An RpoB Mutation Confers Dual Heteroresistance to Daptomycin and Vancomycin in Staphylococcus aureus

Abstract

We have previously reported the establishment of a Staphylococcus aureus laboratory strain, 10*3d1, having reduced susceptibility to daptomycin and heterogeneous vancomycin-intermediate S. aureus (VISA) phenotype. The strain was generated in vitro by serial daptomycin selection (Camargo, I. L., H. M. Neoh, L. Cui, and K. Hiramatsu, Antimicrob. Agents Chemother. 52:4289-4299, 2008). Here we explored the genetic mechanism of resistance in the strain by whole-genome sequencing and by producing gene-replaced strains. By genome comparison between 10*3d1 and its parent methicillin-resistant Staphylococcus aureus (MRSA) strain N315ΔIP, we identified five nonsynonymous single nucleotide polymorphisms (SNPs). One of the five mutations was found in the rpoB gene encoding the RNA polymerase β subunit. The mutation at nucleotide position 1862 substituted the 621st alanine by glutamic acid. The replacement of the intact rpoB with the mutated rpoB, designated rpoB(A621E), conferred N315ΔIP with the phenotypes of reduced susceptibility to daptomycin and hetero-VISA. The rpoB(A621E)-mediated resistance conversion was accompanied by a thickened cell wall and reduction of the cell surface negative charge. Being consistent with these phenotypic changes, microarray data showed that the expression of the dlt operon, which increases the cell surface positive charge, was enhanced in the rpoB(A621E) mutant. Other remarkable findings of microarray analysis of the rpoB(A621E) mutant included repression of metabolic pathways of purine, pyrimidine, arginine, the urea cycle, and the lac operon, enhancement of the biosynthetic pathway of vitamin B2, K1, and K2, and cell wall metabolism. Finally, mutations identified in rplV and rplC, encoding 50S ribosomal proteins L22 and L3, respectively, were found to be associated with the slow growth, but not with the phenotype of decreased susceptibility to vancomycin and daptomycin, of 10*3d1.

Daptomycin, a semisynthetic chemical derived from Streptomyces roseosporus, is a cyclic lipopeptide antibiotic having bactericidal activity against a broad range of aerobic and anaerobic Gram-positive bacteria (1, 27, 29, 48). The exact mechanism of action for daptomycin has not been fully elucidated yet. However, it is generally accepted that daptomycin binds to the cytoplasmic membrane of Gram-positive bacteria via calcium-dependent binding (27). Once bound, the lipopeptide tail of the molecule is inserted into the bacterial cell membrane. This tail serves as an ion channel through which an efflux of potassium, and potentially other ions, can pass through, thereby causing the bacterial cell to rapidly depolarize. Depolarization results in multiple failures in the DNA, RNA, and protein synthesis of the bacteria, ultimately resulting in a rapid bacterial cell death (27, 47). As daptomycin has a rapid bactericidal activity, it is approved in the United States and the European Union as an alternative option for the treatment of infections caused by S. aureus (44, 49).

Although daptomycin has a potent bactericidal activity against methicillin-resistant S. aureus (MRSA), it tends to have less antimicrobial activity against vancomycin-intermediate S. aureus (VISA) clinical strains (17, 43). Various groups reported that MRSA with reduced susceptibility to vancomycin and/or daptomycin were generated during the treatment with either vancomycin or daptomycin (21, 26, 32, 42). Recently, we reported the establishment of a laboratory MRSA strain 10*3d1 having dual heteroresistance to daptomycin and vancomycin by serial daptomycin selection (6). Despite its selection by daptomycin alone, this strain expressed raised resistance to both daptomycin and vancomycin, a thickened cell-wall, and a partially overlapped transcription profile with that of hetero-VISA (6). Here we explored the genetic mechanism for the dual heteroresistance of the strain 10*3d1. Using whole-genome sequencing and gene replacement experiments, we proved that a nonsynonymous mutation in rpoB, the gene that codes for the ß subunit of the bacterial DNA-dependent RNA polymerase, was responsible for the phenotypic conversion.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Staphylococcus aureus strain N315ΔIP is a derivative of strain N315, in which the repressor mecI was inactivated and the beta-lactamase plasmid was cured (31). The genotype of inactivated mecI and absence of the beta-lactamase plasmid is characteristic of recent Japanese methicillin-resistant S. aureus (MRSA) (23). 10*3d1 is an N315ΔIP derivative with a daptomycin-nonsusceptible and vancomycin-heteroresistant phenotype, obtained by gradual daptomycin exposure as described previously (6). Clinical VISA isolates Mu50 and MI, their vancomycin-susceptible in vitro derivatives Mu50-P35 and MI-P84, and respective VISA revertants Mu50-PR and MI-PR described elsewhere previously (13, 14) were used as control strains in some experiments. Among these two isogenic triple sets of VISA phenotype-associated S. aureus strains, all VISA strains are daptomycin nonsusceptible (17). All the strains were grown in brain heart infusion (BHI) broth (Difco, Detroit, MI) at 37°C with aeration if not otherwise indicated. For each experiment, an overnight culture was diluted 100-fold in prewarmed fresh BHI broth and further incubated with aeration to ensure exponential growth condition before sampling. The cell growth was monitored by measuring the optical density of the culture at 578 nm (OD₅₇₈) with a spectrophotometer (Phamacia LKB Biotechnology, Inc., Uppsala, Sweden).

Genome sequencing and SNP detection of daptomycin-nonsusceptible S. aureus strain 10*3d1 with short reads.

10*3d1 genome sequencing was performed using the Solexa/Illumina genome analyzer (Illumina, Inc., San Diego, CA), a recently introduced highly parallel genome sequencer. The chromosomal genome sequence of N315 (accession number BA000018) was used as a scaffold to assemble and orient the reads. In a duplicated run of this instrument, a total of 5,833,472 36-bp-long reads were collected, giving a total coverage of an about 74.5 genome equivalent. Short reads were then aligned to the N315 genome using a short-read mapping program of MetaGenomeGambler PRO (MGG), version 2.1.2 (In Silico Biology, Inc., Yokohama, Japan) with parameters of a seed size of 12 bp and a maximum gap size allowed on a read of 3 bp to generate primary single nucleotide polymorphism (SNP) calls. Only uniquely mapped reads were retained by setting values of −1 for cost to both open and extend a gap and of 15 for cutoff overlap length. We filtered SNP calls and combined them into a single list using the Assemble program of MGG, and the resulting SNPs were verified by manually inspecting multiple alignments of all short reads mapping to each SNP locus. The identified SNPs were then verified by resequencing of the corresponding loci of N315 and 10*3d1 genomic DNAs, together with resequencing the corresponding loci of N315ΔIP (its whole-genome sequence was determined recently [28]), a parent strain of 10*3d1. The resequencing was performed using an Applied Biosystems 3730 capillary sequencer (Applied Biosystems Japan Ltd.) with forward and reverse primers for each locus.

Complete-genome comparisons of N315ΔIP and 10*3d1 using NimbleGen microarray.

Complete genome comparison of 10*3d1 and its parent strain N315ΔIP was performed using an array-based service (CGS) provided by Roche NimbleGen, Inc. (Madison, WI). Briefly, the test (10*3d1) and reference (N315ΔIP) genome DNA samples were separately cleaved to pools of low-molecular-weight fragments and labeled with fluorescent dyes, Cy3 and Cy5, respectively. The labeled samples were then competitively hybridized to a NimbleGen CGS whole-genome tiling array, which was generated with the S. aureus N315 genome sequence as a reference. Resulting hybridization signals were analyzed using NimbleScan, version 2.5, software, and the signal ratio of the number of reference samples to the number of test samples for all probes were plotted versus N315 genomic position. The locations of probes along the genome, which have a significant ratio shift between reference and test probes for both strands, represent the possible sequence differences, including SNPs, deletions, sites of insertion, endpoints of inversions, or translocations. Those locations indicated were checked by PCR and resequencing of both test and reference genomic DNA.

Electron microscopic evaluation of cell wall thickness.

Preparation and examination of S. aureus cells by transmission electron microscopy were performed as described previously (15). Morphometric evaluation of cell wall thickness was performed using photographic images at a ×30,000 final magnification. Thirty cells of each strain with nearly equatorial cut surfaces were measured for the evaluation of cell wall thickness, and results were expressed as mean value ± standard deviation (SD).

Antimicrobial susceptibility testing.

The MICs of several antimicrobials were measured for the constructed mutants and parent strains using Etest strips (AB Biodisk, Sweden). For the daptomycin MIC, the medium was supplemented with 50 mg/liter Ca²⁺ calcium according to Clinical and Laboratory Standards Institute (CLSI) criteria (7). A sterile cotton swab was immersed in a 0.5 McFarland standard of tested bacterial culture before streaking on sterile Mueller-Hinton (MH) agar plates. The antimicrobial strips were applied after 10 min. Plates were then incubated at 37°C and read after 24 h.

Analysis of daptomycin- and vancomycin-resistant subpopulations (population analysis).

The population analysis shows how many cells in a fixed number of cells (usually about 10⁷ CFU) of each strain are resistant to various concentrations of antibiotics. Overnight cultures of tested strains in BHI broth were adjusted to an OD₅₇₈ of 0.3. Tenfold dilutions of these cell suspensions were then prepared, and 0.1 ml of each suspension was spread on BHI agar plates containing various concentrations of daptomycin or vancomycin. Plates were incubated for 48 h at 37°C, and the log number of CFU was plotted against daptomycin or vancomycin concentration.

Growth curve and doubling time.

A portion of fresh bacterial culture was diluted to about 1 × 10⁵ CFU/ml bacterium cells in 10 ml fresh BHI broth and grown at 37°C with 25-rpm shaking in a photo-recording incubator (TN-2612; Advantec, Tokyo, Japan). The OD was monitored automatically every 2 min, and cells were grown to an OD₆₀₀ of 1.0. For growth curve and doubling time determination, OD versus time in the exponential growth phase of each strain was plotted. Doubling times were then calculated as follows: doubling time = [(t₂ − t₁) × log 2]/[log OD₆₀₀ at t₂ − log OD₆₀₀ at t₁].

Time-kill assay.

Time-kill methodology was used to test the strain's susceptibility to daptomycin. A final inoculum of about 1 × 10⁶ CFU/ml bacterium cells was used. Prior to inoculation, each tube of fresh cation-adjusted Mueller-Hinton broth with 50 μg/ml CaCl₂ was supplemented with daptomycin at concentrations of 0.5, 1, 2, 4, and 8 μg/ml. A tube without antibiotic was used as a growth control. The culture tubes were incubated in the TN-2612 incubator at 37°C for growth curve recording. Viability counts were performed at 0, 2, 4, 8, and 24 h by plating serially diluted cultures on MH agar plate. The growth inhibition effect of daptomycin was also evaluated by plotting the bacterial growth curve.

Introduction of chromosomal point mutations by a targeted sequence substitution method.

Plasmid pKOR1 constructed by Bae and Schneewind is an Escherichia coli/S. aureus shuttle vector which has been described as facilitating allelic replacement in S. aureus without the use of antibiotic markers (3). It permits rapid cloning via lambda recombination and ccdB selection in E. coli, while selection of knockout clones without plasmids in S. aureus is accomplished by anhydrotetracycline-mediated induction of pKOR1-encoded secY antisense transcripts, as secY is essential for S. aureus growth (3). To introduce the chromosomal point mutations identified in the drug-resistant mutant into its drug-sensitive parent strain, this method was utilized with some innovation as described previously (16, 38). Briefly, instead of cloning the up- and downstream fragments of a target gene into pKOR1 (as done in knockout experiments [3]), the gene in-frame containing the intended nucleotide sequence was amplified from sequence-donor strain 10*3d1 using primers that contain attB1 (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCT-) and attB2 (5′-GGGGACCACTTTGTACAAGAAAGCTGGG-) sites on the respective up- and downstream sequences. This fragment was then cloned into pKOR1. Following cloning and ccdB selection in E. coli, the constructed plasmid was then introduced into strain N315ΔIP by electroporation using Bio-Rad GenePulser Xcell (Bio-Rad Laboratories, Inc., Hercules, CA) with the setting parameters described previously (31). Overnight culture of plasmid-carrying clones at 43°C selects for single crossover mutants that carry both mutated and corrected nucleotides. Single crossover mutants were then cultured in drug-free broth to facilitate plasmid excision and subjected to anhydotetracycline induction, whereby only nonplasmids carrying mutants could survive. To check for successful introduction of the mutations, resulting mutants were checked for the target sequences via sequencing the corresponding region with forward and reverse primers which were located 50 to 100 bp beyond the primers used in cloning. By using this strategy, the chromosomal point mutations identified in drug-resistant strain 10*3d1 were introduced into drug-sensitive parent strain N315ΔIP.

Determination of whole-cell zeta potential.

Zeta potential is a useful indicator of electronic charge of colloidal suspensions or emulsions (5, 51). The zeta potential was determined using zeta potential analyzer DelsaNano HC (Beckman Coulter, Inc., CA). Suspensions of 10⁵ cells/ml in 10 mM phosphate-buffered saline (PBS), pH 7.4, were applied for triplicate analysis in each experiment. Briefly, the microelectrophoresis chamber was filled with a bacterial suspension, and a voltage difference of 100 V was applied over the chamber. Data represent the means and standard deviations from three independent experiments and are presented by the ratio of change in the zeta potential of resistant strain versus control strain. The zeta potential change ratio (percentage) was calculated as follows: {1 − [(Zt − Zc)/Zc]} × 100, where Zt and Zc represent zeta potential of test and control, respectively.

RNA preparation and microarray analysis.

Bacteria to be tested were grown in 10 ml BHI broth to exponential phase (OD₆₀₀ = 0.6) before harvest. Pellets were then suspended with precooled T₁₀E₁₀ buffer (10 mM Tris-HCl [pH 8.0]), and lysostaphin solution (WAKO, Japan) at 3.0 μg/ml final concentration was added. The suspensions were then incubated at 37°C for 3 min until complete cell lysis occurred. Immediately, 7 ml of acidic phenol (pH 5.2, equilibrated with 20 mM sodium acetate) was added, and 600 μl 3 M sodium acetate was added. The samples were then frozen and thawed 3 times at −80°C and 65°C, respectively. Phenol-chloroform extraction and ethanol precipitation were then carried out. After that, the resulting RNA pellet was subjected to digestion with RNase-free DNase I (Roche, Mannheim, Germany) at 37°C for 30 min. The RNA samples were then purified again with phenol-chloroform extraction and ethanol precipitation. Pellets were then resuspended in 25 μl diethyl pyrocarbonate (DEPC)-treated water. The construction and analysis of the DNA microarray have been described previously (13).

Statistical analysis of data.

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

Microarray data accession number.

Transcriptional profiles of RpoB_621E mutation-related strains may be found under CIBEX accession no. CBX135.

RESULTS AND DISCUSSION

Comparison of the chromosomes of 10*3d1 and N315ΔIP.

To study the mechanism of daptomycin resistance in S. aureus, we have established MRSA strain 10*3d1 with reduced susceptibility to daptomycin (6). The strain was derived from N315ΔIP, a daptomycin- and vancomycin-susceptible laboratory MRSA strain (6, 28, 31) by serial daptomycin selection. The strain unexpectedly showed decreased susceptibility to both daptomycin and vancomycin and expressed some of the VISA characteristics of thickened cell wall and decreased growth rate, as well as the transcription profile partially overlapping that of hetero-VISA (6). Since we noticed cross-resistance of VISA clinical strains to vancomycin and daptomycin in early 2006 (17), the unexpected acquisition of the dual resistance by 10*3d1 prompted us to carry out a whole-genome comparison of 10*3d1 and N315ΔIP and a gene replacement experiment to clarify the genetic basis for the phenomenon. Our strategy was (i) whole-genome sequencing of the 10*3d1 chromosome and identifying SNPs in comparison with its isogenic parent strain N315ΔIP and (ii) introduction of the SNPs into the N315ΔIP chromosome by single-gene replacement to test the effect of the mutations on vancomycin and daptomycin susceptibilities.

We employed two methods for SNP detection: whole-genome sequencing by using a Solexa/Illumina genome analyzer and competitive DNA hybridization using the NimbleGen microarray provided by Roche NimbleGen, Inc. These two methods were complementary to each other in terms of the SNP detection sensitivity and accuracy and genome scale coverage. Table 1 lists the SNPs detected by the above two methods and confirmed by manual validation resequencing. A total of 5 SNPs located in 4 distinct genes were identified, all of which caused amino acid substitution.

TABLE 1.

List of nucleotide differences between N315ΔIP and 10*3d1 chromosomes

SNP no.a	Position on N315 genome	Nucleotide			N315 genome information			Amino acid change (N315→10*3d1)b
		N315	N315ΔIP	10*3d1	ID	Gene	Product
S-1	581481	C	C	A	SA0500	rpoB	RNA polymerase beta chain	A621E
S-2	2065641	G	G	T	SA1826		Hypothetical protein	A14E
S-3	2305476	C	C	T	SA2042	rplV	50S ribosomal protein L22	c
S-4	2307902	C	C	G	SA2047	rplC	50S ribosomal protein L3 (BL3)	G120D
S-5	2307947	C	C	T	SA2047	rplC	50S ribosomal protein L3 (BL3)	G135A

^a“S” stands for single nucleotide polymorphism (SNP).

^bThe number denotes the position of amino acid substitution. The left amino acid is that of N315, and the right that of 10*3d1. A, alanine; E, glutamic acid; G, glycine; D, aspartate.

^cThe initiation codon is changed from ATG to ATA.

Exploring the effect of SNPs of 10*3d1.

To test which SNP is responsible for the raised vancomycin and daptomycin resistance, we introduced the SNPs one by one into the N315ΔIP chromosome. We used a targeted sequence substitution method for the SNP introduction (16, 38), whereby the SNPs were introduced directly on the chromosome without insertion or deletion of any additional sequence or antibiotic selection markers.

(i) The effect of the SA1826 mutation.

Open reading frame (ORF) SA1826 encodes a conserved hypothetical protein with its two domains similar to penicillinase repressor and type III secretion needle MxiH. We first introduced the SA1826 mutation into N315ΔIP, constructing N315ΔIP-SA1826_14E and N315ΔIP-SA1826_14A (Table 2). The latter was a revertant from the recombination intermediate strain regaining the original intact SA1826 gene. Such revertants were obtained in every allelic replacement procedure and used as isogenic control strains throughout this study (Table 2). The constructed mutant strains were then determined for their susceptibilities to daptomycin and vancomycin. Etest determinations found no difference of daptomycin and vancomycin susceptibilities between N315ΔIP-SA1826_14E and N315ΔIP-SA1826_14A (Table 2), indicating that the SA1826 mutation was not responsible for the altered drug resistance phenotype in 10*3d1.

TABLE 2.

SNP type, antibiotic susceptibility profile, and doubling time of S. aureus strains used in this study

Strain	Sequence typea					MIC (mg/liter) by Etestb				Doubling time (min)	Reference for strains
	RpoB621	SA182614	rplV1∼3	RplC120	RplC135	Daptomycin	Vancomycin	Oxacillin	Rifampin
N315ΔIP	A	A	ATG	G	G	0.38	0.5	48 (256)c	0.006	35.73	6
10*3d1	E	E	ATA	D	A	2	3	>256	0.006	61.17	6
N315ΔIP-rpoB621A	A	A	ATG	G	G	0.25	0.25	>256	0.006	38.22	This study
N315ΔIP-rpoB621E	E	A	ATG	G	G	1.5	3	>256	0.006	39.31	This study
N315ΔIP-SA182614A	A	A	ATG	G	G	0.25	0.5	>256	0.006	36.75	This study
N315ΔIP-SA182614E	A	E	ATG	G	G	0.25	0.5	>256	0.006	36.83	This study
10*3d1-SA182614A	E	A	ATA	D	A	1.5	3	>256	0.006	66.46	This study
10*3d1-SA182614E	E	E	ATA	D	A	1.5	3	>256	0.006	70.54	This study
10*3d1-SA182614A-rpoB621A	A	A	ATA	D	A	0.25	0.75	>256	0.006	59.61	This study
10*3d1-SA182614A-rpoB621E	E	A	ATA	D	A	1.5	3	>256	0.006	72.12	This study

^aSubstituted amino acids are given, except for rplV, for which the mutated initiation codon (first three nucleotides) is given. A, alanine; E, glutamic acid; G, glycine; D, aspartate.

^bEtest was performed with Mueller-Hinton agar, and the results were read after 24 h of incubation at 37°C.

^cMIC for the minority subpopulations is given in parentheses.

(ii) The effect of the rpoB mutation.

Next, we introduced the mutated rpoB into N315ΔIP, obtaining strains N315ΔIP-rpoB_621E and N315ΔIP-rpoB_621A (Table 2). To validate if the rpoB(A621E) mutation is a unique genetic alteration during the N315ΔIP-rpoB_621E construction, we performed the whole-genome sequence comparison between N315ΔIP-rpoB_621E and N315ΔIP using the NimbleGen microarray. We found no detectable mutation except for rpoB(A621E) in the N315ΔIP-rpoB_621E in comparison to N315ΔIP (data not shown), supporting the hypothesis that phenotype changes of N315ΔIP- rpoB_621E were caused by the rpoB(A621E) mutation. rpoB encodes the β subunit of the bacterial RNA polymerase. RpoB is a well-known target of action of rifampin, and its mutation is frequently associated with resistance to rifampin. The position of the mutation in 10*3d1, however, was found outside the previously identified rifampin resistance regions of RpoB known to confer rifampin resistance (2, 36). In fact, the mutation was not accompanied by rifampin resistance in either 10*3d1 or N315ΔIP-rpoB_621E (Table 2). However, susceptibility tests showed that the rpoB(A621E) mutation caused a simultaneous increase of daptomycin and vancomycin MICs; the increment of daptomycin MIC was from 0.25 mg/liter to 1.5 mg/liter, and that of the vancomycin MIC was from 0.25 mg/liter to 3.0 mg/liter (Table 2). The extent of MIC changes were comparable to that between N315ΔIP and 10*3d1; N315ΔIP had a daptomycin MIC of 0.38 mg/liter and a vancomycin MIC of 0.5 mg/liter, whereas it was 2.0 mg/liter and 3.0 mg/liter, respectively, for 10*3d1 (Table 2 and Fig. 1).

FIG. 1.

Daptomycin (A) and vancomycin (B) MICs were determined by Etest for N315ΔIP-rpoB_621E and its relevant strains. Note that N315ΔIP-rpoB_621E has reduced daptomycin and vancomycin susceptibilities similar to those of 10*3d1.

(iii) The effect of rplV and rplC mutations.

All our trials failed to introduce the mutations identified in rplV and rplC of 10*3d1 into N315ΔIP with unknown reason. To evaluate the effect of the two mutations in the ribosomal protein genes collectively, we changed the strategy of experiments. Instead of introducing mutations into N315ΔIP, we sequentially cured the SA1826 and rpoB mutations of 10*3d1, obtaining the strain 10*3d1-SA1862_14A-rpoB_621A which harbors mutations only in the two rpl genes compared to N315ΔIP. The obtained mutant strain retained the slow growth phenotype of 10*3d1. Thus, it appears that the rplV and rplC mutations were the cause for the slow growth of 10*3d1. As shown in Table 2, restoration of the rpoB mutation but not the SA1862 mutation of 10*3d1 led to a significant decrease of both daptomycin and vancomycin resistance to the susceptible levels of N315ΔIP, while there was no susceptibility change in the control strain 10*3d1-SA1862_14A-rpoB_621E. This clearly indicated that the mutations of rplV and rplC alone had no significant effect on vancomycin and daptomycin resistance. This experiment also confirmed that the rpoB(A621E) mutation alone is responsible for the dual heteroresistance to vancomycin and daptomycin.

Slow growth has been repeatedly reported for VISA strains and is considered one of the factors associated with VISA phenotype expression in S. aureus (10, 14, 19, 46). In our previous study, 10*3d1 showed a significantly decreased growth rate compared to its daptomycin- and vancomycin-susceptible parent strain N315ΔIP (6). We considered that the slow growth was associated with the expression of the resistance phenotype (6). However, as shown in Table 2 and Fig. 2, the rpoB-restored strain 10*3d1-SA1862_14A-rpoB_621A retained the slow growth phenotype of 10*3d1 after having lost the vancomycin and daptomycin dual resistance. Thus, the slow growth rate itself did not influence the MIC values in any significant way.

FIG. 2.

Growth curve characteristics of strains exhibiting two groups of different growth rates. Note that strains having rplC and rplV mutations (see Table 2) showed significantly slower growth than those without rplC and rplV mutations.

Although the slow growth phenotype did not raise MIC values, it may affect the bactericidal activities of daptomycin and vancomycin. To test this possibility, we evaluated the growth inhibition and time-kill effect of daptomycin against N315ΔIP and 10*3d1-SA1862_14A-rpoB_621A. The two strains had the same susceptibility to bactericidal effects of daptomycin irrespective of the difference in the growth rates (Fig. 2 and Fig. 3). Thus, the slow growth rate was not a major contributor in the reduced susceptibility to daptomycin and vancomycin of 10*3d1.

FIG. 3.

Growth inhibition (A) and time-killing effect (B) of daptomycin against strains N315ΔIP and 10*3d1-SA1826_14A-rpoB_621A. Killing activity is presented as log₁₀ CFU/ml.

The rpoB(A621E) mutation confers daptomycin and vancomycin heteroresistance to N315ΔIP.

The introduction of the rpoB(A621E) mutation in N315ΔIP raised MICs for daptomycin and vancomycin as shown in Table 2. We previously showed that the types of daptomycin and vancomycin resistance expressed by strain 10*3d1 were heterogenous (6). In order to determine if the resistance pattern conferred by the rpoB(A621E) mutation was heterogeneous or not, we analyzed the resistant subpopulation profile of the rpoB(A621E)-carrying strain against vancomycin and daptomycin. Figure 4 shows the results. Population analysis showed that N315ΔIP-rpoB_621E, compared to N315ΔIP-rpoB_621A and parent strain N315ΔIP, had a significant increase in resistant subpopulations for both daptomycin and vancomycin, and the pattern of the population curves were similar to that of 10*3d1. Therefore, the rpoB(A621E) mutation alone could cause the heterogenous resistance phenotype of 10*3d1 against daptomycin and vancomycin.

FIG. 4.

Analysis of daptomycin- and vancomycin-resistant subpopulations of N315ΔIP-rpoB_621E and its isogenic strains. The number of colonies on plates containing various concentrations of daptomycin (A) and vancomycin (B) was counted after 48 h of incubation at 37°C.

Friedman et al. (18) recently reported a finding of two strains with RpoB mutations, Ile₉₅₃→Ser₉₅₃ and Ala₁₀₈₆→Val₁₀₈₆, in a series of laboratory-derived S. aureus strains with decreased susceptibility to daptomycin. However, the two strains with the rpoB mutation also carried additional mutations in mprF and yycG. Furthermore, since they did not find the rpoB mutations in a collection of clinical isolates with decreased susceptibility to daptomycin, the role of the rpoB mutation was not explored any further. In this study, by using a gene replacement method, we unequivocally proved that an rpoB mutation can be a cause of dual heteroresistance to vancomycin and daptomycin.

rpoB(A621E) mutation causes thickening of the cell wall.

How does the rpoB mutation cause reduced susceptibility to daptomycin and vancomycin? The most significant feature of the VISA phenotype is cell wall thickness (11, 14, 22). The feature is directly associated with the “peptidoglycan-clogging mechanism” that prevents the passage of vancomycin across the thickened peptidoglycan layers (12). Unlike vancomycin, daptomycin does not have binding sites in the peptidoglycan, so the clogging effect may not apply to daptomycin resistance. However, daptomycin has a relatively bigger molecular size than vancomycin (molecular weight, 1,620.67 versus 1,485.7). Moreover, daptomycin forms oligomers with calcium ions, which dissociates only when the oligomers come close to the bacterial membrane (45). Therefore, the thickened peptidoglycan layer may serve as an obstacle for the access of daptomycin to the cell membrane. In agreement with our hypothesis (17), 10*3d1 showed a significantly thickened cell wall compared to that of its drug-susceptible parent strain N315ΔIP (6). Figure 5 shows the electronmicrograph showing that the strain N315ΔIP-rpoB_621E harboring the rpoB(A621E) mutation had a much more thickened cell wall than that of the control strain N315ΔIP-rpoB_621A. The extent of the cell wall thickening was 1.60 times, which was comparable to the case of 10*3d1 (whose cell wall was thicker by 63% than that of its parent strain N315ΔIP) (Fig. 5). Therefore, the rpoB mutation is considered to contribute to the heteroresistance to vancomycin and daptomycin by thickening the cell wall.

FIG. 5.

Transmission electron microscopic observation of N315ΔIP-rpoB_621E and its isogenic strains. Transmission electron microscopy was carried out on N315ΔIP-rpoB_621E (D), which was generated from N315ΔIP (A) by substitution of its rpoB gene with that of 10*3d1 (B), and experimental control strain N315ΔIP-rpoB_621A (C), which retained the intact rpoB gene. Magnification is ×30,000. The values given under each picture are the means and standard deviations of each strain's cell wall thickness in nanometers. Note that the cell walls of N315ΔIP-rpoB_621E and 10*3d1 were much thicker than those of the control and parent strains.

Decreased negative cell surface charge associated with the rpoB mutation.

Decreased negative cell surface charge has been reported as one of the contributing factors for the increased daptomycin and vancomycin resistance in S. aureus (25, 33, 52), based on the charge repulsion mechanism assumption. Daptomycin and vancomycin are relatively positively charged at a neutral pH (4, 8, 20, 24). Therefore, if a bacterial cell generates a less negative cell surface, it would reduce the binding of vancomycin and daptomycin molecules to the cell (25, 33). In the case of daptomycin, however, there has been a discrepant observation that the increased negative cell surface charge is observed in in vitro-selected daptomycin-nonsusceptible mutants (34). Therefore, the charge repulsion mechanism might not be generally applicable or not in its simplest form. In any case, to test the contribution of the cell surface charge in our in vitro-selected strains, we determined the whole-cell surface charge of the constructed mutants with some control strains. The cell surface charge was measured by determining the bacterial zeta potential, which is an indicator of the electronic charge of colloidal suspensions (5, 51). A total of two groups of strains were used for the zeta potential determination. The first group consisted of N315ΔIP-rpoB_621E, N315ΔIP-rpoB_621A, 10*3d1, and N315ΔIP. The second group consisted of clinical VISA strains Mu50 and MI and their derivative substrains, passage-derived vancomycin-susceptible strains Mu50-P35 and MI-P84, and Mu50-PR and MI-PR, which were derived from the vancomycin-susceptible derivative strains by selection with 4 mg/liter of vancomycin (14). Their vancomycin and daptomycin MICs were as follows: the vancomycin MICs for Mu50, Mu50-P35, and Mu50-PR were 6, 2, and 6 mg/liter, respectively, and those for MI, MI-P84, and MI-PR were 7, 1, and 6 mg/liter, respectively. The daptomycin MICs for them were 2, 0.5, and 2, and 2, 0.5, and 2 mg/liter, respectively (14, 17). All the vancomycin-susceptible strains used in this experiment were susceptible to daptomycin, while the vancomycin-resistant strains were also resistant to daptomycin (17). Figure 6 shows the summary of decreased negative charge calculations of drug-resistant strains against their drug-susceptible isogenic control strains. Zeta potential values for control strain N315ΔIP, Mu50-P35, and MI-P84 were −28 ± 1.4, −30 ± 1.6, and −35 ± 1.5 mV, respectively. We calculated the ratio (percentage) of negative charge decrement in resistant strains to their respective daptomycin- and vancomycin-susceptible counterparts. It was noted that all the daptomycin- and vancomycin-resistant strains had a decreased negative charge compared to that of their susceptible controls (P < 0.01), and there was a positive correlation between the reduction of negative charge and the reduced susceptibility to daptomycin and vancomycin (Fig. 6). The results indicated that the decreased negative surface charge might contribute to the daptomycin and vancomycin resistance in 10*3d1 as well as in clinical VISA strains Mu50 and MI.

FIG. 6.

Decrease of whole-cell negative charge in daptomycin- and vancomycin-resistant strains. Whole-cell zeta potential, an indicator of the cell electronic charge, was determined at pH 7.4, and the zeta potential change ratio (percentage) (box bar) of the test strain (filled box) to its respective control (open box) was calculated as follows: {1 − [(Zt − Zc)/Zc] × 100}, where Zt and Zc represent the Zeta potential of tthe est and control, respectively. Values are means ± standard deviations.

Rifampin resistance associated with rpoB point mutations is a well-recognized phenomenon in S. aureus (2, 39, 50). The mutations are clustered predominantly in a conserved region of RpoB, designated rifampin resistance-determining region (RRDR), which spans amino acid residues ∼463 to 550. Mutations in the region result in a lessening of the hydrophobic interaction between RpoB and rifampin and a decrease in the binding of rifampin to RNA polymerase holoenzyme (40). The rpoB mutation identified in 10*3d1 is located outside the RRDR, and this coincides with the observation that the rpoB(A621E) mutation does not affect rifampin susceptibility of the strain (Table 2). However, the effect of this mutation on RNA polymerase activity and the detailed mechanism causing heteroresistance to both daptomycin and vancomycin remains to be elucidated.

rpoB mutations are also found in clinical VISA strain Mu50 (38) and JH9 (37). Our finding of the contribution of an rpoB mutation to daptomycin and vancomycin dual resistance in this study indicates that the rpoB mutations also contribute to the VISA phenotype of clinical MRSA strains.

Microarray transcriptional analysis.

In our previous attempt to understand the global regulatory mechanism for the expression of daptomycin resistance in 10*3d1, we performed microarray analysis on 10*3d1 and its parent strain N315ΔIP. The results, including raw signal data, were disclosed at the CIBEX site (http://cibex.nig.ac.jp/index.jsp) under the CIBEX accession number CBX67. By comparing the transcriptional profiles of 10*3d1and N315ΔIP, we identified dozens of specifically up- and downregulated genes in 10*3d1, including those involved in cell wall and cell membrane metabolism (6). However, since 10*3d1 had 5 single point mutations in 4 distinct genes compared to N315ΔIP (Table 1) and all of these mutations might have some influence on the transcriptional profiles of 10*3d1, the alterations of transcriptional profiles of 10*3d1 might not have directly reflected the effect of the rpoB(A621E) mutation. To better delineate the effect of the rpoB(A621E) mutation causing dual resistance to daptomycin and vancomycin, we performed microarray analysis of three combinations of strains: N315ΔIP-rpoB_621E versus N315ΔIP, 10*3d1-SA1826_14A-rpoB_621E versus 10*3d1-SA1826_14A-rpoB_621A, and 10*3d1 versus N315ΔIP (Table 2). There is only one mutation difference [rpoB(A621E)] within each of the first two pairs. The last pair represents the cumulative effects of the five mutations (6).

All transcriptional profiles and comparison data are available on the CIBEX site under the accession number CBX135, and Table 3 shows the extracted representative data. The most significant alterations apparently caused by the rpoB(A621E) mutation were (i) repression of the genes in the purine and pyrimidine metabolic pathways; (ii) repression of urease genes; (iii) enhancement of the biosynthetic pathway genes for riboflavin (vitamin B2), phylloquinone (vitamin K1), and menaquinone (vitamin K2); and (iv) enhancement of several genes involved in the cell envelope metabolism and structure (Table 3). It is not clear yet how the rpoB(A621E)-mediated altered gene transcription profile mentioned above is correlated with the phenotypic changes such as cell wall thickening, decreased cell surface negative charge, and reduced susceptibility to vancomycin and daptomycin. The products of the dlt operon genes are reported to have a function of attaching positively charged d-alanine residues onto the negatively charged phosphate groups in the backbone of teichoic acids in S. aureus (9, 30, 41). An enhanced expression of the dlt operon associated with an increased cell surface positive charge was observed in daptomycin-resistant MRSA (34). The singly introduced rpoB(A621E) mutation in N315ΔIP did enhance the expression of dlt genes. However, curiously, the enhancement was not apparent in 10*3d1 (Table 3). The same was true with the sigB gene, whose overexpression is reported to thicken the cell wall (35). sigB was upregulated by the rpoB(A621E) mutation in N315ΔIP, but the effect was not apparent in the parent strain 10*3d1 (Table 3). Besides these genes, however, there were quite a number of genes involved in cell envelope metabolism enlisted as the genes up- or downregulated by the rpoB(A621E) mutation (Table 3). The altered expression of such genes as sle1, atl, stcA, SA1898 (sceD homolog), SA2097 (ssaA homolog), capABC, SA0126, SA0127, SA2284, SA2285, mgrA, fmtB, and sarUT may modify the cell envelope physiology and structure in a way that evades the attack of vancomycin and daptomycin.

TABLE 3.

Transcriptional profiles of RpoB_A621E mutation-related strains

ORF ID	Gene	Products	Ratio of signal intensitya
			N315ΔIP-rpoB621E/N315ΔIP	10*3d1-SA182614A-rpoB621E/10*3d1-SA182614A-rpoB621A	10*3d1/ N315ΔIP
Purine metabolism
SA0016	purA	Adenylosuccinate synthase	0.36	0.64	0.76
SA0373	xprT	Xanthine phosphoribosyltransferase	0.64	0.40	0.12
SA0374	pbuX	Xanthine permease	0.55	0.42	0.10
SA0375	guaB	Inositol-monophosphate dehydrogenase	0.90	0.50	0.18
SA0376	guaA	GMP synthase (glutamine hydrolyzing)	0.88	0.65	0.23
SA0468		Hypoxanthine-guanine phosphoribosyltransferase homologue	1.51	1.18	1.53
SA0916		Similar to phosphoribosylaminoimidazole carboxylase PurE	0.84	0.80	0.72
SA0917	purK	Phosphoribosylaminoimidazole carboxylase carbon dioxide-fixation chain PurK homolog	0.93	0.54	0.43
SA0918	purC	Phosphoribosylaminoimidazolesuccinocarboxamide synthetase homolog	1.00	0.57	0.43
SA0919		Conserved hypothetical protein	0.96	0.69	0.45
SA0920	purQ	Phosphoribosylformylglycinamidine synthase I PurQ	1.10	0.52	0.32
SA0922	purF	Phosphoribosylpyrophosphate amidotransferase PurF	0.81	0.52	0.26
SA0923	purM	Phosphoribosylformylglycinamidine cyclo-ligase PurM	0.77	0.54	0.29
SA0924	purN	Phosphoribosylglycinamide formyltransferase	0.79	0.56	0.30
SA0925	purH	Bifunctional purine biosynthesis protein PurH	0.75	0.59	0.33
SA0926	purD	Phosphoribosylamine-glycine ligase PurD	0.61	0.55	0.36
SA1052	gmk	Guanylate kinase homolog	0.81	0.91	0.81
SA1172		Similar to GMP reductase	0.34	0.37	0.57
SA1173		Conserved hypothetical protein	0.28	0.33	0.58
SA1461	apt	Adenine phosphoribosyl transferase	0.94	1.30	1.41
Pyrimidine metabolism
SA0440	tmk	Thymidylate kinase homolog	0.64	0.60	0.74
SA0719	trxB	Thioredoxine reductase	0.45	0.46	0.68
SA1041	pyrR	Pyrimidine operon repressor chain A	0.17	0.35	0.82
SA1042	pyrP	Uracil permease	0.14	0.21	0.41
SA1043	pyrB	Aspartate transcarbamoylase chain A	0.13	0.22	0.41
SA1044	pyrC	Dihydroorotase	0.15	0.25	0.39
SA1045	pyrAA	Carbamoyl-phosphate synthase small chain	0.18	0.28	0.40
SA1046	pyrAB	Carbamoyl-phosphate synthase large chain	0.18	0.32	0.38
SA1047	pyrF	Orotidine-5-phosphate decarboxylase	0.22	0.40	0.40
SA1048	pyrE	Orotate phosphoribosyltransferase	0.20	0.43	0.41
SA1049		Hypothetical protein	0.22	0.42	0.47
SA1439	udk	Uridine kinase	0.68	0.74	1.32
SA2375		Similar to dihydroorotate dehydrogenase	0.62	0.75	0.56
Arginine metabolism and urea cycle
SA0175		Conserved hypothetical protein	0.34	0.42	0.24
SA0176		Similar to N-acetylglutamate 5-phosphotransferase	0.63	0.76	0.64
SA0177	argJ	Arginine biosynthesis bifunctional protein homolog	0.61	0.76	0.64
SA0178	argC	N-acetylglutamate gamma-semialdehyde dehydrogenase	0.59	0.73	0.65
SA0179		Ornithine aminotransferase	0.64	0.80	0.82
SA0821	argH	Argininosuccinate lyase	0.44	0.54	0.27
SA0822	argG	Argininosuccinate synthase	0.48	0.59	0.29
SA2082	ureA	Urease gamma subunit	0.37	0.26	0.28
SA2083	ureB	Urease beta subunit	0.23	0.19	0.24
SA2084	ureC	Urease alpha subunit	0.30	0.21	0.24
SA2085	ureE	Urease accessory protein UreE	0.26	0.19	0.24
SA2086	ureF	Urease accessory protein UreF	0.30	0.27	0.29
SA2087	ureG	Urease accessory protein UreG	0.44	0.33	0.32
SA2088	ureD	Urease accessory protein UreD	0.39	0.30	0.32
SA2428	arcA	Arginine deiminase	2.82	2.75	1.45
SA2429		Similar to arginine repressor	3.62	6.20	2.27
Galactose metabolism
SA0123		Similar to UDP-glucose 4-epimerase (gale-1)	0.34	0.45	0.41
SA0236		Similar to PTS fructose-specific enzyme IIBC component	0.68	0.83	0.82
SA0237		Similar to PTS, galactitol-specific IIB component	0.65	0.71	0.93
SA0238	gatC	Probable PTS galactitol-specific enzyme IIC component	0.64	0.81	1.28
SA1521	pfk	6-Phosphofructokinase	0.77	0.60	0.53
SA1991	lacG	6-Phospho-beta-galactosidase	0.33	0.73	1.75
SA1992	lacE	PTS system, lactose-specific IIBC component	0.35	0.63	1.40
SA1993	lacF	PTS system, lactose-specific IIA component	0.32	0.62	1.38
SA1994	lacD	Tagatose 1,6-diphosphate aldolase	0.39	0.59	1.25
SA1995	lacC	Tagatose-6-phosphate kinase	0.38	0.56	1.22
SA1996	lacB	Galactose-6-phosphate isomerase LacB subunit	0.42	0.53	1.17
SA1997	lacA	Galactose-6-phosphate isomerase LacA subunit	0.42	0.54	1.10
SA2288	gtaB	UTP-glucose-1-phosphate uridyltransferase	0.63	0.69	0.53
Nitrogen metabolism
SA0008	hutH	Histidine ammonia-lyase	1.27	1.50	1.25
SA0171	fdh	NAD-dependent formate dehydrogenase	0.81	0.78	1.18
SA0774		ABC transporter ATP-binding protein homolog	2.85	2.95	3.28
SA0775		FeS assembly protein SufD	2.84	3.05	3.71
SA0776		Aminotransferase NifS homolog	2.76	3.12	3.41
SA0777		Similar to nitrogen fixation protein NifU	2.58	3.15	3.44
SA0778	sufB	FeS assembly protein SufB	2.27	2.91	3.23
SA0779		Hypothetical protein	1.91	2.11	2.30
SA0819	gudB	NAD-specific glutamate dehydrogenase	1.78	2.14	1.75
SA1241		Similar to nitric-oxide reductase	2.37	1.42	2.29
SA1310	ansA	Probable l-asparaginase	1.31	1.36	1.18
dlt operon
SA0792		Hypothetical protein	1.56	2.12	0.98
SA0793	dltA	d-alanine-d-alanyl carrier protein ligase	1.61	2.27	0.88
SA0794	dltB	DltB membrane protein	1.60	2.54	0.81
SA0795	dltC	d-alanyl carrier protein	1.55	2.60	0.83
SA0796	dltD	Poly(glycerophosphate chain) d-alanine transfer protein	1.44	2.52	0.84
sigB operon
SA1869	sigB	Sigma factor B	2.96	2.05	1.19
SA1870	rsbW	Anti-sigmaB factor	3.15	1.90	1.14
SA1871	rsbV	Anti-sigmaB factor antagonist	2.64	1.74	1.22
SA1872	rsbU	SigmaB regulation protein RsbU	2.08	1.62	1.22
Riboflavin (vitamin B2) metabolism
SA1586	ribH	6,7-Dimethyl-8-ribityllumazine synthase	1.75	1.88	2.03
SA1588	ribB	Riboflavin synthase alpha chain	1.92	1.94	2.20
SA1589	ribD	Riboflavin-specific deaminase	2.03	2.04	2.08
SA1587	ribA	Riboflavin biosynthesis protein	2.18	1.99	2.16
Phylloquinone (vitamin K1) and menaquinone (vitamin K2) metabolism
SA0898	menB	Naphthoate synthase	1.08	1.30	1.33
SA1303	gerCB	Menaquinone biosynthesis methyltransferase	1.14	1.00	1.16
SA0895		Similar to menaquinone-specific isochorismate synthase	1.79	1.44	2.66
SA1616		Hypothetical protein	1.85	2.12	2.37
SA0897		Similar to prolyl aminopeptidase (EC 3.4.11.5)	1.91	1.74	1.91
SA0896	menD	Menaquinone biosynthesis protein	1.93	1.65	2.33
SA1615	menE	O-succinylbenzoic acid-coenzyme A (CoA) ligase	2.12	1.76	2.18
SA1614	menC	O-succinylbenzoic acid (OSB) synthetase	2.14	1.65	1.78
Citrate cycle (TCA cycle)
SA0944	phdB	Pyruvate dehydrogenase E1 component beta subunit	1.25	0.97	0.90
SA0945	pdhC	Dihydrolipoamide S-acetyltransferase component of pyruvate dehydrogenase complex E2	1.28	1.02	0.97
SA0946	pdhD	Dihydrolipoamide dehydrogenase component of pyruvate dehydrogenase E3	0.90	1.04	1.11
SA0994	sdhC	Succinate dehydrogenase cytochrome b-558	1.72	1.13	1.50
SA0995	sdhA	Succinate dehydrogenase flavoprotein subunit	1.65	1.31	1.50
SA0996	sdhB	Succinate dehydrogenase iron-sulfur protein subunit	1.05	1.58	1.88
SA1088	sucC	Succinyl-CoA synthetase (beta subunit)	1.46	1.54	1.59
SA1089	sucD	succinyl-CoA synthetase (alpha subunit)	1.11	1.52	1.62
SA1131		Similar to 2-oxoacid ferredoxin oxidoreductase, alpha subunit	1.49	1.25	1.77
SA1132		Similar to 2-oxoacid ferredoxin oxidoreductase, beta subunit	1.39	1.37	1.74
SA1184	citB	Aconitate hydratase	1.07	1.11	1.59
SA1244	odhB	Dihydrolipoamide succinyltransferase	1.66	1.74	1.44
SA1245	odhA	2-Oxoglutarate dehydrogenase E1	1.36	1.45	1.24
SA1349		Dihydrolipoamide dehydrogenase	1.54	0.93	1.16
SA1517	citC	Isocitrate dehyrogenase	1.17	1.27	2.02
SA1518	citZ	Citrate synthase II	1.03	1.18	2.11
SA1609	pckA	Phosphoenolpyruvate carboxykinase (ATP)	1.27	0.82	1.10
Cell envelope and cellular processes
SA0111	sirA	Lipoprotein	2.32	2.19	0.65
SA0126		Similar to capsular polysaccharide synthesis protein 14H	0.30	0.44	0.44
SA0127		Similar to capsular polysaccharide synthesis protein 14H	0.36	0.48	0.43
SA0265	lytM	Peptidoglycan hydrolase	2.15	2.06	1.72
SA0295		Similar to outer membrane protein precursor	3.29	3.73	1.03
SA0423	sle1	Hypothetical protein, similar to autolysin (N-acetylmuramoyl-l-alanine amidase)	2.76	2.75	2.89
SA0457	gcaD	UDP-N-acetylglucosamine pyrophosphorylase	2.00	1.59	1.25
SA0523		Similar to poly(glycerol-phosphate) alpha-glucosyltransferase (teichoic acid biosynthesis)	3.25	4.10	1.72
SA0620	stcA	Secretory antigen SsaA homolog	1.81	3.16	3.05
SA0905	atl	Autolysin (N-acetylmuramyl-l-alanine amidase and endo-β-N-acetylglucosaminidase)	3.28	1.89	1.88
SA1023	ftsL	Cell division protein	0.49	0.45	0.70
SA1026	murD	UDP-N-acetylmuramoylalanine-d-glutamate ligase	2.16	1.37	1.29
SA1027	div1b	Cell division protein, FtsQ homolog	2.28	1.32	1.19
SA1212	opp-2D	Oligopeptide transport ATPase	1.90	2.31	2.01
SA1267	ebhA	Similar to streptococcal adhesin Emb	0.30	0.12	0.64
SA1898		Similar to SceD precursor (lytic transglycosylase)	2.35	2.24	4.50
SA2097	ssaA homolog	Similar to secretory antigen precursor SsaA	8.16	8.04	6.00
SA2356	isaA	Immunodominant antigen A	0.38	0.60	1.22
SA2442		Preprotein translocase secA homolog	2.57	1.88	1.23
SA2446		Similar to preprotein translocase secY	2.01	2.08	2.10
SA2455		Capsular polysaccharide biosynthesis, capC	0.37	0.41	0.44
SA2456		Capsular polysaccharide biosynthesis, capB	0.38	0.38	0.40
SA2457		Capsular polysaccharide biosynthesis, capA	0.40	0.36	0.41
Other functions
SA0128	sodM	Superoxide dismutase	1.51	2.04	2.12
SA0222	coa	Staphylocoagulase precursor	0.46	0.64	0.14
SA0309	geh	Glycerol ester hydrolase; lipase	2.29	2.03	0.83
SA0519	sdrC	Ser-Asp rich fibrinogen-binding	3.35	3.04	1.15
SA0520	sdrD	Ser-Asp rich fibrinogen-binding	6.25	9.58	0.58
SA0521	sdrE	pfam04650, YSIRK-type signal peptide	3.21	3.30	0.52
SA0641	mgrA	Transcriptional regulator, MarR family	1.81	4.00	1.74
SA0723	clpP	ATP-dependent Clp protease proteolytic subunit homolog	0.46	0.71	1.00
SA0743	graB	Similar to staphylocoagulase precursor	0.23	0.22	0.31
SA0746		Staphylococcal nuclease	2.35	4.54	0.74
SA1145		Hfq, an abundant, ubiquitous RNA-binding protein	0.30	0.41	0.70
SA1396	bex	GTP-binding protein Era homolog	2.02	1.39	1.48
SA1452	csbD	SigmaB-controlled gene product	3.04	1.82	1.60
SA1549		Similar to serine proteinase Do, heat shock protein HtrA	0.48	0.65	0.56
SA1559		Similar to smooth muscle caldesmon	2.79	1.44	0.93
SA1577	sasC	Hypothetical protein, similar to FmtB protein	0.41	0.38	0.83
SA1617		Similar to latent nuclear antigen	4.69	3.16	1.90
SA1709		Similar to ferritin	1.53	2.24	2.84
SA1941	dps	General stress protein 20U	1.76	3.99	4.28
SA1964	fmtB	FmtB protein	0.30	0.25	0.47
SA1984	asp23	Alkaline shock protein 23, ASP23	3.26	2.00	1.25
SA2284		Similar to accumulation-associated protein	0.05	0.45	0.02
SA2285		Similar to accumulation-associated protein AAP	0.05	0.50	0.02
SA2286	sarT	Helix-turn-helix multiple antibiotic resistance protein	0.45	1.03	0.36
SA2287	sarU	Helix-turn-helix multiple antibiotic resistance protein	0.07	0.92	0.08
SA2290	fnbB	Fibronectin-binding protein homolog	5.69	2.93	1.67
SA2423	clfB	Clumping factor B	4.07	3.79	1.08
SA2430	aur	Zinc metalloproteinase aureolysin	6.02	7.32	1.90
SA2431	isaB	Immunodominant antigen B	2.28	4.16	1.07

^aSignal intensity ratios of strains with RpoB_621E versus those with RpoB_621A.

In conclusion, although detailed studies are needed to clarify the role of rpoB with an A621E mutation, the data presented in this study clearly show that the rpoB mutation does confer S. aureus cells heteroresistant to both vancomycin and daptomycin. Besides the rifampin resistance phenotype, rpoB mutation may have an important role in the multidrug-resistance phenotype of MRSA.