JN Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Neurophysiol 78: 2372-2387, 1997;
0022-3077/97 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Scott, M. L.
Right arrow Articles by Weiss, K. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Scott, M. L.
Right arrow Articles by Weiss, K. R.

The Journal of Neurophysiology Vol. 78 No. 5 November 1997, pp. 2372-2387
Copyright ©1997 by the American Physiological Society

Ion Currents and Mechanisms of Modulation in the Radula Opener Muscles of Aplysia

Marsha L. Scott, Vladimir Brezina, and Klaudiusz R. Weiss

Department of Physiology and Biophysics and The Fishberg Research Center in Neurobiology, Mount Sinai School of Medicine, New York, New York 10029

    ABSTRACT
Abstract
Introduction
Methods
Results
Discussion
References

Scott, Marsha L., Vladimir Brezina, and Klaudiusz R. Weiss. Ion currents and mechanisms of modulation in the radula opener muscles of Aplysia. J. Neurophysiol. 78: 2372-2387, 1997. Numerous studies of plasticity in the feeding behavior of Aplysia have shown that substantial plasticity is due to peripheral neuromodulation of the feeding musculature. Extensive previous work focusing on the accessory radula closer (ARC) muscle has led to the realization that a major function of the modulation in that muscle may be to ensure efficient coordination between its contractions and those of its antagonist muscles. For a more complete understanding, therefore, we must study these muscles also. Here we have studied the radula opener muscles I7-I10. Using single isolated muscle fibers under voltage clamp, we have characterized ion currents gated by voltage and by the physiological contraction-inducing neurotransmitter acetylcholine (ACh) and the effects of the physiological modulators serotonin, myomodulins A and B, and FMRFamide. Our results explain significant aspects of the electrophysiological behavior of the whole opener muscles, as well as why the opener and ARC muscles behave similarly in many ways yet differently in some key respects. Opener muscles express four types of K currents: inward rectifier, A-type [IK(A)], delayed rectifier [IK(V)], and Ca2+-activated [IK(Ca)]. They also express an L-type Ca current [ICa] and a leakage current. ACh activates a positive-reversing cationic current [IACh(cat)] and a negative-reversing Cl current [IACh(Cl)]. The opener muscles differ from the ARC in that, in the openers, activation of IK(A) occurs ~9 mV more positive and there is much less IACh(Cl). In both muscles, IACh(cat) most likely serves to depolarize the muscle until ICa activates to supply Ca2+ for contraction, but further depolarization and spiking is opposed by coactivation of IK(A), IK(V), IK(Ca), and IACh(Cl). Thus the differences in IK(A) and IACh(Cl) may well be key factors that prevent spikes in the ARC but often allow them in the opener muscles. As in the ARC, the modulators enhance ICa and so potentiate contractions. They also activate a modulator-specific K current, which causes hyperpolarization and depression of contractions. Finally, in the opener muscles but not in the ARC, the modulators activate a depolarizing cationic current that may help phase-advance the contractions. Each modulator exerts these effects to different degrees and thus has a distinct effect on voltage and contraction size and shape. The overall effect then will depend on the specific combinations of modulators released in different behaviors. By understanding the modulation in the opener muscles, as well as in the ARC, we are now in a position to understand how the behavior of the two muscles is coordinated under a variety of circumstances.

    INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References

A major goal of neuroscience is to understand the plasticity of behavior. In recent years, considerable insight into the mechanisms as well as the integrative significance of behavioral plasticity has come from studies of relatively simple behaviors in tractable invertebrate preparations. This has been possible because even quite stereotyped, rhythmic behaviors exhibit great plasticity when required. The central pattern generator (CPG) for the behavior may itself be plastic, but it has now become clear that another major site of plasticity is in the periphery, at the neuromuscular junctions and muscles that convert the output of the CPG into behavior (e.g., Bishop et al. 1987; Brezden et al. 1991; Erxleben et al. 1995; Evans and Myers 1986; Groome et al. 1994; Jorge-Rivera and Marder 1996; Laurienti and Blankenship 1997; Muneoka and Kamura 1982; Van Golen et al. 1996; Zoran et al. 1989; reviewed by Calabrese 1989). This has been studied most extensively in Aplysia buccal muscles (e.g., Church et al. 1993; Cropper et al. 1987a, 1990a, 1994; Fox and Lloyd 1995, 1996; Lloyd et al. 1984; Weiss et al. 1978), which generate consummatory feeding movements.

Feeding in Aplysia consists of alternating protraction and retraction, coordinated with opening and closing, of the radula, a hand-like organ that grasps food (Kupfermann 1974). These movements exhibit plasticity with changes in the motivational state of the animal---most notably, they become faster and stronger during food-induced arousal (Kupfermann 1974; Susswein et al. 1978)---or in response to variations in the qualities of the food (Hurwitz and Susswein 1992). In extensive studies focusing on one representative buccal muscle, the accessory radula closer (ARC or I5) muscle (Cohen et al. 1978), much of the plasticity has been found to be due to the actions of numerous neuromodulators released as cotransmitters from the muscle's motoneurons or from modulatory neurons that also innervate the muscle (reviewed in Vilim et al. 1996a,b; Weiss et al. 1992). A great deal is now known about the mechanisms by which these modulators act within the ARC muscle to modify its contractions. However, this work also has suggested that a major function of the modulation may be to coordinate the contractions of the ARC with those of its antagonist muscles. Specifically, by altering the relationship between contraction size and duration, the modulators may preserve the energetically efficient alternation of contractions of opposing muscles when the feeding movements increase in strength and frequency (Weiss et al. 1992). For a more complete understanding, therefore, we must study not just the ARC but also its antagonist muscles.

A complex of radula opener muscles, I7-I10, recently has been identified (Evans et al. 1996; Scott et al. 1991). In many respects, these muscles resemble the ARC, but there are also some clear differences. Both muscles contract when depolarized by acetylcholine (ACh), the primary transmitter of their motoneurons (Cohen et al. 1978; Evans et al. 1996). However, whereas the ARC depolarizes and contracts in a purely graded fashion, the opener muscles can generate spikes. (They also can depolarize, generate spikes, and contract in response to stretch.) The main modulators in the ARC system are serotonin (5-HT) and peptides of the small cardioactive peptide (SCP), myomodulin (MM), buccalin, and FRF/FMRFamide families (e.g., Brezina et al. 1995; Cropper et al. 1987a,b, 1988, 1994; Lloyd et al. 1984; Weiss et al. 1978). Except for the SCPs and buccalins, the same modulators have been detected in the opener system (Evans et al. 1993; C. G. Evans, F. S. Vilim, E. C. Cropper, and K. R. Weiss, unpublished data). On the ARC muscle, the actions of the modulators can best be categorized (Brezina et al. 1995) into three primary effects: adenosine 3',5'-cyclic monophosphate (cAMP)-dependent potentiation of contraction amplitude (Cropper et al. 1990a, 1991; Lloyd et al. 1984; Weiss et al. 1978, 1979); hyperpolarization and depression of contraction amplitude (Cropper et al. 1991, 1994); and cAMP-dependent acceleration of the relaxation rate of the contractions (Cropper et al. 1987a,b; Probst et al. 1994). Different modulators have different combinations of these three effects. Thus 5-HT and the MMs elevate cAMP, potentiate contractions, and accelerate the relaxation rate. Simultaneously, however, they also depress the contractions to different degrees. Some (5-HT, MMB) depress the contractions only slightly, yielding net potentiation; others (MMA) depress much more, resulting in net potentiation at low modulator concentrations and net depression at higher concentrations. The FRF/FMRFamide peptides do not elevate cAMP and are pure depressors. In all of these respects, these modulators appear to act on the opener muscles just as they do on the ARC muscle (Evans et al. 1993; C. G. Evans, F. S. Vilim, E. C. Cropper, and K. R. Weiss, unpublished data). However, 5-HT and the MMs (but not FMRFamide) also depolarize the opener muscles, a fourth effect that is not seen in the ARC muscle.

Our understanding of the behavior of the whole ARC-muscle system has been advanced greatly by cellular-level studies of single isolated ARC muscle fibers that examined their intrinsic electrophysiological properties, excitation-contraction mechanisms, and actions of the modulators (Brezina and Weiss 1995a,b; Brezina et al. 1994a-e; Cropper et al. 1994; Kozak et al. 1996). For example, the potentiation of contractions has been found to be mediated, almost certainly, by cAMP-dependent enhancement of the Ca current in the muscle (Brezina et al. 1994d) and the depression of contractions by activation of a modulator-specific K current (Brezina et al. 1994e; Cropper et al. 1994). The acceleration of relaxation rate appears to be a cAMP-dependent effect on the contractile machinery (Probst et al. 1994). This knowledge has allowed higher-level insights into the integrative significance of the modulation (Brezina et al. 1995, 1996). Before the ARC and the opener muscles can be studied together as a coordinated neuromuscular system, similar knowledge will be required about the opener muscles. In this paper, we examine single isolated opener muscle fibers under voltage clamp. We characterize their basal (unmodulated) ion currents as well as currents activated by ACh and the physiological modulators 5-HT, MMA, MMB, and FMRFamide. As expected from the similarities between the whole opener and ARC muscles, many of the currents are like those that were described in the ARC fibers. We therefore use the ARC findings as a point of departure, emphasizing primarily the differences between the ion currents of the two muscles. Both the similarities and the differences are important in explaining how the two muscles behave similarly in many ways, yet differently in some key respects.

Abstracts of this work have appeared (Scott et al. 1995, 1996).

    METHODS
Abstract
Introduction
Methods
Results
Discussion
References

Isolated opener muscle fibers

Opener muscles were dissected from 100-300 g Aplysia californica (Marinus, Long Beach, CA; Aplysia Resource Facility of the University of Miami, FL; or Marine Specimens Unlimited, Pacific Palisades, CA), and single fibers were isolated by collagenase treatment (Brezina et al. 1994a). Generally, the whole opener muscle complex I7-I10 connected by the collostylar cap (Evans et al. 1996; Scott et al. 1991) was treated together as a unit, as in preliminary experiments we saw no obvious differences between fibers isolated separately from the various component muscles.

In some experiments, fibers were immobilized in agarose gel for mechanical stability (Brezina et al. 1994a), but most often fibers were held suspended by the recording electrode in free solution (cf. Brezina 1994; Brezina et al. 1994a). In all experiments, solution was superfused continuously at 1-3 ml/min at room temperature (20-24°C).

Voltage-clamp techniques

These were essentially as in the ARC studies (Brezina et al. 1994a-e). Fibers were impaled with a single 12-25 MOmega sharp microelectrode containing 3 M K-acetate plus 30 mM KCl to prevent drift. The bath was grounded with an Ag/AgCl pellet via a 3 M KCl agar bridge. The discontinuous single-electrode voltage-clamp mode was used. Currents were probed, as appropriate, with slow (14 mV/s) voltage ramps or with voltage steps. Any disparity between the nominal (commanded) and the actual membrane voltage was corrected by recording the actual voltage in parallel with the membrane current and using the actual voltage in constructing I-V relations. Ramps were filtered at 20-30 Hz, steps at 0.3-1 kHz. Further filtering was applied during generation of figures, and in cases where successive traces were essentially identical, up to 10 traces were sometimes averaged.

A primary concern with voltage-clamp methods is the issue of space clamp---the ability to adequately clamp the membrane potential over the entire cell surface, something that may be particularly difficult with long, thin cells such as muscle fibers. Brezina et al. (1994a) discussed this issue and performed several tests of space clamp in the ARC fibers. Space clamp appeared to be quite good except perhaps when current amplitudes were very large. This problem was limited and space clamp was generally optimized by selecting smaller and shorter fibers for study. In the present study, the range of sizes of opener fibers used (400-900 µm long, 10-25 µm in diameter) was similar to that in the ARC studies. Furthermore, the I-V relations of opener fibers were found to be very similar to those of ARC fibers, and current amplitudes similar or smaller (see RESULTS). Therefore the quality of space clamp achieved in the present experiments was considered to be adequate.

Solutions and drugs

Normal artificial sea water (ASW) contained (in mM) 460 Na+, 55 Mg2+, 11 Ca2+, 10 K+, 602 Cl-, and 10 N-(2-hydroxyethyl)piperazine-N'-(2-ethanesulfonic acid; HEPES) buffer at pH 7.6. Other solutions were made by equimolar substitution of this formula. Na+ was replaced completely with N-methyl-D-glucamine (0 Na ASW); Ca2+ was replaced completely with extra Mg2+ (0 Ca ASW); K+ was elevated at the expense of Na+ (100 K ASW). Sometimes two of these substitutions were made simultaneously (0 Na/0 Ca ASW, 100 K/0 Na ASW). Cl- was reduced from 602 to 142 mM by replacement with isethionate (142 Cl ASW). Ca/tetraethylammonium (TEA)/4-aminopyridine (4-AP) ASW, Ba/TEA/4-AP ASW, and Co/TEA/4-AP ASW were made by replacing all Na+ with TEA, Ca2+ with Ba2+ or Co2+ as appropriate, and adding 10 mM 4-AP. Finally, 380 Ca ASW consisted of 380 Ca2+, 10 K+, and 10 HEPES. Other chemicals and drugs (including 4-AP, <= 10 mM, and TEA, <= 50 mM) were simply added without substitution. Chemicals were obtained from Sigma, except forskolin (Calbiochem) and the peptides [Peninsula (Belmont, CA), Applied Biosystems (Foster City, CA), or Nuros (San Jose, CA)].

Group data are presented as means ± SE.

    RESULTS
Abstract
Introduction
Methods
Results
Discussion
References

Basic properties of opener muscle fibers

Single fibers isolated from the opener muscle complex were smooth and spindle-shaped, tapering toward the ends, although some fibers bifurcated or had wide, flat ends. This basic appearance was already noted by Evans et al. (1996); it is very similar to that of fibers from the ARC (Brezina et al. 1994a) and other molluscan smooth muscles (e.g., Mytilus: Ishii and Takahashi 1982; Philine: Dorsett and Evans 1991; A. brasiliana: Laurienti and Blankenship 1996a). The range of fiber sizes was large (length 100-2,000 µm, diameter 5-25 µm), although only the shorter fibers (400-900 µm in length, 10-25 µm in diameter) were used for experiments.

The opener complex is small compared with the ARC muscles. The wet weight of the ARC muscles relative to the opener complex was roughly 3:1 (29 ± 2.2 vs. 11 ± 0.7 mg, n = 10). Since the single opener fibers were as large as or perhaps only slightly smaller than ARC fibers (cf. Brezina et al. 1994a), the opener complex probably contains many fewer fibers than the ARC muscle. (However, an unknown, and possibly different, fraction of the weight of the muscles is due to a connective-tissue matrix.) Indeed, on dissociation, each opener complex yielded many fewer fibers than did an ARC muscle.

Most opener fibers contracted when injected with sufficient depolarizing current and relaxed as the membrane potential returned to rest. The resting membrane potential (-77 to -55 mV) was similar to the values measured in the intact opener muscle (Evans et al. 1996) as well as in ARC fibers (Brezina et al. 1994a,c).

Basal currents

TOTAL CURRENT-VOLTAGE RELATIONS OF OPENER FIBERS. The quasi-steady-state total current-voltage (I-V) relation of single opener fibers typically was S-shaped, with a region of inward rectification below about -70 mV and a region of outward rectification above about -50 mV, separated by a plateau region of high, or occasionally even negative, slope resistance (Figs. 1, 7, 9, and 10). These features were essentially the same as in ARC fibers (Brezina et al. 1994a), already suggesting that the underlying currents were probably very similar in the two muscles.


View larger version (17K):
[in this window]
[in a new window]
  		FIG. 1. Inwardly rectifying K current, IK(IR). I-V relations were generated using slow voltage ramps from nominally -100 to -30 (A) or 0 mV (B). K+ and Ba2+ concentrations are in mM. A: typical I-V relations recorded in normal extracellular K+ (10 K) before and after addition of 1 mM Ba2+. The net Ba2+-sensitive difference current (Delta ), taken to be IK(IR), reversed in this experiment at -83 mV, rectified inwardly below -70 mV, but turned off at more positive voltages, forming a plateau region in the I-V relation. B: in elevated extracellular K+ (100 K), IK(IR) was much larger in amplitude and its reversal potential was shifted positive, close to the new value of EK predicted by the Nernst equation. At 100 K, IK(IR) was not completely blocked by 1 mM Ba2+ but was blocked by 10 mM Ba2+.


View larger version (14K):
[in this window]
[in a new window]
  		FIG. 7. Acetylcholine (ACh)-activated cationic and Cl currents, IACh(cat) and IACh(Cl). I-V relations were generated using slow voltage ramps from nominally -100 to 0 mV in normal ASW. Application of 10 µM ACh activated a large inward current that reversed relatively positive, was blocked almost completely by further addition of 100 µM hexamethonium, and so was predominantly IACh(cat). In this experiment, there was a small residual component of hexamethonium-insensitive, negative-reversing IACh(Cl).


View larger version (24K):
[in this window]
[in a new window]
  		FIG. 9. Modulator-activated K and cationic currents,IMod(K) and IMod(cat). In all of these experiments, I-V relations were generated using slow voltage ramps from nominally -100 to -30 mV in normal ASW. A: voltage ramps were given every 10 s while 10 µM 5-HT and then also 10 µM FMRFamide (FMRFa) were applied. A1 plots the absolute current amplitudes measured at the lowest and highest voltages, nominally -100 and -30 mV, reached during successive ramps; A2 shows individual I-V relations obtained at the times indicated in A1. 5-HT activated mainly IMod(cat), FMRFamide only IMod(K). B: IMod(K) and IMod(cat) activated by a single modulator. One hundred nanomolar MMA activated both IMod(K) and IMod(cat), with an intermediate combined reversal potential. When the MMA concentration was then increased to 1 µM, the additional current was mainly IMod(K), with a negative reversal potential. However, this current desensitized, revealing IMod(cat), with a positive reversal potential. C: 1 mM CPT-cAMP plus 100 µM Ro 20-1724 had relatively little effect, and in their presence, 10 µM 5-HT still was able to activate IMod(cat) as usual.


View larger version (12K):
[in this window]
[in a new window]
  		FIG. 10. Ionic basis of IMod(K) and IMod(cat). I-V relations were generated using slow voltage ramps from nominally -100 to -30 (A) or 0 mV (B). A: IMod(cat) was carried primarily by Na+. Effect of 10 µM 5-HT in normal ASW (A1) and, in the same fiber, in 0 Na ASW (A2). Note the large negative shift in reversal potential. The 5-HT-activated current remaining in 0 Na ASW was probably near-pure IMod(K). B: IMod(K) was carried by K+. In 100 K/0 Na ASW (plus 30 mM Cs+ to block IK(IR)) (see Brezina et al. 1994e), IMod(K) activated by 10 µM FMRFamide reversed much more positive than in normal, 10 mM K+ ASW (cf. e.g., Fig. 9A2).

HYPERPOLARIZATION-ACTIVATED, INWARDLY RECTIFYING K CURRENT. As in ARC fibers, the inwardly rectifying and the plateau regions of the I-V relation were formed by a classical inwardly (anomalously) rectifying K current, IK(IR) (Fig. 1). Its characteristics were indistinguishable from those of IK(IR) in ARC fibers (Brezina et al. 1994a). It was blocked selectively by 1-10 mM Ba2+ (Fig. 1) or <30 mM Cs+. Below about -70 mV, the current increased linearly with hyperpolarization. Between -70 and -50 mV, the current turned off with depolarization, creating the high-resistance plateau region in the I-V relation. In normal 10 mM K+ solution, the current reversed at -85 ± 1.2 mV (n = 6). [This provided a good estimate of the normal value of the K+ equilibrium potential, EK, in the opener fibers (cf. Brezina et al. 1994a).] When extracellular K+ was elevated to 100 mM, the reversal potential shifted much more positive (in Fig. 1B, for example, to -33 mV), as expected for a current carried mainly by K+. Hyperpolarizing voltage steps revealed no obvious time-dependent kinetics of activation, and no inactivation during steps as long as 3 s (e.g., Fig. 11A). IK(IR) amplitudes ranged between 0.1 and 0.4 nAat -100 mV.


View larger version (23K):
[in this window]
[in a new window]
  		FIG. 11. Voltage-dependent kinetics of IMod(K) and IMod(cat). All experiments were done in normal ASW. A: currents before, during, and after a 3-s-long hyperpolarizing step from -40 to -120 mV under control conditions, after addition of 10 µM 5-HT, and after further addition of 10 µM FMRFamide. 5-HT activated IMod(cat), which appeared as an inwardly relaxing current at -120 mV, whereas FMRFamide activated IMod(K), the outward current at -40 mV. B: voltage-dependent inward relaxation of IMod(cat) on hyperpolarization. These are net 5-HT-activated difference currents from an experiment like that in A with steps from -40 mV to a range of hyperpolarized potentials. C: voltage-dependent outward relaxation of IMod(K) on depolarization. C1 shows currents elicited by a depolarizing step from -90 to 0 mV before and after addition of 10 µM FMRFamide; C2 shows the net FMRFamide-activated difference current. D: experiment as in A (except with 2-s-long steps). Ten micromolar µM FMRFamide, then also 100 µM 4-AP, were applied.

DEPOLARIZATION-ACTIVATED CURRENTS. The total current activated by depolarizing voltage steps from a negative holding potential such as -90 mV had complex kinetics (e.g., Figs. 2A1, 3, and 5A1), suggesting the presence of multiple components. Indeed, by altering the holding potential, changing the ionic composition of the extracellular solution, and using specific pharmacological blockers, we were able to dissect the total current into a number of components: three outward K currents, specifically a fast transient A-type current, a slower delayed rectifier, and a Ca2+-activated K current; an inward Ca current; and a leakage current. Normally, the large outward K currents completely masked the smaller inward Ca current, creating the outward rectification at positive voltages in the total I-V relation.


View larger version (28K):
[in this window]
[in a new window]
  		FIG. 2. A-type and delayed-rectifier K currents, IK(A) and IK(V), and the leakage current, ILeak. All of these experiments were done in Ca2+-free solution to eliminate ICa and IK(Ca). A1: two distinct peaks were evident in currents elicited by steps from a holding potential of -90 mV to a range of depolarized potentials in 0 Ca artifical sea water (ASW). A2: shifting the holding potential to -40 mV preferentially inactivated the faster current, IK(A), leaving the slower current, IK(V) (plus ILeak). A3: difference currents obtained by subtraction of the currents in A2 from those in A1. The current inactivated by the shift in holding potential was mainly IK(A) but evidently some IK(V) too. A4: ILeak, recorded with the same steps as in A1 but in Co/tetraethylammonium (TEA)/4-aminopyridine (4-AP) ASW, in which IK(A) and IK(V) were blocked. B: I-V relations of the net (leak-subtracted) IK(A) and IK(V), and of ILeak, from the records in A, 1-4.

Again, this complement of currents was the same as that found in ARC fibers (Brezina and Weiss 1995a; Brezina et al. 1994a-c), as were the major properties of each current, except for one notable difference in the A-type K current.

A-TYPE K CURRENT, IK(A). This current was responsible for the early outward peak of depolarization-activated current seen in records such as those in Fig. 2A1, obtained in Ca2+-free solution. Thus IK(A) did not depend on extracellular Ca2+. In Ca2+-containing solution, however, a second component of the early peak was contributed by IK(Ca) (see e.g., Fig. 5A1 and below).


View larger version (17K):
[in this window]
[in a new window]
  		FIG. 5. Ca2+-activated K current, IK(Ca). A1: currents elicited by depolarizing steps from -90 to 0 mV in normal ASW (Ca) and 0 Ca ASW. (Note that the latter current is the composite current IK(A) + IK(V) + ILeak that was already studied in Fig. 2A.) A2: subtraction of the two records in A1 yields the net Ca2+-sensitive current, ICa + IK(Ca). B1: currents elicited as in A1 but in the presence of 10 mM TEA, which blocked IK(Ca). B2: subtraction of the records in B1 now yields only ICa. C: finally, subtraction of ICa in B2 from ICa + IK(Ca) in A2 isolates IK(Ca).

As expected for an A-type current, IK(A) was inactivated completely when the holding potential was shifted from -90 to -40 mV (Fig. 2A, 2 and 3). Furthermore, it was completely blocked by 10 mM 4-AP, but was only slightly affected by 50 mM TEA (Fig. 3). On depolarization, IK(A) reached peak amplitude within a few milliseconds, then inactivated nearly completely within 100 ms. Both the activation and inactivation occurred more rapidly with increasing depolarization. Peak IK(A) amplitudes measured at 0 mV ranged from 2 to 7 nA.


View larger version (27K):
[in this window]
[in a new window]
  		FIG. 3. Differential block of IK(A) and IK(V) by the K-current blockers TEA and 4-AP. Currents were elicited by depolarizing steps from -90 mV to potentials ranging from -70 to +40 mV in 10-mV increments. A: currents elicited in normal ASW and after addition of 50 mM TEA, which blocked most of IK(V). B: currents elicited in normal ASW and after addition of 10 mM 4-AP, which blocked IK(A).

In all of these ways, the current resembled IK(A) in ARC fibers (Brezina et al. 1994b). There was, however, one surprising difference. The opener IK(A) appeared to activate significantly more positive (it typically was seen first at test potentials between -40 and -30 mV: Fig. 2B) than the ARC IK(A) (first seen between -50 and -40 mV). (In contrast, the voltage range of inactivation did not seem significantly different in the two fiber types.)

To confirm and quantify this observation, we directly compared the voltage-dependence of activation of IK(A) in a group of opener and a group of ARC fibers. In Fig. 4 we have plotted the peak A-current conductance activated at various potentials around the apparent threshold in the two fiber types. The data have been fitted with Boltzmann relations (see Fig. 4 legend). Although the best fits to the opener and the ARC data had essentially identical slopes, they differed considerably---by ~9 mV---in their position along the voltage axis. In the opener fibers, the A-current conductance was estimated to be half-maximal at -20.8 ± 0.9 mV (n = 10), and in the ARC fibers at -29.5 ± 1.5 mV (n = 7), a statistically significant difference (Student's t-test, P < 0.01). Thus indeed, the opener muscle possesses a more positive-activating IK(A) than does the ARC muscle. It also appears that the opener IK(A) is more homogeneous than the ARC IK(A); an interesting possibility is that the large variability in the voltage-dependence of the ARC current seen in Fig. 4 reveals the presence of multiple IK(A) subtypes.


View larger version (16K):
[in this window]
[in a new window]
  		FIG. 4. Voltage dependence of activation of IK(A) differs between opener and accessory radula closer (ARC) fibers. In 10 opener and 7 ARC fibers, IK(A) was isolated by subtraction of currents remaining in the presence of 4 mM 4-AP from control currents elicited by depolarizing steps from -90 mV to a range of potentials between -55 and -15 mV. Peak IK(A) was converted to conductance, gK(A), assuming ohmic permeation with reversal at the presumed EK, -85 mV in the opener fibers (see HYPERPOLARIZATION-ACTIVATED, INWARDLY RECTIFYING K CURRENT) and -81 mV in the ARC fibers (Brezina et al. 1994a). gK(A) values from individual fibers were fit with the Boltzmann function g = gmax/{1 + exp[(V - V1/2)/k]}, where gmax is the maximal conductance, V1/2 is the voltage at which the conductance is half-maximal, and k is the slope factor. Data from each fiber then were scaled to the mean fitted opener or ARC gmax to normalize for different fiber sizes. Shown are the pooled normalized data, as well as curves representing the means of the opener and ARC fits. Mean fitted parameter values were gmax = 21.0 nS, V1/2 = -20.8 mV, and k = -4.3 mV for the opener fibers and gmax = 34.3 nS, V1/2 = -29.5 mV, and k = -4.53 mV for the ARC fibers. Also shown is the opener fit scaled to the same gmax as the ARC fit.

DELAYED-RECTIFIER K CURRENT, IK(V). This current was indistinguishable from its counterpart in ARC fibers (Brezina et al. 1994b). It was a more slowly activating and inactivating current than IK(A) and was responsible for the second outward peak in Fig. 2A1. It too was Ca2+ independent. IK(V) activated (with an apparent threshold around -20 mV: Fig. 2B) as well as inactivated more positive than IK(A). At a holding potential of -40 mV, where IK(A) was inactivated completely, only a small fraction of IK(V) was inactivated (compare Fig. 2A2 with the subtracted difference current in Fig. 2A3). IK(V) was blocked substantially by 50 mM and completely by 460 mM TEA, but only slightly, if at all, by 10 mM 4-AP (Fig. 3). IK(V) was thus much more sensitive than IK(A) to TEA but less sensitive to 4-AP. Peak IK(V) amplitudes in opener fibers ranged from 1 to 5 nA at 0 mV.

Ca2+-ACTIVATED K CURRENT, IK(Ca). In normal Ca2+-containing solution, depolarizing steps activated a further substantial component of outward current in addition to IK(A) and IK(V) (Fig. 5A1). This current appeared to be very similar to the IK(Ca) described in ARC fibers by Brezina and Weiss (1995a). Thus the current disappeared on removal of extracellular Ca2+ or addition of 1 mM Cd2+; it was noticeable only when the steps reached voltages more positive than about -40 mV, and it was very sensitive to TEA, being virtually completely blocked at 10 mM (Fig. 5). The isolated IK(Ca) usually consisted of a large early transient peak (6-9 nA), sometimes followed by smaller secondary oscillations, superimposed on a sustained current (2-4 nA; Fig. 5C). A very similar complex time course is seen in ARC fibers (Brezina, unpublished observations).

As in ARC fibers (Brezina and Weiss 1995a), the activation of IK(Ca) presumably followed influx of extracellular Ca2+ through voltage-gated Ca channels (see below). However, in ARC fibers much of the IK(Ca) appears not to be activated by this Ca2+ directly but rather by Ca2+ secondarily released from intracellular stores by Ca2+-induced Ca2+ release (CICR). If this was the case in opener fibers too, the spatiotemporal complexity of the intracellular Ca2+ dynamics probably accounted for the complex time course of IK(Ca). Alternatively, the possibility remained that there was more than one type of Ca2+-activated K current present.

Ca CURRENT, ICa. When the same depolarizing steps that normally activated IK(A), IK(V), and IK(Ca) were given in the presence of high TEA and 4-AP, which completely blocked those currents, an underlying Ca current was revealed. This appeared to be primarily a high-voltage-activated, dihydropyridine-sensitive L-type Ca current similar to that found in ARC fibers (Brezina et al. 1994c; Ram and Liu 1991).

The current was carried by Ca2+ (ICa) or Ba2+ (IBa), and was blocked completely when these ions were replaced with Co2+ (Fig. 6, A and B). The current first appeared with steps to about -40 mV, was largest between -10 and 0 mV (IBa) or 0 and +10 mV (ICa), then decreased again until it apparently reversed between +40 and +60 mV (Fig. 6C). On depolarization, the current activated to peak within a few milliseconds (activation became faster with increasing depolarization), then inactivated relatively slowly. The inactivation was much more pronounced for ICa than for IBa (Fig. 6, A and B). In ARC fibers, the main reason for this was shown to be the fact that the current inactivates by a Ca2+-dependent mechanism that is more strongly activated by Ca2+ than by Ba2+. For this reason and presumably also because of the intrinsically greater permeability of L-type channels to Ba2+ (see Brezina et al. 1994c), IBa was significantly (1.3-2.8 times) larger than ICa (at 0 mV, generally 2-5.5 nA vs. 1.9-3.7 nA).


View larger version (30K):
[in this window]
[in a new window]
  		FIG. 6. Ca current, ICa. All of these experiments were done in Ca(Ba, Co)/TEA/4-AP ASW to eliminate all other currents (except ILeak). A: ICa, IBa (current carried through the Ca channels by Ba2+), and ICo (i.e., ILeak) elicited by depolarizing steps from -90 to 0 mV. B: IBa, ICa, and ICo elicited by steps from -90 mV to a range of depolarized potentials. C: I-V relations of peak leak (i.e., ICo)-subtracted ICa and IBa as well as of ICo itself from the records in B. D1: block of (leak-subtracted) IBa, elicited by steps from -90 to 0 mV, by increasing concentrations of nifedipine. D2: summary plot (means ± SE, n = 13) of the percentage block of IBa, elicited by steps from -90 to 0 mV, at the time of the peak current and at the end of the voltage step (after 200 ms), by 10 µM nifedipine.

Finally, ICa or IBa was blocked substantially by the dihydropyridine Ca-channel antagonist nifedipine. As in ARC fibers, the block exhibited a characteristic time dependence, becoming progressively stronger with time into the depolarized voltage step (Fig. 6D). In opener fibers, the block by 10 µM nifedipine was variable, ranging from 15 to 100% at the peak of ICa and from 63 to 100% at the end of the step. Therefore, although in most respects the opener muscle ICa appeared identical to that of the ARC, because of the variability in nifedipine sensitivity, we cannot rule out an additional, non-L-type, component in the opener ICa.

LACK OF DEPOLARIZATION-ACTIVATED Na CURRENT. As in ARC fibers (Brezina et al. 1994c), we failed to find one major class of current in the opener fibers, namely any voltage-gated Na current. When the three depolarization-activated K currents and the Ca current were blocked, there remained only the leakage current.

LEAKAGE CURRENT, ILeak. This was defined operationally to be the current that remained when all identified currents were blocked as completely as possible, i.e., in Co/TEA/4-AP ASW (see METHODS). As in ARC fibers (Brezina et al. 1994b,c), ILeak was small, without significant kinetics, and voltage independent except at very positive voltages (Figs. 2, A4 and B, and 6, B3 and C). ILeak routinely was subtracted to obtain identified currents in pure form.

ACh-activated currents

Next, we examined currents activated by ACh, the primary transmitter of the opener motoneurons (Evans et al. 1996). When ACh was applied rapidly (e.g., puffed from a pressure-ejection pipette) onto a voltage-clamped opener fiber, it activated a large inward current (see e.g., Fig. 13A). At -100 mV, 10 µM ACh activated currents with peak amplitudes ranging from 1.2 to 11 nA, but in most fibers ~2-3 nA. The ACh-activated current reached peak amplitude in 100-200 ms, though very likely this value simply reflected the speed of ACh application; the intrinsic rate of activation of the current was probably faster. (Rates of activation of the ACh and modulator effects are further examined and compared in a later section.) After the peak, the current desensitized with a time course of seconds while ACh was present and then deactivated rapidly once ACh was removed.


View larger version (26K):
[in this window]
[in a new window]
  		FIG. 13. Comparison of activation rates of the ACh and modulator effects. A: ACh-activated current [mainly IACh(cat)]. Ten micromolar ACh was puffed onto the fiber held at -100 mV in normal ASW. B: enhancement of ICa. ICa (peak, leak-subtracted current plotted) was elicited by steps from -90 to 0 mV every 5 s in Ca/TEA/4-AP ASW and was enhanced with 10 µM 5-HT. C: IMod(cat). Ten micromolar 5-HT was puffed onto the fiber held at -100 mV in normal ASW. D: IMod(K). Ten micromolar FMRFamide was puffed onto the fiber held at -30 mV in normal ASW. E: rising phases of the responses in A-D scaled to the same inward or outward maximum and superimposed for comparison. All responses begin at time 0. IMod(cat) and especially ICa(mod) were so slow that they failed to reach significant amplitude within the time window shown. F: summary plot (means ± SE, from 3 fibers each, usually averaging 3-4 responses per fiber) of the activation time constants, tau activation (the rising phase fitted with a single exponential), of the four responses.

I-V relations obtained with voltage ramps (Fig. 7) showed that the bulk of the ACh-activated current closely resembled the cationic ACh-activated current, IACh(cat), described in ARC fibers (Kozak et al. 1996; Liu and Ram 1994). Because Na+ predominates in normal extracellular solution, IACh(cat) is carried primarily by Na+, but small contributions by other cations, in particular K+ and Ca2+, are likely (Kozak et al. 1996). As in ARC fibers, the opener IACh(cat) reversed around 0 mV (the most positive voltage reached with our ramps) or even more positive (Fig. 7; n = 15). Furthermore, the opener current, like the ARC current, was blocked readily by the classical cholinergic antagonist hexamethonium.

In ARC fibers, Kozak et al. (1996) found hexamethonium to be a very useful tool with which to dissect the total ACh-activated current, as it completely blocks IACh(cat) but does not affect a second component, an ACh-activated Cl current, IACh(Cl). In ARC fibers, each of the two components contributes a substantial fraction (IACh(cat) perhaps slightly larger) of the total ACh-activated current. In contrast, in 9 of 14 opener fibers, hexamethonium completely blocked the whole ACh-activated current, and in the other 5 fibers blocked a very large fraction of it (65-93% was blocked by saturating, 100 µM, hexamethonium). The remaining hexamethonium-insensitive current indeed reversed at -57 ± 6 mV (n = 5) as expected for IACh(Cl) (Kozak et al. 1996) (Fig. 7).

The large predominance of IACh(cat) in the opener fibers was confirmed in two other ways. First, the total ACh-activated current always reversed around 0 mV or even more positive, close to the presumed reversal potential of pure IACh(cat) (Fig. 7). In ARC fibers, in contrast, with a large admixture of IACh(Cl), the total ACh-activated current reverses around -30 mV. Second, suberyldicholine, which selectively activates IACh(Cl) in ARC fibers (Kozak et al. 1996), had relatively little effect in the opener fibers. Of 13 fibers tested, it had no effect at all in 8, and in the remainder it activated only very small currents of ~50 pA. As expected, these were hexamethonium-insensitive.

Thus we were able to conclude that although the opener muscle does possess both IACh(cat) and IACh(Cl), it possesses much less IACh(Cl)---probably in terms of absolute current density but certainly relative to IACh(cat)---than does the ARC muscle. This, then, was the second clear difference that we found between the ion currents of the two muscles.

Effects of modulators

As mentioned in the INTRODUCTION, the effects of modulators on the whole opener muscle can, by analogy with their analyzed effects on the ARC muscle, be explained in terms of four primary effects: cAMP-mediated potentiation of contractions, hyperpolarization and depression of contractions, cAMP-mediated acceleration of the relaxation rate of contractions, and depolarization (which has no counterpart in the ARC muscle). As we now describe, in the isolated opener fibers we found likely candidate mechanisms for the first, second, and fourth effects. We identified enhancement of ICa and activation of a modulator-specific K current as probably underlying the first and second effects, respectively, as in the ARC muscle. As the likely mechanism of the fourth effect, we identified a novel modulator-activated cationic current.

ENHANCEMENT OF ICa. The same modulators---5-HT, MMA, MMB---that elevate cAMP in the ARC muscle and potentiate its contractions also have these effects in the opener muscles (Evans et al. 1993; C. G. Evans, F. S.Vilim, E. C. Cropper, and K. R. Weiss, unpublished data). In the ARC muscle, there is strong evidence that they act via cAMP-mediated enhancement of ICa (Brezina et al. 1994d). We naturally asked whether the same was true in the opener muscles.

Indeed, application of 5-HT or the MMs to isolated opener fibers enhanced ICa or IBa very much as in ARC fibers (Fig. 8, A-C). The enhancement began to be evident at modulator concentrations around 10 nM and saturated by 10 µM. Even with saturating concentrations, the effect developed very slowly, over minutes (see e.g., Fig. 13B, and ACTIVATION RATES OF ACH AND MODULATOR RESPONSES), and did not desensitize. The maximal enhancement averaged 50-70%, although it varied considerably between fibers (range 8-450%, n = 57). The maximal effects of 5-HT and the MMs were of similar size and generally occluded each other(Fig. 8C).


View larger version (17K):
[in this window]
[in a new window]
  		FIG. 8. Enhancement of ICa by modulators and adenosine 3',5'-cyclic monophosphate (cAMP). Currents were elicited by depolarizing steps from -90 to 0 mV in Ca(Ba)/TEA/4-AP ASW and then leak-subtracted. A and B: enhancement of IBa and ICa by 10 µM serotonin (5-HT). C: enhancement of IBa by 10 µM MMA was saturating, as 100 µM MMA had no additional effect, and occluded the effect of 10 µM 5-HT. D: 1 mM 8-chlorophenylthio-cAMP (CPT-cAMP) plus 100 µM Ro 20-1724 likewise enhanced IBa and occluded the effect of 10 µM 5-HT.

We saw no obvious difference between the properties of the modulator-enhanced current and the basal current. The enhanced current had similar kinetics, also was blocked completely when Ca2+ or Ba2+ was replaced with Co2+, and showed similar sensitivity to nifedipine (10 µM nifedipine blocked 54 ± 10% of the peak current, 84 ± 5.5% of the current at the end of the voltage step, n = 6; compare Fig. 6D2). Thus most probably, the effect was truly an enhancement of the preexisting ICa rather than activation of a second, distinct Ca current (cf. Brezina et al. 1994d).

As in ARC fibers, the enhancement was mimicked by cAMP. Application of 0.5-1 mM 8-chlorophenylthio-cyclic AMP (CPT-cAMP) plus 100 µM Ro 20-1724 (a phosphodiesterase inhibitor) or of 100 µM forskolin (an adenylyl cyclase activator) enhanced ICa or IBa maximally. This increase was of the same magnitude as that produced by the modulators, and it occluded their effects (Fig. 8D).

FMRFamide, which does not elevate cAMP or potentiate contractions either in the ARC or in the opener muscle (C. G. Evans, F. S.Vilim, E. C. Cropper, and K. R. Weiss, unpublished data), usually had no effect on ICa or IBa, although occasionally it appeared to decrease the current by 10-20%. In the presence of FMRFamide (even when it had decreased the current), 5-HT or the MMs still were able to enhance the current normally. Very similar findings were made, with FMRFamide as well as the FRFs, in ARC fibers (Cropper et al. 1994; Brezina, unpublished observations).

MODULATOR-ACTIVATED K AND CATIONIC CURRENTS, IMod(K) AND IMod(cat). When quasi-steady-state I-V relations were generated with slow voltage ramps in normal solution while the modulators were applied, each of the modulators tested---5-HT, MMA, MMB, and FMRFamide---activated a somewhat different, complex current (Fig. 9). In each case, however, the current could be interpreted as the sum of the same two elementary components, an outward current, IMod(K), and an inward current, IMod(cat). The two components had very different properties and under some circumstances could be activated independently of one another, by two different modulators (Fig. 9A) or even the same modulator at different concentrations or times (Fig. 9B). Each modulator activated different relative amplitudes of the two components, giving a combined current with a distinctive I-V relation and combined reversal potential. This will be described later. First we present the individual properties of the two elementary currents.

PROPERTIES OF IMod(K). Some IMod(K) was activated by all of the modulators tested. Because FMRFamide activated a large amplitude of IMod(K) and, more importantly, did not activate any contaminating IMod(cat), it was used in most experiments to examine IMod(K). The results were confirmed with the other modulators.

The opener IMod(K) very closely resembled the corresponding current activated by the same modulators in ARC fibers (Brezina et al. 1994e; Cropper et al. 1994). As is evident in Fig. 9A2, the current was outwardly rectifying, indeed significantly more so than predicted simply from the asymmetry between the intracellular and extracellular K+ concentrations by the Goldman-Hodgkin-Katz constant-field equation (cf. Brezina et al. 1994e). IMod(K) was thus moderately voltage dependent, activated by depolarization. In normal 10 mM K+ solution, IMod(K) reversed (or simply disappeared, due to the voltage dependence, as in Fig. 9A2) at -83 ± 2 mV (n = 5; see Fig. 12B, FMRFamide values). This was very close to the estimated value of EK (see HYPERPOLARIZATION-ACTIVATED, INWARDLY RECTIFYING K CURRENT). Furthermore, when the extracellular K+ concentrationwas elevated to 100 mM, the reversal potential shiftedto -23 ± 2 mV (n = 7) when the current was activated by FMRFamide (Fig. 10B) and -33 ± 3 mV (n = 7) when it was activated by MMA, again close to the new value of EK. (As can be seen in Fig. 10B, the voltage dependence of IMod(K) was even more obvious under these conditions.) In contrast, the reversal potential did not change when all extracellular Na+ was removed (see Fig. 12B, FMRFamide values). Thus IMod(K) was truly a K current.


View larger version (24K):
[in this window]
[in a new window]
  		FIG. 12. Comparison of the relative amplitudes of IMod(K) and IMod(cat) activated by the different modulators (A) and of the combined reversal potentials (B). Plotted are means ± SE from the indicated numbers of fibers. All modulators were applied at 10 µM, i.e., saturating concentrations. In A, we measured the absolute modulated current amplitude at -100 mV, reflecting mainly IMod(cat), and at -30 mV, reflecting mainly IMod(K), in each case normalized to the unmodulated current amplitude to allow for different fiber sizes. Measurements were made in normal ASW and in B also in 0 Na ASW.

The voltage dependence of IMod(K) was associated with time-dependent relaxations of the current after a depolarizing voltage step. First the current activated moderately slowly, causing an outward relaxation over the first ~100 ms after the step (Fig. 11C). Then the current appeared to inactivate somewhat, over 1 or 2 s (for example, in Fig. 11, A and D, the inactivation was apparently relieved by a 2- or 3-s-long hyperpolarization). On a still slower time scale, the response desensitized.

In ARC fibers, a distinctive property of IMod(K) is its unusually high sensitivity to block by 4-AP, with Ki around 10 µM. At 100 µM, 4-AP blocks IMod(K) essentially completely, except in some fibers where there may be a minor component of IMod(K) that differs from the major component only in being unaffected by 4-AP, even at concentrations as high as 10 mM (Brezina et al. 1994e; Cropper et al. 1994). The results in the opener fibers were consistent with this picture, although the proportion of the putative 4-AP-insensitive component appeared to be larger than in ARC fibers and more variable from fiber to fiber. Sometimes IMod(K) was blocked almost completely by 100 µM 4-AP (Fig. 11D); on the other hand, in as many as 18% of opener fibers tested with FMRFamide and 24% tested with MMA, 100-200 µM 4-AP did not block IMod(K) at all.

Each of the modulators began to activate measurableIMod(K) at concentrations around 10 nM, and saturation was reached between 1 and 10 µM. With rapid, puffed application of the modulators, it was evident that the IMod(K) response took several hundred milliseconds to reach peak amplitude (see e.g., Fig. 13D, and ACTIVATION RATES OF ACH AND MODULATOR RESPONSES). If the modulator remained present, the response desensitized to a variable extent (0-100%, but in most cases partially) with a time course of seconds (e.g., Figs. 9, A1 and B, and 13D).

Although we saw no obvious differences in IMod(K) activated by the different modulators, we were unable to show convincingly through occlusion experiments that all of the modulators activated the same IMod(K). Sometimes, indeed, there clearly was not complete occlusion even between the large responses to saturating MMA and FMRFamide. Unfortunately, these experiments were difficult, and in general IMod(K) was difficult to dissect, because of the very different amplitudes of the responses to the different modulators (see below) and large variability between fibers in the amplitude and 4-AP sensitivity of IMod(K) and in the extent of desensitization. Thus the possible contribution of multiple channel subtypes to IMod(K) will best be evaluated using single-channel methods. [These have already begun to be applied in ARC fibers (Erxleben and Brezina 1996).]

PROPERTIES OF IMod(cat). In addition to the outward IMod(K), 5-HT, MMA, and MMB (but not FMRFamide) also activated an inward current, IMod(cat). This current is never seen---even with the same modulators---in ARC fibers. This, then, was the third major difference that we found between the ion currents of the two muscles.

There was, unfortunately, no modulator that activated only IMod(cat), without any contaminant IMod(K). However, 5-HT activated a large amplitude of IMod(cat) and relatively little IMod(K) (Fig. 9A). It therefore was used in most experiments to examine IMod(cat); the results were confirmed with the other modulators.

IMod(cat) could be seen to be inward, under favorable circumstances, to voltages as positive as -30 mV (Fig. 9, A and B). Almost certainly, however, this still represented the combined reversal potential of a mixture of positive-reversing IMod(cat) and negative-reversing IMod(K). The true reversal potential of IMod(cat) was therefore likely to be more positive yet.

When all extracellular Na+ was removed, IMod(cat) disappeared, leaving just IMod(K) (Fig. 10A; see also Fig. 12B). Thus IMod(cat) appeared to be carried primarily by Na+. It was not primarily a K current, as it did not reverse anywhere near EK. Its reversal potential was closer to ECl [which was about -60 mV, judging by the reversal potential of IACh(Cl) (cf. Kozak et al. 1996)], but reducing extracellular Cl- from the normal 602 to 142 mM had no effect on IMod(cat) (or on IMod(K)). The possibility remained, however, that the IMod(cat) channels were relatively nonspecifically permeable to cations, as in the case of IACh(cat). In that case, Na+ was the main current carrier under normal conditions simply because of its predominance externally. This possibility, which we acknowledge by referring to the current as IMod(cat), was difficult to test because a background of contaminant IMod(K) was always present, against which any changes in IMod(cat) had to be resolved. Apart from Na+, the most physiologically important putative current carrier was Ca2+, as even a small fraction of the current carried by Ca2+ could contribute sufficient Ca2+ to significantly affect contraction of the muscle (see DISCUSSION). Unfortunately, we were unable to resolve any changes in IMod(cat) when extracellular Ca2+ was removed, or when it was elevated (up to 380 mM).

IMod(cat) appeared to be somewhat voltage dependent, activated by hyperpolarization. After a hyperpolarizing voltage step, the current relaxed inwardly over the first several hundred milliseconds; this relaxation became faster with increasing hyperpolarization (Fig. 11, A and B).

Each of the modulators began to activate measurableIMod(cat) at concentrations between 1 and 10 nM, and saturation was reached between 1 and 10 µM. Puffed application of the modulators showed that the IMod(cat) response was relatively slow, taking 1 or 2 s to reach peak amplitude (see e.g., Fig. 13C and ACTIVATION RATES OF ACH AND MODULATOR RESPONSES). In contrast to IMod(K), IMod(cat) did not desensitize (Fig. 9, A1 and B). Saturating concentrations of 5-HT appeared to activate IMod(cat) fully, as the responses to MMA and MMB were occluded. However, the MMs, which activated smaller amplitudes of IMod(cat) (see below), did not always occlude the response to 5-HT.

DIFFERENT MODULATORS ACTIVATE DIFFERENT PROPORTIONS OF IMod(K) AND IMod(cat). The occlusion experiments, even when only partially successful, as well as the general similarity of the currents suggested that on balance, pending final resolution of this question through single-channel experiments, we could conclude that all of the modulators activated the same IMod(K) and IMod(cat). However, each modulator activated the two currents in different proportions. This is summarized in Fig. 12A. The upward bars represent IMod(K), the downward bars IMod(cat). It can be seen that FMRFamide and MMA activated large amplitudes of IMod(K), whereas MMB and 5-HT activated considerably smaller amplitudes. (The maximal amplitude of IMod(K) activated by saturating FMRFamide or MMA was on the order of 1 nA at -30 mV.) On the other hand, 5-HT activated the largest amplitude of IMod(cat) (on the order of 0.3 nA at -100 mV), MMA and MMB smaller amplitudes, and FMRFamide no IMod(cat) at all.

The proportions of IMod(K) and IMod(cat) activated by each modulator were well reflected in the reversal potential ofthe combined current. This is summarized in Fig. 12B. In normal solution, 5-HT, with the largest ratio of IMod(cat) toIMod(K), gave the most positive reversal potential, on average -46 mV. MMA and MMB gave more negative values, and FMRFamide, which activated only IMod(K), gave the most negative reversal potential, on average -83 mV, very close to EK. When extracellular Na+ was removed, so that IMod(cat) was eliminated and just IMod(K) remained, the reversal potentials of all of the other modulators also shifted negative, close to EK, and became essentially indistinguishable from each other.

As expected from this apparent simple additivity of reversal potentials, predictable shifts in reversal potential were seen as the proportions of IMod(K) and IMod(cat) changed with different concentrations of a modulator or at different times during the response, in particular as IMod(K) desensitized (Fig. 9B). Similarly, predictable shifts were seen when two different modulators were applied in succession. For example, 5-HT or MMB applied after MMA or FMRFamide shifted the reversal potential in the positive direction, whereas MMA or FMRFamide applied after 5-HT or MMB shifted it in the negative direction (Fig. 9A2).

ACTIVATION RATES OF ACh AND MODULATOR RESPONSES. The three modulator responses---enhancement of ICa, activation of IMod(K), and activation of IMod(cat)---had relatively slow, and distinctly different, time courses (Fig. 13). To demonstrate this, we compared the rise times of the three responses and that of the current activated by ACh after rapid puffed application of the modulators or ACh. Figure 13E shows typical rise times superimposed for comparison; Fig. 13F summarizes measurements from a number of fibers. As noted above, the rise time of the ACh-activated current, with a time constant of ~150 ms, was probably a good indication of the rate of arrival of the puffed substance at the fiber. However, each of the modulator responses activated significantly slower than this (Student's t-test, P < 0.05), and in fact their time constants were also significantly different from one another. The time constant of the IMod(K) response was ~400 ms, that of the IMod(cat) response ~1.3 s, and that of the ICa enhancement very slow, ~80 s.

ACTIVATION OF IMod(K) AND IMod(cat) IS cAMP INDEPENDENT. The slow activation rates of the modulator responses suggested the possibility of second-messenger involvement (see DISCUSSION). One second messenger, cAMP, indeed was implicated already in the enhancement of ICa. To test whether cAMP mimicked also either of the other two responses, we applied 0.5-1 mM CPT-c