New Insights into Molecular Characterization, Antimicrobial Resistance and Virulence Factors of Methicillin-Sensitive Coagulase-Positive Staphylococcus spp. from Dogs with Pyoderma and Otitis Externa

: The first Tunisian national molecular survey of coagulase-positive staphylococci (CoPS) isolated from dogs with pyoderma and otitis externa was conducted to evaluate the prevalence of CoPS and identify its phenotypic and genotypic diversities. A total of 99 out of the 195 samples collected from 39 sick dogs were identified across multiple sites as methicillin-susceptible CoPS belonging to the species S. pseudintermedius (64.4%), S. aureus (20.2%), S. coagulans (10.1%), and S. hyicus (5%). Fifteen sampled dogs carried more than one Staphylococcus species. Their antibiotic resistance and virulence factors were determined using conventional and molecular methods. Of the S. pseudintermedius isolates found, 17.4% were multidrug-resistant, whereas high rates of virulence genes were observed among the S. aureus isolates. On polystyrene surfaces, 75% of S. aureus isolates were biofilm producers, of which 15% were classified as strong producers. The capsular polysaccharide cap8 genotype was predominant among them. A MultiLocus Sequence Typing (MLST) analysis clustered the S.aureus isolates into five distinct sequence types (STs), with four assigned for the first time. Our findings highlight the spread of CoPS among diseased dogs and, especially, the emergence of S. hyicus, S. coagulans , multidrug-resistant S . pseudintermedius and S. aureus isolates with high genetic variability. The precise characterization of these strains, as well as their continuous monitoring, is necessary for the implementation of preventive strategies given the significant public health risk.

Considered the most relevant coagulase-positive pathogenic staphylococci, S. aureus and S. pseudintermedius have been of interest to human and veterinary medicine [7].Both species are responsible for various infections in different animal species, such as dogs, cats, horses, cattle, rabbits, poultry, fish, and primates.Dogs are highly colonized by CoPS, especially S. pseudintermedius and, to a lesser extent, S. coagulans and S. aureus, which are responsible for several diseases, namely pyoderma, dermatitis, otitis, and systemic infections within the urinary, respiratory, and reproductive tracts [8][9][10][11].
S. hyicus commonly infects pigs and is responsible for porcine exsudative epidermitis [12,13].It can also cause bovine mastitis [14], polyarthritis [15], and skin infections in cattle, horses, and goats.S. hyicus is not considered a zoonotic species, but it has been isolated from humans with severe clinical conditions, namely sepsis [16][17][18].A few studies have reported the isolation of this species in dogs [19].
It is well known that CoPS are versatile pathogens that express various potential virulence factors and have the ability to form a biofilm, allowing them to attach and adhere to host cells, occupy a niche (colonization and tissue invasion), escape or break down host immune shields (immune evasion), generate toxin-mediated syndromes and induce toxinosis.CoPS are of significant concern as; in addition to their pathogenicity, morbidity, and mortality rates, their zooanthroponotic transmission is compounded by the frequent emergence of multidrug resistance (MDR) and the widespread methicillin-resistant (MR) strains [27].
Consequently, the emergence of drug-resistant CoPS across the globe has become a serious threat due to their poor prognosis and the lack of novel drugs, as well as the complications and limitations of their treatment [28].
Methicillin-susceptible staphylococcus aureus (MSSA) and pseudintermedius (MSSP) have also been found in healthy dogs and are as prevalent and virulent as methicillin-resistant staphylococcus aureus (MRSA) and pseudintermedius (MRSP) isolates [29][30][31], suggesting that dogs may have a pertinent role in transmitting infections and could contribute to disease perpetuation.These data highlight that the potential risks of infection with methicillinsusceptible strains should be considered seriously.However, limited data are available on the occurrence of methicillin-susceptible strains in diseased dogs.
The current study is the first national molecular survey of CoPS isolated from dogs with pyoderma and otitis externa in Tunisia and was conducted in order to broaden our understanding of phenotypic and genotypic characteristics of methicillin-sensitive CoPS.

Sampling and Bacteriological Analyses
Samples were collected from 39 diseased dogs presenting with pyoderma, dermatitis, or otitis.These animals came from Tunis and several nearby districts and were brought to the National School of Veterinary Medicine by their owners for a clinical consultation.Each dog's nasal, rectal, auricular, cutaneous, and oral mucosal surfaces were sampled (n = 195).Commercial sterile cotton-tipped swabs were used; they were rubbed against the mucosal surface for approximately 5-10 s.
Swabs were directly inoculated in a brain-heart infusion broth (bioMérieux, Craponne, France) for 24 h at 37 • C, with 6.5% sodium chloride (Sigma-Aldrich, Saint-Louis, MO, USA) and 10% mannitol added.A loopful of each broth inoculum was streaked on a selective medium, mannitol salt agar (bioMérieux, Craponne, France), and the culture plates were incubated at 37 • C for 24-48 h.Cultures positive for staphylococci were subjected to identification procedures based on standard bacteriological methods, including colony morphology, Gram staining, catalase testing (Pharmaghreb, Tunis, Tunisia), and coagulase production (Biorad, Hercules, CA, USA).The strains identified as belonging to the staphylococcus genus were subjected to species-specific polymerase chain reaction (PCR) assays using universal primers.The sequences of primers used for the species' identification, as well as amplification conditions, are mentioned in the Supplementary Table S1.

DNA Extraction and PCR Conditions
Bacterial DNA from Staphylococcus spp.isolates was extracted from a fresh overnight culture on a nutritive agar plate (bioMérieux, Craponne, France).A single colony was picked, suspended homogeneously in 100 µL of a TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0), heated at 99 • C for 15 min to release the DNA, and centrifuged at 1000× g for 5 min at 14 • C. The supernatant was then used as a template for the PCRs.All PCRs were performed once in a total volume of 25 µL using 12.5 µL Taq DNA polymerase 2X-preMix (GenOn, Ludwigshafen am Rhein, Germany), 0.75 µL of each primer at a concentration of 10 µM, 1-10 µg/mL of genomic DNA, and up to 25 µL of sterile distilled water.For each reaction, negative and positive controls were added.

Genotyping of Capsular Polysaccharide Types
The detection of different capsular polysaccharide types was performed based on the amplification of two involved genes by PCR (i.e., cap5, cap8).The sequences of primers used and amplification conditions are mentioned in Table S1.

Biofilm Production Assay of S. aureus Isolates
The assessment of the S. aureus isolates' biofilm formation was achieved using the method described by Stepanovic et al. [32].In brief, sterile 96-well flat-bottomed polystyrene plates (Merck, Darmstadt, Germany) were filled, in triplicate, with the dilution (1/100) of an overnight bacterial culture in Tryptic Soy Broth (bioMérieux, Craponne, France) supplemented with 1% glucose.A negative control was established that contained only a growth medium.Following their incubation at 37 • C for 24 h without shaking, the plates were washed three times with PBS and dried at room temperature before adding a 1% crystal violet solution (Sigma-Aldrich, Saint-Louis, MO, USA) (100 µL/well).The plates were subsequently incubated at room temperature for 30 min and then washed again.After drying for 1 h at 60 • C, 150 µL of ethanol (95%) was added to each well as a mixture of 50% ethanol-50% acetic acid (polychem, Agra, India), and the absorbance was measured at 570 nm using a spectrophotometer.
The results are reported based on average OD values and the cut-off value ODc (mean OD of negative control-3 SDs of negative control).Strains were grouped into the following four categories: not a biofilm producer (OD ≤ ODc); weak biofilm producer (ODc < OD ≤ 2× ODc); moderate biofilm producer (2 ≤ ODc <OD × 4 ODc); and strong biofilm producer (4 × ODc < OD).

MultiLocus Sequence Typing of S. aureus Isolates
MLST was accomplished using the amplification of seven housekeeping genes (arcC, aroE, glF, gmK, pta, tpi, and yqil), and the subsequent assignment of clonal complexes (CCs) was performed for five representative S. aureus strains, as recommended (https://pubmlst.org/accessed on 14 July 2023) via PCR and sequencing.The primers and conditions are detailed in Table S1.The allele sequences were compared with the NCBI nucleotide database (http://www.ncbi.nlm.nih.gov/accessed on 3 July 2023), and ST numbers were allocated based on the key table in the database for the MLST of S. aureus.New STs were assigned by the curator of the MLST database, Vincent Perreten (vincent.perreten@vbi.unibe.ch).
Table 1.Colonization rates of CoPS species isolated from diseased dogs.Fifteen dogs (38.46%) carried more than one Staphylococcus species: five carried both S. pseudintermedius and S. aureus, and four carried S. pseudintermedius and S. coagulans, three were co-infected by S. pseudintermedius and S. hyicus, one carried S. aureus and S. hyicus, and one carried S. pseudintermedius, S. coagulans and S. hyicus.

S. hyicus (n) Total
S. intermedius and S. delphini were not identified in any of the tested isolates.

Antibiotic Susceptibility Profiles
The antimicrobial susceptibility of the CoPS isolates is summarized in Figure 1 and Tables 2 and 3.

Detection of Antimicrobial Resistance Genes
The molecular detection of the mecA gene via PCR revealed that all staphy isolates were susceptible to methicillin.

Detection of Antimicrobial Resistance Genes
The molecular detection of the mecA gene via PCR revealed that all staphylococci isolates were susceptible to methicillin.
In contrast, all MSSA strains carried at least six virulence genes: twelve isolates harbored the highest number of virulence genes (more than 10 genes).A summary of the virulence gene profiles of the MSSA strains is detailed in Table 3.
Of the exfoliative toxin-encoding genes, eta and siet were identified in 25% and 30% of isolates, respectively.Conversely, none of the isolates carried the etb or tsst-1 toxin genes.Fourteen isolates (70%) harbored the leukocidin (lukE/D) gene, whereas the PVL (lukS/F PV) and lukM toxin genes were not identified in any of the isolates.
Additionally, the prevalence of ebp and cbp was moderate (30-40%), while that of bsbp was the lowest (5%).In contrast, all MSSA strains carried at least six virulence genes: twelve isolates harbored the highest number of virulence genes (more than 10 genes).A summary of the virulence gene profiles of the MSSA strains is detailed in Table 3.
Of the exfoliative toxin-encoding genes, eta and siet were identified in 25% and 30% of isolates, respectively.Conversely, none of the isolates carried the etb or tsst-1 toxin genes.Fourteen isolates (70%) harbored the leukocidin (lukE/D) gene, whereas the PVL (lukS/F PV) and lukM toxin genes were not identified in any of the isolates.
We also observed high heterogeneity and many combinations of the genes encoding staphylococcal enterotoxins (ETs) in the MSSA strains.The majority of isolates possessed at least two ET genes, and the highest number, with six ETs, was found in only one isolate.The seccan variant was found in 60% (n = 12) of isolates, followed by seh in 25%, seg and sej in 20%, sea in 15%, and seb, sem, sen in 10% of isolates.The sei gene was detected in a minority of isolates (n = 1, 5%).Surprisingly, the four isolates did not harbor any previously cited enterotoxin-encoding genes.
The sed, see, and seo enterotoxin-and the edin exotoxin-encoding genes were absent in all the investigated isolates.

Biofilm Production
Of the twenty S. aureus isolates analyzed, fifteen (75%) were biofilm producers.Of these, three isolates (20%) were classified as strong biofilm producers, eleven (55%) were considered to be moderate biofilm producers, and one (5%) was a weak biofilm producer.We also observed high heterogeneity and many combinations of the genes encoding staphylococcal enterotoxins (ETs) in the MSSA strains.The majority of isolates possessed at least two ET genes, and the highest number, with six ETs, was found in only one isolate.The sec can variant was found in 60% (n = 12) of isolates, followed by seh in 25%, seg and sej in 20%, sea in 15%, and seb, sem, sen in 10% of isolates.The sei gene was detected in a minority of isolates (n = 1, 5%).Surprisingly, the four isolates did not harbor any previously cited enterotoxin-encoding genes.
The sed, see, and seo enterotoxin-and the edin exotoxin-encoding genes were absent in all the investigated isolates.

Biofilm Production
Of the twenty S. aureus isolates analyzed, fifteen (75%) were biofilm producers.Of these, three isolates (20%) were classified as strong biofilm producers, eleven (55%) were considered to be moderate biofilm producers, and one (5%) was a weak biofilm producer.Five isolates (25%) did not produce biofilm.

MLST Typing of S. aureus Isolates Detected in This Study
The MLST analysis clustered the MSSA isolates into five major STs, four of which were assigned for the first time.The new MLST genotypes were submitted and registered on the MLST database of S. aureus as ST36548, ST36549, ST36550, and ST3651.The known ST was ST5896 (n = 8), which belongs to the CC15 complex (Table 3).

Discussion
Pyoderma (or skin infections) and otitis externa are the main diseases seen in dogs brought to veterinary clinics.These infections are predominantly caused by coagulasepositive staphylococci strains and mainly by S. pseudintermedius, which is known as an opportunistic pathogen and a common colonizer of animals' skin and mucosal cavities.This species is followed in prevalence by other species, including S. cogulans, S. aureus, and, to a lesser extent, S. hyicus [33].In this context, this study aimed to be the first in Tunisia to identify the occurrence of CoPS in dogs with clinical conditions and investigate their antimicrobial resistance profiles while analyzing the molecular structure of the detected isolates since the only two Tunisian studies previously published both focused on healthy animals [34,35].
Our results demonstrate that S. pseudintermedius isolated from sick dogs had the highest colonization rate (64%), which is consistent with the rates found in several other studies in which this species' prevalence ranged from 60% to 96% [36][37][38][39][40][41], confirming that this is a key opportunistic pathogen that is responsible for most cases of pyoderma in dogs.A Tunisian study reported a similar prevalence (55%) in healthy dogs [35].
Although S. aureus is not a frequent commensal in the skin or mucous cavities of dogs, a relatively high prevalence of this species (20%) was observed among sick dogs in our study.Relatively lower prevalences, from 8.8 to 16%, were found in other studies conducted on dogs with skin infections [42,43].Only one Tunisian study was published in this context and revealed a low prevalence (4%) of S. aureus carriage in healthy dogs [34].
Staphylococcus coagulans is known to be the second most common CoPS species, after Staphylococcus pseudintermedius, and is isolated from dogs.We detected this species at a rate of 10%, which is consistent with other studies that found it at rates ranging from 6% to 14% [44][45][46].In a study conducted on 89 pyoderma-related staphylococcus isolates, 27 (30.33%)were identified as S. coagulans [47].The three species described above were found at all body sites sampled.Their presence across multiple body sites may be due to the fact that dogs, during pyoderma, cause themselves trauma by rubbing and licking their hair and skin, promoting the spread of bacteria to other sites.
Additionally, S. hyicus has been isolated from healthy animals and animals infected with septic arthritis and mastitis [14,15], namely from pigs, poultry, cattle, goats, and horses.In dogs, a case of mastitis and lymphadenitis caused by S. hyicus has been reported [16], and a few studies are available on its occurrence in dogs with pyoderma and/or otitis.In our study, we report a S. hyicus species prevalence of 5% in dogs with pyoderma and/or otitis.The isolated strains were present in their nasal cavity (10%), perineum (8%), and skin (7.14%), with complete absence in the mouth and ear.A recent study detected one strain (2.32%) of S. hyicus in skin samples collected from a dog in an animal shelter in Timisoara [48].Another study also found, among 50 screened sick dogs, only one strain of S. hyicus isolated from the skin [19].In healthy dogs, Vani et al. reported the same low incidence rate (5.8%) as ours, but the most frequent carriage sites they reported were the skin (44.8%) and ear (37.9%) [49].
It is noteworthy that, in the current study, dual or triple carriages (in one case) of multiple species were found on the same animal.Effectively, the interactions of competitive species depend on host factors and/or the antagonism between colonizing bacteria.This competitive interference impacts their virulence and colonization abilities, especially when the species are S. aureus and S. pseudintermedius.In fact, niche competition might contribute to species concurrence, driving S. aureus to antagonize S. pseudintermedius colonization (in a dog's natural microbiota)-via the negative regulation of their virulence gene expressions-and predominate in oral and cutaneous sites [50].
Over the years, the emergence and spread of antibiotic resistance among the strains described above have posed significant challenges to the treatment of these infections in dogs, particularly methicillin-resistant (MR) strains.
In our study, there was no occurrence of methicillin resistance in the screened dogs, which is compliant with previously reported results, including those in the two Tunisian studies [34,35,51].In contrast, other investigations have suggested considerably higher rates of MR in both healthy and diseased dogs [19,39,41,52].
On the other hand, a high occurrence of β-lactamase (penicillinase), which is responsible for resistance to penicillin (blaZ), was distinguished among the tested Staphylococcus isolates but at a lower frequency than previously reported rates in both diseased and healthy dogs [35].
The antimicrobial susceptibility of the 64 investigated MSSPs showed a high-tomoderate resistance to non-beta-lactam antibiotics.In a past Tunisian study conducted by Gharsa et al. in 2013 [34] on healthy dogs, resistance to chloramphenicol, erythromycin, and clindamycin was low compared to our study, while higher rates of resistance to sulfamethoxazole-trimethoprim and fusidic acid were found.Additionally, they detected the same low prevalence of resistance to fluoroquinolones, and all their MSSP strains were gentamycin-susceptible, while a resistance rate of 6% was found for gentamycin.
A discrepancy between phenotype and genotype characterization was noticed in some cases in our study since one fusA-negative isolate was found to be resistant to fusidic acid.This may be explained by the presence of other mechanisms involving other genes related to fusidic acid resistance.A few publications have evaluated the mechanisms of fusidic acid resistance in S. pseudintermedius from dogs with pyoderma and otitis externa [53,54].The carriage of fusC has been shown to be the predominant mechanism of fusidic acid resistance, and, often, so has that of fusB or fusD.The acquisition of fusB family genes (fusB, fusc, and fusD) that encode cytoplasmic proteins protects the drug target site [55,56].Additionally, this antibiotic is recommended for the topical management of canine pyoderma [57].
As such, substantial epidemiological studies are required to outline the prevalence and the mechanism of fusidic acid resistance in the staphylococci from both healthy and diseased dogs.
Similarly, the streptomycin adenylyl-nucleotidyltransferase gene ant(6)-Ia was found in only 40% of streptomycin-resistant isolates.This mismatching could be due to other resistance genes, such as the aad(6 ′ )/aph(2 ′′ ) and ant(4 ′ )-Ia genes, as described in other outcomes.Indeed, it has been reported that the most prevalent aminoglycoside-resistance gene targeted among MRSP strains is aph(3)-IIIa, followed by the aad(6 ′ )/aph(2 ′′ ) and ant(4 ′ )-Ia genes [58][59][60].Over the last few years, the number of aminoglycoside-resistant and aminoglycoside-modifying enzymes (AMEs) has increased.AMEs are considered to have potent synergistic effects with other classes of antibiotics, such as beta-lactams, and a significant association with isolates' resistance to methicillin and aminoglycosides [61].This could contribute to the spread of MRSA or MRSP and MDR strains, complicate the curing of infections, and limit the effective choice of antibiotics.
The tetM gene was also found in S. pseudintermedius isolates with resistant phenotypes.One S. hyicus isolate, harboring the tetM gene, was isolated from a dog co-infected with tetM-positive S. pseudintermedius, which may prove that there is horizontal gene transfer.
In general, the absence of the gene of interest in isolates showing phenotypic resistance is likely due to the curation of other resistance genes, the loss of the respective primerbinding site, or other unknown resistance mechanisms.Conversely, some other isolates have appeared phenotypically susceptible while exhibiting the related resistance gene (as with the case of the tetM-positive S. coagulans isolate, which was phenotypically susceptible).This means that the fact that a strain harbors the gene without expressing it may be due to an error in the considered breakpoints or the presence of a silent resistance gene.
Alarmingly, the increasing trends of multidrug resistance in MSSP isolated from Tunisian dogs, compared to other Staphylococcus species, could result in the emergence of harmful Staphylococcus within the community.Thus, continued resistance surveillance targeting both safe and sick dogs would be an effective way to identify resistance in the neighborhood.
Globally, these broad variations might be attributed to geographic locations, the source, and size of specimens, the breed of dogs, the presence of the disease, the methods used, the monitoring period, and antimicrobial use policies.
Furthermore, the presence of virulent factors enhances the pathogenic potential of these bacteria.In this context, the ability to form biofilm is one of the most important virulence determinants and survival mechanisms of S. aureus infections, and it is also involved in antimicrobial resistance [63].Several biofilm-associated genes are behind these infections' adherence to the host's extracellular matrix and abiotic surfaces, including the binding proteins (eno, fnbA/B, fib, ebp, cbp), the clumping factors, hemolysins, and the capsule proteins [64].All these genes were prevalent in the MSSA isolated in the current study.The capsule genotypes were closely associated with S. aureus strain lineages.Indeed, capsular type 5-which was expected to produce more biofilm-was less predominant among our MSSA isolates.The same observation was also reported by Verdier et al. [65].Fifty percent of biofilm-positive MSSA isolates were identified as agr type II, and most biofilm-negative MSSA isolates are agr type I.It is noteworthy that a significant correlation was noted between biofilm formation, capsular expression, group agr, and intracellular survival [66].It was hypothesized that strains adapted to their niche, suggesting that intracellular chronic infection may correspond to the cap5-agrI group and that the cap8-agrII group may correspond to strains better adapted to the extracellular niche, leading to a low invasion rate and acute infection.Hence, it is remarkable that one isolate carried multiple capsule types (a carriage of both capsule types 5 and 8).A similar result was reported by Zhang et al. [67], who verified the positive association of this genotype with agr-positive isolates.
The hla gene, encoding the major virulence factor of S. aureus, α toxin (alpha-hemolysin), was frequently detected (85%) in our study, whereas the hlb (beta-hemolysine) gene was less often found (30%).The combination of hla and hlb has been reported to enhance biofilm formation [68,69].Also, the contribution of hld to biofilm formation has been demonstrated [70].
In the case of S. aureus, 35%, 20%, and 15% of isolates displayed a combined occurrence of (a, d, and g), (a, b, d, and g), and (b, d, and g) hemolysin toxin genes, respectively.Profiles with combinations of the two patterns (a and d) and (d and g) were also found in 20% and 5% of the isolates, respectively.The hemolysin combination of α and γ is the most common in human strains, while hemolysins α, β, and γ are mostly found in animal strains, which is consistent with our results [71,72].The variable pattern of hemolysins reflects their possible contamination of animal or human products.
The same is true for the leukocidin lukE/D gene, considered an important virulence factor [73,74], which was evident in all tested Staphylococcus species.It has been previously reported in S. aureus, including MSSA and MRSA strains [75][76][77].Overall, pore-forming toxins, including hemolysins and leukocidin, as bacterial invasive factors, contribute to many staphylococcal diseases, such as dermonecrosis and impetigo [78], which can therefore explain the skin lesions observed.
In our study, the eta-encoding gene, recognized as an epidermolytic toxin, was detected in 25% of S. aureus isolates.While it is involved in superficial skin infections, such as bullous impetigo [79], low rates (or even its total absence) have been previously reported among the S. aureus strains isolated from sick [19,80,81] and healthy dogs [82].
The siet toxin gene was detected in more than 96% of all tested staphylococcal species.It was found to be produced by almost all the S. pseudintermedius isolates in dogs with pyoderma or chronic otitis [83][84][85][86][87][88] and in healthy dogs [35,80].Siet has been reported to be potentially involved in the etiopathogenesis of scaly and purulent skin infections; it especially causes crusting, erythema, impetigo, exfoliation, and/or superficial pyoderma in dogs [89,90].
Intriguingly, none of the tested SEs were detected in the S. pseudintermedius, S. hyicus, or S. coagulans isolates, while almost all S. aureus isolates harbored at least one SEs gene, and 20% carried up to three genes.This is in contrast to a previous Tunisian study, which found that MSSP isolates from healthy dogs harbored many enterotoxin genes, including sea, seb sec, sed, sei, sej, sek, ser, and sec can [35].
The SEs of the MSSA isolates were randomly distributed without a constant profile.Among the enterotoxins associated with food poisoning (FP), sea and seh and seb, seg, and sei (often associated with FP) were detected in our S. aureus isolates.The sec can variant gene was the most predominant emetic SE (60%) found in our S. aureus isolates, similar to the earlier results found in S. pseudintermedius from dogs with pyoderma and otitis [91,92].Its frequent association with atopic dermatitis and pyoderma in dogs underlines its importance in the survival and pathogenesis of staphylococcus [83,93].
Although the egc SE genes belong to an operon of the egc enterotoxin gene cluster, which contains five enterotoxin genes (seg, sei, sem, sen, and seo) and two pseudogenes, an incomplete egc locus form lacking one or more genes was observed among our S. aureus isolates.The combination of various SE genes seemed to not be directly related to the virulence factors of pyoderma but could advance the occurrence and gravity of S. aureus diseases.
Overall, we noticed a negative trend between the resistance and virulence of primarily the S. aureus and S. pseudintermedius strains.In particular, S. pseudintermedius strains harbored a large reservoir of antibiotic resistance genes, while S. aureus strains carried a broad pool of virulence genes.It has been documented that increased antibiotic resistance almost certainly has a negative effect on the relationship between resistance and virulence in environments where there is no direct selective antibiotic pressure (low antibiotic concentrations) [94].
Altogether, the relationship between resistance and virulence among bacteria depends on the bacterial species, the specific mechanisms of their resistance and virulence, their ecological niche, and the host.
Lastly, the genotyping analysis of S. aureus isolates showed a high diversity of sequence types (STs), with four new STs assigned: ST3658, ST36549, ST3650, and ST3651.All STs were neither phylogenetically related to each other nor to the other STs described in the database.Moreover, their MLST lineages seemed not to be associated with a specific AMR and/or virulence gene profile.
The known ST5896 found in four of the eleven sick dogs belonged to clonal complex 15 (CC15).CC15 has been identified mostly in MSSA strains with nasal colonization [95,96] but also in MRSA from clinical infections in the Middle East, Iran, Kuwait, and Italy [97][98][99][100] and from retail meat products in the Middle East [101].It has been demonstrated that the complex CC15 carries agr2 and completely lacks superantigen (SAg) genes.Unfortunately, our CC15 isolates carried sea, sec can , seg, she, sei, sem, and eta superantigens.

Conclusions
These findings highlight the emergence of S. hyicus in diseased dogs in Tunisia for the first time, as well as multidrug-resistant S. pseudintermedius and S. aureus pathogens with distinct STs.Our study reveals the need for further investigative studies aiming to understand the phylogenetic evolutionary relationship of these CoPS, their biofilm formation ability and the mechanisms underlying them, as well as the need to implement powerful preventive strategies.
, N: nasal mucosa, O: oral, R: oral mucosa, S: skin mucosa, and n: number of isolates.

Table 2 .
Characteristics of non-aureus CoPS isolates recovered from 39 diseased dogs.

Table 3 .
The antibiotic resistance patterns, genetic lineages and virulence-associated genes of the MSSA isolated from diseased dogs.