CRISPR-Cas9 Genome Editing Uncovers the Mode of Action of Methoprene in the Yellow Fever Mosquito, Aedes aegypti

Abstract

Methoprene, a juvenile hormone (JH) analog, is widely used for insect control, but its mode of action is not known. To study methoprene action in the yellow fever mosquito, Aedes aegypti, the E93 (ecdysone-induced transcription factor) was knocked out using the CRISPR-Cas9 system. The E93 mutant pupae retained larval tissues similar to methoprene-treated insects. These insects completed pupal ecdysis and died as pupa. In addition, the expression of transcription factors, broad complex and Krüppel homolog 1 (Kr-h1), increased and that of programmed cell death (PCD) and autophagy genes decreased in E93 mutants. These data suggest that methoprene functions through JH receptor, methoprene-tolerant, and induces the expression of Kr-h1, which suppresses the expression of E93, resulting in a block in PCD and autophagy of larval tissues. Failure in the elimination of larval tissues and the formation of adult structures results in their death. These results answered long-standing questions on the mode of action of methoprene.

Introduction

The yellow fever mosquito, Aedes aegypti, is a vector of numerous viruses that cause diseases such as zika, dengue, chikungunya, and yellow fever, which together puts more than 40% of the world's population at risk of exposure.^1–3 Insect development is regulated by several hormones, including ecdysteroids (the most active form being 20-hydroxyecdysone, 20E) and juvenile hormones (JH).^4–6 While 20E initiates and coordinates molting and metamorphosis, JH maintains juvenile status by preventing metamorphosis.⁷ Thus, insects commit to metamorphosis when JH titers decline.⁸ Many analogs of ecdysone and juvenile hormone (JHAs) are being used to control a variety of insect pests.^9,10 Since adult mosquitoes transmit diseases, insect growth regulators that mimic JH (JH analogs, JHA) are widely used to prevent their development from larva to adult stages. The biological actions of JHA are well studied, but the molecular mode of action of these compounds is not well understood.

The epidermis of lepidopterans makes larval, pupal, and adult cuticles sequentially, and exogenously applied JH can prevent these changes.¹¹ However, the pupal epidermis (except for the abdominal region) of Drosophila melanogaster and other higher insects is derived from imaginal discs, and the application of JH cannot prevent metamorphosis even when given throughout the larval life.¹² However, JH application during the final larval instar or prepupal period of D. melanogaster prevents normal differentiation of abdominal epidermis, whose cells arise from the proliferation of histoblasts.¹² In mosquitoes, a continuous population of cells in the abdomen produces the cuticles of all three stages. The epidermal cells are polyploid in larval stages, and reductional divisions occur in the pupa.¹³ Therefore, the production of larval, pupal, and adult cuticles and their hormonal regulation are different among insects.

Studies in our laboratory showed that methoprene action in Heliothis virescens (Lepidoptera) and Ae. aegypti (Diptera) is different. In H. virescens, the application of methoprene to the final instar larvae affected midgut remodeling, production of pupal cuticle, and larval–pupal ecdysis.¹⁴ In contrast, in Ae. aegypti, methoprene did not interfere with the production of pupal cuticle and pupal ecdysis, but it did interfere with midgut remodeling.¹⁵ In Ae. aegypti, the leg imaginal disc growth and differentiation, respiratory trumpet morphogenesis, and midgut remodeling are initiated on the first day of the last larval stage and are continued throughout this stage.¹⁶ During this time, Ae. aegypti larvae feed and grow, doubling in size. Thus, in Ae. aegypti, larval growth occurs alongside metamorphic differentiation.

This is different from higher dipterans such as D. melanogaster, where larval growth is separated from metamorphosis. In D. melanogaster, imaginal discs grow during the first, second, and early third instar larval stages. Before puparium formation, third instar larvae stop feeding. Most of the differentiation in imaginal discs¹⁷ and proliferation of abdominal histoblasts¹⁸ and degeneration of larval tissues^19,20 occurs after puparium formation. These differences in Ae. aegypti and D. melanogaster are most likely due to the differences in the timing of synthesis and release of hormones and their effects on gene expression during larval–pupal–adult metamorphosis. Therefore, hormone-regulated development is complex and variable among insects. Understanding the molecular action of developmental hormones may help us to improve the efficiency of insect growth regulators, such as JHAs.

Ecdysone-induced helix-turn-helix transcription factor, E93 discovered in D. melanogaster, contributes to stage-specific ecdysone regulation of adult tissue development.²¹ In this insect, the E93 was shown to be required for the development of adult appendages.^22–25 E93 mutants fail to undergo programmed cell death (PCD) in larval salivary glands, and the expression of E93 facilitated PCD in larval salivary glands of E93 mutant pupae.²⁶

Studies in Blattella germanica, Tribolium castaneum, Bombyx mori, and other insects showed that E93 suppresses the expression of transcription factors, Krüppel homolog 1 (Kr-h1 and BR-C), that play key roles in larval and pupal development.²⁷ Interestingly, in T. castaneum, upregulation of E93 by knocking down Kr-h1 triggered premature metamorphosis independent of threshold size and revealed that Kr-h1 repressed the expression of E93 during later larval stages and prevented metamorphosis.²⁸ Knockdown of E93 in B. mori²⁷ or its knockout in Spodoptera frugiperda²⁹ caused an arrest in development during larval–pupal metamorphosis. These studies demonstrated critical roles for E93 in larval to adult transition.

Methoprene-treated mosquito larvae become pupae and die during this stage. Confocal microscopy observations revealed that methoprene treatment resulted in the persistence of larval cells in tissues such as foregut, midgut, hindgut, salivary glands, gastric ceca (GC), and central nervous system (CNS), resulting in a block in their remodeling during larval–pupal metamorphosis. Reverse transcriptase–quantitative real-time polymerase chain reaction (RT-qPCR) analysis showed that the expression of Kr-h1 and broad complex (BR-C) increased and the expression of E93 decreased in methoprene-treated insects. As mentioned above, E93 triggers PCD and regulates tissues remodeling during larval–pupal metamorphosis.

To characterize the link between methoprene and E93, a transgenic Cas9/sgRNA system^30,31 was used to knock out the E93 gene. The phenotypes of E93 mutants are similar to the phenotypes induced by methoprene treatment. Larval tissues persisted in pupae and the insects died during the pupal stage. The mRNA levels of Kr-h1 and BR-C increased, whereas the expression levels of PCD and autophagy genes decreased in E93 knockout mutants. These data suggest that methoprene represses the expression of E93 gene through the methoprene-tolerant-Krüppel homolog 1-E93 (MEKRE93) pathway³² in killing mosquitoes during the pupal stage.

Materials and Methods

Insect rearing and microinjection

The Ae. aegypti AAEL010097-Cas9 and Liverpool IB12 (LVPIB12) strains were maintained at 27°C ± 2°C with a photoperiod of 14 h light:10 h dark and 60–70% relative humidity. The mosquito larvae were given bovine liver powder and maintained in water at 26°C. Pupae were sexed by size and kept in separate containers. After emergence, the adults were fed on 10% sucrose solution. For egg collection, around 200 females and 200 males were kept in each cage for mating. After 4 days, the females were artificially fed with defibrinated sheep blood to induce oviposition. The mosquito embryonic microinjections were performed using methods described previously.³³ Briefly, the mosquito eggs were collected within 20 min after oviposition, aligned on a glass slide, then microinjected. After injection, the eggs were kept at 26°C. On the third day after injection, they were dropped into water for hatching. The hatched larvae were maintained using the same protocol described above.

Hormone treatment

Methoprene [isopropyl (E,E)-(RS)-11-methoxy-3,7,11-trimethyldodeca-2,4-dienoate] and pyriproxyfen (4-phenoxyphenyl (R/S)-2-(2-pyridyloxy) propyl ether 2-[1-(4-phenoxyphenoxy) propan-2-yloxy] pyridine) were gifts from Wellmark International (Dallas, TX). Technical grade methoprene or pyriproxyfen was dissolved in dimethyl sulfoxide (DMSO) at 1000 × concentration and added to the diet. Technical grade stable ecdysone agonist, RH-102240 [2-ethyl-3-methoxybenzoic acid 2-(3,5-dimethylbenzoyl)-2-(1,1-dimethylethyl)hydrazide] was dissolved in ethanol. Preliminary dose–response experiments were conducted, and the minimum dose that causes 90% effect was selected for further studies. To screen the morphological change under the fluorescence light, the hr5ie1-egpf-U6-E93-sgRNA transgenic line was used to do the treatment.

U6-sgRNA strain development

The CRISPR optimal target finder³⁴ was used to design sgRNA target sites. The potential off-target efficiency of the target site was evaluated by screening Ae. aegypti genome sequences (AaegL5.0 GCA_002204515.1).³⁴ The U6-sgRNA construction was processed according to the protocol described previously.³⁶ The Gibson Assembly (New England Biolabs [NEB], MA) method was used for U6-sgRNA preparing constructs. Briefly, the 5′ untranslated regions (UTR) and 3′ UTR of the U6 promoter sequence were amplified using the target-specific primers (Supplementary Table S1).

Then, the pBac-hr5ie1-egfp vector³⁵ was linearized by using the SacII (NEB) and, the 5′ UTR and 3′ UTR of the U6 promoter were cloned into this vector to get pBac-hr5ie1-egfp-u6 vector. Subsequently, the NcoI (NEB) was used to linearize the pBac-hr5ie1-egfp-u6 vector. The sgRNA fragments were amplified using the target-specific primers (Supplementary Table S1) and cloned into the linearized pBac-hr5ie1-egfp-u6 vector using the Gibson Assembly method. The final plasmid DNA was extracted using the Plasmid Plus Midi Kit (Qiagen) following the manufacturer's instructions. The transposase mRNA was prepared using the mMESSAGE mMACHINE T7 Transcription Kit (Thermo Fisher, MA) using the XhoI (NEB) linearized IFP2 vector.³⁵

The Liverpool IB12 strain was used for genetic transformation. The U6-sgRNA constructs were microinjected into fresh embryos. After injection, the hatched larvae were maintained and sexed during the pupal stage. After emergence, the adults were mated with wild-type adults of the opposite sex to obtain the G1. Then, the G1 offspring were screened under the green fluorescent protein (GFP) filter to obtain positive transgenic insects.

RNA isolation and cDNA synthesis

To perform the differential gene expression of wild type, methoprene-treated insects, and E93 mutants, the animals were collected, and total RNA was extracted by using TRIzol. The RNA was converted to cDNA and used in RT-qPCR to quantify relative mRNA levels. 40s ribosomal protein S7 (RpS7) (AAEL009496) was used as a reference gene. The cDNA synthesis and RT-qPCR were performed following the methods described in our publication.⁶

Mutant screen and analysis

To knockout the target gene, the U6-sgRNA lines were mated with AAEL010097-Cas9 line³⁰ to obtain the G1. The G1 offspring were screened using the GFP and DsRed fluorescence signals. Only larvae that were positive for green and red fluorescence were selected and maintained in separate containers. The phenotypes were recorded during the pupal stage. The genomic DNA isolation was performed using DNeasy Blood & Tissue Kit (Qiagen) following the manufacturer's instructions. The PCR product purification was performed using QIAquick PCR Purification Kit (Qiagen) following the manufacturer's instructions. The mutagenesis was confirmed by T7 endonuclease I (T7E1) assay and TA clone sequencing, as previously described.³⁶

Briefly, for T7E1 assay, 200 ng of PCR products was hybridized in NEB buffer 2 under the following conditions: 95°C for 5 min, 95–85°C at −2°C/s, 85–25°C at −1°C/s, and held at 4°C. Then, the mixtures were digested by 10 units of T7E1 enzyme (NEB) in 37°C for 15 min. Immediately, the reaction solution was checked on 2% agarose gels containing GelRed (Biotium, CA), and the mutation events indicated by DNA bands were captured under ultraviolet. The PCR products were cloned into TOPO pCR4 vector (Thermo Fisher) and sequenced by the Sanger method.

Histology

The tissues from larvae and pupae that developed from untreated control and methoprene-treated insects were dissected in 1 × phosphate-buffered saline (PBS). Nuclear staining was carried out with DAPI (4′,6-diamidino-2 phenylindole; Sigma) at 1:1000 dilution of 1 mg/mL stock for 10 min. After nuclear staining, the tissues were washed in 1 × PBS twice and photographed under a dissection microscope (Zeiss; Diagnostic Instruments) equipped with AxioCam MRC 5 using AxioVision 4.0 software. For cross sections, the midguts were dissected in 1 × PBS and fixed in 4% paraformaldehyde overnight at 4°C. After fixation and washing in 1 × PBS twice, the midguts were dehydrated through a series of grades of ethanol (25%, 50%, 75%, 95%, and 100% in 1 × PBS), infiltrated through a series of grades of Xylene/CitriSolv (Fisher) (25%, 50%, 75%, and 100% in ethanol), and finally embedded in three successive baths of Paraplast Plus (56°C; Tyco Healthcare).

Paraffin sections were cut (10 μm thick, room temperature) using a microtome (Leica RM 2135, Germany), mounted on poly-l-lysine-coated slides and dried at 42°C overnight. The slides were stored at 4°C in a dry atmosphere until used for the experiments. The sections were deparaffinized through successive baths of xylene, rehydrated through serial grades of ethanol, water, and 1 × PBS. After nuclear staining with DAPI, the slides were washed with 1 × PBS twice and mounted with 50% glycerol. The sections were observed under Leica TCS-SP laser scanning confocal microscope and documented using Leica TCS software. For confocal screen, the larval and pupal midguts of wild type and E93 mutant were dissected in 1 × PBS and fixed in 4% (vol/vol) paraformaldehyde and rinsed with PBS-T (0.3% [vol/vol] Triton X-100 in PBS), followed by DAPI staining and photographed under a Leica SP5 confocal microscope with excitation filters at 488 and 358 nm to detect green and blue, respectively.

Results and Discussion

Larval tissues persist in pupae that developed from methoprene-treated Ae. aegypti larvae

Irrespective of the time of methoprene application during the larval stage, more than 90% of the treated Ae. aegypti larvae become pupae and die during the pupal stage.¹⁵ To confirm these results, newly molted third instar Ae. aegypti larvae were treated with 100 ng/mL JHAs, methoprene or pyriproxyfen. More than 85% of methoprene- or pyriproxyfen-treated larvae became pupae and died during the pupal stage. In contrast, 100% of control larvae treated with DMSO successfully completed larval and pupal stages and became adults.

To determine whether the Ae. aegypti pupae that developed from methoprene-treated larvae contain larval cells in tissues such as salivary glands, midgut, foregut, hindgut, and CNS, we examined the morphology of these tissues in staged pupae that developed from methoprene-treated larvae or control larvae treated with DMSO. As shown in Figure 1A, the alimentary canal dissected from the pupae that developed from methoprene-treated larvae is similar to that in control larvae at 24 h after ecdysis into the final instar larval stage. The midgut from pupae that developed from methoprene-treated larvae contained larval midgut cells with larger nuclei, whereas the midgut from the untreated 12 h-old pupae showed only pupal midgut cells with smaller nuclei (Fig. 1B).

FIG. 1.

Methoprene prevents tissue remodeling during larval–pupal metamorphosis in Aedes aegypti. (A) The pupae that developed from methoprene-treated larvae contain larval tissues. The alimentary canal dissected from 24-h-old wild-type L4 larva (WT_L4_24h), 12-h-old wild-type pupa (WT_pupa_12h), and 12-h-old pupa developed from methoprene-treated larva (Meth_pupa_12h). The photographs taken under bright light are shown in the top panel, and the photographs taken under the green fluorescent light are in the bottom panel. The scale bar is 500 μm. Images show representative alimentary canal dissected from 10 to 15 larvae in each group. The experiment was repeated three times using independent biological materials. (B) The midguts dissected from WT_L4_24h, WT_pupa_12h, and Meth_pupa_12h were fixed and photographed under a confocal microscope. The larval midgut cells are indicated by white arrows. The scale bar is 200 μm. The photographs show representative midgut images from four larvae for each treatment. The experiment was repeated twice with independent biological materials. (C) The foreguts dissected from WT_L4_24h, WT_pupa_12h, and Meth_pupa_12h were fixed and photographed under a confocal microscope. The scale bar is 200 μm. The photographs show representative foregut images from four larvae for each treatment. The experiment was repeated twice with independent biological materials. (D) The larval salivary glands of WT_L4_24h and Meth_pupa_12h were fixed and photographed under a confocal microscope. The scale bar is 50 μm. The photographs show representative salivary gland images from four larvae for each treatment. The experiment was repeated twice with independent biological materials. (E) The hindguts dissected from WT_L4_24h, WT_pupa_12h, and Meth_pupa_12h were fixed and photographed under a confocal microscope. The larval hindgut cell nuclei are indicated by white arrows. The scale bar is 200 μm. The photographs show representative hindgut images from four larvae for each treatment. The experiment was repeated twice with independent biological materials. (F) The CNS of WT_L4_24h, WT_pupa_12h, and Meth_pupa_12h were fixed. The photographs taken under bright light are shown in the top panel, and the photographs taken under the green fluorescent light are in the bottom panel. The numbers 1–6 or 1–8 indicate the numbers of the abdominal ganglion. The scale bar is 500 μm. Images show representative nervous system dissected from 10 to 15 larvae in each group. The experiment was repeated three times using independent biological materials. (G) The differential expression of genes in methoprene-treated insects is shown as a heat map. The heat map was generated using Morpheus (https://software. broadinstitute.org/morpheus). The mRNA levels were measured using four replicates per treatment. The color code indicates the log2-fold change in mRNA levels. The names of the genes are shown on the right side. The scale bar on the right side shows the relative color code for the log2-fold change in gene expression. CNS, central nervous systems; GC, gastric ceca; HG, hindgut; MT, Malpighian tubules; SG, salivary glands; TG, thoracic ganglion. Color images are available online.

The foregut of larvae is composed of the pharynx, salivary glands, and esophagus. The foregut is remodeled during metamorphosis. Larval salivary glands die, and adult salivary glands develop during the pupal stage before adult emergence. The foregut epithelium is remodeled into two sucking pumps: the cibarial pump and pharyngeal pump. The dorsal diverticulum, a characteristic of the adult foregut, develops beginning at 10 h after larval–pupal ecdysis.³⁷ The GC present at the junction of foregut and midgut degenerate at the end of the larval stage. The foregut dissected from pupae that developed from methoprene-treated larvae is similar to the larval foregut. Both salivary glands and GC are present in these foreguts (Fig. 1C, D). In contrast, both salivary glands and GC are absent in the foregut from pupae that developed from untreated larvae (Fig. 1C, D). These data show that methoprene prevents remodeling of the foregut and degeneration of larval salivary glands and GC.

The larval hindgut also undergoes remodeling during metamorphosis.^37,38 The hindguts dissected from control larvae at 24 h after ecdysis into the final larval stage or from the pupae that developed from methoprene-treated larvae showed similar morphology. They differed from the hindguts dissected from pupae that developed from control larvae (Fig. 1A). The pylorus seen in the larval hindgut is round and bulbous compared with the elongated tubular shape seen in the pupal hindgut (Fig. 1A). Under the confocal microscope, the hindgut from pupae that developed from methoprene-treated larvae showed similar size nuclei (large polyploid nuclei) to that in control larvae at 24 h after ecdysis to fourth instar larvae (L4). The hindgut in control pupae at 12 h after pupation showed smaller diploid nuclei compared with those found in pupal hindgut or hindgut from pupae that developed from methoprene-treated larvae (Fig. 1E). Malpighian tubules do not undergo remodeling, and hence, methoprene did not affect them (Fig. 1A).

The CNS also undergoes remodeling during metamorphosis.^37,38 The optic lobes grow and are seen as well-defined structures in the pupal brain compared with those in the larval brain. Soon after the ecdysis into the pupal stage, the subesophageal ganglion moves forward and becomes more closely associated with the brain, the thoracic ganglion, and the first abdominal ganglion fuse and form a single ganglion. The seventh and eighth abdominal ganglions also fuse to form a single ganglion. As shown in Figure 1F, the CNS dissected from pupae that developed from untreated larvae showed a remodeled pupal/adult CNS.

In contrast, the CNS dissected from untreated larvae at 24 h after ecdysis into the final larval stage, or the CNS dissected from pupae that developed from methoprene-treated larvae showed larval characteristics (i.e., no growth in optic lobes and fusion of ganglia). These data showed that methoprene prevents the remodeling of the CNS. We also dissected pupae from pyriproxyfen-treated larvae and found that the remodeling of salivary glands, CNS, and the alimentary canal was blocked in these insects (Supplementary Fig. S1).

Unlike in most other insects where methoprene blocks larval–pupal metamorphosis, methoprene-treated Ae. aegypti larvae undergo larval–pupal metamorphosis and die during the pupal stage. Previous studies showed that methoprene blocks the death of midgut larval cells but not the proliferation and differentiation of stem cells to pupal/adult midgut epithelium in mosquitoes.³⁹ As a result, the pupae developed from methoprene-treated larvae contain two layers of the midgut epithelium. Other tissues such as the foregut, hindgut, salivary glands, and nervous system undergo remodeling during metamorphosis to accommodate the changes in feeding and behavior between larval and adult stages of Ae. aegypti.^37,38

The data included here showed that methoprene blocks the degeneration of larval cells in salivary glands, GC, and remodeling of the foregut, hindgut, and nervous system. The cumulative effect of problems associated with remodeling of important organs such as the alimentary canal and CNS results in a block in the development during pupal–adult transformation and eventual death. A previous study showed that E93 is required for remodeling of midgut during larval–pupal metamorphosis.²⁶ To determine if methoprene mediates its action through E93, we employed the CRISPR-Cas9 method to knockout E93 in Ae. aegypti.

E93 mutants die during the pupal stage

Genome editing has been successfully performed in Ae. aegypti by injecting sgRNA and Cas9 into embryos.^{30,31,40–46} However, homozygous lines need to be produced for the functional characterization of target genes, which is time-consuming and laborious. Previously, we showed that the use of multiple sgRNAs induced loss-of-function phenotypes of JH receptor in Ae. aegypti in G0 allowing the collection of information of gene function in the first generation.³⁶ Since the E93 mutants did not show any phenotypes until they die during the pupal stage, it was challenging to characterize mutants using the multiple sgRNA method. The transgenic CRISPR-Cas9 system was used to knockout E93 gene.

A sgRNA (5′-GGACATCCTGAATGGGAAACTGG-3′) targeting exon 7 of E93 gene with no predicted off-target effects was used to generate a U6-E93-sgRNA construct (Fig. 2A, B). Six hundred Ae. aegypti Liverpool IB12 (LVP) strain eggs were injected with a mixture of 200 ng/μL U6-E93-sgRNA construct and 300 ng/μL piggyBack transposase mRNA. The G0 larvae (∼150) that hatched from injected eggs were maintained until the adult stage and mated with wild-type LVP adults of the opposite sex. The G1 progeny were screened, and 30 larvae with enhanced green fluorescent protein (EGFP) fluorescence were identified. These larvae reared to the adult stage and mated with LVP adults of the opposite sex. Two hundred EGFP G2-positive larvae were identified. The transgenic U6-E93-sgRNA stain was maintained through mating with the opposite sex LVP adults followed by a selection of EGFP positive larvae.

FIG. 2.

Knockout of E93 gene using the transgenic CRISPR-Cas9 system in Aedes aegypti. (A) Diagram of Ae. aegypti AAEL004572 (E93) gene. The black boxes indicate the exons. The sgRNA target site sequences located in exon 7 are shown at the bottom. (B) The components of U6-E93-sgRNA construct used to produce transgenic U6-E93-sgRNA Ae. aegypti strain. (C) Crossing female AAEL010097-Cas9 with male U6-E93-sgRNA produced offspring expressing both EGFP and DsRed, suggesting the inheritance of both Cas9 and U6-E93-sgRNA constructs. The photographs of pupae that developed from the progeny of female AAEL010097-Cas9 with male U6-E93-sgRNA crosses taken under bright light, green, and red fluorescence are shown. The scale bar is 1000 μm. (D) T7E1 assay was performed to detect mutations in E93 gene. The top gel image shows the PCR products, and the black arrow pointing to the DNA bands is on the right. The bottom gel image shows the T7E1 assay products, the two black arrows on the right point to the expected size of T7E1 digested products. 1–9, nine mutants of E93 gene. L, 1 kb plus DNA ladder. The numbers 100 and 200 on the left indicate 100 and 200 bp fragments in DNA ladder. (E) The mutations were confirmed by sequencing. Of the eight clones sequenced, five of them showed mutations. The wild-type sequences are shown at the top, with the target site marked by red color letters. In mutant sequences, deletions are shown as dashes and insertions as yellow lowercase letters. The net change in length is marked on the right of each sequence (+, insertion; −, deletion). The bottom allele was identified three times from the sequencing results as indicated by × 3. DsRed, red fluorescent protein; EGFP, enhanced green fluorescent protein; PCR, polymerase chain reaction; T7E1, T7 endonuclease I. Color images are available online.

The U6-E93-sgRNA-positive G2 adults were mated with the opposite sex AAEL010097-Cas9 adults³⁰ as described previously.^31,47 The G3 larvae were checked for the presence of EGFP and DsRed during the third instar larval stage, and the larvae that showed EGFP and DsRed fluorescence were selected for observation. Most of these larvae developed to the pupal stage and died (Fig. 2C). To confirm the mutations in E93 gene, the genomic DNA from progeny showing both EGFP and DsRed fluorescence was extracted, and the T7E1 assay was performed. As shown in Figure 2D, all nine randomly selected insects showed mutations in the E93 gene. Of the eight clones sequenced, six of them showed mutations, including different InDels (Fig. 2E).

To reconfirm the specificity of phenotypes produced by E93 mutation, we designed four sgRNAs targeting Exon 3 of E93 (Supplementary Fig. S3A) and injected them into AAEL010097-Cas9 eggs. About 30% of the hatched larvae displayed similar E93 mutant phenotypes detected in progeny of crosses between transgenic sgRNA and Cas9 lines described above (Supplementary Fig. S3B). The mutations were confirmed by sequencing; long fragments deletions and InDels were detected adjacent to the target sites (Supplementary Fig. S3C). These data reconfirmed E93 mutant phenotypes detected in transgenic CRISPR-Cas9 experiments, and therefore, we used this method to produce E93 mutants for the rest of the experiments.

E93 mutant pupae contain larval tissues

Interestingly, we found that the phenotypes detected in E93 mutants are similar to those observed in methoprene-treated insects. Since the tissue remodeling is completed by 12 h after ecdysis into pupal stage, we dissected tissues from 12 h-old E93 mutant pupae (E93-12h-P), 12 h-old wild-type pupae (WT-12h-P), and 48 h-old wild-type L4 larvae (WT-48h-L4) and compared their morphology. As shown in Figure 3A, the E93-12h-P still contains larval salivary glands that are similar to WT-48h-L4. But the larval salivary glands degenerated in the WT-12h-P. Unlike in D. melanogaster,⁴⁸ E93 is required for the degeneration of larval salivary glands in Ae. aegypti.

FIG. 3.

E93 mutant pupae contain larval tissues. (A) The photographs of alimentary canals dissected from 48-h-old wild-type L4 larva (WT_L4_48h), 12-h-old wild-type pupa (WT_pupa_12h), and 12-h-old E93 mutant pupa (E93-M_pupa_12h) are shown. The photographs taken under bright light are shown in the top panel, and the photographs taken under the green fluorescent light are in the bottom panel. The scale bar is 500 μm. Images show representative alimentary canal dissected from 10 to 15 larvae in each group. The experiment was repeated five times using independent biological materials. (B) The photographs of salivary glands dissected from WT_L4_48h and E93-M_pupa_12h taken under a confocal microscope. The scale bar is 50 μm. The photographs show representative midgut images from four larvae for each treatment. The experiment was repeated twice with independent biological materials. (C) The photographs of the foreguts of WT_L4_48h and E93-M_pupa_12h taken under a confocal microscope. The scale bars 200 μm. The photographs show representative midgut images from four larvae for each treatment. The experiment was repeated twice with independent biological materials. (D) The photographs of the hindguts of WT_L4_48h and E93-M_pupa_12h taken under a confocal microscope. The white arrows indicated the larval hindgut cell nucleus. The scale bar is 200 μm. The photographs show representative midgut images from four larvae for each treatment. The experiment was repeated twice with independent biological materials. (E) The photographs of the cross section of midgut from wild-type and E93 mutant pupae. Nuclear staining by DAPI revealed nuclei of different sizes. The white arrows indicate the pupal cells in the wild type; the red arrows indicate the larval cells E93 mutant pupa. The scale bar, 50 μm; 10–15 larvae in each group. The experiment was repeated five times using independent biological materials. (F) The photographs of the CNS dissected from WT_L4_48h, WT_pupa_12h, and E93_pupa_12h. The photographs taken under bright light are shown in the top panel, and the photographs taken under the green fluorescent light are in the bottom panel. The numbers 1–6 or 1–8 indicate the numbers of abdominal ganglion. The scale bar, 500 μm. 10–15 larvae in each group. The experiment was repeated five times using independent biological materials. DAPI, 4′,6-diamidino-2 phenyl indole; E93-M, E93 mutant; TG, thoracic ganglion. Color images are available online.

Also, the Malpighian tubules that did not undergo remodeling are similar in three groups of insects compared (Fig. 3A). Under a confocal microscope, the nuclei of salivary glands from E93 mutant pupae and WT-48h-L4 are larger and similar in size (Fig. 3B). In WT-12h-P, the larval GC completely disappeared, but the E93-12h-P still had GC that are similar to those in WT-48h-L4 (Fig. 3A). The foregut in E93-12h-P also displayed a similar structure to that in WT-48h-L4 (Fig. 3C). In addition, the E93-12h-P and WT-12h-L4 contained similar hindgut, which is different from WT-12h-P (Fig. 3A). The nuclei in the hindgut of the E93-12h-P and WT-48h-L4 showed similar morphology (Fig. 3D).

Cross sections of midgut dissected from E93 mutant pupae showed both larval cells with larger nuclei and pupal/adult cells with smaller nuclei (Fig. 3E). In contrast, the sections of midgut from wild-type larvae showed larval cells, and those from wild-type pupae showed pupal/adult cells. Interestingly, we also found that the CNS of E93-12h-P is similar to that of WT-48h-L4 containing eight abdominal ganglia (Fig. 3F). In comparison, the CNS in wild-type pupae had six abdominal ganglia. The external morphology of E93 mutant pupae or the pupae that developed from methoprene-treated larvae is similar to that of wild-type pupae (Supplementary Fig. S4). The 48 h-old E93 mutant pupae showed adult structures such as eyes, antennae, and legs.

Both methoprene treatment and E93 knockout affect the expression of genes involved in ecdysone signaling, PCD, and autophagy

To understand the molecular mode of action of methoprene, we performed RT-qPCR to quantify mRNA levels of ecdysone receptors (EcR and USP isoforms), ecdysone-induced transcription factors (E93, BR, E74, E75 isoforms, and Hr3), and genes involved in PCD and autophagy in methoprene-treated larvae and pupae. RT-qPCR analysis showed that methoprene treatment suppressed the E93 expression during larval and pupal stages (Fig. 4A). Also, stable ecdysteroid analog induced E93 gene in larvae (Fig. 4B). The mRNA levels of Kr-h1 and four isoforms of broad complex, BR-Z1, BR-Z2, BR-Z3, and BR-Z4, were significantly higher in the fourth instar larvae (36 and 48 h old) and pupae (0 and 12 h old) that developed from methoprene-treated larvae when compared with their levels in control pupae and larvae exposed to DMSO (Fig. 4C and Supplementary Fig. S2).

FIG. 4.

Methoprene suppresses PCD and autophagy through E93. (A) Methoprene treatment suppressed the E93 expression during the fourth instar larval and pupal stages of Aedes aegypti. Liverpool strain third instar larvae were treated with 100 ng/mL methoprene. Total RNA from the fourth instar larvae (36 and 48 h old) and pupae (0 and 12 h old) was isolated and used to determine E93 mRNA levels using RT-qPCR. Mean ± SE (six replicates) are shown. One-way ANOVA Tukey's post-test was used to assess the significance of the changes in mRNA levels. Bars with the same letters are not significantly different at 95% CI. (B) Ecdysone agonist RH-102240 induces the expression of E93. Liverpool strain fourth instar larvae were treated with 100 ng/mL of SEA. E93 mRNA levels were determined by RT-qPCR. * Denotes the significant differences in the mRNA levels of target gene between control and treatment at p < 0.05 analyzed using the one-way ANOVA. (C) Heat map depicts the differential expression of genes involved in hormone signaling, apoptosis, and autophagy. Liverpool strain third instar larvae were treated with 100 ng/mL methoprene. Total RNA isolated from the fourth instar larvae (36 and 48 h old) and pupae (0 and 12 h old) was used to determine relative mRNA levels of select genes by RT-qPCR. The mRNA levels of the RPS7 were used for normalization. The average relative expression levels of six biological replicates were log2 transformed and used to generate the heat map. Details on the relative mRNA levels of each gene are provided in the Supplementary Figure S7. CI, confidence interval; PCD, programmed cell death; RPS7, ribosomal protein subunit 7; RT-qPCR, reverse transcriptase–quantitative real-time polymerase chain reaction; SE, standard error; SEA, stable ecdysone agonist. Color images are available online.

These data suggest that methoprene-induced BR-C is sufficient for larval–pupal metamorphosis in Ae. aegypti. Additionally, methoprene treatment also significantly reduced the expression levels of the key ecdysone response genes, E74, E75A, and Hr3. Notably, the genes involved in PCD—Ark, reaper, hid and initiator caspase 8 and executor caspase 3—were suppressed after methoprene treatment. Also, the autophagy genes—ATG2, ATG7A, ATG7B, ATG8, and ATG18—were suppressed by methoprene treatment (Fig. 4C). However, no effect was observed on the expression ecdysone receptor isoforms, EcR-A/EcR-B and USP-A/USP-B in methoprene treatment.

A previous study showed that E93 is required for the induction of apoptosis genes that are involved in remodeling midgut during larval–pupal metamorphosis.²⁶ Methoprene treatment suppressed the E93 suggesting that methoprene may mediate its action through E93. To test this hypothesis, we assessed the expression levels of key genes involved in ecdysone response, apoptosis, and autophagy in wild type and E93 mutants. After knocking out E93, the mRNA levels of Kr-h1 and four isoforms of broad complex, BR-Z1, BR-Z2, BR-Z3, and BR-Z4, significantly increased in E93 mutant larvae (Fig. 4C and Supplementary Figs. S5 and S7).

Furthermore, the higher expression levels of Kr-h1 and broad complex isoforms continued in E93 mutant pupae when compared with their levels in wild-type pupae where their expression levels decreased by 12 h after pupation. This result reiterates that for successful larval–pupal metamorphosis in Ae. aegypti, increase in E93 levels is not required, and induction of broad complex is sufficient. It is also possible that E93 perhaps suppresses the broad complex isoforms in Ae. aegypti (Fig. 5), which is similar to the situation in B. germanica.⁴⁹ Disruption of the E93 gene caused a decrease in the mRNA levels of early ecdysone response genes (E74 and E75A) and early late gene, Hr3. Ecdysone-induced E93 disruption by E93 knockout severely hampered ecdysone signaling and the mutants died during the pupal stage.

FIG. 5.

Proposed mode of action of methoprene through MEKRE93 pathway. The green box (left) shows the 20E action during the larval–pupal metamorphosis. The yellow box (middle) shows the methoprene action during larval–pupal metamorphosis. The blue box (right) shows the effect of E93 gene knockout. The red cross sign on E93 indicates that the E93 gene was knocked out. The process indicated in red font is blocked in methoprene-treated and E93 knockout insects. 20E, 20-hydroxyecdysone; BR-C, broad complex; E93, ecdysone-induced protein 93F; JH, juvenile hormone; JHA, juvenile hormone analog; Kr-h1, Krüppel homolog 1. Color images are available online.

As mentioned above, the PCD was blocked after knockout of E93 gene. Similar to methoprene treatment, E93 knockout also reduced the mRNA levels of the PCD genes, Ark, reaper, hid and initiator caspase 8 and executor caspase 3, in E93 mutants compared with their levels in wild-type insects (Fig. 4C and Supplementary Fig. S6). Ecdysone-regulated genetic hierarchy was shown to be involved in the PCD of larval tissues^50–53 and proliferation and differentiation of imaginal cells in D. melanogaster.^54,55 Decrease in JH titers and increase in 20E levels at the onset of metamorphosis induces E93 expression, which in turn upregulates genes involved in PCD leading to the degeneration of larval tissues. When Ae. aegypti larvae are exposed to methoprene, JH-induced transcription factor Kr-h1 expression is induced, which then suppresses the expression of E93, resulting in a block in PCD and persistence of larval tissues in pupae. When the E93 gene was knocked out using CRISPR-Cas9 method, the same phenotype of persistence of larval tissues in the pupae was detected.

In addition to apoptosis, autophagy also occurs during metamorphosis to remove larval midgut. The mRNA levels of the key autophagy genes, ATG2, ATG7A, ATG7B, ATG8 and ATG18, were significantly reduced in both methoprene treatment and E93 knockout mutants (Fig. 4C and Supplementary Fig. S7). A recent study showed that E93 governs autophagy-induced termination of vitellogenesis in the adult female mosquito, Ae. aegypti.⁵⁶ Our data suggest that suppression of E93 by methoprene or lack of functional E93 protein in E93 mutants affects both apoptosis and autophagy that result in the retention of larval cells in methoprene-treated insects or E93 knockouts.

Conclusions

In summary, the Ae. aegypti E93 mutants obtained by transgenic Cas9/sgRNA system showed the phenotypes such as the persistence of larval tissues in the pupa that are similar to those detected in the pupae that developed from methoprene-treated larvae. In addition, both E93 mutants and methoprene-treated larvae died during the pupal stage. Our studies showed that methoprene works through its receptor Met and induces Kr-h1 expression, which then suppresses E93 expression.⁵⁷ The lack of E93 decreases the expression of genes involved in the death of larval cells resulting in the persistence of larval tissues in pupae that developed from methoprene-treated larvae (Fig. 5).

Taken together, these data showed that MEKRE93 is the major molecular pathway methoprene uses to manifest its action. The data included in this study help in advancing our understanding of the molecular mode of action of methoprene. This knowledge could help in the management of resistance development against this insecticide and to identify more potent JHA for controlling insect pests and disease vectors.

Authors' Contributions

S.R.P. and G.-H.Z. designed the research. S.C.G., Y.J. and J.K. conducted the experiments and analyzed the data. G.-H.Z. and S.R.P. wrote the article.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by a grant from the National Institutes of Health (1R21AI163561-01) to Subba Reddy Palli.

Supplementary Material

Supplementary Table S1

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Figure S7