The pst Operon of Bacillus subtilis Is Specifically Induced by Alkali Stress

Abstract

To cope with a sudden increase in the external pH value to 8.9, Bacillus subtilis cells induce about 80 genes which can be divided into two classes. Most of these genes are members of the σ^W regulon, while some are under the control of so-far-unknown transcriptional regulators. The genes of the pst operon belong to the second class. Here, we attempted to answer the questions of why and how the genes of this operon are induced. Using transcriptional fusions to two of the five genes of this operon, we confirmed their induction after alkali stress. Furthermore, a Northern blot experiment revealed that the complete operon was alkali inducible, that the transcriptional start site used was identical to that used after phosphate starvation, and that induction was prevented in a phoR background. Most interestingly, increasing the phosphate concentration within the medium prevented alkali induction of the pst operon, and phoA, another member of the PhoRP regulon, did not respond to alkali stress. In the end, we showed that alkali treatment completely prevented phosphate uptake. These results are discussed to explain alkali induction of the pst operon.

Bacterial cells are equipped with a variety of genetic mechanisms allowing them to cope with stressful situations such as a sudden change in temperature, pH values, and oxidative stress (see the book edited by Storz and Hengge-Aronis [12]). These mechanisms include a sensor, a transcriptional regulator (an alternative sigma factor, a transcriptional activator, or a repressor), and target genes whose transcription becomes transiently induced upon exposure to the stress regimen. While the heat shock response has been extensively studied in Bacillus subtilis (11), analysis of the alkali shock response is still in its infancy. Using the DNA macroarray technique, we showed that more than 80 genes were induced at least threefold after the external pH was increased to 8.9 by the addition of NaOH (15). These alkali-inducible genes could be classified into two groups. While most of them are dependent on the alternative sigma factor σ^W, the rest are under the control of so-far-unknown regulator(s). Among the genes induced by alkali treatment are those of the pst operon.

The pst operon is a member of the PhoP-PhoR two-component signal transduction system, which controls the phosphate deficiency response in B. subtilis (for a recent review, see reference 2). When cells are starved for phosphate, several genes are either activated or repressed by the phosphorylated response regulator, among them those of the pst operon and the phoA gene (4, 8). While the pst operon codes for a high-affinity phosphate transport system (8), the phoA gene encodes an alkaline phosphatase (3).

The objective of the present study was to find out why the pst operon is alkali inducible and to identify the regulator involved. To this end, we showed that alkali treatment of cells led to the induction of the pst operon, but not phoA, and that this induction can be prevented by increasing the phosphate concentration within the medium. Increased expression of the pst operon is caused by impaired uptake of phosphate, which thereby mimics phosphate starvation for the cells.

Transcriptional analysis of the pst operon.

DNA macroarray experiments have shown that the genes of the pst operon are alkali inducible (15). This operon consists of the five genes pstS (formerly yqgG or yzmB), pstC (yqgH or yzmC), pstA (yqgI or yzmD), pstBA (yqgJ or yzmE or pstB), and pstBB (yqgK or yzmF). To confirm these results by an independent experimental approach, we made use of transcriptional fusions between pstS (strain AA01) and pstBA (strain AA02). Both strains were grown in Luria-Bertani (LB) medium at 37°C until they reached the mid-exponential growth phase. Then, the cultures were divided into two halves and while one half was further grown untreated, the other half was alkali stressed by the addition of NaOH to raise the external pH to 8.9. Aliquots were taken before and at different times after alkali challenge, and the β-galactosidase activities were determined. In the absence of alkali treatment, a very low level of activity was measured throughout the growth (Fig. 1A). In contrast, the reporter enzyme activity started to increase at 15 min after alkali challenge and reached about 50 U after 120 min (Fig. 1A). A comparable pattern was found when the pstBA::lacZ operon fusion was analyzed, though only about 30 U was measured after 60 min (Fig. 1B).

FIG. 1.

Influence of the external pH value on expression of genes of the Pho regulon. Cells were grown in LB medium at pH 7.4 to mid-exponential phase and divided into two halves, of which one was further incubated at the neutral pH and the second was alkali induced by the addition of 1 N NaOH to raise the pH value to 8.9. Aliquots from both cultures were withdrawn just before treatment and at the time points indicated for β-galactosidase measurements as described previously (5). The results for noninduced (open bars) and alkali-induced (solid bars) cultures of strains AA01 pstS::lacZ (A), AA02 pstBA::lacZ (B), and AA03 phoA::lacZ (C) are shown. All three strains are derivatives of the wild-type strain 1012 (9), into which the pstS::lacZ and the phoA::lacZ operon fusions integrated at the amyE locus (6) had been introduced by transformation. The pstBA::lacZ operon fusion was constructed by first ligating an internal part of pstBA into pMUTIN4 and then using this region of homology to insert the complete recombinant plasmid into pstBA as described previously (13).

These data suggest that all genes of the pst operon are alkali inducible. To prove this assumption, a Northern blot experiment was carried out. Total RNA was prepared before and at different times after alkali challenge, subjected to Northern blotting, and probed with pstBA antisense RNA. While no signal was obtained in the absence of alkali shock, a signal started to appear at 5 min after induction and continued to increase for at least 60 min (Fig. 2). The size of this signal was determined to be about 4.4 kb, which is in good agreement with the length of the pst operon (4.4 kb). We conclude from this result that the complete pst operon is indeed induced by an alkali shock.

FIG. 2.

Northern blot analysis of the pst operon. Total RNA was prepared from strain 1012 before and 15 min after alkali induction as described previously (1). After RNA electrophoresis in a 1.2% agarose gel and blotting, the nylon membrane was hydridized to digoxigenin-labeled pstBA antisense RNA.

Next, we asked whether the transcription start point used after alkali shock is identical to the one determined after phosphate starvation (8) or whether a different one is used. To answer this question, a primer extension experiment was carried out. Total RNA prepared before and 15 min after alkali challenge was reverse transcribed using the primer ON1, and the same oligonucleotide was also used to prime the DNA sequencing reactions. While no signal was obtained in the absence of alkali shock, a strong one became visible after alkali challenge (Fig. 3). This signal corresponds to a C residue in the codogenic strand and is identical to the transcription start site determined after phosphate starvation (8). Therefore, the same transcriptional start site is used both after alkali stress and phosphate starvation, suggesting that the alkali shock mimics phosphate starvation.

FIG. 3.

Primer extension analysis. Mapping of the 5′ end of the pst operon mRNA was carried out with the ³³P-labeled oligonucleotide ON1 (5′-ACTTTCTCCTGCATTTCC-3′) and 10 μg of RNA prepared before (−) and 15 min after (+) alkali induction as outlined previously (14). DNA sequencing reactions utilizing the same primer and a PCR fragment were performed in parallel with chromosomal DNA as template, and the reaction products were separated on the same gel (lanes A, C, G, and T). The mapped 5′ end of the pst mRNA is marked by an asterisk.

Alkali induction of the pst operon depends on the PhoP-PhoR two-component system.

Expression of the pst operon is under positive control of the Pho two-component signal transduction system (8). Therefore, our finding prompted us to investigate whether alkali induction of this operon is also dependent on the Pho system. A phoR::tet allele was crossed into the pstBA::lacZ background, and the resulting strain AA04 was analyzed for β-galactosidase activity in the absence and presence of alkali. It turned out that in the absence of the sensor kinase, the enzymatic activity did not increase (Fig. 4). We conclude that the two-component Pho system is necessary for the induction of the pst operon under conditions of alkali induction.

FIG. 4.

The pstBA gene is not induced in a phoR background. Strain AA04 (phoR::Tc and pstBA::lacZ) was grow in LB medium in the absence of (open bars) and after (solid bars) alkali challenge. Aliquots were withdrawn at the time points indicated and analyzed for β-galactosidase activities. Strain AA04 was obtained by transfer of the phoR::Tc allele (6) into AA02 by transformation.

Alkali induction is specific for the pst operon.

To find out whether this behavior is specific for the pst operon or whether it affects all members of the Pho regulon, we analyzed a phoA::lacZ fusion (strain AA03). This fusion did not respond to an alkali shock (Fig. 1C). To summarize, the pst operon was induced by an alkali challenge, while expression of the phoA gene was not influenced at all by the external pH. We conclude that this induction behavior is indeed specific for the pst operon and does not affect any other members of the Pho regulon. This conclusion is also supported by our DNA macroarray experiments, which failed to detect induction of other members of this regulon (data not shown).

Addition of phosphate prevents alkali induction of the pst operon.

Proceeding on the basis of our determination that an alkali shock mimics phosphate starvation to the B. subtilis cells, we explored whether increasing the phosphate concentration within the LB medium could prevent alkali induction of the pst operon. The B. subtilis strains AA01 and AA02 carrying the pstS::lacZ and pstBA::lacZ fusions were grown in LB medium with phosphate added to 1 mM, and the β-galactosidase activity was measured in the absence of and after alkali induction. It turned out that after the external pH was increased to 8.9, the higher phosphate concentration within the medium indeed completely prevented induction of the two operon fusions (Fig. 5). This result suggests to us that alkali shock as carried out in our experiments leads to induction of the pst operon which can be overcome by the addition of phosphate.

FIG. 5.

A high phosphate concentration prevents alkali induction of the pst operon. Strains AA01 (A) and AA02 (B) were grown in LB medium either without (open bars) or with (solid bars) the addition of 1 mM phosphate to the mid-exponential phase and then alkali stressed. The phosphate concentration in the fresh LB medium was determined to be 0.24 mM without and 0.54 after the addition of 1 mM phosphate, while the pH value was adjusted to 8.9 as described before. β-Galactosidase activities were measured in aliquots from both cultures taken at the time points indicated.

An alkali shock reduces uptake of phosphate.

What could be the explanation for the finding that an alkaline pH within the medium signals phosphate starvation conditions to the cells? One possibility is that the alkaline pH directly influences the activity of a low-affinity phosphate uptake system. To investigate whether such a low-affinity phosphate uptake system is indeed affected by the external pH, phosphate uptake was measured in the absence and presence of alkali challenge. The results presented in Fig. 6A demonstrate a rapid uptake of ³²P at neutral pH, whereas the uptake was completed abolished at an external pH of 8.9. These data are in agreement with our assumption that the external alkaline pH can affect a low-affinity phosphate uptake system, which can be overcome by increasing the phosphate concentration within the medium. Because of its similarity to genes encoding known low-phosphate transport systems, the pit gene (ykaB) has been suggested to code for such a system in B. subtilis (http://genolist.pasteur.fr/SubtiList). Next, we measured phosphate uptake in strain AA13 (pit::pMUTIN4), which was obtained by transfer of the insertion mutation from strain BSF1802 (10) into strain 1012 by transformation. It turned out that this mutant strain was still able to transport ³²P at neutral pH, albeit at a reduced rate compared to that of the wild-type strain (Fig. 6B), and this uptake was almost completely prevented at pH 8.9. We conclude that the cells of a pit knockout are still able to transport phosphate at neutral pH by at least one unknown system which is also affected by the external alkaline pH.

FIG. 6.

An alkaline pH in the growth medium prevents uptake of added radioactive phosphate. Uptake of radioactive phosphate (³²P) was measured essentially as described previously (6). Strains 1012 and AA13 (pit::pMUTIN4) were grown in low-phosphate medium to the stationary phase. Then, cells were harvested by centrifugation, washed twice in low-phosphate medium without phosphate, resuspended in the original volume of 20 ml, divided into two halves, and further shaken to an optical density at 578 nm of 0.7 to induce maximal phosphate uptake. Next, 995 μl of cells was mixed with 5 μl of ³²P (2 × 10⁴ MBq) in a final concentration of 10 μM P_i. Probes were taken at 15-s intervals, and the amount of radioactive phosphate taken up by the cells was measured. Phosphate uptake was measured at pH 7.4 (◊) and 8.9 (⧫). All experiments were carried out at least three times and yielded comparable results.

Conclusions.

We have presented data to explain why the pst operon is alkali inducible. Our results suggest that the effect of alkali is indirect and operates by impairing the low-affinity phosphate uptake system. We propose that an external pH of 8.9 affects the Pit protein (and/or some other transporter) directly, most probably by interfering with the three-dimensional structure and thereby lowering the affinity to phosphate. Reduced uptake of phosphate in turn signals phosphate starvation to the cells, which leads to the induction of the pst operon. Why is it only this operon and not the other members of the Pho regulon that is induced under these conditions? We propose that differential induction of the members of the Pho regulon is dependent on the actual phosphate concentration within the cell. While a certain concentration leads to the induction of the pst operon, it does not result in the induction of the other members. Furthermore, the pst operon codes for a high-affinity phosphate uptake system. When this system provides the cells with a sufficient amount of phosphate, the other genes are not induced, and this situation was present in our experiments. And last but not least, it has been shown that the pstS operon promoter has a much higher affinity for PhoP∼P than the phoA promoter (7), which further corroborates our explanation of why only the pst operon and not the other members of the Pho regulon is induced under these conditions.