A male steroid controls female sexual behaviour in the malaria mosquito – Nature

Rearing of A. gambiae mosquitoes

The A. gambiae G3 strain was reared under standard insectary conditions (26–28 °C, 65–80% relative humidity, 12:12 h light/darkness photoperiod). Larvae were fed on powdered fish food (TetraMin Tropical Flakes, Koi Pellets and Tetra Pond Sticks in a 7:7:2 ratio). Adult mosquitoes were fed ad libitum on 10% glucose solution and weekly on human blood (Research Blood Components). Virgin mosquitoes were obtained by separating the sexes at the pupal stage after microscopic examination of the terminalia. Males carrying the DsRed transgene have been previously described²⁵.

Forced and natural mating

Forced mating experiments were conducted according to protocols described previously¹³. For natural mating, 4-day-old virgin females were kept for two nights with sexually mature virgin males at a 1:3 ratio. For experiments using dsEPP-injected males, co-caging coincided with days 3–4 after injection, when phosphatase activity was maximally silenced (Extended Data Fig. 2b).

HPLC–MS/MS sample preparation for ecdysteroids

Mosquito tissues, remaining carcasses (rest of body) or whole bodies were dissected into 100% methanol and homogenized with a bead beater (2-mm glass beads, 2,400 r.p.m., 90 s). The number of tissues and volume of methanol were as follows: rest of body, 50 in 1,000 µl; MAGs, 50–100 in 80 µl; female LRT, 25–50 in 80 µl. Methanol extraction was performed a second time on the pellet using the same volume of methanol. Cellular debris was removed by centrifugation. Methanol from two extractions was combined and dried under nitrogen flow and resuspended in the following volumes of 80% methanol in water: rest of body, 50 µl; MAGs and female LRT, 30 µl.

HPLC–MS/MS analysis

Samples were analysed on a mass spectrometer (ID-X, Thermo Fisher) coupled to an LC instrument (Vanquish, Thermo Fisher). Five microlitres of sample was injected onto a 3-µm, 100  ×  4.6  mm column (Inspire C8, Dikma) maintained at 25 °C. The mobile phases for the LC were A (water, 0.1% formic acid) and B (acetonitrile, 0.1% formic acid). The LC gradient was as follows: 1 min at 5% B followed by an increase to 100% B in 11 min. After 8 min at 100%, the column was re-equilibrated at 5% B for 4 min. The flow rate was 0.3 ml min⁻¹. Ionization in the MS source was done by heated electrospray ionization in positive and negative modes.

The mass spectrometer measured data in full-MS mode at 60,000 resolution over an m/z range of 350 to 680. The MS/MS data were acquired on [M + H]⁺ (all targets), [M − H₂O + H]⁺ (all targets) and [M − H]⁻ (phosphorylated targets). The MS/MS data were used to confirm the ecdysteroid nature of the targets for which no standards were available. To identify untargeted ecdysteroids, the MS/MS data for all HPLC peaks of >15% relative abundance were analysed. Quantification was done using standard curves created from either pure standards (20E, 3D20E) to calculate absolute amounts or dilutions of one specific sample (all other targets) to calculate their equivalence to the amount found in one male. For 3D20E, quantification was performed with the sum of the following adducts: [M + TFA]⁻, [M + COOH]⁻, [M + Na]⁺, [M + Cl]⁻, [M + NO₃]⁻. Data were extracted and quantified using Tracefinder (version 4.1). MS/MS data were analysed using Xcalibur (version 4.4). The MS spectra of E, 20E and 3D20E were compared to respective standards. 3D20E22P was analysed by derivatization with Girard’s reagent. 20E22P was analysed by m/z ratio.

HPLC–MS/MS purification of 3D20E22P

3D20E22P was purified from MAGs. The purification was performed at analytical scale under the same LC conditions as for HPLC–MS/MS analysis using an ultra-performance liquid chromatography instrument (Acquity, Waters) coupled with a quadrupole mass-based detector (QDa, Acquity, Waters). Fraction collection was triggered when the m/z corresponding to 3D20E22P was detected at the same retention time as that previously identified. The purity of the extracted compounds was then checked by HPLC–MS/MS as described above.

RNA extraction, cDNA synthesis and RT–qPCR

Total RNA was extracted with TRI reagent (Thermo Fisher) from pools of 10–12 reproductive tissues or the rest of the body (headless) following the manufacturer’s instructions. RNA was treated with TURBO DNase (Thermo Fisher). cDNA was synthesized using Moloney murine leukaemia virus reverse transcriptase (M-MLV RT; Thermo Fisher) following the manufacturer’s instructions. Primers used for quantitative PCR with reverse transcription (RT–qPCR; Extended Data Table 2) were previously published²⁴ or were designed using Primer-BLAST²⁶ with preferences for products 70–150 bp in size and for primers spanning exon–exon junctions or primer pairs on separate exons. cDNA samples from three to four biological replicates were diluted four-fold in water for RT–qPCR. Quantification was performed in 15-µl duplicated reactions containing 1× PowerUp SYBR Green Master Mix (Thermo Fisher), primers and 5 µl of diluted cDNA. Reactions were run on a QuantStudio 6 Pro Real-Time PCR System (Thermo Fisher), and the data were collected and analysed using Design and Analysis (version 2.4.3). Relative quantities were normalized against the ribosomal gene RpL19 (AGAP004422), the expression of which does not change significantly with blood feeding²⁷ or mating³, as confirmed in this study.

RNA-seq analysis

RNA quality was checked with an Agilent bioanalyser 2100 Bioanalyzer (Agilent). Illumina paired-end libraries were prepared and run at the Broad Institute of MIT and Harvard. Sequencing reads were aligned to the A. gambiae genome (PEST strain, version 4.12) using HISAT2 (version 2.0.5) with the default parameters. Reads with mapping quality (MAPQ) scores <30 were removed using Samtools (version 1.3.1). The numbers of reads mapped to genes were counted using htseq-count (version 0.9.1) with the default parameters. Calculation of normalized read counts and analysis of differential gene expression was performed using the DESeq2 package (version 1.28.1) in R (version 4.0.3).

Bioinformatic pipeline

Ecdysteroid-modifying gene candidates were identified by first searching the A. gambiae genome with the PSI-BLAST algorithm (https://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/2.8.1/) using the default parameters with the following query protein sequences: EcK from Bombyx mori (accession NP_001038956.1), Musca domestica (accessions XP_005182020.1, XP_005175332.1 and XP_011294434.1) and Microplitis demolitor (accessions XP_008552646.1 and XP_008552645.1); EPP from B. mori (accession NP_001036900), Drosophila melanogaster (accession NP_651202), Apis mellifera (accession XP_394838) and Acyrthosiphon pisum (accession: XP_001947166); and EO from B. mori (accessions NP_001177919.1 and NP_001243996.1) and D. melanogaster (accession NP_572986.1) (step 1). Next, hit genes were filtered on the basis of high mRNA expression (>100 fragments per kilobase of exon per million mapped reads (FPKM) or >85th percentile) in reproductive tissues (female LRT or MAGs) of A. gambiae (step 2). For improved specificity, we selected against candidate enzymes also expressed in the reproductive tissues of A. albimanus, an anopheline species that does not synthesize or transfer ecdysteroids during mating⁷; candidate genes were filtered on the basis of low expression (<100 FPKM or <85th percentile) in reproductive tissues of A. albimanus (step 3). As a final filter (step 4), candidate genes needed to satisfy at least one of the following: (1) significant upregulation after mating (P < 0.05) according to the analysis of differentially expressed genes and (2) lack of expression in non-reproductive tissues (<85th percentile or <100 FPKM).

Whole-female stable isotope labelling

We modified a previously described method^28,29,30 to achieve whole-organism isotopic labelling. In brief, wild-type Saccharomyces cerevisiae type II (YSC2, Sigma) was grown in medium containing (wt/vol) 2% glucose (G7528, Sigma), 1.7% yeast nitrogen base without amino acids and ammonium sulfate (BD Difco, DF0335) and 5% ¹⁵N ammonium sulfate (NLM-713, >99%, Cambridge Isotope Laboratories) as the only source of nitrogen. Yeast was recovered by centrifugation and fed ad libitum to mosquito larvae until pupation. Powdered fish food was supplemented (0.5 mg per 300 larvae) to prevent death at the fourth instar stage. Only females were then used in mating experiments with unlabelled males to analyse the male proteome transferred during copulation.

Male ejaculate proteome (ejaculome)

¹⁵N-labelled virgin females of 4–6 days old were force-mated to age-matched unlabelled virgin males. Successful mating was verified by detection of a mating plug under an epifluorescence microscope. At 3, 12 and 24 h.p.m., the atria of 45–55 mated females were dissected into 50 µl of ammonium bicarbonate buffer (pH 7.8) and homogenized with a pestle. The homogenate was centrifuged, and the supernatant was combined with 50 µl of 0.1% RapiGest (186001860, Waters) in 50 mM ammonium bicarbonate. The supernatant and pellet from each sample were snap-frozen on dry ice and shipped overnight to the MacCoss Lab at the University of Washington, where sample preparation for LC–MS/MS was completed. The pellets were resuspended in 50 µl of 0.1% RapiGest in 50 mM ammonium bicarbonate and sonicated in a water bath. Protein concentration was measured for both pellets and supernatants by BCA assay, and the samples were reduced with 5 mM dithiothreiotol (DTT; Sigma) alkylated with 15 mM iodoacetamide (Sigma) and digested with trypsin for 1 h at 37 °C (1:50 trypsin:substrate ratio). RapiGest was cleaved with the addition of 200 mM HCl followed by 45 min of incubation at 37 °C and centrifugation at 14,000 r.p.m. for 10 min at 4 °C to remove debris. Samples were cleaned by dual-mode solid-phase extraction (Oasis MCX cartridges, Waters) and resuspended in 0.1% formic acid at a final protein concentration of 0.33 µg µl⁻¹. Unlabelled MAG proteomes were similarly analysed from virgin males. Two analytical replicates were analysed for each sample. Next, 1 µg of each sample digest was analysed using a 25-cm fused silica 75-μm column and a 4-cm fused silica Kasil1 (PQ) frit trap loaded with Jupiter C12 reverse phase resin (Phenomenex) with a 180-min LC–MS/MS run on a Q-Exactive HF mass spectrometer (Thermo Fisher) coupled with a nanoACQUITY UPLC system (Waters). Data-dependent acquisition data generated from each run were converted to mzML format using Proteowizard (version 3.0.20287) and were searched using Comet³¹ (version 3.2) against a FASTA database containing protein sequences from A. gambiae (VectorBase release 54), Anopheles coluzzi Mali-NIH (VectorBase release 54), S. cerevisiae (Uniprot, March 2021), three-frame translations of A. gambiae RNA-seq and known human contaminants. Peptide-spectrum match FDRs were determined using Percolator³² (version 3.05) at a threshold of 0.01, and peptides were assembled into protein identifications using protein parsimony in Limelight³³ (version 2.2.0). Relative protein abundance was estimated using a normalized spectral abundance factor (NSAF) calculated for each protein within each run as previously described²⁸. The NSAF relative to each protein was averaged across samples from two different biological replicates. ¹⁵N labelling successfully masked the female proteome, although a small number of unlabelled proteins could be detected from labelled virgins. We documented the detection at reduced amounts (1–5 spectra) of male proteins in female virgin samples only in technical runs in which the virgin samples were run after male/mated samples, as a result of HPLC ‘carryover’. Proteins that were occasionally found as ‘contaminants’ from labelled virgins are listed in Supplementary Table 1.

Customized antibody and western blotting

Two antigenic peptides, QTTDRVAPAPDQQQ (within isoform PA) and MESDGTTPSGDSEQ (within both isoforms PA and PB), from the EPP protein sequence were identified on the basis of predictions of antigenicity, surface exposure probability and hydrophilicity provided as a part of the customized antibody service at Genscript. The two peptides were pooled and were then conjugated with carrier protein KLH and injected into New Zealand rabbits. The rabbits were killed after the fourth injection, and total IgG was isolated with affinity purification. IgG from the most EPP-specific rabbit was used in further western blotting.

For western blotting, MAGs (n = 10, where n indicates the number of biologically independent mosquito samples) and female LRTs (n = 30) were dissected from 4-day-old virgin males and virgin or force-mated females (<10 min after mating), respectively, into protein extraction buffer (50 mM Tris, pH 8.0; 1% NP-40; 0.25% sodium deoxycholate; 150 mM NaCl; 1 mM EDTA; 1× protease inhibitor mixture (Roche)). The samples were immediately homogenized after dissection with a bead beater (2-mm glass beads, 2,400 r.p.m., 90 s). The insoluble debris was removed by centrifugation at 20,000g at 4 °C. Protein was quantified by Bradford assay (Bio-Rad). Then, 20 µg of MAG protein, 40 µg of LRT protein and 20 µg of rest of body protein were denatured and separated by 10% Bis–Tris NuPAGE using MOPS buffer. Protein was transferred to a polyvinylidene difluoride membrane using an iBlot2 transfer system (Thermo Fisher). The membranes were washed twice in 1× PBS-T (0.1% Tween-20 in PBS) and were then blocked for 1 h at 22 °C in Odyssey Blocking Buffer (Li-Cor). The membranes were incubated with the custom rabbit anti-EPP polyclonal primary antibody (1:700 in blocking buffer) and rat anti-actin monoclonal primary antibody MAC237 (Abcam; 1:4,000) with shaking overnight at 4 °C. The membranes were washed with PBS-T and were then incubated with secondary antibodies (donkey anti-rabbit 800CW and goat anti-rat 680LT (Li-Cor), both at 1:20,000) in blocking buffer with 0.01% SDS and 0.2 % Tween-20 for 1 h at 22 °C. The membranes were washed with PBS-T and imaged with an Odyssey CLx scanner. Images were collected and processed in Image Studio (version 5.2). No specific band corresponding to the EPP-RA isoform (82 kDa) was detected.

Recombinant protein and enzymatic assays

The coding regions of EPP (as isoform AGAP002463-RB containing the histidine phosphatase domain, NCBI Conserved Domain Search³⁴) and EcK2 (AGAP002181) were cloned into the pET-21a(+) plasmid (Novagen Millipore Sigma); primers are listed in Extended Data Table 2. Eight GS4 linkers (in tandem) were inserted before the C-terminal 6×His tag in the pET-21a(+)-EcK2 construct. Recombinant proteins were produced with NEBExpress cell-free Escherichia coli protein synthesis reactions (New England BioLabs). The recombinant proteins were purified with NEBExpress Ni spin columns (New England BioLabs). The dihydrofolate reductase (DHFR) control protein was produced using the DNA template from the NEBExpress cell-free E. coli protein synthesis kit. Proteins were stored in 50% glycerol in PBS at −20 °C for up to 3 months.

The phosphatase activity of EPP and tissue extracts was measured using 4-nitrophenyl phosphate (pNPP; Sigma-Aldrich). The reaction buffer contained 25 mM Tris, 50 mM acetic acid, 25 mM Bis–Tris, 150 mM NaCl, 0.1 mM EDTA and 1 mM DTT. Tissues were homogenized in reaction buffer, and cellular debris was removed by centrifugation. The reaction was initiated by adding enzyme or tissue extract to the reaction buffer containing 2.5 mg ml⁻¹ pNPP. The reaction mixture was incubated at room temperature in darkness, and the amount of pNP converted from pNPP was quantified by measuring the absorption at 405 nm at various times.

For in vitro EcK activity, proteins were incubated with 0.2 mg of 20E or 3D20E in 200 µl of buffer (pH 7.5) containing 10 mM HEPES–NaOH, 0.1% BSA, 2 mM ATP and 10 mM MgCl₂ at 27 °C for 2 h. The reaction was terminated by adding 800 µl of methanol and was then chilled for 1 h at −20 °C followed by centrifugation at 20,000g for 10  min at 4 °C. The supernatant was then analysed by HPLC–MS/MS. To heat-inactivate proteins used in the control groups, the proteins were incubated at 95 °C for 20 min in 50% glycerol in PBS.

For in vitro EPP activity, proteins were incubated with 3D20E22P (equivalent to the amount found in 18 pairs of MAGs, purified by HPLC–MS/MS) in 100 µl of buffer (pH 7.5) containing 25 mM Tris, 50 mM acetic acid, 25 mM Bis–Tris, 150 mM NaCl, 0.1 mM EDTA and 1 mM DTT at 27 °C for 3 h. The reaction was terminated by adding 400 µl of methanol and was chilled for 1 h at −20 °C followed by centrifugation at 20,000g for 10 min at 4 °C. The supernatant was analysed by HPLC–MS/MS.

dsRNA production

PCR fragments of EPP (362 bp), EcK1 (AGAP004574, 365 bp) and EcK2 (556 bp) were amplified from cDNA prepared from mixed-sex headless mosquito carcasses. A PCR fragment of the eGFP control (495 bp) was amplified from pCR2.1-eGFP described previously¹¹; the PCR primers are listed in Extended Data Table 2. PCR fragments were inserted between the inverted T7 promoters on the pL4440 plasmid. Plasmid constructs were recovered from NEB 5-alpha Competent E. coli (New England Biolabs) and were verified by DNA sequencing before use (see Supplementary Data 1 for insert sequences). A primer matching the T7 promoter (Extended Data Table 2) was used to amplify inserts from pL4440-based plasmids. PCR product sizes were confirmed by agarose gel electrophoresis. dsRNA was transcribed from the PCR templates using the Megascript T7 transcription kit (Thermo Fisher) and was purified following the manufacturer’s instructions with modifications previously described³⁵.

Thorax microinjection

For dsRNA injection, 1,380 ng of dsRNA (dsGFP, dsEcK1, dsEcK2, dsEPP) was injected (Nanoject III, Drummond) at a concentration of 10 ng nl⁻¹ into the thorax of adult males or females within 1 day of eclosion. Gene knockdown levels were determined in at least three biological replicates by RNA extraction, cDNA synthesis and RT–qPCR. For ecdysteroid injections, depending on the experimental design, 4-day-old virgin or 6-day-old virgin blood-fed females were injected (Nanoject III, Drummond) with 0.13, 0.21 or 0.63 µg of 20E or 3D20E at a concentration of 1.3, 2.1 or 6.3 ng nl⁻¹, respectively. Ten percent (vol/vol) ethanol in water was injected at a volume of 100 nl; 3D20E22P in 10% ethanol was injected at a volume of 100 nl (equivalent to 75% of the amount found in a pair of MAGs). Mosquitoes were randomly assigned to injection groups.

Oviposition and remating assays

For oviposition assays, 3-day-old females were blood-fed ad libitum on human blood. Partially fed or unfed mosquitoes were removed. Depending on the treatment, after a minimum of 48 h after the blood meal, females were placed into individual oviposition cups for four nights. Eggs were counted under a stereoscope (Stemi 508, Zeiss); for mated females, eggs that hatched into larvae were scored as fertile.

For mating assays, depending on the treatment, females were allowed a minimum of 2 days to develop refractoriness to mating, and wild-type age-matched males were subsequently introduced to the same cage. After two nights, spermathecae were dissected from females, and the genomic DNA was released by freeze–thawing and sonication in a buffer containing 10 mM Tris–HCl, 1 mM EDTA and 25 mM NaCl (pH 8.2). The samples were incubated with proteinase K (0.86 µg µl⁻¹) for 15 min at 55 °C followed by 10 min at 95 °C. The crude genomic DNA preparation was diluted 10-fold and was subjected to qPCR detection of a Y-chromosome sequence; the primers are listed in Extended Data Table 2. The absence of a Y-chromosome sequence is indicative of no mating.

For remating assays, force-mated females were checked for the presence of mating plugs to confirm mating status and were allowed 2 days to develop refractoriness to mating in the absence of males, as described previously³⁶. Males carrying DsRed transgenic sperm were subsequently introduced to female cages. After two nights, spermathecae were dissected from females, and the genomic DNA was prepared as described above and subjected to qPCR detection of the DsRed transgene; the primers are listed in Extended Data Table 2. The absence of the DsRed transgene indicates no occurrence of remating.

3D20E synthesis

3D20E was synthesized as previously described³⁷. In brief, 10 mg of 20E (Sigma-Aldrich) was dissolved in 10 ml of water, and 30 mg of platinum black (powder form, Sigma-Aldrich) was then added. A gentle stream of O₂ gas was bubbled continuously into the reaction mixture, which was stirred at room temperature. After 6 h, the reaction was terminated by adding 30 ml of methanol. The mixture was centrifuged, and the catalyst pellet was removed. The supernatant was evaporated to dryness under vacuum at room temperature. The dried reaction product was dissolved in 10% ethanol for injection and methanol for HPLC–MS/MS analysis. The conversion rate (from 20E to 3D20E) was approximately 97% (Fig. 4b), and the MS spectra of the synthesized 3D20E matched those of 3D20E found in mated females (Fig. 4c).

Statistical analysis

The figure legends contain specific details of the statistical tests performed. GraphPad (version 9.0) was used to perform Fisher’s exact tests, Mantel–Cox tests and Student’s t-tests. Cochran–Mantel–Haenszel tests were performed using a customized R script (available at https://github.com/duopeng/mantelhaen.test). The normality of data distribution was tested with the Shapiro–Wilk test using a significance threshold of 0.05. Mann–Whitney tests were performed when data failed to pass the normality test. Survival data were analysed with the Mantel–Cox test. The DESeq2 package (version 1.28.1) was used to perform RNA-seq gene-level differential expression analysis. Horizontal bars on graphs represent medians. A significance value of P = 0.05 was used as the threshold in all tests.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this paper.