Results

VIRUS ISOLATION
CPE indicative of viral replication were observed in cell culture from nine samples representing tissues from all 11 animals; two of the samples each consisted of pooled organs from 2 separate salamanders while the remaining seven samples were from individual animals (Table 1). Ranavirus was not identified. However, a bullet-shaped virus characteristic of rhabdovirus (Figure 1) was observed by transmission electron microscopic analysis of a representative isolate (202238). Based on the observed morphology, PCR testing was conducted on this isolate using a rhabdovirus-specific assay targeting the polymerase gene.26 An amplicon of the expected size (260 nt) was generated, and sequence analysis showed closest similarity with the genus Sprivivirus in the family Rhabdoviridae (99% identity with SVCV A1, GenBank DQ097384). All four animals previously pooled for virus isolation purposes were individually positive using a SVCV-specific rRT-PCR test (Table 1).13

CHARACTERIZATION OF THE SALAMANDER ISOLATES
To further characterize the nine viral isolates, a 714-nt fragment of the viral glycoprotein (G) gene was amplified by PCR.12 The amplified G gene fragments exhibited 100% identity to one another over the region sequenced. One viral isolate (202238) was also chosen for whole-genome sequencing. A near complete contig of 10 991 nt that encoded five genes (nucleoprotein, N; phosphoprotein, P; matrix, M; glycoprotein, G; and large polymerase, L) in an order that is characteristic of the genus Sprivivirus was obtained following assembly of the genome (GenBank accession KU230365, Figure 2). There was no nonvirion gene at the G–L junction as would be characteristic of vesiculoviruses and perhabdoviruses. The conserved transcriptional start signal of 3′-UUGUC-5′ and termination/polyadenylation signal of 3′-AUA CUU UUU UU-5′ were both present in all five genes, and their presence is characteristic (although not exclusively) of Sprivivirus.27 Overall, the salamander virus was assigned to SVCV in accordance with the International Committee on Taxonomy of Viruses (ICTV).27
Bayesian and maximum likelihood phylogenetic analyses of the G gene fragments indicated that the salamander isolates clustered most closely with SVCV genogroup I (Figure 1). Median-joining network analysis further classified the virus to genogroup Ia (Figure 2). The salamander SVCV was distinct from previous isolates of SVCV from North America but was closely related to recent isolates from China. Although relatively few whole-genome sequences for SVCV are available for comparison, the salamander SVCV protein coding regions (N, P, M, G and L) were all closely related (98.2%–99.8% identical) at the nucleotide and amino-acid levels to corresponding genes in SVCV strain SH140501 (Table 2), which was isolated from a goldfish in Shanghai sampled in April 2014.28 This finding further supported an Asian origin for the salamander SVCV isolates.

PHYLOGENETIC ANALYSIS OF HOST SPECIES
To confirm the species identity of the salamanders infected with SVCV, a portion of the host mitochondrial NADH dehydrogenase subunit two gene (and adjacent tRNAs) was sequenced from a subset of four animals. The portion of DNA analyzed was 100% identical between the four animals examined. A representative sequence was deposited in GenBank (NWHC#26876-003; GenBank accession number KU647189). Phylogenetic analyses resulted in trees with the same topology and were similar to those of Wu et al.19 The salamanders that were positive for SVCV were part of a well-supported clade (Bayesian analysis: 100% posterior probability, maximum likelihood analysis: 100 bootstrap support; Figure 3) representative of C. orientalis. Specifically, imported salamanders grouped (Bayesian analysis: 100% posterior probability, maximum likelihood analysis: 100 bootstrap support) with C. orientalis that originated from the vicinity of Hangzhou, Zhejiang province, China, suggesting that the infected salamanders may have been collected from wild populations in that area.19