Staphylococcus aureus From Goats Are Genetically Heterogeneous and Distinct to Bovine Ones

Staphylococcus aureus is one of the major pathogens responsible for intramammary infections in small ruminants, causing severe economic losses in dairy farms. In addition, S. aureus can contaminate milk and dairy products and produce staphylococcal enterotoxins, being responsible for staphylococcal food poisoning. Currently, data on the population structure and the virulence gene patterns of S. aureus strains isolated from goat milk is limited. Therefore, this study aimed at defining Ribosomal Spacer PCR (RS-PCR) genotypes, clonal complexes (CC), spa types, and virulence gene profiles of S. aureus isolated from goat milk samples from Lombardy region of Italy. A total of 295 S. aureus isolates from 65 goat bulk tank milk samples were genotyped by RS-PCR. spa typing and virulence gene patterns of a subgroup of 88 isolates were determined, and MLST was performed on a further subgroup of 39 isolates, representing all the spa types identified during the analysis. This study revealed 7 major genotypic clusters (CLR, CLAA, CLZ, CLAW, CLBW, CLS, and CLI), of which S. aureus CLR (19.8%) was the most common. A total of 26 different spa types were detected, the most prevalent types were t1773 (24%), t5428 (22.7%), and t2678 (12.5%). Overall, 44.3% of all isolates harbored at least one enterotoxin gene. The most prevalent was the combination of sec-sel genes (35.2%). Based on their MLST, isolates were assigned to 14 different CC, with majority grouped as CC133 (24%), CC130 (19.6%), and CC522 (19.6%). The caprine S. aureus population was depicted with a minimum spanning tree and an evolutionary analysis based on spa typing and MLST, respectively. Then, the variability of such strains was compared to that of bovine strains isolated in the same space-time span. Our results confirmed that S. aureus isolates from goats have wide genetic variability and differ from the bovine strains, supporting the idea that S. aureus from small ruminants may constitute a distinct population.


INTRODUCTION
Staphylococcus aureus is one of the major pathogens responsible for clinical and subclinical mastitis in goats, causing significant economic losses due to the decrease of quality and quantity of milk and to the treatment losses for staphylococcal mastitis. Furthermore, S. aureus can contaminate both raw milk and dairy products, being a source of food poisoning (1,2).
The prevalence of S. aureus in dairy goat herds, estimated by analyzing bulk tank milk, varies between 16.7 and 96.2% in different countries (3)(4)(5). In Italy, the prevalence of S. aureus in dairy goat herds was reported to be 43.1% in the Lombardy Region (6) and 76.9% in Sardinia (7).
Ribosomal Spacer PCR (RS-PCR) is an accurate, rapid, inexpensive genotyping method with high discriminatory power (8). It has been used to classify S. aureus isolates into genotypes, that are inferred from the electrophoresis profile using the Mahal software, according to Fournier et al. (9) and Graber (10). In particular, an electrophoretic profile differing in more than one band from all identified genotypes is considered as a new genotype. Genotypes are named and extended leading to the genotypes GTA to GTZ, followed by the genotypes GTAA to GTAZ, GTBA to GTBZ, and so on. Genotypic variants, which differ in only one band from the genotype after the electrophoretic analysis, were indicated with roman superscripts (e.g., GTR I ) (10). RS-PCR genotypes, in turn, can be grouped into clusters (CL), each of which consists of a genotype and all its variants (e.g., GTR and GTR I , GTR II , GTR III , etc.).
Several studies performed RS-PCR to genotype S. aureus strains isolated from bovine milk, and showed a great genotypic variety, differing in their contagiousness (8,9,11,12). S. aureus CLB was found to be highly contagious and has been associated with high within-herd prevalence of intramammary infections (IMI), together with CLR and CLS (8,9,12). In contrast, S. aureus CLC appears to be less problematic because it generally affects individual cows and only one quarter of the mammary gland is affected (11,13). Moreover, each genotype presented a different pattern of virulence genes (9,14,15). In spite of this detailed knowledge on dairy-associated S. aureus genotypes, no data on S. aureus genotypes isolated from goats is currently available.
S. aureus produces a wide range of virulence factors, which are essential for a successful infection. At least 25 different toxins (e.g., enterotoxins SEA to SEQ, toxic shock syndrome toxin-1 tst1, exfoliative toxins ETA and ETB), 15 microbial surface components recognizing adhesive matrix molecules (such as clumping factor A clfA), 20 immune evasion molecules (such as protein A, coagulase, hemolysins, and leukocidins), among others are known S. aureus virulence factors (19). The most prevalent enterotoxin genes detected in S. aureus isolated from small ruminants were sec and sel, whereas the sea gene was found exclusively in caprine isolates (15,17). Merz et al. (17) reported that all the isolates harboring the tst1 gene also harbored the combination of genes sec-sel and were assigned exclusively to CC130 and CC133. Significantly higher prevalence rates of sec, sel, and tst1 genes were observed in S. aureus from small ruminants than among bovine isolates (15,17).
While the population structure and the genomic characteristics of S. aureus isolated from bovine milk are very well-described, similar data on caprine S. aureus is very limited. Therefore, this study aimed at defining RS-PCR genotypes, clonal complexes (CC), spa types, and virulence gene profiles of S. aureus isolated from goat milk in the Lombardy region in northern Italy.

Bacterial Strains and DNA Extraction
S. aureus isolated from goat bulk tank milk (BTM) collected by Cortimiglia et al. (6) was used for this study, thus the sampling criteria and bacteriological analysis have already been described. Briefly, BTM samples were collected from 197 different dairy goat farms located in Lombardy (northern Italy) between July and October 2012. After collection, milk samples were transported at +4 • C to the laboratory and stored at −20 • C till analysis, which were conducted within 3 months. Bacteriological analysis was performed with standard techniques for S. aureus isolation, as previously described by Cortimiglia et al. (6). In particular, milk samples were spread on blood agar supplemented with 5% defibrinated sheep blood and on Baird Parker agar supplemented with rabbit plasma fibrinogen (BP-RPF; Oxoid Ltd., Basingstoke, UK), plates were incubated at 37 • C and analyzed after 24 and 48 h. Colonies developing a typical coagulase halo on BP-RPF agar were considered suspected S. aureus, as were hemolytic colonies on BA testing positive in a tube coagulase test (Coagulase plasma-EDTA, Biolife srl, Milan, Italy).
DNA was extracted using the DNA isolation system kit (Clonit, Medical System, Genova, Italy) according to the manufacturer's guidelines. DNA quality and quantity were measured using a NanoDrop ND-1000 spectrophotometer (Nano-Drop Technologies, Wilmington, DE), and stored at −20 • C till further use.

RS-PCR
If present, five S. aureus isolates from each BTM positive sample were genotyped by RS-PCR, according to Fournier et al. (9) and Graber (10). When different cultural morphologies were detected (pigmentation and hemolysis on blood agar or coagulase and lecitinase halos on Baird Parker agar), four colonies per morphology were selected in each sample.
The PCR products were analyzed using the miniaturized electrophoresis system DNA 7500 LabChip [Agilent Technologies, Santa Clara, CA; (9,10)]. Genotypes were inferred from the electrophoresis profile using the Mahal software, which is freely available online (10) 1 .

Spa Typing
Typing of spa was performed on a subset of strains, representing all the S. aureus RS-PCR genotypes found in each milk sample (i.e., herd). In detail, one or more isolates were selected from each BTM sample, depending on the number of different genotypes identified with the RS-PCR. Typing was performed according to Shopsin et al. (20), and spa types were assigned using the code system described on Ridom SpaServer 2 .
To depict the frequency and genetic relatedness among the isolates, the spa typing results were analyzed, and a minimum spanning tree (MST) was constructed using the BioNumerics 7.6 software (Applied Maths, Sint-Martens-Latem, Belgium).

Virulence Factors
PCR amplifications were performed on the same strains selected for spa typing to investigate the presence of 22 virulence factors contributing to S. aureus pathogenicity: enterotoxins (from sea to see, and from seg to sel) (21,22), toxic shock syndrome (tst) and exfoliative (eta, etb) toxins (21), PVL (lukS-lukF) (23), leucocidin M (lukM), and leukotoxin ED (lukE-lukD) (24), cell-wall associated protein clfA (clumping factor A) (21), cna (collagen-binding protein) (25), and fmtB (cell-wall protein), scn, chp, and sak belonging to the immune evasion cluster (IEC) (26). The primers and protocols are listed in the Supplementary Table. MLST Multi Locus Sequence Typing (MLST) was performed as previously described (27) on one representative strain for each combination of spa typing and RS-PCR genotyping result. Allele numbers and sequence type (ST) were assigned as per the MLST database 3 Clonal analysis of the STs was performed using the clustering algorithm e-BURST 4 An evolutionary analysis was carried out using goeBURST 5 (Algorithm 3.0), a java implementation of the eBURST algorithm rules proposed by Feil et al. (28), using a graphic matroid approach that ensures an optimal solution for the placement of links between STs.

Statistical Analysis
The degree of similarity between the distributions of RS-PCR genotypes among caprine strains and bovine S. aureus strains in the same geographical area and during the same time period (12) was estimated using the Czekanowski index or Proportional Similarity Index (PSI). It is calculated by in which p i and q i represent the proportion of genotypes in the two distinct populations (caprine and bovine S. aureus strains), while i represents the total numerosity of the two populations. The values for PSI range from 1 for identical frequency distributions of the variable of interest to 0 for no similarities between the data sets (29).
Frontiers in Veterinary Science | www.frontiersin.org  Using the observed spa types, a MST was constructed. MST revealed 3 major spa types (t1773, t2678, and t5428) that were linked to a few secondary and tertiary spa types. In between there was a set of rarely observed spa types, some of them were involved in connecting the 3 main types (Figure 1).

MLST
Thirty-nine isolates were selected to perform a MLST analysis. The isolates were assigned to 23 different ST, which belonged to 14 different CC. One ST was a singleton ( Table 2). The most frequent CC were CC133 (n = 10; 25.6%), CC130 (n = 9; 23.1%), and CC522 (n = 7; 17.9%). CC398 was detected in 5.1% of isolates, whereas each of the remaining CC was assigned to only one isolate.
The evolutionary relationships between the different goat S. aureus strains are displayed in Figure 2. Three central ST were observed (ST133, ST522, and ST700), corresponding to CC133, CC522, and CC130, respectively. These clones are the founders of additional single locus variants (Figure 2), reflecting a further evolution away from the central clones. All the other ST/clones remain unrelated.

Statistical Analysis
PSI was estimated by comparing the RS-PCR genotypes of the strains (18 different S. aureus CL), isolated from the 65 caprine BTM samples in our study, with the 471 strains (44 different CL) isolated from 398 bovine BTM samples in a recent study (12). The index was 0.311, indicating a very low correlation between the distribution of caprine and bovine S. aureus genotypes.

DISCUSSION
Several studies used RS-PCR to genotype S. aureus strains isolated from bovine milk, yet no information on S. aureus from goat milk is currently available. For the first time, we genotyped via RS-PCR 295 S. aureus isolates, which grouped in 18 genotypic clusters. The major S. aureus CL have also been detected in bovine milk, except for CLAW. Other minor CL were here identified and reported for the first time, such as CLW, CLAL, CLBS, and CLBX. Previous studies demonstrated that S. aureus CLB, which is the most prevalent among bovine isolates, is highly contagious and associated with high withinherd prevalence of IMI (8,11,12). On the contrary, we isolated S. aureus CLB in a single BTM from a dairy goat farm located in the province of Sondrio. Furthermore, S. aureus CLR and CLS have also been reported to be associated with high withinherd prevalence of IMI (8,12). S. aureus CLR (27.7%) was the most frequently isolated among the analyzed herds. Albeit the lack of details about the sampled farms, we cannot exclude the possibility that some of these farms reared goats and cattle in close proximity, facilitating the spread between the two animal populations.
Most of the isolates were assigned to CC133, CC130, and CC522, so that these lineages may be the major CC among caprine S. aureus isolates in Italy as well. Our results are in agreement with previous studies reporting the predominance of these CC in S. aureus isolated from milk of small ruminants (4,(15)(16)(17).
Several studies consider S. aureus from goats and sheep to be part of the same population based on CC, since they show pronounced adaptation to small ruminants, but not especially to sheep or goats (17,30). Nevertheless, different S. aureus spa types are reported to be associated with the two species. For example, among CC133 S. aureus, spa types t544, t3583, and t7304 were predominantly associated with caprine mastitis, whereas t2678 and t9088 with ovine mastitis (16,18,30). However, these spa types are not exclusively associated with one host species (30). Actually, CC133/t2678 was one of the most prevalent spa types among our strains, whereas we did not find any CC133/t544 or CC133/t3585. Previous studies have described CC133/t544 isolates in sheep as well (16,18). These contrasting results support the idea that S. aureus lineages found among sheep and goats are highly similar (17,31) and this might be due to the fact that in many countries sheep and goats are farmed together, as previously suggested by (31).
Bar-Gal et al. (15) demonstrated that S. aureus isolates from small ruminants differ from the bovine strains in their virulence gene patterns. In fact, cow isolates showed a higher rate of the tested virulence genes than goat and sheep isolates. In our study, 44.3% of all isolates carried at least one enterotoxin gene. All isolates harboring tst also carried the genes sec and sel, and were assigned exclusively to CC130 and CC133, in agreement with previous findings (17,32). These genes are located on the S. aureus pathogenicity island SaPIov1 (33). Genes coding for toxins usually associated with human diseases, such as eta and etb (responsible for the staphylococcal scalded-skin syndrome), or pvl (leukocidin produced by S. aureus that causes leukocyte destruction and tissue necrosis) were rarely or never detected. Moreover, we found high prevalence of leukocidins or genes related to host adhesion and invasion, suggesting their important role as virulence factors for S. aureus strains in small ruminants.
The comparison between S. aureus strains isolated from goat and bovine milk showed that S. aureus from goat milk have wide genetic variability and differ from the bovine strains.
Caprine S. aureus is characterized by 3 major spa types and a set of rarely observed types, representing a limited genetic variation among caprine strains. Except for t002 and t005, all the observed spa are observed frequently, yet are rarely present in the spa type database [Ridom SpaServer (see text footnote 2)] implying some uniqueness of the strains analyzed in the present study. Indeed, except for t002 and t005, all remaining spa types do not match with those typically observed for human and bovine strains (34,35). We suspect that t002 and t005 are likely a result of human milk contamination that can occasionally occur during milking and/or equipment cleaning operations. The most prevalent human spa types include t032, t003, t002, t008, t011, and t127 (see text footnote 2), whereas those most often observed in bovine milk include t529, t2953, t1403, t524, t024, and t164 (35).

CONCLUSION
Our study supports the idea that S. aureus from small ruminants constitutes a distinct population in northern Italy, in agreement with previous findings reported worldwide. It also provides for the first time insights to the S. aureus RS-PCR genotypes circulating in goat herds in this area, allowing a comparison to bovine strains and adding useful data for epidemiological studies. Further information about S. aureus strains isolated from small ruminants is required, as they are a major problem for animal health and may also represent a serious concern for human health, as many dairy products are made from raw milk.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

INTERPRETATIVE SUMMARY
Staphylococcus aureus is the major pathogen responsible for intramammary infections in goats, causing severe economic losses in dairy farms. Furthermore, it can contaminate milk and dairy products, being a source of food poisoning. Since data on the genetic traits and population structure of S. aureus from goats are still very limited, our work aimed at determining genetic lineages and virulence traits of S. aureus isolated from goat milk in northern Italy. Previous studies used the rapid screening method RS-PCR to genotype S. aureus strains from bovine milk, showing a great variety of genotypes differing in their contagiousness. For the first time, we performed RS-PCR on S. aureus isolated from goat milk, in addition to morerecognized genotyping methods, such as spa typing, MLST, and we investigated their virulence profiles. S. aureus isolates showed a wide genetic variability and they differed from bovine strains. Our results support the idea that S. aureus from goats may constitute a distinct population, in agreement with previous findings reported worldwide. Our study provides for the first time an overview of the S. aureus RS-PCR genotypes circulating in goat herds in northern Italy, adding useful data for further epidemiological studies.

AUTHOR CONTRIBUTIONS
ML conceived and planned the experiments. AR, VB, CC, PC, and AG carried out the experiments. AM and FV contributed to the experiments and the interpretation of the results. HG and AR performed the statistical and population study analyses. AG and AR wrote the manuscript in consultation with ML and PC. All authors discussed the results, critically revised, and approved the final manuscript.