MRSA epidemic linked to a quickly spreading colonization and virulence determinant.

The molecular processes underlying epidemic waves of methicillin-resistant Staphylococcus aureus (MRSA) infection are poorly understood(1). Although a major role has been attributed to the acquisition of virulence determinants by horizontal gene transfer(2), there are insufficient epidemiological and functional data supporting that concept. We here report the spread of clones containing a previously extremely rare(3,4) mobile genetic element–encoded gene, sasX. We demonstrate that sasX has a key role in MRSA colonization and pathogenesis, substantially enhancing nasal colonization, lung disease and abscess formation and promoting mechanisms of immune evasion. Moreover, we observed the recent spread of sasX from sequence type 239 (ST239) to invasive clones belonging to other sequence types. Our study identifies sasX as a quickly spreading crucial determinant of MRSA pathogenic success and a promising target for therapeutic interference. Our results provide proof of principle that horizontal gene transfer of key virulence determinants drives MRSA epidemic waves.

S. aureus remains a dangerous pathogen, causing a multitude of serious and sometimes life-threatening diseases around the globe. This is mostly due to widespread antibiotic resistance, which considerably complicates the treatment of S. aureus infections[5]. MRSA strains in particular represent a major problem for public health systems, as they combine resistance to methicillin and other β-lactam antibiotics with relatively minor fitness costs, resulting in high infiltration of hospital and community settings[6]. Remarkably, in the U.S., the estimated number of deaths due to MRSA infections exceeds that due to HIV/AIDS[7]. While molecular typing and genome sequencing have provided better insight into MRSA epidemiology in recent years[8], we have a severe lack of knowledge regarding the mechanisms by which specific molecular determinants cause the spread and pathogenic success of MRSA clones.

Only a very limited number of MRSA clones, belonging to five clonal complexes, are responsible for the majority of MRSA infections worldwide[1]. Specific STs are characteristic for a given geographical location. In China, as in most Asian countries, ST239 is predominant[9]. Recently, genome sequencing of an ST239 strain revealed a novel gene of unknown function (SATW20_21850), which is present in only three of the 43 global ST239 strains that have been sequenced[3,4,10], and absent from all other known S. aureus genomes[3]. Based on its amino acid sequence, which contains a signal peptide and an LPXTG surface-anchoring motif, this gene encodes a secreted, surface-anchored protein of 15 kDa in its processed form. It is located at the 3' end region of a 127.2-kb ΦSPβ-like prophage, which lacks other virulence-associated genes[3]. There is no similarity of the surface protein to other proteins found in the NCBI databases except for the highly similar, ΦSPβ-encoded SesI of S. epidermidis RP62A[11]. We termed the gene sasX and showed that it is absent from other major global MRSA strains by analyzing a series of strains from divergent clonal and geographical backgrounds ( online).

Prompted by the epidemiological connection of sasX to MRSA ST239, we here evaluated the hypothesis that sasX represents a key factor determining the spread of MRSA in China[9]. To that end, we determined ST distribution and frequency of sasX-positive clones among > 800 randomly selected S. aureus isolates from patients with S. aureus infections at three large teaching hospitals located in eastern China between 2003 and 2011 (). The situation in China in terms of MRSA infection rate and dominance of the ST239/SCC-III (Brazil/Hungary) clone is largely characteristic of that in most other Asian countries except Japan and South Korea, and Australia[9,12-15]. Accordingly, ST239 clones were also predominant among the isolates investigated in our study, with ST5 being the second most frequently isolated ST ( online). Notably, the frequency of sasX-positive invasive S. aureus clones increased significantly from 2003 to 2011 (from 19% to 31%, P = 0.0026). Among MRSA isolates, the frequency of sasX increased from 21% to 39% (P = 0.0028), while the increase among methicillin-sensitive S. aureus (MSSA) was not significant (). Furthermore, sasX was almost exclusively found among ST239 strains in 2003-2005, while its frequency among isolates of several other STs considerably increased since then (increase from 5% to 28% of non-ST239 clones among sasX-positive S. aureus between 2003-2005 and 2009-2011, P = 0.0048; ). Moreover, presence of sasX was extremely rare (only one positive clone among 169 tested) among a collection of community isolates from healthy individuals in the same areas ( online), indicating that sasX-positive clones spread predominantly within the hospital setting. Finally, we did not detect ST239 clones among community isolates, indicating that this ST is linked to hospital infections. Analysis of genes surrounding sasX in randomly selected clones of different STs showed that the horizontal transfer of sasX always occurs linked to the ΦSPβ-like prophage with a conserved phage insertion site, although occasional deletions within the phage were observed ( online). Together, these findings demonstrate an epidemic wave of hospital-associated sasX-positive MRSA in China. Furthermore, they indicate events of horizontal gene transfer resulting in the spread of sasX from ST239 to other STs and a causative link between the presence of sasX and invasiveness.

To understand the link between sasX and infection, we set out to elucidate the biological role of sasX. We first demonstrated that the SasX protein is expressed at the S. aureus cell surface using immunofluorescence and immunoblot detection (). Surface proteins in S. aureus fulfill a variety of functions that include first and foremost adhesion to host tissue[16]. In particular, adhesion to nasal epithelial cells is of major importance for the colonization, spread and virulence potential of S. aureus strains, inasmuch as the nares are the predominant location of S. aureus colonization in the human body[17] and there is an epidemiological link between nasal carriage and infection[18]. However, the molecular basis of nasal colonization is incompletely understood and there are only a few molecular factors of S. aureus that have been identified as potentially responsible for nasal colonization, such as teichoic acids[19] and specific surface proteins[20-22]. To measure the impact of sasX on nasal colonization, we constructed isogenic sasX deletion mutants in representative ( online) ST239 and ST5 strains. Adhesion of the sasX mutant strains to human nasal epithelial cells in vitro was significantly impaired compared to the wild-type strains, and restored in sasX-complemented mutant strains (). Furthermore, recombinant SasX protein blocked adhesion of S. aureus to human nasal epithelial cells (; online). Moreover, sasX had a significant impact on in-vivo nasal colonization in separate and competitive colonization models (). These data demonstrate that the SasX surface protein promotes nasal colonization.

In addition to facilitating adhesion to host tissue, S. aureus surface proteins may cause intercellular bacterial aggregation, a phenotype with multiple consequences, such as increased biofilm formation and immune evasion capacity[23-25]. Indeed, we found that sasX facilitated the formation of large bacterial aggregates ( online) and promoted biofilm formation (). Interestingly, sasX also facilitated primary attachment to an abiotic surface (), which is in keeping with our finding that sasX expression was highest during early growth, and decreased afterwards ( online). Moreover, phagocytosis of the wild-type strain by human neutrophils was significantly lower than that of the sasX mutant strain (). Accordingly, survival in human blood () and lysis of human neutrophils as a sign of prolonged bacterial survival after phagocytosis () were significantly increased in the wild-type strain compared to the sasX deletion mutant. Notably, these effects are most likely caused predominantly by the aggregation phenotype and not by a gene regulatory effect, as sasX did not impact expression of the ica genes encoding biofilm exopolysaccharide biosynthesis, the biofilm and virulence regulator agr[26], or a series of cytolysins and other surface protein genes ( online). Together, these data show that sasX facilitates intercellular aggregation, leading to a significant enhancement of S. aureus immune evasion mechanisms.

S. aureus most frequently causes skin and respiratory infections[27]. Correspondingly, these were also the most abundant types of infection detected in our study ( online). Among infections caused by sasX-positive clones, respiratory and pleural-cavity infections as well as abscesses were most prominent (significant difference by type of infection, P = 0.0139, chi square test). Therefore, to determine whether sasX is a virulence factor during S. aureus infection, we selected to perform murine abscess and lung infection models. Mice infected with the wild-type strains developed abscesses that were significantly larger than those in mice infected with the sasX mutant strains (). Furthermore, microscopic evaluation showed increased infiltration of inflammatory cells in the abscesses caused by the wild-type strain (). Of note, the strong impact of sasX on abscess size is similar to that observed with potent S. aureus toxins such as α-toxin[28] or phenol-soluble modulins[29]. In addition, lung wet weight/body weight ratios, degree of inflammation in the lung as assessed by tumor necrosis factor αconcentration, and histological examination showed that sasX is also a major factor influencing S. aureus pathogenesis in the lungs (). Thus, our data classify sasX as a key determinant of MRSA skin and lung infection.

In conclusion, we identified sasX as a critical factor promoting nasal colonization, immune evasion and virulence, and a probable main driving force of the Asian MRSA epidemic. Furthermore, our data provide strong support to the notion that acquisition of specific molecular determinants via horizontal gene transfer represents a major mechanism enhancing the pathogenic potential and epidemiological success of MRSA clones. Moreover, based on our data, it is to be expected that the frequency of sasX among Asian and international MRSA clones will increase further. We thus propose drug development or vaccination efforts aimed at SasX to prevent MRSA colonization and infection.

Methods

For additional procedures, see Supplementary Methods (.

Bacterial isolates

807 S. aureus isolates were randomly selected from different inpatients with S. aureus infections at three different teaching hospitals in China: 100 each from 2004, 2007, and 2010 from Huashan Hospital, Fudan University, Shanghai; 107 from 2006-2008 and 79 from 2009-2010 from Shandong Provincial Hospital, Shandong University, Jinan; 104 from 2003-2005, 105 from 2006-2008 and 112 from 2009-2011 from the first affiliated hospital of Wenzhou Medical College, Wenzhou. 49 S. aureus isolates from nasal swabs among more than 500 healthy volunteers in Shanghai and 120 from nasal swabs among more than 1000 healthy volunteers in Wenzhou were also collected as part of a population-based community prevalence study. Characteristics of all isolates are listed in online. Bacteria were identified as staphylococci by classic microbiological methods: Gram's stain, catalase and coagulase activity on rabbit plasma. S. aureus strains were further categorized by biochemical characterization using the Api-Staph test (BioMérieux, Lyon, France). All strains and plasmids used in this study are listed in online. All oligonucleotides used for cloning or polymerase chain reaction procedures are listed in online.

Nasal colonization model

All animal work was approved by the ethics committee of Fudan University. Female ICR mice were used for the nasal colonization model (ten mice per group). All mice were six to eight weeks of age at the time of use and received drinking water containing penicillin at 100 μg ml-1. S. aureus strains were grown to mid-exponential growth phase, washed, and resuspended in sterile phosphate-buffered saline (PBS) at 107 colony-forming units (CFU) per μl. Mice were anesthetized with isoflurane. The inoculum, which contained 108 CFU in 10 μl of PBS or PBS alone, was pipetted slowly into the nares of the anesthetized mice without touching the nose with the pipette tip. Seven days after inoculation, ten mice per strain were euthanized and evaluated for nasal carriage of S. aureus. The nasal region was wiped externally with 70% ethanol, and the nasal tissue was homogenized in 0.5 ml of tryptic soy broth (TSB). The total number of S. aureus CFU per nose was assessed by plating 100 μl diluted nasal suspensions on TSB agar containing penicillin (100 μg ml-1).

For the competitive nasal colonization model, 15 mice per group were used. S. aureus ST239 (HS663) and its isogenic sasX deletion mutant were grown to mid-exponential growth phase, washed, and resuspended at equal CFUs in sterile PBS at a total of 2×106 CFU per μl. The following steps were similar to above, except that bacteria isolated from the noses at day seven were plated on penicillin and penicillin/tetracycline plates to distinguish wild-type from mutant bacteria.

Lung infection model

Inoculation was performed as described for the colonization model, but with 109 CFU per 20 μl. S. aureus cell suspensions were pipetted into the nares of the anesthetized mice (n=10 per group) without touching the nose. At 5 days after inoculation, mice were euthanized. The lungs from selected groups of animals were excised, washed with saline, and one lobe was fixed in 10% formalin (Sigma). Paraffin embedding and hematoxylin & eosin (HE) staining were performed as previously described[30]. The other lobe was homogenized in 0.5 ml of TSB; homogenized lung tissue and serum samples were used for TNF-α detection.

Skin abscess model

Female outbred, immune-competent, hairless mice were used for the abscess model. All mice were between four and six weeks of age at the time of use. S. aureus strains were grown to mid-exponential phase, washed once with sterile PBS, then resuspended in PBS at 1 × 109 CFU per 100 μl. Mice were anesthetized with isoflurane and inoculated with 100 μl PBS containing 109 live S. aureus or saline alone in the right flank by subcutaneous injection. We examined test animals at 24-h intervals for a total of 10 d with a caliper. We applied length (L) and width (W) values to calculate the area of abscesses with the formula L × W. Abscesses as shown in Fig. 4a developed into scarred lesions during the time of the experiment. The sizes of these lesions showed similar differences as the closed abscesses between the wild-type and sasX mutant groups and remained until the end point of the experiment ( online). However, only the sizes of closed abscesses were measured. All animals were euthanized after completion of the entire procedure.

Figure 4

SasX is a key virulence determinant during MRSA skin and lung infection. (a,d) Skin infection. Outbred, immune-competent, hairless mice were inoculated with 109 CFU of the ST239 or ST5 wild-type and isogenic sasX mutant strains. (b,c,e) Lung infection. Female ICR mice were inoculated in the noses with 109 CFU/20 μl of the ST239 or ST5 wild-type and isogenic sasX mutant strains. Mice were euthanized at day five post infection for subsequent analyses. (a) Skin infection model, developing abscess areas (see representative images on the right). Differences between corresponding wild-type and mutant abscess sizes were statistically significant at every time point (unpaired t-tests, P < 0.05). (b) Lung infection model, lung wet weight/body weight ratios. (c) Lung infection model, concentration of tumor necrosis factor α (TNF-α) in lung tissue samples. (d) Skin infection model, hematoxylin/eosin (HE) stained tissue samples. (e) Lung infection model, HE stained tissue samples. Note increased number of infiltrating inflammatory cells in wild-type sample pictures in (d) and extensive inflammation accompanied by hemorrhagic infiltration and disruption of pulmonary architecture in wild-type sample pictures in (e). Statistical analyses in (b,c) are by unpaired t-tests.