INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

MANY VISCERAL TISSUES are innervated by motoneurons that do not utilize adrenergic or cholinergic transmitters (nonadrenergic noncholinergic neurotransmission) (for reviews see Refs. 20 and 35). Recent evidence suggests that nitric oxide (NO) is a major neurotransmitter mediating nonadrenergic noncholinergic inhibition in many autonomically innervated tissues (9, 28, 34); however, other transmitters, such as ATP, vasoactive intestinal polypeptide (VIP), and pituitary adenylyl cyclase-activating peptide (PACAP), are coexpressed in inhibitory neurons and may be coreleased (10, 18, 29, 31). Many studies of gastrointestinal (GI) muscles have described at least two components of hyperpolarization [i.e., inhibitory junction potentials (IJPs)] in postjunctional cells in response to inhibitory nerve stimulation (19, 23, 36, 38). Typically, IJPs consist of an initial fast component of hyperpolarization that is usually attributed to the release of ATP. The fast component of the IJP is variably sensitive to apamin (20). In some preparations the fast component is totally blocked by 100 nM apamin, but in other preparations the fast component is far less sensitive (23). Superimposed on, and extending beyond, the fast component is a slower component that is attributed to release of NO and, possibly, VIP (19, 23, 38). Different species and tissues manifest different degrees of the slow and fast components, suggesting considerable heterogeneity in the postjunctional receptors and mechanisms linked to neurally evoked inhibition. For example, in the canine colon, enteric inhibitory inputs are mediated primarily by NO (13, 39) via activation of a variety of K⁺ channels (24), and apamin-sensitive responses are not present (37). In mouse and human GI muscles, apamin-sensitive and -insensitive components are present (23, 36). Although there has been speculation that the apamin-sensitive responses are mediated by ATP, the conductance(s) responsible for the fast component of IJPs has not been described. Therefore, it is unclear how ATP causes smooth muscle hyperpolarization.

It is thought that IJPs are due to the activation of K⁺ channels (40), and many authors have speculated that ATP activates small-conductance Ca²⁺-dependent K⁺ (SK) channels similar to those expressed in skeletal muscles (7), brain (25), and hepatocytes (11), because IJPs and responses to ATP are reduced by apamin (1), an inhibitor of SK channels. Apamin-sensitive and -insensitive isoforms of SK channels have been identified (25), and therefore it is possible that the apamin-insensitive portion of ATP responses and fast IJPs could be mediated via apamin-insensitive SK channels. The molecular structure of SK channels has recently been determined, and functional channels have been cloned from mammalian brain (25). SK channels are Ca²⁺ sensitive and voltage independent and have unitary conductances of 5-20 pS. These are characteristic properties that could be used to identify SK channels in GI smooth muscles.

In the present study we examined the effects of ATP on outward currents of murine colonic muscle cells to identify the conductance(s) that may mediate ATP-dependent inhibitory responses in the GI tract. We have used the patch-clamp technique and compared responses in dialyzed and "perforated-patch" cells to test the involvement of intracellular signaling mechanisms. Single channel recordings were also performed to identify specific K⁺ channels activated by ATP.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Cell preparation. Colonic smooth muscle cells were prepared from 20- to 30-day-old Balb/C mice of either sex. Mice were anesthetized with chloroform and killed by cervical dislocation, and the proximal colon was quickly removed. Colons were cut open along the longitudinal axis, pinned out in a Sylgard-lined dish, and washed with Ca²⁺-free Hanks' solution containing (in mM) 125 NaCl, 5.36 KCl, 15.5 NaHCO₃, 0.336 Na₂HPO₄, 0.44 KH₂PO₄, 10 glucose, 2.9 sucrose, and 11 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES). After removal of the mucosa and submucosa, pieces of muscle were incubated in a Ca²⁺-free Hanks' solution containing 4 mg/ml fatty acid-free bovine serum albumin (Sigma Chemical), 14 U/ml papain (Sigma Chemical), 230 U/ml collagenase (Worthington), and 1 mM dithiothreitol (Sigma Chemical). Tissues were incubated at 37°C in enzyme solution for 8-12 min and then washed with Ca²⁺-free Hanks' solution. Tissue pieces were gently triturated to create a cell suspension. Dispersed cells were stored at 4°C in Ca²⁺-free Hanks' solution supplemented with minimum essential medium for suspension culture (Sigma Chemical) and (in mM) 0.5 CaCl₂, 0.5 MgCl₂, 4.17 NaHCO₃, and 10 HEPES. The cells were allowed to adhere to the bottom of a recording chamber on an inverted microscope for 5 min before commencement of experiments. All experiments were performed within 6 h of dispersion of cells.

Voltage-clamp experiments. Dialyzed whole cell, perforated-patch whole cell, and single channel voltage-clamp experiments were performed on murine colonic smooth muscle cells. Currents were amplified with a List EPC-7 amplifier and digitized with a 12-bit analog-to-digital converter (model TL-1, DMA interface, Axon Instruments). Data were stored on videotape or digitized on-line using pCLAMP software (version 5.5.1 or 6.03, Axon Instruments). Data were sampled at 1-5 kHz and low-pass filtered at 0.2-1 kHz using an eight-pole Bessel filter. Probability density plots were obtained by scaling the amplitude histograms so that the total area under the curve was equal to 1.0. In the experiments where the Ca²⁺ sensitivity of the small-conductance K⁺ channels was studied in excised patches, a sampling rate of 1 kHz and a filtration rate of 0.2 kHz were used. Variations in the baseline were corrected with ASCD software (Dr. G. Droogmans, KU, Louvain, Belgium). After baseline adjustment, all-points amplitude histograms were constructed and open probability was determined. A sampling period of 2 min was used for these analyses.

Solutions. For the recordings of K⁺ currents with the dialyzed whole cell technique, the external Mn²⁺-containing physiological saline solution (MnPSS) consisted of (in mM) 5 KCl, 135 NaCl, 2 MnCl₂, 10 glucose, 1.2 MgCl₂, and 10 HEPES adjusted to pH 7.4 with tris(hydroxymethyl)aminomethane (Tris). In some experiments, MnCl₂ was replaced with CaCl₂ [Ca²⁺-containing physiological saline solution (CaPSS)]. Composition of the internal solution for dialyzed cells was (in mM) 110 potassium gluconate, 20 KCl, 5 MgCl₂, 2.7 K₂ATP, 0.1 Na₂GTP, 2.5 creatine phosphate disodium, 5 HEPES, and 1 ethylene glycol-bis(-aminoethyl ether)-N,N, N',N'-tetraacetic acid (EGTA) adjusted to pH 7.2 with Tris. For perforated whole cell patch-clamp experiments, composition of the pipette solution was (in mM) 110 potassium gluconate, 20 KCl, 0.5 EGTA, and 5 HEPES adjusted to pH 7.2 with Tris. Amphotericin B (90 mg/ml, Sigma Chemical) was dissolved with dimethyl sulfoxide, sonicated, and diluted in the pipette solution to give a final concentration of 270 µg/ml. External solution for these experiments was the same as for the dialyzed whole cell patch-clamp experiments. In some experiments, 3 × 10⁷ M apamin (Sigma Chemical) was added to the external solution. Freshly prepared K₂ATP, UTP, and 2-methylthioadenosine 5'-triphosphate (2-MeS-ATP) were applied to cells via addition to the bath solution and superfusion.

5

9

5

4

	RESULTS

Top Abstract Introduction Methods Results Discussion References

ATP effects in dialyzed whole cell configuration. The effects of ATP on outward current were first tested by external application of 1 mM ATP to cells voltage clamped with the conventional, dialyzed, whole cell configuration of the patch-clamp technique. EGTA (1 mM) was added to the pipette solution to buffer intracellular Ca²⁺. In the presence of external MnPSS (see METHODS), outward currents were generated by depolarizing test potentials from a holding potential of 80 mV. The outward currents reached a peak and inactivated during 150-ms test depolarizations. Data generated from inactivation protocols were fitted with a Boltzmann equation: y = 1/{1 + exp[(V  V_h)/k]}, where k represents the slope factor and V_h is voltage of 50% inactivation, which averaged 44 ± 1 mV (n = 5, data not shown). The time-dependent decay of the outward current was fit with a single time constant ( = 62 ± 2 ms at +20 mV; data not shown). At the end of 150-ms test potentials, the outward current had inactivated to 30% of the peak current. Application of external ATP had no effect on the outward currents elicited in dialyzed cells. Similar observations were made when CaPSS was used as the external solution: 1) inactivating outward currents were elicited by step depolarizations from 80 mV (Fig. 1, A and B); 2) the sustained current was 26% of the peak current at the end of 150-ms test pulses; and 3) ATP did not alter the amplitude of the peak or the sustained current (n = 4; Fig. 1, B-D).

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Effect of ATP on outward currents in dialyzed cells. A: current responses of a dialyzed cell to voltage-clamp steps from 80 to +80 mV in 10-mV increments. Depolarization evoked large, inactivating outward currents. B: addition of 1 mM ATP in presence of Ca2+-containing physiological salt solution (PSS) had no effect on outward current. C: summary of peak current responses from 4 cells. , Cells under control conditions; , responses after addition of ATP. Data show lack of effect of ATP on peak outward current in dialyzed cells. D: summary of sustained current responses (at end of 150-ms pulses); symbols as in C. Symbols and error bars from data sets before and after addition of ATP are essentially superimposed.

Effects of ATP on cells using the perforated-patch, whole cell configuration. In the presence of MnPSS, cells voltage clamped with the perforated-patch technique (34) showed a pattern of current responses similar to the currents observed in experiments on cells using the dialyzed, whole cell configuration (i.e., peak and sustained currents). Application of 1 mM ATP activated outward currents in cells in which the perforated-patch technique was used (Fig. 2, A and B). ATP produced a sustained outward component of current, as shown by difference currents (Fig. 2B). For example, at +40 mV the outward current was enhanced 18 ± 4% at the end of a 500-ms test depolarization (n = 6). The difference currents elicited by ATP were well fit by the Goldman-Hodgkin-Katz (GHK) equation for K⁺ (r² = 0.9726, n = 6; Fig. 2C). ATP induced a K⁺ permeability of 8.3 × 10¹⁵ cm · s¹ · cm².

View larger version (17K): [in this window] [in a new window]   Fig. 2.   ATP (1 mM) and 100 µM 2-methylthioadenosine 5'-triphosphate (2-MeS-ATP) increase outward currents in amphotericin-perforated cells. A: current records in response to depolarization from 80 to 40, 0, and +40 mV. , current records before ATP; , currents in presence of 1 mM ATP. B: difference currents (i.e., control currents subtracted from currents in presence of ATP); outward currents activated by ATP are described. C: summary current-voltage (I-V) curve for difference currents obtained from 6 cells. Data points of I-V relationship were fit by Goldman-Hodgkin-Katz (GHK) equation for K+ gradient. From fit of data, it was calculated that ATP induced a K+ permeability of 8.3 × 1015 cm · s1 · cm2. ATP-dependent current was resolvable at potentials negative to range in which voltage-dependent currents were activated (see Fig. 1C). D: currents generated in a cell bathed with Mn2+-containing PSS using permeabilized-patch technique. Control responses () and responses of same cell in presence of 100 µM 2-MeS-ATP () are superimposed. E: difference currents describing outward currents activated by 2-MeS-ATP. F: I-V relationship for difference currents in E (n = 4). Data points are fit with GHK equation for K+.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Effects of ATP on perforated-patch cells in presence of extracellular Ca2+. A: control currents generated by depolarization of an amphotericin-permeabilized cell bathed in Ca2+-containing PSS. B: current records in same cell after addition of 1 mM ATP. C: difference currents; ATP increased outward current during depolarization steps. Small tail currents noted on repolarization are suggestive of a Cl or nonselective cation conductance that may also have been activated by ATP. D: summary plot of I-V relationship of difference currents obtained from 5 cells. Data were fit with GHK equation for K+. Data were not fit well by GHK equation, suggesting that ATP activated more than a single conductance in Ca2+-containing PSS.

Effect of apamin on ATP-activated currents. Apamin (3 × 10⁷ M), an inhibitor of SK channels, was applied to the bath solution to test the pharmacological sensitivity of the current induced by ATP. Apamin itself had no effect on the outward current elicited in dialyzed cells (data not shown). Cells with permeabilized patches were stepped from a holding potential of 80 to +80 mV in 20-mV steps (Fig. 4). ATP (1 mM) increased outward currents, and apamin partially attenuated the response to ATP (32 ± 5%, n = 5; Fig. 4, A and B). Under control conditions, peak outward current averaged 760 ± 2 pA and ATP increased current amplitude to 889 ± 2 pA. In a series of time control experiments the outward current activated by ATP declined very little during 11 min of exposure (i.e., to 862 ± 4 pA, n = 5). In another group of five cells with average outward currents of 756 ± 4 pA, ATP increased the outward current to 884 ± 4 pA. Apamin decreased the outward current in these cells to an average of 827 ± 3 pA (n = 5).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Apamin attenuated outward current activated by ATP. A: currents generated by depolarization of a cell bathed in Mn2+-containing PSS using permeabilized-patch technique. Trace a: control record; trace b, increase in outward current produced by ATP; trace c: reduction in outward current after addition of 3 × 107 M apamin in continued presence of ATP in another cell. Apamin partially blocked ATP-induced current. B: peak currents in response to repetitive steps from 80 to 0 mV. , Time course of average responses to sustained exposure to ATP; , reduction in current caused by application of 3 × 107 M apamin. Points are averages ± SE from 5 cells exposed continuously to ATP (for at least 800 s) and 5 cells in which apamin was added during continuous exposure to ATP.

K⁺ channels activated by ATP. Single channel studies were performed on 198 patches to identify K⁺ channels activated by ATP. Application of 1 mM ATP to the external solution increased the open probability of a small-conductance channel in cell-attached patches (32 of 52 patches tested with this protocol). At a patch potential of 60 mV in the presence of 10⁷ M extracellular free Ca²⁺, few spontaneous channel openings were noted. However, brief openings of a small-conductance channel (mean open time = 2 ± 0.5 ms) were occasionally observed. Because of the brief open times of these channels, we could not fully resolve the amplitude of the currents (Fig. 5A). ATP increased open channel probability of the small-conductance channels. The amplitude of the unitary conductance was 0.35 pA at 60 mV. After 3 min of exposure to ATP, open probability was 0.14 ± 0.05 (Fig. 5, B and F), and, after 7 min, open probability reached a maximum level of 0.26 ± 0.07 (Fig. 5C). The increase in open probability resulted from an increase in mean open time (i.e., to 29 ± 3 ms), and multiple channel openings were also observed. Open probability of the channels decreased when ATP was removed from the external solution. Within the 1st min after removal of ATP, open probability decreased to 0.13 ± 0.06 (Fig. 5D). Several minutes of washout were required to return open probability to the control level (Fig. 5E). Figure 5F shows the changes in open probability elicited by ATP as a function of time (n = 7 patches in which the entire exposure and washout of ATP were accomplished).

View larger version (46K): [in this window] [in a new window]   Fig. 5.   ATP activated a small-conductance channel in cell-attached patches. A-E: representative excerpts of a continuous on-cell patch-clamp recording before and during exposure to ATP. A: current under control conditions, i.e., symmetrical 140 mM K+ and holding potential of 60 mV. Very brief openings of a small-conductance channel were observed (downward deflections). B: 1 mM ATP increased open probability, and a peak level was reached in C. D and E: effect of ATP on channel activity was reversible on washout. F: plot of channel activity, expressed as open channel probability (NPo), as a function of time (n = 7); letters (A-E) indicate times at which traces in A-E were recorded.

I-V relationship of ATP-induced current. To further characterize the small-conductance channels activated by ATP, we performed steady-state recordings and step protocols on inside-out excised patches. Ionic conditions included symmetrical (140 mM K⁺ in pipette and bath solution) and asymmetrical (5 mM K⁺ in pipette and 140 mM K⁺ in bath solution) K⁺ gradients with 10⁷ M extracellular free Ca²⁺. With the inner surface of the patch exposed to this concentration of Ca²⁺, spontaneous openings of the small-conductance channels were observed. Amplitudes of the currents were plotted as a function of potential (Fig. 6C), and the slope conductance in symmetrical K⁺ solutions was determined by linear regression to be 5.3 ± 0.02 pS (n = 4). Because of the small amplitude of the currents, it was difficult to resolve single channel openings at potentials positive to 40 mV in symmetrical K⁺ solutions (Fig. 6A).

View larger version (25K): [in this window] [in a new window]   Fig. 6.   I-V relationship of small-conductance K+ channel in excised membrane patches. A: steady-state recordings from an inside-out patch with symmetrical 140 mM K+ and Ca2+ buffered to 107 M. Solid line and c indicate 0 current when channels were closed; channel openings are represented by downward deflections. B: recordings made in asymmetrical K+ (5 mM in pipette and 140 mM in bath solution at inside surface of patch). c, Closed; o, open. Data were sampled at 1 kHz and filtered at 0.2 kHz. C: I-V relationship for currents in experiments depicted in A () and B (). Data points were fitted with GHK equation. Note shift in extrapolated reversal potentials that would be predicted if channels were K+ selective.

Ca²⁺ sensitivity of the 5-pS K⁺ channels. The Ca²⁺ sensitivity of the 5.3-pS K⁺ channels was assessed by exposing the intracellular surface of excised membrane patches held at 60 mV to several concentrations of free Ca²⁺ (n = 4). When Ca²⁺ was increased from 10⁹ to 10⁸ M, there was no change in open probability (Fig. 7, A and B, traces a and b). At 10⁷ M Ca²⁺, very brief channel openings were observed (Fig. 7, A and B, trace c), but open probability was low (Fig. 7C). When Ca²⁺ was increased to 10⁶ M, two to three channels were activated and open time increased [53 ms vs. 2 ms (control); Fig. 7, A, B, trace d, and D]. At 10⁵ M free Ca²⁺ the number of channel openings increased additionally (Fig. 7, A and B, trace e). The effects of increased Ca²⁺ concentration were fully reversible, and reducing Ca²⁺ back to very low concentrations (10⁸ M) decreased open probability in a concentration-dependent manner. Figure 7C shows a plot of open probability vs. Ca²⁺ concentration, and the 50% effective concentration, calculated from the averaged data, was 4.9 × 10⁷ M Ca²⁺.

View larger version (35K): [in this window] [in a new window]   Fig. 7.   Ca2+ sensitivity of small-conductance channel in inside-out excised patches. Intracellular surface of an inside-out membrane patch was exposed to 109-105 M free Ca2+. A: excerpts of current records while free Ca2+ concentration in bath solution was changed to levels indicated above and below current traces. Open probability of 5.3-pS channel was increased as free Ca2+ increased. B: parts of trace in A are expanded (as indicated by a-h in A), and unitary conductance levels are estimated by dotted lines (solid line and c indicate closed state). At 106 and 105 M Ca2+, openings of at least 2 channels were evident. C: plot from 4 experiments of open probability vs. logarithm of free Ca2+ concentration ([Ca2+]); 50% effective concentration (EC50) was calculated to be 4.9 × 107 M. D: all-points histogram obtained from records taken in presence of 106 M free Ca2+; opening of at least 2 channels is demonstrated.

Activation of 5-pS K⁺ channels by caffeine. The fact that open probability of the 5-pS K⁺ channels was increased by intracellular Ca²⁺ suggests that entry of Ca²⁺ or release of Ca²⁺ from intracellular stores could mediate the activation of these channels by ATP in the cell-attached configuration. Therefore, we tested the effects of caffeine to determine whether release of stored Ca²⁺ might activate the 5.3-pS channels. Caffeine (500 µM) was applied in the external bath solution, and single channel activity was monitored while stepping from 80 to 0 mV in the cell-attached configuration with an asymmetrical K⁺ gradient (i.e., 5 mM K⁺ in the pipette and 140 mM K⁺ in the bath solution). Three K⁺ channels, with conductances of 250, 13, and 5.3 pS (Fig. 8A), were observed on step depolarizations. The open probability of 13-pS channels, likely to be delayed rectifier channels (see above), was unaffected by ATP or caffeine. The 250-pS channels, identified as BK channels, were activated by caffeine treatment (open probability = 0.01 ± 0.01 and 0.12 ± 0.02 for control and caffeine treatment, respectively, n = 4). The increase in BK open probability was taken as an indication that the caffeine treatment released intracellular Ca²⁺. Caffeine also increased open probability of the 5.3-pS channels (Fig. 8, B and B'), which occurred in a biphasic manner. After application of caffeine, there was an initial increase in the openings of the 5.3-pS channels; however, this effect decreased within 3 min. After exposure to caffeine (8 min), addition of ATP had no effect on 5.3-pS channels (data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Caffeine activated 5.3-pS and large-conductance K+ (BK) channels in on-cell patches. A: control recordings from an on-cell patch. Patch was stepped from 80 to 0 mV. Three unitary conductances were noted in this patch. Trace a: opening of 5.3-pS channel; trace b: opening of 13-pS channel (o, open state; c, closed state). Openings of BK channels (large transitions) can also be observed. B: 500 µM caffeine, which releases Ca2+ from intracellular stores, increased open probability of BK channels (confirming that caffeine increased intracellular Ca2+) and 5.3-pS channels. Caffeine did not affect 13-pS channels. A' and B': all-points amplitude histograms before and after caffeine treatment, respectively. In A', histogram was well fit with a Gaussian distribution (solid line); histogram in B' was not well fit by same function as a result of activation of 5.3-pS channels.

Effects of 2-MeS-ATP and UTP on 5.3-pS K⁺ channels. To observe the effects of more specific agonists of P_2y receptors, we also tested whether 2-MeS-ATP and UTP enhanced the open probability of the 5.3-pS channels. 2-MeS-ATP at 10 µM (n = 3) and 100 µM (n = 4) increased the open probability of the 5.3-pS channel in cell-attached patches. Figure 9 also shows all-points amplitude histograms for control conditions and the response to 2-MeS-ATP stimulation (open probability = 0.55 ± 0.11, n = 4) at 60 mV. UTP (1 mM; n = 3) also increased the open probability of the 5.3-pS channels; however, 100 µM UTP was without effect (n = 3; data not shown).

View larger version (32K): [in this window] [in a new window]   Fig. 9.   Activation of 5.3-pS channels by 2-MeS-ATP. A: representative excerpts from an on-cell recording held at 60 mV. B: continuous recording after addition of 100 µM 2-MeS-ATP. This caused a significant enhancement in open probability of small-conductance channels (closed state indicated by solid line). A' and B': current amplitude histograms for recordings in A and B, respectively. Note significant increase in open probability (Po) for 5.3-pS channel.

Effects of PPADS on activation of 5.3-pS K⁺ channels. Because 2-MeS-ATP increased the open probability of the 5.3-pS channel, we tested whether PPADS, a P₂ receptor blocker (27), inhibited this effect. PPADS (10⁵ or 10⁴ M) was added to the external solution, and single channel activity was recorded in the cell-attached configuration. PPADS itself did not change the activity of the 5.3-pS channels under control conditions (Fig. 10, A and B). In the presence of PPADS, application of 1 mM ATP or 100 µM 2-MeS-ATP did not increase the open probability of 5.3-pS K⁺ channels (Fig. 10C). Caffeine was able to activate 5.3-pS channels in the presence of PPADS (open probability = 0.18 ± 0.07, n = 4; Fig. 10D).

View larger version (33K): [in this window] [in a new window]   Fig. 10.   Pyridoxal phosphate 6-azophenyl-2',4'-disulfonic acid tetrasodium (PPADS) prevented enhancement in 5.3-pS channel open probability in response to 2-MeS-ATP. A: 4 traces under control conditions from a cell-attached patch held at 60 mV. Note small downward deflections in records, which indicate brief openings of small-conductance K+ channels. B: continuation of record after addition of 100 µM PPADS. PPADS alone did not alter channel activity. C: PPADS prevented stimulatory effect of 100 µM 2-MeS-ATP on 5.3-pS channel. D: addition of 500 µM caffeine increased channel activity in presence of PPADS. Right: all-points amplitude histograms.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Stimulation of intrinsic enteric inhibitory nerves leads to hyperpolarization and relaxation of GI muscles (6, 20). There has been considerable work to determine the transmitters that mediate enteric inhibitory responses in these muscles, and ATP, VIP, PACAP, and NO have been considered candidates. In fact, enteric inhibitory neurotransmission appears to be accomplished by at least two transmitters released in parallel from the same population of inhibitory motoneurons. The postjunctional mechanisms that mediate the inhibitory effects are not well understood, but this question has importance, because the postjunctional receptors and effectors (i.e., ion channels or other cellular mechanisms) may be excellent targets for therapeutic agents to regulate GI motility. Recent work has described several K⁺ channels in canine colonic muscles that are activated by NO (24), and the present study has addressed the conductance activated by purinergic agonists. ATP and the P_2y receptor agonist 2-MeS-ATP activated 5.3-pS Ca²⁺-activated K⁺ channels. These channels had properties similar to SK channels, the small-conductance Ca²⁺-activated K⁺ channels recently cloned from mammalian brain (25). Besides their high sensitivity to Ca²⁺, SK channels are voltage and time independent. These channels are ideal candidates for activation of outward current at the negative resting potentials of GI muscles. ATP was found to activate outward current at all potentials positive to the K⁺ equilibrium potential (Fig. 2). A portion of the outward current activated by ATP and 2-MeS-ATP was blocked by apamin, and apamin is known to reduce the fast component of IJPs in GI muscles of many species (20). Thus channels with properties consistent with SK channels are expressed by murine colonic myocytes, and these channels may mediate the fast component of IJPs in GI muscles. The data from the present study and the work of others are consistent with the following hypothesis for ATP-dependent hyperpolarization: 1) ATP activates P_2y receptors; 2) P_2y receptors lead to mobilization of internal Ca²⁺; 3) increased intracellular Ca²⁺ (most likely within the submembranous compartment) activates SK channels; and 4) enhanced K⁺ conductance brings membrane potential transiently closer to the K⁺ equilibrium potential.

It is quite clear that the BK channel, another type of Ca²⁺-activated K⁺ conductance channel, is not responsible for ATP-dependent hyperpolarization, because blockers of this channel do not affect IJPs (S. Ward, personal communication). Ca²⁺-activated K⁺ channels with conductances smaller than BK channels have been found in the rabbit portal vein (22), taenia coli (21), and jejunum, guinea pig mesenteric artery (4), and canine colon (24), but the channel conductances (63-150 pS) are much greater than the conductance activated by ATP in murine colonic smooth muscles (i.e., 5.3 pS). The medium-conductance Ca²⁺-activated K⁺ channels have been reported to be insensitive to apamin (8, 47). Apamin blocks some isoforms of SK channels (25) and reduces responses to inhibitory transmitters and ATP in a variety of GI muscles, including guinea pig stomach (26, 43), taenia coli (30, 43, 46), and ileum (48). This suggests that a portion of the responses to enteric inhibitory transmitters may be due to activation of SK channels. The present study documents the expression of a small-conductance, Ca²⁺-activated K⁺ channel with properties consistent with SK channels in GI muscles.

ATP released from nerves could activate different types of purinergic receptors expressed by smooth muscle cells. Others have documented excitatory effects of ATP released from extrinsic sympathetic neurons in colonic muscle, and these effects are mediated by P_2x receptors (41). P_2x receptors are ligand-gated nonselective cation channels (5, 16, 44). Therefore, it is unlikely that P_2x receptors could mediate IJP responses, because activation of nonselective cation channels would result in inward current and depolarization (14, 32). P_2y and P_2u receptors are G protein coupled and mediate inositol trisphosphate (IP₃)-induced release of Ca²⁺ from internal stores (2). By initiating the release of small amounts of Ca²⁺ near the plasma membrane, P_2y and P_2u occupation could lead to the activation of Ca²⁺-dependent outward currents. In the present study we found that ATP and the P_2y agonist 2-MeS-ATP activated small-conductance K⁺ channels. This observations suggests that P_2y receptors could mediate ATP-dependent hyperpolarization. We cannot entirely rule out involvement of P_2u receptors, but we found UTP to be less effective than 2-MeS-ATP in activating SK channels. Nothing is known about the expression of P_2u receptors in colonic muscles, and the effects of UTP could have been mediated by nonspecific activation of P_2y receptors. We also found that the P₂ receptor antagonist PPADS (27) blocked responses to ATP and 2-MeS-ATP, but we view this observation with caution, because a recent study has suggested that PPADS may nonselectively inhibit IP₃ receptors, rather than block P_2y receptors (42). It should be clearly stated that the exact nature of the receptor(s) that mediates the ATP responses we have observed is difficult to determine, since highly selective antagonists for these receptors are not available.

ATP had no effect on currents in cells with use of the standard, dialyzing configuration of whole cell recording. This finding suggests that intracellular second messengers, such as Ca²⁺ or other factors, are required for ATP-induced activation of outward current. The cells were dialyzed with solutions containing EGTA in these experiments, and it is possible that changes in intracellular Ca²⁺ are important in mediating the response to ATP. When the perforated-patch technique, which preserves intracellular signaling mechanisms (34), was used, ATP activated a K⁺ conductance (i.e., the current was outward and the I-V relationship of the conductance was well fit by the GHK equation) when an extracellular solution containing Mn²⁺ (to replace Ca²⁺) was used. When standard Ca²⁺-containing extracellular solution was used, the currents generated by ATP were not well described by the GHK equation, and inward tail currents were observed on repolarization. Thus ATP may generate a mixed current consisting of the K⁺ current elicited in Mn²⁺ solution plus Ca²⁺-dependent Cl or nonselective cation conductances. The characteristics of these inward conductances are not a subject of this study. The fact that a mixed current is generated on perfusion of cells with ATP suggests that release of this substance from nerves in situ must be directed at specialized populations of receptors [i.e., specialized motor end plate regions, as demonstrated morphologically in some purinergically innervated tissues (17)]. Otherwise it would be difficult to explain how release of ATP from extrinsic sympathetic nerves results in excitation (41) whereas release of ATP from intrinsic inhibitory neurons results in hyperpolarization and inhibition (20).

Binding of P_2y receptors is coupled via G proteins to activation of phospholipase C, IP₃ formation, and release of Ca²⁺ from internal stores (15). We suggest that this may be the major pathway of coupling between ATP binding and activation of SK channels, because 1) ATP had no effect on cells dialyzed with EGTA to buffer intracellular Ca²⁺, 2) ATP activated an outward current in cells in which influx of Ca²⁺ was blocked by replacement of external Ca²⁺ with Mn²⁺, 3) caffeine treatment also activated SK independent of purine receptors, and 4) after an initial activation of SK with caffeine, open probability waned, possibly from depletion of sarcoplasmic reticulum Ca²⁺ stores. During this refractoriness, further addition of ATP failed to activate SK channels. The fact that PPADS also blocked SK channel activation in response to ATP and 2-MeS-ATP may also support the notion that IP₃-dependent release of Ca²⁺ is responsible for activation of SK channels, since PPADS may block this mechanism (42).

It might seem odd that the effects of a compound reputed to be an inhibitory neurotransmitter are mediated by the release of Ca²⁺ in smooth muscle cells. However, recent studies have shown that Ca²⁺ release from internal stores can be highly directional and cause local increases in subsarcolemmal Ca²⁺ concentration without affecting global Ca²⁺ (33). A number of Ca²⁺-dependent conductances are expressed in smooth muscle cells (12), and the issue of how Ca²⁺-dependent inhibitory or excitatory responses might be preferentially activated should be considered. One level of selectivity may result from the relative Ca²⁺ sensitivity and voltage dependence of inward and outward currents. For example, SK channels are more sensitive to Ca²⁺ (7) than most of the other conductances present. We found that SK channels were activated at <100 nM Ca²⁺, even at negative membrane potentials. Therefore, small increases in Ca²⁺ near the membrane may selectively activate SK channels, and agonists that release focalized, small amounts of Ca²⁺ may preferentially cause activation of SK channels. Higher concentrations of Ca²⁺, perhaps supplemented by entry of Ca²⁺ or more potent stimulation of Ca²⁺ release, may facilitate nonselective cation channels and activate Ca²⁺-dependent Cl conductances and BK channels (for review see Ref. 12). Opening of BK channels tends to be constrained by the negative membrane potential range over which smooth muscle cells operate, but this conductance appears to provide important feedback regulation of membrane potential in some smooth muscles (8). Activation of Cl and nonselective cation channels provides a depolarizing influence that is reinforced by positive feedback as Ca²⁺ enters via voltage-dependent Ca²⁺ channels. Ca²⁺-dependent conductances could also be selectively regulated by the physical proximity of receptors and channels. For example, the location of P_2y receptors near SK channels may facilitate Ca²⁺-dependent coupling, because local production of IP₃ might produce a "sparklike" event that might affect nearby channels without significantly affecting Ca²⁺ levels near more distant channels. We were able to activate SK channels within the pipette in the "on-cell" configuration by adding purines to the external bath solution. This observation does not support an extremely close obligatory association between P_2y receptors and SK channels and suggests that IP₃ production resulting from P_2y binding may raise local Ca²⁺ within a diameter of 1 µm.

Responses to ATP were sustained for several minutes in whole cell and single channel experiments. Sustained hyperpolarization and inhibition of spontaneous electrical activity have also been noted when murine colonic muscles are exposed to exogenous ATP (45). If ATP-dependent activation of SK channels is due to release of Ca²⁺ from the sarcoplasmic reticulum, then the sustained nature of these responses suggests that ATP stimulation increases the rate of Ca²⁺ release rather than causes a massive dumping of the Ca²⁺ stores. A graded increase in the open probability of Ca²⁺ release channels may facilitate local increases in Ca²⁺ in the vicinity of SK channels without having an impact on global Ca²⁺. Graded release of Ca²⁺ from intracellular stores has been previously demonstrated by responses to caffeine. In the first reported measurements of spontaneous transient outward currents (STOCs), now considered a hallmark of Ca²⁺ release from stores in smooth muscles, caffeine had a differential response depending on dose: a high concentration of caffeine caused a burst of STOCs and then quiescence; a lower concentration caused a sustained increase in the frequency of STOCs (3). These observations are now known to be due to the effects of caffeine on ryanodine receptors (33). A high concentration of caffeine greatly increases the open probability of ryanodine receptors and "dumps" the Ca²⁺ stores. After the stores are dumped, there are no release events and no STOCs. A lower caffeine concentration increases the open probability of ryanodine receptors to a lesser extent, and the stores are not rapidly emptied. Thus the increase in the frequency of STOCs is sustained. Something of this nature (i.e., graded Ca²⁺ release via an IP₃-dependent mechanism) may explain the sustained effects of ATP on SK channels in murine colonic cells.

We are unable to say with certainty that the 5.3-pS single channel currents activated by ATP and 2-MeS-ATP are the basis for the whole cell currents and hyperpolarization responses in intact muscles activated by these agonists. We could not make clear recordings of such small channels in "outside-out" patches, and therefore we could not adequately test the effects of apamin on these channels. Even if this experiment were practical, it may not be definitive, because apamin reduced peak whole cell current by only about one-third. ATP may activate more than a single type of K⁺ conductance in whole cell experiments, and it is also possible that apamin-sensitive and -insensitive isoforms of SK channels could contribute to the whole cell response to ATP. It would be very difficult to demonstrate the latter simply by addition of apamin to the pipette solution. A tentative link between the single channels and whole cell currents can be made inferentially on the basis of the following observations: 1) the current activated by ATP was partially blocked by apamin; 2) apamin blocks some isoforms of SK channels (25); 3) whole cell currents were not resolved when cell interiors were dialyzed and buffered for Ca²⁺; 4) SK channels are Ca²⁺ dependent (7); and 5) the single channels activated by ATP and Ca²⁺ had properties consistent with SK channels. Other conductances, not manifest in our single channel recordings, could also contribute to the whole cell response. In the future, it will be interesting to test the hypothesis that apamin-sensitive and -insensitive isoforms of SK channels are expressed by GI muscles using Northern analysis and isoform-specific cDNA probes.