Endogenous retroviruses (ERVs) are "a unique combination of pathogen and selfish genetic element ." I discuss ERVs often on the blog because they truly intrigue me. A couple of weeks ago I talked about retrotransposons, which are transposable sequences derived from ancient viral infections through the integration of the viral genome into the germline. A recent PNAS paper  states that ERVs that lose the env gene behave like retrotransposons.
"ERVs can replicate both as transposable elements (TEs) and viruses. Some lineages replicate by an entirely intracellular mechanism and are functionally indistinguishable from the class of TEs called LTR-retrotransposons, whereas others do so within the host germline using cell reinfection in the same manner as the copying within somatic cells of exogenous retroviruses (XRVs)."I now realize that I had missed one important point about ERVs. I thought that, once inserted into one of the germline cells, it had to be that same infected cell that got fertilized in order for the viral sequence to became part of the host's genome. My new understanding, in light of this last paper I read, is that once integrated into the germline, the virus can replicate within the cell line, making its presence in any future fertilized egg a much more likely event.
The protein responsible for viral cell entry is the envelope, coded by the env gene. Whether the ERV will replicate as a retrotransposons or by reinfection is determined by the integrity of its env protein.
"We can assume that an ERV lineage with a functional env is reinfecting, whereas an ERV lineage with a disintegrated env is retrotransposing (whether reinfection can include germline cells in other host individuals of the same or other species is not known). Some retroviruses with a defective env are able to reinfect by “hitchhiking” the functional env of a coinfecting retrovirus, a mechanism known as “complementation”. However, complementation does not appear to be common in ERVs."In , Magiorkinis et al. look at different ERV lineages within 38 mammalian genomes and ask what controls the relative abundance of ERV lineages. They look at a relatively young group of ERVs, called IAPs. Though these sequences were originally discovered with a degraded env, and hence behaved as retrotransposons, later similar loci with an intact env within the same group were identified, and one in particular was shown to be able to reinfect cells.
"We find repeated transformations from reinfecting into retrotransposing ERVs and show that this transformation results in a rapid proliferation within the genome. Considering our results together with those from studies of transmission diversity in infectious disease epidemics, we propose that retrotransposition is the trait that leads ERVs to become genomic superspreaders."The results presented in  suggest that, once integrated, ERVs initially expand in the host through reinfection, though eventually the env loses its functionality and they become intracellular retrotransposons. A legitimate question is whether the loss of env is a cause or a consequence of the shift to retrotransposition: at least in mouse IAPs, loss of env has been shown to be a consequence. Env degradation enhances intracellular mobility but lowers the chances of interhost transmission. Even though in the literature there have been cases where ERVs with a degraded env were able to capture the functional env of a coinfecting retrovirus, none were found among the IAPs in this study.
The most interesting result in this study is the positive correlation between loss of env and proliferation. This could be due to the fact that coinfection is more likely to confer a disadvantage to the host, thus favoring the shift from coinfection to retrotransposition. It could also be that, compared to coinfection, retrotransposition favors integration into the germline.
 Magiorkinis, G., Gifford, R., Katzourakis, A., De Ranter, J., & Belshaw, R. (2012). From the Cover: Env-less endogenous retroviruses are genomic superspreaders Proceedings of the National Academy of Sciences, 109 (19), 7385-7390 DOI: 10.1073/pnas.1200913109