| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Center for Neurobiology and Behavior, Department of Pharmacology, Howard Hughes Medical Institute, Columbia University, New York City, New York 10032
Submitted 6 January 2003; accepted in final form 18 April 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
). In
contrast, somatostatin, acting through 5-lipoxygenase metabolites of
arachidonic acid and 
-adrenergic agonists, acting through the cyclic AMP
(cAMP) cascade, have been shown to enhance the M current
(Lammers et al. 1996).
In cardiac muscle, epinephrine and acetylcholine exert opposing actions on
the L-type Ca2+ channel. -adrenergic agonists
enhance the Ca2+ current by cAMP-dependent
phosphorylation (Sako et al.
1998). Acetylcholine, acting through muscarinic receptors,
inhibits the Ca2+ current through production of cyclic
GMP (cGMP) (Sperelakis et al.
1996), which stimulates cAMP phosphodiesterase (PDE), leading to a
fall in cAMP levels (Fischmeister and
Hartzell 1991). A similar antagonistic modulation of
Ca2+ current by cAMP and cGMP has been demonstrated in
hippocampal pyramidal neurons (Doerner and
Alger 1988). Here again, cGMP opposes the action of cAMP via
activation of a cAMP PDE.
In the mechanosensory neurons of Aplysia, serotonin (5-HT) and the
small cardioactive peptide (SCPB) increase the excitability of the
sensory neuron and cause presynaptic facilitation
(Belardetti and Siegelbaum
1988; Byrne and Kandel
1996). In contrast, the neuropeptide FMRFamide decreases sensory
neuron excitability and causes presynaptic inhibition
(Belardetti and Siegelbaum
1988). 5-HT and SCPB enhance excitability, in part, by
causing prolonged all-or-none closure of the S-type K+ channel
through cAMP-dependent phosphorylation
(Shuster et al. 1985).
FMRFamide decreases excitability, in part, by enhancing the S channel open
probability, an action mediated by 12-lipoxygenase metabolites of arachidonic
acid (Belardetti et al. 1987;
Buttner et al. 1989).
FMRFamide also reverses S channel closure produced by 5-HT or cAMP
(Belardetti et al. 1987) by
stimulating protein dephosphorylation
(Endo et al. 1995;
Sweatt et al. 1989). Although
opposing modulation of the S-type K+ channel has been described in
most detail, Aplysia sensory neurons contain a variety of additional
currents that are modulated by these transmitters, including the N and L-type
Ca2+ currents
(Edmonds et al. 1990), a
calcium-activated K+ current
(Walsh and Byrne 1989), and
the delayed rectifier K+ current
(Baxter and Byrne 1989;
Baxter et al. 1999).
A distinct modulatory target of 5-HT and FMRFamide in other identified
Aplysia neurons is a hyperpolarization-activated chloride current.
5-HT inhibits this current in several Aplysia neurons through a
cAMP-dependent mechanism (Lotshaw and
Levitan 1987). In the Aplysia neuron L2, FMRF-amide
enhances this current, although a second messenger has not been identified
(Thompson and Ruben 1988).
Here, we show that in Aplysia sensory neurons, SCPB and
FMRFamide modulate this Cl– current in opposing directions.
SCPB closes the Cl– current through cAMP-dependent
phosphorylation, whereas FMRFamide activates this current via the cGMP
cascade. Moreover, the effect of cGMP is mediated through the recruitment of a
cAMP PDE, which enhances the Cl– current by lowering the
intracellular concentration of cAMP, similar to the antagonistic modulation of
the cardiac L-type Ca2+ channel by adrenergic and
cholinergic agonists.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Aplysia sensory neurons were dissected from pleural ganglia and
grown in cell culture (Bailey et al.
2000). Recordings were obtained 1–7 days after culture. The
modulation of the Cl– current studied here was similar
throughout the culture period.
Electrophysiological recording
The patch-clamp technique was used to obtain whole cell voltage-clamp
recordings with a List EPC7 patch-clamp amplifier (List-Medical). Patch
electrodes were pulled from borosilicate glass micropipettes (Rochester
Scientific, Rochester, NY) with a two-stage vertical puller (David Kopf
Instruments, Tujunga, CA, or List-Medical). Electrodes were then fire-polished
with a heating filament of platinum-iridium wire-coated with a bead of
borosilicate glass (Narishige). Electrode resistances of 2–3 M
were used.
For whole-cell voltage-clamp recordings, the bath solution contained normal artificial sea water (ASW) solution (in mM) composed of 460 NaCl, 10 KCl, 11 CaCl2, 55 MgCl2, and 10 HEPES (pH 7.6); the pipette contained a solution of 550 KCl, 1 MgCl2, 1 CaCl2, 10 EGTA, 5 MgATP, 0.1 GTP, and 10 HEPES (pH 7.3). In experiments requiring K+ current blockade, both the bath and pipette solutions were adjusted by substituting CsCl for KCl.
After breaking into the cell, the extent of dialysis was monitored by the
progressive increase in a slow inward current relaxation elicited by 15-s
pulses from –40 to –80 mV, applied once every 100–200 s.
Dialysis was judged to have reached completion when four successive voltage
pulses yielded current relaxations of similar magnitude. This typically
required 
30–45 min of dialysis. The pulse protocol was then
switched to one of the two listed in the following text and the experiment
begun.
Two types of whole cell voltage-clamp experiments were performed. First, current-voltage relationships were obtained using voltage-clamp ramps to provide a rapid, qualitative assay of drug-induced ionic current responses. Second, time-dependent currents were studied using 15-s-long steps to a single voltage. In the first type of experiment, the steady-state membrane potential was clamped at –40 mV, and 500-ms ramps from –80 to –5 mV were given once every 20 s. In the second type of experiment, the membrane potential was clamped at –40 mV, and 15-s steps to –80 mV were given once every 100 s. In some of these experiments, voltage steps were also given sequentially to: –70, –60, –50, –30, and –20 mV. Current traces obtained in response to voltage ramps or steps were digitized and stored on computer. Difference currents were then acquired by subtracting the average of three to four control current traces (absence of drug) from a similar average of experimental traces (presence of drug).
For reversal potential (Erev) measurements involving Na+ substitution, current responses were first recorded to drug applied in ASW in which Cs+ was substituted for K+. Voltage-clamp ramps were given throughout the experiment. Ramp current-voltage relationships thus generated were then compared with those obtained in response to drug applied in a low-Na+ solution containing (in mM) 300 N-methylglucamine, 170 NaCl, 11 CaCl2, 10 CsCl, 55 MgCl2, and 10 Na HEPES (pH 7.6).
For Erev experiments involving Cl– substitution, a similar procedure was employed except that the external solution contained either 600 mM Cl– or 200 mM Cl– with 400 mM methane sulfonate. In addition, 100 µM TTX was added to the external solutions to block the voltage-dependent Na+ current, allowing adequate clamp control at positive voltages. In these experiments, the pipette solution contained 350 Cs methane sulfonate, 200 CsCl, 1 MgCl2, 1 CaCl2, 10 EGTA, 5 MgATP, 0.1 GTP, and 10 HEPES (pH 7.3). A prepulse protocol, in which ramps were preceded by 5-s pulses to –80 mV, was utilized to enhance the Cl– current.
Drugs and perfusion
Drugs were applied directly to intact sensory neurons using continuous microperfusion. In addition, the culture dish as a whole was simultaneously perfused with the vehicle solution using a macroperfusion system run from a peristaltic pump (Rainin Instrument, Woburn, MA). The bath temperature was maintained between 16 and 18°C with a refrigerated circulating pump (NESLAB Instruments, Portsmouth, NH) attached to a copper cooling plate that served as the base for the culture dish. Bath temperature was recorded with a digital thermometer (Sensortek).
Aliquots of the peptides SCPB, FMRFamide, and YGGFMRFamide (Peninsula) were prepared at a concentration of 1 mM in 0.1 N acetic acid, lyophilized in a speed vacuum, and stored at –20°C. On the day of each experiment, aliquots were resuspended in vehicle at a concentration of 10 µM. Solutions containing 5-HT, 8-chlorophenylthiocyclic AMP (8-CPT-cAMP), 8-bromo-cyclic GMP (8-Br-cGMP), or 8-bromo-cyclic inosine monophosphate (8-Br-cIMP) were prepared fresh from solid on the day of the experiment (Sigma). Similarly, solutions containing 3-isobutyl-1-methylxanthine (IBMX), RO 20-1724, or K252a (CalBiochem) were prepared fresh from solid on the day of the experiment. IBMX experiments were conducted at a concentration of either 100 µM (no DMSO) or 200 µM (0.1% DMSO). RO 20-1724 was dissolved in DMSO (Sigma) at 100–250 mM and applied at a final concentration of 100–250 µM (0.1% DMSO). K252a was dissolved in DMSO in the dark at a concentration of 10 mM. Aliquots of stock solution were added to vehicle at a final concentration of 10 µM K252a and 0.1% DMSO. Experiments with K252a were conducted in the dark to ensure stability of the drug. Once prepared all solutions were vortexed and loaded into the microperfusion system.
After obtaining a stable recording, cells were superfused with control vehicle solution, and test drugs were subsequently applied. For peptides and 8-CPT-cAMP, drug application lasted as along as required to reach a peak response. For inhibitor experiments, K252a was perfused for 10–20 min prior to testing its effect on the current response to other drugs.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Previous studies, using a two-microelectrode voltage-clamp or cell-attached
patch recordings, have shown that Aplysia mechanoreceptor sensory
neurons, at a holding potential of –40 mV, typically respond to
application of 5-HT or SCPB with an inward current and respond to
FMRFamide with an outward current, effects due to the respective closure and
opening of the S-type K+ channel (see
Belardetti and Siegelbaum
1988). In our studies using whole cell voltage clamp (WCVC) at a
holding potential of –40 mV, we also observed these typical inward and
outward current responses to SCPB and FMRFamide, respectively,
immediately after achieving WCVC mode (Fig.
1A). However, after dialysis of the sensory neuron with
the intracellular pipette solution (which contained 550 mM KCl) for periods
>10 min, the current responses to SCPB and FMRFamide reversed.
SCPB now produced an outward current at –40 mV, whereas
FMRFamide produced an inward current (Fig.
1B). The 5-HT response reversed in the same fashion as
the SCPB response (data not shown). An analog of FMRFamide present
in mollusks, YGGFMRFamide (Leung and
Stefano 1983), elicited this inward current more reliably than did
FMRFamide, itself, when responses were compared within the same cell. Although
we did not quantitate this difference, the probability of observing an inward
current and the magnitude of response tended to be greater with YGGFMRFamide
than with FMRFamide.
|
I-V curves for current responses to SCPB and YGGFMRFamide
The outward current response to SCPB displayed a nearly ohmic I-V curve in response to 500-ms voltage ramps. In the example of Figures 1B and 2A, 10 µM SCPB induced an outward current at the holding potential, –40 mV, and the peak outward current during the voltage ramp occurred at the most hyperpolarized potentials (–70 to –80 mV). The extrapolated Erev for the difference current was –8 mV. On average, 10 µM SCPB elicited an outward current of 630 ± 320 pA at –70 to –80 mV (n = 18).
|
Similar to the currents evoked by SCPB, FMRFamide or YGGFMRFamide produced current responses characterized by a near linear I-V curve. However, the shape and slope of these curves were opposite to those produced by SCPB. In the example shown in Fig. 2A, 500 nM YGGFMRFamide evoked an inward current, whose peak value was reached at the most hyperpolarized potential, –80 mV. The extrapolated Erev for the difference current was +7 mV. On average, 0.5–10 µM FMRFamide or YGGFMRFamide elicited an inward current of 310 ± 200 SE pA at –80 mV (n = 13). Although YGGFMRFamide tended to be somewhat more potent than FMRFamide (see preceding text), the data sets for these peptides were pooled because they produced similar responses. In two additional experiments using a constant holding potential of –40 mV, 10 µM FMRFamide produced a large holding current (Fig. 1B), whose mean amplitude in these two cells was 900 pA.
The fact that the direction and voltage dependence of the currents elicited by FMRFamide and SCPB were opposite to one another suggested that these peptides modulated the same current but in opposing directions. The finding that the two I-V curves were nearly linear (although with opposite slopes), indicated that there was voltage-gating on the time scale of the 500-ms voltage ramps (see also Fig. 9). Furthermore, the positive slope of the FMRFamide difference I-V curve indicated that it produced an increase in conductance, whereas the negative slope of the SCPB I-V curve indicated that it produced a decrease in conductance. During repeated experiments, the extrapolated reversal potentials for both I-V curves was approximately equal to 0 mV. (The small variations in the precise shape of the I-V curve and extrapolated reversal potential among different experiments was likely to be due to relatively small variations in the size of the large, transmitter-insensitive background current during the long time course of these experiments). In the following text we show that this novel current response was due to the modulation of a Cl– current, whose magnitude was greatly enhanced upon dialysis with the high intracellular Cl– concentration during the whole cell recordings. The high Cl– concentration also caused ECl to lie near 0 mV, close to the extrapolated reversal potential of the ramp I-V curves. SCPB inhibited the Cl– current, whereas FMRFamide enhanced the current.
|
SCPB inhibited the Cl– current through cAMP-dependent protein phosphorylation
SCPB, like 5-HT, elevates the levels of cAMP in Aplysia
sensory neurons, thereby leading to S channel closure
(Jarrard et al. 1993).
However, SCPB has a simpler action than 5-HT in that the
neuropeptide does not recruit a protein kinase C pathway utilized by 5-HT
(Braha et al. 1990;
Sacktor and Schwartz 1990).
Similar to the modulation of the S current, the induction of the outward
current response by SCPB described in the preceding text appears to
occur via activation of the cAMP cascade. Thus application of 8-CPT-cAMP, a
membrane-permeable analogue of cAMP, elicited an outward current at negative
potentials and yielded a difference ramp I-V curve whose shape,
slope, and reversal potential were similar to those elicited by
SCPB (Fig.
2B). On average 1 mM 8-CPT-cAMP produced an outward
current of 434 ± 330 pA at –80 mV (n = 4).
The effects of SCPB and 8-CPT-cAMP were dependent on protein
kinase activation. Thus application of the membrane permeable kinase
inhibitor, K252a, inhibited the outward current induced by SCPB
(Fig. 3A) and by
8-CPT-cAMP (Fig. 3B).
On average 10 µM K252a produced a 60% inhibition of the peak current
response at –80 mV to either 10 µM SCPB (n = 4)
or 1 mM 8-CPT-cAMP (n = 3). This inhibitory effect was consistently
observed in all seven experiments. Perfusion with 10 µM K252a (0.1% DMSO)
also had no consistent effect on sensory neuron holding current (n =
4). Perfusion with 0.1% DMSO, the vehicle used to dissolve the kinase
inhibitor, had no effect on the sensory neuron current responses (n =
2). The inhibitory action of K252a is consistent with the idea that the
actions of SCPB and 8-CPT-cAMP are mediated by PKA. Because the
inhibition was incomplete (60%), we cannot rule out some effect due to a
direct action of 8-CPT-cAMP on the channel
(Fesenko et al. 1985).
|
Action of FMRFamide is mediated by cGMP
Unlike SCPB, whose actions appear to be universally dependent on
cAMP, FMRFamide has been reported to regulate a diverse array of ion channels
through diverse mechanisms, including activation of S-K+ channels
through arachidonic acid metabolites
(Piomelli et al. 1987b) and
subunits of G proteins (van
Tol-Steye et al. 1999), antagonism of PKA actions through
phosphatase activation (Endo et al.
1995; Sweatt et al.
1989), inhibition of voltage-gated calcium currents through
activation of Go (Man-Son-Hing
et al. 1992) and direct activation of a ligand-gated
Na+ channel (Lingueglia et al.
1995). In Aplysia neuron R14, FMRFamide induces an inward
current similar to that studied here, and this current is also activated by
cGMP injection (Ichinose and McAdoo
1988,
1989). We therefore explored
the role of cGMP in the inward current response to FMRFamide in
Aplysia sensory neurons.
Application of a membrane-permeable analog of cGMP, 8-Br-cGMP, to dialyzed sensory neurons mimicked the inward current response to FMRFamide (Fig. 2C). Thus difference currents in response to 8-Br-cGMP yielded I-V curves with the same shape and slope characteristic of the FMRFamide-activated inward current. On average 1 mM 8-Br-cGMP produced an inward current of 225 ± 112 pA at –80 mV (observed in 4 of 7 cells; 3 cells showed no response).
Cyclic GMP activates a variety of intracellular effectors, including
cGMP-dependent protein kinase, cGMP-regulated PDEs, and cyclic
nucleotide-gated ion channels. To distinguish among some of these effector
mechanisms, we studied responses to a membrane permeant analogue of cIMP, a
cyclic nucleotide that is much more effective at stimulating cGMP-activated
PDE than at activating cGMP dependent protein kinase
(Miller et al. 1973;
Wexler et al. 1998). 8-Br-cIMP
also produced a large and consistent enhancement of the Cl–
current. Figure 2C
illustrates an experiment in which 1 mM 8-Br-cIMP evoked a peak inward current
at –80 mV of 500 pA, whereas the response to 8-Br-cGMP was 250 pA. On
average, 1 mM 8-Br-cIMP produced an inward current of 340 ± 120 pA at
–80 mV (n = 22 of 22 cells). The large response to 8-Br-cIMP is
consistent with the view that Cl– current activation was
mediated, at least in part, by the stimulation of a cGMP-activated PDE that
hydrolyzed cAMP and thus relieved an inhibitory effect of basal cAMP-dependent
phosphorylation.
Cyclic IMP-induced inward current occluded the FMRFamide response
We found that application of cGMP analogues not only simulated the action of FMRFamide, it also occluded the ability of FMRFamide to activate the Cl– current, consistent with the view that cGMP and FMRFamide target the same channel. Thus in the experiment of Fig. 4A, 10 µM YGGFMRFamide alone elicited an inward current of 200 pA at VH = –40 mV and 1 mM 8-Br-cIMP alone produced an inward current of 300–400 pA. However, when YGGFMRFamide was applied in the presence of 8-Br-cIMP, no further response was observed. In five experiments, the FMRFamide-induced and cGMP analogue-induced inward currents occluded one another by 100% at a holding potential of –40 mV.
|
The preceding results showed that exogenous cGMP analogues reproduced the
effect of FMRFamide to activate the Cl– current, whereas
exogenous 8-CPT-cAMP reproduced the effect of SCPB to decrease this
current. Biochemical studies have demonstrated that the sensory neurons are
capable of synthesizing cAMP (Bernier et
al. 1982). Do the sensory neurons possess the requisite machinery
to synthesize cGMP? Previous studies suggested the importance of cGMP for the
induction of long-term hyperexcitability of Aplysia sensory neurons
in response to noxius stimulation (Lewin
and Walters, 1999). Our attempts to measure cGMP in the sensory
neurons using a radioimmunoassay were inconclusive, probably due to the small
amounts of available tissue. Instead, we approached this question indirectly
using zaprinast, a specific inhibitor of the phosphodiesterase that hydrolyzes
cGMP (cGMP PDE) (Choi et al.
2002). Zaprinast alone produced a substantial inward current
response that had a similar slope, shape, and reversal potential as the
current activated by FMRFamide or cGMP analogues (data not shown). Moreover,
nitric oxide (NO), known to activate soluble guanylate cyclase, mimics these
same responses (Armitage et al.
1991). These results indicate that the sensory neurons contain an
active guanylate cyclase.
Effects of inhibitors of cAMP phosphodiesterase
If FMRFamide and 8-Br-cIMP activate the Cl– current by stimulating a PDE that hydrolyses cAMP, inhibitors of the cAMP PDE should inhibit the effects of both FMRFamide and 8-Br-cIMP to elicit the inward current. Indeed, we found that IBMX, an inhibitor of cAMP PDE, reduced the inward current response to FMRFamide. In the experiment of Fig. 4B, 100 µM IBMX (no DMSO) inhibited the peak response to 500 nM YGGFMRFamide at –80 mV by 80%. Both 100 and 200 µM IBMX produced similar and robust inhibition of the responses to FMRFamide or YGGFMRFamide, decreasing them on average by 78% (n = 4). This inhibitory effect of IBMX was observed in all experiments.
IBMX also inhibited the inward current response to 8-Br-cIMP. Figure 5A illustrates an experiment in which 1 mM 8-Br-cIMP induced a peak inward current of 630 pA at –80 mV. After perfusion of the sensory cell with 200 µM IBMX for 5–6 min, the peak response to 8-Br-cIMP was reduced to 280 pA. On average, 100 µM IBMX produced a 26% inhibition (n = 2) and 200 µM IBMX produced a 51% inhibition (n = 2) of the 8-Br-cIMP-induced difference current at –80 mV. The inhibitory effect of IBMX on the inward current responses to either neuropeptides or 8-Br-cIMP was consistently observed in all experiments where examined.
|
A more specific inhibitor of cAMP PDE, RO 20-1724
(Purcell and Carew 2001;
Wells and Miller 1988), also
inhibited the response to 8-Br-cIMP (Fig.
5B). On average, 100 µM RO 20-1724 inhibited the
8-bromo-cIMP difference current at –80 mV by 28% (n = 2); 250
µM RO 20-1724 produced an average inhibition of 67 ± 28% (n
= 7). RO 20-1724 had little or no effect on the holding current in these
experiments.
In contrast to its inhibitory effect on the 8-Br-cIMP-induced current, RO
20-1724 had little or no effect on the magnitude of the response to
SCPB. Figure
5C illustrates an experiment in which application of 10
µM SCPB evoked a peak outward current of 1.4 nA at –80 mV.
After perfusion with 250 µMRO 20-1724 (0.1% DMSO), the SCPB
difference current at –80 mV was virtually unchanged (1.4 nA;
Fig. 5C). On average,
the magnitude of the SCPB difference current after 10 min of
perfusion with 250 µM RO 20-1724 was 91 ± 14% of the response prior
to application of RO 20-1724 (n = 4). However, although the
magnitude of the SCPB response was unchanged upon inhibition of
cAMP-PDE, the time course of the current response was prolonged
threefold, from 2–3 to 8–9 min
(Fig. 5D), consistent
with a prolonged elevation of cAMP levels upon inhibition of hydrolysis. In
three other experiments, the holding current response to SCPB never
fully recovered after inhibition of cAMP-PDE.
Ionic basis of the current responses
The extrapolated reversal potentials for the SCPB and FMRFamide
I-V curves are 0 mV, as described in the preceding text. One
possibility, then, is that the current they modulate is a nonselective cation
current. Several groups have indeed described an inward
Na+-dependent current that is modulated in opposing ways by
FMRFamide and 5-HT (Ichinose and McAdoo
1988; Ruben et al.
1986; Taussig et al.
1989). According to this view, FMRFamide would activate this
Na+ current and SCPB would inhibit it. If this is indeed
the case, external perfusion with a low-Na+ solution should reduce
the amplitude of the inward current activated by FMRFamide (or cGMP analogues)
and shift its reversal potential (Erev) to more
hyperpolarized voltages. Similarly, the low-Na+ solution should
decrease the magnitude of the outward SCPB sensitive current at
negative potentials (because there would be a smaller initial inward
Na+ current for SCPB to deactivate) and also shift its
reversal potential to more negative potentials. In contrast to these
expectations, we found that perfusion with a low-external-Na+
solution (reduced to 37% of normal) caused little or no change in
Erev (if anything, producing a slight positive shift) and
slightly enhanced the outward current response to 10 µM SCPB
(Fig. 6A). The average
Erev in the presence of low-Na+ solution was +7
mV (n = 4) relative to Erev in normal ASW,
whereas the expected shift for a Na+-selective current under these
conditions is –25 mV. On average, responses to 10 µM SCPB
were increased 1.7-fold during perfusion with low-Na+ solution
(n = 4). The basis for the enhancement of the SCPB-induced
current by the low-Na+ solution is unclear.
|
We next determined the sensitivity of the inward current response to the external Na+ concentration (Fig. 6B). In these experiments, we elicited the inward current with 8-Br-cIMP, which activates the same current as FMRFamide (Fig. 4A) and 8-Br-cGMP (Fig. 2C). Although perfusion with a low-Na+ solution (37%) did inhibit the inward current response to 1 mM 8-Br-cIMP, it did not significantly shift the Erev. On average, the magnitude of the current response to 1 mM 8-Br-cIMP was inhibited to 47% of its initial level during perfusion with low-Na+ solution (n = 3). In contrast, the average shift in Erev was +2 mV (n = 3), whereas the expected shift in Erev for a Na+ current under these conditions is –25 mV. Thus although changes in the external Na+ concentration do affect the amplitude of the currents evoked by SCPB and 8-Br-cIMP, the reversal potentials remained largely unchanged. Therefore Na+ could not be the charge carrier for these currents.
Cl– was a second candidate for the charge carrier since its equilibrium potential was –2 mV under our whole cell recording conditions ([Cl–]i = 555 mM and [Cl–]o = 600 mM). We measured Erev for difference current I-V curves elicited in 600 and 200 mM Cl–o. To keep the reversal potential negative, which is necessary to ensure a large current due to its inward rectification, the internal Cl– concentration was set to 200 mM (using methane sulfonate as the replacement anion). To further enhance the magnitude of the Cl– current, we used a prepulse protocol (see following text) in which a 5-s hyperpolarizing pulse to –80 mV from a holding potential of –40 mV preceded the voltage-clamp ramp (500-ms ramp from –80 to +20 mV). 100 µM TTX was also included in the external solution to allow control of the voltage ramp at depolarized potentials.
Reduction of external Cl– from 600 to 200 mM altered both the magnitude of the 8-CPT-cAMP- and 8-Br-cIMP-sensitive currents and shifted their reversal potentials to more positive potentials, in a manner consistent with a Cl–-selective current (Fig. 7). Thus the reduction of external Cl– increased the positive current response to 8-CPT-cAMP at negative potentials (Fig. 7A) and shifted the reversal potential by +29 mV (n = 2), similar to the shift predicted for a Cl– selective current under these conditions (+27.5 mV). Similarly, reduction of external Cl– enhanced the magnitude of the negative current elicited by 8-Br-cIMP (Fig. 7B) and shifted the reversal potential by +24.5 mV (n = 2).
|
FMRFamide antagonized the action of SCPB
The preceding data suggested that SCPB and FMRFamide modulated the same Cl– current in opposing directions via the up- and downregulation of cAMP levels. This hypothesis led to the prediction that FMRFamide should antagonize the action of SCPB. In addition, the effect of FMRFamide to activate the Cl– current in the absence of SCPB should require that resting levels of cAMP be sufficiently elevated so that the Cl– current would be tonically inhibited. Figure 8 illustrates an experiment that was consistent with this model.
|
In this particular cell, 10 µM SCPB initially elicited a large outward current, due to Cl– current inhibition (Fig. 8B). In the same cell, an initial application of 10 µM YGGFMRFamide failed to elicit any current response (Fig. 8A). However, during a prolonged application of SCPB to the same cell, application of YGGFMRFamide now succeeded in eliciting a large increase in the Cl– current (Fig. 8C) that reversed the response to SCPB. Perfusion of the sensory neuron with RO 20-1724 had no effect on the magnitude of the Cl– current inhibition with SCPB (Fig. 8D), consistent with our previous results (Fig. 5C). However, RO 20-1724 did inhibit the action of YGGFMRFamide to increase the Cl– current in the presence of SCPB (Fig. 8E). These results were consistent with the view that cAMP levels in this cell were initially too low to cause significant tonic inhibition of the Cl– current. As a result, there was no preexisting tonic inhibition for YGGFMRFamide to relieve. However, once Cl– channels were inhibited by elevated cAMP levels in response to SCPB, YGGFMRFamide was able to enhance the Cl– current by relieving this inhibition; this effect required PDE activation.
Effects of SCPB and 8-Br-cIMP on a slow time-dependent inward rectifier
Several Aplysia neurons have been shown to contain slow
voltage-dependent Cl– currents, which increase with
hyperpolarization and decrease with depolarization, with kinetics that are on
the time scale of many seconds
(Chesnoy-Marchais 1983;
Lotshaw and Levitan 1987;
Thompson and Ruben 1988).
Moreover, hyperpolarization-activated Cl– currents are
inhibited by cAMP (Lotshaw and Levitan
1987) and enhanced by Cl– loading
(Chesnoy-Marchais 1983),
leading us to ask whether the Cl current modulation in the sensory neurons
that we have studied here is due regulation of a hyperpolarization-activated
Cl conductance. We therefore used prolonged hyperpolarizing voltage steps to
determine whether such a current was present in the sensory neurons and
whether it was subject to antagonistic modulation.
Hyperpolarizing steps from –40 to –80 mV did indeed activate a slow inward current in the sensory neurons (Fig. 9). We further found that the magnitude of this current was greatly enhanced during whole cell dialysis with elevated intracellular Cl– (data not shown) and was sensitive to changes in external Cl–, consistent with this current representing the hyperpolarization-activated Cl– current. Importantly, we found that this current was modulated in an identical manner to the Cl– current characterized above with voltage ramps. Thus application of SCPB inhibited the slow time-dependent inward current, whereas 8-Br-cIMP enhanced the current (Fig. 9). On average, 10 µM SCPB inhibited the slow inward current relaxations by 21 ± 8% (n = 3), whereas 1 mM 8-Br-cIMP caused a 30% enhancement. Thus the modulation of the slow time-dependent inward rectifier current mirrors that seen for the Cl– current present at the resting potential of –40 mV and during subsequent voltage ramps. SCPB inhibited both currents and 8-Br-cIMP enhanced both currents, leading us to tentatively conclude that the modulation of the same hyperpolarization-activated Cl– channel underlies both sets of currents.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
). The voltage dependence and ion selectivity of the current
responses we have studied here indicate that they represent opposing
modulation of a hyperpolarization-activated Cl– current.
Because this Cl– current is highly sensitive to levels of
intracellular chloride, it becomes particularly prominent when the internal
Cl– concentration is high, as in our whole cell recording
conditions. Identification of a cyclic nucleotide regulated, hyperpolarization-activated Cl– current
The demonstration of Cl– current modulation in the sensory
neurons suggests that modulation of the inward rectifier Cl–
current may be more widespread than previously realized. The current itself
appears to be ubiquitous in its phylogenetic distribution and may be quite old
in evolutionary terms. For example, the existence of a slow
hyperpolarization-activated Cl– current has been reported in
a wide range of tissues including green algae, Torpedo electroplax,
molluscan neurons, toad skin, crayfish and frog muscle, rabbit kidney, and rat
sympathetic and hippocampal pyramidal neurons
(Chesnoy-Marchais 1983;
Madison et al. 1986;
Sansom et al. 1990). Although
the molecular identity of the Aplysia Cl– current
remains unknown, its slow hyperpolarization-dependent activation and strong
dependence on internal Cl– concentration suggest that it is a
member of the ClC family of ion channels
(Jentsch et al. 2002). While
the current itself seems to be prevalent in a wide range of tissues,
relatively few studies have reported its modulation
(Lotshaw and Levitan 1987;
Madison et al. 1986;
Thompson and Ruben 1988).
The current has several unusual properties that may contribute to its
modulation having been previously overlooked or the current misidentified.
First, the current often requires internal Cl– loading for
strong expression, and this may vary from preparation to preparation. Thus
Cl– loading is mandatory for expression of
hyperpolarization-activated Cl– currents in some neurons
(Chesnoy-Marchais 1983;
Lotshaw and Levitan 1987), but
not in other neurons (Madison et al.
1986; Thompson and Ruben
1988). In those cells that require high internal
Cl–, attempts to measure reversal potential changes with a
low external Cl– solution may be thwarted by depletion of
internal Cl–, which will inhibit the current
(Chesnoy-Marchais 1983;
Thompson and Ruben 1988).
Another unusual feature of the Cl– current response is its
sensitivity to extracellular Na+. Thus we found that the response
to both FMRFamide and 8-Br-cIMP was inhibited upon lowering the external
Na+ concentration. Although the reversal potential was not
Na+ sensitive, the effect of external Na+ could lead to
the false conclusion that the current was indeed carried by Na+.
Ichinose and McAdoo (1988,
1989) reported that FMRFamide
and cGMP induce a slow voltage-dependent inward current in Aplysia
neuron R14, similar to the effect of FMRFamide in the sensory neurons reported
here. As in the case of the sensory neurons, reduction of
Na+o inhibited the FMRFamide-induced current in R14,
leading the authors to conclude that the inward current is likely to be
carried by Na+. However, the current was also sensitive to changes
in external Cl–, raising the possibility that it represents
the same current we have studied here.
Role of cAMP and cGMP in antagonistic Cl– current regulation
The effect of SCPB to decrease the Cl– current
depends on the stimulation of synthesis of cAMP
(Jarrard et al. 1993). Thus
the effects of SCPB were mimicked by 8-CPT-cAMP, a membrane
permeable cAMP analogue, and prolonged by RO 20-1724, an inhibitor of cAMP
PDE. These results are in agreement with a previous study in Aplysia
by Lotshaw and Levitan (1987)
showing that 5-HT decreases the Cl– current through a cAMP
cascade. Their conclusion was based on the ability of forskolin to mimic the
response to 5-HT. We also provided evidence for the involvement of the
AMP-dependent protein kinase, based on the ability of the non-specific kinase
inhibitor, K252a, to antagonize the actions of both SCPB and
8-CPT-cAMP.
Furthermore, the ability of FMRFamide to activate the Cl–
conductance appears to be due to its ability to antagonize the cAMP cascade,
an effect that is mediated by cGMP. Thus we found that cGMP analogues and
zaprinast (an inhibitor of cGMP PDE) mimic the effects of FMRFamide. Moreover,
cGMP analogues occlude the response to FMRFamide. Ichinose and McAdoo
(1989) have obtained similar
evidence for a role of cGMP in mediating the effects of FMRFamide to increase
an inward current in Aplysia neuron R14. We have also previously
found that nitric oxide can stimulate cGMP synthesis in Aplysia
sensory cells (Armitage et al.
1991).
Our results further suggest that the most likely mode of action of cGMP is
to stimulate the cAMP PDE, leading to a decrease in cAMP and thus an increase
in Cl– current. Thus 8-Br-cGMP, which elicited
Cl– current responses that mimicked the response to
FMRFamide, is an agonist for several cGMP targets, including cGMP-dependent
protein kinase (Corbin et al.
1988a,b).
Evidence that this effect was due to PDE activation came from studies of
8-Br-cIMP, which is more selective for the cGMP-activated PDE and which
produced a large and robust Cl– current activation. Further
support for this view was provided by the findings that IBMX and RO 20-1724,
two inhibitors of cAMP PDE, inhibited the responses to FMRFamide and
8-Br-cIMP. Similar cGMP-dependent actions due to cAMP PDE stimulation have
been reported for Ca2+ channel modulation in the heart
(Fischmeister and Hartzell
1987) and hippocampal neurons
(Doerner and Alger 1988).
The mechanism by which FMRFamide stimulates cGMP synthesis was not
addressed by our experiments. Two possibilities are stimulation of soluble
guanylate cyclase through nitric oxide
(Garthwaite 1991) or direct
activation of a membrane receptor guanylate cyclase
(Wedel and Garbers 2001).
However, neuronal NO synthase is a calcium-dependent enzyme and FMRFamide has
been shown to reduce, not elevate, intracellular calcium in Aplysia
sensory neurons (Blumenfeld et al.
1990). In some systems, arachidonic acid or its metabolites can
activate guanylate cyclase and thereby elevate levels of cGMP
(Kiesel and Catt 1987).
12-HPETE and 5-HPETE metabolites of arachidonic acid, but not 12-HETE or
5-HETE, stimulate cGMP formation in Aplysia neural tissue
(Piomelli et al. 1987a).
Moreover, FMRFamide stimulates the synthesis of 12-HPETE in the sensory
neurons (Piomelli et al.
1987b). Thus arachidonic acid-induced elevation of cGMP is an
attractive mechanism for the activation of the inward rectifier
Cl– current by FMRFamide.
One problem with this model is that FMRFamide has not been so far found to
reduce cAMP levels in the sensory neurons
(Ocorr and Byrne 1985;
Sweatt et al. 1989). However,
these previous studies were carried out in the presence of IBMX, which would
have eliminated any contribution of PDE activation. Moreover, the effect of
PDE activation may be more pronounced in cells where cAMP has previously been
elevated (Fischmeister and Hartzell
1987) (see Fig. 6).
It is also possible that FMRFamide may cause the direct dephosphorylation of
the Cl– channel, independent of changes in PDE activity
(Sweatt et al. 1989) due to
either phosphatase activation (Ichinose et
al. 1990) or direct kinase inhibition
(Piomelli et al. 1989).
Finally, FMRFamide may also recruit other phosphorylation-independent
mechanisms for modulating the Cl– channel
(Buttner et al. 1989;
Fesenko et al. 1985).
The activation of a PDE by FMRFamide may well contribute to it's modulation
of other sensory neuron proteins in addition to the Cl–
channel. For example, PDE activation may act in parallel with phosphatase
activation to mediate FMRFamide's antagonism of S K+ channel
closure by 5-HT or cAMP (Belardetti et al.
1987; Endo et al.
1995). Similarly, PDE activation and consequent inhibition of
CREB1 phosphorylation may act in parallel with CREB2 activation to mediate
FMRFamide's inhibition of long-term facilitation by 5-HT
(Guan et al. 2002). However,
it is also possible that any effect of PDE activation may be limited to the
Cl– channel, given the existence of local signaling complexes
(Davare et al. 2001) and
restricted cyclic nucleotide microdomains
(Rich et al. 2001).
Physiological role for Cl– current regulation
The physiological role of the inward rectifier Cl– current
in Aplysia sensory neurons at present is not clear. Previous studies
have shown that under physiological conditions, the predominant effects of
cAMP modulation are a decrease in resting K+ conductance, due to
closure of the S-K+ channel
(Goldsmith and Abrams 1992;
Klein et al. 1982; Siegelbaum et al.
1982), a decrease in a delayed rectifier K+ current
(Goldsmith and Abrams 1992),
and modulation of a calcium-activated K+ current
(Walsh and Byrne, 1989). In
addition, cAMP also enhances the magnitude of an L-type voltage-gated
Ca2+ current
(Edmonds et al. 1990). In
contrast, FMRFamide is known to enhance the S-K+ current
(Belardetti et al. 1987),
inhibit a dihydropyridine-insensitive Ca current important for transmitter
release (Edmonds et al. 1990)
and to antagonize cAMP-mediated protein phosphorylation through activation of
a phosphoprotein phosphatase (Endo et al.
1995; Sweatt et al.
1989). In this context of pleiotropic actions of these
transmitters, a natural question arises as to the physiological significance
of the Cl– current modulation we have identified here,
especially given the necessity of using internal Cl– loading
to observe a significant effect.
The role of a hyperpolarization-activated Cl– current is
perhaps most clearly understood in Torpedo where its subcellular
localization contributes to the formation of a voltage gradient across the
cell, creating a "battery" for the electric organ. Madison et al.
(1986) proposed that a similar
current may be specifically localized to the dendritic membrane in hippocampal
pyramidal cells and thereby may regulate dendritic excitability. Although
Cl– loading is required in some preparations
(Chesnoy-Marchais 1983;
Lotshaw and Levitan 1987), in
other cases a large inward rectifier current is clearly observed in the
absence of Cl– loading for expression
(Madison et al. 1986;
Thompson and Ruben 1988). One
interesting possibility is that the Cl– current may be more
important at early developmental stages, where the internal
Cl– concentration is high due to a lack of expression of a
Cl– pump at this developmental stage
(Rivera et al. 1999). The
current may also become activated during intense bouts of inhibitory synaptic
input, which can cause Cl– loading of cellular processes
(Chesnoy-Marchais 1983;
Lotshaw and Levitan 1987). The
hyperpolarization-activated Cl– current may also,
paradoxically, predominate when a cell is strongly excited. Thus
Chesnoy-Marchais (1983) found
that in Aplysia cerebral ganglion A neurons, Cl–
loading is no longer required to activate the Cl– current
after a short train of depolarizing pulses. This feature of the current may be
particularly relevant in the Aplysia sensory cells where
activity-dependent modulation has been shown to contribute to plasticity
(Small et al. 1989).
Even under conditions where the Cl– current may be
functional, it is difficult to predict the relative contribution of
Cl– current modulation to the actions of SCPB and
FMRFamide. Although the magnitude of the Cl– current varied
considerably from preparation to preparation, its modulation typically
produced smaller amplitude currents than modulation of the S K+
current. The contribution of Cl– current modulation relative
to that of other sensory neuron currents
(Baxter et al. 1999;
Goldsmith and Abrams 1992)
will depend on a variety of factors, including
[Cl–]i and cellular activity (see preceding text),
the levels of resting cAMP and cGMP, and the membrane voltage. In normal
resting neurons, EK and ECl lie near
the resting potential. Thus opening of both K+ and
Cl– channels tend to oppose excitability and help stabilize
the cell by contributing to repolarization. As a result, the decrease in input
resistance produced by up modulating the Cl– current with
FMRFamide may decrease cellular excitability, an effect synergistic with the
inhibitory actions of FMRFamide to activate K+ current
(Belardetti et al. 1987) and
inhibit Ca2+ current
(Edmonds et al. 1990). In
contrast, the effect of SCPB to inhibit Cl–
conductance may cause an increase in cellular excitability, also producing a
synergistic effect with the up-modulation of Ca2+
current and down-modulation of S-K+ current produced by this
neuropeptide in these cells.
The coordinate targeting of these different channels by modulatory pathways may serve to extend the range of potentials over which such pathways can exert regulatory control. Thus the Cl– current may be dominant at hyperpolarized voltages because of its voltage dependence and because of the proximity of EK to such negative potentials. In contrast, the S-K+ current tends to be the predominant background conductance at more depolarized voltages because it does not inactivate with depolarization and because EK is normally very negative. The ability of individual neurotransmitters to produce divergent actions through modulation of multiple ion channels simultaneously can thus provide a means to influence cellular activity over a wide range of conditions. In addition, our results elucidate the way in which the actions of multiple transmitters can converge on a single ion channel, in this case the hyperpolarization-activated Cl– channel. These convergent properties are important in allowing the cell to integrate information from a number of sources.
|
DISCLOSURES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Present address of N. Buttner, Laboratory for Structural Neuroscience, Dept. of Psychiatry, McLean Hospital, Belmont, MA 02478.
|
FOOTNOTES |
|---|
Address for reprint requests: S. A. Siegelbaum, Center for Neurobiology and Behavior, Columbia University, 722 W. 168 St., New York, NY 10032 (E-mail: sas8{at}columbia.edu).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Ascher P, Marty A, and Neild TO. Effects of procaine on voltage relaxation of acetylcholine induced currents in Aplysia neurons [proceedings]. J Physiol 263: 121P–122P, 1976.[Medline]
Bailey CH,
Giustetto M, Zhu H, Chen M, and Kandel ER. A novel function for
serotonin-mediated short-term facilitation in aplysia: conversion of
a transient, cell-wide homosynaptic hebbian plasticity into a persistent,
protein synthesis-independent synapse-specific enhancement. Proc
Natl Acad Sci USA 97:
11581–11586, 2000.
Baxter DA and
Byrne JH. Serotonergic modulation of two potassium currents in the pleural
sensory neurons of Aplysia. J Neurophysiol
62: 665–679,
1989.
Baxter DA,
Canavier CC, Clark JW Jr, and Byrne JH. Computational model of the
serotonergic modulation of sensory neurons in Aplysia. J
Neurophysiol 82:
2914–2935, 1999.
Belardetti F, Kandel ER, and Siegelbaum SA. Neuronal inhibition by the peptide FMRFamide involves opening of S K+ channels. Nature 325: 153–156, 1987.[Medline]
Belardetti F and Siegelbaum SA. Up- and down-modulation of single K+ channel function by distinct second messengers. Trends Neurosci 11: 232–238, 1988.[ISI][Medline]
Bernier L, Castellucci VF, Kandel ER, and Schwartz JH. Facilitatory transmitter causes a selective and prolonged increase in adenosine 3':5'-monophosphate in sensory neurons mediating the gill and siphon withdrawal reflex in Aplysia. J Neurosci 2: 1682–1691, 1982.[Abstract]
Blumenfeld H, Spira ME, Kandel ER, and Siegelbaum SA. Facilitatory and inhibitory transmitters modulate calcium influx during action potentials in aplysia sensory neurons. Neuron 5: 487–499, 1990.[ISI][Medline]
Braha O, Dale
N, Hochner B, Klein M, Abrams TW, and Kandel ER. Second messengers
involved in the two processes of presynaptic facilitation that contribute to
sensitization and dishabituation in Aplysia sensory neurons.
Proc Natl Acad Sci USA 87:
2040–2044, 1990.
Brown BS and Yu SP. Modulation and genetic identification of the M channel. Prog Biophys Mol Biol 73: 135–166, 2000.[ISI][Medline]
Buttner N, Siegelbaum SA and Volterra A. Direct modulation of Aplysia S-K+ channels by a 12-lipoxygenase metabolite of arachidonic acid. Nature 342: 553–555, 1989.[Medline]
Byrne JH and
Kandel ER. Presynaptic facilitation revisited: state and time dependence.
J Neurosci 16:
425–435, 1996.
Chesnoy-Marchais D. Characterization of a chloride conductance
activated by hyperpolarization in Aplysia neurons. J
Physiol 342:
277–308, 1983.
