GIN transposons: genetic elements linking retrotransposons and genes.

In a previous work, we characterized a gene, called Gypsy Integrase 1 (GIN1), which encodes a protein very similar to the integrase domains present in Gypsy/Ty3 retrotransposons. I describe here a paralog of GIN1 and GIN2 and show that both genes are present in multiple vertebrates and that a likely homolog is found in urochordates. Surprisingly, phylogenetic and structural analyses support the counterintuitive idea that the GIN genes did not directly derive from retrotransposons but from a novel type of animal-specific DNA transposons, the GIN elements. These elements, described for the first time in this study, are characterized by containing a gene that encodes a protein that is also very similar to Gypsy/Ty3 integrases. It turns out that the sequences of the integrases encoded by GIN1 and GIN2 are more similar to those found in GIN elements than to those detected in retrotransposons. Moreover, several introns are in the same positions in the integrase-encoding genes of some GIN elements, GIN1 and GIN2. The simplest explanation for these results is that GIN elements appeared early in animal evolution by co-option of the integrase of a retrotransposon, they later expanded in multiple animal lineages, and, eventually, gave rise to the GIN genes. In summary, GIN transposons may be the "missing link" that explain how GIN genes evolved from retrotransposons. GIN1 and GIN2 may have contributed to control the expansion of GIN elements and Gypsy/Ty3 retrotransposons in chordates.