All animal procedures were conducted in accordance with Indonesia’s national guidelines for animal welfare and biosafety for transboundary animal diseases. The study protocol, including field sampling, animal handling, and laboratory processing of infectious materials, was reviewed and approved by the Ethical Clearance Committee of the National Research and Innovation Agency (BRIN), Indonesia (approval no. 065/KE.02/SK/10/2022, October 6, 2022). Only naturally infected animals identified during the FMD outbreak were included; no experimental infection, disease induction, or invasive procedures were performed. Oropharyngeal epithelium and lesion scrapings were collected by licensed veterinarians and trained animal health officers following World Organisation for Animal Health (WOAH) guidelines for safe FMD sampling. Animals were handled using minimal-restraint techniques without sedation to avoid pain or distress. Farmers or animal owners provided prior informed consent, and all activities were supervised by the District Livestock Services in West Java and South Sumatra. Sample collection was conducted using appropriate personal protective equipment, and transport complied with national regulations for infectious substances. Laboratory work, including RNA extraction, cDNA synthesis, polymerase chain reaction (PCR) amplification, and sample storage, was performed in a BSL-2 facility with Class II biological safety cabinets. Waste materials were decontaminated by autoclaving or chemical disinfection. The study adhered to Animal Research: Reporting of In Vivo Experiments 2.0, the WOAH Terrestrial Manual (Chapter 3.1.8), and the Three Rs principles.

Samples were obtained from suspected FMD cases based on clinical signs (hypersalivation and vesicular lesions of the oral cavity and feet) in West Java and South Sumatra, Indonesia, during 2022–2023 (Figure 1). Sampling was conducted on smallholder cattle farms raising local breeds (Bos indicus × Bos taurus; Peranakan Ongole, crossbred, and Bali), with herd sizes of approximately 50–100 animals.

The methodology used in this study is presented in Figure 2. Animals with clinical scores of 2 and 3 (moderate and severe, respectively) presenting vesicles and ulceration of the oral cavity and interdigital spaces of the feet, hypersalivation, and lameness were selected. A total of 28 clinical samples (oral and foot epithelium) were collected and preserved in RNA Shield (Zymo Research, Irvine, CA, USA) (catalog no. R1100-50). The samples were transported in insulated containers with ice packs at approximately 4°C, with temperature validation using temperature logs, over 48–72 h. RNA Shield (Zymo Research) was used to protect nucleic acids at ambient temperature for up to 7 days. All samples were stored at −20°C in the laboratory for further analysis.

Total RNA extraction was performed according to the manufacturer’s instructions using the Quick-RNA Miniprep Kit (Zymo Research) (catalog no. R1054), with a total RNA elution volume of 30 µL. RNA quality and concentration were evaluated using a NanoDrop spectrophotometer, and RNA with an A260/280 ratio of 1.8–2.0 was selected for further processing. cDNA synthesis was performed using the RevertAid cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) (catalog no. K1622) to generate complementary DNA. Negative controls (RNase/DNase-free water) were included to exclude contamination. The reaction setup involved incubation at 65°C for 5 min, followed by cDNA synthesis at 25°C for 5 min, 42°C for 60 min, and enzyme inactivation at 70°C for 5 min.

The cDNA was amplified by conventional PCR using DreamTaq PCR Master Mix (Thermo Fisher Scientific) (catalog no. K1081) with 3D primers for FMD diagnosis [7]. PCR conditions included initial denaturation at 95°C for 5 min, followed by 35 cycles of denaturation at 95°C for 10 s, annealing at 60°C for 20 s, and extension at 72°C for 30 s, with a final extension at 72°C for 5 min. The capsid genes in positive samples were analyzed using primers designed based on the FMDV serotype O reference sequence (MN095357.1) (Table 1). Gene-specific annealing temperatures were applied for VP1, VP2, and VP3. PCR amplification was performed in triplicate, and amplicons were pooled. RNase/DNase-free water was used as a negative control. Amplicons were confirmed on 1.5% agarose gels and purified using a PCR Clean-Up Kit (Geneaid, New Taipei City, Taiwan) (catalog no. DFC100). Purified products were sequenced using Sanger sequencing (BigDye Terminator v3.1, 1st BASE), and sequence identity was confirmed using BLAST against the NCBI GenBank database.

Sequence data were analyzed using MEGA-X v10.2.6. Low-quality regions were trimmed based on chromatogram inspection, Phred quality scores <Q20 were excluded, and ambiguous terminal bases were removed. Sequences were aligned using the MUSCLE algorithm. All complete VP1 sequences were deposited in GenBank on October 2, 2025 (accession nos. PX406137–PX406155).

Aligned VP1 nucleotide and amino acid sequences were analyzed using MEGA-X v10.2.6 and compared with reference FMDV isolates [8]. Phylogenetic trees were constructed using the maximum likelihood method under the generalized time-reversible model with the Kimura 2-parameter substitution model. Bootstrap analysis was conducted with 1000 replicates, and gaps were treated using partial deletion.

Translated VP1 amino acid sequences were analyzed using IEDB tools to predict MHC class I–binding epitopes employing NetMHCpan 4.1 EL [9]. BoLA alleles were prioritized. Linear B-cell epitope prediction was performed using IEDB B-cell tools (threshold 0.5) [10]. Experimentally validated high- and low-immunogenic peptides were used as positive and negative controls, respectively.

Antigenicity was evaluated using VaxiJen v2.0 [11]. Allergenicity was assessed using AllerTOP v2.1 and AllergenFP v1.1 [12, 13]. Toxicity prediction was conducted using ToxinPred [14]. Predicted epitopes were benchmarked against experimentally validated epitopes.

Ligand-binding site prediction and 3D visualization were performed using PrankWeb [15]. VP1 structural models were generated using Swiss-Model and assessed using standard quality parameters. Binding pockets were ranked based on calibrated probability scores.

Molecular docking between VP1 and TLR7 was performed using the HDOCK server [16]. Docking scores and confidence scores were used to evaluate binding probability, and docking complexes were visualized using Swiss-PdbViewer v4.1.0.

Local FMDV serotype O strains were obtained from naturally infected cattle in South Sumatra and West Java during the 2022–2023 outbreak period (Table 2). PCR amplification was performed using serotype O–specific primers targeting the capsid gene coding regions VP1, VP2, and VP3. Sequencing quality was evaluated using Phred quality scores, and only reads meeting a Q20 threshold were included for downstream analysis. Comparative sequence analysis demonstrated that the VP1, VP2, and VP3 genes showed reduced nucleotide identity relative to the most recent Indonesian reference strain, FMDV serotype O isolate ISA/1/2022 (Table 3). For VP1, samples Depok 2, Palembang 8, Sukabumi 1, and Sukabumi 2 exhibited the lowest sequence identity (98%). A similar pattern was observed for VP2, with Sukabumi 1 and Sukabumi 2 also showing 98% identity. In contrast, VP3 identity decreased to 98% in most samples, except for Depok 1 and Bogor 1, which retained higher similarity (99%).

Phylogenetic reconstruction based on nucleotide substitution models classified the recent FMDV serotype O isolates into three principal clusters. Cluster I comprised Depok 2, Sukabumi 1, Sukabumi 2, and reference FMDV serotype O strains from the database (O/ME-SA/Ind-2001d, Myanmar, India, Oman, Kuwait, Iran, and Bhutan). Cluster II included Bandung 9 and Palembang 1, 2, 3, 4, 6, 7, and 8. Cluster III consisted of Bogor 1, Bandung 1, 3, 4, 5, 8, and Depok 1, which showed close genetic relatedness to Indonesian reference isolates ISA/Gresik/2022 and ISA/Banjarnegara/2022. In contrast, phylogenetic analysis based on amino acid substitutions resolved only two main clusters: Cluster I included Depok 2, Sukabumi 1, Sukabumi 2, and related database isolates, whereas Cluster II comprised the remaining samples together with ISA/Gresik/2022 and ISA/Banjarnegara/2022 (Figure 3).

Comparative analysis of VP1 sequences identified multiple nucleotide variations relative to Indonesian and other serotype O reference sequences, particularly at positions 10, 37, 84, 102, 103, 171, 219, 225, 386, 408, 451, 459, 462, 465, 515, 526, 582, 589, 615, and 627 (Table 4). These substitutions resulted in amino acid changes at positions 34, 113, 129, 172, and 197 (Table 5).

Based on MHC class I–binding predictions, the VP1 protein yielded eight peptides capable of binding T-cell epitopes associated with BoLA-1, BoLA-2, BoLA-3, BoLA-4, BoLA-5, BoLA-6, BoLA-T5, and BoLA-T7 alleles. Compared with Ind/2022, ISA/Banjarnegara/Gresik, and other FMDV serotype O isolates, samples from Bandung 1, 3, 4, 5, 8, 9, and Palembang 8 exhibited distinct T-cell epitope binding profiles. Samples Depok 2 and Sukabumi 1 and 2 differed from Ind/2022 and ISA/Banjarnegara/Gresik only at amino acid positions 4 (alanine) and 13 (tyrosine), localized within residues 2–15 of VP1 corresponding to the BoLA-T7 allele. No epitope differences were observed for other alleles. All predicted peptides were antigenic and non-toxic, although several were classified as probable allergens (Table 6).

B-cell epitope prediction indicated that Sukabumi 1, Sukabumi 2, and Depok 2 differed from ISA/2022 and ISA/Banjarnegara/Gresik by an alanine-to-threonine substitution at position 13 within the 6–30 epitope region. These profiles were otherwise consistent with other FMDV serotype O database isolates. Notably, Palembang 8 exhibited a glutamic acid–to–lysine substitution at position 196 within the 192–205 epitope region, distinguishing it from Ind/2022, ISA/Banjarnegara/Gresik, and other reference isolates. Peptides spanning positions 93–107 were predicted to be non-antigenic across all samples, whereas other regions demonstrated antigenic potential. Several peptides showed predicted allergenicity, while all remained non-toxic (Table 7).

Protein–ligand binding site analysis of VP1 revealed notable variation among isolates (Table 8). Palembang 8 exhibited the highest number of predicted binding pockets, enriched in threonine, tyrosine, glutamine, and asparagine residues capable of forming hydrogen bonds. Differences in pocket composition between ISA/Gresik/2022, ISA/Banjarnegara/2022, and the field samples suggested altered ligand-binding specificity. Structural visualization showed five high-probability binding pockets in Palembang 8, whereas other isolates displayed fewer than five pockets (Figure 4). Binding pockets ranked 1 or 2 were prioritized as the most probable functional sites.

Molecular docking analysis demonstrated variable interaction strength between VP1 and TLR7. Palembang 8 exhibited the lowest docking score and highest confidence score, indicating the strongest predicted interaction with TLR7 (Table 9). A more negative docking score reflected thermodynamically favorable binding, while confidence scores >0.7 indicated a high-probability of interaction; scores between 0.5 and 0.7 suggested possible binding, and values <0.5 indicated unlikely interactions. Palembang 8 formed a dominant docking cluster comprising 65 structures, indicating a stable and reliable binding mode. The associated Z-score (−1.2) confirmed significantly higher stability compared with other clusters. Energy decomposition analysis revealed strong electrostatic and van der Waals contributions, consistent with dense hydrogen-bonding networks and optimal shape complementarity. Residue-level interaction analysis further showed that Palembang 8 had the highest proportion of hydrogen bonds with TLR7, supporting its enhanced binding specificity (Table 10). Visualization of the VP1–TLR7 complex confirmed that key binding epitopes were surface-exposed, facilitating effective receptor engagement (Figure 5).

FMD remains the most damaging viral disease affecting cattle production in Indonesia. Although a national FMD control program has been implemented using an inactivated vaccine containing serotype O strains O/Mya98/XJ/2010, O1 Campos, O1 Manisa, and IND/O/R2/75 (Decree of the Minister of Agriculture of the Republic of Indonesia No. 738/KPTS/PK.300/M/10/2022), FMD continues to circulate within the country. The vaccination campaign has reduced the frequency of clinical disease; however, unregulated animal movement facilitates the ongoing circulation of FMDV strains. In addition, Indonesia’s geographic position increases the risk of incursion of circulating strains from other Asian countries. Consequently, serotype O remains endemic and continues to circulate in Indonesia [6].

This study presents the first molecular analysis of VP1, VP2, and VP3 sequences from the 2022–2023 outbreaks in South Sumatra and West Java. The nucleotide identity of VP1 decreased to 98.26–99.05% in several samples, while VP3 identity declined to 98.48% in multiple isolates. These findings suggest ongoing antigenic drift that may influence vaccine matching. Phylogenetic analysis based on VP1 nucleotide and amino acid sequences demonstrated that the collected samples formed clusters distinct from recent Indonesian reference isolates ISA/Banjarnegara/2022 and ISA/Gresik/2022. Specifically, three clusters were identified using nucleotide variation and two clusters using amino acid sequences. This pattern contrasts with previous reports in which 2022 Indonesian isolates clustered tightly within Ind-2001d or Mya-98 lineages, highlighting novel intra-country antigenic diversification within a single outbreak period. Multiple nucleotide and amino acid substitutions were detected relative to ISA/Banjarnegara/2022 and ISA/Gresik/2022, including changes at amino acid positions 4, 13, 129, 172, and 197 (Table 6). These substitutions are located near known immunogenic regions, including the G–H loop, B–C loop, and N-terminus of VP1, and may influence receptor binding and antigen recognition.

The observed genetic heterogeneity of serotype O FMDV across different regions is likely driven by unrestricted livestock movement. Endemic co-circulation of multiple genotypes may facilitate recombination when closely related strains co-infect a host [17]. As an RNA virus lacking proofreading mechanisms, FMDV evolves rapidly, with mutation rates ranging from 10^-3 to 10^-5 substitutions per nucleotide site per year [18]. Each viral population exists as a quasispecies comprising genetically and antigenically diverse variants [19]. Because the principal immunogenic region of FMDV is located within the 1D region, VP1 remains the most informative gene for investigating viral variation and evolutionary relationships among isolates [20].

FMDV evolution predominantly occurs through point mutations, and the extent of genetic change influences viral adaptability, transmissibility, and pathogenicity [21, 22]. The emergence of new strains may compromise disease control in endemic regions when vaccine strains fail to provide adequate cross-protection [23]. Moreover, current vaccines do not prevent the establishment of a carrier state, as infected cattle and buffaloes may excrete the virus for extended periods regardless of vaccination status. Although mortality is uncommon, FMD causes severe disease, and vaccination primarily reduces clinical severity rather than preventing infection or halting transmission.

Epitope prediction analysis of VP1 revealed variation in peptide binding to host T- and B-cell receptors among the collected samples compared with database sequences. All predicted peptides were antigenic and non-toxic, although some showed potential allergenicity based on in silico analysis. The BLV envelope protein (UniProt: P03380) served as a positive control and was predicted to be antigenic, non-allergenic, and non-toxic, supporting the robustness of the analytical pipeline. Certain isolates exhibited unique T-cell binding shifts, such as BoLA-T7–associated changes linked to A4T and A13T substitutions. Given the genetic diversity of Indonesian cattle populations, allele-specific immune responses may influence vaccine performance at regional levels. These findings indicate host-allele-dependent immunogenic divergence, which has not been previously reported for Indonesian FMD outbreaks. Key B-cell antigenic sites of FMDV are located within surface-exposed structural motifs of the capsid, including the N-terminus, B–C, E–F, F–G, G–H loops, and the C-terminus [24]. The conserved RGD motif within the G–H loop of VP1 is essential for integrin-mediated viral entry [25], although linear peptides spanning this region provide limited protection in natural hosts [26]. While neutralizing antibodies are central to protective immunity, T-cell responses are also required for robust immune protection [27], mediated through peptide presentation by MHC molecules to T-cell receptors [28].

Previous epitope studies of serotype O have primarily focused on neutralizing domains of VP1 at amino acid positions 133, 143–144, 147–149, 151–153, 161, and 164–167 [29], as well as positions 50 and 52 [30] and 199–202 [31]. Computational analyses have predicted multiple B- and T-cell epitopes for serotypes A and O [32], while secondary structure and antigenicity profiling of VP1 has further identified dominant B-cell epitopes [33, 34]. In the present study, ligand-binding pocket analysis revealed notable differences in binding specificity among Indonesian isolates. Palembang 8 exhibited the highest number of binding pockets, enriched with threonine, tyrosine, glutamine, and asparagine residues, distinguishing it from ISA/2022 and other reference isolates. These findings represent the first structural binding-pocket mapping of Indonesian FMDV serotype O isolates.

Protein–protein docking analysis demonstrated that Palembang 8 showed the lowest docking and confidence scores for interaction between VP1 and TLR7, indicating a higher probability of binding [39]. This isolate exhibited extensive hydrogen-bonding, strong electrostatic and van der Waals interactions, and high structural complementarity, suggesting enhanced engagement with innate immune receptors. To our knowledge, this is the first report describing the VP1–TLR7 binding potential of FMDV isolates, providing new insights into innate immune recognition mechanisms. Nevertheless, this study is limited by its reliance on in silico predictions without experimental neutralization assays or empirical antigenicity testing. Future work should incorporate wet-lab validation, refined structural modeling, and epitope conservancy monitoring to strengthen these findings.