Since its discovery in 1973, synaptic plasticity has garnered the attention and imagination of neuroscientists as a potential cellular basis of learning, memory and development. As a result, the intense study devoted to the prototypic CA1 long-term potentiation (LTP) in the hippocampus has generated an unparalleled amount of knowledge and made it our best understood model of activity-dependent plasticity in the brain (see here for a discussion of the postsynaptic mechanisms of synaptic plasticity). Thus, it comes as a surprise that controversy still divides the field as to whether presynaptic changes are also involved in its induction and expression. Interest in presynaptic mechanisms of synaptic plasticity has resurged in the past decade as forms of synaptic plasticity that are predominantly presynaptically induced and/or expressed were discovered (Citri & Malenka, 2008). It is now understood that presynaptic changes, such as activity- or Ca2+-dependent modifications of release machinery and retrograde signaling, can powerfully influence synaptic efficacy. As research in the field is very much still in its infancy, only the presently two best studied models of presynaptic plasticity, mossy fibre LTP and endocannabinoid LTD, will be discussed in some detail pertaining to their molecular mechanisms and functional significances. Nevertheless, a review of presynaptic involvements in behaviourally-relevant forms of synaptic plasticity is crucial to characterizing it as yet another site amenable to activity-dependent modification, and will add to our understanding of the brain’s capacity for information storage and intricate cognition.



Mossy Fibre LTP


High-frequency stimulation can also elicit a form of LTP in the CA3 of the hippocampus, in the synapse between the dentate gyrus granule cell mossy fibres and the CA3 pyramidal neurons (Fig. 1) (Nicoll & Malenka, 1995). Unlike its CA1 counterpart, the induction and expression of the thus named mossy fibre LTP (mf-LTP) are predominantly presynaptic and do not require NMDA receptor activation.


external image F1.large.jpg
Figure 1. Schematic of the hippocampal trisynaptic circuit.

Activity-dependent activation of presynaptic glutamatergic autoreceptors

MF-LTP.png

Figure 2. Schematic of mf-LTP induction. (Nistico et al., 2011).

Current model of mf-LTP induction (Fig.2) holds that repeated stimulation of the presynaptic terminal first opens L-type voltage-gated Ca2+ channels to allow a rapid Ca2+ influx that elicits glutamate release. If successful, the high ambient level of glutamate in the synaptic cleft would consequently activate the group I metabotropic glutamate autoreceptors (mGluR1 and mGluR5) and Ca2+-permeant GluK1 kainate autoreceptors (KAR) found on the presynaptic terminal. Through their coupled G proteins, mGluR1 and mGluR5 activate the production of inositol triphosphate (IP3) from membrane phospholipids, which then diffuse through the cytoplasm to opent the ligand-gated Ca2+ channels on the endoplasmic reticulum (ER) membrane. This releases the Ca2+ ions stored in the ER, serving to elevate intracellular Ca2+ levels (Laurie et al., 2003). Opening of the ionotropic KARs further enhance the Ca2+ transient (Lerma, 2003). As a result of this positive feedback loop of glutamate release, autoreceptor activation and rise in intracellular Ca2+ that further enhances glutamate release, the massive elevation in intracellular Ca2+ activates Ca2+/calmodulin-sensitive adenylyl cyclase (AC) and consequently up-regulates the activity of the cAMP-dependent protein kinase A (PKA). As PKA has a number of substrates that are key players in the active zone exocytotic machinery (Nisticò et al., 2009), it is very likely that the enhancement of its activity would ultimately result in an increase in the probability of neurotransmitter release, thus strengthening neurotransmission at the synapse.

Enhanced release probability through cAMP/PKA-dependent modification of active zone proteins


external image ne270509.f5.jpeg
Figure 3. Putative protein-protein interactions at the active zone mediating synaptic vesicle attachment. (Südhof, 2004).

The expression of mf-LTP is believed to involved cAMP/PKA dependent modification of an array of synaptic vesicle docking and release proteins, as summarized in Figure 3. Presently the best characterized candidate molecule that mediates the expression of mf-LTP is RIM1α, an active zone scaffolding protein responsible for priming docked synaptic vesicles for release (Han, Kaeser, Sudhof & Schneggenburger, 2011). A substantial body of evidences has shown that RIM1α is critical for regulating the probability of release. First, RIM1α knockout mice display a 50% reduction in the probability of release at the synapses between the Schaffer collateral and the CA1 region (Schoch et al., 2002) and consequently do not display mf-LTP (Castillo et al., 2002). Calakos and coworkers (2004) later revealed that these knockout mice have a smaller readily releaseable pool of synaptic vesicles, possibly due to the loss of RIM1α-mediated organization. A similar role has been demonstrated in C. elegans (Koushika et al, 2001). In addition, in vitro studies have demonstrated that RIM1α interacts with many other active zone and synaptic vesicle proteins critical for release, including the Ca2+ sensor for vesicle exocytosis synaptomtagmin 1, voltage-gated Ca2+ channels and the SNARE complex protein SNAP25 (Coppola et al., 2001). In particular, RIM1α’s interaction with the synaptic vesicle protein Rab3A is required for mf-LTP (Lonart et al., 1998). Finally, RIM1α is positively regulated by Ca2+ influx and PKA phosphorylation (Lonart et al., 2003), making it very likely that it lies downstream of the cAMP/PKA pathway to ultimately enhance the stimulation-secretion coupling of neurotransmitter release. Recently, Yang and Calakos (2011) have identified Munc13-1, yet another priming factor that is required for efficient vesicle exocytosis (Ho, Lee & Martin, 2011), as a new mediator of mf-LTP through its interaction with RIM1α. While it has long been known that the Munc13 family of proteins are essential for neurotransmission (Rizo & Rosenmund, 2008), their roles in long-term synaptic plasticity ironically remain poorly understood, precisely because their importance in mediating synaptic transmission made knockout models inviable (Breustedt et al., 2010). Based on previous findings that RIM1α bind to the regulatory N-terminus C2A domain of Munc13-1 (Deng, Kaeser, Xu & Sudhof, 2011), the authors employed a novel molecular technique by stereotaxically injecting viral vectors containing cDNAs of truncated RIM1α or Munc13-1sequences into the mice’s brains (Yang & Calakos, 2010), thus disrupting the protein interactions but leaving intact their individual wild-type functions. It was found that while mf-LTP was abolished in animals expressing mutated copies of either protein, restoring this protein interaction in RIM1α knockout mice was both required and sufficient to rescue it. The extensive involvements of synaptic vesicle priming proteins in mf-LTP induction and expression thus suggest that vesicle priming is a major control point in the probability of neurotransmitter release. As priming is the process where vesicles and SNARE proteins are brought into close contact so that Ca2+ influx can elicit their release, it is conceivable that the amount of primed vesicles would be proportional to the probability of release (Hanse & Gustafsson, 2001). Such an hypothesis fits nicely with observations of increased quantal content (Xiang et al., 1994) and glutamate release (Kawamura et al., 2004), and decreased release failure rates (Lopez-Garcia et al., 1996) in synapses expressing mf-LTP.

Yang and Calakos’ work is one of several recent studies that suggest a resurgent interest in mf-LTP. Curiously, since the mid-2000’s there has been a stall in progresses made in the field, as seen by the rather dated references cited even in recent reviews by Mellor (2010) and Kobayashi (2010), possibly due to technical limitations that made examination of the presynaptic terminal very difficult. However, with the advent of molecular and imaging techniques, several papers published in the last year have taken to once again address the mechanisms underlying this phenomenon. The 2011 study by Nistico and coworks is especially noteworthy for its elucidation of synergistic interactions between mGluRs and GluK1 KARs in the induction of mf-LTP. While presynaptic glutamate autoreceptors are generally believed to be involved in the induction of mf-LTP, there have been long-standing controversies over which ones are involved and just how important their really are. Both mGluR antagonists and knockouts have given mixed results, with some studies reporting impairment of mf-LTP induction (Yeckel et al., 1999; Hsia et al., 1995) and others failing to find any effect (Mellor & Nicoll, 2001; Conquet et al., 1994). The role of GluK1 KARs have also been unclear due to the different ambient Ca2+ levels used in the different experiment, for it has been found that its absence can be compensated for by L-type voltage-gated Ca2+ channels (Lauri et al., 2003) if the bath has a high enough (4mM) Ca2+ concentration. Nistico and coworkers used the granule cell morphological marker Alexa 594 and Ca2+-sensitive dye Fluo-4, techniques that were only developed recently (Dargan et al., 2009), to trace mossy fiber axons into the CA3 and study the effects that various glutamate receptor antagonists had on Ca2+ transients evoked by repeated stimulation of the presynaptic terminal. It was found that mf-LTP was abolished only when GluK1 KARs and both mGluR1 and mGluR5 were simultaneously blocked by their respective antagonists, reflected in a significant reduction in the size of Ca2+ influx observed after evoking five action potentials at 20 Hz. However, mf-LTP was not affected if either mGluRs remain active, prompting the authors to conclude that the activation of mGluR1 or mGluR5 are interchangeable as long as GluK1 KARs are also activated. More interestingly, blockage of both mGluR1 and mGluR5 by their respective antagonists MPEP and LY367385 preferentially depressed Ca2+ transients evoked by action potentials that occurred later in the train, completely supporting the autoreceptor functions that have been proposed for these glutamate receptors. Thus, existing findings may now be re-examined in light of this co-activation requirement, for divergent experimental conditions could have accounted for their apparently contradictory nature.

Role in working memory and pattern recognition

Behaviourally, it has been proposed that mf-LTP might mediate working memory functions that facilitate pattern recognition and reactivity to novel environmental stimuli. While disruption of mf-LTP did not seem to impair the animals’ ability to acquire learning, D’Adamo and coworkers (2004) have found that Rab3A knockout mice displayed impaired spatial working memory and reference memory, which stores information about consistent patterns that might have been detected. The dentate gyrus has been implicated in neuronal pattern detection, for it has been observed that even small changes in the environment can significantly alter activity patterns among place-modulated granule cells in exploring rats (Leutgeb et al., 2007). Consequently, the dentate gyrus may input such recognized patterns directly into the CA3 to facilitate associative learning (Myers & Scharfman, 2011). Recently it was reported that even though Rab3A knockout mice had no problem acquiring the freezing response in a contextual fear conditioning paradigm, the behaviour was no longer context-specific when tested the day after training (Ruediger et al., 2011). Thus this “cognitive flexibility” may possibly be an evolutionary measure to allow rapid acquisition of avoidance behaviour to aversive stimuli in the environment, which ultimately serves to promote survival of the organism.


eCB-LTD


Endocannabinoids (eCB) are a family of naturally-occurring neuromodulatory molecules that are thus named because they activate the G protein-coupled cannabinoid receptors that mediate the effects of delta-9-tetrahydrocannabinol (THC), the major psychoactive ingredient of marijuana (De Petrocellis & Di Marzo, 2009). The two most common eCBs in the body are arachidonoyl ethanolamide (anadamide) and 2-arachidonoyl glycerol (2-AG), the latter being the most abundant in the brain (Ohno-Shosaku et al., 2011). As retrograde signaling molecules, eCBs are produced and released postsynaptically in an activity-dependent manner to act most commonly on the type 1 cannabinoid (CB1) receptors on the presynaptic terminal (Kawamura et al., 2006). Since their discovery (Devane et al., 1992), evidences quickly emerged demonstrating eCBs’ importance in a variety of processes such as analgesia, appetite and neurogenesis (Maccarrone, 2010).


Figure 4. The mechanism of endocannabinoid action.

However, studies into eCB’s involvements in synaptic plasticity did not begin until 2002, when Gerdeman and coworkers identified 2-AG as the retrograde messenger that reconciled the postsynaptic induction but presynaptic expression of a form of long-term depression (LTD) in the dorsal striatum (Choi & Lovinger, 1997). Since then, mechanistically similar forms of eCB-LTD have been observed in glutamatergic and GABAergic synapses (Sidhpura & Parsons, 2011) throughout the brain, suggesting it to be an important mechanism for rapid feedback signaling and stabilization of neural circuits.

Activity-dependent endocannabinoid mobilization and retrograde signalling

ECB.png
Figure 5. Induction and expression of eCB-LTD. (Ohno-Shosaku et al., 2011)

For eCB-LTD research is still in its infancy, it is not yet clear why induction protocols seem to differ depending on the brain region examined (Adermark & Lovinger, 2009). Nevertheless, a tentative induction mechanism has been proposed (Fig.5) (Ohno-Shosaku et al., 2011) based on evidences demonstrating that eCB-LTD induction requires several pre- and postsynaptic events to occur within a narrow temporal window. First, repetitive activation of the presynaptic terminal results in the release of glutamate that activates the group I metabotropic glutamate receptors (mGluR1/5) on the postsynaptic terminal. Through their couplings with Gq/11 proteins, the mGluRs then activate a downstream signaling pathway that has to converge with a concurrent elevation in postsynaptic Ca2+ concentration for 2-AG production to occur (Di Marzo & Matias, 2005), thus providing associativity between presynaptic neurotransmitter release and postsynaptic depolarization. Then through unknown transport mechanisms, for it is lipophilic and thus cannot be stored in vesicles, 2-AG is released to diffuse across the synaptic cleft and bind to presynaptic CB1 receptors. While the activated CB1 receptors can act alone through their coupling with Gi/o family of G proteins to inhibit AC and thus PKA activity (Turu & Hunyady, 2010), evidences show that its activation has to coincide with a rise in presynaptic Ca2+ for eCB-LTD to occur. This integration of postsynaptic eCB release and presynaptic activity serves two purposes. First, it confers additional synaptic specificity to eCB-LTD induction, for eCB can diffuse to act on nearby presynaptic terminals (Heifets, Chevaleyre & Castillo, 2008). But also, the Ca2+ transient is required to activate the Ca2+-sensitive phosphatase calcineurin (CaN) (Singla, 2007), which in conjunction with the inhibition of PKA activity would tilt the kinase-phosphatase balance toward dephosphorylation and inactivation of presynaptic targets.

CB1 receptor-dependent modification of presynaptic ion channels and release machinery

Two very likely targets of dephosphorylation are presynaptic ion channels and release machinery proteins, for common to all forms of eCB-LTD is a persistent reduction in
the probability of neurotransmitter release (Mackie, 2006). CB1 receptor-dependent inhibition of voltage-gated Ca2+ channels have been described in the calyx of Held synapse, where a depression of presynaptic Ca2+ current can be recorded in the large presynaptic terminal after CB1 receptor activation (Kushmerick et al., 2004). Conversely, the activation of A-type K+ channels (Mackie et al., 1995) and inwardly rectifying K+ channels (Kreitzer, Carter & Regehr, 2002) have also been demonstrated. Together, these changes likely weaken the excitation-secretion coupling by hyperpolarizing the presynaptic terminal and depressing the amplitude and duration of the Ca2+ transient. Interestingly, perhaps indicative of its importance in regulating neurotransmitter release, RIM1α is also required for the induction of eCB-LTD in the hippocampus and amygdala (Chevaleyre et al., 2007). As discussed in the previous section, RIM1α’s role in mediating in vesicle priming and exocytosis is positively regulated by Ca2+ influx and PKA phosphorylation to result in enhanced neurotransmitter release and LTP (Lonart et al., 2003). Thus it is very likely that the reduction in Ca2+ current and PKA activity would negatively regulate RIM1α activity during eCB-LTD to inhibit vesicle release (Chevaleyre & Castillo, 2003). However, further study is warranted to elucidate RIM1α’s precise role as Kaeser and coworkers (2008) had demonstrated that, contrary to expectations, phosphorylation of RIM1α at the serine-413 residue is not required for eCB-LTD induction in the hippocampus.

Role in extinction learning and addiction

The functional significances of eCB-LTD have gained recent interest as the importance of experience-dependent plasticity in inhibitory synapses is beginning to emerge, for eCB-mediated depression of transmission at GABAergic synapses can disinhibit postsynaptic cells and facilitate the induction of LTP during learning (Lafourcade, 2009), therefore providing an entirely new level of synaptic regulation. However, while eCB-LTD has been implicated in specific behaviours, interpretation of the evidences is difficult for its roles seem to vary according to the nature of the task, stage of learning and level of CB1 receptor activation. Administrations of CB1 receptor agonist and antagonists in the basolateral amygdala have been shown to be able to bidirectionally regulate the acquisition of fear memory, demonstrating CB1 receptor function to be essential in this form of learning (Tan et al., 2011). Similarly, depression of CB1 receptor function with the inverse agonist rimonabant prevented the acquisition of reward-related memory (Yu et al., 2011). However, in a modified version of the Morris water maze task, Niyuhire and coworkers (2007) have shown that while acquisition of spatial memory was intact in CB1 receptor knockout mice, they were unable to suppress this learning when they had to find the platform again after it had been moved to a new location. This deficit in extinction learning has also been reported in aversive and fear learning models, using both CB1 receptor knockout mice and administration of CB1 receptor antagonists (Lin et al., 2009). Abush and Akirav (2010) provided a possible explanation for these conflicting evidences by demonstrating that the effects of CB1 receptor activation and inhibition on spatial learning do not follow a simple pattern of up- and down-regulation, respectively. Thus, the eCB system seems to have a much more intricate effect on synaptic connectivity and cognition, warranting further study. The inconsistencies have also evoked criticisms that no studies have been done to examine whether the stimulus frequencies used in in vitro induction of eCB-LTD are in fact physiological (Heifets & Castillo, 2009). If eCB-LTD is entirely a laboratory phenomenon, then such observations only illustrate aspects of CB1 receptor function rather than the role of eCB-LTD in learning and memory.

Reward_circuit.jpg
Figure 6. Reward circuit of the human brain.

eCB-LTD is best studied in regions of the brain that are involved in reward and addiction (Fig. 6), such as the ventral tegmental area and nucleus accumbens (Luscher & Malenka, 2011). Thus, given that chronic substance abusers display CB1 receptor desensitization and mGluR5 down-regulation (Pan et al., 2008), it has been hypothesized that dysregulation of the eCB signalling system contributes to the pathological form of associative learning that is addiction (Marsicano & Lafenetre, 2009). In particular, the environmental stimuli that come to be strongly associated with the powerful emotional and physical effects of drug use also gradually become strong cues that induce cravings and drug-seeking behaviour (Haj-Dahmane & Shen, 2010). The dangers of addictive behaviours lie in the inability to extinguish such an association during periods of attempted abstinence, so that encounter of such conditioned environmental cues almost invariably results in relapses (Serrano & Parsons, 2011). As eCB dysfunctions are manifested in animal models as impairments in extinction learning, researchers have hypothesized that a similar deficit in chronic substance users may account for the failure of extinction therapy to weaken the motivational effects of these conditioned cues by repeatedly presenting them in the absence of drug rewards (this drug-induced deficit in plasticity is discussed in further detail from the perspective of impaired metaplasticity). Thus it has been suggested that the restoration of proper eCB functions should be used to supplement extinction therapy in order to more effectively treat addictive behaviour (Sidhpura & Parsons, 2011).


References


Abush, H. & Akirav, I. (2010). Cannabinoids modulate hippocampal memory and plasticity. Hippocampus, 20(10), 1126-1138.

Adermark, L. & Lovinger, D.M. (2009). Frequency-dependent inversion of net striatal output by endocannabinoid-dependent plasticity at different synaptic inputs. J Neurosci, 29(5), 1375-1380.

Breustedt, J. et al. (2010). Munc13-2 differentially affects hippocampal synaptic transmission and plasticity. Cereb Cortex, 20(5), 1109-1120.

Castillo, P.E., Schoch, S., Schmitz, F., Sudhof, T.C. & Malenka, R.C. (2002). RIM1alpha is required for presynaptic long-term potentiation. Nature, 415(6869), 327-330.

Calakos, N. et al. (2004). Multiple roles for the active zone protein RIM1alpha in late stages of neurotransmitter release. Neuron, 42(6), 889-896.

Chevaleyre, V. & Castillo, P.E. (2003). Heterosynaptic LTD of hippocampal GABAergic synapses: a novel role of endocannabinoids in regulating excitability. Neuron, 38(3), 461-472.

Chevaleyre, V. et al. (2007). Endocannabinoid-mediated long-term plasticity requires cAMP/PKA signaling and RIM1alpha. Neuron, 54(5), 801-812.

Choi, S. & Lovinger, D.M. (1997). Decreased frequency but not amplitude of quantal synaptic responses associated with expression of corticostriatal long-term depression. J Neurosci, 17(21), 8613-8620.

Citri, A. & Malenka, R.C. (2008). Synaptic plasticity: multiple forms, functions, and mechanisms. Neuropsychopharmacology, 33(1), 18-41.

Conquet, F. et al. (1994). Motor deficit and impairment of synaptic plasticity in mice lacking mGluR1. Nature, 372(6503), 237-243.

Coppola, T. et al. (2001). Direct interaction of the Rab3 effector RIM with Ca2+ channels, SNAP-25, and synaptotagmin. J Biol Chem, 276(35), 32756-32762.

D’Adamo, P. et al. (2004). Mice deficient for the synaptic vesicle protein Rab3a show impaired spatial reversal learning and increased explorative activity but none of the behavioral changes shown by mice deficient for the Rab3a regulator Gdi1. Eur J Neurosci, 19(7), 1895-1905.

Dargan, S.L. & Amici, M. (2009). Role of kainate autoreceptors in short-term plasticity at hippocampal mossy fiber synapses. J Neurosci, 29(18), 5713-5715.

Deng, L., Kaeser, P.S., Xu, W. & Sudhof, T.C. (2011). RIM proteins activate vesicle priming by reversing autoinhibitory homodimerization of Munc13. Neuron, 69(2), 317-331.

Devane, W.A. et al. (1992). Isolation and structure of a brain constituent that binds to the cannabinoid receptor. Science, 258(5090), 1946-1949.

Di Marzo, V. & Matias, I. (2005). Endocannabinoid control of food intake and energy balance. Nat Neurosci, 8(5), 585-589.

Haj-Dahmane, S. & Shen, R.Y. (2010). Regulation of plasticity of glutamate synapses by endocannabinoids and the cyclic-AMP/protein kinase A pathway in midbrain dopamine neurons. J Phsyiol, 2010,, 588(14), 2589-2604.

Han, Y., Kaeser, P.S., Sudhof, T.C. & Schneggenburger, R. (2011). RIM determines Ca²+ channel density and vesicle docking at the presynaptic active zone. Neuron, 69(2), 304-316.

Hanse, E. & Gustafsson, B. (2001). Vesicle release probability and pre-primed pool at glutamatergic synapses in area CA1 of the rat neonatal hippocampus. J Physiol, 531(2), 481-493.

Heifets, B.D. & Castillo, P.E. (2009). Endocannabinoid signaling and long-term synaptic plasticity. Annu Rev Physio, 2009, 71 283-306.

Heifets, B.D., Chevaleyre, V. & Castillo, P.E. (2008) . Interneuron activity controls endocannabinoid-mediated presynaptic plasticity through calcineurin. Proc Natl Acad Sci U S A, 105(29), 10250-10255.

Ho, V.M., Lee, J.A. & Martin, K.C. (2011). The Cell Biology of Synaptic Plasticity. Science, 334(6056), 623-628.

Hsia, A.Y. et al. (1995). Evidence against a role for metabotropic glutamate receptors in mossy fiber LTP: the use of mutant mice and pharmacological antagonists. Neuropharmacology, 34(11), 1567-1572.

Kaeser, P.S. et al. (2008). RIM1alpha phosphorylation at serine-413 by protein kinase A is not required for presynaptic long-term plasticity or learning. Proc Natl Acad Sci U S A, 105(38), 14680-14685.

Kawamura, Y. et al. (2004). Glutamate release increases during mossy-CA3 LTP but not during Schaffer-CA1 LTP. Eur J Neurosci, 19(6), 1591-1500.

Kawamura, Y. et al. (2006). The CB1 cannabinoid receptor is the major cannabinoid receptor at excitatory presynaptic sites in the hippocampus and cerebellum. J Neurosci, 26(11), 2991-3001.

Kobayashi (2010) Chapter 4 – Hippocampal MossyFiber Synaptic Transmission and Its Modulation

Koushika, S.P. et al. (2001). A post-docking role for active zone protein Rim. Nat Neurosci, 4(10), 997-1005.

Kreitzer, A.C., Carter, A.G. & Regehr, W.G. (2002). Inhibition of interneuron firing extends the spread of endocannabinoid signaling in the cerebellum. Neuron, 34(5), 787-796.

Kushmerick, C. et al. (2004). Retroinhibition of presynaptic Ca2+ currents by endocannabinoids released via postsynaptic mGluR activation at a calyx synapse. J Neurosci, 24(26), 5955-5965.

Lafourcade, C.A. (2009). Presynaptic mechanisms of endocannabinoid-mediated long-term depression in the hippocampus. J Neurophysiol, 102(4), 2009-2012.

Lauri, S.E. et al. (2003). A role for Ca2+ stores in kainate receptor-dependent synaptic facilitation and LTP at mossy fiber synapses in the hippocampus. Neuron, 39(2), 327-341.

Lerma, J. (2003).Roles and rules of kainate receptors in synaptic transmission. Nat Rev Neurosci, 4(6), 481-495.

Leutgeb, S. & Leutgeb, J.K. (2007). Pattern separation, pattern completion, and new neuronal codes within a continuous CA3 map. Learn Mem, 14(11), 745-757.

Lin, H.C., Mao, S.C., Su, C.L. & Gean, P.W. (2009). The role of prefrontal cortex CB1 receptors in the modulation of fear memory. Cereb Cortex, 19(1), 165-175.

Lonart, G. et al. (2003). Phosphorylation of RIM1alpha by PKA triggers presynaptic long-term potentiation at cerebellar parallel fiber synapses. Cell, 115(1), 49-60.

Lopez-Garcia, J.C., Arancio, O., Kandel, E.R., Baranes, D. (1996). A presynaptic locus for long-term potentiation of elementary synaptic transmission at mossy fiber synapses in culture. Proc Natl Acad Sci U S A, 93(10), 4712-4717.

Luscher, C. & Malenka, R.C. (2011). Drug-evoked synaptic plasticity in addiction: from molecular changes to circuit remodeling. Neuron, 69(4), 650-663.

Maccarrone, M. (2010). Endocannabinoid signaling in healthy and diseased brain. Exp Neurol, 224(1), 1-2.

Mackie, K., Lai, Y., Westenbroek, R. & Mitchell, R. (1995). Cannabinoids activate an inwardly rectifying potassium conductance and inhibit Q-type calcium currents in AtT20 cells transfected with rat brain cannabinoid receptor. J Neurosci, 15(10), 6552-6561.

Mackie, K. & Stella, N. (2006). Cannabinoid receptors and endocannabinoids: evidence for new players. AAPS J, 8(2), E298-E306.

Marsicano, G. & Lafenetre, P. (2009). Roles of the endocannabinoid system in learning and memory. Curr Top Behac Neurosci, 1, 201-230.

Mellor, J. & Nicoll, R.A. (2001). Hippocampal mossy fiber LTP is independent of postsynaptic calcium. Nat Neurosci, 4(2), 125-126.

Mellor, J. (2010). "Synaptic plasticity at hippocampal synapses" in Hippocampal Microcircuits. New York: Springer.

Nicoll, R.A. & Malenka, R.C. (1995). Contrasting properties of two forms of long-term potentiation in the hippocampus. Nature, 377(6545), 115-118.

Nisticò, R. et al. (2009). GLUK1 receptor antagonists and hippocampal mossy fibre function. Int Rev Neurobiol, 85, 13-27.

Nistico, R. et al. (2011). Synergistic interactions between kainate and mGlu receptors regulate bouton Ca signalling and mossy fibre LTP. Sci Rep, 1(103),

Niyuhire, F. et al. (2007). The disruptive effects of the CB1 receptor antagonist rimonabant on extinction learning in mice are task-specific. Psychopharmacology, 191(2), 223-231.

Ohno-Shosaku, T., Tanimura, A., Hashimotodani, Y. & Kano, M. (2011). Endocannabinoids and Retrograde Modulation of Synaptic Transmission. Neuroscientist,

Pan, B., Hillard, C.J. & Liu, Q.S. (2008). Endocannabinoid signaling mediates cocaine-induced inhibitory synaptic plasticity in midbrain dopamine neurons. N Neurosci, 28(6), 1385-1397.

De Petrocellis, L. & Di Marzo, V. (2009). An introduction to the endocannabinoid system: from the early to the latest concepts. Best Pract Res Clin Endocrinol Metab, 23(1), 1-15.

Rizo, J. & Rosenmund, C. (2008). Synaptic vesicle fusion. Nat Struct Mol Biol, 15(7), 665-674.

Ruediger, S. et al. (2011). Learning-related feedforward inhibitory connectivity growth required for memory precision. Nature, 473(7348), 514-518.

Schoch, S. et al. (2002). RIM1alpha forms a protein scaffold for regulating neurotransmitter release at the active zone. Nature, 415(6869), 321-326.

Serrano, A. & Parsons, L.H. (2011). Endocannabinoid influence in drug reinforcement, dependence and addiction-related behaviors. Pharmacol, 132(3), 215-241.

Sidhpura, N. & Parsons, L.H. (2011). Endocannabinoid-mediated synaptic plasticity and addiction-related behavior. Neuropharmacology, 61(7), 1070-1087.

Singla, S., Kreitzer, A.C. & Malenka, R.C. (2007). Mechanisms for synapse specificity during striatal long-term depression. J Neurosci, 27(19), 5260-5264.

Tan, H. et al. (2011). Cannabinoid transmission in the basolateral amygdala modulates fear memory formation via functional inputs to the prelimbic cortex. J Neurosci, 31(14), 5300-5312.

Turu, G. & Hunyady, L. (2010). Signal transduction of the CB1 cannabinoid receptor. J Mol Endocrinol, 44(2), 75-85.

Xiang, Z., Greenwood, A.C., Kairiss, E.W. & Brown, T.H. (1994). Quantal mechanism of long-term potentiation in hippocampal mossy-fiber synapses. J Neurophysiol, 71(6), 2552-2556.

Myers, C.E. & Scharfman, H.E. (2011). Pattern separation in the dentate gyrus: a role for the CA3 backprojection. Hippocampus, 21(11), 1190-1215.

Yang, Y. & Calakos, N. (2010). Acute in vivo genetic rescue demonstrates that phosphorylation of RIM1alpha serine 413 is not required for mossy fiber long-term potentiation. J Neurosci, 30(7), 2542-2546.

Yeckel, M.F., Kapur, A. & Jognston, D. (1999). Multiple forms of LTP in hippocampal CA3 neurons use a common postsynaptic mechanism. Nat Neurosci, 2(7), 625-633.

Yang, Y. & Calakos, N. (2011). Munc13-1 is required for presynaptic long-term potentiation. J Neurosci, 31(33), 1 053-12057.

Yu, L.L. et al. (2011). Effects of cannabinoid CB₁ receptor antagonist rimonabant on acquisition and reinstatement of psychostimulant reward memory in mice. Behav Brain Res, 217(1), 111-116.
Working memory and pattern recognition