Development of Pgrac100-based expression vectors allowing high protein production levels in Bacillus subtilis and relatively low basal expression in Escherichia coli

Background In general, fusion of recombinant genes to strong inducible promoters allowing intracellular expression in Bacillus subtilis is a two-step process. The ligation products are transformed into Escherichia coli, followed by identification of the correct plasmid, and this plasmid is subsequently transformed into B. subtilis. This raises the problem that basal level of expression of the recombinant gene could be harmful for E. coli cells. Based on the Pgrac promoter, we optimized the UP element, the -35, 15, -10 and the +1 region to enhance the promoter activity in B. subtilis after induction. However, detailed investigations for a promoter to develop expression vectors that allows high protein production levels in B. subtilis and a relatively low basal expression levels in E. coli has not been studied yet. Results We screened the previously constructed library of E. coli – B. subtilis shuttle vectors for high level expression in B. subtilis and low basal level in E. coli. Promoter Pgrac100 turned out to meet these criteria, in which ß-galactosidase expression level of Pgrac100-bgaB is about 9.2 times higher than Pgrac01-bgaB in B. subtilis and the ratio of those in induced B. subtilis over un-induced E. coli from Pgrac100-bgaB is 1.3 times higher than Pgrac01-bgaB. Similarly, GFP expression level of Pgrac100-gfp is about 27 times higher than that of Pgrac01-gfp and the ratio from Pgrac100-gfp is 35.5 times higher than Pgrac01-gfp. This promoter was used as a basis for the construction of three novel vectors, pHT253 (His-tag-MCS), pHT254 (MCS-His-tag) and pHT255 (MCS-Strep-tag). Expression of the reporter proteins BgaB and GFP using these expression vectors in B. subtilis at a low IPTG concentration were measured and the fusion proteins could be purified easily in a single step by using Strep-Tactin or IMAC-Ni columns. Conclusions This paper describes the construction and analysis of an IPTG-inducible expression vector termed Pgrac100 for the high level production of intracellular recombinant proteins in B. subtilis and a relatively low basal expression level in E. coli. Based on this vector, the derivative vectors, Pgrac100-His-tag-MCS, Pgrac100-MCS-His-tag and Pgrac100-MCS-Strep-tag have been constructed.


Background
The production of heterologous proteins in different microbial systems has revolutionized biotechnology. Most expression systems are based on an inducible promoter, and addition of the appropriate inducer leads to the production of the heterologous protein in most cases intracellularly. Microbial expression systems have been described for bacteria, yeast, filamentous fungi, and unicellular algae. All these systems have advantages and disadvantages, which have been extensively discussed [1][2][3].
Escherichia coli is the most widely used bacterial host to synthesize recombinant proteins for biochemical and functional studies. E. coli cells are easy to culture since they have a very short doubling time in rich media, and they are easy to manipulate genetically. Three disadvantages related to E. coli are: (1) low expression of some heterologous genes; (2) some heterologous proteins are insoluble and form inclusion bodies; and (3) contamination of the heterologous proteins by the endotoxin LPS [4,5].
Bacillus subtilis is an attractive alternative host for heterologous protein production and engineering because of the following reasons: (1) it can secrete proteins efficiently, especially homologous proteins up to 20 g/l; (2) it is nonpathogenic; (3) and it has been granted the GRAS (generally regarded as safe) status by the American Food and Drug Administration [6][7][8]. The authors have developed several plasmid-based expression vectors exhibiting structural stability [9], where induction can be accomplished by addition of xylose [10], IPTG [11], glycine [12] or by a cold shock [13].
Promoter P spac is one of the most-popular promoters used for expression of heterologous proteins in B. subtilis, but it is rather weak [14]. The IPTG-inducible P grac promoter is 50 times stronger than P spac , and it has been derived from the B. subtilis groESL promoter and the E. coli lac operator [15,16]. To further improve protein expression levels, we created a library of a second generation P grac promoters by either introducing promoter mutations in the consensus regions resulting in stronger promoters [11] or by applying the mRNA controllable stabilizing elements (CoSE) [17]. However, enhancing the protein expression levels in B. subtilis also leads to higher basal expression levels in E. coli. In addition, some normal genes are not supposed to be harmful for E. coli, but it can inhibit the growth at high background expression levels, for examples ß-galactosidases, BgaB from Geobacillus stearothermophilus and LacZ from Escherichia coli [18]. Therefore, expression vectors harboring promoters that control high protein production levels in B. subtilis after induction and allow a low basal level of expression in E. coli are of utmost importance. This study aims the development of expression vectors for B. subtilis based on a promoter that allows high inducible protein production in B. subtilis and relatively low basal level in E. coli.

Screening for an appropriate promoter
Identification of a suitable inducible promoter controlling high production levels of recombinant proteins in B. subtilis and, at the same time, retaining relatively low basal levels in E. coli in the absence of the inducer is an important requirement during construction of expression vectors for B. subtilis. To accomplish this goal, we used the Pgrac-promoter library described [11,17] and screened for low BgaB expression in E. coli by using the method described for B. subtilis [18]. During screening of a library of 84 different promoters, we analyzed the BgaB expression levels based on the blue color of the colonies on the X-gal plates in the absence of IPTG for E. coli and in the presence of 0.01 mM IPTG for B. subtilis. This IPTG concentration was used based on our previous results showing that IPTG and BgaB expression levels (activity) were linear for Pgrac01 and promoters stronger than Pgrac01 at IPTG concentrations from 0.0025 to 0.025 mM [18]. As examples, we analyzed the three promoters Pgrac01, previously called Pgrac [11], Pgrac100 and Pgrac212. Pgrac01 is at least 50-times stronger than Pspac [14] based on BgaB activities and allowed BgaB protein accumulation up to 9.1 % of total the cellular proteins [11,15]. Pgrac212 is structurally similar to Pgrac01 containing modifications at the controllable stabilizing element (CoSE)the region from +1 up to the RBSresulting in BgaB levels within the same range as compared to Pgrac100 [17]. Pgrac100 is different from Pgrac01 at the UP-element (−44-TCTTATCT-37 -> −44-AAAAATCT-37), the −35 motif (TTGAAA -> TTGACA), and the −15 region (−16-TCT-14 -> −16-ATG-14) (Fig. 1a). The negative control plasmid, Pgrac01 without the bgaB gene exhibited white colonies for both B. subtilis and E. coli on X-gal plates (Fig. 2a). When the strength of the 84 different promoters was analyzed on X-gal plates, Pgrac100-bgaB and Pgrac212-bgaB exhibited a stronger blue color in comparison to Pgrac01-bgaB in B. subtilis in the presence of IPTG (Fig. 2a). When these plasmids were analyzed in E. coli in the absence of the inducer, Pgrac212-bgaB exhibited the strongest blue color, followed by Pgrac100-bgaB and Pgrac01-bgaB (Fig. 2a). E. coli colonies sometimes showed that only a part of the colonies were blue. However, this is not an indication that the plasmids were structurally unstable. The stability of the plasmid backbone derived from pHT01-bgaB was confirmed previously [16]. Calculation of the grey values from these colonies confirmed the result observed by eyes. By screening the 84-promoter library, Pgrac100 appeared to be the most appropriate one that met the criteria for an optimal inducible promoter. It has a relatively low background level in E. coli and a high inducible expression level in B. subtilis.

Choice of promoter Pgrac100
To have a clearer picture of the Pgrac100 promoter, we measured the ß-galactosidase (BgaB) activities of potential promoter candidates from cells grown in liquid LB medium for both E. coli in the absence of IPTG and B. subtilis after addition of the inducer. The ratios of ß-galactosidase activities obtained with B. subtilis and E. coli were calculated, representing the promoter strengths in both species. High activity in B. subtilis and high ratio numbers indicate a better promoter. We used Pgrac01 (formerly Pgrac) as the reference. The B. subtilis and E. coli cells containing pHT01 (Pgrac without bgaB, negative control) do not produce detectable ß-galactosidase activity. As an example, Fig. 2b shows that Pgrac100-bgaB has a higher ratio than Pgrac01-bgaB and more than three times higher ratio than Pgrac212-bgaB. In addition, Pgrac100-bgaB is about 9.2 times higher than Pgrac01-bgaB after induction at 0.01 mM IPTG (Table 1). When these two values were compared with those obtained with other promoters in our library, the BgaB activities indicated that Pgrac100 is the most appropriate candidate that controls high production levels of recombinant proteins in B. subtilis and at the same time maintains a relative low background expression in E. coli (data not shown).
The results in Table 1 also showed that Pgrac100-bgaB seems to be characterized from high basal expression (222 ± 68 units) in B. subtilis. If low basal expression in E.coli is important to facilitate the cloning of toxic genes, the presence of basal expression in B. subtilis could make difficult the plasmid transformation. However, it also indicated that we could use Pgrac100 promoter for high production levels of recombinant protein at low concentration of IPTG inducer. If we consider low background expression levels in E. coli and B. subtilis, selection of Pgrac01 [16] could be an option.
In comparison with other systems, we transformed PxylA-bgaB (pHCMC04-bgaB) and Pspac-bgaB (pHCMC05-bgaB) [9] into E. coli and spread transformants on X-gal plates, and the E. coli colonies developed blue color. Colonies of Pspac-bgaB were within the same range as those from Pgrac01-bgaB, while colonies from PxylA-bgaB were deeper blue than the others (Fig. 2c). When the E. coli cells were growth in liquid LB medium in the absence of the inducers, BgaB activities from Pspac-bgaB were equal to that of Pgrac01-bgaB, while those from PxylA-bgaB were within the same range as those from Pgrac100-bgaB (Fig. 2d). In B. subtilis, the BgaB expression levels of the two constructs, Pspac-bgaB and PxylA-bgaB in the presence of inducers were within the same range [9] and 50 times lower than Pgrac01-bgaB [15,16]. Though Pspac expressed lower basal levels than and PxylA as high as Pgrac100 in E. coli, the expression levels in B. subtilis was also very low in the presence of inducer. Therefore, these promoters are not appropriate to be used for over-production of recombinant proteins in B. subtilis.

Important factors of Pgrac100 in controlling GFP expression
Though BgaB is a popular reporter protein for B. subtilis, it has heterogeneous properties in E. coli [19]. In order to confirm the properties of Pgrac100, we replaced the bgaB gene by gfp + (pHT100-gfp; Table 2) and analyzed for GFP expression. The background expression level of GFP from Pgrac100-gfp in E. coli is 37 RFU (Relative Fluorescence Unit), while that of Pgrac01-gfp is 68 RFU. In addition, the ratio of GFP activities of the  [16]; b, DNA sequences of the promoters present in Pgrac01 and Pgrac100, where the differences between these two promoters are underlined (UP element, −35, −15 regions) [11]; c, DNA sequence downstream of the RBS of plasmid pHT254, including the multi-cloning sites, the start codon, the His-tag and the stop codon (BamHI-Start codon-XbaI-AatII-His-tag-Stop codon/TAA-SmaI) background expression level in E. coli and in B. subtilis cells induced with 0.01 mM IPTG in for Pgrac100-gfp turned out to be 15.3 while that of Pgrac01-gfp was 0.5 ( Table 1) and that of Pgrac212-gfp 0.2. These results clearly confirm that the promoter Pgrac100 is able to tightly control protein expression in E. coli at the same range as compared with Pgrac01.
The expression levels of GFP of Pgrac100-gfp increased after addition of IPTG and reached up to 568 RFU at 0.01 mM IPTG, 27-fold higher than that of Pgrac01-gfp (Table 1) and 4.7-fold higher than that of Pgrac212-gfp. In addition, we calculated the induction factor and the ratio of the activities of induced and uninduced samples. Pgrac100 exhibited an induction factor of 9 at 0.01 mM IPTG and of 25 at 0.1 mM IPTG (Table 1), while those of Pgrac01 were 2.6 and 24.7, respectively ( Table 1) and those of Pgrac212 6.3 and 77 (data not shown). Similar results using BgaB as reporter were also observed for Pgrac100 (Table 1). The substantial differences in protein expression levels between BgaB and GFP might be because they come from two different organisms, BgaB from G. stearothermophilus and GFP from Aequorea victoria, and the sequences of the genes might influence the transcription and/or translation efficiency in E. coli and B. subtilis. These results demonstrate that Pgrac100 not only tightly controls the background expression level in E. coli, but also allowed high protein production levels at low IPTG concentrations. subtilis 1012 on X-gal plates and in liquid medium. a, Bacterial cells containing pHT01 (Pgrac, negative control), pHT01-bgaB (Pgrac01-bgaB), pHT100 (Pgrac100-bgaB) and pHT212 (Pgrac212-bgaB) were spotted on X-gal LB agar plates containing appropriate antibiotics and 0.01 mM IPTG for B. subtilis and without IPTG for E. coli at 30°C for 48 h. Then, pictures were taken and single colonies are shown. b, The bacterial cells were grown in liquid LB medium at 37°C to the mid-logarithmic growth phase, and then induced with 0.01 mM IPTG for B. subtilis and kept un-induced for E. coli. The cells were collected after 4 h of induction, and the BgaB activities were measured. The ratio of β-galactosidase activities of the samples were calculated from induced B. subtilis cells and un-induced E. coli cells. The ratio was set as one when the BgaB activities from both E. coli and B. subtilis were identical [4,5]. c, E. coli cells containing pHT01 (Pgrac, negative control), pHT100 (Pgrac100-bgaB), pHCMC04-bgaB (PxylA-bgaB), pHCMC05-bgaB, pHT01-bgaB (Pgrac01-bgaB were spotted on X-gal LB agar plates containing ampicillin. d, the E. coli cells were grown in liquid LB medium at 37°C to the mid-logarithmic growth phase, then the growing cells collected and the BgaB activities were measured. The ratio of β-galactosidase activities of the samples were calculated from different constructs to Pgrac01-bgaB In summary, promoter Pgrac100 is an excellent choice for the construction of inducible expression vectors for B. subtilis.

Construction of basic expression vectors
The above result demonstrated that promoter Pgrac100 allowed high protein production levels in B. subtilis and low background expression levels in E. coli by using two reporter proteins, BgaB and GFP. To generate tagging expression vectors, we removed bgaB from pHT100 [11] and added the DNA fragments containing start codon-His-tag-BamHI-XbaI-AatII-SmaI, BamHI-start codon-XbaI-AatII-His-tag-stop codon/TAA or BamHI-start codon-XbaI-AatII-Strep-tag-stop codon/TAA, resulting pHT253, pHT254, and pHT255, respectively ( Table 2). Fig. 1c shows the DNA sequence of the multi-cloning site, the His-tag, the start and the stop codon from pHT254. The other plasmids were adapted appropriately to meet these requirements. The full sequences of these three plasmids were similar to pHT01 [16] except for the promoter regions and the multi-cloning sites with different tags. The map of plasmid pHT254 is shown in Fig. 1a. Fig. 2b indicates the differences between promoter Pgrac01 and Pgrac100 at the UP element and the −35 and −15 regions. Target genes can be introduced using restriction enzymes BamHI, XbaI, AatII or SmaI as fusions or non-fusions with either a His-tag or a Streptag at the N-or C-terminus.

Evaluation of the expression vectors in B. subtilis
To evaluate the basic expression vectors pHT253, pHT254, and pHT255, we introduced gfp + or bgaB as translational fusions with 8xHis-or Strep-tags resulting in pHT1169 (8xHis-gfp), pHT1170 (gfp-8xHis), pHT1171 (gfp-Strep) and pHT1178 (8xHis-bgaB), pHT1179 (bgaB-8xHis) and pHT1180 (bgaB-Strep) ( Table 2). Plasmids containing the Strep-tag at the N-terminus fused with the reporters were also constructed, but the production levels were very low (data not shown). Fig. 3 shows expression of BgaB and of GFP fused to the His-or Strep-tag under control of Pgrac100 after induction with 0.1 mM IPTG. The His-tag at the N-terminus in plasmid pHT253 drastically reduced the expression levels of BgaB, reaching 6.2 % of the total cellular proteins (Fig. 3a) and GFP (Fig. 3d) compared to the fusions at the C-terminus and in the absence of any tag. The expression levels of BgaB and GFP in these constructs are equal to those in pHT01-bgaB (Pgrac01-bgaB) and pHT10-gfp (Pgrac01-gfp) [16] in terms of their activities. These results indicate the expression levels of Pgrac100 with the His-tag at the N-terminus are comparable to Pgrac synthesizing BgaB and GFP.
The fusions, BgaB-His (Fig. 3b), BgaB-Strep (Fig. 3c), GFP-His (Fig. 3e) and GFP-Strep (Fig. 3f ) are produced at levels comparable to those without a purification tag, BgaB (from pHT100-bgaB) and GFP (from pHT100-gfp) deduced from SDS-PAGE gels. The BgaB expression levels could reach up to 30 % of total cellular proteins [11], while the tagged versions accumulated 24 % using 0.1 mM IPTG (Fig. 3b and c) and up to 30 % using 1 mM IPTG. Similarly, the untagged and the C-tagged versions of GFP could be produced at 15 % of total cellular proteins on the average (Figure 3e and f ). However, the expression levels of the fusions at low concentrations of IPTG were lower than the untagged constructs. Besides in B. subtilis 1012, we also checked the expression in B. subtilis WB800N [20], a derivative of WB800 [21], a The data for BgaB and GFP activity presented have been obtained with pHT01-bgaB (Pgrac01-bgaB), pHT10-gfp + (Pgrac01-gfp) pHT212 (Pgrac212-bgaB) and pHT100-gfp (Pgrac100-gfp). BgaB activity is shown in Miller units while GFP indicated as activity is relative fluorescence unit (RFU). All experiments were carried out from at least three different colonies, and standard errors were calculated derivative of strain 168. The expression levels in WB800N were similar to those in 1012 (data not shown). These results indicate that the expression vectors pHT253, pHT254, pHT255 could be used for overproduction of recombinant proteins to high levels in different B. subtilis strains.

Conclusions
We show that the artificial promoter Pgrac100 could be used for the construction of His-or Strep-tagged versions, pHT253, pHT254 and pHT255 for B. subtilis. These three new expression vectors provide two advantages: (i) allowing high production levels of recombinant proteins in B. subtilis after induction and (ii) maintaining relatively low background expression levels in E. coli.

Methods
Bacterial strains, plasmids and growth conditions E. coli strain OmniMAX (Invitrogen) was used as the recipient in all cloning experiments and to determine the expression levels. B. subtilis strains, 1012 [22] and WB800N [20] were used to analyze expression of the bgaB and gfp+ genes. A list of the plasmids and oligonucleotides used in this study is shown in Table 2. Cells were routinely grown in Luria broth (LB) at 37°C under shaking at 200 rpm. Antibiotics were added where appropriate (ampicillin at 100 μg/mL for E. coli and chloramphenicol at 10 μg/mL for B. subtilis).

Construction of plasmids
The plasmid pHT100 [11] carrying promoter Pgrac100 fused to the reporter gene bgaB was used as backbone.

Measurement of the BgaB and GFP production levels in E. coli and in B. subtilis
Three colonies were cultured in 0.5 ml LB medium containing the appropriate antibiotic in a 96 well-block (Eppendorf block) and shaken overnight at 200 rpm at room temperature (25°C). The pre-culture of each clone (75 μl) was transferred to 3 ml LB medium containing the appropriate antibiotic in a 24-well-block. The block was incubated at 37°C with shaking at 200 rpm. When the OD 600 of the culture reached 0.6 -1, the cells were induced by addition of IPTG at final concentration of 0 mM, 0.001 mM, 0.01 mM and 0.1 mM. The cells were harvested after 2 or 4 h of induction. The cells were collected in Eppendorf tubes at an OD 600 of 2.5 after centrifugation. Samples were prepared for activity measurements or SDS-PAGE. The cells were lysed by lysozyme and sample buffer was added to 150 μl, and 8 μl each were applied to SDS-PAGE. ß-galactosidase activities were measured as described [23]. For E. coli, the GFP cells were re-suspended in 300 μl BPS, 12 μl chloroform, and 6 μl SDS 0.1 % were added followed by shaking for 1 h. For B. subtilis, the GFP cells were lysed in 300 μl PBS containing 1 mg/ml were grown in LB medium to mid-log phase, and production of the recombinant proteins was induced by the addition of 0.1 mM IPTG. Cells were lysed, and aliquots were analyzed by SDS-PAGE (lane T, total cellular protein). The cellular extracts were applied to appropriate affinity columns, washed extensively and the bound protein was eluted as described under Materials and Methods. E1, E2 and E3 indicate the first, the second and the third elution step, respectively lysozyme and incubated at 37°C for 2 h. The samples were centrifuged at 10 000 rpm for 5 min and used for determination of the activities. GFP activities were measured by using a Synergy HT Multi-mode Microplate Reader and 384 W plate (Black) with an excitation wavelength at 485 (+/−20) nm and an emission wavelength at 520 (+/−20) nm. The experiment was carried with at least three different colonies, and standard errors were calculated.
Affinity purification of the fusion proteins B. subtilis 1012 carrying different plasmids were grown in LB medium to mid-log phase, and production of the recombinant proteins was induced by addition of 0.1 mM IPTG. The cells were collected by centrifugation and re-suspended in the desired buffers with lysozyme (0.25 mg/ml) and disrupted by sonification. For His-tag fusion proteins, the protocol with recommended buffers for Ni-NTA Spin Columns (Qiagen) was applied, in which the washing buffer contain 40 mM imidazole. For Strep-tag fusion proteins, the Strep-Tactin Spin column kit (IBA) was used.