Gene Regulatory Variation in Drosophila melanogaster Renal Tissue

Abstract

Genetic variation influencing levels of gene expression is abundant in natural populations, and may exert its effects through complex mechanisms that depend on an organism's genetic background and the tissue in which expression is measured. We investigated natural variation in gene expression in the Malpighian tubules of three inbred Drosophila melanogaster strains and their F₁ hybrids. One of the strains was from a population in the species' ancestral range (Zambia), while the other two were from a more recently derived population (Sweden). Although closely related, the two Swedish strains differed greatly in terms of their expression inheritance when hybridized with the Zambian strain, with one Swedish strain showing a large excess of genes with recessive expression inheritance, as well as a large number of genes with overdominant inheritance. Although most expression variation could be attributed to trans-regulation, there were ∼200 genes that showed allele-specific expression differences in each of the between-population hybrids, indicating that cis-regulation contributes as well. The cis-regulated genes were enriched with cytochrome P450 genes, and the upstream regions of six of these genes were incorporated into transgenic reporter gene constructs to test their effects on expression. Differential expression was observed for five of the six reporter genes in the Malpighian tubule, suggesting that a large proportion of cis-regulatory variation lies directly upstream of the affected gene. In most cases, the differential expression was specific to the Malpighian tubule or greater in this tissue than in the rest of the body, highlighting the importance of single-tissue studies of gene expression variation.

Keywords: cis-regulation; cytochrome P450; gene expression; trans-regulation.

Figure 1

Gene expression divergence among genotypes. The numbers above the diagonal indicate the number of DE genes for each pairwise comparison (FDR = 5%). The numbers below the diagonal indicate the expression divergence as determined by subtracting the Spearman correlation coefficient (ρ) over all genes from one. DE, differential expression; FDR, false discovery rate.

Figure 2

The structure of expression variation among genotypes. (A) PC analysis of gene expression counts for all six genotypes. The biological replicates of each genotype are shown in the same color. (B) Numbers of genes that show a significant expression difference (FDR = 5%) between both of the parental strains given on the top line of the x-axis and the strain shown underneath. DE, differential expression; FDR, false discovery rate; PC, principle component.

Figure 3

Genetic differentiation in upstream regions and the effect of upstream variation on expression divergence. (A) Schematic of the reporter gene constructs. The upstream regions of six cytochrome P450 genes from the S58 and Z418 strains were cloned in front of the lacZ gene and integrated into the D. melanogaster genome. Lines indicate the polymorphic sites between S58 and Z418, with different colors indicating which strain has the ancestral variant (red = Z418, blue = S58, and gray = unknown). The total number of polymorphisms in each category is shown in the same color scheme to the right of the upstream region. (B) Relative expression of each endogenous gene in Malpighian tubules (MTs) of the S58 and Z418 parental strains (blue) and their alleles in the S58×Z418 hybrid (red) as determined by RNA-sequencing. Also shown is the relative expression of the respective reporter genes in MTs (orange), whole flies (white), and carcasses with the MTs removed (gray) as determined by β-galactosidase activity. Error bars represent SEM. For reporter genes, the significance of expression differences between the two upstream regions was assessed using a Student’s t-test. (C) Expression divergence in the MT between the Z418 and S58 strains partitioned according to type of regulatory variation. The contributions of cis variation within the tested upstream region (red), cis variation in other regions (gray), and trans variation (blue) were estimated from the RNA-sequencing and reporter gene data. For genes showing a nonadditive cis × trans interaction (Cyp12a5 and Cyp6a20), the total cis variance was assumed to be the maximum possible, i.e., the observed expression divergence between the parental strains. (D) The proportion of highly differentiated SNPs within the tested upstream regions relative to that of their respective chromosome arms. Values greater than one indicate an enrichment of SNPs with F_ST falling within the top 10% (red) or top 5% (blue) of the chromosomal F_ST distribution. Significance was assessed using a χ²-test. * P < 0.05, ** P < 0.01, and *** P < 0.005.