Friday, February 08, 2008
Our adaptive immune system (thought to have evolved around the time of the earliest jawed vertebrates) functions by recognizing things in our bodies that aren't us and attacking them, which is why transplants and grafts of tissues that are different from our own tend to get rejected by our bodies. But this poses an interesting problem for the evolution of placental mammals (first pointed out by Peter Medawar in 1953): The fetus is genetically different from the mother, so before she can start carrying her progeny around inside her for long stretches of time there would have to be some mechanism in place to prevent her immune system from going into attack mode on it.
There are a few different ways this could plausibly be accomplished, but the one that evolution actually seems to have hit on is pretty neat, I think. One way we know it doesn't happen is by the mother somehow recognizing that the fetus carries half her genes, because otherwise IVF blastocysts implanted into surrogate mothers would spontaneously abort. So whatever is going on here is a "kin-blind" adaptation.
A significant chunk of our DNA had its origins as retroviral DNA. Most of these are now inactive, but a tiny portion actually appear to still code proteins. It's been found in mice, sheep and humans (and presumably generalizes to all placental mammals) that a particular kind of endogenous retrovirus is highly expressed in the outermost layer of the blastocyst (see e.g. Venables et al. 1995 for the human example). Furthermore, when you inhibit the expression of these genes the result is uniform spontaneous abortion immediately following implantation (Dunlap et al. 2006).
Most retroviruses are immunosuppressive, the most infamous example being HIV. Connecting the dots, it's quite plausible that these particular ancient retroviruses have been recruited into the mammalian genome and serve as local immunosuppressors in the uterus during development. In fact, we already know that syncytin, a protein crucial in placenta formation, is the product of a retroviral gene (Knerr et al. 2004), so there's nothing at all far-fetched about this. (In fact talking about these genes as if they were viruses just clouds the issue: The fact that they're now propagated in exactly the same way as the rest of your nuclear genome means that they're just as much your genes as any other bit of your DNA.)
The idea that viruses played a crucial role in the evolution of placental mammals is pretty nifty, but this is just the best-investigated case and there's circumstantial evidence suggesting that retroviruses have been involved in other major evolutionary innovations too. For instance, it turns out that eukaryotic DNA polymerases bear a closer structural resemblance to viral DNA polymerases than they do to those of eubacteria, suggesting that perhaps the genes of DNA viruses were recruited in the evolution of eukaryotic cellular machinery (Villarreal & Filippis 2000).
But around here we're more interested in human evolution, and there's some suggestive data on that score: Turns out human endogenous retroviruses are expressed in a wide range of tissues during development (Andersson et al. 2002; Muir, Lever & Moffett 2004); that retroviral promoters, enhancers & silencers inserted near genes can alter gene expression (Thornburg, Gotea & Makalowski 2006; Dunn, Medstrand & Mager 2003; Ting et al. 1992); and that sequence & phylogenetic analysis suggests they may be responsible for a significant portion of large-scale deletions and insertions on the genome (Huges & Coffin 2001). We're used to thinking of predators and parasites as indirect drivers of evolutionary change in organisms, but when the parasites can obtain direct access to their host's DNA this gets taken to a whole deeper level that's only recently been appreciated.