Intracellular ion activities in Malpighian tubule cells of Rhodnius prolixus: evaluation of Na⁺-K⁺-2Cl^-cotransport across the basolateral membrane

The fluid secreted by the upper segment of Malpighian tubules of Rhodnius prolixus during diuresis consists of approximately 100mmoll^-1 NaCl and 80mmoll^-1 KCl(Maddrell and Phillips, 1975). Secretion of ions and osmotically obliged water by tubules of this and other species is driven primarily by an apical vacuolar-type H⁺-ATPase(Wieczorek et al., 1989; Maddrell and O'Donnell, 1992). It is suggested that electrogenic transport of H⁺ from the cell to the lumen energizes amiloride-sensitive exchange of cytoplasmic K⁺and/or Na⁺ for luminal H⁺.

Entry of Na⁺, K⁺ and Cl^- through a basolateral Na⁺-K⁺-2Cl^- cotransporter has been proposed for tubules of Rhodnius prolixus on the basis of the effects of bumetanide, Na⁺-free saline and Cl^--free saline on fluid secretion and transepithelial potential(O'Donnell and Maddrell, 1984; Ianowski and O'Donnell, 2001). Na⁺-K⁺-2Cl^- cotransport has also been implicated in basolateral entry of ions into Malpighian tubules of other species, including Aedes aegypti(Hegarty et al., 1991), Formica polyctena (Leyssens et al., 1994), Manduca sexta(Audsley et al., 1993; Reagan, 1995), Teleogryllus oceanicus (Xu and Marshall, 1999) and Locusta migratoria(Al-Fifi et al., 1998).

In vertebrates, the cation-Cl^- cotransport superfamily includes two Na⁺-K⁺-Cl^- (NKCC) isoforms, one Na⁺-Cl^- (NCC) isoform (bumetanide-insensitive,K⁺-independent) and four K⁺-Cl^- (KCC)isoforms. The Na⁺-K⁺-Cl^- cotransporter is an electroneutral transporter with a stoichiometry of 1Na⁺:1K⁺:2Cl^- in the overwhelming majority of cases (Haas and Forbush,2000). Nevertheless, a cotransporter with a stoichiometry of 2Na⁺:1K⁺:3Cl^- has been described(Russell, 1983).

The direction of net ion transport by the Na⁺-K⁺-2Cl^- cotransporter may be into or out of the cell depending on the sum of the chemical potential gradients of the transported ions, and the transporter can be inhibited by loop diuretics such as furosemide or bumetanide (Haas and Forbush, 2000). Several alternative routes for Cl^-and/or K⁺ entry have been proposed in tubules of other insects,including a K⁺-Cl^- cotransporter in Drosophila melanogaster (Linton and O'Donnell,1999), Ba²⁺-sensitive K⁺ channels in Formicapolyctena(Leyssens et al., 1994) and the Na⁺/K⁺-ATPase in Locusta migratoria(Anstee and Bowler, 1979; Anstee et al., 1986) and Drosophila melanogaster (Linton and O'Donnell, 1999). It is important to point out that our earlier studies (O'Donnell and Maddrell,1984; Ianowski and O'Donnell,2001) did not preclude the possible involvement of other transporters such as K⁺-Cl^- or Na⁺-Cl^- cotransporters, K⁺ channels or the Na⁺/K⁺-ATPase.

Critical evaluation of the possible roles of cotransporters, exchangers and ion channels requires measurement of membrane potential and the intracellular activities of the ions involved so that electrochemical potentials can be calculated for each ionic species. The directions of net ion movements for a particular transporter can then be predicted by summing the electrochemical potentials of all the participating ions to calculate the netelectrochemical potential. For example, proposals of Na⁺-K⁺-2Cl^- cotransport during Malpighian tubule fluid secretion assume that a favourable electrochemical potential for Na⁺ influx will drive the coupled influx of Cl^- and K⁺ against their electrochemical potentials(Xu and Marshall, 1999; Ianowski and O'Donnell, 2001). However, in spite of the central role of Na⁺ in fluid secretion by Malpighian tubules of blood feeders as well as many other insects, there is to date a single report of intracellular Na⁺ activity in tubules of the weta Hemideina maori (Neufeld and Leader, 1998), whereas activities of K⁺ and Cl^- have been measured in Malpighian tubules of Locusta migratoria (Morgan and Mordue,1983), Formica polyctena(Leyssens et al., 1993; Dijkstra et al., 1995) and Hemideina maori (Neufeld and Leader, 1998). Measurements of intracellular Na⁺activity are also critical to evaluation of the links between the transport of Na⁺ and other inorganic ions (H⁺, Ca²⁺) or organic solutes (e.g. sugars, amino acids and organic acids).

This paper describes experiments in which basolateral membrane potential and intracellular activities of Na⁺, K⁺ and Cl^- were measured simultaneously in Malpighian tubules of Rhodnius prolixus. The results support a role for Na⁺-K⁺-2Cl^- cotransport and rule out significant contributions of K⁺-Cl^- cotransport,Na⁺/K⁺-ATPase or K⁺ channels to net ion entry during serotonin-stimulated fluid secretion.

Fifth-instar Rhodnius prolixus Stål were used 1-4 weeks after moulting in all experiments. Animals were obtained from a laboratory colony maintained at 25-28°C and 60% relative humidity in the Department of Biology, McMaster University. Experiments were carried out at room temperature (20-25°C).

Animals were dissected under the appropriate saline(Table 1) with aid of a dissecting microscope. Only the fluid-secreting upper Malpighian tubule, which comprises the upper two-thirds (approximately 25 mm) of the tubule's length,was used. In contrast to tubules of dipterans, the upper tubule of Rhodnius prolixus consists of a single cell type whose secretory properties are uniform along its length(Collier and O'Donnell,1997).

Secretion assays were performed as described previously(Ianowski and O'Donnell,2001). Briefly, upper segments of Malpighian tubules were isolated in 100 μl droplets of bathing saline under paraffin oil. The cut end of the tubule was pulled out of the saline and wrapped around a fine steel pin pushed into the Sylgard base of a Petri dish. After stimulation with serotonin(10^-6 moll^-1), secreted fluid droplets formed at the cut end of the tubule and were pulled away from the pin every 5 min for 60-90 min using a fine glass probe. Secreted droplet volume was calculated from droplet diameter measured using an ocular micrometer. Secretion rate was calculated by dividing the volume of the secreted droplet by the time over which it formed.

An isolated upper Malpighian tubule was attached to the bottom of a custom-built superfusion chamber pre-coated with poly-L-lysine to facilitate adherence of the tubules under saline(Ianowski and O'Donnell,2001). The fluid in the chamber was exchanged at 6 ml min^-1, sufficient to exchange the chamber's volume every 3 s.

Intracellular ion activity and basolateral membrane potential were measured simultaneously in single cells using ion-selective double-barrelled microelectrodes (ISMEs). The ISMEs were fabricated from borosilicate double-barrelled `Piggy-back' capillary glass (WPI, Sarasota, USA). The capillary glass was washed for 30 min in nitric acid, then rinsed with deionized water and baked on a hot plate at 200°C for 30 min. The capillaries were then removed from the hot plate, and the smaller barrel filamented was filled with a 2-3 cm column of deionized water before pulling on a vertical micropipette puller (PE-2, Narishige, Japan). Retention of the hydrophobic ionophore cocktails requires silanization of the interior of the ion-selective barrel. For this purpose, approximately 300 μl of dimethyldichlorosilane (Fluka) was placed in a glass vial. A 23 gauge syringe needle was passed from inside to outside through the plastic cap of the vial. The syringe needle was placed in the lumen of the larger unfilamented barrel. The glass vial was then placed on a hot plate at 200°C for approximately 8 s to produce a stream of dimethyldichlorosilane vapour through the end of the syringe needle so as to silanize the interior of the larger capillary barrel. The water in the smaller barrel prevented silanization of its interior. The double-barrelled capillary was then removed from the syringe needle and baked for 45 min at 200°C. Finally, a short column of liquid ion exchanger was introduced into the larger barrel and it was backfilled with the appropriate solution. The smaller barrel remained hydrophilic and was filled with the appropriate reference electrode solution (see below).

In some cases, the resistance of the ion-selective electrode was above 10¹¹ Ω, resulting in very slow response times and unstable voltages. Electrode resistance was therefore reduced by controlled submicrometre tip breakage. The tip of the electrode was touched to the tubule surface or to the surface of a piece of tissue paper under saline, as described by O'Donnell and Machin(1991). This process of controlled tip breakage permitted a two- to fourfold reduction in tip resistance and consequent improvement in response time without compromising the quality of subsequent impalements. Electrodes were used for experiments only when the 90 % response time of the ion-selective barrel to a solution change was less than 30 s and when the response of the ion-selective barrel to a 10-fold change in ion activity was more than 49 mV. Approximately 40 % of K⁺ electrodes, 30 % of Na⁺ electrodes and 30 % of Cl^- electrodes met these criteria.

K⁺-selective microelectrodes were based on potassium ionophore I, cocktail B (Fluka). The K⁺-selective barrel was backfilled with 500 mmol l^-1 KCl. The reference barrel was filled with 1 mol l^-1 sodium acetate near the tip and shank and 1 mol l^-1KCl in the rest of the electrode. The K⁺-selective electrode was calibrated in solutions of (in mmol l^-1) 15 KCl:135 NaCl and 150 KCl. The mean slope of the K⁺ electrodes used in this study was 52±1 mV per decade change in K⁺ activity (mean ±S.E.M., N=22).

Na⁺-selective microelectrodes were based on the neutral carrier ETH227 (sodium ionophore I, cocktail A, Fluka). The Na⁺-selective barrel was backfilled with 500 mmol l^-1 NaCl and the reference barrel was filled with 1 mol l^-1 KCl. Na⁺-selective electrodes were calibrated in solutions of (in mmol l^-1) 15 NaCl:135 KCl and 150 NaCl. The mean slope of the Na⁺ electrodes used was 57±1.5 mV per decade change in Na⁺ activity (mean± S.E.M., N=21). Since Ca²⁺ is known to interfere with the Na⁺ neutral carrier ETH227, the bathing saline in these experiments was initially Ca²⁺-free saline(Table 1). When the microelectrode electrode had impaled the cell, the bathing saline was replaced with control saline. Preliminary experiments showed that exposure of the tubule to Ca²⁺-free saline did not affect transepithelial ion transport; secretion rates of serotonin-stimulated upper Malpighian tubules were identical in control saline and in Ca²⁺-free saline.

Cl^--selective microelectrodes were based in ionophore I,cocktail A (Fluka). Both Cl^--selective and reference barrels were backfilled with 1 mol l^-1 KCl. The electrode was calibrated in 100 mmol l^-1 KCl and 10 mmol l^-1 KCl. The mean slope of the Cl^- electrodes used was 53±1.5 mV per decade change in Cl^- activity (mean ± S.E.M., N=22).

Potential differences from the reference (V_ref) and ion-selective (V_i) barrels were measured by a high-input-impedance differential electrometer (FD 223, WPI). V_ref was measured with respect to a Ag/AgCl electrode connected to the bath through a 0.5 mol l^-1 KCl agar bridge. V_i was filtered through a low-pass RC filter with a time constant of 1 s to eliminate noise resulting from the high input impedance(approximately 10¹⁰Ω) of the ion-selective barrel. V_ref and the difference V_i-V_ref were recorded using an A/D converter and data-acquisition system (Axotape, Axon Instruments, Burlingame,CA, USA).

Intracellular recordings were acceptable if the potential was stable to within 1 mV for 30 s or longer. In addition, recordings were acceptable only if the potential of each electrode in the bathing saline after withdrawal differed from the potential before impalement by less than 3 mV. In preliminary experiments using fine-tipped single-barrelled microelectrodes, we established that mean values of basolateral membrane potential before and after serotonin stimulation were -58 mV (95 % confidence interval -61 to -55 mV) and -63 mV (95 % confidence interval -65 to -60 mV) respectively. In experiments using double-barrelled ion-selective electrodes, values of basolateral membrane potential (V_bl) less negative than-55 mV in unstimulated tubules and -60 mV in stimulated tubules were considered indicative of poor-quality impalements, and the data were therefore discarded.

The ion activity in the calibration solution was calculated as the product of ion concentration and the ion activity coefficient. The activity coefficients for the single electrolyte calibration solutions are 0.77 and 0.901 for 100 mmol l^-1 KCl and 10 mmol l^-1 KCl,respectively (Hamer and Wu,1972). For the solutions containing 0.15 mol l^-1 KCl or NaCl and mixed solutions of KCl and NaCl with constant ionic strength (0.15 mol l^-1), the activity coefficient is 0.75, calculated using the Debye-Huckel extended formula and Harned's rule(Lee, 1981).

K⁺ and Na⁺ activities of secreted droplets were measured using single-barrelled ion-selective microelectrodes as described previously (Maddrell and O'Donnell,1992; Maddrell et al.,1993; O'Donnell and Maddrell,1995). The K⁺- selective and Na⁺-selective microelectrodes were silanized using the procedures of Maddrell et al.(1993). Filling and calibration solutions of single-barrelled ion-selective and reference electrodes were the same as those described above for double-barrelled microelectrodes.

Ion flux (nmol min^-1) was calculated as the product of secretion rate (nl min^-1) and ion activity (mmol l^-1) in the secreted droplets.

Thermodynamic evaluation of a particular ion transporter involves calculation of the net electrochemical potential(Δμ_net/F)(Schmidt and McManus, 1977; Haas et al., 1982; Loretz, 1995).

A positive value of Δμ_net/F favours net movement of ions from cell to bath, whereas a negative value would tend to promote a net movement from bath to cell. WhenΔμ _net/F=0 mV, there is no net force operating on the cotransporter system (Schmidt and McManus, 1977; Haas et al.,1982; Loretz,1995).

To study the role of K⁺ channels on K⁺ transport, the effect of Ba²⁺ on V_bl was studied. Electrodes were pulled from filamented single-barrelled capillary pipettes (WPI,Sarasota, FL, USA) filled with 3 mol l^-1 KCl and connected to an electrometer (Microprobe system M-707A, WPI, Sarasota, FL, USA). Microelectrode resistance was typically 20-40 MΩ.

Stock solutions of bumetanide (Sigma) were prepared in ethanol so that the maximum final concentration of ethanol was ≤0.1 % (v/v). Previous studies have shown that Malpighian tubule secretion rate is unaffected by ethanol at concentrations ≤1 % (v/v) (Ianowski and O'Donnell, 2001). Serotonin and ouabain (Sigma) were dissolved in the appropriate saline solution (Table 1).

Results are expressed as means ± S.E.M. Significant differences were evaluated using unpaired Student's t-tests (P<0.05).

Stimulation of Malpighian tubule cells for 30 min with 10^-6 mol l^-1 serotonin produced significant (P<0.05) changes in V_bl and intracellular ion activities. Basolateral membrane potential hyperpolarized from -59±0.1 mV (N=34) to-63±0.5 mV (N=22). Intracellular Na⁺ activity increased by 12 mmol l^-1 after serotonin stimulation(Table 2). Interference by K⁺ on Na⁺ microelectrodes was estimated using the Nicolsky-Eisenman equation (Ammann,1986). These calculations showed that interference due to intracellular K⁺ would cause a negligible overestimation of a_Naⁱ of less than 0.4 mmol l^-1.

The negative value of Δμ_Na/F indicates that the electrochemical potential favoured movement of Na⁺ from the bathing saline into the cell in both unstimulated and stimulated tubules(Table 2; Fig. 1A).

In unstimulated tubules, the K⁺ electrochemical potential favoured K⁺ movement from the cell to the bath(Table 2). Stimulation with serotonin reduced both intracellular K⁺ activity andΔμ _K/F (Table 2; Fig. 1B). For stimulated tubules, Δμ_K/F was not significantly different from 0 mV.

In contrast, serotonin stimulation did not affect intracellular Cl^- activity or Δμ_Cl/F. The electrochemical potential favoured Cl^- movement from the cell to the haemolymph in both stimulated and unstimulated tubules(Table 2; Fig. 1C).

To determine whether other anions in the cell interfered with the Cl^- electrode, the effect of replacing Cl^- in the bath with SO₄^2- on intracellular Cl^- activity was measured. After 10 min in Cl^--free saline, intracellular Cl^- activity was reduced to 5±0.8 mmol l^-1(N=3) in stimulated tubules. Thus, the interference of other anions on intracellular Cl^- measurements was small.

Net electrochemical potentials were calculated for the three cation-Cl^- cotransporters of interest. The results indicated that in both unstimulated and serotonin-stimulated tubules the electrochemical potentials favoured movement of Na⁺, K⁺ and Cl^- from the bath into the cell through a cotransporter with a stoichiometry of Na⁺-K⁺-2Cl^-(Fig. 2A). The data were also consistent with net movement of ions from the bath into the cell through a Na⁺-Cl^- cotransporter(Fig. 2B), but not through a K⁺-Cl^- cotransporter. K⁺-Cl^-cotransport would have produced net movement of ions in the opposite direction, from the cell to the bath (Fig. 2C).

Intracellular Cl^- activity in serotonin-stimulated tubules (30 min in 10^-6 mol l^-1 serotonin) showed a significant decrease from 30±4 mmol l^-1 to a minimum of 8±3 mmol l^-1 (N=4) when the bathing saline was changed from 14.5K saline (containing 137.1 mmol l^-1 Na⁺) to Na⁺-free saline (Table 1) for 1-2 min (Fig. 3A). After the initial rapid decline, a_Clⁱ subsequently increased on average by 6±2 mmol l^-1 (N=4) during sustained exposure to Na⁺-free saline (Fig. 3A). Possible explanations for the decline in a_Clⁱ and this subsequent small increase are considered in the Discussion. Intracellular Cl^- activity also showed a significant decrease from 32±3 mmol l^-1 in 14.5K saline to 10±3 mmol l^-1 (N=5) in K⁺-free saline (Table 1) after 1-2 min(Fig. 3B). Intracellular Cl^- activity recovered to 43±2 mmol l^-1(N=4) and 37±6 mmol l^-1 (N=5) when Na⁺ or K⁺ concentration, respectively, was restored in the bathing fluid (Fig. 3A,B). The salines used in these experiments(Table 1) were slightly different from those used in previous experiments to permit depletion of a single cation (i.e. Na⁺ or K⁺) without altering the concentration of the remaining ions in solution.

Intracellular Cl^- activity declined from 33±3 mmol l^-1 (N=5) in control saline to a minimum of 8±0.6 mmol l^-1 (N=5) in saline containing 10^-5 mol l^-1 bumetanide. The effect of Na⁺-free or K⁺-free saline on intracellular Cl^- activity was greatly reduced in the presence of 10^-5 mol l^-1 bumetanide. After addition of bumetanide, Cl^- activity was further reduced by only approximately 2 mmol l^-1 from approx. 8 to approx. 6 mmol l^-1 (N=2) when the tubules were exposed to Na⁺-free saline (Fig. 4A). Similarly, K⁺-free saline produced a further decrement in intracellular Cl^- activity of only 3 mmol l^-1 (from 9±0.8 to 6±0.9 mmol l^-1, N=3) in tubules treated with 10^-5 mol l^-1bumetanide (Fig. 4B).

Fluid secretion rates, K⁺ flux and Na⁺ flux were all reduced by the addition of 10^-5 mol l^-1 bumetanide to serotonin-stimulated Malpighian tubules. Fluid secretion was reduced by 72%within 30 min of addition of bumetanide(Fig. 5A). Over the same period, K⁺ flux and Na⁺ flux were reduced by 69% and 87%, respectively (Fig. 5B,C).

In contrast, the addition of ouabain 10^-4 moll^-1 did not affect either fluid secretion rate(Fig. 6A) or K⁺ flux(Fig. 6B) in serotonin-stimulated tubules.

To evaluate the possible role of K⁺ channels in vectorial movement of K⁺ across the basolateral membrane and into the cell,the effects of the K⁺ channel blocker Ba²⁺ on fluid secretion and basolateral membrane potential (V_bl) were studied. During this experiment, NaH₂PO₄ was omitted from the control saline to prevent the precipitation of barium phosphate.

The addition of 6 mmoll^-1 Ba²⁺ had no effect on the fluid secretion rate of Rhodnius prolixus Malpighian tubules stimulated with 10^-6 moll^-1 serotonin(Fig. 7). Addition of 6 mmoll^-1 Ba²⁺ caused V_bl to depolarize slightly but significantly by 7±1 mV (N=5)(Fig. 8), which is consistent with the presence of basolateral K⁺ channels.

This paper reports the first simultaneous measurements of intracellular Na⁺ activity and V_bl in the Malpighian tubules cells of an insect. In conjunction with measurements of intracellular K⁺ and Cl^- activities and membrane potential, a critical thermodynamic evaluation of basolateral cation-Cl^- cotransport in Malpighian tubules is now possible for the first time.

Intracellular activities of Na⁺, K⁺ and Cl^- in the Malpighian tubule cells of Rhodnius prolixusfall within the range of activities for these ions seen in other insect epithelia studied using ISMEs. The intracellular Na⁺ activity measured in Rhodnius prolixus tubule cells is very similar to that of 32 mmoll^-1 for Hemideina maori, measured in saline of similar osmolality (Neufeld and Leader,1998). Values of a_Naⁱ range from 8 mmoll^-1 in rectal cells of Schistocerca gregaria(Hanrahan and Phillips, 1984)to 17 mmoll^-1 in unstimulated salivary duct cells of Periplaneta americana (Lang and Walz, 2001).

Intracellular Cl^- activities of 32 mmoll^-1 measured in Rhodnius prolixus tubules are very similar to values of 38 mmoll^-1 in Locusta migratoria tubules(Morgan and Mordue, 1983) and 35 mmoll^-1 in tubules of Formica polyctena(Dijkstra et al., 1995).

Similarly, the intracellular K⁺ activity in unstimulated Rhodnius prolixus Malpighian tubules (86 mmoll^-1) is comparable with the values of 71 mmoll^-1 measured in Locusta migratoria tubules (Morgan and Mordue, 1983) and 61 mmoll^-1 in Formica polyctena tubules (Leyssens et al.,1993). For comparisons with studies that reported ion concentrations, a_Naⁱ, a_Kⁱ and a_Clⁱ in other tissues have been calculated assuming an activity coefficient of 0.75.

The values for a_ionⁱ measured directly with ISMEs in the present study are also similar to those estimated from total concentration measurements obtained by X-ray microanalysis(Gupta et al., 1976). Estimated activities for Na⁺, Cl^- and K⁺ in the main cytoplasm of unstimulated Rhodnius prolixus tubules were 10 mmoll^-1, 23 mmoll^-1 and 77 mmoll^-1,respectively. After serotonin stimulation, estimated activities for Na⁺, Cl^- and K⁺ were 32 mmoll^-1,45 mmoll^-1 and 76 mmoll^-1, respectively(Gupta et al., 1976).

Intracellular levels of Na⁺ and K⁺ but not Cl^- are altered by stimulation with serotonin. A dramatic rise in a_Naⁱ from 17 to 69 mmoll^-1 is also seen when Periplaneta americana salivary ducts are stimulated with dopamine (Lang and Walz,2001). The increase in a_Naⁱ in the present study is of similar magnitude to the corresponding decrease in a_Kⁱ in response to serotonin stimulation. It is worth noting that a_Kⁱ declines from 84 to 30 mmoll^-1 when Periplaneta americana salivary ducts are stimulated with dopamine (Lang and Walz,2001). In both Rhodnius prolixus tubules and Periplaneta americana salivary ducts, therefore, the sum of a_Naⁱ and a_Kⁱremains constant when the cells are stimulated, but the Na⁺:K⁺ activity ratio increases.

The simultaneous measurement of a_Naⁱ and V_bl permits accurate calculation of the electrochemical potential (Δμ_Na/F) for Na⁺ across the basolateral membrane. Importantly, these calculations show that passive Na⁺ entry into the cell is highly favoured in both unstimulated and stimulated cells.

The electrochemical potential for Cl^-(Δμ_Cl/F), in contrast, is outwardly directed. This indicates that Cl^- activity in the cell is higher than expected on the basis of passive distribution of Cl^- across the basolateral membrane in both unstimulated and serotonin-stimulated cells. Cl^- must therefore be actively transported across the basolateral membrane into the cell. Outwardly directed electrochemical potentials for Cl^- have also been reported in Locusta migratoriaMalpighian tubules (Morgan and Mordue,1983). In contrast, Δμ_Cl/F in Formica polyctena Malpighian tubule cells favours Cl^-movement into the cell across the basolateral membrane(Dijkstra et al., 1995).

The electrochemical potential for K⁺(Δμ_K⁺/F) across the basolateral membrane is outwardly directed in unstimulated tubules, whereas values ofΔμ _K/F near 0 mV are found after serotonin stimulation. Similarly, values of Δμ_K/F across the basolateral membrane not different from zero have been measured for Malpighian tubule cells of Locusta migratoria(Morgan and Mordue, 1983) and Formica polyctena (Leyssens et al., 1993).

The feasibility of K⁺-Cl^-,Na⁺-K⁺-2Cl^- and Na⁺-Cl^-cotransporters in ion movement across the basolateral membrane was evaluated by calculating the net electrochemical potential using the electrochemical potentials for each participating ion(Schmidt and McManus, 1977; Haas et al., 1982; Loretz, 1995). The results show that a K⁺-Cl^- cotransporter would drive net movement of K⁺ and Cl^- from cell to bath, i.e. in the opposite direction to that during fluid secretion. It is important to point out that our data do not preclude the presence of a K⁺-Cl^- cotransporter and its involvement in other functions such as cell volume regulation in hypo-osmotic media(Lauf et al., 1992). However,our findings do show that a K⁺-Cl^- cotransporter could not be involved in vectorial ion transport during fluid secretion by the Rhodnius prolixus Malpighian tubule.

Both Na⁺-K⁺-2Cl^- and Na⁺-Cl^- cotransporters favour coupled movement of Na⁺, K⁺ and Cl^- or of Na⁺ and Cl^-, respectively, from bath to cell in both unstimulated and serotonin-stimulated tubules. Na⁺-Cl^- cotransport,however, is inconsistent with the finding that bumetanide decreases transepithelial fluxes of both Na⁺ and K⁺, rather than just that of Na⁺.

Our results also show that maintenance of a_Clⁱ is dependent upon the presence of both Na⁺ and K⁺ and is bumetanide-sensitive. Intracellular Cl^- activity declines in response to bumetanide, K⁺-free saline or Na⁺-free saline. Moreover, our results suggest that intracellular Cl^- activity is near equlibrium across the apical membrane. For example, in serotonin-stimulated cells with an a_Clⁱ of 32 mmol l^-1, the measured apical membrane potential is -31 mV, cell-negative(Ianowski and O'Donnell,2001). The latter value is very close to the Nernst equilibrium potential for Cl^- (E_Cl) across the apical membrane (-34mV), assuming a luminal Cl^- activity of approximately 124 mmol l^-1. Moreover, when the apical membrane potential increases (i.e. as the transepithelial potential becomes more lumen-positive)in Na⁺-free saline or K⁺-free saline or in the presence of bumetanide (Ianowski and O'Donnell,2001), a_Clⁱ declines, as observed in the present study (Fig. 3). We suggest that the subsequent small increase in a_Clⁱ after the initial rapid decline in Na⁺-free saline (Fig. 3A) reflects changes in apical membrane potential. Earlier studies(O'Donnell and Maddrell, 1984)have shown that transepithelial potential increases to a lumen-positive value in Na⁺-free saline, then gradually declines. Taken together, our results suggest that basolateral Na⁺-K⁺-2Cl^-cotransport and apical Cl^- channels together play a primary role in setting the level of intracellular Cl^- activity. Future studies will examine electrochemical gradients across the apical membrane in detail and will also address possible contributions from other transporters (e.g. basolateral Cl^-/HCO₃^- exchangers).

There is also molecular biological evidence for cation-Cl^-cotransporters in Malpighian tubules of an insect. A putative Na⁺-K⁺-2Cl^- cotransporter cloned from Manduca sexta tubules (MasBSC)(Reagan, 1995) shares 40-43%sequence identity with the shark, rat and mouse bumetanide-sensitive Na⁺-K⁺-2Cl^- cotransporter, 40% with human and mouse thiazide-sensitive Na⁺-Cl^- cotransporters and 25-26% sequence identity with mouse K⁺-Cl^-cotransporters. MasBSC appears to be one of the oldest members of the family of Na⁺-(K⁺)-Cl^- transporters reported to date(Reagan, 1995; Mount et al., 1998). MasBSC also shares 52% amino acid sequence identity with the CG2509 gene product of Drosophila melanogaster, suggesting that this putative Na⁺-K⁺-2Cl^- cotransporter may occur in other insects as well.

The involvement of the Na⁺/K⁺-ATPase in serotonin-stimulated fluid secretion can be rejected on the grounds that ouabain does not reduce fluid secretion rate or K⁺ flux. We can also rule the contribution of K⁺ channels to transport of K⁺ from bath to lumen during fluid secretion by Rhodnius prolixus tubules on the basis of two sets of evidence. First, treatment with Ba²⁺ did not affect fluid secretion, suggesting that K⁺ channels are not a significant component of transepithelial K⁺ transport at physiological concentrations of extracellular K⁺. Second, basolateral membrane potential depolarised after addition of Ba²⁺. This effect is consistent with blockage of K⁺ leakage into the bath, since the basolateral membrane potential appears to be determined primarily by the K⁺ conductance (i.e. V_m is very similar to E_K). The depolarisation of the basolateral membrane potential after addition of Ba²⁺ is consistent with a small but positive value forΔμ _K/F that favours movement of K⁺ from cell to bath through ion channels. Although the magnitude ofΔμ _K/F indicated in Table 2 (0±2mV) is indistinguishable from zero, this may reflect the limitations inherent in the use of double-barrelled ion-selective microelectrodes for measurement of very small (<1mV) changes in potential. Importantly, since net K⁺transport during fluid secretion is in the opposite direction, from bath to cell, passive K⁺ movement through channels is not involved in vectorial ion transport during serotonin-stimulated fluid secretion. It is also worth noting that, in insect epithelia where basolateral K⁺channels play a role in transepithelial K⁺ secretion driven by apical H⁺-ATPases and K⁺/H⁺ exchangers,Ba²⁺ results not in a depolarisation of V_bl but in a substantial hyperpolarisation, as discussed in detail by Weltens et al.(1992) for Formica polyctena tubules by Moffett and Koch(1992) for the lepidopteran midgut.

Taken together, our results are consistent with a cardinal role for a bumetanide-sensitive Na⁺-K⁺-2Cl^-cotransporter during serotonin-stimulated fluid secretion by Malpighian tubules of Rhodnius prolixus. We also found a favourable net electrochemical potential for this transporter in unstimulated tubules. The operation of the Na⁺-K⁺-2Cl^- cotransporter at a low rate in unstimulated tubules has been suggested previously(Maddrell and Overton, 1988). In unstimulated tubules, K⁺ may enter cells from the bath both through a ouabain-sensitive Na⁺/K⁺-ATPase and a low level of activity of a Na⁺-K⁺-2Cl^-cotransporter (Maddrell and Overton,1988). Addition of ouabain reduces one path for K⁺entry and blocks the transport of Na⁺ from cell to bath, with a resulting increase in Na⁺ transport from cell to lumen(Maddrell and Overton,1988).

Stimulation with serotonin results in a nearly 1000-fold increase in the rate of transepithelial ion transport through stimulation of apical ion transporters as well as the basolateral Na⁺-K⁺-2Cl^- cotransporter(Maddrell and Overton, 1988; Maddrell, 1991). Because ion flux through the basolateral cotransporter is so much greater than that through the Na⁺/K⁺-ATPase, fluid secretion in stimulated tubules is insensitive to ouabain. Moreover, if the rates of ion transport through K⁺ channels and the Na⁺/K⁺-ATPase are negligible relative to the rate of ion influx through the Na⁺-K⁺-2Cl^- cotransporter, then levels of Na⁺ and K⁺ in the cell would tend to become equal. However, it is well known that the apical transporters (i.e. the combined effects of the H⁺-ATPase and the alkali cation/H⁺exchangers) have a preference for Na⁺ over K⁺, resulting in selective transfer of Na⁺ into the lumen(Maddrell, 1978; Maddrell and O'Donnell, 1993). Under these conditions then, one might expect intracellular Na⁺activity to increase, but not to the level of a_Kⁱ. This was the pattern of changes in a_Naⁱ and a_Kⁱobserved in the present study.

The finding of a large inwardly directed electrochemical potential for Na⁺ across the basolateral membrane of Rhodnius prolixustubules suggests that the Na⁺ gradient may be utilised for other Na⁺-coupled transporter systems in addition to the Na⁺-K⁺-2Cl^- cotransporter. K⁺-coupled transporters for uptake of organic molecules such as amino acids have been well described in epithelia from species such as Manduca sexta with low levels of Na⁺ in the haemolymph(Castagna et al., 1997; Liu and Harvey, 1996; Bader et al., 1995). Future studies of Rhodnius prolixus Malpighian tubules will address the role of the basolateral Na⁺ gradient in processes such as solute uptake(amino acids, sugars, organic acids) or pH regulation through Na⁺/H⁺ exchange (see Petzel, 2000).

The authors are grateful to the Natural Sciences and Engineering Research Council (Canada) for financial support and to Dr Emmi Van Kerkhove for helpful discussions of the manuscript.