Stabilizing multicellularity through ratcheting

Abstract

The evolutionary transition to multicellularity probably began with the formation of simple undifferentiated cellular groups. Such groups evolve readily in diverse lineages of extant unicellular taxa, suggesting that there are few genetic barriers to this first key step. This may act as a double-edged sword: labile transitions between unicellular and multicellular states may facilitate the evolution of simple multicellularity, but reversion to a unicellular state may inhibit the evolution of increased complexity. In this paper, we examine how multicellular adaptations can act as evolutionary 'ratchets', limiting the potential for reversion to unicellularity. We consider a nascent multicellular lineage growing in an environment that varies between favouring multicellularity and favouring unicellularity. The first type of ratcheting mutations increase cell-level fitness in a multicellular context but are costly in a single-celled context, reducing the fitness of revertants. The second type of ratcheting mutations directly decrease the probability that a mutation will result in reversion (either as a pleiotropic consequence or via direct modification of switch rates). We show that both types of ratcheting mutations act to stabilize the multicellular state. We also identify synergistic effects between the two types of ratcheting mutations in which the presence of one creates the selective conditions favouring the other. Ratcheting mutations may play a key role in diverse evolutionary transitions in individuality, sustaining selection on the new higher-level organism by constraining evolutionary reversion.This article is part of the themed issue 'The major synthetic evolutionary transitions'.

Keywords: evolution; major transition; multicellularity; ratcheting; stability.

Figure 1.

Schematic showing the effects of evolution in an E_G environment on the fitness of I and G cells in environments E_G and E_I. (a) Evolution of G cells in an E_G environment leads to increased fitness in both E_G and E_I environments, though the effect is smaller in E_I. These fitness changes have no consequences on the fitness of I cells in either environment. (b) The addition of ratcheting effects couples increases in G cell fitness with decreases in I cell fitness in both E_I and E_G. Ultimately, the effect is that the relative advantage of I cells (derived from G cells by mutation) in E_I is significantly decreased while the relative advantage of G cells in E_G is increased.

Figure 2.

Ratcheting type 1 increases the stability of multicellularity. (a) The duration of G cells in an E_I environment is shown as a function of the duration of growth in the E_G environment. Each point is the median of 100 simulations. If type 1 ratcheting mutations do not occur (red) then the duration in E_G has only a small effect on the stability of multicellularity by removing all pre-existing I cells from the population. By contrast, if ratcheting type 1 mutations occur (blue) there is a much larger increase in the stability of the multicellular form. Increased duration of growth in E_G leads to increased accumulation of ratcheting traits and greater multicellular stability. (b) An empirical cumulative distribution function plot shows the effect of the duration of growth in E_G on the variation in the persistence of multicellularity when ratcheting mutations occur. Depending on the magnitude and number of ratcheting mutations that fix in the population, the stability of multicellularity can be three to five times greater than the median. (c) For comparison, a similar plot is shown when there are no ratcheting mutations.

Figure 3.

The case when I cells become less fit than G cells in the E_I environment. (a) As a result of G cells evolving in an E_G environment, the evolution of ratcheting traits drives the fitness of I cells in E_I below G cells. (b) The consequence of this is that once such mutations fix, there is no selective benefit for G cells to revert back to I cells even when grown in an E_I environment. The time it takes for I cells to occupy 99% of the population is shown by the blue curve. Each point is the median of 100 simulations. Simulations were run for only 300 rounds so a value of 300 means that G cells are present for the entire duration of the simulation. For comparison, the red curve shows the case without type 1 ratcheting mutations. (c) An empirical cumulative distribution function plot shows the variation in the stability of multicellularity for different durations of growth in E_G. The value of each curve at 300 shows the percentage of simulations in which I cells eventually dominated the population. Those that do not reach 100 correspond to simulations in which G cells remained present.

Figure 4.

Selection for lower probability of switching. The probability of switching between I and G cells is shown as a function of the number of rounds grown in E_G. Each curve is the median of 10 evolved simulations and colours correspond to different c values—fitness differences between I and G cells—such that blue is c = 0.1, red is c = 0.2 and black is c = 0.9. All populations evolve lower probabilities of switching, starting at p = 10⁻¹ and evolving close to p = 10⁻³, which is the same value as the probability that a mutation changes the probability of switching.

Figure 5.

Combining ratcheting types. (a) Type 1 ratcheting can promote type 2 ratcheting. The fraction of maximal growth rate, as determined by the largest eigenvalues of equation (2.3), is shown as a function of the switch rate p for different values of c_i (c_g is fixed at 0.1). The blue curve shows that when c_g = c_i = 0.1, the optimal switch rate is p ≈ 0.2. When c_g > c_i, as a consequence of ratcheting type 1 mutations, then the optimal switch rate is p < 10⁻⁶. The red (c_i = 0.07), green (c_i = 0.05) and black (c_i = 0.01) curves show that as the fitness asymmetry increases there is stronger selection against switching frequently. (b) Type 2 ratcheting can promote type 1 ratcheting. The probability of finding a beneficial mutation to overcome a fitness gap of c is shown as a function of the switch rate p for different values of c. Each curve represents a different fitness gap (blue is c = 0.1, red is c = 0.2, green is c = 0.3 and black is c = 0.5) and is scaled by the probability of finding a beneficial mutation when p = 1, i.e. the worst case scenario. Thus, the vertical axis shows the factor of improvement when switching is lowered from p = 1. The chance of finding a beneficial mutation to overcome c increases as the switch rate is lowered, which can result from type 2 ratcheting.