The animal model, Aedes aegypti (the Rockefeller strain) was maintained in the laboratory in a low-temperature illuminated incubator (model 818, Thermo Electron Corporation, Waltham, MA, USA) at 26°C and 80% humidity according to standard rearing procedures (Xiao et al., 2014). A. aegypti Aag2 cells were cultured at 28°C in Schneider's Drosophila medium for maintenance and microbial challenge. The cell media were supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, and 100 U/mL each of penicillin and streptomycin. The E. coli ST515 strain, which is equipped with a GFP reporter and spectinomycin resistance, was cultured on LB plates or LB medium with 100 μg/mL spectinomycin at 37°C. B. subtilis TH4545 strain was cultured on LB plates or LB medium with 5 μg/mL kanamycin at 37°C. S. aureus BAA-1696 strain was cultured on LB plates or LB medium with 25 U/mL polymyxin B at 37°C. C. albicans was cultured on YPD plates or YPD medium with 100 μg/mL ampicillin and 50 μg/mL kanamycin at 30°C. S. marcescens, E. faecium and Leucobacter spp. were isolated from adult mosquito midgut (Pang et al., 2016) and cultured on LB plates or LB medium without any antibiotics at 37°C.

The genes of A. aegypti Rel1A (AAEL007696), Rel1B (AAEL006930), and STAT (AAEL009692) were isolated from an A. aegypti cDNA library, and then cloned into pAc5.1-V5-His A vector (Invitrogen, Cat. No# V4110-20). The recombinant expression plasmids were named as pAc.5.1-Rel1A-V5, pAc5.1-Rel1B-V5, and pAc5.1-STAT-V5, respectively. For the truncation assay, the reporter plasmids were constructed by inserting the truncations of Gambicin promoter into a pGL3-Basic plasmid (Promega, Cat. No# E1751). The inserted promoter regions were followed down-stream by a firefly luciferase gene. A renilla luciferase gene was inserted into pAc5.1-V5-His A (pAc5.1-Renilla) was transfected as an internal control. The plasmids with mutants (M1–M4) were constructed by the pGL3 plasmids with 1000 bp promoter region (pGL3-1k) via a Fast MultiSite Mutagenesis System (Transgen, Cat. No# FM201).

The monolayer cells without aggregation were suitable for transfection. Briefly, the Aag2 cells were seeded at 3 × 10⁶ cells/mL per well in a 6-well plate. The cells formed a monolayer after 12 h of culture. Then, 2 μg of dsRNA was premixed with Effectene® (Qiagen, Cat. No# 301425) according to the manufacturer's instructions, and consequently added to the cells. After 6–18 h of transfection, the medium was replaced with fresh medium. The cells were cultured for the following investigation.

The monolayer Aag2 cells were seeded as described above. Microbe cells cultured to logarithmic phase were collected by centrifuge and washed with sterile PBS twice. The bacteria resuspended by PBS was then added into the cultured cell medium. The final microbial concentration in the cell medium was 0.05 OD₆₀₀ for each microorganism. 12 h later, the stimulated Aag2 cells were collected to isolate total RNA that was synthesized into cDNA for qPCR detection.

Adult female mosquitoes were kept on ice for 15 min, and then transferred to a cold tray to receive a systemic injection of 300 nL of 5 OD₆₀₀ microbial cells at logarithmic phase (Cheng et al., 2010; Liu et al., 2014; Xiao et al., 2014). 6 h later, the inoculated mosquitoes were sacrificed to isolate total RNA for quantitative AMP mRNA analysis.

The samples of animal models or cells were homogenized in Buffer I of an RNeasy Mini Kit (Qiagen, Cat. No# 74106) with a Pestle Grinder System (Fisher Scientific, Cat. No# 03-392-106). The detailed procedure of total RNA isolation is described in the RNeasy Kit manual. cDNA was randomly reverse-transcribed using an iScript™ cDNA Synthesis Kit (Bio-Rad, Cat. No# 1708891). The AMP expression was then relatively quantified with qPCR by SYBR® Green II method. The primers are shown in the Supplementary Table S2. The amount of AMP expression was normalized with A. aegypti actin (AAEL011197). Fold changes were calculated by comparison to the corresponding controls using the comparative Ct method (2^−ΔΔCt). All the primer pairs used in this study were reported in the previous research (Xiao et al., 2014, 2015; Pang et al., 2016) and the specific amplification reactions were confirmed by melt curve analysis.

The recombinant pGL3 plasmids with Gambicin promoter regions were mixed with pAc5.1-Renilla (19:1 w/w), and the plasmid mixture were subsequently transfected into the Aag2 cell with Effectene® Reagent (Qiagen, Cat. No# 301425). 36 h later, 0.05 OD₆₀₀ E. coli cells at logarithmic phase were used for bacterial challenge for 12 h. And then, the treated cells in a well of 48-well plate (about 6.5 × 10⁴ cells) were harvested for lysis by 100 μL per well 1 × Passive Lysis Buffer (PLB) supplied with the Dual-luciferase Report System Kit (Promega, Cat. No# E1910). The lysates were centrifuged, and subsequently the supernatant was transferred into a 96-well white polystyrene assay plate (Corning, Cat. No# 3922). 50 μL of each LAR-II and Stop&Glo solution were subsequently added into the wells for detection of firefly and renilla luciferase signals. The fluorescence was detected by a Varioskan® Flash reader (Thermo-Fisher, Cat. No# 5250030). The value of firefly luciferase was normalized by that of renilla luciferase.

The EMSA is used to validate the binding of specific nuclear regulator proteins and the specific potential regulation sites in the promoter area. This assay includes two steps: Nuclear protein extraction and EMSA assay.

The Aag2 cells were transfected by the plasmids of pAc.5.1-Rel1A-V5 / pAc5.1-Rel1B-V5 (1:1 mix) or pAc5.1-STAT-V5 for 48 h, respectively. Then infected the cells with logarithmic-phase E. coli at a final concentration of 0.05 OD₆₀₀ for 12 h as described above. Subsequently, the transfected and infected cells were washed twice and collected in cold PBS buffer. The cell nuclear was extracted by a NE-PER nuclear extraction kit (Thermo-Fisher, Cat. No# 78833). Briefly, the transfected cells in a well of 6-well plate (about 8 × 10⁵ cells) were lysed by 200 μL of CER-I buffer containing 1 × complete EDTA-Free protease inhibitor (Roche, Cat. No# 04 693 132 001). Then, 11 μL ice-cold CER-II was added to the tubes. After centrifugation (16,000 × g, 5 min at 4°C), the pellets were resuspended in 100 μL ice-cold NER reagent. The samples were placed on ice and continued vortexing for 15 s every 10 min for 4 times. And then, the tubes were centrifuged at 16,000 × g for 10 min. The supernatant fractions (nuclear protein extracts, NPE) were immediately transfer to a clean pre-chilled tube. The nuclear protein extracts were stored at −80°C freezer until using.

The EMSA assay was performed with a LightShift® EMSA Kit (Thermo-fisher, Cat No# 20148) according to the manufacturer's instruction. Briefly, the nuclear protein extracts, biotin-labeled oligonucleotide probes and 1 μL of anti-V5 antibody (MBL, Cat. No# M167-3, Lot. No. 005), were added into a 20 μL reaction system, including 1 × binding buffer (1 mM Tris, 50 mM KCl, 1 mM DTT; pH 7.5), 0.05% Non-idet P-40, 2.5% glycerol and 50 ng/μL poly (dI:dC). Unlabeled or mutant oligonucleotide served as the competitors. The unlabeled probe can eliminate the shift by compete the binding of the probe and the regulator protein; the mutant oligonucleotide cannot compete the binding, so they can not eliminate the shift band. After incubation at room temperature for 20 min, the reaction mixture was run on 5% polyacrylamide gel, and then the bounds were transferred to a Nylon membrane (Bio-Rad, Cat. No# 162-0153) via a Trans-blot® SD Simi-Dry Transfer System (Bio-Rad, Cat. No# 170-3940). Consequently, the DNA was cross-linked to the membrane for 15 min. The membrane was replaced in 20 mL blocking buffer for 15 min, and then additionally incubated in another 20 mL blocking buffer containing 66.7 μL stabilized streptavidin-HRP conjugate (1:300 dilution) for 1 h. After 4 washings, the membrane was transferred to a new container with 30 mL substrate equilibration buffer and incubated for 5 min. After treated by a substrate working solution provided by the EMSA kit, the membrane was imaged by a ChemiDoc™ Imaging System (Bio-Rad, Cat. No# 1708251).

Mosquitoes were randomly allocated into different groups. Mosquitoes that died before measurement were excluded from analysis. The investigators were not blinded to the allocation during the experiments or to the outcome assessment. All experiments were performed independently at least 2 times. Descriptive statistics are provided in the figure legends. A Kruskal–Wallis analysis of variance was conducted to detect any significant variation among replicates. If no significant variation was detected, the results were pooled for further comparison. Given the nature of the experiments and the type of samples, differences in continuous variables were assessed with the non-parametric Mann–Whitney test. Differences in AMP fold changes were analyzed by using t-test with Welch's correction. All results are expressed as mean ± S.E.M. from independent experiments. P-values < 0.05 were considered significant (^*P < 0.05, ^**P < 0.005, ^***P < 0.0005 and ^****P < 0.0001). All analyses were performed using GraphPad Prism® statistical software.

The Aag2 cell line is an A. aegypti cell lineage of embryonic origin (Peleg, 1968). This cell lineage is immuno-competent and has similar immune responses to that of live A. aegypti mosquitoes. Therefore, it is widely used as an immune cell model for studies of mosquito immunity (Gao et al., 1999; Fallon and Sun, 2001; Sim and Dimopoulos, 2010; Barletta et al., 2012).

We examined the patterns of AMP regulation in various microbe infections. We added Escherichia coli (Gram-negative), Serratia marcescens (Gram-negative), Staphylococcus aureus (Gram-positive), Enterococcus faecium (Gram-positive), Leucobacter spp. (Gram-positive), and Candida albicans (fungi) at 0.05 OD₆₀₀ onto a confluent monolayer of the Aag2 cells, respectively. Uninfected cells served as negative controls. 12 h later, the treated cells were collected to isolate total RNA for AMP detection by quantitative PCR (qPCR). 3 Defensins (Def A, C and D), 6 Cecropins (Cec A, D, E, F, G, and N), and Gambicin were dramatically induced, while the other AMPs showed a modest induction or no response to the E. coli and S. marcescens infections (Figures 1A,B). A similar pattern of AMPs with a modest induction presents by infection with the Gram-positive bacteria S. aureus, E. faecium, and Leucobacter spp. (Figures 1C–E). However, infection with a fungi C. albicans rarely induced AMP expression in the Aag2 cells (Figure 1F). We examined the cell death after incubated with microbes. There were no significant differences between the microbes treated groups and the untreated group, suggesting that the induction of AMP genes is not due to a response of killed cells (Supplementary Figure S1).

We next determined the AMP expression pattern after infection with Bacillus subtilis, a Gram-positive bacterium. The regulatory spectrum of AMPs caused by B. subtilis infection was similar to that caused by Gram-negative bacterial infection (Figures 1G,H). This phenomenon may be reasoned from that Bacillus genus share the similar DAP-type peptidoglycans as those in Gram-negative bacteria (Nguyen-Huy et al., 1976) which enables to activate the Imd pathway in Drosophila (Lemaitre et al., 1997; Leulier et al., 2003). Then, we assessed the in vivo AMP expression pattern in live A. aegypti mosquitoes, 6 h following infection through intrathoracic injection of various microbes. The in vivo AMP expression pattern in live mosquitoes shares similarity to that of Aag2 cells (Supplementary Figure S2). This may be due to the complication of mosquito immune system in which other mechanism, such as phagocytosis, encapsulation, and complement-like factors, may be induced after infection (Hillyer and Christensen, 2002; Waterhouse et al., 2010; Barletta et al., 2012).

In Drosophila, AMPs are regulated by both Toll and Imd pathways. However, the regulatory mechanism of AMPs expression in A. aegypti is still unclear. We have found that infections with E. coli, S. marcescens, and B. subtilis are capable of dramatically activating AMP expression in Aag2 cells by a similar manner (Figure 1H). Besides, It was known that Bacillus bacteria activates the Imd pathway by its DAP-type peptidoglycans (Nguyen-Huy et al., 1976). We therefore determined whether the induction of AMPs is controlled by the Imd pathway. Expression of the key components in the Imd [Imd and Rel2 (refer to Relish in Drosophila)], Toll [Myeloid Differentiation Factor 88 (MyD88) and Rel1A], and JAK-STAT [Domeless (Dome) and STAT] pathways were silenced by double-stranded RNA (dsRNA) transfection in the Aag2 cells (Figure 2A). Green fluorescent protein (GFP) dsRNA was used as a negative control. Subsequently, E. coli at 0.05 OD₆₀₀ was added to the transfected cells. AMP expression was determined by qPCR 12 h after the bacterial challenge. Knockdown of the Imd pathway components (Imd and Rel2) dramatically impaired the induction of most AMPs (fold change more than 2) in the Aag2 cells (Figure 2B, Supplementary Figure S3 and Supplementary Table S1); however, genetic manipulation of the Toll and JAK-STAT components showed only a modest change (fold change less than 2) for a few of the AMP transcripts (Figures 2C,D, Supplementary Figure S3 and Supplementary Table S1), suggesting induction of AMPs is predominantly controlled by the Imd pathway. In the Imd-mediated signaling cascade, the cleaved Imd activates a MAPK kinase kinase, TAK1 (transforming growth factor β-activated kinase 1), which is responsible for further activating both the JNK (c-Jun N-terminal kinase), and IKK (inhibitor of κB kinase)/Relish branches of the Imd pathway (Silverman et al., 2003; Kleino and Silverman, 2014). We therefore determined the AMPs regulation in the JNK-silencing Aag2 cells (Supplementary Figure S4A). Impairment of JNK expression did not show any reduction, however showed enhancement in some AMP transcripts (fold change more than 2), such as Def E, Cec B, Cec I, Cec J, and Dpt, in the Aag2 cells (Supplementary Figure S4B), suggesting that the JNK signaling is likely to play a negative-regulatory role in AMP expression in mosquitoes. Therefore, we show a different AMPs regulatory pattern between A. aegypti and Drosophila. The Drosophila AMPs are regulated both Toll and Imd pathway (Meister et al., 1997). However, the AMPs in A. aegypti are mostly regulated by Imd pathway (Guillen et al., 2014; Yadav et al., 2015). Nevertheless, one AMP–Gambicin–is uniquely regulated by all the three pathways in Aag2 cells, which let us to focus on this peptide.

The induction of Gambicin was apparently reduced by interrupting either of the three pathways in the Aag2 cells (Figure 2), suggesting that Gambicin transcription may be combinatorially regulated by these three pathways. To investigate the regulation of Gambicin, we next cloned the 1000 bp promoter region which contains a TATA box and an arthropod initiator sequence (Cherbas and Cherbas, 1993) upstream of the Gambicin gene into a pGL3-Basic vector (Supplementary Figure S5). We designed and constructed the truncations of the Gambicin promoter shown in Figure 3A. A plasmid constitutively expressing renilla luciferase (pAc5.1-Renilla), which was co-transfected with the promoter-inserted recombinant plasmids, served as an internal control. The transfected cells were then challenged by E. coli at 0.05 OD₆₀₀ and assayed for luciferase activation. There was no change in luciferase activation with the deletion of the −1000 to −600 bp region of the Gambicin promoter. However, the activity was apparently impaired by deletion of the region from −600 to −500 bp, indicating that one or more key regulatory site(s) might exist in this region (Figure 3B). In the signal cascades of the mosquito Imd, Toll, and JAK-STAT pathways, the activated of transcription factors, such as Rel1 for Toll pathway, Rel2 for Imd pathway, and STAT for JAK-STAT pathway, bind to specific regulatory sites on the promoters to initialize the transcription of downstream genes. The sequences of these regulatory sites have been characterized by previous studies (Yan et al., 1996; Ehret et al., 2001; Lin et al., 2004; Shin et al., 2005) (Figure 3C). We therefore predicted the regulatory sites within the region from −600 to −500 bp for the three pathways. Intriguingly, a STAT regulatory site (−577 to −568 bp) and a Rel2 regulatory site (−542 to −531 bp) were predicted in this region. We therefore mutated these sites in the 1000-bp Gambicin promoter (pGL-1k) and designated them as M1 (the STAT −577 to −568 bp mutant) and M2 (the Rel2 −542 to −531 bp mutant) (Figure 3D; Supplementary Figure S5). We then transfected luciferase plasmids with these mutants into the Aag2 cells for bacterial stimulation. Compared to the effects of the wild type promoter, both mutants significantly reduced the level of bacteria-mediated luciferase activation (Figure 3E), suggesting that both regulatory sites are independently essential for Gambicin induction.

Deletion of the region from −600 to −500 bp impaired the bacteria-mediated luciferase activation. Nonetheless, we did not rule out additional key regulatory sites present in the down-stream region beyond −500 bp. A potential Rel2 and a potential Rel1 regulatory sites were then predicted (−477 to −468 bp and −388 to −379 bp, respectively). We next mutated these two sites, which were designated as M3 (the Rel2 −477 to −468 bp mutant) and M4 (the Rel1 −388 to −379 bp mutant) (Figure 3F and Supplementary Figure S5). Compared to the effects of the wild type promoter, transfection by the recombinant plasmids with the M4, but not the M3, mutant repressed bacteria-mediated luciferase activation (Figure 3G), indicating that the region from −388 to −379 bp is a functional Rel1 regulatory site for Gambicin induction.

We identified regulatory sites on the Gambicin promoter for transcription factors representing three signaling pathways. We next selected the STAT (−577 to −568 bp) and Rel1 (−388 to −379 bp) regulatory sites to validate the binding affinity between the sequences and their transcription factors. The A. aegypti Rel1 and STAT genes were cloned into pAc5.1-V5-His-A expression vectors, and subsequently, the recombinant plasmids were transfected into the Aag2 cells. The expression of Rel1 and STAT was confirmed by western blot analysis with an anti-V5 antibody (Figure 4A). We silenced the Rel2 gene by dsRNA transfection to avoid a non-specific cross-reaction between the Rel1 regulatory site and the Rel2 factor. In the Rel1-transfected, Rel2-silenced Aag2 cells, the recognition between the Rel1 regulatory sequence and the ectopically expressed Rel1 transcription factor was determined via a super-shift in EMSA assay (Figures 4B,C). In the STAT-transfected Aag2 cells, a shift of the probe at the STAT regulatory site was clearly detected using the same approach (Figures 4B,D). The V5 antibody-mediated super-shift indicated specific binding between the probes and these ectopically expressed transcription factors, demonstrating the specific recognition between the transcription factors and their regulatory binding sites.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.