Horizontal gene transfer events have played a major role in the evolution of microbial species, but their importance in animals is less clear. Here, we report horizontal gene transfer of cytolethal distending toxin B (cdtB), prokaryotic genes encoding eukaryote-targeting DNase I toxins, into the genomes of vinegar flies (Diptera: Drosophilidae) and aphids (Hemiptera: Aphididae). We found insect-encoded cdtB genes are most closely related to orthologs from bacteriophage that infect Candidatus Hamiltonella defensa, a bacterial mutualistic symbiont of aphids that confers resistance to parasitoid wasps. In drosophilids, cdtB orthologs are highly expressed during the parasitoid-prone larval stage and encode a protein with ancestral DNase activity. We show that cdtB has been domesticated by diverse insects and hypothesize that it functions in defense against their natural enemies.

Environments often vary across a life cycle, imposing fluctuating natural selection across development. Such fluctuating selection can favor different phenotypes in different life stages, but stage-specific evolutionary responses will depend on genetic variance, covariance, and their interaction across development and across environments. Thus, quantifying how genetic architecture varies with plastic responses to the environment and across development is vital to predict whether stage-specific adaptation will occur in nature. Additionally, the interaction of genetic variation and environmental plasticity (GxE) may be stage-specific, leading to a three-way interaction between genotype, environment, and development or GxDxE. To test for these patterns, we exposed larvae and adults of Drosophila melanogaster isogenic lines derived from a natural population to extreme heat and cold stress after developmental acclimation to cool (18°C) and warm (25°C) conditions and measured genetic variance for thermal hardiness. We detected significant GxE that was specific to larvae and adults for cold and heat hardiness (GxDxE), but no significant genetic correlation across development for either trait at either acclimation temperature. However, cross-development phenotypic correlations for acclimation responses suggest that plasticity itself may be developmentally constrained, though rigorously testing this hypothesis requires more experimentation. These results illustrate the potential for stage-specific adaptation within a complex life cycle and demonstrate the importance of measuring traits at appropriate developmental stages and environmental conditions when predicting evolutionary responses to changing climates.

Organisms with complex life cycles demonstrate a remarkable ability to change their phenotypes across development, presumably as an evolutionary adaptation to developmentally variable environments. Developmental variation in environmentally sensitive performance, and thermal sensitivity in particular, has been well documented in holometabolous insects. For example, thermal performance in adults and juvenile stages exhibit little genetic correlation (genetic decoupling) and can evolve independently, resulting in divergent thermal responses. Yet, we understand very little about how this genetic decoupling occurs. We tested the hypothesis that genetic decoupling of thermal physiology is driven by fundamental differences in physiology between life stages, despite a potentially conserved Cellular Stress Response. We used RNAseq to compare transcript expression in response to a cold stressor in Drosophila melanogaster larvae and adults and used RNAi (RNA interference) to test whether knocking down nine target genes differentially affected larval and adult cold tolerance. Transcriptomic responses of whole larvae and adults during and following exposure to -5°C were largely unique both in identity of responding transcripts and in temporal dynamics. Further, we analyzed the tissue-specificity of differentially-expressed transcripts from FlyAtlas 2 data, and concluded that stage-specific differences in transcription were not simply driven by differences in tissue composition. In addition, RNAi of target genes resulted in largely stage-specific and sometimes sex-specific effects on cold tolerance. The combined evidence suggests that thermal physiology is largely stage-specific at the level of gene expression, and thus natural selection may be acting on different loci during the independent thermal adaptation of different life stages.

Organisms with complex life cycles demonstrate a remarkable ability to change their phenotypes across development, presumably as an evolutionary adaptation to developmentally variable environments. Developmental variation in environmentally sensitive performance, and thermal sensitivity in particular, has been well documented in holometabolous insects. For example, thermal performance in adults and juvenile stages exhibit little genetic correlation (genetic decoupling) and can evolve independently, resulting in divergent thermal responses. Yet, we understand very little about how this genetic decoupling occurs. We tested the hypothesis that genetic decoupling of thermal physiology is driven by fundamental differences in physiology between life stages, despite a potentially conserved Cellular Stress Response. We used RNAseq to compare transcript expression in response to a cold stressor in Drosophila melanogaster larvae and adults and used RNAi (RNA interference) to test whether knocking down nine target genes differentially affected larval and adult cold tolerance. Transcriptomic responses of whole larvae and adults during and following exposure to -5C were largely unique both in identity of responding transcripts and in temporal dynamics. Further, we found no relationship between stage-specificity and tissue-specificity of transcripts, suggesting that the differences are not simply driven by differences in tissue composition across development. In addition, RNAi of target genes resulted in largely stage-specific and sometimes sex-specific effects on cold tolerance. The combined evidence suggests that thermal physiology is largely stage-specific at the level of gene expression, and thus natural selection may be acting on different loci during the independent thermal adaptation of different life stages.