Support for the reproductive ground plan hypothesis of social evolution and major QTL for ovary traits of Africanized worker honey bees (Apis mellifera L.)

Abstract

Background: The reproductive ground plan hypothesis of social evolution suggests that reproductive controls of a solitary ancestor have been co-opted during social evolution, facilitating the division of labor among social insect workers. Despite substantial empirical support, the generality of this hypothesis is not universally accepted. Thus, we investigated the prediction of particular genes with pleiotropic effects on ovarian traits and social behavior in worker honey bees as a stringent test of the reproductive ground plan hypothesis. We complemented these tests with a comprehensive genome scan for additional quantitative trait loci (QTL) to gain a better understanding of the genetic architecture of the ovary size of honey bee workers, a morphological trait that is significant for understanding social insect caste evolution and general insect biology.

Results: Back-crossing hybrid European x Africanized honey bee queens to the Africanized parent colony generated two study populations with extraordinarily large worker ovaries. Despite the transgressive ovary phenotypes, several previously mapped QTL for social foraging behavior demonstrated ovary size effects, confirming the prediction of pleiotropic genetic effects on reproductive traits and social behavior. One major QTL for ovary size was detected in each backcross, along with several smaller effects and two QTL for ovary asymmetry. One of the main ovary size QTL coincided with a major QTL for ovary activation, explaining 3/4 of the phenotypic variance, although no simple positive correlation between ovary size and activation was observed.

Conclusions: Our results provide strong support for the reproductive ground plan hypothesis of evolution in study populations that are independent of the genetic stocks that originally led to the formulation of this hypothesis. As predicted, worker ovary size is genetically linked to multiple correlated traits of the complex division of labor in worker honey bees, known as the pollen hoarding syndrome. The genetic architecture of worker ovary size presumably consists of a combination of trait-specific loci and general regulators that affect the whole behavioral syndrome and may even play a role in caste determination. Several promising candidate genes in the QTL intervals await further study to clarify their potential role in social insect evolution and the regulation of insect fertility in general.

Figure 1

The relationship between ovary size and degree of ovary activation under queenless conditions in Africanized backcross workers that are characterized by large ovaries. Almost all workers have developed their ovaries to some extent. The workers with the largest ovaries only had an activation score of three in contrast to some workers with smaller ovaries that contained fully developed eggs.

Figure 2

In both parallel backcrosses one major QTL for transgressive ovary size in worker honey bees was identified. However, these major QTL were on different chromosomes in the ABC3 backcross (a) and the ABC5 backcross (b).

Figure 3

In both parallel backcrosses one significant QTL for ovary asymmetry in worker honey bees was identified. The ABC3 QTL was located on the 4^th chromosome (a) and the ABC5 QTL was located on the 5^th chromosome (b), indicating no overlap between the backcrosses.

Figure 4

An additional significant QTL for ovary size was identified in the ABC5 backcross population. This QTL is identical to the behavioral pln1 QTL, demonstrating pleiotropy as predicted by the reproductive ground plan hypothesis of social evolution in honey bees.

Figure 5

Possible genetic model of the inheritance of the observed transgressive ovary sizes, focusing on the two major QTL on chromosome 6 and chromosome 11 in the two backcrosses ABC3 and ABC5. Possibly different alleles are indicated by subscripts. The crossing scheme does not allow for different alleles from the Africanized ancestor to segregate at one locus. Thus, the differences between ABC3 and ABC5 must be caused by dominance or epistasis effects.