• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. Jul 2011; 193(14): 3588–3597.
PMCID: PMC3133333

Substitutions in the Presumed Sensing Domain of the Bacillus subtilis Stressosome Affect Its Basal Output but Not Response to Environmental Signals [down-pointing small open triangle]

Abstract

The stressosome is a multiprotein, 1.8-MDa icosahedral complex that transmits diverse environmental signals to activate the general stress response of Bacillus subtilis. The way in which it senses these cues and the pathway of signal propagation within the stressosome itself are poorly understood. The stressosome core consists of four members of the RsbR coantagonist family together with the RsbS antagonist; its cryo-electron microscopy (cryoEM) image suggests that the N-terminal domains of the RsbR proteins form homodimers positioned to act as sensors on the stressosome surface. Here we probe the role of the N-terminal domain of the prototype coantagonist RsbRA by making structure-based amino acid substitutions in potential interaction surfaces. To unmask the phenotypes caused by single-copy rsbRA mutations, we constructed strains lacking the other three members of the RsbR coantagonist family and assayed system output using a reporter fusion. Effects of five individual alanine substitutions in the prominent dimer groove did not match predictions from an earlier in vitro assay, indicating that the in vivo assay was necessary to assess their influence on signaling. Additional substitutions expected to negatively affect domain dimerization had substantial impact, whereas those that sampled other prominent surface features had no consequence. Notably, even mutations resulting in significantly altered phenotypes raised the basal level of system output only in unstressed cells and had little effect on the magnitude of subsequent stress signaling. Our results provide evidence that the N-terminal domain of the RsbRA coantagonist affects stressosome function but offer no direct support for the hypothesis that it is a signal sensor.

INTRODUCTION

In Bacillus subtilis and its close relatives, the stressosome is a large cytoplasmic complex that converts diverse environmental signals, such as acid, ethanol, heat, or salt stress, to an output that activates the general stress factor σB (reviewed in references 12, 20, and 30). Key stressosome components are encoded by clustered genes in a variety of bacteria, including many that lack the σB transcription factor; these clusters are found in contexts suggesting involvement in a range of signaling pathways (29). The stressosome therefore represents a sensory machine whose output can be adapted to different tasks. However, there is limited understanding of the internal operation of the stressosome, even in the B. subtilis model.

Genetic and biochemical studies have shown that the B. subtilis stressosome comprises at least three protein species: the RsbT serine/threonine kinase, the RsbS antagonist, and one or more members of the RsbR coantagonist family (1, 9, 13, 24, 32). In vivo each stressosome appears to contain a mixture of the four partially redundant RsbR coantagonists, RsbRA, RsbRB, RsbRC, and RsbRD, and in vitro these coantagonists can dynamically exchange into the complex (13, 24). Cryo-electron microscopy (cryoEM) images of a minimal stressosome formed in vitro from purified RsbRA, RsbS, and RsbT are consistent with a structure in which the STAS (sulfate transporter anti-anti-σ) domains of 20 RsbRA dimers and 20 RsbS monomers form an essentially icosahedral core that can bind 20 RsbT monomers, suggesting a mass of 1.8 MDa (26). According to the current model of the signaling network, RsbT is bound to RsbS in unstressed cells and is released when stress is sensed (Fig. 1). Free RsbT then binds and activates the RsbU environmental phosphatase, inducing the stress response via a signaling cascade that ultimately frees σB from its cognate anti-σ factor (20, 30).

Fig. 1.
σB regulatory network. (A) Model of signaling pathways that regulate σB (20, 30). Environmental and energy pathways converge on RsbV anti-anti-σ and RsbW anti-σ, which directly regulate σB. The multiprotein stressosome ...

The stressosome is thought to regulate the kinase activity of RsbT during the response, and RsbT phosphorylates the STAS domains of the RsbS and RsbR proteins on conserved serine and threonine residues, contributing to RsbT release (10, 16, 18, 35). Alteration of the conserved serine in RsbS (S59A) significantly weakens activation of the response, suggesting that S59 phosphorylation is important but not essential for signaling in otherwise wild-type cells (23). Similar alterations to conserved threonines in RsbRA, the prototype of the RsbR family, support the inference that phosphorylation of T171 is a prerequisite for signaling but does not by itself trigger the response; in contrast, the additional phosphorylation of T205 appears to attenuate the response in the face of extreme environmental stress (16, 24).

How might the stressosome sense input signals? In contrast to the RsbS antagonist, whose structure consists of a single STAS domain, the RsbR coantagonists each have two domains: a C-terminal STAS domain and an N-terminal, nonheme globin domain (26, 28). The isolated N-terminal domain of RsbRA dimerizes in vitro, and cryoEM images of wild-type and mutant stressosomes indicate that these N-terminal dimers form turret-like projections that extend outward from the icosahedral core. Marles-Wright et al. (26) therefore proposed that the N-terminal, nonheme globin domains of the RsbR proteins are positioned to act as sensors, passing the signal to their C-terminal STAS domains via a 13-residue helical connector. Presumably, conformational changes in the STAS domains of the RsbR proteins are communicated to the STAS domains of neighboring RsbS proteins, promoting RsbT release.

Although this proposal is attractive, there is presently limited experimental evidence to support it. In one study, Murray et al. (28) found that some substitutions within the N-terminal domain of RsbRA prevented RsbT binding to the stressosome in vitro, whereas other substitutions did not. However, none of these was assayed to determine the phenotype in vivo. In another study, Reeves and Haldenwang (31) identified one substitution in the N-terminal region of RsbRA that elicited high σB activity in unstressed cells, but this change appeared to have only minor influence on environmental signaling.

Here we probe the in vivo function of the N-terminal domain of RsbRA by assaying the same substitutions that emerged from these prior studies, using a genetic background in which their effects would be clearly apparent. We included in our study additional alterations to the domain surface that might be expected to affect dimerization or interaction with other signaling components. Although some of the tested substitutions had a striking influence on function in unstressed cells, none significantly affected environmental signaling.

MATERIALS AND METHODS

Bacterial strains and genetic methods.

B. subtilis strains are shown in Table 1; plasmids used for strain construction are shown in Table 2. We used standard recombinant DNA methods (33) and performed natural transformation of B. subtilis strains as described by Dubnau and Davidoff-Abelson (14). Deletion and point mutations were introduced into B. subtilis strains using the I-SceI-mediated allele exchange method of Janes and Stibitz (22). For these exchanges, pSS4332 was the source of I-SceI restriction enzyme, and we constructed the companion integrative plasmid pTG5916 by adding an I-SceI site to pUS19. pTG5916 served as the basis for all plasmid constructions containing mutated rsb genes, as follows. (i) In-frame deletions in rsbRA, rsbRB, and rsbRD were created with four-primer PCR (21), and the appropriate fragments were cloned into pTG5916. (ii) Site-directed mutagenesis used a QuikChange Lightning kit (Stratagene, La Jolla, CA) to alter the pTG5923 template, which carried an rsbRA fragment cloned into pTG5916. (iii) The RsbS S59A substitution was introduced using pTG6009, which carried a mutant rsbS amplified by PCR from strain PB470 and cloned into pTG5916. In these strains, σB activity was measured indirectly using a single-copy transcriptional fusion between the well-characterized ctc promoter and a lacZ reporter (8).

Table 1.
Bacillus subtilis strains
Table 2.
Plasmids used for strain construction

β-Galactosidase accumulation assays.

Shake cultures were grown at 37°C to mid-exponential phase in buffered Luria broth lacking salt (BLB) (7) and then diluted 1:25 into fresh BLB. Culture densities were monitored with a Klett-Summerson colorimeter equipped with a number 66 (red) transmission filter. Samples were collected from unstressed cells during exponential growth up to a density of 20 Klett units, when different amounts or kinds of stressors were added to the final concentrations indicated in the figures. All samples were treated essentially according to Miller (27), as previously described (24). Protein concentrations were determined with a Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, CA), and enzyme activity was defined as ΔA420 × 1,000 min−1 mg−1. Assays were done under moderate white light (3 to 4 μmol m−2 s−1) from fluorescent room illumination, which would normally saturate the YtvA blue-light activator of σB (4). Light intensity was measured using a Black Comet model CXR-SR spectroradiometer equipped with a CR2 UV-Vis-NIR cosine receptor (Stellar Net Inc., Tampa, FL).

Basal activity was defined as the activity in unstressed cells sampled at 20 Klett units; stress activation was defined as the difference between this basal value and maximum activity after ethanol or salt stress. The PB1078 parent used to assess the effects of RsbRA substitutions (and bearing null rsbRB, rsbRC, and rsbRD alleles) had an 8-fold-higher basal activity than the PB198 wild-type control (with its full complement of four RsbR coantagonists): 84 versus 10 units. However, the PB1078 parent was similar to the PB198 wild type with respect to activation in response to 4% ethanol stress: 1,282 versus 1,224 units. These activities are the averages from at least five experiments.

Detection of RsbRA by Western blotting.

Mouse monoclonal anti-RsbRA antibody was kindly provided by William Haldenwang (15). Antibody specificity was verified and Western blots were done as previously described (24). Briefly, cells were grown at 37°C to mid-exponential phase in BLB, harvested by centrifugation, and broken by sonication. Protein samples (40 μg) from wild-type and mutant extracts were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Bio-Rad). Detection of cross-reacting material using primary and secondary antibody and an ECL Plus kit (Amersham Pharmacia Biotech, Piscataway NJ) was carried out as previously described (24).

Selection of candidate interaction sites on the N-terminal domain of RsbRA.

In addition to examining residues previously identified by others, we manually inspected the dimer crystal structure (Protein Data Bank [PDB] code 2BNL [28]) for surface features characteristic of protein-protein contacts. Specifically, we looked for exposed nonpolar surfaces and nearby concentrations of charged side chains. Five separate surfaces were targeted; the asterisk within each category indicates a polar or amphiphilic residue substituted with alanine or a nonpolar residue substituted with arginine. (i) L55* and V57* comprise a solvent-accessible hydrophobic patch along with Y8, A12, L54, and A62; (ii) Y35* is prominently displayed in the so-called CD corner (28), with (iii) a nearby hydrophobic patch composed of L37, V41*, and I77; (iv) on the periphery, far from the dimer groove, K47* forms a salt bridge with D51 and is bounded by nonpolar L19, I50, L54, and L55; and (v) near the dimer interface and on the opposite side from the α-helical linker to the STAS domain, E108* flanks a disordered charged loop and is adjacent to E111. The three-dimensional locations of these residues are displayed in Fig. 2.

Fig. 2.
Locations of altered residues within the homodimer formed by the nonheme globin domain of RsbRA. The three-dimensional structure (PDB code 2BNL) is from Murray et al. (28). The surface of one monomer is shown in gold, the other in silver. (Top) View showing ...

RESULTS

Phenotypes of mutants with dimer groove substitutions do not correlate with predictions of previous in vitro assays.

Based on the structural similarities between the N-terminal, nonheme globin domain of RsbRA and related domains of other proteins, such as KaiA and HemAT, Murray et al. (28) proposed that the prominent groove formed by an N-terminal homodimer (Fig. 2) provides a site of interaction both for hypothetical effector proteins and for the RsbT kinase. Effector binding was suggested to displace RsbT from the stressosome and activate environmental signaling. Consistent with this proposal, Hardwick et al. (19) noted structural similarities between the N-terminal domain of RsbRA and the established site of interaction of RsbT with its alternate binding partner, RsbU. Moreover, any of three substitutions located on the surface of the N-terminal dimer groove (E60A, K82A, and E126A) prevented RsbT binding to the stressosome in vitro, whereas two other groove substitutions (T86A and N129A) were indistinguishable from wild-type RsbRA in this regard (28).

We further tested this proposal by determining the in vivo phenotypes caused by the same five N-terminal substitutions studied by Murray et al. (28). Our experimental approach considered two salient features of the system. First, rsbRA mutations must be evaluated in their normal chromosomal context, in which the rsbRA-rsbS-rsbT genes are cotranscribed and translated (31). Second, the negative function of RsbRA is partly redundant with its paralogs RsbRB, -C, and -D (24, 32). Therefore, to reveal the true phenotype caused by each N-terminal substitution, we exchanged the allele of interest with the wild-type, chromosomal copy of rsbRA in a strain missing the other three members of the coantagonist family. We then measured the effect of each substitution using a single-copy transcriptional fusion whose expression was fully dependent on σB. Contrary to expectations, predictions from the earlier in vitro assays were not confirmed by the phenotypes we observed.

Substitutions unable to bind RsbT in vitro would be expected to generate high σB activity in vivo, even in unstressed cells, whereas those that avidly bound RsbT would behave like the parent strain. However, we found only modest phenotypes, and these had little correspondence to the results of the earlier study. For example, the T86A and N129A substitutions, which effectively bound RsbT when formed into stressosomes in vitro (28), nonetheless manifested a 2- to 3-fold increase in σB activity in unstressed cells (Fig. 3 A). This elevated activity was greater than that elicited by the E60A and E126A substitutions, which were unable to bind RsbT in vitro. In contrast, the K82A substitution, grouped by Murray et al. (28) with E60A and E126A, gave the strongest phenotype, with a 4- to 5-fold increase in unstressed cells. We conclude that the phenotype of elevated σB activity in unstressed cells was not simply correlated with the previously measured biochemical property of mutant stressosomes to bind RsbT in vitro.

Fig. 3.
Effects of dimer groove substitutions on σB activity in stressed and unstressed cells. (A) Relative basal activity in unstressed cells, with the white bar showing the parent strain (PB1078, encoding only the RsbRA coantagonist) taken as 1. Light-gray ...

We also tested the effects of each of the five alleles on response to 4% ethanol stress. Here also, there was no correlation with the previous in vitro assays. Moreover, the mutant alleles generally had a weaker and nonparallel influence on stress response (Fig. 3B) than they did on basal activity in unstressed cells (Fig. 3A). This was particularly evident for the K82A allele, which had the strongest effect on basal activity but no significant impact on stress response. To better illustrate this result, we show in Fig. 3C results from a typical full assay of the mutant bearing K82A. Here the maximum response to 4% ethanol stress was shifted only by the difference in basal level between the K82A strain and its parent, and the stress-induced increase in σB activity was the same in both strains. The results of multiple similar assays are summarized in the bar graphs of Fig. 3A and B, both for the K82A mutant and for the other mutants. We conclude that the ability to respond to 4% ethanol stress in vivo is not correlated with the previously measured ability of the mutant stressosomes to bind RsbT in vitro.

We should note that, in addition to RsbRA, the strains in these experiments contained another RsbR paralog within their stressosomes—the YtvA blue-light sensor (3, 17). Genetic analysis suggests that YtvA does not function as a coantagonist like RsbRA, -B, -C, and -D and instead has only a positive role in σB activation (1, 3, 5, 17). To determine if the presence of YtvA somehow contributed to the discrepancy between our in vivo results and the in vitro results of Murray et al. (28), we introduced a ytvA deletion into the strain background and repeated the in vivo assays. As expected, σB activity decreased in all strains (data not shown). However, the relative activities in strains bearing the N-terminal substitutions remained the same as those in the assays with results shown in Fig. 3. Because the presence of YtvA improved the sensitivity of the assay without otherwise affecting the relative order of the results, we used strains wild type at the ytvA locus for all remaining experiments.

Substitutions predicted to affect dimerization significantly influence signal output in unstressed cells.

The five substitutions examined thus far all lie near one another on the surface of the dimer groove (Fig. 2). Inspection of the available crystal structure suggested that these residues also form an intradimer hydrogen-bonded network. K82 is central to this network and makes extensive intersubunit hydrogen-bonded contacts with E60, E126, and either T125 or P122. K82 positioning may be further stabilized by intra- and intersubunit hydrogen bonds with T86 and N129, respectively. The intricacy of this network suggests that the K82A substitution would significantly affect dimer stability or structure.

We therefore wished to explore the effect of another substitution predicted to affect dimer strength but not associated with the K82 network. K93 lies on the surface of the dimer groove and makes a buried, intersubunit salt link with D117; it is separated from K82 by 17 Å and lies in a different molecular context (Fig. 2). When assayed in unstressed cells of a strain encoding only the RsbRA coantagonist, the K93A substitution increased σB activity more than 4-fold relative to the parent—about the same as the K82A substitution (Fig. 4 A). Moreover, the effect of combining K93A with K82A was essentially multiplicative, with a 15-fold increase over the parent activity.

Fig. 4.
Other substitutions with significant effect on basal activity had little impact on stress response. (A) Relative basal activity in unstressed cells, with the white bar showing the parent strain (PB1078, encoding only the RsbRA coantagonist) taken as 1. ...

For comparison, we also examined a strain bearing the E136K substitution, identified by Reeves and Haldenwang (31) during a random screen for dominant mutations that affect stressosome function. Prior to our study, the E136K mutant was the only N-terminal substitution mutant whose phenotype had been examined in vivo, and this substitution was found to affect σB activity primarily in unstressed cells. However, these earlier assays were done in a strain encoding all RsbR paralogs, and the authors pointed out that any impact on stress signaling could be masked by the redundant function of other family members.

We therefore assayed the E136K phenotype in a strain in which RsbRA was the only functional coantagonist and found that it increased σB activity 20-fold in unstressed cells (Fig. 4A). This is similar to the effect that Reeves and Haldenwang observed in a strain with all four coantagonists. E136 is the last residue of the N-terminal, nonheme globin domain (28), and it lies adjacent to the 13-residue α-helical linker that connects the N- and C-terminal domains (26). Given its position in the crystal structure, E136K is not expected to affect dimerization of the N-terminal domain, and the proximal basis of its dysfunction likely differs from that for K82A or K93A (Fig. 2). Nonetheless, strains bearing either E136K or the K82A K93A double mutation manifested similarly high σB activities in unstressed cells (Fig. 4A). These high basal activities represent about one-third of the activity of the fully deregulated system, defined as the activity observed in strains missing all four RsbR coantagonists or the RsbS antagonist (data not shown).

Notably, even the E136K and K82A K93A substitutions, with their strong impact on basal activity, had less than a 2-fold effect on response to stresses of different strengths and kinds (Fig. 4B to D). These results extend the generality of the phenotypes observed in experiments with results shown in Fig. 3: maximum stress response was shifted only by the difference in basal level between the mutant and its parent, and the stress-induced increases in σB activity were similar in all strains. This phenomenon is underscored in the representative salt stress assay with results shown in Fig. 4E. Here the basal levels of the three strains differ strikingly, with even the maximum responses of the parent and K82A mutant remaining well below the basal level of the E136K mutant, and yet all three have comparable stress responses. Because the tested N-terminal substitutions manifested similar amplitudes of fusion expression following stress, we conclude that none significantly affected the sensitivity of stress detection.

Suppression analysis indicates that the K82A or the E136K protein can form functional stressosomes.

The results in Fig. 3 and and44 show that the tested substitutions had a differential effect on the two states examined: basal output of the system in unstressed cells increased, whereas amplitude of the stress response remained largely unaffected. The RsbRA regulator is known to have both positive and negative roles (2). The positive function is thought to reflect the ability of RsbRA to enhance the phosphorylation of RsbS by RsbT during the stress response, which is associated with RsbT release (10, 18, 25). In contrast, the negative function reflects the requirement for RsbRA to act as a coantagonist with RsbS to sequester RsbT within the stressosome; RsbS alone is unable to effectively bind RsbT in vitro or prevent constitutively high signaling in vivo (9, 24). The N-terminal substitutions could conceivably affect either the positive or the negative function of RsbRA. For example, they could alter the positive function by increasing the phosphorylation of RsbS, leading to greater RsbT release in unstressed cells. On the other hand, they could alter the negative function if the mutant RsbRA proteins were present at less than wild-type levels or were less capable of forming stressosomes in vivo; these defects might impact the ability of the stressosome to bind RsbT and keep basal output low. We performed a genetic suppression experiment to distinguish these possibilities.

Alteration of the conserved serine in RsbS to alanine (S59A) prevents phosphorylation by RsbT (35). We combined the RsbS S59A substitution with each of the four strongest rsbRA mutations in a background in which RsbRA was the only coantagonist present and found that the basal output of K82A- and E136K-bearing strains returned to the low level of the parent (Fig. 5 A). Notably, RsbS S59A did not significantly diminish the ability of these strains to respond to 4% ethanol stress (Fig. 5B). A similar lack of RsbS S59A influence on stress response was reported in a strain engineered to express another single RsbR paralog, RsbRC (24). Thus, the suppression of the basal phenotypes noted in Fig. 5A was not due to a general signaling defect introduced by the RsbS substitution.

Fig. 5.
Genetic suppression analysis indicates that RsbRA proteins with K82A or E136K substitutions form functional stressosomes. (A) Relative basal activities elicited by the indicated RsbRA substitution are shown by light-gray bars; activities conferred by ...

Because RsbS S59A completely suppressed the K82A and E136K phenotypes, we infer that these substitutions increase basal output by increasing the phosphorylation level of RsbS S59 in unstressed cells. Although we cannot eliminate the possibility that RsbS S59A acts as a bypass suppressor, counteracting the effects of K82A and E136K by a means unrelated to their primary defects, its ability to fully suppress two such quantitatively different phenotypes argues that RsbS S59 lies on the signaling pathway directly affected by K82A or E136K. More importantly, because neither wild-type RsbS nor its S59A form can by itself reverse the high level of σB activation caused by loss of RsbR coantagonist function (1, 24, 32), this correction of the basal phenotype indicates that the stressosome complement in the suppressed strains was fully capable of binding RsbT and holding the system in a nonsignaling state. We draw the strong inference that K82A and E136K do not adversely affect RsbRA levels or the ability to form functional stressosomes. This inference is in accord with the earlier immunological and biochemical analysis of Reeves and Haldenwang (31), who found that E136K had no observable effect on these properties.

In contrast, the K93A phenotype was not suppressed by the presence of RsbS S59A, and the K82A K93A phenotype was only partly suppressed (Fig. 5A). This partial suppression was consistent with the ability of S59A to correct the K82A but not the K93A defect. These results suggest that K93A acts by a different mechanism than K82A or E136K. However, they provide no information regarding the ability of strains bearing the K93A or K82A K93A substitutions to form functional stressosomes.

In the absence of a positive suppression result for K93A or K82A K93A, we next asked if these substitutions affected RsbRA synthesis or stability. Estimating RsbRA levels by probing cell extracts with a monoclonal anti-RsbRA antibody, we found that strains encoding K93A or E136K manifested about the same RsbRA signal as the parent strain (Fig. 5C). However, strains encoding K82A or K82A K93A had no detectable signal. This result was unexpected because the K82A phenotype was fully corrected by the RsbS S59A suppressor and K82A K93A was partly corrected (Fig. 5A). The suppression results indicate that RsbRA was present in these strains at levels sufficient to form stressosomes capable of binding RsbT. The simplest explanation for the negative Western result is that K82A removes the epitope recognized by the monoclonal antibody.

Other surface substitutions have little effect on system properties.

We wished to explore the notion that other surfaces of the N-terminal domain might be important for environmental signaling, perhaps by providing sites of interaction with other cellular components. The N-terminal domains of all four RsbR coantagonists found in B. subtilis are predicted to be structurally similar, but they share low sequence similarity (28). In the absence of any clues provided by sequence conservation, we manually inspected the crystal structure of the RsbRA N-terminal domain and chose five features for further investigation (see Materials and Methods). These included a tyrosine residue that forms a prominent part of the CD corner, suggested by others (28) as a region of interest, and also residues whose modification was designed to disrupt either a charged cluster or a hydrophobic patch (Fig. 2). When assayed in a strain missing RsbRB, -C, and -D, none of these RsbRA substitutions by themselves had a significant effect on either basal output or stress response (Table 3).

Table 3.
Effect of N-terminal surface alterations on σB activitya

We also attempted to determine the overriding function of the N-terminal domain by means of a large, in-frame deletion that removed the region coding for the nonheme globin and the adjacent 13-residue linker (residues 3 to 145). The remaining STAS domain of RsbRA (residues 146 to 274) is able to assemble into a stressosome when complexed with RsbS in vitro (26). However, we were unable to establish whether this STAS domain could assemble into a stressosome in vivo. We replaced wild-type rsbRA with the allele encoding only the STAS domain in the genetic background in which the other three coantagonists were absent and found that the resulting strain manifested extremely high reporter activity (data not shown). This activity could not be corrected by the RsbS S59A suppressor, nor could we detect any RsbRA signal in Western blots of cell extracts, which was not unanticipated if the deleted K82 in fact represents an epitope required for recognition. Absent any evidence that the mutant protein is present at normal levels or can form a functional stressosome, we cannot address the direct effect the N-terminal deletion might have on signaling.

Mutant phenotypes are masked in the presence of other RsbR paralogs.

Our experimental approach—use of a strain expressing only RsbRA—was designed to uncover the effect of each substitution without interference from the other RsbR coantagonists. For comparison, we also determined the phenotypes caused by selected substitutions in a strain that was otherwise wild type, encoding all of the RsbR paralogs. Notably, the phenotype caused by the K82A substitution, which was the strongest of the five originally studied by Murray et al. (28), was completely masked by the presence of RsbRB, -C, and -D in the wild-type background (Fig. 6). In contrast, the phenotype caused by the E136K substitution was much the same in both backgrounds. This was not surprising, considering that E136K was originally identified and characterized in a wild-type strain (31). To distinguish whether this difference in response in the wild type reflected a qualitative or quantitative effect of the altered residue, we also examined the K82A K93A double substitution. This is akin to K82A in terms of its potential effect on dimerization and groove formation and akin to E136K in terms of the magnitude of its effect on σB activity in unstressed cells. In the wild-type background, the K82A K93A phenotype was also completely masked by the presence of RsbRB, -C, and -D (Fig. 6). Thus, of the substitutions we examined, only the E136K phenotype was apparent in otherwise wild-type cells.

Fig. 6.
The wild-type background masks the phenotypes caused by two representative N-terminal substitutions. Relative basal activity in unstressed cells, with the white bar showing the level for the wild-type strain (PB198, encoding all four members of the RsbR ...

DISCUSSION

Genetic, biochemical, and structural analyses indicate that a minimal stressosome comprises RsbRA, RsbS, and RsbT (9, 23, 24, 26). Moreover, a bioinformatics analysis found homologs of RsbRA, RsbS, and RsbT encoded by contiguous genes in a wide array of bacteria, suggesting that these three proteins form a sensory and transmission module that can be coupled to different signaling networks (29). However, only in B. subtilis and Listeria monocytogenes is the physiological role of the module known, i.e., activation of the σB general stress factor in response to environmental signals (30, 34). We investigated the mechanism of signal transmission in the Bacillus model by making substitutions within the N-terminal, nonheme globin domain of RsbRA, the presumed sensing domain of the module (26). Our study has three findings.

First, even the strongest phenotypes elicited by our rsbRA mutations were masked by the presence of the other RsbR coantagonists (Fig. 6). An exception was the E136K phenotype, identified earlier by Reeves and Haldenwang (31). On the other hand, these authors had indicated uncertainty regarding the effect of E136K on stress activation when assayed in a strain bearing the full complement of RsbR paralogs. Thus, our experimental approach—using a strain expressing only RsbRA—was key to revealing the phenotype caused by each substitution.

Second, the phenotypes elicited by substitutions within the N-terminal dimer groove of RsbRA did not agree with predictions resulting from biochemical analysis of the same mutant proteins assembled into stressosomes in vitro. As shown in Fig. 3, the in vivo assay was required to capture the effect of each substitution on stressosome output. We conclude that the basic ability of wild-type and mutant stressosomes to bind RsbT in vitro, which figured prominently in earlier studies (10, 28), has limited predictive power in vivo. These earlier studies used size exclusion chromatography to qualitatively assess RsbT binding to a stressosome core consisting of RsbRA and RsbS. However, such assays could not control other factors that may affect stressosome function, such as the in vivo concentrations of its constituent proteins, their cellular environment, or their in vivo level of phosphorylation.

In contrast, we assayed the effects of each N-terminal substitution on system output, which takes into account its influence on all elements in the signaling pathway. In this regard, the suppression results shown in Fig. 5 indicate that the elevated basal output elicited by the K82A substitution in RsbRA can be completely corrected by the S59A substitution in RsbS. This result calls into question the earlier interpretation that the K82A substitution significantly interferes with RsbT binding to the N-terminal domain of RsbRA (28) and implies that it acts instead by increasing the phosphorylation levels of RsbS. Consistent with this revised interpretation, Murray et al. (28) were unable to detect interaction between RsbT and the N-terminal domain, suggesting that determinants external to the domain contributed more to the strength of RsbT binding to the stressosome. Moreover, structural studies of a static stressosome showed RsbT positioned over RsbS, not RsbRA (26). Thus, other than the in vitro assay, the interpretation of which is now open to question, there are presently no experimental data to support the model in which the N-terminal domain of RsbRA provides important contacts for RsbT binding, which is then displaced by competing effector proteins (28).

Third, substitutions within the N-terminal region that manifested a significant phenotype elevated basal output of the system only in unstressed cells and had little impact on subsequent stress signaling (Fig. 3 and and4).4). Thus, the stressosome functions we examined—basal output and stress signaling—appear to be genetically separable. One explanation is that a distinct stress-signaling pathway was indeed untouched by our substitutions. In this view, the true stress-signaling pathway remains to be discovered and may not in fact initiate within the N-terminal domain. However, another possibility is that the stressosome itself has considerable signaling capacity and that our substitutions impacted only a fraction of that reserve. Distinguishing these alternatives may involve isolating rsbRA mutants that are unable to signal. The locations within the RsbRA protein of the alterations they encode would also address the question of which regions are involved in signal sensing and which are involved in signal transmission.

Despite this uncertainty, the present analysis indicates that (i) substitutions within the N-terminal domain of the RsbRA coantagonist can influence stressosome function, as reflected by their significant impact on basal output, and (ii) the effects of these substitutions are likely communicated to the RsbS antagonist, as reflected by the ability of the S59A alteration within RsbS to suppress a subset of them. However, the results thus far provide no support for the hypothesis that the N-terminal domain functions as a stress sensor.

ACKNOWLEDGMENTS

We thank William Haldenwang for providing anti-RsbRA antibody, Scott Stibitz for pSS4332, and William Burkholder for pUS19. We also thank Anu Thinda for her assistance in the site-directed mutagenesis and Valley Stewart for his helpful discussions.

This research was supported by Public Health Service grant GM42077 from the National Institute of General Medical Sciences.

Footnotes

[down-pointing small open triangle]Published ahead of print on 20 May 2011.

REFERENCES

1. Akbar S., et al. 2001. New family of regulators in the environmental signaling pathway which activates the general stress transcription factor σB of Bacillus subtilis. J. Bacteriol. 183:1329–1338 [PMC free article] [PubMed]
2. Akbar S., Kang C. M., Gaidenko T. A., Price C. W. 1997. Modulator protein RsbR regulates environmental signalling in the general stress pathway of Bacillus subtilis. Mol. Microbiol. 24:567–578 [PubMed]
3. Ávila-Pérez M., Hellingwerf K. J., Kort R. 2006. Blue light activates the σB-dependent stress response of Bacillus subtilis via YtvA. J. Bacteriol. 188:6411–6414 [PMC free article] [PubMed]
4. Ávila-Pérez M., van der Steen J. B., Kort R., Hellingwerf K. J. 2010. Red light activates the σB-mediated general stress response of Bacillus subtilis via the energy branch of the upstream signaling cascade. J. Bacteriol. 192:755–762 [PMC free article] [PubMed]
5. Ávila-Pérez M., et al. 2009. In vivo mutational analysis of YtvA from Bacillus subtilis: mechanism of light activation of the general stress response. J. Biol. Chem. 284:24958–24964 [PMC free article] [PubMed]
6. Benson A. K., Haldenwang W. G. 1993. Regulation of σB levels and activity in Bacillus subtilis. J. Bacteriol. 175:2347–2356 [PMC free article] [PubMed]
7. Boylan S. A., Redfield A. R., Brody M. S., Price C. W. 1993. Stress-induced activation of the σB transcription factor of Bacillus subtilis. J. Bacteriol. 175:7931–7937 [PMC free article] [PubMed]
8. Boylan S. A., Rutherford A., Thomas S. M., Price C. W. 1992. Activation of Bacillus subtilis transcription factor σB by a regulatory pathway responsive to stationary-phase signals. J. Bacteriol. 174:3695–3706 [PMC free article] [PubMed]
9. Chen C. C., Lewis R. J., Harris R., Yudkin M. D., Delumeau O. 2003. A supramolecular complex in the environmental stress signalling pathway of Bacillus subtilis. Mol. Microbiol. 49:1657–1669 [PubMed]
10. Chen C. C., Yudkin M. D., Delumeau O. 2004. Phosphorylation and RsbX-dependent dephosphorylation of RsbR in the RsbR-RsbS complex of Bacillus subtilis. J. Bacteriol. 186:6830–6836 [PMC free article] [PubMed]
11. Cybulski R. J., Jr., et al. 2009. Four superoxide dismutases contribute to Bacillus anthracis virulence and provide spores with redundant protection from oxidative stress. Infect. Immun. 77:274–285 [PMC free article] [PubMed]
12. de Been M., Francke C., Siezen R. J., Abee T. 2011. Novel σB regulation modules of Gram-positive bacteria involve the use of complex hybrid histidine kinases. Microbiology 157:3–12 [PubMed]
13. Delumeau O., Chen C. C., Murray J. W., Yudkin M. D., Lewis R. J. 2006. High-molecular-weight complexes of RsbR and paralogues in the environmental signaling pathway of Bacillus subtilis. J. Bacteriol. 188:7885–7892 [PMC free article] [PubMed]
14. Dubnau D., Davidoff-Abelson R. 1971. Fate of transforming DNA following uptake by competent Bacillus subtilis. I. Formation and properties of the donor-recipient complex. J. Mol. Biol. 56:209–221 [PubMed]
15. Dufour A., Voelker U., Voelker A., Haldenwang W. G. 1996. Relative levels and fractionation properties of Bacillus subtilis σB and its regulators during balanced growth and stress. J. Bacteriol. 178:3701–3709 [PMC free article] [PubMed]
16. Eymann C., et al. 2011. In vivo phosphorylation patterns of key stressosome proteins define a second feedback loop that limits activation of Bacillus subtilis σB. Mol. Microbiol. 80:798–810 [PMC free article] [PubMed]
17. Gaidenko T. A., Kim T. J., Weigel A. L., Brody M. S., Price C. W. 2006. The blue-light receptor YtvA acts in the environmental stress signaling pathway of Bacillus subtilis. J. Bacteriol. 188:6387–6395 [PMC free article] [PubMed]
18. Gaidenko T. A., Yang X., Lee Y. M., Price C. W. 1999. Threonine phosphorylation of modulator protein RsbR governs its ability to regulate a serine kinase in the environmental stress signaling pathway of Bacillus subtilis. J. Mol. Biol. 288:29–39 [PubMed]
19. Hardwick S. W., et al. 2007. Structural and functional characterization of partner switching regulating the environmental stress response in Bacillus subtilis. J. Biol. Chem. 282:11562–11572 [PubMed]
20. Hecker M., Pané-Farré J., Völker U. 2007. SigB-dependent general stress response in Bacillus subtilis and related gram-positive bacteria. Annu. Rev. Microbiol. 61:215–236 [PubMed]
21. Ho S. N., Hunt H. D., Horton R. M., Pullen J. K., Pease L. R. 1989. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51–59 [PubMed]
22. Janes B. K., Stibitz S. 2006. Routine markerless gene replacement in Bacillus anthracis. Infect. Immun. 74:1949–1953 [PMC free article] [PubMed]
23. Kang C. M., Brody M. S., Akbar S., Yang X., Price C. W. 1996. Homologous pairs of regulatory proteins control activity of Bacillus subtilis transcription factor σB in response to environmental stress. J. Bacteriol. 178:3846–3853 [PMC free article] [PubMed]
24. Kim T. J., Gaidenko T. A., Price C. W. 2004. A multicomponent protein complex mediates environmental stress signaling in Bacillus subtilis. J. Mol. Biol. 341:135–150 [PubMed]
25. Kim T. J., Gaidenko T. A., Price C. W. 2004. In vivo phosphorylation of partner switching regulators correlates with stress transmission in the environmental signaling pathway of Bacillus subtilis. J. Bacteriol. 186:6124–6132 [PMC free article] [PubMed]
26. Marles-Wright J., et al. 2008. Molecular architecture of the “stressosome,” a signal integration and transduction hub. Science 322:92–96 [PubMed]
27. Miller J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
28. Murray J. W., Delumeau O., Lewis R. J. 2005. Structure of a nonheme globin in environmental stress signaling. Proc. Natl. Acad. Sci. U. S. A. 102:17320–17325 [PMC free article] [PubMed]
29. Pané-Farré J., Lewis R. J., Stülke J. 2005. The RsbRST stress module in bacteria: a signalling system that may interact with different output modules. J. Mol. Microbiol. Biotechnol. 9:65–76 [PubMed]
30. Price C. W. 2010. General stress response in Bacillus subtilis and related Gram-positive bacteria, p. 301–318 In Storz G., Hengge R., editors. (ed.), Bacterial stress responses, 2nd ed. ASM Press, Washington, DC
31. Reeves A., Haldenwang W. G. 2007. Isolation and characterization of dominant mutations in the Bacillus subtilis stressosome components RsbR and RsbS. J. Bacteriol. 189:1531–1541 [PMC free article] [PubMed]
32. Reeves A., Martinez L., Haldenwang W. 2010. Expression of, and in vivo stressosome formation by, single members of the RsbR protein family in Bacillus subtilis. Microbiology 156:990–998 [PubMed]
33. Sambrook J., Fritsch E. F., Maniatis T. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY
34. Shin J. H., Brody M. S., Price C. W. 2010. Physical and antibiotic stresses require activation of the RsbU phosphatase to induce the general stress response in Listeria monocytogenes. Microbiology 156:2660–2669 [PMC free article] [PubMed]
35. Yang X., Kang C. M., Brody M. S., Price C. W. 1996. Opposing pairs of serine protein kinases and phosphatases transmit signals of environmental stress to activate a bacterial transcription factor. Genes Dev. 10:2265–2275 [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...