Intraspecific Transcriptome Variation and Sex-Biased Expression in Anopheles arabiensis

Abstract

The magnitude and functional patterns of intraspecific transcriptional variation in the anophelines, including those of sex-biased genes underlying sex-specific traits relevant for malaria transmission, remain understudied. As a result, how changes in expression levels drive adaptation in these species is poorly understood. We sequenced the female, male, and larval transcriptomes of three populations of Anopheles arabiensis from Burkina Faso. One-third of the genes were differentially expressed between populations, often involving insecticide resistance-related genes in a sample type-specific manner, and with the females showing the largest number of differentially expressed genes. At the genomic level, the X chromosome appears depleted of differentially expressed genes compared with the autosomes, chromosomes harboring inversions do not exhibit evidence for enrichment of such genes, and genes that are top contributors to functional enrichment patterns of population differentiation tend to be clustered in the genome. Further, the magnitude of variation for the sex expression ratio across populations did not substantially differ between male- and female-biased genes, except for some populations in which male-limited expressed genes showed more variation than their female counterparts. In fact, female-biased genes exhibited a larger level of interpopulation variation than male-biased genes, both when assayed in males and females. Beyond uncovering the extensive adaptive potential of transcriptional variation in An. Arabiensis, our findings suggest that the evolutionary rate of changes in expression levels on the X chromosome exceeds that on the autosomes, while pointing to female-biased genes as the most variable component of the An. Arabiensis transcriptome.

Keywords: Anopheles arabiensis; faster-X effect; functional diversification; sex-biased gene expression; transcriptome variation.

Fig. 1.

Collection sites of gravid females An. Arabiensis in Burkina Faso. Nanoro is located in the Sudano-Sahelian zone, which extends throughout much of the central part of the country. In this zone, vegetation is constituted of an arid savannah and the climate is characterized by a short rainy season from July to September. Bobo Dioulasso, or Bobo, and Kodeni are situated in the Sudano-Guinean zone at the Southwest of the country. In this zone, the type of vegetation is humid savannah, and the rainy season extends from June to October.

Fig. 2.

Magnitude and patterns of interpopulation differentiation of the transcriptome of An. Arabiensis across sample types. (A) Number of differentially expressed genes at a 1% FDR and fold-change ≥2 for females, males, and larvae. The total number of unique differentially expressed genes across contrasts for each sample type is indicated in parenthesis. (B) Venn diagrams showing the degree of overlap among genes differentially expressed across the three contrasts performed for each sample type. (C) Hierarchical clustering of differentially expressed genes for different sample types. The genes included correspond to those in the union set of the Venn diagrams in (B). B, Bobo; K, Kodeni; and N, Nanoro. Unlike Nanoro, Bobo, and Kodeni are associated with the same ecotype zone.

Fig. 3.

Commonalities in functional enrichment patterns across different functional rubrics are limited among sample types. (A) Degree of overlap among sets of differentially expressed genes in at least one population contrast at a 1% FDR across the three types of samples assayed for GO terms and pathways. GO term and pathway enrichment was assessed using g:Profiler (Raudvere et al. 2019) and VectorBase (Giraldo-Calderon et al. 2015), respectively. (B) Degree of overlap among sets of differentially expressed genes tagged as top contributors to patterns of functional enrichment for GO terms (without distinguishing among Biological Process, Molecular Function, and Cellular Component) and pathways.

Fig. 4.

Differentially expressed genes are nonrandomly distributed across the An. Arabiensis chromosomal element. (A) Per sample type. (B) Main contributors to functional enrichment patterns. Top, different chromosomes, with L and R denoting the left and right arms of the metacentric chromosomes 2 and 3, respectively. Only genes with FC ≥ 2 in expression at a 1% FDR in at least one of the population contrasts performed are considered. Each of these genes is indicated with a vertical solid line. Significant enrichment (↑) or depletion (↓) at 1% FDR-corrected P-value (*) or at 5% FDR-corrected P-value (•) is shown.

Fig. 5.

Abundant and inconsistent sex-biased gene expression across populations of An. Arabiensis. (A) The violin plots show the proportion and range of fold-change values of differentially expressed genes between males and females in the three surveyed populations. (B) For sex-biased gene expression, female (red) or male (blue) bias might not be found in all three populations (no consistently category). In fact, the percentage of genes showing the same sex bias across populations (consistently category) is only 12.54% (4.32% + 8.22%). (C) Sankey charts showing the lability in the sex-biased expression between populations. The seven delineated categories in table 1 are color-coded. Bobo and Nanoro are more closely related to each other based on the patterns of sex-biased expression than either of them to Kodeni. See supplementary table S12, Supplementary Material online for the precise number of genes associated with different pairwise combinations of categories based on the magnitude and type of sex-biased expression between the two populations involved in each case. F, female-biased; M, male-biased; No DE, no differentially expressed between the sexes, that is, nonsex-biased.

Fig. 6.

Magnitude of variation in sex expression ratio of sex-biased genes. The violin plots show the proportion and range of the variation in sex expression ratio across the three populations of An. Arabiensis investigated. For each population, all sex-biased genes, or subsets of them based on the magnitude of the sex bias, are plotted side by side. The magnitude of variation in sex expression ratio was calculated as the standard deviation (SD) of log2(male/female) expression. Only in 2 out of 12 contrasts is there evidence of significantly higher variation between the sexes; in all cases male-biased genes show significantly higher variation than female-biased genes with an effect size >0.3 and statistical significance (***) at P_adj < 0.001 (supplementary table S15, Supplementary Material online).