Anne-Merethe Hanssen, Department of Microbiology and Virology, Institute of Medical Biology, University of Tromso, Tromso, Norway. Tel.: +47 77 64 57 52; fax: +47 77 64 53 50; e-mail: email@example.com
Staphylococcal cassette chromosome (SCC) elements are, so far, the only vectors described for the mecA gene encoding methicillin resistance in staphylococci. SCCmec elements are classified according to the type of recombinase they carry and their general genetic composition. SCCmec types I–V have been described, and SCC elements lacking mecA have also been reported. In this review, we summarize the current knowledge about SCC structure and distribution, including genetic variants and rudiments of the elements. Its origin is still unknown, but one assumes that staphylococcal cassette chromosome is transferred between staphylococci, and mecA-positive coagulase-negative staphylococci may be a potential reservoir for these elements. Staphylococcal genomes seem to change continuously as genetic elements move in and out, but no mechanism of transfer has been found responsible for moving SCC elements between different staphylococcal species. Observations suggesting de novo production of methicillin-resistant staphylococci and horizontal gene transfer of SCCmec will be discussed.
The genetic determinant of methicillin resistance (mec) has been localized on the chromosome of Staphylococcus aureus (Sjostrom et al., 1975; Kuhl et al., 1978). There is no allelic equivalent of mec in methicillin-susceptible S. aureus (MSSA). The mecA gene encodes an additional penicillin-binding protein (PBP), PBP2a, which has a low affinity for all β-lactam antibiotics (Hartman & Tomasz, 1981; Archer & Niemeyer, 1994). PBP2a is a high-molecular weight class B transpeptidase that catalyzes the formation of cross-bridges in bacterial cell wall peptidoglycan (Goffin & Ghuysen, 1998; Berger-Bachi & Rohrer, 2002). Assisted by the transglycosylase domain of the native PBP2 of S. aureus (Pinho et al., 2001) it takes over the function of cell wall biosynthesis in the presence of β-lactam antibiotics when normally occurring PBPs are inactivated by ligating β-lactams. mecA is located on a mobile genetic element called staphylococcal cassette chromosome (SCC) in S. aureus (21–67 kb fragment) (Ito et al., 1999; Katayama et al., 2000).
In this review, we describe the SCCmec element in different staphylococci and discuss its role in horizontal gene transfer between staphylococci. There are several reasons why we should pay attention to staphylococci and the mobile genetic element SCC.
1Staphylococci are part of our normal flora and they are potential pathogens that have become resistant to all known antibiotics. In Norway, 70–80% of methicillin-resistant CoNS (MR-CoNS) and <1% of methicillin-resistant S. aureus (MRSA) can be found in hospitals and in the community. MRSA infections increase the cost and the risk of mortality each year worldwide.
2Not much is known about the clonality of MRSA and MR-CoNS in Norway. Basic studies of staphylococci on the genetic level are important in understanding the development of resistance and the mechanisms of its spread. It is important to distinguish between the spread of resistant isolates (clonal) and the spread of resistance genes (horizontal). Only horizontal gene transfer of mec has been observed, and the mechanisms for transfer are still unknown.
3The SCC element, carrying resistance or virulence, is a mobile genetic element that can spread between different species.
4Norway has a low prevalence of antibiotic resistance, and we are in a unique position for studying resistance development and horizontal gene transfer at an early stage.
This review mainly deals with the diversity and distribution of SCCmec in staphylococci. In addition, the de novo generation of methicillin-resistant staphylococci and the horizontal gene transfer of SCCmec are discussed.
Mobile genetic elements in staphylococci
Pathogens are able to spread, establish ecological reservoirs, colonize and cause disease. At the same time they can acquire resistance genes and adjust resistance levels to increasing concentrations of antimicrobial agents. Resistance genes themselves, like the elements that carry them, can have a very broad host range. Mobile genetic elements involved in the spread of resistance and virulence in staphylococci are genomic islands, bacteriophages, pathogenicity islands, chromosomal cassettes, plasmids, insertion sequences and transposons (Baba et al., 2002; Ito et al., 2003; Holden et al., 2004; Lindsay & Holden, 2004; Gill et al., 2005).
Integration of the element is sequence specific, i.e. at a unique site (bacterial chromosomal attachment site, attBSCC) located near the S. aureus origin of replication (close to pur and spa). attBSCC is found downstream of an open reading frame (ORF) of unknown function, designated orfX, that is well conserved among clinical strains of S. aureus (Ito et al., 2001). orfX is present in both methicillin-resistant and -susceptible strains. attBSCC contains a 15 bp sequence that, when SCCmec is integrated in the chromosome, is found at both chromosome-SCCmec junctions. One of the two repeat sequences is located within SCCmec at the right end. Incomplete inverted repeats are also present at both ends of SCCmec (Ito et al., 2001). These repeated sequences appear to be recognized by SCCmec-specific recombinases during the integration and excision of SCCmec to and from the chromosome (Hiramatsu et al., 2001). The exact integration site of SCCmec was inferred to within a maximum ambiguity of four bases (Ito et al., 1999).
SCC does not contain any genes coding for bacteriophage head and tail proteins, and it lacks the tra genes necessary for conjugative transfer (Ito et al., 1999). In addition to the ccr and mec genes, the elements contain one or more antibiotic resistance genes (Hiramatsu et al., 2001), which are carried on various transposons and integrated copies of plasmids (Katayama et al., 2000).
The different types of SCCmec are called SCCmec type I (class B mec gene complex and ccrAB type 1); SCCmec type II (class A mec gene complex and ccrAB type 2); SCCmec type III (class A mec gene complex and ccrAB type 3); SCCmec type IV (class B mec gene complex and ccrAB type 2), and SCCmec type V (class C2 mec gene complex and ccrC) (Fig. 1). The remaining part of SCCmec, besides ccr and mec, is called the J-region (J for Junkyard), meaning that these are nonessential components of SCCmec (Ito et al., 2003). Variants of each SCCmec type are defined by differences in the J regions. SCCmec type IV has been subdivided into subtypes SCCmec IVa and SCCmec IVb (Ma et al., 2002), and SCCmec IVc (Ito et al., 2003) based on differences in the J-region called J1 (Okuma et al., 2002). SCCmec subtype IVd has also been described (GenBank acc. no. AB097677).
A simplex PCR strategy (Ito et al., 2001, 2004; Okuma et al., 2002) has been developed for the detection of ccr and mec complexes in addition to the J-region. A multiplex PCR strategy has been developed for the rapid identification of structural types and variants of the mec elements in MRSA (Oliveira & Lencastre, 2002), where six main loci (A–F) are included. By this method one can detect variant differences. SCCmec type variant IA differs from SCCmec type I by the presence of an integrated copy of pUB110 downstream of the mec complex (Fig. 1). SCCmec IIIA differs from SCCmec type III by the absence of pT181 and its flanking IS431 elements (Fig. 1). SCCmec IIIB lacks the integrated copies of Tn554, pT181 and the mer operon with its associated insertion sequences (Oliveira & Lencastre, 2002). SCCmec type IVA has also been described and it differs from type IV by the presence of a copy of pUB110 (Oliveira & Lencastre, 2002).
Recently, seven novel variants of SCCmec types II and IV were reported – SCCmec IIA, IIB, IIC, IID, IIE, IVE and IVF (Shore et al., 2005). In addition, an SCCmec I variant lacking the pls gene was observed, plus an SCCmec III variant lacking pI258 and Tn554 (Shore et al., 2005).
Staphylococci contain one to four copies of IS431, at least one of which is located downstream of mecA (IS431mec) (Archer & Niemeyer, 1994) (Fig. 2). There is a higher prevalence of IS431 in CoNS than in S. aureus (Kobayashi et al., 2001a), especially in Staphylococcus haemolyticus (Kobayashi et al., 1999). IS431 elements trap and cluster resistance determinants with similar IS elements through homologous recombination and this explains the multiple drug resistance phenotype that is characteristic of methicillin-resistant staphylococci. IS431mec contains an open reading frame of a putative transposase gene and 14–22 bp terminal inverted repeats (Kobayashi et al., 2001b).
SCCmec types I, IV and V do not contain any antibiotic resistance genes, with the exception of mecA (Okuma et al., 2002), but SCCmec subtype IVc carries Tn4001 encoding bifunctional AAC/APH protein (aacA-aphD), conferring resistance to most of the aminoglycosides except arbekacin (Gillespie et al., 1987). SCCmec type II carries Tn554 (Murphy et al., 1985), which encodes erythromycin (ermA) and spectinomycin (spc) resistance. pUB110 is flanked by a pair of IS431 elements (McKenzie et al., 1986) (Fig. 1), and encodes kanamycin and tobramycin (aadD)/bleomycin (ble) resistance (Ito et al., 1999). SCCmec type III contains Tn554 (MLS resistance), pseudo ψTn554, encoding cadmium resistance (cad), pUB110, an integrated copy of pT181 (tetK, tetracycline resistance) and pI258, carrying mercury resistance (Ito et al., 1999) (Fig. 1). Unique to SCCmec type V are hsdR, hsdS and hsdM, which encode a type I restriction-modification system that might play a role in the stabilization of the element (Ito et al., 2004).
SCC is a conveyor not only of methicillin resistance and other antibiotic resistance genes, but also of virulence genes. At least four SCC non-mec types have been described so far in S. aureus and CoNS. S. aureus SCCcap1 is located at the same chromosomal site as all SCCmec elements, and it contains a virulence factor called capsular polysaccharide 1, which makes the strain more resistant to phagocytosis (Luong et al., 2002). It is defective in mobilization because it lacks a ccrA homolog and contains a ccrB homolog with a nonsense mutation (ϕccr-complex). SCCcap1 resembles SCCmec type III. The Staphylococcus hominis ATCC 27844 SCC element contains neither antibiotic resistance genes nor mobile genetic elements (Katayama et al., 2003a,b). It contains ccrAB1 with an intact ccrB1 gene, and not a frameshift mutation as reported in ccrB1 in S. aureus (Ito et al., 2001). ATCC 27844 carries several homologs of restriction modification genes for stable maintenance, stabilization or defense of SCC DNA upon entry into restrictive host bacteria (Ito et al., 2003; Katayama et al., 2003a,b).
Two new members of the SCC family have been found in Staphylococcus epidermidis ATCC 12228, namely SCCcomposite island and SCCpbp4 (Mongkolrattanothai et al., 2004). The unique SCCcomposite island (57 kb) lacks mecA but carries two pairs of type 2 and 4 ccrAB recombinase genes. It is flanked by 28 bp SCC specific terminal repeat sequences. The SCCcomposite island also contains a smaller SCC element, SCCpbp4 (19 kb), nested within the right side of the SCCcomposite island and carrying a homolog of the gene encoding PBP4 (pbp4), which encodes a teichoic acid biosynthesis protein (tagF) and three stretches of DNA that are highly homologous to those found in S. aureus SCC elements (Mongkolrattanothai et al., 2004). The ccrAB genes in SCCpbp4 are closely related to the type 4 ccrAB identified in MRSA strain HDE288 (Oliveira et al., 2001). SCC476 has been found in MSSA strain 476, which carries ccrAB with closest similarity to ccrAB1 of S. hominis ATCC27844 (Katayama et al., 2003a,b), and a fusidic acid-resistance determinant Far1. Our studies have also shown that the cassette might exist without mec in CoNS (Hanssen et al., 2004).
Genetic variants and rudiments of SCC
The MSSA strain ATCC 25923 carries a 5877 bp fragment inserted at attBSCC downstream of orfX (Ito et al., 2001). It has a structural characteristic similar to SCCmec at both ends, i.e. incomplete inverted repeats and direct repeats of 15 bp, but contains no drug resistance gene or ccr genes. It seems to be a remnant of SCC/SCCmec that was integrated in its complete form and then afterwards deleted with ccr and mec (Ito et al., 2001). In addition, SCC fragments, such as dcs (a conserved region in SCCmec types I, II and IV between the IS431 copy and orfX) and ccrAB1, have been reported in an MSSA strain (Corkill et al., 2004).
All the variants of SCCmec and nontypeable SCCmec observed in various studies indicate that the reservoir of SCCmec is rather large in CoNS. The mec locus exists in the absence of the known types of ccr genes, both in MRSA and in CoNS (Hanssen et al., 2004, 2005). The explanation for the absence of ccr genes in SCC may be:
1there are unrecognized ccr types because the ccr genes are distantly related to the known ccrABs and ccrC;
2a high mutational rate in the binding sequence of the ccr primers;
3homologous recombination between ccr genes may generate new isotypes of ccr complexes that are not detectable by PCR;
4the ccr genes have been deleted from SCC; or
5mecA is transferred independently of ccr.
The ccr-untypeable strains may contain novel types of ccr genes, or SCCmec without ccr genes (Okuma et al., 2002). Most of the S. haemolyticus strains (ccrAB2, no mec-regulators and no IS1272) and S. hominis strains (ccrAB1 with original version of ccrB1, but not IS1272) observed in one of our studies are likely candidates for carrying novel types of SCCmec (A.-M. Hanssen et al., unpublished observations).
There are reports of multiple SCC insertions (Ito et al., 2001; Mongkolrattanothai et al., 2004), e.g. SCCmec type III seems to be composed of two SCC elements and it contains both ccrAB3 and a ccr pseudo-version (Ito et al., 2001). Composite versions of SCC might be explained by sequential integration of two copies of SCC followed by deletion(s) of internal parts. The recombinant SCCmec type IVE may have arisen by recombination between SCCmec type IVc and another SCCmec element, while SCCmec type IVF may have arisen by recombination between a SCCmec II variant and SCCmec IVE or between SCCmec type IVb and SCCmec IVE (Shore et al., 2005). The high variability of SCC structures, deleted and original versions of ccrB1, pseudo-ccr genes and multiple copies of SCC elements indicate that these elements may be hot spots for recombination, a possible survival strategy for the bacteria. The mec- and ccr-gene complexes go through complex recombination and rearrangement processes in the genomes of CoNS, generating novel types of SCCmec elements. Most likely, only a small fraction of the diversity in SCC elements is displayed in S. aureus (Ito et al., 2004).
Distribution of SCCmec types
The distribution of SCCmec in nature is limited to relatively few clonal complexes of related MRSA (Katayama et al., 2005). SCCmec IV has been found in diverse genetic backgrounds, which suggests that type IV has an increased mobility compared with their larger SCCmec counterparts (Daum et al., 2002). It is the dominant SCCmec type in community-acquired MRSA (CA-MRSA) and is rarely seen in health-care associated MRSA strains (H-MRSA) (Okuma et al., 2002; Ito et al., 2003). The majority of epidemic H-MRSA carry SCCmec types I, II or III (Ito et al., 2001; Enright et al., 2002), except for the multiply susceptible MRSA from Australian hospitals. The nonmultiresistant oxacillin-resistant S. aureus (NORSA) strains carry SCCmec type IVa or IVb (Merlino et al., 2002; Okuma et al., 2002). SCCmec subtype IVc is described as hospital-acquired in France (Vandenesch et al., 2003), but SCCmec subtype IVc has also been reported among CA-MRSA in Sweden (Berglund et al., 2005) and Norway (Hanssen et al., 2005). Most isolates with unknown SCCmec types are CA-MRSA (Berglund et al., 2005). SCCmec types IV and V are present in diverse genetic backgrounds among MRSA from outpatients in Australia (Coombs et al., 2004). Less is known about the distribution of SCCmec in CoNS, but type IV SCCmec with diverse J1-regions are distributed in more than 30% of CA-S. epidermidis (Ito et al., 2004, unpublished data). As many as 10 different structural types of SCCmec have been observed in methicillin-resistant S. epidermidis (MRSE) (Miragaia et al., 2005). SCCmec type V is also distributed among CoNS (Ito et al., 2004, unpublished data). In a report on hospital isolates of S. epidermidis, 36% were SCCmec type IV, 34% SCCmec type II, 28% SCCmec type III and only 2% SCCmec type I (Wisplinghoff et al., 2003).
Genes on the move
The genetic pool and horizontal gene transfer
The genetic pool is a diverse assemblage of mobile genetic elements. Every organism is in theory potentially able to take up DNA, and the probability of an organism taking up DNA is limited by: (1) the ability to take up DNA; (2) the willingness to deliver DNA and (3) the organisms being in near to each other at the same time in the same environment, or free DNA circulating in the environment. All organisms sharing the same ecological niche, or bacterial pathogens that are just passing through, are potential donors and recipients. Bacteria from different sites also appear to exchange genes. Even though it is very difficult to show that such transfers can and do happen, there is evidence supporting this.
Conjugation and conjugative mobilization have been described among staphylococcal isolates (Forbes & Schaberg, 1983; Projan & Archer, 1989; Archer & Scott, 1991; Thomas & Archer, 1992). Broad-host range conjugative plasmids with aminoglycoside resistance have been shown to transfer at low frequencies from streptococci to staphylococci (Macrina & Archer, 1993) and a self-transmissible plasmid transfers between CoNS and Staphylococcus aureus (Archer & Johnston, 1983). It has been speculated that the same conjugative transposon harboring tetracycline resistance determinant (tetM) is transmissible between S. aureus and Clostridia across the genus barrier (Ito et al., 2003). Tn4001, found on the chromosome and multiresistance plasmids in S. aureus and CoNS, is thought to be largely responsible for the emergence of linked resistance to the aminoglycosides gentamicin, tobramycin and kanamycin in staphylococci (Rouch et al., 1987). Transposons closely related to Tn4001 are also found in enterococci (Hodel-Christian & Murray, 1991) and streptococci (Horaud et al., 1996). IS elements can jump into and out of phage genomes and conjugative plasmids. IS431 is involved in the transfer of genes or entire plasmids into other replicons or the chromosome. It is associated with antibiotic resistance genes and is widely distributed in staphylococci (Kobayashi et al., 2001b). Gill and coworkers (Gill et al., 2005) found evidence of gene transfer and movement of mobile elements between the staphylococci and other low-GC-content Gram-positive bacteria. The possible interspecies transfer of the vancomycin resistance determinant vanA between Enterococcus faecalis and S. aureus (Weigel et al., 2003) is further evidence that there may be a common gene pool available for many different species. Elements involved in dissemination of resistance determinants among Gram-positive bacteria are numerous, and novel elements are still being discovered. Horizontal gene transfer between bacteria in the environment is necessary for the generation of genetic diversity and is a means of natural selection and evolution (Solomon & Grossman, 1996). There is a pool of virulence- and antibiotic-resistance genes in the environment in the form of large elements available for transfer between strains. Acquisition of DNA might provide the necessary genetic material for bacteria to survive in a quickly changing environment. S. aureus has been described as containing a skeleton where all S. aureus are the same and there are regions that pop in and out (Holden et al., 2004).
Origin and reservoirs of SCCmec
The origin of SCCmec is unknown. It is believed that mecA in all staphylococci descend from one common ancestor, and Staphylococcus sciuri may have been the evolutionary precursor of the structural gene of PBP2a (Couto et al., 1996). Evidence suggests horizontal gene transfer of mec DNA between staphylococcal species and genera of the mecA gene between different Gram-positive genera (Archer & Niemeyer, 1994). One assumes that ccr and mec genes were brought together in CoNS from an unknown source (Suzuki et al., 1993; Archer et al., 1994; Wu et al., 1996; Hiramatsu et al., 2001), where deletion of the mec regulatory genes occurred, and then the genes were transferred into S. aureus to generate MRSA (Musser & Kapur, 1992; Suzuki et al., 1993; Archer et al., 1994). The mec-determinant probably entered S. aureus after the introduction of methicillin into medical practice. Resistant strains of CoNS may serve as a reservoir for antibiotic-resistant genes that can possibly be transferred to other Gram-positive organisms, including strains of S. aureus (Mongkolrattanothai et al., 2004). Information about these reservoirs of antibiotic resistance genes is thus important.
The hypothesis for the transfer of SCCmec between Staphylococcus epidermidis and S. aureus has been supported by several lines of evidence (Wisplinghoff et al., 2003). SCCmec type IV in S. epidermidis shows a 98–99% homology with SCCmec type IVa in S. aureus. In the rest of the S. epidermidis genome only 17% of the ORFs have at least 80% identity to S. aureus, suggesting that interspecies exchange of DNA has occurred (Wisplinghoff et al., 2003). Our studies have shown that CoNS strains contain ccrAB genes that are 100% identical to S. aureus (Hanssen et al., 2004, 2005). Another line of evidence is that IS1272 is present in Staphylococcus haemolyticus and S. epidermidis but is less frequently found in S. aureus (Archer et al., 1996). Staphylococcus aureus and S. epidermidis have identical IS1272–ΔmecR1 junction sequences in SCCmec type I and IV (Wisplinghoff et al., 2003). Recombination may have occurred in CoNS between type I sequences and additional sequences and generated SCCmec type IV, and the type IV sequence was subsequently transferred into S. aureus (Archer et al., 1996; Kobayashi et al., 1999; Wisplinghoff et al., 2003). However, IS1272 appears to be intact at the many insertion locations in S. haemolyticus, although it usually contains deletions on locations in S. epidermidis and S. aureus. It is assumed that S. haemolyticus was one of the definitive hosts for IS1272 and was only acquired secondarily by S. epidermidis and S. aureus (Archer et al., 1996). A third line of evidence is that methicillin resistance is highly prevalent among S. epidermidis isolates and is less common among S. aureus isolates; more than 70% of S. epidermidis hospital isolates are methicillin resistant. Therefore, the reservoir of SCCmec is likely to be larger in S. epidermidis (Archer et al., 1994; Wisplinghoff et al., 2003). Also, type IV SCCmec was highly prevalent among S. epidermidis from the 1970s and has not been found among MRSA isolates recovered during that time. The first S. aureus carrying SCCmec IV was recovered in the early 1980s (Wisplinghoff et al., 2003). The SCCcomposite island in S. epidermidis (Mongkolrattanothai et al., 2004) contains stretches of homology with other SCCs found in MRSA isolates, and it has been suggested that S. epidermidis acts as reservoir for DNA sequences that have been horizontally transferred and recombined within SCC elements in S. aureus by homologous recombination (Mongkolrattanothai et al., 2004). On the other hand, de Sousa & de Lencastre (2004) propose that the main source of SCCmec could be MRSA itself and that the acquisition of mecA from CoNS is an infrequent event. This hypothesis needs further investigation.
Generally, we are unlikely to detect gene transfers that produce no change in the pathogen or are deleterious to the pathogen. Hospitals are convenient ecosystems for gene transfer because of the many patients, continuous change, reservoirs, and selective antibiotic pressure. Recognition of gene transfer is focused on disease and treatment, and usually only gene transfers involving pathogens are detected. The transfer event itself is hardly ever observed (Tauxe et al., 1989). For many years, S. epidermidis has been unappreciated as a nosocomial pathogen, but this species is likely to be a reservoir of genetic elements for relevant pathogens. The suggestion that the skin may be the site for gene transfer was put forward in the 1970s (Noble & Naidoo, 1978). The prevalence of a particular gene or element would not have to be very high among CoNS for it to serve as a reservoir for genes capable of dissemination (Archer & Scott, 1991).
The basic hypothesis for all our work (Hanssen et al., 2004, 2005) has been that ‘mec is transferred horizontally between staphylococci.’ This is based on the observation that we have a large population of MR-CoNS in Norway, but a low prevalence of MRSA. Could it be that CoNS strains are reservoirs for SCCmec? MRSA only appear as pop-ups in Norway, and if they belong to the international clones, they would be genetically identical. One of our working hypotheses has been that ‘MRSA in northern Norway arise locally as S. aureus acquires SCCmec from other staphylococcal derivatives’ (Hanssen et al., 2004). If so, we would see greater genetic relationship among staphylococci from Norway than between Norwegian isolates and international staphylococci. We have found evidence of horizontal gene transfer of SCCmec between local CoNS and S. aureus, but we cannot say anything about the direction of transfer or the mechanisms involved. Strains from the same geographical region have identical ccr genes that differ from those of strains from other regions. Type 1 and 2 ccrAB genes are different (≥97% homologous) from the published sequences, whereas type 3 genes are 100% identical to the published sequences. The Norwegian strains contain a local variant of ccrAB2 with 99% identity to ccrAB2 in SCCmec type IVc and 96% identity to ccrAB2 in S. aureus strain N315. The major differences are found in the ccrA2 gene. The sequence homology within the ccr subtypes in genomically unrelated methicillin-resistant staphylococcal strains indicates a horizontal transmission of the SCC element as a major mode of dissemination.
Horizontal gene transfer of SCCmec into lineages of MSSA from members of the CoNS is a possible explanation for the discovery of MRSA isolates in the community in Northern Norway, where no obvious connection to hospitals or recent travel abroad could be found (Hanssen et al., 2004, 2005). This MRSA strain may not be successful as an endemic clone, but selective pressure by β-lactam antibiotics can promote survival of this strain. The existence of CoNS in nosocomial settings where a selective antibiotic pressure exists can perhaps provide a background for new MRSA lineages.
Mechanisms of SCC transfer
To understand the spread of methicillin resistance it is important to identify how the resistance genes are dispersed. So far, there has been no report of SCCmec in other bacteria than staphylococci (Hiramatsu et al., 2001). Neither the mechanisms responsible nor the organisms involved in mecA transfer are known, and in general very little is known about the CoNS genome. It is only recently that we got an insight into the S. epidermidis genome (Mongkolrattanothai et al., 2004; Gill et al., 2005). Most experiments are based on what is known about SCCmec in S. aureus. Lacey (1972) failed to transfer mec between S. aureus strains by conjugation, whereas Cohen & Sweeney (1973) showed that mec is transferable between S. aureus cells by bacteriophage-mediated generalized transduction. Trees & Iandolo (1988) reported that mec could be mobilized from the chromosome to a penicillinase plasmid, pI524, and suggested that mec may be a part of a transposable genetic element. But no transducing phage capable of transferring genetic information across the staphylococcal species barrier has been described (Katayama et al., 2000). Wielders and coworkers (Wielders et al., 2001) reported a possible horizontal in vivo transfer of mecA between S. epidermidis and S. aureus during antibiotic treatment, and a sporadic MRSA emerged de novo. The way in which SCCmec functions remains to be determined, and whether the mechanism involved is phage transduction, transformation, conjugation or another genetic transfer system.
One might expect a frequent exchange of genetic material between S. aureus and CoNS species that are genetically closely related and that inhabit the same ecological niches. Surprisingly, this does not seem to be the case regarding mecA. The incidence of gene transfer between organisms depends on the degree of interaction between organisms, selective pressures in the environment, the mechanism of gene transfer (conjugation, transformation, transduction), environmental conditions and host restriction (Tauxe et al., 1989). Several systems for the introduction of genetic material into staphylococci have been developed, i.e. protoplast transformation, electroporation, and generalized transduction, but these have limitations. The reasons for the difficulty introducing mec into staphylococci may be that different environmental pressures are affecting the transfer process, or that specific gene transfer mechanisms for different elements may be operative in the various staphylococci. There may also be restriction barriers affecting the shuttle introduction into staphylococcal recipients (Katayama et al., 2003a,b), e.g. the genetic background may affect the stability of mecA in S. aureus (Katayama et al., 2005).
De novo production of methicillin-resistant staphylococci
By using multilocus-sequence typing (MLST) and SCCmec typing as methods we can better understand the origin and spread of MRSA clones. In addition, they are excellent tools for describing the movement of SCCmec among S. aureus strains. We have suggested that there is both horizontal transfer of SCCmec type IV and clonal spread of successful MRSA strains in northern Norway (Hanssen et al., 2005). Two MRSA clones, ST8-MRSA-IV and ST80-MRSA-IV, predominate in northern Norway, but international epidemic MRSA strains have also been observed (Hanssen et al., 2005).
There are several reports that support the transfer of SCCmec into an MSSA genetic background converting S. aureus to MRSA (Hiramatsu et al., 2001; Enright et al., 2002; Robinson & Enright, 2004). Using MLST and additional sets of highly variable S. aureus surface protein (sas) genes, Robinson & Enright (2003) made evolutionary models that indicated that there were 16 acquisitions of SCCmec by an MSSA. Wielders et al. (2001) observed the result of in vivo horizontal transmission of mecA to an infecting strain of MSSA. The origin of strains containing SCCmec IV is suggested to be hospital-acquired MRSA (H-MRSA) that have undergone deletions of antibiotic resistance genes under lower antibiotic selective pressure in the community (Mongkolrattanothai et al., 2003). It is unclear which staphylococcal species donated the five SCCmec types found among MRSA, but the presence of five types suggests multiple introductions into S. aureus, and their presence in the same sequence type (ST) indicates that horizontal transfer of mec genes is relatively frequent within S. aureus (Enright et al., 2002).
The North-Norwegian ST8-MRSA-IV isolates (Hanssen et al., 2005) are represented by EMRSA-2 and EMRSA-6, which have been identified in several European countries and in the United States (Enright et al., 2000; Aires & de Lencastre, 2003). These Norwegian ST8-MRSA-IV strains may be remnants of the early MRSA isolates, i.e. similar to the old, classical, antibiotic-susceptible MRSA types, often with a relatively low level of methicillin resistance, lack of the regulation gene mecI, and closely related PFGE profiles (Crisostomo et al., 2001). The early MRSA isolates may represent the progeny of MSSA that was the original recipient of the heterologous mec element (Oliveira et al., 2001). MSSA strains ready to participate as recipients in the horizontal spread of mecA were already present among MSSA isolates recovered around the early 1960s. These strains may have remained in circulation, acquiring various forms of the mec element and providing the genetic backgrounds for some of the contemporary MRSA clones (Crisostomo et al., 2001; Oliveira et al., 2001). The North-Norwegian ST8-IV isolates are closely related so they could recently have diverged from a common ancestor, for example by horizontal gene transfer of SCCmec-IV to a local methicillin-susceptible S. aureus strain (Hanssen et al., 2005). The strains are also generally susceptible to antibiotics, indicating that there might not have been enough time for the selection of multidrug-resistant clones of MRSA. MLST on MSSA strains in northern Norway would add valuable information about the relationship and evolutionary origin of MSSA and MRSA in the region.
Why do we not see more MRSA in Norway? International clones of MRSA are entering the country, but it seems as if they do not gain ‘ground.’ There are several factors that may contribute to this: temperate climate, low population density, a high social standard, cultural differences, good hospital hygiene and overall low consumption of antibiotics, or perhaps Norwegian S. aureus strains do not ‘need’ the SCCmec carrying multiresistance to the same extent as S. aureus in other countries.
MRSA strains in Norway have always been considered sporadic because they appear as pop-ups and the epidemic ends within short time (Andersen et al., 2002). Diverse ST and PFGE-pattern, SCCmec type IV and low-level antibiotic resistance provide support for the sporadic nature of these MRSA strains. The sporadic MRSA have been described as resembling community-acquired MRSA isolates with a relatively limited multiresistance pattern, faster growth rates, a diversity of genetic backgrounds, and a frequent association with SCCmec type IV (Aires & de Lencastre, 2003). There may be a continuous new synthesis of MRSA with few resistance determinants compared to most other epidemic MRSA strains, possibly replacing previously circulating multiresistant MRSA in northern Norway. This is supported by Wannet et al. (2004), who suggested that horizontal transmission of SCCmec type IV from an unknown source, MR-CoNS or MRSA, into ST45 MSSA resulted in ST45 MRSA, the predominant clone in the Netherlands. This clone has low-level oxacillin resistance and a novel SCCmec type (Wannet et al., 2004). The sensitive MRSA strains are often detected in settings with low antimicrobial pressure and low patient turnover (Andersen et al., 2002). Bacterial competition and the lack of strong selection may limit the community spread of MRSA and account for its sporadic distribution. Also, the ability of an MRSA to colonize and replace a pre-existing MSSA strain could be inhibited by bacterial interference (Shopsin et al., 2000).
The North-Norwegian ST8-IV MRSA strains are highly related, suggesting clonal expansion in northern Norway (Hanssen et al., 2005). However, variation in the resistance pattern and local variants of ccrAB2-sequences in ST8-IV suggest that they have evolved for some time separated from the origin and may have acquired SCCmec from local staphylococcal strains (Hanssen et al., 2005). The ST80-IV, ST5-IV and ST12-IV MRSA strains that we have observed do not contain the local variant of ccrAB2, but instead ccrAB2, typical for SCCmec type IVc. It remains to be shown whether a local variant of ccrAB2 is a typical trait for the North-Norwegian ST8-MRSA-IV strains only, or whether this variant is found in all ST8-IV strains from other countries.
Our data indicate that in northern Norway CA-MRSA predominate (Hanssen et al., 2005). This is based on the presence of SCCmec type IV, a low level of resistance, a low level of oxacillin resistance, diverse PFGE and diverse STs, and the fact that the isolates are obtained from people in the community. But H-MRSA have also been observed, with high-level resistance, and containing SCCmec types I, II or III. There appear to be two distinct ways that CA-MRSA originated: displacement of some epidemic hospital clones into the community (where the H-MRSA have undergone deletions of antibiotic resistance genes under lower antibiotic pressure in the community), and horizontal transmission of SCCmec type IV into MSSA (Enright et al., 2002). There is no definitive evidence that any CA-MRSA isolates have arisen from commensal MSSA (Eady & Cove, 2003), except for the demonstration of the possible in vivo horizontal transmission of SCCmec into an infecting strain of MSSA, from which a sporadic MRSA emerged de novo (Wielders et al., 2001).
Are all staphylococci feeding from a local gene pool? Almost all sporadic MRSA isolates from northern Norway have ST8 and a local variant of ccrAB2. There are CoNS isolates from the same environment that have identical nucleotide ccrAB-sequences (Hanssen et al., 2004, 2005). These data suggest either that MRSA are spread clonally in Norway or that they obtained their ccrAB genes from the local gene pool. The near identity between ccr sequences from MRSA and CoNS suggests a very short time of divergence. It is highly probable that the ccr genes were horizontally transferred between species some time in the past.
The staphylococcal genome is under continuous change, with genetic elements moving in and out. The species-independent conservation of SCCmec suggests horizontal transfer between staphylococci, and the extensive rearrangements observed in ccr- and mec-gene complexes indicate frequent exchange of genetic material within SCCmec. SCC has been described as a major gene acquisition machine serving as a genetic shuttle between staphylococci.
New variants of SCCmec will emerge that are not detected by the existing simplex or multiplex PCR methods. This will lead to a revision of PCR methods every time a new variant is discovered. Why should SCCmec typing continue? Detailed SCCmec typing is unsuitable for the routine laboratory, but important for epidemiological purposes, especially since SCCmec is part of the nomenclature of international MRSA clones. The significance of SCC variants should also be investigated, perhaps by SCCmec-typing larger collections of isolates. All this clearly indicates that a standardized typing method and nomenclature of SCCmec is needed.
Very little is known about the potential donors, recipients, mechanisms or direction of SCC transfer. We do not know which SCC came first –ccr, mecA, SCC or SCCmec – and whether horizontal transfer of SCCmec is a frequent or a rare event. Traditionally, CoNS were thought to be of little clinical importance. With the knowledge we have about SCCmec today, it may be worthwhile to focus on CoNS. They may be reservoirs of antibiotic resistance genes, and may perhaps be the driving force for the generation of new MRSA strains. There is a need to continue the search for the origin of SCCmec and to identify the route and mechanism of transfer as there is a common agreement that SCCmec in MRSA has been transferred from CoNS. In Norway, we are rather privileged because the antimicrobial resistance level is low, and this is probably due to the relatively low consumption of antibiotics. Having low MRSA endemicity but significant levels of MR-CoNS generates an excellent setting for ‘gene pool sharing studies.’ But how much longer can countries such as Norway escape the resistance problems observed in other countries? Whole genome sequencing has allowed us a glimpse into the world of the staphylococci and may improve our understanding of how resistant clones and resistance elements spread among staphylococci.
We would like to thank Dr Jaana Vuopio-Varkila and Professor Alex van Belkum for fruitful discussions. This work has been supported by funding from the University of Tromso, Norway; the University Hospital of North Norway, Norway; and Helse Nord, Norway.