Department of Pharmacology, Centre of Biomedical Research Excellence (COBRE), University of Nevada School of Medicine, Reno, Nevada, USA
Corresponding author N. Leblanc: Department of Pharmacology/Mail Stop 318, Center of Biomedical Research Excellence (COBRE), Manville Sciences Building, University of Nevada School of Medicine, Reno, Nevada 89557-0270, USA. Email: firstname.lastname@example.org
The role of the Ca2+-dependent phosphatase calcineurin (CaN) in the modulation of Ca2+-dependent Cl- channels (ClCa) was studied in freshly isolated rabbit coronary arterial myocytes. Immunocytochemical experiments showed that calmodulin-dependent protein kinase II (CaMKII) and CaN were distributed evenly throughout the cytoplasm of coronary myocytes at rest and translocated to the plasmalemma when intracellular Ca2+ was increased. ClCa currents (ICl(Ca)) elicited by cell dialysis with fixed intracellular Ca2+ levels up to 500 nm were inhibited by 10 μm cyclosporin A (CsA), a specific inhibitor of CaN, in a voltage-dependent manner, whereas currents evoked by 1 μm Ca2+ were not affected. Inhibition of CaN with CsA also led to a significant reduction in Ca2+ sensitivity of the channel at +50 mV; half-maximal activation increased from 363 ± 16 nm Ca2+ in control to 515 ± 40 nm Ca2+ in the presence of CsA. Similar effects were observed on ICl(Ca) when a specific peptide fragment inhibitor of CaN (CaN-AF, 5 μm) was dialysed into the cell via the pipette (500 nm Ca2+). Application of KN-93 (10 μm), a specific inhibitor of CaMKII, enhanced ICl(Ca) in myocytes dialysed with 1 μm Ca2+ but produced no significant effect on this current when the cells were dialysed with 350 or 500 nm Ca2+. These results are consistent with the notion that in coronary arterial cells, the activity of ClCa is enhanced by dephosphorylation of the channel or a closely associated regulatory protein. Moreover the balance of CaN and CaMKII regulating ICl(Ca) is dependent on the level of Ca2+ used to activate ICl(Ca).
Ca2+-activated chloride channels (ClCa) are widely expressed in smooth muscle cells and represent a major excitatory mechanism (Large & Wang, 1996). In most studies a membrane current generated by the opening of these channels (ICl(Ca)) is elicited either as a consequence of Ca2+ influx or release from intracellular Ca2+ stores. However, we have recently characterised ICl(Ca) in myocytes from different vascular preparations that has been activated by pipette solutions containing Ca2+ clamped at specific concentrations (Greenwood et al. 2001). In rabbit portal vein and pulmonary and coronary artery, myocytes currents with distinctive voltage-dependent kinetics were evoked by pipette solutions containing 500 nm free Ca2+ but not by pipette solutions containing 100 nm free Ca2+ (Greenwood et al. 2001). The reversal potential of the evoked currents was close to the theoretical Cl− equilibrium potential and was shifted to more negative potentials by substitution of external Cl− with the more permeable thiocyanate, consistent with the activation of a Ca2+-dependent Cl− current. This technique has been used to study ICl(Ca) in various other cell types including lacrimal gland (Evans & Marty, 1986), parotid secretory cells (Ishikawa & Cook, 1993; Arreola et al. 1996), endothelial cells (Nilius et al. 1997) and Erlich ascites (Papassotiriou et al. 2001). With this technique the kinetics of the channel are recorded at a fixed level of Ca2+ and putative modulators can be studied without an additional effect on Ca2+ influx or Ca2+ release channels (Greenwood et al. 2001).
Ca2+-dependent Cl− channels in smooth muscle cells appear to be gated directly by Ca2+ binding to the channel protein (Piper & Large, 2003), similar to Cl− currents in mammalian secretory cells (Arreola et al. 1996) and Xenopus oocytes (Callamaras & Parker, 2000; Kuruma & Hartzell, 2000). This is in comparison to Ca2+-activated Cl− channels in epithelial cells that are relatively insensitive to Ca2+ until the channel is phosphorylated by calmodulin-dependent protein kinase II (CaMKII) suggesting that it is Ca2+-dependent phosphorylation that activates the channel (Fuller et al. 1994). However, recent evidence suggests that Ca2+-activated Cl− channels in smooth muscle cells are modulated by phosphotransferase reactions. In tracheal and aortic smooth muscle cells, ICl(Ca) evoked by Ca2+ released from intracellular Ca2+ stores decayed more quickly than the concomitant Ca2+ transient (Wang & Kotlikoff, 1997; Hirakawa et al. 1999). However, the discrepancy between the time course of Ca2+ transient and ICl(Ca) disappeared after inhibition of CaMKII, a serine/threonine kinase (Wang & Kotlikoff, 1997). These findings led Wang & Kotlikoff (1997) to suggest that phosphorylation by CaMKII inactivates ICl(Ca) in equine tracheal myocytes. More recently, our group also showed that ICl(Ca) recorded from rabbit pulmonary and coronary artery but not portal vein cells are suppressed by CaMKII (Greenwood et al. 2001).
The possible phosphorylation of ICl(Ca) channels or of an accessory protein by CaMKII implies that these channels are likely to be regulated by at least one phosphatase. Among the different types of phosphatases, calcineurin (CaN), also known as PP-2B, is a serine/threonine phosphatase which is activated within the physiological range of intracellular Ca2+. Calcineurin has been implicated in a wide array of cellular functions which include the regulation of smooth muscle contraction and various ion channels (for a review see Klee et al. 1998; or Rusnak & Mertz, 2000). The major goal of this study was to test the hypothesis that CaN regulates ICl(Ca) in coronary artery smooth muscle cells. Our data suggest that CaN modulates coronary arterial ICl(Ca) in a Ca2+-dependent manner which opposes those mediated by CaMKII as previously reported (Greenwood et al. 2001). Preliminary findings have been reported in abstract form (Ledoux et al. 2000).
Isolation of vascular myocytes
Cells were prepared from arteries isolated from New Zealand white rabbits (2-3 kg) that had been killed by anaesthetic overdose (pentobarbitone, 2 mg kg−1) in accordance with Canadian and US regulations. All animal handling protocols received the approval of local ethics committees. Arterial myocytes were isolated from the left descending and circumflex coronary arteries (CA). After dissection and removal of connective tissue the CA were cut into small strips and placed in a physiological salt solution (PSS) containing no added Ca2+ and 100 μm EGTA at 22°C for 30 min. CA were then incubated in a PSS containing 10 μm Ca2+(no EGTA) and 1 mg ml−1 collagenase type 2 (Worthington, Lakewood, NJ, USA) and 0.05 mg ml−1 protease type I or type XXVII (Sigma Chemical Co., St Louis, MO, USA) for 20-25 min at 35°C. In all cases, cells were released by gentle agitation with a wide-bore Pasteur pipette. Cells were stored at 4°C and used within 6 h.
All currents were recorded in the whole-cell voltage clamp mode using pClamp software (version 8.2, Axon Instruments Inc., Foster City, CA, USA) and Axopatch-1D or Axopatch 200A amplifier (Axon Instruments Inc.). Analysis was performed using the applicable software as well as Origin (version 5.0, Microcal, Northampton, MA, USA). Current tracings, graph plots and confocal images were all exported to CorelDraw (version 9, Corel Corporation, Ottawa, Ontario, Canada) for processing of the figures.
The effect of calcineurin (CaN) on the activity of ICl(Ca) was investigated using two different and specific inhibitors of this phosphatase, namely cyclosporin A (CsA) and a synthetic peptide CaN-AF. CaN is a heterodimer composed of a catalytic subunit (CaN A) comprising auto-inhibitory and calmodulin-binding domains, bound to a regulatory Ca2+-binding subunit (CaN B). CsA is a neutral lipophilic cyclic undecapeptide isolated from the fungus Hypocladium inflatumgams that binds to cyclophilin A, an endogenous cytosolic peptidyl-proline-cis-trans isomerase, to form a complex that suppresses CaN activity with high affinity. The complex CsA-cyclophilin A inhibits CaN by binding specifically to CaN A (Klee et al. 1998; Rusnak & Mertz, 2000). The synthetic peptide CaN-AF inhibits CaN by mimicking the action of the auto-inhibitory domain of calcineurin A on the catalytic domain. CsA was applied to vascular smooth muscle cells either in the pipette solution or in the external solution and there was no significant difference in the inhibitory effect produced by these different methods. CaN-AF was dissolved in the pipette solution at a concentration of 5 μm. Experiments were performed on CA myocytes using the above pipette solution containing CaN-AF (test solution) alternated by the pipette solution alone (control).
For all whole-cell experiments, the external solution contained (mm): NaCl (130), NaHCO3 (10), TEA-Cl (5.4), MgCl2 (0.5), glucose (5.5), Hepes-NaOH (10; pH 7.35), CaCl2 (1.8) and nifedipine (0.001). The pipette solution contained (mm): cesium aspartate (100), CsCl (20), TEA-Cl (20), Hepes-CsOH (5; pH 7.2), EGTA (10), MgATP (5) and disodium GTP (0.2). Free Ca2+ concentration was adjusted to fixed levels by adding the appropriate amounts of CaCl2 and MgCl2, with free Mg2+ concentration set to 1 mm, as determined by the software WinMaxC (v.2.1, http://www.stanford.edu/~cpatton). For 350 nm Ca2+ (mm): MgCl2 (0.89) and CaCl2 (7.01); for 500 nm (mm): MgCl2 (0.77) and CaCl2 (7.7); for 1 μm Ca2+ (mm): MgCl2 (0.74) and CaCl2 (8.7). The desired Ca2+ concentrations were independently verified using a Ca2+-sensitive electrode (Thermo Orion, Model 93-20, Beverly, MA, USA) and calibrated Ca2+ solutions available from a commercial source (World Precision Instruments, Inc., CALBUF-2, Sarasota, FL, USA). In view of the fairly good correlation between the measured and calculated Ca2+ concentrations, we have used the latter throughout our study. The calculated equilibrium potential for Cl− with the three different sets of solution were: 350 nm Ca2+, −24.0 mV; 500 nm Ca2+, −23.4 mV; 1 μm Ca2+, −22.8 mV. CsA, CaN autoinhibitory fragment (CaN-AF), niflumic acid and all enzymes were purchased from Sigma Chemical Company (St Louis, MO, USA).
Freshly isolated coronary smooth muscle cells were fixed for 1 h at 37°C (5 % CO2) on coverslips pre-coated with Laminin (15 mg ml−1; Sigma Chemical Co.). Cells were initially stored in low [Ca2+] dissecting solution and were then incubated in one of two different external solutions for 15 min. One external solution was identical to that used in electrophysiological experiments except that TEA was replaced with 4.2 mm KCl and 1.2 mm KH2PO4 contained 1.8 mm Ca2+. The other bathing solution was used to raise intracellular Ca2+ to mirror internal conditions to those of electrophysiological experiments. This solution contained 80 nm ionomycin to permeabilise the membrane selectively to Ca2+ and the free Ca2+ concentration was buffered at 1 μm by 10 mm EGTA and appropriate amounts of CaCl2 and MgCl2 (as determined by the WinMaxC program). ML-7 (3 μm), an inhibitor of myosin light chain kinase, was added to suppress cell contraction prior to fixation. Myocytes were then fixed with freshly prepared paraformaldehyde 3 % (pH 7.3) and permeabilised with Triton X-100 0.2 %. Following three washes in PBS, the coverslips were then incubated in 5 % normal goat serum (Jackson Immunoresearch Laboratories Inc.) and 2.5 % BSA for 30 min at room temperature. After washing in PBS, cells were incubated with a monoclonal murine anti-CaN antibody (Sigma Chemical Co.) at a dilution of 1:200, and a polyclonal rabbit anti-CaMKII antibody (Calbiochem, La Jolla, CA, USA) at a dilution of 1:100 for 1 h at room temperature. The primary antibodies were diluted in PBS containing 2 % BSA and 1 % normal goat serum. Negative control experiments were performed by repeating the above steps in the absence of the primary antibodies. Coverslips containing the cells were washed three times in PBS and exposed with a fluorescein isothiocyanate (FITC)-coupled anti-mouse antibody (Jackson Immunoresearch Laboratories Inc.) at a dilution of 1:500 and a Cy5-coupled anti-rabbit antibody (Jackson Immunoresearch Laboratories Inc.) at a dilution of 1:500 for 1 h in the dark at room temperature. Both secondary antibodies were grown in goat and diluted in 2 % BSA and 1 % normal goat serum (Jackson Immunoresearch Laboratories Inc.). After washing three times in PBS, coverslips were mounted on standard microscope slides using diazabicyclol [2.2.2] octane (DABCO) 0.2 % diluted 1:1 in glycerol. Coronary myocytes were viewed using a ×60 oil objective (NA = 1.4) on a Zeiss Axiovert 100M microscope coupled to a LSM510 confocal scanning system equipped with image acquisition and analysis softwares (Iena, Gemany). Argon (488 nm) and helium-neon (633 nm) laser lines were used to visualise the FITC- and Cy5-conjugated secondary antibodies, respectively.
All data are the means ±s.e.m. of n cells. Microcal Origin (v.5.0, Northampton, MA, USA) was used to perform curve fitting and estimate parameters. Statistica for Windows 99 (version 5.5, Tulsa, OK, USA) was used to determine statistical significance between means using unpaired Student's t test when two groups were compared, or one- or two-way ANOVA (MANOVA) test followed by a Fisher LSD post hoc multiple range test for repeated measures in multiple group comparisons. P < 0.05 was considered to be statistically significant.
Properties of Ca2+-activated Cl− currents in coronary arterial smooth muscle cells
In all experiments, arterial myocytes were dialysed for 5-8 min with a pipette solution containing either 10 mm EGTA and no added Ca2+ or an elevated fixed Ca2+ concentration (range: 350 nm to 1 μm) while current was continuously recorded by the application of repetitive steps to +70 or +90 mV from a holding potential of −60 mV. This method minimised the possible impact of agents which might indirectly influence ICl(Ca) through changes in intracellular Ca2+ concentration (Greenwood et al. 2001). Even though cell contraction occurred within seconds and generally attained a steady-state in less than 2 min, the instantaneous and time-dependent components of ICl(Ca) gradually increased to reach a stable level after 5 min, with little, if any, evidence of run-down for up to 20 min. Besides displaying faster kinetics of activation and deactivation, ICl(Ca) recorded at near physiological temperature (30°C) exhibited qualitatively similar properties to those previously described at room temperature by our group (Greenwood et al. 2001) which include activation by intracellular Ca2+, outward rectification of both the instantaneous and late current components, and a reversal potential (Erev) near the predicted equilibrium potential for Cl−. For 350 nm, 500 nm and 1 μm Ca2+, Erev values were respectively −19 ± 5 mV (n= 6), −14 ± 1 mV (n= 23) and −15 ± 1 mV (n= 41). As reported previously (Greenwood et al. 2001), niflumic acid (100 μm) abolished the time-dependent current and a fraction of the instantaneous component of ICl(Ca) in cells dialysed with 350 nm (n= 3), 500 nm (n= 4) and 1 μm Ca2+ (n= 5).
Cell dialysis with a solution containing 10 mm EGTA and no added Ca2+ (0 Ca2+) led to very small time- and voltage-independent currents which were essentially composed of leak current (Fig. 1Aa). While the magnitude of ICl(Ca) was very similar when comparing sample recordings obtained in cells dialysed with 500 nm and 1 μm Ca2+ (respectively Fig. 1Ac and Ad), currents recorded from cells dialysed with 350 nm were smaller and exhibited slower kinetics of activation at positive potentials (Fig. 1Ab). This is better reflected in Fig. 1B which shows the mean current-voltage (I-V) relationships for the instantaneous (Fig. 1Ba) and late (Fig. 1Bb) currents registered from cells dialysed with Ca2+-free pipette solution or either one of three Ca2+ concentrations. Consistent with the Ca2+ dependence of ICl(Ca), the late current component recorded at the end of 1 s steps from −60 to +50 mV in cells dialysed with 350 nm Ca2+ was significantly reduced relative to those from cells exposed to 500 nm or 1 μm Ca2+, but was significantly higher than that recorded with Ca2+-free pipette solution. Both the instantaneous and late currents in cells dialysed with 500 nm and 1 μm Ca2+ were indistinguishable, similar to our previous observations at room temperature (Greenwood et al. 2001), a result consistent with the Ca2+ dependence of ICl(Ca) in rat portal vein myocytes (Pacaud et al. 1992).
Immunocytochemical detection of CaN and CaMKII
Figure 2A and B shows coronary smooth muscle cells respectively labelled with specific anti-CaN (green) and anti-CaMKII antibodies (red). In panels Aa and Ba, the cells had been incubated for 15 min in normal external solution containing 1.8 mm Ca2+ before fixation. Whereas the centrally located nucleus displayed little staining for CaN, intense labelling of CaMKII was apparent in this organelle. CaMKII was also detectable in the cytoplasm with modestly intense spots. Immunolabelling of CaN was also apparent in the cytoplasm but with a relatively more homogeneous distribution than CaMKII. Figure 2Ca shows a composite image of those depicted in Fig. 2Aa and Ba. At the locations indicated by the white arrows, cross-sectional images (bottom images of Fig. 2Aa, Ba and Ca) reconstructed from serial z-stack confocal fluorescent images also reflected a relatively more uniform cytoplasmic distribution of CaN than CaMKII. Both the longitudinal and cross-sections of the composite images displayed in Fig. 2Ca also support the predominance of CaMKII in the nuclear region, and a relatively uniform distribution of the two enzymes in the cytoplasm. In either case, immunofluorescence was not due to non-specific labelling by the secondary antibodies or cross-talk between the two laser excitation lines (Fig. 2E).
We next examined whether elevation of intracellular Ca2+ to levels approximating those imposed by cell dialysis with elevated Ca2+ concentration to record ICl(Ca) in the following patch clamp experiments induces translocation of the two enzymes. Before fixation, coronary myocytes were exposed for 15 min to a solution containing the Ca2+ ionophore ionomycin (80 nm), 1 μm free Ca2+ and 3 μm ML-7 to prevent cell contraction. The inclusion of ML-7 was important as separate experiments carried out with ionomycin alone led to cell balling up which precluded a proper visualisation of the cellular distribution of CaMKII and CaN. Experiments performed in the presence of ML-7 alone (1 μm free Ca2+) yielded similar results to those described in Fig. 2A and B. Figure 2Ab and Bb clearly shows that both enzymes translocated to near the membrane when the cells were stimulated, and the composite images revealed a high degree of co-localisation (Fig. 2Cb). While nuclear CaMKII labelling remained intense, CaN immunofluorescence was now apparent in the nucleus, mostly clustered in some areas, consistent with the well known nuclear translocation of CaN following elevation of cytosolic Ca2+ level (Klee et al. 1998; Rusnak & Mertz, 2000). Evidence in support for a co-localisation of the two enzymes was also provided from single labelling experiments with CaN (Fig. 2Da) or CaMKII (Fig. 2Db). In the absence of the CaMKII primary antibody, CaN immunofluorescence was much more intense both in the nucleus and near the membrane at identical laser power and gain. These results suggest that binding of the CaMKII primary and secondary antibodies, likely to be due to spatial constraint, lessened the ability of the CaN antibodies to reach their respective epitope.
Effects of cyclosporin A on ICl(Ca)
The effect of CaN on the activation of ICl(Ca) in coronary artery myocytes was investigated initially by studying the effect of CsA on currents evoked by 350 nm, 500 nm and 1 μm Ca2+. As the IC50 for the inhibition of CaN by CsA is 10−9-10−8m (Rusnak & Mertz, 2000), we used a concentration of CsA (10 μm) that would be predicted to abolish the activity of CaN. Application of CsA either in the pipette solution or in the perfusate inhibited ICl(Ca) in a Ca2+-dependent manner. Internal CsA (10 μm) significantly reduced ICl(Ca) in myocytes dialysed with 350 or 500 nm Ca2+ (Figs 3, 4A and 4B). For example with 350 nm Ca2+ in the pipette solution, late ICl(Ca) recorded at +100 mV was 14.4 ± 2.3 pA pF−1 (n= 6) in the absence and 5.6 ± 0.8 pA pF−1 (n= 5) in the presence of internal CsA (P < 0.001). Similarly, late ICl(Ca) measured at +50 mV was respectively 6.1 ± 0.8 pA pF−1 (n= 23) and 3.6 ± 0.3 pA pF−1 (n= 24) in the absence and presence of CsA (P < 0.001) in myocytes dialysed with 500 nm Ca2+. In contrast, the fully activated ICl(Ca) evoked by dialysis with 1 μm Ca2+ was not influenced by internal CsA (Fig. 4C and D). At +50 mV, late ICl(Ca) was 6.8 ± 0.8 pA pF−1 (n= 41) in control and 7.7 ± 1.4 pA pF−1 (n= 18) with CsA in the pipette solution (P= 0.985). The reversal potential of the current was not significantly altered by including CsA in the pipette solution for any of the Ca2+ tested: with 350 nm Ca2+: control, −19 ± 5 mV (n= 6); CsA, −27 ± 5 mV (n= 5); 500 nm Ca2+: control, −14 ± 1 mV (n= 23); CsA, −13 ± 1 mV (n= 24); 1 μm Ca2+: control, −15 ± 1 mV (n= 41); CsA, −15 ± 1 mV (n= 18). Although not systematically analysed, niflumic acid (100 μm) also inhibited the instantaneous and late components of ICl(Ca) in CsA-treated myocytes dialysed with either 500 nm (n= 1) or 1 μm Ca2+ (n= 5).
We also analysed whether modulation of ICl(Ca) by CaN was voltage- and Ca2+-dependent. Figure 5Aa, Ab and Ac shows graphs illustrating the effects of CsA on the chord conductance of the fully activated ICl(Ca) (gCl(Ca)) measured at the end of steps to various potentials for cells dialysed with 350 nm, 500 nm and 1 μm Ca2+, respectively. CsA significantly attenuated gCl(Ca) in cells dialysed with 350 nm (Fig. 5Aa) and 500 nm Ca2+ (Fig. 5Ab) but not with 1 μm Ca2+ (Fig. 5Ac). The impact of CsA was also noticeably larger at positive potentials with 350 nm Ca2+ than with 500 nm Ca2+ and this is better reflected in Fig. 5Ad showing a graph illustrating the voltage dependence of CsA-induced inhibition of gCl(Ca) for these two sets of experiments. In both cases, CsA significantly altered the conductance of Ca2+-activated Cl− channels in a voltage-dependent manner. The effect of CsA exhibited opposite relationships as a function of voltage in the two groups: with 350 nm Ca2+, the percentage block by CsA increased linearly from ≈45 % at −60 mV to ≈65 % at +100 mV (P= 0.0197; r2= 0.44); with 500 nm Ca2+, this parameter declined following a linear relationship from ≈52 % at −60 mV to ≈43 % at +50 mV (P < 0.0001; r2= 0.78). These results indicate that the modulation of ClCa channels by CaN is influenced by voltage although the voltage dependence of this mechanism is relatively modest.
The data presented above suggest that the voltage dependence of the modulation of ICl(Ca) may also be influenced by intracellular Ca2+. To assess this possibility, we plotted the magnitude of late ICl(Ca) at +50 mV measured in the absence or presence of CsA (data from Fig. 3 and Fig. 4) as a function of pipette Ca2+ concentration (Fig. 5B). In control (open circles), intracellular Ca2+ activated ICl(Ca) with an EC50 of 363 ± 16 nm. Consistent with a CaN-mediated alteration in Ca2+ sensitivity, CsA (filled circles) induced a rightward shift of the Ca2+ dependence of ICl(Ca) resulting in an EC50 of 515 ± 40 nm.
We previously reported that inhibition of CaMKII enhanced ICl(Ca) in rabbit pulmonary and coronary myocytes by altering channel gating mechanisms (Greenwood et al. 2001). In pulmonary myocytes dialysed with 500 nm Ca2+, 5 μm KN-93 was shown to reduce the time constant of activation of the time-dependent current during the depolarising step and prolong the tail current decay following repolarisation to the holding potential. In coronary myocytes, both the time-dependent and tail currents could be best fit by the sum of two exponential components. Figure 6 shows mean fast (τfast) and slow (τslow) time constants of the time-dependent and tail currents for experiments carried out with 500 nm (Fig. 6A) or 1 μm Ca2+ (Fig. 6B), in the absence (open columns) or presence of CsA (filled columns). With 500 nm Ca2+ (Fig. 6A), CsA significantly enhanced τfast at +60 mV; in contrast, τslow was apparently reduced by CsA, an effect which was just at the limit of significance. Both τfast and τslow tended to be reduced by CsA, although not significantly at −70 mV. In contrast, CsA had no effect on any kinetic parameter of ICl(Ca) measured at both +60 and −70 mV in cells dialysed with 1 μm Ca2+ (Fig. 6B), a result consistent with the lack of effect of CsA on ICl(Ca) at this Ca2+ concentration (Fig. 4C and D).
Effects of a specific peptide inhibitor of calcineurin on ICl(Ca) in coronary arterial myocytes
To test our main hypothesis further, we also examined the effects of intracellular application of a synthetic peptide fragment (CaN-AF, 5 μm) at a concentration near the IC50 for inhibiting the PR-II peptide fragment of CaN in vitro (IC50≈10 μm; Perrino et al. 2002). Figure 7A shows the mean I-V relationships for the instantaneous (Fig. 7Aa) and late (Fig. 7Ab) ICl(Ca) recorded after 5 min of cell dialysis in the absence (control) or presence of 5 μm CaN-AF in coronary myocytes at 30°C. This dialysis time was sufficient to observe significant effects on ICl(Ca) with autocamtide-2, a 13 amino acid specific inhibitory peptide of CaMKII, or the much larger autophosphorylated and constitutively active form of this enzyme in an earlier study (Greenwood et al. 2001). As for the CsA experiments, currents from separate cells were recorded during the same day of isolation by alternating between a pipette solution either containing or not containing the peptide inhibitor. Intracellular application of CaN-AF with 500 nm Ca2+ significantly reduced ICl(Ca) to an extent comparable to that exerted by cell dialysis with CsA (compare with Fig. 4B). Figure 7B shows two sample experiments in which late current measured at +90 mV was monitored during the external application of 10 μm CsA (indicated by arrows) in cells dialysed with or without CaN-AF. While CsA reduced ICl(Ca) within 2-3 min in the absence of CaN-AF in the pipette solution (Fig. 7B, top), this compound failed to inhibit the current when the peptide inhibitor was included in the pipette solution (Fig. 7B, bottom). Figure 7C shows a bar graph reporting mean data for the late component of ICl(Ca) pooled from 7 to 11 experiments similar to those shown in Fig. 7B. While an external application of CsA had no effect on ICl(Ca) in myocytes dialysed with the peptide, it significantly decreased (P < 0.01) the instantaneous (data not shown) and late (Fig. 7C) ICl(Ca) components. The Erev of the measured current was again not significantly affected by the presence of CaN-AF in the pipette solution: control, Erev=−13 ± 1 (n= 7); CaN-AF, Erev=−12 ± 1 (n= 11). These results support the idea that CaN enhances the activation of ICl(Ca) in arterial smooth muscle.
Effects of CaMKII inhibition on ICl(Ca)
In a recent study by our group (Greenwood et al. 2001), we reported that at room temperature, KN-93, a specific inhibitor of CaMKII (IC50≈370 nm; Sumi et al. 1991), enhanced ICl(Ca) in coronary arterial smooth muscle cells dialysed with 1 μm Ca2+ but not with 500 nm Ca2+. Since temperature could have an impact on such effects, we decided to re-examine the effects of KN-93 on ICl(Ca) elicited at 30°C. Figure 8A shows that ICl(Ca) recorded from myocytes dialysed with 500 nm Ca2+ failed to respond to KN-93 as previously found in the same preparation at room temperature (Greenwood et al. 2001). However, 10 μm KN-93 (5 min) significantly enhanced the instantaneous (Fig. 8Aa) and late (Fig. 8Ab) components of ICl(Ca) in myocytes dialysed with 1 μm Ca2+ similar to our previous study (Greenwood et al. 2001). To further test the hypothesis, we also evaluated the effects of KN-93 in cells dialysed with 350 nm Ca2+. KN-93 (10 μm) produced no effect on the magnitude of ICl(Ca). At +90 mV, the instantaneous component of ICl(Ca) was 5.7 ± 0.9 pA pF−1 in control and 5.8 ± 1.2 pA pF−1 in the presence of 10 μm KN-93 (n= 3; P= 0.8), and the late component was 12.2 ± 3.1 pA pF−1 in control and 12.9 ± 3.3 pA pF−1 after exposure to KN-93 (n= 3; P= 0.11). We also analysed the effects of KN-93 on the reversal potential of the current and found that it was not significantly influenced by this compound: with 500 nm Ca2+: control, −14 ± 2 (n= 8), KN-93, −15 ± 2 mV (n= 8); 1 μm Ca2+, −17 ± 2 (n= 12), KN-93, −19 ± 2 mV (n= 12). As previously documented (Greenwood et al. 2001), niflumic acid (100 μm) inhibited ICl(Ca) in KN-93-treated cells dialysed with 500 nm (n= 1) and 1 μm Ca2+ (n= 5). Niflumic acid-induced suppression of ICl(Ca) enhanced by KN-93 in cells dialysed with 1 μm Ca2+ and the lack of change in the reversal potential of the current by this inhibitor are consistent with the hypothesis that the increased current results from the activity of Ca2+-activated Cl− channels. Taken together, our results suggest that CaN appears to play a more important role in regulating ICl(Ca) at modest to intermediate levels of Ca2+ whereas a significant modulation of these channels by CaMKII only becomes apparent at high Ca2+.
The principal finding of this study is that the activation of Ca2+-dependent Cl− channels in rabbit coronary smooth muscle cells is enhanced by the Ca2+-sensitive phosphatase calcineurin by increasing the sensitivity of the underlying channel to intracellular Ca2+. Inhibition of CaN by internal or external application of CsA, or by intracellular dialysis of a specific inhibitory peptide, attenuated ICl(Ca) in myocytes dialysed with 350 or 500 nm Ca2+ but not with 1 μm Ca2+. Immunocytochemical experiments revealed that CaMKII and CaN were both present in coronary artery cells and both enzymes translocated towards the membrane under conditions that elevated intracellular Ca2+. Overall the data of the present study and Greenwood et al. (2001) show that Ca2+-dependent Cl− channels in rabbit coronary smooth muscle cells is regulated in a reciprocal manner by the Ca2+-dependent enzymes CaMKII and CaN. A hypothetical working model to explain the differential Ca2+-dependent contribution of these two enzymes in the regulation of ICl(Ca) is discussed.
Immunolabelling of CaMKII and CaN in coronary myocytes
CaMKII and CaN are both expressed in rabbit coronary arterial smooth muscle cells as revealed by immunocytochemistry and high-resolution confocal imaging. The polyclonal antibody raised in rabbit against CaMKII does not distinguish among the various isoforms that have been isolated so far (Singer et al. 1996). In smooth muscle, the γ and δ isoforms were shown to predominate (Zhou & Ikebe, 1994; Singer et al. 1996). Similarly, the monoclonal CaN antibody raised against an epitope on CaN A does not discriminate among the two known isoforms of this phosphatase (CaN A-α and CaN A-β, Klee et al. 1998; Rusnak & Mertz, 2000). An interesting and novel feature of the cellular localisation of these proteins is that in cells superfused with normal external solution, the two enzymes displayed a relatively uniform distribution in the cytosol, with some clusters of CaMKII near the membrane. While intense labelling of CaMKII was apparent in the nucleus, immunodetection of CaN was weak in this organelle. In contrast, a more intense labelling of the two enzymes near the membrane was observed in myocytes exposed to the Ca2+ ionophore ionomycin to elevate intracellular Ca2+. These results show that CaMKII and CaN translocate to the membrane when intracellular Ca2+ levels increase, and thus influence the activity of channels, transporters and proteins involved in signal transduction by modulating their phosphorylation state. Dual labelling showed a high degree of membrane co-localisation of the two proteins, which was not an artifact due to cross-talk between the two excitation-emission laser lines. In addition, immunostaining of CaN was more apparent in the nucleus in dual-labelling experiments. This observation is consistent with the well characterised Ca2+-dependent translocation of this phosphatase and CaMKII described in other cell types (Chantler, 1985; Klee et al. 1998; Shen & Meyer, 1999; Rusnak & Mertz, 2000). It is tempting to suggest that the regulation of ICl(Ca) by CaMKII and CaN may be exerted by a membrane-delimited interaction whereby elevated intracellular Ca2+ levels would favour the juxtaposition of the channel and the two opposing regulatory enzymes. These results have to be interpreted with caution in view of the relatively low resolution of confocal imaging due to the use of light in the visible wavelength range (Stricker & Whitaker, 1999), and perhaps the effect of signal amplification due to clustering of secondary antibodies. It is nevertheless possible that membrane targeting of CaN and CaMKII amplifies their potency at modulating ClCa activity, membrane potential and tone during signal transduction.
CaN enhances ICl(Ca)
Since CaMKII has been shown to down-regulate ICl(Ca) in arterial (Greenwood et al. 2001) and airway (Wang & Kotlikoff, 1997) smooth muscle cells, it would be expected that the underlying channel is also modulated by at least one phosphatase in manner opposite to that mediated by CaMKII. CaN-induced dephosphorylation has been shown to regulate various plasmalemmal and intracellular ion channels including neuronal N-type Ca2+ channels (Zhu & Yakel, 1997) and post-synaptic NMDA receptor-coupled Ca2+ channels (Ghetti & Heineman, 2000), ryanodine (Park et al. 1999) and inositol 1,4,5-trisphosphate (IP3) receptors (Cameron et al. 1995), insulin-dependent Cl− channels (Marunaka et al. 1998), L-type Ca2+ (Schuhmann et al. 1997) and ATP-dependent K+ (Wilson et al. 2000) channels in smooth muscle cells. The data of the present study suggest that ICl(Ca) in rabbit coronary arterial smooth muscle cells is enhanced by CaN (PP-2B) under moderately elevated intracellular Ca2+ levels.
Ca2+-activated Cl− currents elicited in myocytes dialysed with 350 or 500 nm Ca2+ in the presence of CsA were significantly smaller than those evoked in cells untreated with CsA. Micromolar concentrations of CsA have been used in many studies to investigate the potential effects of CaN on numerous cellular functions. Schuhmann et al. (1997) used 1 μg ml−1 of CsA which corresponds to 0.83 μm to investigate the potential dephosphorylating effect of CaN on L-type Ca2+ channels recorded from vascular smooth muscle cells. Marunaka et al. (1998) used 1 μm CsA to investigate the role of CaN in mediating some of the effects of insulin on Cl− channels from renal epithelium. Wilson et al. (2000) used an identical concentration of CsA to ours to study the effects of CaN on ATP-dependent K+ current recorded from rat aortic smooth muscle cells. We do not know the reason for the high potency of CaN-AF at inhibiting ICl(Ca) in our experiments. An alternative explanation might also relate to the fact that two isoforms of CaN A expressed in smooth muscle cells, CaN A-α and CaN A-β, exhibit different Km values towards the CaN auto-inhibitory peptide and CsA-cyclophilin complex as substrates (Perrino et al. 2002), which could influence the efficacy of CaN-AF to inhibit ICl(Ca). In spite of this, external application of CsA significantly attenuated ICl(Ca) in control myocytes but this compound produced no effect on the remaining ICl(Ca) evoked in myocytes dialysed with CaN-AF supporting the hypothesis that CsA and CaN-AF, which are structurally unrelated, both interact with the same protein.
Our present study indicates that the modulation of ICl(Ca) by CaN is complex and occurs in a Ca2+-dependent manner similar to the effect of CaMKII reported previously (Greenwood et al. 2001). With 350 nm Ca2+, the impact of CaN in regulating channel conductance increased with membrane depolarisation. Interestingly, a reverse situation was observed when the cells were dialysed with 500 nm Ca2+. We cannot determine from our data whether the phosphorylation status of the channel bears any influence on permeation. However, CaN does appear to modulate gating mechanisms. CsA induced significant changes in kinetics of the time-dependent current during depolarising steps and tail current relaxation upon repolarisation to the holding potential. These changes were also qualitatively opposite to those evoked by CaMKII (Greenwood et al. 2001). We speculate that CaN-mediated dephosphorylation sensitises the channel to voltage at any given intracellular Ca2+ concentration (Arreola et al. 1996). At relatively low internal Ca2+ levels, a leftward shift of the activation curve by CaN would result in an apparent increase in the percentage block of gCl(Ca) with membrane depolarisation as the range of voltages examined probably lay around the lower end portion of the steady-state activation curve. On the other hand, the decline of the percentage block of gCl(Ca) by CsA as a function of voltage in cells dialysed with 500 nm Ca2+ might be due to the fact that the activation curve is shifted further towards negative potentials. In this scenario, the impact of CaN would apparently decrease as a function of voltage due to the fact that the voltage steps tested would span the upper end of the steady-state activation curve. In addition to voltage-dependent displacement of the steady-state activation curve, we cannot dismiss the possibility that dephosphorylation may also recruit channels which have been completely silenced by phosphorylation. Finally, our analysis also revealed that CaN-mediated dephosphorylation of the channel or regulatory protein enhances the sensitivity of the channels to intracellular Ca2+ as CsA increased the value of half-maximal activation by Ca2+ 152 nm at +50 mV. This is not surprising in view of the Ca2+ and voltage dependence of these channels in vascular smooth muscle (Large & Wang, 1996; Greenwood et al. 2001) as well as other cell types (Arreola et al. 1996; Nilius et al. 1997; Callamaras & Parker, 2000; Kuruma & Hartzell, 2000). Whether CaN-induced dephosphorylation influences voltage-dependent Ca2+ binding (Arreola et al. 1996) or channel deactivation (Kuruma & Hartzell, 2000) will require substantially more quantitative analysis.
CaMKII and CaN exert opposite regulatory effects on ICl(Ca)
Our study clearly highlighted contrasting effects of CaMKII and CaN on ICl(Ca). Inhibition of CaMKII by KN-93 accelerated the kinetics of activation and delayed deactivation of ICl(Ca) in pulmonary arterial myocytes whereas the inhibition of ICl(Ca) in coronary myocytes by CsA was associated with a slowing of the activation kinetics at +60 mV. Moreover, the Ca2+ dependence of the effects of KN-93 and CsA were markedly different. KN-93 enhanced ICl(Ca) only when the channel was stimulated by 1 μm Ca2+ while CsA-mediated inhibition of CaN significantly reduced this current with Ca2+≤ 500 nm.
One explanation of our data is that dephosphorylation of autophosphorylated CaMKII is mediated by CaN. However, biochemical studies suggest that inactivation of CaMKII is exerted by PP1 and PP2A (Strack et al. 1997) or a novel CaMK phosphatase (Ishida et al. 1998). An alternative explanation for this apparent discrepancy might be that CaN is more sensitive to intracellular Ca2+ concentration than CaMKII. First, CaN is subjected to regulation by a dual mechanism: (1) direct activation of CaN by high affinity binding of Ca2+ to CaN B which leads to a modest but significant increase in Vmax, and (2) stimulation by Ca2+-calmodulin (CaM) enhances Vmax by more than 20-fold (Klee et al. 1998). Second, at saturating Ca2+ levels, the KD for CaM binding to CaN is in the range of 0.1 to 1 nm while that for CaMKII ranges from 20 to 100 nm. In vascular smooth muscle, the in situ Ca2+ dependence to generate autophosphorylated CaMKII is also highly cooperative (Hill coefficient = 3) with an EC50 near 700 nm (Abraham et al. 1996). Since CaN is more sensitive to Ca2+ than CaMKII, it would be expected that the regulation of ICl(Ca) at intermediate Ca2+ concentrations (350-500 nm) is driven primarily by CaN which increases ICl(Ca) activity by dephosphorylation of the channel or regulatory element. In other words, with 350 or 500 nm Ca2+, the ability of CaMKII to down-regulate ICl(Ca) would be overwhelmed by the more potent dephosphorylation exerted by CaN. When intracellular Ca2+ increases further, ICl(Ca) channels are probably subjected to a more balanced regulation by the two enzymes.
Direct Ca2+ gating and the influence of Ca2+-dependent enzymes
One salient question that remains unanswered, largely due to lack of molecular information on these channels in native smooth muscle cells and other excitable cells, is whether ICl(Ca) channels are directly gated by Ca2+ or require Ca2+-dependent phosphorylation for their activation similar to Cl− channels in epithelial cells (Fuller et al. 1994). We have shown in a previous study (Greenwood et al. 2001) that introduction of constitutively active CaMKII to pulmonary artery myocytes attenuated the activation of ICl(Ca). Recent inside-out patch studies were carried out in rabbit and mouse aortic (Hirakawa et al. 1999) and rabbit pulmonary artery (Piper & Large, 2003) smooth muscle cells where unitary Cl− currents could be activated by Ca2+ applied directly to the internal side of the membrane. In pulmonary myocytes, Ca2+ was shown to modify gating as well as anion permeation in a concentration-dependent manner; with 500 nm Ca2+, three distinct sub-conductance states of 1.2, 1.8 and 3.5 pS exhibiting voltage-dependent behaviour could be resolved (Piper & Large, 2003). In this study the EC50 value for activation of the channel by Ca2+ ranged from 8 to 250 nm. These values are lower than that recorded in the present study (363 nm) and in rat portal vein myocytes (365 nm; Pacaud et al. 1992). This discrepancy may reflect the minimal channel phosphorylation that would be expected from excised patch recordings. Although it would be tempting to conclude that Ca2+ directly gates ClCa channels and their activity is only modulated by CaMKII and CaN, we must be careful when extrapolating these results. Recent evidence suggests that Ca2+-dependent facilitation and inactivation of P/Q-type Ca2+ channels are conferred by Ca2+ binding on different lobes of calmodulin, which is thought to bind to an ‘IQ-like’ domain on the carboxy terminus of the α1A Ca2+ channel subunit (DeMaria et al. 2001). Local Ca2+ activation of small conductance Ca2+-dependent K+ channels (Xia et al. 1998) and Ca2+-dependent inactivation of L-type Ca2+ channels (α1C subunit; Pitt et al. 2001) by CaM tethering to the pore forming subunit have also been documented. Finally, enhancement of the slow cardiac delayed rectifier IKs by β-adrenergic stimulation requires the targeting of 3′,5′-monophosphate-dependent protein kinase and protein phosphatase 1 to the channel regulatory subunit hKCNQ1 through the protein yotiao (Marx et al. 2002). It thus remains possible that the gating of Ca2+-dependent Cl− channels in vascular myocytes is dictated by a Ca2+-binding protein and/or enzyme which may be permanently associated with one of the channel subunits or recruited during intracellular Ca2+ elevation.
Calcineurin-mediated dephosphorylation of Ca2+-activated Cl− channels or regulatory subunit is likely to play an important role in regulating resting membrane potential and tone. Within the physiological range of membrane potentials, CsA reduced the conductance of ICl(Ca) at −50 mV by 44 % in coronary artery cells dialysed with 500 nm Ca2+. As the input resistance of these smooth muscle cells is high (Leblanc et al. 1994), even a small change in Cl− conductance would impact on the resting membrane potential. Differential Ca2+-dependent regulation of ICl(Ca) by CaMKII and CaN could provide fine tuning of the membrane depolarisation caused by various levels of occupancy by constricting agonists on their receptor. At low levels of stimulation leading to small or moderate elevation of intracellular Ca2+ levels, a more important role of CaN would induce a sustained activation of ICl(Ca) and membrane depolarisation. As intracellular Ca2+ increases further, the progressively more potent activation of CaMKII by the Ca2+-calmodulin complex would increase phosphorylation, leading to down-regulation of ICl(Ca), and allow for a reduction in steady-state Ca2+ entry through L-type Ca2+ channels and reduced tone. This hypothesis is consistent with our recent finding that niflumic acid-sensitive relaxation of pressurised resistance-sized rabbit mesenteric arteries pre-constricted with the α1-agonist phenylephrine was inversely related to the concentration of this agonist (Remillard et al. 2000).
In conclusion, our data show that in coronary artery myocytes the amplitude of ICl(Ca) at modest to intermediate levels of intracellular Ca2+ is determined more by CaN-induced dephosphorylation than CaMKII-dependent phosphorylation. Our data suggest that cytosolic Ca2+ governs Cl− channel activity in arterial smooth muscle cells in a complex manner involving the recruitment of Ca2+-dependent enzymes which were shown to translocate towards the membrane during their activation.
I.G. is a Wellcome Trust Research Career Development Fellow and was additionally supported by a Royal Society Travel Grant. J.L. was supported by a Doctoral Studentship Award from the Canadian Institutes of Health Research (CIHR). The work was also supported by grants to N.L. from the CIHR, the Québec Heart and Stroke Foundation, The Montréal Heart Institute Fund and the Center of Biomedical Research Excellence (COBRE), University of Nevada School of Medicine, Reno, Nevada, USA.