Characterization of new endogenous geminiviral elements in yam (#Dioscorea# spp.) genomes

Geminivirus sequences integrated into the genomes of various plant species (Nicotiana species, Lactuca sativa, Malus domestica and Populus trichocarpa) have been reported. The best studied of these integrated geminivirus sequences are the so-called "geminivirus-related DNA" or GRD elements within the genomes of various Nicotiana species. Our study aimed at assessing whether such integrated geminiviral sequences are present within the genomes of yam species. We have used a polyphasic approach combining a first step of plant genome screening (in silico EST analysis and viral metagenomics) followed by a second step of molecular characterization of the integrated sequences (Long-PCR and Inverse-PCR). We have found two new phylogenetically related endogeneous geminivirus (EGV) elements within the genome of the majority of Asian Dioscorea spp. section Enantiophyllum. Based on multiple lines of evidence, such as (i) seed transmission of the two EGV sequences, (ii) no amplification by rolling circle amplification, (iii) worldwide presence of the two EGV sequences in a range of asympomatic Asian Dioscorea species and (iv) lack of sequences of coat protein genes (i.e. they are not likely to be insect transmissible), we hypothesized that those two EGV elements were likely to have become integrated within an ancestor of the Asian Dioscorea species. We then confirmed the presence of the two EGV elements within D. alata and D. nummularia genomes by fluorescence in situ hybridization. The predicted proteins expressed by the two EGV sequences had detectable homology to begomovirus replication enhancer and replication associated proteins but only one out of the two EGV sequences contains a GC rich sequence resembling the conserved hairpin structures found at geminivirus virion strand origins of replication (including the TAATATTAC sequence in the likely loop region) and is organized as 2.6 Kb tandem repeats. The discovery and analysis of such fossilized geminivirus sequences within plant genomes would certainly help to more accurately date events deep in the evolutionary history of the Geminiviridae family such as the most recent geminivirus common ancestor, and the origins of the various geminivirus genera. (Texte integral)