K⁺ transport in Malpighian tubules of Tenebrio molitor L.: a study of electrochemical gradients and basal K⁺uptake mechanisms

The mechanisms underlying ion transport in insect Malpighian tubules have been studied extensively in a number of species (for reviews, see Maddrell and O'Donnell, 1992; Nicolson, 1993; Pannabecker, 1995; Beyenbach, 1995). Consensus has now been reached on a model for ion and fluid transport, which is driven primarily by an apical vacuolar-type H⁺-ATPase. In Malpighian tubule cells, the electrogenic transport of H⁺ into the tubule lumen establishes a proton gradient and an electrical potential difference across the apical membrane, which drives movement of alkali cations from the cell to the lumen through apical Na⁺/nH⁺ and/or K⁺/nH⁺ antiporters(Zhang et al., 1994). In addition, the proton pump generates a favourable electrical gradient for anion movement (Cl^-), which may be either transcellular or through the paracellular shunt (Dÿkstra et al., 1994; Beyenbach, 1995). With the exception of some insects that make use of sodium (e.g. Glossina: Gee 1976; Libellula: Nicholls, 1985), potassium acts as the primary driving force of urine formation in Malpighian tubules. Blood-sucking species can switch from a mixed Na⁺/K⁺ in basal conditions to a preferentially Na⁺-driven fluid secretion after a blood meal (e.g. Rhodnius: Maddrell, 1980).

Basolateral K⁺ transport systems must permit adequate K⁺ uptake to maintain fluid secretion. The existence of Ba²⁺-sensitive K⁺ channels has been confirmed in tubules of Onymacris (Nicolson and Isaacson, 1987), Formica(Weltens et al., 1992), Aedes (Masia et al.,2000) and Locusta(Hyde et al., 2001). Uptake mechanisms for K⁺ and/or Na⁺ at the haemolymph side may differ according to the species. In tubules of the forest ant Formica polyctena, alternative routes for basal K⁺ entry appear to be implicated over different ranges of bathing saline [K⁺]. In the presence of high [K⁺], entry occurs via high-conductance,Ba²⁺-sensitive channels. At lower [K⁺],K⁺/Cl^- and/or Na⁺/K⁺/2Cl^- cotransporters become functional(Leyssens et al., 1994; Van Kerkhove, 1994). Although there has been some controversy about the presence of a basolateral Na⁺/K⁺-ATPase in insect epithelia, this ouabain-sensitive electrogenic pump has been implicated in facilitating K⁺ transport in Locusta(Anstee et al., 1986) and Na⁺ transport in Rhodnius(Maddrell and Overton, 1988)tubule cells. However, fluid secretion rates of Drosophila melanogaster (Dow et al.,1994) and Formica(Leyssens et al., 1994)tubules remain unaffected by ouabain and thus, if present, the Na⁺/K⁺-ATPase does not appear to play a role in fluid secretion.

The present study focuses on possible K⁺ uptake pathways across the basolateral membrane of Tenebrio tubule cells and demonstrates the impact of the apical membrane on passive K⁺ uptake viathe Ba²⁺-sensitive K⁺ channels. Fluid secretion rates and potentials across basolateral and apical membranes have been measured at different bath [K⁺] and in the presence and absence of various blockers. The results obtained indicate the direction of the electrochemical gradient for K⁺, the influence of the apical membrane potential on this gradient and the relative importance of the alternative K⁺uptake routes. This study provides a basis for future studies in which (1) the nature of the K⁺ channels involved in basolateral K⁺uptake is further investigated and (2) the various K⁺ uptake mechanisms are implicated as sites for regulation by endogenous factors.

Tenebrio molitor L. larvae used in this study were maintained in dry bran cultures at room temperature (20-23°C). The diet was supplemented weekly with apple as a source of moisture. Care was taken in selecting mealworms of similar size for all experiments.

The composition of the bathing solutions is summarised in Table 1. Solutions were freshly prepared each week, filtered through 0.22 μm Millipore filters and kept at 2°C until use. The pH was measured daily before use. In experiments containing Ba²⁺, NaH₂PO₄ was omitted from all salines to maintain constant osmolality and prevent precipitation. Control experiments in which NaH₂PO₄ was omitted showed no change in secretion rate or electrical profile. In low Na⁺experiments, salines contained a maximum of 6 mmol l^-1Na⁺.

The following pharmacological substances were tested on Malpighian tubule preparations: barium chloride (Sigma, Bornem, Belgium), ouabain (Fluka Buchs,Switzerland) and bumetanide (Sigma).

Malpighian tubules from larval Tenebrio molitor were isolated as described by Wiehart et al.(2002). These were the free segments of the tubules, severed near the midgut and the rectal complex. Droplets of physiological saline (50 μl) were placed in a Petri dish coated with Sylgard (10:1 base to curing agent) and covered with water-saturated liquid paraffin. Two tubules were placed into each saline drop. The two ends of each tubule were pulled out of the bathing fluid and wrapped around Minuten pins, where they continued to secrete, and the urine was collected as discrete droplets in the liquid paraffin. Secreted drops were removed with a fine glass pipette and their diameters measured with a calibrated eyepiece graticule. The volume, and therefore the rate of secretion, was determined assuming that the droplets were spherical. The tubules were allowed to equilibrate for 20 min before three control readings were made at 15 min intervals. The bathing solution was then replaced with the experimental solution containing a high or low K⁺ concentration and/or the test substances. For the purpose of this study, the effect of the various K⁺ concentrations and/or the test substances was tested on unstimulated tubules. Measurements were then taken over a 45-60 min period. Rates of secretion were expressed as a percentage of the third control rate reading.

Immediately after dissection, a portion of a Malpighian tubule (3-5 mm) was suspended in a Ringer bath between two holding pipettes. The peritubular bath(volume 300 μl) was perfused with Ringer solution at a rate of 1 ml min^-1. Intracellular (V_bl) and transepithelial(V_te) potentials were measured with 3 mol l^-1KCl-filled microelectrodes (borosilicate filament glass, Harvard; o.d. 1.2 mm,i.d. 0.69 mm; tip diameter <0.5 μm; resistance 20-40 MΩ) connected to a Micro Probe System M-707 electrometer (World Precision Instruments, New Haven, USA) via Ag/AgCl wire. The reference electrode was a coarse,low-resistance glass electrode (1 MΩ) filled with 3 mol l^-1KCl/agar (2%) connected to earth via Ag/AgCl wire. Cell impalement was accepted if a sudden drop in potential occurred, if the potential was stable for at least a few minutes and if the electrode potential differed by not more than 3 mV from the baseline after withdrawal. Transepithelial potential was measured by advancing the microelectrode through the cell layer into the lumen of the tubule. The apical membrane potential(V_ap) was calculated as the difference between the measured transepithelial and basolateral potentials.

Results are presented as means ± S.E.M., with the number of tubules(N) or number of measurements (n) in parentheses. The statistical significance of differences in fluid secretion or electrode potentials was evaluated by paired or unpaired Student's t-tests(two-tailed). A value of P<0.05 was accepted as statistically significant.

Under control conditions, tubules secreted spontaneously at a rate of 3.79±0.6 nl min^-1 (N=72; measured within the first hour after dissection). The mean V_bl was -24±0.43 mV (n=166), with reference to the bathing medium, and the mean V_te was 24±1.44 mV (n=72), lumen positive. The V_ap was not measured directly but was derived by subtracting the mean V_bl from the mean V_te. Therefore, V_ap in control Ringer was 48 mV, cell negative, with reference to the lumen.

Fig. 1 summarises the change in fluid secretion rates observed after changing [K⁺] and[Na⁺] in the bath. An increase in [K⁺] from 50 mmol l^-1 (control) to 140 mmol l^-1 (Na⁺ replaced by K⁺) increased the fluid secretion rate from 5.38±0.58 nl min^-1 to 7.64±0.60 nl min^-1 (P<0.02, N=8). Replacing the high [K⁺] Ringer with control Ringer reversed this effect. With a 10-fold drop in K⁺ concentration(K⁺ replaced by Na⁺), secretion rates dropped from 6.38±0.95 nl min^-1 to 1.48±0.52 nl min^-1(P=0.01, N=8). Fluid secretion rates did not recover after the low K⁺ medium was replaced with control Ringer, and some tubules stopped secreting altogether, illustrating the importance of K⁺ to the normal secretion of Tenebrio Malpighian tubules.

Fig. 2A shows the response of V_bl to varying K⁺ concentrations. The microelectrode was then advanced into the lumen and the effect of the same series of K⁺ concentrations was measured on V_te. The basolateral membrane was clearly sensitive to the bath [K⁺]. It depolarized from -24 mV in control Ringer (50 mmol l^-1 K⁺) to -13 mV in the presence of a high K⁺ concentration (140 mmol l^-1) and hyperpolarized to approximately -50 mV in 5 mmol l^-1 K⁺. The response of V_bl to a change in bath [K⁺] was immediate and is consistent with the presence of a significant K⁺ conductance in the basolateral membrane of the tubule cells, with the change in V_bl being 28 mV decade^-1(Fig. 2B).

Transepithelially, there was an increase in potential as the K⁺concentration increased from 5 mmol l^-1 to 140 mmol l^-1K⁺ (Fig. 2A). Fig. 3 summarises the results. From the electrical potential profile presented in this figure, it is clear that there is a correlation between V_ap and V_bl. The overall result of increasing bath [K⁺]from 5 mmol l^-1 to 140 mmol l^-1 was depolarisation of both V_bl and V_ap, the change in V_ap being 60-77% of the change observed in V_bl.

The presence of K⁺ channels in the basolateral membrane of Tenebrio tubules and their relative importance in fluid secretion were investigated by the addition of 6 mmol l^-1 Ba²⁺, a known K⁺ channel blocker, to the control bathing solution (50 mmol l^-1 K⁺). Fluid secretion rates decreased significantly from 4.17± 0.88 nl min^-1 to 0.76±0.24 nl min^-1 within 45 min in the presence of Ba²⁺(Fig. 4A; N=8, P<0.004). This drop in secretion rate was immediate, but the return of secretion rates to control levels after washout of Ba²⁺took approximately 45 min.

Fig. 4B shows a typical response of V_bl to Ba²⁺. Adding 6 mmol l^-1 Ba²⁺ to control bath Ringer caused a hyperpolarization of the basolateral membrane from -24±0.4 mV to-52±1.9 mV and a slight decrease of V_te (not shown)from 24±1.4 mV to 21±1.9 mV (n=47 and 34,respectively). The hyperpolarization of V_bl was sudden in some cells and more sluggish in others, the mean time being 4.6±0.4 min(n=47). The effect of Ba²⁺ on the potential profile was completely reversible after washout for 5-8 min, with V_bland V_te not significantly different from the initial control potentials. The addition of 6 mmol l^-1 Ba²⁺ to 140 mmol l^-1 K⁺ Ringer hyperpolarized V_bl significantly from -9±1.2 mV to -40±3.6 mV over a period of 11.6±2.2 min and caused a small but non-significant decrease in V_te from 30±7 mV to 27±7.8 mV(n=7 and 5, respectively; results not shown).

Two different responses were seen when 6 mmol l^-1Ba²⁺ was added to a low bath [K⁺] (5 mmol l^-1). In eight of the 21 impalements, the V_blhyperpolarized significantly from a mean value of -52±2 mV to-62±3 mV (P<0.01), while in the remaining 13 impalements a significant depolarisation from -60±2 mV to -46±3 mV was seen(P<0.001). Fig. 5illustrates both these responses. Possible reasons for this will be discussed later. The changes in the electrical profiles of Malpighian tubule cells bathed in the different K⁺ concentrations in the presence of Ba²⁺ are summarised in Fig. 6.

To further investigate the nature of the potassium conductance of the basolateral membrane, K⁺ ions were replaced by rubidium, a commonly used `permeant' substitute for potassium. However, replacing 40 mmol l^-1 K⁺ of the 50 mmol l^-1 K⁺control Ringer solution by 40 mmol l^-1 rubidium caused a hyperpolarization of the V_bl of 35 mV (results not shown). The addition of 6 mmol l^-1 rubidium to control Ringer had no effect on V_bl (n=4). Clearly, rubidium is not able to substitute for potassium in the Malpighian tubules of Tenebrio.

The possibility of K⁺ entering the cell through the Na⁺/K⁺/2Cl^- cotransporter was investigated by means of the loop diuretic bumetanide. Usually, a concentration of 10 μmol l^-1 bumetanide is sufficient to block the Na⁺/K⁺/2Cl^- cotransporter, and higher concentrations are needed to block the K⁺/Cl^-cotransporter (Palfrey and O'Donnell,1992). Bumetanide (10 μmol l^-1) significantly decreased fluid secretion rates of unstimulated tubules from 2.08±0.22 nl min^-1 to 0.55±0.09 nl min^-1(P<0.01, n=14). This inhibitory effect was reversible and after a washout period of 45 min, fluid secretion rates had recovered to 1.56±0.16 nl min^-1 (Fig. 7). The effect of bumetanide on V_bl and V_te was investigated in the presence and absence of Ba²⁺. None of the bumetanide treatments showed any significant effect on V_bl or V_te (n=5 and 7, respectively).

The contribution of the basolateral Na⁺/K⁺-ATPase to K⁺ uptake was investigated by blocking this ATP-dependent pump with 1 mmol l^-1 ouabain. K⁺ ions and ouabain compete for the same binding sites (Baker and Willis,1970); therefore, a high bath [K⁺] would decrease the effect of ouabain. Because, as previously shown, a low [K⁺] (5 mmol l^-1) irreversibly inhibits fluid secretion rates, the secretion assay was carried out in control Ringer (50 mmol l^-1K⁺). The secretion rates of unstimulated tubules decreased from 5.6±0.93 nl min^-1 to 3.0±0.6 nl min^-1after 15 min in the presence of 1 mmol l^-1 ouabain(Fig. 8). A further decrease to 2.26±0.36 nl min^-1 was observed after an additional 45 min(n=8). The inhibitory effect of ouabain on tubule secretion rates was irreversible. Basolateral and transepitelial membrane potentials showed no visible changes 10 min after the addition of 1 mmol l^-1 ouabain(n=5).

We have shown that fluid secretion in Malpighian tubules of Tenebrio appears to be primarily dependent on the presence of K⁺ in the bathing solution. In a low K⁺ (5 mmol l^-1), high Na⁺ solution, secretion was greatly reduced and the effect was largely irreversible. Tubules of the migratory locust Locusta migratoria responded in a similar fashion(Anstee and Bell, 1975). An increase in bath [K⁺] significantly increased the rate of secretion by Tenebrio Malpighian tubules, but a low Na⁺ (6 mmol l^-1) solution (in the presence of a high K⁺concentration) did not affect fluid secretion. This has previously also been found in Formica polyctena (Van Kerkhove et al., 1989) and the stick insect Carausius morosus (Pilcher, 1970). Haemolymph ion concentrations are very variable in insect species (see table 2 in Van Kerkhove et al., 1989),and these ion substitution experiments may mirror physiological conditions.

Sodium-dependent fluid secretion is primarily found in blood-sucking insects (e.g. Rhodnius and Aedes). These insects take huge,but infrequent, blood meals and need to rid themselves as soon as possible of the high NaCl and water load, which render them sluggish and easy prey. However, when not stimulated, tubules of Rhodnius secrete a K⁺-rich fluid containing only low levels of sodium ions(Maddrell and O'Donnell,1992).

The potential profile in Malpighian tubules of Tenebrio under control conditions is V_bl -24 mV, with reference to the bath, and V_te 24 mV, lumen positive. The result is a mean V_ap of 48 mV, cell interior negative. This potential profile closely resembles that described for tubules of Onymacris plana, another tenebrionid beetle(Nicolson and Isaacson,1987).

V_bl was very responsive to the K⁺concentration in the bath. Decreasing bath [K⁺] caused an immediate hyperpolarization, which was followed by an instant depolarization when the K⁺ concentration was again increased. Although V_bl recovered, fluid secretion of tubules previously treated with a low [K⁺] did not(Fig. 1). A possible reason for this could be cell shrinking, but there is no microscopical evidence for this. It must be remembered that V_bl only measures the electrical effects of a low [K⁺] at the level of the basolateral membrane, but for fluid secretion in general a low [K⁺] could affect numerous transport mechanisms. Lack of availability of K⁺ions could slow down the putative apical V-ATPase, which would in turn affect the performance of the putative apical cation/nH⁺ antiporter. During fluid secretion, fluid is secreted over the entire tubule length and it would therefore take a longer period of time for the mechanisms to recover and produce enough fluid within the tubule lumen to actually be measurable.

Basal K⁺ permeability is high in Malpighian tubule cells of most insect species studied so far (for a review, see Nicolson, 1993) and only the tubule cells of the yellow fever mosquito Aedes aegypti show a detectable Na⁺ permeability, which is further increased in the presence of cAMP (Sawyer and Beyenbach,1985).

When V_bl was plotted as a function of log[K⁺], a linear function was found with a slope of 28 mV decade^-1 (Fig. 2B)and not 58 mV decade^-1, as expected if K⁺ alone determined the membrane potential. Either other ions play a role and/or the intracellular K⁺ concentration drops in low [K⁺]. If other ions (e.g. Na⁺) had some importance, the dependence of V_bl to log [K⁺] would be expected to level off at lower [K⁺]. This is not the case. Thus, a drop in intracellular[K⁺] seems more likely. Leyssens et al.(1993) have shown that the intracellular [K⁺] in tubule cells of Formica can drop to 75% of the initial value when the [K⁺] in the bath is decreased 10-fold. In Tenebrio cells, a changing intracellular [K⁺]dependent on the bath K⁺ concentration thus seems feasible, but this will have to be confirmed by ion-selective microelectrode measurements. On the other hand, the intracellular [K⁺] in Malpighian tubule cells of Locusta seemed to remain fairly constant and was not affected by the bath [K⁺](Baldrick et al., 1988).

Furthermore, changing the bath K⁺ concentration indicated that V_ap closely followed V_bl (60-77% of the change observed in V_bl). This has also been reported for tubules of Formica (Leyssens et al., 1993). Weltens et al.(1992) have measured the resistance of the apical and basolateral membranes in Formica tubule cells and found that the resistance ratio of the apical membrane over the basolateral membrane is high. From the electrochemical model of an epithelium,it can be understood that the circular current, caused by the different electromotive forces across the basolateral and apical membranes and across the paracellular shunt, will contribute to the actual measured basolateral and apical voltage differences (Weltens et al., 1992). The effect will be small in the membrane with the lowest resistance and high in the membrane with the relative higher resistance. The basolateral membrane is dominated by a high K⁺conductance (low resistance) so that the circular current will have little impact on the measured potential and V_bl approaches the equilibrium potential for K⁺. The apical membrane, on the other hand, will be influenced appreciably by a change in electromotive force (and thus of circular current) and will `follow' this change. This mechanism may also be operative in the tubules of Tenebrio.

Barium is a frequently used K⁺ channel blocker and has been shown to reduce K⁺ movement across the basolateral membrane in Malpighian tubule cells of Formica(Weltens et al., 1992), Rhodnius (lower tubule; Haley and O'Donnell, 1997), Drosophila(O'Donnell et al., 1996), Aedes (Masia et al.,2000) and Locusta(Hyde et al., 2001). Applying barium will increase the basolateral resistance and therefore make the effect of the epithelial circular current more `visible' on this membrane, as explained in the previous section. The electromotive force of the apical V-ATPase sends a positive current into the lumen, making the cell interior more negative. In control [K⁺], the electrochemical gradient seems to be cell inward and, in the presence of barium, it becomes more difficult for K⁺ to cross the basolateral membrane and to compensate for this loss of positive ions from the cell. Also, the inward current across this larger resistance will cause a bigger potential drop across the membrane,hyperpolarizing it. This was substantiated by the study of Weltens et al.(1992) where the authors have shown that the hyperpolarization of the basolateral membrane in the presence of Ba²⁺ was abolished by bafilomycin A₁, an inhibitor of the V-ATPase.

In the present study, Ba²⁺ reversibly reduced fluid secretion by 83%. In control saline, basolateral and apical membranes responded by a marked hyperpolarization and V_te decreased slightly. This is most probably due to (1) blocking of the K⁺ channels in the basolateral membrane, (2) the increase in electrical potential difference created across the apical membrane, possibly by a putative proton pump (V-ATPase), and (3)increases in the basolateral resistance and the hyperpolarizing effect of the circular current as explained above.

The hyperpolarization of the basolateral membrane in some cells of Tenebrio tubules was notably slower than in others. This could indicate impeded access of Ba²⁺ to its site of interaction,possibly along the lateral spaces.

Different K⁺ channel types are mostly classified according to their single channel conductance or ligands rather than by their gating kinetics. The basolateral barium-sensitive K⁺ channels found in Tenebrio tubule cells appear to be specific for K⁺ ions and are impermeable to rubidium. This became evident when the equimolar substitution of K⁺ by rubidium caused a hyperpolarization of the basolateral membrane potential, similar to the hyperpolarization found in the presence of a low [K⁺]. If rubidium were able to substitute for K⁺, a slight depolarization of the basolateral membrane potential would have been expected after the addition of 6 mmol l^-1 rubidium to control Ringer, as this would have been sensed as an increase in total K⁺ concentration, but no effect was seen. In comparable studies,substitution of K⁺ ions with rubidium caused a 50% decrease in fluid secretion of Locusta tubules and a hyperpolarization of V_bl (Hyde et al.,2001; Pivovarova et al.,1994). This is in contrast to K⁺ channels found in the midgut of the tobacco hornworm Manduca sexta, where rubidium substituted for K⁺ ions to a greater extent(Schirmanns and Zeiske, 1994),and in the Malpighian tubules of the black field cricket Teleogyllus oceanicus, where a concentration of 8.6 mmol l^-1 rubidium caused a 10% increase in fluid secretion rates with rubidium almost completely replacing K⁺ in the secreted fluid(Marshall and Xu, 1999). In Locusta tubule cells, an increase in intracellular rubidium was seen when K⁺ was replaced by rubidium, but the rubidium was not transferred to the lumen via the apical K⁺/H⁺exchanger, indicating that the selectivity to K⁺ ions does not lie within the K⁺ channel but rather the putative apical K⁺/H⁺ exchanger responsible for transporting K⁺ to the lumen (Pivovarova et al., 1994). In contrast to the K⁺/H⁺exchanger of Teleogyllus tubule cells, this exchanger in Locusta cells appears to have a much higher affinity for K⁺ ions than for rubidium. Whether this is the same for tubule cells of Tenebrio has yet to be determined.

In the presence of a low bath [K⁺] (5 mmol l^-1),Ba²⁺ either hyperpolarized or depolarized V_bl. Providing that V_bl follows the Nernst potential for K⁺, for a V_bl of -52 mV, K⁺ would be at equilibrium if the intracellular [K⁺] were 40 mmol l^-1. This value is close to that found by Leyssens et al.(1993) in tubule cells of Formica. Any intracellular concentration below 40 mmol l^-1K⁺ would cause K⁺ to move into the cell. Blocking this inward movement with barium would result in a hyperpolarization (see Weltens et al., 1992). Cells that show a depolarization of the basolateral membrane in the presence of barium most probably have an outward electrochemical gradient for K⁺, the favourable electrochemical gradient created by the apical proton pump being too low to draw K⁺ ions into the cell. The addition of barium blocks the K⁺ ions from leaving the cell via the K⁺ channels and results in a depolarization of V_bl. This has also been found in the lower tubules of Rhodnius (Haley and O'Donnell,1997), which are responsible for ion reabsorption, and in the upper secreting tubule of the same insect(Ianowski et al., 2002). Secretion by the upper Malpighian tubules of Rhodnius is not inhibited by Ba²⁺ because it is driven by Na⁺ rather than K⁺ (Ianowski et al.,2002). Clearly, at physiological (50 mmol l^-1) and higher bath K⁺ concentrations, which we have shown to increase fluid secretion, the hyperpolarization of the basolateral membrane indicates that the net movement of K⁺ ions is from the bath into the cell.

Basolateral entry of ions via the Na⁺/K⁺/2Cl^- cotransporter has been implicated in Malpighian tubules of other insect species. In the present study, we tested the effect of bumetanide, which blocks this cotransporter, on fluid secretion rates of Tenebrio Malpighian tubules and found strong inhibition in non-stimulated tubules. Bumetanide also inhibits fluid secretion in cAMP-stimulated tubules of Rhodnius(O'Donnell and Maddrell, 1984; Ianowski et al., 2002) and Aedes (Hegarty et al.,1991), in stimulated and unstimulated tubules of Drosophila (Linton and O'Donnell,1999) and in unstimulated tubules of Locusta(Baldrick et al., 1988) and Formica (Leyssens et al.,1994). However, due to the relatively high concentration of bumetanide required to partially inhibit fluid secretion in Drosophila, Linton and O'Donnell(1999) suggested that bumetanide inhibited a K⁺/Cl^- cotransporter rather than an Na⁺/K⁺/2Cl^- cotransporter. According to the results of the fluid secretion assay, it seems highly likely that K⁺ ions cross the basolateral membrane of Tenebrio tubules via the Na⁺/K⁺/2Cl^- cotransporter,in addition to passage through channels.

Bumetanide affects electroneutral transport mechanisms, and the lack of a visible effect on membrane potentials supports this characteristic. This result is in accordance with findings of previous studies on Locusta(Baldrick et al., 1988) and Formica (Leyssens et al.,1994).

The role of an Na⁺/K⁺-ATPase in the Malpighian tubules of Tenebrio has been substantiated, given that ouabain, a specific inhibitor of the Na⁺/K⁺-ATPase, decreased fluid secretion irreversibly by 52%. Fluid secretion by unstimulated tubules of Aedes (Hegarty et al.,1991) and Locusta(Anstee and Bowler, 1979) is similarly affected in the presence of 1 mmol l^-1 ouabain. By contrast, ouabain stimulates fluid secretion in tubules of Rhodnius(Maddrell and Overton, 1988)and Drosophila (Linton and O'Donnell, 1999). Inhibiting the Na⁺/K⁺-ATPase disrupts its active role of transporting three Na⁺ ions from the cell interior to the outside and two K⁺ ions from the outside into the cell(De Weer, 1992), resulting in an increase in intracellular Na⁺ concentration(Maddrell and O'Donnell,1992). In stimulated tubules of Rhodnius, when fluid secretion is predominantly driven by the presence of a high [Na⁺],the presence of ouabain increases the secretion rate as well as the Na⁺ concentration in the secreted fluid(Maddrell and Overton, 1988). In both Rhodnius and Drosophila, V_bl depolarized in the presence of ouabain, indicative of an increase in intracellular Na⁺ concentration and, possibly, a decrease in intracellular K⁺ levels. Considering that the Na⁺/K⁺-ATPase provides a route of K⁺ entry into the tubule cells, blocking this pump decreases fluid secretion in insects, where K⁺ is the main player in driving fluid secretion.

In this study we have shown that basolateral K⁺ uptake is an important factor determining fluid secretion rates of TenebrioMalpighian tubules. K⁺ ions are transported across the basolateral membrane via barium-sensitive K⁺ channels and viathe electroneutral Na⁺/K⁺/2Cl^- cotransporter and the Na⁺/K⁺-ATPase. The nature of the basolateral K⁺ channels and the possible regulation of the various K⁺ uptake mechanisms by endogenous factors are subjects for further investigation.

Financial support was provided by a bilateral award (Bil98/53) under the Flemish—South African agreement on scientific and technological cooperation and by the South African National Research Foundation and the University of Pretoria.