Autoinduction of Bacillus subtilis phoPR Operon Transcription Results from Enhanced Transcription from Eσ^A- and Eσ^E-Responsive Promoters by Phosphorylated PhoP

Abstract

The phoPR operon encodes a response regulator, PhoP, and a histidine kinase, PhoR, which activate or repress genes of the Bacillus subtilis Pho regulon in response to an extracellular phosphate deficiency. Induction of phoPR upon phosphate starvation required activity of both PhoP and PhoR, suggesting autoregulation of the operon, a suggestion that is supported here by PhoP footprinting on the phoPR promoter. Primer extension analyses, using RNA from JH642 or isogenic sigE or sigB mutants isolated at different stages of growth and/or under different growth conditions, suggested that expression of the phoPR operon represents the sum of five promoters, each responding to a specific growth phase and environmental controls. The temporal expression of the phoPR promoters was investigated using in vitro transcription assays with RNA polymerase holoenzyme isolated at different stages of Pho induction, from JH642 or isogenic sigE or sigB mutants. In vitro transcription studies using reconstituted Eσ^A, Eσ^B, and Eσ^E holoenzymes identified P_A4 and P_A3 as Eσ^A promoters and P_E2 as an Eσ^E promoter. Phosphorylated PhoP (PhoP∼P) enhanced transcription from each of these promoters. Eσ^B was sufficient for in vitro transcription of the P_B1 promoter. P₅ was active only in a sigB mutant strain. These studies are the first to report a role for PhoP∼P in activation of promoters that also have activity in the absence of Pho regulon induction and an activation role for PhoP∼P at an Eσ^E promoter. Information concerning P_B1 and P₅ creates a basis for further exploration of the regulatory coordination or overlap of the PhoPR and SigB regulons during phosphate starvation.

Inorganic phosphate (P_i) is the limiting nutrient for biological growth in the soil, the natural habitat of Bacillus subtilis. To thrive in this environment where P_i levels are often 2 to 3 orders of magnitude lower than levels of other required ions (29), B. subtilis has evolved complex regulatory systems for utilization of this limiting nutrient. At least three global regulatory systems are responsible for changes in gene expression upon phosphate deprivation. One set of genes is controlled either positively or negatively by the PhoP-PhoR two-component regulators, genes referred to as the Pho regulon genes (for review, see reference 12). Other genes that are induced upon phosphate limitation are dependent on SigB (1), an alternative stress sigma factor. A third class of genes is expressed under phosphate-limiting growth conditions that are independent of either SigB or PhoP-PhoR (1). The regulatory coordination between these three sets of genes is unclear, although up-regulation of certain Pho regulon genes has been reported in a sigB mutant strain (12, 33).

Pho regulon genes are the most extensively studied set of phosphate-regulated genes in B. subtilis. Identification of genes of known function that are directly regulated by PhoP-PhoR provides insight into one strategy B. subtilis may use to deal with conditions of limiting phosphate. A high-affinity P_i transport system (25, 34, 36) (PstS system) is induced for the uptake of inorganic phosphate, while a family of alkaline phosphatases, PhoA, PhoB, and PhoD (5, 6, 14, 15), are secreted whose activity may function to supply the decreasing P_i pool. Anionic cell wall polymer turnover (2) is controlled by PhoP-PhoR, as phosphorylated PhoP (PhoP∼P) directly represses tag genes (23, 35) that are required for synthesis of the high-phosphate anionic polymer, teichoic acid (27), and activates the tua genes (24, 35) responsible for synthesis of a non-phosphate-containing polymer (39), teichuronic acid, under phosphate-limiting conditions. One might say that B. subtilis carries its phosphate reserve on its back, as teichoic acid is turned over as the teichuronic acid replaces it. The secreted phosphodiesterases and phosphomonoesterases, PhoD, PhoB, and PhoA, are believed to have a role in the teichoic acid degradation, providing an additional phosphate supply for uptake via the PstS high-affinity transport system. Other genes that require PhoP-PhoR for activation that may be directly regulated by PhoP∼P include glpQ (1), encoding a glycerophosphodiesterase; glnQ (28), encoding a glutamine ABC transporter; ykoL (37), a peptide of unknown function; and additional genes of unknown function, yhaX, yhbH, yttP (33) and yycp, ydbH, and yjdB (28).

The Pho regulon response is controlled at two levels: at the level of phoPR operon transcriptional regulation and by the signal that results in autophosphorylation of PhoR and the subsequent activation of PhoP by phosphorylation via PhoR. Studies reported here focused on transcriptional regulation of the phoPR operon. Previous reports showed that the phoPR operon was expressed at low levels during phosphate-replete growth but was induced two- to threefold upon P_i limitation (16). That the induced transcription level of phoPR in the wild-type (WT) strain was dependent on the phosphate starvation signal and PhoPR suggested that the operon was autoregulated, perhaps directly. These data raised a question about the preinduction transcription of phoPR, as previously characterized Pho regulon promoters which are directly activated by PhoP (phoA, phoB, phoD, tuaA, and pstS) are silent in vivo under P_i-replete conditions. Further, neither artificially elevating phoPR transcription under phosphate-replete conditions via an inducible promoter (4) nor chromosomal mutation (A. Puri and F. M. Hulett, unpublished data) initiates the Pho response, presumably because the signal is missing. Studies reported here were initiated to determine if phoPR transcription were directly regulated by PhoPR and, if so, what mechanism accounts for expression of phoPR during P_i-replete growth when other Pho regulon promoters are silent.

Our data suggest that the phoPR operon is directly autoregulated by PhoP-PhoR. This regulation is accomplished by up-regulation of Eσ^A and Eσ^E promoters responsible for transcription of the phoPR operon. Two additional phoPR promoters are not PhoP regulated. This is the first report of PhoP activation of an Eσ^E promoter or of a role for PhoP in up-regulation of promoters that have some activity in the absence of Pho regulon induction.

MATERIALS AND METHODS

Bacterial strains and plasmids.

All strains and plasmids used in this work are listed in Table 1. Plasmids pSB5 and pSB38 were constructed by amplifying phoPR promoter regions from JH642 (pheA1 trpC2) chromosomal DNA using a 5′ primer containing an EcoRI site and one of two 3′ primers each containing a BamHI site. Primers FMH202 (5′-GTGAATTC⁻³⁰⁰TCATTGAACTTGAACTG⁻²⁸²-3′) and FMH079 (5′-GTGGATCC⁺⁹²GTAATGACATCATAGCCT⁺⁷⁵-3′) or FMH312 (5′-TTGGATC⁺²⁴CACAACTAAAATTTTCTTGTTC⁺³-3′) were used to amplify two phoPR promoter fragments that were each cloned into pCR2.1, creating pSB5 and pSB38, respectively. (Superscript numbers identify base pair positions 5′ [−] or 3′ [+] of the PhoP translational start site.) The phoPR promoter fragments in pSB5 (392 bp) or pSB38 (324 bp) were sequenced, released from the vector by BamHI and EcoRI digestion, and cloned into the EcoRI/BamHI sites upstream of the promoterless lacZ in pDH32 to create pSB40, which contains the full-length phoPR promoter fusion, and pSB39, which contains the same 5′ promoter sequence with a 3′ coding region deletion. pSB40 and pSB39 were linearized by Pst1 digestion, transformed into JH642 or MH5600 (phoP ΔEcoRI), selecting for Cm^r and screened for an amyE⁻ phenotype. Representative pSB40 transformants containing a single copy of the full-length phoP-lacZ promoter fusion at the amyE locus in JH642 or MH5600 (phoP ΔEcoRI) were called MH5562 and MH5565, respectively. Representative pSB39 transformants containing the 3′-truncated phoPR-lacZ promoter fusion in JH642 or MH5600 (phoP ΔEcoRI), were called MH5559 and MH5567, respectively. MH5580 was constructed by transforming chromosomal DNA containing the sigE::Erm^r from strain EU8701 into MH5562 and selecting for Erm-resistant transformants. MH6200 was constructed by transforming chromosomal DNA from PB344 (sigB::Spc^r) into MH5562 and selecting for Spc^r transformants.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Genotype	Source or reference
B. subtilis strains
JH642	pheA1 trpC2	J. A. Hoch
EU8701	pheA1 trpC2 ΔsigE::Ermr	C. P. Moran
PB344	trpC2 sigBΔ3::Spcr	C. Price
MH5636	pheA1 trpC2 rpoCΩ pYQ52 Cmr	Ying Qi
MH5654	pheA1 trpC2 rpoCΩ pYQ52 Cmr ΔsigE::Ermr	Ying Qi
MH5559	trpC2 pheA1 amyE::phoP39-lacZ Cmr	This study
MH5562	trpC2 pheA1 amyE::phoP40-lacZ Cmr	This study
MH5565	trpC2 pheA82::Tn917 LErphoP ΔEcoRI amyE::phoP40-lacZ Cmr	This study
MH5567	trpC2 pheA82::Tn917 LErphoP ΔEcoRI amyE::phoP39-lacZ Cmr	This study
MH5580	trpC2 pheA1sigE::ErmramyE::phoP40-lacZ Cmr	This study
MH6200	pheA1 trpC2 rpoCΩ pYQ52 Cmr ΔsigB3::Spcr	This study
Plasmids
pCR2.1	Ampr Kanr	Invitrogen
pSB5	Ampr Kanr full-length phoPR promoter in pCR2.1	This study
pSB38	Ampr Kanr full-length phoPR promoter without the coding region in pCR2.1	This study
pSB39	Ampr Cmr 326-bp BamHI/EcoRI fragment from pSB38 subcloned into pDH32	This study
pSB40	Ampr Cmr 396-bp BamHI/EcoRI fragment from pSB5 subcloned into pDH32	This study
pSP200	pET 16b::sigB Ampr	This study
pSP201	pET 16b::sigE Ampr	This study

Media and enzyme assays.

For phosphate starvation induction of Pho reporter enzymes, alkaline phosphatases (APases), or the phoPR promoter fusions, cells were cultured in low-phosphate defined medium (LPDM) as described previously (13). For sporulation induction conditions, the cells were grown in modified Schaeffer's sporulation medium with glucose (SSG) (21). β-Galactosidase specific activity was determined by the method of Ferrari et al. (9). β-Galactosidase specific activity was expressed in units per milligram of protein. The unit used was equivalent to 0.33 nmol of ortho-nitrophenol produced per min. APase specific activity was determined as previously described (13); the units were micromoles of p-nitrophenol produced per minute at 37°C.

RNA preparation and primer extension analysis.

Total RNA was isolated from B. subtilis cells grown in either LPDM or SSG medium. Two volumes of RNAprotect bacterial reagent (QIAGEN) was mixed with 1 volume of bacterial culture and incubated for 5 min at room temperature. The mixture was centrifuged at 5,000 × g for 10 min. The total RNA was extracted from the above pellet using the RNeasy Midi kit (QIAGEN). A total of 50 μg of RNA was used in each primer extension reaction mixture. The primer extension reactions were performed as described previously (5) using primer FMH079 (see Fig. 2). A sequencing ladder was produced by end labeling the primer FMH079 with [γ-³²P]ATP and with pSB5 as template using Sequenase (U.S. Biochemical Corp.) according to the instructions of the manufacturer.

FIG. 2.

Transcription start sites and PhoP binding sites on the phoPR promoter sequence and 5′ PhoP coding sequence. Gray shading identifies sequence protected by both PhoP and PhoP∼P. Stippled shading identifies sequence protected only by PhoP∼P. Transcriptional start sites for P_B1, P_E2, P_A3, P_A4, or P₅ are indicated by bold sequence base pairs that are identified by a bent arrow followed by the promoter number. The −10 consensus sequence for each promoter is underlined by a slender rectangle marked −10 followed by the promoter number. The −35 consensus sequence for each promoter is overlined by a slender rectangle marked −35 followed by the promoter number. The consensus repeats for PhoP dimer binding [TT(A/C/T)A(C/T)A] are underlined with the sequence in bold print. The translational start codon, ATG, is boxed and identified by a bent arrow marked +1. Sequence numbering is relative to the A of ATG as +1. Arrows with half arrowheads identify primers used to amplify sequences in the two promoter fusion constructs analyzed below in Fig. 3. The asterisk identifies the transcription site of P_BX1.

DNase I footprint assays.

The phoPR promoter fragment from pSB5 (Table 1) was digested with either BamHI, for the coding strand, or EcoRI for the noncoding strand and was end labeled with Klenow fragment in the presence of [α-³²P]dATP. The insert was then released by digestion with either EcoRI or BamHI. Purification of the probes and the DNase I footprinting experiments were performed according to the methods of Liu and Hulett (24). In each reaction mixture, 1.4 μg of a truncated form of PhoR (*PhoR) and various amounts of PhoP were used. A final concentration of 4 mM ATP was added for reactions requiring PhoP∼P. The concentration of PhoP in the reaction mixtures was 55 nM, 275 nM, 1.38 μM, and 6.7 μM.

Overexpression and purification of proteins. (i) σ^B.

A DNA fragment containing the entire coding region of σ^B was amplified by PCR using chromosomal DNA of JH642 as template. Oligonucleotide primers were FMH492 (5′-TGCATATG TTGATCATGACACAACCATCAAAAACT-3′) and FMH493(5′-ATGGATCCTTACATTAACTCCATCGAGGGATCTT-3′). These primers contained NdeI and BamHI sites, respectively. The PCR product was cloned into pET16b (Novagen) at the same sites, to generate pSP200. Escherichia coli BL21(DE3)pLysS cells containing pSP200 were grown in Luria-Bertani medium (1,000 ml) containing ampicillin (100 μg/ml) at 30°C. When the optical density at 540 nm was 0.6, isopropyl-β-d-thiogalactopyranoside (1 mM) was added to the culture and the cells were collected by centrifugation at 8,000 × g for 15 min after a 3-h incubation period. The pellet fraction was suspended in 30 ml of sonication buffer (50 mM Tris [pH 8], 500 mM NaCl, 5 mM MgCl₂, and 20% glycerol), to which 1 mM phenylmethylsulfonyl fluoride was added directly before the cells were disrupted by sonication and separated by centrifugation at 120,000 × g for 1 h at 4°C. The supernatant fraction was applied to a 2.5-ml nickel-nitrilotriacetic acid-agarose (QIAGEN) affinity column (the Ni-nitrilotriacetic acid resin was previously equilibrated with sonication buffer in a 1.0- by 10-cm Econo column [Bio-Rad]). The column was sequentially washed with the sonication buffer (20 times with 2.5 ml) followed by 30 mM imidazole in sonication buffer (twice with 2.5 ml) at 4°C. The bound protein was eluted using a stepwise imidazole concentration gradient from 100 to 500 mM in the sonication buffer at 4°C. The eluted proteins were dialyzed overnight against 2× storage buffer (10 mM Tris [pH 8.0], 10 mM MgCl₂, 100 mM KCl, 0.1 mM EDTA, and 50% glycerol) at 4°C. The protein concentration was determined with the Bio-Rad protein assay (Bio-Rad Laboratories) using bovine serum albumin as the standard.

(ii) σ^E.

A DNA fragment that contains the mature σ^E protein-coding region (sigE) without the N-terminal 27-amino-acid-coding region of pro-σ^E (20a) was generated by PCR using JH642 chromosomal DNA as template. The following primers were made with the restriction sites for NdeI and BamHI: FMH490 (5′-TGCATATGGGCGGGAGTGAAGCCCTGCCGCCTCCAT-3′) and FMH491(5′-CTGGATCCTTACACCATTTTGTTGAACTC-3′). The PCR product was cloned into pET16b at the same site (Novagen), generating pSP201. pSP201 was transformed into E. coli BL21(DE3) pLysS, and a representative transformant was used as a σ^E-overexpressing strain. The σ^A-overexpressing strain was provided by M. Fujita and Y. Sadaie. σ^E and σ^A were overexpressed and purified as described above.

(iii) PhoP and *PhoR.

PhoP and *PhoR were purified as previously described (22). *PhoR is a soluble, truncated form of PhoR (38).

(iv) RNAP and core polymerase.

B. subtilis MH5636 (34) or B. subtilis MH5654 was grown in either LPDM or SSG medium, and the RNA polymerase (RNAP) and the core polymerase were purified as described previously (34).

In vitro transcription.

Linear template DNA used in the in vitro transcription assays was released from pSB5 by EcoRI digestion, releasing a 409-bp DNA fragment containing the full-length phoPR promoter region. These DNA fragments were purified from a 1% agarose gel with a QIAquick gel extraction kit (QIAGEN) according to the manufacturer's directions. The transcription reaction mixture (20-μl final volume) consisted of a 2 nM concentration of template, various concentrations of PhoP or PhoP and *PhoR, 1 mM ATP, and 0.4 pmol of purified B. subtilis RNAP (34). The transcription buffer contained 100 mM potassium glutamate, 10 mM Tris (pH 8.0), 0.1 mM EDTA, 50 mM KCl, 1 mM CaCl₂, 5 mM MgCl₂, 10 μg of bovine serum albumin per ml, 1 mM dithiothreitol, and 5% glycerol. Either PhoP alone or a mixture of PhoP-*PhoR (equal molar) and ATP (1.0 mM) was incubated with the template at 37°C for 10 min. RNAP or the core polymerase containing required sigma factors was then added to the reaction mixture, and incubation continued at 37°C for 15 min. A single round of transcription was initiated by the addition of a transcription buffer containing ATP, GTP, and CTP at 100 μM each, 10 μM UTP, 5 μCi of [α-³²P]UTP (Amersham), and heparin at 50 μg/ml. After incubation at 37°C for 15 min, reactions were stopped by the addition of 10 μl of loading dye (7 M urea, 100 mM EDTA, 5% glycerol, 0.05% xylene cyanol, and 0.05% [wt/vol] bromophenol blue). Samples were analyzed on 8 M urea-6% polyacrylamide gels. Dried gels were analyzed by using a PhosphorImager (Molecular Dynamics).

RESULTS

PhoP or PhoP∼P binds to three sites in the phoPR promoter region and one site in the coding sequence for PhoP.

Previous data showed that induction of the phoPR operon upon phosphate-limited growth was dependent on PhoP and PhoR. DNase I footprinting experiments were performed to determine whether regulation of the phoPR operon by the PhoP-PhoR two-component system might be direct. Either PhoP (in the presence of *PhoR but the absence of ATP) or PhoP∼P (in the presence of *PhoR and ATP) protected multiple regions positioned similarly on the coding and noncoding strands (Fig. 1 and 2). Phosphorylated PhoP extended the PhoP-protected region primarily on the noncoding strand between two PhoP binding regions from −150 to −213 and directly 5′ of the PhoP-protected region on the coding strand within the PhoP-coding sequence (+39 to +25). Only PhoP∼P protected a region on the coding strand between −9 and −22 or a region between −245 and −280 on the noncoding strand. All regions protected by both phosphorylated and unphosphorylated PhoP contained appropriately spaced (4 to 6 bp apart) repeated consensus sequences for PhoP dimer binding (6), TT(A/C/T)A(C/T)A (Fig. 2). The consensus repeats positioned 5′ of the coding region were on the noncoding strand, while the repeat within the coding region was on the coding strand. A number of DNase-hypersensitive sites were evident on the coding and the noncoding strands upon PhoP binding (Fig. 1).

FIG. 1.

DNase I footprint analysis of the phoPR promoter bound by PhoP and PhoP∼P. Various amounts of PhoP, incubated with *PhoR (1.4 μg) in the presence or absence of 4 mM ATP, were mixed with the 407-bp phoPR promoter labeled on either the coding or noncoding strand and treated with DNase I. The concentration of PhoP used in each reaction mixture was, from left to right, 0 nM, 55 nM, 275 nM, 1.38 μM, and 6.7 μM. Lanes with PhoP∼P are labeled +ATP, and those with unphosphorylated PhoP are labeled −ATP. Lanes F, PhoP-free lanes; lane G, the G-sequencing reaction lane used as a reference. The thick black vertical lines represent the PhoP and PhoP∼P binding regions, while the thin lines represent sites bound only by PhoP∼P. The hypersensitive sites are marked with a dark arrowhead. Base pairs are numbered relative to the translation start site (as +1).

The PhoP binding site in the PhoP-coding region is required for full induction of the phoPR promoter during phosphate deprivation.

Three PhoP-activated Pho regulon promoters have secondary binding sites in addition to a core binding region between approximately −20 and −60, relative to the transcription start site, that binds two PhoP dimers. The secondary binding sites are located either >177 bp 5′ of the transcription start site (6) or 3′ within the coding region (25). In the phoD promoter, a 5′ secondary binding site was essential for 95% of the promoter function. Two other PhoP-activated promoters, phoA and pstS, had PhoP and/or PhoP∼P binding regions within the coding region of the activated gene that were required for full expression of either promoter. To assess the importance of the 3′ PhoP binding site for phoPR promoter expression, phoPR promoter activity in JH642 (parental strain, MH5562) or a phoP mutant strain (MH5565) containing a full-length phoP-lacZ promoter fusion was compared to that of a JH642 strain (MH5559) or a phoP mutant strain (MH5567) with a phoP-lacZ promoter fusion containing a deletion of the 3′ binding site, as shown in Fig. 2 (deletion of bp +25 to +92). Figure 3A shows low expression from the full-length phoPR promoter in JH642 (MH5562) during exponential growth under phosphate-replete conditions (1 to 4 h) followed by induction (5 to 8 h), initiated as the culture entered stationary phase due to P_i limitation. The same promoter fusion in the phoP mutant strain (MH5565) showed little induction upon phosphate limitation, but lacZ expression increased slightly during late stationary phase (10 to 12 h). Expression of the phoP promoter fusion with the 3′ truncation in JH642 (strain MH5559) or in the phoP mutant background (MH5567) was reduced >5-fold compared to the full-length promoter in JH642, indicating the importance of this PhoP binding site within the PhoP coding sequence to phoPR operon promoter function during phosphate starvation.

FIG. 3.

The roles of PhoP or PhoP binding regions in phoPR transcription differ during phosphate-limited and phosphate-replete growth. (A) Growth and phoPR expression in LPDM. An arrow marks the induction of APase in phoP⁺ strains MH5562 and MH5559. (B) Growth and phoPR expression in SSG. Filled symbols represent growth; open symbols represent expression of various phoP-lacZ promoter fusions. Circle, MH5562 (JH642 strain; phoP-lacZ); square, MH5565 (phoP strain; phoP-lacZ); triangle, MH5559 (JH642 strain; phoP-lacZ fusion containing a deletion of bp +25 to +92 that removed the 3′ PhoP binding site, phoP_Δ25-92-lacZ); inverted triangle, MH5567 (Δ phoP; phoP_Δ25-92-lacZ); diamond, MH5580 (sigE phoP-lacZ).

The same four strains plus a sigE mutant strain (MH5580) containing the full-length phoPR-lacZ promoter fusion were cultured in a high-phosphate medium (SSG; 43 mM P_i), which was designed to induce sporulation and development, to assess post-exponential phoP promoter expression independent of P_i limitation, a condition where PhoP would be predicted to be unphosphorylated. The phoP-lacZ expression pattern (Fig. 3B) from the full-length promoter fusion, either in the parent strain (MH5562) or in the phoP mutant strain (MH5565) was similar. Expression in either strain was low during the first 7 h of growth, followed by a threefold induction that peaked between 9 and 10 h. Because there was no difference in the β-galactosidase accumulation in the phoP⁺ versus the phoP strain, it would appear that there is no significant role for PhoP in phoPR transcription under these conditions. The full-length promoter fusion in the sigE mutant background (MH5580) failed to induce during late stationary growth. Induction of the 3′-truncated phoP promoter-lacZ fusion in the phoP mutant strain (MH5567) was similar to that of the complete promoter, but expression was reduced in JH642 (MH5559), suggesting a possible repressor role for the unphosphorylated PhoP. Expression of the 3′-deleted promoter in either the WT or Δ phoP background was reduced during the first 7 h compared to the full-length promoter.

The difference in PhoP requirement under different phoPR induction conditions might be explained by multiple promoters, as was determined for phoB (encoding APase B), which was shown to have a vegetative promoter that required PhoP under P_i-limiting growth conditions and a second promoter for induction during sporulation (5).

The phoPR operon is transcribed from multiple promoters.

Primer extension was performed to identify the promoter(s) responsible for expression of the phoPR operon. Figure 4A shows the results of the primer extension analysis on RNA isolated during Pho regulon expression under phosphate starvation conditions. Three 5′ ends (labeled P₁, P₃, and P₄) were identified (Fig. 4A, lane 1) by using RNA isolated from cells approximately 1 h after phoPR induction, T₁. An additional 5′ end (P₂) was observed (Fig. 4A, lane 2) by using RNA isolated from cells 3 h into phosphate starvation induction, T₃. The concentration of P₂ increased relative to P₃ and P₄ concentrations in RNA from cells 4 h after phoPR induction (Fig. 4A, lane 3), while P₁ continued to increase but remained the least abundant of the 5′ ends. Because we show below that the same 5′ ends were found in vitro using purified RNAP, we will refer to them as transcription start sites.

FIG. 4.

Primer extension identified four mRNA 5′ ends in the phoPR promoter region. (A) Primer extension analysis of the phoPR promoter region. The end-labeled primer FMH079 was annealed to RNA from transition or post-exponential-stage cultures. (A) Lane 1, RNA isolated from LPDM-grown cells during early induction (T₁); lanes 2 and 3, RNA isolated from LPDM-grown cells during later Pho induction (T₃ and T₄); lanes T, C, G, and A, sequencing ladders generated by annealing the same end-labeled primer to a plasmid (pSB5) containing the full-length promoter region of the phoPR operon and extending it with Sequenase (U.S. Biochemical Corp.). Arrowheads labeled P_B1, P_E2, P_A3, and P_A4 identify the mRNA 5′ ends. (B) Comparison of phoP 5′ ends in RNA isolated from postexponential cells grown under phosphate starvation or phosphate-replete sporulation conditions. Lane 1, RNA isolated from LPDM-grown cells during late Pho induction, T₄; lane 2, DNA isolated from SSG-grown cells at sporulation stage T₄. (Labeling is as in panel A.) (C) Putative promoter −10 and −35 consensus regions for P_B1, P_E2, P_A3, and P_A4 compared to sigma factor consensus sequences (10). Bold letters in the P_B1, P_E2, P_A3, and P_A4 sequences represent matches to the sigma binding consensus sequence. In the consensus sequence, capital letters indicate highly conserved positions and lowercase letters indicate less-conserved positions. R = A or G; W = A or T.

To explore the transcriptional regulation of the phoPR operon under sporulation conditions, we performed primer extension analysis of the phoPR operon with total RNA isolated from post-exponential-stage cells grown in SSG at sporulation stage 4, T₄. Under sporulation conditions, the major 5′ end for the phoPR operon (Fig. 4B, lane 2) was identical to the above P₂ promoter identified in RNA from cells that entered stationary phase due to P_i starvation (above and Fig. 4B, lane 1). A low concentration of P₁ was also observed.

The transcription start sites P₁, P₂, P₃, and P₄ are located −23, −34/−37, −48/−49, and −69 bp upstream of the translational start site (ATG), respectively (Fig. 2). The −10 and −35 regions of each promoter were analyzed for sequence similarity to established sigma factor binding consensus sequences (10). Sequence alignments (Fig. 4C) provided putative promoter assignments for P₁, P₂, P₃, and P₄ as σ^B, σ^E, σ^A, and σ^A, respectively. Hereafter, we refer to the four promoters as P_B1, P_E2, P_A3, and P_A4. Putative −10 and −35 sequences for each promoter are indicated in Fig. 2.

Temporal expression of the phoPR promoters investigated using in vitro transcription assays with RNAP isolated at different times during induction.

RNAP was purified from B. subtilis (MH5636, His-tagged rpoC strain) grown in LPDM as the cells transitioned from exponential growth to stationary phase at T₀ and 3 or 4 h later (T₃ and T₄), or from strain MH5654 (sigE rpoC His tagged) at stage T₄. In vitro transcription reactions were carried out with each RNAP in the presence of PhoP∼P (Fig. 5A). The in vitro transcription pattern differed considerably depending on the stage of growth of the cells from which the RNAP was isolated (Fig. 5A).

FIG. 5.

Growth stage-specific RNAP shows temporal expression of in vitro phoPR promoter transcripts and the absence of P_E2 with RNAP from a sigE mutant strain. (A) In vitro transcription of the phoPR promoter with RNAP isolated from stage T₀, T₃, and T₄ cells grown in LPDM. The in vitro transcription reactions were carried out as described in Materials and Methods. M, RNA marker. In vitro transcripts were generated using RNAP from LPDM-grown MH5636 cells harvested at T₀ (lane 1), T₃ (lane 2), or T₄ (lane 3) or from MH5654 (sigE) cells at T₄ (lane 4). All reaction mixtures contained 5 pmol each of PhoP and *PhoR plus 1 mM ATP. (B) In vitro transcription products identified by primer extension. Markings and procedures were the same as for Fig. 4A, except the mRNA was generated by in vitro transcription (panel A, lanes 3 and 4).

Primer extension (Fig. 5B) was used to identify the T₄ RNAP (WT or sigE) in vitro transcript start sites (Fig. 5A, lanes 3 and 4). P_A4 and P_B1 transcripts were identified in the in vitro-generated mRNA using T₄ RNAP isolated from the sigE deletion strain. P_B1 and P_E2 were identified in the T₄ RNAP from the WT strain. Both in vitro-generated mRNAs identified a transcription start site (P_Bxl) not observed in total RNA from cells cultured under the conditions previously tested. The P_BX1 transcription start site and putative −10 and −35 sequences for σ^B are indicated in Fig. 2.

The in vitro-generated P_A4 and P_A3 transcripts (Fig. 5A, lane 1) decreased in reactions using later-stage T₃ or T₄ RNAP from a WT strain (lanes 2 and 3) but were most prominent (Fig. 5A, lane 4) in the reaction using stage T₄ sigE RNAP. Conversely, the P_E2 transcript that was absent in the reaction using WT T₀ RNAP (Fig. 5A, lane 1) was apparent in T₃ RNAP reactions and increased dramatically in reactions using stage T₄ WT RNAP (Fig. 5A, lane 3). P_E2 was not transcribed by the T₄ RNAP missing the SigE subunit (Fig. 5A, lane 4), suggesting that P_E2 is dependent on SigE (directly or indirectly). The quantity of both the P_B1 and P_Bxl transcripts increased with later-stage RNAP but showed no difference with RNAP isolated from WT or sigE stage T₄ cells (Fig. 5A, lanes 3 and 4), suggesting that the sigE mutation did not affect the form of RNAP required for their transcription. Thus, the number of transcripts obtained varied, as did the relative concentration of each transcript, depending on the growth stage and the strain from which the RNAP was isolated.

Promoters P_E2 and P_A4 require phosphorylation of PhoP (PhoP∼P) for maximum expression.

To determine the role of PhoP and PhoP∼P in transcription from P_B1, P_E2, P_A3, and P_A4, in vitro transcription reactions were done using the full-length promoter as template and WT T₄ RNAP or sigE T₄ RNAP in the absence of PhoP or with varying concentrations of PhoP or PhoP∼P. Figure 6A shows the results of the in vitro transcription using WT T₄ RNAP. Lanes 1 and 5 showed that significant amounts of P_E2 and P_xl transcripts were generated in the absence of PhoP. Reaction mixtures with increasing PhoP concentrations from 1 to 5 pmol (Fig. 6A, lanes 2 to 4) indicated that these concentrations of PhoP did not significantly affect transcription from P_E2 and P_Bxl. Similar reactions that included *PhoR and ATP for phosphorylation of PhoP (lanes 6 to 8) indicated that PhoP∼P (1 to 5 pmol) enhanced P_E2 transcription but not transcription of P_BXl.

FIG. 6.

PhoP∼P enhances transcription from P_E2 and P_A4. (A) PhoP phosphorylation affects RNAP T₄ phoPR promoter transcript P_E2. Symbols are the same as for Fig. 5. Lanes 1 and 5 contain no PhoP; lanes 2 to 4 and lanes 6 to 8 contain increasing concentrations of PhoP (1, 2.5, and 5 pmol). For phosphorylation of PhoP, equal molar concentrations of PhoP and *PhoR were in reaction mixtures applied to lanes 6 to 8. (B) T₄ RNAP from a sigE strain yielded P_B1, P_A3, and P_A4 transcripts but no P_E2 transcript; PhoP∼P enhanced P_A4 transcription. Lane 1 contains an in vitro transcription reaction identical to lane 8 in panel A for a direct comparison between reactions using T₄ RNAP from JH642 and from a mutant strain. Lanes 2 and 5 contain no PhoP; lanes 3 and 4 and lanes 6 and 7 contain increasing concentrations of PhoP (2.5 and 5 pmol, respectively). Lanes 6 and 7 each contain equal molar amounts of PhoP and *PhoR plus ATP for phosphorylation of PhoP.

Similar experiments were carried out with sigE T₄ RNAP (Fig. 6B) to examine P_B1, P_A3, and P_A4, as the data in Fig. 5A (lane 4) had shown the highest transcription levels of these promoters with that sample of RNAP. A control to mark the position of P_E2 and P_Bxl was included in lane 1 from a reaction mixture identical to that in Fig. 6A, lane 8. Lanes 2 and 5 contained transcripts generated from the phoPR promoter by sigE T₄ RNAP alone. Unphosphorylated PhoP (Fig. 6B, lanes 3 and 4) did not significantly affect transcription of P_B1, P_A3, or P_A4. PhoP∼P (2.5 to 5 pmol) increased the P_A4 transcript severalfold (Fig. 6B, lanes 6 and 7). P_B1 and P_A3 showed little enhanced transcription by PhoP∼P (Fig. 6B, lanes 6 and 7). PhoP∼P did not affect transcription from the P_BX1 promoter.

In vitro transcription using core RNAP plus purified sigma factors identifies σ^A, σ^B, and σ^E phoPR operon promoters.

Data from Fig. 4 and 5 suggested that the different phoPR promoters likely required different forms of RNAP holoenzymes for transcription. To reconstitute specific RNAP holoenzymes, B. subtilis sigma factors were expressed in E. coli and purified as described in Materials and Methods, and core polymerase was prepared from RNAP holoenzyme as described previously (34).

Figure 7A shows phoPR promoter transcripts generated using the reconstituted Eσ^A. An in vitro transcription reaction using core RNAP, PhoP∼P, and the phoPR promoter template yielded no transcripts (lane 6). The reaction with reconstituted Eσ^A (lanes 1 and 7) identified P_A4 and P_A3 as σ^A promoters. The P_A4 promoter showed enhanced transcription with increasing concentrations of PhoP∼P (lanes 8 to 13) but little change with unphosphorylated PhoP (lanes 2 to 5). The P_A3 promoter was very weak with PhoP (lanes 2 to 5) or without (lanes 1 and 7) but showed enhanced transcription with PhoP∼P (lanes 8 to 13).

FIG. 7.

PhoP∼P activates transcription from P_A4 and P_E2 by using different forms of RNAP. (A) Expression of P_A3 and P_A4 requires Eσ^A. Lane M contains the 200-nucleotide marker. Lanes 1 to 5 and 7 to 13 contain reconstituted Eσ^A. Lane 6 contains core RNAP with no sigma factor added. Lanes 1 and 7 contain no PhoP or PhoP∼P. Lanes 2 to 5 contain increasing amounts of PhoP (0.1 to 5.0 pmol). Lane 6 contains 5.0 pmol of PhoP∼P. Lanes 8 to 13 contain increasing amounts of PhoP∼P (0.1 to 5 pmol). Lanes 8 to 13 contain equal molar amounts of PhoP and *PhoR plus ATP for phosphorylation of PhoP. (B) PhoP∼P specifically activates the P_E2 Eσ^E promoter of the phoPR operon. The 100-nucleotide marker is in the lane marked M. Lanes 1 to 8 contain reconstituted Eσ^E plus the phoPR template. Lanes 2 to 4 contain increasing amounts (0.5, 1, and 5 pmol) of PhoP. Lanes 5 to 8 contain increasing amounts (0.25, 0.5, 1, and 5 pmol) of PhoP∼P. Lanes 9 to 13 contain reconstituted Eσ^E plus the spoIIID template. Lanes 10 and 11 contain 1 and 5 pmol of PhoP, respectively. Lanes 12 and 13 contain 1 and 5 pmol of PhoP∼P, respectively.

The P_E2 promoter is a σ^E promoter that is enhanced by PhoP∼P (Fig. 7B). The reaction using the same promoter template and same the core enzyme as above for reconstituted σ^E RNAP holoenzyme resulted in transcription from the P_E2 promoter (lane 1) that was little affected by increasing PhoP concentrations between 0.5 and 5 pmol (lanes 2 to 4). Reactions containing PhoP∼P (lanes 5 to 8) yielded increasing P_E2 transcripts with increasing PhoP concentrations between 0.25 and 5 pmol, indicating that the Eσ^E P_E2 promoter is PhoP∼P activated. A control experiment (Fig. 7B, lanes 9 through 12) that was carried out with a well-characterized Eσ^E promoter (41, 42), spoIIID, indicated that the spoIIID transcripts were not affected by PhoP or PhoP∼P. These data suggest that PhoP∼P activation of Eσ^E promoters is specific to the phoPR P_E2 promoter.

Reconstituted Eσ^B in reactions with the phoPR template (Fig. 8A) yielded transcripts from P_B1 and P_BX1 (lane 1). Neither promoter appeared to require PhoP (lanes 2 to 5) or PhoP∼P for transcription (lanes 6 to 9), as neither promoter showed a dose-dependent transcription increase, and any variation in transcription appeared to be within experimental error.

FIG. 8.

Core plus SigB is sufficient to transcribe from P_B1 or P_(x1); RNAP from a sigB mutant strain cannot transcribe from P_B1 or P_(x1). (A) Eσ^B transcription from the P_B1 or P_(x1) promoter does not require PhoP or PhoP∼P. Lane M contains the 100-nucleotide marker. Reaction mixtures in lanes 1 to 9 contained the phoPR promoter template and reconstituted Eσ^B. Lanes 2 to 5 and 6 to 9 contained increasing amounts (0.1, 0.5, 1, and 5 pmol) of PhoP or PhoP∼P, respectively. (B) Neither T₀ RNAP nor T₄ RNAP from a sigB mutant strain can transcribe from P_B1. Lanes 1 to 7 contain the phoPR template. The reaction mixture in lane 1 contained T₀ sigE RNAP plus PhoP∼P and identified the migration positions of P_(BX1), P_B1, P_A3, and P_A4. Lane 2 contains core RNAP plus σ^B as in lane 1 of panel A. Reaction mixtures in lanes 3 and 4 contained T₀ RNAP, and those in lanes 5 to 7 contained T₄ RNAP from a sigB mutant strain. Reaction mixtures in lanes 3 and 5 contained PhoP, and lanes 4, 6, and 7 contained PhoP∼P.

RNAP holoenzyme isolated from a sigB mutant strain cannot catalyze P_B1 transcription (Fig. 8). To further test P_B1 and P_BX1 promoter dependence on SigB, we isolated RNAP from a sigB mutant strain (MH6200) at stages T₀ and T₄ for in vitro transcription studies (Fig. 8B). Lane 1 shows transcription products from the phoPR promoter template using RNAP from a sigE mutant as position markers for P_BX1, P_B1, P_A3, and P_A4. Lane 2 shows P_B1 and P_BX1 transcripts from the same template using reconstituted Eσ^B. Neither P_B1 nor P_BX1 was transcribed when using the SigB-deficient T₀ RNAP with PhoP or PhoP∼P (lanes 3 and 4, respectively), in marked contrast to that observed in the control reactions using T₀ RNAP from a sigE mutant strain (lane 1) or reconstituted Eσ^B (lane 2). The SigB-deficient T₀ RNAP yielded an increased level of P_A4 transcript with PhoP∼P (lane 4) compared to that with unphosphorylated PhoP (lane 3). No P_B1 transcript was detected using T₄ RNAP holoenzyme from a sigB mutant (lanes 5 to 7), while P_A4 and P_E2 transcripts were enhanced with PhoP∼P, consistent with previous experiments. Interestingly, a transcript was detected at the P_BX1 position using T₄ RNAP holoenzyme from a sigB mutant.

Together, these studies suggest that the four promoters identified by primer extension using in vivo total RNA from JH642 (Fig. 4) included two σ^A promoters (P_A4 and P_A3), one σ^E promoter (P_E2), and one σ^B promoter (P_B1). The reconstituted RNAP studies suggest that PhoP∼P enhances transcription from Eσ^A promoters, P_A4 and P_A3, and from the Eσ^E promoter, P_E2, but not from the Eσ^B promoter, P_B1.

The P_E2 promoter was not transcribed in a sigE strain; P₅ was identified in RNA from a sigB mutant strain.

To further analyze phoPR promoter expression under phosphate starvation in a sigE (EU8701) or sigB (PB344) mutant strain, RNA was isolated at various times during promoter induction. Primer extension analysis indicated that P_E2 was not expressed in the sigE mutant strain (Fig. 9), consistent with the fact that in vitro transcription studies using RNAP holoenzymes from a sigE mutant strain failed to transcribe P_E2 and that in experiments using reconstituted Eσ^E the only transcript from the phoPR template was P_E2.

FIG. 9.

RNA from a sigE mutant strain contains no P_E2 transcript, and RNA from a sigB mutant strain shows temporal regulation of phoPR promoter transcripts and identifies a 5′ mRNA terminus upstream of P_A4. Primer extension and generation of sequencing ladders were the same as described in the legend for Fig. 4. WT lanes used RNA for primer extension studies that was isolated from the parental strain JH642 at T₀, T₂, or T₄. σ^E lanes used RNA for primer extension studies that was isolated from the sigE mutant strain EU8701 at T₀, T₂, or T₄. σ^B lanes used RNA for primer extension studies that was isolated from sigB mutant strain PB344 at T₀, T₁, T₂, T₃, and T₄.

The role of SigB in phoPR transcription was more complex. Primer extension located an additional 5′ end of a message in RNA from the sigB mutant that was located upstream of P_A4 and was the most abundant transcript at T₀ (identified as P₅ in Fig. 9). By T₁ the relative abundances of P₅, P_A4, and P_A3 were nearly equal, as relative concentrations of P₅ decreased compared to products of the two SigA promoters (P_A3 and P_A4). By T₂ the P_E2 Eσ^E transcript was most abundant and continued to be through T₄. The form of RNAP required for P₅ transcription is not known. Although a sequence similar to a SigH consensus was seen upstream of P₅ (Fig. 2), the P₅ primer extension product was observed by using RNA from a sigH sigB double mutant, suggesting that it is not transcribed by Eσ^H (data not shown). Further complicating the SigB analysis, a 5′ end of a message was detected by primer extension analysis at approximately the same position as the SigB-dependent transcript, P_B1, that was identified in vitro using reconstituted Eσ^B and that failed to be transcribed in vitro using RNAP holoenzymes that were isolated from a sigB mutant strain. It is not clear if this accurately represents the 5′ end of P_B1 transcription initiation or if it results from message processing from one of the upstream phoPR promoters, or if it is the product of premature termination of the reverse transcriptase reaction.

DISCUSSION

Analysis of autoregulation of the phoPR operon identified two new roles for PhoP in promoter activation.

Because PhoP∼P was required for full induction of the phoPR operon during P_i limitation (Fig. 3A) (15, 16), it was important to determine if the regulation were direct and, if so, which promoter(s) was involved. Analysis of data presented here suggests that the mechanism of PhoP autoregulation differs from that required for activation of other Pho regulon promoters in two important ways.

Previous data that established a direct role for PhoP∼P at a particular promoter also showed that Eσ^A holoenzyme was required for transcription from that promoter (34). Here we show that PhoP∼P can also function with Eσ^E holoenzyme to enhance transcription at the P_E2 promoter of phoPR. Two additional B. subtilis response regulators, ResD and Spo0A, are known to function with multiple RNAP holoenzymes. ResD, a paralogue of PhoP, activates two ctaA promoters; one is a Eσ^A promoter and the second promoter requires a developmental sigma factor (30) that we have recently shown to be σ^E (S. Paul and F. M. Hulett, unpublished data). Spo0A∼P activates the spoIIA promoter, whose transcription depends on Eσ^H, and also activates the sigE or spoIIE promoters, whose transcription depends on Eσ^A (19, 43, 44).

Secondly, PhoP was essential for any detectable promoter function in vivo (on-off switch), and PhoP∼P was required for any transcription regulation in vitro at previously studied PhoP-activated promoters (34, 35). In contrast, the role of PhoP∼P in autoregulation is to enhance the otherwise-lower transcription from three phoPR promoters, P_E2, P_A3, and P_A4. Two of these promoters, P_A4 and P_E2, have well-conserved sequences at both the −10 and −35 sequences for SigA and SigE, respectively, which may explain the PhoP-independent transcription. In vivo, SigE-dependent stationary-phase induction of phoPR in SSG was independent of PhoP (Fig. 3), supporting the in vitro transcription data, which showed that Eσ^E was sufficient for P_E2 transcription and that the increase in P_E2 expression by PhoP was phosphorylation dependent. The in vivo data are consistent with the absence of a phosphate deficiency signal for Pho regulation in this high-phosphate medium, SSG, and with the identification of phoPR among genes controlled by Eσ^E in a recent genome-wide study (8).

At least part of the temporal expression pattern for each promoter was explained by the identification of the RNAP holoenzyme required by that promoter using previous knowledge concerning when these RNAP holoenzymes function and how null mutations in one sigma factor affect the RNA holoenzyme pool (17; for review see reference 20). Prolonged σ^A promoter (P_A3 and P_A4) transcription levels in the sigE mutant strain (Fig. 5) are consistent with the observation that if σ^E is not made, σ^A remains associated with the core, whereas in the WT strain when σ^E is activated in the mother cell most of the σ^A is no longer associated with the core RNAP. Similarly, stationary-phase transcription from P_E2 (Fig. 5) is consistent with the timing of σ^E activation in the mother cell during development (11, 18, 26, 31).

The PhoP binding pattern for autoregulation shows similarities and differences when compared to the binding pattern at other PhoP-regulated promoters.

PhoP binding to the phoPR promoter shared certain characteristics observed in PhoP binding patterns at other activated Pho regulon promoters, such as (i) binding unphosphorylated or phosphorylated PhoP to certain promoter regions with extension of DNA protection adjacent to these regions by phosphorylated PhoP (7, 25), (ii) having tandemly repeated consensus sequences for PhoP dimer binding in sequences protected by both PhoP and PhoP∼P (7) or (iii) possessing PhoP binding sites within the coding sequence of the promoter-proximal gene that affect promoter function (25). As with the phoA or pstS promoters, the PhoP binding site within the PhoP coding region was very important for phoPR induction during P_i limitation (Fig. 3A, LPDM), but not for postexponential induction during development (Fig. 3B, SSG) under phosphate-replete conditions.

The PhoP binding pattern upstream of PhoP-stimulated promoters (P_E2 or P_A4) is different than that observed for other Pho regulon-activated promoters (tuaA, phoA, phoB, pstS, or phoD), where PhoP or PhoP∼P protected a core binding region from approximately −20 to −60 that contained two dimer binding consensus repeats on the coding strand (6, 24). PhoP or PhoP∼P protected the P_A4 Eσ^A promoter upstream of −35 in a region that contained a single PhoP dimer consensus repeat on the noncoding strand.

The P_E2 promoter, which has a higher enhanced transcription in vitro with PhoP∼P compared to P_A4, differs from PhoP regulon-activated promoters not only in PhoP binding pattern but also in the holoenzyme required for transcription, Eσ^E. As with P_A4, the P_E2 PhoP binding consensus repeats are on the noncoding strand, but the PhoP-protected region extends from −1 to −35 upstream of the P_E2 transcription start site. Transcription of this promoter during development (T₃ in SSG) was the same in a phoP mutant strain as in the parent strain, indicating that the level of transcription was not dependent on PhoP (Fig. 3) under these phosphate-replete conditions.

The P_A3 promoter is protected by PhoP and PhoP∼P from −23 to +10, with PhoP consensus binding sites on the noncoding strand opposite the +1 site for transcription and the −10 promoter sequence. The P_A3 promoter has a very poor σ^A −35 consensus and appears to be a relatively stronger promoter in vivo than in vitro, suggesting that an additional unknown protein may function in vivo that is absent from our in vitro experiments. This could be a transcription activator or a DNA binding protein that changes the DNA conformation to enhance P_A3 transcription. It occurred to us that ResD might be that activator, but in vitro transcription with ResD or ResD∼P did not increase the P_A3 transcript (Paul and Hulett, unpublished).

Thus, none of the three phoPR promoters that are activated by PhoP∼P have the usual core binding region for PhoP between −20 and −60 relative to their transcription start site. These differences in PhoP binding pattern during autoregulation suggest that the mechanism for PhoP activation of these promoters may be different from that for other Pho regulon promoters and may involve differences in the PhoP-RNAP interaction.

Regulatory coordination between phosphate deficiency response global regulators, PhoP-PhoR and SigB.

Results reported here provide insight into the interdependent regulation between these two global regulators, but more investigation is required to fully characterize the promoters involved. SigB is activated via the energy stress pathway during phosphate-limited growth; thus, both the PhoPR operon and SigB contribute to the B. subtilis phosphate deficiency response. It is likely that the stress from P_i limitation is increased in the sigB mutant strain due to the absence of SigB-regulated genes. Our data suggest that this additional stress is responsible for induction of P₅. The dramatic appearance of an upstream 5′ mRNA end (referred to as P₅) in RNA isolated from a sigB mutant strain during phosphate starvation may account for the increased transcription of phoPR observed in a sigB mutant during P_i limitation (33). Assuming that this 5′ end identifies a fifth phoPR operon promoter, the sigma factor for the putative P₅ promoter is in question. If P₅ expression required only a sigma factor that is present in a sigB mutant strain during phosphate deficiency stress, then one might expect in vitro transcription from P₅ using RNAP holoenzyme isolated from a sigB mutant strain. That P₅ was not expressed in vitro using sigB RNAP, with or without PhoP or PhoP∼P, may suggest that P₅ requires an activator protein that is not PhoP.

In vitro data for transcription with Eσ^B RNAP holoenzyme or RNAP holoenzyme from a sigB mutant strain indicate that P_B1 is a sigB promoter. That mRNA 5′ends were mapped to the P_B1 position in RNA from a sigB mutant strain places the in vitro data in question and requires further experimentation for clarification.

A recent report concerning phoPR transcription (32) contains elements that both agree and differ with the work presented here. The two σ^A promoters Pragai et al. identified correspond to P_A3 and P_A4. Why only two promoters were observed is not clear. Strain differences cannot be the reason, as we have observed all four promoters, including P_B1 and P_E2, in primer extension studies using RNA from B. subtilis 168 (data not shown) in addition to JH642. Differences observed in PhoP footprints and PhoP DNA binding affinity to the phoPR promoter in this and the previous study (32) have logical explanations. The phoPR promoter fragment used in the previous study (32) does not include either the 3′ or 5′ PhoP/PhoP∼P binding sites shown in Fig. 1 and 2. The very high concentrations of PhoP/PhoP∼P required for phoPR promoter protection and differences in the PhoP protection pattern are consistent with the absence of the 3′ and 5′ PhoP binding sites, which were found here and in earlier studies (6, 25, 34) to be important for in vivo promoter activity, PhoP binding affinity, and cooperative binding between PhoP dimers at other Pho regulon promoters.

In conclusion, the data presented in this study reveal a complex phoPR promoter, the complexity of which likely evolved as a consequence of the limited phosphate availability in the soil. The multifaceted transcriptional control suggests the importance of this two-component signaling system to cellular physiology under a wide range of conditions that include phosphate starvation during growth (P_A4 and P_A3) and development (P_E2) as part of development under phosphate-replete conditions (P_E2) and as part of the energy stress response (P_B1 and P₅). The data presented here provide a basic understanding of phoPR transcriptional control onto which additional levels of regulation are likely layered. As such, it should prove an invaluable basis for exploring the proposed roles of ResD (40), AbrB (40), CcpA (3, 12), and SigB (12, 33) in Pho regulation, should they act directly at the transcriptional level of phoPR or affect the Pho regulon signal that in turn affects the transcriptional level of phoPR via autoregulation.