sequence tree for the S gene (Figure 3), equine strains E17 and E19 segregated amongst BCoV strains from calves of Southeastern region of Brazil, identified as USP and again diverged from EqCoV strains (AY316300, NC.010327, and EF446615).
The cluster with these two equine strains also contains BCoV strains described by Brandao et al.
and aa identities for S gene, the lowest identity found amongst the two equine studied strains and groups of BCoV was 89.27% or 89.60% and the highest identities were 99.92% and 99.83%, respectively.
The tree for N gene (Figure 4) showed that both equine strains segregated in the same cluster of Kakegawa BcoV (AB354579), similar to that found in the nt.
Regarding the identities of nucleotides and amino acids for the N gene, the lowest identities found amongst the equine strains E17 and E19 and BCoV strains groups were 96.94% and 97.18% and the highest was 97.73% or 97.5%, respectively.
Nonetheless the genes analysed, unexpectedly, clustered with BCoV strains.
The presence of coronaviruses similar to BCoV in hosts other than cattle has already been reported in buffalos [17, 18], lamas and alpacas [19, 20], deer , and giraffes , demonstrating that this virus can adapt to other herbivores, including horses, as found in the present study, a fact not reported previously.
Furthermore, the 18 nt./ six aa deletion in S1, already described for BCoV strains , was also detected in strains E17 and E19, possibly allowing for changes in the spike glycoprotein that could reduce crossed immunity with other BCoV strains .
Regarding the HE gene, strains E17 and E19 both clustered with Kakegawa BCoV strain (Figure 2; Genbank ID: AB354579).
Alternatively, BCoV strains from bovine resulted in three main clusters for HE: (a) strains from adult cows studied herein; (b) strains from dairy cows with winter dysentery previously reported in Brazil (Genbank); and (c) strains from cattle from other countries retrieved from the GenBank.
This model of segregation might represent a phylogeographic pattern rather than temporal and/ or host-specific patterns, since it is known that there are no markers to differentiate strains of BCoV from calves and adult cattle [24, 25] and either for temporal changes [26-28].
Considering the N gene tree (Figure 4), the strain E17 has diverged from E19 and BCoV strains, owing to a single nucleotide substitution, leading to aa substitution Val92Leu, increasing, for this strain, the number of nonsynonymous substitutions, which might mean that strain E17 is adapting to the equine host, and increasing its divergence from an ancestor BCoV strain.