Activation of N- and P/Q-type voltage-gated calcium channels triggers neurotransmitter release at central and peripheral synapses. These channels are targets for regulatory mechanisms, including inhibition by G-protein-linked receptors. Inhibition of P/Q-type channels has been less well studied than the extensively characterized inhibition of N-type channels, but it is thought that they are inhibited by similar mechanisms although possibly to a lesser extent than N-type channels. The aim of this study was to compare the inhibition of the two channel types.
Calcium currents were recorded from adrenal chromaffin cells and isolated by the selective blockers ω-conotoxin GVIA (1 μm) and ω-agatoxin IVA (400 nm). The inhibition was elicited by ATP (100 μm) or intracellular application of GTP-γ-S. It was classified as voltage-sensitive (relieved by a conditioning prepulse) or voltage-insensitive (present after a conditioning prepulse). The voltage-insensitive inhibition accounted for a 20% reduction of both currents, whereas the voltage-sensitive inhibition reduced the N-type current by 45% but the P/Q-type current by 18%. However, the voltage dependence of the inhibition, the time course of relief from inhibition during a conditioning prepulse, and the time course of reinhibition after such a prepulse showed few differences between the N- and P/Q-type channels. Assuming a simple bimolecular reaction, our data suggest that changes in the kinetics of the G-protein/channel interaction alone cannot explain the differences in the inhibition of the N- and P/Q-type calcium channels. The subtle differences in inhibition may facilitate the selective regulation of neurotransmitter release.
- calcium channel
- patch clamp
- N-type calcium channel
- P/Q-type calcium channel
Of the multiple subtypes of voltage-gated calcium channel current (I Ca), it has been shown that calcium influx via the N-, P, and Q-subtypes triggers neurotransmission at central and peripheral synapses (Luebke et al., 1993; Takahashi and Momiyama, 1993; Regehr and Mintz, 1994; Wheeler et al., 1994; Waterman, 1996; Wright and Angus, 1996). N-type channels are identified pharmacologically as being blocked irreversibly by ω-conotoxin GVIA (ω-Cgtx GVIA) (McCleskey et al., 1987; Plummer et al., 1989) and are encoded for by the class B α1 subunit gene (Dubel et al., 1992; Williams et al., 1992). P-type channels are blocked potently by ω-agatoxin IVA (ω-Aga IVA) (Mintz et al., 1992); Q-type channels also are blocked by ω-Aga IVA but with a somewhat lower affinity (Zhang et al., 1993; Randall and Tsien, 1995). Biophysically, the P-type channels exhibit little or no inactivation during prolonged depolarizations, whereas Q-type channels show substantial inactivation even after 100 msec (Llinás et al., 1989; Mintz et al., 1992;Randall and Tsien, 1995). It remains unclear which α subunit gene encodes for the P- and Q-type channels. The class A α1subunit produced inactivating currents that more closely resembled the Q-type current (Zhang et al., 1993), but it also may encode for channels with properties similar to the P-type (Stea et al., 1994; Liu et al., 1996). Until these issues are resolved, many researchers in the field have adopted the terminology “P/Q-type” current when referring to either component.
Synaptic transmission can be regulated by neurotransmitter modulation of voltage-gated calcium channels, such as the well documented inhibition of N-type channels by activation of G-protein-linked receptors (Hille, 1992, 1994; Dolphin, 1995). Several pathways, most of which are membrane-delimited, converge on the N-type channels. There is general consensus that this involves a direct effect of the activated G-protein subunit or subunits on the calcium channel itself, and recent evidence indicates a key role for the G-protein βγ subunits (Herlitze et al., 1996; Ikeda, 1996). The inhibition exhibits characteristic gating shifts manifested as slowed activation kinetics, a diminution of the inhibition at positive membrane potentials, and partial relief from inhibition by conditioning prepulses (Bean, 1989;Elmslie et al., 1990; Penington et al., 1991). These effects have been incorporated into models in which the channels exhibit two functional gating states, one in the presence (“reluctant”) and another in the absence (“willing”) of inhibition (Bean, 1989; Elmslie et al., 1990; Boland and Bean, 1993; Golard and Siegelbaum, 1993).
Recently, it has become apparent that P/Q-type channels are inhibited by similar mechanisms (Mintz and Bean, 1993), but there has been little detailed comparison with N-type current inhibition. However, a few reports suggest that the N-type current is inhibited to a greater extent than the P/Q-type current (Mintz and Bean, 1993; Bayliss et al., 1995; Bourinet et al., 1996; Currie and Fox, 1996). Such differential targeting may have a role in the selective regulation of neurotransmission. Therefore, we have compared the similarities and differences in the inhibitory modulation of N- and P/Q-typeI Ca in bovine adrenal chromaffin cells.
MATERIALS AND METHODS
Culture of cells. Chromaffin cells were prepared by digestion of bovine adrenal glands with collagenase and purified by density gradient centrifugation, as previously described (Artalejo et al., 1992a). The cells were plated on collagen-coated glass coverslips (22 × 22 mm) at a density of ∼0.15 × 106 cells/cm2 and maintained in an incubator at 37°C in an atmosphere of 92.5% air/7.5% CO2 with a relative humidity of 90%. Fibroblasts were suppressed effectively with cytosine arabinoside (10 μm), leaving relatively pure chromaffin cell cultures. Although mixed, the cultures were enriched somewhat for epinephrine-containing over norepinephrine-containing cells. One-half of the incubation medium was exchanged every day. This medium consisted of DMEM/F12 (1:1) supplemented with fetal bovine serum (10%), glutamine (2 mm), penicillin/streptomycin (100 U · ml−1/100 μg · ml−1), cytosine arabinoside (10 μm), and 5-fluorodeoxyuridine (10 μm).
Electrophysiology. Chromaffin cells were voltage-clamped in the whole-cell configuration of the patch-clamp technique (Hamill et al., 1981) with an Axopatch 1C amplifier (Axon Instruments, Foster City, CA) at a holding potential of −80 mV, andI Ca was activated by step depolarizations. Current–voltage curves were generated by voltage ramps of 100 msec duration from the holding potential (−80 mV) to +100 mV. Leak currents were generated by averaging 16 hyperpolarizing sweeps (steps or ramps). All of the data reported in this paper were capacitance- and leak-subtracted. The data were filtered at 2 kHz and then digitized at 100 μsec per point. Series resistance was compensated partially (≈ 80%) by using the series resistance compensation circuit of the Axopatch-1C amplifier. Electrodes were pulled from microhematocrit capillary tubes (Drummond, Broomall, PA) and coated with SYLGARD (Dow Corning, Midland, MI). After fire polishing, final electrode resistances when filled with the CsCl-based patch pipette solution (see below) were ∼1.5–2.0 MΩ. Voltage protocols and data analysis were performed in AxoBasic. Data are reported as mean ± SEM, and statistical significance was determined with paired or independent Student’s t test.
Solutions. Electrodes were filled with (in mm): CsCl 110, MgCl2 4, HEPES 20, EGTA 10, GTP 0.35, ATP 4, and creatine phosphate 14, pH 7.3 (adjusted by CsOH); osmolality, ≈ 310 mOsm. In some experiments 0.07 GTP-γ-S replaced an equal amount of GTP. The NaCl-based extracellular recording medium contained (in mm): NaCl 140, KCl 2, glucose 10, HEPES 10, CaCl2 10, and tetrodotoxin (TTX) 2 μm, pH 7.3 (adjusted with NaOH); osmolality, ≈ 315 mOsm. In some experiments the NaCl was replaced by choline Cl. Nisoldipine (1 μm) was present in all extracellular solutions to block any facilitation (L-type) I Ca. ATP for extracellular application was prepared as a stock solution in distilled water and kept on ice before dilution to final concentrations in extracellular recording medium. ω-Conotoxin GVIA (Alomone Labs, Jerusalem, Israel) and ω-agatoxin IVA (gift from Dr. N. Saccomano, Pfizer, Groton, CT) were stored at −20°C either lyophilized or as concentrated stocks in distilled water. Final concentrations were prepared daily by dilution in extracellular recording medium. Bovine serum albumin (BSA; (1 mg/ml) was included in the recording medium to prevent nonspecific binding of the peptides. BSA itself was found to have no effect on I Ca .
After the formation of a seal, the cell was perfused continually by an Adams & List (Westbury, NY) DAD-12 superfusion system. In brief, this consisted of a “sewer pipe” arrangement in which a quartz capillary of 100 μm diameter is placed close to the cell. The sewer pipe was connected to six reservoirs, and flow from these was controlled by valves operated by the AxoBasic software. This enabled the cell to be perfused continually with fresh solution and the rapid exchange of solutions for application of toxins or ATP. The exchange time for switching between two solutions was <1 sec. The bath volume was kept relatively constant by using an outlet connected to a vacuum. All solutions perfused in this way contained 1 mg/ml BSA. All experiments were performed at room temperature (∼23°C).
We have shown previously that ATP inhibits both N-type and P/Q-type calcium currents in adrenal chromaffin cells. The aims of this study were to investigate the similarities and differences in the inhibitory response of the N- and P/Q-type calcium currents. These experiments included prepulse depolarizations to very positive potentials. Chromaffin cells possess L-type calcium channels, which are normally quiescent but can be recruited by similar depolarizing prepulses, neurotransmitters such as dopamine that elevate cAMP, or rapid repetitive depolarizations in the physiological range (Fenwick et al., 1982; Artalejo et al., 1990, 1992b). To avoid the complication of recruitment of these channels, we included 1 μmnisoldipine, which selectively blocks these L-type channels (Artalejo et al., 1991) in all extracellular solutions.
In the presence of nisoldipine (1 μm) theI Ca was composed almost totally of N-type and P/Q-typeI Ca. Perfusion of 1 μm ω-Cgtx GVIA to block the N-type current reduced the peak inwardI Ca amplitude by 48 ± 2% (n = 20). Subsequent perfusion of 400 nmω-Aga IVA to identify the P/Q-type I Ca blocked most of the remaining current, equivalent to 45 ± 2% (n = 17) of the total current amplitude (Fig.1 A). A small component of the total current (7 ± 1%; n = 17) remained after application of both toxins. There was no difference in the percentage block by the two toxins if the order of application was reversed (i.e., ω-Aga IVA applied before ω-Cgtx GVIA), indicating that there was no overlap in the pool of channels blocked by each toxin. In this case ω-Aga IVA blocked 43 ± 3% and ω-Cgtx GVIA blocked 47 ± 3% of the peak I Ca (n = 9). Some cells bathed in the NaCl-based recording solution exhibited a small TTX-resistant inward sodium current. This current decayed rapidly (3–4 msec) during the depolarizing steps used to activateI Ca and was fully inactivated either by depolarizing prepulses or by changing the holding potential to −50 mV. The current also disappeared when the cells were perfused with choline Cl-based recording solution. It did not interfere with the measurement of peak I Ca, because changing the recording solution from NaCl to choline Cl-based solution had no effect on the measured peak inward I Ca amplitude.
ATP inhibits N-type ICa to a greater extent than P/Q-type ICa
To compare the inhibition of the N- and P/Q-type currents, we applied a supramaximal dose of ATP (100 μm) to cells before and after application of ω-Cgtx GVIA (Fig.1 A). The I Ca that remained after application of ω-Cgtx GVIA was termed the P/Q-typeI Ca. Subtraction of the currents obtained after ω-Cgtx GVIA application from those obtained before ω-Cgtx GVIA application yielded a difference current that was pure N-typeI Ca. Figure 1 B shows that the N-type I Ca was inhibited to a greater extent than the P/Q-type I Ca and that the activation kinetics of both currents were slowed. After very long steps there was often little or no inhibition, because it was slowly relieved during the depolarization. Thus all current amplitudes from the same cell were measured at a fixed time after activation of the current (typically ∼10 msec) corresponding to the peak of the controlI Ca. Although arbitrary, it is common practice to measure the inhibition in this manner. Activation of N-type channels in chromaffin cells is slightly slower than that of P/Q-type channels (Artalejo et al., 1992a), but we were careful to measure the amplitudes at a point at which both currents had peaked.
The mean percentage of inhibition by ATP was 65 ± 1.6% for the N-type current and 39 ± 1.2% (n = 35;p < 0.001) for the P/Q-type current. The same pattern of differential inhibition was seen if the order of toxin application was reversed so that ATP was applied before and after application of ω-Aga IVA (Fig. 1 C). In this case the current remaining after block with ω-Aga IVA was the N-typeI Ca, and the P/Q-typeI Ca was the subtracted difference current. In nine cells recorded in this way the N-type current was inhibited by 60 ± 4.3%, and the P/Q-type current inhibition was 36 ± 3.0% (p < 0.001).
Differences in the voltage sensitivity of the inhibition of N- and P/Q-type currents
The ATP-mediated inhibition of I Caexhibited voltage sensitivity; it was diminished at very positive test potentials and was relieved in part by depolarizing prepulses to very positive potentials (Currie and Fox, 1996). The two current components were isolated in the same manner as shown in Figure 1 to determine whether the inhibition of the N- and P/Q-typeI Ca was equally voltage-sensitive. Depolarizing voltage step commands to +20 mV for 20 msec were used to activate the currents in control conditions and with ATP present to elicit the inhibition. Then, still in the presence of ATP,I Ca was activated again, but the test pulse was preceded by a depolarizing prepulse to +100 mV for 50 msec (Fig.2 A). Such prepulses relieved a significantly greater proportion of the N-type current inhibition (69 ± 3%) than the P/Q-type current inhibition (47 ± 3%;n = 17; p < 0.001). Note that, in control conditions in the absence of inhibition, identical prepulses had no effect on I Ca, because in chromaffin cells these channels show no voltage-dependent inactivation (Artalejo et al., 1992a). Thus under control conditions the peak amplitude of those currents preceded by a prepulse was 98.4 ± 0.4% (n = 23) of those not preceded by a prepulse.
Figure 2 B shows the pooled data for these experiments in which the total percentage of inhibition of the peak current is subdivided into the voltage-sensitive component (the component that was relieved by a depolarizing prepulse) and the voltage-insensitive component (the component of the inhibition that was not relieved by a depolarizing prepulse). The voltage-insensitive component accounted for approximately the same amount of inhibition for both currents. In contrast, approximately twice as much N-type current inhibition was voltage-sensitive as compared with the P/Q-type current. The same pattern of inhibition was seen if the data were obtained before and after ω-Aga IVA rather than ω-Cgtx GVIA to isolate the currents (Fig. 2 B). It should be noted that in using this protocol there will be some reblock of the voltage-sensitive inhibition during the 10 msec interval between the prepulse and test pulse (see below). However, this is likely to be <10% and to be similar for the two current types.
Intracellular GTP-γ-S mimics the differential inhibition produced by ATP
GTP-γ-S is an analog of GTP that is resistant to hydrolysis, and therefore irreversibly activates G-proteins. By including GTP-γ-S in the patch pipette solution, the inhibition ofI Ca could be elicited without the need for receptor activation by ATP. In all experiments reported here, after inhibition of I Ca by the GTP-γ-S, ATP was applied to the cells to ensure that the G-proteins were fully activated and that the inhibition was maximal. In most cases the response to ATP was occluded, although in some cells it did produce a small increase in the inhibition.
The GTP-γ-S-containing patch pipette solution was identical to the normal solution except that 70 μm (one-fifth) of the GTP was replaced by GTP-γ-S. Once the whole-cell configuration was obtained, the GTP-γ-S slowly dialyzed into the cell and activated the G-proteins, causing an inhibition of the calcium current over a period of 5–7 min (Fig. 3 A). In many cells the GTP-γ-S appeared to activate transiently an inward current over the same time course (data not shown). This inward current decayed back to baseline within a few minutes in almost all cells, and those cells in which this was not the case were discarded. The ionic nature of this current was not investigated in this study. Figure 3 A shows that the inhibition of I Ca produced by GTP-γ-S was similar to that produced by ATP. The currents on the left were recorded shortly after breaking into the cell before the GTP-γ-S had time to diffuse into the cell and activate the G-proteins. A prepulse given at this time had no effect on the amplitude or kinetics ofI Ca. The currents on the right were recorded 7–8 min later after the GTP-γ-S had inhibited the current. A prepulse given at this time sped the activation kinetics and increased the amplitude of I Ca due to relief of the voltage-sensitive component of the inhibition.
When GTP-γ-S was used to produce the inhibition, the differential targeting of the N- and P/Q-type I Ca was still apparent. This was manifested in the percentage of current blocked by ω-Cgtx GVIA (Fig. 3 B). In cells loaded with GTP-γ-S-containing patch pipette solution, ω-Cgtx GVIA blocked 25 ± 1.8% (n = 27) of the current, and in cells loaded with control patch pipette solution containing no GTP-γ-S, it blocked 48 ± 2.4% (n = 20; p < 10−9). This is consistent with a higher proportion of the N-type current being inhibited by the GTP-γ-S, as compared with the P/Q-type current.
The voltage-sensitive component of the inhibition produced by GTP-γ-S was investigated by using a prepulse protocol like that shown in Figure2 A. The N- and P/Q-type I Cawere isolated by using ω-Cgtx GVIA as before. Because dialysis of GTP-γ-S into the cell starts immediately on entering the whole-cell configuration, there was no way to obtain reliable control (uninhibited) data. Thus, the amount of voltage-sensitive inhibition was quantified by determining the increase in current amplitude produced by a prepulse relative to the amplitude of the current remaining after full inhibition (Fig. 3 A). This potentiation of current amplitude was 2.51 ± 0.11 for the N-typeI Ca and 1.46 ± 0.04 (n = 21; p < 10−9) for the P/Q-typeI Ca, illustrating the greater degree of voltage-sensitive inhibition of the N-type current relative to the P/Q-type current. The same analysis performed on currents inhibited by ATP application produced similar results (Fig. 3 C). The potentiation was 2.20 ± 0.12 for the N-type current and 1.33 ± 0.03 (n = 15; p < 10−4) for the P/Q-type current.
Comparison of the voltage-sensitive inhibition
Because the voltage-sensitive component of the inhibition accounted for the differential targeting of the two current types, it was investigated more closely by varying the parameters of the prepulse protocol. The voltage protocols were repeated before and after application of ω-Cgtx GVIA to isolate the N- and P/Q-typeI Ca. GTP-γ-S patch pipette solution was used to stimulate the inhibition to minimize any complications with desensitization of the agonist response, but experiments also were done using ATP to stimulate the inhibition. Any cells that showed excessive desensitization or rundown over the course of the protocol were discarded.
Figure 4 shows the data in which the time course of relief from inhibition was investigated. Prepulses were to either +100 or +140 mV and were separated from the test pulse by an interval of 10 msec. The duration of the prepulse was varied from 2 to 50 msec. The current increase caused by each prepulse (over currents activated with no prepulse preceding them) was normalized, and the data were fit with a single exponential to determine the time constant for the relief from inhibition. At +100 mV (Fig. 4 A) the time constant for the N-type current was 9.8 ± 0.7 msec (n = 12), and for the P/Q-type current there was a small but significant increase in the time constant to 13.1 ± 1.0 msec (n = 12; p = 0.01). When this was repeated using ATP to produce the inhibition, there was a small difference in the time constants, but it was not statistically significant (N-type = 9.9 ± 1.0 msec and P/Q-type = 10.8 ± 1.8 msec; n = 4). At +140 mV (Fig.4 B) there was no difference in the time course for relief from inhibition with time constants of 11.5 ± 1.0 and 11.7 ± 0.8 msec (n = 5) for the N- and P/Q-type currents, respectively.
Figure 5 explores the time course for channel reinhibition after a prepulse to +100 mV. When the cell is repolarized after the prepulse, the channels become reinhibited, so by varying the duration of the interpulse between the prepulse and test pulse, the time course of this reinhibition was determined. This was repeated at potentials of −60, −80, and −100 mV. The prepulse was to +100 mV for 50 msec, and the interpulse duration was varied from 5 to 300 msec. The increase in current produced by the prepulses was normalized, and the data were fit with a single exponential decay. Figure 5 Ashows that at −60 mV the time constants for reinhibition of the N- and P/Q-type currents were virtually identical at 108 ± 11 and 105 ± 11 msec, respectively (n = 5). At −80 mV the time constants again were not significantly different when the inhibition was elicited either by GTP-γ-S-containing patch pipette solution (Fig. 5 B) (N-type = 112 ± 17 msec; P/Q-type = 91 ± 8 msec; n = 7) or by ATP application (N-type = 110 ± 7 msec; P/Q-type = 87 ± 10 msec; n = 5). Similarly, at −100 mV (data not shown) there was no statistically significant difference between the time constants of reinhibition for the N-type (143 ± 17 msec) and the P/Q-type currents (125 ± 11 msec; n = 5).
The voltage dependence of the relief from inhibition was studied by varying the potential to which the prepulse was stepped (Fig.6). Prepulses to increasingly positive potentials relieve more of the inhibition, so for a prepulse of a fixed duration the voltage dependence of the relief can be compared for the two current types. The prepulse duration was 10 msec, as was the interpulse duration. Prepulse potential was changed in the range between −10 mV and +120 mV. The increase in current amplitude for each prepulse was normalized, and the data were fit with a sigmoidal curve (Fig. 6). The midpoint of the curve (representing 50% of maximal relief from inhibition) was found to be shifted significantly in the hyperpolarized direction for the P/Q-type current (+40 ± 1.9 mV) relative to the N-type current (+51 ± 1.8 mV; n = 6;p < 0.001). The slight change in the slope of the curve was not significant. Similar data were obtained using ATP to inhibit I Ca. The midpoint of inhibition was +40 ± 2 mV for the P/Q-type current and +56 ± 2 mV (n = 6; p < 0.001) for the N-type current. The experiments were repeated with a prepulse duration of 100 msec rather than 10 msec. The data were shifted slightly to the left for both current types, but there remained a significant difference between the two, with a midpoint for the P/Q-type current of +31 ± 2 mV and for the N-type current of +47 ± 2 mV (n = 4; p = 0.03).
Comparison of the current–voltage relationship of N- and P/Q-typeICa
The hyperpolarized shift in the voltage dependence of relief from inhibition could be attributable to shifts in the gating of the P/Q-type current relative to the N-type current. This was investigated by comparing the current–voltage relationship for the two current types, using ramp depolarizations from −80 to +100 mV. These experiments were performed in choline Cl-based recording solution to minimize the contribution of TTX-resistant sodium channels if present. Ramps were performed before and after application of either ω-Cgtx GVIA (to isolate the N-type current) or ω-Aga IVA (to isolate the P/Q-type current). Application of ω-Cgtx GVIA reduced the amplitude of I Ca and produced a small but statistically significant leftward shift in the I/V relationship relative to control (Fig. 7 A). Application of ω-Aga IVA also reduced the amplitude and shifted the I/V curve by a similar amount but in the opposite direction (Fig. 7 B), indicating that the shift was due to genuine differences in the properties of the N- and P/Q-type currents. The mean data revealed a statistically significant hyperpolarized shift (7–10 mV) in the peak of the I/V curve for the P/Q-type current relative to the N-type current (Table 1).
The percentage inhibition produced by ATP was examined by using rampI/V = curves (n = 5), and it was found that there was no significant difference in the extent of the inhibition over a 30 mV voltage range spanning the peak for both current types (+10 to +30 mV for the P/Q-typeI Ca; +20 to +40 mV for the N-typeI Ca). The inhibition was proportionally less for both current types at potentials both negative and positive to those stated above. Hence, the differential inhibition reported in this paper using step depolarizations (to +20 or +30 mV) were not artifacts generated by the shift in gating properties for the two current types.
N-type channels are preferentially targeted by ATP
Cultured bovine adrenal chromaffin cells provide a good model in which to compare the inhibition of N- and P/Q-typeI Ca because they are voltage-clamped readily, express the currents in an approximately 1:1 ratio, and we previously have characterized the inhibition produced by ATP (Currie and Fox, 1996). This paper clearly demonstrates that the N-typeI Ca was inhibited to a greater extent than the P/Q-type I Ca. A supramaximal dose of ATP (100 μm; EC50 = 0.5 μm) was used to ensure full activation of the receptors before and after application of either ω-Cgtx GVIA or ω-AGA IVA to isolate the two currents. This ruled out the possibility that the difference was due to cell variability or the level of G-protein activation and was confirmed by using GTP-γ-S to maximally stimulate the inhibition in an irreversible manner.
A small component of current in the chromaffin cells was insensitive to either ω-Cgtx GVIA or ω-Aga IVA. Although not thoroughly investigated, ATP inhibited this current in some cells (Currie and Fox, unpublished data). To ensure that the toxin-insensitive current did not account for the differential inhibition of the N- and P/Q-typeI Ca, we used both ω-Cgtx GVIA and ω-Aga IVA to isolate the currents. Thus, in some cases the residual toxin-insensitive current was included in the N-typeI Ca, and in others it was included in the P/Q-type I Ca, but this had little effect on the data because in all cases the residual current was extremely small.
Using conditioning prepulses, we classified the inhibition as either voltage-sensitive (that component relieved by a prepulse) or voltage-insensitive (that component present after a prepulse). This classification is purely functional, because it is by no means clear whether the two components represent different signaling pathways or whether there is only partial reversal of a single inhibitory mechanism. Nevertheless, it was striking that the voltage-insensitive component of the inhibition accounted for the same percentage reduction of the N- and P/Q-type currents (≈ 20%) and that the difference in the inhibition was accounted for by the voltage-sensitive component (≈ 45% Vs ≈ 18%).
Properties of the voltage-sensitive inhibition
Both the N- and P/Q-type currents exhibited the typical characteristics of this type of inhibitory modulation, including slowed activation kinetics and partial reversal by prepulses to very depolarized potentials. These characteristics have been incorporated into models (Bean, 1989; Elmslie et al., 1990; Boland and Bean, 1993;Golard and Siegelbaum, 1993) in which the channels exhibit two functional gating states, one from which they readily open on depolarization (willing) and one from which they open slowly (reluctant). Activation of the inhibitory pathway (perhaps binding of the G-protein βγ subunit to the channel) shifts a proportion of the channels into the reluctant state, but strong depolarizations can still open these channels and overcome the inhibition, possibly by promoting the dissociation of the G-protein subunit or subunits from the channel (but see Kasai, 1992).
Despite the 2.5-fold difference in the magnitude of the voltage-sensitive inhibition, there were only small differences in the time course for relief from inhibition or for reinhibition of the two channel types. There was a significant shift in the voltage dependence of the relief from inhibition (Fig. 6), but this may be accounted for by the similar shift in the current–voltage relationships of the two channels (−10 mV for the P/Q- relative to the N-type). In this study the activation in the presence of inhibitor could not be well fit even with multiple exponentials and was complicated further by somewhat different Ca2+-dependent inactivation kinetics of the N- and P/Q-type currents. This prevented a quantitative comparison, but qualitatively the slowing was similar for the N- and P/Q-typeI Ca. Thus, all of our data point to there being little difference in the voltage-sensitive inhibition of the two channel types in terms of the kinetics or voltage dependence.
What accounts for the differential inhibition of the two channel types?
For the purposes of this discussion we will assume that the voltage-sensitive inhibition is produced by direct interaction of the activated G-protein subunit or subunits with the calcium channel, as represented by the simple binding scheme below:
In Equation 1 above, CaCh represents either the N-type or P/Q-type calcium channel, G* the activated G-protein, andk 1 and k -1 are the forward and back rate constants, respectively. The interaction site or sites seem to be on the α1 subunit of the calcium channel, because channels expressed in Xenopus oocytes by either α1A or α1B subunits alone are susceptible to G-protein-mediated inhibition (Roche et al., 1995;Bourinet et al., 1996). Binding sites for G-protein βγ subunits recently have been identified in the I–II linker of the α1A and α1B subunits (De Waard et al., 1997; Zamponi et al., 1997), but see Zhang et al. (1996) for a report in which sites in both domain I and the C terminus of the α1 subunit are involved in G-protein modulation. The calcium channel β-subunit also plays a role, because it clearly “antagonizes” the inhibition in both native (Campbell et al., 1995) and recombinant channels (Bourinet et al., 1996).
Assuming a binding model such as (Eq. 1), the fractional occupancy of the channels by G-proteins can be written as: Equation 2The time constant for binding of the G-protein to the channel, for instance during reblock of the channels after a prepulse, is: Equation 3Recently Tsien and colleagues compared the inhibition of cloned α1A and α1B channels expressed inXenopus oocytes (Zhang et al., 1996). They suggested that variations in the off-rate (k -1) of the G-protein from the channel explain the differences in the extent of inhibition (fractional occupancy, Eq. 2) and reinhibition rate (Eq. 3) that they observed for the two channel types. Our data obtained in native channels from chromaffin cells show that the kinetics of relief from inhibition and reinhibition between the two channel types are very similar, suggesting that the off-rate remains unchanged. If there is an increase in the G-protein off-rate (k -1) for the P/Q-type channels relative to N-type channels, it suggests that the local concentration of activated G-proteins near P/Q-type channels is lower, such that the kinetics for the two channel types (Eq. 3) remain similar. Alternatively, the difference in reinhibition rate observed between our study and that of Zhang et al. (1996) may be attributable to the fact that the channels expressed in theXenopus oocyte system are not identical to those in chromaffin cells. These differences may reside in the α1subunit or in accessory subunits. A preliminary report showed that the extent of inhibition of currents produced by expression of α1A or α1B alone was similar but that coexpression of a β3-subunit led to a greater reduction in the inhibition of α1A, as compared with α1B (Roche and Treistman, 1996). This suggests that the influence of the calcium channel β-subunit is critical in leading to the differential inhibition of the two channel types.
Another explanation of our data is that, rather than the affinity of the G-protein/calcium channel interaction, it is the efficacy with which the bound G-protein exerts its actions that is different for the two channel types (off-rates and G-protein concentrations are similar for the two channel types). However, this explanation is not consistent with the data of Zhang et al. (1996).
N- and P/Q-type calcium channels as targets for regulation of neurotransmission
N-type calcium channels are targets of at least five inhibitory pathways (Hille, 1994), and it now seems that P/Q-type channels are inhibited in a similar manner. It remains unclear whether the voltage-dependent relief from inhibition identifies two distinct pathways or whether there is partial relief of a single pathway. Phosphorylation may mediate the voltage-insensitive inhibition in at least some cell types (Diversé-Pierluissi et al., 1995). Functionally, the voltage-sensitive inhibition is greater for the N-type channels and also may provide a more “dynamic” aspect to the inhibition. This would be achieved if there were relief from inhibition during trains of action potentials. However, use of short depolarizing steps to mimic action potentials suggests that for N-type channels this may be confined to cells capable of firing at very high rates (Penington et al., 1991; Williams et al., 1997). Relief of P/Q-type channel inhibition by action potentials may depend on the duration of the action potential waveform (Brody et al., 1997).
The nonlinear relationship between intracellular calcium levels and secretion (Heinemann et al., 1994) means that even small changes in calcium influx could have profound effects on transmitter release. Ultimately, the regulation of neurotransmitter release will be due to the summation of many subtle inputs, including the inhibition ofI Ca investigated in this paper. There is also evidence to suggest that P/Q-type channels may be potentiated selectively in some systems (Mogul et al., 1993; Huang et al., 1996). Hence, one might envisage a situation whereby the relative proportion of N- and P/Q-type channels at synapses would be regulated precisely. Synapses in which P/Q-type channels dominate would be inhibited less and so would operate over a smaller range (between uninhibited and maximally inhibited states) than synapses in which N-type channels dominate, thus providing a fine tuning of the regulation of release.
This work was supported by National Institutes of Health grants to A.P.F. We thank Dr. Zhong Zhou for kindly preparing the chromaffin cells and Dr. Nicholas Saccomano of Pfizer, Incorporated, Groton, CT, for the gift of ω-Aga IVA.
Correspondence should be addressed to Dr. Kevin Currie, The Department of Pharmacological and Physiological Sciences, The University of Chicago, 947 East 58th Street, Chicago, IL 60637.