Mutation of RNA Polymerase β-Subunit Gene Promotes Heterogeneous-to-Homogeneous Conversion of β-Lactam Resistance in Methicillin-Resistant Staphylococcus aureus

S. aureus and plasmids used in this study are listed in Table 1. The cloning and transformation of Escherichia coli strain DH5α were carried out by standard techniques (TaKaRa-Bio Co., Ltd., Shiga, Japan). All S. aureus strains were grown in brain heart infusion (BHI) broth or agar (Becton, Dickinson and Co.[BD], Franklin Lakes, NJ), with aeration at 37°C for over 18 h. The antibiotics ampicillin (Wako Pure Chemical Industries, Osaka, Japan), chloramphenicol (Sigma Chemical Co., St. Louis, MO), methicillin (SmithKline Beecham/GlaxoSmithKline, London, United Kingdom), and imipenem (WAKO) were used in this study.

The standard methods of DNA manipulations were described previously (25). Genomic DNAs and plasmid were prepared with Miniamp and Miniprep kits (Qiagen, Inc., Valencia, CA). Restriction enzymes were used as recommended by the manufacturer (TaKaRa). Routine PCR amplification was performed with an Expand High-Fidelity system (Roche, Mannheim, Germany).

Sequencing of the genome of strain N315ΔIP-H5 (abbreviated as ΔIP-H5) was performed with a Genome Sequencer 20 system, a recently introduced highly parallel genome sequencer from 454 Life Sciences (Branford, CT). The sequence assembly and gap closing were carried out as described previously (20). Since ΔIP-H5 is an in vitro derivative of N315, the resulting sequence of the ΔIP-H5 genome was then compared to that of N315, and then the discovery of genome-wide single nucleotide polymorphisms (SNPs) was achieved by whole-genome alignments with the Mummer (version 3.20) software package (http://mummer.sourceforge.net/). PCR amplification and sequencing of ΔIP-H5 chromosomal DNA found no additional mutations except for an rpoB mutation, which causes the replacement of a single amino acid, Asn 967, with Ile (N967I).

To analyze SNPs of three strains, H5 rpoB(I967N), N315 rpoB(N967I), and N315h4, 90-bp paired-end read sequencing was performed using a Hiseq2000 sequencing platform (Illumina Inc., San Diego, CA) at the infoBio company (Tokyo, Japan). Sequences of 90-bp paired-end reads from each bacterial strain were obtained in a single lane of a flow cell. Image analysis was performed with Illumina's Pipeline Analysis software, version 1.8. The 15.1-Mb, 18.2-Mb, and 15.9-Mb reads were obtained from the chromosomal DNA library of strains H5 rpoB(I967N), N315 rpoB(N967I), and N315h4, respectively. As de novo DNA analysis, each read was mapped to the reference whole-genome sequence of N315, and then mapping of SNPs was extracted by using the software Genome Traveler (in silico Biology Inc., Yokohama, Japan).

For the chromosomal allele replacement of rpoB(N967I) in N315 or rpoB(I967N) in ΔIP-H5, we used the pKOR1 allele replacement system, as described previously (26). In brief, a 3.0-kb of rpoB insert DNA encompassing a 1-kb flanking sequence of a phage attachment site was generated by PCR from chromosomal DNA of strain N315 or ΔIP-H5 by using the following primers: attB1-rpoB (5′-GGGGACAAGTTTGTACAAAAAAAGCAGGCT-3′) and attB2-rpoB (5′-GGGGACCACTTTGTACAAGAAAGCTGGGT-3′). The resulting plasmid, pKO-rpoB(967I) or pKO-rpoB(967N) was introduced into S. aureus N315 or ΔIP-H5 by electroporation, generating transformant N315(pKO-rpoB[967I]) or ΔIP-H5(pKO-rpoB[967N]). The replacement of the rpoB gene was carried out by a two-step procedure. We identified the desired strains H5 rpoB(I967N) and N315 rpoB(N967I) by determination of the nucleotide sequence using the following three primers; 5′-CAGCTGGATGGCAAATAATG-3′, 5′-TGTGGAATTGTGAGCGGATA-3′, and 5′-AATTGCGCTTTACCGCCAAGTGGTTGT-3′. Also we confirmed the absence of other additional mutations in the mutants by SNP analysis, except for a mutation of sa0168(M199V) in N315 rpoB(N967I).

The susceptibility of several antibiotics was examined by Etest (AB Biodisk, Solna, Sweden) and population analysis, as described previously (10). All examinations were performed by using BHI broth or agar plates.

The preparation and the examination of S. aureus cells by transmission electron microscopy were performed as described previously (22, 27). At least 120 cells of each sample with nearly equatorial cut surfaces were measured for the evaluation of cell wall thickness, and the results were expressed as the means ± standard deviations.

The doubling time was calculated, as described previously (22). The growth conditions were 37°C with shaking at 25 rpm in a TN-2612 incubator (Advantec, Osaka, Japan). The optical density at 660 nm (OD₆₆₀) versus time was plotted for each strain in the exponential growth phase. A linear regression curve was obtained from values between the log₂ OD₆₆₀ and the log₂ t, where t is time (R ≥ 0.999). The doubling time was calculated follows: doubling time = [(t₂ − t₁) × log₂]/(log OD₆₆₀ at t₂ − log OD₆₆₀ at t₁), where t₁ and t₂ are the times at the start and the end of the logarithmic growth phase. At least three independent experiments were carried out.

Triton X-100-stimulated autolysin activity in Tris-HCl buffer (pH 7.5) was measured as described previously (28). Cells were grown to mid-exponential phase to an OD₆₆₀ of about 1.5, with a cultivation temperature of 37°C. The culture was rapidly chilled, and cells were washed twice with ice-cold distilled water and suspended to an OD₆₆₀ of 1.5 in 50 mM Tris-HCl buffer supplemented with 0.05% Triton X-100. Autolysis was measured during incubation at 30°C as a decrease in the OD₆₆₀ by using a model TN-2612 biophotorecorder (Advantec Osaka, Japan). All data from the autolysis experiments are reported as percentages of the initial turbidity (at the zero time point) and are representative of three independent experiments.

RNA extraction, cDNA labeling, hybridization, and data analysis for microarray analysis were carried out according to protocols described previously (29).

The strains were assayed for PBP2′ production by Western blotting. S. aureus membrane proteins were prepared from mid-exponential-stage cultures. An overnight culture (0.1 ml each) was inoculated into 15 ml of fresh BHI broth and allowed to grow to an optical density at 660 nm of 1.0. Then each strain was induced by methicillin at 0, 1, 8, or 128 mg/liter for 1 h, as described previously (10). The cells were harvested and washed with buffer (50 mM Tris, 150 mM NaCl, 5 mM MgCl₂ [pH 7.5]) and then resuspended in the same buffer. Lysostaphin, DNase, and RNase were added to final concentrations of 200, 20, and 10 mg/liter, respectively, and then the mixture was incubated at 37°C for 30 min. The lysed cells were centrifuged at 4,400 × g for 10 min, and the supernatant was ultracentrifuged at 110,000 × g for 40 min. The resultant pellet was washed twice and resuspended in 50 mM phosphate buffer (pH 7.0). Membrane proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred as described previously (12). PBP2′ was detected with a mouse anti-PBP2′ monoclonal antibody (a gift of Denka Seiken Co., Ltd., Niigata, Japan) as the primary antibody (diluted 1:100,000) and alkaline phosphatase-conjugated anti-mouse immunoglobulin (Promega, Madison, Wis.). We detected bound antibodies by a color development system (Bio-Rad), as directed by the manufacturer.

The statistical significance of the data was evaluated with Student's t test.

The transcriptional profiles of ΔIP-H5 relative to N315ΔIP and of N315 rpoB(N967I) relative to N315 are deposited in the NCBI under GEO accession number GSE42218.

To find the chr* genetic determinant, we determined the whole genomic sequence of homo-MRSA strain ΔIP-H5 (oxacillin MIC of >256 mg/liter) which had been obtained from hetero-MRSA strain N315ΔIP (oxacillin MIC of 8 mg/liter) by selection with 8 mg/liter of imipenem (30). We have already reported that the imipenem population analysis for N315ΔIP (imipenem MIC of 0.25 mg/liter) was performed, and a selective concentration of imipenem was defined as the concentration which resulted in a reduction of the initial bacterial population by approximately 4 log units (Fig. 1A) (20). The imipenem MIC increased to >32 mg/liter from 0.25 mg/liter in ΔIP-H5 (Table 2). Alignment of a 2,735,083-bp-long chromosome sequence of ΔIP-H5 with that of the parent strain N315ΔIP revealed a single mutation in the rpoB gene with the amino acid change Asn967→Ile (N967I). The amino acid substitution N967I was located outside the rifampin resistance-determining region (RRDR) (31).

To verify the effect of rpoB(N967I), we constructed two mutant strains, N315 rpoB(N967I) and H5 rpoB(I967N), which were derived from pre-MRSA strain N315 and homo-MRSA strain ΔIP-H5, respectively. H5 rpoB(I967N) was the reverse (or back) mutant of ΔIP-H5 that has the rpoB(N967I) mutation. We then determined the antibiograms of these mutants together with their parent strains by Etest and population analysis. As shown in Table 2, the MICs of methicillin (Fig. 2), oxacillin, and imipenem in the back-mutant strain H5 rpoB(I967N) reverted to the level of parent strain N315ΔIP. Also, the pattern of the population analysis curve for methicillin in H5 rpoB(I967N) also returned to the heterogeneous type (Fig. 1B). On the other hand, N315 rpoB(N967I) exhibited Eagle-type resistance with a unique inhibitory zone on the agar plate containing lower concentrations of methicillin (from 2 to 12 mg/liter) and grew better on the agar plates containing high concentrations of methicillin (from 12 to >256 mg/liter) (Fig. 2). The population analysis curve of N315 rpoB(N967I) showed an Eagle-type pattern similar to that of N315h4 (Fig. 1B). Other MICs for vancomycin, rifampin, and daptomycin did not change among all the tested strains, except for a decreased linezolid MIC for the Eagle-type strain N315h4 (Table 2). Eagle-type resistance occurs because of the poor induction of mecA gene expression by methicillin at lower concentrations (10). In this regard, oxacillin seemed to be a much better inducer of mecA gene transcription, and Eagle-type resistance is not clearly demonstrated by population analysis using oxacillin (data not shown). This coincides with the MIC data in Table 2: the double inhibitory zone was observed with methicillin and imipenem in the two Eagle-type strains but not with oxacillin (>256 mg/liter) (Fig. 2).

Eagle-type strain N315h4 is known to exhibit tolerance to the killing activity of high concentrations of methicillin (10). As shown in Fig. 3, N315h4 demonstrated equivalent susceptibility to N315 to 8 mg/liter of methicillin. However, N315h4 exhibited reduced susceptibility to the cytokilling activity of methicillin at 128 mg/liter. This high-dose methicillin tolerance was also observed with N315 rpoB(N967I) (Fig. 3). Homo-MRSA strain ΔIP-H5 is resistant to high concentrations of methicillin, so it increases in cell number in the presence of 128 mg/liter of methicillin. On the other hand, ΔIP-H5 rpoB(I967N) harboring the reverted wild-type rpoB gene lost both resistance and tolerance to methicillin and returned to the pattern of the cytokilling profile of N315ΔIP (Fig. 3). Thus, the mutation rpoB(N967I) was shown to confer on the cells tolerance to high doses of methicillin.

To verify the amount of PBP2′ in all of the tested strains, we performed Western blot analysis for PBP2′. The amount of PBP2′ was measured before and after exposure to 1, 8, or 128 mg/liter of methicillin for 1 h (Fig. 4). We have already reported that N315 produced a very low-level amount of PBP2′ before induction with methicillin because of the mecI gene-mediated repression of mecA gene transcription (12). In contrast, N315ΔIP was found to produce detectable amounts of PBP2′ without methicillin induction because PBP2′ production was derepressed by the absence of mecI gene (Fig. 4). PBP2′ was produced constitutively in homo-MRSA ΔIP-H5 and hetero-MRSA H5 rpoB(I967N) in the absence of the mecI repressor gene of mecA, both before and after induction with methicillin (Fig. 4).

The PBP2′ production in the two Eagle-type MRSA strains N315h4 and N315 rpoB(N967I) was also repressed both before and after induction with 0 and 1 mg/liter of methicillin, as described previously (10). Induction with 8 and 128 mg/liter of methicillin was able to induce PBP2′ production appreciably.

To investigate the genetic alteration responsible for the Eagle-type resistance of N315h4, we determined its whole genome sequence and compared it with that of N315. We found only one mutation in rpoB in the N315h4 chromosome, and no other mutation was found. However, the amino acid position and substitution were different from those of ΔIP-H5, which carried R644H. Thus, certain rpoB mutations are capable of conferring high-dose methicillin tolerance and homo- as well as Eagle-type methicillin resistance.

All three rpoB mutant strains, N315 rpoB(N967I), N315h4, and ΔIP-H5, had prolonged doubling times compared with their parent strain N315 or with N315ΔIP carrying the wild-type rpoB gene (Table 2). Electron microscopic observation revealed that ΔIP-H5 had a thicker cell wall (21.6 ± 1.8 nm) than N315ΔIP (17.7 ± 1.5 nm) and its back-mutant strain H5 rpoB(I967N) (14.2 ± 1.2 nm). The cell wall of Eagle-type strains N315h4 (20.2 ± 1.8 nm) and N315 rpoB(N967I) (17.2 ± 1.8 nm) were thicker than the cell wall of the parent strain N315 (14.1 ± 1.3 nm). The significant difference between the rpoB mutant strains and wild-type rpoB strains was determined by a two-sided Student's t test (P < 0.018). The results indicated that the doubling time prolongation and cell wall thickening were associated with the rpoB(N967I) mutation.

Microarray analysis of ΔIP-H5 and N315 rpoB(N967I) revealed 147 and 163 differentially expressed genes, respectively, compared to the parent strains. Among the genes, 33 downregulated and 38 upregulated genes, respectively, were commonly found in ΔIP-H5 (compared to N315ΔIP) and N315 rpoB(N967I) (compared to N315), except for sa2430 encoding the zinc metalloproteinase aureolysin (Table 3). With reference to the tolerance phenotype of the rpoB(N967I) mutation, we noticed a change in expression of the genes involved in autolysin activity and the cell wall synthesis function. The transcriptional level of cidABC encoding the complex up-regulatory system for autolysin activity (32–36) was decreased by 0.42-, 0.48-, and 0.44-fold in ΔIP-H5 relative to N315ΔIP and by 0.25-, 0.27-, 0.26-fold in N315 rpoB(N967I) relative to N315, respectively. On the other hand, the expression levels of the negative regulators of autolysis, lrgAB (35, 37), were increased by 2.53- and 2.08-fold in ΔIP-H5 relative to N315ΔIP and by 6.60- and 6.57-fold in N315 rpoB(N967I) relative to N315, respectively.

To observe the effect of rpoB(N967I) on the regulation of the genes involved in autolysis, we tested Triton X-100-induced autolysis of the rpoB(N967I) mutant strains (Fig. 5). The two rpoB(N967I) mutants, N315 rpoB(N967I) and ΔIP-H5, had significantly decreased Triton X-100-induced autolysis rates (50% lysis in about 100 to 110 min) compared to their parent strains N315 and N315ΔIP (50% lysis in 40 to 60 min). Moreover, N315h4 carrying rpoB(R644H) also exhibited a decrease in autolysis compared to the parent strain N315. These results indicated that the rpoB(N967I) and rpoB(R644H) mutations were correlated with the level of autolytic activity in S. aureus.