|Home | About | Journals | Submit | Contact Us | Français|
A community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) strain known as pulsed-field type USA300 (USA300) is epidemic in the United States. Previous comparative whole-genome sequencing studies demonstrated that there has been recent clonal emergence of a subset of USA300 isolates, which comprise the epidemic clone. Although the core genomes of these isolates are closely related, the level of diversity among USA300 plasmids was not resolved. Inasmuch as these plasmids might contribute to significant gene diversity among otherwise closely related USA300 isolates, we performed de novo sequencing of endogenous plasmids from 10 previously characterized USA300 clinical isolates obtained from different geographic locations in the United States. All isolates tested contained small (2- to 3-kb) and/or large (27- to 30-kb) plasmids. The large plasmids encoded heavy metal and/or antimicrobial resistance elements, including those that confer resistance to cadmium, bacitracin, macrolides, penicillin, kanamycin, and streptothricin, although all isolates were sensitive to minocycline, doxycycline, trimethoprim-sulfamethoxazole, vancomycin, teicoplanin, and linezolid. One of the USA300 isolates contained an archaic plasmid that encoded staphylococcal enterotoxins R, J, and P. Notably, the large plasmids (27 to 28 kb) from 8 USA300 isolates—those that comprise the epidemic USA300 clone—were virtually identical (99% identity) and similar to a large plasmid from strain USA300_TCH1516 (a previously sequenced USA300 strain from Houston, TX). These plasmids are largely divergent from the 37-kb plasmid of FPR3757, the first sequenced USA300 strain. The high level of plasmid sequence identity among the majority of closely related USA300 isolates is consistent with the recent clonal emergence hypothesis for USA300.
Community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) infections in the United States are largely caused by a strain known as pulsed-field gel electrophoresis type USA300 (USA300) (14, 17, 18). Using comparative whole-genome sequencing, we demonstrated previously that there has been clonal emergence of a subset of USA300 isolates, which comprise the epidemic USA300 clone (13). On average, these USA300 isolates differ by ~50 single-nucleotide polymorphisms (SNPs) in the core genome (13). Although there is limited nucleotide diversity in the core genome among these USA300 clinical isolates, comparative DNA sequencing failed to resolve plasmid DNA content, presumably because of significant nucleotide diversity.
Plasmids carry antibiotic resistance genes and virulence determinants and spread rapidly within and between species. The contribution of plasmids to the emergence of USA300 and clonal subtypes and their potential spread of antibiotic resistance has not been thoroughly investigated, and only a few extrachromosomal elements have been characterized. For example, strain FPR3757, the first USA300 strain for which there was a complete genome sequence, contains a 4-kb plasmid encoding tetracycline resistance that is 99.9% identical to a plasmid in Staphylococcus epidermidis (7). Strain FPR3757 also contains a conjugative 37-kb plasmid (pUSA03) that encodes resistance to mupirocin and high-level resistance to clindamycin, a front-line antibiotic used to treat CA-MRSA infections (7). A recent study by McDougal et al. identified pUSA03-like plasmids in USA300 isolates recovered from individuals with invasive disease (16). In comparison, strain USA300_TCH1516 has a 27-kb plasmid (pUSA300HOUMR) that is not related to pUSA03 but is a “mosaic” of plasmids from other staphylococcal species (11). These observations are consistent with the findings of Tenover et al. (24), who reported that plasmids among USA300 isolates varied in size and composition, although the analysis of nucleotide diversity measured was limited to the presence or absence of four selected antibiotic resistance genes.
Since plasmids are readily exchanged among S. aureus strains, we hypothesized that these mobile genetic elements confer significant diversity among closely related USA300 isolates. To test this hypothesis, we performed de novo DNA sequencing of plasmids from previously characterized USA300 clinical isolates.
The USA300 isolates used in this study (Table (Table1)1) were previously characterized by comparative whole-genome sequencing, mec typing, spa (protein A) typing, and multilocus sequence typing (13).
Plasmids were isolated from overnight cultures using the Qiagen Maxi prep kit (Valencia, CA) and quantified using the Quant-IT Picogreen double-stranded DNA quantification kit (Invitrogen, Carlsbad, CA). DNA and RNA contamination was visualized by agarose gel electrophoresis and ethidium bromide staining. Protein contamination was estimated by measuring the A260/A280 ratio with a SpectraMax 384 spectrophotometer. Plasmids were digested with EcoRI according to the manufacturer's instructions (Roche Diagnostics Corporation, Indianapolis, IN), resolved by agarose gel electrophoresis, and visualized with ethidium bromide.
Plasmid nucleotide sequences were determined with a 454 Life Sciences Genome Sequencer FLX (GS FLX) Titanium (Roche) at Agencourt Biosciences (Danvers, MA) following the manufacturer's recommended procedures. All plasmids were sequenced to a minimum of 10× coverage, and gap closure was performed by designing PCR primers to flanking regions and sequencing the PCR products by capillary Sanger sequencing. The sequencing reads were then assembled into a single contig using DNA Sequencher 4.7 (Gene Codes Corporation, Ann Arbor, MI). The origin of replication for each plasmid was identified using Oriloc software (http://pbil.univ-lyon1.fr/software/Oriloc/oriloc.html) (9), and complete plasmid sequences were annotated by Integrated Genomics (Chicago, IL).
Full-length plasmid sequences were aligned according to a multiple sequence alignment program, MAFFT v6.8.14b (12), and a rooted neighbor-joining method tree with 1,000 bootstraps was generated. For each phylogenetic analysis, the most distantly related strain was used as the outgroup. The numbers at nodes are the percentages of bootstraps that supported the branching pattern shown. The scale bar represents branch length as it relates to the number of substitutions per base position analyzed.
Broth microdilution (BMD) assays were performed in 96-well microtiter plates using serial 2-fold dilutions of antibiotic (0.125 to 64 μg/ml). Susceptibility and resistance levels were defined based on recommendations found in Clinical and Laboratory Standards Institute Supplement M100-S18 (4). Antibiotics were purchased from Sigma Aldrich (St. Louis, MO) unless otherwise stated. To confirm MIC results from BMD assays, analyses with Etest strips (AB Biodisk, Solna, Sweden) were performed for several antibiotics according to the manufacturer's instructions.
Resistance to trimethoprim-sulfamethoxazole, quinupristin-dalfopristin, and tigecycline was monitored by disk diffusion assays on Mueller-Hinton agar. All antibiotic disks were purchased from Fisher, LLC (Pittsburgh, PA). Erythromycin-inducible clindamycin resistance was tested using the double disk diffusion assay (8, 21, 22) and a microbroth assay (23) as previously described. In the disk diffusion assay, inducible clindamycin resistance was monitored as a flattening of the zone of inhibition adjacent to the erythromycin disk. Erythromycin and clindamycin disks were purchased from Remel (Lenexa, KS).
All plasmid sequences were deposited in the NCBI database with the following accession numbers: LAC-p01 (CP002150), LAC-p03 (CP002149), 18805-p01 (CP002133), 18805-p03 (CP002132), 18806-p03 (CP002134), 18807-p01 (CP002136), 18807-p03 (CP002135), 18808-p01 (CP002138), 18808-p03 (CP002137), 18809-p01 (CP002140), 18809-p03 (CP002139), 18810-p01 (CP002142), 18810-p03 (CP002141), 18811-p01 (CP002144), 18811-p03 (CP002143), 18813-p03 (CP002145), 18813-p04 (CP002146), 19321-p01 (CP002148), and 19321-p03 (CP002147).
All USA300 isolates tested contained endogenous plasmids, including at least one large plasmid (27 to 37 kb) (Table (Table1;1; see Fig. S1 in the supplemental material), in accordance with the known plasmid contents of strains FPR3757 and USA300TCH_1516 (7, 11). Based upon agarose gel electrophoresis, there were at least four different plasmid profiles (see Fig. S1 in the supplemental material), indicating that nucleotide sequence diversity existed among plasmids from closely related strains (e.g., FPR3757 and LAC).
As a step toward understanding the extent of USA300 plasmid diversity, we determined the complete nucleotide sequences of all plasmids from a well-characterized group of USA300 clinical isolates (Table (Table1).1). All isolates except 18806 and 18813 had a small 2- to 3-kb plasmid. These small plasmids were nearly identical to pUSA01, a 3.1-kb plasmid of FPR3757 (there were 3 unique plasmid SNPs among all strains) (Fig. (Fig.11 C and D) (7). Strain 19321 contained a 2.4-kb plasmid (19321-p01) 99.7% identical to pWBG738, a plasmid from a multilocus sequence type (MLST) 1 community-associated S. aureus isolate collected from Western Australia (19) (Fig. 1B and C). Plasmid 19321-p01 had three open reading frames (ORFs), including ermC, which encodes resistance to macrolide, lincosamide, and streptogramin antibiotics (Fig. (Fig.1B).1B). As such, strain 19321 was resistant to clindamycin (Table (Table22 ), an antimicrobial agent commonly used to treat CA-MRSA infections (6). Interestingly, plasmid 19321-p01 has 99% sequence identity to the segment of pUSA03 that contains a repL homologue (encoding a plasmid replication protein) and ermC and is located between two transposon elements. This plasmid may be an important means for the acquisition of antibiotic resistance elements by USA300. pWBG738 and pUSA01 have no common DNA elements and thus share little sequence similarity (Fig. 1A and B).
Large plasmids (~27 kb) from eight USA300 isolates—those that comprise the epidemic USA300 clone—were essentially identical (99% sequence identity) and also similar to pUSA300HOUMR (11) (Fig. (Fig.22 and and3).3). These plasmids encode several heavy metal and/or antibiotic resistance elements, including resistance to Cd2+, bacitracin, macrolides, penicillin, kanamycin, and streptothricin (Fig. (Fig.2A2A).
By comparison, large plasmids from strain FPR3757 (pUSA03) (7), 18813 (18813-p03 and 18813-p04) (13), 19321 (19321-p03) (13), USA300 PAnicu (pSAP082A) (16), USA300 GA (pSAP079A) (16), and USA300 TN (pSAP080A) (16) were divergent from those of the other USA300 isolates (Fig. (Fig.22 and and3).3). Strain 18813 was the only isolate that had two large plasmids (18813-p04 and 18813-p03) (Table (Table11 and Fig. Fig.2).2). Plasmid 18813-p03 was 97% identical to pUSA03 but did not contain an ileS (whuch encodes isoleucyl-tRNA synthetase) or ermC gene (Fig. (Fig.2A).2A). As with pUSA03, 18813-p03 encoded the conjugative transfer elements (tra genes) necessary for conjugal transfer of DNA. Plasmid 18813-p04 was a chimera of pN315/pMW2 and plasmids integrated within staphylococcal cassette chromosome (SCC) mec type III (SCCmecIII) of S. aureus strains V14 and 85/2082 (Fig. (Fig.2B).2B). This plasmid contained a cadD-cadX operon (encoding cadmium resistance protein and regulator) and a β-lactamase (bla) resistance cassette also present in pN315 and pMW2 (Fig. (Fig.2B)2B) (1, 15). Plasmid 18813-p04 had several elements that confer resistance to heavy metals (e.g., Cd2+, Hg2+, As2+, and Zn2+) (Fig. (Fig.2).2). These elements are found in plasmids integrated into SCCmecIII of strains V14 and 85/8082 (Fig. (Fig.2B).2B). The mer operon and acr3 (which encodes an arsenical-resistance protein) were flanked by a transposon and a gene encoding a recombinase, respectively, suggesting that a recombination event resulted in the mosaic sequence of the plasmid.
Plasmid 19321-p03 was >99% identical to pSK67 (27,420 out of 27,439 bases were identical, with 16 1-base gaps), an orphan family Bla/heavy metal resistance plasmid from an archaic S. aureus isolate collected from a Melbourne, Australia, hospital in 1949 (10). These observations are in accordance with comparative whole-genome sequence analysis, which indicates 19321 is divergent from the other isolates comprising the epidemic USA300 clone (13). We also discovered that 19321-p03 encoded putative staphylococcal enterotoxins J, P (or D), and R (Fig. (Fig.2A).2A). These enterotoxin genes have been reported previously in plasmids of other S. aureus strains (2, 26).
To determine whether the antibiotic resistance elements present in plasmids contributed to the resistance patterns of each USA300 isolate, we determined the MIC levels of several clinically relevant antibiotics toward all isolates by using broth microdilution assays (Table (Table2).2). As expected, all isolates were resistant to oxacillin (MIC ≥ 16 μg/ml) and erythromycin, although the MIC of erythromycin toward 18813 was considerably lower than that of the other 10 isolates (8 μg/ml for 18813 versus ≥64 μg/ml for all others) (Table (Table2).2). Consistent with these results, plasmids from isolate 18813 lacked macrolide resistance elements (Fig. (Fig.2A).2A). FPR3757 and 19321 were constitutively resistant to clindamycin, likely due to ermC present in pUSA03 and 19321-p01, respectively, and none of the isolates had erythromycin-inducible clindamycin resistance (Table (Table2).2). Strain FPR3757 was the only organism resistant to tetracycline, presumably conferred by the tetracycline resistance element on pUSA02 (MIC ≥ 64 μg/ml for FPR3757 and < 0.125 to 4 μg/ml for all other strains) (Table (Table2)2) (7). However, all isolates were susceptible to minocycline and doxycycline, two antibiotics in the tetracycline family used to treat CA-MRSA infections (Table (Table2).2). All USA300 isolates were susceptible to trimethoprim-sulfamethoxazole, vancomycin, teicoplanin, tigecycline, and linezolid, antimicrobial agents used as frontline therapy for CA-MRSA infection (Table (Table2).2). Strain FPR3757 was the only organism resistant to mupirocin, as pUSA03 contains the ileS gene, which encodes high-level mupirocin resistance (7). Additionally, all isolates were susceptible to rifampin and fusidic acid, a current therapeutic compound used in Europe for the treatment of MRSA infections.
USA300 infections primarily affect skin and soft tissue, but the pathogen can cause severe invasive syndromes, such as osteomyelitis, sepsis, or pneumonia (reviewed in references 3, 5, and 6). Inasmuch as plasmids provide a relatively efficient means for staphylococci to exchange genetic information, comprehensive knowledge of plasmid content among USA300 isolates is critical for a full understanding of pathogen diversity and virulence and the emergence and spread of antibiotic resistance.
Complete nucleotide sequences of plasmids from at least six USA300 isolates are available for comparative analysis, and these plasmids have varied gene contents (7, 11, 16). Given the relative ease with which plasmids are exchanged among staphylococci, there is potential for these mobile elements to contribute significantly to genetic diversity among USA300 strains. Indeed, two plasmids reported by McDougal et al. (16), which were recovered from USA300 isolates with unusual antimicrobial resistance profiles, are similar to pUSA03 (7, 16) and 18813-p03 (Fig. (Fig.3B)3B) but relatively divergent from the earlier USA300 isolates that comprise the epidemic clone (except for FPR3757) (Fig. (Fig.3B).3B). Such plasmid diversity is likely explained by varied selective pressures imparted to the parent strain. For example, acquisition of plasmid-mediated antibiotic resistance in USA300 is far more common in health care-associated isolates than in those from the community (16).
Several USA300 isolates—LAC, USA300_TCH1516, and 18805 to 18811—contained essentially identical plasmids, a finding consistent with the high level of similarity among the core genome sequences of these isolates (11, 13). These data provide further support for the idea that there has been recent clonal emergence of the epidemic USA300 strain. Given these findings, the large plasmid and/or a subset of genes contained within it could be used for molecular epidemiology studies. Whether elements encoded by one or more of the plasmids described here contributed to the emergence or epidemicity of the USA300 clone is unknown. None of the genes encoding proteins with known or putative functions appear uniquely suited to promote the success of USA300. However, there are plasmid open reading frames that encode hypothetical proteins with uncharacterized functions (Fig. (Fig.2).2). Nevertheless, the core genomes of FPR3757 and LAC are virtually identical, but these isolates contain fairly divergent large plasmids (Fig. (Fig.2A2A and and3),3), an observation that suggests the large plasmid is not key to the emergence and epidemicity of USA300.
In addition to facilitating the acquisition of antibiotic resistance elements, plasmids potentially provide an efficient means to transfer virulence factor genes among staphylococci. For example, exfoliative toxin B, a causative agent of staphylococcal scalded skin syndrome, is plasmid encoded (25). Interestingly, plasmid 19321-p03 contained genes encoding SER, SEJ, and SEP (or D) (Fig. (Fig.2A).2A). The enterotoxin genes sed, sej, and ser are carried on plasmids from Staphylococcus aureus isolates associated with outbreaks of food poisoning (2, 20, 26). The contribution of the plasmid-encoded enterotoxins to USA300 pathogenesis, if any, remains to be determined.
This work was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.
The authors declare no conflicts of interest.
Published ahead of print on 13 October 2010.
†Supplemental material for this article may be found at http://jcm.asm.org/.