The repetitive ETn (early transposon) family of sequences represents an active "mobile mutagen" in the mouse genome. The presence of long terminal repeats (LTRs) and other diagnostic features indicate that ETns are retrotransposons but they contain no long open reading frames or documented similarity to the genes of known retroviruses or other retroelements. Thus, the mechanisms responsible for the mobility of this family have been unknown. In this study, we used computer searches to detect a small region of previously unrecognized type D retroviral pol homology within ETn elements. This small region was used to isolate two mouse endogenous proviral elements with gag, pro, and pol genes similar to simian type D viruses. This new family of mouse endogenous proviruses, termed MusD, is present in several hundred copies in the genome. Interestingly, the MusD LTRs, 3 internal region, and the 5 region expected to contain the packaging signal are very closely related to members of the ETn subfamily that have recently transposed. Analysis of different mouse strains indicates that MusD elements predate the existence of the mobile subfamily of ETns. These findings indicate that the ETn family was likely created via recombination events resulting in a near complete substitution of MusD coding sequences with unrelated DNA. Furthermore, these results suggest that ETn transcripts retrotranspose using proteins provided by MusD proviruses.ETn (early transposon) elements were first described in 1983 as a family of middle repetitive sequences transcribed during early mouse embryogenesis (4). ETn expression peaks between 3.5 and 7.5 days and is found primarily in undifferentiated cells of the inner cell mass and embryonic ectoderm (3). These elements were initially classified as retrotransposon-like because they contain long terminal repeats (LTRs) and retrovirus-like primer binding sites and are flanked by target site duplications (11,26). However, sequence analysis of full-length copies revealed no long open reading frames (ORFs) and no significant homology to known retroviral genes (26). Although this enigmatic structure might indicate that these elements are old and extensively mutated, this is not the case. Copies cloned at random can be closely related to each other, which suggests a relatively recent dispersion in the genome. Furthermore, it is evident that some ETn elements remain active as retrotransposons. At least eight mouse mutations at different loci are due to ETn insertions (1,9,10,18,24,27,28) and several somatic insertions have also been reported (17,22,29). Despite the bona fide mutagenic activity of ETns, very little has been done to investigate their mode of retrotransposition. ETn transcripts are presumably recognized by reverse transcriptase and other proteins encoded by another type of endogenous retrovirus or retrotransposon, but the identity of these putative coding-competent elements is unknown. Interestingly, it was noted several years ago (23) that new ETn insertions into the immunoglobulin (Ig) region...
