The Home Environment Is a Reservoir for Methicillin-Resistant Coagulase-Negative Staphylococci and Mammaliicocci

Coagulase-negative staphylococci (CoNS) and mammaliicocci are opportunistic human and animal pathogens, often resistant to multiple antimicrobials, including methicillin. Methicillin-resistant CoNS (MRCoNS) have traditionally been linked to hospitals and healthcare facilities, where they are significant contributors to nosocomial infections. However, screenings of non-hospital environments have linked MRCoNS and methicillin-resistant mammaliicocci (MRM) to other ecological niches. The aim of this study was to explore the home environment as a reservoir for MRCoNS and MRM. A total of 33 households, including households with a dog with a methicillin-resistant staphylococcal infection, households with healthy dogs or cats and households without pets, were screened for MRCoNS and MRM by sampling one human, one pet (if present) and the environment. Samples were analyzed by a selective culture-based method, and bacterial species were identified by MALDI-TOF MS and tested for antibiotic susceptibility by the agar disk diffusion method. Following whole-genome sequencing, a large diversity of SCCmec elements and sequence types was revealed, which did not indicate any clonal dissemination of specific strains. Virulome and mobilome analyses indicated a high degree of species specificity. Altogether, this study documents that the home environment is a reservoir for a variety of MRCoNS and MRM regardless of the type of household.


Introduction
Coagulase-negative staphylococci (CoNS) and Mammaliicoccus spp.(formerly known as the Staphylococcus sciuri group) are a heterogeneous group of skin and mucous membrane commensals and are also opportunistic pathogens responsible for various infections in humans and animals [1,2].They are considered to have a lower pathogenic potential than the more virulent Staphylococcus aureus and Staphylococcus pseudintermedius.Still, CoNS cause a substantial number of infections in immunocompromised as well as in otherwise healthy patients [3].Staphylococcus epidermidis, Staphylococcus hominis and Staphylococcus haemolyticus are significant contributors to septicemia in neonates, while Staphylococcus saprophyticus is one of the most common causative agents in urinary tract infections [4][5][6].
In addition to their opportunistic pathogenic potential, CoNS and mammaliicocci are thought to be important reservoirs for antimicrobial resistance genes (ARGs), including the mecA gene responsible for methicillin resistance, and mobile genetic elements associated with ARGs [1,7].Thus, the potential for horizontal gene transfer of these genes to more pathogenic bacteria, such as S. aureus, is present.
As opposed to human and animal infections caused by either methicillin-resistant S. aureus (MRSA) or methicillin-resistant S. pseudintermedius (MRSP), methicillin-resistant MRCoNS infections are not monitored in Norway.Hence, the knowledge of the occurrence of MRCoNS is largely unknown.Based on reports of high rates of MRCoNS in cases of neonatal septicemia and MRCoNS-carrying health care personnel, we can assume that Norwegian healthcare facilities make up important reservoirs for MRCoNS [6,8].However, non-hospital environments may also serve as reservoirs for these bacteria [9].High rates of MRCoNS have been reported in public transportation systems and humans without previous exposure to health care systems [10,11].In a previous study on the transmission of MR Staphylococcus spp.(MRS) from infected dogs to their owners, we found several species of MRCoNS and MR Mammaliicoccus spp.(MRM) in the same home environments [12].These findings led us to wonder if there was an association between the dogs' infection status and the existence of MRCoNS and MRM in their homes, or if MRCoNS was a standard feature in all kinds of homes.
Thus, this study aims to gain more insight into the home environment's role as a reservoir for MRCoNS by screening the environment, humans and pets in different households for MR bacteria.Furthermore, we examine the distribution of MRCoNS and MRM by identifying the species, sequence types and SCCmec elements in the home environment.Finally, we screen the bacteria for virulence genes, ARGs and mobile genetic elements.

Isolates
A total of 117 verified MRCoNS and MRM isolates constituted the sample material for the occurrence analysis.Of these, 103 isolates were submitted for whole-genome sequencing, of which 75 are presented in the resistome-, virulence-and SCCmec results.Thirty-nine S. epidermidis, S. haemolyticus and S. hominis isolates were included in the in silico MLST analysis, while 57 S. epidermidis, S. haemolyticus, S. hominis and S. saprophyticus isolates are presented in the mobilome analysis.

Occurrence of MRCoNS and MRM
All but three households (30/33) tested positive for a minimum of one species of MRCoNS or MRM (Table 1).The majority of the studied MRCoNS/MRM isolates were detected in the home environments (n = 107/117).Overall, MR S. saprophyticus, S. haemolyticus, S. hominis and S. epidermidis were the most prevalent species in the households.Additionally, some species appeared to be more prevalent in specific households, as follows: MR S. epidermidis (MRSE) in homes with infected dogs or healthy cats and MR S. hominis in homes with infected and/or healthy dogs.Households with infected dogs had a higher mean number of different species (3.13) compared to those with healthy pets (1.79) and without pets (1.55).A one-way ANOVA revealed that there was a statistically significant difference in the mean numbers of species between the household groups (F (2.30) = (3.487)p = 0.04).The Tukey's HSD test found that this difference was between the homes with infected pets and those without pets (p = 0.03).As shown in Table 1, six humans and three dogs carried MRCoNS.Two of these dogs suffered from MRSP infections, in addition to carrying MRSE, while the third dog (household H) had an MRSE infection.Their three owners tested positive for MRSE.Two owners carried MR S. haemolyticus, one owner of an infected dog and one owner of a healthy cat.One owner of a healthy dog carried MR S. warneri.None of the humans in the household without pets tested positive for MRCoNS, while all but one of the eleven home environments tested positive for at least one species of MRCoNS.

Antimicrobial Resistance
Forty-one of the seventy-five isolates were multi-resistant, expressing resistance to three or more classes of antimicrobials.S. cohnii ssp.cohnii, S. hominis and S. haemolyticus were the most resistant species, expressing resistance to means of 4.3, 3.7 and 3.6 classes, respectively (Figure 1).Following resistance to anti-staphylococcal beta-lactams, resistance to macrolides (erythromycin) and fusidanes (fusidic acid) was most frequently observed (Table 2).Worth noticing is the large proportion of S. hominis isolates that expressed no phenotypic resistance to cefoxitin and amoxicillin clavulanic acid.The resistome analysis supported the phenotypical resistance profile, displaying a high prevalence of genes conferring resistance to erythromycin and fusidic acid among the MRCoNS and MRM (Table 3) and further confirmed that all sequenced MRCoNS and MRM possessed the mecA gene.Despite the presence of mecA, we observed a variable phenotypic expression of resistance to beta-lactams among the isolates (Table 2).This was particularly evident for S. hominis, with 1/11 isolates being phenotypically susceptible to oxacillin and 5/11 isolates being susceptible to cefoxitin.Furthermore, we observed phenotypic susceptibility to amoxicillin-clavulanic acid among six MRCoNS and MRM species: S. cohnii ssp.cohnii (2/3), S. epidermidis (8/12), S. haemolyticus (2/16), S. hominis (6/11), S. warneri (1/5) and M. vitulinus (n = 1).In addition to mecA, the beta-lactamase-encoding blaZ gene was present in 16 of the 20 amoxicillin-clavulanic-acid-susceptible isolates.

Antimicrobial Resistance
Forty-one of the seventy-five isolates were multi-resistant, expressing resistance to three or more classes of antimicrobials.S. cohnii ssp.cohnii, S. hominis and S. haemolyticus were the most resistant species, expressing resistance to means of 4.3, 3.7 and 3.6 classes, respectively (Figure 1).Following resistance to anti-staphylococcal beta-lactams, resistance to macrolides (erythromycin) and fusidanes (fusidic acid) was most frequently observed (Table 2).Worth noticing is the large proportion of S. hominis isolates that expressed no phenotypic resistance to cefoxitin and amoxicillin clavulanic acid.The resistome analysis supported the phenotypical resistance profile, displaying a high prevalence of genes conferring resistance to erythromycin and fusidic acid among the MRCoNS and MRM (Table 3) and further confirmed that all sequenced MRCoNS and MRM possessed the mecA gene.Despite the presence of mecA, we observed a variable phenotypic expression of resistance to beta-lactams among the isolates (Table 2).This was particularly evident for S. hominis, with 1/11 isolates being phenotypically susceptible to oxacillin and 5/11 isolates being susceptible to cefoxitin.Furthermore, we observed phenotypic susceptibility to amoxicillin-clavulanic acid among six MRCoNS and MRM species: S. cohnii ssp.cohnii (2/3), S. epidermidis (8/12), S. haemolyticus (2/16), S. hominis (6/11), S. warneri (1/5) and M. vitulinus (n = 1).In addition to mecA, the beta-lactamase-encoding blaZ gene was present in 16 of the 20 amoxicillin-clavulanic-acid-susceptible isolates.

SCCmec Cassettes and Sequence Types
Table 4 shows the predicted SCCmec cassettes based on detected genes and best homology in the sequenced MRCoNS and MRM isolates.For four isolates (one S. epidermidis, one S. haemolyticus and two S. hominis), the SCCmec prediction based on genes deviated from the prediction based on homology.In these cases, the SCCmec cassettes were reported based on the prediction of the genes.The S. epidermidis isolates were assigned three types (II (2a), III (3A) and IV (2b)) in addition to one non-typeable (NT) isolate.In just one household, the same SCCmec cassette was predicted in isolates of different species: an S. epidermidis and an S. saprophyticus isolate of type III (3A).The SCCmec elements of a substantial number of isolates (43/75) were non-typeable.Eleven of the fourteen NT S. haemolyticus isolates showed best homology with SCCmec type V but missed either the ccrC1 or mec class C2, or both.Three of the NT S. saprophyticus isolates had the best predicted homology to SCCmec III (3A) but missed either ccrA3 or both ccrA3 and ccrB3, while two additional S. saprophyticus isolates shared the best homology with SCCmec I (1B) but missed ccrA1 and ccrB1.The sequence types (STs) were predicted by in silico MLST for 25 of the 39 S. epidermidis, S. haemolyticus and S. hominis isolates.The 12 S. epidermidis isolates were assigned to 10 different STs (5,35,57,81,130,218,224,332, 640 (n = 2) and 679) in addition to 1 nontypeable isolate.The two MRSE ST640s were from different households.Nine of sixteen S. haemolyticus were typed to seven STs (1, 3 (n = 2), 30, 42, 49, 52 (n = 2) and 56).Isolates with identical STs were from different households.The remaining S. haemolyticus isolates had combinations of allelic profiles not reported earlier.This was also the case for 6 of the 11 S. hominis isolates, while the remaining 5 were typed to two different STs (1 and 18), all of which were from different households.

Household Analysis of Human and Pet Isolates
In households with infected pets (C, F and H), the dogs and owners concurrently tested positive for MRSE (Table 1).In household H, the dog, owner and home environment tested positive for MRSE ST640 with identical susceptibility profiles, resistance genes and SCCmec elements [12].In contrast, the dog and owner in household C carried two different STs (MRSE ST679 and ST130), presenting different resistance genes and SCCmec types (Supplementary Materials Table S1).MRSE isolates with identical susceptibility profiles to the dog isolate were detected in the bathroom and kitchen.However, no isolates similar to the MRSE found on the owner were recovered from the home environment.We observed similar findings in household F, in which the owner and dog carried MRSE with different STs (ST218 and ST5) and SCCmec cassettes (III (3A) and IVa (2B)).Contrary to household C, we recovered only isolates with identical susceptibility profiles to the human isolate from the home environment.In household D, MR S. haemolyticus ST42 and ST1 were recovered from the owner, while the dog tested negative for these.Only the MR S. haemolyticus ST42 was detected in the home environment.
Two owners of healthy pets carried MRCoNS.The first owner tested positive for a MR S. warneri possessing a SCCmec cassette V ( 5C2 and 5).An MR S. warneri with an identical SCCmec cassette was recovered from the household's kitchen.The second owner carried an MR S. haemolyticus with an NT SCCmec cassette and ST.An MR S. haemolyticus ST52 with different resistance genes was recovered from the home environment.

Virulence Genes
The hits in the virulence gene database are presented in Table S2, and the respective Ha scores for each virulence gene are shown.The exfoliative toxin-encoding gene etc was detected in all sequenced isolates.Furthermore, we observed a high frequency of phenol soluble modulin-encoding genes, thermonuclease-encoding nuc genes and siderophoreencoding genes in most MRCoNS isolates.Overall, the MRCoNS and MRM showed a high degree of species specificity in the virulence gene analysis, apart from S. haemolyticus and S. hominis, which clustered together (Figure 2).We detected a high occurrence of genes involved in adherence in the S. epidermidis isolates (atl, ebh and sdr genes).In addition, we observed two subpopulations of S. epidermidis isolates based on the presence of Type VII secretion-associated genes (Table S2).The tendency of two subpopulations was also evident among the S. haemolyticus isolates.Based on Ha scores of several capsular polysaccharide synthesis enzymes involved in immune evasion (cap genes), the group with high Ha scores consisted of six isolates, of which three were recovered from humans and two were of environmental origin.Isolates with low Ha scores for cap genes were solely detected in the environment.

Mobilome Analysis
The four most prevalent species found in the households, S. saprophyticus, S. epidermidis, S. haemolyticus and S. hominis, were included in the mobilome analysis.Three common gene clusters encoding an IS6 family transposase, the competence protein ComGC and an uncharacterized SPBc2 prophage-derived protein YoqJ (annotated "Common" in Subgrouping based on the Ha scores of cap genes was also observed among the S. hominis and S. saprophyticus isolates.The MRM had, in general, few hits with high Ha scores, apart from the etc gene and the capO and capP genes.

Mobilome Analysis
The four most prevalent species found in the households, S. saprophyticus, S. epidermidis, S. haemolyticus and S. hominis, were included in the mobilome analysis.Three common gene clusters encoding an IS6 family transposase, the competence protein ComGC and an uncharacterized SPBc2 prophage-derived protein YoqJ (annotated "Common" in Figure 3) were identified in all the isolates.Otherwise, the mobilomes were mainly species-specific (Figure 3).We observed a few examples of similar gene sequences in different species at the household level.For instance, S. haemolyticus and S. saprophyticus isolates from the same household carried phage major capsid protein-encoding genes and phage portal protein-encoding genes with 76.8% and 80% amino acid identities, respectively.In addition, a site-specific tyrosine-type recombinase/integrase was shared by S. epidermidis and S. saprophyticus in two households, and an IS256-like transposase was shared by S. haemolyticus and S. hominis in two other homes.

Discussion
MRCoNS are opportunistic pathogens prevalent in hospital environments, often due to their hardy nature, ability to form biofilms and resistance to antimicrobials [13].The newly described genus Mammaliicoccus shares many properties with staphylococci, like habitat and methicillin resistance [14].In a former study of the dissemination of clinical MRS in households with infected pets [12], we detected a broad range of MRCoNS and MRM in the home environments and from pets and their owners.To follow up on this observation, we decided to screen different categories of households for the presence of MRCoNS and MRM.To our surprise, MRCoNS and MRM were nearly ubiquitous in the home environments regardless of the presence of pets or health status.The finding of S. epidermidis, S. haemolyticus, S. hominis and S. saprophyticus as the predominant species in the households is reasonable since these species are known as skin commensals in hu-

Discussion
MRCoNS are opportunistic pathogens prevalent in hospital environments, often due to their hardy nature, ability to form biofilms and resistance to antimicrobials [13].The newly described genus Mammaliicoccus shares many properties with staphylococci, like habitat and methicillin resistance [14].In a former study of the dissemination of clinical MRS in households with infected pets [12], we detected a broad range of MRCoNS and MRM in the home environments and from pets and their owners.To follow up on this observation, we decided to screen different categories of households for the presence of MRCoNS and MRM.To our surprise, MRCoNS and MRM were nearly ubiquitous in the home environments regardless of the presence of pets or health status.The finding of S. epidermidis, S. haemolyticus, S. hominis and S. saprophyticus as the predominant species in the households is reasonable since these species are known as skin commensals in humans.However, they can also cause infections, and the frequent occurrence of methicillin-resistant isolates in home environments is noteworthy.To our knowledge, the home environment has previously not been described as a reservoir for MRCoNS and MRM.
The skin, skin glands and mucous membranes of mammals are considered the main habitats for CoNS [15].However, in most of the households studied, both the human and the pet tested negative for MRCoNS/MRM while the bacteria were present in the home environment.The absence of MRCoNS/MRM in humans and pets may reflect that the sampling sites in the humans and pets were not optimal for detecting some of the CoNS/mammaliicoccal species.For instance, S. saprophyticus is a frequent colonizer of the perineal region, rectum and urethra in humans, and S. hominis and S. haemolyticus are often isolated from axillae and pubic areas high in apocrine glands [1].These sites were not included in the sampling procedures.On the other hand, S. epidermidis is a common human, canine and feline nasal mucosa colonizer [16,17].Therefore, we find it peculiar that we identified relatively few carriers of MRSE, considering that MRSE was present in around one-third of the households.An explanation may be that we only sampled one human member in each household, thus missing possible carriers of the MRCoNS and MRM.Another factor contributing to the high number of MRCoNS/MRS in the households could be that the bacteria had been introduced via visitors, soil or other external sources.
Carriage of MRCoNS in pets was exclusively found in infected dogs.The owners of the three MRSE-positive dogs all tested positive for MRSE.Interestingly, the isolates from the dogs and owners differed in two of the cases, indicating a diversity of MRCoNS not only between households but also within the household.Moreover, we observed that homes with infected pets had a large diversity of MRCoNS species recovered from the home environment.Five of the eight dogs in this group had been treated with beta-lactam antimicrobials within the past three months before sampling, two of which had undergone antimicrobial treatment several times during the past year.The carriership and the diversity may reflect the MRCoNS's and MRM's competitive advantage when exposed to beta-lactam antimicrobials.Furthermore, five dogs in this group had been hospitalized within the past twelve months, and two owners were human health care workers.Hence, it is not unlikely that the pets or owners have been exposed to MRCoNS/MRM in these environments and transmitted them further to their home environment.Still, MRCoNS and MRM were present in many households where neither humans nor dogs had been in contact with health care facilities, again emphasizing that MRCoNS and MRM are indeed found outside clinical environments.
The phenotypic resistance analysis revealed that just over 50% of the MRM and MRCoNS were multidrug-resistant.This is consistent with previous reports on CoNS and MRCoNS in non-clinical settings [9,18].Mobile genetic elements play a central role in spreading ARGs among bacteria [19].Considering that MRCoNS constitute reservoirs for ARGs, we conducted a mobilome analysis primarily to investigate whether the most prevalent species in the households had mobile genetic elements in common, which could indicate genetic exchange at the household level.Nonetheless, the detected mobile genetic elements displayed mainly a species-specific profile rather than a household-related pattern.This could indicate that mobile genetic elements are not easily transmitted between different staphylococcal species.However, it must be emphasized that this analysis is based on short-read data.To gain further insight in the ARGs' location relative to the mobile genetic elements, it would be necessary to combine short-read and long-read data.
The inconsistent phenotypic expression of resistance to cefoxitin among the MRCoNS isolates was noteworthy.EUCAST and CLSI operate with different zone diameters when assessing cefoxitin resistance.By following the CLSI breakpoints rather than the EUCAST breakpoints, eight of the ten cefoxitin susceptible CoNS isolates would have been classified as resistant.On the other hand, two of the MRSE isolates would have been reported susceptible to cefoxitin.According to the EUCAST guidelines, cefoxitin should be used when screening for methicillin resistance in CoNS [20].However, CLSI emphasizes that the cefoxitin disk diffusion test may not perform reliably in detecting methicillin resistance for all CoNS species (e.g., S. haemolyticus) [21].Although cefoxitin is the recommended agent for most CoNS when screening for methicillin resistance, our results show that oxacillin is more reliable than cefoxitin for the purpose.
MRS are considered resistant to most beta-lactam agents, i.e., penicillins, beta-lactam combination agents and cephems, except for ceftaroline [22,23].However, we observed a high frequency (20/75) of amoxicillin-clavulanic-acid-susceptible isolates.Sixteen of the susceptible isolates carried blaZ, which encodes a beta-lactamase that inactivates amoxicillin.Admittedly, this could be due to the lack of official breakpoints for amoxicillin-clavulanic acid disk diffusion.Still, six of these isolates were susceptible to cefoxitin, thus demonstrating that even if the mecA gene is present, it is not necessarily expressed towards cefoxitin and amoxicillin-clavulanic acid in vitro.
We could not predict STs or SCCmec cassettes for most MRCoNS/MRM isolates.The pubMLST database only contains data for S. epidermidis, S. haemolyticus, S. hominis and S. chromogenes, and the missing ST identifications may be due to a lack of characterized environmental genomes in the database.The high proportion of non-typeable SCCmec cassettes is consistent with previous reports [24,25].In many cases, the cassettes shared homology with previously described SCCmec but lacked identifiable ccr genes needed to determine type.This was especially evident for the NT S. haemolyticus cassettes that had the best homology with SCCmec type V.The combination of NT SCCmec elements combined with non-identifiable STs demonstrates the large diversity among the staphylococcal and mammaliicoccal isolates in the home environments and the gaps in knowledge about the epidemiology/ecology of staphylococci from environmental reservoirs.
In general, CoNS and MRM are considered less virulent than S. aureus.Still, CoNS and MRM cause a substantial number of infections, presumably possessing virulence genes enabling them to do so.Virulence genes in CoNS and mammaliicocci are far less studied than the virulence genes of S. aureus.Consequently, we used a database mainly consisting of amino acid sequences from putative and known virulence factors in S. aureus to characterize virulence genes in our MRCoNS and MRM isolates [26].Admittedly, this is not optimal and will cause uncertainty around the hits with low and medium Ha scores.We focused on the highest scores within each species.However, we cannot be certain that hits with lower scores are of limited importance.Overall, the MRCoNS and MRM displayed species-specific virulence gene patterns, apart from the ubiquitous etc gene.The virulence gene patterns revealed subgroups within the S. epidermidis, S. haemolyticus, S. hominis and S. saprophyticus isolates based on the presence of type VII secretion-associated genes for the former and cap genes for the three latter.The isolation source seemed to matter for multiple cap genes only in the S. haemolyticus isolates, as all the human isolates were in this subgroup.

Participants
Participants were recruited through social media and small animal clinics in Oslo and the surrounding areas.All participants signed individual consent forms and completed questionnaires regarding their professions, antimicrobial consumption and hospital admissions within the past 12 months.Thirty-three households participated in the study.Of these, eight were households with dogs diagnosed with an MRS infection; eight were households with clinically healthy dogs; six were households with clinically healthy cats; and eleven were households without pets.The inclusion criteria included cats with MRS infections.However, during the time we recruited participants for the study, no cats with MRS infections were diagnosed in our recruiting clinics.One human and one pet from each home participated in the study.The inclusion criteria for healthy pets were the following: clinically healthy pets without symptoms of infection when examined by a veterinarian.The humans sampled were healthy according to their own statements.

Sampling
The samples were collected in the period from October 2019 to October 2021.The same veterinarian was responsible for sampling all the household environments and the participating pets.Pets diagnosed with an MRS infection were sampled from the infection site, the perineum and the oral mucosa using nylon flocked swabs (Eswab™ 480C, Copan group, Brescia, Italy).These dogs participated parallelly in another study [12].Healthy dogs and cats were sampled from the oral mucosa and perineum.Human participants collected swab samples from their nostrils and throats according to the instructions of the veterinarian present at the time of sampling.The home environments were sampled using cloths (Sodibox ® Swab cloth, Nevez, France) for swabbing of the most relevant areas such as the pets' food bowls and sleeping areas, living room floors, bathrooms (sink faucet and hand towel) and kitchens (kitchen counter, dish towel, cloth and sink faucet).In the households without pets, the three latter locations were sampled.

Culturing and Identification
The samples were cultured as described by Røken et al. [12].Briefly, all samples were enriched overnight in Müller-Hinton (MH) broth supplemented with 6.5% NaCl.Then, 20 µL of MH broth was inoculated on Oxacillin Resistance Screening Agar Base (ORSAB, Oxoid, Basingstoke, UK) supplemented with 2 µg oxacillin and incubated for 24 h at 35 • C. Blue, blue-white and white colonies growing on the ORSAB agar were sub-cultured on 5% bovine blood agar overnight.The species were identified using Matrix-assisted laser desorption/ionization (MALDI-TOF MS) (VITEK ® MS, bioMérieux, Craponne, France).Isolates were tested for the presence of the mecA gene by PCR on a Bio-Rad T100 Thermal cycler (Bio-Rad, Hercules, CA, USA) [27].

Selection of Isolates
One MRCoNS/MRM from each species was included from each sampling location.Based on the phenotypic resistance profiles, species and households, isolates were selected for whole-genome sequencing (WGS).All WGS isolates went through an additional species identification checks using the Microbial Genomes Atlas (MiGA) webserver against the TypeMat database [30] (Table 5).If the species identities differed between the MALDI-TOF and TypeMat databases, we used the TypeMat output.Furthermore, all sequenced matrix values were expressed as Ha scores ranging from 0 to 1 [26].Briefly, the scores were calculated using the following formula: Ha = (pident × length)/(qlen × 100) where "pident" represents the proportion of amino acid sequence identities between the VF query and translated proteins from the bacterial genomes in this study, "length" represents the alignment length of a hit and "qlen" is the length of the query sequence drawn for each VF.
The mobilome analysis was conducted using Anvi'o bioinformatics suite version 7.1.[37].Before the pangenome analysis, we excluded one S. saprophyticus from the dataset due to a high number (>2000) of partial genes.We created the Anvi'o contigs database with the "anvi-gen-contigs-database" program using Prodigal to identify open reading frames [38].The resulting genes were associated with the functions from the NCBI's Clusters of Orthologous Groups (COGs) database [39].The pangenome was computed by the core Anvi'o program "anvi-pan-genome" (default settings), which in turn utilized DIAMOND [40] and MCL [41].Metadata were integrated into the pangenome results with the "anvi-import-misc-data" program.After the pangenome visualization, we extracted all gene clusters annotated with the "Mobilome" COGs category with the "anvi-split" program for further manual inspection.We used COG annotations or an ad hoc protein web-blast search to characterize the gene clusters of the staphylococcal mobilome.

Statistical Analysis
We used one-way ANOVA to compare the number of different MRCoNS and MRM species between the households with infected pets, with healthy pets and without pets.Tukey's honest significant difference (HSD) test was applied to test the pairwise difference between the three household groups.The significance level was set to 0.05.

Conclusions
In conclusion, we have documented that the home environment is a reservoir for MRCoNS and MRM regardless of the type of household and the carrier status of humans and pets.However, homes with infected pets had a larger diversity in MRCoNS and MRM species than households without pets, which might be due to the recent use of antimicrobials and contact with human and veterinary hospitals.The large diversity in SCCmec elements and sequence types among and within the households indicates no clonal spread of specific strains.The limited common virulomes and mobilomes indicate a high degree of species specificity rather than exchanges of genetic elements between species in the home environment.Informed Consent Statement: Written informed consent was obtained from all subjects involved in the study.

Figure 1 .
Figure 1.Number of phenotypic resistance classes in the MRCoNS and MRM isolates, n = 75.The crosses represent the mean number of antimicrobial resistance (AMR) classes, the horizontal lines represent the median number, the boxes represent the quartiles, while the whiskers represent minimum and maximum number or resistance classes.

Figure 1 .
Figure 1.Number of phenotypic resistance classes in the MRCoNS and MRM isolates, n = 75.The crosses represent the mean number of antimicrobial resistance (AMR) classes, the horizontal lines represent the median number, the boxes represent the quartiles, while the whiskers represent minimum and maximum number or resistance classes.

Figure 2 .
Figure 2. Principal component analysis plot of virulence gene similarity in MRCoNS and MRM isolates.The exfoliative toxin gene etc was present in all isolates.The annotations refer to the genes with the highest Ha scores among the isolates.

Figure 2 .
Figure 2. Principal component analysis plot of virulence gene similarity in MRCoNS and MRM isolates.The exfoliative toxin gene etc was present in all isolates.The annotations refer to the genes with the highest Ha scores among the isolates.

Figure 3 .
Figure 3. Mobilome of S. epidermidis, S. haemolyticus, S. hominis and S. saprophyticus, sorted by species.The colors in the "Household" column represent the different households the isolates were isolated from.Gene clusters identified in all isolates are annotated as "Common" and include genes encoding an IS6 family transposase, the competence protein ComGC and an uncharacterized SPBc2 prophage-derived protein YoqJ.

Figure 3 .
Figure 3. Mobilome of S. epidermidis, S. haemolyticus, S. hominis and S. saprophyticus, sorted by species.The colors in the "Household" column represent the different households the isolates were isolated from.Gene clusters identified in all isolates are annotated as "Common" and include genes encoding an IS6 family transposase, the competence protein ComGC and an uncharacterized SPBc2 prophage-derived protein YoqJ.

Funding:
This research was funded by the Faculty of Veterinary medicine, Norwegian University of Life Sciences, and by NORM, grant number 19_05.Institutional Review Board Statement: The study was approved by the Norwegian National Research Ethics Committee (REK Sør-Øst) (Protocol code: 2019/97, 3 April 2019).

Table 3 .
Percentage of the 75 isolates testing positive for antimicrobial resistance genes (ARGs).All numbers are percentages of the number of isolates shown in column 2. A darker shade represents a higher percentage.

Table 4 .
Predicted SCCmec elements based on detected genes in the sequenced MRCoNS and MRM isolates.