Rapid response of fly populations to gene dosage across development and generations

Abstract

It remains unknown how developmental systems evolve in response to variable genetic and environmental conditions. Here, we have examined the evolvability of the classic bicoid network in Drosophila, which is essential for anterior-posterior patterning in the early embryo. This network can be synthetically perturbed by increasing the dosage of bicoid, which causes a posterior shift of the network's regulatory outputs and a decrease in fitness. To directly monitor network evolution across populations with extra copies of bicoid, we performed genome-wide EMS mutagenesis, followed by experimental evolution. After only 8-15 generations, evolved populations have normalized patterns of gene expression and increased survival. Using a phenomics approach, we find that populations normalized through rapid increases in embryo size driven by maternal changes in metabolism and ovariole development. We extend our results to wild populations of flies, demonstrating strong predictability. Together, our results necessitate a broader view of regulatory network evolution at the systems level. This study highlights the power of synthetic evolution using animal systems, a generalizable platform for the dissection of gene regulation and complex genomes.