Abstract
Multiple types of voltage-dependent Ca2+channels are involved in the regulation of neurotransmitter release (Tsien et al., 1991; Dunlap et al., 1995). In the nerve terminals of the neurohypophysis, the roles of L-, N-, and P/Q-type Ca2+ channels in neuropeptide release have been identified previously (Wang et al., 1997a). Although the L- and N-type Ca2+ currents play equivalent roles in both vasopressin and oxytocin release, the P/Q-type Ca2+ current only regulates vasopressin release. An oxytocin-release and Ca2+ current component is resistant to the L-, N-, and P/Q-type Ca2+ channel blockers but is inhibited by Ni2+. A new polypeptide toxin, SNX-482, which is a specific α1E-type Ca2+ channel blocker (Newcomb et al., 1998), was used to characterize the biophysical properties of this resistant Ca2+ current component and its role in neuropeptide release. This resistant component was dose dependently inhibited by SNX-482, with an IC50 of 4.1 nm. Furthermore, SNX-482 did not affect the other Ca2+ current types in these CNS terminals. Like the N- and P/Q-type Ca2+ currents, this SNX-482-sensitive transient Ca2+ current is high-threshold activated and shows moderate steady-state inactivation. At the same concentrations, SNX-482 blocked the component of oxytocin, but not of vasopressin, release that was resistant to the other channel blockers, indicating a preferential role for this type of Ca2+ current in oxytocin release from neurohypophysial terminals. Our results suggest that an α1E or “R”-type Ca2+ channel exists in oxytocinergic nerve terminals and, thus, functions in controlling only oxytocin release from the rat neurohypophysis.
Voltage-dependent channels are responsible for the Ca2+ that enters nerve terminals and elicits vesicular release of neurotransmitters (Augustine et al., 1987). Neurotransmitter release in the CNS is regulated by multiple types of Ca2+ channels (Dunlap et al., 1995). A number of studies have defined several electrophysiologically distinct Ca2+channels on neuronal cell bodies: L-, N-, T-, and P-types (Fox et al., 1987; Bean, 1989; Tsien et al., 1991; Llinas et al., 1992). Other classes of channels, such as the Q- and R-types, have been revealed by molecular cloning (Snutch and Reiner, 1992; Ellinor et al., 1993;Sather et al., 1993; Perez-Reyes et al., 1998) and the use of Ca2+ antagonists (Olivera et al., 1984;Hillyard et al., 1992; Ramachandran et al., 1993; Newcomb et al., 1998). The specific role at CNS terminals of these different types of Ca2+ channels, however, is still unclear.
The N-type channel seems to be involved in classical neurotransmission (Hirning et al., 1988), whereas the L-type is known to regulate the secretion of certain peptides (Cazalis et al., 1987; Dunlap et al., 1995). The class E (α1E) and G (α1G) Ca2+channels have been localized recently to the CNS (Westenbroek et al., 1995; Perez-Reyes et al., 1998). However, the phenotype of the expressed α1E channel is controversial (Snutch and Reiner, 1992; Randall and Tsien, 1995), and the biophysical properties of the class E current in CNS terminals remain to be determined.
To determine any role for class E channels in CNS secretion (see Wu et al., 1998, 1999), we studied the nerve terminals of the rat neurohypophysis. This is a population of relatively homogeneous peptidergic nerve endings that allows comparative study by a number of different techniques. This has been a very useful model system for characterization of nerve terminal Ca2+channels (Lemos and Nowycky, 1989; Wang et al., 1992, 1997a;Wang and Lemos, 1994; Fisher and Bourque, 1995) and of mechanisms underlying depolarization–secretion coupling (Cazalis et al., 1987;Lim et al., 1990; Lindau et al., 1992; Wang et al., 1993b,1997a; Branchaw et al., 1998). We have shown previously that “L”- and “N”-type Ca2+ channels exist in nerve terminals of the neurohypophysis (Lemos and Nowycky, 1989; Wang et al., 1992) and that they control both vasopressin (AVP) and oxytocin (OT) release, except for a significant resistant component (Cazalis et al., 1987; Dayanithi et al., 1988; Wang et al., 1993b). More recently we have shown that a “Q”-type Ca2+ current component also exists in approximately one-half of these CNS terminals (Wang et al., 1997a). Moreover, when blockers of the Q-type Ca2+ current were added to the terminals, the resistant component of AVP release was essentially abolished. In contrast, a similar resistant component of OT release was unchanged by the same concentrations of the Q-type channel blockers.
Most recently we have shown that purified native SNX-482, a specific α1E channel blocker, could inhibit the neurohypophysial Ca2+ current (Newcomb et al., 1998). This led us to examine, using a combination of pharmacological and biophysical techniques, whether class E or R-type Ca2+ channels might also exist on these CNS terminals and functionally contribute to neurosecretion.
MATERIALS AND METHODS
Electrophysiological recordings. As we have described previously (Wang et al., 1997a), after sedation by CO2 the rats were killed by decapitation using a guillotine. The neurohypophysis was then excised, following previous protocols, and homogenized in a solution containing (in mm): sucrose, 270; HEPES-Tris, 10; and K-EGTA, 0.01, pH 7.25 (Cazalis et al., 1987). All chemicals were obtained from Sigma (St. Louis, MO). The isolated neurohypophysial nerve terminals could be identified under an inverted microscope (Nordmann et al., 1987). Normal Locke's solution [containing (in mm), NaCl, 145; KCl, 5; CaCl2, 2.2; MgCl2, 1; Na-HEPES, 10; and glucose, 15, pH 7.35] was then used to perfuse the terminals. Before patch-clamp recordings, the terminals (usually 5–8 μm in diameter) were perfused with the 5 mmBa2+ (replacing CaCl2) Locke's solution, which also contained 1 μm TTX with 0.02% BSA. To obtain perforated-patch (Rae et al., 1991) recordings in the “whole-terminal” configuration (Hamill et al., 1981), freshly made amphotericin B (240–300 μg/ml) was mixed with the pipette solution that contained (in mm): Cs-glutamate, 135; HEPES, 10; glucose, 5; CaCl2, 2; MgCl2, 1; and TEA, 20, pH 7.25.
As reported previously (Wang et al., 1997a), the perforated-patch recording configuration enables us to overcome problems with the rundown of Ca2+ currents that complicated former studies (Lemos and Nowycky, 1989; Wang et al., 1992; Wang and Lemos, 1994; Fisher and Bourque, 1995; Branchaw et al., 1998). Only terminals with perforated-patch access resistances of <10 MΩ were chosen for further recordings. The Ba2+ current (IBa), which was activated by depolarizing from −80 to +10 mV and demonstrated both transient and long-lasting components (see, e.g., Fig. 1A), could be maintained for >1 hr without appreciable rundown. TheIBa was filtered at 3 kHz and sampled at 10 kHz. pClamp (Axon Instruments, Burlingame, CA) was used for acquisition and analysis of data.
Peptide release. Rat neurohypophyses (see Electrophysiological recordings) were homogenized as described previously (Cazalis et al., 1987). The homogenate was centrifuged at 2400 × g for 6 min. The resulting pellet contains highly purified nerve terminals. The nerve endings were loaded onto filters (0.45 μm Acro disk; Gelman Sciences, Ann Arbor, MI) and perfused at 37°C with normal Locke's solution. Four minute fractions of perfusate were collected, and the evoked release was triggered by an 8-min-duration pulse of a depolarizing concentration (50 mm) of K+. The results are given as AVP or OT release per fraction using specific radioimmunoassays (Wang et al., 1997a). The medium before and after the depolarizing period contained (in mm): NaCl, 40; KHCO3, 5;N-methyl-d-glucamine (NMG)-Cl, 100; MgCl2, 1; CaCl2, 2; glucose, 10; and Tris-HEPES, 10, with 0.02% BSA, pH 7.25. Depolarization medium contained 50 mmK+, in which the NMG was reduced to maintain the osmolarity (300–310 mOsm).
Polypeptide toxins. The polypeptide toxins used in this study were synthetic versions prepared by Neurex Pharmaceutical Corporation (Ramachandran et al., 1993). These were termed SNX-482, the synthetic version of a novel 41 amino acid peptide isolated from the venom of the West African tarantula Hysterocrates gigas(Newcomb et al., 1998), SNX-111, the synthetic version of ω-conopeptide MVIIA (Olivera et al., 1994), SNX-194, the methionine-12 to norleucin-12 derivative of SNX-111, and SNX-230, the synthetic version of MVIIC (Hillyard et al., 1992). The synthetic version of ω-AgaIVA (Mintz et al., 1992) was purchased from Peptides International (Louisville, KY) or synthesized as described by Gaur et al., (1994). In the text we refer to the synthetic peptides by their original names or by the Neurex terms.
Data analysis. All results are given as means ± SEM, and the statistical significance of differences in groups was analyzed using SigmaStat (Jandel Scientific, San Rafael, CA) with Tukey'st tests.
RESULTS
Ca2+ channel currents
In the isolated neurohypophysial terminals, the peakIBa, which was activated by depolarizing from −80 to +10 mV, demonstrates both transient and long-lasting components (Fig.1A). As we have reported previously (Wang et al., 1997a), the use of the dihydropyridine (DHP) Ca2+ channel antagonist nicardipine (2.5 μm) selectively inhibits the long-lasting (L-) component of the Ba2+ currents (Fig. 1). Subsequent addition of the N-type Ca2+ channel blocker MVIIA (3000 nm) led to rapid inhibition of a large portion of the isolated transient component (and, to a lesser extent, the long-lasting component) of the Ba2+ current. This concentration has been shown previously to block the N-type component maximally (Wang et al., 1997a).
In ∼50% of the neurohypophysial terminals investigated, subsequent addition of low (36 nm) concentrations of MVIIC inhibited this remaining component, and higher (150 nm) concentrations almost completely abolished it (Wang et al., 1997a). In another group of terminals (∼46%), however, the non-L- and -N-types of Ca2+ currents could not be blocked by the P/Q-type Ca2+channel blockers MVIIC or AgaIVA (Fig. 1). This resistant part of the transient Ca2+ current appeared to be analogous to an R-type Ca2+ channel current (Randall and Tsien, 1995).
Pharmacology of resistant Ca2+channel currents
To test whether this resistant component of the Ba2+ current could indeed be classified as an R-type Ca2+ channel current, Ni2+, a T- and R-type Ca2+ channel blocker, was applied to this terminal. Low concentrations (86μm) of Ni2+ inhibited the resistant current (Fig.1). Because of the low selectivity of Ni2+between Ca2+ channels, however, the identity of the resistant component of the Ca2+ current in the nerve terminal was still unclear.
A newly discovered polypeptide toxin, SNX-482, was found to be a specific blocker of the class E (α1E) Ca2+ channel (Newcomb et al., 1998). This toxin made it possible for us to identify the Ni2+-sensitive type of Ca2+ current and to probe its function in neurohypophysial nerve terminals (Wang et al., 1997b; Dayanithi et al., 1999).
First, the effects of SNX-482 on the long-lasting and transient components of the Ba2+ current of the neurohypophysial terminals were examined (Fig.2A). The isolated, transient component of the Ba2+ current usually includes an N-type and either a P/Q-type or a resistant component of Ca2+ channel currents (Wang et al., 1997a). The IC50 for the undifferentiated transient IBa, obtained from the equation I =Imax{1 − [x/(IC50 + x)]}, is 226 nm (Fig. 2B). This is similar to that for SNX-482 to inhibit the cloned α1B(N-type) Ca2+ channel current but much higher than that (IC50 = 10 nm) to block the heterologously expressed α1E-type currents (Newcomb et al., 1998). The toxin does not affect the DHP-sensitive or L-type current in these terminals. These results indicate that, at high concentrations, SNX-482 could block some combination of N-, P/Q-, and/or class E-type Ca2+ channels in the terminals.
Application of a combination of DHP, MVIIA, and MVIIC or of high concentrations of MVIIA/SNX-194 and MVIIC/AgaIVA allowed us to obtain isolated “resistant” Ba2+ or Ca2+ currents (Fig.3A). SNX-482, in a dose-dependent manner (in a total of seven terminals), inhibited the isolated resistant currents (Fig. 3A,B) with an IC50 of 4.1 nm (Fig.3C), similar to that found for the α1E Ca2+ currents expressed in human embryonic kidney (HEK) cells (Newcomb et al., 1998). The inhibition by SNX-482 of the resistant-type Ba2+ current is reversible (Fig.3D). Furthermore, both SNX-482 and Ni2+ inhibited the same previously resistant component of the Ba2+ current (Fig. 3D).
Any sensitivity of P/Q-type currents in the nerve terminals to SNX-482 was then examined. As shown in Figure4A, in the presence of the L-type blocker nicardipine and the N-type blocker MVIIA, the remaining Ba2+ component was not affected by SNX-482, although it was inhibited by the P/Q-type blocker AgaIVA. This confirmed that SNX-482 is not a P/Q-type or class A channel blocker (Newcomb et al., 1998). The inhibition of the resistant Ba2+ current component, in ∼46% of the neurohypo-physial terminals investigated, by both Ni2+ and SNX-482 lead us to conclude that this channel current most closely resembles that of the α1E Ca2+ channel subunit expressed in HEK cells (Newcomb et al., 1998).
Interestingly, in ∼5% of the terminals investigated (n = 21), in addition to the L- and N-type Ca2+ channel currents, there appears to exist both P/Q- and SNX-482-sensitive-type Ba2+ currents. Figure 4Bis an example of this, showing that the non-L- and non-N-type Ba2+ currents were partially sensitive to both SNX-482 and AgaIVA.
Biophysical properties
Biophysical characterization of the resistant component of the neurohypophysial terminal IBa also favors a class E or R-type Ca2+ channel classification. This component of the current is a transient, high-voltage-activated Ba2+ current with an inactivation rate constant of 21 ± 3 msec (n = 7) during a step to 0 mV (see Fig. 3A). Figure5A illustrates the activation (V1/2 = −14.2 mV) and steady-state inactivation (V1/2 = −58.8 mV) of the SNX-482-sensitive component of the neurohypophysial terminal Ba2+ current. The inactivating rate constant and activation and steady-state inactivation curves (Fig.5A) of this neurohypophysial Ca2+ current component are most consistent with those of the R-type Ca2+ channel in granule cells (Randall and Tsien, 1995). Nevertheless, the other transient Ca2+ current components (N- and P/Q-type) appear to have biophysical properties similar to those of this “R”-type Ca2+ component (Wang et al., 1997a).
The relative permeabilities for Ca2+versus Ba2+ between the total channel currents and the isolated SNX-482-sensitive or R-type current were compared, as shown in Figure 5B. Ba2+ currents were significantly larger than the corresponding Ca2+ currents for both the total and the isolated components. The inactivation rate constant, however, differed. The total current showed slower inactivation with Ba2+ as compared with Ca2+ as the charge carrier, whereas the R-type currents showed no difference in their inactivation with either Ba2+ or Ca2+.
Peptide release
In a previous report (Wang et al., 1997a), we found that a significant portion of OT release could not be inhibited even by simultaneous applications of L-, N-, and P/Q-type Ca2+ channel blockers. To determine whether the class E or R-type Ca2+ channel could play a role in this secretion, we measured both OT and AVP release in the same samples collected from perfused populations of nerve terminals (Fig. 6). Capitalizing on the same pharmacological protocol used to isolate the R-type component electrophysiologically (see Fig. 3A), we revealed a similar resistant component (42.3%) of Ca2+-dependent OT release (Fig.6B). In these experiments, high K+ alone induced OT release of 4258 ± 306 pg (n = 4), and both nicardipine and MVIIA, given in combination, reduced (by 57.7 ± 3.8%) high-K+-stimulated release to 1812 ± 376 pg. SNX-482 (20 nm) completely blocked the remaining stimulated OT release (to basal level, 159 ± 33 pg). In contrast, a similar resistant component (38.4%) of stimulated AVP release (406 ± 30 pg) was essentially unchanged (458 ± 63 pg; n = 4) by the same concentration of this R-type channel blocker (Fig. 6A). As shown previously (Wang et al., 1997a), this resistant component of AVP release was blocked (to basal level, 60 ± 3 pg) by the P/Q-type blocker MVIIC. These results were the same even if the order of drugs was reversed or scrambled (data not shown). Furthermore, stimulated release was stable during prolonged applications of each of the Ca2+ channel blockers, indicating that steady-state effects had been established.
We have also performed a set of experiments to compare quantitatively the SNX- 482 block of OT release with the IC50 of this toxin on R-type calcium channels. As described in the Figure 6legend, the nerve terminals were challenged with 50 mmK+ either in the absence of any channel blocker (control) or in the presence of both 2.5 μmnicardipine (L-type channel blocker) and 1 μm MVIIA (N-type channel blocker) and then subsequently with varying concentrations of SNX-482 (1, 5, 10, 20, 50, or 100 nm). In this batch of experiments, K+ alone evoked 3678 ± 139 pg (n = 3). L- and N-type channel blockers reduced this OT release by 59.7% (to 2194 ± 100 pg). Further addition of 1, 5, 10, 20, 50, or 100 nmSNX-482 suppressed the resistant OT release by 8% (to 2018 ± 30 pg), 32.6% (to 1478 ± 54 pg), 61.4% (to 846 ± 40 pg), 93.4% (to 145 ± 7 pg), 95.6% (to 96 ± 6 pg), and 97.6% (to 52 ± 6 pg), respectively. The IC50calculated from the equation r =Rmax{1 − [x/(IC50 + x)]} for the SNX-482 block of OT release is 6.8 nm, which is comparable with the IC50 for the toxin on R-type calcium currents. Finally, the effects of SNX-482 on OT versus AVP release were significantly (p < .001) different, thus revealing the importance of the class E or R-type component in only OT release.
DISCUSSION
The isolated neurohypophysial terminals are uniquely useful for studying the pharmacological, biophysical, and functional properties of Ca2+ channels at the site of secretion, and they have revealed a surprising pharmacological and functional complexity in the CNS presynaptic Ca2+channel family.
Four different components of Ca2+-dependent neuropeptide release
The regulation of neurotransmission in the mammalian CNS has been characterized by the involvement of multiple types of voltage-dependent Ca2+ channels, each of which might play a specific role in the regulation of neurotransmission. In the mammalian neurohypophysial system, the L- and N-type Ca2+ channels play an equivalent role in both AVP and OT release. This is quite different from the role of Ca2+ channels in classical neurotransmission, in which the N- and P/Q-type, instead of the L-type, Ca2+ channel current are dominant in controlling neurotransmitter release (Hirning et al., 1988; Wheeler et al., 1994; Dunlap et al., 1995). Furthermore, the P/Q-type Ca2+ channel has turned out to be critical for AVP release from neurohypophysial nerve terminals (Wang et al., 1997a). Finally, the present results have demonstrated that an SNX-482-sensitive Ca2+ current is responsible for an important part of OT release.
Identity of the resistant Ca2+ channel in nerve terminals
We have now shown that the SNX-482-sensitive Ca2+ current has similar biophysical properties to that of the class E channel. The phenotype of the expressed class E channel (Zhang et al., 1993) can resemble that of native currents described as either R- (Randall and Tsien, 1995) or T-type (Snutch and Reiner, 1992). The T-type Ca2+ channel in the CNS is a low-voltage-activated channel that is affected by Ni2+ (Tsien et al., 1991; Fisher and Bourque, 1996). The terminal SNX-482-sensitive Ca2+ current, although also blocked by Ni2+, is moderately high voltage activated and more permeable to Ba2+ than to Ca2+ (Fig. 5B). Thus, in terms of its biophysical properties, this channel appears to be different from the T-type Ca2+ channel.
A correspondence between cloned expressed class E calcium channels and various currents described as resistant, or R-type, is suggested by similar electrophysiological properties and resistance to selective antagonists of N, P/Q, and L-type calcium channels (Newcomb et al., 1998). In the absence of a selective antagonist of the class E calcium channel, however, the correspondence between calcium channel classes defined by cDNA sequencing and electrophysiology has been unclear, and the role of class E calcium channels in physiology has not been studied previously (but see Wu et al., 1998, 1999). Our study capitalizes on the recent discovery of a selective class E antagonist from tarantula venom, SNX-482, and it has allowed us to analyze directly the identity, function, and pharmacology of the resistant-type calcium channels in CNS terminals.
The initial studies of the in vitro actions of native SNX-482 have revealed unanticipated diversity in the response of native R-type currents to the peptide (Newcomb et al., 1998). Nevertheless, because low nanomolar concentrations of SNX-482 have no effects on other calcium channel subtypes (see Figs. 2, 4) (Newcomb et al., 1998), the potent block of the neurohypophysial R-type current demonstrates that the resistant current isolated pharmacologically is not simply residual current flowing through incompletely blocked N-, P/Q-, and L-type calcium channels. Thus, the variability of the response of native R-type currents almost certainly indicates pharmacological heterogeneity of the distinct entities, perhaps splice variants, which are responsible for these currents.
These variants could also explain the fact that class E mRNA has, so far, not been detected in the hypothalamic magnocellular somata that project to the neurohypophysis (Gainer and Chin, 1998). In contrast, preliminary evidence (G. Dayanithi, unpublished results), using antibodies raised against class E channels, has localized these channels to isolated neurohypophysial terminals.
The R-type Ca2+ channel controls OT, but not AVP, release
Our data suggest that in one group of terminals, there is a Ni2+- and SNX-482-sensitive Ca2+ channel able to regulate OT release preferentially, whereas in another group of terminals the P/Q-type Ca2+ channel plays a converse role in AVP release. We demonstrate here that the α1E class or R-type Ca2+ channel exists on these neurohypophysial terminals, where it participates in the control of neuropeptide secretion. These results lead to the hypothesis that the R-type channels are preferentially localized on OT peptide-containing nerve terminals and thus do not affect Ca2+ currents in vasopressinergic terminals. Interestingly, some (5%) terminals appear to have both types of channels (Fig. 4B), comparable with the observed percentage of terminals containing both OT and AVP (Wang et al., 1997b). In any case, the data clearly show that the R-type component plays an important role in OT, but not in AVP, secretion from these CNS terminals.
In conclusion, we have demonstrated that an R-type Ca2+ channel exists in at least one type of CNS terminal and is important in depolarization–secretion coupling. This lends support to the idea that R-type channels may play a specific role in synaptic transmission in other CNS synapses (Newcomb et al., 1998; Wu et al., 1998, 1999). The data presented here clarify the specific identities and functional importance of the Ca2+ channels actually located at nerve terminals and point out that the R- and P/Q-channels, at least, may be heterogeneously localized for different functions.
Footnotes
We would like to acknowledge support by the National Institutes of Health Grant NS29470 (J.R.L.). Portions of this work were supported by the Neurex Pharmaceutical Corporation, and we would like to thank members of their synthetic chemistry group for providing the synthetic peptides. We also thank G. Miljanich for support and advice.
Correspondence should be addressed to Dr. José R. Lemos, Department of Physiology and Neuroscience Program, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, MA 01655. E-mail: Jose.Lemos{at}umassmed.edu.
Dr. Wang's present address: Division of Neurobiology, Department of Neurology and Neuroscience, Cornell University Medical College, 411 East 69th Street, New York, NY 10021.