On the Origin and Evolution of the Mosquito Male-determining Factor Nix

Abstract

The mosquito family Culicidae is divided into 2 subfamilies named the Culicinae and Anophelinae. Nix, the dominant male-determining factor, has only been found in the culicines Aedes aegypti and Aedes albopictus, 2 important arboviral vectors that belong to the subgenus Stegomyia. Here we performed sex-specific whole-genome sequencing and RNAseq of divergent mosquito species and explored additional male-inclusive datasets to investigate the distribution of Nix. Except for the Culex genus, Nix homologs were found in all species surveyed from the Culicinae subfamily, including 12 additional species from 3 highly divergent tribes comprising 4 genera, suggesting Nix originated at least 133 to 165 million years ago (MYA). Heterologous expression of 1 of 3 divergent Nix open reading frames (ORFs) in Ae. aegypti resulted in partial masculinization of genetic females as evidenced by morphology and doublesex splicing. Phylogenetic analysis suggests Nix is related to femaleless (fle), a recently described intermediate sex-determining factor found exclusively in anopheline mosquitoes. Nix from all species has a conserved structure, including 3 RNA-recognition motifs (RRMs), as does fle. However, Nix has evolved at a much faster rate than fle. The RRM3 of both Nix and fle are distantly related to the single RRM of a widely distributed and conserved splicing factor transformer-2 (tra2). The RRM3-based phylogenetic analysis suggests this domain in Nix and fle may have evolved from tra2 or a tra2-related gene in a common ancestor of mosquitoes. Our results provide insights into the evolution of sex determination in mosquitoes and will inform broad applications of mosquito-control strategies based on manipulating sex ratios toward nonbiting males.

Keywords: genetic engineering; homomorphic sex chromosome; male-determining factor; mosquito control; sex determination.

Fig. 1.

Background and workflow. a) Schematic showing known signals of the sex-determination pathway in mosquitoes. In the subfamily Culicinae, the primary signal is characterized only in 2 aedine species within the subgenus Stegomyia and it is a predicted RNA-binding protein named Nix. The presence of the primary signal Nix determines the male sex outcome, possibly through an unknown intermediate signal(s) as indicated by a question mark (Hall et al. 2015; Aryan et al. 2020). In the subfamily Anophelinae, the primary signals are characterized only in 2 anopheline species within subgenus Celia. They are Guy1 in An. stephensi and Yob in An. gambiae. Both are 56 amino acid residues in length. However, there is no significant sequence similarity between them (reviewed in Kojin, Compton et al. 2022). In An. gambiae, Yob somehow inactivates Fle, a presumed intermediate signal required for female development, resulting in the male sex (Krzyzinska et al. 2021). fle is present in all anopheline mosquitoes analyzed including An. stephensi (Krzyzinska et al. 2021). Nix and Fle are related and both are predicted RBPs with 3 RNA-recognition motifs (RRMs). Evolutionary context is provided based on the known phylogeny of a number of representative mosquito species. Times of origin of the Culicinae and Anophelinae are according to Zadra et al. (2021). Tree is not drawn to scale. Other time estimates exist and are mentioned in text. Not all genera within the 2 subfamilies are included. Anopheles is the major but not the only genus in subfamily Anophelinae. Ps. columbiae, Psorophora columbiae; T. ambionensis, Toxorhynchites amboinensis; Cx. quinquefasciatus, Culex quinquefasciatus. Protein and RNA illustrations are modified from Biorender Icons. b) Workflow of the study. WGS or RNAseq data from thirteen species in the subfamily Culicinae were used to identify Nix (Table 1). Details on -omics data acquisition, Nix discovery, and annotation are described in supplementary fig. S1, Supplementary Material online. ORF, open reading frame.

Fig. 2.

Multiple sequence alignment of Nix and fle sequences and AlphaFold structure predictions of selected Nix and fle proteins. a) Schematic shows structure of Nix compared to fle. b) Multiple sequence alignment of Nix and fle. Only RRM1, RRM2, and RRM3 are shown. Full alignment is provided as supplementary fig. S3, Supplementary Material online. Boxes surround conserved motifs RNP2 and RNP1. Arrows indicate conserved intron positions with black-boxed residues representing the codon that is split by the intron. Ae.alb.d.205aa is a degenerate copy of Nix, and Ae.det.275aa, Wy.smi.319aa, and To.sp.287aa are from RNAseq data only, therefore intron determination was not possible for these sequences. See Table 1 for full species’ names.

Fig. 3.

a to d) AlphaFold2 structure predictions (images from NCBI iCn3D viewer) are rainbow-colored (N-terminal red, C-terminal purple) with RRM1 (orange/yellow), RRM2 (green/turquoise) and RRM3 (blue/purple) located in center, left, and right, respectively. Some N-terminal and C-terminal disordered sequence is not shown in d.

Fig. 4.

Culicinae Nix forms a monophyletic but highly divergent clade when rooted using fle as the outgroup. Phylogenetic relation of Nix and femaleless (fle) was inferred using a) Maximum Likelihood, b) MrBayes, and c) BOINJ. Clade credibility values are indicated. The scale bar shows substitutions per site. See supplementary fig. S3 and supplementary data S1C, Supplementary Material online for alignment used to infer phylogeny. fle, exclusive to Anopheline mosquitoes, was used to root the tree. Nix has not been found in Culex quinquefasciatus, a species with extensive genomic data. Ae.alb.d represents a degenerate copy of Nix. A fourth but highly truncated Ae. triseriatus copy (Ae.tri.4, see supplementary data S1, Supplementary Material online) was not included in this analysis. See Table 1 for full species’ names.

Fig. 5.

Nix is widely distributed and restricted to culicinae mosquitoes. Cladogram shows relationships between mosquito species based on known mosquito phylogeny (Reidenbach et al. 2009; Soghigian et al. 2017; da Silva et al. 2020; Zadra et al. 2021). Genus names are given at top and the Culicinae and Anophelinae subfamilies are shaded. Species for which Nix is present are indicated at the top of the figure and species for which male-determining function of Nix has been demonstrated are in red and bold.

Fig. 6.

Heterologous expression of transgenic Ae. vexans Nix in Ae.aegypti masculinizes genetic females by altering the sex-determining pathway. a) Images show each of 4 genotype progeny resulting from a cross between transgenic Ae. aegypti males having an Ae. vexans Nix expression construct, and wild-type females. Image at left was captured with white light. Arrows point to typical male morphological features, plumose antennae (upper red arrow), and gonocoxites/gonostyli (lower blue arrow), that are also present in the transgenic intersex individual. Image at right was captured with fluorescent microscope using an EGFP filter, showing transgenic individuals evidenced by EGFP expression in whole body. Left panel and right panel show same individuals in same order. b) Expression of Ae. aegypti endogenous Nix (left panel) and transgenic Ae. vexans Nix (right panel) in Ae. aegypti line Ae.vex.p11 was determined by ddPCR relative to gene AAEL002401 used as a control. Genotype, presence of transgene, and phenotype are indicated at bottom. The expression of the transgene in genetic females and males was not significantly different than the endogenous Nix expression in wild-type (WT) males (Tukey's multiple comparison test). c) Expression of female and male dsx isoforms (dsx^F and dsx^M) was determined by ddPCR relative to an endogenous gene AAEL002401 used as a control. Adult progeny with 4 resulting genotypes from a cross of transgenic males and wild-type females were assayed. X-axis labels: genotypic sex (male, M/m; female; m/m); ± indicates presence/absence of the Ae. vexans Nix transgenic cassette determined by a fluorescent marker; symbols indicate phenotypic sex (male, female, and intersex). In (b) and (c) individual values are shown with the mean and ± standard error of the mean (SEM). Statistically significant differences between transgenic genetic females and WT males/WT females are indicated from 1-way Analysis of Variance (ANOVA) followed by Tukey's multiple comparison test.

Fig. 7.

Transcription of transgenic Ae. polynesiensis and Ae. japonicus Nix compared to native Nix, in Ae. aegypti transgenic lines. a) Expression of Ae. aegypti endogenous Nix (WT) and transgenic Ae. polynesiensis Nix was determined by ddPCR in the Ae. aegypti transgenic line Ae.pol.p2, relative to gene AAEL002401 used as a control. b) Expression of Ae. aegypti endogenous Nix (WT) and transgenic Ae. japonicus Nix was determined by ddPCR in the Ae. aegypti transgenic line Ae.jap.p2, relative to gene AAEL002401 used as a control. In Ae. polynesiensis and Ae. japonicus Nix-expressing transgenic lines, the expression of the transgene in genetic females and males was significantly lower than the endogenous Nix expression in WT males according to 1-way ANOVA followed by the Tukey's multiple comparison test (Ae. polynesiensis, P = 0.0016; Ae. japonicus, P < 0.0001). Individual values are shown with the mean and ± SEM.

Fig. 8.

Nix is widespread in culicine mosquitoes and forms a sister clade to fle which is exclusive to anopheline mosquitoes. Shown are the unrooted phylogenies of Nix, femaleless (fle), and transformer 2 (Tra2) using amino acid sequences from RRM3 from Nix and fle, and the single RRM from Tra2. All sequences are from mosquitoes except the sequences in the “nonmosquito Tra2” clade/group. Clade credibility values are shown for major clades. The labels for clade credibility values for the branch separating Tra2 and Nix/fle clades are larger and bold for emphasis. Scale bar shows substitutions per site. See supplementary fig. S7 and data S1C, Supplementary Material online for the RRM 3 alignment used for phylogenetic inference and supplementary fig. S8, Supplementary Material online for phylogenies showing all clade credibility values. Ae.tri._3 is not included in these trees because it does not have an RRM3. a) Maximum likelihood tree generated by RAxML using 1,000 bootstrap replicates. b) Tree generated by MrBayes with 1 million generations. Nix sequences from 3 species (Ps. col, To. amb., To. sp.) were not grouped in the Nix clade but are in the same bubble as all other Nix sequences. c) Distance tree generated by BIONJ using 1,000 bootstrap replicates. See Table 1 for full species’ names. Nonmosquito species are Bombyx mori, Apis mellifera, Drosophila melanogaster, Drosophila virilis, Musca domestica, Lucilia cuprina, Ceratitis capitata, Bactrocera oleae, Anastrepha fraterculus, and Triboleum castaneum.