Engineering strong and stress-responsive promoters in Bacillus subtilis by interlocking sigma factor binding motifs.

Prokaryotic gene expression is largely regulated on transcriptional levels with the involvement of promoters, RNA polymerase and sigma factors. Developing new promoters to customize gene transcriptional regulation becomes increasingly demanded in synthetic biology and biotechnology. In this study, we designed synthetic promoters in the Gram-positive model bacterium Bacillus subtilis by interlocking the binding motifs of σA for house-keeping gene expression and that of two alternative sigma factors σH and σB which are involved in responding post-exponential growth and general stress, respectively. The developed promoters are recognized by multiple sigma factors and hence generate strong transcriptional strength when host cells grow under normal or stressed conditions. With green fluorescent protein as the reporter, a set of strong promoters were identified, in which the transcription activities of PHA-1, PHAB-4, PHAB-7 were 18.6, 4.1, 3.3 fold of that of the commonly used promoter P43, respectively. Moreover, some of the promoters such as PHA-1, PHAB-4, PHAB-7, PBA-2 displayed increased transcriptional activities in response to high salinity or low pH. The promoters developed in this study should enrich the biotechnological toolboxes of B. subtilis.

Introduction

Bacterial RNA polymerase indispensably requires sigma factor to initiate gene transcription [[1], [2], [3]]. The binding motifs of sigma factors, commonly known as the canonical -35 and -10 elements are considered as the core structure of bacterial promoters. Every bacterial species has a house-keeping or principal sigma factor responsible for the transcription of essential genes and several types of alternative sigma factors governing the expression of genes required at special conditions [3]. For instance, the principal sigma factor σ70 of Escherichia coli regulates genes that are indispensable for cell survival such as TCA cycle and protein synthesis. The alternative sigma factor σ32 (RpoH) governs expression of genes to deal with heat shock and the alternative sigma factor σ38 (RpoS) controls the expression of genes for starvation or stationary phase. Bacillus subtilis also has a principal sigma factor σA (SigA) and 10 characterized alternative sigma factors. Among them, the alternative sigma factor σH (SigH) controls gene expression at post-exponential phase and alternative sigma factor σB (SigB) controls genes for general stress response [4]. Principal sigma factor is constitutively produced while alternative sigma factors are generated conditionally in response to intra- or extra-cellular stimuli [3]. All the sigma factors compete for a limited number of RNA polymerase [3,5].

There are a considerable amount of natural promoters governed by more than one sigma factor. For example, in E. coli there are over eight hundred identified regions bound by both σ70 and σ38 [6]. The commonly used B. subtilis P43 promoter is recognized by both σB and σA [7]. Promoters governed by multiple sigma factors should have stronger transcriptional activities, as the promoter recognition chances would be raised. Developing synthetic promoters composed of both principal and alternative sigma factor binding motifs has very high importance in the field of biotechnology. Such synthetic promoters are more resistant to imperfect growing conditions, such as biotechnological processes that cause cell stress or metabolic burden and generate strong transcription of target genes [8].

As another biotechnologically versatile host strain, B. subtilis has excellent protein secretion capability and a high level of biosafety. B. subtilis has been engineered for production of proteins such as α-amylase and chemicals like vitamins and nucleotides [9]. Robust transcription of target genes in various cultivation conditions is crucial for the applications of B. subtilis in biotechnology. To enrich the transcriptional regulation toolboxes of B. subtilis, we constructed strong and stress-responsive B. subtilis promoters in this research. The engineered promoters PHA-1, PHAB-4, PHAB-7 showed higher activities comparing with P43 [7,10] and Pgrac [11]. More importantly, these promoters were in apparent responses to the commonly encountered growth stresses like high salinity or low pH.

Methods

Medium and cultivation

E. coli and B. subtilis strains were cultivated in Luria-Bertani (LB medium, 10 g/L Typtone, 10 g/L Sodium chloride and 5 g/L Yeast extract, pH 7.0) at 37 °C. Spizizen minimal medium was used to prepare B. subtilis competent cells as described [12]. When necessary 100 μg/mL ampicillin, 10 μg/mL chloramphenicol, 20 mM sodium citrate buffer (pH 4.5) or 0.5 M NaCl were supplemented to the culture (all as the final concentration).

Strains and plasmids

The used strains and plasmids in this studied were listed in Table 1 and the primers were shown in Table 2. E. coli JM109 was used for all plasmid constructions. B. subtilis 168 was used as the host strain to measure the activity of the synthetic promoters. The E. coli - B. subtilis shuttle vector pHT01 [13] was selected as the backbone for promoter library construction and screening. To build the plasmid pHT01-gfp, gfp gene was amplified with primers gfp-F and gfp-R from pMD19-gfp [14], digested with BamHⅠ and SmaⅠ and ligated with the equivalently cleaved pHT01. P43 promoter was amplified from pP43NMK [15] with primers P43–F and P43-R. The pHT01-gfp was linearized via PCR with primers p0-F and p0-R to remove the Pgrac promoter. The linearized pHT01-gfp and the P43 PCR product were combined into pHT01-P43-gfp by T5 exonuclease DNA assembly (TEDA) [16].

Table 1

Stains and plasmids used in this study.

Strains	Feature	Reference
E. coli JM109	E. coli K-12 F' traD36 proA+B+lacIq Δ(lacZ)M15/Δ(lac-proAB) glnV44 e14-gyrA96 recA1 relA1 endA1 thi hsdR17	New England Biolabs
B. subtilis 168	B. subtilis wild type strain	Bacillus Genetic Stock Center (BGSC)
Plasmids
pHT01	E. coli - B. subtilis shuttle vector replicative in B. subtilis, carrying IPTG inducible promoter Pgrac, ampR, catR.	[13]
pP43NMK	E. coli - B. subtilis shuttle vector replicative in B. subtilis, with P43 promoter, ampR, KmR.	[15]
pMD19-gfp	pMD19 T-vector carrying the gfp gene.	[14]
pHT01-gfp	The gfp gene under the control of Pgrac promoter in pHT01.	This study
pHT01-P43-gfp	The gfp gene under the control of constitutive promoter P43 in pHT01.	This study
pHT01-PHA-gfp (Promoter library)	The gfp gene under the control of PHA promoter (library) in pHT01.	This study
pHT01-PBH-gfp (Promoter library)	The gfp gene under the control of PBH promoter (library) in pHT01.	This study
pHT01-PBA-gfp (Promoter library)	The gfp gene under the control of PBA promoter (library) in pHT01.	This study
pHT01-PHAB-gfp (Promoter library)	The gfp gene under the control of PHAB promoter (library) in pHT01.	This study
pHT01-PHBA-gfp (Promoter library)	The gfp gene under the control of PHBA promoter (library) in pHT01.	This study

Table 2

Primers used in this study.

Primer	Sequence (5′–3′)
gfp-F	CGCATGGGTAAGGGAGAAGAACTTTTC
gfp-R	TCCCGATCCACCCGGGTTATTTGTATAGTTCATCCATGCC
PHA-F	NTATAATNNNNNNAAAGGAGGAAGGATCAATGGGTAAGGGAGAAGAACTTTTC
PHA-R	CANATTCGNNNNNNNNTGTCAAAATTCCTGCGCAACGCAATTAATGTGAGTTAAG
PBH–F	NNNCGAATNNNNNNAAAGGAGGAAGGATCAATGGGTAAGGGAGAAGAACTTTTC
PBH-R	NATACCCNNNNAATTCCTNNNNNNAAACCTGCGCAACGCAATTAATGTGAGTTAAG
PBA-F	TTATAATNNNNNNAAAGGAGGAAGGATCAATGGGTAAGGGAGAAGAACTTTTC
PBA-R	TACCCNNNNNNNNNNNTGTCAAAAACCTGCGCAACGCAATTAATGTGAGTTAAG
PHAB-F	AATNNGGGTATNNNNNNAAAGGAGGAAGGATCAATGGGTAAGGGAGAAGAACTTTTC
PHAB-R	ATANNNATTCGNNAAACCTNTGTCAAATTCCTGCGCAACGCAATTAATGTGAGTTAAG
PHBA-F	ATNNNNNNNNTATAATNNNNNNAAAGGAGGAAGGATCAATGGGTAAGGGAGAAGAAC
PHBA-R	ACCCNNNTGTCAATTCGNNNNAAACCTNNNNAATTCCTGCGCAACGCAATTAATGTGAG
P43–F	TGATAGGTGGTATGTTTTCGCTTG
P43-R	TGATCCTTCCTCCTTTGGTACCGCTATCACTTTATATTTTAC
p0-F	AGCGGTACCAAAGGAGGAAGGATCAATGGGTAAGGGAGAAGAACTTTTC
p0-R	CAAGCGAAAACATACCACCTATCAGCGCAACGCAATTAATGTGA

Promoter library construction

Promoter libraries borne by pHT01 vector were constructed by replacing the Pgrac promoter with the DNA comprising interlocking sigma factor binding motifs. Briefly, sigma factor binding motifs and randomized intra- and inter-σ binding motif spacers were firstly included in designated primers (Px-F and Px-R Table 2) synthesized by GENEWIZ (Suzhou, China). PCR was performed with the described primers using pHT01-gfp as template. PCR products were phosphorylated and end to end ligated by Blunting Kination Ligation (BKL) Kit (Takara, Beijing) to generate the final plasmids containing the promoter libraries (Table 1).

Promoter library screening and promoter characterization

For primary screening of the synthetic promoter libraries, B. subtilis cells carrying the plasmids were cultivated in 96-well plates filled with LB medium for 24 h. Cell density (OD600) and fluorescence intensity (excitation wavelength 490 nm, emission wavelength 530 nm, gain 70) were measured with BioTek Cytation Plate Reader. The culture of B. subtilis strain carrying the pHT01 vector was applied to subtract background fluorescence signal. During the second round of screening, 26 strong promoters that did not impair cell growth were re-assayed in shake flask cultures. Sampling was performed every 4 h, and cells were washed with 20 mM phosphate buffered saline (pH 7.0, PBS). After appropriate dilution, cell density and fluorescence intensity were measured as described above.

To test the response of synthetic promoters to stresses including low pH and high salinity, cells were pre-cultivated at normal condition (LB medium, 37 °C, pH 7.0) for 6 h. Afterwards, the cultures were shifted to the designated conditions by supplementing 20 mM final concentration of sodium citrate buffer (pH 4.5) or 0.5 M final concentration of NaCl. Samples were taken every 2 h. Cells were washed with PBS, and the cell density (OD600) and fluorescence intensity were measured as described above.

Promoter sequencing

Promoters with interesting properties were isolated from B. subtilis via plasmid extraction (Sangon plasmid extraction kit). Crude plasmid samples were used to transform E. coli JM109. Single E. coli colonies were picked for plasmid Sanger sequencing by GENEWIZ (Suzhou, China).

Results and discussion

Design of synthetic promoters

The consensus sequences recognized by B. subtilis σA (house-keeping sigma factor), σB (sigma factor for general stress response), and σH (sigma factor for post-exponential and sporulation gene expression) (Fig. 1A) were selectively assembled to interlock each other (Fig. 1B and 1C). The binding motifs of σA and σB have been defined as TTGACA (-35)-N14-TGNTATAAT (-10) and AGGTTT (-35)-N17-GGGTAT (-10) in the DBTBS (a database of transcriptional regulation in B. subtilis) [17]. The binding motifs of σH was defined as AGGAATT (-35)N14-CGAAT (-10) here according to previous report [4]. RNA polymerase would accordingly be guided to the synthetic promoters with the assistance of more than one sigma factor and the frequency of promoter recognition would thus be increased, especially when alternative sigma factors were enriched in response to stresses. The length of the intra σ binding motif spacer (the spacer between the -35 and -10 elements) sequence was not changed, while the nucleotides of the spacer were randomized via PCR with degenerate primers (Fig. 1B and 1C). Under this provision, there are numerous ways to combine the σ binding motifs into synthetic promoters as the length of inter σ binding motif spacer (the spacer between two -35 or two -10 elements), the arrangement of the motifs and the overall number of the assembled motifs are all variants (Fig. 1B).

Fig. 1

Schematic of synthetic promoters composed of interlocking sigma factor binding motifs. (A) In many cases, promoters contain one sigma factor binding motif, including the -35 and -10 elements. The commonly used B. subtilis promoter P43 comprises overlapping binding motifs of σB and σA. Diamond stands for the -35 element while rectangle indicates the -10 element. (B) Inspired by P43, there are numerous possible ways to construct synthetic promoters by interlocking the binding motifs of different sigma factors when considering variations in inter σ binding motif spacer length, motif arrangement or motif compositions. Herein, the length of the intra σ binding motif spacer was fixed as the same to the natural promoters, but the spacer of inter σ binding motif was variable. (C) Nucleotide sequences of natural promoters and synthetic promoters recognized by single or multiple sigma factors. The consensus sequences recognized by σA, σB and σH were indicated with green, violet and yellow; -35 elements and -10 elements of the sigma factor binding motifs were indicated with diamonds and rectangles, respectively. Red letters indicate the nucleotide shared by two adjacent sigma factor binding motifs. N stands for degenerate nucleotide.

In this study, we designed B. subtilis synthetic promoters by selectively combining the motifs recognized by σA, σB and, σH arbitrarily in five ways (Fig. 1C). The aim is to create synthetic promoters with strong transcriptional activities in all grow phases under different cultivation conditions. PHA comprises interlocking binding motifs of σA and σH; PBH comprises interlocking binding motifs of σB and σH; PBA, similar to P43, comprises interlocking binding motifs of σB and σA; PHAB and PHBA comprise interlocking binding motifs of σH, σA and σB. The binding motifs of σB and σA in PHBA promoters are inverted comparing to PHAB. Moreover, in PHAB promoters the three selected -35 elements and -10 elements are adjacent to each other (Fig. 1C).

Preliminary screening of the synthetic promoters

The promoter libraries for PHA, PBH, PBA, PHAB, and PHBA were constructed by replacing the Pgrac promoter of pHT01-gfp with the designated interlocking sigma factor binding motifs. For each library, 384 promoters were analyzed. Relative fluorescence intensity (Fluorescence intensity (au)/OD600) was used to demonstrate the promoter activities. Most PBH, PBA, PHAB and PHBA promoters displayed significantly weaker promoter activities comparing to the commonly used B. subtilis inducible promoter Pgrac, as the upper quartiles of these promoter activities (Fig. 2, upper black dash lines) were much lower than the activity of the Pgrac (Fig. 2, pink dash line). On average PHA promoter library displayed the highest activity with its median value (Fig. 2, black solid line) slightly lower than the activity of Pgrac. Principally PBH promoters should only display its maximal promoter activities at post-exponential phase with growth stresses, when σH and σB were enriched.

Fig. 2

Distribution of promoter activities measured in the primary screening of the libraries. B. subtilis cells carrying the promoter libraries were growth for 24 h in 96-well plates. Fluorescence intensity (au)/OD600 was used to demonstrate the promoter strength. For the violin plot, the median values were marked with solid black lines. The upper and lower quartiles of these promoter activities were indicated with upper and lower black dash lines, respectively. Pink dash line indicated the Pgrac promoter activity. The promoters selected for further characterization were labeled out.

The data here preliminarily suggests that it is possible to raise B. subtilis promoter activities by interlocking sigma factor binding motifs. However, we found more sigma factor binding motifs does not necessarily make the synthetic promoters stronger, as it was shown that PHA library outperformed the three-sigma-factor controlled PHAB and PHBA. Elongated promoter structure may also structurally tangle promoter DNA and decrease transcriptional activity. Moreover, downstream sigma factor binding motifs might also cause the pausing of a transcriptional process initiated from upstream sigma factor binding motifs [18].

Synthetic promotes belonging to the same libraries exhibited vastly different transcriptional activities, for instance, PHA promoters displayed the broadest activity variation (Fig. 2). The estimated activity of the strongest PHA promoter displayed more than 130 fold higher than the weakest PHA promoter (Fig. 2, green numbers). This suggests the spacer sequences are not less important in determining promoter activities. Sigma factor binding motifs is indispensable for the promoter recognition, while the spacer sequences may play critical roles in subsequent structural modulation of RNA polymerase-DNA complex and promoter open complex formation [19]. Previous studies also showed the significance of the spacer sequences between -35 and -10 elements [19,20].

We found a few PBH and PHAB promoters displayed high estimated promoter activities (Fig. 2, top point of the violin plot) but impaired significantly B. subtilis cell growth. The OD600 measured in 96-well plate was about 0.05 after 24 h of cultivation. These promoters were no longer considered. Twenty six strong promoters that did not impair cell growth were selected for further characterization in shake flask (Fig. 2).

Secondary screening of the synthetic promoters

To demonstrate the promoter activities precisely, cells cultivated in shake flasks were washed twice with 20 mM phosphate buffered saline (PBS) (pH 7.0) to remove GFP released from cell lysis. OD600 and fluorescence intensities were measured at 6 h, 12 h, 18 h and 24 h. Relative fluorescence intensity was used again to demonstrate the promoter activities. The inducible promoter Pgrac (activated with 0.1 mM isopropylthio-β-galactoside (IPTG)) and constitutive B. subtilis promoter P43 were used as the control promoters. We found the measured promoter activities herein (Fig. 3) were lower than that measured via 96-well plate cultivation (Fig. 2). This should be partially ascribed to cell washing, which removed GFP released into the medium (Fig. 3).

Fig. 3

Transcriptional strength of the selected strong promoters measured in the second round of screening. B. subtilis cells carrying the promoters were grown in LB medium in shake flasks. Cells were washed twice with PBS and fluorescence intensity and OD600 were measured at 6 h, 12 h, 18 h and 24 h. Pgrac and P43 promoters were used as the control. Transcription driven by Pgrac was induced with 0.1 mM IPTG.

Most promoters displayed stronger promoter activities in the secondary screening than Pgrac and P43, except PHBA-1, PHAB-2. We speculate that PHBA-1 and PHAB-2 might be more sensitive to variation in cultivation conditions. The strongest promoter PHA-1 exhibited significantly higher strength than Pgrac and P43 controls (Fig. 3). The relative fluorescence intensity of PHA-1 at 12 h was 17.6 fold of P43. Moreover, we found the activities of these synthetic promoters formed a gradient (Fig. 3). These promoters were classified into “High activity”, “Medium activity” and “Low activity” groups. These promoters with quite different transcriptional activities should enrich the gene regulation toolbox of B. subtilis.

The sequences of these 26 synthetic promoters were also compared (Table 3). However, we did not find apparent regularity defined the correlation between promoter strength and the randomized inter- and intra-σ binding motif spacers (Fig. 3 and Table 3). Additionally, we found the sequences of PHAB-5, PHAB-6 and PHBA-4 promoters were mutated. The -35 element of σB were changed from “GGGTAT” to “GGGTAA” in PHAB-5. The -35 elements of σH and σA and the -10 element of σB were missing in PHAB-6, while the -10 element of σB and -35 element of σA were deleted in PHBA-4. There is no clue how these truncated promoters, particularly PHAB-6, executed gene transcription initiation, but our finding herein suggests the plasticity of sigma factor reorganizing B. subtilis promoters.

Table 3

Sequences of the 26 selected promoters.

Promoters		Sequence (5′ to 3′)
PHA	PHA-1	AGGAATTTTGACACCCTCACTCGAATGTGCTATAATGGCCACAAAGGAGGAAGGATCA
	PHA-2	AGGAATTTTGACAAGCCGTTTCGAATATGATATAATAAGCGGAAAGGAGGAAGGATCA
	PHA-4	AGGAATTTTGACACCCTAGATCGAATTTGCTATAATGGGGGAAAAGGAGGAAGGATCA
	PHA-11	AGGAATTTTGACAATGCGCCACGAATCTGCTATAATGTGCACAAAGGAGGAAGGATCA
	PHA-12	AGGAATTTTGACAGCGGTACGCGAATGTGATATAATAGTCTAAAAGGAGGAAGGATCA
	PHA-13	AGGAATTTTGACACCCGGCTACGAATATGCTATAATAACTTGAAAGGAGGAAGGATCA
	PHA-14	AGGAATTTTGACACCATATAGCGAATGTGATATAATCGAAGTAAAGGAGGAAGGATCA
PBA	PBA-1	AGGTTTTTGACACGCCGCTCGGTGGGTATTATAATAGACGTAAAGGAGGAAGGATCA
	PBA-2	AGGTTTTTGACAAATATAGACTGGGGTATTATAATGAAACTAAAGGAGGAAGGATCA
	PBA-3	AGGTTTTTGACAGCGATTCCCATGGGTATTATAATAGGGTAAAAGGAGGAAGGATCA
	PBA-4	AGGTTTTTGACAAAGTCCCGGACGGGTATTATAATAGTCCCAAAGGAGGAAGGATCA
	PBA-9	AGGTTTTTGACACATCGCTTCCAGGGTATTATAATCAGAAGAAAGGAGGAAGGATCA
PHAB	PHAB-1	AGGAATTTGACATAGGTTTTTCGAATTCCTATAATGGGGGTATAACCCGAAAGGAGGAAGGATCA
	PHAB-2	AGGAATTTGACACAGGTTTCCCGAATGGGTATAATACGGGTATGACCCGAAAGGAGGAAGGATCA
	PHAB-3	AGGAATTTGACATAGGTTTTCCGAATTCCTATAATAAGGGTATCAAATAAAAGGAGGAAGGATCA
	PHAB-4	AGGAATTTGACACAGGTTTTACGAATCTGTATAATAAGGGTATCAGGGAAAAGGAGGAAGGATCA
	PHAB-5	AGGAATTTGACAAAGGTTTCCCGAATTCGTATAATCGGGGTAAAAGGAGGAAGGATCA
	PHAB-6	AGGAATTTGACATAATGAGGGTATGTCGGGAAAGGAGGAAGGATCA
	PHAB-7	AGGAATTTGACAGAGGTTTCTCGAATACCTATAATTGGGGTATGGTAGGAAAGGAGGAAGGATCA
	PHAB-8	AGGAATTTGACACAGGTTTGCCGAATCACTATAATGGGGGTATGATACCAAAGGAGGAAGGATCA
PHBA	PHBA -1	AGGAATTATAGAGGTTTTTCGCGAATTGACACCTGGGTATCTGGTTGCTATAATTCCTGCAAAGGAGGAAGGATCA
	PHBA -2	AGGAATTAAGGAGGTTTCTTGCGAATTGACAGCCGGGTATAGGTGCGTTATAATTAGTTAAAAGGAGGAAGGATCA
	PHBA -3	AGGAATTTTACAGGTTTCTTCCGAATTGACATACGGGTATGTTCGTGTTATAATTTTGTCAAAGGAGGAAGGATCA
	PHBA -4	AGGAATTGCCTAGGTTTCAAGCGAATTGACAGGCGGGGGTGAAAGGAGGAAGGATCA
	PHBA -5	AGGAATTGACTAGGTTTTCTGCGAATTGACAGCCGGGTATAAAGCTGATATAATTGTAAGAAAGGAGGAAGGATCA
	PHBA -9	AGGAATTGTCAAGGTTTCCCGCGAATTGACACCCGGGTATCGTAGCGATATAATGTAGGGAAAGGAGGAAGGATCA

Ribosome binding site is labeled in italic. Straight, dash and wave underlines stand for the binding motif of σH, σB and σA, respectively.

Characterization of synthetic promoters to different stresses

In the next step, eight synthetic promoters, PHA-1, PHA-2, PHAB-4 (high activity group, Fig. 3); PHAB-6, PHAB-7 and PHBA-5 (medium activity group, Fig. 3); PBA-2 and PBA-3 (low activity group, Fig. 3), were selected as the representatives to test their responses to low pH and high salinity pressures, which were commonly encountered during biotechnological bacteria cultivation. Cells carrying the synthetic promoters or the control promoter Pgrac were cultivated under the normal condition (LB medium, pH 7.0) for 4 h before being exposed to pH 4.5 or 0.5 M NaCl (final concentration). The time-course of the promoter activity was recorded and compared (Fig. 4). We found transcriptions from synthetic promoters PHA-1, PBA-2, PHAB-6, PHAB-7 and PHBA-5 were enhanced at varying degrees in response to 0.5 M NaCl while PHA-2 was repressed by 0.5 M NaCl (Fig. 4). In comparison, PHAB-4 was in positive response to acidic pH. These results indicate that it is feasible to make a promoter stress-responsive by incorporating the binding motifs of alternative sigma factors.

Fig. 4

Response of the synthetic promoters to stresses. Cells were pre-cultivated in LB medium at 37 °C for 6 h. To activate the transcription from Pgrac, 0.1 mM IPTG was added at the beginning of cultivation. The cultivation conditions were shifted to pH 4.5 or supplemented with 0.5 M NaCl at time point 6 h (indicated with arrows). Cells were collected at indicated time points and wash twice with PBS before the measurement of cell density (OD600) and fluorescence intensity.

B. subtilis σB is known for responding to general stresses. Promoters carrying the binding motif of σB, such as PHAB and PHBA were induced as anticipated by the supplement of NaCl or by low pH. B. subtilis σH assistants the transcription of genes specific to post-exponential growth phase or sporulation. The incorporation of σH binding motif to synthetic promoters should be beneficial for strong gene transcription at post-exponential phase. This is in agreement with our finding that in comparing to Pgrac the PHA promoter and PHAB promoters displayed stronger activities after time point 8 h (Fig. 4). High salinity reduces the association of σH [21] and therefore may suppress the σH control gene expression. It explains the reduced activity of PHA-2 in presence of 0.5 M NaCl (Fig. 4). However, we found the PHA-1 promoter lacking σB binding motifs was induced by salt stress. The differences between PHA-2 and PHA-1 promoters were only found in the spacer region. More study is required to explain how the spacer regions could invert the response of a promoter to stresses.

Conclusions

Promoter is the most important regulatory element of gene expression. Developing promoters with novel properties has captured many attentions in the field of synthetic biology or biotechnology [[22], [23], [24], [25]]. Stress-responding capability is one the properties that have been studied extensively. Various synthetic stress responsive promoters were created dominantly by engineering the transcriptional regulators and the cognate cis-DNA elements [[26], [27], [28], [29], [30]], which were considered as the accessory elements of promoters. In this study, we developed stress responsive promoter for application in B. subtilis by modulating the core elements of all bacterial promoters, the binding motif of sigma factor (Fig. 1). Incorporating the binding motifs of different sigma factors would assimilating the endogenous regulation of sigma factors to synthetic promoters and confer promoters stress-responsive capability and strong activity, especially when cells encountering growth transition or environmental changes.

The strongest promoter necessarily containing the binding motif of σA, which is the principal sigma factor of B. subtilis. B. subtilis promoters unrecognizable to σA are unlikely to generate strong transcriptional strength, as indicated by the low activity of PBH (Fig. 2). Comparing to PHA, the PBA showed much weaker activities, which may be a preliminary indication of the σH's advantage over the σB in competing for RNA polymerase. Promoter library construction and large/medium-throughput screening seem inevitable when developing customized promoters as synthetic promoters composed of same organization of sigma factor binding motifs may have strikingly different transcriptional strengths (Fig. 2 and Fig. 3) or different responses towards the same stressor (Fig. 4 and Table 3). In many sophisticated promoter engineering strategies, synthetic promoters designed based on promoter modularity require some irrational optimizations as many factors such as DNA topological structure or post-transcriptional regulation may unpredictably affect the promoter activities.

Future work may expand this promoter engineering method to the other industrial microorganisms such as Corynebactrium glutamicum, Clostridium acetobutylicum, Synechococcus sp. and Streptomyces coelicolor.