METHOD DETAILS

All similarity searches were conducted using BLAST,⁶⁴ with the word size = 2. The search for the A. gambiae dsx splicing factors within the A. gambiae PEST strain genome was conducted using a BLASTp and tBLASTn algorithm, with the expect value 1e-7, and the D. melanogaster Tra2 (FlyBase IDs CG10128-PA) or Aedes aegypti Nix (a distant homolog of D. melanogaster tra2; GenBank accession {"type":"entrez-protein","attrs":{"text":"AHW46195.1","term_id":"607098468","term_text":"AHW46195.1"}}AHW46195.1) protein sequences used as query. The phylogenetic distribution of the fle orthologs was evaluated through a tBLASTn search against the NCBI whole-genome shotgun contigs (wgs) database using as query the amino acid translation of fle. The search for a putative fle paralog was conducted using a local BLASTp search against a database of the A. gambiae AgamP4.12 peptides downloaded from Vectorbase, and for other Anopheles species by implementing BLASTp search against a protein database for a respective species in Vectorbase. Structure of the proteins was derived by searches of the Conserved Domain Database⁷⁴ and through structure prediction using Jpred-4.⁶⁵ Sequence alignments were conducted using ClustalX2.⁶⁶ Maximum likelihood phylogenetic analysis was conducted in MEGA7⁶⁷ using JTT matrix-based model⁷⁵ with Gamma distributed evolutionary rate differences among sites. The sequence logos were generated using WebLogo.⁶⁸

Open reading frame fragments of the A. gambiae or A. stephensi genes were amplified through PCR from a genomic DNA template or in a one-step RT-PCR (Invitrogen) reaction from the pupal total RNA templates using gene-specific primers flanked at their 5′ ends with the T7 promoter sequence (for details on primers see Table S2). The resulting products, cloned into pGEM T-easy vector (Promega), were used directly, or after reamplification with the same primer pairs, as a template to synthesize double-stranded RNA (dsRNA) using the MEGAscript RNAi T7 kit (Life Technologies) according to manufacturer’s recommendation. Similarly, a fragment of β-lactamase (bla) gene was amplified by PCR from the pGEM-T Easy vector using bla-specific primers, each containing the T7 promoter sequence at the 5' end, and dsRNA was synthesized as described above.

Prior to transfection, the A. gambiae Sua5.1 cells were split into new culture flasks and transfection experiments were performed when cells’ confluency reached 60%–80%. Approximately 1 × 10⁶ cells per well were seeded onto 24 well plates and transfected in suspension, using 3 μl of Lipofectamine 2000 transfection reagent (Life Technologies) and 1.5 μg of dsRNA of a tested gene per well. In parallel, cells in a separate set of wells were transfected with a plasmid (0.3 μg per well) containing the eGFP open reading frame under the control of A. gambiae polyubiquitin promoter as a control of transfection efficiency. In addition, non-transfected control cells were cultured in a set of wells in each experiment. After approximately 24 hours, the transfection efficiency was evaluated using fluorescence microscopy. If at least 30% of the plasmid control cells per well were GFP-positive on a given plate, experimental and non-transfected control cells from that plate were harvested 48 hours post transfection to isolate total RNA with PureLink Micro kit (Life Technologies). Transfection experiments were repeated 3 times.

The effect of knockdown of the analyzed genes on the pattern of the dsx splicing was evaluated through RT-PCR using total RNA templates from the transfected Sua5.1 cells and primers dsxF2 and dsxR5m. Similarly, RT-PCR was used to analyze the effect of a stable fle knockdown on the splicing pattern of dsx and fru (the latter using primers Aga_fruF and Aga_fruR) in pupae of selected transgenic A. gambiae lines.

Transient gene silencing experiments were conducted as described earlier.²¹ Briefly, early preblastoderm A. gambiae or A. stephensi embryos of unknown sex were microinjected with a solution containing dsRNA (1-1.5 μg/μl) of either fle or bla (as a control) gene and a plasmid (0.2 μg/μl) with the GFP gene downstream of the Drosophila melanogaster actin 5C promoter, or with plasmid alone as a control. Surviving first instar larvae were screened for the presence of GFP marker in the midgut cells using an M165 FC microscope equipped with a GFP filter. The larvae were sorted into GFP-negative and GFP-positive groups (cf. Figure S9 in Krzywinska et al.²¹). After pupation, sex of individuals was determined based on morphological characters. The experiments in A. gambiae were repeated three times with each of the two dsRNAs (cf. Figure 2A), and in A. stephensi were repeated two times.

To create the transgenic construct we used the p165 plasmid⁶¹ backbone by digesting the plasmid with MluI and NotI and ligating a linker containing the SfiI site and compatible MluI/NotI ends. The resulting plasmid was digested with SfiI and NotI, and two in vitro synthesized inserts were incorporated in a single ligation reaction. One insert, with SfiI and NheI ends, encoded the puromycin resistance gene pac under the control of the AGAP004395 promoter; the other insert, with NheI/NotI ends, encoded SV40 terminator, followed by GFP under the control of the AGAP003430 promoter (both promoters drive expression in the Sua5.1 cells, and the AGAP003430 promoter in vivo throughout development in gastric caeca, anterior and posterior stomach, Malpighian tubules and rectum, and at lower level in the brain, thoracic muscles and anal papillae). The resulting plasmid was digested with MluI and SfiI to clone a PCR-generated A. gambiae polyubiquitin promoter⁷⁶ with MluI/BsaI-FseI ends and a PCR-generated SV40 terminator with FseI/SfiI ends. Finally, that plasmid was digested with BsaI and FseI to clone a fle_3miR fragment encoding a polycistronic transcript designed to silence expression of fle. The fle_3miR fragment contained three microRNAs targeting fle and was constructed by annealing overlapping oligonucleotides (Table S2) and PCR amplification, as previously described.⁷⁷ After cloning into pUAST-attB vector,⁶² the fle_3miR fragment was released using BsaI and FseI. Thus engineered construct was excised using MluI and AsiSI and cloned into a MluI and AsiSI-cut p165-based plasmid backbone flanked by piggyBac arms and a φC31 attB site added to each end in opposite orientation, to create a transformation plasmid pBac_attBs_fle-3miR (GenBank: {"type":"entrez-nucleotide","attrs":{"text":"MW147152","term_id":"1959114186","term_text":"MW147152"}}MW147152).

Early preblastoderm A. gambiae embryos were microinjected with a solution of pBac_pattPs_fle_3miR (0.4 μg/μl) and a helper plasmid pENTR R4-vas2-Transposase-R3 (0.2 μg/μl), containing piggy-Bac transposase reading frame under the control of the vas2 regulatory sequences,⁶³ following previously described methods.⁷⁸ Injection of approximately 1200 embryos yielded 252 G0 larvae, of which 108 individuals, that exhibited a transient fluorescent marker expression, reached the pupal stage. The emerging adults (59 males and 49 females) were placed in 9 same-sex pools for crosses with the wild-type G3 strain mosquitoes. Over 120 transgenic G1 mosquitoes were recovered from four pools of male founders, whereas no transgenic G1 individuals were produced by the female founders. Selected G1 males originating from different founder cages or exhibiting different intensity of fluorescent marker expression (if originating from the same cage; Figure 2I) were crossed to wild-type females to establish 7 independent lines. Progeny from these crosses were screened for inheritance of fluorescent marker to evaluate transgene copy number. In three lines approximately 50% of G2 individuals were transgenic, indicative of single insertions. Four other lines exhibited 60%–89% transgene inheritance, with transgenic G2 individuals representing up to three discernible classes of fluorescence intensity or pattern per line, indicative of multiple insertions. In an effort to isolate single insertion sub-lines, males derived from multiple-insertion lines and representing different fluorescence classes were backcrossed with wild-type females at consecutive generations and the number of fluorescence phenotypes was monitored at each generation. Sub-lines that, after 6 generations, produced more than one fluorescence phenotype, or in which more than 50% of the individuals inherited the transgene, were eliminated. Finally, molecular characterization was used to confirm that each line from the final set possessed a single, unique genomic transgene integration site.

No efforts have been made to generate fle loss-of-function mutants in this study out of feasibility concerns. Tra2 in Drosophila, in addition to female sex determination, is necessary for male germline development – loss of Tra2 function leads to male sterility.⁷⁹ Fle may perform a similar role during the Anopheles spermatogenesis, which, combined with an apparent fle haploinsufficiency in females, could make recovering fle knockout mutants biologically impossible.

The integration sites of the piggyBac element within the genome has been identified using the splinkerette PCR protocol⁸⁰ or by inverse PCR. DNA isolated from individual pupae was used for both approaches. For inverse PCR, the DNA was digested with CviQI, HaeIII, MspI, Sau3AI, or TaqI (NEB), circularized by ligation, and amplified by PCR using primers ITRL1F and ITRL1R for piggyBac left arm, or ITRR1F⁸¹ and InpBacR2R for piggyBac right arm (Table S2) to isolate flanking genomic regions. The products containing genomic sequences flanking the piggyBac elements were sequenced directly, or after cloning, and genomic location of the integration sites was identified by BLAST search.

Adult mosquito abdomens were dissected in phosphate-buffered saline (PBS) to release tergites with the associated musculature. The tissues were fixed in PBS containing 4% paraformaldehyde for 15 min, washed three times for 5 min in PBS, and incubated in ActinGreen 488 ReadyProbes Reagent containing AlexaFluor 488-conjugated phalloidin. After three short washes the tissues were mounted on slides and photographed with a Leica DFC365 FX camera mounted on a Leica M165 FC microscope equipped with a GFP filter. Images were processed with ImageJ.⁶⁹

Total RNA was extracted from individual A. gambiae pupae using PureLink RNA Micro Kit (Invitrogen) according to manufacturer’s recommendations. For each sample, 500 ng of total RNA was used to synthesize cDNA with LunaScript RT SuperMix Kit (NEB). Quantitative PCR was conducted using primer pairs JK1051/JK1052 and JK1053/JK1054 to amplify, respectively, a fragment of fle and of the housekeeping gene encoding ribosomal protein S7 (rpS7, AGAP010592) used to normalize the expression. QuantStudio 3 Real-Time PCR System (Applied Biosystems) was employed to run the reactions using Luna Universal qPCR Master Mix (NEB) at annealing temperature of 59°C. Expression levels were calculated using 2^−ΔΔCt method,⁸² with triple technical and three biological replicates for each sample, and all data normalized to the relative fle/rpS7 expression in the samples of the wild-type female pupae.

Total RNA was extracted using the Trizol method and quality-checked using TapeStation (Agilent). Triplicate samples of female pupae from wild-type G3 line and from transgenic 4M4B line were used for transcriptome sequencing. The TruSeq library preparation protocol (Illumina) was followed by 150 bp paired-end sequencing using NovaSeq 6000 sequencing system (Illumina). The reads were pseudo-aligned to the A. gambiae transcriptome genebuild AgamP4.12 using Kallisto v0.46.⁷⁰ Transcripts per kilobase million (TPM) value was quantified for each transcript and averaged across multiple replicates of the same sample. As a further check of statistical robustness, differential expression analysis was performed using DESeq2,⁷¹ with filtering out transcripts covered by less than 10 reads among all samples, and then shrinking log2-fold changes using apeGLM.⁷²