JOURNAL. Vol. 62.

OF

NEUROPHYSIOLOC;~

No. 3. September

1989. Prruied

IH L..S..d.

Serotonergic Modulation of Two Potassium Currents in the Pleural Sensory Neurons of Aplysia
DOUGLAS Department
SUMMARY

A. BAXTER AND JOHN H. BYRNE ~fNeurohio/ogv e and Anatomy, The University of Texas Medical
AND CONCLUSIONS

School, Houston,

Texas 77225

1. The properties of membrane currents that were modulated bv serotonin (5-HT) were investigated with two-electrode voltage-clamp techniques in sensory neuron somata isolated from the pleural ganglion of Ap/evsia cal(fh-nica. The modulatory effects of 5-HT were revealed by computer subtraction of current responses elicited in the presence of 5-HT from current responses elicited prior to the application of 5-HT. The complexities of the resulting 5-HT difference currents (J5-& suggested that 5-HT modulated more than one component of membrane current. 2. The 5-HT difference currents appeared to have at least two distinct components. One component was clearly evident at membrane potentials more negative than - 10 mV was relatively voltage independent and did not inactivate. A second component was activated at membrane potentials more positive than - 10 mV, had complex kinetics, and was highly voltage dependent. In an attempt to identify the membrane currents that were modulated by 5-HT, we compared the pharmacologic sensitivity of ZSeHT. that of previously described K+ currents. to 3. The two components of 15-H1 had different sensitivities to agents that block KS currents. The relatively voltage-independent component of IS-H7- was not blocked by 2 mM 4-aminopyridine (4-AP) and was relatively insensitive to tetraethylammonium (TEA) (estimated Kd of 92 mM). In contrast, the voltage-dependent component of 15-HT.was blocked by 4-AP (2 mM) and moderate concentrations of TEA (estimated Kd of 5 mM). 4. The KS current blockers that were used to examine 15-HT were also used to examine voltage-activated membrane currents. Externally applied TEA blocked the delayed or voltage-dependent KS current (IK r-) with an estimated dissociation constant (&) of 8 mM and a membrane current similar to the Ca2+-activated K+ current (IK,& with an estimated Kd of 0.4 mM. In addition, externally applied 4-AP (2 mM) blocked IKqI. Thus TEA and 4-AP were equipotent in blocking both IK.l. and the voltage-dependent component of 15-f11.. 5. The suggestion that ls-l-l.l. contained multiple components was supported further by examining the modulatory effects of adenosine 3,5-cyclic monophosphate (CAMP) that mediates some actions of 5-HT on membrane currents in these cells. CAMP difference currents (IcAMP) were similar to the relatively voltageindependent component of JSeHT. The subsequent addition of 5-HT to solutions already containing CAMP resulted in 5-HT difference currents similar to the voltage-dependent component of 15-H,T.Thus the voltage-dependent component of 15+,T did not appear to be modulated by elevated levels of CAMP. 6. Similarly, the modulatory effects of small cardioactive peptide (SCPb) that mimics 5-HT and elevates levels of CAMP revealed difference currents (I s(Ph) that were similar to both IcAMP and the relatively voltage-independent component of ]5-HT. The subsequent addition of 5-HT to solutions already containing SCPh resulted in 5-HT difference currents similar to the voltagedependent component of JSmHT.. Thus SCPb did not appear to modulate a current similar to the voltage-dependent component of &-1,-l *
0022-3077/89 $1 SO Copyright

7. Based on pharmacologic sensitivity and voltage dependence, we conclude that in somata of pleural sensory neurons brief depolarizing voltage-clamp pulses activate at least two currents that could be modulated by 5-HT. The relatively voltage-independent component had properties consistent with the previously described S current. The voltage-dependent component had properties similar to IK,L,.

INTRODUCTION

In Aplysia cafifornica, cellular mechanisms underlying neuronal plasticity have been analyzed extensively in two populations of sensory neurons. One population is located in the abdominal ganglion and mediates the siphon-gill withdrawal reflex. The second population is located in the pleural ganglion and mediates the tail withdrawal reflex. Many analyses of the biophysical mechanisms of plasticity in these sensory neurons have focused on a serotonin (5-HT)-sensitive K+ current, which is termed the S current (Klein et al. 1982; Pollock et al. 1985). Extracellular application of 5-HT reduces the magnitude of the S current via CAMP-dependent protein phosphorylation that closes the channel (Klein et al. 1982; Siegelbaum et al. 1982, 1987; Camardo et al. 1983; Pollock et al. 1985, 1987; Shuster et al. 1985; Belardetti and Siegelbaum 1988). The S current contributes to the resting K+ conductance and to the repolarization of action potentials in these cells. The serotonergic modulation of the S current is postulated to be an important mechanism contributing to presynaptic facilitation of transmitter release from sensory neurons, which in turn is thought to be the cellular basis of several simple forms of learning (for reviews see Byrne 1985, 1987; Carew et al. 1986, 1987; Hawkins et al. 1986; Kandel and Schwartz 1982). Several criteria are used to distinguish the S current from other KS currents that are found in molluscan neurons (e.g., Thompson 1977; Adams et al. 1980a; Byrne 1980), including its voltage dependence, kinetics, and pharmacologic sensitivity. The S current is relatively voltage independent, active over a wide range of membrane potentials, noninactivating, increases in a time-dependent manner during depolarization, and is relatively insensitive to the K+ channel blockers 4-aminopyridine (4-AP) and tetraethylammonium (TEA) (Klein et al. 1980, 1982; PaupardinTritsch et al. 198 1; Siegelbaum et al. 1982, 1987; Camardo et al. 1983; Walsh and Byrne 1984, 1989; Pollock et al. 1985, 1987; Shuster and Siegelbaum 1987; Scholz and Byrne 1987, 1988; Brezina et al. 1987, 1988). More recent
665

0 1989 The American Physiological Society

666

D. A. BAXTER

AND

J. H. BYRNE

results indicate, however, that in addition to the S current other membrane currents are modulated by 5-HT. For example, Walsh and Byrne (1989) reported that in the pleural sensory neurons ofAplysia the modulatory effects of 5-HT on membrane currents were reduced by agents that block a Ca2+-activated K+ current [e.g., low concentrations of extracellular TEA, replacement of extracellular Ca2+ with Ba2+, or intracellular injection of the Ca2+ chelator ethylene glycol-bis(P-aminoethyl ether)-N,N,N,N-tetraacetic acid (EGTA)]. Furthermore, they reported that 5-HT reduced the outward currents that were activated by intracellular injections of Ca2+. Thus it appears that in addition to the S current, 5-HT also modulates a steady-state Ca2+sensitive K+ current that contributes to the resting conductance in these cells. This finding raises the possibility that 5-HT may modulate additional IV currents in the pleural sensory neurons. In the present study, we tested this hypothesis by examining the modulatory effects of 5-HT on membrane currents that were elicited by a series of depolarizing voltageclamp pulses. The modulatory effects of 5-HT were revealed by computer subtraction of current responses that were elicited by brief voltage-clamp pulses in the presence of 5-HT from current responses that were elicited by identical voltage-clamp pulses prior to application of 5-HT. The resulting 5-HT difference currents (15-r-& were found to have at least two distinct components: a relatively voltageindependent component and a highly voltage-dependent component. By comparing the pharmacologic sensitivity, voltage dependence, and kinetics of these two components of 15-HTto those of several K+ currents previously identified in the somata of these sensory neurons and other molluscan neurons (e.g., Connor and Stevens 197 1; Thompson 1977; Byrne 1980; Adams et al. 1980b; Gorman and Thomas 1980; Klein et al. 1982; Hermann and Hartung 1983; Hille 1984; Latorre et al. 1984; Pollock et al. 1985; Deitmer and Eckert 1985; Walsh and Byrne 1985, 1989; Shuster and Siegelbaum 1987; Acosta-Urquidi 1988; Rudy 1988), we concluded that in addition to the S current and a steady-state Ca2+ -sensitive K+ current, 5-HT also modulates the delayed or voltage-dependent K+ current (1k V). Furthermore, a membrane-permeable analogue of cA&IP and small cardioactive peptide (SCPb) were found to modulate a component of membrane current identical to the relatively voltage-independent component of 15-HT, but they did not mimic the action of 5-HT on a the voltage-dependent component of membrane current. Thus the modulation of I k,V by 5-HT did not appear to be mediated via CAMP. The possible mechanism and functional significance of this novel action of 5-HT are discussed. Preliminary reports of these results have been presented (Baxter and Byrne 1986, 1987).
METHODS

Aplysia ca&mica (150-300 g) were obtained from Marine Specimens (Pacific Palisades, CA), Marinus Biomarine (Westchester, CA), Sea Life Supply (Sand City, CA) and Alacrity Marine Biological Services (Redondo Beach, CA). Animals were maintained at 15C in aerated artificial seawater (ASW), which is commercially available (Instant Ocean, Aquarium Systems, Mentor, OH), and were fed a diet of dried seaweed. Dissections were performed after anesthetizing the animals bv the iniection of

a volume of isotonic MgC& equal to about one-half of the animals volume. The ventrocaudal (VC) clusters of the pleural ganglia contain the somata of -200 mechanoafferent neurons that appear to be homogeneous in their biochemical and electrophysiological properties (Walters et al. 1983a,b; Walsh and Byrne 1984, 1989; Ocorr and Byrne 1985, 1986; Pollock et al. 1985, 1987; Scholz and Byrne 1987, 1988). For these experiments, clusters of sensory neuron somata were surgically isolated by undercutting the VC cluster with iridectomy scissors. The isolated clusters were pinned to the Sylgard (Dow-Corning) floor of an experimental chamber with a volume of 300 ~1. The static ASW bathing solution was buffered to pH 7.6 with lo-30 mM Trizma (Sigma) and contained 150 PM tetrodotoxin (TTX; Sigma). During all experiments the bath temperature was regulated to 15 (+ 1C) by a thermoelectric cooling unit placed under the experimental chamber. Sensory neuron somata were impaled with two glass capillary microelectrodes that were filled with 3 M potassium acetate and that had resistances of 2-6 MQ. The sensory neurons were voltage clamped at their resting membrane potential using an Axoclamp2 amplifier (Axon Instruments) and conventional two-electrode voltage-clamp techniques. Membrane conductances were activated by 200 ms voltage-clamp pulses that were preceded by a 25 s, hyperpolarizing prepulse to either -50, -70, or -90 mV. After the 200 ms depolarization, the membrane potential was stepped back to a value of -50 mV for 500 ms. To avoid the cumulative inactivation of membrane conductances (e.g., Aldrich et al. 1979; Byrne 1980), the voltage-clamp pulses were separated by 90 s. During times between the standard voltage-clamp protocol (see above), the cells were returned to their resting membrane potential. Membrane current responses and a calibration pulse were digitized (12 bit resolution) on-line with a laboratory microcomputer and stored for later display and analysis. During each membrane current response, 500 samples were taken at a sampling rate ( 1.67 kHz) that was sufficiently fast to retain the details of the original current waveform. A minimum of three stable membrane current responses were obtained at each voltage in the series of voltageclamp pulses before and after addition of compounds to the bathing solution. These stable responses were averaged off-line into a single current trace for each potential in the series of voltageclamp pulses. This procedure insured that these data represented stable responses to the applied compounds and not nonspecific run-down, increased the signal-to-noise ratio, and minimized the contamination of the difference currents by small random fluctuations in the currents. Although the membrane voltages during the voltage-clamp pulses were not digitized, they were monitored constantly. Current records were accepted only if the step changes in membrane potential were completed in < 1 ms and no voltage sag was detected. The series resistance was estimated under current-clamp conditions from the size of the initial step change in a potential recorded by one intracellular microelectrode when a 10 nA current step was injected via the other microelectrode (Finkel and Gage 1985). The average series resistance for this preparation was 4 kQ (n = 17). No compensation for series resistance was used. The effects on membrane currents of pharmacologic agents added to the bath were revealed by computer subtraction of membrane currents elicited after drug treatment from membrane currents elicited before drug treatment (see Byrne 1980 and Fig. 1). Small (30-50 ~1) aliquots of ASW containing concentrated solutions of either serotonin creatinine sulfate (5-HT; Sigma), tetraethylammonium chloride (TEA; Kodak), 4-aminopyridine (4-AP; Sigma), 8-(4-parachlorophenylthio)-cyclic AMP (8-pcptCAMP; Sigma) or small cardioactive peptide (SCP,; Peninsula Laboratories) were added directly into the experimental chamber. The effective range of concentrations for TEA. 4-AP. and 5-HT

MODULATION

OF IS+ CURRENTS

667

wereestimatedfrom previously reported dose-response relationships(Thompson 1977; Hermann and Gorman 198la,b; Stanfield 1983; Ocorr and Byrne 1985; Deitmer and Eckert 1985; Shusterand Siegelbaum1987; Brezina et al. 1987, 1988; Rudy 1988) and independently confirmed during the courseof these experiments.The concentrationsand the method of bath application are similar to those usedin other studieson the sensory neurons (e.g., Klein et al. 1980, 1982, 1986; Ocorr and Byrne 1985,1986;Pollock et al. 1985;Hochner et al. 1986b;Shusterand Siegelbaum 1987;Walshand Byrne 1989). Recordingsfrom cellswith the following characteristics were included in the data set: 1) stable resting membrane conductances,2) membranecurrentsthat did not showprogressive, nonspecificalterationsduring repetitive voltage-clamppulses, rest3) ing membranepotentials greater than -40 mV [-43 t 3 mV (meant SD)], and 4) input resistances greaterthan 15 MQ (32 t 8). The resultsof this study representsuccessful recordingsfrom 140preparations.
RESULTS

Modulatory eflects of 5-HT on membrane currents We examined the modulatory effects of 5-HT on membrane currents by isolating 5-HT difference currents (&&. An example of the method used to isolate 15-HTis shown in

Fig. 1. The membrane currents that were elicited in ASW by a voltage-clamp pulse from -90 to -20 mV are shown in Fig. IA1 (trace a). The addition of 5-HT (30 PM) reduced the total membrane current (trace b). This reduction is particularly evident at the end of the voltage-clamp pulse. The difference between the current responses elicited in the presence of 5-HT (trace b) and the current responses elicited in ASW (trace a) represents the membrane current that was modulated by 5-HT. Computer subtraction of the 5-HT trace from the ASW trace results in a 5-HT difference current, I 5-n-r(Fig. lA2), in which the 5-HT-mediated reduction in outward current is represented as an upward deflection. Previous work indicates 5-HT reduces outward membrane current, at least in part, by initiating the closure of a distinct class of 5-HT sensitive K+ channels, the S channels (Klein et al. 1982; Siegelbaum et al. 1982, 1987; Camardo et al. 1983; Pollock et al. 1985, 1987). Thus the example of I 5-H-rshown in Fig. IA2 represents the S current and has properties consistent with those previously reported for the S current (see also below). In contrast to the relatively simple modulation of membrane currents by 5-HT shown in Fig. IA, a complex 5-HT difference current was observed at more depolarized memBl Step to +20 mV

Step to -20 mV

b + 5-HT a ASW

20 msec

5 nA

30 nA 20 msec

B2 A2
5-HT Difference Current (a - b)

5-HT Difference Current (a - b)

20 msec

-J

2 nA

FIG. 1. Modulation of membrane currents by 5-HT at 2 different levels of depolarization. Al: membrane currents were elicited by voltage-clamp pulses from -90 to -20 mV in ASW (trace a) and following bath application of 5-HT (30 PM) (trace b). Note that the membrane currents do not return to the preclamp level since the cell was clamped from a holding potential of -90 to -20 mV and then back to -50 mV (see METHODS). A2: the 5-HT difference current (I,-,,) was isolated by subtracting the current response elicited in the presence of 5-HT from the current response elicited in ASW (a - b). BI: membrane currents from the same cell were elicited by voltage-clamp pulses from -90 to +20 mV in ASW (trace a) and following bath application of 5-HT (30 PM) (trace b). B2: 5-HT difference current (I 5+rr) was isolated by subtracting the current response elicited in the presence of 5-HT from the current response elicited in ASW (a - b). Note changes in ordinate scales. The complexity of the waveform of the 5-HT difference current obtained at +20 mV indicates that 5-HT is modulating not only the S current but also some other component of membrane current.

668

D. A. BAXTER

AND

J. H. BYRNE

brane potentials. The membrane current that was elicited in ASW by a voltage-clamp pulse from -90 to +20 mV are shown in Fig. lB2 (trace a). Application of 5-HT (30 PM) appeared to slow the rate-of-rise of the total membrane current and thus markedly decreased the total current early during the voltage-clamp pulse. In addition, 5-HT appeared to slow the inactivation of the outward membrane current and thus increased the outward current later during the voltage-clamp pulse. The difference between the current responses elicited in the presence of 5-HT and the current responses elicited in ASW represents the modulatory effects of 5-HT at more depolarized levels and is shown in Fig. lB2. The early upward peak of 15-HT in Fig. lB2 reflects the initial decrease in the total membrane current (Fig. 1BI), whereas, the downward shift in Fig. lB2 represents the late increase in the total membrane current at the end of the voltage-clamp pulse after application of 5-HT (Fig. 1BI). A comparison of the 5-HT difference currents shown in Fig. 1, A2 and B2 indicates that the

modulatory effects of 5-HT are more complex than simply initiating the closure of the S channel and that other membrane currents are involved. A complete series of 5-HT difference currents that were isolated from a wide range of voltage-clamp potentials is shown in Fig. 2. The membrane currents that were elicited in ASW by voltage-clamp pulses from -90 to -30, -20, and - 10 mV are shown in A2 and the membrane currents that were elicited by voltage-clamp pulses to 0, + 10, and +20 mV are shown in AI (note the change in gain). The membrane currents that were elicited by identical voltageclamp pulses following application of 5-HT are shown in Fig. 2B. The 5-HT difference currents that were isolated by subtracting the current responses elicited in the presence of 5-HT from the current responses elicited in ASW are shown in Fig. 2C. At membrane potentials more negative than 0 mV (Fig. 2C2), I 5-HT increased slowly during the voltage-clamp pulses, and the peak amplitudes increased nearly linearly with voltage (also see Fig. 8). This relatively

Al

ASW

Bl

ASW + 5-HT

20 msec

B2
-10 mV

-20
-30 J 5 nA

20 msec

C 1 5-HT Difference Currents

(A - B)

0 mV

15 nA 20 msec

-I

FIG. 2. Voltage sensitivity of currents modulated by 5-HT. Membrane currents were elicited by voltageclamp pulses from -90 mV to membrane potentials ranging between -30 and +20 mV. A: current responses were elicited in ASW. Small currents elicited with small depolarizations are shown in A2, whereas large currents elicited by larger depolarizations are shown in A 1. B: current responses to potentials corresponding to those shown in Al and A2 were elicited in the same cell following bath application of 5-HT (30 FM). C: 5-HT difference currents were isolated by subtracting the current responses elicited in the presence of 5-HT from the corresponding current responses elicited in ASW (A - B). Results similar to these were obtained in all preparations (n = 33) in which I 5-Hsr was isolated from neurons bathed in ASW. Note changes in ordinate scales. The component of &-HT expressed at small depolarizations (c2) exhibits little voltage-dependence and has simple kinetics, whereas the component expressed at larger depolarizations (CI) is highly voltage dependent and has complex kinetics.

-20 -30

20 msec

2 nA

MODULATION

OF

CURRENTS

669

voltage-independent component of ZsmHT not inactivate did during the 200 ms depolarizing voltage-clamp pulses. Moreover, it did not inactivate during long-duration (1 min) voltage-clamp pulses to -20 mV (n = 9; data not shown). These properties are consistent with the previous descriptions of S current (Klein et al. 1980, 1982; Siegelbaum et al. 1982, 1987; Camardo et al. 1983; Walsh and Byrne 1984, 1989; Ocorr and Byrne 1985; Pollock et al. 1985, 1987: Scholz and Byrne 1987, 1988). In contrast, the magnitude, rate of activation, and rate of inactivation of 5-HT difference currents that were isolated from voltageclamp pulses to membrane potentials more positive than - 10 mV were highly voltage dependent (Fig. 2Cl). The characteristics of lsmH7 at depolarized membrane potentials (Fig. 2Cl) are inconsistent with the those of the S current. These results suggest that 5-HT modulates an additional voltage-dependent membrane conductance that is activated at membrane potentials more positive than - 10 mV.

on IK,Ir. Subsequent increases in the bath concentration of TEA should block the remaining 1k v. Consequently, we examined the membrane currents that are sensitive to low and high concentrations of TEA. The membrane currents that were elicited by voltage-clamp pulses to +20 mV first in ASW, following addition of 2 mM TEA, and following the increase of the concentration of TEA to 50 mM are shown in Fig. 3Al. Although the low concentration of TEA reduced the membrane current at the end of the voltageclamp pulse, the low concentration of TEA had little effect on the peak amplitude of the current response. Increasing the concentration of TEA to 50 mM markedly reduced the peak amplitude of the current response. The membrane currents that were blocked by low concentrations of TEA were isolated by subtracting current responses elicited by voltage-clamp pulses to 0, + 10, and +20 mV in the presence of 2 mM TEA from corresponding current responses elicited in ASW (Fig. 3A2). The current affected by low concentrations of TEA was slow to activate, showed little inactivation, and had a mild voltage dependence. In these Separation c!fK currents respects, it bears some resemblance to the S current and the 5-HT difference currents illustrated in Fig. 2C2. It differs While the relatively voltage-independent component of from 15-H-r observed at small depolarizations, however, in the 15-H7-has properties similar to the S current, the voltits threshold for activation and sensitivity to TEA (see age-dependent component of Ir+.r has not been described below). The sensitivity to low concentrations of TEA sugpreviously. We were interested, therefore, in determining gests that the difference currents in Fig. 3A2 may represent whether this second component represented the modulaa TEA-sensitive form of 1k Ca. The membrane currents that tion bv 5-HT of either the voltage-dependent IS+ current were blocked by high concentrations of TEA were isolated (Ik .,-),?a+-activated K current (ZKCa), or transient IS by subtracting the current responses elicited by voltagecurrent (I,), which are found in these sensory neurons clamp pulses to 0, + 10, and +20 mV in high concentra(Klein et al. 1982; Pollock et al. 1985; Baxter and Byrne tions of TEA from corresponding current responses elicited 1988: Walsh and Bvrne 1989) and other molluscan in low concentrations of TEA (Fig. 3A3). The current afneurons (e.g., Connorand Stevens 197 1; Thompson 1977; fected by high concentrations of TEA began to be activated Aldrich et al. 1979; Adams et al. 1YSOa,b; Byrne 1980: at potentials more positive than - 10 mV, and both its Gorman and Thomas 1980; Hermann et al. 198 I a,b, 1983; activation and inactivation kinetics are voltage dependent. Hockberger and Connor 1984; Strong et al. 1984, 1986; The large magnitude of the high TEA difference current, Thomas 1984; Deitmer and Eckert 1985; Junge 1985; the marked voltage dependence of its peak amplitude and Kehoe 1985: Bezanilla et al. 1986: Acosta-Urquidi 1988: the voltage dependence of its kinetics of activation and also see, Hille 1984; Latorre et al. 1984; Rogawski 1985; inactivation are similar to those reported for & in these Rudy 1988). The method that we chose to investigate this sensory neurons and other Apl-ysia neurons (Thompson possibility was to compare the pharmacologic sensitivity, 1977; Aldrich et al. 1979; Adams and Gage 1980a; Byrne voltage-dependence, and kinetics of the voltage-dependent 1980; Hermann and Gorman 198 1b; Klein et al. 1982; component of 15-H-r to those of Ik I., Jk Ca, and I*. An ex- Pollock et al. 1985; Strong and Kaczmarek 1986). tensive body of literature indicates that& I,, and IKCa can In certain neurons of Aplvsia, low concentrations of be blocked by the use of appropriate concentrations of 4-AP have been reported to block Ik I, with a & of -0.8 TEA and 4-AP and that IA can be selectively inactivated at mM, while having no effect on lk.Ca (Hermann and Gorrather hyperpolarized membrane potentials (e.g., Connor man 198 la; also see Rudy 1988). Thus we examined the and Stevens 197 1; Thompson 1977; Byrne 1980; Hermann membrane currents that are sensitive to low concentrations et al. 198 la,b, 1983; Klein et al. 1982; Stanfield 1983; Hille of 4-AP. An example of the membrane currents that were 1984; Latorre et al. 1984; Rogawski 1985; Blatz and Magelicited first by a voltage-clamp pulse to +20 mV in ASW leby 1987; Shuster and Siegelbaum 1987; Cook 1988: and then with an identical depolarizing step after applying Rudy 1988: Walsh and Byrne 1989). Therefore, we used 1.5 mM 4-AP are shown in Fig. 3BZ. The 4-AP markedly these methods to isolate examples of these three classes of reduced the peak amplitude of the current response but Kt currents and compare their characteristics with those had little effect on the current at the end of the voltageot I>-HTclamp pulse. The membrane currents that were blocked by of 4-AP were isolated by subtracting I- k.I.* In previously studied neurons of Aplwia, TEA has low concentrations I current responses elicited by voltage-clamp pulses to 0, been reported to block IK 1vwith an apparent dissociation constant (Kd) of 6-8 mM and some examples of 1k (Jawith a + 10, and +20 mV in the presence of 4-AP from corresponding current responses elicited in ASW (Fig. 3B2). rcb near 0.5 mM (Thompson 1977; Byrne 1980; Hermann These 4-AP difference currents are similar to the high TEA and Gorman 198 1b; Klein et al. 1982; Deitmer and Eckert con1985; Walsh and Byrne 1989). Thus low concentrations of difference current (Fig. 3A3) and have characteristics TEA should block 1kCa significantly but have little effect sistent with those of 1k .I. .

D. A. BAXTER

AND

J. H. BYRNE

TEA

A1

Step to +20 mV

Bl

Step to +20 mV

+ 2 mM TEA + 50 mM TEA

20 msec

20

nA

A2

/mvijOnA
B2
4-AP Difference Currents
20Zec

Low TEA Difference

Currents

(a - b)

(a - b)

A3

High TEA Difference

Currents

(b - c)
nA

FIG. 3. Pharmacologic sensitivity of I k.1.. A 1: currents were elicited by a voltage-clamp pulse from - 50 to +20 mV in ASW (trace a), following the addition of 2 mM TEA (trace b), and following the increase of the concentration of TEA to 50 mM (trace c). A2: TEA difference currents in this cell were isolated from voltage-clamp pulses to 0, + 10, and +20 mV. The component of membrane current that was blocked by the low concentration of TEA was isolated by subtracting current responses elicited in the presence of 2 mM TEA from corresponding current responses elicited in ASW (e.g., a - b in AI). A3: the component of membrane current that was blocked by higher concentrations of TEA was isolated by subtracting current responses elicited in the presence of 50 mM TEA from corresponding current responses elicited in the presence of 2 mM TEA (e.g., b - c in Al). These TEA difference currents are representative of results in 14 different preparations. BI: in a different cell, membrane currents were elicited by voltage-clamp pulses from -50 mV (a holding potential that inactivated IA, e.g., see Fig. 5A) to +20 mV in ASW (trace a) and following the addition of 1.5 mM 4-AP to the bath (trace b). B2: 4-AP difference currents were isolated from voltage-clamp pulses to 0, + 10, and +20 mV by subtracting current responses elicited in the presence of 4-AP from corresponding current responses elicited in ASW (e.g., a - b in Bl). These difference currents are representative of results obtained in 12 different preparations. The currents blocked by high concentrations of TEA (50 mM, A3) and low concentrations of 4-AP (1.5 mM, B2) are nearly identical in their voltage sensitivity and kinetics of activation and inactivation. These features are characteristic of IK,I.

pica. Walsh and Byrne (1989) have shown that in these sensory neurons low concentrations of TEA (5 mM) block a K current that is activated by intracellular injection of Ca2+ (also see, Hermann et al. 198 lb, 1983; Klein et al. 1982; Stanfield 1983; Hille 1984; Thomas 1984; Deitmer and Eckert 1985; Kehoe 1985; Sawada et al. 1987, 1989). As shown in Fig. 3, externally applied 4-AP (2 mM) blocks I K 2,,but apparently not the low TEA difference current, which resembles Ik,ca (Thompson 1977; Hermann et al. 198 la, 1983). Thus we examined the membrane currents that were insensitive to 4-AP but were sensitive to TEA. Membrane currents were elicited by a voltage-clamp pulse to +20 mV first in ASW containing 4-AP, following the addition of 2 mM TEA, and following the increase of the concentration of TEA to 50 mM (Fig. 4A). Application of a low concentration of TEA markedly reduced the membrane current. Increasing the concentration of TEA to 50 mM had little additional effect on the membrane currents.

This result indicates that an additional component of membrane current can be blocked by low concentrations of TEA even after a significant component of membrane current has been blocked already by 4-AP. This pharmacologic sensitivity is consistent with TEA-sensitive forms of Ik Ca. The membrane current that is blocked by low TEA

(bht not by 4-AP) was isolated by subtracting current responseselicited by voltage-clamp pulsesto 0, + 10, and +20 mV in the presence of low concentrations of TEA (plus 4-AP) from corresponding current responses elicited in ASW that contained 4-AP (Fig. 4B). The low TEA difference currents became significantly activated at membrane potentials more positive than - 10 mV and were slow to develop. With voltage-clamp pulses above +30 mV, the amplitude of the low TEA difference currents began to decrease(data not shown). These low TEA difference currents and those in Fig. 3A2 appear to be Zk Ca,based on the voltage dependence and kinetics of this current and its

MODULATION

OF

K+ CURRENTS

671

Step to +20 mV
a ASW+

1.5mM4-AP

b +2 mMTEA c + 50 mM TEA . J 20 nA -

20 msec

Low TEA Difference Currents

20 mV

(a - b)

potentials is a distinctive characteristic of IA (Connor and Stevens 197 1; Thompson 1977; Adams et al. 1980b; Byrne 1980; Klein et al. 1982; Hille 1984; Latorre et al. 1984; Strong 1984; Junge 1985; Pollock et al. 1985; Rogawski 1985; Acosta-Urquidi 1988; Baxter and Byrne 1988; Rudy 1988). Thus IA can be isolated by subtracting current responses elicited from a holding potential of -50 mV from corresponding current responses elicited from a holding potential of -90 mV (Fig. 5B). Three lines of evidence indicate that modulation of IA by 5-HT cannot account for the voltage-dependent component of JS-nT. First, the voltage-dependent component of 15-HTis present only at membrane potentials more positive that - 10 mV (Fig. ZCl), whereas IA is activated by voltageclamp pulses to membrane potentials more negative than - 10 mV (e.g., Fig. 5B). Second, in sharp contrast to the voltage dependence of I k,C/ (Fig. 3) and the voltage-dependent component of I 5-IIT (Fig. 2C1), the rates of activation and inactivation of IA are relatively voltage independent. Third, the amplitude of IA increases relatively linearly with voltage (Fig. 5B), whereas both Ik band the voltage-dependent Component of&T do not.

A
15 nA
20 msec
currents after preexposure to FIG. 4. Effects of TEA on membrane 4-AP. A: membrane currents were elicited by a voltage-clamp pulse from -70 to +20 mV in ASW containing 1.5 mM 4-AP (trace a) to block IKql. (see Fig. 3B), in the presence of 2 mM TEA (trace b) to block TEA-sensiin the presence of 50 mM TEA (trace c) to block any ti ve IK,Ca T and residual IK, IJ or TEA-sensitive IK,Ca. Application of a low concentration of TEA markedly reduced the membrane current. Increasing the concentration of TEA to 50 mM had little additional effect on the membrane currents. B: low TEA difference currents were isolated from voltage-clamp pulses to 0, + 10, and +20 mV by subtracting the current responses elicited in the presence of 2 mM TEA (and 1.5 mM 4-AP) from the corresponding current responses elicited in ASW containing 1.5 mM 4-AP (e.g., a - b in A). These difference currents are representative of results obtained in 13 different preparations. The currents blocked by low concentrations of TEA (2 mM) are slow to activate, do not inactivate, and are mildly voltage-dependent. These features are characteristic of ZK,ca.

Step to -20 mV

-50 mV -90 mV 2 nA 20 ms

Isolated I K,A
-10 mV

-20 mV

similarity to IK,Ca observed in these sensory neurons and other neurons of Aplysia (Gorman and Thomas 1980; Hermann et al. 198 1a,b, 1983; Klein et al. 1982; Walsh and Byrne 1989). Thus the slow development of the low TEA difference currents (Figs. 3A2 and 4B) probably reflects the time course of the accumulation of intracellular Ca2+ (Gorman and Thomas 1980). The decrease in peak amplitude with large depolarizations is due presumably to the fact that as the membrane potential approaches the Ca2+ equilibrium potential, the Ca2 current decreases and less intracellular Ca2+ is available to activate Ik,Ca (Hermann and Hartung 1983). I~. The membrane currents that were elicited by voltageclamp pulses to -20 mV from holding potentials of -50 and -90 mV are shown in Fig. 5A. The fast, transient component of the outward membrane current that was present during the voltage-clamp pulse from -90 mV was inactivated by the -50 mV holding potential. This complete inactivation at relatively hyperpolarized membrane

-30 mv

2 nA
20 ms
Isolation of IA. A: examples of membrane currents elicited by voltage-clamp pulses to -20 mV from holding potentials of -50 and -90 mV. B: component of outward current that was highly sensitive to the holding potential was isolated by subtracting current responses elicited by voltage-clamp pulses to -30, -20, and - 10 mV from a holding potential -50 mV from the corresponding current responses elicited by depolarizations to the same levels but from a holding potential of -90 mV. This current is relatively voltage independent with respect to both its kinetics of activation and its kinetics of inactivation. The kinetics of activation and inactivation are rapid compared to both lk,,. (e.g., Fig. 3, A.? and B2) and Ca of -I,, . These results indiI K 1 (Fig. 4B). These features are characteristic cate that the properties of IA are different from those of the voltage-dependent component of I 5-HTT and thus, it is unlikely that modulation of 1A by 5-HT could count for the voltage-dependent component of IwHT.
FIG. 5.

672

D. A. BAXTER

AND

J. H. BYRNE

Phurmacologic

sensitivity of5-HT .

difference currents .

While other interpretations are possible, the similarity between the voltage dependence and the complex kinetics of IK,L. and those of the voltage-dependent component of IsmHT suggested to us that this component of IsmHT may be due to a modulation of 1k rjby 5-HT. It appeared that 5-HT slowed the activation and inactivation of Ik v (e.g., Fig. lsl>. Such a modulation would explain both the initial decrease and the later increase in the outward membrane current induced by of 5-HT. Alternatively, the late increase in membrane current may represent an indirect modulation Of 1~ Ca by 5-HT. Boyle et al. (1984) and Spira et al. (1987) found that application of 5-HT increased the concentration of intracellular free Ca2+ as measured by the Ca2+ indicators arsenazo III and fura-2. These results raise the possibility that 5-HT could lead to increased activation by increasing the levels of intracellular Ca2+ (see Of IK.Ca also Sacktor et al. 1986; Braha et al. 1988). Thus the late increase in outward current induced by 5-HT (Fig. IBI) could be due to an increase in 1k,ca. To distinguish between

these two possibilities, we examined the sensitivity of &-ijT to agents that differentially affect Ik v and 1k,ca. As mentioned above (Figs. 3A and 4), Ik I/ and some forms Of 1~ Ca can be distinguished by their dikerent sensitivities to TEA; I k,Cacan be blocked by low concentrations of TEA that have little effect on IK,c. As shown in Fig. 6A, we examined the sensitivity of the voltage-dependent component of 15 to low concentrations of TEA by isolating -u-r 15-H-r solutions containing 2 mM of TEA. Membrane in currents were elicited by voltage-clamp pulsesto +20 mV; first in ASW, following the addition of 2 mM TEA and following addition of 5-HT to the solution already containing TEA (Fig. 6Al). Despite the presence of a low concentration of TEA, I 5-H-rstill displayed the characteristic voltage-dependent component (Fig. 6A2). Thus a concentration TEA that blocks an Ik c,-like current (but not Ik v) does not block the voltage-dependent component of &-iT. (A more complete analysis of the sensitivity of &-HT to TEA is presented in Figs. 7 and 9.) IK,v and 1k,Caalso can be distinguished on the basis of their sensitivity to 4-AP; Ik 3 is blocked by low concentrav

Al

Low TEA
a ASW

Bl

4-AP

c b

(ASW

+ TEA) + 5-HT

+ 2 mM TEA

a ASW b + 1.5mM4-AP c (ASW + 4-AP) + 5-HT

25 nA 20 msec

I 20
C) B2
5-HT Difference Current

nA

A2

5-HT Difference

Current

(b -

(b - c)

20 msec

-J

5 nA

FIG. 6. Effect of TEA and 4-AP on the ability of 5-HT to modulate membrane currents. Membrane currents were elicited by voltage-clamp pulses from -70 to +20 mV. Al: current responses were elicited first in ASW (trace a), following addition of 2 mM TEA (trace b) to block IK.CYa, and following the addition of 5-HT (35 PM) to the bath already containing TEA (trace c). A2: 5-HT difference current was isolated by subtracting the current response elicited in the presence of 5-HT (and low TEA) from the current response elicited in the presence of low concentrations of TEA (b - c). This difference current is representative of results obtained in 16 different preparations. BI: in a different cell, current responses were elicited first in ASW (trace a), following addition of 1.5 mM 4-AP (trace b) to block IK,I., and following the addition of 5-HT (35 pM) to the bath containing 4-AP (trace c). B2: 5-HT difference current was isolated by subtracting the current response elicited in the presence of 5-HT (and 4-AP) from the current response elicited in the presence of 4-AP (b - c). This difference current is representative of results obtained in 6 different preparations. These results indicate that agents (low TEA) that block ZK.ca do not block the voltage-dependent component of I5-HT, whereas, agents (4-AP) that block &I block the Moreover, when the voltage-dependent component of 1s-HT is blocked by 4-AP, the voltage-dependent component of&r. relatively voltage-independent component of I 5_n-r- is revealed. Thus the pharmacologic sensitivity of the voltage-dependent component of I 5-Hr appears to be identical to the pharmacologic sensitivity of IKqI/.

MODULATION

OF

K+ CURRENTS

673

tions of 4-AP, whereas IKCa is not (see Figs. 3B and 4). Therefore, we examined the sensitivity of the voltage-dependent component of 15-HTto 4-AP by isolating 15-Hr in solutions that contained 4-AP (Fig. 6B). Membrane currents were elicited by voltage-clamp pulsesto +20 mV; first in ASW, following addition of 1.5 mM 4-AP and following addition of 5-HT to the solution already containing 4-AP (Fig. 6Bl). As shown in Fig. 6B2, 4-AP blocked the voltage-dependent component of 15-HT(also see Fig. 8). The 5-HT difference current that was isolated in the presence of 4-AP revealed only the residual voltage-independent component that is identical to the S current. [Previous work has shown that the S current is relatively insensitive to 4-AP and has a & for 4-AP of 10 mM (Shuster and Siegelbaum 1987; Brezina 1988).] Thus an agent (4-AP) that blocks Ik L (but not JK,Ca) also blocks the voltage-dependent component of 15-HT. To more completely compare the sensitivities to TEA of Ik Iwand the voltage-dependent component of &iT, we from sensory neurons that were bathed in examined IsmHT ASW (Fig. 7A) and from neurons bathed in ASW that contained various concentrations of TEA (Fig. 7, B-E). In each case, the voltage-clamp pulse was from -70 to +20 mV. Thus, in the experiments of Fig. 7, different preparations were first exposed to a specific concentration of TEA and then the ability of 5-HT to modulate the membrane currents was examined. At concentrations of 2 and 5 mM TEA, the voltage-dependent component of 15-HT was still present (Fig. 7, B and C). These concentrations of TEA can be effective in blocking Ik Ca:,but have lessof an effect on IKsl. (Hermann and Go&an 198 1b; Klein et al. 1982; Walsh and Byrne 1989) (also seeFigs. 3, 4, and 9). Therefore, it is unlikely that the complex kinetics of &-nT can be accounted for by an increase in the amplitude of Zk ca. The voltage-dependent component of &i-r was blocked only at the high concentrations of TEA that are also effective in blocking IK.I. (Fig. 7, D and E; note the change in the value of the calibration bar). 5-HT difference currents that were isolated in ASW containing 100 mM TEA reveal only the residual voltage-independent component that is identical to the S current obtained at less depolarized membrane potentials (compare Figs. 7E and 2C2). Figure 8 compares the current-voltage (Z-I/) relationships of 5-HT difference currents that were isolated from senso neurons under t hree conditions: 1) from sensory ry neurons bathed first in ASW and then in 5-HT, 2) from sensory neurons first bathed in ASW that contained 100 mM TEA (to reduce Zk,caand Ik i) and then in 5-HT, and 3) from sensory neurons bathed first in 2 mM 4-AP (to block Jk.J,)and then 5-HT. In all three cases,the membrane currents were elicited by voltage-clamp pulses from -70 mV to membrane potentials ranging between -40 and +30 mV. The amplitudes of 5-HT difference currents were measured at the end of the voltage-clamp pulses and were plotted such that a positive value represented a decrease and a negative value represented an increase in the outward membrane currents. In cells exposed to ASW and then to 5-HT, the relatively voltage-independent component of 15-H-r was evident during voltage-clamp pulses to -40 mV through ~0 mV (also seeFig. 2C2). The downward shift in the I-V relationship of &-HT at 0 mV reflects the development of the voltage-dependent component of &-HT (also see

ASW

I
B
2 mM TEA

15nA

I
5 mM TEA 50 mM TEA

._

15nA

15nA

5 nA

100 mM TEA

I 5 nA 20msec
FIG. 7. Effects of preexposure to different concentrations of TEA on the ability of 5-HT to modulate membrane currents. Each trace is a 5-HT difference current from a different experiment and in each experiment the cell was maintained in either ASW (A) or ASW containing the indicated concentration of TEA (B-E). All voltage-clamp pulses were from -70 to +20 mV. 5-HT difference currents were isolated by subtracting the current responses elicited following bath application of 5-HT (30 PM) from the current responses elicited prior to the application of 5-HT (see Figs. 1 and 64. Between 5 and 15 preparations were examined for each of the concentrations of TEA indicated, and similar results were obtained in all preparations (also see Fig. 9). Note changes in ordinate scales. These results indicate that increased concentrations of TEA, which increase the block of &., block the voltage-dependent component of 15-H*. In high concentrations of TEA the relatively voltage-independent component of JS-rjT. is revealed.

Fig. 2Cl). Of particular significance is the observation that the concentrations of 4-AP and TEA that block 1k v also block this downward shift in the I- V relationship for &-nT. Since 4-AP does not block 1k,ca, the absenceof the down-

674

D. A. BAXTER

AND

J. H. BYRNE

. A
n

ASW 4-AP TEA (2 mM) (IOOmM)

, I . , , ,

I
,k

r #

Vm (mV)

-3

-L

-1

FIG. 8. Current-voltage relationships for the relatively voltage-independent and voltage-dependent components of 15-ijl.. Current-voltage (I-I,) relationships are shown for 5-HT difference currents that were isolated under 3 conditions: I) from neurons bathed in ASW, 2) from neurons bathed in ASW containing 2 mM 4-AP, and 3) from neurons bathed in ASW containing 100 mM TEA. Amplitudes of the 5-HT difference currents were measured at the end of the voltage-clamp pulses. Data are plotted such that a positive value represents a decrease in the outward membrane current and a negative value represents an increase in the outward membrane current. Each point represents the average of between 4 and 15 experiments. For each average, the coefficient of variation was no greater than 7% of the mean value. In ASW, I 5_ is present at potentials as HT negative as -40 mV and increases with depolarization in a relatively voltage-independent manner up to a potential of ~0 mV. At potentials more positive than 0 mV, the highly voltage-dependent component of &-HT is observed. Pretreatment with 2 mM 4-AP does not affect &-nT in the range of -40 to 0 mV but completely blocks the highly voltage-dependent component of I c_HT normally observed at potentials more depolarized than 0 mV. TEA ( 100 mM) also blocks the highly voltage-dependent component of I 5_i1 I and has a slight effect on the relatively voltage-independent component of &-HT. These results indicate that pharmacologic agents that block I K;.I. also block the voltage-dependent component of 1-H I *

Log D-m (M)

FIG. 9. TEA dose-response relations for IK,<-a, lKsI., and the 2 components of &-HT.. Data are plotted as the percentage of the average control amplitude for each type of Ki current that remains in the indicated concentration of TEA. For 1s-,-17..+ and Zs-H.r, 20, the control amplitudes ?() were the amplitudes measured at the end of the voltage-clamp pulses for direrence currents isolated in ASW (e.g., Fig. 1). Effects of TEA were obtained by isolating ZS-nT.+20 and I 5_u1,-20 from cells preexposed to different concentrations of TEA (e.g., Fig. 7). The average amplitudes of Js-HT,+20 were negative values in all concentrations of TEA below 100 mM (see Fig. 7). (Positive values for the average amplitude of 15-IIT,+ 20 represented the complete block of the voltage-dependent component of&,-, and therefore, were plotted as zero percentage of the control.) For ZK,I., the control amplitude was the average of the peak amplitude of the 4-AP difference current during a voltage-clamp pulse to +20 mV in ASW (e.g., Fig. 3). To generate the dose-response relationship of Ik, r, for TEA, different concentrations of TEA were applied to each cell and the amount of current that could be subsequently blocked by 4-AP was determined. The average of these values for each concentration was then normalized to the average control 4-AP difference current obtained in the absence of TEA. For the I k-c,-like current, the data were normalized to the average of the peak amplitude of the difference current obtained after adding 50 mM TEA to cells that were preexposed to 4-AP and clamped to a potential of +20 mV (e.g., Fig. 4). Thus the Zk ca-like current was presumed to be totally blocked by 50 mM TEA (see below and text). The dose-response relationship for the IK,(.a- like current for different concentrations of TEA was obtained by pretreatment of cells with 4-AP and determining how much additional current could be blocked by various concentrations of TEA. Note that no additional current is blocked as the concentration of TEA is increased beyond 5 mM. A nonlinear parameter estimation program (based on the Gauss-Newton algorithm) was employed to estimate k;, for each K current. The 4 smooth curves for IK.(.ar Is-i,T,t 70, Ik., , and 1s-i,pr. ?() were generated using a form of the Michaelis-Menton equation (see text) and estimated values for Kd of 0.4, 5, 8, and 92 mM, respectively. Each data point on the graph represents the average of between 4 and 15 experiments. For each average, the coefficient of variation was no greater than 9% of the mean value. These results indicate that the TEA dose-response relationship for 1k,j and the voltage-dependent component of &-uT are identical. In contrast, the dose-response curve for &-ii7.. 20 is not similar to that for either the I k,c,-like current or ZK.r..

ward shift in the I- I/relationship for &,- in the presence of 4-AP indicates again that modulation of Ik ca cannot account for the complex kinetics and voltage dependence of 15-HT. It is also interesting to note that this concentration of 4-AP had absolutely no effect on the component of &iT elicited by voltage-clamp pulses to membrane potentials more negative than 0 mV. In contrast, high concentrations of TEA reduced the amplitude of 15-HT elicited at these same membrane potentials. Therefore, it appears that the relatively voltage-independent component of 15-HT is slightly sensitive to TEA and insensitive to 4-AP. This pharmacologic sensitivity of the relatively voltage-independent component of I 5-H-r is identical to that of the S current (Klein et al. 1982; Brezina et al. 1987, 1988; Shuster and Siegelbaum 1987; Walsh and Byrne 1989). Figure 9 illustrates the dose-response relationships for the effects of TEA on an Zk ca-like current, IK v, and the two components of 15-ljT. The relatively voltage-indepen5 dent component of 1_ H-i- was isolated from voltage-clamp pulses to -20 mV (I 5-HT,-20), which did not activate the voltage-dependent component (e.g., Fig. 1). The voltage-

dependent component of &T was isolated from voltageclamp steps to +20 mV (1 5-HT +&. External TEA produced a dose-dependent reduction in the amplitudes of all four classes of currents (Fig. 9). Previous data from sensory (Shuster and Siegelbaum 1987) and other neurons of&@?-

sia (Thompson
Deitmer

1977; Hermann

and Gorman

198 lb;

and Eckert

1985; Brezina et al. 1987, 1988; also

MODULATION

OF

K- CURRENTS

675

see Stanfield 1983; Hille 1984; Rudy 1988) indicate that each K+ channel is blocked by a single molecule of TEA. Thus the block of KS currents by TEA can be fit by a one-to-one binding reaction, which assumes that the percentage of K+ current remaining is equal to the number of KS channels not occupied by a molecule of TEA: % control = 100 - 100X ([TEA]/& + [TEA])), where I& is the apparent dissociation constant. The Ik c,-like current was the most sensitive to TEA with an estimated I& of 0.4 mM. Ik LF and &-r +70 had very similar sensitivities to TEA with estimated Kd of 8 and 5 mM, respectively. Therefore, it appears that I 5_HT,+zo represents the serotonergic modulation of a membrane current that has properties closely resembling Ik, I ,. Finally, 15-H-r -20 had an estimated I& of ~92 mM. These values for Kd are close to those previously reported by others (Thompson 1977; Hermann and Gorman 198 1b; Brezina et al. 1987, 1988; Shuster and Siegelbaum 1987).

A Step to +20 mV
c (ASW + CAMP) + 5HT

+ CAMP

I30

nA

20 msec

Modhtion

of membrane .

currents by CAMP

B Difference Currents

Previous work on sensory neurons has shown that serotonergic modulation of the S current is mediated via cyclic AMP (CAMP) (Siegelbaum et al. 1982, 1987; Camardo et al. 1983; Walsh and Byrne 1984, 1989; Ocorr and Byrne 1985; Pollock et al. 1985; Shuster et al. 1985; Belardetti and Siegelbaum 1988). We were interested, therefore, in determining whether CAMP also mediates the serotonergic If modulation of Ik TY. CAMP does mediate both actions of 5-HT, then the application of a membrane-permeable analogue of CAMP should produce the same modulation of membrane currents as 5-HT (e.g., Fig. l), and preexposure to a saturating concentration of CAMP should occlude any further modulation of membrane currents by 5-HT. Thus we compared the modulatory actions of 8-pcpt-CAMP and 5-HT on membrane currents (Fig. 10). Membrane currents were elicited by voltage-clamp pulses to +20 mV; first in ASW, following addition of 8-pcpt-CAMP (50 PM) and following the addition of 5-HT (30 PM) to the bath already containing 8-pcpt-CAMP (Fig. IOA). Application of 8pcpt-CAMP reduced the total membrane current at the end of the voltage-clamp pulse but had little effect on total membrane current early during the voltage-clamp pulse. The membrane current that was modulated by 8-pcptCAMP was isolated by subtracting the current response elicited in the presence of 8-pcpt-CAMP from the current response elicited in ASW (Fig. lo@. The resulting CAMP difference current (IcAMP) was similar to the relatively voltage-independent component of &-HT that can be isolated in ASW at hyperpolarized membrane potentials (e.g., Fig. 2U), and at depolarized membrane potentials when Ik L/ has been blocked by 4-AP or high concentrations of TEA (e.g., Figs. 682 and 7E). Increasing the bath concentration of 8-pcpt-CAMP by an order of magnitude did not produce any further modulation of membrane currents, and the preexposure to 50 PM of 8-pcpt-CAMP occluded additional modulation of the relatively voltage-independent component of 15-H-r by subsequent application of 5-HT (Baxter and Byrne 1987). Thus 50 PM 8-pcpt-CAMP appeared to produce a maximal effect on membrane currents and appeared to modulate only the relatively voltage-independent component of 15-HT.

FIG. 10. Comparison of the modulatory effects of CAMP and 5-HT on membrane currents. A: membrane currents were elicited by voltage-clamp pulses from -70 to +20 mV in ASW (trace a), following the addition of 8-pcpt-CAMP (50 PM) (trace b), and following the addition of 5-HT (30 PM) to the bath already containing 8-pcpt-CAMP (trace c). B: CAMP difference current (&*& was isolated by subtracting the current response elicited in the presence of 8-pcpt-CAMP from the current response elicited in ASW (a - b). The 5-HT difference current (Js+,.J was isolated by subtracting the current response elicited the presence of 5-HT (and 8pcpt-CAMP) from the current response elicited in ASW containing 8pcpt-CAMP (b - c). This difference current is representative of results obtained in 9 different preparations. These results indicate that CAMP mediates the serotonergic modulation of the S current, but does not mediate the modulation of a voltage-dependent component.

Although 8-pcpt-CAMP mimicked the actions of 5-HT on the S current, 8-pcpt-CAMP failed to modulate Ik .V. The addition of 5-HT to the bath, which still contains the 8-pcpt-CAMP, produced additional modulation of membrane currents (Fig. IOA, trace c). Application of 5-HT reduced the membrane current early during the voltageclamp pulse and increased the membrane current at the end of the voltage-clamp pulse. The modulatory effect of 5-HT was isolated by subtracting the current responseelicited in the presence of 5-HT (plus 8-pcpt-CAMP) from the

676

D. A. BAXTER

AND

J. H. BYRNE

current elicited in ASW and 8-pcpt-CAMP (Fig. 1OB). This I 5-nT waveform displayed the characteristic voltage-dependent component. These results indicate that the modulation of IK L by 5-HT is not mediated by elevated levels of CAMP. Furthermore, these results further support the finding that there are at least two distinct components to 15-H-r. Modulation ofmembrane currents by SCPh To provide further support for the suggestion that the modulation of I K,I, by 5-HT is not mediated via CAMP, we

20 r
A I. , ,, ,

A
.

1CAMP 1 SCPb

/ /I //

Step to +20 mV
) + 5-HT

I 5-HT

a b

ASW +scPt,

FIG. 12. The current-voltage (Z-V) relationships are shown for CAMP and SCPh difference currents that were isolated from neurons bathed in ASW. [The data for 5-HT difference currents isolated in ASW (Fig. 8) are replotted in this figure for the purpose of comparison.] The amplitudes of the CAMP and SCPh difference currents were measured at the end of the voltage-clamp pulses, and the data are plotted such that positive values represent a decrease in the outward membrane current (e.g., see Figs. 10 and 11). Each point represents the average of between 4 and 15 experiments. For each average, the coefficient of variation was no greater than 13% of the mean value.

\ J

20 nA 20 msec

Difference Currents n
1%HT (b - c)

(a- b)

-I

I 5 nA
20 msec

FIG. 11. Comparison of the modulatory effects of SCPh and 5-HT on membrane currents. A: membrane currents were elicited by voltage-clamp pulses from -70 to +20 mV in ASW (trace a), following the addition of SCPb (20 PM) (trace b), and following the addition of 5-HT (20 PM) to the bath already containing SCP,., (trace c). B: SCPb difference current (Iscpr,) was isolated by subtracting the current response elicited in the presence of SCP,, from the current response elicited in ASW (a - b). This difference current is representative of results obtained in 4 different preparations. The 5-HT difference current was isolated by subtracting the current response elicited in the presence of 5-HT (and SCPhj from the current response elicited in ASW containing SCPb (b - c). These results indicate that SCPb, mimics the actions of 5-HT on the S current, but does not modulate an additional voltage-dependent component as does 5-HT.

compared the modulatory effects of small cardioactive peptide b (SCPb) and 5-HT. In sensory neurons, the actions of SCPb parallel the actions of 5-HT by elevating intracellular levels of CAMP, which in turn results in the closure of the S channel (Abrams et al. 1984; Ocorr and Byrne 1985, 1986). Thus we compared the modulatory actions of SCPb and 5-HT on membrane currents. Membrane currents were elicited by voltage-clamp pulses to +20 mV first in ASW, following addition of SCPb (20 PM) and following the addition of 5-HT (20 PM) to the bath already containing SCPb (Fig. 1 lA, trace c). The action of SCPb paralleled that of CAMP and reduced the total membrane current at the end of the voltage-clamp pulse while having little effect on total membrane current early during the voltage-clamp pulse. The application of 5-HT to the bath, which still contained the SCPb, produced additional modulation of membrane currents (Fig. 1 IA). 5-HT reduced the membrane current early during the voltage-clamp pulse and increased the membrane current at the end of the pulse. The membrane current that was modulated by SCPb was isolated by subtracting the current response elicited in the presence of SCPb from the current response elicited in ASW (Fig. 1lB). The resulting SCPb difference current (lsCPb) was similar to JcAMPand the relatively voltage-independent component of &-HT (Figs. 2C2 and 1OB). The modulatory effect of 5-HT was isolated by subtracting the current response elicited in the presence of 5-HT (plus SCPb) from the current elicited in ASW and SCPb (Fig. 1OB). This &-H-r waveform displayed the characteristic voltage-dependent component. The differential effects of 5-HT, CAMP, and SCPb on membrane currents are illustrated by the Z-V relationships of LAMP $ km 9and 15-n-r(Fig. 12). As described above (Fig. 8) in ASW, the downward shift in the I- I relationship of &-nT at membrane potentials above 0 mV indicates the development of the voltage-dependent component of &-HT.

MODULATION

OF

K+ CURRENTS

677

The observation that the 1-V relationships for IcAMP and lscPb for membrane potentials above 0 mV do not show the downward shift (Fig. 12) further supports the finding that there are two distinct components to &-HT, only one of which is modulated via elevated levels of CAMP.

mimic the action of 5-HT on the voltage-dependent component of I 5-n-r (Baxter and Byrne 1987). Furthermore, SCPb, which elevates levels of CAMP, modulates only the relatively voltage-independent component of 15-n-r. These results indicate not only that the modulation of JK v by 5-HT may require another second messenger system: but also provide additional evidence that there are two separate DISCUSSION currents modulated by 5-HT. Although the mechanism by We conclude that in somata of sensory neurons isolated which 5-HT modulates IK,C is not known, a tentative hyfrom pleural ganglia 5-HT modulates at least two KS cur- pothesis can be advanced. We have observed that the relarents that are elicited by brief voltage-clamp pulses. One tionship between the time-to-peak for Ik v and membrane component has properties consistent with the previously voltage is shifted to more positive membrane potentials by described S current (Klein et al. 1980, 1982; Siegelbaum et 5-HT (data not shown). Recent experiments in squid axon al. 1982, 1987; Camardo et al. 1983; Walsh and Byrne have indicated that protein phosphorylation can result in a 1984, 1989; Ocorr and Byrne 1985; Pollock et al. 1985, shift in the voltage dependence of both the activation and 1987; Scholz and Byrne 1987, 1988). It is relatively voltage inactivation of I k,I/ (Bezanilla et al. 1986; see also Lagrutta et al. 1989). A similar phosphorylation could account for independent, activates relatively slowly, does not inactivate, is not blocked by 2 mM 4-AP, is relatively insensitive the modulation of 1k h in the sensory neurons. While it is not yet clear whether both components of to TEA, and under proper pharmacologic conditions can be isolated and observed at membrane potentials ranging 15-nT are modulated during endogenous nervous activity, it between -40 and +30 mV. The other component has is interesting to note that the somata of these sensory neurons are enveloped by serotonergic varicosities (Lo et properties similar to Ik I. It is voltage dependent with complex kinetics, and it is blocked by 4-AP and TEA. al. 1987; Zhang et al. 1988). Physiological stimuli that Therefore, we propose that in addition to the S current, cause release of 5-HT from these varicosities would have 5-HT also modulates Ik cw, possibly by slowing its activa- profound effects on the electrophysiological activity of the tion and inactivation kinetics. sensory neurons. Since the relatively voltage-independent There are several possible reasons why previous investicomponent (S current) does not inactivate and is a prominent current at the resting membrane potential, it reduces gations have not observed the voltage-dependent component of &.I in the sensory neurons of Aplysia. First, this the excitability of sensory neurons by shunting depolarizcomponent, like 1k, hs,is susceptible to cumulative inactiing currents. Inhibiting the S channel by bath application vation during repetitive depolarizations. For example, dur- of CAMP analogues doubles the number of action potening voltage-clamp pulses to +20 mV that were separated by tials stimulated during 1 s depolarizing current pulses, and 90 s and preceded by hyperpolarizations, we observed out- modestly broadens the action potential (Klein et al. 1986; ward membrane currents that reached an average peak Baxter and Byrne 1987). Although & is not activated at amplitude of 148 nA within 20 ms (e.g., Fig. 2A). Eliminatthe resting membrane potential, it increases dramatically ing the hyperpolarizing prepulse and shortening the interwith large depolarizations, becoming the predominant outpulse interval to 10 s reduced the outward current by an ward current. Thus activation of IKsl. contributes signifiorder of magnitude and significantly altered the actions of cantly to the repolarization phase of the action potential. of I k,v by 5-HT causes a significant de5-HT on these attenuated membrane currents (data not The modulation shown). In addition, the cumulative inactivation is tempercrease in this outward current and a threefold increase in ature sensitive, and inactivation is more pronounced at spike duration (Baxter and Byrne 1987 and manuscript in room temperature (22 to 2sC) than at the normal physi- preparation). Thus closure of S channels may account priological temperature of 1SC used in this study (data not marily for the enhanced excitability (Klein et al. 1986; shown). Second, 5-HT often does not significantly reduce Baxter and Byrne 1987), and modulation of Ik crmay acthe maximum conductance of this voltage-dependent com- count primarily for the broadening of the action potential ponent (e.g., Figs. 1, 2, 6, and 11). Therefore, examining that is observed in the pleural sensory neurons during application of 5-HT. the effects of 5-HT on the peak amplitudes of membrane currents would not detect the changes in the kinetics that The pharmacologic methods and the parameters for the were revealed by difference currents. Finally, the possibility voltage-clamp protocol that were used in this study do not that abdominal and pleural sensory neurons respond dif- rule out the possibility that other currents are modulated ferently to 5-HT cannot yet be ruled out. by 5-HT. Indeed, Walsh and Byrne (1989) found that 5-HT Previous patch-clamp studies have shown that in ab- reduces a Ca2+-sensitive K+ current that is blocked by low dominal sensory neurons 5-HT closes the S channel by a concentrations of TEA but that does not appear to be actiCAMP-dependent phosphorylation of a protein closely as- vated by brief voltage-clamp pulses. Although a possibility species of lkCa in these cells sociated with this K channel (Siegelbaum et al. 1982; exists for a TEA-insensitive Shuster et al. 1985). Of the two currents modulated by (e.g., Hermann and Hartung 1983; Deitmer and Eckert 5-HT that are described in this paper, only the relatively 1985; Kehoe 1985), it is unlikely that the modulation by voltage-independent component appears to be sensitive to 5-HT of a TEA-insensitive 1k Cacould account for the voltelevated levels of intracellular CAMP (see also Walsh and age-dependent component of 15-uT. To account for the Byrne 1984, 1989; Pollock et al. 1985). Bath applications of block of the downward shift in the I- V relationship of IS-n1 CAMP analogues mimic the action of 5-HT on the rela- by 4-AP (Fig. 6), such a current would have to be also tivelv voltage-independent component of IqmFIT.but do not sensitive to 4-AP. At present. none of the Ca2+-activated

678

D. A. BAXTER

AND

J. H. BYRNE

K currents so far described have been shown to be blocked by aminopyridines (Thompson 1977; Adams et al. 19SOb; Hermann et al. 198 la, 1983; Stanfield 1983; Hille 1984; Latorre et al. 1984; Thomas 1984; Deitmer and Eckert 1985; Kehoe 1985; Mallart 1985; Blatz and Magleby 1987; Ritchie 1987; Smart 1987; Cook 1988; Rudy 1988). Furthermore, it is unlikely that the kinetics of IKca could account for both the initial upward peak and the late downward shift in the voltage-dependent component of 15-HT (e.g., Fig. lB2). There also remains the possibility that 5-HT may modulate inward currents that are not affected by 4-AP or TEA (e.g., Hockberger and Connor 1984; Pellmar 1984; Paupardin-Tritsch et al. 1986; Edmonds et al. 1987; Tsien 1987; Braha et al. 1988). The results presented in this paper and by others indicate that the modulatory effects of 5-HT on the cellular properties of sensory neurons in Aplvsia involve several components. At least three different I? currents are altered: the S current (Klein et al. 1980, 1982; Pollock et al. 1985), a slowly activating Ca2+ -sensitive K current (Walsh and Byrne 1989) and IK.IIV.There is also an increase in the levels of free intracellular Ca* following application of 5-HT that is not merely secondary to the changes in Kf currents (Boyle et al. 1984). In addition, 5-HT has other actions; including, translocation of the phospholipid-dependent C kinase (Hochner et al. 1986a; Sacktor et al. 1986), stimulation of protein synthesis that is required for the development of long-term presynaptic facilitation and increased excitability (Montarolo et al. 1986; Dale et al. 1987) and mobilization of releasableneurotransmitter (Gingrich et al. 1985, 1987, 1988; Hochner et al. 1986a). Thus there are multiple sites for plasticity within a single neuron, and while complex, all of the effects of 5-HT may act synergistically to facilitate synaptic transmission at the sensory to motor neuron synapse.
We thank Drs. L. Cleary, S. Critz, E. Kandel, M. Klein, and K. Scholz for their helpful comments on an earlier draft of this manuscript. This research was sponsored by the Air Force Office of Scientific Research, Air Force Command, USAF, under Grants AFOSR 87-0274 and National Institute of Mental Health Award K02 MH-00649. Address for reprint requests: D. A. Baxter, Dept. of Neurobiology and Anatomy, The University of Texas Medical School, P.O. Box 20708, Houston. TX 77225. Received 24 August 1987: accepted in final form 10 April 1989.

D. A. AND BYRNE, J. H. Reduction of voltage-activated IS+ by forskolin is not mediated by CAMP in pleural sensory of Aplysia. Sot. Neurosci. Abstr. 14: 153, 1988. BELARDETTI, F. AND SIEGELBAUM, S. Up- and down-modulation of single KS channel function by distinct second messengers. Trends Neurosci. 11: 232-238, 1988. BEZANILLA, F., CAPUTO, C., DIPOLO, R., AND ROJAS, H. Potassium conductance of squid giant axon is modulated by ATP. Proc. Natl. Acad. Sci. USA 83: 2743-2745, 1986. BLATZ, A. L. AND MAGLEBY, K. L. Calcium-activated potassium channels. Trends Neurosci. 10: 463-467, 1987. BOYLE, M. B., KLEIN, M., SMITH, S. J., AND KANDEL, E. R. Serotonin in voltage-clamped sensory increases intracellular Ca transients neurons of Aplysia californica. Proc. Natl. Acad. Sci. USA 8 1: 7642-7646, 1984. BRAHA, L., KLEIN, M., AND KANDEL, E. R. Phorbol ester increases Ca* current in Aplysia sensory neurons. Sot. Neurosci. Abstr. 14: 644, 1988. BREZINA, V. Guanosine S-triphosphate analogue activates potassium current modulated by neurotransmitters in Aplysia neurones. J. Physiol. Lond. 407: 15-40, 1988. BREZINA, V., ECKERT, R., AND ERXLEBEN, C. Modulation of potassium conductances by an endogenous neuropeptide in neurones of Aplysia
BAXTER,

currents neurons

caltfornica. J. Physiol. Lond. 382: 267-290, 1987.

J. H. Analysis of ionic conductance mechanisms in motor cells mediating inking behavior in Aplysia cahfornica. J. Neurophysiol. 43: 630-650, 1980. BYRNE, J. H. Neural and molecular mechanisms underlying information storage in Aplysia: implication for learning and memory. Trends Neurosci. 8: 478-482, 1985. BYRNE, J. H. Cellular analysis of associative learning. Physiol. Rev. 67: 329-439, 1987. CAMARDO, J. S., SHUSTER, M. J., SIEGELBAUM, S. A., AND KANDEL, E. R. Modulation of a specific potassium channel in sensory neurons of Aplysia by serotonin and CAMP-dependent protein phosphorylation. Cold Spring Harbor Symp. Quant. Biol. 48: 2 13-220, 1983. CAREW, T. J. Cellular and molecular advances in the study of Aplysia. In:
BYRNE,

The Neural and Molecular Bases qf Learning: Dahlem Konferenzen,

edited by J.-P. Changeux and M. Konishi. New York: Wiley, 1987, p. 177-203. CAREW, T. J. AND SHALEY, C. L. Invertebrate learning: from behavior to molecules. Annu. Rev. Neurosci. 9: 435-487, 1986. CONNOR, J. A. AND STEVENS, C. F. Voltage clamp studies of a transient outward current in gastropod neural somata. J. Physiol. Lond. 213: 21-30, 1971. COOK, N. S. The pharmacology of potassium channels and their therapeutic potential. Trends Pharmacol. Sci. 9: 2 l-28, 1988. DALE, N., KANDEL, E. R., AND SCHACHER, S. Serotonin produces longterm changes in the excitability of Aplysia sensory neurons in culture that depend on new protein synthesis. J. Neurosci. 7: 2232-2238, 1987. DEITMER, J. W. AND ECKERT, R. Two components of Ca-dependent potassium current in identified neurones of Aplysia californica. Pfluegers Arch. 403: 353-359, 1985. EDMONDS, B. W., KLEIN, M., AND KANDEL, E. R. A component of calcium channel current in Aplysia sensory neurons is blocked by dihydropyridine. Sot. Neurosci. Abstr. 13: 792, 1987. FINKEL, A. S. AND GAGE, P. W. Conventional voltage clamping with two intracellular microelectrodes. In: Voltage and Patch Clamping With Microelectrodes, edited by T. G. Smith, Jr., H. Lecar, S. J. Redman, and P. W. Gage. Washington, DC: Am. Physiol. Sot., 1985, p. 47-94. GINGRICH, K. J., BAXTER, D. A., AND BYRNE, J. H. Mathematic model of cellular mechanisms contributing to presynaptic facilitation. Brain Res. B2411. 1: 5 13-520, 1988. 2 GINGRICH, K. J. AND BYRNE, J. H. Simulation of synaptic depression, posttetanic potentiation, and presynaptic facilitation of synaptic potentials from sensory neurons mediating gill-withdrawal reflex in Aplysia. J. Neurophwysiol. 53: 652-669, 1985. GINGRICH, K. J. AND BYRNE, J. H. Single-cell neuronal model for associative learning. J. Neurophysiol. 57: 1705- 17 15, 1987. GORMAN, A. L. F. AND THOMAS, M. V. Potassium conductance and internal calcium accumulation in a molluscan neurone. J. Physiol. Lond. 308: 287-3 13, 1980. HAWKINS, R. D., CLARK, G. A., AND KANDEL, E. R. Cell biological studies of learning in simple vertebrate and invertebrate systems. In: Handbook of Physiology. The Nervous System. Higher Functions of the Brain. Bethesda, MD: Am. Physiol. Sot., 1986, sect. 1, vol. V, chapt. 2, p. 25-83. HERMANN, A. AND GORMAN, A. L. F. Effects of 4-aminopyridine on

REFERENCES V. F., CAMARDO, J. S., KANDEL, E. R., endogenous neuropeptides modulate the gill and siphon withdrawal reflex in Aplwia by presynaptic facilitation involving CAM P-dependent closure of a serotonin-sensitive potassium channel. Proc. Nutl. Acad. Sci. CSA 8 1 : 7956-7960, 1984. ACOSTA-URQUIDI, J. Modulation of calcium current and diverse K+ currents in identified Hermissenda neurons by small cardioactive peptide B. J. Neurosci. 8: 1694- 1703, 1988. ADAMS, D. J. AND GAGE, P. W. Divalent ion currents and the delayed potassium conductance in an Ap/ysia neurone. J. Physiol. Lond. 304: 297-3 13, 1980a. ADAMS, D. J., SMITH., S. J., AND THOMPSON, S. H. Ionic currents in molluscan soma. Annu. Rev. Neurosci. 3: 14 l- 167, 1980b. ALDRICH, D. J., GE-IIING, P. A., ANI> THOMPSON, S. H. Inactivation of delayed outward current in molluscan neurone somata. J. Physiol. Lond. 29 1: 507-530, 1979. BAXWR, D. A. AND BYRNE, J. H. Serotonin-modulated membrane currents Ap/y.sia tail sensory neurons. Sot. Neurosci. Abstr. 12: 765, 1986. BAXTER, D. A. AND BYRNE, J. H. Modulation of membrane currents and excitability by serotonin and CAMP in pleural sensory neurons of Aplysia. Sot. Neurosci. Abstr. 13: 1440, 1987.
ABRAMS, T. W.. CASWIJNXI, AND LLOYI), P. E. Two

MODULATION potassium 1981a. currents in a molluscan neuron. J.

OF

K+ CURRENTS

679

Gen. Physiol. 78: 63-86,

HERMANN, A. AND GORMAN, A. L. F. Effects of tetraethylammonium on potassium currents in a molluscan neuron. J. Gen. Physiol. 78: 87- 110,
1981b.

ROGAWSKI, M. A. The A-current: how ubiquitous a feature of excitable cells is it? Trends Neurosci. 8: 2 14-2 19, 1985. RUDY, B. Diversity and ubiquity of K channels. Neuroscience 25:
729-749, 1988.

SACKTOR, T. C.,

HERMANN, A. AND HARTUNG, K. Ca activated K+ conductance in molluscan neurons. Ceil Calcium 4: 387-405, 1983. HILLE, B. Ionic Channels ofExcitable Membranes. Sunderland, MA: Sinauer, 1984.

HOCHNER, B., BRAHA, O., KLEIN, M., AND KANDEL, E. R. Distantprocesses in presynaptic facilitation contribute to sensitization and dishabituation in Aplysia: possible involvement of C kinase in dishabituation. Sot. Neurosci. Abstr. 12: 1340, 1986a. HOCHNER, B., KLEIN, M., SCHACHER,S.,ANDKANDEL, E.R.Additional components in the cellular mechanism of presynaptic facilitation contributes to behavioral dishabituation in Aplysia. Pro& Natl. Acad. Sci. USA 83: 8794-8798, 1986b. HOCKBERGER, P. AND CONNOR, J. A. Alteration of calcium conductances and outward current by cyclic adenosine monophosphate (CAMP) in neurons of Limax maximus. Cell. Mol. Neurobiol. 4: 3 19-338, 1984. JUNGE, D. Calcium dependence of A-current in perfused Aplysia neurons. Brain Res. 346: 294-300, 1985. KANDEL, E. R. AND SCHWARTZ, J. H. Molecular biology of learning: modulation of transmitter release. Science Wash. DC 218: 433-442, 1982. KEHOE, J. Synaptic block of a calcium-activated potassium conductance in Aplysia neurones. J. Physiol. Lond. 369: 439-474, 1985. KLEIN, M., CAMARDO, J., AND KANDEL, E. R. Serotonin modulates a specific potassium current in the sensory neurons that show presynaptic facilitation in Apl-vsia. Proc. Natl. Acad. Sci. USA 79: 57 13-57 17, 1982. KLEIN, M., HOCHNER, B., AND KANDEL, E. R. Facilitatory transmitters and CAMP can modulate accommodation as well as transmitter release in Aplysia sensory neurons: evidence for parallel processing in a single cell. Proc. Natl. Acad. Sci. USA 83: 7994-7998, 1986. KLEIN, M. AND KANDEL, E. R. Mechanism of calcium current modulation underlying presynaptic facilitation and behavioral sensitization in Aplysia. Proc. Natl. Acad. Sci. USA 77: 69 12-69 16, 1980. LAGRUTTA, A. A., GOURDON, I., AND HARTZELL, H. C. ,&adrenergic modulation of delayed rectifier K currents in bullfrog atria1 myocytes (Abstract). Biophys. J. 55: 549a, 1989. LATORRE, R., CORONAD, R., AND VERGARA, C. K channel gated by voltage and ions. Annu. Rev. Physiol. 46: 485-495, 1984. Lo, L.-T., BYRNE, J. H., AND CLEARY, L. J. Distribution of three modulatory transmitters within the pleural ganglion ofAplysia. Sot. Neurosci. Abstr. 13: 1440, 1987. MALLART, A. A calcium-activated potassium current in motor nerve terminals of the mouse. J. Physiol. Land. 368: 577-59 1, 1985. MONTAROLO, P. G., GOELET, P., CASTELLUCCI, V. F., MORGAN, R., KANDEL, E. R., AND SCHACHER,S. A critical period for macromolecular synthesis in long-term heterosynaptic facilitation in Aplysia. Science Wash. DC 234: 1249- 1254, 1986. OCORR, K. A. AND BYRNE, J. H. Membrane responses and changes in CAMP levels in Aplysia sensory neurons produced by serotonin, tryptamine, FMRFamide and small cardioactive peptide (SCPB). Neurosci. Lett. 55: 65-72, 1985. OCORR, K. A. AND BYRNE, J. H. Evidence for separate receptors that mediate parallel effects of serotonin and small cardioactive peptideB (SCPs) on adenylate cyclase in Aplysia cakJkrnica. Neurosci. Lett. 70: 283-288, 1986. PAUPARDIN-TRITSCH, D., DETERRE, P., ANDGERSCHENFELD, H.M.Relationship between two voltage-dependent serotonin responses of molluscan neurones. Brain Res. 2 17: 20 l-206, 198 1. PAUPARDIN-TRITSCH, D., HAMMOND,~., AND GERSCHENFELD, H. M. Serotonin and cyclic GMP both induce an increase of the calcium current in the same identified molluscan neurons. J. Neurosci. 6: 27 15-2723, 1986. PELLMAR, T. C. Enhancement of inward current by serotonin in neurons of Ap!ysia. J. Neurobiol. 15: 13-25, 1984. POLLOCK, J. D., BERNIER, L., AND CAMARDO, J.S.Serotonin andcyclic adenosine 3:5-monophosphate modulate the S potassium current in tail sensory neurons in the pleural ganglion of Aplysia. J. Neurosci. 5: 1862-1871, 1985. POLLOCK, J. D. AND CAMARDO, J. S. Regulation of single potassium channels by serotonin in the cell bodies of the tail mechanosensory neurons of Aplysia californica. Brain Res. 4 10: 367-370, 1987. .ITCHIE,A. K. Two distinct calcium-activated potassium currents in a rat anterior pituitary cell line . J. Physiol. Land. 385: 59 l-609, 1987.

OBRIAN, C. A., WEINSTEIN, I. B., AND SCHWARTZ, J. H. Translocation from cytosol to membrane of protein kinase C after stimulation of Aplysia neurons with serotonin. Sot. Neurosci. Abstr. 12: 1340, 1986. SAWADA, M., CLEARY, L., AND BYRNE, J. H. Inositol triphosphate (IP3) and activators of protein kinase (PKC) modulate membrane currents in tail motors neurons of Aplysia. J. Neurophysiol. 6 1: 302-3 10, 1989. SAWADA, M., ICHINOSE, M., AND MAENO, T. Ionic mechanism ofthe outward current induced by intracellular injections of inositol triphosphate into Aplysia neurons. J. Neurosci. 7: 1470- 1483, 1987. SCHOLZ, K. P. AND BYRNE, J. H. Long-term sensitization in Aplysia. biophysical correlates in tail sensory neurons. Science Wash. DC 235: 685-687, 1987. SCHOLZ, K. P. AND BYRNE, J. H. Intracellular injection of CAMP induces a long-term reduction of neuronal K+ currents. Science Wash. DC 240: 1664-1666, 1988. SHUSTER, M.J., CAMARDO, J.S., SIEGELBAUM,~. A., ANDKANDEL, E.R. Cyclic AMP-dependent protein kinase closes serotonin-sensitive K+ channels of Aplysia sensory neurons in cell-free membrane patches. Nature Land. 3 13: 392-395, 1985. SHUSTER, M. J. AND SIEGELBAUM, S. A. Pharmacological characterization of the serotonin-sensitive potassium channel of Aplysia sensory neurons. J. Gen. Physiol. 90: 587-608, 1987. SIEGELBAUM, S. A. The S-current: a background potassium current. In:

Neuromodulation,

The Biochemical Control gf Neuronal Excitability,

edited by L. K. Kaczmarek and I. B. Levitan. New York: Oxford Univ. Press, 1987, p. 187-205. SIEGELBAUM, S. A., CAMARDO, J. S., AND KANDEL, E. R. Serotonin and cyclic AMP close single K+ channels in Aplysia sensory neurons. Nature Land. 299: 413-417, 1982. SMART, T. G. Single calcium-activated potassium channels recorded from cultured rat sympathetic neurones. J. Physiol. Land. 389: 337-360, 1987. SPIRA, M.E., BLUMENFELD, H., SCHACHER,S.,ANDSIEGELBAUM,S. A. Measurement of free intracellular calcium in Aplysia sensory neurons using fura-2. Sot. Neurosci. Abstr. 13: 1238, 1987. STANFIELD, P. R. Tetraethylammonium ions and the potassium permeability of excitable cells. Rev. Physiol. Biochem. Pharmacol. 97: l-68, 1983. STRONG, J. A. Modulation of potassium current kinetics in bag cell neurons of Aplysia by an activator of adenylate cyclase. J. Neurosci. 4: 2772-2783, 1984. STRONG, J. A. AND KACZMAREK, L. K. Multiple components of delayed potassium current in peptidergic neurons of Aplvsia: modulation by an activator of adenylate cyclase. J. Neurosci. 6: 8 14-822, 1986. THOMAS, M. V. Voltage-clamp analysis of a calcium mediated potassium conductance in cockroach (Periplaneta americana) central neurones. J. Physiol. Lond. 350: 159- 178, 1984. THOMPSON, S. J. Three pharmacologically distinct potassium channels in molluscan neurones. J. Phvsiol. Lond. 265: 465-488, 1977. TSIEN, R. W. Calcium currents in heart cells and neurons. In: Neuromodulation. The Biochemical Control qf Neuronal Excitability, edited by L. K. Kaczmarek and I. B. Levitan. New York: Oxford Univ. Press, 1987, p. 206-242. WALSH, J. P. AND BYRNE, J. H. Forskolin mimics and blocks a serotonin-sensitive decreased K conductance in tail sensory neurons of Aplysia. Neurosci. Lett. 52: 7- 1 1, 1984. WALSH, J. P. AND BYRNE, J. H. Analysis of decreased conductance serotonergic response in Aplysia ink motor neuron. J. Neurophysiol 53: 590-602, 1985. WALSH, J. P. AND BYRNE, J. H. Modulation of a steady-state Ca2+-activated, K+ current in tail sensory neurons Aplysia: role of serotonin and CAMP. J. Neurophysiol. 6 1: 32-44, 1989. WALTERS, E. T., BYRNE, J. H., CAREW, T.J., AND KANDEL, E. R. Mechanoafferent neurons innervating tail of Aplysia. I. Response properties and synaptic connections. J. Neurophysiol. 50: 1522- 1542, 1983a. WALTERS, E. T., BYRNE, J. H., CAREW, T.J., AND KANDEL, E. R. Mechanoafferent neurons innervating tail of Aplysia. II. Modulation by sensitizing stimulation. J. Neurophysiol. 50: 1543- 1559, 1983b. ZHANG, Z., CLEARY, L. J., MARSHAK, D. W., AND BYRNE, J. H. Serotonergic varicosities make apparent syn .aptic contacts wi th pleural sensory neurons of Aplysia. Sot. Neurosci. Abstr. 14: 84 1 988.