Genome-wide expression patterns and the genetic architecture of a fundamental social trait

Abstract

Explaining how interactions between genes and the environment influence social behavior is a fundamental research goal, yet there is limited relevant information for species exhibiting natural variation in social organization. The fire ant Solenopsis invicta is characterized by a remarkable form of social polymorphism, with the presence of one or several queens per colony and the expression of other phenotypic and behavioral differences being completely associated with allelic variation at a single Mendelian factor marked by the gene Gp-9. Microarray analyses of adult workers revealed that differences in the Gp-9 genotype are associated with the differential expression of an unexpectedly small number of genes, many of which have predicted functions, implying a role in chemical communication relevant to the regulation of colony queen number. Even more surprisingly, worker gene expression profiles are more strongly influenced by indirect effects associated with the Gp-9 genotypic composition within their colony than by the direct effect of their own Gp-9 genotype. This constitutes an unusual example of an "extended phenotype" and suggests a complex genetic architecture with a single Mendelian factor, directly and indirectly influencing the individual behaviors that, in aggregate, produce an emergent colony-level phenotype.

Figure 1. Expression profiles between S. invicta adult workers with the BB and Bb genotypes of Gp-9.

Expression profiles for 39 differentially expressed genes are depicted (ANOVA, 10% false discovery rate [FDR]). Each row represents data for one gene, and each column represents data for a pool of 7–10 nestmates with the same Gp-9 genotype sampled from each of twenty colonies of each social form. Colonies were collected from Georgia (2004, checkered peach bar; 2006, solid peach bar) and Louisiana (2006, solid red bar), USA (data from each polygyne colony are presented in the identical order for the alternate genotype groupings). Expression levels for each gene are depicted relative to the average level across all experimental samples (blue, low levels; yellow, high levels). Genes are arranged by hierarchical clustering. See Text S1 and Table S1 for confirmation of selected gene expression results with quantitative RT-PCR (qRT-PCR).

Figure 2. Expression profiles between S. invicta adult workers from monogyne and polygyne colonies.

Expression profiles for 91 differentially expressed genes are depicted (ANOVA, 10% FDR). Rows and columns represent data as described in the Figure 1 caption. See Text S1 and Table S1 for confirmation of selected gene expression results with qRT-PCR.