θ frequency stimulation up-regulates the synaptic strength of the pathway from CA1 to subiculum region of hippocampus

Abstract

The subiculum (SB) is the principal target of the axons of the CA1 pyramidal cells and serves as the final relay in the trisynaptic loop between the entorhinal cortex and the hippocampus. We have examined synaptic plasticity in the synaptic pathway between the CA1 pyramidal cells and the SB in hippocampal slices and compared it under the same experimental condition with the synaptic plasticity in Shaffer collateral pathway (CA3-CA1). We find that the frequency response curve of synaptic strength induced by prolonged low-frequency stimulation (1-5 Hz) is systematically up-shifted from Shaffer collateral to the CA1-SB pathway. The up-regulation of synaptic strength is mediated by the activity-dependent modulation by β-adrenergic transmission. Because the CA3-CA1 and the CA1-SB synaptic pathways are in series and the β-adrenergic modulation is region-specific, this modulation seems to be involved in the selective control of signal transmission between the different regions of hippocampus.

The hippocampus is critically involved in the explicit memory storage of mammals. The hippocampus has three major synaptic pathways: the perforant pathway, which projects from the entorhinal cortex to the granule cells of the dentate gyrus; the mossy fiber pathway, which projects from granule cells to the pyramidal cells in the CA3 region, and the Shaffer collateral pathway, which projects from CA3 pyramidal cells to the pyramidal cells in CA1 region. The CA1 pyramidal neuron in turn projects to the pyramidal neurons of subiculum (SB), a relay which conveys the output of the hippocampus to the entorhinal cortex (1-4). Clinical observation and animal experiments have found that the SB and its related circuit is involved in several forms of memory, i.e., working memory, spatial memory, and memory of discriminative avoidance (5-8). Although the synaptic plasticity in the input pathway to the CA1 neurons (the Shaffer collateral pathway, from CA3 to CA1 neurons), has been studied extensively, little is known about the output pathway of the CA1 neurons, the synaptic pathway from CA1 to the neurons of SB.

Toward the end, we examined the synaptic plasticity in the CA1-SB pathway with that of the Shaffer collateral pathway (CA3-CA1) in hippocampal slices of C57B6 mice. We find that the frequency response curve of synaptic plasticity induced by low-frequency stimulation (1-5 Hz) is systematically up-shifted from Shaffer collateral to the CA1-SB pathway. Stimulation (1 Hz for 15 min), which induces long-term synaptic depression in the Shaffer collateral pathway, does not induce any synaptic depression in the CA1-SB pathway. Stimulation (5 Hz for 15 min), which induces a slight depression in the Shaffer collateral pathway, induced a substantial long-term synaptic potentiation (LTP) in the CA1-SB pathway. By contrast, the synaptic strength induced by higher frequency stimulation at 10 Hz and 100 Hz is not different. The LTP induced by prolonged theta frequency stimulation in SB is not dependent on the N-methyl-d-aspartate (NMDA) receptor. It is modulated by β-adrenergic receptor and depends on protein kinase A (PKA) and protein phosphatase. Our results indicate that the low-frequency stimulation induced synaptic plasticity is region specific and that the activity-dependent β-adrenergic modulation may be involved in the up-regulation of synaptic strength in these two synaptic pathways.

Materials and Methods

Horizontal brain slices (400-500 μm) containing the entorhinal cortex, the SB, and the hippocampal formation (Fig. 1) were prepared from C57/B6 mice aged 4-5 wk. Animals were decapitated, and the brains were quickly removed and placed in ice-cold artificial cerebrospinal fluid (ACSF). The tissue block containing the cortex and hippocampus were taken, and the horizontal sections were cut in the slice chopper. Slices were maintained in an interface chamber with continuous perfusion of ACSF at 1.5-2 ml/min and bubbled with 95% O₂/5% CO₂. The composition of ACSF was as follows: 124 mM NaCl/1.2 mM MgSO₄/4 mM KCl/1.2 mM NaH₂PO₄/2 mM CaCL2/26 mM NaHCO₃/10 mM d-glucose. The temperature of the slice was maintained at 28°C. Experiments were started at least 2 h after the slice dissection.

Fig. 1.

Comparison of synaptic plasticity in CA3-CA1 and CA1-SB pathways induced by 15 min of 1- to 5-Hz stimulation. (A) Schematic illustration of the placement of electrodes in two different synaptic pathway on the horizontal brain slice. (Left) The CA3-CA1 pathway. The stimulating electrode was placed on the stratum radiatum of CA1 region to stimulate the Shaffer collateral pathway. The extracellular recording electrode was placed on the stratum radiatum. (Right) The CA1-SB pathway. The stimulating electrode was placed on the Alveus to stimulate the axon of CA1 neuron. The recording electrode was placed on the SB region. An incision was made between the CA1 and SB region to prevent the contamination of other synaptic pathways. The representative waveforms of EPSPs recorded in CA1 and SB was shown on the right side of each panel. Calibration: 2 mV, 5 ms. (B1) Synaptic responses induced by 15 min of 1-Hz (Top), 3-Hz (Middle), and 5-Hz (Bottom) stimulation in the CA3-CA1 synaptic pathway. (B2) Synaptic response induced by 15 min of 1-, 3-, and 5-Hz stimulation in the CA1-SB pathway. On the bottom of 1 and 2, is the representative EPSPs waveforms before and 20 min after 5-Hz stimulation in the CA3-CA1 and CA1-SB pathways. Calibration: 2 mV, 5 ms. (C) A summary of synaptic strength 20 min after 1- to 5-Hz stimulation in the CA3-CA1 (•) and CA1-SB (○) synaptic pathways. (D) Paired-pulse facilitation in the CA3-CA1 (•) and CA1-SB (○) pathways. (E) Input-output relation in the CA3-CA1 (•) and CA1-SB (○) synaptic pathways.

Extracellular recordings were made by using ACSF-filled glass electrodes (1-3 MΩ) Stimuli were delivered through concentric bipolar stainless steel electrodes (25-μm wire diameter, FHC, Bowdoinham, ME). In the recording of synaptic plasticity in Shaffer collateral pathway (CA3-CA1), the stimulating and recording electrodes were placed in the stratum radiatum of the CA1 region (Fig. 1A Left). To study the synaptic transmission in the CA1-SB pathway, the stimulation electrode was placed on the Alveus to stimulate the axon of CA1 pyramidal neuron and the recording electrode was placed on the dendritic layer of SB region (Fig. 1A Right) (3, 9, 10). To avoid the contamination of CA3-SB pathway, an incision was made between the CA1 region and SB region by a microdissection knife in all experiments of CA1-SB pathway (Fig. 1A Right). To measure the baseline synaptic responses, 0.017 Hz (0.05-ms pulse duration) stimulation was used. To examine synaptic plasticity, 1-, 3-, 5-, 10-, and 100-Hz stimulation (0.05-ms pulse duration) was used.

For bath application, the following drugs were made and stored as concentrated stock solution and then diluted 1,000-fold when applied to the perfusion solution: 1 μM KT-5720 (Biomol, dissolved in DMSO, the final concentration of DMSO is 0.1%)/50 μM D-APV [d-(-)-2-amino-5-phosphonopentanotic acid, Sigma]/1 μM propranolol (Sigma)/1 μM isoproterenol (Sigma)/1 μM calyculin A (Sigma).

Results

Low-Frequency Stimulation Induced Synaptic Strength Is Up-Shifted in the CA1-SB Pathway. We first studied the synaptic plasticity induced by 1- to 5-Hz stimulation in the CA1-SB and CA3-CA1 pathways (Shaffer collateral pathway). We found that prolonged 1-Hz stimulation induced a substantial long-term synaptic depression in the Shaffer collateral pathway (76 ± 7% 20 min after stimulation, n = 5; Fig. 1B1 Top), but this 1-Hz stimulation did not induce synaptic depression in the CA1-SB pathway (Fig. 1B2 Top; 97 ± 3%, n = 7, P < 0.05). As we increased the frequency from 1 to 3 Hz, we found that prolonged 3-Hz stimulation, which induced a weak depression in the CA3-CA1 pathway (Fig. 1B1 Middle; 83 ± 7%, n = 5), induced a modest synaptic potentiation of ≈20% in CA1-SB pathway (Fig. 1B2 Middle; 121 ± 13%, n = 5, P < 0.05). Prolonged 5-Hz stimulation, which also induced a weak depression in the CA3-CA1 pathway (Fig. 1B1 Bottom; 86 ± 4%, n = 6), induced a substantial synaptic potentiation of ≈50% in the CA1-SB pathway (Fig. 1B2 Bottom; 151 ± 16%, n = 7, P < 0.01). A summary of synaptic strength 20 min after 1- to 5-Hz stimulation is shown in Fig. 1C.

A brief 5-Hz stimulation (30 sec) elicits a synaptic potentiation in the Shaffer collateral pathway, which declines with an increase in the total number of pulses (11). To determine whether the difference in synaptic plasticity in these two pathways only occurred with prolonged stimulation, we also examined the effect of 5-Hz stimulation using trains of different duration (pulse number). In the CA3-CA1 pathway, 5-Hz stimulation for 30 sec (150 pulse) induced synaptic potentiation; 3 min of 5-Hz stimulation (900 pulse) and 15 min (4,500 pulse) induced only the depression (30 sec: 140 ± 10%, n = 6; 3 min: 92 ± 3%, n = 6; and 15 min: 86 ± 4%, n = 7; Fig. 2A1). In contrast, we found that increasing the number of pulses still leads to synaptic potentiation in CA1-SB pathway. A 5-Hz stimulation for 3 min (900 pulse) and for 15 min (4,500 pulse) all induce synaptic potentiation in the CA1-SB pathway (120 ± 4%, n = 6 in group of 3 min and 151 ± 16%, n = 7 in group of 15 min), which is significantly different from that in CA3-CA1 pathway (Fig. 2A2; P < 0.01).

Fig. 2.

Comparison of synaptic plasticity induced by different duration (30 s-15 min) of 5-,10-, and 100-Hz stimulation. (A1) Changes of synaptic response induced by 30 s, 3 min, and 15 min of 5-Hz stimulation in the CA3-CA1 pathway (open symbols) and CA1-SB pathway (filled symbols). Circles, 30 s; triangles, 3 min; squares, 15 min. (A2) A summary of changes of synaptic strength in CA3-CA1 and CA3-CA1 pathways induced by different duration of 5-Hz stimulation. Each point is the mean of synaptic responses 20 min after 5-Hz stimulation (○, CA3-CA1 pathway; •, CA1-SB pathway). (B1) Changes of synaptic response induced by 90 s, 3 min, and 15 min of 10-Hz stimulation in the CA3-CA1 (open symbols) and CA1-SB (filled symbols) pathways. Triangles, 90 s; squares, 3 min; circles, 15 min. (B2) A summary of synaptic strength change in the CA3-CA1 and CA3-CA1 pathways induced by different duration of 10-Hz stimulation. (○, CA1-SB pathway; •, CA3-CA1 pathway). (C1) LTP induced by a single 100-Hz (1S) stimulation in the CA3-CA1 (•) and CA1-SB (○) pathways. Representative EPSPs before and 20 min after the 100-Hz tetanus in the CA3-CA1 (Left) and CA1-SB (Right) pathways were shown on the top of this panel. Calibration: 1 mV, 10 ms. (C2) A summary of synaptic strength curve induced by 15 min of 1-, 3-, 5-, and 10-Hz stimulation in the CA3-CA1 and CA1-SB synaptic pathways. Each point is the mean (± SEM) of responses 20 min after 15 min stimulation

Because prolonged 5-Hz stimulation induced potentiation in CA1-SB pathway, we were curious to know whether similar potentiation occurred with prolonged 10-Hz stimulation? As shown in Fig. 2B, 90 sec (900 pulse) of 10-Hz stimulation (900 pulse) induced comparable synaptic potentiation in both of CA3-CA1 (128 ± 6%, n = 5) and CA1-SB synaptic pathway (122 ± 6%, n = 5). However, with the increase in pulses number to 3 min (3,000 pulse) and to 15 min (9,000 pulse), no further synaptic potentiation was induced in either CA3-CA1 or CA1-SB pathway.

These results reveal the interesting feature that the up-shifting of synaptic strength only occurs in a certain range of prolonged low-frequency (1-5 Hz) stimulation in the CA1-SB pathway (Fig. 2C2), indicating that the difference in synaptic plasticity between these two pathways is not the result of the difference in general excitability. There also is no significant difference in the input-output curve (Fig. 1D) and no significant difference in the paired-pulse facilitation in these two pathways (Fig. 1E).

We next examined the LTP induced by high-frequency stimulation in the CA1-SB pathway. A single tetanus train (100 Hz, 1 s) induced LTP in the CA1-SB synaptic pathway. Although the posttetanic potentiation in the CA1-SB pathway seems to be smaller than that in the CA3-CA1 synaptic pathway, the amplitude of LTP in the CA1-SB pathway 15-20 min after the 100-Hz tetanus is not significant different from that in CA3-CA1 synaptic pathway (Fig. 2C1; CA1-SB, 136 ± 15%, n = 5 vs. CA3-CA1, 141 ± 17%, n = 5; P > 0.5).

β-Adrenergic Modulation of Low-Frequency Stimulation Induced Synaptic Plasticity in CA1-SB Pathway Is Different from That in the Shaffer Collateral Pathway. In the CA3-CA1 pathway, θ frequency stimulation induces synaptic potentiation that is modulated by the β-adrenergic receptor activation. A long train of 5-Hz stimulation, which did not induce LTP by itself, induced LTP in the presence of the β-adrenergic receptor agonist Isoprotenol (ISO) (11, 12). We now asked: Does β-adrenergic receptor activation also play a role in the synaptic plasticity induced by prolonged low-frequency stimulation in CA1-SB pathway?

We find that in this pathway β-adrenergic receptor activation also modulates synaptic plasticity and it does so in a frequency-dependent manner. In the CA1-SB pathway, the coapplication of β-adrenergic receptor agonists ISO (1 μM) facilitates the induction of LTP induced by 1-Hz stimulation, a frequency that did not otherwise give rise to LTP (control: 93 ± 5%, n = 6; ISO, 137 ± 7%, n = 6, P < 0.01; Fig. 3B). By contrast, the coapplication of ISO depressed LTP induced by prolonged 5-Hz stimulation, itself, in the CA1-SB synaptic pathway. In the presence of β-adrenergic agonist ISO (1 μM), the synaptic potentiation was 110 ± 3% (n = 6), which is significantly different from the synaptic potentiation in the absence of ISO (151 ± 12%, n = 6, 40 min after stimulation, P < 0.01; Fig. 3D). In contrast to the depression effect of ISO on the LTP induced by 5-Hz stimulation in CA1-SB pathway, the application of ISO facilitated the induction of LTP in the CA3-CA1 synaptic pathway induced by 15 min of 5-Hz stimulation (control: 86 ± 4%, n = 6; ISO: 147 ± 14%, n = 6, P < 0.01; Fig. 3C) and 1-Hz stimulation (control: 80 ± 5%, n = 6; ISO: 142 ± 13%, n = 6, 40 min after stimulation, P < 0.01; Fig. 3A) under the same experimental conditions. As a control, the application of ISO alone did not induce any significant enhancement of the synaptic transmission in CA1-SB and CA3-CA1 pathway (n = 4 for each pathway; Fig. 3E).

Fig. 3.

Region-specific and activity-dependent modulation of β-adrenergic agonist ISO. (A) ISO facilitated the induction of LTP induced by 1-Hz stimulation in the CA3-CA1 pathway (control, ○; ISO, •). (B) The 1-Hz stimulation (15 min) did not induce LTP in the CA1-SB pathway (○); coapplication of ISO (1 μM) facilitated the induction of LTP (•). (C) ISO facilitated the induction of LTP induced by 5-Hz stimulation in the CA3-CA1 pathway (control, ○; ISO, •). (D) The 5-Hz stimulation (15 min) by itself induced LTP in the CA1-SB pathway (○); coapplication of ISO (1 μM) depressed the LTP (•). (E) Application ISO alone did not induce significant enhancement of synaptic plasticity (CA1-SB pathway, ○; CA3-CA1 pathway, •). (F) A summary of synaptic strength induced by 15 min of 1- to 5-Hz stimulation in the CA3-CA1 (•) and CA1-SB (○) pathways in the presence of ISO. Each point is the mean (± SEM) of responses 40 min after low-frequency stimulation.

A summary of the synaptic plasticity induced by 1- to 5-Hz stimulation in these two pathways in the presence of ISO is shown in Fig. 3F. This graph illustrates that ISO facilitates the induction of LTP induced by 1-Hz stimulation in both CA3-CA1 and CA1-SB pathways, but ISO has opposite effects in these two pathways on the synaptic plasticity induced by 3 and 5 Hz. One interpretation of these results is that in the CA1-SB pathway β-adrenergic input is not recruited at 1 Hz but does become recruited at 3 and 5 Hz. Adding adrenergic agonists at these higher frequencies desensitizes or occludes the input from the β-adrenergic fibers. Consistent with this idea, we found that the β-antagonist blocks the LTP in CA1-SB pathway produced by prolonged 5-Hz stimulation. In the presence of β-adrenergic receptor antagonist proprenolol (1 μM), 5-Hz (15 min) stimulation failed to induce LTP (101 ± 5%, n = 6), which is significantly different from the LTP produced in the absence of β-adrenergic agonist (145 ± 9%, n = 5, P < 0.01; Fig. 4A). This blockade effect of β-adrenergic antagonist indicated that the LTP induced by prolonged 5-Hz stimulation depends on the endogenous activation of β-adrenergic receptor, as we have argued. Coapplication of 5-Hz stimulation and exogenous agonist (ISO) may saturate the β-adrenergic receptor in the CA1-SB pathway, so that the β-adrenergic receptor activation induced by the prolonged 5-Hz stimulation now only depresses the LTP. Propronolol also blocks the facilitatory effects of ISO on the synaptic plasticity induced by 1-Hz stimulation (ISO control: 140 ± 7%, n = 5; ISO + proprenolol: 100 ± 6%; Fig. 4B). The β-adrenergic modulation on the 5-Hz stimulation-induced potentiation does not depend on the recruitment of inhibitory action of interneuron and still exists in the presence of γ-aminobutyric acid antagonist picrotoxin (20 μM). Proprenolol (1 μM) again depressed the LTP induced by 5-Hz stimulation in the presence picrotoxin (Fig. 4C; control: 136 ± 5%, n = 6, proprenolol: 98 ± 6%, n = 5, P < 0.01).

Fig. 4.

LTP induced by 5-Hz stimulation in the CA1-SB pathway is mediated by β-adrenergic receptor and PKA. (A) β-adrenergic receptor antagonist propranolol (1 μM) depressed the LTP induced by 5-Hz stimulation (□, 5-Hz control; •, propranolol + 5 Hz; ○, propranolol alone). (B) Propranolol reversed the facilitation effect of ISO on the synaptic change induced by 1-Hz stimulation (○, ISO; •, ISO + propranolol). (C) Proprenolol depressed the LTP induced by 5-Hz stimulation in the presence of γ-aminobutyric acid antagonist picrotoxin (○, control; •, proprenolol). (D) NMDA antagonist D-APV (50 μM) did not affect the LTP induced by 5-Hz stimulation (○, control; •, D-APV). (E) PKA inhibitor KT5720 (1 μM) depressed the LTP induced by 5-Hz stimulation (○, control; •, KT5720) (F) KT5720 reversed the facilitation effect of ISO on the induction of LTP induced 1-Hz stimulation (○, 1-Hz stimulation + ISO; •, 1-Hz stimulation + ISO + KT5720). (G) Calyculin A depressed the LTP induced by 5-Hz stimulation. The 5-Hz stimulation (15 min) induced LTP in the CA1-SB pathway (○). Coapplication of calyculin A (1 μM) depressed the LTP (•). Application of calyculin A alone has no effect on the baseline EPSP (▴). (H) Calyculin A facilitated the induction of LTP induced by 1-Hz stimulation. The 1-Hz stimulation (15 min) did not induce any LTP (○). In the presence of calyculin A (1 μM), 1-Hz stimulation (15 min) induced LTP (•).

In further contrast to LTP in Shaffer collateral pathway, the LTP induced by prolonged θ frequency stimulation in the CA1-SB pathway was not dependent on the NMDA receptor. As shown in Fig. 4D, the prolonged 5-Hz stimulation (15 min) induced a substantial synaptic potentiation (145 ± 11%, n = 7) even in the presence of NMDA antagonist D-APV (50 μM). This potentiation is not significantly different from the LTP in the absence of D-APV (P > 0.5), indicating the prolonged θ frequency stimulation induced LTP in the CA1-SB pathway is totally independent to the NMDA receptor.

Prolonged θ Frequency Potentiation in CA1-SB Pathway Is Dependent on PKA and Protein Phosphatase. In several synaptic pathways of hippocampus, β-adrenergic receptor activation leads to an increase in cAMP and the consequent recruitment of PKA. This effect is thought to be via coupling of the receptor to a GTP binding protein (13, 14). We therefore examined the effect of PKA inhibition in the CA1-SB synaptic pathway. In the presence of PKA inhibitor KT5720 (1 μM), 5-Hz stimulation (15 min) induced a slight depression (91 ± 4%, n = 6), which is significantly different from the synaptic potentiation in the control experiments (142 ± 7%, 20 min after stimulation n = 7, P < 0.01; Fig. 4E). The PKA inhibitor also blocked the facilitatory effects of β-adrenergic agonist ISO on the LTP induced by 1-Hz stimulation in this pathway (ISO control: 145 ± 13%, 40 min after stimulation, n = 6; ISO + KT5720: 93 ± 7%, n = 5, P < 0.01; Fig. 4F). KT5720 has no significant effect on the synaptic plasticity induced by 1 Hz (15 min) stimulation (control: 98 ± 4%; KT5720: 92 ± 5%, n = 5, P > 0.5, data not shown).

In studies of β-adrenergic modulation on the 5-Hz-induced LTP in the CA3-CA1 pathway, O'Dell and coworkers (11) found that the modulation is mediated through a PKA-mediated inhibition of protein phosphatases and inhibition of protein phosphatase 1 and 2A facilitated the induction of 5-Hz stimulation-induced LTP. To test this possibility, we applied calyculin A, an inhibitor of protein phosphatase 1 and 2A and found that the effect in the CA1-SB pathway is different from CA3-CA1 pathway. In contrast to its facilitation effect in Shaffer collateral pathway (11), the application of protein phosphatase inhibitor calyculin A (1 μM) depressed the LTP induced by prolonged 5-Hz stimulation in the CA1-SB pathway (Fig. 4G; control: 149 ± 10%, n = 6; calyculin A: 115 ± 4%, n = 6, P < 0.01). The effects of the phosphatase inhibitor is activity-dependent, whereas calyculin A reduces the LTP in response to 5 Hz, it facilitated the induction of LTP induced by 1-Hz stimulation in this pathway (Fig. 4H; control: 96 ± 3%, n = 7, Caliculin A: 127 ± 5% n = 6, P < 0.01).

Discussion

The CA1 pyramidal neurons synapse onto the pyramidal neuron of SB has been studied both morphologically and electrophysiologically (1-4). This monosynaptic pathway is mediated by glutamate released from the CA1 pyramidal cells acting on AMPA receptor in the SB neurons. The AMPA antagonist NBQX completely blocked the excitatory postsynaptic potential (EPSP) elicited in SB neuron by stimulation of the axons of the CA1 neuron (alveus) (3, 10). In rat (in vivo), 1-Hz stimulation did not induce any synaptic depression in CA1-SB pathway but it induced rather a weak form LTP in this pathway (15). The synaptic mechanism underlying this phenomenon however is unknown. We examined systematically the synaptic plasticity of CA1-SB pathway in the slices of mice and compared it under the same experimental condition with the Shaffer collateral pathway. We found that the synaptic strength induced by 1- to 5-Hz frequency stimulation, but not 10- and 100-Hz stimulation is enhanced from Shaffer collateral to CA1-SB pathway. We also found that this region specific difference of synaptic plasticity is mediated by the activity-dependent β-adrenergic modulation.

In the CA3-CA1 pathway, prolonged 5-Hz stimulation elicits LTP when the phosphatase cascade was blocked by the application of PKA agonist or β-adrenergic agonist, which presumably depress the phosphatase cascade by phosphorylating the protein phosphatase inhibitor (11, 12). By contrast, we find that prolonged 5-Hz stimulation could induce a profound LTP by itself in the CA1-SB synaptic pathway without requiring the coapplication of a PKA agonist or an inhibitor of protein phosphatase, suggesting that, in this pathway, prolonged 5-Hz stimulation recruits less activation of protein phosphatase 1 and 2A than in the Shaffer collateral pathway. The 5-Hz stimulation in the CA1-SB pathway may induce a higher level of endogenous PKA activation, which can counteract phosphatase activity more effectively in this synaptic pathway and thereby results in less long-term synaptic depression and more LTP. The effect of PKA on the protein phosphatase evoked by the prolonged 5-Hz stimulation is mediated by the β-adrenergic receptor via a GTP binding protein coupling (13, 14). Coapplication of β-adrenergic receptor agonist ISO enable the induction of prolonged 5-Hz stimulation induced LTP in the CA3-CA1 pathway (11, 12). Interestingly, we find that although β-adrenergic modulation occurs in both pathways, there are differences in how it is recruited in the CA1-SB and CA3-CA1 pathways. In the CA1-SB pathway, β-adrenergic agonists enable the induction of LTP induced by prolonged 1-Hz stimulation, much as they do in the CA3-CA1 pathway. By contrast, β-adrenergic agonists disable the induction of LTP induced by prolonged 5-Hz stimulation in the CA1-SB pathway. The different effect of β-adrenergic modulation in these two pathways may reflect region-specific differences in the β-adrenergic activation level induced by prolonged low-frequency stimulation. Immunohistochemical studies support this idea and show that there is a higher density of β-adrenergic receptor in the CA1-SB region than in the CA3 region (16). It, therefore, is likely that the endogenous β-adrenergic activation induced by prolonged 5-Hz stimulation is already very high in the SB, so that 5-Hz stimulation alone, without coapplication of β-adrenergic receptor agonist could induce LTP in SB and the coapplication of β-adrenergic receptor agonist may simply saturate the enabling effect of β-adrenergic receptor in the SB. As a result, in the SB the LTP induced by the prolonged 5-Hz stimulation is depressed rather than enhanced in the presence of β-adrenergic agonist. By contrast, when the β-adrenergic activation level is low as with 1-Hz stimulation, β-adrenergic agonist still enabled the induction of LTP.

Previous studies have shown that the β-adrenergic agonists facilitate the prolonged low-frequency stimulation in the CA3-CA1 pathway, but not in the CA3-CA3 pathway (17). We now find that the β-adrenergic modulation is even stronger in the CA1-SB pathway, which is downstream of the CA3-CA1 pathway. Our result provided another example for the region-specific and activity-dependent β-adrenergic modulation induced by prolonged low-frequency stimulation The β-adrenergic modulation is not just involved in synaptic plasticity per se but also plays a critical role in the shifting of the amplification of synaptic plasticity from one synaptic pathway in the hippocampus to another. This new feature of β-adrenergic modulation may play an important role in the modulation of learning and memory.