Surveillance of coronaviruses in wild aquatic birds in Hong Kong: expanded genetic diversity and discovery of novel subgenus in the Deltacoronavirus

Abstract

Migratory birds may carry emerging viruses over long distances. Regular surveillance and metagenomic analysis were employed to explore the diversity of avian coronaviruses at Hong Kong's Mai Po Wetland. We tested a total of 3239 samples collected from 2018 to 2024, among which the prevalence rate of viruses of the genus Gammacoronavirus (64.4%) was higher than that of Deltacoronavirus (35.6%). The host species were identified for 79.8% of the coronavirus-positive samples. Two deltacoronaviruses with full-genome sequences and one nearly complete gammacoronavirus genome were identified in faecal samples of three bird species. We also predicted putative transcriptional regulatory sequences and 3CLpro and PLpro cleavage sites for these viruses. Results from our phylogenetic analysis and pairwise amino acid identity comparisons, using the International Committee on Taxonomy of Viruses classification criteria based on the DEmARC framework, indicate that black-faced spoonbill coronavirus (BSCoV, strain MP22-1474) prototypes a new subgenus. Great cormorant coronavirus (GCCoV, strain MP18-1070) and falcated duck coronavirus (FDCoV, strain MP22-196) belong to two previously known species while diverging most profoundly from known viruses of these species. Two recombination events may have contributed to the evolution of FDCoV MP22-196 in genome regions from ORF1b to the S gene and from the M gene to the N gene. The cophylogenetic analysis between avian hosts and coronaviruses provides evidence for a strong linkage between viruses of the genus Gammacoronavirus and the birds of order Anseriformes. This study highlights the importance of ongoing surveillance for coronaviruses in wild migratory birds.

Keywords: Deltacoronavirus; Gammacoronavirus; subgenus; virus surveillance; wild aquatic birds.

Figure 1

Phylogenetic analysis on the RdRp region (360 bp) of coronaviruses. The trees are constructed by IQ-TREE v2.3.5 with the GTR + I + G substitution model. The tree is midpoint rooted. Bootstrap values are set up with 1000 replicates as statistical support. The branch values are bootstrap supports (%). Coronaviruses detected in this study (N = 247) are highlighted in colour and named with a sample number, followed by the corresponding identical host species. Samples related to the same groups of similar nodes have been collapsed and indicated by the isosceles triangles figure, and the total number of sequences in each group is shown next to the triangles. The GenBank accession numbers of retrieved genes are indicated at the end of the reference name. Four main monophyletic clusters corresponding to the genera Alphacoronavirus, Betacoronavirus, Deltacoronavirus, and Gammacoronavirus within the subfamily Orthocoronavirinae are indicated on the right side of this figure.

Figure 2

Genome and domain organization of coronaviruses discovered in this study with night-heron coronavirus HKU19 (NHCoV HKU19, NC_016994), Wigeon coronavirus HKU20 (WiCoV HKU20, NC_016995), and infectious bronchitis virus (IBV, NC_048213). Three coronaviruses are shown in bold. The upper indicated ORF1a and ORF1b regions and downstream of the S, E, M, and N genes are represented by labeled boxes. Putative accessory ORFs located before and after the N gene region are represented by outlined boxes. 3CLpro, NIRAN, RdRp, ZBD, and HEL 1 are annotated within the ORF1a and ORF1b regions. These domain combinations are used for phylogenetic and DEmARC analysis of the family Coronaviridae.

Figure 3

Phylogenetic trees of seven protein domains of coronaviruses. Depicted are trees for amino acid sequences of 3CLpro, NIRAN, RdRp, ZBD, HEL 1, S protein, and N protein of coronaviruses, including GCCoV MP18-1070, BSCoV MP22-1474, and FDCoV MP22-196. Each tree is rooted with Microhyla letovirus 1 (GECV01031551) and Pacific salmon nidovirus (MK611985) as an outgroup. These trees are constructed by IQ-TREE v2.3.5 with the LG + F + G substitution model. Branch support is estimated using SH-aLRT with 1000 replicates. The support values of SH-aLRT are depicted with dots of varying shades at the internal nodes, according to details within the key box in the left corner. The branch values are bootstrap supports (%). Demarcations of species, subgenus, and genus are shown on the right side (see Fig. 4 for classification of the newly identified coronaviruses). For BSCoV MP22-1474, the subgenus is to be named Pladecovirus, and the species name is to be named Deltacoronavirus plataleae.

Figure 3

Phylogenetic trees of seven protein domains of coronaviruses. Depicted are trees for amino acid sequences of 3CLpro, NIRAN, RdRp, ZBD, HEL 1, S protein, and N protein of coronaviruses, including GCCoV MP18-1070, BSCoV MP22-1474, and FDCoV MP22-196. Each tree is rooted with Microhyla letovirus 1 (GECV01031551) and Pacific salmon nidovirus (MK611985) as an outgroup. These trees are constructed by IQ-TREE v2.3.5 with the LG + F + G substitution model. Branch support is estimated using SH-aLRT with 1000 replicates. The support values of SH-aLRT are depicted with dots of varying shades at the internal nodes, according to details within the key box in the left corner. The branch values are bootstrap supports (%). Demarcations of species, subgenus, and genus are shown on the right side (see Fig. 4 for classification of the newly identified coronaviruses). For BSCoV MP22-1474, the subgenus is to be named Pladecovirus, and the species name is to be named Deltacoronavirus plataleae.

Figure 3

Phylogenetic trees of seven protein domains of coronaviruses. Depicted are trees for amino acid sequences of 3CLpro, NIRAN, RdRp, ZBD, HEL 1, S protein, and N protein of coronaviruses, including GCCoV MP18-1070, BSCoV MP22-1474, and FDCoV MP22-196. Each tree is rooted with Microhyla letovirus 1 (GECV01031551) and Pacific salmon nidovirus (MK611985) as an outgroup. These trees are constructed by IQ-TREE v2.3.5 with the LG + F + G substitution model. Branch support is estimated using SH-aLRT with 1000 replicates. The support values of SH-aLRT are depicted with dots of varying shades at the internal nodes, according to details within the key box in the left corner. The branch values are bootstrap supports (%). Demarcations of species, subgenus, and genus are shown on the right side (see Fig. 4 for classification of the newly identified coronaviruses). For BSCoV MP22-1474, the subgenus is to be named Pladecovirus, and the species name is to be named Deltacoronavirus plataleae.

Figure 3

Phylogenetic trees of seven protein domains of coronaviruses. Depicted are trees for amino acid sequences of 3CLpro, NIRAN, RdRp, ZBD, HEL 1, S protein, and N protein of coronaviruses, including GCCoV MP18-1070, BSCoV MP22-1474, and FDCoV MP22-196. Each tree is rooted with Microhyla letovirus 1 (GECV01031551) and Pacific salmon nidovirus (MK611985) as an outgroup. These trees are constructed by IQ-TREE v2.3.5 with the LG + F + G substitution model. Branch support is estimated using SH-aLRT with 1000 replicates. The support values of SH-aLRT are depicted with dots of varying shades at the internal nodes, according to details within the key box in the left corner. The branch values are bootstrap supports (%). Demarcations of species, subgenus, and genus are shown on the right side (see Fig. 4 for classification of the newly identified coronaviruses). For BSCoV MP22-1474, the subgenus is to be named Pladecovirus, and the species name is to be named Deltacoronavirus plataleae.

Figure 4

Hierarchical classification of three avian coronaviruses using DEmARC. Shown PPD of three avian coronaviruses with the most closely related coronaviruses of taxa up to the genus they belong to, using box-and-whisker plots. Outliers are indicated by hollow circular symbols, while solid symbols represent typical values. For GCCoV MP18-1070 and FDCoV MP22-196, their intraspecies PPDs are plotted separately from those with more distantly related viruses within either genus or subgenus, respectively. Intraspecies PPDs were derived from GCCoV MP18-1070 and night-heron coronavirus HKU19 (NHCoV HKU19, NC_016994), and also FDCoV MP22–196 and duck coronavirus isolate DK/GD/27/2014 (DuCoV 2714, NC_048214). Family-wide demarcation PPD thresholds for ranks of species, subgenus, and genus are shown as broken horizontal lines.

Figure 5

Putative recombination events of FDCoV MP22-196 are identified using the RDP method in RDP5 v5.61, and full-genome nucleotide sequence similarities are assessed using Simplot++ v1.3. FDCoV MP22-196 is used as the query sequence (PQ821724, recombinant) and analysed with the reference genome (Table S3) by RDP5 v5.61. Putative major and minor parent sequences are identified by RDP5 and used for the similarity plot analysis of FDCoV MP22-196 by Simplot++. The sites and lines are depicted in different colours according to the details within the key box on the right side. (A) RDP and (B) similarity plot derived from the major parent is bird gammacoronavirus AnasCN24 isolate AvAa-GammaCoV/SH21-SH145 (PP845446), and the minor parent is bird gammacoronavirus AnasCN24 isolate AvAz-GammaCoV/SH19-SH17B (PP845443). (C) RDP method and (D) similarity plot derived from the major parent is bird gammacoronavirus AnasCN24 isolate AvAz-GammaCoV/SH20-SH66 (PP845387), and the minor parent is avian coronavirus isolate MW18 (MK204411).

Figure 6

Tanglegram of cophylogenetic relationships between wild bird hosts and coronaviruses. Two trees are constructed by IQ-TREE v2.3.5 with the GTR + I + G substitution model. The left is the host evolutionary tree (COX1 gene, 670 bp), and the branches in the same colour represent the same order. The host tree is rooted with Apteryx mantelli North Island brown kiwi (EU525317) as an outgroup. The right is the phylogenetic tree of the viral RdRp gene (440 bp), and the branches in the same colour represent the same subgenus. The virus tree is rooted with human coronavirus 229E (NC_002645). The cophylogenetic relationships are performed by the ParaFit and PACo functions in the R program. Solid lines denoted significant coevolution links between coronaviruses and their hosts, and dotted lines denoted nonsignificant links.