Voltage-gated ion channels are expressed in the Malpighian tubules and anal papillae of the yellow fever mosquito (Aedes aegypti), and may regulate ion transport during salt and water imbalance

Mosquitoes are economically important vectors of infectious diseases (e.g. malaria, Zika virus and Ebola), contributing to the annual death toll of ∼700,000 people from, and ∼4 billion people at risk of contracting, mosquito-borne infectious diseases worldwide (Lozano et al., 2012; WHO, 2014). Disease transmission occurs when the female mosquitoes take blood from humans to gain protein needed to produce eggs, and, in the process, transmit infectious disease pathogens to the human host. Following blood feeding, the female mosquito lays eggs in water, which can introduce vectors of infectious diseases into new areas not yet affected by an established mosquito population. Recent incursions into regions that do not usually have permanent breeding populations create a direct threat (Scheffers et al., 2016).

The mosquito life cycle consists of a filter-feeding aquatic larval life stage hatching from eggs laid in water, after which developing larvae pupate and metamorphose into terrestrial adults. Mosquito larvae can be found in diverse aquatic habitats and, in general, demonstrate an extensive range of salinity tolerance compared with most animals (Bradley, 1994). Most Ae. aegypti larvae develop in a constant salinity (mostly freshwater, FW), larvae of females that do choose to lay eggs in brackish water (BW) near coastline may experience fluctuations in salinity, with documented cases ranging from 2 to 18 ppt (Ramasamy et al., 2011), and established survival in up to 10 ppt (e.g. de Brito Arduino et al., 2015). These larvae are often faced with rapid changes in salinity caused by heavy rainfall, evaporation, or incursion of seawater (Bradley, 1987). Worldwide, Ae. aegypti are increasingly present in BW habitats and more northern latitudes, currently reaching as far as Canada in North America, with projections predicting expanding habitat range worldwide (Ramasamy et al., 2011, 2014; Surendran et al., 2012; Lounibos and Kramer, 2016; Ramasamy and Surendran, 2016; Surendran et al., 2018a,b; Parker et al., 2019; Giordano et al., 2020; Khan et al., 2020; Brennan et al., 2021). This includes Southern California, where Ae. aegypti has established breeding populations across the state counties stemming from recent introduction from multiple genetically divergent source populations – it has recently been detected in 72% of households in Los Angeles County, posing increased risk of infectious disease transmission and spread for inhabitants of coastal cities that surround ocean and brackish waters (Lee et al., 2019; Donnelly et al., 2020). Southern California lacks natural FW habitats and mosquito eggs are typically laid by blood-fed females into bodies of standing water created by human activity or in bodies of FW, which often connect with the ocean. Thus, all larvae hatched from eggs laid in FW actively hyper-osmoregulate, keeping hemolymph ion concentrations higher than those of FW. However, larvae may end up drifting downstream towards the ocean through BW. Eggs laid in small human-made standing bodies of water can face rapid evaporative water loss, increasing in salinity and nitrogenous waste content as larvae develop (Bradley, 1987). In fact, the recent spread of Ae. aegypti may be due to its ability to tolerate saline BW that surrounds coastal cities, where the capacity of some populations to complete their life cycle in BW is considered a heritable trait (Lee et al., 2019).

The Malpighian tubules (MTs) are an osmoregulatory organ found in both adult and larval mosquitoes, consisting of a multifunctional epithelium, which together with the hindgut is essential for the maintenance of solute and water homeostasis, acid–base balance and nitrogenous waste excretion, among other functions (Piermarini and Gillen, 2015; Esquivel et al., 2016; Piermarini et al., 2017a). MTs of FW-dwelling larvae produce isosmotic urine, which is modified by the rectum via KCl reabsorption to conserve ions (recycling ions between MTs and the rectum to move excess water out via the urine). In contrast, post-prandial urine production by MTs in blood-fed females is plentiful, but urine is closer to being isosmotic, containing both ions and water from a recent blood meal (Bradley, 1994; Drake et al., 2010; Li et al., 2017). Ion transport is achieved via principal and secondary epithelial cells of the MTs, which carry out cation and anion transport, and thus directly enable the excretory function of the MTs by creating an osmotic gradient that draws water into the tubule (Beyenbach and Piermarini, 2011). Mechanisms of ion transport in the MTs of mosquitoes are well studied, whereas the regulation of ion transport remains largely understudied.

The anal papillae (AP) are exclusively larval osmoregulatory organs that are in direct contact with the external aquatic environment. Classical ultrastructural and functional studies revealed that these sac-like structures are filled with hemolymph and are syncytial in nature (Koch, 1938; Treherne, 1954; Ramsay, 1953; Sohal and Copeland, 1966). In FW habitats, AP actively sequester ions and experience diffusional water uptake from surrounding FW (Wigglesworth, 1933; Marusalin et al., 2012), and are capable of completely replenishing Na⁺ and Cl⁻ content of the hemolymph in ∼3 h (Donini and O'Donnell, 2005; Durant et al., 2021). The larvae are known to survive in more saline habitats, with AP altering in ultrastructure, ion and water transport kinetics, and gene expression, but remaining functional despite fluctuating salinities over the course of larval development (Stobbart, 1971; Donini et al., 2007; Marusalin et al., 2012; Akhter et al., 2017; Surendran et al., 2018a,b; Durant et al., 2021). Molecular mechanisms of ion transport in AP of mosquitoes have recently been described; however, the detailed mechanisms of its regulation remain to be studied (Donini and O'Donnell, 2005; Patrick et al., 2006; Del Duca et al., 2011; Durant et al., 2021).

Larvae of Ae. aegypti maintain salt and water balance during BW exposure using a combination of the hindgut, the MTs and AP, resulting in hemolymph osmolality equivalent to BW, but without active osmoconformation (Bradley, 1994). Multiple studies report that Ae. aegypti can be reared in, or abruptly exposed to, ∼30% seawater, although detailed mechanisms of how short-term variations in environmental salinity are detected and handled by the osmoregulatory tissues of mosquito larvae remain unclear (e.g. Edwards, 1982; Donini et al., 2006, 2007; Jonusaite et al., 2016, 2017; Durant et al., 2021). Recent transcriptomic surveys revealed that voltage-gated ion channels (VGICs) are expressed in all non-innervated non-contractile osmoregulatory epithelia of animals studied to date (Kapoor et al., 2021; Durant et al., 2021; Kolosov and O'Donnell, 2022). VGICs are often ion selective and can respond rapidly to changes in membrane potential (V_m) (Kariev and Green, 2012). In insect MTs, specifically, voltage-gated Ca_V1 and HCN channels have recently been shown to directly regulate ion transport in the MTs of larval lepidopterans (Kolosov and O'Donnell, 2019; Kolosov et al., 2021).

In mosquitoes, MTs and AP are simple epithelial organs with tracheae – neither is innervated or capable of peristaltic movement; thus, all regulation of MT ion transport must be endo/auto/paracrine, or aided by autonomous osmosensing of the organ (Edwards and Harrison, 1983; Coast, 1998). Additionally, lack of innervation or muscle layers eliminates the confounding variables of VGIC contribution to the regulation of ion transport via neuronal/muscular pathways. The objectives of the current study were to: (i) generate MT transcriptomes of FW-acclimated and BW-exposed larvae of Ae. aegypti, (ii) detect all VGICs expressed in the MTs and AP, (iii) investigate whether VGICs alter in mRNA abundance in MTs and AP of larvae exposed to BW, and (iv) investigate whether select VGIC proteins are expressed in these osmoregulatory epithelia. We demonstrate that multiple VGICs are expressed in MTs and AP epithelia of larval Ae. aegypti, and that the abundance of several changes with BW exposure. As Ca_V1 has been shown by our recent study to be involved in the regulation of ion transport in the MTs of larval lepidopterans (Kolosov et al., 2021), we additionally show that Ca_V1 may be specifically linked to ion transport in the MTs of Ae. aegypti.

Eggs of Aedes aegypti (Linnaeus) were obtained from Benzon Research (Carlisle, PA, USA), which provided the basis for the lab Aedes colony. Eggs were hatched in dechlorinated FW ([Na⁺] 19.2±5.6 µmol l⁻¹, [Cl⁻] 44.3±15.5 µmol l⁻¹). Larvae were fed a diet of inactivated yeast and liver powder, water was replaced every 2 days and larval density was maintained throughout development at 50 larvae l⁻¹. Pupae were transferred to bug dorms daily, where adult Ae. aegypti were kept and fed 5% aqueous sucrose solution and warmed-up Alsevers sheep blood (Carolina Biological Supply Company, Burlington, NC, USA). Lab temperature is maintained at 21°C throughout the year, and the light:dark cycle is maintained at 12 h:12 h using timed light fixtures.

For salinity exposure experiments, several groups of 100 larvae were exposed to different salinities. Half the containers were exposed to a typical water change to FW. The other half were exposed to 10 ppt BW ([Na⁺] 132.0±4.2 mmol l⁻¹, [Cl⁻] 151.9±3.1 mmol l⁻¹) for 24 h. Salinity exposure containers were set up in a paired and staggered manner to allow for prompt dissection of tissues of larvae exposed to salinities for ∼24 h. BW was made in the lab using Instant Ocean aquarium salt (United Pet Group Inc., Cincinnati, OH, USA) and double-distilled water. This mixture was measured using a portable refractometer until 10 ppt was reached.

Hemolymph was collected from BW- and FW-exposed larvae under hydrated oil by gently puncturing the exoskeleton with iris microscissors. Larvae were captured and briefly rinsed in double-distilled water to wash environmental salt off the exoskeleton. Collected hemolymph samples were immediately analyzed for ion content using ion-selective microelectrodes constructed according to Donini and O'Donnell (2005). Briefly, K⁺- and Na⁺-selective microelectrodes were constructed by pulling glass microcapillaries (cat. no. B150-110-10, Sutter Instruments, Novato, CA, USA) on a PUL-1000 microelectrode puller (World Precision Instruments, Inc., Sarasota, FL, USA) with a barrel diameter of ∼2–3 µm. Freshly pulled microelectrodes were silanized and baked, and kept in a desiccator until further use. K⁺ Ionophore 1–cocktail A was purchased (cat. no. 99311, Sigma-Aldrich, St Louis, MO, USA). Na⁺ ionophore was prepared in the lab, and consisted of (w/w): 3.5% Na⁺ ionophore X (cat. no. 71747, Sigma-Aldrich), 0.6% potassium tetrakis(4-chlorophenyl)borate (cat. no. 60591, Sigma-Aldrich) and 95.9% 2-nitrophenyl octyl ether (cat. no. 73732, Sigma-Aldrich). On the day of use, ion-selective electrodes were constructed by back-filling the glass microelectrode with 150 mmol l⁻¹ KCl or NaCl, front-filling it with the liquid-state ionophore cocktails mentioned above, and calibrating in the following solutions: K⁺, 1.5 mmol l⁻¹ and 15 mmol l⁻¹; Na⁺, 15 mmol l⁻¹ and 150 mmol l⁻¹. Solid-state Cl- electrodes were constructed according to Kolosov and O'Donnell (2020), and calibrated in 15 and 150 mmol l⁻¹ NaCl. All calibration solutions were balanced to maintain constant ionic strength. Ion-selective microelectrodes and reference electrodes (non-silanized, back-filled with 150 mmol l⁻¹ KCl) were connected to a high-impedance electrometer (ML165 pH Amp, ADInstruments, Colorado Springs, CO, USA), which in turn was connected to a PowerLab 2/26 data acquisition system running LabChart software (ADInstruments, Sydney, NSW, Australia). Hemolymph ion concentration measurements were conducted under oil and ion-selective microelectrode data were acquired and analyzed using LabChart software. Voltage readings from hemolymph samples were used to back-calculate ion concentrations in the samples using the following equation: [ion]_sample=[ion]_standard×10^(dV/S), where [ion]_sample is hemolymph ion concentration, [ion]_standard is ion concentration in the calibration solution, dV is the difference in voltage between the hemolymph and the calibration solution, and S is the Nernstian slope of the electrode measured in response to a 10-fold change in ion concentration.

Microdissections of larval Ae. aegypti were performed under a stereomicroscope in larval saline containing (in mmol l⁻¹) 30 NaCl, 25 Hepes, 3 KCl, 5 NaHCO₃, 0.6 MgSO₄, 5 CaCl₂, 5 glucose, 5 proline, 9.1 glutamine, 8.74 histidine, 14.4 leucine, 3.37 arginine, 5 succinic acid, 5 malic acid and 10 tri-sodium citrate, pH 7.10 (Donini et al., 2007) (all chemicals were purchased from fisher scientific, ThermoFisher Scientific, Carlsbad, CA, USA). Larvae were pinned down by the head and MTs and AP were dissected out in larval Aedes saline. Tissue samples were manipulated and microdissected out using Dumont 5 ultrafine forceps and glass microprobes custom-made in the lab, while carefully and diligently removing tracheal connections and surrounding tissues to ensure no other tissues were harvested together with AP or MTs. The tissue samples were stored in Eppendorf tubes filled with 500 µl of RNAlater^® Stabilization solution (ThermoFisher Scientific) and samples were collected for each osmoregulatory tissue in each salinity type. Each biological replicate contained 200 MTs or 160 AP microdissected out of 40 larvae from the same treatment group. During collection time, previously harvested biological replicates were stored at 4°C. Tissue collection was always performed in a timed and paired manner to ensure ∼24 h of salinity exposure and the side-by-side collection of the two salinity treatment samples.

RNA extraction (RNA precipitation, wash and redissolving) was performed using TRIzol reagent (ThermoFisher Scientific) according to the manufacturer's instructions as described in Kolosov et al. (2019). Briefly, larvae were dissected in larval Aedes saline and tissues collected as described above. Following microdissection and storage in RNAlater^®, tissues were homogenized in Trizol^® using a 26G syringe needle and total RNA was extracted following the manufacturer's protocol. RNAlater was removed by aspiration and tissues were immersed in 0.5 ml Trizol (ThermoFisher Scientific) per biological replicate. Tissues were homogenized in Trizol using a 26G syringe needle and allowed to remain at room temperature for 5 min to dissociate nucleoprotein complexes. Molecular grade chloroform (100 μl) was then added to every sample. Tubes were vigorously agitated by hand and incubated at room temperature for 3 min. Samples were then centrifuged at 12,000 g for 15 min at 4°C to perform phase separation, after which the clear aqueous phase containing total RNA was collected into a set of new, sterile 1.5 ml Eppendorf tubes and kept on ice. Molecular grade isopropanol (250 μl) was added to every sample to precipitate RNA. Tubes were agitated gently and left at room temperature for 10 min, after which they were centrifuged at 12,000 g for 10 min at 4°C to concentrate the precipitate into the pellet. Supernatant was removed from every tube by aspiration and pellets were washed in 75% molecular grade ethanol solution in RNase-free water, and centrifuged at 7500 g for 5 min to immobilize pellets. Ethanol was then removed by aspiration and pellets were left to air-dry for 1 min at room temperature, after which they were dissolved in 12 μl of RNase-free water. Total RNA was purified using a Thermo Scientific GeneJet RNA Cleanup and Concentration Micro Kit (cat. no. K0841, ThermoFisher Scientific). The concentration and quality of purified RNA were determined using a nano-sample spectrophotometer (DeNovix DS-11, Wilmington, DE, USA). Absorbance ratios of A₂₆₀/A₂₈₀=2.11±0.01 and A₂₆₀/A₂₃₀=2.08±0.03 were observed, indicating excellent purity of samples free of contamination. Total purified RNA from 3 biological replicates for each salinity treatment group was sent to University of California Riverside Genomics Facility for Illumina library preparation (see ‘RNAseq library preparation, sequencing, quality control, mapping and differential expression analysis’, below).

Following a similar RNA extraction and purification process, 2 μg of total RNA, treated with DNAse, was used for oligo-dT-aided reverse-transcriptase cDNA synthesis. cDNA was diluted in RNAse/DNAse-free water and used for PCR and qPCR (see ‘PCR/qPCR amplification and gel electrophoresis’, below).

Single-end 75 bp Illumina libraries were prepared by the University of California Riverside Institute for Integrative Genome Biology sequencing facility (UCR IIGB, Riverside, CA, USA) using the following methodology. Total RNA samples were screened with an Agilent Bioanalyzer 2100 to confirm quality and concentration. Each sample was diluted to 250 ng for input into NEBNext Poly(A) mRNA Magnetic Isolation Module (E4790, New England Biolabs), followed by NEBNext Ultra II Directional RNA Library Prep kit for Illumina (E7760). The NEB protocol was followed with the following adjustments: 0.8× bead clean after second strand synthesis, adaptor dilution of 1:30, 0.7× bead clean after ligation, 15 cycles of amplification, and dual bead clean after enrichment. Standard Illumina adaptors were used for all, but each library was barcoded independently with TruSeq indexes. Final libraries were qualified and quantified with Agilent Bioanalyzer 2100, then pooled equimolarly. Sequencing was performed with Illumina NextSeq500 using a High-output Single-end 75 bp kit.

Sequencing (at UCR IIGB) of six strand-specific libraries (3 FW exposed and 3 BW exposed) with 75 bp single-end reads yielded a total of 398,425,646 reads with an average of ∼66.4 million of raw reads per library. Quality control (QC) reports were performed by UCR IIGB and 100% of reads with a read length of exactly 75 bp were subjected to quality control trimming with Trimmomatic using previously described protocols (usegalaxy.org; Goecks et al., 2010; Kolosov et al., 2019). Raw sequencing data are available at SRA BioProject (PRJNA1068135).

The AaegL5.0 A. aegypti transcriptome was accessed at ncbi.nlm.nih.gov and official genome transcript annotation was used for mapping. Mapping of Next Generation Sequencing (NGS) reads to this transcript set was performed using Salmon (Galaxy version 1.3.0, usegalaxy.org) mapping software (Patro et al., 2017). Mapping was performed to the Ae. aegypti transcriptome and annotation supplied with the official genome annotation was used.

Differential expression analysis was performed in R (version 3.4.2) using the DESeq2 package (release 3.7) (Love et al., 2014), where the BW-exposed group was compared with the FW-exposed group. This analysis involved compiling counts tables for all samples generated by Salmon in Galaxy, uploading them to R, creating a metadata table indicating which sample belongs to which treatment group, and running the DESeq2 script in R as instructed in the vignette publication, which estimates size factors and dispersions, determines mean dispersion relationship, and builds a model of differential expression. DESeq2 analysis in R resulted in a list of all differentially expressed transcripts (Table S1) accompanied by the log₂ fold-change (log₂FC), P-value of the change (P) and P-value adjusted (P_adj) for the false discovery rate (FDR) by the DESeq2 Benjamini–Hochberg algorithm. For further information, the reader is directed to Love et al. (2014).

Detailed analysis of NGS data suggested that multiple voltage-gated, ligand-gated and mechanosensitive ion channels (VGICs, LGICs and MSICs, respectively) were expressed in the MTs of larval Ae. aegypti. Expression of transcripts encoding VGICs was determined by RT-PCR in MTs and AP using Platinum or DreamTaq Hot-Start Supermix (ThermoFisher Scientific). Changes in transcript abundance in MTs and AP of BW-exposed larvae were assessed using quantitative real-time PCR (qPCR) with 2× DyNAmo Flash SYBR Green qPCR Mastermix (F415L, ThermoFisher Scientific), a BioRad PCR machine for RT-PCR (PTC-2000; Bio-Rad Laboratories Canada Ltd, Mississauga, ON, Canada) and ThermoFisher Scientific QuantStudio 3 qPCR machine running ABM Design and Analysis software. Primers specific for VGICs were designed based on sequences obtained from RNAseq experiments. Gel electrophoresis was performed on PCR amplicons to confirm their size, and electrophoresed samples were subsequently extracted from the gel, purified and sequenced to ensure that amplicon identity was in line with designed parameters, and submitted to GenBank for annotation (see Table 1 for PCR primer and cycling information).

The following reaction conditions were used: 1 cycle for denaturation (95°C, 4 min), followed by 40 cycles of: denaturation (95°C, 30 s), annealing (see Table 1, 30 s) and extension (72°C, 30 s), with a final extension step (72°C, 10 min). To ensure that a single PCR product was synthesized during reactions, a dissociation curve analysis was carried out after each qPCR run. Transcript abundance was normalized to that of Ae. aegypti actin (act). The use of act for gene of interest normalization in salinity exposure studies was validated by statistically comparing tub threshold cycle values between tissues to confirm that no statistically significant changes occurred (P=0.842 for MTs and P=0.4688 for AP, Student's t-test).

Whole-mount immunohistochemistry (IHC) procedures have been described in detail elsewhere (Kolosov et al., 2018a; Patrick et al., 2006). Briefly, larvae were dissected in larval saline (as described above); adult tissues were dissected in adult saline (in mmol l⁻¹: 150 NaCl, 3.4 KCl, 1.8 NaHCO₃, 1.7 CaCl₂, 1.0 MgSO₄, 25 Hepes and 5.0 glucose; Yu and Beyenbach, 2002). After dissection in appropriate saline, MTs or AP were removed and fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS, pH 7.4) overnight at 4°C. The following day, tissues were washed in PBS and dehydrated (20% v/v, stepwise to 100%) and rehydrated (100% to PBS) in a methanol/PBS series. Tissues were then permeabilized and blocked in PBS containing 0.1% Triton X-100 solution (PBT), containing 2% w/v bovine serum albumin (BSA). After blocking, tissues were incubated with primary antibody overnight at 4°C at 1:100 dilution. The anti-VGIC antibodies were custom-ordered from GenScript (Piscataway, NJ, USA), epitope-affinity purified and directed against the C-terminal region of the following Ae. aegypti proteins encoded by mRNA detected by RNAseq and PCR (NCBI accession no.; epitope): Ca_V1 alpha-1 isoform (XP_021699870.1; CWSEYDPDAKGRIKH), Na⁺ leak channel Nalcn (XP_021710075.1; MKMLGRKQSLKGEPC), voltage-gated K⁺ channel KCNH8 (XP_021699750.1; ILKEFPEELRGDIS) and transient receptor potential TRP/Painless (XP_001652261.2; CLTANDKRPGDDDY). An additional preparation was included with each IHC analysis, where primary antibody was omitted to act as negative control. Tissues were washed after 16–18 h in PBT/1% BSA supplemented with normal goat serum 3 times for 15 min each with constant agitation to remove unbound primary antibody. Following washes, tissues were incubated with goat anti-rabbit TRITC-conjugated secondary antibody at 1:1000 dilution (Jackson ImmunoResearch, Westgrove, PA, USA) in the dark at room temperature for 2 h. Following incubation with secondary antibody, tissue was washed 3 times in PBT/1% BSA for 15 min each time. Tissues were then mounted on slides using a transfer pipet. Slides were then blotted dry using KimWipes and preparations were mounted using ProLong Antifade^® reagent (ThermoFisher Scientific) containing 4,6-diamidino-2-phenylindole (DAPI) as a nuclear stain under coverslips and left to cure in the dark. Images of VGICs were obtained using a Nikon Eclipse CI-S microscope and Nikon DS-Qi2 camera in imaging facilities at the Department of Biological Sciences at California State University San Marcos.

Cross-section immunohistochemistry procedures were performed as follows. Tissues were carefully microdissected out in appropriate saline (see above) and fixed in Bouin's solution for 4 h at room temperature. Following fixation, all tissues were rinsed 3 times in 70% ethanol and stored in 70% ethanol at 4°C until further processing. Fixed tissues were dehydrated through an ascending series of ethanol rinses (70–100%), cleared with xylene and infiltrated and embedded in Polyfin Tissue Embedding Medium (cat. no. 50-279-84, fisher scientific). Then, 5 μm thick sections were cut using an HM 355 S rotary microtome (MICROM International GmbH, Walldorf, Germany), collected on glass slides and incubated overnight at 45°C. Sections were deparaffinized with xylene, rehydrated to water via a descending ethanol series (100% to 50%), and subjected to heat-induced epitope retrieval (HIER). HIER was accomplished by immersing slides in a sodium citrate buffer (10 mmol l⁻¹, pH 6.0) and heating both solution and slides in a microwave oven to 92–95°C. The solution was allowed to cool for 20 min, reheated again and cooled for a further 15 min. Slides were then washed 3 times with PBS (pH 7.4) and quenched for 30 min in 3% H₂O₂ in PBS. Following quenching, slides were successively washed with 0.4% Kodak Photo-Flo 200 in PBS (PF/PBS, 10 min), 0.05% Triton X-100 in PBS (PBT, 10 min), and 10% antibody dilution buffer (ADB: 10% goat serum, 3% BSA in PBT) in PBS (ADB/PBS, 10 min). Slides were incubated overnight at room temperature with one of the rabbit polyclonal anti-VGIC antibodies (anti-NALCN, anti-TRP/Painless, anti-KCNH8, or anti-Ca_V1; 1:100 dilution in ADB). Custom epitope affinity purified anti-VGIC antibodies (see above) were used. As negative controls, two sets of slides were also incubated overnight with ADB alone. Following overnight incubation, sections were successively washed with PF/PBS, PBT and ADB/PBS (10 min each) as described above, and incubated with tetramethylrhodamine isothiocyanate (TRITC)-labeled goat anti-rabbit antibody (1:500 in ADB; Jackson ImmunoResearch) for 2 h at room temperature. Slides were then successively washed with PF/PBS, PBT and PF/PBS (10 min each) and rinsed 3 times with 0.4% PF in distilled water (PF/dH₂O, 1 min each). Slides were air dried for 1 h and mounted with Molecular Probes ProLong Antifade (ThermoFisher Scientific) containing DAPI. Fluorescence images were captured using a Nikon Eclipse CI-S microscope and Nikon DS-Qi2 camera, and merged using ImageJ software (US National Institutes of Health, Bethesda, MD, USA).

Adult female mosquitoes were either fed on sucrose solution or fed on animal blood, and harvested for microscopy 4 h after feeding. Once MTs were microdissected out, tissue samples were processed as above for Ca_V1 immunohistochemistry (see below).

Negative control preparations with primary antibodies omitted were used with every set of IHC samples, did not produce immunofluorescence, and were omitted for brevity.

MTs were isolated in ice-cold physiological saline, transferred to microtubes and stored at −80°C for later analysis. Tissues from 40 larvae were combined in each tube. For examination of Ca_V1 expression, tissue samples were thawed on ice and homogenized in a RIPA homogenization buffer containing 50 mmol l⁻¹ Tris-HCl, pH 7.5, 150 mmol l⁻¹ NaCl, 1% sodium deoxycholate, 1% Triton X-100, 0.1% SDS, 1 mmol l⁻¹ PMSF and 1:200 protease inhibitor cocktail (Sigma-Aldrich). All homogenates were then centrifuged at 13,000 g for 20 min at 4°C, and the protein content of the collected supernatants was determined using the Pierce 660 nm assay (ThermoFisher Scientific) according to the manufacturer's guidelines. Samples (10–20 μg protein) were prepared for SDS-PAGE by heating for 5 min at 100°C in a 6× loading buffer containing 360 mmol l⁻¹ Tris-HCl (pH 6.8), 12% (w/v) SDS, 30% glycerol, 600 mmol l⁻¹ DTT and 0.03% (w/v) Bromophenol Blue. SDS-PAGE electrophoresis and western blot analysis of Ca_V1 were conducted according to a previously described protocol (Kolosov and Kelly, 2017) using anti-Ca_V1 antibody at a 1:1000 dilution. Antigen reactivity was visualized using SuperSignal West Pico PLUS Chemiluminescent Western ECL substrate kit (34577, ThermoScientific, Rockford, IL, USA) and images were captured using a iBright CL750 imaging system (ThermoFisher Scientific). A brightfield image of the blot was acquired to superimpose the ladder onto the final blot figure panel.

Larvae were dissected under larval Ae. aegypti saline (see recipe above). MTs together with the gut were dissected out and MTs were pulled away from the gut and separated from tracheae. MTs were excised and mounted in petri dish pre-coated with poly-l-lysine to aid tissue adherence. Transepithelial potential (V_te) and basolateral membrane potential (V_bl) were measured in separate preparations by impaling the tubule lumen with a microelectrode pulled from BF150-110-10 glass (Sutter Instruments) on a microelectrode puller PUL-1000 (World Precision Instruments, Inc.) where the barrel was filled with 150 mmol l⁻¹ KCl (Kolosov et al., 2021). All potentials were measured with respect to a reference electrode back-filled with 150 mmol l⁻¹ KCl. V_te and V_bl were recorded with a high impedance (>10¹³ Ω) electrometer (ML165 pH Amp, ADInstruments), which in turn was connected to a PowerLab 2/26 data acquisition system running LabChart software (ADInstruments). Apical membrane potential (V_a) was calculated as the difference between V_te and V_bl.

Stock solutions of nifedipine (Sigma-Aldrich), an inhibitor of voltage-gated Ca_V1 channels, was prepared in DMSO and diluted in saline to 0.1 μmol l⁻¹ (final DMSO concentration in the bath did not exceed 0.01% v/v). Addition of inhibitor to the bathing saline for V_te/V_bl was achieved by replacing 1/10th of the volume of bathing saline with freshly prepared inhibitor solution at 10× the final concentration in saline. Nifedipine at concentrations used in this study did not affect the response time V_te/V_bl electrodes. V_te and V_bl measurements were performed in different cells.

Significance of the effect of salinity exposure on hemolymph ion concentration, and mRNA abundance of ion channels was assessed with Student's t-test with a significance limit of P<0.05, following checks for normality and homogeneity of variance (performed in GraphPad Prism 7 statistical software). The effects of pharmacological inhibition of Ca_V1 on basolateral and transepithelial membrane potentials were assessed using a paired t-test in the same software with the same significance limit.

Hemolymph ion concentrations were measured to determine whether 24 h exposure to BW constituted a significant challenge to the salt-and-water balance of the larvae. Significant elevation in hemolymph [Na⁺] and [Cl⁻] was observed in BW-exposed larvae. Hemolymph [Na⁺] increased significantly from 89.0±4.2 mmol l⁻¹ to 132.1±1.6 mmol l⁻¹ (n=9–10, t-test P<0.0001) (Fig. 1A). Hemolymph [Cl⁻] increased significantly from 96.2±3.3 mmol l⁻¹ to 128.0±2.5 mmol l⁻¹ (n=10, t-test P<0.0001) (Fig. 1B). Hemolymph [K⁺] did not change significantly (5.4±0.7 mmol l⁻¹ to 6.7±0.9 mmol l⁻¹, n=10, P=0.2822) (Fig. 1C).

A preliminary literature search and survey of publicly available and collaborator transcriptomic datasets indicated that larval MTs and AP, as well as adult MTs express a variety of VGICs, which change in transcript abundance in response to salt and water imbalance (Table 2). Following this initial analysis, which was particularly useful for identifying targets in the AP of larvae, all analysis aimed at identifying targets in MTs was conducted using transcriptomic data generated in the lab.

Analysis of our own transcriptomic data detected expression of multiple voltage-gated cation-, Na⁺-, K⁺- and Ca²⁺-selective channels, as well as many LGICs and MSICs in the MTs of larval Ae. aegypti (Fig. 2). VGIC assemblage in MT epithelia included: several transient receptor potential (TRP) A, M and V channels, ranging in mean abundance from 2041.66 transcripts per million (TPM) (TRPA/pyrexia) to 0.11 TPM (TRPV5); voltage-gated K⁺ channels KCNS1/shab, KCNH8/elk, KCNQ2, KCNH6/erg2, KCNQ1, KCNH1/eag and K_Ca5.1/slo, ranging in abundance from 550.71 TPM (KCNS1/shab) to 1.00 TPM (K_Ca5.1/slo); voltage-gated Na⁺ channels nalcn (5.11 TPM), 60E (0.74 TPM) and para (1.90 TPM); voltage-gated Ca²⁺ channels Ca_V1 (1.60 TPM) and Ca_V3 (1.60 TPM); and cation-selective HCN2 (0.67 TPM) (see Table S1).

Deseq2 analysis of BW-exposed and FW-acclimated MT transcriptomes revealed that several VGICs altered in transcript abundance in the MT with salinity exposure. Transcripts of voltage-gated TRP/pyrexia, TRP/painless 1 and nalcn were less abundant in the MTs of larvae exposed to BW. In contrast, voltage-gated TRP/painless 2, TRP/painless 3 and KCNH8 were all more abundant in the MTs of BW-exposed larvae (see Table S1).

In order to establish the background of salt-and-water balance transcriptome changes, alterations in transcript abundance of solute and water transporters were quantified as well. Presence and abundance changes of ion pumps, channels and transporters, as well as aquaporins, hormone receptors and transcripts encoding cyclic nucleotide and Ca²⁺ signaling pathways, and septate junction proteins have been provided in the supplementary information and are not discussed in the manuscript for brevity and clarity of focus (see Table S1, ‘General transporters’).

PCR and gel electrophoresis confirmed the expression of several VGICs from different ion channel families using transcript-specific primers, where most channels were found to be expressed in both MTs and AP of Ae. aegypti (Fig. 3). Expression of the voltage-gated calcium channel Ca_V1, the sodium channels nalcn and parax7, the potassium channel KCNQ1, and the transient receptor potential channels TRPA1, TRP/painless 1, 2 and 3, as well as TRP/pyrexia 1 and 2 was detected in both MTs and AP of Ae. aegypti larvae. Expression of the hyperpolarization-activated cyclic nucleotide-gated sodium/potassium channel HCN2×4 was detected only in the MTs.

Transcriptomics did not find any effect of BW exposure on the expression levels of VGICs – RNAseq is prone to false negatives due to a low number of replicates, low expression levels of many genes, insufficient depth of sequencing, and multiple comparison false discovery rate adjustment (e.g. Robert and Watson, 2015). Therefore, we employed qPCR to detect multiple changes in mRNA abundance between FW- and BW-exposed larvae. In AP, mRNA abundance of KCNQ1 and TRP/pyrexia 1 significantly decreased ∼2-fold, while that of nalcn decreased ∼4-fold. TRP/painless 2 significantly increased ∼4-fold, while TRPA1 increased ∼2-fold in mRNA abundance following 24 h BW exposure (Fig. 4). Similarly, in MTs, voltage-gated Ca_V1, TRP/pyrexia 2 and TRPA1 all significantly decreased ∼2-fold in mRNA abundance, while KCNQ1 significantly increased ∼12-fold, and TRP/painless 1 and TRP/painless 3 significantly increased ∼3.5-fold in mRNA abundance (Fig. 5). All P-values regarding data presented in Figs 4 and 5 can be found in Table S1.

Immunohistochemistry was used to demonstrate that Ca_V1 was expressed in the luminal side of AP (Fig. 6A,B), Nalcn was expressed in the water-facing membrane (Fig. 6C,D), while KCNH8 expression was restricted to the distal-most tip of the AP (Fig. 6E,F), and TRP/Painless was expressed in the water-facing membrane of the AP (Fig. 6G,H).

In contrast, expression of all VGICs in the MTs was restricted to the apical (lumen-facing) membrane of the principal cells, which expressed Nalcn (Fig. 7A,B), TRP/Painless (Fig. 7C,D) and KCNH8 (Fig. 7E,F). Western blotting with anti-Ca_V1 antibody identified Ca_V1 as a single protein of expected molecular weight (Fig. 8A); this was detected in the MTs larvae (Fig. 8B) and sugar-fed (Fig. 8C) and blood-fed (Fig. 8D) adults.

Pharmacological inhibition of Ca_V1 significantly depolarized MT epithelia of adults and larvae. In larvae, V_te decreased significantly from 29.02±1.93 mV to 20.16±2.16 mV (P<0.001), but V_bl did not alter significantly (−74.20±6.77 mV to −61.40±10.51 mV) (Fig. 8E), corresponding with a calculated change in V_a from approximately −103 mV to approximately −81 mV. In contrast, both V_te and V_bl altered in adult MTs from 41.38±6.95 mV to 15.40±4.35 mV (P=0.015) and from −32.28±2.19 mV to −20.28±2.05 mV (P=0.003), respectively, corresponding with a calculated change in V_a from approximately −74 mV to approximately −36 mV.

In the current study, we employed data mining from publicly available datasets, as well as in-lab transcriptomics to indicate the presence of VGICs in AP and MTs of larval Ae. aegypti. We then used acute salinity exposure combined with PCR and gel electrophoresis, qPCR, immunohistochemistry and V_m measurements to directly demonstrate the presence of VGICs in osmoregulatory epithelia of larval Ae. aegypti and their potential relevance to salt and water balance in MTs and AP. An acute exposure to increased salinity presented a challenge in osmoregulatory function, as demonstrated by ion loading of the hemolymph. A hypothesis-generating, transcriptomic approach was used to find VGIC genes that might be involved in ion transport and its regulation in the MTs and AP of larval mosquitos, and suggested that multiple VGICs belonging to several ion channel families were expressed in these osmoregulatory organs (Table 2, Fig. 2). Using PCR and gel electrophoresis, we confirmed that voltage-gated Ca²⁺, cation-selective, Na⁺ and K⁺ channels are expressed in both MTs and AP (Fig. 3). Additionally, to investigate the abundance of VGIC transcripts in a quantitative manner, we determined that in BW-exposed larvae, KCNQ1 (∼2-fold), nalcn (∼4-fold) and TRP/pyrexia 1 (∼2-fold) demonstrated lower mRNA abundance in AP (Fig. 4), while Ca_V1 (∼2-fold), Nalcn (∼3-fold), TRP/pyrexia 2 (∼2-fold) and TRPA1 (∼4-fold) were lower in mRNA abundance in the MTs (Fig. 5). In contrast, BW exposure upregulated TRP/painless 2 (∼4-fold) and TRPA1 (∼2-fold) in AP (Fig. 4), and KCNQ1, TRP/painless 1 and TRP/painless 3 (∼3.5-fold) in MTs (Fig. 5). We used immunolocalization to confirm the protein products of target VGICs are in the osmoregulatory epithelia, and found luminal Ca_V1, water-facing Nalcn and TRP/Painless, and distal KCNH8 in AP epithelia (Fig. 6), and all four VGICs in the apical (luminal) membrane of principal cells in MT epithelia (Fig. 7). Of special interest was the pharmacological inhibition of Ca_V1 in larval and adult MTs, in agreement with previous studies on larval lepidopterans, which led to depolarization of V_te in larvae and adults, as well as depolarization of V_bl in adults only (Fig. 8). These data bring forth evidence of the presence of VGICs in non-innervated, non-contractile epithelia of mosquito larvae, and their functional relevance to the regulation of epithelial ion transport and salt and water balance of the animal. Whether detected VGICs directly participate in directional ion transport and/or setting of V_m in MTs and AP of larval mosquitoes, or whether they are simply used respond to V_m changes preceding the reactionary change in ion transport (e.g. environmental, or systemic ion levels that epithelia are exposed to, or mechanosensation cues) will require further mechanistic study.

VGICs may provide a mechanism for rapidly adjusting ion transport in MTs and AP. In Ae. aegypti, ion transport in MTs and AP can alter quite rapidly with salinity exposure (Wigglesworth, 1933; Donini et al., 2007; Surendran et al., 2018a,b). Changes in ultrastructure, ion and water transport needed to compensate for passive diffusional ion loss in the hypo-osmotic medium of FW are minimized in favor of iso-osmotic BW needs of the animal (Durant et al., 2021; Ramasamy et al., 2021). MTs of BW-exposed Ae. aegypti secrete less K⁺ but the same amount of Na⁺ (Donini et al., 2006). In general, fluid secretion by the MTs is regulated by hormones (e.g. kinin, CAPA) and the intracellular second messengers Ca²⁺ and cyclic nucleotides (Donini et al., 2006). Similarly, AP in BW-exposed larvae reduce Na⁺ and Cl⁻ uptake from FW following 6 h of BW exposure (Donini et al., 2007).

Increased hemolymph levels of Na⁺ and Cl⁻ in BW-exposed larvae with no significant change in K⁺ are in line with previous reports of acute BW exposure of freshwater obligate mosquito larvae (Patrick and Bradley, 2000; Patrick et al., 2001; Donini et al., 2006, 2007). Ion loading of the hemolymph has been shown to lead to changes in ions used in diuresis. Alterations of K⁺ transport may be utilized to either conserve Na⁺ under FW (Na⁺-deprived) conditions or eliminate more Na⁺ in saline (Na⁺-rich) conditions (Donini et al., 2006). Simply put, the use of more K⁺ to draw water osmotically into MTs (which may require alterations in K⁺ transport mechanisms) spares hemolymph Na⁺ from being lost during diuresis.

General transcriptomic response of larval MTs to BW exposure indicate significant restructuring of ion transport machinery and its regulation. The roles of many of these proteins (e.g. V-type H⁺-ATPase, aquaporins, cyclic nucleotide and Ca²⁺ signaling, septate junctions) in ion and fluid secretion by MTs of insects are well established (e.g. Patrick et al., 2006; Piermarini et al., 2010; Piermarini and Gillen, 2015; Piermarini et al., 2017a,b; Misyura et al., 2020; Duong et al., 2022; Donini et al., 2006; Beyenbach et al., 2009; Ionescu and Donini, 2012; Tiburcy et al., 2013; Efetova et al., 2013; Gioino et al., 2014; Sajadi et al., 2018; Weng et al., 2008; Calkins and Piermarini, 2017; Jonusaite et al., 2017; Kolosov et al., 2018b). However, to the best of our knowledge, much less is known about the specific roles of many hormones and signaling pathways (e.g. GRK and MAPKK) in the regulation of ion transport in larval mosquitoes (readers interested in these are directed to Table S1 for further details).

In line with previous transcriptomic studies on osmoregulatory epithelia of other insects, multiple VGICs were detected in the MTs and AP of larval Ae. aegypti (Kolosov et al., 2019; Kolosov and O'Donnell, 2019; Kapoor et al., 2021; Durant et al., 2021). In the MTs of larval lepidopterans, Ca_V1 and HCN1 channels have been shown to regulate cation transport – both classes of VGICs are activated by changes in V_m, while HCN channels are additionally gated by cyclic nucleotides. Thus, the presence of these VGICs in MTs and AP of Ae. aegypti larvae may provide an additional connection of ion transport with Ca²⁺ levels, cyclic nucleotide levels and V_m.

Two types of previously undescribed channels were detected in AP and/or MTs in the current study but did not alter in mRNA abundance following BW exposure – HCN2 and para. HCN channels, permeable to both Na⁺ and K⁺, are unique among VGICs in that they have a reverse voltage dependence that leads to activation upon hyperpolarization and are additionally activated by cyclic nucleotides, where the latter overrides the former (Wahl-Schott and Biel, 2009). Cyclic nucleotides are known to enhance and reduce fluid secretion in the MTs of larval and adult Ae. aegypti, respectively (Donini et al., 2006; Sajadi et al., 2018). HCN channels can provide an additional link between direct activation of ion transport and second messenger-based hormone action. In larval Trichoplusia ni, when HCN1 channels are blocked in MTs, ion transport switches direction from K⁺ secretion to K⁺ reabsorption (Kolosov et al., 2019). Thus, HCN2 channels in Ae. aegypti may also be connected with the regulation of ion transport, which will require further detailed investigation.

para (short for paralytic) is a gene encoding a voltage-activated Na⁺ channel of insects (Warmke et al., 1997). Functional voltage-gated Na⁺ channels have been described in other animal epithelia, e.g. human intestinal epithelia, MTs of caterpillars (Barshack et al., 2008; Kolosov and O'Donnell, 2019). The presence of voltage-dependent Na⁺ channels in the MTs and AP may provide a rapid link between V_m and Na⁺ permeability in these osmoregulatory epithelia.

Of special interest is the presence of Ca_V channels in the MTs – these channels are key transducers of V_m changes into intracellular Ca²⁺ concentration changes that initiate intracellular physiological response (Catterall, 2011). At highly polarized V_m, Ca_V1 channels are normally closed. They are activated (opened) at depolarized V_m, allowing Ca²⁺ into the cell and initiating an intracellular response. Intracellular Ca²⁺ is a diuretic second messenger in insect MTs. In MTs of larval T. ni, Ca_V1 channels have been shown to connect cation transport with V_m via Ca²⁺ signaling (Kolosov et al., 2021). Pharmacological inhibition of Ca_V1 channels in the MTs of larval and adult Ae. aegypti in the current study had a similar effect on V_m, depolarizing V_te, to the previously published data in larval T. ni (Kolosov et al., 2021). Interestingly, in larval MTs, V_te decreased significantly without accompanying changes in V_bl, indicating that most of the measured membrane potential change may have taken place in the apical membrane. In contrast, in adults, both V_bl and V_a seemed to have contributed to the measured changes in V_te, offering potential insight into which transporters (apical and/or basolateral) may be regulated by Ca_V1.

KCNQ/K_V7 channels are found in several animal epithelia, including human airway epithelial cells (Mondejar-Parreño et al., 2020), intestinal epithelia (Preston et al., 2010) and kidney (Abbott, 2015), where they are connected with transepithelial K⁺ transport and the generation and maintenance of resting V_m and ionomotive driving force (Demolombe et al., 2001). KCNQ channels are unique in the way that they can be modulated to lose their voltage dependence and remain constitutively active, are inhibited by external K⁺ and can be modulated by intracellular Ca²⁺ and cAMP levels (Schroeder et al., 2000; Abrahamyan et al., 2023; van der Horst et al., 2020). There are pronounced differences in K⁺ transport by the MTs of FW- and BW-reared Ae. aegypti larvae, which are thought to be aimed at conserving Na⁺ levels in the hemolymph (Donini et al., 2006). KCNQ1 mRNA is more abundant in MTs of BW-exposed larvae and AP of FW-reared larvae. Thus, the KCNQ1 channel may contribute to K⁺ secretion in the MTs of BW larvae, helping the larvae rid themselves of extra K⁺ and preventing K⁺ loading with BW exposure. Additionally, it may contribute to the uptake of environmental K⁺ by the AP of larvae in FW, aiding larvae in retention of hemolymph K⁺ in the face of diffusional K⁺ loss to FW.

MTs of FW-reared larvae produce the same amount of fluid with the same amount of Na⁺ to BW-reared larvae, while AP actively take up more environmental Na⁺ from surrounding FW (Donini and O'Donnell, 2005; Donini et al., 2006). Previous research has also shown that Na⁺ is absorbed by the AP from surrounding FW with the help of electrodiffusive entry using inside-negative apical membrane potential (Edwards, 1983), which later was shown to involve V-type H⁺-ATPase (Patrick et al., 2006) and Na⁺ channels (Del Duca et al., 2011) in the apical (water-facing) membrane of AP. This specifically agrees with a previous study by Del Duca et al. (2011), who showed that a Na⁺ channel blocker phenamil stops Na⁺ uptake but that a Na⁺/H⁺ exchanger blocker [5-(N,N-hexamethylene)-amiloride] does not. Nalcn identified in this study can contribute to channel-based absorption of Na⁺ from FW by AP. How exactly Nalcn contributes to Na⁺ secretion in the MTs will require further detailed study. Alternatively, it may not participate in directional Na⁺ transport in the MTs, and instead serve as a Na⁺ sensor, contributing to the maintenance of V_m and connecting it to Na⁺ transport. In general, Nalcn is a VGIC selective for monovalent cations (Na⁺ and K⁺) expressed in excitable and non-excitable tissues (e.g. pancreatic β cells) (Monteil et al., 2023; Swayne et al., 2010; Xie et al., 2020). Heterozygous Nalcn knockout mice demonstrate increased blood [Na⁺] with a possible association between Nalcn and K⁺ excretion in humans (Sinke et al., 2011; Kho et al., 2020).

All six TRP channels detected in MTs and AP of larval Ae. aegypti belong to the TRPA channel subfamily. mRNA abundance of every TRP channel in the current study was significantly different between FW-reared and BW-exposed larvae in at least one of the examined tissues. TRP channels are cation-permeable (Na⁺, K⁺, Ca²⁺) voltage-dependent channels (Yue and Xu, 2021). TRP channels are activated through a range of gating mechanisms (V_m depolarization, ligands, etc.) (Venkatachalam and Montell, 2007). TRPA channels demonstrate wide tissue expression profiles (Cheng and Zheng, 2021), connecting them to the function of many excitable and non-excitable tissues, e.g. nerve endings of human urothelium (Vanneste et al., 2021); human lung epithelia (Ko et al., 2020), non-neuronal pancreatic islet cells, airway epithelia, and skin epidermis (Nilius et al., 2012), where it has been connected with inflammatory and cytokine release; and enteroendocrine cells of pig gut and Drosophila gut epithelia (Van Liefferinge et al., 2020; Gong et al., 2023). In human and rat colonic mucosa, TRPA1 has been connected with ion secretion (Talavera et al., 2020).

TRPA channels have been detected in MTs of T. ni (Kolosov et al., 2019), Pieris rapae (Mao et al., 2020) and Bactrocera dorsalis (Su et al., 2018). TRPA1, TRP painless and TRP pyrexia are more permeable to K⁺ than to Na⁺, with well-established roles in nociception and thermotaxis in excitable tissues (Tracey et al., 2003; Lee et al., 2005; Corfas and Vosshall, 2015). In contrast, and to the best of our knowledge, the roles of TRP channels in epithelia of MTs and AP have not been explored to date.

Aedes aegypti is an obligate FW mosquito. Increased appearance of Aedes larvae in coastal BW regions worldwide in recent years is a concern in terms of the increased spread of arboviral diseases (Lee et al., 2019; Surendran et al., 2018b). Aedes aegypti completing their life cycle in BW may provide a perennial reservoir of arboviral transmission during dry seasons worldwide (Surendran et al., 2018b). This trend has been exacerbated by global climate change (Ramasamy and Surendran, 2012). The genomic and physiological basis of salinity adaptation and blood feeding, as well as short-term exposure to changing salinity in coastal waters are just beginning to be explored by the scientific community (Esquivel et al., 2016; Li et al., 2017; Ramasamy et al., 2021; Durant et al., 2021). Completing their life cycle in BW is advantageous as it presents larvae with an opportunity to relax osmotic/ionic stress. Acute exposure to BW, however, does require adjustment of the osmoregulatory apparatus, and thus constitutes osmoregulatory stress.

Larvae living in water of rapidly changing salinity may benefit from having a mechanism in their excretory tissues for rapidly detecting such a change. Bioelectrical signals govern the cell biology of all tissues, including non-excitable tissues, playing important roles in processes such as regeneration and ion transport (McLaughlin and Levin, 2018; Kapoor et al., 2021). Studies on animal epithelia have reported expression of Ca²⁺-, Na⁺- and K⁺-selective, as well as non-selective and cation-permeable VGICs in epithelia of the lung, intestine (Barshack et al., 2008), kidney (Siroky et al., 2017) and skin (Pitt et al., 2021). Many of these channels are connected with intracellular Ca²⁺ signaling, osmotic stress response, extracellular ion sensing and modulation of directional ion transport (Abbott, 2015; Bleich and Warth, 2000; Demolombe et al., 2001; Morera et al., 2015; Nilius and Droogmans, 2001; Schönherr et al., 2000; Shi et al., 1997; Siroky et al., 2017; Yang and Cui, 2015; Zhu et al., 2010; Kolosov et al., 2021). The presence of VGICs in epithelia may provide a mechanism for the rapid detection of depolarizing stimuli (e.g. mechanosensation of fluid flow, changing ion concentrations) that leads to changes in V_m. MT epithelia of larval lepidopterans (Kolosov et al., 2019; Kolosov and O'Donnell, 2019) and mosquito larvae (current study) have been shown to express high levels of mechanosensitive Piezo channels, which activate in response to membrane stretch. Their activation may allow for the activation of VGICs, and amplification of the signal via TRP channels.

VGICs of mosquitoes are often targeted with insecticides, repellents and anti-feedants because of their established role in excitable tissues of insects (Salgado, 2017; Mack et al., 2021; Inocente et al., 2018). However, the presence of VGICs in the excretory and osmoregulatory epithelia of larvae and adults is rarely considered. VGICs in osmoregulatory epithelia may provide mosquito larvae with a mechanism for rapidly sensing systemic and environmental disturbances in salt and water content. Detailed examination of what every class of VGICs does in the mosquito MTs and AP will require mechanistic study. Some VGICs may directly participate in directional ion transport, whereas others may act as ion sensors or V_m setters. Whether VGICs remain voltage sensitive when expressed in non-excitable epithelia will require further mechanistic study using heterologous expression models.

The authors would like to thank Holly Clark and Clay Clark at IIGB Genomics Core at University of California, Riverside for synthesizing RNAseq libraries and completing sequencing.