Since microbial production of L-glutamic acid was started in 1957 in Japan, various amino acids production with microorganisms has been developed and almost all protein-constitutive amino acids become able to be produced by microbial biotechnology, fermentation, or enzymatic method. This chapter summarizes the microbial biotechnology which was developed and industrialized in Japan. The amino acids include L-alanine, L-cysteine, L-DOPA, L-glutamic acid, D-p-hydroxyphenylglycine, hydroxy-L-proline, L-lysine and L-threonine.
The water extract of a marine algae “kelp” (Laminaria japonica) has been used in a number of Japanese recipes as a kind of soup to flavor cooking. The tasty (“umami” in Japanese) factor in the marine algae was identified as monosodium glutamate by Prof. Kikunae Ikeda in 1908. Soon thereafter, Ajinomoto Co. Ltd. started to produce monosodium glutamate by extraction from wheat protein after hydrolysis with hydrochloric acid.
In 1956, Kyowa Hakko Kogyo Co. Ltd. succeeded in producing sodium glutamate by using a bacterium (Corynebacterium glutamicum). Then, production using microbial methods of various amino acids (including of l-alanine, l-aspartic acid, l-arginine, l-citrulline, l-cysteine, l-DOPA [3,4-dihydroxyphenylalanine], l-glutamic acid, l-glutamine, glutathione, l-histidine, d-hydroxyphenylglycine, l-hydroxyproline, l-isoleucine, l-lysine, l-ornithine, l-phenylalanine, d-phenylglycine, l-polylysine, l-proline, l-serine, l-threonine, l-tryptophan, and l-tyrosine) was investigated and successfully manufactured on an industrial scale. Glycine is produced by chemical methods because the molecule has no chiral center, and methionine is also produced by chemical methods in its racemic form because the main use of the amino acid is as feedstuff. d-Methionine is metabolized in animals by the action of d-amino acid oxidase. These amino acids were useful as sources of medicines, food additives, feedstuffs, and starting materials for chemical synthesis.
This chapter describes the microbial production of some amino acids including the producing strains, production method, product usage, and industrial production.
The annual world production of l-alanine is about 500 t. This amino acid is useful as an enteral and parenteral nutrient and as a food additive, which has a sweet taste and bacteriostatic properties.
l-Alanine production by fermentation is difficult because bacteria usually have an alanine racemase to racemize the product. Fermentative production of l-alanine with racemase-deficient strains of Corynebacterium glutamicum, Brevibacterium flavum, and Arthrobacter oxydans (Hashimoto and Katsumata 1999) has been investigated, and good yields have been reported, although the method is not yet industrially applicable.
Tanabe Seiyaku Co. in Japan applied on an industrial scale the enzymatic method described above, using aspartic acid produced by immobilized enzymes as the starting material. A bacterial strain selected for its strong aspartate β-decarboxylase activity was identified as Pseudomonas dacunhae (Chibata et al. 1965). The bacterial cells with high activity were immobilized on κ-carrageenan, a polysaccharide obtained from seaweed, and packed in a column. The l-alanine was produced by this column system continuously. To prevent the evolution of carbon dioxide gas, a closed column reactor was designed and used for the production. In this column, the enzyme reaction proceeds under high pressure (Chibata et al. 1986a).
The substrate, l-aspartate, is produced from fumarate by an enzyme system involving aspartase, as described in the section on l-aspartate. To produce l-alanine directly from fumarate, the l-alanine-producing column was connected in tandem to an l-aspartate-producing column. In this tandem column system, side reactions caused by fumarase in Escherichia coli and alanine racemase in P. dacunhae reduced the yield. Then, both bacterial cells were separately treated with high temperature and low pH, respectively, and the enzymes responsible for the side reactions were inactivated. Immobilization of these two kinds of bacterial cells with κ-carrageenan resulted in the production of l-alanine in a single reactor without the production of the side products, malate and d-alanine (Takamatsu et al. 1982; Chibata et al. 1986b).
The annual world production of l-aspartate is estimated to be 7,000 t. l-Aspartate is used as an enteral and parenteral nutrient, a food additive, and a starting material for the low-calorie sweetener aspartame (aspartylphenylalanine methyl ester). It is also used as a raw material to synthesize detergent and for chelating or water treatment agents.
l-Aspartate is produced by the reaction of fumarate and ammonia catalyzed by aspartase.
l-Aspartate production began in 1960 using a batchwise process involving E. coli cells containing high aspartase activity. In 1973, Chibata and collaborators at the Tanabe Seiyaku Co. started producing l-aspartate using a continuous reaction system consisting of an immobilized enzyme column. In the system, aspartase extracted from E. coli cells was immobilized on ion exchange resin. Escherichia coli cells were immobilized by trapping in acrylamide gel, and then, the column was used in industrial production (Tosa et al. 1973). In 1978, this matrix was changed to κ-carrageenan. The production of l-aspartate was greatly improved by this method, and the yield became 100 t/month using a 1 kl bioreactor (Chibata et al. 1986b). In the United States, E. coli cells with high aspartase activity immobilized on polyurethane and polyazetidine were reported, and the latter was shown to have high aspartase activity, producing aspartate at the rate of 55.9 mol/h/kg cell (wet weight; Fusee et al. 1981).
Annual world production of l-cysteine is 1,500 t. Its uses are as an enteral and parenteral nutrient, food additive, constituent of hair treatment preparations, and starting material of constituents used in cosmetics.
S-Carboxymethyl-l-cysteine is also produced by the same enzymatic method from the corresponding starting material (Yokozeki et al. 1988).
In screening for high-yield producers, the bacterial strain that produced the most l-cysteine from DL-ATC was isolated from soil and designated “Pseudomonas thiazolinophilum.” The enzymes responsible for the conversion were inducible, and the addition of DL-ATC to the culture medium was essential for enzyme activity. Addition of Mn+2 and Fe+2 to the medium also contributed to increasing the enzyme activity. The reaction proceeds by adding cells with high enzyme activities but no cysteine-desulfhydrase (an l-cysteine-degrading enzyme) to the reaction mixture containing DL-ATC. l-Cysteine produced in the reaction mixture is oxidized to L-cystine by aeration and precipitated as crystals. This increases the efficiency of l-cysteine production, which is 31.4 g/l obtained from 40 g/l of DL-ATC, i.e., 95 % product yield by molar ratio.
The annual world production of l-DOPA is around 250 t. l-DOPA (the precursor of the neurotransmitter dopamine) is useful as a treatment for Parkinson’s disease. It had been mainly produced by a chemical synthetic method that included eight reaction steps including an optical resolution step.
l-DOPA is produced from pyrocatechol, pyruvate, and ammonia by a one-step enzyme reaction using tyrosine phenol-lyase (TPL). Ajinomoto Co. Ltd. began using Erwinia TPL for enzymatic l-DOPA production (by a simple one-step method and one of the most economical processes to date) in 1993.
Erwinia herbicola was selected as the most favorable strain for the l-DOPA production out of 1,041 microbial strains tested. No enzyme activity was found in yeasts, fungi, and actinomycetes. Culture conditions for the preparation of cells containing high TPL activity and reaction conditions for the synthesis of l-DOPA were optimized with Erwinia herbicola. Additions of yeast extract, meat extract, polypeptone, and the hydrolyzate of soybean protein to the basal medium enhanced cell growth as well as the formation of TPL. Catabolite repression of biosynthesis of TPL was observed on adding glucose, pyruvate, and α-ketoglutarate to the medium at high concentrations. Glycerol was a suitable carbon source for cell growth as well as for the accumulation of the enzyme in growing cells. TPL is an inducible enzyme, and the addition of l-tyrosine to the medium is essential for formation of the enzyme. l-Phenylalanine is not an inducer of TPL biosynthesis but works as a synergistic agent for the induction by l-tyrosine. The activity of TPL increased five times by the addition of l-phenylalanine together with l-tyrosine to the medium. Cells of E. herbicola with high TPL activity were prepared by growing them at 28 °C for 28 h in a medium containing 0.2 % KH2PO4, 0.1 % MgSO4·7H2O, 2 ppm of Fe+2 (FeSO4·7H2O), 0.01 % pyridoxine-HCl, 0.6 % glycerol, 0.5 % succinic acid, 0.1 % DL-methionine, 0.2 % DL-alanine, 0.05 % glycine, 0.1 % l-phenylalanine, and 12 ml of hydrolyzed soybean protein in 100 ml of tap water, with the pH controlled at 7.5 throughout cultivation. Under these conditions, TPL was efficiently accumulated in the cells of E. herbicola and made up about 10 % of the total soluble cellular protein (Yamada and Kumagai 1975).
The enzymatic synthesis reaction of l-DOPA is carried out in a batchwise system with cells of E. herbicola containing high activity of TPL. Since pyruvate, one of the substrates, was unstable in the reaction mixture at high temperature, low temperature was used for the synthesis of l-DOPA. The reaction was carried out at 16 °C for 48 h in a reaction mixture containing various amounts of sodium pyruvate, 5 g of ammonium acetate, 0.6 g of pyrocatechol, 0.2 g of sodium sulfite, 0.1 g of EDTA, and cells harvested from 100 ml of broth in a total volume of 100 ml. The pH was adjusted to 8.0 by the addition of ammonia. At 2-h intervals, sodium pyruvate and pyrocatechol were added to the reaction mixture to maintain the initial concentrations. The maximum synthesis of l-DOPA was obtained when the concentration of sodium pyruvate was kept at 0.5 %. The addition of substrates, pyrocatechol and pyruvate, was separated by a time interval to prevent the denaturation of TPL and to prevent by-product formation. Sodium sulfite is added to keep the reactor in a reductive state and to prevent the oxidation of product l-DOPA. The l-DOPA is not soluble in the reaction medium, so it forms a crystalline precipitate (reaching 110 g/l) during the reaction (Yamada and Kumagai 1975; Kumagai 1999a, b).
Induction and repression mechanisms of TPL in E. herbicola were studied. It was found that TPL biosynthesis is regulated at the transcriptional level. Tyrosine phenol-lyase mRNA was increased by the addition of tyrosine and decreased by the addition of glucose in the medium. TyrR box and operator-like regions were found in the 5′ flanking region of its gene, tpl. TyrR box is a typical binding site on DNA where a regulator protein TyrR binds and controls transcription of the regulon of enzyme genes or transporter genes responsible for biosynthesis of aromatic amino acids or transport through cell membrane (Suzuki et al. 1995; Katayama et al. 1999). Katayama et al. reported three point mutations in the tyrR gene that caused high-level expression of lacZ in E. coli and tpl in E. herbicola (Katayama et al. 2000). The functions of the product of tutB gene and the gene itself, located just downstream of tpl in E. herbicola, were analyzed. It was elucidated that tutB encodes a tyrosine-specific transporter and this is essential for maximum induction of TPL in E. herbicola cells (Katayama et al. 2002).
World production of monosodium l-glutamate using so-called coryneform bacteria is around 1 million tons per annum. Monosodium glutamate is used as a flavor enhancer and an intermediate material for chemical synthesis of medicines. Its ester is used as a detergent and the polymer as artificial skin. Two Japanese companies, Ajinomoto and Kyowa Hakko Kogyo, built factories and produced it in other countries, mainly in Southeast Asia. China, Korea, and Taiwan also are producing large amounts of monosodium l-glutamate.
Glutamic acid is produced by Corynebacterium glutamicum in the presence of high concentrations of sugar and ammonium, appropriate concentrations of minerals, and a limited concentration of biotin under aerobic conditions (Kikuchi and Nakao 1986). In 2–3 days, around 100 g of l-glutamate per liter accumulates in the medium.
Various glutamic acid-producing strains were reported after the first report on Corynebacterium glutamicum. They are Brevibacterium flavum, Brevibacterium lactofermentum, Brevibacterium thiogenitalis, and Microbacterium ammoniaphilum, and all strains are Gram-positive, nonsporulating, nonmotile, or rodlike cocci and all require biotin for growth. Currently, these strains are thought to belong to the genus Corynebacterium.
The carbon source most commonly used as a starting material is glucose, which is obtained by enzymatic hydrolysis of starch from corn, potato, and cassava. Waste molasses is also used since it is cheap. Acetic acid and ethanol are also good carbon sources to produce glutamate. A high concentration of a nitrogen source is necessary to accumulate glutamate, and ammonia gas, its solution, the inorganic salt, or urea is used in actual production.
Coryneform bacteria generally show high rates of sugar assimilation and highly active glutamate dehydrogenase, which is responsible for glutamate biosynthesis. Glucose incorporated in the cell is metabolized through the Embden-Meyerhof pathway (EMP) and a part of the tricarboxylic acid (TCA) cycle, and 2-oxoglutarate formed in the cycle is aminated to glutamate by the action of glutamate dehydrogenase.
Biotin is an important factor regulating the growth of the bacterium and glutamic acid production. Its suboptimal addition is essential to produce a large amount of glutamic acid in the medium (Clement and Lanneelle 1986). To use a starting material such as waste molasses, which contains excess biotin, the addition of penicillin to the medium during growth was found to be effective. In the production of glutamic acid, several saturated fatty acids or their esters were also found to function similarly to penicillin. A glycerol-requiring mutant of Corynebacterium alkanolyticum was used to produce glutamic acid in appreciable amounts without the addition of penicillin and without the need to control biotin concentration (Kikuchi and Nakao 1986).
Since these treatments are essential for the glutamate fermentation, it has been suggested that the cell surface of the bacteria is damaged under such conditions, and consequently leaking of glutamate takes place. This leaking theory has been accepted for a long time. But recently another published theory of excretion of glutamate suggested an exporter protein of glutamate was present on the cell surface of the bacterium (Kraemer 1994).
The amount of 2-oxoglutarate dehydrogenase complex (ODHC), which catalyzes the conversion of 2-oxoglutarate to succinyl-CoA as the first step of succinate synthesis in the TCA cycle, was reported to be decreased in glutamate-producing bacterial cells. And recently, the enzyme activity was confirmed to be very low in the presence of detergent, or limited amounts of biotin or penicillin (Kawahara et al. 1997). These results suggest that one of the main causes for the glutamate overproduction is the decrease of 2-oxoglutarate dehydrogenase activity (ODH). A disrupted (ODH) gene-bearing bacterial strain produced as much glutamate as the wild-type strain under conditions of glutamate overproduction.
Furthermore, a novel gene dtsR was cloned, which rescues the detergent sensitivity of a mutant derived from a glutamate-producing bacterium, Corynebacterium glutamicum (Kimura et al. 1996). The authors found that this gene encodes a putative component of a biotin-containing enzyme complex and has something to do with fatty acid metabolism. The disruption of this gene causes constitutive production of glutamate even in the presence of excess biotin. The authors suggested that the overproduction of glutamate is caused by an unbalance of the coupling between fatty acid and glutamate synthesis (Kimura et al. 1997). They successfully showed that inducers of glutamate overproduction such as Tween 40 and limited amounts of biotin reduced the level of DtsR, which then triggered overproduction by decreasing the activity of ODHC (Kimura et al. 1999).
Kyowa Hakko Kogyo Co. Ltd., the Research Institute for Innovative Technology for Earth in Japan, and Degussa in Germany completed the analysis of the genomic DNA nucleotide sequence of Corynebacterium glutamicum.
Kaneka Co. Ltd. started the enzymatic production of d-p-hydroxyphenylglycine (d-HPG) in 1980 in Singapore, and the immobilized d-carbamoylase reactor was introduced in this process in 1995. The annual production of d-HPG by this method is around 2,000 t. d-HPG is a starting material for the production of semisynthetic penicillins and cephalosporins, such as amoxicillin and cefadroxil. d-HPG is produced from DL-p-hydroxyphenylhydantoin (DL-HPH) by a two-step enzymatic method (Takahashi 1986).
d-Hydantoin hydrolase activity was found in some bacteria belonging to the genera Bacillus, Pseudomonas, Aerobacter, Agrobacterium, and Corynebacterium and in actinomycetes belonging to the genera Streptomyces and Actinoplanes. d-Carbamylase activity was found in various bacteria belonging to the genera Agrobacterium, Pseudomonas, Comamonas, and Blastobacter. The genes of these two enzymes were cloned, and an E. coli strain transformed by this gene was used as the practical enzyme source. To obtain stable d-carbamoylase for repeated use, a random mutation technique was applied to the Agrobacterium d-carbamoylase. Three heat stable mutant enzymes were obtained. These mutations were found at His 57, Pro203, and Val236. These mutations were combined in one molecule, and the mutant enzyme containing the triple mutation His57Tyr, Pro203Glu, and Val236Ala had 19 °C higher heat stability than did the wild-type enzyme (Ikenaka et al. 1999). Escherichia coli cells containing this mutant enzyme were immobilized and used for practical industrial production of d-HPG with the simultaneous use of immobilized d-hydantoinase on line. This immobilized mutant d-carbamoylase reactor can be used for 1 year without any supply of new enzyme.
d-Phenylglycine is also produced by the same enzymatic method using the corresponding starting material.
Ozaki et al. developed a specific hydroxyproline detection method with high-performance liquid chromatography (Ozaki et al. 1995) and screened strains for microbial proline hydroxylase activity. l-Proline 4-hydroxylase was found in some etamycin-producing actinomycetes belonging to the genera Streptomyces, Dactylosporangium, or Amycolatopsis (Shibasaki et al. 1999). l-Proline 3-hydroxylase was found in some telomycin-producing actinomycetes belonging to the genus Streptomyces and in bacteria belonging to Bacillus (Mori et al. 1996).
The genes of these proline hydroxylase-producing organisms were cloned in E. coli cells, respectively, and the cells overexpressing the enzyme were used as the enzyme source in the industrial process of l-hydroxyproline production. Since the genes obtained from actinomycetes had some difficulty in being highly expressed in E. coli cells, the genetic codons corresponding to the N-terminal of the enzyme protein were changed to match the codon usage in E. coli. Furthermore, the promoter of trp operon was introduced twice at the promoter site of the gene in the plasmid to achieve the overexpression. These transformants expressed 1,400 times higher activity of proline 4-hydroxylase and 1,000 times higher activity of proline 3-hydroxylase in comparison with the original strain.
2-Oxoglutarate, one of the substrates of hydroxylation, is made from glucose in the reaction medium via the EMP pathway and TCA cycle in E. coli, and the product succinate is recycled. The mutant strain of E. coli lacking the l-proline-degrading enzyme was obtained and used for the host cells in the production of l-hydroxyproline.
Using E. coli as the host cells in l-proline production, the direct production of l-hydroxyproline from glucose became possible. In this case, the derepressed genes of the l-proline biosynthetic pathway were introduced into E. coli cells together with the gene of l-proline hydroxylase.
The estimated annual world production of l-lysine is around 500,000 t, almost all supplied by Ajinomoto, Kyowa Hakko Kogyou, Archer Daniels Midland (ADM), and Badische Anilin- und Soda-Fabrik (BASF). l-Lysine (an essential amino acid for swine and poultry) is useful as an additive to feeds such as grains and defatted soybeans, which contain less lysine.
l-Lysine is produced by some mutants derived from wild strains of glutamate-producing bacteria including Corynebacterium glutamicum, Brevibacterium lactofermentum, and B. flavum in the presence of high concentration of sugar and ammonium, at neutral pH, and under aerobic conditions (Tosaka and Takinami 1986).
These characteristics of lysine production were combined to make strains that were much more efficient producers. In addition, the introduction of a leucine-requiring mutation also increases the amount of lysine, since in the mutant dihydrodipicolinate synthase is released from repression by leucine.
The precursors of lysine synthesis include phosphoenolpyruvate, pyruvate, and acetyl CoA. Many mutations are induced in lysine producer cells to supply sufficient amounts of these precursors in good balance. These are deletion mutants of pyruvate kinase, those that show low activity of pyruvate dehydrogenase, etc. Furthermore, an alanine requirement was also reported to be effective in increasing the lysine amount.
The genes of the enzymes responsible for the biosynthesis of lysine in Corynebacterium have been cloned, and their nucleotide sequences are known. These genes include aspartokinase, aspartate semialdehyde dehydrogenase, dihydrodipicolinate synthase, dihydrodipicolinate reductase, tetrahydrodipicolinate succinylase, succinyl diaminopimelate desuccinylase, diaminopimelate dehydrogenase, and diaminopimelate decarboxylase (Tosaka and Takinami 1986). A host-vector system of Corynebacterium was established, and the introduction of some genes that encode the enzymes responsible for lysine biosynthesis (i.e., aspartokinase and dihydrodipicolinate synthase) was found to be effective in increasing the amount of lysine produced (Cremer et al. 1991).
A new gene ldc which encodes lysine decarboxylase (formerly known as cadA in E. coli) has been identified, and the enzyme was purified from the overexpressing strain. The lysine decarboxylase encoded by ldc is constitutively produced by E. coli cells, although lysine decarboxylase encoded by cadA is inducible (Kikuchi et al. 1997). It is interesting to note that this new lysine decarboxylase is present in lysine-producing Corynebacterium and to investigate the effects of deleting the gene on l-lysine production.
Vrljic et al. (1996) cloned a new gene lysE from Corynebacterium glutamicum and showed that it encodes a specific l-lysine exporter. Recently, they analyzed the membrane topology of the gene product and showed that it is a member of a protein family found in some other bacteria such as E. coli, Bacillus subtilis, Mycobacterium tuberculosis, and Helicobacter pylori. The authors suggested that LtsE superfamily members would be shown to catalyze export of a variety of biologically important solutes including amino acids (Aleshin et al. 1999; Vrljic et al. 1999; Zakataeva et al. 1999).
The annual worldwide production of l-threonine is around 13,000 to 14,000 t. l-Threonine is an essential amino acid for humans and some livestock animals, such as pigs and poultry. It is used as an additive for animal feed, medicines, food, and cosmetics.
l-Threonine is produced by some auxotrophic mutants or threonine-analogue-resistant mutants, and those are created by genetic engineering techniques. The bacteria used are Escherichia coli, Corynebacterium glutamicum, Brevibacterium lactofermentum, B. flavum, Serratia marcescens, and Proteus rettgeri (Nakamori 1986). l-Threonine production by fermentation was started in the 1970s. The auxotrophic mutant and analogue-resistant mutant strains obtained for this purpose were cultured in the presence of amino acids required by the mutant.
Auxotrophic mutants of l-lysine, diaminopimelate, or l-methionine were found to produce l-threonine in the culture medium, but the amount was not high enough to justify their use in practical production. A resistant mutant to an l-threonine analogue, α-amino-β-hydroxyvaleric acid (AHV), was obtained and shown to be an l-threonine producer. In this strain, homoserine dehydrogenase was insensitive to the feedback inhibition by l-threonine (Fig. 4.8 ). The much stronger l-threonine-producing strains were obtained by the combination of the auxotrophic mutations and AHV-resistant mutation. l-Threonine-producing mutant of S. marcescens was induced by the techniques of phage transduction. The strain has the following properties: deficiency of l-threonine-degrading enzymes; a mutation in the aspartokinase and homoserine dehydrogenase genes, making them insensitive to feedback inhibition by l-threonine; mutations in genes for l-threonine biosynthetic enzymes, releasing them from repression by l-threonine; a mutation in the aspartokinase gene, making it insensitive to feedback inhibition by l-lysine; and a mutation in the aspartokinase and homoserine dehydrogenase genes, releasing them from the repression by l-methionine.
Recombinant DNA techniques were employed to improve the l-threonine producer. Genes of the threonine operon obtained from AHV-resistant and feedback-insensitive mutants were introduced into a threonine-deficient mutant of E. coli to amplify the expression of enzymes and to increase the amount of l-threonine. Escherichia coli mutant strains were also constructed with amplified genes of the threonine operon (obtained from AHV-resistant and feedback-insensitive mutants) by the action of Mu phage on the chromosomal DNA. This strain is used in France for l-threonine production. Okamoto et al. constructed an l-threonine hyper-producing E. coli mutant that can produce l-threonine (100 g/l) in 77 h. They suggested that uptake of l-threonine in this strain is impaired (Okamoto et al. 1997).