In addition to these major promoter revisions, the expansion of the A-type promoter array is known to have gone through several stages of increasing promoter number (Schichman, et al., 1993). Of course, if major changes to promoters correlate to interesting phenomena such as generation of new lineages, then so could point mutations. So there is a lot of room to explore adaptive changes in LINE-1 promoter sequences.
The human LINE-1 promoter is often considered as yet another separate promoter. The degree of divergence expected between mouse and human is such that an ancestral relationship would not be detectable, however, there is now data on LINE-1 promoters in other orders tending to support that LINE-1 promoter exchanges may be a general phenomenon. The human LINE-1 promoter is a monomer and there are not yet any recognized alternative promoters on human LINE-1s. However, the time scale of events for human LINE-1 is about 10x slower than for mouse LINE-1, and so it might take reconstruction of much more divergent human LINE-1 sequences to recover an otherwise comparable history of promoter switches. Reconstruction of ancient human (primate) LINE-1 has been undertaken by Smit (Smit et al., 1995), but only in a very small portion of the polymerase domain.
Any slow phase in the expansion could be delineated by looking at Mus species that split out during the period. There had to be a period before the oldest MuERVC sequence sampled wherein the population frequency of the element rose from an initial introduction into one animal to levels approaching 1/genome. If this period developed slowly, then other Mus (or rat) species might have captured completely separate clades of MuERVC.
Endogenous retroviral families often are associated with a larger solo LTR family. We haven't looked yet to get a count of solo LTRs for MuERVC, but this would obviously be an important correction for how much excision has altered the perception of what copy numbers were like during the radiation.
Whether or not trans-mobilization and recombination with defective variants can have been a factor in the decline of the family depends on how close to equilibrium it was in the first place. If there is vigorous expansion, then one would expect the selfish selection to be able to select out effectively replicating sequences regardless of these other burdens. On the other hand, if the replication of the element is being held close to an equilibrium point by other forces, then an additional burden of entanglement with a growing population of defective variants might be enough to bring the family to an end.
Of course there's still a question of whether the family really is at an end, or if the frequency of nondefective sequences has just declined. All of the sequences that we have sampled have frame defects. It would be possible to screen for a nondefective surviving copy by dropping PCR products into a reporter system such that frame integrity would be monitored.
Variation in output from specific lineages
Some of the best evidence for lineages suddenly expanding or dying out comes from work by Carol Casavant and Holly Wichman in the Peromyscus species group (Casavant et al., 1996).
Exploring the MuERVC Radiation
One question is: how rapid was this amplification? One can imagine an almost instantaneous amplification, especially if horizontal transfer (animal to animal infection) played a major role. Alternatively, the amplification could have been more similar to the pace of transposon amplification. The ectopic retroviruses show an efficiency of making new inserts that might not be too far off from the LINE-1 amplification rate (details). However, the immune system supplies part of the restraint in that system. So one could imagine that the initial spread through the population was rapid and consisted mainly of infection of non immune pregnant females. Once the element reached an endogenous level of one per genome, females might become more efficiently immunized by endogenous activation and hence a slow phase of amplification might set in.