In this study, an FCoV type-specific, recombinant protein-based IFA was established. The expressed recombinant RBD region of the type II FCoV S protein, bearing one-tenth of the authentic S protein critical in both virus neutralization and cell attachment, is highly conserved among FCoV, TGEV and CCoV . The alignment of the amino acid sequences of this region revealed relative conservation among prototypic viral strains within the same serotype (i.e., 86.8-95.4% and 91.4-98.3% of type I and II FCoV, respectively) but were distinct between different serotypes (20.7-23.0%) (Figure 1a). As the expressed partial S proteins contained four N-linked glycosylation sites for both type I (551-NVT-553, 559-NDT-561, 591-NGS-593 and 596-NVT-598) and type II (520-NKS-522, 536-NIT-539, 558-NIT-560 and 566-NNT-568) FCoV (Figure 1a), the baculovirus expression system was chosen to provide post-translational modification and complex folding, including glycosylation [18,19]. A similar strategy was applied to severe acute respiratory syndrome (SARS)-CoV, and the recombinant RBD maintained authentic antigenicity that induced a strong RBD-specific antibody response . In addition, baculovirus-expressed, truncated SARS-CoV S protein-based IFAs were found to be highly sensitive and specific compared with conventional whole virus-based IFAs [21,22].
Based on different testing methods, the prevalence of FCoV in field cats varies among countries, e.g., 50% in Switzerland , 34% in Australia , 22% in Japan  and 13.7% in Korea . In the present study, FCoV-specific antibodies detected by IFA based on proteins derived from a local FCoV isolate revealed a seroprevalence of 34.5%. Previous studies revealed that >80% of seropositive cats are suspected to be infected with type I FCoV [9-11]. In the present study, 70.4% of local cats were infected with type I FCoV. Infection with type II FCoV in Taiwan (21.3%, 61/287) is higher than in other countries, i.e., 4.4% in Switzerland and 10.1% and 2% in Japan in 1992 and 2007, respectively. The dense population and close contact between cats and other animals in Taiwan could be reasons contributing to the higher rate of recombination between FCoVs and CCoVs or other alphacoronaviruses, leading to a relatively high type II FCoV infection rate. This relatively high antibody titer against type II virus might also result from other alphacoronavirus infection, as feline cells are susceptible to infection with FCoV, CCoV, TGEV and HCoV-229E  and given that seroconversion was observed in cats experimentally infected with CCoV , TGEV  or HCoV-229E .
In a prior study, a competitive ELISA was used to discriminate the serotype of infection . However, certain sera could not be typed once inhibition reached 30-80%. In the present study, serotypes were determined by the direct observation of a fluorescent signal specifically indicating two types of recombinant protein, which decreased the ambiguity. In another IFA study using type I or type II FCoV-infected cells, 23% of cats displayed equal titers against type I and II FCoV. The phenomenon was suggested by the authors to result either from antibodies against common antigenic epitopes in other structural protein or from co-infection with the two viruses . Our type-specific S protein-based IFA can avoid the sharing of common antigenicity, and co-infection can be readily identified.
This is the first report to correlate viral infection with the specific antibody evoked. We compared the serotypes and the genotypes of FCoV-infected animals. Most type I FCoV-infected cats with FIP harbored anti-type I FCoV antibodies, and three of them displayed antibodies against both types of FCoV, indicating co-infection with FCoV and/or other CoVs. Additionally, one cat with type II FIP (FIP-24) displayed an antibody response to type I and II FCoV simultaneously. This finding matches the recombination theory that states that type II FCoV arises from the mutation of type I FCoV. However, we found that three cats with type II FIP possessed antibodies solely against type II FCoV. Among those cats, FIP-23 (cat 11) was living in a shelter that had experienced a recent FIP outbreak caused by type II FCoV . The presence of antibodies solely against type II FCoV in cat FIP-23 provides an additional clue about the horizontal transmission of type II FIP. In the case of cat FIP-22, sequencing analysis of the C-terminal one-third of the viral genome indicated a virus bearing high resemblance to CCoV, but not to FCoV. Combining the serological and genetic findings, this FIP could have resulted from a CCoV infection. Further genetic analysis regarding this speculation is currently under investigation.
Among healthy cats with untypable FCoV, 84.6% were found to have a type I FCoV infection. Although the serotype of the antibodies detected did not always match the virus present at the time of sampling, infection with FCoVs, and especially with type I, is persistent and can last for several years in a multicat environment [14,30]. Moreover, type I virus has been reported to be more genetically diverse than type II virus in the genotyped region, and this diversity might hinder PCR amplification . Despite the better distinguishability, 8.3% of the cats in the current study possessed anti-FCoV antibodies that could not be serotyped. This issue could have been due to the lower antibody titers against FCoV, or polymorphisms may exist in the target region in S protein.
FCoV antibody titers when accurately performed are thought to be of some value in distinguishing enteric infection from FIP. Cats with IFA titers ≥1:3200 are highly suggestive of FIP . In our study using the serotyping method established, type-specific FCoV antibody could be titrated and the titer in some of our FIP cats reached relatively high level as well (≥1:1600). Our recent study revealed that the infection of serotype II FCoV correlated significantly with the occurrence of FIP . Also horizontal transmission of a serotype II virus was clearly demonstrated to responsible for a FIP outbreak . Continuous survey of the antibodies against two serotype of FCoV, especially in the multi-cat environment could help in identifying the invasion of any type II FCoV hence isolating the animals to prevent future FIP occurrence.