The AalNix3&4 isoform is required and sufficient to convert Aedes albopictus females into males

Aedes albopictus is one of the most invasive insect species in the world and an effective vector for many important arboviruses. We reported previously that Ae. albopictus Nix (AalNix) is the male-determining factor of this species. However, whether AalNix alone is sufficient to initiate male development is unknown. Transgenic lines that express each of the three AalNix isoforms from the native promoter were obtained using piggyBac transformation. We verified the stable expression of AalNix isoforms in the transgenic lines and confirm that one isoform, AalNix3&4, is sufficient to convert females into fertile males (pseudo-males) that are indistinguishable from wild-type males. We also established a stable sex-converted female mosquito strain, AalNix3&4-♂4-pseudo-male. The pseudo-male mosquitoes can fly and mate normally with wild-type female, although their mating competitiveness is lower than wild-type. This work further clarifies the role of AalNix in the sex determination pathway and will facilitate the development of Ae. albopictus control strategies that rely on male-only releases such as SIT and sex-ratio distortion.

Introduction Aedes albopictus is one of the most invasive and aggressive insect species in the world [1]. It is a competent vector for several globally-important arboviruses, including dengue (DENV), Zika (ZIKV), yellow fever and chikungunya (CHIKV) viruses [2][3][4][5]. Female mosquitoes require blood meals (hematophagy) from terrestrial vertebrates for both reproduction and nutrition. The process of blood feeding allows pathogens carried by the mosquitoes to be transmitted to vertebrate hosts and this poses a serious threat to human health [6].
In some Dipteran insects, including mosquitoes and non-Drosophilid flies, the presence of products encoded by a dominant male determining factor (M-factor) is the primary initiator of sex determination. Females in these species lack this factor and are designated m/m and males are heterogametic, M/m. The Nix gene was first identified as the M-factor in the yellow Fever mosquito, Aedes aegypti, and the AaeNix ortholog is linked to the 1 st chromosome and is both required and sufficient to initiate male development [7]. Disruption of the Ae. albopictus ortholog, AalNix, can result in the feminization of males [8]. Guy1 and Yob/gYG2 are the candidate M-factors in the malaria vector mosquitoes, Anopheles stephensi and An. gambiae, respectively [9][10][11]. The Mdmd gene is the M-factor of the house fly, Musca domestica, and is indispensable for normal male development [12]. MoY is necessary and sufficient for male development and the M-factor of the fruit fly, Ceratitis capitata [13].
A cascade of genes in the sex-determination pathway relay the information from the primary signal to achieve terminal morphological and physiological differentiation [14]. In a previous study, we determined that the AalNix gene product is the primary signal for the Ae. albopictus sex-determination pathway [8]. However, whether AalNix alone can be used to initiate male determination requires further investigation. We also had shown that the male-specific doublesex gene isoform (dsx M ) is the default splicing pattern in Aedes mosquito, which means that one or more potential dsx splicing enhancers link NIX and DSX [15]. The sex-specific splicing of dsx orthologues in Drosophila species is controlled by transformer (tra) [16][17][18]. Our working model is that accumulation of AalNix, a pre-mRNA splicing factor, is sufficient to regulate the alternative splicing of the unknown splicing enhancer, and cause this splicing factor to lose its ability to regulate dsx splicing, enabling dsx to recover its male-default splicing. However, no ortholog of tra has been found yet in mosquito families, most likely because it evolves rapidly resulting in significant sequence divergence [19,20]. The link of the M factor with the downstream dsx is still largely unexplored in mosquitoes.
Our previous work with AalNix showed that the gene has four exons (1-4) [8]. These exons are spliced alternately to produce four isoforms designated AalNix1 (contains all four exons), AalNix2 (exons 1, 3 and 4), AalNix3 (exon 1, intron 1 and 14 base-pairs (bp) of the 5'-end of exon 2) and AalNix4 (exon 1, intron 1 and 59 bp of the 5'-end of exon 2). AalNix1 has two regions homologous to the Ae. aegypti ortholog, AaeNix, that encode two RNA-binding motifs (RRM). AalNix2 encodes one RRM resulting from an exon 2 skip. AalNix3 and AalNix4 share the same stop codon due to retention of a short intron. Using mosquito transgenesis and expression constructs for each of the known AalNix isoforms, we show that AalNix3&4 alone is required and sufficient to convert females into fertile males with phenotypes similar to wildtype males. We also mated pseudo-males derived from genotypically m/m females carrying only the AalNix3&4 transgene to wild-type female mosquitoes and obtained stable lines. These studies support the further development of AalNix transgenes, as the male-determining factor of Ae. albopictus, for applications to sterile insect technologies (SIT) for mosquito control and provide the basis for the next step in the implementation of Cas9/guide RNA-based gene-drive systems in Ae. albopictus. Theoretically, gene-drive systems should result in the offspring of drive-system males and wild-type female mosquitoes being 50% gene-drive males, and the other 50% intersex or pseudo-males due to the introduced AalNix gene. The intersex do not produce offspring and pseudo-males, as good as the ones of line AalNix3&4-♂4 do produce some offspring including new pseudo-males, while wild-type males inheriting the gene-drive transgene have the same mating ability as wild-type males, and can mate with wild-type females to transmit the gene drive system resulting in population suppression. Our work confirms and extends a recent contribution by Lutrat et al [21].

Generation of transgenic mosquitoes for each AalNix isoform
We made individual transgenes that linked the endogenous AalNix promoter to three distinct abbreviated forms of the coding region to see which of the four exons play roles in establishing male sex determination (Figs 1; S5). The transgenes were constructed to contain exons 1, 2, 3 and part of 4 (AalNix1), exons 1, 3 and part of 4 (AalNix2) or exon 1, intron1 and 59 bp of the 5'-end of exon 2 (AalNix3&4), and were inserted into a piggyBac transformation vector carrying a dominant fluorescent protein marker gene (OpIE2-DsRed). Following microinjection of embryos, matings of surviving injected adults (G 0 ) to wild-type mosquitoes and screening of G 1 progeny for the marker gene, three males and females each were recovered for AalNix1 ( Table 1). The line derived from one of the males, AalNix1-♂2, was characterized further by inverse gene amplification procedures (iPCR) and shown to be inserted into an intergenic region at AaloF1:JXUM01S001974: 263598-263603 (S1A Fig). Three transgenic males and two females were recovered in the AalNix2 G 1 progeny, and AalNix2-♂3, inserted intergenically at Aalbo_primary.1 scaffold_701: 209295-209300, was established for further analyses (S1B Fig). The G 1 progeny arising from the AalNix3&4 transgenesis protocols included a number of mosquitoes with an intersex phenotype arising most likely from apparent conversion of females to males and providing preliminary evidence for the activity of this isoform combination in the sex-determination pathway (Table 1). We selected 21 phenotypically-normal males, six phenotypically-normal females expressing DsRed and four intersex individuals, and two lines, AalNix3&4-♂4 (Aalbo_primary.1 scaffold_64: 2942049-2942054) and AalNix3&4-♂15 (Aalbo_primary.1 scaffold_1: 15815977-15815982), were established for further analyses (Figs 2; S1C). Both of these insertions were intergenic based on the genome annotation and did not disrupt any coding sequence, and the insertion site of AalNix3&4-♂4 is shown in (Figs 2C; S1). Nonetheless, we do not rule out the possibility that inverse PCR cannot detect all transgene insertions in a transgenic line.

AalNix3&4 is sufficient alone to convert females into males
We designed two pairs of oligonucleotide primers and used gene amplification analyses (PCR) to explore the inheritance profiles of the transgenic AalNix isoforms (S1 Table;  The pseudo-males (m/m; AalNix, DsRed + ) display complete and normal-appearing malespecific external and internal reproductive organs including gonocoxites, gonostyli, testes and accessory glands. In addition, they display the plumose antennae characteristic of wild-type males (Figs 2; S3). We selected randomly a G 2 AalNix3&4-♂4 pseudo-male, mated it with  wild-type females and established a stable line, AalNix3&4-♂4-pseudo-male, that has been propagated for >17 generations and remains stable. Remarkably, there is a complete absence of an M-locus in this cross, which was thought previously essential for the sexual development of male mosquitoes [8]. The wing-length measurements of the pseudo-male offspring of this line are comparable in size to wild-type males, supporting the conclusion that the converted males are indistinguishable from their wild-type counterparts unless viewed with fluorescence microscopy, where they display the DsRed phenotype ( Fig 2B; S5 Table). All of the pseudo-males recovered in subsequent generations (G 3 -G 10 ) are DsRed + , while the females are not, and the sex-ratio is~1:1 (p = 0.865) (S6 Table). These data confirm that the AalNix3&4 isoform is sufficient alone to convert genotypic females (m/m) to fully phenotypic males.
The AalNix3&4-♂15 line yielded four phenotypes during its establishment, normal females, intersexes, converted and the wild-type males (S7 Table). The differences in the characteristics of this line when compared to AalNix3&4-♂4 could result from insertion site effects from it being integrated at a different location in the genome. Also, there were multiple piggyBac transgene insertions in these lines, and some of these insertions (at least one) were not masculinizing and produced transgenic females. Furthermore, genotypically-female (m/m), transgenic (Nix, DsRed + ) converted males were rare (~0.5%, 21 of 4102) and the strain went extinct at 10 generations.

Transgene-expressed AalNix alters doublesex (dsx) and fruitless (fru) splicing
Endogenous AalNix expression is first detected in embryos between 4h and 8h after oviposition and thereafter throughout subsequent life stages. Importantly, transcription of AalNix is limited to male mosquitoes and regulates the splicing of the zygotic Aaldsx M isoform [8]. We investigated the impact of exogenously-supplied AalNix on dsx and fru to determine its impact in regulating the sex determination pathway in the transgenic mosquitoes.
AalNix expression in AalNix1-♂2 and AalNix2-♂3 transgenic males was 1.62-fold and 2.40-fold higher than wild-type males, respectively, while as expected, AalNix transcripts are not found in transgenic females (Fig 3). As a downstream gene of sex determination pathway, the female and male isoforms of dsx and fru do not change in the AalNix transgenic individuals, compared to wild-type individuals (Fig 3).
The expression levels of AalNix gene transcripts in AalNix3&4-♂4 (carrying the extra copy of the expressed first exon and 24 bp of intron 1) intersex individuals, pseudo-males and male were 0.27-, 0.9-and 2.48-fold greater, respectively, than wild-type males (Fig 3A-3D). The female-specific dsx F splice variant was the main isoform detected in wild-type females, and the male-specific splicing variant, dsx M was predominant in males. Interestingly, dsx F transcripts are abundant in intersex mosquitoes, which are phenotypically more similar to wild-type females. Furthermore, the transcript accumulation of dsx F in pseudo-males is 0.16-fold lower than wild-type females and is closer to that of wild-type males. The transcript accumulation levels of the dsx M isoform in intersexes is 17% that of wild-type male mosquitoes and 24-fold higher than of wild-type females. However, dsx M accumulates to similar levels in pseudo-males and genotypic males (M/m) carrying the AalNix3&4 transgene and these are lower than in wild-type males. The expression trends of fru F and fru M are similar to dsx F and dsx M (Fig 3A-3D).

Mating competitiveness of transgenic male mosquitoes
Competition assays in cages involved exposing ten wild-type females to mixed groups of pseudo-males, transgenic males expressing both native AalNix and exogenous AalNix3&4 and wild-type males. Following exposure to a 1:1 ratio of five pseudo-males to five wild-type males, the percentage of females that produced broods that contained transgenic progeny was 15% (9 of 51) ( Table 2) and the percentage of transgenic progeny was 9.08% (251 of 2,512) (S8 Table). Although the visible morphological phenotypes of the transgenic males are not distinguishable from those of wild-type males (Fig 2), they contribute significantly fewer progeny than the wild-types to the next generation (p < 0.001; Table 3).
A competition experiment in cages (six replicates) involving exposing ten wild-type females to a 1:1 ratio of wild-type: heterozygous transgenic males (M/m; Nix/+) resulted in 58.3% (35 of 25) ( Table 2) of the females producing broods with transgenic progeny of which 31.89% (603 of 1,288) (p < 0.001) were transgenic (Table 4; S9 Table). This data support the further development of these or similar transgenic strains for application to future population replacement or population suppression genetic control strategies.

Discussion
The results of these experiments demonstrate that the expression of exogenously-derived Aal-Nix gene products can convert genotypically (m/m) Ae. albopictus females into males. Transgenic females carrying and expressing the first AalNix exon can develop into functional adult males capable of mating with wild-type females to produce normal offspring. These data are consistent with a recent report and confirm that the Nix gene alone is sufficient to initiate male development in Ae. albopictus, and lays the foundation for the next step to link the Nix gene with a CRISPR/Cas9 gene-drive system for control of vector-borne diseases.
Aedes albopictus and Ae. aegypti have a sex-determining locus, M, on their 1 st chromosomes. This chromosome behaves generally in all other aspects as an autosome. There are two annotated protein-encoding genes in the Ae. aegypti M-locus, including Nix, and a myo-sex gene that has a role in the flight of adult males [22]. However, only one gene orthologous with AaeNix has been found in the M-locus of Ae. albopictus and it is designated AalNix [23]. Unlike AaeNix, which has only a single isoform, AalNix has four isoforms (AalNix1-4). The coding sequence of AalNix3 and AalNix4 are nearly identical and contain only the first RRM motif found in AaeNix [8]. Intron 1 and its 5'-end donor site location is conserved among Aae-Nix and AalNix (S4 Fig). It is worth noting that AalNix3 and AalNix4 differ from other isoforms in that their CDS includes 24 bases of the AalNix gene intron1. As an RRM motifcontaining protein, NIX is critical for dsx and fru male-specific transcript splicing in the sexdetermination pathway. We constructed three AalNix isoform expression plasmids and were able to establish the transgenic strains of each isoform. Expression of exogenously-provided AalNix1 and AalNix2 did not result in sex conversion in the females in the corresponding transgenic lines, but we did not recover a larger number of lines. In contrast, genetic females (m/m) of AalNix3&4 lines showed conversion to morphologically-normal wild-type (M/m) males. The different AalNix3&4 insertion sites in two distinct transgenic lines produced genetic females (m/m) with different degrees and proportions of male conversion, most likely resulting from an insertion-site effects. In addition, it is known that using piggyBac is likely to produce multiple-copy insertions [24], and inverse PCR mapping in lines most-likely detected only one of the multiple insertions, not necessarily the masculinizing one, results in the differences observed between AalNix3&4-♂4 and AalNix3&4-♂15. As seen with previous work with AalNix gene [8], the AalNix3&4 isoform was found in male pupae, male adults and C6/36 cells, supporting the conclusion that it plays a major role in the development of sex characteristics of Ae. albopictus males. While it cannot be ruled out that the insertion-sites affect AalNix1 and AalNix2 function because both contain exon 1, AalNix3&4 has the extra 25 bp of intron 1 and contains only one RRM site. It is possible that AalNix1 and AalNix2 may have roles assisting AalNix3&4 in the male sex determination pathway. As shown in the mating competition experiment, wild-type males carrying the transgene were as competitive as those without the transgene, but the mating competitiveness of pseudo-males was reduced.
We also isolated a completely gender-converted female from the AalNix3&4-♂4 strain, and following mating it with wild-type females to establish a stable strain, all subsequent male progeny of outcrosses to wild-type females inherited the AalNix3&4 transgene (DsRed + ), while the females did not (DsRed -). Thus, we have inserted an artificial M-locus into the female (m/m) genome and produced fully-functional males through sex conversion. The strain carrying the male sex-linked fluorescent markers are ideally suited for automated fluorescence sex-sorting needed for technologies such as SIT where it is important to obtain large numbers of male mosquitoes. Furthermore, since Nix alone completely converts females to males, combining it with a gene-drive system could produce a strain in which all resulting progeny are males following outcrossing [19,25]. We are working now to build such a genedrive system in Ae. albopictus based on gender conversion and this could help achieve the goal of eliminating breeding female mosquitoes by producing pseudo-male mosquitoes.

Mosquitoes
The Ae. albopictus Foshan strain was a kind gift from the Centers for Disease Control of Guangdong Province. This strain was isolated from Foshan, Guangdong, PRC and has been established in the laboratory since 1981. All mosquitoes are reared in an insect breeding room with constant temperature and humidity (27 ± 1˚C, relative humidity 70%-80%) under a 14 L:10D photoperiod. The larvae were reared in plastic pots and fed ground turtle food (INCH-GOLD, Shenzhen, China). Adult mosquitoes were kept in 20 X 20 X 30 cm nylon mesh cages and allowed access to a cotton wick soaked in a 10% glucose solution and mated freely. Adult females were allowed to take blood meals from defibrinated sheep blood (Solarbio Life Sciences, Beijing, China) using a Hemotek membrane feeding system (Discovery Workshops, Accrington, UK).

Plasmid construction
All oligonucleotide primers are listed in S1 Table. To generate the pAalNix constructs, the pig-gyBac backbone plasmid AAEL010097-Cas9 (a kind gift from Omar Akbari [Addgene plasmid # 100707; http://n2t.net/addgene:100707]) [3] was digested and ligated to a linearized AalNix cassette (containing the CDS of AalNix1, AalNix2, AalNix3) at Not I/Asc I sites using the In-Fusion HD Cloning Kit (Takara, Dalian, China) [26]. To construct the piggyBac-AalNix1 plasmid, a 3,967-bp sequence containing the 2,478-bp AalNix native promoter (S5 Fig) (including the 133-bp 5 0 -end UTR), the 999-bp Nix1 ORF (the 24-bp Strep-tag II fusion to the C-terminus of AalNix1), and a 239 bp 3 0 -end UTR region followed by the bGH poly(A) signal, was amplified from pAalNix1 [15] using the primers Aal01 and Aal02 and cloned into the linearized backbone vector described above. To construct the piggyBac-AalNix2 plasmid, the backbone of the AAEL010097-Cas9 Vector and two DNA fragments were then joined in the following order: 1) AalNix native promoter and exon1 sequence (674bp) was amplified from pAalNix1 using the primer Aal03 and Aal04; 2) the sequence of the remaining 247-bp AalNix2 ORF region (Strep-Tag II fusion to the C-terminus of AalNix2) and 159 bp 3 0 -end UTR region followed by bGH poly(A) signal (this fragment was synthesis in vitro (Invitrogen, Guangzhou, China) using primers Aal05 and Aal06. These two fragments were purified and joined using the In-Fusion HD Cloning Kit (Takara), resulting in the pAalNix2 plasmid. The pAalNix3&4 plasmid was constructed using the same method for the pAalNix2 plasmid: 1) the AalNix native promoter and exon1 sequence were amplified from pAalNix1 using the primer Aal07 and Aal08; 2) the sequence of the remaining 25-bp AalNix4 ORF region (Strep-Tag II fusion to the C-terminus of AalNix4) and 138 bp 3 0 UTR region followed by bGH poly-(A) signal sequence (Invitrogen Gene Synthesis service) using primers Aal09 and Aal10. These two fragments were purified and inserted into pAalNix3&4 plasmid using the In-Fusion HD Cloning Kit (Takara). All amplification steps were performed using Q5 Hot Start High-Fidelity 2X Master Mix (NEB, Ipswich, MA) and the amplicons were analyzed by agarose gel electrophoresis and, then purified using the GeneJET Gel Extraction Kit (Thermo Fisher Scientific, Carlsbad, CA, USA). All pAalNix plasmid sequences were verified using Shenggong Biotech Sanger sequencing services (Shenggong Biotechnology Co. Ltd., Guangzhou, China). Plasmids with the correct sequence were used as donors for construction of transgenic mosquitoes.

Characterization insertion site of the transgenic line
Inverse PCR was performed to characterize the insertion site as described previously [3,22]. Genomic DNA was extracted from adult females and males from each line, using a DNeasy Blood and Tissue kit (Qiagen, Hinden, Germany) according to the supplier's protocol. One microgram of DNA was digested for 8h at 37˚C with Sau3AI (5'-end reaction) or HinP1I (3'end reaction) restriction enzymes (NEB), and then purified using the GeneJET PCR Purification Kit (Thermo Fisher Scientific) in accordance with the manufacturer's protocol. Approximately 200 ng DNA fragments were circularized with T4 DNA ligase (NEB) at 16˚C overnight in a 200 uL reaction volume. Finally, the ligation product was purified and subjected to two rounds of PCR reaction using Q5 Hot Start High-Fidelity 2X Master Mix (NEB) with primers AE049-AE056 [3]. Primers Aal049/ Aal050 and primers Aal053/ Aal054 were used for the first round of PCR for 5' -end reaction and 3'-end reaction respectively. Primers Aal051/ Aal052 and primers Aal055/ Aal056 were used for nested PCR for 5'-end reaction and 3'-end reaction, respectively. The nested PCR products were directly loaded on agarose gels, and were extracted using the GeneJET Gel Extraction Kit (Thermo Fisher Scientific), ligated using the CloneJET PCR Cloning Kit (Thermo Fisher Scientific), and transformed into DH5α competent cells (Takara). Positive clones were purified (EZNA Plasmid Mini Kit I, Omega) and sent for sequencing (ShengGong Biotechnology Co. Ltd., Shanghai, China).

Identification of endogenous and transgenic AalNix
To ensure that all transgene cassettes were stably inherited to the next generation, male phenotype individuals with DsRed fluorescence were selected and a part of the leg cut off to extract DNA, using the Extract-N-Amp Tissue PCR Kit (Sigma-Aldrich Inc., St. Louis, MO) according to the manufacturer's protocol. PCR was performed to characterize the native AalNix and the exogenous transgenic AalNix with the primers Nix_i2-F/ Nix_i2-R and the primers Nix_E1F/ strep tag-R, using DreamTaq Hot Start Green PCR Master Mix (2X) (Thermo Fisher Scientific, Waltham, MA, USA) according to the user manual. Males containing the endogenous AalNix and exogenous AalNix genes were mated with wild-type females to established lines.

Wing-length measurement and statistical analysis
Wing-length values were measured for 30 individuals from wild-type females and males and AalNix3&4-♂4-pseudo-males. The wings of one mosquito were removed and placed under a laboratory microscope (connected to a computer and a camera) to take photographs. Imagepro Plus software was used to measure the wing length of adult mosquitoes. The wing length as determined is the distance from the axillary incision to the apical margin (excluding the wing-margin fringe hairs). The average length of a pair of wings from each mosquito was recorded as the wing length.

RT-qPCR analysis of transgenic mosquitos
The relative expression levels of AalNix, Aalbdsx and Aalfru in transgenic mosquitos were quantified by real-time PCR using the SuperReal PreMix Plus kit (SYBR Green) (Tiangen Biotech Co., Ltd., Beijing, China) according to the manufacturer's instructions. Total RNA was extracted from whole adult transgenic mosquitoes from each phenotype of each line, using TRIzol (Life Technologies, Carlsbad CA, USA) according to the manufacturer's directions. Each sample contains nine adult mosquitos assessed in triplicate and normalized with AalRpS7 mRNA. The qPCR results were analyzed using the 2 −ΔΔCT method.

Mating competition assays
To assess mating competitiveness between wild type and transgenic males (m/m♂; nix/+), five WT males and five transgenic males were placed into a 1.5-L container and allowed to adapt to the environment. A total 10 virgins females previously fed with a blood meal were put into the container to mate. After two day of full mating, the female mosquitoes in each container were placed in a single oviposition cup to lay eggs and these were allowed to hatch. When the hatched larvae reached the third-fourth instar, the were screened with fluorescence microscopy (Nikon SMZ1000) and counted. Six replicate cages were set up and assayed for each line as described.

Statistical analysis
All data were represented as means ± standard deviation (SD) from at least three independent experiments. Analysis of differences between groups were analyzed by unpaired Student's ttest. One-Way ANOVA of variance was used to evaluate spatiotemporal expression profiles of AalNix. Statistical significance was set at P < 0.05 and was determined using GraphPad Prism Version 8.02 program (GraphPad Software, CA, USA) and IBM SPSS Statistics 20.0 (SPSS Software, IL, USA). Gene expression was normalized using the deltaCT method against AalRpS7 as a reference gene.