Allele specific expression and bacterial community composition in the Drosophila gut

The pervasiveness of gene expression variation and its contribution to phenotypic variation and evolution is well known. This gene expression variation is context dependent, with differences in regulatory architecture often associated with intrinsic and environmental factors, and is modulated by regulatory elements that can act in cis (linked) or in trans (unlinked) relative to the genes they affect. So far, little is known about how this genetic variation affects the evolution of regulatory architecture among closely related tissues during population divergence. To address this question, we analyzed gene expression in the midgut, hindgut, and Malpighian tubule as well as microbiome composition in the two gut tissues in four Drosophila melanogaster strains and their F1 hybrids from two divergent populations: one from the derived, European range and one from the ancestral, African range. In both the transcriptome and microbiome data, we detected extensive tissue- and genetic background-specific effects, including effects of genetic background on overall tissue specificity. Tissue-specific effects were typically stronger than genetic background-specific effects, although the two gut tissues were not more similar to each other than to the Malpighian tubules. An examination of allele specific expression revealed that, while both cis and trans effects were more tissue-specific in genes expressed differentially between populations than genes with conserved expression, trans effects were more tissue-specific than cis effects. Despite there being highly variable regulatory architecture, this observation was robust across tissues and genetic backgrounds, suggesting that the expression of trans variation can be spatially fine-tuned as well as or better than cis variation during population divergence and yielding new insights into cis and trans regulatory evolution. Total RNA was extracted from female 6-day-old Drosophila melanogaster midgut (20 per replicate) and hindgut (60 per replicate) samples for 4 isofemale strains (2 from a derived Swedish population and 2 from an ancestral African population) as well as F1 hybrids between the Swedish and African strains. 3 biological replicates were performed per strain or F1 hybrid (48 samples in total). High throughput RNA sequencing (RNA-seq) and amplicon sequencing were then performed using the same RNA extractions. RNA-seq was performed using Poly-A selection and 150-bp paired reads.