The Umbravirus is a genus of single-stranded positive-sense RNA plant viruses classified in the subfamily Calvusvirinae of the Tombusviridae [1]. Umbraviruses have 4 - 4.2 kb genomes containing four open reading frames (ORFs) but lack a capsid protein (CP) gene, which makes them dependent on a ‘helper’ virus, usually from the Luteoviridae, for genome encapsidation and plant-to-plant transmission [1, 2]. In recent years, several viral RNA sequences discovered from different hosts have been regarded as “umbra-like” viruses based on sequence homology to the RNA-dependent-RNA-polymerase (RdRp) [3-6]. However, biological and genomic differences warrant a further examination to determine whether these viruses should be included within the Umbravirus or belong to a distinct genus of umbra-related viruses. For example, papaya virus Q (PpVQ) and papaya meleira virus 2 (PMeV-2), two umbra-like viruses found in papaya plantings of Ecuador and Brazil, respectively, possess a ~ 4.4 kb genome containing two non-overlapping ORFs (ORF1 and ORF2) followed by an 1,800-nt-long non-coding region (NCR) [3, 4]. While no evidence for a luteovirid helper virus was found for PpVQ [7], a double-stranded RNA (dsRNA) virus, known as papaya meleira virus (PMeV), was found as the helper virus for PMeV-2 [4]. Interestingly, a similar umbra-like virus was reported in commercial papaya plants from Mexico and referred to as papaya meleira virus -Mexican variant (PMeV-Mx) [5]. As for genome features, citrus yellow vein associated virus (CYVaV), an unpublished umbra-like sequence record (GenBank Accession NC_040311), along with three recently described virus sequences namely opuntia umbra-like virus (OpUV), Ethiopian maize associated virus (EMaV) and sugarcane umbra-like virus (ScUV) possess shorter 2.6 – 2.9 kb genomes arranged similar to typical umbravirus genomes [6, 8]. Here, we report the complete genomic sequence of an umbra-like virus discovered from a mountain papaya species known as babaco (Vasconcellea x heilbornii, Fam. Caricaceae) and analyze its genomic and evolutionary relationships with other recently described umbra-like viruses.
High-throughput sequencing (HTS) was applied on an asymptomatic babaco plant (Vasconcellea x heilbornii cv. Pentagona), which was being prepared for tissue culture propagation as part of a virus-free certification project. Double-stranded RNA (dsRNA) was extracted from fresh leaf tissue as described [3] and subjected to a melting step (98C for 10 min) prior to generation of complementary DNA (cDNA) library [9]. HTS was done on a NextSeq 500 Illumina platform as single 75 bp reads. Sequences were trimmed and filtered using Trimmomatic v. 0.39 [10] and assembled using metaspades from SPAdes v3.13 [11]. The assembled contig was compared to NCBI GenBank using the Basic Local Alignment Search Tool (BLAST) and re-sequenced by the Sanger method using overlapping primers. Terminal sequences were confirmed by RACE using total RNA as a template for cDNA, and specific primers near the ends. Detection primers F: CGTGTGCTTGCTGGTTTTCGTTC and R: CAACGGGAAACCCATACACCTGG were designed to amplify a 1,055 nt fragment of the 3’ NCR and used to verify the presence of the virus-like RNA in the original plant as described [3].
The genome of the babaco umbra-like virus consisted of 4,584 nt (Genbank acc. numb. MT113181) with terminal sequences 5’-GGAUCCCUU and AUGCGCCC-3’. A 9-nt-long NCR was observed at the 5’ end, followed by two non-overlapping ORFs (ORF 1 and 2) with a 152 nt-long intergenic region. ORF 1 (frame 1, nt positions: 10 - 1,083) encodes a putative protein (P1) of 38 kDa for which no homology was found in the database. ORF 2 (frame 3, nt positions: 1,236 – 2,741) encodes a 57 kDa protein with RdRp motifs highly conserved across members of the Tombusviridae [12]. The heptameric motif GUCUUUU was identified 12 nt upstream ORF 1 stop codon, suggesting a potential -1 ribosomal frameshift (FS) mechanism for the translation of ORF 1 – ORF 2 as a fusion protein of 101 kDa. This FS mechanism has been shown to operate in umbraviruses, whose ORF 2 is characterized by the absence of an AUG start codon [13]. No additional ORFs were predicted downstream ORF 2; hence an 1,843 nt-long genome fragment remained as the 3’ NCR (Fig. 1a), similar to the genomes of PpVQ and PMeV-2.
RdRp sequence identities between the babaco- and the recently reported umbra-like viruses ranged from 40 to 70% at the nucleotide level, and from 41 to 73% at the amino acid level, with PpVQ as the closest relative. Considering the species demarcation criteria for umbraviruses –less than 75% aa identity for the viral RdRp [1] – we tentatively consider the babaco umbra-like virus as a distinct species from PpVQ, and herein use provisionally the name babaco virus Q (BabVQ) to refer to this virus.
Evolutionary relationships between BabVQ and members of the Tombusviridae were inferred by a Bayesian and maximum likelihood phylogenetic analysis of the RdRp. A total of 85 RdRp homologs were aligned and analyzed using BEAST v1.10.4 [14] and MEGA-X [15] under the Le-Gascuel (LG) evolutionary model plus gamma distribution and invariable sites. The topology of the tree revealed the existence of two well defined clades that included the umbra- and umbra-like viruses, respectively. Furthermore, two groups were formed within the umbra-like clade, one containing BabVQ and the papaya umbra-like viruses PpVQ, PMeV-2 and PMeV-Mx, and a second one containing CYVaV, OpUV, EMaV and ScUV (Fig. 1b).
We compared the genomes of the two closely related groups and found two striking similarities: i) the non-overlapping nature of ORFs 1 and 2 and ii) the presence of the potential ribosomal slippery signal located in a similar position (except for PMeV-2) with respect to the stop codon of ORF 1. However, two important differences were also observed between the two groups: i) BabVQ and closest relatives lack discernible ORFs 3 and 4 and ii) the genomes of the CYVaV group are considerably smaller (2.6 – 2.9 kb) (Fig. 2). Lastly, HTS failed to identify a potential helper virus either in the Luteoviridae, as is the case for typical umbraviruses, or any additional virus that may serve as CP assistor for BabVQ. This communication adds to growing evidence for the expansion of the Tombusviridae and the proposed creation [6] of a new genus for the classification of umbra-related capsidless viruses.