CLASSIFICATION AND IDENTIFICATION OF GEMINIVIRUSES USING SEQUENCE COMPARISONS

The genomes and ORFs of 36 geminiviruses were compared to obtain phylogenetic trees and frequency distributions of all possible pairwise comparisons with an objective to classify geminiviruses. Such comparisons show that geminiviruses form two distinct clusters of leafhopper-transmitted viruses that infect monocots (subgroup I) and whitefly-transmitted viruses that infect dicots (subgroup III), irrespective of the part of the genome considered. Of the two leafhopper-transmitted viruses that infect dicots, tobacco yellow dwarf virus has a sequence most similar to subgroup I viruses, and that of beet curly top virus differed depending upon the ORF considered. The distributions of identities within subgroups are significantly different suggesting that the taxonomic status of a particular isolate within a subgroup can be quantified. All the recognized strains of any one virus have greater than 90 % sequence identity. It was observed that the 200 nucleotide intercistronic regions of geminiviruses are more variable than the remainder of the genome. The amino acid sequences of the coat protein (CP) of subgroup III viruses are more conserved than the remainder of the genome. However, a short N-terminal region (60-70 amino acids) of the CP is more variable than the rest of the CP sequence and is a close representation of the genome. PCR primers based on conserved sequences can be used to clone and sequence the N-terminal sequences of the CP of the geminiviruses; this sequence is sufficient to classify a virus isolate. A possible taxonomic structure for geminiviruses is proposed after considering the sequence comparisons and biological properties.