Molecular Epidemiology of Methicillin-Susceptible and Methicillin-Resistant Staphylococcus aureus in Wild, Captive and Laboratory Rats: Effect of Habitat on the Nasal S. aureus Population

Rats are a reservoir of human- and livestock-associated methicillin-resistant Staphylococcus aureus (MRSA). However, the composition of the natural S. aureus population in wild and laboratory rats is largely unknown. Here, 144 nasal S. aureus isolates from free-living wild rats, captive wild rats and laboratory rats were genotyped and profiled for antibiotic resistances and human-specific virulence genes. The nasal S. aureus carriage rate was higher among wild rats (23.4%) than laboratory rats (12.3%). Free-living wild rats were primarily colonized with isolates of clonal complex (CC) 49 and CC130 and maintained these strains even in husbandry. Moreover, upon livestock contact, CC398 isolates were acquired. In contrast, laboratory rats were colonized with many different S. aureus lineages—many of which are commonly found in humans. Five captive wild rats were colonized with CC398-MRSA. Moreover, a single CC30-MRSA and two CC130-MRSA were detected in free-living or captive wild rats. Rat-derived S. aureus isolates rarely harbored the phage-carried immune evasion gene cluster or superantigen genes, suggesting long-term adaptation to their host. Taken together, our study revealed a natural S. aureus population in wild rats, as well as a colonization pressure on wild and laboratory rats by exposure to livestock- and human-associated S. aureus, respectively.


Introduction
Staphylococcus aureus (S. aureus) is a major opportunistic pathogen in human medicine and it is increasingly recognized as a zoonotic pathogen. The nasal cavity of approximately 20% of adult humans is persistently and asymptomatically colonized with S. aureus [1][2][3]. These bacteria can cause a wide variety of illnesses, ranging from skin and soft tissue infections (e.g., abscesses) to life-threatening diseases (e.g., endocarditis and sepsis) [4]. The emergence of multiresistant (methicillin-resistant S. aureus, MRSA; vancomycin-resistant S. aureus, VRSA) and highly virulent (community acquired MRSA, CA-MRSA) strains makes S. aureus a prominent example of the antibiotic resistance crisis and a major public health concern worldwide [5][6][7]. This medical predicament is especially heightened by the fact that to date, no S. aureus vaccine is available [8].
Besides humans, S. aureus is also able to colonize and infect numerous other species, including companion animals, livestock as well as wild animals [9][10][11][12][13]. Transmission of MRSA occurs primarily upon person-to-person contact, but MRSA can also spread between domestic animals and people [14,15]. The recent detection of MRSA in rats suggests that pest animals might be an additional reservoir for MRSA [16][17][18][19][20]. For instance, wild rats caught on farms were carrying typical livestock-associated (LA)-MRSA strains (e.g., clonal complex (CC) 398 and sequence type (ST) 97) [18,21], while urban rats carried the same MRSA lineages prevalent in the respective local human or animal population (e.g., USA300-ST8) [19]. The natural S. aureus population in rats, which is likely dominated by methicillin-susceptible S. aureus (MSSA), is largely unknown [21]. In view of the emerging zoonotic potential of certain S. aureus lineages [22][23][24], an effective "One Health" strategy against the spread of S. aureus and especially MRSA has to take these bidirectional transmissions between animals and humans into account [25,26].
The ability of S. aureus to colonize various host species is favored by its plastic genome, which encompasses the core genome (75%), the core variable genome (10%) and the accessory genome (15%).
The core variable genome contains the information for surface proteins and transcription regulators, while the accessory genome consists mainly of mobile genetic elements (MGEs), and thus represents the most variable part of the genome [27][28][29]. Unsurprisingly, the accessory genome plays a major role in adaptation processes of S. aureus [30], allowing it to adapt to changing environments, such as different host species. Prominent examples include: (i) lack of antibiotic resistance genes and the highly human-specific immune evasion gene cluster (IEC) in animal isolates; (ii) lack of superantigen genes in murine S. aureus isolates; (iii) acquisition of variants of the von Willebrand factor-binding protein in ruminant and equine strains, carried by highly mobile pathogenicity islands (SaPIs); as well as (iv) a single nucleotide mutation in the dltB gene, a gene involved in lipoteichoic acid biosynthesis, which enabled a human strain to infect rabbits [10,11,[31][32][33][34].
These host adaptation processes frequently restrict the host range that can be successfully colonized and infected by a certain S. aureus lineage. In consequence, certain CCs seem to have a restricted host spectrum, like ST5 in poultry or ST433 in pigs [35,36], while others, like CC8, CC22 and CC398, display an extended host spectrum and are therefore called Extended-Host-Spectrum Genotypes (EHSGs) [24,37,38].
Along with mice, rats are a common animal model for studying S. aureus pathogenicity and new intervention strategies [39][40][41]. While investigating the epidemiology of S. aureus in both wild and laboratory mice in previous studies, we have found distinct clonal lineages and hints for host adaptation [11,34,42]. Moreover, certain mouse-and vole-adapted S. aureus strains turned out to be suitable tools to improve murine infection models by allowing persistent intranasal colonization or strongly reducing the required inoculation dose in infection models [42,43]. Despite their importance in infection models, the prevalence of S. aureus in laboratory rats, and its mechanisms of host adaptation, has yet to be investigated.
The major aim of this study was to extend our knowledge of S. aureus (esp. MRSA) carriage in rats. This was achieved by collecting S. aureus strains from free-living and captive wild rats, as well as laboratory rats, followed by determining the population structure and antibiotic resistance profiles of these S. aureus isolates. Moreover, we looked for signs of host adaptation by screening for bacterial genes known to be involved in this process.

Laboratory Rats and Wild Rats Are Colonized with S. aureus
This study aimed at determining the prevalence and molecular epidemiology of S. aureus, including MRSA, in free-living and captive wild rats, as well as laboratory rats. Therefore, 145 free-living wild rats captured between 2009 and 2017 in three German federal states as well as in the Czech Republic were analyzed for S. aureus nasal carriage (Tables S1 and S2). In addition, a total of 188 captive wild rats originating from three German federal states and held in laboratory housing or large enclosures for several years were investigated for S. aureus nasal carriage (Tables S3 and S4). Furthermore, 114 laboratory rats, raised as experimental or feeder animals, from four German federal states were included (Tables S3 and S5). Finally, S. aureus isolates from 52 laboratory rats held in Germany, USA, Japan and Canada were analyzed (Table S5).
Overall, the S. aureus carriage rate was similar among free-living wild rats (37/145; 25.5%) and captive wild rats (41/188; 21.8%), but lower among laboratory rats (14/114; 12.3%) ( Table 1). However, within the different categories, the prevalence differed strongly according to the geographical location. In free-living wild rats, the highest prevalence was seen in the Moravian-Silesian Region (MSR, 62.1%), followed by the German states Mecklenburg-Western Pomerania (MV, 22.2%), and then North Rhine-Westphalia (NRW, 17.3%) ( Table 1 and Table S6). The lowest prevalence was observed in pest animals from a zoo in Baden-Württemberg (BW, 5.9%). This divergence was even more pronounced among captive wild rats, where breeding procedures and hygiene measures can promote or prevent the spread of S. aureus, respectively. For instance, S. aureus prevalence was very high (77.1%) in captive wild rats in Berlin (BE) ( Table 1, Tables S4 and S7). In contrast, one wild rat population caught at a farm and afterwards held in husbandry in NRW was completely S. aureus-free. Moreover, we observed S. aureus-positive (e.g., Neufels, Tilbury) and -negative (e.g., BSR, WPHR) rat strain populations in the same husbandry in Brandenburg (BB) (Tables S4 and S7). This suggests that the transmission between strains is effectively blocked by hygiene measures. Colonization rates among laboratory rats were also diverse and ranged from 0.0% (BW, NRW) to 33.3% (MV). 1 Free-living wild rats: rats that were collected in the wild between 2009 and 2017 using live, electric as well as snap traps, in addition to bait poisons; captive wild rats: offspring of wild rats that were bred in a laboratory setting or a large enclosure for several years or even decades; laboratory rats: domesticated rats raised for animal experiments or as feeder animals. 2 For captive wild rats, the state where the animal husbandry is located is provided. 3  The S. aureus population structure in our described rat cohorts was resolved by spa typing. Free-living wild rats were predominantly colonized with two S. aureus lineages: CC130 (N = 21) and CC49 (N = 15) ( Figure 1A and Table S6). In addition, a single CC30 isolate was detected in a free-living wild rat caught in a German town.
Interestingly, captive wild rats held in laboratories or large enclosures for several years were colonized either with the above described rat-related lineages CC130 (N = 7) and CC49 (N = 1), the wild mouse-related lineage ST890 (N = 5) or with the livestock-associated lineage CC398 (N = 28) ( Figure 1A and Table S7) [11]. For instance, wild rats whose predecessors were captured on a pig farm in 2015 (strain Neufels) and kept under laboratory settings in two separate husbandries (BB and BE) were mostly (28/41; 68.3%) colonized with CC398 (spa type t011) (Tables S4 and S7). However, while the BB rats were colonized with CC398-MRSA (mecA-positive; see below), the BE rats carried CC398-MSSA. These data imply that contact with livestock is a significant risk factor for the acquisition of LA-S. aureus isolates. In contrast, the offspring of wild rats captured in a city were colonized with CC130-MSSA and ST890-MSSA. S. aureus lineages commonly found in humans were not detected in these animals.
On the other hand, laboratory rats raised as experimental animals or feeder animals harbored a broad range of lineages ( Figure 1A and Table S8). The most frequently detected lineage was CC15 (N = 15), followed by CC8 (N = 10), CC88 (N = 9), CC7 (N = 6), CC1 (N = 5), CC5 (N = 5), and CC188 (N = 5). Many of those are common in the human population, e.g., CC1, CC5, CC7, CC8, CC15, and CC20 [44]. In addition, we detected a novel multilocus sequence type (MLST) in this cohort in a laboratory rat from BW (MLST5598; spa type t2091). Notably, there was no overlap between the S. aureus populations in laboratory rats and wild rats (both free living and captive). we detected a novel multilocus sequence type (MLST) in this cohort in a laboratory rat from BW (MLST5598; spa type t2091). Notably, there was no overlap between the S. aureus populations in laboratory rats and wild rats (both free living and captive).  Table  1. In addition, 52 S. aureus strains isolated from laboratory rats from Germany, USA, Japan and Canada and characterized in this study were included in the analysis. The S. aureus population structure was resolved by spa typing; related spa types were grouped into CCs. (A) Isolates were clustered according to their origin (free-living wild rats, captive wild rats, laboratory rats). S. aureus isolates from free-living wild rats with direct contact with livestock or from captive wild rats whose ancestors have had contact with livestock are depicted in a lighter shade. (B) Prevalence of penicillin resistance in the different lineages. Minimum inhibitory concentration (MIC) values for penicillin were determined with the broth microdilution method.
To determine whether there are differences in the geographical distribution of the S. aureus lineages, all determined lineages were assigned to a map. For captive wild rats, the location of the husbandry, rather than their capture location, is depicted. Among wild rats (both free-living and captive), CC130 was most widely spread ( Figure 2A). This lineage was detected in several rats from five German federal states (BB, BE, BW, MV, and NRW), as well as in the Czech Republic (Moravian-Silesian Region, MSR). Similarly, CC49 was detected in both free-living and captive wild rats originating from BE, NRW and the Czech Republic (MSR).
Among the laboratory rats, CC8, CC15 and CC188 were most widely spread, with CC8 being detected in rats from three continents (Figure 2A, B). CC1, CC7, and CC88 were observed in two locations, while the remaining lineages were detected only at a single location. The validity of these data is, however, restricted by the limited sampling size in some geographical locations (Table 1).

Figure 1.
Wild rats and laboratory rats differ in their colonizing S. aureus population and the prevalence of penicillin-resistant S. aureus. S. aureus isolates were obtained from free-living as well as captive wild rats, in addition to laboratory rats. For a definition of the different rat categories, please refer to the legend of Table 1. In addition, 52 S. aureus strains isolated from laboratory rats from Germany, USA, Japan and Canada and characterized in this study were included in the analysis. The S. aureus population structure was resolved by spa typing; related spa types were grouped into CCs. (A) Isolates were clustered according to their origin (free-living wild rats, captive wild rats, laboratory rats). S. aureus isolates from free-living wild rats with direct contact with livestock or from captive wild rats whose ancestors have had contact with livestock are depicted in a lighter shade. (B) Prevalence of penicillin resistance in the different lineages. Minimum inhibitory concentration (MIC) values for penicillin were determined with the broth microdilution method.
To determine whether there are differences in the geographical distribution of the S. aureus lineages, all determined lineages were assigned to a map. For captive wild rats, the location of the husbandry, rather than their capture location, is depicted. Among wild rats (both free-living and captive), CC130 was most widely spread ( Figure 2A). This lineage was detected in several rats from five German federal states (BB, BE, BW, MV, and NRW), as well as in the Czech Republic (Moravian-Silesian Region, MSR). Similarly, CC49 was detected in both free-living and captive wild rats originating from BE, NRW and the Czech Republic (MSR).
Among the laboratory rats, CC8, CC15 and CC188 were most widely spread, with CC8 being detected in rats from three continents (Figure 2A,B). CC1, CC7, and CC88 were observed in two locations, while the remaining lineages were detected only at a single location. The validity of these data is, however, restricted by the limited sampling size in some geographical locations (Table 1).

Penicillin Resistance Was Low in Wild Rats but High in Rats with Livestock Contact
We have previously reported that murine S. aureus isolates show a lower prevalence of penicillin resistance than human strains [34]. To extend these findings, we determined the MIC of penicillin for all rat-derived S. aureus isolates using the broth microdilution method. Overall, 95/144 (66.0%) isolates were penicillin-resistant, however the prevalence varied significantly between the groups (Chi-Squared test, p < 0.001). While the resistance rate was low in S. aureus isolates from free-living wild rats (15/37; 40.5%), it averaged 73.2% (30/41) in captive wild rats and 75.8% (50/66) in laboratory rats. The highest rate of penicillin resistance (96.3%; 26/27) was observed in rats with livestock contact (Tables 1, S6 and S7). This pattern is also reflected by varying rates of penicillin resistance in S. aureus lineages associated with either free-living wild rats (CC49, CC130), captive wild rats (CC130, ST890), laboratory rats (various CCs) and captive wild rats with previous livestock contact (CC398) ( Figure 1A, B).

MRSA Was Detected among Wild Rats, but Not among Laboratory Rats
In general, several wild rats were colonized with MRSA, while the laboratory rat population was free of MRSA (Table 2). In total, 6/188 (3.2%) captive wild rats and 2/145 (1.4%) free-living wild rats were colonized with MRSA. Five out of six MRSA-positive captive wild rats were colonized with CC398-MRSA. Interestingly, all affected rats were derived from the same rat strain (Neufels) in BB within the same year. These animals originated from black rats (Rattus rattus) caught at a livestock farm in Southern Germany and have been bred in BB since 2015. All five strains belonged to the same spa-type (t011), carried the

Penicillin Resistance Was Low in Wild Rats but High in Rats with Livestock Contact
We have previously reported that murine S. aureus isolates show a lower prevalence of penicillin resistance than human strains [34]. To extend these findings, we determined the MIC of penicillin for all rat-derived S. aureus isolates using the broth microdilution method. Overall, 95/144 (66.0%) isolates were penicillin-resistant, however the prevalence varied significantly between the groups (Chi-Squared test, p < 0.001). While the resistance rate was low in S. aureus isolates from free-living wild rats (15/37; 40.5%), it averaged 73.2% (30/41) in captive wild rats and 75.8% (50/66) in laboratory rats. The highest rate of penicillin resistance (96.3%; 26/27) was observed in rats with livestock contact (Table 1, Tables  S6 and S7). This pattern is also reflected by varying rates of penicillin resistance in S. aureus lineages associated with either free-living wild rats (CC49, CC130), captive wild rats (CC130, ST890), laboratory rats (various CCs) and captive wild rats with previous livestock contact (CC398) ( Figure 1A,B).

MRSA Was Detected among Wild Rats, but Not among Laboratory Rats
In general, several wild rats were colonized with MRSA, while the laboratory rat population was free of MRSA (Table 2). In total, 6/188 (3.2%) captive wild rats and 2/145 (1.4%) free-living wild rats were colonized with MRSA. Five out of six MRSA-positive captive wild rats were colonized with CC398-MRSA. Interestingly, all affected rats were derived from the same rat strain (Neufels) in BB within the same year. These animals originated from black rats (Rattus rattus) caught at a livestock farm in Southern Germany and have been bred in BB since 2015. All five strains belonged to the same spa-type (t011), carried the methicillin-resistance gene mecA and showed an identical resistance pattern, suggesting the maintenance and spread of a single MRSA clone during captive breeding (Table 2). Moreover, we detected two mecC-encoding CC130 isolates (spa type t843) in wild rats. One of the rats was caught in a zoo as pest rodent, while the other was a captive wild rat. In addition, a Norway rat positive for CC30-MRSA encoding the mecA gene was caught in a city. All three isolates were phenotypically resistant to cefoxitin, but sensitive to oxacillin, suggesting a heterogenous expression of the PBP2a protein. mecBand mecD-positive S. aureus isolates were not detected in this study.

S. aureus Isolates from Rats Show Features of Host Adaptation
We have previously reported that murine S. aureus isolates from laboratory and wild mice and voles have adapted to their rodent host by eliminating human-specific phage-carried immune evasion (IEC) genes as well as MGE-carried SAg genes [11,34,45]. To test whether S. aureus isolates from rats show similar features of host adaptation, we compared the presence of IEC genes and the SAg gene repertoire with CC-matched human isolates.
Sa3int phages harbor the human-specific IEC encoding staphylokinase (SAK), staphylococcal complement inhibitor (SCIN), chemotaxis inhibitory protein of S. aureus (CHIPS) as well as staphylococcal enterotoxins A or P (SEA, SEP) [46,47]. IEC-encoding prophages are prevalent in human S. aureus isolates, but frequently absent in animal-adapted strains, including murine and vole S. aureus isolates [11,34,44]. In line with this, only 1/37 (2.7%) isolates from free-living wild rats, none of the S. aureus isolates from captive wild rats (0/41) and 39/66 (59.1%) of laboratory rats harbored the phage-carried IEC genes sak, chp and/or scn (Tables S6-S8). A CC-wise comparison of the most prevalent rat S. aureus lineages with matched human strains revealed that IEC genes were significantly less frequent or even absent in the rat-derived CC8, CC49, CC88, and CC398 isolates (Table 3).  SAg genes are prevalent in human S. aureus isolates (ca. 80%, not considering selX in the core genome) [44]. In contrast, only 1/37 (2.7%) isolates from free-living wild rats, none of the S. aureus isolates from captive wild rats (0/41) and 21/66 (31.8%) isolates of laboratory rats harbored SAg genes. In 12 of these 22 strains, these SAg genes were located on MGEs (Tables S6-S8). A CC-matched analysis revealed that MGE-carried SAg genes were significantly less frequent in rat CC8 and CC88 isolates than their human counterparts (Table 3). For instance, seven out of 10 human CC8 isolates harbored the plasmid-carried staphylococcal enterotoxin genes sed/sej/ser, the phage-carried sea/sep, and/or the SaPI-carried seb/sek/seq ( Table 3 and Table S9). In contrast, all ten rat CC8 isolates lacked SAg genes (Table S8). Similarly, 50% of human CC88 isolates were positive for SAg genes, while all nine rat CC88 isolates were negative (Tables S8 and S9). Taken together, S. aureus isolates from laboratory rats and wild rats rarely harbor phage-carried IEC genes and MGE-carried SAg genes, suggesting long-term adaptation to their hosts.

Both CC and Origin of S. aureus Determine Its Coagulation Behavior
The coagulation of the host's plasma is a typical virulence trait of host-adapted S. aureus strains [31]. None of the tested strains was able to completely coagulate plasma obtained from Sprague Dawley rats ( Figure 3). Nevertheless, the formation of small to large clots was observed, and after 24 h of incubation, human CC88 strains formed the biggest clots. Looking at the kinetics of coagulation, rat CC49 strains coagulated rat plasma faster than all other tested strains whose coagulation behavior was similar within the first 5 h. Moreover, the rat CC49 strains were the only ones able to reduce the coagulation score, i.e., disintegrate the already formed clots, over time.

Both CC and Origin of S. aureus Determine Its Coagulation Behavior
The coagulation of the host's plasma is a typical virulence trait of host-adapted S. aureus strains [31]. None of the tested strains was able to completely coagulate plasma obtained from Sprague Dawley rats ( Figure 3). Nevertheless, the formation of small to large clots was observed, and after 24 h of incubation, human CC88 strains formed the biggest clots. Looking at the kinetics of coagulation, rat CC49 strains coagulated rat plasma faster than all other tested strains whose coagulation behavior was similar within the first 5 h. Moreover, the rat CC49 strains were the only ones able to reduce the coagulation score, i.e., disintegrate the already formed clots, over time.

Discussion
While wild rats are a known reservoir for MRSA, the natural S. aureus population in wild and laboratory rats has not been described in detail [18,19,21]. Here, we determined the prevalence of S. aureus including MRSA in free-living and captive wild rats, as well as laboratory rats (Table 1). We observed striking differences between these three major cohorts. Free-living wild rats as well as those captured and held in husbandry were mainly colonized with the S. aureus lineages CC49 and CC130. In contrast, laboratory rats were colonized with various lineages, mostly of human origin (Figure 1). Rats with contact with livestock were frequently colonized with the livestock-associated lineage CC398; some of the isolates were even methicillin-resistant ( Figure 1, Table 2). S. aureus isolates from rats showed features of host adaptation on both the genetic as well as the phenotypic level (Table 3). These results underline the importance of rats as reservoir for MSSA as well as MRSA. Figure 3. The S. aureus lineage CC49 originating from rats shows a different coagulation behavior in rat plasma than human CC49, human CC88 and rat CC88 strains. (A) The graph shows the coagulation of rat plasma by both human and rat CC49 and CC88 isolates (N = 4-5/group) over a time course of 24 h. The medium control did not induce any coagulation (data not shown). The graph depicts representative data from one of four consecutive experiments with concordant results. (B) Differences in the coagulation potential between the groups (rat CC49 vs. human CC49; rat CC88 vs. human CC88; human CC49 vs. human CC88, rat CC49 vs. rat CC88) were statistically evaluated per time point by a Kruskal-Wallis test with Dunn's multiple comparisons test. The asterisks represent the adjusted p value: *, p value < 0.05; **, p value < 0.01.

Discussion
While wild rats are a known reservoir for MRSA, the natural S. aureus population in wild and laboratory rats has not been described in detail [18,19,21]. Here, we determined the prevalence of S. aureus including MRSA in free-living and captive wild rats, as well as laboratory rats (Table 1). We observed striking differences between these three major cohorts. Free-living wild rats as well as those captured and held in husbandry were mainly colonized with the S. aureus lineages CC49 and CC130. In contrast, laboratory rats were colonized with various lineages, mostly of human origin (Figure 1). Rats with contact with livestock were frequently colonized with the livestock-associated lineage CC398; some of the isolates were even methicillin-resistant ( Figure 1, Table 2). S. aureus isolates from rats showed features of host adaptation on both the genetic as well as the phenotypic level (Table 3). These results underline the importance of rats as reservoir for MSSA as well as MRSA.

Wild Rats Are Predominantly Colonized with The S. aureus Lineages CC49 and CC130
One-quarter (25.5%) of the examined wild rats were positive for S. aureus. These S. aureus-positive rats were detected in different areas of Germany and the Czech Republic over a time period of eight years, suggesting that S. aureus colonization of rats is widely distributed. Hence, rats can be considered as a natural host of S. aureus. Compared to other species, e.g., humans, the diversity of the detected lineages is quite low, although this might be affected by the small sample size in some cohorts [44]. Interestingly, the two dominating lineages, CC130 and CC49, could also be found in wild mice and voles, as we have reported previously [11]. The lineage CC130 has already been isolated from a Norway rat [10] and shows a broad host spectrum, especially in wildlife. It has been detected in red foxes, hedgehogs, brown hares, wild boar, rabbits and magpies all over Europe and Tunisia [10,[48][49][50]. In contrast to our study, most of the previously described CC130 isolates were MRSA, harboring the mecC locus.
CC49 strains show a less diverse host spectrum and have mostly been isolated from wild animals such as bank voles, squirrels and wild boar in Central Europe and Spain [10,11,[51][52][53]. Usually, these isolates show an MSSA phenotype, but MRSA (both mecA and mecC positive) were also found in pigs, horses, and humans [54][55][56]. This suggests that SCCmec elements are predominantly acquired and maintained in environments with significant selective pressure due to the use of antibiotics [57,58].

Rats with Contact with Livestock Frequently Carry CC398
Captive wild rats whose predecessors were captured on a pig farm in 2015 (strain Neufels) and kept under laboratory settings in two separate husbandries (BB and BE) were mostly (90.3%) colonized with CC398-MSSA (Tables S4 and S7). However, while the BB rats were colonized with CC398-MRSA (mecA-positive), the BE rats carried CC398-MSSA. CC398-MRSA have been repeatedly reported from rats trapped on farms and even within inner city centres [18,19,21]. Since two of these studies focused only on MRSA, epidemiological data on CC398-MSSA are rather scarce.
CC398-MRSA are highly prevalent in German livestock, but CC398-MSSA have also been reported [59,60]. Thus, it can be assumed that the CC398 strains were acquired from livestock and spread within the rat population, as this type of acquisition and spread has already been described for other lineages and host species [30,35,61]. There are several possible scenarios: First, CC398-MRSA could have been transmitted from livestock to rats and subsequently lost the SCCmec element in BE husbandry. Second, CC398-MSSA could have been transferred from pigs to rats and the BB subpopulation subsequently acquired the SCCmec element. Third, both CC398-MSSA and -MRSA could have been transmitted to rats with the former being maintained in the BE and the latter in the BB rat population. The detection of a LA-S. aureus lineage in captive wild rat populations years to decades after acquisition of the ancestral animals suggests a stable colonization of rats and an effective transmission to their offspring, similar to what has been observed in laboratory mice [45]. Since CC398 belongs to the EHSG, it is still unclear whether this effective acquisition of LA-S. aureus strains holds true for all S. aureus lineages or just to certain ones. Nevertheless, we can conclude that wild rats can act as reservoir for LA-S. aureus, including LA-MRSA, making the containment of those pathogens even more challenging than previously assumed.

Laboratory Rats Carry Various Lineages, Mostly of Human Origin
Laboratory rats carried a high diversity of S. aureus lineages-many of which are commonly found in humans, e.g., CC1, CC5, CC8, CC15 and CC101 [44]. Considering the close contact of laboratory rats with humans over many decades, it is only reasonable to assume that these strains have originally been transmitted from humans to rats. This might be supported by the fact that there is a significant overlap with the S. aureus population in laboratory mice [34], implying that humans could be the prime source of new S. aureus strains in laboratory animals. Transmission of S. aureus from humans to animals has been described several times [62]. However, it seems likely that not all human-adapted lineages are able to colonize rats. For instance, the most common human lineages (CC30 and CC45) are almost completely absent in the rat population and have also only rarely been detected in laboratory mice [34,44].
On the other hand, colonizing lineages such as CC88 and CC188 have already been described in other rodents and are rarely found in humans, suggesting that they might represent rodent-specific lineages [11,34]. Their origin, however, remains unclear. They might have originated from the human population and introduced into the rat colony by animal caretakers, with subsequent adaptation to their new rodent host over time. Alternatively, they might have already been present in the ancestry rat population at the time when commercialized breeding was established. Nevertheless, the low prevalence of phage-carried IEC genes and MGE-carried SAg genes in all rat-derived S. aureus lineages suggests a long-term colonization of these hosts.

Captive Wild Rats Maintain Their S. aureus Population
Three of the analyzed rat husbandries (located in the German federal states BB, BE, and NRW) harbored wild rats that have been maintained in captivity for several years. Two of these husbandries were S. aureus-positive (19.4% in BB, 77.1% in BE), while in the NRW cohort, S. aureus was completely absent. Within the same husbandry in BB both S. aureus-positive (e.g., Neufels, Tilbury) and -negative (e.g., BSR, WPHR) rat strains were detected (Table S4). This suggests that a transmission between different rat strains was effectively prevented by the implemented hygiene measures. This finding is supported by similar observations in laboratory mice [45].
In addition, the captive wild rats are of interest, because they allow some conclusions about the stability of the S. aureus population after adaptation to a new environment with no further contact with wildlife, but increased contact with humans. The S. aureus-positive cohorts of captive wild rats (located in BB, and BE) were still colonized with S. aureus lineages common in wild rats (CC49, CC130) and livestock (CC398), while typical human S. aureus lineages, which were common in laboratory rats (see above), were not detected, even though they included animals crossed with laboratory rats (Table S5). This may be explained by a high stability of the S. aureus population in the rat cohort [63]. The absence of typical human S. aureus lineages suggests that several factors, including the existing microbiome [64][65][66], the limited contact of caretakers to animals, as well as high hygienic standards (including wearing coats and gloves) successfully prevent the introduction of human-adapted S. aureus strains. Moreover, the rat-adapted S. aureus population may be able to outcompete introduced human-adapted S. aureus strains. Indeed, we have previously observed that the S. aureus population in laboratory mice is stable over years, and that these bacteria are very effectively transmitted to their offspring [45]. Whether this holds true for rats as well needs to be further investigated.

S. aureus Isolates from Rats Likely Adapt to Their Host by Eliminating MGEs Carrying Human-Specific Virulence Factors
A comparison of rat and CC-matched human S. aureus isolates showed that the rat strains are probably adapted to their host on both genetic and phenotypic levels. The observed scarcity of the phage-carried IEC genes, especially among S. aureus isolates from wild rats, indicates a strong selection against these genetic elements. Several phage-encoded immune evasion factors show no or little activity on rat cells [67][68][69][70]. For example, both CHIPS and SCIN efficiently block complement activation in humans but show no activity in mice and rats [67,70]. In contrast, SAK shows thrombolytic activity also in rats [71]. In addition, the Sa3int phage integrates into and thereby inactivates a known staphylococcal virulence gene, the sphingomyelinase hemolysin beta (hlb) gene. Hlb lyses erythrocytes, lymphocytes and keratinocytes and contributes to skin colonization in mouse models [72]. Thus, eliminating the Sa3int phage and thereby restoring the hlb gene could be advantageous for colonizing rats. In line with this, we have previously observed that this prophage is rare in S. aureus populations from laboratory and wild mice [11,34,45]. Indeed, the lack of the phage-carried IEC genes is a well-known marker for S. aureus isolates of animal origin [46,73,74].
Another genetic correlate of host adaptation was the absence of MGE-carried SAg genes in the lineages CC8 and CC88. This confirms previous observations in wild and laboratory mice, where S. aureus isolates also lacked MGE-carried SAg genes. S. aureus SAgs activate human T cells in the picomolar concentration range in vitro but show a 100-1000-fold reduced activity on rat T cells and a 10-100-fold reduced activity on murine T cells [69,75,76]. Elimination of SAg genes in S. aureus isolates from rats, whenever their genetic location permits this, underlines their expendability in rats.

Rat-derived CC49 Isolates Show Enhanced Procoagulatory Activity on Rat Plasma
Blood coagulation is a major virulence trait of S. aureus, and some pro-or anticoagulatory factors are known to be host specific [31]. Therefore, we compared the pro-coagulatory activity of CC49 and CC88 isolates from rats with their matched human counterparts. Rat CC49 isolates coagulated rat plasma faster than matched human CC49 isolates. In contrast, we observed no difference in our test system for the CC88 isolates. In line with our data, Viana et al. reported that ruminant-derived strains, but not human isolates, had the capacity to stimulate clotting of ruminant plasma due to a unique MGE-encoded paralogue of the von Willebrand factor binding protein [33]. The molecular basis for the observed differences in our rat-derived S. aureus isolates will be clarified by whole-genome sequencing in future studies.

Rats Carrying Human-Derived or LA-MRSA Present a Human Health Risk
Considering the growing public health relevance of LA-MRSA, i.e., CC398-MRSA, in human health and livestock farming [13,77,78], there is an urgent need to elucidate potential transmission routes. S. aureus/MRSA is not only transmitted via person-to-person contact, but can also spread among and between domestic animals, livestock and people via direct contact or inhalation of contaminated air and dust in stables [79,80]. Moreover, pest species, including rats, are a potential source of MRSA [18,19,21]. Taking the "One Health" concept into consideration, measures to control the spread of S. aureus, and in particular MRSA, must take these different reservoirs and transmission routes into account.
Rats actively explore a number of different habitats, and therefore have a unique opportunity to encounter a variety of human and livestock pathogens [16,[81][82][83]. Indeed, there is growing evidence that rats play a role in the spread and persistence of MRSA on pig farms [18,21]. Their ability to cover considerable distances allows rats to spread pathogens from one herd to another [84,85]. Besides farm animals, rats might also transfer livestock-derived S. aureus from farms to domestic animals or pets, and ultimately to humans through contact with these animals or their excrements. In addition, they might also serve as a 'mixing vessel', facilitating the transfer of genetic elements encoding resistance genes either between different S. aureus strains or between S. aureus and other bacterial species [19]. Thus, strategies to control MRSA on farms should take pest control into account.
Rats are not only a reservoir for LA-MRSA but also HA-MRSA. In our study, we observed a single CC30-MRSA-positive rat caught within a city. Similarly, Himsworth et al. reported an MRSA-colonization rate of 3.5% among tested urban rats, mostly involving human MRSA common in that area, such as USA300, but also CC398-MRSA [19]. While the risk of rat-to-human MRSA transmission outside the farm setting is very low, it could become relevant in impoverished, inner-city neighborhoods, where socio-economic factors might promote rat infestations and hence rat-to-human contact [19,81,86]. Moreover, the occupational risk of encountering MRSA-colonized rats is quite considerable in case of sewage workers or workers on landfill sites.
In addition to the known infection risk due to colonized free-living wild rats, we report here that captive wild rats may also be a source of MRSA. To avoid transmission and subsequent infection of caretakers, strict hygiene measures, including the use of gloves and face masks, as well as proper disinfection of cages should be implemented.
The observed elimination of human-specific virulence genes, i.e., IEC-carried immune evasion genes as well as SAg genes, in rat-derived S. aureus isolates raises the question, whether they are capable of colonizing and infecting humans. We propose that the potential to colonize and/or infect humans depends on a lineage-specific repertoire of host-specific or broadly-reactive adhesins, immune evasion as well as virulence factors. On the one hand, human colonization and infection with CC49 has only rarely been reported [44,56,87], suggesting a reduced fitness and virulence of these strains in humans. In line with this, caretakers in animal breeding facilities were not colonized with the endogenous rat-or mouse-adapted S. aureus strains (personal communication, S. Holtfreter) [45].
On the other hand, zoonotic CC398-MRSA are commonly encountered among farm workers or other individuals with professional contact with livestock, thus promoting the spread of CC398-MRSA in the general population and introducing this lineage in hospitals and other healthcare facilities. A recent study demonstrated that these strains are able to cause all types of infections attributed to S. aureus; nevertheless, the disease-burden of LA-MRSA seems to be lower compared to other MRSA lineages [78]. The factors driving the spread of this zoonotic lineage in the human population are not understood yet. Interestingly, the livestock-associated CC398 originated from a human CC398-MSSA. Upon host switch, it eliminated the IEC-carrying Sa3int phages and acquired methicillin resistance [88]. Conversely, one could assume that a re-acquisition of these phages promotes the spread of zoonotic CC398-MRSA in the human population. However, a recent study demonstrated that the acquisition of IEC-encoding Sa3int phages is no major driver for the re-adaptation of CC398-MRSA to the human host [89].

Conclusions
S. aureus, particularly MRSA, has a major impact not only on human health, but also on farm and companion animals as well as wildlife. Rats are key players in the ecology of MRSA on farms and may also transmit human MRSA. Overall, our data show that rats harbor a natural S. aureus population, consisting mainly of CC49-and CC130-MSSA. However, contact with humans or livestock poses a colonization pressure on rats, reflected in the uptake and long-term carriage of S. aureus isolates of human or livestock origin. Laboratory rats represent a special case, because they lack the lineages CC49 and CC130 and instead carry lineages that are likely human-derived. Upon transfer to rats, S. aureus strains seem to adapt to their new host by removing futile genetic elements carrying human-specific immune evasion genes or SAg genes. Future studies should further investigate S. aureus ecology in different settings, e.g., farms versus inner-city environments, as well as among different rat species, i.e. Rattus rattus versus Rattus norvegicus, and should consider sampling all possible hosts including wild animals (particularly rats), livestock, pets and humans as well as the abiotic environment (e.g., dust samples), to gain a better understanding of transmission and maintenance dynamics. In light of the "One Health" concept, such investigational studies could provide the conceptual foundation for developing effective measures to contain these dangerous pathogens in the future.
The category "free-living wild rats" comprises rats that were collected in the wild between 2009 and 2017 using live, electric as well as snap traps, in addition to bait poisons, at different locations in Mecklenburg-Western Pomerania (N = 18), North Rhine-Westphalia (N = 81), and Baden-Württemberg (N = 17) [90] as well as in the Moravian-Silesian Region in the Czech Republic (N = 29) (Figure 1). Details on the geographical origin, habitat, and capture method of these free-living wild rats as well as on their characteristics and nasal S. aureus colonization are described in Tables S1 and S2, respectively.
The category "captive wild rats" comprises offspring of wild rats that were bred in a laboratory setting or a large enclosure for several years or even decades. Details on the geographical origin, source, original habitat, housing conditions and hygiene measures in these animal facilities, as well as on the rat characteristics and nasal S. aureus colonization are described in Tables S3 and S4, respectively. The category "laboratory rats" comprises domesticated rats raised for animal experiments or as feeder animals (Tables S3 and S5).
Samples were collected using approved methods, according to relevant legislation and by permission of the responsible State authorities whenever necessary (Supplementary Tables S1 and S3). Wild rats were captured by professional pest control operators, and hence no capture permit was required. The only exception were wild rats from NRW (NRW_2; Supplementary Table S1), which were captured using live traps following approval by the North Rhine-Westphalia State Agency for Nature, Environment and Dead animals were immediately frozen and stored at −20 • C until dissection. Their noses were aseptically removed from the body and frozen again at −20 • C.
S. aureus isolates from laboratory rats: We also included S. aureus isolates in this study, which were obtained from laboratory rats by collaborators. The CR strains (N = 38) were provided by Charles River's Research Animal Diagnostic Services (Wilmington, USA). Strains were submitted by different Charles River facilities, as well as from Charles River customers (pharmaceutical companies, universities, or research institutes). 13 S. aureus isolates were obtained during routine health monitoring or from euthanized rats suffering from S. aureus infections (e.g., abscesses) at the animal facilities of the German Cancer Research Center (DKFZ, Deutsches Krebsforschungszentrum, Heidelberg, BW, Germany) in Heidelberg, Germany between 1983 and 2014. A single S. aureus isolate was obtained from the caecum of R. norvegicus during routine health monitoring at the animal facilities of the University of Ulm (Tierforschungszentrum, Ulm, BW, Germany).
S. aureus isolates from humans: Fifty CC-matched human S. aureus strains were derived from several S. aureus colonization studies (T, SH, SHIP) [44,91,92]. The T and SH strains were obtained from healthy blood donors in Northern Germany in 2002 and 2005-2006, respectively. Spa types as well as agr type, SAg gene patterns and S. aureus integrase (Saint) phage groups of these strains were previously reported [91,92]. The SHIP studies (SHIP-2, SHIP-Trend-0) are population-based studies in Western Pomerania and are described in depth elsewhere [44,93]. Human CC49, CC88, and CC130 isolates were derived from strain collections of the Robert Koch-Institute and the Institute for Medical Microbiology and Hygiene, Technical University of Dresden.

Sample Preparation and Screening for S. aureus
The rat noses were thawed at room temperature for 1 h and then transferred into tubes containing 1 mL BBL™ Phenol Red Mannitol Broth (MSB: 10 g/L pancreatic digest of casein; 75 g/L NaCl; 5 g/L D-mannitol; 18 mg/L phenol red in double-distilled water). The noses were homogenized using 2 rounds of spinning (6000/min, 2 × 20 s; 15 s breaks) in the presence of 1.4 and 2.8 mm Zirconium oxide beads (Bertin Precellys 24 homogenizer, VWR, Darmstadt, Germany). Aliquots of 100 µL of the homogenate were plated onto mannitol salt agar (MSA) plates (Becton, Dickinson and Co., Franklin Lakes, NJ, USA) and incubated for 48 h at 37 • C. The rest of the homogenate was subsequently transferred to a 50 mL tube and cultured aerobically in MSB (total volume 5 mL) for 48 h at 37 • C under agitation for bacterial enrichment.
Three random samples from MSA plates containing golden-yellow colonies were used for a mixed-colony PCR to detect S. aureus presence by testing for gyrase (gyr) and nuclease (nuc) genes. In brief, bacterial cells were resuspended in reaction mix (25 µL) containing 1× GoTaq ® Flexi buffer, 100 µM deoxynucleoside triphosphates (dATP, dCTP, dGTP, and dTTP; buffer and nucleotides from Thermo Fisher Scientific Baltics UAB, Vilnius, Lithuania), 5 mM MgCl 2 , 320 nmol/L of each primer, 1.0 U GoTaq ® G2 Flexi DNA polymerase (Promega, Mannheim, Germany) in DNase-and RNase-free water. PCRs were performed as follows: an initial denaturation step at 95 • C for 10 min was followed by either 30  A positive result indicating the presence of S. aureus was followed by subculturing 3 distinct colony morphotypes on sheep blood agar (Becton, Dickinson and Co, Franklin Lakes, NJ, USA; 37 • C; 24 h). The colonies were further tested using the S. aureus-specific latex agglutination test (Prolex TM Staph Xtra Latex Kit, Pro-Lab Diagnostics, Richmond Hill, ON, Canada). Finally, S. aureus isolates were subjected to colony PCR (gyr, nuc). DNA isolation was performed on all PCR-positive S. aureus isolates using the DNeasy Blood and Tissue Kit (Qiagen, Venlo, the Netherlands) according to the manufacturer's instructions, but with an addition of 0.1 mg/ mL lysostaphin (Sigma-Aldrich, St. Louis, MI, USA) to the lysis buffer [11]. S. aureus isolates were stored as glycerol stocks.

Spa Genotyping and Multilocus Sequence Typing (MLST)
Based on the allelic profile of seven house-keeping genes, the species S. aureus can be divided into STs, which again are grouped into CCs [95]. Spa typing and MLST were performed as described elsewhere [96,97]. Spa typing was performed on all S. aureus isolates, and isolates were assigned to CCs based on spa typing results [34]. MLST was performed on selected isolates to confirm the spa clustering results. Profiles were submitted to the SpaServer and the MLST database (https://pubmlst.org/saureus/).

MIC of Penicillin against S. aureus Strains Using the Broth Microdilution Method
The MIC of penicillin G (Merck KGaA, Darmstadt, Germany) was determined in a total volume of 0.1 mL of Mueller-Hinton broth (MHB: casein acid hydrolysate 17.5 g/L; Beef extract 3 g/L; Starch 1.5 g/L; Merck KGaA, Darmstadt, Germany) containing approximately 5 × 10 5 colony-forming units (CFU)/mL in microdilution trays with round-bottom wells (CELLSTAR ® , greiner bio-one) according to the CLSI broth microdilution method [99]. The following control strains were included: S. aureus ATCC ® 43300 (MRSA) and S. aureus ATCC ® 29213 (bla + ) as positive controls; S. aureus ATCC ® 25923 (bla -) as a negative control. In addition, medium sterility as well as growth controls were included in the assay. The penicillin range tested was 0.001-16 µg/mL, whereas the resistance breakpoint was set to ≥0.25 µg/mL according to CLSI guidelines [99]. The inoculated microdilution trays were sealed and incubated at 37 • C for 24-48 h in an ambient air incubator. The MIC values were determined as the lowest penicillin concentration that completely inhibits growth of the organism in the microdilution wells as detected by the unaided eye.

Coagulation Assay
Rat CC88 isolates (N = 5) and human CC88 isolates (N = 5), as well as rat CC49 isolates (N = 5) and human CC49 isolates (N = 4) were grown for 7 h at 37 • C and 200 rpm in tryptic soy broth (TSB) to reach the early stationary phase. The OD of the TSB cultures was adjusted to 4. Seventy microliters of the adjusted bacterial cultures (approx. 2.0 × 10 9 CFU/mL) were mixed in 10 mL glass tubes with 500 µL of rat heparinized plasma from Sprague Dawley rats (Equitech-Bio, Kerrville, TX, USA) and incubated at 37 • C without agitation. The coagulation state was visually examined in a blinded fashion after 0.5 h, 1 h, 2 h, 3 h, 4 h, 5 h, 19 h and 24 h using a modified coagulation score [103]. No coagulation was reported as 0, cloudiness of plasma as 1, small coagulation flakes as 2, a medium sized clot as 3, a large clot as 4 and complete coagulation (tube can be inverted) as 5 [45].

Statistical Analyses
Categorical data were assessed by using Pearson's Chi-squared test. p values of ≤0.05 were considered statistically significant.
Differences in the coagulation potential between the groups (rat CC49 vs. human CC49; rat CC88 vs. human CC88; human CC49 vs. human CC88, rat CC49 vs. rat CC88) were statistically evaluated per time point by a Kruskal-Wallis test with Dunn's multiple comparisons test.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6651/12/2/80/s1. Table S1: Description of the geographical origin, habitat and capture method of free-living wild rats (N = 145). Table S2: Overview of the wild rats included in this study, their origin, characteristics and nasal S. aureus colonization. Table S3: Description of the geographical origin, original habitat, housing conditions and hygiene measures in husbandries of captive wild rats (N = 188) as well as laboratory rats (N = 114) from Germany. Table  S4: Overview of the captive wild rats included in this study, their origin, characteristics and nasal S. aureus colonization. Table S5: Overview of the laboratory rats included in this study, their origin, characteristics and nasal S. aureus colonization. Table S6: Genotype, virulence genes, phage patterns, and penicillin resistance of S. aureus isolates from wild rats. Table S7: Genotype, virulence genes, phage patterns, and penicillin resistance of S. aureus isolates from captive wild rats. Table S8: Genotype, virulence genes, phage patterns, and penicillin resistance of S. aureus isolates from laboratory rats. Table S9: Genotype, virulence genes, and phage patterns of selected CC-matched human S. aureus isolates.