Molecular Basis of Regulating High Voltage-Activated Calcium Channels by S-Nitrosylation*

Nitric oxide (NO) is involved in a variety of physiological processes, such as vasoregulation and neurotransmission, and has a complex role in the regulation of pain transduction and synaptic transmission. We have shown previously that NO inhibits high voltage-activated Ca2+ channels in primary sensory neurons and excitatory synaptic transmission in the spinal dorsal horn. However, the molecular mechanism involved in this inhibitory action remains unclear. In this study, we investigated the role of S-nitrosylation in the NO regulation of high voltage-activated Ca2+ channels. The NO donor S-nitroso-N-acetyl-dl-penicillamine (SNAP) rapidly reduced N-type currents when Cav2.2 was coexpressed with the Cavβ1 or Cavβ3 subunits in HEK293 cells. In contrast, SNAP only slightly inhibited P/Q-type and L-type currents reconstituted with various Cavβ subunits. SNAP caused a depolarizing shift in voltage-dependent N-type channel activation, but it had no effect on Cav2.2 protein levels on the membrane surface. The inhibitory effect of SNAP on N-type currents was blocked by the sulfhydryl-specific modifying reagent methanethiosulfonate ethylammonium. Furthermore, the consensus motifs of S-nitrosylation were much more abundant in Cav2.2 than in Cav1.2 and Cav2.1. Site-directed mutagenesis studies showed that Cys-805, Cys-930, and Cys-1045 in the II-III intracellular loop, Cys-1835 and Cys-2145 in the C terminus of Cav2.2, and Cys-346 in the Cavβ3 subunit were nitrosylation sites mediating NO sensitivity of N-type channels. Our findings demonstrate that the consensus motifs of S-nitrosylation in cytoplasmically accessible sites are critically involved in post-translational regulation of N-type Ca2+ channels by NO. S-Nitrosylation mediates the feedback regulation of N-type channels by NO.

High voltage-activated (HVA) Ca 2ϩ channels play obligatory roles in diverse physiological functions, including regulation of gene expression, synaptic transmission, and muscle contraction. These channels are heteromeric protein complexes composed of ␣ 1 , ␤, and ␣ 2 ␦ subunits (1,2). The ␣ 1 subunit (Cav␣ 1 ) contains the channel pore and is the principal component of HVA Ca 2ϩ channels. Both Cav␣ 1 and cytosolic auxiliary ␤ subunits (Cav␤) carry out essential gating functions (1,3,4). Cav␣ 1 has the most drug-binding sites of the subunits, whereas Cav␤ is essential for regulating the surface expression of the channel complex (1).
There exist several types of HVA Ca 2ϩ channels, including L-, N-, P/Q-, and R-types. L-type Ca 2ϩ channels are not only present in cardiac and smooth muscles but also expressed in neurons and endocrine cells where they regulate a multitude of processes, including the release of hormones and neurotransmitters and gene expression (5,6). N-type channels are distributed in the brain and peripheral nervous system and are the major subtypes present in nociceptive dorsal root ganglion (DRG) 2 neurons (7-10). P/Q-type channels are involved in neurotransmitter release at synaptic terminals (11,12). The Cav␤ subunits (Cav␤1-Cav␤4) are abundantly expressed in excitable tissues such as the brain, heart, and muscles. Cav␤3 is the predominant partner of Cav2.2 (N-type), and Cav␤4 is the most prevalent partner of Cav2.1 (P/Q-type) in the brain (3,13). HVA Ca 2ϩ channels are essential for the nociceptive transmission (14,15), and changes in HVA Ca 2ϩ channel properties in DRG neurons may be involved in the development of chronic neuropathic pain (16 -18).
Nitric oxide (NO) plays a complex role in the modulation of pain. Although some studies suggest that NO is pro-nociceptive (19,20), NO can also reduce pain and nociceptive transmission (21)(22)(23). Indeed, our previous study has shown that NO inhibits HVA Ca 2ϩ channel currents in DRG neurons and attenuates excitatory synaptic transmission in the spinal dorsal horn (24). However, the molecular mechanism underlying this inhibitory action is still unclear. NO exerts ubiquitous signaling via post-translational modification of cysteine residues, a reaction termed S-nitrosylation. Because nitrosothiols are exceptionally labile because of their reactivity with intracellular reducing reagents such as ascorbic acid, S-nitrosylation functions as a reversible post-translational modification analogous to phosphorylation. S-Nitrosylation is involved in regulation of NMDA receptors (25) and ryanodine receptor/Ca 2ϩ release channels (26). In DRG neurons, NO-induced HVA Ca 2ϩ channel inhibition is resistant to the guanylate cyclase inhibitor (24). However, N-ethylmaleimide, a specific alkylating agent of cysteine sulfhydryl, prevents the inhibitory effect of the NO precursor L-arginine and the NO donor S-nitroso-N-acetyl-DLpenicillamine (SNAP) on HVA Ca 2ϩ channels (24). Thus, NO likely inhibits HVA Ca 2ϩ channels through cGMP-indepen-* This work was supported, in whole or in part, by National Institutes of Health dent pathways such as the nitrosylation of free sulfhydryl groups of cysteine residues.
In this study, we determined the role of S-nitrosylation in the effects of NO on various types of reconstituted HVA Ca 2ϩ channels. We present new molecular evidence showing cysteine modification as a distinct mechanism for regulation of N-type channels by NO.
Cell Surface Protein Isolation and Cav2.2 Surface Expression-HEK293 cells transfected with V5-tagged Cav2.2, Cav␤3, and ␣ 2 ␦1 were treated with SNAP for 5 min and quickly washed with PBS. Cell surface biotinylation and surface protein isolation were carried out using the cell surface protein isolation kit (Pierce Biotechnology, Rockford, IL) according to the manufacturer's instructions. The surface protein samples were probed with the anti-V5 antibody (1:1,000 dilution; Life Technologies, Grand Island, NY) and Na ϩ /K ϩ -ATPase antibody (1:1,000 dilution; EMD Millipore, Billerica, MA) using Western blot analysis. Na ϩ /K ϩ -ATPase, a plasma membrane protein, was used as a loading control. The amounts of Cav2.2 proteins were normalized by the protein band of Na ϩ /K ϩ -ATPase. The mean value of Cav2.2 protein levels in HEK293 cells treated with the vehicle was defined as 1 (27).
Site-directed Mutagenesis-Point mutations were performed using the QuikChange Site-directed Mutagenesis Kit (Stratagene, Santa Clara, CA) or the In-Fusion HD Cloning Kit (Clontech) according to the manufacturer's instructions.
Data Analysis-The HVA Ca 2ϩ current data were analyzed using Origin 8 software (Microcal Software, Northampton, MA). Conductance-voltage (G-V) curves were fitted using the Boltzmann equation, where G is the membrane conductance, V m is the test potential, and V rev is the reversal potential. V 0.5 is the voltage for 50% activation or inactivation of HVA Ca 2ϩ currents, and k is a voltage-dependent slope factor. Western immunoblotting data were quantified by LI-COR Image Studio software (LI-COR Biosciences, Lincoln, NE). We used the Student's t test to compare the SNAP or MTSEA effects on HVA Ca 2ϩ currents between two groups. For comparing the differences in the SNAP effects on N-type channel currents between the wild-type and various mutant groups, we used the -square test with Yates correction for continuity. Values are given as mean Ϯ S.E., with n indicating the number of cells used for the electrophysiological recording or the number of independent repeats for biochemical experiments. Differences were considered statistically significant if the p value was less than 0.05.
Bath application of SNAP produced only a small inhibitory effect on P/Q-type currents in HEK293 cells coexpressing with Cav␤1 (7.95 Ϯ 2.53%, n ϭ 8) or Cav␤3 (8.72 Ϯ 1.69%, n ϭ 6) and had no significant effect on P/Q-type currents reconstituted with Cav␤2 or Cav␤4 (Fig. 2). Also, SNAP inhibited L-type currents reconstituted with Cav␤1 (15.41 Ϯ 0.87%, n ϭ 7) but had no significant effect on L-type currents reconstituted with Cav␤2, Cav␤3, or Cav␤4 (Fig. 2). Thus, compared with P/Q-type and L-type channels, N-type channels reconstituted with Cav␤1 or Cav␤3 were much more sensitive to inhibition by NO. In the following experiments, we focused our analysis on NO modulation of N-type channels.
SNAP Causes a Depolarizing Shift of Voltage-dependent Activation of N-type Channels-To examine the SNAP effect on voltage-dependent activation of N-type channels, the membrane potential was held at Ϫ90 mV. I Ba currents were elicited by a series of depolarizing pulses for 150 ms from Ϫ70 to 50 mV with an interval of 10 mV. HEK293 cells expressing Cav2.2 ϩ ␣ 2 ␦1 and Cav␤1, Cav␤3, or Cav␤4 subunits were tested. Bath application of 100 M SNAP caused a significant depolarizing shift of voltage-dependent N-type channel activation (i.e. reducing the channel sensitivity to depolarizing voltages) and slightly increased the slope factors when Cav2.2 was coexpressed with Cav␤1 or Cav␤3 (Fig. 3A, Table 1). SNAP had no significant effect on voltage-dependent N-type channel activation in HEK293 cells coexpressing Cav2.2, ␣ 2 ␦1, and Cav␤4 (Fig. 3A, Table 1).
The voltage-dependent inactivation of N-type channels was assessed by using a series of pre-pulses from Ϫ90 to 10 mV for 500 ms followed by depolarization of the cell to 0 mV for 150 ms (27). Bath application of SNAP had no significant effect on voltage-dependent inactivation of N-type channels in HEK293 cells expressing Cav2.2 ϩ ␣2␦1 and Cav␤1, Cav␤3, or Cav␤4 (Fig. 3B, Table 1).
SNAP Has No Effect on the Membrane Surface Expression of Cav2.2-Because the inhibitory effect of NO on N-type currents may be associated with reduced Cav2.2 trafficking to the plasma membrane, we next analyzed the effect of SNAP on Cav2.2 surface expression. HEK293 cells transfected with V5-tagged Cav2.2, Cav␤3, and ␣ 2 ␦1 were treated with 100 M SNAP or vehicle for 5 min. Cell surface proteins were then isolated and subjected to Western blotting analysis. Na ϩ /K ϩ -ATPase, a plasma membrane protein, was used as a loading control. The Ca2.2 surface protein levels did not differ significantly between SNAP-and vehicle-treated groups (Fig. 4).
SNAP Inhibits N-type Channels through S-Nitrosylation-The SNAP effect on HVA Ca 2ϩ channels in DRG neurons and transfected HEK293 cells is independent of cGMP pathways (24,29). To determine whether S-nitrosylation is involved in the inhibitory effect of NO on N-type currents, we used MTSEA, a membrane-permeable sulfhydryl-specific modifying reagent (25), which can rapidly react with thiols to form mixed disulfides, thereby preventing subsequent S-nitrosylation of proteins. HEK293 cells expressing Cav2.2, Cav␤3, and ␣ 2 ␦1 subunits were pretreated with 2.5 mM MTSEA for 1 min before whole cell recordings. Treatment with MTSEA alone did not significantly change the amplitude of N-type currents. MTSEA treatment completely blocked the inhibitory effect of 100 M SNAP on the amplitude of N-type channels (Fig. 5). These results suggest that NO inhibits N-type channels through S-nitrosylation. Intracellular Cysteine Residues of N-type Channels Are Involved in S-Nitrosylation Modification-Our results with the sulfhydryl-specific modifying reagent indicate that NO may interact with cysteine residues to produce its effects on N-type channels. We next used site-directed mutagenesis to identify the possible cysteine residues involved in S-nitrosylation modification of N-type channels by NO. We inspected amino acid sequences of Cav␣ 1 and Cav␤ subunits and searched for the candidate S-nitrosylation sites. According to a previous report (30), cysteine (Cys) residues followed by glutamic acid (Glu) or aspartic acid (Asp) are the critical sites (i.e. the consensus motif of S-nitrosylation) involved in NO redox reactions. In N-type channels, there are seven CD/CE pairs: two (Cys-277 and Cys-1688) in the pore loops, three (Cys-805, Cys-930, and Cys-1045) in the II-III intracellular loop, and the remaining two (Cys-1835 and Cys-2145) in the C terminus of the channel (Fig. 6A). In contrast, there is only one CD pair in P/Q-type channels (Fig.  6B), and there are three CD/CE pairs in L-type channels (Fig. 6C).
Sequence alignment analysis showed that the positions of CE pairs in Cav␤ subunits are much more conserved than those in Cav␣ 1 subunits. Cav␤1, Cav␤2, and Cav␤4 subunits have two CE pairs: one in the SH3 domain and the other in the GK domain. For the Cav␤3 subunit, only one CE pair (Cys-346) exists in the GK domain. Because the Cav␤ 3 subunit is the predominant partner of Cav2.2 in neurons (13,28), in the following experiments, we attempted to identify the critical cysteine residues in both Cav2.2 and Cav␤3 subunits involved in S-nitrosylation modifications using site-directed mutagenesis.
In contrast, mutating Cys-1065, a non-predicted cysteine residue in the II-III intracellular loop, to alanine did not significantly alter the inhibitory effect of SNAP (Fig. 7B, Table 2). Furthermore, SNAP had no significant effect on voltage-dependent activation or inactivation of N-type currents reconstituted with the C805A or C930A in HEK293 cells (Fig. 8, Table 1).
Mutating Cys-346 to alanine in the Cav␤3 subunit generated wild-type-like N-type currents in HEK293 cells coexpressing Cav2.2 and ␣ 2 ␦1. However, the inhibitory effect of SNAP on N-type currents reconstituted with the C346A mutant was reduced by about 50% compared with that in wild-type Cav␤3 (16.02 Ϯ 3.13% versus 32.09 Ϯ 3.21%; Fig. 7, Table 2). Taken together, our findings reveal that NO inhibits N-type channels through S-nitrosylation of intracellular cysteine residues in both ␣ 1 and Cav␤ subunits.

Discussion
NO is a pleiotropic cell signaling molecule that controls diverse biological processes. S-Nitrosylation is the covalent modification of a protein cysteine thiol by an NO group to generate an S-nitrosothiol, which can occur rapidly without the action of any enzymes (25,31,32). In the cardiovascular system, S-nitrosylation seems to be involved in post-translational modification of several ion channels, including Nav1.5, L-type channels, Kv1.5, and Kv4.3 (33). NO can inhibit NMDA receptor currents through S-nitrosylation in the central nervous system (34), and mutating Cys-399 on the GluN2A subunit removes NO sensitivity (25). However, it has not been clear whether S-nitrosylation is involved in regulating N-type and P/Q-type Ca 2ϩ channels by NO. In this study, we systematically determined the role of S-nitrosylation in the inhibition of HVA Ca 2ϩ currents by NO. Compared with L-type and P/Q-type channels, N-type channels coassembled with Cav␤1 or Cav␤3 subunits showed prominent sensitivity to NO. The NO donor SNAP rapidly inhibited HVA Ca 2ϩ currents, and this effect was readily reversed by DTT. Also, SNAP failed to inhibit HVA Ca 2ϩ currents in the presence of MTSEA, which covalently modifies protein sulfhydryl groups. The sulfhydral side chain (-SH) of cysteine forms a covalent disulfide bond with sulfur in MTSEA (35), and this chemical reaction can prevent subsequent S-ni-trosylation by NO. Notably, the modification by MTSEA does not affect ion channel function if the exposure time is less than 8 min (25, 29, 36). These findings demonstrate that N-type

. Effect of SNAP on voltage-dependent activation and inactivation of N-type channels.
A, voltage-dependent activation curves of N-type channels reconstituted with Cav2.2, ␣ 2 ␦1, and different Cav␤ subunits. Voltage-dependent activation curves were obtained by plotting the normalized conductance as a function of the command potential recorded. B, voltage-dependent inactivation curves of N-type channels reconstituted with Cav2.2, ␣ 2 ␦1, and different Cav␤ subunits. Voltage-dependent inactivation curves were obtained using inactivation protocols. Data points were fitted using the Boltzmann equation. V 0.5 and slope factors are shown as the mean Ϯ S.E. listed in Table 1.  For the vector-only group, HEK293 cells were cotransfected with pcDNA6 V5-FLAG vector ϩ Cav␤3 ϩ ␣ 2 ␦1 and were used as a blank control; for the SNAP group, HEK293 cells were cotransfected with Cav2.2-V5 ϩ Cav␤3 ϩ ␣ 2 ␦1 and treated with 100 M SNAP for 5 min before cell surface biotinylation. B, summary data show that SNAP had no significant effect on the surface protein levels of Cav2.2-V5 (n ϭ 4 independent experiments). The amounts of Cav2.2 proteins were normalized by Na ϩ /K ϩ -ATPase proteins on the same gel. The mean value of Cav2.2 treated with the vehicle was defined as 1. channels with Cav␤1 or Cav␤3 subunits are particularly sensitive to NO via S-nitrosylation modulation.
We found that in N-type channels co-assembled with Cav␤1 or Cav␤3 subunits, the inhibitory effect of SNAP was associated with a significant depolarizing shift of voltage-dependent activation of N-type channels. However, SNAP treatment had no significant effect on the Cav2.2 protein abundance of the membrane surface. Cav␤ subunits are critical for channel trafficking and surface expressions (1,3). The trafficking of Ca 2ϩ channels is controlled by numerous processes, including co-assembly with auxiliary subunits, ubiquitin ligases, and interactions with other membrane proteins (37). Ca 2ϩ channel trafficking typically takes more than 10 min and does not change the channel kinetic properties (38,39). Because the inhibitory effect of SNAP on HVA Ca 2ϩ channels occurred rapidly and associated with the depolarizing shift of channel activations (i.e. reduced voltage-dependent activation) in our study, NO likely inhibits N-type Ca 2ϩ channels through S-nitrosylation-induced changes in channel conformation and gating properties. Also, because SNAP failed to alter voltage-dependent activation of N-type channels when cysteine S-nitrosylation sites (i.e. C805A and C930A mutants) were mutated, our data suggest that the shift in voltage dependence likely accounts for all of the inhibitory effect of NO on N-type channels.
In this study, we also identified key structural motifs essential for the inhibitory gating of N-type Ca 2ϩ channels via cysteine S-nitrosylation. Sequence alignment analysis indicated that the consensus motifs of S-nitrosylation are much more abundant in Cav2.2 than in Cav1.2 and Cav2.1. This structural distinction is consistent with our recording data showing that SNAP had a much more pronounced effect on N-type channels than on P/Q-type and L-type channels. Thus, the prominent sensitivity to NO is a characteristic of N-type channels. We found that several intracellular cysteine residues, including Cys-805 and Cys-930 in the II-III loop and Cys-1835 and Cys-2145 in the C terminus, critically mediate the S-nitrosylation modification of the N-type channels. Mutating Cys-1045 in the II-III loop also significantly reduced NO sensitivity of N-type channels. In concord with our findings, polynitrosylation modification is known to contribute to the activation of the ryanodine receptor (40). Multiple consensus motifs of S-nitrosylation also exist in NMDA receptors (25). The cytoplasmically accessible cysteine sites may produce an additive effect and transmit cysteine modifications to functionally critical domains of N-type channels to elicit their conformational changes. In addition, site-directed mutagenesis analysis showed that mutating Cys-346 of the Cav␤3 subunit decreased the NO inhibitory effect by ϳ50% compared with the wild-type Cav␤3. These data indicate that in addition to Cav␣ 1 subunits, the Cav␤ subunit is also involved in redox modification of the N-type channel by NO. Our findings indicate that polynitrosylation modification is an important  feature in the negative feedback regulation of N-type channels by NO.
In summary, our study provides new information for our understanding of HVA Ca 2ϩ channel regulation by S-nitrosylation at the molecular level. Our findings reveal that cytoplasmically accessible cysteine residues serve as "NO sensors" of HVA Ca 2ϩ channels. This mechanism is important for the feedback regulation of Ca 2ϩ signals by NO in neurons. N-type channels are expressed at high levels in both DRG neurons and their presynaptic terminals and are critically involved in triggering the neurotransmitter release from primary afferent terminals (9,10,41). N-type Ca 2ϩ channels are also recognized as a crucial target for treating painful conditions. Both Cav2.2 and Cav␤3 subunits whose cysteine residues undergo S-nitrosylation probably conduct NO-induced inhibition of N-type channels in native DRG neurons. Thus, S-nitrosylation may play a critical role in feedback regulation of N-type channels by NO following stimulation of endogenous neuronal NO synthase in native primary sensory neurons (21,24). The S-nitrosylation sites on N-type channels could be targeted for pain treatment.

FIGURE 8. Effect of SNAP on voltage-dependent activation and inactivation of N-type currents reconstituted with Cav2.2 mutants.
A, voltage-dependent activation curves of N-type channels reconstituted with C805A or C930A, Cav␤3, and ␣ 2 ␦1 subunits. Voltage-dependent activation curves were obtained by plotting the normalized conductance as a function of the command potential. B, voltage-dependent inactivation curves of N-type currents reconstituted with C805A or C930A, Cav␤3, and ␣ 2 ␦1 subunits. Voltage-dependent inactivation curves were obtained using inactivation protocols. Data points were fitted using the Boltzmann equation. V 0.5 and slope factors are shown as the mean Ϯ S.E. listed in Table 1.