Keywords

Amino acid fermentation is a huge industry. More than 5 million metric tons of amino acids (more than 2 million metric tons of glutamate, around 2 million metric tons of lysine, etc.) are manufactured annually all over the world [1]. Amino acids find application in a wide range of fields: seasoning, animal feed, medicine, and starting material for chemicals including pharmaceuticals. A number of scientific paper and patent on amino acid fermentation have been published every year. These studies have been contributing to widen our knowledge on the science of metabolism, genetics, and physiology.

The discovery of glutamate fermentation by Corynebacterium glutamicum in 1956 was the inception of amino acid fermentation. This discovery was a major breakthrough not only for glutamate production but also for research, demonstrating that amino acid fermentation is possible. In this chapter, the history of amino acid fermentation (glutamate fermentation is mainly focused on) before and after its discovery is reviewed.

1 Prehistory of Glutamate Fermentation

Although glutamate was isolated from gluten in 1866, its major applicational value was found in 1908 by Kikunae Ikeda [2]. He studied taste presenting substance derived from kelp, whose soup has traditionally been used as a basic seasoning in Japan, and identified the monosodium salt of l-glutamate as an umami substance. An entrepreneur, Saburosuke Suzuki, collaborated with Dr. Ikeda and commercialized the discovery; he began selling monosodium glutamate (MSG) as a new seasoning “AJI-NO-MOTO®” in 1908 (Fig. 1).

Fig. 1
figure 1

The first product of umami seasoning, “AJI-NO-MOTO®.” The photo is reprinted under the kind permission of Ajinomoto Co., Inc.

Full size image

Because umami is a traditional taste component in Japan, the new product got a great sales success. However, there were several difficulties in manufacturing MSG, for which wheat gluten was hydrolyzed using hydrochloric acid and l-glutamate was isolated following the recrystallization as the monosodium salt. Because there were scarce materials that could tolerate acidic conditions under high temperature during hydrolyzation in the early twentieth century, obtaining suitable vessels for hydrolyzation was a big issue. After several trials and errors, a certain type of ceramic pot was found to be suitable. Hydrogen chloride gas released during the hydrolyzation process was very hazardous. Furthermore, the formation of a large volume of waste (the remaining fraction of the hydrolyzate) presented additional issue [3].

Therefore, there was a strong demand for a new manufacturing process for MSG or l-glutamate. During the mid-1950s, a decade after the end of World War II, the economy and life of people in Japan was returning to normal. Corresponding with the social situation, the demand for MSG had increased rapidly; the production of MSG by Ajinomoto Co., Inc jumped from 6,662 metric tons in 1955 to 13,586 metric tons in 1959 [4].

Several attempts were made to establish a new method for manufacturing l-glutamate. Because of the necessity of optical resolution, chemical synthesis was not advantageous while it was used commercially for a particular period. Chemical synthesis coupled with enzymatic resolution was a feasible method [5]. Izaki et al. have reported specific degradation of d-glutamate by Aerobacter [6]. The optical resolution of N-acyl-dl-glutamate by d-specific acylase activities derived from Aspergillus tamarii and Penicillium vinaceum presented additional approach [7].

Since glutamate is biosynthesized from 2-oxoglutarate through amination, it was natural to pursue the process that consisted of microbial production of 2-oxoglutarate and microbial amination of the oxoacid.

The earliest report on fermentative production of 2-oxoglutarate can be found in 1946; Pseudomonas fluorescens accumulated 16–17 g of the acid per 100 g of glucose [8]. The yield was increased to 41 g of the acid per 100 g of glucose [9].

Formation of small amounts of glutamate in the presence of 2-oxoglutarate and ammonium salts was reported in Clostridium [10], Escherichia coli [11], Bacillus subtilis [12], P. fluorescens [13], and Pseudomonas ovalis [14]. Attempts to perform the conversion using amino acids as amino donors were also reported [13, 15].

The combination of 2-oxoglutarate fermentation and amination of the oxoacid appears to be just one step away from direct fermentation of glutamic acid (direct fermentation refers to the process of producing the product from a low-cost carbon source, such as sugar, and a nitrogen source, such as ammonia, through the cultivation of microorganisms). It was known that accumulation of small amounts of glutamate occurred in the cultivation medium of certain bacteria. Morrison and Hinshelwood and Dagley et al. have observed very small amount of accumulation of the amino acid in the cultivation broth of E. coli and Aerobacter aerogenes [16, 17]. Thorne et al. have reported that B. anthracis formed glutamate (2 g/L) under the conditions in which polyglutamate formation was hampered [18]. Asai et al. screened for a glutamate producer and found Micrococcus varians to be the best producer although the accumulation was far lower (2.9 g/L at the optimal efficiency) for industrial application [19].

2 Discovery and Industrialization of Glutamic Acid Fermentation

M. J. Johnson, an emeritus professor at the University of Wisconsin, described the situation in 1955, “there is, in many quarters, great interest in fermentative glutamic acid production” [20]. However, there was a strong notion that it is irrational to expect a microorganism to accumulate a large quantity of the amino acid extracellularly because (1) glutamate is an essential component for the organism, (2) excretion of glutamate is an economical loss for the cell because the biosynthesis of the amino acid is energetically expensive, and (3) secretion of glutamate would not be expected to have suppressing effect against surrounding microorganisms like antibiotic dose. Researchers in Kyowa Hakko Kogyo Co., Ltd. have overcome this conceptual obstacle.

Dr. Udaka, a researcher of Kyowa Hakko Kogyo Co., Ltd. at that time, has set the following screening method [21]. Bacteria isolated from the environment were replicated on nutrient agar plates and several types of defined medium plates (test plates). After colony formation, the test plates were exposed to UV radiation to kill the organisms. Then, soft agar medium containing the glutamate auxotroph Leuconostoc mesenteroides was overlaid on the test plates. The halo of growth development around the colony on the test plates indicated glutamate excretion by the colony (Fig. 2).

Fig. 2
figure 2

Bioassay screening of glutamate-producing microorganisms. Glutamate productivity of the test strain can be estimated by the scale of the halo formed around the strain. The photo is reprinted under the kind permission of Kyowa Hakko Bio Co. Ltd.

Full size image

After screening only approximately 500 isolates, they found the superior strain, Corynebacterium glutamicum (originally reported under the name of Micrococcus glutamicus). The strain accumulated 10.3 g of glutamate per liter when cultivated in a flask with liquid synthetic medium with 5% glucose [22], and the accumulation was easily increased to >30 g/L [23] with >25% yield against glucose input, indicating that the strain is applicable for industrial glutamate fermentation.

When the study was scaled up, however, the researchers faced a puzzle; the bacterium produced only trace amounts of glutamate. Several months of intensive research revealed important characteristics of glutamate fermentation: the bacterium is a biotin auxotroph and produces glutamate only under biotin-limited conditions. It was assumed that the cells grew in screening and flask cultivation using a trace amount of biotin carried over from the pre-culture and, thus, resulted in biotin-limited conditions.

The first commercial fermentation of glutamate was conducted in 1958 at a plant of Kyowa Hakko Kogyo. There were still problems to be solved for commercialization, such as the downstream process, particularly the control of the crystal form. In that way amino acid fermentation took the first step.

3 Establishment of Glutamate Fermentation

Glutamate fermentation by C. glutamicum was first presented at the International Symposium on Enzyme Chemistry held in Japan in 1957. As the congress was the first major international scientific meeting held in Japan after World War II, the presentation attained great attention from Japanese researchers in the field of life sciences. Once the discovery of glutamate fermentation was public, several researchers and companies rushed into the research. It looks like a “gold rush,” which is observed in Figs. 3 and 4.

Fig. 3
figure 3

Trends of scientific paper publication on glutamate fermentation in English. For each decades (except for 1957–1970 and 2001–2014), the number of papers are counted by country. Reviews, patents, and oral presentations are not included

Full size image
Fig. 4
figure 4

Numbers of Japanese patents (black bar) and oral presentations (white bar) in the annual meeting of Japan Society for Bioscience, Biotechnology, and Agrochemistry from 1957 to 1967

Full size image

Until 1970, Japanese researchers enthusiastically conducted research and held a monopoly on the scientific papers on glutamate fermentation (Fig. 3). Japanese research activity appeared to steady down by the 1980s; however, it showed resurgence during the 2000s. In contrast, reports from other Asian countries, particularly from China and India, showed an increase from the 1980s. Several of these studies have outlined the discovery of a new strain (most strains are under Corynebacterium sp.), the application of a new raw material, and improvement of the production system, suggesting a strong economical demand of glutamate fermentation in these areas. Research from Europe began to emerge from the 1990s. Most of the European research focused on the mechanism of glutamate fermentation and played an essential role in understanding the mechanism.

Figure 4 shows the “gold rush” in Japan from a different angle. A rapid increase of patents in 1960 suggests that companies began research on glutamate fermentation after the announcement of the discovery of C. glutamicum. Representative strains and carbon source are shown in Table 1. Most of the important aspects of the fermentation appeared before 1980, which are summarized below. These insights are closely or directly related to the elucidation of the mechanism of glutamate production discussed in [35].

Table 1 Glutamate fermentation from different carbon sources by different microorganisms
Full size table

3.1 Producer Strain

Many species were claimed as glutamate-producing microorganisms such as Brevibacterium lactofermentum, Brevibacterium flavum, Corynebacterium callunae, Corynebacterium lilium, etc. Most of them were aerobic, gram-positive, nonacid-fast, nonspore-forming, rod-shaped, and biotin-requiring bacteria. These “new” species were categorized in the spices of Corynebacterium glutamicum by thorough taxonomic investigation in later years [36–38]. Thus, the name of C. glutamicum is used for these bacteria hereinafter. It became a shared sense within several years after the discovery of the bacterium that the bacterium (even wild-type strain) produces glutamate at a yield of >40% against input sugar (Table 1) under appropriate conditions.

Bacteria capable of producing glutamate from sugar and not belonging to C. glutamicum have also been reported. Chao and Foster have reported productivity of 13.5 g/L from 3% glucose by Bacillus megaterium [39]. It is interesting that this strain is also biotin auxotroph. Some Arthrobacter [40] and Streptomyces [41] strains were also reported to produce glutamate. Recently, it was reported that metabolically engineered E. coli, Enterobacter agglomerans, Klebsiella planticola, and Pantoea agglomerans produce significant amounts of glutamate [42, 43].

3.2 Carbon Source

As C. glutamicum readily utilizes glucose, fructose, and sucrose, these sugars are initially used as a carbon source. Molasses, a more economically desirable source, was mainly used after the discovery of an alternative method of biotin limitation because it contains excess amount of biotin. Ethanol and acetic acid have also been reported to provide the bacterium a good productivity. Recently, it was demonstrated that the spectrum of usable sugar of C. glutamicum can be expanded by expressing heterologous genes [44].

Carbohydrate obtained by petrochemistry was studied as an alternative carbon source. Shiio and Uchio have reported several kinds of bacteria capable of forming glutamate from paraffin [32]. C. hydrocarboclustus was reported to have relatively high productivity (Table 1). Ghosh and Banerjee have described the production from n-alkane by the Serratia marcescens strain [45]. In addition, aromatic compounds, such as benzoate, have been investigated as a carbon source [33, 34, 46].

Glutamate fermentation from methanol has been attempted since the 1970s [29, 47]. As C. glutamicum cannot assimilate methanol, various microorganisms such as Methanomonas methylovora [29], Methylobacillus glycogenes [48], and B. methanolicus [30] have been screened and used. The B. methanolicus strain M168-20(pHP13) appears to be the highest producer from methanol (Table 1). Glutamate production through photosynthesis (carbon source is CO2) has also been investigated using algae [49, 50].

3.3 Production Conditions

As described above, biotin limitation was the crucial factor of glutamate fermentation by C. glutamicum. This feature made it impossible to use some economically desirable raw materials, such as molasses. Attempts to clarify the underlying mechanism and finding alternative methods of biotin limitation have been made since a very early period. Shiio et al. first reported that biotin limitation causes the change of cellular permeability of amino acids including glutamate [51]. Reports that followed confirm the association between biotin limitation and glutamate permeability [52, 53].

Several alternative methods of biotin limitation were devised in the 1960s. One of the major breakthroughs was the addition of penicillin; adding an appropriate amount of penicillin at an early stage of cultivation triggers glutamate production under biotin excess conditions [54]. Addition of surfactant was an alternative method. Because the effect of surfactant depends on its chemical composition, specific surfactants were chosen, such as polyoxyethylene sorbitan monostearate [55, 56] and cetyltrimethylammonium bromide [57].

Penicillin inhibits cell wall synthesis. The first action point of a surfactant should be the cell surface. Putting these and the biotin effect together, the “permeability hypothesis” or the “leak model” was claimed to be the mechanism of glutamate fermentation; change in cellular permeability of glutamate caused by some treatment triggers the leak out of the amino acid, which shifts intracellular metabolism toward glutamate formation [53, 58]. The finding of the relationship between fatty acid composition in the cell membrane and glutamate productivity [59, 60] supported the notion. It was also supported by the finding that an oleic acid-requiring mutant produces glutamate under biotin-sufficient conditions [26, 61]. Based on these lines of evidences, the mechanism of glutamate fermentation seemed to have been settled around 1970. However, further contention emerged later (see below and [35]).

Conditions other than biotin were investigated in detail, but only effects of oxygen and pH are mentioned here. Glutamate production by C. glutamicum requires aerobic conditions. Under oxygen-insufficient conditions, the bacterium produces succinic acid and/or lactic acid [62]. Medium pH should be maintained slightly above 7.0. When pH is controlled to be acidic, around pH 5.5, the bacterium produces mainly glutamine.

4 Recent Development in Glutamic Acid Fermentation

4.1 In C. glutamicum

In general, (a) genetic modification(s) is necessary for forcing a microorganism to overproduce a certain amino acid, for example, deregulation of aspartokinase for lysine production. Glutamate production by C. glutamicum is different. The wild-type strain exerts high productivity under appropriate conditions as described above. Although there have been a lot of patents claiming mutations beneficial for glutamate production by the bacterium, none of them appear to improve the productivity significantly. However, since the 1990s, genes essential for glutamate fermentation have been identified, which shed new light on the mechanism. Since the mechanism is discussed in Chap. 4, a very rough sketch is presented below.

There have been two predominant hypotheses on the mechanism. The first idea is the “permeability hypothesis” as described above. The other is the “flux hypothesis,” claiming that change in carbon metabolism is the major cause of the overproduction, which is a quite common concept among other amino acid fermentations. Facts supporting each hypothesis are presented below; however, it should be noted that the two hypotheses are not necessarily exclusive.

Before mentioning on the flux hypothesis, the biosynthetic pathway of glutamate is briefly reviewed. Glutamate is formed by amination of 2-oxoglutarate. Glutamate dehydrogenase (GDH) and the coupled reactions of glutamine synthetase and glutamine synthase (GS/GOGAT system) are involved in the amination [63–67]. While cell growth can be sustained either by GDH or GS/GOGAT system, GDH is responsible for glutamate overproduction [67]. Carbon from sugar is metabolized to 2-oxoglutarate through glycolysis and part of the TCA cycle. Enzymes contributing to the flux have been clarified, although the flux around oxaloacetate is complicated. Characteristics and regulation of these enzymes have also been reported [68–73]. 2-Oxoglutarate is the important branch point, the glutamate-forming direction by amination and the glutamate-degrading direction by oxidative decarboxylation to succinate with 2-oxoglutaratedehydrogenase complex (ODHC). ODHC is the focal point in flux control.

During the 1960s, it was assumed that C. glutamicum had no or very low activity of ODHC [74]. This was the basis of the flux model during the early 1960s, but the permeability hypothesis became dominant as described above. Later, it was confirmed that C. glutamicum possessed certain ODHC activity [75]. The flux hypothesis was revived by the finding that deletion of the odhA gene (encoding E1o subunit of ODHC) conferred the bacterium glutamate overproductivity ([76], oral presentation was in 1996). Simultaneously it was reported that ODHC activity is reduced under glutamate-producing conditions [77]. Kim et al. have reported the supportive results [78]. Recently, it was revealed that ODHC activity is regulated by OdhI/PknG [79] and that OdhI, the inhibitor of ODHC, is induced under glutamate-producing conditions [80].

In parallel with the revival, supportive findings for the permeability hypothesis have also been accumulated. In 1989, Hoischen and Kramer have demonstrated by biochemical analysis that glutamate excretion is mediated by an active efflux system and not by simple leakage [81]. They further demonstrated the importance of alternation of membrane tension [82, 83]. Changes in the membrane components cause such alternation of membrane tension, and there have been several reports indicating the association between change in membrane composition and glutamate overproductivity [84–86].

Genetic findings supporting the permeability hypothesis have also been accumulated. Kimura et al. have described the dtsR gene that restores detergent (Tween 40) sensitivity [87]. DtsR has a high similarity to the β-subunit of methylmalonyl-CoA carboxyltransferase, and it is suggested to be the biotin enzyme involved in fatty acid biosynthesis. The disruption of dtsR causes oleic acid auxotrophy and glutamate production [88]. The ltsA gene was found to code a gene whose disruption causes the cell to be lysozyme sensitive and glutamate productive [89, 90]; however, the physiological role of the gene product is not known. Nakamura et al. found a mechanosensitive channel Ncgl1221 for glutamate excretion [91–95]. Mutation of the gene caused glutamate overproduction even in the presence of intact OdhA [92, 94].

The production mechanism still remains elusive, but it will be comprehensively understood in the near future.

Because glycolysis and the TCA cycle are the major carbon metabolic pathways to glutamate, the relation between energy metabolism and glutamate production is a point to be investigated. In this respect, Yokota et al. have reported that mutants decreasing H-ATPase activity at below 1/4 the level of the wild type produce less glutamate [96].

4.2 In Other Bacteria

It was a strong support of the flux hypothesis that the disruption of odhA confers E. coli glutamate productivity [43, 74]. Since glutamate productivity by the disruption of the gene has known in other bacteria [42], at least, reducing the flux through ODHC appears to be generally applicable.

Glutamate excreted in the medium is reimported by the glutamate import system in C. glutamicum [97, 98] or other bacteria, such as E. coli [99]. Thus, an increase in glutamate in the medium exerts a negative effect on glutamate synthesis through the regulation of biosynthetic enzymes. When glutamate dissolved in the medium forms crystals, it no longer has further effect on the regulation. Since isoelectric point of glutamate is around pH 5.5, glutamate concentration dissolved in the medium is lowest at the pH area. Thus, if a bacterium that grows and produces glutamate in this pH, it must be advantageous for glutamate production by reducing the negative effect of glutamate in the medium. Based on this idea, several bacteria were found and constructed as the producers [100]. The entire genome sequence of one strain has recently been published [101].

5 Amino Acid Fermentation

Whether it is a natural expectation or not, researchers in Kyowa Hakko Kogyo Co., Ltd. thought that C. glutamicum might produce other amino acids immediately after the discovery of glutamate fermentation. They started delivering mutants for amino acid producers. The attempt quickly resulted in the second amino acid fermentation; a mutant requiring arginine accumulated ornithine [24]. This was the first report releasing regulation on amino acid biosynthesis by amino acid auxotrophy. Further, in the next year, the researchers reported lysine-producing mutants of C. glutamicum based on the same idea [102]. Earliest reports on each of the amino acid fermentations are listed in Table 2.

Table 2 Earliest scientific reports on each of the amino acid fermentations
Full size table

The other important method liberating regulation is delivering analogue-resistant mutants. Although it was known since the 1950s that analogue resistance confers amino acid overproductivity [106, 108], the first clear and quantitative example was presented by Sano and Shiio [103]. They demonstrated that an S-(2-aminoethyl)-l-cysteine-resistant mutation makes aspartokinase of C. glutamicum insensitive to feedback regulation by lysine, resulting in lysine production. Since obtaining resistant mutants (positive screening) is easier than obtaining auxotrophic mutants (negative screening), this method was a very powerful tool for amino acid fermentation.

Producer breeding was not limited to C. glutamicum. Other bacteria, such as E. coli and S. marcescens, were manipulated and used as amino acid producers. In these enterobacteria, destroying the degrading activity of the target amino acid was important at some times. For example, the construction of the threonine producer of S. marcescens was initiated by getting a mutant incapable of metabolizing threonine [116]. Destruction of the amino acid-degrading ability and deregulation of the biosynthetic pathway can be regarded as the general strategy of strain improvement.

Since the 1980s, recombinant DNA technology has been available and applicable for strain improvement. Gene dosage effect is the most instant use of the technology. A gene coding for a key enzyme is amplified by cloning on a multi-copy plasmid. An increase in the expression results in the increase of the activity, enhancing the metabolic flux to the desired direction. Combining conventional mutation and recombinant DNA technology enables the construction of a producer strain for the amino acid whose biosynthesis is under multiple and complex regulation, such as tryptophan [117].

Discovery of an amino acid exporter was noteworthy. The lysine exporter in C. glutamicum was first predicted using biochemical analysis [118, 119] and then confirmed genetically [120]. The finding is interesting not only because it is counterintuitive but also because it raises a new key concept in amino acid fermentation. Currently, over eight kinds of amino acid exporters have been identified in C. glutamicum [121, 122] and E. coli [123–127].

Knowing carbon flux in the producer microorganism is important for strain improvement. Carbon distributions between the hexose monophosphate pathway (HMP) and the pentose monophosphate pathway (PPP) and around the oxaloacetate supply have been the focal points of research interest. 13C-NMR analysis combined with metabolic balance analysis has been widely used for this purpose, and much insight has been accumulated [128–130]. One of the earliest finding of the analysis is that the carbon distribution through HMP and PPP in C. glutamicum is different between the glutamate producer and the lysine producer [131]. The carbon flux distribution between HMP and PPP was 8:2 in the glutamate producer, whereas it was 4:6 in the lysine producer, suggesting that the flux is controlled by the necessity of NADPH. Subsequently, this finding led to cofactor engineering [132–134].

6 Feature Prospective

At the end of the twentieth century, the whole genome sequence determination was initiated in various kinds of organisms including E. coli [135] and C. glutamicum [136]. Genome data enabled a new way of strain improvement termed as “genome breeding” [137–139] that creates a minimal set of mutations beneficial for production by (1) comparative analysis of whole genomes of the wild-type strain and a producer strain obtained from the wild-type via successive classical mutations, (2) defining beneficial mutations, and (3) assembling them in the wild-type background. This enables the construction of a producer strain with high productivity without undesirable traits of the classical producer mutant, such as slow growth and stress sensitiveness.

Glutamate fermentation has been used mainly for production of MSG, a seasoning. However, it may take part in wider areas of application because glutamate is expected to be one of the basic chemicals from renewable resources [140]. Thus, glutamate fermentation yet retains enough potential for contributing for the quality of human life in addition to its historical role in modern fermentation industry and applied microbiology.