A co-opted gypsy-type LTR-retrotransposon is conserved in the genomes of humans, sheep, mice, and rats

One subset of sequences present within mammalian genomes is the retroelements, which include endogenous retroviruses and retrotransposons [1]. While there are typically thousands of copies of endogenous retroviruses within mammalian hosts, almost no LTR-retrotransposon-like sequences have been identified [2-4]. Here, we report the presence of a remarkably intact and conserved gypsy-type LTR-retrotransposon sequence within the genomes of several mammals, including humans and mice. Each host probably contains a single orthologous element, indicating that the original, ancestral gypsy LTR-retrotransposon first integrated into mammals over 70 million years ago. It is thus the first described example of a near-intact orthologous retroelement within humans and mice and is one of the most ancient retroelement sequences described to date. Despite their extreme age, the orthologs within each species examined contain a large ORF, between 4.0 and 5.2 kb in length, encoding proteins with sequence similarity to LTR-retrotransposon-derived Capsid (CA), Protease (PR), Reverse Transcriptase (FIT), RibonucleaseH (RNaseH), and Integrase (IN). Calculation of nonsynonymous and synonymous nucleotide substitution frequencies indicated that the encoded proteins are under purifying selection, suggesting that these elements have, in fact, been co-opted by their hosts. A possible function for these elements, involving gypsy LTR-retrotransposon restriction in mammals, is discussed.