S -Nitrosation Controls Gating and Conductance of the a 1 Subunit of Class C L-type Ca 2 1 Channels*

Modulation of smooth muscle, L-type Ca 2 1 channels (class C, Ca V 1.2b) by thionitrite S -nitrosoglutathione (GSNO) was investigated in the human embryonic kid-ney 293 expression system at the level of whole-cell and single-channel currents. Extracellular administration of GSNO (2 m M ) rapidly reduced whole-cell Ba 2 1 currents through channels derived either by expression of a 1C-b or by coexpression of a 1C-b plus b 2a and a 2- d . The non-thiol nitric oxide (NO) donors 2,2-diethyl-1-nitroso-oxhydrazin (2 m M ) and 3-morpholinosydnonimine- hydrochloride (2 m M ), which elevated cellular cGMP levels to a similar extent as GSNO, failed to affect Ba 2 1 currents significantly. Intracellular administration of copper ions, which promote decomposition of the thionitrite, antagonized its inhibitory effect, and loading of cells with high concentrations of dithiothreitol (2 m M ) prevented the effect of GSNO on a 1C-b channels. Intracellular loading of cells with oxidized glutathione (2 m M ) affected neither a 1C-b channel function nor their modulation by GSNO. Analysis of single-channel behavior revealed that GSNO inhibited Ca 2 1 channels mainly by reducing open probability. The development of GSNO-induced inhibition was associated with the transient occurrence of a reduced conductance state of the channel. Our results demonstrate that GSNO modulates the a 1 subunit of smooth muscle L-type Ca 2 1 channels by an intracellular mechanism that is independent of NO release and stimulation of guanylyl cyclase. We suggest S -nitrosation of intracellularly located sulfhydryl groups as an important determinant of Ca 2 1 channel gating and conductance. DEA/NO) M HCl. sam-ples NaOH, Diethylenetriamine-pen-taacetic

The identification of nitric oxide (NO) 1 as the major endothelium-derived relaxing factor led to the discovery of a variety of NO-mediated signal transduction mechanisms (1)(2)(3)(4). NOmediated control of vascular functions involves the modulation of various ion transport systems including the high voltage-activated L-type Ca 2ϩ channels (5)(6)(7). Both increases (5) and decreases (6,7) of L-type Ca 2ϩ current in response to NO donors have been reported. Cellular regulation of Ca 2ϩ current was suggested to depend at least in part on NO-induced activation of guanylyl cyclase and subsequent modification of the phosphorylated state of channel proteins (8 -11). However, Snitrosothiols (thionitrites) were recently demonstrated to exert a potent inhibitory effect on class C, L-type Ca 2ϩ channels, which was suggested to be independent of intracellular cGMP accumulation (12). Thus, a more direct modulation of L-type Ca 2ϩ channels by NO donors appears likely. So far it is unclear whether the inhibition of Ca 2ϩ channels by thionitrites requires the release of NO and how single-channel properties are affected by this mechanism of regulation. Modulation of Ca 2ϩ channels by the S-nitrosothiol nitrosoglutathione (GSNO) appears to be of particular physiological interest, because GSNO is the most abundant endogenous thionitrite and has been suggested as a potential NO storage site or transport species (11,13,14). GSNO decomposes slowly to generate NO, a reaction which is catalyzed by traces of metal ions (14 -17). Alternatively GSNO is able to modify protein thiol residues via nitrosation or glutathionylation reactions. In the present study we investigated whether Ca 2ϩ channel inhibition by GSNO involves (i) decomposition of the thionitrite to yield NO, (ii) transnitrosation reactions, or (iii) the formation of mixed disulfides. Moreover, the changes in single Ca 2ϩ channel function associated with GSNO-induced inhibition were characterized. Our results strongly suggest intracellular S-nitrosation as a mechanism involved in physiological control of Ca 2ϩ channel function and demonstrate a unique modification of Ca 2ϩ channel function by GSNO, involving joint alteration of gating and permeability of the pore-forming ␣1C subunit.

EXPERIMENTAL PROCEDURES
Cells-␣1C-b channels 2 were stably expressed in HEK293 cells (18). In some sets of experiments, cells were transfected to express ␤2a and ␣2-␦ subunits 3,4 , in addition to the pore-forming subunit. Positively transfected cells were identified because of the expression of the human T-cytotoxic cell cluster of differentiation (CD8) as a marker gene (19). The cells were cultured at 37°C in Dulbecco's modified Eagle's medium ϩ 10% fetal calf serum and 0.25 mg/ml G418 (Sigma).
For single-channel cell-attached recordings, the whole-cell bath solution was used as pipette solution. In experiments with single ␣1C-b * This work was supported in part by the Fonds zur Förderung der Wissenschaftlichen Forschung (Spezialforschungsbereich Biomembranes F715 and P12667 to K. G., F708 to W. S., and P12728 to C. R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ To whom correspondence should be addressed. Tel.: 43-316-380-5570; Fax: 43-316-380-9890; E-mail: klaus.groschner@kfunigraz.ac.at. 1 The abbreviations used are: NO, nitric oxide; GSNO,S-nitrosoglutathione; HEK, human embryonic kidney; P o (open probability), probability of a single channel to be found in an open state; P s (availability), probability of a single channel to open during depolarization; SIN-1, 3-morpholinosydnonimine-hydrochloride; DEA/NO, 2,2-diethyl-1-nitroso-oxhydrazin; DTPA, diethylenetriamine-pentaacetic acid; DTT, dithiothreitol. channels, the pipette solution was supplemented with S-Bay-K 8644 (1 M) to increase open times to a level that allows reasonable analysis of conductance and gating. The bath solution in these experiments contained (in mM) 110 potassium aspartate, 50 Tris, 10 KCl, 3 EGTA, and 2 MgCl 2 at a pH of 7.4 adjusted with N-methyl-D-glucamine. Pipettes (2-5 megohm) were pulled from borosilicate glass (Clark Electromedical Instruments, Pangbourne, United Kingdom) and were coated with Sigmacoat (Sigma) for single-channel recordings. Experiments were performed in a 200 l bath chamber with a gravity-driven perfusion system. Currents were recorded at room temperature using an EPC-7 patch clamp amplifier (List, Darmstadt, Germany). Signals were low pass-filtered at 1 kHz and digitized with 5 kHz. Voltage clamp protocols (depolarizing pulses from Ϫ70 to ϩ30 mV or voltage ramps from Ϫ100 to ϩ80 mV/0.6 V/s; 0.2 Hz) were controlled by pClamp software using a Digidata 1200 computer interface (Axon Instruments, Foster City, CA).
Analysis of channel gating (open probability (P o ) and availability (P s )) was performed with custom-made software (20). Statistical analysis was performed using Student's t test considering a level of p Ͻ 0.05 as significant. For analysis of effects on whole-cell currents, the slow run-down of Ba 2ϩ current was accounted for by averaging the values immediately before administration and after washout of drugs to obtain the reference (control) value.
Measurement of cGMP Accumulation-HEK293 stably expressing ␣1C-b channel protein were subcultured on 24-well plates and incubated for 15 min at 37°C in a buffer containing (in mM) 110 potassium aspartate, 50 Tris, 10 KCl, 3 EGTA, and 2 MgCl 2 at a pH of 7.4 adjusted with N-methyl-D-glucamine. Cells were incubated in the presence of 1 mM 3-isobutyl-1-methylxanthine with NO donors (GSNO, SIN-1, or DEA/NO) for 4 min, and the incubation was stopped by removal of medium and addition of 0.01 M HCl. cGMP was measured in the samples by radioimmunoassay (3).
Solutions and Compounds-GSNO, as well as 8-Br-cGMP, was dissolved freshly in the bath solution. For administration of SIN-1 and DEA/NO, 100 mM stock solutions were prepared in water or 10 mM NaOH, respectively. Neocuproine (2, 9-dimethyl-1,10-phenanthroline) was directly dissolved in the pipette solution. Diethylenetriamine-pentaacetic acid (DTPA) was dissolved in 10 mM NaOH and further diluted in pipette solution. All NO donors were obtained from Alexis (San Diego, CA), and all other compounds were from Sigma.

RESULTS
Inhibition of ␣1C-b Currents by GSNO Is Independent of NO Release and Involves an Intracellular Site of Action- Fig. 1 compares the effects of GSNO, DEA/NO, and SIN-1 on wholecell Ba 2ϩ currents through ␣1C-b channels. Extracellular administration of GSNO (2 mM) rapidly inhibited the inward Ba 2ϩ currents as illustrated in Fig. 1, A and B. The current to voltage relation was not affected by GSNO (not shown). GSNO was tested at concentrations between 50 M and 2 mM, which inhibited ␣1C-b currents in the range of 13 Ϯ 5% (50 M; n ϭ 3) to 38 Ϯ 6% (2 mM; n ϭ 17). Neither DEA/NO nor SIN-1 mimicked the action of GSNO. DEA/NO (2 mM) moderately promoted the Ba 2ϩ current run-down (Fig. 1C), and SIN-1 (2 mM) failed to affect ␣1C-b currents (Fig. 1D). Similarly, channels comprised of ␣1C-b, ␤2a, and ␣2-␦, corresponding to channel complexes expressed in cardiovascular tissues (21,22), were rapidly inhibited by GSNO (n ϭ 5; see Fig. 2) but barely affected by DEA/NO or SIN-1 (data not shown). ). These measurements of cGMP levels were performed in the presence of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (1 mM). When 3-isobutyl-1-methylxanthine was omitted, significant rises in cellular cGMP were detectable only with DEA/NO (1.3-fold) but not with GSNO or SIN-1. 5 Thus the effect of the NO donors were not correlated with cGMP accumulation. To test whether intracellular decomposition of GSNO to yield NO was required for the observed Ca 2ϩ channel modulation, the action of GSNO was studied under conditions that promote decomposition of the thionitrite within the cell. Copper ions are well known to effectively promote homolytic cleavage of GSNO and were therefore considered as a suitable tool to deliberately increase the decomposition rate. reduced metal ions, in particular Cu ϩ , were reported to catalyze decomposition of GSNO (15,23,24), both CuSO 4 (50 M) and the reductant DTT (500 M) were included in the pipette solution. If decomposition of GSNO within the cell was essential to Ca 2ϩ channel inhibition, loading of the cells with copper ions would be expected to enhance inhibition, otherwise this treatment should suppress the action of the thionitrite. Fig. 4A shows that the effects of GSNO (2 mM) on whole-cell Ba 2ϩ currents were completely abolished in cells that were loaded with a pipette solution containing 50 M CuSO 4 plus DTT (500 M). To confirm that this inhibition is dependent on metal ions we performed experiments using the metal chelators neocuproine (50 M) plus DTPA (50 M) to remove catalytically active metal ions, in particular Cu ϩ . Addition of neocuproine plus DTPA recovered the inhibitory effects of GSNO to the extent observed in presence of DTT (500 M) alone (Fig. 4, B and C). Thus the intracellular presence of metal ions significantly antagonized the inhibitory action of GSNO. Nonetheless, the inhibitory effects of GSNO observed in the presence of DTT (500 M) plus metal chelators or in the presence of DTT (500 M) alone were significantly blunted as compared with control (see Figs. 1 and 2), indicating that DTT by itself interferes with the action of the thionitrite.
Inhibition of ␣1C-b Currents by GSNO Involves Nitrosation of Critical Thiol Residues-An excess of DTT is expected to compete with cellular thiols in terms of transnitrosation or oxidation and is therefore expected to inhibit thionitrite effects that are based on these reactions. Indeed, GSNO-induced inhibition of whole-cell Ba 2ϩ currents in cells dialyzed with 500 M DTT plus metal chelators was clearly less pronounced (15 Ϯ 4%; n ϭ 6) as compared with control conditions (38 Ϯ 6%; n ϭ 17). Therefore we tested whether loading of the cells with high concentrations of DTT (2 mM) by itself is sufficient to prevent the action of GSNO. Fig. 5 illustrates that GSNO barely inhibits ␣1C-b currents in cells that were dialyzed with 2 mM DTT (n ϭ 6). Prevention of GSNO effects by excess of DTT was not dependent on metal ions as metal chelators did not recover inhibitory effects of GSNO in the presence of 2 mM DTT. Inhibition of GSNO effects by intracellular DTT suggests an intracellular transnitrosation or glutathionylation reaction as the chemical basis of GSNO effects. To test whether oxidation of thiols and/or the formation of mixed disulfides can account for the effects of GSNO, we performed experiments in which the cells were loaded with 2 mM oxidized glutathione (GSSG). As shown in Fig. 5B, dialysis with GSSG failed to suppress ␣1C-b currents (n ϭ 3) and did not affect the inhibitory action of GSNO. Thus GSNO-induced modulation of ␣1C-b channels is suggested to be based on the nitrosation of critical thiol residues. 6). Note that in this recording configuration, access of bathapplied GSNO to the channels in the patch requires transport across the cell membrane. Overall channel activity in cellattached patches (mean patch current) decreased to values between 40 and 10% of control within a 2-to 10-min exposure to GSNO (2 mM), respectively. Prolonged exposure to GSNO resulted in complete loss of channel activity in most experiments. The inset in Fig. 6A shows that the control activity in cell-attached patches was fairly stable during prolonged perfusion with bath solution. Fig. 6B summarizes the inhibitory effects of GSNO observed in cell-attached experiments. A similar inhibition was observed in cells expressing ␣1Cb*␤2a*␣2-␦ as illustrated in Fig. 7. Neither DEA/NO (2 mM) nor SIN-1 (2 mM) or 8-Br-cGMP (2 mM) exerted a significant inhibitory modulation (data not shown). In an attempt to gain further information on Ca 2ϩ channel modulation by GSNO, we analyzed the changes in single-channel gating properties induced by the thionitrite. Because most of the cell-attached patches contained more than one channel, we analyzed singlechannel P o and P s , as well as the number of active channels in the patches, by use of a recently developed method (20) that allows the analysis of multichannel data. GSNO inhibited ␣1C-b, as well as ␣1C-b*␤2a*␣2-␦, channels mainly by reduction of P o (Fig. 8). In addition we observed a slight reduction of P s from 0.48 Ϯ 0.02 to 0.31 Ϯ 0.03 for ␣1C-b, which was statistically significant, and from 0.39 Ϯ 0.05 to 0.31 Ϯ 0.06 for ␣1C-b*␤2a*␣2-␦ channels. It is of note that the single-channel experiments with ␣1C-b were performed in the presence of S-Bay-K 8644 (1 M) to prolong open times and to enable a reasonable analysis of the data. A detailed inspection of singlechannel currents during the onset of the inhibitory modulation revealed that GSNO induced the transient occurrence of a reduced conductance state. This GSNO-induced subconductance level was detectable in the early phase of GSNO-induced channel modulation in 1 of 5 patches with ␣1C-b*␤2a*␣2-␦ and in 3 of 9 patches with ␣1C-b channels. A typical experiment is illustrated in Fig. 9. The GSNO-induced conductance state exhibited relatively long dwell times at a current amplitude of ϳ0.4 pA (at 40 mM Ba 2ϩ and 0 mV) corresponding to a subconductance level of ϳ50% of the full conductance state. These results provide evidence for a combined modulation of gating and permeability properties of the ␣1C-b subunit by GSNO as the basis of its inhibitory effects on L-type Ca 2ϩ channels.

DISCUSSION
The results of the present study suggest that nitrosothiols such as GSNO affect central functions of the pore-forming ␣1 subunit of class C L-type Ca 2ϩ channels, i.e. gating, as well as ion permeation, because of S-nitrosation of a critical thiol group.
Molecular Basis of L-type Ca 2ϩ Channel Modification by GSNO-Three potential NO donors (GSNO, DEA/NO, and SIN-1) were compared for their effects on whole-cell Ba 2ϩ currents in HEK293 cells stably transfected with the smooth muscle Ca 2ϩ channel ␣1C-b subunit. All three compounds are well known to increase cGMP levels in various cellular systems (6,10,11,25). Despite this principle similarity, the three compounds favor distinct chemical reactions that have been suggested to mediate (patho-) physiological effects of NO (26,27). DEA/NO releases NO, which is readily available to stimulate soluble guanylyl cyclase, whereas SIN-1 decomposes to produce both NO and superoxide (28), two species that combine rapidly to form peroxynitrite, a well known nitrating and oxidizing species. It is unlikely that peroxynitrite contributes to the effects of GSNO in our whole-cell experiments, because HEPES, which has recently been demonstrated to react with peroxynitrite to form NO donors (29), was used as a buffer system. GSNO, by contrast, not only decomposes slowly to release NO but is, in addition, able to react with protein thiol groups to form S-nitrosothiols or mixed disulfides (30). Consistent with the view that all three compounds release NO under certain conditions, they all elevated cGMP levels of HEK293 cells, with DEA/NO being most effective. In clear contrast to its superior action as a stimulator of soluble guanylyl cyclase, DEA/NO exerted rather moderate inhibitory effects on L-type Ca 2ϩ channels, indicating that cGMP does not control ␣1C-b channels in the HEK293 expression system. The notion that the observed Ca 2ϩ channel inhibition by GSNO does not involve cGMP was further corroborated by the observation that 8-Br-cGMP (2 mM) did not affect ␣1C-b channels in cellattached experiments. These results are in line with a previous study demonstrating that inhibition of L-type Ca 2ϩ channels by S-nitrosothiols is insensitive to inhibition of soluble guanylyl cyclase (12). Thus GSNO appears to exert a specific, cGMPindependent effect on L-type Ca 2ϩ channels. To test whether this effect of GSNO is based on a direct interaction of the thionitrite with intracellular protein thiol residues or requires the release of NO, we studied its effect in cells loaded with copper ions to promote decomposition of GSNO (13,14,17,23,24) or excess of DTT, which competes with endogenous thiols for S-nitrosation and oxidation (31)(32)(33). Both conditions clearly blunted or even eliminated the action of GSNO, supporting the concept that Ca 2ϩ channel modulation by GSNO involves a reaction of the thionitrite with intracellular sulfhydryls. Two types of reaction between GSNO and protein thiol residues were considered, transnitrosation and glutathionylation. Our observation that even excessive loading of cells with GSSG failed to affect channel function, which argued against glutathionylation as a mechanism of channel modulation. This conclusion is further supported by our previous experiments with the lipophilic oxidant tert-butylhydroperoxide, which is known to induce oxidation of cellular thiols and formation of mixed disulfides but failed to affect ␣1C-b channels in the HEK293 expression system. 5 Thus we suggest nitrosation of critical protein thiol groups as a mechanism of L-type Ca 2ϩ channel regulation. The modest inhibitory effects observed with the NO donor DEA/NO may, therefore, be because of nitrosation of thiol groups, a reaction that was shown to occur at high concentrations of NO in the presence of oxygen (28).
It is tempting to speculate about the modification of a critical cysteine residue of the Ca 2ϩ channel ␣1 subunit itself as the regulatory principle mediating the effects of GSNO. Although modification of a yet unknown regulatory protein cannot be excluded, our experiments demonstrate that Ca 2ϩ channel regulation by GSNO does not require expression of any of the known auxiliary subunits of the channel. Thus the ␣1C protein appears to be the most likely target for GSNO modulation. An intracellular localization of the regulatory thiol residue is suggested by the following lines of evidence: (i) the effects of GSNO were sensitive to intracellular administration of copper ions and DTT, and (ii) GSNO inhibited L-type Ca 2ϩ channels not only in the whole-cell but also in the cell-attached configuration of the patch clamp technique. A truncated form of ␣1C that lacks the cytoplasmic C-terminal 438 amino acids including 13 cysteine residues was also sensitive to inhibition by GSNO. 5 The sequence of this truncated ␣1C includes 11 cysteine residues that are potentially accessible from the cytoplasmic side and may be considered as candidates for a regulation by transnitrosation.
Protein Thiol Nitrosation as a Unique Regulatory Mechanism That Affects both Gating and Conductance of L-type Ca 2ϩ Channels-Our analysis of the inhibitory effect of GSNO on single-channel parameters revealed that the reduction of whole-cell currents was mainly because of suppression of single-channel open probability. In addition we observed a slightly reduced channel availability and number of active channels in the patches already within 5-10 min of exposure to GSNO, and channel activity was completely lost during prolonged (Ͼ10 min) exposure. A detailed inspection of the single-channel behavior during the development of GSNO inhibition uncovered a striking phenomenon. A defined and reproducible subconductance level of about 50% of the full unitary conductance was recorded in about 30% of the experiments. This conductance state was observed only transiently during the development of the GSNO-induced inhibition of channel activity. Thus we report here on a subconductance state of ␣1C-b channels that is induced by a potential (patho-) physiologic regulator of L-type Ca 2ϩ channels. The transient nature of the GSNO-induced subconductance state precluded a more detailed analysis of this phenomenon. Nonetheless, our results strongly suggest that GSNO exerts a modulatory effect on both gating and permeation in the ␣1C-b protein. The observed GSNO-induced modulation of gating and conductance was independent of the expression of auxiliary ␤2a and ␣2-␦ subunits and may well explain inhibitory effects of nitrosothiols on native cardiovascular Ca 2ϩ channels (5,34,35).
In summary, we provide evidence for inhibitory modulation by GSNO of the principle pore-forming subunit of smooth muscle L-type Ca 2ϩ channels by intracellular S-nitrosation. GSNOmediated regulation of Ca 2ϩ channels involves unique changes in the functional properties of the pore-forming ␣1 subunit and may represent an important mechanism that links NO signaling to Ca 2ϩ channel functions in the cardiovascular system.