ABSTRACT
The hindgut of the desert locust possesses an unusual chloride transport system. The isolated locust rectum absorbs chloride from the mucosal (lumen) to the serosal (haemolymph) side at a rate which is equal to the short-circuit current (Isc). Net chloride transport persists in nominally Na-free or HCO3(CO2)-free saline, is insensitive to normal inhibitors of NaCl co-transport and anion exchange, and is independent of the net electrochemical gradient for sodium across the apical membrane. However, active chloride transport is strongly dependent on mucosal potassium (K2 = 5·3mm-K). Chloride entry across the apical membrane is active, whereas the net electrochemical gradient across the basal membrane favours passive Cl exit from the cell. Although mucosal potassium directly stimulates ‘uphill’ chloride entry, there is no evidence for coupled KCL co-transport, nor would co-entry with potassium be advantageous energetically.
Net chloride absorption and IK are stimulated by a peptide hormone from the central nervous system which acts via cyclic-AMP. Cyclic-AMP increases Isc and approximately 1000% and transepithelial conductance (Gt) ˜100%. Approximately half of the ΔGt during stimulation results from increased Cl conductance at the basal cell border. This increase is also reflected in a shift of the basal membrane e.m.f. towards the Nemst potential for chloride. The remainder of the cAMP-induced ΔGt is due to an elevation of apical membrane K conductance, which causes a 400% increase in transepithelial potassium permeability as estimated by radiotracer diffusion. Because of this stimulation of K conductance, potassium serves as the principal counterion for active chloride transport under open-circuit conditions. Very high luminal levels of K oppose the stimulatory actions of cAMP on active Cl transport and K conductance. These and other results have been incorporated into a cellular model for KCL absorption across this insect epithelium.
INTRODUCTION
There is evidence for active chloride transport across six epithelia from 17 different insect species; however, little is known regarding the mechanisms which are responsible, or how these are controlled at the cellular level. In epithelia from vertebrate animals, models for active chloride transport generally involve uphill entry into cell by co-transport with sodium or entry of chloride in exchange for internal bicarbonate or hydroxyl ions. However, insect epithelia are unusual in that they often absorb chloride from external fluids that contain little sodium but which are extraordinarily rich in potassium [e.g. the hindgut of the desert locust (140mm-K; Phillips, 1964; Hanrahan, 1982), proximal segment of blowfly salivary gland (190mm-K; Oschman & Berridge, 1970), proximal segment otRhodnius Malpighian tubules (70 mm-K; Maddrell & Phillips, 1975) and pupal integument of lepidopterans (164mm-K; Jungreis, 1978; Cooper, Deaton & Jungreis, 1980)]. Recently, we studied active Cl absorption across the locust rectum by testing predictions of vertebrate Na-and HCCh-coupled mechanisms (Hanrahan & Phillips, 1982a). We consistently obtained negative or quantitatively unfavourable results and concluded that locust hindgut has an electrogenic, K-dependent mechanism for Cl transport. Both active Cl absorption and epithelial K conductance are apparently stimulated by a peptide hormone via cAMP. This paper briefly reviews the function of the rectal epithelium in locusts, describes evidence for this unusual Cl transport mechanism, and then discusses the regulation of KCL absorption with reference to a simple equivalent circuit model.
REVIEW OF TRANSPORT IN LOCUST RECTUM
As in most other insects, the hindgut of the desert locust is an important renal organ which functions in ionic and osmotic regulation (see Phillips, 1981). The locust excretory system is shown diagrammatically in Fig. 1. Approximately 200 blind-ended Malpighian tubules secrete a K-rich, primary urine into the hindgut lumen. The rate of tubular secretion is stimulated during feeding and flight by diuretic hormone, a small peptide (Mr about 2000) either from the corpus cardiacum (Mordue, 1970, and personal communication) or the suboesophageal ganglion (Proux, Rougon & Cupo, 1982). Most of the secreted fluid, containing wastes, ions and metabolites, moves posteriorly to the rectum, where it is modified by selective reabsorption of ions, amino acids and water (Phillips, 1964, 1981; Balshin & Phillips, 1971).
Locust rectum is a columnar epithelium consisting mainly of large cells (∼ 15 μm diameter, ∼ 100μm length) organized into six discrete pads. A thin layer of cuticle loosely covers the luminal (mucosal) surface and attaches to the tissue at the margins of each pad. The apical (lumen-facing) membrane does not form a brush-border as such; however, both apical and lateral membranes are highly infolded and are associated with numerous mitochondria. The rectal epithelium is richly supplied with trecheoles for gas exchange. Ion transport strongly depends on aerobic metabolism, and preferentially oxidizes proline which is available in Malpighian tubule fluid (Chamberlin & Phillips, 1982).
Phillips (1964) showed that chloride, sodium and potassium are actively reabsorbed from the lumen of ligated recta in situ. Transepithelial potential (Vt) and the rate of Cl transport in tissues mounted in Ussing-type chambers are similar to those measured in intact locusts (Williams, Phillips, Prince & Meredith, 1978). Williams et al. found that net chloride flux from the lumen (mucosal) side to haemocoel (serosal) side , and short-circuit current (Isc -uncorrected for saline resistance both declined exponentially from approximately 6·0 μequiv cm−2h−1 (after 30 min) to 3·5 μequivcm−2h−1 and 3·0μequivcm−1h−1, respectively after 3h in vitro. In contrast, Na absorption was ˜4pequivcm−2h−1 and did not change with time. The early decay of ISC was attributed to a decline in electrogenic Cl transport because high initial values of Isc were not observed when recta were bathed by chloride-free salines in which chloride was replaced by nitrate or sulphate. Further in vitro evidence for the electrogenicity of chloride absorption was obtained by Spring (Spring, Hanrahan & Phillips, 1978; Spring & Phillips, 1980a,b) who found that adding homogenates of the corpus cardiacum (CC) 1 ·5 h after dissection restored both Isc and to their initial values. Similar stimulations were later observed using haemolymph from recently-fed locusts (Hanrahan, 1978; Spring & Phillips, 1980c). A peptide hormone named CTSH (Chloride Transport-Stimulating Hormone; Mr=8000) has been isolated from corpus cardiacum and haemolymph (Phillips, Mordue, Meredith & Spring, 1980; Phillips et al. 1981). The actions of CTSH are apparently mediated by the second messenger cAMP because; (i) addition of cAMP increases Isc and (Spring & Phillips, 1980a; Hanrahan & Phillips, 1982a,b,c), (ii) exposure of recta to CTSH causes a three-fold elevation of tissue cAMP levels (Spring & Phillips, 1980a) and (iii) theophylline (which inhibits phosphodiesterase activity and thus cAMP degradation) is also stimulatory (Hanrahan, 1978, 1982). Fig. 2 shows the response of Isc (corrected for saline resistance) and mucosa-to-serosa Cl flux after addition of 1 mm-cAMP to the serosal side.
These studies provide strong evidence that Cl is actively absorbed and that transport is stimulated in vitro by CTSH and cAMP, but they give little information regarding the cellular mechanism of transport or its regulation. Before describing recent findings in locust rectum, it would be useful first to outline some models that have been proposed for active chloride transport in other tissues.
VERTEBRATE MODELS FOR EPITHELIAL CHLORIDE TRANSPORT
Many models of chloride absorption and secretion in vertebrates propose that ‘uphill’ entry of Cl into the cell is driven by entry of Na down its net electrochemical gradient through co-transport with Na (see Frizzell, Field & Schultz, 1979) or Na and K (Gregor, 1981; Musch et al. 1982). The inward sodium gradient is maintained by the ubiquitous Na/K exchange pump at the basal membrane. In those instances where transepithelial chloride transport is electrogenic, sodium that co-enters with Cl may recycle back to the mucosal side through a Na-selective, paracellular pathway (Field et al. 1978). Alternatively, Cl may enter via Cl/OH or CI/HCO3 exchange mechanisms, driven by primary active transport through the action of a membrane-bound ATPase (Bornancin, DeRenzis & Naon, 1980; Herrera et al. 1978; Humphreys Piou, 1979; Komnick, Schmitz & Hinssen, 1980; but see Bonting, de Pont, van Amelsvoor & Schrijen, 1980) or by secondary active transport, driven by exit of HCO3 or OH down their respective electrochemical gradients, or perhaps by parallel Na/H exchange. The anion exchange mechanism postulated for vertebrate epithelia is usually inhibited by thiocyanate (Epstein, Maetz & DeRenzis, 1973), disulphonic stilbenes such as SITS (4-acetamide-4&-isothiocyanostilbene-2,2&-disulphonic acid, e.g. Perry, Haswell, Randall & Farrell, 1981), and may be blocked indirectly by acetazolamide (e.g. Maetz & Garcia-Romeu, 1964), which inhibits carbonic anhydrase activity so that HCO3 supply becomes rate-limiting.
Some predictions of vertebrate models
If chloride co-enters with sodium, then (i) Isc and should be abolished or greatly reduced in nominally Na-free saline and (ii) they should be correlated with the level of Na contamination, (iii) There should be a 1:1 relationship between unidirectional influxes of 22Na and 36C1, and (iv) the net electrochemical gradient for Cl across the apical membrane should vary with and perhaps reverse when favours Na exit from cell to mucosa, (v) determined by measuring tracer fluxes, should depend strongly on (vi) Finally, the diuretic furosemide should inhibit Cl transport by blocking NaCl co-entry (Candia, 1973). Ouabain should reduce by inhibiting the basal membrane Na pump which would lead to a reduction .
If rectal Cl absorption involves anion exchange entry (which seems unlikely in view of the electrogenicity of transepithelial transport), then (i) removing exogenous CO2 and HCO3 should inhibit Cl transport unless metabolic CO2 is generated and converted to HCO3 at a sufficiently high rate. In this case, (ii) metabolic CO2 production must equal or exceed the rate of Cl entry, (iii) Because of the efficiency which would be required for CO2 hydration, one would expect involvement of carbonic anhydrase and thus sensitivity of Cl transport to the inhibitor acetazolamide, (vi) Finally, Cl entry should be blocked by anion exchange inhibitors such as SITS.
Testing the predictions of vertebrate models
Sodium
Recta were mounted in Ussing-type chambers and stimulated with 1 mm-cAMP. Bilateral replacement of normal saline (containing 115 · 3±0 · 9mm-Na) with nominally Na-free saline (0 049 mm-Na, N-methyl-D-glucamine substituted) had no effect on IK during the first 75 min exposure (Fig. 3A; P > > 0 · 2). After 5 · 5 h equilibration in nominally Na-free saline, serosal addition of 1 mm-cAMP caused large increases (> 5 μ equivcm−2h−1) in Isc and Results of longterm exposure to Na-free con ditions were complicated by other side effects: stimulations after 5 · 5 h exposure were variable and not different statistically from those observed in normal saline, although the mean values of Isc and during maximal stimulation were approximately 20% and 40% lower, respectively.
A gradual loss of transport under Na-free conditions would be expected because the prolonged viability of locust rectum in vitro depends on amino acids, and amino acid absorption is Na-coupled in this tissue (Balshin & Phillips, 1971). Consistent with this bypothesis, lower , Isc and Rt are also observed when recta are exposed for similar periods to amino acid-free saline containing normal levels of Na ( 114 mm) and glucose (10 mm; Hanrahan, 1982). Other explanations for poor viability during longterm exposure to Na-free conditions might include disruption of cell volume or intracellular pH regulation. No correlation was observed between final trace levels of sodium and cAMP-stimulated and Isc (5 – 9μequivcm−2h−1) which were measured at the end of flux experiments under nominally Na-free conditions (<1– 200 μM-Na).
Initial rates of 36C1 and 22Na influx into the tissue of everted rectal sacs (mucosal side out) were measured with corrections for extracellular radioactivity using 3H-mannitol space. Injection of 10 mm-cAMP into sacs had no effect on the low rate of 22Na influx (>0·8nequivmg−1 h−1). In contrast, 36C1 influx increased from 0·8 to 4·9 ± 1·6 nequivmg−1 h−1. This six-fold elevation in 36C1 influx without a concomitant stimulation of æNa influx is not consistent with obligatory NaCl co-entry across the apical cell border.
In a separate series of experiments, intracellular Cl and Na activities and potentials were measured using double-barrelled ion-sensitive microelectrodes while sodium concentration was varied on the mucosal side. Transepithelial potential was clamped at 0 mV during these experiments and the net electrochemical potentials for Na and Cl across the apical membrane were calculated. When muco sal Na concentration was lowered from 115 mm to 49 μM, the calculated declined, reversing when mucosal Na concentration was less than 1 mm (see Hanrahan & Phillips, 19826). In contrast, remained relatively constant and was 38 mV above equilibrium when favoured Na exit rather than entry into the cell. Clearly, NaCl co-entry from the mucosal side is not required in order to maintain intracellular Cl above equilibrium.
To test directly the effects of varying on transepithelial chloride transport, recta were mounted in Ussing-type chambers and Vt was clamped at 0 mV. Unidirec tional fluxes of 36C1 were measured at 15-min intervals, first in normal saline (115·3 ± 0·9mm-Na) and then when Na was reduced stepwise on the mucosal side to 2·18 ± 0·74 mm(N = 14) by replacement with N-methyl-D-glucamine. In other recta, double-barrelled microelectrodes were used to measure intracellular Na activity and potential under conditions that were identical to those during tracer flux experiments. Reducing the net inward gradient for sodium across the apical membrane from 118 ±0·9 mV to 16 + 1·3 mV had no significant effect on (from 8·97 ± 1·03 to 9·02± 1·36μequivcm−2h−1; x̄±S.E.; N = 7P>0· 2) and , decreased slightly (from 2·06 ± 0·48 to 1·39 ± 0·24 μequivcm−2h−1; X±S.E.; N= 7; P<0·l).
Mucosal addition of the inhibitor furosemide had no effect on Isc after 1·5 h ex posure (P> 0·2; Fig. 3B). Note that the cuticular intima was dissected away from the luminal surface to ensure access of the furosemide to the apical membrane in these experiments. Exposure to 1 mm ouabain for 1 h did not reduce the stimulation of and Isc by cAMP and actually caused a small but significant stimulation of Isc and net (0·34 and 0·28μequivcm−2h−1; P<0·01). This small, ouabain-induced change in IK is in the direction expected for inhibition of electrogenic Na absorption. Many insect epithelia are relatively insensitive to ouabain (see Harvey et al. 1983, this volume); however, homogenates of locust rectum have considerable ouabain sensitive Na/K ATPase activity (K, = 106M; Peacock, 1981), and intracellular Na and K activities under Isc conditions in the presence of cAMP (8·01 ± 0·41 and 74·8 ± l·2mm, respectively) are similar to those for most other cells that possess a basolateral Na/K exchange pump (Hanrahan & Phillips, 1982b). These results provide no evidence for a Na-dependent Cl transport mechanism in locust rectum.
Bicarbonate
Transepithelial 36C1 fluxes were measured across short-circuited recta after equilibration in normal saline containing l0mm-HCO3 (stirred with 95% O2/5% CO2), or nominally HCO3-free saline (stirred with 100% O2). Fig. 4 shows that Cl fluxes were not affected by exposure of recta to HCO3,CO2-free saline after 7h. Moreover, there is no reason to propose anion exchange because Δ Isc equals Δ after stimulation.
Could enough HCO3 be derived from metabolic CO2 to supply an anion exchange machanism? Chamberlin (1981) measured O2 consumption by cAMP-stimulated recta under virtual short-circuit conditions. Assuming that locust rectum has a maximal respiratory quotient of 1 ·0 and that all metabolically-derived CO2 is conver ted to HCO3 and is made available to an apical membrane exchange process, the supply of HCO3 would be ˜3 ·2 μequiv cm−2 h−1 or only 40% of under nominally HCO3,CO2-free conditions (Fig. 4).
Finally, exposure of cAMP-stimulated recta to 1 mm-acetazolamide for 1 h in nor mal saline had no effect on Isc. Cyclic AMP-stimulated Isc was also insensitive to 1 mm-SITS when added to both sides (Hanrahan & Phillips, 1982b).
In summary, data obtained from a variety of different techniques suggest that the major fraction of transrectal Cl transport does not involve a standard Na-or HCO3-coupled mechanism.
ACTIVE CHLORIDE TRANSPORT IS K-DEPENDENT IN LOCUST RECTUM
Serosal addition of 1 MM-CAMP to recta equilibrated in nominally K-free saline (5 ·-5 h) stimulated and Isc to 30% of controls (Fig. 5A). When K was restored to both sides by bilateral addition of 10mm-K-methylsulphate, Isc and increased to normal levels. We measured the effects of adding K-methylsulphate to only one side of locust rectum and then corrected for potassium diffusion by measuring currents produced by adding K-methylsulphate asymmetrically under nominally Cl-free conditions and to azide-poisoned tissues. Low levels of K (0 – 10 nw) were stimulatory only when added on the mucosal side. The concentration-dependence of K-stimulated Isc was measured by equilibrating recta for 4 h in K-free saline under Isc conditions, stimulating Cl transport with 1 mm-cAMP on the serosal side, and then adding K-methylsulphate bilaterally. The K-dependent component of Isc was determined by comparing the effects of bilateral K-methylsulphate addition with those of Na-methylsulphate addition. A Lineweaver-Burke plot revealed that ˜ 5 · 3 mm external K caused half-maximal stimulation of Isc. Fig. 5B shows the net electrochemical gradient for chloride across the apical membrane as calculated from intracellular potential and Cl activity measured (i) in the absence of cAMP and K, (ii) after addition of 1 mm-cAMP and (iii) after addition of 2 mm-K to both sides in the presence of cAMP. Short-circuit current measured in parallel experiments under identical conditions is shown in Fig. 5C. Net Cl flux and opposing Cl entry both increase during exposure to cAMP and are further stimulated by addition of 2 mm-K. Thus, both these agents apparently stimulate active Cl entry directly. While Cl transport is dependent on mucosal K, there is no evidence for KCL co-entry: Firstly, is only 0 · 1 mV at 10 mm external K (compared to an opposing of 40 mV) and is in the wrong direction to permit passive K recycling back to the mucosal side at higher concentrations (40 – 140 mm-K). Recycling must be postulated if KCL co-transport occurs to explain why is several times under short-circuit conditions (Hanrahan & Phillips, 1982a). Secondly, co-transport with potassium would not supply energy for ‘uphill’ Cl entry (i.e. although an electroneutral cotransport step would effectively remove the electrical potential opposing chloride entry, KCL influx would then be opposed by an equally large potassium gradient). Lastly, due to the very high rate of Cl transport during cAMP stimulation (˜ 10 μ equivcm−2h−1), a large K flux should pass through the postulated KCL pathway if Cl entry occurs by KCL co-transport: thus transepithelial K permeability would be partially Cl-dependent. However, chloride removal had no effect on transepithelial 42K fluxes under Isc conditions (see below and Fig. 7B). At present, the simplest explanation for the effects of potassium on transepithelial Cl transport is that external K enhances electrogenic Cl entry, perhaps by activating an ATPase in the apical membrane, without being co-transported. This scheme would be analagous to enzyme activation in which the Ka for mucosal potassium is 5 · 3 mm.
Transepithelial chloride absorption is describable by Michaelis-Menten kinetics, but Kt and depend on external K concentration. In nominally K-free saline, the transepithelial Kt was 22 · 7 ± 4·0mm-Cl and = 3·5 ± 0·7 µequivcm−2h−1. When the external K concentration was raised to 100 mm, K increased 4·-5-fold to 99·6 ± 13·4 mm-Cl and was 23·1 ± 5·-3μequivcm−2h−1. Locust rectum may possess a single Cl transport mechanism that is stimulated by K but which also operates slowly under K-free conditions. Alternatively, there may be two systems, one which is K-insensitive and the other made up of pump sites that respond to external K in a graded way or are gradually recruited as [K] is elevated. Regardless of the molecular details, these results show that the major fraction of Cl absorption across locust rectum occurs by an unusual mechanism which is stimulated by mucosal potassium.
CYCLIC-AMP REGULATES PASSIVE PERMEABILITY IN LOCUST RECTUM
There is evidence that locust rectum is a tight epithelium having low electrical resistance (Hanrahan, Phillips & Steeves, 1982). Qualitative evidence for tightness was obtained by briefly exposing tissues to transepithelial single-salt gradients and recording the deflection in transepithelial potential (ΔVt). If locust rectum is leaky, ΔVt would be similar regardless of whether the gradient was directed from mucosa-to-serosa or serosa-to-mucosa, because the paracellular pathway does not normal specify. Isosmocity was maintained by the addition of sucrose to low-salt solutions, and all salines were vigorously oxygenated, and contained 1 mm-cAMP and amino acids at their haemolymph concentrations. Results were consistent with high cation selectivity at the mucosal surface and only slight selectivity for K over Cl on the serosal side. This asymmetry would be consistent with transcellular diffusion across apical and basal membranes having different ionic permeabilities rather than through a single paracellular pathway.
The absence of non-selective shunts is further suggested by very low transrectal permeability to 35SO4 (<3·5 × 10−7 cms−1). This value is less than 1% of that expected from measurements of 42K fluxes after correcting for K and Cl mobilities in free solution, on the assumption that K and SO4 diffuse through the same non-selective shunt pathway (i.e. edge damage). Voltage scanning the epithelial surface while passing large transepithelial currents did not reveal significant shunts between cells, rectal pads or at the edge of the tissue where it attaches to the chamber; however, holes made by deliberately damaging the epithelium were easily detected.
Finally, membrane and paracellular resistances were measured by flat-sheet cable analysis after demonstrating cell-cell coupling with fluorescent dye and electrical methods (Hanrahan et al. 1982). Intraepithelial spread of current was measured by injecting current pulses into one cell and measuring the resulting deflections in apical membrane potential at variable distances using a second microelectrode. The data were plotted semi-logarithmically against interelectrode distance and fitted to a Bessel function (see Fromter, 1972). Assuming that locust rectum may be treated as a syncytium (which is supported by demonstration of electrical coupling between cells; Hanrahan, 1982) and may be modelled as a thin flat sheet, apical and basal membrane resistances were calculated to be 220 ± 30·3 and 193 ± 12 Ωcm2, respectively, and paracellular resistance was 742 ± 141 Ωcm2. The low macroscopic resistance of rectal cell membranes probably results from extensive infolding, which, according to electron micrographs, should cause between 9-and 200-fold amplification of surface area in various parts of the cell. During exposure to 1 MM-CAMP in normal saline, both apical and basal membrane resistances declined by ˜80% to 44 and 36 Ωcm2, respectively, without significant change in paracellular resistance. Consequently, the fraction of transepithelial ionic diffusion through paracellular pathways declines from 40% to ˜5% during exposure to cAMP so that locust rectum effectively becomes ‘tighter’ during stimulation.
Fig. 6A shows the effects of adding 1 mm-cAMP on transepithelial potential (Vt), and apical and basal membrane potentials (Va and Vb, respectively) when bathed by normal saline. Deflections in Va, Vb, and Vb were produced by passing constant-current pulses from mucosa-to-serosa: they suggest that transepithelial resistance declines during stimulation with cAMP as a result of reductions in both apical and basal membrane resistances as described above. The decline in basal membrane resistance during cAMP exposure (Fig. 6B) was abolished under Cl-free conditions (compare with Fig. 6C), suggesting that cAMP increases basal membrane Cl conduc tance. The increase in apical membrane conductance was not dependent on chloride and therefore was not associated with electrogenic ion transport. Furthermore, cAMP reduced Rt under nominally Cl-free conditions when no change in Isc was observed below, Fig. 7A).
We examined the possibility that cAMP increases apical membrane K permeabil ity. Under Cl-free conditions, the cAMP-induced decline in Rt coincided with ˜ 400% stimulation of transepithelial 42K fluxes (Fig. 7A, B). Restoration of chloride did not stimulate or although Isc increased ten-fold to control levels. These results indicate that cAMP enhances rectal K permeability in addition to stimulating basal membrane Cl conductance. Consistent with this notion, cAMP did not reduce Ra when K was omitted from ‘Cl-free’ saline (Hanrahan, 1982).
Hormonal stimulation of potassium conductance would allow K to serve as a more effective counterion for electrogenic Cl absorption. Net Cl absorption is ten-fold higher than under Isc conditions; however, they are similar under open-circuit conditions (4 · 3 ±0 · 9 and 4 · 5 ± 0 · 4 μ equivcm−2h−1, respectively; N.S., P > 0 · 2). The actions of cAMP on electrogenic Cl transport and passive K conductance are thus complementary, and would provide much tighter control over KCL absorption under open-circuit (in vivo) conditions than would be possible by stimulating active trasport alone.
There is some net absorption of K from mucosa-to-serosa during cAMP stimulation under Isc conditions, but this is low compared to net chloride absorption (μequivcm−2h−1, μequivcm−2h−1). It was not possible to determine whether the active step in K transport occurred at the apical or basal membranes because net electrochemical gradients across both membranes favour K flux in the mucosa-to-serosa direction under open-circuit conditions, whereas potassium is near equilibrium across both cell borders under Isc conditions. In any case, it should be emphasized that the major effect of cAMP on K transport is the stimulation of apical membrane K permeability. Under identical conditions, the measured rate of Na transport is low (1 μ equivcm−2h−1) and is unaffected by cAMP (K. Black & J. E. Phillips, unpublished observations).
Local control mechanisms in this epithelium are discussed in detail elsewhere (Hanrahan & Phillips, 1982b) but deserve brief mention to complete this summary of our present understanding of locust rectum. Above 100 mm and in the presence of cAMP, K has an increasingly inhibitory effect on Cl transport, and apical membrane K permeability varies inversely with luminal concentration of this cation. Thus excess reabsorption of KCL is automatically prevented should K levels be high in midgut and Malpighian tubule fluid entering the rectum. Finally, the rectal contents can become acidic (pH ˜ 4 · 5) and strongly hyperosmotic (1200mosmoll−1) in situ (Phillips, 1964) and both these conditions inhibit Cl-dependent Isc across cAMP-stimulated locust rectum (Hanrahan & Phillips, 19826; Hanrahan, 1982).
AN EQUIVALENT CIRCUIT MODEL FOR THE LOCUST RECTAL
EPITHELIUM
The model shown in Fig. 8A is based on results obtained using a variety of techniques. It features three cAMP-regulated elements: (i) an electrogenic Cl pump at the apical membrane, (ii) an apical membrane K conductance and (iii) a basal membrane Cl conductance. Also shown is a K conductance pathway in the basal membrane which is presumably the route of K exit for transepithelial K absorption under open-circuit conditions and for most K entering the cell from the serosal side through the Na/K exchange pump. Fluctuation analysis has shown this conductance to result from K channels (˜ 150 × 10 −6 channels per cm2 macroscopic area) which are blocked by barium on the serosal side (Ki= 2 · 9mm-Ba; Hanrahan, Wills & Lewis, 1983).
If this model is to be consistent with previous ion-sensitive microelectrode and tracer flux data, it should predict the electrical behaviour of locust rectum; for example, the effect of cAMP on Vt under open-circuit conditions, and the intracellular potential of cAMP-stimulated recta under Isc conditions.
The effects of cAMP on Ea and Eb are probably due to stimulation of active Cl transport rather than K permeability because Ea reaches a value during cAMP stimulation that is 12 mV more negative than the Nernst potential for K across the apical membrane . Furthermore, Cl-removal in the continued incsence of cAMP restores Ea and Eb to approximately their pre-cAMP values (56·3and 53 mVa respectively). Finally, cAMP stimulations of Vt during these experimewpr were similar to those observed previously, when ten-fold stimulations of were measured using tracers and Vt was obtained from short-circuited preparations during brief interruptions in Isc (from 6·4 ± 0·6 mV to 27·1 ± 1·4 mV; x̄ ±S.E., N = 20 recta) or from continuously open-circuited preparations (from 8·3 ±1·2mV to 24·7 ± 2·2mVa x̄ ±S.E., N = 12 recta).
From Table 1 and equations 3 and 4, we calculate that Ea would increase from 59·8 mV to 69·5 mV and Eb would decrease from 54·1 mV to 37·0 mV during exposure to cAMP. These values are in good agreement with those based solely on electrical measurements (Table 1).
In summary, the circuit models shown in Fig. 8 are internally consistent in that e.m.f. values calculated from ion activities and tracer fluxes agree well with those calculated strictly from electrical data. These models also seem realistic since they accurately predict the effects of cAMP on transepithelial potential and the intracellular potential in short-circuited tissues.
CONCLUSIONS
Fig. 9 shows the mechanism of KCl absorption proposed for locust rectum. The rate of KCl transport depends on mucosal K concentration because K directly stimulates active Cl entry across the apical membrane (Fig. 5A, B; Hanrahan & Phillips, 19826). Interestingly, mucosal potassium indirectly enhances its own passive reabsorption by stimulating electrogenic Cl transport, since K is the counterion for active Cl absorption under open-circuit conditions (Hanrahan & Phillips, 1982a). Simultaneous hormonal control of active Cl transport and passive K permeability by the hormone CTSH would provide remarkably efficient control over salt reabsorption and presumably ionic regulation in the desert locust.
ACKNOWLEDGEMENTS
We thank Dr S. A. Lewis for helpful discussions and Dr N. K. Wills for critically reading the manuscript. This work was supported by a postgraduate scholarship to J. W. H. and operating grants to J. E. P. from the Natural Sciences and Engineering Research Council (Canada).