ROLE OF AMPA RECEPTOR ENDOCYTOSIS IN SYNAPTIC ...

REVIEWS ROLE OF AMPA RECEPTOR ENDOCYTOSIS IN SYNAPTIC PLASTICITY Reed C. Carroll*, Eric C. Beattie‡§, Mark von Zastrow‡ and Robert C. Malenka§ Activity-mediated changes in the strength of synaptic communication are important for the establishment of proper neuronal connections during development and for the experiencedependent modification of neural circuitry that is believed to underlie all forms of behavioural plasticity. Owing to the wide-ranging significance of synaptic plasticity, considerable efforts have been made to identify the mechanisms by which synaptic changes are triggered and expressed. New evidence indicates that one important expression mechanism of several long-lasting forms of synaptic plasticity might involve the physical transport of AMPA-type glutamate receptors in and out of the synaptic membrane. Here, we focus on the rapidly accumulating evidence that AMPA receptors undergo regulated endocytosis, which is important for long-term depression.

*Department of Neuroscience, Albert Einstein College of Medicine, Bronx, New York 10461, USA. ‡ Departments of Psychiatry and Cellular and Molecular Pharmacology, University of California, San Francisco, California 94143, USA. § Nancy Pritzker Laboratory, Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Palo Alto, California 94304, USA. Correspondence to R.C.M. e-mail: [email protected]

Through the course of evolution, biological systems have developed several conserved mechanisms by which cells can rapidly modify their responsiveness to extracellular stimuli. These include biochemical and allosteric alterations in the properties of key signalling molecules, such as ion channels, transmembrane receptors and enzymes. Several such modifications have been identified as being particularly important for various forms of synaptic plasticity. For example, activation/inhibition of protein kinases and phosphatases1–3 and the consequent phosphorylation/dephosphorylation of critical synaptic substrates, such as receptors4, are crucial for activitydependent changes of synaptic efficacy. Owing to its key role in modulating the responsiveness of a wide range of cell types, the regulated endocytosis of membrane receptors is another mechanism that has received a great deal of attention over the past decade. Many G-protein-coupled receptors (GPCRs), receptor tyrosine kinases and ligand-gated ion channels are transported into the cell via apparently similar processes in response to agonist stimulation. Recently, it has been shown that glutamate receptors of the AMPA subtype (AMPARs) are also internalized in response to various extracellular stimuli. Because of its potential importance for the rapid attenuation of neuro-

transmission during synaptic plasticity, this finding has generated much interest, and reinforces the general importance of receptor endocytosis in the control of neuronal function. Here, we discuss the role of endocytosis in the regulation of AMPARs, and how the control of this cellular process might contribute to synaptic plasticity. We also briefly discuss evidence on how the delivery of AMPARs to the synaptic plasma membrane can be controlled by activity. Endocytosis of signalling receptors

Interest in the possibility that AMPARs are subject to regulated endocytosis developed, in part, out of the knowledge that this process modulates the expression of several families of cell-surface signalling receptors. For example, it has been known for nearly two decades that certain members of the GPCR superfamily undergo desensitization in response to extended exposure to ligand. The activation of several GPCRs results in the stimulation of specific G-protein-coupled receptor kinases, or GRKs, which phosphorylate the receptor5,6. Phosphorylated receptors can bind arrestins, cytosolic proteins that disrupt the activity of the receptor and lead to its internalization by linking the receptor to proteins of the endocytic machinery, such as the

NATURE REVIEWS | NEUROSCIENCE

VOLUME 2 | MAY 2001 | 3 1 5

© 2001 Macmillan Magazines Ltd

REVIEWS

Ligand binding

Receptor phosphorylation

P

Adaptor interaction

Association with clathrincoated pits

Narrowing of neck

Dynamindependent fission

Endocytosis

Uncoating

P P

P

Recycling P P

Lysosome Degradation Signalling

Figure 1 | Endocytosis of cell-surface signalling receptors. Several basic steps have been identified in the regulated endocytosis of a number of membrane signalling receptors. Following binding of the receptor by ligand, activation of downstream signalling pathways results in the phosphorylation of the receptor. This modification allows for the interaction of the receptor with adaptor proteins that couple it to the clathrin endocytic machinery. Clathrin-coated pits that contain the receptor subsequently invaginate and bud off from the cell surface. Dynamin is believed to be involved in the fission of the invaginated pits. After endocytosis, the receptors can recycle back to the plasma membrane, be targeted for degradation in lysosomes or continue to serve some signalling function. Many receptor tyrosine kinases dimerize and autophosphorylate as a result of ligand binding. CLATHRIN-associated

CLATHRIN

A major structural component of coated vesicles that are implicated in protein transport. Clathrin heavy and light chains form a triskelion, the main building element of clathrin coats. AP2

A heterotetrameric complex that serves as an adaptor, linking membrane receptors to clathrincoated pit endocytic machinery. BENZODIAZEPINES

Pharmacologically active molecules with sedative and anxiolytic effects. They act by binding to the GABA receptor γ subunit and potentiating the response elicited by the transmitter. ENDOSOME

Organelle that carries materials ingested by the cell and passes them to lysosomes for degradation or recycles them to the cell surface.

316

adaptor AP2 (REFS 7–9). Although it seems that GPCRs can be internalized by more than one mechanism, many GPCRs are endocytosed via association with clathrin-coated pits10 (FIG. 1). Members of the receptor tyrosine kinase family are also subject to agonist-induced internalization. For example, insulin receptors, epidermal growth factor receptors (EGFRs) and the nerve growth factor receptor TrkA are endocytosed via clathrin-coated pits after exposure to their respective agonist ligands11–16. This might involve a mechanism analogous to that used by GPCRs, as indicated by the finding that ligand-induced phosphorylation of EGFRs promotes receptor interaction with AP2 (REF. 17). EGFR-dependent activation of Src-family kinases also stimulates endocytosis by promoting the assembly of clathrin-coated pits18. The first ionotropic receptors found to be regulated by agonist-induced endocytosis were the GABAA (γaminobutyric acid) receptors (GABAARs). Initial evidence for this came from the observation that extended treatment of cortical neurons with GABA or BENZODIAZEPINES causes a significant increase in the proportion of internalized GABAARs19. Additionally, there is an association of GABAARs with clathrin-coated pits in vivo20, which is increased following chronic benzodiazepine treatments21. Similar to other receptor types, the endocytosis of GABAARs via clathrin-coated pits has been found to first involve an increased association of the receptor with the adaptor protein AP2 (REF. 22). Although the function of regulated GABAAR endocytosis is unknown, blockade of constitutive endocytosis in hippocampal neurons leads to an increase in GABAARmediated current in these cells, indicating that turnover of these receptors contributes to the maintenance of GABA-mediated inhibition22. Of possible clinical relevance, chronic treatments with benzodiazepines that cause decreases in GABAAR-mediated currents23 and increases in the association of GABAARs with the endocytic machinery21 can result in tolerance to these drugs.

Recent work indicates that internalization of some membrane receptors might do more than simply terminate their activation. In fact, many receptors have been found to have important signalling functions that depend on their endocytosis. For example, internalization of certain GPCRs is thought to mediate signal transduction via the formation of an ENDOSOME-associated multiprotein ‘signalling complex’ that contains activated receptors and components of the MAP kinase signalling pathway24,25. Similarly, after their endocytosis in neuronal processes, neurotrophin receptors are believed to exert trophic actions as a consequence of their delivery to the soma via signalling endosomes, which contain the activated receptors as well as associated signalling proteins14,16,26. In addition, GABAAR endocytosis might trigger a reduction in the synthesis of new GABAARs27. The growing knowledge about regulated receptor endocytosis and its functional importance has led to the logical question of whether AMPARs could be similarly internalized and, if so, what the physiological consequences are. The possibility of regulated AMPAR endocytosis has been of particular interest because of a series of findings that have implicated the transport of AMPARs in the expression of synaptic plasticity. AMPAR transport during synaptic plasticity

Indirect evidence. In an effort to understand the mechanisms responsible for synaptic plasticity, attention has focused on NMDA (N-methyl-D-aspartate)-receptordependent long-term potentiation (LTP) and longterm depression (LTD) at synapses on pyramidal cells in the CA1 region of the hippocampus. Although there is experimental support for both pre- and postsynaptic modifications contributing to LTP and LTD28,29, for several years there has been great interest in one particular model of plasticity that involves postsynaptic modifications of glutamate receptors themselves. The model proposes the existence of synapses that express only functional NMDA receptors (NMDARs), so-called

| MAY 2001 | VOLUME 2

www.nature.com/reviews/neuro


REVIEWS

mGluR

AMPAR

AMPAR

VGCC NMDAR

Insulin receptor

Dynamin Calcineurin

P

Depolarization

Clathrin

? Ca2+

P

PKC

???

Ca2+ store

Figure 2 | Activation of several membrane receptors can trigger AMPAR endocytosis. Calcium influx through NMDARs or through voltage-gated calcium channels (VGCCs) activated by AMPAR-mediated membrane depolarization drives the internalization of AMPARs via the stimulation of calcineurin. Activation of insulin receptors might also trigger AMPAR endocytosis in a calcineurin-dependent manner (perhaps by activating calcium release or through another unidentified pathway). Metabotropic glutamate receptor (mGluR)-dependent activation enhances AMPAR endocytosis, perhaps via activation of protein kinase C (PKC) in Purkinje cells (and possibly also in hippocampus). PKC might act directly or indirectly to influence the phosphorylation state of the GluR2 AMPAR, altering its affinity for cytosolic interacting proteins that might in turn allow for endocytosis.

SILENT SYNAPSE

A synapse that contains NMDA receptors but no AMPA receptors and therefore is functionally silent during low-frequency, basal synaptic transmission. EPITOPE-TAGGED MOLECULE

A protein to which the immunological determinant of an antigen has been artificially added, allowing for its subsequent detection with the corresponding antibody. QUANTAL SIZE

The response of the postsynaptic membrane to the transmitter contained in a single synaptic vesicle. DOMINANT-NEGATIVE

A mutant protein that can form a heteromeric complex with the normal molecule, knocking out the activity of the entire complex. DYNAMIN

A small GTP-binding protein that is essential for clathrinmediated endocytosis. It is believed to be involved in fission after invagination of endocytic vesicles.

‘SILENT SYNAPSES’30. These synapses are termed silent as, under conditions of normal synaptic transmission, they are unable to respond to synaptically released glutamate owing to the voltage-dependent block of NMDARs by Mg2+ and the lack of functional AMPARs. In the context of plasticity, the silent-synapse hypothesis proposes that functional AMPARs can either be introduced into previously silent synapses to potentiate neurotransmission following LTP, or be lost from AMPAR-expressing synapses, resulting in decreased synaptic strength during LTD. The existence of silent synapses has been supported by both immunocytochemical and electron-microscopic studies, which have shown that a proportion of synapses exist throughout development that seem to express only NMDARs31–36. Furthermore, electrophysiological studies have demonstrated the existence of synapses that express NMDAR-mediated synaptic responses but no AMPAR-mediated responses37–39. Most importantly, these studies showed that with the induction of LTP, the putative NMDAR-only synapses were ‘unsilenced’ by the rapid appearance of AMPAR-mediated currents. Together, these results provided strong evidence that the activity-dependent insertion of AMPARs into silent synapses could be an important mechanism for the expression of LTP. By analogy, the loss of AMPARs from AMPAR- and NMDAR-expressing synapses and the subsequent generation of silent synapses might be involved in the production of LTD.

SYNAPTONEUROSOME

The presynaptic terminal isolated in conjunction with the postsynaptic spine after subcellular fractionation. This structure retains the anatomical integrity of the synapse.

Activity-dependent redistribution of AMPARs. The first demonstrations that neural activity could regulate the distribution of AMPARs at the synapse used chronic pharmacological agents to manipulate the network excitability of cultured neurons. When synaptic activity

was increased for several days in hippocampal cultures through a reduction of GABA-mediated inhibition, the proportion of synapses containing EPITOPE-TAGGED surface AMPARs was greatly decreased, with no detectable effect on the level of NMDAR expression at synapses40. A similar phenomenon was observed both in spinal and cortical cultures in which pharmacologically induced increases in synaptic activity were accompanied by decreases in the QUANTAL SIZE of AMPAR-mediated responses41,42. Moreover, when AMPAR antagonists were applied for hours to days, there was a resulting increase in the surface expression of synaptic AMPARs in hippocampal and spinal cultures31,42 and a consequent decrease in the proportion of anatomically defined silent synapses31. Whereas experiments using pharmacological methods to modify basal synaptic activity provided evidence that the localization of AMPARs could be modified, the redistribution of AMPARs to and away from synapses occurred in response to treatments that lasted from hours to days, making their relevance to rapidly induced synaptic plasticity uncertain. Evidence that receptors could be rapidly redistributed over a time course more consistent with the expression of LTP or LTD came from the demonstration that brief application of glutamate to hippocampal cultures could cause a profound loss of immunocytochemically detected AMPARs but not NMDARs from synaptic sites43. Subsequent studies showed that a similar rapid loss of AMPARs can be triggered by the activation of multiple receptor types, including AMPARs, NMDARs, insulin receptors and metabotropic glutamate receptors44–49 (FIG. 2). Immunocytochemical studies that specifically distinguish surface and intracellular receptors have further proven that the loss of AMPARs from synapses is the result of internalization of the receptors from the membrane surface. Similar to GPCRs and other membrane receptors, AMPARs are internalized via clathrin-coated pits (FIG. 3). This was established by experiments showing that the regulated internalization of AMPARs was associated with an increased colocalization and interaction with the clathrin adaptor protein AP2 (REFS 44,46). Furthermore, regulated AMPAR internalization was completely blocked when clathrin-mediated endocytosis was inhibited either by high concentrations of extracellular sucrose, expression of a DOMINANT-NEGATIVE form of DYNAMIN, or the disruption of the dynamin–amphiphysin complex that is essential to this form of internalization44,46,50. The ability of AMPARs to undergo rapid endocytosis, particularly in response to glutamate receptor activation, is consistent with a contribution of this mechanism to the reduction in synaptic strength observed with LTD. Direct experimental support for this idea came from a study of the redistribution of AMPARs accompanying LTD in hippocampal cultures. NMDAR-dependent LTD triggered by electrical stimulation of a network of neurons resulted in a significant decrease in the expression of immunocytochemically and electrophysiologically detectable surface AMPARs at synapses51. Hippocampal LTD in vivo is also associated with a decrease in the expression of AMPARs in SYNAPTONEUROSOMES, further


VOLUME 2 | MAY 2001 | 3 1 7


REVIEWS a Pre

PSD Post

Spine apparatus

b

Figure 3 | Electron micrographs of coated vesicles in dendritic spines. a | A coated vesicle adjacent to the spine apparatus. b | A coated vesicle that is in contact with the plasma membrane adjacent to the POSTSYNAPTIC DENSITY (PSD). These vesicles represent a possible ultrastructural correlate for the organelles responsible for the endocytosis (that is, clathrin-coated pits) and perhaps for delivery of AMPARs. (Reproduced with permission from REF. 96 © (2000) Macmillan Magazines Ltd.)

BAPTA-AM

A derivative of the Ca2+ chelator BAPTA with its four carboxylate groups masked by esterifying groups, making it membrane permeable. Upon cleavage by cellular esterases, BAPTA is unable to pass back out of the cell. BAPTA-AM allows buffering of intracellular Ca2+ changes. PDZ DOMAIN

A peptide-binding domain that is important for the organization of membrane proteins, particularly at cell–cell junctions, including synapses. They can bind to the carboxyl termini of proteins, or can form dimers with other PDZ domains. PDZ domains are named after the proteins in which these sequence motifs were originally identified (PSD95, Discs-large, zona occludens-1). POSTSYNAPTIC DENSITY

An electron-dense thickening underneath the postsynaptic membrane at excitatory synapses that contains receptors, structural proteins linked to the actin cytoskeleton and signalling machinery, such as protein kinases and phosphatases.

318

suggesting a role for receptor redistribution in synaptic plasticity52. Finally, in both CA1 pyramidal neurons and cerebellar Purkinje cells, LTD was blocked following the inhibition of clathrin-mediated endocytosis by a peptide designed to disrupt dynamin function in the postsynaptic cell50,53. Inhibitors of endocytosis also blocked the effect of insulin, which causes a depression of synaptic currents that occludes LTD46,54. Although these results have provided solid support for the importance of endocytosis in LTD, subsequent studies of the signalling pathways and mechanisms that regulate AMPAR endocytosis have established further links with plasticity. Regulation of AMPAR endocytosis

Pathways involved in clathrin-mediated endocytosis. The involvement of AMPAR endocytosis in several forms of LTD suggests that similar intracellular signalling pathways should mediate both AMPAR transport and plasticity. Several recent studies have investigated this possibility. The predominant hypothesis for the triggering of NMDAR-dependent LTD proposes that LTD requires an NMDAR-dependent rise in postsynaptic calcium, which preferentially activates a protein-phosphatase cascade that includes calcineurin (PP2B) — a Ca2+/calmodulindependent phosphatase — and protein phosphatase 1 (PP1)55–58. Consistent with this hypothesis, NMDARinduced internalization of AMPARs was inhibited by

preventing increases in intracellular calcium concentration (either by removing extracellular calcium or application of BAPTA-AM), as well as by specific calcineurin inhibitors45,47. Calcineurin was also found to be involved in promoting AMPAR endocytosis in response to application of insulin54 or AMPA45, suggesting that this signalling pathway might be widely involved in the internalization of AMPARs (FIG. 2). The mechanism by which calcineurin affects AMPAR endocytosis is unknown. However, by analogy with work on the activity-dependent enhancement of synaptic vesicle endocytosis, an attractive hypothesis is that calcineurin acts as a calcium sensor, facilitating endocytosis via its association with dynamin/amphiphysin and the subsequent dephosphorylation of components of the endocytic machinery45,59,60. Compared with calcineurin, the role of PP1 in triggering endocytosis of AMPARs is less clear. One study found that inhibition of PP1 blocks the internalization of AMPARs elicited by NMDA application47. Other researchers, however, have found that inhibition of PP1 apparently enhances detectable AMPAR endocytosis45,54. As these studies used different methods to detect internalized AMPARs, it is conceivable that some of the differences might lie in the populations of receptors that these protocols identify. Further complexity with regard to the signalling pathways triggering AMPAR endocytosis comes from the study of LTD in cerebellar Purkinje cells. Unlike NMDAR-dependent LTD in the hippocampus, the triggering of this form of plasticity requires activation of protein kinase C (PKC)61–63, which seems to drive the internalization of AMPARs in these cells64,65 (FIG. 4a). Thus the regulation of AMPAR endocytosis seems to be cell-type specific, perhaps because the subunit compositions of AMPARs differ, and consequently so would the receptor-associated proteins that might differentially regulate endocytosis (see below). Two additional forms of AMPAR endocytosis have been identified that are unlikely to be directly involved in the expression of synaptic plasticity. Constitutive endocytosis of AMPARs has been observed electrophysiologically53, immunocytochemically49,54 and biochemically47. The rate of this process is rapid, resulting in an apparent turnover of 40–50% of surface receptors in tens of minutes. In addition, ligand binding to AMPARs, either by competitive antagonists or AMPA in the absence of receptor activation, is sufficient to drive a form of AMPAR endocytosis that does not seem to require either calcium influx or activation of calcineurin47,54. Constitutive and activation-independent endocytosis of AMPARs is further distinguished from regulated endocytosis by studies of the transport of mutant AMPARs in a heterologous expression system. In HEK293 cells, deletion of a membrane-proximal segment of the GluR2 carboxyl terminus disrupted constitutive endocytosis but not a form of regulated endocytosis triggered by insulin. By contrast, mutations of the last 15 amino acids of the carboxyl terminus, which include the PDZ-binding domain (see below), had the opposite effect54, indicating that independent mechanisms regulate these two forms of AMPAR endocytosis.




REVIEWS AMPAR-interacting proteins in endocytosis. An important aspect of the regulation of AMPAR endocytosis that has been under intense investigation is the role of proteins that directly interact with AMPARs. Most of these proteins bind to the carboxyl terminus of the GluR2/3 AMPAR subunits, and they include the N-ETHYLMALEIMIDESENSITIVE FACTOR (NSF) and the PDZ-domain-containing

N-ETHYLMALEIMIDE-SENSITIVE FACTOR

An ATPase that is a key component of the membrane fusion machinery.

a Cerebellum Ca2+ mGluR1 channel

AMPAR

Glutamate

Dynamin

P

GRIP/ ABP

PICK Clathrin coat

PICK

PKC

Ca2+ PICK

b Hippocampus NMDAR

AMPAR Dynamin Calcineurin

P

P

Clathrin coat

Ca2+ NSF

P

PKC PICK

PICK Mobile

GRIP/ PP1 ABP

GRIP/ ABP Sequestered

P

Degradation

Figure 4 | Model of regulated AMPAR endocytosis during long-term depression in two systems. a | In cerebellar Purkinje cells, mGluR1-dependent activation of protein kinase C (PKC) results in the phosphorylation of serine 880 on the GluR2 AMPAR subunit. This is associated with a decrease in the affinity of GRIP/ABP for GluR2, as well as an increase in PICK1–GluR2 interactions at synaptic sites. GRIP/ABP might therefore act to stabilize AMPARs at the synapse, whereas PICK1 might somehow enhance the endocytosis or internal stability of AMPARs. A rise in calcium due to voltage-dependent calcium channels is also required for longterm depression. This could be required for the activation of PKC, or might enhance endocytosis directly via other signalling pathways. b | In the hippocampus, activation of NMDA (N-methylD-aspartate) receptors elevates cytosolic calcium levels and stimulates calcineurin, which is associated with, and might activate, the clathrin endocytic machinery. Other unknown signalling pathways might also modify the receptor to enhance its association with clathrin-coated pits. AMPARs are subsequently internalized and are either transported for degradation or could possibly become stabilized intracellularly by association with GRIP, a process that requires the dephosphorylation of S880, perhaps through the actions of PP1. The PKC-dependent phosphorylation of S880 mobilizes the AMPARs, allowing their return to the membrane surface, perhaps in an NSF-dependent manner. Association of GluR2 with PICK1 could either be the result or the cause of the PKC-dependent phosphorylation of S880; the function of PICK1 is still uncertain. (GRIP, glutamate-receptor-interacting protein; mGluR, metabotropic glutamate receptor; NSF, N-ethylmaleimide-sensitive factor; PICK, protein that interacts with C kinase; PP, protein phosphatase.)

proteins GRIP1 (glutamate-receptor-interacting protein), GRIP2/ABP and PICK (protein that interacts with C kinase)66,67. Other proteins such as SAP97 and NARP can also interact with AMPARs. SAP97 binds to the carboxyl terminus of GluR1 via a PDZ domain68, and NARP binds to the extracellular amino terminus of GluR1, 2 and 3 (REF. 69). However, the influence of SAP97 and NARP on AMPAR endocytosis has not been studied and these proteins are not discussed further here. NSF, which interacts with the carboxyl terminus of GluR2 proximal to the transmembrane region70–72, was originally identified for its role in presynaptic vesicle fusion and neurotransmitter release73. It was therefore surprising that NSF was the first protein to be implicated in the regulation of transport of AMPARs (TABLE 1). Postsynaptic introduction of peptides that specifically disrupt the interaction of GluR2 with NSF causes a rapid run down of synaptic currents in hippocampal neurons53,71,74,75. This is accompanied by a significant reduction in the expression of surface AMPARs as measured by immunocytochemistry53,75 and by reduced sensitivity of neurons to local application of AMPA71. Taken together, these results indicate that normal synaptic expression of AMPARs might require the NSF–GluR2 interaction. However, the exact function of this interaction remains unknown. NSF might act to stabilize a population of AMPARs at the synapse and make them relatively resistant to constitutive and/or regulated endocytosis. Alternatively, it might be required for the exocytotic delivery of AMPARs to synapses (see below). Postsynaptic NSF might also be important in synaptic plasticity. This was first indicated by a study showing that disruption of postsynaptic NSF function with N-ethylmaleimide or blockade of its interaction with SNAP (synaptosomal-associated protein) by a peptide both impair LTP76. Subsequently, peptide-mediated disruption of the NSF–GluR2 interaction was found to occlude NMDAR-dependent LTD in hippocampal neurons74,75. Conversely, saturation of LTD occluded any further reduction in AMPAR-mediated synaptic currents by the peptide. Evidence also exists for an important, albeit still undefined, role for PDZ-domain-containing proteins that bind to the carboxyl terminus of GluR2/3 (REFS 66,67) in regulating AMPAR transport. These include GRIP1, a protein that contains seven PDZ domains, ABP and GRIP2, which are splice variants with either six or seven PDZ domains, and the PKC-interacting protein PICK1, which has one PDZ domain (TABLE 1). The first evidence for the potential importance of these proteins in the normal surface expression of AMPARs came from the finding that overexpression of the carboxy-terminal 50 amino acids of the GluR2 subunit, which should disrupt the interaction of endogenous AMPARs with GRIP/ABP and PICK, resulted in a decrease in the clustering of AMPARs in primary cultures of spinal cord neurons77. A subsequent study found that removal of the PDZ-binding domain of GluR2, which also eliminates its interaction with all of these proteins, did not disrupt the targeting of recombinant GluR2 to the synaptic plasma membrane in hippocampal neurons.


VOLUME 2 | MAY 2001 | 3 1 9


REVIEWS

Table 1 | Function of GluR2-interacting proteins Protein

Cell system

Method of disruption

Effect on basal activity

Effect on plasticity

Surface receptor expression

References

NSF

Hippocampus

Peptide

Decrease

LTD reduced

Decreased over time

53,71,75

Hippocampus

Antibody

Decrease

Hippocampus

N-ethylmaleimide

Decrease

LTP reduced

Hippocampus

Peptide

Increase (fraction of cells)

Blocks LTD and dedepression

Hippocampus

GluR2 mutagenesis

Cerebellum

Peptide

No effect

Blocks LTD

65

GRIP/ABP

PICK1

71 53,76 79 Decreased accumulation

78

Spinal neurons

Peptide

No effect

Blocks facilitation

83

Hippocampus

Peptide

No effect

No effect

79

Cerebellum

Peptide

No effect

Blocks LTD

65

Cerebellum

Antibody

No effect

Blocks LTD

65

GRIP, glutamate-receptor-interacting protein; LTD, long-term depression; LTP, long-term potentiation; NSF, N-ethylmaleimide-sensitive factor; PICK, protein that interacts with C kinase.

However, the mutant subunit accumulated at synapses significantly less than wild-type GluR2 (REF. 78). This deficit in synaptic accumulation seems to be due specifically to the loss of GRIP/ABP (but not PICK1) binding, and might be due to an increase in the degree of constitutive GluR2 endocytosis78. These findings indicate that the normal function of the GluR2–GRIP/ABP interaction might be to limit endocytosis, perhaps by stabilizing the receptor in the plasma membrane. This conclusion contrasts with observations that acute postsynaptic introduction of peptides that disrupt the interaction of GluR2 with GRIP/ABP actually results in an increase in AMPAR-mediated synaptic responses in a sub-population of hippocampal CA1 pyramidal cells79 (TABLE 1). This observation led to the alternative hypothesis that the binding of GluR2/3 to GRIP/ABP in the hippocampus stabilizes AMPARs in an intracellular pool (not in the plasma membrane) and prevents their reinsertion into the synapse (FIG. 4b). These apparently contradictory results were obtained over very different time courses, using different means to disrupt GluR2/3 protein interactions and might indicate that GRIP/ABP subserves several functions in the delivery, stabilization and endocytosis of synaptic AMPARs. Indeed, as GRIP/ABP might be found both at the membrane and in the cytosol of neurons, it is possible that it could stabilize receptors in both locations. Additional complexity for the roles of GRIP/ABP and PICK1 is provided by the observation that, although these proteins interact with the same site on GluR2, phosphorylation of a specific GluR2 serine residue (S880) in its PDZ-binding domain disrupts binding to GRIP/ABP but not to PICK1 (REFS 64,80). Phosphorylation of S880 is enhanced by PKC activation, which also leads to a decreased association of GRIP with GluR2 in vitro 80, in cultured HEK293 cells80 and in Purkinje cells81 (FIG. 4a). In hippocampal cultures, phosphorylation of S880 by PKC is accompanied by a redistribution of PICK1 to synaptic sites80, whereas in cultures of cortical neurons, activation of PKC by a phorbol ester leads to the internalization of GluR2 receptors80. This suggests that regulated phosphorylation of S880 and the resulting shift in

320

GluR2/3 protein associations might provide a mechanism for actively modulating AMPAR localization. This prediction seems to be confirmed by investigations of the role of PICK1 and GRIP/ABP in synaptic plasticity. LTD in the cerebellum, which requires PKC activation, is also accompanied by increased phosphorylation of S880 and a subsequent dissociation of AMPARs from GRIP, possibly increasing association with PICK1 as a consequence81. It was proposed that this rearrangement leads to disruption of AMPAR clusters and eventually to receptor internalization81. Consistent with this hypothesis, the postsynaptic introduction of peptides that disrupt GRIP/ABP and PICK1 interactions with GluR2/3 significantly reduces LTD in cerebellar cultures65. Although a peptide that blocks GRIP/ABP and PICK1 interaction with GluR2 also blocks LTD in hippocampal CA1 pyramidal cells79, this blockade seems to occur by a different mechanism than in cerebellar Purkinje cells. As mentioned above, in the hippocampus, the postsynaptic introduction of the peptide caused an increase in synaptic strength in a subset of cells, and in most cells following the prior induction of LTD. These results are consistent with the hypothesis that disrupting the interaction of GRIP/ABP with GluR2 blocks LTD by preventing the stabilization of endocytosed receptors in intracellular pools, thereby allowing the rapid return of receptors to the synaptic surface, an event that seems to require PKC79. By contrast, a peptide designed to specifically disrupt the PICK1–GluR2/3 interaction had no effect on LTD or on basal synaptic strength79. In cultured Purkinje cells, however, PICK1 seems to be much more important in LTD. Peptides that disrupt the PICK1– GluR2/3 interaction, antibodies directed against the PDZ domain of PICK1 and heterologously expressed mutant PICK1–GST fusion proteins all impaired LTD when introduced into the postsynaptic cell65. Although these experiments do not rule out a role for GRIP, they could mean that PICK1 binding to GluR2 in Purkinje cells is necessary to prime receptors for endocytosis, or to stabilize the endocytosed receptors in intracellular pools65. These discrepancies might reflect important cell-typespecific differences in the mechanisms that regulate




REVIEWS

NMDAR

Ca2+

AMPAR

Continuous recycling of receptor is GluR2-subunit dependent

NSF

CaMKII SAP97

Golgi/ endosome In early development, receptor delivery is GluR4-subunit dependent

Regulated insertion is GluR1-subunit dependent

Figure 5 | Model of AMPAR cycling and insertion. AMPARs might be inserted into the synaptic membrane through several pathways. Long-term potentiation involves the delivery of GluR1-containing AMPARs, a process that requires activation of CaMKII and perhaps interaction with SAP97. The final levels of AMPARs at the synapse are maintained by the rapid constitutive cycling of the receptors that might require AMPARs containing GluR2. During early postnatal development, GluR4-containing AMPARs might be required for the activity-dependent delivery of newly synthesized receptors. (CaMKII, calcium/calmodulin-dependent kinase II; NSF, N-ethylmaleimide-sensitive factor.)

AMPAR transport and hence synaptic plasticity. Furthermore, these results indicate that understanding the exact functions of GluR2/3-interacting proteins will provide an important key to understanding the expression of several forms of synaptic plasticity. Exocytosis of AMPARs

RECTIFICATION

The property whereby current through a channel does not flow with the same ease from the inside as from the outside. In inward rectification, for instance, current into the cell flows more easily than out of the cell through the same population of channels.

Reinsertion of AMPARs into the synaptic membrane. Clearly, the functional consequence of AMPAR endocytosis on the surface expression of AMPARs is dependent not only on the extent and rate of internalization, but also on the rate of insertion and recycling of AMPARs back to the synaptic plasma membrane. Therefore, a review of the functions and mechanisms of AMPAR endocytosis would not be complete without mentioning the converse process that probably involves some form of exocytosis of AMPARs. In addition, as LTP and LTD are often thought of as related but opposite phenomena, the regulated insertion of AMPARs into the synaptic membrane would be the predicted mechanism by which synaptic strength can be rapidly increased during LTP if AMPAR endocytosis is mediating LTD. Several significant findings over the past few years have provided ample evidence that both constitutive and regulated exocytosis of AMPARs have a significant role in determining synaptic strength. The steady-state surface expression of synaptic AMPARs seems to depend on a rapid constitutive cycling of receptors (FIG. 5). Blockade of clathrin-mediated endocytosis in hippocampal neurons results in a rapid increase in synaptic currents, whereas inhibition of exocytosis results in a run-down of synaptic responses53. Constitutive AMPAR insertion has also been detected immunocytochemically in hippocampal neurons and biochemically in cortical neurons47. The estimated rate of reinsertion of internalized receptors is tightly coupled to

the rate of endocytosis, resulting in a relatively constant level of surface AMPAR expression47. This balance between internalization and reinsertion is apparently dynamic, in that increasing the basal activity of cultures increases both the overall rate of endocytosis and the rate of receptor reinsertion to comparable levels47. Biochemical analysis in cortical neurons has shown that AMPARs can also recycle back to the membrane surface after the triggering of regulated receptor endocytosis47. The amount of receptor recycling following agonist-induced AMPAR internalization, however, seems to depend on the stimulus that drives endocytosis. Application of NMDA led to rapid endocytosis of AMPARs, which were then recycled back to the surface over time. However, AMPARs endocytosed after AMPA treatment were for the most part not returned to the surface, but instead transported to lysosomes for degradation. Although the mechanisms that regulate receptor recycling are unknown, it has been suggested that GRIP/ABP might be involved. As mentioned above, after the induction of LTD in CA1 pyramidal cells, disruption of the GluR2–GRIP/ABP interaction led to a rapid, PKC-dependent increase in synaptic responses. This has been interpreted as GRIP acting to anchor internalized receptors in endosomes, thereby preventing their reinsertion until phosphorylation by PKC activation releases them79 (FIGS 4,5). PKA phosphorylation, perhaps acting on GluR1, has similarly been suggested to be important for receptor reinsertion47. Implications for plasticity. The earliest indication that AMPARs could be actively driven into synaptic membranes, causing an increase in surface receptor levels, came from the electrophysiological studies of silent synapses described above. Similar electrophysiological results have been obtained throughout the central nervous system82–84, indicating that the regulated transport of AMPARs is probably a ubiquitous property of excitatory synapses. Initial evidence that these observations were actually due to transport of AMPARs came from the finding that loading postsynaptic cells with peptides of the GluR2 carboxyl terminus, which should prevent protein interactions with GluR2, was sufficient to prevent the serotonin-induced potentiation of silent synapses in spinal neurons85. Recent elegant studies have provided direct evidence for the regulated insertion of AMPARs into the synapse. By directly visualizing a fusion protein consisting of GFP (green fluorescent protein) and the AMPAR subunit GluR1, investigators observed that after the induction of LTP in cultured hippocampal slices, GluR1 became redistributed into synaptic spines86. Furthermore, after LTP induction, the RECTIFICATION properties of synaptic currents were changed in GluR1–GFP-expressing CA1 pyramidal cells87. This confirmed that LTP caused the delivery of the GluR1–GFP homomers to the synaptic plasma membrane, as currents mediated by these receptors show robust inward rectification, whereas the endogenous GluR2-containing AMPARs do not. Importantly, the recombinant GluR1 homomers were not delivered constitutively to synapses, but required


VOLUME 2 | MAY 2001 | 3 2 1


REVIEWS

BREFELDIN A

A fungal metabolite that affects membrane transport and the structure of the Golgi apparatus. TETANUS TOXIN

The causative agent of tetanus. Tetanus toxin blocks transmitter release as a result of synaptobrevin proteolysis.

322

either the induction of LTP or the overexpression of calcium/calmodulin-dependent kinase II (CaMKII), which is known to be important in triggering of LTP28. Surprisingly, mutagenesis of the carboxy-terminal tail of GluR1 revealed that phosphorylation of serine 831, a known CaMKII site88, was not required for the regulated synaptic delivery of GluR1 but that the PDZ-binding domain at the carboxyl terminus of GluR1 was required. This result points to a potentially important role in synaptic plasticity for PDZ-domain-containing proteins that interact with GluR1, such as SAP97 (REF. 68). Interestingly, GluR2-containing AMPARs seem capable of being delivered to the synapse without a requirement for CaMKII or LTP89, providing evidence that this receptor subunit might be involved in constitutive receptor delivery90. Furthermore, the delivery of AMPARs to synapses during early postnatal development seems to require GluR4 and to use an activity-dependent mechanism distinct from that used during LTP89. So, the roles of specific AMPAR subunits in regulated endocytosis and activity-dependent exocytosis/delivery might differ, presumably because of the difference in their specific protein–protein interactions. In addition to SAP97, recent work has identified the protein stargazin as critically important for the delivery of AMPARs to the synaptic plasma membrane, suggesting that this or other AMPAR-interacting proteins might be involved in the regulated or constitutive exocytosis of AMPARs91. Several studies have examined the regulated membrane insertion of endogenous AMPARs rather than recombinant AMPAR subunits. Two of these series of experiments used a chemical means of potentiating synaptic strength through brief, direct activation of NMDARs. In one study92, subcellular fractions, prepared from control and NMDA-treated hippocampal slices, were analysed for relative expression of AMPARs. NMDA-potentiated slices were found to have greater levels of GluR1 and GluR2 in the synaptic membrane fractions. This increase required the activity of CaMKII and the calcium-dependent protease calpain, and was blocked by BREFELDIN A (BrfA), raising the possibility that AMPAR-containing vesicles are being secreted from the Golgi during regulated receptor insertion. BrfA had no effect on basal transmission, suggesting that constitutively recycling receptors and actively inserted receptors are not, at least initially, in the same pool92. A second study used electrophysiological and immunocytochemical methods to study the regulated insertion of AMPARs in cultured hippocampal neurons93. Following treatment with glycine to activate synaptic NMDARs, an increase in synaptic AMPARmediated currents could be detected. TETANUS TOXIN, which blocks exocytosis, prevented the enhancement when loaded into postsynaptic cells. Treatment with glycine also caused an increase in the surface expression of AMPARs found at synaptic sites. Most importantly, using a clever immunocytochemical method for detecting only newly inserted AMPARs, the investigators observed an increased delivery of AMPARs to synaptic sites. Further support for the activity-dependent synaptic delivery of endogenous AMPARS comes

from the finding that after the induction of LTP in vivo in the adult hippocampus, there is an increase in the amount of GluR1 and GluR2 in synaptoneurosomes52. Concluding remarks

We have presented evidence that the regulated endocytosis and delivery of AMPARs to synaptic sites provides an elegant symmetric mechanism for the activity-dependent regulation of synaptic strength. Furthermore, a balance of constitutive exocytosis and endocytosis of receptors seems to keep the level of synaptic AMPARs relatively constant. This balance can be disrupted by signals that drive either an increased insertion, which seems to depend partly on the recruitment of newly synthesized receptors, or removal of synaptic AMPARs, a process that is enhanced by increased receptor degradation. The resulting alteration in synaptic strength is then presumably maintained by constitutive cycling of receptors. It is important to remember, however, that increased insertion of AMPARs might also occur via a reduction in constitutive endocytosis and conversely, a decrease in AMPAR surface expression can occur via a decrease in constitutive delivery. So, the endocytotic and exocytotic mechanisms that are used for the transport of synaptic AMPARs are inextricably intertwined. Although the processes that are responsible for the expression of LTD are far from completely understood, the findings discussed in this review begin to identify some of the key players and their putative roles in this phenomenon. A possible model that unites the available observations is as follows (FIGS 4,5). Basal levels of AMPARs are maintained at the membrane via a constant constitutive cycling of receptors. NSF has a critical role in this process, either by assisting the insertion of receptors or stabilizing a population of synaptic receptors at the membrane surface. GRIP might also act to stabilize receptors at the membrane surface, the evidence for this being most compelling in the cerebellum. In Purkinje cells, activation of PKC could lead to the phosphorylation of S880, disrupting the binding of GRIP to the AMPAR GluR2 subunit and increasing the association with PICK1, leading to internalization of the receptor. In hippocampal CA1 pyramidal cells, the role of GRIP and PKC phosphorylation in triggering AMPAR endocytosis is uncertain. Instead, activation of calcineurin by NMDARs or other receptors leads to the clathrin-dependent internalization of AMPARs. Once internalized, AMPARs might be stabilized in intracellular vesicles through binding to GRIP, an interaction that requires S880 in GluR2 to be maintained in a dephosphorylated state via maintained activity of PP1. Disruption of GRIP binding to the AMPAR as a result of, or in conjunction with, PKC phosphorylation of GluR2 S880 enables the internalized receptors to be freed up and recycle back to the membrane surface. The phosphorylation state of S845, a protein kinase A (PKA) site on GluR1, could also influence AMPAR transport during LTD47,94,95. The molecular mechanisms responsible for the delivery of AMPARs to the synaptic plasma membrane are much less clear. What can be stated is that an exocytotic




REVIEWS process that involves SNARE-mediated membrane fusion seems to be important and that activity can modulate AMPAR insertion, in part via activation of CaMKII as well as independent signalling pathways. Over a period of just a few years, an impressive amount of information has been gathered regarding the means of regulation of AMPAR expression at the synaptic surface. The finding that the transport of AMPARs can be dynamically regulated by activity opens up numerous possibilities as to the role that these transport processes might have in the development, plasticity and ageing of the neuron. An intriguing possibility is that changes in the level of AMPAR expression at synapses provide the first step towards a more complete restructuring of the synapse96. As has always been the case in

1.

2. 3.

4.

5.

6.

7.

8.

9.

10.

11. 12.

13.

14.

15.

16.

17.

18.

19.

20.

Bliss, T. V. & Collingridge, G. L. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361, 31–39 (1993). Malenka, R. C. Synaptic plasticity in the hippocampus: LTP and LTD. Cell 78, 535–538 (1994). Nicoll, R. A. & Malenka, R. C. Contrasting properties of two forms of long-term potentiation in the hippocampus. Nature 377, 115–118 (1995). Swope, S. L., Moss, S. I., Raymond, L. A. & Huganir, R. L. Regulation of ligand-gated ion channels by protein phosphorylation. Adv. Sec. Mess. Phosphopro. Res. 33, 49–78 (1999). Pitcher, J. A., Freedman, N. J. & Lefkowitz, R. J. G proteincoupled receptor kinases. Annu. Rev. Biochem. 67, 653–692 (1998). Zhang, J. et al. Molecular mechanisms of G protein-coupled receptor signaling: role of G protein-coupled receptor kinases and arrestins in receptor desensitization and resensitization. Receptors Channels 5, 193–199 (1997). Krupnick, J. G. & Benovic, J. L. The role of receptor kinases and arrestins in G protein-coupled receptor regulation. Annu. Rev. Pharmacol. Toxicol. 38, 289–319 (1998). Lefkowitz, R. J., G protein-coupled receptors. III. New roles for receptor kinases and β-arrestins in receptor signaling and desensitization. J. Biol. Chem. 273, 18677–18680 (1998). Laporte, S. A., Oakley, R. H., Holt, J. A., Barak, L. S. & Caron, M. G. The interaction of β-arrestin with the AP-2 adaptor is required for the clustering of β2-adrenergic receptor into clathrin-coated pits. J. Biol. Chem. 275, 23120–23126 (2000). Tsao, P. & Von Zastrow, M. Downregulation of G proteincoupled receptors. Curr. Opin. Neurobiol. 10, 365–369 (2000). DiGuglielmo, G. M. et al. Insulin receptor internalization and signalling. Mol. Cell. Biochem. 182, 59–63 (1998). Ceresa, B. P. & Schmid, S. L. Regulation of signal transduction by endocytosis. Curr. Opin. Cell. Biol. 12, 204–210 (2000). Vieira, A. V., Lamaze, C. & Schmid, S. L. Control of EGF receptor signaling by clathrin-mediated endocytosis. Science 274, 2086–2089 (1996). Wiley, H. S. & Burke, P. M. Regulation of receptor tyrosine kinase signaling by endocytic trafficking. Traffic 2, 12–18 (2001). Beattie, E. C., Howe, C. L., Wilde, A., Brodsky, F. M. & Mobley, W. C. NGF signals through TrkA to increase clathrin at the plasma membrane and enhance clathrinmediated membrane trafficking. J. Neurosci. 20, 7325–7333 (2000). Grimes, M. L. et al. Endocytosis of activated TrkA: evidence that nerve growth factor induces formation of signaling endosomes. J. Neurosci. 16, 7950–7964 (1996). Sorkin, A. & Carpenter, G. Interaction of activated EGF receptors with coated pit adaptins. Science 261, 612–615 (1993). Wilde, A. et al. EGF receptor signaling stimulates SRC kinase phosphorylation of clathrin, influencing clathrin redistribution and EGF uptake. Cell 96, 677–687 (1999). Tehrani, M. H. & Barnes, E. M. Agonist-dependent internalization of γ-aminobutyric acidA/benzodiazepine receptors in chick cortical neurons. J. Neurochem. 57, 1307–1312 (1991). Tehrani, M. H. & Barnes, E. M. Identification of GABAA/ benzodiazepine receptors on clathrin-coated vesicles

the synaptic plasticity field, exciting new findings have been accompanied by conflicting results that are initially difficult to interpret. This is inevitable when dissecting any complex phenomenon. However, important discussions have emerged out of the confusion and they have rapidly driven the field forward and will continue to do so in the immediate future. Links DATABASE LINKS AMPARs | insulin receptor | EGFR | TrkA | AP2 | GABAARs | NMDARs | dynamin | amphiphysin | PP2B | PP1 | PKC | GluR2 | NSF | GRIP1 | PICK | SNAP | GluR1 | CaMKII | GluR4 | calpain | PKA ENCYCLOPEDIA OF LIFE SCIENCES AMPA receptors

from rat brain. J. Neurochem. 60, 1755–1761 (1993). 21. Tehrani, M. H. & Barnes, E. M. Sequestration of γ-aminobutyric acidA receptors on clathrin-coated vesicles during chronic benzodiazepine administration in vivo. J. Pharmacol. Exp. Ther. 283, 384–390 (1997). 22. Kittler, J. T. et al. Constitutive endocytosis of GABAA receptors by an association with the adaptin AP2 complex modulates inhibitory synaptic currents in hippocampal neurons. J. Neurosci. 20, 7972–7977 (2000). 23. Gallager, D. W., Lakoski, J. M., Gonsalves, S. F. & Rauch, S. L. Chronic benzodiazepine treatment decreases postsynaptic GABA sensitivity. Nature 308, 74–77 (1984). 24. DeFea, K. A. et al. Beta-arrestin-dependent endocytosis of proteinase-activated receptor 2 is required for intracellular targeting of activated ERK1/2. J. Cell. Biol. 148, 1267–1281 (2000). 25. McDonald, P. H. et al. Beta-arrestin 2: A receptor-regulated MAPK scaffold for the activation of JNK3. Science 290, 1574–1577 (2000). 26. Riccio, A., Pierchala, B. A., Ciarallo, C. L. & Ginty, D. D. An NGF-TrkA-mediated retrograde signal to transcription factor CREB in sympathetic neurons. Science 277, 1097–1100 (1997). 27. Barnes, E. M. Use-dependent regulation of GABAA receptors. Int. Rev. Neurobiol. 39, 53–76 (1996). 28. Malenka, R. C. & Nicoll, R. A. Long-term potentiation — a decade of progress? Science 285, 1870–1874 (1999). 29. Kullmann, D. M. & Siegelbaum, S. A. The site of expression of NMDA receptor-dependent LTP: new fuel for an old fire. Neuron 15, 997–1002 (1995). 30. Malenka, R. C. & Nicoll, R. A. Silent synapses speak up. Neuron 19, 473–476 (1997). 31. Liao, D., Zhang, X., O’Brien, R., Ehlers, M. D. & Huganir, R. L. Regulation of morphological postsynaptic silent synapses in developing hippocampal neurons. Nature Neurosci. 2, 37–43 (1999). 32. Petralia, R. S. et al. Selective acquisition of AMPA receptors over postnatal development suggests a molecular basis for silent synapses. Nature Neurosci. 2, 31–36 (1999). 33. Nusser, Z. et al. Cell type and pathway dependence of synaptic AMPA receptor number and variability in the hippocampus. Neuron 21, 545–559 (1998). 34. Takumi, Y., Ramírez-León, V., Laake, P., Rinvik, E. & Ottersen, O. P. Different modes of expression of AMPA and NMDA receptors in hippocampal synapses. Nature Neurosci. 2, 618–624 (1999). References 32–34 used immunogold labelling of AMPARs to analyse quantitatively the AMPAR content at individual synapses in the CA1 region of the hippocampus, and found that a proportion of synapses contain no or very small numbers of AMPARs, thus providing anatomical support for the silent synapse hypothesis. 35. Racca, C., Stephenson, R. A., Streit, R., Roberts, J. D. & Somogyi, P. NMDA receptor content of synapses in stratum radiatum of the hippocampal CA1 area. J. Neurosci. 20, 2512–2522 (2000). 36. Gomperts, S. N., Rao, A., Craig, A. M., Malenka, R. C. & Nicoll, R. A. Postsynaptically silent synapses in single neuron cultures. Neuron 21, 1443–1451 (1998). 37. Isaac, J. T., Nicoll, R. A. & Malenka, R. C. Evidence for silent synapses: implications for the expression of LTP. Neuron 15, 427–434 (1995).


38. Liao, D., Hessler, N. A. & Malinow, R. Activation of postsynaptically silent synapses during pairing-induced LTP in CA1 region of hippocampal slice. Nature 375, 400–404 (1995). 39. Durand, G. M., Kolvachuk, Y. & Konnerth, A. Long-term potentiation and functional synapse induction in developing hippocampus. Nature 381, 71–75 (1996). References 37–39 provided electrophysiological evidence for the existence of silent synapses and showed that during LTP, silent synapses are converted to functional synapses, presumably by the addition or modification of AMPARs. 40. Lissin, D. V. et al. Activity differentially regulates the surface expression of synaptic AMPA and NMDA glutamate receptors. Proc. Natl Acad. Sci. USA 95, 7097–7102 (1998). 41. Turrigiano, G. G., Leslie, K. R., Desai, N. S., Rutherford, L. C. & Nelson, S. B. Activity-dependent scaling of quantal amplitude in neocortical neurons. Nature 391, 892–896 (1998). 42. O’Brien, R. J. et al. Activity-dependent modulation of synaptic AMPA receptor accumulation. Neuron 21, 1067–1078 (1998). 43. Lissin, D. V., Carroll, R. C., Nicoll, R. A., Malenka, R. C. & von Zastrow, M. Rapid, activation-induced redistribution of ionotropic glutamate receptors in cultured hippocampal neurons. J. Neurosci. 19, 1263–1272 (1999). 44. Carroll, R. C. et al. Dynamin-dependent endocytosis of ionotropic glutamate receptors. Proc. Natl Acad. Sci. USA 96, 14112–14117 (1999). 45. Beattie, E. C. et al. Regulation of AMPA receptor endocytosis by a signaling mechanism shared with LTD. Nature Neurosci. 3, 1291–1300 (2000). 46. Man, H.-Y. et al. Regulation of AMPA receptor-mediated synaptic transmission by clathrin-dependent receptor internalization. Neuron 25, 649–662 (2000). 47. Ehlers, M. D., Reinsertion or degradation of AMPA receptors determined by activity-dependent endocytic sorting. Neuron 28, 511–525 (2000). References 44-47 showed that the endocytosis of AMPARs occurs via a process involving dynamin and clathrin-coated pits and that calcium-dependent activation of protein phosphatases (references 45, 47) might be important in the enhancement of AMPAR endocytosis owing to NMDAR or AMPAR activation. 48. Snyder, E. M. et al. Group I metabotropic glutamate receptor activation initiates internalization of AMPA receptors in cultured hippocampal neurons. Soc. Neurosci. Abstr. 134, 15 (2000). 49. Zhou, Q., Xiao, M. & Nicoll, R. A. Contribution of cytoskeleton to the internalization of AMPA receptors. Proc. Natl Acad. Sci. USA 98, 1261–1266 (2001). 50. Wang, Y. T. & Linden, D. J. Expression of cerebellar longterm depression requires postsynaptic clathrin-mediated endocytosis. Neuron 25, 635–647 (2000). 51. Carroll, R. C., Lissin, D. V., von Zastrow, M. Nicoll, R. A. & Malenka, R. C. Rapid redistribution of glutamate receptors contributes to long-term depression in hippocampal cultures. Nature Neurosci. 2, 454–460 (1999). The first demonstration that the expression of longterm depression is caused, at least in part, by a loss of AMPARs from synapses. 52. Heynen, A. J., Quinlan, E. M., Bae, D. C. & Bear, M. F. Bidirectional, activity-dependent regulation of glutamate receptors in the adult hippocampus in vivo. Neuron 28, 527–536 (2000).

VOLUME 2 | MAY 2001 | 3 2 3


REVIEWS 53. Lüscher, C. et al. Role of AMPA receptor cycling in synaptic transmission and plasticity. Neuron 24, 649–658 (1999). 54. Lin, J. W. et al. Distinct molecular mechanisms and divergent endocytotic pathways of AMPA receptor internalization. Nature Neurosci. 3, 1282–1290 (2000). 55. Lisman, J. A mechanism for the Hebb and the anti-Hebb processes underlying learning and memory. Proc. Natl Acad. Sci. USA 86, 9574–9578 (1989). 56. Mulkey, R. M., Herron, C. E. & Malenka, R. C. An essential role for protein phosphatases in hippocampal long-term depression. Science 261, 1051–1055 (1993). 57. Mulkey, R. M., Endo, S., Shenolikar, S. & Malenka, R. C. Involvement of a calcineurin/inhibitor-1 phosphatase cascade in hippocampal long-term depression. Nature 369, 486–488 (1994). 58. O’Dell, T. J. & Kandel, E. R. Low-frequency stimulation erases LTP through an NMDA receptor-mediated activation of phosphatases. Learn. Mem. 1, 129–139 (1994). 59. Lai, M. M. et al. The calcineurin–dynamin 1 complex as a calcium sensor for synaptic vesicle endocytosis. J. Biol. Chem. 274, 25963–25966 (1999). 60. Slepnev, V. I., Ochoa, G. C., Butler, M. H., Grabs, D. & DeCamilli, P. Role of phosphorylation in regulation of the assembly of endocytic coat complexes. Science 281, 821–824 (1998). 61. Linden, D. J. & Connor, J. A. Participation of postsynaptic PKC in cerebellar long-term depression in culture. Science 254, 1656–1659 (1991). 62. Crepel, F. & Jaillard, D. Protein kinases, nitric oxide and longterm depression of synapses in the cerebellum. Neuroreport 1, 133–136 (1990). 63. Hartell, N. A. cGMP acts within cerebellar Purkinje cells to produce long-term depression via mechanisms involving PKC and PKG. NeuroReport 5, 833–836 (1994). 64. Matsuda, S., Mikawa, S. & Hirai, H. Phosphorylation of serine-880 in GluR2 by protein kinase C prevents its C terminus from binding with glutamate receptor-interacting protein. J. Neurochem. 73, 1765–1768 (1999). 65. Xia, J., Chung, H. J., Wihler, C., Huganir, R. L. & Linden, D. J. Cerebellar long-term depression requires PKC-regulated interactions between GluR2/3 and PDZ domain-containing proteins. Neuron 28, 499–510 (2000). References 64 and 65 present evidence that PKCmediated phosphorylation of GluR2 and the consequent disruption of GluR2 interactions with PDZ-domain-containing proteins leads to the internalization of AMPARs during long-term depression at parallel-fibre–Purkinje-cell synapses. 66. Scannevin, R. H. & Huganir, R. L. Postsynaptic organization and regulation of excitatory synapses. Nature Rev. Neurosci. 1, 133–141 (2000). 67. Braithwaite, S. P., Meyer, G. & Henley, J. M. Interactions between AMPA receptors and intracellular proteins. Neuropharmacology 39, 919–930 (2000). 68. Leonard, A. S., Davare, M. A., Horne, M. C., Garner, C. G. & Hell, J. W. SAP97 is associated with the α-amino-3hydroxy-5-methylisoxazole-4-propionic acid receptor GluR1 subunit. J. Biol. Chem. 273, 19518–19524 (1998).

324

69. O’Brien, R. J. et al. Synaptic clustering of AMPA receptors by the extracellular immediate-early gene product Narp. Neuron 23, 309–323 (1999). 70. Osten, P. et al. The AMPA receptor GluR2 C terminus can mediate a reversible, ATP-dependent interaction with NSF and α- and β-SNAPs. Neuron 21, 99–110 (1998). 71. Nishimune, A. et al. NSF binding to GluR2 regulates synaptic transmission. Neuron 21, 87–97 (1998). 72. Song, I. et al. Interaction of the N-ethylmaleimidesensitive factor with AMPA receptors. Neuron 21, 393–400 (1998). References 70–72 showed that NSF interacts with the intracellular tail of GluR2 and that disruption of this interaction causes a decrease in synaptic efficacy that subsequently (references 53, 74) was shown to be due to the loss of synaptic AMPARs. 73. Rothman, J. E., Mechanisms of intracellular protein transport. Nature 372, 55–63 (1994). 74. Luthi, A. et al. Hippocampal LTD expression involves a pool of AMPARs regulated by the NSF–GluR2 interaction. Neuron 24, 389–399 (1999). 75. Noel, J. et al. Surface expression of AMPA receptors in hippocampal neurons is regulated by an NSF-dependent mechanism. Neuron 23, 365–376 (1999). 76. Lledo, P. M., Zhang, X., Sudhof, T. C, Malenka, R. C. & Nicoll, R. A. Postsynaptic membrane fusion and long-term potentiation. Science 279, 399–403 (1998). Presents evidence that postsynaptic membrane fusion is important for the expression of long-term potentiation. 77. Dong, H. et al. GRIP: a synaptic PDZ domain-containing protein that interacts with AMPA receptors. Nature 386, 279–284 (1997). 78. Osten, P. et al. Mutagenesis reveals a role for ABP/GRIP binding to GluR2 in synaptic surface accumulation of the AMPA receptor. Neuron 27, 313–325 (2000). 79. Daw, M. I. et al. PDZ proteins interacting with C-terminal GluR2/3 involved in a PKC-dependent regulation of AMPA receptors at hippocampal synapses. Neuron 28, 873–886 (2000). Presents evidence that the interactions of GluR2/3 with GRIP/ABP might be important for the intracellular retention of AMPARs after they are internalized and that the return of AMPARs to the membrane surface requires activity of PKC. 80. Chung, H. J., Xia, J., Scannevin, R. H., Zhang, X. & Huganir, R. L. Phosphorylation of the AMPA receptor subunit GluR2 differentially regulates its interaction with PDZ domaincontaining proteins. J. Neurosci. 20, 7258–7267 (2000). 81. Matsuda, S., Launey, R., Mikawa, S. & Hirai, H. Disruption of AMPA receptor GluR2 clusters following long-term depression induction in cerebellar Purkinje neurons. EMBO J. 19, 2765–2774 (2000). 82. Isaac, J. T., Crair, M. C., Nicoll, R. A. & Malenka, R. C. Silent synapses during development of thalamocortical inputs. Neuron 18, 269–280 (1997). 83. Li, P. & Zhuo, M. Silent glutamatergic synapses and nociception in mammalian spinal cord. Nature 393, 695–698 (1998).


84. Rumpel, S., Hatt, H. & Gottmann, K. Silent synapses in the developing rat visual cortex: evidence for postsynaptic expression of synaptic plasticity. J. Neurosci. 18, 8863–8874 (1998). 85. Li, P. et al. AMPA receptor-PDZ interactions in facilitation of spinal sensory synapses. Nature Neurosci. 2, 972–977 (1999). 86. Shi, S. H. et al. Rapid spine delivery and redistribution of AMPA receptors after synaptic NMDA receptor activation. Science 284, 1811–1816 (1999). 87. Hayashi, Y. et al. Driving AMPA receptors into synapses by LTP and CaMKII: requirement for GluR1 and PDZ domain interaction. Science 287, 2262–2267 (2000). References 86 and 87 present an elegant series of experiments showing that long-term potentiation involves the CaMKII-dependent delivery of AMPARs to the synaptic plasma membrane. 88. Barria, A., Derkach, V. & Soderling, T. Identification of the Ca2+/calmodulin-dependent protein kinase II regulatory phosphorylation site in the α-amino-3-hydroxyl-5-methyl-4isoxazole-propionate-type glutamate receptor. J. Biol. Chem. 272, 32727–32730 (1997). 89. Zhu, J. J., Esteban, J. A., Hayashi, Y. & Malinow, R. Synaptic potentiation during early development: delivery of GluR4containing AMPA receptors by spontaneous activity. Nature Neurosci. 3, 1098–1106 (2000). 90. Malinow, R., Mainen, Z. F. & Hayashi, Y. LTP mechanisms: from silence to four-lane traffic. Curr. Opin. Neurobiol. 10, 352–357 (2000). 91. Chen, L. et al. Stargazin regulates synaptic targeting of AMPA receptors by two distinct mechanisms. Nature 408, 936–943 (2000). Identifies a role for the protein stargazin in the synaptic targeting of AMPARs in cerebellar granule cells. This protein can interact both with AMPARs, which allows targeting of the receptors to the membrane surface, and with PSD95, which seems to be necessary for the synaptic clustering of the AMPARs. 92. Broutman, G. & Baudry, M. Involvement of the secretory pathway for AMPA receptors in NMDA-induced potentiation in hippocampus. J. Neurosci. 21, 27–34 (2001). 93. Lu, W. Y. et al. Activation of synaptic NMDA receptors induces membrane insertion of new AMPA receptors and LTP in cultured hippocampal neurons. Neuron 29, 243–254 (2001). 94. Kameyama, K., Lee, H. K., Bear, M. F. & Huganir, R. L. Involvement of a postsynaptic protein kinase A substrate in the expression of homosynaptic long-term depression. Neuron 21, 1163–1175 (1998). 95. Lee, H.-K., Barbarosie, M., Kameyama, K., Bear, M. F. & Huganir, R. L. Regulation of distinct AMPA receptor phosphorylation sites during bidirectional synaptic plasticity. Nature 405, 955–959 (2000). 96. Luscher, C., Nicoll, R. A., Malenka, R. C. & Muller, D. Synaptic plasticity and dynamic modulation of the postsynaptic membrane. Nature Neurosci. 3, 545–550 (2000).

