2.1.2. Exploring alternative carbon resource for GFAA production

In addition to glucose, other alternative carbon feedstocks, including xylose, arabinose, glucosamine (GlcN), N-acetyl-glucosamine (GlcNAc), sucrose, molasses, and glycerol, have been intensively applied to produce GFAAs (Fig. 4) [3]. Xylose, which is a cheap and attractive feedstock, has been applied to produce l-ornithine [113,204] and GABA [7,187]. Overexpression of the xylAB operon (encoding xylose isomerase and xylulokinase originating from Xanthomonas campestris) by employing a recombinant plasmid pXMJ19-xylAB generates the recombinant strain C. glutamicum SO29, which produces l-ornithine at a titer of 18.9 g/L [204]. Heterologous expression of E. coli gadB (encoding glutamate decarboxylase, GAD) and xylAB resulted in engineered C. glutamicum H36GD1852 that produced 35.47 g/L GABA during batch culturvation [7]. Similarly, the GABA-producing strain was engineered to utilize alternative carbon sources, such as GlcN and GlcNAc. The uptake of GlcN was identified to be transported by the PTS^Glc system and repressed by NanR (encoded by nanR) [179]. Seventeen Corynebacterium strains were tested for possibility to grow on GlcNAc, which revealed that only Corynebacterium glycinophilum DSM45794 harboring an EII permease for the assimilation of GlcNAc [110]. Deletion of NanR and overexpression of nagE in C. glutamicum generated the mutant strain GABA6C, which produced 3.9 g/L of GABA [72]. Arabinose is a pentose that comprises much of rice straw hydrolyzate. By expressing araBAD from E. coli MG1655, C. glutamicum strains were engineered to produce l-glutamate, l-ornithine, and l-arginine at a titer of 37 ± 10 mM, 124 ± 35 mM, and 25 ± 2 mM, respectively, from arabinose [148]. Sucrose and molasses, as industrial raw materials, are highly attractive for the fermentative production of bio-products because they are inexpensive and readily available [84,140]. Sucrose-6-phosphate hydrolase (encoded by sacB) in the genome of C. glutamicum can hydrolyze sucrose to form fructose and glucose-6-phosphate [33]. β-Fructofuranosidase (encoded by sacC), an extracellular enzyme from Mannheimia succiniciproducens is favorable for engineering microorganisms, has been successfully applied to the E. coli K-12 strain to produce l-threonine [88]. Recombinant C. glutamicum harboring the sacC gene from M. succiniciproducens was constructed and produced 22.0 g/L and 27.0 g/L of l-ornithine from sucrose and molasses, respectively [216]. Recently, with the rapid development of the biodiesel industry, more than 1.6 million tons of waste glycerol has been generated per year [114]. To utilize glycerol in C. glutamicum, coexpression of the glpFKD operon from E. coli were performed in C. glutamicum WT (pVWEx1-glpFKD^Eco), C. glutamicum ORN1(pVWEx1-glpFKD^Eco) and C. glutamicum ARG1(pVWEx1-glpFKD^Eco), which produced l-glutamate, l-ornithine, and l-arginine at yields of 7.2 ± 0.2 mM, 17.9 ± 0.4 mM, and 23.6 ± 0.9 mM, respectively [114].

Open in a separate window

Fig. 4

The utilization of alternative carbon sources to produce GFAAs. nagB, encoding glucosamine 6-phosphate deaminase; nagA, encoding N-acetyl-glucosamine 6-phosphate deacetylase; nanR, encoding transcriptional regulator; nanE, encoding N-acetylglucosamine-specific transporter; glpF, encoding glycerol facilitator; glpK, encoding glycerol kinase; glpD encoding glycerol-3-phosphate dehydrogenase; xylA, encoding xylose isomerase; xylB, encoding xylulokinase; araA, encoding arabinose isomerase; araB, encoding ribulokinase; araD, encoding ribulose-5-phosphate-4-epimerase; sacC, encoding β-fructofuranosidase.

In summary, slow glucose uptake rate frequently leads to a decrease in GFAA productivity [107]. Manipulating the glucose uptake system is a feasible strategy to improve GFAA production. In addition, it is emphasized that using glucose in fermentation industry has the problem of competing with the food industry and human nutrition. To avoid this competition, xylose, arabinose, sucrose, GlcNAc, and glycerol may serve as environmentally friendly and renewable natural carbon resources for the biotechnological production of GFAAs.

2.2.2. Metabolic engineering of PPP and EMP for improving GFAA production

It is widely acknowledged that flux distribution in the EMP and the PPP plays a crucial role in maintaining the balance of cellular metabolism. Under l-glutamate-producing conditions, the carbon metabolic flux distribution in the EMP and PPP was estimated at 80:20, whereas it was 30:70 and 40:60 during l-arginine and l-lysine fermentation, respectively [78]. Channeling the metabolic flux to PPP can directly provide more NADPH for the biosynthesis of GFAAs. For instance, coexpression of zwf and gnd in the strain CgΔargG resulted in 287% and 363% improvement in intracellular NADPH content, which stimulated l-citrulline accumulation [102]. In addition, optimization of intracellular NADPH supplementation by downregulating the expression of pgi and upregulating the expression of the tkt operon is a feasible strategy to improve l-arginine production (as shown in Fig. 6) [129]. Similar strategies have been applied for the development of l-ornithine and l-proline-producing strains. Replacing native start codons of pgi and native promoter of tkt operon resulted in an engineered strain that produced l-ornithine at a yield of 51.5 g/L, with an overall productivity of 1.29 g/L/h [77]. Simultaneously, modulation of zwf and gnd to improve the supplementation of NADPH is the preferred genetic engineering strategy for l-proline production [191].

Open in a separate window

Fig. 6

Genetic modification of EMP and PPP to produce GFAA in C. glutamicum. gapC, encoding NADP-dependent glyceraldehyde-3-phosphate dehydrogenase; rocG, encoding NADH-dependent glutamate dehydrogenase; G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; FBP, fructose-1,6-bisphosphate; DHAP, Dihydroxyacetone phosphate; GA3P, glyceraldehyde-3-phosphate; 1,3BPG, 1,3-bisphosphate-glycerate; 3 PG, 3-phosphate-glycerate; 2 PG, 2-phosphate-glycerate; PEP, phosphoenolpyruvate; Pyr, pyruvate; 6PGL, gluconolactone-6-phosphate; 6 PG, gluconate-6-phosphate; Ru5P, ribulose-5-phosphate; Xu5P, xylulose-5-phosphate; R5P, ribose-5-phosphate; S7P, sedoheptulose-7-phosphate; E4P, erythrose-4-phosphate; α-KG, 2-oxoglutarate; Glu, l-glutamate.

Overexpression of enzymes in glycolysis for optimization of carbon metabolic velocity and ATP supply has been applied in the construction of engineered strains (Fig. 6). For instance, insertion of a strong P_eftu promoter into the upstream region of pfkA and gapA resulted in engineered strains SO7 and SO9 that produced 26.5 and 22.8 g/L of l-ornithine, which was 11.2% and 9.7%, respectively, higher than the parent strain. In addition, to conjugate the glycolytic pathway with the generation of NADPH, gapC from Clostridium saccharobutylicum was introduced into C. glutamicum strainΔAPER, which showed a 277% increase in intracellular NADPH content and a 10% improvement in l-ornithine production titer [70]. This useful strategy was further confirmed by its application in C. glutamicum strain KBJ11, which produced 88.26 g/L of l-ornithine during fed-batch culturvation [27].

In addition to EMP and PPP, the gluconate bypass pathway frequently consumes the mid-metabolite, which reduces the generation of NADPH. Therefore, approaches aimed at the gluconate bypass pathway have been used to improve the NADPH pools for GFAA production. Deletion of gntK enables the absence of gluconate kinase activity in strain SJC8039, which showed a 51.8% increase in intracellular NADPH content and a 49.9% increase in the yield of l-ornithine [58]. Inactivation of glucose dehydrogenase by chromosomal in-frame deletion of NCgl0281, NCgl2582, and NCgl2053 led to the improvement in specific G6DH and 6-PGD activities, which promoted the supplement of intracellular NADPH and l-ornithine accumulation [59].

In summary, metabolic engineering of EMP and PPP to increase NADPH availability and metabolic velocity, has been applied to four categories to improve GFAA production: (1) Redistribution of the carbon flux between the EMP and PPP, (2) enhancing the expression of key enzymes involved in energy metabolism, (3) blocking the gluconate-bypass pathway, and (4) the heterologous expression of NADP -or NAD-dependent enzymes. These strategies increase intracellular NADPH supplementation, which is universally applicable to anabolism.

5.1.2. Modification strategies for developing l-ornithine, l-citrulline, and l-arginine-producing strains

l-arginine and its derivatives, l-ornithine and l-citrulline, possess the same anabolic pathways in C. glutamicum and C. crenatum, indicating that uniform modification strategies can be applied to the development of the corresponding strains. The typical modification strategies related to anabolic pathways of l-ornithine, l-citrulline, and l-arginine for the construction of engineered strains are composed of four sections as follows: a. Removal of feedback inhibition of l-arginine on NAGK; b. removal of feedback repression of ArgR and FarR; c. overexpression of argGH or argCJBDFR operons; and d. blocking the competing and degradation metabolic pathways.

• (1)

  Modification strategies focused on the anabolic pathway for l-arginine production

NAGK is a key rate-limiting enzyme that is inhibited by l-arginine in C. glutamicum. To obtain l-arginine high-producer strains, the removal of feedback inhibition through site-directed mutagenesis is a precondition. To obtain NAGK with high thermostability and specific activity, the variant NAGK EH3 (including E19Y, I74V, F91H, and K234T) was generated and introduced into C. crenatum SYPA5-5, which resulted in the recombinant strain SYPA-argB_EH3. The 96-h fermentation indicated that l-arginine production reached 61.2 g/L, represents a 27.9% increase as compared with the control strain SYPA5-5 [213]. In addition, NAGK M4 with substitutions D311R, D312R, E19R, and H26E was generated and integrated into the chromosome of C. crenatum MT, resulting in a 26.2% increase in l-arginine production [206]. Furthermore, l-arginine-insensitive NAGK with substitutions H268 N or R209A [193] was introduced into the genome of C. crenatum SYPA5-5, resulting in the recombinant strains SYPA5-5-CcNAGK_H268N and SYPA5-5-CcNAGK_R209A. Compared with the initial CcNAGK, the specific activities of CcNAGK_H268N and CcNAGK_R209A remained at 98.2% and 94.3% respectively in the presence of 15 mM l-arginine, whereas the NAGK enzyme activity in the control group declined to zero [217]. These studies suggested that feedback inhibition of CcNAGK could be removed by site-directed mutagenesis, which is a crucial strategy for developing l-arginine-producing strains.

ArgR and FarR are global repressor proteins in C. glutamicum and are responsible for the transcription inhibition for arg operon expression. Hence, the deletion of argR and farR is an appropriate approach for improving l-arginine accumulation. For instance, the C. glutamicum strain AR2 was formed by the double inactivation of argR and farR, based on strain AR1. The results of fed-batch fermentation indicated that the AR2 strain produced 61.9 g/L l-arginine, which was 82.6% higher than that produced by the control strain (33.9 g/L) [129]. Similarly, ArgR and FarR repressors were inactivated by removing argR and farR from the genome of C. glutamicum SNK 118, which generated the C. glutamicum strain JML05. After 72 h fed-batch cultivation, the l-arginine production titer of mutant strain JML05 reached up to 51.96 g/L, represents a 25.2% increase as compared with the parent strain [203]. Deletion of the above two repressors significantly improved l-arginine accumulation, indicating that stable expression of the arg operon promotes l-arginine anabolism. Overexpression of argCJBDFR and argGH operons is a common conventional strategy that is widely applied in l-arginine-producing strains. For instance, the C. glutamicum strain AR6 was constructed by replacing the native promoter of argCJBDFR. Fermentation results showed that AR6 produced 92.5 g/L l-arginine with a yield of 0.40 g/g, which are 12.8% and 14.3% higher than those of the control strain AR5 [129]. In addition, the argGH operon was overexpressed in C. crenatum strain H-7 by recombinant plasmid pDXW-10-argGH, which resulted in the mutant strain H-7 GH. Then, the mutant strain produced 67.92 g/L l-arginine after 72 h fed-batch cultivation [217].

To further optimize l-arginine biosynthesis, increasing the supplementation of carbamyl phosphate facilitates the conversion of l-ornithine into l-citrulline. Carbamyl phosphate is converted from l-glutamine catalyzed by carbamoyl phosphate synthetase (encoded by carAB). Hence, overexpression of carAB by insertion of a P_sod promoter in C. glutamicum strain AR4 improved the biosynthesis of carbamyl phosphate, which resulted in the engineered strain AR5. The titer of l-arginine reached up to 82 g/L with a yield of 0.35 g/g by fed-batch cultivation of strain AR5, which was 14.3 and 12.9% higher than that of the start strain AR4 [129].

• (2)

  Modification strategies focused on anabolic pathways for improving l-ornithine production

l-ornithine is an intermediate metabolite of the urea cycle, which can be catalyzed to form l-citrulline and l-arginine by ornithine carbamoyltransferase, argininosuccinate synthase, and argininosuccinate lyase. Thus, blocking the urea cycle by deletion of argF is necessary for the biosynthesis of l-ornithine. For instance, inactivation of argF in C. glutamicum S9114 generates the mutant strain Sorn1 that produced l-ornithine up to 7.97 g/L, representing 16-fold increase as compared to the parent strain (0.46 g/L) [209]. Similarly, ornithine carbamoyltransferase was inactive in C. glutamicum ATCC13032, which generated the mutant strain SJ8000, which produced approximately seven-fold more l-ornithine than the control strain [56]. Although l-ornithine-producing strains can be constructed by inactivation of ArgF, the accompanying problem of l-arginine auxotrophy requires exogenous addition of l-arginine, which not only increases the cost of fermentation production, but also leads to feedback inhibition of NAGK [210]. To avoid this drawback, attenuating the expression of argF, achieved by insertion of a T4 terminator, is an effective approach. Thus, C. glutamicum strain CO-9 was successfully constructed on the basis of strain CO-1, which not only downregulated the expression of argF, but also improved l-ornithine by 42.8% as compared with strain CO-1 harboring argF deletion [210].

Similar to l-arginine, the enzymes that catalyze the conversion of l-ornithine are coexpressed as an argCJBDFR operon, which is repressed by ArgR. Therefore, argR is deleted in almost all l-ornithine-producing strains [218]. Typically, C. glutamicum strain SJ8039 was constructed via deletion of argR, which increased the production titer of l-ornithine from 6.78 to 9.65 mg/g dry cell weight [56]. Additionally, deletion of argR in C. glutamicum strain ΔAPE6937 generated the mutant strain ΔAPE6937R42. The l-ornithine production titer of this mutant strain reached up to 17.3 ± 0.4 g/L, which was 27% higher than that of the control group [69]. Similar to the genetic modification of higher l-arginine-producing strains, enhancement of the l-ornithine biosynthesis pathway is an effective approach to improve l-ornithine production. Currently, two approaches are used to reinforce the l-ornithine biosynthesis pathway. The biosynthesis of l-ornithine could be enhanced by overexpression of the endogenous argCJBD operon. In particular, stable expression of argJ is the crucial step for the biosynthesis of l-ornithine in C. glutamicum. Thus, plasmid-based overexpression of argJ was introduced into strain 1006ΔargR, generating the mutant strain 1006ΔargR-argJ. During shake flask cultures, this mutant strain produced 31.6 g/L l-ornithine, suggests a 11.7% increase as compared to the initial strain [218]. In addition, cg3035 has a catalytic function in glutamate acetylation, which is similar to argJ [132]. Overexpression of the cg3035 gene by insertion of the P_tac strong promoter increased the yield of l-ornithine by 12.6% [208]. Furthermore, plasmid overexpression of ncgl0462, encoding a class II aminotransferase, promotes l-ornithine production in C. glutamicum [75]. Coexpression of ncgl0462 and mutated argCJBD genes in the C. glutamicum strain SJC8514 resulted in the recombinant strain SJC8514 (pEK-CJBD_mut). The fermentation results showed that the titer of the l-ornithine recombinant strain SJC8514 (pEK-CJBD_mut) was 12.48 g/L, exhibits 22.7% increase than that of the control group [75]. On the other hand, introducing an artificial “linear” pathway into C. crenatum can significantly promote l-ornithine accumulation. The propitious introduction of this novel artificial “linear” pathway requires highly active antifeedback inhibition of NAGS. After the experimental screening, heterologous expression of argA from E. coli and argE from Serratia marcescens was adopted. The fermentation results showed that the resulting strain Cc-QF-4 allowed the production of 40.4 g/L of l-ornithine, which was 48.1% higher than that obtained with the initial strain Cc-QF-1 (23.5 g/L) [161].

Although the inactivation of argF can block the conversion from l-ornithine into l-arginine and prevent the intracellular accumulation of l-arginine, supplementation of l-arginine in the medium is required for cell growth, which causes feedback inhibition of NAGK. Thus, l-arginine-insensitive NAGK, derived from C. glutamicum ATCC 21831 with the M64V mutation in argB, was introduced into the l-ornithine-producing strain, which significantly promotes l-ornithine accumulation [33,41]. Similarly, overexpression of NAGK, harboring E19R, H26E, and H268 N mutation, in strain ORN1 reached the production of 12.1 ± 0.5 g/L l-ornithine in CgXII medium with 4% glucose, which was 6.1% higher than that accumulated by ORN1(pEKEx3-argB) in the same medium [67].

• (3)

  Modification strategies focused on anabolic pathways for l-citrulline production

l-citrulline is a mesostate of the l-arginine anabolic pathway. Theoretically, the genetic modification strategy for the construction of l-ornithine-high-producing strains can be completely applied to l-citrulline, and the genetic modification for l-citrulline-high-producing strains was as follows: First, blocking the conversion of l-citrulline into l-arginine. To avoid the conversion of l-citrulline into l-arginine, deletion of argG (encoding argininosuccinate synthetase, AS) is an essential measure for accumulation of l-citrulline [46,102]. For instance, argG was deleted in C. glutamicum ATCC13032, generating the recombinant strain CIT1, which produced 2.52 ± 0.10 g/L l-citrulline, representing 16.8-fold increase in the yield compared to the initial strain [46]. In addition, inactivation of the repressor ArgR not only increases l-ornithine and l-arginine production titer, but also improves l-citrulline production titer [46]. Deletion of argR in the mutant strain CIT1 resulted in the production of 5.43 ± 0.16 g/L l-citrulline, which was two-fold higher than that produced by the initial strain [46]. Furthermore, feedback inhibition is removed. The activity of the enzyme OATase (encoded by argJ) decreased by 50% in the presence of 30 mM l-citrulline, indicating that OATase was feedback inhibited by l-citrulline [46]. To facilitate the biosynthesis of l-citrulline, argJ was overexpressed in C. glutamicum, which improved the yield of l-citrulline by 57.2% [46].

In summary, genetic modifications, including removal of feedback repression by argR inactivation, the release of feedback inhibition by NAGK mutation, and enhancement of the anabolic pathway by arg operon overexpression, are universal strategies for developing l-arginine, l-ornithine, and l-citrulline (Fig. 12). The accumulation of l-ornithine and l-citrulline as intermediate metabolites of the urea cycle requires the deletion of downstream metabolic pathways. However, in practice, the “blocking” strategy results in insufficient biosynthesis of l-arginine, thus affecting cell growth, which requires excessive supplementation of l-arginine, which does not only cause feedback inhibition, but also increases the fermentation cost. To solve this problem, attenuating the expression of argF or argG by insertion of a terminator or screening for a feedback-resistance enzyme has been proven to be a feasible approach for improving l-ornithine or l-citrulline production.

Open in a separate window

Fig. 12

Modification strategies for anabolic pathway of l-ornithine, l-citrulline, and l-arginine. carAB, encoding carbamoyl phosphate synthetase; proB, encoding γ-glutamyl kinase; ncgl1469, encoding Ncgl1469 displaying N-acetylglutamate synthase; cg3035, encoding Cg3035 displaying N-acetylglutamate synthase; ncgl0462, encoding Ncgl0462 displaying N-acetylornithine aminotransferase activity; argR, encoding a transcriptional regulator; NAGK, N-acetylglutamate kinase; FarR, a transcriptional regulator.

5.2.2. Modification strategies for the anabolic pathway of l-proline production

Based on these metabolic pathways, the corresponding genetic modification strategies were used to improve l-proline accumulation. The first strategy is to remove feedback inhibition of γ-glutamyl kinase. Thus, codon saturation mutation of G149 in γ-glutamyl kinase was carried out using Cpf1-assisted ssDNA recombineering, thereby generating the mutant strain ZQJY-1, which produced 4.47 ± 1.15 g/L l-proline, representing approximately four-fold increase compared to the model strain ATCC 13032 (1.35 ± 0.95 g/L) [212]. To increase the yield of l-proline, feedback-resistant γ-glutamyl kinases with the inducible promoter P_tac or the constitutive promoter P_eftu were processed to promote the biosynthesis of l-proline [212]. The second strategy was to block the degradation pathway of l-proline. Proline dehydrogenase and 1-pyrroline-5-carboxylate dehydrogenase (both encoded by putA) are responsible for catalyzing the conversion of l-proline into l-glutamate [120]. It has been reported that blocking this process can effectively improve l-proline production [214]. To promote the biosynthesis of l-proline, the putA gene was mutated, resulting in strain ZQJY-2, which produced l-proline at a titer of 6.05 ± 0.63 g/L, indicating a 35.3% increase as compared with the control strain ZQJY-1 (4.47 ± 1.15 g/L) [212]. The third strategy indirectly modifies the l-arginine anabolic pathway. It should be noted that this strategy is effectively carried out on the basis of l-ornithine higher producing strain. Overexpression of ocd can accelerate the conversion of l-ornithine into l-proline. For this purpose, heterologous expression of ocd from P. putida resulted in the recombinant strain JJ004 that produced l-proline at a titer of 10.0 ± 0.1 g/L, indicating a five-fold increase compared with the parent strain. Combining it with the overexpression of feedback-resistant NAGK resulted in strain JJ006 that produced l-proline at a yield of 12.7 ± 0.3 g/L—an increase of 27% compared with strain JJ004 [68]. The fourth strategy is the genome-scale metabolic model (GEMM)-assisted genetic modulation. A comprehensive GEMM of C. glutamicum ATCC13032 and iCW773, was applied to predict targets for the construction of high l-proline-producing strains. The predicted targets of iCW773 include the upregulation of ten genes, downregulation of one gene, and deletion of one gene. Then, those relevant targets were selected for genetic modification, generating strain Pro6, which produced 66.43 g/L l-proline during fed-batch fermentation, reaching a yield of 0.26 g/g glucose [68]. In conclusion, these genetic modification strategies are based on the l-proline anabolic pathway and l-arginine anabolic pathway, respectively (Fig. 13). Compared with these genetic modification strategies, modulating the l-proline anabolic pathway merely involves three enzymatic reactions and one spontaneous reaction, which is more economical and feasible for the biosynthesis of l-proline [68]. Simultaneously, to improve l-proline accumulation, genetic manipulations are required to remove feedback inhibition and block the degradation pathway, which enables l-proline production to reach up to 120.18 g/L at the 3 L fermentor level [212].

Open in a separate window

Fig. 13

Anabolic pathways of proline and its modification strategies in C. glutamicum. proB, encoding γ-glutamyl kinase; proA, encoding l-glutamate-5-semialdehyde dehydrogenase; proC, encoding pyrroline-5-carboxylate reductase; putA, encoding Δ1-pyrroline-5-carboxylate dehydrogenase; ocd, encoding ornithine cyclodeaminase.