S. xylosus FFCShyA4 was isolated from the exocarp of avocado (Persea nubigena var. guatemalensis) as described by Angamarca et al. (2023). The strain was cultured in Brain Heart Infusion (BHI) broth (Merck Millipore, MA, USA). Genomic DNA and total RNA were extracted using the Illumina DNA Prep Kit (Illumina Inc., San Diego, CA, USA), following quality control procedures. DNA concentration was measured using a NanoDrop™ 2000 spectrophotometer (Thermo Fisher Scientific, USA), and libraries were prepared according to the manufacturer’s protocol.

Genome sequencing was conducted using the Illumina HiSeq X Ten platform (Macrogen Inc., Seoul, Korea). Libraries were prepared by fragmenting DNA or cDNA samples and ligating 5′ and 3′ adapters, as per the manufacturer’s protocol. Adapter-ligated fragments were PCR-amplified, gel-purified, and loaded onto a flow cell for cluster generation via bridge amplification or ExAmp cluster amplification on patterned flow cells. Sequencing was performed using Illumina SBS technology, a reversible terminator-based method that minimizes incorporation bias and reduces sequence errors, even in repetitive or homopolymer regions, by ensuring natural competition among all four dNTPs. Raw sequencing data were analyzed, and FastQC (v0.11.5) was used to assess quality, base count, GC content, and other metrics. De novo assembly was carried out with SPAdes v3.15.5 using multiple k-mers, and the best assembly was chosen based on parameters like contig number, total contig bases, and N50 (Bankevich et al., 2012). Assembly completeness was evaluated using BUSCO v5.1.3 against the bacterial or eukaryotic database, based on near-universal single-copy orthologs. BLAST analysis was used to identify the species corresponding to each scaffold.

ANI analysis was performed by aligning the reference sequence (Taxon ID: GCF_007992815.1_S. xylosus_NBRC_109,770) with contig 1 using BLASTN (custom assay project, Macrogen Inc., Seoul, Korea). A circular genome map was generated via the PROKSEE server¹ (Grant et al., 2023). Genome sequence data underwent further analysis through the Type (Strain) Genome Server (TYGS) for a comprehensive genome-based taxonomic study (Meier-Kolthoff and Göker, 2019). The closest related strain genomes were identified using MASH distance comparisons, with precise distances calculated by the Genome BLAST Distance Phylogeny (GBDP) method under the “coverage” algorithm (Meier-Kolthoff et al., 2013). Phylogroups were classified using in silico Clermont Phylotyper EzClermont (Waters et al., 2020). Synteny analysis was conducted via multiple genome alignment using Mauve with default parameters (minimum Localized Collinear Blocks (LCBs) of 1,000, island size of 50, backbone size of 50, and maximum gap of 50). The draft genome of FFCShyA4 was aligned relative to the S. xylosus CCM2738 and S. xylosus NBRC 109770, reference genomes—originally isolated from human skin (Darling et al., 2004).

The identification of coding sequences (CDS), rRNA, tRNA/tmRNA, signal leader peptides, and noncoding RNA was conducted following the methodology outlined by Molina et al. (2024). Gene annotation was performed using Prokka v1.14.6 (Seemann, 2014), while plasmid annotation was carried out with pLannotate (McGuffie and Barrick, 2021). Functional annotation was achieved using InterProScan v5.0 (Jones et al., 2014), which evaluates sequence similarity at the family level and matches sequences against member databases within InterPro, including Pfam, the Conserved Domain Database (CDD), and TIGRFAM, a curated database focusing on prokaryotic protein families (Selengut et al., 2007). Furthermore, the EggNOG database (Evolutionary Genealogy of Genes: Non-supervised Orthologous Groups) was utilized for additional annotations (Huerta-Cepas et al., 2019), with psi-BLAST applied to align predicted protein sequences to the EggNOG database.

For the FFCShyA4 strain, a total of 2,972,650,628 bases were generated, producing 19,686,428 pair-end reads. The total number of contigs was 2,856,035 bp. The GC content was calculated at 32.63%, with a Q30 value of 92.75%. The estimated genome size was 3,093,530 bp. In addition, the strain harbored one plasmid of 32,035 bp. The general genome features of FFCShyA4 are shown in Supplementary Table S3. The circular genome and plasmid maps are shown in Figures 1A,B. Following assembly of the complete or draft genome, BLAST analysis was conducted to identify species-level similarities for each scaffold. Genus-level classification revealed that 90.91% of the scaffolds matched Staphylococcus, while 9.09% aligned with Mammaliicoccus. ANI analysis revealed a 99.97% nucleotide identity and 94.19% alignment coverage with S. xylosus NBRC109770 (human skin) (Lamers et al., 2012). The five most closely related genomes identified based on ANI results are presented in Supplementary Table S4. Hierarchical clustering of ANI data was visualized as two-dimensional dendrograms, using simple linkage for both ANI percentage identity and ANI alignment coverage (Supplementary Figures S1A,B). BLASTN analysis revealed that plasmid pFFCShyA4 shares 99.26% sequence identity with plasmid pDMSX03-1 from S. xylosus strain DMSX03. Phylogenetic analysis identified S. xylosus strain CCM2738 and S. pseudoxylosus strain S044009 as the closest related strains as shown by TYGS analysis (Supplementary Figure S2). The Mauve alignment of FFCShyA4, and the two references S. xylosus NBRC109770 and S. xylosus CCM2738 reveals a mix of conserved and variable regions, highlighting both evolutionary stability and divergence (Figure 2). The conserved regions, mainly found at the genomes terminal ends and encompassing essential housekeeping genes, suggest a high degree of preservation among the three strains (González et al., 2010). However, extensive genomic rearrangements, including inversions and translocations, are observed between 1,000,000–1,800,000 bp, suggesting evolutionary divergence (Shukla et al., 2009). Notable insertion/deletion (indel) events occur within 600,000–1,000,000 bp and 2,200,000–2,600,000 bp, with FFCShyA4 displaying unique genomic islands absent in the reference strains, potentially linked to adaptation in food-related environments (Vermassen et al., 2016). Additionally, the high density of intersecting lines observed in the 1,200,000–1,600,000 bp region indicates potential hotspots for horizontal gene transfer (HGT), likely involving MGEs (Ghaly et al., 2024). These regions frequently harbor genes linked to antibiotic resistance, responses to metal and oxidative stress (e.g., copper and arsenic resistance genes, superoxide dismutase), as well as carbohydrate metabolism and fermentation. The presence of HGT-associated markers, including transposases and prophage-related genes, underscores the strain’s metabolic versatility and genomic plasticity (Coll et al., 2025). However, the alignment highlights conserved core regions while also identifying unique genomic segments in FFCShyA4 that may contribute to its distinct functional capabilities, including potential virulence and resistance traits (Zhou et al., 2020). These results support the hypothesis that FFCShyA4 shares a common lineage with NBRC 109770 and CCM2738 but has undergone genetic modifications that could enhance its survival in food-associated environments or human-related settings, warranting further investigation into its role in public health and food safety (Vermassen et al., 2014). These findings suggest that the FFCShyA4 strain may possess unique traits enhancing its environmental adaptability and resilience, particularly in response to stressful or competitive conditions such fruit environment.

The genome comprises 2,720 genes, including 2,665 coding sequences (CDS), 47 tRNAs, 7 rRNAs, and 1 tmRNA (Supplementary Table S3). Gene locations were predicted using Prokka, while BLAST analysis was performed to determine the functions and identify the assembled sequences by comparing them against nucleotide and protein sequence databases. The predicted genes were subsequently aligned with multiple databases to obtain functional annotations using specific aligners, as summarized in Table 1. Among the total proteins, 2,576 matched EggNOG DB proteins (2,538 Single EggNOG and 38 Multi EggNOG proteins) and 89 proteins with no hit (Figure 3A). Similarly, 2,476 proteins matched COG DB (2,210 Single COG category and 266 Multi COG category) while 189 showed no hit (Figure 3B). The number of genes associated with KEGG (2,580 genes) functional annotation categories is shown in Supplementary Figure S3. The plasmid annotation (Figure 1B) includes several key features: arsB (100% identity) from S. xylosus, involved in arsenical resistance and likely forming the channel of an arsenite pump; arsR (99% identity) from S. aureus, a transcriptional repressor for the ars operon, which regulates its own expression and is derepressed by oxyanions of arsenic, antimony, and bismuth in the +III oxidation state, as well as arsenate (As(V)); arsC (95.4% identity) from S. aureus strain N315, encoding an enzyme that reduces arsenate [As(V)] to arsenite [As(III)]; bin3 (83.2% identity) from S. aureus, a potential DNA invertase; crtN (66.5% identity) and crtP (65% identity) from S. aureus, involved in staphyloxanthin biosynthesis, a carotenoid contributing to virulence by protecting against oxidative stress; and ADH1 terminator (100% identity), a transcription terminator for the S. cerevisiae alcohol dehydrogenase 1 (ADH1) gene. These features highlight roles in resistance, enzymatic activity, and oxidative stress protection. ADH1 is not naturally found in Staphylococcus species (Makhlin et al., 2007). In the context of plasmid annotation, the ADH1 terminator is used as a transcription terminator, and possible serves as a genetic element for controlling gene expression.

Three CRISPR sequences were identified in the FFCShyA4 genome, located at positions 79,658 to 79,743, 868,840 to 868,928, and 47,184 to 47,284, respectively. Each sequence includes a short spacer sequence flanked by degenerate repeats (consensus DRs), with repeat conservation of 96.68, 96.65, and 100%, and spacer conservation of 100%. No Cas genes were found. One intact prophage was found in the genome (Supplementary Table S5). The detection of Phage Staphy_47 in the FFCShyA4 genome is significant as it highlights the potential influence of prophages on the genetic and functional diversity of the host bacterium. This phage can contribute to HGT, providing the host with adaptive advantages, including virulence factors, resistance genes, or enhanced survival mechanisms under stress conditions (Deghorain and Van Melderen, 2012). Additionally, a total of 138 GIs of 126,044 bp were predicted with IslandViewer using as a reference genome of non-coagulase S. xylosus strain DMSX03 isolated from fermented soybean (Heo et al., 2021). Many hypothetical proteins and enzymes, such as proteases, and virulence factors (mazE, PemK/MazF), were identified in the GIs. These genomic regions are crucial for the rapid evolution, diversification, and adaptation of Staphylococcus due to their frequent rearrangements, excisions, transfers, and the acquisition of additional DNA (Supplementary Table S6). Among the 121 genes identified from the virulence database, the thermonuclease gene (nuc), the intercellular adhesion protein C involved in polysaccharide intercellular adhesin (PIA) synthesis (icaC), and the ATP-dependent Clp protease proteolytic subunit (clpP) were detected, although they exhibited imperfect matches (identity < 85%, with the input sequence length shorter than the full virulence gene length, and 60% alignment coverage). In addition, the elongation factor Tu (tufA) and phosphopyruvate hydratase (eno) were annotated showing 74 and 79.8% sequence identity, respectively. The detected virulence factors are summarized in Supplementary Table S7. Interesting, the presence of nuc but not icaC, sdrE, and hlg genes was confirmed by PCR (Supplementary Figure S4), as these genes typically are associated with S. aureus. The detection of the nuc gene in CNS strain may be attributed to HGT, genomic variability within CNS, or the presence of a divergent gene variant with partial homology, highlighting the potential for genetic exchange and adaptation in diverse environments (Canning et al., 2020). Besides, several virulence factors such as surface-displayed alpha-enolase (eno), adaptor protein MecA (mecA), type II toxin-antitoxin system PemK/MazF family toxin, a virulence associated protein E (virE), were annotated with EggNOG. Moreover, within FFCShyA4 genome, a total of 133 MGEs were identified, including 17 related to insertion/excision, 74 associated with replication/recombination/repair, 21 linked to phages, 6 involved in stability/transfer/defense, and 15 related to transfer. Analysis of the 34 matched pathogenic families (accounting for 1.28% of the proteome) indicated that the FFCShyA4 isolate is a human pathogen with a likelihood of 0.982. This finding was further supported by the presence of an inhibition zone on blood agar, indicative of hemolytic activity (Supplementary Figure S5). Additionally, the gelatinase test for FFCShyA4 was positive, consistent with previous research showing that gelatinase is a common virulence factor among Staphylococcus spp. (Bertelloni et al., 2021). Using CARD protein IDs, 61 predicted genes were classified by drug class, resistance mechanism, and AMR gene family, with the distribution of genes shown in Supplementary Figure S6. Besides, 11 ARGs including vancomycin (vanY gene in vanA cluster), lincosamide (salD), and fosfomycin (fosBx1) resistance were detected according to the RGI tool criteria. The presence of acquired ARGs indicates that this strain was likely exposed to selective pressures, such as environmental antimicrobial use, which may have promoted the acquisition of these resistance genes through horizontal gene transfer (HGT) or other mechanisms (Chen et al., 2022). Based on EggNOG annotation, we confirmed the presence of several key antibiotic resistance-related genes in the FFCShyA4 genome, including fusA, pbp4, and ileS. The fusA gene encodes elongation factor G (EF-G), which is essential for protein synthesis and serves as a target for antibiotics like fusidic acid, suggesting a potential mechanism for resistance to this antibiotic (Resch et al., 2008). The pbp4 gene encodes penicillin-binding protein 4 (PBP4), which plays a critical role in bacterial cell wall synthesis and is a key target of β-lactam antibiotics like penicillin (Satishkumar et al., 2021). Alterations in PBP4 can reduce the binding affinity of these antibiotics, contributing to resistance. However, its role in S. xylosus was not yet identified. Additionally, ileS, which encodes isoleucine-tRNA ligase, is involved in protein synthesis and may influence resistance indirectly by impacting bacterial growth and response to translation-targeting antibiotics (Grundy et al., 1997). These findings indicate that the FFCShyA4 strain exhibits a high level of adaptability to antibiotic pressure, its genomic features suggest it may serve as a reservoir of ARGs across environmental interface. These genomic features suggest that FFCShyA4 possesses ecological adaptability and may function as a reservoir of ARGs, with potential relevance to One Health pathways of AMR dissemination across human, animal, and environmental interfaces (Robinson et al., 2016).

Two bacteriocin clusters (Area of interest, AOI) were annotated within the FFCShyA4 genome as follows: contig 5.2 (AOI_01) (started at 18217, end at 38349) of auto-inducing peptide III (AIP III) class, and contig 5.2 (AOI_02) (started at 85346, end at 105367) of the sactipeptides class (ribosomally synthesized peptides) (Figure 4). BLASTP analysis revealed that AIP III shares 100% amino acid sequence identity with the cyclic lactone autoinducer peptide AgrD found in several Staphylococcus species, including S. xylosus. The agr locus was originally characterized in S. aureus as a regulatory element that controls the production of exoproteins involved in virulence (Dufour et al., 2002). Besides, AIP III plays a critical role in quorum sensing within Staphylococcus species, particularly S. aureus (Tamai et al., 2023). By regulating gene expression in response to population density, AIP III influences virulence factors, biofilm formation, and antibiotic resistance. This peptide signals the bacteria to produce toxins and surface proteins that aid in immune evasion and infection persistence, while biofilms protect the bacteria from antimicrobials (Tamai et al., 2023). Disrupting AIP III-mediated quorum sensing presents a potential therapeutic strategy to reduce bacterial virulence and biofilm formation (Gonzales et al., 2024). In addition, the sactipeptides, a class of ribosomally synthesized and post-translationally modified peptides (RiPPs), exhibit diverse biological activities, including antibacterial, spermicidal, and hemolytic properties (Chen et al., 2021). Analysis of the FFCShyA4 genome using the antiSMASH web tool identified nine distinct biosynthetic regions (Table 2). These include a terpene biosynthetic cluster (region 1.1) with heme D1 as the closest known cluster (11% similarity), an NI-siderophore cluster (region 2.1) with 100% similarity to a known siderophore cluster, and regions encoding Type III polyketide synthase (T3PKS) and non-ribosomal peptide synthase (NRPS) on contigs 2 and 3. Based on gene similarity analysis with the MiBIG database, a 54% sequence similarity was observed with staphylopherrin A found in S. aureus subsp. aureus NCTC 8325 (Supplementary Table S8), which plays a vital role in iron acquisition (Battaglia and Garrett-Sinha, 2023). Staphylopherrin A aids the bacteria by binding to iron (Fe3+) in the environment and facilitating its transport into the bacterial cell (Battaglia and Garrett-Sinha, 2023). Early genomic analysis of the S. xylosus C2a strain reveals the presence of genes encoding staphylopherrin A (Vermassen et al., 2014). This iron acquisition mechanism enables S. aureus to survive and thrive within the host, enhancing its virulence and ability to cause infections. This suggests that S. xylosus may utilize similar mechanisms as S. aureus to acquire iron from the environment. Notably, one NRPS cluster predicted within FFCShyA4 genome shared 8% similarity with surfactin. Additionally, an opine-like metallophore cluster located on contig 3 displayed 100% similarity to the staphylopine biosynthesis cluster. Finally, a cyclic lactone autoinducer biosynthetic region (region 5) showed 4% similarity to the kijanimicin cluster. Staphylopine, a versatile metallophore produced by the prominent human pathogen Staphylococcus aureus, is essential for the uptake of transition metals and contributes significantly to bacterial virulence (Song et al., 2018). These findings suggest that S. xylosus may employ virulence-like strategies like S. aureus, underscoring the need to assess its role in antimicrobial resistance dissemination and cross-domain adaptation (Battaglia and Garrett-Sinha, 2023). Further research should explore the functional roles of the AIP III and sactipeptide bacteriocin clusters in FFCShyA4, particularly their involvement in quorum sensing, biofilm formation, and microbial competition within foodborne ecosystems.

S. xylosus FFCShyA4 exhibited β-hemolytic activity and tested positive for gelatinase, like the S. aureus reference strains ATCC1026 and ATCC43300, indicating a potential adaptive trait rather than a direct virulence marker (Supplementary Figure S5). Genome annotation revealed the presence of a hemolysin III family protein in FFCShyA4; however, the failure to amplify hemolysin genes via PCR suggests potential sequence variations leading to primer mismatches or regulatory mechanisms influencing gene expression. Alternatively, hemolytic activity may be mediated by proteases or phospholipases rather than classical hemolysins (Vandenesch et al., 2012). It is also possible that homologous hemolysin genes or functional analogs were acquired through horizontal gene transfer (HGT) or retained for ecological adaptation. Unlike S. aureus, where hemolysins are key virulence factors, S. xylosus likely utilizes β-hemolysis for ecological fitness, particularly in food-associated environments such as plant surfaces. While this activity may enhance colonization potential, it does not necessarily indicate pathogenicity in healthy individuals, underscoring the phenotypic diversity and evolutionary plasticity of staphylococcal species (Eltwisy et al., 2020). The FFCShyA4 strain contained several putative resistance genes linked to resistance to lincosamide (1), tetracycline (4), fosfomycin (1), fluoroquinolone (16), macrolide (3), mupirocin (1), and salicylic acid (1) antimicrobials (Supplementary Figure S7). The antibiotic resistance profile of FFCShyA4 was compared with S. aureus ATCC1026 and S. aureus ATCC43300, highlighting key differences in their susceptibility patterns (Figure 5). FFCShyA4 shows resistance to several key antibiotics, particularly TE30 (tetracycline), MET5 (methicillin), CXM30 (cefuroxime), VA30 (Vancomycin), AX25 (amoxycillin), AM10 (ampicillin), OX1 (oxacillin), and LZD30 (linezolid), while displaying intermediate resistance to CIP5 (ciprofloxacin), C30 (chloramphenicol), and DA2 (clindamycin) (Supplementary Table S9). The MAR index was 0.67 for FFCShyA4, 0.70 for S. aureus ATCC 1026, and 0.62 for S. aureus ATCC 43300. Typically, a MAR index above 0.2 is considered significant, as it suggests exposure to antimicrobials in environments where misuse or overuse occurs, such as hospitals or livestock settings (Tang et al., 2023). The high MAR index and resistance to critical antibiotics, including methicillin and vancomycin, reflect possible exposure to antimicrobials in agricultural or post-harvest settings, raising concerns over antimicrobial misuse beyond clinical boundaries (Iwu et al., 2020). From a One Health perspective, the detection of multidrug-resistant S. xylosus on plant-based food products underscores the interconnectedness of human, animal, and environmental health, emphasizing the need for integrated surveillance of AMR across non-clinical reservoirs such as agricultural commodities, food supply chains, and market environments (Robinson et al., 2016). Compared to the reference S. aureus strains, FFCShyA4 exhibited phenotypic resistance to tetracycline, corroborating the resistance determinants identified through genomic analysis. The EggNOG annotation of FFShyA4 revealed the presence of tpiA and bla genes, indicating its potential for methicillin and β-lactam resistance, aligning with the observed multidrug resistance in the antibiogram. The bla gene suggests β-lactamase production, leading to resistance to penicillin and related antibiotics. Although the FFCShyA4 strain exhibited phenotypic resistance to methicillin, complementary PCR analysis using mecA and mecC primers failed to detect these genes. This discrepancy may be attributed to primer mismatches arising from sequence variations or the genomic location of mecA, potentially affecting primer binding efficiency. Besides, the presence of tpiA confirms the species identity, as it is a conserved housekeeping gene in S. xylosus. Early investigations into methicillin-resistant coagulase-negative staphylococci revealed that resistance to most β-lactam antibiotics arises primarily from the acquisition of the mecA or mecC genes, which encode alternative penicillin-binding proteins (PBPs) with reduced β-lactam binding affinity (Schwarz et al., 2018). The absence of mecA/mecC genes, common in livestock-associated Staphylococcus, does not exclude an animal origin but may indicate a less direct or recent exposure. Besides, the multidrug efflux pumps such as NorA which was annotated in the FFCShyA4 genome may contribute to methicillin resistance by actively exporting β-lactam antibiotics, reducing their intracellular concentration (Dashtbani-Roozbehani and Brown, 2021). Further comparative genomic and epidemiological studies would be necessary to confirm the specific origin of this strain. Another common resistance mechanism against penicillin involves enzymatic hydrolysis of the antibiotic by β-lactamases encoded by the blaZ or blaARL genes (Schwarz et al., 2018). The absence of mecA/mecC alongside the detection of alternative resistance determinants (bla, NorA) supports the idea that non-traditional or environmental pathways may be driving resistance evolution in underregulated sectors, such as local markets. Early studies indicate that antimicrobial resistance in Staphylococcus spp. isolated from various food products varies significantly depending on the species and source of isolation (Resch et al., 2008). Among S. xylosus strains, resistance rates ranged from 22% for tetracycline to 69% for penicillin, with 93% of isolates originating from meat starter cultures. Notably, all coagulase-negative staphylococci (CNS) strains were fully susceptible to clinically relevant antibiotics such as chloramphenicol, clindamycin, cotrimoxazole, gentamicin, kanamycin, linezolid, neomycin, streptomycin, and vancomycin (Resch et al., 2008). Additionally, S. xylosus strains isolated from fermented soybean products were reported to be susceptible to chloramphenicol, erythromycin, gentamicin, kanamycin, lincomycin, oxacillin, tetracycline, and trimethoprim (Kong et al., 2022). However, 23 strains exhibited acquired phenotypic resistance to erythromycin, lincomycin, and tetracycline. Minimum inhibitory concentration (MIC) testing showed continuous or unimodal distribution patterns for these antibiotics, which did not align with the resistance profiles, suggesting possible discrepancies. These findings indicate that the current CLSI breakpoint values for ampicillin and penicillin G in S. xylosus may need to be re-evaluated. Moreover, the presence of acquired resistance to erythromycin, lincomycin, and tetracycline underscores the importance of performing antibiotic susceptibility testing before utilizing these strains in food-related applications (Kong et al., 2022). These observations underscore the necessity for coordinated, cross-sectoral surveillance strategies encompassing food production and environmental interfaces to monitor the emergence and dissemination of antimicrobial resistance, particularly in regions with insufficient regulation of antimicrobials use in agri-food systems.