Keywords

1 Introduction

Vitamin C, also named as L-ascorbic acid, is a kind of essential vitamin and antioxidant for human beings and some primates (Zhang et al. 2011). Vitamin C is widely used in pharmaceutical, foods, beverages, cosmetics and feeds industries (Table 12.1). Along with the expansion of vitamin C in different field, the market is still growing. The world market for vitamin C and its direct derivatives (such as calcium salt, potassium salt, sodium salt, and glycosylation derivatives) is beyond 0.5 billion US dollars.

Table 12.1 Application of vitamin C and its derivatives
Full size table

The earliest commercial vitamin C production process began from the Reichstein route (Table 12.2), which is developed by Reichstein in 1934, Germany. Reichstein developed a strategy for conversion of glucose to 2-keto-L-gulonic acid (2-KLG) with five steps of chemical reactions and one step of bio-reaction (Reichstein and Grussner 1934). The 2-KLG was further converted into vitamin C with esterification (Fig. 12.1). The process was industrialized and dominated the vitamin C market for over 60 years. However, the Reichstein process has many disadvantages, such as high-energy consuming, high-amount of organic solvent requirement and serious environment pollution. Therefore, many researchers find out several different processes to improve the performance of classical Reichstein process. One of the most successful routes is the “two-step” fermentation process, which was applied in industrial scale for more than 40 years. Along with the development of industrial biotechnology, researchers further developed several other processes to improve the economic and environmental aspects of the two-step fermentation process.

Table 12.2 Manufacturing methods of vitamin C
Full size table
Fig. 12.1
figure 1

Reichstein process for vitamin C production. The D-glucose was hydrogenated to form D-sorbitol. The D-sorbitol was converted into L-sorbose by acetic bacteria. The L-sorbose was further oxidized with protection to form 2-KLG. The 2-KLG was then esterified and lactonized to form vitamin C

Full size image

According to the production process, vitamin C production by biotechnology can now be divided into two-step fermentation and one-step fermentation. The two-step fermentation including tandem fermentation process that uses glucose as substrate and the fermentation process that uses D-sorbitol as substrate (Fig. 12.2). For the former, D-glucose is converted into 2,5-diketo-D-gluconate (2,5-DKG) by Erwinia sp., which is subsequently converted into 2-KLG by Corynebacterium. The latter was studied the earliest and most intensively. In this process, L-sorbose is formed from D-sorbitol by Gluconobacter oxydans, and then by Ketogulonicigenium vulgare with Bacillus megaterium as the associated bacterium, L-sorbose is converted into 2-KLG (Zhang et al. 2011). K. vulgare is a 2-KLG producing bacterium, its growth is a lengthy, inefficient process with a low 2-KLG yield when cultured alone. B. megaterium is an associated bacterium which can promote K. vulgare growth and acid production but it does not produce 2-KLG (Liu et al. 2011a).

Fig. 12.2
figure 2

Classical two-step fermentation process. The D-glucose was hydrogenated to form D-sorbitol. The D-sorbitol was converted into L-sorbose by acetic acid bacteria. The L-sorbose was further oxidized with a mixture culture system with B. megaterium and K. vulgare to form 2-KLG. The 2-KLG was then esterified and lactonized to form vitamin C. The only difference between the classical two-step process and the Reichstein process is the replacement of low efficient protective oxidation with a fermentation process. *The G. oxydans here was further identified to be K. vulgare

Full size image

The optimization of the fermentation process has been continuing ever since, tremendous work has been done to improve the process stability and yield, such as the medium optimization, the relationship between two bacteria, the ecological regulation, etc. (Zhang et al. 2010). One-step fermentation is based on the bacteria of the two-step fermentation, it aims to achieve one-step fermentation from glucose or sorbitol to 2-KLG by using genetic engineering techniques to construct the genetically engineered bacteria.

2 Elucidation of the Mechanisms in Mutualism Between Bacillus megaterium and Ketogulonicigenium vulgare

2.1 Previous Research on the Classical Two-Step Fermentation Process

Because of co-culture of K. vulgare and B. megaterium during the classical two-step fermentation period of vitamin C, the optimal culture condition not only very suitable for growing condition of two bacteria but the excretion of acid by K. vulgare (Zhang et al. 2011). Zhou et al. discovered that carbon source, nitrogen source, dissolved oxygen, growth factors of seed culture medium and environmental factors etc., could regulate the growth and metabolism of strains effectively, and make the organisms grow under an optimal condition (Zhou et al. 2002). Therefore, there would be considerable impacts on the fermentation. The metabolism of mixed bacteria in vitamin C fermentation was more complex compared with that of a single strain.

The commonly used urea and corn steep liquor were nitrogen source in the industrial production of vitamin C by the classical two-step fermentation process. The metabolic properties were listed as follow: (1) There were two kinds of roles with urea addition, one was as a physiological alkaline substance to adjust pH, the other was providing the nitrogen sources for bacterial metabolism; (2) The protein content of the system was increasing with fermentation time; (3) The 17 kinds of amino acids in corn steep liquor were divided into three categories (Li et al. 1996). The large number of unclear natural components and the different batches of corn steep liquor can also cause the instability of 2-KLG production and fermentation cycle. The specific growth factors of cell growth and acid production were a prerequisite to control the stability of 2-KLG production.

Dissolved oxygen was also an important factor in the fermentation process of vitamin C. Giridhar and Srivastava (2000) found that in the batch and fed-batch fermentation processes of D-sorbitol to L-sorbose, adding 4% of the oxygen vector (N-hexadecane) could shorten the period for 2 and 5 h, and increase the production by 17 and 26%, respectively.

2.2 Elimination the Inhibition Effects of Substrate to Improve the Production and Yield of 2-KLG

The microbial transformation of D-sorbitol to L-sorbose is the most important industrial fermentation process of vitamin C in “Reichstein process” and two-step fermentation. Microorganisms are inhibited severely by high concentration of D-sorbitol. The oxidation rate of D-sorbitol decrease drastically, which lead to the low concentration of L-sorbose in broth. Thereby, it will increase the downstream cost of the separation and purification. Control of initial concentration of sorbitol with 200 g/L, the production and yield of L-sorbose were 200 g/L and 14.2 g/L/h (Giridhar and Srivastava 2002). Besides this, with constant speed feeding of 600 g/L D-sorbitol, the concentration of L-sorbose was 1.6 times greater than batch fermentation; whereas 272.37 g/L L-sorbose achieved with linear feeding strategy (Giridhar and Srivastava 2001); exponential feeding of 700 g/L D-sorbitol after 10 h batch fermentation with 200 g/L D-sorbitol made 290 g/L L-sorbose (Srivastava and Lasrado 1998).

When L-sorbose was further converted to 2-KLG, huge amount of L-sorbose will make the bacteria metabolize abnormally, meanwhile the production and yield reduced. With the fermentation of 80 g/L sorbose for 35–40 h, the concentration of 2-KLG was 65–75 g/L. In the industrial process, constant feeding strategy was used widely. During the whole fermentation process, the curve of acids producing rate shaped a saddle while an ideal curve should be only a peak. Feeding sorbose according to the curve of fermentation rate would make the curve of acids producing fit the ideal model closer. Feeding L-sorbose at 10 and 20 h during fermentation, L-sorbose was 100 and 140 g/L, the concentration of 2-KLG reached 120–135 g/L, and the transformation efficiency was 90%.

2.3 Improvement of High Temperature Tolerance

Previous researches showed that B. megaterium could grow very well under 33 °C. However, K. vulgare could only grow under 30 °C. Therefore, the energy cost could be significantly improved by improving the higher temperature tolerance of K. vulgare. Because the mechanisms inside the tolerance of bacteria to high temperature are obscure, the rational metabolic engineering of K. vulgare to improve its tolerance to higher temperature is difficult. Current improvement of the high temperature tolerance of K. vulgare is mainly focused on the random mutagenesis (Yan et al. 2006) or de novo screening from nature (Moonmangmee et al. 2000). Yan et al. improved the performance of the mix-culture system consists of K. vulgare G0 and B. megaterium B0 under 33 °C by obtaining of mutant GI13 by ion implantation. The conversion ratio of the mutant GI13 at 33°C was improved to 94%. Besides, the L-sorbose dehydrogenase (SDH) activity of the mutant strain was further improved by 100% compared to the wild-type strain G0 (Yan et al. 2006).

G. oxydans, which is used for the conversion of the D-sorbitol to L-sorbose, could only grow under 32 °C (Yang and Lim 1997). Moonmangmee et al. (2000) screened a G. oxydans CHM54 from plants. The strain could grow under 37 °C well and has a conversion ratio of 80%.

2.4 Inhibition of the Degradation Pathway of 2-KLG

During the fermentation process of the mix-culture of Pseudomonas putida ATCC 21812 and Gluconobacter melanogenus IFO 3293 for the 2-KLG production, it was discovered that there was an enzyme, which needs NAD(P)H as cofactor, could further convert the 2-KLG to L-idonic acid. Furthermore, a similar enzyme was also found in K. vulgare WB0104 (Yang et al. 2006). Besides, the 2-ketoaldonate reductase A and B in Erwinia herbicola SCB125 could also degrade 2,5-DKG and 2-KLG. These downstream pathways could inhibit the further accumulation of 2-KLG. Therefore, knockout of these genes is essential for the further improvement of 2-KLG production. Chen et al. (2000) knocked out the thrA gene with a streptomycin resistance gene and successfully decreased the 2-KLG degradation.

2.5 Effects of the Relationship Between Two Bacteria in Mixed Culture on Vitamin C Production

The two-step fermentation process which uses D-sorbitol as the substrate is a classic example where a mixed-culture method produces a metabolic product. The second step fermentation is completed by mixed fermentation of 2-KLG-producing bacteria K. vulgare and associated bacteria B. megaterium. K. vulgare is difficult to culture alone and 2-KLG production capacity is very low simultaneously. B. megaterium does not produce 2-KLG, but it can promote the growth and acid production of K. vulgare. There are many kinds of associated bacteria, In addition to the commonly used B. megaterium, Bacillus cereus, Bacillus subtilis, Bacillus licheniformis, Bacillus thuringiensis, and some yeast strains also possess the role.

It is a collaborative symbiotic relationship between acid producing bacteria and associated bacteria (Zhou et al. 2002). The enzyme activity that synthesizes of 2-KLG is directly proportional to the amount of K. vulgare cells and has nothing to do with the amount of B. megaterium cells. Further studies have shown that the 30–50 KDa (Feng et al. 2000) and >100 KDa (Lv et al. 2001) biological active substances of B. megaterium released in the metabolic process, participating in the growth and acid production process of K. vulgare. The extracellular active substances are proteins.

Three bacteria were involved in the classical two-step fermentation process, for example G. oxydans, B. megaterium and K. vulgare. The denomination of the bacteria for G. oxydans and K. vuglare is highly speculative because of the lack of efficient strain identification methods. Furthermore, there are also some updates on the two strains. In early times, G. oxydans were named as G. melanogenus, Acetobacter suboxydans, or A. melanogenus. Because most of these stains could not be sufficiently distinguished for further confirmation, all these strains that could efficiently catalyze D-sorbitol to L-sorbose are now designated as G. oxydans. The species classified as K. vulgare were finally defined as K. vulgare. The bacterium was previously known as G. oxydans or some other Acetobacter sp. strains. However, the strain that could only grow well with B. megaterium or other Bacillus sp. is now referred to as K. vulgare.

In the classical two-step fermentation process, the D-sorbitol was synthesized by hydrogenation from D-glucose by chemical reaction. The D-sorbitol was then converted into L-sorbose by the sorbitol dehydrogenase (encoded by sldh) in G. oxydans. The culture broth containing L-sorbose was then transferred to another bioreactor, added with other raw materials and sterilized for the second time. Then the L-sorbose was converted in to L-sorbosone by the L-sorbose dehydrogenase (encoded by sdh) in K. vulgare. The L-sorbosone was further converted into 2-KLG by L-sorbone dehydrogenase (encoded by sndh) in K. vulgare.

The fantastic phenomenon in the classical two-step fermentation process is the second step, which consists of the B. megaterium and K. vulgare. The K. vulgare itself could not grow well after supplement of different kinds of substrates. The addition of B. megaterium, B. thuringiensis or B. cereus could significantly enhanced both the cell growth and 2-KLG production by the K. vulgare. There are many reports aiming to demonstrate the mutualism between the two bacteria. In the past few decades, much effort has been devoted to elucidating the detailed mechanism by which B. megaterium enhances K. vulgare growth and 2-KLG production (Bremus et al. 2006; Shinjoh et al. 1995; Sugisawa et al. 2005; Tsukada and Perlman 1972). Studies have showed that both internal and external metabolites from B. megaterium, through biochemical and molecular methods and certain proteins or amino acids, play a role in this process (Zhang et al. 2011; Zhao et al. 2008). Further research has shown that B. megaterium secretes two kinds of proteins of about 30–50 kDa and >100 kDa that improve K. vulgare growth and increase 2-KLG productivity (Lu et al. 2003; Xu et al. 2004; Zhao et al. 2008). However, researchers have not specifically determined which proteins play a role in promoting K. vulgare growth and 2-KLG production. Based on these results, Zhang et al. (2010) used lysozyme to damage the B. megaterium cell wall structure to release the intracellular components, as a consequence, the growth rate of K. vulgare, its sorbose consumption rate, and 2-KLG productivity increased 27.4, 37.1, and 28.2%, respectively. Because most of the companion strains for K. vulgare are Bacillus sp., it was proposed that the spore formation of these Bacillus strains should have some impact on the mutualism process. Zhu et al. (2012) first revealed that not only the spore formation, but also the spore stability, played key role in the mutualism process.

Based on the development of next-generation high-throughput sequencing technologies, all the genomes of three strains involved in the classical two-step fermentation process were obtained in last several years. One of the B. megaterium was sequenced. The 4.14-Mb genome of B. megaterium WSH-002 contains four replicons, a circular chromosome (4.04 Mb) encoding 5,186 predicted open reading frames (ORFs), and three circular plasmids, named pBME_100 (0.074 Mb), pBME_200 (9,699 bp), and pBME_300 (7,006 bp), with mean GC contents of 39.1, 36, 32.2, and 33.2%, respectively. There are 5,482 protein-encoding genes, 99 tRNAs, and 10 rRNA operons. Among them, 2,460 functional descriptions, 1,327 gene abbreviations, and 856 EC numbers were assigned to the WSH-002 genome by function annotation. Furthermore, 782 genes were assigned according to the Kyoto Encyclopedia of Genes and Genomes metabolic pathways (Liu et al. 2011c).

There are two K. vulgare strains independently. According to phylogenetics, both K. vulgare Y25 and WSH-001 has the same ancestor. The two strains were used by the two main vitamin C manufactures in China, North China Pharmaceutical Group Corporation and Jiangsu Jiangshan Pharmaceutical Company, respectively. However, the multiple-mutagenesis with different methods may lead to some significant differences in these two strains. The genome of K. vulgare Y25 consists of a circular chromosome and two plasmids. The chromosome is composed of 2,776,084 bp, with a GC content of 61.72%. One plasmid contains 268,675 bp, with a GC content of 61.35%, and the other contains 243,645 bp, with a GC content of 62.63%. There are a total of 3,290 putative ORFs (2,807 [chromosome], 256 [pYP1], and 227 [pYP2]) using Glimmer, giving a coding intensity of 91.05% (Xiong et al. 2011). The complete genome sequence of K. vulgare WSH-001 is composed of a circular, 2,766,400-bp chromosome and two circular plasmids named pKVU_100 (267,986 bp) and pKVU_200 (242,715 bp) with mean GC contents of 61.69, 61.33, and 62.58%, respectively. There are 2,604 protein-encoding genes, three rRNA operons, and 51 tRNA-encoding genes in the chromosome and 246 and 215 protein-encoding genes in plasmids pKVU_100 and pKVU_200, respectively. 2497 functional descriptions, 1,279 gene abbreviations, and 820 EC numbers were assigned in the WSH-001 genome by function annotation. Among them, the genes for the 2-KLG synthesis pathway from L-sorbose were annotated; four genes encoding L-sorbose/L-sorbosone dehydrogenase, responsible for converting L-sorbose to L-sorbosone, are highly homologous to ssdA1 (AB092515), ssdA2 (AB092516), ssdA3 (AB092517), and ssdB (AB092518) of K. vulgare strain DSM 4025. The sndh gene that is responsible for the conversion of L-sorbosone to 2-KLG was located in plasmid pKVU_200 (Liu et al. 2011b). It seems that the genome of K. vulgare (∼2.7 M) is much smaller than other common bacteria. Function annotation indicated that the K. vulgare lacks of most of the genes or gene clusters for biosynthesis of many kinds of amino acids, nucleotides and cofactors.

Based on the release of genome sequences for K. vulgare and B. megaterium, and the integration of other -omics data, the understanding of the mechanisms in the mutualism were further illuminated (Liu et al. 2011b, c; Xiong et al. 2011). Based on the global pathway analysis, Zhang et al. (2011) reported most of the essential amino acids for the dependent growth of K. vulgare and developed a definite culture medium for both K. vulgare growth and 2-KLG production (Zhang et al. 2011). Furthermore, Liu et al. (2011a) reconstructed the biosynthesis pathways for amino acids in K. vulgare and found out many gaps in the amino acid metabolic pathways, and consequently developed a novel strategy with gelatin supplement, which contains the several kinds of key amino acids that could not be de novo biosynthesized in K. vulgare, and significantly enhanced the K. vulgare growth and 2-KLG productivity.

Zhou et al. (2011) found that the microorganisms interact by exchanging a number of metabolites with time-of-flight mass spectrometry. Both intracellular metabolism and cell-cell communication via metabolic cooperation were essential in determining the population dynamics of the ecosystem. The contents of amino acids and other nutritional compounds in K. vulgare were rather lower in comparison to those in B. megaterium, but the levels of these compounds in the medium surrounding K. vulgare were fairly high, even higher than in fresh medium. Erythrose, erythritol, guanine, and inositol accumulated around B. megaterium were consumed by K. vulgare upon its migration. The oxidization products of K. vulgare, including 2-KLG, were sharply increased. Upon co-culturing of B. megaterium and K. vulgare, 2,6-dipicolinic acid (the biomarker of sporulation of B. megaterium), was remarkably increased compared with those in the monocultures. Therefore, the interactions between B. megaterium and K. vulgare were a synergistic combination of mutualism and antagonism.

3 Metabolic Engineering Based on the Two-Step Vitamin C Fermentation Process

The classical two-step fermentation process is the most successful route for vitamin C production for its high yield of 2-KLG on D-sorbitol. Though there are two fermentation process, the yield of L-sorbose on D-sorbitol and the yield of 2-KLG on L-sorbose could achieve to more than 99.5 and 97%, respectively. Few of the industrial process could achieve this level. Therefore, the metabolic engineering on the classical two-step fermentation process is always undergoing.

3.1 Interruption of the Downstream Metabolism Pathways

It has already been reported that the 2-KLG synthesized by K. vulgare could be further degraded into iduronic acid by 2-KLG reductase in the bacteria. Therefore, the earliest metabolic engineering for the classical two-step fermentation process is to knockout the gene that encodes the 2-KLG reductase. Saito et al. (1997) obtained a K. vulgare strain NB6939 (it was named as G. oxydans in the original literature) with interrupted L-iduronic acid by mutagenesis using N-methyl-N’-nitro-N-nitrosoguanidine, the 2-KLG concentration of the resultant strain was improved to 31 g/L, which was 200% of the wild-type strain.

3.2 Enhancement of the Cofactor Metabolism Pathway

Conversion of D-glucose or D-sorbitol to 2-KLG or vitamin C involves a number of dehydrogenases that need different kinds of cofactors. These dehydrogenase include D-sorbitol dehydrogenase (SLDH), L-sorbose dehydrogenase (SDH), L-sorbosone dehydrogenase (SNDH), 2,5-diketo-D-gluconic acid reductase (2,5-DKGR), 2-keto-L-gulonic acid reductase (2-KLGR). The SDH needs FAD or Pyrroloquinoline Quinone (PQQ) as cofactor in different strains. SNDH from G. melanogenus UV10 use NAD(P) as cofactor. 2,5-DKGR in Corynebacterium ATCC 31090 (Anderson et al. 1985) and 2-KLGR in K. vulgare (former G. oxydans) (Jiang et al. 1997) use NAD(P)H as the cofactor.

Therefore, rational regulation of the existence, concentration or preference of cofactors could improve 2-KLG production. Banta et al. (2002) showed that there were two kinds of 2,5-DKGR in Corynebacterium sp., i.e., 2,5-DKGR A and 2,5-DKGA B. The 2,5-DKGR A could use both NADH and NADPH as cofactor. However, the dissociation constant (K d ) of the enzyme with NADH is 260 times of that with NADPH as cofactor. Therefore, the 2,5-DKG A tend to use NADPH as cofactor. However, the NADH concentration is three times higher than the NADPH and has higher stability. Sanli et al. decreased the K d of 2,5-DKGR A with NADH to 1/3 compared to the wild-type enzyme using double mutagenesis of F22Y and A272G (Sanli et al. 2001). The tetra-mutagenesis of the 2,5-DKGR A (F22Y/K232G/R238H/A272G) decreased the K d to only 1/4 of the wild-type enzyme (Banta et al. 2002). This will significantly decreased the cost for 2-KLG production by 2,5-DKGR A in vitro.

Primary analysis showed that the K. vulgare could not synthesize folate, which referred to a series of folic acid derivatives that is distinguished by the state of oxidation, one-carbon substitution of the pteridine ring, and the number of glutamate residues (Sybesma et al. 2003). Folate is a kind of important cofactor, which donates one-carbon compounds in many reactions, and is crucial in the de novo biosynthesis of amino acids, purines, and pyrimidines (Sybesma et al. 2003). Leduc et al. (2004) showed that the derivatives of folate could enhanced the cell growth and 2-KLG production of K. vulgare LMPP-20356, while the vitamins, glutathione, hemin, PQQ, pyrimidines could not achieve the same results. By introducing a folate biosynthesis genes cluster found in Lactococcus lactis MG1363, which encodes enzymes catalyzing reactions involved in the assembly of folate and its various derivatives, into K. vulgare, both the cell density and 2-KLG production by the classical two-step fermentation process were significantly enhanced (Cai et al. 2012).

4 One-Step Vitamin C Fermentation Process

Previously, researchers from both industry and academia have invariably noted the inherent disadvantages of the two-step fermentation vitamin C production process, such as long period fermentation, additional sterilizing, control of the mix-culture system, which lead to increased consumption for both raw materials and energy. Compared to other similar large block fermentation products, the manufacturing cost for vitamin C is much higher than others according to its final concentration of more than 100 g/L. The higher cost is mostly caused by its two-step fermentation process.

Therefore, if the vitamin C or its direct precursor, 2-KLG, could be produced by one single step fermentation process, the final cost of the vitamin C will be significantly decreased. Current research on the one-step fermentation process is focused mainly on the following three aspects: (1) Classical two-step based one-step fermentation process for 2-KLG production; (2) Innovative two-step based one-step fermentation process 2-KLG; (3) Direct production of vitamin C from glucose.

4.1 Classical Two-Step Based One-Step Fermentation Process

The K. vulgare contains the genes and proteins for the sdh and sndh, which could convert the L-sorbose to 2-KLG. The cofactor requirements of the sdh and sndh is somehow confused. However, it could be confirmed that at least one of the dehydrogenase is PQQ dependent. The PQQ is a kind of cofactor that was discovered by Hauge as the third redox cofactor after NAD+ and FAD+ in bacteria (Hauge 1964). Anthony and Zatman (1967) found the unknown redox cofactor in alcohol dehydrogenase and named it methoxatin (Anthony and Zatman 1967). In 1979, Salisbury and colleagues as well as Duine and colleagues extracted this prosthetic group from methanol dehydrogenase of methylotrophs and identified its molecular structure (Salisbury et al. 1979; Westerling et al. 1979). The existence of PQQ in Acetobacter sp. was first reported by Adachi’s group (Ameyama et al. 1981).

Because the G. oxydans has sldh gene and the complete PQQ de novo biosynthesis and regeneration genes, if the sdh and sndh from K. vulgare was introduced into the G. oxydans, the resultant G. oxydans strain should produce 2-KLG from D-sorbitol by one-step fermentation process. Early research had already found out that the G. oxydans T100 could produce 2-KLG from sorbitol directly. However, both the final 2-KLG concentration and yield of 2-KLG on sorbitol was relatively low; 7.0 g/L and 13.1%, respectively. Overexpression of the sorbitol dehydrogenase gene sldh and sorbitone dehydrogenase gene sndh from G. oxydans T100 into the strain G. oxydans G624 with expression vector pSDH155 could significantly improve the 2-KLG production (Fig. 12.3). The final concentration of 2-KLG and the yield of 2-KLG on sorbitol was improved to 16 g/L and 30%, respectively (Saito et al. 1997). Furthermore, by inhibition of the L-idose pathway in G. oxydans by chemical mutagenesis and the replacement of tufB1 promoter from E. coli, the recombinant strain could finally produce 2-KLG with a concentration of 88 g/L and a yield of 2-KLG on sorbitol of 82.6% (Saito et al. 1998).

Fig. 12.3
figure 3

Classical two-step based one-step fermentation process (*Recombinant strains with different dehydrogenase). The D-glucose was hydrogenated to form D-sorbitol. The D-sorbitol was directly converted into 2-KLG with a G. oxydans strain or a E. coli strain (or any other potential strains) with the three different kinds of dehydrogenases. The 2-KLG was then esterified and lactonized to form vitamin C

Full size image

G. oxydans is a kind of Gram-positive bacteria and lacks of investigation in the genetic engineering for both multi-gene overexpression and gene knockout. Another choice is to metabolic engineer other model microorganisms, such as E. coli or S. cerevisae. Because most of the dehydrogenase genes are located on the cell membrane, the difference between prokaryotes and eukaryotes makes S. cerevisiae a poor choice. Therefore, the aim would be to express all of the dehydrogenase and PQQ metabolism related pathways in E. coli for one-step biosynthesis of 2-KLG from D-sorbitol. Previous reports had already achieved the de novo biosynthesis of PQQ in E. coli by overexpression of the PQQ biosynthesis gene cluster pqqABCDE from G. oxydans (Yang et al. 2010). However, the regeneration system of the PQQ is still not well elucidated. Therefore, current works are mainly focused on the bacteria that has its own PQQ metabolism system, such as G. oxydans and Paracoccus denitrificans (Xia et al. 2003). To solve this problem, it is essential to elucidate the PQQ regeneration system on gene level.

However, after the report of the innovative two-step fermentation process, the research on the classical two-step based one-step fermentation process seems to be suspended. Few literatures about metabolic engineering of G. oxydans for one-step vitamin C production could be found after then.

4.2 Innovative Two-Step Based One-Step Fermentation Process

Though the classical two-step fermentation process has got highly promoted results, it still needs D-sorbitol as substrate, which need additional hydrogenation step by chemical process. Therefore, the direct fermentation of D-glucose to 2-KLG is still presumed. There have been some attempts to add an additional step that could convert the D-glucose to D-sorbitol, in order to supplement the classical two-step fermentation process. However, the lack of efficient enzyme makes this impossible.

Therefore, an innovative two-step fermentation process were discovered to resolve this problem. In the new innovative two-step fermentation process, the D-glucose was converted into 2,5-diketo-gluonic acid (2,5-DKG) by Erwinia herbicola ATCC 21988 or some similar strains with glucose dehydrogenase, gluconate dehydrogenase and 2-keto-D-gluconate dehydrogenase. Then the 2,5-DKG could be transformed into 2-KLG by Corynebacterium ATCC 31090 or other similar strains with 2,5-DKG reductase. With this innovative two-step fermentation process, the D-glucose could be transformed into 2-KLG within two steps (Anderson et al. 1985) (Fig. 12.4). However, 2,5-DKG is highly unstable during the fermentation process. Additional sterilization could completely destroy this chemical.

Fig. 12.4
figure 4

Innovative two-step based one-step fermentation process. The D-glucose was converted into 2,5-DKG by Erwinia sp. The 2,5-DKG was converted into 2-KLG by Corynebacterium sp. The 2-KLG was then esterified and lactonized to form vitamin C. The innovative two-steps fermentation process need not chemical hydrogenation process

Full size image

Based on this innovative two-step fermentation process, Anderson et al. (1985) has expressed a 2,5-DKG reductase from Corynebacterium ATCC 31090 into the Erwinia herbicola ATCC 21988 with the expression vector ptrp1-35. The final recombinant E. herbicola strain could produce 1 g/L of 2-KLG from saturated D-glucose solution. By the protoplast fusion, Lin et al. (1999) fused an Erwinia herbicola and a Corynebacterium strain. The resultant strain could produce 2.07 g/L of 2-KLG (Fig. 12.5).

Fig. 12.5
figure 5

Innovative two-step based one-step fermentation process. The D-glucose was directly converted into 2,5-DKG by a metabolic engineered Erwinia sp. strain containing a 2,5-DKGR from Corynebacterium sp. The 2-KLG was then esterified and lactonized to form vitamin C

Full size image

4.3 Direct Production of Vitamin C from D-Glucose

The vast majority of animals and plants are able to synthesize vitamin C through a sequence of four enzyme-driven steps, which convert D-glucose to vitamin C (Fig. 12.6). In eukaryotes, vitamin C biosynthesis occurs via two highly divergent biochemical pathways. In rats and other mammals, glucose is converted into L-gulono-1,4-lactone via a complex pathway involving uronic acid derivatives and the ‘inversion’ of the original carbon skeleton (Isherwood and Cruickshank 1954; Touster 1962). The widespread occurrence of this pathway in the animal kingdom is deduced by the ubiquity of its terminal enzyme L-gulono-1,4-lactone oxidase (GUL oxidase). The gene coding for this enzyme is mutated in humans and some other animals, resulting in the loss of vitamin C synthesizing capacity.

Fig. 12.6
figure 6

Vitamin C production in higher eukaryotes

Full size image

Vitamin C biosynthesis generally exists in most of the plant cells (Zhang et al. 2007). It plays a crucial role in the detoxification of peroxide, ozone and free radicals and is essential for photosynthetic activity via the regeneration of membrane-soluble antioxidants (α-tocopherol), zeaxanthin and the pH-mediated modulation of PS II activity. Plant cells can accumulate large amount of vitamin C, particularly in green tissues and storage organs (Smirnoff 2000). Because plant cells have the whole pathway for the synthesis of vitamin C from D-galactose, attempts have been made for exploiting plant cell or microalgae cultures for the commercial production of vitamin C via fermentation.

Running et al. (2002) isolated Chlorella pyrenoidosa strains that were capable of producing over 70-fold more vitamin C than wild-type strains (0.64 mg/g dry cell weight for wild-type compared with 45 mg/g dry cell weight for the most-productive mutants) (Running et al. 2002). These improvements were achieved through repeated rounds of chemical mutagenesis and optimization of fermentation process conditions. Further claims were made regarding a method for the extracellular stabilization of vitamin C by reduction of medium pH (Running et al. 2002), resulting in the majority of vitamin C becoming harvestable in the fermentation medium. Yields of vitamin C microalgal cultures are not yet competitive with the traditional processes and it is likely that full economic exploitation of these systems for one-step fermentation strategies for the synthesis of vitamin C will require genetic engineering of the pathway. However, progress in the definition of the vitamin C biosynthetic pathway in plants has been singularly slow. Original models for the biosynthetic pathway in plants were based on the animal model (Isherwood and Cruickshank 1954). However, plant cells do not contain GUL oxidase and do not invert the glucose carbon skeleton during its conversion into vitamin C (Smirnoff 2000). Over the past years several pathways were proposed to accommodate these observations, but direct evidence of operation of any of these pathways has not been obtained. Only recently, with the identification of vitamin C-deficient Arabidopsis mutants (Conklin et al. 1996), the pathway in plants has been resolved (Wheeler et al. 1998). This made the direct fermentation production from glucose or even starch or cellulose ­achievable. However, because most of the genes from plants or animals had weaker reactive dynamics characteristics, how to improve the productivity by involving these pathways from higher organisms is the key bottleneck for this route.

Furthermore, though the direct production of vitamin C from D-glucose seems to be an ideal process, vitamin C is renowned for its instability and susceptibility to oxidation. Therefore, a ventilation fermentation process may destroy most of the vitamin C. Some vitamin C derivatives, such as 2-O-alpha-D-glucopyranosyl-L-ascorbic acid (AA-2G), are relatively stable and perform almost identical physiological functions. In consequence, AA-2G is now widely used in luxury cosmetics and some of beverages. It can survive for a longer period than in subjects administered ascorbic acid 2-phosphate, which is a conventional vitamin C derivative. Therefore, the simultaneously glycosylation of vitamin C during the fermentation process may be a direct route for vitamin C production.

5 Perspectives

Based on the current reports, though there are many researches on the strain screening and mutagenesis, process optimization, relationships between B. megaterium and K. vulgare, metabolic engineering of K. vulgare, development of new two-step fermentation process to avoid mix-culture of B. megaterium and K. vulgare, development of one-step fermentation process based on both classical two-step fermentation process on new two-step fermentation process, the significant proceedings in the new processes are still sought for industrial applications for different reasons. It seems that evolution of the classical two-step fermentation process of vitamin C seems to have encountered many problems.

Based on current results, the further development of vitamin C production by microorganisms with metabolic engineering should be focused on the following area: (1) Demonstration of mutual relationships between the B. megaterium and K. vulgare and therefore enhancing cell growth and 2-KLG production by K. vulgare; (2) Elucidation of the genes involved in the biosynthesis and transportation of 2-KLG and its intermediate metabolites in both classical two-step and new two-step fermentation steps; (3) Investigation of the mechanisms of PQQ-dependent incompletion oxidation dehydrogenases and the regeneration of PQQ; (4) Discovery of efficient metabolic pathways for the direct fermentation of vitamin C and even its stable derivatives.