Isolation and Characterization of Bacillus subtilis sigB Operon Mutations That Suppress the Loss of the Negative Regulator RsbX

Abstract

ς^B, a transcription factor that controls the Bacillus subtilis general stress response regulon, is activated by either a drop in intracellular ATP or exposure to environmental stress. RsbX, one of seven ς^B regulators (Rsb proteins) whose genes are cotranscribed with ς^B, is a negative regulator in the stress-dependent activation pathway. To better define the interactions that take place among the Rsb proteins, we analyzed sigB operon mutations which suppress the high-level ς^B activity that normally accompanies the loss of RsbX. Each of these mutations was in one of three genes (rsbT, -U, and -V) which encode positive regulators of ς^B, and they all defined amino acid changes which either compromised the activities of the mutant Rsbs or affected their ability to accumulate. ς^B activity remained inducible by ethanol in several of the RsbX⁻ suppressor strains. This finding supports the notion that RsbX is not needed as the target for ς^B activation by at least some stresses. ς^B activity in several RsbX⁻ strains with suppressor mutations in rsbT or -U was high during growth and underwent a continued, rather than a transient, increase following stress. Thus, RsbX is likely responsible for maintaining low ς^B activity during balanced growth and for reestablishing ς^B activity at prestress levels following induction. Although RsbX likely participates in limiting the ς^B induction response, a second mechanism for curtailing unrestricted ς^B activation was suggested by the ς^B induction profile in two suppressor strains with mutations in rsbV. ς^B activity in these mutants was stress inducible but transient, even in the absence of RsbX.

ς^B is a Bacillus subtilis transcription factor which directs RNA polymerase to the promoters of a large number of genes that are activated at the end of exponential growth or after exposure to any of a number of environmental stresses (6, 8, 10, 17, 18, 21, 30, 34, 37). Induction of this regulon is controlled at the level of ς^B activation (7, 9, 10). ς^B and at least two additional B. subtilis ς factors (ς^F and ς^G) are regulated by specific anti-ς proteins that sequester ς factors, thus blocking their availability to RNA polymerase (3, 9, 11–13, 15, 21, 24, 27). The anti-ς^B protein (RsbW) is one of seven ς^B regulators whose genes (rsb) are cotranscribed with the ς^B structural gene (sigB) (20, 35). These eight genes are arranged in an operon with the sequence P_A rsbR rsbS rsbT rsbU P_B rsbV rsbW sigB rsbX (20, 35). All eight genes are constitutively expressed from a promoter (P_A) that is recognized by the housekeeping ς (ς^A). An internal ς^B-dependent promoter (P_B) positioned between rsbU and rsbV elevates the levels of the downstream four genes during periods of ς^B activity.

A model for the regulation of ς^B by the seven Rsb proteins is depicted in Fig. 1. The anti-ς^B protein (RsbW) can form mutually exclusive complexes with either ς^B or an alternative target, RsbV (2, 9, 12, 34). RsbV availability for binding to RsbW appears to be the critical factor driving the activation of ς^B. RsbW is also a protein kinase which can phosphorylate RsbV and convert it to a form that is no longer able to associate with RsbW (2, 12). It is believed that the relatively high ATP levels which occur during exponential growth favor the phosphorylation of RsbV. This leads to the formation of RsbW-ς^B complexes and the silencing of ς^B-dependent transcription (2, 34). When cells become nutritionally depleted, the drop in ATP causes inefficient phosphorylation of RsbV, the persistence of RsbV-RsbW complexes, and the release of ς^B. The stress-dependent activation of ς^B occurs with little regard for ATP levels (34). Instead, stress activation relies on the reactivation of phosphorylated RsbV (RsbV-P) by an RsbV-P-specific phosphatase (33). Dephosphorylation of existing RsbV-P occurs during both the stress-induced and stationary-phase-induced activation of ς^B; however, it is essential only for induction of the former pathway (33). Both dephosphorylation pathways require one or more of the rsbR, -S, and -T gene products, but only the stress activation pathway appears to require RsbU (29, 33–35). RsbU is believed to be the RsbV-P phosphatase of the stress pathway (37). The stationary-phase pathway’s phosphatase is unknown. In vitro evidence suggested how the remaining Rsb proteins could participate in a stress activation pathway (37). RsbT, a positive regulator of the stress pathway, was found both to stimulate RsbU-dependent dephosphorylation of RsbV-P and to catalyze the phosphorylation of a negative regulator, RsbS. RsbX, an additional negative regulator, could dephosphorylate RsbS-P. It was hypothesized that RsbS is a direct inhibitor of RsbT and that following environmental stress, RsbT inactivates RsbS by phosphorylation (37). Freed from RsbS-dependent inhibition, RsbT then activates RsbU, which in turn dephosphorylates RsbV. RsbX could reset the system by reactivating the RsbS-P to again inhibit RsbT. The function of RsbR is unknown, although recent evidence suggests that it may play a structural role, facilitating interactions between various other Rsb proteins (1).

FIG. 1.

Model of ς^B regulation. (1) The anti-ς^B protein, RsbW (W), can form mutually exclusive complexes with either ς^B or its antagonist, RsbV (V). When bound to RsbW, ς^B is unable to form an RNA polymerase holoenzyme (E-ς^B). (2) RsbW can phosphorylate RsbV, using ATP as a phosphate donor. RsbV-P (V-P) does not bind to RsbW. During growth, relatively high ATP levels favor the phosphorylation and inactivation of RsbV, leaving RsbW free to bind ς^B. When ATP levels fall, the phosphorylation of RsbV may be inefficient, leading to the persistence of active RsbV, the formation of stable RsbV-RsbW complexes, and the release of ς^B. (3) The magnitude of the ς^B activation during low-ATP conditions (e.g., entry into stationary phase) is enhanced by the dephosphorylation of a portion of the preexisting RsbV-P. The mechanism responsible for this is unknown. (6) Environmental stress (e.g., heat shock, osmotic shock, or ethanol treatment) activates an RsbV-P phosphatase, RsbU, which generates a pool of active RsbV, regardless of ATP levels. (4) The phosphatase activator, RsbT (T), is believed to be normally inactive due to an association with its negative regulator, RsbS (S). RsbR (R), an additional regulatory protein, can bind to RsbS or RsbT and is speculated to play a structural role in the stress pathway. When exposed to stress, (5) RsbT is believed to phosphorylate and inactivate RsbS and (6) activate the RsbU phosphatase. (7) RsbS-P is dephosphorylated and reactivated by a phosphatase, RsbX (X), encoded by one of the genes downstream of the sigB operon’s ς^B-dependent promoter. RsbX levels thus increase with increasing ς^B activity. This may serve to limit the activation process and return RsbT to an inactive complex with RsbS.

To better define the interactions that occur during stress-induced activation of ς^B, we attempted to isolate mutations which could suppress the high level of ς^B activity that occurs if B. subtilis loses RsbX function. By the current model, the loss of RsbX should eliminate the cell’s ability to reactivate RsbS-P and lead to elevated activity of ς^B due to RsbT-dependent processes. This communication describes the suppressor mutations that we have identified within the sigB operon. In support of the in vitro model, we isolated mutations in the genes for each of the three positive regulators (RsbT, RsbU, and RsbV) of the stress activation pathway. The profiles of ς^B activity in several of the mutant strains argue that at least some stress induction of ς^B can occur in the absence of RsbX and that RsbX plays a role both in maintaining low ς^B activity during growth and in reestablishing prestress ς^B activity in cells where ς^B activity had been induced. A second, RsbX-independent pathway for limiting ς^B activation was implied by the ς^B induction profile in two suppressor strains with mutations in rsbV. Stress-induced ς^B activity peaked and then fell in these strains, even though RsbX was absent.

MATERIALS AND METHODS

Bacterial strains, plasmids, and cultivation of bacteria.

The B. subtilis strains and plasmids used are listed in Table 1. All BSA strains are derivatives of PY22, which was obtained from P. Youngman (University of Georgia). Bacteria were grown in LB (23) at 37°C. The cells were exposed to ethanol stress during exponential growth by adding ethanol to a final concentration of 4%. Escherichia coli TG2 was used as the host for cloning.

TABLE 1.

B. subtilis strains and plasmids used in this study

Strain or plasmid	Relevant genotype or features	Source, reference, or constructiona
Strains
PY22	trpC2	P. Youngman, University of Georgia
BSA46	trpC2 SPβ ctc::lacZ	6
BSA132	trpC2 rsbV312 rsbX::pWH25b SPβ ctc::lacZ	34
BSA272	trpC2 sigB::Δ HindIII-EcoRV::cat SPβ ctc::lacZ	34
XS1	trpC2 rsbX::spc rsbV74GD SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS2	trpC2 rsbX::spc rsbU44PR SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS4	trpC2 rsbX::spc rsbV85LP SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS5	trpC2 rsbX::spc rsbU80Yterm SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS6	trpC2 rsbX::spc rsbT107VG SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS15	trpC2 rsbX::spc rsbU194VA SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS16	trpC2 rsbX::spc rsbT15IS SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS33	trpC2 rsbX::spc rsbT63Qterm SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS35	trpC2 rsbX::spc ΔrsbST SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS38	trpC2 rsbX::spc rsbU228GR SPβ ctc::lacZ	pML7/X::Spec → BSA46
XS11	trpC2 aph3′5"/sigB rsbX::spc rsbV74GD SPβ ctc::lacZ	pRR7 → XS1
XS21	trpC2 aph3′5"/sigB rsbX::spc rsbU44PR SPβ ctc::lacZ	pRR7 → XS2
XS41	trpC2 aph3′5"/sigB rsbX::spc rsbV85LP SPβ ctc::lacZ	pRR7 → XS4
XS51	trpC2 aph3′5"/sigB rsbX::spc rsbU80Yterm SPβ ctc::lacZ	pRR7 → XS5
XS61	trpC2 aph3′5"/sigB rsbX::spc rsbT107VG SPβ ctc::lacZ	pRR7 → XS6
XS151	trpC2 aph3′5"/sigB rsbX::spc rsbU194VA SPβ ctc::lacZ	pRR7 → XS15
XS161	trpC2 aph3′5"/sigB rsbX::spc rsbT15IS SPβ ctc::lacZ	pRR7 → XS16
XS331	trpC2 aph3′5"/sigB rsbX::spc rsbT63Qterm SPβ ctc::lacZ	pRR7 → XS33
XS351	trpC2 aph3′5"/sigB rsbX::spc ΔrsbS-T SPβ ctc::lacZ	pRR7 → XS35
XS381	trpC2 aph3′5"/sigB rsbX::spc rsbU228GR SPβ ctc::lacZ	pRR7 → XS38
XS12	trpC2 aph3′5"/sigB rsbV74GD SPβ ctc::lacZ	XS11 → BSA272
XS22	trpC2 aph3′5"/sigB rsbU44PR SPβ ctc::lacZ	XS21 → BSA272
XS42	trpC2 aph3′5"/sigB rsbV85LP SPβ ctc::lacZ	XS41 → BSA272
XS52	trpC2 aph3′5"/sigB rsbU80Yterm SPβ ctc::lacZ	XS51 → BSA272
XS62	trpC2 aph3′5"/sigB rsbT107VG SPβ ctc::lacZ	XS61 → BSA272
XS152	trpC2 aph3′5"/sigB rsbX::pWH25 rsbU194VA SPβ ctc::lacZ	XS151 → BSA272
XS162	trpC2 aph3′5"/sigB rsbT15IS SPβ ctc::lacZ	XS161 → BSA272
XS332	trpC2 aph3′5"/sigB rsbX::pWH25 rsbT63Qterm SPβ ctc::lacZ	XS331 → BSA272
XS352	trpC2 aph3′5"/sigB ΔrsbST SPβ ctc::lacZ	XS351 → BSA272
XS382	trpC2 aph3′5"/sigB rsbU228GR SPβ ctc::lacZ	XS381 → BSA272
Plasmids
pBluescript	Apr	26
pUK19	Apr Kanr	19
pML7	Apr Cmr PBrsbV rsbW sigB rsbX	6
pML7/X::Spec	Apr Cmr Spr PBrsbV rsbW sigB rsbX::spec	7
pUR1	Apr Kanr 520-bp DNA fragment with upstream sigB operon region	This study
pRR7	Apr Kanr PA::rsbR	This study
pRS11	Apr Kanr PA::rsbR rsbS	This study
pRT2	Apr Kanr PA::rsbR rsbS rsbT	This study
pRU13	Apr Kanr PA::rsbR rsbS rsbT rsbU	This study
pRV6	Apr Kanr PA::rsbR rsbS rsbT rsbU PB::rsbV	This study
pDG148	Apr Kanr PspaclacI	P. Stragier, Institut de Biologie Physico-Chimie, Paris, France
pBT1	AprrsbT	This study
pDT11	Apr Kanr Pspac::rsbT lacI	This study
pBU1	AprrsbU	This study
pDU3	Apr Kanr Kanr Pspac::rsbU lacI	This study
pACT2	Apr, vector with ADc of GAL4	Clontech Laboratories, Inc.
pAS2-1	Apr, vector with BDd of GAL4	Clontech Laboratories, Inc.
pUV46	Apr, protein fusion of RsbW with BD of GAL4	32
pUV70	Apr, protein fusion of RsbS with AD of GAL4	32
pUV134	Apr, protein fusion of RsbR with AD of GAL4	32
pUV145	Apr, protein fusion of RsbU with BD of GAL4	32
pUV166	Apr, protein fusion of RsbT with AD of GAL4	32
pUV187	Apr, protein fusion of RsbU with AD of GAL4	32
pUV207	Apr, protein fusion of RsbV with AD of GAL4	32
pTV13	Apr, protein fusion of RsbV74GD with AD of GAL4	350-bp BamHI-XhoI fragment of a PCR product generated with the primers RSBV5MM and RSBV3MM (Table 2) cloned into pACT2
pTV13	Apr, protein fusion of RsbV74GD with AD of GAL4	350-bp BamHI-XhoI fragment of a PCR product generated with the primers RSBV5MM and RSBV3MM (Table 2) cloned into pACT2
pST61	Apr, protein fusion of RsbT107VD with BD of GAL4	400-bp BamHI-PstI fragment of a PCR product generated with the primers ALPC6 and ALPC7 (Table 2) cloned into pAS2-1
pSU256	Apr, protein fusion of RsbU44PR with BD of GAL4	1,000-bp SmaI-SalI fragment of a PCR product generated with the primers RSBU5MM and RSBU3MM (Table 2) cloned into pAS2-1
pSU1521	Apr, protein fusion of RsbU194VA with BD of GAL4	1,000-bp SmaI-SalI fragment of a PCR product generated with the primers RSBU5MM and RSBU3MM (Table 2) cloned into pAS2-1

^aAn arrow indicates the construction of the strain by transformation of DNA from the source on the left of the arrow into the strain on the right of the arrow. 

^bpWH25 was integrated into the B. subtilis chromosome by a Campbell-like recombination that did not disrupt rsbX. 

^cAD, activation domain. 

^dBD, DNA binding domain. 

Construction of plasmids for mapping suppressor mutations.

A DNA fragment encoding 360 bases upstream of the sigB operon, plus P_A and the first 160 bp of rsbR, was amplified by PCR using primers UDINDE and UPOPNDE (Table 2). This DNA fragment was inserted into the NdeI site of pUK19, thereby placing the B. subtilis DNA upstream of the plasmid’s kanamycin resistance (Km^r) gene (kan). A clone in which P_A was oriented toward kan was selected (pUR1) and used as a vector for further cloning. Immediately downstream of kan is pUK19’s multiple cloning site. DNA fragments that encoded P_A and increasing lengths of the sigB operon (Fig. 2) were amplified by PCR using the primers described in Fig. 2 and Table 2. These amplified DNAs were inserted into the BamHI and SphI sites of the multiple cloning site of pUR1 to create plasmids pRR7, pRS11, pRT2, pRU13, and pRV6 (Fig. 2). These plasmids were used to introduce wild-type sigB operon genes into the B. subtilis suppressor strains by double-crossover events between linearized plasmids and the suppressor strains’ chromosomes. The appearance of Km^r, small dark colonies following transformation with a particular plasmid indicated that the suppressor mutation was encoded within the group of genes carried by that plasmid.

TABLE 2.

Primers used in construction of mapping plasmids

Primer	Sequence 5′-OH to 3′-OH
UDINDE	CTTCAGCCCATATGCTATTGTT
UPOPNDE	TGAATGTCCATATGTTCTTTAGA
RDIBAM	GCCTTACGGATCCGTTTGGCA
ROPSPH	GATGTCTCAGCATGCTATTCCC
SOPSPH	TTGGTCGTGCATGCGCTATTCC
TOPSPH	CGTTCTTGCATGCTACCGAAGC
UOPSPH	CACGTCTGGCATGCGTTAAC
4LPCVN	CGTTACGCATGCAGCATTATTCTTCATTGC
RSBT5END	AGAGAAAGCTTCTTGAGACATTGAAGCGGG
RSBT3END	AGAGAGGATCCTACCGAAGCCATTTGATGGC
NAT5	ATCATCGGTGCAATTAAACGT
XOPNDE	TCTGGTTTCATATGACAGCTGT
RSBV5MM	AGAGAGAGGATCCGTATGAATATAAATGTTGATGTG
RSBV3MM	AGAGAGCTCGAGTCGACTCATTGCACTCCACCTTC
ALPC6	GTTTCCGGATCCATATGAACGACCAATCCTGTG
ALPC7	ATACCCCTGCAGCTACCGAAGCCATTTGATG
RSBU5MM	AGAGCCCGGGGGTGGATTTTAGGGAGGTT
RSBU3MM	AGAGGAATTCGTCGACGTTAAACCTTTCTCCG
TSEQ2	GGCTAGAATTACAACGGC
ALVHIS	TAAAACAAGCTTATACGTTATGATGATGATGATGATG
	AACCTTTCTCCGCAAAAC

FIG. 2.

Physical map of sigB operon plasmids used for suppressor mapping. The rectangles depict the open reading frames for rsbR, rsbS, rsbT, rsbU, and rsbV. P_A and P_B indicate the ς^A- and ς^B-dependent promoters, respectively. Only the restriction sites used in the construction of recombinant plasmids and the primers used for their amplification are shown. Abbreviations: Sp, SphI; B, BamHI.

Construction of replicating plasmids permitting overexpression of rsbT and rsbU in B. subtilis cells.

rsbT was amplified by PCR using primers RSBT5END and RSBT3END (Table 2). This DNA fragment was inserted into the BamHI and HindIII sites of pBluescript to create plasmid pBT1. Then rsbT was cut from pBT1 by using the XbaI and HindIII sites and inserted downstream of P_spac in the B. subtilis vector pDG148 to create plasmid pDT11. rsbU was amplified by PCR using primers ALVSHIS and TSEQ2 (Table 2) and then cloned into the EcoRI and HindIII sites of pBluescript. The XbaI-HindIII fragment with rsbU from this plasmid (pBU1) was inserted into pDG148 to create pDU3. Multicopy plasmids pDU3 and pDT11 produce elevated levels of RsbU and RsbT, respectively, even in the absence of isopropyl-β-d-thiogalactopyranoside.

Reconstruction of intact rsbX in the suppressor strains.

A Km^r gene was inserted upstream of the sigB operon in the chromosomes of the suppressor strains by transformation with linearized plasmid pRR7 (Fig. 2) to generate strains XS11, XS21, XS41, XS51, XS61, XS151, XS331, XS351, and XS381 (Table 1). Chromosomal DNA from these strains was transformed into either BSA272 (sigB::ΔHindIII-EcoRV::cat) or BSA132 (rsbV312). The transformants were selected on the plates with 5 μg of kanamycin and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) per ml. Blue, Km^r colonies could arise only if the integrated DNA incorporated the Km^r gene that is upstream of the sigB operon and, depending on the recipient, a functional copy of either rsbV or sigB. All of the intervening sigB operon genes (i.e., rsbR, -S, -T, and -U) would have a high likelihood of being transferred from the suppressor strain to the RsbX⁺ recipient. Km^r Spc^s (spectinomycin-sensitive) blue colonies were screened by PCR using primers NAT5 and XOPNDE (Table 2). The presence of RsbX in isolated clones was verified by Western blotting. The suppressor strains constructed with an intact rsbX are listed in Table 1.

PCRs and sequencing DNA.

PCR was performed with AmpliTaq DNA polymerase (Perkin-Elmer) according to standard protocols, using the primers depicted in Table 2.

Sequencing of mutant genes was performed in both directions by the UTHSCSA Center for Advanced Technologies using an Applied Biosystems 373 DNA sequencer, with an ABI PRISM TM Dye Terminator Cycle Sequencing Ready Reaction kit and AmpliTaq DNA polymerase, FS (Perkin-Elmer).

Yeast assays.

Mutant or wild-type rsb genes were cloned into vectors (Matchmaker two-hybrid system; Clontech Laboratories Inc., Palo Alto, Calif.) that created translational fusions between them and the activation or DNA binding domain of the yeast Gal4 activator protein (Table 1). The vectors were transformed into yeast strain Y190 in appropriate combinations (25). Interactions between the various Rsb proteins were assayed by determining Gal4-dependent β-galactosidase activity (25). The yeast strains were grown to exponential phase; 1 ml of each culture was harvested by centrifugation and stored at −20°C. β-Galactosidase was assayed as previously described (32), using chloroform and sodium dodecyl sulfate (SDS) to permeabilize the cells and o-nitrophenol-β-d-galactoside as the substrate. The assays were done on two clones of each pairing in duplicate. The values were calculated as Miller units (1,000 × A₄₂₀/[volume × time × A₆₀₀]) and are given as a ratio of the activity in mutant versus wild-type pairing.

General methods.

SDS-polyacrylamide gel electrophoresis (PAGE), Western blot analysis, and the Bacillus β-galactosidase assays were performed as previously described (13). All DNA manipulations and transformation of E. coli were done according to standard protocols. Transformation of natural competent B. subtilis cells with plasmid and chromosomal DNA was carried out as described by Yasbin et al. (38).

RESULTS

Isolation of mutants.

RsbX is a negative regulator of the stress induction pathway. In the absence of RsbX, ς^B activity becomes high enough to impair growth and result in dark blue pinpoint colonies when such strains, carrying a ς^B-dependent lacZ gene, are plated on LB supplemented with X-Gal (6, 18, 20, 29). B. subtilis BSA46 (SPβ ctc::lacZ) was transformed with linearized plasmid DNA (pML7) carrying a Spc^r marker in rsbX (29) and plated on medium containing spectinomycin and X-Gal. Spontaneous suppressor mutations were seen as larger colonies among the small dark blue rsbX::spc transformants. Although RsbX is a specific regulator of the stress-induced pathway, its loss can be suppressed by any mutation that reduces ς^B activity, including mutations in sigB itself and its principal positive regulator RsbV (6). ς^B is activated by either a stress-dependent or an ATP-responsive pathway (Fig. 1) (34). To heighten our chances of picking interesting mutants, we restricted our analysis to colonies which were at least as blue as control colonies with null mutations in an essential component (e.g., rsbU) of the stress pathway but with an intact ATP-responsive pathway. These would be less likely to contain loss of function mutations in rsbV or sigB.

Chromosomal DNA was prepared from the large blue colony transformants and used to transform wild-type B. subtilis to Spc^r (i.e., RsbX⁻). Cotransformation of the suppressor phenotype with Spc^r was taken as an indication that the suppressor mutation was likely to be within sigB operon genes. Ten mutants which displayed linkage of the suppressor phenotype with Spc^r were selected for further study.

As an initial characterization of the mutant strains, cultures of each mutant were grown to early stationary phase in LB (optical density at 540 nm [OD₅₄₀] of 1.0) and examined by Western blotting for the presence of the eight sigB operon proteins (Fig. 3; summarized in Table 3). As expected, all of the mutant strains were missing RsbX, but several had additional alterations in sigB operon proteins. Three of the mutants (XS1, XS4, and XS15) had elevated levels of sigB operon proteins whose genes are downstream of the operon’s internal ς^B-dependent promoter. This would occur if the suppressor mutations only partially blocked the effect of the loss of RsbX on the level of ς^B activity (i.e., the stress induction pathway is still partially active and responding to the loss of RsbX). Additional changes observed in some of the mutants included the absence of RsbT (XS33), RsbS and -T (XS35), or RsbU (XS5), as well as reduced levels of RsbU (XS38) or RsbV (XS4). Although the Western blot data do not identify the sites of the mutations, they suggest that at least two of the suppressor strains carry mutations in rsbU (XS5 and XS38), two others have changes in rsbT or rsbV (XS33 and XS4, respectively), and another (XS35) has a probable rsbST deletion.

FIG. 3.

Western blot analysis of RsbX⁻ suppressor cell extracts. The cells were grown in LB to an OD₅₄₀ of 1.0, harvested, resuspended in buffer (50 mM Tris HCl [pH 8.0], 0.1 mM EDTA, 5% phenylmethylsulfonyl fluoride), and disrupted by passage through a French press. The extracts were fractionated by SDS-PAGE, transferred to nitrocellulose, and probed by Western blotting using monoclonal antibodies raised against RsbV, RsbW, ς^B, RsbX, RsbR, RsbS, RsbT, and RsbU. The extracts (from left to right) were from strains XS1, XS2, XS4, XS5, XS6, XS15, XS16, XS33, XS35, XS38, and BSA46 (wild type).

TABLE 3.

Western blot analysis of Rsb proteins in RsbX⁻ suppressor strains

Strain	Protein levela
	RsbR	RsbS	RsbT	RsbU	RsbV	RsbW	ςB	RsbX
BSA46	++	++	++	++	++	++	++	++
XS1	++	++	++	++	+++	+++	+++	−
XS2	++	++	++	++	++	++	++	−
XS4	++	++	++	++	+	+++	+++	−
XS5	++	++	++	−	++	++	++	−
XS6	++	++	+	++	++	++	++	−
XS15	++	++	++	++	+++	+++	+++	−
XS16	++	++	+	++	++	++	++	−
XS33	++	++	−	++	++	++	++	−
XS35	++	−	−	++	++	++	++	−
XS38	++	++	++	+	++	++	++	−

^aIn comparison to protein level in wild-type BSA46. ++, wild-type BSA46 level; +++, increased level; +, reduced level; −, undetectable. 

Genetic characterization of the suppressor mutations.

A series of specialized plasmids was constructed as a tool to map the suppressor mutations. A Km^r gene was placed within a segment of DNA that is homologous with the region upstream of the sigB operon (Materials and Methods). The sigB operon’s P_A promoter and increasingly larger segments of the sigB operon were placed downstream of this Km^r gene. Each plasmid added a sigB operon gene to the genes present on the plasmid that preceded it in the series (Fig. 2). When this collection of plasmids was linearized and introduced into each of the suppressor strains, Km^r transformants arose by homologous recombination before and after the Km^r gene. The promoter-distal gene on the smallest plasmid in the series which gave rise to pinpoint colonies was inferred to be the site of the suppressor mutation. The results of this exercise are presented in Table 4. Based on the reappearance of the RsbX⁻ phenotype, four of the ten suppressor mutations are in rsbU, four are in rsbT, and two are in rsbV.

TABLE 4.

Transformation mapping of suppressor strains

Recipient	Absence or presence of pinpoint colonies with indicated donor DNAa							Geneb	Allelec
	pRR7d	pRS11d	pRT2d	pRU13d	pRV6d	pDT11e	pDU3e
XS1	−	−	−	−	+	+	−	rsbV	rsbV74GD
XS2	−	−	−	+	+	−	+	rsbU	rsbU44PR
XS4	−	−	−	−	+	+	−	rsbV	rsbV85LP
XS5	−	−	−	+	+	−	+	rsbU	rsbU80Yterm
XS6	−	−	+	+	+	0	−	rsbT	rsbT107VG
XS15	−	−	−	+	+	+	+	rsbU	rsbU194VA
XS16	−	−	+	+	+	0	−	rsbT	rsbT15IS
XS33	−	−	+	+	+	0	−	rsbT	rsbT63Qterm
XS35	−	−	+	+	+	0	−	rstT	ΔrsbST
XS38	−	−	−	+	+	−	+	rsbU	rsbU228GR

^a−, absence; +, presence; 0, no transformant colonies. 

^bPredicted site of mutation. 

^cAlterations detected by DNA sequencing. Numbers following the gene designations indicate the amino acid residues at which the changes occurred, followed by the original and new amino acids at these sites (e.g., rsbV74GD is a glycine-to-aspartate change at position 74 of rsbV). Changes which created nonsense codons are designated “term.” ΔrsbST is an rsbS-to-rsbT deletion. 

^dLinearized plasmid. 

^eIntact replicating plasmid for overexpression of rsbT (pDT11) or rsbU (pDU3). 

To verify the sites of the mutations, as well as to test the degree to which the alteration blocked the stress activation pathway, each of the suppressor mutants was transformed with replicating plasmids which expressed either rsbT or rsbU at elevated levels (pDT11 and pDU3, respectively). Free RsbT is normally limiting for the stress-dependent pathway of ς^B activation. Providing additional RsbT from pDT11 leads to a fatal RsbU-dependent ς^B activation in wild-type B. subtilis (28, 37). When pDT11 was introduced into each of the suppressor strains, no transformants were observed from the mutants which were believed to have defects in rsbT (XS6, XS16, XS33, and XS35) (Table 4). This is the expected lethal outcome of high RsbT levels in strains with intact regulators that are downstream of RsbT in the stress pathway (Fig. 1). The three strains with mutations in either rsbV (XS1 and XS4) or rsbU (XS15), which reduced the activity of the stress induction pathway, gave pinpoint transformants after acquiring pDT11. This finding is consistent with suppressor mutations which partially impair the stress pathway at a point downstream of rsbT (i.e., in rsbV or rsbU [Fig. 1]). The three rsbU mutant strains (XS2, XS5, and XS38) (Table 4) that blocked RsbX⁻ activation of ς^B (Fig. 3) formed normal-sized colonies following transformation with pDT11 (Table 4).

As a test of whether the phenotypes of the four strains with apparent mutations in rsbU are due to an explicit loss of RsbU function, we transformed those, as well as the other suppressor strains, with a replicating plasmid (pDU3) which provides an additional source of RsbU. Unlike the case for RsbT, overproduction of RsbU has no effect on ς^B activity in the absence of stress (28). As expected, all four of the putative rsbU mutant strains, but none of the others, formed pinpoint colonies upon transformation with pDU3 (Table 4).

Having localized the mutations to particular genes or regions within the sigB operon, we used PCR techniques to clone and sequence the regions in question. XS35, missing both RsbS and RsbT (Fig. 3, lane 9), had a deletion of rsbST, and the two strains with undetectable RsbU (XS5) and RsbT (XS33) (Fig. 3, lanes 4 and 8, respectively) had chain-terminating mutations in these genes. The remaining seven suppressor mutations were missense changes in the genes implicated by the mapping study. The changes identified are indicated by the allele designations listed in Table 4.

Effect of the RsbX⁻ suppressor mutations on ς^B activity.

Wild-type B. subtilis and RsbX⁻ suppressor strains, each carrying a ctc::lacZ reporter gene, were grown in LB and harvested during exponential growth, at 1 h after entry into stationary phase, and at various times after exposure to 4% ethanol. The ς^B-dependent β-galactosidase activity that was measured in each of these strains is listed in Table 5. Wild-type B. subtilis (BSA46) has little ς^B activity during logarithmic growth; however, its activity rose approximately 10-fold following ethanol stress or the onset of stationary phase (Table 5). All of the strains with suppressor mutations in genes whose products are believed to function only in the stress response pathway (i.e., rsbT and rsbU [Fig. 1]) had similar 10- to 15-fold increases in ς^B activity upon entry into stationary phase but variable inducibility by stress. In addition to the nonsense (rsbU80Yterm and rsbU63Qterm) and deletion (ΔrsbST) mutant strains, two strains with missense mutations in rsbU (rsbU44PR and rsbU228GR) and one with a missense mutation in rsbT (rsbT107VG) were uninducible by stress (Table 5). Apparently, these mutations inactivate the gene products involved. One mutation in rsbU (rsbU194VA) and one in rsbT (rsbT15IS) only partially inhibited the stress pathway, resulting in relatively low ς^B activity during growth in the absence of RsbX but elevated ς^B activity following ethanol treatment (Table 5). ς^B activity during growth of both of the RsbX⁻ strains with mutations in rsbV (rsbV75GD and rsbV85LP) was approximately 10-fold higher than that found in the wild-type strain (Table 5). The ς^B activity in these mutants increased during both stationary phase and ethanol stress. The absolute level of stress-induced ς^B activity was greater in the mutant cells than in wild-type B. subtilis; however, the proportional increase was less.

TABLE 5.

ς^B activity in RsbX⁻ suppressor strains during growth and stress

Strain	β-Galactosidase (ctc::lacZ) activitya (Miller units)
	Stationary phase	Log phaseb	At indicated time (min) after ethanol additionc
			10	20	30	40	50
BSA46 (wild type)	55	4.2	34	51	45	39	32
XS1 (rsbV74GD)	92	60	51	88	90	81	73
XS2 (rsbU44PR)	31	1.7	1.6	1.5	1	0.8	0.8
XS4 (rsbV85LP)	146	40	21	48	57	57	57
XS5 (rsbU80Yterm)	57	2.7	1.8	2	1.4	1.1	1.3
XS6 (rsbT107VG)	50	2.2	1.3	1.7	1.5	1.6	3
XS15 (rsbU194VA)	124	9	14	29	33	34	36
XS16 (rsbT15IS)	55	4.3	3.5	7.8	33	33	38
XS33 (rsbU63Qterm)	26	1.6	1.6	1.5	1.2	1.2	1
XS35 (ΔrsbST)	35	1	1	1	1	1.1	1
XS38 (rsbU228GR)	24	2.2	0.2	0.4	0.4	0.3	0.9

^aβ-Galactosidase activity assays were performed as previously described (13). 

^bβ-Galactosidase activity was measured on cells harvested in the middle of log phase (OD₅₄₀ = 0.250). 

^cEthanol was added to a final concentration 4% to cells growing in LB (OD₅₄₀ = 0.1). 

The induction of ς^B activity in the four mutant strains which were still stress responsive was examined in greater detail (Fig. 4). Consistent with the notion that the suppressor mutations should lessen the activity of the stress pathway, all four of the mutant strains had less than wild-type levels of stress-induced ς^B activity in the presence of a wild-type rsbX allele (Fig. 4A to D versus E). The strains with mutations in genes (rsbU and rsbT) whose products participate only in the stress pathway (Fig. 1) displayed normal stationary-phase induction of ς^B (Fig. 4A and B), while the two rsbV mutants, having lesions in a gene which is needed in both pathways (Fig. 1), had virtually no ς^B activation upon entry into stationary phase (Fig. 4C and D).

FIG. 4.

ς^B Induction in selected suppressor strains. B. subtilis carrying SPβ ctc-lacZ was grown in LB and either exposed to 4% ethanol during exponential growth (time zero) or allowed to enter stationary phase. β-Galactosidase activity is expressed in Miller units (see Materials and Methods). (A) B. subtilis XS16 (rsbT15IS); (B) XS15 (rsbU194VA); (C) XSS1 (rsbV74GD); (D) XSS4 (rsbV85LP); (E) BSA46 (wild type). Open and closed symbols represent unstressed and stressed cultures, respectively; squares depict RsbX⁻ suppressor strains, while triangles illustrate congenic RsbX⁺ strains. Wild-type B. subtilis (BSA46) is represented by circles in panel E. The growth profiles of BSA46 and the mutant strains were similar.

The profile of ς^B activity in the RsbX⁻ variants of these mutants displayed several interesting features. First, ς^B activity in all four of the mutant strains is clearly inducible by ethanol, even though RsbX is lacking (Fig. 4A to D). This finding is consistent with the notion that stress-dependent activation of ς^B does not merely involve the release of negative regulation imposed by RsbX. Next, stress-induced ς^B activity is transient in all of the RsbX⁺ strains (Fig. 4A to E) but persists in the RsbX⁻ strains carrying the rsbT15IS and rsbU194VA alleles (Fig. 4A and B). The drop in stress-induced ς^B activity in the RsbX⁺ strains and its persistence in the rsbT rsbX and rsbU rsbX double-mutant strains support the model that RsbX plays a role in reestablishing prestress ς^B activity levels. Finally, the patterns of uninduced ς^B activity in the RsbX⁻ strains are different, depending on whether the suppressor mutations are in the stress pathway genes (rsbU and rsbT) or in rsbV. In the former instance (Fig. 4A and B), ς^B activity remained relatively constant until the culture entered stationary phase, while in the latter case (Fig. 4C and D), ς^B activity continually increased during growth. The rsbU and rsbT mutations would be expected to reduce the induction potential of the stress pathway, while the rsbV mutations should lessen the consequences of an inappropriately active stress pathway on ς^B activity (Fig. 1). The rsbV mutant profiles indicate that during growth in the absence of RsbX, the stress pathway is biased toward the activation of ς^B. Thus, RsbX not only participates in poststress recovery of ς^B activity levels but also is involved in maintaining low ς^B activity in unstressed cells.

Suppressor mutations in rsbV.

The rsbV alleles were isolated as suppressors of the high ς^B activity that results from elevated RsbU phosphatase activity. Therefore, their products are likely to be either less effective substrates for the RsbU phosphatase or less able to interact with RsbW to cause the release of ς^B, or both (Fig. 1). As seen in Fig. 4C and D, both of the mutant rsbV rsbX strains have high ς^B activity and are therefore at least partially activatable by RsbU. In addition, stationary-phase ς^B induction, a process independent of RsbU (Fig. 1), does not occur in the RsbX⁺ variants of these strains (Fig. 4C and D). These results argue that the principal defect in the mutant RsbV proteins is likely to be impaired ability to block RsbW binding to ς^B. In the case of the rsbV85LP allele, RsbV instability could be sufficient to explain the lower RsbV activity (Fig. 3, lane 3); however, the defect in the RsbV75GD, which accumulates to wild-type levels (Fig. 3, lane 1), is less clear.

We had previously noted that Rsb interactions could be detected in yeast by using translational fusions of rsb genes to the separated Gal4 DNA binding and activator domains (32). Reporter gene activity in this system has been proposed as a measure of the degree to which the fused proteins interact (16). We therefore attempted to use Gal4-dependent promoter activity as a measure of how RsbV74GD associates with RsbW and RsbU. We created a translational fusion between rsbV74GD and the yeast Gal4 activator domain and then examined the ability of each of the chimeric mutant proteins to activate Gal4-dependent transcription in yeast when paired with either rsbU or rsbW fused to the Gal4 DNA binding domain. The rsbV74GD-RsbW and -RsbU interactions were approximately 30 and 2%, respectively, as effective in promoting Gal4-dependent transcription as the same associations with wild-type RsbV (Table 6). This finding argues that the rsbV74GD mutation affects both the RsbV-RsbU and RsbV-RsbW interactions, with the RsbV-RsbU interaction being the most seriously compromised.

TABLE 6.

Rsb interactions in yeast dihybrid system

Mutant allele	Mutant/wild/type interactionsa
	T:R	V:W	T:U	U:V	T:T	V:U	U:T	T:S
rsbV74GD		0.289				0.016
rsbT107VD	0.099		0.047		0.123			0.081
rsbU44PR				0.027			0.165
rsbU194VA				0.021			0.002

^aAssays were performed on two clones from each pairing in duplicate, and the results are presented as a ratio of the mutant value divided by the value obtained in a similar pairing with Gal4 fusions to the wild-type alleles. 

We next examined the phosphorylation state of RsbV74GD following ethanol stress or entry into stationary phase. Extracts were prepared, fractionated by isoelectric focusing (IEF) to separate RsbV from RsbV-P, and analyzed by Western blotting using an anti-RsbV monoclonal antibody as a probe (33). Unphosphorylated RsbV74GD, with a glycine-to-aspartate substitution, should approximate the charge of the phosphorylated form of wild-type RsbV and migrate to this position on the gel matrix. This is seen in Fig. 5. The unphosphorylated species (V*) was very evident in the RsbX⁻ strain (XS1) and still present, although to a lesser extent, in the strain with RsbX (XS12). In addition to unphosphorylated RsbV74GD, the phosphorylated form (V*-P) was also seen in the extracts. The RsbV and RsbV-P profiles in both the wild type (BSA46) and rsbV74GD mutant (XS12) were quite similar. RsbV and Rsb-V* were major RsbV components at 5 to 10 min after ethanol stress but became minor elements, compared to RsbV-P and RsbV*-P, by 30 min poststress (Fig. 5). As expected from the failure of the reporter gene assay to detect ς^B activity in the rsbV74GD strain with an intact rsbX (Fig. 4), the RsbV74GD proteins, expressed from the sigB operon’s ς^B-dependent promoter, are barely visible in extracts of stationary-phase cells (Fig. 5, lanes stat1 and stat2). The IEF data illustrate both the phosphorylation of RsbV74GD and its stress-dependent dephosphorylation. Apparently, even though the rsbV74GD mutation significantly alters the RsbU-RsbV interaction that is detected in the two-hybrid system, this is not enough to block RsbU’s stress-dependent dephosphorylation of RsbV74GD-P. This observation is consistent with the strain’s inducibility by stress and high ς^B activity in the absence of RsbX (Fig. 4C). We suspect that the principal defect of the rsbV74GD mutation is an impaired ability of RsbV74GD to effectively antagonize RsbW-ς^B complex formation, even though there was a significant interaction between RsbV74GD and RsbW in the Gal4 reporter system. RsbV interacts with RsbW, both as a substrate for RsbW’s kinase reaction and as an initiator of ς^B release (Fig. 1). The presence of phosphorylated RsbV74GD in our extracts demonstrates that the interaction of RsbV74GD with RsbW, as kinase, remains intact. This may be the basis of the significant residual RsbW-RsbV interaction that we observed in yeast.

FIG. 5.

IEF analysis of RsbV from wild-type and rsbV74GD mutant B. subtilis. BSA46 (wild type), XS1 (rsbV74GD rsbX::spc), and XS12 (rsbV74GD) were grown in LB and treated with ethanol (4%, final volume). Samples were harvested prior to ethanol treatment (lane 0) and at 5, 10, and 30 min thereafter (lanes 5, 10, and 30). Extracts were prepared, subjected to IEF, transferred to nitrocellulose, and probed with an anti-RsbV monoclonal antibody. Samples of XS12 were also collected at the onset of stationary phase (OD = 0.68; lane stat1) and at an OD of 1.55 (lane stat 2). The positions of the wild-type (V and V-P) and mutant (V* and V*-P) RsbV and RsbV-P on the IEF gel are indicated.

Suppressor mutations in rsbU.

RsbU is a phosphatase which activates RsbV in stressed cells and allows it to drive ς^B release from RsbW (Fig. 1). Four of the RsbX⁻ suppressor mutations mapped to rsbU. One (rsbU80Yterm) is a nonsense mutation, but the other three are missense mutations at different sites along RsbU (Table 4). One of the rsbU missense mutations (rsbU228GR), a glycine-to-arginine substitution at position 228, reduced the ability of the mutant RsbU to accumulate (Fig. 3, lane 10). The remaining two alterations (Table 4), a proline-to-arginine substitution at position 44 and a valine-to-alanine change at position 194, altered the activities of the mutant RsbUs but not their ability to accumulate (Fig. 3, lanes 2 and 6). The rsbU44PR and rsbU228RG mutations totally eliminated stress inducibility of ς^B (Table 3), while the rsbU194VA allele reduced it three- to fourfold (Fig. 4B). When the mutant rsbU alleles (rsbU44PR and rsbU194VA), whose products accumulated to normal levels, were cloned as translational fusions to the yeast Gal4 DNA binding domain and paired with Gal4 activator domain fusions to the proteins (RsbV or RsbT) with which RsbU is known to interact, both mutant alleles exhibited a reduced capacity to activate the yeast reporter system with either fusion (Table 6). rsbU194VA was especially poor in its interaction with RsbT. Apparently, the substitutions in rsbU44PR and rsbU194VA affect both of the known RsbU interactions.

Suppressor mutations in rsbT.

RsbT is an essential activator protein in the stress response pathway (Fig. 1). Four of the suppressor mutations lie in RsbT (Table 4). All four of the mutations altered RsbT’s abundance (Fig. 3). Two of these eliminated the rsbT protein, either by a deletion of the rsbST region (ΔrsbST) or by the introduction of a premature termination codon (rsbT63Qterm). The remaining two mutations (rsbT107VG and rsbT15IS) are missense changes which also significantly reduced RsbT levels (Fig. 3, lanes 5 and 7). The rsbT107VG strain had normal ATP-responsive ς^B activation but no detectable stress induction (Table 5). Yeast dihybrid analysis of rsbT107VG (Table 6) showed a decrease in activation capacity with each of the proteins with which RsbT is known to associate. No interaction appeared to be uniquely compromised by the rsbT107VG alteration. Inasmuch as the Western blot analysis failed to detect RsbT107VG in B. subtilis extracts, we assume that decreased stability, caused by the valine-to-glycine change, is its principal defect.

The rsbT15IS mutant has a curious phenotype. The rsbT15IS rsbX::spc double mutant displayed relatively low ς^B activity during growth, thereby implying that the stress pathway is inactivated by the rsbT15IS mutation. Nevertheless, ς^B activity in this strain was still ethanol inducible (Fig. 4A; Table 5). Apparently, ethanol stress alters the system and adds a feature that is lacking in the unstressed bacteria. The stress induction, although significant, was delayed relative to that seen in either the wild-type strain or any of the other suppressor mutants (Fig. 4). We had previously noted that stress leads to a modest increase in the level of wild-type RsbT (13). Given that RsbT15IS seemed to accumulate less well than the wild-type RsbT, we wondered if stress might be significantly affecting the abundance of RsbT15IS. We therefore examined RsbT and RsbT15IS levels by Western blot analyses of extracts that were prepared from unstressed and ethanol-stressed bacteria (Fig. 6). Stress elevated the levels of both wild-type RsbT (Fig. 6A) and RsbT15IS (Fig. 6B), with the RsbT15IS becoming markedly more abundant in the stressed bacteria than in unstressed cells (Fig. 6B). It is possible that the environment within stressed bacteria or the associations that RsbT undergoes during stress stabilize RsbT. This stabilization may be especially critical to the RsbT15IS and allow it to accumulate to effective levels.

FIG. 6.

Western blot analysis of RsbT in BSA46 and XS16 (rsbX::spc rsbT15IS). Extracts prepared from BSA46 (A) and XS16 (B) cultures that had been harvested approximately 40 min after entry into stationary phase (lanes a) or 110 min after exposure to ethanol (lanes b) were fractionated by SDS-PAGE, transferred to nitrocellulose, and probed with an anti-RsbT monoclonal antibody. The position of RsbT is indicated. The slower-migrating band in lanes a is a stationary-phase B. subtilis protein which cross-reacts with the anti-RsbT antibody (13).

DISCUSSION

The loss of RsbX results in a growth-inhibiting activation of ς^B (6, 10, 18, 20, 29). This is thought to occur by a protein cascade (Fig. 1) which ultimately dephosphorylates and reactivates the anti-ς^B antagonist RsbV (33, 37). RsbT, -S, and -X appear to be key regulators, with RsbU the enzyme that actually catalyzes the RsbV-P dephosphorylation (37). RsbV-P can be dephosphorylated in vitro by RsbU in a reaction that is stimulated by RsbT (37). RsbS, a negative regulator of RsbT, can be phosphorylated in vitro by RsbT and dephosphorylated by RsbX (37). An attractive model proposes that stress allows RsbT to phosphorylate and inactivate its inhibitor RsbS and then, freed from RsbS inhibition, trigger the activation of RsbU (37). RsbX reactivates the RsbS to quench the induction. Thus, stress activation may depend on the relative abundance of phosphorylated RsbS, just as ς^B activity depends on the abundance of unphosphorylated RsbV.

It is formally possible that an increase in the RsbS-P/RsbS ratio could occur by a stress-dependent activation of the RsbT kinase, the inactivation of the RsbX phosphatase, or both. The stress inducibility of several of the RsbX⁻ suppressor mutants (Fig. 4) argues that stress induction of ς^B can occur in the absence of RsbX-dependent processes; however, it remains undetermined whether a stress-dependent effect on RsbX activity contributes to the induction process in wild-type cells.

The sites of the RsbX⁻ suppressor mutations were found in each of the three essential proteins of the stress activation pathway (Fig. 1): RsbV, the effector of ς^B release from the anti-ς^B protein (6, 10, 12); RsbU, the phosphatase that converts RsbV-P into its active form (32, 37); and RsbT, the putative RsbU activator (37). The nonsense and deletion mutations which were isolated in the genes encoding these proteins reinforce the view that RsbT, -U, and -V are positive regulators in the pathway inhibited by RsbX. The missense mutations in the genes illustrate specific changes which affect the abilities of these proteins to accumulate or interact with the other Rsbs.

RsbV is a member of a group of proteins, including SpoIIAA, RsbS, RsbR, and ORF₁₀₈ (an RsbV homolog from Staphylococcus aureus), which are believed to be phosphorylated by regulatory kinases (2, 3, 11, 12, 21, 35, 37). In the case of RsbV, and probably the other members of this group, phosphorylation alters its ability to bind and inactivate a target protein (2, 11, 12, 37). RsbV’s target is the anti-ς^B factor RsbW. In the absence of RsbV, ς^B is not released from RsbW and remains inactive (6, 10, 12). Two suppressor mutations were localized to rsbV. One is a leucine-to-proline substitution at residue 85. This mutation lies in a region rich in amino acids which normally participate in α-helix formation and where hydrophobic amino acids are present in the other members of this group (Fig. 7A). Since this mutant protein was not detected by Western blotting (Fig. 3, lane 3), it is likely that the proline substitution destabilized RsbV85LP’s structure and facilitated its turnover. The second mutation is a glycine-to-aspartate substitution at position 74. This substitution of a relatively large acidic amino acid for a small neutral one does not affect the mutant protein’s ability to accumulate (Fig. 3, lane 1) but does alter RsbV activity (Fig. 4C; Table 5). Small, uncharged amino acids are found at this position in the other B. subtilis RsbV homologs; however, glutamate, a residue similar to the amino acid substituted by the suppressor mutation, lies at this position in the S. aureus RsbV counterpart (Fig. 7A). Apparently, the B. subtilis and S. aureus RsbVs have unique requirements at this site.

FIG. 7.

Alignments of Rsb proteins with homologous proteins at the sites of suppressor mutations. Rsb proteins and their related protein segments are displayed for regions where suppressor mutations were found. Each amino acid (single-letter code) on a black background is identical to the Rsb amino acid at that position; gray shading indicates a similar amino acid. Each number at the left indicates the residue number of the first amino acid depicted for that protein segment. The amino acid changed in the mutant Rsb is in bold type, with the substituted amino acid and its residue number appearing above it. (A) RsbV changes in comparison with the S. aureus RsbV homolog (ORF₁₀₈), the RsbV homolog from the B. subtilis ς^F system (SpoIIAA), and the related Rsb proteins RsbS and -R. (B) RsbU changes, along with the similar regions from the S. aureus RsbU homolog (ORF₃₃₃), RsbX, and the corresponding RsbU-like phosphatase from the ς^F system (SpoIIE). (C) RsbT changes above homologous regions from the E. coli NtrB kinase, RsbW, and the RsbW homolog in the ς^F system (SpoIIAB).

A homolog (ORF₃₃₃) of the RsbU phosphatase has been identified in the sigB operon region of S. aureus (36). In addition, two other B. subtilis regulatory phosphatases, RsbX (21, 37) and SpoIIE (4, 5, 14), have recognizable homologies with RsbU. One of the rsbU missense mutations (rsbU228GR) reduced the ability of the mutant protein to accumulate (Fig. 3, lane 10). The glycine residue that becomes arginine in this mutant is conserved in all three of the rsbU homologs (Fig. 7B) and probably has structural significance. The remaining two rsbU mutations, rsbU44PR and rsbU194PA, altered the activities of the mutant RsbUs (Table 5) but not their ability to accumulate (Fig. 3, lanes 2 and 6). The rsbU44PR mutation occurs in a region of RsbU that is not conserved in the homologous proteins; however, the valine, changed to alanine in rsbU194VA, is conserved in the S. aureus RsbU homolog and occurs at a site where hydrophobic residues are found in the related B. subtilis proteins (Fig. 7B).

The two missense mutations that were isolated in the rsbT kinase/activator gene reduced RsbT levels (Fig. 3, lanes 5 and 7). One, an isoleucine-to-serine change near the RsbT amino terminus (rsbT15IS), lies in a region that is not conserved among the B. subtilis regulatory kinases but is immediately upstream of a region of identity with the E. coli regulatory kinase NtrB (Fig. 7C). The other mutation, a valine-to-glycine change (rsbT107VF) in the carboxy end of the molecule, is within a region where hydrophobic residues are found in the related B. subtilis proteins (Fig. 7C).

The profiles of ς^B activity in strains where the suppressor mutations lie in rsbT or -U differed from the profiles of the rsbV mutants. Consistent with the notion that the stress-induced and ATP-responsive pathways are separable, the strains with suppressor mutations in rsbU or rsbT had relatively normal stationary-phase induction (Fig. 4A and B; Table 5) but altered stress induction. In contrast, the rsbV mutants displayed restricted activity in both pathways (Fig. 4C and D; Table 5). The ongoing increase in ς^B activity during growth in the RsbX⁻ strain with an intact RsbT/RsbU pathway (i.e., suppressor mutation in rsbV) can be interpreted as an indication that RsbX plays a role in maintaining proper ς^B activity levels under conditions where no obvious outside stress (e.g., ethanol or heat shock) is imposed. Therefore, either the stress pathway is biased to hold RsbT in a partially active state at all times or actively growing cultures are constantly receiving stress inputs for ς^B activation. Regardless of which circumstance is true, RsbX is apparently needed to restrict RsbU activity in the absence of obvious stress stimulation.

An intriguing aspect of the ς^B activity profiles in the RsbX⁻ strains with suppressor mutations in rsbV is the transience of their ς^B activities following stress induction. We had previously speculated that RsbX plays a role in reestablishing the steady-state level of ς^B activity following stress (31). We still feel that this is likely to be true; however, the drop in ς^B activity in the rsbV rsbX double-mutant strains (Fig. 4C and D) suggests that there are additional mechanisms that can limit stress induction of ς^B. Perhaps a device exists in the stress induction pathway that adapts the system to a particular level of stress and restricts further ς^B activation. The failure to detect this inhibition in the rsbT rsbX and rsbU rsbX strains (Fig. 4A and B) implies that if such a mechanism exists, it requires on an intact RsbT or -U.

Although much has been learned about ς^B regulation, there are still many unanswered questions. Particularly intriguing is the mechanism by which diverse stresses communicate with the system. Which of the Rsb proteins are the targets for these signals, and what is the mechanism used in this signaling? Ongoing genetic and biochemical studies in several laboratories are likely to soon shed light on this.