- Email: [email protected]
Synaptic augmentation by 5-HT at rested aplysia sensorimotor synapses: Independence of action potential prolongation
Neuron,
Vol. 13, 159-166,
July, 1994, Copyright
0 1994 by Cell Press
Synaptic Augmentation by 5-HT at Rested Aplysia Sensorimotor Synapses: Independence of Action Potential Prolongation Marc Klein Laboratory of Neurobiology Clinical Research Institute and Centre de Recherche University of Montreal Montreal, Quebec H2W Canada
and Behavior of Montreal en Sciences Neurologiques lR7
Summary Short-term augmentation of synaptic transmission at sensory neuron synapses of Aplysia contributes to behavioral sensitization and is one of the current models for a cellular mechanism of learning. This neuromodulatory process, mediated at least in part by the facilitatory neurotransmitter serotonin (S-HT) acting through CAMP, has been thought to result largely from prolongation of the sensory neuron action potential (API. The quantitative contribution of AP prolongation to synaptic augmentation was examined using a new culture preparation that is favorable for controlling the voltage at the presynaptic terminals. Preventing AP prolongation by using unvarying voltage-clamp commands in place of triggered APs did not reduce augmentation significantly, and pharmacological prolongation of APs caused by a high concentration of 5-HT led to a negligible increase in the synaptic response. Together with earlier evidence against the involvement of changes in Caz+ current, these results suggest that synaptic augmentation may result from modulation of steps in the secretory process that lie distal to the flow of ion currents across the nerve terminal membrane. Introduction In the marine mollusk Aplysia, neuromodulation of synaptic transmission at sensory neuron-to-motor neuron synapses underlies several simple forms of learning that involve changes in defensive reflexes upon presentation of different stimuli (Pinsker et al., 1970; Kandel and Schwartz, 1982; Byrne, 1987). The facilitatory neurotransmitter serotonin (5-HT) contributes to short-term facilitation of neuronal transmission at these synapses-and thus to the behavioral processes of sensitization and dishabituation-through at least two presynaptic mechanisms involving at least two second-messenger systems (Hochner et al., 1986b; Braha et al., 1990; Ghirardi et al., 1992; Klein, 1993). The most recent version of the model for the enhancement of synaptic transmission during learning in this system states that at relatively rested synapses 5-HT increases transmitter release from the sensory neurons by causing, through an increase in intracellular CAMP, phosphorylation of molecules that decrease K+ currents in the presynaptic terminals. Depressing K+ currents slows the repolar-
ization of the action potential (AP) and thereby prolongs the presynaptic Ca*+ influx (Klein et al., 1980; Hochner et al., 1986a). In contrast to this process, called 5-HT-mediated synaptic augmentation (Klein, 1993), other processes independent of the configuration of the AP increase release at synapses that have been activated more extensively and that have consequently undergone homosynaptic depression, the endogenous decrease in transmitter release that is believed to contribute to habituation of the reflexes. These facilitatory processes, termed synaptic restoration (Klein, 1993), may be mediated by protein kinase C (PKC; Sacktor and Schwartz, 1990; Braha et al., 1990; Sossin and Schwartz, 1992; Sugita et al., 1992; Ghirardi et al., 1992). The evidence supporting the role of AP broadening in synaptic augmentation has largely been indirect because it has not been possible to reliably control the voltage at the presynaptic terminals from the recording site in the neuronal cell body. In the present study, a new preparation of Aplysia synapses in cell culture was used, one in which the voltage at the terminals could be well controlled with electrodes in the cell body, allowing a direct examination of the role of changes in the AP shape during synaptic augmentation. The results of this study indicate that augmentation can occur without prolongation of the AP and that the contribution of AP prolongation to synaptic augmentation is, at most, a minor one. These and other observations suggest that synaptic augmentation and synaptic restoration, although mediated through different biochemical pathways, both result from changes in release processes-such as mobilization of synaptic vesicles-that lie distal to the activation and modulation of ion channels in the presynaptic membrane. Involvement of CAMP-dependent protein kinase in synaptic augmentation thus appears to parallel the direct modulation of secretion by CAMP at the neuromuscular junction of the crayfish (Dixon and Atwood, 1989; Delaney et al., 1991), in pancreatic B cells (Hughes et al., 1989; Ammali et al., 1993), and in pituitary lactotrophs (Sikdar et al., 1990) and may be an expression of a general modulatory function of protein kinase A (PKA) in regulated secretion.
Results Transmitter Release Can Be Well Controlled by Voltage Clamping “Soma-to-Soma” Synapses in Culture When individual sensory neurons and motor neurons from Aplysia are cultured under appropriate conditions (see Experimental Procedures), they completely absorb their neurites and become approximately spherical in appearance. When one of these sensory neurons and one of these motor neurons are paired, synapses form between the tightly apposed cells with-
Neuron 160
Figure
1. Structure
of the
Sensory
(Left) Gross morphology of several (Right) Ultrastructure of presumed and postsynaptic neurons and the
Neuron-Motor
Neuron
Pairs
Used
in the Study
soma-soma pairs prepared as described in Experimental Procedures. sites of synaptic contact. Note the reciprocal interdigitation of short processes normal appearance of the synaptic structures. Bar, 50 urn (left); 200 nm (right).
out any outgrowth of processes onto the substrate (Figure 1, left). Although these cultures appear to develop direct soma-to-soma synaptic connections, ultrastructural examination reveals reciprocal invagination and interdigitation of fine processes between the cells resembling a miniature neuropil (Figure 1, right). Presumably as a consequence of the short length of these processes, it is possible to control transmitter release from the sensory neurons in a graded manner by varying the amplitude of the presynaptic depolarization under voltage clamp (Figure 2). As expected for release sites close to the site of recording, the dependence of release on the sensory neuron voltage closely parallels the voltage dependence of the Ca*+ current in these neurons (Boyle et al., 1984; Edmonds et al., 1990). Augmentation of Synaptic Transmission by 5-HT Is Unaffected by Prevention of AP Prolongation in the Sensory Neuron To test the contribution of changes in AP configuration to augmentation, it is necessary to block all changes in AP shape caused by the facilitatory transmitter while at the same time monitoring transmitter release by recording postsynaptically. This can be
between
the
pre-
achieved byvoltageclamping the presynaptic sensory neuron and eliciting transmitter release with voltageclamp commands in the shape of unvarying APs, adding 5-HT, and then determining whether the resulting facilitation differs from the facilitation induced with triggered APs in non-clamped cells. Since the success of this experiment depends critically on the quality of the clamp control of presynaptic voltage, it is importanttodemonstrateatthesametimethatincreases in release caused by artificial prolongation of the presynaptic AP can in fact be blocked. The first series of experiments was therefore designed to test, at the same synapse, the effect that voltage clamping the presynaptic sensory neuron had on augmentation induced by 5-HT and on the increase in release induced by prolongation of the AP by the K+ channel blocker tetraethylammonium chloride (TEA). Augmentation of the postsynaptic current (PSC) by a low concentration of 5-HT (0.5 PM) was examined by applying 5-HT after the second of two APs elicited at 1 min intervals. After two additional APs, TEA was added, and the additional increase in the PSC associated with the prolongation of the AP caused by TEA was measured. In the voltage-clamp part of the experiment, synapses were activated with voltage-clamp
AP Prolongation 161
2.5
in Facilitation
r
2.0 g
1.5
-
E l.O-
5 z g
0
0.5
-
0.0
-
-0.51 -30
/ P‘\ f p
/ o/O’O I
’
I
0
1
’
1
1
-20
-10
0
10
20
30
40
50
Voltage Figure 2. Voltage Soma Contacts
Dependence
of Transmitter
Release
at Soma-
(A) Five millisecond voltage-clamp steps from a holding potential of -50 mV to different levels (v pre) elicit graded changes in postsynaptic currents (I post). Presynaptic currents are shown in the middle records. Release was first detected at this synapse with a step to -10 mV. Note the ongoing homosynaptic depression that is evident on comparison of successive presentations of the samevoltage step. Here and in Figure3 inward presynaptic currents are carried predominantly by Na’. Calibrations, from top to bottom, are 3 nA, 20 nA, and 40 mV. (B) Summary of voltage dependence of release from four experiments like the one in (A); error bars represent SEM where they are greater than the radius of the symbols. In each experiment the values of the postsynaptic current were normalized to the average of those elicited by preceding and following steps to +I0 mV; the step to +I0 mV was repeated several times in each experiment in order to allow correction for homosynaptic de pression during the progress of the experiment.
commands in the shape of sensory neuron APs that had been previously recorded on tape, and the same protocol of application of 5-HT and TEA was repeated. If control of the presynaptic voltage was adequate to block the increase in release caused by TEA-which acts by prolonging triggered APs-then any increase caused by 5-HT would be attributable to factors other than a change in the configuration of the presynaptic depolarization. To determine the magnitude of the effects of TEA alone, a separate set of experiments was done in non-clamped cells with application of 5-HT but without application of TEA. Synaptic augmentation by 5-HT was completely independent of AP prolongation in these experiments: application of 5-HTcaused identical increases in PSCs elicited with voltage commands in the shape of APs and in PSCs elicited with triggered APs (Figure 3). In contrast, the increase in the PSC caused by TEA when triggered APs were evoked was completely blocked
when voltage-clamp commands were used instead (Figure 3). Since these experiments were intended to examine the contribution of AP prolongation to synaptic augmentation, it was also necessary to determine how much prolongation occurred when the neurons were not voltage clamped. Surprisingly, adding 5-HT in these experiments actually caused a slight narrowing of the triggered APs (to 96.5% + 1.0% of the value before 5-HT, measured from the peak of the AP to 50% of the peak amplitude on the falling limb; n = 17). This observation already implies that 5-HT can cause significant synaptic augmentation without an increase in the AP duration. It is therefore not surprising that using a voltage-clamp protocol to block broadening did not change the augmentation. It might nonetheless be possible that with higher concentrations of 5-HT AP broadening does contribute significantly to facilitation. The experiment was therefore repeated with a higher concentration of 5-HT, with which measurable AP prolongation did occur. Increasing the concentration of 5-HT to 5 uM (n = 3) caused an increase of the peak and the area of the PSC to 298% + 35% and 408% f 66% of the control, respectively, but still narrowed the AP by about 5%. The experiment was therefore performed using 5-HT at 50 PM, a concentration that caused broadening of 4.8% + 1.2% (n = 7). The protocol of this set of experiments was the same as that used above, except that the higher concentration of 5-HT was added and that thecontrolwithoutTEAadditionwasomitted. In these experiments,theeffectsof5-HTonthePSCwere15%20% greater with triggered APs than with voltageclamp commands in the shape of APs; however, the difference was not significant (p > .5; Figure 4). As earlier, adding TEA increased the PSC when triggered APs were used but not when voltage-clamp commands were used. As a further internal check on voltage control of release, the voltage clamp was turned off after two stimuli delivered in the presence of 5-HTand TEA, and natural APs were triggered in order to see whether releasing the clamp would cause the PSC to increase as a result of the broadening of the AP caused by the combination of TEA and 5-HT. When the clamp was released and an AP triggered in the continuation of the earlier experiments, the PSC increased (by 17% f 4% for the peak and 38% + 11% for the area; n = 6);intheexperimentsusingtriggeredAPsthroughout, the corresponding PSCs decreased as a result of homosynaptic depression (by 25% + 4% for the peak and 19% + 7% for the area; n = 6; pp& < .OOl, result shows both that the PSC of P area < .005). This the clamped neurons could have been increased by broadening had it been permitted to occur and that the voltage clamp was effective in preventing broadening. The absence of a significant difference between 5-HT-induced augmentation at clamped and nonclamped synapses argues that AP prolongation is not
NelJlDll 162
Figure 3. Augmentation by 0.5 uM 5-HT at Clamped and at Non-Clamped Sensory Neuron Synapses
5HT
TEA
SHT TEA
5HT
(A) In representative records from two experiments, 5-HTat 0.5 uM applied after PSC #!Z augments the PSC (uppermost traces in each set) elicited by triggered APs (top) and by voltage-clamp commands with the waveforms of naturai APs (bottom). In contrast,addingTEAafter PSC#4enhances the PSConly in the non-clamped synapse (top), at which AP prolongation is caused by the TEA. Note that the presynaptic voltageclampcurrent (bottom traces in the CLAMP TEA records) moves inward in response to the TEA, whereas the PSC is not increased because of the prevention of AP prolongain the presynaptic current records; the slow initial rise of Calibration is 5 nA for the PSCs, 25 mV for the voltage
tion. No corrections for leak or capacitive components have been made the command voltage precludes the appearance of a capacitive transient. records, and 25 nA for the presynaptic currents. (B) Summary of the effects of 0.5 uM 5.HT and 10 mM TEA on PSCs at clamped and at non-clamped synapses. Change in the PSC is expressed relative tothe PSC before addition of 5-HTor TEA in the respective experiments, i.e., (PSC posttreatment/PSC pretreatment)-I. Open bars show values for PSCs elicited with triggered APs, hatched bars are for PSCs elicited with voltage-clamp commands, and closed bars are for control PSCs elicited with triggered APs without addition of TEA. Error bars show SEM. The values are 1.24 * 0.19 (n = 15, APs) and 1.24 I 0.19 (n = 9, clamp commands) for PSC peaks in response to 5-HT and 1.63 + 0.26 (APs) and 1.64 f 0.18 (clamp commands) for PSC areas in the same experiments. Responses to TEA were 0.42 + 0.07 (n = 9, APs), -0.04 L 0.11 (n = 9, clamp commands), and -0.03 k 0.16 (n = 8, control APs without TEA) for PSC peaks; and 0.54 + 0.09 (APs), -0.02 f 0.13 (clamp commands), and 0.03 f 0.21 (APs without TEA) for PSC areas. For the TEA responses, synaptic responses to triggered APs with TEA differed from both the clamped synapses (p < .005 for both peaks and areas) and the controls without TEA (p < .02 for peaks and p < .05 for areas). The clamped synapses and the controls were not different from each other (p > .5 for both peaks and areas).
a major contributor to facilitation by5-HT. In addition, the ability of 5-HT to cause greater than 2-fold facilitation at 0.5 uM and 5 PM, at which the AP actually narrows, supports this conclusion. However, with higher concentrations of 5-HT, at which AP broadening does occur, a small contribution of broadening might be difficult to detect simply because of the relatively largevariability in the augmentation responses. The final set of experiments was done in order to achieve an estimate of the enhancement in the PSC that would be expected as a result of the observed AP prolongation caused by 5-HT. The Relation between AP Prolongation and Enhancement of Transmitter Release Is linear with a Slope of Approximately 1 In earlier studies rectangular voltage-clamp depolarizations of varying duration were used to assess the relation between the duration of the presynaptic depolarization and the postsynaptic response (Hochner et al., 1986a, 198613). However, there is no obvious way to extrapolate from the dependence of release on the duration of rectangular steps to the dependence on AP duration, especially if the release caused by the rectangular steps falls in a region that is below what normally occurs with release induced by APs. In particular, it might be expected that in the range of duration that is just above release threshold the slope of the relation might be much steeper than it is at durations that cause more normal release. In addition, without doing detailed simulations, it is impossible to know how the precise configuration of the depolarization will affect release. The best way to examine
the relation between fore, is to use stimuli APs to evoke release 1990).
AP duration and release, therewith the waveforms of natura! (Llinas et al., 1982; Augustine,
3.5 3.0 0. .-c & 5 r 0
2.0 1.5 1.0 0.5
z
0.0 2.5 /
-0.5 1 5HT
TEA
5HT
TEA
Figure 4. Augmentation by 50 uM 5.HT at Clamped Clamped Sensory Neuron Synapses
and at Non-
Change in the PSC is expressed relative to the PSC before addition of 5-HTor TEA in the respective experiments, i.e., (PSC posttreatment/PSC pretreatment)-I. Open bars show values for PSCs elicited with triggered APs, and hatched bars are for PSCs elicited with voltage-clamp commands. Error bars show SEM. Values are 1.94 k 0.48 (n = 7, APs) and 1.63 ?r 0.23 (n = 6, clamp commands) for PSC peaks in response to 50 uM 5-Hi; and 3.24 -i- 0.68 (APs) and 2.82 f 0.38 (clamp commands) for areas in the same experiments (p > .5 for both peaks and areas). In response to subsequent addition of IO mM TEA, values are 0.43 + 0.20 (APs) and -0.14 2 0.04 (clamp commands) for peaks (p < .OS); and 1.23 + 0.26 (APs) and -0.17 + 0.10 (clamp commands) for areas (p < .OOl).
AP Prolongation 163
in Facilitation
Table 1. Increase in AP Duration and PSC Area Caused TEA in the Absence and in the Presence of 5-HT AP Duration (Post-TEA)
PSC Area (Post-TEA)
n
(PreTEA)
(Pm-TEA)
Ratio
0
6 8 9
1.58 1.78 1.62
1.54 1.86 1.38
0.97 1.04 0.85
0.5
9 9
1.74 1.47
1.54 1.29
0.89 0.88
5
3
1.85
1.61
0.87
50
5 6
1.82 1.78
2.33 1.70
1.28 0.96
2.5 5-HT Concentration
(NW
1.0
0.5 0.0
~0.0
0.5
1.0
1.5
2.0
2.5
AP Duration Figure
5. Relation
Between
PSC Area
and AP Duration
in TEA
See text for details of the experiment. APs broadened by TEA were shortened by injection of hyperpolarizing current pulses and then permitted to broaden again in successive stimuli. PSC area as a proportion of the pm-TEA PSC and corrected for homosynaptic depression was plotted against AP duration for 47 measurements at 17 synapses (8 with tail sensory neurons and 9 with siphon sensory neurons). In 9 preparations examined without TEA for purposes of correction for homosynaptic depression, average values for successive PSCs were 0.61,0.60, and 0.51 for the second, third, and fourth PSCs, respectively, relative to the first (synapses with tail sensory neurons). For the synapses with siphon sensory neurons, the corresponding values were taken as 0.50,0.45, and 0.40 based on comparison with the experiments using the tail sensory neurons. The least-squares regression line has a slope of 0.913 with correlation coefficient 0.581 (p < .Ol).
The dependence of synaptic release on presynaptic AP duration was measured in the following manner: the sensory neuron member of a soma-soma couple was fired once with an intracellular depolarizing stimulus; TEA was then added, and the increases in AP duration and in the PSC were measured in response toasecond stimulus. Forthe next stimulus, a hyperpolarizing current step was imposed on the sensory neuron beginning at the point where the AP in TEA diverged from the control AP; this was done in order to shorten the AP, thereby attenuating the PSC (see Gingrich and Byrne, 1985). The final stimuluswasonce again an AP thatwas not shortened by a hyperpolarization. Thus, each experiment gave control values for AP duration and PSC amplitude and area, as well as three additional values for each measure, two with broadened APs in TEA and a third with an artificially shortened AP, also in the presence of TEA. To correct for the decline in successive PSCs caused by homosynaptic depression, measurements were made in a separate set of control synapses without adding TEA or imposing a hyperpolarizing step. The values of the experimental PSCs were normalized to the corresponding control values. The relation between the postsynaptic response and the duration of the AP can be approximated by a straight line with a slope of about 1 (Figure 5; Table 1). Applying this relation to the AP prolongation caused
tion
by 5-HT
alone
leads
accounted
to the
for
conclusion
no more
that
than
a 5%
prolonga-
increase
by
For each set of experiments, the average AP duration and PSC area were expressed as a proportion of the pm-TEA value, and the ratio between the two averages was calculated as a measure of the relation between AP prolongation and PSC increase. PSCs were corrected for homosynaptic depression; correction factors ranged from 1.00 to 1.35 in different sets of experiments. The values were derived from all the data presented in the text and figures, with one additional set of values for TEA alone (n = 6 for 0 uM 5-HT).
in the PSC when 5-HT was applied at 50 PM. The fact that this concentration of 5-HT caused an increase of around 300% in the PSC area means that AP prolongation contributed negligibly to the facilitation in this case. Furthermore, the possibility that AP duration might have had a greater influence on the PSC in the presence of 5-HT is addressed by the data in Table 1, which shows that the slope of the relation between PSC area and AP duration was not changed by 5-HT. It is important to point out that TEA appears to have had no direct effect on synaptic transmission here. That TEA acts only by AP prolongation follows from the fact that preventing prolongation by voltage clamping in the presence of TEA blocked the increase in the PSC(Figure3; Figure4).Theequivalentobservation in the present set of experiments is that the fitted line of Figure 5 passes close to the point (l.O,l.O), i.e., that in the presence of TEA an AP that is of the same duration as the control will elicit a PSC that is also the same as the control. Discussion In contrast to the present findings, several earlier reports concluded that AP prolongation contributes significantly to synaptic augmentation (Hochner et al., 1986a, 1986b; Eliot et al., 1993). There are several possible reasons for this discrepancy. First, as discussed above, the relation between the duration of presynaptic depolarization and the postsynaptic response was examined with rectangular voltage-clamp pulses whose relation to normal APs is unclear and may therefore have led to an overestimate of the steepness of this relation. Second, augmentation in the earlier studies was tested with high concentrations of 5-HT
(IO-100 PM) that, combined with repeated firing and consequent K+ current inactivation, caused much greater AP prolongation than in the present study but a similar increase in the synaptic response. It is not clear whether lower concentrations of 5-HT and lower stimulation frequencies causing far less AP broadening would have given less augmentation. The effect of 5-HT on AP broadening is very sensitive to the frequencyofAPfiring: inasmall numberofexperiments, 50 uM 5-HTcaused prolongation of 2.7% + 1.3% when stimulated at 1 min intervals (n = 3), whereas stimulation at 30 and 15 s intervals caused increases of 14.5% + 4.5% (n = 2) and 11.3% f 2.1% (n = 4), respectively. Finally, the one study in which TEA was used to examine the relation between AP duration and the postsynaptic response in the intact abdominal ganglion (Gingrich and Byrne, 1985) demonstrated a nearlinear relation at shorter durations similar to that observed here. Studies in the squid (Augustine, 1990) and in the crayfish (Delaney et al., 1991) also demonstrate linear or near-linear relations between AP duration and transmitter release, albeit somewhat steeper than that shown here. There is no reason to suspect that the differences between the conclusions of the present study and earlier work is the result of some fundamental difference between the cultures used here and the preparations used previously. At the ultrastructural level, these synapses and their surroundings are indistinguishable from those in the intact nervous system (compare Figure 1, right, with Bailey et al., 1979). In particular, the synapses are found at fine processes of the presynaptic neuron rather than coming directly off the cell body, suggesting that cytoskeletal architecture as well as gradients and accumulation of ions will be similar to those in intact ganglia. Furthermore, the amplitudes of the synaptic responses are comparable to those in earlier studies. They show normal homosynaptic depression, and the facilitation induced by 5-HTboth augmentation and restoration is similar to that observed in other studies (Klein, 1993; Ghirardi et al., 1992; Eliot et al., 1993). It is therefore likely that the conclusions reached here are equally applicable to synapses in the nervous system. The results presented here are consistent with earlier experiments showing that the PSC can be augmented by direct activation of adenylyl cyclase without an increase in AP duration (Klein, 1993). The present results show that augmentation by low concentrations of 5-HT also can occur without AP prolongation. In addition, blocking the AP prolongation caused by a high concentration of 5-HT has no significant effect on augmentation, and mimicking this amount of broadening by other means causes only a minimal increase in the synaptic response. The sum total of these findings argues that AP prolongation contributes minimally to synaptic augmentation by 5-HT. There are two possible mechanisms for the facilitating action of PKA that are consistent with the present
results. The first is that PKA activation leads to direct modulation of the Ca*+ influx through voltageactivated Ca*+ channels. Previous work has demonstrated that S-HT increases one type of Ca2’ current in isolated sensory neurons, butthat this current does not appear to contribute to normal transmitter release at synapses (Edmonds et al., 1990; Braha et al., 1993). On the other hand, a current that does contribute to release does not appear to be modulated by 5-HT (Edmonds et al., 1990). Nonetheless, the possibility that there may exist additional Ca*+ currents which are modulated by PKA and play a role in transmitter releasecannotyet beexcluded, norcan the possibility that modulation of currents in isolated sensory neurons differs from that in neurons which have formed synaptic connections. The second possible mode of action of PKA in augmenting transmitter release is through an effect on components of the release process that come into play subsequent to Ca*+ entry. Earlier work showed that restoration of depressed synaptic transmission by 5-HT is largely independent of AP prolongation (Hochner et al., 1986b). Later studies proposed that synaptic augmentation at rested synapses is mediated by activation of CAMP-dependent protein kinase (Braha et al., 1990; Ghirardi et al., 1992; Klein, 1993), whereas synaptic restoration may be caused by activation of PKC (Braha et al., 1990; Sacktor and Schwartz, 1990; Ghirardi et al., 1992; Sugita et al., 1992). The increase in spontaneous release of transmitter caused by 5-HT can be mimicked by activation of PKC and blocked by an inhibitor of PKC, but corresponding manipulations of the PKA system do not affect spontaneous release (Ghirardi et al., 1992). Since the increase in spontaneous release can be dissociated from Ca*’ influx, it has been proposed that the increase reflects adirect modulation of the secretory apparatus by5-HT and PKC (Dale and Kandel, 1990). The present report is consistent with the idea that activation of PKA may also modulate release directly. However, the difference between the actions of the two kinase systems on spontaneous release suggests that they are involved at different steps in the secretory process. A study of the inhibitory actions of the tetrapeptide FMRFamide in Helisoma neurons concluded that this peptide directly modulates the secretory machinery (Man-Son-Hing et al., 1989). FMRFamide inhibits both spontaneous and evoked release from Aplysia sensory neurons as well (Belardetti et al., 1987; Dale and Kandel, 1990). It thus appears likely that FMRFamide in Aplysia directly inhibits transmitter release also. Since FMRFamide antagonizes the actions of 5-HT in both biochemical and physiological assays (Sweatt et al., 1989; Critz et al., 1991; lchinose and Byrne, 1991; Pieroni and Byrne, 1992), it is probable that it acts at some of the same sites as 5-HT. Molecular loci for the direct modulatory actions of 5-HT and FMRFamide on synaptic transmission have not been identified.An attractivecandidatefor regulation would be a synapsin-like molecule that can con-
AP Prolongation 165
in Facilitation
trol the availability of synaptic vesicles for release as a function of its state of phosphorylation (LlinBs et al., 1991). Ultrastructural work of Bailey and Chen (1988) on sensorimotor synapses of Aplysia demonstrated a local depletion of synaptic vesicles from the vicinity of the presynaptic release sites with short-term homosynaptic depression. This work suggests that the mobilization of vesicles to docking sites at the presynaptic membrane is under the tight control of elements in the nerve terminal,whether cytoplasmic, cytoskeletal, or vesicular in nature; and the later studies therefore suggest that interactions among these molecules may be regulated by PKC in synaptic restoration. The recent identification of some of the molecular components of the secretory apparatus (see Jesse11 and Kandel, 1993, for review) has begun to provide a number of candidates for the modulation of transmitter release that underlies the various forms of short-term synaptic and behavioral plasticity. The actions of CAMP at Aplysia and at crayfish synapses (Dixon and Atwood, 1989; Delaney et al., 1991), as well as on Ca2+ currents and on Ca*+ current-independent modulation of secretion in b cells of the mammalian pancreas (Hughes et al., 1989; WmmB;Iti et al., 1993) and in mammalian lactotrophs (Sikdar et al., 1990), suggest that modulation by protein kinases of processes involved in secretion involves analogous components of the secretory systems of different cell types with physiological consequences ranging from glucose and Ca2+ regulation to the induction of short-term memory processes.
Experimental
Procedures
Tail sensory neurons and LFS motor neurons were dissociated as described (Klein, 1993; a number of siphon sensory neurons were included in the experiment of Figure 5) and then maintained in 10% Aplysia hemolymph for l-2 days until all neuronal processeswere absorbed (see also Haydon, 1988). Individual sensory and motor neurons were then paired and left in 10% hemolymph for 24 hr or less to form synapses. For recording, single pairs were transferred to uncoated plastic culture dishes with recording medium (Leibovitz L15 supplemented with salts; see Klein, 1993), where they adhered tightly to the bottom of the dish without further treatment. Pairs not used immediatelywere kept at 4OC until use. Sensory neurons were impaled with two micropipettes filled with 2 M potassium acetate for intracellular recording or voltage clamping (using an Axoclamp 2A; Axon Instruments) and held at -50 mV in all experiments. IFS neurons were voltage clamped with a single 2 M potassium acetate micropipette at -80 mV with a second Axoclamp 2A in the discontinuous voltage-clamp mode set at a sampling frequency of 6 kHz. All recordingsweredone in standard recording medium (above); no modificationswere made to improvevoltage control. TEA and 5-HT (creatinine sulfate; both from Sigma) were applied directly onto the cells in a volume of 50 ~1 from a hand-held Pipetman; the final concentration was estimated from separate dye dilution experiments to be approximately half the concentration in the added aliquot. When TEA was added after 5-HT application, it wasapplied in asolutioncontainingtheestimatedconcentration of 5-HT already present; in the control experiments of Figure 3, a second application of 5-HT alone in this manner caused no change in the synaptic response, confirming the accuracy of the estimate. Records were stored on videotapewith a PCM recorder and analyzed using the Spike program (Hilal Associates). Stu-
dent’s two-tailed t test was used for all statistical data are expressed as mean + SEM.
comparisons;
Acknowledgments This work was supported by grants from the National Institutes of Health (MH45397) and the Natural Sciences and Engineering Research Councilof Canada(OGP0138426)and bya5loan Fellowship. I would like to thank V. Castellucci, J. Koester, J. H. Schwartz, W. Sossin, and L.-E. Trudeau for their comments on an earlier version of the manuscript and Annie Campbell for the electron microscopy. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC Section 1734 solely to indicate this fact. Received
February
18, 1994; revised
May
4, 1994.
References ammlll, C., Ashcroft, F. M., and Rorsman, P. (1993). Calciumindependent potentiation of insulin release by cyclic AMP in single p-cells. Nature 363, 356-358. Augustine, G. (1990). Regulation squid giant synapse by presynaptic current. J. Physiol. 437, 343-364.
of transmitter release at the delayed rectifier potassium
Bailey, C. H., and Chen, M. (1988). Morphological basis term habituation in Aplysia. J. Neurosci. 8, 2452-2459.
of short-
Bailey, C. H., Thompson, E. B., Castellucci, V. F., and Kandel, E. R. (1979). Ultrastructure of the synapses of sensory neurons that mediate the gill-withdrawal reflex in Aplysia. J. Neurocytol. 8, 415-444. Belardetti, F., Kandel, E. R., and Siegelbaum, 5. A. (1987). Neuronal inhibition by the peptide FMRFamide involves opening of 5 K+ channels. Nature 325, 153-156. Boyle, M. B., Klein, M., Smith, 5. J., and Kandel, E. R. (1984). Serotonin increases intracellular Ca2+ transients in voltageclamped sensoryneuronsof Aplysiacalifornica. Proc. Natl.Acad. Sci. USA 87, 7642-7646. Braha, O., Dale, N., Hochner, B., Klein, M., Abrams, T. W., and Kandel, E. R. (1990). 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. Braha, O., Edmonds, B., Sacktor, T., (1993). The contributions of protein C to the actions of 5-HT on the L-type neurons in Aplysia. J. Neurosci. 73, Byrne, J. H. (1987). Cellular iol. Rev. 67, 329-439.
analysis
Kandel, E. R., and Klein, M. kinase A and protein kinase CaZ+ current of the sensory 1839-1851.
of associative
learning.
Phys-
Critz, 5. D., Baxter, D. A., and Byrne, J. H. (1991). Modulatory effects of serotonin, FMRFamide, and myomodulin on the duration of action potentials, excitability, and membrane currents in tail sensory neurons of Aplysia. J. Neurophysiol. 66, 19121926. Dale, N., and Kandel, E. R. (1990). Facilitatoryand inhibitorytransmitters modulate spontaneous transmitter release at cultured Aplysia sensorimotor synapses. J. Physiol. 427, 203-222. Delaney, K., Tank, D. W., and Zucker, R. 5. (1991). Presynaptic calcium and serotonin-mediated enhancement of transmitter release at crayfish neuromuscular junction. J. Neurosci. 77,26312643. Dixon, D., and Atwood, H. L. (1989). Conjoint action of phosphatidylinositol and adenylate cyclase systems in serotonin-induced facilitation at the crayfish neuromuscular junction. 1. Neurophysiol. 62, 1251-1259. Edmonds, B., Klein, M., Dale, N., and Kandel, E. R. (1990). Contributions of two types of calcium channels to synaptic transmission and plasticity. Science 2.50, 1142-1147.
Neuron 166
Eliot, L. S., Kandel, E. R., Siegelbaum, (1993). Imaging terminals of Aplysia strates role of enhanced Ca*+ influx Nature 367, 634-637.
S. A., and Blumenfeid, H. sensory neurons demonin presynaptic facilitation.
Ghirardi, M., Braha, O., Hochner, B., Montarolo, P. G., Kandel, E. R., and Dale, N. (1992). Roles of PKA and PKC in facilitation of evoked and spontaneous transmitter release at depressed and nondepressed synapses in Aplysia sensory neurons. Neuron 9, 479489. Gingrich, K. J., and Byrne, J. H. (1985). Simulation of synaptic depression, posttetanic potentiation, and presynaptic facilitation of synaptic potentials from sensory neurons mediating the gill-withdrawal reflex in Aplysia. J. Neurophysiol. 53, 652-669. Haydon, P. C. (1988). The formation tween cell-cultured neuronal somata.
of chemical J. Neurosci.
synapses be8, 1032-1038.
Hochner, B., Klein, M., Schacher, S., and Kandel, E. R. (1986a). Action potential duration and the modulation of transmitter release from the sensory neurons of Aplysia in presynaptic facilitation and behavioral sensitization. Proc. Natl. Acad. Sci. USA 83, 8410-8414. Hochner, B., Klein, M., Schacher, S., and Kandel, E. R. (1986bj. Additional component in thecellular mechanism of presynaptic facilitation contributes to behavioral dishabituation in Aplysia. Proc. Natl. Acad. Sci. USA 83, 8794-8798. Hughes, S. J., Chalk, J. G., and Ashcroft, S. J. H. (1989). Effect of secretagogues on cytosolic free Ca2+ and insulin release at different extracellular Ca2+ concentrations in the hamster clonal B-cell line HIT-T15. Mol. Cell. Endocrinol. 65, 35-41. Ichinose, M., and Byrne, J. H. (1991). Role of protein phosphatases in the modulation of neuronal membrane currents. Brain Res. 549, 146-150. Jessell, T. M., and Kandel, E. R. (1993). bidirectional and self-modifiable form tion. Cell 72/Neuron 70 (SuppI.), I-30. Kandel, learning: 443.
Synaptic transmission: of cell-cell communica-
a
E. R., and Schwartz, J. H. (1982). Molecular biology of modulation of transmitter release. Science 278, 433-
Klein, M. (1993). Differential cyclic AMP dependence of facilitation at Aplysia sensorimotor synapses as a function of prior stimulation: augmentation versus restoration of transmitter release. J. Neurosci. 73, 3793-3801. Klein, M., Shapiro, and the modulation
E., and Kandel, E. R. (1980). Synaptic plasticity of the Ca*+ current. J. Exp. Biol. 89,117-157.
Llinas, R., Sugimori, M., and by spike-like depolarizations Acad. Sci. USA 79, 2415-2419.
Simon, in the
S. M. (1982). Transmission squid synapse. Proc. Natl.
Llinls, R., Cruner, J. A., Sugimori, M., McCuinness, T. L., and Greengard, P. (1991). Regulation by synapsin I and Ca2+calmodulin-dependent protein kinase Ii of transmitter release in squid giant synapse. 1. Physiol. 436, 257-282. Man-Son-Hing, H., Zoran, M. J., Lukowiak, K., and Haydon, P. G. (1989). A neuromodulator of synaptic transmission acts on the secretory apparatus as well as on ion channels. Nature 347, 237-239. Pieroni, J. P., and Byrne, J. H. (1992). Differential effects of serotonin, FMRFamide, and small cardioactive peptide on multiple, distributed processes modulating sensorimotor synaptic transmission in Aplysia. J. Neurosci. 12, 2633-2647. Pinsker, H., Kupfermann, I., Castellucci, E. R. (1970). Habituation and dishabituation reflex in Aplysia. Science 767, 1740-1742.
V. F., and Kandel, of the gill-withdrawal
Sacktor,T.C.,andSchwartz,J. H.(1990).Sensitizingstimulicause translocation of protein kinase C in Aplysia sensory neurons, Proc. Natl. Acad. Sci. USA 87, 2036-2039. Sikdar, S. K., Zorec, facilitates Ca-induced Lett. 273, 150-154.
R., and Mason, W. T. (1990). CAMP exocytosis in bovine lactotrophs.
directly FEBS
Sossin, W. S., and Schwartz, J. H. (1992). Selective activation Ca*‘-activated PKCs in Aplysia neurons by 5.HT. J. Neurosci. 1160-1168.
of 72,
Sugita, S., Goldsmith, 1. R., Baxter, D. Involvement of protein kinase C in broadening and synaptic facilitation tions of Aplysia. 1. Neurophysiol. 68,
A., and Byrne, J. H. (19923. serotonininduced spike in sensorimotor connec643-651.
Sweatt, J. D., Volterra, A., Edmonds, B., Karl, K. A., Siegelbaum, S. A., and Kandel, E. R. (1989). FMRFamide reverses protein phosphorylation produced by5-HTand CAMP in Aplysia sensory neurons. Nature 342, 275-278.
Vol. 13, 159-166,
July, 1994, Copyright
0 1994 by Cell Press
Synaptic Augmentation by 5-HT at Rested Aplysia Sensorimotor Synapses: Independence of Action Potential Prolongation Marc Klein Laboratory of Neurobiology Clinical Research Institute and Centre de Recherche University of Montreal Montreal, Quebec H2W Canada
and Behavior of Montreal en Sciences Neurologiques lR7
Summary Short-term augmentation of synaptic transmission at sensory neuron synapses of Aplysia contributes to behavioral sensitization and is one of the current models for a cellular mechanism of learning. This neuromodulatory process, mediated at least in part by the facilitatory neurotransmitter serotonin (S-HT) acting through CAMP, has been thought to result largely from prolongation of the sensory neuron action potential (API. The quantitative contribution of AP prolongation to synaptic augmentation was examined using a new culture preparation that is favorable for controlling the voltage at the presynaptic terminals. Preventing AP prolongation by using unvarying voltage-clamp commands in place of triggered APs did not reduce augmentation significantly, and pharmacological prolongation of APs caused by a high concentration of 5-HT led to a negligible increase in the synaptic response. Together with earlier evidence against the involvement of changes in Caz+ current, these results suggest that synaptic augmentation may result from modulation of steps in the secretory process that lie distal to the flow of ion currents across the nerve terminal membrane. Introduction In the marine mollusk Aplysia, neuromodulation of synaptic transmission at sensory neuron-to-motor neuron synapses underlies several simple forms of learning that involve changes in defensive reflexes upon presentation of different stimuli (Pinsker et al., 1970; Kandel and Schwartz, 1982; Byrne, 1987). The facilitatory neurotransmitter serotonin (5-HT) contributes to short-term facilitation of neuronal transmission at these synapses-and thus to the behavioral processes of sensitization and dishabituation-through at least two presynaptic mechanisms involving at least two second-messenger systems (Hochner et al., 1986b; Braha et al., 1990; Ghirardi et al., 1992; Klein, 1993). The most recent version of the model for the enhancement of synaptic transmission during learning in this system states that at relatively rested synapses 5-HT increases transmitter release from the sensory neurons by causing, through an increase in intracellular CAMP, phosphorylation of molecules that decrease K+ currents in the presynaptic terminals. Depressing K+ currents slows the repolar-
ization of the action potential (AP) and thereby prolongs the presynaptic Ca*+ influx (Klein et al., 1980; Hochner et al., 1986a). In contrast to this process, called 5-HT-mediated synaptic augmentation (Klein, 1993), other processes independent of the configuration of the AP increase release at synapses that have been activated more extensively and that have consequently undergone homosynaptic depression, the endogenous decrease in transmitter release that is believed to contribute to habituation of the reflexes. These facilitatory processes, termed synaptic restoration (Klein, 1993), may be mediated by protein kinase C (PKC; Sacktor and Schwartz, 1990; Braha et al., 1990; Sossin and Schwartz, 1992; Sugita et al., 1992; Ghirardi et al., 1992). The evidence supporting the role of AP broadening in synaptic augmentation has largely been indirect because it has not been possible to reliably control the voltage at the presynaptic terminals from the recording site in the neuronal cell body. In the present study, a new preparation of Aplysia synapses in cell culture was used, one in which the voltage at the terminals could be well controlled with electrodes in the cell body, allowing a direct examination of the role of changes in the AP shape during synaptic augmentation. The results of this study indicate that augmentation can occur without prolongation of the AP and that the contribution of AP prolongation to synaptic augmentation is, at most, a minor one. These and other observations suggest that synaptic augmentation and synaptic restoration, although mediated through different biochemical pathways, both result from changes in release processes-such as mobilization of synaptic vesicles-that lie distal to the activation and modulation of ion channels in the presynaptic membrane. Involvement of CAMP-dependent protein kinase in synaptic augmentation thus appears to parallel the direct modulation of secretion by CAMP at the neuromuscular junction of the crayfish (Dixon and Atwood, 1989; Delaney et al., 1991), in pancreatic B cells (Hughes et al., 1989; Ammali et al., 1993), and in pituitary lactotrophs (Sikdar et al., 1990) and may be an expression of a general modulatory function of protein kinase A (PKA) in regulated secretion.
Results Transmitter Release Can Be Well Controlled by Voltage Clamping “Soma-to-Soma” Synapses in Culture When individual sensory neurons and motor neurons from Aplysia are cultured under appropriate conditions (see Experimental Procedures), they completely absorb their neurites and become approximately spherical in appearance. When one of these sensory neurons and one of these motor neurons are paired, synapses form between the tightly apposed cells with-
Neuron 160
Figure
1. Structure
of the
Sensory
(Left) Gross morphology of several (Right) Ultrastructure of presumed and postsynaptic neurons and the
Neuron-Motor
Neuron
Pairs
Used
in the Study
soma-soma pairs prepared as described in Experimental Procedures. sites of synaptic contact. Note the reciprocal interdigitation of short processes normal appearance of the synaptic structures. Bar, 50 urn (left); 200 nm (right).
out any outgrowth of processes onto the substrate (Figure 1, left). Although these cultures appear to develop direct soma-to-soma synaptic connections, ultrastructural examination reveals reciprocal invagination and interdigitation of fine processes between the cells resembling a miniature neuropil (Figure 1, right). Presumably as a consequence of the short length of these processes, it is possible to control transmitter release from the sensory neurons in a graded manner by varying the amplitude of the presynaptic depolarization under voltage clamp (Figure 2). As expected for release sites close to the site of recording, the dependence of release on the sensory neuron voltage closely parallels the voltage dependence of the Ca*+ current in these neurons (Boyle et al., 1984; Edmonds et al., 1990). Augmentation of Synaptic Transmission by 5-HT Is Unaffected by Prevention of AP Prolongation in the Sensory Neuron To test the contribution of changes in AP configuration to augmentation, it is necessary to block all changes in AP shape caused by the facilitatory transmitter while at the same time monitoring transmitter release by recording postsynaptically. This can be
between
the
pre-
achieved byvoltageclamping the presynaptic sensory neuron and eliciting transmitter release with voltageclamp commands in the shape of unvarying APs, adding 5-HT, and then determining whether the resulting facilitation differs from the facilitation induced with triggered APs in non-clamped cells. Since the success of this experiment depends critically on the quality of the clamp control of presynaptic voltage, it is importanttodemonstrateatthesametimethatincreases in release caused by artificial prolongation of the presynaptic AP can in fact be blocked. The first series of experiments was therefore designed to test, at the same synapse, the effect that voltage clamping the presynaptic sensory neuron had on augmentation induced by 5-HT and on the increase in release induced by prolongation of the AP by the K+ channel blocker tetraethylammonium chloride (TEA). Augmentation of the postsynaptic current (PSC) by a low concentration of 5-HT (0.5 PM) was examined by applying 5-HT after the second of two APs elicited at 1 min intervals. After two additional APs, TEA was added, and the additional increase in the PSC associated with the prolongation of the AP caused by TEA was measured. In the voltage-clamp part of the experiment, synapses were activated with voltage-clamp
AP Prolongation 161
2.5
in Facilitation
r
2.0 g
1.5
-
E l.O-
5 z g
0
0.5
-
0.0
-
-0.51 -30
/ P‘\ f p
/ o/O’O I
’
I
0
1
’
1
1
-20
-10
0
10
20
30
40
50
Voltage Figure 2. Voltage Soma Contacts
Dependence
of Transmitter
Release
at Soma-
(A) Five millisecond voltage-clamp steps from a holding potential of -50 mV to different levels (v pre) elicit graded changes in postsynaptic currents (I post). Presynaptic currents are shown in the middle records. Release was first detected at this synapse with a step to -10 mV. Note the ongoing homosynaptic depression that is evident on comparison of successive presentations of the samevoltage step. Here and in Figure3 inward presynaptic currents are carried predominantly by Na’. Calibrations, from top to bottom, are 3 nA, 20 nA, and 40 mV. (B) Summary of voltage dependence of release from four experiments like the one in (A); error bars represent SEM where they are greater than the radius of the symbols. In each experiment the values of the postsynaptic current were normalized to the average of those elicited by preceding and following steps to +I0 mV; the step to +I0 mV was repeated several times in each experiment in order to allow correction for homosynaptic de pression during the progress of the experiment.
commands in the shape of sensory neuron APs that had been previously recorded on tape, and the same protocol of application of 5-HT and TEA was repeated. If control of the presynaptic voltage was adequate to block the increase in release caused by TEA-which acts by prolonging triggered APs-then any increase caused by 5-HT would be attributable to factors other than a change in the configuration of the presynaptic depolarization. To determine the magnitude of the effects of TEA alone, a separate set of experiments was done in non-clamped cells with application of 5-HT but without application of TEA. Synaptic augmentation by 5-HT was completely independent of AP prolongation in these experiments: application of 5-HTcaused identical increases in PSCs elicited with voltage commands in the shape of APs and in PSCs elicited with triggered APs (Figure 3). In contrast, the increase in the PSC caused by TEA when triggered APs were evoked was completely blocked
when voltage-clamp commands were used instead (Figure 3). Since these experiments were intended to examine the contribution of AP prolongation to synaptic augmentation, it was also necessary to determine how much prolongation occurred when the neurons were not voltage clamped. Surprisingly, adding 5-HT in these experiments actually caused a slight narrowing of the triggered APs (to 96.5% + 1.0% of the value before 5-HT, measured from the peak of the AP to 50% of the peak amplitude on the falling limb; n = 17). This observation already implies that 5-HT can cause significant synaptic augmentation without an increase in the AP duration. It is therefore not surprising that using a voltage-clamp protocol to block broadening did not change the augmentation. It might nonetheless be possible that with higher concentrations of 5-HT AP broadening does contribute significantly to facilitation. The experiment was therefore repeated with a higher concentration of 5-HT, with which measurable AP prolongation did occur. Increasing the concentration of 5-HT to 5 uM (n = 3) caused an increase of the peak and the area of the PSC to 298% + 35% and 408% f 66% of the control, respectively, but still narrowed the AP by about 5%. The experiment was therefore performed using 5-HT at 50 PM, a concentration that caused broadening of 4.8% + 1.2% (n = 7). The protocol of this set of experiments was the same as that used above, except that the higher concentration of 5-HT was added and that thecontrolwithoutTEAadditionwasomitted. In these experiments,theeffectsof5-HTonthePSCwere15%20% greater with triggered APs than with voltageclamp commands in the shape of APs; however, the difference was not significant (p > .5; Figure 4). As earlier, adding TEA increased the PSC when triggered APs were used but not when voltage-clamp commands were used. As a further internal check on voltage control of release, the voltage clamp was turned off after two stimuli delivered in the presence of 5-HTand TEA, and natural APs were triggered in order to see whether releasing the clamp would cause the PSC to increase as a result of the broadening of the AP caused by the combination of TEA and 5-HT. When the clamp was released and an AP triggered in the continuation of the earlier experiments, the PSC increased (by 17% f 4% for the peak and 38% + 11% for the area; n = 6);intheexperimentsusingtriggeredAPsthroughout, the corresponding PSCs decreased as a result of homosynaptic depression (by 25% + 4% for the peak and 19% + 7% for the area; n = 6; pp& < .OOl, result shows both that the PSC of P area < .005). This the clamped neurons could have been increased by broadening had it been permitted to occur and that the voltage clamp was effective in preventing broadening. The absence of a significant difference between 5-HT-induced augmentation at clamped and nonclamped synapses argues that AP prolongation is not
NelJlDll 162
Figure 3. Augmentation by 0.5 uM 5-HT at Clamped and at Non-Clamped Sensory Neuron Synapses
5HT
TEA
SHT TEA
5HT
(A) In representative records from two experiments, 5-HTat 0.5 uM applied after PSC #!Z augments the PSC (uppermost traces in each set) elicited by triggered APs (top) and by voltage-clamp commands with the waveforms of naturai APs (bottom). In contrast,addingTEAafter PSC#4enhances the PSConly in the non-clamped synapse (top), at which AP prolongation is caused by the TEA. Note that the presynaptic voltageclampcurrent (bottom traces in the CLAMP TEA records) moves inward in response to the TEA, whereas the PSC is not increased because of the prevention of AP prolongain the presynaptic current records; the slow initial rise of Calibration is 5 nA for the PSCs, 25 mV for the voltage
tion. No corrections for leak or capacitive components have been made the command voltage precludes the appearance of a capacitive transient. records, and 25 nA for the presynaptic currents. (B) Summary of the effects of 0.5 uM 5.HT and 10 mM TEA on PSCs at clamped and at non-clamped synapses. Change in the PSC is expressed relative tothe PSC before addition of 5-HTor TEA in the respective experiments, i.e., (PSC posttreatment/PSC pretreatment)-I. Open bars show values for PSCs elicited with triggered APs, hatched bars are for PSCs elicited with voltage-clamp commands, and closed bars are for control PSCs elicited with triggered APs without addition of TEA. Error bars show SEM. The values are 1.24 * 0.19 (n = 15, APs) and 1.24 I 0.19 (n = 9, clamp commands) for PSC peaks in response to 5-HT and 1.63 + 0.26 (APs) and 1.64 f 0.18 (clamp commands) for PSC areas in the same experiments. Responses to TEA were 0.42 + 0.07 (n = 9, APs), -0.04 L 0.11 (n = 9, clamp commands), and -0.03 k 0.16 (n = 8, control APs without TEA) for PSC peaks; and 0.54 + 0.09 (APs), -0.02 f 0.13 (clamp commands), and 0.03 f 0.21 (APs without TEA) for PSC areas. For the TEA responses, synaptic responses to triggered APs with TEA differed from both the clamped synapses (p < .005 for both peaks and areas) and the controls without TEA (p < .02 for peaks and p < .05 for areas). The clamped synapses and the controls were not different from each other (p > .5 for both peaks and areas).
a major contributor to facilitation by5-HT. In addition, the ability of 5-HT to cause greater than 2-fold facilitation at 0.5 uM and 5 PM, at which the AP actually narrows, supports this conclusion. However, with higher concentrations of 5-HT, at which AP broadening does occur, a small contribution of broadening might be difficult to detect simply because of the relatively largevariability in the augmentation responses. The final set of experiments was done in order to achieve an estimate of the enhancement in the PSC that would be expected as a result of the observed AP prolongation caused by 5-HT. The Relation between AP Prolongation and Enhancement of Transmitter Release Is linear with a Slope of Approximately 1 In earlier studies rectangular voltage-clamp depolarizations of varying duration were used to assess the relation between the duration of the presynaptic depolarization and the postsynaptic response (Hochner et al., 1986a, 198613). However, there is no obvious way to extrapolate from the dependence of release on the duration of rectangular steps to the dependence on AP duration, especially if the release caused by the rectangular steps falls in a region that is below what normally occurs with release induced by APs. In particular, it might be expected that in the range of duration that is just above release threshold the slope of the relation might be much steeper than it is at durations that cause more normal release. In addition, without doing detailed simulations, it is impossible to know how the precise configuration of the depolarization will affect release. The best way to examine
the relation between fore, is to use stimuli APs to evoke release 1990).
AP duration and release, therewith the waveforms of natura! (Llinas et al., 1982; Augustine,
3.5 3.0 0. .-c & 5 r 0
2.0 1.5 1.0 0.5
z
0.0 2.5 /
-0.5 1 5HT
TEA
5HT
TEA
Figure 4. Augmentation by 50 uM 5.HT at Clamped Clamped Sensory Neuron Synapses
and at Non-
Change in the PSC is expressed relative to the PSC before addition of 5-HTor TEA in the respective experiments, i.e., (PSC posttreatment/PSC pretreatment)-I. Open bars show values for PSCs elicited with triggered APs, and hatched bars are for PSCs elicited with voltage-clamp commands. Error bars show SEM. Values are 1.94 k 0.48 (n = 7, APs) and 1.63 ?r 0.23 (n = 6, clamp commands) for PSC peaks in response to 50 uM 5-Hi; and 3.24 -i- 0.68 (APs) and 2.82 f 0.38 (clamp commands) for areas in the same experiments (p > .5 for both peaks and areas). In response to subsequent addition of IO mM TEA, values are 0.43 + 0.20 (APs) and -0.14 2 0.04 (clamp commands) for peaks (p < .OS); and 1.23 + 0.26 (APs) and -0.17 + 0.10 (clamp commands) for areas (p < .OOl).
AP Prolongation 163
in Facilitation
Table 1. Increase in AP Duration and PSC Area Caused TEA in the Absence and in the Presence of 5-HT AP Duration (Post-TEA)
PSC Area (Post-TEA)
n
(PreTEA)
(Pm-TEA)
Ratio
0
6 8 9
1.58 1.78 1.62
1.54 1.86 1.38
0.97 1.04 0.85
0.5
9 9
1.74 1.47
1.54 1.29
0.89 0.88
5
3
1.85
1.61
0.87
50
5 6
1.82 1.78
2.33 1.70
1.28 0.96
2.5 5-HT Concentration
(NW
1.0
0.5 0.0
~0.0
0.5
1.0
1.5
2.0
2.5
AP Duration Figure
5. Relation
Between
PSC Area
and AP Duration
in TEA
See text for details of the experiment. APs broadened by TEA were shortened by injection of hyperpolarizing current pulses and then permitted to broaden again in successive stimuli. PSC area as a proportion of the pm-TEA PSC and corrected for homosynaptic depression was plotted against AP duration for 47 measurements at 17 synapses (8 with tail sensory neurons and 9 with siphon sensory neurons). In 9 preparations examined without TEA for purposes of correction for homosynaptic depression, average values for successive PSCs were 0.61,0.60, and 0.51 for the second, third, and fourth PSCs, respectively, relative to the first (synapses with tail sensory neurons). For the synapses with siphon sensory neurons, the corresponding values were taken as 0.50,0.45, and 0.40 based on comparison with the experiments using the tail sensory neurons. The least-squares regression line has a slope of 0.913 with correlation coefficient 0.581 (p < .Ol).
The dependence of synaptic release on presynaptic AP duration was measured in the following manner: the sensory neuron member of a soma-soma couple was fired once with an intracellular depolarizing stimulus; TEA was then added, and the increases in AP duration and in the PSC were measured in response toasecond stimulus. Forthe next stimulus, a hyperpolarizing current step was imposed on the sensory neuron beginning at the point where the AP in TEA diverged from the control AP; this was done in order to shorten the AP, thereby attenuating the PSC (see Gingrich and Byrne, 1985). The final stimuluswasonce again an AP thatwas not shortened by a hyperpolarization. Thus, each experiment gave control values for AP duration and PSC amplitude and area, as well as three additional values for each measure, two with broadened APs in TEA and a third with an artificially shortened AP, also in the presence of TEA. To correct for the decline in successive PSCs caused by homosynaptic depression, measurements were made in a separate set of control synapses without adding TEA or imposing a hyperpolarizing step. The values of the experimental PSCs were normalized to the corresponding control values. The relation between the postsynaptic response and the duration of the AP can be approximated by a straight line with a slope of about 1 (Figure 5; Table 1). Applying this relation to the AP prolongation caused
tion
by 5-HT
alone
leads
accounted
to the
for
conclusion
no more
that
than
a 5%
prolonga-
increase
by
For each set of experiments, the average AP duration and PSC area were expressed as a proportion of the pm-TEA value, and the ratio between the two averages was calculated as a measure of the relation between AP prolongation and PSC increase. PSCs were corrected for homosynaptic depression; correction factors ranged from 1.00 to 1.35 in different sets of experiments. The values were derived from all the data presented in the text and figures, with one additional set of values for TEA alone (n = 6 for 0 uM 5-HT).
in the PSC when 5-HT was applied at 50 PM. The fact that this concentration of 5-HT caused an increase of around 300% in the PSC area means that AP prolongation contributed negligibly to the facilitation in this case. Furthermore, the possibility that AP duration might have had a greater influence on the PSC in the presence of 5-HT is addressed by the data in Table 1, which shows that the slope of the relation between PSC area and AP duration was not changed by 5-HT. It is important to point out that TEA appears to have had no direct effect on synaptic transmission here. That TEA acts only by AP prolongation follows from the fact that preventing prolongation by voltage clamping in the presence of TEA blocked the increase in the PSC(Figure3; Figure4).Theequivalentobservation in the present set of experiments is that the fitted line of Figure 5 passes close to the point (l.O,l.O), i.e., that in the presence of TEA an AP that is of the same duration as the control will elicit a PSC that is also the same as the control. Discussion In contrast to the present findings, several earlier reports concluded that AP prolongation contributes significantly to synaptic augmentation (Hochner et al., 1986a, 1986b; Eliot et al., 1993). There are several possible reasons for this discrepancy. First, as discussed above, the relation between the duration of presynaptic depolarization and the postsynaptic response was examined with rectangular voltage-clamp pulses whose relation to normal APs is unclear and may therefore have led to an overestimate of the steepness of this relation. Second, augmentation in the earlier studies was tested with high concentrations of 5-HT
(IO-100 PM) that, combined with repeated firing and consequent K+ current inactivation, caused much greater AP prolongation than in the present study but a similar increase in the synaptic response. It is not clear whether lower concentrations of 5-HT and lower stimulation frequencies causing far less AP broadening would have given less augmentation. The effect of 5-HT on AP broadening is very sensitive to the frequencyofAPfiring: inasmall numberofexperiments, 50 uM 5-HTcaused prolongation of 2.7% + 1.3% when stimulated at 1 min intervals (n = 3), whereas stimulation at 30 and 15 s intervals caused increases of 14.5% + 4.5% (n = 2) and 11.3% f 2.1% (n = 4), respectively. Finally, the one study in which TEA was used to examine the relation between AP duration and the postsynaptic response in the intact abdominal ganglion (Gingrich and Byrne, 1985) demonstrated a nearlinear relation at shorter durations similar to that observed here. Studies in the squid (Augustine, 1990) and in the crayfish (Delaney et al., 1991) also demonstrate linear or near-linear relations between AP duration and transmitter release, albeit somewhat steeper than that shown here. There is no reason to suspect that the differences between the conclusions of the present study and earlier work is the result of some fundamental difference between the cultures used here and the preparations used previously. At the ultrastructural level, these synapses and their surroundings are indistinguishable from those in the intact nervous system (compare Figure 1, right, with Bailey et al., 1979). In particular, the synapses are found at fine processes of the presynaptic neuron rather than coming directly off the cell body, suggesting that cytoskeletal architecture as well as gradients and accumulation of ions will be similar to those in intact ganglia. Furthermore, the amplitudes of the synaptic responses are comparable to those in earlier studies. They show normal homosynaptic depression, and the facilitation induced by 5-HTboth augmentation and restoration is similar to that observed in other studies (Klein, 1993; Ghirardi et al., 1992; Eliot et al., 1993). It is therefore likely that the conclusions reached here are equally applicable to synapses in the nervous system. The results presented here are consistent with earlier experiments showing that the PSC can be augmented by direct activation of adenylyl cyclase without an increase in AP duration (Klein, 1993). The present results show that augmentation by low concentrations of 5-HT also can occur without AP prolongation. In addition, blocking the AP prolongation caused by a high concentration of 5-HT has no significant effect on augmentation, and mimicking this amount of broadening by other means causes only a minimal increase in the synaptic response. The sum total of these findings argues that AP prolongation contributes minimally to synaptic augmentation by 5-HT. There are two possible mechanisms for the facilitating action of PKA that are consistent with the present
results. The first is that PKA activation leads to direct modulation of the Ca*+ influx through voltageactivated Ca*+ channels. Previous work has demonstrated that S-HT increases one type of Ca2’ current in isolated sensory neurons, butthat this current does not appear to contribute to normal transmitter release at synapses (Edmonds et al., 1990; Braha et al., 1993). On the other hand, a current that does contribute to release does not appear to be modulated by 5-HT (Edmonds et al., 1990). Nonetheless, the possibility that there may exist additional Ca*+ currents which are modulated by PKA and play a role in transmitter releasecannotyet beexcluded, norcan the possibility that modulation of currents in isolated sensory neurons differs from that in neurons which have formed synaptic connections. The second possible mode of action of PKA in augmenting transmitter release is through an effect on components of the release process that come into play subsequent to Ca*+ entry. Earlier work showed that restoration of depressed synaptic transmission by 5-HT is largely independent of AP prolongation (Hochner et al., 1986b). Later studies proposed that synaptic augmentation at rested synapses is mediated by activation of CAMP-dependent protein kinase (Braha et al., 1990; Ghirardi et al., 1992; Klein, 1993), whereas synaptic restoration may be caused by activation of PKC (Braha et al., 1990; Sacktor and Schwartz, 1990; Ghirardi et al., 1992; Sugita et al., 1992). The increase in spontaneous release of transmitter caused by 5-HT can be mimicked by activation of PKC and blocked by an inhibitor of PKC, but corresponding manipulations of the PKA system do not affect spontaneous release (Ghirardi et al., 1992). Since the increase in spontaneous release can be dissociated from Ca*’ influx, it has been proposed that the increase reflects adirect modulation of the secretory apparatus by5-HT and PKC (Dale and Kandel, 1990). The present report is consistent with the idea that activation of PKA may also modulate release directly. However, the difference between the actions of the two kinase systems on spontaneous release suggests that they are involved at different steps in the secretory process. A study of the inhibitory actions of the tetrapeptide FMRFamide in Helisoma neurons concluded that this peptide directly modulates the secretory machinery (Man-Son-Hing et al., 1989). FMRFamide inhibits both spontaneous and evoked release from Aplysia sensory neurons as well (Belardetti et al., 1987; Dale and Kandel, 1990). It thus appears likely that FMRFamide in Aplysia directly inhibits transmitter release also. Since FMRFamide antagonizes the actions of 5-HT in both biochemical and physiological assays (Sweatt et al., 1989; Critz et al., 1991; lchinose and Byrne, 1991; Pieroni and Byrne, 1992), it is probable that it acts at some of the same sites as 5-HT. Molecular loci for the direct modulatory actions of 5-HT and FMRFamide on synaptic transmission have not been identified.An attractivecandidatefor regulation would be a synapsin-like molecule that can con-
AP Prolongation 165
in Facilitation
trol the availability of synaptic vesicles for release as a function of its state of phosphorylation (LlinBs et al., 1991). Ultrastructural work of Bailey and Chen (1988) on sensorimotor synapses of Aplysia demonstrated a local depletion of synaptic vesicles from the vicinity of the presynaptic release sites with short-term homosynaptic depression. This work suggests that the mobilization of vesicles to docking sites at the presynaptic membrane is under the tight control of elements in the nerve terminal,whether cytoplasmic, cytoskeletal, or vesicular in nature; and the later studies therefore suggest that interactions among these molecules may be regulated by PKC in synaptic restoration. The recent identification of some of the molecular components of the secretory apparatus (see Jesse11 and Kandel, 1993, for review) has begun to provide a number of candidates for the modulation of transmitter release that underlies the various forms of short-term synaptic and behavioral plasticity. The actions of CAMP at Aplysia and at crayfish synapses (Dixon and Atwood, 1989; Delaney et al., 1991), as well as on Ca2+ currents and on Ca*+ current-independent modulation of secretion in b cells of the mammalian pancreas (Hughes et al., 1989; WmmB;Iti et al., 1993) and in mammalian lactotrophs (Sikdar et al., 1990), suggest that modulation by protein kinases of processes involved in secretion involves analogous components of the secretory systems of different cell types with physiological consequences ranging from glucose and Ca2+ regulation to the induction of short-term memory processes.
Experimental
Procedures
Tail sensory neurons and LFS motor neurons were dissociated as described (Klein, 1993; a number of siphon sensory neurons were included in the experiment of Figure 5) and then maintained in 10% Aplysia hemolymph for l-2 days until all neuronal processeswere absorbed (see also Haydon, 1988). Individual sensory and motor neurons were then paired and left in 10% hemolymph for 24 hr or less to form synapses. For recording, single pairs were transferred to uncoated plastic culture dishes with recording medium (Leibovitz L15 supplemented with salts; see Klein, 1993), where they adhered tightly to the bottom of the dish without further treatment. Pairs not used immediatelywere kept at 4OC until use. Sensory neurons were impaled with two micropipettes filled with 2 M potassium acetate for intracellular recording or voltage clamping (using an Axoclamp 2A; Axon Instruments) and held at -50 mV in all experiments. IFS neurons were voltage clamped with a single 2 M potassium acetate micropipette at -80 mV with a second Axoclamp 2A in the discontinuous voltage-clamp mode set at a sampling frequency of 6 kHz. All recordingsweredone in standard recording medium (above); no modificationswere made to improvevoltage control. TEA and 5-HT (creatinine sulfate; both from Sigma) were applied directly onto the cells in a volume of 50 ~1 from a hand-held Pipetman; the final concentration was estimated from separate dye dilution experiments to be approximately half the concentration in the added aliquot. When TEA was added after 5-HT application, it wasapplied in asolutioncontainingtheestimatedconcentration of 5-HT already present; in the control experiments of Figure 3, a second application of 5-HT alone in this manner caused no change in the synaptic response, confirming the accuracy of the estimate. Records were stored on videotapewith a PCM recorder and analyzed using the Spike program (Hilal Associates). Stu-
dent’s two-tailed t test was used for all statistical data are expressed as mean + SEM.
comparisons;
Acknowledgments This work was supported by grants from the National Institutes of Health (MH45397) and the Natural Sciences and Engineering Research Councilof Canada(OGP0138426)and bya5loan Fellowship. I would like to thank V. Castellucci, J. Koester, J. H. Schwartz, W. Sossin, and L.-E. Trudeau for their comments on an earlier version of the manuscript and Annie Campbell for the electron microscopy. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC Section 1734 solely to indicate this fact. Received
February
18, 1994; revised
May
4, 1994.
References ammlll, C., Ashcroft, F. M., and Rorsman, P. (1993). Calciumindependent potentiation of insulin release by cyclic AMP in single p-cells. Nature 363, 356-358. Augustine, G. (1990). Regulation squid giant synapse by presynaptic current. J. Physiol. 437, 343-364.
of transmitter release at the delayed rectifier potassium
Bailey, C. H., and Chen, M. (1988). Morphological basis term habituation in Aplysia. J. Neurosci. 8, 2452-2459.
of short-
Bailey, C. H., Thompson, E. B., Castellucci, V. F., and Kandel, E. R. (1979). Ultrastructure of the synapses of sensory neurons that mediate the gill-withdrawal reflex in Aplysia. J. Neurocytol. 8, 415-444. Belardetti, F., Kandel, E. R., and Siegelbaum, 5. A. (1987). Neuronal inhibition by the peptide FMRFamide involves opening of 5 K+ channels. Nature 325, 153-156. Boyle, M. B., Klein, M., Smith, 5. J., and Kandel, E. R. (1984). Serotonin increases intracellular Ca2+ transients in voltageclamped sensoryneuronsof Aplysiacalifornica. Proc. Natl.Acad. Sci. USA 87, 7642-7646. Braha, O., Dale, N., Hochner, B., Klein, M., Abrams, T. W., and Kandel, E. R. (1990). 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. Braha, O., Edmonds, B., Sacktor, T., (1993). The contributions of protein C to the actions of 5-HT on the L-type neurons in Aplysia. J. Neurosci. 73, Byrne, J. H. (1987). Cellular iol. Rev. 67, 329-439.
analysis
Kandel, E. R., and Klein, M. kinase A and protein kinase CaZ+ current of the sensory 1839-1851.
of associative
learning.
Phys-
Critz, 5. D., Baxter, D. A., and Byrne, J. H. (1991). Modulatory effects of serotonin, FMRFamide, and myomodulin on the duration of action potentials, excitability, and membrane currents in tail sensory neurons of Aplysia. J. Neurophysiol. 66, 19121926. Dale, N., and Kandel, E. R. (1990). Facilitatoryand inhibitorytransmitters modulate spontaneous transmitter release at cultured Aplysia sensorimotor synapses. J. Physiol. 427, 203-222. Delaney, K., Tank, D. W., and Zucker, R. 5. (1991). Presynaptic calcium and serotonin-mediated enhancement of transmitter release at crayfish neuromuscular junction. J. Neurosci. 77,26312643. Dixon, D., and Atwood, H. L. (1989). Conjoint action of phosphatidylinositol and adenylate cyclase systems in serotonin-induced facilitation at the crayfish neuromuscular junction. 1. Neurophysiol. 62, 1251-1259. Edmonds, B., Klein, M., Dale, N., and Kandel, E. R. (1990). Contributions of two types of calcium channels to synaptic transmission and plasticity. Science 2.50, 1142-1147.
Neuron 166
Eliot, L. S., Kandel, E. R., Siegelbaum, (1993). Imaging terminals of Aplysia strates role of enhanced Ca*+ influx Nature 367, 634-637.
S. A., and Blumenfeid, H. sensory neurons demonin presynaptic facilitation.
Ghirardi, M., Braha, O., Hochner, B., Montarolo, P. G., Kandel, E. R., and Dale, N. (1992). Roles of PKA and PKC in facilitation of evoked and spontaneous transmitter release at depressed and nondepressed synapses in Aplysia sensory neurons. Neuron 9, 479489. Gingrich, K. J., and Byrne, J. H. (1985). Simulation of synaptic depression, posttetanic potentiation, and presynaptic facilitation of synaptic potentials from sensory neurons mediating the gill-withdrawal reflex in Aplysia. J. Neurophysiol. 53, 652-669. Haydon, P. C. (1988). The formation tween cell-cultured neuronal somata.
of chemical J. Neurosci.
synapses be8, 1032-1038.
Hochner, B., Klein, M., Schacher, S., and Kandel, E. R. (1986a). Action potential duration and the modulation of transmitter release from the sensory neurons of Aplysia in presynaptic facilitation and behavioral sensitization. Proc. Natl. Acad. Sci. USA 83, 8410-8414. Hochner, B., Klein, M., Schacher, S., and Kandel, E. R. (1986bj. Additional component in thecellular mechanism of presynaptic facilitation contributes to behavioral dishabituation in Aplysia. Proc. Natl. Acad. Sci. USA 83, 8794-8798. Hughes, S. J., Chalk, J. G., and Ashcroft, S. J. H. (1989). Effect of secretagogues on cytosolic free Ca2+ and insulin release at different extracellular Ca2+ concentrations in the hamster clonal B-cell line HIT-T15. Mol. Cell. Endocrinol. 65, 35-41. Ichinose, M., and Byrne, J. H. (1991). Role of protein phosphatases in the modulation of neuronal membrane currents. Brain Res. 549, 146-150. Jessell, T. M., and Kandel, E. R. (1993). bidirectional and self-modifiable form tion. Cell 72/Neuron 70 (SuppI.), I-30. Kandel, learning: 443.
Synaptic transmission: of cell-cell communica-
a
E. R., and Schwartz, J. H. (1982). Molecular biology of modulation of transmitter release. Science 278, 433-
Klein, M. (1993). Differential cyclic AMP dependence of facilitation at Aplysia sensorimotor synapses as a function of prior stimulation: augmentation versus restoration of transmitter release. J. Neurosci. 73, 3793-3801. Klein, M., Shapiro, and the modulation
E., and Kandel, E. R. (1980). Synaptic plasticity of the Ca*+ current. J. Exp. Biol. 89,117-157.
Llinas, R., Sugimori, M., and by spike-like depolarizations Acad. Sci. USA 79, 2415-2419.
Simon, in the
S. M. (1982). Transmission squid synapse. Proc. Natl.
Llinls, R., Cruner, J. A., Sugimori, M., McCuinness, T. L., and Greengard, P. (1991). Regulation by synapsin I and Ca2+calmodulin-dependent protein kinase Ii of transmitter release in squid giant synapse. 1. Physiol. 436, 257-282. Man-Son-Hing, H., Zoran, M. J., Lukowiak, K., and Haydon, P. G. (1989). A neuromodulator of synaptic transmission acts on the secretory apparatus as well as on ion channels. Nature 347, 237-239. Pieroni, J. P., and Byrne, J. H. (1992). Differential effects of serotonin, FMRFamide, and small cardioactive peptide on multiple, distributed processes modulating sensorimotor synaptic transmission in Aplysia. J. Neurosci. 12, 2633-2647. Pinsker, H., Kupfermann, I., Castellucci, E. R. (1970). Habituation and dishabituation reflex in Aplysia. Science 767, 1740-1742.
V. F., and Kandel, of the gill-withdrawal
Sacktor,T.C.,andSchwartz,J. H.(1990).Sensitizingstimulicause translocation of protein kinase C in Aplysia sensory neurons, Proc. Natl. Acad. Sci. USA 87, 2036-2039. Sikdar, S. K., Zorec, facilitates Ca-induced Lett. 273, 150-154.
R., and Mason, W. T. (1990). CAMP exocytosis in bovine lactotrophs.
directly FEBS
Sossin, W. S., and Schwartz, J. H. (1992). Selective activation Ca*‘-activated PKCs in Aplysia neurons by 5.HT. J. Neurosci. 1160-1168.
of 72,
Sugita, S., Goldsmith, 1. R., Baxter, D. Involvement of protein kinase C in broadening and synaptic facilitation tions of Aplysia. 1. Neurophysiol. 68,
A., and Byrne, J. H. (19923. serotonininduced spike in sensorimotor connec643-651.
Sweatt, J. D., Volterra, A., Edmonds, B., Karl, K. A., Siegelbaum, S. A., and Kandel, E. R. (1989). FMRFamide reverses protein phosphorylation produced by5-HTand CAMP in Aplysia sensory neurons. Nature 342, 275-278.