Causal Role of Single Nucleotide Polymorphisms within the mprF Gene of Staphylococcus aureus in Daptomycin Resistance

Abstract

Single nucleotide polymorphisms (SNPs) within the mprF open reading frame (ORF) have been commonly observed in daptomycin-resistant (DAP^r) Staphylococcus aureus strains. Such SNPs are usually associated with a gain-in-function phenotype, in terms of either increased synthesis or enhanced translocation (flipping) of lysyl-phosphatidylglycerol (L-PG). However, it is unclear if such mprF SNPs are causal in DAP^r strains or are merely a biomarker for this phenotype. In this study, we used an isogenic set of S. aureus strains: (i) Newman, (ii) its isogenic ΔmprF mutant, and (iii) several in trans plasmid complementation constructs, expressing either a wild-type or point-mutated form of the mprF ORF cloned from two isogenic DAP-susceptible (DAP^s)-DAP^r strain pairs (616-701 and MRSA11/11-REF2145). Complementation of the ΔmprF strain with singly point-mutated mprF genes (mprF_S295L or mprF_T345A) revealed that (i) individual and distinct point mutations within the mprF ORF can recapitulate phenotypes observed in donor strains (i.e., changes in DAP MICs, positive surface charge, and cell membrane phospholipid profiles) and (ii) these gain-in-function SNPs (i.e., enhanced L-PG synthesis) likely promote reduced DAP binding to S. aureus by a charge repulsion mechanism. Thus, for these two DAP^r strains, the defined mprF SNPs appear to be causally related to this phenotype.

INTRODUCTION

Daptomycin (DAP) is a cyclic lipopeptide antibiotic which is active against a wide range of Gram-positive organisms, including methicillin-resistant Staphylococcus aureus (MRSA), vancomycin (VAN)-intermediate S. aureus (VISA), and VAN-resistant S. aureus (VRSA) strains (1–4). Multiple studies have shown that there has been no substantive DAP “MIC creep,” even after widespread clinical use of this agent over the last decade (5–7). However, increasing reports have described the evolution of DAP resistance (DAP^r) in vivo in association with DAP clinical treatment failures in infections with Staphylococcus aureus, especially MRSA (8–14).

MprF is a lysyl-phosphotidylglycerol (L-PG) synthase which transfers positively charged lysine molecules from lysyl-tRNA and adds them to phosphotidylglycerol (PG) within the S. aureus cell membrane (CM). In addition, MprF functions as an inner-to-outer CM translocase for L-PG (15–17). L-PG is a positively charged phospholipid (PL) which is unique to the S. aureus CM, accounting for ∼10 to 30% of its total CM PL content (18–20). The amount and asymmetry of L-PG in the CM likely contribute to the relative positive charge properties of the S. aureus cell surface (16, 17, 19). Hence, mprF deletion mutants exhibit an absence of L-PG in their CM and reduced cell surface positive charge, resulting in increased susceptibility to a number of cationic antimicrobial peptides (CAMPs) (10, 18–20). Although DAP is an anionic molecule, its antimicrobial activity is absolutely dependent on it undergoing extensive complexing with calcium (19). This event renders DAP a de facto CAMP (1, 4).

Over the past several years, a number of laboratories, including ours, have linked the presence of single nucleotide polymorphisms (SNPs) within the mprF locus with the DAP^r phenotype (10, 21–26). These SNPs within the mprF open reading frame (ORF) have been observed in both clinically derived and in vitro-passage-derived DAP^r S. aureus strains (10, 12, 21, 22, 25–28). Such SNPs are usually associated with a gain-in-function phenotype, in terms of either increased synthesis or enhanced translocation of L-PG (10, 22, 23, 25, 26, 29). These findings led us to investigate whether the mprF SNPs identified in DAP^r strains directly related to, and were solely responsible for, the DAP^r phenotype in vitro.

In the current study, we used a well-defined and previously well-characterized DAP-susceptible (DAP^s) and methicillin-susceptible S. aureus (Newman) strain, its isogenic mprF knockout strain, and several in trans complementation constructs within the mprF knockout strain. For complementation strategies, we employed a plasmid system expressing either a wild-type (WT) or a singly point-mutated form of mprF ORFs derived from two distinct isogenic DAP^s-DAP^r strain pairs. We investigated the impact of such SNPs on in vitro susceptibilities to DAP and a prototypical CAMP felt to be important in innate host defense against S. aureus infections (neutrophil-derived hNP-1 [10, 19]). Moreover, CM phospholipid profiles, net surface positive charges, and DAP whole-cell binding were analyzed to identify if the mprF SNPs were causally related to the phenotypic changes that might affect in vitro DAP or CAMP resistance.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The bacterial strains used in this study are listed in Table 1. We used the DAP^s S. aureus Newman strain (30), its isogenic mprF knockout strain (ΔmprF), and several in trans complementation constructs within the Newman ΔmprF strain, employing a low-copy-number plasmid (pRB474 [31]) carrying either a WT or a point-mutated form of mprF cloned from two isogenic DAP^s-DAP^r strain pairs (methicillin-susceptible S. aureus [MSSA] 616-701 [10, 25] and MRSA11/11-REF2145 [24, 26]). It should be underscored that this plasmid only expresses in trans complementation constructs during exponential growth (the maximal point of mprF expression [10, 25]). The DAP MICs (Etest) for the pairs 616-701 and MRSA11/11-REF2145 were 0.5 to 2 μg/ml and 1 to 4 μg/ml, respectively, as previously reported (10, 24). The two isogenic pairs were selected for mprF cloning based on the location of SNPs within mprF gene. Each DAP^r strain has a defined SNP within the central transmembrane segment bordering the synthase-translocase interface domain (S295L in 701 [25] and T345A in the REF2145 strain [24, 25]). Parental Newman and Newman ΔmprF strains bearing the empty plasmid were used as controls.

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Descriptiona	Reference or source
S. aureus
Newman	MSSA; wild-type strain isolated from a human infection in 1952	30
Newman ΔmprF	Newman mprF::Em; Emr	19
CNewman	Newman ΔmprF strain expressing mprF cloned from Newman	This study
C616	Newman ΔmprF strain expressing mprF cloned from 616	This study
C701	Newman ΔmprF strain expressing mprF cloned from 701	This study
CMRSA11/11	Newman ΔmprF strain expressing mprF cloned from MRSA11/11	This study
CREF2145	Newman ΔmprF strain expressing mprF cloned from REF2145	This study
E. coli DH5α	Host strain for construction of recombinant plasmids	49
Plasmids
pRB474	Shuttle vector carrying B. subtiltis vegII promoter; Cmr	34
pCR2.1	E. coli plasmid; Ampr	Invitrogen

^aAbbreviations: Em^r, erythromycin resistance; Cm^r, chloramphenicol resistance; Amp^r, ampicillin resistance.

All S. aureus strains were grown in either tryptic soy broth (TSB; Difco Laboratories, Detroit, MI) or Mueller-Hinton broth (MH; Difco Laboratories) for individual experiments. Liquid cultures were grown in Erlenmeyer flasks at 37°C with shaking (250 rpm) in a volume that was no greater than 10% of the flask volume. All strains were maintained at −70°C until thawed before each experimental run.

DNA manipulations and mprF cloning.

Genomic DNA was isolated from S. aureus using the method of Dyer and Iandolo (31). Plasmid DNA purification was performed using Wizard Plus kits from Promega, Inc. (Madison, WI). Preparation and transformation of Escherichia coli DH5α were accomplished using the method described by Inoue et al. (32). Electroporation of plasmid DNA into S. aureus was carried out using the procedures of Schenk and Laddaga (33).

Complementation of the mutated or nonmutated forms of mprF in Newman ΔmprF was achieved by PCR amplifying the mprF ORFs with primers mprF-F-bamHI (5′-CCCGGATCCAATTAGAATTGATGTGAAAAAATG-3′) and mprF-R-sphI (5′-CCCGCATGCAGCGCTTCAGGCATAACTGT-3′) using genomic DNA samples from strains Newman, 616, 701, MRSA11/11, and REF2145. The resulting PCR products were ligated into the BamHI and SphI sites of the Gram-positive expression vector pRB474 (34). This placed the expression of mprF under the control of the vegII promoter, a vegetative promoter from Bacillus subtilis that only expressed during the exponential growth phase (35). DNA sequencing of the mprF ORFs cloned into pRB474 was kindly performed at City of Hope, Duarte, CA, as described before (25, 28).

To assess expression levels of cloned mprF genes in Newman ΔmprF, quantitative reverse transcription-PCR (qRT-PCR) analyses were performed using primers qRT-mprF-F and qRT-mprF-R as described previously (36). The qRT-PCR analyses confirmed that all the cloned mprF genes were expressed in trans at levels comparable to that of the wild-type Newman strain during exponential growth (data not shown).

Antimicrobial agents and MIC testing.

DAP powder was provided by Cubist Pharmaceuticals (Lexington, MA). Vancomycin (VAN) powder was purchased from Sigma Chemical Co. (St. Louis, MO). The MICs of DAP, VAN, and oxacillin (OX) were determined by standard micro-Etest according to the manufacturer's recommended protocols. A minimum of two independent experimental runs was performed.

Population analysis.

Population analyses of the strain sets were performed with DAP and VAN as described before (28, 37). Briefly, the range of concentrations tested was 0 to 16 μg/ml to encompass sublethal to lethal drug levels, using an initial inoculum of ∼10⁹ CFU/ml. DAP population analysis was performed in the presence of 50 μg/ml of calcium (Ca²⁺), as recommended by the manufacturer. To provide a quantification of the population analyses, area-under-the population analysis curves (AUCs) were calculated using the trapezoidal rule and statistically compared as previously detailed (Microcal Origin 5.0; Northampton, MA [26]). A minimum of three independent experimental runs was performed.

CAMP susceptibility testing.

Standard MIC testing in nutrient broth may underestimate CAMP activities (20, 38). Accordingly, in vitro bactericidal assays were carried out with hNP-1 as described previously using a 2-h microdilution method in Eagle's minimal essential medium (10, 20). hNP-1, a prototypical α-defensin, was purchased from Peptide International (Louisville, KY). For the killing assays, we used a final inoculum of 5 × 10³ CFU of overnight-grown cells and hNP-1 at 10 μg/ml. This hNP-1 concentration was selected based on extensive pilot studies showing its inability to substantially reduce starting inocula of parental strain Newman over the 2-h exposure period. Data were calculated and expressed as the relative percent surviving CFU (±standard deviation [SD]) of CAMP-exposed versus CAMP-unexposed cells. A minimum of three independent runs was performed.

CM PL contents.

The three major S. aureus CM phospholipids (PLs) are L-PG, PG, and cardiolipin [CL] (15, 19, 21). To quantify the relative proportions of these three PLs in our strain sets, CM PLs were extracted from S. aureus cell pellets as described previously (18). The target PLs were separated and identified by using two-dimensional thin-layer chromatography (2D-TLC) by their mobilities and ninhydrin staining properties (10, 18), removed from the plates, and then quantified spectrophotometrically by a previously described chemical assay (10, 18). The proportion of synthesized L-PG which was translocated to the outer CM leaflet was quantified spectrophotometrically as detailed before, using the L-PG outer CM-impermeable UV probe fluorescamine (18, 23). Data were expressed as the percent proportionalities (±SDs) of the three PLs. At least three independent experiments were performed to analyze the PL contents.

Net cell surface charge.

To quantify relative cell surface charge in our study strains, we employed the well-established cytochrome c binding assays. The binding of cytochrome c (Sigma) was measured spectrophotometrically by quantifying the amount of this polycation remaining within reaction mixture supernatants following 10 min of exposure to the study strains (∼10⁹ CFU); the larger the amounts of residual cytochrome c in the supernatants, the more relative net positive surface charge (10, 18, 26, 39). Data were calculated and mathematically converted to express the percentage of cytochrome c bound to the cell. Data shown for surface charge assays are the means (±SDs) of three independent experiments.

DAP binding assays.

To determine relative profiles of whole-cell DAP binding to the S. aureus study strains, 6 μg/ml of DAP was added to 10⁸ CFU of each strain, as previously detailed (10) Supernatants were then analyzed for residual unbound peptide by a radial-diffusion assay; using a standard curve technique as described previously (10, 40), the amount of bound peptide (±SD) was calculated. These assays were performed at least four times for each strain on separate days.

Statistical analysis.

The Kruskal-Wallis ANOVA test with the Tukey post hoc correction for multiple comparisons was utilized where indicated. Significance was determined at a P value of <0.05.

RESULTS

DAP and CAMP susceptibility profiles.

As shown in Table 2, complementation of the Newman ΔmprF strain with plasmids expressing singly point-mutated forms of mprF (mprF_S295L or mprF_T345A) resulted in increased DAP MICs versus those for the strains expressing the wild-type mprF genes (mprF₆₁₆ or mprF_MRSA11/11). For example, the Newman ΔmprF strain had a DAP MIC of 0.125 μg/ml; incorporation of the complementation plasmid with wild-type mprF from either the parental Newman, 616, or MRSA11/11 strain caused a 2-fold increase (0.25 μg/ml) in DAP MICs. In contrast, mutant forms of mprF (mprF_S295L or mprF_T345A) within the Newman ΔmprF strain caused an 8- to 16-fold increase in DAP MICs compared to that for the Newman ΔmprF strain. There was no impact of mprF point mutations on VAN or OX MICs.

Table 2.

In vitro susceptibilities of the study strains to DAP and the host defense CAMP, hNP-1

Strain	mprF SNP	MIC (μg/ml)			% survival (mean ± SD) after 2 h of exposure to: hNP-1 (10 μg/ml)
		DAP	VAN	OX
Newman		0.5	3	0.38	47.48 ± 11.32
Newman ΔmprF		0.125	3	0.38	9.72 ± 5.77a
CNewman		0.25	3	0.38	27.19 ± 10.46
C616 (DAPs)	WT	0.25	3	0.38	34.95 ± 8.51
C701 (DAPr)	S295L	1	3	0.38	39.78 ± 17.39
CMRSA11/11 (DAPs)	WT	0.25	3	0.38	37.21 ± 10.82
CREF2145 (DAPr)	T345A	2	3	0.38	48.48 ± 8.19

^aP < 0.01 versus WT Newman strain.

As also shown in Table 2, deletion of mprF in the Newman parental strain resulted in significantly increased susceptibility to the innate host defense CAMP, hNP-1 (P < 0.01). Complementation of the Newman ΔmprF strain with a plasmid expressing the WT mprF gene cloned from the Newman strain partially restored susceptibilities to hNP-1 toward the parental level. However, complementation of the Newman ΔmprF strain with plasmids expressing either mprF_S295L or mprF_T345A showed a near-parental-level or parental-level restoration of hNP-1 susceptibility profiles. However, these susceptibility difference data did not reach statistical significance compared with the Newman ΔmprF strains complemented with either of the parental forms of mprF.

Population analyses.

For the Newman ΔmprF strains expressing either of the two point-mutated mprF genes (mprF_S295L or mprF_T345A), the DAP^r population curves were substantially shifted to the right compared to those for strains complemented with the parental mprF genes (Fig. 1). Area-under-the curve (AUC) values for DAP^r population analyses were ∼4-fold and ∼7-fold greater for the two strains complemented with point-mutated forms of mprF genes than for the strains complemented with the parental mprF genes, respectively (mean ± SD, 1.90 ± 0.02 for C₆₁₆ versus 7.36 ± 0.29 for C₇₀₁ and 2.93 ± 0.78 for C_MRSA11/11 versus 21.62 ± 0.62 for C_REF2145; P < 0.0001 for both comparisons). As expected, AUC values for S. aureus Newman and the C_Newman construct (ΔmprF strain complemented with mprF_Newman; 5.68 ± 0.03 and 5.15 ± 0.3, respectively) were substantially higher than for the Newman ΔmprF strain (1.85 ± 0.04) (P < 0.01). Results of VAN^r population analyses of the strain sets were virtually identical regardless of which mprF genes were employed for complementation (data not shown).

Fig 1.

Population analyses of study strains upon exposure to a range of DAP concentrations. These data represent the means (±SDs) for three separate assays.

Synthesis and translocation of L-PG.

The proportion of L-PG synthesized within the overall PL content was significantly increased in the two strains expressing point-mutated mprF genes compared to that of each strain containing the respective wild-type mprF genes (Table 3). This was mainly related to relative enhancement of L-PG synthesis rather than increased translocation profiles.

Table 3.

Asymmetry of L-PG and DAP binding assays

Strain	% of total PL content (mean ± SD)			Amt (μg) of bound DAP (6 μg)
	Inner L-PG	Outer L-PG	Total L-PG
Newman	20.94 ± 3.90	1.87 ± 0.24	22.81 ± 3.66	0.74 ± 0.08
Newman ΔmprF	0	0	0	1.12 ± 0.11a
CNewman	11.32 ± 3.62	2.86 ± 0.08	14.18 ± 3.54	0.88 ± 0.14
C616	1.87 ± 1.48	2.85 ± 1.95	4.71 ± 3.43	0.89 ± 0.22
C701	11.88 ± 0.51a	2.80 ± 0.47	14.68 ± 0.04a	0.58 ± 0.20a
CMRSA11/11	1.93 ± 1.88	3.28 ± 1.58	5.21 ± 3.46	0.99 ± 0.21
CREF2145	10.13 ± 5.61a	2.42 ± 1.55	12.55 ± 7.16a	0.61 ± 0.23a

^aP < 0.05 versus the Newman WT, C₁₆₁, and C_MRSA11/11 strains.

Net surface positive charge.

As expected, the Newman ΔmprF mutant had a significantly reduced positive surface charge compared to that of the Newman parental strain (Fig. 2). This perturbation was corrected toward the parental level by complementation with the Newman WT mprF. Of note, the net positive surface charges of both the strains expressing point-mutated mprF were significantly higher than those of the strains complemented with their respective wild-type mprF.

Fig 2.

Binding of positively charged cytochrome c to S. aureus whole cells. The graph shows percentage of cytochrome c bound after 10 min of incubation with each study strain at room temperature. Data represent the means (±SDs) from three independent experiments. *, P < 0.01.

DAP whole-cell binding.

The Newman ΔmprF construct bound significantly more DAP than the Newman parental strain (Table 3); complementation with Newman mprF normalized this construct to parental-level DAP binding. Complementation of the Newman ΔmprF strain with either mprF_S295L or mprF_T345A resulted in significantly reduced DAP binding versus that of the strains expressing the wild-type mprF genes (Table 3).

DISCUSSION

MprF controls the overall production and translocation of lysyl-phosphatidylglycerol (L-PG), the unique positively charged S. aureus cell membrane (CM) phospholipid (PL); this protein thus affects susceptibilities to a variety of host defense CAMPs and bacterium-derived lantibiotics, as well as other positively charged antibiotics, such as calcium-DAP (15, 17, 41). This “resistance” phenomenon is presumably via a charge repulsion mechanism (10, 16, 17, 25, 28). In support of this notion, S. aureus mprF deletion mutants demonstrate increased susceptibility to a number of CAMPs, including thrombin-induced platelet microbicidal proteins and hNP-1, compared to that of relevant parental strains (17, 19). Also, deletion of mprF resulted in a reduced virulence in experimental infection models, including experimental infective endocarditis, presumably related to enhanced clearance of mprF mutants from sites of endovascular infection by platelet- and/or white cell-derived CAMPs which are prevalent at such sites (17, 19).

Recently, Friedman et al. demonstrated the progressive accumulation of a series of SNPs induced by in vitro exposure of MRSA strains to DAP, including point mutations in mprF, yycG, rpoB, and rpoC (42). In all instances, the mprF mutation was the earliest SNP acquired (42). Of note, similar point mutations in the mprF ORF have been documented to emerge in clinical S. aureus isolates from patients failing DAP treatment (10, 12, 21, 22, 25–28, 42–44). It is important to emphasize that although the acquisition of mprF SNPs (with or without concomitant yycG SNPs) is frequently found in DAP^r strains (22), the latter phenotype may either occur in the absence of such SNPs or be unrelated to MprF gains in function despite the presence of such SNPs (12, 28). The exact mechanism(s) by which SNPs within the yyc operon cause the DAP^r phenotype is not known.

Interestingly, in DAP^r staphylococcal isolates, mprF mutations have mainly occurred in well-defined hot spots within the mprF ORF, usually within the central bifunctional domain (a region controlling both synthase and translocase functionalities). The current study yielded cogent findings that support a causal role of mprF SNPs in the DAP^r phenotype. First, the incorporation of singly point-mutated forms of mprF (mprF_S295L or mprF_T345A) into the complementation plasmid yielded increased DAP MICs above the Newman parental levels (0.5 μg/ml) and into the near-DAP^r (1 μg/ml) or DAP^r (2 μg/ml) range. In addition, there were significant rightward shifts in DAP^r population analysis AUCs in the latter constructs compared to Newman ΔmprF, as well as to ΔmprF strains complemented with parental forms of the mprF genes (C₆₁₆ or C_MRSA11/11). These data indicate that individual and distinct point mutations within mprF can roughly recapitulate the DAP^r phenotype observed in donor strains (701 and REF2145; DAP MICs of 2 and 4 μg/ml, respectively) (10, 26). Second, deletion of mprF provided a construct (Newman ΔmprF) which was hypersusceptible to the host defense CAMP, hNP-1. Complementation of this construct with the native Newman mprF or either of the other parental mprF forms returned the hNP-1 susceptibility profile somewhat toward the parental level. In contrast, complementation with the mutated forms of mprF yielded near-parental-level hNP-1 susceptibility profiles. Thus, the mprF locus appears to be required, although not entirely sufficient, for rendering strains relatively resistant to killing by hNP-1. The lack of total restoration of parental-level hNP-1 resistance profiles may be related to (i) the presence of heterologous forms of mprF within the Newman genetic background and/or (ii) restriction of mprF expression to only the exponential growth phase in the specific plasmid utilized in this investigation. To the latter point, we have previously shown that one of the important genetic correlates of DAP^r is dysregulation of mprF expression in DAP^r mutants, featuring maintained mprF expression during stationary growth phases (25, 28).

Third, recently published studies by our group and others have suggested that there are two prominent DAP/CAMP-adaptive phenotypic mechanisms in S. aureus: (i) increased positive surface charge (10, 17, 25, 28) and (ii) “recalibration” of CM biophysical order (fluidity/rigidity) (18, 23, 45). In the current study, we found that the proportional amount of L-PG synthesized within the overall PL content was significantly increased in the C₇₀₁- and C_REF2145-complemented strains compared to that in their parental mprF constructs, C₆₁₆- and C_MRSA11/11-complemented strains, respectively. However, the C₇₀₁ and C_REF2145 strains exhibited levels of outer leaflet L-PG similar to those of their parental C₇₀₁- and C_REF2145-complemented strains, indicating that the L-PG synthase activity, but not the translocase activity, was enhanced by the observed SNPs (mprF_S295L or mprF_T345A). However, it is not currently clear why complementation of the Newman ΔmprF strain with parental or mutated forms of mprF genes failed to restore the capacity of such strains to synthesize L-PG to parental Newman levels. We speculate that, as described above, the complementation plasmid used (pRB474) (34), which expresses cloned mprF genes only at exponential growth phase, may have blunted the “normal” biphasic expression profiles of mprF and resultant CM PL phenotypes (25, 46). Nonetheless, the enhanced synthesis of L-PG in C₇₀₁- and C_REF2145-complemented strains versus those of the respective C₆₁₆- and C_MRSA11/11-complemented strains correlated with enhanced net positive surface charge and reduced DAP binding in those strains, supporting, at least in part, a charge repulsion mechanism for DAP^r being in play (10, 17, 47, 48). In addition, we investigated the CM fluidity/rigidity properties and pigment content among the strain sets but found no substantial differences among the strain sets that could be ascribed to the presence or absence of parental versus mutated forms of mprF (data not shown).

In summary, our data suggest that (i) individual and distinct point mutations within the mprF ORF can roughly recapitulate phenotypes observed in mprF donor strains (i.e., changes in DAP MICs, surface positive charge, L-PG profiles, and susceptibility profiles to host defense CAMPs), and (ii) these gain-in-function SNPs (i.e., enhanced L-PG synthesis) likely promote reduced DAP binding to S. aureus cells by a surface charge repulsion mechanism. Thus, for these two DAP^r strains, the defined mprF SNPs appear to be causally related and major contributors to this phenotype. Our current findings nicely complement those of Rubio et al. (29) which showed that mprF antisense RNA knockdown strategies were able to reverse the DAP^r phenotype in vitro.

Current studies are in progress to further examine the relationship of mprF SNPs and DAP^r, to better define the genetic regulatory pathway(s) that influences DAP^r, and to characterize the breadth of host defense CAMP cross-resistance among DAP^r strains.