Rapid Evolution of Ovarian-Biased Genes in the Yellow Fever Mosquito (Aedes aegypti)

Abstract

Males and females exhibit highly dimorphic phenotypes, particularly in their gonads, which is believed to be driven largely by differential gene expression. Typically, the protein sequences of genes upregulated in males, or male-biased genes, evolve rapidly as compared to female-biased and unbiased genes. To date, the specific study of gonad-biased genes remains uncommon in metazoans. Here, we identified and studied a total of 2927, 2013, and 4449 coding sequences (CDS) with ovary-biased, testis-biased, and unbiased expression, respectively, in the yellow fever mosquito Aedes aegypti The results showed that ovary-biased and unbiased CDS had higher nonsynonymous to synonymous substitution rates (dN/dS) and lower optimal codon usage (those codons that promote efficient translation) than testis-biased genes. Further, we observed higher dN/dS in ovary-biased genes than in testis-biased genes, even for genes coexpressed in nonsexual (embryo) tissues. Ovary-specific genes evolved exceptionally fast, as compared to testis- or embryo-specific genes, and exhibited higher frequency of positive selection. Genes with ovary expression were preferentially involved in olfactory binding and reception. We hypothesize that at least two potential mechanisms could explain rapid evolution of ovary-biased genes in this mosquito: (1) the evolutionary rate of ovary-biased genes may be accelerated by sexual selection (including female-female competition or male-mate choice) affecting olfactory genes during female swarming by males, and/or by adaptive evolution of olfactory signaling within the female reproductive system (e.g., sperm-ovary signaling); and/or (2) testis-biased genes may exhibit decelerated evolutionary rates due to the formation of mating plugs in the female after copulation, which limits male-male sperm competition.

Keywords: Aedes aegypti; Genetics of Sex; gonads; molecular evolution; ovary-biased genes; protein sequence divergence.

Figure 1

The average GC content at synonymous third codon positions (GC3s) and the GC content of introns (GCi). Results are shown for the set of coding sequences (CDS) with high (highest 5%), intermediate (between the 5th and 95th percentile), or low (lowest 5%) expression in A. aegypti embryos. GCi for short (≤130 bp) and long (>130 bp) introns (after removal of 50 bp at each end) per expression class are shown. Different letters (a, b, or c) above the bars showing mean values of GC3s, GCi (short) and GCi (long) within each expression class indicate a statistically significant difference (P < 0.05) using paired Mann-Whitney U (MWU) tests (the P-value was <0.001 in each contrast). Note that the GC3s values were also statistically significant different between the three expression classes using paired contrasts (MWU-test P < 0.001). Error bars are standard errors.

Figure 2

The average expression level in testes and ovaries of CDS with sex-biased expression in A. aegypti. The testis-biased and ovary-biased CDS were each subdivided with respect to fold bias, and the expression level in each tissue type for the biased genes is shown. (A and B) testis-biased short and long CDS. (C and D) ovary-biased short and long CDS. Within (A and B), all paired contrasts between fold class categories for testis-expressed and for ovary-expressed CDS were statistically significant (P < 0.05) with Mann-Whitney U tests P < 0.001 in each contrast. For (C), all paired contrasts between fold bias class were statistically significant for ovary expression (P < 0.001 for each contrast), but not for testis expression (all P > 0.86). For (D), all paired contrasts between fold classes were statistically significant for ovary expression, and between the ≥ 2 to 5 vs. the ≥ 5 to 10 and ≥ 10 classes for testis expression (P was < 0.001). Note the y-axes are in log scale. CDS, coding sequences; FPKM, frequency per kilobase million.

Figure 3

Box and whisker plots of dN/dS (A) and Fop (B) showing the distribution of all testis-biased, ovary-biased, and unbiased CDS in Aedes. Different letters above any two bars for the testis-biased, ovary-biased, and unbiased genes within each figure indicate a statistically significant (P < 0.05) difference in values using Mann-Whitney U tests (P was < 0.001 in all contrasts). N_{Testis-Biased} = 2927; N_Ovary-Biased = 2013, and N_Unbiased = 4449. Fop was measured using A. aegypti, and dN/dS measured using A. albopictus as a reference. CDS, coding sequences; dN/dS, nonsynonymous to synonymous substitution rates; Fop, frequency of optimal codons.

Figure 4

(A) The number of genes expressed in testes, ovaries, and embryos in A. aegypti among the 9389 genes under study (328 CDS had no expression). (B). The mean dN/dS of genes for tissue-specific (single tissue) and genes coexpressed (two tissue types or all three), and those not expressed in any tissues (none). Standard errors are shown in error bars. Different letters above any two bars indicate a statistically significant difference (P < 0.05) between those dN/dS values, using a paired Mann-Whitney U (MWU)-test (note that P was < 0.01 for all paired contrasts). *The ovary-specific and testis-ovary dN/dS did not differ across all genes per group (MWU-test P > 0.05), but the former had higher values using contrasts of values above the median (P < 0.001). No difference was observed between CDS coexpressed in ovary–embryo vs. testis–ovary as denoted by the same letter (d) in both categories. **No expression (No exp.) exhibited no difference in dN/dS as compared to testis-specific CDS, but was different from embryo-specific CDS. CDS, coding sequences; dN/dS, nonsynonymous to synonymous substitution rates.