Regulatory logic and transposable element dynamics in nematode worm genomes

Abstract

Genome sequencing has revealed a tremendous diversity of transposable elements (TEs) in eukaryotes but there is little understanding of the evolutionary processes responsible for TE diversity. Non-autonomous TEs have lost the machinery necessary for transposition and rely on closely related autonomous TEs for critical proteins. We studied two mathematical models of TE regulation, one assuming that both autonomous tranposons and their non-autonomous relatives operate under the same regulatory logic, competing for transposition resources, and one assuming that autonomous TEs self-attenuate transposition while non-autonomous transposons continually increase, parasitizing their autonomous relatives. We implemented these models in stochastic simulations and studied how TE regulatory relationships influence transposons and populations. We found that only outcrossing populations evolving with Parasitic TE regulation resulted in stable maintenance of TEs. We tested our model predictions in Caenorhabditis genomes by annotating TEs in two focal families, autonomous LINEs and their non-autonomous SINE relatives and the DNA transposon Mutator. We found broad variation in autonomous - non-autonomous relationships and rapid mutational decay in the sequences that allow non-autonomous TEs to transpose. Together, our results suggest that individual TE families evolve according to disparate regulatory rules that are relevant in the early, acute stages of TE invasion.

Fig. 1.

Autonomous TEs are characterized by competent recognition sequences including target site duplications (TSDs) and terminal inverted repeats (TIRs). They also encode full-length enzymes, here transposase. The transposase protein must be shared between autonomous and non-autonomous relatives and becomes a limiting resource in transposition. Non-autonomous TEs have independent and combined deletions of the sequences necessary to produce competent enzymes and TSD and TIR recognition sequences.

Fig. 2.

Under A) a parasitic model of regulatory logic autonomous TEs self-attenuate transposition while their non-autonomous relatives continue to increase transposition rates as genome-wide transposition increases. Under B) a competitive model of regulatory logic both autonomous and non-autonomous transposons increase transposition rates as genome-wide transposition increases. Both autonomous and nonautonomous TEs share the same function but are plotted here with a small displacement for visualization.

Fig. 3.

When autonomous and non-autonomous transposons operate according to (A) the same competitive regulatory logic the deterministic model predicts a small number of transposons will increase until they stabilize in the population. The two TEs have identical numerical trajectories and are plotted on top of one another. (B) When the non-autonomous transposon parasitizes their autonomous relative, the transposons stabilize in the population but it takes more than twice as many generations. At a stable equilibrium the non-autonomous transposon is more abundant than the autonomous transposon. Here, $u_i=0.01,{\nu }_i=0.001$ and the selective effect of each TE is s = 0.001.

Fig. 4.

Across the range of population genetic parameters the simulated populations evolved according to 4 trajectories: 1) Early TE Loss; 2) Eventual TE Loss; 3) TE proliferation, fitness loss and eventual population extinction; or 4) Stable Maintenance

Fig. 5.

When the population is large (here, N=1000 individuals) and dioecious, populations evolving with (A-B) competitive TE regulation gradually lose TEs. In contrast, populations evolving with (C-D) parasitic TE regulation stably maintain a low number of TEs.

Fig. 6.

Large self-fertile populations evolving with competitive TE regulation evolved according to different trajectories. (A-B) Ten of the 100 populations retained both TEs for a period of time, eventually losing both autonomous and non-autonomous TEs. (C-D) Thirty-eight of the populations retained their autonomous TE but lost their non-autonomous TE early in the evolution of the simulated populations. Those populations experienced TE proliferation, fitness loss and eventual population crashes.

Fig. 7.

Autonomous LINE elements are more abundant than mutated non-autonomous SINE elements in both androdioecious and dioecious Caenorhabditis genomes.

Fig. 8.

The majority of LINE, SINE, Mutator and Mutator-origin MITE sequences are truncated fragments of less than 500bp in length.