Molecular epidemiological surveillance of the feline coronavirus reveals novel recombinant FCoV strains with evidence of ongoing viral evolution

Highlights

• •Whole-genome sequencing identified three FCoV viral genomes (FIPV-170, FIPV-193, and FIPV-246) that belong to FCoV type-I.
• •These FCoV viral genomes shared about 90% nucleotide similarity with FCoV type-I strains and approximately 83% similarity with FCoV type-II strains, indicating substantial genomic drift.
• •Substantial amino acid substitutions were found in the S1 region of the spike glycoprotein, suggesting potential changes in viral attachment and pathogenicity.
• •The Recombination pattern in these isolates closely resembles that reported in the recent Cyprus FIP outbreak, supporting the possibility of continued emergence of recombination FCoV strains in the United States.

Abstract

Feline Infectious Peritonitis Virus (FIPV) is a virulent biotype of feline coronavirus (FCoV) that causes fatal systemic disease in cats and continues to emerge through mutation and recombination. We conducted demographic data collection, molecular surveillance and genomic characterization of FCoV circulating in feral and shelter cats. Among 139 cats screened, 22.3% tested positive for FCoV. Lower weight and Body Condition Scores indicate higher likelihood for FCoV. Whole-genome sequencing identified three novel FCoV (FIPV-170, FIPV-193, and FIPV-246) belonging to FCoV type I. The viral genomes shared ≈90% nucleotide similarity among themselves and ≈83% similarity with local type II strains. Spike gene analysis revealed 76 amino acid substitutions in the S1 region, including 12 mutations in the receptor-binding domain. Accessory gene mutations included a premature stop codon in NSP-3a and truncation of NSP-7a. Recombination analysis identified mosaic genomes derived from feline coronavirus types I and II and canine coronavirus lineages, with breakpoints in ORF1b, spike, and NSP-3abc regions, resembling patterns reported during the recent 2023 FCoV outbreak in Cyprus (FCoV-23). Continued regional molecular and genomic surveillance is advised.

1. Introduction

Feline infectious peritonitis (FIP) is a highly fatal systemic disease of cats caused by virulent variants of feline coronavirus (FCoV) (Lopes et al., 2024), an enveloped positive-sense RNA virus belonging to the genus Alphacoronavirus within the family Coronaviridae (Attipa et al., 2025). FIP infection is widespread among domestic and feral cats and are primarily transmitted through the fecal route (Stranieri et al., 2020) in environments characterized by high animal density, such as shelters, breeding facilities, and feral colonies (Shah et al., 2025).

While most FCoV infections remain subclinical or cause mild enteric disease, mutations and genomic rearrangements within feline enteric coronavirus (FECV) can lead to the emergence of the virulent biotype known as feline infectious peritonitis virus (FIPV) (Andrew, 2000).

The evolution of FIPV is strongly influenced by the high mutation rate and recombination capacity typical of coronaviruses. Recombination events between different FCoV lineages and even between feline and canine coronaviruses have been increasingly recognized as key drivers of viral diversification and emergence of pathogenic strains (Attipa et al., 2025; Decaro et al., 2021).

In recent years, the global epidemiology of FIP has gained renewed attention following reports of unusually large outbreaks associated with recombinant FCoV variants. One notable example occurred in Cyprus, where a rapidly spreading epidemic caused high mortality among domestic and stray cats and was linked to the emergence of recombinant FCoV strains exhibiting unique genomic signatures (Attipa et al., 2025).

The viral genomes of most coronaviruses including FECV and FIPV have poor proof-reading RNA polymerase enzymes (Ba Abduallah and Hemida, 2021; Hemida and Alnaeem, 2019; Hemida and Ba Abduallah, 2020; Hemida et al., 2023) . Their genomes are prone for frequent changes due to some point mutations, genetic drifts or shifts in addition to the possibility of recombination with other coronaviruses particularly canine coronavirus (Ba Abduallah and Hemida, 2021; Hemida and Alnaeem, 2019; Hemida and Ba Abduallah, 2020; Hemida et al., 2023). The next generation sequencing (NGS) advanced our knowledge of the genetic materials of the viruses affecting various species of companion, and livestock animals as well as birds (Ba Abduallah and Hemida, 2021; Al-Mubarak et al., 2024; Al-Mubarak et al., 2023; Al-Mubarak et al., 2023; Gomaa et al., 2008; Hemida et al., 2020; Hemida and Alnaeem, 2022; Hemida et al., 2019; Hemida et al., 2020; Hemida et al., 2017; Hemida et al., 2019; Hemida et al., 2018). The NGS enables us to monitor these viral genomes and identify any alteration in the genetic material of these viruses. This approach also enables us to link any change in the virulence/tropism/pathogenesis of any of these viruses with the genetic materials of these viruses. It will also help in fine-tuning the current vaccines/diagnostic assays and the management protocols of these viruses for the better results.

The integration of the machine learning, generative artificial intelligence tools and the genome sequence decoded using the NGS showed promising trends in the design of novel antiviral therapy and vaccines against many viral diseases of animals and birds including the FIPV (Al-Kubati et al., 2023; Goud et al., 2025; Khan et al., 2025; Khan et al., 2025; Zhang et al., 2014).

Despite the importance of these events, limited information is available regarding the genomic diversity and evolutionary dynamics of circulating FIPV strains in North American feral cat populations (Shah et al., 2025; McKay et al., 2020). Feral and shelter-associated cats represent important viral reservoirs where high-density living conditions facilitate viral transmission, mutation, and recombination (Shah et al., 2025; McKay et al., 2020; An et al., 2011) . Consequently, these populations may serve as sources for the emergence of novel FIPV variants with enhanced pathogenicity or transmission potential (Shah et al., 2025; McKay et al., 2020; An et al., 2011).

In this study, we conducted a comprehensive molecular surveillance and genomic characterization of FCoV detected in feral and shelter-associated cats in the downstate New York region of the United States. Through whole-genome sequencing, phylogenetic reconstruction, and recombination analysis, we sought to identify the genetic determinants underlying the evolution of circulating strains of FCoV. Importantly, we report that the recombinant patterns observed in the identified viral genomes share striking similarities with those previously associated with the Cyprus FIP outbreak (Attipa et al., 2025). These findings highlight the ongoing evolution of FCoV and underscore the need for continuous genomic surveillance to detect emerging virulent strains that may pose significant risks to feline populations.

2. Materials and methods

2.1. Demographic data collection

Clinical samples and demographic information were collected from 139 cats, on-site at multiple veterinary clinics. The cats were sourced from many cat shelters, rescues, owners, and trappers throughout the downstate New York area, ranging from the 5 boroughs and throughout the region of Long Island, New York. Prior to a cat’s appointment, an informed consent form was given, verbally explained, and signed off on by the clients. All data collected from the demographic risk-associated factors were as follows: collection date, collection clinic, name, clinic ID number, town, state, zip code, county, gender, approximate age, breed, weight in pounds, body condition score out of 5, microchip status, surgical history, pregnancy status, endoscopic parasite status, vaccination for Rabies, vaccination for Feline Viral Rhinotracheitis C – Calicivirus P – Panleukopenia (FVRCP), and owned or feral status (Supplementary Table 1).

2.2. Samples collection, processing and preservation

Oropharyngeal and rectal swab samples were collected from all 139 cats. All swabs samples were collected in a sterile 4 mL of viral transport media (VTM) consisting of 1% penicillin/streptomycin antibiotic (Thermo Fisher Scientific, REF# 15,140–122), 1% antibiotic-antimycotic antifungal 100X (Thermo Fisher Scientific, REF# 15,240–096), 2% heat inactivated Horse Serum (HS) (Thermo Fisher Scientific, REF# 26,050–088), and 96% Dulbecco's Modified Eagle Medium (DMEM) (Corning, CAT# 10013CV). All swab samples were stored in the −80 °C freezer until further use.

2.3. RNA extraction and qRT-PCR

Total RNA was extracted from combined oropharyngeal swab rectal swab fluid from each individual cat’s and were treated with TRIzol LS Reagent (Invitrogen, REF# 10,296,010) following the manufacturer’s guidelines. The RNA was converted into complementary DNA (cDNA) after subsequently using the High-Capacity Reverse Transcription Kit (Applied Biosystems, LOT# 2902,953 s). Next, qRT-PCR was performed using the PowerUp™ SYBR™ Green Master Mix (Applied Biosystems, REF# A25742) following manufacturer’s protocol. The oligonucleotides used for qRT-PCR for FCoV, and β-actin are listed in Table 1. The β-actin was used as the housekeeping gene to normalize viral genome expression. The qRT-PCR amplification was performed using the QuantStudio3 System (Applied Biosystems). Samples negative for FCoV were used as a control. The genomic viral expression was normalized following the 2−ΔΔCt method (Livak and Schmittgen, 2001).

Table 1.

Information of the oligonucleotides used for the molecular surveillance of the FCoV in the field cat samples by the qRT-PCR.

Virus	Direction		Primer Sequence
FCoV		Forward	5′-AGCAACTACTGCCACRGGAT-3´
		Reverse	5′-GGAAGGTTCATCTCCCCAGT-3´
β-actin		Forward	5´-ACTGGGACGACATGGAGAAG-3´
		Reverse	5´-TGGGGTGTTGAAAGTCTCGA-3´

2.4. The full-Length genome sequences of the FCoV viral genomes decoding by the generation sequencing (NGS)

The swab samples with high FCoV viral load via qRT-PCR were subjected to NGS. Total RNA samples were treated with DNase I (New England Biolabs; Lot #10,213,692) following manufacturer’s instructions. The NGS procedure was carried out by Indiana University, Molecular biology lab. Briefly, RNA Libraries were prepared by the Illumina Stranded mRNA Prep protocol (catalog no 20,040,534) with the poly-A selection step omitted to reduce bias towards the 3′ end. The RNA is fragmented into smaller pieces to prepare them for cDNA synthesis. Using random primers and a reverse transcriptase enzyme, the RNA fragments are copied into first-strand cDNA. For second-strand cDNA synthesis, thymine is replaced with uracil using dUTP, which enables the identification of the RNA transcript's original strand and ensures strand specificity. An adenine base is added to the end of the fragmented cDNA, and then adapter sequences are ligated, providing the necessary primer binding sites for downstream PCR and sequencing. The adapter-ligated cDNA is purified to remove excess reagents, and then it is amplified using PCR to create enough sequencing-ready library. The libraries were analyzed by Agilent 4200 TapeStation, then pooled and loaded on a NextSeq 1000/2000 P2 XLEAP-SBS Reagent Kit (100 Cycles) flow cell (catalog no 20,100,987) configured to generate 2 × 61 nt paired-end reads. The demultiplexing of the reads is performed using bcl2fastq, version 2.20.0. Sequenced reads were processed for adapter removal and quality filtering using Trimmomatic version 0.38 (Bolger et al., 2014), employing a cutoff that required an average base quality score of at least 20 within a sliding window of three bases, and retaining only sequences trimmed to a minimum length of 20 bases (Parameters: ILLUMINACLIP:adapters.fa:2:20:7 LEADING:20 TRAILING:20 SLIDINGWINDOW:3:20 MINLEN:20). Subsequently, the filtered read pairs were assembled using SPAdes version 3.15.5 (Bushmanova et al., 2019). Among the resulting assembled transcripts, viral sequences were distinguished from host transcripts by aligning them to the complete genome sequence of Feline infectious peritonitis virus (NC_002306.3) using NUCmer version 3.1 from the MUMmer package (Kurtz et al., 2004) and were visualized using the (mummerplot version 3.5: https://github.com/mummer4/mummer/issues/40). Full length of the three FCoV identified in this study were submitted to NCBI with the accession numbers PPX999723, PX999722, and PX999724 for FIPV-170, FIPV-193, and FIPV-246, respectively. The terms FIPV in FIPV-170, FIPV-193, and FIPV-246 are identifiers and do not indicate clinically confirmed FIPV viral genomes.

2.5. Multiple sequence alignment, pairwise comparison and phylogenetic analysis of the novel FCoV viral genomes reported in this study

Total of 110 isolates, including three novel FCoV viral genomes, FCoV type-I/II, FCoV-23, canine coronavirus (CCoV) typeI/II, transmissible gastroenteritis virus (TGEV), porcine respiratory coronavirus (PRCV), and human coronaviruses were used for multiple sequence alignment (MSA) and phylogenetic tree construction (Supplementary Table 2). All the reference sequences were obtained from the NCBI database. These sequences were selected based on FCoV type (type-I/II), country of origin, and year of isolation. The MSA was performed using Geneious Prime V.11.

Phylogenetic tree was constructed using MEGA12 software (Kumar et al., 2024), following maximum likelihood and the Tamura-Nei model (Tamura and Nei, 1993). The bootstrap was set to 1000 replicates for each branch. Nexius file of phylogenetic analysis from MEGA12 were used to construct the final trees using the online software iTOL (Letunic and Bork, 2024). Human coronavirus, TGEV, and PRCV were used as an out-group in all MSA and phylogenetic analyses.

The pairwise comparison based on the nucleotide sequences of three FIPV/FCoV identified in this study was conducted with other reference strains/isolates for full length genome, ORF1ab, spike, NSP-3abc, envelope, membrane, nucleocapsid, and NSP-7ab was performed using Geneious Prime V.11. Heat map showing the nucleotide similarities was constructed using GraphPad Prism v9 software.

2.6. Recombination analysis of the novel FCoV viral genomes

The recombination analysis of FIPV-170, FIPV-193, FIPV-246 was performed separately with reference strains, including FCoV-I strain UG-FH8, FIPV-II strain 79–1146, FCoV-23 strain 2-C11 Re, CCoV-I strain 23/03, and CCoV-IIa CB/05 using RDP5 software (Martin et al., 2021). Each analysis was performed using seven detection methods (RDP, GENECONV, Bootscan, Chimaera, SiScan, MaxChi, and 3Seq). The recombination events positive by at least five methods with a p-value threshold of 0.01 and further supported by the phylogenetic tree analysis were considered for further analysis. The phylogenetic trees were constructed using the maximum likelihood method, with a bootstrap of 1000 replicates using RDP5 software. The likely recombination events were verified by Bootscan and Similarity analysis in SimPlot Version 3.5.1 (Lole et al., 1999). Parameters fro RDP5 and Simplot were set to a window size of 400 bp, step size of 40 bp, GapStrip was On, Kimura (2-parameter) substitution model, and T/t was 2.0.

2.7. The 3D structure prediction of the spike glycoproteins

Protein structural prediction for FIPV-170 spike glycoprotein and receptor binding domain (RBD) was modeled using the SWISS-MODEL structural prediction and analysis server (Waterhouse et al., 2018), using default settings. The spike glycoprotein amino acid sequences of FCoV/CCoV reference strains were obtained from NCBI and used to compare with FCoV identified in this study through multiple sequence alignment using Geneious Prime V.11; (https://www.geneious.com/download-previous-versions).

2.8. Statistical analysis

The descriptive statistics included calculating the proportion of positives, the distribution of strong, weak, and negative tests, and the distribution of Ct-values for all samples. For the proportion of positives, the test results were coded as binary outcomes (positive, Ct-value 〈30, and negative, Ct-value 〉 30), and the exact binomial method with 95% confidence intervals was used to calculate the positive proportion. In addition, the distribution of strong positive (<25), weak positive (Zhang et al., 2014; McKay et al., 2020; An et al., 2011; Livak and Schmittgen, 2001; Bolger et al., 2014), and negative (>30) test results were calculated and illustrated in a graph. Lastly, the distribution of Ct-values for all samples was displayed in a graph to compare it to the positive test threshold (Ct value <30).

At the second step, logistic regression models were used to assess the association between test positivity and predictors, including the cats’ gender, age, weight, and body condition score. Results were reported as odds ratios (ORs) with 95% confidence intervals and p-values. Significant predictors (p ≤ 0.05) were plotted.

At the last step, the samples were geocoded to the 5-digit ZIP Code level, and the number of positive tests at each location was calculated and illustrated in choropleth point maps. In addition, Inverse Distance Weighted (IDW) interpolation was used to estimate the FCoV risk in unsampled locations, based on the proximity of known cases. The interpolated surface was visualized to show the distribution of FCoV cases across the study area and to identify areas with higher concentrations (hot spots) of FCoV cases.

3. Results

3.1. Epidemiology and pathology of FCoV in long island, New York

Among the 139 cats, a positivity rate of 22.30% for FCoV (mean=0.223; 95% CI: 0.157 - 0.301) was observed (Fig. 1A). Most of the samples tested were negative, and only a small proportion (< 12%) were strong or weak positives (Fig. 1B). The distribution of Ct-values for all tested samples indicated that the majority were higher than 30 (Fig. 1C). Univariable logistic regression identified two variables with borderline associations with FCoV positivity. Cats with lower weight had higher odds of testing positive for FCoV (OR = 0.85, 95% CI: 0.71–0.99; p = 0.05). Lower score of the body condition also had a marginal association with increased odds of positivity (OR = 0.20, 95% CI: 0.03–0.88; p = 0.057) (Fig. 1D). Fig. 1E displays the distribution of FCoV‑positive samples in cats across Long Island, New York. Positive cases were detected in both Nassau and Suffolk counties, with higher concentrations in the western portion of the island. Most locations showed one to two positive samples (green points), and a smaller number of sites reported three to four positives (red points). Eastern Long Island had fewer positive locations overall. The map indicates a geographically broad but uneven distribution of FCoV‑positive cases across the region. Fig. 1F shows the inverse distance weighted (IDW) interpolation of FCoV‑positive samples across Long Island, NY. Higher predicted intensities, shown in red and orange, appeared in western and central Long Island, corresponding to areas with clusters of positive tests. Lower predicted values (blue regions) were concentrated in eastern Long Island.

Fig. 1.

Statistical analysis of FCoV breakout in New York. (A) The proportion of FCoV positive samples among 139 cats. The PCR Positivity was determined by using a Ct < 30 threshold. (B) The distribution of negative, strong positive, and weak positive tests in feline diagnostic samples. The FCoV PCR Ct values defined strong positive (<25), weak positive (Zhang et al., 2014; McKay et al., 2020; An et al., 2011; Livak and Schmittgen, 2001; Bolger et al., 2014), and negative (>30) samples. (C) Distribution of FCoV PCR Ct values compared to the diagnostic cutoff of positivity (Ct value <30). (D) Results of univariable logistic regression models evaluating the association between FCoV positivity and body condition score and weight. (E) Distribution of FCoV positive cats across Long Island, New York. (F) Spatial Interpolation of FCoV PCR Positive Test in Cats Across Long Island Using inverse distance weighting (IDW).

3.2. Genome organization of FCoV viral genomes detected in this study

Among the 139 infected feline samples tested for FCoV via qRT-PCR, three samples (170, 193, and 246) were subjected to NGS analysis. The resultant sequences were aligned with reference strains of FCoV type I isolate “Black” (accession number EU186072) and FCoV type II isolate “79–1146” (accession number NC_002306) to organize the genome sequences. The sequences were assembled and deposited to the NCBI GenBank database under the accession numbers FIPV-170 (PX999723), FIPV-193 (PX999722), and FIPV-246 (PX999724). The FIPV-170 was 29,260 nucleotides (nt) in length with a 38.6% GC ratio. The complete size of FIPV-193 was 29,253 nt and FIPV-246 was 29,255 nt with GC percentages of 38.3 and 38.7, respectively. The alignment results showed that the FCoV Black strain was 90.67% like FIPV-170, 90.79% similar to FIPV-193, and 90.52% similar to FIPV-246. Whereas the FCoV 79–1146 strain was 84.53% similar to FIPV-170, 84.32% similar to FIPV-193, and 84.70% similar to FIPV-246. These findings indicates that these three viral genomes identified in this study belong to FCoV type-I. The genomes of these FCoV viral genomes are composed of stranded FCoV/FIPV genomic structure, which includes ORF1ab, four structural proteins (spike, envelope, membrane, and nucleocapsid), and five non-structural proteins (NSP) (3a, 3b, 3c, 7a, and 7b) (Table 2). A truncated NSP3a was observed in FIPV-170, which contains an early stop codon, making a small protein of 63 amino acids (aa) in length.

Table 2.

Genome organization of FCoV viral genomes.

Gene	FIPV-170				FIPV-193				FIPV-246
	Start	Stop	Size (nt)	Size (aa)	Start	Stop	Size (nt)	Size (aa)	Start	Stop	Size (nt)	Size (aa)
5′UTR	1	298	298	-	1	313	313	-	1	296	296	-
ORF1ab	299	20,370	20,073	6690	314	20,388	20,076	6691	297	20,362	20,067	6688
Spike	20,367	24,770	4404	1467	20,385	24,785	4401	1466	20,359	24,762	4404	1467
NSP 3a	24,782	24,970	189	62	24,797	25,009	213	70	24,774	24,986	213	70
NSP3b	24,942	25,163	222	73	24,957	25,178	222	73	24,934	25,155	222	73
NSP 3c	25,160	25,873	714	237	25,175	25,888	714	237	25,152	25,865	714	237
Envelope	25,860	26,108	249	82	25,875	26,123	249	82	25,852	26,100	249	82
Membrane	26,119	26,904	786	261	26,134	26,925	792	263	26,111	26,902	792	263
Nucleocapsid	26,917	28,047	1131	376	26,938	28,071	1134	377	26,915	28,042	1128	375
NSP 7a	28,052	28,357	306	101	28,076	28,330	255	84	28,047	28,352	306	101
NSP 7b	28,362	28,982	621	206	28,335	28,955	621	206	28,357	28,977	621	206
3′UTR	28,983	29,260	278	-	28,956	29,253	298	-	28,978	29,255	278	-

3.3. Phylogenetic analysis of full length of FCoV viral genomes (170, 193, and 246) with reference strains

The phylogenetic analysis was constructed using FCoV viral genomes identified in this study (170, 193, and 246) together with reference strains from FCoV/FIPV type-I, type-II, FCoV from Cyprus in 2023 (FCoV-23), and CCoV strains type-I and type-II. Other alpha coronaviruses, including porcine respiratory coronavirus (PRCV), transmissible gastroenteritis virus (TGEV), and human coronavirus enteric strain 4408, were used as an outgroup. The phylogenetic analysis results showed that the FIPV-170, FIPV-193, and FIPV-246 belong to FCoV type-I and clustered together with FCoV strain UG-FH8 from Belgium in 2015 (Fig. 2).

Fig. 2.

Phylogenetic tree of 110 full-length genome nucleotide sequences. The phylogenetic tree was constructed of 110 full length genomes, including FCoV type-I, FCoV type-II, FCoV-23, CCoV type-I/II, and other alpha CoVs (PRCV, TGEV, and HCoV). The maximum log likelihood of the tree was −541,429.87 with the bootstrap of 1000 replicates. The FCoV viral genomes identified in this study are highlighted in red and indicated with an asterisk. All the strains/isolates are separated based on their genomic types, indicated at the left side of the tree. The HCoV 4408, and 229E were used as an outgroup. The tree was generated using MEGA12 software. The final circular tree and group annotation were constructed using the online software iTOL.

3.4. Phylogenetic analysis of spikes, NSP-3abc, and NSP-7ab of the novel FCoV viral genomes with reference strains

Variations in the spike and accessory genes NSP-3abc and NSP-7ab of FCoV play an important role in virus tropism and infection spread in different biotypes (Pedersen, 2009; Gao et al., 2023). The nucleotide sequences of the spike glycoprotein of the FIPV-170, 193, and 246 showed approximately 85% similarity with FCoV Black strain (EU186072), and only 56% similarity with FIPV strain 79–1146 (NC_002306). Therefore, we performed nucleotide phylogenetic analysis of spike, NSP-3abc, and NSP-7ab of 110 strains/isolates including FIPV-170, 193, and 246, together with reference FCoV/FIPV type-I, type-II, FCoV-23, CCoV-I/II, and other alpha coronaviruses shown earlier. Results showed that the spike of FIPV-170 and FIPV-193 clustered together with FCoV UU30 (HQ392472), and FIPV-246 cluster together with FCoV Black strain (Fig. 3A, 3B). The nucleotide sequence of NSP-3abc of FIPV-246 clusters together with FCoV Black strain (Fig. 3C). Unlike spike, the NSP-3abc of FIPV-170 and FIPV-193 clustered together with FCoV HU-2 (LC880185) from Japan (Fig. 3C). However, the NSP-3abc of FIPV-170 and FIPV-193 showed close resemblance (approximately 94%) with FCoV-23 isolates derived from the Cyprus FCoV outbreak in 2023 (Fig. 3C). The FCoV detected in sample 193 has a short NSP-7a and shows 78.76% similarity to FIPV-170 and 78.10% similarity to FIPV-246. However, all three viral genomes clustered together with the FIPV Aqua strain (LC852817) from Japan in 2024, showing 89.66%, 83.09%, and 87.88% similarity with FIPV-170, 193, and 246, respectively (Fig. 3D).

Fig. 3.

Nucleotide sequence analysis of spikes, NSP-3abc, and 7ab. (A) Schematic representation of the complete structure and organization of FCoV. (B) Phylogenetic tree of spike nucleotide sequences of 111 strains/isolates showing the highest log likelihood of −102,294.56. (C) Phylogenetic tree of NSP-3abc nucleotide sequences of 109 strains/isolates showing the highest log likelihood of −16,530.30. (D) Phylogenetic tree of NSP-7ab nucleotide sequences of 96 strains/isolates showing the highest log likelihood of −12,039.89. All the trees were constructed with the bootstrap value of 1000 replicates. The FCoV viral genomes identified in this study are highlighted in red and indicated with an asterisk. All the strains/isolates are separated based on their genomic types, indicated at the left side of the tree. The trees were generated using MEGA12 software. The final circular tree and group annotation were constructed using the online software iTOL.

3.5. Pairwise comparison of FIPV-170, 193, and 246 with FCoV and CCoV reference strains and other alpha coronaviruses

The nucleotide pairwise comparison of the complete genome, ORF1ab, spike, 3abc, E, M, N, and 7ab was performed between 28 strains/isolates belonging to FCoV type-I, FCoV type-II, FCoV-23, CCoV type-I/II, and other alpha CoVs. The FIPV-170 was used as a reference for comparison with other isolates. The three FCoV viral genomes identified in this study (170, 193, and 246) share only approximately 91% similarity among each other. The full-length genome of these three viral genomes shares about 90% similarity with other FCoV type-I isolates, and about 84% similarity with isolates from FCoV type-II and FCoV-23 groups (Fig. 4). However, they share 79.36% with CCoV type-I (strain 23/03) and about 76% similarity with other CCoV type-II isolates (Fig. 4). The nucleotide sequence of FIPV-170, 193, and 246 spike shares about 86% similarity with each other, about 84% with other FCoV type-I isolates, 56% with FCoV type-II and FCoV-23 isolates (Fig. 4). The spike of these three viral genomes shares 70.85% similarity with CCoV type-I (strain 23/03) and about 55% with other CCoV type-II isolates (Fig. 4) (Supplementary Figure S1A). The NSP-3abc of novel FCoV share about 90–97% similarity with FCoV type-I isolates, 65–89% with FCoV type-II, and 90–95% with FCoV-23 isolates (Fig. 4) (Supplementary Figure S1B, S1C, S1D). These results indicated that the nucleotide sequences of FIPV-170, 193, and 246 share more similarity with FCoV type-I and CCoV type-I, whereas NSP-3abc share more similarity with FCoV type-I and FCoV-23, indicating the possibility of novel recombination between FCoV type-I, FCoV-23, and CCoV type-I.

Fig. 4.

Nucleotide Pairwise analysis of FIPV-170, FIPV-193, and FIPV-246 with reference FCoV/CCoV strains. Pairwise comparison was performed for the full-length genome, ORF1ab, spike, NSP-3a, NSP-3b, NSP-3c, envelope, membrane, nucleocapsid, NSP-7a, and NSP-7b between 28 alpha coronaviruses. The comparison was done between three FCoV viral genomes identified in this study, together with seven isolates from the type-I FCoV group, nine from the type-II FCoV group, two from FCoV-23, four from CCoV type-I/II, and other alpha CoVs including PRCV, and TGEV. The FIPV-170 was used as a standard. The pairwise homology was calculated using Geneious Prime v11 and heatmap was constructed using GraphPad Prism v9 software. The similarity level was indicated by a color gradient.

3.6. Recombination analysis of FIPV-170, 193, and 246

The pairwise comparisons of FIPV-170, 193, and 246 with other FCoV and CCoV strains suggested the presence of potential recombination events in these novel viral genomes. To further explore this possibility, recombination analysis of these novel viral genomes was performed with reference strains, including FCoV-I strain UG-FH8, FCoV-II strain 79–1146, FCoV-23 strain 2-C11 Re, CCoV-I strain 23/03, and CCoV-IIa strain CB/05. Pairwise sequence alignment revealed that the full length of FIPV-170, 193, and 246 shares approximately 90% nucleotide identity with each other. Therefore, recombination analysis was conducted separately for each viral genome with selected reference strains. The possible recombination events were further verified by BootScan and Similarity analysis using Simplot.

The notable recombination events of FIPV-170 were reported at ORF1a (position 3781–5172 nt), ORF1b (position 17,491–18,740), and between nucleocapsid and NSP-7a (position 27,890–28,266 nt). These events showed CCoV-IIa strain CB/05 as a major parent-like and CCoV-I strain 23/03 as a minor parent-like (Supplementary Figure S2A, S2B, S2C). The phylogenetic analysis of FIPV-170 recombination events showed FCoV-I strain UG-FH8 and FCoV-23 strain 2-C11 sequences with a potentially related recombination pattern (Supplementary Figure S2D).

Recombination events of FIPV-193 were reported at ORF1a (position 4293–5106 nt), and at ORF1b (position 16,789–18,883). Like the recombination events in FIPV-170, FIPV-193 also showed CCoV-IIa strain CB/05 as a major parent-like, CCoV-I strain 23/03 as a minor parent-like, and FCoV-I strain UG-FH8, and FCoV-23 strain 2-C11, with evidence of a potentially related recombination pattern (Supplementary Figure S3A, S3B, S3C, S3D).

In FIPV-246, the most important recombination events were reported at position 19,289 – 26,950 nt, which covers a small part of ORF1b, spike, NSP-3abc, envelope, and half membrane gene, which show FCoV-II strain 79–1146 as a major parent-like, CCoV-IIa strain CB/05 as a minor parent-like (Fig. 5A, 5B, 5C, 5E). The phylogenetic analysis showed FCoV-I strain UG-FH8 together with FCoV-23 strain 2-C11, exhibited evidence of a potentially related recombination pattern as FIPV-246 (Fig. 5D). Together, these findings revealed that the novel FCoV field viruses possess multi-segment genome, shaped by recombination of FCoV-II and CCoV-IIa related linages. Furthermore, these three novel FCoV share potentially similar recombination events with FCoV-I and FCoV-23 strain detected from Cyprus in 2023 (Fig. 5A, 5B, 5C, 5E).

Fig. 5.

Recombination analysis of full-length genome of FIPV-246 with reference strains. (A) Bootstrap support, distance plot, (B) pairwise identity analysis, and (C) similarity analysis of FIPV-246 recombination with reference strains. Bootstrap support and pairwise identity analysis were performed using RDP5 software, while similarity analysis was performed using Simplot software. The FCoV genomic map is shown at the top, with the recombination region highlighted in red. The potential recombinant (FIPV-246), potential major parent-like (FIPV-II), and potential minor parent-like (CCoV-IIa) are shown at the right side and distinguished by different colors. The potential recombination region is marked with a red dotted box. (D) The phylogenetic tree of the recombinant region (position 19,289 – 26,950 nt) was constructed using the maximum likelihood method with a bootstrap of 1000 replicates. The potential recombinant is shown in red, potential major parent-like is shown in green, and minor parent-like are highlighted in grey, and the sequence with evidence of potentially related recombination pattern are shown in pink in the tree and indicated in the legends. (E) Schematic diagram of major recombination observed in FCoV-I, FCoV-II, FCoV-23, and CCoV-II following the previous study (Attipa et al., 2025), and recombination events reported in FIPV-170, FIPV-193, and FIPV-246.

3.7. Mutations in the spike glycoprotein of novel field FCoV viral genomes

The pairwise comparisons results showed that FIPV-170, 193, and 246 possess about 86% similarity between each other in spike glycoprotein (Fig. 4) (Supplementary Figure S1A). Therefore, we observed amino acid changes among these three viral genomes and with other reference strains from FCoV-I, FCoV-II, FCoV-23, and CCoV-I/II, especially in the S1 of the spike glycoprotein (Fig. 6) (Supplementary Table 3). Multiple sequence alignment results showed 76 novel amino acid substitutions in FIPV-170, 193, and 246 viral genomes when compared to other reference strains. A total of 12 amino acid substitutions were reported in receptor binding domain (RBD) of spike glycoprotein (Fig. 6A, 6B). Among these substitutions, three mutations (P/S571T, S575A, and T594S) were reported in the RBD of FIPV-170 (Fig. 6B, 6C, 6D). Five amino acid substitutions (K561E, D564N, E572K, V/N664F, and Y677F) were reported in the RBD of FIPV-193 (Fig. 6B, 6D). Moreover, four novel amino acid substitutions (L596I, K608R, L/V624I, and A681S) were reported in the RBD of FIPV-246 (Fig. 6B, 6D).

Fig. 6.

Identification of variations in the spike glycoprotein of novel FCoV field viruses. (A) Schematic representation of FCoV spike glycoprotein and its genome organization. The map was constructed following FIPV-170 spike glycoprotein amino acid sequences. The S1 subunit of spike glycoprotein is composed of N-terminal domain (NTD), and C-terminal domain (CTD), with receptor binding domain (RBD) region in between CTD. The S2 subunit contain fusion domain with fusion peptide (FP), Heptad Repeat 1 (HR1), Heptad Repeat 2 (HR2), and transmembrane (TM) domain at the C-termina end. The S1/S2 furin cleavage site and S2 cleavage site are shown with red arrow head. (B) Multiple sequence showing the amino acid variations in the RBD region and TM domain. (C) Protein structure of spike glycoprotein of FIPV-170. The RBD region is indicated in black dotted box. (D) Protein model of RBD showing the amino acids variations. The amino acids variations in FIPV-170 are shown in green, FIPV-193 in red, and FIPV-246 in blue. The protein models were generated using online SWISS-MODEL server.

3.8. Amino acid sequences in the S1/S2 furin-cleavage site of the novel FCoV field viruses

The furin-cleavage site (FCS) at the S1/S2 domain is the main determinant of FCoV biotype switching and disease development (Jaimes et al., 2020; Zehr et al., 2023). Our multiple sequence alignment showed that the S1/S2 FCS of novel feline field viral genomes are closely related to other reference strains in the FCoV type-I (Fig. 7A). Specifically, the FIPV-170 and FIPV-193 possess 788-RRSRR-793, similar to the FCoV-I UU2 strain. In contrast, the CCoV-IIa strain CB/05, and American CCoV isolates identified in 2024 (CCoV 1–71 and UCD-1) share a similar sequence with FCoV-II at S1/S2 FCS (Fig. 7A). Whereas CCoV-I strain 23/03 and huntervalley_6 exhibit an S1/S2 FCS pattern more closely related to FCoV-I isolates. The S1/S2 FCS of FIPV-246 show substation of S to A at 790, and T to V at position 794, similar to CCoV-I (Fig. 7A). Consistent with this observation, sequence similarity between CCoV-I and FIPV-246 was also observed in the S1-NTD and S1-RBD (Supplementary Table 3) (Fig. 6B). Furthermore, multiple sequence alignment was performed on 111 strains/isolates, including FIPV-170, FIPV-193, FIPV-246, FCoV-I, FCoV-II, FCoV-23, and CCoV-I/II, to identify the frequency of amino acids at S1/S2 FCS. The most prevalent motif identified was 788-PNFYYYS-794, followed by 788-RRSRRST-794 (Fig. 7B) (Supplementary Figure S4).

Fig. 7.

Comparison of S1/S2 furin-cleavage site of FCoV/CCoV strains/isolates. (A) Multiple alignment of 17 strains/isolates belongs from FCoV type-I/II, FCoV-23, CCoV type-I/II and FCoV viral genomes identified in this study (FIPV-170, 193, 246). (B) Motif logo showing the frequency of the amino acids at the S1/S2 furin-cleavage site conducted in 111 strains/isolates. The diagram was generated using an online server SPlot (Tang et al., 2023).

4. Discussion

In this study, a positivity rate of 22.30% among 139 cats in Long Island, New York, was observed, which is comparable to other studies conducted in North America and worldwide (Moyadee et al., 2024; Izes et al., 2020; Kokkinaki et al., 2023). Most cats in our study were tested in multi-cat environments (shelters), and cats with lower body weight and body condition scores had higher odds of testing positive for the virus. These findings suggest that malnourished cats might have a weakened immune system that predisposes them to viral infections (Solikhah et al., 2024), and living in a multi-cat setting with high viral loads might facilitate the viral transmission (Felten and Hartmann, 2019).

A spatial heterogeneity of FCoV-positive cases was observed in this study, with higher concentrations in the western portion of the island. Most of the sampling sites showed one to two positive samples and a smaller number of sites reporting three to four positives, suggesting an uneven distribution and different transmission patterns of the virus. This finding might be explained by the variations in the FCoV risk factors, including differences in cat population density, disease prevention practices in multi-cat settings (shelters) (Drechsler et al., 2011), environmental conditions (Reissner et al., 2023), and geographical variations in viral strains (Lauzi et al., 2020).

The FCoV viral genomes identified in this study (170, 193, and 246) are FCoV strains carrying FIPV-associated genomic features rather than definitively confirmed FIPV isolates. Previously, we identified two FIPV isolates, FIPV-A15 (PQ192636) and FIPV-A37 (PQ202302), in New York (Shah et al., 2025). The FIPV-A15 and FIPV-A37 isolates belong to the FCoV type-II and shared 99.83% similarity with each other. The FIPV-A15 and FIPV-A37 were confirmed as confirmed FIPV isolates through comprehensive clinical evaluation, histopathology, immunohistochemistry, and molecular characterization. In the present study, these three FCoV viral genomes were isolated from the same region of New York, and the infected cats showed the same clinical signs as FIPV-A15 and FIPV-A37. Therefore, FIPV-170, 193, and 246 were considered FCoV strains carrying FIPV-associated genomic features. Unlike FIPV-A15 and FIPV-A37, these three viral genomes belong to the FCoV type-I and share approximately only 83% nucleotide similarity with FIPV-A15 and FIPV-A37. More interestingly, FIPV-170, 193, and 246 have huge differences in the nucleotides between each other and share only about 90% similarity. Indicates that recombination between feline and canine coronaviruses happened most frequently, but their spreading frequency is low, as shown previously in several reports (Attipa et al., 2025; Ntafis et al., 2012; Decaro et al., 2007).

Mutations in the spike, NSP-3abc, and NSP-7ab correlate with the biotype switch between FECV to FIPV (Gao et al., 2023; Jaimes et al., 2020; Zehr et al., 2023; Brown, 2011), resulting in tissue tropism and disease spread. The furin-cleavage site (FCS) at the S1/S2 interface of spike glycoprotein is the main factor of FCoV biotype switching, viral transmission, and disease development (Jaimes et al., 2020; Zehr et al., 2023). Isolates belonging to the FCoV-I group exhibit distinct sequence variations at the S1/S2 FCS, which can be tissue-specific (Healey et al., 2022). In this study, FIPV-170 and FIPV-193 showed standard 788-RRSRRS-793, while FIPV-246 contained 788-RRARRS-793 with a substitute S790A at the poly-basic residue. However, the FCS region is absent in FCoV-II and FCoV-23 (Attipa et al., 2025). A study reported that the CCoV strain HuPn isolated from Malaysia in 2018 has no FCS and was experimentally found to be un-cleaved (Tortorici et al., 2022). The un-cleaved spike is more stable than the cleaved spike and is reported to provide more stability (Wrobel et al., 2020). The S1/S2 FCS facilitates the movement of NTD and helps RBD in receptor binding. Another study showed that Sialoside binding is an essential step in coronavirus infection, and it may help in the intracellular spread of the disease (Qing et al., 2020). However, several studies have shown that an absence or mutations in the S1-NTD sialic acid binding domain can result in an alternative process for the virus to encode its own sugar cleavage enzyme and modify glycoprotein binding (Attipa et al., 2025; Buchanan et al., 2022). In this study, we found 76 amino acid mutations in the S1-NTD of FIPV-170, 193, and 246 when compared with other reference strains from FCoV-I, FCoV-II, FCoV-23, and CCoV-I/II (Supplementary Table 3). This may increase the binding efficacy at the RBD and enhance the flexibility of the spike.

A study reported that two amino acid mutations in the spike protein at M1058L and S1060A are responsible for the virus shift from FECV to FIPV (Chang et al., 2012). Another study reported a similar M1058L mutation in 16 out of 18 FIPV-infected cats (Decaro et al., 2021). However, several studies found that the substitution of these two amino acids may not be related to all FIPV cases (Porter et al., 2014; Barker et al., 2017; Felten et al., 2017; Jahne et al., 2022). Consistently, we did not find substitutions in spike M1058L and S1060A in FIPV-170, 193, and 246. A study demonstrated that M1058L and S1060A are involved in the tissue tropism and its adaptability to monocytes and macrophages and enhance virulence and pathogenicity (Chang et al., 2012). However, another study suggested mutations in HR1 and HR2 are responsible for tropism in macrophages and trigger FIP infection (Rottier et al., 2005). Furthermore, in the RBD of spike, we reported a total of twelve novel amino acid substitutions, in which three amino acid substitutions were in FIPV-170, five in FIPV-193, and four in FIPV-246. Alterations in the spike may also contribute to immune evasion and promote tissue tropism and virulence in the host (Jaimes and Whittaker, 2018). These results suggest that the mutations in the S1-NTD and RBD could be the possible reason for tissue tropism in these novel FCoV field viruses.

Mutations in the NSP-3abc and NSP-7ab encoded accessory proteins are also associated with the genomic drift from FECV to FIPV (Borschensky and Reinacher, 2014; Lutz et al., 2020). Another study proposed an essential role of NSP-3c protein in enteric tissue tropism, but not the main cause of infection in FECV or FIPV (Bank-Wolf et al., 2014). The NSP-7ab are Alpha-CoV-specific proteins and are absent in Beta- and Gamma-CoVs (Ortego et al., 2002). A recent study showed that some specific mutations in NSP-7b are correlated with FECV to FIPV development in Persian cats (Kennedy et al., 2008). In this study, we reported that FIPV-170 have an early stop codon in NSP-3a, resulting in a short protein of 63 amino acids instead of 71 amino acids. Whereas FIPV-193 have short NSP-7a protein of 85 amino acids instead of 105 amino acids, which could be the reason that NSP-7ab did not cluster with other isolates in phylogenetic analysis (Fig. 3D). However, the function of NSP-3a and NSP-7a in FIPV/FCoV tissue tropism and genomic drift is not well studied and needs further investigation.

Among the 110 strains/isolates analyzed, the nucleotide multiple sequence alignment and pairwise comparison showed the maximum similarity of no >91% with any of the isolates, suggesting substantial genetic drift. Therefore, we performed recombination analysis with reference strains, including FCoV-I strain UG-FH8, FCoV-II strain 79–1146, FCoV-23 strain 2-C11 Re, CCoV-I strain 23/03, and CCoV-IIa strain CB/05. Results showed that both FIPV-170 and FIPV-193 share the same potential recombination region at ORF1b, 17,491 – 18,740 and 16,789 – 18,882, respectively (Supplementary Figure S2, S3). Whereas FIPV-246 showed recombination events at position 19,289 – 26,950 spanning ORF1b, spike, and NSP-3abc genes (Fig. 5A, 5B). More interestingly, the main recombination events at FIPV-170 and FIPV-246 showed the same major and minor parent-like, CCoV-IIa strain CB/05 and CCoV-I strain 23/03, respectively.

However, recombination events in FIPV-246 showed FCoV-II strain 79–1146 as a major parent-like, and CCoV-IIa strain CB/05 as a minor parent-like. The CCoV infection are typically limited to the enteric tract, but CCoV NA/09 and CB/05 are recently shown to spread into internal organs including the intestine, lungs, spleen, liver, kidney, lymph nodes, brain, T-cells, and B-cells (Ntafis et al., 2012; Decaro et al., 2007; Buonavoglia et al., 2006; Marinaro et al., 2010). The CCoV CB/05 is closely related to FCoV-23 strains from Cyprus (Attipa et al., 2025) and is reported to small outbreaks (Buonavoglia et al., 2006). Notably, the phylogenetic analysis of recombination events of all three FCoV viral genomes identified in this study showed evidence of the potential similar recombinant events with FCoV-I strain UG-FH8 from Belgium and FCoV-23 strain 2-C11 from Cyprus. Together, these results suggest frequent genomic shifts in cats in the New York region, with the identification of FCoV type-II isolates in 2024 (FIPV-A15 and FIPV-A37) and FCoV type-I in 2025 (FIPV-170, 193, and 246) from the same region. Recombination analysis further supports this notion, revealing a potentially similar recombination pattern to FCoV-23 and showing resemblance to the recent outbreak of a novel FCoV-23 strain reported in Cyprus and the UK (Maggi et al., 2022).

Funding

This study was funded by a grant from the United States Department of Agriculture National Institute of Food and Agriculture (Grant# NI26AHDRXXXXG063).

Ethics statement

All animal handling and sample collection in this study was approved by the Long Island University Animal care committee (IACUC# 2024–004).

Biography for the first Author

Dr. Abid Shah is a postdoctoral fellow at Lewyt College of Veterinary Medicine, Long Island University, NY, USA. His research focuses on the molecular biology of viruses affecting various species of animals and birds. author.

Ethics declaration.

Animal subject.

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.

This study was approved by the LIU-IACUC comittee.

(Approval No 2024–004).

CRediT authorship contribution statement

Abid Ullah Shah: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Methodology, Formal analysis, Data curation. Csaba Vagra: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Resources, Methodology, Formal analysis, Data curation. Nadine Moawad: Writing – review & editing, Writing – original draft, Software, Resources, Project administration, Methodology, Formal analysis, Data curation. Megan Rosen: Writing – review & editing, Writing – original draft, Validation, Resources, Methodology, Formal analysis. Diane Levitan: Writing – review & editing, Writing – original draft, Resources, Data curation. Maged Gomaa Hemida: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no conflicts of interest.

Data availability

Data will be made available on request.