SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The WalRK (YycFG) two-component system co-ordinates cell wall metabolism with growth by regulating expression of autolysins and proteins that modulate autolysin activity. Here we extend its role in cell wall metabolism by showing that WalR binds to 22 chromosomal loci in vivo. Among the newly identified genes of the WalRK bindome are those that encode the wall-associated protein WapA, the penicillin binding proteins PbpH and Pbp5, the minor teichoic acid synthetic enzymes GgaAB and the regulators σI RsgI. The putative WalR binding sequence at many newly identified binding loci deviates from the previously defined consensus. Moreover, expression of many newly identified operons is controlled by multiple regulators. An unusual feature is that WalR binds to an extended DNA region spanning multiple open reading frames at some loci. WalRK directly activates expression of the sigIrsgI operon from a newly identified σA promoter and represses expression from the previously identified σI promoter. We propose that this regulatory link between WalRK and σI RsgI expression ensures that the endopeptidase requirement (CwlO or LytE) for cell viability is fulfilled during growth and under stress conditions. Thus the WalRK and σI RsgI regulatory systems cooperate to control cell wall metabolism in growing and stressed cells.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The cell wall of Gram-positive bacteria is a macromolecular mesh-type sacculus surrounding the cell membrane that is composed of peptidoglycan and teichoic acid in approximately equal amounts. It determines bacterial cell shape, protects the bacterium against internal osmotic pressure and environmental insult and provides a platform for exposure of surface proteins. The cell wall is a dynamic structure that is continuously remodelled by synthesis and turnover during growth, septation and cell separation. Therefore, balancing synthesis and turnover is essential for cell wall integrity and for cell viability.

Peptidoglycan synthesis in Bacillus subtilis occurs in a dispersed pattern along the cell cylinder in growing cells and at the septum in dividing cells (Stewart, 2005; Carballido-López and Formstone, 2007; den Blaauwen et al., 2008). Studies show there to be separate synthetic machineries at each location. Septal synthesis is carried out by the divisome, a protein complex whose assembly is initiated by a ring of FtsZ at the mid-cell division site while cylindrical synthesis is carried out by a protein complex associated with actin-like proteins MreB, MreBH and Mbl. Each complex has a distinctive spectrum of associated penicillin binding proteins (PBP) and autolysins for peptidoglycan synthesis and remodelling (Daniel and Errington, 2003; Carballido-López and Formstone, 2007; den Blaauwen et al., 2008). Peptidoglycan expansion is effected through cleavage of the glycan strands and cross-linking peptides by an array of murolytic enzymes thereby generating novel substrates for insertion of new wall material by the synthetic activities of PBPs (Smith et al., 2000; Vollmer et al., 2008). It is interesting that the extracellularly located autolysins and PBPs are the most redundant activities of the peptidoglycan synthetic complex: for example B. subtilis encodes 35 murolytic enzymes and 16 PBPs (Smith et al., 2000; Scheffers et al., 2004; Sauvage et al., 2008; Vollmer et al., 2008). Autolysin and PBP expression is highly regulated so that different physiological states display a characteristic spectrum of each activity (Scheffers et al., 2004; Vollmer et al., 2008; Botella et al., 2011). The regulatory complexity of peptidoglycan metabolism stems from the requirement to balance synthesis and turnover in a manner that is responsive to the prevailing environmental and nutritional conditions. To achieve this, peptidoglycan precursors must be delivered in sufficient quantities to two synthetic machineries at different cellular locations where the existing cell wall material has been cleaved to produce novel substrates at a level adequate for synthesis but avoiding cell lysis.

The WalRK (YycFG) two-component signal transduction system (TCS) plays a central role in co-ordinating peptidoglycan synthesis with cell growth and division in B. subtilis (Dubrac et al., 2008; Devine, 2012). Unusually for a TCS, WalRK expression is essential for viability, with depletion leading to cell lysis (Fabret and Hoch, 1998). WalRK activates expression of autolysins (YocH, CwlO, LytE) and a cell-wall-associated protein (YdjM) while repressing modulators of autolysin activity (IseA, PdaC) in response to a signal that emanates from the divisome (Howell et al., 2003; Bisicchia et al., 2007; Fukushima et al., 2008; 2011; Yamamoto et al., 2008; Hashimoto et al., 2012). The model of WalRK function proposes that it co-ordinates cylindrical wall synthesis with cell division: under conditions conducive for growth it directs expression of enzymes required for cylindrical cell wall synthesis: under the obverse conditions, these enzymes are not expressed and extant enzyme activity is inhibited by derepression of autolysin activity inhibitors (IseA) and modulators (PdaC) (Bisicchia et al., 2007; Dubrac et al., 2008; Fukushima et al., 2008; Kobayashi et al., 2012).

It is now well established that B. subtilis cells require the endopeptidase activity of one or other of the CwlO or LytE autolysins for viability (Bisicchia et al., 2007; Hashimoto et al., 2012). The short cell phenotype seen upon cellular depletion of both autolysins signifies that this endopeptidase activity is required for cylindrical wall synthesis (Bisicchia et al., 2007; Hashimoto et al., 2012). A recent study shows a similar requirement for the activity of one of three redundant DD-endopeptidases (Spr, YdhO, yebA) for peptidoglycan synthesis in Escherichia coli (Singh et al., 2012; Vollmer, 2012). CwlO expression is absolutely dependent on WalRK activation with WalR∼P binding to cognate recognition sequences positioned in the vicinity of the −35 position of a σA promoter (Bisicchia et al., 2007). The lytE gene has both σA and σI promoters whose activity is regulated by the WalRK TCS and the σI RsgI sigma factor/antisigma factor regulatory systems respectively (Bisicchia et al., 2007; Tseng et al., 2011). The activity of σI is induced by heat stress with LytE expression playing an important role in cell survival under this condition. An unusual feature of the σI RsgI-mediated control of lytE expression is that an initiation point of transcription could not be identified for the σI promoter (Tseng et al., 2011). This led the authors to suggest that σI, acting either alone or in an RNA polymerase holoenzyme complex, may function as a co-activator of expression in cooperation with the WalR response regulator, directing lytE expression from the WalRK regulated σA promoter (Tseng et al., 2011). In view of these observations on CwlO and LytE expression, it is important to establish how the requirement for endopeptidase activity is fulfilled under different conditions. WalRK is activated primarily in exponentially growing cells, hence CwlO is expressed only under this condition (Howell et al., 2006; Bisicchia et al., 2007). It remains to be established therefore how the requirement for endopeptidase (CwlO or LytE) activity is fulfilled when WalRK is not activated, e.g. during stationary phase or under stress conditions (Bisicchia et al., 2007).

The known WalRK regulon of B. subtilis was determined by experimental approaches that identify differentially transcribed genes under conditions of WalRK depletion or modification (Fukuchi et al., 2000; Howell et al., 2003; Bisicchia et al., 2007; Dubrac et al., 2008; Devine, 2012). It is difficult to assess how comprehensive this regulon is because of technical difficulties associated with investigating essential genes. For this reason, we performed a ChIP on chip analysis using high-density arrays, to identify and delimit the chromosomal regions to which WalR∼P binds in vivo. We show that WalR binds to 22 chromosomal regions in vivo, extending its role in the regulation of cell wall metabolism. WalRK also regulates expression of the σI RsgI regulatory system that controls expression of cell-wall-associated genes (MreBH, lytE, bcrC) under stress conditions. Together our results show how the requirement for viability of one or other of the CwlO or LytE endopeptidase-type autolysins is fulfilled under growth and stress conditions by the combined and cooperative activities of the WalRK and σI RsgI regulators.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The WalR bindome determined by ChIP on chip analysis

A ChIP on chip analysis using high-density arrays was performed to establish the chromosomal regions to which WalR binds in vivo. Samples were harvested from exponentially growing and phosphate limited cells, conditions under which WalR binding to DNA should be maximal and greatly reduced respectively. Our results show that WalR binds specifically to 22 regions of the B. subtilis chromosome, a majority of which (14 of the 22 sites) have a known involvement in cell wall metabolism (Table 1, Figs 1 and S1). The threshold of biologically relevant chromosomal binding was set at lytE because of the independent body of evidence that shows its expression to be controlled by WalRK (Bisicchia et al., 2007). Confidence in the biological relevance of these 22 sites is supported by the following controls: (i) growth and gene expression in the strain expressing the FLAG-tagged WalR protein are normal, (ii) the specificity of immunoprecipitating only WalR-bound DNA fragments is confirmed by the absence of peaks at these chromosomal loci when WalR is not FLAG-tagged, (iii) binding to these loci occurs mainly in exponentially growing cells and is greatly reduced or absent in phosphate limited cells (with the exceptions noted below) consistent with previous studies on WalRK activation (Table 1, Figs 1 and S1), and (iv) all known WalR regulon genes are included in the WalR bindome (Table 1, Fig. S1A). We conclude that these 22 sites constitute the complete biologically relevant WalR bindome.

figure

Figure 1. WalR binding in vivo to B. subtilis chromosomal DNA. Binding of WalR to chromosomal DNA in vivo was determined by ChIP on chip analysis using a tiled array. Chromosomal loci are named in each panel. The genetic organization of the region is represented by blue arrows. The light blue vertical lines represent 100 bases of DNA. The intensity (peak height) and extent (peak length) of WalR binding is represented by the red peak whose position corresponds to that region of the chromosomal locus (blue arrows). Four binding profiles are shown in each panel: Profile 1: WalR binding to the (+) strand in exponentially growing cells. Profile 2: WalR binding to the (−) strand in exponentially growing cells. Profile 3: WalR binding to the (+) strand in phosphate limited cells. Profile 4: WalR binding to the (−) strand in phosphate limited cells. The binding peaks are assigned to the following promoters: (A) wapA; (B) sigIrsgI; (C) pbpH; (D) oppA; (E) yybNMLKIJ; (F) yfmC.

Download figure to PowerPoint

Table 1. WalR binding to chromosomal DNA in vivo determined by ChIP on chip analysis
 Gene(s)ScoreaMax (log2)bSize (bp)cScoreaMax (log2)bSize (bp)cRegulationFunction
Exponential growthPhosphate limitation
  1. The asterisk (*) signifies genes previously identified as members of the WalRK regulon; slash (/) signifies the genes are divergently transcribed; omega (Ω) signifies the genes are convergently transcribed; hyphen (-) signifies the genes are contiguous; number symbol (#) indicates that there are a number of peaks and that the values given are those for the peak associated with the gene outside the parenthesis. [n] signifies the number of structural genes covered by WalR∼P. Gene functions are those ascribed in the Subtiwiki and PubMed databases. Calculation of the score and the maximum level are described in Experimental procedures.

  2. a

    Score is the sum of the signals (log ratios) of each probe within the designated peak.

  3. b

    Max (log2) is the highest measure of binding signal within the peak.

  4. c

    Size (bp) the length of chromosomal DNA covered by the probes that have a signal (log ratios).

 1iseA* (yoeB)409.92.158332201.853862WalRKInhibitor of autolysin activity
 2pdaC* (yjeA)322.82.154973181.31.72720WalRKPeptidoglycan deacetylase
 3wapA144.21.41271087.11.162200YvrGHb, DegSUWall-associated protein
 4cwlO*69.01.17162516.80.621113WalRKAutolysin (endopeptidase)
 5yybN-yybJ [5]50.50.623257    ABC transporter
 6yfiY / yfiZ31.70.85111324.70.731144FurABC transporter for siderophores (schizokinen & arthrobactin)
 7pbpH / ykwD24.40.73111925.80.761037AbrBPenicillin binding protein H (PbpH)
 8sigI-rsgI13.90.641073   RsgISigma factor I (SigI) and regulation of SigI activity (RsgI)
 9ydjM*-ydjN11.40.541067   WalRKExpansin-like (YdjM)
10oppA10.80.561048   ScoC, TnrAOligopeptide transport
11ykvT*10.20.55904   WalRKAutolysin (YkvT)
12yocH*9.60.6770   WalRK, PrkCAutolysin (transglycosylase)
13yyd(BC)D# [3]7.50.491298     
14ggaA(B)# [2]6.80.471043    Biosynthesis of minor teichoic acid
15cyeA / yoaU4.10.46529    Putative cysteine permease (CyeA); LysR-type regulator YoaU
16yyd(FGH)I-yyzN# [6]3.90.421013   AbrB, RoKControl of LiaRS activity
17btr (ybbB)3.60.45817   FurTranscriptional activator (AraC family)
18dacA1.80.43521    Pbp5 (carboxypeptidase) (DacA)
19lytE*1.00.4461   WalRK σIAutolysin (endopeptidase) (LytE)
20yfmB-yfmC   24.90.741026FurIron-citrate ABC transporter (binding protein) (YfmC)
21cspB / yhcJ   9.20.53742 CspB, major cold shock protein (RNA chaperone)
22kbaA Ω pdaB (ybaN)   4.40.52443σE of pdaBinner membrane protein involved in KinB activation (KbaA); polysaccharide deacetylase involved in sporulation (PdaB)

These analyses extend the role of WalRK in regulating cell wall metabolism. Among the newly identified genes that encode proteins with a cell wall function are wapA (Wall Associated Protein A), pbpH (penicillin binding protein H), dacA (penicillin binding protein 5) and ggaAB (minor teichoic acid synthetic enzymes). WalRK also binds to the promoter region of the sigIrsgI operon that encodes the stress responsive σI RsgI sigma/anti-sigma factor regulatory system (Tseng et al., 2011). WalR binding is higher in phosphate-limited cells than in exponentially growing cells at only three loci (yfmBC, cspB/yhcJ and kbaB/pdaB), one of which encodes the major cold shock protein (CspB) thereby reinforcing the association with cell stress (Table 1, Figs 1F and S1B). Interestingly, neither transcriptome nor bindome studies show an association with ftsAZ as reported by Fukuchi et al. (2000). A further notable feature is that the promoter regions of some newly identified bindome members (e.g. pbpH and oppA) encode non-coding RNAs (Fig. S1B, C and Table 1, Irnov et al., 2010).

A majority of the WalR DNA binding profiles are typical of that expected for binding of a conventional transcriptional regulator to its promoter (e.g. wapA, sigI, oppA, Fig. 1A, B and D). A notable feature is the significantly greater peak height (maximum fold enrichment) and width (the extent of the DNA region immunoprecipitated) for genes that are negatively regulated (e.g. iseA, pdaC) than for those positively regulated (e.g. yocH, cwlO) by WalRK (Fig. S1A). Moreover, WalR binding to some genes is observed in phosphate-limited cells (e.g. iseA, pdaC, wapA, cwlO), albeit with a reduced DNA footprint when compared with that of exponentially growing cells.

A different binding profile is found at three chromosomal loci (yydFGHIJ, 5 genes spanning 4.6 kilobases; yybNMLKJ, 5 genes spanning ∼ 3.3 kilobases and ggaAB, 2 genes spanning 4.5 kilobases) where WalR binds to an extended DNA region encompassing several structural genes (Figs 1E and S1). The biological relevance of such binding is supported by the observation that it is very greatly reduced, or abolished in phosphate limited cells (Table 1). Notably, the yydFGHIJ operon encodes a peptide (YydF) that induces the activity of the LiaRS TCS when modified (Butcher et al., 2007), again showing the relationship between WalRK and stress.

Features of WalR binding to DNA

We performed a bioinformatic analysis on the complete set of WalR bindome promoters (i.e. excluding the loci showing extended binding) to assess why only five were identified in transcriptome studies (Howell et al., 2003; Bisicchia et al., 2007). Twenty-nine potential WalR binding sites were identified in 12 newly identified promoters. A WalR DNA binding consensus derived from the complete set of binding sites is very similar to that previously established, with the TGT bases of each six base pair direct repeat being highly conserved (Fig. 2). However, the spacing between the six base pair direct repeats differs in genes identified by transcriptome and bindome studies. The 5 base pair spacing was found in 8 of the 10 potential binding sites associated with genes identified by transcriptome analysis, but only in 7 of the 29 potential binding sites associated with the genes newly identified by ChIP on chip analysis. Of the remaining 22 sites, seven have a 3 base pair spacing, two have a 4 base pair spacing, eleven have a 6 base pair spacing and two have a 7 base pair spacing (Fig. 2). This suggests that WalR binding to the newly identified sites differs in some way, perhaps functioning to modulate rather than directly activate or repress gene expression. In this context, it is interesting to note that many of the newly identified promoters are controlled by other regulators including TCSs (YvrGHb, PhoPR, DegSU), alternative sigma factors (σI) and transition state (AbrB) and other regulators (Table 1).

figure

Figure 2. A graphic depiction of the WalR DNA binding consensus. Putative WalR binding sequences were identified in the promoter regions of the members of the bindome: 29 from promoters of newly identified genes in the bindome and 10 from the constituent genes of the previously established transcriptome. The graphic was prepared by WebLogo. The frequency of each spacer length (3–7 nt) in the 39 WalR binding sequences is presented between the direct repeats in bar-graph format.

Download figure to PowerPoint

WalR does not bind to PhoP DNA recognition sequences in vivo

None of the PhoPR regulon genes are members of the WalRK bindome, despite the similarity in the WalR and PhoP consensus DNA binding sequences (Bisicchia et al., 2010). The tagAB promoter region has two almost consensus WalR binding sites while all other PhoP regulated promoters have at least one WalR binding sequence, albeit in inverse complement orientation relative to the initiation point of transcription (Bisicchia et al., 2010). To investigate the specificity of WalR binding, we constructed hybrid ydjM, yocH and cwlO promoters in which the natural WalR binding sequences are precisely replaced by a DNA segment containing PhoP binding sequences from the tuaA promoter in reversed orientation (Fig. S2). However, none of these hybrid promoters showed any activity (data not shown). We conclude that additional specificity determinants that distinguish WalR and PhoP DNA binding sites in vivo remain to be established.

WalR control of some newly identified bindome members

WalR binding to, and transcriptional regulation of, selected newly identified promoters was assessed by electrophoretic mobility shift assays (EMSA) and by quantification of RNA transcripts (by qRT PCR) in cells expressing and not expressing WalRK. EMSA assays were performed on seven WalR target regions (yfiY/Z, wapA, oppA, pbpH, sigI, yybN, yydI). Mobility shifts were observed for wapA, oppA, pbpH, yybN (Fig. 3) and sigI (Fig. 5B) but not for the yfiY/Z and yydI promoter regions (data not shown). The effect of WalRK binding on the expression of 14 newly identified genes was assessed in strain AH9912 (PspacwalRKHIJ) by quantifying their transcript level in cells expressing (+IPTG) and not expressing (−IPTG) the WalRK TCS. With the exception of the wapA and sigIrsgI operons (see below), these genes showed little difference in transcript levels under the two conditions. In summary these analyses support WalR binding to many of the newly identified genes. However, WalRK does not appear to be the primary transcriptional regulator for a majority of the newly identified bindome members under the conditions tested, consistent with their not being identified in previous transcriptome analyses (Howell et al., 2003; Bisicchia et al., 2007).

figure

Figure 3. Electrophoretic mobility shift assays (EMSA) of selected newly identified WalR binding targets. The identity of the promoter-containing DNA fragments for each gene tested is shown above each panel. Each reaction contained 2 ng of a biotin-labelled fragment of promoter DNA, 0.5 μg of non-specific herring sperm DNA and 4 μg of BSA. The non-migrating bands in the pbpH and yybN lanes are probably biotinylated single stranded DNA generated by PCR with an excess of the biotinylated primer to which WalR will not bind (Albano et al., 2005). The concentration of phosphorylated WalR used in each reaction is indicated below the figure (0–2 μM).

Download figure to PowerPoint

WalRK-mediated control of wapA

WapA is a large (∼ 258 kDa) cell-wall-associated protein of unknown function, that contains distinct repeats at both the amino- and carboxy-terminal regions and that is present in only some Bacilli. Its expression is regulated by the YvrGHb and DegSU TCS (Dartois et al., 1998; Serizawa et al., 2005). All the regulatory elements controlling wapA expression are located within 75 base pairs of the initiation point of transcription (Fig. 4A, data not shown). The ChIP on chip analysis shows that WalR binds to the wapA promoter in exponentially growing and phosphate limited cells with a higher score under both conditions than that of the genes (yocH, cwlO, ydjM, lytE) whose expression is known to be activated by WalRK (Fig. 1, Table 1). There is a putative WalR recognition sequence positioned within the 17 base pair spacer region of the σA promoter: both direct repeats deviate from the highly conserved TGT at one position although they are separated by the canonical 5 bp (Fig. 4A). To establish whether the observed WalR binding has an effect on expression, we measured wapA transcripts in cells expressing different levels of WalRK in exponentially growing and phosphate limited cells. This approach was chosen because mutation of the WalR boxes embedded within the σA promoter might damage its activity leading to difficulties in interpreting the results. Cells of strain AH9912 (PspacwalRKHIJ) were grown through a phosphate limitation cycle in LPDM (low phosphate-defined medium) without IPTG addition (white column, WalRK depleted), containing 75 μM IPTG (grey column, the level of WalRK expression required for normal growth, Howell et al., 2003) and 1 mM IPTG (black column, WalRK overexpression). As a control we confirmed that walRK expression is inducible by IPTG: walR transcript levels vary ∼ 3–7-fold in exponentially growing (T−0.5) and phosphate limited (T+0.5, T+1) cells depending on the level of inducer added (compare white, grey and black bars Fig. 4B). In contrast, there is an inverse relationship between the cellular level of wapA transcript and the amount of IPTG inducer added, showing that WalR represses its expression consistent with the location of the proposed WalR binding sequence (Fig. 4A and C). Furthermore, the effect of WalR on expression appears to differ in exponentially growing and phosphate limited cells. Only an approximate twofold difference in wapA transcript is observed in cells growing exponentially with zero and 1 mM IPTG added perhaps explaining why this gene was not identified in previous transcriptome analysis (Fig. 4C, Bisicchia et al., 2007). However, there is up to a fivefold difference in wapA transcript levels in phosphate limited cells growing under similar conditions (Fig. 4C). This is the first report of a role for WalRK in non-exponentially growing cells. We conclude that WalR represses wapA expression, supporting the view that the ChIP on chip analysis accurately reflects in vivo WalRK activity.

figure

Figure 4. Control of wapA expression by WalRK. The intergenic region containing the wapA promoter is shown (A). The motifs of the SigA promoter identified by Dartois et al. (1998) are shown in bold and the initiation point of transcription is indicated with a bent arrow. The wapA ribosome binding site and initiation codon (ATG) are also shown in bold. The putative WalR binding sequence is shown in underlined large capitals.

B. Measurement of walR transcript level in strain AH9912 (PspacwalRKHIJ) by quantitative RT-PCR.

C. Measurement of wapA transcript level in strain AH9912 (PspacwalRKHIJ) by quantitative RT-PCR. Strain AH9912 was grown in LPDM in the absence of IPTG (white bars), in the presence of 75 μM IPTG (grey bars) and in the presence of 1 mM IPTG (black bars). The level of transcript was normalized relative to the level of transcript for the +IPTG sample at time point T1, which was assigned a value of 1 (data not shown). T0 is the point at which cells enter the phosphate limited state.

Download figure to PowerPoint

WalRK-mediated control of sigIrsgI operon expression

Binding of WalR to the sigIrsgI promoter region merited further investigation since this regulatory system controls expression of proteins (LytE, MreBH, BcrC) involved in cell wall metabolism. There is a σI-type promoter (Fig. 5A, −35 and −10 motifs shown in bold capitals) in the control region of the sigIrsgI operon, identified by Tseng and Shaw (2008). However, we have established by 5′ RACE that this region also contains a σA-type promoter positioned between the σI promoter and the sigI open reading frame (Fig. 5A, −35 and −10 motifs shown in bold capitals). In addition there are two potential WalR binding motifs (capitals underlined) located overlapping and downstream of the initiation point of transcription of the σI promoter and 28 base pairs upstream of the −35 region of the σA promoter (Fig. 5A). Such a location suggests that WalR represses expression of the σI promoter. Both WalR binding sites were mutated (mutated sequences in capital italics beneath the WalR boxes) to assess their role in controlling expression of the sigIrsgI operon. EMSA analysis shows that mutating the WalR boxes significantly attenuates WalR∼P binding to the promoter fragment (Fig. 5B). Analysis of expression (by qRT PCR) shows that, despite their location relative to the σA promoter (28 bp upstream), mutating the WalR boxes causes a 2–3-fold decrease in the level of sigIrsgI transcript in exponentially growing cells indicating that WalR∼P binding stimulates its expression under these conditions (Fig. 5C). To assess the effect of WalR∼P binding on σI promoter activity, we constructed a strain in which the σI promoter is constitutively active (ΔrsgI::kan) and WalRK expression is controlled by the IPTG inducible Pspac promoter (strain LSB279 rsgI::kan PspacwalRKyycHIJ). We distinguished between the activities of the σA and σI promoters by using primers located within the sigI open reading frame (cumulative transcript from both promoters) and upstream of the σA promoter (transcript from the σI promoter only) for qRT PCR. We confirmed that walRK expression is inducible in strain LSB279: walR transcripts differ ∼ 10-fold in cells growing without (white boxes) and with (black boxes) IPTG (Fig. 6A). Our results show an inverse relationship between the level of walRK expression and the activity of the constitutive σI promoter (Fig. 6). When walRK expression is low (Fig. 6A, white boxes) the level of sigIrsgI transcript directed from both the σA and σI promoters (white boxes, Fig. 6B) and from the σI promoter only (white boxes, Fig. 6C) is elevated. When walRK expression is high (Fig. 6A, black boxes) transcripts directed from both the σA and σI promoters (Fig. 6B, black boxes) and the σI promoter only (Fig. 6C, black boxes) is reduced. Thus we conclude that WalR∼P represses the σI promoter activity. These data show that WalR∼P activates the σA promoter and inhibits σI promoter of the sigIrsgI operon in exponentially growing cells.

figure

Figure 5. Control of sigIrsgI expression by WalRK.

A. The DNA sequence of the sigI promoter region: the consensus −35 and −10 regions of the σI and σA promoters are shown in bold, the initiation points of transcription by bent arrows and the ribosome binding site (RBS) and the initiation codon (GTG) in bold. The WalR binding boxes are underlined in big bold capitals and the sequence to which each motif is mutated is indicated in bold capital italics below each WalR motif.

B. Electrophoretic mobility shift assay for sigI promoter region: sigI, the DNA fragment containing the wild-type WalR boxes; sigI*, the DNA fragment containing the mutated WalR boxes and htrA, the negative control. Each reaction contained 2 ng of biotin-labelled promoter DNA fragment. The concentration of phosphorylated WalR used in each reaction is indicated below the figure (0–2 μM).

C. The levels of sigI transcripts were determined by qRT PCR in wild-type cells (black bars) and in cells of strain LSB126 in which the WalR binding sequences are mutated as shown in (A). Cells were grown in LPDM and the point of phosphate limitation is designated T0. Transcript levels are normalized relative to the level of transcript for the wild-type sample at time point T1.5, which is assigned a value of 1.

Download figure to PowerPoint

figure

Figure 6. Measurement of sigI transcript levels in strain LSB279 (ΔrsgI::kan PspacwalRKHIJ) by quantitative RT-PCR. Strain LSB279 was grown in LPDM in the absence (white column) or the presence of 75 μM IPTG (black column): (A) Transcript levels of walR; (B) Transcript levels of sigI measured using primers located within the open reading frame (cumulative transcript from the σI and σA promoters); (C) Transcript levels of sigI transcript measured with primers located upstream of σA promoter (i.e. sigI transcript directed from the σI promoter only). The level of transcript was normalized relative to the level of transcript of the −IPTG samples at time point T1.5, which was assigned a value of 1. T0 is the point at which cells enter the phosphate limited state.

Download figure to PowerPoint

Null mutations in cwlO and sigIrsgI are synthetically lethal

The control of sigIrsgI expression by WalRK is interesting when viewed in the context of the requirement for viability of either the CwlO or LytE endopeptidase activities (Bisicchia et al., 2007; Hashimoto et al., 2012). CwlO expression is completely dependent on WalRK activation whereas LytE expression is controlled both by WalRK and SigI. Therefore, we sought to establish the relative contributions of WalRK and σI RsgI to the control of CwlO and LytE expression by establishing the frequency with which single and double mutant strains can be obtained. Strains singly mutated in any of the three loci (sigIrsgI, cwlO, lytE) are readily obtained as previously reported (Margot et al., 1998; Zuber et al., 2001; Yamaguchi et al., 2004; Bisicchia et al., 2007). However, we consistently obtain null lytE mutants at an approximately fivefold higher frequency (∼ 105 μg−1 DNA) than null cwlO mutants (∼ 2 × 104 μg−1 DNA). Strains in which a null lytE mutation is combined with a deletion of either sigIrsgI or of rsgI only are also are readily obtained, and with equal transformation frequencies (∼ 105 μg−1 DNA). Thus WalRK-mediated control of cwlO expression is sufficient for viability as expected. No transformants were ever obtained in repeated attempts (> 5) to combine a null mutation in cwlO with the sigIrsgI deletion showing that WalRK-mediated control of lytE expression is insufficient for viability in the absence of CwlO. However, a strain containing a cwlO null mutation and a deletion of rsgI only [LSB195 (ΔcwlO ΔrsgI) in which σI is constitutively active] is viable and is obtained at a high transformation frequency (∼ 2 × 104 μg−1 DNA) showing that SigI must be active under conditions where WalRK is not activated (i.e. no CwlO produced). Together these results show B. subtilis cell viability requires that either the WalRK or the σI sigma factor be active.

To establish whether the viability of strain LSB195 (ΔcwlO ΔrsgI) is due to σI-directed expression of lytE only, or whether expression of other SigI regulon members (e.g. mreBH or bcrC) contribute, we sought to combine a null cwlO mutation with a deletion in sigIrsgI in a strain where lytE expression is controlled by a xylose-inducible promoter (PxylA). Cells combining null mutations in cwlO and sigIrsgI are viable, and such a strain is obtained at high frequency (∼ 1 × 104 μg−1 DNA), when lytE is expressed (presence of xylose inducer) but no transformants are obtained when lytE is not expressed (absence of xylose inducer). This shows that only σI-directed expression of lytE is required for viability in the absence of cwlO. In summary, these data show that cell viability of B. subtilis requires that either WalRK or SigI must be active to fulfil the requirement for expression of CwlO or LytE. WalRK-directed expression of cwlO is sufficient for cell viability. However, when WalRK is not active, σI activation must occur to direct LytE expression. Therefore, the control of sigIrsgI expression by WalRK shown in the previous section provides a link between these two regulators to ensure that adequate levels of CwlO or LytE are produced to maintain cell viability under different environmental conditions.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Cell envelope metabolism is the predominant theme of the WalRK bindome

This study extends the role of WalRK in B. subtilis showing it to be part of a regulatory network that makes cell wall synthesis responsive to the prevailing growth and environmental conditions. WalR binds to 22 chromosomal loci in vivo (the WalRK bindome), significantly more than were previously identified (seven genes) by in silico and transcriptome analysis (Howell et al., 2003; Bisicchia et al., 2007). Cell wall metabolism and cell envelope stress are the predominant themes of the WalRK bindome with 14 of the 22 members known to be engaged in one or other of these processes. Newly identified loci involved in cell wall metabolism include wapA that encodes the major wall-associated protein WapA, pbpH and dacA that encode PBPs PbpH and Pbp5 and ggaAB that encodes the GgaAB enzymes for synthesis of the minor teichoic acid. The link with cell envelope stress is revealed by the inclusion of yydFGHIJyyzN and sigIrsgI in the WalRK bindome. The yydFGHIJ operon encodes the YydF peptide, the enzyme to modify it (YydG), a protease (YydH) and an ABC-type transporter (YydIJ). The modified YydF peptide induces the activity of the LiaRS TCS that is also responsive to stress induced by cell-wall-associated antibiotics (Mascher et al., 2004; Jordan et al., 2006; Butcher et al., 2007). The σI RsgI regulatory system is induced by heat stress and phosphate limitation (Tseng et al., 2011; L.I. Salzberg, E. Botella and K.M. Devine, unpubl. results). SigI activity is controlled by RsgI, an antisigma factor that spans the cell membrane and is responsive to the stress stimulus. Furthermore, by regulating σI RsgI expression, WalRK indirectly controls expression of the actin-like protein MreBH, the LytE autolysin and the Lipid I recycling enzyme BcrC, all of which execute central roles in cell wall metabolism. Other notable features of the WalR bindome are its association with iron metabolism, three bindome members (yfiYZ, btr, yfmBC) are controlled by Fur, and the location of non-coding RNAs in the oppA and pbpH promoter regions that may confer more global regulatory effects on WalR function (Irnov et al., 2010).

Integration of WalRK function with other transcriptional regulators

This study shows that viewing WalRK as the primary transcriptional regulator of only five genes (yocH, cwlO, ydjM, iseA, pdaC) in exponentially growing cells, is an incomplete picture of its function in vivo. WalRK clearly functions under other conditions in cooperation with other regulators. Eleven of the 22 bindome members, and nine of the newly identified loci, are regulated by additional regulators (Table 1) including two-component systems (YvrGHb, PhoPR, DegSU), transition state regulators (AbrB, Rok), sigma factors (σI) and other miscellaneous transcription factors (Fur, ScoC, TnrA). Thus WalRK function is likely to be integrated with, and probably affected by, the activities of other regulators and may be manifest only under certain conditions. For example, WalR binds to the wapA promoter in both exponentially growing and phosphate limited cells (Fig. 1A). However, WalRK-mediated control of wapA expression is more evident in phosphate limited cells than in cells growing exponentially (Fig. 4C). This may be due to the fact that control by the YvrGHb TCS predominates in exponentially growing cells (Serizawa et al., 2005). In addition, while WalRK binds both the lytE and sigIrsgI promoters, their expression is also regulated by the σI sigma factor whose activity is responsive to stress (this point is expanded below). Moreover, WalRK binds to three loci (yfmC, cspB and kbaA) at a higher level in phosphate limited cells that in cells growing exponentially. Together these observations indicate that WalRK function is integrated with the activity of other transcriptional regulators, modulating gene expression under a wider variety of conditions than previously thought.

In a more general context, this ChIP on chip analysis suggests that the cellular roles of many transcriptional regulators may be underestimated. The constituent gene cohort of most regulons is assigned by differential transcription in the presence and absence of the regulator. However, our study suggests that such analysis may identify only those genes for which the transcription factor is the primary transcriptional regulator and may not identify genes that are co-regulated, either directly or indirectly, by the regulator under study.

The DNA sequence requirements for WalR binding and transcriptional regulation in vivo are complex

Significant complexity is evident in the DNA sequence requirements for WalR binding and transcriptional activity in vivo. The DNA binding consensus deduced from transcriptome analysis shows the TGT motifs to be highly conserved (Fig. 2, Bisicchia et al., 2007; Dubrac et al., 2008). Putative WalR binding motifs were identified in the promoter regions of all newly identified bindome members but they are distinguished by increased deviation from the consensus, especially in the second six base pair repeat, and show significantly more variance (3–7 base pairs) in the spacer region (Fig. 2). An intriguing feature is that negatively controlled genes (e.g. iseA, pdaC) bind WalR with significantly higher scores (see explanation in Experimental procedures) than do the positively regulated genes (cwlO, ydjM, yocH): for example, the score for WalR binding to the wapA promoter region is similar to that of other negatively regulated genes (i.e. iseA and pdaC, Table 1) despite the putative DNA binding sequence deviating from the consensus at the most highly conserved positions (Fig. 3A). This difference in binding scores between positively and negatively regulated genes may reflect the fact that repression requires the presence of WalR at the promoter for longer than does activation. A second intriguing feature is that WalR does not bind to the tagAB promoter in either exponentially growing or phosphate limited cells, despite there being two canonical WalR binding sites located in the −35 region of its promoter (Liu et al., 1998). PhoP∼P binds at these motifs to repress tagAB expression (∼ 5×) at the onset of phosphate limitation, but non-phosphorylated PhoP does not bind at these sites (Botella et al., 2011, L.I. Salzberg, D. Noone, E. Botella and K.M. Devine, unpubl. obs.). Therefore, PhoP does not prevent WalR binding to these sites in exponentially growing cells. Furthermore, despite all other PhoPR regulon genes having canonical WalR boxes within their promoters, albeit positioned in inverse orientation relative to the initiation point of transcription, none bind WalR in exponentially growing cells. We show experimentally that simply inverting PhoP binding motifs (thereby generating WalR consensus binding sequences) does not result in WalR-mediated control of expression (Fig. S2), indicating that WalR binding in vivo requires features or conditions not yet identified. A further interesting feature is that WalR binds to an extended DNA region (up to 4.5 kilobases) spanning several open reading frames (yybNMLKIJ, yydBCD, dacA, ggaAB) at several loci, suggesting that perhaps WalR plays a structural role in DNA organization within these regions. We conclude that features additional to the defined DNA consensus sequence are required for WalR binding and regulation of expression in vivo.

WalRK regulates σI and RsgI expression in exponentially growing cells

The sigIrsgI operon promoter is a newly identified locus (8th highest score) to which WalR binds in vivo in exponentially growing, but not in phosphate limited, cells (Table 1, Fig. 4). The σI sigma factor is a member of the σA family whose activity is regulated by the RsgI anti-sigma factor. This regulatory system is highly conserved among Bacillales and regulates cell wall metabolism (lytE, bcrC, mreBH) in response to heat and phosphate limitation stresses (Zuber et al., 2001; Asai et al., 2007; Tseng and Shaw, 2008; Tseng et al., 2011; this study). We show that the sigIrsgI operon has a σA promoter in addition to the σI promoter previously reported. Moreover, we show that WalR activates the σA promoter but represses the σI promoter (Figs 5 and 6). The WalR boxes overlap the σI promoter consistent with repression but are located 28 base pairs upstream of the σA promoter, a position not usual for promoter activation. However, WalR activation of sigIrsgI expression is consistent with the results of Dominguez-Cuevas et al. (2012), who report increased sigIrsgI expression in strain PDC134 that has the WalR* partial gain of function variant. The positioning of the WalR boxes relative to the σA promoter of sigIrsgI suggests that WalR binding may increase expression by a looping mechanism or may induce a localized conformational change in DNA that facilitates interaction between WalR and the RNA polymerase holoenzyme. The rationale for this regulatory arrangement is compelling, displaying features common to positive autoregulatory systems. Under non-stress conditions, a low level of σI and RsgI proteins is maintained by WalR activation of the σA promoter. However, the positive autoregulatory loop is not operative because (i) σI is inactive due to RsgI binding, and (ii) WalR represses the σI promoter. In response to stress stimuli sensed by RsgI, an active σI is released, the stimulus is amplified by positive autoregulation of the sigIrsgI operon in the absence of repression by WalR, leading to increased expression of σI regulon genes.

WalR controls sigIrsgI expression to fulfil the requirement for either CwlO or LytE during growth or under stress conditions

The rationale for WalRK-mediated control of sigIrsgI expression emerges from the finding that B. subtilis viability requires the activity of one or other of the CwlO or LytE endopeptidases (Bisicchia et al., 2007; Hashimoto et al., 2012). Expression of cwlO is dependent on WalRK activation, its only known regulator. Expression of lytE is more complex: the evidence suggests that expression is controlled directly by WalRK from a σA-type promoter and by the σI sigma factor, perhaps with WalR as a co-regulator (Bisicchia et al., 2007; Tseng et al., 2011). The question then arises as to how the endopeptidase requirement (i.e. CwlO or LytE) is fulfilled during growth (WalRK is active) and under stress (WalRK is not active) conditions. WalRK-mediated control of σI RsgI expression is crucial to ensuring that CwlO and/or LytE are appropriately expressed under each condition. WalRK activated control of cwlO expression is sufficient for cell viability. However, WalRK activated control of lytE expression is not sufficient for viability. Null mutations in cwlO and sigIrsgI are synthetically lethal, signifying that active σI is required for viability under conditions where WalRK is not activated (i.e. cwlO is not expressed). However, cells with cwlO and rsgI deleted (σI is constitutively expressed) are viable, showing that some gene under σI control can suppress the synthetic lethality. We show that expression of lytE is the sole requirement for viability in the absence of cwlO since xylose induced lytE expression suppresses the synthetic lethality of deleting cwlO and sigIrsgI. Therefore, the endopeptidase requirement for cell viability is fulfilled by placing CwlO and LytE expression under the control of two regulatory systems each of which responds to different stimuli: WalRK is active during exponential growth directing CwlO expression while σI directs LytE expression during heat stress (Zuber et al., 2001), phosphate limitation (this study) and stationary phase (data not shown).

An integrated model for control of CwlO and LytE expression by WalRK and σI RsgI in exponentially growing and stressed cells

The working model presented in Fig. 7 integrates the roles of WalRK and σI RsgI in the control CwlO and LytE expression in exponentially growing (left panel) and stressed cells (right panel), based on experimental evidence presented here and in other studies (Bisicchia et al., 2007; Tseng et al., 2011). Two fundamental observations underlying this model are that cell viability requires either CwlO or LytE endopeptidase activity and that CwlO expression is absolutely dependent on WalRK activation, its only known regulator (Bisicchia et al., 2007). Therefore, conditions under which WalRK is not (sufficiently) activated require a back-up system for endopeptidase expression. We propose that σI RsgI controlled expression of LytE is the back-up regulatory system that fulfils the endopeptidase requirement under stress conditions. The regulatory connection between WalRK and σI RsgI revealed in this study is a central feature of this model. A basal level of σI and RsgI protein is maintained in exponentially growing non-stressed cells by WalR∼P mediated activation of the σA promoter and by repression of the σI promoter. However, the σI produced is maintained in an inactive state by the cognate RsgI anti-sigma factor. In this state, WalR∼P directs expression of CwlO and LytE, but only the former is produced at a level sufficient for viability. Under stress conditions, reduced WalRK activation is compensated by σI activation and signal amplification by positive autoregulation of sigIrsgI expression through the σI promoter. Thus the σI directed synthesis of LytE compensates for reduced CwlO synthesis. This model implies that although both enzymes can provide the endopeptidase requirement for viability (in fact strains with cwlO and lytE individually deleted have no discernable phenotype) CwlO and LytE activities are adapted to different physiological states (growth and stress respectively). This may account for differences in their expression (Botella et al., 2011), cellular localization, cell wall binding (NLPC/P60 and LysM domains respectively, Yamamoto et al., 2003) and dispensability (we obtain lytE mutants more readily than cwlO mutants). Our studies in differentiating their contributions to cell physiology under different conditions are ongoing. Although WalRK and σI are activated by opposite signals (normal growth and stress), their respective contributions to viability may vary, each being activated to a greater or lesser extent under a particular condition. In this way WalRK and σI cooperate to integrate a diversity of signals to ensure that the endopeptidase requirement (CwlO or LytE) for cell viability is fulfilled, by making their expression responsive to the prevailing environmental and nutritional conditions.

figure

Figure 7. Model integrating the roles of WalRK and σI RsgI in fulfilling the endopeptidase (CwlO and LytE) requirement for cell viability in exponentially growing and stressed B. subtilis cells. A model is presented to illustrate the cooperation between the WalRK and σI RsgI regulatory systems in exponentially growing (left panel) and stressed (right panel) cells. Peptidoglycan chains are represented by alternating light and dark blue rectangles that are linked by peptide chains (bent lines): aberrant peptidoglycan is represented by angled chains (right panel). The cytoplasmic cell membrane is represented by horizontally packed grey phospholipids in which WalK (two transmembrane domains) and RsgI (one transmembrane domain) are embedded. Signals are represented by daggers of different sizes to represent different intensities and with brackets to indicate signal reduction/absence. WalR phosphorylation by WalK is signified by a half-circle arrow whose size represents the signal intensity. Genes are shown as coloured arrows with an associated promoter (P) and the identity of the sigma factor indicated in subscript. Transcription is represented by full arrows of different weights emanating from WalR∼P (reduced activation WalR∼p) and σI to represent different transcription levels. Broken lines represent translation into protein (represented by barrel and square) of different sizes. The plus (+) sign indicates the positive autostimulatory loop generated by σI activation.

Download figure to PowerPoint

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Bacterial strains and growth conditions

Bacterial strains used in this study are listed in Table 2. E. coli strains EC101 (repA+) (Law et al., 1995) and TG-1 (Sambrook et al., 1989) were used for propagating plasmids pG+host4 and pXT in E. coli. E. coli strains were grown in Luria–Bertani (LB) medium (Sambrook et al., 1989). B. subtilis strains were grown in LB, high phosphate-defined medium (HPDM) or LPDM (Müller et al., 1997). Media were supplemented as appropriate with antibiotics at the following concentrations: spectinomycin 100 μg ml−1, erythromycin 150 μg ml−1 (for strain EC101 only), chloramphenicol 5 μg ml−1, kanamycin 10 μg ml−1 and MLS (macrolide lincomycin streptogramin B resistance; erythromycin 1 μg ml−1 + lincomycin 25 μg ml−1). OD600 was measured using a UVmini-1240 UV-VIS spectrophotometer (Shimadzu Scientific Instruments).

Table 2. Strains and plasmids used in this study
Strain or plasmidDescriptionSource or reference
  1. a. LFH signifies long flanking homology PCR.

E. coli strain  
EC101E. coli JM101 with repA from pWV01 integrated into the chromosome (Kmr)Law et al. (1995)
TG-1supE hsdΔ5 thi Δ(lac-proAB) F(traD36 proAB lacIqlacZΔM15)Sambrook et al. (1989)
B. subtilis strains  
168trpC2Laboratory stock
AH9912trpC2 walR::pAH022 (PspacwalRKHIJhtrC; PwalR-lacZ)Howell et al. (2003)
LSB003trpC2 walR::pLIS001 (PwalR-walR-3XFLAG)Bisicchia et al. (2010)
LSB089trpC2 ΔlytE::catLFHa [RIGHTWARDS ARROW] 168
LSB095trpC2 ΔrsgI::kanLFHa [RIGHTWARDS ARROW] 168
LSB096trpC2 ΔsigIrsgI::kanLFHa [RIGHTWARDS ARROW] 168
LSB103trpC2 ΔcwlO::catLFHa [RIGHTWARDS ARROW] 168
LSB110trpC2 ΔsigIrsgI::kan, thrC::pLIS015 (PxylAlytE)pLIS015 [RIGHTWARDS ARROW] LSB096
LSB115trpC2 ΔrsgI::kan, thrC::pLIS015 (PxylAlytE)pLIS015 [RIGHTWARDS ARROW] LSB095
LSB126trpC2 in which WalR boxes in PsigI have been mutated to restriction sitesExcision of pLIS021
LSB279trpC2 ΔrsgI::kan walR::pAH022 (PspacwalRKHIJhtrC; PwalR-lacZ) 
Plasmids  
pG+host4Ts derivative of pGK12 (Ermr)Maguin et al. (1992)
pXTOverexpression vector (PxylA) that integrates at thrC locus (Spcr)Derré et al. (2000)
pDG1661E. coli plasmid for ectopic intergration at amyE containing Cmr cassetteGuérout-Fleury et al. (1996)
pDG646E. coli plasmid containing Ermr cassetteGuérout-Fleury et al. (1995)
pDG780E. coli plasmid containing Kanr cassetteGuérout-Fleury et al. (1995)
pAH022pMUTIN4 derivative for inducible expression of the walR operon (Ermr)Howell et al. (2003)
pLIS015lytE cloned into pXT (Spcr)This work
pLIS021pG+host4 derivative containing mutated WalR boxes in sigI promoter region (Ermr)This work

Strain and plasmid construction

Oligonucleotides used in this study are shown in Table S1. To generate strain LS126, in which the WalR binding boxes within the sigIrsgI promoter are mutated as shown in Fig. 5A, two fragments were amplified with Phusion polymerase using the primer pairs PsigI up fwd MB (H)/sigI EMSA up rev and sigI EMSA do fwd/PsigI do rev MB (B) and B. subtilis 168 chromosomal DNA as template. The fragments were fused by strand overlap extension PCR and cloned into BamHI-HindIII digested pG+host4. The resulting plasmid pLIS021 was propagated in E. coli EC101. The mutated sequences were introduced into the B. subtilis chromosome by integration and excision of plasmid pLIS021 as previously described (Bisicchia et al., 2010). The excisants with the desired mutations in the WalR boxes in the sigIrsgI promoter were identified by digestion of PCR-generated promoter fragments with EcoRI, PstI, NcoI and XhoI.

Strains LSB089 (ΔlytE::cat), LSB095 (ΔrsgI::kan), LSB096 (ΔsigIrsgI::kan) and LSB103 (ΔcwlO::cat) were generated using an adaptation of the long flanking homology (LFH) polymerase chain reaction protocol (Wach, 1996). In brief, DNA fragments encoding the desired antibiotic resistance cassettes (cat and kan) were amplified using an appropriate vector as template (Guérout-Fleury et al., 1995; 1996). Two primer pairs were designed to amplify ∼ 750 base pair DNA fragments flanking the locus to be deleted. Each fragment has a unique ∼ 25 base pair sequence at its 5′ and 3′ ends that are homologous to the corresponding ends of the chosen antibiotic cassette. Primer pairs lytE up fwd/lytE up rev (cat) and lytE do fwd (cat)/lytE do rev were used to generate the DNA fragments required to construct strain LSB089 (ΔlytE::cat); primer pairs OE91/yvcE up rev (cat) and yvcE do fwd (cat)/OE94 were used to generate the DNA fragments required to construct strain LSB103 (ΔcwlO::cat); primer pairs rsgI up fwd/rsgI up rev (kan) and rsgI do fwd (kan)/rsgI do rev were used to generate the DNA fragments required to construct strain LSB095 (ΔrsgI::kan) and primer pairs sigI up fwd/sigI up rev (kan) and rsgI do fwd (kan)/rsgI do rev were used to generate the DNA fragments required to construct strain LSB096 (ΔsigIrsgI::kan). Approximately 150–200 ng of the flanking DNA fragments and 250–300 ng of DNA fragment encoding the desired resistance cassette were joined together using Expand Long Template PCR System (Roche) using the appropriate up-forward and down-reverse primers. The resulting PCR product was used to transform B. subtilis 168 by the method of Anagnostopoulos and Spizizen (1961). The desired transformants were identified by direct screening of colonies using PCR using the up forward- and reverse primers that anneal within the resistance cassette. Strain LSB279 (ΔrsgI::kan walR::pAH022 PspacwalRKHIJhtrC) was generated by transforming strain LSB095 with plasmid pAH022 and selecting for MLS resistance.

To generate strains LSB110 (ΔsigIrsgI::kan thrC::pLIS015 PxylAlytE) and LSB115 (ΔrsgI::kan thrC::pLIS015 PxylAlytE) a DNA fragment encoding a promoterless lytE gene was generated with Phusion polymerase using the primer pair lytE pXT fwd (B)/lytE pXT rev (E) and B. subtilis 168 chromosomal DNA as template. The fragment was cloned into BamHI-EcoRI digested pXT, a derivative of pDG1731 that places the cloned gene under the control of a xylose-inducible promoter (PxylA) (Derré et al., 2000). The resulting pLIS015 plasmid, was propagated in E. coli TG-1. Strains LSB096 (sigIrsgI::kan) and LSB095 (rsgI::kan) were transformed with ScaI-linearized pLIS015 selecting for spectinomycin resistance. The resultant strains LSB110 (ΔsigIrsgI::kan, thrC::pLIS015 PxylAlytE) and LSB115 (ΔrsgI::kan thrC::pLIS015 PxylAlytE) have the lytE gene under xylose inducible control located at the thrC locus.

Chromatin immunoprecipitation using a tiled array and data analysis (ChIP on chip)

A culture of strain LSB003 (walR::pLIS001 PwalRwalR-3xFLAG) was grown at 37°C in LPDM containing erythromycin (1 μg ml−1), lincomycin (25 μg ml−1) and IPTG (100 μM). Cells of strain 168 [for the control (mock) precipitation] were grown in LPDM containing 100 μM IPTG at 37°C. A 200 ml culture sample was harvested at the mid-exponential phase of growth (T1, OD600 0.2) and a 50 ml culture sample was harvested 2 h after the onset of phosphate starvation state (T2). Cross-linking was performed by the addition of formaldehyde to a final concentration of 1% followed by a 20 min incubation at room temperature with slow shaking. The cultures were then quenched with glycine (0.36 M final concentration) for 5 min at room temperature. Cells were collected by centrifugation at 4°C and the pellets were washed twice with 50 ml ice-cold Buffer A (10 mM Tris-HCl pH 7.5, 150 mM NaCl) before being snap frozen in a dry ice/ethanol bath.

Pellets were thawed at 37°C and resuspended in 0.5 ml buffer B (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.2 mM EDTA, 0.1% Triton X-100) supplemented with 3 mg ml−1 lysozyme and 0.1 mg ml−1 RNase A (final concentrations) and incubated for 30 min at 37°C. The volume of the sample was then adjusted to 1 ml with buffer B and samples placed on ice for 10 min. Genomic DNA was sheared by sonication (Diagenode Bioruptor Twin sonicator) into fragments between 0.2 and 1 kilobase in size. Sonicated samples were centrifuged for 10 min at 20 000 g (4°C) to remove insoluble debris. Fifty microlitres of the cleared lysates was removed to serve as the ‘input sample’ and 50 μl of 10× buffer C (10 mM Tris-HCl pH 8, 10% SDS, 10 mM EDTA) and 400 μl buffer B were added and the sample was stored on ice. Forty microlitres of anti-FLAG M2-agarose resin (Sigma-Aldrich) was added to each of the remaining cleared lysates, which were equilibrated with ice-cold buffer B (according to manufacturer's instructions) and incubated for 2 h at 4°C with mild agitation. The resin beads were then collected by centrifugation (5000 g for 1 min at 4°C) and were washed three times with 0.5 ml ice-cold buffer B. To reverse the cross-linking, beads were transferred to new 1.5 ml eppendorf tubes, resuspended in 0.5 ml 1× buffer C and incubated for ∼ 20 h at 65°C with shaking (950 r.p.m.). The input samples were similarly treated. DNA fragments that were immunoprecipitated were collected by centrifugation for at 13 300 r.p.m. for 1 min at 4°C. DNA from the chromatin immunoprecipitations and ‘input samples’ was purified using the QIAquick PCR purification kit (Qiagen), eluted in 50 μl H2O and treated with 0.02 mg ml−1 RNase A (Roche) for 30 min at 37°C. RNase was removed using the QIAquick PCR purification kit and samples were eluted in 30 μl H2O.

Equal amounts of immunoprecipitated and input DNA were used as template and amplified using the GenomePlex complete Whole Genome Amplification (WGA) kit (Sigma-Aldrich) according to the manufacturer's instructions, purified using the QIAquick PCR purification kit and eluted in 30 μl H2O. Ten microliters of each of the purified amplified samples was amplified a second time and purified as previously described. Aliquots of 3.5–4 μg of twice-amplified immunoprecipitated and ‘input sample’ DNA prepared from cultures containing both tagged and non-tagged WalR at a concentration of ∼ 250 ng μl−1 (A260/A280 ≥ 1.7 and A260/A230 ≥ 1.6) from three separate ChIP experiments were sent to NimbleGen. Immunoprecipitated DNA was labelled with Cy5 dye and ‘input sample’ DNA was labelled with Cy3 dye and samples were hybridized to the BaSysBio B. subtilis T2 385 K array (Rasmussen et al., 2009).

The ChIP-chip data were processed with Ringo, a Bioconductor package for the analysis of ChIP-on-chip data from the NimbleGen platform (Toedling et al., 2007). After reading in the data into the tool, images were generated for quality assessment. The data were normalized using the Tukey-biweight scaling procedure. This method is recommended by NimbleGen to scale ChIP-chip readouts so that the data are centred on zero. Peaks were smoothed by computing medians in a running window approach. The window size was set to 700 bp to correlate with the average fragment size. During this step the values from three replicates were combined. The smoothed control data were then subtracted from the smoothed sample data to remove background noise. After calculating a noise-threshold for each antibody on the smoothed and background-subtracted data the significant peaks were called with a probability threshold of 99.99%. Only peaks with a significant signal on more than 30 contiguous tiles were considered. Parameters given in Table 1 are as follows: the score assigned to a peak is the sum of the log ratios (signals) for the probes within the peak; the Max(log2) value is the highest measure of binding signal within the peak and the size of the peak is the length of DNA covered by the probes that have a signal above the background threshold. It is important to visualize binding peak morphology (Figs 1 and S1) since a particular score could represent either a narrow peak with a very high signal or a broad peak with a low signal. Submission of the ChIP on chip data to GEO is in progress.

Gel mobility shift DNA binding assays

His6x-tagged WalR and ′WalK proteins were purified from strain BL21(DE3) as previously described (Bisicchia et al., 2010). DNA fragments spanning the promoter regions of the following genes were prepared by PCR using B. subtilis strain 168 chromosomal DNA and the stated primers: wapA, a 227 base pair DNA fragment amplified using wapA EMSA fwd/wapA EMSA rev primers; pbpH, a 459 base pair DNA fragment amplified using pbpH EMSA fwd/pbpH EMSA rev); sigI, a 289 base pair DNA fragment amplified using sigI EMSA fwd/sigI EMSA rev); oppA, a 475 base pair DNA fragment amplified using oppA EMSA fwd/oppA EMSA rev); yybN, a 442 base pair DNA fragment amplified using yybN EMSA fwd/yybN EMSA rev and htrA, a 300 base pair DNA fragment amplified using htrAbiotin/htrAGSR) primers. The forward primer was biotinylated in each case. To create the 289 bp DNA fragment encoding the sigI promoter with mutated WalR boxes (replaced with restriction sites), upstream and downstream DNA fragments were amplified with Phusion polymerase using primers sigI EMSA/sigI EMSA up rev and sigI EMSA do fwd/sigI EMSA rev respectively. The two fragments were then fused by strand overlap extension using Phusion polymerase. The biotinylated fragments were gel purified according to standard procedures. Binding assays were performed as follows: increasing amounts of WalR were incubated with 1 μM ′WalK in the presence of 1 mM ATP, 10 mM Tris-HCl (pH 7.4), 50 mM NaCl, 5% glycerol, 1 mM EDTA, 4 mM DTT, 0.5 μg sheared herring sperm DNA and 4 μg BSA. The phosphorylation of WalR was allowed to proceed for 15 min at room temperature. Then 2 ng of biotinylated probe was added and the reaction was incubated for a further 10 min at room temperature. The reactions were then loaded on 6% polyacrylamide gels made with Tris-acetate buffer and electrophoresis proceeded at 4°C. Bands were detected after electrotransfer to BioDyne membrane (Pall) using the Phototope-Star Detection kit (NEB) according to manufacturer's instructions.

RNA extraction and qRT PCR

Cells were depleted for WalRK as follows: a fresh colony of newly constructed strain AH9912 was used to inoculate LPDM containing erythromycin (1 μg ml−1), lincomycin (25 μg ml−1) and 75 μM IPTG and incubated until the culture was growing exponentially. This culture was used to inoculate three flasks of fresh LPDM (containing the same antibiotics) one of which contained IPTG at a concentration of 1 mM IPTG, one at 75 μM and one without the IPTG addition (cells in this culture will become depleted for WalRK). To prepare total cellular RNA from strains 168 and LS126, flasks containing LPDM were inoculated with and overnight culture grown in HPDM. Cultures were grown at 37°C with shaking at 220 r.p.m. in an orbital shaker (New Brunswick, Edison, NJ). Samples were harvested at designated times, centrifuged for 2–3 min at 4°C and cell pellets were snap frozen in a dry ice–ethanol bath. Cell pellets were either stored at −70°C or processed immediately. Total RNA was prepared using the GenElute Mammalian Total RNA Miniprep Kit (Sigma) the following manufacturer's instructions with the following modifications: cells were first broken using a Fastprep shaker (Bio101) as previously described (Noone et al., 2000) and DNase was added on the column (10U RQ1 DNase, Promega) and incubated for 15 min at 37°C. cDNA synthesis was conducted using the Transcriptor High Fidelity cDNA synthesis kit (Roche) using 1 μg DNase-treated total RNA and a random hexamer primer according to manufacturer's instructions. As a control to monitor removal of DNA, parallel reactions lacking reverse transcriptase were also carried out. Quantification of transcript levels by qRT PCR was conducted using the LightCycler 480 SYBR Green I Master kit (Roche) on a LightCycler 480 instrument (Roche). Reactions were set up in duplicate and crossing points (Cp) were determined using the Second Derivative Maximum Method of the LightCycler 480 software (version 1.5.0). The level of 16S ribosomal RNA was used as a reference to normalize samples and calculate relative expression ratios were calculated using the 2−ΔΔCT method (Livak and Schmittgen, 2001). Primers for quantitative PCR were designed using PrimerExpress 3.0 software and a melt curve cycle was included for every primer set to check the specificity of each amplification.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The authors would like to thank Jeff Errington for the gift of strains and Philippe Noirot for discussion on ChIP on chip analysis. This work was funded by Science Foundation Ireland Award 08/IN.1/B1859. Data visualization was performed on a server maintained by the Trinity Centre for High Performance Computing that is funded by grants from Science Foundation Ireland. Karsten Hokamp is funded by Science Foundation Ireland.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
mmi12092-sup-0001-si.pdf682K

 

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.