The first steps of transposable elements invasion: parasitic strategy vs. genetic drift

Abstract

Transposable elements are often considered as selfish DNA sequences able to invade the genome of their host species. Their evolutive dynamics are complex, due to the interaction between their intrinsic amplification capacity, selection at the host level, transposition regulation, and genetic drift. Here, we propose modeling the first steps of TE invasion, i.e., just after a horizontal transfer, when a single copy is present in the genome of one individual. If the element has a constant transposition rate, it will disappear in most cases: the elements with low-transposition rate are frequently lost through genetic drift, while those with high-transposition rate may amplify, leading to the sterility of their host. Elements whose transposition rate is regulated are able to successfully invade the populations, thanks to an initial transposition burst followed by a strong limitation of their activity. Self-regulation or hybrid dysgenesis may thus represent some genome-invasion parasitic strategies.

Figure 1.—

Loss frequency and transposition rate. The solid line represents the analytical result of Kaplan et al. (1985) for an infinite population. Solid circles, simulation results under the same conditions, with a population size N = 50; open circles, simulation results when TEs are deleterious (s = −0.01). When the transposition rate tends to 0, the expected loss frequencies are respectively 1 − 2u for an infinite population, 1 − 1/2N = 0.99 for the neutral case, and 1 − (1 − exp(−2s))/(1 − exp(−4Ns)) = 0.9968 for the selection case.

Figure 2.—

Different burst and regulation models used in this work: the transposition rate u_n depends of the copy number n. (a) No transposition burst (u_n = u₀); (b) threshold self-regulation (u_n = u_b if n ≤ k_t; here k_t = 3, u_n = u₀ otherwise); (c) continuous self-regulation [ with k_c = 2].

Figure 3.—

Examples of copy number dynamics during the first 20 generations with different transposition rates, u = 10⁻⁴ (a and b), u = 1 (c and d), and u = 10 (e). Populations were initialized with only one copy. The left (a, c, and e) gives examples of simulations where TEs have been lost, and the right (b and d) gives examples of successful invasions. Over 1000 simulations, no invasion was reported for u = 10. For the two other transposition rates, the percentage of TE loss among 1000 repetitions is indicated on the central diagrams. Each small colored rectangle represents a single individual, and the color of the rectangle corresponds to its TE copy number (see the color scale on the bottom). For each of the 20 generations, the 50 individuals of the population are plotted in an arbitrary order. When the transposition rate is low (a and b), the element frequency evolves mainly through genetic drift, and the TE is lost in most cases. With intermediate transposition rates (u = 1), the probability of invasion increases; the element can still be lost by genetic drift during the very first generations (c), but the copy number increase due to the high transposition rate leads to a quick diffusion of the TE in all individuals (d). In the last situation (u = 10), the TE copy number in the few individuals carrying them is so high (>100) that they are sterile (denoted by *), and the element is not transmitted to the next generation.

Figure 4.—

(a) Evolution of the mean copy number per individual in successfully invaded populations for the different transposition burst models. Error bars represent standard errors. (b) Evolution of the mean transposition rate in the same set of populations. For both a and b, means and standard errors were computed on 672 successfully invaded populations (over 100,000 repetitions) for the “no-burst” model, 271 populations (1000 repetitions) for the random-burst model, 291 populations (1000 repetitions) for the hybrid dysgenesis model, 464 populations (1000 repetitions) for the threshold self-regulation model (k_t = 3), and 518 repetitions (1000 repetitions) for the continuous self-regulation (k_c = 2) model. Other simulation parameters are N = 50, u₀ = 10⁻⁴, u_b = 5, v = 0, and s = −0.01.

Figure 5.—

Maintenance frequencies after 1000 generations for the four models. (a) Random burst model; (b) hybrid dysgenesis model; (c) threshold model; (d) continuous regulation model. Two parameters are allowed to vary, the burst transposition rate (u_b) and the basal transposition rate (u₀). Population size N is fixed to 50, and selection coefficient s = −0.01.