ABSTRACT
The presence of excess glucose in growth media prevents normal sporulation of Bacillus subtilis. The crsA47mutation, located in the gene for the vegetative phase sigma factor (ςA) results in a glucose-resistant sporulation phenotype. As part of a study of the mechanisms whereby the mutation in ςA overcomes glucose repression of sporulation, we examined the expression of genes involved in sporulation initiation in the crsA47 background. The crsA47 mutation had a significant impact on a variety of genes. Changes to stage II gene expression could be linked to alterations in the expression of thesinI and sinR genes. In addition, there was a dramatic increase in the expression of genes dependent on the minor sigma factor ςH. This latter change was paralleled by the pattern of spo0H gene transcription in cells with thecrsA47 mutation. In vitro analysis of RNA polymerase containing ςA47 indicated that it did not have unusually high affinity for the spo0H gene promoter. The in vivo pattern of spo0H expression is not predicted by the known regulatory constraints on spo0H and suggests novel regulation mechanisms that are revealed in the crsA47background.
The activation of genes that are induced late in the growth cycle of Bacillus subtilisinvolves a network of interacting regulatory pathways. These pathways control the cellular response to conditions that include nutritional stress and high cell density (reviewed in references 13, 19, 24,36, and 47). Under the appropriate environmental stimuli,B. subtilis will differentiate to form dormant endospores. The key factor in initiating sporulation is the accumulation and phosphorylation of the response regulator and transcription factor Spo0A (reviewed in references 13, 19, and 24). Phosphorylation of Spo0A takes place through a multicomponent pathway (the phosphorelay) (3) that appears to integrate multiple signals that act positively or negatively to regulate sporulation (24). Once a sufficient level of phosphorylated Spo0A (Spo0A∼P) is reached, a complex series of feedback loops will drive differentiation forward (19, 24, 36).
One critical component of sporulation is the minor sigma factor, ςH, encoded by a gene originally found as a stage zero sporulation mutant, spo0H (12, 20, 46). ςH is required for the transcription of a variety of sporulation genes, including spo0A, spo0F, andkinA (20, 41). The activity of ςH, which increases as cells enter stationary phase, is under complex, still-undefined regulatory controls (19, 45,46). Transcription of the spo0H gene is repressed by the transition state regulator AbrB (11, 48, 50), and so the induction of spo0H seen at the transition between log growth and sporulation is influenced by Spo0A∼P repression of theabrB gene (4, 11). Transcriptional and translational regulation of spo0H by the presence of nutrients has also been reported (8, 14).
A variety of experiments have shown that the increase in ςH activity as measured by transcription of ςH-dependent genes does not match the accumulation of the ςH protein, implying the existence of posttranslational controls (8, 16, 17, 21, 23, 45, 50). Furthermore, evidence has been presented that external pH (8) and the activity of the tricarboxylic acid cycle (26) affect ςH activity and that ςH protein levels are affected by a Lon-type protease during the onset of sporulation and during stress responses (30). Recently, ClpX has been implicated in ςH-dependent transcription activity. In vivo and in vitro experiments have suggested that ClpX may interact directly with RNA polymerase containing ςH and stimulate transcription (31, 37). The level of ςH-dependent transcription increases for 1 to 2 h after sporulation initiation and then declines (4, 11,46). The decline of activity has been linked to another Clp protein, ClpP, and the decline is associated with the loss of ςH protein (38).
Excess glucose in the growth medium represses sporulation, and some of the effects on sporulation have been linked to specific repression of ςH activity (1, 7, 8, 16, 17, 50, 52). ThecrsA47 mutation renders sporulation resistant to repression by glucose (49). The molecular basis for this phenotype is unknown, but the crsA47 mutation is located within the gene for the ςA subunit of RNA polymerase (sigA orrpoD) (27).
In this report, we describe experiments that suggest that thecrsA47 mutation in ςA affects the expression of ςH-dependent genes through its effect on the expression of spo0H. The effects include overexpression of the spo0H gene during the onset of sporulation and unusual extended expression past the time of normal shutoff during differentiation. As the pattern cannot be explained solely by increased promoter affinity, these effects suggest novel regulation of thespo0H gene.
MATERIALS AND METHODS
Bacterial strains and plasmids.The strains used in this study are shown in Table 1. The promoter-lacZ fusions used were created using a B. subtilis amyE integrative vector, with the exception of thekinA-lacZ fusions that were inserted in the kinAgene. Promoter-lacZ fusions provided in the JH642 background were transferred into GBS10 (containing the crsA47 mutation in the rpoD gene but otherwise isogenic to JH642) by transformation with chromosomal DNA, and selection for both the antibiotic resistance and the amyE mutant phenotype was conferred with the acquisition of the construct. The plasmid pGS0H was created by the ligation of theEcoRI/BamHI-digested Vent polymerase (New England Biolabs, Inc.) PCR product generated from chromosomal DNA and the primer pair 5′-AAGGATCCTGTTTCTGGCGAGTAG-3′ and 5′-ACGAATTCGGCACGGACGTTAGAA-3′ that targets thespo0H gene into the EcoRI/BamHI sites of the B. subtilis integrative vector pDH32 (44). The pDH32-based clone was linearized withPstI prior to transformation into B. subtilis. Chloramphenicol-resistant transformants generated were confirmed to be amylase negative by using 1% starch agar plates prior to β-galactosidase analysis.
B. subtilis strains used in this study
Bacterial transformation and growth conditions. B. subtilis transformations were performed by the method of Hoch (24) with 1 to 2 μg of plasmid DNA or 20 to 100 ng of chromosomal DNA. Transformants were selected on Schaeffer sporulation agar plates supplemented with 5 μg of chloramphenicol or kanamycin/ml. Bacillus cultures used to determine sporulation frequency were grown in Schaeffer spore media (SSM), pH 7.5, supplemented with tryptophan and phenylalanine (each 10 μg/ml), chloramphenicol or kanamycin (5 μg/ml), and, when appropriate, SSM with 1% glucose (SSMG). Cells were grown for 22 to 24 h, serially diluted in SSM, and plated before and after treatment with 0.1 volume of chloroform to obtain a viable cell count and a spore count.
β-Galactosidase assay. B. subtilis strains used for analysis of promoter-lacZ activity were inoculated into SSM, pH 7.5, containing 5 μg of the appropriate antibiotic/ml and supplemented with tryptophan and phenylalanine (each, 10 μg/ml) and, when appropriate, 0.2% glucose. Aliquots (0.5 ml) were removed hourly, the cells were collected by centrifugation, and cell pellets were frozen at −70°C until analyzed. β-Galactosidase assays and designation of time zero (T 0) in sporulation were done as previously described (15). Enzyme specific activity was expressed in Miller units (35). Assays of a minimum of three independent cultures were performed, and one representative pattern for each strain is shown. Each point in the assays is an average of duplicate samples that differed by no more than 5% from the mean.
In vitro transcription assay.The RNA polymerase preparations were isolated as described by Dobinson and Spiegelman (10) from logarithmic-stage cultures of JH642 and GBS10, except that the heparin-Sepharose column set was eliminated. Transcription assays used fractions from the glycerol gradient. Two DNA templates were used. The plasmid pUCA2trpA (6) that contains the bacteriophage φ29A2 promoter was treated withPvuII, and the 600-bp fragment containing the promoter was isolated by electrophoresis through agarose and extracted from the agarose with a GeneClean kit from Qiagen. The spo0H promoter was isolated by amplifying a DNA fragment using 20 pmol (each) of two specific primers described above, 800 ng of chromosomal DNA from JH642, and an amplification protocol as follows: preincubation of the template and primers at 95°C for 5 min, and after the addition of 2.5 U ofTaq polymerase (Promega), 30 cycles of 95°C for 1 min, 61°C for 1 min (with descending annealing temperatures of 0.3°C per cycle), and 72°C for 1 min. The product was purified by electrophoresis through 1% agarose and extracted from the gel as described above for the A2 promoter. The extracted product was digested with HindIII to provide a fixed endpoint for the transcription assays and then precipitated. The concentrations of promoter fragments were determined by measuring the absorbance at 260 nM.
The in vitro transcription assay followed published protocols (6). In brief, reaction tubes containing 16 μl of reaction mixture with template, transcription buffer (6), ATP, and [α-32P]GTP (3 μCi/reaction mixture) were warmed to 37°C. The polymerase (2 μl of an appropriate dilution) was added. Two minutes later, 2 μl of a mixture of heparin (100 μg/ml, final concentration), UTP, and CTP was added to inactivate noninitiated RNA polymerase molecules and allow those that had initiated to elongate either to the terminator (in the case of the φ29A2 promoter) or to the end of the DNA fragment (in the case of thespo0H promoter). After 5 min, a stop buffer containing 7 M urea was added, and the reaction products were separated from free nucleotides on 5% polyacrylamide gels containing 7 M urea and 0.5× Tris-borate-EDTA (TBE). The gels were exposed to a Molecular Dynamics PhosphorImager screen, and the data were collected and analyzed with the ImageQuant 1.0 software on the instrument.
RESULTS
The crsA47 mutation results in overexpression of ςH-dependent spo0 genes in glucose-containing media.To investigate the mechanism by which the crsA47mutation resulted in catabolite-resistant sporulation, we examined the expression of a variety of genes important in the sporulation initiation pathway in the presence and absence of glucose. Among these genes, the ones with ςH-dependent promoters showed very unusual profiles of activity. Figure 1shows the expression of the kinA and spoVGpromoter fusions in wild-type (JH642) and crsA47 (GBS10) backgrounds. The kinA promoter has not been previously shown to be subject to regulation other than that imposed by ςHactivity (1, 41). The spoVG gene is preceded by a ςH promoter that is repressed during logarithmic growth by the transition state regulator AbrB (43, 48).
The effect of the crsA47 mutation on expression of kinA and spoVG. β-Galactosidase activities in strains carrying kinA-lacZ (A) orspo0VG-lacZ (B) fusions were measured as described in Materials and Methods. T 0 represents the onset of sporulation. Strains contained the wild-type ςA gene (squares) or the crsA47 mutation (diamonds) and were grown in either SSM (open symbols) or SSMG (filled symbols). (A) JH12664 (kinA-lacZ) and GBS107 (crsA47 kinA-lacZ); (B) GBS110 (spoVG-lacZ) and GBS109 (crsA47 spoVG-lacZ).
Three observations were common to the activity of the kinAand spoVG promoters. First, in a wild-type ςAbackground promoter activity was depressed by the presence of glucose. Second, in a crsA47 background the expression was elevated by the presence of glucose compared to that seen in the wild type in the absence of glucose, with transcription levels persisting long after the point of maximum activity in the wild type. Third, promoter activity in a crsA47 background in the absence of glucose was only marginally affected compared to the effect in the wild type. The onset of promoter activity was not affected in the same way for the two promoters. kinA promoter activity (Fig. 1A) in GBS107 grown in the absence of glucose began earlier than in JH12664, whereas the timing of spoVG transcription activity (Fig. 1B) appeared similar in both GBS110 and GBS109. The logarithmic-phase repression of spoVG by AbrB did not differ between wild-type and crsA47 mutants. The suggestion that AbrB regulation was not altered in the GBS109 was supported by the expression of anabrB-lacZ fusion which showed similar patterns in GBS10 and JH642 (data not shown).
The expression patterns of spo0F promoter-lacZfusions are shown in Fig. 2. Thespo0F gene is preceded by dual ςAςH promoters, with the ςH-dependent promoter requiring Spo0A∼P as a transcription activator (28,41, 51). In JH12862, grown in SSM, expression from thespo0F promoter-lacZ fusion increased during late-exponential-phase growth, peaked at roughlyT 1, and decreased thereafter. The addition of glucose to the media resulted in a decrease in overall expression. In GBS105, the expression from spo0F began earlier and peaked at higher levels than were seen in JH12862, both with and without added glucose. As was seen with both the kinA and spoVGpromoters (Fig. 1), transcription of spo0F in GBS105 in cells grown with added glucose appeared to be stimulated afterT 0 compared to levels seen in cells grown without glucose.
The effect of the crsA47 mutation on expression of spo0F. β-Galactosidase activities in strains carrying spo0F-lacZ fusions were measured as described in Materials and Methods. T 0 represents the onset of sporulation. Strains contained the wild-type ςA gene (JH12862; squares) or the crsA47 mutation (GBS105; diamonds) and were grown in either SSM (open symbols) or SSMG (filled symbols).
The crsA47 mutation results in the expression ofspoII genes despite the presence of glucose. The late- and post-exponential phase expression of ςH-dependentspo0 genes appeared to be elevated in the cells containing the crsA47 mutation (Fig. 1 and 2), suggesting that ςH was active in these cells despite the presence of glucose. The activation of ςH during the transition state has been shown to be critical for the initiation of sporulation and is required for the maximal expression of genes encoding phosphorelay proteins and accumulation of a sufficient level of Spo0A∼P to activate the transcription of stage II spo genes. As a test for whether or not Spo0A∼P was fully activated, we examined the expression of the stage II genes spoIIA andspoIIG (Fig. 3).
The effect of the crsA47 mutation on expression of spoIIA and spoIIG. β-Galactosidase activities in strains carrying spoIIA-lacZ(A) or spoIIG-lacZ (B) fusions were measured as described in Materials and Methods. T 0 represents the onset of sporulation. Strains contained the wild-type ςA gene (squares) or the crsA47 mutation (diamonds), and were grown in either SSM (open symbols) or SSMG (filled symbols). (A) JH16124 (spoIIA-lacZ) and GBS106 (crsA47 spoIIA-lacZ); (B) JH16304 (spoIIG-lacZ) and GBS101 (crsA47 spoIIG-lacZ).
In strain JH16124 cells grown in SSM (Fig. 3A), the expression from the ςH-dependent spoIIA promoter-lacZfusion began roughly at T 0 and peaked atT 2, dropping thereafter. In the presence of glucose, spoIIA-lacZ expression was depressed in stationary phase. In GBS106 grown in SSM (Fig. 3A), transcription activity from the spoIIA promoter began to increase at the same time as that observed in JH16124 and peaked at levels not substantially different from that in JH16124. In the presence of glucose, the expression of spoIIA-lacZ in GBS106 began immediately after the onset of stationary phase and rose to a level more than twice that seen in JH16124 grown in the absence of glucose.
Figure 3B depicts the expression from the ςA-dependentspoIIG promoter-lacZ fusion in the wild type (JH16304) and cells containing the crsA47 mutation (GBS101). As seen in Fig. 3A for the spoIIA-lacZ fusion, promoter activity from the spoIIG promoter was repressed by glucose in the wild type but not repressed in GBS101. Since effective expression of these two stage II operons directly requires high Spo0A∼P levels, these cells must contain high levels of Spo0A∼P. The spo0A gene is expressed from two promoters, one ςA dependent and one ςH dependent. It has been shown elsewhere that expression for the ςH-dependent promoter is enhanced in crsA47 mutants (9).
The presence of the crsA47 mutation results in an alteration in the pattern of transcription from sinI andsinR promoters in the presence of glucose.SinR inhibits the expression of several spo genes, includingspo0A (34), spoIIG, andspoIIA (5, 32, 33). The sinR gene is constitutively expressed from a ςA-dependent promoter throughout exponential and post-exponential growth of B. subtilis (18). SinR inhibition of transcription is negatively regulated by Spo0A∼P levels, which stimulate increased transcription of the ςH-dependent sinIgene (the gene directly upstream of sinR) (18). SinI sequesters SinR via protein-protein interaction preventing SinR-mediated repression of promoter activity (2, 29). Gaur et al. showed that the presence of excess glucose in the media inhibits the transcription of the sinIgene (18). Presumably, inadequate transcription ofsinI results in a SinI/SinR protein ratio insufficient to fully sequester SinR and relieve repression of sporulation genes (2, 29).
The data in Fig. 3 showed that ςH- and ςA-dependent stage II promoters were deregulated in cells containing the crsA47 mutation grown in glucose. Because SinR is central to regulation of both the spoIIG andspoIIA operons, we examined the regulation of thesinR and sinI promoters. The data shown in Fig.4 indicate that in cells containing wild-type ςA (GBS114) grown in the absence of glucose,sinI transcription levels increased throughout late logarithmic growth to peak at T 0 (Fig. 4A). Transcription of sinR increased throughout late logarithmic growth and into stationary phase (GBS116) (Fig. 4B), presumably in part due to readthrough from the sinI promoter (18) as well as from the sinR promoter. The expression patterns in Fig. 4 are similar to those observed by others (18,33). When glucose was added, sinI transcription remained relatively low during stationary phase in GBS114 (Fig. 4A), whereas sinR transcription in GBS116 (Fig. 4B) remained roughly the same as in the absence of glucose.
The effect of the crsA47 mutation on expression of sinI and sinR. β-Galactosidase activities in strains carrying sinI-lacZ (A) orsinR-lacZ (B) fusions were measured as described in Materials and Methods. T 0 represents the onset of sporulation. Strains contained the wild-type ςA gene (squares) or the crsA47 mutation (diamonds) and were grown in either SSM (open symbols) or SSMG (filled symbols). (A) GBS114 (sinI-lacZ) and GBS113 (crsA47 sinI-lacZ); (B) GBS116 (sinR-lacZ) and GBS115 (crsA47 sinR-lacZ).
If transcription of the sinI and sinR genes reflects protein levels, then cells with wild-type ςAgrown in the absence of glucose would contain a roughly 20-fold excess of SinI over SinR. Since these cells sporulate efficiently, this ratio should indicate the level of SinI needed to complex SinR betweenT 0 and T 1.5. In the presence of glucose, expression of the sinI promoter in cells with wild-type ςA was reduced, with the implication that the ratio of SinI to SinR would not block SinR repression of sporulation.
sinI promoter activity in cells containing thecrsA47 mutation (GBS113) is also shown in Fig. 4A. In these cells, the activity from the sinI promoter in the absence of glucose rose slowly during logarithmic growth to peak atT 0.5 at levels 25 to 30% of that seen in GBS114. The sinI-lacZ activity was also altered in GBS113 grown in the presence of glucose (Fig. 4A), with the observed pattern of transcription similar to that seen in other ςH-dependent promoters examined in the crsA47background. Transcription from the sinI promoter rose fromT −1.5 to peak after T 2at levels five times higher than was seen in cells with wild-type ςA.
Figure 4B shows sinR-lacZ activity in cells containing thecrsA47 mutation (GBS115). Without excess glucose, transcription of the sinR gene was reduced from that seen in cells containing wild-type ςA (GBS116), peaking at roughly T 0 and decreasing thereafter. In the presence of glucose, transcription rose sharply fromT −2 to peak at T 0 at levels similar to those achieved in GBS116 in the presence or absence of glucose.
Continuing the assumption that the activity of the sinI andsinR promoter fusions reflects protein levels, then even though SinR and SinI levels were reduced in cells containing thecrsA47 mutation (in media without excess glucose), the ratio would be similar to that seen in wild-type cells. Thus, SinR activity would be blocked by SinI during early stationary phase. When glucose was added, the large induction of the sinI promoter would further reduce SinR activity. Thus, unlike cells with wild-type ςA where the addition of glucose decreased the ratio of expression of sinI/sinR, in the cells with thecrsA47 mutation the ratio of sinI/sinR would increase with added glucose. This alteration in sin operon transcription in the crsA47 mutant could contribute to the ability of these cells to express the spoIIA andspoIIG operons in the presence of glucose.
The importance of SinR negative regulation in glucose repression of sporulation was examined by determining the sporulation efficiency of a ΔsinR mutant, as shown in Table2. In the crsA47 mutant (GBS10), the sporulation efficiencies of cells grown in the presence and absence of glucose were comparable, clearly indicating a glucose-resistant sporulation phenotype. In JH642, the addition of excess glucose to the medium resulted in a 104-fold decrease in sporulation efficiency. However, in GBS112 (sigA+ ΔsinR) the sporulation efficiency in medium with excess glucose was only three-fold less than that observed in the absence of glucose. These results, combined with those shown in Fig. 2 through 4, suggest that a decrease in SinR repression whether by sinR gene deletion or by altering the ratios of sinI and sinR transcription increased expression of spoII genes and contributed to the glucose-resistant sporulation phenotype.
Sporulation efficiencies of sinl-lacZ-, sinR-lacZ-, and ΔsinR-containing strains
The crsA47 mutation results in increased transcription from the spo0H gene in late- and post-exponential-phase growth.The results shown in Fig. 3 and 4 demonstrate increased expression of both ςA- and ςH-dependent stage II genes in the crsA47 background when the cells were grown in the presence of excess glucose. The implication from the analysis of the sinI-sinR operon was that the increase in stage II gene expression reflected the increase in ςHactivity, which allowed SinI to inactivate SinR. Thus, overexpression of the spoIIG operon (ςA dependent) has the same fundamental mechanism as overexpression of the spoIIAoperon (ςH dependent); that is, increased activity of ςH. However, the reason for the high levels of stationary-phase transcription from all of the ςH-dependent promoters examined in the crsA47mutant in the presence of glucose was not clear. Previous studies have shown higher-than-normal ςH-dependent promoter activity during the stationary phase either by using an increased copy number ofspo0H or by placing spo0H under the control of the isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible PSPAC promoter (21). Thus, the increased ςH activity could be due to overexpression of thespo0H gene in the crsA47 mutant, and this was examined using a spo0H-lacZ fusion in the presence and absence of added glucose.
Figure 5 shows the expression patterns of the ςA-dependent spo0H-lacZ fusion in strains containing wild-type sigA (GBS151) or the crsAmutation (GBS150). In GBS151 the activity of the spo0Hpromoter began to increase at T −2 to peak at roughly T 1, with the maximum activity not substantially different in the presence or absence of glucose. This pattern of transcription is similar to that seen elsewhere (1,50). In GBS150 the timing of transcription from thespo0H promoter was the same as in GBS151, but peak activity was roughly eight times higher in the absence of glucose and 19 times higher in the presence of glucose than was seen in the GBS151.
Effect of the crsA47 mutation on the expression of spo0H. β-Galactosidase activities in a strain carrying spo0H-lacZ fusions were measured as described in Materials and Methods. T 0represents the onset of sporulation. Strains contained the wild-type ςA gene (GBS151; squares) or the crsA47mutation (GBS150; diamonds) and were grown in either SSM (open symbols) or SSMG (filled symbols).
The high level of transcription of the spo0H gene seen in GBS150 grown in the presence of glucose could lead to excess ςH protein and thus explain the unusual patterns of ςH-dependent transcription seen in cells containing thecrsA47 mutation. These data raise the question whether the extended transcription from the ςA-dependent promoter of the spo0H gene reflected a general phenomenon or whether it was specific to the spo0H promoter. The only known regulator of the spo0H promoter is AbrB. We would not expect AbrB levels to play a role in this regulation of spo0H beyondT 1 as the gene is repressed earlier, and this repression was not changed in cells containing the crsA47mutation (9). However, it was possible that thecrsA47 mutation altered the reduction of ςAactivity that is normally seen in sporulation (46).
As part of the characterization of the crsA47 mutation, we examined several other ςA-dependent promoters. TherapA-encoded phosphatase removes phosphate from the phosphorelay by dephosphorylating phosphorylated Spo0F (reviewed in reference 40). Transcription from the rapA gene promoter also showed a crsA47- and excess-glucose-specific increase in transcription after T 0 (Fig.6A). In other work, the same effect for the ςA-dependent spo0APv promoter has been seen (9). In contrast, expression of a second ςA-dependent gene encoding a phosphatase,rapB, did not show increased transcription in strains containing the crsA47 mutation and grown in excess glucose (Fig. 6B). Thus, the effect of the crsA47 mutation was promoter specific as well as being specific to growth conditions.
Effect of the crsA47 mutation on the expression of rapA and rapB. β-Galactosidase activities in strains carrying rapA-lacZ (A) orrapB-lacZ (B) fusions were measured as described in Materials and Methods. T 0 represents the onset of sporulation. Strains contained the wild-type ςA gene (squares) or the crsA47 mutation (diamonds) and were grown in either SSM (open symbols) or SSMG (filled symbols). (A) JH12961 (rapA-lacZ) and GBS104 (crsA47 rapB-lacZ); (B) JH12866 (rapB-lacZ) and GBS103 (crsA47 rapB-lacZ).
In vitro analysis of transcription from the spo0Hpromoter by RNA polymerase containing ςA47.One possible explanation for the increase in ςH activity seen in strains containing the crsA mutation is that the mutation in ςA increases the affinity of the polymerase for thespo0H gene promoter. To directly measure this affinity, we purified RNA polymerase from wild-type and GBS10 strains and tested them in vitro in single-round transcription assays (as described in Materials and Methods) (Fig. 7). As a means of measuring the specific activity of the RNA polymerase preparations, we compared the activity of the two preparations on a promoter from bacteriophage φ29 that has been cloned into the plasmid pUCA2trpA (Fig. 7A). Aliquots of the polymerase preparations were mixed with DNA and initiating nucleotides (ATP plus GTP) and challenged with a mixture of heparin and CTP plus UTP. The heparin inactivates noninitiated RNA polymerase molecules while the CTP plus UTP allow enzymes that have initiated to elongate. Over a range of DNA concentrations of 0.5 to 8 nM, the wild-type RNA polymerase was three- to five-fold more active than the preparation from GBS10. We then repeated this experiment using a DNA fragment containing thespo0H promoter. This DNA fragment was produced by amplification of the promoter region. Transcription from this promoter produced a single transcript (data not shown). The activity of both polymerase preparations was significantly lower on the spo0Hpromoter than on the phage promoter. However, as with the φ29A2 promoter, the polymerase containing ςA47 was three- to five-fold less active on this template than was the wild-type polymerase. By this assay, the change in ςA caused by thecrsA mutation did not increase the affinity of the polymerase for the spo0H promoter.
In vitro transcription activity of RNA polymerase isolated from JH642 and GBS10. In vitro transcription reactions were carried out using a control promoter (pUCA2trpA) (A) or a DNA fragment containing the spo0H gene promoter (B). Reaction mixtures contained a constant amount of RNA polymerase from either JH642 (wild type; squares) or GBS10 (crsA; circles) and increasing amounts of template. Transcription products were separated by electrophoresis through polyacrylamide gels containing 7 M urea. The level of product produced was determined using a Molecular Dynamics PhosphorImager and ImageQuant 1.0 software and is reported in arbitrary units.
DISCUSSION
We began the study of the crsA47 mutation to uncover the mechanism by which it makes sporulation resistant to the presence of excess glucose in the growth medium. During the course of these studies, we observed the consistent overexpression of ςH-dependent genes in strains carrying crsA47that we have described in this paper. The overexpression occurred afterT 0 and only in strains grown in the presence of excess glucose. The overexpression peaked 2 to 3 h later than the normal peak of ςH-dependent expression and then declined. This combination of features suggests a novel feature of ςH transcription regulation that is revealed in thecrsA47 genetic background.
The regulation of ςH activity appears to be complex. There is a low level of transcription of the spo0H gene during vegetative growth, and at least a few ςH-dependent genes are transcribed during this time (19, 42). In wild-type cells, transcription dependent on ςH is induced during the transition stage, and it normally peaks within 1 to 2 h after the onset of sporulation and decays after that time (46). Comparison of the levels of ςH protein with ςH-dependent transcription in vegetative growth and in early sporulation indicates the presence of posttranscriptional regulation of ςH (1, 8, 16, 17, 21, 23, 42,50). Later in sporulation, the decrease in ςHactivity is associated with loss of ςH protein due to the activity of ClpC protease (38). Another Clp family protein, ClpX, influences the induction of activity of ςHduring early sporulation, although not through changes in the level of ςH protein. It has been reported that ClpX stimulates ςH-dependent transcription by direct interaction with the polymerase containing ςH (31, 37). It is known that ςH activity is regulated at transcription, since spo0H gene transcription is repressed by AbrB (11, 43). It has been recently shown that ςH-dependent expression of a limited set of genes is enhanced by amino acid starvation and that this enhancement requires the relA gene product (14). While this finding is yet another illustration of the complexity of controls over ςH activity, it is probably not related to our findings, which appeared to be specific to the presence of glucose.
It seems likely to us that the overexpression of ςH-dependent genes in GBS10 and its derivatives grown in the presence of glucose cannot be explained by loss of repression by AbrB, since AbrB repression was normal in GBS10 (9), and ςH-dependent transcription continues to increase after the time when AbrB is repressed (9). Given the similarity of the expression patterns of the spo0H gene (Fig. 5) and the ςH-dependent genes (Fig. 1 and 3), we suggest that overexpression of spo0H in cells containing thecrsA47 mutation is a sufficient explanation for the unusually high transcription of ςH-dependent genes.
In a direct in vitro test of the activity of RNA polymerase isolated from GBS10 and wild-type cells (Fig. 7), there was no evidence that thecrsA47 mutation increased the activity at thespo0H promoter. Furthermore, several lines of in vivo evidence support the finding that the crsA47 mutation does not simply increase the affinity of the polymerase for thespoH promoter. First, GBS10 does not grow unusually slowly, as might be expected if the crsA47 mutation altered the polymerase specificity. Second, the overexpression of thespo0H promoter in GBS10 happened only in stationary phase and only in the presence of excess glucose in the growth medium, suggesting a specific regulatory mechanism. Third, two other ςA-dependent promoters studied (for the abrBand rapB genes) did not show unusual expression patterns in GBS10 (9). We note in passing that the regulation of AbrB, which is known to control expression of a number of genes (48), must be particularly important to the overall physiology of cells with a crsA47 mutation, because it was found that a crsA47 abrB double mutant grew so poorly even in rich media that regulation in the strain could not be studied (9).
Our results indicate that there must be a regulator of thespo0H promoter (and possibly other similar promoters) whose activity is changed in GBS10 grown in excess glucose. We cannot provide any indication whether the regulator is an activator of thespo0H promoter that is hyperactive or a repressor whose activity is reduced. The latter seems more likely, since the extended high expression can be viewed as a lack of shutoff of spo0Htranscription. The existence of such a regulator implies that other controls of ςH are yet to be discovered.
A consequence of ςH overexpression in the presence of glucose was the change in the SinI-to-SinR ratio, with the predicted result that SinR repression of stage 0 and stage II sporulation genes would be reduced. There is little question that reduction of SinR in the cell would increase the glucose resistance of sporulation (32). This is likely to be a major contributor to the glucose-resistant sporulation in GBS10 and illustrates the redundant pathways that regulate entry into sporulation.
ACKNOWLEDGEMENTS
We thank M. Perego, J. A. Hoch, and I. Smith for their continuous generosity in providing strains and insights during the course of this study.
The research was supported by grants from the Natural Sciences and Engineering Research Council of Canada and the Canadian Institutes of Health Research to G.B.S.
FOOTNOTES
- Received 22 November 2000.
- Accepted 30 May 2001.
- Copyright © 2001 American Society for Microbiology
