A Search for Parent-of-Origin Effects on Honey Bee Gene Expression

Abstract

Parent-specific gene expression (PSGE) is little known outside of mammals and plants. PSGE occurs when the expression level of a gene depends on whether an allele was inherited from the mother or the father. Kin selection theory predicts that there should be extensive PSGE in social insects because social insect parents can gain inclusive fitness benefits by silencing parental alleles in female offspring. We searched for evidence of PSGE in honey bees using transcriptomes from reciprocal crosses between European and Africanized strains. We found 46 transcripts with significant parent-of-origin effects on gene expression, many of which overexpressed the maternal allele. Interestingly, we also found a large proportion of genes showing a bias toward maternal alleles in only one of the reciprocal crosses. These results indicate that PSGE may occur in social insects. The nonreciprocal effects could be largely driven by hybrid incompatibility between these strains. Future work will help to determine if these are indeed parent-of-origin effects that can modulate inclusive fitness benefits.

Keywords: gene expression; genomics; honey bees; imprinting; kinship theory; social insects.

Figure 1

Experimental design. A reciprocal cross design was used to uncouple parent-of-origin effects from lineage-of-origin effects. Two colonies were produced with Africanized maternity (AE) and two colonies with European maternity (EA). As a result, F1 workers from both colonies will be genotypically identical, but the parent-of-origin for each allele varies across colonies. Predicted allelic expression is shown for sites with no bias, a parent-of-origin bias, or a lineage-of-origin (e.g., allelic) bias.

Figure 2

Numbers of transcripts with parentally biased and lineage-biased gene expression. The final row includes the total numbers of tested transcripts and the number of SNPs included in each test are in parentheses. In all cases, there were significantly more maternally biased than paternally biased transcripts (Storer-Kim tests, P < 0.001) and more European-biased than Africanized-biased transcripts (Storer-Kim tests, P < 0.001). In total, there were 46 transcripts with parent-of-origin effects. Note that some of the transcripts were significant in more than one sampling group (see Table 1).

Figure 3

Parent-of-origin effects on gene expression. A heatmap of the 46 parentally biased transcripts demonstrates that reciprocal crosses have somewhat consistent patterns of parental bias across developmental and behavioral states. Each line represents a cufflinks-predicted transcript with a significant parent-of-origin effect on gene expression. Parental bias for each significant transcript is shown for larvae, adults, and brains. AE: cross between Africanized queen and European-derived drone. EA: the reciprocal cross. Numbers under each cross denote replicates. Blues represent a paternal bias; reds represent maternal biases. *The locus was confirmed with validation datasets. ^§The locus was tested, but the bias was not confirmed. See Table S1 for detailed results.