Amino acid–dependent regulation of insulin-like peptide signaling is mediated by TOR and GATA factors in the disease vector mosquito Aedes aegypti

Significance Aedes aegypti female mosquitoes are vectors of human viral diseases because they require vertebrate blood for their reproduction. Using CRISPR-Cas9 gene modification, transgenesis, and RNA interference in the mosquito, we identified the molecular mechanism underlying the nutritional pathway, in which the serine/tyrosine kinase target of rapamycin (TOR) mediates the amino acid (AA) signaling through GATA transcription factors and affects insulin-like peptides (ILP). GATA repressor (GATAr) inhibits the transcription of ilp4ilp6, and ilp7 genes. FoxO, a downstream TF in the insulin pathway, prevents the TOR-GATAr-mediated repression of these ilp genes. After the blood meal–mediated AA influx, GATA activator elevates the expression of ilp1ilp2ilp3ilp5, and ilp8 genes. Targeting AA-dependent reproductive processes may provide effective means in controlling mosquito population.

Hematophagous female mosquitoes are important vectors of numerous devastating human diseases (1). Mosquito-borne diseases are among the world's leading causes of death and illness today, and mosquito control is of the utmost importance for public health (2)(3)(4). The distribution of mosquito-borne diseases is determined by a complex set of social, demographic, and environmental factors. The Aedes aegypti mosquito is the major vector of Dengue fever, yellow fever, Chikungunya, Zika, and other vector-borne viral human diseases (5)(6)(7)(8)(9). Every female mosquito produces a batch of about 120 eggs at one time, and three or more batches in its lifespan, mounting an impressive reproductive capacity. Each reproductive cycle of a female mosquito consists of two phases: a previtellogenic phase (posteclosion; PE), during which female's tissues are preparing for egg development, and a vitellogenic phase (post blood meal; PBM), during which egg development occurs following blood feeding. The first phase lasts about 3 to 4 d, during which a female feed on nectar until it finds a host. The second, the vitellogenic phase, is initiated following blood feeding and usually lasts 72 h or until a female finds another host and feeds on blood to initiate another gonadotrophic cycle. Thus, the two phases of each gonadotropic cycle are drastically different in dietary intake-nectar feeding (carbohydrate) during the PE and blood feeding (protein) during PBM (10)(11)(12). Feeding on carbohydrate during the PE phase facilitates mating, flight, host seeking, and preparation for egg production (11). Elucidation of the molecular mechanisms underlying nutrient and metabolic needs in support of female mosquito reproduction is essential for implementation of vector population-control approaches.
Insulin is the major hormone regulating nutrition activating the uptake of glucose, fatty acids, and amino acids (AAs) along with their storage in the form of glycogen, lipid, and protein. Dysregulation of the insulin pathway leads diabetes and other metabolic disorders in humans (13,14). Insect insulin-like peptides (ILPs) are 6-to 8-kDa molecules with the insulin fold motif that belong to the insulin superfamily and are encoded by multiple genes (15,16). ILPs are produced by the insect brain and other tissues and function in the coordination of metabolism, development, and reproduction (16,17). Insects store glycogen and triacylglycerides (TAGs) as energy reserves in the adipocytes of the fat body for growth and reproduction. There are eight different ilps genes (ilps [1][2][3][4][5][6][7][8] in the genome of the Ae. aegypti mosquito. Previous studies have revealed that CRISPR-Cas9 disruption of each ILP in female mosquitoes caused differential degrees of deficiency in growth, nutrient storage, carbohydrate/lipid balance, and egg development (18,19).
AAs from a blood meal are necessary for anautogenous female mosquitoes to complete a gonotrophic cycle (10). The level of free AAs rises sharply in the mosquito hemolymph as digestion of the vertebrate blood meal occurs in the midgut (20). AAs are absorbed and carried to the mosquito fat body cells by specific transporters (21,22). The fat body, an insect organ analogous to the vertebrate liver and adipose tissue combined, serves as the nexus for nutrient sensing, lipid storage, and endocrine signaling to the brain and reproductive organs (23,24). AAs provide building blocks for protein synthesis and are also used for energy expenditure; in addition, free AAs are known as regulators of cellular signaling to coordinate metabolism (20,25). However, how insulin signals coordinate their action with the AA signaling during mosquito reproduction remains largely unknown. This knowledge is vital for a thorough understanding of the mechanisms controlling anautogenous reproduction in mosquitoes and may provide the background for the utilization of these peptide hormones or their regulators for management of mosquitoes and vector-borne diseases.
The ILP and the target of rapamycin (TOR) pathways are two main nutrient sensors (26)(27)(28). AAs are critical signals for TOR activation (29). In the mosquito, after receiving the nutrition signal from AA-TOR, the activated GATAa (GATA activator) displaces the GATAr (GATA repressor) isoform to enhance gene expression (30). GATAr or GATAa knockouts in mosquitoes are lacking because of challenges related to alternative splicing of the GATA gene pre-mRNA. We examined the role of TOR-mediated AA signaling in ilp expression by comparing the analysis of AA infusion, RNA-interference (RNAi) silencing of TOR, and isoform-specific CRISPR-Cas9 genomic editing of GATAr or GATAa. We found that TOR differentially regulates ilp expression in female mosquitoes. Further, we showed that the expression of ilp4, ilp6, and ilp7 was inhibited by the GATAr isoform in response to low AA-TOR signaling, while the expression of ilp1, ilp2, ilp3, ilp5, and ilp8 was activated by the GATAa isoform in response to the increased AA-TOR signaling. Importantly, FoxO, a downstream transcription factor (TF) in the insulin signaling pathway, is involved as an interrupter role in TOR-GATAr-mediated repression of ilp4, ilp6, and ilp7 expression. Thus, our results reveal a signaling mechanism of how ILPs are controlled to coordinate their actions with the nutrient availability and reproductive needs in mosquitoes.

Differential Actions of AA and TOR on ilp Gene Expression.
Our previous study has demonstrated that ilps 4, 6, and 7 are upregulated during the PE phase and down-regulated after blood feeding. In contrast, ilps 1, 2, 3, 5, and 8 exhibited high levels of expression during the PBM phase (SI Appendix, Fig. S1) (31). Thus, in addition to hormonal input, identified in the abovementioned work, a direct nutritional signaling maybe involved in controlling ilp gene expression. Signaling through the TOR pathway is the key to activation of vitellogenesis. The AA transporter Slimfast is critical for this signaling (10,21,32). To examine whether AA signaling is engaged in regulating ilp gene expression, we first conducted AA infusion into the hemocoel of the adult female mosquito during the PE phase. It has been shown that the addition of a balanced AA mixture into the hemocoel of previtellogenic mosquitoes stimulates oogenesis (33). We introduced the solution containing 17 AAs into 3-d-old female A. aegypti adults for 24 h (0.083 μL/h × 24 h = 2.0 μL/24 h) or 48 h (4.0 μL/48 h) using the infusion method reported before (33) (SI Appendix, Fig. S2 A  and B). The transcript levels of each of the eight ilps were measured using real-time quantitative PCR (RT-qPCR) analysis. After the 24-h AA infusion ilps 4, 6, and 7 were down-regulated, while no effect was noticed on the expression of ilps 1, 2, 3, 5, and 8. This indicates that although the low level of AAs was sufficient for reducing the expression level of ilps in the first group, it was not sufficient for ilp activation of the second group (Fig. 1A). However, these five ilps (1, 2, 3, 5, and 8) were significantly up-regulated in mosquitoes infused with AAs for 48 h (Fig. 1B). Control females infused with a saline solution exhibited the same pattern of ilp gene expression as untreated mosquitoes. Thus, the AA signaling plays a critical role in regulating ilp gene expression in reproducing female mosquitoes.
To understand whether TOR is involved in mediating the AA signaling in coordinating ilp expression, we conducted TOR RNAi in a combination of AA infusion. Female mosquitoes were injected with double-stranded TOR RNA (dsTOR) or dsLuciferase (dsLuc) as a control at 6 h PE after adult emergence (SI Appendix, Fig. S2C). After 3 d of recovery, the dsRNA-injected females were infused with AAs for 24 h or 48 h (0.083 μL/h). The fat body tissue was dissected and used for RT-qPCR analysis. The transcript levels of AA-down-regulated ilp genes (ilps 4, 6, and 7) increased in dsTOR RNAi mosquitoes infused for 24 h, while no response was observed for the AA-up-regulated ilp genes (ilps 1, 2, 3, 5, and 8) (Fig. 1C). After 48-h infusion, no response was detected for AA-down-regulated ilp genes (ilps 4, 6, and 7) in dsTOR RNAi mosquitoes, while the AA-up-regulated ilp genes (ilps 1, 2, 3, 5, and 8) were down-regulated (Fig. 1D). These results clearly indicate that TOR is requited for mediating the AA signaling. Moreover, it mediates the repressive effect of AAs on ilps 4, 6, and 7 under the low AA level condition as well as the activating AA effect on ilps 1, 2, 3, 5, and 8 under the high-AA-level condition.
To investigate whether this pattern of the AA-TOR regulation extends to ILPs at the protein level, we generated CRISPR-Cas9 epitope-tagged ILPs in Ae. aegypti. As previously described, hemagglutinin (HA) and FLAG tags were fused into the B-and A-chains in each of eight ILPs using CRISPR-Cas9-mediated homology-directed repair and single-stranded oligodeoxynucleotides (34). We measured the hemolymph level of each epitope-tagged ILP by enzyme-linked immunosorbent assay (ELISA) in the CRISPR-Cas9 modified mosquito females. The hemolymph levels of ILPs 4, 6, and 7 were down-regulated after the 24-h AA infusion in the epitope-tagged 3-d-old female adults, while those of ILPs 1, 2, 3, 5, and 8 were up-regulated after the 48-h AA infusion (Fig. 2). Next, we performed AA infusion in combination with TOR RNAi. dsTOR was injected to the epitope-tagged females at 6 h PE and after 3 d of recovery, these mosquitoes were infused with AA (24 h or 48 h; 0.083 μL/h). We found that the hemolymph levels of ILPs 4, 6, and 7 were higher in AA-dsTOR RNAi mosquitoes (Fig. 2). In contrast, the hemolymph levels of ILPs 1, 2, 3, 5, and 8 were lower than those observed with dsLuc as a control (Fig. 2). The amounts of ILP proteins in the circulating hemolymph varied from 0.4 to 80 pg/μL.

GATAa and GATAr Are Critical for the Control of ilp Gene
Expression. Nutrition signals, particularly AAs, are reportedly transduced through TOR to GATA-type TFs, which regulate the target genes (30,33,35). To elucidate the mechanism and physiological function of GATA regulation of ilp gene expression in the mosquito, we utilized the CRISPR-Cas9 genome-editing approach to precisely disrupt GATAr and GATAa genes, respectively. In A. aegypti mosquitoes, these two GATA isoforms have distinct fifth exons. For generating isoform-specific CRISPR-Cas9 gene editing of GATAa and GATAr, we designed a pair of single-stranded guide RNAs (sgRNAs) (5a-s1, 5a-s2) flanking the GATAr-specific fifth (5a) exon and another pair (5b-s1 and 5b-s2) flanking the GATAa-specific exon-5b by scanning DNA sequences around the fifth exon of the gata gene ( Fig. 3 A and B). The sgRNA target sites were complementary to the spacer part of the CRISPR RNA and had a protospacer adjacent motif of NGG (the recognition site for Streptococcus pyogenes Cas9) (SI Appendix, Fig. S2). Each synthetic sgRNA (40 ng/μL) and Cas9 protein (333 ng/μL) were injected into mosquito embryos to generate genomic disruption, as previously described (18). Sanger sequencing revealed polymorphic mutations and depletion of GATAr and GATAa fifth exon sequences (SI Appendix, Fig. S2).
The GATAr-and GATAa-mutated mosquitoes at 72-h PE and 24-h PBM were used for RT-qPCR analysis. The transcript levels of ilps in either mutations or wild-type (WT) controls were measured. The levels of the AA-down-regulated ilps (ilps 4, 6, and 7) were elevated but the up-regulated ilps had no change in GATAr-disrupted mosquito females at 72-h PE relative to WT. Levels of the AA-up-regulated ilps (ilps 1, 2, 3, 5, and 8) were lower and the down-regulated ilps had no change in GATAa-disrupted mosquito females at 24-h PBM relative to WT (Fig. 3C). This experiment shows the specificity of obtained CRIPSR-CAS9 mutants for GATAa and GATAr. Moreover, it clearly indicated the differential action of GATAa and GATAr in controlling the ilp gene expression.
GATAr and GATAa Affect Egg Development. We dissected the ovaries from WT, GATAr and GATAa mutant vitellogenic mosquitoes at 24-h PBM, when the ovary had developed follicles. We found that 81% of the GATAr mutants and 72% of GATAa mutants displayed dramatic defects in ovarian development and egg maturation. Examination of the GATAr mutant ovaries revealed much smaller follicles at 24 h PBM and fewer laid eggs by the end of the mosquito gonadotrophic cycle than with WT mosquitoes (SI Appendix, Fig. S3 and Fig. 4). The GATAa mutant ovaries were half melanized, displayed several bigger follicles at 24-h PBM and produced fewer eggs than WT (SI Appendix,  To evaluate the effect of GATAr and GATAa in regulating nutrient metabolism during mosquito reproduction, we stained the lipid droplets in the fat bodies of WT and mutants (with abnormal ovaries) at 24-h PBM with Nile red. The lipid droplets were dramatically smaller and less abundant in the GATAr than in WT control. On the contrary, they were considerably bigger and more abundant in GATAa mutant females than in WT control. (Fig. 4). Mosquito females store lipid energy reserves as TAGs in the fat body cells during the sugar-feeding (PE) phase; TAGs are required for subsequent reproductive events (36,37). A significant difference in TAG levels in GATAr and GATAa mutants was detected compared with those in WT control mosquitoes (Fig. 4). GATAr Hemolymph protein levels of ILPs 4, 6, and 7 were lower after the 24-h AA infusion (AA1), but higher after the 24-h AA infusion and RNAi-TOR (dsTOR+AA1), than in control SS1 (C) Hemolymph protein levels of ILPs 1, 2, 3, 5, and 8 were higher after the 48-h AA infusion (AA2), but lower after the 48-h AA infusion and RNAi-TOR (dsTOR+AA2), than in control SS2. Hemolymph ILP-HA/FLAG content (pg/μL) was determined using ELISA in tagged females after AA infusion or RNAi treatments and respective controls (infusion of saline solution or injection of dsLuc). Data represent three biological replicates with five individuals in each and are shown as mean ± SEM; *P <0.05, **P < 0.01, ***P < 0.001. Schematic of sgRNA (upstream 5a-s1 and downstream 5a-s2) cleavage used for deletion of the fifth exon "exon-5a" of GATAr upon sgRNA (upstream 5b-s1 and downstream 5b-s2) cleavage of GATAa "exon-5b". (C) Isoform-specific knockout of GATAr (ΔGATAr) activated ilp transcripts 46, and 7 but displayed no significant effect on the other ilps before blood feeding (PE 72 h); isoform-specific knockout of GATAa (ΔGATAa) repressed ilp transcripts 1235, and 8, but no significant effects on the other three ilps after a blood meal (PBM 24 h). Wild-type females were used as control. Data represent three biological replicates with 30 individuals in each and are shown as mean ± SEM; *P < 0.05, **P < 0.01, ***P < 0.001. mutant females had highly reduced TAG levels, while the GATAa mutant females displayed elevated TAG levels at 24 h PBM (Fig. 4). However, glycogen levels were greater in GATAr and lower in GATAa mutants (Fig. 4). The GATAr and GATAa mutants created by CRISPR-Cas9 emphasize crucial and differential roles of these GATA isoforms in ovarian development and in lipid and sugar homeostasis.
GATAr Binds to ilps 4, 6, and 7, whereas GATAa Binds to ilps 1, 2, 3, 5, and 8. Mosquito GATAr or GATAa belong to GATA-type TFs. These are zinc-finger proteins that bind the consensus DNA sequence (A/T)GATA(A/G) in target gene regulatory regions and control the activity of these genes (30,35). Putative GATA-binding sites are present in regulatory regions of all eight ilp genes of the A. aegypti mosquito (Fig. 5A). To confirm authenticity of these bindings sites, we conducted chromatin immunoprecipitation (ChIP) analysis (see SI Appendix, Supplemental Section for the method details). GATA binding was enriched at the AA-downregulated, GATAr-repressed ilp genes (ilps 4, 6, and 7) in WT female mosquitoes at 72-h PE. This binding enrichment was diminished to a background level after the CRISPR-Cas9 knockout of GATAr (Fig. 5B). GATA binding was at the background levels in the AA-up-regulated, GATAa-activated ilp genes (ilps 1, 2, 3, 5, and 8) at 72-h PBM. In contrast, the GATA binding at this group of ilps was elevated after a blood meal (24 h PBM). The CRISPR-Cas9 disruption of GATAa diminished this binding enrichment (Fig. 5C). GATA binding was at the background levels in GATArrepressed ilp genes (ilps 4, 6, and 7) in WT as well in the GATAa knockouts at 24 h PBM (Fig. 5C). These experiments revealed that GATAr and GATAa binds specifically for different ilps.

Differential Effects of TOR, GATAr, and GATAa on FoxO
Localization in the Fat Body Cells. FoxO is a TF downstream of the insulin signaling pathway. Insulin binds to the insulin receptor and affects FoxO phosphorylation and cellular localization to fulfill its function (38). To study the effect of nutritional signaling on FoxO localization, we utilized the CRISPR-Cas9 homologous recombination to link the HA tag into the mosquito endogenous FoxO protein at the carboxy terminus, as previously described (18,34). To further explore the regulation of ILP signaling and action of the TOR and GATA factors, we observed HA-tagged FoxO intracellular localization in the fat body of female mosquitoes after RNAi knockdown of TOR or CRISPR-Cas9 knockouts of GATAr or GATAa. In mosquitoes with HA-tagged FoxO in the WT background, HA-tagged FoxO is localized in fat body nuclei at 72 h PE, while predominantly in the cytoplasm at 24-h PBM (Fig. 7). The CRISPR-Cas9 mutation of GATAr in female mosquitoes with HA-tagged FoxO with HA-tagged FoxO at 72-h PE resulted in its cytoplasmic retention. In contrast, the loss of either TOR or GATAa at 24-h PBM caused FoxO-HA nuclear translocation (Fig. 7). These results indicate that the GATAr and GATAa factors have an opposite effect on FoxO cellular localization in the fat body. FoxO is activated in a GATAr-dependent manner at 72-h PE, while its action is inhibited at 24-h PBM and this is GATAa dependent. It has been shown that FoxO binds to promoters of several carbohydrate and lipid metabolic enzyme genes in A. aegypti, and FoxO-RNAi elevates the transcript levels of several metabolic enzyme genes, including succinyl-coA synthetase, trehalose-6-phosphate synthase, malate dehydrogenase, lipase, and fatty acid synthases (34). The GATAr knockout also elevated the transcript levels of these metabolic enzyme genes, which is the opposite of the effects of the GATAa knockout or TOR-RNAi (SI Appendix, Fig. S5). To investigate the possibility of a direct effect of GATA on the expression of metabolic enzyme genes, we searched regulatory regions of these metabolic genes. GATA binding motifs were identified in the 5′ upstream regulatory regions of the above-mentioned metabolic enzyme genes (SI Appendix, Fig. S6). To determine whether GATA binds to these gene promoters, ChIP analysis was done in combination with qPCR to show GATA-bound regions. The GATA binding was undetectable at the promoters of tested metabolic enzyme genes in WT or GATA-RNAi females (SI Appendix, Fig. S6). This experiment revealed no direct genomic interaction between GATA factors and metabolic genes.

FoxO Prevents GATAr Binding to ilps 4, 6, and 7 Permitting Their
Expression at the PE Phase. AA-TOR dramatically activates GATAa gene expression after a blood meal, while the expression of the GATAr gene was down-regulated below a detectable level between 18-h and 24-h PBM when AAs increased to a maximal concentration in hemolymph (30,35). To investigate the GATAa and GATAr signaling in response to TOR RNAi, we inserted HA tag at the N terminus (sequence in common except for exon-5) through CRISPR-Cas9 for intracellular imaging of mosquito GATAr and GATAa (SI Appendix, Fig. S7). We detected HAtagged GATA intracellular localization in the fat body of female mosquitoes after RNAi TOR knockdown. No change in GATA signaling was observed with the loss of TOR in female mosquitoes with HA-tagged GATA at 72-h PE. In contrast, the loss of TOR at 24-h PBM caused GATA-HA reduction that correspond the regulation of TOR on GATAa while GATAr had no change (SI Appendix, Fig. S7). Given the absence of any effect of TOR on GATA protein levels at 72-h PE, to determine whether FoxO is involved in the GATAr-binding of ilp promoters, we performed ChIP-qPCR analysis to specifically amplify GATA-bound regions before and after blood feeding in FoxO-RNAi (dsFoxO) or FoxO-overexpressed (Vg-FoxO) female mosquitoes (SI Appendix, Fig. S10).
We generated the binary transgenic line Vg-Gal4/UAS-FoxO to overexpress FoxO after a blood meal in female fat bodies according to the previously described method (39). The UAS-FoxO construct containing the foxo coding sequence was inserted into the pBac[3×P3-DsRed] plasmid at the AscI restriction site. The responder line UAS-FoxO was created by injecting transformation vector pBac[3×P3-DsRed, UAS-FoxO] and helper into preblastodermal embryos. The driver Vg-Gal4 line was made previously (39). The binary transgenic mosquitoes Vg-Gal4/UAS-FoxO were produced by crossing the Vg-Gal4 female (green eyes with EGFP) with the UAS-FoxO male (red eyes with DsRed) (SI Appendix, Fig. S8). FoxO overexpression during PBM in the fat body of the Vg-Gal4/UAS-FoxO females was detected by means of RT-qPCR and results compared with those of WT, Vg-Gal4 and UAS-FoxO lines at 72-h PE, 12-h and 24-h PBM, as controls (SI Appendix, Fig. S8). GATA-binding was present and increased at the GATAr-repressed ilp genes (ilps 4, 6, and 7) after RNAi knockdown of FoxO before blood feeding (SI Appendix, Fig. S9A). GATA-binding was present at all ilp genes after blood meal at 12-h PBM, but FoxO overexpression diminished the binding enrichment at the GATAr-repressed ilp genes (ilps 4, 6, and 7) (SI Appendix, Fig. S9B). In contrast, no change was observed at the GATAa-activated ilp genes (ilps 8-1-3, 2, and 5) at 12-h PBM or 24-h PBM after FoxO overexpression when compared with the control UAS-FoxO female mosquitoes (SI Appendix, Fig. S9 B and C). Thus, the FoxO action is essential for preventing GATAr binding to ilps 4, 6, and 7 and permitting these ilps to be expressed at the PE phase.
To gain further confirmation of the involvement of FoxO in repression of GATAr-ilp binding, we constructed the pAc-FoxO-Flag expression plasmid. The luciferase activities were recovered when S2 cells were cotransfected with the pAc-FoxO-Flag, pAc-GATAr-Myc, and the ilp4-Fluc, ilp6-Fluc, or ilp7-Fluc vectors and compared with the control group transfected with only pAc-GATAr-Myc and the ilp4-Fluc, ilp6-Fluc, or ilp7-Fluc vectors. This experiment demonstrated that FoxO blocked GATAr binding to these ilps (Fig. 8).

Discussion
Anautogeny is a fundamental phenomenon underlying the vectorial capacity of mosquitoes (40). The gonadotrophic cycle of anautogenous female mosquitoes is separated into two phases: PE and PBM. During this cycle, the female mosquito shifts its feeding from nectar during PE to vertebrate blood in PBM to coordinate the nutrient intake and utilization with the reproductive needs for egg production (10,11). A blood meal is digested and absorbed by mosquito females to provide nutrients as well as signaling AAs for activating the TOR pathway (29). TOR mediates AA signaling in mosquito anautogeny as its RNAi knockdown is significantly impacting mosquito egg development (41). Blood feeding activates the brain to release ovary ecdysteroidogenic hormone (OEH) and ILP3 for initiating reproductive events, and ILP3 but not OEH stimulates digestion of the blood meal (42). ILP signaling plays a critical role in regulating multiple physiological processes such as growth, metabolism, reproduction, and lifespan in animals (16,43). Our study focused on the control of ILPs due to their importance as regulators linking metabolism and reproduction in the disease vector A. aegypti.
Here, we investigated the effect of the AA-TOR signaling on ILPs during the female mosquito reproductive cycle. ilp 4, 6, and 7 are up-regulated during the PE phase and down-regulated after blood feeding, while ilps 1, 2, 3, 5, and 8 exhibited high levels of expression during the PBM phase (SI Appendix, Fig. S1) (31). Infusion of a balanced AA mixture into PE mosquitoes has been reported to be sufficient to activate egg development (32). Our results indicate that the AA-TOR signaling pathway regulates eight Aedes ILPs differentially in reproducing female mosquitoes. We show that infusion of the low level of AA down-regulated the expression of ilp 4, 6, and 7 genes, but this AA level was not sufficient to up-regulate the expression of other ilps (1, 2, 3, 5, and 8). The activation of the latter group of the ilp genes required higher levels of AAs. RNAi knockdown of TOR caused the opposite effect up-regulating ilps 4, 6, and 7 and down-regulating ilps 1, 2, 3, 5, and 8. This indicated that TOR mediates the AA signaling.
Using the CRIPSR-Cas9 approach, we successfully obtained mosquito knockout mutants for the GATA isoforms GATAr and GATAa. This study has revealed that the GATA isoforms GATAr and GATAa differentially affect the expression of ilp genes. GATAr and GATAa factors directly interact with promoters of the ilp genes; GATAr specifically binds ilps 4, 6, and 7, while GATAa -ilps 1, 2, 3, 5, and 8. Our previous studies have shown that CRISPR-Cas9-mediated mutation on different ilp genes causes a differential degree of deficiency in growth, nutrient storage, carbohydrate/lipid balance and egg development (18,19). Here, we demonstrate that ilp4, ilp6, and ilp7 are negatively regulated by the AA-TOR signaling through GATAr factor, where GATAr inhibits the expression of these genes at the low levels of AAs. In contrast, the GATAa factor activated the expression of ilp1, ilp2, ilp3, ilp5, and ilp8 genes by directly interacting with their promoters at the high influx of AAs during the PBM phase. Depletion of GATAr by the isoform-specific CRISPR-Cas9 resulted in the formation of small ovarian follicles and little lipid accumulation. The situation was reversed in the isoform-specific GATAa-depleted mosquitoes. Mosquito reproductive defects with enlarged ovarian follicles and excessive lipid storage were observed in the GATAa CRISPR-Cas9 mutants. Furthermore, we found that GATAr and GATAa affected the sugar reserves. Therefore, the TOR-GATA axis plays a key role in regulating adaptive physiological responses to nutrient availability by controlling ILPs in mosquitoes, and the GATA isoform-specific switch differentially mediates the AA-TOR signaling to the ilp genes supporting ovarian development and energy reserve homeostasis.
The secretion of brain ILPs is regulated by several TOR-dependent factors (44)(45)(46). We determined whether the levels of circulating ILPs were regulated by the AA-TOR-GATA pathway in female mosquitoes. We generated CRISPR-Cas9 epitope-tagged ILPs in A. aegypti female mosquitoes to analyze the abundance of each ILP protein in hemolymph. Our experiments demonstrated that indeed the levels of circulating ILPs in mosquitoes are tightly controlled by AAs, TOR, and GATA factors.
The FoxO TF activity correlates with its intracellular localization, which is controlled by the insulin signaling (47,48). In the fat body of female adult mosquitoes, we found that FoxO was in the nuclei at 72-h PE and in the cytoplasm at 24-h PBM, when the AA level was high. This demonstrates that FoxO was active during the PE phase and inactive during the PBM one. CRISPR-Cas9 disruption of GATAr-isoform caused FoxO retardation in the cytoplasm before blood feeding at 72-h PE GATAa on FoxO. GATAa-isoform disruption resulted in FoxO nuclear localization after a blood meal at 24-h PBM. RNAi TOR knockdown caused FoxO nuclear localization like that of GATAa. Thus, GATAr and GATAa relay the AA-TOR signaling in opposite ways affecting the FoxO intracellular localization and, therefore, its activity. We showed that the knockout of GATAr isoform increases the expression of FoxO-repressed metabolic enzyme genes, while the knockout of GATAa decreases these genes. However, no direct interaction of GATAa or GATAr with these metabolic genes was detected.
Our experiments further revealed the critical role of FoxO in regulating GATAr binding. We show that when FoxO is active during the PE phase, it blocks binding of GATAr to the promoters of ilp 3, 4, and 7, thereby preventing their inhibition and permitting expression. The influx of AAs, even at the low dose, deactivate FoxO, and GATAr blocks the expression of these ilp genes. The high AA level triggers activation of ilps 1, 2, 3, 5, and 8. This AA trigger is mediated by TOR and GATAa. The diagram summarizing the model of interactions of these factors during mosquito reproduction is presented in Fig. 9.
In conclusion, we have identified the TOR-GATA switch mechanism for the AA signaling control on the ilp gene expression and circulating ILPs, coordinating nutrient metabolism during reproductive cycles of the disease vector A. aegypti.

Materials and Methods
A detailed description of the materials and methods used for this study is provided in (SI Appendix, Supplemental Materials and Methods). A. Aegypti mosquitoes and their genetics were used. RNAi, CRISPR-Cas9 knockin, gene-tagging, isoformspecific CRISPR-Cas9 knockout, piggyBac transformation, AA infusion, ChIP, cell luciferase, ELISA, and immunofluorescence were performed. dsRNA injection, mosquito embryo microinjection, and cell culture were described previously.  Fig. 9. Schematic diagram of the AA-TOR-GATA axis regulating expression of ilp genes. FoxO is active during the PE phase, it blocks binding of GATAr to the promoters of ilps 3, 4, and 7, thereby preventing their inhibition and permitting expression. The influx of AAs, even at the low dose, deactivate FoxO, and GATAr blocks the expression of these ilp genes. The high AA level triggers activation of ilps 1, 2, 3, 5, and 8. This AA trigger is mediated by TOR and GATAa.