Improved Plasmid-Based Inducible and Constitutive Gene Expression in Corynebacterium glutamicum

Corynebacterium glutamicum has been safely used in white biotechnology for the last 60 years and the portfolio of new pathways and products is increasing rapidly. Hence, expression vectors play a central role in discovering endogenous gene functions and in establishing heterologous gene expression. In this work, new expression vectors were designed based on two strategies: (i) a library screening of constitutive native and synthetic promoters and (ii) an increase of the plasmid copy number. Both strategies were combined and resulted in a very strong expression and overproduction of the fluorescence protein GfpUV. As a second test case, the improved vector for constitutive expression was used to overexpress the endogenous xylulokinase gene xylB in a synthetic operon with xylose isomerase gene xylA from Xanthomonas campestris. The xylose isomerase activity in crude extracts was increased by about three-fold as compared to that of the parental vector. In terms of application, the improved vector for constitutive xylA and xylB expression was used for production of the N-methylated amino acid sarcosine from monomethylamine, acetate, and xylose. As a consequence, the volumetric productivity of sarcosine production was 50% higher as compared to that of the strain carrying the parental vector.


Introduction
C. glutamicum was discovered in the 1960s as a natural L-glutamate producer [1]. Since then, both its genetic toolbox [2] and its number of heterologous pathways [3,4] have been extended. On the one side, production of value-added compounds such as amino acids [5,6], organic acids [7,8], and terpenoids [9,10] has been established. Recently, the production of sarcosine (N-methylglycine) was enabled by overexpression of the imine reductase DpkA from Pseudomonas putida in a glyoxylate-overproducing C. glutamicum strain by providing monomethylamine as the methyl-donor [11]. On the other side, several approaches were followed in order to establish a flexible feedstock concept that allows C. glutamicum production strains to grow/produce on the basis of a variety of non-food competitive substrates such as industrial or agricultural/aquatic side streams [12]. The access to glycerol, the stoichiometric byproduct of biodiesel production, was enabled and applied to various production strains [13,14]. Moreover, recent attempts have aimed to establish the methylotrophy in C. glutamicum for methanol utilization [15]. Besides the sugar polymers starch [16] and cellulose [17], the pentose sugars xylose and arabinose [18,19] that derive from hemicellulose can be used as alternative substrates for a variety of high-value products including the fragrance compound patchoulol [20] and the potential antipsychotic compound sarcosine [11].
Many of these production strain-engineering efforts rely on gene expression vectors, which represent a powerful tool not only for metabolic engineering, but also for in depth analysis of basic metabolic principles in C. glutamicum that facilitate the development of new metabolic engineering strategies. For C. glutamicum, a wide range of different

Construction of New Expression Vectors
The new expression plasmids were constructed in E. coli DH5α. First, target promoters or genes were amplified by a high-fidelity PCR (All-in HiFi, highQu, Kraichtal, Germany) and cloned into digested or backbone-amplified expression vectors by Gibson-Assembly [44]. The oligonucleotides are listed in Table 2 and were delivered by Metabion (Planegg/Steinkirchen, Germany). The PCR amplificates were purified with a PCR and gel extraction kit (Macherey-Nagel, Düren, Germany). E. coli DH5α-competent cells were prepared with CaCl 2 and transformed via heat shock. Recombinant clones were screened by colony-PCR and plasmids were isolated with a miniprep kit (GeneJET, Thermo Fisher Scientific, Schwerte, Germany). New expression vectors were confirmed by sequencing. C. glutamicum cells were transformed by electroporation as described elsewhere [45].
For construction of the pVWEx4 expression vectors, a site-directed mutagenesis in the repA gene was performed on pVWEx1 for an increased plasmid copy number. Site-directed mutagenesis was conducted via plasmid backbone amplification (HA36 + HA37) with Pfu Turbo DNA Polymerase. The vector pVWEx6 was constructed via digestion of pVWEx4 with KpnI and MauBI and insertion of the Psyn promoter including a lac operator motif (N105 + N106).

Fluorescence Analysis
GfpUV fluorescence was measured on a FACS GalliosTM (Beckmann Coulter GmbH, Krefeld, Germany) with 405 nm excitation from a blue solid-state laser. Forward-scatter characteristics (FSC) and side-scatter characteristics (SSC) were monitored as small-and large-angle scatters of the 405 nm laser. Fluorescence was detected using a band-pass filter (500/50 nm). The produced GfpUV protein possesses characteristic emission at 509 nm with an excitation wavelength of 385 nm. C. glutamicum WT harboring the newly constructed plasmids or the control plasmid were harvested in stationary growth, washed once in phosphate-buffered saline, and the optical density was adjusted to OD 600nm~0 .1. C. glutamicum WT was used to determine autofluorescence. Median fluorescence intensities of 20,000 cells were calculated from each culture.

SDS-PAGE
C. glutamicum WT cells carrying the newly constructed plasmids were grown at 30 • C in 1 mL in Biolector ® flowerplate system (m2p-labs GmbH, Baesweiler, Germany). Cells were harvested by centrifugation and stored at −20 • C. For cell extract preparation, thawed cells were re-suspended in KPI buffer (100 mM disodium hydrogenphosphate and 100 mM sodium dihydrogenphosphate with pH 6.9). Cells were disrupted by ultrasonification using Hielscher UP200S2 (Teltow, Germany) with an amplitude of 60% and a pulsing cycle of 0.5 (power discharge 0.5 s, pause 0.5 s) for 9 min. After ultracentrifugation (1 h, 45,000× g, 4 • C) the protein concentration was determined with Bradford Reagent (Sigma-Aldrich, Germany). Then, a 10 µg protein sample was mixed with Lämmli-Buffer and loaded on an SDS-PAGE, comprising a 4% stacking gel and 10% running gel. Gels were initially run at 50 V and then at 100 V. Protein gels were stained in 0.1% Coomassie blue solution (30% methanol; 10% acetic acid).

Enzyme Assay for Xylose Isomerase XylA
Cells of the strains WT (pECXT99A-xylAB) and WT (pECXT_Psyn-xylAB) were inoculated from a fresh LB overnight culture into a LB main culture with antibiotics and IPTG (1 mM) if applicable. Cultures were inoculated with an initial OD 600 nm of 0.6 and grown until OD 600 nm was 4. All cells from each culture were harvested at 4 • C and stored at −20 • C till further use. Cells were washed in Tris-HCl Buffer (100 mM pH 7.5) and resuspended in 2 mL buffer prior to ultrasonification using Hielscher UP200S2 (Teltow, Germany) with an amplitude of 60% and a pulsing cycle of 0.5 (power discharge 0.5 s, pause 0.5 s) for 9 min. After ultracentrifugation (1 h, 45,000× g, 4 • C) the protein concentration was determined with Bradford Reagent (Sigma-Aldrich, Germany). Enzymatic activity of XylA was measured from the raw extract in a total volume of 1 mL at 30 • C: 100 mM Tris-HCl pH 7.5, 10 mM MgCl 2 , 0.17 mM NADH, and 1 U sorbitol dehydrogenase. Then, 200 µL D-xylose (2.5 M) was added and absorption was measured for an additional 3 min using a Shimadzu UV-1650 PC photometer (Shimadzu, Duisburg, Germany). The assay was performed with three appropriate protein concentrations each, applying 5/10/15 µL of crude extracts from strains WT (pECXT99A-xylAB) and WT (pECXT_Psyn-xylAB) with protein concentrations of 5.8 mg/mL and 1.72 mg/mL, respectively. Specific activities were calculated as µmol min −1 (mg protein) −1 , defined as one unit (U/mg).

Sarcosine Quantification
Sarcosine quantification was performed from the clear supernatant of culture samples. Samples were derivatized with 9-fluorenylmethyl chlorocarbonate (FMOC) as decribed elsewhere [46]. Identification and separation was performed on a reversed phase HPLC.

Screening of Strong Constitutive Promoters in the pECXT99A-Based Vector System
A selection of native and synthetic promoters was cloned in the pECXT99A-backbone (lacking the lacI q gene to allow for constitutive expression) in order to test the different promoter strengths using the promoterless reporter gene gfpUV. The promoters from the endogenous genes tuf (cg0587), ilvC (cg1437), sodA (cg3237), gapA (cg1791), and pgk (cg1790) were amplified from the genome of C. glutamicum WT. The synthetic promoters PH36 and Psyn were generated by gene synthesis, whereas P45 was generated by oligonucleotide insertion. GfpUV reporter fluorescence in C. glutamicum WT carrying the plasmids was monitored 18 h after inoculation in CGXII with 4% glucose without any inducer. The strain carrying the parental plasmid pECXT99A-gfpUV (with lacIq) was chosen as a reference (cultured in the presence of 1 mM IPTG), as this vector served as the starting point for the cloning of the promoter library. All tested promoters showed activity and, thus, allowed for constitutive gene expression without the need to add an inducer to the medium (Figure 1). The synthetic promoter PH36 (norm. MFI: 2.7 ± 13%) and the endogenous promoters Ppgk (norm. MFI: 4.0 ± 12%) and PilvC (norm. MFI: 7.0 ± 5%) showed significant expression in comparison to that of the autofluorescence (norm. MFI: 1.0 ± 10%). However, the expression was relatively low in comparison to that of the other vectors. The endogenous promoters PsodA (norm. MFI: 10.8 ± 1%), PgapA (norm. MFI: 19 ± 9%), and Ptuf (norm. MFI: 50 ± 15%) showed increasing expression levels with the latter one showing comparable expression values to that of the synthetic P45 (norm. MFI: 56 ± 7%). Thus, gfpUV expression from the constitutive P45 was as strong as IPTG-induced gfpUV expression from Ptrc in pECXT99A (with lacIq and Ptrc; norm. MFI: 56 ± 6%). Notably, gfpUV expression from the synthetic promoter Psyn exceeded this level by about a factor of five (norm. MFI: 286 ± 4%). Thus, the newly constructed pECXT_Psyn showed the highest gfpUV expression of the compared vectors ( Figure 1).

Improving Plasmid-Borne Inducible Gene Expression in C. glutamicum by Combination of a Stronger Promoter with Increased Plasmid Copy Number
First, the properties of the commonly used IPTG-inducible expression vectors pEKEx3, pVWEx1, and pECXT99A were compared with regard to gfpUV gene expression by FACS analysis (Figure 2) and protein abundance by SDS-PAGE ( Figure 3). Induction factors were determined from reporter fluorescence measured in the absence and presence of 1 mM IPTG. With vector pEKEx3-gfpUV, a maximal norm. MFI of 64 ± 13% was reached, but expression was leaky, thus, an induction factor of only 6 resulted ( Figure 2). The vector pECXT99A-gfpUV supported a maximal norm. MFI of 48 ± 3% with an induction factor of 47. Expression from pVWEx1-gfpUV without IPTG was as tight as that observed for pECXT99A-gfpUV. With higher maximal norm. MFI of 99 ± 3% an induction factor of 76 was observed with pVWEx1-gfpUV. SDS-PAGE revealed lower GfpUV protein abundance for pEKEx3-gfpUV as compared to that of pECXT99A-gfpUV, which was highest among the three vectors for pVWEx1-gfpUV ( Figure 3).
Next, we sought to improve inducible gene expression vector pVWEx1, which possesses the origin of replication pHM1519. It has been show that pHM1519-based vectors are maintained at plasmid copy numbers of approximately 140 in C. glutamicum cells [47]. Importantly, it has recently been shown that the copy number of vectors with a pHM1519 replicon can be increased about 5-fold to about 800 [30]. This finding prompted us to increase the plasmid copy number of pVWEx1. Therefore, the repA gene encoding an initiator protein for the pHM1519 replicon was mutated (base transition from G to A at nucleotide 1286) to exchange Gly at position 429 of RepA protein by Glu via site-directed mutagenesis, which was designated as copA1 [30]. The resulting vector based on pVWEx1 with the mutation RepA G429D was named pVWEx4. FACS analysis revealed an approximately 1.5-fold increased IPTG-induced gfpUV gene expression using vector pVWEx4-gfpUV as compared to that using pVWEx1-gfpUV (norm. MFI of 145 ± 7% as compared to 99 ± 3%). Since expression without IPTG was almost as tight for pVWEx4 as for pVWEx1, the induction factor with pVWEx4 was higher than that with pVWEx1 (121 as compared to 76). Thus, introduction of the copA1 mutation into the repA gene increased the plasmid copy number as expected, and allowed for a very strong and more than 100-fold inducible target gene expression.  Third, since Ptrc was weaker than Psyn (Figure 1), we hypothesized that replacing Ptrc in plasmid pVWEx4 by the stronger Psyn, while maintaining the lacIq and lac operator sequences for IPTG-inducible gene expression, would lead to increased target gene expression that could be modulated by titrating IPTG concentrations. The resulting vector was named pVWEx6. This vector allowed for an approximately 50-fold induction with 1 mM IPTG, leading to a maximal norm. MFI of 410 ± 1% (Figure 2 and Table 3). Moreover, we sought to find a lower IPTG concentration for pVWEx6 that would achieve comparable expression as that of the fully induced pVWEx1, pEXCT99A, and/or pEKEx3. Our choice of 0.025 mM and 0.050 mM IPTG as intermediate IPTG concentrations was guided by prior experience with vectors pVWEx1 and pEKEx3 for expression of genes for membrane proteins or transporter proteins [48][49][50]. Overexpression of genes for membrane proteins or transporter proteins proved difficult as too high levels reduced growth, probably by altering membrane integrity. However, with 0.025 mM or 0.050 mM IPTG as the intermediate IPTG concentrations, functional overexpression of genes coding for membrane proteins or transporter proteins was achieved without observed growth impairment [48][49][50]. Here, intermediate GfpUV fluorescence was observed with 0.025 and 0.05 mM IPTG. SDS-PAGE analysis of GfpUV reporter protein abundance (Figure 3) confirmed the FACS results of very strong expression (Figure 2). In fact, the IPTG-induction factor was as high as those obtained with pVWEx1 and pECXT99A, while the fully induced expression strength supported by pVWEx6 was one magnitude higher than those achieved with pEKEx3, pECXT99A, and pVWEx1.

Fast Production of Sarcosine from Xylose by Application of the Newly Constructed pECXT_Psyn
As an application test, sarcosine production by C. glutamicum was chosen ( Figure 4A). C. glutamicum was engineered to produce the N-methylated amino acid sarcosine in the enzyme reaction catalyzed by the Pseudomonas putida KT2440 imine reductase DpkA using glyoxylate as the 2-oxo acid substrate and monomethylamine as the N-alkyl donor [11]. To enable xylose utilization, the WT carrying malate synthase gene deletion ∆aceB, a start codon exchange from ATG to GTG for isocitrate dehydrogenase gene icd and the vector pVWEx1-dpkA_RBSopt for IPTG-inducible expression of dpkA, was transformed with pECXT99A-xylAB as the second vector. Notably, the resulting C. glutamicum strain SAR3 produced sarcosine with higher yield coefficients than that with glucose, however, the volumetric productivity was low due to slow growth with xylose [11]. To test if stronger expression of xylAB using the newly constructed vector pECXT99A_Psyn-xylAB accelerates xylose-based sarcosine production, strain SAR3* that carried pECXT99A_Psyn-xylAB instead of pECXT99A-xylAB was constructed ( Figure 4A).
First, to score xylose catabolism an enzyme assay of the heterologous xylose isomerase XylA was performed. To this end, crude extracts of C. glutamicum WT carrying either pECXT99A_Psyn-xylAB or pECXT99A-xylAB were prepared and assayed spectrophotometrically for XylA activity ( Figure 4B). A more than 3-fold increase in the specific XylA activity indicated that pECXT99A_Psyn-xylAB provided higher xylAB expression than pECXT99A-xylAB ( Figure 4B). Next, sarcosine production by strains SAR3 and SAR3* was compared in media containing monomethylamine, potassium acetate, and xylose. When grown in the presence of acetate, C. glutamicum is known to carry a high flux from acetyl-CoA into the TCA cycle and to glyoxylate [7,51] (Figure 4A). This is due to transcriptional regulation of the isocitrate lyase gene aceA (and malate synthase gene aceB that is deleted in SAR3) by the transcriptional activator RamA and the transcriptional repressor RamB [52,53]. Xylose served as the primary carbon and energy source for growth under these conditions. Growth of SAR3* was 1.3-fold faster than growth of SAR3 (growth rates of 0.13 ± 0.01 h −1 as compared to 0.10 ± 0.01 h −1 , respectively). Importantly, after 20 h, sarcosine accumulated to titers of 1.0 ± 0.08 g L −1 for SAR3 and 1.5 ± 0.13 g L −1 for SAR3* ( Figure 4C). As a consequence, the volumetric productivity was improved by 50% (0.077 g L −1 h −1 for SAR3*) as compared to SAR3 (0.051 g L −1 h −1 ). Thus, pECXT99A_Psyn-xylAB proved useful to improve the xylose-based sarcosine production by recombinant C. glutamicum.

Discussion
Expression vectors play a crucial role in basic research to investigate endogenous metabolism and to establish microbial bio-production by metabolic engineering. However, the development and/or optimization of expression systems is rarely the focus of scientific work. In this study, we focused on the optimization of the two well-established expression vectors pECXT99A and pVWEx1 by application of a strong promoter and by an increased plasmid copy number, respectively. Due to different C. glutamicum origins of replication and antibiotic resistance markers (Table 3), these vectors can co-exist: pEKEx3 plus pECXT99A or pECXT_Psyn plus pVWEx1 or pVWEx4 or pVWEx6.
The expression level of a vector system is dependent on a wide range of aspects that affect either transcription or translation of an overexpressed gene or the protein stability itself. For example, plasmid-driven gene expression relies on the number of gene copies (plasmid-copy number) and the promoter strength. Both aspects were part of this work. In addition, the translational efficiency of target genes may be optimized by changing the translational start codon [54] or by changing ribosome binding sites as a trade-off between an optimal RBS (consensus C. glutamicum: 5 -GAAAGGAGG-3 ) and the prevention of secondary mRNA structures [55][56][57]. Independent of vector characteristics, the target gene (and gene product) may be optimized. Protein stability can be improved either by adding tags, by construction of fusion enzymes [58,59], or by truncations to improve protein solubility [60,61] while taking into account dimerization domains, catalytic centers, and cofactor binding sites [62]. If their termini are freely accessible, membrane proteins may be fused as shown in the fusion of the cytosolic C terminus of CrtZ to the cytosolic N terminus of CrtW, which resulted in improved astaxanthin production by C. glutamicum [10]. The improved vectors developed here (pVWEx4, pVWEx6, and pECXT99A_Psyn) add to the toolbox available to microbiologists studying the physiology of C. glutamicum as well as for metabolic engineering in strain and process development. Moreover, the expression characteristics obtained in comparative fluorescence reporter gene expression experiments will guide the choice of vectors according to the task of interest.
To compare plasmid-borne target gene expression, the endogenous promoters of the genes pgk, ilvC, sodA, gapA, and tuf were chosen in this study. The promoters of the glycolytic genes pgk [63] and gapA [64], encoding phosphoglycerate kinase and glyceraldehyde 3-phosphate dehydrogenase, respectively, as well as of the superoxide dismutase gene sodA [65], the ketol-acid reductoisomerase gene ilvC [65], and the transcription elongation factor gene tuf [66] have already been described as strong constitutive promoters. In a different vector background, it was shown that PilvC and PsodA showed strong expression in a GFP reporter assay with comparable expression levels as that of Ptac in the case of PilvC and around a 2.5-fold higher expression in the case of PsodA [65]. In the study presented here, expression levels of both promoters were lower than that of Ptrc. As discussed above, the promoter is one of several factors influencing reporter gene expression, making comparisons of reporter gene expression using different vector backbones, ribosome binding sites, etc. difficult. Nevertheless, the previously described vectors along with those described here allow for low, but constitutive expression and are of interest for the production of membrane proteins, transporter proteins, or enzymes that lead to the formation of toxic byproducts and for balancing of engineered pathways [48].
To maximize target gene expression, the design and usage of synthetic promoters for genetic engineering in C. glutamicum is on the rise, as these usually short promoters can achieve high transcription levels [37,38]. Moreover, the utilization of artificial promoters allows researchers to circumvent the challenges of regulatory interference in metabolic engineering approaches for overproduction. The PH36 promoter was identified by Yim et al. 2013, where it showed a 60-fold higher expression level in the pCES plasmid background in comparison to that of the Ptrc promoter [37]. In contrast, here, PH36 showed the weakest expression in the fluorescence assay. The synthetic Psyn was designed and identified as it showed high promoter activity in a β-galactosidase assay [38]. Although this result is difficult to compare with the fluorescence assay used here, we have shown that the short synthetic promoter Psyn (5 -TTGACATTAATTTGAATCTGTGTTATAATGGTTC-3 ) enabled tunable and strong expression in combination with the lac operator sequence in cis (5 -CTGGAATTGTGAGCGGATAACAATTC-3 ) and lacIq in trans. The newly constructed pVWEx6 vector showed a gene expression strength that is one magnitude higher than that of other commonly used expression vectors. Thus, the application of pVWEx6 can reduce the costs for IPTG, as 25 µM IPTG lead to comparable expression patterns as that of pVWEx1, pECXT99A, and pEKEx3 induced with 1 mM IPTG. Potentially, leakiness of these vectors can be further minimized as shown very recently for pEKEx2 by restoring lacIq gene sequences and by avoiding duplicate DNA sequences [67].
As an application case, we transferred the findings about elevated gene expression to fermentative sarcosine production with C. glutamicum. Strain SAR3* was based on the newly designed pECXT_Psyn for expression of xylose utilization gene xylAB and it showed a volumetric productivity for sarcosine with 0.077 g L −1 h −1 as compared to that of the reference strain (0.051 g L −1 h −1 ) [11]. Our previous work on sarcosine production [11] identified slow growth with xylose as the limitation of the volumetric productivity for sarcosine. Here, we have shown that accelerating growth with xylose improved the volumetric productivity of SAR3* by 50% as compared to that of SAR3. We deliberately did not use pVWEx4 or pVWEx6 for dpkA expression here, since N-methylation of glyoxylate to yield sarcosine as catalyzed by DpkA does not limit sarcosine production. This was deduced from our previous work [68], where pVWEx1-dpkA supported a volumetric productivity of up to 0.35 g L −1 h −1 for N-methylation of pyruvate to N-methylalanine.
Since it was shown that doubling of the MMA concentration as well as optimization of the carbon source composition (5 g/L xylose and 30 g/L actetate) improved the sarcosine production by SAR3 about 3-fold, combination of the newly constructed strain SAR3* with those culture conditions might further improve sarcosine production with C. glutamicum.

Data Availability Statement:
The data presented in this study are available in article.