Postsynaptic M1 and M3 receptors are responsible for the muscarinic enhancement of retrograde endocannabinoid signalling in the hippocampus

Postsynaptic M1 and M3 receptors are responsiblefor the muscarinic enhancement of retrogradeendocannabinoid signalling in the hippocampus

Takako Ohno-Shosaku,1 Minoru Matsui,2,3 Yuko Fukudome,1 Jumpei Shosaku,1 Hiroshi Tsubokawa,4 Makoto M.Taketo,3,5 Toshiya Manabe2,6 and Masanobu Kano1

1Department of Cellular Neurophysiology, Graduate School of Medical Science, Kanazawa University, 13-1 Takara-machi, Kanazawa920-8640, Japan2Division of Neuronal Network, Department of Basic Medical Sciences, The Institute of Medical Science, The University of Tokyo,Tokyo, Japan3Laboratory of Biomedical Genetics, Graduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo, Japan4National Institute for Physiological Sciences, Okazaki, Japan5Department of Pharmacology, Graduate School of Medicine, Kyoto University, Kyoto, Japan6Division of Cell Biology and Neurophysiology, Department of Neuroscience, Faculty of Medicine, Kobe University, Kobe, Japan

Keywords: depolarization-induced suppression of inhibition (DSI), inhibitory transmission, mouse, rat, synaptic modulation

Abstract

The cholinergic system is crucial for higher brain functions including learning and memory. These functions are mediated primarily bymuscarinic acetylcholine receptors (mAChRs) that consist of five subtypes (M1–M5). A recent study suggested a novel role ofacetylcholine as a potent enhancer of endocannabinoid signalling that acts retrogradely from postsynaptic to presynaptic neurons. Inthe present study, we further investigated the mechanisms of this cholinergic effect on endocannabinoid signalling. We made pairedwhole-cell recordings from cultured hippocampal neurons, and monitored inhibitory postsynaptic currents (IPSCs). The postsynapticdepolarization induced a transient suppression of IPSCs (DSI), a phenomenon known to involve retrograde signalling by endocanna-binoids. The cholinergic agonist carbachol (CCh) markedly enhanced DSI at 0.01–0.3 mM without changing the presynaptic cannabinoidsensitivity. The facilitating effect of CCh on DSI was mimicked by the muscarinic agonist oxotremorine-M, whereas it was eliminated bythe muscarinic antagonist atropine. It was also blocked by a non-hydrolizable analogue of GDP (GDP-b-S) that was appliedintracellularly to postsynaptic neurons. The muscarinic enhancement of DSI persisted to a substantial degree in the neurons preparedfrom M1-knockout and M3-knockout mice, but was virtually eliminated in the neurons from M1/M3-compound-knockout mice. CCh stillenhanced DSI significantly under the blockade of postsynatpic Kþ conductance, and did not significantly influence the depolarization-induced Ca2þ transients. These results indicate that the activation of postsynaptic M1 and M3 receptors facilitates the depolarization-induced release of endocannabinoids.

Introduction

The cholinergic system in the central nervous system (CNS) is

involved in various neural functions including cognition, learning

and memory (Bartus et al., 1982; Nilsson et al., 1992; Aigner,

1995; Jerusalinsky et al., 1997; Segal & Auerbach, 1997). These

functions are mediated largely by members of the muscarinic acet-

ylcholine receptor (mAChR) family, which belongs to the superfamily

of G-protein-coupled receptors (Hulme et al., 1990; Wess, 1996). Five

subtypes of mAChRs (M1–M5) have been identified so far. M1, M3 and

M5 receptors couple positively to phospholipase C (PLC) through Gaq/

11, whereas M2 and M4 receptors couple negatively to adenylyl cyclase

through Gai/o (Caulfield, 1993; Wess, 1996; Caulfield & Birdsall,

1998; Felder et al., 2000). In the hippocampus, four of the subtypes

(M1–M4) are expressed abundantly. Individual subtypes feature dis-

tinct cellular and subcellular distributions, indicating their different

roles in physiological functions (Vilaro et al., 1993; Levey et al., 1995;

Rouse et al., 1999). Activation of mAChRs modulates various neural

functions by changing membrane potential, cytoplasmic Ca2þ con-

centration and synaptic transmission (Nicoll, 1985; Nicoll et al., 1990;

Irving & Collingridge, 1998; Kimura, 2000).

It is now generally accepted that endogenous cannabinoids

(endocannabinoids) mediate a retrograde signal from postsynaptic

neurons to presynaptic terminals (Kano et al., 2002; Kreitzer &

Regehr, 2002; Wilson & Nicoll, 2002). The endocannabinoids are

released from postsynaptic neurons in response to depolarization or

activation of group I metabotropic glutamate receptors (mGluRs), and

then suppress the transmitter release through activating presynaptic

cannabinoid receptor type 1 (CB1) (Kreitzer & Regehr, 2001b;

Maejima et al., 2001; Ohno-Shosaku et al., 2001; Wilson & Nicoll,

2001). Depolarization-induced suppression of inhibition (DSI) is the

best-studied example of synaptic modulation that depends on

endocannabinoid-mediated retrograde signalling. Several studies have

shown that cholinergic activation facilitates hippocampal DSI. Bath-

applied carbachol (CCh), a cholinergic agonist, or synaptically

European Journal of Neuroscience, Vol. 18, pp. 109–116, 2003 � Federation of European Neuroscience Societies

doi:10.1046/j.1460-9568.2003.02732.x

Correspondence: Professor Masanobu Kano, as above.

E-mail: [email protected]

Received 17 February 2003, revised 24 April 2003, accepted 28 April 2003

released acetylcholine enhances the DSI of spontaneous inhibitory

postsynaptic currents (IPSCs) (Pitler & Alger, 1994; Martin & Alger,

1999; Martin et al., 2001). This enhancement has been ascribed to

selective activation of DSI-sensitive GABAergic interneurons by CCh

(Martin et al., 2001). GABAergic interneurons are heterogeneous in

DSI sensitivity. CCh may selectively elevate the firing rate of DSI-sen-

sitive interneurons and increase the proportion of DSI-sensitive spon-

taneous IPSCs, which results in the enhancement of DSI. In addition, a

recent study demonstrates that CCh enhances DSI of evoked IPSCs

and suggests a facilitating effect of CCh on endocannabinoid sig-

nalling itself (Kim et al., 2002). Several important questions, however,

remain unanswered. These include the site of action of the cholinergic

agonist, the receptor subtypes involved and whether the enhancement

is in fact the result of an increase in endocannabinoid release.

Here, we offer robust evidence that activation of postsynaptic

mAChRs significantly enhances the depolarization-induced release

of endocannabinoids. Using three strains of knockout mice enabled us

to demonstrate successfully that M1 and M3 receptors mediate the

enhancement of endocannabinoid signalling. The study presented here

provides clear answers to the aforementioned questions and can thus be

considered to represent an important contribution to our knowledge of

the muscarinic enhancement of endocannabinoid signalling.

Materials and methods

All experiments were carried out in accordance with the guidelines laid

down by the animal welfare committees of Kanazawa University, The

Institute of Medical Science at The University of Tokyo and The

National Institute for Physiological Sciences.

Cultures

Newborn rats or mice were decapitated under deep ether anaesthesia,

their brains were quickly removed and hippocampal neurons were

cultured as described previously (Ohno-Shosaku et al., 2001). The

cultures were kept at 36 8C in 5% CO2 for 10–14 days before use.

Electrophysiology

The external solution for the experiments contained (in mM), 140

NaCl, 2.5 KCl, 1 MgCl2, 2 CaCl2, 10 N-2-hydroxyethylpiperazine-N0-2-ethanesulphonic acid (HEPES), 10 glucose and 1 kynurenic acid

(pH 7.3 adjusted with NaOH). The recording chamber was perfused

with the external solution at a flow rate of 1–3 mL/min. The internal solu-

tion contained (in mM): 120 K-gluconate, 15 KCl, 6 MgCl2, 5 ethylene

glycol-bis(b-aminoethyl ether)N,N,N0,N0-tetraacetic acid (EGTA), 10

HEPES, 20 KOH, 5 Na2ATP and 0.2 mM Na3GTP (pH 7.3 adjusted

with KOH). In some experiments, we also used the internal solution

without Na2GTP. Because the results obtained with these two solutions

showed no significant differences, the data for both of them were

pooled. In some experiments to examine the effects of intracellular

injection of Csþ, CsCl was substituted for K-gluconate. The electrode

resistance ranged from 3 to 5 MV when filled with the internal

solution. All experiments were performed at room temperature.

Whole-cell currents from one pair of neurons were recorded simul-

taneously with the voltage-clamp method using a patch-clamp ampli-

fier (EPC-9/2 or EPC-9/3, HEKA, Germany). The presynaptic neuron

was stimulated by applying positive voltage pulses (from �80 to 0 mV,

2 ms, repeated at 0.5 Hz). The evoked IPSCs were recorded at �80 mV

from the postsynaptic neuron.

Evaluation of DSI

DSI was induced by applying a depolarizing voltage pulse (from �80

to 0 mV, 0.3–5 s) to the postsynaptic neuron. Its magnitude was

calculated as the percentage reduction of the mean IPSC amplitude

after depolarization (at 4–18 s) compared with that before. Averaged

data from different experiments are shown as mean � SEM.

Ca2þ measurements

Neurons were loaded for at least 10 min with a Ca2þ indicator dye

(fura-2, 200mM) through patch pipettes. Fluorescence signals for exci-

tations of 340 nm (F340) and 380 nm (F380) were measured at 0.2 Hz

at the soma and the proximal dendrite by means of a cooled-CCD

camera system (Acquacosmos, Hamamatsu Photonics, Japan). Intra-

cellular Ca2þ levels were calculated by converting fluorescent ratios

(F340/F380) to intracellular Ca2þ concentrations ([Ca2þ]i) with the

use of the following formula: [Ca2þ]i¼Kd[(R�Rmin)/(Rmax�R)]Fo/Fs,

where R is the ratio value, Rmin is the ratio for a Ca2þ-free solution

(20 mM BAPTA), Rmax is the raio for a saturated Ca2þ solution (2 mM

Ca2þ), Kd is the dissociation constant for fura-2 (224 nM), Fo is the

intensity of the Ca2þ-free solution at 380 nm, and Fs is the intensity of

the saturated Ca2þ solution at 380 nm.

Generation of mutant mice

For disruption of the M1 receptor gene, a targeting vector, pChrm1-N1,

was constructed by using the mouse M1 genomic fragments (Matsui

et al., 1999). The SacI–KpnI (8.2 kb) and SacI–BamHI (1.4 kb) frag-

ments were placed upstream and downstream, respectively, of the

inverted PGK-neo-bpA cassette (Soriano et al., 1991), followed by

insertion of the PGK-DTA cassette (Yagi et al., 1990) at the down-

stream end. It should be noted that the genomic segment to be deleted

in this targeting design was the same as that which was mutated in

another mutant mouse line for the M1 receptor gene (Hamilton et al.,

1997). The gene targeting in embryonic stem (ES) cells (RW4,

Genome Systems) was performed as described elsewhere (Matsui

et al., 2000). G418-resistant cells were screened by PCR using the

primers PR1 (50-CAG ACT GCC TTG GGA AAA GC-30) and MR9

(50-TCC TGC CTA AAC GCA AAC AG-30), to amplify a 1.5-kb

fragment. Two independent clones (1N-8 and 1N-11) to be employed

for the homologous recombination were verified by using an M1 probe

and a neo probe. The 1N-11 clone differentiated into the germ line that

harboured the mutant allele. A mutant mouse line was established in a

mixed background of 129/SvJ and C57BL/6. The M1�/� mice

appeared healthy, as in another study of a different mouse line lacking

M1 (Hamilton et al., 1997).

The genotyping protocol for the M3 allele has been described

elsewhere (Matsui et al., 2000). The procedure for the M1 allele

was similar, except that the primers used were MR25 (50-GGG

TCA CTG AGA AGT AAC GG-30), PR1 (50-CAG ACT GCC TTG

GGA AAA GC-30) and MF29 (50-GCT CGT GTC CTT TCT TTT

CAG-30).For the generation of M1�/�M3�/� mice, the M1 (N3 generation)

and M3 mutant mice (N8 generation) (Matsui et al., 2000) were

crossed to obtain M1þ/–M3þ/– mice. Cross-breeding of these mice

yielded pups with various genotypes for M1 and M3 alleles, including

M1�/�M3�/� mutants. The descendants of these mutant mice were

used for further phenotypic analysis. Because the pups lacking M3

were hypophagic around weaning, we fed them with hydrated paste

food to improve their growth (Matsui et al., 2000).

Chemicals

WIN55,212-2, AM281, pirenzepine and oxotremorine-M were pur-

chased from Tocris Cookson (UK). Atropine and gallamine were from

Nacalai Tesque (Japan), and carbachol (CCh) was from Sigma-Aldrich

(USA).

110 T. Ohno-Shosaku et al.

� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 18, 109–116

Results

Whole-cell recordings were obtained from pairs of neurons in hippo-

campal cultures prepared from newborn rats. In about half of the

neuron pairs, postsynaptic depolarization (to 0 mV, for 5 s) induced

a transient suppression of IPSCs, termed depolarization-induced

suppression of inhibition (DSI) (Llano et al., 1991; Pitler & Alger,

1992b) (Fig. 1A). DSI is known to be mediated by endocannabinoids

that are released from depolarized postsynaptic neurons in a Ca2þ-

dependent manner, act retrogradely onto presynaptic CB1 and sup-

press the transmitter release (Kreitzer & Regehr, 2001a; Ohno-Sho-

saku et al., 2001; Wilson & Nicoll, 2001; Diana et al., 2002; Yoshida

et al., 2002). For the subsequent experiments, we used the neuron pairs

in which 5-s depolarization induced DSI with a magnitude of more

than 15%.

Enhancement of DSI by activation of postsynaptic mAChRs

First, we examined the effects of a cholinergic agonist, CCh, on DSI.

The magnitude of DSI induced by 5-s depolarization (5-s DSI) ranged

from 55.9% to 97.9% (77.4� 4.2%, n¼ 8). In the same pairs, 1-s

depolarization produced no or significantly smaller DSI (1-s DSI)

ranging from 0.3% to 43.7% (17.0� 4.8%) (Fig. 1A). Addition of CCh

to the external solution enhanced the 1-s DSI in a dose-dependent

manner (Fig. 1B). This effect of CCh was significant even at a low

concentration of 0.03 mM (Fig. 1C). The effect of 1-s depolarization in

the presence of 0.3 mM CCh was almost equivalent to that of 5-s

depolarization without CCh (Fig. 1C). As reported previously (Ohno-

Shosaku et al., 2001), the 5-s DSI was completely blocked by treating

the neurons with a CB1 antagonist, AM281 (0.3 mM) (Fig. 1D and E).

This antagonist also suppressed the 5-s DSI in the presence of 0.3 mM

CCh (Fig. 1D and E). These results indicate that CCh enhances DSI by

facilitating the endocannabinoid signalling.

We then examined the pharmacology of cholinergic receptors

responsible for the enhancement of DSI. As exemplified in Fig. 2A

and summarized in Fig. 2B, a non-selective muscarinic antagonist,

atropine (1 mM), and an M1-preferring antagonist, pirenzepine (1 mM),

blocked the effect of 0.3 mM CCh on DSI completely. By contrast, the

CCh effect persisted in the presence of a high dose (10 mM) of an M2-

preferring antagonist, gallamine. A muscarinic agonist, oxotremorine-

M (0.3 mM), mimicked the effect of CCh on DSI. When a non-

hydrolysable analogue of GDP (GDP-bS) was applied intracellularly

to the postsynaptic neuron through a patch pipette, CCh had no

significant effect on DSI (Fig. 3A and B). The summary data in

Fig. 3C show that the CCh-induced enhancement of DSI was

eliminated by the effect of GDP-b-S on the postsynaptic neurons.

Fig. 1. Facilitation of DSI by CCh. (A) An example of a neuron pair demon-strating prominent DSI to 5-s depolarization but not to 1-s depolarization. Theamplitude of IPSCs is plotted as a function of time. The postsynaptic neuronwas depolarized at the time indicated by the arrows. IPSC traces acquired at theindicated points are shown on the right. Each trace represents the average ofseveral consecutive IPSCs. (B) A representative experiment demonstrating adose-dependent enhancement of DSI by CCh. DSI was induced by 1-sdepolarization. In this and the following figures exemplifying DSI, IPSC tracesacquired before and after depolarization have been superimposed for each DSIinduction trial. (C) Averaged data for the enhancement of DSI by CCh. CChconcentrations are shown below the columns. DSI was induced by either 5-s(left) or 1-s (four columns on the right) depolarization. The asterisks representstatistically significant differences from the control (1-s depolarization, 0 CCh).��P< 0.01; ��P< 0.001 (paired t-test). (D and E) An example (D) and theaveraged data (E) showing the blockade of DSI by the CB1 antagonist AM281(0.3mM) in the presence or absence (Control) of CCh (0.3mM). DSI was inducedby 5-s depolarization. For the experiments shown in this figure, the neuron pairswere used that exhibited more than 50% reduction in IPSC amplitude after 5-sdepolarization.

Fig. 2. Gallamine-resistant mAChRs are involved in the DSI enhancement.Examples of IPSCs (A) and summary data (B) for the enhancement of DSI bycholinergic agonists and its inhibition by muscarinic antagonists. DSI wasinduced by 0.3–5-s depolarization. The enhancement of DSI by CCh (0.3mM)was antagonized by atropine (atr, 1mM) and pirenzepine (pir, 1mM), but not bygallamine (gall, 10mM). Oxotremorine-M (Oxo-M, 0.3mM) effectively enhancedDSI, similar to enhancement by CCh. �P< 0.05; ��P< 0.001 (paired t-test).


Muscarinic enhancement of endocannabinoid release 111

These results suggest that CCh enhances DSI by activating gallamine-

resistant mAChRs on the postsynaptic neurons.

CCh enhances depolarization-induced endocannabinoidrelease

The enhancement of DSI can be caused by either an increase in pre-

synaptic cannabinoid sensitivity or facilitation of endocannabinoid

release from postsynaptic neurons. To test the former possibility, we

examined whether CCh affected the sensitivity of IPSCs to the

cannabinoid agonist WIN55,212-2. CCh (0.3mM) had no effect on the

suppression of IPSCs caused by 0.3 nM, 3 nM and 30 nM WIN55,212-2

(Fig. 4A), indicating that CCh enhances DSI mainly by facilitating

endocannabinoid release.

We then examined whether endocannabinoid release is facilitated by

CCh. Although our electrophysiological technique is very sensitive in

detecting the actions of cannabinoids on synaptic transmission, it is not

so for measuring the released endocannabinoids directly. We therefore

took an alternative approach to estimate the released endocannabinoids

indirectly. Previously, we examined the dose–inhibition relationship

between IPSCs and WIN55,212-2 (Ohno-Shosaku et al., 2002b).

Using this relationship, we estimated the amount of endocannabinoids

released during DSI as an equivalent concentration of WIN55,212-2

(Ohno-Shosaku et al., 2002a). By applying the same method, we

attempted to estimate the amount of endocannabinoids released from

depolarized neurons before and during application of CCh. In 17

neuron pairs tested, application of 0.3 mM CCh itself had no effect on

IPSCs in nine pairs (98.7� 2% of the control, Fig. 4B, left, ‘group 1’)

but caused a significant reduction of the IPSC amplitude in the

remaining eight pairs (64.4� 6.3% of the control, Fig. 4B, right,

‘group 2’).

In group 1, depolarization reduced the IPSC amplitude to 81.9%

and 30.5% before and during CCh application, respectively (Fig. 4B,

left). The degrees of suppression generated by depolarization alone

(Fig. 4C, depol-1) and with CCh (Fig. 4C, (CCh þ depol)-1) are

equivalent to those induced by, respectively, 0.4 nM and 3.3 nM

WIN55,212-2. This demonstrates that CCh caused about an eight-fold

enhancement of endocannabinoid release, as an equivalent concentra-

tion of WIN55,212-2.

In group 2, depolarization and 0.3 mM CCh reduced the IPSC ampli-

tude to 81.4% and 64.4% of the control, respectively (Fig. 4B, right).

Fig. 3. The CCh-induced enhancement of DSI is blocked by intracellularapplication of GDP-b-S to postsynaptic neurons. (A) Examples of IPSCsshowing the effects of 0.3mM CCh on DSI with (right) or without (left)GDP-b-S. DSI was induced by 1-s depolarization. (B and C) The individual(B) and averaged data (C) obtained before, during and after application of0.3mM CCh. DSI was induced by 0.3-s (closed circles), 0.5-s (open circles), 1-s(closed squares) or 2-s (open squares) depolarization. ��P< 0.001 (paired t-test). The internal solution for postsynaptic neurons contained 1.5 mM LiCl[GDPbS(–)] or 0.5 mM GDP-b-S in the form of a tri-lithium salt [GDPbS(þ)].

Fig. 4. Estimation of depolarization-induced endocannabinoid release beforeand during CCh application. (A) The suppression of IPSCs by 0.3 nM, 3 nM and30 nM WIN55,212-2 in the absence (open columns) and presence (shadedcolumns) of 0.3 mM CCh. (B) The suppression of IPSCs by 0.3mM CCh (CCh-1,CCh-2), depolarization (depol-1, depol-2), and by the two combined [(CCh þdepol)-1 (CCh þ depol)-2]. Data are summarized for the two population ofneuron pairs: those with no (group 1) and significant (group 2) suppression by0.3mM CCh. (C) Dose-dependent suppression of IPSCs by the cannabinoidagonist WIN55,212-2. The same set of data is shown as that in a previous reportof ours (Ohno-Shosaku et al., 2002a, Figure 6). The data were fitted by asigmoid curve. Horizontal dotted lines indicate the levels of suppressioninduced by depolarization (depol-1, depol-2), 0.3 mM CCh (CCh-2), and acombination of the two [(CCh þ depol)-1 (CCh þ depol)-2] for the two groupsof pairs shown in B. Because CCh caused no significant suppression of IPSCs ingroup 1 (CCh-1), the endocannabinoid concentration should be 0. Therefore,the level of suppression by CCh-1 is not plotted. The curve was fitted accordingto the equation: y¼ 100/[1þ (x/EC50)n], where x is the concentration, EC50 isthe concentration of agonist resulting in 50% of the response and n is the slope.EC50 and n of the fitting curve are 1.582 and 1.097, respectively.



These values are equivalent to those induced by, respectively, 0.4 nM

and 0.9 nM WIN55,212-2 (Fig. 4C, depol-2 and CCh-2). The CCh-

induced suppression could be caused by both cannabinoid-dependent

and -independent mechanisms (Kim et al., 2002). Thus, the actual

concentration of endocannabinoids released by 0.3 mM CCh should be

lower than the equivalent of that resulting from 0.9 nM WIN55,212-2.

In the presence of 0.3 mM CCh, depolarization reduced IPSC amplitude

to 19.5% of the control (Fig. 4B, right), equivalent to that induced by

5.8 nM WIN55,212-2 (Fig. 4C, (CCh þ depol)-2). This estimated value

(5.8 nM) was much higher than the algebraic sum of the estimated

value for depolarization alone (equivalent to 0.4 nM WIN55,212-2) and

that for CCh alone (equivalent to 0.9 nM WIN55,212-2). These data

clearly demonstrate that CCh greatly enhances the depolarization-

induced release of endocannabinoids.

Involvement of M1 and M3 receptors in DSI enhancement

To identify the subtypes of mAChRs involved in DSI enhancement, we

used three strains of knockout mice lacking M1 (Fig. 5), M3 (Matsui

et al., 2000) or M1 plus M3. In the hippocampal neuron pairs cultured

from wild-type mice, oxotremorine-M (Oxo-M, 0.3 mM) enhanced DSI

significantly (Fig. 6A, WT; Fig. 6B and C, M1þ/M3þ). Oxotremorine-

M had much less effect on the neurons prepared from M1-knockout

mice, but it still enhanced DSI significantly (Fig. 6A, M1-KO; Fig. 6B

and C, M1–/M3þ). In the neuron pairs from M3-knockout mice, the

DSI-enhancing effect of oxotremorine-M was reduced but persisted to

a substantial degree (Fig. 6A, M3-KO; Fig. 6B and C, M1þ/M3–). In

contrast, DSI enhancement by oxotremorine-M was virtually elimi-

nated in the neuron pairs from M1/M3-compound-knockout mice

(Fig. 6A, M1/M3-KO; Fig. 6B and C, M1–/M3–). These reductions

and eliminations of the enhancing effect are unlikely to result from the

impairment of intracellular signalling downstream of Gaq/11 because

group I mGluRs, another family of Gaq/11-coupled receptors, func-

tioned normally in all three mouse strains. We found that the enhance-

ment of DSI by the group I mGluR agonist DHPG (Varma et al., 2001;

Ohno-Shosaku et al., 2002a) was normal in all these knockout mice

(Fig. 6A–C, DHPG). These results clearly indicate that both M1 and

M3 receptors are involved in the cholinergic enhancement of DSI.

Changes in Kþ conductance or depolarization-induced Ca2þ

transients cannot account for DSI enhancement

DSI could be enhanced through changes in postsynaptic Kþ conduc-

tance that may be influenced by muscarinic activation (Cole & Nicoll,

1984; Nicoll et al., 1990). However, this was not the case because

we found that CCh (0.3 mM) enhanced DSI significantly (from

Fig. 5. Generation of M1�/�mice. (A) Targeting strategy for generation of M1�/�mice. The box marked ‘Chrm1’ corresponds to a coding region of the M1 receptor(‘Chrm1’ is an approved gene symbol for the mouse M1 gene.). Arrowheads PR1 and MR9 indicate the PCR primers used for homologous recombinant screening,whereas MF29, PR1 and MR25 indicate those used for genotyping. BamHI (B), KpnI (K), PstI (P), EcoRI (R), SacI (S) and SalI (Sl) sites are shown together with theexpected sizes hybridizable to the M1 and the neo probes. pBS, pBluescript. (B) Hybridization with the M1 probe showing a 5.9-kb band specific to the targeted alleleand a 4.7-kb band derived from the wild-type allele. 1N-8 and 1N-11 are proper homologous recombinant clones, whereas RW4 is a parental ES cell clone.

Fig. 6. Cholinergic enhancement of DSI involves M1 and M3 receptors. Exam-ples of IPSCs (A) and summary data (B and C) for the effects of oxotremorine-M (Oxo-M, 0.3 mM) and DHPG (5mM) on DSI in hippocampal neurons culturedfrom wild-type (WT), M1-deficient (M1-KO), M3-deficient (M3-KO) andM1/M3-deficient (M1/M3-KO) mice. DSI was induced by 0.3–5-s depolariza-tion. Open and shaded columns in B indicate the magnitude of DSI in theabsence and presence of oxotremorine-M (0.3mM, left) or DHPG (5mM, right),respectively. The same data are shown in C as the per cent increases in DSImagnitude generated by oxotremorine-M or DHPG relative to control.�P< 0.05; ��P< 0.01; ��P< 0.001 (paired t-test for B; unpaired t-test for C).



29.9� 9.5% to 62.3� 12.3%, n¼ 7, P< 0.01) even when the post-

synaptic Kþ conductance was blocked substantially by intracellularly

applied Csþ. Another possibility is that CCh may enhance DSI by

potentiating the depolarization-induced Ca2þ transient, which triggers

the production and release of endocannabinoids. This is also unlikely

because CCh did not significantly influence the basal level of cyto-

plasmic Ca2þ concentration ([Ca2þ]i) (n¼ 15, data not shown) or the

depolarization-induced Ca2þ transient. In the absence of CCh, the

basal levels of [Ca2þ]i at the soma and the dendrite were 98.6� 7.6 nM

and 133.8� 14.1 nM, respectively (n¼ 12). The 1-s depolarization

increased [Ca2þ]i by 56.4� 3.3 nM at the soma and 60.7� 7.0 nM at

the dendrite (n¼ 12). In the presence of CCh (0.3 mM), the peak

amplitude of Ca2þ transient induced by 1-s depolarization was

98.6� 5.3% at the soma (n¼ 12) and 104.1� 7.8% at the dendrite

(n¼ 12), compared with that obtained in the absence of CCh. By

contrast, the 5-s depolarization in the absence of CCh induced a

much larger Ca2þ transient (475.0� 50.1% at the soma, n¼ 8, and

446.9� 60.2% at the dendrite, n¼ 8), indicating that the fluorescent

signal was not saturated. These results strongly suggest that musca-

rinic activation enhances endocannabinoid release presumably by

up-regulating endocannabinoid production downstream of the Ca2þ

elevation.

Discussion

The study presented here has provided evidence for the action of

acetylcholine as a potent enhancer of endocannabinoid release from

postsynatpic neurons. This effect was mimicked by mAChR agonists,

blocked by mAChR antagonists and eliminated by intracellular appli-

cation of GDP-b-S to postsynaptic neurons. The cholinergic enhance-

ment of endocannabinoid signalling was virtually eliminated in

M1/M3-compound-knockout mice. These results clearly indicate that

both the M1 and the M3 receptors on the postsynaptic neurons are in-

volved in the muscarinic enhancement of endocannabinoid signalling.

CCh-induced DSI enhancement in hippocampal slices

Several previous studies have demonstrated that CCh enhances depo-

larization-induced suppression of spontaneous IPSCs (sIPSC-DSI) in

hippocampal slices (Pitler & Alger, 1992a; Martin & Alger, 1999;

Martin et al., 2001). This enhancement has been ascribed to CCh-

induced depolarization of presynaptic GABAergic interneurons (Mar-

tin et al., 2001). It is known that GABAergic neurons are hetero-

geneous in expression of terminal CB1 receptors and the resulting DSI

sensitivity (Ohno-Shosaku et al., 1998, 2001; Martin et al., 2001;

Wilson et al., 2001). Thus, the selective activation of DSI-sensitive

neurons by CCh could increase the DSI-sensitive component of

sIPSCs, resulting in the apparent enhancement of sIPSC-DSI.

Very recently, Kim et al. (2002) reported that CCh (0.2–0.5mM)

significantly enhanced DSI of stimulus-evoked IPSCs (eIPSC) in CA1

pyramidal neurons of rat hippocampal slices. This DSI enhancement

was blocked by atropine and the cannabinoid antagonist AM251, and

was not associated with an increase in depolarization-induced Ca2þ

transients. They also confirmed that both DSI and the CCh-induced

enhancement of DSI were deficient in CB1-knockout mice. All of

these findings are consistent with our results and suggest that mus-

carinic activation enhances endocannabinoid signalling. Evoked

IPSCs recorded from a single CA1 neuron in slices are usually derived

from multiple presynaptic interneurons including cannabinoid-sensi-

tive and cannabinoid-insensitive populations. This raises the possibi-

lity that the magnitude of the eIPSC-DSI may not directly reflect the

amount of released endocannabinods. It is well established that

GABAergic basket interneurons in the hippocampus consist of two

distinct populations: cholecystokinin (CCK)-positive neurons, which

abundantly express CB1 receptors on their axon terminals, and par-

valbumin (PV)-positive neurons, which lack CB1 receptors (Katona

et al., 1999; Hajos et al., 2000). The CCK-positive and PV-positive

basket cells have been thought to correspond to the DSI-sensitive and

DSI-insensitive GABAergic neurons, respectively (Wilson et al., 2001;

Wilson & Nicoll, 2002). If CCh excites DSI-sensitive neurons pre-

ferentially, as previously suggested (Martin et al., 2001), the DSI-

sensitive component of eIPSCs would increase and the DSI magnitude

would apparently become larger. Alternatively, if CCh preferentially

suppresses the DSI-insensitive component of eIPSCs, DSI would also

be enhanced apparently. Immunocytochemical studies of the hippo-

campus show that M2 receptors are located in both cholinergic and

non-cholinergic terminals (Rouse et al., 2000). It is also reported that

PV coexists with M2 receptors in many of the nerve terminals on

pyramidal cell somata, whereas CCK is rarely colocalized with M2

receptors (Hajos et al., 1998). It is therefore likely that CCh suppresses

preferentially the DSI-insensitive component of eIPSCs through the

activation of M2 receptors on the PV-positive basket cells.

In this study, we have measured unitary IPSCs evoked by single

presynaptic neurons and have demonstrated unequivocally that DSI is

markedly enhanced by postsynaptic activation of mAChRs. We have

also estimated quantitatively the amount of endocannabinoids released

from depolarized neurons before and during CCh application. There-

fore, the present study has extended the recent results obtained in slice

preparation (Kim et al., 2002) and represents an important contribution

to our knowledge of the muscarinic enhancement of endocannabinoid

signalling.

Depression of IPSCs by muscarinic activation

Several previous studies have shown that inhibitory synaptic transmis-

sion is depressed by CCh or other cholinergic/muscarinic agonists in

the hippocampus and other brain regions (Sugita et al., 1991; Pitler &

Alger, 1992a; Behrends & Ten Bruggencate, 1993; Kimura, 2000). We

also observed that application of cholinergic agonists caused a

decrease in IPSC amplitudes in some neuron pairs, although the extent

of the reduction varied from pair to pair. The reduction of IPSC

amplitude is associated with a decrease in mIPSC frequency (Behrends

& Ten Bruggencate, 1993) but not with changes in the mIPSC

amplitude (Behrends & Ten Bruggencate, 1993) nor with postsynaptic

sensitivity to GABA (Sugita et al., 1991; Behrends & Ten Bruggen-

cate, 1993). These results suggest that the suppression is due to the

reduction in transmitter release. According to a classic view, it has

simply been assumed that cholinergic agonists act directly on pre-

synaptic mAChRs and suppress the transmitter release. However, there

is an alternative possibility that cholinergic agonists activate post-

synaptic mAChRs and suppress GABA release indirectly via endo-

cannabinoid-mediated retrograde signalling. A recent study (Kim

et al., 2002) has suggested that both these mechanisms may be

involved in the CCh-induced suppression of eIPSC.

Possible physiological significance of endocannabinoidsignalling

It has been reported that activation of postsynaptic group I mGluRs

facilitates endocannabinoid signalling (Maejima et al., 2001; Varma

et al., 2001; Ohno-Shosaku et al., 2002a). Both M1/M3 receptors and

group I mGluRs couple to PLC through Gaq/11 and share intracellular

cascades. This overlap of the signalling cascade suggests that other

neurotransmitters/modulators activating PLC-coupled receptors can

also facilitate endocannabinoid signalling.

It is now widely accepted that endocannabinoid signalling is an

important mechanism by which the activity of postsynaptic neurons



can retrogradely influence presynaptic functions (Kreitzer & Regehr,

2001a, b; Maejima et al., 2001; Ohno-Shosaku et al., 2001; Varma

et al., 2001; Wilson & Nicoll, 2001; Diana et al., 2002; Ohno-Shosaku

et al., 2002a, b; Yoshida et al., 2002). Postsynaptic depolarization and

activation of group I mGluRs or M1/M3 receptors co-operate to trigger

endocannabinoid release. The released endocannabinoids in turn

suppress the transmitter release from presynaptic terminals through

activating the presynaptic CB1 receptors. This form of endocannabi-

noid-mediated retrograde suppression is reversible and short lasting.

By contrast, recent studies suggest that the endocannabinoid system is

also involved in long-term synaptic plasticity. In one of these studies,

the induction of long-term potentiation in the hippocampus was

facilitated by endocannabinoid released during DSI (Carlson et al.,

2002). In others, long-term depression in the striatum (Gerdeman et al.,

2002; Robbe et al., 2002) and basolateral amygdala (Marsicano et al.,

2002) was suppressed by cannabinoid antagonists and was deficient in

CB1-knockout mice. Furthermore, the extinction of certain forms of

learning is reported to be impaired in CB1-knockout mice (Marsicano

et al., 2002; Varvel & Lichtman, 2002). These studies indicate that the

endocannabinoid system is involved in certain aspects of learning and

memory. It has been well established that the cholinergic system is

involved in cognition, learning and memory (Bartus et al., 1982;

Nilsson et al., 1992; Aigner, 1995; Jerusalinsky et al., 1997; Segal

& Auerbach, 1997). In view of these findings, the results presented

here suggest the possibility that various higher brain functions exerted

by the cholinergic system may involve the modulation of endocanna-

binoid signalling.

Acknowledgements

We thank T. Tabata for valuable comments on the manuscript and Y. Araki, S.Kobayashi, N. Matsubara, H. Karasawa, S. Takahashi and D. Motomura fortechnical assistance in generating M1 and M3 knockout mice. This research wassupported by grants-in aid for scientific research (T.O., M.M., M.M.T., T.M. andM.K.) and special coordination funds for promoting science and technology(T.M. and M.K.) from the Ministry of Education, Science, Sports, Culture andTechnology of Japan, by the Sumitomo Foundation (T.O. and T.H.), by the CellScience Research Foundation (M.K.), by Vehara Memorial Foundation (T.M.),by Terumo Life Science Foundation (T.M.), by RISTEX (T.M.), by JapanScience and Technology Corporation (JST; T.M.), by an industrial technologyresearch grant programme from the New Energy and Industrial TechnologyDevelopment Organization (NEDO) of Japan (M.M.) and by a grant from theOrganization for Pharmaceutical Safety and Research, Japan (M.M.T.).

Abbreviations

[Ca2þ]i, intracellular Ca2þ concentration; CB1, cannabinoid receptor type 1;CCh, carbachol; DHPG, (RS)-3,5-dihydroxyphenylglycine; DSI, depolariza-tion-induced suppression of inhibition; GDP-b-S, guanosine 50-O-(2-thio-diphosphate); IPSC, inhibitory postsynaptic current; mAChR, muscarinicacetylcholine receptor; mGluR, metabotropic glutamate receptor; PLC, phos-pholipase C.

References

Aigner, T.G. (1995) Pharmacology of memory: cholinergic–glutamatergicinteractions. Curr. Opin. Neurobiol., 5, 155–160.

Bartus, R.T., Dean, R.L., 3rd, Beer, B. & Lippa, A.S. (1982) The cholinergichypothesis of geriatric memory dysfunction. Science, 217, 408–414.

Behrends, J.C. & Ten Bruggencate, G. (1993) Cholinergic modulation ofsynaptic inhibition in the guinea pig hippocampus in vitro: excitation ofGABAergic interneurons and inhibition of GABA-release. J. Neurophysiol.,69, 626–629.

Carlson, G., Wang, Y. & Alger, B.E. (2002) Endocannabinoids facilitate theinduction of LTP in the hippocampus. Nature Neurosci., 5, 723–724.

Caulfield, M.P. (1993) Muscarinic receptors – characterization, coupling andfunction. Pharmacol. Ther., 58, 319–379.

Caulfield, M.P. & Birdsall, N.J. (1998) International Union of Pharmacology.XVII. Classification of muscarinic acetylcholine receptors. Pharmacol. Rev.,50, 279–290.

Cole, A.E. & Nicoll, R.A. (1984) The pharmacology of cholinergic excitatoryresponses in hippocampal pyramidal cells. Brain Res., 305, 283–290.

Diana, M.A., Levenes, C., Mackie, K. & Marty, A. (2002) Short-term retrogradeinhibition of GABAergic synaptic currents in rat Purkinje cells is mediated byendogenous cannabinoids. J. Neurosci., 22, 200–208.

Felder, C.C., Bymaster, F.P., Ward, J. & DeLapp, N. (2000) Therapeuticopportunities for muscarinic receptors in the central nervous system. J.Med. Chem., 43, 4333–4353.

Gerdeman, G.L., Ronesi, J. & Lovinger, D.M. (2002) Postsynaptic endocan-nabinoid release is critical to long-term depression in the striatum. NatureNeurosci., 5, 446–451.

Hajos, N., Katona, I., Naiem, S.S., MacKie, K., Ledent, C., Mody, I. & Freund,T.F. (2000) Cannabinoids inhibit hippocampal GABAergic transmission andnetwork oscillations. Eur. J. Neurosci., 12, 3239–3249.

Hajos, N., Papp, E.C., Acsady, L., Levey, A.I. & Freund, T.F. (1998) Distinctinterneuron types express m2 muscarinic receptor immunoreactivity on theirdendrites or axon terminals in the hippocampus. Neuroscience, 82, 355–376.

Hamilton, S.E., Loose, M.D., Qi, M., Levey, A.I., Hille, B., McKnight, G.S.,Idzerda, R.L. & Nathanson, N.M. (1997) Disruption of the m1 receptor geneablates muscarinic receptor-dependent M current regulation and seizureactivity in mice. Proc. Natl Acad. Sci. USA, 94, 13311–13316.

Hulme, E.C., Birdsall, N.J. & Buckley, N.J. (1990) Muscarinic receptorsubtypes. Annu. Rev. Pharmacol. Toxicol., 30, 633–673.

Irving, A.J. & Collingridge, G.L. (1998) A characterization of muscarinicreceptor-mediated intracellular Ca2þ mobilization in cultured rat hippocam-pal neurones. J. Physiol. (Lond.), 511, 747–759.

Jerusalinsky, D., Kornisiuk, E. & Izquierdo, I. (1997) Cholinergic neurotrans-mission and synaptic plasticity concerning memory processing. Neurochem.Res., 22, 507–515.

Kano, M., Ohno-Shosaku, T. & Maejima, T. (2002) Retrograde signaling atcentral synapses via endogenous cannabinoids. Mol. Psychiatry, 7, 234–235.

Katona, I., Sperlagh, B., Sik, A., Kafalvi, A., Vizi, E.S., Mackie, K. & Freund,T.F. (1999) Presynaptically located CB1 cannabinoid receptors regulateGABA release from axon terminals of specific hippocampal interneurons.J. Neurosci., 19, 4544–4558.

Kim, J., Isokawa, M., Ledent, C. & Alger, B.E. (2002) Activation of muscarinicacetylcholine receptors enhances the release of endogenous cannabinoids inthe hippocampus. J. Neurosci., 22, 10182–10191.

Kimura, F. (2000) Cholinergic modulation of cortical function: a hypotheticalrole in shifting the dynamics in cortical network. Neurosci. Res., 38, 19–26.

Kreitzer, A.C. & Regehr, W.G. (2001a) Cerebellar depolarization-inducedsuppression of inhibition is mediated by endogenous cannabinoids. J.Neurosci., 21, RC174.

Kreitzer, A.C. & Regehr, W.G. (2001b) Retrograde inhibition of presynapticcalcium influx by endogenous cannabinoids at excitatory synapses ontoPurkinje cells. Neuron, 29, 717–727.

Kreitzer, A.C. & Regehr, W.G. (2002) Retrograde signaling by endocannabi-noids. Curr. Opin. Neurobiol., 12, 324–330.

Levey, A.I., Edmunds, S.M., Koliatsos, V., Wiley, R.G. & Heilman, C.J. (1995)Expression of m1-m4 muscarinic acetylcholine receptor proteins in rat hippo-campusandregulationbycholinergic innervation.J. Neurosci.,15, 4077–4092.

Llano, I., Leresche, N. & Marty, A. (1991) Calcium entry increases thesensitivity of cerebellar Purkinje cells to applied GABA and decreasesinhibitory synaptic currents. Neuron, 6, 565–574.

Maejima, T., Hashimoto, K., Yoshida, T., Aiba, A. & Kano, M. (2001)Presynaptic inhibition caused by retrograde signal from metabotropic glu-tamate to cannabinoid receptors. Neuron, 31, 463–475.

Marsicano, G., Wotjak, C.T., Azad, S.C., Bisogno, T., Rammes, G., Cascio,M.G., Hermann, H., Tang, J., Hofmann, C., Zieglgansberger, W., Di Marzo,V. & Lutz, B. (2002) The endogenous cannabinoid system controls extinctionof aversive memories. Nature, 418, 530–534.

Martin, L.A. & Alger, B.E. (1999) Muscarinic facilitation of the occurrence ofdepolarization-induced suppression of inhibition in rat hippocampus. Neu-roscience, 92, 61–71.

Martin, L.A., Wei, D.S. & Alger, B.E. (2001) Heterogeneous susceptibility ofGABA (A) receptor-mediated IPSCs to depolarization-induced suppressionof inhibition in rat hippocampus. J. Physiol. (Lond.), 532, 685–700.

Matsui, M., Araki, Y., Karasawa, H., Matsubara, N., Taketo, M.M. & Seldin,M.F. (1999) Mapping of five subtype genes for muscarinic acetylcholinereceptor to mouse chromosomes. Genes Genet. Syst., 74, 15–21.

Matsui, M., Motomura, D., Karasawa, H., Fujikawa, T., Jiang, J., Komiya,Y., Takahashi, S. & Taketo, M.M. (2000) Multiple functional defects



in peripheral autonomic organs in mice lacking muscarinic acetyl-choline receptor gene for the M3 subtype. Proc. Natl Acad. Sci. USA, 97,9579–9584.

Nicoll, R.A. (1985) The septo-hippocampal projection: a model cholinergicpathway. Trends Neurosci., 8, 533–536.

Nicoll, R.A., Malenka, R.C. & Kauer, J.A. (1990) Functional comparison ofneurotransmitter receptor subtypes in mammalian central nervous system.Physiol. Rev., 70, 513–565.

Nilsson, O.G., Leanza, G., Rosenblad, C., Lappi, D.A., Wiley, R.G. & Bjork-lund, A. (1992) Spatial learning impairments in rats with selective immu-nolesion of the forebrain cholinergic system. Neuroreport, 3, 1005–1008.

Ohno-Shosaku, T., Maejima, T. & Kano, M. (2001) Endogenous cannabinoidsmediate retrograde signals from depolarized postsynaptic neurons to pre-synaptic terminals. Neuron, 29, 729–738.

Ohno-Shosaku, T., Sawada, S. & Yamamoto, C. (1998) Properties of depolar-ization-induced suppression of inhibitory transmission in cultured rat hip-pocampal neurons. Pflugers Arch., 435, 273–279.

Ohno-Shosaku, T., Shosaku, J., Tsubokawa, H. & Kano, M. (2002a) Coopera-tive endocannabinoid production by neuronal depolarization and group Imetabotropic glutamate receptor activation. Eur. J. Neurosci., 15, 953–961.

Ohno-Shosaku, T., Tsubokawa, H., Mizushima, I., Yoneda, N., Zimmer, A. &Kano, M. (2002b) Presynaptic cannabinoid sensitivity is a major determinantof depolarization-induced retrograde suppression at hippocampal synapses.J. Neurosci., 22, 3864–3872.

Pitler, T.A. & Alger, B.E. (1992a) Cholinergic excitation of GABAergicinterneurons in the rat hippocampal slice. J. Physiol. (Lond.), 450, 127–142.

Pitler, T.A. & Alger, B.E. (1992b) Postsynaptic spike firing reduces synapticGABAA responses in hippocampal pyramidal cells. J. Neurosci., 12,4122–4132.

Pitler, T.A. & Alger, B.E. (1994) Depolarization-induced suppression ofGABAergic inhibition in rat hippocampal pyramidal cells: G protein invol-vement in a presynaptic mechanism. Neuron, 13, 1447–1455.

Robbe, D., Kopf, M., Remaury, A., Bockaert, J. & Manzoni, O.J. (2002)Endogenous cannabinoids mediate long-term synaptic depression in thenucleus accumbens. Proc. Natl Acad. Sci. USA, 99, 8384–8388.

Rouse, S.T., Edmunds, S.M., Yi, H., Gilmor, M.L. & Levey, A.I. (2000)Localization of M (2) muscarinic acetylcholine receptor protein in choli-

nergic and non-cholinergic terminals in rat hippocampus. Neurosci. Lett.,284, 182–186.

Rouse, S.T., Marino, M.J., Potter, L.T., Conn, P.J. & Levey, A.I. (1999)Muscarinic receptor subtypes involved in hippocampal circuits. Life Sci.,64, 501–509.

Segal, M. & Auerbach, J.M. (1997) Muscarinic receptors involved in hippo-campal plasticity. Life Sci., 60, 1085–1091.

Soriano, P., Montgomery, C., Geske, R. & Bradley, A. (1991) Targeteddisruption of the c-src proto-oncogene leads to osteopetrosis in mice. Cell,64, 693–702.

Sugita, S., Uchimura, N., Jiang, Z.G. & North, R.A. (1991) Distinct muscarinicreceptors inhibit release of gamma-aminobutyric acid and excitatory aminoacids in mammalian brain. Proc. Natl Acad. Sci. USA, 88, 2608–2611.

Varma, N., Carlson, G.C., Ledent, C. & Alger, B.E. (2001) Metabotropicglutamate receptors drive the endocannabinoid system in hippocampus. J.Neurosci., 21, RC188.

Varvel, S.A. & Lichtman, A.H. (2002) Evaluation of CB1 receptor knockoutmice in the Morris water maze. J. Pharmacol. Exp. Ther., 301, 915–924.

Vilaro, M.T., Mengod, G., Palacios, G. & Palacios, J.M. (1993) Receptordistribution in the human and animal hippocampus: focus on muscarinicacetylcholine receptors. Hippocampus, 3 (Spec. No.), 149–156.

Wess, J. (1996) Molecular biology of muscarinic acetylcholine receptors. Crit.Rev. Neurobiol., 10, 69–99.

Wilson, R.I., Kunos, G. & Nicoll, R.A. (2001) Presynaptic specificity ofendocannabinoid signaling in the hippocampus. Neuron, 31, 453–462.

Wilson, R.I. & Nicoll, R.A. (2001) Endogenous cannabinoids mediate retro-grade signalling at hippocampal synapses. Nature, 410, 588–592.

Wilson, R.I. & Nicoll, R.A. (2002) Endocannabinoid signaling in the brain.Science, 296, 678–682.

Yagi, T., Ikawa, Y., Yoshida, K., Shigetani, Y., Takeda, N., Mabuchi, I.,Yamamoto, T. & Aizawa, S. (1990) Homologous recombination at c-fynlocus of mouse embryonic stem cells with use of diphtheria toxin A-fragmentgene in negative selection. Proc. Natl Acad. Sci. USA, 87, 9918–9922.

Yoshida, T., Hashimoto, K., Zimmer, A., Maejima, T., Araishi, K. & Kano, M.(2002) The cannabinoid CB1 receptor mediates retrograde signals fordepolarization-induced suppression of inhibition in cerebellar Purkinje cells.J. Neurosci., 22, 1690–1697.



Documents

Postsynaptic M1 and M3 receptors are responsible for the muscarinic enhancement of retrograde endocannabinoid signalling in the hippocampus