Evidence that Rh proteins in the anal papillae of the freshwater mosquito Aedes aegypti are involved in the regulation of acid–base balance in elevated salt and ammonia environments

Like many aquatic invertebrate and vertebrate animals, the larvae of most mosquito species inhabit freshwater systems which can vary remarkably in water chemistry, including in ionic composition and pH (Patrick et al., 2002a, b). While freshwater NaCl concentrations can range between 0 and 8 mmol l⁻¹, mosquito larvae maintain much higher hemolymph ion levels in comparison to the surrounding environment ([Na⁺]=85.1±2.2 mmol l⁻¹ and [Cl⁻]=65.8±5.2 mmol l⁻¹) (Donini and O'Donnell, 2005; Patrick et al., 2002b). This presents a challenge to hemolymph ion homeostasis in a relatively hypo-osmotic environment. Strategies of counteracting ion loss and water gain have been investigated in aquatic dipteran larvae, which have proven to be effective osmoregulators in varying environmental salinity (Akhter et al., 2017; Jonusaite et al., 2011; Nguyen and Donini, 2010; Patrick et al., 2001). Furthermore, aquatic invertebrates, and insects in particular, are among the most pH-tolerant animals on the planet (Matthews, 2017). Mosquito larvae have been shown to tolerate chronic exposure to highly alkaline or acidic water (pH 4–11) in nature and under laboratory conditions with minor effects on hemolymph pH, growth and development (Clark, 2004). Acidified water coupled with ion-poor conditions can be particularly challenging, as the excretion of acid–base equivalents is coupled to the uptake of certain ions that are limited (Vangenechten et al., 1989).

The regulation of hemolymph ion composition and pH, which are often linked, is essential for homeostasis and survival in freshwater systems with varying water chemistry. In addition to ventilation and buffering strategies, aquatic insects regulate their hemolymph pH and osmolarity through the active uptake and excretion of ions and acid–base equivalents internally between the hemolymph and gut lumen, or with the external environment (Matthews, 2017). For most air-breathing aquatic insects, such as mosquito larvae, ions and acid–base equivalents are absorbed or excreted directly between the hemolymph and the surroundings through ion-permeable epithelia. This is unlike water-breathing aquatic insects, such as the aquatic nymphs of mayflies, which possess ionocyte-dense tracheal gills for active ion transport (Nowghani et al., 2017).

Many air-breathing aquatic dipteran larvae possess dedicated ionoregulatory and osmoregulatory organs known as anal papillae (Koch, 1938; Wigglesworth, 1932). The four anal papillae in larval Aedes aegypti are elongate sac-like projections of the cuticle that surround the anal opening, with a lumen that is continuous with the hemocoel of the body (Copeland, 1964; Sohal and Copeland, 1966). The thin cuticle, syncytial epithelium and relatively small surface area of the anal papillae restricts ion and water movement to this organ, while a relatively impermeable integument covers the rest of the body (Wigglesworth, 1932). These morphological features enable the anal papillae to function as major ion exchange organs. The anal papillae are the site of Na⁺, Cl⁻ and K⁺ uptake and the excretion of acid–base equivalents (H⁺, NH₄⁺, HCO₃⁻) in mosquito larvae (Donini and O'Donnell, 2005; Patrick et al., 2001; Stobbart, 1965, 1971). However, until recently there remained limited evidence of a role for the anal papillae in pH homeostasis (Clark et al., 2007).

Ammonia (NH₃/NH₄⁺) is generated through metabolic processes, primarily the deamination of amino acids, and can be used as an acid–base equivalent (Onken and Moffett, 2017). At physiological pH (∼7.4; pK_a=9.4), only 1.7% of total ammonia is present as NH₃ (Weiner and Verlander, 2017). The predominant ionic form, NH₄⁺, acts as an H⁺ equivalent that can alter intracellular and extracellular pH when transported (Weihrauch and Allen, 2018). However, the extent to which ammonia transport is used to mitigate changes in pH of internal fluids of aquatic invertebrates remains poorly understood. Evidence of a link between ion transport, pH homeostasis and ammonia excretion in the anal papillae epithelium of A. aegypti larvae has been shown through electrophysiology and pharmacological studies. Upon exposure of A. aegypti larvae to 30% seawater, hemolymph Na⁺, Cl⁻ and H⁺ concentrations increased within hours, coupled with a decrease in Na⁺ and Cl⁻ uptake at the anal papillae (Donini et al., 2007). Independent uptake of Na⁺ and Cl⁻ was also suggested based on differences in measured uptake kinetics of each ion. The inhibition of basolateral Na⁺/K⁺-ATPase (NKA), which functions to maintain a negative cell potential and low [Na⁺] in the cytosol, causes significant reductions in NH₄⁺ efflux at the anal papillae (Chasiotis et al., 2016; Patrick et al., 2006). The efflux of H⁺ driven by apical V-type H⁺-ATPase (VA) contributes to the negative intracellular voltage generated by NKA, facilitating NH₃ efflux likely through an ammonia-trapping mechanism at the apical membrane whereby NH₃ combines with H⁺ on the external side, which maintains an NH₃ gradient favoring excretion (Chasiotis et al., 2016; Durant et al., 2017; Weihrauch et al., 2009). Inhibition of VA also results in a significant reduction in ammonia excretion at the anal papillae epithelium, implicating VA and NKA in the process of ammonia excretion, which is likely to stem from establishing the negative electrical potential of the cytosol and ammonia trapping. Additionally, the negative cytosolic voltage generated by NKA and VA is important in energizing Na⁺ uptake in other aquatic invertebrates (Onken, 2003; Onken and Riestenpatt, 2002). Interestingly, inhibition of an apical Na⁺/H⁺ exchanger 3 (NHE3, SLC 9 family) decreases NH₄⁺ excretion at the anal papillae in A. aegypti larvae, suggesting that this transporter may either contribute to ammonia trapping or perhaps directly conduct NH₄⁺ across the apical membrane (Chasiotis et al., 2016).

Aedes aegypti possess two homologs of the vertebrate Rhesus glycoproteins (Rh proteins), AeRh50-1 and AeRh50-2, which have been localized to the apical and basolateral membranes of the anal papillae epithelium (Durant et al., 2017). It has been suggested that Rh proteins may transport NH₃ as well as CO₂; however, the specific transport substrate(s) of Rh proteins remains to be determined (Kustu and Inwood, 2006; Weihrauch et al., 2004; Weiner and Verlander, 2017). In A. aegypti larvae, knockdown studies of both AeRh50-1 and AeRh50-2 proteins results in a significant reduction in ammonia efflux (Durant et al., 2017). A decrease in [NH₄⁺] in the hemolymph coupled with an acidification of hemolymph was also observed in response to AeRh50-1 knockdown and it was suggested that this may occur if AeRh50-1 was capable of CO₂ transport. The partial pressure gradient of CO₂ plays a significant role in the acid–base balance of body fluids of aquatic animals (Nawata and Wood, 2008), and in this respect it was speculated that AeRh50-1 may mediate CO₂ excretion through the anal papillae with dsRNA-mediated knockdown, resulting in accumulation of CO₂ in the hemolymph; however, no measurements of CO₂ have been performed in relation to these studies (Durant and Donini, 2018).

There is currently little known about the transport mechanisms used by mosquito larvae to achieve their acid–base balance, with some evidence demonstrating a role of the anal papillae in pH homeostasis (Clark et al., 2007). Given the lack of information on the contribution of ammonia transport and the function of the anal papillae in acid–base balance in mosquito larvae, as well as the link between ion uptake and excretion of acid and base equivalents, we examined the transport rates and hemolymph levels of Na⁺, NH₄⁺ and pH in response to NH₄Cl (high environmental ammonia, HEA) and moderate levels of NaCl (approximately 10 times higher than freshwater levels). We hypothesized that hemolymph pH regulation is partly achieved through modulation of NH₄⁺ excretion, an H⁺ equivalent, by the anal papillae. If this occurs, then increased NH₄⁺ excretion at the anal papillae in conjunction with acidification of the hemolymph might be observed. Furthermore, changes in the magnitude of NH₄⁺ excretion by anal papillae may be representative of changes in ammonia transporter expression.

Aedes aegypti larvae (Liverpool strain) were obtained from a colony reared in the Department of Biology, York University (Toronto, ON, Canada). Larvae were reared in dechlorinated tap water (deH₂O) at room temperature (21°C) on a 12 h:12 h light:dark cycle. Larvae were fed daily with a 1:1 solution of liver powder and yeast in water. Rearing water was refreshed every other day. For rearing treatments employed in this study, larvae were hatched in deH₂O (µmol l⁻¹: [Na⁺] 590, [Cl⁻] 920, [Ca²⁺] 760, [K⁺] 43, pH 7.35) and transferred to either 5 mmol l⁻¹ NaCl (in deH₂O, pH 7.35) or 5 mmol l⁻¹ NH₄Cl (in deH₂O, pH 7.35) at 2 days post-hatching (Nguyen and Donini, 2010). Control larvae were maintained in deH₂O. Aedes aegypti larvae were reared in the respective treatments following the protocol above, until reaching fourth instar (approximately 7 days). Fourth instar larvae from each treatment were used 24 h post-feeding for all physiological and molecular studies.

Partial mRNA sequences for A. aegypti NHE3 (AAEL001503) were used for primer design (forward primer: 5′- CTACCTGGCGTATCTGAATGC-3′; reverse primer: 5′-CGTATTTGATGGTCGTGTGC-3′; 131 bp amplicon size, 58°C annealing temperature). The purified RT-PCR product, resolved by gel electrophoresis, was sequenced at The Centre for Applied Genomics (TCAG), The Hospital for Sick Children, to confirm NHE3 sequence specificity. Three biological samples, each consisting of a pool of 200 anal papillae from 50 larvae were isolated and collected in cold lysis buffer with 1% 2-mercaptoethanol (Ambion, Austin, TX, USA). Total anal papillae RNA was extracted using the Purelink RNA mini kit (Ambion) and genomic DNA was removed with an RNase-free DNase using the TURBO DNA-free^TM Kit (Applied Biosystems, Streetsville, ON, Canada). cDNA was synthesized using the iScript^TM synthesis kit (Bio-Rad, Mississauga, ON, Canada) with 0.75 µg of total RNA for each reaction.

qPCR using the primers described above was used to examine the relative mRNA abundance of NHE3 in the anal papillae of A. aegypti larvae. qPCR reactions were carried out using the CFX96™ real-time PCR detection system (Bio-Rad) and SsoFast™ Evagreen^® Supermix (Bio-Rad) according to the manufacturer's protocol. Ribosomal protein 49 (rp49) RNA served as the reference gene utilizing primers that have been previously designed and utilized (Paluzzi et al., 2014). A melting curve analysis was performed after each cycle to confirm the presence of a single product. Quantification of relative transcript abundance was determined according to the Pfaffl method (Pfaffl, 2004). The mRNA abundance of NHE3 in the anal papillae of larvae from each rearing condition (NaCl and HEA) was expressed relative to the control deH₂O, which was assigned a value of 1.0 after normalizing to rp49 transcript abundance.

Changes in the protein abundance of AeAmt1, AeAmt2 and AeRh50 s in response to rearing in HEA (5 mmol l⁻¹ NH₄Cl) have been previously documented (Durant and Donini, 2018). Here, the protein abundance of AeAmt1, AeAmt2 and AeRh50s in the anal papillae of fourth instar larvae in response to 5 mmol l⁻¹ NaCl was examined using western blotting following an established protocol (Chasiotis and Kelly, 2008). Three biological samples consisting of pooled anal papillae were isolated from 30 larvae in A. aegypti saline (Donini et al., 2007) and were sonicated (3×10 s) on ice in a homogenization buffer [50 mmol l⁻¹ Tris-HCl pH 7.4, 1 mmol l⁻¹ PMSF, 150 mmol l⁻¹ NaCl, 1% sodium deoxycholate, 1% Triton X-100, 0.1% SDS and 1:200 protease inhibitor cocktail (Sigma-Aldrich, Oakville, ON, Canada)], then centrifuged at 13,000 g for 10 min at 4°C, and the supernatant was collected and stored at −80°C. The preparation of samples consisted of 5 µg of protein from anal papillae (Bradford Assay, Bio-Rad) in radioimmunoprecipitation assay (RIPA) homogenization buffer and 6× loading buffer [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] for a total sample volume of 25 µl. Samples were prepared for SDS-PAGE by heating for 5 min at 100°C, and then were electrophoretically separated by SDS-PAGE (12% polyacrylamide). Western blot analysis of AeAmt1, AeAmt2 and AeRh50s was conducted according to an established protocol (Chasiotis and Kelly, 2008; Chasiotis et al., 2016; Durant et al., 2017). Custom-synthesized polyclonal antibodies raised in rabbit against AeAmt1, AeAmt2 and AeRh50s were used at dilutions of 1:500, 1:5000 and 1:2000, respectively, and have been previously validated (Chasiotis et al., 2016; Durant and Donini, 2018; Durant et al., 2017). The AeRh50 antisera is presumed to detect both AeRh50-1 and AeRh50-2 because of high epitope sequence similarity (Durant et al., 2017). Following chemiluminescent detection of ammonia transporter expression using a Gel Doc XR⁺ system (Bio-Rad), blots were stripped using stripping buffer (20 mmol l⁻¹ magnesium acetate, 30 mmol l⁻¹ KCl, 0.1 mol l⁻¹ glycine, pH 2.2) and total protein analysis was carried out using Coomassie total protein staining as a loading control (0.1% Coomassie Brilliant Blue R250, 50% methanol, 50% double-distilled H₂O, ddH₂O) (Eaton et al., 2013). Blots were incubated in Coomassie stain for 1 min followed by de-staining for 3–5 min in de-stain solution (50% ethanol, 10% acetic acid, 40% ddH₂O). Blots were rinsed in ddH₂O for 1 min, dried completely and visualized using a Gel Doc XR+system (Bio-Rad). Densitometric analysis of AeAmt1, AeAmt2, AeRh50s and Coomassie total protein was conducted using ImageJ 1.50i software (National Institutes of Health, Bethesda, MD, USA).

Hemolymph was collected by gently blotting A. aegypti larvae on filter paper to remove excess water, submerging the larvae in paraffin oil (Sigma-Aldrich) and gently tearing the cuticle with fine forceps to release a droplet of hemolymph into the oil. Na⁺, NH₄⁺ and pH concentrations in the hemolymph were determined by measuring the activities of the free ions (Na⁺, NH₄⁺, H⁺) using ISMEs (Donini and O'Donnell, 2005). The following ionophore cocktails (Fluka, Buchs, Switzerland) and back-fill solutions (in parentheses) were used: NH₄⁺ ionophore I cocktail A (100 mmol l⁻¹ NH₄Cl); Na⁺ ionophore II cocktail A (100 mmol l⁻¹ NaCl); and H⁺ ionophore I cocktail B (100 mmol l⁻¹ NaCl/100 mmol l⁻¹ sodium citrate, pH 6.0). The ISMEs were calibrated after every five hemolymph measurements in the following solutions (mmol l⁻¹): NH₄⁺: 0.1, 1, 10 NH₄Cl; Na⁺: 30 NaCl+270 LiCl and 300 NaCl; and H⁺: 100 mmol l⁻¹ NaCl and 100 mmol l⁻¹ sodium citrate at pH 6.0, 7.0 and 8.0 (using NaOH or HCl). Voltages were recorded and analyzed in LabChart 6 Pro software (ADInstruments Inc., Sydney, NSW, Australia).

The SIET system protocol used in this study to measure NH₄⁺ and Na⁺ flux at the anal papillae of larvae has been detailed previously (Chasiotis et al., 2016). Larvae were mounted alive in a Petri dish using beeswax and submerged in 4 ml of bath solution (2.5 mmol l⁻¹ NH₄Cl and 5 mmol l⁻¹ NaCl in ddH₂O), leaving the anal papillae exposed and immobilized for measurements. ISMEs were calibrated after every four larvae measurements in 0.1, 1 and 10 mmol l⁻¹ NaCl in ddH₂O for Na⁺ and 0.1, 1 and 10 mmol l⁻¹ NH₄Cl for NH₄⁺ in ddH₂O. Voltage gradients over an excursion distance of 100 µm were recorded adjacent to the papillae with both Na⁺ and NH₄⁺ microelectrodes (see ISME above) at the exact same site at the same time. The sampling protocol utilized here has been outlined previously (Chasiotis et al., 2016). Readings were taken along the lateral to distal portion of the anal papillae at five equidistant sites. Background voltage gradients were taken 3000 µm away from the anal papillae using the same sampling protocol and were subtracted from the voltage gradients recorded at the papillae. To calculate Na⁺ and NH₄⁺ flux, voltage gradients were used to calculate Na⁺ and NH₄⁺ concentration gradients, and the Na⁺ and NH₄⁺ flux was then calculated using the concentration gradients with Fick's law of diffusion. For SIET measurements, a single biological replicate is defined as the average flux from four repeated measurements at each of five sites along a single anal papilla from a single larva.

The NKA and VA activity assay used in this study was adapted from McCormick (1993) and has been outlined previously in detail (Jonusaite et al., 2011). This assay is dependent on the enzymatic coupling of NKA inhibitor ouabain (Sigma-Aldrich) or VA inhibitor bafilomycin (LC Laboratories, Woburn, MA, USA)-sensitive hydrolysis of adenosine triphosphate (ATP) with the oxidation of reduced nicotinamide adenine dinucleotide (NADH), with NADH disappearance being directly measured in a microplate spectrophotometer. Briefly, anal papillae from 30 larvae were collected in 1.5 ml centrifuge tubes, flash-frozen in liquid nitrogen and stored at −80°C. Samples were later thawed on ice, and 100 µl of homogenizing buffer (four parts SEI, composed of 150 mmol l⁻¹ sucrose, 10 mmol l⁻¹ Na₂EDTA, 50 mmol l⁻¹ imidazole, pH 7.3, and one-part SEID, composed of 0.5% deoxycholic acid in SEI) was added to each tube. Samples were sonicated on ice for 10 s at 5 W using an XL 2000 Ultrasonic Processor (Qsonica), and were then centrifuged at 10,000 g for 10 min at 4°C. The supernatants of each sample (100 µl) were transferred to 1.5 ml centrifuge tubes and stored at −80°C.

where ΔATPase is the difference in ATP hydrolysis in the absence and presence of inhibitors ouabain or bafilomycin, S is the slope of the ADP standard curve and [P] is the protein concentration of the sample (Bradford assay, Sigma-Aldrich). The final ATPase activity was expressed as µmoles of ADP mg⁻¹ protein h⁻¹.

Statistical analyses for all experiments were computed using Prism^® 7.00 (GraphPad Software, La Jolla, CA, USA), detailed in the caption of each figure. For analysis of multiple groups using a one-way ANOVA, pairwise multiple comparisons were performed using the Tukey multiple comparisons test, unless specified otherwise, and the adjusted P-value is indicated.

Changes in protein expression of ammonia transporters in the anal papillae were examined in response to rearing of A. aegypti larvae in 5 mmol l⁻¹ NaCl (Fig. 1). AeAmt1 and AeAmt2 protein abundance in the anal papillae did not differ between larvae reared in deH₂O and 5 mmol l⁻¹ NaCl (Fig. 1A,B). In contrast, a 5-fold increase in protein abundance of AeRh50s in the anal papillae was observed compared with deH₂O control levels in response to rearing in 5 mmol l⁻¹ NaCl (Fig. 1C; P=0.0335).

Na⁺ and NH₄⁺ flux at the anal papillae of fourth instar larvae reared in deH₂O, 5 mmol l⁻¹ NaCl or 5 mmol l⁻¹ NH₄Cl (HEA) was measured using SIET (Fig. 2). While there was an apparent trend towards decreased Na⁺ absorption at the anal papillae of larvae reared in NaCl and an apparent reversal of Na⁺ flux for larvae reared in NH₄Cl, compared with the deH₂O control, no statistically significant change in Na⁺ flux was observed in response to the respective rearing conditions (Fig. 2). In contrast, A. aegypti larvae reared in 5 mmol l⁻¹ NaCl had significantly higher NH₄⁺ efflux rates (∼7-fold increase) compared with the average flux measured for deH₂O-reared larvae (Fig. 2). Larvae reared in 5 mmol l⁻¹ NH₄Cl showed similar NH₄⁺ flux at the anal papillae to that of deH₂O control larvae (Fig. 2).

The activity of the primary ionomotive pumps NKA and VA in the anal papillae of fourth instar larvae reared in deH₂O, 5 mmol l⁻¹ NaCl or 5 mmol l⁻¹ NH₄Cl was examined (Fig. 3). The activities of NKA and VA in the anal papillae of larvae reared in NaCl did not change when compared those in the anal papillae of control deH₂O-reared larvae. In contrast, NKA activity in the anal papillae was significantly reduced in larvae reared in HEA compared with activity levels in deH₂O and NaCl (Fig. 3; P=0.0357). Furthermore, VA activity in the anal papillae of larvae reared in HEA was significantly higher than that in the papillae of larvae reared in deH₂O and NaCl (P≤0.005).

Hemolymph Na⁺, NH₄⁺ and pH levels in the hemolymph of fourth instar A. aegypti larvae reared in deH₂O, 5 mmol l⁻¹ NaCl and 5 mmol l⁻¹ NH₄Cl were measured using ISMEs (Fig. 4). No significant difference in hemolymph Na⁺ levels was observed between larvae reared in deH₂O and NaCl, or deH₂O and HEA (Fig. 4A). However, larvae reared in NaCl had a significantly higher Na⁺ hemolymph concentration than larvae reared in HEA (Fig. 4A). The hemolymph NH₄⁺ concentration decreased in NaCl-reared larvae compared with the deH₂O control and HEA larvae (Fig. 4B). While there was a trend towards increased NH₄⁺ levels in the hemolymph of HEA-reared larvae compared with the deH₂O control, this change was not significant (P=0.0922). Aedes aegypti larvae reared in NaCl and HEA had significantly lower hemolymph pH levels in comparison to larvae reared in deH₂O (Fig. 4C).

The mRNA abundance of NHE3 in the anal papillae of larvae reared in deH₂O, 5 mmol l⁻¹ NaCl and 5 mmol l⁻¹ NH₄Cl was measured (Fig. 5). NHE3 transcript abundance, normalized to ribosomal protein 49 (rp49) mRNA abundance in the anal papillae, did not change in larvae reared in HEA in comparison with larvae reared in deH₂O (Fig. 5). A significant decrease in NHE3 mRNA was observed in larvae reared in NaCl compared with those reared in deH₂O (an approximate 75% reduction) and HEA.

This study describes the novel physiological mechanisms used by A. aegypti larvae to accomplish ammonia excretion, acid–base balance and hemolymph ion homeostasis when faced with varying aquatic environments. Based on our previous work, we hypothesized that ammonia excretion at the anal papillae through ammonia transporters could be utilized for acid–base regulation. We found that hemolymph acidification was a common consequence of exposing larvae to HEA or moderate levels of NaCl. This was coupled to alterations in ammonia transporter expression in the anal papillae, and in the case of NaCl exposure, elevated ammonia excretion rates at the anal papillae. The results support the hypothesis that ammonia excretion at the anal papillae may be used to excrete acid when hemolymph acidification occurs.

In a recent study, we examined the effects of rearing A. aegypti larvae in HEA on the expression of ammonia transporters in the anal papillae (Durant and Donini, 2018). We reported a decrease in the protein abundance of apical AeAmt2 and the AeRh50s, while the basolateral AeAmt1 protein abundance did not change compared with that of control larvae reared in deH₂O. While we do not know whether NHE3 protein abundance is altered, here we add that the mRNA abundance of apical NHE3 remains unchanged in response to rearing in HEA (Figs 5), while the activity of basolateral NKA decreases and that of the apical VA increases in the anal papillae of HEA-reared larvae (Fig. 3). Although the transporters involved in ammonia excretion and ion transport in the anal papillae of mosquito larvae are similar to those found in other aquatic invertebrates and vertebrates (e.g. fish), the alterations in expression reported here demonstrate differences in how these transporters are utilized for ammonia excretion. For example, in the freshwater goldfish and common carp gills, exposure to HEA results in an increase in apical Rhcg mRNA expression and NKA activity (Sinha et al., 2013, 2016). Similar responses have been observed in a freshwater shrimp and flatworm, where NKA activity increases and the expression of a Rh protein increases (Pinto et al., 2016; Weihrauch et al., 2012a). These responses are the opposite of those observed in the anal papillae of mosquito larvae in the present study, which are consistent with findings in the freshwater leech, whereby Rh protein expression and NKA activity both decrease upon HEA exposure (Quijada-Rodriguez et al., 2015). A decrease in expression of apical AeAmt2 and AeRh50s may be a response to limit the entry of ammonia from the water in the face of a gradient that no longer favors excretion into the water. In fact, we measured a hemolymph ammonia level of ∼1.4 mmol l⁻¹ in HEA-exposed larvae (Fig. 4B), which is considerably lower than the external HEA treatment of 5 mmol l⁻¹ and would therefore favor ammonia entry; however, it must be noted that we did not measure cytosolic ammonia levels in the syncytial anal papillae epithelium. In the marine crab Metacarcinus magister, a 7 day exposure to 1 mmol l⁻¹ ammonia caused an increase in hemolymph ammonia to levels close to 1 mmol l⁻¹ and a similar response was observed for the freshwater leech Nephelopsis obscura (Martin et al., 2011; Quijada-Rodriguez et al., 2015). It was suggested that this increase in hemolymph ammonia permits ammonia excretion in HEA by minimizing the ammonia gradient that would otherwise lead to ammonia entry. This is in contrast to our findings in the mosquito larvae, which appear to maintain hemolymph ammonia levels that are lower than the external ammonia levels. The unaltered basal expression of AeAmt1, which has been proposed to transport NH₄⁺ from hemolymph to cytosol and is driven by the cytosol negative potential generated by basal NKA and apical VA, could maintain cytosol ammonia levels higher than those in the hemolymph, thus supporting ammonia excretion across the apical membrane (Chasiotis et al., 2016). In this case, the electrical potential is more important than the ammonia concentration gradient. Although NKA activity decreases in HEA, the elevated VA activity may compensate to maintain the cytosol negative potential in the anal papillae. Ammonia excretion is still evident in papillae of HEA-exposed larvae because ammonia excretion rates similar to those recorded from papillae of control larvae were measured in water with 2.5 mmol l⁻¹ NH₄Cl (and 5 mmol l⁻¹ NaCl), which represents an inward-directed ammonia gradient relative to the hemolymph ammonia levels (Fig. 2).

A reduction in Na⁺ uptake at the anal papillae in a bath containing 5 mmol l⁻¹ NaCl and 2.5 mmol l⁻¹ NH₄Cl was previously demonstrated (Weihrauch et al., 2012b). Larvae exposed to HEA showed a trend towards a lower hemolymph [Na⁺] (Fig. 4A, P=0.059, Tukey's multiple comparisons test) as well as a lower hemolymph pH compared with control larvae (Fig. 4C). Similar findings were reported for some crustaceans exposed to HEA (Chen and Chen, 1996; Harris et al., 2001; Young-Lai et al., 1991). The prevailing theory for a reduction in hemolymph [Na⁺] caused by ammonia exposure is the impairment of an apical Na⁺/NH₄⁺ transport system (Chen and Chen, 1996; Harris et al., 2001; Romano and Zeng, 2007), which may involve a transport metabolon composed of VA, NHE3 and Rh proteins (Shih et al., 2012, 2013; Wright and Wood, 2009; Wu et al., 2010). The findings of a trend towards reduced hemolymph [Na⁺] in HEA larvae, coupled with the observation that NH₄⁺ fluxes from anal papillae in a bath containing NaCl are statistically similar to those measured from control larvae, do not support a mechanism of Na⁺/NH₄⁺ exchange at the anal papillae. Furthermore, papillae from HEA-exposed larvae showed a mean Na⁺ excretion in tandem with NH₄⁺ excretion, suggesting that ammonia excretion and Na⁺ uptake mechanisms are not linked when larvae are reared in HEA (Fig. 2). Additionally, the loss of Na⁺ at the anal papillae is likely responsible for the apparent decrease in hemolymph [Na⁺] and the decrease in NKA activity in HEA, which is important for Na⁺ uptake, also suggests that Na⁺/NH₄⁺ exchange is unlikely at the anal papillae.

Lastly, the lower hemolymph pH of larvae reared in HEA could be the result of deprotonation of NH₄⁺ whereby NH₃ diffuses across cellular membranes and H⁺ remains in the hemolymph, leading to acidification. The NH₄⁺ levels in the hemolymph of HEA-exposed larvae were ∼1.4 mmol l⁻¹ compared with levels in control larvae of ∼1 mmol l⁻¹ (P=0.09 Tukey's multiple comparisons test). This apparent elevation in hemolymph ammonia may have been sufficient to acidify the hemolymph. However, the relative amount of deprotonation of NH₄⁺ would be minimal at a pH of ∼7.7. An alternative explanation is that larvae intentionally decrease their hemolymph pH from ∼7.9 to ∼7.7 in order to limit deprotonation of the elevated NH₄⁺ in the hemolymph. This has been suggested in copepods that accumulate high levels of ammonia in the hemolymph during diapause (Schründer et al., 2013). Another possible explanation is that the acidification of the hemolymph is a consequence of the downregulation of AeRh50 expression. A previous study noted that RNAi-mediated knockdown of AeRh50-1 in the anal papillae resulted in hemolymph acidification (Durant et al., 2017). This is similar to the effect of HEA rearing whereby AeRh50 expression is downregulated and hemolymph acidification occurs (Durant and Donini, 2018). We previously speculated that this acidification could arise if AeRh50-1 is capable of transporting CO₂ and is located on the basal membrane of the anal papillae epithelium. If this were the case, then downregulation of AeRh50-1 in response to HEA would result in a buildup of bicarbonate and protons in the hemolymph; however, at this time there is no conclusive evidence that Rh proteins transport CO₂ and studies to determine the transport substrate of these transporters continue to be very important.

Rearing of A. aegypti larvae in 5 mmol l⁻¹ NaCl did not affect the abundance of either AeAmt1 or AeAmt2 protein in the anal papillae but increased the abundance of AeRh50 protein by 5 times the levels detected in deH₂O-reared larvae (Fig. 1). These observations on the expression of ammonia transporters in the anal papillae coincide with significantly higher NH₄⁺ excretion from the anal papillae of NaCl-reared larvae and decreased hemolymph NH₄⁺ levels, suggesting that the elevated ammonia excretion occurs through AeRh50 proteins. As previously mentioned, Rh proteins are proposed gas channels which were shown to transport NH₃ by recruitment of NH₄⁺ and subsequent deprotonation (Baday et al., 2015; Kustu and Inwood, 2006). We propose that the electrical component of the electrochemical gradient drives NH₄⁺ into the cytosol of the anal papillae epithelium through AeAmt1 and that ammonia exits primarily through an apical AeRh50 protein, which can recruit NH₄⁺ that is accumulating in the cytosol and transport NH₃ to the water. The resulting protons in the cytosol can be removed through VA, which functions with the apical AeRh50 in an ammonia-trapping mechanism. In the present study, the observed increase in ammonia excretion and AeRh50 expression at the anal papillae of NaCl-reared larvae was unexpected; however, it may be explained by ammonia excretion being utilized to remove acid from the hemolymph. This has been proposed for crustaceans (Fehsenfeld and Weihrauch, 2017).

Acidification of the hemolymph in aquatic animals exposed to salt is thought to occur because of a downregulation of Na⁺ uptake, which is coupled with H⁺ secretion (Henry and Cameron, 1982; Maxime et al., 1990; Truchot, 1981; Truchot and Nonnotte, 1990). This acidification of the hemolymph was shown to occur in A. aegypti larvae that were transferred to relatively high salt levels (Donini et al., 2007). Evidence that Na⁺ uptake is associated with H⁺ secretion in A. aegypti has been presented (Patrick et al., 2002b; Stobbart, 1967) and we have shown that Na⁺ uptake at the papillae is likely to occur through a putative Na⁺ channel driven by the electrical gradient established by the apical VA, which secretes H⁺ into the water (Del Duca et al., 2011). We have since shown the expression of NHE3 on the apical side of the anal papilla epithelium and cannot rule out the possibility of Na⁺ uptake coupled to H⁺ secretion via this transporter (Chasiotis et al., 2016). Although there is a trend towards a decrease in Na⁺ uptake at the anal papillae of NaCl-reared larvae, the VA activity was no different from that found in control larvae. The mRNA abundance of the NHE3 was reduced to ∼1/4 the levels in controls and although it is tempting to speculate that the trend towards a decrease in Na⁺ uptake and the acidification of the hemolymph may be the result of reduced NHE3 expression at the anal papillae, the protein levels of NHE3 need to be measured.

The acidification of the hemolymph in NaCl-reared larvae is also likely to release more CO₂ in the hemolymph because of the buffering properties of bicarbonate. As a result, the larvae would require an outlet for the increased CO₂ levels in the hemolymph. The increased expression of AeRh50s in the anal papillae might provide a means to eliminate the excess CO₂ if these transporters were capable of CO₂ transport. In fish, algae and bacteria there is some evidence that Rh proteins may be capable of transporting CO₂ (Perry et al., 2010; Soupene et al., 2002, 2004). If A. aegypti Rh proteins are able to transport CO₂, then they may also be utilized for acid–base balance in this manner; however, more research is required to ascertain the transport substrates of these transporters.

In summary, in HEA, the apical ammonia transporters in the anal papillae of A. aegypti appear to be detrimental and are downregulated in order to limit ammonia entry into the cytosol, whereas basal AeAmt1 expression remains unaltered, with NH₄⁺ excretion being maintained against an inward gradient. In moderate levels of salt, we propose that the main physiological challenge of A. aegypti larvae is hemolymph acidosis, similar to other aquatic organisms (Fehsenfeld and Weihrauch, 2017). We show that elevated AeRh50 protein abundance coupled with an increase in excretion of NH₄⁺ at the anal papillae and a hemolymph acidosis occur in response to salt, suggesting an important role for Rh proteins in acid–base regulation. These findings provide evidence for the hypothesis that NH₄⁺ excretion at the anal papillae is utilized to remove acid from the hemolymph.

We thank Dr Scott Kelly for use of facilities to run enzyme assays.