The origin and evolution of geminivirus-related DNA sequences in Nicotiana.

A horizontal transmission of a geminiviral DNA sequence, into the germ line of an ancestral Nicotiana, gave rise to multiple repeats of geminivirus-related DNA, GRD, in the genome. We follow GRD evolution in Nicotiana tabacum (tobacco), an allotetraploid, and its diploid relatives, and show GRDs are derived from begomoviruses. GRDs occur in two families: the GRD5 family's ancestor integrated into the common ancestor of three diploid species, Nicotiana kawakamii, Nicotiana tomentosa and Nicotiana tomentosiformis, on homeologous group 4 chromosomes. The GRD3 family was acquired more recently on chromosome 2 in a lineage of N. tomentosiformis, the paternal ancestor of tobacco. Both GRD families include individual members that are methylated and diverged. Using relative rates of synonymous and nonsynonymous nucleotide substitutions, we tested for evidence of selection on GRD units and found none within the GRD3 and GRD5 families. However, the substitutions between GRD3 and GRD5 do show a significant excess of synonymous changes, suggesting purifying selection and hence a period of autonomous evolution between GRD3 and GRD5 integration. We observe in the GRD3 family, features of Helitrons, a major new class of putative rolling-circle replicating eukaryotic transposon, not found in the GRD5 family or geminiviruses. We speculate that the second integration event, resulting in the GRD3 family, involved a free-living geminivirus, a Helitron and perhaps also GRD5. Thus our data point towards recurrent dynamic interplay between geminivirus and plant DNA in evolution.