Staphylococcus aureus is a major human pathogen that produces many virulence factors in a temporally regulated manner controlled by at least two global virulence regulatory loci (agr and sarA). We identified previously a two-component system, ArlS–ArlR, that modifies the activity of extracellular serine protease and may be involved in virulence regulation. Here, we show that mutations in either arlR or arlS increase the production of secreted proteins [α-toxin (Hla), β-haemolysin, lipase, coagulase, serine protease (Ssp)] and especially protein A (Spa). Furthermore, the pattern of proteins secreted by both mutants was strikingly different from that of the wild-type strain. Transcriptional fusions showed that expression of hla, ssp and spa was higher in both mutants than in the wild-type strain, indicating that the arl operon decreases the production of virulence factors by downregulating the transcription of their genes. The arl mutation did not change spa expression in an agrA mutant or in a sarA mutant, suggesting that both the sarA and the agr loci are required for the action of arl on spa. Northern blot analyses indicated that the arl mutation increased the synthesis of both RNA II and RNA III, but decreased sarA transcription. Finally, arl was not autoregulated, but its expression was stimulated by agr and sarA. These results suggest that the Arl system interacts with both agr and sarA regulatory loci to modulate the virulence regulation network.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Staphylococcus aureus is a major human pathogen that causes a wide spectrum of infections, from superficial lesions to systemic and life-threatening infections, such as osteomyelitis, endocarditis, pneumonia and septicaemia. S. aureus produces a range of virulence factors that contribute to its pathogenicity (Novick, 2000). These factors include secreted proteins, such as serine protease (Ssp), nuclease, haemolysins [α-toxin (Hla), β-haemolysin], enterotoxins, lipase and coagulase, and proteins exposed on the cell surface, such as protein A (Spa) and fibrinogen-, fibronectin- and collagen-binding proteins.
The agr locus, which consists of five genes (agrB, agrD, agrC, agrA and hld), encodes two divergent transcripts (RNA II and RNA III), the synthesis of which is initiated by two different promoters, P2 and P3 respectively. RNA III, which overlaps the hld open reading frame (ORF), is the effector molecule of the agr locus (Novick et al., 1993; Morfeldt et al., 1995). RNA III synthesis is highly dependent on the activation of the agrBDCA genes, which are co-transcribed on RNA II. AgrA and AgrC constitute the response regulator and histidine kinase components, respectively, of a two-component system (Novick et al., 1995). RNA III synthesis is regulated by quorum sensing. Bacteria produce and secrete inducer molecules that trigger RNA III synthesis when they reach a threshold concentration. Known autoinducers of RNA III include the agr-encoded autoinducing peptides (AIP; encoded by agrD and possibly processed by agrB) (Balaban and Novick, 1995; Ji et al., 1995) and the potential RNA III-activating protein (RAP) (Balaban and Novick, 1995) (Fig. 1).
The second global regulatory locus, sar, located within a 1.2 kb fragment, contains the major 372 bp sarA ORF that encodes the 15 kDa protein SarA (Cheung and Projan, 1994). This ORF is transcribed by three overlapping transcripts (sarA, sarC and sarB) (Fig. 1). These transcripts have a common 3′ end, but are initiated from three different promoters, sarA (580 bp) at P1, sarC (840 bp) at P3 and sarB (1150 bp) at P2 (Bayer et al., 1996; Manna et al., 1998) (Fig. 1).
Tegmark et al. (2000) showed that another gene, sarH1, is involved in the regulation of virulence factors in S. aureus. sarH1 encodes a 29 kDa protein, consisting of two homologous halves that display a high degree of similarity to SarA. The interaction between agr, sarA and sarH1 in the co-ordinated regulation of secreted virulence factors and cell wall adhesins of S. aureus is complex. The transcription of genes encoding secreted proteins (α-toxin, serine protease) is activated by RNA III, whereas that of genes encoding cell wall-associated proteins, such as protein A, is repressed by RNA III (Novick, 2000). SarA activates α-toxin gene transcription, but represses transcription of the genes for serine protease and protein A (Cheung and Ying, 1994; Chan and Foster, 1998) (Fig. 1). SarA acts partly through the agr regulatory pathway by binding to agr promoters, stimulating the transcription of agr (Heinrichs et al., 1996; Morfeldt et al., 1996; Chien and Cheung, 1998; Rechtin et al., 1999). However, SarA also affects the expression of certain virulence genes directly, independently of its effect on agr. Thus, SarA is a general transcriptional factor, one of the targets of which is the agr locus (Blevins et al., 1999; Wolz et al., 2000). Finally, SarH1 is responsible for the agr- and sarA-dependent repression of protein A synthesis (Tegmark et al., 2000) (Fig. 1).
The SrrA–SrrB two-component system regulates the production of both exotoxin (toxic shock syndrome toxin 1) and surface-associated (protein A) virulence factors in response to environmental oxygen levels. This regulation is mediated in part by the agr locus. SrrA–SrrB may act in anaerobic repression of staphylococcal virulence factors (Yarwood et al., 2001).
Yet another locus, sae, which also encodes a two-component system, stimulates the production of α-toxin, β-haemolysin and coagulase by a pathway that does not involve agr or sarA (Giraudo et al., 1994; 1996). Thus, the production of individual virulence factors seems to depend on the activities of at least five different regulators that interact to stimulate or repress target gene transcription.
We recently identified a new two-component system, ArlS–ArlR (Fournier and Hooper, 2000) involved in several cell activities. Production of a multidrug resistance efflux pump, NorA, was increased in an arlS transposon insertion mutant (Fournier et al., 2000). The arlS mutant exhibited dramatic autolysis as a result of increased peptidoglycan hydrolase activity. The arlS mutant formed a biofilm on polystyrene surfaces, probably because of altered activity of secreted peptidoglycan hydrolases. Serine protease activity was low in the arlS mutant, suggesting that this two-component system is involved in the regulation of virulence factor production in S. aureus (Fournier and Hooper, 2000).
In this study, we determined the effect of the ArlS–ArlR system on the production of virulence factors. We demonstrate that the arl locus is involved in the regulation of several virulence factors, mainly protein A, and some secreted proteins (α-toxin, β-haemolysin, coagulase, lipase). The arl locus exerts its effects on virulence factors mostly via the agr and/or sarA regulatory pathway.
Construction of an arlR mutant
We studied previously the effects of the arlR–arlS locus on adhesion, autolysis and extracellular proteolytic activity in a mutant containing a Tn917LTV1 insertion in the arlS gene (Fournier and Hooper, 2000; Fig. 2). To study the role of arlR–arlS in the regulation of virulence gene expression, we constructed an arlR mutant in which part of arlR was deleted and replaced by the cat gene (see Experimental procedures). The deletion resulted in the elimination of 126 amino acids (57%) from the predicted ArlR protein (Fig. 2A). This mutation was then transferred to the wild-type strain, RN6390, and to other S. aureus strains (Table 1).
Table 1. Bacterial strains and plasmids used in this study.
the attP site of phage L54a; replicon of pMB1, Apr (E. coli);
temperature-sensitive replicon of pE194, Tcr (S. aureus)
300 bp fragment containing spa promoter
cloned upstream of the lacZ gene of pBF50
280 bp fragment containing ssp promoter
cloned upstream of the lacZ gene of pBF50
270 bp fragment containing hla promoter
cloned upstream of the lacZ gene of pBF50
300 bp fragment containing arlR-arlS promoter
cloned upstream of the lacZ gene of pBF50
To check that arlR was inactivated, total RNA from wild-type (RN6390), and arlR mutant (BF21) strains was analysed by Northern blotting, using two specific probes (one for arlR and one for arlS) (Fig. 3). Both probes hybridized with a 2.7 kb RNA fragment, the expected size of a combined arlR–arlS transcript. This suggested that arlR and arlS are in a bicistronic operon. Neither arlR or arlS transcripts were detected in the arlR mutant, confirming that arlR was deleted and that replacement of arlR by the cat gene, which is transcribed divergently, had a polar effect on the transcription of arlS. Therefore, the arlR deletion mutant was named the arlRS mutant. A second 1.5 kb band that was detected in the wild-type strain, but not in the arlRS mutant, might result from the use of an alternative promoter or terminator site or from processing of the 2.7 kb transcript.
Thus, two different mutants (arlS– and arlRS–) were used to characterize the effects of the arlR–arlS locus on virulence factor gene expression.
Phenotypic characterization of arlS and arlRS mutants
The effect of inactivation of arlR or arlS on different virulence factor proteins (secreted and cell wall associated) is shown in Table 2.
Table 2. Phenotypic characterization of S. aureus strains carrying the atl mutationa.
β-Lactamaseb (µg ml−1)
Extracellular protein (µg ml−1)
DNased (AU ml−1)
Lipased (AU ml−1)
Protein Ad (AU mg−1 protein)
a. The phenotypes were characterized as described in the text. Quantitative results are given as a mean of at least two independent determinations. ND, not determined. As we have previously studied the effect of the arl mutation in 8325 derivatives ( Fournier and Hooper, 2000), we compared two different backgrounds. Strains BF28 and BF29 are derivatives of 8325, which carries the three prophages, whereas strains BF23, BF24 and BF26 are derivatives of 8325-4, which is cured of these three prophages (see text and Table 1).
β-Lactamase activity was determined by MICs of ampicillin.
HU, haemolytic units.
AU, arbitrary units (see Experimental procedures).
+ to 2+ indicates the relative strength of signals on skimmed milk plates (protease). 0, no detectable activity.
36 ± 5
30 ± 3
11 ± 1
0.6 ± 0.2
14 ± 4
98 ± 13
115 ± 5
30 ± 3
1.0 ± 0.4
39 ± 13
149 ± 2
142 ± 29
101 ± 1
1.4 ± 0.5
40 ± 8
68 ± 8
6.0 ± 1.4
4.8 ± 1.2
68 ± 12
173 ± 17
5.5 ± 0.7
0.9 ± 0.3
34 ± 5
The level of concentration of total extracellular proteins was three or four times higher in the arlS and arlRS mutants (BF24, BF26) than in the parent strain (BF23) (Table 2). To avoid autolysis, we used strains carrying the atl mutation. Atl is the main autolysin in S. aureus (Foster, 1995). S. aureus strains carrying both arlS and atl mutations have a very low rate of autolysis, as does the single atl mutant (Fournier and Hooper, 2000). In strains containing a transcriptional fusion between the atl promoter region and lacZ (BF23, BF24 and BF26), β-galactosidase activity was similar in the culture supernatants of the parent strain (BF23) and the two mutants (BF24 and BF26; data not shown), indicating that the increased release of extracellular proteins was not the result of bacterial lysis. Furthermore, the inactivation of arlR and arlS resulted in different patterns of extracellular proteins (Fig. 4A). In both arlS and arlRS mutants, the levels of most proteins were higher than those in the wild type, but the levels of some proteins were lower (particularly those smaller than 31 kDa). The wild-type pattern was restored in the mutants complemented with arlRS (BF25 and BF27).
The activity of several secreted proteins was assessed in the wild-type strain and in the two mutants in two different genetic backgrounds: ISP794, a derivative of strain 8325, which carries three different prophages φ11, φ12 and φ13; and RN6390, a derivative of strain 8325-4, which does not carry these prophages (Table 2). Interestingly, the arlS mutation did not have the same effect in these two backgrounds. In derivatives of strain 8325, the levels of several secreted proteins (DNase, lipase and protease) were significantly lower in the arlS mutant (BF29) than in the parent strain (BF28), as demonstrated previously for protease (Fournier and Hooper, 2000). The level of extracellular β-lactamase, which is not a virulence factor, was also much lower in the arlS mutant (one-fifth of that in the wild type). In contrast, most of the tested secreted virulence factors (α-toxin, β-haemolysin, lipase, coagulase and serine protease) in derivatives of strain 8325-4 were synthesized in reproducibly greater amounts in the arlS and arlRS mutants (BF24 and BF26 respectively) than in the parent strain (BF23), whereas β-lactamase production was unaffected. Thus, in the 8325-4 strain background, only the production of virulence factors was increased by the inactivation of arl genes, whereas in the 8325 strain background, the production of other secreted proteins was also impaired as a result of the presence of either prophages or undefined mutations in 8325. We therefore studied the action of the Arl system in the 8325-4 background.
We also found that the level of production of the cell wall-associated protein A was much higher in both mutants than in the parent strain or in the mutant complemented with arlRS, in which it was very low (Fig. 5, Table 2). Both cell wall-associated (Fig. 5A) and extracellular (Fig. 5B) protein A levels were higher in the mutants.
Expression of virulence factor genes in arlS and arlRS mutants
As inactivation of the arl genes modified the production of virulence factors, we investigated the effect of arl mutations on the expression of virulence factor genes. To determine whether the arl mutations affected transcription, we constructed chromosomal transcriptional–reporter gene fusions with the lacZ gene and the virulence determinant genes spa, hla and ssp. We introduced these fusions into the wild-type and various S. aureus mutant strains and assayed for β-galactosidase activity.
The level of expression of spa was 530 times higher in both arlS and arlRS mutants (BF19 and BF21 respectively) than in the wild-type strain (RN6390) at an OD600 of 1.0 (Fig. 6A). The level of expression of hla and ssp was about 2.5 times higher in the mutants (BF19 and BF21) than in the wild-type strain at an OD600 of 4.0 (Fig. 6B and C). These results confirm the phenotypes documented in Table 2 and indicate that the arl operon decreases the production of virulence factors by downregulating the transcription of their genes.
Interaction of ArlS–ArlR with the regulatory loci agr, sarA and sarH1
The best-characterized regulatory loci of the virulence factors are agr, sarA and sarH1. As the Arl system modifies the transcription of virulence factor genes, we determined the effect of the inactivation of arl genes on the expression of virulence factor genes in agrA, sarA and sarH1 mutant backgrounds. It was not necessary to introduce the atl mutation before measuring the levels of extracellular proteins in agrA or sarA mutants, because these two mutations abolished autolysis in arl mutants (data not shown). Thus, the lysis observed in the arlS mutant (Fournier and Hooper, 2000) requires the agr or sarA loci. β-Galactosidase activity in the culture supernatants of strains ALC488 and BF31, which carried a chromosomal transcriptional ssp:lacZ fusion, and strains RN6112 and BF34, which carried a chromosomal transcriptional spa:lacZ fusion, were not significantly different (data not shown), indicating that the extracellular proteins were not released by autolysis.
The pattern of extracellular proteins was similar in the arl–agrA mutants and the agrA mutant, and in the arl–sarA mutants and the sarA mutant (Fig. 4B and C). This suggests that arl represses target gene expression through agr and/or sarA. Furthermore, serine protease activity was higher in the double agrA–arl mutant than in the agrA mutant. In contrast, the serine protease activity of the sarA–arl mutant was similar to that of the sarA mutant (Table 3). This suggests that the effect of the arl mutation on serine protease production depends on sarA but not agr. Finally, production of cell wall-associated protein A was not or only slightly affected by the introduction of arl mutations in the sarA mutant or in the agrA mutant. In contrast, the addition of arl mutations in the sarH1 mutant increased the production of protein A (Table 3). Transcriptional fusions also showed that the level of expression of spa in the arlS–agrA double mutant (BF34) was three times higher than that in the agrA mutant (RN6112) at an OD600 of 0.1. The level of spa expression in the double arlS–sarA mutant (BF31) was similar to that in the sarA mutant (ALC488) (Fig. 7A). The level of spa expression in the double arlRS–sarH1 mutant (BF40) was 90 times higher than that in the sarH1 mutant (BF38) at an OD600 of 0.1 (Fig. 7B). Thus, the effect of the arl mutations on the transcription of spa observed in a wild-type background (Fig. 6A) was completely abolished by the sarA mutation, was dramatically decreased by the agrA mutation and was unaffected by the sarH1 mutation. This suggests that sarA and, to a lesser extent, agrA are required for the action of arl on spa.
Table 3. Phenotypic characterization of S. aureus strains carrying agrA, sarA and sarH1 mutationsa.
The phenotypes were characterized as described in the text. Quantitative results are given as a mean of at least two independent determinations.
HU, haemolytic units.
AU, arbitrary units (see Experimental procedures).
2+ to 4+ indicates the relative strength of signals on skimmed milk plates (protease). 0, no detectable activity.
5 ± 1
0.8 ± 0.1
1.0 ± 0.3
20 ± 1
0.5 ± 0.1
0.06 ± 0.02
37 ± 1
0.06 ± 0.02
24 ± 9
6.5 ± 0.7
1.2 ± 0.3
13.0 ± 2.9
11 ± 1
47 ± 4
9.0 ± 0.1
1.8 ± 0.2
9.0 ± 0.9
10 ± 1
104 ± 3
16 ± 3
3.2 ± 1.1
2.9 ± 0.6
12 ± 1
ArlR–arlS modifies expression of the agr and sarA loci but not that of the sarH1 loci
As the arlR–arlS locus seems to affect the expression of virulence factor genes by interacting with agr and/or sarA, we analysed the expression of both regulators in the wild-type strain (RN6390), the arlS mutant (BF19) and the mutant complemented with arlRS (BF20). Northern blot analyses using total RNA from late-log phase cells were carried out with probes specific for the different regulators (Fig. 8). RNA II synthesis was higher in the arlS mutant than in the wild-type strain and its mutant complemented with arlRS(Fig. 8A). Similarly, the production of RNA III was slightly higher in the arlS mutant than in the wild-type and the mutant complemented with arlRS(Fig. 8B). However, analysis of sar transcripts revealed that levels of sarB (1150 bp) and sarC (840 bp) transcripts were unaffected, whereas sarA (580 bp) transcript levels were lower in the arlS mutant than in the wild-type strain and its mutant complemented with arlRS(Fig. 8C). Finally, levels of sarH1 transcripts were unaffected by the arlS mutation (Fig. 8D). Thus, the arl locus modifies the synthesis of RNA II and RNA III from the agr locus and of the sarA transcript from the sar locus, suggesting that arl acts on virulence gene expression through agr and sar.
SarA and agr stimulate expression of the arlR–arlS locus
As the arl locus acts on the transcription of sarA and agr, we also studied the effect of sarA and agrA mutations on expression of the arlR–arlS locus. We first determined the transcriptional start site of the arl operon, which enabled us to construct a lacZ transcriptional fusion with a fragment containing the arl promoter region. Primer extension analysis with two different primers (data not shown) revealed that the transcription start site of the arlR–arlS locus was 141 nucleotides (nt) upstream from the predicted translation start site of arlR(Fig. 2B). This predicted start site corresponds to a promoter similar to the SigA-dependent consensus sequence (Deora and Misra, 1996)Q1.
Transcription of the arl locus increased slightly during the exponential phase up to the post-exponential phase, suggesting that activation of the promoter of this operon was growth phase dependent (Fig. 9). The arlR and arlS mutations did not alter the expression of the arl operon, indicating that this two-component system was not autoregulated. However, inactivation of sarA and agrA decreased expression of the arl locus (by a factor of two to six), indicating that both regulators affect transcription of the arl operon. Thus, agr and sarA activate the transcription of arl.
We have shown that the two-component system, ArlS–ArlR, is involved in the regulation of transcription of some virulence genes. The most dramatic effect is the repression of protein A gene expression. This system also decreases the synthesis of several other virulence factors to a lesser extent (α-toxin, β-haemolysin, lipase, serine protease and coagulase). To our knowledge, this is the only regulator of S. aureus that downregulates the expression of all the virulence factors. Interestingly, some of the effects of arl are mediated by sarA and/or agr.
The regulation of protein A (SpA) synthesis is complex and involves at least three different loci (agr, sarA and sarH1). SarA and RNA III are repressors (Cheung et al., 1997), whereas SarH1 activates spa transcription (Tegmark et al., 2000) (Fig. 1). SarA seems to repress spa directly by binding to a specific motif in the spa promoter (Cheung et al., 1997; Chien et al., 1999). However, a recent study (Tegmark et al., 2000) suggested that sarA and agr repress spa transcription by repressing sarH1 transcription. SarH1 seems to activate spa transcription directly by binding to its promoter region. In both arl mutants, spa transcription (Fig. 6A) and the production of extracellular and cell wall-associated protein A (Fig. 5) were dramatically increased. The level of expression of spa was 120 times higher in both arlS and arlRS mutants than in the wild-type strain and 90 times higher in the arlRS–sarH1 mutant than in the sarH1 mutant at an OD600 of 0.1, suggesting that the increase in spa expression as a result of the arl mutations was similar in the sarH1 mutant and in the wild-type strain. Moreover, production of protein A in the double arl–sarH1 mutants is three to 23 times higher than that in the sarH1 mutant (Table 3), and production of protein A in the arl mutants is eight to 25 times higher than that in the wild-type strain (BF23) (Table 2). Differences measured by quantification of Western blots are underestimated because spots of mutants were overexposed as a result of the great amount of protein A. Finally, the arlS mutation does not modify the level of sarH1 transcripts (Fig. 8D). Taken together, these data indicate that sarH1 does not significantly alter the effect of the arl mutations on spa expression and that sarH1 is thus not involved in the action of arl on spa. The level of expression of spa in the arlS–agrA double mutant was three times higher than that in the agrA mutant, whereas the arl mutation did not alter spa expression in the sarA mutant background. Therefore, sarA and, to a lesser extent, agrA are required for the action of arl on spa transcription. The Arl system increases the level of sarA transcript (Fig. 8C). SarA has a slight suppressive effect on spa transcription, independently of sarH1 (Tegmark et al., 2000). Thus, we can speculate that arl acts on spa expression through the activation of sarA.
The regulation of serine protease (Ssp), which represents another class of virulence factors (Novick, 2000), differs from that of protein A. RNA III activates ssp transcription, whereas SarA strongly represses it (Cheung et al., 1992; Chan and Foster, 1998; Lindsay and Foster, 1999) (Fig. 1). The doubling in Ssp levels observed in the atl mutant background (Table 2) was also observed in the agrA mutant background, whereas no obvious increase was observed in the sarA mutant background (Table 3). This suggests that arl requires sarA, but not agr, to alter the expression of ssp. As SarA downregulates ssp, arl probably affects ssp transcription through the activation of sarA(Fig. 1).
Thus, SarA is an important part of the regulatory pathway that involves arl in the expression of virulence factors. The control of the sarA locus is complex. Activator and repressor proteins bind to the sar promoters to modulate the expression of sarA. SarA is likely to be an activator for its own expression. In contrast, activation of SigB probably leads to a downregulation in sarA expression. The 14 kDa SarR also represses the expression of sarA (Manna et al., 2001). The activator (SarA) or downmodulators (SigB and SarR) of sarA expression could be the target(s) of the Arl system.
The arl locus also modifies agr expression. Studies on the regulation of agr transcription revealed that it is autoregulated. Indeed, AgrA and AgrC belong to a two-component system. Mutations in agrA or agrC decrease agr transcription (Novick et al., 1993; 1995; Ji et al., 1995), suggesting that AgrA acts directly on agr promoters. However, the phosphorylated regulator AgrA has not been shown to bind to the agr promoter region. Only SarA has been shown to bind to this region (Heinrichs et al., 1996; Morfeldt et al., 1996). SarA activates the synthesis of both RNA II and RNA III (Bayer et al., 1996; Chien et al., 1998) (Fig. 1). We found that arl decreased transcription of the RNA II operon (Fig. 8A) and, to a lesser extent, that of RNA III (Fig. 8B). arl also increased sarA transcription (Fig. 8C). Thus, arl does not modify agr transcription through sarA. If this were the case, arl would increase agr transcription. Thus, it remains unclear how arl affects agr transcription. As the Arl system has a major effect on exoprotein production, this system may modify the level of the secreted inducer molecules (AIP) required to activate the agr operon.
This study demonstrates the existence of a new virulence regulator in S. aureus. The arl locus modifies the expression of virulence factor genes. The effects of the Arl system on virulence factors depend mainly on the regulatory loci sarA or agr. The synthesis of several other unidentified exoproteins appears to be either stimulated or suppressed by arl(Fig. 4A). This is not surprising, because two-component systems are known to act on several target genes (Stock et al., 1989). The modification of the exoprotein pattern (Fig. 4B and C) and autolysis (data not shown) observed in arl mutants is not observed in the arlS–sarA and arlS–agrA mutants. This suggests that sarA and agr are the key loci for the action of the arl regulon. Thus, identification of the Arl system provides new directions for future research in the virulence regulation of S. aureus.
Bacterial strains and growth conditions
The bacterial strains used in this study are listed in Table 1. Experiments were carried out with an S. aureus agrA mutant (Emr) (RN6112) (Table 1) rather than an agr mutant (Tetr) (RN6911) because the agrA mutant was erythromycin resistant and thus compatible with the resistance marker genes of the other mutations and plasmids used in this study. RNA II and RNA III are not detectable in the agrA::Tn551 strain RN6112 (Morfeldt et al., 1988; Novick et al., 1989); thus, the agrA mutant is similar to the agr mutant.
Escherichia coli strains were used for cloning experiments. They were grown in Luria–Bertani (LB) medium at 37°C. Staphylococci were grown in trypticase soy broth (TSB) and plated on trypticase soy agar (TSA) at 37°C, unless otherwise stated.
DNA was manipulated and CaCl2-competent E. coli cells transformed according to standard procedures (Sambrook et al., 1989). Plasmid DNA was isolated with the Qiagen midiprep kit. S. aureus was transformed with plasmid DNA by electroporation (Fournier and Hooper, 1998). Chromosomal DNA from S. aureus was prepared as described by Stahl and Pattee (1983). Transformation with high-molecular-weight chromosomal DNA was carried out as described previously using phage φ55 (Stahl and Pattee, 1983). The conditions for polymerase chain reaction (PCR) were: 5 min at 94°C, followed by 25 cycles of 94°C for 30 s, 45–51°C for 30 s and 72°C for 30 s to 2 min and, finally, 7 min at 72°C. The promoter regions were amplified with the Expand High Fidelity PCR system (Boehringer Mannheim). For all other PCRs, Taq polymerase (Pharmacia) was used.
Plasmids and plasmid construction
The plasmids used in this work are listed in Table 1. To construct an integrative promoterless transcriptional lacZ fusion vector, the lacZ gene from pHT304-18Z (Agaisse and Lereclus, 1994) was cut with KpnI and BamHI and introduced into pUC18 (New England Biolabs) to give pUHT. pUHT was cut with KpnI, treated with the Klenow fragment and cut with EcoRI to give a 5.7 kb fragment. To remove the BamHI site from pSK950 (Niemeyer et al., 1996), the plasmid was cut with BamHI, the ends were filled in with the Klenow fragment and the plasmid religated. The attP site from this plasmid was cut with EcoRI and SalI to give a 350 bp fragment. To obtain the thermosensitive replicon of pE194 and the tetracycline resistance gene, pLTV1 (Camilli et al., 1990) was cut with XbaI, the ends filled in with the Klenow fragment and the resulting fragment cut with SalI to give a 6 kb fragment. These three fragments were ligated together to give pBF50, which carried the sequences of pUC18 (replicon and ampicillin resistance in E. coli), the attP site of phage L54a, the thermosensitive replicon of pE194, the S. aureus tetracycline resistance gene and three unique sites upstream from the promoterless lacZ gene (HindIII, XbaI and BamHI) (Table 1). pBF50 can specifically integrate into the chromosomal attB site located in the geh gene, which encodes staphylococcal lipase. Integration is facilitated by the presence of pYL112Δ19 (Table 1), which carries the L54a int gene, encoding integrase.
To construct transcriptional fusions of various promoters with the lacZ gene of pBF50, PCR-generated DNA fragments were inserted between the HindIII and XbaI sites of pBF50. PCR was performed with the following primers (restriction enzyme sites are underlined): 5′-AAATTAAAGCTTAGCACATTCA-3′ and 5′-TACCCTCTAGATGTATTTGTAAAGT-3′ (spa promoter); 5′-CCAAACAATTAAGCTTCAAAAGTTA-3′ and 5′-AAACCTCTAGAAAATTTATTTACAA-3′ (ssp promoter); 5′-TTAATCAATAAGCTTAGCTATGTCT-3′ and 5′-CTATTTTCTAGAACGATTTGAGGAA-3′ (hla promoter) and 5′-TTCATTCTGCAGTAGTGAAAAGTCA-3′ and 5′-TACACCTCTAGATACGACTTTTTCTAATAA-3′ (arl promoter). For the arl promoter, the PCR product was first inserted between the PstI and XbaI sites of pUC18 and then subcloned by insertion between the HindIII and XbaI sites of pBF50. The plasmids carrying the spa promoter (pBFSpa), ssp promoter (pBFSsp), hla promoter (pBFHla) and arl promoter (pBFArl) were first introduced, at 30°C into a restriction-deficient strain, S. aureus RN4220 (Table 1), which carried pYL112Δ19 and was selected on tetracycline plates (3 µg ml−1). Transcriptional fusion plasmids were then integrated into the lipase gene by a shift to 42°C. The site-specific integration of plasmids into the S. aureus chromosome was confirmed by loss of lipase activity, resulting from disruption of the lipase gene (Lee et al., 1991). Chromosomal DNA of RN4220 strains carrying the integrated plasmids was then used to transform other S. aureus strains, using phage φ55 and selection on 3 µg ml−1 tetracycline.
Construction of an arlR deletion mutant
As the arlS mutant grows more slowly than the wild-type strain (Fournier and Hooper, 2000), we used a strain in which the arlR–arlS locus was reintroduced into the lipase gene to complement the arlR deletion. The arlR–arlS locus was amplified by PCR and inserted into pSK950 to generate pSKarl (Fournier and Hooper, 2000). The PstI fragment, containing pE194, was deleted to remove the erythromycin resistance gene, resulting in pCLarl. pCLarl was introduced into RN4220 at 37°C and selected on 3 µg ml−1 tetracycline to give strain BF38. Integration into the lipase gene was verified by loss of lipase activity.
A deletion/replacement mutant of arlR was constructed as follows: a 2.1 kb DNA fragment containing the arlR–arlS promoter region and adjacent upstream chromosomal DNA was amplified from ISP794 chromosomal DNA using primers 5′-GCTAAACTGCAGACCTAAAGAGAA-3′ containing a PstI site (underlined) and 5′-TACACCTCTAGATACGACTTTTTCTAATAA-3′ containing a XbaI site (underlined) from preliminary sequence data from the University of Oklahoma Genome sequencing project. The cat gene of pC194 was amplified by PCR using the following primers: 5′-CCTTAGGATCCAGATAAGAAAGAAA-3′ containing the BamHI site (underlined) and 5′-CGGCATTATCTAGAATTATAAAAGCCA-3′ containing the XbaI site (underlined). A 1.7 kb fragment, containing the 3′ terminus of arlR and adjacent downstream chromosomal DNA, including arlS, was obtained from pSKarl by restriction at the HincII site present in the arlR gene and the EcoRI site of the polylinker. These three fragments were cloned successively into pUC18. Finally, pE194 (Shivakumar et al., 1980), which contained a thermosensitive replicon and an erythromycin resistance gene, was inserted into the PstI site. The resulting plasmid was introduced into the derivative of strain RN4220 carrying the arlR–arlS locus in the lipase gene (strain BF38) at 30°C by selection on chloramphenicol (5 µg ml−1) and was integrated by two shifts to 42°C. To promote a second recombination event, a single colony was used to inoculate TSB supplemented with 5 µg ml−1 chloramphenicol and cultured at 30°C. The culture was diluted and plated out on TSA medium to yield isolated colonies. The colonies were then screened for Ems and Cmr. We checked that the arlR gene had been deleted by PCR amplification and Northern blot analysis. The chromosomal DNA of this strain, carrying the arlR deletion, was then used to transform other S. aureus strains, using phage φ55 and selection on 5 µg ml−1 chloramphenicol.
Cells were assayed for β-galactosidase (LacZ) activity using either the colorimetric method with ONPG (Miller, 1972) or the Aurora Gal XE chemiluminescent reporter assay system (ICN). Bacterial cells were grown to different OD600, harvested, washed and resuspended in lysis medium supplemented with 50 µg ml−1 lysostaphin and 15 µg ml−1 DNase. For the colorimetric method, 5 mM dithiothreitol (DTT) was also added. The mixture was incubated for 30 min at 37°C, and the β-galactosidase and protein concentrations of the lysis supernatant were then determined. The chemiluminescent method involved measurement of β-galactosidase activity in a LB 9501 luminometer (Berthold) with a 100 µl automatic injector and a 5 s integration time. β-Galactosidase activity measured by colorimetry is expressed as specific activity [nmol of ONP (o-nitrophenol) min−1].
Secreted protein analysis
pI258 (Table 1) was introduced into strains BF23, BF24, BF28 and BF29 by electroporation and selected on 0.1 µg ml−1 ampicillin. β-Lactamase activity was estimated by determining the minimal inhibitory concentrations (MICs) of TSA supplemented with serial 1:2 dilutions of ampicillin (Fournier and Hooper, 1998).
To determine the amount of extracellular protein in the culture supernatant, cells were grown to an OD650 of 3.5. Supernatants were filtered and stored at −20°C. Protein concentration was determined by the Bradford method (Bio-Rad). Ten millilitres of culture supernatant was concentrated by precipitation with 5% trichloroacetic acid, separated by 10% (w/v) SDS–PAGE and stained with Coomassie brilliant blue.
α-Toxin and β-haemolysin were assayed as described previously (Vandenesch et al., 1991). Briefly, a serial 1:2 dilution of the supernatants was carried out, and each dilution was added to 0.5% whole rabbit or sheep blood, respectively, used as substrates. Haemolytic activity (HU) was determined by measuring residual turbidity at 540 nm. The 50% lysis points were calculated by interpolation. Activities (haemolytic units) are the reciprocal of the dilution that gave 50% lysis. Strain DU1090, which does not produce α-toxin, and strain ISP2094, which does not produce β-haemolysin, were used as controls.
Nuclease (DNase) activity was assayed as described previously (Smeltzer et al., 1993). Briefly, salmon sperm DNA (1 mg ml−1) was mixed with culture supernatants. The mixture was incubated for 30 min at 37°C, and the DNA was then precipitated by adding trichloroacetic acid to a final concentration of 25%. The nuclease activity of the supernatant was determined by measuring the optical density at 260 nm. The control consisted of TSB instead of culture supernatant. Arbitrary units (AU) of DNase activity are defined as OD260(sample)–OD260(TSB), adjusted for the volume of supernatant used.
Lipase activity was determined using tributyrin as a substrate (Smeltzer et al., 1992). The reaction was monitored by measuring the decrease in optical density at 450 nm as a result of the hydrolysis of emulsified tributyrin. Arbitrary units of lipase activity were calculated as the linear slope of a plot of OD450 versus time × 100, adjusted for the volume of supernatant.
Protease activity was detected as clear zones surrounding colonies on nutrient agar plates supplemented with 2% skimmed milk powder.
Coagulase activity was determined by mixing 300 µl of 1:2 serial dilutions of whole-cell cultures with 300 µl of rabbit plasma (Difco). Arbitrary units of coagulase activity are expressed as the reciprocal of the lowest dilution at which coagulation disappeared.
Protein A analysis
Cultures were grown to an OD650 of 1.0 and centrifuged. The cell pellets were used to determine the concentration of cell wall-associated protein A. The culture supernatants were filtered and used to determine the amount of secreted protein A. Cell wall-associated proteins were extracted with lysostaphin in a hypertonic medium (30% raffinose), as described previously (Cheung and Fischetti, 1988). For secreted protein A, filtered supernatants were concentrated 100-fold by centrifugation with a Millipore Centriprep concentrator. Cell wall-associated proteins (1–25 µg) and secreted proteins (0.1–1.5 µg) were resolved on a 10% SDS–polyacrylamide gel, electroblotted onto nitrocellulose Hybond-C Pure and probed with rabbit anti-staphylococcal protein A antibody (Sigma) at a 1:15 000 dilution. Bound antibody was detected with donkey anti-rabbit immunoglobulin G conjugated to peroxidase (Pharmacia) and the ECL Western blotting detection system (Amersham) (1:20 000). The protein A-deficient mutant, DU5723, was used as a control. Quantification of signals from Western blots was performed by densitometric analysis of the autoradiograms using the public domain National Institutes of Health image program (version 1.62). Arbitrary units (AU) correspond to the integrated density measured by the program.
RNA was extracted from late-exponential phase (OD600 of 1.0) or stationary phase (OD600 of 3.0) cultures. Culture (25 ml) was centrifuged, and the cells were disrupted in a FastPrep disintegrator with 500 mg of glass beads, 400 µl of 2% Macaloid, 40 µl of 10% SDS and 500 µl of phenol–choroform–isoamyl alcohol (Derréet al., 1999). The RNA was precipitated with ethanol, collected by centrifugation and resuspended in water. The concentration of RNA was determined by measuring absorbance at 260 nm.
Northern blot analysis
Samples containing 3–25 µg of total RNA from late-exponential phase cells were analysed by Northern blotting. Samples were denatured, separated on a 1.5% agarose–formaldehyde gel and transferred to Hybond-XL nylon membrane. Internal fragments of the genes corresponding to hu (5′-CAGATTTAATCAATGCAGTTGCAGA-3′ and 5′-TAATGCTTTACCAGCTTTGAATGCT-3′), RNA III (5′-CAGAGATGTGATGGAAAATAGTTGA-3′ and 5′-ATTAAGGGAATGTTTTACAGTTATT-3′), agrA (5′-CAAAGAGAAAACATGGTTACCATTA-3′ and 5′-CGATGCATAGCAGTGTTCTTTATTT-3′), sarA (5′-ATGATTGCTTTGAGTTGTTATCAAT-3′ and 5′-ACTCAATAATGATTCGATTTTTTTA-3′), sarH1 (5′-ATAGTGTTTGATAATGTCATTTATTCA-3′ and 5′-TGTAAATGATCTTTATCTGCTAAT A-3′), arlR (nt 354–807; Fournier and Hooper, 2000) and arlS (nt 1263–2326; Fournier and Hooper, 2000) were amplified by PCR, radiolabelled with [α-32P]-dCTP using a random-primed labelling kit (Boehringer Mannheim) and used as probes. These probes are indicated in Figs 1 and 2. The filters were hybridized and washed in stringent conditions, as described previously (Sambrook et al., 1989).
The hu transcript was used as an internal control for the amount of RNA (Chien et al., 1999). Its sequence (see primers above) was obtained on the basis of similarity between HU protein from S. aureus and the histone-like protein, Hbsu, from Bacillus subtilis (a homologue of the E. coli HU proteins). The sequence was obtained from preliminary sequence data from the University of Oklahoma Genome sequencing project. The B. subtilis HBsu protein is equally synthesized during growth in B. subtilis vegetative cells (Micka et al., 1991).
We would like to thank Staffan Arvidson for providing S. aureus KT200, Ambrose L. Cheung for strains ALC488 and RN6112, Timothy J. Foster for strains DU1090 and DU5723, David C. Hooper for pI258, and John J. Iandolo for strain ISP2094. This work was supported by research funds from the Institut Pasteur, Centre National de la Recherche Scientifique, Université Paris 7, and Fondation pour la Recherche Médicale. B.F. received a fellowship from the Fondation pour la Recherche Médicale.