Molecular characterization of resistance to Rifampicin in an emerging hospital-associated Methicillin-resistant Staphylococcus aureus clone ST228, Spain

Background Methicillin-resistant S. aureus (MRSA) has been endemic in Hospital Universitari de Bellvitge, Barcelona, since 1990. During the 1990-95 period the Iberian clone (ST-247; SCCmec-I) was dominant. Isolates of clonal complex 5 (ST-125; SCCmec-IV) gradually replaced the Iberian clone from 1996 to 2003. A new multiresistant MRSA phenotype showing rifampicin resistance emerged in 2004 and rapidly increased from 25% in 2004 to 45% in 2006. The aims of this study were i) the molecular characterisation of rifampicin resistant MRSA isolates, ii) the study of the rifampicin resistance expression by disk diffusion, microdilution and E-test, and iii) the analysis of the rpoB gene mutations involved in rifampicin resistance. Results A sample of representative 108 rifampicin-resistant MRSA isolates belonged to a single PFGE genotype, ST-228, SCCmec type I and spa type t041. Of 108 isolates, 104 (96%) had a low-level rifampicin resistance (MICs, 2 to 4 mg/L) and 4 a high-level rifampicin resistance (MICs, 128 - ≥ 256 mg/L). Disk diffusion and E-test methods failed to identify a low-level rifampicin resistance in 20 and 12 isolates, respectively. A low-level rifampicin resistance was associated with amino acid substitution 481His/Asn in the beta-subunit of RNA polymerase. Isolates with a high-level rifampicin resistance carried additional mutations in the rpoB gene. Conclusions The emergence of MRSA clone ST228-SCCmecI, related to the Southern Germany clone, involved a therapeutical challenge for treating serious MRSA infections. Decreased susceptibility to rifampicin in MRSA strains of ST228-SCCmecI was associated with one or two specific mutations in the rpoB gene. One fifth of isolates with low-level rifampicin-resistance were missed by the diffusion methods.


Background
Methicillin resistant Staphylococcus aureus (MRSA) is an important pathogen in Spanish hospitals. The percentage of patients infected or colonised by MRSA among patients with nosocomial S. aureus has been estimated between 20.2% and 30.5% in nation-wide multicenter studies [1,2].
In the Hospital Universitari de Bellvitge MRSA has been endemic since 1990. The majority of strains isolated during the 1990-95 period belonged to the multiresistant Iberian clone. By multilocus sequence typing (MLST) the Iberian clone showed an allelic profile or sequence type (ST) 247, carrying the staphylococcal cassette chromosome mec (SCCmec) type I [3]. Isolates of the Iberian clone exhibited resistance against almost all antibiotics available for MRSA therapy including clindamycin, erythromycin, gentamicin, tobramycin, tetracycline, ciprofloxacin and rifampicin. From 1996 to 2003, the Iberian clone was gradually replaced by isolates of Clonal Complex 5 (ST125 and variants; SCCmec type IV) related to the Paediatric clone (ST5; SCCmec type IV) [4]. Unlike the Iberian clone, these strains showed only consistent resistance to tobramycin and ciprofloxacin combined with variable resistance to clindamycin and/or erythromycin. Similar trends have been observed in other hospitals in Spain and in other countries such as France, Germany, Belgium or Portugal, with involvement of different clonal lineages [5][6][7][8][9][10]. MRSA isolates resistant to clindamycin, erythromycin, gentamicin, tobramycin, and ciprofloxacin were detected in 2004. These isolates showed reduced susceptibility to rifampicin (RIF-R), according to the Clinical and Laboratory Standards Institute (CLSI) criteria [11]. This new phenotype of multiresistance differed from that of the Iberian clone on the low level RIF-R and on the tetracycline susceptibility. The frequency of the RIF-R MRSA iso- Rifampicin cannot be used as a single agent to treat MRSA infections because of the rapid selection of resistant mutants [12,13]. However, combinations of rifampicin with other anti-staphylococcal agents such as quinolones [14] or fusidic acid [15] could prevent the emergence of rifampicin resistance during therapy [16]. Rifampicin interacts specifically with the RNA polymerase beta-subunit encoded by the gene rpoB [12]. Rifampicin resistance in S. aureus, as in other bacteria, is associated with mutations in particular regions (cluster I and II) of the gene rpoB [13,17].
The objectives of the present study were: i) to characterise a collection of MRSA isolates expressing this new multiresistant pattern, and to determine whether they represented a novel genotype or they were the current representatives of a previously detected clone, ii) to determine the different levels of the rifampicin resistance by disk diffusion, microdilution and E-test, and iii) to analyse mutations in the rpoB gene related to rifampicin resistance.

Hospital setting
The Hospital Universitari de Bellvitge in Barcelona, Spain, is a nearly 900-bed tertiary care teaching centre. It is the reference hospital for a geographical area with a population of approximately 1 million inhabitants. The Hospital Universitari de Bellvitge in Barcelona provides medical and surgical care for adult patients with an average of 25,000 admissions per year. It has six intensive care units with a total of 60 beds and an active organ transplant program.
The control of MRSA in our institution is based on the active screening of patients at risk and contact isolation of infected or colonised patients. The origin of the strains was from blood cultures or catheter-related sites (n = 38), wound swabs (n = 28), respiratory samples (n = 24), exudates (n = 12), nasal swabs (n = 4) and sterile fluids (n = 2). Oral informed consent was given by all patients before taking the clinical specimen. The patient acquisition of MRSA infection or colonisation was prospectively assessed. Five strains with the same resistance pattern but fully susceptible to rifampicin (RIF-S) (MIC 0.012 mg/L) were included in this study. This RIF-S pattern represented about 4% of all MRSA isolated between 2004 and 2006.

Antimicrobial susceptibility testing
Susceptibility testing of primary MRSA isolates is performed routinely by the disk-diffusion method on Mueller-Hinton 2 agar plates (MH2, bioMérieux) to the following antibiotics: penicillin ( recommendations [11], and by E-test (AB biodisk, Solna, Sweden). Isolates were interpreted as susceptible or resistant, according to the CLSI criteria [11].

Determination of spontaneous mutation frequency for rifampicin resistance
The determination of spontaneous mutation frequency for rifampicin resistance was aimed at identifying whether the presence of a first mutation conferring low level rifampicin resistance facilitated the acquisition of supplementary mutations responsible for increasing rifampicin MICs. The rifampicin mutation frequency was calculated in reference strain ATCC700698 (MIC 0.006 mg/L) and in two RIF-R MRSA strains carrying the low level resistance mutation His481/Asn (rifampicin MICs of 1.5 and 2 mg/L, respectively). Bacterial strains were cultured in a shaking incubator at 37°C in Luria-Bertani (LB) broth (BD Diagnostics, Spain) until the exponential growth phase. Each strain was plated on the selective and non-selective LB agar plates and incubated at 37°C. Rifampicin selecting concentrations were 2 and 20 mg/L for the reference strain, and 20 mg/L for the RIF-R MRSA strains. In these experimental conditions OD 620 = 0.125 corresponded to 5 × 10 7 cfu/ml. The equivalent to 10 7 , 10 8 and 10 9 cfu were spread on selective plates, and appropriated diluted samples were plated on non-selective plates. After 24 h to 36 h, colonies that grew on selective and non-selective plates were counted and mutation frequencies were calculated. Three independent experiments were performed to ensure reproducibility.

Molecular typing
Pulsed Field Gel Electrophoresis (PFGE) was performed after SmaI restriction of chromosomal DNA according to Chung et al. [20]. Pulses run from 5 s to 15 s for 10 h for block 1, and from 15 s to 60 s for 13 h for block 2 [21]. Isolates with PFGE patterns differing in four or less restriction fragments were considered to be subtypes of a single genotype. Isolates with differences in more than four fragments were ascribed to distinct genotypes [22].

Rifampicin resistance levels and associated rpoB mutations
The majority (n = 104, 96%) of the 108 RIF-R MRSA isolates, showed rifampicin MICs between 2 and 4 mg/ L. Two isolates had rifampicin MICs of 128 mg/L and the remaining two had MICs ≥ 256 mg/L. Corresponding E-test and disk diffusion results are shown in table 1. On the basis of these results and following other authors' categorisation [13,17,27] the strains were classified into categories of rifampicin susceptible (MICs, ≤ 0.5 mg/L), low-level rifampicin resistance (MICs, 1 to 4 mg/L), and high-level rifampicin resistance (MICs, ≥ 8 mg/L). Interestingly, 20 strains with rifampicin MICs of 2 mg/L showed inhibition zones between 20 and 23 mm, borderline to the susceptible CLSI breakpoint (inhibition zones ≥ 20 mm). The five RIF-S MRSA isolates, with the same multi-resistance pattern, had rifampicin MICs of 0.012 mg/L and inhibition zones > 30 mm.
The mutations in the rifampicin resistance-determining region of rpoB gene were studied in 32 RIF-R and in 5 RIF-S MRSA strains. Results are shown in table 2. All 32 strains presented the mutational change 481His/Asn, determined by a mutation in cluster I of rpoB gene, conferring a low-level rifampicin resistance. The four isolates with MIC≥ 128 mg/L had an additional amino acid substitution: 468Gln/Lys (n = 1), 477Ala/Thr (n = 2) or 527Ile/Leu (n = 1), associated with a high level rifampicin resistance. Mutational changes 468 and 477 were determined by mutations located in cluster I and substitution 527 was determined by a mutation located in cluster II. RIF-S MRSA isolates, had no mutations related to rifampicin resistance. All isolates, including RIF-S isolates, and 3 (ATCC29213, ATCCBAA44, and PER88) out of 4 control strains, presented a silent mutation in amino acid 498 with the substitution Ala(GCG) per Ala(GCT).

Frequency of spontaneous mutation for rifampicin resistance
The rifampicin mutation frequency was calculated in reference strain ATCC700698 (MIC, 0.006 mg/L) and in two RIF-R MRSA strains carrying the low-level resistant amino acid substitution 481His/Asn (rifampicin MICs, 1.5 and 2 mg/L, respectively). Rifampicin high level resistant mutants occurred with frequencies of around 10 -7 to 10 -8 in the RIF-R MRSA strains after selection by rifampicin concentration of 20 mg/L. An identical mutational ratio was found in the control strain ATCC700698 at both selective concentrations (2 and 20 mg/L).

Discussion
The multi-resistant nature of most MRSA clones found in hospitals represents a therapeutical challenge for treating serious MRSA infections. The burden that the Iberian clone posed in Spanish hospitals in the early 90 s [3,28], shifted to other clones susceptible to more antibiotics, which have been dominant in recent years [8,29]. In this paper, we described the emergence and spread of a MRSA clone resistant to clindamycin, erythromycin, gentamicin, tobramycin, ciprofloxacin and rifampicin which has reduced substantially the number of effective antibiotics for treatment of serious MRSA infections.
Rifampicin is an antibiotic of substantial interest in the rise of MRSA infections, but cannot be used as a single agent to treat such infections because rapid emergence of resistance can occur, even during therapy [16,30,31]. Although the role of rifampicin as adjunctive therapy is controversial [31], the combined therapy seems beneficial as long as the bacteria exhibit susceptibility to the antibiotics combined [30]. The distinctive phenotypic feature in the particular clone of ST-228 described here was the borderline resistance to rifampicin that could be missed by some methods of antimicrobial susceptibility testing (i.e. disk diffusion or E-test). Hence our interest in studying whether this low-level RIF-R was an adaptive phenomenon or to the contrary, known rpoB mutations underlay such phenotype.
Almost all isolates belonging to this multi-resistant MRSA clone (104/108) showed a low-level rifampicin resistance (MICs, 1 to 4 mg/L) and carried the amino acid substitution 481His/Asn in the RNA polymerase. Only 4 isolates showed additional substitutions known to be involved in a high-level rifampicin resistance: two isolates (MICs, 128 mg/L) carried mutational change 477Ala/Thr, and one isolate (MIC, ≥ 256 mg/L) 468Gln/Lys [13,17,27,32]. The fourth isolate (MIC, ≥ 256 mg/L) showed substitution 527Ile/Leu, the only one which mutation was found in the rifampicin resistance-determining cluster II, described recently among Japanese MRSA isolates [32]. It is noteworthy that 20 isolates (19%) of the RIF-R isolates, carrying rpoB mutation resulting in amino acid substitution in position 481, were detected as rifampicin susceptible by the diskdiffusion test. However, the inhibition zones of these strains were between 20 and 23 mm, closer to the susceptibility breakpoint established by CLSI (susceptibility ≥ 20 mm) than inhibition zones among RIF-S MRSA isolates that were usually ≥ 30 mm. Therefore, if screening for rifampicin resistance is made only by disk diffusion, special attention needs to be paid to strains borderline to the CLSI susceptibility breakpoint to avoid reporting false susceptibility results. MICs by E-test failed to detect rifampicin resistance following CLSI guidelines [11] in a group of 12 strains (MICs, 0.75-1 mg/L). These isolates showed MICs by microdilution of 2 mg/L and carried the rpoB mutation responsible for amino acid substitution in position 481. Thus, and according to other authors, it would be advisable to apply ≤ 0.5 and ≥ 8 mg/L as new breakpoints to classify rifampicin susceptibility or resistance in S. aureus clinical isolates [13,17].
High-level rifampicin resistance could be attributable to double mutations within rpoB, as previously described [27]. We did not find in this particular clone that the presence of a prior mutational change (481His/Asn) increased the frequency of acquisition of additional mutations responsible for a higher level of rifampicin resistance, when compared to a reference strain.
The multi-resistance pattern exhibited by the Iberian clone, dominant lineage in our hospital during the 90's, also included resistance to tetracycline. The new RIF-R MRSA isolates were resistant to clindamycin, erythromycin, gentamicin, tobramycin, ciprofloxacin and susceptible to tetracycline. However, molecular typing showed that the Iberian clone and the new RIF-R MRSA clone had different genetic backgrounds represented by ST-247 and ST-228, respectively, with only a single locus in common. Although both clones carried a SCCmec element type I, PFGE patterns and spa-types were clearly different.
All strains with the multi-resistant phenotype described in this work, showing resistance or decreased susceptibility to rifampicin, belonged to ST-228, carried a SCCmec element type I and were spa-type t041. This clone seems to be related to the Southern Germany clone (ST-228, SCCmec type I, spa-type t001 or spatype t041) reported in Germany in 1997-98 [21,33]. In the same period, strains of ST-228 and SCCmec type I were reported at several hospitals located in seven Italian cities [34], although these isolates also showed resistance to multiple antibiotics, rifampicin resistance was not stated. Recently, strains of ST-228 have spread epidemically in Finland in 2002-2004 and in Hungary in  [9,29,38]. However, none of these reports (from Spain or other countries) analysed the decreased susceptibility to rifampicin among representative strains of ST-228. During the 2004-2007 period, we did not find significant changes in the rifampin consumption in our institution, which was on average 0.5 DDD/100 patients-days for intravenous and 1.0 DDD/100 patientsdays for oral administration. A set of 5 strains resistant to clindamycin, erythromycin, gentamicin, tobramycin, ciprofloxacin, but fully susceptible to rifampicin with MICs of 0.012 mg/L were included in this study. On average, this RIF-S pattern represented 4% of all MRSA isolated between 2004 and 2006, however this resistance phenotype can be traced back to 1999 in our hospital. The RIF-S isolates were classified as ST-228, the same as the RIF-R MRSA. Isolates of ST-228 (MLST, arcc 1, aroe 4, glpf 1, gmk 4, pta 12, tpi 24, and yqi 29) belong to the Clonal Complex 5, as well as isolates of ST-125 (MLST, arcc 1, aroe 4, glpf 1, gmk 4, pta 12, tpi 1, and yqi 54) which was the dominant MRSA clone in Hospital Universitari de Bellvitge from 1996 to 2003. Consequently, an alternative hypothesis to explain the emergence of multiresistant clone ST-228, SCCmec type I would be the SCCmec type I transfer from ST-247 to the ST-125 background, with an intermediate stage where RIF-S isolates could be found belonging to ST-228.

Conclusion
Our study describes the hospitalary spread of an MRSA clone (ST-228, SCCmec-I, spa-t041), related to the Southern-Germany clone (ST-228, SCCmec type I, spatype t001 or spa-type t041) [21,33]. In this particular case, the studied strains were resistant to many more antibiotics than any previous MRSA clone spread in our institution, with the exception of the Iberian clone. In addition, the study of the rpoB mutations demonstrated that rifampin was not a suitable option for treatment of infections caused by this clone.