Ca2+ Entry via postsynaptic voltage-sensitive Ca2+ channels can transiently potentiate excitatory synaptic transmission in the hippocampus

Neuron,

Vol. 9,1175-1183,

December,

1992, Copyright

0 1992 by Cell Press

Ca2+Entry via Postsynaptic VokqySensitive Ca2+ Channels Can Transiently Potentiate Excitatory Synaptic Transmission in the Hippocampus Dimitri M. Kullmann,* David j. Perkei,* Toshiya Manabe,* and Roger A. Nicoll Departments of Pharmacology and Physiology University of California, San Francisco San Francisco, California 941438450

Summary We have studied the role of Ca*+ entry via voltagesensitive Ca*+ channels in long-term potentiation (LTP) in the CA1 region of the hippocampus. Repeated depolarizing pulses, in the presence of the NMDA receptor antagonist D-APV and without synaptic stimulation, resulted in a potentiation of excitatory postsynaptic potentials (EPSPs) or currents (EPSCs). This depolarization-induced potentiation was augmented in raised extracellular Ca*+and was blocked by intracellular BAPTA, a Ca*+ chelator, or by nifedipine, a Ca*+ channel antagonist, indicating that the effect was mediated by Ca*+ entry via voltage-sensitive Ca*+ channels. Although the peak potentiation could be as large as 3-fold, the EPSPfOs decayed back to baseline values within approximately 30 min. However, synaptic activation paired with depolarizing pulses in the presence of t%APV converted the transient potentiation into a sustained form. These results indicate that a rise in postsynaptic Ca*+ via voltage-sensitive Caz+ channels can transiently potentiate synaptic transmission, but that another factor associated with synaptic transmission may be required for LTP.

entering via NMDA receptors may be segregated from that entering via voltage-sensitive Ca2+ channels, so that Ca*+ entering via voltage-sensitive CaH channels may not have access to the subsynaptic region (Wigstram and Gustafsson, 1985; Nicoll et al., 1988; Madison et al., 1991). Data supporting this suggestion have recently been presented (Guthrie et al., 1991; Muller and Connor, 1991). However, recent experiments have also suggested that under certain conditions, Ca2+ entry via voltage-sensitive Ca2+ channels can indeed reach levelssufficienttotrigger LTP. Specifically, it has been reported that tetanic stimulation of presynaptic fibers at 200 Hz, which is a higher frequency than that normally used to induce LTP, can evoke a Ca2+-dependent, NMDA receptor-independent potentiation (Grover and Teyler, 1990). A similar potentiation can be induced by blockade of K+ channels with a brief application of tetraethylammonium (Aniksztejn and Ben-Ari, 1991). However, in both of these studies synaptic stimulation accompanied the conditioning procedure, and thus the potentiation observed may have resulted from effects of synaptic activation in addition to the rise in postsynaptic Ca2+. We have found conditions in which activation of voltage-sensitive Ca2+ channels in the absence of synaptic activity results in a potentiation of excitatory postsynaptic potentials or currents (EPSP(C)s). We have explored the properties of this potentiation and conclude that, while Ca*+ alone can potentiate excitatory synaptic transmission, synaptic activity in addition to Ca*+ may be required to obtain stable longlasting potentiation.

Introduction Results Long-term potentiation (LTP) in the CA1 region of the hippocampus is defined as a stable potentiation of excitatory synaptic responses induced bycoactivation of both presynaptic and postsynaptic elements. There is general agreement that the induction of LTPof excitatory synaptic transmission in the CA1 region of the hippocampus requires the activation of the N-methylo-aspartate (NMDA) subtype of glutamate receptor and that Cap, most likely entering via the NMDA receptor, serves as the trigger for this potentiation (Bliss and Lynch, 1988; Collingridge and Singer, 1990; Gustafsson and Wigstrom, 1990; Madison et al., 1991). Photorelease data show directly a Ca2+-triggered enhancement (Malenka et al., 1988). It is also well established that dendritic depolarization results in a large, widespread influx of CaZc via voltage-sensitive Ca*+ channels (Regehr et al., 1989; Jaffe et al., 1992). To reconcile thisfindingwiththe known specificityof LTPtostimulated synapses, it has been postulated that the Ca2+ *This paper authors.

represents

an equal contribution

by the first three

Repeated Postsynaptic Depolarization in the Absence of Synaptic Stimulation Can Potentiate Synaptic Transmission To prevent Ca*+ entry via NMDA receptors, all experiments, unless otherwise stated, were carried out in the presence of 25-50 PM o-2-amino-S-phosphonovaleric acid (o-APV). Under these conditions, a steady depolarization (to 0 mV, 100 s duration; 2.5 mM extracellular Ca2+) of CA1 pyramidal cells in the absence of synaptic stimulation had little or no effect on excitatory synaptic transmission as measured with 0.1 Hz stimulation of Schaffer collateral-commissural afferents in the stratum radiatum (Figures IA and IB,) (see Malenka et al., 1989). However, when the steadydepolarization was replaced with repeated depolarizing pulses (to 0 mV, 3 s duration, 0.2 Hz, 20 pulses), a clear potentiation of synaptic transmission generally occurred. A comparison in the same cell of the effect of steady depolarization and multiple depolarizing pulses on EPSPs recorded with an intracellular microelectrode is shown in Figure IA. Superficially, the po-

NC3lVXl

1176

Figure 1. Effects of Steady Depolarization and Repeated Depolarizing Pulses on the Size of EPSP(C)s (A) shows an experiment using an intracellular microelectrode in which the cell was first depolarized to +I0 m V and held at this potential for 100 s (Steady depol.), and later depolarizing pulses were applied (Depol. pulses). Traces represent averages of IO15 consecutive EPSPs obtained at the times marked by the small letters. Scale bars: 5 mV, 20 ms. Baseline stimulation frequency was 0.1 Hz. Each point represents the average slope of 3 consecutive EPSPs. D-APV (25 PM) was present throughout the experiment, and unlessotherwisestated, stimulation was stopped during the depolarization. (B) compares results from two groups of cells recorded with whole-ceil techniques. In (B,) the cells were depolarized to 0 m V for 100 s; in (BJ a series of 20 depolarizing voltage pulses to 0 m V were given. The traces, which are averages of 20 records from two representative experiments, show the response to a 5 mV, 30 ms hyper10 , , I 550 polarizing voltage pulse (to monitor series Time (min) resistance and input resistance) followed by an EPSC. Records were obtained at the times marked by the small letters and are superimposed. Scale bars: 100 pA, 20 ms. Note that no change in the recording conditions occurred during the experiment. o-APV (50 PM) was present in both sets of experiments. This and all subsequent figures, unless otherwise indicated, summarize results obtained from all cells whether or not potentiation was seen.

A

tentiation seen with depolarizing pulses would appear to be at variance with the lack of effect of repeated Ca*+ spikes described by Malenka et al. (1989). However, the Ca2+/Mg2+ ratio was higher in the present study, and stronger, longer-lasting depolarizing pulses were used. This protocol would be expected to enhance Ca” influx. The potentiation was also readily evoked with whole-cell recording techniques. FigurelB,which summarizes resultsfrom several cells, shows that following a steady depolarization to0 mV, no potentiation was seen (n = 6), whereas in a separate group of cells following repeated depolarizing pulses, the EPSCs increased in size (n = 8). W ith 2.5 mM extracellular Ca2+, this potentiation was observed in 5 out of 8 cells. This figure also illustrates thetypical timecourseof the potentiation. It required a few minutes to reach a peak following the voltage pulses (see also Figures 2-4). The peak potentiation with whole-cell recording averaged 4 8 % f 1 3 % (n = 8) in 2.5 mM Ca2+ and was not accompanied by any change in input resistance or series resistance. The EPSC amplitude typically decayed to control values in approximately 15-30 min. A transient depression of EPSP(C)s was often seen immediately after the depolarization butwas unrelated to the potentiation, since it could be seen in experiments not showing potentiation (Figure IB; Figure 2B; Figure 4). If the time course of the potentiation was divided by that of the depression observed under conditions that blocked the potentiation, a delay was still observed. Thus, while the origin of this depression is unclear, it cannot fully explain the delay in the potentiation.

Thedepolarization-induced potentiation depended on the experimental conditions. W ith conventional intracellular recording the potentiation could be evokedrepeatedlyovertime(Figure2A),whereaswith whole-cell recording it could be evoked only within approximately 20-30 min of breaking into the wholecell mode (Figure 2B). W h e n the whole-cell pipette solution contained an ATP regenerating system (Forscher and Oxford, 1985), the potentiation was changed in a number of ways. First, the potentiation was more reliable, occurring in over 9 0 % of the cells examined. Second, it was possible to evoke the potentiation at later times (Figure 2C), although eventually (-90 min) the phenomenon could not be evoked. Third, the magnitude of the potentiation was considerably increased and could be as large as 3-fold. Finally, potentiation could be evoked with 2.5 mM extracellular Ca2+ by steady depolarization without the need to apply multiple depolarizing pulses (Figure 2C). The enhancement of this phenomenon by the regenerating system may reflect the known modulation of voltage-sensitive Ca2+channels by Mg2+-nucleotides (Osterrieder et al., 1982; O ’Rourke et al, 1992). Alternatively, intracellular Ca2+ may be more effective at enhancing transmission when the regenerating system is present. Depolarization-Induced Potentiation Is Mediated by Postsynaptic Caz+ W e have carried out experiments to determine whether the observed potentiation results from activation of voltage-sensitive Ca2+ channels. First, we

Depolarizing 1177

A

Pulse-Induced

Potentiation

Microelectrode

250

B

a

20

40

60

80

Time (min) Figure 2. Effects of Recording Induced Potentiation

Conditions

on Depolarization-

(A) shows records obtained with an intracellular microelectrode, indicating that the depolarizing pulse-induced potentiation can be repeated over time. (6) shows records obtained with whole-cell recording and indicates thattheabilityofdepolarizingpulsestoevokethepotentiation is lost with time. (C) shows that with an ATP regenerating system (ATPRS) in the pipette solution the potentiation is larger and can be repeated. In this experiment a steady depolarization was used toelicitthe potentiation.TheextracellularCa2’concentration was 5 m M in (A), 4 m M in (B), and 2.5 m M in (C). Each of the graphs shows the results from an individual experiment.

compared the effect of different concentrations of extracellular Ca*+. With intracellular recording it was possible to make this comparison in the same cell (Figure 3A). Depolarizing current pulses were first applied with 2.5 mM extracellular Ca*+, and in this cell no potentiation was detected. The Ca2+ concentration was then increased to 5 mM and the pulses were repeated, now evoking a clear potentiation. A comparison of the results obtained in 2.5 mM Ca*+ and 5 mM Ca2+ (n = 5) is shown in Figure 38, confirming that the magnitude of the potentiation was dependent on the concentration of extracellular Ca*+. We performed twoadditional experiments todefine the role of Ca*’ in this potentiation. First, we compared the effects of depolarization on EPSCs recorded with pipettes containing 0.2 mM EGTA, which buffers Caz+ weakly (Figure 4, Control) (n = 8), with those obtained when the internal solution contained IO mM BAPTA, which buffers Ca*’ much more strongly (Figure 4, BAPTA) (n = 5). The presence of BAPTA completely blocked the potentiation. These results clearly establish that intracellular Caz+ is essential for the po-

tentiation. A likely mechanism for this Ca*+ involvement is that the depolarization activates voltagesensitive Ca*+ channels. This was confirmed by showing that the potentiation could also be blocked when the experiment was carried out in the presence of the Ca*+ channel antagonist nifedipine (20 uM) (Figure 4, Nifedipine) (n = 4). In a separate set of experiments using intracellular recording, potentiation elicited by depolarizing pulses in the presence of 20 uM nifedipine averaged 25% f 12% (n = IO), as compared with 59% f 10% in control conditions (n = 8). In 3 of these cases depolarizing pulses were applied before and after theapplication of nifedipine, allowing acomparison within cells. Do Voltage-Sensitive Caz+ Channels Contribute to LTP? Since nifedipine has a profound effect on depolarization-induced potentiation, this raises the question whether Ca*+entryvia voltage-sensitive Cati channels plays a role in NMDA receptor-dependent LTP. When the ATP regenerating system was used, pairing synaptic activity with steady depolarization in the absence of o-APV resulted in a large potentiation (Figure 5A) (n = 8). When the same experiment was performed in the presence of 20 uM nifedipine (n = 4), the transient component of the potentiation was absent, but there was no effect on the potentiation observed at 20 min. Thus, under these conditions, which considerablyenhance potentiation due to Ca*+ influx, dihydropyridinesensitiveCaechannelsdocontributetotheearly transient potentiation evoked by pairing, but not to that seen at later times. Is this also the case with LTP generated using more conventional stimulus patterns? Extracellular field EPSPs were recorded in response to stimulation of two independent pathways. LTP was elicited in one pathway in the absence of D-APV with tetanic stimulation (100 Hz, 1 s, repeated 4 times). After LTP had stabilized, 20 uM nifedipine was applied for 20 min, and the second pathway was then tetanized (n = 6). In Figure 5B the LTP evoked in the control pathway is superimposed on the LTP evoked in the presence of nifedipine, showing that it had no detectable effect on LTP. Nifedipine also had no effect on baseline synaptic transmission. These results suggest that Ca*+ entry via voltage-sensitive Ca*+ channels does not contribute to LTP induced by conventional methods. Does Depolarization-Induced Potentiation Interact with NMDA Receptor-Dependent LTP? One means of addressing whether depolarizationinduced potentiation shares mechanisms with LTP is to test for occlusion between the two phenomena. We therefore determined whether the depolarization-induced potentiation was reduced by prior induction of NMDA receptor-dependent LTP. Two pathways were monitored in this experiment, and LTP was induced in one pathway by pairing synaptic stimulation (2 Hz) with steady membrane depolarization in

NC3lV3l 1178

. . . ..f**

.'. .

..*. .-v. . .a.

;*

Figure 3. Elevating Extracellular hances the Depolarization-Induced tiation

5 m M calcium

2.5 m M calcium

*

:i

l *.**.’ -0

.*

.*

. . .

l .

.

l

.

.

- l .w%*‘+-ycr .-

l

0.

l

*.

Cali EnPoten-

(A) shows an example of the effect of elevating extracellular Ca*+ on the potentiation induced by depolarizing pulses. In this case, voltage pulses, recorded with an intracellular microelectrode, had little effect when given in the presence of 2.5 m M Ca*+, but when repeated in 5 m M Cal+ caused a potentiation. (B) compares the depolarization-induced potentiation evoked in 5 cells recorded with intracellular microelectrodes. (B,) shows potentiation evoked in the presence of 2.5 m M extracellular Ca’+. (B,) shows potentiation in the same cells in the presence of 5 m M extracellular CaZ+.

Time (min) the absence of D-APV. An ATP regenerating system was included in the pipette solution to maximize subsequent depolarization-induced potentiation. After a period of about 20-25 min, depolarizing pulses were applied, and the potentiation in the control pathway

A

500 -

z

400 -

Control 0 Nifedipine l

% .g 5.

300-

I

200-

Time (min) Control 0 Nifedipine

l

t

91

0

15

30

45

60

Time (min) Figure 5. The Contribution of Voltage-Sensitive to NMDA Receptor-Dependent LTP

Figure 4. Depolarization-Induced tracellular BAPTA or Extracellular

Potentiation Nifedipine

Is Blocked

by In-

The upper panel (Control) shows the effect of depolarization on EPSCs in control conditions (n = 8). The middle panel (BAPTA) shows results from another series of cells recorded with patch pipettes containing 10 m M BAPTA (n = 5). The lower panel (Nifedipine) shows results from another series of cells bathed in 20 uM nifedipine (n = 4). In these experiments an ATP regenerating system was included in the whole-cell pipette solution and steady depolarization (60 s duration, 0 mV) was used rather than pulses. o-APV was not present in these experiments.

Ca*+ Channels

(A) EPSCs were recorded with whole-cell pipettes containing an ATP regenerating system. o-APV was omitted in this series of experiments. LTPwas induced by delivering synaptic stimulation at 2 Hz during a 60 s depolarization to 0 mV. Under these conditions, a very large transient potentiation was followed by stable potentiation. In the presence of nifedipine (20 PM) the transient component was completely blocked while the sustained potentiation was unaltered. (B) Field EPSPs recorded in the stratum radiatum in response to stimulation of two independent pathways. Comparison of tetanus-induced potentiation in one pathway to subsequent tetanusinduced potentiation evoked in the other pathway in the presence of nifedipine (20 PM).

Depolarizing

Pulse-Induced

Potentiation

1179

200r

0 Paired l

75' -5

0

I I

1 5

10

Control

15

I 20

Time (min) Figure 6. Depolarization-Induced by Previously Established NMDA

Potentiation Is Not Occluded Receptor-Dependent LTP

Two independent pathways were stimulated in the absence of o-APV. In one pathway 2 Hz stimulation was applied for 60 s during a steady depolarization to 0 mV, whereas stimulation was stopped in the other pathway. Between 20 and 25 min after inducing LTP, voltage pulses were applied in the absence of synaptic stimulation, and the potentiation in the two pathways was compared.

was compared with that in the pathway that had undergone LTP (Figure 6). To ease the comparison between the two pathways, the baselines before the pulses were renormalized to 100%. Only those cells that exhibited NMDA receptor-dependent LTP (n = 7 out of IO) were included in the summary (average EPSC amplitude after potentiation was 229% f 18% of control). Figure 6 shows that there was no obvious difference between the control (closed circles) and the potentiated (open circles) pathways, indicating that LTP did not prevent subsequent depolarizationinduced potentiation. While some suggestion of interaction was observed in other experiments at earlier times after LTP induction, the presence of a decaying baseline complicated the analysis. Synaptic Activity Paired with Depolarizing Pulses Evokes a Sustained Potentiation A highly consistent finding was that the depolarization-induced potentiation was transient, even though with high extracellularCa*+and/orwith theATP regenerating system the peak potentiation could be at least 3-fold (see Figure 2C). We therefore inquired whether synaptic activity in the absence of NMDA receptor activation could, when combined with a rise in postsynaptic Ca*+, convert the transient potentiation into sustained LTP. Theeffect of depolarizing pulses alone in one pathway was compared with voltage pulses plus synaptic activation (2 Hz stimulation) in another pathway. o-APV (50 vM) was present throughout to prevent NMDA receptor activation. Conditions were chosen to favor potentiation, e.g., 4 mM extracellular Ca*+ and/or an ATP regenerating system in the pipette solution. Two key questions were specifically addressed with this experiment. First, does the potentiation with depolarizing pulses alone return entirely to

baseline, and second, is the potentiation in the paired pathway clearly different from that in the unpaired pathway? The potentiation of synaptic strength with depolarizing pulses alone decayed completely within 30 min (Figure 7, closed circles). In contrast, when given with continuous 2 Hz synaptic stimulation, the pulses induced a sustained component to the potentiation (Figure 7, open circles). A similar result was obtained even when synaptic stimulation was applied only during the depolarizing pulses to the reversal potential of the EPSP(C)s (about 0 mV) and not between the pulses. This result indicates that the effect of synaptic stimulation cannot be explained by providing additional Ca2+ entry via voltage-sensitive Ca2+ channels. In another series of control experiments (n = 14) we examined the consequence of 2 Hz stimulation in the absenceof membrane depolarization. In this situation a small enhancement (16% f 9%) occurred, with synaptic strength returning to baseline within 15 min. Discussion The precise role of Ca2+ in NMDA receptor-dependent LTP has been a major issue for many years. The blockade of LTP by intracellular Ca*+ chelators (Lynch et al., 1983; Malenka et al., 1988) indicates a need for Ca2+ in the induction of LTP. Two lines of evidence have suggested that a rise in postsynaptic Ca2+ may be sufficient to potentiate excitatory synaptic transmission. First, activation of NMDA receptors by exogenous NMDA, which would be expected to increase the intracellular Ca2+ concentration, can enhance EPSPs (Collingridge et al., 1983; Kauer et al., 1988a).

l

?,I

0

10

Time

20

Unpaired

30

40

(min)

Figure 7. Pairing Synaptic Stimulation with Depolarizing Pulses in the Presence of 50 NM D-APV Produces Sustained Potentiation Two independent pathways were stimulated in the presence of o-APV (n = 27). Depolarizing pulses were delivered to the ceil, during which stimulation of one pathway was interrupted (closed circles) and the other pathway was stimulated 120 or 200 times at2 Hz (open circles). Idehtical results wereobtained using intracellular microelectrodes (n = 13) and whole-cell pipettes (n = I+, and the results from these two groups have been averaged together.

NWRXl 1180

Second, the intracellular release of caged Ca2+ from Nitr 5 enhances EPSPs (Malenka et al., 1988). Application of NMDA typically produces a transient potentiation (Collingridge et al., 1983; Kauer et al., 1988a), although under certain conditions, e.g., raised extracellular Ca*+, the potentiation can be longer lasting (Thibault et al., 1989; Malenka, 1991; Manabe et al., 1992). The potentiation of transmission following the release of Ca*+from Nitr 5 can be relatively long lasting (Malenka et al., 1988), but synaptic stimulation accompanied the rise in Caz+. Therefore, whether a rise in Ca*+ is entirely sufficient to induce LTP is currently debatable. It has recently been reported that in the absence of NMDA receptor activation, very high frequency tetanization (Grover and Teyler, 1990) or a brief application of tetraethylammonium (Aniksztejn and Ben-Ari, 1991) can potentiate EPSPs and that this potentiation appears to depend on the activation of voltagesensitive Ca*+ channels. This potentiation was sustained, but since synaptic stimulation accompanied the inducing stimuli, factors other than Ca*+ could have contributed to these phenomena. We have found that, when repeated depolarizing pulses are given in the absence of synaptic stimulation, potentiation of EPSP(C)s occurs. This potentiation is most likely due to Caz+ entry via voltagesensitive Ca*+ channels since it is dependent on the concentration of extracellular Ca*+, blocked by intracellular Ca2+ chelators, and blocked by nifedipine. Given that Ca*+ entry via voltage-sensitive Ca*+ channels can potentiate EPSP(C)s, we asked three related questions. First, does CaH entry via this route normally contribute to NMDA receptor-dependent LTP? Second, does depolarization-induced potentiation share a mechanism with NMDA receptor-dependent LTP? Third, is a rise in postsynaptic Ca*+ sufficient to generate sustained potentiation? Do VoltageSensitive Caz+ Channels Normally Contribute to NMDA Receptor-Dependent LTP? If synaptic stimulation was delivered coincident with depolarization applied via an electrode containing an ATP regenerating system when NMDA receptorswere functional, a large transient potentiation was seen followed by a sustained potentiation. Nifedipine entirely blocked the transient potentiation in the control pathway (data not shown) and an early transient component, but not the sustained component, in the paired pathway (compare Figure 4, Nifedipine, and Figure 5A). Although voltage-sensitive Ca*+ channels can contribute to the early potentiation evoked under these conditions, we could find no evidence that these channels participate in LTP evoked by more standard procedures. Specifically, LTP evoked in field potential recordings by 100 Hz tetanic stimulation was entirely unaltered by nifedipine, in agreement with previous results with Ca*+ channel antagonists (Taube and Schwartzkroin, 1986). This could indicate either that, with 100 Hz stimulation, the Ca*+ which enters

via voltage-sensitive Ca 2+channels fails to gain access to the site for potentiation or that Ca*+ entering via NMDA receptors saturates the mechanism involved in the potentiation. Do Depolarization-Induced Potentiation and NMDA Receptor-Dependent LTP Share a Common Mechanism? Under certain conditions, such as NMDA-induced potentiation (Collingridge et al., 1983; Kauer et al., 1988a; Asztely et al., 1991) or for a period following a tetanus (Gustafsson et al., 1989; Malenka, 1991), a short-term potentiation can be recorded. Whether this early potentiation shares the same exact mechanism with LTP is not entirely clear. We have addressed this question by examining whether NMDA receptor-dependent LTP occludes subsequent depolarization-induced potentiation. While some suggestion of interaction was observed shortly after NMDA receptor-dependent LTP, the presence of a decaying baseline complicated the analysis. When examined at 20-25 min after inducing LTP, there was no obvious difference between the amount of depolarization-induced potentiation in a control pathway and one that had undergone LTP. Assuming that depolarizing pulses activate the same processes as NMDA, the lack of interaction is consistent with earlier workon NMDA-induced potentiation (Kauer et al., 1988a), although more recently an interaction has been reported (Asztely et al., 1991). While mutual occlusion of the two phenomena would suggest a shared mechanism, a lack of interaction is less informative. For instance, if a portion of the synapses in the stimulated pathwayfail to undergo LTP because they lack NMDA receptors (Bekkers and Stevens, 1989) but nonetheless possess the appropriate Ca2+-activated machinery necessary for the potentiation, the depolarization-induced potentiation at these synapses would not show interaction with synaptically induced potentiation. Depending on the relative proportions of these types of synapses, it may be difficult to demonstrate interaction between the two forms of potentiation, despite underlying commonality of mechanism. Is a Rise in Postsynaptic Caz+ Sufficient to Evoke Sustained Potentiation? While the present results confirm that Ca*+ alone is sufficient to potentiate EPSP(C)s, this potentiation declines to baseline values even when every effort is taken to maximize the increase in intracellular Ca*+. Thus, despite peak potentiations of 3-fold under these conditions, the EPSP(C)s still return to control values within approximately 30 min. We have found that, with NMDA receptors blocked, pairing synaptic activation with multiple depolarizing pulses converts the transient potentiation into a sustained form. If the concentration of Ca*+ required to produce sustained potentiation when paired with synaptic activity was less than that required to evoke the transient potentiation on its own, the present findings

Depolarizing 1181

Pulse-Induced

Potentiation

could explain the recent result of Grover and Teyler (1992) that the o-APV-resistant LTP induced by tetanic stimulation in CA1 is pathway specific. In this case the rise in Ca2+ via voltage-sensitive Caz+ channels, which should be widespread, would be below threshold to potentiate EPSPs directly, but when paired with synap tic activation, would allow potentiation to occur. This raises the question why pairing synaptic activation with NMDA application fails to evoke sustained potentiation (Collingridgeet al., 1983; Kaueret al., 1988a). A possible explanation is that high concentrations of NMDAcausea strong presynaptic inhibition of synaptic transmission (Collingridge et al., 1983). What might synaptic activity provide to convert transient into sustained potentiation? The finding that LTP can be generated between pairs of neurons (Malinow, 1991) indicates that activity in these synapses alone is sufficient to induce LTP and that coactivation of other, nonglutamatergic synapses is not required. The candidate mechanisms for inducing sustained potentiation may be listed as follows. First, synaptically released glutamate may be entirely responsible. Residual NMDA receptor activity was ruled out by verifying that 50 uM o-APV, the concentration used in this study, entirely blocked the NMDA receptor-mediated synaptic currents. The effect of synaptic stimulation cannot be mediated by increased Ca*+entrycaused by additional opening of voltage-sensitive Caz+ channels, since the effect could be seen when synaptic stimulation was delivered onlywhen the membrane potential was beyond the EPSP(C) reversal potential. Our inability to evoke clear sustained potentiation with voltage pulses alone, despite efforts to maximize Ca*+ entry, makes it unlikely that synaptic activity exerts its effect by providing extra Ca2+ via either non-NMDA receptors (lino et al., 1991) or metabotropic glutamate receptors. Other possibilities include Na+entryvia nonNMDA receptors (or additionally NMDA receptors under normal conditions) or protein kinase C activation by metabotropic glutamate receptors, analogous to cerebellar long-term depression (Linden and Connor, 1991). A number of studies have implicated the activation of protein kinase C in LTP (Malenka et al., 1986; Anwyl, 1989; Linden and Routtenberg, 1989; Malinow et al., 1989; Wang and Feng, 1992). It has recently been reported that activation of metabotropic glutamate receptors alone can cause a long-lasting potentiation in the dorsolateral septal nucleus (Zheng and Gallagher, 1992) and the hippocampus(Bortolottoand Collingridge, 1992), although previous studies in the hippocampus failed to observe any potentiation (Baskys and Malenka, 1991; McGuinness et al., 1991; Aniksztejn et al., 1992). The inability of glutamate application to evoke long-lasting potentiation (Kauer et al., 1988a) may have resulted from an inability to reach adequate concentrations in the synaptic cleft because of strong uptake mechanisms. A second possibility is that some synaptic process in addition to the release of glutamate is responsible for sustained potentiation. A cotransmitter released

with glutamate from the same synapses is one possibility. For instance, ATP, which is present in high concentration in many types of vesicle (Lagercrantz, 1976; Volknandt and Zimmermann, 1986), might act postsynaptically. In considering a cotransmitter mechanism it is important to realize that it must be released with low frequency stimulation, since pairing low frequency synaptic activation with depolarization evokes sustained LTP (Gustafsson et al., 1987; Kauer et al., 1988b). Alternatively, some aspect of presynaptic activity independent of neurotransmitter release may prime these synapses so that they are receptive to some retrograde factor. In conclusion, we have found that a rise in postsynaptic Ca2+ due to activation of voltage-sensitive Ca2+ channels is sufficient to potentiate synaptic transmission. However, the potentiation differs from LTP in that it is transient and was not in our hands occluded by established LTP. After the blockade of NMDA receptors, presynaptic activation coincident with postsynaptic voltage pulses evokes sustained potentiation, suggesting that both synaptic activity and a rise in postsynaptic Caz+ are necessary for LTP. Such a dual requirement illustrates the need for some coincidence detection mechanism in addition to that provided by the NMDA receptor. Experimental

Procedures

Hippocampal slices (500 urn thick) were cut from 3-to 5-week-old male Hartley guinea pigs and placed in a holding chamber for at least 1 hr. A single slice was then transferred to the recording chamber and held between two nylon nets, submerged beneath a continuously superfusing medium that had been pregassed with 95% Oz and 5% CO* (Nicoll and Alger, 1981). The composition of the medium, unless otherwise stated, was 119 m M NaCI, 2.5 m M KCI, 1.3 m M MgSO,, 2.5 m M CaClz, 1.0 m M NaHzPO+ 26.2 m M NaHCOI, 11 m M glucose. Picrotoxin (100 PM) was present in all experiments. All experiments were done at room temperature. Recordings were made with field electrodes (3 M NaCI), with conventional intracellular electrodes (3 M CsCI), or with the whole-cell patch-clamp technique (Blanton et al., 1989). The whole-cell pipette solution typically contained 122.5 m M cesium gluconate, 17.5 m M CsCI, 10 m M HEPES, 0.2 m M EGTA, 8 m M NaCI, 2 m M Mg-ATP, and 0.3 m M Nap-CTP (pH 7.2, osmolarity 290-300 mOsm). In some experiments 10 m M BAPTA was added to the pipette solution, in which case the cesium gluconate was reduced to maintain the osmolarity constant. In some experiments the pipettes were filled with an ATP regenerating system which was the same as the standard solution with the following exceptions: 100 m M cesium gluconate, 2 m M Naz-ATP, 2 m M Mg-ATP, 20 m M phosphocreatine, and 50 U/ml creatine phosphokinase. Stimuli were delivered through fine, bipolar, stainless steel electrodes placed in the stratum radiatum. In most experiments a stimulating electrode was placed on either side of the recording electrode so that two independent pathways could be stimulated. The membrane potential of the recorded cell was held at -90 mV unless otherwise stated. The voltage pulse protocol was as follows: 20 depolarizing pulses to 0 mV, 3 s duration, 0.2 Hz. For conventional intracellular recording, an Axoclamp 2A amplifier was used. When the cell was depolarized, the discontinuous current-clamp mode was used and the headstage voltage was continuously monitored (switching frequency 0.5-1.5 kHz). For wholecell recording, an Axopatch ID amplifier was used and the clamp current was filtered at 1 kHz, digitized at 2 kHz, and stored on an IBM AT compatible computer equipped with

NWVXl 1182

a Labmaster A/D board. The series resistance was monitored throughout the experiment, and if it changed by more than 15%, the data were discarded. The results are expressed as the mean f SEM. A key assumption in these experiments, especially those presented in Figure 7, is that 50 PM D-APV, which was present throughout, did entirely eliminate NMDA receptor activity. In a separate set of experiments (n = 5) &cyano-7-nitroquinoxaline2,3dione (CNQX; 10 PM) was added to the superfusion medium to block non-NMDA receptors. The membrane potential was held at +30 mV to relieve the Mg* block of the NMDA receptors. In every case, o-APV (50 FM) completely blocked the EPSC remaining in CNQX within 5 min. Drugs were applied by addition to the superfusion medium and were all purchased from Sigma, except for o-APV and CNQX (Cambridge Research Biochemicals) and BAPTA (Molecular Probes). Acknowledgment We thank Drs. Shaul Hestrin and Michael Strykerfor their invaluable comments on the manuscript. D. M. K. was supported by the Human Frontiers Science Program and the Fulbright Commission. This work was supported by the NIMH and the NIH (R. A. N.). R. A. N. is a member of the Keck Center for Integrative Neuroscience and the Silvio Conte Center for Neuroscience Research. 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

August

3, 1992; revised

September

P., and Oxford, C. S. (1985). Modulation of calcium by norepinephrine in internally dialyzed avian sensory J. Gen. Physiol. 85, 743-763.

Grover, L. M., and Teyler, T. J. (1990). Two components of longterm potentiation induced by different patterns of afferent activation. Nature 347, 477-479. Grover, L. M., and Teyler, T. J. (1992). N-methyl-o-aspartate recep tor independent long-term potentiation in area CA1 of rat hippocampus: input specific induction and preclusion in a nontetanized pathway. Neuroscience 49, 7-11. Gustafsson, B., and Wigstrom, H. (1990). Basic features of longterm potentiation in the hippocampus. Sem. Neurosci. 2, 321333. Gustafsson, B., Wigstrom, H., Abraham, W. C., and Huang, Y.-Y. (1987). Long-term potentiation in the hippocampus using depolarizing current pulses as the conditioning stimulus to single volley synaptic potentials. J. Neurosci. 7, 774-780. Custafsson, B., Asztely, F., Hanse, E., and Wigstrom, H. (1989). Onset characteristics of long-term potentiation in theguinea-pig hippocampal CA1 region in vitro. Eur. J. Neurosci. 7, 382-394. Guthrie, P. B., Segal, M., and Kater, S. B. (1991). Independent regulation of calcium revealed by imaging dendritic spines. Nature 354, 76-80. lino, M.,Ozawa, S.,and Tsuzuki, K. (1991). Permeationof calcium through excitatory amino acid receptor channels in cultured rat hippocampal neurones. 1. Physiol. 4.24, 151-166. Jaffe, D. B., Johnston, D., Lasser-Ross, N., Lisman, J. E., Miyakawa, H., and Ross, W. N. (1992). The spread of Na+ spikes determines the pattern of dendritic Ca2+ entry into hippocampal neurones. Nature 357, 244-246 Kauer, J. A., Malenka, R. C., and Nicoll, R. A. (1988a). NMDA application potentiates synaptic transmission in the hippocampus. Nature 334, 250-252.

16, 1992.

References

Kauer, J. A., Malenka, R. C., and Nicoll, R. A. (1988b). A persistent postsynaptic modification mediates long-term potentiation in the hippocampus. Neuron 7, 911-917.

Aniksztejn, L., and Ben-Ari, Y. (1991). Novel form of long-term potentiation produced by a K’channel blocker in the hippocampus. Nature 349, 67-69. Aniksztejn, L., Otani, S., and Benhri, Y. (1992). Quisqualate metabotropic receptors modulate NMDA currents and facilitate induction of long-term potentiation through protein kinasec. Eur. J. Neurosci. 4, 500-505. Anwyl, R. (1989). Protein kinase C and long-term potentiation the hippocampus. Trends Pharmacol. Sci. 70, 236-239.

Forscher, channels neurons.

in

Asztely, F., Hanse, E., Wigstrom, H., and Custafsson, C. (1991). Synaptic potentiation in the hippocampal CA1 region induced by application of N-methyl-o-aspartate. Brain Res. 558, 153-156. Baskys, A., and Malenka, R. C. (1991). Agonists at metabotropic glutamate receptors presynaptically inhibit EPSCs in neonatal rat hippocampus. J. Physiol. 444, 687-701. Bekkers, J. S., and Stevens, C. F. (1989). NMDA and non-NMDA receptors are colocalized at individual excitatory synapses in cultured rat hippocampus. Nature 347, 230-233. Blanton, M. G., LoTurco, J. J., and Kriegstein, A. R. (1989). Whole cell recording from neurons in slices of reptilian and mammalian cerebral cortex. J. Neurosci. Meth. 30, 203-210. Bliss, T. V. P., and Lynch, M. A. (1988). Long-term potentiation of synaptic transmission in the hippocampus: properties and mechanisms. In Long-Term Potentiation: From Biophysics to Behavior, P. W. Landfield and S. A. Deadwyler, eds. (New York: Liss), pp. 3-72. Bortolotto, Z. A., and Collingridge, G. L. (1992). Activation of glutamate metabotropic receptors induces long-term potentiation. Eur. J. Pharmacol. 274, 297-298. Collingridge, C. L., and Singer, W. (1990). Excitatory amino acid receptors and synaptic plasticity. Trends Pharmacol. Sci. 77,290296. Collingridge, G. L., Kehl, S. J., and McLennan, H. (1983). Excitatory amino acids in synaptic transmission in the Schaffer collateralcommissural pathway of the rat hippocampus. J. Physiol. 334, 33-46.

Lagercrantz, H. (1976). On the composition and function of large dense cored vesicles in sympathetic nerves. Neuroscience 7, 81-92. Linden, D. J., and Connor, J. A. (1991). Participation of postsynap tic PKC in cerebellar long-term depression in culture. Science 254,1656-1659. Linden, D. J., and Routtenberg, A. (1989). Role of protein kinase C in long-term potentiation: a testable model. Brain Res. Rev. 74, 279-296. Lynch, G., Larson, J., Kelso, S., Barrionuevo, C., and Schottler, F. (1983). Intracellular injections of EGTA block induction of hippocampal long-term potentiation. Nature 305, 719-721. Madison, D. V., Malenka, R. C., and Nicoll, R. A. (1991). Mechanisms underlying long-term potentiation of synaptic transmission. Annu. Rev. Neurosci. 74, 379-398. Malenka, R. C. (1991). Postsynaptic factors control the duration of synaptic enhancement in area CA1 of the hippocampus. Neuron 6,53-60. Malenka, R. C., Madison, D. V., and Nicoll, R. A. (1986). Potentiation of synaptic transmission in the hippocampus by phorbol esters. Nature 3.27, 695-697. Malenka, R. C., Kauer, J. A., Zucker, R. J., and Nicoll, R. A. (1988). Postsynaptic calcium is sufficient for potentiation of hippocampal synaptic transmission. Science 24.2, 81-84. Malenka, R. C., Kauer, J. A., Perkel, D. J., and Nicoll, R. A. (1989). The impact of postsynaptic calcium on synaptic transmissionits role in long-term potentiation. Trends Neurosci. 72,444-450. Malinow, R. (1991). Transmission between slice neurons: quanta1 levels, oscillations 722-724.

pairs of hippocampal and LTP. Science 252,

Malinow, R., Schulman, H., and Tsien, R. W. (1989). Inhibition of postsynaptic PKC or CaMKl I blocks induction but not expression of LTP. Science 245, 862-866.

Depolarizing 1183

Pulse-Induced

Potentiation

Manabe, T., Renner, P., and Nicoli, R. A. (1992). Postsynaptic contribution to long-term potentiation revealed by the analysis of miniature synaptic currents. Nature 355, 50-55. McCuinness, N., Anwyl, R., and Rowan, M. (1991). The effects of trans-ACPD on long-term potentiation in the rat hippocampal slice. NeuroReport 2, 688-690. Miiller, W., and Connor, J. A. (1991). Dendritic spines as individual neuronal compartments for synaptic Ca?+ responses. Nature 354, 73-76. Nicoll, R. A., and Alger, cording from submerged 156.

B. E. (1981). A simple chamber for rebrain slices. J. Neurosci. Meth. 4,153-

Nicoll, R. A., Kauer, j. A., and Malenka, excitement in long-term potentiation.

R. C. (1988). The current Neuron 7, 97-103.

O’Rourke, B., Backx, P. H., and Marban, E. (1992). Phosphorylation-independent modulation of L-type calcium channels by magnesium-nucleotide complexes. Science 257, 245-248. Osterrieder, W., Brum, G., Hescheler, J., and Trautwein, W. (1982). Injection of subunits of cyclic AMP-dependent protein kinase into cardiac myocytes modulates Ca*+ current. Nature 298, 576-578. Regehr, W. G., Connor, J. A., and Tank, D. W. (1989). Optical imaging of calcium accumulation in hippocampal pyramidal cells during synaptic activation. Nature 347, 533-536. Taube, J. S., and Schwartzkroin, P. A. (1986). Ineffectiveness organic calcium channel blockers in antagonizing long-term tentiation. Brain Res. 379, 275-285.

of po-

Thibault, O., joly, M., Muller, D., Schottler, F., Dudek, S., and Lynch, G. (1989). Long-lasting physiological effects of bath applied N-methyl-o-aspartate. Brain Res. 476, 170-173. Volknandt, W., and Zimmermann, H. (1986). Acetylcholine, ATP, and proteoglycan are common to synaptic vesicles isolated from the electric organs of electric eel and electric catfish as well as from rat diaphragm. J. Neurochem. 47, 1449-1462. Wang, J.-H., and Feng, D.-P. (1992). Postsynaptic protein kinase C essential to induction and maintenance of long-term potentiation in the hippocampal CA1 region. Proc. Natl. Acad. Sci. USA 89, 2576-258Ci. Wigstrom, H., and Gustafsson, B. (1985). On long-lasting potentiation in the hippocampus: a proposed mechanism for its dependence on coincident pre and postsynaptic activity. Acta Physiol. Stand. 723, 519-522. Zheng, F., and Gallagher, J. P. (1992). Metabotropic glutamate receptors are required for the induction of long-term potentiation. Neuron 9, 163-172.