Differential regulation drives plasticity in sex determination gene networks

Abstract

Background: Sex determination networks evolve rapidly and have been studied intensely across many species, particularly in insects, thus presenting good models to study the evolutionary plasticity of gene networks.

Results: We study the evolution of an unlinked gene capable of regulating an existing diploid sex determination system. Differential gene expression determines phenotypic sex and fitness, dramatically reducing the number of assumptions of previous models. It allows us to make a quantitative evaluation of the full range of evolutionary outcomes of the system and an assessment of the likely contribution of sexual conflict to change in sex determination systems. Our results show under what conditions network mutations causing differential regulation can lead to the reshaping of sex determination networks.

Conclusion: The analysis demonstrates the complex relationship between mutation and outcome: the same mutation can produce many different evolved populations, while the same evolved population can be produced by many different mutations. Existing network structure alters the constraints and frequency of evolutionary changes, which include the recruitment of new regulators, changes in heterogamety, protected polymorphisms, and transitions to a new locus that controls sex determination.

Figure 1

Evolution of mutation pair f→f^-/a→A. The mutant A is introduced at low frequency in the genotype a/A;m/f^-. Through interbreeding, a range of other genotypes are produced in subsequent generations (represented by arrows). Phenotypic sex (i.e. indicated above the arrows) of these novel genotypes depends on the values of k and h (see Figure 2(e)). Three potential outcomes are possible: transition to a new sex determining locus (Region I), transition coupled with a change in heterogamety (Region II), or recruitment of allele A without change in the sex determining locus (Region III).

Figure 2

Evolutionary outcomes. Evolutionary outcomes depend heavily on k (the level of constitutive expression across both loci) and h (the steepness of the sigmoid function σ of Eq (2)) which jointly determine the change in gene expression at the D locus (${\widehat{S}}_D$) caused by the A regulatory allele at the R locus. For each mutation pair, we map regions in which there are transitions in sex determination locus (T), recruitment without transition (R), and changes in heterogamety (H).

Figure 3

Constraints on fitness. The graph shows male (W_M) and female (W_F) fitness values for recruitment and transition events for the mutation pair a→A/f→f^- (Region III, Figure 2a). Note that both male-favoring (W_M > 1, W_F < 1) and female-favoring (W_M < 1, W_F > 1) sexual conflict are observed in this case. The global constraint (W_M - 1) + (W_F - 1) > 0 limits permissible values for recruitment.

Figure 4

Protected polymorphisms. The distribution of mutations leading to protected polymorphisms, across the parameter space (k and h), is shown with curves delimiting each region as in Figure 2. Each point represents a particular simulation run that results in protected polymorphism.

Figure 5

D locus expression and sex. In these examples, we illustrate how ${\widehat{S}}_{\text{D}}$ defines sex for five different cases: (a) a→A/f→f^-, Region II, (b) a→A/f→f^-, Region III, (c) f→f^-/a→A, Region II, (d) f→f⁺/a→A, Region I (* indicates this genotype could be a/A;m/m or A/A;m/m as both genotypes have the same ${\widehat{S}}_{\text{D}}$ value), and (e) f→f^-/a→A, Region III. In (a) and (b) the A allele is homozygous in all genotypes (not shown).

Figure 6

Protected polymorphism and fitness. The mutation pair a→A/f→f^- in Region II (Figure 5(a)) either leads to recruitment of f^- (black points) when selection favors higher expression in males (w_M > 0) or protected polymorphism (green points) when selection favors lower expression in males (w_M < 0).