Social dilemma in the external immune system of the red flour beetle? It is a matter of time

Abstract

Sociobiology has revolutionized our understanding of interactions between organisms. Interactions may present a social dilemma where the interests of individual actors do not align with those of the group as a whole. Viewed through a sociobiological lens, nearly all interactions can be described regarding their costs and benefits, and a number of them then resemble a social dilemma. Numerous experimental systems, from bacteria to mammals, have been proposed as models for studying such dilemmas. Here, we make use of the external immune system of the red flour beetle, Tribolium castaneum, to investigate how the experimental duration can affect whether the external secretion comprises a social dilemma or not. Some beetles (secretors) produce a costly quinone-rich external secretion that inhibits microbial growth in the surrounding environment, providing the secretors with direct personal benefits. However, as the antimicrobial secretion acts in the environment of the beetle, it is potentially also advantageous to other beetles (nonsecretors), who avoid the cost of producing the secretion. We test experimentally if the secretion qualifies as a public good. We find that in the short term, costly quinone secretion can be interpreted as a public good presenting a social dilemma where the presence of secretors increases the fitness of the group. In the long run, the benefit to the group of having more secretors vanishes and becomes detrimental to the group. Therefore, in such seminatural environmental conditions, it turns out that qualifying a trait as social can be a matter of timing.

Keywords: Tribolium castaneum; public goods; social evolution.

Figure 1

Defining secretors versus nonsecretors: The individual quinone levels where assessed for 30 randomly sampled individuals from each replicate are represented as a dot. In total, the aim was to get 150 individuals from each treatment. In some cases, the population size was not large enough for these measurements. Even in cases where there are no secretors, there is a basal level of expression. As a conservative assumption, we set a level twice that of the maximum basal expression (given by the horizontal dashed line) to differentiate between secretors and nonsecretors. We then extrapolate this measure to the population sizes given in Figure 3

Figure 2

Comparing the observed quinone levels from the experiment with the expected levels calculated under linearity. If the secretor do not compensate for the nonsecreting beetles, then the amount of quinone expected in the populations would be linear as to the fraction of secretors in each population. The quinone levels lie between 0 and 1 as forming the ends of a continuum where 0 corresponds to q _ns = 1.82, the mean quinone level of a nonsecreting population and 1 to q _s = 57.75, the mean quinone level of a secreting population as from the seeding parental generation (see Section Results). Note that the expected quinone levels (fraction of secretors) are not constant as (0, 0.25, 0.5, 0.75, 1) but change with the fraction of secretors after 1 and 3 months as a result of the population dynamics (Figure 4). The expectation is based on the assumption that the production of quinones is unaffected or cannot be actively modulated by the secretor in response to the presence/absence of nonsecretors. The strong linear relationship between the shows that the secretors do not compensate/modulate their secretion for the presence of nonproducers. The solid lines are a linear fit to the data with a 95% confidence region shaded around it

Figure 3

Population size dynamics in the three treatments for the different initial ratios of secretors: nonsecretors. In the first month, the all‐secretors group achieves the largest population size in all three treatments, with the growth being monotonic in the number of secretors at the initial conditions. After 3 months, however, in all three treatments, the all secretors suffer population crashes. This could be due to the population reaching its ecological limit the earliest as compared to others due to the initially fast growth. Also, it is possible that the amount of quinone produced cannot be controlled individually by the beetles and thus a population crash results due to excessive quinones becoming toxic. Within each treatment (and control) panel, we performed an ANOVA between the distributions. The final increase (after 1 month) or decrease (after 3 months) is always significant. For the control, heat‐killed and Beauveria bassiana treatments after 1 month, the results were significant at p < 0.05 level F _4,20 = 22.39, p = 3.7 × 10⁻⁷, F _4,20 = 22.39, p = 4.8 × 10⁻⁶ and F _4,20 = 22.39, p = 5 × 10⁻⁸, respectively, indicated by like letters. After 3 months, the decrease in population size is significant again with F _4,20 = 16.71, p = 3.6 × 10⁻⁶, F _4,20 = 11.49, p = 5.2 × 10⁻⁵ and F _4,20 = 10.04, p = 1.2 × 10⁻⁴ for control, heat‐killed and B. bassiana treatments, respectively

Figure 4

Population dynamics is usually excluded from the analysis of evolutionary dynamics. The experiment begins at the points indicated by circles, the 1‐month time point represented by stars, and the 3‐month time point is given by squares. However, when we look at ecological and evolutionary dynamics together we see that all treatments experience, a growth in population size, which reaches around 300 individuals in almost all populations except with all secretors (and the one with 75% secretors in the heat‐killed treatment) before reducing. This might indicate that for the given environment, a population limit is reached with about 300 individuals. Furthermore, the more initial secretors the stronger the drop down, potentially indicating that the upper quinone‐threshold was crossed. This in combination with expending the environment finally results in extinction, whereas populations which start with less secretors grow slower and thus still survive after 3 months