Mechanisms of transcriptional regulation in Anopheles gambiae revealed by allele-specific expression

Abstract

Malaria control relies on insecticides targeting the mosquito vector, but this is increasingly compromised by insecticide resistance, which can be achieved by elevated expression of detoxifying enzymes that metabolize the insecticide. In diploid organisms, gene expression is regulated both in cis, by regulatory sequences on the same chromosome, and by trans acting factors, affecting both alleles equally. Differing levels of transcription can be caused by mutations in cis-regulatory modules (CRM), but few of these have been identified in mosquitoes. We crossed bendiocarb-resistant and susceptible Anopheles gambiae strains to identify cis-regulated genes that might be responsible for the resistant phenotype using RNAseq, and CRM sequences controlling gene expression in insecticide resistance relevant tissues were predicted using machine learning. We found 115 genes showing allele-specific expression (ASE) in hybrids of insecticide susceptible and resistant strains, suggesting cis-regulation is an important mechanism of gene expression regulation in A. gambiae. The genes showing ASE included a higher proportion of Anopheles-specific genes on average younger than genes with balanced allelic expression.

Keywords: allele; cis-regulation; insecticide; resistance; transcript.

Figure 1.

Crossing, DNA and RNA extraction schema. (a) Thirteen Kisumu females (blue) were crossed to 13 Nagongera males (red). Females mate only once. Genomic DNA from Nagongera males was extracted and sequenced. (b) In the reciprocal cross, 13 Nagongera females (red) were crossed to 13 Kisumu males (blue). Genomic DNA was extracted and sequenced from Kisumu males. (c) Individual mated Kisumu females were transferred to laying cups. Genomic DNA was extracted and sequenced following egg laying. (d) Individual mated Nagongera females were transferred to laying cups. Genomic DNA was extracted and sequenced following egg laying. (e) F1 progeny (purple) from the three Kisumu mothers sequenced at step (c) were raised to adulthood. Genomic DNA was extracted and sequenced from individual males (f). F1 progeny (purple) from the three Nagongera mothers sequenced at step (d) were raised to adulthood. Genomic DNA was extracted and sequenced from individual males (g,h). Female F1 from each of the six mothers were raised to adulthood. RNA was extracted and sequenced from pools of 10 F1 females 3–5 days after eclosion.

Figure 2.

Reciprocal crosses show similar overall gene expression. (a) Volcano plot of log₂ fold change against −log₁₀ p-value comparing gene expression in the F1 progeny of reciprocal crosses between Kisumu and Nagongera strains. Blue points: genes downregulated in progeny of Kisumu mothers; orange points: genes upregulated in progeny of Kisumu mothers. (b) Volcano plot of log₂ fold change against −log₁₀ p-value comparing gene expression between F1 progeny of Nagongera and Kisumu with the Kisumu parental strain. Blue points: genes downregulated in the Kisumu compared with cross progeny; orange points: genes upregulated in Kisumu compared with cross progeny.

Figure 3.

ASE in progeny of crosses between strains. Plots of total read count against ASE at each SNP for SNPs with at least 100 reads. Cross and source of SNPs used to count reads are indicated at the top of each plot. For crosses B5, K2 and K6 the SNP ASE indicates maternal reads/total reads at SNP, and for crosses B1 and B3 SNP ASE indicates reference reads/total reads at SNP. Cross K6 is shown with ASE inferred using SNPs from both the parents of cross K4 and for the initially assumed parents of cross K6.

Figure 4.

Intersection of genes showing ASE between crosses. UpSet plot with the number of genes showing ASE for each cross and the intersection of these genes between crosses. Only the first 28 sets of overlaps are shown.

Figure 5.

Ages of genes showing ASE or not showing ASE. Bar plots comparing the age of genes showing ASE (black bars) and without ASE (red bars) in the progeny of six crosses between Nagongera and Kisumu strains, using the Wagner parsimony method. The cross name is indicated at the top left of each plot, Fisher’s exact test p-values for the difference in fraction of genes in each age between ASE/non-ASE genes are displayed above the bars. *0.001 < p < 0.05; ***p ≤ 0.001.