Redefining the Role of the β-Lactamase Locus in Methicillin-Resistant Staphylococcus aureus: β-Lactamase Regulators Disrupt the MecI-Mediated Strong Repression on mecA and Optimize the Phenotypic Expression of Resistance in Strains with Constitutive mecA Expression

Abstract

In response to β-lactam chemotherapy, Staphylococcus aureus has acquired two resistance determinants: blaZ, coding for β-lactamase, which confers resistance to penicillins only, and mecA, coding for an extra cell wall cross-linking enzyme with reduced affinity for virtually all other β-lactams. The transcriptional control of both resistance determinants is regulated by homologous repressors (BlaI and MecI, respectively) and sensor inducers (BlaR1 and MecR1, respectively). There is a cross-talk between the two regulatory systems, and it has been demonstrated that bla regulators stabilize the mecA acquisition. In a recent study, we have unexpectedly observed that in most MRSA strains, there was no significant change in the resistance phenotype upon the overexpression in trans of a MecI repressor, whereas in those few strains negative for the bla locus, there was a massive decrease of resistance (D. C. Oliveira and H. de Lencastre, PLoS One 6:e23287, 2011). Here, we demonstrate that, contrary to what is currently accepted, the bla regulatory system efficiently disrupts the strong MecI-mediated repression on mecA, enabling the optimal expression of resistance. This effect appears to be due to the formation of MecI::BlaI heterodimers that might bind less efficiently to the mecA promoter and become nonfunctional due to the proteolytic inactivation of the BlaI monomer. In addition, we have also observed that the presence of bla regulators may enhance dramatically the expression of β-lactam resistance in MRSA strains with constitutive mecA expression, compensating for the fitness cost imposed by the large β-lactamase plasmid. These observations point to important unrecognized roles of the bla locus for the expression of the methicillin-resistant S. aureus (MRSA) phenotype.

INTRODUCTION

Methicillin-resistant Staphylococcus aureus (MRSA) is a leading cause of infections in hospitals in many countries and has also become an important community- and livestock-associated pathogen (1–3). MRSA isolates are cross-resistant to virtually all β-lactam antibiotics, one of the most clinically relevant classes of antimicrobial agents. The characteristic MRSA phenotype is due to an extra penicillin-binding protein (PBP2A) coded by the mecA gene (4), plus the frequent presence of the β-lactamase gene coding for penicillin resistance (5, 6), which is present in >95% of strains (7).

The mecA gene is part of a large polymorphic DNA fragment, the staphylococcal cassette chromosome mec element (SCCmec), which has integrated in the chromosome (8, 9). In many MRSA strains, the SCCmec contains, upstream of the mecA gene, the divergent mecR1-mecI regulatory genes coding for a sensor inducer and a repressor of mecA transcription, respectively (10). This genetic organization is similar to that of the β-lactamase (bla) locus, which consists of the structural gene blaZ and the homologous blaR1-blaI regulatory genes, and there is a cross-talk between the regulatory systems, as each one is able to control the transcription of mecA and blaZ (11–13). In fact, both purified MecI and BlaI have been shown to protect the same region of the mecA-blaZ promoter sequences (14–16), although the inducers apparently are not interchangeable (15). Moreover, it has been clearly demonstrated that the presence of the bla locus enhances dramatically the mecA domestication (17, 18), presumably due to the capacity of bla regulators to efficiently control the mecA transcription.

In recent studies, we have demonstrated that the mecA regulatory locus is a three-component system that, in addition to mecR1-mecI, contains the previously unrecognized mecR2 gene, coding for an antirepressor. MecR2 interacts with MecI, disturbing its binding to the mecA promoter and fostering its inactivation by proteolytic cleavage (19). The function of MecR2 as an antirepressor accounts for the unexpected lack of effect on the phenotypic expression of β-lactam resistance upon overexpression in trans of a wild-type copy of MecI observed in most MRSA strains (20). Nevertheless, in a few strains negative for the bla locus, there was a massive decline of the resistance phenotype upon the overexpression of MecI in trans. Taking into account the well-recognized role of bla genes in controlling mecA transcription, we hypothesized that, contrary to what is accepted in the current model for the regulation of mecA, BlaI-BlaR1 could interfere directly with the MecI-mediated repression of the mecA gene.

In this study, we demonstrate that the presence of the bla locus disrupts the effect of the MecI overexpression in those MRSA strains negative for the bla locus. Analysis of the expression of wild-type and mutant MecI suggests that this effect is explained by the formation of MecI::BlaI heterodimers, which might bind less efficiently to the mecA promoter and/or become nonfunctional due to the BlaR1-mediated hydrolysis of BlaI. As a consequence, this process also fosters MecI proteolysis, presumably by native cytoplasmic proteases, although it is not essential for the bla-mediated induction of mecA. In addition, we find that in those strains with constitutive mecA expression and low-level resistance to β-lactams, the presence of the bla locus contributes to the optimal expression of the MRSA phenotype, compensating for the fitness cost associated with carrying the β-lactamase plasmid. Together, these observations suggest an unrecognized role for the bla locus as an enhancer of the MRSA phenotype.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The plasmids and bacterial strains used in this study are listed in Tables 1 and 2, respectively. S. aureus strains were routinely grown at 37°C with aeration in tryptic soy broth (TSB; Difco) or on tryptic soy agar plates (TSA; Difco). E. coli strains were grown with aeration at 37°C in Luria-Bertani broth (LB; Difco) or in Luria-Bertani agar (LA; Difco). Recombinant S. aureus strains were selected and maintained with tetracycline at 5 mg/liter and/or chloramphenicol at 20 mg/liter. Recombinant E. coli strains were selected and maintained with ampicillin at 100 mg/liter. Selection for carriage of the β-lactamase plasmid in S. aureus strains was performed with ampicillin at 100 mg/liter. Phenotypic analysis of β-lactam resistance in S. aureus parental and recombinant strains was performed by diffusion disks containing 1 mg of oxacillin (Oxa) (21) and by population analysis profiles (PAPS) at 30°C for 24 to 48 h, as previously described (22, 23).

Table 1.

Plasmids used in this study

Plasmid	Relevant characteristic(s)	Source
pGC-2	E. coli-S. aureus shuttle vector, high copy number, insert expression driven by bacteriophage promoters SP6 and T7, Apr, Cmr	P. Matthews
pSPT181	E. coli-S. aureus shuttle vector, insert expression driven by bacteriophage promoter SP6, Apr, Tcr	40
pGC::mecI	pGC2 with mecI gene from strain N315, overexpression driven by T7 promoter	20
pST::mecI	pSPT181 with mecI gene from strain N315, overexpression driven by SP6 promoter	20
pST::blaR1blaI	pST181 with blaR1-mecI genes from β-lactamase plasmid of strain MW2, expression driven by blaR1 native promoter	This study
pSP::blaR1	pST181 with blaR1 gene from β-lactamase plasmid of strain MW2, expression driven by blaR1 native promoter	This study
pGC::mecIMUT	pGC2 with mecI gene from strain N315 mutated at the N101-F cleavage site	P. Arêde

Table 2.

Parental and recombinant S. aureus strains used in this study

Strain	Relevant characteristiccoevolved	Source/reference(s)
RN4220	Restriction-deficient derivative of reference strain NCTC8325-4	R. Novick
COL	Prototype MRSA strain, homogeneous high-level Oxar, ΔmecR1 (no C-terminal inducer domain, ΨIS1272 insertion), mecI negative, mecR2 negative, β-lactamase negative, clone ST250-I	A. Tomasz (41)
N315	Prototype MRSA strain, heterogeneous Oxar, wild-type mecR1-mecI, mecR2 positive, β-lactamase positive, clone ST5-II	K. Hiramatsu (20, 42)
MW2	Prototype CA-MRSA strain, heterogeneous low-level Oxar, ΔmecR1 (no C-terminal inducer domain, ΨIS1272 insertion), mecI negative, mecR2 negative, β-lactamase positive, clone ST1-IV	43
VNG17	Heterogeneous low-level Oxar, ΔmecR1 (no C-terminal inducer domain, ΨIS1272 insertion), mecI negative, mecR2 negative, β-lactamase negative, clone ST5-IV	20, 44
RJP17	Heterogeneous low-level Oxar, ΔmecR1 (no C-terminal inducer domain, ΨIS1272 insertion), mecI negative, mecR2 negative, β-lactamase negative, clone ST5-IV	20, 44
HT0350	Heterogeneous low-level Oxar, deleted of mecR1-mecI (IS431 insertion), mecR2 negative, β-lactamase negative, clone ST377-V	J. Étienne (45)
HU25	Homogeneous high-level Oxar, wild-type mecR1, truncated mecI, mecR2 positive, β-lactamase positive, clone ST239-III	20, 46
COL + pGC::mecI	COL transformed with pGC2::mecI	20
COL + pGC::mecI + pbla	COL co-transformed with pGC2::mecI and pbla	This study
VNG17 + pGC::mecI	VNG17 transformed with pGC2::mecI	20
VNG17 + pbla	VNG17 transformed with the β-lactamase plasmid from MW2	This study
VNG17 + pGC::mecI + pbla	VNG17 cotransformed with pGC2::mecI and pbla	This study
RJP17 + pGC::mecI	RJP17 transformed with pGC2::mecI	20
RJP17 + pbla	RJP17 transformed with the β-lactamase plasmid from MW2	This study
RJP17+ pGC::mecI + pbla	RJP17 cotransformed with pGC2::mecI and pbla	This study
HT0350 + pST::mecI	HT0350 transformed with pST::mecI, overexpressing MecI	20
HT0350 + pbla	HT0350 transformed with the β-lactamase plasmid from MW2	This study
HT0350 + pST::mecI + pbla	HT0350 cotransformed with pST::mecI and pbla	This study
COL + pGC::mecI + pST::blaR1blaI	COL cotransformed with pGC2::mecI and pSPT181::blaR1blaI	This study
RJP17 + pST::blaR1blaI	RJP17 transformed pSPT181::blaR1blaI	This study
COL + pGC::mecIMUT	COL transformed with pGC2::mecIMUT	This study
COL +pGC::mecIMUT + pST::blaR1blaI	COL cotransformed with pGC2::mecIMUT and pSPT181::blaR1blaI	This study
VNG17 + pGC::mecIMUT	VNG17 transformed with pGC2::mecIMUT	This study
VNG17 +pGC::mecIMUT + pST::blaR1blaI	VNG17 cotransformed with pGC2::mecIMUT and pSPT181::blaR1blaI	This study
RJP17 + pGC::mecIMUT	RJP17 transformed with pGC2::mecIMUT	This study
RJP17 + pGC::mecIMUT + pST::blaR1blaI	RJP17 cotransformed with pGC2::mecIMUT and pSPT181::blaR1blaI	This study
HU25 + pGC::mecIMUT	HU25 transformed with pGC2::mecIMUT	This study
N315::ΔmecR2 mut	N315 mecR2-null mutant intermediate, β-lactamase negative	19
N315::ΔmecR2 + pbla/N315	N315 mecR2 deletion backcross, β-lactamase positive	19
N315::ΔmecR2 + pbla/MW2	N315 mecR2-null mutant β-lactamase negative intermediate transformed with pbla of strain MW2	This study

DNA manipulations.

DNA manipulations were performed by standard methods (24, 25). Total DNA from S. aureus was isolated from bacterial cultures with the Wizard genomic DNA purification kit (Promega) according to the manufacturer's recommendations and using lysostaphin (0.5 mg/ml) and RNase (0.3 mg/ml) in the lysis step. Plasmid DNA was isolated from bacterial cultures with the High Pure plasmid isolation kit (Roche). For plasmid DNA isolation from S. aureus strains, the culture pellets were resuspended in suspension buffer supplemented with 0.1 mg/ml of lysostaphin and incubated at 37°C for 30 to 60 min. Restriction enzymes were used as recommended by the manufacturer (New England BioLabs). Routine PCR was performed with GoTaq Flexi DNA polymerase (Promega). PCR amplification of cloning inserts was performed by high-fidelity PCR (Phusion high-fidelity DNA polymerase; New England BioLabs). DNA purification from PCR was performed with a High Pure PCR product purification kit (Roche). For ligation protocols, the inserts and linearized plasmids were resolved in a low-melting-point agarose gel (1%) (Invitrogen) and DNA bands were purified with a Gene Clean Turbo kit (MP Biomedicals) by following the manufacturer's recommendations. Vector arms and insert ligation was performed with a Rapid DNA ligation kit (Roche) according to the manufacturer's recommendations.

Construction of recombinant strains.

All recombinant plasmids used in this work were first constructed and stabilized in E. coli, electroporated into S. aureus restriction-deficient strain RN4220, and finally transduced by the 80α phage to the desired parental strain, as previously described (26, 27). The integrity of plasmid inserts was confirmed by restriction analysis, PCR, and DNA sequencing, performed by STAB Vida.

To transform the native β-lactamase plasmid (pbla) of strain MW2 into the β-lactamase-negative MRSA strains COL, VNG17, RJP17, and HT0350, a cell suspension of the donor strain was infected with phage 80α and transducing lysate was prepared. Subsequently, cell suspensions of receptor strains were infected with the transducing lysate, and the acquisition of pbla was selected for with ampicillin at 100 mg/liter. The presence of pbla was confirmed by PCR detection of internal fragments of blaZ, blaR1, and bla using the primer sets previously described (28).

To clone the blaR1-blaI genes, a fragment containing both genes and the native blaR1 promoter was amplified by high-fidelity PCR with primers blaR30 (5′ ATAT GGATCC GCA TAA ACA CCA ATA TGA GC 3′; the BamHI recognition sequence is underlined) and blaF30 (5′ ATAT CTGCAG ATT TTC TGT ACA CTC TCA TC 3′; the PstI recognition sequence is underlined), using a pDNA preparation from prototype strain MW2 as the template. After double digestion with BamHI and PstI, the fragment was cloned into pSPT181 divergently from the SP6 promoter, originating the recombinant plasmid pST::blaR1-blaI, in which blaR1-blaI is expressed from its native promoter.

Western blotting.

The proteolysis of MecI under induced conditions (sub-MIC oxacillin at 0.05 mg/liter) was evaluated by Western blotting using a polyclonal antibody raised against purified MecI, as previously described (19).

Competition assays.

In order to evaluate the fitness cost associated with the maintenance of β-lactamase plasmid (pbla), competition assays were performed between the parental strains and recombinant strains transformed with pbla from prototype strain MW2. Briefly, the densities of overnight cultures were adjusted to an optical density at 620 nm (OD₆₂₀) of 1, and cultures were mixed (5 μl plus 5 μl) in a test tube with 5 ml of TSB. Daily, 5 μl of the mixed cultures was inoculated in 5 ml fresh TSB. The assays were conducted for 4 to 10 days, and cultures were exposed to an extended incubation of 60 h to simulate starvation conditions. The assays were performed with and without added oxacillin at a sub-MIC of 0.5 mg/liter. The number of CFU was determined by plating serial dilutions of the cultures onto TSA and TSA supplemented with ampicillin at 100 mg/liter. The numbers of CFU carrying pbla was estimated based on the number of colonies on TSA plates supplemented with ampicillin, whereas the number of CFU without pbla was estimated based on the number of colonies on TSA plates minus the number of colonies on the TSA plates supplemented with ampicillin.

RESULTS AND DISCUSSION

The β-lactamase plasmid (pbla) disrupts MecI-mediated repression and optimizes expression of resistance in strains with constitutive mecA expression.

In previous studies, upon the overexpression in trans of MecI in a representative collection of MRSA strains, we have unexpectedly observed that in most cases there was no significant effect on the phenotypic expression of β-lactam resistance (20). Follow-up studies have demonstrated that, for those strains with functional mecR1-mecI regulatory loci, the observed phenomenon was due to the presence of a previously unrecognized third regulatory gene, mecR2, coding for a potent antirepressor (19). Nevertheless, many of the tested MRSA strains had no functional mecR1-mecI regulatory loci due to the presence of insertion sequences and were negative for the mecR2 gene. However, in three out of four strains negative for the bla locus, we could observe a massive decrease of resistance levels upon overexpression of mecI in trans. Based on these observations and taking into account all of the published evidence demonstrating that the bla locus can control the transcription of mecA (11–13), we hypothesized that the bla locus could also interfere directly with MecI-mediated repression.

In order to explore that hypothesis, we sought to transform the four bla-negative MRSA strains (COL, VNG17, RJP17, and HT0350; Table 2) with the native pbla of prototype community-associated MRSA (CA-MRSA) strain MW2, which has a nonfunctional mecR1-mecI locus and is negative for mecR2; however, being an epidemic community-associated strain, it is presumably well adapted to the expression of β-lactam resistance. We have also cotransformed the four bla-negative MRSA strains with pbla and the recombinant plasmid overexpressing mecI (pGC::mecI). As illustrated in Fig. 1, the presence of pbla completely attenuated the effect of the overexpression of MecI, and the resistance levels of parental strains were restored. In addition, in the three strains with low-level resistance (VNG17, RJP17, and HT0350), the presence of pbla caused a massive boost in the phenotypic expression of oxacillin resistance. Parental strain COL was also transformed with pbla, but due to its high-level resistance, no effect was detected (data not shown).

Fig 1.

Effect of pbla on the phenotypic expression of β-lactam resistance. The native β-lactamase plasmid (pbla) of prototype CA-MRSA strain MW2 was transformed into parental strains COL, VNG17, RJP17, and HT0350 and respective recombinants overexpressing MecI. The oxacillin resistance levels were evaluated by diffusion disks containing 1 mg of oxacillin. WT, wild-type parental strain.

These observations suggested two important unrecognized roles for the bla locus. First, contrary to what is accepted in the literature, it seems that the bla regulatory proteins BlaR1 and BlaI can interfere directly with MecI function. As a matter of fact, the formation of MecI::BlaI heterodimers has been demonstrated (14, 15). Moreover, Llarrull et al. have demonstrated that at cellular levels, BlaI dimers bind less efficiently than BlaI monomers (29), and as such, MecI::BlaI heterodimers may display reduced binding affinity to the mecA promoter and become more susceptible to proteolytic inactivation. Of note, although strains COL, VNG17, and RJP17 have a partially deleted mecR1 gene still with the N-terminal cytoplasmic inducer domain, strain HT0350 has a much more extended deletion without any functional domain; therefore, the role of MecR1 can be excluded from these observations. Second, our observations suggest that in those strains with constitutive mecA expression, the transcriptional control mediated by blaR1-blaI, enhances the expression of resistance, most likely by optimizing the transcription levels of mecA and, hence, the production levels of PBP2a. This hypothesis is supported by the reported lack of correlation between mecA transcript/PBP2a levels and the resistant phenotype (i.e., high expression levels do not necessarily correlate with high-level resistance) (30, 31). In addition, because PBP2a cooperates functionally and colocalizes with the native PBP2, which seems to be part of a large multiprotein complex (32), constitutive mecA expression may originate high levels of PBP2a that disturb that delicate protein complex, interfering with cell wall synthesis and, as a consequence, with the phenotypic expression of β-lactam resistance.

pbla fosters the proteolysis of MecI.

Previously, we demonstrated that the antirepressor MecR2, by interacting with MecI, fosters its proteolytic cleavage, which appears to be essential for the optimal expression of β-lactam resistance (19, 33). The MecI proteolysis occurs independently from MecR1, and as such, it is presumably mediated by native cytoplasmic proteases. Here, we thought to explore the hypothesis that pbla acts in a similar way, i.e., the formation of MecI::BlaI heterodimers destabilizes MecI and fosters its inactivation by proteolysis. For that purpose, we analyzed the proteolysis of MecI by Western blotting in strain RJP17 transformed with pGC::mecI only and cotransformed with pGC::mecI and pbla. As illustrated in Fig. 2, in the presence of pbla the MecI cellular levels decrease dramatically. Therefore, we speculate that in the presence of pbla, the inactivation of MecI by proteolytic cleavage is enhanced. This seems to be a very efficient process, taking into account that in these experiments, MecI was artificially overexpressed in trans from a high-copy-number plasmid with strong bacteriophage promoters, whereas bla genes were expressed from its native promoter and location.

Fig 2.

Effect of pbla on MecI proteolysis. Western blotting of MecI expression under inducing conditions (Oxa at 0.05 mg/liter). Lanes: 1, strain N315 without oxacillin (negative control); 2, N315 with oxacillin; 3, RJP17 + pGC::mecI + pbla; 3, RJP17 + pCG::mecI.

Regulatory genes blaR1-blaI mediate the pbla effects on the MRSA phenotype.

The β-lactamase plasmid is large and, in addition to the bla and maintenance loci, contains other genes coding for identified (e.g., resistance to cadmium) and unidentified functions. In order to confirm that the pbla effects on the MRSA phenotype were indeed mediated by the bla regulators blaR1 and blaI, both genes together with the native blaR1 promoter were cloned into an E. coli-S. aureus shuttle plasmid, pSPT181, originating recombinant plasmid pST::blaR1-blaI. Although pSPT181 carries an SP6 promoter flanking the cloning site, the blaR1-blaI locus was inserted in a divergent direction, so that the cloned genes were expressed from the native blaR1 promoter. The recombinant plasmid was transformed into the recombinant COL strain overexpressing mecI (COL + pGC::mecI) and to parental strain RJP17. As illustrated in Fig. 3, the presence of blaR1-blaI genes was sufficient to disrupt the effects of MecI overexpression on the phenotypic expression of oxacillin resistance on strain COL and to enhance the resistant phenotype of parental strain RJP17. Therefore, the bla regulatory genes account for the effects observed upon the transformation of parental and recombinant strains with the whole native β-lactamase plasmid of strain MW2.

Fig 3.

Effect of blaR1-blaI regulatory genes on the phenotypic expression of β-lactam resistance. The regulatory genes blaR1-blaI from the β-lactamase plasmid of strain MW2, under the control of the blaR1 promoter, were cloned and transformed into the recombinant COL strain overexpressing mecI (COL + mecI) and parental strain RJP17. The oxacillin resistance levels were evaluated by diffusion disks containing 1 mg of oxacillin. WT, wild-type parental strain.

pbla effects on the MRSA phenotype are independent of MecI proteolysis.

Since previous studies suggested that BlaR1 could not act directly on MecI (15), we sought to clarify the role of MecI proteolysis. For this purpose, we evaluated the effect of pbla on a mutant variant of MecI for the cleavage site N¹⁰¹-F (MecI^MUT), which we have recently shown to be resistant to the MecR2 action as an antirepressor (33). Recombinant strains COL + pbla, VNG17 + pbla, and RJP17 + pbla were transformed with a recombinant plasmid overexpressing MecI^MUT (pGC::mecI^MUT). As a control experiment, we have also transformed the prototype strain HU25, which, in spite of being mecR1 and mecR2 positive, expresses a nonfunctional truncated MecI protein due to a conserved nonsense point mutation (20). As illustrated in Fig. 4, even in the presence of a mutant variant of MecI for the cleavage site, the presence of pbla was sufficient for optimal β-lactam resistance levels. This was also true in prototype strain HU25, suggesting that the bla locus of this strain, which is integrated at the chromosome (34, 35), is functionally similar to pbla of MW2. These observations suggest that either the MecI proteolysis is not essential for the induction of mecA by blaR1-blaI or, less likely, that BlaR1 cleaves MecI in a different position. To explore these hypotheses, we evaluated the proteolysis of MecI^MUT by Western blotting. As illustrated in Fig. 5, in sharp contrast to what was observed in Fig. 2, we could not detect any proteolytic degradation of MecI^MUT, which suggests that the proteolysis of MecI is not required for the bla-mediated induction of mecA. Thus, the effect of bla on MecI must rely on the formation of MecI::BlaI heterodimers which presumably have decreased affinity for the mecA promoter and are inactivated by the BlaR1-mediated proteolysis of BlaI only. The observed proteolytic degradation of MecI, as illustrated in Fig. 2, most likely is a consequence of the destabilization of MecI dimers, as previously observed for MecR2, but is not essential for the optimal expression of β-lactam resistance induced by bla. As a matter of fact, it has been suggested that BlaI heterodimers containing an intact monomer and a cleaved monomer do not bind to DNA (16, 36, 37); hence, heterodimers consisting of intact MecI monomer and cleaved BlaI monomer might be compatible with an induced state and optimal expression of β-lactam resistance.

Fig 4.

Effect of MecI^MUT on the induction of β-lactam resistance by pbla. The native β-lactamase plasmid (pbla) of prototype CA-MRSA strain MW2 was transformed into strains COL, VNG17, RJP17, and HU25, overexpressing a mutant variant of MecI for the cleavage site (MecI^MUT). The oxacillin resistance levels were evaluated by diffusion disks containing 1 mg of oxacillin. WT, wild-type parental strain.

Fig 5.

Effect of pbla on the proteolysis of a MecI mutant variant for the cleavage site (MecI^MUT). Western blotting of MecI expression under inducing conditions (Oxa at 0.05 mg/liter). Lanes: 1, strain N315 (positive control); 2, N315 without oxacillin (negative control); 3, VNG17 + pGC::mecI^MUT; 3, VNG17 + pGC::mecI^MUT + pbla; 4, RJP17 + pGC::mecI^MUT; 5, RJP17 + pCG::mecI^MUT + pbla.

Polymorphisms on pbla are functionally important.

Prompted by the results from the overexpression of MecI^MUT in prototype strain HU25, we wondered why we have previously detected a massive effect on the phenotypic expression of β-lactam resistance upon deletion of mecR2 in prototype strain N315 and in other strains (19). In fact, strain N315 also carries a β-lactamase plasmid, and as such, based on the data from this study, the presence of bla regulators should compensate for the absence of mecR2. Taking advantage of the fact that the strategy used to generate the mecR2-null mutant in strain N315 generated a bla-negative intermediate, we transformed this strain with pbla from strain MW2. As illustrated in Fig. 6, the presence of pbla of strain MW2 compensates for the deletion of mecR2. These observations suggest that bla and mec regulatory systems have coevolved in the S. aureus population and have overlapping or redundant functions regarding the control of mecA expression and the optimization of the β-lactam-resistant phenotype.

Fig 6.

Polymorphisms on pbla are functionally important. (A) pbla of strain MW2 compensates for the mecR2 deletion in strain N315. The strategy used to generate the mecR2 null mutant of strain N315 generated a bla-negative intermediate which was transformed with pbla from strain MW2 (N315::ΔmecR2 + pbla/MW2). In contrast to the backcross mecR2 null mutant positive for pbla of the parental strain (N315::ΔmecR2 + pbla/N315), the presence of pbla of MW2 compensates for the mecR2 deletion. The oxacillin resistance levels were evaluated by diffusion disks containing 1 mg of oxacillin. (B) Phylogenetic analysis of BlaR1 variants from representative MRSA strains/lineages, sequences of which are available at GenBank. Strain codes and/or genetic lineages and GenBank accession numbers of the BlaR1 protein variants are the following: USA300a, strain TCH959, clone ST8-IV, YP_001569086; USA300b, strain TCH1516, clone ST8-IV, YP_001569057; USA300c, strain LAC, clone ST8-IV, ADM29210; MW2, clone ST1-IV, NP_863282; ST45-V, strain WBG8404, YP_06937210; N315, clone ST5-II, NP_395547; MRSA252, clone ST36-II, YP_041216; TW20, clone ST239-III, YP_005732670; and ST398, strain S0385, CAQ50407. USA300a-c, MW2, and ST45-V are CA-MRSA strains; N315, MRSA252, and TW20 are HA-MRSA; ST398 is of animal origin. The tree was constructed online at www.phylogeny.fr.

We have analyzed the pbla sequences of strains MW2 and N315, and we found polymorphisms in the coding region of blaR1. These polymorphisms generate point mutations in regions important for the signal transduction mechanism mediated by BlaR1, such as the transmembrane domain TM2 and loops L2 and L3 (38), and as such, they may account for the observed functional differences. Since MW2 is a CA-MRSA strain and N315 an health care-associated MRSA (HA-MRSA) strain, we wondered if any correlation could be established between pbla types and the epidemiological setting of strains. However, in line with previous observations (28, 39), no correlation could be established, as illustrated in Fig. 6B. For instance, BlaR1 protein sequences of three strains belonging to the same epidemic CA-MRSA clone in the United States, clone USA300 (ST8-IV), do not cluster together, and epidemiologically and clonally unrelated strains, such as strains N315 and USA300, may carry homologous BlaR1 proteins.

Constitutive PBP2a expression may interfere with the optimal expression of resistance.

S. aureus plasmids containing the β-lactamase locus are large (20 to 37 kb); therefore, they impose a fitness cost. In this study, this was clearly demonstrated by the differences in the densities of overnight cultures between parental strains VNG17, RJP17, and HT0350 and the respective recombinant strains transformed with pbla from prototype strain MW2 (data not shown). Nevertheless, pbla is maintained within the S. aureus population, presumably because it confers resistance to penicillins and other compounds, such as cadmium. Based on the fact that the presence of pbla enhanced the phenotypic expression of oxacillin resistance in the three strains with constitutive mecA expression tested, we reasoned that the transcriptional control mediated by the bla locus could confer a selective advantage by modulating the expression levels of mecA. To test this hypothesis, we performed competition assays between parental strains and pbla transformants, taking advantage of the penicillin-resistant phenotype encoded by pbla. As illustrated in Fig. 7, the prevalence of pbla-positive cells decreased under normal growth conditions and also if cultures were incubated for extended periods (60 h), simulating starvation stress conditions. However, in the presence of sub-MIC oxacillin at 0.5 mg/liter, the number of pbla-positive cells remains constant and may even increase after a starvation period. These observations suggest that in the presence of oxacillin, bla regulators, by modulating the amounts of mecA transcript and PBP2a protein, compensate for the perturbations in the cell wall physiology caused by the combined effects of the β-lactam antibiotic and the excess of PBP2a in clinical MRSA strains with no functional mecA regulatory locus. From this perspective, prototype strain COL is unique, because it expresses full oxacillin resistance in spite of the constitutive high-level mecA expression.

Fig 7.

Competition assays. (A) Strain VNG17 versus VNG17 + pbla; (B) strain RJP versus RJP17 + pbla; (C) strain HT0350 versus HT0350 + pbla. Culture mixes were inoculated daily in fresh medium with or without sub-MIC oxacillin (0.5 mg/liter). The assays were conducted for 4 to 10 days, and cultures were exposed to an extended incubation of 60 h to stimulate starvation conditions (gray box). For panel C, two experiments were performed. In one, oxacillin was added at the onset of the assay, and in the other, it was added after the extended incubation of 60 h.

Concluding remarks.

It has been recognized for a long time that the β-lactamase locus (bla) plays an important role in MRSA evolution, namely, by facilitating the acquisition and stabilization of the mecA gene (17, 18). Moreover, blaZ and mecA are regulated by highly homologous repressors and sensor inducers, which are interchangeable (11–13), although subtle differences do exist between the two regulatory systems, such as the existence of an antirepressor in the mecA locus and the specificity of repressor proteolysis (19, 38). In this study, we have demonstrated that bla also acts as a potent enhancer of the MRSA phenotype by disrupting the strong MecI-mediated repression on mecA and, in those strains with constitutive mecA expression, by modulating the expression levels of the resistance gene. These observations point to unrecognized, important functions of the bla locus for the MRSA phenotype. First, our study demonstrates that bla regulatory proteins BlaR1 and BlaI disrupt the strong repression of mecA mediated by MecI. Presumably, this effect is due to the formation of MecI::BlaI heterodimers. The formation of MecI::BlaI heterodimers has been demonstrated (15), and based on kinetic studies for BlaI binding to the operator sequences (29), we speculate that these heterodimers have reduced affinity for the mecA promoter and become more susceptible to the BlaR1-mediated proteolytic inactivation of BlaI. This process, similar to what happens with the interaction between MecI and MecR2, may foster the proteolysis of MecI as well, although according to our observations, this is not strictly required for an efficient bla-mediated induction of mecA. Second, our study demonstrates that the bla locus, by modulating the expression of mecA, not only stabilizes its acquisition but also fully optimizes the phenotypic expression of resistance, presumably by maintaining the delicate levels of protein complexes and synthesis and turnover pathways at the cell wall, the cellular target of β-lactams. In short, two regulatory systems, blaR1-blaI and mecR1-mecI-mecR2, with redundant and overlapping functions, ensure optimal expression of the MRSA phenotype. Taking into account that clinical MRSA strains with constitutive mecA expression are extremely rare, strategies targeting those two regulatory systems may be important for the effective control of MRSA infections.