Get Complete Project Material File(s) Now! »
Delivery of AMPARs to the Synapse
AMPARs need to be delivered to the somatodendritic compartment of the neuron where they are inserted into the plasma membrane by exocytosis. Direct visualization of the intracellular transport of GFP-GluA1 and GFP-GluA2 using FRAP (fluorescence recovery after photobleaching) measurements revealed that AMPARs are transported at rates comparable with fast axonal transport and move in a predominantly, but not exclusively, proximal to distal direction (Perestenko and Henley, 2003). This suggests that the vesicular transport of AMPAR using microtubules requires both motor proteins, kinesin and dynein. AMPAR-containing vesicles can also use actin filaments via the Ca++-sensitive motor proteins Myosin Va (Correia et al., 2008) and Myosin Vb (Lisé et al., 2006). An alternative for the AMPAR delivery to appropriate synaptic destinations is local protein synthesis. An interesting study using quantitative high-resolution in situ hybridization in cultured hippocampal neurons demonstrated that an important number of synapses contains GluA1 or GluA2 mRNAs (Grooms et al., 2006), suggesting that granules containing AMPAR mRNAs are positioned close to the spines. GRIP1 and ABP are localized at the PSD and also in some intracellular pools. It has been shown that GRIP1 can bind the heavy chain of the motor protein kinesin, KIF5 (Kinesin-like protein 5) and the protein liprin α (Setou et al., 2002; Shin et al., 2003; Wyszynski et al., 2002). In parallel, liprin α interacts with KIF1A (Kinesin-like protein 1A, also known as axonal transporter of synaptic vesicles or microtubule-based motor KIrF 1A), which can co-immunoprecipitate with AMPAR (Shin et al., 2003). Taken together, this indicates that the AMPAR/GRIP/ABP complex can be transported along dendrites by several kinesin motor proteins attaching to microtubules with the help of KIF5, KIF1A and liprin α.
AMPARs Membrane Insertion
After AMPAR transport at the delivery place, SNARE-mediated fusion events at the plasma membrane lead to the exocytosis of the AMPARs, as illustrated in Figure 10 (reviewed in (Jurado, 2014)).
Figure 10: AMPAR exocytosis. Sequential steps in the SNARE-mediated AMPAR exocytosis.
It is known that the kinetics for exocytosis are different between AMPAR subunits. GluA1 insertion is constitutively slow and accelerated by activity, whereas GluA2 exocytosis is rapid under basal conditions (Passafaro et al., 2001). A GluA2 mutant that cannot bind NSF present slower exocytosis kinetics than GluA2. A GluA3 mutant including an NSF binding site, present the same kinetics as GluA2 (Beretta et al., 2005). This well-grounded experiment suggests that NSF is required for fast AMPAR incorporation into the synaptic plasma membrane. The protein 4.1N is required for GluA1 insertion, and its binding to GluA1 is enhanced by PKC-dependent phosphorylation of the S816 and S818 of GluA1 CTD. Because of the overlapping binding sites, when GluA1 interacts with 4.1N, it cannot bind AP-2 for endocytosis. This suggests an opposite role for 4.1N and AP-2. Disruption of 4.1N by shRNA leads to a decreased surface expression of the GluA1-containing AMPARs and a defective hippocampal long-term potentiation of synapses (Lin et al., 2009). This suggests that the interaction between GluA1 and 4.1N is required for AMPAR insertion both constitutively and in an activity-dependent manner.
A complex formed by SAP97 and AKAP79 (PKA anchoring molecule) tethers PKA for GluA1 phosphorylation (Colledge et al., 2000). Phosphorylation of the GluA1 CTD residue S845 by PKA perturbs hippocampal LTD, and dephosphorylation facilitates GluA1 activity-dependent internalization (Lee et al., 2000). This suggests that the GluA1/SAP97/AKAP79 interaction blocks LTD by preventing GluA1 internalization. In agreement, one study showed that overexpression of SAP97 increased the amplitude of AMPAR-EPSC and potentiated hippocampal LTP (Nakagawa et al., 2004). Together, these data suggest that the interaction between GluA1 and SAP97 is needed for the maintenance of surface GluA1 and the control of LTP. However, SAP97 conditional KO mice in another study, presented normal hippocampal LTP (Howard et al., 2010), and thus the role of SAP97 in activity-dependent AMPAR exocytosis remains unclear. The phosphorylation of S831 (GluA1) by CaMKII has been shown to be necessary for hippocampal LTP (Lee et al., 2000), indicating that this regulation promotes GluA1-containing AMPAR activity-dependent insertion.
Some experiments have shown that postsynaptic SNARE-mediated membrane fusion is required for activity-dependent AMPAR insertion (Kennedy et al., 2010; Lu et al., 2007). Research has focused on ubiquitous SNARE proteins involved in both constitutive and activity-dependent AMPAR exocytosis. A role for complexin, a postsynaptic protein that regulates neurotransmitter release, has been suggested in activity-dependent AMPAR exocytosis but not constitutive AMPAR insertion. Complexin is required for NMDAR-triggered insertion of AMPAR in hippocampal cultured neurons (Ahmad et al., 2012). Complexin exhibits a strong binding affinity to SNARE complexes containing syntaxin-1A, -2 or -3 but does not bind to SNARE complexes containing syntaxin-4 that characterizes the spine exocytic zone (Pabst et al., 2000). These data suggest that the complexin-mediated activity-dependent AMPAR insertion takes place at the dendritic shaft (Kennedy et al., 2010).
It is uncertain where the AMPARs exocytic zones are positioned in relationship to the PSDs. One study using photoreactive AMPAR antagonists and electrophysiology proposed that AMPARs are inserted into the plasmalemma at the level of the soma and laterally diffuse across very long distances to synapses (Adesnik et al., 2005). By imaging superecliptic pHluorin (SEP)-labeled AMPARs under total internal reflection fluorescence microscopy, insertion events were observed in both the soma and dendritic shafts, but never in spines (Lin et al., 2009). One study using very sensitive real-time imaging has suggested that AMPARs are inserted in the dendritic shaft, close to, but not at, dendritic spines (Yudowski et al., 2007). It was suggested that NMDAR-dependent LTP recruits GluA1 from the internalized pool (only 20% of the total recruitment), primarily onto the dendritic shaft (Makino and Malinow, 2009). In discordance, Corera and collaborators examined the AMPAR trafficking in isolated synaptosomes during LTP induction. They found an increased level of GluA1 and GluA2 at the PSD with no changes in the total levels, suggesting that exocytosis can happen close to the PSD (Corera et al., 2009). It has recently been proposed that there are two different sites for AMPAR insertion: one in the spine and another in the dendritic shaft. Combining two-photon glutamate uncaging with two-photon imaging of GFPGluA1 or SEPGluA1 in transfected CA1 hippocampal neurons from slices, Patterson et al. monitored individual AMPAR exocytosis events (Figure 11A and B). They describe exocytosis in both, dendrites and spines, under basal conditions, and they observe a ~ 5-fold increase in exocytosis in spines and dendrites during LTP induction (Patterson et al., 2010). A recent study used high-resolution live cell imaging to visualize activity-dependent exocytosis (Figure 11C). It was revealed that activity triggers a massive exocytosis of AMPAR-containing endosomes in dendritic spines, more precisely, in submicron membrane clusters of syntaxin 4 positioned laterally to the PSD (Kennedy et al., 2010). In conclusion, it seems clear that there is an area of AMPAR exocytosis at the dendritic shaft. Other observations would seem to suggest that this AMPAR exocytic zone is insufficient and support the existence of other exocytosis zones, placed at the dendritic spine and enriched in syntaxin 4. The hypothesis is that both exocytic zones could work under basal conditions, and that activity could increase the exocytosis at the dendritic spine.
Figure 11: AMPAR exocytosis events at the dendritic shaft and at the spine. (A-B) mCherry (upper) and SEPGluA1 (lower) images of spines undergoing stimulation. Stimulated regions are shown by open arrowhead. Exocytosis is shown by closed arrowhead. From (Patterson et al., 2010). (A) Example of transient spine exocytosis (left) and example of sustained spine exocytosis (right). (B) Example in a dendrite immediately beneath a spine, showing movement of fluorescence into the stimulated spine (left). Another example that is 2 µm away (right). Scale bar: 1 µm. (C) Syntaxin 4 marks sites of exocytosis at the spine. TfR-SEP (top row) and surface syntaxin 4-HA (middle row) signals were imaged during spontaneous exocytic events. Scale bar: 500 nm. From (Kennedy et al., 2010).
Surface AMPARs: Stabilization, Anchoring and Clustering
It has been shown that the protein 4.1N is required not only for GluA1 insertion (Lin et al., 2009), but also for GluA1 membrane stabilization. The GluA1 subunit can bind with the protein 4.1N and form a complex (GluA1/4.1N) that favors the stabilization of AMPARs at the plasma membrane. The protein 4.1N interacts with the actin filaments, making a link between the AMPAR and the cytoskeleton. Overexpression of 4.1N or the use of latrunculin, which inhibits actin polymerization in filaments, all reduce GluA1 surface expression (Shen et al., 2000). The interaction between GluA2 and NSF could also contribute both to the insertion and to membrane stabilization of AMPARs. NSF does not perturb GluA2 surface expression (Beretta et al., 2005; Lee et al., 2002), but a run-down of EPSCs are recorded in the presence of a peptide disrupting GluA2/NSF binding (Lee et al., 2002). A GluA2 mutant that cannot interact with NSF cannot be delivered to synapses in cultured hippocampal neurons (Shi et al., 2001). Indeed, the addition of the NSF binding site to GluA3 is sufficient to direct membrane insertion at synapses (Beretta et al., 2005). Other studies have indicated that the protein kinase M zeta (PKMζ) promotes the interaction of GluA2-containing AMPARs with NSF and enhances AMPAR responses at Schaffer collateral/CA1 synapses (Yao et al., 2008). The protein 4.1N mediates the interaction with the transmembrane GluA2 and the cytoskeleton, while NSF and PKMζ have a role in the insertion of GluA2 in the proper place.
AMPARs can be theoretically anchored at PSDs on their own (GluA2L and GluA4 have a PDZ binding site) or via AMPAR interactors (particularly stargazin and, in general, other TARPs or Shanks containing a PDZ binding site). It was shown that the application of dimeric stargazin CTD peptides in the patch pipette produces a reduction of 50% in the amplitude of AMPAR-mediated EPSCs (Sainlos et al., 2011), suggesting that stargazin is necessary for AMPAR anchoring at the PSD. The interaction of GluA2-containing AMPAR with GRIP1/ABP that both contain several PDZ domains, also seems to be necessary for the anchoring and is modulated by phosphorylation at S880. The unphosphorylated GluA2 prefers the GRIP1/ABP interaction rather than the one with PICK1. The interaction between AMPAR and ABP, that inhibits the S880 phosphorylation, favors the GRIP/ABP binding (Seidenman et al., 2003). After phosphorylation, GluA2 might dissociate from GRIP/ABP and bind PICK1 (Reviewed in (Collingridge et al., 2004)). Another phosphorylation of the AMPAR CTD at the residue Y876 by the Src tyrosine kinase reduces the binding with GRIP/ABP but not the binding of PICK1 (Hayashi and Huganir, 2004). GluA2 constructs that disrupt the interaction with GRIP1/ABP induce a dramatic decrease of surface GluA2. The authors hypothesized that the PDZ protein GRIP1 is only present in the plasma membrane, proposing a role for GRIP1 in AMPAR surface anchoring (Osten et al., 2000). There are roles for other synaptic proteins in the anchoring of AMPAR. The suppression of the neuronal K-Cl cotransporter KCC2 limits the aggregation of GluA1 with other transmembrane proteins at the surface (Gauvain et al., 2011). This effect likely involves KCC2 interaction with submembrane actin cytoskeleton through its C-terminal domain but not its ion transport function.
Table of contents :
PART 1: SYNAPTIC DEVELOPMENT NEVER STOPS
1. The Formation and Maturation of Excitatory Synapses
2. Synaptic Plasticity
2.1. Short-Term Synaptic Plasticity
2.2. Long-Term Synaptic Plasticity
3. AMPA-type Receptor Turnover
3.1. AMPA-type Receptors
3.2. AMPA-type Receptor Turnover
3.2.1. Delivery of AMPARs to the Synapse
3.2.2. AMPARs Membrane Insertion
3.2.3. Surface AMPARs: Stabilization, Anchoring and Clustering
3.2.4. Lateral Diffusion of AMPARs
3.2.5. Endocytosis of the AMPAR: Internalization
3.2.6. AMPARs Recycling
3.2.7. Degradation of AMPARs
3.2.8. Colophon
PART 2: THE OLIVOCEREBELLAR NETWORK AS A MODEL TO STUDY EXCITATORY SYNAPSE FORMATION AND FUNCTION
1. Organization of the Olivocerebellar Network
1.1. Global Structure of the Cerebellum and Inferior Olive
1.2. Histology of the Cerebellar Cortex
1.3. Cerebellar Efferences and Afferences
1.4. Afferences and Efferences of the Inferior Olive
2. The Life of Purkinje Cell Excitatory Synapses
2.1. The Parallel Fiber/Purkinje Cell Synapse
2.2. The Climbing Fiber/Purkinje Cell Synapse
3. Cerebellar Functions
3.1. Sensorimotor Cerebellar Functions
3.2. Cognitive Cerebellar Functions
PART 3: SUSHI DOMAIN PROTEINS IN THE NERVOUS SYSTEM
1. The Sushi Domain
2. Sushi Domain-containing Proteins are Evolutionarily Conserved in the Nervous System
2.1. Caenorhabditis elegans: LEV-9
2.2. Drosophila melanogaster: Hig and Hasp
2.3. Vertebrata
2.3.1. GABAB Receptors
2.3.2. SEZ-6
2.3.3. SRPX2
2.3.4. SUSD Family
2.3.5. CSMD Family