Search for a command to run...
Two hundred seventeen field isolates of cucumber mosaic cucumovirus (CMV), sampled from 11 natural populations, were typed by RNase protection assay (RPA) using probes from the genomic RNAs of strains in subgroup I and in subgroup II of CMV strains. Most (85%) of the analyzed isolates belonged to subgroup I. For these subgroup I isolates, only two clearly different RPA patterns, A and B, were found for each of four probes representing RNA1, RNA2, and each of the two open reading frames in RNA3. On the basis of these RPA patterns for each probe, different haplotypes were defined. The frequency composition for these haplotypes differed for the various analyzed populations, with no correlation with place or year of sampling. This genetic structure corresponds to a metapopulation with local extinctions and recolonizations. Most subgroup I isolates (73%) belonged to haplotypes with RPA pattern A (type 1) or B (type 2) for all four probes. A significant fraction of subgroup I isolates (16%) gave evidence of mixed infections with these two main types, from which genetic exchange could occur. Genetic exchange by segment reassortment was seen to occur: the fraction of reassortant isolates was 4%, reassortment did not occur at random, and reassortants did not become established in the population. Thus, there is evidence of selection against reassortment between types 1 and 2 of subgroup I isolates. Aphid transmission experiments with plants doubly infected with type 1 and type 2 isolates gave further evidence that reassortment is selected against in CMV. Genetic exchange by recombination was detected for RNA3, for which two RPA probes were used. Recombinant isolates amounted to 7% and also did not become established in CMV populations. Sequence analyses of regions of RNA1, RNA2, and RNA3 showed that there are strong constraints to maintain the encoded sequence and also gave evidence that these constraints may have been different during divergence of types 1 and 2 and, later on, during diversification of these two types. Constraints to the evolution of encoded proteins may be related to selection against genetic exchange. Our data, thus, do not favor current hypotheses that explain the evolution of multipartite viral genomes to promote genetic exchange.