In silico prediction and experimental validation of virus-derived small interfering RNAs and their putative host mRNA targets in infected tomato plants.

RNA silencing (RS) is a conserved mechanism in a broad range of eukaryotes. In plants, RS acts as an antiviral system and a successful virus infection requires suppression or evasion of the induced silencing response. Small interfering RNAs (siRNAs) accumulate in plants infected with RNA and DNA viruses and provide specificity to this RNA mediated immune system. High-throughput sequencing has contributed to expanding our knowledge of siRNA populations better describing their abundance, complexity and diversity in infected tissues. Virusderived siRNAs (vsiRNAs, 21-24 nt) from virus-infected plants are extraordinarily abundant and diverse. However, certain regions of viral genomes ("hot spots") are usually more represented than others in sequenced vsiRNA populations. Potato virus Y (PVY) is an important pathogen of solanaceous crops belonging to the largest plant virus family, Potyviridae. The PVY genome is a single-stranded, positive-sense RNA of about 10 kb. PVYC-to and PVY-SON41 are two PVY isolates that induce different disease symptoms on tomato (Solanum lycopersicum): while the former provokes severe leaf distortion, the latter produces in the same host no visible phenotype. This study propose an innovative in silico approach, which consisted in mining genomic regions of PVY isolates and looking at possible PVY vsiRNAs putatively able, by sequence complementarity, to targeting and suppressing accumulation of host messenger RNA (mRNA) as predicted by RS mechanisms, leading to dysfunctional biological processes that could explain isolate-specific disease phenotypes. A computational pipeline was implemented, which allowed the retrieval within the viral genome of 21-nt vsiRNAs from PVYC-to and PVY-SON41 isolates complementary to tomato predicted mRNA sequences (database Solgenomics, release ITAG2.3). The pipeline was based on five bioinformatic algorithms and a relational database (DBMS MySQL) for management of the results obtained in each steps. The comparative study for identifying the putative tomato target genes was performed with NCBI blast+ package 2.2 (option -task 'blastn-short' identity > 94%, max 2 mismatch or gap, alignment length > 19 bp). RandFold was employed for searching in the PVY genomes secondary structures containing putative vsiRNAs identified in previous steps. With this tool, five regions in the viral genome showing potential tRNA-like or microRNA-like secondary structures were identified, that might account for as many "hot spots" of vsiRNAs accumulation.