SarA Is a Repressor of hla (α-Hemolysin) Transcription in Staphylococcus aureus: Its Apparent Role as an Activator of hla in the Prototype Strain NCTC 8325 Depends on Reduced Expression of sarS

Abstract

In most Staphylococcus aureus strains, inactivation of sarA increases hla transcription, indicating that sarA is a repressor. However, in S. aureus NCTC 8325 and its derivatives, used for most studies of hla regulation, inactivation of sarA resulted in decreased hla transcription. The disparate phenotype of strain NCTC 8325 seems to be associated with its rsbU mutation, which leads to σ^B deficiency. This has now been verified by the demonstration that sarA repressed hla transcription in an rsbU⁺ derivative of strain 8325-4 (SH1000). That sarA could act as a repressor of hla in an 8325-4 background was confirmed by the observation that inactivation of sarA in an agr sarS rot triple mutant dramatically increased hla transcription to wild-type levels. However, the apparent role of sarA as an activator of hla in 8325-4 was not a result of the rsbU mutation alone, as inactivation of sarA in another rsbU mutant, strain V8, led to increased hla transcription. Northern blot analysis revealed much higher levels of sarS mRNA in strain V8 than in 8325-4, which was likely due to the mutation in the sarS activator, tcaR, in 8325-4, which was not found in strain V8. On the other hand, the relative increase in sarS transcription upon the inactivation of sarA was 15-fold higher in 8325-4 than in strain V8. Because of this, inactivation of sarA in 8325-4 means a net increase in repressor activity, whereas in strain V8, inactivation of sarA means a net decrease in repressor activity and, therefore, enhanced hla transcription.

Staphylococcus aureus is a common human pathogen which colonizes the nares and skin of about one-third of all healthy people. The types of infection caused by this organism range from superficial cutaneous infections to life-threatening bacteremias. The pathogenesis of S. aureus is very complex, and virulence depends on the production of an array of extracellular toxins and enzymes, as well as adhesins, which are regulated by a number of global regulators, e.g., agr (accessory gene regulator) and the sarA (staphylococcal accessory regulator) family of regulators (reviewed in references 1, 9, and 50). In addition, the alternative sigma factor, sigma B (σ^B), seems to be involved in virulence gene expression in addition to its role in regulating stress responses and intermediary metabolism (4, 12, 37, 38). Notably, among the genes controlled by sigma B are the virulence regulators sarA and sarS (4, 20, 67). The activity of σ^B is regulated by the anti-sigma factor RsbW, which binds σ^B, thereby inhibiting its association with the RNA polymerase (3, 46). The ability of RsbW to bind σ^B depends on the phosphorylation status of RsbV, which is determined by the activity of the phosphatase RsbU (22, 68, 70).

Most S. aureus strains produce α-hemolysin (alpha-toxin, hla), which is a 33-kDa pore-forming protein that can lyse a wide range of human cells, including lymphocytes and keratinocytes (21, 31, 69). In addition, α-hemolysin may induce apoptosis in T lymphocytes (31) and has also been shown to be required for biofilm formation (10). Similar to most secreted toxins and enzymes in S. aureus, the expression of α-hemolysin (hla) is positively regulated by agr (29, 47, 58). The effector molecule of the agr-dependent regulation is RNAIII, which is transcribed from the agr P3 promoter at the end of the exponential phase of growth in response to autoactivation of the agr P2 promoter. Autoactivation is achieved via the accumulation of a secreted pheromone, autoinducing peptide, encoded by the agr operon (30, 39, 42). RNAIII appears to increase α-hemolysin expression by three different routes, via a direct interaction with the hla transcript that stimulates translation (48) and via rot- and sarS-dependent pathways. The rot gene locus was first recognized as a suppressor mutation which partially restored α-hemolysin (hla) expression in an agr null mutant (45). Gene array methodology has revealed that rot acts as a global regulator, affecting the transcription of many virulence genes, e.g., hlb, spa, geh, and sspA (59). The central role of rot in S. aureus virulence gene regulation was recently demonstrated in a rabbit endocarditis model of infection (44). Based on the observation that the inactivation of rot had no effect on hla expression in an agr⁺ background and that the transcription of rot was not affected by the agr mutation, it has been suggested that RNAIII acts by sequestering Rot, thereby preventing it from binding to target gene promoters (45, 59). In a recent study, some evidence was presented that RNAIII may even promote the degradation of Rot via a clpXP-dependent route (23). As rot stimulates the transcription of sarS (56, 59), which is a strong repressor of hla (67), the suppressive effect of rot on hla transcription is probably indirect, mediated by sarS. Whether rot is also a direct repressor of hla transcription is not known. The transcription of hla is also positively regulated by saeSR, which encodes a two-component signal transduction system (25). Transcription of the sae locus was clearly reduced in both agr and sarA mutants during the post-exponential phase of growth, indicating that sae is downstream of both RNAIII and sarA in the regulatory pathway (53). However, the level of hla transcription was approximately 10 times higher in the sarA than in the agr mutant (11), indicating that sarA and sae are also epistatic.

In the original S. aureus sarA mutant isolated by Cheung and coworkers (14), the production of α-hemolysin was increased in comparison to that in the parental strain. The same results were obtained when sarA was inactivated in a number of other clinical S. aureus strains (8, 32). However, when the sarA mutation was transferred to the prototype S. aureus strain NCTC 8325, it resulted in reduced α-hemolysin production, indicating that sarA acted as an activator of hla transcription (11, 16, 67). Most further studies of hla regulation in S. aureus have been carried out with strains derived from NCTC 8325. In this strain, the sarA-dependent stimulation of hla transcription seems to be mediated by sarS, which is negatively regulated by sarA and acts as a repressor of hla (67). However, it has been reported that sarA also enhances the transcription of hla via agr. This effect seems to be direct, as sarA has been shown to bind to the agr promoter region (13, 18, 27, 65), but might also be mediated by sarT and sarU (41, 63).

Although the reason for the disparate phenotype of strain NCTC 8325 is not known, it seems to be associated with the σ^B deficiency of this strain and its derivatives (8), which results from a mutation in the rsbU gene (38). In addition, strain 8325-4 has a mutation in the tcaR gene locus that could indirectly affect hla expression, as tcaR stimulates the transcription of sarS (a repressor of hla) (43). Restoration of the σ^B activity in strain NCTC 8325 derivatives dramatically decreased hla expression levels relative to those in the parental strains (24, 28). This was likely sarA independent, as the overall levels of sarA transcription in strains 8325-4 and SH1000 (8325-4 rsbU⁺) were very similar (28). However, other studies, using rsbU⁺ derivatives of 8325-4, revealed increased sarA expression during the post-exponential phase of growth, from the σ^B-dependent promoter (P3) (4, 34). Interestingly, preliminary experiments in our laboratory revealed that the inactivation of sarA in SH1000 (8325-4 rsbU⁺) resulted in increased hemolysin production on rabbit blood agar, suggesting that sarA acts as a repressor of hla in an rsbU⁺ background, contrary to its effect in the parental rsbU-deficient strain, 8325-4. The present study was undertaken to understand why sarA acts as an activator of hla transcription in 8325-4 while being a repressor in the rsbU⁺ derivative (SH1000) and in most clinical strains.

MATERIALS AND METHODS

Bacterial strains, plasmids, and cultivation conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. S. aureus strains were grown on nutrient agar plates (Difco). For screening of protease and α-hemolysin production, nutrient agar plates supplemented with casein or rabbit erythrocytes were used (5, 6). S. aureus strains were precultured in tryptic soy broth (Difco) for 16 to 18 h. When required, 10 μg ml⁻¹ tetracycline, 50 μg ml⁻¹ kanamycin, 5 μg ml⁻¹ erythromycin, or 5 μg ml⁻¹ lincomycin was added to the culture medium. Cells from precultures were centrifuged and inoculated in 100 ml of brain heart infusion (Difco) in 1-liter baffled flasks (culture/flask volume ratio, 1:10) to give an optical density at 600 nm (OD₆₀₀) of 0.5 and incubated on a rotary shaker (180 rpm) at 37°C. For induction of xylA promoter constructs, xylose was added to the culture media to a final concentration of 0.5%. Escherichia coli strains were cultivated on Luria-Bertani (LB; Difco) agar plates or in LB medium, supplemented with 100 μg ml⁻¹ ampicillin when required.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain/plasmid	Relevant characteristic(s)a	Source or reference
Strains
E. coli DH5α	E. coli strain used for propagation of all plasmid constructs	26
S. aureus
8325-4	Prototype S. aureus strain, rsbU	52
DU1090	8325-4 hla::ermB (Emr)	55
KT3001	SH1000 sarA::km (Kmr)	This study
PC1839	8325-4 sarA::km (Kmr)	11
PM466	RN6390, agr null	45
PM466 Δrot::tet	RN6390, agr null, rot::tet (Tcr)	P. J. McNamara
RN4220	Restriction-deficient mutant of 8325-4	36
RN6390	Laboratory S. aureus strain that is derived from 8325-4 rsbU	54
SH1000	8325-4 with functional rsbU	28
V8	High-level-protease-producing strain from which the four major staphylococcal proteases (sspA, sspB, aur, and scp) were originally characterized	ATCC 49775
WA501	8325-4 sarS::ermB (Emr)	This study
WA502	SH1000 sarS::ermB (Emr)	This study
WA504	8325-4 sarA::km sarS::ermB (Kmr Emr)	This study
WA505	8325-4(pIK64) sarA::km sarS::ermB (Kmr Emr Tcr)	This study
WA507	SH1000 sarA::km sarS::ermB (Kmr Emr)	This study
WA519	WA502(pKT210) (Tcr)	This study
WA845	V8 sarA::km (Kmr)	This study
WA885	SH1000 sarS::ermB rot::tet (Emr Tcr)	This study
WA900	SH1000 sarS::ermB sarA::km rot::tet (Emr Kmr Tcr)	This study
WA1029	RN6390, agr null, sarS::ermB rot::tet (Emr Tcr)	57
WA1049	RN6390, agr null, sarS::ermB rot::tet sarA::km (Emr Tcr Kmr)	This study
WA1215	KT3001 hla::ermB (Emr)	This study
WA1217	RN6390, agr null, sarS::ermB (Emr)	57
WA1428	RN6390, agr null, sarS::ermB sarA::km (Emr Kmr)	This study
WA1430	RN6390, agr null, rot::tet sarA::km (Tcr Kmr)	This study
WA1434	8325-4(pKT601) sarS::ermB (Emr Tcr)	This study
Plasmids
pAW16	pKT4 with upstream and downstream fragments of the sarS locus inserted at either side of ermB (Emr)	This study
pAW17	pAW16 with Tcr cassette from pT181 (Tcr Emr)	This study
pGEM-T Easy	E. coli cloning vector for PCR products (Ampr)	Promega
pKT4	pGEM-T Easy with ermB of Tn551 (Ampr Emr)	67
pIK64	S. aureus plasmid with sigB under the control of the xylA promoter (Tcr)	38
pKT210	pSPT245 with xylR and sarS under the control of the xylA promoter (Tcr)	67
pKT601	S. aureus plasmid with sarA under the control of the xylA promoter (Tcr)	66
pSPT245	Shuttle vector (Tcr)	49
pT181	Tetracycline-resistant plasmid from S. aureus (Tcr)	35

^aAmp^r, resistance to ampicillin; Km^r, resistance to kanamycin-neomycin; Em^r, resistance to erythromycin-lincomycin; Tc^r, resistance to tetracycline.

Construction of strains and plasmids.

All plasmids were constructed and cloned in E. coli DH5α, as described previously (60), and subsequently transferred to the restriction-deficient S. aureus strain RN4220 by electroporation (61) before transfer to appropriate S. aureus strains using phage φ11 (51). For construction of the sarS allelic-replacement mutant WA501, PCR fragments (1,032 and 1,057 bp) flanking the sarS gene were synthesized using the primer pairs sarH1-forward (5′-TATATAGCATGCCAAATACGCCCGACAACACTC-3′) and sarH1-reversed (5′-TATATACCGCGGCATATCGACTTCAGGCTTGAC-3′) and spaF (5′-CTGTTAGAGCTCTCAATAATTTAAAAAAGC-3′) and sarH1-back (5′-TATATAGAGCTCGTTATCAGCTTTCGGTGCTTG-3′), respectively. The forward primers contained SphI and SacI restriction sites and the reverse primers contained SacII and SacI restriction sites (underlined sequences). The PCR fragments were inserted in two steps at either side of the ermB cassette in the plasmid pKT4. In the resulting construct (pAW16), a tetracycline resistance cassette from plasmid pT181 was inserted between the ApaI and AatII sites to generate pAW17. This plasmid was then transferred to S. aureus RN4220 by electroporation, and integration in the sarS locus was confirmed by PCR analysis. The sarS::ermB allele was transferred to S. aureus 8325-4 via φ11-mediated transduction in order to obtain the double crossover that replaced the sarS gene with the ermB cassette. Erythromycin-lincomycin-resistant transductants were screened for loss of tetracycline resistance, and the sarS::ermB allele in one selected clone, WA501, was verified by PCR analysis. The hla::ermB allele of strain DU1090 was moved to KT3001, generating the derivative WA1215. The rot::tet allele from strain PM466 Δrot::tet was transferred to WA502, generating WA885. The sarA::km allele from PC1839 was transferred to SH1000, V8, WA885, WA1029, WA1217, and PM466 Δrot::tet, generating the derivatives KT3001, WA845, WA900, WA1049, WA1428, and WA1430, respectively. The sarS::ermB allele from WA501 was transferred to SH1000, PC1839, and KT3001, resulting in the derivatives WA502, WA504, and WA507, respectively. The construct pKT210, containing the sarS gene under the control of the xylA promoter, was transferred to WA502, resulting in the strain WA519. The construct pKT601, containing the sarA gene under the control of the xylA promoter, was transferred to WA501, generating WA1434. The construct pIK64, containing sigB under the control of the xylA promoter, was transferred to strain WA504, generating WA505.

Sequence analysis of the tcaR operon.

The tcaR gene locus of strain V8 was amplified by PCR and sequenced in both directions using the oligonucleotide primers tcaR1 (5′-CAATGATTGCTGAGGACAGCTGATAAAAATAGTAATTAG-3′) and tcaR2 (5′-CAATTCTGTTCCTCTTCAGCTGTTATATGTATATCAC-3′), which were based on the tcaR sequence of strain COL (GenBank accession no. AY008833). DNA sequence analysis was performed using an ABI PRISM BigDye Terminator cycle sequencing kit v. 2 (Applied Biosystems) and an Applied Biosystems 377 DNA sequencer.

Northern blot analysis.

Total S. aureus RNA was prepared using the FAST RNA-blue kit (Bio 101) according to the instructions of the manufacturer. The concentration of RNA was determined by measuring the absorbance at 260 nm. Samples containing 10 μg of total RNA were analyzed by Northern blotting as described previously (47). For Northern hybridization, internal fragments of hla (nucleotides [nt] 315 to 621; GenBank accession number X01645), RNAIII (nt 1047 to 1572; accession number X52543), sarA (nt 843 to 1260; accession number U46541), sarS (nt 70 to 642, locus SACOL0096; accession number CP000046), and 16S rRNA (nt 11 to 1022; accession number X68417) were amplified by PCR, radiolabeled with [α-³²P]dCTP (Amersham) using a random-prime labeling kit (Roche Molecular Biochemicals), and used as probes. Radioactivity was detected using a radioisotope imaging system (PhosphorImager 445SI; Molecular Dynamics), and quantified using the ImageQuant software (Molecular Dynamics). Each experiment was repeated three times.

Analysis of α-hemolysin in culture supernatants.

For Western blot analysis of α-hemolysin, S. aureus culture supernatants were harvested at the post-exponential phase (6 h) and an amount of supernatant corresponding to a bacterial density of 0.12 OD₆₀₀ units (equal to 1 ml) was separated by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride-based membranes. The membranes were incubated with monoclonal mouse anti-α-hemolysin antibodies (64), and bound antibodies were subsequently detected by using horseradish peroxidase-conjugated sheep anti-mouse antibodies (Amersham).

Nucleotide sequence accession number.

Our nucleotide sequence data for the tcaR locus in strain V8 have been deposited in the EMBL database under accession number AM237209.

RESULTS

Whether sarA acts as a repressor or activator of hla in strain 8325-4 depends on the rsbU activity.

Inactivation of sarA in several clinical S. aureus strains resulted in increased α-hemolysin production, indicating that sarA is a repressor of hla (8, 14). However, in strain 8325-4 (RN6390), sarA appeared to act as an activator of hla transcription, as inactivation of sarA resulted in decreased hla expression (11, 16, 67). It has been suggested that the disparate phenotype of 8325-4 could be due to the rsbU mutation of this strain, resulting in σ^B deficiency (8). To analyze this further, we inactivated the sarA gene locus of SH1000 (8325-4 rsbU⁺) by allelic replacement, generating strain KT3001. As indicated by slightly increased hemolysis on rabbit blood agar (Fig. 1A), the expression of hla seemed to be upregulated in KT3001 relative to that in SH1000. Immunoblotting and inactivation of the hla locus in KT3001 confirmed that the increased zone of hemolysis was indeed due to the production of α-hemolysin (Fig. 1B and C). The enhanced α-hemolysin production of KT3001 (SH1000 sarA) was accompanied by elevated (4- to 12-fold) hla mRNA levels in KT3001 relative to the level in SH1000 during the post-exponential phase of growth (Fig. 2), indicating that sarA acts as a repressor of hla transcription in SH1000. Taken together, our findings suggest that the apparent role of sarA as an activator of hla transcription in strain 8325-4 depends on the rsbU activity. The immunoblot analysis (Fig. 1C) also revealed partial degradation of α-hemolysin in KT3001, likely a result of derepressed protease expression in the absence of SarA, consistent with previous observations (7, 40). The 55-kDa protein band in supernatants from SH1000 (Fig. 1C) represents the extracellular fraction of protein A, as indicated by its reactivity with nonspecific immunoglobulin G and by its pronounced upregulation in an agr mutant (data not shown). The lack of protein A in KT3001 was probably due to the increased production of proteases, which is consistent with previous results showing that protein A is degraded by extracellular proteases (33). The reduced protein A production in strain WA502 (SH1000 sarS) is in agreement with sarS being an activator of spa (67).

FIG. 1.

(A) Hemolysin production by strains 8325-4 (rsbU) and SH1000 (rsbU⁺) and their corresponding sarA (PC1839 and KT3001), sarS (WA501 and WA502), and sarA sarS (WA504 and WA507) mutant derivatives, grown on a rabbit blood agar plate. wt, wild type. (B) Hemolysin production by KT3001 (SH1000 sarA) and WA1215 (SH1000 sarA hla), grown on a rabbit blood agar plate. (C) Immunoblot analysis of α-hemolysin in culture supernatants of PC1839 (8325-4 sarA) (lane 1), 8325-4 (lane 2), SH1000 (lanes 3 and 4), KT3001 (SH1000 sarA) (lane 5), and WA502 (SH1000 sarS) (lane 6).

FIG. 2.

Northern blot analysis of sarS, sarA, RNAIII, 16S rRNA, and hla in strain SH1000 (rsbU⁺) and its derivatives WA502 (sarS) and KT3001 (sarA) and in strain 8325-4 (rsbU) and its derivative PC1839 (sarA). The main sarA transcript seen is the sigma B-dependent sarA P3 transcript, which is markedly enhanced in SH1000. The bottom panel (hla) shows a reduced exposure. Samples were taken at the specified time points during the growth of a representative culture. wt, wild type.

Whether sarA acts as a repressor or activator of hla depends on the relative activity of sarS.

To test whether the effect of rsbU on hla regulation by sarA in 8325-4 is a general phenomenon, we analyzed another naturally σ^B-deficient S. aureus strain, V8, which has an insertion element inserted in the rsbU gene (34). In this strain, inactivation of sarA resulted in a four- to fivefold increase in hla transcription during the post-exponential phase of growth (4 h and 6 h) (Fig. 3A), indicating that it is not σ^B activity per se that determines whether sarA acts as a repressor or activator of hla transcription in a given strain but that other factors are also involved. As a result of the σ^B deficiency, strains 8325-4 and V8 had low levels of the σ^B-dependent sarA P3 transcript during the post-exponential phase of growth (34). As sarA-dependent regulation of hla transcription is partly mediated by sarS, which is a repressor of hla (67), the level of sarS mRNA was determined in 8325-4, SH1000, and V8 and in their corresponding sarA mutants (PC1839, KT3001, and WA845). In agreement with previous findings (67), sarS mRNA levels were almost undetectable in 8325-4 (Fig. 2 and 3B). On the other hand, we observed a 5- to 15-fold increase in sarS mRNA levels in SH1000 relative to the level in 8325-4 (Fig, 2). This is consistent with the observation that the transcription of sarS is enhanced by σ^B (4, 67). However, we did not detect the σ^B-dependent sarS transcript (P2) in SH1000 but only increased levels of the σ^A-dependent sarS transcript (P1, the same sarS transcript that was upregulated in the 8325-4 sarA mutant PC1839 [67]) (Fig. 2). This indicates that the upregulation of sarS in SH1000 is not a direct effect of the σ^B activity. As shown in Fig. 3B, the level of sarS transcription was much higher in strain V8 than in 8325-4. A possible explanation would be that the tcaR gene locus, which encodes an activator of sarS transcription, is inactive in 8325-4 (43). DNA sequence analysis (see Materials and Methods) confirmed that the tcaR allele in strain V8 was intact and identical to that of strain COL (GenBank accession no. AY008833). Consistent with previous results (67), the inactivation of sarA in 8325-4 (strain PC1839) resulted in highly elevated (more than 30-fold) sarS transcription during the post-exponential phase of growth (Fig. 2 and 3B). Interestingly, inactivation of sarA in SH1000 or in V8 resulted only in an approximately twofold increase in sarS transcription (Fig. 2 and 3B). Therefore, whether the inactivation of sarA will increase or decrease hla transcription seems to depend on the relative levels of sarA and sarS expression in the parental strain and to what extent sarS is upregulated in a cognate sarA mutant. The reason why sarA appears to be an activator of hla transcription in 8325-4 may thus be that the levels of both sarA and sarS are low (essentially no repression of hla transcription) and that the inactivation of sarA results in a very dramatic increase in sarS transcription and therefore suppression of hla. To confirm that sarA acts as a repressor of hla transcription even in 8325-4, a plasmid carrying sarA behind a xylose-inducible promoter (pKT601) was transferred to a sarS mutant derived from strain 8325-4, resulting in WA1434. As shown in Fig. 4, induction of sarA expression with 0.05% xylose, which did not affect bacterial growth (data not shown) and reduced RNAIII production only very slightly at 2 h and 4 h, completely repressed hla transcription. This is in agreement with the observation that SarA bound to the hla promoter in vitro (19, 67) and repressed hla transcription in an in vitro assay (2), supporting a model in which sarA can also suppress hla transcription in a direct way.

FIG. 3.

(A) Northern blot analysis of hla and RNAIII in strains V8 and WA845 (V8 sarA). (B) Northern blot analysis of levels of sarS and 16S rRNA in strains PC1839 (8325-4 sarA), 8325-4, WA845 (V8 sarA), and V8. Samples were taken at the specified time points during growth.

FIG. 4.

Northern blot analysis of hla, RNAIII, and 16S rRNA in WA1434 (8325-4 sarS, containing pKT601 with xylose-inducible sarA). Samples were taken at the indicated time points during growth in liquid media with or without 0.05% xylose (xyl; final concentration).

High hla transcription can be obtained in an agr mutant if sarA, sarS, and rot are inactivated.

Our present results indicate that hla transcription is suppressed by three regulators, sarA, sarS, and rot. As Rot, which is an activator of sarS transcription (59), is supposed to be neutralized by RNAIII (45), the relative impact of all three repressors was analyzed with an agr mutant. For this, we constructed agr sarS sarA and agr sarA rot triple mutants (WA1428 and WA1430, respectively) and an agr sarS rot sarA quadruple mutant (WA1049). As shown in Fig. 5A, hla mRNA levels were equally low, relative to the level in the parental strain (RN6390), in WA1428 (agr sarS sarA), in WA1430 (agr sarA rot), and in an agr sarS rot triple mutant (WA1029). Interestingly, the quadruple mutant (WA1049) had clearly increased (up to 15-fold) hla mRNA levels relative to those in the triple mutants during the post-exponential phase of growth (4 h and 6 h) (Fig. 5A). These results indicate that sarA, sarS, and rot are equally potent and sufficient to suppress hla transcription. As rot appeared to suppress hla transcription independently of sarS and sarA, this regulatory effect might be direct. It can also be concluded that RNAIII per se was not required for a high level of transcription of hla in the absence of repression by sarA, sarS, and rot, as the levels of hla transcription were very similar in the quadruple mutant (WA1049) and the parental strain, RN6390. However, it must be pointed out that the quadruple mutant produced much less α-hemolysin than the parental strain (Fig. 5B and data not shown), confirming that RNAIII is still essential for the translation of the hla mRNA (48).

FIG. 5.

(A) Northern blot analysis of hla and 16S rRNA in WA1049 (agr sarS rot sarA), WA1029 (agr sarS rot), WA1430 (agr rot sarA), WA1428 (agr sarS sarA), and RN6390 (wild type [wt]), grown in liquid cultures with samples taken at the indicated time points. (B) Hemolysin production by strains, grown on a rabbit blood agar plate, from left to right as follows: row 1, 8325-4 (wt) and WA1217 (agr sarS); row 2, PM466 Δrot::tet (agr rot) and WA1049 (agr sarS rot sarA); and row 3, PM466 (agr) and WA1029 (agr sarS rot).

The σ^B-mediated reduction in hla transcription in strain SH1000 is independent of sarA, sarS, and rot.

Considering that both sarA and sarS can act as repressors of hla transcription, the reason for the low hla mRNA levels in strain SH1000 would be the increased σ^B-dependent expression of these regulators during the post-exponential phase of growth. To further investigate this, the relative impact of sarA and sarS on hla transcription in SH1000 was analyzed by comparing the hla mRNA levels in the sarA and sarS single mutants and the sarA sarS double mutant derived from SH1000 (Fig. 2 and data not shown). As hla transcription levels remained very low in all mutants compared to the level in 8325-4, we concluded that sarA and sarS were not the sole factors responsible for the low hla transcription in SH1000. In agreement with this, induction of σ^B expression in an 8325-4 sarA sarS double mutant containing a xylose-inducible sigB gene (strain WA505) resulted in severe suppression of hla transcription (Fig. 6), confirming that sarA and sarS were not required for the σ^B-dependent repression of hla. Notably, RNAIII production was impaired (three- to fivefold) upon the induction of σ^B in WA505 (Fig. 6), suggesting that the reduction in hla transcription was the result of low RNAIII levels. A similar reduction of the RNAIII level was observed in SH1000 relative to the level in its parental strain (28) (Fig. 2). As RNAIII was not required for the transcription of hla in derivatives of 8325-4 lacking sarA, sarS, and rot (Fig. 5A), the impact of the reduced RNAIII levels in SH1000 on hla transcription was analyzed by the construction of a sarA sarS rot triple mutant of SH1000 (WA900). As the level of hla transcription was hardly upregulated in WA900 relative to the level in SH1000 (Fig. 7), we concluded that the σ^B-mediated reduction in hla transcription in SH1000 was independent of sarA, sarS, and rot and of reduced RNAIII levels. Notably, although the inactivation of sarS in SH1000 derivatives had no marked effect on hla mRNA levels (Fig. 2 and data not shown), these strains (WA502 and WA507) were completely nonhemolytic on rabbit blood agar (Fig. 1A) and lacked detectable α-hemolysin (Fig. 1C and data not shown). This is consistent with the reduced (up to 13-fold) RNAIII levels in these strains (Fig. 2 and data not shown), as RNAIII is required for the translation of the hla mRNA (48). As complementation of WA502 with pKT210, encoding sarS, restored RNAIII expression to parental (SH1000) levels (Fig. 8), we concluded that sarS enhanced RNAIII production in SH1000, in contrast to its effect in 8325-4.

FIG. 6.

Northern blot analysis of RNAIII and hla in WA505 (8325-4 sarA sarS, containing pIK64 with xylose-inducible sigB). Samples were taken at the indicated time points during growth in liquid media with or without 0.05% xylose (xyl; final concentration).

FIG. 7.

Northern blot analysis of hla and 16S rRNA in WA900 (sarS sarA rot triple mutant of SH1000) and 8325-4. Samples were taken at the indicated time points during the growth of representative cultures.

FIG. 8.

Effect of complementation of the sarS mutant derived from SH1000. Northern blot analysis of RNAIII in WA519 (SH1000 sarS with pKT210 carrying sarS), WA502 (SH1000 sarS), and SH1000. Samples were taken at the indicated time points during the growth of representative cultures.

DISCUSSION

The present study was undertaken to find an explanation of why sarA appears to be an activator of hla transcription in the prototype strain NCTC 8325 and its derivatives, while being a repressor of hla in most other S. aureus strains (8, 14). Our results clearly show that the role of sarA as an activator of hla transcription in 8325-4 is associated with the rsbU mutation in this strain, which leads to σ^B deficiency, as sarA acted as a repressor of hla in an rsbU⁺ derivative of 8325-4 (SH1000) (Fig. 1 and 2). However, this was not a direct effect of the σ^B deficiency, as the inactivation of sarA in another σ^B-deficient strain, V8, resulted in increased hla transcription (Fig. 3A). The possibility that strains V8 and 8325-4 respond differently to sarA inactivation because of their different rsbU mutations cannot be excluded. Our results indicate that whether the inactivation of sarA leads to increased or decreased hla transcription depends on the relative activities of sarA and sarS, i.e., their basal levels of expression, and how much sarS transcription is elevated when sarA is inactivated. In the parental strain, 8325-4, the basal post-exponential-phase levels of sarA and sarS mRNAs are low due to the impaired σ^B activity (σ^B enhances sarA and sarS transcription [4, 67]) and the mutation in tcaR (tcaR stimulates sarS transcription [43]). In strain V8, which was found to have an intact tcaR locus, the level of sarS mRNA was much higher than the level in strain 8325-4 (Fig. 3B), while the post-exponential-phase levels of sarA expression were essentially the same in these strains (34). We assume that the levels of sarA in these strains are sufficient to partly repress hla transcription but that the suppressive effect of sarS is significant only in strain V8, which has a lower level of hla mRNA than 8325-4 (our unpublished data). However, inactivation of sarA resulted in an ∼15-fold-higher increase in sarS transcription in 8325-4 than in V8 (Fig. 2 and 3B). A likely explanation of why this led to reduced hla transcription in strain 8325-4 and increased hla transcription in strain V8 is that the loss of repression of hla due to the inactivation of sarA in 8325-4 was counteracted by the massive increase in sarS but that the level of repression of hla in response to the increase in sarS transcription in the V8 sarA mutant (WA845) was less than the derepression of hla due to the loss of sarA activity.

Our present findings strongly suggest that sarA is a repressor of hla transcription in S. aureus strains, including 8325-4 and its derivatives. Repression of hla transcription by sarA seems to be independent of agr (RNAIII), sarS, and rot. As indicated in Fig. 9, the transcription of hla is negatively controlled by three regulators, sarA, sarS, and rot, all of which have to be inactivated or neutralized to achieve full hla transcription. It has been reported that hla transcription is also negatively controlled by sarT (63). However, it is not clear whether this effect is direct or mediated by sarS, which is stimulated by sarT (62). On the other hand, as sarT is strongly repressed by both sarA and agr (63), the possible direct repression of hla by sarT seems not to be of major importance, as very similar hla mRNA levels were obtained in the wild type (sarT repressed by both sarA and agr) and in an agr sarA sarS rot quadruple mutant (Fig. 5A). The positive effect of RNAIII on hla transcription seems to be due mainly to its capability of downregulating sarS transcription via the neutralization of Rot and repression of sarT, both of which are required for sarS transcription (45, 56, 59, 63). This is consistent with the observation that RNAIII was not required for hla transcription in the absence of sarA, sarS, and rot (Fig. 5A). Interestingly, the inactivation of these regulators in strain SH1000 (8325-4 rsbU⁺) did not result in elevated hla transcription (Fig. 7), indicating that another σ^B-dependent factor(s) (indicated by X with an asterisk in Fig. 9) is responsible for the low hla transcription in this strain. The observed requirement for an intact sarS allele for RNAIII production in SH1000 (Fig. 2 and 8) might be explained by sarS acting as a negative regulator of a σ^B-dependent repressor of agr (RNAIII) expression. This and additional σ^B-dependent factors involved in the regulation of hla transcription might be found among the hypothetical proteins identified in a microarray-based analysis of the σ^B regulon (4). Although the much-studied and genetically well characterized strain 8325-4 and its derivatives seem unrepresentative of most clinical S. aureus isolates because of the mutations in rsbU and tcaR, they are still very useful in the exploration of the regulatory networks governing virulence gene expression if compared to other strains.

FIG. 9.

Schematic overview of regulatory interactions involving sarA, sarS, sarT, rot, agr (RNAIII), and σ^B in the control of hla transcription. The arrows indicate stimulation, and the bars indicate repression. The asterisks indicate genes known to be upregulated by σ^B. In the exponential phase of growth, hla transcription is suppressed by sarA, sarS, and rot (45, 67) (Fig. 5). The transcription of sarS is stimulated by rot (56, 59) and sarT (62) and is suppressed by sarA (15, 67). The transcription of sarT is repressed by sarA and agr (RNAIII) (63). RNAIII expression is enhanced by sarA (13, 17, 65). In the post-exponential phase, RNAIII is supposed to neutralize Rot activity (23, 45), whereas the transcription of sarA and sarS is σ^B dependent (4, 34, 67) (Fig. 2). In addition, σ^B seems to suppress hla transcription via mechanisms not involving sarA, sarS, and rot (referred to as X) (Fig. 7). For further explanation, see the text.