Classification of new morbillivirus and jeilongvirus sequences from bats sampled in Brazil and Malaysia

Abstract

As part of a broad One Health surveillance effort to detect novel viruses in wildlife and people, we report several paramyxovirus sequences sampled primarily from bats during 2013 and 2014 in Brazil and Malaysia, including seven from which we recovered full-length genomes. Of these, six represent the first full-length paramyxovirid genomes sequenced from the Americas, including two that are the first full-length bat morbillivirus genome sequences published to date. Our findings add to the vast number of viral sequences in public repositories, which have been increasing considerably in recent years due to the rising accessibility of metagenomics. Taxonomic classification of these sequences in the absence of phenotypic data has been a significant challenge, particularly in the subfamily Orthoparamyxovirinae, where the rate of discovery of novel sequences has been substantial. Using pairwise amino acid sequence classification (PAASC), we propose that five of these sequences belong to members of the genus Jeilongvirus and two belong to members of the genus Morbillivirus. We also highlight inconsistencies in the classification of Tupaia virus and Mòjiāng virus using the same demarcation criteria and suggest reclassification of these viruses into new genera. Importantly, this study underscores the critical importance of sequence length in PAASC analysis as well as the importance of biological characteristics such as genome organization in the taxonomic classification of viral sequences.

Fig. 1

Histograms of paramyxovirid sequences submitted to GenBank by submission year (left) and sequence length (right). Trends are shown for all Paramyxoviridae (top) and Orthoparamyxovirinae only (bottom). The number of sequences submitted is increasing each year, but a significant majority of sequences submitted are of short PCR fragments (~500 bp), and very few are of full genomes (~15-16 kb).

Fig. 2

Nucleotide-sequence-based phylogeny with maximum-likelihood bootstrap values of the orthoparamyxovirins for which full- or nearly full-length RdRp gene sequences are available. The genus Respirovirus is represented by a single sequence, that of Sendai virus, as an outgroup. Branches are colored according to their current genus classifications, which are also indicated at the right. Colored highlighting indicates the taxonomic suggestions discussed here, including reclassifying Tupaia virus and Mòjiāng virus each into its own new genus and establishing the subclades A-D within the genus Jeilongvirus.

Fig. 3

Genome organization of representative members of each genus included in this study. Sequences are organized according to their RdRp phylogeny, which is shown on the left. The new genome sequences described here are labeled with red font, with the exception of the bat morbilliviruses (PDF-3137 and PBZ-1381), which share the same genome arrangement with other morbilliviruses. Where more than one genome arrangement is present in a single genus, a representative of each type is shown. ORFs are represented by colored polygons, and the black lines represent the entire length of the genome, both of which are drawn to scale. ORFs are colored as follows: nucleocapsid (N gene), blue; phosphoprotein (P gene), yellow; matrix (M gene), green; fusion (F gene), peach; ATUs, magenta; receptor binding protein (RBP gene), purple; RNA-dependent RNA polymerase (RdRp gene), orange. Asterisks on RBP ORFs indicate premature stop codons.

Fig. 4

Nucleotide-sequence-based phylogenies for all genes, excluding ATUs. Clades are color-coded with the same scheme as in Fig. 2: Morbillivirus, blue; Henipavirus, orange; Jeilongvirus “subclade A”, green; Jeilongvirus “subclades B” and “C”, yellow; Jeilongvirus “subclade D”, red; Narmovirus, purple. Where sequences are incomplete, no sequence is shown in the tree for that gene. For clades that are not monophyletic by genus in one or more of the trees (i.e., for paraphyletic sequences previously proposed as “shaanviruses” and for Tupaia virus), branches are shown as dotted lines.

Fig. 5

PAASC histograms of pairwise amino acid comparisons based on the RdRp gene. Scores on the x-axis were calculated by scoring alignments with the BLOSUM62 matrix and dividing by the total possible score (100% identity). Values on the y-axis indicate frequency. Gold boxes correspond to PIDs between sequences that do not conform to the identified cutoff values based on their established classifications. The cutoff values used are shown as dotted lines, with red for full-length RdRp sequences and blue for short fragments (Tong-PanPMV and Tong-RMH).