INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The targets for -lactam antibiotics are penicillin-binding proteins (PBPs), enzymes that function to cross-link peptidoglycan in the bacterial cell wall (7). Staphylococci can resist the action of -lactam antibiotics by producing a new PBP, PBP2a, with reduced -lactam affinity that is able to mediate essential cell wall construction when -lactam-susceptible PBPs have been inactivated (14, 15). The gene that encodes PBP2a, mecA, was acquired from an unknown donor bacterium, together with 30 to 50 kb of additional DNA, and inserted at a specific chromosomal site (1, 6, 19). The mec complex also contains a two-gene operon, mecR1-mecI, that is divergently transcribed from and has overlapping promoter-operator regions with mecA (31, 34). mecR1 encodes a signal sensor/transducer with antirepressor activity, while mecI encodes a repressor of mecA transcription (21). In the absence of induction via MecR1 by a -lactam antibiotic, mecA transcription is tightly repressed by mecI (31, 33).

Staphylococci can also resist the action of certain -lactam antibiotics by producing an enzyme, -lactamase, that inactivates these antibiotics by hydrolyzing the -lactam ring. The gene encoding -lactamase, blaZ, is usually plasmid encoded. -Lactamase production is also regulated by two divergently transcribed genes, blaR1 and blaI, whose gene products have amino acid homology to mecR1 and mecI, respectively (9, 17, 38). There are similarities between the promoter-operator regions of the mecA and blaZ regulons, and both purified MecI and BlaI have been shown to bind to bla promoter-operator sequences in DNase protection assays (9).

On the basis of the similarities between blaZ and mecA regulatory sequences, it is reasonable to assume that each regulon may modulate the transcriptional expression of mecA. Previous studies provide evidence that mecA transcription and PBP2a production can be affected by BlaI. Hackbarth and Chambers (10) reported that PBP2a production was converted from unregulated and constitutive to inducibly repressed by the introduction of a plasmid containing blaR1 and blaI. In addition, Ryffel et al. (32)found that bla-mediated repression of mecA transcription was less stringent at baseline and more rapidly inducible than mec-mediated repression. However, the staphylococcal strains used in these studies contained uncharacterized mec regulatory sequences or were nonisogenic, and mecA transcription was not quantified.

In this study, we used constructs that differed only in terms of defined mutations in mec and bla regulatory sequences in the presence of mec-reporter gene fusions to quantify the effect of these regulatory sequences on mecA expression. In addition, we investigated the biochemical basis of mecA coregulation by comparing the binding of purified MecI and BlaI to mecA promoter-operator sequences and by assessing MecI-BlaI protein-protein interactions.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

All strains and plasmids used in this study are listed in Table 1. For routine culture, S. aureus strains were grown at 37°C in either Trypticase soy or brain heart infusion broth. Escherichia coli strains were grown in Luria broth. For solid media, agar-agar was added to broth at a final concentration of 1.5%. All media used for growth and maintenance of bacteria strains were purchased from Difco Laboratories, Detroit, Mich. For the culture of yeast strain Y190 (Clontech, Palo Alto, Calif.) and its derivatives, yeast extract-peptone-dextrose (YPD) or dropout base supplemented with complete synthetic medium lacking leucine, tryptophan, or both (Bio 101, Vista, Calif.) was used. Antibiotics were used at the following concentrations: ampicillin (Amp), 50 µg/ml for E. coli; chloramphenicol, 10 µg/ml; erythromycin, 10 µg/ml; minocycline, 1 µg/ml; and gentamicin, 1 µg/ml. All antibiotics and chemicals were purchased from Sigma Chemical Company, St. Louis, Mo. Restriction endonucleases and other enzymes were purchased from New England Biolabs, Beverly, Mass.

Cloning, transformations, and DNA manipulations. All restriction endonuclease digestions and ligations were performed as directed by the manufacturer. E. coli DH5 and Able K competent cells, obtained from Gibco-BRL (Grand Island, N.Y.) and Stratagene (La Jolla, Calif.), respectively, were transformed according to the manufacturer's directions. Yeast strain Y190 was transformed by the lithium acetate method as described by Clontech (Clontech Matchmaker System Manual; Clontech Laboratories, Palo Alto, Calif.). Small-scale isolation of plasmid DNA from E.coli and S. aureus was done by the method of Birnboim and Doly (3) and the hexadecyltrimethylammonium bromide extraction method of Townsend et al. (35), respectively. Large-scale plasmid DNA isolations were obtained using Qiagen Midi Prep affinity columns (Qiagen, Chatsworth, Calif.). Plasmid DNA was isolated from Saccharomyces cerevisiae using the RPM yeast plasmid isolation kit (Bio 101, Vista, Calif.). Recombinant plasmids were transferred between S. aureus strains by transduction as described by Sharma et al. (33).

-Galactosidase activity. An overnight culture diluted in warmed BHI or BHI supplemented with -lactam inducer was adjusted to an optical density at 600 nm (OD₆₀₀) of approximately 0.050 and shaken at 37°C until an OD₆₀₀ of 0.600 was reached (2 h). Cells were harvested by centrifugation and stored overnight at 80°C. The thawed pellet was resuspended in 1 ml of Z buffer (0.06 M Na₂HPO₄, 0.04 M NaH₂PO₄, 0.01 M KCl, 0.001 M MgSO₄, 0.05 M -mercaptoethanol, pH 7.0), and 500 µl of the suspension plus an additional 500 µl of Z buffer was added to 2.5 g of 1-mm Zirconia beads (Biospec Products, Bartlesville, Okla.). Cell lysates were produced by bead beating for 5 min at 4°C using the Biospec Products (Bartlesville, Okla.) bead beater, and lysates were centrifuged for 10 min at 26,895 × g. A 50- to 100-µl sample was taken for assay of -galactosidase activity using ONPG (o-nitrophenyl--D-galactoside) as the substrate. Activity was expressed as Miller units per milligram of protein. Protein concentration was determined with the Bio-Rad protein assay (Bio-Rad, Hercules, Calif.)

DNase I footprinting. DNA fragments used for DNase I footprinting were prepared by PCR amplification, and the procedure was performed as described by Sharma et al. (33).

Construction of Gal4 protein fusion plasmids and production of yeast extracts. A 422-bp BamHI fragment from plasmid pGO443 containing the mecI coding sequence was cloned into the BamHI site of pACT2 and pAS1. The blaI coding sequence was obtained from plasmid pI6187 by PCR amplification with the following primers: blaI upstream primer, 5'AA ATG GGT GTT TCA AAT G 3' (bold C denotes a T in the wild-type sequence, change made to prevent early termination), and blaI downstream primer, 5'GAC ACT GAA TTT GCT C 3'. The blaI PCR products were purified by gel electrophoresis and cloned into the SmaI site of pAS1 and pACT2. To confirm the presence of the in-frame junction from both fusion constructs, recombinant plasmids were sequenced using the Dye Terminator cycle sequencing kit (Applied Biosystems, Foster City, Calif.) and the ABI 377 DNA sequencer.

Western blot analysis. For Western immunoblot analysis, S. aureus whole-cell lysates were prepared as described by Gregory et al. (9).We resolved 80 µg of total protein in 16.5% Tris-tricine minigels by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. The membranes were blocked with 5% skim milk in TBST (20 mM Tris-HCl, 500 mM NaCl, 0.05% Tween 20, pH 7.5) for 1 h and incubated with rabbit anti-MecI polyclonal antiserum diluted 1:100 with the blocking solution overnight at 4°C. The secondary antibody was horseradish peroxidase-conjugated donkey anti-rabbit immunoglobulin (Amersham Pharmacia Biotech Inc., Piscataway, N.J.). Bound antibodies were detected by color development using nitroblue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate (Bio-Rad Laboratories, Richmond, Calif.) or by chemiluminescence with the ECL Western blot detection system (Amersham Pharmacia Biotech Inc., Piscataway, N.J.).

Constructs used for mecA transcriptional analysis. mecA transcription was assessed by analysis of -galactosidase activity from a mecA-lacZ fusion. The previously described plasmid pGO630 (33) contains the 5' 50 bp of mecA transcriptionally fused to lacZ. mecA is preceded by the mecA-mecR1 intergenic region, which contains both the mecA and mecR1 promoter-operator sequences, followed by the mecR1 and mecI regulatory genes. This plasmid was introduced into the chromosomal phage L54a att site of methicillin-susceptible S. aureus strain GO450TF, interrupting the lipase (geh) gene. This strain was designated GO450TF::630.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Binding of MecI and BlaI to mec promoter-operator sequences. MecI and BlaI were purified as glutathione-S-transferase fusion proteins and cleaved from the fusion as previously described (33). Equal amounts of cleaved, purified MecI and BlaI proteins, as estimated by Coomassie blue staining of polyacrylamide gels, were used in DNase protection assays. A 210-bp DNA fragment containing the mecA-mecR1 intergenic region, 16 bp of the 5' end of mecA, and 6 bp of the 5' end of mecR1 was used as the mec target. As previously described (33), MecI bound to a 43-bp region of mec DNA that included a 30-bp palindrome with 15 bp of dyad symmetry, the mecA 10 promoter sequence, and the mecR1 35 sequence (Fig. 1). As shown in Fig. 1, BlaI protected the same sequences as MecI.

View larger version (138K): [in this window] [in a new window]   FIG. 1.   DNase I protection analysis. DNase I footprint of the 210-bp target fragment containing the region between the mecA and mecR1 translational start sites. MecI and BlaI protected regions are shown. Lanes 1 through 7 contain 0, 0.001, 0.01, 0.05, 0.15, 0.4, and 1 mg of MecI or BlaI, respectively. C, A, T, and G are the labeled nucleotides of the sequencing ladder generated from each target fragment by using primers complementary to the 5' end. The sequence is read from bottom (5'end) to ttop (3' end). This image was scanned from the original film by using an AlphaImager 1000 (Alpha Innotech) and was cropped and labeled by using Canvas 5 graphics software (Deneba, Miami, Fla). The same systems were used to prepare Fig. 2.

Repression of mecA transcription. Maximal unregulated transcription from a chromosomal mecA-lacZ fusion was determined after inactivation of mecR1 and mecI by allelic replacement with the tetM gene, as described previously (26). As shown in Table 2, -galactosidase activity from a mecR1 mecI mutant (GO450TF::630I) was 66-fold greater than the activity from the regulated chromosomal gene (GO450TF::630). A gene dosage effect of MecI-mediated repression was shown when plasmid-encoded MecI (pGO621C) was introduced into the strain containing the regulated chromosomal mecA-lacZ fusion (GO450TF::630). -Galactosidase activity decreased more than sixfold in the presence of plasmid-encoded MecI. Similarly, there was an almost twofold decrease in transcription in the presence of high-copy blaI (pCH2278). Thus, there was additive repression provided by the increased gene dosage of plasmid-encoded repressors over that seen when only the chromosomal copy of mecI was present. However, there was more repression mediated by the MecI-containing plasmid constructs than by those containing BlaI. This was despite the observation, shown in the Western blot in Fig. 2, that approximately the same amounts of BlaI and MecI were produced by each of the plasmids. The increased repression of MecI over BlaI was also seen when a plasmid encoding each of the repressors was introduced into the strain containing an unregulated chromosomal mecA-lacZ fusion (GO450TF::630I; Table 2).

View larger version (52K): [in this window] [in a new window]   FIG. 2.   Western immunoblot analysis showing relative affinity of anti-MecI polyclonal antiserum for MecI and BlaI. In the first and second lanes, equal amounts of purified MecI and BlaI, as determined by Coomassie staining, were probed to demonstrate the equal affinity of anti-MecI antiserum for both repressors. The other lanes illustrate the production of repressor protein from plasmid constructs: GO450::630I alone, third lane; GO450::630I plus pGO621C (MecI), fourth lane; and GO450::630I plus pCH2278 (BlaI), last lane.

Analysis of MecI and BlaI interactions. The gene dosage effect, by which increasing amounts of repressor led to increasing transcriptional repression, may have been due in part to protein-protein interactions. A previous study suggested that optimal repression was dependent on the ability of MecI to form dimers and higher-0order oligomers (33). We used the yeast two-hybrid assay to assess the formation of MecI-MecI homodimers and MecI-BlaI heterodimers.

Induction of mecA transcription. MecI-repressed mecA transcription was induced from the chromosomal copy of mec regulators (GO450TF::630) with CBAP (2-[2'-carboxyphenyl] benzoyl--aminopenicillanic acid) at concentrations of 1, 5, 10, and 20 µg/ml. Maximal induction without growth inhibition was achieved at 10 µg/ml. As shown in Table 2, after 2 h of induction, CBAP at 10 µg/ml produced only a 12-fold induction of MecI-repressed mecA transcription, a value that represented only 19% of unregulated transcription seen in the mecI mutant (GO450TF::630I). There was a much greater induction from the plasmid-encoded copy of mec regulators (pGO621C) due to increased baseline repression, but induction still reached only 18% of the unregulated value. In contrast, induction by CBAP at 10 µg/ml of BlaI-repressed mecA transcription in the mecR1 mecI mutant resulted in a 44-fold increase in mecA transcription that was 81% of maximal.

View larger version (31K): [in this window] [in a new window]   FIG. 3.   Time course comparing induction of mecA expression via MecR1 and BlaR1. -Galactosidase activity is in Miller units per milligram of protein per hour. Samples were taken at the indicated times after induction with 10 µg of CBAP per ml. The solid bars represent CBAP induction of strain GO450TF::630I (mecR1 mecI) plus pGO621C (mecR1+ mecI+); the striped bars represent induction of strain GO450TF::630I (mecR1 mecI) plus pCH2278 (blaR1+ blaI+).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Investigators have previously shown that the genes regulating production of -lactamase can also regulate PBP2a production or mecA transcription (10, 31, 36). We have also recently demonstrated, using Northern blots in clinical isolates containing -lactamase plasmids and mecI allelic replacement knockouts, that BlaI also regulates mecA transcription in Staphylococcus epidermidis. (5). We extended these data by demonstrating that purified MecI and BlaI also bound to the same mecA-mecR1 promoter-operator sequences. Gregory et al. (9) demonstrated that both MecI and BlaI also bound the same bla promoter-operator sequences, and Lewis and Dyke (25) showed that both could regulate -lactamase production. While the overall sequence homology of the mecA-mecR1 and blaZ-blaR1 intergenic regions is not high (59%), there is 81% homology between the arms of the respective mec and bla promoter-operator dyads.

In order to compare the activities of MecI and BlaI as repressors, we assessed the ability of plasmid-encoded regulators to affect transcription of a mecA-lacZ transcriptional fusion. We have previously shown that mecA can be regulated by MecI and BlaI in methicillin-resistant S. aureus isolates using Northern blots (T. Dickinson and G. Archer, unpublished observations). All comparisons were made in the same genetic background, and both plasmid-encoded repressors were provided in similar amounts, as shown by Western blot analysis. Several observations were notable. First, there was a gene dosage effect of MecI, so that when repressor was provided on a high-copy plasmid there was an increase in mecA transcriptional repression over that seen when MecI was encoded on the chromosome. Furthermore, there was additive repression between either plasmid-encoded MecI or BlaI and chromosomal MecI.

The explanation for the direct association of mecA transcription with repressor quantity may be progressive saturation of unsaturated binding sites. The amount of MecI produced from the chromosomal gene may be insufficient, due to autoregulation, to completely saturate binding sites. This provides sufficient low-level transcription of mecA to mediate -lactam resistance and of mecR1 to mediate induction. As repressor quantity increases, protein-protein interactions may also promote target saturation. Repressors with amino-terminal helix-turn-helix motifs like MecI and BlaI bind to DNA as dimers (13). Dimerization is energetically coupled to operator binding, and monomers and dimers are in dynamic equilibrium (2, 20). We used the yeast two-hybrid assay to demonstrate the ability of MecI and BlaI to form both homodimers and heterodimers with equal efficiency. In addition, we have previously associated formation of higher-order protein oligomers with optimum MecI repression (33), a finding also noted for the lambda cI repressor (20). Thus, increasing quantities of either MecI or BlaI could shift the equilibrium from monomers to dimers and higher-order oligomers, promoting more rapid and complete saturation of repressor binding sites.

Whereas the MecI and BlaI repressors were virtually interchangeable with regard to target interactions and activity, there was a marked difference between the two inducers. First, induction of mecA transcription was more rapid and more complete through BlaR1 than through MecR1. Second, there was a difference in the relative responses of the two inducers to low concentrations of structurally different -lactam antibiotics: CBAP, a penicillin, and cefoxitin, a cephalosporin. Finally, there was a lack of cross-induction by inducer (BlaR1) for the heterologous repressor (MecI).

The differential response of the two inducers may be due to a difference in amino acid composition between the sensor regions of BlaR1 and MecR1. BlaR1 from Bacillus licheniformis is a 606-amino-acid transmembrane protein with its carboxy terminus external to the cell membrane. The terminal 250 amino acids contain motifs for -lactam binding and have been shown to complex with -lactam antibiotics (22, 39). While the -lactam binding motifs are identical between sequences from Bacillus and staphylococcal BlaR1 and MecR1, there is only 48% amino acid identity between the entire staphylococcal BlaR1 and MecR1 carboxy-terminal sensor regions. This low homology could easily explain differences in the rapidity of induction and -lactam selectivity of the two proteins.

The specificity of each inducer for its homologous repressor may also be related to sequence differences between the predicted cytoplasmic portions of BlaR1 and MecR1. Zhang et al. (38) recently showed that induction through BlaR1 produced a series of reactions whereby BlaR1 first underwent autoproteolysis, converting itself into a protease. Activated BlaR1 then inactivated BlaI by site-specific proteolytic cleavage. We have evidence that MecI is cleaved at a site very similar to the one in BlaI following -lactam induction through MecR1 (J. M. Bosilevac and G. Archer, unpublished data). Signal transduction is thought to occur through four membrane-spanning segments, inducing conformational change in two cytoplasmic domains, one of which contains the zinc metalloprotease site (38). The existence of additional chromosomal genes that are required for induction has also been postulated (4). Since there is only 13% homology between the MecR1 and BlaR1 cytoplasmic domains outside the metalloprotease motif, the specificity for induction by homologous inducers is probably due to specific interactions between inducer and repressor or inducer and an intermediate molecule. While there are examples of proteolytic cleavage as a mechanism of gene regulation, such as in the inactivation of alternative sigma factors in recovery from the heat shock response (16), response regulation through repressor-specific proteolytic inactivation, as exemplified by BlaR1 and MecR1, constitutes a unique mechanism for signal transduction that may have homologues among other bacteria.