Mechanisms underlying the modulation of L-type Ca²⁺ channel by hydrogen peroxide in guinea pig ventricular myocytes

Abstract

Although Cav1.2 Ca²⁺ channels are modulated by reactive oxygen species (ROS), the underlying mechanisms are not fully understood. In this study, we investigated effects of hydrogen peroxide (H₂O₂) on the Ca²⁺ channel using a patch-clamp technique in guinea pig ventricular myocytes. Externally applied H₂O₂ (1 mM) increased Ca²⁺ channel activity in the cell-attached mode. A specific inhibitor of Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) KN-93 (10 μM) partially attenuated the H₂O₂-mediated facilitation of the channel, suggesting both CaMKII-dependent and -independent pathways. However, in the inside-out mode, 1 mM H₂O₂ increased channel activity in a KN-93-resistant manner. Since H₂O₂-pretreated calmodulin did not reproduce the H₂O₂ effect, the target of H₂O₂ was presumably assigned to the Ca²⁺ channel itself. A thiol-specific oxidizing agent mimicked and occluded the H₂O₂ effect. These results suggest that H₂O₂ facilitates the Ca²⁺ channel through oxidation of cysteine residue(s) in the channel as well as the CaMKII-dependent pathway.

Introduction

An alteration in the cell’s redox state such as increased reactive oxygen species (ROS) production is associated with pathology [1–3]. In the heart, ROS as highly reactive compounds accumulate in tissues during myocardial ischemia/reperfusion and cause peroxidation of lipids and proteins [4] which play an important role in the pathogenesis of ischemia/reperfusion abnormalities, including myocardial stunning, irreversible injury, and reperfusion arrhythmias [5]. ROS-induced Ca²⁺ overload is one of the major causes of cardiomyocytes injury during ischemia/reperfusion [6]. Ca²⁺ overload induced by oxidation is thought to be mediated by increased Ca²⁺ release from the sarcoplasmic reticulum (SR) through the Ca²⁺ release channel (ryanodine receptor 2, RyR2) and decreased Ca²⁺ uptake by inhibiting Ca²⁺-ATPase (SERCA) activity [7].

L-type (Cav1.2) Ca²⁺ channels (LTCCs) in the myocardium sarcolemma are the main route for Ca²⁺ influx into cells. Different from skeletal muscle, in cardiac myocytes, Ca²⁺ influx through LTCCs triggers Ca²⁺ release, thus determining the Ca²⁺ dynamics in the cardiac myocytes. Accumulating evidence shows that basal activity of LTCCs is modulated by cytoplasmic factors including protein kinase-mediated phosphorylation [8, 9], phosphatase-mediated dephosphorylation [8, 9], and the interaction with Ca²⁺ and Mg²⁺ [10], lipids [11] and proteins [12, 13]. Recent studies suggest that the function of LTCC is crucially modulated by ROS during ischemia/reperfusion [14, 15]. Exposure of myocytes to high concentration of H₂O₂ results in alteration of Ca²⁺ channel activity and cellular Ca²⁺ homeostasis [16]. However, the effects of oxidation are so far controversial, since both inhibition and facilitation of LTCCs by oxidation have been suggested [17–20]. The mechanism of the ROS effect on LTCCs also remains elusive. For example, H₂O₂ has been suggested to facilitate LTCCs by activation of Ca²⁺/CaM-dependent protein kinase II (CaMKII) through oxidation of methionin residues in CaMKII or increasing Ca²⁺ release through RyR (20). On the other hand, Tang et al. [19] have reported that H₂O₂-induced facilitation of LTCCs is mediated by glutathionylation of LTCCs.

To explore the effect of oxidation on cardiac LTCCs and the underlying mechanisms, the inside-out mode of the patch-clamp technique is beneficial since the internal side of LTCC can be well controlled. We have previously found that LTCC activity is maintained with calmodulin (CaM) and ATP without run-down of the channel in the inside-out patches [12, 21–26]. In this study, using this method, we have investigated the effect of H₂O₂ on the current through LTCCs. H₂O₂ was found to increase Ca²⁺ channel activity in both the cell-attached mode and the inside-out mode. The specific CaMKII inhibitor KN-93 partially attenuates the facilitation effect in the cell-attached mode, but had no effect in the inside-out mode. Our results suggest that H₂O₂ facilitates LTCCs through CaMKII-dependent and -independent pathways.

Materials and methods

Materials

MgATP, hydrogen peroxide (H₂O₂) tablet, and 5,5′-dithiobis (2-nitrobenzoic acid) (DNTB) (cysteine residue oxidation reagent) were purchased from Sigma-Aldrich (St. Louis, MO, USA), KN-93 (CaMKII inhibitor) and KN-92 (inactive analog of KN-93) were from Calbiochem (San Diego, CA, USA), and Bay K8644 (Ca²⁺ channel agonist) was from Wako (Osaka, Japan).

Preparation of single cardiac myocytes

Single ventricular myocytes were isolated from adult guinea pig hearts by collagenase dissociation as described previously [27]. In brief, a female guinea pig (weight 300–500 g) was anesthetized with pentobarbital sodium (30 mg/kg, i.p.), and the aorta was cannulated in situ under artificial respiration. The dissected heart was mounted on a Langendorff apparatus and perfused with Tyrode solution for 3 min, then with nominally Ca²⁺-free Tyrode solution for 5 min, and finally with Ca²⁺-free Tyrode solution containing collagenase (0.08 mg/ml; Yakult) for 7–15 min. The collagenase was washed away with a high K⁺, low Ca²⁺ solution (storage solution). The single ventricular myocytes were dispersed and filtered through a stainless steel mesh (105 μm). Then, 0.05 mg/ml protease (Type XIV, Sigma) and 0.02 mg/ml DNase I (Type IV, Sigma) were incubated with the myocytes to improve the success rate in attaining a gigaohm seal. The enzyme-treated myocytes were then washed twice by centrifugation (800 rpm for 3 min) and stored at 4 °C.

The experiments were carried out under the approval of the Committee of Animal Experimental, Kagoshima University.

Solutions

The tyrode solution contained (in mM) 135 NaCl, 5.4 KCl, 0.33 NaH₂PO₄, 1.0 MgCl₂, 5.5 glucose, 1.8 CaCl₂, and 10 HEPES–NaOH buffer (pH 7.4). The storage solution was composed of (in mM) 70 KOH, 50 glutamic acid, 40 KCl, 20 KH₂PO₄, 20 taurine, 3 MgCl₂, 10 glucose, 10 HEPES, and 0.5 EGTA; pH was adjusted to 7.4 with KOH. The pipette solution contained (in mM) 50 BaCl₂, 70 tetraethylammonium chloride, 0.5 EGTA, 0.003 BAY K8644, and 10 HEPES-CsOH buffer (pH 7.4). The basic internal solution consisted of (in mM) 90 potassium aspartate, 30 KCl, 10 KH₂PO₄, 1 EGTA, 0.5 MgCl₂, 0.5 CaCl₂, and 10 HEPES–KOH buffer (pH 7.4; free Ca²⁺ 80 nM, pCa 7.1). CaM and MgATP were dissolved in basic internal solution in the inside-out patch mode.

Preparation of CaM

The cDNA of human CaM cloned into the pGEX6P-3 vector (GE Healthcare Bioscience, Uppsala, Sweden) was expressed as glutathione-S-transferase (GST) fusion protein in Escherichia coli BL21 and purified using glutathione–Sepharose 4B (GE Healthcare). The GST region was cleaved by PreScission Protease (GE Healthcare). The purity of CaM was confirmed by SDS-PAGE and quantified by the Bradford method (Thermo Fisher Scientific, Rockford, IL, USA) with bovine serum albumin as the standard and a correction factor of 1.69.

Patch clamp and data analysis

Barium current through LTCCs was recorded in the cell-attached and inside-out mode using the patch-clamp technique. For recording in the cell-attached mode, the myocytes were perfused with the basic internal solution at 31–35 °C using a patch pipette (2–4 MΩ) containing 50 mM Ba²⁺ and 3 μM Bay K8644. Bay K8644, a Ca²⁺ channel agonist, was used to prolong the open time of the channel to facilitate the experiments. Barium currents through LTCCs were elicited by depolarizing pulses from −70 to 0 mV for 200 ms duration at a rate of 0.5 Hz. The current were recorded with a patch-clamp amplifier (EPC-7; List, Darmstadt, Germany) and fed to a computer at a sampling rate of 3.3 kHz. The mean current (I) was measured and divided by the unitary current amplitude (i) to yield NPo, based on the equation I = N × Po × i, where N is the number of channels in the patch and Po is the time-averaged open-state probability of the channels. Data are presented as the mean ± S.E. Student’s t test or ANOVA with post hoc Tukey HSD test was used to estimate statistical significance and values of P < 0.05 were considered as significant.

Results

H₂O₂ facilitated Ca²⁺ channel in the cell-attached mode via CaMKII-dependent and independent pathways

We first examined the effect of H₂O₂ on the current through LTCCs in the cell-attached mode in guinea pig ventricular myocytes. After recording the current for 2 min as a control, 1 mM H₂O₂ was applied in the perfusion solution. As shown in Fig. 1a, b, Ca²⁺ channel activity was rapidly increased without a change in the unitary current amplitude. In an average of six patches, Ca²⁺ channel activity was increased to 206 ± 32 % of the control (Fig. 1d). This result confirmed the facilitating effect of H₂O₂ on LTCCs.

Fig. 1

H₂O₂ facilitates L-type Ca²⁺ channel activity in cell-attached mode. a Time course of channel activity (NPo) recorded in the cell-attached mode before and after application of 1 mM H₂O₂. b Examples of current traces of the Ca²⁺ channels before (a) and after (b) application of H₂O₂ taken at the times indicated in (a). c Effect of 1 mM H₂O₂ on channel activity (NPo) recorded in cell-attached mode in the presence of 10 μM KN-93. d Summary of the normalized activity of the Ca²⁺ channel treated with H₂O₂ with no drug (n = 6), and with 10 μM KN-93 (a specific CaMKII inhibitor) (n = 9) or KN-92 (an inactive form of KN-93) (n = 5). Mean channel activity (60 traces) in each patch was normalized to the corresponding control value, averaged in the same group, and shown as mean ± SE. *P < 0.05 and **P < 0.01 versus control (Student’s t test), and ^# P < 0.05 and NS not significant versus H₂O₂ without drug (ANOVA and Tukey HSD test)

Facilitation of LTCCs by glutathionylation [19] or phosphorylation mediated by activated CaMKII [20] during oxidative stress has been proposed. To evaluate the possible CaMKII-dependent effect of H₂O₂, we incubated the cardiomyocytes with 10 μM KN-93, a specific CaMKII inhibitor, for 10 min before recording the current. Under this condition, it has been reported that the activity of CaMKII is nearly completely inhibited by KN-93 [28, 29]. As shown in Fig. 1c, KN-93 significantly attenuated H₂O₂-mediated facilitation [132 ± 15 % (n = 9) vs. 206 ± 32 % with no drug, P < 0.05], while KN-92, an inactive analog of KN-93, partially attenuated the facilitation but it was not statistically significant [156 ± 13 % (n = 5) vs. 206 ± 32 % with no drug, P = 0.30). These results suggested that not only CaMKII-dependent but also independent pathways were involved in H₂O₂-mediated facilitation of LTCCs.

H₂O₂ facilitated Ca²⁺ channel in the inside-out mode independently of CaMKII

To explore the mechanism of CaMKII-independent facilitation of LTCCs produced by H₂O₂, we investigated the H₂O₂ effect on LTCCs in the inside-out patches in which the Ca²⁺ channel activity was maintained by application of 1 μM CaM together with 3 mM ATP [12, 21–26]. In the inside-out patch mode, LTCCs were disconnected with cytoplasmic factors and perfused with an artificial solution with known composition. This was quite beneficial to examine direct effects of external reagents on LTCC. After the patch was excised and moved to a small inlet in the perfusion chamber, which was connected to a microinjection system, CaM/ATP was immediately applied to induce Ca²⁺ channel activity, the single channel current was recorded for 3 min as a control current, then 1 mM H₂O₂ was added to the CaM/ATP solution. As shown in Fig. 2a, H₂O₂ significantly increased the CaM-induced Ca²⁺ channel activity in the inside-out mode. This facilitatory effect of H₂O₂ was concentration-dependent up to 1–2 mM, and higher concentrations of H₂O₂ conversely inhibited Ca²⁺ channel activity presumably due to a non-specific deteriorating effect of H₂O₂ (Fig. 2b). These results suggested that H₂O₂ (<2 mM) facilitated Ca²⁺ channel activity in the inside-out patches via a direct modification of LTCC and/or its closely-located proteins such as CaM and CaMKII. To assess the possibility that CaMKII was still located near the channel and modulated channel activity in the excised patches, we examined the effect of KN-93 in the inside-out patches. As shown in Fig. 2c, KN-93 had only a small effect on the H₂O₂-mediated facilitation of LTCCs in the inside-out mode. In summary, channel activity was modulated by H₂O₂ from 144 ± 32 % (control) to 272 ± 70 % in the absence of KN-93 (n = 5), whereas the change was from 139 ± 21 to 231 ± 29 % in the presence of KN-93 (n = 6). Although KN-93 seemed to slightly attenuate the increasing effect of H₂O₂ on channel activity, this difference was statistically insignificant. Thus, KN93 did not significantly affect the H₂O₂-mediated facilitation of the Ca²⁺ channel in the inside-out mode.

Fig. 2

H₂O₂-mediated facilitation of Ca²⁺ channel activity in inside-out mode. a, c Time course of channel activity (NPo) recorded in the cell-attached (c.a.) mode followed by the inside-out (i.o.) mode, in which channel activity was maintained with 1 μM CaM + 3 mM ATP, and then 1 mM H₂O₂ without (a) or with (c); 10 μM KN-93 was additionally applied as indicated by the boxes in each graph. ATP was included throughout the experiments in the i.o mode. b Concentration-dependent effect of H₂O₂. Normalized channel activity was plotted against concentration of H₂O₂. Data (n = 4–6) were fitted with a combined Hill’s equation as: $(100 + A \cdot [{\text{H}}_{2} {\text{O}}_{2} ]^{{n_{\text{f}} }}/({\text{Kd}}_{\text{f}}^{{n_{\text{f}} }} + [{\text{H}}_{2} {\text{O}}_{2} ]^{{n_{\text{f}} }} )) \cdot ({\text{Kd}}_{\text{i}}^{{n_{\text{i}} }} /({\text{Kd}}_{\text{i}}^{{n_{\text{i}} }} + [{\text{H}}_{2} {\text{O}}_{2} ]^{{n_{\text{i}} }} ))$, where [H₂O₂] was the concentration of H₂O₂, A the extent of facilitation, Kd_f and Kd_i apparent dissociation constants, and n _f and n _i Hill’s numbers for facilitation and inhibition, respectively. The fitted curve was drawn with A = 147, Kd_f = 0.68 mM, n _f = 2, Kd_i = 3.11 mM, and n _i = 3 (r ² = 0.983). d Summary of the normalized channel activity in the presence of CaM + ATP in i.o. mode before and after addition of H₂O₂ ± KN-93 (n = 5–6). Data are shown as mean ± SE. *P < 0.05 versus CaM (t test), NS not significant (ANOVA)

H₂O₂ may also be able to oxidize CaM and modulate the effect of CaM on Ca²⁺ channel activity. To assess this possibility, we examined the effect of oxidized CaM pretreated with 1 mM H₂O₂ at room temperature for 30 min. As shown in Fig. 3a, b, after Ca²⁺ channel activity was maintained by intact (untreated with H₂O₂) CaM + ATP in the inside-out patches, we substituted the oxidized CaM for the untreated CaM. Channel activity did not change significantly, suggesting that CaM was not oxidized or that oxidized CaM, if any, had similar potency as the untreated CaM on activity of LTCC (Fig. 3c). This result suggested that oxidation of CaM was not involved in the H₂O₂-mediated facilitation of LTCCs, and thus a direct oxidation of LTCCs might be a possible mechanism for the facilitation in the inside-out mode.

Fig. 3

Facilitation of Ca²⁺ channel by H₂O₂ is not due to oxidation of CaM. CaM was pre-incubated with 1 mM H₂O₂ for 30 min and then applied to the Ca²⁺ channels. a Time course of channel activity recorded first in the cell-attached (c.a.) mode followed by the inside-out (i.o.) mode with 1 μM non-treated (intact) CaM + 3 mM ATP, followed by substitution with H₂O₂-treated CaM (1 μM). b Example of current traces for the control in c.a. mode (a), with CaM + ATP in i.o mode (b), and H₂O₂-treated CaM + ATP (c), at the time period indicated in (a). c Summary of normalized channel activity induced by CaM + ATP (n = 7) and H₂O₂-treated CaM + ATP (n = 6). Data are shown as mean ± SE. ***P < 0.001 versus control (c.a.), NS not significant (ANOVA and Tukey HSD test)

Cysteine residues in Ca²⁺ channel are involved in H₂O₂-mediated facilitation

The α1C subunit of LTCC contains 38 cysteine residues and 36 methionine residues in the cytoplasmic chains, which are potentially subject to oxidation modification. To identify the amino acid residue which was oxidized by H₂O₂ and responsible for the H₂O₂-mediated facilitation of LTCCs, a specific oxidizing agent of cysteine residues 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB) was applied in place of H₂O₂. As shown in Fig. 4, 1 mM DNTB significantly increased Ca²⁺ channel activity maintained by CaM (from 129 ± 22 up to 184 ± 21 %, n = 5), suggesting that oxidation of cysteine residue(s) was responsible, at least partially, for the H₂O₂-mediated facilitation of LTCCs. Since there is no specific oxidizing agent of methionine residues available, we examined the effects of H₂O₂ on LTCC after application of DTNB. Application of H₂O₂ + CaM after DTNB + CaM only slightly increased Ca²⁺ channel activity and was statistically insignificant (Fig. 5). These results suggested that oxidation of cysteine residue(s) was the major cause of the H₂O₂-mediated facilitation of LTCCs in the inside-out patches.

Fig. 4

DTNB, a specific oxidant of cysteine residues, facilitates Ca²⁺ channel activity in inside-out mode. a Time course of channel activity recorded first in the cell-attached (c.a.) mode followed by the inside-out (i.o.) mode. After patch excision, 1 μM CaM + 3 mM ATP was applied to maintain channel activity, and then DTNB (1 mM) was additionally applied. b Example of current traces for control in c.a. mode (a), with CaM + ATP in i.o mode (b), and with CaM + ATP + 1 mM DTNB (c) at the time period indicated in (a). c Summary of normalized channel activity in the presence of CaM + ATP and CaM + ATP + DTNB (n = 5). Data are shown as mean ± SE. *P < 0.05 versus control (t test)

Fig. 5

a Time course of channel activity recorded first in the cell-attached (c.a.) mode followed by the inside-out (i.o.) mode. After patch excision, 1 μM CaM + 3 mM ATP was applied to maintain channel activity, then DTNB (1 mM) was added, and finally H₂O₂ (1 mM) was additionally applied. b Summary of channel activity induced by CaM + ATP + DTNB (n = 5) and CaM + ATP + H₂O₂ (n = 6), normalized to the activity values obtained for conditions of CaM + ATP. Data are shown as mean ± SE. NS not significant (t test)

Discussion

In the present study, we investigated the effect of hydrogen peroxide (H₂O₂) on the L-type Ca²⁺ channel (LTCC) of cardiac myocytes in the cell-attached and the inside-out mode. We found that H₂O₂ facilitates cardiac LTCCs through two pathways: (1) direct oxidation of cysteine residue(s) of the channel, and (2) indirect pathways via activation of CaMKII.

Changes of the redox state in the cardiac myocytes play an important role in heart diseases. LTCC as a major regulator of cardiac function is subject to redox modification. Although there is accumulating evidence supporting that ROS modulate the function of LTCCs, the results of these studies are somehow controversial. Oxidizing agents are reported to inhibit the human and rabbit cardiac LTCC expressed in HEK293 cells [17, 18]. In isolated guinea pig ventricular myocytes, oxidation decreased the current through LTCC [30]. On the other hand, a decrease in cellular superoxide and H₂O₂ is associated with a decrease in the current of native [31, 32] and expressed LTCCs [17, 33], while the thiol-specific oxidizing agent, DNTB, increases the LTCC current [34]. The complicated interactions between LTCC and uncontrolled cytoplasmic factors may be partially responsible for the uncertainty surrounding the effects of oxidation on LTCCs. The present study took advantage of the inside-out mode of patch-clamp technique in which most of the cytoplasmic factors were washed out [12, 21–26]. Our results show that H₂O₂ facilitates LTCC at concentrations up to 2 mM and inhibits the channel at higher concentrations. This finding may partly account for the diverse results in the previous studies.

The underlying mechanisms of modulation of the Ca²⁺ channel by H₂O₂ are so far not completely clear. Song et al. [20] have reported H₂O₂-mediated facilitation of the Ca²⁺ channel through activation of CaMKII which can be activated by either Ca²⁺/CaM or oxidation of methionine residues in CaMKII. Thus, H₂O₂ is involved in both activation processes: (1) H₂O₂ enhances Ca²⁺ release from SR by increasing RYR activity; (2) oxidation of methionine residues (281/282 in mouse) in CaMKII protein sustains the kinase activity independent of Ca²⁺/CaM [3]. However, Tang et al. [19] suggest that H₂O₂ facilitates the Ca²⁺ channel through direct glutathionylation of the channel protein. It is difficult to distinguish the direct effect of H₂O₂ from an indirect one when examining whole-cells. In the present study, we employed a method to record LTCC current in the inside-out mode, in which channel activity was maintained by CaM/ATP in the internal solution [12, 21–26]. Our results show that the CaMKII-specific inhibitor, KN-93, did not completely attenuate H₂O₂-mediated facilitation of LTCC in the cell-attached mode, suggesting that H₂O₂-mediated facilitation is mediated not only by a CaMKII-dependent pathway but also by CaMKII-independent pathways. The finding that KN-93 does not inhibit the H₂O₂ effect in the inside-out mode indicates that the CaMKII-mediated facilitation is absent in the inside-out mode, and hence implies that CaMKII might not be attached to LTCC or, if present, it might not be in a state sensitive to oxidation in our inside-out patches.

Thus, our results indicate that, in addition to the CaMKII-dependent pathway, there is an additional CaMKII-independent pathway for the H₂O₂-mediated facilitation of LTCC. Since most intracellular proteins are washed out in the inside-out patches, it is likely that direct oxidation of LTCC or its associated proteins by H₂O₂ might be involved in the facilitation of LTCC. Since H₂O₂-pretreated CaM does not mimic the facilitatory effect of H₂O₂, oxidation of CaM does not account for the mechanism of facilitation. This is consistent with the fact that human CaM does not contain any cysteine residues. Thus, it seems most likely that the Ca²⁺ channel protein itself undergoes direct oxidation by H₂O₂ as the CaMII-independent pathway of the LTCC facilitation.

Both cysteine and methionine residues are subject to oxidation by H₂O₂. The pore-forming subunit α1C of cardiac LTCC is rich in cysteine and methionine residues in the cytoplasmic chains [35]. Our findings that the specific cysteine oxidizing agent DTNB mimics the H₂O₂ effect and that the effect of subsequently applied H₂O₂ is largely occluded suggest that cysteine residue(s) are involved in the H₂O₂-mediated facilitation. However, this does not exclude a possible involvement of methionine residue(s). Future work should focus on determining the oxidation sites responsible for the H₂O₂-mediated modulation of LTCC. In conclusion, LTCC may undergo ROS-mediated modification via the direct oxidation of LTCC as well as the indirect pathways involving CaMKII activation. This would be relevant for the understanding of ROS-mediated regulation of ion channels and Ca²⁺ overload and arrhythmogenesis during oxidation stress on the heart.