INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE TRANSMEMBRANE MOVEMENT of ions in the cells of the colonic epithelium is involved in two distinct processes: the transepithelial movement of fluid and the regulation of cell volume. The first process is responsible for the overall fluidity of the colonic luminal content, which is determined by the balance between fluid secretion and absorption occurring in the crypt and surface epithelial cells. The second process is based on the uptake or efflux of osmotically active compounds and is responsible for the rapid compensation of cell volume changes resulting from fluctuating entry or exit of ions and osmotically obliged water and from variations of the osmotic pressure in the luminal compartment of the colon.

Current understanding of the mechanisms underlying transepithelial fluid movements through the colon is based on observations of the opposing processes of Cl secretion within the crypts and Na⁺ and Cl uptake by the surface epithelial cells. Several types of Cl channels have been characterized on the apical membrane of isolated rat enterocytes (9, 14) and human colon carcinoma cells (20) and on the surface cells of intact human mucosa (9, 39). Most of these channels belong to the family of outwardly rectifying Cl channels (ORCC) and, along with the so-called cystic fibrosis transmembrane conductance regulator Cl channel, are thought to share a major role in Cl secretion. In the basolateral membrane, a Cl channel with a linear current-voltage (I-V) relationship and a slope conductance of 29 pS has been reported in isolated crypts from rat distal colon (9), and ORCC have been observed in isolated crypts of mouse jejunum (4) and guinea pig enterocytes (36), but the role of these basolateral Cl channels has not been clearly identified.

Similarly, various studies on intestinal epithelial cells, e.g., isolated enterocytes (32) or intact crypt cells (8, 35, 37), have implicated Cl conductances in regulatory volume decrease (RVD) following hypotonic shock. In this experimental situation, the data of Worrell et al. (52) and Kubo and Okada (24) implicated a role of ORCCs in the volume regulatory process.

On the basis of this, it appears that members of the large family of ORCCs may be involved in both volume regulation and in ion secretion or absorption. A Cl channel located on the basolateral membrane of crypt cells could provide a pathway for the exit of Cl ions during RVD and/or be involved in the basolateral Cl exit step during NaCl absorption. However, there is a marked lack of information concerning conductive Cl movements through the basolateral membrane of colonic cells.

The present study used the patch-clamp technique for recording single ion channel activity, with the aim of identifying and characterizing the Cl channels present under steady-state conditions in the basolateral membrane of enterocytes in intact mouse distal colon crypts. We report the presence of a Cl channel characterized by a strong outward rectification, sensitive to DIDS, 5-nitro-2-(3-phenylpropylamino)benzoate (NPPB), 9-anthracenecarboxylic acid (9-AC), and GTP. Preliminary data have been presented in abstract form (34).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolation of colonic crypts. Isolated crypts were prepared by a modification of the procedure described by Siemer and Gögelein (47). Female mice weighing 20-30 g were anesthetized (ketamine, 10-20 mg/ml) and killed by cervical dislocation. The distal colon was taken above the pelvic brim, dissected, and rinsed in ice-cold NaCl solution containing (in mM) 140 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 10 HEPES-Tris buffer, 10 glucose, and 1 dithiothreitol (DTT), pH 7.4. The intact colon was everted and filled with a Ca²⁺-free solution containing (in mM) 96 NaCl, 1.5 KCl, 10 HEPES-Tris, 27 Na-EDTA, 55 sorbitol, 44 sucrose, and 1 DTT, pH 7.4, and incubated for 20 min, then transferred into a high-Ca²⁺ NaCl solution containing (in mM) 140 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 10 HEPES-Tris, and 10 glucose, pH 7.4. The crypts were released into solution by gentle shaking, washed twice by centrifugation at 600 g for 2 min, and stored on ice until use. Before patch-clamp experiments, the crypts were fixed to the glass bottom of an experimental chamber (volume 0.4 ml) coated with poly-L-lysine (0.01% wt/vol). Patch-clamp experiments were performed at room temperature (20-22°C).

Experimental solutions and drugs. The composition of solutions used in patch pipettes and bathing solutions is given in Table 1. The Ca²⁺ concentration used in the bathing and the pipette solutions was adjusted using EGTA to pCa 3 in cell-attached configuration, and it was adjusted to pCa 8 in the bathing solutions in the excised inside-out configuration. All solutions were equilibrated in air, filtered through 0.2-µm Millipore cellulose disks, and had a final osmolality of 310 mosmol/kgH₂O. Osmolality was determined by vapor pressure osmometry (Wescor). A set of reservoirs connected to perfusion pipettes was used to test the effects of different solutions on channel activity in excised patches. Solution changes were performed within a few seconds by manual switching between reservoirs. The permeability of the ORCC to anions other than Cl was assessed using appropriate potassium salts (KX), where X = SCN, I, Br, F, NO₃, SO₄², or gluconate (see Table 1). NPPB was obtained from Research Biochemicals International. All other chemicals were from Sigma.

Current recordings. Characterization of the single channel current was performed in both the cell-attached and excised inside-out patch configurations. Single channel currents were recorded by the method of Hamill et al. (17) using a RK400 patch clamp amplifier (Biologic, Claix, France) filtered at 0.3 or 1 kHz, digitized (48 kHz), and stored on a digital audio tape (DTR 1204, Biologic). For analysis, the data were played back, transferred to a computer, and analyzed by the PAT computer program (Dempster, Strathclyde Electrophysiology Software). Patch pipettes (tip resistance ranging between 10 and 15 M) were prepared from borosilicate glass capillaries (GC 150F, Clark), pulled, and polished on a programmable puller (DMZ, Werner Zeitz Augsburg, Germany). Three to five gigaohm seals were obtained by calibrated suction using a syringe connected to the patch pipette. Under these conditions, the success rate of obtaining gigaohm seals was 59%. The sign of the clamped voltage (V_p) refers to the pipette solution with respect to the bath, and outward currents (positive charges flowing across the patch membrane into the pipette) are shown as an upward deflection in the current traces. In the excised configuration, the imposed membrane potential (V_m) is referred to as V_p. I-V curves were constructed by plotting the mean current amplitude for each clamped potential.

Liquid junction potentials. The liquid junction potential (LJP) was defined as the potential of the bath solution with respect to the pipette solution (2), and the V_m was calculated as V_m = V_p + LJP, where V_p is the reading provided by the patch-clamp amplifier. When bath solutions of different composition were successively applied to the patch membrane, the corresponding changes in LJPs were corrected using the Henderson equation (JPCalc computer program)
where
and u, C, and z represent the mobility, concentration, and valency of each ion species (i), and R, T, and F are the gas constant, temperature, and Faraday constant, respectively. Subscripts b and p denote bath and pipette solutions, respectively.

Data analysis. Open probability (P_o) was determined as the fraction of digitized points above a threshold set midway between the closed and open peaks of current-amplitude histograms. In these conditions, P_o was defined as the ratio of the total time spent in the open state to the total time of the complete record. P_o was determined from stable recordings immediately after identification of the channels. Except for NP_o calculation, analysis was confined to patches containing a single channel. When more than one channel open state was observed, N was determined as the maximum number of channels observed. The NP_o value was determined from the amplitude histograms for each record. Patches often contained more than one channel (the average number of ORCCs observed simultaneously was 1.96 ± 0.08, n = 120). To determine this, the V_m was clamped to 70 mV for 10-20 min, and when channel activity was detected this potential was maintained for 5-10 more minutes to check that all channels present under the patch were activated. When multiple channels were present, they always activated before this time. The I-V curve was first obtained to confirm that the channel was the ORCC. Where only a single channel was found, this was confirmed by holding the V_m at 70 mV for several minutes before performing any kinetic analysis at either +50 or 50 mV with data collected first at +50 mV. Moreover, at the end of each experiment, and especially when only one channel was observed, the V_m was always clamped to a high positive voltage to check that no additional channels were present. For all of the data relevant to single channel open state, only the data in which this protocol was fully accomplished were considered. The probability of simultaneous opening of 1-5 channels was also calculated as described by Colquhoun and Hawkes (6). Calculations were made on continuous data, filtered at 1 KHz, that began with a closed-open transition and ended with an open-closed transition. From these records, we calculated the mean open time (MOT), mean closed time (MCT), and the number of open states per second (NOS). Conventional 50% threshold analyses yielded distribution of dwell times that were fit by multiexponential or power functions consistent with multiple open and closed states. Gaps were defined as closed intervals of relatively long duration. To get the best fit, the gaps were defined as closed intervals >100 ms. Bursts consisted of all channel activity between gaps. Within the bursts we calculated the burst MOT, the burst MCT, and the NOS. The I-V plots obtained were fit with a second-order polynomial equation, and values for reversal potentials (E_r) were obtained from the fitted curves.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The dissociation protocol used in the present experiments eliminated connective tissue, and cells with basolateral membranes free of basal lamina were routinely obtained, resulting in a 59% success rate for obtaining gigaohm seals. Observation of the isolated preparations indicated that the crypts remained viable for at least 4 h after dissociation, after which some cell rounding occurred. Patch-clamp experiments were carried out before changes in cell shape became visible.

Channel activation in intact cells. Spontaneous Cl channel activity was only rarely found in the cell-attached mode. Of 735 successful seals, only 7 cell-attached patches showed spontaneous single channel activity consistent with Cl -selective channels. Figure 1A gives an example of these recordings obtained with 145 mM KCl in the pipette and with isotonic Ringer solution in the bath at a range of applied potentials in cell-attached patches. The single channel I-V relationship is presented in Fig. 1B. Under these conditions, the single channel current exhibited outward rectification and reversed polarity at the resting V_m (40 mV as measured with microelectrodes; Dr. J. P. Pennec, personal communication). This is very close to the Nernst potential for Cl (E_cl = 36 mV) calculated in these cells with 35 mM intracellular Cl concentration (51).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   The outward rectifier Cl channel (ORCC) in cell-attached configuration. A: representative single channel current tracings of Cl channels in cell-attached patches at the indicated holding potentials. The bath contained isotonic Ringer solution, and the pipettes were filled with 145 mM KCl. The closed state is shown by the dashed lines. Upward deflection at positive clamp potentials indicates the flow of anions from pipette to cell interior. Vp, clamped voltage. B: current-voltage (I-V) relationship, under similar conditions to those described in A, from the mean of 7 experiments in which the Cl channel was spontaneously active in the membrane patch. Inset: open probability of multiple channels (NPo)-voltage relationship indicating the dependency of channel activity on membrane potential (Vm).

Kinetic analysis of the channel in intact cells. Figure 1B shows that NP_o was a direct function of the imposed potential (V_p) with the highest values for positive potentials. In addition, as illustrated in Fig. 1A, the number of channels simultaneously active (N) was an increasing function of V_m.

Cl channels in excised inside-out patches. Spontaneous channel activity always disappeared immediately after excision of active cell patches. After excision in the inside-out configuration from quiescent cell-attached patches, an activity consistent with Cl channels was observed in 120 out of 735 seals. However, such activity usually occurred only after 10-20 min and application of depolarizing voltage steps of +70 mV (V_p). Table 2 summarizes all mean values ± SE of the E_r and slope conductances at E_r (gE_r), between 100 and 40 mV (g), between +40 and +100 mV (g⁺), calculated from single channel currents in the excised inside-out configuration, with different bathing and pipette-filling solutions. Figure 2A shows an example of the current records obtained with K_int solution (pCa 8) in the bath and 145 mM KCl (pCa 3) in the pipette at a range of imposed potentials. As shown in Fig. 2B, the I-V relationship showed the strong outward rectification characteristic of ORCC. The channel slope conductance was 23.4 ± 1.2 pS (n = 25) between 100 and 40 mV and increased to 90.2 ± 1.7 pS (n = 25) between +40 and +100 mV. Under these conditions, the channel activity reversed at a V_m of 2.5 ± 0.2 mV (n = 25). Table 2 shows that the ORCC was distributed along the crypts from the base to the apex. For all of the following experiments, recordings were performed at the crypt base. Seals were generally easier to achieve and crypt structure integrity was better conserved in this region.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   ORCC in inside-out configuration. A: representative tracings from single channel currents of ORCCs in excised inside-out patches at the indicated holding potentials. The bath contained Kint (pCa 8) solution, and the pipettes were filled with 145 mM KCl (pCa 3). The closed state is shown by the dashed lines. Upward deflections at positive clamp potentials indicate the flow of anions from pipette to cell interior. B: I-V relationship, under similar conditions to those described in A, from the mean of 25 experiments. Inset: open probability (Po) as a function of Vm (n = 6).

Ion selectivity of the ORCC. Table 2 and Fig. 3 show the effects of substituting NaCl or K-gluconate for KCl in the bath or in the pipette. Replacement of KCl by NaCl had no significant (P < 0.05) effect on the I-V relationship (Table 2), whereas in the presence of K-gluconate in the pipette, the reversal potential shifted to 34.3 ± 1.9 mV (Fig. 3, n = 6). The anionic vs. cationic selectivity was determined by changing from K_int to half-strength KCl solution (K_int1/2), with only half of the concentration of KCl on the cytosolic side of the patch. This maneuver significantly (P < 0.01) shifted the I-V curve to the left and the reversal potential was 14.0 ± 1.5 (n = 9) mV, consistent again with E_Cl (17.9 mV). The relative permeability (P_Cl/P_cations) derived from the Goldman-Huxley-Katz relation was 14.6 ± 4.7 (n = 9). The measured conductances of the channel (g⁺, g, and gE_r) are also significantly reduced (P < 0.01) in the presence of K_int1/2.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Anion selectivity of the ORCC in inside-out configuration. I-V relationship, corresponding to the data contained in Table 2, of the ORCC in excised inside-out patches at different Vm. Data were constructed from the means ± SE of experiments in which the bathing solution contained Kint solution with pipettes containing 145 mM KCl (dashed line) or 145 mM K-gluconate () (n = 6) and where the bathing solution contained half-strength KCl (Kint1/2) solution with pipettes containing 145 mM KCl () (n = 9).

Relative anion permeability. The relative permeability of various anions compared with Cl was assessed from measurements of current reversal potentials in ion substitution experiments. KCl 145 mM, initially present in the bathing solution, was replaced by KCl 72.5 mM + KX 72.5 mM (X being the anion to be tested), and the Cl concentration in the pipette was kept constant at 145 mM. The shift of the reversal potential was expressed with respect to the KCl solution after correction for the calculated junction potential. Ion selectivity was calculated from

Table 3 gives the permeability ratios P_anion/P_Cl and conductances calculated with different anions. As shown in Table 3, the permeability sequence was I > SCN > Br > Cl > NO₃ > F SO₄²  gluconate. The calculated permeability ratios for SO₄² and gluconate were indistinguishable from zero. Moreover, the conductance for these two anions measured between 40 and 100 mV was lower than the equivalent conductance measured with K_int1/2 in the bath, suggesting that they may also block this ORCC.

Kinetic analysis of the channel in excised inside-out patches. Figure 2 shows that P_o was an increasing linear function of V_m. Moreover, in spite of the relatively low frequency of channel occurrences, several identical channels (up to five) were simultaneously present in active membrane patches. For example, a total of 206 multiple ORCCs were observed in 120 recordings. In addition, Fig. 1A clearly shows that for a given membrane patch, the number of channels simultaneously active was proportional to the V_m with the highest values seen at positive potentials. However, the observed increase in open states (also observed in the cell-attached configuration) most likely results from the voltage dependence of P_o. The probability of simultaneous opening of 1-5 channels was also calculated as described by Colquhoun and Hawkes (6). This probability was in good agreement with that predicted for independent channels by a binomial distribution with a single channel P_o of 0.57. Dwell time analysis was performed on patches containing only one Cl channel, and the kinetic analysis was made at a holding potential of +50 or 50 mV. At these potentials, the channel consistently displayed burst/gap behavior. The ORCC kinetics were characterized by very low P_o at negative potentials (P_o= 0.23 ± 0.07, n = 11) compared with positive potentials (P_o= 0.57 ± 0.04, n = 19), corresponding to longer MCT (30.7 ± 8.4 ms, n = 11, vs. 7.6 ± 0.9 ms, n = 19). Within bursts, the channel kinetics were characterized by a lower P_o at negative potentials (0.32 ± 0.01, n = 413 bursts) compared with positive potentials (0.65 ± 0.01, n = 195 bursts), as shown in Table 4. Open time durations were fitted best by the sum of two exponential distributions at +50 mV (Table 4) and by a single exponential distribution at 50 mV. Closed time distribution was always fitted by the sum of three exponentials.

8

3

Effect of channel activators or blockers in intact cells. Several attempts were made to induce channel activity in the cell-attached configuration with the aid of agents known to increase intracellular pH, Ca²⁺, or cAMP or to directly activate protein kinases A or C. All trials were made either by incubating the crypts in Ringer solutions containing the drugs at 30, 60, or 90 min before patch clamping (n = 15-20 for each drug concentration) or by addition to the bath after obtaining the gigaohm seal (n = 15-20 for each drug concentration). Forskolin (10 or 20 µM), histamine (10 µM), carbachol (50 µM), and A-23187 (1 or 100 µM) failed to induce any channel activity consistent with a Cl conductance. In the same patches, ORCC could be activated by voltage after transition to the excised inside-out configuration. Activators of protein kinase C (phorbol 12-myristate 13-acetate, 1, 2, and 4 µM; phorbol 12,13-dibutyrate, 1, 2, and 3 µM) also failed to activate ORCC in cell-attached patches.

Effect of channel activators or blockers in excised patches. Figure 4 shows the effects of 50 µM NPPB on the channel activity. Addition of the blocker to the bathing solution led within seconds to a 95% inhibition of ORCC activity. The blockade was characterized by a reduction (94.0 ± 1.9%) in NP_o from 0.77 ± 0.12 (n = 10) to 0.03 ± 0.01 (n = 10) in parallel with a decrease in the unitary current (control = 2.58 ± 0.08 pA and NPPB = 1.45 ± 0.19 pA; n = 10). This unitary current decrease results from the flickery block induced by NPPB. Subsequent washout of NPPB from the bathing solution slowly led to 51.7 ± 10.4% recovery of the channel NP_o.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Effect of 5-nitro-2-(3-phenylpropylamino)benzoate (NPPB) on ORCC activity. A: representative single channel currents of ORCC recorded from excised inside-out membrane patches at the holding potential of 50 mV before and after exposure of the cytosolic patch face to 50 µM NPPB. The bathing solution contained Kint, and pipettes contained 145 mM KCl. B: NPo of current trace shown in A where N is the number of channels. Records presented in A are located with arrows.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Effect of 9-anthracenecarboxylic acid (9-AC) on ORCC activity. A: representative single channel currents of ORCC recorded from excised inside-out membrane patches at the holding potential of 50 mV before and after exposure of the cytosolic patch face to 50 and 100 µM 9-AC. The bathing solution contained Kint, and pipettes contained 145 mM KCl. B: typical recording of NPo evolution before and after exposure of the cytosolic patch face to 50 and 100 µM 9-AC, where N is the number of channels.

Absence of channel activation by phosphorylation in excised patches. Previous studies have shown that protein kinase A stimulates ORCC in different types of epithelial cells (3, 28, 44). However, in the mouse colonic crypt cells in the excised inside-out configuration, all attempts (n = 10) to activate the channel with the catalytic subunit of bovine or porcine cAMP-dependent protein kinase (50 or 100 nM) in the presence of its cofactor Mg²⁺-ATP, failed to show any effect. This implies that the Cl channel is not regulated by protein kinase A phosphorylation or that such regulation requires some additional factor lost on excision.

G protein regulation. To determine whether G proteins regulate ORCC activity, we examined the effects of guanosine 5'-O-(3-thiotriphosphate) (GTPS) and guanosine 5'-O-(2-thiodiphosphate) (GDPS) on the P_o and amplitude of the single channel current in inside-out patches. GTPS (5, 10, 15, or 20 µM) added to the solution bathing the cytoplasmic side of the patch reduced P_o and single channel activity by ~50% (Fig. 6, A and C) with no apparent significant concentration dependence (at least in the range of 5-20 µM) observed. This lack of concentration dependency could reflect the fact that 5 µM GTPS already induced a maximal effect on the channel P_o and unitary current. Because the observed decrease in the channel conductance induced by GTPS could result from a flickery block, the unitary current in the presence of this nucleotide was determined for different filter frequencies (500 Hz and 1, 2, and 5 kHz). No increase in the measured current was observed with increasing filtration frequency, suggesting that a flickery block was not involved, although the possibility that the frequency of block exceeded the fastest filter rate used (5 kHz) cannot be definitively excluded.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   G protein regulation of ORCC activity. Representative single channel currents of ORCC recorded from excised inside-out membrane patches at the holding potential of 50 mV. A: before and after exposure of the cytosolic patch face to 10 µM guanosine 5'-O-(3-thiotriphosphate) (GTPS). B: before and after exposure of the cytosolic patch face to 50 µM guanosine 5'-O-(2-thiodiphosphate) (GDPS), without or with previous exposure to 10 µM GTPS. C: diagrams showing the dose dependency of the reductions of recorded currents and relative Po following exposure to GTPS, calculated from 6 experiments. The bathing solution contained Kint and pipettes contained 145 mM KCl.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The purpose of the present study was to identify and characterize the Cl channels present in the basolateral membrane of enterocytes in intact mouse distal colonic crypts. Compared with other methods using isolated enterocytes, the technique of dissociation of intact crypts has the obvious advantage of maintaining a clear differentiation between the apical and basolateral poles of the cells, along with the base-to-apex topology of the crypt.

Only one type of Cl channel was found, and this was an ORCC. The ORCC demonstrates an intermediary conductance (20-90 pS), strong outward rectification, and inhibition by DIDS, NPPB, 9-AC, and GTP. Outwardly rectifying anion channels have been described in a wide variety of epithelial cells and also in nonepithelial cell membranes. In most of these reports, the occurrence of such ORCCs was found to be extremely low in the cell-attached configuration, consistent with our findings of spontaneous activity in only ~1% of the patches under steady-state conditions. An exception to this is the intermediate conductance ORCC from chicken colon epithelial cells described by Fischer et al. (12). Although this channel was reported to show significant spontaneous activity, the study involved isolated cells, so it was not possible to determine whether the channel was located in the apical or basolateral membrane. In intact rat colon epithelium, Diener et al. (9) described an ORCC (13.5-40 pS) on the apical membrane that was spontaneously active in cell-attached patches and that remained active after excision of the patch. Some reports have suggested that ORCC activity can be induced in intact cells by cAMP (13, 16, 18, 20). However, this is not a general feature of ORCC, since, as in the present study, there are several reports of cAMP failing to influence ORCC activity (9, 49).

Another consistent property of ORCC from a variety of tissues is so-called excision activation (25, 48). At room temperature, ORCCs are usually quiescent after excision, and activation requires application of depolarizing membrane voltages (70-100 mV). In contrast, at 37°C the channels may activate after excision without application of membrane voltage. In the present study, channel activation at 37°C could not be studied because the isolated crypts rapidly lost their morphological integrity. At room temperature, the protocol for successful channel activation in this study (15- to 20-min waiting time and 70-mV depolarization steps) is very similar to that reported in the literature for other ORCCs (11, 13, 16, 28, 44). Under these conditions, channel activity was recorded in 16.5% of all seals. The observation of activation after excision has led to the suggestion that ORCCs might be tonically inhibited as long as the cell membrane is in contact with the cytoplasm and that the cell might contain a cytosolic inhibitor that diffuses away after patch membrane excision (11, 23, 26).

As with the effect of cAMP on ORCC activity in intact cells, channel activation by exposure to purified cAMP-dependent protein kinase catalytic subunit has been reported, but with varying success (3, 28, 44). As an example, in HT-29 cells, the number of patches presenting channel activity increased from 3 to 42% after exposure of the excised inside-out patches to protein kinase A + ATP (20). On the contrary, no activation was seen in T84 cell patches (48). Similarly, no activation by the protein kinase A catalytic subunit was observed in excised patches in the present study, although it was subsequently shown by voltage activation that the patches did contain ORCC.

The permeability order for the mouse colon basolateral ORCC described here was found to be I > SCN > Br > Cl > NO₃ > F SO₄²  gluconate, corresponding to the Eisenmann I sequence (53). Similar sequences have been reported for other Cl channels from enterocytes, including ORCCs (13, 24, 36). However, a higher relative permeability to nitrate than Cl has been reported for several other ORCCs (1, 27), including those in intestinal cells (15, 38), and this characteristic appears to be a common feature for this type of Cl channel. Like respiratory cells (25) or rabbit parietal cells (41), a higher permeability of Cl over nitrate was observed for the ORCC described in these crypt intestinal cells. The higher permeability for larger ions indicates that small ionic size does not favor flow through the channel. This may indicate that an important factor for permeation is the energy necessary to dehydrate the anion, because large anions have lower hydration energies and move more easily from the aqueous phase to the cationic sites located inside the channel (16). Partial substitution of Cl by other anions modified single channel conductance as well as the E_r. A similar anionic selectivity is obtained when determined from the shift in E_r or from the conductance measured at E_r. However, in contrast to observations for other ORCCs (13, 15, 24, 27), the selectivity sequence determined from conductance measured between 40 and 100 mV is different from the sequence derived from relative permeabilities.

In addition, the data show that stimulation of protein kinase C failed to activate the basolateral ORCC in mouse colon crypt cells, although this has been reported to be successful in the ORCC of airway epithelial cells (29). The absence of response to histamine, carbachol, and A-23187, all of which are known to cause increases in intracellular Ca²⁺ levels, demonstrates that Ca²⁺ is not a trigger for ORCC activation. This is further supported by the fact that changing the pCa value in the bath solution in the excised inside-out configuration did not modify single channel activity. A similar insensitivity to intracellular Ca²⁺ levels has been previously demonstrated for other ORCCs (4, 13, 14, 24, 36).

In contrast to the above, GTPS had pronounced effects on the activity of the mouse colon basolateral ORCC, decreasing single channel P_o and conductance by 50%. In addition, this could be readily reversed by addition of GDPS. The indication is that G proteins likely play a key role in the regulation of the basolateral ORCC of crypt cells. Inhibitory effects of G proteins on other ORCC have been previously described (19, 45), as well as on high-conductance (21, 31) and low-conductance Cl channels (42, 43). On the other hand, activation was seen in inwardly rectifying Cl channels (22, 50), as well as for high-conductance Cl channels (33, 46) or for Cl channels of very low conductance (30). As yet, the precise mechanisms underlying these stimulatory or inhibitory effects of G proteins remain unclear. It was suggested (45) that the inhibition of ORCC activity could occur through an activation of phospholipase A2 and subsequent production of arachidonic acid or from the inhibition of adenylate cyclase and the subsequent decreased pool of intracellular cAMP. It would seem that the latter mechanism is most unlikely for the mouse colon basolateral ORCC, since we never observed any activation of ORCC by cAMP-dependent protein kinase. Alternatively, the activity of the ORCC may be regulated by a phosphorylation/dephosphorylation process that is influenced by a G protein-dependent activation of phosphatases (40, 45). In the present study, activation of G proteins decreases the unitary current of the ORCC as well as its P_o. In a number of studies, the inhibitory effect of the activation of G proteins on Cl channels is via a decrease of the P_o without affecting the single channel conductance (21, 31, 43, 45). To the best of our knowledge, only one report of an effect of modulating G protein activity on channel conductance has been published. In this case, the inhibition of the unitary conductance by pertussis toxin, an agent known to inactivate G_i proteins, increased the P_o of an immunopurified ORCC incorporated in planar bilayer but also conferred linearity to the I-V relationship of this channel (19). The authors suggested that the rectification of the channel partially involved its interaction with the G protein. However, they also showed that the addition of GTP by itself had no effect on the channel conductance. The direct effect of GTP and GDP on the ORCC activity described in the present study seems to be a unique feature. G protein regulation of both P_o and unitary current provides a more efficient regulation of the channel activity than an effect on P_o alone. Clearly, the regulation of the activity of the mouse colon basolateral ORCC and the possible role of G proteins in such regulation is worthy of further investigation.

As to the possible physiological role(s) of this basolateral ORCC in the mouse colon, it is clear that such channels could be involved in the transport of Cl through the basolateral membrane of colonic crypts during the process of NaCl absorption. However, the stimulation of NaCl absorption, for example by somatostatin or by increased absorption of short-chain fatty acids (7, 10), is known to induce cell swelling and the induction of the subsequent process of regulatory volume decrease. Moreover, in intestinal cell lines, activation of ORCC was detected during cell volume regulation following hypotonic shock (24, 52). In some preliminary experiments, we attempted to examine changes in channel activity following exposure to hypotonic medium in the native mouse colon basolateral ORCC using the cell-attached configuration. However, these proved to be unsuccessful because of the fragility of the seal due to the large variations of membrane tension and to unavoidable movements of the intact crypt in the perfusion flow.

In conclusion, we have been able to demonstrate for the first time the presence of a Cl channel characterized by a strong outward rectification and sensitive to DIDS, NPPB, 9-AC, and GTP on the basolateral membrane of crypt cells from mouse distal colon. The presence of this ORCC could be demonstrated in both cell-attached and excised inside-out patch-clamp configurations. The existence of such a basolateral Cl channel is frequently postulated in the models for NaCl absorption in colonic cells, and our findings are therefore consistent with such a physiological role. Furthermore, the possible involvement of this basolateral ORCC in volume regulation processes following hypotonic shock should be considered, particularly as we have previously demonstrated a key role of a Cl exit pathway sensitive to NPPB and 9-AC during RVD in these cells (35). Clearly, further investigations are needed to precisely define the physiological role, or roles, of this channel in the basolateral membrane.