Direct nanopore RNA sequencing of umbra-like virus-infected plants reveals long non-coding RNAs, specific cleavage sites, D-RNAs, foldback RNAs, and temporal- and tissue-specific profiles

Abstract

The traditional view of plus (+)-strand RNA virus transcriptomes is that infected cells contain a limited variety of viral RNAs, such as full-length (+)-strand genomic RNA(s), (–)-strand replication intermediate(s), 3′ co-terminal subgenomic RNA(s), and viral recombinant defective (D)-RNAs. To ascertain the full complement of viral RNAs associated with the simplest plant viruses, long-read direct RNA nanopore sequencing was used to perform transcriptomic analyses of two related umbra-like viruses: citrus yellow vein-associated virus (CY1) from citrus and CY2 from hemp. Analysis of different timepoints/tissues in CY1- and CY2-infected Nicotiana benthamiana plants and CY2-infected hemp revealed: (i) three 5′ co-terminal RNAs of 281 nt, 442 nt and 671 nt, each generated by a different mechanism; (ii) D-RNA populations containing the 671 fragment at their 5′ends; (iii) many full-length genomic RNAs and D-RNAs with identical 3′end 61 nt truncations; (iv) virtually all (–)-strand reads missing 3 nt at their 3′ termini; (v) (±) foldback RNAs comprising about one-third of all (–)-strand reads and (vi) a higher proportion of full-length gRNAs in roots than in leaves, suggesting that roots may be functioning as a gRNA reservoir. These findings suggest that viral transcriptomes are much more complex than previously thought.