simultaneous determination of D-amino acids by the coupling method of D-amino acid oxidase with high-performance liquid chromatography

An enzymatic assay system of D-amino acids was established using the D-amino acid oxidase of Schizosaccharomyces pombe. In this method, the enzyme converts the D-amino acids to the corresponding α-keto acids, which are then reacted with 1,2-diamino-4,5-methylenedioxybenzene (DMB) in an organic solvent. The resultant fluorescent compounds are separated and quantified by high-performance liquid chromatography (HPLC). Use of an organic solvent following the α-keto acid modification with DMB prevents the non-enzymatic deamination of L-amino acids, which are generally present at much higher concentrations than D-amino acids in biological samples. With this method, D-Glu, D-Asn, D-Gln, D-Ala, D-Val, D-Leu, D-Phe, and D-Ile can be quantified in the order of micromolar, and other D-amino acids except D-Asp can be assayed within a sensitivity range of 50-100 μM. The established enzymatic method was used to analyze the d-amino acid contents in human urine. The concentration of D-Ser obtained using this enzymatic method (223 μM) was in good agreement with that obtained using the conventional HPLC method (198 μM). The enzymatic method also demonstrated that the human urine contained 5.45 μM of d-Ala and 0.91 μM of D-Asn. Both D-amino acids were difficult to be identified using the conventional method, because the large signals from L-amino acids masked those from d-amino acids. The enzymatic method that we have developed can circumvent this problem.     

Enzymatic reduction of dehydrocholic acid to 12-ketochenodeoxycholic acid with NADH regeneration

12-Ketochenodeoxycholic acid, an essential intermediate in the synthesis of chenodeoxycholic acid, has been enzymatically prepared from dehydrocholic acid. The specific reduction of dehydrocholic with NADH was catalysed by 3α-hydroxysteroid dehydrogenase (3α-hydroxysteroid: NAD(P)⁺ oxidoreductase, EC 1.1.1.50) and 7α-hydroxysteroid dehydrogenase (7α-hydroxysteroid:NAD⁺ 7-oxidoreductase, EC 1.1.1.159). Cofactor regeneration was obtained through the formate dehydrogenase (formate:NAD⁺ oxidoreductase, EC 1.2.1.2) catalysed oxidation of formate. Complete transformation of dehydrocholic acid to the 12-keto derivative was achieved with a coenzyme turnover number up to 1200. No steroid by-products were detected by high performance liquid chromatography and thin layer chromatography. The process yielded 9 g product l^?1 in 66–84 h. The high purity of the enzymatically prepared 12-ketochenodeoxycholic acid should drastically reduce the formation of the toxic by-product lithocholic acid, which occurs in the synthesis of chenodeoxycholic acid when using chemical methods alone.     

Effect of reducing-equivalent disposal and NADH/NAD on deamination of amino acids by intact rumen microorganisms and their cell extracts.

When mixed rumen microorganisms were incubated in media containing the amino acid source Trypticase, both monensin and carbon monoxide (a hydrogenase inhibitor) decreased methane formation and amino acid fermentation. Both of the methane inhibitors caused a significant increase in the ratio of intracellular NADH to NAD. Studies with cell extracts of rumen bacteria and protozoa indicated that the ratio of NADH to NAD had a marked effect on the deamination of reduced amino acids, in particular branched-chain amino acids. Deamination was inhibited by the addition of NADH and was stimulated by methylene blue, an agent that oxidizes NADH. Neutral and oxidized amino acids were unaffected by NADH. The addition of small amounts of 2-oxoglutarate greatly enhanced the deamination of branched-chain amino acids and indicated that transamination via glutamate dehydrogenase was important. Formation of ammonia from glutamate was likewise inhibited by NADH. These experiments indicated that reducing-equivalent disposal and intracellular NADH/NAD ratio were important effectors of branched-chain amino acid fermentation.     

Effect of reducing-equivalent disposal and NADH/NAD on deamination of amino acids by intact rumen microorganisms and their cell extracts

When mixed rumen microorganisms were incubated in media containing the amino acid source Trypticase, both monensin and carbon monoxide (a hydrogenase inhibitor) decreased methane formation and amino acid fermentation. Both of the methane inhibitors caused a significant increase in the ratio of intracellular NADH to NAD. Studies with cell extracts of rumen bacteria and protozoa indicated that the ratio of NADH to NAD had a marked effect on the deamination of reduced amino acids, in particular branched-chain amino acids. Deamination was inhibited by the addition of NADH and was stimulated by methylene blue, an agent that oxidizes NADH. Neutral and oxidized amino acids were unaffected by NADH. The addition of small amounts of 2-oxoglutarate greatly enhanced the deamination of branched-chain amino acids and indicated that transamination via glutamate dehydrogenase was important. Formation of ammonia from glutamate was likewise inhibited by NADH. These experiments indicated that reducing-equivalent disposal and intracellular NADH/NAD ratio were important effectors of branched-chain amino acid fermentation.