Variability and phylogeny of the L1 capsid protein gene of human papillomavirus type 5: contribution of clusters of nonsynonymous mutations and of a 30-nucleotide duplication.

We analyzed the variability and established the phylogeny of the L1 capsid protein gene of 33 isolates of human papillomavirus type 5 (HPV5) obtained from epidermodysplasia verruciformis patients from different continents. By comparing the sequences of a 419-bp fragment with those published for two Japanese isolates, we found 12.9% variable nucleotide positions, defining 25 variants with mutation rates ranging from 0.2 to 8.8%. Such a high intratypic diversity is unusual among HPVs. Nine of the 139 encoded amino acids were variable and 12 protein variants were identified. Fifteen of the 16 substitutions observed were clustered in two short regions. A 9-amino-acid insert, already reported for the Japanese HPV5b isolate, was found within one of the regions in five isolates. Our data support that the insert arose from the duplication of a 30-nucleotide sequence. Phylogenetic trees distributed the DNA variants into three subtypes (a to c) with a divergence higher than 4.5% and allowed the recognition of European and African lineages. By contrast with the trees based on the HPV5 E6 gene, HPV5a DNA variants and the HPV5b variants lacking the insert constituted a single group in the L1 amino acid tree, probably reflecting different levels of structural constraints for the HPV5 L1 and E6 proteins. In that respect, the short variable L1 sequences should represent less constrained regions.