Integrative expression vectors with Pgrac promoters for inducer-free overproduction of recombinant proteins in Bacillus subtilis

Highlights • The new inducer-free integrative expression vectors could repress the reporter gene expression in the E. coli cloning strain, thereby facilitating the cloning step.• The expression vectors carrying IPTG-inducible Pgrac promoters allow the production of the recombinant protein at high levels in B. subtilis in the absence of the inducer.• The single-copy expression levels of integrative constructs, Pgrac01-bgaB, Pgrac100-bgaB, Pgrac212-bgaB could reach to % and 8%, 20.9 % and 42 % of total cellular proteins after 12 h incubation, respectively.• The double integration of Pgrac212-bgaB into both amyE and lacA loci resulted in BgaB expression up to 53.4 %.


Introduction
B. subtilis is presently the best-characterized Gram-positive bacterium [1]. Its biochemistry, physiology and genetics have been studied intensively for more than fifty years [2]. As a result, a great deal of vital information concerning its transcription and translation mechanisms, genetic manipulation, and large-scale fermentation has been acquired [3]. In addition, it has a long history of industrial use because of its excellent growth on cheap carbon sources, and robustness under industrial conditions [4]. Another advantage is that B. subtilis is regarded as a Generally Recognized as Safe (GRAS) organism that lacks endotoxins and is non-pathogenic [5,6]. Furthermore, it has no significant bias in codon usage [3]. Therefore, it has become an ideal bacterial 'factory' for recombinant protein production and a large variety of expression vectors have been created. The vectors may be inducible or inducer-free (constitutive or autoinducible) [7][8][9]. Inducer-free expression vectors have recently been gaining increased popularity because of the high costs [10] of many inducer compounds and problems with their toxicity [11].
Traditional expression systems utilized high copy number plasmids introduced into B. subtilis to create recombinant strains for heterologous protein expression [3]. However, plasmid-less engineered B. subtilis strains are preferred in industrial applications due to their stability and lower ecological risk [12]. Formerly, plasmid-free recombinant B. subtilis strains were constructed by homologous recombination between the target sequence in the chromosome and the homologous flanking sequences sandwiching the fragment of interest mediating ectopic insertion of the desired genes into the bacterial chromosome [3]. Recently, several new vectors have been developed resulting in more efficient integration. Some of them allow production of recombinant proteins by incorporating many copies of the recombinant gene at different sites in the bacterial genome [5,13]. For example, an integrated vector engineered with the strong inducer-free promoter NBP3510 exhibited β-galactosidase (BgaB) and GFP (green fluorescent protein) production of 43 % and 30 % of the total cellular proteins, respectively [14].
The unsolved problem in the generation of useful vectors allowing robust B. subtilis expression systems is their leaky expression in E. coli. Most of the expression vectors for B. subtilis are shuttle-vectors because the cloning steps must be carried out in E. coli [3], which can sometimes result in unexpectedly high protein expression levels [15,16], and, in the worst case, kill the E. coli cells. Previous publications did not mention whether strong promoters for B. subtilis also cause high background expression in E. coli. Here, we constructed inducerfree expression vectors allowing the integration of the recombinant gene into the B. subtilis genome. These vectors carried strong promoters and their expression level in B. subtilis was comparable to that from high copy number replicative plasmids while expression in E. coli remained relatively low. In addition, double-copy insertion strains were generated by the integration of the recombinant gene into two different neutral loci. The best resulting strains had an expression level of over 50 % of the total intracellular proteins.

Bacterial strains, plasmids and growth conditions
Plasmids, oligonucleotides, and strains used in this study are shown in Table 1. The E. coli strain OmniMAX (Invitrogen) was used as the recipient in all cloning experiments and to determine the background expression levels. All recombinant B. subtilis strains used to analyze the expression of the bgaB gene were derived from B. subtilis 1012. Cultures were initiated from single-colony inocula grown on LB agar plates. Cells were routinely grown in Luria broth (LB) at 37 C with shaking at 200 rpm. Where necessary, the antibiotics ampicillin at 100 mg/mL for E. coli and chloramphenicol at 10 mg/mL for B. subtilis were added to recombinant strains harboring replicative plasmids.

Expression vectors integrating into the lacA locus
We first created the three basic inducer-free integrative expression vectors, pHT2171, pHT2184 and pHT2188, and then, the bgaB gene was introduced into these vectors to obtain pHT2172, pHT2185 and pHT2189. pHT2171 was constructed by inserting the cassette containing the lacO3 operator and the Pgrac01 promoter (amplified from pHT2134 as the template, a derivative of pHT2071 [19] carrying a new MCS, with ON2195/ ON2194) into the pHT1305 empty vector which contains sequences to allow integration at the lacA locus on the B. subtilis genome without any promoter. pHT2184 was constructed by insertion of the cassette harboring lacO3 and the Pgrac100 promoter into the pHT1326 backbone. This backbone has no promoter but contains the homologous sequences allowing integration into the lacA locus on the B. subtilis genome. pHT2184 was first cleaved with SacI and BamHI to remove Pgrac100. Then, it was ligated to Pgrac212 amplified using the pHT2080 template to create pHT2188.

Expression vectors able to integrate into the amyE locus
To construct plasmids pHT2170, pHT2176, and pHT2177, we removed lacI together with the lacO3 sequence from plasmids pHT2115, pHT2118, and pHT2119, respectively, and inserted the lacO3 sequence between the SnaBI and SacI restriction sites by using the two complementary oligonucleotides ON1975 and ON1976.

Generation of E. coli and B. subtilis recombinant strains
Expression vectors were confirmed by DNA sequencing. The correct vectors were transformed into E. coli OmniMAX as described in [22] and into B. subtilis 1012 as described elsewhere [23]. Recombinant B. subtilis strains generated by double crossover events were screened by PCR using specific oligonucleotide pairs. Fig. 2A and D shows specific primers with the length of the PCR products for the strains carrying Pgrac01 that integrate into amyE and lacA loci. The same PCR approach has been used to confirm the integration of the expression cassettes containing Pgrac100 and Pgrac212.

Measurement of BgaB expression levels in E. coli and B. subtilis
Recombinant strains were streaked on LB agar plates with the appropriate antibiotic: ampicillin at 100 mg/mL for E. coli and chloramphenicol at 10 mg/mL for B. subtilis harboring replicative plasmids, and no antibiotic for B. subtilis strains carrying integrated sequences. A single colony was inoculated into a culture tube containing 5 mL LB medium and antibiotic and shaken overnight at 200 rpm at 37 C. Cultures of each strain were replicated using three separate colonies. The OD 600 of the preculture was measured and an appropriate volume of pre-culture of each clone was transferred to 30 mL LB medium containing the appropriate antibiotic in 100 mL shake flasks to give an OD 600 of 0.1 and incubated with shaking at 37 C. When the OD 600 of the culture reached 0.8-1, the cells were divided into two sub-cultures and one of them was induced by the addition of IPTG to a final concentration of 1 mM. Cells were collected by centrifugation at 0 h just before induction and at 2, 4, 6, 8, 10, 12 h after induction. The OD 600 of all sub-cultures was monitored and a volume equivalent to an OD 600 of 2.4 was pipetted into 1.5 mL Eppendorf tubes, centrifuged, and the supernatants removed. Samples were prepared for activity measurements and for SDS-PAGE analysis. BgaB activity was measured as described in [19].
For SDS-PAGE analyses, cell pellets were lysed by the addition of 100 mL lysis buffer (25 mM SDS, 250mM sucrose) with an addition of 2.5 mL lysozyme at 50 mg/mL. The mixtures were vortexed thoroughly and incubated at 37 C for 5 min. After that, 25 mL of 5X sample buffer (10 mL Tris ÀHCl pH 6.8, 1.54 g dithiothreitol, 0.4 g SDS, 4 mL glycerol, bromophenol blue, and dH 2 O up to 20 mL) was added. The samples were mixed well and heated at 95 C for 5 min followed by centrifugation at 15871 rcf for 5 min. Aliquots of 8 mL of each sample were applied to each well on SDS-PAGE gels [24]. DlacI, deletion of partial or full lacI gene; Pgrac01 (another name is Pgrac) [20]; Pgrac100 [21] and Pgrac212 [18] are the names of different promoters. Fig. 4D shows the alignment of the promoter sequences.

Background expression levels of Pgrac01-bgaB constructs in E. coli
Expression vectors for B. subtilis are shuttle-vectors, and the cloning steps are carried out in E. coli. However, a vector that could drive a high level of expression in B. subtilis often results in a high level of expression in E. coli even in the absence of inducers. Previously, we reported a new strategy to construct inducer-free plasmids based on IPTG-inducible promoters. By deleting part of the lacI gene in IPTG-inducible pHT vectors, the resulting plasmids could express the target proteins in the absence of IPTG in B. subtilis while still repressing the background expression in E. coli through the chromosomal lacI gene [19]. To generate expression vectors for B. subtilis which express the recombinant protein at high levels in B. subtilis while performing very low expression in E. coli, we used this strategy to develop inducer-free integrative expression vectors based on Pgrac promoters as reported earlier [19]. The promoters are flanked by two lac operators, lacO1 at the downstream and lacO3 at upstream of the promoters, which promote the repression of leaky expression by DNA-loop formation in the presence of the LacI repressor [25]. Besides, to increase the stability of the vectors as well as to lower the background expression in E. coli, we introduced the rop gene into the vectors. The Rop protein decreases the copy number of ColE1-like plasmids by stabilizing the RNAI-RNAII duplex [26].
An integrative expression vector harboring the Pgrac01 promoter and the bgaB reporter gene was constructed, and the leaky expression level was compared with that of some other replicative plasmids.
While pHT01-bgaB (Pgrac01-bgaB, containing lacI) could repress leaky expression in E. coli OmniMAX about 16-fold, pHT2071 (Pgrac01-bgaB, DlacI) could repress it only about 3-fold, and pHT2170 (Pgrac01-bgaB, DlacI, rop) could repress it about 24fold (Fig. 1A). These results demonstrate that the combination of two lac operators together with the rop gene in the pHT2170 plasmid created optimal repression of background expression in E. coli. The leaky expression of the bgaB gene from the inducer-free integrative expression vectors was the lowest among these vectors analyzed (Fig. 1B). Moreover, the leaky expression from pHT2170 was approximately similar to pHT1379 which served as the negative control which did not express bgaB. These results demonstrated that pHT2170 could repress the leaky expression in E. coli thereby allowing the cloning steps. The conceptual figure (Fig. 1C) shows the repression of the Pgrac promoter by plasmid and chromosomal lacI and control of plasmid copy via rop gene in E. coli cloning strain.

Inducer-free expression of Pgrac01-bgaB integrated into the B. subtilis genome
The expression cassette of the vector pHT2170 consisted of two homologous sequences of the amyE gene from B. subtilis genome, the spc R gene for selection, the Pgrac01 promoter and the bgaB reporter gene. This vector does not include an origin of replication for B. subtilis; therefore, when it is transformed into B. subtilis cells in the presence of spectinomycin, the expression cassette is integrated into the genome by either a single or a double crossover event. The latter was confirmed by PCR ( Fig. 2A, B) and the resulting strain was named B. subtilis HT2170. The expression cassette was stably maintained in the genome of this strain even in the absence of spectinomycin. The bgaB expression level in B. subtilis HT2170 in the absence of IPTG and antibiotics was tested and compared with those of some other replicative plasmids carrying the same or a different promoter. HT2170 synthesized the BgaB protein in a manner different from strains carrying pHT01-bgaB or pHCMC05-bgaB, but in the same way as with pHT2071. While pHT01-bgaB and pHCMC05-bgaB expressed bgaB only in the presence of IPTG, pHT2170 synthesized BgaB similarly in the absence or presence of IPTG (Fig. 3A). SDS-PAGE analyses (Fig. 3B) confirmed this result. These results indicate that pHT2170 can efficiently express bgaB in an inducer-free manner. The best performance of pHT2170 was 6.66 Â 10 4 methylumbelliferyl β-Dgalactopyranoside (MUG) units while that of pHCMC05-bgaB with the Pspac promoter reached only 0.47 Â 10 4 MUG units. The BgaB activity of pHT2170 was about 14-fold higher than that obtained with the multi-copy replicative Pspac plasmid. The SDS-PAGE results (Fig. 3B) also confirmed that the expression level of the inducer-free integrative vector pHT2170 was much higher than that of pHCMC05-bgaB. However, the BgaB activity of pHT2170 was lower than that obtained with replicative plasmids carrying the same promoter. The BgaB activity of pHT2170 was about 80 % of pHT01-bgaB after induction with 1 mM IPTG and about 50 % of that obtained with pHT2071 (Fig. 3A). In addition, SDS-PAGE analysis revealed that BgaB expressed by pHT2170 accounted for only 8% of the total protein while pHT2071 expressed 14 %, and pHT01-bgaB about 10 % (Fig. 3B).
The expression of B. subtilis HT2170 was 1.25 or 1.75 times less than plasmid expression carrying the same promoter, most probably because of the lower copy number. Therefore, to generate integrative vectors whose expression level was comparable with that of multi-copy plasmids, we next constructed integrative vectors with stronger promoters.

High expression levels of the bgaB reporter gene using stronger Pgrac promoters
Next, we aimed to construct inducer-free integrative vectors with high-level expression by using the strongest promoters from the Pgrac library. On example is the engineered Pgrac100 promoter whose UP element, the -35 and -15 sequences have been mutated allowing intracellular accumulation of BgaB up to 30 % [27]. Pgrac212 is structurally similar to Pgrac01 containing modifications at the controllable stabilizing element (CoSE; the region from +1 to the RBS) [15] resulting in BgaB levels within the same range as compared to Pgrac100 [21]. Therefore, Pgrac100 and Pgrac212 were chosen. The highest BgaB activities were 6.9 Â 10 4 MUG units for the vector carrying the Pgrac01 promoter, 9.1 Â10 4 MUG units for those with Pgrac100, and 14.4 Â 10 4 MUG units for those with Pgrac212 (Fig. 4A). Using the stronger promoters, bgaB expression increased 1.3 up to 2.1-fold. Analysis of Fig. 4B and 4C by Alpha Ease 4.0 showed that BgaB expressed by Pgrac100 vectors accounted for 20.9 % and Pgrac212 vectors accounted for 42 % of the total intracellular proteins after 12 h. BgaB expressed by HT2176 -Pgrac100 accounted for 9% of the total cellular protein at 2 h, 12 % at 4 h and 20.9 % at 12 h (Fig. 4B). Similarly, the amount of BgaB produced by HT2177-Pgrac212 was low at the early stage of the culture, but rose rapidly with time-11 % of the total protein at 2 h, 32.7 % after 4 h, finally reaching 42 % after 12 h, or about 4-fold over time in culture (Fig. 4C).

Increasing expression levels by integrating Pgrac212-bgaB into both amyE and lacA loci
Another way to increase the expression level in B. subtilis strains carrying the expression unit in the genome is to increase the copy number of the expression unit. Several methods are available for multi-copy insertions of the target gene. Among them, integration at two and more different chromosomal sites may be the most stable [28], so the expression cassette was inserted into both the amyE and the lacA loci to increase the magnitude of expression of the target gene. The vector was first transformed into competent B.
subtilis cells and the integration occurred at the amyE locus in the presence of spectinomycin. Next, a second vector with homologous sequences to the lacA locus was transformed into the strain already carrying one copy of the expression unit at the amyE locus in the presence of neomycin. The recombinant strains possessed two expression cassettes in the genome were confirmed by PCR (Fig. 2C, D), and bgaB expression was evaluated by the MUG assay and by SDS-PAGE.
Integration of the bgaB gene into both the amyE and the lacA loci doubled the copy number of the gene and the synthesis of BgaB from both loci was much higher than the expression from either the amyE or the lacA locus under control of the same promoter. Strains with expression cassettes containing Pgrac01, Pgrac100, or Pgrac212 integrated at the two loci expressed BgaB at levels of 23.4 %, 24 %, and 53.4 %, of total proteins, respectively. The expression of BgaB was increased from 1.1 to 1.3-fold (Fig. 5A). The strain with Pgrac212 expression cassettes inserted into both loci synthesized    BgaB at 11 % after 2 h of induction, 34.7 % after 4 h and continued increasing to 53.4 % at 12 h (Fig. 5B).

Discussion
Due to its many favorable characteristics, B. subtilis serves as an excellent cell factory for the production of heterologous proteins. However, competent B. subtilis cells for cloning experiments are present only in low numbers resulting in poor transformation efficiency. The competence problem has been overcome by performing the cloning steps in E. coli using B. subtilis -E. coli shuttle vectors [3], but leaky protein expression in E. coli can hamper the cloning efficiency. The strong promoters that allow a high level of expression in B. subtilis also drive leaky expression to high levels in E. coli. Many regulatable vectors for strong protein synthesis in B. subtilis result from the incorporation of a promoter with the lac operator [29][30][31][32]. This type of hybrid promoter is used in IPTG-inducible vectors to prevent unwanted expression in E. coli. However, the difference between the presence of the lacI gene in B. subtilis and in the E. coli genome allows lacO-hybrid promoters to be used to create vectors that are inducer-free in B. subtilis but remain controllable in E. coli. In a previous study, we reported a new strategy for generating inducer-free replicative plasmids [19] to achieve high-level expression in B. subtilis while suppressing leakiness in E. coli; but the level of background expression in E. coli still remained high. These vectors were based on ColE1-like plasmids and the rop gene could be used to reduce the expression of leaky promoters in E. coli by lowering the plasmid copy number. The gene encoding the Rop protein is removed in most vectors to increase their copy number in E. coli [33], but we chose to introduce it into our vectors because they were constructed for recombinant protein synthesis in B. subtilis where low plasmid copy number in E. coli would not be a problem. On the contrary, it helps to reduce background expression in this host. In this study, the combination of the rop gene and the two lac operators in our vectors led to 24-fold repression in recombinant E. coli carrying pHT2170 with the Pgrac01 promoter compared to only 3-fold with the previously reported strategy, significantly decreasing the background expression. After 4 h, the leaky BgaB was only around 230 MUG units, 19-fold lower than with pHT2071 (Pgrac01, replicative, inducer-free, high copy number) and 4-fold lower than pHT01-bgaB (Pgrac01, replicative, IPTG-inducible, high copy number). The conceptual figure (Fig. 1C) shows the background in E. coli and the inducible or inducer-free expression in B. subtilis in the host.
Replicative plasmids can have problems with stability and safety however, so bioengineers are turning more to vectors integrated into the host's genome. A single integrated expression cassette may not produce the desired expression level, so efforts are being made to increase the copy number in the genome in order to boost expression. Ten copies of an mpr B.amy cassette in which the GSP gene was placed between the promoter of the B. amyloliquefaciens rplU-rpmA gene and the Rho-independent transcription terminator were ectopically inserted into designated (3 copies) and random (7 copies) sequences into the recipient's DNA. The resulting bacterial strain produced approximately 0.5 g/L of secreted GSP after cultivation in flasks with starch-containing media, and its performance was comparable to an analogous strain in which the mpr B.amy cassette was carried on a multi-copy plasmid [12]. In another study, nine copies of a arg R.pyc cassette containing the Rummeliibacillus pycnus arginase gene regulated by the strong promoter P43 were inserted into the recipient's genome. Tests showed that the highest arginase activity (14.5 U/mL) was obtained from flask cultures, and this segregation-stable strain could efficiently hydrolyze L-arginine with a 97.2 % molar yield, suggesting a potential application for the food industry [28]. A strong promoter was engineered that allowed synthesis of BgaB and sfGFP to levels of 43 % and 30 % of intracellular proteins, respectively. It was also used to allow the secretion of methyl parathion hydrolase (MPH) and chlorothalonil hydrolytic dehalogenase (Chd) to a level of 0.3 g/L (144 U/mL) and 0.27 g/l (4.4 U/mL) using shake-flask culture conditions [14]. In our study, the strong Pgrac promoters were used to generate effective B. subtilis -E. coli inducer-free integrative vectors. The best performance of plasmid- less strain with a single genomic copy of the BgaB expression cassette produced target protein to 42 % of total protein. With ectopic insertion into both amyE and lacA loci, the BgaB yield reached 53.4 % of total protein. A series of different integrative vectors with a variety of expression levels were created to meet different protein expression needs.
Inducer-free expression vectors (constitutive and auto-inducible) avoid the need to add an inducer to the culture medium thereby lowering the production cost. Constitutive promoters are not suitable for the production of toxic proteins, but autoinducible promoters are ideal for large-scale commercial protein production. Such promoters induce expression of the target gene from the late log phase to the stationary phase with no requirement for an inducer, which facilitates high-yield production of heterologous proteins at low cost [34]. The silencing of the lacI gene in designated vectors allowed constitutive expression in B. subtilis. The best performance was obtained with vectors expressing the target protein at low levels during early culture stages but switching to high production when host cells reach the late log phase. As demonstrated here, B. subtilis carrying the Pgrac212 cassette expressed BgaB protein only up to 11 % during the first two hours after induction, then the yield increased to 30 % over the next two hours when the cells begin to enter stationary phase. These positive results show the potential commercial value of our newly constructed vectors. These efforts also constitute an extension of our previous investigation of B. subtilis as a vaccine delivery vector. We showed that the expression of small amounts of LTB in the cytoplasm or anchored on the cell surface by a sortase [35,36] could induce a humoral immune response in mice [37]. The use of a vector carrying Pgrac212 linked to a gene encoding the human rhinovirus 3C protease resulted in the production of recombinant protein up to 16 % in B. subtilis after IPTG induction [38]. Recent reports using different expression systems showed the expression of various recombinant proteins in B. subtilis such as α-amylase, PhoA, single-chain variable antibody fragment, RNase barnase, trehalose synthase, human FGF21 [39][40][41]. Therefore, our inducer-free expression system could be a costeffective solution for synthesizing recombinant proteins or vaccines for animals.

Declaration of Competing Interest
The authors report no declarations of interest.