Mosquito natriuretic peptide identified as a calcitonin-like diuretic hormone in Anopheles gambiae (Giles)

Female mosquitoes ingest blood meals that are equivalent to more than twice their unfed body mass, and engorged animals can fly only with great difficulty, making them more prone to predation(Roitberg et al., 2003). The blood meal serves primarily as a source of protein for egg development, but it also contains considerable amounts of unwanted salts and water that threaten haemolymph homeostasis. To counter this, a rapid natriuresis and diuresis commences even before the blood meal is completed, and ∼40% of the imbibed plasma volume and salt are voided within 1–2 h of feeding(Williams et al., 1983). Indeed, within 20 min of completing the meal, a volume of fluid equal to that of the haemolymph of an unfed insect is absorbed from the midgut and expelled through the anus.

Endocrine control of this post-prandial diuresis has been extensively studied in the yellow fever mosquito, Aedes aegypti(Beyenbach, 2003). Between blood meals, the female conserves water and voids little or no urine, but,during the peak phase of diuresis, clear drops of NaCl-rich urine are eliminated from the anus every 12–15 s(Coast et al., 2002; Williams et al., 1983). The stimulus for this diuresis and natriuresis is mosquito natriuretic peptide(MNP), which is released into the haemolymph from structures within the head when the insect takes a blood meal(Beyenbach and Petzel, 1987). MNP acts via cyclic AMP to stimulate secretion of Na⁺-rich primary urine by the five Malpighian (renal) tubules. Its effects are duplicated by a membrane permeant analogue of cyclic AMP (dibutyryl-cyclic AMP), which has been shown to depolarise the basolateral membrane voltage(V_b) of Malpighian tubule principal cells and to hyperpolarise the transepithelial voltage (V_t) by a similar amount (Beyenbach,2003; Sawyer and Beyenbach,1985). These changes are accompanied by reductions in transepithelial resistance and in the fractional resistance of the basolateral membrane and have been attributed to an increase in the Na⁺conductance of the basolateral membrane(Beyenbach, 2003). In addition,dibutyryl-cyclic AMP activates a bumetanide-sensitive Na⁺/K⁺/2Cl^– cotransporter(Hegarty et al., 1991). The actions of cyclic AMP make more Na⁺ available for transport into the tubule lumen via cation/proton antiports in the principal cell apical membrane. This additional Na⁺ (along with Cl^– as a counterion) is accompanied by an osmotically equivalent volume of water, and tubule secretion is elevated >7-fold while the Na⁺:K⁺ concentration([Na⁺]:[K⁺]) ratio of the secreted urine increases from unity to approximately 10 (Beyenbach,2003).

MNP is thought to belong to the corticotropin releasing factor(CRF)-related family of insect diuretic hormones, which use cyclic AMP as their second messenger (Beyenbach,2003). In support of this, a CRF-related diuretic hormone(Culsa-DH) from the salt-water mosquito Culex salinarius stimulates cyclic AMP production by Ae. aegypti Malpighian tubules(Cady and Hagedorn, 1999b) and has diuretic and natriuretic activity(Clark et al., 1998b), although the latter effects are small in comparison with those of MNP and exogenous cyclic AMP. Additionally, Culsa-DH causes a biphasic change in the transepithelial voltage of isolated perfused Malpighian tubules(Clark et al., 1998a), which resembles the response to MNP, and stimulates urine production in vivo (Cady and Hagedorn,1999a).

Among the regulatory peptide genes identified in the genome of the malaria mosquito Anopheles gambiae (Holt et al., 2002; Riehle et al.,2002) are those encoding orthologues of peptides known to have diuretic activity, namely CRF-related DH, myokinins, calcitonin (CT)-like DH,CAP_2b (capa) and tachykinins(Schooley et al., 2005). Of particular interest for the identification of MNP are the CRF-related and CT-like peptides (Anoga-DH₄₄ and Anoga-DH₃₁,respectively) since their Drosophila melanogaster homologues act via cyclic AMP in stimulating Malpighian tubule secretion(Cabrero et al., 2002; Coast et al., 2001). Here, we describe the effects of synthetic Anoga-DH₄₄ and Anoga-DH₃₁ on fluid and electrolyte (Na⁺ and K⁺) transport by Malpighian tubules of female An. gambiaeand Ae. aegypti. We show that the CT-like diuretic hormone is the mosquito natriuretic factor described by Petzel et al.(1985), later designated MNP(Beyenbach and Petzel,1987).

Newly emerged females of the G3 strain of An. gambiae (Giles) were obtained from the London School of Hygiene and Tropical Medicine (LSHTM). This strain originated from McCarthy Island in Gambia (West Africa) and was colonised at LSHTM in 1975. Mosquitoes were reared according to methods developed at the LSHTM. The LSHTM also supplied newly emerged females of their AeAe strain of Aedes aegypti (L.). Adult mosquitoes were held at 27°C under a 16 h:8 h light:dark regime and fed a 2.5% solution of glucose.

Malpighian tubules were removed from adult females at 3–10 days post-emergence. Insects were anaesthetised by chilling at 4°C and decapitated prior to removing the Malpighian tubules under Aedessaline with the following composition (in mmol l^–1): NaCl,150; NaHCO₃, 1.8; KCl, 3.4; CaCl₂, 1.7;MgSO₄, 4.1; glucose, 5; Hepes(N-2-hydroxyethylpiperazine-N′2-ethanesulphonic acid),25; pH adjusted to 7.1 with 1 mol l^–1 NaOH(Clark et al., 1998b). Tubule secretion assays were performed as described by Clark et al.(1998b) but using 10 μl drops of bathing fluid. A number of different salines were tested for their ability to support fluid secretion (results not shown), and the bathing fluid selected for An. gambiae Malpighian tubules was one that comprised a 1:1 mixture of K⁺-free Drosophila saline(O'Donnell et al., 1996) and Schneider's insect medium (Gibco™; Invitrogen Ltd, Paisley, UK). The Drosophila saline had the following composition (in mmol l^–1): NaCl, 117.5; NaHCO₃, 10.2;NaH₂PO₄, 4.3; CaCl₂, 2.0; MgCl₂,8.5; Hepes, 8.6, glucose, 20.0; pH adjusted to 7.0 with NaOH. Test substances were dissolved in the bathing fluid at twice their desired final concentration, and 5 μl were added to the incubation medium, having first removed a 5 μl sample. Similar procedures were used for Ae. aegypti Malpighian tubules, but the bathing medium comprised a 1:1 mixture of Aedes saline and Schneider's insect medium.

Electrolyte concentrations (Na⁺ and K⁺) in urine droplets held beneath water-saturated paraffin oil were measured using ion-selective microelectrodes (Coast et al., 2001; Ianowski and O'Donnell, 2004). Potassium electrodes were based on potassium ionophore I, cocktail A (Fluka, Buchs, Switzerland) and were backfilled with 500 mmol l^–1 KCl. Sodium electrodes were based on sodium ionophore II, cocktail A (Fluka) and were backfilled with 500 mmol l^–1 NaCl. For both measurements, the reference electrode was filled with 1 mol l^–1 LiCl. Potassium electrodes (mean slope 56.1±0.4 mV; N=36) were calibrated in mixed solutions of 200 mmol l^–1 KCl and NaCl, whereas Na⁺ electrodes(mean slope 54.0±0.3 mV; N=26) were calibrated in mixed solutions of 200 mmol l^–1 NaCl and LiCl. Potassium is known to interfere with Na⁺ measurements and this was corrected for as described by Ianowski and O'Donnell(2004).

Isolated Malpighian tubules were attached to poly-d-lysine coated cover slips before being transferred to a small (250 μl) chamber that was perfused at 1 ml min^–1 with Aedes saline containing the following amino acids (in mmol l^–1): glycine,1.7; l-proline, 7.0; l-glutamine, 6.16; l-histidine, 0.95; l-leucine, 0.55; l-lysine,4.5; l-valine, 1.3. Use of this amino-acid-replete saline, which contains amino acids at the same concentrations as those in the bathing fluid for the fluid secretion assay, resulted in more stable voltages and facilitated the placement of a microelectrode in the tubule lumen. Perfusion was stopped prior to the addition of test compounds and restarted to wash-off. Microelectrodes (120–140 MΩ resistance when filled with 0.1 mol l^–1 KCl) were fabricated from 1 mm o.d. filament glass tubing(GC100F-75; Clark Electromedical Instruments, Pangbourne, UK) using a microprocessor-controlled vertical pipette puller (World Precision Instruments, Sarasota, FL, USA). They were backfilled with 0.1 mol l^–1 KCl and connected to a high-impedance electrometer(M-707A; World Precision Instruments) via an Ag/AgCl half-cell (World Precision Instruments). An Ag/AgCl reference electrode was placed in a 3 mol l^–1 KCl reservoir connected to the perfusion chamber via a Ringer agar bridge. Basolateral membrane(V_b) and transepithelial voltages (V_t)were measured close to the distal (closed) end of tubules, which were observed using an inverted phase contrast microscope (Nikon TMS, Tokyo, Japan). Large principal cells were selected for recording V_b, and microelectrodes were advanced in steps of 3–7 μm at an oblique angle using a piezoelectric drive (PM-10; World Precision Instruments) until a sudden jump in potential indicated that the basolateral membrane had been impaled. Recordings were deemed successful if the potential remained stable(±2 mV) for >30 s and returned to 0±2 mV after withdrawal of the electrode. Similar criteria were adopted for recording V_t after the microelectrode had been advanced through the apical membrane into the tubule lumen. Results were recorded digitally using a data acquisition system (Datacan V; Sable Systems, Henderson, NV, USA).

Production of cyclic AMP and cyclic GMP by isolated Malpighian tubules was measured as described by Coast et al.(2001). Malpighian tubules were transferred to Eppendorf tubes (5 or 10 per tube) containing Drosophila saline. After incubating for 10 min at room temperature,saline containing 5 mmol l^–1 3-isobutyl-1-methylxanthine(IBMX; final concentration 0.5 mmol l^–1) was added to each tube. This was followed 10 min later by the addition of IBMX saline alone(controls) or IBMX saline containing either Anoga-DH₃₁ or Anoga-DH₄₄ (final concentration 1 μmol l^–1). The IBMX solution was freshly prepared from a 250 mmol l^–1stock solution in DMSO (final DMSO concentration 0.2%). The incubation was terminated after 0–10 min by quenching the tubes in liquid nitrogen. Samples were stored at –80°C until shipped on dry ice to the University of North Wales for the measurement of cyclic AMP and cyclic GMP by radioimmunoassay (RIA; Coast et al.,2001). The cyclic AMP assay was modified by the inclusion of 10μl crustacean saline (Webster,1986) in the RIA, which improves the slope of the standard curve(S.G.W., unpublished observation).

The An. gambiae CT-like diuretic hormone was located using the Ensembl genome browser with Drome-DH₃₁(Coast et al., 2001) as a query. The 31-residue peptide Anoga-DH₃₁ has the predicted sequence TVDFGLSRGYSGAQEAKHRMAMAVANFAGGP-NH₂ (84% identical to Drome-DH₃₁) and is encoded on a 95 amino acid precursor peptide with accession number XM_321755. The An. gambiae CRF-related diuretic hormone, a 44-residue peptide with the predicted sequence TKPSLSIVNPLDVLRQRIILEIARRQMRENTRQVELNKALLREI-NH₂ (82% identical with Drome-DH₄₄) was also found with Ensembl using Drome-DH₄₄ (Cabrero et al.,2002) as a query. There is an intron in the encoding region for Anoga-DH₄₄, located between Glu³³ and Val³⁴;the gene encoding it is incomplete on both ends, probably due to sequencing errors. Both peptides were synthesized using Nα-9-fluorenylmethoxycarbonyl (FMOC) chemistry with an Applied Biosystems 431A synthesiser (Foster City, CA, USA). For each synthesis, Rink MBHA amide resin (Novabiochem, San Diego, CA, USA) was used on a 0.1 mmol scale. We activated FMOC–amino acids with 1-hydroxybenzotriazole in 1-methyl-2-pyrrolidinone in the presence of dicyclohexylcarbodiimide for amino acid activation, with a 10-fold molar excess of acylating species. Six 22 min coupling cycles were employed (King,1996; http://www.abrf.org/ABRFNews/1966/December1996/Long_PepSyn.html);no double couplings were performed. Protecting groups were Arg-[2,2,4,6,7-pentamethylbenzo-(2,3-dihydro)-furansulphonyl], Asn(Trityl),Asp(OtBu), Gln(Trityl), Glu(OtBu), His(Trityl), Lys(Boc), Ser(OtBu), Thr(OtBu)and Tyr(OtBu). The dry resin–peptide was cleaved using Reagent K(King et al., 1990). Crude peptides were precipitated and then washed with methyl-t-butyl ether. The dried crude peptides were purified to homogeneity with a ThermoSeparations P4000 liquid chromatograph (San Jose, CA, USA). Crude peptide was first purified by cation exchange on a Polysulfoethyl A column (25×2.1 cm;Poly LC, Columbia, MD, USA) eluted at 10 ml min^–1 using 20 mmol l^–1 sodium acetate, pH 4.5, and a gradient from 0 to 1 mol l^–1 NaCl. This removed Arg deletion peptides, which are difficult to separate by reversed phase. Purified fractions were re-purified with a YMC 25×2.0 cm C₈ column (Waters Co., Milford, MA, USA)eluted at 10 ml min^–1 with an ethanol–water–0.1%trifluoroacetic acid gradient.

Muscakinin (Musdo-K) was a generous gift from Dr R. J. Nachman (USDA,College Station, TX, USA). Unless otherwise stated, all other chemicals were obtained from Sigma-Aldrich.

Tests for significance were performed using GraphPad Instat v.3.06(GraphPad Software, San Diego, CA, USA), with P<0.05 being accepted as significant. Dose–response curves with variable slope were fitted using Prism™ v.4.02 (GraphPad Software).

Prior to examining the effects of Anoga-DH₃₁ and Anoga-DH₄₄ on An. gambiae Malpighian tubules, we determined whether exogenous cyclic AMP had effects similar to those reported in Ae. aegypti. Fluid secretion and tubule fluid Na⁺ and K⁺ concentrations were measured over 30 min under control(unstimulated) conditions and then at 10–15 minintervals in the presence of a membrane permeant cyclic AMP analogue (8-bromo-cyclic AMP; 1 mmol l^–1). The cyclic AMP analogue stimulated Malpighian tubule secretion 10-fold, and the [Na⁺]:[K⁺] ratio of the primary urine increased from 0.7 to 7.5(Fig. 1A). Transepithelial Na⁺ transport (the product of ion concentration and rate of secretion) increased >18-fold compared with only a 2.6-fold increase for K⁺, which declined after 10 min(Fig. 1B). The effect of 8-bromo-cyclic AMP on ion and fluid transport by An. gambiaeMalpighian tubules thus reproduced the diuretic and natriuretic activities of the membrane permeant dibutyryl analogue of cyclic AMP in Ae. aegypti.

Likewise, the effects of 8-bromo-cyclic AMP on Malpighian tubule electrophysiology in An. gambiae were similar to those reported for dibutyryl-cyclic AMP in Ae. aegypti. Absolute values for V_t (33.7±3.7 mV; N=28) and principal cell V_b (–68.5±1.7 mV; N=54) varied considerably between tubules under control conditions. The addition of 100μmol l^–1 8-bromo-cyclic AMP resulted in an immediate depolarisation of V_b by 55.8±5.3 mV (N=9)while V_t hyperpolarised by 39.1±10.8 mV(N=5; Fig. 2). These values are not significantly different (P=0.142; unpaired t-test), which suggests that the apical membrane voltage(V_a=V_t–V_b)remains relatively constant. However, given the variability of V_t and V_b, which were measured in different Malpighian tubules, a small change in V_a might have gone undetected.

Fluid secretion and tubule fluid Na⁺ and K⁺concentrations were measured over 30 min under control conditions and then in the presence of 1 μmol l^–1 of either Anoga-DH₄₄or Anoga-DH₃₁. After 60 min, all tubules were challenged with 1 mmol l^–1 8-bromo-cyclic AMP to establish maximum rates of ion and fluid transport. Tubule secretion was increased 3-fold by Anoga-DH₄₄ (Fig. 3A)but then more than doubled after the addition of 8-bromo-cyclic AMP. Importantly, the [Na⁺]:[K⁺] ratio of the secreted urine fell from 0.70±0.09 to 0.45±0.04 (N=5; P<0.05) in response to Anoga-DH₄₄ but increased to 2.98±0.30 after the addition of 8-bromo-cyclic AMP(Fig. 3A). This reflects the selective stimulation of transepithelial Na⁺ transport by 8-bromo-cyclic AMP, but not by Anoga-DH₄₄(Fig. 3B).

Fig. 4 shows the results of the same experiment with Anoga-DH₃₁. Fluid secretion was increased 10-fold by Anoga-DH₃₁ and was not stimulated further by the addition of 8-bromo-cyclic AMP (Fig. 4A). The response to Anoga-DH₃₁ was associated with a marked increase in the [Na⁺]:[K⁺] ratio of the secreted urine, from 0.64±0.08 (N=5) to 5.91±1.69. The ratio increased to 6.30±1.08 in the presence of 8-bromo-cyclic AMP, but this was not significantly different (P=0.604; paired t-test)from that obtained with Anoga-DH₃₁ alone. Thus, in marked contrast to Anoga-DH₄₄, the CT-like diuretic hormone selectively stimulated transepithelial Na⁺ transport, which was increased 16-fold compared with only a 2.4-fold increase for K⁺(Fig. 4B).

To demonstrate further the different effects of the two diuretic hormones on ion and fluid transport, a group of Malpighian tubules were challenged over 30 min intervals with first Anoga-DH₄₄ alone and then together with Anoga-DH₃₁; the final concentration of both peptides was 1 μmol l^–1. Under control conditions, tubule secretion was 0.36±0.01 nl min^–1 (N=5) and the[Na⁺]:[K⁺] ratio of the secreted urine was 1.17±0.36. Anoga-DH₄₄ produced a 3-fold increase in fluid secretion but had no effect on the urine [Na⁺]:[K⁺]ratio, whereas the ratio increased >7-fold after the addition of Anoga-DH₃₁ to tubules already exposed to Anoga-DH₄₄, and there was a further 2-fold increase in the rate of secretion(Fig. 5). The change in[Na⁺]:[K⁺] ratio reflects the markedly different effects of the two peptides on transepithelial Na⁺ and K⁺transport. In the presence of Anoga-DH₄₄, Na⁺ transport increased from 35±4 pmol min^–1 to a maximum of 109±17 pmol min^–1 but reached 399±55.2 pmol min^–1 after the addition of Anoga-DH₃₁. On the other hand, K⁺ transport was increased from 36±7 pmol min^–1 to 105±29 pmol min^–1 by Anoga-DH₄₄ but then fell to 47±7 pmol min^–1following the addition of Anoga-DH₃₁.

Fig. 6 shows real-time recordings of the K⁺ concentration in urine secreted by tubules challenged with either Anoga-DH₄₄ or Anoga-DH₃₁. The Na⁺ concentration is calculated on the basis of the sum of[Na⁺] and [K⁺] being virtually constant at 200 mmol l^–1. Following a 30 min equilibration period, a drop of secreted fluid was removed, and a K⁺ selective microelectrode,together with a reference electrode, was positioned in the newly secreted urine. After 2–3 min, time to establish the K⁺ concentration of the urine under control conditions, tubules were challenged with either Anoga-DH₄₄ or Anoga-DH₃₁ (1 μmol l^–1final concentration). The representative recording in Fig. 6A shows the effect of the CRF-related peptide Anoga-DH₄₄. The K⁺ concentration of the secreted urine was initially 75 mmol l^–1 but fell to 65 mmol l^–1 within 10 min of adding the CRF-related diuretic hormone. This reflects a small increase in Na⁺ transport relative to K⁺, but this was not sustained and the K⁺concentration was constant over the next 20 min. Fig. 6B shows similar data for a tubule stimulated with Anoga-DH₃₁. The K⁺concentration of the secreted urine fell exponentially from about 102 mmol l^–1 to 30 mmol l^–1 over 30 min. The reciprocal increase in the Na⁺ concentration of the secreted urine reflects the natriuretic activity of the calcitonin-like diuretic hormone.

Representative recordings of the effects of 100 nmol l^–1Anoga-DH₄₄ and 100 nmol l^–1 Anoga-DH₃₁on the basolateral membrane voltage of Malpighian tubule principal cells are shown in Fig. 7. Addition of Anoga-DH₄₄ resulted in a triphasic change in membrane voltage(Fig. 7A), which initially hyperpolarised by 6.1±1.3 mV (N=15; Phase 1) and then depolarised by 35.4±4.7 mV (Phase 2) before returning to close to its control value in the continued presence of the peptide (Phase 3). The transepithelial voltage changed in a similar manner (data not shown), first depolarising by 13.5±4.1 mV (N=8) and then hyperpolarising by 28.8±6.5 mV before returning to close to the control value. The magnitude of the changes in transepithelial voltage was not significantly different from that in the basolateral membrane voltage during both Phase 1(P=0.124, unpaired t-test with Welch correction) and Phase 2(P=0.418, unpaired t-test) of the response, which suggests that the apical membrane voltage is not affected by Anoga-DH₄₄.

The effect of Anoga-DH₃₁ on the voltage across the basolateral membrane was very different in that the initial hyperpolarising phase (Phase 1) was absent (Fig. 7B). Instead, there was an immediate depolarisation of 41.9±2.5 mV(N=30) followed by a variable degree of repolarisation, which left the membrane depolarised relative to its control value until the peptide was washed off. Similar changes were recorded in the transepithelial voltage (data not shown), which hyperpolarised by 37.2±4.9 mV (N=14) before dropping back to a plateau that was higher than during the control period. The magnitude of the initial change in basolateral and transepithelial voltages did not differ significantly (P=0.347), and the voltage across the apical membrane is therefore likely to be unaffected.

Differences between the actions of Anoga-DH₄₄ and Anoga-DH₃₁ on tubule electrophysiology are highlighted in Fig. 8, which shows the basolateral membrane voltage recorded in a single principal cell challenged sequentially with 100 nmol l^–1 of each of the peptides and then with 100 μmol l^–1 8-bromo-cyclic AMP. The triphasic response to Anoga-DH₄₄ is readily distinguishable by the brief hyperpolarising phase that precedes the depolarisation and by the membrane potential repolarising before the peptide is washed off. The initial hyperpolarisation is absent from the responses to Anoga-DH₃₁ and cyclic AMP, and the membrane remains depolarised until the secretagogues are washed off.

The diuretic activity of Anoga-DH₃₁ was investigated on An. gambiae tubules at concentrations ranging from 1 nmol l^–1 to 1000 nmol l^–1. Fluid secretion was measured over 45 minunder control conditions and then for 30 min in the presence of Anoga-DH₃₁. Finally, 1 mmol l^–18-bromo-cyclic AMP was added to all tubules and fluid secretion measured over a further 30 min. Diuretic activity was calculated as the difference between rates of fluid secretion in the control and experimental periods. The length of tubule within the bathing fluid drop varied considerably and the diuretic activity of Anoga-DH₃₁ was therefore normalised by expressing it as a percentage of the diuretic activity of 1 mmol l^–18-bromo-cyclic AMP.

Fluid secretion was significantly stimulated (P<0.05; paired t-test) by concentrations of Anoga-DH₃₁ of ≥10 nmol l^–1, and the response at 1000 nmol l^–1 did not differ from that obtained with 1 mmol l^–1 8-bromo-cyclic AMP (P=0.564). The apparent EC₅₀ was 50 nmol l^–1 with 95% confidence limits of 41–59 nmol l^–1.

The production of cyclic AMP was measured over 10 min by Malpighian tubules(five per tube) isolated from An. gambiae and incubated in Drosophila saline containing 500 μmol l^–1 IBMX. Cyclic AMP levels in the saline controls were below the level of detection in the RIA, which was 15 fmol tube^–1. However, significant(P<0.01) amounts of cyclic AMP were detected in the presence of 1μmol l^–1 Anoga-DH₃₁, reaching the equivalent of 32.7±7.3 fmol tubule^–1 (N=8). Cyclic AMP levels in tubules challenged with 1 μmol l^–1Anoga-DH₄₄ were more variable (12.7± 7.5 fmol tubule^–1; N=8) and not significantly different from zero (P=0.134). The experiment was therefore repeated with 10 Malpighian tubules per assay tube and incubating for 0, 2, 5 and 10 min in saline containing 500 μmol l^–1 IBMX with or without the addition of 1 μmol l^–1 Anoga-DH₄₄. After incubating for 2 min in the presence of Anoga-DH₄₄, cyclic AMP levels had increased significantly (P<0.001; Mann–Whitney test) from 7.0±1.7 fmol tubule^–1 (N=10) to 54.2±12.7 fmol tubule^–1 (N=5). Cyclic AMP levels were not increased further by incubating for 5 min (32.6± 7.5 fmol tubule^–1; N=5) and 10 min (40.7±7.8 fmol tubule^–1; N=5) but were significantly higher(P<0.01; Mann–Whitney test) than the corresponding controls,which were 2.2±2.2 fmol tubule^–1 (N=5) and 5.4±1.1 fmol tubule^–1 (N=5),respectively.

Following a 10 min incubation in IBMX saline containing either 1 μmol l^–1 Anoga-DH₄₄ or 1 μmol l^–1Anoga-DH₃₁, cyclic GMP levels were equivalent to 0.3±0.1 fmol tubule^–1 (N=4) and 0.3±0.1 fmol tubule^–1 (N=5), respectively, which were not significantly different (P=0.965) from the saline controls(0.3±0.1 fmol tubule^–1; N=6).

Anoga-DH₃₁ has both diuretic and natriuretic activity in An. gambiae, and this CT-like peptide is the most likely candidate for the MNP described in Ae. aegypti by Petzel et al.(1985). We sought to confirm this by testing both peptides on Malpighian tubules from the yellow fever mosquito. Secretion was measured over 30 min under control conditions and for 30 min periods in the presence of either Anoga-DH₄₄ or Anoga-DH₃₁ (final concentration 1 μmol l^–1)alone and then in combination with 1 mmol l^–1 8-bromo-cyclic AMP. The results were very similar to those obtained with An. gambiaeand are summarised in Table 1. Anoga-DH₄₄ stimulated fluid secretion but had no effect on the[Na⁺]:[K⁺] ratio of the secreted urine. Both parameters were increased >3-fold after the addition of 8-bromo-cyclic AMP, reflecting a dramatic rise in transepithelial Na⁺ transport, whereas K⁺ transport was unchanged. In marked contrast, the CT-like peptide Anoga-DH₃₁ stimulated fluid secretion and transepithelial Na⁺ transport to the same extent as exogenous cyclic AMP. To further illustrate the differing effects of the two peptides, one batch of tubules was challenged with 1 μmol l^–1Anoga-DH₄₄ alone and then in combination with 1 μmol l^–1 Anoga-DH₃₁(Table 1). As shown previously,Anoga-DH₄₄ increased fluid secretion but had an insignificant effect on the urine [Na⁺]:[K⁺] ratio. Fluid secretion was further accelerated by the addition of Anoga-DH₃₁ and this was associated with a marked increase in the [Na⁺]:[K⁺]ratio of the secreted fluid, which reflected the specific stimulation of transepithelial Na⁺ transport.

We have shown that both Anoga-DH₄₄ and Anoga-DH₃₁stimulate cyclic AMP production by Malpighian tubules isolated from An. gambiae but that only the CT-like peptide has pronounced natriuretic activity. Conceivably, this might be because Anoga-DH₃₁ is more effective than Anoga-DH₄₄ in elevating intracellular levels of cyclic AMP. To investigate this, the peptides were tested on Malpighian tubules after cyclic AMP phosphodiesterase activity had been inhibited with IBMX in order to elevate intracellular levels of the second messenger. The results are shown in Fig. 9. Tubule secretion rose during the 50 min incubation with 100 μmol l^–1 IBMX, while the [Na⁺]:[K⁺] ratio of the secreted urine was unchanged. The addition of 1 μmol l^–1 Anoga-DH₄₄ to the IBMX saline resulted in a further increase in tubule secretion but had no significant effect(P>0.05; Tukey–Kramer Multiple Comparisons Test) on the[Na⁺]:[K⁺] ratio. However, with the inclusion of Anoga-DH₃₁ (final concentration 1 μmol l^–1) in the bathing fluid, the urine [Na⁺]:[K⁺] ratio increased dramatically, which, taken together with the further acceleration of tubule secretion, represented a 9-fold increase in Na⁺ transport (from 17.1±2.1 pmol min^–1 to 180.2± 31.1 pmol min^–1) compared with only a 2-fold increase in K⁺transport (from 64.5±8.7 pmol min^–1 to 114.2±15.0 pmol min^–1).

The CRF-related peptides Anoga-DH₄₄ and Culsa-DH initially depolarise the transepithelial voltage of Malpighian tubules from An. gambiae (present study) and Ae. aegypti(Clark et al., 1998a),respectively. Clark et al.(1998a) attribute this initial depolarisation to the Ca²⁺-dependent opening of a paracellular Cl^– conductance, which characterises the response to the kinin family of insect diuretic peptides and results in a non-selective increase in Na⁺ and K⁺ transport(Beyenbach, 1995). Conceivably,this non-selective response to activation of the Cl^–conductance pathway could outweigh a cyclic AMP-dependent stimulation of Na⁺ transport and thus account for the different effects of Anoga-DH₃₁ and Anoga-DH₄₄ on cation transport. To test this hypothesis, we determined what effect activating the Cl^–conductance pathway using a diuretic/myotropic kinin had on the response to Anoga-DH₃₁. Since none of the An. gambiae kinins were available to us, we used the housefly (Musca domestica) kinin,Musdo-K.

Fig. 10 shows a recording of the basolateral membrane voltage from an An. gambiae principal cell challenged separately with first 100 nmol l^–1Anoga-DH₃₁, then with 100 nmol l^–1 Musdo-K and finally with a combination of the two peptides. The basolateral membrane was depolarised by Anoga-DH₃₁ and hyperpolarised by Musdo-K, which is consistent with the effect of kinin stimulation in Ae. aegyptiMalpighian tubules (Pannabecker et al.,1993). A combination of the two peptides produced a triphasic response similar to that obtained with Anoga-DH₄₄ (cf. Fig. 7A), although the initial hyperpolarisation was more pronounced.

Given that the electrophysiological signature of Anoga-DH₄₄ was similar to that obtained with a combination of Anoga-DH₃₁ and Musdo-K, we investigated how these two peptides influenced fluid secretion and ion transport. Musdo-K alone (1 μmol l^–1 final concentration) produced a 4-fold increase in fluid secretion but had no significant effect on the urine [Na⁺]:[K⁺] ratio(Fig. 11). By contrast, the[Na⁺]:[K⁺] ratio was increased >6-fold when Anoga-DH₃₁ (final concentration 1 μmol l^–1) was also added to the same group of tubules. This was accompanied by a 6-fold stimulation of fluid secretion and therefore represented an 18-fold increase in Na⁺ transport, from 38±5 pmol min^–1(N=8) to 677±93 pmol min^–1, compared with a 3-fold increase in K⁺ transport, from 128±20 pmol min^–1 to 378±58 pmol min^–1.

The mosquito natriuretic peptide identified in an Ae. aegypti head extract by Petzel et al.(1985) acts viacyclic AMP to stimulate secretion of Na⁺-rich urine by Malpighian tubules (Petzel et al., 1987). CRF-related diuretic hormones are known to use cyclic AMP as a second messenger (Coast and Kay,1994), and Beyenbach(2003) recently suggested that MNP belongs to this peptide family, based upon results from a cross-species study using Culsa-DH. In the present paper, we report the actions of Anoga-DH₄₄, the CRF-related peptide in the malaria mosquito, on Malpighian tubules from An. gambiae and Ae. aegypti. We show that it has diuretic activity but does not selectively stimulate Na⁺ transport, which is the defining feature of MNP. On the other hand, Anoga-DH₃₁, a CT-like peptide that is common to An. gambiae and Ae. aegypti, has both diuretic and natriuretic activities, stimulating fluid secretion and transepithelial Na⁺transport to the same extent as exogenous 8-bromo-cyclic AMP. We therefore propose that the previously uncharacterised MNP belongs to the CT-like family of insect diuretic hormones.

Anoga-DH₄₄ and Anoga-DH₃₁ produce characteristic changes in transepithelial voltage, which are due largely to voltage changes at the principal cell basolateral membrane. The response to the CRF-related peptide is very similar to that reported previously for MNP (fraction III in Petzel et al., 1985) and Culsa-DH (fig. 2 in Clark et al.,1998a), in that the transepithelial voltage briefly depolarises before hyperpolarising. Interestingly, this transient is not seen with either Anoga-DH₃₁ or membrane-permeant analogues of cyclic AMP (present study; Sawyer and Beyenbach,1985) and yet they, rather than Anoga-DH₄₄, mimic the natriuretic activity of MNP. Clark et al.(1998b) attribute the depolarising and hyperpolarising phases of the response to Culsa-DH as being due to the Ca²⁺-dependent opening of a paracellular Cl^– conductance and the cyclic AMP-mediated stimulation of active cation transport, respectively. Our findings are consistent with this interpretation in that a combination of Anoga-DH₃₁ and Musdo-K, a member of the kinin family of diuretic peptides that use Ca²⁺ as a second messenger in activating a Cl^– conductance pathway(O'Donnell et al., 1998; Yu and Beyenbach, 2002), mimic the electrophysiological signature of Anoga-DH₄₄. One way of accounting for the biphasic effect of MNP on transepithelial voltage is to suggest that fraction III (which was dubbed MNP by Petzel et al., 1985) contained Anoga-DH₃₁ together with a kinin. In support of this, gel filtration chromatography of fraction III indicated a molecular mass of 1862 Da (Petzel et al., 1986),which is more similar to Aedae-K-I (1593 Da) than to either CT-like or CRF-related diuretic hormones.

A paradox to arise from the present study is that, although Anoga-DH₃₁ and Anoga-DH₄₄ both stimulate cyclic AMP production by isolated Malpighian tubules, only the former has natriuretic activity. As yet, we cannot explain this, but we are able to rule out a number of possibilities. First, the peptides could be acting on functionally different populations of principal cells, only one of which responds to a rise in the level of intracellular cyclic AMP with a selective increase in transepithelial Na⁺ transport. However, such functional diversity among principal cells appears unlikely, because recordings of the basolateral membrane voltage from a single cell (Fig. 8) show that it responds in characteristic fashion to both Anoga-DH₃₁ and Anoga-DH₄₄.

Secondly, a quantitative difference in the cyclic AMP signal obtained in response to Anoga-DH₃₁ and Anoga-DH₄₄ could determine the extent to which Na⁺ transport is stimulated. In this context,it is worth noting that the CRF-related diuretic hormone of D. melanogaster (Drome-DH₄₄) stimulates cyclic AMP production by Malpighian tubule principal cells and also activates a cyclic AMP-specific phosphodiesterase (PDE), which will curtail the second messenger response(Cabrero et al., 2002). This was not investigated in the present study, which measured cyclic AMP production by Malpighian tubules after PDE activity was inhibited by IBMX. However, a quantitative difference in cyclic AMP signalling is unlikely to explain why only Anoga-DH₃₁ has natriuretic activity, because Anoga-DH₄₄ does not selectively stimulate Na⁺ transport by Malpighian tubules in which PDE activity was inhibited by IBMX.

The third possibility to be examined was whether the activation of both paracellular and transcellular pathways by Anoga-DH₄₄ (see above)militates against a natriuretic response. This also appears unlikely because Anoga-DH₃₁ selectively stimulates transepithelial Na⁺transport by Malpighian tubules in which the paracellular pathway is first activated by Musdo-K. Similarly, differences in signalling exist forβ ₂ and β₁ adrenergic receptors in vertebrate heart tissue, which are believed to be due to differences in compartmentation of cyclic AMP in the tissue. Three principles of compartmentation of cyclic AMP signalling have been postulated: not all cyclic AMP gains access to all protein kinase A; not all protein kinase A has access to all possible substrates; and not all cyclic AMP has access to all cellular phosphodiesterases (Steinberg and Brunton,2001). There is evidence that most β₂ adrenergic receptors are localized in caveolae, whereas most β₁adrenergic receptors are not; this difference in localization in the plasma membrane may explain differences in the cyclic AMP response observed on stimulation of these receptors (Steinberg and Brunton, 2001). It is likely that the receptors for the CRF-like DH, vs the CT-like DH, are localized differently in the cell membrane, giving rise to differences in their signalling. It is also possible that these receptors couple to different G proteins; Culsa-DH leads to an apparent activation of both intracellular Ca²⁺ and cyclic AMP,depending on the concentration of the ligand(Clark et al., 1998b). By contrast, our data point to Anoga-DH₃₁ leading to only elevation of cyclic AMP.

In conclusion, the diuretic and natriuretic activities of Anoga-DH₃₁, a CT-like diuretic hormone common to An. gambiae and Ae. aegypti, are consistent with it being the MNP identified by Petzel et al.(1985). However, we have yet to show that Anoga-DH₃₁ is released into the circulation when the insect takes a blood meal.

We are grateful to Shahida Begum (London School of Hygiene and Tropical Medicine) for the supply of mosquitoes and to Alan Tyler for technical assistance. This work was partially supported by a NIH grant (GM 48172) to D.A.S.