High diversity of picornaviruses in rats from different continents revealed by deep sequencing

Abstract

Outbreaks of zoonotic diseases in humans and livestock are not uncommon, and an important component in containment of such emerging viral diseases is rapid and reliable diagnostics. Such methods are often PCR-based and hence require the availability of sequence data from the pathogen. Rattus norvegicus (R. norvegicus) is a known reservoir for important zoonotic pathogens. Transmission may be direct via contact with the animal, for example, through exposure to its faecal matter, or indirectly mediated by arthropod vectors. Here we investigated the viral content in rat faecal matter (n=29) collected from two continents by analyzing 2.2 billion next-generation sequencing reads derived from both DNA and RNA. Among other virus families, we found sequences from members of the Picornaviridae to be abundant in the microbiome of all the samples. Here we describe the diversity of the picornavirus-like contigs including near-full-length genomes closely related to the Boone cardiovirus and Theiler's encephalomyelitis virus. From this study, we conclude that picornaviruses within R. norvegicus are more diverse than previously recognized. The virome of R. norvegicus should be investigated further to assess the full potential for zoonotic virus transmission.

Figure 1

The distributions of unique reads are shown for RNA (A) and DNA (B) libraries in absolute numbers of unique reads, as well as proportions of unique reads from RNA (C) and DNA (D) libraries. Unique Illumina reads were mapped to contigs that were mapping to complete reference genomes from viruses, R. norvegicus (Rn5), protists, phages, fungi, vectors, bacteria, archaic viruses and Archaea. Sample locations in Denmark are abbreviated as indicated: Amager East (AE), Egedal (EM), Botanical Garden of Copenhagen (BGC), Faculty of Health and Medical Sciences (FHMS) and Copenhagen University Hospital (CUH). Additional sampling locations included Hong Kong, Kuala Langat and Kuala Lumpur.

Figure 2

Viral diversity in urban wild R. norvegicus. Each dot represents ≥1 contig(s) from a specific sample that maps to a specific reference genome within the virus families indicated (colour coded). The dot size indicates the number of non-overlapping contigs mapping to each reference genome (scale bar). The x-axis shows the average contig length, while the y-axis shows the mean identity of the contigs mapping to viral reference genomes. Thresholds corresponding to contig lengths exceeding 500 nt and percentage identity <70% are delimited by the horizontal and vertical lines. (A) and (B) show RNA and DNA sequencing results, respectively.

Figure 3

Abundance and diversity of virus families. Distribution and taxonomic classification (family) of reads mapping to assembled viral contigs. Colour intensity represents log2-transformed counts of reads mapped to the contigs resembling a viral family for each sampling location shown on the y-axis. Sample locations in Denmark are abbreviated as indicated: Amager East (AE), Egedal (EM), Botanical Garden of Copenhagen (BGC), Faculty of Health and Medical Sciences (FHMS) and Copenhagen University Hospital (CUH).

Figure 4

Maximum-likelihood phylogenetic tree of reference sequences from the Cardiovirus genus, with subsequently placed picornavirus ORF contigs identified in the present study. Picornavirus-like contigs were added to the tree using SEPP software (see Materials and Methods), including four contigs published elsewhere^, (denoted Sachsenröder et al. and Firth et al., respectively). Contigs from samples from different geographical locations are colour coded as indicated in the legend. The data set from FHMS was excluded, as no cardiovirus-like contigs were identified. Monophyletic clusters of multiple reference sequences were collapsed (for example, equine rhinitis A virus). The major nodes of the underlying ML tree were supported by bootstrap values ≥69.