Construction of a Vitreoscilla Hemoglobin Promoter-Based Tunable Expression System for Corynebacterium glutamicum

Corynebacterium glutamicum is an industrial strain used for the production of valuable chemicals such as L-lysine and L-glutamate. Although C. glutamicum has various industrial applications, a limited number of tunable systems are available to engineer it for efficient production of platform chemicals. Therefore, in this study, we developed a novel tunable promoter system based on repeats of the Vitreoscilla hemoglobin promoter (Pvgb). Tunable expression of green fluorescent protein (GFP) was investigated under one, four, and eight repeats of Pvgb (Pvgb, Pvgb4, and Pvgb8). The intensity of fluorescence in recombinant C. glutamicum strains increased as the number of Pvgb increased from single to eight (Pvgb8) repeats. Furthermore, we demonstrated the application of the new Pvgb promoter-based vector system as a platform for metabolic engineering of C. glutamicum by investigating 5-aminovaleric acid (5-AVA) and gamma-aminobutyric acid (GABA) production in several C. glutamicum strains. The profile of 5-AVA and GABA production by the recombinant strains were evaluated to investigate the tunable expression of key enzymes such as DavBA and GadBmut. We observed that 5-AVA and GABA production by the recombinant strains increased as the number of Pvgb used for the expression of key proteins increased. The recombinant C. glutamicum strain expressing DavBA could produce higher amounts of 5-AVA under the control of Pvgb8 (3.69 ± 0.07 g/L) than the one under the control of Pvgb (3.43 ± 0.10 g/L). The average gamma-aminobutyric acid production also increased in all the tested strains as the number of Pvgb used for GadBmut expression increased from single (4.81–5.31 g/L) to eight repeats (4.94–5.58 g/L).

To engineer C. glutamicum for biochemical production, its metabolic pathway has been manipulated using a plasmid-based expression system, which establishes the synthetic pathway for the production of biochemicals [5]. This is an important step for evaluating the success of the constructed pathway and for establishing the target metabolite production in recombinant strains. This also helps identify which key reactions can be improved in order to drive metabolic flux toward optimized chemical production [1,4,5]. For example, the production of GABA using C. glutamicum strains was established by using a synthetic promoter-based expression vector system (P L26 < P I16 < P H36 ) capable of low, intermediate, and high-strength glutamate decarboxylase (GadB mut ) expression. It was demonstrated that the use of the high-strength promoter P H36 in the pHGmut strain (5.89 ± 0.35 g/L) enabled higher production of GABA than in pIGmut (5.32 ± 0.04 g/L) and pLGmut (4.87 ± 0.15 g/L) strains with expression of GadB mut under intermediate (P I16 ) and low (P L26 ) strength promoters, respectively [26]. Cadaverine production using a synthetic promoter-based expression system (P L10 < P L26 < P I1 6 < P I64 < P H30 < P H36 ) was also evaluated, and it was observed that a high level of protein expression using the strongest promoter (P H36 ) does not necessarily result in the highest level of metabolite production [28,29]. In case of batch fermentation for cadaverine production using recombinant C. glutamicum strains, the use of P H30 (23.8 g/L) for the expression of lysine decarboxylase produced a higher titer than that when the P H36 promoter was used (21.3 g/L) [28]. However, the repertoire of vector systems capable of different levels of protein expression in C. glutamicum is still limited [5]. Most promoters in plasmids derived from E. coli plasmids such as P tac , P trc , P lacUV5 , P R , and P L have been evaluated [4,5]. However, despite the adoption of the promoters from E. coli plasmids, the variety of genetic engineering tools applicable to C. glutamicum is still less compared to the tools available for E. coli [1,4,5]. Therefore, the discovery or creation of a new vector system and evaluation of its capability for tunable protein expression in a target host strain are important factors in developing recombinant strains for the production of biochemicals in biorefineries [4,5].
Recently, the use of five repeats of the P tac promoter successfully enabled stable overexpression of phaCAB genes in recombinant E. coli and enhanced poly(R-3-hydroxybutyrate) (PHB) accumulation 5.6 times that of the control strain, which had a single copy of the P tac promoter [32]. In another report, the use of the promoter from Vitreoscilla hemoglobin protein (P vgb ) was successfully demonstrated in E. coli. It was found that the expression vectors with increasing repeats of the P vgb promoter enabled tunable expression of the PHB synthesis operon (phaCAB) and allowed enhanced accumulation of poly(hydroxybutyrate) (PHB) [33]. The highest accumulation of 90% PHB in 5.37 g/L CDW was achieved by the recombinant strain with eight repeats of the P vgb promoter. The use of the P vgb promoter system was also successfully demonstrated in recombinant C. glutamicum. It was used to investigate the effect of VHb protein expression on L-glutamate and L-glutamine production in the recombinant strains. However, it was observed that the expression of the VHb protein under the P tac promoter was better than that with the use of a single repeat of P vgb promoter (1.71 ± 0.08 nmol/g > 0.69 ± 0.10 nmol/g) [34]. Therefore, in this study, new tunable gene expression systems were developed based on repeats of P vgb . The strength of the P vgb promoter-based expression systems with increasing repeats of P vgb were investigated by evaluating the expression level of GFP in recombinant strains. To demonstrate the use of the new tunable P vgb promoter-based expression system as a novel platform for metabolic engineering of C. glutamicum, 5-AVA and GABA production were established in different strains of C. glutamicum. 5-AVA and GABA production were selected as model compounds, because they are derived from L-lysine and L-glutamate, respectively.

Construction of P vgb Promoter-Based Tunable Expression Systems for C. glutamicum and Evaluation of Green Fluorescent Protein in the Recombinant Strains
To investigate the tunability of protein expression using P vgb in recombinant C. glutamicum, several expression vector systems based on pCES208 plasmids were constructed. GFP was expressed in C. glutamicum under the control of increasing repeats of P vgb (P vgb , P vgb4 , P vgb8 ). The strength of protein expression in the three constructs was investigated by measuring the intensity of fluorescence in the recombinant strains, by using the fluorescent activated cell sorting analysis (FACS) (Figure 1) [35]. The resulting recombinant C. glutamicum EGFPV1, EGFPV4, and EGPV8 strains harbored one (P vgb ), four (P vgb4 ), and eight (P vgb8 ) repeats of P vgb , respectively. C. glutamicum KCTC 1857 was cultured and used as the negative control. Colonies were randomly picked and GFP expression in each cell was evaluated by using FACS. The measured fluorescence intensity obtained from colonies with P vgb -promoter-based vectors ranged from 10 2 -10 3 , whereas the negative control cells did not show significant fluorescence. The fluorescence by clones with P vgb8 were more intense than the clones with P vgb4 and P vgb , after 24 h of cultivation ( Figure 2). However, after 48 h, no significant difference was observed in GFP expression between clones with P vgb4 and those with P vgb8 (data not shown). The increasing number of P vgb in the constructed expression system was directly proportional to the intensity of the GFP fluorescence in the resulting recombinant strains ( Figure 2). These observations were similar to the PHB production enhancement in recombinant E. coli strains harboring the PHB operon (phaCAB) under the control of P vgb8 compared to the use of P vgb . Based on these results, we have demonstrated that the strength of protein expression using P vgb was tunable by modulating the number of its copies in the constructed expression system. To further demonstrate tunability of the P vgb promoter-based expression system in recombinant C. glutamicum, the developed constructs were used to produce 5-AVA and GABA from L-lysine and L-glutamate, respectively. 5-AVA and GABA were selected as model compounds to test the application of the P vgb promoter-based expression system because their production in C. glutamicum has been extensively studied in recent times [6,25,26,[29][30][31].

5-AVA Production Using Recombinant C. glutamicum with Tunable P vgb Promoter-Based Expression Systems
To demonstrate that the P vgb promoter-based expression system is a new platform for engineering better C. glutamicum strains, its feasibility to produce 5-AVA in C. glutamicum KCTC 1857 was evaluated [6,25,30]. Tunability of protein expression by P vgb was verified by developing C. glutamicium 5AVA1 and 5AVA8 strains for the expression of lysine 2-monooxygenase (DavB) and delta-aminovaleramidase (DavA) under the control of single (P vgb ) and eight repeats (P vgb8 ) of P vgb [6,25,30]. 5-AVA production using the recombinant strains was evaluated in flask cultivation under low and high aeration conditions, as it was previously demonstrated that protein expression using P vgb was enhanced under microaerobic conditions in recombinant E. coli strains [34]. The production of 5-AVA by the recombinant C. glutamicum 5AVA1 (3.43 ± 0.10 g/L) and 5AVA8 (3.68 ± 0.07 g/L) strains was better under high aeration conditions than at low aeration conditions (2.10-1.76 g/L) (Figure 3). Higher accumulation of L-lysine and glutarate was also detected after flask cultivation of 5AVA1 and 5AVA8 strains under high aeration conditions (4.41-4.46 g/L of L-lysine and 0.71-0.74 g/L of glutarate) than that at low aeration (3.99-4.58 g/L of L-lysine and 0.49-0.57 g/L of glutarate) (Figure 3c,d). 5-AVA production by all the recombinant strains were higher under high aeration conditions than that under flask cultivation with low aeration conditions. These results were similar to the trend of L-glutamate and L-glutamine production by recombinant C. glutamicum strains, which expressed the Vitreoscilla hemoglobin gene under P vgb and P tac , wherein chemical production was higher under high aeration than at low aeration [34]. This is because C. glutamicum is an obligate aerobic microorganism, and it requires high aeration for efficient amino acid production [1,4,5]. Additionally, because DavB requires oxygen as a co-substrate, higher aeration conditions are preferred for 5-AVA production [30].
Furthermore, the effect of eight repeats of P vgb on 5-AVA production was evaluated to demonstrate the tunability of protein expression using the constructed P vgb promoter-based vector system ( Figure 3). Different levels of 5-AVA production were achieved when the number of P vgb increased from single (P vgb ) to eight repeats (P vgb8 ) ( Figure 3). It was observed that 5-AVA production increased based on the number of repeats of P vgb . 5-AVA production by the 5AVA8 strain (Figure 3d, 3.69 ± 0.07 g/L), which expressed DavBA under P vgb8 , was higher than that of the 5AVA1 strain (Figure 3c, 3.43 ± 0.10 g/L)), which harbored P vgb . 5-AVA production by 5AVA1 and 5AVA8 strains (3.43-3.69 g/L) were comparably higher than that of the H30_AVA strain, which expressed DavBA under the strong synthetic promoter P H30 (1.4 ± 0.3 g/L g/L) (Table S1) [30]. This effect was similar to the observed trend of enhanced PHB accumulation in recombinant E. coli strains when the number of P tac and P vgb repeats increased to five and eight, respectively [32,33]. Finally, the effect of his-tagged DavB on 5-AVA production was also investigated, because we have previously reported that it enhanced 5-AVA production in C. glutamcium H30_AVA His . In our previous study, we demonstrated enhanced 5AVA production by expressing his-tagged DavB along with DavA under the strong synthetic promoter P H30 in H30_AVA His strain, which produced 4.2 ± 0.9 g/L of 5-AVA [30]. This significantly increased 5-AVA production compared to that by the H30_AVA strain, which expressed DavB without his-tag (4.2 ± 0.9 g/L >1. 4 ± 0.3 g/L) (Figure 4) [30]. Therefore, to investigate the effect of this additional his-tag on 5-AVA production under the control of the P vgb promoter system, recombinant strains 5AVA1 His and 5AVA8 His were constructed by inserting his-tagged DavB into the P vgb promoter system with single (P vgb ) and eight (P vgb8 ) repeats, respectively. 5-AVA production by the resulting recombinant strains were only evaluated under high aeration conditions, as the low aeration condition did not increase 5-AVA production in 5AVA1 and 5AVA8 strains (Figure 3). Under high aeration conditions, the 5AVA8 His (3.31 ± 0.08 g/L) strain produced higher concentrations of 5-AVA than the 5AVA1 His (2.82 ± 0.03 g/L) strain ( Figure 4, Table S1). As shown in Figure 4, 5-AVA production increased as the number of P vgb repeats increased from single to eight in 5AVA1 His and 5AVA8 His (3.31 ± 0.08 g/L > 2.82 ± 0.03 g/L). Glutaric acid production also increased in both 5AVA1 His (0.64 ± 0.01 g/L) and 5AVA8 His (0.67 ± 0.05g/L). Interestingly, both L-lysine accumulation and 5-AVA production increased as the number of repeats of P vgb increased from a single repeat in 5AVA1 His (4.00 ± 0.06 g/L) to eight repeats in the 5AVA8 His (4.38 ± 0.03 g/L) strain. Based on these observations, the P vgb promoter system was successfully used as a new platform for 5-AVA production in C. glutamicum.

Gamma-Aminobutyric Acid Production Using Recombinant C. glutamicum with Tunable P vgb Promoter-Based Expression Systems
The constructed P vgb promoter-based vector system was also used for the expression of mutated glutamate decarboxylase (GadB mut ) in order to investigate GABA production in the C. glutamicum strain ATCC 13032 and in high-L-glutamate producing strains C. glutamicum KCTC 1447 and C. glutamicum KCTC 1852 as host strains ( Figure 5, Table S2) [26,31]. C. glutamicum KCTC 1447 and C. glutamicum KCTC 1852 were used because they are capable of high production of L-glutamate, an important precursor for GABA production [26,31]. C. glutamicum ATCC 13032 was also used as control host strain [30]. The effect of the eight repeats of P vgb on GABA production was evaluated under high aeration conditions because we have previously demonstrated that low aeration did not increase chemical production by recombinant strains expressing DavBA under single and eight repeats of P vgb during shake flask cultivation (Figure 3). It was observed that the average GABA production by C. glutamicum KCTC 1852-derived strains (5.31 g/L-5.58 g/L) was higher than that by recombinant C. glutamicum ATCC 13032 and C. glutamicum KCTC 1447 strains (4.82-5.22 g/L) ( Figure 5, Figure S3). This is because the host strain, C. glutamicum KCTC 1852, is naturally capable of higher L-glutamate production than C. glutamicum KCTC 1447 and C. glutamicum ATCC 13032 [30]. Higher GABA production was attributed to the efficient production of L-glutamate, an important pre-requisite for the development of strains for GABA production [26,31]. The level of GABA production in all the tested strains increased as the number of P vgb repeats increased ( Figure 5). For example, GABA production in V1GD1852 (5.31 ± 0.16 g/L) and V8GD1852 (5.58 ± 0.27 g/L) strains increased when GadB mut was expressed under P vgb and P vgb8 , respectively. The same tendency was observed in V1GD13032 (4.82 ± 0.09 g/L) and V1GD1447 (5.09 ± 0.64 g/L) strains with respect to GABA production, under the control of the P vgb promoter, and by V8GD13032 (4.94 ± 0.27 g/L) and V8GD1447 (5.22 ± 0.60 g/L) strains under the control of the P vgb8 promoter. Regarding GABA production using C. glutamicum ATCC 13032 strains (4.82-4.94 g/L), similar GABA production was achieved by pLGmut (4.87 ± 0.15 g/L), V1GD13032 (4.82 ± 0.09 g/L), and V8GD13032 (4.94 ± 0.27 g/L) strains, which expressed GadB mut under the synthetic promoters P L26 , P vgb , and P vgb8 , respectively ( Figure 4, Table S2) [26]. The level of GABA production using the constructed P vgb promoter-based expression system (4.82-5.09 g/L) was comparable to the titers achieved by recombinant strains pHGmut (5.89 ± 0.35 g/L) and pIGmut (5.32 ± 0.04 g/L). The recombinant strains expressed GadB mut under synthetic promoters of high (P H36 ) and intermediate (P I16 ) strength [31]. The level of GABA production by the C. glutamicum KCTC 1852-derivative strains V1GD1852 (5.31 ± 0.16) and V8GD1852 (5.58 ± 0.27), expressing GadB mut , were similar to the level of GABA production achieved when the strong synthetic promoter P H36 was used in the H36GM1852 strain (8.47 ± 0.06 g/L) (Table S2). This shows that the strength of GadB mut expression by the P vgb promoter-based expression system is also tunable like previously used synthetic promoters (P L26 , P I16 , P H36 ) (Table S2) [26,31]. Based on these results, we concur that the P vgb -based expression system was capable of tunable expression of key proteins for 5-AVA (DavBA) and GABA (GadB mut ) production by increasing the number of P vgb repeats. The addition of the newly constructed P vgb promoter-based expression system into the repertoire of plasmids available for metabolic engineering of C. glutamicum provides an alternative method of fine-tuning protein expression levels for validating synthetic metabolic pathways and improving biochemical production in biorefineries.

Bacterial Strains and Plasmids
All the bacterial strains and plasmids used in this study are listed in Table 1. E. coli XL1-Blue (Stratagene, La Jolla, CA, USA) was used for the general gene cloning studies. C. glutamicum ATCC 13032, C. glutamicum KCTC 1447, C. glutamicum KCTC 1852, and C. glutamicum KCTC 1857 strains were purchased from the Korean Collection for Type Cultures (KCTC, Daejeon, Korea). The pCES208-based plasmids: pCES208H30:DavBA, pCES208H30:DavB His A and pHG mut were constructed as previously described [26,30].

Plasmid Construction
All DNA manipulations were performed following standard procedures [36]. Polymerase chain reaction (PCR) was performed with the C1000 Thermal Cycler (Bio-Rad, Hercules, CA, USA). The primers used in this study were synthesized at Bioneer (Daejeon, Korea). The pHG mut plasmid was cut at KpnI and BamHI sites to replace the synthetic promoter P H36 with P vgb promoter. The subsequent repeats of the P vgb promoter were inserted into the pCES208V:eGFP vector at BamHI/BglII sites to obtain pCES208V4:eGFP and pCES208V8:eGFP plasmids. The pCES208V:eGFP, pCES208V4:eGFP, and pCES208V8:eGFP plasmids for the expression of enhanced GFP were constructed by inserting GFP at the BamHI/NotI sites of modified pCES208-based plasmids with one, four, and eight repeats of P vgb , respectively. The plasmids for the expression of DavBA, DavB His A, and GadB mut were also inserted at the BamHI/NotI sites of pCES208V:eGFP and pCES208V8:eGFP (Table 1).

Culture Conditions
E. coli XL1-Blue, used for general gene cloning experiments, was cultured at 37 • C in Luria-Bertani (LB) medium (10 g/L tryptone, 5 g/L yeast extract, and 5 g/L NaCl). Flask cultures of the recombinant strains of C. glutamicum were obtained in triplicates by culturing at 30 • C and 250 rpm in a rotary shaker. High aeration conditions for flask cultivation were maintained by adding 20 mL of the appropriate culture medium into a 250-mL baffled flask. Low aeration conditions for flask cultivation were maintained by adding 100 mL of the culture medium into a 250-mL glass flask. Seed cultivation of C. glutamicum strains was carried out in 14-mL round-bottomed tubes containing 2 mL of Recovery Growth (RG) medium (10 g/L of glucose, 40 g/L of brain heart infusion, 10 g/L of beef extract, and 30 g/L of D-sorbitol) with incubation overnight at 30 • C and 250 rpm [29]. The main-flask cultures for 5-AVA production were grown in 250 mL baffled flasks containing 20 mL of CG50 medium for 120 h at 30 • C and 250 rpm. The CG50 medium for flask cultivation contained (per liter) 50 g glucose, 15 g yeast extract, 15 g (NH4) 2 SO 4 ·7H 2 O, 0.5 g KH 2 PO4, 0.5 g MgSO 4 ·7H 2 O, 0.01 g MnSO 4 ·H 2 O, 0.01 g FeSO 4 ·7H 2 O, and 20 µg/L of kanamycin (Km) for plasmid maintenance [30]. The GP1 medium optimized in our previous study [26] was used for GABA production. The GP1 medium for flask cultivation contained (per liter) 50 g (NH 4 ) 2 SO 4 ·7H 2 O, 1 g K 2 HPO 4 , 3 g urea, 0.4 g MgSO 4 · 7H 2 O, 50 g peptone, 0.02 g FeSO 4 ·7H 2 O, 0.007 g MnSO 4 ·H 2 O, 200 µg thiamine, and 1 mM of pyridoxal 5 -phosphate hydrate (PLP) [26]. PLP was added to the culture medium as it is a cofactor of glutamate decarboxylase. Moreover, 0.1 mM of PLP was the optimum concentration for prolonging GABA production using recombinant C. glutamicum strains [26]. Kanamycin and biotin were added to the GP1 culture medium at 25 and 50 µg/L, respectively. Only 50 µg /L of biotin was used in flask cultivation for GABA production because biotin-limited condition promotes L-glutamate accumulation [26]. CaCO 3 was added to the culture medium at 10 g/L to minimize the pH change during cultivation.

Analysis
The concentrations of glucose and organic acids were determined by high performance liquid chromatography (HPLC). The standard and sample concentrations of 5-AVA, GABA, L-lysine, and L-glutamate were determined by HPLC using an Optimapak C18 column (RStech, DaeJeon, Korea) as previously reported [37].

Fluorescence-Activated cell Sorting Analysis (FACS) for Measuring GFP Expression by Recombinant Strains
FACS analysis was used to investigate the GFP expression by using the constructed P vgb system. The recombinant EGFPV1, EGFPV4 and EGFPV8 strains were grown in Brain Heart Infusion (BHI) media for 24 h at 30 • C. The cells were then collected and diluted using PBS buffer. FACS analysis (BD Biosciences, San Jose, CA, USA) was performed for 100,000 clones of each samples using argon ion laser (blue, 488 nm) and band-pass filter (530 nm ± 15 nm) [35].

Conclusions
The P vgb promoter-based expression system constructed in this study was capable of tunable expression of green fluorescent protein in recombinant C. glutamcium strains, when the repeats of P vgb promoter increased from one (P vgb ) to four (P vgb4 ) to eight (P vgb8 ). Furthermore, GABA and 5-AVA production by recombinant C. glutamicum strains also increased when the expression of DavBA and GadB mut in the P vgb promoter-based expression system increased from single to eight repeats of the P vgb promoter. This shows that the strength of protein expression using the P vgb promoter-based tunable system was comparable to that of previously established synthetic promoters (P L26 , P I16 , P H30 P H36 ) [26,[28][29][30]. Based on the different levels of 5-AVA and GABA production by all the tested strains, the P vgb promoter-based expression system was capable of tunable expression of DavBA and GadB mut by mere manipulation in the number of P vgb repeats. The newly constructed P vgb promoter-based expression system developed in this study expands the repertoire of plasmids available for metabolic engineering of C. glutamicum and provides another method for fine-tuning levels of protein expression for convenient and rapid validation of synthetic metabolic pathways, ultimately improving biochemical production in biorefineries.