Introduction

The members of the Staphylococcus genus are part of the normal microbiota of the nose and skin of humans and animals (including many avian species) (Szczuka et al. 2023). In addition, coagulase-positive staphylococci could occasionally cause clinical diseases mediated by highly potent virulence genes (Aqel et al. 2023). However, not every staphylococcal virulence gene is expressed. Instead, the expression of the genes is usually restricted to times and places and regulated by bacterial factors (Grazul et al. 2023). Over the last years, some coagulase-negative staphylococci (CoNS) species (as S. epidermidis, S. haemolyticus, or S. hominis) emerged as a cause of opportunistic infections such as those in septicaemic children, or in patients with immunosuppression or medical implants, among others (França et al. 2021; Heilmann et al. 2019). Most of other CoNS species are unfrequently implicated in human or animal infections, being often highly susceptible to antimicrobial agents (Merrild et al. 2023; Santoiemma et al. 2020; Argemi et al. 2019). However, there have been sporadic reports of some S. pasteuri causing endocarditis, whereas S. hyicus, S. chromogenes, S. lentus, and S. sciuri are considered etiological agents of exudative epidermitis with zoonotic potentials (Kirk et al. 2022; Kalai et al. 2021; Li et al. 2021). Moreover, S. saprophyticus contracted from contaminated food have long been implicated in urinary tract infections in young teenagers (Lawal et al. 2021a, b). Much more recently, whole-genome sequence data of CoNS species have led to the identification and characterization of numerous putative virulence factors (Argemi et al. 2019). Furthermore, CoNS could acquire clinically relevant and critical antimicrobial resistance (AMR) genes and transmit them across other species and hosts through various mobile genetic elements (mobilome) (Rossi et al. 2020). Specifically, S. haemolyticus has been ranked as the most antibiotic-resistant species among the CoNS (Kranjec et al. 2021). The transferability of AMR genes between different Staphylococcus species has been strongly suggested by the sequence similarity of their associated mobilome, especially plasmids (Souza-Silva et al. 2022).

The mecA gene, its staphylococcal cassette chromosome (SCCmec) carrying element, and the arginine catabolic mobile element (ACME) originated from CoNS were acquired by S. aureus (Shokrollahi et al. 2022). mecC-carrying CoNS have also been reported from many countries but in very low frequencies (Loncaric et al. 2019). Previously thought to be a wildlife MRSA trait, the continuous spread of the SCCmec-bound mecC gene in CoNS highlights their potential role in the evolutionary origin and genetic transfer to MRSA (Abdullahi et al. 2023a).

Most methicillin-resistant CoNS (MRCoNS) strains are often found to be resistant to other non-betalactam antibiotics except for glycopeptides, which have long been utilized in the treatment of staphylococcal infections (Chajęcka-Wierzchowska et al. 2023). As the AMR epidemic keeps expanding, the few methicillin-resistant staphylococcal infections that are treated using oxazolidinones (Gostev et al. 2021) could have promoted the emergence, spread, and persistence of linezolid resistance, as some mechanisms mediated by ARGs (cfr, poxtA, and optrA) are carried by plasmids (Bai et al. 2019; Dortet et al. 2018). However, high-level linezolid resistance could be caused by non-transferable mechanisms mediated by mutations in the 23S rDNA, and in the ribosomal proteins L3, L4, and L22 (Ruiz-Ripa et al. 2021).

The clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated proteins (Cas) are RNA-based adaptive immunity to protect and are utilized by many bacteria against invading mobile genetic elements (MGEs) (Tao et al. 2022a). Hence, the CRISPR–Cas system might be a potential means to prevent the acquisition of plasmid and phage invasion and even horizontal transfer of AMR genes in staphylococci (Murugesan and Varughese 2022). There are two categories of CRISPR-Cas, which are based on their proteins’ structures, constituents, and modes of action (Nishimasu and Nureki 2017). The Class 1 CRISPR-Cas uses multiple protein effector complexes to break down nucleotides and can be subdivided into types -I, -III, and -IV, whereas the Class 2 CRISPR-Cas utilizes single-protein effector complexes to break down nucleotides, of which it is subdivided into types -II, -V, and -VI (Shmakov et al. 2015; Makarova et al. 2015). The types II-CRISPR-Cas systems have largely been studied and have successfully been used to delete antimicrobial resistance genes (ARGs) due to their relatively simple structures (Tao et al. 2022a). Moreover, the Type I CRISPR-Cas systems have been developed and manipulated to eliminate ARGs (Tao et al. 2022a). In this regard, certain CRISPR-Cas system prevents foreign nucleotides (such as plasmids and phages) from evading the bacteria thereby limiting the acquisition of ARGs (Tao et al. 2022a).

The genetic characterization of CoNS is necessary to understand their evolution and source distribution, reservoir hosts, and vectors of AMR transmission. In this regard, certain animal hosts such as the pigs and human workers in pig farm environments are believed to be under high antibiotic pressure and carry staphylococci presenting a high-level multidrug resistance (MDR) phenotype. However, animals in the wildlife may be at low antibiotic pressure as they are rarely exposed to antimicrobial agents (Abdullahi et al. 2021). It is worth mentioning that the ecology and epidemiology of AMR in CoNS could be different from that of S. aureus because the CoNS species could present different and diverse AMR profiles. In this study, the genomic characterization of 26 multidrug resistant-CoNS (resistant to ≥ four classes of antimicrobial agents) previously obtained from the nasal cavity of healthy humans without animal contact, nestling storks, pigs and pig owners, as well as dogs and their owners from Spain were performed by Illumina technology.

Materials and methods

Coagulase-negative staphylococci strains in this study

A total of 516 non-repetitive CoNS strains were obtained in previous studies (Abdullahi et al. 2023a; b; c; 2024a) from nasal samples of healthy animals and healthy humans with different types of animal contact: (a) healthy nestling storks (NS) (268 isolates); (b) healthy pigs (H-P) and pig farmers (H-PF) (75 isolates); (c) healthy dogs (H–D) and dog owners (H-DO) (130 isolates); and (d) healthy humans who had no contact with animals (HH−) (113 isolates). The antimicrobial susceptibility of these isolates was previously determined by disk diffusion tests, and the presence of ARGs by PCR (Abdullahi et al. 2023a; b; c; 2024a). From this collection, 26 CoNS isolates of 10 species (S. borealis, S. saprophyticus, S. sciuri, S. hominis, S. epidermidis, S. pasteuri, S. hyicus, S. simulans, S. haemolyticus, and S. arlette) were selected to be further characterized in the present study by whole genome sequencing (WGS), and they were of the following origins: NS (n = 4); H-P and H-PF (n = 14); H–D and H-DO (n = 4); and HH− (n = 4). The selection criteria of the strains included were as follows: (i) CoNS that presented an MDR phenotype for four or more classes of antibiotics, selecting one species each per host carrying this resistance phenotype; and (ii) MDR-CoNS isolates with similar AMR genes detected from humans and animals in the same ecological niche to detect potential transmission events.

The study protocols in which these isolates were recovered were reviewed and approved by the ethical research committees of the University of Zaragoza, the University of La Rioja and the University of Castilla La Mancha (Spain).

Genome sequencing, assembly, and phylogenetic analyses

Whole genome sequencing of the selected 26 CoNS isolates was carried out on the Illumina NextSeq platform. The MagNA Pure 96 DNA Multi-Sample Kit (Life Technologies, Carlsbad, CA, USA, 4413021) was used to extract genomic DNA according to instructions provided by the manufacturers. The Qubit 1X dsDNA HS Assay Kit (Thermo Fisher Scientific, Scoresby, VIC, Australia) was used for DNA quantification, while Sequencing libraries were prepared using the Illumina Nextera XT DNA Library Preparation Kit (Illumina, San Diego, CA, USA, FC-131–1096) and sequenced on the NextSeq 500 platform (Illumina, San Diego, CA, USA) using a 300-cycle kit to obtained paired-end 150 bp reads, as previously described (Abdullahi et al. 2023d).

All the genomes analyzed in this study were de novo assembled using SPAdes (v.3.15.5), performing the in silico typing with the settings of a minimum of 90% coverage and 80% identity. First, core-genome single nucleotide polymorphisms (SNPs) between the eight S. epidermidis strains in this study were detected with the NASP pipeline v.1.0.0 (Sahl et al. 2016) after they were mapped together with a reference strain ATCC 14990 (GenBank accession number: GCA_006094375) and 31 previously published S. epidermidis genomes from different countries with similar genetic lineages from the PubMLST database (https://pubmlst.org/bigsdb?db=pubmlst_sepidermidis_strains&page=query&genomes=1) (identification [id] numbers: 32110, 32113, 32116, 41749, 42109, 43340, 43421, 43426, 43427, 43436, 43455, 43466, 43518, 43568, 43636, 43643, 43656, 43697, 43720, 43770, 43771, 43774, 43786, 43800, 43816, 43823, 43921, 44294, 44298, 44496, 44521) to obtain an S. epidermidis phylogenetic trees. GATK (v.4.2.2) was used to call SNPs and excluded positions featuring < 90% unambiguous variant calls and < 10 depth. IQ-TREE (v.2.1.2) was used to construct the phylogenetic trees using ModelFinder with 100 bootstraps. The graphical data was added to the phylogenies with iTOL v.6.6 (Letunic and Bork 2021). To determine the relatedness of the S. saprophyticus from a pig and pig farmer, we used a web-based CSI phylogeny database to obtain the SNPs by mapping the genomes to a reference S. saprophyticus ATCC 15305 (GenBank accession no. AP008934.1) with the default parameter, except for the minimum distance between SNPs which was disabled. Also, the SNPs of the S. borealis from four pigs were determined by comparing them with 16 additional publicly available genomes of S. borealis strains available from NCBI (GenBank accession numbers: GCA_030362885, GCA_030362875, GCA_003580835, GCA_003580835, GCA_034103225, GCA_024580895, GCA_030501495, GCA_035788295, GCA_035791815, GCA_035791575, GCA_013345165, GCA_009735325, GCA_013345185, GCA_013345175, GCA_013345205, GCA_013345195) mapped with a reference strain 7067_4#69 (GenBank accession number: GCA_001224225.1) by using the web-based CSI phylogeny database following settings similar to the ones used for S. saprophyticus.

Annotation, typing, and in silico analysis of the CoNS genomes

The sequence types (STs) were determined with MLST v.2.16 (Jolley et al. 2018). Virulence factors, plasmid replicons, and antimicrobial resistance genes were identified using ABRicate v.0.9.0 and the respective databases VFDB, Plasmidfinder, and Resfinder databases from the Center for Genomic Epidemiology. Mutations associated with AMR were identified using ResFinder v4.1 (Bortolaia et al. 2020) and PointFinder (Zankari et al. 2017). Biocide and heavy metal resistance genes were identified using BACMET (Pal et al. 2014). Phaster was used to identify all prophage elements (Arndt et al. 2016). The SCCmec types were assigned using SCCmecFinder 1.2 (https://cge.food.dtu.dk/services/SCCmecFinder/). The genetic environment of the ermT, cfr, and mecC genes was illustrated in comparison with the reference strains using the EasyFig software.

Determination of the CRISPR-Cas system of coagulase-negative staphylococci

The CrisprCasFinder (https://crisprcas.i2bc.paris-saclay.fr/) was used to identify the numbers of CRISPR, Cas proteins, and spacers of all the MDR-CoNS (Couvin et al. 2018). Specifically, the size of the flanking region and other parameters were set to default values. Moreover, three CoNS strains that contained larger sequences than CrisprCasFinder could handle were analyzed by the CRISPRCasMeta (https://crisprcas.i2bc.paris-saclay.fr/CrisprCasMeta/Index) applying all the default settings.

Genome availability

All the raw genome reads generated from this study have been deposited at the European Nucleotide Archive under Study Accession no. PRJNA1023081.

Statistical analysis

Data generated from this study reported frequencies and were presented in tables. Univariate logistic regression was to compute the odd ratio (OR) at a 95% confidence interval (95%CI) between the presence of MDR-CoNS genomes, and various mobilome with the ecological niches. Significant association at p < 0.05 was considered.

Results and discussion

CoNS have long been considered reservoirs of ARGs; however, very few genomic studies have elucidated the influence of different ecological niches on the levels of ARGs and their MGEs. Moreover, there is a paucity of phylogenomic data on the transmission pathways of CoNS species and their ARGs between humans and animals.

Resistome, mobilome, and relatedness of the 26 CoNS analyzed in this study

The phenotypes of resistance of the 26 CoNS isolates characterized in this study are shown in Supplementary Table S1, and their resistome, virulome, genetic lineages, and mobile genetic elements are represented in Table 1. As identified, all the isolates presented an MDR phenotype to 4 to 9 classes of antimicrobial agents. In this regard, the CoNS isolates with the least were those from nestling storks and with the highest those from pigs and pig farmers (Table 1). The mechanisms of resistance to most of the antibiotics were mediated by combinations of multiple antibiotic resistance genes (ARGs).

Table 1 Antimicrobial resistome, and metal and biocide resistance determinants in the 26 MDR-CoNS isolates and their associated mobile genetic elements
Full size table

Concerning the genetic lineages of S. epidermidis, STs belonging to the clonal complexes CC2 and CC5 were identified. For the S. haemolyticus, the three isolates were of the lineage ST30 and ST68. Moreover, S. sciuri-ST212 and S. hominis-ST33 were detected. The genetic lineages of other species were not identified as no MLST scheme has been developed and validated for them yet.

Multiple virulence genes that mediate host immune evasion, adhesion, and haemolysis among others were identified (Table 1). It is important to remark on the detection of an S. epidermidis strain that carries the enterotoxin genes, sec and sel.

Relatedness of the coagulase-negative staphylococci strains

The phylogenetic analysis identified clusters of related strains of various CoNS species with other countries. Specifically, the cfr-carrying S. epidermidis-ST16 strain (X5485) was related to an S. epidermidis-ST16 strain from a human blood sample (SNP = 70) from Canada (id-41749). The S. epidermidis-ST35 from a dog owner is related to an human strain from Portugal (id-43340) (SNP = 90). Moreover, the S. epidermidis-ST297 from a healthy human in our study is related to three human strains from Germany, the UK, and Switzerland (SNP < 80) (Fig. 1). Also, the S. epidermidis-ST173 strain (X9066) was related to an animal strain in Thailand (id-44496) (SNP = 76) (Fig. 1). Furthermore, the sec/sel carrying-S. epidermidis-ST595 strain (X4430) is not related (> 3000 SNPs) to previously described sec/sel-carrying strains from Portugal and Italy (ID-43921, Id-43401). It is important to remark that despite the few SNP differences (< 85 SNPs) between some strains from Portugal (id-43340) and Canada (id-41749) with our two linezolid-resistant strains (X5485 and X6049b), none of them from the two countries was linezolid-resistant. This suggests that our strains might have acquired the gene and mutation following antibiotic pressure in the livestock niche. These findings highlight the international circulation of related S. epidermidis strains between humans and animals as confirmed by the phylogenetic analysis (SNP < 100) (Fig. 1, Supplementary Table S2).

Fig. 1
figure 1

Phylogenetic tree based on core genome SNP analysis of eight S. epidermidis isolates of this study with 31 publicly available genomes with similar lineages

Full size image

Aside from the S. epidermidis strains, we found related S. borealis (SNP < 10) between pigs (Fig. 2, Supplementary Table S3). However, the relatively high SNP (n = 346) between the S. saprophyticus strains from a pig and pig farmer on the same farm suggests that they are unlikely related (Supplementary Table S4).

Fig. 2
figure 2

SNIP-based phylogenetic tree of the four S. borealis isolates of this study mapped with all the 16 publicly available genomes

Full size image

Mobilome-bound antimicrobial resistance in coagulase-negative staphylococci

Generally, MDR was the criteria of selection, and so, all isolates need to be resistant to at least 4 classes of antibiotics (Supplementary Table S1). In this sense, the resistome profile of the strains was mainly to beta-lactam, macrolide-lincosamide-streptogramin-B (MLSB), tetracyclines, sulfamethoxazole-trimethoprim, aminoglycosides, and phenicols as previously detected by PCR and presented in our previous study (Abdullahi et al. 2023a). MDR to a maximum of 5 antibiotic classes was found from the previous study on CoNS strains from nestling storks (NS) (Abdullahi et al. 2023a). For the healthy dogs (H–D), healthy dog owners (H-DO), pigs (H-P), and pig farmers (H-PF), resistance to a maximum of nine antibiotic classes was obtained. In the case of isolates of  healthy humans without animal contact (HH−) resistance to a maximum of seven antibiotic classes were obtained. Resistance genes already detected by PCR were found (Abdullahi et al. 2023a, b, c, 2024a, b, c), but others not previously identified such as lsaB, lsaE, vgaA(LC), vga(E), bleO, str, and dfrC were identified (Table 1). Concerning plasmid bound-AMR genes, all the MRCoNS from pigs and pig farmers had mecA genes carried by SCCmec type-Vc except the two S. saprophyticus strains that had mecA in SCCmec type-IVb. The predominance of the SCCmec type-Vc in these isolates strongly suggests the interspecies transmission of mecA gene by the same SCCmec element. Thus, it has been speculated that the SCCmec type-Vc in LA-MRSA originated from MR-CoNS carried in the same ecological niche (such as nostrils in this case) (Matuszewska et al. 2022). This is corroborated by the SCCmec types carried by LA-MRSA-CC398 isolates from the same pigs and pig farmers (Abdullahi et al. 2024b). Whereas the SCCmec type-IV (a common community SCCmec type in MRSA) in S. saprophyticus from the pig and pig farmer suggests community-associated strains brought to the pig farm. Moreover, S. saprophyticus is known to cause uncomplicated urinary tract infections in the community (Lawal et al. 2021a, b). In nestling storks, the MDR-S. arlettae and S. epidermidis isolates were methicillin-susceptible, whereas the MR-S. haemolyticus carried mecA gene located in SCCmec type-V. Moreover, the MR-S. lentus carried the mecA/mecC genes located in SCCmec-mecC hybrid. It is important to remark that the mecA gene might be intrinsic in S. lentus (Saber et al. 2017). Of the MR-CoNS strains from dogs/owners and healthy humans, both the classical hospital and community-associated SCCmec elements were detected. This shows the SCCmec type in these hosts aside from pigs/farmers has no categorical predilection.

Concerning genes that encode MLSB resistance, the ermB, ermC, erm45, vgaA(LC), and vga(E) genes were identified in single or in combination among over 50% of the CoNS isolates (Table 1). Specifically, the ermC gene in most of the ermC-positive strains was located in small plasmids that were 99.8% identical to those previously described in an S. aureus isolate, plasmid pMSA16 (GenBank accession number: JQ246438.1) and in an S. saprophyticus isolate, pSES22 (GenBank accession number: AM159501.1). Moreover, it is important to remark on the detection of the unusual ermT gene in two staphylococcal species: S. borealis (carried by plasmid repUS18) and S. hyicus (with no associated plasmid). The ermT gene is not a common mechanism for MLSB resistance in CoNS. It appears ermT gene is silently evolving in CoNS causing a constitutive MLSB resistance phenotype.

Tetracycline resistance was found in all the pigs’ and pig farmers’ isolates and mediated by different combinations of genes. In this regard, tet(K), tet(L), tet(M), and tet(45) were found in most of the pigs/pig farmers isolates (Table 1). Moreover, the tet(L) gene was found in one S. epidermidis (X6293) isolate from a dog owner. It is important to mention that the tet(L) gene was located in plasmid rep22 in all the pigs and farmers isolates. However, no MGE was detected to be associated with all the tet(M) and tet(45) carrying MDR-CoNS strains. The absence of MGE associated with tet(M) gene in the MDR-CoNS is different from the transposon-linked tet(M) gene found in the S. aureus strains (Abdullahi et al. 2024b), and this is subject to further investigations to unravel the reasons for the differences. Perhaps, this plasmid rep22-located tet(L) gene is coincidentally predominant in the pig farm niche. The tet(K) gene in most of the CoNS isolates was located in rep7a while in only one strain (X6049b) was located in plasmid rep20, and three others from pigs were not associated with this plasmid replicon (Table 1). It is important to highlight that all the plasmid bound-tet(L) genes were linked with the dfrK gene in similar plasmid repUS12. A similar observation was reported in an MRSA-CC398 strain from a pig (GenBank accession number: FM207105). However, tet(L) was not found to be located in any plasmid in one of the S. hyicus strains from a pig (X5069) carrying a Tn559-bound dfrK. This denotes the difference in the pattern of acquisition of tet(L) gene and potential inter-species transmission in CoNS and S. aureus in a pig farm setting.

Aside from these plasmid-bound AMR genes, other genes that mediate resistance to aminoglycosides (such as ant4′ and bleO, located in plasmid repUS12), clindamycin (e.g., lnuA, in rep22), and sulfamethoxazole-trimethoprim (e.g., dfrK, in repUS12 and rep22) were occasionally identified. In some instances, these AMR genes were not associated with any plasmid. We cannot categorically infer the reason some AMR genes are located in plasmids in some CoNS strains while in the bacterial chromosome of others. It could be that the bacteria lost the plasmids during horizontal transfer but the recipient bacteria retained the AMR genes (Dimitriu 2022). The similarity in plasmids that carry many AMR genes in all the CoNS strains demonstrates their impact on bacterial fitness for survival and capability to transfer these resistant genes intra-species (the same species), interspecies, and between different hosts. The transferability of AMR genes between different Staphylococcus species has been strongly suggested by the sequence similarity of their associated mobilome, especially plasmids (Souza-Silva et al. 2022). Moreover, some plasmids appeared to carry multiple AMR genes from different classes of antibiotics (such as repUS12 and rep22).

Aside from these mobilome-bound AMR genes, the aminoglycoside and MLSB resistance genes ant9′ and ermA were also carried by Tn554 in an S. epidermidis strain from a dog owner (X3617). Similar findings (i.e., Tn544-linked ant9′ and ermA genes) was reported but in a different CoNS species, S. lugdunensis (Chang et al. 2019, 2021). This suggests potential inter-staphylococcal species transmission of the ARGs. Chloramphenicol resistance is an important phenotypic marker for linezolid resistance, especially in pig farm settings. Chloramphenicol has long been prohibited for the treatment of animal and human infections in Spain. However, florfenicol is still used for livestock. The fexA and fexB genes confer resistance to both florfenicol and chloramphenicol and could be responsible for the frequent co-resistance to chloramphenicol found in CoNS strains from pigs and pig farmers. In this study, only fexA which was carried by Tn554 and Tn558 was identified in four pigs’ strains and this illustrates the influence of pig farm setting on the persistence of phenicol resistance genes especially the fexA that could be carried by two different transposons. Of clinical and public health concern is that other critical AMR genes such as those that mediate transferable linezolid resistance could be co-selected. In this regard, two cfr-carrying S. epidermidis and S. saprophyticus isolates from a pig previously identified were identified (Abdullahi et al. 2023b). Upon genomic characterization, the cfr gene in S. saprophyticus strain was located in a plasmid rep15, while in S. epidermidis was not associated with any plasmid but was flanked by ISSau9 (Table 1).

Antimicrobial resistance mediated by chromosomal point mutations

Twelve of the 26 CoNS isolates analyzed (46.2%) carried one or more mechanisms of ciprofloxacin resistance mediated by DNA topoisomerase IV point mutations at GyrA (S84L) and DNA gyrase at GrlA (S80F) (Table 1). Interestingly was the detection of 21 non-synchronous mutations on the GyrA on one S. simulans strain from a healthy pig (X5777) (Table 1). A major difference in the ciprofloxacin resistance rate was observed between the isolates from the pigs and pig farmers and those of the other hosts: 7 (50%) of the CoNS isolates from pigs and pig farmers showed one or more of the mutations on quinolone-resistance-determining region, whereas three CoNS isolates from healthy humans (S. epidermidis and S. haemolyticus) and one S. arlettae isolate from a nestling stork exhibited this mutation (Table 1). These highlight the influence of pig farm antibiotic pressure on ciprofloxacin resistance on the CoNS isolates. Moreover, mutation-mediated AMR related to linezolid resistance was found in ribosomal proteins L3, L4, and L22 in a S. epidermidis-ST15 strain from a dog owner, as previously identified by PCR-sequencing (Abdullahi et al. 2023c).

Plasmid-bound biocide and metal resistance among the CoNS isolates

Concerning biocide resistance, various plasmid bound-biocide resistance genes (such as qacA [rep20, rep22], qacC [rep13], qacG [rep21], and qacJ [rep21]) were detected in 34.6% of the 26 MDR-CoNS isolates characterized in this study. The acquisition of qacG gene carried on plasmid rep21 was previously found in the majority of S. aureus strains from our previous study (Abdullahi et al. 2024b). This plasmid-bound resistance to quaternary ammonium compounds could facilitate the persistence and co-selection of MDR in CoNS, as these genes make it very difficult for their eradication (Seier-Petersen et al. 2015). In addition, smr gene that encodes resistance against cationic antiseptic compounds (Damavandi et al. 2017) was identified in four strains (Table 1). Two genes that encode for cadmium and zinc resistance (cadD and czrC, respectively) were identified, of which czrC predominated (42.3%).

Metal resistance has previously been hypothesized to co-select for AMR and they are often linked to SCCmec elements (Lawal et al. 2021b) and plasmids in LA-MRSA, S. epidermidis, S. saprophyticus, S. haemolyticus, etc. (Lawal et al. 2021a; Argudín and Butaye 2016; Schijffelen et al. 2010). Specifically, determinants of copper (copA) and zinc (czrC) resistance were widespread among our MR-CoNS isolates of the pigs and pig farmers, but absent or minimal in other hosts. This could denote the potential selection of resistance to these metals due to their persistence in pig farm settings (e.g., in pig feed) especially when plasmid-linked (Huang et al. 2021; Slifierz et al. 2015). Moreover, the cadmium resistance gene (cadD) suggests the involvement of environmental pollution where these staphylococci originated (Rebelo et al. 2021).

Genetic environment of the unusual antimicrobial resistance gene in CoNS strains

The in silico analysis of the ermT sequences of three CoNS strains of two different species (S. borealis and S. hyicus) from healthy pigs revealed major differences in their genetic environment (Fig. 3). The ermT gene is in the opposite direction respect to ant9′ and both are located in plasmid repUS18 in S. borealis strain. However, the ermT gene in the other two S. hyicus strains (X5447 and X5069) is not associated with any plasmid, perhaps it is chromosomally located. The ermT gene in the three strains produces an erythromycin-clindamycin resistance phenotype of constitutive character and highlights their evolution in MLSB resistance among CoNS.

Fig. 3
figure 3

Genetic environment of ermT gene of three CoNS isolates of this study (X5447, X5069, and X5409) in comparison with those of four reference strains

Full size image

The in silico analysis of S. lentus strain (X4638) showed that it carried a hybrid SCCmec-mecC, which is 100% similar to an S. sciuri strain from bovine infection in the UK (Harrison et al. 2014). Specifically, the SCCmec-mecC hybrid consisted of a class C1 mec complex located immediately downstream of a SCCmec type-VII element. Moreover, the cadA, cadC, and cadD genes are included in the system (Fig. 4). It has been previously described that most CoNS that carry the mecC gene are within a hybrid SCCmec element comprised of mecA included in SCCmec type VII and a mecC region consisting of the class E mec complex (de Moura et al. 2023; Belhout et al. 2023; Paterson 2020). However, blaZ-SCCmec XI was initially found to be associated with mecC in our S. lentus X4630 strain by PCR and amplicon sequencing by Sanger (Abdullahi et al. 2023a). Following WGS, the mecC gene of the S. lentus X4638 strain was noted to be quite different from the classical SCCmec type XI that was first demonstrated in S. aureusLGA251 (accession number FR821779). The reason for this variation is subject of further analysis. But, it could be hypothesized that a recombination event took place between the SCCmec type III (intrinsic for most MR-S. lentus) of the mecA gene and SCCmec type XI of the mecC to produce the SCCmec-mecC hybrid (i.e., the SCCmec type VII). In this regard, there is a need for caution in the use of PCR-based assays for the detection of SCCmec types in mecC-carrying non-aureus staphylococci. To the best of our knowledge, this report represents the first description of a mecC in an S. lentus strain from a wild bird. This suggests the expansion of this mechanism of methicillin resistance in CoNS across various ecological niches including wild animals, which were previously proposed to be the major reservoirs of the mecC gene in S. aureus (Abdullahi et al. 2021).

Fig. 4
figure 4

The environment of the mecC gene of S. lentus (X4638) in comparison with previously described mecC-carrying coagulase-negative staphylococci and the S. aureusLGA251 strain

Full size image

The linezolid resistance mechanism mediated by plasmid pURX4944 (41.6 Kb) (Fig. 5) carrying the cfr gene located upstream of lsaB was identified in S. saprophyticus X4944 strain and it was 96% identical to the plasmid of a clinical S. epidermidis strain from Italy (GenBank accession number: KR230047.1). Nevertheless, the cfr gene of our S. epidermidis-ST16 strain was not associated with a plasmid but was flanked by IS256 upstream of lsaB (Fig. 6). It has been suggested that the emergence and dissemination of the cfr gene in animals that have never used any of the oxazolidinones might be due to the selective pressure by the high use of florfenicols, lincosamides, tetracyclines, and pleuromutilins in the livestock sector (Gostev et al. 2021).

Fig. 5
figure 5

Circular representation of the plasmid-carrying the cfr gene in S. saprophyticus

Full size image
Fig. 6
figure 6

Environment of the cfr gene of S. epidermidis (X5485) and S. saprophyticus (X4944) in comparison with previously described cfr-carrying coagulase-negative staphylococci and S. aureus

Full size image

Virulome profile of the coagulase-negative staphylococci strains

We investigated the frequency and distribution of virulence genes among the different CoNS isolates from the four hosts. About 96.1% of the MDR-CoNS strains harbored one or more of diverse adherence, exoenzymes, haemolysin, or immune evasion genes (Table 2). Enterotoxins constitute important virulence determinants of the genus Staphylococcus, of which they are rarely detected in CoNS (França et al. 2021). Enterotoxins are the most implicated in food-borne gastroenteritis (Grispoldi et al. 2021). Moreover, other virulence factors could be responsible for a range of staphylococcal-related infections that are rarely detected in non-aureus staphylococci (Nanoukon et al. 2018). However, it is important to highlight the detection of a sec- and sel-carrying S. epidermidis strain of the lineage ST595. Similar studies have previously reported these virulence genes and their associated pathogenicity islands in S. epidermidis (Lin et al. 2021; Nasaj et al. 2020; Banaszkiewicz et al. 2019). Moreover, it has been suggested that only S. epidermidis from animals or food but not from humans may typically produce S. aureus-related enterotoxins (Podkowik et al. 2016; Veras et al. 2008; Stach et al. 2015; Nanoukon et al. 2018). However, some sec and sel genes have been identified in association with plasmids, phages, and pathogenicity islands. Thus, they can be horizontally transmitted between any host, including humans. It appears that the sec and sel-carrying S. epidermidis from nestling stork are not transferable as they were not associated with a mobile genetic element. Moreover, simultaneous colonization of the nostril by several Staphylococcus spp could promote the transfer of enterotoxin genes from S. aureus to commensal S. epidermidis (Nanoukon et al. 2018).

Table 2 Virulence determinants and prophages in the 26 MDR-CoNS isolates
Full size table

Aside from the toxins, many CoNS harbored genes such as the capB and capC (encode capsules) and adsA, galE, wbtE, wbtP genes that facilitate immune evasion by CoNS (Naushad et al. 2019; Li et al. 2015). Furthermore, the icaABC operon and its icaR were present in five strains (19.2%) (Table 2). This denotes that some of the CoNS species could easily adhere to the mucosa and inanimate surfaces and serve as a fundamental step in colonization and persistence on environmental surfaces and fomites (Idrees et al. 2021).

CRISPR-Cas system distribution among the coagulase-negative staphylococci

Complete CRISPR-Cas system was detected in 19.2% of the CoNS strains, of which cas-1, -2, and -9 predominated in S. borealis (75%). In other species, Cas3-type I CRISPR was identified in two S. epidermidis strains (X6590 and X6049b) from humans. Furthermore, the mecC-carrying S. lentus harbored Cas2-type I and Cas9-type II (Table 3). The low frequency of CRISPR-Cas positive strains identified in our study is closely similar to the 12.3% rate by Rossi et al. (2017) which consisted of mainly class 1 type IIIA and class 2 type IIC systems. Considering that most CRISPR-Cas reduces or eliminates mobile genetic elements such as plasmids, the low frequency of CRISPR-Cas elements among MDR-CoNS isolates in this study could explain the reason why all the strains had ARGs carried by multiple plasmids. However, large-scale genome-based studies including isolates with different profiles of antibiotic resistance are necessary to better understand the roles of the CRISPR-Cas system on AMR genes and their plasmids among S. borealis.

Table 3 CRISPR-Cas system distribution among the CoNS isolates
Full size table

Conclusion

These findings showed that various healthy ecological niches harbor CoNS that presented MDR phenotype mediated by multiple ARGs with several mobile genetic elements with relatively low frequency of intact CRISPR-Cas systems. Furthermore, our findings highlight the potential geographical dissemination of some lineages of CoNS species across various hosts. Collectively, our findings underscore the need to strengthen the genomic epidemiological approach and inclusion of MDR-CoNS from all hosts to adequately control the global fight against AMR and potentially pathogenic ones as identified in the sec- and sel-carrying S. epidermidis.