CHAPTER 8

SINGLE-CHANNEL RECORDINGS FROM THE CRYPT BASOLATERAL MEMBRANE.

Whole-cell patch clamp studies have revealed both inwardly rectifying and outwardly rectifying K⁺ currents in guinea pig villus enterocytes (Fargon et al., 1990) and Necturus enterocytes respectively (Giraldez et al., 1990). With the advent of single-channel recording techniques and the ability to form cell-free membrane patches (Horn & Patlak, 1980; Hamill et al., 1981) it has become possible to characterise the ion channels underlying the `whole-cell' membrane conductances. Single channel recording techniques have revealed a diversity of K⁺-selective ion channels in many epithelia (Dawson, 1987). Large conductance Ca²⁺- and voltage-independent K⁺ channels have been reported in studies with guinea-pig villus enterocytes (Mintenig et al., 1992) and in the basolateral membrane of rat duodenal crypts (Fraser et al., 1991); whilst large conductance Ca²⁺- and voltage-activated K⁺ channels (Maxi-K⁺) have been reported in enterocytes isolated from rat (Morris et al., 1986) and Necturus (Sheppard et al., 1988b) small intestines. However, small conductance K⁺ channels have also been characterised in rabbit (Sepulveda & Mason, 1985) and Necturus enterocytes (Sheppard et al., 1988a). The physiological roles of these K⁺ channels in small intestinal transport is not well understood, and thus more physiological approaches may be required to establish the location, function and mechanisms of regulation of these ion channels.

This aim of the work presented in this chapter was to investigate the nature of the ion channels present in the crypt basolateral membrane to determine the single channel correlates of the macroscopic K⁺ currents observed in the resting and agonist-stimulated small intestinal crypt. Recordings obtained with KCl-rich pipette solutions have revealed the presence of both cation non-selective and anomalously inwardly-rectifying channel activities. Both these channels are activated upon patch excision and are Ca²⁺- and voltage-dependent in their activity. Carbachol addition activates both the cation non-selective channel and a small conductance channel that is infrequently observed in unstimulated cell-attached patches. With the use of Na⁺ gluconate rich pipette solutions 2-4 pS K⁺-selective channel activity was observed upon carbachol addition and patch depolarisation. It is proposed that this small conductance K⁺ channel underlies the carbachol-activated whole-cell K⁺ conductance.

Crypts were prepared as described previously in section 4.2.1..

8.2.1. Solutions.

Carbachol was prepared as a 100mM stock in water and stored frozen. The concentrations of Ca²⁺ and EGTA required to buffer free Ca²⁺ in the bath to the desired concentration were calculated according to the stability constants of Martell & Smith (1974). The composition of solutions for ion substitution experiments is given in table 3.1.. The low Cl^- solution was prepared by an equimolar replacement of NaCl with Na gluconate. No correction for Ca²⁺ chelation by gluconate was made when preparing the low Cl^- Hanks medium. Hanks containing 5mM Ba²⁺ and 5mM TEA were prepared as for normal Hanks, but omitting 5mM NaCl. Changes in [K⁺]_o were achieved by equimolar changes of KCl with NaCl. The composition of all other bath and pipette solutions used are given in table 8.1. All bath and pipette solutions were filtered through Dynaguard 0.2m m syringe filters prior to use.

Electrophysiological recordings of single channel currents were performed as described in section 3.2.3 with the following modifications. Small mouth patch-pipettes used for single channel recording were fabricated from Boralex glass capillaries (Rochester Scientific Co., New York, U.S.A.) and hard typical resistances of 8-10 MW when filled with K⁺ or Cl^- rich pipette solutions. Single channel currents were recorded as outlined by Hamill et al. (1981) using a List EPC-7 amplifier (List Electromedical, Germany) or a Biologic RK-300 amplifier (Biologic, Claix, France). Junction potentials arising from entry into the bathing solution were compensated prior to seal formation. Single channel activity was recorded from the basolateral membrane of intact crypt enterocytes once an adequate seal resistance (>10 GW ) had been obtained, and the fast transient capacitative current had been compensated. To form cell-free patches the electrode was quickly retracted from the cell and the excised (inside-out) membrane patch was passed through the air-water interface to prevent vesicle formation. Cells were voltage clamped at a pre-determined holding potential and single-channel current events were recorded in response to depolarising and hyperpolarising voltage steps. In the cell-attached configuration the trans-membrane potential of the patch (V_patch) is determined by that of the cell interior (V_m) as well as that of the inside of the patch pipette (V_ref) and is given by the relation:

Usually V_m is unknown. The normal sign convention was observed throughout with positive (outward) currents corresponding to upward deflections of the illustrated current records.

Single channel currents were acquired using a voltage pulse generator programme (VGEN, J.Dempster, 1987). Stored single channel records were analysed using a semi-automated analysis programme (PAT, J.Dempster, 1987). Relay records were low pass filtered at 0.5-1.0 kHz (-3db) and digitised at 2-4 kHz as described in 3.2.4. Single channel currents were measured from generated unitary current amplitude distribution histograms. Channel open- and closed-state events were identified using a detection threshold crossing method with the discriminator positioned midway between the fully open and closed levels.

Channel open-state probability (P_o), defined as the fraction of the total time the channel occupies the open state, was computed from the duration of channel open times determined using the detection threshold method. When patches contained more than one channel, the detection threshold was reset for each successive single channel level to determine the time intervals when 1, 2, 3,...,n channels were open simultaneously. Channel open probability for a patch containing more than one channel (nP_o) was calculated as P_o (level 1) + P_o (level 2) ... + P_o (level n). If the patch contained a known number of channel open levels the P_o could be derived by dividing nP_o by n.

Only membrane patches containing a single channel were used for stochastic analyses. Histograms of channel residence times were constructed from current records exceeding 40 s duration. Mean open and closed times represent arithmetic means of all observed open or shut times. Illustrated traces represent records digitised at 2-4 kHz (filtered at 0.5-1.0 kHz (-3db), Bessel 8-pole low-pass filter) using an IBM-AT computer with appropriate software (PAT, J.Dempster, 1987) and then imported as ASCII files into a graphics programme (Sigmaplot 5.0, Jandel Corporation, 1992).

Current-voltage relationships were analysed assuming that they obeyed the Goldman-Hodgkin-Katz constant field theory (GHK), which is a good approximation of the behaviour of many channels not exhibiting anomalous rectification. The data were fitted to equations of the form:

i = GV [C_o-C_iexp(FV/RT)] / {C_sym[1-exp(FV/RT)}]

where i is the single channel current, G is the single-channel conductance, C_o and C_i are, respectively, the ion concentrations bathing the extracellular and intracellular sides of the membrane patch, V is the command potential in the case of excised patches or the sum of the command potential and membrane potential in the case of cell-attached patches, R is the gas constant, T is the absolute temperature, and F is Faraday's constant. Conductances measured in the presence of asymmetrical ion distributions are normalised to the symmetrical case by dividing by C_sym.

8.3.1. Inwardly-rectifying single channel activity of the crypt basolateral membrane.

Cell-attached recordings were obtained from the basolateral membrane of the crypt mid-region. With KCl-rich pipette solutions spontaneously active inwardly-rectifying single channel activity was found in 29% of cell-attached patches (19/66 patches). The channel displayed a maximal slope conductance of 33 ± 0.8 pS at a holding potential of -80 mV and of close to 16pS when the patch was held at the spontaneous membrane potential. Fig.8.1 shows traces of single channel activity recorded from a cell-attached patch with 145 mM KCl in the pipette. The current-voltage (I-V) relationship gave a slope conductance of 32-34pS at a holding potential of -80 mV and an apparent reversal potential (E_REV) of +70mV. The current-voltage relation in both the cell-attached and excised configurations rectified anomalously and could not be described by the GHK constant field equation. The cell attached I-V reversal potential is consistent with channel selectivity for cations over anions, as [Cl^-]_i would have to be in excess of 322 mM to account for such a reversal. Small outward current levels were apparent in the cell-attached configuration only when the patch was depolarised by at least 100mV (Fig.8.1). The inwardly rectifying channel was observed to be spontaneously active in cell-attached patches and only one channel open level was detected in 89% of active patches.

Inwardly rectifying channel activity was seen with a greater frequency in excised inside-out patches (39%, 12/31 patches) than in cell-attached patches. In two excised patches the channel inactivated irreversibly during recording. Following patch excision into Na⁺-rich Hanks medium there was a characteristic reduction in the slope conductance at hyperpolarised potentials, with a conductance of 25-29pS at a holding potential of -60mV (n=12, Fig.8.1.). Under these conditions small inward current levels were observed at a holding potential of 0 mV, indicative of a positive E_REV, although outward currents could not be clearly discerned (Fig.8.3.). When the intracellular face of the patch was bathed with 145 KCl (symmetrical) solution with free [Ca²⁺]_i buffered to 10^-6M, the I-V relation appeared to reverse around 0mV with an inward conductance of 8-9pS and an outward conductance of 5-6pS (n=4, Fig.8.1.). The I-V relation reversed at 0mV with 145mM KCl in the bath, whilst inward current levels were seen at this holding potential with NaCl rich Hanks medium, consistent with a cation-selective channel that is at least partially selective for K⁺ over Na⁺ ions despite its anomalous rectification (Fig.8.1.).

Kinetic analysis was performed on recordings of at least 30s in duration obtained from cell-attached patches with a single channel open level (Fig.8.1.). The channel kinetics exhibit a strong dependence on the holding potential in cell-attached patches, with a high channel P_o `plateau' observed between holding potentials of 0 (P_o=0.88) and -60 mV (P_o=0.67). In contrast only infrequent opening bursts being seen at holding potentials more negative than -60 mV (Fig.8.1C, n=2), the channel exhibiting time-dependent inactivation at holding potentials more negative than -60mV. Once inactivated by holding at hyperpolarised potentials the channel could not subsequently be reactivated at any holding potential between -100 mV and 20 mV in cell-attached patches. Holding at depolarised potentials notably decreased channel P_o, although the apparent decline in channel activity was reversible. In many patches in which inwardly-rectifying single channel activity had been previously inactivated by holding at hyperpolarised potentials, channel activity reappeared following patch excision into bathing media containing high (1mM) [Ca²⁺]_i (Fig.8.3). Excision of the membrane patch into NaCl rich Hanks medium containing 1mM Ca²⁺ led to an apparent loss of voltage-dependence of inwardly rectifying channel activity, with the P_o remaining more or less constant at a range of holding potentials from -100mV to -20mV.

Experiments were then performed to determine the dependency of channel activity upon the cytosolic free [Ca²⁺]_i in excised patches. When the medium bathing the intracellular face of the patch was changed from a solution containing 10^-6M [Ca²⁺]_i to <10^-8M no single channel activity was observed until the Ca²⁺ concentration was restored to 10^-6M (n=2). The Ca²⁺-dependence of single channel activity is shown in Fig.8.2. No spontaneous channel activity was observed in a patch bathed with a free [Ca²⁺]_i of 10^-8M, although channel activity was observed at 10^-3M and 10^-6M. Both the mean channel open probability (P_o) and open time increase as [Ca²⁺]_i is increased from 10^-6M to 10^-3M, although the mean closed time appears to be relatively unchanged. In conclusion the inwardly-rectifying channel is clearly dependent upon intracellular Ca²⁺ for its activity and appears to be sensitive to changes in cytosolic [Ca²⁺]_i.

A second single channel activity was infrequently observed in 6% of cell-attached patches (4/66 patches) that was characterised in the unstimulated crypt by brief channel openings that became more frequent as the patch was hyperpolarised. This channel gave a slope conductance of around 18pS at 0mV and of 26-27pS at a holding potential of -80mV in cell-attached recordings with KCl-rich pipette solutions (Fig.8.4). The cell-attached current-voltage relation gave a reversal potential of around +50 mV (Fig.8.4), consistent with selectivity to cations over anions as [Cl^-]_i would need to be between 150 and 180mM to account for such a reversal.

This cation-selective single channel activity was found in 19% (6/31) of patches when excised into solutions containing high (10^-3M) intracellular free [Ca²⁺]_i, concomitant with a spontaneous increase in channel activity from both quiescent (n=3) and previously active cell-attached patches (n=2, Fig. 8.4). In the cell-attached mode only single open levels were apparent at a range of holding potentials, whilst multiple channel open levels were observed in excised patches at hyperpolarised holding potentials, associated with an increase in single channel P_o.

The ionic composition of the medium bathing the intracellular face of excised patches was varied to determine the selectivity of this channel activity. When the intracellular face of the patch was exposed to 145mM (symmetrical) KCl with a free [Ca²⁺]_i of 10^-3M the channel exhibited an approximately linear I-V relation (Fig.8.4) reversing at 0mV with a slope conductance of 17-18pS (n=4). When bathing the intracellular face of the patch with NaCl-rich Hanks medium or a NaGluconate rich solution the channel still appeared to reverse through 0mV with a linear slope conductance of 18-20pS (n=4, Fig.8.5). With 40mM KCl/105mM NaCl in the bath (Fig.8.5) the channel I-V relation appeared to show a slight outward rectification with outward current levels apparent at a holding potential of 0mV (n=3). This outward rectification was described by the GHK equation with a slope conductance of 18pS and a reversal potential of -10mV indicative of a P_Na/P_K of 1.7:1 (n=3).

Experiments were performed to establish the Ca²⁺-dependence of cation non-selective channel activity in excised inside-out patches (Fig.8.6). When the `cytosolic' free Ca²⁺ was buffered to <10^-8M channel activity was almost completely abolished at both hyperpolarised and depolarised holding potentials (n=2), although activity was readily restored upon the reintroduction of 10^-6M Ca²⁺ to the bathing medium. At 10^-6M free [Ca²⁺]_i channel activity was observed at holding potentials from -100 to 100 mV, with channel P_o increasing sharply as the patch was hyperpolarised. Cation channel P_o was greatest between -60 and -80 mV, with a progressive increase in P_o observed at hyperpolarised potentials (Fig.8.6).

K⁺ channel inhibitors were added to the medium bathing the intracellular face of the patch containing non-selective channel activity. Fig.8.7. shows current amplitude distribution histograms obtained from an excised inside-out patch with 3 channel open levels bathed with Hanks medium containing 1.3mM Ca²⁺. Cation channel P_o values were determined for a range of holding potentials from -100 to 100 mV. With 5mM Ba²⁺ in the bathing medium the channel P_o appeared to be reduced at hyperpolarised potentials (range of 0.3 against 0.5-0.7 for Hanks control), although at depolarised holding potentials the channel P_o was similar to that observed for the Hanks control (0.2-0.3). Upon the addition of 5mM TEA to the bathing medium a small decrease in P_o at hyperpolarised potentials (0.4-0.5) but channel activity was completely abolished at depolarised holding potentials.

8.3.3. Effects of carbachol upon single channel activity in the crypt basolateral membrane.

Carbachol was added to the medium bathing the crypts to determine the identity of the single channels underlying the carbachol-induced hyperpolarisation. In cell-attached patches obtained with KCl-rich pipette solutions the addition of 100m M carbachol evoked large increases in single channel activity in patches which were previously quiescent. This carbachol-evoked increase in single channel activity was most notable at hyperpolarised holding potentials. Fig.8.8. shows traces of single channel activity recorded immediately before and at 20 and 40 second intervals after the addition of 100m M carbachol to the bathing medium. The frequency and duration of bursts of channel activity was clearly increased after carbachol application, with maximum channel activity observed within 30-60s of agonist addition. Also shown in Fig.8.8 is the mean single channel P_o from 3 separate patch recordings determined for a range of holding potentials from -100 mV to 100 mV before and after carbachol addition. Single channel P_o increases relatively little at the spontaneous resting membrane potential or at depolarised holding potentials. However P_o increased dramatically at hyperpolarised holding potentials in the presence of carbachol. The increase in channel activity appeared to decline within 2-3 minutes of agonist application. Addition of 100m M carbachol increased intermediate-conductance single channel activity in 64% (14/22) of patches.

In 5 (22%) of these patches there was no obvious channel activity present prior to agonist addition.

Two distinct single channel conductances were activated by carbachol in cell-attached patches with KCl rich pipette solutions, one with an intermediate slope conductance of 24-27pS at 0mV and 30pS at -80mV and the other with a linear slope conductance of 3-4pS (Fig.8.9). The activation of the intermediate conductance channel by carbachol was associated with a 12mV shift in the reversal potential of the channel I-V relation (+36mV to +48mV). Small conductance single channel activity was present spontaneously in at least 31% of cell-attached patches (20/65), although channel openings were infrequent. Carbachol addition increased small conductance channel activity in 45% of patches to which it was applied (9/20). In one cell-attached patch recording small conductance channel activity showed a gradual 25-fold increase in nP_o was observed following carbachol addition at a holding potential of 0mV (not shown). The small conductance channel I-V relation appeared to reverse at potentials between +50 and +75 mV when the pipette solution contained 145mM KCl. Such a reversal is consistent with cationic selectivity. No clear evidence of small conductance channel activity was noted in excised patches (n=31), although this is difficult to interpret as the channel activity may have been too small to detect. No large conductance (>50pS) channel activity was observed with KCl-rich pipette solutions in cell-attached patches obtained from unstimulated crypts (0/66 patches), following carbachol addition (0/22 patches) or in excised inside-out patches (0/31 patches) at any holding potential.

In 4 cell-attached patches in which no increase in single channel P_o was observed following carbachol addition (Fig.8.3.), inwardly-rectifying single channel activity was already present. Similarly in 4 patches in which carbachol evoked no increase in single channel activity, inwardly-rectifying single channel activity appeared following patch excision into a bath containing 1mM cytosolic free [Ca²⁺]_i. These observations are inconsistent with the inward-rectifier being activated by carbachol or its intracellular mediators.

8.3.4. Single channel recordings obtained with Na⁺ Gluconate rich pipette solutions.

Pipette solutions that contain low concentrations of KCl and high concentrations of Na Gluconate were employed to identify the single channel K⁺-selective conductance underlying the carbachol-induced hyperpolarisation. A recording obtained from a cell-attached patch in which no single channel activity had been observed prior to the addition of 100m M carbachol is shown in Fig.8.10 (upper trace). Single channel activity was induced within seconds of carbachol addition at the spontaneous membrane potential. The single channel current amplitude was observed to increase as the membrane patch was depolarised. The E_REV could be extrapolated to indicate a reversal potential around 50mV more hyperpolarised than the spontaneous membrane potential. The application of 100m M carbachol resulted in the appearance of small conductance channel activity in 6/18 (33%) of cell-attached patches. The carbachol-induced channel activity appeared within 3-5s of agonist application in these patches, a time-course consistent with the current-clamp hyperpolarisations of E_m. No large conductance K⁺-selective single channel activity was observed in response to carbachol (0/18). In 3 patches in which carbachol clearly activated small outward currents in the cell-attached configuration, no K⁺-selective channel activity was observed when the patch was subsequently excised. In 18 excised inside-out patches obtained with Na Gluconate rich pipette solutions no clear evidence of open levels attributable to K⁺-selective single channel activity were observed, even with 1m M [Ca²⁺]_i and 145mM K⁺ in the bathing medium.

Outward current levels were also activated in 3 cell-attached patches following membrane patch depolarisation, without the prior application of carbachol. Figure 8.11 shows recordings of outward current levels that were activated by briefly depolarising the patch by 20mV. Recordings of single channel activity at potentials 20 and 40 mV more depolarised than the spontaneous membrane potential are shown together with unitary current amplitude distribution histograms (Fig.8.11). The current-voltage relation shown in Fig.8.11. was fitted to the GHK equation (section 8.2.3.) assuming a membrane potential of -45 mV and intracellular K⁺ and Na⁺ activities of 125 and 20 mM respectively. The fit to the GHK equation indicated a reversal potential of -70mV with a slope conductance of 3.7 pS and a P_Na/P_K of less than 0.05. In this recording, as in several others, K⁺-selective channel activity was not observed at hyperpolarised potentials. From three separate recordings of this depolarisation-activated channel current-voltage relations could be derived with mean slope conductances of 2.0, 2.8 and 3.7pS (2.8 ± 0.5pS). Corresponding reversal potentials between -70 and -75 mV more negative than the spontaneous membrane potential were extrapolated from these three recordings, consistent with K⁺-selectivity. These low conductance channels appeared to be similar to those activated by carbachol in their low conductance and reversal potential.

8.4.1. Inwardly-rectifying single channel activity.

Single-channel recording has revealed the diversity of K⁺- and cation non-selective channels present in many secretory and absorptive epithelia (Latorre et al., 1989; Partridge & Swandulla, 1988). The ubiquitous Ca²⁺-activated cation non-selective channel (CAN conductance) and Ca²⁺-activated K⁺ channels have both been proposed to play a role in the mechanism of sustained fluid and electrolyte secretion in exocrine epithelia (Maruyama & Petersen, 1984; Petersen & Maruyama, 1984). The multitude of K⁺ channels present in epithelia are important for the recycling of cations across the basolateral membrane, in the determination of the resting membrane potential and the maintenance of an electrochemical gradient favourable to sustained Cl^- secretion (Dawson, 1987). Current-clamp measurements from the crypt have indicated that basolateral K⁺ conductance(s) underlie both the resting membrane potential and the carbachol-induced hyperpolarisation.

A strongly inwardly-rectifying single channel activity with a unitary conductance of 32-34 pS is the most frequently observed single channel activity in the unstimulated crypt basolateral membrane, and is found to be spontaneously active in 29% of cell-attached patches. Although reversal potentials obtained with KCl-rich pipette solutions are consistent with the inward-rectifier being K⁺-selective, the channel rectifies anomalously, the channel conductance increasing with hyperpolarisation and decreasing with depolarisation, a phenomenon first described in skeletal muscle by Katz more than 40 years ago (Katz, 1949). Inwardly-rectifying K⁺ channels with a similar conductance (27-31 pS) have been described in other cell types, including guinea pig ventricular myocytes (Sakmann & Trube, 1984) and rabbit osteoclasts (Kelly et al., 1992).

The inward-rectifier has a high P_o at the spontaneous resting membrane potential and at mildly hyperpolarised potentials (-20 to -60 mV), suggesting that it might contribute to the cationic permeability of the unstimulated crypt basolateral membrane. Ion substitution experiments suggest that the inward-rectifier gates outward K⁺ current when bathed in symmetrical KCl more readily than it does when Na⁺ is the predominant intracellular cation. This observation, along with the greater conductance that is observed when [Na⁺]_i is increased and [K⁺]_i is decreased, suggests that the relative intracellular cation activities may serve to regulate the conductance of the inward-rectifier (IR) under physiological conditions.

Although the inwardly rectifying channel activity is clearly Ca²⁺-dependent the [Ca²⁺]_i values at which increases in channel P_o were induced were unphysiologically high and no evidence exists to suggest that increases in intracellular [Ca²⁺]_i will regulate channel activity in vivo. In support of this contention no increase in inwardly-rectifying channel activity was discerned in the cell-attached configuration following the addition of carbachol, indicating that the IR probably does not underlie the macroscopic carbachol-activated basolateral K⁺ conductance. The possibility that inwardly rectifying single channel activity is sensitive to changes in free [Ca²⁺]_i that may occur in the unstimulated enterocyte remains to be explored.

The ubiquitous inwardly-rectifying K⁺ channel has been proposed to regulate the resting membrane potential in many cell types (e.g., Kelly et al., 1992). This function is readily attributable to a constitutively active, voltage-dependent and inwardly rectifying K⁺-selective channel. However, the high frequency of inwardly-rectifying channel activity observed in the crypt basolateral membrane does not concord with the outwardly-rectifying nature of the crypt whole-cell conductance. In bovine pigmented ciliary epithelial cells the inwardly-rectifying K⁺ conductance has been proposed to mediate outward current flow at the spontaneous membrane potential and to repolarise the resting membrane potential following cation loading, the first step in the secretion of aqueous humour (Stelling & Jacob, 1992). However, in these cells solution flow over the cell has been shown to inactivate the IR in whole-cell recordings, possibly due to a reduction in the local [K⁺]_o. Thus the concentration of extracellular K⁺ may regulate inwardly-rectifying channel activity in vivo. This observation may explain the high frequency with which such channel activity is found in the crypt basolateral membrane using KCl rich pipette solutions. Another interesting proposal is that an enhanced basolateral K⁺ conductance may result in a local accumulation of K⁺ ions in the vicinity of the crypt basolateral membrane, and thus the local E_K may be more positive than the whole-cell membrane potential, allowing the re-uptake of K⁺ ions via the IR (Kotera et al., 1991). Furthermore the conductance of the guinea-pig IR seems to be dependent upon the cation composition of the bathing medium at the intracellular face of the patch.

It might be expected that a K⁺-selective channel that determines the resting membrane potential would be activated by membrane depolarisation, although the converse appears to be the case for the guinea-pig IR. Further evidence against the involvement of the IR in the maintenance of the resting membrane potential comes from the absence of any significant effect of 5mM TEA upon crypt membrane potential, since TEA has been shown to be a potent blocker of the IR at this concentration in bovine pigmented ciliary epithelial cells (Stelling & Jacob, 1992). However the guinea-pig IR does appear to gate outward K⁺ currents when bathed in symmetrical KCl and in the cell-attached configuration. It may be possible that [K⁺]_o and the relative concentrations of [Na⁺]_i and [K⁺]_i are fundamental to the regulation of inwardly-rectifying single channel activity, as has been suggested by others (Stelling & Jacob, 1992; Harvey & Ten Eick, 1988). Thus the reuptake of K⁺ ions across the basolateral membrane would appear to be the most likely role of the guinea-pig IR, although whether the IR allows a bidirectional movement of K⁺ ions across the basolateral membrane under physiological conditions remains to be established.

8.4.2. Cationic non-selective single channel activity.

A second single channel activity with a 26-27 pS slope conductance is infrequently observed in cell-attached patch recordings with KCl-rich pipette solutions. This channel activity is characterised by occasional bursts of activity that become slightly more frequent at hyperpolarised potentials. The reversal potentials obtained with KCl-rich pipette solutions in cell-attached and excised inside-out patches with low Cl^- bathing solutions suggest that this channel activity has a low permeability to Cl^- ions. Substitutions of cytosolic K⁺ for Na⁺ in excised patches indicate that this channel activity discriminates poorly between cations with a P_Na/P_K of approximately 1.7:1. However this cation non-selective channel activity also exhibits inward rectification in cell-attached, but not excised, membrane patches. This rectification, although not as pronounced as that observed for the IR, may favour Na⁺ entry into the cytosol following channel activation.

The cation non-selective channel appears to be activated following patch excision into bathing media containing 10^-3M free [Ca²⁺]_i, as the frequency with which cation-selective channels are found in excised patches being three times that of cell-attached patches. This cation non-selective channel may occur in clusters, as single open levels were not observed in excised patches. However, only single open levels were usually apparent in cell-attached patches inferring a change in channel behaviour following patch excision. Cationic non-selective channel activity is dependent upon the presence of cytosolic Ca²⁺ in excised patches and exhibits a characteristic increase in channel P_o as the membrane patch is hyperpolarised when cytosolic Ca²⁺ is buffered at 1m M. This Ca²⁺-dependence and apparent inability to distinguish between cations is consistent with the properties of other calcium-activated non-selective cation (CAN) channels, first described by Kass and colleagues in 1978.

Many CAN channels have been characterised which have a slope conductance in the range of 15-40 pS, that discriminate poorly between Na⁺ and K⁺ ions and whose activity is dependent upon the cytosolic free [Ca²⁺]_i. CAN channels have been described in many secretory tissues including mouse and guinea-pig pancreatic acinar cells (Maruyama & Petersen, 1982, 1984; Suzuki & Petersen, 1988), rat lacrimal gland cells (Marty et al., 1984), cultured rat insulinoma cells (Sturgess et al., 1987), rat pancreatic duct cells (Hamill et al., 1981) and cultured neonatal mouse mandibular glands (Cook et al., 1990). Most CAN channels so far described appear to be voltage-independent, although of the few that have been shown to exhibit a voltage-dependent P_o, only aortic endothelial CAN channels share the hyperpolarisation-induced activation with the non-selective conductance of the small intestinal crypt basolateral membrane (Fichtner et al., 1987). Although the crypt CAN channel exhibits Ca²⁺-dependent behaviour in excised patches, the Ca²⁺-sensitivity of the mouse pancreatic acinar cell CAN channel has been shown to be reduced following patch excision (Maruyama & Petersen, 1984), and thus experiments in the cell-attached mode are necessary to demonstrate that this channel is regulated by physiological changes in intracellular Ca²⁺.

There are several lines of evidence to suggest that the crypt CAN channel is activated by changes in intracellular Ca²⁺ in cell-attached patches. Firstly, both the carbachol-activated intermediate conductance channel and the CAN channel activity present in excised patches show a clear increase in P_o at hyperpolarised holding potentials (Figs. 8.6, 8.7 and 8.8). Secondly, the cell-attached reversal potential of the carbachol-activated intermediate conductance channel is consistent with selectivity for cations over anions, as is the positive shift in reversal potential observed following carbachol addition, given the hyperpolarisations evoked by carbachol. Thirdly, the 24-27 pS slope conductance measured for the intermediate conductance carbachol-activated conductance in cell-attached patches is in agreement with the 26-27 pS slope conductance that could be derived from the channel activity present in cell-attached patches prior to excision.

The addition of either 5mM Ba²⁺ or TEA⁺ to the cytosolic face of the crypt CAN channel resulted in a voltage-dependent inhibition of channel P_o. Intracellular TEA⁺ completely blocked outward CAN channel currents at depolarised potentials, consistent with the presence of a binding site for TEA⁺ within the channel pore. The blockade of channel pores by TEA at depolarised potentials was first demonstrated by Armstrong with the delayed rectifier K⁺ channel, whilst at hyperpolarised potentials inward K⁺ flow was observed to relieve channel blockade, presumably by the physical expulsion of TEA from the internal binding site in the channel pore (Armstrong, 1971). The CAN channel is constitutively active in the presence of cytosolic Ca²⁺, which may be expected to allow TEA⁺ direct access to the channel pore. Both 5mM Ba²⁺ and TEA⁺ reduce the P_o of the CAN channel at hyperpolarised potentials, possibly by modulating the hyperpolarisation-dependent gating mechanism or by competing with Ca²⁺ for its binding site on the CAN channel protein.

The crypt CAN channel may serve to limit the extent of the carbachol-induced hyperpolarisation of crypt membrane potential, a function consistent with its activation by membrane hyperpolarisation and elevations in cytosolic Ca²⁺. Such a mechanism would be consistent with its permeability to Na⁺ ions, activation at hyperpolarised potentials in the presence of carbachol and slight inward rectification in cell-attached patches. The role of the CAN channel in the maintenance of the resting membrane potential is not clear, as hyperpolarisations in the absence of carbachol do not greatly increase channel P_o. It is also possible that the crypt CAN channel may possess a significant permeability to Ca²⁺ ions and therefore mediate the entry of extracellular Ca²⁺ following carbachol stimulation, although to date no sizeable Ca²⁺ permeability has been reported for a CAN channel in exocrine cells (e.g. Petersen & Maruyama, 1982).

Agonists that mobilise intracellular Ca²⁺ have been demonstrated to activate CAN channels in lacrimal gland cells (Marty et al., 1984) and pancreatic acinar cells (Maruyama & Petersen, 1982) which appear to be important for sustained fluid and electrolyte secretion in these epithelia. A cationic conductance present in bovine pigmented ciliary epithelial cells (Jacob, 1989) reportedly mediates the entry of cations into the cell without strongly depolarising the cell membrane potential, thus loading the cell with cations as a first step in the secretion of aqueous humour (Stelling & Jacob, 1992). It is difficult however to reconcile how the activation of a Na⁺ permeability pathway might enhance fluid and electrolyte secretion, as any significant entry of Na⁺ into the exocrine cell might be expected to impair the efficiency of Na⁺-coupled Cl^- cotransport and to reduce the electrochemical gradient for Cl^- exit (Petersen & Maruyama, 1984; Young et al., 1987). It has been proposed by Marty that exocrine secretion is driven by the movement of K⁺ across the apical membrane, and thus Na⁺ entry through cationic channels would increase Na⁺/K⁺ ATPase activity and maintain an intracellular K⁺ concentration favourable to continued electrolyte secretion (Marty, 1987). The activation of cation channels may also serve to maintain the electrochemical gradient for K⁺ efflux by holding the membrane potential below E_K during carbachol stimulation.

8.4.3. Identification of the carbachol-activated K⁺ conductance.

In addition to the intermediate conductance cationic channel, carbachol also activates a small 3-4 pS channel activity in cell-attached patches with a reversal potential that is also consistent with cationic selectivity, although it has not been determined whether this 3-4pS channel activity is selective for K⁺ ions. This single channel conductance could not be detected in excised patches, irrespective of the cytosolic K⁺ or free Ca²⁺ concentration. The composition of the pipette solution was therefore modified to allow the identification of K⁺ channel activity in the cell-attached configuration without the requirement for patch excision by employing pipette solutions containing 145mM NaGluconate and low concentrations of K⁺ and Cl^- ions. Assuming that under physiological conditions intracellular ionic activities range between 10 and 40mM for Na⁺, 70 and 140mM for K⁺ and 30 and 60mM for Cl^-, with an observed range of E_m values from -10 to -70mV we can calculate the probable range of driving forces on each ion across the patch of membrane in the cell-attached configuration. Thus, assuming a negligible permeability to gluconate^-, we can predict a driving force of between 60 and 100mV for Cl^- exit (inward current by convention); of between 40 and 140mV for Na⁺ influx, and for K⁺ ions a driving gradient of between 50 and 120mV for K⁺ efflux. Therefore the appearance of outward current levels within a `physiological' range of membrane potentials and intracellular ionic activities can be conclusively interpreted as the efflux of K⁺ ions, especially if the single channel reversal potential approaches E_K. This allows the identification of low conductance K⁺-selective channels in the cell-attached configuration without the necessity for patch excision to determine channel selectivity, a manoeuvre that has been demonstrated to alter the behaviour of many ionic channels (Kunzelmann et al., 1989, 1991).

The addition of carbachol to the bathing medium rapidly induced the appearance of small outward currents with a slope conductance of 2-3pS at the spontaneous membrane potential in 33% of cell-attached patches (Fig.8.10). This small conductance K⁺ channel activity appeared to inactivate when the patch was depolarised by 60 mV or more. The cell-attached reversal potential was around 70 mV more hyperpolarised than the spontaneous membrane potential (around -60 mV in the presence of carbachol), close to the predicted value of E_K of between -110 and -125 mV, suggesting that this channel is selective for K⁺ ions. This small conductance carbachol-activated K⁺ (SK) channel activity in the crypt basolateral membrane indicates that this channel amy belong to the class of small conductance Ca²⁺-activated K⁺ channels (SK_Ca) first characterised at the single channel level in cultured rat skeletal muscle (Blatz & Magleby, 1986) and in excised patches from guinea pig hepatocytes (Capiod & Ogden, 1987), which demonstrated half-maximal activation at 200-500 and 600 nM cytosolic Ca²⁺ respectively.

Depolarisation of the membrane patch in cell-attached recordings by 20 to 40 mV often induced the appearance of small outward currents with a slope conductance of 2-4 pS. Channel activity was not observed at potentials 60 to 80 mV more depolarised than the spontaneous membrane potential or at hyperpolarised holding potentials. As the cAMP mobilising agents VIP and forskolin depolarise the crypt, it is tempting to speculate that the activation of the SK by membrane depolarisation may limit the limit the magnitude of the agonist-induced depolarisation and thus maintain the driving gradient for Cl^- exit. The extrapolated cell-attached reversal potential for these depolarisation-activated channels was 60 to 70 mV more hyperpolarised than the spontaneous membrane potential, again indicative of a single channel conductance highly selective for K⁺ ions. The apparent selectivity for K⁺ over Na⁺ ions, channel inactivation at more depolarised potentials and the similarity in single channel conductance suggest that this channel is identical to the one activated by carbachol. Small and intermediate conductance K⁺ channels that are activated by membrane depolarisation and elevations in intracellular free Ca²⁺ have been previously described in adrenal chromaffin cells (Marty & Neher, 1985), MCF-7 human breast carcinoma cells (Wegmen et al., 1991) and Necturus enterocytes following exposure to L-alanine (Sheppard et al., 1988a). In contrast, SK_Ca channels that are weakly voltage-dependent have also been reported in thymic lymphocytes (Mahaut-Smith & Mason, 1991) and in isolated rabbit enterocytes (Sepulveda & Mason, 1985).

The apparent absence of Ca²⁺-dependent small conductance K⁺ channel activity in excised patches precludes studies of the sensitivity of this channel to cytosolic Ca²⁺ or its intracellular mediators. However it might be predicted that the SK channel is sensitive to increases in free [Ca²⁺]_i following stimulation by carbachol, as can be inferred from the dependence of the carbachol-induced hyperpolarisation upon extracellular Ca²⁺. Current-clamp recordings also suggest that the channel underlying the carbachol-induced hyperpolarisation is relatively quinine insensitive, as 200m M quinine did not appear to reduce the magnitude of the carbachol-induced hyperpolarisation. The 23 pS Ca²⁺-activated K⁺ channel of MCF-7 cells has been demonstrated to be insensitive to extracellular quinidine, whereas quinine has been shown to potently displace apamin binding from SK_Ca channels in guinea pig hepatocytes (Cook & Haylett, 1985), suggesting that the quinine sensitivity of SK_Ca channels may vary substantially between cell types.

Large conductance Ca²⁺- and depolarisation-activated `maxi-K⁺' channels were not seen been observed in patches obtained from the crypt basolateral membrane, inconsistent with a role of the maxi-K⁺ channel in mediating the carbachol-induced hyperpolarisation of small intestinal crypt membrane potential. Although inwardly-rectifying K⁺ channel activity is activated by elevations of intracellular free Ca²⁺ in Hela carcinoma cells (Sauve et al., 1986) and in the avian salt gland (Richards et al., 1989), the inward rectifier of the small intestinal crypt basolateral membrane, although Ca²⁺-dependent in excised patches, does not appear to be activated by carbachol in cell-attached patches. Thus it would appear that the SK channel is the only detectable K⁺-selective channel activated by carbachol present in the crypt basolateral membrane, although the possibility that maxi-K⁺ channels are present in the apical or paracellular membrane domains cannot be precluded. The identification of the single channels mediating the entry of extracellular Ca²⁺ and underlying the outwardly-rectifying K⁺ conductance that dominates the resting membrane potential of small intestinal crypts may require alternative strategies.

Return to top
Return to table of contents
Return to science homepage
References
Next chapter

Table 8.1. Composition of solutions.

All concentrations are given in mM and solutions titrated to pH 7.2 with Tris. 5mM Ba²⁺ Hanks and 5mM TEA Hanks were prepared by equimolar substitution for NaCl.