Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Ca²⁺-activated K⁺ (K_Ca) channels are located on the surface of a variety of cells, including airway smooth muscle (ASM) cells (1, 2), where they are involved in the regulation of membrane polarization and, in turn, of muscle tone (3). The activity of this type of channel is under the control of many substances, including some used frequently for clinical treatments (4). One of these substances, nitric oxide (NO·), was initially identified as the endothelium-derived relaxing factor (EDRF) (5, 6). In the airways, NO· is released principally by inhibitory nonadrenergic, noncholinergic (NANC)_i neurons or by epithelial cells to modulate the tone of adjacent smooth muscle cells (7). The mechanisms by which NO· induces its relaxing effect have been extensively studied in the vascular system (11, 12), but much less investigated in ASM, despite their clinical relevance (13). NO· is thought to act through activation of the soluble guanylate cyclase enzyme that catalyzes a transient increase in the cytoplasmic level of guanosine 3',5'-cyclic monophosphate (cGMP). This increase in cGMP activates a cGMP-dependent protein kinase (PKG) that phosphorylates different key proteins implicated in smooth muscle relaxation (14). Previous pharmacological studies performed on ASM have shown that specific inhibitors of K_Ca channels antagonize the relaxation induced by either endogenous (8) or exogenous NO· (11, 15), suggesting that this type of channel is one of the effectors by which NO· induces its relaxing effect. Moreover, it has been reported that K_Ca channels in airways account for almost 50% of NO-mediated relaxation (8, 11), compared with only 20- 40% in vascular smooth muscle (VSM). These observations indicate the existence of differences in the molecular mechanisms that lead to K_Ca-channel activation in VSM and ASM. Indeed, previous patch-clamp experiments revealed two modes of K_Ca-channel activation by NO· in VSM: a cGMP-dependent mechanism, involving activation of guanylate cyclase and a PKG-dependent phosphorylation step (16), and a cGMP-independent mechanism (19). Despite frequent reports that the activation of K_Ca channels mediates relaxation elicited by NO· in ASM (7, 20, 21), it has never been shown whether or not the effect of NO was due to a cGMP-independent mechanism. However, in a recent study, we demonstrated that native and purified K_Ca channels, derived from bovine tracheal smooth muscle cells and reconstituted into planar lipid bilayers (PLB), were activated by a PKG-dependent phosphorylation of the K_Ca channel in the presence of cGMP, Mg-ATP, and PKG (22). Biochemical assays using [-³²P]ATP also revealed that PKG phosphorylates a protein identified as being part of the -subunit of the K_Ca channel (23, 24). Thus, in ASM, one mode of K_Ca-channel activation would be cGMP-dependent.

The mechanisms by which NO·, released from nerve terminals via NANC stimulation and/or airway epithelial cells, triggers ASM relaxation are still not fully understood. On the basis of results of previous studies, at least three intracellular pathways have been proposed to explain how NO· can decrease smooth muscle tone. The first two pathways would depend on an increase in the level of intracellular cGMP, owing to activation of the soluble guanylate cyclase by NO· (11, 15). In ASM, these pathways result in the phosphorylation of either regulatory proteins of the contractile elements or surface-membrane channels. Bolotina and colleagues (19) recently demonstrated that in VSM, NO·-induced relaxation might be independent of an increase in the level of cGMP. Their patch-clamp measurements supported a direct activation of K_Ca channels by NO·. Our working hypothesis assumed that this third pathway may also exist in ASM, but it had not yet been investigated in in vitro experiments. Therefore, the aim of our investigation was to test whether NO· can activate K_Ca channels by directly regulating single K_Ca channels into PLB. Direct measurements of K_Ca-channel activities were made with 3-morpholinosydnonimine hydrochloride (SIN-1) as an NO· donor in reconstitution experiments under free-[Ca²⁺] and voltage-clamp conditions. Pharmacomechanical measurements were made in parallel to evaluate the sensitivity of precontracted smooth muscle strips to NO-induced relaxation. We demonstrated that the activation of K_Ca channels mediates part of NO-induced ASM relaxation. Moreover, this relaxation appeared to result in part from a direct cGMP-independent activation of K_Ca channels by NO·, involving a nitrothiosylation reaction that could change the gating of the K_Ca channel without affecting its Ca²⁺ sensitivity.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

NO· donor Preparations

NO· is highly unstable in solution. Therefore, we used the relatively more stable compounds 1-propanamine-3-(2-hydroxy-2-nitroso-1-propylhydrazine (PAPA NONOate) or SIN-1 as sources of NO·. PAPA NONOate was prepared as a 9-mM basic stock solution in 10 mM NaOH, pH 11, and stored at 20°C. Before each experiment, a sample of this stock solution was thawed and diluted 3-fold in the appropriate KCl buffer, pH 8, saturated with argon gas. An aliquot of this 3-mM PAPA NONOate solution was then added to the 3-ml experimental chamber to achieve the final concentrations. SIN-1 degrades to SIN-1A and then to SIN-1C after releasing O₂·, NO·, and a proton (25). When O₂· is liberated as a free radical, superoxide dismutase (SOD) has to be present to eliminate O₂· in order to prevent or at least to limit the production of peroxynitrite (ONOO) that would result from the oxidation of NO· in the presence of trace quantities of metals. Thus, all experiments were performed in the presence of ethyleneglycol- bis-(-aminoethyl ether)-N,N'-tetraacetic acid (EGTA) and/or ethylenediaminetetraacetic acid (EDTA) to chelate these trace contaminants.

SIN-1 was diluted in the following solution (in mM): 50 KCl, 20 potassium-4-(2-hydroxyethyl)-1-piperazine-N'-2-ethanesulfonate (K-Hepes), and 0.002 free Ca²⁺ (109 µM CaCl₂ + 110 µM K-EGTA), pH 7.4, containing 50 U/ml SOD (Sigma Chemical Co., St. Louis, MO) to prevent NO· breakdown (26) and peroxynitrite formation (27, 28). The solution was prepared 15 min before being tested experimentally, and was not used thereafter for longer than 1 h after the time of its preparation. It was also continually kept protected from ultraviolet (UV) irradiation by natural light. During all experiments, the pH, temperature, and susceptibility of solutions to O₂ were controlled in order to keep quantitatively constant the amount of NO· released from SIN-1 (25). The capacity of SIN-1 to release NO· was verified biologically by its ability to induce a sustained relaxation of bronchial smooth muscle for at least 1 h. Furthermore, NO· production in our KCl buffer system was ascertained by direct chemiluminescence measurements made with a Sievers 270B NO· analyzer (Sievers Institute, Boulder, CO) and standard protocols previously described (29).

Bronchial Smooth Muscle Strips Preparation and Isotension Measurements

Male rats (Sprague-Dawley) weighing 225-250 g were killed by exsanguination after intraperitoneal injection of sodium pentobarbital (50 mg/kg body weight). The lungs and airways were quickly removed and placed in Krebs solution. The main bronchial tissues were dissected and cut helically as previously described (30), and the epithelial cells (EC) were removed to eliminate any possible contribution of endogenous NO· or epithelial-derived hyperpolarizing factor (EDHF), both of which have been shown to modulate smooth muscle tone (31, 32). The absence of EC was confirmed histologically. Each bronchial strip was mounted in a 5-ml jacketed organ bath containing Krebs-bicarbonate solution (KBS) composed of (in mM): 118.1 NaCl, 4.7 KCl, 1.2 MgSO₄ · 7 H₂O, 1.2 KH₂PO₄, 25 NaHCO₃, 2.5 CaCl₂, and 11 glucose, pH 7.4, gassed with 95%O₂-5%CO₂ at 37°C. Contractions and relaxations were measured isometrically with a Grass Polygraph (Model 7D; Grass Instruments Co., Quincy, MA) as changes in tension. The tissues were subjected to an initial loading tension of 1 g and allowed to equilibrate for 60 min (with changes of bath medium every 15 min) before the experiments were begun (33). Following incubation in the presence of 10 µM methylene blue (MB), the bronchial strips were precontracted with 0.2 µM carbamylcholine chloride (Sigma) after which single doses or cumulative doses of SIN-1 (Biomol Research Laboratories, Inc., Plymouth, PA) were applied. The relaxation response to a given dose of SIN-1 was expressed as a percentage of the maximum tension induced by carbachol.

Preparation of Tracheal Smooth Muscle Microsomal Fractions and Channel Recording

The crude microsomal fraction was prepared as described previously (33). The reconstitution system consists of two experimental chambers, denoted cis and trans, separated by a septum with a 250-µm-diameter hole. The cis chamber was connected to an operational amplifier (Model 8900; Dagan, Minneapolis, MN) and the trans chamber was connected to virtual ground by means of two low-resistance MERE 2 electrodes (World Precision Instruments, Inc., Sarasota, FL). The hole was pretreated with a mixture of phospholipids at a concentration of 25 mg/ml chloroform in a ratio of 3:2:1 (wt/wt) of phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine (Avanti Polar Lipids, Alabaster, AL). PLB were formed from the application of the same phospholipid solution dissolved in decane. The experimental chambers contained the following solutions (in mM): 250 KCl cis/50 KCl trans, plus 20 K-Hepes and 0.01 free Ca²⁺ (109 µM CaCl₂ + 100 µM K-EGTA), pH 7.4. Microsomal fractions enriched in sarcolemmal vesicles (10-60 µg of proteins) were fused into the PLB from the cis chamber in such a way that the cis chamber corresponded to the extracellular side and the trans chamber to the cytoplasmic side of the channel. Currents generated by the reconstituted K_Ca channels were filtered (cutoff frequency, 10 kHz) and recorded on videotape (DAS/VCR 900; Toshiba, Minneapolis, MN). Currents were displayed on-line on a chart recorder (DASH II Model MT; Astro-Med, Inc., West Warwick, RI) for trace illustrations, or were played back, filtered at 1 kHz, and digitalized at 2 kHz for storage in 2-min files on a hard disk. All reconstitution studies were performed at room temperature (22 ± 2°C) (mean ± SEM). K_Ca-channel activities were analyzed in terms of current amplitudes and open-channel probability (Po). Multiple-channel activities were routinely recorded (~ 70%). Po values were determined as described previously (22, 33) and were expressed as NPo, where N is the total number of channels recorded experimentally. The number of channels functionally active within the bilayer was determined at the beginning of each recording, in the presence of 10 µM free Ca²⁺ (trans) cytoplasmic and at a holding potential (+30 mV) under which the Po of K_Ca is maximal (4, 22). K_Ca-channel activation was determined as NPo in the presence of NO· donor divided by NPo under control (low [Ca²⁺]) conditions over the same period of time.

Statistical Analysis

Results are means ± SEM (n = 3-7). Mean values were compared through paired Student's t tests, using the Sigma Plot program (Jandel Scientific Software, San Rafael, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Measurements of NO· Release from SIN-1 Degradation

To verify that SIN-1 was able to release NO· under our experimental conditions, we made specific measurements with an NO· analyzer (29). Figures 1a through 1c illustrate the detection signal for NO· production in the absence and presence of SOD. The stock solution was able to release NO· for up to 2 h after solubilization of SIN-1 and storage at 20°C (data not shown). However, the solution was always used within an hour of solubilization of SIN-1, as indicated in the first paragraph of the MATERIALS AND METHODS section. Hence, 90% of NO· was released within 10 min of addition of SIN-1 to the experimental solution at 22°C (Figures 1a through 1c). This time (10 min) correlates fairly well with the activation periods observed for the K_Ca channels in the PLB, which typically occurred within 2 to 7 min (see the following discussion). Note that in the presence of SOD, NO· was released over a longer period of time (Figure 1b).

View larger version (62K): [in this window] [in a new window]   Figure 1.   SIN-1 releases NO· in KCl buffer systems. (a) Addition of three calibrated standard amounts of 400 pmol NO (peaks 1, 2, and 3, respectively) in a 50 mM KCl, 10 µM free Ca2+, 20 mM Hepes, and 6.7 mM NaI buffer system of pH 7.4 at room temperature. NO· release from SIN-1 degradation was measured upon addition of 250 nmol of SIN-1 in absence of SOD (peak 4). Vertical and horizontal calibration bars = 0.6 mV and 3 min, respectively. (b) A single addition of 250 nmol of SIN-1 in the presence of 50 U/ml of SOD. Note the long-lasting generation of NO· over a 30-min period. (c) Repetitive calibration peaks generated by additions of 400 pmol of NO-gaz (peaks 1 and 2) and sequential additions of 250 nmol of SIN-1 under the same conditions as in (b) at various times after solubilization of SIN-1 powder to obtain a 5-mM stock solution. Note that the recorded traces in (c) were pasted from two data files separated by a lag time during which computer files were renamed and transferred.

Recording of Ca²⁺-sensitive Large Conducting K⁺ Channels

Single-channel activity was recorded in an asymmetric (50 mM trans/250 mM cis) KCl buffer system at 0 mV as a function of the free [Ca²⁺] in the cis chamber. Figure 2a shows that upon cumulative addition of EGTA in the trans chamber, the free [Ca²⁺] was decreased sequentially from 10 µM to 2 µM, 0.3 µM, and 0.1 µM. The single-channel current amplitude was unaffected by the reduction of free [Ca²⁺] in the trans chamber; rather, this resulted in a stepwise decrease of the open probability (Po). The Ca²⁺ sensitivity of the large conducting K⁺ channels (245 pS) is quantified in Figure 2b. At submicromolar free [Ca²⁺], the channel activity was very low. However, a sigmoidal relationship was obtained by plotting steady-state Po values against increasing free [Ca²⁺] in the trans chamber (n = 4). The channel was maximally activated (Po  0.7 at 0 mV) at a free [Ca²⁺] higher than 2 µM, with half-maximal activation observed at a free [Ca²⁺] of 0.40 µM. This type of conducting and gating behavior is similar to the behavior described for large conducting K_Ca channels from various tissues, including smooth muscle cells, upon patch-clamp recording in the cell-attached or inside-out configurations (34). It also corresponds to the functional fingerprint of the reconstituted K_Ca channel found in our previous studies (22, 31).

View larger version (36K): [in this window] [in a new window]   Figure 2.   Ca2+ sensitivity of the reconstituted large conducting K+ channel from airway smooth muscle. (a) Single K+ channel currents as a function of the trans free Ca2+ concentration. Current traces were obtained in asymmetric KCl buffer of 50 mM trans/250 mM cis, plus 20 mM K-Hepes and 10 µM free Ca2+ (109 µM CaCl2 + 100 µM K-EGTA, pH 7.4). The holding potential was 0 mV. Following cumulative addition of EGTA, free [Ca2+] in the trans chamber was decreased from 10 µM to 0.1 µM. The Po was reduced at submicromolar [Ca2+]. In all figures, closed arrowheads and open arrowheads indicate closed and open channel states, respectively. (b) Dependence of the KCa-channel open probability (Po) on free Ca2+ levels. Data were fitted with the Hill equation: Po = (Pomax XnH)/(Ka + XnH). Half-maximal activation was obtained at 0.4 µM [Ca2+] with nH = 2.49. Po values were calculated from analysis of amplitude histogram (mean ± SEM, n = 4).

NO·, But Not the End-Products of SIN-1, Relaxes Bronchial Smooth Muscle and Activates Reconstituted K_Ca Channels

In solution, SIN-1 generates NO· and other secondary products (25, 27). To demonstrate that NO· but no other degradation products of SIN-1 were involved in bronchial smooth-muscle (BSM) relaxation, as well as in activating reconstituted K_Ca channels, a set of control experiments was performed as shown in Figure 3. SIN-1 solutions, prepared and kept for up to a week at 22°C in order for the secondary products to accumulate, very slightly relaxed precontracted BSM strips (< 5%), even at high concentrations (> 200 µM). In contrast, 50 µM of freshly prepared SIN-1 (see the MATERIALS AND METHODS section) induced 55% relaxation (Figure 3a), indicating that NO· was the sole active substance responsible for BSM relaxation with the SIN-1 solution.

View larger version (43K): [in this window] [in a new window]   Figure 3.   Contribution of NO· in SIN-1-induced relaxation of precontracted BSM and activation of reconstituted KCa channels. (a) Inset: Neither the vehicle (V) nor 50 µM of degraded SIN-1 (O for "old") was able to relax carbachol-precontracted BSM, whereas 50 µM of freshly prepared SIN-1 (F) induced a significant increase in the channel activity. Quantification of data from three different experiments indicated that NO· was responsible for more than 90% of the tissue response. (b) Multiple-channel recordings of tracheal smooth muscle KCa channels reconstituted in planar lipid bilayers show a direct activation of the channels by NO·. The effect of SIN-1 was tested on the activity of these channels at 0 mV under our standard conditions (see MATERIALS AND METHODS). With 10 µM free Ca2+, the activity of four channels was detected from the current recording (first trace). This activity was reduced by decreasing free cytoplasmic [Ca2+] to less than 0.5 µM with 20 µM EGTA (second trace). Under these conditions, neither the vehicle (third trace) nor the SIN-1 degradation products (fourth trace) induced channel activation, whereas freshly prepared SIN-1 significantly activated the KCa channels (lowest trace). These results demonstrate that NO· is the sole component in the SIN-1 solution responsible for the activation of KCa channels.

Figure 3b illustrates a typical experiment in which the K_Ca currents were recorded in 50 mM KCl trans/250 mM KCl cis and a free [Ca²⁺] of 10 µM. At a holding potential of 0 mV, the channels were highly active (Figure 3b, first trace). Their activity was greatly diminished after reducing the free [Ca²⁺] below 0.5 µM on the cytoplasmic side of the channel (trans chamber; Figure 3b, second trace). Variations in [Ca²⁺] were achieved by addition of precalibrated amounts of EGTA, calculated according to Fabiato (35). Note that these changes in trans (cytosolic face) free Ca²⁺ concentrations also allow one to ascertain the orientation (sidedness) of the channels within the PLB. Neither addition of the vehicle (Figure 3b, third trace) nor that of 50 µM old SIN-1 solution (Figure 3b, fourth trace) modified the activity of the channels. When 50 µM of freshly prepared SIN-1 was added, all the channels were simultaneously activated, without any change in their current amplitudes or unitary conductance (Figure 3b, lower trace), indicating that NO· directly activates the K_Ca channel by means of an interaction with the channel protein.

Dose-dependent Activation of Reconstituted K_Ca Channels by NO· through a cGMP-independent Mechanism

To ascertain further the specific and direct effect of NO· on the activity of reconstituted ASM K_Ca channels within the PLB, multiple-channel activities were recorded. These channels exhibited a high degree of activity in the presence of 10 µM Ca²⁺ at 0 mV, as shown in Figure 3b. To visualize better their activation by SIN-1, the activity of K_Ca channels was decreased by reducing [Ca²⁺]_trans to 0.3 µM (Figure 4a, first trace). Upon addition of cumulative doses of SIN-1, the steady-state activities of the channels increased in a dose-dependent manner (Figure 4a, second and following traces). Quantitative analysis of the activation of K_Ca channels by NO· is reported in Figure 4b. The dose of SIN-1 that caused half of the maximal effect was estimated to be 30 µM. The maximal stimulation of K_Ca channels was obtained with 100 µM SIN-1, which corresponded to an ~ 5-fold activation. It is noteworthy that NO-induced K_Ca-channel activation never reached the initial activity recorded in the presence of 10 µM Ca²⁺, suggesting that NO· most likely activated K_Ca channels by a direct interaction. However, since NO· and Ca²⁺ have different physicochemical properties, the presence of NO-specific interaction sitesdifferent from the binding sites for Ca²⁺-induced activationwithin the K_Ca-channel proteins can be postulated.

View larger version (41K): [in this window] [in a new window]   Figure 4.   Concentration-dependent activation of KCa channels by NO· released from SIN-1. (a) Multiple-channel current recordings at different doses of SIN-1. At low (~ 0.3 µM) free cytoplasmic [Ca2+], the channel activities were relatively low (first trace). Addition of cumulative SIN-1 concentrations of 10 µM, 50 µM, and 100 µM, corresponding to the second, third, and fourth traces, respectively, resulted in a concentration-dependent increase in channel activity. (b) Fold activation of KCa channels was reported as a function of SIN-1 concentration (n = 3). The solid line shows a sigmoid fit of the data from which half-maximal activation (EC50) was estimated at 30 µM SIN-1 and maximal activation was reached at 100 µM SIN-1. The saturation effect of the SIN-1 on the activity of KCa channels supports a direct activation by NO·.

Time Analysis of NO· Activation

To test the way in which NO· modifies K_Ca-channel gating and thus increases channel Po, open- and closed-time analyses were performed on single-channel recordings. Figure 5a displays the results obtained from a typical experiment. Channel activities were sequentially recorded in the presence of high (10 µM) and low (0.3 µM) trans [Ca²⁺], following the addition of 100 µM SIN-1 (NO·), and upon addition of 5 mM dithiothreitol (DTT). The channel's open probability was reduced by lowering the trans [Ca²⁺] whereas SIN-1 (NO·) and DTT slightly reversed this effect. To quantify the effects of [Ca²⁺], NO·, and DTT on the gating behavior of the K_Ca channel, time analyses of the open and closed events were performed by computation of cumulative-time histograms and corresponding curve fitting of the various distributions, as previously done (22, 33). Data obtained from three separate experiments were compiled and are shown in Figure 5b. The values of the mean open times (Figure 5b, upper panel) show that SIN-1 (NO·) slightly increased the short time constant (_op1) when compared with the average values measured in a high or low trans [Ca²⁺]. In contrast, the unsuspected activating effect of DTT, which will be further discussed subsequently, resulted in a lengthening of the mean "short" (_op1) and "long" (_op2) open events.

View larger version (43K): [in this window] [in a new window]   Figure 5.   Effects of low [Ca2+], NO·, and DTT on single-channel mean open and closed time intervals. (a) Typical recordings of a single KCa channel under various experimental conditions. The channel activity was initially obtained at 0 mV, in asymmetric 50 mM trans/250 mM cis KCl buffer in the presence of 10 µM free Ca2+ on both sides of the PLB (upper trace). The free [Ca2+] in the trans chamber was then decreased by adding EGTA, which induced a steady-state Ca2+-dependent inactivation (second trace). Addition of 100 µM SIN-1 partly reactivated the channel (third trace). However, 5 mM DTT failed to reverse the activation induced by NO· (lowest trace). (b) Following time analysis of various data files in terms of open- and closed-time distributions, the average mean open (op1 and op2, upper panel) and closed (Cl1, and Cl2, lower panel) time constants (n = 3) are reported in bar diagrams. The prevalent effect of NO· addition was a significant decrease in Cl1 and Cl2, whereas subsequent addition of 5 mM DTT mainly resulted in an increase in op1 and op2.

The main effect of SIN-1 (NO·) consisted of a decrease in the mean duration of closed events compared with the values determined with a low trans [Ca²⁺] (Figure 5b, _Cl1 and _Cl2, lower panel). Thus, SIN-1 appeared to affect the gating behavior of K_Ca channels by reducing the mean duration of shut intervals and increasing the number of short open events, which is consistent with the observations made from multiple-channel recordings (Figures 3b and 4a). Note that the slight overall increase in Po measured following the addition of DTT was the result of an increase in mean open times. DTT had no further effect on mean closed times when added after SIN-1.

Effect of Reducing and Alkylating Agents

Because it has been proposed and shown that NO· might react with sulfhydryl groups (9), we tested the effect of NO· before and/or after addition of various reducing and alkylating agents. Table 1, rows a and b, summarizes the data obtained in various experiments in which either DTT (5 mM) or reduced glutathione (5 mM GSH) was initially used to reverse the activation induced by 100 µM SIN-1 (NO·) at a low free [Ca²⁺] in the trans chamber (intracellular side of the channels). Note that 5 mM DTT or GSH did not reverse the positive effect of NO· on the open probability of the channel. In fact, the addition of the reducing agents resulted in slight activation, which was quite the opposite of the suspected effect. Table 1, row c, shows that 5 mM N-ethylmaleimide (NEM) activated the channels in the presence of a low free [Ca²⁺] in the trans chamber, but more importantly prevented further activation by NO· following the addition of SIN-1. It is noteworthy that in all these experiments, subsequent addition to the trans chamber of an amount of Ca²⁺ that was precalibrated to yield 10 µM Ca (Table 1, rows a through c, right column) reactivated the K_Ca channels with Po values close to values determined under control conditions (Table 1, rows a through c, left column), suggesting that the site of Ca²⁺ activation of the K_Ca channels is independent of the site of NO· activation. Similar results were obtained when DTT was added in the trans chamber prior to NO· challenge, as shown in Figures 6a and 6b on a multiple-channel recording. The channels were fully activated by 10 µM Ca²⁺ (Figure 6a, upper trace) until the Po was reduced by lowering the trans free [Ca²⁺] (second trace). Adding 5 mM DTT trans partially activated K_Ca channels at a low free [Ca²⁺]_trans, but prevented further activation by NO· following the addition of 100 µM SIN-1 in the same chamber. DTT activation of K_Ca channels was observed in 11 of 13 trials. Furthermore, the DTT effect was dose-dependent between 0.3 and 10 mM (data not shown), suggesting that more than one cysteine (SH) or disulfhydryl (S-S) group are involved in channel-protein activation. However, as shown in Figure 6b, DTT effects on K_Ca channels in the presence of SIN-1 did not prevent their reactivation by Ca²⁺ (Figure 6b, far-right bar histogram), suggesting the existence of two independent activation processes, in addition to the voltage-dependent activation, which was preserved (data not shown).

View larger version (46K): [in this window] [in a new window]   Figure 6.   DTT activates KCa channels but prevents further activation by NO·. (a) Representative recordings obtained at +30 mV in the presence of 10 µM free [Ca2+] display the activity of three functional channels (first trace). The free cytosolic [Ca2+] was decreased to 0.5 µM by addition of EGTA in the trans chamber, thus reducing channel activities (second trace). Subsequent addition of 5 mM DTT in the same compartment resulted in a 2.38-fold increase in the steady-state Po of the channels (third trace). Adding 100 µM SIN-1 to the cytosolic side of the proteins had no effect in the presence of 5 mM DTT (lower trace). (b) Data quantification derived from four similar experiments involving multiple-channel recordings. NPo bars describe the average results from ongoing experiments. DTT at 5 mM (trans) significantly increased NPo (*P < 0.03), whereas SIN-1 (NO·) failed to further enhance NPo (**NS). However, complete reactivation of the channels was observed when the free [Ca2+] was restored to 10 µM, its initial value.

Sidedness of NO· Effect

Our initial results showed that the addition of SIN-1 on either side of the PLB activated K_Ca channels, an expected observation considering that NO· molecules readily diffused across the lipid bilayer. It was of interest to determine the location of the action of these NO· molecules, and the following experiments were aimed at demonstrating the sidedness of the NO· activation.

Complementary experiments were performed wherein the reducing agent was tested on the cis side of the experimental chamber prior to addition of SIN-1 on the same side (Figure 7). First, the orientation of the GK_Ca channel in the PLB was ascertained by addition of EGTA, which resulted in a decrease in the free [Ca²⁺] in the trans chamber or cytoplasmic side of the proteins (Figure 7a, upper traces). Adding 5 mM DTT in the cis chamber failed to produce any effect on the gating or conducting behavior of K_Ca channels, which contrasted with the results shown in Figures 6a and 6b, where DTT had been added to the opposite chamber. However, addition of 100 µM SIN-1 in the cis chamber partially reactivated the channels present in the bilayer. At a constant low free [Ca²⁺], this reactivation might be due to NO· molecules diffusing through the artificial membrane toward their activation site(s). Figure 7b illustrates quantitatively the behavior of the channels during such an experiment. Channel-open probability was largely decreased when the free [Ca²⁺] was reduced on the cytoplasmic (trans) side, and DTT had no effect from the cis side even if used for up to 20 min. On the other hand, SIN-1 was able to increase steady-state NPo values after a delay of 3 min. Note that gaps in NPo measurements correspond to stirring periods and delays due to current/voltage measurements under the various conditions. Following this protocol, the channels preserved their functional Ca²⁺-activation sites, since they recovered their activity upon restoration of the free [Ca²⁺] to 10 µM in the trans chamber (data not illustrated).

View larger version (32K): [in this window] [in a new window]   Figure 7.   Unlike DTT, SIN-1 can activate KCa channels from the extracellular side. (a) The activities of three channels were recorded at +30 mV in the presence of 10 µM free [Ca2+] (first trace). The usual orientation of KCa channels was ascertained by reducing the trans free [Ca2+] with EGTA (second trace). Addition of 5 mM DTT in the cis chamber did not produce any change in channel activity (third trace), which contrasts with the results shown in Figure 5. Yet under these conditions, 100 µM SIN-1 added to the cis chamber induced a partial reactivation of the channel. This effect might have been due to NO· diffusing across the bilayer prior to nitrothiosylation of specific SH groups. (b) Time-dependent behavior of KCa channels under specific conditions, as illustrated below the graph. Normalized NPo values were calculated from 1-min-long files. Note that time segments lacking data points correspond to stirring artifacts, and that I/V curves were generated during these time segments.

PAPA NONOate, an NO· Donor, Activates K_Ca Channels in the Presence of EDTA and/or K-urate

To ascertain further that NO· but not ONOO activates K_Ca channels, two sets of experiments were performed. First, the effects of PAPA NONOate were tested at low (0.1 µM) free [Ca²⁺] in the presence of EDTA, which was added to chelate traces of heavy metals, mainly Cu²⁺ and Fe²⁺, that might induce oxidation of reactive species.

Figure 8a shows that under control conditions, the channel Po was very low (< 0.01). Sequential additions of PAPA NONOate induced a concentration-dependent activation of the single channel without affecting the unit current amplitude or the unitary conductance (Figures 8c and 8d). Figure 8e shows the quantitative analysis of the effects of PAPA NONOate on multiple-channel recordings in the absence and in the presence of K-urate. K-urate was used to prevent the formation of ONOO, a highly reactive species. The large activation of the channels observed in the presence of 100-200 µM PAPA NONOate was related to the low NPo value in the control (low free [Ca²⁺]) condition. Note that similar activations were observed in the presence of 8 µM K-urate and identical PAPA NONOate concentrations. In two experiments, we also verified that the activations induced by 100 µM PAPA NONOate were not reversed by 2 mM DTT. Nevertheless K_Ca channels were fully reactivated upon addition of 10 µM free Ca²⁺ in the trans chamber, as observed with SIN-1 (see the previous section).

View larger version (28K): [in this window] [in a new window]   Figure 8.   PAPA NONOatean NO· donorinduces a concentration-dependent activation of KCa channels. (a) Single channel recording obtained at +20 mV in symmetrical 250 mM trans/250 mM cis KCl buffer and at low free [Ca2+] (~ 0.1 µM) in the presence of 100 µM EGTA + 20 µM EDTA to chelate Ca2+ and traces of heavy metal such as Cu2+ and Fe2+. (b through d) Activation of KCa channel upon addition of cumulative concentrations of PAPA NONOate of 10 µM, 50 µM, and 100 µM, respectively. Similar results were obtained in five independent experiments involving single (as illustrated previously) or, more frequently, multiple-channel recordings as quantified subsequently. (e) Normalized concentration-activation curve for PAPA NONOate in the presence of 20 µM EDTA. NPo values were calculated from amplitude histograms. Closed circles: PAPA NONOate-induced activation of KCa channels at low (0.1 µM) free [Ca2+]. The open triangles represent the activation induced by PAPA NONOate in the presence of 20 µM EDTA and 8 µM of K-urate, a peroxynitrite scavenger. Note that the addition of K-urate had no effect on the channel behavior either at high or low [Ca2+].

Altogether, these results suggest that NO· generated by the degradation of PAPA NONOate in the absence of heavy metals and/or in the presence of K-urate can activate the reconstituted K_Ca channels.

NO· Had No Effect on ASM Cl Channel Activity

To rule out the possibility that the effect of NO· generated by SIN-1 was due to nonspecific action or indirect action on the channel proteins, we tested the effect of NO· on the behavior of Cl channels, which have been shown to be present with K_Ca channels in the same membrane preparations (36). Using PLB of identical composition to those previously used to record K_Ca-channel activities, and a symmetrical CsCl buffer systemCs⁺ being a poorly permeant cation through K_Ca channels (36)we measured the activity of the unitary Cl channel as an internal control to test NO· specificity. Figure 9a shows that cumulative additions of SIN-1 in the cis chamber (up to 250 µM) did not affect the amplitude or the gating behavior of ASM Cl channels under steady-state conditions, consisting of constant voltage (40 mV), pH, and free [Ca²⁺]. In another set of experiments, three concentrations of SIN-1 were used at various holding potentials. Figure 9b shows that regardless of the applied voltage, the addition of cumulative doses of SIN-1 (NO·) did not modify the open probability (Po) of this anion-selective channel. Similarly, addition of PAPA NONOate (up to 200 µM) had no effect on the gating of Cl channels. Moreover DTT (up to 1 mM) had no effect on the conducting or gating behavior of the Cl channel (data not illustrated), thus confirming that Cl channels from the ASM membrane are insensitive to DTT, to NO·, and to byproducts of SIN-1. In contrast, K_Ca channels are rather sensitive to the direct action of oxidizing (H₂O₂) and reducing agents (37).

View larger version (44K): [in this window] [in a new window]   Figure 9.   Absence of effect of SIN-1 on the reconstituted ASM Cl channel. SIN-1 failed to affect the open probability of Cl channels (Po) when used at a concentration of up to 250 µM. (a) Upper trace: chloride current recording under control conditions (i.e., symmetrical 250 mM CsCl solutions and low [Ca2+] [0.1 µM], pH 7.4). Middle trace: addition of 150 µM SIN-1 to the cis chamber did not modify channel activity as measured in terms of Po. Lower trace: Further increase in concentration of SIN-1 to 250 µM failed to alter the channel activity measured as either the amplitude of the unitary current or the open probability (Po). All current traces were obtained at a holding potential of 40 mV. Channel openings are displayed as downward deflections from the closed-channel baseline. (b) Three-dimensional representation of the mean Po values computed at various holding potentials as a function of SIN-1 concentration (i.e., 0 µM, 50 µM, 100 µM, and 150 µM). SIN-1 had no significant effect on channel Po, regardless of the voltage applied across the lipid bilayer (P < 0.05, n = 5).

Charybdotoxin Decreased the Sensitivity of BSM to SIN-1

Three kinds of potassium channels were reported to be responsible for conducting K⁺ ions across ASM sarcolemmal membranes: delayed rectifying; ATP-sensitive; and large conducting, Ca²⁺-activated K⁺ channels (1, 2). To prove that these last channels were involved in the control of ASM tone, additional pharmacomechanical experiments were performed on BSM in the absence or in the presence of charybdotoxin (ChTX), a rather specific K_Ca-channel inhibitor (38), while inhibiting guanylate cyclase activity with MB (38). As shown in Figure 10a, addition of 100 nM ChTX 10 min before challenging the tissues diminished the tension elicited by carbachol and reduced the amplitude of the relaxation caused by 100 µM SIN-1, suggesting that the ChTX-sensitive pathway (i.e., K_Ca channel) mediates part of the effect of NO·. To assess the specificity of the effect, tension measurements were made at different doses of SIN-1. In the presence of 10 µM MB and 100 nM ChTX, the dose-response curve was displaced toward higher concentrations of SIN-1, suggesting that blockage of K_Ca channels by ChTX modified the muscle fibers' sensitivity to NO·. The estimated EC₅₀ value for SIN-1 was increased from 70 to 130 µM in the presence of ChTX (Figure 10b), thus indicating that the K_Ca channels of BSM contribute significantly to NO-induced relaxation even in the presence of MB. A possible contribution of K_ATP channels to NO-mediated relaxation, as recently reported by Murphy and Brayden (41) for mesenteric arteries, was ruled out, since 1 µM glibenclamide, a K_ATP inhibitor, had no effect on SIN-1-induced BSM relaxation (data not shown).

View larger version (16K): [in this window] [in a new window]   Figure 10.   Effect of charybdotoxin, a specific KCa-channel inhibitor, on SIN-1-induced relaxation of bronchial smooth muscle. (a) Tension measurements following the addition of a single dose of 100 µM SIN-1 before and after ChTX addition. Tissues were precontracted with 0.2 µM carbachol in the presence of 105 M MB or 105 M MB plus 100 nM ChTX. The relaxation induced by 100 µM SIN-1 was reduced after pretreatment with ChTX. W = wash-out. (b) Dose-dependent response curves (n = 5) measured in the absence (closed circles) and presence (closed squares) of ChTX. The EC50 values for SIN-1 was increased from 70 to 130 µM. These results ascertain the role of KCa channels in mediating NO-induced BSM relaxation.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we provide direct evidence that NO·, rather than the oxidizing peroxynitrite, modifies the gating behavior of K_Ca channels derived from ASM, and that this effect is prevented by reducing and alkylating agents. We also investigated the role of K_Ca channels in NO-mediated relaxation of ASM, in an attempt to provide some insight into the mechanisms implicated in this process, using PAPA NONOate or SIN-1 in the presence of SOD as sources of NO· (25).

Previous studies have suggested that different pathways may mediate the effect of endogenous NO· on vascular and nonvascular smooth-muscle tone (5, 6, 8, 9). The different hypotheses put forward are summarized in Figure 11; a cGMP-dependent phosphorylation of contractile elements and other regulatory proteins (P1), a cGMP-dependent phosphorylation of cell-surface ion channels (P2), and a direct activation of K_Ca channels by NO· (P3) appear to play a role in NO-induced relaxation. There is growing evidence that the effect of NO· on ASM tone parallels an increase in cGMP level (7, 11). The first two pathways (P1 and P2) involve the activation of soluble guanylate cyclase by NO· or its derivatives (39), causing an increase in the production of cGMP. This transient increase in cGMP would activate the cGMP-dependent protein kinase known to phosphorylate (P1) regulatory proteins of contractile elements (14), and/or (P2) would activate K_Ca channels (16, 22, 42). The third proposed pathway involves a direct chemical interaction of NO· or its highly reactive derivative peroxynitrite with the K_Ca-channel protein, which would bypass the activation of soluble guanylate cyclase. The resulting effect would be an increase in K⁺ permeability, a hyperpolarization, and a subsequent relaxation of the tissue, as demonstrated by Bolotina and colleagues (19) in VSM. Alternatively, direct chemical modification of the channel protein by peroxynitrite oxidation or nitrothiosylation, followed by oxidation of vicinal thiols, has received little attention.

View larger version (39K): [in this window] [in a new window]   Figure 11.   Schematic representation of the different pathways (P1, P2, and P3) possibly mediating NO-induced relaxation of ASM. P1 and P2 are cGMP-dependent, whereas P3 is cGMP-independent. Abbreviations: NANC = nonadrenergic, noncholinergic system; GC = guanylate cyclase; PKG = cGMP-dependent protein kinase; MB = methylene blue; ChTX = charybdotoxin; Vm = membrane potential; KCa = large conductance Ca2+- and voltage-dependent K+ channels; PDE IV = phosphodiesterase type IV; BC = bronchoconstrictor agent; R = receptor; G = G-protein; PLC = phospholipase C; InsP3 = inositol-1,4,5-trisphosphate; ER = endoplasmic reticulum; Ca2+-CaM = Ca-calmodulin complex; MLCK = myosin-light chain kinase. Symbols:  decrease,  increase,  pump.

It has been reported previously that ASM uses cGMP-dependent routes (17) that usually result in relaxation (20, 21). We now show the existence of a cGMP-independent pathway that mediates the effect of NO. Additionally, inhibition of the cGMP-dependent pathways P1 and P2 with the guanylate cyclase inhibitor MB (38) partly reduced the tissue sensitivity to NO·, as shown in Figure 10. The implication of K_Ca channels in the cGMP-independent pathway (P3) was demonstrated with the known specific K_Ca-channel inhibitors iberiotoxin (IbTX) and ChTX (11, 40). Addition of ChTX and MB before BSM were contracted with carbachol substantially reduced the tissue sensitivity to SIN-1, indicating that NO· or ONOO might activate the K_Ca channels. This observation correlates well with previously reported findings that IbTX, ChTX, and tetraethylammonium (TEA) (all K_Ca-channel blockers) consistently decreased the sensitivity to S-nitroso-N-penicilliamine (SNAP) (another NO· donor) in left pulmonary artery as well as in trachea (11). A direct effect of NO· on the K_Ca channel is shown in our reconstitution experiments (Figures 3b, 4, and 8), in which NO· induced an increase in Po, mainly by enhancing the number of channel openings. These results provide further evidence of the ways by which NO· mediates relaxation of precontracted smooth muscles, as reported by Bialecki and Fisher (11). Possible implication of K⁺_ATP channels, as reported in the rabbit mesenteric vein by Murphy and Brayden (41), was ruled out, since glibenclamide had no effect on NO-mediated relaxation of BSM.

To assess the direct effect of NO· on the activity of reconstituted K_Ca channels of ASM, bovine tracheal sarcolemma-enriched vesicles were fused into PLBs. The main advantages of this technique are its resolution and the fact that it allows easy control and manipulation of the experimental conditions on either side of the bilayer. Hence, this is one of the few systems in which one can directly study the effects of reactive species on single-channel proteins. We have previously used this technique to describe the biophysical and biochemical properties of K_Ca channels (22, 33); with an average conductance of 250 pS in 250 mM KCl cis/50 mM KCl trans, the channels show voltage-dependent open probability (22) and are sensitive to changes in free [Ca²⁺] in the trans chamber, which corresponds to the cytoplasmic side of the channels (Figures 2a and 2b). In the experiments of interest in this regard, the channel activity was first reduced by decreasing the cytoplasmic Ca²⁺ concentration from 10 µM to levels below 0.5 µM, using EGTA and/or EDTA to chelate Ca²⁺ and traces of heavy metals. Following this, cumulative doses of SIN-1 produced a concentration-dependent activation of the channels without affecting their conductance, suggesting that in the presence of SOD and a heavy-metal chelator, NO·more likely than ONOOmay react with specific sulfhydryl groups that would be involved in modulation of the gating behavior of K_Ca channels. To this effect, a time-distribution analysis of K_Ca-channel kinetics revealed that NO· donors increased overall Po by decreasing mean closed-time intervals, but had little effect on the duration of open-channel events. In contrast, DTT exerted a predominant lengthening effect on the channel's open-time intervals, leaving the closed-time distribution unaffected. Considering these observations, we suggest that NO·, generated either by PAPA NONOate or by SIN-1 + SOD, acts more potently on the K_Ca channel when the latter is in a closed state, whereas DTT is effective when interacting with the channel in its open conformation. At this point we have to stress that the activating effect of NO· on the Po for the K_Ca channel was consistently prevented but never reversed by reducing agents such as DTT or GSH. Instead, both DTT and GSH, as well as NEM, an alkylating agent, had stimulating effects on channel Po (Figures 6 and 7 and Table 1). In contrast, NEM was reported to inactivate K_Ca channels in VSM (19). This discrepancy between the effects induced by NEM in aortic and airway smooth muscles will have to be assessed under identical experimental conditions and related to the relative position of specific sulfhydryl groups.

The various NO-related species produced by NO· synthase, NO-generating drugs, or NO· donors, as well as their main biologic targets and reactants, have been clearly identified by Stamler and colleagues (43; Table 1). NO· can react with thiols, metals, and oxygen, and ONOO only with thiols. NO· can induce nitrosylation only in the presence of an electron acceptor, and this reaction is therefore highly dependent on the experimental conditions and the redox state of the environment (43).

An alternative mechanism to explain the activation by SIN-1 of K_Ca channels would be through peroxynitrite production, would the peroxynitrite in turn oxidize or nitrate a single or multiple critical residues. However, oxidation of the K_Ca channel has been shown to reduce channel Po (37, 44). Furthermore, our results in the PAPA NONOate experiments (Figure 8) argue against a peroxynitrite effect. Thus, S-nitrosylation of specific cysteine residue(s) by NO· (or nitrosonium [NO⁺]) remains a plausible mechanism for K_Ca-channel activation.

Taking into account the number (i.e., 29) of cysteine residues in the proposed sequence of the  subunit of the bovine ASM K_Ca channel, which is related to the "slowpoke" family of K⁺ channels (45, 46), it is perplexing to pinpoint a specific residue that could be susceptible to activation by NO·. However, an interaction with either cysteine 277 or with methionines 282 or 285, which are localized in the P loop of the pore-forming structure of the bovine K_Ca channel (47), is unlikely, since neither PAPA NONOate nor SIN-1 affected the channel's unitary conductance. The results of the reconstitution experiments in our study (Figures 6-8) strongly suggest that the site of NO· activation would be on the cytosolic side of the  subunit, which is known to contain numerous cysteine residues (up to 16) in its internal loops (45), most of them in the C-terminal region (47). Furthermore, 5 mM DTT, 5 mM GSH, and 5 mM NEM on the trans side of the experimental chamberall of which prevented activation by SIN-1 as well as by PAPA NONOate (NO·)did not affect the Ca²⁺ sensitivity of the K⁺ channels, as shown in Figure 7b and Table 1, suggesting that the Ca²⁺ activation sites are independent as well as different from the sites of NO· activation. It is also unlikely that NO· interacted with the  subunit of K_Ca-channel protein complex, since this subunit displays a hairpin shape, with only three cysteine residues on its single extracellular loop (48). Moreover, these residues are not flanked by the consensus sequences (K, R, D, E) C (D, E) identified by Stamler and colleagues for various proteins known to be subjected to S-nitrosylation, such as hemoglobin, cyclooxygenase, and guanylate cyclase (43).

Given these sequential requirements, a closer look at the -subunit amino-acid sequence suggests possible interference by cysteine nitrothiosylation. C975, which is a conserved residue in bovine and drosophila K_Ca channel (47), is flanked by residues that might be compatible with the mechanisms proposed herein (43). Site-specific alterations of the channel would be the only way to confirm whether this particular residue is involved in activation of the K_Ca channel by NO·.

It should be noted that peroxynitrite formation was systematically minimized under our experimental conditions. For instance, activation with PAPA NONOate was obtained in the presence of EDTA, a heavy metal chelator, as well as in the presence of K-urate, which is known to scavenge ONOO. Moreover, SOD and EGTA were present in all experiments involving SIN-1 activation of K_Ca channels. On the other hand, while assessing the effects of modifying the channel redox state on modulation of the expressed human brain K_Ca channel (hslo), DiChiara and Reinhart (37) demonstrated that DTT in the same concentration range as that used in the present work increased K_Ca-channel open probability, which basically corroborates the results reported and discussed previously (Figures 6 and 7 and Table 1) pertaining to K_Ca channels from ASM. Recenlty, DTT and -mercaptoethanol were also shown to activate maxi-K channels from equine tracheal smooth muscle cells (4). In contrast, it was recently shown that hydrogen peroxide (H₂O₂), as well as diamide, decreased the open probability of K_Ca channels by an oxidative mechanism (37, 49).

The physiologic relevance of the pathways investigated in the present study is unknown, but its independence of cGMP may underlie the need for a rapid cellular response to changes in the redox state of the environment or to very labile molecules such as NO· (12), whereas a covalent modification could result in a long-lasting effect, as reported for NANC inhibition of ASM tissue (7). Under physiopathologic conditions, the steady-state activation of K_Ca channels would lead to ASM repolarization or hyperpolarization, thus inhibiting Ca²⁺ entry via L-type, voltage-gated Ca²⁺-selective channels, a step that would facilitate ASM relaxation and would explain the observed BSM relaxation induced by NO· donors (44).

In conclusion, the endogenous release of NO· from nerve terminals via a NANC system, and/or from epithelial cells, as well as the exogenous application of NO donors, appear to activate several molecular mechanisms that synergetically induce ASM relaxation. Although caution must be taken not to extrapolate to in vivo situations, the results obtained with reconstituted proteins in artificial membranes, the present results strongly support our working hypothesis that NO·, but not peroxynitrite, may directly activate K_Ca channels. This hypothesis was further supported by the increased channel-open probability that was achieved under experimental conditions limiting peroxynitrite formation. Moreover, pharmacologic tension measurements suggest that the K_Ca channel is one of the effectors of the relaxant action of NO· on ASM strips. Hence, the modulation of ion-channel proteins by cellular redox potential has emerged as a significant determinant of channel function (37). Alternatively, we have shown that the NO· activation of K_Ca channels, which is prevented by DTT, is highly dependent on the redox state of the system, as already proposed for other ionotropic systems such as the N-methyl-D-aspartate (NMDA) receptor (49). Nevertheless, the precise chemical modification(s) of the channel proteins induced by nitrogen monoxide (NO), nitric oxide (NO·), or less likely by derivatives of NO· such as peroxynitrite, remain to be formally demonstrated.