Spontaneous transmitter release is critical for the induction of long-term and intermediate-term facilitation in Aplysia

Abstract

Long-term plasticity can differ from short-term in recruiting the growth of new synaptic connections, a process that requires the participation of both the presynaptic and postsynaptic components of the synapse. How does information about synaptic plasticity spread from its site of origin to recruit the other component? The answer to this question is not known in most systems. We have investigated the possible role of spontaneous transmitter release as such a transsynaptic signal. Until recently, relatively little has been known about the functions of spontaneous release. In this paper, we report that spontaneous release is critical for the induction of a learning-related form of synaptic plasticity, long-term facilitation in Aplysia. In addition, we have found that this signaling is engaged quite early, during an intermediate-term stage that is the first stage to involve postsynaptic as well as presynaptic molecular mechanisms. In a companion paper, we show that spontaneous release from the presynaptic neuron acts as an orthograde signal to recruit the postsynaptic mechanisms of intermediate-term facilitation and initiates a cascade that can culminate in synaptic growth with additional stimulation during long-term facilitation. Spontaneous release could make a similar contribution to learning-related synaptic plasticity in mammals.

Spontaneous transmitter release was discovered 60 y ago by Fatt and Katz (1), who found that it represents the quantal unit of transmitter release evoked by a presynaptic action potential. However, until recently, relatively little has been known about other possible functions of spontaneous release. In the last few years, we have learned that spontaneous release can contribute to postsynaptic firing (2, 3), regulation of postsynaptic kinase pathways (4), and maintenance of postsynaptic dendritic spines and receptors (5, 6). Spontaneous release has also been found to contribute to some cases of homeostatic scaling of synaptic strength (7–14). Here we report a role of spontaneous transmitter release in the induction of a learning-related form of synaptic plasticity, long-term facilitation produced by serotonin (5HT) in Aplysia.

Long-term plasticity can differ from short-term in recruiting the growth of new synaptic connections, a process that requires the participation of both the presynaptic and postsynaptic components of the synapse (15–21). Because short-term plasticity often involves only one component of the synapse (22–24), the question arises: How does information about synaptic plasticity spread from its site of origin to recruit the other component of the synapse? Studies of synaptic growth during development have revealed a fairly elaborate program of pre- and postsynaptic changes involving a variety of orthograde and retrograde messengers (25), including activity-dependent or spontaneous release of the transmitter itself (26–30). We have now investigated the possible role of spontaneous transmitter release from the presynaptic neuron as an orthograde signal for recruiting the postsynaptic mechanisms of long-term plasticity.

In this paper, we report that spontaneous release is a critical signal for the induction of long-term facilitation. In addition, we find that this signaling is engaged quite early, during an intermediate-term stage that is the first to involve postsynaptic as well as presynaptic molecular mechanisms (22, 23). In a companion paper in this issue of PNAS (31), we show that spontaneous release acts as an orthograde signal to recruit the postsynaptic mechanisms of intermediate-term facilitation, and may thereby initiate a cascade that can culminate in synaptic growth with additional 5HT exposure during long-term facilitation. These results thus elucidate a mechanism by which synaptic plasticity can spread from one component of the synapse to recruit the other component during long-term plasticity.

Results

Induction of Long-Term Facilitation Is Accompanied by Increases in the Frequency and Amplitude of Miniature Excitatory Postsynaptic Currents.

To examine the possibility that spontaneous release might have a role in long-term facilitation, we first recorded miniature excitatory postsynaptic currents (mEPSCs) or potentials (mEPSPs) interleaved with long-term facilitation of the evoked EPSP induced by five pulses of serotonin (10 μM) each 5 min in duration with a 15-min interstimulus interval (5× 5-min 5HT) in Aplysia cocultures (Fig. 1 A and B). Because the cultures contained only a single sensory neuron and motor neuron, and the motor neuron does not form autapses (Fig. S1), the mEPSCs came from the same presynaptic neuron as the evoked EPSPs. There was significant facilitation of the evoked EPSP both 0–50 min (F[1,36] = 9.02, P < 0.01 compared with saline control) and 24 h (F = 4.34, P < 0.05) after washout of 5× 5-min 5HT (Fig. 1C1). There was also a substantial increase in the frequency of mEPSCs (F = 8.66, P < 0.01) and a more modest increase in their amplitude (F = 8.62, P < 0.01) 0–50 min after washout of the 5HT (Fig. 1C2). The increases were maintained at a fairly constant level for 50 min after washout (nonsignificant correlation with delay in each case), but were not significant 24 h later. These results demonstrate that the induction of long-term facilitation is accompanied by increases in the frequency and amplitude of mEPSCs.

Fig. 1.

The induction of long-term facilitation by 5× 5-min 5HT is accompanied by increases in the frequency and amplitude of mEPSCs. (A) The experimental protocol for recording mEPSCs interleaved with long-term facilitation of the evoked EPSP (eEPSP). (B) Examples of mEPSCs recorded before, 0–50 min, and 24 h after washout of 5× 5-min 5HT or saline control. The recordings are AC-coupled. (Insets) Representative mEPSCs on a faster timescale to illustrate their shape. (C) Average results from experiments like the ones shown in B. There were significant overall effects of 5HT (n = 11), compared with control (n = 9), for facilitation of the eEPSP (F[1,18] = 8.20, P < 0.01) (C1) and the frequency (F = 3.72, P < 0.05 one-tail) and amplitude (F = 4.96, P < 0.05) of mEPSCs (C2) in two-way ANOVAs with one repeated measure (time). There was also a significant within-group correlation between the absolute frequency of mEPSCs 0–50 min after the 5HT application and facilitation of the eEPSP 24 h later (C3). The data have been normalized to the value on the pretest (Before) in each experiment. The average pretest values were 27.1 mV for the eEPSP, 2.4 min⁻¹ for mEPSC frequency, and 7.5 pA for mEPSC amplitude, and the average delay for mEPSC recording after washout of the 5HT was 20 min. In this and subsequent figures, the error bars indicate SEMs; *P < 0.05, **P < 0.01, ⁺P < 0.05 one-tail vs. control, ^#P < 0.05, ^##P < 0.01 vs. no inhibitor.

We next examined correlations between facilitation of the evoked EPSP and mEPSCs measured in the same experiments. We reasoned that if mEPSCs contribute to the induction of facilitation, the absolute frequency or amplitude of mEPSCs during induction should correlate with the percent increase in the evoked EPSP at a later time. Consistent with this hypothesis, the absolute frequency of mEPSCs 0–50 min after washout of the 5HT correlated with the percent facilitation of the evoked EPSP 24 h later (within-group correlation = 0.50, P < 0.05) (Fig. 1C3). Furthermore, when the correlation with mEPSC frequency was factored out in an analysis of covariance (ANCOVA), there was no longer significant long-term facilitation of the evoked EPSP, suggesting that much of the variance in facilitation might be accounted for by spontaneous release during induction.

We also examined the correlation between the percent increase in mEPSC frequency and the facilitation of the evoked EPSP measured at the same time. There were significant correlations both 0–50 min (0.68, P < 0.01) and 24 h (0.56, P < 0.05) after washout of the 5HT. These results suggest that expression of the increases in mEPSCs and the evoked EPSP may share some common mechanisms, such as an increased number of synapses (15) or probability of release (32).

Induction of Intermediate-Term Facilitation Is also Accompanied by Increases in the Frequency and Amplitude of mEPSCs.

When does spontaneous release begin to contribute to facilitation? To address this question, we next recorded spontaneous mEPSCs or mEPSPs interleaved with intermediate-term facilitation of the evoked EPSP induced by 10-min 5HT (20 μM) (Fig. 2 A and B). We chose that paradigm because it has been used most extensively to study postsynaptic mechanisms of facilitation in Aplysia (23, 33, 34), and we wished to investigate how those mechanisms are recruited (31). The results were generally similar to those for long-term facilitation, except that there was a larger decrease in the test-alone control EPSPs due to homosynaptic depression, which is quite reliable at these synapses at stimulation intervals of 10 min or less. There was significant facilitation of the evoked EPSP both during (F[1,23] = 28.72, P < 0.01 compared with saline control) and after washout (F = 3.08, P < 0.05, one-tail test) of the 5HT (Fig. 2C1). Consistent with previous studies (35, 36), there was also a substantial increase in the frequency of mEPSCs during the 5HT application (F = 21.12, P < 0.01), which was then maintained at a lower level after washout (F = 5.42, P < 0.05) (Fig. 2C2). In addition, 10-min 5HT produced a modest increase in the amplitude of mEPSCs during the 5HT application (F = 3.58, P < 0.05, one-tail).

Fig. 2.

The induction of intermediate-term facilitation by 10-min 5HT is also accompanied by increases in the frequency and amplitude of mEPSCs. (A) The experimental protocol for recording mEPSCs interleaved with intermediate-term facilitation of the evoked EPSP. (B) Examples of mEPSCs recorded before, during, and after washout of 10-min 5HT or control. (C) Average results from experiments like the ones shown in B. There were significant overall effects of 5HT (n = 17), compared with control (n = 8), for facilitation of the eEPSP (F[1,23] = 29.94, P < 0.01) (C1) and the frequency (F = 15.91, P < 0.01) and amplitude (F = 4.92, P < 0.05) of mEPSCs (C2). There was also a significant within-group correlation between the absolute amplitude of mEPSCs during the 5HT application and the facilitation of the eEPSP after washout of the 5HT (C3). The average pretest values were 22.3 mV for the eEPSP, 8.2 min⁻¹ for mEPSC frequency, and 10.5 pA for mEPSC amplitude, not significantly different between the 5HT and control groups.

The absolute amplitude of mEPSCs during 10-min 5HT correlated with the percent facilitation of the evoked EPSP after washout of the 5HT (within-group correlation = 0.41, P < 0.05) (Fig. 2C3), consistent with the idea that spontaneous release may contribute to the induction of the facilitation. Furthermore, when that correlation was factored out in an ANCOVA, there was no longer significant intermediate-term facilitation of the evoked EPSP, again suggesting that much of the variance in facilitation might be accounted for by spontaneous release during induction.

These results suggest that spontaneous transmitter release begins to contribute to facilitation in 10 min or less, during induction of an intermediate-term stage. Because these early events could be critical first steps in a sequence that can lead to longer-term facilitation (31), we decided to focus our analysis on them initially.

Presynaptic Manipulations That Reduce Spontaneous Transmitter Release also Reduce Intermediate-Term Facilitation of the Evoked EPSP.

To provide a more direct test of the idea that spontaneous release contributes to induction of intermediate-term facilitation, we examined whether manipulations that reduce spontaneous release also reduce facilitation. As a first step, we injected the sensory neuron with the light chain of botulinum toxin D (BoTx D), which interferes with transmitter release by cleaving the vesicle-associated protein VAMP/synaptobrevin (37). Different members of the botulinum toxin family have different effects on spontaneous and evoked release (38), perhaps because the two forms of release involve different Ca²⁺ sensors (39). Compared with other members of the family, BoTx D has a greater effect on spontaneous release. Presynaptic injection of a low concentration of BoTx D (1 μM in the electrode) reduced intermediate-term facilitation of the evoked EPSP by 10-min 5HT (F[1,16] = 43.93, P < 0.01 compared with vehicle), especially after the 5HT, without significantly affecting test-alone homosynaptic depression or reducing the pretest EPSP (Fig. 3A). As expected, presynaptic BoTx D also reduced the frequency of mEPSPs before, during, and after 10-min application of 5HT (F[1,16] = 6.26, P < 0.05 compared with vehicle control overall), with no significant effect on mEPSP amplitude (Fig. 3B 1 and 2). Similarly, presynaptic injection of the slow Ca²⁺ chelator EGTA (100 mM in the electrode), which also reduces spontaneous release (35), reduced facilitation by 10-min 5HT (F[1,14] = 5.10, P < 0.05 compared with vehicle). These results support the idea that spontaneous transmitter release contributes to intermediate-term facilitation of the evoked EPSP.

Fig. 3.

Presynaptic manipulation that reduces spontaneous transmitter release also reduces intermediate-term facilitation of the evoked EPSP. (A1) Injection of BoTx D (n = 5) into the sensory neuron (SN) reduced intermediate-term facilitation of the evoked EPSP, compared with vehicle control (n = 5). BoTx D did not have a significant effect on test-alone depression (n = 5), compared with vehicle (n = 5). There was a significant BoTx D–5HT interaction (F[1,16] = 18.85, P < 0.01). (A2) The average pretest value was 10.4 mV, not significantly different between vehicle and BoTx D. (B1) Examples of mEPSPs following injection of BoTx D or vehicle into the sensory neuron. (B2) Average results from experiments like the ones shown in B1. Injection of BoTx D into the sensory neuron (n = 7) reduced the frequency of mEPSPs before, during, and after washout of 10-min 5HT but had no effect on their amplitude, compared with vehicle (n = 11). (B3) Injection of BoTx D into the sensory neuron did not have a significant effect on the percent increase in frequency of mEPSPs during and after washout of 5HT.

Although presynaptic BoTx D reduced the overall frequency of mEPSPs, it did not reduce the percent increase in mEPSP frequency during the 5HT application (Fig. 3B3). These results suggest that the absolute level, rather than the increase, of mEPSPs during induction is critical for subsequent facilitation of the evoked EPSP, in agreement with the correlations shown in Figs. 1C3 and 2C3. In addition, the BoTx D results demonstrate a dissociation between the percent increase in mEPSPs and the facilitation of the evoked EPSP. This result is consistent with the idea that although the increases in mEPSPs and the evoked EPSP may share some mechanisms (such as an increased probability of release), intermediate-term facilitation of the evoked EPSP also involves additional mechanisms (15, 23, 33, 34). According to this idea, presynaptic BoTx D reduces the absolute level of mEPSPs to below some threshold for recruiting those additional mechanisms.

Manipulations That Increase Spontaneous Release Enhance Intermediate-Term Facilitation.

To examine whether the increase in spontaneous transmitter release contributes to the induction of intermediate-term facilitation, we used two manipulations that act more selectively than 5HT does. For example, 5HT can stimulate postsynaptic as well as presynaptic receptors (15, 34). To increase spontaneous release without also directly affecting the postsynaptic neuron, we expressed in the sensory neuron an Aplysia octopamine receptor (OAR). This receptor, which is not normally expressed in sensory neurons, is positively coupled to adenylyl cyclase and production of cAMP. Brief application of octopamine to cocultures with OAR-expressing sensory neurons reproduces many of the cAMP-dependent effects of 5HT (40), which can include an increase in spontaneous release (36). Ten-minute application of octopamine (20 μM) to cocultures with OAR-expressing sensory neurons produced intermediate-term facilitation of the evoked EPSP that was roughly similar in both amplitude and duration to the facilitation by 10-min 5HT (F[1,43] = 7.73, P < 0.01 compared with no OAR expression, and F = 7.40, P < 0.01 compared with no octopamine) (Fig. 4A). Neither OAR expression alone nor octopamine alone had a significant effect compared with the test-alone (depression) control, suggesting that octopamine does not act by stimulating endogenous receptors on the postsynaptic neuron.

Fig. 4.

Manipulations that increase spontaneous transmitter release enhance intermediate-term facilitation of the evoked EPSP. (A) Ten-minute octopamine produced intermediate-term facilitation in cocultures with OAR-expressing sensory neurons (OA, OAR SN, n = 14), whereas cultures with no OAR expression (OA, control, n = 6) or no octopamine (Depression, OAR SN, n = 5) were not significantly different from the test-alone control (Depression, control, n = 22). There was a marginally significant OA–OAR interaction (F[1,43] = 3.87, P < 0.05 one-tail) in a two-way ANOVA. The average pretest value was 9.2 mV, not significantly different between the groups. (B1) Examples of mEPSCs before, during, and after washout of 10-min octopamine in cultures with expression of OA receptors in the sensory neuron or controls with no OAR expression. (B2) Average results from experiments like the ones shown in B1. Ten-minute octopamine produced increases in the frequency and amplitude of mEPSCs in cultures with OAR-expressing sensory neurons (n = 13), compared with controls (n = 9). There were significant overall effects of OAR expression for mEPSC frequency (F[1,20] = 15.13, P < 0.01) and amplitude (F = 4.84, P < 0.05). The average pretest values were 2.6 min⁻¹ for mEPSC frequency and 13.1 pA for mEPSC amplitude, not significantly different between OAR expression and control. (C) Ten-minute N4C mutant α-latrotoxin (n = 6) did not produce facilitation, compared with controls (n = 6), but mutant α-latrotoxin combined with 10 μM 5HT (n = 21) produced greater facilitation than 5HT alone (n = 17). There was a marginally significant 5HT–α-latrotoxin interaction (F = 3.00, P < 0.05 one-tail). The average pretest value was 17 mV, not significantly different between the groups.

Ten-minute application of octopamine also produced a substantial increase in the frequency of spontaneous mEPSCs in cocultures with OAR-expressing sensory neurons (F[1,20] = 20.29, P < 0.01 compared with no OAR expression), and this increase was maintained at a lower level after washout of the octopamine (F = 5.43, P < 0.05) (Fig. 4B). We obtained similar results in the presence of tetrodotoxin (TTX; 50 μM), which blocks action potentials, suggesting that the increase in frequency of mEPSCs was not due to spontaneous firing of the sensory neuron (816% of pretest during the octopamine, F[1,28] = 21.69, P < 0.01 compared with no OAR expression). Ten-minute octopamine also produced a more modest increase in the amplitude of mEPSCs during the octopamine application (F = 7.86, P < 0.01). As controls, expression of OAR in the sensory neuron did not have a significant effect on the frequency or amplitude of mEPSCs before application of octopamine. Collectively, these results suggest that intermediate-term facilitation can be initiated presynaptically and may be expressed both pre- and postsynaptically in 10 min or less, and that spontaneous transmitter release contributes to induction of the facilitation.

To examine the role of spontaneous release in another way, we used α-latrotoxin (LaTx), which stimulates the release of docked vesicles from presynaptic terminals (41) and produces a substantial increase in the frequency of spontaneous mEPSCs with no increase in mEPSC amplitude (Fig. S2B). Because α-latrotoxin can also form Ca²⁺ pores, we used a mutant LaTx that does not form pores (LaTx^N4C) (42). Ten-minute application of LaTx^N4C (20 nM) did not by itself produce facilitation of the evoked EPSP (not significant, compared with control) (Fig. 4C). However, when we applied LaTx^N4C together with a low concentration of 5HT (10 μM), which by itself produces modest facilitation (F[1,46] = 3.89, P < 0.05 one-tail compared with control), the combination produced significantly greater facilitation than the 5HT alone (F = 9.74, P < 0.01). We obtained similar results with wild-type α-latrotoxin (1 nM) (Fig. S2A).

Together with the results described above, these findings suggest that mechanisms recruited by spontaneous release are necessary for intermediate-term facilitation and act synergistically with additional mechanisms (activated, for example, by presynaptic cAMP) to produce the facilitation.

Spontaneous Release Is also Required for and Contributes to Induction of Long-Term Facilitation.

We next examined whether spontaneous release plays a similar role in long-term facilitation. Like intermediate-term facilitation, presynaptic injection of a low concentration of BoTx D (1 μM in the electrode) reduced long-term facilitation of the evoked EPSP by 5× 5-min 5HT (F[1,32] = 10.01, P < 0.01 compared with vehicle) without significantly affecting the test-alone control or reducing the pretest EPSP (Fig. 5A). Conversely, 5× 5-min application of octopamine (20 μM) to cocultures with OAR-expressing sensory neurons produced long-term facilitation of the evoked EPSP, which was roughly similar to the facilitation by 5× 5-min 5HT (F[1,21] = 14.28, P < 0.01 compared with no OAR expression, and F = 10.39, P < 0.01 compared with no octopamine) (Fig. 5B). As controls, neither OAR expression alone nor octopamine alone had a significant effect compared with the test-alone control. These results suggest that spontaneous release is necessary for long-term as well as intermediate-term facilitation, and that mechanisms recruited by spontaneous release contribute to the induction of both forms of facilitation.

Fig. 5.

Spontaneous transmitter release is also required for and contributes to the induction of long-term facilitation of the evoked EPSP. (A1) Injection of BoTx D (n = 10) into the sensory neuron reduced long-term facilitation by 5× 5-min 5HT, compared with vehicle control (n = 13). BoTx D did not have a significant effect on the test-alone control (n = 7), compared with vehicle (n = 6). There was a significant BoTx D–5HT interaction (F[1,32] = 4.32, P < 0.05) in a two-way ANOVA. (A2) The average pretest value was 27 mV, not significantly different between vehicle and BoTx D. (B) Five pulses of 5-min octopamine produced long-term facilitation in cocultures with OAR-expressing sensory neurons (OA, OAR SN, n = 9), whereas cultures with no OAR expression (OA, control, n = 5) or no octopamine (Control, OAR SN, n = 6) were not significantly different from the test-alone control (Control, Control, n = 5). There was a significant OA–OAR interaction (F[1,21] = 5.49, P < 0.05). The average pretest value was 18 mV, not significantly different between the groups.

Discussion

Our results suggest that spontaneous transmitter release is critical for the induction of two learning-related forms of synaptic plasticity in Aplysia, long-term and intermediate-term facilitation. In a companion paper (31), we have investigated mechanisms by which spontaneous release contributes to those forms of plasticity, and find that it recruits postsynaptic mechanisms of intermediate-term facilitation including increased IP3, Ca²⁺, and membrane insertion and recruitment of clusters of AMPA-like receptors, which may be first steps in synapse assembly during long-term facilitation. Spontaneous transmitter release thus acts as an orthograde signal that conveys information from one component of the synapse to the other component during the induction of long-term plasticity.

Recent studies have shown that spontaneous release also contributes to several other synaptic effects, including some cases of homeostatic scaling (7–14). Scaling in those cases has additional similarities to intermediate- and long-term facilitation: Both types of plasticity can involve postsynaptic Ca²⁺ (8, 23, 33), protein synthesis (8, 23, 34), AMPA receptor insertion (7, 8, 16, 33), and modulation of the presynaptic probability of release (10–12, 15). However, scaling has an opposite sign of action compared with facilitation (during scaling, spontaneous release acts to decrease protein synthesis and synaptic strength). One possible explanation is that a brief, large increase in spontaneous release and consequently postsynaptic Ca²⁺ produces facilitation, whereas a long, small increase produces scaling, similar to the patterns that produce long-term potentiation (LTP) and long-term depression (43). Consistent with this idea, reducing background spontaneous release with botulinum toxin tended to increase the pretest EPSP (Figs. 3A2 and 5A2).

Although spontaneous release is enhanced by activation of presynaptic 5HT receptors during learning-related synaptic plasticity in Aplysia, different types of presynaptic receptors could play an analogous role in mammals. Stimulation of presynaptic dopaminergic and nicotinic receptors in hippocampus, prefrontal cortex, and ventral tegmental area enhances spontaneous release (3, 44–48) and contributes to synaptic plasticity including both early- and late-phase LTP (15, 49–55) (Fig. 6). Thus, in addition to contributing to learning-related synaptic plasticity in Aplysia, spontaneous release may play a similar role in synaptic plasticity in mammalian brain regions involved in learning, memory, and reward, and could be involved in disorders that affect plasticity in those regions including Alzheimer’s disease (56–58), schizophrenia (59, 60), and addiction (44, 61, 62).

Fig. 6.

The role of spontaneous release in cellular and molecular mechanisms of intermediate- and long-term facilitation in Aplysia (Top), compared with its hypothesized role in learning-related synaptic plasticity in hippocampus and prefrontal cortex (PFC) (Middle) and ventral tegmental area (VTA) (Bottom). 5HT, serotonin (green); AC, adenylyl cyclase; ACh, acetylcholine (yellow); D1R, D1 dopamine receptor; DA, dopamine (green); Glu, glutamate (red); nAR, nicotinic acetylcholine receptor.

Materials and Methods

Cell culture and electrophysiological methods were generally the same as described previously (23, 63–65) (SI Materials and Methods). Cocultures consisting of an L7 gill motor neuron and one or two pleural sensory neurons were used 4–6 d after plating. In experiments on spontaneous release, we plated a single sensory neuron and a motor neuron. The DNA construct for OAR was obtained from B. K. Kaang (Seoul National University, Seoul, Korea) and cloned into the Aplysia expression vector pNEX3 (66). Purified plasmid DNA was microinjected into the sensory neuron, which was examined 1 d later to confirm expression of the fluorescent protein. Drugs were added to the perfusion 30 min before the start of an experiment and were present for the remainder of experiments with 10-min 5HT, or until washout of 5× 5-min 5HT. Drug and control treatments were always compared in the same batch, usually on the same day. 5HT, octopamine, TTX, 6,7-dinitroquinoxaline-2,3(1H,4H)-dione (DNQX) (Sigma), wild-type α-latrotoxin (Alomone Labs), and the N4C mutant (a gift from Thomas Sudhof, Stanford University, Stanford, CA) were prepared as stock solutions in water and diluted in artificial seawater before use. In experiments with wild-type α-latrotoxin and their controls, LaCl₃ (20 μM) was added to attempt to block α-latrotoxin pores (67). In some experiments, inhibitors were injected intracellularly into the sensory neuron 30 min before the start of an experiment, and were always compared with vehicle injections. EGTA (Calbiochem) and the light chain of botulinum toxin D (List Biological Laboratories) were diluted in the vehicle solution (0.5 M KAc with 10 mM Tris⋅HCl, pH 7.4, and 0.2% Fast Green to visualize the injection).

The data were normalized to the pretest value in each experiment and analyzed with a two-way ANOVA with one repeated measure (time), followed by planned comparisons of the experimental treatments. The overall ANOVA results are reported in the figure legends, and planned comparisons of interest are reported in Results. Data on mEPSC or mEPSP frequency (and EPSP amplitude when facilitation in individual experiments was greater than 1,000%) were log-transformed to make them more normally distributed. Control groups from experiments in the same figure were pooled if they were not significantly different from each other.