Influence of a Functional sigB Operon on the Global Regulators sar and agr in Staphylococcus aureus

Abstract

The growth phase-dependent activity profile of the alternate transcription factor ς^B and its effects on the expression of sar and agr were examined in three different Staphylococcus aureus strains by Northern blot analyses and by the use of reporter gene fusion experiments. Significant ς^B activity was detectable only in the clinical isolates MSSA1112 and Newman, carrying the wild-type rsbU allele, but not in the NCTC8325 derivative BB255, which is defective in rsbU. ς^B activity peaked in the late exponential phase and diminished towards the stationary phase when bacteria were grown in Luria-Bertani medium. Transcriptional analysis and a sarP1-sarP2-sarP3 (sarP1-P2-P3)-driven firefly luciferase (luc+) reporter gene fusion demonstrated a strong ς^B activity- and growth phase-dependent increase in sar expression that was totally absent in either rsbU or ΔrsbUVWsigB mutants. In contrast, expression of the agr locus, as measured by RNAIII levels and by an hldp::luc+ fusion, was found to be higher in the absence of ς^B activity, such as in rsbU or ΔrsbUVWsigB mutants, than in wild-type strains. Overexpression of ς^B in BB255 derivatives resulted in a clear increase in sarP1-P2-P3::luc+ expression as well as a strong decrease in hldp::luc+ expression. The data presented here suggest that ς^B increases sar expression while simultaneously reducing the RNAIII level in a growth phase-dependent manner.

Staphylococcus aureus is a major human pathogen causing a variety of infections, ranging from minor skin and wound infections to life-threatening diseases (42). Pathogenicity in S. aureus is based on a wide range of cell wall-associated and extracellular proteins that are regulated in a coordinate and growth phase-dependent manner. These virulence determinants are controlled among others by the accessory gene regulator agr and the staphylococcal accessory regulator sar (48). Mutations in either agr or sar result in mutants that are strongly attenuated in virulence compared to their corresponding parental strains (1, 8, 15, 31).

The agr locus regulates the expression of cell wall-associated proteins and secreted exoproteins in response to the density of the bacterial population (37). The proposed function of this regulatory system is to enhance the production of wall-associated adhesins, which interact with the host's matrix proteins, and potential defense factors (protein A) in the early stages of infection. This is followed by the expression of excreted invasion factors, such as hemolysins, proteases, and lipases, that are suggested to be involved in the dissemination of the organism from the primary site of infection once the infection has been established (58). The agr locus comprises two divergent transcriptional units, RNAII and RNAIII, which are transcribed from the agrP2 and agrP3 promoters, respectively (Fig. 1B) (reviewed in reference 46). RNAII encodes a four-gene operon, including a two-component signal transduction system that responds to the concentration of a secreted and processed peptide pheromone, which is encoded within the operon itself. The primary function of the RNAII gene products is to activate the agrP2 and agrP3 promoters, significantly aided by SarA. Transcription from the agrP3 promoter results in a 510-nucleotide RNA molecule (RNAIII), which appears to be the effector molecule of the positive and negative regulation of virulence genes that are controlled by the agr locus (36, 49). RNAIII is thought to regulate most target genes at the level of transcription but has also been shown to influence the translation of some genes (45, 49) and contains a small open reading frame coding for delta hemolysin (hld).

FIG. 1.

Genetic organization of the sar and agr loci of S. aureus. Genetic organization of the sar locus (A) and the agr locus (B) of S. aureus and schematic representation of the integration of sarP1-P2-P3::luc+ or hldp::luc+ fusion constructs into the S. aureus chromosome by single crossover. For a description of construction of the plasmids pECsarP1-P2-P3-luc+ and pEChldp-luc+ and integration of the constructs into the S. aureus chromosome, see Materials and Methods. Open reading frames, promoters, and respective transcripts are indicated.

SarA, the major functional protein encoded by the sar locus, is generally believed to be required for the activation of expression of the agr locus (20, 21, 46) and influences the regulation of several virulence factors independently from agr (7, 11, 14, 19, 41, 60). SarA is essentially involved in the capacity of S. aureus to survive inside of polymorphonuclear neutrophils (33), and the ability of S. aureus to enter mammalian cells and induce apoptosis is supposed to be dependent on factors regulated by sar and agr (58). SarA expression itself is controlled by three different, tandemly arranged promoters (Fig. 1A) in a growth phase-dependent manner (3, 23, 43). Although one of these transcripts, sarC, was shown to be controlled by the alternative transcription factor ς^B in vitro (10, 28, 43), the inadvertant use of an rsbU mutant to compare sar expression in a sigB mutant may have wrongly suggested that ς^B was involved neither in the transcriptional control of the sar locus nor in agr expression (12, 14, 17). The strains used in those studies were recently shown to possess almost no ς^B activity due to a mutation in the rsbU gene, which encodes a positive regulator of ς^B (30). The transcription factor ς^B itself, organized in the rsbUVWsigB operon, is supposed to be activated by a cascade encompassing RsbU, an RsbV-specific phosphatase, the anti-anti-sigma factor RsbV, and the anti-sigma factor RsbW.

In this study we demonstrate by the use of transcriptional analyses and reporter gene fusion experiments in three different genetic backgrounds that transcription of both the sar and agr loci are clearly influenced by the ς^B activity in S. aureus strains harboring a functional sigB operon.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. S. aureus was routinely grown in Luria-Bertani (LB) medium at 37°C and at 200 rpm. Antibiotics were used at the following concentrations: for erythromycin and tetracycline, 10 μg ml⁻¹; for ampicillin, 50 μg ml⁻¹.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant genotype or phenotypea	Reference or source
Strains
E. coli
DH10B	F− φ80dlacZΔM15 recA1	Gibco
S. aureus
RN4220	NCTC8325-4, r− m+ (restriction minus, modification plus), rsbU	38
BB255	Essentially the same as NCTC8325, rsbU	4
Newman	ATCC25904; clinical isolate, high level of clumping factor, rsbU+	24
MSSA1112	Clinical isolate, bla rsbU+	25
GP266	RN4220, rsbU+sigB1(Am) Tcr	6
GP268	BB255, rsbU+V+W+sigB+ Tcr	30
IK181	BB255, ΔrsbUVWsigB::erm(B) Emr	40
IK184	Newman, ΔrsbUVWsigB::erm(B) Emr	40
MB32	Newman, asp23+asp23p::pECasp23p-luc+ Emr	30
MB33	BB255, rsbU asp23+asp23p::pECasp23p-luc+ Emr	5
MB39	MSSA1112, ΔrsbUVWsigB::erm(B) Emr	This study
MB49	GP268, asp23+asp23p::pECasp23p-luc+ Tcr Emr	30
MB69	Newman, ΔrsbUVWsigB::erm(B) asp23+asp23p::pECasp23p-luc+ Tcr Emr	30
MB70	MSSA1112, ΔrsbUVWsigB::erm(B)asp23+asp23p::pECasp23p-luc+ Tcr Emr	This study
MB73	MSSA1112, asp23+asp23p::pECasp23p-luc+ Emr	This study
MB90	BB255, ΔrsbUVWsigB::erm(B) asp23+asp23p::pECasp23p-luc+ Tcr Emr	30
MB94	BB255, rsbU+hld+hldP::luc+ Tcr Emr	This study
MB95	BB255, rsbU hld+hldp::pEChldp-luc+ Emr	This study
MB97	Newman, hld+hldp::pEChldp-luc+ Emr	This study
MB98	BB255, rsbU sar+sar-P1-P2-P3::pECsar-P1-P2-P3-luc+ Emr	This study
MB100	Newman, sar+sar-P1-P2-P3::pECsar-P1-P2-P3-luc+ Emr	This study
MB101	Newman, sigB1(Am) sar+sar-P1-P2-P3::pECsar-P1-P2-P3-luc+ Tcr Emr	This study
MB102	BB255, rsbU+sigB1(Am) sar+sar-P1-P2-P3::pECsar-P1-P2-P3-luc+ Tcr Emr	This study
MB103	Newman, sigB1(Am) hld+hldp::pEChldp-luc+ Tcr Emr	This study
MB104	MSSA1112, hld+hldp::pEChldp-luc+ Emr	This study
MB105	MSSA1112, sar+sar-P1-P2-P3::pECsar-P1-P2-P3-luc+ Emr	This study
MB112	MSSA1112, sigB1(Am) hld+hldp::pEChldp-luc+ Tcr Emr	This study
MB113	MSSA1112, sigB1(Am) sar+sar-P1-P2-P3::pECsar-P1-P2-P3-luc+ Tcr Emr	This study
Plasmids
pSP-luc+	Apr, firefly luciferase casette vector	Promega
pEC1	Apr Emr, 1.45-kb ClaI erm(B) fragment of Tn551 in pUC18	9
pTX15	TcrPxyl, staphylococcal origin of replication	50
pIK64	Tcr, pTX15 derivative, containing sigB under the control of the xylose-inducible promoter Pxyl	40
pSPsarP1-P2-P3	Apr, 867-bp PCR fragment of sar promoter from strain RN4220 in pSP-luc+	This study
pSPhldp	Apr, 1-kb PCR fragment of hld promoter from strain RN4220 in pSP-luc+	This study
pECsarP1-P2-P3-luc+	Emr, 2.5-kb KpnI-EcoRI sarP1-P2-P3-luc+ fragment of pSPsarP1-P2-P3 in pECI, S. aureus integration vector that inserts into the sar promoters (sarP1-P2-P3)	This study
pEChldP::luc+	Emr, 2.6-kb KpnI-EcoRI hldp-luc+ fragment of pSPhldP in pEC1, S. aureus integration vector that inserts into the hld promoter (hldp)	This study

^aAbbreviations are as follows: Ap^r, ampicillin resistant; Em^r, erythromycin resistant; Tc^r, tetracycline resistant. 

General methods.

All DNA manipulations, basic molecular methods, and handling of Escherichia coli were performed in accordance with standard protocols (54). Genetic manipulation of S. aureus was done as described earlier (39). The general transducing phage 80α was used for transductions.

Construction of pECsarP1-P2-P3-luc+ and pEChldp-luc+.

A DNA fragment covering 867 bp of the sar promoter region of S. aureus RN4220 was generated by PCR using an upstream primer (5′-CGGTACCGTTGATTTGGGTAGTATGC-3′) including a KpnI linker (underlined) and a downstream primer (5′-TTGCCATGGTTAAAACCTCCC-3′) including a NcoI site (underlined), with italic nucleotides corresponding to positions 5 to 24 and 852 to 872 of the sequence found under GenBank accession no. U46541, respectively. For hldp::luc+, a DNA fragment covering 1 kb of the agr locus of S. aureus RN4220 was generated by PCR using an upstream primer (5′-GTGCCATGGAAATCACTCCTTCC-3′) including a NcoI site (underlined) and a downstream primer (5′-TGGTACCTCAACTTCATCCATTATG-3′) including a KpnI site (underlined), with italic nucleotides corresponding to positions 397 to 419 and 1348 to 1372 of the sequence found under GenBank accession no. AF230358, respectively. The PCR products obtained were digested with KpnI and NcoI and cloned in frame with the 5′ end of the luciferase gene of plasmid pSP-luc+. Sequence analysis and comparison confirmed the identity of the constructs to the RN6390 sequence or RN4220 sequence, respectively. A 2.5-kb KpnI-EcoRI fragment, including the sar promoter region fused to the luciferase coding region, or a 2.6-kb KpnI-EcoRI fragment, including the hld promoter region fused to the luciferase coding region, was subsequently cloned into the suicide plasmid pEC1 (9) to obtain the plasmids pECsarP1-P2-P3-luc+ (Fig. 1A) and pEChldp-luc+ (Fig. 1B), respectively. The plasmids obtained were transformed by electroporation into RN4220 and transduced into different S. aureus genetic backgrounds.

Northern blot analyses.

Isolation of total RNA was done as described by Cheung et al. (16). Eight micrograms of total RNA of each sample was electrophoresed through a 1.5% agarose–0.66 M formaldehyde gel in morpholinepropanesulfonic acid running buffer (20 mM morpholinepropanesulfonic acid, 10 mM sodium acetate, 2 mM EDTA [pH 7]). Blotting of RNA onto a positively charged nylon membrane (Roche, Basel, Switzerland) was performed with a vacuum blotter (Pharmacia, Uppsala, Sweden). The intensities of the 23S and 16S rRNA bands stained with ethidium bromide were verified to be equivalent in all the samples before transfer. Labeling and hybridization were done by the use of the digoxigenin labeling and detection kits according to the manufacturer's instructions (Roche). The following specific primers were used to generate the digoxigenin-labeled DNA probes by PCR labeling: SasarA+, 5′-AGGGAGGTTTTAAACATGGC-3′; SasarA−, 5′-CTCGACTCAATAATGATTCG-3′ (nucleotides 851 to 870 and 1177 to 1196 of the sequence found under GenBank accession no. U46541); RNAIII+, 5′-GTGATGGAAAATAGTTGATGAG-3′; RNAIII−, 5′-GTGAATTTGTTCACTGTGTCG-3′ (nucleotides 453 to 474 and 333 to 353 of the sequence under GenBank accession no. AF230358).

Luciferase assay.

Bacterial cells from overnight cultures containing the appropriate antibiotic were diluted with fresh LB medium to an optical density at 600 nm (OD₆₀₀) of 0.01. Freshly diluted cells were incubated without antibiotics at 37°C and at 200 rpm. S. aureus cells, obtained at different growth stages, were harvested by centrifugation at 11,000 × g during 1 min at room temperature, and the cell pellets were resuspended in phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na₂HPO₄, 1.4 mM KH₂PO₄ [pH 7.3]) to an OD₆₀₀ of 10. Luciferase activity was that determined by rapidly mixing PBS-resuspended cells (10 μl) with an equal volume of Luciferase Assay Substrate (Promega, Madison, Wis.). Luminescence was measured on a Turner Designs TD-20/20 Luminometer (Promega) for a time period of 10 s with a delay of 2 s.

Fibronectin binding assay.

Binding of S. aureus to fibronectin was measured quantitatively in microtiter plates, by a slight modification of a previously described method (44). Briefly, 50 μl of human fibronectin (Sigma, Buchs, Switzerland) was serially diluted twofold from a starting concentration of 500 μg/ml in PBS. Bacterial cells grown to exponential or stationary phase were harvested by centrifugation and washed in PBS, and 20 μl of a suspension (5 × 10⁸ CFU) was added to the fibronectin dilutions. The lowest concentration of fibronectin triggering clumping after overnight incubation at 4°C was recorded as the titer.

RESULTS

ς^B activity in S. aureus.

The ς^B activities of the two genetically distinct strains Newman and MSSA1112 and their respective ΔrsbUVWsigB mutants, as well as those of the rsbU strain BB255 and its derivative GP268, transformed with the rsbU wild-type allele, were analyzed during growth by the use of the asp23 reporter gene system (30). The two clinical isolates MSSA1112 and Newman possessed quite similar ς^B activity profiles in LB medium, with a maximal ς^B activity in late exponential growth phase followed by a significant decrease thereafter (Fig. 2B and C). The ς^B activity profile obtained for strain GP268 (MB49) was comparable to those found for strains MSSA1112 (MB73) and Newman (MB32), while its parental strain produced almost no ς^B activity throughout the whole growth cycle (Fig. 2A). Additionally, all three ΔrsbUVWsigB mutants were unable to produce any ς^B-dependent activity at all. The unexpected strong decrease in ς^B activity that was observed from the onset of stationary phase was confirmed by monitoring the transcription of the ς^B-dependent genes asp23 and sigB in Northern blot analyses. Both transcripts were found to be the most abundant during late exponential growth phase, and their levels were drastically reduced during stationary phase (data not shown).

FIG. 2.

ς^B activity during growth of S. aureus. Expression of asp23::luc+ during growth of S. aureus strains BB255 (A), MSSA1112 (B), Newman (C), and their respective sigB mutants. Strains were grown in LB medium at 37°C. Bacterial growth was measured as the OD₆₀₀ (solid symbols). ς^B transcriptional activity was determined by measuring the luciferase activity of Luc+ (open symbols), the product of the luc+ reporter gene fused to the ς^B-dependent promoters of asp23 (asp23p). (A) Squares, S. aureus strain MB33 (BB255, asp23p::luc+); triangles, strain MB49 (BB255, rsbU+ asp23p::luc+); circles, strain MB90 (BB255, ΔrsbUVWsigB asp23p::luc+). (B) Squares, S. aureus strain MB73 (MSSA1112, asp23p::luc+); circles, strain MB70 (MSSA1112, ΔrsbUVWsigB asp23p::luc+). (C) Squares, S. aureus strain MB32 (Newman, asp23p::luc+); circles, strain MB69 (Newman, ΔrsbUVWsigB asp23p::luc+).

Influence of ς^B on the expression of the sar locus.

Northern blot analyses with the sarA gene as a probe showed strong sarA transcription, originating from the ς^A-dependent sarP1 promoter, during exponential growth (OD₆₀₀, 0.3 to 1.5) and declining with the onset of stationary phase in all strains analyzed (Fig. 3). A similar time course but much weaker transcription was observed in all strains for sarB, originating from the ς^A-dependent sarP2 promoter. In contrast, ς^B-dependent transcription of sarC from the sarP3 promoter was detectable only in the wild-type strains MSSA1112 and Newman, not in BB255 or any of the ΔrsbUVWsigB mutants. sarC-specific transcripts were detectable abundantly from late exponential growth phase up to stationary phase (OD₆₀₀, 1.5 to 5.0). Reporter gene fusion experiments with the luciferase gene luc+ fused to the sarP1, sarP2, and sarP3 (sarP1-P2-P3) promoters suggested an increased SarA production in strains MSSA1112 and Newman with the beginning of late exponential growth phase that paralleled the time course of the overall sar transcripts observed in those strains. No such increase in luciferase activity was detectable with the rsbU mutant BB255 (Fig. 4A) or in any of the ΔrsbUVWsigB mutants. Plasmid pIK64 (P_xyl::sigB)-mediated xylose-induced overexpression of ς^B in the BB255 derivative MB98 resulted in a strong increase in sarP1-P2-P3::luc+ expression (Fig. 5A), which was not observed with the control plasmid pTX15 (Fig. 5B), unambiguously proving ς^B to influence sar expression directly or indirectly.

FIG. 3.

Northern blot analyses of the sar locus. Total RNAs (8 μg/lane) of S. aureus strains BB255 (A), MSSA1112 (B), Newman (C), and their respective sigB mutants were blotted onto positively charged nylon membranes and subjected to Northern blot analyses. RNAs were obtained from cells grown in LB medium at 37°C and harvested at different growth stages (indicated as OD₆₀₀ values [numbers above the lanes]). The blotted membranes were hybridized using a digoxigenin-labeled DNA probe specific for sarA (for details on the construction, see Materials and Methods). The RNA molecular weight marker I (Roche) was used as a size marker. Relevant transcript signals are indicated.

FIG. 4.

Role of ς^B in the regulation of sarP1-P2-P3::luc+ expression during growth. S. aureus derivatives of strains BB255 (A), MSSA 1112 (B), Newman (C), and their respective sigB mutants were grown in LB medium at 37°C. Bacterial growth was measured by OD₆₀₀ (closed symbols). sarP1-P2-P3::luc+ expression was determined by measuring the luciferase activity of the reporter gene luc+ (open symbols). (A) Squares, S. aureus strain MB98 (BB255, sarP1-P2-P3::luc+); circles, strain MB102 (BB255, sigB sarP1-P2-P3::luc+). (B) Squares, S. aureus strain MB105 (MSSA1112, sarP1-P2-P3::luc+); circles, strain MB113 (MSSA1112, sigB sarP1-P2-P3::luc+). (C) Squares, S. aureus strain MB100 (Newman, sarP1-P2-P3::luc+); circles, strain MB101 (Newman, sigB sarP1-P2-P3::luc+).

FIG. 5.

Effect of overexpressed ς^B on the expression of sarP1-P2-P3::luc+. S. aureus derivatives of strain MB98 (BB255, sarP1-P2-P3::luc+), harboring plasmid pIK64 (P_xyl::sigB) (A) or control plasmid pTX15 (P_xyl) (B) were grown in LB medium at 37°C. Growth was measured as the OD₆₀₀ (closed symbols). Growing cultures were split into equal parts at the OD₆₀₀ of 1, and overexpression of ς^B was induced in one of the parts by supplementing 0.5% xylose (diamonds), while the control part was left without addition (squares). The arrow denotes the time point of supplementation. sarP1-P2-P3::luc+ expression (open symbols) was determined by measuring the luciferase activity of the reporter gene luc+ as described in Materials and Methods.

Influence of ς^B on the expression of the agr locus.

Comparison of the growth phase-dependent transcription of RNAIII in Northern blot analyses revealed low transcription levels and a delayed induction of RNAIII expression in the wild-type strains MSSA1112 and Newman compared to results for their corresponding ΔrsbUVWsigB mutants (Fig. 6B and C). The RNAIII expression profile of strain GP268 (BB255 rsbU⁺) paralleled the transcription profiles found for strains MSSA1112 and Newman, with expression being significantly delayed and at a lower level than that for BB255 (Fig. 6A). Reporter gene fusion experiments, using the luciferase gene luc+ fused to the hld gene, carried by RNAIII (Fig. 1B), confirmed these data. The RNAIII-representing luciferase activity profiles of strains carrying the wild-type rsbU allele were significantly lower than those for ΔrsbUVWsigB mutants and BB255, respectively (Fig. 7). Overexpression of ς^B in the BB255 derivative MB95, harboring plasmid pIK64, resulted in a strong decrease in hldp::luc+ expression from that of the control (Fig. 8), corroborating the negative regulatory effect of ς^B on agr expression.

FIG. 6.

Northern blot analyses of RNAIII. Total RNAs (8 μg/lane) of S. aureus strains BB255 and GP268 (BB255, rsbU⁺) (A), MSSA1112 and MB39 (MSSA1112, ΔrsbUVWsigB) (B), and Newman and IK 184 (Newman, ΔrsbUVWsigB) (C) were blotted onto positively charged nylon membranes and subjected to Northern blot analyses. RNAs were obtained from cells grown in LB medium at 37°C and harvested at different growth stages (indicated as OD₆₀₀ values above the lanes). The blotted membranes were hybridized using a digoxigenin-labeled DNA probe specific for RNAIII (for details on construction, see Materials and Methods). The RNA molecular weight marker I (Roche) was used as the size marker. Relevant transcript signals are indicated.

FIG. 7.

Role of ς^B in the regulation of hldP::luc+ expression during growth. S. aureus derivatives of strains BB255 (A), MSSA1112 (B), and Newman (C) were grown in LB medium at 37°C. Bacterial growth was measured by OD₆₀₀ (closed symbols). hldp::luc+ expression was determined by measuring the luciferase activity of the reporter gene luc+ (open symbols) as described in Materials and Methods. (A) Squares, S. aureus strain MB95 (BB255, hldp::luc+); diamonds, strain MB94 (BB255, rsbU⁺ hldp::luc+). (B) Squares, S. aureus strain MB104 (MSSA1112, hldp::luc+); circles, strain MB112 (MSSA1112, sigB hldp::luc+). (C) Squares, S. aureus strain MB97 (Newman, hldp::luc+); circles, strain MB103 (Newman, sigB hldp::luc+).

FIG. 8.

Effect of overexpressed ς^B on the expression of hldP::luc+. S. aureus derivatives of strain MB95 (BB255 hldp::luc+), harboring plasmid pIK64 (P_xyl::sigB) (A) or control plasmid pTX15 (P_xyl) (B), were grown in LB medium at 37°C. Growth was measured as the OD₆₀₀ (closed symbols). Growing cultures were split into equal parts at the OD₆₀₀ of 1, and overexpression of ς^B was induced in one of the parts by supplementing 0.5% xylose (diamonds), while the control part was left without addition (squares). The arrow denotes the time point of supplementation. hldp::luc+ expression (open symbols) was determined by measuring the luciferase activity of the reporter gene luc+ as described in Materials and Methods.

Fibronectin binding activity.

Since sar and agr are known to influence the ability of S. aureus to bind to fibronectin (55, 60), the effects of RsbU on the fibronectin binding capacity were determined with strain BB255 and its derivatives (Table 2). Strain GP268, carrying an intact sigB operon, showed a more than 100-fold-lower fibronectin-clumping titer than its rsbU-defective parent, BB255, irrespective of the growth phase, while no difference was apparent between BB255 and its ΔrsbUVWsigB mutant.

TABLE 2.

Titration of fibronectin clumping of S. aureus

Strain (genotype)	Fibronectin-clumping titera
	Exponential growth phase	Stationary growth phase
GP268 (rsbU+V+W+sigB+)	1	1
BB255 (rsbU)	250	125
BB255 (ΔrsbUVWsigB)	125	125

^aClumping titer, lowest fibronectin concentration (in μg/ml) showing clumping. Results are expressed as the median of results from five to nine independent experiments. 

DISCUSSION

Most of our knowledge of the regulation of ς^B in S. aureus has been adapted from the well-characterized ς^B regulon of the closely related soil bacterium Bacillus subtilis (reviewed in reference 34). In this organism, ς^B has been shown to function as a stress- and stationary phase-specific transcription factor. In B. subtilis, activation of ς^B appears to be basically dependent on the activity of the two RsbV-specific phosphatases, RsbU and RsbP, with the latter being essential for the stationary phase and energy stress-dependent activation of ς^B (56). RsbU was found to be of importance only for the environmental stress activation of this sigma factor (57, 59, 61). The data presented here and elsewhere (6, 30) instead suggest ς^B of S. aureus to be a transcription factor with the main activity in late exponential growth phase rather than in stationary phase. Additionally, the data clearly demonstrated that the natural rsbU mutant BB255 was almost completely unable to express ς^B activity, illustrating the importance of RsbU for the overall activity level of ς^B in S. aureus. Accidentally, nearly all studies of the influence of ς^B on the expression of sar and agr have been carried out with NCTC8325 isogenic backgrounds, harboring the mutation in rsbU (12, 14, 23, 43). In consequence, the findings presented here call into question the ς^B-dependent results obtained from such strains.

Transcriptional data and reporter gene fusions of the sar locus revealed a strong ς^B-dependent transcription of sarC in S. aureus strains harboring an intact sigB operon, while no such transcription was detectable in the rsbU-deficient strain BB255 nor in any of the ΔrsbUVWsigB mutants (Fig. 3 and 4). This results are in contrast to the findings of Bayer et al. (3) and Manna et al. (43), who detected significant amounts of sarC transcripts during late exponential phase in the closely related rsbU-deficient strain RN6390. It is noteworthy that the reporter gene fusion experiments of Manna et al. (43) performed with strain RN6390 revealed only little activity for the sarP3 promoter, the level of which was approximately 50-fold lower than that for the sarP1 promoter. Interestingly, the maximum activity obtained in this study for the sarP3 promoter was comparable to that found for the sarP2 promoter. A comparison of the reporter gene data obtained from Manna et al. (43) with the intensities of the different sar transcripts presented here in Fig. 3 led us to the conclusion that the RN6390 strain used by Manna et al. also possessed almost no ς^B activity. This conclusion is further strengthened by the findings of Cheung et al. (17), who reported neither in vitro nor in vivo activity of the sarP3 promoter in strain RN6390, in accordance with our finding that the luciferase activity profile of BB255 was indistinguishable from that of its ΔrsbUVWsigB mutant. Both findings fit well with our deduction that normal levels of ς^B-dependent transcription of sarC occur only in the presence of a functional RsbU phosphatase. Our data demonstrate that expression of the sar locus is significantly upregulated by ς^B in a growth phase-dependent manner in S. aureus strains harboring an intact sigB operon. Thus, we postulate that ς^B positively contributes to the overall level of SarA in S. aureus. This hypothesis is strengthened by recent findings of Gertz et al. (29), who reported significantly lower SarA levels in a ΔrsbUVWsigB mutant of the rsbU⁺ strain COL.

The influence of SarA on the expression of the agr locus has been the topic of several studies (13, 20, 21, 35, 46, 53). Even though a factor(s) other than SarA (e.g., ORF3, RAP, and RIP) is suggested to participate in controlling agr-related transcription (2, 13, 20, 21), SarA is currently believed to stimulate the expression from both the agrP2 and agrP3 promoters, leading ultimately to the upregulation of RNAIII (21; for a review, see reference 48). The ς^B-dependent upregulation of sar expression observed in rsbU⁺ strains would be expected therefore to result in an increase in agr expression. However, the data presented here revealed a weaker RNAIII transcription for the rsbU⁺ strains MSSA1112 and Newman than for their ΔrsbUVWsigB mutants, suggesting that agr expression is negatively influenced by ς^B activity, irrespective of the positive effect of ς^B on sar expression. This finding is supported by a recent study of Chakrabarti and Misra (10), demonstrating an inhibitory influence of SarA on transcription from both the agrP2 and agrP3 promoters in vitro. The authors suggest that either SarA, together with an as-yet-uncharacterized cellular factor(s), activates transcription of the agr operon, or SarA regulates expression of one or more factors which then activate agr expression. In line with this hypothesis, this as-yet-uncharacterized cellular factor(s) involved in the activation of agr expression may be positively regulated by SarA but dominated negatively by ς^B activity.

Many potential virulence factors have been shown to be regulated by SarA and RNAIII in a cooperative way (11, 13, 18, 27, 52), but the expression of several virulence factors was found to be upregulated by one regulator but repressed by the other (55, 60). Additionally, some virulence factors are influenced by one of the two loci but unaffected by the other one (7, 26, 32). Thus, the ς^B-mediated increase in SarA, accompanied by the decrease in RNAIII, is very likely to enhance these phenomena, resulting in severe growth phase-dependent differences in the expression profiles of some virulence factors. One possible role of ς^B in this scenario is to prolong the production of cell surface proteins, such as fibronectin binding proteins, that are positively influenced by SarA and negatively influenced by agr (55, 60). Simultaneously, ς^B may down-regulate RNAIII-specific activities, i.e., the repression of protein A and upregulation of exoproteins (36, 49) or the production of capsular polysaccharides (22, 51). Consistent with this hypothesis is our finding that the rsbU⁺ derivative GP268 possessed a significantly lower fibronectin-clumping titer than its rsbU-defective parent, BB255, signalling the presence of larger quantities of fibronectin binding proteins in GP268 than in BB255.

The impact of either SarA and/or RNAIII on the expression of virulence factors in S. aureus is well documented. On account of the studies performed with NCTC8325 derivatives, it is unquestionable that sigB mutants are still able to produce sufficient amounts of those two global regulators to be virulent, which led to the conclusion that ς^B has no essential function in the virulence and pathogenicity of S. aureus (12). This conclusion is further strengthened by the findings of Nicholas et al. (47), who observed no differences between the clinical isolate WCUH29 and its isogenic ΔsigB mutant in their ability to cause infections in three distinct animal infection models. However, we agree with Gertz et al. (29), who question whether the infection models analyzed so far really reflect the natural situation in the host. The findings that both agr and sar expression are significantly influenced by ς^B in an rsbU⁺ genetic background should be reason to reevaluate if and how ς^B is involved in the virulence and pathogenicity of S. aureus.