In this study, we reported unusual plasmid-mediated antimicrobial resistance among USA300 isolates causing invasive disease. The frequency of isolates with unusual resistance was low, ranging from 6.2% (i.e., clindamycin resistance) to less than 1% (i.e., gentamicin resistance, TMP-SMZ resistance, and doxycycline resistance). In the past, USA300 isolates were predominantly from community-associated infections, but data from the ABCs MRSA project and other studies have demonstrated that isolates of this lineage are found with increasing frequency in health care-associated infections (21
). It has been hypothesized that as these isolates move into the health care setting, they will acquire resistance determinants more commonly seen in health care-associated MRSA strains, such as USA100 (23
). In support of this hypothesis, our data indicate that a majority (84.5%) of invasive USA300 isolates with unusual resistance phenotypes were isolated from patients who had one or several contacts with the health care system. While only 15.5% of isolates with unusual resistance were from community-associated infections, the presentation of these isolates in the community could complicate appropriate empirical therapy for potential community-associated MRSA infections.
A well-described mechanism of resistance transfer in S. aureus
is the pSK41-like conjugative plasmid, and USA300 isolates with pSK41 plasmid-mediated resistance were reported previously (11
). These plasmids carried ermC
-mediated clindamycin resistance and mupA
-mediated mupirocin resistance, and they were identified in limited geographical areas (i.e., San Francisco, CA, and Boston, MA) (10
). In this study, unusual resistance determinants in USA300 were frequently associated with pSK41-specific genes traE
, and repA
. Specifically, pSK41-like genes were identified in 25 (35.2%) of the USA300 isolates with unusual resistance, particularly in those isolates that were positive for resistance determinants aac6
, and dfrA
, conferring resistance to gentamicin, mupirocin, and trimethoprim, respectively. Conjugation studies confirmed the location of these resistance determinants on a pSK41-like plasmid. These results suggest that pSK41 is an important vector for the dissemination of resistant determinants in USA300. Also, evidence of a pSK41-like plasmid was found in USA300 isolates from five of eight states where surveillance was conducted, indicating that these plasmids are more geographically widespread than was previously recognized.
We looked for evidence of identical pSK41-like plasmids in USA300 and USA100 isolates with similar resistance profiles from the same geographical location. This analysis was limited to gentamicin-resistant isolates, since gentamicin resistance in S. aureus is usually associated with hospital acquisition. Indistinguishable plasmids were identified both in USA100 isolates with different PFGE patterns and in USA300 isolates with different PFGE patterns in the same geographical location. In addition, indistinguishable plasmids were identified in a USA100 isolate and a USA300 isolate from the same geographical location. These results suggest horizontal plasmid transfer between isolates of the same lineage and between isolates of different lineages.
Not all of the unusual plasmid-mediated resistance determinants identified in this study were associated with pSK41-like plasmids. No pSK41-like plasmid was present in USA300 isolates for which the only unusual resistance was clindamycin. These isolates were positive for ermC
by PCR, and the gene was most likely carried on a small plasmid, as had been determined previously in clindamycin-resistant USA300 isolates (35
). Three USA300 isolates from health care-associated infections were doxycycline intermediate, tetracycline resistant, and tetM
positive. Two of these isolates were associated with pSK41-like plasmids and other resistance determinants, and one isolate was negative for pSK41-like genes. The location of tetM
was not investigated.
This is the first report of TMP-SMZ resistance in USA300 isolates. This is significant because TMP-SMZ therapy has been recommended for empirical treatment of infections originating in the community when coverage for MRSA is desired (6
). The mechanism and epidemiology of TMP-SMZ resistance were interesting. A known S. aureus
resistance determinant has been identified only for trimethoprim, dfrA
. No resistance determinant is known for resistance to sulfamethoxazole, but the mechanism is likely due to mutations in the chromosomal dihydropteroate synthase, which inhibits the modification of p
-aminobenzoic acid to dihydrofolate, blocking folic acid synthesis (15
). Several TMP-SMZ-susceptible isolates in this study were positive for dfrA
, a resistance gene carried by pSK41-like plasmids. This is of concern, because it signals the emergence of transferable trimethoprim resistance in a MRSA lineage that has proven to be a successful pathogen.
As the CDC database of MRSA strains has expanded, we have identified isolates with similarities to both USA300 and USA500. Five TMP-SMZ-resistant isolates, originally identified as USA300 by PFGE, were excluded from the study because they were PVL negative and did not contain SCCmec
IVa. In addition to being related to USA500 isolates by PFGE analysis, these isolates were also arcA
negative and dfrA
negative. USA500 isolates, primarily seen in health care-associated infections, are characteristically resistant to TMP-SMZ (23
) and generally negative for dfrA
, suggesting another mechanism of trimethoprim resistance.
In summary, this is the first report describing gentamicin, TMP-SMZ, and doxycycline resistance in USA300 isolates, although there have been reports of sporadic cases of gentamicin and doxycycline resistance in USA300 isolates (10
). We also describe USA300 isolates with resistance to clindamycin and mupirocin in more diverse geographical regions than have previously been reported (10
). These isolates are epidemiologically associated with health care exposures; as USA300 isolates emerge in health care settings, these isolates may acquire additional antimicrobial resistance, with pSK41-like plasmids being an important vector for transmitting resistance. If USA300 isolates become more resistant, as described here, empirical therapy for community-associated MRSA infections may become more complicated. These findings highlight the need for ongoing assessment of phenotypes as they relate to empirical treatment guidelines, as well as the need for evaluation of methods to minimize cross-transmission of resistance elements.