Discovery of a novel bottlenose dolphin coronavirus reveals a distinct species of marine mammal coronavirus in Gammacoronavirus

Abstract

While gammacoronaviruses mainly comprise infectious bronchitis virus (IBV) and its closely related bird coronaviruses (CoVs), the only mammalian gammacoronavirus was discovered from a white beluga whale (beluga whale CoV [BWCoV] SW1) in 2008. In this study, we discovered a novel gammacoronavirus from fecal samples from three Indo-Pacific bottlenose dolphins (Tursiops aduncus), which we named bottlenose dolphin CoV (BdCoV) HKU22. All the three BdCoV HKU22-positive samples were collected on the same date, suggesting a cluster of infection, with viral loads of 1 × 10(3) to 1 × 10(5) copies per ml. Clearance of virus was associated with a specific antibody response against the nucleocapsid of BdCoV HKU22. Complete genome sequencing and comparative genome analysis showed that BdCoV HKU22 and BWCoV SW1 have similar genome characteristics and structures. Their genome size is about 32,000 nucleotides, the largest among all CoVs, as a result of multiple unique open reading frames (NS5a, NS5b, NS5c, NS6, NS7, NS8, NS9, and NS10) between their membrane (M) and nucleocapsid (N) protein genes. Although comparative genome analysis showed that BdCoV HKU22 and BWCoV SW1 should belong to the same species, a major difference was observed in the proteins encoded by their spike (S) genes, which showed only 74.3 to 74.7% amino acid identities. The high ratios of the number of synonymous substitutions per synonymous site (Ks) to the number of nonsynonymous substitutions per nonsynonymous site (Ka) in multiple regions of the genome, especially the S gene (Ka/Ks ratio, 2.5), indicated that BdCoV HKU22 may be evolving rapidly, supporting a recent transmission event to the bottlenose dolphins. We propose a distinct species, Cetacean coronavirus, in Gammacoronavirus, to include BdCoV HKU22 and BWCoV SW1, whereas IBV and its closely related bird CoVs represent another species, Avian coronavirus, in Gammacoronavirus.

FIG 1

Genome organizations of BdCoV HKU22 and representative CoVs from each genus. Orange boxes, papain-like proteases (PL1^pro, PL2^pro, and PL^pro), chymotrypsin-like protease (3CL^pro), and RNA-dependent RNA polymerase (RdRp); green boxes, hemagglutinin esterase (HE), spike (S), envelope (E), membrane (M), and nucleocapsid (N) proteins; blue boxes, putative accessory proteins.

FIG 2

Amino acid differences and selection pressure analysis of S proteins of BdCoV HKU22 and BWCoV SW1. (A) Distribution of amino acid changes in the S proteins of BdCoV HKU22 and BWCoV SW1. The positions of the amino acid changes are depicted by vertical lines, and deletions are marked by asterisks. (B) Distribution of positively selected sites in S protein identified using random effects likelihood (REL) (39). dN > dS, more nonsynonymous than synonymous substitutions.

FIG 3

Phylogenetic analyses of the ORF1b polyprotein and the S and N proteins of BdCoV HKU22. The trees were constructed by using the neighbor-joining method in the JTT substitution model with a gamma-distributed rate variation and bootstrap values calculated from 1,000 trees. Bootstrap values below 70% are not shown. A total of 2,716, 1,495, and 379 amino acid positions in ORF1b polyprotein, S protein, and N protein, respectively, were included in the analyses. The tree was rooted to Breda virus (AY_427798). For the ORF1b polyprotein, the scale bars indicate the estimated number of substitutions per 10 amino acids. For the S and N proteins, the scale bars indicate the estimated number of substitutions per 5 amino acids. The three strains of BdCoV HKU22 characterized in this study are in boldface.

FIG 3

Phylogenetic analyses of the ORF1b polyprotein and the S and N proteins of BdCoV HKU22. The trees were constructed by using the neighbor-joining method in the JTT substitution model with a gamma-distributed rate variation and bootstrap values calculated from 1,000 trees. Bootstrap values below 70% are not shown. A total of 2,716, 1,495, and 379 amino acid positions in ORF1b polyprotein, S protein, and N protein, respectively, were included in the analyses. The tree was rooted to Breda virus (AY_427798). For the ORF1b polyprotein, the scale bars indicate the estimated number of substitutions per 10 amino acids. For the S and N proteins, the scale bars indicate the estimated number of substitutions per 5 amino acids. The three strains of BdCoV HKU22 characterized in this study are in boldface.

FIG 4

Western blot analysis of purified recombinant BdCoV HKU22 N-protein antigen. Lanes 1 to 4, 5 to 8, and 9 to 12, serial serum samples collected from the three bottlenose dolphins RT-PCR positive for BdCoV HKU22, respectively. Prominent immunoreactive protein bands of about 42 kDa (arrowheads) were visible 4 to 8 weeks after the fecal samples of the three bottlenose dolphins tested positive for BdCoV HKU22 (lanes 3 and 4, 7 and 8, and 11 and 12). Only very faint bands were observed for serum samples obtained before and within a few days after their fecal samples tested positive for BdCoV HKU22 (lanes 1 and 2, 5 and 6, and 9 and 10). The dates of serum collection are indicated below each lane. The dates that the dolphins were RT-PCR positive and RT-PCR negative for BdCoV HKU22 are also shown.

FIG 5

Estimation of date of divergence of BdCoV HKU22 and BWCoV SW1. (A) Regression of root-to-tip distances against the date of sampling of four genomic sequences to estimate the rate of evolution and the time of the most recent common ancestor. (B) Maximum likelihood tree with time scale using the estimated rate of evolution.

FIG 6

Model of CoV evolution with a distinct species, Cetacean coronavirus, in Gammacoronavirus.