In Bacillus subtilis, the WalRK (YycFG) two-component system controls peptidoglycan metabolism in exponentially growing cells while PhoPR controls the response to phosphate limitation. Here we examine the roles of WalRK and PhoPR in peptidoglycan metabolism in phosphate-limited cells. We show that B. subtilis cells remain viable in a phosphate-limited state for an extended period and resume growth rapidly upon phosphate addition, even in the absence of a PhoPR-mediated response. Peptidoglycan synthesis occurs in phosphate-limited wild-type cells at ∼27% the rate of exponentially growing cells, and at ∼18% the rate of exponentially growing cells in the absence of PhoPR. In phosphate-limited cells, the WalRK regulon genes yocH, cwlO(yvcE), lytE and ydjM are expressed in a manner that is dependent on the WalR recognition sequence and deleting these genes individually reduces the rate of peptidoglycan synthesis. We show that ydjM expression can be activated by PhoP∼P in vitro and that PhoP occupies its promoter in phosphate-limited cells. However, iseA(yoeB) expression cannot be repressed by PhoP∼P in vitro, but can be repressed by non-phosphorylated WalR in vitro. Therefore, we conclude that peptidoglycan metabolism is controlled by both WalRK and PhoPR in phosphate-limited B. subtilis cells.
The cell wall of Gram-positive bacteria is composed of approximately equal amounts of peptidoglycan and anionic polymers (polyol phosphates, e.g. wall teichoic acid and lipoteichoic acid) and about 20% protein (for reviews on cell wall structure and synthesis see Bhavsar and Brown, 2006; Barreteau et al., 2008; Sauvage et al., 2008; Vollmer et al., 2008). It is an enigmatic structure requiring apparent opposing properties for its function: it is rigid, determining the shape of the cell and can withstand turgor pressure, yet it is sufficiently flexible to allow cell growth, cell division and responses to osmotic stress; it acts as a barrier between the cell and its environment, yet it is permeable to nutrient ingress, and to egress of large proteins and other macromolecules; its integrity is essential for viability, yet cell wall degradation must occur during growth, elongation of the cell cylinder and cell division. Such attributes signify that the cell wall is a dynamic structure whose synthesis, maintenance and functions display considerable complexity.
Its location external to the cytoplasmic membrane requires there to be communication between the cell wall and the internal cellular milieu. Peptidoglycan and teichoic acid precursors, penicillin binding proteins, and autolysins are synthesized intracellularly, and must be exported to appropriate sites for synthesis and assembly of the mature cell wall to occur (Bhavsar and Brown, 2006; Barreteau et al., 2008; Sauvage et al., 2008; Vollmer et al., 2008). Therefore, intracellular production of cell wall precursors and proteins must be spatially and temporally coordinated with their external utilization. Our knowledge of how communication occurs between the cell wall and the cytoplasm is growing and is now known to involve transmembrane protein complexes, two-component signal transduction systems (TCS) and sigma factor/anti-sigma factor systems. Recent studies suggest that peptidoglycan and teichoic acid synthesis in the cell cylinder is associated with a multiprotein complex that spans the cytoplasmic membrane thereby linking processes that occur in the outer and inner cellular compartments (Carballido-Lopez and Formstone, 2007; Formstone et al. 2008). The constituent proteins of this complex include the membrane-embedded RodA, MreC and MreD proteins, the actin-like MreB, MreBH and Mbl proteins that form filaments at the cytoplasmic face of the membrane and the PBP2a and PbpH penicillin binding proteins located outside the cytoplasmic membrane (Daniel and Errington, 2003; Lee and Stewart, 2003; Wei et al., 2003). This complex is also proposed to participate in localization of the LytE autolysin to its site of action within the cell wall (Carballido-lopez et al., 2006). The peptidoglycan and teichoic acid synthetic complexes are distributed in a helical manner along the cell cylinder through their association with the triplet of actin-like filaments, resulting in a helical and punctate mode of cell wall synthesis along the cell cylinder (Jones et al., 2001; Daniel and Errington, 2003; Formstone et al., 2008). A transmembrane multiprotein complex, but with a different protein composition also forms at the septum to effect septation and cell division (for review see Carballido-Lopez and Formstone, 2007).
Communication between the cell wall milieu and the cytoplasm is also mediated by TCS. The WalRK (YycFG) TCS regulates cell wall metabolism in growing cells by controlling expression of autolysins [YocH, CwlO(YvcE) and LytE] and modulators of autolysin activity [IseA(YoeB) and YjeA]. The IseA protein inhibits autolysin activity directly while YjeA deacetylates peptidoglycan, thereby modulating its susceptibility to autolytic digestion (Bisicchia et al., 2007; Salzberg and Helmann, 2007; Yamamoto et al., 2008). Significantly, cell elongation requires either the CwlO or LytE endopeptidase-type autolysin, emphasizing the crucial role of WalRK in cell wall metabolism (Bisicchia et al., 2007). Hoch et al. have shown recently that WalK is septally located forming a complex with YycH, YycI and FtsZ (Szurmant et al., 2005; 2007; Fukushima et al., 2008). There are indications that WalK senses multiple signals: WalK activity is enhanced by the presence of a septum (Fukushima et al., 2008); a survey of relevant transcriptomic studies in Bacillus subtilis (performed in this laboratory by P.B. and K.M.D.) suggests WalK may sense some aspect of Lipid II (perhaps the D-Ala-D-Ala moiety), a crucial intermediate in cell wall synthesis (Dubrac et al., 2008) while the association between WalK and YycH and YycI suggests that WalK may respond to signals emanating from within the cell membrane (Szurmant et al., 2007; 2008; Fukushima et al., 2008).
Two-component systems and ECF-type sigma factor/anti-sigma factor pairs also mediate communication between the cell wall and cytoplasmic compartments in detecting any compromise of cell wall integrity and in executing an appropriate response. In B. subtilis four TCS (LiaRS, BceRS, YvcPQ and YxdJK) and a number of ECF-type sigma factors (for example, SigW, SigX and SigM) are known to function in mediating appropriate responses to a wide variety of conditions and agents that damage the cell wall (Helmann, 2002; 2006; Mascher, 2006; Mascher et al., 2006; 2007; Jordan et al., 2008; Staron et al., 2009). In addition, the PhoPR TCS, which mediates one of the cellular responses to phosphate limitation in B. subtilis, plays a role in cell wall homeostasis under this condition (Hulett, 2002). Many PhoPR regulon genes function in scavenging phosphate, e.g. alkaline phosphatases (phoA, phoB, phoD) and a high-affinity phosphate transporter (pstSCAB1B2). However, PhoPR also controls the teichoic acid/teichuronic acid transition that occurs in response to phosphate limitation by repressing the tagABC operon and activating expression of the tuaABCDEFGH operon. It is clear therefore that there are a multiplicity of mechanisms to enact bidirectional communication between the cell wall and the intracellular milieu.
Previous studies show many similarities and associations between the WalRK and PhoPR TCS in B. subtilis. They cluster (with ResDE) in a phylogenetic tree of TCS from B. subtilis (Fabret and Hoch, 1998). The PhoP′-′WalR and WalR′-′PhoP hybrid regulator proteins are both functional and capable of being activated by the cognate kinase of the receiver domain (Howell et al., 2003; 2006). Moreover, unidirectional cross-talk can occur between these TCS, with the PhoR sensor kinase being capable of phosphorylating both PhoP and the non-cognate WalR in vitro (Howell et al., 2006). In addition, we have noted that the consensus WalR DNA recognition sequence [TGT(A/T)A(A/T/C)-N5-TGT(A/T)A(A/T/C)] is very similar to the reverse complement of the PhoP consensus [TT(A/T/C)ACA-N4–6-TT(A/T/C)ACA] (Hulett, 2002). Thus the DNA recognition sequences of WalR and PhoP are in opposite orientations in their respective cognate promoters (Hulett, 2002; Howell et al., 2003; 2006). This has added significance in view of the following observations: (i) WalR can bind to the phoPR promoter, (ii) cells depleted for WalRK cannot mount a normal PhoPR-dependent phosphate limitation response and (iii) cells depleted for WalRK lyse during growth in Luria–Bertani (LB), whereas lysis does not occur when WalRK depletion occurs concomitantly with phosphate limitation (Howell et al., 2003; 2006). Despite these observations, the reported roles of WalRK (i.e. peptidoglycan metabolism in exponentially growing cells) and PhoPR (i.e. phosphate scavenging and anionic polymer metabolism in phosphate-limited cells) appear quite dissimilar.
In this study we investigate the involvement of WalRK and PhoPR in cell wall metabolism in B. subtilis. We show that peptidoglycan synthesis occurs in phosphate-limited cells, even in the absence of a PhoPR-mediated response and that autolysins and other members of the WalRK regulon play a role in this process. Moreover, we show that PhoP∼P can activate expression of the WalRK regulon gene ydjM in vitro suggesting that both WalR∼P and PhoP∼P play a role in peptidoglycan metabolism in phosphate-limited cells.
Bacillus subtilis cells are viable during phosphate limitation even in the absence of a PhoPR-mediated response
To better characterize the PhoPR-mediated response in B. subtilis, we determined cell growth and viability during phosphate limitation. When wild-type cells become phosphate-limited, growth is strongly attenuated but cells continue growing at a reduced rate for an extended period (Fig. 1). In contrast growth of AH024 (ΔphoPR) cultures ceases and cultures remain static when cells encounter phosphate-limiting conditions (Fig. 1). However, despite this difference in response, both cultures retain 100% viability even after exposure to phosphate-limiting conditions for up to 3 h (data not shown). A microscopic analysis did not reveal any difference in cell morphology between phosphate-limited wild-type (strain 168) and phoPR (strain AH024) mutant cells (data not shown). When phosphate was added after 3.5 h of phosphate limitation (arrow in Fig. 1), both cultures rapidly resumed growth but with different profiles. Strain AH024 (ΔphoPR) resumed growth with a significantly higher growth rate than did the wild-type culture (Fig. 1). These observations show that phosphate-limited cells are viable and can rapidly resume growth when phosphate becomes available. Surprisingly the PhoPR-mediated response is not necessary for the maintenance of viability and may even place cells at a disadvantage when phosphate becomes available.
Cell wall synthesis occurs in phosphate-limited cells but is reduced in a phoPR mutant
The viability of wild-type and phoPR mutant cells during phosphate limitation prompted us to examine cell wall synthesis during this period. The contribution of the WalRK and PhoPR two-component systems was of particular interest because of their documented roles in peptidoglycan and teichoic acid metabolism (Qi and Hulett, 1998; Hulett, 2002; Howell et al., 2003; Bisicchia et al., 2007). Cell wall synthesis was monitored by measuring [14C]-N-acetylglucosamine accumulation in cells of strains 168 (wild-type) and AH024 (ΔphoPR). Label was added to exponentially growing cultures at OD600 0.1 and at 3 h after the onset of phosphate limitation (T3). Accumulation of [14C]-N-acetylglucosamine was measured at 15 min intervals in exponentially growing cultures (Fig. 2A) and at 20 min intervals during phosphate limitation (Fig. 2B) and all experiments were performed at least twice. To verify that cell-associated [14C]-N-acetylglucosamine reflects the level of peptidoglycan synthesis, the proportion of total cell radioactivity present in cell walls was determined. Between 60% and 70% of total cell radioactivity was present in the purified cell wall fractions, a level that was similar in 168 (wild-type) and AH024 (ΔphoPR) strains (data not shown). The rate of [14C]-N-acetylglucosamine accumulation in exponentially growing wild-type cells was ∼2.2 × 105 cpm OD600−1 h−1 (Fig. 2A) and decreased to ∼6 × 104 cpm OD600−1 h−1 (Fig. 2B) during the T3–T5 period of phosphate limitation. Thus peptidoglycan continues to be synthesized in phosphate-limited cells at ∼27% the rate of exponentially growing cells. To gauge the effect of the PhoPR two-component system on peptidoglycan synthesis this experiment was repeated with strain AH024 (ΔphoPR). There was no difference in the rates of [14C]-N-acetylglucosamine accumulation in exponentially growing cells of strains 168 (wild-type) and AH024 (ΔphoPR) cultures (Fig. 2A). However, the rate of accumulation in phosphate-limited cells of strain AH024 (ΔphoPR) was 4 × 104 cpm OD600−1 h−1, a ∼33% reduction in the rate observed in phosphate-limited wild-type strain 168. Thus assigning 100% to the rate of peptidoglycan synthesis in exponentially growing cells of strain 168 (wild-type), the rate of synthesis in phosphate-limited cells of strain 168 is reduced to ∼27% and is further reduced to ∼18% in phosphate-limited cells of strain AH024 (ΔphoPR). Therefore, we conclude that cell wall synthesis occurs at a significant rate in phosphate-limited cells in a manner that is affected by PhoPR.
Autolysins of the WalRK regulon contribute to peptidoglycan metabolism during phosphate limitation
The WalRK and PhoPR two-component systems regulate peptidoglycan and anionic polymer metabolism respectively (Hulett, 2002; Bisicchia et al., 2007). However, the results presented above suggest that PhoPR may also have a role, either directly or indirectly, in controlling peptidoglycan metabolism. Therefore, we asked if genes of the WalRK regulon are expressed during phosphate limitation, and whether PhoPR influences their expression. The results presented in Fig. 3 show the relative RNA levels of WalRK regulon genes in exponentially growing and phosphate-limited cells of strains 168 (wild-type) and AH024 (ΔphoPR). The increased level of phoA and pstS transcripts at T0.5 confirms that the PhoPR response is activated at this time. In strain 168 (wild-type) the levels of yocH, cwlO, lytE and ydjM transcript are high during exponential growth, decrease at the point of phosphate limitation but then increase gradually during the 5 h phosphate limitation period. However, expression of these genes in strain AH024 (ΔphoPR) differs from the profile seen in strain 168 (wild-type) primarily during the phosphate limitation period (Fig. 3). In exponentially growing cells, transcript levels in strain AH024 (ΔphoPR) are similar to those observed in strain 168 (wild-type). However, in phosphate-limited cells of strain AH024 (ΔphoPR), the transcript level of all four genes decreases significantly at T0.5 and remains at this low level for the duration of the 5 h phosphate limitation period.
As the WalRK regulon genes are expressed throughout phosphate limitation, we then asked if they contribute to peptidoglycan metabolism during this period. To address this question, the accumulation of [14C]-N-acetylglucosamine was measured in strains AH023 (ΔyocH) and BP079 (ΔcwlO) during the T3–T5 period of phosphate starvation (squares and x-symbols, respectively, in Fig. 2B). The results show decreased rates of [14C]-N-acetylglucosamine accumulation in the cells of both strains, with the observed rates being intermediate between those of strains 168 (wild-type) and AH024 (ΔphoPR). Also the cwlO mutation has a greater effect on peptidoglycan synthesis than does the yocH mutation. A similar experiment performed with strains BP068 (ΔydjM) and L16638 (ΔlytE) showed decreased rates of [14C]-N-acetylglucosamine accumulation in phosphate-limited cells, reduced in both strains to the same extent as observed in strain AH023 (ΔyocH; data not shown). Thus four members of the WalRK regulon, three autolysins (YocH, CwlO and LytE) and a cell wall-associated protein (YdjM) contribute to peptidoglycan synthesis in phosphate-limited cells.
These results led us to predict that the endogenous autolytic activity of phosphate-limited AH024 (ΔphoPR) cells would be lower than that of strain 168 (wild-type) cells. This prediction was tested by monitoring the autolysis of exponentially growing, and phosphate-limited cells of strains 168 (wild-type) and AH024 (ΔphoPR) for up to 3 h as outlined in Experimental procedures. The results are shown in Fig. 4. The rates of cell autolysis of exponentially growing wild-type and AH024 (ΔphoPR) cells are very similar (Fig. 4A). However, phosphate-limited wild-type cells (diamonds) lysed at a higher rate than did phosphate starved phoPR mutant cells (circles), a trend that was observed in three separate experiments (Fig. 4B). The same results were obtained in a separate series of experiments in which cell autolysis of wild-type and phoPR mutated cells was triggered by treatment with NaN3 (data not shown). Taken together the results show that the YocH, CwlO, LytE and YdjM proteins of the WalKR regulon are involved in peptidoglycan metabolism in phosphate-limited cells. Moreover, they also demonstrate that the PhoPR TCS plays a role in controlling peptidoglycan metabolism during phosphate limitation.
WalRK and PhoPR are expressed throughout exponential growth and phosphate limitation
The finding that YocH, CwlO, LytE and YdjM proteins of the WalRK regulon are expressed and contribute to peptidoglycan metabolism in phosphate-limited cells prompted us to ask how their expression is controlled during phosphate limitation. We therefore determined the relative transcript and protein levels of both WalR and PhoP in exponentially growing and phosphate-limited cells. The results (Fig. 5A) show the relative level of WalR and PhoP RNA transcripts normalized to the level phoP transcript at T−2.5. It is evident that the phoP transcript is present during exponential growth, increases approximately two- to three-fold at the point of phosphate limitation and remains at an elevated level throughout the remainder of the experiment, consistent with previous reports (Sun et al., 1996; Pragai et al., 2004). The walR transcript is also present throughout growth and phosphate limitation with the level in exponentially growing cells being approximately twofold higher than in phosphate-limited cells (Fig. 5A). Surprisingly, the level of phoP transcript is two- to three-fold higher than the level of walR transcript during exponential growth and up to 10-fold higher during phosphate limitation. WalR and PhoP protein levels were compared in exponentially growing and phosphate-limited cells by performing Western analysis on extracts from strains LSB002 and LSB003 that expressed PhoP and WalR proteins, respectively, tagged at their C-termini with the same 3xFLAG epitope, with expression of the downstream walK and phoR genes under the control of the IPTG-inducible Pspac promoter. As WalR and WalK are essential, the dependence of strain LSB003 growth on the presence of ≥ 100 µM IPTG shows (i) that WalR-3xFLAG is functional and (ii) that the downstream walK gene is being expressed (Howell et al. 2006 reported that 100 µM IPTG was the minimum level required for growth of strain AH9913 in which walRK expression is under Pspac control). Likewise PhoP-3xFLAG is functional because strain LSB002 has an essentially normal phosphate limitation response in the presence of IPTG, showing increased expression of the PtuaAgfp transcriptional fusion and elevated alkaline phosphatase levels. The results of the western analysis are shown in Fig. 5B. The level of WalR-3xFLAG is higher than that of PhoP-3xFLAG during exponential growth, in contrast to the transcript levels in this period. The level of PhoP-3xFLAG increases at T0 to the point where the levels of both proteins are comparable, consistent with the increased phoP mRNA level. Interestingly, WalR-3xFLAG and PhoP-3xFLAG proteins accumulate to comparable levels in phosphate limited cells (T1) despite the difference in their relative transcript levels. Together these results show that the WalRK and PhoPR two-component systems are both expressed in exponentially growing and phosphate-limited cells and suggest that their expression might be controlled post-transcriptionally.
WalR and PhoP can both bind to the promoter regions of the WalRK and PhoPR regulon genes
We previously reported that the WalR and PhoP consensus DNA binding sequences are the reverse complement of each other (Hulett, 2002; Howell et al., 2006). This is illustrated by comparing the sequences of the yocH (WalRK regulon) and tuaA (PhoPR regulon) promoters (Fig. 6A). WalR binding sequences are in red while PhoP binding sequences are in blue. The TGT and ACA residues of the direct repeats within each binding motif are the most highly conserved bases of the WalR and PhoP DNA binding sequences respectively. We noted that the WalR DNA binding motif of all WalRK regulon genes is in the orientation that places the TGT sequence on the top strand, while the PhoP binding motif of all PhoPR regulon genes is in the orientation that places the ACA sequence on the top strand (Fig. 6A). Thus the WalR and PhoP DNA binding motifs are very similar but are in opposite orientations within WalRK- and PhoPR-recognized promoters. In view of these observations we predicted that the WalR and PhoP proteins should bind to the promoters of both regulons. To test this hypothesis, we performed gel shift analysis on six promoters of WalRK regulon genes (yocH, cwlO, ydjM, yjeA, lytE and iseA) and two PhoPR regulon genes (phoA and phoPR), testing their ability to be bound by both phosphorylated WalR and PhoP proteins. The results are presented in Fig. 6B. It is evident that PhoP∼P directly binds to the promoter regions of all six WalRK regulon genes. In addition WalR∼P binds to the promoter regions of the two PhoPR regulon genes. Further examination reveals interesting differences in the ability of individual promoters to be bound by PhoP∼P and WalR∼P: for example, PhoP∼P and WalR∼P protein levels in the 2.5–5 µM range are required to first observe a mobility shift of the phoPR promoter, approximately 3–10-fold higher than any of the other promoters. In addition, WalR and PhoP proteins have different affinities for individual promoters. PhoP∼P binds to the yocH promoter DNA with a higher affinity than does WalR∼P: the promoter fragment is almost completely shifted with 0.72 µM PhoP∼P but no discernable shift is observed with this level of WalR∼P. Similarly PhoP∼P binds to the phoA promoter with a higher affinity than does WalR∼P. Conversely WalR∼P has a higher affinity for the lytE and iseA promoters than does Pho∼P. These data demonstrate that promoters of WalRK regulon genes can be bound by PhoP∼P at levels comparable to that observed for WalR∼P.
PhoP can activate transcription of ydjM but cannot repress iseA transcription in vitro
The results presented thus far show that the WalRK regulon genes are expressed and participate in peptidoglycan metabolism in phosphate-limited cells; that WalR and PhoP proteins are both present at comparable levels during phosphate limitation and that WalR and PhoP can both directly bind to the promoters of the constituent genes of both regulons. This prompts the question: is expression of WalRK regulon genes (i.e. yocH, cwlO, ydjM, lytE, iseA and yjeA) controlled by WalRK or PhoPR during phosphate limitation?
To address this question we first asked if expression of WalRK regulon genes in phosphate-limited cells is dependent on the presence of a WalR DNA binding sequence. Strains EL63 and EL64 were constructed in which one hexameric motif of the WalR binding sequence within the yocH and ydjM promoters was mutated (Fig. 7A). Such mutations were shown to reduce WalR-mediated control of gene expression during growth in LB (Bisicchia et al., 2007). Importantly both strains contain the wild-type PhoPR and WalRK proteins. The expression profiles were determined by Northern analysis and are presented in Fig. 7A. The yocH and ydjM genes are expressed in cells of strain 168 (wild-type) during exponential growth and phosphate limitation, with decreased expression during the intervening period, consistent with previous results (Fig. 7A, top panel; Bisicchia et al., 2007). However, there is a significant reduction in yocH transcript accumulation when the WalR binding sequence is mutated (Fig. 7A, middle panel; strain EL63). Crucially expression is reduced both in exponentially growing (T−1.5–T0) and in phosphate-limited (T0–T2.5) cells. Similarly mutation of the WalR binding sequence within the ydjM promoter reduces the level of ydjM transcript during both exponential growth and phosphate limitation – in this case no transcript was observed at either growth stage even upon prolonged exposure (Fig. 7A, middle panel; strain EL64). The profile of phoA transcript accumulation in these same RNA preparations (Fig. 7A, bottom panel) shows that the PhoPR-dependent phosphate limitation response was induced as expected in both cultures. These expression profiles were confirmed by measurement of ydjM and yocH transcripts by quantitative RT-PCR is a separate set of RNA samples (data not shown). These results show that expression of WalRK regulon genes yocH and ydjM is dependent on the WalR DNA binding sequence in phosphate-limited cells.
We then tested whether PhoPR can regulate transcription of ydjM and iseA in vitro in a manner similar to WalRK. These two genes were chosen to represent the activation (ydjM) and repression (iseA) capabilities of WalRK. The results are shown in Fig. 7B. Transcription of the tuaA gene was used to confirm PhoPR activity in vitro. A tuaA run-off transcript of the expected size was produced only when PhoP∼P (i.e. in the presence of PhoP and ′PhoR) was present in the in vitro reactions, confirming that the PhoP and ′PhoR proteins are active (data not shown). With the ydjM template, a low barely detectable level of run-off transcript was observed upon addition of RNA polymerase holoenzyme (Eσ) alone. Transcript increased threefold upon WalR addition, and up to eightfold when both WalK and ′WalR were present, confirming that expression of ydjM is WalRK dependent (Fig. 7B and C). The level of ydjM transcript increased threefold upon PhoP addition (comparable to that observed when WalR was added) but crucially, up to sixfold when both PhoP and ′PhoR were added to the reaction (Fig. 7B and C). These results show that ydjM transcription can be specifically activated by the PhoPR TCS in vitro. However, the situation with iseA is different. Consistent with iseA being repressed by WalRK in vivo, a high level of run-off transcript was observed in vitro upon addition of the RNA polymerase holoenzyme (Eσ) alone, as expected (Fig. 7B, right panel, lanes F–J). Addition of WalR to this reaction causes a significant decrease (to 80%) in iseA transcript that is further decreased (to 40%) when both WalK and ′WalR were added, confirming that WalR∼P represses iseA transcription, as shown in vivo (Fig. 7B, Bisicchia et al., 2007). However, addition of either PhoP alone or of both PhoP and ′PhoR does not cause any decrease in iseA run-off transcript level showing that neither PhoP nor PhoP∼P can repress iseA transcription (Fig. 7B, right panel, lanes I, J). We then determined the level of ydjM and iseA run-off transcripts in vitro in the presence of both the WalR′K and PhoP′R TCS a situation that mimics that of phosphate-limited cells (see Fig. 5). The results presented in Fig. 7B are confirmed in this repeat set of reactions (with the exception of lane I where WalR addition alone lead to a greater than expected decrease) that are included for internal comparison (Fig. 7C). Significantly addition of both WalR′K and PhoP′R proteins to the reaction resulted in a ninefold increase in ydjM transcript level, compared with a sevenfold increase observed with WalR′K alone (Fig. 7C, compare lanes C and E). The results for iseA expression are striking (Fig. 7C, right panel, lanes H–N): the level of repression achieved by WalR′K is not affected by the addition of either PhoP or PhoP′R, showing that neither PhoP nor PhoP∼P interferes or competes with WalR∼P for binding to the iseA promoter (Fig. 7C, lanes K and L). These results show that PhoP′R can activate transcription of ydjM in vitro, albeit to a lower extent than WalR′K, but that PhoP′R cannot repress iseA transcription nor does it interfere with WalRK-mediated repression of iseA in vitro.
PhoP occupies the ydjM and iseA promoters in phosphate-limited cells
To ascertain the physiological relevance of these in vitro transcription results, we determined the extent to which the WalR and PhoP proteins occupy the promoters of ydjM, iseA, phoA and tagA in vivo in exponentially growing and phosphate-limited cells. Strains LSB003 and LSB002 express 3xFLAG-tagged WalR and PhoP, respectively, in a manner that places expression of downstream genes under the control of the IPTG-inducible Pspac promoter, and both proteins are functional as previously discussed. Cross-linking of proteins to DNA and chromatin immunoprecipitation (ChIP) with a FLAG-specific antibody was carried out as described in Experimental procedures. The enrichment of ydjM, iseA, phoA and tagA promoter DNA fragments in precipitated DNA was measured by quantitative PCR. The same procedure was carried out with wild-type strain 168 (no FLAG-tagged proteins, termed ‘mock precipitation’) to control for, and ascertain the level of, non-specific precipitation of DNA fragments. The results are shown in Fig. 8. There is significantly increased occupancy of the iseA promoter by WalR in exponentially growing cells (Fig. 8A). While occupancy of the ydjM promoter is somewhat elevated over that observed in the ‘mock’ precipitation, the levels observed are close to the threshold of significance (Fig. 8A). However, in exponentially growing cells, there is no increased occupancy of the phoA and tagA promoters by WalR, nor does PhoP occupy the promoters of any of the four genes, consistent with expectations from other studies. In phosphate-limited cells, PhoP occupies the phoA and tagA promoters, but WalR does not occupy either promoter in such cells as expected (Fig. 8B). In general promoter occupancy was higher for negatively regulated (iseA and tagA) than for positively regulated (ydjM and phoA) genes (Fig. 8). Importantly in phosphate-limited cells, WalR and PhoP both occupy the ydjM and iseA promoters at levels significantly higher than observed in ‘mock’ precipitations (Fig. 8B). Some caution is warranted, however, because ydjM promoter occupancy by WalR and PhoP is close to the significance threshold and phosphorylation of PhoP (in phosphate-limited cells) does not result in significantly increased occupancy of ydjM and iseA promoters as is observed with the two promoters (phoA and tagA) of the PhoPR regulon (Fig. 8B). We conclude that PhoP is bound to the promoters of WalR regulon genes at a low but significant level in phosphate-limited cells. Moreover, WalR is bound at elevated levels to the ydjM and iseA promoters in phosphate-limited cells, but it is not bound to phoA or tagA promoters in these cells.
Cell wall metabolism is a major component of the B. subtilis response to phosphate limitation. Teichoic acid is a phosphate-rich polymer that comprises about 40% of the walls of exponentially growing cells and this is replaced by teichuronic acid, a non-phosphate-containing polymer, when phosphate becomes limiting. This transition is part of the PhoPR-mediated phosphate limitation response in B. subtilis: PhoP∼P represses teichoic acid synthesis while activating teichuronic acid synthesis in phosphate-limited cells (Hulett, 2002). How synthesis of peptidoglycan, the second major structural component of cell walls, is regulated in phosphate-limited cells and how its synthesis is coordinated with the teichoic/teichuronic acid transition is poorly understood. Here we show that both WalRK and PhoPR TCS are involved in control of peptidoglycan metabolism in phosphate-limited cells and present evidence that coordinate regulation of peptidoglycan and anionic polymer synthesis may derive from ancestral features and capabilities retained by these proteins after their divergence.
A crucial finding of this study is that phosphate-limited cells continue to divide at a reduced rate, remain viable for an extended period in a phosphate-limited state and rapidly resume growth when phosphate becomes available. We show that the rate of peptidoglycan synthesis in phosphate-limited wild-type cells is ∼27% the rate of that of exponentially growing cells. Thus the teichoic/teichuronic acid transition that occurs in the cell walls of phosphate-limited cells is accompanied by a high rate of peptidoglycan synthesis. Moreover, in phosphate-limited AH024 (ΔphoPR) cells that do not detectably divide or undergo the teichoic/teichuronic acid transition, peptidoglycan synthesis occurs at approximately one-sixth (∼18%) the rate of exponentially growing wild-type cells. Together these data show that cells in the phosphate-limited state have an active cell wall metabolism, even in the absence of a PhoPR-mediated response, and are capable of rapid resumption of growth when phosphate becomes available.
Bisicchia et al. (2007) showed that the WalRK controls peptidoglycan metabolism in exponentially growing cells by regulating expression of autolysins (YocH, CwlO, LytE), a cell wall-associated protein (YdjM) and modulators of autolysin activity (IseA and YjeA). Here we show that the yocH, cwlO, lytE and ydjM genes are also expressed in phosphate-limited cells and that yocH and ydjM expression is dependent on the WalR DNA binding sequence. Crucially, we show that deleting these genes individually results in decreased rates of peptidoglycan synthesis in phosphate-limited cells. Collectively these data show that peptidoglycan synthesis in exponentially growing and phosphate-limited cells requires similar activities. This prompted us to examine the roles of WalRK and PhoPR in controlling peptidoglycan metabolism in phosphate-limited cells and to establish whether similarities between these two-component systems are important for this process. Several observations are pertinent to addressing these issues. First, the WalR and PhoP DNA binding sequences are the complement of each other (Fig. 6A) and promoters of genes from both regulons can be bound by both proteins (Fig. 6B). In effect the WalR and PhoP binding sequences are in opposite orientations within their respective cognate promoters. Second, the WalR and PhoP proteins are both present throughout growth and phosphate limitation – in fact WalR and PhoP proteins are present at comparable levels in phosphate-limited cells (Fig. 5B; Howell et al., 2006). Nevertheless, WalK activation is thought to the confined to exponentially growing cells with septal location playing an important role (Fukushima et al., 2008) while PhoR activation is confined to phosphate-limited cells (Hulett, 2002). Third, the ′PhoR kinase has the ability to phosphorylate WalR in vitro, but ′WalK cannot phosphorylate PhoP (Howell et al., 2006). These observations suggest several possible mechanisms by which yocH, cwlO, lytE and ydjM expression may be controlled in phosphate-limited cells and why their expression levels are reduced in this condition (Fig. 3): (i) control is exerted by WalRK alone, with reduced expression reflecting diminished WalK activation due to a decreased growth rate, (ii) control is exerted by WalR that is phosphorylated by activated PhoR, with reduced expression reflecting lower WalR phosphorylation by the less efficient non-cognate PhoR kinase, (iii) control is exerted by PhoPR alone with reduced expression reflecting the lower affinity of PhoP∼P binding to WalR regulon promoters or (iv) by some combination of these mechanisms. Four approaches were employed to establish whether the WalRK or PhoPR TCS activates yocH, cwlO, lytE and ydjM expression in phosphate-limited cells. Genetic analysis, comparing yocH, cwlO, lytE and ydjM expression in phosphate-limited cells of strains 168 (wild-type) and AH024 (ΔphoPR), shows that their expression is reduced two- to three-fold when phoPR is deleted. While this result is consistent with PhoPR-mediated control of expression, it is qualified by preliminary transcriptome analysis showing that expression of many genes not directly regulated by PhoPR decreases by a similar order of magnitude (K. M. Devine et al., unpublished). Gel shift analysis revealed that WalR and PhoP can bind to the promoters of all genes tested from each regulon (Fig. 6B). Again this is suggestive of PhoPR-mediated control of their expression. However, a possible caveat to this conclusion emanates from the unusual relationship between the WalR and PhoP DNA binding sequences (similar sequence in inverted orientation), such that binding of the proteins to these promoters may reflect a physical capability in vitro that does not have a physiological relevance. The in vitro transcription analysis shows that PhoP∼P can activate ydjM expression approximately sixfold compared with eightfold activation by WalR∼P. However, PhoP∼P is unable to repress iseA expression in vitro (Fig. 7). The specificity and level of ydjM transcription in vitro are perhaps the strongest and most unequivocal evidence supporting PhoPR-mediated control of WalRK regulon genes in phosphate-limited cells. Interestingly, non-phosphorylated WalR is an effective repressor of iseA transcription in vitro (Fig. 7). The ChIP/qPCR analysis shows that ydjM and iseA promoter fragments are enriched in samples harvested from phosphate-limited cells of strains expressing tagged WalR and tagged PhoP (Fig. 8). Again this supports the view that ydjM expression is controlled by PhoP in vivo. Possible caveats to this conclusion arise from the observations that PhoP occupancy of the ydjM promoter is close to the significance threshold, that PhoP phosphorylation does not lead to increased occupancy of the ydjM promoter and that PhoP binds to the iseA promoter at a similar level but does not repress its expression in vitro. Two additional observations are noteworthy: (i) WalR and PhoP occupancy was higher at promoters they repressed than those they activated and (ii) WalR does not bind to phoA or tagA promoters even when phosphorylated. The most parsimonious interpretation of these data, despite the attendant caveats, is that both WalRK and PhoPR regulate ydjM expression in phosphate-limited cells. However, determining their relative contributions requires further investigation, a difficult task in view of WalRK essentiality. Importantly control of iseA expression differs from that of ydjM in several respects: (i) non-phosphorylated WalR is an effective repressor of iseA transcription in vitro and (ii) PhoP∼P binds to the iseA promoter in vivo but is incapable of repressing its expression in vitro. These observations may be rationalized on the basis that IseA is an autolysin antagonist (Salzberg and Helmann, 2007; Yamamoto et al., 2008), an inappropriate activity in phosphate-limited cells where cell division and peptidoglycan synthesis occur at significant rates. Therefore, iseA repression by non-phosphorylated WalR may be required to compensate for the fact that the level of phosphorylated WalR∼P is reduced in phosphate-limited cells (due to decreased growth rate) and that phosphorylated PhoP∼P cannot repress iseA transcription. A further implication might be that because WalR∼P levels are reduced in phosphate-limited cells, expression of the YocH, CwlO and LytE autolysins might be lower than that required for peptidoglycan metabolism, requiring their expression to be augmented by the activated PhoPR TCS.
This study reveals further linkage between the WalRK and PhoPR TCS in B. subtilis and how they control cell wall metabolism in phosphate-limited cells. The many similarities between these TCS points to a common ancestry, one that is shared with the PhoBR TCS of Escherichia coli that controls the response of this bacterium to phosphate limitation (see Lamarche et al. 2008 for review). This relationship with the E. coli homologue is illustrated by (i) the consensus DNA binding sequence of PhoB from E. coli is the same as the WalR DNA binding consensus from B. subtilis (Blanco et al., 2002) and (ii) that WalR expressed in E. coli can repress expression of phoA, a member of the PhoBR regulon (Okajima et al., 2008). The roles of WalRK and PhoPR in B. subtilis have diverged with the former assuming a primary role in peptidoglycan metabolism in exponentially growing cells while the latter functions to execute the anionic polymer transition and direct synthesis of phosphate scavenging activities. In this study we have presented evidence showing that PhoPR also has the capability to regulate peptidoglycan metabolism, to augment and perhaps coordinate peptidoglycan with anionic polymer metabolism in phosphate-limited cells.
Bacterial strains and growth conditions
Bacterial strains used in this study are listed in Table 1. E. coli strain TG-1 (Sambrook et al., 1989) was used for routine cloning and E. coli strain BL21(DE3) (Studier and Moffatt, 1986) for protein overexpression. Strain EC101 (repA+) (Law et al., 1995) was used for propagating plasmid pG+host4 in E. coli. E. coli strains were grown in LB medium (Sambrook et al., 1989). B. subtilis was grown in LB, high phosphate-defined medium (HPDM) or low phosphate-defined medium (LPDM) (Muller et al., 1997). Growth profiles of wild-type strain 168 and strain AH024 were established by growing cells in HPDM overnight at 37°C and then diluting them to an OD600 of 0.08 in 100 µl of LPDM medium in a 96 well culture plate (CELLSTAR®, Greiner Bio-one). The plate was incubated at 37°C with constant shaking (slow) in a Biotek Synergy2™ for 8 h. Four hours after phosphate starvation (T4), LPDM has been converted to HPDM by the injection of 5 µl of 73 mM phosphate buffer solution, raising the phosphate concentration to 3.65 mM in the medium. OD600 was measured every 10 min throughout the experiment. Antibiotics were added to the medium at the following concentrations per ml as appropriate: ampicillin 100 µg; kanamycin 50 µg; erythromycin 150 µg. For B. subtilis, antibiotic concentrations were as follows: tetracycline 13 µg; kanamycin 10 µg; erythromycin 1 or 3 µg; chloramphenicol 3 µg; spectinomycin 100 µg, lincomycin 25 µg.
Plasmids and oligonucleotide primers used in this study are listed in Tables 1 and 2 respectively. Standard procedures were used for DNA manipulations (Sambrook et al., 1989). The plasmid pG+host4 was used to mutate the WalR binding boxes in the yocH and ydjM promoters at their respective chromosomal loci (Biswas et al., 1993). The mutations replaced the one hexameric binding-motif with the HindIII sequence to facilitate identification of the desired clones. Two fragments were amplified for each promoter using primer pairs yocHPROM-1/YOCH4R and YOCH4F/yocHPM-2 with template pBP058 for yocH and primer pairs oBP106/YDJM4R and YDJM4F/oBP107 with template pBP058 for ydjM. The fragments were fused by strand overlap extension and cloned into EcoRI-BamHI-digested pG+host4. The resulting plasmids, pEL59 (carrying mutation yocH4) and pEL60 (carrying mutation ydjM4), were propagated in E. coli repA+ (Law et al., 1995). These plasmids were integrated into the B. subtilis chromosome as described by Biswas et al. (1993). Briefly, B. subtilis was transformed using the standard protocol (Anagnostopoulos and Spizizen, 1961) and single cross-over integrants were confirmed by PCR (strains EL61 and EL62 carrying integrations at the yocH and ydjM loci respectively). Overnight cultures of the transformants were grown in the absence of antibiotic at 37°C. The next day the overnight culture was diluted 1/50 and grown at 28°C to promote plasmid excision. After 4 and 8 h of growth at 28°C serial dilutions of each culture were plated without antibiotic selection. Colonies were then transferred to LB and LB+Em150 µg ml−1 plates to screen for erythromycin-sensitive clones that are the required excisants. The desired excisants containing the WalR box mutations were identified by digestion of PCR-generated promoter fragments with HindIII. Both mutations were confirmed by sequencing of the PCR fragments. The resulting strains EL63 and EL64 carry mutations yocH4 and ydjM4 in the yocH and ydjM promoters respectively.
Strains LSB002 (PphoP-phoP-3xFLAG) and LSB003 (PwalR-walR-3xFLAG) containing C-terminally 3xFLAG-tagged copies of phoP and walR under control of their native promoters were generated by transforming 168WT with plasmids pLIS003 and pLIS001 respectively, and selecting for erythromycin plus lincomycin [MLS (macrolide-lincomycin-streptogramin B)]-resistant transformants. PCR amplification confirmed correct integration of the plasmids. Plasmids pLIS003 and pLIS001 were constructed as follows: the 400 nucleotides at the 3′-end of the phoP and walR genes (without the stop codons) were amplified using primers LIS3 & LIS4 (phoP) and LIS5 & LIS6 (walR). These fragments were ligation-independently cloned into pMUTIN-SPA-LIC (gift of Mark Fogg); 0.2 pmol of each fragment was treated with T4 DNA polymerase and 2.5 mM dTTP, while the pMUTIN-SPA-LIC plasmid was linearized using AscI, gel purified and treated with T4 DNA polymerase and 2.5 mM dATP. Mixes of 10 ng of T4 treated vector and 30 ng of T4 treated insert were used to transform E. coli TG-1 (Aslanidis and de Jong, 1990; Bonsor et al., 2006; Fogg and Wilkinson, 2008). The resulting plasmids were extracted from E. coli and sequenced prior to transformation into B. subtilis.
Measurement of incorporation of 14C-labelled N-acetylglucosamine into B. subtilis cell walls
Bacillus subtilis cells were cultured in LPDM medium at 37°C and 220 r.p.m. Cell wall synthesis of exponentially growing cells was measured by growing cells until the OD600 had reached a value of 0.1, and then transferring 8 ml of culture to a screw cap plastic flasks containing 20 µl of 14C-labelled N-acetylglucosamine (7.4 MBq ml−1; Amersham) and 12.8 µl of unlabelled N-acetylglucosamine (5 ng ml−1; Sigma Chemical) to give a final concentration of 6.4 mg l−1. Growth was monitored turbidimetrically at OD600 for 2 h and 400 µl of samples of the cell suspension was harvested in duplicate every 15 min. Samples were then processed as previously described (Bisicchia et al., 2007). Cell wall synthesis of phosphate starved cells was measured as outlined above with the radioactivity substrate being added 3 h after the onset of phosphate limitation, and samples were harvested every 20 min.
Measurement of autolysis rates of exponentially growing and phosphate-limited B. subtilis cells
Bacillus subtilis strains were grown in parallel in LPDM medium in shake flasks at 37°C and 220 r.p.m. either until early log-phase (OD600 = 0.5), or until 2 h after the outset of the phosphate-starvation-induced stationary phase. Cells were then washed once in 10 mM Tris-HCl (pH 8.0) and resuspended in an equal volume of 10 mM Tris-HCl (pH 8.0). The OD600 of different cultures was adjusted to the same original value (OD600 = 0.5 for exponentially growing cells and OD600 = 1.1 for phosphate starved cells) and in order to perform three technical replicates for each experimental condition, nine 90 µl aliquots of each culture were transferred to a Microtest Tissue Culture 96 wells Plate (Falcon). For measurement of autolysis rates, 10 µl of 10 mM Tris-HCl (pH 8.0) was added to three of these aliquots, to give a final volume of 100 µl. The plate was incubated at 37°C in a plate reader (Biotek synergy 2) with slow orbital shaking, and cell lysis was monitored by measuring the decrease in OD600 at 10 min intervals. Results showed a very high level of reproducibility among the three technical replicates for each condition.
His6x-tagged WalR′, WalK, PhoP and ′PhoR were purified from strain BL21(DE3) as previously described (Howell et al., 2006). The intracellular region of WalK was cloned in NdeI/BamHI-digested pET21b after amplification with primers YYCG5CNPET and YYCGPETbamh6 to generate plasmid pDN1101. PhoP was cloned in NdeI/XhoI-digested pET21b after amplification with primers PHOP5PET and PHOP3PET to generate plasmid pDN1102. The intracellular region of PhoR was cloned in NcoI/XhoI-digested pET21d after amplification with primers PHORPETF2 and PHOR3PET to generate plasmid pDN1103.
Gel mobility shift DNA binding assays
DNA fragments spanning the promoter regions of yocH (yocH-biotin/yocH PM-2), yvcE (YvcE-biotin/YvcE-GSR), ydjM (YdjM-biotin/YdjM-GSR), yjeA (YjeA-biotin/oBP55), lytE (oBP165-biotin/oBP177), iseA (YoeB-biotin/YoeB-GSR), phoA (PhoA-biotin/PhoArev1), phoPR (PhoPR-biotin/PhoPRrev1) were amplified from B. subtilis 168 chromosomal DNA by PCR using Phusion polymerase (NEB) and the indicated biotinylated oligonucleotide pairs. Biotinylated fragments were gel purified according to standard procedures. Binding assays were performed as follows: increasing concentrations of WalR (or PhoP) were incubated with 1 µM WalK (or PhoR) in the presence of 1 mM ATP and the phosphorylation reaction took place for 15 min at room temperature. Addition of the probe (2 ng) followed and the binding reactions took place also at room temperature for 30 min. Reactions were stopped by addition of loading dye and electrophoresis took place at 4°C under the conditions described in Bisicchia et al. (2007). All reactions contained 0.5 µg of poly-[d(I-C)] as competitor DNA.
Strains used to establish transcription profiles by Northern analysis were processed as described previously (Howell et al., 2003). Briefly, overnight cultures of the strains were grown in HPDM and the following day inoculated and grown in freshly prepared LPDM. Samples were collected throughout growth to monitor gene expression. RNA isolation, electrophoresis, blotting and hybridization were performed as detailed in Howell et al. (2003). DNA probes for detection of yocH, ydjM, iseA, cwlO, lytE and phoA were synthesized according to the DIG-labelling protocol (Roche) using B. subtilis 168 chromosomal DNA as template and primer pairs YOCH-PMA/YOCH-PM2, YDJM-GSFw/oBP064, YOEB PROBE-1/YOEB PROBE-2, oBP061/oBP062, LYTEF1/LYTER1 and IJ99F/IJ99R respectively. Twenty-five micrograms of total RNA was used in each reaction.
RNA for qPCR analysis was isolated as previously described (Howell et al., 2003) with the addition of a DNase treatment step, 1.5U RQ1 DNase (Promega) for 15 min at 37°C followed by a final acid phenol : chloroform (5:1) extraction and precipitation. cDNA synthesis was undertaken with the Transcriptor cDNA synthesis kit (Roche) using 1 µg DNase-treated total RNA and random primer following the recommendations of the manufacturer. Parallel reactions lacking reverse transcriptase were carried out and used as templates to ensure successful DNA removal. qRT-PCR analysis of transcript levels was undertaken using the Lightcycler 480 SYBR I green master kit from Roche on a Lightcycler 480 instrument. All reactions were set up at least 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. Primers were designed using PrimerExpress 3.0. A melt curve cycle was included for every primer pair to check the specificity of each amplification. Relative expression ratios were calculated using the 2–ΔΔCT method (Livak and Schmittgen, 2001).
In vitro transcription
In vitro transcription reactions were performed as per Paul et al. (2004) with the following modifications: 5 pmol of WalR and PhoP response regulator proteins and ′WalK and ′PhoR histidine kinases were added to reactions as appropriate; the concentration of ATP was 50 µM; 0.1 pmol gel purified linear template was added to each reaction. Templates for ydjM and iseA reactions were generated using primers ydjMshortF/ydjMlongR and yoeBELF1/yoeBELR1 respectively. Core E. coli RNA polymerase [Epicentre Biotechnologies (Madison, WI)] was used at 2 pmol per reaction and was combined with purified B. subtilisσA at a molar ratio of 1:4 (8 pmol) to form holoenzyme on ice for 15 min, prior to addition to the reaction (final volume 20 µl). Reactions were incubated for 15 min at 37°C and stopped by addition of 10 µl stop buffer. Reactions were electrophoresed through 8.3 M urea-6% polyacrylamide gels. The results were detected and analysed by autoradiography or using a phosphoimager and run-off transcripts were quantified utilizing a Fuji FLA-3000 imager and Fuji Multi Gauge software (V2.2).
Western blot analysis
Cells were grown in LPDM with the addition of 1 µg ml−1 erythromycin plus 25 µg ml−1 lincomycin (MLS) and 100 µM IPTG at 37°C. Samples were harvested by centrifugation at 7100 g for 2 min at 4°C. Cell pellets were resuspended in 100 ml lysis buffer (10 mM Tris pH 8, 0.5 mM EDTA, 100 µg ml−1 lysozyme, 10 µg ml−1 DNase and Calbiochem protease inhibitor cocktail set III) and incubated at 37°C for 30 min. One hundred millilitres of 2× SDS-PAGE (100 mM Tris pH 6.8, 4% SDS, 20% glycerol) loading buffer along with one fiftieth volume of β-mercaptoethanol was added to the lysates, the contents were thoroughly vortexed then boiled for 10 min. Protein concentration was determined using a non-interfering protein assay kit (Calbiochem) using a standard curve generated with BSA. Samples of 30 mg (total protein) were separated on a 12% SDS-PAGE Tris-Glycine polyacrylamide gel and transferred to a PVDF membrane (Roche) by electroblotting. The membrane was blocked with 4.5% skim milk in Tris buffered saline (pH 7.5) plus 0.05% Tween, incubated with 1:7500 dilution of primary Monoclonal ANTI-FLAG M2 antibody produced in mouse (Sigma-Aldrich, F1804) for 2 h, washed and then incubated with 1:30 000 dilution of rabbit anti-mouse horseradish peroxidase antibody conjugate (Sigma-Aldrich, A9044) for 1 h. Protein–antibody complexes were detected with the SuperSignal West Dura Extended Duration substrate (Thermo Scientific, 34075).
Chromatin immunoprecipitation and qPCR
Cells were grown in LPDM with the addition of 1 µg ml−1 erythromycin plus 25 µg ml−1 lincomycin (MLS) and 100 mM IPTG at 37°C. Samples were harvested at mid-exponential phase (OD600 = 0.2; 200 ml) and at 2 h after the onset of phosphate starvation state (T2; 50 ml). Cross-linking was performed by the addition of formaldehyde to a final concentration of 1% followed by 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 RNaseA (final concentrations) and incubated for 30 min at 37°C, then the sample volume was 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 kb. Sonicated samples were centrifuged for 10 min at 20 000 g (4°C) to remove insoluble debris. Fifty microlitres of aliquot of the cleared lysates was removed to serve as the input sample, to which 50 µl of 10× buffer C (10 mM Tris-HCl pH 8, 10% SDS, 10 mM EDTA) and 400 µl buffer B was added and the sample was stored on ice. The remaining cleared lysates were each added to 40 µl anti-FLAG M2-agarose resin (Sigma-Aldrich; A2220) equilibrated with ice-cold buffer B (according to manufacturers' instructions) and incubated for 2 h at 4°C with mild agitation. Beads were then collected by centrifugation for 1 min at 5000 g (4°C) and 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 along with the input samples for ∼20 h at 65°C with shaking (950 r.p.m.). ChIP samples were collected by centrifugation for 1 min at 13 300 r.p.m. (4°C). DNA from both ChIP and input samples was purified using the QIAquick PCR purification kit (Qiagen, 28106), eluted in 50 µl H2O and treated with 0.02 mg ml−1 RNaseA (Roche, 11119915001) for 30 min at 37°C. RNase was eliminated using the QIAquick PCR purification kit and samples were eluted in 30 µl H2O. For quantitative PCR, input and ChIP samples were diluted 1:20 and 1:4 respectively. Two microlitres of DNA was added to 5 µl of LightCycler 480 SYBR Green I Master mix (Roche), 1 µl of each primer (5 pmol stock) and 1 µl H2O (primers listed in Table 2). qPCR was performed on a Roche Lightcycler 480 following manufacturers' instructions. Cp values were determined using the second derivative maximum method.
This work was supported by Science Foundation Ireland Principal Investigator Awards 03/IN3/B409 and 08/IN.1/B1859 to K.M.D. P.B. was partly supported by an IRCSET award and E.B. and S.H. are funded by BaSysBio Grant LSHG-CT-2006-037469. The authors would like to thank: Mark Fogg and Peter Lewis for the gift of plasmids; our colleagues Nathalie Pigeonneau and Olivier Delumeau (Philippe Noirot's laboratory) in the BaSysBio consortium, Amanda Daly and Gerard Brien for discussion and expert advice on ChIP; Peter Lewis, Xiao Yang and Elecia Johnston for advice on in vitro transcription. We also thank John Helmann, Gabriele Bierbaum, Colin Harwood and the anonymous reviewers for helpful discussion.