Posté : 12 févr.13, 20:36
Hello, Wayell. I am the author of the site Tiel linked you. You mention the restriction capabilities of KAP-1, and propose that it accounts for the shared ERVs between various lineages. This simply isn't the case. KAP-1 is a tripartite motif-containing antiviral protein (specifically TRIM28), and while TRIMs do restrict infection, they do not eradicate it. Further, the integration bias of any elements that do manage bypass it and integrate in the host genome remains unaffected. The most it could do is influence the degree to which a given lineage is infected with a particular retrovirus; not to cause parallel infections to result in identically positioned insertions. Locus specificity would still be required to account for identical ERV positioning.Wayell a écrit :Un dernier mot autours des Rétro virus endogènes.
Ce genre de virus dit "rétroéléments" exprime ces gènes par une protéine très spéciale, la KAP-1. Par rétroéléments, on entend un processus complexe activé des le début de embryogenèse. Cette protéine est la base du génome du virus endogène. Sans elle, les mutation sont irréfutablement et incontestablement délétères. L'infection est soit défective ou inactive chez le génome des eucaryotes. De ce fait, ils subsistent sur plusieurs milliers d'années, voire plus, par leurs attachement aux cellules somatiques ainsi que des cellules germinales. Le génome du rétrovirus peut être transmis à la génération suivante et persister dans la descendance au sein du génome de l'organisme hôte sans avoir d'incidence notable sur son développement (ou les cas problématiques seraient éliminés par la sélection naturelle).
Voilà comment on retrouve ce genre de ERVs à deux (plusieurs) époques différentes sur deux (plusieurs) espèces différentes.
Cela étant dit, la question de leurs origine reste en suspend !
In fact, even if integrase was locus specific, TRIMs cannot account for the hierarchical sharing of ERVs. Take TRIM5α, for instance. The first figure shows the restriction specificity of hominoid and a few OWM TRIM5αs for reconstructed PtERV1:
The second figure shows the distribution of PtERV1 elements (also among hominoids and a few OWMs). Notice that restriction efficiency doesn't remotely correlate with PtERV1 distribution; while humans are efficient/PtERV1-lacking and gorillas and baboons are inefficient/PtERV1-possessing, chimpanzees are efficient/PtERV1-possessing and orangutans and gibbons are inefficient/PtERV1-lacking. Far from restriction specificity predicting element endogenization/fixation, literally every combination is present in the data.
There simply isn't any indication that the innate immune system can account for the observed hierarchical grouping of shared ERVs (let alone their sharing in the first place).
It is a matter of observation that base pair reuse is low; not a convenient grasping of straws. Here, I'll post a kind of general overview I wrote a while back on how ERVs demonstrate common ancestry:Wayell a écrit :Je n'ai jamais pu comprendre ce besoin de croire à la contingence, faisant du hasard la clé de toutes chose. Et je ne comprends toujours pas cette manière de se dérober des faits scientifiques avérés.
- As for target site preference in general, it need only be rare that insertion occur at the same loci twice (i.e. that integrase reuse base pairs twice) for identical ERV distribution between chimpanzees and humans to demonstrate common ancestry. This is because of the thousands of insertions in the human genome, ~0.1% are lineage specific; the remainder (~99.9%) are in identical loci (IHGS Consortium, 2001; CSA Consortium, 2005).
And all that is required for the hierarchical distribution of ERVs among catarrhines to demonstrate common ancestry is for it to be rare that integrase reuses base pairs many times (4+), since there are numerous ERVs at orthologous loci in all catarrhines (Lebedev et al, 2000; Hughes & Coffin, 2005).
But how rare is retroviral base pair reuse? The most charitable study I am aware of is Wang et al. (2007), which found that of 40,569 insertions, only 41 were observed in the same spot twice (i.e. reused a base pair twice), and none reused a base pair three or more times. Now, only half of all events were observed, since retroviruses insert in either direction (sense or antisense) in equal proportion and without bias (Zhang et al, 2008; Brady et al, 2009), but that still puts the 2x reuse rate at 0.2%.
So even with this generous figure of 0.2% 2x reuse rate, chimpanzee-human sharing should be 0.2%. But remember, it's actually ~99.9%. And there should be no 3x+ reuse, yet there are many.
Simply put, retroviruses very rarely insert in the same place twice, yet almost all of the tens of thousands of human ERVs are in the exact same spot in chimpanzees. And (even in a huge sample of 40 thousand+) retroviruses don't insert in the same place three or more times, yet there are many ERVs in the same spot in many different primates (including humans).