Construction of strains to investigate α-CTD activity in vivo

To examine the in vivo activities of α-CTD, we constructed strains in which the expression of RpoA (we use “RpoA” to indicate the product of the rpoA gene, instead of “α-subunit") could be switched from an intact RpoA (RpoA^int) to a C-terminally truncated RpoA (RpoA^del). RpoA^del lacked the folded domain of α-CTD, which is composed of four α-helices(α1 ~ α4, 66 amino acids), while retaining the four amino acids at the C-terminal end (RKDD, outside of the folded domain) (S1 Fig) [30]. The expressions of RpoA^int and RpoA^del were respectively controlled by the inducible promoters, Pspac, which is repressed by the LacI repressor and activated by IPTG (isopropyl β-D-1-thiogalactopyranoside), and Pxyl which is repressed by the XylR repressor and activated by xylose (Fig 1). The rpoA gene is co-transcribed with infA (encoding a translation initiation factor), rpmJ, rpmM, rpmK and rplQ (encoding ribosome subunits) (Fig 1A) [31]. The transcription of this transcriptional unit (TU) initiates upstream of infA, while the transcriptional terminator follows after the termination codon of rplQ [31]. We introduced the Pspac promoter with lacI (encoding LacI) into the intergenic region between rpoA and rpmK on the B. subtilis chromosome (Fig 1A), thus, terminating transcription of infA upstream of rpoA and putting the expression of rpoA and rplQ under the control of the Pspac promoter. In addition, a DNA fragment containing the Pxyl promoter followed by the genes encoding RpoA^int or RpoA^del and xylR (encoding XylR) was introduced into the amyE locus of the B. subtilis chromosome (Fig 1B). The rplQ gene is essential for B. subtilis cell viability [31]. The expressions of rpoA and rplQ via the Pspac promoter were simultaneously repressed by LacI when our constructed strain was grown in LB medium without IPTG. To compensate for the depletion of rplQ, we introduced rplQ downstream of the Pxyl promoter and rpoA into the amyE locus. Our constructed strain, which harbored an IPTG-inducible intact rpoA (rpoA^int) gene and a xylose-inducible truncated rpoA gene (rpoA^del), was unable to grow on the LB plate containing 1% xylose and lacking IPTG, whereas the strain harboring xylose-inducible rpoA^int (instead of rpoA^del) at the amyE locus was viable on the same LB plate (S2 Fig). This confirms that α-CTD is essential for the cell viability of B. subtilis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schematic representation of the loci used to switch rpoA expression from IPTG-to xylose-inducible.

Open thick arrows represent genes, while thick gray arrows represent genes whose expression levels were induced by IPTG or xylose via the Pspac and Pxyl promoters, respectively. The dashed line (with arrowhead) represents the region in which transcription was initiated from the original promoter of the operon; this region includes rpoA and rplQ, which are expected to be located upstream of infA [31]. Gray solid lines represent the transcribed regions of the new transcriptional units (Pspac-rpoA^int-rplQ, Pxyl-rpoA^int-rplQ, Pxyl-rpoA^del-rplQ or Pxyl-rplQ) regulated by the Pspac and Pxyl promoters. Native rpoA locus (A) and amyE locus (B) are shown.

https://doi.org/10.1371/journal.pone.0131588.g001

We designated the constructed strains as SMS05 (Pspac-rpoA^int-rplQ, Pxyl-rpoA^int-rplQ), SMS06 (Pspac-rpoA^int-rplQ, Pxyl-rpoA^del-rplQ) and SMS07 (Pspac-rpoA^int-rplQ, Pxyl-rplQ) and used this system to examine the function of α-CTD in B. subtilis.

Growth of cells following the induction of RpoA^del

We first examined the effects of the RpoA^del expression/RpoA^int depletion on the cell growth. B. subtilis cells were cultivated to log phase in LB medium containing 1 mM IPTG (LB^IPTG), to trigger the expression of RpoA^int. The cells were, then, washed with LB medium to remove the IPTG, and re-cultivated in LB medium containing 1% xylose (LB^xyl). The growth curves of the rpoA^int- and the rpoA^del-expressing strains in LB^xyl are shown in Fig 2A. After the induction of rpoA^del expression, the growth of both strains continued for at least 6 hours. There was little difference in the growth rates during the initial 2 hours of the cultivation in LB^xyl (Fig 2A, blue and pink lines). However, the growth rate of rpoA^del-expressing cells over the total 6-hour cultivation (Fig 2A, pink line) was significantly reduced, compared with that of rpoA^int-expressing cells (Fig 2A, blue line); the doubling time of rpoA^del expressing cells was 1.9 hours, while that of the rpoA^int-expressing cells was 1.0 hour. Meanwhile, SMS07, in which rplQ was expressed, but RpoA was depleted, showed growth arrest in LB^xyl (Fig 2A, green line), suggesting that rpoA^del-expressing cells had sufficient RNAP activity to maintain cell growth for several hours.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The effects of replacing RpoA^int with RpoA^del in the RNAP complexes of rpoA^del-expressing cells.

(A) Growth curves of strains that expressed RpoA^int (SMS05, blue line), RpoA^del (SMS06, pink line) and RplQ without RpoA (SMS07, green line), upon xylose induction. (B and C) Pull-down assays of RpoA assembled in RNAP complexes. RNAP complexes were purified using His-tagged RpoC as bait, separated by SDS-PAGE and detected by fluorescent staining with Flamingo dye (Bio-Rad), which is characterized by a high sensitivity and a low background, allowing accurate quantification [32]. The band positions corresponding to RpoB and C, RpoA^int and RpoA^del in the SDS-PAGE gel are indicated by arrows (right). The duration of cultivation in LB^xyl is shown at the bottom of each panel. (D) Time course of alteration in the relative amounts of RpoA^int and RpoA^del in the RNAP complexes of rpoA^int- and rpoA^del-expressing cells grown in LB^xyl. The blue line indicates the relative ratio (see below) of the amount of RpoA^int in the RNAP complexes of rpoA^int-expressing cells at each time point. For normalization, the amount of RpoA^int was divided by that of RpoB, C in each lane as follows: normalized signal intensity of RpoA = [signal intensity of the RpoA band] / [signal intensity of the RpoB, C band]. The amount of RpoA^int at each time point was then divided by the amount of RpoA^int at time 0 as follows: relative ratio (RpoA^{int (X hour)}) = [normalized signal intensity of RpoA^{int (X hour)}] / [normalized signal intensity of RpoA^{int (0 hour)}]. The orange line represents the relative ratios of RpoA^int in the RNAP complexes of rpoA^del-expressing cells at each time point, calculated as described above. The black line shows the relative ratios of the amount of RpoA^del in the RNAP complexes of rpoA^del-expressing cells at each time point against to the amount of RpoA^del at 6 hours after the beginning of the RpoA^del induction. Here, the relative ratio was calculated as: relative ratio (RpoA^{del (X hour)}) = [normalized signal intensity of RpoA^{del (X hour)}] / [normalized signal intensity of RpoA^{del (6 hour)}]. Each relative ratio was calculated using the average values obtained from triplicate experiments.

https://doi.org/10.1371/journal.pone.0131588.g002

In the RNAP complex, RpoA^int is replaced by RpoA^del in rpoA^del-expressing cells

In order to compare the activity of RpoA^del with that of RpoA^int in transcriptional regulation, the expressed RpoA^del must be correctly assembled into the RNAP complex. To test whether this was the case, we constructed strains in which RpoC (the β’ subunit of RNAP) was expressed with a C-terminal histidine-tag. We have demonstrated that the C-terminal histidine-tagged RpoC (RpoC-His) correctly assembled with other RNAP subunits and formed competent RNAP complex, which could interact with the elongation factors GreA and NusA, and showed elongation activity [33]. In addition, the B.subtilis cells expressing the tagged RpoC showed normal growth [34]. These cells were designated SMS08 (Pspac-rpoA^int-rplQ, Pxyl-rpoA^int-rplQ, rpoC-His12) and SMS09 (Pspac-rpoA^int-rplQ, Pxyl-rpoA^del-rplQ, rpoC-His12).

Western blotting using the anti-RpoB antibody and crude extracts prepared from rpoA^int- and rpoA^del-expressing cells indicated that the amount of RNAP complex was constant in both cell types during the 6 hours of growth in LB^xyl (S3 Fig). This result indicated that α-CTD deficiency did not affect the amount of RNAP complex in B. subtilis cells grown under our experimental conditions. The RNAP complexes were affinity purified via the C-terminal histidine-tag of RpoC (RpoC-His) and resolved by SDS-PAGE. To evaluate alterations in the amount of RpoA^int and RpoA^del assembled into the RNAP complex, the amounts of RpoA^int were determined at each time point relative to that at the beginning of the RpoA^del induction (0 hour), while those of RpoA^del at each time point were determined relative to that at the end of the 6-hour study period (Fig 2B and 2C). Our results revealed that the relative amount of RpoA^int in the RNAP complexes of rpoA^int-expressing cells was constant during the 6-hour cultivation in LB^xyl (Fig 2D, blue line), while the relative amount of RpoA^int in the RNAP complexes of rpoA^del-expressing cells gradually decreased to about 30% of the initial amount (0 hour) over the first 3 hours of cultivation in LB^xyl and more moderately decreased thereafter (Fig 2D, orange line). This is consistent with the slow turnover of RpoA in RNAP complexes, as previously observed in E. coli cells [8]. In parallel with this reduction in RpoA^int, the amount of RpoA^del assembled in the RNAP complexes increased between 0 and 3 hours of cultivation in this system (Fig 2D, black line). It is possible that the remaining signal at the RpoA^int position after 3 hours of cultivation reflected background noise. Regardless, our results indicate that at least 70% of the RpoA^int assembled into RNAP complexes was replaced by RpoA^del in rpoA^del-expressing cells cultivated in LB^xyl.

RpoA^del-containing RNAP complex can bind DNA and form elongation complexes in B. subtilis

We next questioned whether RNAP assembled with RpoA^del lost DNA-binding activity. We used ChAP-chip analysis of RpoC and RpoA to evaluate the DNA-binding activity of the RpoA^del-containing RNAP toward the B. subtilis chromosome [34]. In addition to the RpoC-His-expressing strains, we constructed strains expressing RpoA with a C-terminal histidine tag (RpoA^int-His and RpoA^del-His) and designated them SMS18 (Pspac-rpoA^int-rplQ, Pxyl-rpoA^int-His-rplQ) and SMS19 (Pspac-rpoA^int-rplQ, Pxyl-rpoA^del-His-rplQ).

Given the above results reflecting the replacement of RpoA^int with RpoA^del in the RNAP complexes of rpoA^del-expressing cells (SMS09) cultivated for 3 hours in LB^xyl, cells were cultured for 3 hours in LB^xyl and then treated with formaldehyde to crosslink the chromosomal DNA and RNAP. ChAP-chip analyses of RpoC-His and RpoA-His were performed as described previously [33,34].

Because there are multiple and highly homologous rRNA and tRNA genes, raising the possibility that cross hybridization could prevent the accurate quantification of their RNAP bindings, we compared the RNAP binding intensities with respect to protein-coding regions on B. subtilis genomic DNA in the rpoA^del- and rpoA^int-expressing cells. The RNAP (RpoC-His) binding intensities in the rpoA^del (SMS09) and rpoA^int (SMS08)-expressing cells were positively correlated (S4A Fig, r > 0.90), as were the binding intensities of RpoA^del-His and RpoA^int-His in the rpoA^del (SMS19) and rpoA^int-expressing cells (SMS18) (S5 Fig, r > 0.96). These results clearly indicated that the RNAP complex assembled with RpoA^del could bind to the B. subtilis chromosome and form an elongation complex. Detailed analysis of the ChAP-chip results discussed later.

The impact of α-CTD deficiency on the B. subtilis transcriptome

To explore the impact of α-CTD deficiency on global transcriptional gene regulation in B. subtilis, the transcriptomes of rpoA^int-expressing cells (SMS08) and rpoA^del-expressing cells (SMS09) were compared using the Affymetrix Genechip B. subtilis genome array (Affymetrix). We used the same growth conditions used in the ChAP-chip analysis and determined the transcriptomes of rpoA^int- and rpoA^del-expressing cells grown in LB^xyl for 0, 1, 2 and 3 hours.

The differences between the transcriptomes of rpoA^int-expressing cells and rpoA^del-expressing cells gradually increased during their cultivation in LB^xyl. The correlation coefficient, therefore, gradually decreased over the same period (Fig 3). To identify the genes whose expression levels were affected by α-CTD deficiency, we extracted data for genes whose expression levels differed by > 4- or < -0.25 fold (> 2 or < -2 in log₂ scale) in the rpoA^del-expressing cells compared to the rpoA^int-expressing cells and which had a false discovery rate (q-value) of < 0.2. After 3 hours cultivation, we identified 53 down-regulated genes and 27 up-regulated genes differentially expressed between the rpoA^int-expressing cells (SMS08) and the rpoA^del-expressing cells (SMS09) (S5 Table). Amongst them, there were 20 down- and 5 up-regulated genes that were also differentially expressed after 2 hours of cultivation (S2 Table). At 2 hours, we identified four additional down-regulated genes (yhfH, yhzC, ykbA and yokG) and one additional up-regulated gene (ybcP) in the rpoA^del-expressing cells. At 3 hours, these genes were excluded from the list of differentially expressed genes for the following reasons: yhzC showed a low expression level; yhfH had a high q-value (log₂ fold = -4.13, q = 0.201); ykbA and yokG had moderate fold changes or a high q-value (ykbA, log₂ fold = -1.95, q = 0.14; yokG, log₂ fold = -2.47, q = 0.24); and ybcP had a high q-value (log₂ fold = 2.64, q = 0.24). In general, however, the up- and down-regulated genes overlapped at 2 and 3 hours. At 0 and 1 hour, no up- or down-regulation was identified by the criteria used in this study. We detected the known down-regulation of the srf operon that was previously associated with α-CTD deficiency [20], but not the known down-regulation of flagellar genes, hag, fliD and motA [25,26]. These results might indicate that the transcriptional regulation of hag, fliD and motA have more factors affecting them than simple UP element dependent regulation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Transcriptome analysis of rpoA^int-expressing cells (SMS08) and rpoA^del-expressing cells (SMS09).

A scatter plot (log₂ scale) of the transcriptional signal intensity (averaged from duplicate experiments) of each gene in rpoA^del-expressing cells (vertical axis) versus rpoA^int-expressing cells (horizontal axis) at 0 hour (A) and at 1 hour (B), 2 hours (C) and 3 hours (D) after the beginning of the RpoA^del induction. The correlation coefficients between the transcriptomes of rpoA^int and rpoA^del-expressing cells are indicated as (r) in each panel. The average signal intensities from two independent experiments are plotted. For each gene plotted, the sum of the signal intensities for all experiments performed at the same time point (two experiments each for rpoA^int- and rpoA^del-expressing cells) was > 400; this avoided the inclusion of minimally expressed genes in our analysis. We analyzed a total of 2755 (0 hour), 2855 (1 hour), 2921 (2 hour) and 2956 (3 hour) genes. The genes found to be down-regulated in rpoA^del-expressing cells compared to rpoA^int-expressing cells at 3 hours after the beginning of the RpoA^del induction are shown as blue dots, and the up-regulated genes are shown by red dots.

https://doi.org/10.1371/journal.pone.0131588.g003

α-CTD deficiency increases the transcription of genes with alternative σ factor-dependent promoters

In the DBTBS and BsubCyc databases [35,36], there is information about promoters for nine of the genes found to be up-regulated in rpoA^del-expressing cells. Of them, six are known to interact with alternative σ factors, while the other three are known to be σ^A dependent (Fig 4 and S1 Table). Therefore, some of the up-regulation caused by α-CTD deficiency might be explained by the activation of alternative σ factors. However, no unique alternative σ factor was found to be responsible for up-regulation in rpoA^del-expressing cells (S1 Table), making it impossible to estimate the stresses and environmental signals required to activate these promoters under α-CTD deficiency. In addition, the three up-regulated TUs found to have σ^A-dependent promoters were also differed from each other in terms of their functions; the skf operon is involved in the transition state response [37]; proH-proJ is involved in stress response [38]; and the yydF operon is involved in cell wall metabolism [39]. Hence, we did not estimate the potential biological relevance of the transcriptional up-regulation by α-CTD deficiency. In addition, the up-regulation induced by α-CTD deficiency was relatively weaker than the corresponding down-regulation (Fig 3C and 3D). Hence, we focused the following analyses on the genes and transcriptional units found to be down-regulated in α-CTD deficient cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Genome-wide map of the genes up- and down-regulated, and highly reduced RNAP binding in rpoA^del-expressing cells (at 3 hours).

From the outermost ring, the genes with highly reduced RNAP binding identified by ChAP-chip analysis (dark-blue bars, gene names corresponding to those bars are indicated by dark-blue letters outside of rings), down-regulated genes identified by transcriptome analysis (sky-blue bars, gene names corresponding to those bars are indicated by sky-blue letters outside of rings), CDSs (orange bars, an outer and inner ring indicates clockwise and counterclockwise CDSs, respectively) and up-regulated genes identified by transcriptome analysis (red bars in innermost ring, gene names corresponding to those bars are indicated by red letters) are shown as bars at their corresponding genome coordinates. The pink letters indicate the genes which were identified as the genes both down-regulated and with highly reduced RNAP binding. The GenomeMatcher was used to draw rings [40].

https://doi.org/10.1371/journal.pone.0131588.g004

Functions and transcriptional regulation of the genes down-regulated under α-CTD deficiency

To examine the potential biological relevance of the α-CTD-dependent activation of gene expression in B. subtilis cells, we investigated the functions of the 53 genes found to be down-regulated after 3 hours of growth in LB^xyl (Fig 4 and Table 1, in which the result of ChAP-chip is also summarized [the ChAP-chip result is discussed below]). The functions and other data regarding these down-regulated genes and transcriptional units (TUs) are presented in Table 2. Interestingly, we identified a large proportion of genes known to be up-regulated in response to changes in nutrient availability. More than half of the genes were related to the transition-state response (13 genes), including the quorum-sensing-related genes (srfAA, AB, AC, AD [srf operon], rapA, rapC and pel) [41,42], and biofilm-formation-related genes (epsB, A and tasA) [43], or the utilization of secondary carbon sources (16 genes), including genes related to the utilization of phosphoenolpyruvate-carbohydrate phosphotransferase system (PTS) sugars and other carbon sources (rbsR, C, B [the rbs operon], gmuB, A, C, D, F [the gmu operon], gamA, P, manP, A, manR and mtlA, D) [44–49] (Fig 4 and Table 2). These results strongly suggest that α-CTD plays an important role in the activation of genes involved in the adaptation to nutrient availability.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Functional categories of the genes that showed down-regulation and/or highly reduced RNAP binding in rpoA^del-expressing cells.

https://doi.org/10.1371/journal.pone.0131588.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of the genes identified by transcriptomic analysis as being down-regulated in rpoA^del-expressing cells.

https://doi.org/10.1371/journal.pone.0131588.t002

ChAP-chip analysis suggests the down-regulation of genes related to utilization of secondary carbon sources and ribosome synthesis in the rpoA^del-expressing cells

To further investigate the α-CTD-dependent activation of gene expression in B. subtilis cells, we extracted the genes whose RNAP bindings were highly reduced in rpoA^del-expressing cells compared to rpoA^int-expressing cells, based on our RpoC-His ChAP-chip result. We calculated the relative ratio of the RNAP binding intensities obtained for each gene from the rpoA^del- and rpoA^int-expressing cells and extracted the 50 genes that were found to be most highly reduced in their RNAP binding (≈ top 1% of all genes), based on the rank of genes in order of the level of reduction (the rank is the average of ranks in duplicate experiments, see S3 Table). The genes highly reduced in their RNAP binding are summarized with up- and down-regulated genes identified by transcriptome in Fig 4 and Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Summary of the genes that showed highly reduced RNAP binding in the rpoA^del-expressing cells, as identified by ChAP-chip analysis.

https://doi.org/10.1371/journal.pone.0131588.t003

Among the genes that showed highly reduced RNAP binding, nine were included in four TUs identified in our transcriptome analysis as being down-regulated in rpoA^del-expressing cells (the srf operon, mtlAFD, mtnKA, and rapA-phrA) (Fig 4; indicated by pink letters, Table 1, Table 3 and S3 Table). Interestingly, the RNAP binding in rpoA^del-expressing cells was also found to be highly reduced for a number of additional genes related to the utilization of secondary carbon sources, including treP, treA, treR (TU: trePAR), cggR, gapA (cggR-gapA-pgk-tpiA-pgm-eno), xylA, B (xylAB), ackA (ackA), glpK, glpF (glpFK), glpD (glpD) and trpP (trpP) (Fig 4, Table 3 and S3 Table). Although these secondary carbon metabolism-related genes were not identified in our transcriptome analysis, our RNAP-binding results suggest that their expression levels may be regulated by α-CTD (the reason for the small overlapping of the transcriptome and ChAP-chip results is discussed in the next sub-section).

Greatly reduced RNAP binding was also observed for 16 genes encoding ribosomal proteins (13 essential and three non-essential) and three encoding essential proteins (secY, adk, mapA) from the five clusters of ribosomal protein genes (Fig 4, Tables 1 and 3 and S3 Table; see also S6 Fig (1, 5 through 7) for ribosomal protein gene clusters infC-rplT, rpsT, rpmGA, rpsJ-rplQ and rplJ-rplL). Although we have only limited information about these TUs and their promoters, and were thus unable to predict whether UP elements or activators contribute to activating these genes in B. subtilis, our results suggest that certain mechanisms may govern the α-CTD-dependent regulation of transcription at these ribosomal protein gene clusters. The α-CTD-dependent activation of essential ribosomal protein genes might explain why α-CTD is essential for cell viability in B. subtilis.

Transcriptome and ChAP-chip analyses are complementary in detecting the impact of the α-CTD deficiency in B. subtilis cells

The signal intensities determined by transcriptome analysis represent the RNAs accumulated in the cells, while the RNAP binding intensities determined by ChAP-chip analysis indicate the amount of transcribing RNAPs present on a given gene. This might explain the differences in our detection of down-regulation in rpoA^del-expressing cells using these two platforms. To examine this possibility, we evaluated the distribution of these genes in a scatter plot generated from our transcriptome analysis (S7 Fig). We found that the genes with highly reduced RNAP binding tended to be represented by fewer transcripts in rpoA^del-expressing cells versus the rpoA^int-expressing cells (S7 Fig, dark-blue dots), but that genes with higher transcript signal intensities were less often identified as being down-regulated by the criteria used in our transcriptome analysis. This suggests that RNA accumulation may have obscured some of the down-regulated genes in our transcriptome analysis, especially if the genes are highly expressed.

In addition, high background signal in the ChAP-chip data might also be the reason for the small overlapping of down-regulation detected by the transcriptome and ChAP-chip analysis. The background signal reduced the sensitivity of detection of differences in RNAP binding to protein coding regions on B. subtilis genomic DNA in rpoA^del- and rpoA^int- expressing cells, especially if the genes were expressed at low levels.

In fact, the signal intensities of the genes with highly reduced RNAP binding identified in ChAP-chip (S7A Fig, dark-blue dots in a scatter plot generated from our transcriptome analysis) tended to be higher than those of the down-regulated genes identified in transcriptome (S7A Fig, sky-blue dots). In addition, the signal intensities of genes identified as down-regulated and with highly reduced RNAP binding were intermediate (S7A Fig, pink dots). This clearly supports our hypothesis that ChAP-chip analysis could mainly detect the difference in RNAP binding to relatively highly expressed genes in the rpoA^del- and rpoA^int-expressing cells. In contrast, transcriptome analysis appears to be more suitable for detection of differential expression in relatively poorly expressed genes in rpoA^del- and rpoA^int- expressing cells.

In addition, the intensities of RNAP binding to the up- and down-regulated genes tended to be increased and decreased, respectively, in rpoA^del-expressing cells compared to rpoA^int-expressing cells (S4A Fig, red and blue dots). This tendency was weak but statistically significant (S4B Fig).

Taken together, we concluded that both transcriptome and ChAP-chip results reflected the impact of α-CTD-deficiency on the global transcriptional regulation of B. subtilis.

Transcriptional regulation of the genes down-regulated and highly reduced RNAP binding under α-CTD deficiency

We next investigated the transcriptional-regulator binding sites related to the down-regulated genes and the genes that had highly reduced RNAP binding in rpoA^del-expressing cells, using the information on the TUs and their transcriptional regulators deposited in the DBTBS and Bsubcyc databases [35,36]. In E. coli, the activators are classified as Class I or Class II based on the location of their binding sites and the mode through which they interact with RNAP to activate gene expression. Class I activators bind upstream of the -35 elements of promoters to interact with α-CTD, while Class II activators interact with the σ subunit via a binding sites that overlap with the -35 element of promoters [1]. From the databases, we found nine TUs, including the genes down-regulated and/or highly reduced RNAP binding in rpoA^del-expressing cells, whose activator binding sites are located upstream of the -35 elements of promoters (S1 and S3 Tables). The activators, ComA (regulating three TUs; the srf operon, rapA-phrA, rapC-phrC), RemA (two TUs; eps and yqxM-sipW-tasA), ManR (two TUs; manPA-yjdF and manR), YufM (one TU; maeN) and YycF(WalR) (one TU; cwlO) might have the potential to directly regulate the expression levels of the relevant TUs as Class I activators via their interaction with α-CTD (the down-regulation of cwlO was identified by ChAP-chip analysis, others were identified by transcriptome). Meanwhile, the transcriptional regulators, CcpA (regulating one TU; rbs operon), SinR (one TU; yqxM-sipW-tasA), AbrB (one TU; spo0E), CtsR (one TU; clpE), LmrA (one TU; lmrAB), YycF(WalR) (two TUs; wapA-yxxG and yoeB), CggR (one TU; cggR-gapA-pgk-tpiA-pgm-eno) and XylR (one TU; xylAB) might bind to promoter regions and/or transcriptional start sites of the nine TUs (two TUs, the down-regulation of cggR-eno and xylAB were identified by ChAP-chip analysis, others were identified by transcriptome) (S1 and S3 Tables). Hence, these transcriptional regulators might have the potential to directly repress the expression levels of these TUs. The α-CTD mediated activation might compete with the direct action of the repressors on the promoter region, and therefore, the TUs might be further repressed in the absence of the α-CTD.

UP elements are located upstream of the -35 promoter elements of genes that show down-regulation and highly reduced RNAP binding in rpoA^del-expressing cells

As the residues responsible for the interaction with the UP elements found in the α-CTD are known to be highly conserved in bacterial RNAPs [4], we next evaluated the contribution of the UP element to regulating the genes that showed down-regulation of transcription and/or highly reduced RNAP binding in rpoA^del-expressing cells.

We first extracted the -41 to -57 sequences of their known promoters [35,36] (35 promoters for 33 TUs, covering 55 genes in total that showed down-regulation of transcription and/or highly reduced RNAP binding in rpoA^del-expressing cells) and compared them with the UP element consensus sequence proposed for E. coli [4,13,14]. Twelve of the promoters were found to contain UP-element-like sequences in which > 60% of the nucleotides coincided with the consensus sequence [4,13,50] (S8 Fig).

We then further evaluated the sequences upstream of the -35 elements of promoters with respect to the transcriptional start sites (TSSs) of B. subtilis, which were experimentally determined by new-generation sequencing [51]. We extracted the -38 to -57 sequence of each promoter (corresponding to the TSS) and searched for conserved nucleotides using the Weblogo program [52]. A conserved AT rich sequence was found in two separate regions upstream of the genes that showed down-regulation of transcription and/or highly reduced RNAP binding in rpoA^del-expressing cells (S9 Fig); centred at the -43 and -53 positions, these sequences corresponded to the proximal and distal subsets of the UP elements proposed in E. coli promoters [4,13]. Little conservation of AT-rich sequences was detected upstream of the -35 promoter elements of the other tested genes. These results suggest that although it is rare for the entire UP element sequence to be conserved upstream of the -35 element of the promoter of a putative α-CTD-dependent gene, such elements are partially conserved, potentially contributing to the activation of gene expression in B. subtilis. This is consistent with the previous observation that half of an UP element is partially able to activate transcription in E. coli [4].

ChAP-chip analysis reveals two types of α-CTD dependent transcriptional activation in B. subtilis

To examine the effects of α-CTD deficiency on RNAP binding, we assessed the RNAP binding profiles of the genes that showed highly reduced RNAP binding in rpoA^del cells upon visual inspection of the RNAP (especially RpoC-His) distributions. The typical distributions of RNAP (revealed by using RpoC-His or RpoA-His) on the B. subtilis chromosome are shown in Fig 5A, and the binding profiles of RNAP throughout the chromosome of B. subtilis (as assessed in duplicate experiments) are presented in S10 Fig.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. RNAP binding effects following the replacement of RpoA^int with RpoA^del in the RNAP complex.

(A) Typical RNAP binding profiles obtained from rpoA^int- and rpoA^del-expressing cells; rpsU and yqeY are indicated, with the RNAP binding signal of each probe mapped to the corresponding position in the B. subtilis chromosome. The binding intensity (shown by vertical bars) was determined as the relative ratio of the signal intensities obtained for the hybridization of labeled DNA fragments prepared from the affinity purification with RpoC or RpoA (ChAP DNA) versus whole cell extract (control DNA) fractions in each experiment. The RNAP binding intensities determined by affinity purification with RpoC as bait are shown in lanes 1 and 2. The RNAP binding intensities determined by affinity purification with RpoA as bait are shown in lanes 3 and 4. The RNAP binding intensities in the rpoA^int-expressing cells are shown in lanes 1 (SMS08) and 3 (SMS18), and those in rpoA^del-expressing cells are indicated in lanes 2 (SMS09) and 4 (SMS19). The arrangement of genes in the presented chromosomal region is indicated by thick arrows at the top of the figure. The RNAP binding profiles obtained from one (Exp. 1) of duplicate experiments are shown as representative. (B) Binding to the mtl operon is shown as a typical example of the reduced RNAP binding observed in rpoA^del-expressing cells. (C) Binding to the srf operon is shown as an example of the unique RNAP binding profile observed in rpoA^del-expressing cells, in which reductions were seen in the protein coding regions but not in the promoter proximal regions.

https://doi.org/10.1371/journal.pone.0131588.g005

The RNAP (RpoC-His) binding signals were generally similar in the rpoA^del- and rpoA^int-expressing cells (Fig 5A, lanes 1 and 2), but they were reduced across the promoters and protein-coding regions for most of the TUs identified as having highly reduced RNAP binding in rpoA^del-expressing cells (Fig 5B, lanes 1 and 2, and S6 Fig (1 through 7). Thus, α-CTD deficiency appears to inhibit the recruitment of RNAP to many promoters in B. subtilis.

Interestingly, we observed unique RNAP binding profiles for cggR-gapA, odhAB, qox and the srf operon in rpoA^del-expressing cells: the RNAP (RpoC-His) binding signals were reduced in the coding regions, but peaked across the promoters or promoter-proximal regions of these genes (Fig 5C, lanes 1 and 2; S6 Fig (8 and 9). These results suggest that RNAP could be recruited to these gene promoters in the absence of α-CTD, but that α-CTD is needed to form the productive complex or overcome pausing or termination signals located in the promoter-proximal regions. While α-CTD could be involved in anti-termination activity or another regulatory activity governing transcriptional elongation, we believe that it is more likely to be involved at the transcriptional initiation step. A previous study showed that the E. coli transcriptional activator, CRP, interacts with α-CTD, and that this enhanced formation of the productive open complex at the malT, lac and gal promoters; in the absence of CRP, in contrast, RNAP formed non-productive (malT) or less active (lac and gal) RNAP-promoter complexes [53]. The RNAP binding profiles determined herein for the cggR-gapA, odhAB, qox and srf operons suggest that α-CTD may also enhance formation of the productive complex in B. subtilis by interacting with transcriptional activators.

Construction of bacterial strains

The bacterial strains and primers used in this study are listed in S4 and S5 Tables, respectively. The B. subtilis strains used in this study were derived from B. subtilis strain 168. Strain SMS01 (168 rpoA::cat-Pspac-rpoA^int-rplQ) was constructed by transformation of strain 168 with the DNA fragment encoding: the IPTG-inducible promoter (Pspac), the lacI and chloramphenicol resistance (cat) genes, flanked by sequences homologous to the 3’ region of rpsK with its downstream sequence and the 5’ region of the rpoA gene with its upstream sequence. The latter enabled the double-crossover recombination of the DNA fragment with the B. subtilis chromosome, thereby introducing the Pspac promoter into the intergenic region between rpsK and rpoA. The stepwise PCR-based procedures used to construct the recombinant strains are schematically shown in S12 Fig. We initially amplified various short DNA fragments as follows: the 3’ region of the rpsK gene with its downstream region was amplified from genomic DNA of strain 168 using primers SMP01 and SMP02 (fragment A); cat, lacI and the Pspac promoter were amplified from the genomic DNA of 168spacZ (S. Ishikawa, unpublished) using primers SMP03 and SMP04 (fragment B); and the 5’ region of the rpoA gene with its upstream region was amplified from strain 168 genomic DNA with primers SMP05 and SMP06 (fragment C). Fragment A–C were joined together to generate the DNA fragment that was used to transform strain 168 by recombinant PCR. After transformation, chloramphenicol-resistant colonies were selected to obtain SMS01. SMS02 (168 amyE::Pxyl-rpoA^int-rplQ-km) was similarly constructed by the transformation of strain 168 with a DNA fragment harboring the xylose-inducible promoter (Pxyl), the intact rpoA gene, the rplQ gene and the kanamycin resistance gene (km), flanked by sequences homologous to the 5’ region of the xylR gene and the 3’ region of the amyE gene. Double crossover recombination thus introduced the rpoA-rplQ genes along with the Pxyl promoter into the amyE locus of the B. subtilis chromosome (S12 Fig). The DNA Pxyl promoter was amplified from LY111 genomic DNA [61] using primers SMP07 and SMP08 (fragment D); full-length rpoA and rplQ were amplified from strain 168 using primers SMP09 and SMP12 (fragment E); the kanamycin-resistant fragment was obtained by PCR using plasmid pDG780 [62] and primers SMP13 and SMP14 (fragment F); and the amyE 3’ region was amplified from LY111 genomic DNA using primers SMP15 and SMP16 (fragment G). These four DNA fragments were concatenated by recombinant PCR, and the resulting DNA fragment was used to transform LY111 cells (S12 Fig). The DNA fragment used to construct strain SMS03 (168 amyE::Pxyl-rpoA^del-rplQ-km) was generated by recombinant PCR using DNA fragments D, F and G, plus additional fragments H and I, which were PCR amplified from strain 168 genomic DNA using primers SMP09 and SMP17, and SMP11 and SMP12, respectively (S12 Fig). The amplified DNA fragments carried a truncated rpoA gene consisting of 1–733 bp and 931–945 bp (the 3’ end) of rpoA, rplQ, and the intact intergenic region between rpoA and rplQ, all located downstream of the Pxyl promoter. The generated fragment was introduced into the amyE locus of the LY111 chromosome. The DNA fragment used to construct SMS04 (168 amyE::Pxyl-rplQ-km) consisted of fragments D, F and G, and additional fragment J, which carried rplQ and the intact intergenic region between rpoA and rplQ (including the termination codon of the rpoA gene). SMS04 was obtained by the transformation of LY111 with the amplified DNA fragment. Kanamycin-resistant colonies were selected to obtain strains SMS02, -03 and -04. These strains were then transformed with genomic DNA purified from SMS01 to construct SMS05 (168 rpoA::cat-Pspac-rpoA^int-rplQ, amyE::Pxyl-rpoA^int-rplQ-km), SMS06 (168 rpoA::cat-Pspac-rpoA^int-rplQ, amyE::Pxyl-rpoA^del-rplQ-km) and SMS07 (168 rpoA::cat-Pspac-rpoA^int-rplQ, amyE::Pxyl-rplQ-km), respectively. SMS05, SMS06 and SMS07 were transformed with genomic DNA purified from 168 rpoC-His, and erythromycin-resistant colonies were obtained and designated SMS08 (168 rpoA::cat-Pspac-rpoA^int-rplQ, amyE::Pxyl-rpoA^int-rplQ-km, rpoC::pMUTinHis-Em ΔrpoC), SMS09 (168 rpoA::cat-Pspac-rpoA^int-rplQ, amyE::Pxyl-rpoA^del-rplQ-km, rpoC::pMUTinHis-Em ΔrpoC), respectively. To construct SMS16 (168 amyE::Pxyl-rpoA^intHis-rplQ-tet) and SMS17 (168 amyE::Pxyl-rpoA^delHis-rplQ-tet), we initially constructed SMS14 (168 amyE::Pxyl-rpoA^int-km) and SMS15 (168 amyE::Pxyl-rpoA^del-km). For the construction of SMS14, we used fragments D, G, additional fragment K, which was amplified from strain 168 genomic DNA using primers SMP09 and SMP10, and fragment L, which was amplified from plasmid pDG780 using primers SMP27 and SMP14. These fragments were concatenated by recombinant PCR and transformed to LY111 cells to generate SMS14. For the construction of SMS15, we used fragments D, G, L, and additional fragment M, which was amplified from strain 168 genomic DNA using primers SMP09 and SMP17 (S12 Fig). For the construction of SMS16, we used fragment N, which included the 3’ region of the rpoA gene and was amplified from strain 168 genomic DNA using primers SMP28 and SMP29, fragment O, which encoded a 12 x Histidine tag and was amplified from strain 168 genomic DNA using primers SMP30 and SMP31, and fragment P, which carried the tetracycline resistance gene (tet) and the amyE 3’ region and was amplified from LY111 genomic DNA using primers SMP32 and SMP16. Strain SMS14 was transformed with the generated fragment and tetracycline was used to obtain SMS16. Similarly, strain SMS17 was constructed by the transformation of strain SMS15 with the DNA fragment consisting of fragment Q, which included the truncated 3’ region of the rpoA gene and was amplified from strain 168 genomic DNA using primers SMP33 and SMP34, fragment O and fragment P (S12 Fig). Transformants were selected based on tetracycline resistance. Finally, SMS18 (168 rpoA::cat-Pspac-rpoA^int-rplQ, amyE::Pxyl-rpoA^intHis-rplQ-tet) and SMS19 (168 rpoA::cat-Pspac-rpoA^int-rplQ, amyE::Pxyl-rpoA^delHis-rplQ-tet) were generated by transforming strains SMS16 and SMS17 with the chromosomal DNA of strain SMS01.

Growth conditions

All experiments were performed under the same culture conditions. B. subtilis cells were grown overnight at 37°C on LB plates containing 1 mM IPTG (to induce the expression of RpoA^int) and antibiotics, if needed. For pre-culture, B. subtilis cells were suspended in 50 mL of fresh LB^IPTG [Luria-Bertani (LB) medium containing 1 mM IPTG] at OD₆₀₀ ≈ 0.01 and grown to mid-log phase (OD600 ≈ 0.4) at 37°C under aerobic conditions. The cells were harvested by centrifugation (6000 rpm for 10 min) and washed twice with fresh LB medium. The cells were then resuspended in 400 mL of fresh LB^xyl (LB containing 1% xylose) at OD₆₀₀ ≈ 0.03, and experimental cultivations were performed at 37°C under aerobic conditions.

Pull-down purification of RNAP complexes

B. subtilis cells expressing Histidine-tagged RpoC (SMS08 cells expressing rpoA^int and SMS09 expressing rpoA^del) were grown in LB^xyl at 37°C under aerobic conditions, as described above. The cells were cultivated in LB^xyl, harvested by centrifugation hourly from 0 to 6 hours, and resuspended in 1.5 mL of binding buffer (0.1 M Tris-HCl, pH 7.5, 20% glycerol, 1 mM β-mercaptoethanol, 50 mM imidazole, 0.5 M NaCl, 1 mM PMSF), and sonicated for a total of 10 min (4-second pulses with 10-second intervals). The lysates were then centrifuged (8000 rpm for 10 min at 4°C), and the supernatants (taken as whole-cell extracts) were mixed with 50 μL of nickel-affinity beads (His-Tag Isolation & Pulldown, Invitrogen). The mixtures were incubated for 30 min at 4°C, and the beads were removed by a brief centrifugation and the use of a magnetic rack. After the beads were subjected to five washes with binding buffer, the immobilized RNAP complexes were eluted with 50 μL of elution buffer (0.1 M Tris-HCl, pH 7.5, 20% glycerol, 1 mM β-mercaptoethanol, 0.5 M imidazole, 0.1 M NaCl). The recovered RNAP complexes were resolved on a 5–20% SDS-PAGE gradient gel, and the results were visualized with the Flamingo (Bio-Rad) fluorescent stain. Gel densitometry was performed using the ImageJ software (NIH). The bands corresponding to RpoC-His, RpoA^int and RpoA^del were confirmed by western blotting using anti-His-tag and anti-RpoA antibodies.

Western blotting

Western blotting was performed as described previously [33]. The rabbit poly-clonal anti-RpoA antibody was generated against a mixture of chemically synthesized peptides corresponding to residues 143 to 159 (QRGRGYTPADANKRDDQ), 233 to 250 (HAEIMVEKEEDQKEKVLE), and 297 to 314 (EEVKAKLEELGLGLRKDD) of RpoA. The rabbit immunizations, serum preparation and antibody purification were performed by SIGMA Genosys (Japan). The poly-clonal anti-His-tag antibody and mono-clonal anti-RpoB antibody were purchased from MBL (Japan) and Neoclone (USA), respectively.

Transcriptome analysis

Total RNA was isolated from rpoA^int- and rpoA^del-expressing cells harvested hourly from 0 to 3 hours after the change of medium (from LB^IPTG to LB^xyl), cDNAs were prepared from each of these eight RNA samples (5 μg), and fragmented cDNAs were subjected to terminal labeling, as previously described [33]. Hybridization and Genechip scanning were performed using a Genechip B. subtilis genome array and the GCOS software (Affymetrix); an appropriate hybridization condition (at 45°C for 16 hours) and wash and staining protocol for the array, ProkGE-WS2, according to the manufacture’s instruction (Affymetrix). All experiments were performed in duplicate. The signal intensities obtained from all probes were normalized using the GCOS software, with the target signal intensity set to 500 (Affymetrix). To exclude genes with low expression levels, we selected the probes corresponding to the genes with adequate signal intensities (sum of the signal intensities in duplicate experiments, each involving two hybridizations (from dupricate experiments) with SMS08 cells and two with SMS09 cells > 400). The genes that were up- and down-regulated in SMS09 cells versus SMS08 cells were those having relative signal intensity ratios < 0.25 or > 4 and q < 0.2 by Welch’s t-test. The raw transcriptome data have been deposited in Arrayexpress under accession, E-MTAB-3383.

ChAP-chip analysis

ChAP-chip analysis was performed as previously described [33]. Briefly, after the medium was changed from LB^IPTG to 400 ml LB^xyl, cells were grown with aeration at 37°C for 3 hours, and then incubated with formaldehyde (final concentration, 1%) for 30 min at 37°C to crosslink the RNAP complex to the chromosomal DNA. The cells were then pelleted by centrifugation and resuspended in 3 ml of UT buffer (100 mM HEPES, 50 mM imidazole, 8 M urea, 0.5 M NaCl, 1% Triton X-100, and 10 mM β-mercaptoethanol, pH 7.4). The affinity purification of the RNAP-DNA complexes, purification of DNA from these RNAP-DNA complexes and original Whole-cell extract, and amplification of purified DNA for labeling were all performed as previously described [33]. A custom Affymetrix tiling array for B. subtilis [33] was used, and signal intensities were acquired using the GCOS software in CEL format. The In Silico Molecular Cloning Program, Array Edition (In Silico Biology, Japan) was used to determine the binding intensities and visualize the binding profiles of RNAP on the B. subtilis chromosome. The signal intensities of DNA affinity-purified with RNAP (ChAP DNA) and purified from whole-cell extracts (control DNA) were adjusted to confer a signal average of 500 for normalization, and the RNAP (RpoC and RpoA) binding intensities of all probes were determined based on their enrichment of DNA fragments by the affinity purification of RNAP. This was calculated by dividing the signal intensities of ChAP DNA by those of control DNA [33]. All experiments were performed in duplicate. The raw ChAP-chip data have been deposited in Arrayexpress under accession numbers, E-MTAB-3384 and -3385.