Serial Daptomycin Selection Generates Daptomycin-Nonsusceptible Staphylococcus aureus Strains with a Heterogeneous Vancomycin-Intermediate Phenotype

Bacterial strains.N315ΔIP is a MRSA strain that has been described previously (31). It is a derivative of pre-MRSA strain N315 whose mecI gene is intact and in which methicillin resistance is not expressed (31). N315ΔIP represents the genotype of hospital-acquired MRSA strains in Japan whose mecI gene is inactivated by mutations. The penicillinase plasmid of N315 was further eliminated from N315ΔIP to make the strain more pertinent as a model for the tracing of the development of vancomycin-intermediate resistance in Japanese hospital-acquired MRSA strains, since the representative heterogeneous VISA and VISA strains Mu3 and Mu50, respectively, lacked the penicillinase plasmid. N315ΔIP was classified as sequence type 5, staphylococcal chromosomal cassette mec type IIa, and its agr type is II. Strain 10*3d1 is daptomycin nonsusceptible and a derivative of N315ΔIP obtained in this study. It was isolated after exposure to daptomycin in two steps, as follows. A total of 10⁷ CFU of N315ΔIP was spread on Mueller-Hinton agar (MHA) plates supplemented with 50 μg/ml calcium and 0.5 μg/ml daptomycin (Cubist Pharmaceuticals, Lexington, MA). Colonies were recovered, and those with increased MICs were spread as described above on plates containing 1 μg/ml daptomycin. Colonies were recovered, and the MICs were determined. The colony with the highest MIC was subcultured for 6 consecutive days on plates containing 1 μg/ml daptomycin to stabilize the strain. Strain 10*3d1 has an MIC of 3 μg/ml and is considered nonsusceptible by the CLSI (9). Strain 10*3d1-10 is a daptomycin-susceptible derivative strain obtained by serially cultivating 10*3d1 in 4 ml of drug-free Mueller-Hinton broth (MHB) at 37°C. After 24 h, 4 μl was transferred to 4 ml of fresh medium. This procedure was repeated for 10 consecutive days. All strains were checked for their isogenicity by pulsed-field gel electrophoresis (data not shown) and were stocked at −80°C in 40% glycerol.

Antibiotic susceptibility tests.The daptomycin MICs were determined by the broth microdilution method with concentrations 0.125, 0.25, 0.50, 1.0, 2.0, 4.0, and 6.0 μg/ml. The test medium was MHB supplemented with calcium, according to the criteria of the CLSI (9). MICs were also determined by Etest (AB Biodisk, Solna, Sweden) in MHA, as recommended by the manufacturer, since Etest results have already been demonstrated to have a good correlation with those of the CLSI method (24). In order not to miss a slowly growing resistant cell subpopulation, the MICs were evaluated not only at an incubation time of 24 h but also at an incubation time of 48 h (22). The MICs of vancomycin, teicoplanin, linezolid, oxacillin, imipenem, bacitracin, fosfomycin, and colistin were determined by Etest in MHA; with fosfomycin, however, MHA was supplemented with 25 μg/ml d-glucose-6-phosphate. The Etest with daptomycin was also performed in the presence of gadolinium (III) chloride at 200 μM in MHA. Gadolinium chloride is a well-known mechanosensitive channel blocker (18) that can decrease the membrane potential of the cell and that could give us new insight into the importance of the high membrane potential of daptomycin-resistant strains.

Population analysis.Analysis of the daptomycin-nonsusceptible subpopulation of strain 10*3d1 was performed by comparing it to the population of parental train N315ΔIP, as described previously (20). An appropriately diluted overnight culture was spread onto MHA plates supplemented with 50 μg/ml calcium and daptomycin at various concentrations. The plates were then incubated at 37°C for 48 h to 72 h before the colonies were enumerated. The population curve was drawn by calculating and plotting the number of resistant cells theoretically contained in 50 μl of the undiluted culture.

Growth curve.Overnight cultures of the test strains were diluted 1/1,000 in 10 ml fresh MHB and were grown at 37°C with shaking at 25 rpm in a photorecording incubator (TN-2612; Advantec, Tokyo, Japan). The optical density (OD) was automatically monitored every 2 min up to an OD at 600 nm (OD₆₀₀) of 0.8. For growth curve and doubling time determinations, the OD versus time was plotted for each strain in the exponential growth phase. The doubling times were then calculated as follows: [(t₂ − t₁) × log 2]/(log OD₆₀₀ at t₂ − log OD₆₀₀ at t₁), where t₁ is sampling time 1 and t₂ is sampling time 2.

Membrane potential measurement.The membrane potential was determined by flow cytometry by using a FACSAria apparatus (BD, Biosciences, San Jose, CA) with 3,3′-diethyloxacarboncyanine iodide [DiOC₂(3); BacLight bacterial membrane potential kit; Molecular Probes, Eugene, OR], as described before (48), except for the following modifications. Strains were grown to exponential phase (OD₆₀₀ = 0.3) in MHB supplemented with 50 μg/ml calcium and diluted 1:100 in 1 ml of the component C provided by the manufacturer. Negative controls consisted of cells treated with carbonyl cyanide 3-chlorophenylhydrazone (CCCP), which was provided with the kit, and gadolinium (III) chloride (Sigma, St. Louis, MO), which was used at 200 μM (final concentration). In order to distinguish depolarized cells from cells with damaged membranes, the impermeant DNA-binding dye TO-PRO-3 (concentration, 500 nM) was combined with the cells, and an additional control consisting of N315ΔIP heat-killed cells was included. A total of 30 μM DiOC₂(3) was added to each sample, and the tubes were incubated at room temperature for 4 min. Flow cytometry was performed with 10,000 cells per run by using a 488-nm beam from an argon ion laser to excite the DiOC₂(3). The green and red fluorescences were detected as described previously (48), and the ratios of these values obtained for each strain allowed us to compare the membrane potentials of these strains. TO-PRO-3 was excited with a 633-nm beam from a helium-neon laser, as recommended (48), and the detection of an elevated far-red fluorescence with this dye (detected through a 695-nm long-pass filter) was indicative of membrane damage (48). The membrane potential was measured as the fluorescence ratio (red/green) by using the population mean fluorescence intensities.

Cell membrane labeling and fluorescence polarization measurements.Membrane fluidity was determined by measuring fluorescence polarization with 1,6-diphenyl-1,3,5-hexatriane (DPH; Sigma) as a probe, as described by Bayer et al. (3). In brief, an overnight culture of S. aureus was washed two times with cooled 0.05 M Tris-HCl (pH 7.6) buffer by centrifugation (at 5,000 × g for 15 min), and the culture was suspended in digestion buffer with a bacterial density of about 10⁹ CFU/ml. The cell wall-free staphylococcal protoplasts were prepared as described previously (10). For cell membrane labeling with DPH, a 2 mM solution of the dye was prepared in tetrahydrofuran, and 100 μl of the solution was added to 50 ml of 0.05 M Tris-HCl (pH 7.6). Excess tetrahydrofuran was removed by flushing the solution with nitrogen. The protoplasts were then incubated in the DPH solution at a final concentration of 2 μM for 60 min at 30°C. Fluorescence polarization was measured after incubation of the labeled sample at 35°C for 30 min with a fluorescence spectrophotometer (F-4500-plus-FP; Hitachi, Tokyo, Japan). The excitation and emission wavelengths were 360 and 426 nm, respectively. The corrected fluorescence was determined by subtracting the emissions of the unlabeled control protoplasts from those of the labeled protoplasts. The degree of fluorescence polarization, or the polarization index (PI), was calculated from the following formula, as described by Bayer et al. (3): [I_V − I_H (I_HV/I_HH)]/[I_V + I_H (I_HV/I_HH)], where I is the corrected fluorescence intensity and subscripts V and H indicate the values obtained with vertical or horizontal orientation of the analyzer, respectively. The lower that the PI value is, the more fluid the membrane is (3).

Electron microscopic evaluation of cell wall thickness.The preparation of S. aureus cells for transmission electron microscopy and the examination of S. aureus cells by transmission electron microscopy were performed as described previously (12). Morphometric evaluation of the cell wall thickness was performed by using photographic images at a final magnification of ×30,000. Thirty cells of each strain with nearly equatorial cut surfaces were measured for the evaluation of cell wall thickness, and the results were expressed as the mean value ± standard deviation.

Autolysis assay.The autolysis assay was performed as described previously (39). Basically, overnight cultures of the test strains were diluted 1/1,000 in 10 ml fresh LB broth (pH 7.2) and were grown at 37°C with shaking at 25 rpm in the photorecording incubator indicated above. The OD was monitored automatically every 2 min, and the cells were grown to an OD₆₀₀ of 0.7. Afterwards, the cells were pelleted by centrifugation at 7,500 rpm for 5 min at 4°C. The cells were then washed twice and resuspended in 0.01 M sodium phosphate buffer (pH 7.0). The cell suspension was then incubated at 37°C with continuous shaking at 25 rpm. The decrease in the OD was monitored every 2 min for 20 h with the same photorecording incubator described above.

RNA preparation and microarray analysis.The bacteria were grown in 10 ml MHB to exponential phase (OD₆₀₀ = 0.6) before they were harvested. Pellets were then suspended with precooled T₁₀E₁₀ buffer (10 mM Tris-HCl [pH 8.0]), followed by addition of lysostaphin solution (Wako, Japan) at a final concentration of 3.0 μg/ml. The suspensions were then incubated at 37°C for 3 min until complete cell lysis occurred. Seven milliliters of acidic phenol (pH 5.2, equilibrated with 20 mM sodium acetate) was immediately added, followed by addition of 600 μl of 3 M sodium acetate. The samples were then frozen and thawed three 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 by phenol-chloroform extraction and ethanol precipitation. The pellets were then resuspended in 25 μl diethyl pyrocarbonate-treated water. The construction and analysis of the DNA microarray have been described previously (13). To confirm the reproducibility of the assay, RNA extraction and hybridization were performed in duplicate for each experiment. The results are given as the ratio of the average signal intensity of 10*3d1 to that of N315ΔIP to compare nonsusceptible and parental strains and the ratio of the average signal intensity of 10*3d1-10 to that of N315ΔIP to compare what changed and how similar the transcriptional profiles of the susceptible and the parental strain are.

Construction of overexpression plasmid and electroporation.Plasmid DNA was purified from Escherichia coli by using an SV Minipreps DNA purification system (Promega Corporation, Madison, WI), according to the manufacturer's instructions. The isolation of DNA from S. aureus cells, restriction endonuclease digestion, ligation reactions, and DNA cloning were carried out as described previously (13). Electroporation of S. aureus was performed with a Bio-Rad gene pulser with a pulse controller, as described previously (31).

Overexpression plasmids were constructed by cloning the PCR-amplified response regulator genes SA2179 and SA2151 from 10*3d1 into plasmid vector pRIT5 and cloning pRIT5 in N315ΔIP. SA2179 was amplified with the primers 5′-AAATAGAATTCGCAGAAGTGATAGATGCGGGAAATGGTTTTG-3′ and 5′-AAATAGTCGACGGTACCGAAAGGAGAAAGGATTAGGTTCCATTC-3′, in which the underlined nucleotides in the two primers represent added EcoRI and SalI sites, respectively. SA2151 was amplified with the primers 5′-ATAATGGATCGCGTCAGTATAAGGTTGAAATATGAACTCAATATGAAC-3′ and (5′-ATAATCTGCAGGTGTCTTCATGATTTTCGCGTCATTAGATGC-3′), in which the underlined nucleotides in the two primers represent added BamHI and PstI sites, respectively. The integrities of the cloned genes were ascertained by sequencing each overexpression plasmid. The resultant plasmids, p2179 and p2151, respectively, were then introduced into N315ΔIP, resulting in strains ΔIP(2179) and ΔIP(2151), respectively. Overexpression was carried out overnight in MHB plus 10 μg/ml chloramphenicol. Afterwards, the culture was spread at an OD₆₀₀ of 0.3 onto MHA for daptomycin MIC determination by Etest.

Microarray data accession number.The profiles of strains 10*3d1 and 10*3d1-1 may be found under CIBEX accession no. CBX67.

Daptomycin-nonsusceptible and -susceptible strains derived from N315ΔIP.After two steps of daptomycin exposure, a daptomycin-nonsusceptible strain, strain 10*3d1, whose MICs were 2.0 μg/ml by the broth microdilution method and 3.0 μg/ml by Etest was isolated. Its phenotypic revertant (which was daptomycin susceptible), strain 10*3d1-10, was then obtained by serial passage of 10*3d1 for 10 days in drug-free MHB. Its daptomycin MIC was 1.0 μg/ml by both Etest (Table 1) and the broth microdilution method, and the strain was considered susceptible to daptomycin by the CLSI criteria (9).

Before the following experiments were started, the strains were confirmed to be isogenic by their identical pulsed-field gel electrophoresis patterns (data not shown). The MICs of several drugs for this isogenic set were determined by Etest (Table 1), and the values were read after 48 h at 37°C. It was noted that besides the increase in the daptomycin MIC from 0.5 to 3 μg/ml in strain 10*3d1, its vancomycin MIC increased from 0.75 to 3.0 μg/ml, its teicoplanin MIC increased from 0.5 to 6.0 μg/ml, and its imipenem MIC increased from 3.0 to >32 μg/ml (Table 1). The daptomycin MICs decreased in passage-derived susceptible strain 10*3d1-10, and the MICs of vancomycin and teicoplanin also decreased (Table 1). On the other hand, decreases in the MICs of beta-lactam antibiotics were not observed.

Nonsusceptible and phenotypic revertant strains grow slowly.Both strain 10*3d1 and strain 10*3d1-10 had very slow growth rates compared to the growth rate of parental strain N315ΔIP. The doubling time of 10*3d1 was 57.25 min, while that of N315ΔIP was 21.36 min. Susceptible (phenotypic revertant) strain 10*3d1-10 recorded a doubling time of 44.63 min: although it grew faster than its parent, strain 10*3d1, it still grew much slower than N315ΔIP (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 1.

Growth curves of N315ΔIP and derivative strains. Circles, N315ΔIP; squares, 10*3d1; triangles 10*3d1-10.

Population analysis shows daptomycin heteroresistant subpopulations in 10*3d1.To analyze whether strain 10*3d1 has homogeneous resistance, we performed a population analysis and compared the results with those for strain N315ΔIP (Fig. 2A). Because of the slow growth of 10*3d1, in the population analysis the colonies were counted only after 72 h instead of after 48 h. Strain 10*3d1 showed a heterogeneous population curve and was judged to have a heterogeneous daptomycin resistance phenotype. The strain also showed heterogeneous vancomycin resistance, as illustrated in Fig. 2B.

• Open in new tab
• Download powerpoint

FIG. 2.

Population analysis of N315ΔIP (open circles) and 10*3d1 (black squares). The numbers of colonies on plates containing various concentrations of daptomycin (A) and vancomycin (B) were counted after 72 h of incubation at 37°C due the slow growth of the nonsusceptible strain.

Overview of gene expression.We compared the microarray transcription profile of 10*3d1 with that of N315ΔIP in drug-free medium to figure out which genes are correlated with daptomycin resistance. Afterwards, we compared the microarray transcription profile of 10*3d1-10 with that of N315ΔIP to determine how similar gene expression in 10*3d1-10 and N315ΔIP was. With these results, we were also able to compare the profiles of 10*3d1 and 10*3d1-1 (CIBEX accession no. CBX67) and see which genes have changed their expression levels after 10 days of serial passage in drug-free medium.

Transcription analysis showed upregulation of some enzymes of fatty acid metabolism and repression of phospholipid biosynthesis.Because the major target of daptomycin described so far is the membrane, we carefully observed the genes encoding the enzymes involved in fatty acid and phospholipid biosynthesis whose expression changed in strains 10*3d1 and 10*3d1-10 compared to their expression in strain N315ΔIP.

The expression of the enzymes involved in the metabolic process of fatty acid (SA0223 to SA0226) was enhanced in strain 10*3d1 (ratios > 2.0) but not in strain 10*3d1-10 (Table 2). On the other hand, vraA and vraB, which encode the long-chain fatty acid coenzyme A (CoA) ligase and acetyl-CoA c-acetyltransferase, respectively, along with vraC, which is included in this operon, were repressed only in 10*3d1-10. These data suggest that enhanced fatty acid metabolism may be correlated with enhanced resistance to daptomycin.

The expression of some genes involved in phospholipid biosynthesis in strains 10*3d1 and 10*3d1-10 also changed. The gene encoding cardiolipin synthase (cls) was repressed in the two strains (0.47 and 0.30, respectively), and mprF (or fmtC) was slightly repressed in both strains (0.61 and 0.51, respectively). mprF was sequenced, and there was no mutation in 10*3d1 (data not shown). These results show that the alterations in enzymes involved in the biosynthesis of phospholipids persisted even in the strain cultivated for 10 days in drug-free medium.

Besides fatty acid metabolism and phospholipid synthesis, the genes required for glycolipid synthesis, lipoteicoic acid anchoring (SA0875 [ypfP], SA0874 [ltaA]), and teichoic acid synthesis (SA0592 [tagA]) were also repressed. Teichoic acid gene repression was accentuated in the susceptible strain (Table 2). The expression of the ltaS (SA0674) gene, which encodes polylglycerol phosphate synthase (LtaS), which is required for lipoteichoic acid synthesis, was not altered in strain 10*3d1 compared to that in N315ΔIP, but it was repressed in 10*3d1-10 (Table 2). The autolysin gene (atl), however, was upregulated in both strains.

Regulators.Some two-component systems were expressed differently in strain 10*3d1 and its susceptible derivative strain compared to their expression in strain N315ΔIP. The vraS and vraR genes were upregulated only in the daptomycin-nonsusceptible strain (Table 2). The expression of the graS gene was slightly increased in the nonsusceptible strain, strain 10*3d1 (Table 2), and the graR and graS genes did not show any transcriptional difference in susceptible strain 10*3d1-10 compared to that in N315ΔIP. The entire operon containing vraSR and the graRS genes were sequenced, and no mutations were found.

The other two-component systems, SA2152 and SA2179, which have not yet been explored by our group, had only one of their components increased in strain 10*3d1 but not in strain 10*3d1-10 (Table 2). However, the introduction of the SA2179 and SA2151 genes on multicopy plasmids into N315ΔIP did not raise the daptomycin MICs, which were between 0.38 and 0.50 μg/ml by Etest.

Genes related to the electron transport system.When an electron transport system is enhanced, it increases the membrane potential by pumping protons out of the cell. Several enzymes correlated with the electron transport system were upregulated in strains 10*3d1 and 10*3d1-10 compared to their expression in strain N315ΔIP. NAD-dependent formate dehydrogenase, encoded by fdh (SA0171), was the enzyme whose transcription level increased the most in 10*3d1, reaching a ratio of 82.28 when its transcription level was compared with that in N315ΔIP. This gene was still upregulated in the susceptible strain to levels considered high (ratio = 3.18); however, its upregulation was much less so than that in the resistant strain (ratio = 82.28). Other sets of the genes involved in the electron transport system, such as succinate dehydrogenases and quinol oxidases, were upregulated in both strains (Table 2). Among these, the open reading frames (ORFs) SA0204, which encodes an NAD(P)H dehydrogenase homologue, and SA2311, which encodes a hypothetical protein similar to NAD(P)H flavin oxidoreductase, were upregulated only in strain 10*3d1.

MscL is a mechanosensitive channel with a large conductance that acts as an emergency valve in the bacterial cytoplasmic membrane by preventing excessive turgor pressure, which can cause the lysis of bacterial cells (32). This channel was already described as being involved in the calcium release induced by membrane depolarization in cyanobacterial cells (38). Strains 10*3d1 and 10*3d1-10 presented increased levels of transcription of mscL (3.36 and 4.58, respectively).

Transporters and cell wall-related genes.The expression of the following transporters was upregulated only in resistant strain 10*3d1. ORF SA0172, which encodes a membrane protein similar to LmrP, a multidrug transporter, was upregulated only in strain 10*3d1 (ratio = 2.32), as were the vraFG genes (Table 2).

Finally, we observed an increased level of transcription of the cell wall-related gene murI only in strain 10*3d1. None of the penicillin-binding proteins was altered in the strains tested in this study (data not shown). Almost all genes in the capsule operon were found to be upregulated only in strain 10*3d1-10 (Table 2).

Membrane potential and membrane fluidity.Strain 10*3d1 presented a higher membrane potential than the parental strain (fluorescence ratios = 10.52 and 7.19, respectively). The membrane potential of strain 10*3d1-10 decreased to 9.19 after it was cultured for 10 days in drug-free MHB, although it was still higher than that of strain N315ΔIP (Fig. 3). It is already known that gadolinium (III) chloride can inhibit MscL (18), and since this channel was upregulated in the S. aureus strains evaluated in our study, we inhibited MscL by adding 200 μM GdCl₃ solution to the culture. Studies of the viability of cells incubated in MHB with 200 μM GdCl₃ showed that this substance did not affect the total number of CFU per ml on MHA plates (data not shown). In the presence of 200 μM GdCl₃, strains N315ΔIP and 10*3d1 presented lower membrane potentials (fluorescence ratios = 1.75 and 2.03, respectively), although they were still higher than those in the presence of the proton ionophore CCCP (Fig. 3). Thus, we determined the daptomycin MICs of 10*3d1 and N315ΔIP strains on MHA in the presence of GdCl₃, but the daptomycin MICs did not change.

• Open in new tab
• Download powerpoint

FIG. 3.

Membrane potential determination. The membrane potential increases with the fluorescence ratio (green/red). Gd, gadolinium (III) chloride (final concentration, 200 μM); 10*3d1 10th, 10*3d1-10.

Jones et al. reported an increase in membrane fluidity in daptomycin-nonsusceptible strains (23). However, we did not observed any significant difference in membrane fluidity among strains 10*3d1, 10*3d1-10, and N315ΔIP. The PIs from six independent measurements at 35°C were 0.296 ± 0.022, 0.296 ± 0.043, and 0.276 ± 0.019 for N315ΔIP, 10*3d1-10, and 10*3d1, respectively.

Thickness of cell wall and autolysis assay.Although the autolysin gene was upregulated in strains 10*3d1 and 10*3d1-10, in comparison to the autolytic activity in N315ΔIP, these strains presented little increase in autolytic activity during 2 h of incubation (Fig. 4). We also measured the cell wall thicknesses of these strains in order to investigate if any change in the cell wall occurred by exposing the strain to daptomycin. Impressively, 10*3d1 had a very thick cell wall of 41.81 nm, while N315ΔIP presented a cell wall thickness of only 26.65 nm (Fig. 5). After 10 days of passage in drug-free medium, 10*3d1-10 showed a thinner (37.51 nm) cell wall than 10*3d1.

• Open in new tab
• Download powerpoint

FIG. 4.

Autolysis assay. Strain 209P (closed circles) was used as a control, since it has a high level of autolysin activity. Open circles, N315DIP; open squares, 10*3d1 10th; open triangles, 10*3d1-10.

• Open in new tab
• Download powerpoint

FIG. 5.

Representative transmission electron micrographs and cell wall thicknesses of the strains studied. The values given under each panel are the means (nm) and standard deviations of the cell wall thickness from 30 cells of each strain determined at a magnification of ×25,000.