Seasonality can induce coexistence of multiple bet-hedging strategies in Dictyostelium discoideum via storage effect

Abstract

The social amoeba Dictyostelium discoideum has been recently suggested as an example of bet-hedging in microbes. In the presence of resources, amoebae reproduce as unicellular organisms. Resource depletion, however, leads to a starvation phase in which the population splits between aggregators, which form a fruiting body made of a stalk and resistant spores, and non-aggregators, which remain as vegetative cells. Spores are favored when starvation periods are long, but vegetative cells can exploit resources in environments where food replenishes quickly. The investment in aggregators versus non-aggregators can therefore be understood as a bet-hedging strategy that evolves in response to stochastic starvation times. A genotype (or strategy) is defined by the balance between each type of cells. In this framework, if the ecological conditions on a patch are defined in terms of the mean starvation time (i.e. time between the onset of starvation and the arrival of a new food pulse), a single genotype dominates each environment, which is inconsistent with the huge genetic diversity observed in nature. Here we investigate whether seasonality, represented by a periodic, wetdry alternation in the mean starvation times, allows the coexistence of several strategies in a single patch. We study this question in a non-spatial (well-mixed) setting in which different strains compete for a common pool of resources over a sequence of growth-starvation cycles. We find that seasonality induces a temporal storage effect that can promote the stable coexistence of multiple genotypes. Two conditions need to be met in our model. First, there has to be a temporal niche partitioning (two well-differentiated habitats within the year), which requires not only different mean starvation times between seasons but also low variance within each season. Second, each season's well-adapted strain has to grow and create a large enough population that permits its survival during the subsequent unfavorable season, which requires the number of growth-starvation cycles within each season to be sufficiently large. These conditions allow the coexistence of two bet-hedging strategies. Additional tradeoffs among life-history traits can expand the range of coexistence and increase the number of coexisting strategies, contributing towards explaining the genetic diversity observed in D. discoideum. Although focused on this cellular slime mold, our results are general and may be easily extended to other microbes.3

Eco-evolutionary significance of “loners”

“…To do so, we embedded our population-partitioning model into an ecological framework of pairwise strain competition for resources over successive growth-starvation cycles [11,12,41] (S9 Fig; see Materials and Methods). Recent empirical and theoretical work has revealed several life-history traits that are likely to have ecological consequences, such as division rates, spore germination success, cell size, and the existence of loners [10][11][12]17,41,42]. Here, we aim to investigate only whether this newly proposed mechanism of loner allocation and its consequent developmental interactions could have an ecological impact and should therefore be included into ecological frameworks that explore D. discoideum diversity.…”

Eco-evolutionary significance of 'loners'

“…S8; see Methods). Although empirically we did not investigate fitness differences between strains, here we assume that such differences exist and depend on the environmental conditions as in prior work (Martínez-García & Tarnita, 2016, 2017; Tarnita et al, 2015). The environment is characterized by the mean time between nutrient replenishment events.…”

“…S4). Mathematically, the growth dynamics is given by a Monod-like equation: where the dot indicates the time derivative, X i is the population size for strain i , R is the amount of resources, and R1/2 is the abundance of resources at which the growth rate is half of its maximum Here, for simplicity, we assumed that both strains have the same maximum growth rate, although relaxing this assumption constitutes an important expansion towards a more complete understanding of life history traits and tradeoffs in slime molds (Martínez-García & Tarnita, 2016, 2017; J.B.B. Wolf et al, 2015).…”

“…Spores die at a constant and low rate d , whereas loners have a survival probability, S , that decays with time and reaches zero at a maximum survival time T sur . Analogous with (Martínez-García & Tarnita, 2017), we fit this maximum survival time as well as the functional shape of the survivorship curve using experimental data (Dubravcic, 2013) to obtain …”

“…Are loners mistakes—merely inevitable byproducts of large scale synchronization attempts—or are they a variable trait that selection can shape with potential ecological consequences? Theoretical investigations of such loner behaviors have been sparse, but the handful of existing studies have suggested that, at least in some systems, they could be a means of spatio-temporal niche-partitioning (Dubravcic, van Baalen, & Nizak, 2014) that promotes diversity (Martínez-García & Tarnita, 2017; Tarnita, Washburne, Martínez-García, Sgro, & Levin, 2015). However, despite this theoretically established potential, variability and heritability of loner behaviors have not been characterized in natural populations.…”