Contrasting Evolutionary Rates between Social and Parasitic Bumblebees for Three Social Effect Genes

Abstract

Reproductive division of labour, characterizing eusocial insect societies, is predicted to reduce efficacy of selection on genes involved in worker traits as selection will be mainly effective in the reproductive caste. Indirect selection acting through a share of genes due to relatedness might be a weak force. Thus, all being equal, population genetic theory adapted to genes with social effects on fitness predicts a higher polymorphism within and between populations on such genes compared to genes with direct effects on fitness. In order to test this prediction empirically, we use social and non-social bumblebee species, all belonging to the genus Bombus. Non-social bumblebees are social parasites of the social species, invading host nests during an early phase of colony development. The parasitic bee, lacking a worker caste, kills the resident queen and starts reproducing, utilizing the host workers in brood care. Three social-effect genes, foraging (for), salivary gland secretion 3 (sgs3) and vitellogenin (vg), were chosen as well as some control genes to be amplified in two social bumblebee species and in two parasitic (non-social) species. High coverage amplicon re-sequencing was used to determine levels of intra- and inter-specific polymorphism. Social effect genes show contrasting rates of evolution. While vg shows a pattern of slightly positive selection, but no sign of social pleiotropy, sgs3 is expected to be a socially pleiotropic gene. The signature of high intraspecific polymorphism, low divergence and the adaptive fixation of derived alleles, along with its expression pattern, predict a socially pleiotropic effect. By contrast, for shows high levels of conservation between social species, suggesting a role in worker behaviour but also functions in queens. This empirical test of evolutionary theory suggests that social effect genes might also operate in both castes resulting in difficulties to interpret effects on molecular evolution. No general molecular pattern of social effect genes could be detected without taking into account their function in each caste and the direction of selection they experienced. Nevertheless, social effect genes experienced a stronger selection pressure in social than in parasitic species.