Identification and characterization of an endogenous pararetrovirus in Gentiana triflora plants.

Deep sequencing and de novo assembly of small RNAs (sRNA analysis) provide a powerful approach to detect plant RNA and DNA viruses. Here, using sRNA analysis of Gentiana triflora (gentian) plants affected with kobu-sho disease, a serious disease of gentian, we found caulimovirus-like sequences, provisionally designated as gentian caulimovirus-like sequence (GCVS). Analysis of the complete nucleotide sequence indicated that GCVS is a novel virus with a genomic structure and organization typical of members of genus Caulimovirus (family Caulimoviridae). Based on Southern blot hybridization and rolling circle amplification analyses, GCVS seems to be integrated into the genomes of a wild gentian plant and three F₁ gentian cultivars and their respective parents and without any episomal viral forms in plants. Consistently, no episomal virus was detected in gentian plants with kobu-sho disease. To confirm GCVS integration in the gentian genome, we used a BLASTn analysis of the gentian draft genome (4,374 contigs; total: 3,657,985,820 nucleotides [nt]) using the GCVS sequence as a query and identified 1,929 GCVS-derived fragments (total: 3,116,680 nt) across 280 contigs. All fragments were shorter than the full-length GCVS genome (8,125 nt), and more than 90% were less than 4,000 nt long. Open reading frames in these fragments encoded proteins containing numerous amino acid substitutions and were frequently disrupted by premature stop codons. Collectively, these features indicate that GCVS represents a non-infectious endogenous pararetrovirus, defined as a pararetroviral sequence that has been integrated into the host genome and is no longer capable of autonomous replication or infection.