Promoter Screening from Bacillus subtilis in Various Conditions Hunting for Synthetic Biology and Industrial Applications

Abstract

The use of Bacillus subtilis in synthetic biology and metabolic engineering is highly desirable to take advantage of the unique metabolic pathways present in this organism. To do this, an evaluation of B. subtilis’ intrinsic biological parts is required to determine the best strategies to accurately regulate metabolic circuits and expression of target proteins. The strengths of promoter candidates were evaluated by measuring relative fluorescence units of a green fluorescent protein reporter, integrated into B. subtilis’ chromosome. A total of 84 predicted promoter sequences located upstream of different classes of proteins including heat shock proteins, cell-envelope proteins, and proteins resistant against toxic metals (based on similarity) and other kinds of genes were tested. The expression levels measured ranged from 0.0023 to 4.53-fold of the activity of the well-characterized strong promoter P43. No significant shifts were observed when strains, carrying different promoter candidates, were cultured at high temperature or in media with ethanol, but some strains showed increased activity when cultured under high osmotic pressure. Randomly selected promoter candidates were tested and found to activate transcription of thermostable β-galactosidase (bgaB) at a similar level, implying the ability of these sequences to function as promoter elements in multiple genetic contexts. In addition, selected promoters elevated the final production of both cytoplasmic bgaB and secreted protein α-amylase to about fourfold and twofold, respectively. The generated data allows a deeper understanding of B. subtilis’ metabolism and will facilitate future work to develop this organism for synthetic biology.

About the Authors:

Yafeng Song

Contributed equally to this work with: Yafeng Song, Jonas M. Nikoloff

Affiliations Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China, Key Laboratory of Systems Microbial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China

Jonas M. Nikoloff

Contributed equally to this work with: Yafeng Song, Jonas M. Nikoloff

Affiliation: Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China

Gang Fu

Affiliations Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China, Key Laboratory of Systems Microbial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China

Jingqi Chen

Affiliations Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China, Key Laboratory of Systems Microbial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China

Qinggang Li

Affiliations Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China, Key Laboratory of Systems Microbial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China

Nengzhong Xie

Affiliations National Engineering Research Center for Non-food Biorefinery, State Key Laboratory of Non-food Biomass Energy and Enzyme Technology, Nanning 5300074, P. R. China, Guangxi Biomass Industrialization Engineering Institute, Guangxi Academy of Sciences, Nanning 530007, P. R. China

Ping Zheng

Affiliations Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China, Key Laboratory of Systems Microbial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China

Jibin Sun

Affiliations Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China, Key Laboratory of Systems Microbial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China

Dawei Zhang

* E-mail: [email protected]

Affiliations Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China, Key Laboratory of Systems Microbial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, P. R. China

Introduction

Metabolic engineering seeks to increase the synthesis of desired products de novo and by modification of existing metabolic pathways or by optimization of appropriate genetic elements [1–4]. The use of different biological elements may help to tune expression to achieve the desired production level, provided that cell metabolism remains co-ordinated [5–8]. Some organizations, e.g. the Genetically Engineered Machine (iGEM) Foundation, have tested the efficiency of sequence elements to function as interchangeable components or biobricks that can be used to build biological systems. The goal is to assemble libraries of these elements that can be applied for the engineering of living cells.

Current efforts in synthetic biology have focused on the evaluation of cis-sequences including promoters, ribosome binding sites (RBS), and terminators in both Escherichia coli and Saccharomyces cerevisiae [9–12]. There are several successful examples of changing promoters to alter expression in prokaryotic and eukaryotic cells. Hal Alper et al constructed a promoter library based on the bacteriophage P_L-λ promoter that was generated using error-prone PCR. The library was tested for promoter strength by measuring the downstream expression of gfp and the chloramphenicol acetyltransferase gene, and those that exhibited a linear relationship between promoter strength and reporter were selected. They then expressed a series of promoter-dxs (deoxy-xylulose-P synthase) constructs in a recombinant E. coli strain overexpressing genes (ispFD and idi) of the isoprenoid synthesis pathway, and observed the linear response of lycopene yield to promoter strength. [13]. In a second example, a strong and tunable promoter library was obtained that showed a range of 400-fold at the mRNA level [14]. This library was created by combining various copies of upstream activation sequences with the native promoter AOX1 from Yarrowia lipolytica. The final expression of humanized Renilla GFP (hrGFP) in Y. lipolytica was increased eightfold over the activity of the original endogenous promoter [14]. In this system, both a high output of heterologous protein and a strictly controlled metabolic pathway were engineered [14]. In another report, a set of insulated promoters, differing in strength and context-independent behavior, was designed and applied for controlled protein production. The properties of those promoter devices in one test context were predictive of those properties in a new context allowing steady-state protein production regulated by transcriptional regulation [15]. 195 native or synthetic promoters and 192 RBSs were characterized for their ability to drive the expression of superfold GFP (sfGFP) in Streptomyces venezuelae [16]. Next, an insulator sequence, RiboJ, was introduced to reduce interference between promoters and RBSs and these combinations were again tested by examination of sfGFP levels. The insulator element RiboJ is a DNA sequence that contains both a ribozyme and hairpin acts to help expose the RBS [16]. These synthetic modular regulatory elements were then inserted upstream of the lycopene biosynthetic cluster in S. avermitilis. The correlation between lycopene production and promoter strength, elevated lycopene titer as well, confirmed the utilization and feasibility of these expression cassettes and paved the way for further application of these promoter elements in system biotechnology [17].

B. subtilis shows significant advantages as a host for protein expression due to its efficient secretion capacity and because it is a generally recognized safe organism. It is used for applications in the detergent, textile, and pharmaceutical industries [18–20]. However, due to unsuitable vectors containing weak promoters, RBS sequences, terminators, incompatible antibiotic genes, or insufficient plasmid copy numbers [21], many heterologous proteins are produced with only low yields using these vectors. In addition, overexpressing genes of secretory components is time-consuming and inefficient, particularly due to the complexity of the secretion mechanisms [22,23]. There are recent reports from large-scale omic-studies of this Gram-positive bacterium [24–26] that discussed the variation of proteins or transcripts after different conditions of cultivation. However, they compared only a few examples for specific promoter strength and did not perform a comprehensive analysis. Some promoters were discovered or modified by optimization of their key elements [27] for application in target expression systems, but no systematic promoter candidate strength evaluation has been performed using experimental assessment of reporter gene transcription. Some accessible omic-data about promoter candidates was from B. thuringiensis rather from B. subtilis. In that study, Wang et al systematically identified 1203 active promoter candidates through analysis of genome-wide transcription start sites based on RNA-seq data. Additionally, they further evaluated the characteristics of 20 highly active promoters combined with the corresponding 5’-UTR to screen the highly active promoter-5’-UTR DNA region complex by directing the expression of reporter gene lacZ [28]. Therefore, an efficient and facile approach for achieving desired production goals is selection of a suitable promoter element to strictly control related protein expression levels and to allow precise and functional modularity [29].

Here, we constructed a promoter-probe vector with GFP as a detection target and measured expression levels. We wanted to identify stress-activated promoters that could direct high-level expression under specific stress conditions, like heat shock, high salt, or ethanol to stimulate membrane stress. More than 80 promoter candidates that showed high transcription levels under multiple conditions were selected and classified according to the encoded proteins located downstream from the promoters [30]. Promoter candidate activities, relative to that of the constitutive strong promoter of P43, [31], ranged from 0.0023 to 4.53, which spanned a ~1960-fold range. After heat shock, salt, or ethanol treatment, they showed changed RFU levels when compared with normal cultivation, however, the changes were less than expected. The strong promoter PtrnQ exhibited higher activity than P43 in driving the transcription of both cytoplasmic and secretory proteins (about four- and two-fold, respectively), enabling normalized measurement of the promoter candidates’ strengths.

Materials and Methods

Strains and plasmids

E. coli DH5α was used as cloning host. The plasmid-free strain B. subtilis 1A751 was used as the main host for the promoter library. Tested promoter candidates were amplified from B. subtilis 168 genomic DNA. The function and relevant properties of strains and plasmids are listed in Table 1.

[Figure omitted. See PDF.]

Table 1. Strains and plasmids used in this study.

https://doi.org/10.1371/journal.pone.0158447.t001

Construction of the promoter-probe vector and vectors carrying different promoter candidates and pMA5 containing PtrnQ

To simplify construction and utilize a strategy that did not rely on restriction enzymes, we used prolonged-overlap-PCR to replace the bgaB gene of pDL with gfpmut1 [32]. Primers used in this procedure are listed in Table 2.

[Figure omitted. See PDF.]

Table 2. Oligonucleotides used in this study.

https://doi.org/10.1371/journal.pone.0158447.t002

Promoter candidate fragments were prepared by PCR amplification using B. subtilis 168 genomic DNA as template, PrimerSTAR Max DNA Polymerase mix containing dNTP and DNA polymerase (TaKaRa, Japan), and primers (Genwiz, China) listed in Table 2, according to the instructions of the manufacturer. 2000 bp upstream of the start codon of the target protein were amplified as the promoter candidates. The enzyme digestion sites in the vector 5’ and 3’ ends were BamHI and KpnI (BamHI/EcoRI, EcoRI/KpnI). After purification of the digested promoter candidates (E.Z.N.A.^™ Cycle Pure Kit 200, Omega Bio-tek, Inc., USA) and the vector enzyme digestion products, the ligation was performed using T4 ligase (Thermo Fisher Scientific, USA) according to the instructions of the manufacturer and transformed into E. coli DH5α. Putative candidate colonies carrying cloned promoter fragments were identified by colony PCR before being validated by sequencing (Genwiz, China).

Culture media and expression conditions

Cells were incubated aerobically at 37°C and 220 rpm, in LB media. Antibiotics were added where appropriate (ampicillin at 100 μg/mL, kanamycin at 40 μg/mL, chloramphenicol at 5 μg/mL). Cells for the GFP fluorescence intensity assays were incubated in LB in 96-well micro-plates.

Whole cell fluorescence measurements

Expression of GFPmut1 was monitored using whole-cell fluorescence under the multimode microplate reader (SpectraMax M5) at an optical density of 0.4–0.6 for λ = 600 nm (OD₆₀₀) for the different transformants. Cells were centrifuged at 4000 rpm for 10 minutes, the supernatant was discarded, and cells were resuspended in an equal volume of double distilled water. The extinction and emission wavelength were set at 488 nm and 523 nm, respectively. B. subtilis 1A751 without chromosomal gfp was defined as the negative control. Standard deviations are based on a minimum of three statistically independent experiments.

Fluorescence-activated cell sorter (FACS) analysis

Fluorescence-activated cell sorter (FACS) can process tens of thousands of individual cells within a few seconds and obtain many parameters at the same time. In order to guarantee the accuracy of the detected results, we measured the RFU of some selected samples using FACS. For the FACS analysis, cells were grown for 6.5 h at 37°C and then were centrifuged at 12000 rpm for 2 minutes. After discarding the supernatant, cells were resuspended in phosphate-buffered-saline (PBS). Samples were analyzed using a FACS Aria Cell Sorting System (Beckman, MoFlo XDP, USA) with a 100 mW, 488 nm coherent sapphire solid state laser for GFP. B. subtilis 1A751 without chromosomal gfp was defined as the negative control.

RNA isolation, reverse transcriptase PCR (RT-PCR), and real-time fluorescence quantitative PCR

Overnight cultures were inoculated at a ratio of 1:100 into 5 mL LB. After growing for 4 h and 45 min at 37°C, cells were collected for RNA isolation. For those cells containing promoter candidates from heat shock proteins, cultures were grown at 43°C instead of 37°C for 15 min before testing. Similarly, cells containing promoter candidates from cell-envelope proteins were tested by incubation in LB media for 4 h and 30 min followed by addition of ethanol to a final concentration of 4% (v/v). Cells containing promoter candidates from proteins that confer resistance to toxic metals were tested by incubation in LB media for 4 h and 30 min followed by addition of CoCl₂ to a final concentration of 8 mM. 5 mL cultures were used for RNA extraction according to the manufacturer’s instructions (TIANGEN, China) and the PrimeScript RT reagent Kit (TaKaRa) was used to extract cDNA. Real-time fluorescence quantitative PCR (qRT-PCR) was performed with FastStart Universal SYBR Green Master (ROX) (Roche, Germany) using a 7500 system (Applied Biosystems, USA) following the manufacturer’s instructions. Based on the 2^-ΔΔCt method, we calculated the gene expression level by using housekeeping gene gap as reference gene. PCR conditions were the following: reverse transcription at 50°C for 2 min, then denaturation at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 s and then annealing and extension at 60°C for 1 min. The primer sequences for gap and gfp amplification are listed in Table 2.

Measurement of BgaB activity and α-amylase activity

The BgaB activity assay was carried out as previously described [33]. Samples were collected after incubating in LB media for 4 h and 45 min with an inoculation ratio of 1:100. The α-amylase activity assay was performed as described previously [13]. 1 mL supernatants of culture samples after centrifugation were collected to perform sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (NuPAGE 10% Bis-Tris Gel, Commassie Brilliant Blue, Invitrogen Life Technologies, USA).

Results

Strength of different promoter candidates from B. subtilis

To measure promoter candidates’ strengths precisely, it is ideal if single copies of the reporter gene can be inserted into the genome to avoid complications due to heterogeneous numbers of plasmids in each bacterium. An inaccurate distribution of plasmid copy numbers may even occur within the same population of cells, thus single copy gene replacements or insertions are preferred over plasmids [34]. pDL is a promoter-probe vector carrying sequences corresponding to the genomic sequences upstream and downstream of amyE, which can be used to facilitate double over-crossing with the chromosome [35]. Loss of amyE has no detectable effect on growth. However, the reporter gene in pDL is bgaB, which requires complex detection procedures. Thus, to simplify screening, we exchanged this gene with gfp [36]. The pDL plasmid carrying the cloned promoter fragment upstream of GFP cassette was linearized with PstI and integrated into the B. subtilis’ genome through homologous recombination. Prolonged overlap extension PCR approaches were used to generate the pDL-GFP vector (Fig 1) which obviated the need to look for compatible restriction enzyme sites and allowed high efficiency transformation [37].

[Figure omitted. See PDF.]

Fig 1. Scheme of the constructed promoter-probe vector pDL-GFP showing multiple cloning sites (MCS) and restriction enzyme recognition sites.

The boxes show the 5’ and 3’-end enzymes used for ligation of promoter candidates (PC). The left side shows the backbone of the integrated plasmid pDL-GFP and the main genes. On the right side at the bottom the sequences and enzyme sites within it are shown.

https://doi.org/10.1371/journal.pone.0158447.g001

PCR fragments (2 kb) corresponding to the sequences upstream of 84 genes were amplified with the oligos shown in S1 Table and, after digestion with the enzymes indicated, inserted into the pDL-GFP vector. These 84 were chosen because they were candidates to be expressed by specialized sigma factors (S2 Table) and/or under stress conditions (S3 Table). The promoter-gfp constructs were subsequently inserted into the B. subtilis chromosome at the amyE locus in single copy. The strengths of the cloned promoters were measured in a multimode microplate reader and the RFU of GFP signal from different promoters ranged from 0.0023 to 4.53 when compared to GFP fluorescence promoted by P43 (Fig 2). When we calculated the relative strengths of these promoter candidates, background expression level of promoterless-gfp was subtracted. Of these promoter candidates, PtrnQ was more than four-fold stronger than the P43 promoter, used as a strong reference promoter. trnQ codes for arginine-specific tRNA. PsigX and PgroES exhibited 3.03 and 1.55 times of P43’s strength in expression, respectively. In addition, PsecDF and PugtP showed slightly lower strengths than P43. There were also many promoters presenting low activities, many of them upstream of hypothetical proteins with unknown functions.

[Figure omitted. See PDF.]

Fig 2. Strength of different promoter candidates measured by microplate reader.

The relative expression level is the activity of each promoter compared to the strong constitute promoter P43. The activities of most promoters were less than 50% the activity of P43. Promoters with activities greater than 50% P43 were shown in the inset histogram, and P43 activity was set as 100%. Data are means ± standard deviation for three independent experiments.

https://doi.org/10.1371/journal.pone.0158447.g002

Sequence analysis of strong promoters from B. subtilis

We selected seven strongest promoter candidates and predicted the -35, -10 elements and the lengths of the spacers (Fig 3A) using Softberry Inc., (http://linux1.softberry.com/berry.phtml?topic=bprom&group=programs&subgroup=gfindb). These promoter candidates exhibited conserved -10 boxes, which were similar to the classical “TATAAT” sequence (Fig 3B), which was determined by applying Weblogo (http://weblogo.berkeley.edu/) [38,39]. However, the -35 sequences from these promoters were not consistent with the classical “TTGACA” (Fig 3C). Only 3 of the 7 promoters showed a canonical 17 bp spacer region. The differences in the lengths of the spacers could explain why we did not obtain obvious conserved sequences at particular positions, e. g. the -35 boxes, when we fixed the -10 boxes (Fig 3D). Additionally, the predicted -10 and -35 elements did not always match the experimentally validated actual -10 and -35 boxes. Use of another software tool for the same prediction gave varied results. An alternative program predicted promoters in the same region without precisely identifying -35 and -10 elements (results not shown). The predicted -10 and -35 boxes are listed in S2 Table as well as their regulated sigma factors. In B. subtilis, RNA polymerase consists of αα’ββ’ and one of several σ initiation factors which determines the specificity of RNA polymerase [40]. σ^A regulates the largest number of operons and is the primary “housekeeping” sigma factor, and the alternative sigma factor σ^B is activated in response to general stress or energy depletion [41]. σ^F, σ^E, σ^G, and σ^K play important roles in sporulation [42]. Among extracytoplasmic function (ECF) sigma factors, σ^M, σ^W, σ^X, and σ^V confer resistant against cell envelope active compounds, and σ^M, σ^W, σ^X, σ^Y, and σ^V control cell envelope stress proteins [43]. In addition, some genes had multiple promoters and overlapping recognition was observed due to the similar promoter recognition properties of some of these ECF sigma factors.

[Figure omitted. See PDF.]

Fig 3. Sequence logo of the seven relatively strong promoter candidates from B. subtilis.

A: Sequences of predicted -35 and -10 boxes and the spacers between the two elements. About half of the spacers between -35 and -10 boxes were around 17 base pairs. These three sequence logos were constructed using Weblogo3. B: Sequence logo of -10 hexamers; similar to cardinal “TATAAT” sequence. C: Sequence logo of -35 hexamer, not consistent with the classical “TTGACA” sequences. The -35 logo was very consistent with the consensus logo (5/6 positions were majoritarily correct). This was better than the -10 logo where only 4/6 positions were mostly the consensus base. D: Sequence logo of promoter candidates when the -10 box was fixed; the different spacing between the promoter elements may have contributed to difficulty in assigning the -35 box. This was presumably because the best-fit -10 sequences were used to make the logo. Since the varying lengths of the spacers have not been taken into account, there was no consensus -35 box.

https://doi.org/10.1371/journal.pone.0158447.g003

Fluorescence-activated cell sorter (FACS) analysis of promoter candidates from B. subtilis

FACS analysis was performed on bacteria growing in the logarithmic phase. Data analysis confirmed the fluorescence of GFP, positioned downstream of 10 different promoter candidates. Selected promoter candidates activated transcription at high, medium, and low levels. There was a good correlation between the data generated by the microplate reader and the FACS analysis, confirming the relative strengths of the candidate promoters (Fig 4).

[Figure omitted. See PDF.]

Fig 4. Comparison of promoter strengths from the two measuring methods.

The X and Y axis represent results from microplate reader and fluorescence-activated cell sorter (FACS) analysis, respectively. A: Transformants of different promoter candidates showed different results. B: Three transformants of the same promoter candidates showed similar results. C: Relationship between different promoter candidates’ strengths measured by microplate reader and FACS.

https://doi.org/10.1371/journal.pone.0158447.g004

Changes in transcriptional level for different strains under stress conditions

Since many of these predicted promoters were originally located in front of heat shock proteins (PsigI, PgroES, and PftsH) and proteins that respond to stress conditions (PyusI, PywrK, PycbR, PypuA, PytpA, and PmurF), we speculated that their expression would increase after the appropriate stress treatments. First, we measured gfp transcriptional levels by real-time fluorescence quantitative PCR for 9 different promoter candidates after treatments, such as heat-shock or high osmotic pressure treatments. Bacteria were cultured in 5 mL of LB media at 37°C for 4.5 h, and then subject to stress conditions. PsigI, PgroES, and PftsH cultures were subject to 15 min of heat shock at 43°C. This sudden temperature upshock is sufficient to induce the synthesis of heat shock proteins including both molecular chaperones such as GroES and ATP-dependent protease such as FtsH [44]. sigI encodes the RNA polymerase sigma factor σ^I, and is required for growth at higher temperatures [45]. Cell envelope stress can result from chemical or genetic impairment, and chemical stress may include salt, ethanol, and superoxide [46]. Cultures of PypuA, PytpA and PmurF in LB were supplemented with ethanol at a final concentration of 4% (v/v) for 15 min. yupA encodes a hypothetical protein, and its promoter was previously reported to be responsive to increases in the stress-responsive extracytoplasmic function ECF sigma factor sigM [47]. ytpA encodes a protein homologous to lysophospholipase, and sporulation was impaired when ytpA was blocked [48]. murF encodes UDP-N-acetylmuramoyl-tripeptide—D-alanyl-D-alanine ligase, which is responsible for the last step in the synthesis of cell wall peptidoglycan. PyusI, PycbR, PywrK were grown and treated with CoCl₂ to a final concentration of 8 mM for 15 min, since proteins encoded by yusI, ycbR and ywrk were predicted to confer resistance to toxic metals, based on similarity. ycbR and yusI encode proteins that are similar to toxic cation resistance protein and arsenate reductase [49], respectively. ywrK encodes an arsenical membrane pump. There were rapid alterations in transcription levels after the treatments, confirmed by qRT-PCR results (Fig 5A). Surprisingly, PftsH and PmurF levels decreased dramatically, PywrK and PypuA were slightly reduced, and the other promoters showed increased activities. We next measured the expression levels of gfp after these stress treatments.

[Figure omitted. See PDF.]

Fig 5. Changes of activity of promoter candidates at the transcriptional and expression level under different conditions.

A: Level of changes of mRNA level of GFP measured by real-time fluorescence quantitative PCR after specific condition treatments corresponding to equivalent cultures incubated in LB media at 37°C. Strains carrying promoter candidates PsigI, PgroES, PftsH were transferred to 43°C for 15 min after incubating at 37°C for 4.5 h. Strains carrying promoter candidates PyupA, PytpA, PmurF were incubated in LB media with 4% (v/v) ethanol for 15 min after incubating in LB media for 4.5 h. Strains carrying promoter candidates PyusI, PycbR and PywrK were incubated in LB media with 8 mM CoCl₂ for 15 min after culturing in LB media for 4.5 h. B: Relative activity of GFP expression level of different promoter candidates when cultured in LB media at 37°C for 5.5 h and heat shock at 43°C for another one hour. C: Relative activity of GFP expression level of different promoter candidates when cultured in LB media and LB media with 4% (v/v) ethanol for 6.5 h.

https://doi.org/10.1371/journal.pone.0158447.g005

Strength of B. subtilis promoter candidates under different stress conditions

We measured the transcriptional levels as the first step to assess potential activity changes for the specific promoters under different conditions. Despite appreciable changes, both increases and decreases, in mRNA levels (as measured by qRT-PCR) (Fig 5A), we did not detect significant changes in GFP activity (S1B Fig). We present the comparison of changes of selected promoters for both transcription and protein expression after the different treatments in S1 Fig. We assumed that after stress treatments, tested promoter candidates would respond by increasing expression level. Unexpectedly, some of these heat shock proteins’ promoters exhibited decreased but not increased activities in driving gfp expression when incubated under the stress conditions of 48°C or 43°C for 6.5 h (Data not shown). This may be because too long time at this high temperature may inhibit vegetative growth. In particular, the temperature treatment of 48°C caused a majority of bacteria to show a slow-growth phenotype. Necessary enzymes involved in central carbon metabolic and cell division might be inhibited under this stress. If the time for heat shock was short, e.g. 15 min, there were sharp shifts in transcriptional levels detected through qRT-PCR (Fig 5A). However, longer time (e.g. one hour) was essential to allow accumulation of proteins to test the response to heat shock (Fig 5B). However, after incubation at 37°C for 5.5 h and heat shock at 43°C for another one hour, any changes in GFP expression were minor (Fig 5B) compared to the mRNA changes measured for the subset of 9 promoters studied by qRT-PCR (Fig 5A). The heat shock stimulon is quite complex, and involves several classes of heat shock genes controlled by transcriptional repressors, alternative sigma factors, and two-component signal transduction system. Different temperatures, heat shock periods, growth media and strains may influence the final results [44].

Next, we investigated whether promoter candidates respond to stress by changing the growth conditions. For this, we inoculated these bacteria into LB media containing 4%, 6%, or 8% (v/v) ethanol. At ethanol concentrations of 6% and 8%, bacterial growth was slowed. However, we detected no significant changes in GFP expression after treatment with 4% ethanol (Fig 5C). 11 of 46 promoter candidates showed a relatively sharp response after incubation in LB with 0.8 M NaCl (Fig 6). However, of those, strains carrying PmreB, PpssA, PoatA, PytpA, PyuaF, PydaH, PyjoB, and PywbO showed no changes in expression at 0.4 M or 0.6 M NaCl (Fig 6). At final concentrations of 1.2 M NaCl and 6.5 h of incubation, bacterial growth was inhibited.

[Figure omitted. See PDF.]

Fig 6. Strength of different promoter candidates under different conditions.

The strength of promoters was measured when strains were cultured in LB media alone, or supplemented with 0.4 to 0.8 M NaCl for 6.5 h. Only some promoters showed a slight response with 0.8 M NaCl. Error bars represent standard deviations of biological triplicates. Activities of most promoter candidates did not change relative to the activities in LB.

https://doi.org/10.1371/journal.pone.0158447.g006

Correlation between BgaB activity and GFP fluorescence for promoter candidates

It is important that activities of promoters we measured are context-independent, i. e., that promoter activities remain constant regardless of the downstream gene being regulated [50]. To investigate this, we compared the GFP activity from plasmid pDL-GFP with that of the same promoter in the original plasmid pDL, with the reporter gene, bgaB, encoding thermostable β-galactosidase. We selected 10 promoter candidates spanning a wide range of strengths, including PradA, PmreBH, PmurF, PywnJ, PyxjI, PyxzE, PtrnQ, PsigX, PsecDF, and PugtP. Except for the different promoter fragments, the backbone of pDL carrying bgaB was the same as the one carrying gfp. There was a linear correlation between the measured GFP and BgaB activities (Fig 7). Thus, the effects we tested were independent of the reporter gene.

[Figure omitted. See PDF.]

Fig 7. Relationship of thermostable β-galactosidase (BgaB) and GFP expression levels.

The same promoter candidates were tested upstream of β-galactosidase (BgaB) and GFP and showed similar results. ONP was detected at OD₄₂₀ indicating that ONPG, the substrate of BgaB, was hydrolyzed.

https://doi.org/10.1371/journal.pone.0158447.g007

Increased secretion production of α-amylase in B. subtilis

To further validate the promoter activities, we substituted the P43 promoter upstream of the secreted protein α-amylase with PtrnQ in the pMA5 plasmid. B. subtilis 1A751 was chosen as the host strain for the expression of this shuttle-plasmid. During 72 h of fermentation in 250 mL triangular flasks, α-amylase was secreted into media and the activity was analyzed every 12 h. The enzyme activity peak time occurred close to 60 h. The α-amylase production driven by PtrnQ was only twice as high as that from the P43 promoter. This is less than the fourfold observed when gfp and bgaB were used to monitor PtrnQ and P43 activities (Figs 2 and 8). The difference could be because α-amylase is a secretory protein and some secretion bottlenecks might hinder the efficient transportation of this protein. However, despite the differences in the fold activation, this result showed that PtrnQ is a strong promoter compared to P43 and could be used to improve the production of secreted proteins.

[Figure omitted. See PDF.]

Fig 8. Comparison of P43 and PtrnQ driving α-amylase in pMA5.

A: Enzyme activity of α-amylase that secreted into the medium during fermentation in flasks for 72 h. Error bars represent standard deviations of biological triplicates. B: SDS-PAGE analysis of α-amylase in the supernatant secreted by B. subtilis 1A751. 10 μL of each supernatant sample was loaded on the gel. Lane M: molecular weight marker; Lane 1: pMA5 without α-amylase; Lane 2, 4, 6 represented pMA5 containing P43 after incubation for 36 h, 48 h, and 60 h, respectively. Lane 3, 5, 7 represent pMA5 containing PtrnQ after incubation for 36 h, 48 h, and 60 h, respectively.

https://doi.org/10.1371/journal.pone.0158447.g008

Discussion

Promoters are fundamental genetic materials that determine transcription and expression. The use of these elements as biobricks will allow their use in directed synthetic biology to obtain fine-tuned and balanced expression of genes. The Registry of Standard Biological Parts (RSBP) (http://parts.igem.org/Promoters) continues to collect an increasing number of genetic parts for use as components of biological devices and systems. However, until now, most promoter information has focused on sequences that work in E. coli. There have been some attempts to discover and optimize suitable promoters for B. subtilis, for instance, the Ludwig Maximilian University (LMU) of Munich team similarly tried to “open Bacillus subtilis and its unique biology to the iGEM world” by measuring promoter activity with reporter genes (http://2012.igem.org/Team:LMU-Munich/Data). However, only a small number of the promoters they used were native to B. subtilis. Other reports have tested individual specific promoters, but there has been no comprehensive analysis of different promoter candidates derived from diverse proteins. Hence, we determined to evaluate promoters of interest and normalize their strength to the well-characterized P43.

If genes are expressed at a high level in favorable, unfavorable or threatening circumstances, they are likely to contain strong promoters or promoters that may respond to certain environmental signals (http://science.sciencemag.org/content/suppl/2012/02/29/335.6072.1103.DC1). We selected 84 promoter sequences mainly located upstream of heat shock proteins, cell envelope proteins, and proteins resistant against toxic metals. Tested promoter candidates showed a range of strengths relative to P43, and several of these were tested for function in different genetic circuits, indicating their use in a variety of genetic contexts to allow gene fine-tuning. One approach is using pools of thousands of synthetic, codon-usage variants to try to regulate protein production, but this process is inefficient and expensive [51]. An alternative approach is to control metabolic pathways by precisely controlled gene transcription using validated promoters, particularly when only inefficient high-throughput screening methods are available. A successful example of the application of a promoter library is the genetic module created for the synthesis of fatty acid ethyl ester (FAEE) by dynamic sensor regulated systems in E. coli [52].

Prior to evaluation of these promoter candidates for their response to various stress treatments, we expected that some of the applied stresses would produce significant changes in gene expression. However, only 11 candidates (PmreB, PpssA, PpspA, Pspo0M, PyqeZ, PoatA, PytpA, PyuaF, PydaH, PyjoB and PywbO) exhibited appreciable shifts when incubating in LB with 0.8 M NaCl at 37°C for 6.5 h. In addition, the putative stress responsive promoters were actually weaker than the constitutive promoter PtrnQ (Fig 2). Several reasons may explain this finding. Firstly, our small sample size may not have included many promoter candidates that are significantly responsive to these tested stress conditions continuously. Secondly, the profiles of mRNA and protein expression may not be strictly correlated, for example if significant mRNA processing occurs [53]. Thus, some promoter candidates may have responded to stress treatments by increasing their transcription levels, but that increased transcription may not have led to increased translation under these conditions [54]. Our experiment, Figs 5A and S1, showed this was the case for at least 9 of the promoters. Finally, during stress, in addition to changes of stress-specific protein expression levels, rearrangements in the composition and structure of cell envelope occur in order to obtain high salt-resistant capacity. This may affect genes involved with lipid and fatty acid composition of the cytoplasmic membrane [54]. For example, ydaH and ytpA encode lipid II flippase and phospholipase, respectively, proteins involved in lipid metabolism. Both YuaF and YqeZ belong to the NfeD family, and YuaF plays a role in maintaining membrane integrity during conditions of cellular stress [55]. In general, many promoters may maintain their relative activity levels under different growth conditions. The changes we observed, however, derived from global effects but not specific responses [56]. Before application of these so-called condition-inducible promoters, further validation was necessary to demonstrate specific regulated response rather than global regulation, as well as validating the specific sequences recognized by different sigma factors [57].

The promoter library constituted of intrinsic biological elements gives us a more clear understanding of their expression patterns when exposed to adverse culture conditions, which is beyond the reach of a synthetic promoter library. This can be a good complement to artificial promoters for synthetic biology and metabolic engineering. The promoter PtrnQ was strong when positioned upstream of gfp and bgaB, but performed less well when we tried to improve the production of α-amylase. Though pMA5 contains an efficient native signal peptide of α-amylase SP_amyl [23], other bottlenecks such as a lack of chaperones may have limited the secretion efficiency of α-amylase. Overall, the yield of a secretory protein likely does not always depend linearly on promoter strength.

Conclusions

We constructed an integrated plasmid to measure the activities of promoter candidates by using a gfp reporter. These promoter candidates were selected from four categories of proteins belonging to the group of heat shock proteins, stress response proteins, or proteins which are resistant against toxic metals (based on similarity) and other kind of genes. We compared these 84 promoter candidates to the well-characterized P43 to determine the promoters’ relative strength. Tested promoters can now be considered for use as biobrick elements in B. subtilis. One strong promoter, PtrnQ, increased the secretion of α-amylase to 2.1 fold that of P43 driven expression. Other inducible promoters will be useful for fundamental research or application in industrial production.

Supporting Information

[Figure omitted. See PDF.]

S1 Fig. Comparison of changes of GFP on the level of both transcription and protein expression after different stress treatments.

A: Level of changes of mRNA level of GFP after specific condition treatments corresponding to equivalent cultures incubated in LB media at 37°C. Strains carrying promoter candidates PsigI, PgroES, PftsH were heated at 43°C for 15 min after incubating at 37°C for 4.5 h. Strains carrying promoter candidates PyupA, PytpA, PmurF were incubated in LB media with 4% (v/v) ethanol for 15 min after incubating in LB media for 4.5 h. Strains carrying promoter candidates PyusI, PycbR and PywrK were incubated in LB media with 8 mM CoCl₂ for 15 min after culturing in LB media for 4.5 h. B: Level of changes of expression level of GFP after specific condition treatments corresponding to equivalent cultures incubated in LB media at 37°C. Strains carrying promoter candidates PsigI, PgroES, PftsH were heated at 43°C for 1 hour after incubating at 37°C for 5.5 h. Strains carrying promoter candidates PyupA, PytpA, PmurF were incubated in LB media with 4% (v/v) ethanol for 6.5 h. Strains carrying promoter candidates PyusI, PycbR and PywrK were incubated in LB media with 8 mM CoCl₂ for 1 hour after culturing in LB media for 5.5 h.

https://doi.org/10.1371/journal.pone.0158447.s001

(TIF)

S1 Table. Primers used to amplify promoter candidates in this study.

https://doi.org/10.1371/journal.pone.0158447.s002

(DOCX)

S2 Table. Predicted -10 boxes, -35 boxes, spacers and regulated sigma factors of these promoter candidates.

https://doi.org/10.1371/journal.pone.0158447.s003

(DOCX)

S3 Table. Functions of proteins encoded by genes downstream of these promoter candidates and their categories.

https://doi.org/10.1371/journal.pone.0158447.s004

(DOCX)

Acknowledgments

Authors would like to express their thanks to financial supports from National Nature Science Foundation of China (31370089, 21506244, 31400079), State Key Development 973 Program for Basic Research of China (2013CB733601), Nature Science Foundation of Tianjin City (CN) (16JCYBJC23500, 15JCQNJC09500) and the Guangxi Science and Technology Development Project (14125008-2-22).

Author Contributions

Conceived and designed the experiments: YS GF DZ. Performed the experiments: YS JN GF JC. Analyzed the data: YS JN GF JC QL NX. Wrote the paper: YS JN JC PZ JS DZ.

Citation: Song Y, Nikoloff JM, Fu G, Chen J, Li Q, Xie N, et al. (2016) Promoter Screening from Bacillus subtilis in Various Conditions Hunting for Synthetic Biology and Industrial Applications. PLoS ONE 11(7): e0158447. https://doi.org/10.1371/journal.pone.0158447

References

1. Kang SY, Choi O, Lee JK, Ahn JO, Ahn JS, Hwang BY, et al. (2015) Artificial de novo biosynthesis of hydroxystyrene derivatives in a tyrosine overproducing Escherichia coli strain. Microb Cell Fact 14: 78. pmid:26055892

2. Kang SY, Choi O, Lee JK, Hwang BY, Uhm TB, Hong YS. (2012) Artificial biosynthesis of phenylpropanoic acids in a tyrosine overproducing Escherichia coli strain. Microb Cell Fact 11: 153. pmid:23206756

3. Ng CY, Khodayari A, Chowdhury A, Maranas CD (2015) Advances in de novo strain design using integrated systems and synthetic biology tools. Curr Opin Chem Biol 28: 105–114. pmid:26177080

4. Qi H, Li BZ, Zhang WQ, Liu D, Yuan YJ (2015) Modularization of genetic elements promotes synthetic metabolic engineering. Biotechnol Adv.

5. Wang G, Bai L, Wang Z, Shi T, Chen T, Zhao X. (2014) Enhancement of riboflavin production by deregulating gluconeogenesis in Bacillus subtilis. World J Microbiol Biotechnol 30: 1893–1900. pmid:24477882

6. Duan YX, Chen T, Chen X, Zhao XM (2010) Overexpression of glucose-6-phosphate dehydrogenase enhances riboflavin production in Bacillus subtilis. Appl Microbiol Biotechnol 85: 1907–1914. pmid:19779711

7. Wang Z, Chen T, Ma X, Shen Z, Zhao X (2011) Enhancement of riboflavin production with Bacillus subtilis by expression and site-directed mutagenesis of zwf and gnd gene from Corynebacterium glutamicum. Bioresour Technol 102: 3934–3940. pmid:21194928

8. Tojo S, Satomura T, Kumamoto K, Hirooka K, Fujita Y (2008) Molecular mechanisms underlying the positive stringent response of the Bacillus subtilis ilv-leu operon, involved in the biosynthesis of branched-chain amino acids. J Bacteriol 190: 6134–6147. pmid:18641142

9. Bienick MS, Young KW, Klesmith JR, Detwiler EE, Tomek KJ, Whitehead TA. (2014) The interrelationship between promoter strength, gene expression, and growth rate. PLoS One 9: e109105. pmid:25286161

10. Chen J, Zhu X, Tan Z, Xu H, Tang J, Xiao D, et al. (2014) Activating C4-dicarboxylate transporters DcuB and DcuC for improving succinate production. Appl Microbiol Biotechnol 98: 2197–2205. pmid:24323285

11. Yamanishi M, Ito Y, Kintaka R, Imamura C, Katahira S, Ikeuchi A, et al. (2013) A genome-wide activity assessment of terminator regions in Saccharomyces cerevisiae provides a ''terminatome'' toolbox. ACS Synth Biol 2: 337–347. pmid:23654277

12. Curran KA, Karim AS, Gupta A, Alper HS (2013) Use of expression-enhancing terminators in Saccharomyces cerevisiae to increase mRNA half-life and improve gene expression control for metabolic engineering applications. Metab Eng 19: 88–97. pmid:23856240

13. Alper H, Fischer C, Nevoigt E, Stephanopoulos G (2005) Tuning genetic control through promoter engineering. Proc Natl Acad Sci U S A 102: 12678–12683. pmid:16123130

14. Blazeck J, Liu L, Redden H, Alper H (2011) Tuning gene expression in Yarrowia lipolytica by a hybrid promoter approach. Appl Environ Microbiol 77: 7905–7914. pmid:21926196

15. Davis JH, Rubin AJ, Sauer RT (2011) Design, construction and characterization of a set of insulated bacterial promoters. Nucleic Acids Res 39: 1131–1141. pmid:20843779

16. Lou C, Stanton B, Chen YJ, Munsky B, Voigt CA (2012) Ribozyme-based insulator parts buffer synthetic circuits from genetic context. Nat Biotechnol 30: 1137–1142. pmid:23034349

17. Bai C, Zhang Y, Zhao X, Hu Y, Xiang S, Miao J, et al. (2015) Exploiting a precise design of universal synthetic modular regulatory elements to unlock the microbial natural products in Streptomyces. Proc Natl Acad Sci U S A.

18. Liu L, Liu Y, Shin HD, Chen RR, Wang NS, Li J, et al. (2013) Developing Bacillus spp. as a cell factory for production of microbial enzymes and industrially important biochemicals in the context of systems and synthetic biology. Appl Microbiol Biotechnol 97: 6113–6127. pmid:23749118

19. Westers L, Westers H, Quax WJ (2004) Bacillus subtilis as cell factory for pharmaceutical proteins: a biotechnological approach to optimize the host organism. Biochim Biophys Acta 1694: 299–310. pmid:15546673

20. Ferreira LC, Ferreira RC, Schumann W (2005) Bacillus subtilis as a tool for vaccine development: from antigen factories to delivery vectors. An Acad Bras Cienc 77: 113–124. pmid:15692682

21. Jones KL, Kim SW, Keasling JD (2000) Low-copy plasmids can perform as well as or better than high-copy plasmids for metabolic engineering of bacteria. Metab Eng 2: 328–338. pmid:11120644

22. Chen J, Gai Y, Fu G, Zhou W, Zhang D, Wen J. (2015) Enhanced extracellular production of alpha-amylase in Bacillus subtilis by optimization of regulatory elements and over-expression of PrsA lipoprotein. Biotechnol Lett 37: 899–906. pmid:25515799

23. Chen J, Fu G, Gai Y, Zheng P, Zhang D, Wen J. (2015) Combinatorial Sec pathway analysis for improved heterologous protein secretion in Bacillus subtilis: identification of bottlenecks by systematic gene overexpression. Microb Cell Fact 14: 92. pmid:26112883

24. Overkamp W, Ercan O, Herber M, van Maris AJ, Kleerebezem M, Kuipers UP. (2015) Physiological and cell morphology adaptation of Bacillus subtilis at near-zero specific growth rates: a transcriptome analysis. Environ Microbiol 17: 346–363. pmid:25367190

25. Bernhardt J, Weibezahn J, Scharf C, Hecker M (2003) Bacillus subtilis during feast and famine: visualization of the overall regulation of protein synthesis during glucose starvation by proteome analysis. Genome Res 13: 224–237. pmid:12566400

26. Kohlstedt M, Sappa PK, Meyer H, Maass S, Zaprasis A, Hoffmann T, et al. (2014) Adaptation of Bacillus subtilis carbon core metabolism to simultaneous nutrient limitation and osmotic challenge: a multi-omics perspective. Environ Microbiol 16: 1898–1917. pmid:24571712

27. Trang TPP, Nguyen D, Schumann W (2012) Development of a strong intracellular expression system for Bacillus subtilis by optimizing promoter elements. Journal Of Biotechnology 157: 167–172. pmid:22100269

28. Wang J, Ai X, Mei H, Fu Y, Chen B, Yu Z, et al. (2013) High-throughput identification of promoters and screening of highly active promoter-5'-UTR DNA region with different characteristics from Bacillus thuringiensis. PLoS One 8: e62960. pmid:23675447

29. Smolke CD, Silver PA (2011) Informing biological design by integration of systems and synthetic biology. Cell 144: 855–859. pmid:21414477

30. Nicolas P, Mader U, Dervyn E, Rochat T, Leduc A, Pigeonneau N, et al. (2012) Condition-dependent transcriptome reveals high-level regulatory architecture in Bacillus subtilis. Science 335: 1103–1106. pmid:22383849

31. Wang PZ, Doi RH (1984) Overlapping promoters transcribed by Bacillus subtilis sigma 55 and sigma 37 RNA polymerase holoenzymes during growth and stationary phases. J Biol Chem 259: 8619–8625. pmid:6330116

32. You C, Zhang XZ, Zhang YH (2012) Simple cloning via direct transformation of PCR product (DNA Multimer) to Escherichia coli and Bacillus subtilis. Appl Environ Microbiol 78: 1593–1595. pmid:22194286

33. Koga K, Ikegami A, Nakasone K, Murayama R, Akanuma G, Natori Y, et al. (2006) Construction of Bacillus subtilis strains carrying the transcriptional bgaB fusion with the promoter region of each rrn operon and their differential transcription during spore development. J Gen Appl Microbiol 52: 119–124. pmid:16778356

34. Wong Ng J, Chatenay D, Robert J, Poirier MG (2010) Plasmid copy number noise in monoclonal populations of bacteria. Phys Rev E Stat Nonlin Soft Matter Phys 81: 011909. pmid:20365401

35. Yuan G, Wong SL (1995) Regulation of groE expression in Bacillus subtilis: the involvement of the sigma A-like promoter and the roles of the inverted repeat sequence (CIRCE). J Bacteriol 177: 5427–5433. pmid:7559325

36. Cormack BP, Valdivia RH, Falkow S (1996) FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173: 33–38. pmid:8707053

37. Zhang XZ, You C, Zhang YH (2014) Transformation of Bacillus subtilis. Methods Mol Biol 1151: 95–101. pmid:24838881

38. Schneider TD, Stephens RM (1990) Sequence logos: a new way to display consensus sequences. Nucleic Acids Res 18: 6097–6100. pmid:2172928

39. Crooks GE, Hon G, Chandonia JM, Brenner SE (2004) WebLogo: a sequence logo generator. Genome Res 14: 1188–1190. pmid:15173120

40. Haldenwang WG (1995) The sigma factors of Bacillus subtilis. Microbiol Rev 59: 1–30. pmid:7708009

41. Rollenhagen C, Antelmann H, Kirstein J, Delumeau O, Hecker M, Yudkin MD. (2003) Binding of sigma(A) and sigma(B) to core RNA polymerase after environmental stress in Bacillus subtilis. J Bacteriol 185: 35–40. pmid:12486038

42. Higgins D, Dworkin J (2012) Recent progress in Bacillus subtilis sporulation. FEMS Microbiol Rev 36: 131–148. pmid:22091839

43. Kingston AW, Liao X, Helmann JD (2013) Contributions of the sigma(W), sigma(M) and sigma(X) regulons to the lantibiotic resistome of Bacillus subtilis. Mol Microbiol 90: 502–518. pmid:23980836

44. Schumann W (2003) The Bacillus subtilis heat shock stimulon. Cell Stress Chaperones 8: 207–217. pmid:14984053

45. Asai K, Ootsuji T, Obata K, Matsumoto T, Fujita Y, Sadaie Y. (2007) Regulatory role of RsgI in sigI expression in Bacillus subtilis. Microbiology 153: 92–101. pmid:17185538

46. Eiamphungporn W, Helmann JD (2008) The Bacillus subtilis sigma(M) regulon and its contribution to cell envelope stress responses. Mol Microbiol 67: 830–848. pmid:18179421

47. Jervis AJ, Thackray PD, Houston CW, Horsburgh MJ, Moir A (2007) SigM-responsive genes of Bacillus subtilis and their promoters. J Bacteriol 189: 4534–4538. pmid:17434969

48. Tamehiro N, Okamoto-Hosoya Y, Okamoto S, Ubukata M, Hamada M, Naganama H, et al. (2002) Bacilysocin, a novel phospholipid antibiotic produced by Bacillus subtilis 168. Antimicrob Agents Chemother 46: 315–320. pmid:11796336

49. Guo X, Li Y, Peng K, Hu Y, Li C, Xia B, et al. (2005) Solution structures and backbone dynamics of arsenate reductase from Bacillus subtilis: reversible conformational switch associated with arsenate reduction. J Biol Chem 280: 39601–39608. pmid:16192272

50. Haynes KA, Silver PA (2009) Eukaryotic systems broaden the scope of synthetic biology. J Cell Biol 187: 589–596. pmid:19948487

51. Quan J, Saaem I, Tang N, Ma S, Negre N, Gong H, et al. (2011) Parallel on-chip gene synthesis and application to optimization of protein expression. Nat Biotechnol 29: 449–452. pmid:21516083

52. Zhang F, Carothers JM, Keasling JD (2012) Design of a dynamic sensor-regulator system for production of chemicals and fuels derived from fatty acids. Nat Biotechnol 30: 354–359. pmid:22446695

53. Brewster RC, Jones DL, Phillips R (2012) Tuning promoter strength through RNA polymerase binding site design in Escherichia coli. PLoS Comput Biol 8: e1002811. pmid:23271961

54. Hahne H, Mader U, Otto A, Bonn F, Steil L, Bremer E, et al. (2010) A comprehensive proteomics and transcriptomics analysis of Bacillus subtilis salt stress adaptation. J Bacteriol 192: 870–882. pmid:19948795

55. Dempwolff F, Moller HM, Graumann PL (2012) Synthetic motility and cell shape defects associated with deletions of flotillin/reggie paralogs in Bacillus subtilis and interplay of these proteins with NfeD proteins. J Bacteriol 194: 4652–4661. pmid:22753055

56. Keren L, Zackay O, Lotan-Pompan M, Barenholz U, Dekel E, Sasson V, et al. (2013) Promoters maintain their relative activity levels under different growth conditions. Mol Syst Biol 9: 701. pmid:24169404

57. Gerosa L, Kochanowski K, Heinemann M, Sauer U (2013) Dissecting specific and global transcriptional regulation of bacterial gene expression. Mol Syst Biol 9: 658. pmid:23591774