Bovine-like coronaviruses isolated from four species of captive wild ruminants are homologous to bovine coronaviruses, based on complete genomic sequences

Abstract

We sequenced and analyzed the full-length genomes of four coronaviruses (CoVs), each from a distinct wild-ruminant species in Ohio: sambar deer (Cervus unicolor), a waterbuck (Kobus ellipsiprymnus), a sable antelope (Hippotragus niger), and a white-tailed deer (Odocoileus virginianus). The fecal samples from the sambar deer, the waterbuck, and the white-tailed deer were collected during winter dysentery outbreaks and sporadic diarrhea cases in 1993 and 1994 (H. Tsunemitsu, Z. R. el-Kanawati, D. R. Smith, H. H. Reed, and L. J. Saif, J. Clin. Microbiol. 33:3264-3269, 1995). A fecal sample from a sable antelope was collected in 2003 from an Ohio wild-animal habitat during the same outbreak when a bovine-like CoV from a giraffe (GiCoV) was isolated (M. Hasoksuz, K. Alekseev, A. Vlasova, X. Zhang, D. Spiro, R. Halpin, S. Wang, E. Ghedin, and L. J. Saif, J. Virol. 81:4981-4990, 2007). For two of the CoVs (sambar deer and waterbuck), complete genomes from both the cell culture-adapted and gnotobiotic-calf-passaged strains were also sequenced and analyzed. Phylogenetically, wild-ruminant CoVs belong to group 2a CoVs, with the closest relatedness to recent bovine CoV (BCoV) strains. High nucleotide identities (99.4 to 99.6%) among the wild-ruminant strains and recent BCoV strains (BCoV-LUN and BCoV-ENT, isolated in 1998) further confirm the close relatedness. Comparative genetic analysis of CoVs of captive wild ruminants with BCoV strains suggests that no specific genomic markers are present that allow discrimination between the bovine strains and bovine-like CoVs from captive wild ruminants; furthermore, no specific genetic markers were identified that defined cell cultured or calf-passaged strains or the host origin of strains. The results of this study confirm prior reports of biologic and antigenic similarities between bovine and wild-ruminant CoVs and suggest that cattle may be reservoirs for CoVs that infect captive wild ruminants or vice versa and that these CoVs may represent host range variants of an ancestral CoV.

FIG. 1.

Map of the BCoV genome isolated from CWRs. The overlapping ORF1a and ORF1b encode proteins of the replicative complex; the approximate positions of putative functional domains (identified according to homology predicted from MHV-A59 domains from the sequence NC_001846) are given above the BCoV genome scheme: AC-PL1-PL2-X-TM1 contains the N-terminal acidic (AC) domain, papain-like proteinase domains PL1-PRO and PL2-PRO, ADP-ribose 1′-phosphatase (formerly known as “X-domain”), and transmembrane domain 1 (TM1). The overall identity with the group 2a MHV genome in this region is 66%. For the other domains, the proteins and their identities to MHV (in parentheses) are as follows. TM2 (78.1%) and TM3 (78.5%) are transmembrane-spanning domains, 3CL (76.9%) is the main protease, and SS-RNA binding protein (81.2%) is the last domain before the ribosomal slippage site which leads to −1 frameshift and synthesis of ORF1a and ORF1b as a single polyprotein, ORF1ab. RdRp (85.1%) is the RNA-dependent RNA polymerase, the key and most-conserved protein in the CoV replicative complex; next are the metal-binding and NTPase/helicase domains (83.1%), nuclease ExoN homolog (80.7%), uridylate-specific endoribonuclease NendoU (79.6%), and 2′-O-ribose methyltransferase (79.9%). BCoV genes in the 3′ region of the genome encode the next proteins, abbreviated as follows: 32kDa, 32-kDa nsp; HE, hemagglutinin-esterase protein; S, spike protein; 4.8kDa, 4.9kDa and 12.7kDa, nsp's by size; E, small membrane/envelope protein; M, membrane protein; and N, nucleocapsid protein.

FIG. 2.

Phylogenetic analysis of the WbCoV, SDCoV, WtDCoV, and SACoV isolates to identify their relatedness with known reference strains of BCoV or bovine-like CoVs. The range of rates of nucleotide substitution per 100 nucleotides is indicated below. The year of isolation is given for each strain. The Quebec and Mebus BCoV reference strains were attenuated by serial passages in cell culture. The LUN and ENT BCoV reference strains were also adapted to cell culture before sequencing, but their attenuation status is unknown.

FIG. 3.

Amino acid variations observed in structural proteins and nsp's of CWR CoV and reference BCoV sequences. Differences from consensus sequence amino acids are marked in bold. Amino acid substitutions common for CWR bovine-like CoVs from the same outbreak are boxed with solid lines. Amino acid substitutions distinguishing TC- or GnC-adapted viruses from the original strain are boxed with dashed lines. Three amino acid substitutions in the S1 domain of WtDCoV spike protein that are potentially responsible for altering the virus-neutralizing antibody binding site are highlighted with gray and boxed with solid lines. Notable differences in sequence variations include the following. (i) The BCoV-Quebec strain has a 10-aa insertion (TILRQWLPAG) after the G2509 position in ORF1b. (ii) The BCoV-Quebec strain has several nucleotide deletions, starting from codon 2551 in ORF1b (nucleotide positions on the BCoV-Mebus sequence are 20995, 21004, 21007, 21014 to 21015, 21022, 21027, 21030, 21037 to 21038, 21042, and 21054). This leads to a shorter amino acid sequence length for the Quebec strain in the region 2551 to 2570, 16 aa instead of 20 as in all other strains, and difference in residues due to frameshifts (Quebec has the sequence LLLDIGVHVVRCSYIH, while other CoVs have PITKNIGEYNVSKDGFFTYI). (iii) The BCoV-Quebec strain has a single-nucleotide deletion (cytosine in position 21272 in the Mebus genome) in ORF1b, causing frameshift and early translation termination at codon 41 upstream compared to the Mebus sequence end (stop codon at aa position 2676). (iv) In the 32-kDa nsp, a single-nucleotide deletion in BCoV-Quebec (guanine 21666 in BCoV-Mebus genome) causes frameshift and mismatch between Quebec and other BCoV sequences in the region of aa 55 to 71. The guanine insertion after position 21714 (according to the Mebus sequence) restores the match between the Quebec sequence and other sequences. (v) A single-nucleotide deletion in the BCoV-Quebec strain (adenine 22259 in the BCoV-Mebus genome) causes frameshift and mismatch between the Quebec sequence and other BCoV sequences in the region of aa 253 to 261. A 2-nucleotide deletion (cytosine 22285 and adenine 22286 in the BCoV-Mebus genome) restores the match between the Quebec sequence and other sequences. (vi) The Y507S substitution in the S1 subunit of spike might have contributed to the failure of SACoV replication in cell culture and in Gn calves. (vii) The guanine-to-cytosine substitution in the start codon of the E protein of SACoV probably shifts the start of protein synthesis to codon 3, also methionine. *, stop codon.