IntroductionThe mosquito midgut is the first barrier confronting many mosquito-borne disease pathogens, including Plasmodium, dengue virus, and so on1,2. Besides its function of separating pathogens physically and secreting antimicrobial peptides (AMPs), the mosquito midgut creates Reactive Oxygen Species (ROS) to kill pathogens efficiently. The ROS is created by Nicotinamide adenine dinucleotide phosphate (NADPH) Oxidase family enzymes, including NADPH Oxidase (NOX) and Dual Oxidase (DUOX)3. Theoretically, DUOX and NOX may sense cellular stress and pathways related to calcium imbalance4, such as endoplasmic reticulum stress, which are induced by infection or damage5. Structurally, both NOX and DUOX enzymes have EF domains that accept calcium information inside the cell6 and facilitate electron transmission between their Flavin Adenine Dinucleotide (FAD), NADPH binding domains, which allow the transmembrane domain to create ROS, including superoxide and hydrogen peroxide to outside of the cell7,8. DUOX has an additional peroxidase domain that creates more toxic ROS, such as hypochlorous acid9,10. DUOX and NOX are expressed differently in different organs and may have different killing potential based on the location and ROS creation types11,12,13. Studies are exploring their role against mosquito pathogens or symbiotic bacteria under different conditions.Duox has a dual role in killing and protecting bacteria. Usually, mosquito ingests pathogens in a blood meal accompanied by peritrophic protein matrix formation and an increase in microbiome diversity14,15. Duox expression increases after a blood meal, and this suppresses the microbiome16. Duox knockdown disrupts the peritrophic protein matrix and then helps maintain pathogenic or conventional bacteria from inducing midgut AMPs instead17. In Drosophila, which don’t have the blood-induced peritrophic matrix, Duox stills have a dual role, combating infection and tolerating symbiotic bacteria10,18 achieved by a feedback loop pathway. In detail, under ECC15 (Erwinia carotovora 15) infection, an insect pathogen, Duox-derived ROS is not induced in the midgut. But it can be transferred to the midgut from the malpighian tubules, and it is necessary for midgut cell repair12. These ROS cause cell damage and activate JAK/STAT and MAPK pathways, which are cellular stress pathways associated with cell proliferation19,20 and inducible by unbalanced calcium from cell damage21. It is not clear so far how Duox knockdown could affect mosquito pathogenic bacteria invasion without the blood meal effect or microbiome and midgut cell regeneration. Nox is also involved in JAK/STAT and MAPK-mediated cell regeneration induced by ECC15 in Drosophila. It usually reacts to broader stresses, including chemical-induced damage and symbiotic bacterial infection directly than Duox9,22,23. In mosquitoes, Nox helps combat viruses and activates the Toll immune pathway24 after a blood meal. It seems Nox also works on mosquito pathogens. But generally, it is not clear what the function of Nox is, the difference between Nox and Duox in confronting pathogenic bacteria infection and symbiotic bacteria. Cell stress pathways, including JAK/STAT, MAPK, and cell regeneration, may be involved.In this study, we characterized the role of Duox and Nox in eliminating ECC15 and shaping the midgut microbiome. We further determined the transcriptomic changes upon ECC15 oral infection controlled by Duox and Nox and their effect on midgut cell regeneration. We show that Nox, instead of Duox, was necessary for eliminating ECC15 and suppressed more rare species in the microbiome, though both genes affect ECC15 infection and microbiome composition. Transcriptionally, we found that Nox controls more gene responses than Duox upon ECC15 infection; the heat shock protein pathway may be the main stress pathway discriminating Nox and Duox. It may regulate stress signals that induce immune AMP, JAK/STAT, and MAPK pathway genes expression. As a result, only Nox was found necessary for robust cell regeneration after ECC15 infection. Collectively, our study suggests that the Nox instead of Duox gene may be the central coordinator in the ROS-related pathway regulating cell transcriptome, repair, and immunity against both pathogenic and symbiotic bacteria in the mosquito midgut.ResultsNox (Nicotinamide adenine dinucleotide phosphate Oxidase) instead of Duox (Dual Oxidase) was necessary for Aedes aegypti midgut ECC 15 (Erwinia carotovora 15 ) elimination, and dominant microbe speciesAedes aegypti Nox (NADPH Oxidase) family genes consist of a single Nox and Duox gene. They accept both cellular calcium signals and synthesize ROS. Nox may be induced in broader conditions, while Duox may create much toxic ROS. In order to understand which of them controls pathogenic bacteria elimination in the midgut, we knocked down Nox, Duox, or lacz gene as the control in mosquitoes before orally infecting them with ECC15 (Erwinia carotovora 15) OD600 = 100 and monitored their survival and midgut colony-forming unit (CFU). As shown in Fig. 1A, after Lacz knock-down, midgut CFU started to decrease at 36 h post ECC15 infection, and a lot of them were zero at 96 h post infection. The Nox knock-down group midgut CFUs were maintained at a similar level at 0, 6, 36, and 96 h post-infection. The Duox knock-down group midgut CFUs decreased at 36 h post-infection but were higher than Lacz group. Both Nox and Duox decreased mosquito survival in the following 14 days (around 60% at 14 days) post-infection at a similar level compared to Lacz (around 85% at 14 days). Nox, but not Duox, was necessary for ECC15 elimination. We also determined if either of the two genes could affect the Aedes aegypti microbiome by detecting CFU and survival post-knockdown without infection. Neither Nox nor Duox changed microbiome CFUs very significantly at 72, 96, and 120 h post knock-down, though Duox midgut CFU decreased a little 72 h (Fig. 1C). All three knock-down groups of mosquitoes had similar survival (around 95%, Fig. 1D). So, we further checked if microbiome compositions were changed at 72 h post-knockdown and 48 h by 16s rRNA sequencing, as the CFU change may indicate early compositional change. All mosquito midguts were dominated (above 50% relative abundance) by Asaia bacteria at 48 h post knock-down (Fig. 1E). There was no obvious compositional difference between the groups. At 72 h post knock-down, Duox and Lacz group mosquito midguts were dominated by Serratia or Enterobacteria, while in the Nox group, Asaia remains the main genus, and a lot of low-abundant genera grew (Fig. 1E, Supplementary Data 1). Alpha-diversity analysis by Shannon and Simpson indexes with emphasis on low-abundance and dominant species showed that Nox had higher indexes at 72 h post-knockdown (Fig. 1F) than Lacz and Duox groups. The efficiency of gene silencing was confirmed by qPCR quantification at 24 h (Supplementary F1). In summary, Nox but not Duox was necessary for Aedes aegypti midgut to eliminate ECC15 and maintain a microbiome dominated by some species.Fig. 1: Nox, but not Duox is necessary for ECC15 elimination and suppresses microbiome rare genus.A Mosquito midgut CFU (colony-forming-unit) on LB plates with 100 ug/ml rifampicin at 0, 6, 36, and 96 h post ECC15 infection in Duox, Lacz, and Nox knockdown mosquitoes (2 days after RNAi), each dot represents one gut, n = 3. B The survival of knockdown mosquitoes after ECC15 infection in the following 2 weeks, n = 3. C The mosquito midgut microbiome CFU on MRS plates at 72, 96, and 96 h post dsRNA injection, n = 3. D The survival of knockdown mosquitoes without ECC15 infection in the following weeks post-injection. E The mosquito midgut microbiome composition at 48, 72 h post RNAi. F The Shannon and G Simpson indices of mosquito microbiome diversity at the genus level at 48, 72 h post dsRNA injection. Statistical differences were done by Kruskal–Wallis test for CFU and the Cox test for survival. The significances were labeled as different letter above the bar or on the end of the curve P