Electrolyte Transport Through a Cation-Selective ion Channel in Large Intestinal Enterocytes of Xenopus Laevis

In numerous vertebrates, much of the transepithelial electrical potential difference across the large intestine is generated by electrogenic Na⁺ absorption (Groot and Bakker, 1988), so that sodium ions diffuse into the epithelial cell through specific apical Na⁺ channels, which are highly sensitive to the epithelial Na⁺ channel blocker amiloride (Benos, 1986). The sodium ions are then extruded to the interstitium by the basolateral Na⁺/K⁺-ATPase. In amphibians, this transport has been intensively studied in vitro in both toads (Bufo sp.) (Cofré and Crabbé, 1967; Dawson and Curran, 1976) and frogs (Rana sp.) (Krattenmacher and Clauss, 1988, 1989; Krattenmacher et al. 1988). Electrogenic Na⁺ transport in the colon of these amphibians follows the general model mentioned above. However, in the African clawed toad (Xenopus laevis) the apical entry step differs considerably from this model. A previous study (Krattenmacher et al. 1990) demonstrated that, in Xenopus, Na⁺ passes the apical membrane of the colonic epithelial cells through an amiloride-insensitive ion channel. Furthermore, the amount of Na⁺ absorption is increased by about 120 % in the absence of mucosal calcium. Thus, electrogenic Na⁺ absorption in the amphibian colon does not follow a general and uniform mechanism.

In intestinal epithelia a similar Ca²⁺ sensitivity has only been reported in the rumen of sheep (Martens et al. 1990) and it has not been described in the large intestine. In non-intestinal epithelia it was described in the skin of larval (Hillyard et al. 1982) and adult frogs (Van Driessche and Zeiske, 1985) and in toad skin (Aelvoet et al. 1988) and urinary bladder (Van Driessche, 1987; Van Driessche et al. 1987; Aelvoet et al. 1988; Das and Palmer, 1989). In these amphibian tissues, micromolar concentrations of ‘mucosal’ calcium blocked this channel, which, in the absence of mucosal calcium, was also permeable to other monovalent cations. Van Driessche (1987) speculated that this channel might represent an epithelial Ca²⁺ channel, because similar properties have been described for Ca²⁺ channels in frog muscle fibres (Aimers et al. 1984; Aimers and McCleskey, 1984; Hess et al. 1986). A recent patch-clamp study by Das and Palmer (1989) indicated that the channel seems to be a non-selective cation channel that may be involved in Na⁺-absorbing or K⁺-secreting processes. However, the physiological function of this Ca²⁺-sensitive channel is unclear. In contrast, in the colon of Xenopus, Krattenmacher et al. (1990) demonstrated that the Ca²⁺-sensitive pathway plays a significant role in electrogenic Na⁺ absorption, even in the presence of mucosal calcium. In this study we have looked at the physiological role of this channel by investigating the Ca²⁺-sensitive pathway in the colon of Xenopus in more detail.

Breeding adult frogs (Xenopus laevis; Horst Kähler, Hamburg, FRG) of either sex were used for the experiments. They were kept at room temperature in tap water and fed once a week with commercial cat food (Brekkies, EFFEM GmbH Verden, FRG). The frogs were pithed, the colon was quickly removed, opened along the mesenteric border and rinsed with frog Ringer’s solution. The tissue was stretched and glued (Histoacryl Blue, Braun Melsungen, FRG) serosal side downwards onto a Lucite ring. The inner diameter of the ring was 8 mm. The preparations were mounted in an Ussing chamber which was specially designed to avoid edge damage (De Wolf and Van Driessche, 1986). Silicone grease (Bayer Silicone) was used to seal the edges on both sides of the mounted tissue. Effective tissue area was 0.5 cm². The mucosal and serosal chamber compartments were continuously perfused with frog Ringer’s solution (room temperature) at a flow rate of about 5 ml min⁻¹.

The spontaneous transepithelial potential difference was clamped to zero with a voltage-clamp unit (Elke Nagel, Biomedical Instruments, München, FRG). The resulting short-circuit current (I_sc) and the transepithelial conductance (G_T) were continuously recorded by a standard stripchart recorder. G_T was calculated by the voltage-clamp unit from deflections in transepithelial voltage and current, which were induced by superimposed voltage pulses of 10 mV amplitude and 200 ms duration. Pulse frequency was 0.5 Hz. Pulse polarity had no effect on G_T values.

The serosal side of the tissue was always exposed to a NaCl-Ringer’s solution with the following composition (in mmoll⁻¹): Na⁺, 115; K⁺, 2.5; Ca²⁺, 1.0; Cl⁻, 117; HCO₃⁻, 2.5 and Hepes buffer, 8 {pH 8.0, adjusted with Trizma base [Tris-(hydroxymethyl)-aminomethane]}. The normal mucosal Ringer’s solution was of the same composition. However, for selectivity experiments, the main cation Na⁺ was substituted by equimolar amounts of Li⁺, K⁺, Rb⁺ or Cs⁺, added as chloride salts. For Ca²⁺-free Ringer’s solutions 1 mmol I⁻¹ CaCl₂ was omitted and 0.5mmoll⁻¹ EGTA was added. In experiments where pH dependency was investigated, the pH of the mucosal Ca²⁺-free Ringer’s solutions (6, 6.5,7, 7.5 and 8) was adjusted with Trizma base. According to Portzehl et al. (1964), free (ionized) calcium concentrations ([Ca²⁺]_free) were controlled in the micro-and submicromolar range by utilizing the buffering ligands EDTA (at pH 6.5) and HEEDTA (N-hydroxyethylethylenediamine-triacetic acid, at pH8.0). These ligands have four negative charges (L⁴⁻) which may bind H⁺ and/or Ca²⁺. Only the ligands with four (L⁴⁻) and three (HL³⁻) negative charges were considered in the calculations concerning Ca²⁺ binding because the affinity of Ca²⁺ for the ligands H₂L²⁻ and H₃L⁻ can be neglected. The true association constants of Ca²⁺ for the ligands L⁴⁻ and HL³⁻ were designated as K_CaL and K_CaHL.

Quinidine was dissolved in dimethyl sulphoxide (DMSO) and added to the mucosal solution at a final concentration of 0.1–3.2 mmol I⁻¹. Verapamil was dissolved in water and used in the mucosal solution in the concentration range 25–500 μmol I⁻¹. The solubility of verapamil was facilitated by adding about 10 μmol I⁻¹ Tween 80 [polyoxyethylen-(20)-sorbitanmono-oleate]. Tween 80 was without effect on short-circuit current. All chemical compounds, except standard salts (Fluka, Buchs, CH), were obtained from Sigma (München, FRG).

Results are expressed as mean±standard error of the mean (S.E.M.). N designates the number of experiments. Statistical analysis was done using Student’s t-test (if possible for paired observations) with a significance level of P⩽0.05.

After about 30-45 min of equilibration of the voltage-clamped tissue in Camcontaining (1 mmoll⁻¹) NaCl-Ringer’s solution (pH8.0), short-circuit current (I_sc) and transepithelial conductance (G_T) reached stable values (Table 1). When Ca²⁺ was removed from the mucosal solution after the equilibration time, I_sc increased significantly by about 80%. This increase was accompanied by a significant increase (about 70%) in G_T. Both effects were fully reversible on the re-addition of 1mmoll⁻¹ Ca²⁺. This phenomenon has already been described by Krattenmacher et al. (1990), who demonstrated that the current increase is fully explained by a stimulation in electrogenic Na⁺ transport from the mucosal to the serosal side of the epithelium. Interestingly, the amount of this Ca²⁺-sensitive stimulation is strongly dependent on the mucosal pH. Experiments in which the pH of the mucosal Ca²⁺-free solution was increased in steps (6.0, 6.5,7.0,7.5, 8.0) showed a similar stepwise increase in I_sc (Fig. 1). A plot of the Ca²⁺-sensitive current versus the mucosal pH revealed a strong linear relationship (Fig. 2). Because this effect of pH had already been observed in preliminary experiments, all solutions used in this study contained 8 mmol I⁻¹ Hepes buffer to achieve constant pH values. Although in some experiments pH also had a small effect on I_sc, even in the presence of mucosal calcium, a clear pH dependency of the Ca²⁺-insensitive current was not found. Thus, the pH-induced current changes were regarded as changes in Ca²⁺-sensitive current.

For further characterization of the Ca²⁺-sensitive current, we blocked the stimulated short-circuit current in a dose-dependent manner by increasing mucosal Ca²⁺-concentrations. The time course of changes in the I_sc in such an experiment is shown in Fig. 3. Accurate micromolar Ca²⁺-concentrations were obtained using the Ca²⁺-buffers EDTA (at pH 6.5) and HEEDTA (at pH 8) (for details see Materials and methods). Analysis of the dose-response curves using the direct linear plot (DLP) showed that the stepwise inhibition of the I_sc follows Michaelis-Menten-type kinetics (Fig. 4). Using this method we were able to calculate the Michaelis constant of the Ca²⁺-induced inhibition ⁠. Because of the pH dependency demonstrated in Fig. 2, we calculated at pH values of 6.5 and 8.0. The results of this analysis are given in Table 2. The Michalis constant was less than 1μmoll⁻¹ at both pH values and there was no significant difference between the two. The direct linear plot method gives a calculated value for maximal ⁠. Because this calculated value [indicated by (DLP) in Table 2] is similar to the measured value of [indicated by (measured)], the direct linear plot method can be regarded as an appropriate method for calculating in this study. Although clearly increases with increasing mucosal pH, as shown in Fig. 2, the difference between at pH 6.5 and that at pH 8.0 in Table 2 is not significant. This might be explained by the small number of experiments and the wide scatter of the values, at least at pH 8.0.

To investigate whether divalent cations other than Ca²⁺ can block this channel, we substituted all mucosal Ca²⁺ with Mg²⁺. Fig. 5 shows that the short-circuit current did not change when the mucosal Ringer’s solution was changed from Ca²⁺-Ringer (1 mmoll⁻¹ Ca²⁺ as the divalent cation) to Mg²⁺-Ringer’s solution (1 mmol I⁻¹ Mg²⁺ as the divalent cation). In both cases, the removal of either Ca²⁺ or Mg²⁺ induced a similar increase in I_sc. Furthermore, the stimulation of I_sc by removal of mucosal Ca²⁺ could be fully reversed by the addition of 1 mmol I⁻¹ Ba²⁺ (Fig. 6). These experiments clearly show that the current flowing through this channel is sensitive not only to calcium but also to other divalent cations such as magnesium and barium.

To investigate the permeability of the Ca²⁺-sensitive pathway, a series of experiments was performed in which the main mucosal cation (usually Na⁺) was replaced by other monovalent cations of the alkali metal group, such as Li⁺, K⁺, Rb⁺ or Cs⁺. Table 3 gives the values of the short-circuit current and the corresponding transepithelial conductance measured in mucosal Ringer’s solutions with one of these ions as the main monovalent cation. The main cation of the serosal Ringer’s solution was in all cases Na⁺. The mucosal change from NaCl-Ringer’s solution to RbCl-or LiCl-Ringer resulted in a significant decrease of short-circuit current (Table 3). In the case of CsCl-Ringer, I_sc also decreased in five of the six experiments, but because the results showed a wide scatter, the decrease was not found to be significant (Table 3). In contrast, the mucosal change from NaCl-to KCl-Ringer left I_sc relatively unchanged. It is not possible to establish the real selectivity of the ion channel for monovalent cations because the transepithelial ion movements include not only the apical but also the basolateral membrane. However, in all cases, except with lithium, a significant calcium sensitivity was found. I_sc increased significantly when mucosal Ca²⁺ was removed and either Na⁺, K⁺, Rb⁺ or Cs⁺ was the main mucosal cation, but the extent of stimulation varied considerably (Table 3). in the K⁺-Ringer was about twice as high as in the Na⁺-Ringer experiments. In contrast, when Cs⁺ was the main cation, was only about 50% of the value when Na⁺ was present. Concomitant with the increase in I_sc, an enhanced transepithelial conductance (GT) was measured (Table 3). After removal of mucosal Ca²⁺, G_T was significantly higher in the Na⁺-, K⁺- and Rb⁺-containing solutions. In the Cs⁺-containing solution, the increase in G_T was not significant. In contrast, in the Li⁺-containing solution, where I_sc was not Ca²⁺-sensitive, G_T also remained unaffected. Obviously, this Ca²⁺-sensitive channel is, at least in the absence of mucosal Ca²⁺, not a specific Na⁺ channel but a cation channel, permeant for several other monovalent cations, particularly K⁺.

For a further pharmacological characterization of the Ca²⁺-sensitive channel, we added several organic ion-channel blockers to the mucosal Ringer’s solution. The Ca²⁺-sensitive part of the short-circuit current was completely blocked by quinidine (Fig. 7), which is known to block basolateral K⁺ channels (Germann et al. 1986; Dawson et al. 1988), and by the Ca²⁺ channel blocker verapamil (Fig. 8) (Fleckenstein, 1977; Hagiwara and Byerly, 1981). The dose-dependent inhibition kinetics for both blockers seemed to be of the Michaelis-Menten type. The concentrations that induced a half-maximal inhibition were 38.4 and 49.7μmoll⁻¹ for verapamil in two experiments, and 600±185μmoll⁻¹ (N=4) for quinidine. The effects of both quinidine and verapamil were not reversible, although reversibility would be expected for a Michaelis-Menten type of inhibition. Reversibility was only observed in one experiment when the Cabsensitive current was blocked by a single dose of 500μmolI⁻¹ verapamil. Surprisingly, in some experiments where the blockers were added to the serosal solution, I_sc decreased. However, these effects were inconsistent and small. Possibly both blockers penetrate the cell membranes and are accumulated inside the cells, which may explain the observed irreversibility and the serosal effects. Furthermore, the possibility that quinidine, and perhaps also verapamil, induces intracellular effects such as an increase in cytoplasmic Ca²⁺ concentration (Windhager and Taylor, 1983) cannot be excluded. These pharmacological data, therefore, must be interpreted with care.

In a previous study (Krattenmacher et al. 1990) we demonstrated that electrogenic Na⁺ absorption in the colon of the African clawed toad (Xenopus laevis) differs considerably from that in other vertebrates, and even that in other toads or frogs. In Xenopus, Na⁺ passes the apical membrane of colonic epithelial cells not through the classical amiloride-sensitive Na⁺ channel (Benos, 1986), but through a channel that is sensitive to mucosal Ca²⁺. In the present study we investigated the properties of this channel in more detail, especially with regard to its ion selectivity and its physiological role. The results of selectivity experiments of an apical ion channel have to be carefully interpreted from transepithelial measurements, because the transported ion has to pass both the apical and basolateral membranes of the epithelial cell. Furthermore, the electrochemical driving force across the apical membrane varies considerably for different ions. Thus, in accordance with common practice (Hillyard et al. 1982; Aelvoet et al. 1988), we use the term ‘apparent selectivity’ in this study, representing the selectivity of the epithelium for transepithelial ion-transport processes.

Our experiments clearly showed that the short-circuit current (I_sc) increased with the removal of mucosal Ca²⁺ when Na⁺, K⁺, Rb⁺ or Cs⁺ was the principal monovalent cation in the mucosal solution. Only when Li⁺ was the main cation was there no significant Ca²⁺ sensitivity of I_sc. The results indicate that the junctional resistance remains unaffected by mucosal Ca²⁺ removal, provided that, in the presence of a chemical gradient, the paracellular pathway shows no cation selectivity, otherwise changes in transepithelial conductance (G_T) would also be expected in the Li⁺ experiments, where no increase in Ca²⁺-sensitive current occurred. However, although in these experiments a chemical gradient existed for Li⁺ (mucosa to serosa) and Na⁺ (serosa to mucosa), G_T was not increased by the change to Ca²⁺-free mucosal solution. Thus, it seems very likely that, for the various cations, the induced current increase caused by the removal of mucosal Ca²⁺ is not due to changes in the junctional resistance but by an increased cellular conductivity. As previously shown (Na⁺ as main cation), the Ca²⁺-sensitive current increase is fully explained by an increase in electrogenic Na⁺ transport (Krattenmacher et al. 1990). Therefore, it can be assumed that ⁠, in the present experiments in which Na⁺ was replaced by one of the other monovalent cations, is also caused by the principal cation in the solution. With respect to the size of (Table 3), we found an apparent selectivity sequence of K⁺>Na⁺ = Rb⁺>Cs⁺>Li⁺. It is conceivable that Li⁺, which is, in its hydrated form, the largest of the alkali metals (Frey-Wyssling, 1953), is too large to pass one of the cell membranes. However, from these experiments it cannot be concluded that the apical ion channel is the limiting barrier for this ion.

In the toad urinary bladder, Van Driessche et al. (1987) described an apical Ca²⁺-sensitive pathway, which is permeable to Na⁺, K⁺, Rb⁺, Cs⁺, Li⁺ and NH₄⁺, provided that the mucosal Ringer’s solution contained no Ca²⁺. The authors found similar currents flowing through this pathway when Na⁺, K⁺ or Rb⁺ was the principal mucosal cation, whereas Cs⁺ and Li⁺ induced about half as much current. The apparent ion-selectivity of this channel was K⁺>Rb⁺ = Na⁺>Cs⁺>Li⁺ (Aelvoet et al. 1988). We found the same sequence in the Xenopus colon and, therefore, a similarity between the channels seems likely. A further similarity is the strong dependence of the Ca²⁺ sensitivity on mucosal pH (Figs 1,2). For the toad urinary bladder, this increase in with decreasing proton concentration was interpreted as modulation of the channel gating (open-closed) by protons (Aelvoet et al. 1988); at high pH values (low proton concentration) the channel gating may be interrupted, with the channel remaining in the open state. In the Xenopus colon, it is interesting to note that the inhibition kinetics of mucosal Ca²⁺ seems to be unaffected by the proton modulation of the channel gating, because the Michaelis constant of the Ca²⁺-induced inhibition was similar at pH6.5 and 8.0 (Table 2). Futhermore, we have shown that, in the Xenopus colon, this channel is not only blocked by Ca²⁺, but also by other divalent cations such as Ba²⁺ and Mg²⁺ (Figs 5,6), providing one more similarity to the ‘toad bladder’ channel. A further parallel has already been described in a previous study (Krattenmacher et al. 1990): the Na⁺ current flowing through this channel does not saturate with increasing mucosal Na⁺ concentration, but depends linearly on it. All these similarities contribute to the suggestion that the Ca²⁺-sensitive channel type present in the apical membrane of the colonic epithelium of Xenopus is very similar, if not identical, to the Ca²⁺-sensitive channel type existing in the apical membrane of the epithelial cells of the toad urinary bladder. However, two differences should be mentioned: (1) whereas in the toad bladder this channel is sensitive to antidiuretic hormone (oxytocin) or its second messenger cyclic AMP (Van Driessche et al. 1987), these hormones have no effect in the colon of Xenopus (Krattenmacher et al. 1990); (2) in Xenopus, the channel has a considerable conductance even in the presence of mucosal Ca²⁺ (Krattenmacher et al. 1990), whereas in the toad urinary bladder the current flowing through this channel is very low when the channel is blocked by Ca²⁺.

The physiological role of this Ca²⁺-sensitive channel is still unclear. In recent years it has been suggested that this pathway may represent an epithelial Ca²⁺ channel (Van Driessche, 1987; Van Driessche et al. 1987; Aelvoet et al. 1988). This idea is based on the observation that the epithelial Ca²⁺-sensitive channel has several features in common with Ca²⁺ channels found, for example, in frog muscle membranes (Aimers et al. 1984; Aimers and McCleskey, 1984; Hess et al. 1986): conductivity for monovalent cations in the absence of extracellular Ca²⁺; current flow through these Ca²⁺ channels that does not saturate with increasing extracellular concentration of the permeating cation; a conductance blockable by micromolar Ca²⁺ concentrations; and sensitivity to cyclic AMP. Furthermore, Van Driessche (1987) showed that the Ca²⁺-sensitive channel in the urinary bladder of the toad could conduct Ca²⁺ in the presence of nmol I⁻¹ concentrations of Ag⁺. However, the Ca²⁺-sensitive pathway in the toad urinary bladder was recently investigated in more detail by Das and Palmer (1989), who used an appropriate method (patch-clamp) to measure the real ion selectivity at the single-channel level. They found a selectivity with the sequence Rb⁺=K⁺>Na⁺>Li⁺, which differs from the apparent selectivity described by Aelvoet et al. (1988). They demonstrated that the Ca²⁺ sensitivity originates from the outward rectifying characteristic of the channel: the single-channel conductance for monovalent cations was 5–6 times larger for ion movements in the cell-to-mucosa direction than in the mucosa-to-cell direction. This rectification only occurred in the presence of mucosal Ca²⁺. In other words, the presence of extracellular Ca²⁺ favours cation secretion over cation absorption. The authors concluded that this channel is not a Ca²⁺ channel, but the correlate of an outwardly rectifying, amiloride-insensitive apical K⁺ conductance, as previously described by Palmer (1986).

For these reasons, the possibility that the channel in the Xenopus colon is also involved in K⁺-secreting processes cannot be excluded. This hypothesis may be supported by our finding that, in the sequence of apparent ion selectivity, K⁺ predominates (Table 3). Furthermore, quinidine, a blocker of basolateral K⁺ channels (Germann et al. 1986; Dawson et al. 1988), completely blocked the Cabsensitive short-circuit current in our experiments (Fig. 7). Das and Palmer (1989) demonstrated at the single-channel level that quinidine is a potent blocker of the Ca²⁺-sensitive pathway in toad bladder cells. This direct action of quinidine is an important finding because, in frog muscle cells, it is known that quinidine increases sarcoplasmic Ca²⁺ concentrations by reducing the Ca²⁺ uptake into intracellular stores or by inducing Ca²⁺ release from these stores (Windhager and Taylor, 1983). Although an effect of quinidine on intracellular Ca²⁺ concentration cannot be excluded, we assume that in our experiments quinidine acts directly at the Ca²⁺-sensitive channel and not via changes in the intracellular Ca²⁺ concentration. This assumption is supported by the rapid action of quinidine when added to the mucosal Ringer’s solution (Fig. 7). Thus, the quinidine response of the short-circuit current in the colon of Xenopus indicates a pharmacological similarity to either the Ca²⁺-sensitive cation channel in the toad urinary bladder or the basolateral K⁺ channel of the turtle colon. Interestingly, verapamil, which is considered to be a specific Ca²⁺ antagonist, also fully inhibited the Ca²⁺-sensitive current (Fig. 8). However, this does not necessarily indicate that the Cabsensitive channel represents an epithelial Ca²⁺ channel, because the action of verapamil is not confined to Ca²⁺ channels, at least at higher (⩽0.1 mmol I⁻¹) concentrations (Fleckenstein, 1977; Atlas and Adler, 1981). In contrast to La³⁺, which is known to block Na⁺ transport completely (Krattenmacher et al. 1990), both verapamil and quinidine only blocked the Ca²⁺-sensitive current. This may lead to reappraisal of the Ca²⁺-sensitive pathway. The possibility that electrogenic Na⁺ transport occurs through two different pathways cannot be excluded. It is conceivable that two Na⁺-conducting pathways are located in the apical cell membrane: one that is not sensitive to mucosal Ca²⁺ but is blocked by La³⁺ and is always operational, and another that is totally blocked by di- or trivalent cations (Ca²⁺, Ba²⁺, Mg²⁺, La³⁺⁾, verapamil or quinidine.

In conclusion, at present two possibilities must be considered concerning the Ca²⁺-sensitive short-circuit current. If there is only one Na⁺-conducting ion channel in the apical cell membrane it is conceivable that the effect of Ca²⁺ may be related to selectivity changes resulting from Ca²⁺ binding at or near the mouth of the channel. If there are two Na⁺-conducting pathways, the Ca²⁺-insensitive one may represent a new pathway for electrogenic Na⁺ absorption in the vertebrate colon, although the physiological role of the Ca²⁺-sensitive channel remains unclear. However, we exclude the possibility that mucosal Ca²⁺ may regulate the channel under physiological conditions because the channel conductivity is sensitive to such extremely low extracellular Ca²⁺ concentrations. Although we have not measured the Ca²⁺ concentration of the colonic faeces, it can be assumed that, under physiological conditions, it is higher than micromolar levels.

This study was supported by Deutsche Forschungsgemeinschaft (Cl 63/4-2), Europäische Gemeinschaft [ST2J-0392-C (EDB)] and the H. Wilhelm Schaumann Stiftung zur Förderung der Agrarwissenschaften (RV). The results of this study are part of the doctoral thesis of RV. The authors gratefully acknowledge the excellent technical assistance of Mrs Britta Habura and Mrs Marion Ganz.