Abstract
We have identified two novel, closely related subfamilies of non-long-terminal-repeat (non-LTR) retrotransposons in Drosophila melanogaster, the Waldo-A and Waldo-B subfamilies, that are in the same lineage as site-specific LTR retrotransposons of the R1 clade. Both contain potentially active copies with two large open reading frames, having coding capacities for a nucleoprotein as well as endonuclease and reverse transcriptase activities. Many copies are truncated at the 5' end, and most are surrounded by target site duplications of variable lengths. Elements of both subfamilies have a nonrandom distribution in the genome, often being inserted within or very close to (CA)(n) arrays. At the DNA level, the longest elements of Waldo-A and Waldo-B are 69% identical on their entire length, except for the 5' untranslated regions, which have a mosaic organization, suggesting that one arose from the other following new promoter acquisition. This event occurred before the speciation of the D. melanogaster subgroup of species, since both Waldo-A and Waldo-B coexist in other species of this subgroup.
