Forty-five swab samples of cormorant feces were collected using sterile cotton swabs and identified by experts in bird identification during the collection process. The swabs were dipped into Dulbecco phosphate-buffered saline (DPBS) with a volume of 0.5 mL and vigorously vortexed for 10 min. They were then centrifuged, and each supernatant was collected in a 1.5 mL centrifuge tube and stored at −80°C until use.

Five pools with the volume of 500 μL were created, with each pool containing nine supernatants. To remove eukaryotic and bacterial cell-sized particles, the samples underwent centrifugation at 12,000 × g for 20 min at 4°C, followed by filtration through a 0.45 μm filter. The filtrate was then treated with DNase and RNase at 37°C for 60 min before extracting the total nucleic acid using the QIAamp MinElute Virus Spin Kit (Qiagen), according to the manufacturer’s instructions. The Illumina Nextera XT DNA Sample Preparation Kit was used to construct five libraries, and the Miseq Illumina platform was used to create 250 base-pairing ends. Illumina’s vendor software was employed to clip the barcodes from the raw sequences. The data was analyzed using a 32-node Linux cluster. An in-house analysis pipeline was employed, which eliminated cloned readings, trimmed low sequencing quality tails by applying a minimum threshold of phred quality score of 10, and removed adapters using NCBI VecScreen. Bacterial reads were subtracted by mapping them to the bacterial blast NT database using Bowtie2. The readings were assembled using SOAPdenovo2 r240, with a kmer size of 63 and default settings. The assembled contigs and singlets were aligned with an in-house viral proteome database using blastx (v.2.2.7) with an E-value cut-off of less than 10⁻⁵. To compile the viral blastx database, the NCBI virus reference proteome’s fasta file was used based on the annotation classification in the virus Kingdom.

The putative open reading frames (ORFs) of the assembled virus genome were predicted by Geneious software (Version 11.1.2) and further checked by comparing them to related viruses by Blastp in NCBI. Conserved protein motifs were searched using Geneious software (Version 11.1.2).

The phylogenetic tree was constructed using MrBayes v3.2.7, with the parameters “lset nst = 6 rates = invgamma.” This setting applied the GTR substitution model with gamma-distributed rate variation across sites and a proportion of invariable sites (“GTR + I+Г”). Additionally, “prset aamodelpr = mixed” was employed to enable the program to use the ten built-in amino acid models. The maximum number of generations was set to be ten million, and sampling occurred at every 50 generations, with the first 25% of Markov chain Monte Carlo (mcmc) samples being discarded as burn-in. Convergence was confirmed when the standard deviation of split frequencies was below 0.01.

To investigate the prevalence of calicivirus in cormorants, two sets of primers were designed in the RdRp region of Cormcali01. The first round of RT-PCR used the forward primer “5-CCC​CTA​ACA​GGC​ACA​GAA​GG-3,” and the reverse primer “5-GGT​TGA​TTT​GTT​GCG​GGA​CC-3” to amplify the target region. The second round of PCR used the forward primer “5-ATC​TGA​TGC​TTG​CAG​ACG​CT-3,” and the reverse primer “5-ATT​CAC​CAC​CTG​TGC​ATC​GT-3” to screen for Cormcali01.

A total of 4,609,598 raw reads were generated. Among them, 805 reads had significant BLAST hits to calicivirus [raw sequencing data were deposited to the Genome Sequence Archive (Chen et al., 2021) in National Genomics Data Center (Xue et al., 2022), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA010897) that are publicly accessible ¹ ]. Using the assemble program in the Geneious software (version 11.1.2), one complete bird calicivirus was retrieved from the raw reads and named Cormcali01. It was deposited into NCBI with the accession number OR062614.

The Cormcali01 genome is 8,561 base pairs (bp) long, and it significantly differs from all known calicivirus nucleotide sequences. However, it has the highest amino acid identity (47.34%) to the unclassified Ruddy turnstone calicivirus A. The G+C content of Cormcali01 was 49.66%, which is lower than that of Ruddy turnstone calicivirus A (G+C, 51.77%).

For Cormcali01, two ORFs were identified using Geneious software. The ORF1 consists of 7,326 bp (2,442 aa, 268.9 kDa), which encodes the polyprotein that will be further cleaved into five nonstructural proteins and VP1 (Figure 1). Based on alignment with amino acid sequences from goose calicivirus, temminck’s stint calicivirus (Liao et al., 2014; Matsvay et al., 2022), and caliciviral 3C-like protease cleavage site preferences (Sosnovtsev et al., 2006; Oka et al., 2007), the ORF1 polyprotein of Cormcali01 was predicted to be cleaved at five sites: Q³⁷⁵/S, Q⁸⁰²/A, Q¹¹⁰⁵/A, E¹¹⁸⁸/G, and E¹⁸²⁶/G, which would generate N-term (375 aa, 42.8 kDa), NTPase (427 aa, 47.3 kDa), NS4 (303 aa, 34.1 kDa), VPg (83 aa, 9.3 kDa), Pro-Pol (638 aa, 70.2 kDa), and VP1 (615aa, 65.2 kDa). The 519-bp ORF2 was predicted to encode 173 aa VP2 with a molecular weight of 18.2 kDa that had 3 nt overlapped with ORF1 at its 3′-end. A translational upstream ribosome binding site motif (⁷⁸⁴⁹ CAUGGGA ⁷⁸⁵⁵) was identified at 35 nt upstream of the ORF2 start codon, which is similar to other avian caliciviruses (Wang et al., 2017). It indicates that VP2 of Cormcali01 has the same translation initiation mechanism as other avian caliciviruses. VP2 is considered as the minor structural protein in caliciviruses, but it plays a crucial role in productive replication, leading to the formation and maturation of infectious virions (Sosnovtsev et al., 2005). The variability of VP2 sequences between genera is considerably high, with only 10% amino acid similarity. Moreover, the length of these proteins varies significantly, ranging from 106 to 268 amino acids. Compared to the genomes of its closest avian caliciviruses, Cormcali01 demonstrates a shorter length and lacks homology with them. This suggests possible variations in its replication and maturation mechanisms in relation to other caliciviruses.

The Geneious software was used to search for conserved protein motifs, and the results showed that the polyprotein of Cormcali01 contains characteristic motifs that are conserved in caliciviruses. These include NTPase motif GXPGXGKT (⁶⁰⁹GPPGYGKT⁶¹⁶), ⁶⁸⁷KGKVFTSKLIIATTN⁷⁰¹; VPg motif: EYXEX (¹⁰⁸³EYAEW¹⁰⁸⁷); NSpro motifs G (D/Y)CGXP (¹²⁶⁶GDCGRP¹²⁷¹), DY(S/K)(K/G)WDST (¹⁵⁴⁹DYSKWDST¹⁵⁵⁶), ¹⁶⁰⁴GLPSG¹⁶⁰⁸, as well as ¹⁶⁵²YGDD¹⁶⁵⁵ and ¹⁶⁹⁹FLKR¹⁷⁰², which were 100% conserved among all caliciviruses. VP1 motifs ¹⁹⁸⁰PPG¹⁹⁸² and FXXVXPP (²⁰⁶⁷FCLLKEP²⁰⁷³) were also identified.

To determine the taxonomic relationship between Cormcali01 and the caliciviruses currently acknowledged, we performed a search for homologous VP1 proteins using the BLASTp algorithm. According to the International Committee on Taxonomy of Viruses (ICTV), the criterion for genera demarcation in the Caliciviridae family is the divergence of the VP1 amino acid sequence of more than 60% (Vinje et al., 2019). The divergence of the VP1 amino acid sequence between Cormcali01 and representatives of each accepted genera were calculated (Figure 2). Results showed that Cormcali01 had the highest aa identity of 43.90% with unassigned Ruddy turnstone calicivirus A (MH453861) and had less than 40% similarity with other nacoviruses or bavoviruses. Thus, according to the accepted criterion, Cormcali01 cannot be categorized into any of the approved genera. To ascertain the evolutionary linkage between Cormcali01 and other members of the Caliciviridae family, bayesian inference tree were constructed based on the amino acid sequences of VP1 proteins of bird caliciviruses retrieved from the NCBI protein database. Results showed that VP1 of Cormcali01 clustered with unassigned caliciviruses, including Ruddy turnstone calicivirus (MH453861), duck calicivirus (MH453811), goose calicivirus (KY399947), and Temminck stint calicivirus (ON815296) with a high posterior probability of >99% to form a clade (Figure 3). Other unassigned avian caliciviruses, such as grey teal calicivirus (MK204392), pink-eared duck calicivirus (MK204415), and avocet calicivirus (MH453803), together with two strains of the Nacovirus genus, turkey calicivirus (NC043516) and goose calicivurs (KJ473715) formed a clade. All the chicken caliciviruses, including an unassigned chicken calicivirus clustered together in the genus Bavovirus. This suggests that Cormcali01 is genetically closer to a subset of unclassified avian caliciviruses, forming a separate evolutionary branch, and is further from other defined nacoviruses and bavoviruses. It appears to be affiliated with the proposed genus Sanovirus (Wang et al., 2017; Matsvay et al., 2022).

To investigate the prevalence of calicivirus in cormorants, two sets of primers were designed in the RdRp region of Cormcali01. The results showed that 22.22% of the fecal samples (10 out of 45) tested positive for Cormcali01. This prevalence was lower than that of Bavovirus and Nacovirus, which were previously reported in chicken farms (83% for Bavovirus, 46% for Nacovirus), as well as in wild ducks such as mallards (40%) and American black ducks (26.9%) (Wolf et al., 2012; Canuti et al., 2019).

In summary, a novel bird calicivirus, named Cormcali01, was identified in the fecal samples collected from cormorants, and its genome and encoding proteins were characterized. Cormcali01 possesses the conserved motifs found in caliciviruses, including two NTPase motifs, one VPg motif, five NSpro motifs, and two VP1 motifs. Divergence and phylogenetic analysis of VP1 indicated that Cormcali01 is closely related to unassigned bird caliciviruses proposed to be in the genus Sanovirus. Cormcali01 had a prevalence of 22.22% in cormorant tested, which indicates a potential outbreak of bird calicivirus in wild birds.