![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
Received 28 February 2006/ Accepted 26 April 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In the gram-positive bacterium Bacillus subtilis, several processes are known to be regulated by extracellular peptide signaling, including the initiation of genetic competence (the ability to incorporate exogenous DNA from the environment) (41, 70), sporulation (59, 70), production of degradative enzymes (9, 52) and exopolysaccharides (9, 77), and antibiotic synthesis (9, 41, 70). Three types of secreted peptide signaling molecules have been identified: a modified 5- to 10-amino-acid peptide, ComX, that interacts extracellularly with its receptor (41, 62, 76); lantibiotic peptides, such as subtilin, which interact extracellularly with their receptors (reviewed in reference 73); and unmodified pentapeptides, known as Phr peptides, that are internalized to inhibit the activity of their target proteins, known as Rap proteins (reviewed in references 31 and 56).
B. subtilis encodes a family of 8 Phr peptides (PhrA, PhrC, PhrE, PhrF, PhrG, PhrH, PhrI, and PhrK) and a family of 11 Rap proteins (RapA to RapK) (24, 30). Each Phr peptide is encoded in an operon with a Rap protein (30), and each characterized Phr inhibits the activity of its cotranscribed Rap (6, 27, 52, 59, 70). The PhrC peptide (also known as competence- and sporulation-stimulating factor [70]) also inhibits the activity of an unpaired Rap protein, RapB (55). It is possible that the other unpaired Rap proteins are also inhibited by noncognate Phr peptides.
In addition to expression from the upstream rap promoter (32, 59), most phr genes are also expressed from a promoter upstream of phr that is recognized by RNA polymerase containing the alternative sigma factor, 
H (32, 42, 47). This regulation by H causes the level of each phr gene to increase as cells transition from exponential growth to stationary phase (32, 42).
The primary phr gene products are pre-Phr peptides that are 38 to 57 amino acids in length. Pre-Phr peptides are exported and cleaved to form the mature Phr pentapeptides (reviewed in references 31 and 56). The oligopeptide permease (Opp), an ATP-binding cassette (ABC) transporter that imports small peptides (58, 65), transports the Phr peptides into the cell, where they can inhibit the activities of Rap proteins (34, 55).
Several Rap proteins regulate a variety of processes by antagonizing the activities of response regulator proteins, either by stimulating dephosphorylation of the response regulator (RapA, RapB, and RapE) (27, 57, 78) (Table 1) or by binding to the response regulator and interfering with DNA binding (RapC, RapF, and RapG) (6, 10, 52). One Rap protein, RapI, regulates the mobility of the genetic element ICEBs1, possibly by antagonizing the activity of the element-encoded immunity repressor (2). RapG and RapH were recently shown to inhibit expression of some ComA-activated genes, as well as the expression of DegU-activated genes (24). The remaining Rap proteins (RapD, RapJ, and RapK) have not been characterized.
|
In order to more fully understand the roles that Rap proteins and Phr peptides play in modulating ComA-activated processes, we characterized the roles that several Rap proteins and Phr peptides play in regulating ComA-dependent gene expression. We monitored the effects of overproduction of several uncharacterized Rap proteins on global gene expression and found that overproduction of RapK inhibits the expression of genes activated by ComA. Furthermore, deletion of phrK resulted in increased expression of ComA-dependent genes. This effect on ComA-dependent gene expression was dependent upon rapK, indicating that PhrK stimulates ComA's activity by inhibiting RapK's activity. We also observed that PhrC, PhrF, and PhrK stimulate ComA-dependent gene expression to different magnitudes and that all three peptides are required for full expression of ComA-dependent genes. Based on these results as well as previously published observations, we conclude that signaling through the PhrC, PhrF, and PhrK peptides integrates multiple physiological cues to modulate the levels and timing of processes regulated by ComA.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Strains and alleles.
The strains used in this study are listed in Table 2. All B. subtilis strains were derived from the parental strain JH642 (60). The E. coli strain used for cloning is an MC1061 derivative with F'(lacIq) lacZM15 Tn10 (tet). Standard techniques were used for cloning and strain construction (23, 66). Pspank(hy) (2), 
rapC::cat (70), phrC::erm (70), and amyE::srfA-lacZ
682 (3) were previously described.
|
rapF-phrF was deleted by replacing (from the start codon) nucleotides +542 of rapF to +149 of phrF with a tetracycline resistance gene derived from pDG1513 (19). rapK was deleted by replacing nucleotides +36 to +980 of rapK with the chloramphenicol resistance gene from pGEM-cat (83); cat was replaced with erm by integration of the plasmid pCm::Er (74). phrF was deleted by replacing nucleotides +38 to +103 of phrF with the chloramphenicol resistance gene from pGEM-cat. phrK was deleted by replacing nucleotides +64 to +100 of phrK with spectinomycin resistance derived from pDL55 (4).
oppBCDE::spc (spo0KBCDE::spc) was created by replacing the RsrII/ClaI fragment in plasmid pDR9 (65) (contains oppBCDE) with the BglI/NdeI fragment containing spc from pUS19 (4). Both plasmid and insert DNA was recessed/filled-in by the Klenow fragment of E. coli DNA polymerase I prior to ligation. The recombinant plasmid was integrated into the B. subtilis chromosome by homologous recombination.
sigH::cat::spc (spo0H::cat::spc) was created by integrating plasmid pJL62 (35) by homologous recombination into a strain containing the sigH::cat (26) mutation.
The Ppel-lacZ promoter fusion was generated by cloning the DNA from nucleotides –371 to +39 of pel upstream of the promoterless lacZ in the vector pDG793 (18), followed by integration into the thrC locus by homologous recombination. A similar fusion at amyE was previously described (9).
DNA microarrays.
Pspank(hy)-rapF (JMA26), Pspank(hy)-(rapHphrH) (JMA27), Pspank(hy)-rapJ (JMA29), and Pspank(hy)-rapK (JMA30) cells were grown in defined minimal medium for at least four generations to an optical density at 600 nm (OD600) of 
0.5. IPTG was added to half of the cultures, and samples were collected from induced and uninduced cultures 30 min later. Wild-type (JH642), phrC (RSM121), phrF (JMA163), and phrK (CAL7) cells were grown in defined minimal medium for at least four generations to an OD600 of 1, when samples were collected.
Cells were harvested and total RNA was prepared as described previously (7). RNA from each sample was reverse transcribed and labeled as described previously (2). In the experiments monitoring gene expression in cells overexpressing the indicated rap gene, labeled cDNAs from induced samples (with IPTG) and uninduced samples (without IPTG) were cohybridized to cDNA microarrays as described previously (2). In the experiments monitoring gene expression in wild-type, phrC, phrF, and phrK cells, labeled cDNA from each experimental sample was hybridized with a labeled reference cDNA sample to 65-mer oligonucleotide arrays as described previously (2).
Arrays were scanned and analyzed as described previously (2). Iterative outlier analysis (7, 37) was used as described previously (2) to identify genes whose expression changed significantly with 95% or greater confidence. The mean ratio for a set of three independent experiments is reported. Lists of significant genes were arranged into known or putative operons based on the prediction of coorientation of transcription and the absence of predicted Rho-independent terminators. If a gene or genes that were part of a known or predicted operon changed significantly, the average fold changes in gene expression for the other genes in the operon were also assessed. If the expression of those genes changed similarly but these changes were below the significance threshold of the analysis, the values of these fold changes were included in Fig. 1 and in Table S1 in the supplemental material.
|
Microarray data accession number. Complete microarray results are available in Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo). The accession number for the series record is GSE4670.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Each rap gene was overexpressed from the IPTG-inducible promoter Pspank(hy) during exponential growth in defined minimal medium. RNA transcript levels in induced and uninduced cells were compared 30 min after induction. We analyzed the results of three independent experiments to identify those genes whose expression changed significantly in response to overproduction of each Rap. We compared the results of these experiments to the published genomewide analyses of several response regulator regulons (9, 14, 29, 40, 46, 54) in order to identify response regulators whose activities were potentially regulated by these Rap proteins. In particular, we were interested in identifying any Rap proteins that affect ComA-dependent gene expression.
Overproduction of each Rap protein resulted in changes in the expression of several genes. RapH overproduction caused small changes in expression of 15 genes; there was no significant overlap between those genes affected by RapH overproduction and characterized response regulator regulons (Fig. 1A; see Table S1 in the supplemental material). These observations are in contrast to the results recently reported by Hayashi et al., who observed that rapH overexpression inhibited srfA, rapF, and sacB expression (24). Reanalysis of the nucleotide sequence of our rapH overexpression construct, which was constructed based on the previously published sequence for the rapH region (30) and not the newly reported sequence of Hayashi et al. which identified multiple sequence errors in this region (24), revealed that our construct overexpresses both rapH and phrH. Therefore, we believe that the small changes in gene expression that were observed were due to efficient inhibition of RapH's activity by the cooverexpressed PhrH peptide.
RapJ overproduction affected the expression of 39 operons, including 20 that are known to be regulated by Spo0A. Expression of 17 Spo0A-activated operons decreased in cells overproducing RapJ, and expression of 3 Spo0A-repressed operons increased (Fig. 1A; see Table S1 in the supplemental material). It is not known whether RapJ directly or indirectly inhibits the activity of Spo0A; further studies will be needed to determine how RapJ affects Spo0A-dependent gene expression.
Overproduction of RapF or RapK inhibited the expression of genes known to be activated by ComA (Fig. 1A to C and Table 3). The effects of RapF overproduction on ComA-dependent gene expression are consistent with previously published data (6). To understand how multiple Rap proteins and Phr peptides affect the activity of ComA, we further explored the roles that RapF, RapK, PhrF, and PhrK play in regulating ComA-dependent gene expression.
|
0.8), when expression of ComA-dependent genes is still increasing to its maximal level. Gene expression changes were measured under these conditions to avoid analyzing cells transitioning from exponential-phase growth to stationary-phase growth, when patterns of gene expression change dramatically. RapF may also affect transcription independently of ComA, because four of the operons whose expression changed in response to RapF overproduction are not regulated by ComA. Effects of RapK overproduction on gene expression. The expression of 37 operons changed significantly in response to RapK overproduction (Fig. 1A and C and Table 3; see Table S1 in the supplemental material). Fourteen of the 30 operons whose expression decreased are activated by ComA, including eight operons that are also activated by Spo0A. RapK overproduction inhibited the expression of three additional operons that are activated by Spo0A and activated expression of four operons that are repressed by Spo0A. Based on these results, we infer that RapK inhibits the activity of ComA, either directly or indirectly, as the majority of operons whose expression demonstrated the largest decreases in expression are activated directly by ComA and indirectly by Spo0A (see Table S1 in the supplemental material). However, several genes whose expression changed in response to overexpression of rapK are not part of the ComA regulon (Fig. 1A and C and Table 3; see Table S1 in the supplemental material), indicating that RapK likely affects the activity of additional regulators. Spo0A is likely another direct or indirect target of RapK, as several of the operons affected by rapK overexpression are known to be regulated by Spo0A. Nevertheless, the effect of RapK overproduction on the expression of ComA-regulated genes is unlikely to be due solely to effects on Spo0A's activity, as many of the ComA-regulated genes whose expression changed in response to overproduction of RapK did not change in response to overproduction of RapJ, which also inhibited Spo0A-dependent gene expression (Fig. 1A; see Table S1 in the supplemental material).
Overproduction of RapF and RapK inhibits expression of Ppel-lacZ. In order to further investigate the roles that RapF and RapK play in regulating ComA-dependent gene expression, we monitored the effects of RapF and RapK overproduction on the expression of the ComA-activated gene pel by using a fusion of the pel promoter to the reporter gene, lacZ. In contrast to the microarray experiments, which monitor gene expression at only a single time, the Ppel-lacZ fusion provides a dynamic picture of the effects of RapF or RapK overproduction on gene expression throughout growth. We compared expression of Ppel-lacZ in cells overexpressing rapF or rapK to that in control cells with an empty overexpression vector (Fig. 2). Consistent with previous observations (9), pel expression was initially low and increased with increasing cell density in control cells (Fig. 2). Other ComA-regulated genes exhibit similar patterns of expression (32, 41) (Fig. 3). Overproduction of RapF or RapK prevented the density-dependent increase in pel expression (Fig. 2). A similar effect was observed when rapC, the other known ComA inhibitor, was overexpressed (Fig. 2). These data further demonstrate that RapF and RapK inhibit ComA-dependent gene expression when overproduced and indicate that RapF and RapK could play significant roles in regulating ComA-dependent gene expression.
|
|
phrF and phrK mutants to mRNA levels in wild-type cells. We expected that loss of phrF or phrK would result in decreased ComA-dependent gene expression, as the activity of RapF or RapK should increase due to the absence of their inhibitory peptides. We also tested the effects of deletion of phrC on global gene expression. The expression of several ComA-regulated genes (srfA, rapA, and rapC) is known to decrease in phrC mutants (10, 32, 70); this is due to increased RapC activity (10, 70). We found that deletion of all three phr genes resulted in decreased expression of genes activated by ComA, with phrF and phrK mutants having the largest and smallest decreases in magnitude of expression of ComA-dependent genes, respectively (Fig. 1A and D to F; see Table S1 in the supplemental material).
(i) phrC.
Deletion of phrC significantly changed the expression of 66 operons (Fig. 1A and D and Table 3; see Table S1 in the supplemental material). Twenty-one of the 54 operons whose expression decreased are activated by ComA; these include 11 operons that are activated by Spo0A. Deletion of phrC also resulted in decreased expression of 10 additional operons that are activated by Spo0A. The remaining operons whose expression changed in response to deletion of phrC were either regulated by other response regulators or not known to be regulated by a response regulator (Fig. 1A and D and Table 3; see Table S1 in the supplemental material).
(ii) phrF.
Deletion of phrF resulted in significant changes in the expression of 72 operons (Fig. 1A and E and Table 3; see Table S1 in the supplemental material). Twenty-four of the 69 operons whose expression decreased in phrF cells are known to be activated by ComA; these include 11 that are also activated by Spo0A. Deletion of phrF also resulted in decreased expression of 10 additional operons that are activated by Spo0A. The remaining operons whose expression changed in response to deletion of phrC were either regulated by other response regulators or not known to be regulated by a response regulator (Fig. 1A and E and Table 3; see Table S1 in the supplemental material).
(iii) phrK.
Deletion of phrK resulted in significant changes in the expression of 40 operons (Fig. 1A and F and Table 3; see Table S1 in the supplemental material). Of the 38 operons whose expression decreased in phrK cells, 16 are activated by ComA, including 8 that are also activated by Spo0A. Deletion of phrK resulted in decreased expression of 14 additional operons that are activated by Spo0A, whereas the remaining operons affected by deletion of phrK are not known to be regulated by response regulators (Fig. 1A and F and Table 3; see Table S1 in the supplemental material).
Targets for Phr peptides and Rap proteins.
Consistent with previous studies and our results from overproduction of RapF and RapK, our analysis of phrC, phrF, and phrK mutants demonstrates that all three Phr peptides play a role in activating expression of genes regulated by ComA. PhrC and PhrF are known to stimulate the expression of ComA-dependent genes by inhibiting RapC and RapF, respectively (6, 70). RapC and RapF have been previously shown to directly interact with ComA and to inhibit its ability to bind to DNA (6, 10). PhrK likely stimulates ComA-dependent gene expression by inhibiting the activity of RapK. RapK inhibits ComA-dependent gene expression either by acting on ComA directly or by affecting the activity of another regulator that is known to modulate ComA-dependent gene expression (14, 20, 32, 46, 67).
RapK may also play a role in directly or indirectly regulating the expression of genes activated by Spo0A, because a number of Spo0A-controlled genes were affected by overproduction of RapK and deletion of phrK. There are two classes of Spo0A-regulated genes: those genes whose transcription is activated or repressed by intermediate levels of phosphorylated Spo0A (Spo0AP), which include genes involved in biofilm formation, competence development, and antibiotic production, and those genes whose transcription is activated or repressed by high levels of Spo0AP, many of which are involved in sporulation (15, 16, 20-22). This regulation is thought to allow cells to progressively respond to decreasing nutrient availability as the amount of Spo0AP increases (8, 15, 16). Under the conditions tested, we expected that the majority of genes whose expression would change would be those that were regulated by intermediate levels of Spo0AP. This is what we observed, as the majority of Spo0A-regulated genes whose expression changed in response to overexpression of rapK or deletion of phrK were those that respond to intermediate levels of Spo0AP. Further work is needed to determine if RapK can inhibit Spo0A under conditions favorable for regulation of genes dependent upon high levels of Spo0AP.
Although one might expect that overexpression of each rap gene and deletion of each cognate phr would have similar effects on gene expression, this is not what we observed. This is likely due to the levels of active Rap protein that were present in each condition. Overexpression of rapF had very modest effects on global gene expression, whereas deletion of phrF had large effects on gene expression. As described above, the modest effects of RapF overproduction on global gene expression were likely due to decreased expression from the native copy of rapF, resulting in a relatively small increase in the ratio of RapF to PhrF. In contrast, any RapF protein that was produced in phrF mutant cells should be active, as it is in excess over its inhibitory peptide. Although overexpression of rapK had a larger effect on ComA-dependent gene expression than did deletion of phrK, these results may be partially explained by the fact that transcription of rapK from its native promoter was likely low at the time at which changes in gene expression were measured, as expression of rapK is thought to be indirectly activated by Spo0A, which activates expression of genes at the end of exponential phase and in early stationary phase (14, 17). This regulation likely explains most of the differences in the magnitude of effects observed in response to overexpression of rapF or rapK and deletion of phrF or phrK.
In addition to genes regulated by ComA, deletion of each phr gene also resulted in changes in the expression of several other genes. Many of these genes were not detected in the rap overexpression experiments. These differentially regulated genes may be regulated by both the rap and phr genes and reflect differences in the way the experiments were performed, i.e., transient overexpression of each rap gene compared to absence of each phr gene throughout growth. Alternatively, each Phr peptide may affect the activity of a protein or proteins in addition to its cognate Rap protein. The PhrC and PhrG peptides affect the activities of proteins in addition to their cognate Rap proteins (52, 55, 70). PhrC interacts with two additional proteins (70), one of which is the indirect Spo0A antagonist RapB (55). Therefore, the effects of the phrC mutation on Spo0A-regulated gene expression may reflect uninhibited activity of RapB. The other target of PhrC has not been identified, so it is possible that some of the gene expression changes observed in the phrC mutant could reflect changes in the activity of this protein.
PhrC, PhrF, and PhrK play different roles in stimulating ComA-dependent gene expression.
The results of our microarray analysis of the phrC, phrF, and phrK mutants indicated that the phrF mutation had the largest effect on the expression of genes activated by ComA under the conditions tested, i.e., mid-to-late exponential growth in defined minimal medium (OD600 of 1). However, as peptide levels change throughout growth, possibly affecting gene expression to different extents, we used a fusion of lacZ to the ComA-dependent promoter srfA (PsrfA-lacZ) to monitor the effects of phr null mutations throughout growth (Fig. 3A).
As described previously (70), we found that srfA expression was 2-fold lower in phrC mutants than in the wild type and that the initial increase in srfA expression was slightly delayed (Fig. 3A). In phrF mutant cells, srfA expression was reduced 10- to 20-fold and the initial increase in srfA expression was delayed several generations relative to that in wild-type cells. Furthermore, significant differences in the level of srfA expression were observed in phrF cells at lower cell densities, a condition where there were much smaller differences in the magnitude of srfA expression between wild-type, phrC, and phrF cells (Fig. 3A, inset). In the phrK mutants, srfA expression was 75% of that in wild-type cells and was slightly delayed.
These results provide further evidence that all three Phr peptides are required for full levels of ComA-dependent gene expression and that PhrC, PhrF, and PhrK stimulate ComA-dependent gene expression to different magnitudes. Furthermore, the effects of the phrC, phrF, and phrK mutations on srfA expression were dependent upon the presence of their cognate rap genes, as srfA expression was not inhibited when both rap and phr were inactivated [(rapF phrF) (Fig. 3B) and rapK phrK (data not shown)]. These observations, as well as previously published results (70), indicate that ComA-dependent gene expression decreases in phrC, phrF, and phrK mutants due to the increased activities of RapC, RapF, and RapK, providing further evidence that RapC, RapF, and RapK regulate ComA-dependent gene expression.
Multiple Phr peptides act independently to inhibit srfA expression.
We also examined the effects of multiple phr deletions on srfA expression to determine if each phr gene stimulated ComA-dependent gene expression independently. We found that srfA expression was less in phrC phrK double mutant cells than in phrC or phrK single mutants (Fig. 3A), indicating that PhrC and PhrK act independently to stimulate ComA-dependent gene expression. Adding additional phr mutations (phrC, phrK, or phrC and phrK) to the phrF cells had little effect on srfA expression (Fig. 3C), indicating that in the absence of phrF, neither PhrC nor PhrK significantly stimulates ComA-dependent gene expression. These data provide further evidence of the important role that PhrF plays in stimulating ComA-dependent gene expression and the modest roles played by the PhrC and PhrK peptides.
Deletion of rapF or rapK has no effect on ComA-dependent gene expression.
As overexpression of rapF or rapK or deletion of phrF or phrK resulted in decreased ComA-dependent gene expression, we monitored the effects of deletion of rapF or rapK on ComA-dependent gene expression. Deletion of rapC in otherwise wild-type cells results in increased expression of ComA-dependent genes (10, 70). However, we found that deletion of rapF or rapK had no detectable effect on srfA expression under the conditions tested: growth in minimal medium (Fig. 3B), nutrient broth sporulation medium (DSM) (data not shown), or DSM supplemented with glucose and glutamate (described previously [11]; data not shown). Similarly, Bongiorni et al. (6) did not observe changes in expression of the ComA-dependent gene rapA in rapF mutant cells.
We were also unable to detect additional effects on srfA expression when the rapC mutation was combined with the rapF, rapK, or both the rapF and rapK mutations (Fig. 3B and data not shown). This is in contrast to the results reported by Bongiorni et al. (6), who observed that combining the rapF and rapC mutations resulted in a further increase in expression of rapA. These conflicting observations could be due to differences in the promoters assayed (rapA compared to srfA) or the growth conditions (DSM, in which ComA-dependent gene expression increases at the end of exponential growth, versus minimal glucose medium, in which ComA-dependent gene expression increases during exponential growth) (41, 51). However, because we observed that overproduction of RapF or RapK or deletion of PhrF or PhrK results in decreased ComA-dependent gene expression, we suspect that RapF and RapK have roles in regulating ComA-dependent gene expression under conditions that we have not yet identified.
Rap proteins inhibit ComA-dependent gene expression in strains defective for synthesis and uptake of Phr peptides.
The inability to detect effects of rapF and rapK mutations on srfA expression in cells grown in minimal media was surprising, as our previous results indicated that under these growth conditions, RapC, RapF, and RapK all actively repress srfA expression in the absence of their inhibitory peptides. Therefore, we looked for additional insights into the roles that RapC, RapF, and RapK play in regulating ComA-dependent gene expression by examining the effects of rapC, rapF, and rapK mutations in the presence of mutations, opp or sigH, that affect ComA-dependent gene expression, likely by affecting import and synthesis of Phr peptides, respectively.
(i) opp.
Previous work had shown that the oligopeptide permease (Opp) is required for competence development, sporulation, and expression of srfA (20, 65). The requirement for opp for sporulation is bypassed by deleting both rapA and rapB, indicating that the major (and probably only) role for Opp in sporulation is to inhibit the activities of the RapA and RapB proteins, likely by importing the PhrA and PhrC peptides, respectively (59). However, the role for Opp in competence development and srfA expression is less clear. Null mutations in rapC do not bypass the need for opp for expression of srfA or competence development. It seemed likely that opp null mutants had decreased expression of ComA-activated genes due to uninhibited activity of Rap proteins other than or in addition to RapC (34, 70).
Based on our observations that RapC, RapF, and RapK all inhibit expression of srfA and other genes activated by ComA, we thought it was likely that the role that Opp played in regulating srfA expression was to import the peptides that antagonize the activities of RapC, RapF, and RapK. Therefore, we tested the ability of rapC, rapF, and rapK deletions to suppress the defects in srfA expression that occurred in an opp mutant. We found that deletion of rapC, rapK, or both rapC and rapK was not able to suppress the defect in srfA expression observed in opp cells, as srfA was expressed at the same low level in opp, opp rapC, opp rapK, and opp rapC rapK mutant cells (Fig. 3D and data not shown). However, deletion of rapF in opp mutant cells resulted in a significant restoration of srfA expression (Fig. 3D). srfA expression was further enhanced in opp rapF cells by deletion of rapC, rapK, or both rapC and rapK (Fig. 3D). srfA expression in opp rapC rapF rapK cells was slightly higher than that in wild-type cells.
These results indicate that the primary reason that opp mutant cells exhibit low levels of srfA expression is due to uninhibited activity of RapF. In addition, the similarly low levels of srfA expression in phrF and opp mutant cells indicate that the increase in RapF activity in opp mutant cells is primarily due to the inability to import the PhrF peptide. The changes in the timing and level of srfA expression in the opp rapF rapK and opp rapC rapF mutants also correlate well with the changes in the timing and level of srfA expression in the phrC and phrK mutants. Taken together, these results indicate that in the absence of Opp and the import of the PhrC, PhrF, and PhrK peptides, RapF, and to a lesser extent RapC and RapK, inhibit ComA-dependent gene expression.
(ii) sigH.
H directs some of the transcription of all phr genes except phrA and phrH (24, 32, 42, 47). H is also required for full activation of srfA expression (26) and seems to play a role in the posttranslational production of mature PhrC peptide (32). We tested whether the decrease in srfA expression that occurs in sigH mutants was due to the increased activity of RapC, RapF, or RapK as a result of decreased inhibition by their cognate peptides. We found that deletion of rapF, rapK, or both rapF and rapK in sigH mutants had no effect on srfA expression (Fig. 3E and data not shown). However, deletion of rapC in sigH mutants restored srfA expression to levels slightly higher than wild-type levels (Fig. 3E). Deletion of rapF or rapK in addition to rapC in sigH cells resulted in levels of expression similar to those as observed in sigH rapC cells (data not shown), whereas deletion of rapC, rapF, and rapK in sigH mutants resulted in a further increase in the levels of srfA expression (Fig. 3E). These results indicate that the primary defect in srfA expression in sigH mutants is due to increased RapC activity but that increased RapF and RapK activities also contribute to the lower levels of srfA expression in sigH mutant cells in the absence of RapC.
These results, compared to the results observed with deletion of rap genes in opp mutant cells, also indicate that there may be different requirements for H activity among the Phr peptides, with some Phr peptides being more dependent upon H activity, at the level of either phr transcription or posttranslational processing. phrA and phrH do not possess H-dependent promoters and instead appear to be transcribed only from the upstream rap promoters (24, 42, 59). phrE and phrI are transcribed from at least one promoter in addition to the H-dependent promoter (42). As the role that H plays in the posttranslational processing of PhrC and potentially other peptides has not been elucidated, it is currently not clear whether different Phr peptides may be more dependent upon H activity for posttranslational processing.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
The reason that RapF has such profound effects on ComA-dependent gene expression in the absence of PhrF, but not in its presence, may be due to higher intracellular levels of the PhrF peptide at low cell densities, higher sensitivity of RapF protein to low levels of PhrF, a combination of these two factors, or an additional factor that inhibits the activity of RapF under these conditions. If PhrF peptides accumulate in the cell at a lower cell density than PhrC peptides, then RapF activity could be inhibited at lower cell densities than those observed for RapC. Accumulation of PhrF peptides at lower cell density could be due to higher levels of PhrF transcription, processing, or import into the cell. Differences in the transcription or processing of PhrC and PhrF may explain the differences in the sensitivity of sigH mutants to the activities of RapC and RapF (Fig. 3E); differences in the import of Phr peptides may explain the observation that mutations in the oligopeptide permease that do not respond to the PhrC peptide but still allow more significant levels of ComA-dependent gene expression than in opp cells can be obtained (69). Similarly, if RapF were more sensitive to low levels of Phr peptide than RapC, then RapF's activity could be inhibited at lower cell densities than those observed for RapC. Consistent with these hypotheses, we observed that srfA expression was lower in phrF mutants than in phrC mutants or wild-type cells at the earliest time points assayed (Fig. 3A, inset).
Integration of multiple signals for complex regulation of gene expression.
The involvement of multiple peptides in the activation of ComA provides the opportunity for a variety of physiological signals to modulate the magnitude and timing of ComA-dependent gene expression (Fig. 4). Although ComX production appears to occur at a consistent rate throughout growth (3), the abundance of Rap proteins and Phr peptides is regulated at the level of transcription by proteins that respond to different cellular signals (Fig. 4). Transcription of rapC and rapF is activated by ComA (9, 32, 54), which establishes a negative autoregulatory loop for transcription (Fig. 4). rapC transcription is also repressed by CodY, a protein that is active when cellular pools of branched-chain amino acids and GTP are high (reviewed in reference 72) (Fig. 4). rapK is thought to be activated indirectly by Spo0A (14, 46). Several factors are known to regulate the transcription and activity of Spo0A: CodY represses transcription of Spo0A, RNA polymerase containing H provides additional transcription of Spo0A, high population density signals and ongoing DNA replication stimulate the activation of Spo0A, and DNA damage inhibits the activity of Spo0A (reviewed in references 8 and 17) (Fig. 4). Furthermore, transcription from H-dependent promoters affects the levels of phrC, phrF, and phrK transcripts (32, 42, 47), in addition to the factors that regulate expression of their upstream rap genes (Fig. 4). H is also regulated at the level of transcription and activity by a variety of physiological signals, including indirect activation of sigH transcription by Spo0A and inhibition of H activity by certain carbon sources and low pH (reviewed in reference 8) (Fig. 4).
Differential regulation could also occur during production of the active PhrC, PhrF, and PhrK peptides. Processing of Phr peptides to their mature forms is thought to rely upon the secretion machinery and at least one extracellular protease (31, 56, 75); these proteins may exhibit different specificities towards each peptide. In addition, differential import into the cell through the oligopeptide permease may also contribute to variation in the ability of Phr peptides to antagonize the activities of their partner Rap proteins. Therefore, it is likely that this regulatory network serves to modulate the magnitude and timing of the ComA response under a variety of different conditions.
Similarly, the activity of the Spo0A protein, which activates expression of genes involved in sporulation and other post-exponential-phase processes, is regulated indirectly by multiple Phr peptides that act through Rap proteins that inhibit the activity of Spo0F (27, 55, 57, 59). The involvement of multiple Rap proteins and Phr peptides allows integration of additional signals into this regulatory network, as rapA and rapE expression is controlled by ComA, rapB is expressed during exponential phase, and transcription of phrC and phrE is stimulated by RNA polymerase containing H (27, 32, 42, 47, 49, 57).
The involvement of multiple quorum-sensing signals in coordinating biological responses is not unique to B. subtilis. In Pseudomonas aeruginosa, a complex network involving at least three quorum-sensing signals modulates expression of several virulence genes (reviewed in references 28, 68, and 80). In several of the Vibrio species, two or three quorum-sensing signals control specific biological responses, including bioluminescence in V. harveyi and V. fischeri (see references 25 and 38 and references therein) and virulence in V. harveyi and V. cholerae (25, 44). In both cases, it is thought that the involvement of multiple signals plays a role in fine-tuning the level and timing of responses to specific conditions (36, 39, 45, 61, 68, 80). In all these systems, the utilization of quorum-sensing systems that integrate multiple signals provides the cells the ability to modulate specific biological responses under a variety of conditions.
|
ACKNOWLEDGMENTS |
|---|
J.M.A. was supported in part by an NIH predoctoral training grant, and this work was supported in part by Public Health Service grant GM50895 from the NIH to A.D.G.
|
FOOTNOTES |
|---|
Supplemental material for this article may be found at http://jb.asm.org/.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
) (wild type [wt]), Pspank(hy)-rapC (
) (rapC++), Pspank(hy)-rapF (
) (rapF++), and Pspank(hy)-rapK (
) (rapK++) cells is shown. The inset shows data replotted with the y axis from 0 to 1.5 units of ß-galactosidase specific activity.
), 
), 
), 
), and 
), and 