Correspondence: Cong-Ming Wu, Beijing Key Laboratory of Detection Technology for Animal-Derived Food Safety, College of Veterinary Medicine, China Agricultural University, No. 2 West Yuanmingyuan Road, Beijing 100193, China. Tel.: +86 010 62733378; Fax: +86 010 62731032; e-mail: email@example.com
The polymorphic mutation frequencies for 154 Staphylococcus aureus isolates from Chinese bovine clinical mastitis cases were investigated. We found that nearly 29% of the isolates presented as weak mutators, while only two (1.3%) strong mutators were detected. Of the 15 weak mutators that exhibited ciprofloxacin resistance phenotypes, only one isolate was found to be mutS deficient. All of the ciprofloxacin-resistant isolates had the classic ciprofloxacin resistance mutations at codon 80 within the ParC subunit of topoisomerase IV and codon 84/88 within the GyrA subunit of DNA gyrase. The proportion of ciprofloxacin-resistant isolates among the weak mutators (34.1%) was significantly higher than that found in the normomutators (11.4%) and hypomutators (0%) (P < 0.001, Fisher's exact test), suggesting a positive correlation between weak mutators and ciprofloxacin resistance.
Increasing fluoroquinolone (FQ) resistance among populations of Staphylococcus aureus is an issue of major concern. Although several plasmid-mediated resistance determinants have been described (Robicsek et al., 2006; Kim et al., 2009; Strahilevitz et al., 2009; Wang et al., 2009), the great majority of the genetic alterations associated with such resistance are mutations in chromosomal genes. FQ resistance in S. aureus is primarily caused by mutations within the quinolone resistance-determining region (QRDR) in this species; such mutations include the GyrA and GyrB subunits of DNA gyrase and the ParC and ParE subunits of topoisomerase IV (Hooper, 2002; Schmitz et al., 2002). Thus, in bacteria, relatively increased mutation frequencies may be of importance for the evolution of drug resistance by increasing capacity to acquire newly arising mutations.
The focus of interest regarding the influence of mutators on bacterial evolution has been on those strongly hypermutable bacteria (Oliver et al., 2000; Giraud et al., 2002; Blazquez, 2003; Chopra et al., 2003; Macia et al., 2005), which has been proposed to be a major source of emergence of mutational antibiotic resistance (Giraud et al., 2002). However, as most mutations are neutral or deleterious, bacteria with hypermutable phenotypes probably possess tremendous evolutionary cost (Taddei et al., 1997; Denamur & Matic, 2006). Although several experimental and theoretical studies have indicated that the mutational load of the strong mutators makes them less competitive than the weak mutators in the long run (Taddei et al., 1997; Giraud et al., 2001; Shaver et al., 2002), unfortunately, less attention has been paid to weak mutators, which have only been reported in limited studies on Escherichia coli, Stenotrophomonas maltophilia, and very recently Salmonella enterica (Baquero et al., 2004; Le Gall et al., 2009; Turrientes et al., 2010). As yet, no information on weak mutators is available for S. aureus. Hence, this is the first study to describe the polymorphic mutation frequencies in S. aureus clinical isolates.
Materials and methods
A collection of 154 S. aureus isolates obtained from bovine clinical mastitis cases in China in 2005 were used in this study. Isolate identification was performed with the API ID32 Staph test kit (bioMerieux, Marcyl' Etoile, France) and conventional tests. The strains were also analyzed for their species-specific 23S rRNA gene using previously described methods (Cremonesi et al., 2005).
Determination of mutation frequencies
Mutation frequencies were investigated using the rifampin-supplemented culture method as described previously (Prunier et al., 2003; Hall & Henderson-Begg, 2006). Briefly, a single bacterial colony was suspended in 20 mL of trypticase soy broth (TSB: Luqiao, Beijing, China) and grown to a maximum viable cell count at 37 °C. The bacterial cells were collected at 450 g for 5 min and resuspended in 1 mL of TSB. A 100 μL sample from this cell suspension and successive dilutions of it were plated onto trypticase soy agar (TSA: Luqiao) plates, with and without rifampin (100 μg mL−1; China Institute of Veterinary Drug Control, Beijing, China). After 48-h incubation at 37 °C, colony counting was performed, and the mutation frequencies determined by dividing the total number of mutants on rifampin-supplemented agar by the total viable cell count on antibiotic-free agar. All experiments were performed in triplicate, and their mean values recorded.
Antimicrobial susceptibility testing
The susceptibilities of the S. aureus isolates to seven FQ drugs, including ciprofloxacin, ofloxacin, enrofloxacin, norfloxacin, lomefloxacin, difloxacin, and pefloxacin, were tested using the standardized broth microdilution method. All tests were performed and results evaluated according to the Clinical and Laboratory Standards Institute guidelines (CLSI) document M31-A3 (2008). For agents with no breakpoints recommended by the CLSI M31-A3, the results were evaluated according to CLSI M100-S21 (2011).
Analysis of the methyl mismatch repair and QRDR genes
Genomic DNA from the S. aureus isolates was extracted using a TIANamp bacteria DNA kit (Tiangen, Beijing, China). PCR assays for the mutS and mutL genes were performed using primers described previously (Prunier & Leclercq, 2005). Mutations in the QRDR genes gyrA, gyrB, parC, and parE were investigated as described by Trong et al. (2005). The DNA sequences obtained from the isolates were compared with those in GenBank using the blast program (http://blast.ncbi.nlm.nih.gov/).
Results and discussion
Figure 1 shows the overall distribution of the rifampin mutation frequencies in the 154 S. aureus isolates examined in this study, using the phenotype categories previously established for E. coli: hypomutators (f ≤ 8× 10−9), normomutators (8 × 10−9 < f < 4 × 10−8), weak mutators (4 × 10−8 ≤ f < 4 × 10−7), and strong mutators (f ≥ 4 × 10−7) (Baquero et al., 2004). A sharp modal peak in the frequency distribution was found at 1 × 10−8, which is the same as that previously observed with E. coli and S. maltophilia isolates (Baquero et al., 2004; Turrientes et al., 2010). In this case, the main peak of the normomutators accounted for 57.1% of the isolates. Nearly 29% of the isolates presented as weak mutators, whereas with the E. coli, S. maltophilia, and S. enterica isolates from various sources, the values ranged from 3.3% to 38%. In addition, hypomutators represented 13.0% of the S. aureus isolates, and only two (1.3%) strong mutators were detected. In contrast, our unpublished data indicated that the population of S. aureus from a bovine subclinical mastitis collection was considerably enriched with strong mutators (c. 15%).
Since defects in the methyl mismatch repair system are considered to be the major cause of hypermutation in bacterial isolates (Horst et al., 1999; Oliver et al., 2002; Chopra et al., 2003), we analyzed the entire mutS (2619 bp) and mutL (2010 bp) genes from the two strong mutators and 15 of the weak mutators that displayed ciprofloxacin-resistant phenotypes, using previously described primers (Prunier & Leclercq, 2005). Overall, three isolates showed alternations in their deduced MutS and/or MutL amino acid sequences, including two strong mutators and one weak mutator. A premature termination (V463Stop) in the C-terminal region of MutL due to a 1-bp deletion-caused frameshift was observed in the two strong mutators. Additionally, one of them harbored another six amino acid substitutions (N181H, N373D, T415M, G789D, V800I, and L811S) in MutS. Interestingly, three of these six amino acid substitutions (N181H, N373D and V800I) were shared with the only weak mutator that was found to be mutS deficient. These results confirm the role of the mutS and mutL genes in the strong hypermutability. However, the absence of defects in the mutS/mutL genes from most of the weak mutators suggests that other mutator genes might be responsible for the slightly increased mutation frequencies.
We sought to determine the FQ susceptibility profiles of the 154 S. aureus isolates for seven FQ-based compounds. The overall resistance values ranged from 6.5% for pefloxacin to 31.2% for norfloxacin, whereas 22.7%, 21.4%, 18.2%, 17.5%, and 13% of the isolates exhibited resistance to lomefloxacin, ofloxacin, enrofloxacin, ciprofloxacin, and difloxacin, respectively. Notably, none of these drugs (except ciprofloxacin and enrofloxacin) were used for the treatment of the studied bovine clinical mastitis cases; however, the 27 ciprofloxacin-resistant isolates all appeared to exhibit ofloxacin- and norfloxacin-resistant phenotypes, suggesting a high level of cross-resistance to FQ drugs.
Mutations in the gyrA, gyrB, parC, and parE genes were investigated in the 27 ciprofloxacin-resistant isolates. In addition, the mutations of these four genes in three isolates with the highest mutation frequencies (> 1 × 10−7) among those that exhibited ciprofloxacin-susceptible or intermediate phenotypes were sequenced. Thirteen different combinations of amino acid mutations in QRDR were identified in 28 of the isolates, including one of the ciprofloxacin-susceptible ones (Table 1). From these 28 isolates, 14 (50.0%) contained two mutations, 13 (46.4%) had more than three mutations, while only one contained a single mutation. In contrast with the previous notion that multiple mutations are associated with higher MICs for ciprofloxacin than single-point mutations (Wang et al., 1998), our study showed no such correlation. We found that all of the 27 ciprofloxacin-resistant isolates combined the classic mutations at codon 80 within ParC and codon 84/88 within GyrA that have been shown to play major roles in the development of FQ resistance (Schmitz et al., 1998). Interestingly, a single substitution (S80Y) within ParC was also identified in the susceptible isolate, despite a previous finding that single mutations in ParC appear to be sufficient to generate resistance to ciprofloxacin (Hooper, 2002). No amino acid mutations were found within ParE. With respect to ParC, a novel mutation (Y372N) was found, but its role, if any, in FQ resistance remains speculative.
Table 1. Amino acid alternations within QRDR and corresponding ciprofloxacin MICs for 154 isolates of Staphylococcus aureus
Amino acid alternation(s)
No. of isolates with the following ciprofloxacin MIC (μg mL−1)
A positive association between mutation frequencies and ciprofloxacin resistance in the isolates was observed in this study. Figure 2 shows that the weak mutators are evenly distributed across the full MIC range, while none of the hypomutators exhibit MICs > 2 μg mL−1. We found that five most resistant isolates (with ciprofloxacin MICs values of up to 128 μg mL−1) contained three weak and two strong mutators, showing relatively high mutation frequencies (Fig. 2a). Furthermore, as shown in Fig. 2b, the proportion of resistant isolates among the weak mutators (34.1%) was significantly higher than that found in the normomutators (11.4%) or hypomutators (0%) (P < 0.001, Fisher's exact test). Also, significantly higher MICs were obtained for the weak mutators than for the normomutators or hypomutators (P < 0.001, Mann–Whitney U-test). In contrast to our observations, Baquero et al. (2004) failed to find any significant association between mutator phenotypes and ciprofloxacin resistance in E. coli. Nevertheless, Orlen & Hughes (2006) demonstrated that even small increases in the E. coli mutation rate can have a positive and measurable effect on the evolution of FQ resistance in vitro.
Due to the evolutionary cost in terms of deleterious mutations, weak mutators might have longer persistence time and better chances of evolving antibiotic resistance than strong mutators in the long run. Weak mutators are, therefore, likely to be a continuing and important factor in the evolution and emergence of antibiotic resistance in bacteria, and further work is needed to gain clear understanding of this subject.
This work was supported by Chinese Special Fund for Agro-scientific Research in the Public Interest (201203040), Chinese Key Projects in the National Science & Technology Pillar Program during the 25-year Plan Period (2012BAK01B02) and National Natural Science Funds of China (21207123).