Introduction

Staphylococcus aureus is an important pathogen which is able to develop resistance to multiple antimicrobial agents. This microorganism produces large amounts of virulence factors, and some of them can be classified as leukocidins [as Panton-Valentine leukocidin (PVL)], pyrogenic toxin superantigens ( PTSAgs), haemolysins, or exfoliatins [1].

In 1961, an important public health problem arose when the first methicillin-resistant S. aureus (MRSA) strain emerged. The mecA gene, which encodes a modified penicillin-binding protein with low affinity to beta-lactams (PBP2a), confers resistance to methicillin but also to most other beta-lactams that leads to a serious therapeutic problem. Additionally, studies carried out in methicillin-susceptible S. aureus (MSSA) and MRSA have shown higher diversity of clonal lineages among MSSA in relation to MRSA. Thus, according to the results obtained with different molecular typing methods, it has been observed that some MRSA clones tend to occur more in certain countries or areas [2, 3] while other clones are disseminated worldwide, threatening public health [4]. These typing methods allow us to know more about the epidemiology, evolution, and spread of this microorganism.

A differentiation between hospital-acquired (HA) and community-acquired (CA) MRSA was initially proposed. However, these distinctions have diminished and today it is often difficult to distinguish them. Recent studies have further shown that some HA-MRSA clonal lineages have been superseded by CA-MRSA clones [4]. In addition, new associated lineages in farm animals, known as livestock-associated (LA) MRSA, have also been found in healthcare settings [5].

Several surveillance studies have been conducted in European countries to determine the predominant MRSA clones. So far, the data obtained in Spain indicate that, while the Iberian clone (CC247-ST247-SCCmecI-agrI) was the dominant one in our country in the 1990s, a derived clone of pediatric origin (CC5-ST125-SCCmecIV-agrII) is the one the most commonly found in the 21st century [6–9]. The aim of our study was to identify circulating MRSA clones in a Spanish hospital, and to determine their changes, studying two periods distant in time. Besides the genetic lineage diversity, antibiotic resistance mechanisms and virulence traits were also analyzed to understand the evolution of MRSA in our country.

Materials and methods

Selection of isolates

A total of 204 MRSA isolates were included in this study. These isolates were recovered from different patients in San Pedro Hospital of La Rioja (Northern Spain) in two periods. One hundred and three of these isolates were obtained from January to June 2001, and the remaining 101 from January to April 2009. The criterion of choice was to study approximately the first 100 MRSA isolates causing any type of infection in each of these years. In the first period, 29 % of the obtained S. aureus were MRSA and in the second period the proportion was 27 %. The collection included isolates from different origins such as wounds, nasal fluid, urine, sputum, blood, bronchial and tracheal aspirates, abdominal and pleural fluids, ears, and catheters.

Confirmation of species and mecA gene detection

All MRSA isolates were typed by the Microscan system (Microscan, Alphen aan den Rijn, The Netherlands) in the hospital and were subsequently confirmed by a specific duplex polymerase chain reaction (PCR) of the nuc gene (S. aureus thermonuclease gene) and the mecA gene (methicillin-resistance gene) [10].

Molecular typing

Single-locus DNA sequencing of S. aureus protein A (spa) [11] was carried out following standard methodology and the obtained sequences were analyzed by Ridom Staph-Type software version 1.5.21 (Ridom, Würzburg, Germany). Determination of agr types was performed by two multiplex PCRs as described elsewhere [12]. SCCmec-typing was carried out by multiplex PCRs using sets of region-specific primers, as previously described [13]. Identification of the new recently reported variant of SCCmec IV (IVNv) was performed by PCR [14]. Multilocus sequence typing (MLST) was implemented by PCR and sequencing in 11 selected MRSA isolates (one isolate of each spa-type detected) (http://www.saureus.mlst.net). The clonal complex (CC) of the isolates was assigned according to the ST determined (in 11 isolates) or according to the spa-type detected (in the remaining 193 isolates) [15]. Pulsed field gel electrophoresis (PFGE) with the SmaI enzyme was performed in 18 of the MRSA t067 isolates and in one MRSA isolate of each of the other spa-types detected [16].

Antimicrobial susceptibility testing

Susceptibility testing for 17 antimicrobial agents was carried out by the disk-diffusion method [17]. The antibiotics tested were as follows (in μg/disk): penicillin (10 U), oxacillin (1), cefoxitin (30), erythromycin (15), clindamycin (2), lincomycin (15), gentamicin (10), kanamycin (30), tobramycin (10), tetracycline (30), ciprofloxacin (5), chloramphenicol (30), trimethoprim-sulfamethoxazole (1.25 + 23.75), vancomycin (30), teicoplanin (30), mupirocin (5 and 200), and linezolid (30). Clinical and Laboratory Standards Institute (CLSI) breakpoints were used for all antimicrobial agents, except for lincomycin, for which the breakpoints recommended by the Société Française de Microbiologie were considered (http://www.sfm.asso.fr).

Resistance genotypes

The presence of the erm(A), erm(B), erm(C), erm(F), erm(T), msr(A)/msr(B), mph(C), lnu(A), vga(A), vga(C), lsa(C), tet(K), tet(L), tet(M), tet(O), aac(6′)-Ie–aph(2′′)-Ia, ant(4′)-Ia, aph(3′)-IIIa, dfrS1, dfrD, dfrG, dfrK, mupA, fexA, cfr, cat(pC194), cat(pC221), and cat(pC223) resistance genes was tested by PCR [18, 19]. Mutations in quinolone targets were determined by sequence analysis of the grlA and gyrA genes in the 13 quinolone-resistant MRSA isolates from blood samples [10].

Virulence factors

The presence of the genes encoding PVL (lukS/F-PV), toxic shock syndrome toxin (TSST)-1 (tst-1), exfoliative toxins A, B, and D (eta, etb, etd), and haemolysins alpha-, beta-, delta-, gamma- and gamma-variants (hla, hlb, hld, hlg, hlg v ) was determined by PCR [1, 10, 20]. The 18 genes that encode staphylococcal enterotoxins (SEs) (sea, seb, sec, sed, see, seg, seh, sei, sej, sek, sel, sem, sen, seo, sep, seq, ser, and seu) were investigated by three multiplex PCRs [21].

Results

Molecular typing of MRSA isolates

The occurrence of CC-agr-SCCmec and the spa-types detected in each period are shown in Fig. 1. The most frequent spa-type detected in both periods was t067, being more prevalent in 2001 (93 %) than in 2009 (71 %). Another detected spa-type in both periods was t008, although an important increase in its frequency was observed in 2009 in respect to 2001 (16 and 1 %, respectively) (Fig. 1). The spa-types detected only in the MRSA isolates of 2001 were t012, t051, t109, and t1318 (6 %) and in the MRSA isolates of 2009 they were t002, t024, t032 t190, and t2849 (13 %). MLST was implemented in 11 selected MRSA isolates and the following STs were detected (spa-types): ST8 (t008, t024, t190, and t2849), ST22 (t032), ST30 (t012), ST125 (t067), ST146 (t002), ST228 (t109 and t1318), and ST247 (t051). These STs belonged to six different CCs, and in the remaining isolates (in which MLST was not performed) the CC was assigned based on the spa-type detected. Thus, according to these criteria, the following CCs were found (%2001/%2009): CC5 (93/76), CC8 (1/22), CC22 (0/2), CC30 (1/0), CC228 (3/0), and CC247 (2/0) (Fig. 1). Therefore, most of the isolates presented spa-types associated with CC5, and a remarkably high percentage of MRSA isolates with spa-types related to CC8 was found in 2009 (Fig. 1).

Fig. 1
figure 1

Clonal complex (CC)-agr-SCCmec and Staphylococcus aureus protein A (spa)-types detected in each studied year

Full size image

The most frequently detected SCCmec type was SCCmec IVc (≥94 % in both periods) and it was associated with isolates of CC5, CC8, and CC22. Other subtypes of SCCmec type IV were detected in very low percentages: the SCCmec IVa was detected in two MRSA CC8 (t008 and t024) isolates and the new variant of SCCmec IVNv was detected in one MRSA CC5-t002 isolate. The MRSA CC30 isolate showed SCCmec III, and the three MRSA CC228 and the two MRSA CC247 isolates had SCCmec I. Three types of agr were detected (%2001/%2009): agr I (3/24), agr II (96/76), and agr III (1/0).

In addition, 28 MRSA isolates were studied by PFGE using the SmaI restriction enzyme (18 MRSA CC5-t067 isolates and one isolate of each of the other spa-types detected). Seven unrelated pulsotypes were identified (A-G); moreover, 5 subtypes were revealed in pulsotype A (A1-A5) and 2 subtypes were shown in pulsotype C (C1 and C2). The distribution of pulsotypes was as follows: A (A1-A5)/CC5-t067, B/CC5-t002, C1/CC8-t008-t190-t2849, C2/CC8-t024, D/CC22-t032, E/CC30-t012, F/CC228-t109-t1318, and G/CC247-t051.

Origin of MRSA isolates

Most of the MRSA were isolated from wounds, representing 39 % in 2001 and 63 % in 2009. Samples of other origins in which MRSA isolates were found were as follows (%2001/%2009): nasal (14/16), urine (20/7), sputum (10/8), blood (9/4), and others (8/2). The correlation between the origin of the isolates and the CC is shown in Fig. 2.

Fig. 2
figure 2

Origins of the isolates and clonal lineages detected in 2001 and 2009. The category ‘Others’ includes bronchial and tracheal aspirates, abdominal and pleural fluid, ears, and catheters. MRSA methicillin-resistant Staphylococcus aureus

Full size image

Resistance phenotypes and genotypes

In general, the level of resistance to quinolones, aminoglycosides, macrolides, lincosamides, and tetracycline remained more or less constant in both periods, although it seemed that there could have been a slight downward trend in 2009 (Fig. 3). Resistance to trimethoprim-sulfamethoxazole was only detected in one MRSA isolate in 2001 and resistance to mupirocin and chloramphenicol was only detected in MRSA isolates in 2009. Moreover, in both years, some isolates (33 % in 2001 and 34 % in 2009) showed a multiresistant phenotype (resistance to three or more antibiotic families in addition to beta-lactams), with these isolates presenting a high number of resistance genes. Table 1 shows the different resistance patterns and the resistance genes detected in the 204 MRSA isolates.

Fig. 3
figure 3

Percentages of resistant isolates detected in the MRSA isolates in each year in this study. CIP ciprofloxacin, TOB tobramycin, KAN kanamycin, GEN gentamicin, ERY erythromycin, CLI clindamycin, TET tetracycline, MUP mupirocin, CHL chloramphenicol, SXT trimethoprim-sulfamethoxazole

Full size image
Table 1 Resistance profiles, resistance genotypes, and CC-spa types of the studied MRSA isolates
Full size table

More different resistance profiles were found in the isolates in 2001 than in isolates in 2009 (Table 1). Quinolone resistance was the one most frequently found among the 204 MRSA isolates. The 13 quinolone-resistant MRSA isolates from blood samples showed the amino acid changes Ser80Phe in GrlA and Ser84Leu in GyrA. Aminoglycoside resistance was also very common, the ant(4′)–Ia gene being the one most frequently found. The levels of resistance to tobramycin and kanamycin in 2001 and 2009 were very high, while the resistance to gentamicin was moderate. The level of macrolide resistance was similar in isolates from both periods, but the percentage of isolates showing clindamycin resistance was higher in 2001 (30 %) than in 2009 (17 %) (Fig. 3). So, it was not surprising that the erm genes were more commonly found in isolates from 2001 than in isolates from 2009.

Tetracycline resistance appeared in a very low percentage of isolates, and different combinations of tet(K), tet(L), and tet(M) genes were detected (Table 1). The only one trimethoprim-sulfamethoxazole-resistant MRSA isolate presented the dfrS1 gene. Additionally, it is relevant to note the detection of 11 mupirocin-resistant isolates in 2009 which harbored the mupA gene. Resistance to chloramphenicol was only detected in one isolate and was associated with the gene cat(pC221).

Virulence gene profiles

Different virulence gene profiles were detected in the 204 MRSA isolates (Table 2). The majority of the MRSA isolates (99 %) were positive for at least one staphylococcal enterotoxin (SE) gene, and all of them carried some haemolysin gene. An association between the SE gene profiles and CC was detected in most of the cases. The SE genes located in the egc-cluster (seg, sei, sem, sen, and seo) were found in isolates with spa-types associated with CC5, CC22, CC30, and CC228. In some cases, the egc-cluster appeared in isolates which also showed the seu (called egc-like cluster), sea, and/or sep genes. The egc-cluster and the sed, sej, and ser genes were identified in one isolate. Interestingly, the egc-like cluster was contained in a higher number of CC5-t067 isolates in 2001 (31 %) than in 2009 (4 %). The combination of the sed, sej, and ser genes and the sea gene was observed in several CC8 isolates. The tst-1 gene was identified in four isolates (one CC30 and three CC5) from 2001. Two MRSA CC8 isolates were positive for lukS/F-PV. Both isolates were obtained in 2009 and were the only ones with SCCmec type IVa. Lastly, the two MRSA CC247 isolates showed the sea gene.

Table 2 Combinations of detected virulence genes in the 204 studied MRSA isolates according to clonal lineages and the year of isolation
Full size table

Discussion

Two hundred and four MRSA isolates obtained from a Spanish hospital in two different years (2001 and 2009) were included in this study. The prevalence of MRSA isolates was slightly higher in 2001 (29 %) than in 2009 (27 %). In some studies carried out in Spain a progressive increase in the oxacillin-resistance level was detected from 1986 (2 %) to 2002 (31 %) [22], and the level seemed to be stabilized in 2006 (30 %) [23]. It seems probable that the implementation of several control measures in the different Spanish hospitals has been able to avoid an increase in these values. In this regard, there have been several attempts to establish a consensus document [24, 25].

In Europe, according to the latest data of the European Antimicrobial Resistance Surveillance System (EARSS), seven countries reported decreasing trends for invasive MRSA, whereas an increasing trend was observed in four other countries [26]. In any case, the rates above 25 % detected in more than one-fourth of European countries, among them Spain [26], are worrisome, with the MRSA problem being a public health priority.

It has been observed that, in contrast to MSSA, MRSA spa-types principally have a regional distribution in European countries [3]. In our work, the t067 spa-type was the most prevalent one in both periods. This spa-type is the one most frequently found in Spain [9] and it is also quite common in Finland, as recently reported [27]; however, it has been found at a very low frequency in other European countries, in the United States, and in South America [3, 28] (http://spa.ridom.de/spa-t067.shtml). Additionally, the spa-type t067 is usually associated with ST125 and, in some cases, with ST5 (both included in CC5), SCCmec type IV, and agr type II. The most common subtypes of SCCmec IV are normally IVa and IVc, which have been detected in variable percentages according to different studies [6, 9]. All our MRSA CC5-t067 isolates presented the same subtype (SCCmec type IVc).

Of note, the association between MRSA CC5-t067 and ciprofloxacin, tobramycin/kanamycin and erythromycin resistance has been described. The resistance genes responsible for these tobramycin/kanamycin and erythromycin phenotypes are mainly the ant(4′)-Ia and msr(A)/msr(B) genes, respectively [9]. In our study, a high percentage of CC5-t067 isolates harbored the ant(4′)-Ia gene, and in regard to erythromycin, a moderate percentage of the CC5-t067 isolates showed resistance to this antibiotic and many of them contained the msr(A)/msr(B) genes. However, these resistance mechanisms, and mainly the ant(4′)-Ia gene, were also found in most of the MRSA non-CC5-t067 isolates in our study.

Another spa-type (t002) related to CC5 was identified in our isolates from 2009. This spa-type has already been detected in other studies carried out in Spanish hospitals [6, 9, 29]. It is interesting to remark that one MRSA CC5-t002 isolate presented the new variant of SCCmec type IV, which had been previously described in isolates with the same spa-type in Argentina [14].

The higher prevalence of MRSA CC8 in 2009 with respect to 2001 is outstanding. While in 2001 only one CC8 isolate was identified (1 %), 22 % of the isolates were of this lineage in 2009. Some STs belonging to CC8 have been considered predominantly as CA-MRSA [2, 30]. Most of our MRSA CC8 isolates were obtained from wound samples, and CA-MRSA usually causes skin and soft-tissue infections. The substitution of HA-MRSA isolates by CA-MRSA isolates has already been reported by others [4, 31–33].

Interestingly, differences between HA-MRSA and CA-MRSA are less clear with time [31]. Thus, most of our MRSA CC8 isolates showed resistance, in addition to beta-lactams, also to aminoglycosides, quinolones, mupirocin, macrolides, or lincosamides. Moreover, only two isolates harbored the PVL genes.

Interestingly, some of our CC8 and CC5 isolates of 2009 showed mupirocin resistance. The emergence of this resistance has already been described in other hospitals [34]; this resistance could be produced by the use of mupirocin for treating nasal MRSA carriers.

Other detected CCs (CC22, CC30, CC228, and CC247) in our study have been described previously in studies carried out in our country [6, 8, 9, 29]. CC22 is a British clone (EMRSA-15) which was detected in Spain in 1999, and it has always been found at a very low percentage (<2 %) [8]. MRSA CC30 (EMRSA-16) has usually been described as showing SCCmec II or IV [9, 29]. However, our MRSA CC30 isolate presented SCCmec III and, in the only description of this type of SCCmec in CC30, an association with Tn6072 was established [35]. CC228 (Southern German clone) is, in addition to CC5 (pediatric clone), one of the major HA-MRSA clones widely disseminated around the world [2]. Our two MRSA CC247 isolates were the only tetracycline-resistant ones in 2001 and this CC is known as the Iberian clone. In several studies performed in Spanish hospitals this CC was found to be the dominant one in the 1990s. However, its presence has been decreasing in subsequent studied years [7, 8]. Interestingly, in our study, CC247 was only detected in two isolates from 2001, and it was not identified in any isolates from 2009.

We note that some spa-types associated with CC5 (t003, t041, t062, t088, t837, t1154, t2222, t2226), CC8 (t148), and CC30 (t018), which have been detected previously by others in Spain [9], were not identified in the present study. Moreover, other major HA-MRSA and CA-MRSA clones, such as CC45, CC59, and CC80 [2], which are common in other countries, were not detected in our hospital.

We detected several isolates that presented a multiresistant phenotype; infections caused by this microorganism are difficult to treat. Remarkably, in some cases, multiple genes encoding the same resistance were identified. However, about 65 % of the isolates were resistant to fewer than three antimicrobials, besides oxacillin, and this finding could be explained by the high prevalence of SCCmec IV, as suggested by others [9]. We note that there appeared to have been a slight downward trend in the level of resistance in 2009. A significant decrease in the levels of resistance to gentamicin and clindamycin had already been observed [9, 23].

The tst-1 gene was detected in four isolates from 2001; three of them belonged to CC5-t067 and the other one to CC30-t012. CC5-t067 isolates harboring the tst-1 gene have been detected in Spain [6] and a TSST-1 CC5 clone has recently been described in France [36]. Moreover, an association between TSST-1 and CC30 has been reported previously [37].

There seems to be a linkage between SE genes and the clonal background [38]. In our study, the egc cluster was detected in isolates belonging to CC5, CC22, CC30, and CC228. This is in accordance with the results obtained by others, where egc was detected in CC5, CC22, and CC30 isolates, but not in CC8 isolates [10, 38]. The combination of the sed, sej, and ser genes was identified in 15 CC8 isolates (t008 and t2849, spa-types) and in two CC5-t002 isolates. These genes are located on the plasmid pIB485 [39] and their presence in CC8 and CC5 isolates has been described previously [40]. Moreover, a strong linkage between the plasmid-borne SE genes (sed, sej, and ser) and CC8 has been suggested [38]. Other SE genes identified were sea and sep, both located on bacteriophages [41, 42]. In the present study, the sea gene was mainly found in CC8, CC228, and CC247 isolates and the sep gene in CC5 (but also in one CC8 isolate). The detection of the sea gene in several CCs is explained by the broad distribution of this gene [38]. Moreover, it has been observed that the sep gene is more frequent in MRSA agr II isolates than in MSSA isolates [6].

In conclusion, in the present study, the most commonly found spa-type in both periods was t067 associated with CC5. This CC5-t067 lineage is still predominant in our country and it is associated with resistance to quinolones, aminoglycosides, and to a lesser extent, with resistance to macrolides. The possible emergence of CC8 was identified in the second period with two lukS/F-PV-positive MRSA isolates. The presence of four tst-1-positive isolates in the first period was also remarkable. The level of resistance among MRSA isolates remained more or less constant in both periods, although it seemed that there could have been a slight downward trend in 2009. This fact would be very positive and requires further research. The surveillance of genetic lineages, antibiotic resistance mechanisms, and virulence traits of MRSA isolates in our hospitals is highly important to better understand the behavior of this microorganism.