A central goal of biology is to uncover the genetic basis for the origin of new phenotypes. A particularly effective approach is to examine the genomic architecture of species that have secondarily lost a phenotype with respect to their close relatives. In the eusocial Hymenoptera, queens and workers have divergent phenotypes that may be produced via either expression of alternative sets of caste-specific genes and pathways or differences in expression patterns of a shared set of multifunctional genes. To distinguish between these two hypotheses, we investigated how secondary loss of the worker phenotype in workerless ant social parasites impacted genome evolution across two independent origins of social parasitism in the ant genera Pogonomyrmex and Vollenhovia. We sequenced the genomes of three social parasites and their most-closely related eusocial host species and compared gene losses in social parasites with gene expression differences between host queens and workers. Virtually all annotated genes were expressed to some degree in both castes of the host, with most shifting in queen-worker bias across developmental stages. As a result, despite >1 My of divergence from the last common ancestor that had workers, the social parasites showed strikingly little evidence of gene loss, damaging mutations, or shifts in selection regime resulting from loss of the worker caste. This suggests that regulatory changes within a multifunctional genome, rather than sequence differences, have played a predominant role in the evolution of social parasitism, and perhaps also in the many gains and losses of phenotypes in the social insects.

No abstract

No abstract

Persistent cooperation between unrelated reproductives occurs rarely in mature eusocial insect societies, and when present, is frequently geographically constrained. Here we present genetic and behavioral evidence showing that primary polygyny occurs in some, but not all populations of the honey ant Myrmecocystus mendax. Specifically, we found that all mature colonies sampled in a population in the Sierra Ancha Mountains of central Arizona (USA) were polygynous with a relatively high number of queens (average = 6.27), while the majority of mature colonies sampled in the Chiricahua Mountains of southeastern Arizona were monogynous. Field and laboratory observations showed that Chiricahua foundresses are primarily haplometrotic, whereas Sierra Ancha foundresses can be either haplometrotic or facultatively pleometrotic. Nestmate relatedness of mature Sierra Ancha field colonies suggests that the reproductive individuals within these colonies are unrelated, consistent with primary polygyny. In the laboratory, Sierra Ancha foundresses cooperatively established incipient colonies without queen reduction, and colonies with multiple queens produced more minims and workers that may serve the role of repletes (honeypots) than haplometrotic colonies. This was in stark contrast to foundresses from the Chiricahua population, which showed strong aggression when forced to found colonies together in the laboratory. When brood raiding was experimentally induced between laboratory Sierra Ancha colonies, queens from colonies with more workers had a higher survival probability, although in some cases the competing colonies fused and queens from both colonies continued to reproduce. Fusion between incipient ant colonies is a rare phenomenon, but could contribute to the high frequency of polygyny and high queen number in mature colonies in the Sierra Ancha population.

No abstract