A total of 1,199 primate samples from the WaNPRC facility was tested for the presence of MRSA between May and August 2015. The animals [M. fascicularis, Macaca mulatta, and M. nemestrina] were in-house animals, as well as, out-of-state macaques shipped to the facility. The isolates were previously collected as part of the general care of the animals approved by the Institutional Animal Care and Use Committee at the University of Washington, United States, and the American Society of Primatologists (ASP) Principles for the Ethical Treatment of Nonhuman Primates (Soge et al., 2016). In addition, other animals were obtained from different commercial vendors and different sources outside the United States and were investigated shortly after their arrival at the WaNPRC during the quarantine period. MRSA-positive animals were given baths with chlorhexidine scrub for five consecutive days. The chlorhexidine was applied to the entire body and scrubbed with a surgical scrub brush with extra time spent cleaning axillary, perianal and preputial areas. In addition, animals received nasal application of mupirocin ointment 2% given twice daily for 5 days at the same time. Animals were sampled again at two and four weeks after initial MRSA positive culture and chlorhexidine and mupirocin treatment and retreated if still MRSA positive. All animals in the colony had initial nasals cultures done, while wound and/or skin infections were also sampled when present. All samples were taken from ketamine-sedated animals using standard microbiological swabs; BD BBL CultureSwab Plus Amies Medium (Becton Dickinson, Franklin Lakes, NJ, United States) and/or Starplex Starswab II (Starplex Scientific, Etobicoke, ON, Canada) as previously described (Roberts et al., 2011; Soge et al., 2016). For the current study, colonies were identified as S. aureus by production of β-hemolysis on blood agar plates and a positive Staphaurex^® test following manufacturer’s instructions (Remel, Lenexa, KS, United States; Soge et al., 2016). No isolate was selected unless they met these criteria (Soge et al., 2016). The presence of the alternative PBP2’ was determined with the Thermo Scientific PBP2’ latex agglutination test kit^® using instruction from the manufacturer (Thermo Fisher Scientific Remel Products, Lenexa, KS, United States). MRSA isolates were stored at -80°C. Isolates were selected without knowledge of the host primate species. This included 15 of 56 MRSA ST188 SCCmec IV isolates obtained from 36 animals and selected from various sample sites including animals that appeared refractory to mupirocin topical treatment. The 15 ST188 isolates came from ten M. nemestrina hosts and included three skin samples, and 12 nasal samples (Table 1). From M. nemestrina Z1242, three different nasal isolates Z1242N1, Z1242N2, Z1242N3, were selected taken on Feb 2, April 24, and June 5, 2015 to determine if the same strain was present over the 5 month time period. This animal was treated with chlorhexidine scrub and nasal application of mupirocin ointment between samplings. One M. nemestrina [Z121] had paired nasal Z121N and skin Z121S isolates taken May 29, 2015, while M. nemestrina Z123 had two isolates from two nasal samples [Z123N1 and Z123N2] isolated May 15 and 29, 2015 and a skin sample [Z123S] isolated May 29, 2015 (Table 1). This animal was treated with chlorhexidine scrub and mupirocin ointment when first identified as MRSA positive in May 2015. All these animals were from the WaNPRC facility.

Nine of the 21 ST3268 SCCmec V isolates were selected from animals representing different commercial vendors and out-state-location sources for the macaques. There were seven nasal samples and two wound samples. The isolates were selected without knowledge of the host primate species and included two isolates [A1404N nasal, A1404S skin sample both taken on July 20, 2015] from a SIV-positive M. mulatta [A1404] from WaNPRC. M. mulatta A1404 had close contact with the SIV-positive animal A140 [A140 nasal] and was also from the WaNPRC (Table 1). Both animals had a compromised immune system and bite wounds. The other six MRSA ST3268 isolates originated from macaques shipped from other United States primate sites, macaques shipped from two different commercial vendors [A1408, A1535] or directly shipped from China and having been in quarantine for 6 months in California before shipping to the WaNPRC [A1524] (Table 1). These nine isolates came from two M. mulatta [nasal isolates A140, A1404N, and one wound isolate A1404W], two M. fascicularis [nasal isolates A1524, A1525] and three M. nemestrina [two nasal A109, Z1403, one wound site isolate K990W] (Table 1).

All isolates were grown on Brucella agar (Difco Laboratories, Division BD Sparks, MD, United States) slants and shipped by courier to Germany for further molecular testing.

The Alere StaphyType^® DNA microarray was used for all isolates as previously described (Monecke et al., 2011, 2016). The microarray includes 334 target sequences and ∼170 separate genes and allelic variants including species markers, SCCmec, capsule, agr group typing markers, common antibiotic resistance genes, toxins and microbial surface components recognizing adhesive matrix molecules [MSCRAMM] genes. The latter genes comprise among others clfA and clfB (encoding clumping factors A and B), fnbA and fnbB (encoding fibronectin binding proteins A and B), fib (encoding fibrinogen binding protein), eno (encoding laminin binding protein), and cna (encoding collagen binding protein), the gene products of which play a role in the initial attachment of bacteria to host tissue. The detailed protocol as well as the sequences of primers and probes have previously been published (Monecke et al., 2011).

The clonal complexes (CCs) were determined by automated comparison of the microarray hybridization profiles to a database of previously characterized isolates (Monecke et al., 2011, 2016). The spa typing was performed according to Harmsen et al. (2003). The spa types were determined using the Ridom website.

The MLST typing was done using PCR and sequencing and the SCCmec typing was performed as previously described prior to being sent to Germany (Soge et al., 2016).

The antimicrobial susceptibility testing was performed for 30 antimicrobial agents by broth microdilution according to the Clinical and Laboratory Standards Institute (Clinical Laboratory Standard Institute [CLSI], 2018). The microtiter plates (MCS Swalmen, Netherlands) included penicillins (penicillin, ampicillin, amoxicillin/clavulanic acid, oxacillin), carbapenems (imipenem), a macrolide (erythromycin), a lincosamide (clindamycin), tetracyclines (tetracycline, doxycycline), aminoglycosides (gentamicin, streptomycin), a quinolone (ciprofloxacin), an oxazolidinone (linezolid), a glycopeptide (vancomycin), a streptogramin combination (quinupristin/dalfopristin), a phenicol (florfenicol), a pleuromutilin (tiamulin), and the combination trimethoprim/sulfamethoxazole. The aminoglycoside kanamycin was tested by broth macrodilution (Clinical Laboratory Standard Institute [CLSI], 2018, Supplementary Table S1). As there are no CLSI-approved clinical breakpoints applicable to primates other than humans, we used the human clinical breakpoints as listed in the CLSI document M100, 28th edition (Clinical Laboratory Standard Institute [CLSI], 2018). The breakpoints for the categories susceptible (S), intermediate (I) and resistant (R), are as follows: penicillin (S ≤ 0.12 μg/mL, R ≥ 0.25 μg/mL), oxacillin S ≤ 2 μg/mL, R ≥ 4 μg/mL, ciprofloxacin and quinupristin/dalfopristin (S ≤ 1 μg/mL, I = 2 mg/mL, R ≥ 4 μg/mL), gentamicin, doxycycline and tetracycline (S ≤ 4 μg/mL, I = 8 μg/mL, R ≥ 16 μg/mL), erythromycin (S ≤ 0,5 μg/mL, I = 1–4 μg/mL, R ≥ 8 μg/mL), clindamycin (S ≤ 0.5 μg/mL, I = 1–2 μg/mL, R ≥ 4 μg/mL), linezolid (S ≤ 4 μg/mL, R ≥ 8 μg/mL), trimethoprim/sulfamethoxazole (S ≤ 2/38 μg/mL, R ≥ 4/76 μg/mL), and vancomycin (S ≤ 2 μg/mL, I = 4–8 μg/mL, R ≥ 16 μg/mL) (Clinical Laboratory Standard Institute [CLSI], 2018, Supplementary Table S1). There are no clinical breakpoints for S. aureus for ampicillin, amoxicillin-clavulanic acid and imipenem, but if S. aureus strains are classified as resistant to oxacillin they are also considered as resistant to other β-lactams. Since there are no CLSI approved kanamycin breakpoints available, isolates with MICs of ≥ 64 μg/mL were tentatively considered as resistant (Feßler et al., 2010). Florfenicol and tiamulin are not used in human medicine and thus no breakpoints are available.

Susceptibility testing of the biocides benzalkonium chloride, chlorhexidine, glutardialdehyde, and isopropanol was also performed by broth macrodilution. For this, a bacterial suspension was prepared in a tryptone-saline-diluent (TSD; 1 g tryptone-peptone, 8.5 g sodium chloride in 1 L purified water) in a concentration of in 1 ×10⁸–1 × 10⁹ cfu/mL from 16 to 24 h old cultures on tryptic soy agar (TSA) (Roth, Karlsruhe, Germany). This suspension was diluted 1:10. From this dilution, 20 μl were added per each ml double concentrated tryptic soy broth (2× TSB) (Roth, Karlsruhe, Germany). One ml of this inoculum was added to a 2-fold benzalkonium chloride dilution series prepared in 1 mL volumes. The test ranges were as follows: benzalkonium chloride 0.00005–0.0008%, chlorhexidine 0.000025–0.0008%, glutardialdehyde 0.03–1%, and isopropanol 4 to 12%. The results were read after 24 h incubation at 37°C (Feßler et al., 2018).

SmaI macrorestriction analysis with subsequent pulsed-field gel electrophoresis was performed as previously described (Murchan et al., 2003) and the gels were analyzed according to the criteria Tenover et al. (1995) and (Deng et al., 2017).

Previously, nasal cultures were performed on 596 primates and 105 (17.6%) were MRSA positives. With the exception of four animals all in-house primates carried the MRSA ST188, while the MRSA ST3258 was associated with animals that were shipped into WaNPRC from other primate facilities and commercial breeders (Soge et al., 2016). M. nemestrina represent 75% of the primates in the WaNPRC. All ST188 and ST3268 isolates were positive for the species markers (rrnD1, gapA, katA, coA, nuc1, spa, sbi), capsule and agr alleles and consistent with an identification as S. aureus. All fifteen ST188 isolates selected for the study came from M. nemestrina hosts and were verified to have the ST188 MLST profile (3-1-1-8-1-1-1). All but one had spa type t189 (07-23-12-21-17-34), while the remaining isolate [Z143] had spa type t3887 (07-23-12-12-34). The nine ST3268 isolates had a MLST profile of 1-14-430-214-10-303-329. Two different spa types were identified, t13638 (n = 5) and t15469 (n = 4). The two spa types differed by the presence of an additional repeat 17 in spa type t15469 (210-23-02-34-17-34-34-17-17-23-34) compared to spa type t13638 (210-23-02-34-17-34-34-17-23-34) (Table 1). The spa type t13638 isolates were cultured from M. fascicularis and M. nemestrina. This spa type was first described in a methicillin-susceptible S. aureus from the United Kingdom². The spa type t15469, cultured from M. mulatta, is a novel spa type, first described in these primate isolates² (Table 1). In the ST3268 isolates, the spa types correlated with the host macaque species. The four isolates from M. mulatta hosts were spa type t15469, while the two M. fascicularis and M. nemestrina isolates were spa type t13638 (Table 1).

Nine of 15 ST188 [L091 (nasal), Z121N (nasal), Z121S (skin), Z123N1 (nasal) and Z123N2 (nasal), Z123S (skin), Z1304 (nasal), Z131S (skin), and Z143 (nasal)], from six M. nemestrina had indistinguishable PFGE patterns [A]. Five ST188 isolates originating from three animals [K062 (nasal), Z1242N1, Z1242N2 and Z1242N3 (nasal), and Z130 (nasal) shared PFGE sub-pattern [A1], while the ST188 isolate A112 (nasal) had a second PFGE sub-pattern [A2] (Table 1 and Supplementary Figure S1).

Of the nine MRSA ST3268 isolates, seven [A140 (nasal), A1404N (nasal), A1404W (wound), A1408 (nasal), A1524 (nasal), A1525 (nasal), and K990W (wound)] had the same PFGE pattern [B]. The isolates A1404N (nasal) and A1404W (wound) were cultured from the same animal eleven days apart and were indistinguishable in their PFGE patterns, their resistance pheno- and genotypes, as well as, their virulence genes (Table 1). The two sub-patterns B1 and B2 were found in found in single isolates A109 (nasal) and Z1403 (nasal), respectively (Table 1 and Supplementary Figure S1).

All 21 MRSA isolates were resistant to penicillin and oxacillin. They carried the mecA gene and the β-lactamase gene blaZ. All isolates were also resistant to ciprofloxacin. In addition, all ST188 isolates were resistant to macrolides and lincosamides via the erm(B) gene and carried the aminoglycoside resistance gene aacA-aphD mediating gentamicin and kanamycin resistance. The aacA-aphD gene was only present in five of the ST3268 isolates, which exhibited high kanamycin MICs (≥ 256 mg/L) and were classified as resistant or intermediate to gentamicin. The nine MRSA ST3268 isolates were all tetracycline resistant and carried the tet(K) gene, while only five MRSA ST188 isolates (K062, Z1242N1, Z1242N2, Z1242N3, and Z130), from 3 M. nemestrina, were resistant to tetracycline and carried the tet(K) gene (Table 1).

From some of the animals, several isolates taken at different time points [Z1242N2, Z1242N3, Z123N1, and Z123N2] were included. However, even after one or more rounds of mupirocin topical treatment and chlorhexidine baths, the MRSA isolates either persisted in the noses of these juvenile animals or the animals were re-infected or re-colonized. Treatment success was measured by MRSA-negative cultures at two and four weeks after treatment. If the animal was still MRSA-positive, it was considered as treatment failure. If this happened, the animal was retreated with mupirocin and chlorhexidine baths. This primarily happened in juvenile animals. Because this “treatment failure” was limited to juvenile animals the veterinarian staff felt that it suggested that the animals were refractory to clearance of the isolate, the isolate may have become resistant to mupirocin due to acquisition of the mupirocin gene mupA or an alternative resistance mechanism, or other characteristic of being a juvenile M. nemestrina rather than clearance and reinfection since there was no sign of clearance in two and four week samples (Table 1). However, none of these isolates or any of the other isolates in the study were resistant to mupirocin nor did they carry the mupA gene (Table 1).

All the isolates were tested for reduced susceptibility to benzalkonium chloride, while no change was seen with chlorhexidine, glutardialdehyde, or isopropanol. Some isolates including ST188 isolates K062, Z1242N1, Z1242N2, Z1242N3, Z130, and ST3268 isolates A1524, A1525, and K990W, had a benzalkonium chloride MIC of 0.0004% and carried the qacC gene. All other isolates, that did not harbor the qacC gene, had benzalkonium chloride MICs of 0.0001% (Table 1). No other change in the MIC of disinfectants were observed.

The nine ST3268 isolates had the enterotoxin gene cluster egc [seg, sei, selm, seln, selo, selu] and the additional enterotoxin genes sec and sel. In contrast, none of the ST188 harbored the enterotoxin gene cluster egc, sec or sel genes (Table 1). The fifteen ST188 and nine ST3268 isolates carried the hlgA locus [comprising of hlgA/lukF/lukS], leukocidin genes [lukD/E and lukX/Y], the aureolysin gene [aur], and the protease genes sspA, sspB, and sspP. The gene for the S. aureus surface protein G [sasG] was present among the ST3268 isolates but absent in the ST188 isolates. Two isolates were additionally tested with a new array and both A1403 and Z140 were positive for the carotinoid pigment gene cluster [crtM/N/O/P]. Other isolates were not tested.

In contrast, the enterotoxin H gene [entH], ORF CM14, and splE were absent in all isolates (Table 1). The collagen-binding adhesin [cna] and the protease genes splA, splB were present in the ST188 isolates but were not detected among the ST3268 isolates. None of the 21 isolates carried PVL genes, the toxic shock syndrome toxin 1 gene [tst1], exfoliative toxin genes [etA, etB, etD], or genes associated with β-haemolysin converting phages (sea, see, scn, chp) (Table 1).

Two ST3268 SCCmec V isolates, A140 from a M. mulatta and Z1403 from a M. nemestrina, were further tested for SCCmec accessory genes. The following genes were identified in both mvaS, cstB-SCC2, ydhK, D1GU38, Q4LAG7, czrC, “ccrAA” (a recombinase homologue associated with ccrC), and a SCCmec terminus type 2 (Monecke et al., 2016). This is consistent with the presence of SCCmec VT+czrC composite elements as described for the CC398 strain SO385 (GenBank accession number AM990992.1), a livestock-associated MRSA strain from Western Europe (Schijffelen et al., 2011).

All isolates from the same animal shared indistinguishable PFGE patterns, regardless of whether nasal samples were taken at different times, or nasal and skin samples taken at the same time from the same M. nemestrina. As shown below, isolates from the same animal were also indistinguishable with respect to their resistance pheno- and genotypes, and other genes including enterotoxin, hemolysin, leukocidin, or PVL genes (Table 1), suggesting the presence of the same or a closely related strain in different locations of the animal and/or the persistence of that strain over time.