Informed consent and permission for sample collection were obtained from animal owners and veterinary facilities before sampling. The Research Ethics Committee of the Faculty of Veterinary Medicine, Universitas Gadjah Mada, Yogyakarta, Indonesia, approved the collection and use of animal isolates (Approval No. 143/EC-FKH/Int./2024).

This study used a cross-sectional, laboratory-based observational design to investigate the genotypic characteristics of AMR in S. aureus isolated from animals. Sampling was conducted from June 2024 to August 2025. Laboratory analyses were performed at the Clinical Pathology Laboratory, Faculty of Veterinary Medicine, and the Animal Science Learning Center at Universitas Gadjah Mada in Yogyakarta, Indonesia.

A total of 121 bacterial isolates, comprising 30 from bovine milk, 61 from cats, 18 from dogs, 7 from rabbits, and 5 from goat milk, were used in this study. Bovine and goat milk samples were obtained from dairy farms in Boyolali, Central Java, Indonesia, whereas companion animal samples were collected from veterinary clinics and hospitals in Yogyakarta and Semarang.

Animals exhibiting clinical signs suggestive of bacterial infection, such as pyogenic skin lesions, rhinitis, pyrexia, lethargy, or mastitis (dairy animals) were included in the study. Swab samples were collected from the oropharynx, nasal cavity, ears, and skin lesions of companion animals. Milk samples were aseptically collected from dairy animals with clinical or subclinical mastitis. Animals that had received systemic antibiotics within 2 weeks before sampling or samples showing mixed bacterial growth were excluded from the analysis.

Continuous sample collection was conducted throughout the study period to minimize seasonal or temporal bias. All samples were aseptically transported and processed immediately upon arrival at the laboratory. Specimens were cultured on 5% defibrinated sheep blood agar and incubated aerobically at 37°C for 24 h. Colonies with typical S. aureus morphology were selected for further identification by Gram staining, hemolytic activity testing, and subculture onto mannitol salt agar. Biochemical confirmation was performed using catalase and coagulase tests following standard protocols.

All S. aureus isolates were inoculated into tryptic soy broth (Merck, Germany) and incubated at 37°C for 24 h. The culture was washed with sterile phosphate-buffered saline (PBS) and diluted with sterile PBS to a turbidity equivalent to a 0.5 McFarland standard (HiMedia, USA). Susceptibility testing was performed using the Kirby–Bauer disk diffusion method in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI M100, 2024) [11].

The culture was standardized to a 0.5 McFarland standard, and 70 µL was spread on Mueller-Hinton agar (MHA; Merck, Germany) using a glass spreader. Antibiotic disks were placed on MHA at a specified distance from one another and incubated at 37°C for 24 h. The antibiotic disks (Oxoid, UK) used in this study were tetracycline (30 µg), gentamicin (10 µg), erythromycin (15 µg), penicillin G (10 µg), cefoxitin (30 µg), ciprofloxacin (5 µg), and clindamycin (10 µg). Inhibition zone diameters were measured and interpreted as susceptible, intermediate, or resistant according to CLSI M100 (2024) criteria. Multidrug resistance (MDR) was defined as resistance to ≥1 agent in ≥3 antimicrobial classes, as defined by Magiorakos et al. [12].

Genomic DNA was extracted from S. aureus isolates using the Geneaid Bacteria DNA Kit (Geneaid, Taiwan) according to the manufacturer’s instructions. Molecular identification was performed by polymerase chain reaction (PCR) amplification of the 23S rRNA and nuc genes, and the coa gene was amplified to characterize S. aureus isolates.

Detection of antibiotic resistance genes, including mecA (methicillin/oxacillin resistance), blaZ (penicillin G resistance), aacA-D (aminoglycoside resistance), ermA, msrB, and linA (macrolide-lincosamide-streptogramin B resistance), tetK and tetM (tetracycline resistance), and norA (ciprofloxacin resistance), was performed using specific primers and PCR conditions listed in Table 1. Each 10 μL PCR reaction mixture contained 0.5 μL of forward primer (10 pmol), 0.5 μL of reverse primer (10 pmol), 5 μL of PCR mix, 3.5 μL of nuclease-free water, and 0.5 μL of DNA template. MyTaq HS Red Mix (Bioline, UK) was used for the amplification of all target genes except nuc, for which PowerPol 2× PCR Mix with Dye (ABclonal, USA) was used. Briefly, the PCR tubes were centrifuged and placed in a thermal cycler under the corresponding conditions. All primers used in this study were selected from previously published studies and validated for specificity and sensitivity against S. aureus (Table 1) [13–21]. For quality assurance, each PCR run included a positive control consisting of DNA from a well-characterized S. aureus strain, previously confirmed phenotypically, genotypically, and by whole-genome sequencing, and a negative control containing nuclease-free water instead of template DNA. All PCR assays were performed in duplicate to ensure reproducibility, and consistent results were obtained across repeated runs.

The PCR products were separated by electrophoresis on a 1.5% (w/v) agarose gel (GeneDireX, USA) in 1× TBE buffer and stained with Safe-Red™ Gel Stain (ABM, Canada). Bands were visualized using a UV transilluminator and compared with a 100-bp DNA ladder (Smobio, Taiwan).

We ensured quality control by including a well-characterized S. aureus isolate that had been previously confirmed phenotypically and genotypically and further validated by whole-genome sequencing [9]. This isolate served as an internal positive control for biochemical identification, antimicrobial susceptibility testing, and PCR assays. These measures ensured the accuracy, reliability, and reproducibility of the phenotypic and molecular results.

Descriptive statistics were used to summarize the distribution of S. aureus isolates and resistance genes across animal species (bovine milk, cats, dogs, rabbits, and goat milk). A comparative analysis of resistance gene prevalence and antimicrobial resistance patterns among host species was performed using Fisher’s exact test at a significance level of p < 0.05. The agreement between phenotypic AMR and the corresponding resistance genes was evaluated descriptively by assessing the concordance between disk diffusion results and PCR detection.

A total of 55 isolates were confirmed as S. aureus based on colony morphology, biochemical characteristics, and PCR amplification of the 23S rRNA and nuc genes. The isolates were obtained from various animal sources, including bovine milk (24/30), cats (18/61), dogs (5/18), rabbits (6/7), and goat milk (2/5).

Typical S. aureus colonies were round, smooth, convex, glistening, and grayish to golden-yellow on sheep blood agar (5%). Most isolates exhibited a creamy-to-opaque appearance with distinct hemolytic zones. All isolates were Gram-positive, catalase-positive cocci arranged in clusters. Mannitol fermentation was observed in 98.2% (54/55) of the isolates, and 43.6% (24/55) were coagulase-positive. Variable hemolysis patterns (α-, β-, and γ-hemolysis) were observed among the isolates obtained from different animal sources. In addition, 5 isolates from bovine milk (2/24, 8.3%), dog (2/5, 40%), and goat milk (1/2, 50%) exhibited atypical small-colony variant (SCV)-like morphology, characterized by pinpoint colonies smaller than those of wild-type S. aureus (Figure 1). These morphological variants were rare but consistently reproducible upon subculture. The phenotypic results were consistent with the general characteristics of S. aureus and supported the identification of the species before molecular analysis. Table 2 summarizes the detailed biochemical and hemolytic profiles.

Molecular confirmation of all isolates was performed by PCR amplification of the 23S rRNA and nuc genes. Characterization of S. aureus isolates was performed by detecting the coa gene. The amplicon sizes of the 23S rRNA, nuc, and coa genes were approximately 1250, 279, and 204 bp, respectively (Figure 2). All S. aureus isolates from dogs, rabbits, and goat milk were positive for the coa gene, whereas 95.8% (23/24) of bovine milk isolates and 72.2% (13/18) of cat isolates were positive.

All S. aureus isolates confirmed by phenotypic and genotypic tests were evaluated for antimicrobial susceptibility using the Kirby–Bauer disk diffusion method with the following antibiotics: tetracycline (TE, 30 µg), gentamicin (CN, 10 µg), erythromycin (E, 15 µg), penicillin G (P, 10 µg), cefoxitin (FOX, 30 µg), ciprofloxacin (CIP, 5 µg), and clindamycin (DA, 10 µg).

Table 3 and Figure 3 summarize the overall resistance rates among S. aureus isolates from various animal sources. The isolates exhibited variable resistance patterns, with the highest rates of resistance to penicillin (20%–72%) and tetracycline (17%–28%), whereas most isolates remained susceptible to ciprofloxacin and cefoxitin. No significant differences in the distribution of antibiotic resistance were observed among animal sources (Fisher’s exact test, p > 0.05). In contrast, erythromycin showed a statistically significant difference (Fisher’s exact test, p < 0.05).

All 55 S. aureus isolates from bovine milk, cats, dogs, rabbits, and goat milk were screened for antibiotic resistance–encoding genes using PCR. The genes analyzed were tetK, tetM, ermA, blaZ, aacA-D, norA, linA, mecA, and msrB. Overall, tetK was detected in all isolates across all host species, indicating widespread tetracycline resistance. High detection frequencies were also observed for blaZ, norA, linA, and mecA, although prevalence varied among animal sources. In contrast, ermA and msrB were detected at relatively low frequencies and exhibited host-specific distribution patterns. Table 4 summarizes the detailed detection rates by host species, while Figures 4 and 5 show representative PCR amplicons and comparative visualizations of resistance gene profiles. Significant differences were observed for tetM, blaZ, aacA-D, norA, and msrB (p < 0.05), whereas no significant differences were observed for ermA, linA, and mecA. The tetK gene was excluded because it was detected uniformly across all sources.

The MDR pattern of S. aureus isolates varied by animal source. The most frequent phenotypic combinations included resistance to β-lactams (penicillin, oxacillin, ampicillin, amoxicillin, and cefoxitin), macrolides (erythromycin), and tetracycline (Figure 6). Although several MDR patterns were more common in specific animal sources, none of the observed differences reached statistical significance using Fisher’s exact test. The molecular detection of resistance genes revealed diverse genotypic profiles that matched the observed phenotypes (Figure 7). Several MDR patterns were observed among S. aureus isolates from different animal sources, with certain patterns showing significantly higher prevalence in specific sources, particularly dogs and cats. In contrast, the majority of cases occurred at low frequencies and did not differ statistically (p < 0.05). The coexistence of multiple resistance genes was common among S. aureus isolates from all animal sources and was dominated by combinations involving tetK, often accompanied by tetM, norA, linA, and mecA. Overall, gene profiles with ≥ 4 resistance genes accounted for the majority of MDR isolates, with combinations including tetK–tetM and tetK–norA–linA observed in more than 50% of isolates across host species. More complex gene constellations (≥ 6 genes) were particularly frequent in bovine milk and companion animal isolates, underscoring extensive MDR at the genotypic level. The presence of multiple resistance genes in several isolates indicated the potential for MDR among circulating strains across animal sources.

Based on the biochemical characteristics and PCR amplification of the 23S rRNA and nuc genes, a total of 55 isolates from bovine milk, cats, dogs, rabbits, and goat milk were confirmed to be S. aureus. The 23S rRNA and nuc genes are widely accepted molecular markers for S. aureus identification [9], whereas the coa gene provides additional strain-level specificity [22]. However, several isolates showed discordance between molecular detection of the coa gene and phenotypic coagulase testing. Similar inconsistencies have been reported in regional and international studies, indicating that phenotypic coagulase tests alone may underestimate the prevalence of S. aureus, particularly in animal-derived isolates, and should be complemented with molecular confirmation [23–25]. Mutations affecting coa expression or regulation have been proposed as underlying mechanisms for such discrepancies [24, 26–27].

A small proportion of isolates from bovine milk, cats, dogs, and rabbits were catalase-negative, a rare phenotype in S. aureus. Catalase-negative S. aureus (CNSA) strains have been linked to katA gene mutations. Although they are increasingly reported in both human and veterinary settings, their potential for virulence remains unclear [28]. Their presence in animal sources underscores the importance of molecular diagnostics in preventing misidentification and underreporting.

Hemolytic activity varies by host origin, reflecting host-adaptive virulence traits. Beta-hemolysis predominated among isolates from cats, consistent with studies implicating β-hemolysin in skin and soft-tissue infections in companion animals [29, 30]. In contrast, gamma-hemolysis was more common among bovine milk isolates, consistent with reports of mastitis-associated S. aureus strains in Southeast Asia and elsewhere, where non-hemolytic phenotypes are frequently observed [31–34]. These findings support the concept that host niche and infection ecology influence hemolysin expression.

In this study, five isolates exhibited atypical growth, forming pinpoint-sized colonies smaller than those of wild-type S. aureus (Figure 1), resembling SCVs. SCVs are a subpopulation of S. aureus characterized by phenotypic alterations that enable persistence under adverse environmental conditions. SCVs have been increasingly reported among livestock- and companion-animal-associated S. aureus, particularly in the context of chronic infections and prolonged antimicrobial exposure [35]. SCVs exhibit enhanced biofilm formation, altered metabolism, and increased antibiotic tolerance, thereby facilitating persistence within the host and environment [36, 37]. Although SCVs were not explicitly characterized in this study, their associations with biofilm formation, intracellular persistence, and antimicrobial tolerance are highly relevant in veterinary settings. SCVs may contribute to chronic or recurrent infections in animals and complicate treatment outcomes, particularly in the context of MDR S. aureus [35]. The interplay among methicillin resistance, biofilm production, and host adaptation highlights MRSA as a significant One Health concern, facilitating transmission among animals, humans, and the environment and complicating infection control efforts. The recognition of these phenotypes underscores the need for further investigation into adaptive survival strategies of S. aureus in animal sources and their implications for long-term infection control.

In this study, the antimicrobial susceptibility profiles showed that resistance patterns were broadly comparable across animal sources for most antibiotics tested, with no statistically significant differences. This suggests a relatively uniform distribution of resistance phenotypes across animal sources, which may reflect similar antimicrobial exposure or shared environmental and management conditions. Erythromycin was the only antimicrobial for which a significant difference was observed among animal sources. Resistance to macrolides may be influenced by host-associated factors or differences in usage patterns. The majority of macrolide use was in food-producing animals; in the U. S. in 2017, macrolides were the most commonly used in cattle [38].

In this study, S. aureus isolates from different animal sources exhibited variable resistance patterns, consistent with regional and global reports on antimicrobial resistance in S. aureus from livestock and companion animals [39–42]. These differences largely reflect geographic variation in resistance patterns shaped by antimicrobial use practices, regulatory frameworks, and animal production systems. The widespread empirical use of penicillins, tetracyclines, and aminoglycosides in food-producing animals has been identified as a major driver of the emergence of resistance in Indonesia and other Southeast Asian countries [42–45]. This study extends previous Indonesian reports on S. aureus by providing a comparative, multi-host analysis of resistance gene profiles across livestock and companion animals within the same geographic regions. Unlike earlier studies that focused on single host species or on phenotypic resistance alone [8, 9, 25, 33], our findings demonstrate shared resistance gene patterns across animal sources, supporting the existence of interconnected AMR reservoirs in a One Health context.

Bacteria were classified as MDR if they showed resistance or intermediate resistance to at least one antibiotic from three or more antimicrobial classes [12, 13]. MDR S. aureus strains were primarily detected in bovine milk and cat isolates (Figure 6). The consistent statistical significance across multiple MDR patterns reflects a structured distribution of resistance combinations rather than independent random events. Many MDR patterns share common antibiotics, resulting in correlated resistance profiles that manifest as significant global associations across multiple combinations. The occurrence of MDR strains in dairy cattle has been strongly associated with repeated antibiotic use for mastitis management, suboptimal milking hygiene, and environmental contamination [42]. Similar trends have been reported in major dairy-producing regions across Asia, underscoring the zoonotic risk posed by MDR S. aureus throughout the food chain [45]. Close contact with humans and frequent veterinary interventions in companion animals further increase the potential for interspecies transmission.

MDR S. aureus isolates were confirmed by the detection of several antibiotic resistance genes (Figure 7). The frequent co-occurrence of resistance genes, such as mecA, blaZ, and tetK, suggests a potential role for horizontal gene transfer. However, the underlying genetic platforms (e.g., plasmids or SCCmec types) remain uncharacterized and warrant further investigation using whole-genome approaches. Similar findings have been reported in several countries, indicating that MDR S. aureus poses a significant challenge to both animal and public health due to its zoonotic potential [45, 46]. A comparison with studies from Malaysia [45] indicates broadly similar resistance trends, reinforcing the regional relevance of our findings. MDR transmission can occur within veterinary clinics through contact among personnel, animals, and owners and may contribute to nosocomial spread or localized outbreaks. The predominance of shared gene combinations across animal hosts further indicates interconnected resistance reservoirs consistent with the One Health framework. Therefore, continuous surveillance and prudent antibiotic use are essential to minimize the emergence and dissemination of MDR S. aureus in both livestock and companion animals.

PCR amplification detected several antibiotic resistance genes, including tetK, tetM, ermA, blaZ, aacA-D, norA, linA, mecA, and msrB, in S. aureus isolates from diverse animal sources. Genotypic profiles were compared with phenotypic susceptibility results obtained using the Kirby–Bauer disk diffusion method. Discrepancies between resistance genotypes and phenotypes were observed among S. aureus isolates obtained from different animal sources. In several cases, isolates exhibited phenotypic resistance despite the absence of corresponding resistance genes. Some isolates phenotypically resistant to oxacillin were mecA-negative. This discrepancy may reflect the presence of mecC, an analog of mecA located on the SCCmec type XI element, which encodes a penicillin-binding protein (PBP2c) that confers β-lactam resistance but has distinct structural and thermosensitive properties [47–50]. Similarly, erythromycin-resistant isolates lacking ermA suggest the involvement of other macrolide resistance determinants, particularly ermC [51, 52]. The aacA-D-negative isolates could also exhibit adaptive resistance [53].

Conversely, several isolates harbored resistance genes but remained phenotypically susceptible. MecA-positive but oxacillin-susceptible isolates may be explained by insufficient PBP2a expression or compensatory native PBP activity [54]. The lack of phenotypic β-lactam resistance in some blaZ-positive isolates, as well as susceptibility in isolates carrying tetK, tetM, aacA-D, linA, norA, or msrB, is consistent with previous reports describing gene silencing, regulatory suppression, sequence variation, or gene activation [55–59]. In addition, the presence of the blaZ gene did not always indicate β-lactam susceptibility, as mecA may also mediate resistance [55]. The phenotypic susceptibility of some linA-positive isolates may result from mutations in the 50S ribosomal subunit that permit antibiotic binding [57]. Similarly, norA detection in ciprofloxacin-susceptible isolates may reflect sequence variation or modulation by global regulators such as MgrA [58]. Finally, the discordance between msrB detection and erythromycin susceptibility may be attributable to gene mutations or functional overlap with other MLSB resistance determinants, including the erm and linA genes [59].

The overall MDR genotypic pattern (Figure 7) corresponded well with the phenotypic resistance profiles. MDR S. aureus was detected across multiple animal sources, particularly in isolates from bovine milk and cats. The emergence of MDR strains is strongly linked to inappropriate and excessive antibiotic use in both livestock and companion animals. In veterinary practice, commonly used antimicrobial classes include fluoroquinolones, aminopenicillins (with or without β-lactamase inhibitors), cephalosporins, tetracyclines, sulfonamides, macro-lides, and glycopeptides for the treatment of bacterial infections [60]. In many regions, antibiotic administration often occurs without laboratory diagnosis, adherence to withdrawal periods, or veterinary prescription, as reported in Rwanda, where over 85% of treatments were empirical or extra-label [61]. These practices significantly contribute to the selection and spread of MDR S. aureus strains. The detection of multiple resistance genes in phenotypically susceptible isolates suggests the presence of latent resistance reservoirs that may be activated under selective pressure from antimicrobials. Silent or low-expression genes represent an important ecological risk, particularly in environments with frequent or inappropriate antibiotic exposure. The shared distribution of resistance genes across livestock and companion animals likely reflects overlapping ecological niches and common selective pressures, reinforcing the role of animals as interconnected reservoirs of AMR within a One Health framework.