Characterization of the Ca²⁺ Current in Isolated Terminals of Crustacean Peptidergic Neurons

Ca²⁺ influx through voltage-operated Ca²⁺ channels plays a crucial role in regulating neurosecretion, and yet the small size and inaccessibility of most terminals has impeded the characterization of channels located at the release sites. Notable exceptions include the squid giant synapse (Llinás et al. 1981a,b; Augustine et al. 1985a,b), the rat neurohypophysis (Lemos et al. 1994), the chick ciliary ganglion (Stanley and Goping, 1991) and the goldfish retinal bipolar cells (Heidelberger and Matthews, 1992). In particular, little is known about the voltage-dependence and kinetics of activation and inactivation of the Ca²⁺ channels present in nerve terminals.

One preparation amenable to this kind of analysis is the crab neurohemal organ, the X-organ–sinus gland, which contains peptide-secreting neurons with exceptionally large terminals. This system has been used extensively to study aspects of excitation–secretion coupling (for a review, see Stuenkel and Cooke, 1988). The Ca²⁺ current (I_Ca) in isolated X-organ somata and growth cones has been well characterized (Lemos et al. 1986; Meyers et al. 1992; Meyers, 1993; Richmond et al. 1995), but little is known of I_Ca at the peptide release sites in the terminals of these neurons, although regenerative, tetrodotoxin (TTX)-resistant depolarizations indicative of Ca²⁺ spikes have been recorded from terminals using intracellular electrodes (Cooke, 1985; Nagano and Cooke, 1987). Recently a terminal preparation was developed from the sinus gland, producing isolated nerve endings routinely measuring 5–10 μm in diameter which can be on-cell patch-clamped and whole-terminal voltage-clamped (Lemos et al. 1986; Stuenkel et al. 1990). Consistent with the calcium hypothesis of neurosecretion, one component of the inward current in these terminals has been shown to involve Ca²⁺ channels on the basis of Cd²⁺ sensitivity, resistance to TTX and evidence of run-down (Lemos et al. 1986). Given the relative scarcity of preparations in which terminals can be voltage-clamped and the interest in the biophysical and pharmacological properties of terminal Ca²⁺ channels, we have investigated I_Ca in sinus gland terminals.

The sinus gland of the crab Cardisoma carnifex Herbst was separated from the adjoining eyestalk tissue in normal crab saline (NCS), consisting of (in mmol l⁻¹): NaCl, 440; KCl, 11.3; CaCl₂, 13.3; MgCl₂, 26; Na₂SO₄, 23; Hepes, 10; pH adjusted to 7.4 with NaOH, and mechanically dissociated by trituration in defined medium (DM, as described in Cooke et al. 1989). The resulting isolated terminals were plated in Primaria culture dishes (Becton Dickinson) and left to adhere to the dish for a minimum of 1 h prior to experimentation.

The DM was exchanged with filtered extracellular solution immediately prior to experimentation. Any further changes in extracellular milieu were achieved by pressure-ejection of solutions onto individual terminals.

Voltage-clamp recordings were obtained in the whole-terminal patch-clamp configuration with the use of an EPC9 amplifier. Data acquisition and storage were performed by HEKA software (Instrutech) run on a Macintosh Centris 650. Analysis and graphics were achieved with Igor Pro software (Wavemetrics). Capacitance-compensated current signals were leak-subtracted with four scaled pulses obtained at hyperpolarizing command potentials (filtered with a four-pole Bessel filter, corner frequency 2.9 kHz). Pipettes used to obtain tight-seal whole-terminal recordings were pulled from Kimax thin-walled glass capillaries (1.5–1.8 mm o.d.) on a vertical puller (David Kopf Instruments, TW 150F-4). Pipettes were coated with dental wax to reduce capacitance and fire-polished with a microforge (Narishige, model MF-83). Typically pipettes filled with the intracellular solution and immersed in the bath had resistances ranging from 1.5 to 6 MΩ. All experiments were conducted at room temperature (24–26 °C).

I_Ca in terminals was isolated using an extracellular solution that contained (in mmol l⁻¹): N-methyl-D-glucamine methane sulphonate (NMG-MeSO₃), 240; MgCl₂, 24; CaCl₂, 52; Hepes, 10; CsCl, 100; tetraethylammonium bromide (TEABr), 20; 4-aminopyridine 3; TTX, 0.0005; pH 7.4. The intracellular solution, applied through the pipette contained (in mmol l⁻¹): NMG-MeSO₃, 200; CsCl, 100; NaCl, 10; MgATP, 5; bis-(o-aminophenoxy)-N,N,N′,N′-tetraacetic acid (BAPTA) (Cs⁺ salt), 20; Hepes, 50; TEABr, 10; pH 7.4. The concentration of chelator was varied in some experiments as indicated in the Results. In Ba²⁺ substitution experiments, NMG-MeSO₃ and CsCl were replaced with NaCl in the extracellular solution to prevent precipitation of Ba²⁺ salts. The tonicity of all solutions was adjusted with sucrose to 1100 mosmol l⁻¹.

A nifedipine (Sigma) stock solution dissolved in 95 % ethanol, stored in the dark at 4 °C, was diluted to a final concentration of 1 μmol l⁻¹ in external solution, protected from light during use and applied through pressure-ejection.

Since successful whole-terminal recordings were acquired infrequently, it proved impractical to test limited supplies of conotoxins and agatoxin on isolated terminals. Therefore, we examined the potential effects of these toxins on terminal I_Ca indirectly, by monitoring the Ca²⁺-dependent release of crustacean hyperglycemic hormone (CHH) from an isolated nerve tract–sinus gland preparation (as previously described; Keller et al. 1994).

Release of CHH was chosen as it and its co-localized gene products represent 90 % of the peptide content of the sinus gland and account for the bulk of peptide released in response to high-K⁺ stimulation (Stuenkel and Cooke, 1988). CHH is present in the majority of the X-organ neurons and terminals of the sinus gland (Dircksen et al. 1988). A sensitive enzyme-linked immunosorbent assay (ELISA) has been developed to detect levels of CHH release from individual sinus glands. CHH release (Keller et al. 1994) and the secretion of other peptides (Stuenkel and Cooke, 1988) from isolated X-organ–sinus gland preparations have been shown to be abolished by removal of external Ca²⁺ or blocking with Mn²⁺ or Cd²⁺, indicating that Ca²⁺ influx through terminal voltage-operated Ca²⁺ channels (VOCCs) is necessary for CHH secretion.

The nerve tract was retained and pinned to the base of a small perfusion chamber to hold the sinus gland in place during release experiments. Constant perfusion of the micro-chamber was achieved with a peristaltic pump. Reduced-Ca²⁺ (5 mmol l⁻¹) and reduced-Mg²⁺ salines (2.6 mmol l⁻¹) were used in these experiments to facilitate toxin binding. Release was stimulated by perfusing a high-K⁺ saline (50 mmol l⁻¹) onto the preparation. Fractions of the perfusate were collected every 2 min and subsequently analyzed for CHH content using an ELISA (Keller et al. 1994).

50 μl of 5 mmol l⁻¹ Ca²⁺, 2.6 mmol l⁻¹ Mg²⁺ saline containing ω-conotoxin (ω-Ctx) GVIA (10 μmol l⁻¹), ω-agatoxin (ω-Aga) IVA (500 nmol l⁻¹) and ω-Ctx MVIIC (5 μmol l⁻¹) was applied directly to the 25 μl micro-chamber after the perfusion had been halted and was left on for 2 min. The perfusion system was then restarted and the sinus gland was stimulated with K⁺. The same protocol was used in control experiments with the omission of the toxins.

The acutely isolated terminals from X-organ neurons were examined under conditions designed to isolate Ca²⁺ currents and minimize outward current contamination. Breaking into the terminals following seal formation often caused the terminal to rupture, severely limiting the number of successful whole-terminal voltage-clamp recordings. The quantity of releasable peptide contained in successfully voltage-clamped terminals was not determined.

From a holding potential (V_hold) of −50 mV, 10 mV incrementing depolarizing steps beyond −40 mV produced inward currents in the terminals which peaked at approximately +20 mV, which is typical of high-voltage-activated (HVA) I_Ca (Fig. 1A). An equation incorporating a Boltzmann and a linear term was used to fit the normalized current–voltage (I/V) data from six terminals and gave a value for the V_1/2 of activation of +3.7 mV, a valence (z) of 3.5e and a reversal potential (E_Ca) of +69.7 mV. A hyperpolarizing prepulse of 200 ms to −90 mV failed to reveal more current (Fig. 1B), and the characteristics of the terminal Ca²⁺I/V relationship (V_1/2=+4.2 mV, z=3.8e, E_Ca=62.1 mV) showed no significant difference from those at a V_hold of −50 mV, implying that the terminals do not have a transient, low-voltage-activated I_Ca. The I/V relationships from acutely isolated terminals are indistinguishable from those previously obtained from X-organ somata (Richmond et al. 1995), which also lack a low-threshold-activated Ca²⁺ current.

The average peak amplitude for the terminal I_Ca in 52 mmol l⁻¹ Ca²⁺ was 43.2±4.2 pA with an average membrane capacitance of 4.8±0.7 pF, giving a current density of 10.5±4.9 μA cm⁻² (mean ± S.E.M., N=9, assuming a capacitance of 1 μF cm⁻² for the cell membranes). Interestingly, the current density of acutely dissociated X-organ somata in 52 mmol l⁻¹ Ca²⁺ has previously been shown to be 28.8±6.3 μA cm⁻² (N=6), more than twice that of the terminals (Richmond and Penner, 1994).

The terminal I_Ca showed considerable time-dependent inactivation immediately following the peak. The decay was best fitted in most traces by a single exponential function. Fig. 2A plots the time constant (τ) of inactivation as a function of the command potential. The τ of inactivation was fastest at +20 mV (18.6±1.22 ms, N=6), the potential producing maximal Ca²⁺ influx, suggesting that inactivation is current-dependent. To test the Ca²⁺-dependence of inactivation, a double-pulse protocol was used, in which the effect of varying a 200 ms prepulse potential from −80 mV to +60 mV on the test current amplitude at +20 mV was investigated. The average test-pulse I_Ca amplitude (normalized to the peak amplitude with a prepulse of −80 mV) was plotted against prepulse potential (Fig. 2B). Maximal inactivation coincided with peak Ca²⁺ entry, with partial recovery at more positive potentials, indicative of Ca²⁺-dependent inactivation. This was further investigated by substituting 52 mmol l⁻¹ Ba²⁺ or Sr²⁺ for Ca²⁺ as the charge carrier. Ba²⁺ largely removed inactivation, and Sr²⁺ caused a marked reduction in the extent of inactivation compared with Ca²⁺ (Fig. 2C). A comparison of the double-pulse inactivation curves from three terminals illustrates the relative current-dependent inactivation produced by the different divalent cations (Fig. 2D).

We also examined the effects on inactivation of altering the level of Ca²⁺ chelator in the internal pipette solution. We compared the percentage of current remaining at the end of a 50 ms step depolarization to +20 mV in terminals loaded with 0 chelator, 5 mmol l⁻¹ EGTA, 5 mmol l⁻¹ BAPTA or 20 mmol l⁻¹ BAPTA. There was no significant difference between the extent of inactivation in terminals loaded with 5 mmol l⁻¹ EGTA, 5 mmol l⁻¹ BAPTA or 20 mmol l⁻¹ BAPTA; therefore, these results were grouped and compared with the inactivation in 0 chelator. As exemplified in Fig. 3A,B, inactivation was significantly higher in terminals lacking intracellular chelator compared with terminals with either BAPTA or EGTA present (P=0.0002, Student’s t-test). A possible explanation for the similarity of action of BAPTA and EGTA is that, at such high concentrations of each (5 mmol l⁻¹), the faster binding kinetics and higher affinity of BAPTA over EGTA are no longer discernible. In support of this speculation, 20 mmol l⁻¹ BAPTA was no more effective in reducing inactivation than 5 mmol l⁻¹ BAPTA. In Fig. 3B, the average percentage of inactivation at the end of a 50 ms pulse in the presence or absence of chelator is summarized.

The effects of the L-type Ca²⁺ channel blocker nifedipine were tested on I_Ca elicited by a command potential to +20 mV. Nifedipine at a concentration of 1 μmol l⁻¹ had no effect on the kinetics (Fig. 4A) or the peak amplitude of the terminal I_Ca (Fig. 4B). To ensure that the negative result with nifedipine was not due to the voltage-dependence of blocking effects, a standing holding potential of −10 mV was implemented prior to ramp commands from −100 to +100 mV. Under these circumstances, the current amplitude over the entire voltage range of the ramp command was unaffected by nifedipine.

These results indicate that the terminal Ca²⁺ channels are not of the dihydropyridine-sensitive L-type.

Owing to the very low success rate in obtaining whole-cell recordings from terminals, we used an indirect approach to test for the presence of N-and P/Q-type Ca²⁺ channels. The isolated sinus gland was used to measure Ca²⁺-dependent release of CHH from terminals. By monitoring the K⁺-stimulated release of CHH in control and toxin-treated preparations, we were able to study, indirectly, whether any block of Ca²⁺ channels occurred in CHH-containing terminals. ω-Ctx GVIA (10 μmol l⁻¹) was used to block N-type channels in combination with ω-Aga IVA (500 nmol l⁻¹) and ω-Ctx MVIIC (5 μmol l⁻¹) to block P/Q-type channels. In both control and toxin-treated terminals, CHH release was stimulated twice with 50 mmol l⁻¹ K⁺ and the second release amplitude was normalized to the first. The toxin combination was added to the chamber containing the terminals for 2 min in a saline containing low concentrations of divalent cations prior to the second K⁺ stimulation. Averages of three control and three toxin-treated preparations revealed no significant difference in the extent of CHH release (Fig. 5). These data suggest that the Ca²⁺ channels involved in CHH release from the terminals are insensitive to a combination of mammalian N-and P/Q-type channel blockers. In one experiment, it was also shown that both the L-type channel agonist BayK and the blocker nifedipine (1 μmol l⁻¹) had no effect on CHH secretion compared with controls.

We have examined I_Ca in the isolated peptidergic terminals of the crab neurohemal organ, the X-organ–sinus gland. As demonstrated by the I/V relationship, these terminals have high-voltage-activated (HVA) Ca²⁺ channels, with no evidence of low-threshold I_Ca. Both the features of the I/V relationship and the dominant Ca²⁺-dependent inactivation of the terminal Ca²⁺ channels closely resemble those of the X-organ somata from which the terminals originate (Meyers et al. 1992; Richmond et al. 1995).

The heterogeneity of HVA Ca²⁺ channels has been demonstrated in an increasingly large number of neurons. In many instances, the voltage-dependent properties of these channels are sufficiently similar that the whole-cell current of mixed populations of channel types appears as a single V-shaped I/V relationship, as seen in the sinus gland terminals. Pharmacological tools have been useful in some preparations to reveal the presence of heterogeneous channel populations. In some terminal preparations, mixed populations of HVA channels have been found; for example, the neurohypophyseal terminals contain L-type, N-type and P-type channels (Lemos et al. 1994). The I_Ca in chick ciliary ganglia is insensitive to dihydropyridines and is predominantly, although incompletely, blocked by ω-Ctx GVIA (Yawo and Momiyama, 1993). In retinal bipolar cells, however, the somata and terminals both appear to contain only L-type Ca²⁺ channels (Heidelberger and Matthews, 1992). In the present study, no effects of specific Ca²⁺ channel blockers and toxins were observed. Similar results were obtained previously in the X-organ somata (Richmond et al. 1995). The lack of any pharmacological distinction between terminal and somata I_Ca in the X-organ–sinus gland raises the possibility that these neurons have a homogeneous population of Ca²⁺ channels. However, this possibility remains tentative until specific blockers for these channels become available. Furthermore, while toxins have been tested directly on the somata of X-organ neurons, the paucity of successful terminal recordings necessitated an indirect assessment of terminal VOCC pharmacology through CHH secretion. If mixed populations of VOCCs exist in these terminals and each subtype is able to support CHH release fully, then inhibition of one channel type would not be expected to alter the level of CHH release. However, since we have shown that the terminals do not respond to the L-type antagonist nifedipine and that in the simultaneous presence of N-and P/Q-type blockers release is unaffected, at least one VOCC subtype supporting release of CHH must be pharmacologically distinct from these channel types.

Few studies have addressed the pharmacology of crustacean Ca²⁺ channels. Similarly negative results for the effects of nifedipine were found on crab pyloric neurons (Golowasch and Marder, 1992) and the lobster neuromuscular junction (Grossman et al. 1991), although the crayfish giant axon reportedly has a nifedipine-sensitive I_Ca (Nishio et al. 1993). Indirect evidence, based on a reduction in synaptic transmission, indicates that Ca²⁺ channels sensitive to ω-Ctx GVIA and ω-Aga IVA are present at presynaptic loci of the lobster abdominal neuromuscular junction (Grossman et al. 1991) and the crayfish leg opener muscle (Araque et al. 1994), respectively. It appears, therefore, that crustacean neuromuscular terminals have Ca²⁺ channels pharmacologically similar to the N-type channels found at several vertebrate synapses (Dayanithi et al. 1988; Hirning et al. 1988) and the P-type channels found in other invertebrate and mammalian synapses (Llinás et al. 1989, 1992; Uchitel et al. 1992). Since no effect of a combined N-and P/Q-type block was observed on CHH release, the pharmacological profiles of the X-organ, CHH-secreting terminals may differ from those found at crustacean neuromuscular synapses.

It has been shown previously using intracellular electrodes that Ca²⁺-dependent depolarization is restricted to the sinus gland terminals and somata and is not detectable in the adjoining axons of X-organ neurons (Nagano and Cooke, 1987), as has been found in other systems (Stockbridge and Ross, 1984). Interestingly, the isolated X-organ somata have more than double the channel density of the terminals (see Richmond et al. 1995) in contrast to the goldfish retinal bipolar cells where the reverse pattern is observed (Heidelberger and Mathews, 1992). Whereas blockade of the terminal Ca²⁺ channels inhibits CHH release (Keller et al. 1994), supporting their role in secretion, the functional significance of a high Ca²⁺ channel density in the somata is less well understood (Richmond et al. 1995). It is also unclear whether the Ca²⁺ channels of the terminals are equally distributed on the terminal membrane or are clustered at release sites, giving a locally higher current density within microdomains, as has been found at the squid giant synapse (Llinás et al. 1992).

The sinus gland terminals exhibit several characteristics which indicate that the predominant mode of Ca²⁺ channel inactivation is Ca²⁺-rather than voltage-dependent. The U-shaped relationship between prepulse potential and test-pulse current amplitude is a characteristic of Ca²⁺-dependent inactivation. The Ca²⁺ influx during the prepulse produces inactivation and hence a reduction in test-pulse amplitude which is greatest at prepulse potentials producing the largest I_Ca (approximately +20 mV in these terminals). Similarly, the rate of inactivation exhibits a voltage-dependence, peaking at +20 mV. Substituting Sr²⁺ or Ba²⁺ for Ca²⁺ as the charge carrier progressively attenuated the extent of inactivation in the terminals, further indicating that inactivation is primarily regulated by a Ca²⁺-dependent mechanism in which Sr²⁺ but not Ba²⁺ can partially imitate Ca²⁺. Finally, a small but significant (P=0.0002) increase in the extent of inactivation was observed when the terminals were internally dialyzed without the Ca²⁺ chelators BAPTA or EGTA.

The presence of Ca²⁺-dependent inactivation is expected to have important functional consequences in terms of regulating the secretory output of the terminals, in that it would act as a negative feedback loop closing Ca²⁺ channels that have been extensively stimulated. The CHH-containing neurons exhibit sustained release, which is somewhat at odds with the idea that their terminals may exhibit Ca²⁺-dependent inactivation (Keller et al. 1994). One explanation for this discrepancy may be that although Ca²⁺-dependent inactivation was present in all terminals studied under voltage-clamp none of these terminals was of the CHH phenotype. Interestingly, it is known that the release of the much less prevalent (3 % total peptide content) red pigment concentrating hormone rapidly inactivates in the presence of a maintained secretory stimulus (Stuenkel and Cooke, 1988).

Alternatively, we have previously shown that, in identified CHH-containing X-organ somata, hyperpolarization partially reverses Ca²⁺-dependent inactivation (Richmond et al. 1995). Therefore, hyperpolarization during the repolarizing phase of the action potentials known to occur in the terminals during secretion could alleviate inactivation. In addition, the physiological Ca²⁺ concentration of the crab hemolymph of 13 mmol l⁻¹ would be expected to produce less Ca²⁺-dependent inactivation than that observed in the present experiments in which [Ca²⁺] was elevated fourfold to enhance current amplitude. This, combined with the unknown capacity of the intact terminal to sequester Ca²⁺, may allow sufficiently prolonged I_Ca activation to account for the sustained pattern of CHH release observed.

We thank Beverly Haylett and Barbara Reichwein for participation in CHH secretion experiments, Jana Labenia and Jacqueline Tellei for technical assistance, David Featherstone for generating analysis macros and Emanuele Sher and Marc Rogers for critical reading of the manuscript. This work was supported by NIH grant RO1 NS-15453 and NSF grant BNS 89 10432 (to I.M.C.), by funds from the Ida Russell Cades Fund of the University of Hawaii, by the Hermann und Lilly-Schilling-Stiftung to R.P. and by grants from the Deutsche Forshungsgemeinschaft to R.K.