Prevalence of Clonal Complexes and Virulence Genes among Commensal and Invasive Staphylococcus aureus Isolates in Sweden

Staphylococcus aureus encodes a remarkable number of virulence factors which may contribute to its pathogenicity and ability to cause invasive disease. The main objective of this study was to evaluate the association between S. aureus invasiveness and bacterial genotype, in terms of the presence of virulence genes and affiliation to clonal complexes. Also, the significance of different virulence genes, mainly adhesins, for the development of infective endocarditis was investigated. DNA microarray technology was used to analyze 134 S. aureus isolates, all methicillin-susceptible, derived from three groups of clinically well-characterized patients: nasal carriers (n=46), bacteremia (n=55), and bacteremia with infective endocarditis (n=33). Invasive isolates were dominant in four of the major clonal complexes: 5, 8, 15, and 25. Of the 170 virulence genes examined, those encoding accessory gene regulator group II (agr II), capsule polysaccharide serotype 5 (cap5), and adhesins such as S. aureus surface protein G (sasG) and fibronectin-binding protein B (fnbB) were found to be associated with invasive disease. The same was shown for the leukocidin genes lukD/lukE, as well as the genes encoding serine protease A and B (splA/splB), staphylococcal complement inhibitor (scn) and the staphylococcal exotoxin-like protein (setC or selX). In addition, there was a trend of higher prevalence of certain genes or gene clusters (sasG, agr II, cap5) among isolates causing infective endocarditis compared to other invasive isolates. In most cases, the presence of virulence genes was linked to clonal complex affiliation. In conclusion, certain S. aureus clonal lineages harboring specific sets of virulence genes seem to be more successful in causing invasive disease.


Introduction
S. aureus commonly colonizes the human skin and mucosal membranes without causing disease. However, it is also a wellknown pathogen causing a broad spectrum of infections ranging from superficial skin infections to invasive disease such as life-threatening bacteremia and infective endocarditis (IE). An increased incidence of S. aureus bacteremia has been reported [1], and it constitutes the most common etiology of IE in Sweden with increased incidence in recent years (unpublished data from the Swedish national infective endocarditis quality register). The bacterium produces several virulence factors which may contribute to its invasive potential [2], including surface-associated adhesins such as MSCRAMMs (Microbial Surface Components Recognizing Adhesive Matrix Molecules) as well as secreted virulence factors like exotoxins and enzymes. Capsular polysaccharides and regulators may also contribute to pathogenicity [2,3]. Since S. aureus is a commensal and acts as an opportunistic pathogen, the predisposing factors of the host, together with virulence factors of the microorganism might play a significant role for invasiveness.
Previous studies have investigated the importance of S. aureus clonality for invasive disease [4,5], as well as associations between different virulence genes and S. aureus disease [6][7][8][9][10][11][12], though the results have shown some contradictions. MSCRAMMs are among the factors of interest, as they are known to have the capacity to bind extra-cellular matrix proteins such as fibrinogen, fibronectin, and elastin, all of which are potentially important for the invasive capacity of S. aureus. More specifically, the association of adhesins such as clumping factors (Clf) and fibronectin-binding proteins (FnbP) and development of IE has previously been examined, although mainly in experimental studies [13][14][15]. Still, additional clinical research is needed to investigate the significance of bacterial genotype for S. aureus disease.
The present study aimed to evaluate the potential association between S. aureus invasive disease and bacterial genotype in terms of the presence of genes encoding adhesins and other virulence factors as well as affiliation to clonal complexes (CCs). Furthermore, it sought to investigate whether the prevalence of certain MSCRAMM genes is associated with S. aureus IE. To achieve these aims, we used DNA microarray technology to analyze S. aureus isolates derived from three groups of clinically well-characterized patients: nasal carriers, bacteremia, and bacteremia with IE.

Ethics statement
The study was conducted in accordance with the ethical guidelines of Declaration of Helsinki and was approved by the regional ethical committee in Örebro (Dnr 543/88) and Uppsala (Dnr 2011/3349). Written informed consent to participate was provided from all patients comprising the bacteremia group without IE (Dnr 543/88). Regarding the IE patients, the regional ethical committee in Uppsala approved the use of these clinical data entered in the Swedish quality register of infective endocarditis (Dnr 2011/3349). These IE patients were written informed that clinical data were recorded in a national quality register and could be used for research purposes.
Nasal isolates were obtained from swabs of anonymized patients intended for elective orthopedic surgery who were verbally informed regarding the purpose according to a Chairman decision of the regional ethical committee in Örebro 1992. For the IE patients and the anonymized patients whose nasal swabs were used, the regional ethical committee waived the need for written consent.

Bacterial isolates
A total of 134 methicillin-susceptible S. aureus (MSSA) were included in the study: 46 nasal carriage and 88 bacteremic isolates. Within the bacteremia group, 33 isolates originated from patients with IE. The isolates were evaluated according to origin (nasal carriage, bacteremia with and without IE).
All isolates were derived from patients treated at Örebro University Hospital, and included in the study after ethical approval. Most of the bacteremic isolates (61/88) and corresponding clinical data were prospectively collected from hospitalized patients with a diagnosis of S. aureus bacteremia during 1988-1992 [16]. This group included 6 patients with concomitant IE. An additional 27 S. aureus blood isolates from patients with IE were also included. These isolates originated from patients who had been treated for S. aureus IE during 2008-2011 and registered in the Swedish national IE quality register, a national database recording relevant clinical data, diagnostic procedures, treatment, and outcome. Nasal carriage isolates were obtained from nasal swabs of patients intended for elective orthopedic surgery during 1992. For each patient, only one episode of S. aureus bacteremia was included, with the exception of one patient who had two separate episodes of bacteremia with genotypically different isolates.
The S. aureus isolates were identified by routine microbiological methods such as coagulase and DNase tests. Isolates were stored at -70°C in preservation medium (yeast extract; Difco Laboratories, Sparks, MD, USA; and horse serum added to trypticase soy broth; BBL, Sparks, MD, USA).

Definitions
The diagnosis of S. aureus bacteremia was verified by at least one positive blood culture performed with the Bactec system (Becton Dickinson, USA)In order to evaluate the importance of different patient characteristics, the patients' medical records were carefully surveyed for underlying conditions such as diabetes mellitusrenal insufficiency requiring hemodialysis or peritoneal dialysis, present malignant diseaseand immunosuppression due to treatment with corticosteroids, chemotherapyIE: infective endocarditisFnbP: fibronectin-binding proteinsMLST: multilocus sequence typingCCs: clonal complexesCP: capsular polysaccharidePIA: polysaccharide intercellular adhesionSTs: sequence types or immunosuppressive disease such as leukemia. Presence of indwelling catheters (e.g. central venous catheter), pacemakers, cardiac valves, prosthetic joint devices, and osteosynthesis material were recorded. In addition we registered hematogenous complications such as IE, acute osteomyelitis, septic arthritis, deep seated abscesses, and meningitis. IE was defined according to the modified Duke criteria [17].

DNA microarray based genotyping
Genotyping was performed with the Alere StaphyType DNA microarray test (Alere Technologies GmbH, Jena, Germany), according to protocols and procedures described in detail elsewhere [18,19]. This DNA microarray includes 333 target sequences corresponding to approximately 170 distinct genes and their allelic variants, covering species markers, markers for the recognition of SCCmec, and capsule types as well as agr groups, resistance genes, exotoxins, MSCRAMM genes, and others. Primer and probe sequences have been published previously [19,20].
In short, monoclonal S. aureus cultures were grown on Columbia blood agar overnight at 37°C. Harvested cultures were enzymatically lysed using lysozyme, lysostaphin, and RNAse. DNA was prepared using commercially available spin columns (Qiagen, Hilden, Germany). The resulting RNA-free unfragmented DNA preparations were used as templates in a multiplexed primer elongation with only a single reverse primer per target. During that step, biotin-16-dUTP was incorporated into the amplicons. Labeled amplicons were hybridized to the microarray. After washing and blocking steps, horseradishperoxidase-streptavidin conjugate was added. Following further incubation and washing, hybridizations were visualized using a precipitating dye. An image of the microarray was recorded using a designated reader (Alere Technologies GmbH, Jena, Germany). Normalized intensities of the spots were calculated based on their average intensities and further analyzed as described previously [20].
The assignment of isolates to CCs or sequence types (STs), as defined by multilocus sequence typing (MLST) [21], was determined by an automated comparison of hybridization profiles to a collection of reference strains previously subjected to MLST [19].

Statistics
Proportions were compared with Fisher's exact test, using version 17 of the SPSS software package. P-values <0.05 were considered statistically significant.

Patient characteristics
Clinical data for patients with bacteremia, with and without IE, are shown in Table 1. No clinical data were available for the nasal carriage population. All 33 patients with IE fulfilled the criteria of definite IE according to the modified Duke criteria [17]. Of patients with IE, 19 of 33 (58%) had one or more factors predisposing for IE (underlying heart disease n=5; prosthetic heart valves n=3; previous history of IE n=4; injection drug use n=12). All patients except one were examined by echocardiography (28 transesophageal and 4 transthoracic). The diagnosis was confirmed by autopsy in two patients, one of whom did not undergo echocardiographic investigation. Twenty-three patients showed vascular embolization (osteomyelitis n=1; septic arthritis n=6; embolization to central nervous system/meningitis n=6; septic pulmonary infarcts n=11; skin and soft tissue n= 4). Eight patients underwent surgical intervention due to IE, and four patients relapsed shortly after the end of treatment or operation.

Agr groups
The distribution of isolates among agr groups I-IV according to carrier status and invasive disease is shown in Table 3, along with affiliation to CCs. Affiliation to agr groups was strongly linked to bacterial clonality in that all isolates assigned to a specific CC clustered to the same agr group (Table 3). A total of 46% (62/134) of the MSSA isolates were assigned to agr group I and only 6 isolates were assigned to agr group IV. Neither group showed differences regarding the clinical origin of bacterial isolates. However, within agr group II, 26 of 31 isolates were of invasive origin and 13 of these were from patients with IE, but in agr group III, 18 of 35 isolates were carriage isolates. Thus, agr group II was associated with invasive disease (p=0.017), while agr group III was linked with carriage status (p=0.022).

Surface-associated polysaccharides
All isolates carried capsular polysaccharide (CP) serotype 5 (29%) or 8 (71%); no CP1 was found. There was an association between invasiveness and CP5 (p=0.044); 15 of the 39 CP5 isolates (38%) were from patients with IE. The presence of cap genes was in accordance with affiliations to CCs (Figure 1), but there were no correlation between cap genes and agr groups.

MSCRAMMs
Genes encoding 15 different MSCRAMMs were analyzed (Table 4), and two of these (fnbB and sasG) were more prevalent among invasive isolates than carriage isolates (p=0.008 and p=0.004, respectively), while the converse was observed for cna (p=<0.0001). Carriage of sasG and cna was in concordance with CC affiliation (Figure 1). The genes bbp, ebh, map, and sdrD were found in most isolates (88-99%), regardless of clinical origin, while the remaining genes (clfA, clfB, ebpS, eno, fib, fnbA, sdrC, and vwb) were present in all isolates. In addition, bap, encoding a biofilm-associated protein was abscent in all isolates.

Genes encoding toxins
The hemolysin genes hla, hld, hl, and hlIII (loci and gene positions in Table 4) were present in all isolates except one hla-negative ST49 isolate and two hlIII-negative isolates within CC22. The hlb gene was found in 110/134 isolates (82%), with no difference between the invasive and the nasal carriage group. Of leukocidins, only one isolate carried the PVLencoding genes (lukF-PV, lukS-PV). However, lukD/lukE were more prevalent among invasive isolates (46/88) compared to nasal carriage isolates (12/46), and were thus significantly associated with invasive disease (p=0.006). Genes encoding gamma-toxin (lukF/lukS/hlgA) as well as leukocidin homologue family protein (loci and gene position in Table 4) were ubiquitous.
Genes encoding exfoliative toxins were absent (etB) or rare (etA 5%; etD 8%), and all etD-positive isolates were assigned to CC25. The tst1-gene was found in 22% of isolates, and was more prevalent among nasal carriage isolates compared to invasive isolates (p=0.005). Of enterotoxin genes, those within the egc cluster (seg+sei +sem+sen+seo+seu) were the most frequent (64%) ones, being linked to bacterial clonality ( Figure 1) and significantly associated with carriage isolates (p=0.034). The remaining enterotoxin genes were found in 0-19% of the isolates with no difference in prevalence with regard to clinical origin.
The presence of genes encoding proteases, splA and splB (loci and gene position in Table 4) was associated with invasive isolates (p=0.002), but there was no difference concerning splE. SspA, sspB, and sspP (loci and gene position in Table 4) were ubiquitous among isolates. SetC (also known as selx, locus and gene position in Table 4) was also found to be associated with invasive disease (p=0.004). Genes encoding HLB-converting phages (sak, chp, scn) were present in 80-95% of isolates but genes encoding epidermal cell differentiation inhibitors were absent (edinA, edinC) or rare (edinB; 8%). All isolates harbored the gene isdA, encoding transferrin-binding protein.

Genes associated with invasive disease
A further comparison between invasive non-IE isolates and IE isolates was performed in virulence genes which demonstrated a significant association with invasive disease (Table 5).

Discussion
S. aureus encodes a remarkable number of surfaceassociated as well as extracellular virulence factors. However, previous studies have shown contradictory results regarding the importance of these different virulence genes for invasiveness. Based on 134 MSSA isolates from three wellcharacterized patient populations, the aim of the present study was to evaluate the potential association between S. aureus disease and bacterial genotype with a focus on clonality and genes encoding MSCRAMMs. We used DNA microarray based analysis, which offers the great advantage of a simultaneous and fast analysis of a large number of virulence genes as well as assignment to clonal lineages at a very reasonable cost compared to multiple PCRs or whole genome sequencing.
The S. aureus isolates were distributed between six major CCs, mainly in concordance with previous studies from Europe [5,10,22] and the USA [4]. CC30 and CC45 comprised the largest number of isolates. Within the other four major CCs (5, 8, 15, and 25), invasive isolates dominated. In contrast, some previous studies have failed to show an association between invasive disease and clonality [5,9]. By studying less heterogeneous S. aureus infections, Fowler et al [4] could show an association between invasive disease with hematogenous complications and isolates within CC5 and CC30. The results of Fowler et al are in accordance with our results for CC5 but not for CC30. However, in their study it was the MRSA isolates within CC30 that were significantly associated with invasiveness. Our study only involved MSSA isolates, and CC30-MRSA are rare to virtually non-existent in Sweden. Thus epidemiological differences between the places where the studies were performed may explain this apparent discrepancy. Of isolates assigned to CC5, the majority originated from patients with IE. Although the total number of isolates within CC5 was limited, this result may suggest an association between CC5 and serious invasive S. aureus disease.
Due to polymorphisms in the gene encoding an autoinducing peptide in the agr locus, the S. aureus isolates are assigned to agr groups I to IV. As described in previous studies, the agr groups are strongly linked to bacterial clonality [8,19], although both the present study and previous work [19] showed that isolates within the same agr group may belong to genetically diverse CCs. In our study, isolates within agr II were significantly associated with invasive disease, and 50% of invasive isolates within agr II originated from patients with IE. Since agr II included isolates assigned to CC5 and CC15, and invasive isolates dominated within these clones, it is not possible to conclude whether this association is a consequence of clonality or assignment to agr group. A previous study by Jarraud et al also showed a positive relationship between agr II and IE [8].
As expected, all isolates carried CP serotype 5 or 8 [23]. Among the isolates that displayed CP5, invasive isolates dominated which is consistent with a previous study in which a CP5-positive strain in a mouse model showed a higher bacteremia level as well as enhanced capacity to persist in the bloodstream, compared to a CP8-positive strain [24]. It is also worth noting the even stronger correlation between CP5 and IE. However, as for agr group, CP type is linked to the affiliation to CCs, and so the association between CP5 and invasive disease as well as IE could be a consequence of clonality. The absence of correlation between capsule serotypes and agr groups is also in line with results in a previous study [19]. Due to the capacity of MSCRAMMs to bind extracellular matrix proteins, the significance of MSCRAMM genes for invasive disease and IE has been thoroughly examined previously [7,[11][12][13][14][15]25]. The present study showed a significant association between the presence of fnbB as well as sasG and invasive disease, in accordance with previous studies [26,27]. There was also a trend towards a higher prevalence of sasG among isolates from IE patients compared to other invasive isolates. The presence of sasG, but not fnbB, was in accordance with affiliation to CCs.
The results are however difficult to interpret, due to the overlapping properties of many adhesins. Besides binding fibronectin, both FnbA and FnbB have the capacity to bind elastin; although unlike FnbA, FnbB has not been shown to promote platelet aggregation [28] or bind fibrinogen. The overall capacity to bind fibronectin is not crucial based on whether the isolate carries one or both fnb genes [27]. This may indicate that the additional role of FnbB, such as binding elastin, may be important for S. aureus invasive capacity. The lack of evidence for FnbB promoting platelet aggregation is consistent with our findings, which showed no difference in fnbB gene prevalence between the invasive non-IE isolate group and the IE isolate group.
The SasG protein shows opposing characteristics; it prevents adhesion to extracellular matrix components such as fibronectin, cytokeratin 10, and IgG, while at the same time it promotes adhesion to nasal epithelial cells as well as biofilm formation [29]. The gene sasG shares sequence similarity with a plasmin sensitive protein encoded by the gene pls, in which mutation has been correlated with reduced invasion of host cells [30]. It is thus plausible that sasG also plays an important role for S. aureus invasiveness. Unexpectedly, and contrary to previous studies [7,31,32], cna was more common in nasal carriage isolates than in invasive isolates. An explanation for this finding could be that cna is also a CC-associated marker. Since CC30 is overrepresented among carrier isolates and since all CC30 isolates also carry cna, the apparent association of this gene with carriage might be a function of the clonality of the strains in question. Furthermore, cna is the most important collagen-binding adhesin protein of S. aureus [33], and this feature may also be important for adherence to the nasal epithelial cells.
Several other MSCRAMM genes shown to be important for invasiveness and development of IE are highly conserved in the S. aureus genome [34]. This is in accordance with our results, where all isolates harbored the genes clfA, clfB, fnbA, and sdrC. On the other hand, bap, which encodes a biofilmassociated protein, was absent in all isolates, which is consistent with the fact that this gene has so far only been reported in animal strains [35].
As expected, hla was present in almost all isolates, with some occasional exceptions probably due to single mutations. Similarly, hlb was also present at high frequency, though its function depends on whether the hlb-converting prophage is integrated or not. In contrast to hemolysin genes, those encoding exfoliative toxins were rare, which was expected since patients with exfoliative staphylococcal disease were not specifically included in the present study. Egc was the most prevalent among enterotoxin genes, and as shown before linked to certain CCs and negatively associated with invasive disease [10,36]. Among leukocidins, the prevalence of lukD +lukE was high, linked to CCs and significantly associated with invasive disease, also shown by Eiff et al [37]. The splA/splB genes were significantly associated with invasive disease, as well as setC.
A limitation of our study is that we have merely investigated the presence of genes harbored by S. aureus, rather than their expression, and that analysis was restricted to those alleles of genes covered by the DNA microarray. Moreover, specific patient-related risk factors such as immunosuppression, intravenous drug use, and indwelling medical devices may predispose for invasive S. aureus disease. However, we included isolates from patients with these risk factors, as it might be a reason why not all patients with predisposing medical conditions get invasive disease. Conversely, a major advantage of the study is the well-characterized patient population.
In conclusion our study indicates that invasive S. aureus isolates are related to certain CCs. We also found a significant association between invasiveness and genes encoding CP type as well as specific MSCRAMMs such as fnbB and sasG. Moreover, our results suggest a trend toward even higher prevalence of certain virulence genes among isolates causing IE compared to other invasive isolates. However, in most cases the presence of virulence genes was linked to CC affiliation. It is also known that certain clonal lineages carry specific sets of virulence genes [20]. Even though S. aureus isolates from all CCs can cause invasive disease, it is reasonable to assume that certain S. aureus clonal lineages harboring specific sets of virulence genes are more successful at causing an invasive disease.
Against the background of increased life expectancy, implementation of more advanced medical interventions, and the emergence of antibiotic resistance, continuous surveillance of the incidence of S. aureus bacteremia is important. There may be continuous increase in the incidence of S. aureus bacteremia as well as changes in the structure and composition of the bacterial population. A study to further investigate any changes in the molecular epidemiology of invasive S. aureus isolates over the last 3 decades is currently ongoing.