Editor's evaluation: Regulation of different phases of AMPA receptor intracellular transport by 4.1N and SAP97

Abstract

Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Changes in the number of synaptic AMPA receptors underlie many forms of synaptic plasticity. These variations are controlled by an interplay between their intracellular transport (IT), export to the plasma membrane (PM), stabilization at synapses, and recycling. The cytosolic C-terminal domain of the AMPAR GluA1 subunit is specifically associated with 4.1 N and SAP97. We analyze how interactions between GluA1 and 4.1N or SAP97 regulate IT and exocytosis in basal conditions and after cLTP induction. The down-regulation of 4.1N or SAP97 decreases GluA1 IT properties and export to the PM. The total deletion of its C-terminal fully suppresses its IT. Our results demonstrate that during basal transmission, the binding of 4.1N to GluA1 allows their exocytosis whereas the interaction with SAP97 is essential for GluA1 IT. During cLTP, the interaction of 4.1N with GluA1 allows its IT and exocytosis. Our results identify the differential roles of 4.1N and SAP97 in the control of various phases of GluA1 IT. Editor's evaluation This important study by Bonnet et al. addresses the question of how AMPA receptor numbers at the synapse are regulated during basal conditions and during chemically induced Long Term Potentiation. Specifically, the study aims to determine the molecular mechanisms that contribute to the intracellular trafficking of AMPA receptors and determine their insertion into the synaptic plasma membrane. Using compelling methodology, the authors dissect the distinct roles of two proteins that bind to the C-terminal domain of the AMPA receptor subunit GluA1: 4.1N and SAP97. The findings will be of interest to anyone trying to understand molecular events contributing to synaptic plasticity in health and disease, and more broadly, the method could be adapted for tracking intracellular movements of a wide range of proteins. https://doi.org/10.7554/eLife.85609.sa0 Decision letter Reviews on Sciety eLife's review process Introduction AMPA-type glutamate receptors (AMPAR) are ionotropic tetrameric receptors activated by glutamate, the main excitatory neurotransmitter of the central nervous system. Their synaptic targeting, clustering, and immobilization in the post-synaptic density in front of glutamate release sites are crucial for efficient excitatory synaptic transmission. The number of neurotransmitter receptors, and particularly AMPAR, present at the synapse is regulated by a complex set of interdependent mechanisms going from biogenesis, IT (Díaz-Alonso and Nicoll, 2021; Hiester et al., 2018), externalization at the PM, lateral diffusion (Choquet and Triller, 2013), stabilization at synapses and trafficking in and out synaptic sites (Groc and Choquet, 2020). In highly polarized and spatially extended neurons, IT is of fundamental importance to distribute cargo over hundreds of micrometers and is likely finely tuned and balanced to control receptor distribution. Accordingly, the IT of neosynthesized AMPAR plays a crucial role to transport them down the dendrites from the Golgi apparatus or Golgi outposts where they are matured after synthesis in the ER. After being released from the Golgi, the secretory vesicles containing the newly-synthesized AMPAR are trafficked to the PM through interaction with adaptor proteins and molecular motors to be finally exocytosed at the PM (Schwenk et al., 2019). We and others provided evidence that AMPAR IT is highly modulated by neuronal activity and this suggests that regulation of IT might be a core constituent of the control of synaptic strength during various forms of synaptic plasticity in neurons (Hangen et al., 2018; Hoerndli et al., 2015). However, despite its potential key role in synaptic regulation, and probable involvement in synaptopathies such as Huntington or Alzheimer diseases (Mandal et al., 2011), the molecular mechanisms that are involved in the regulation of AMPAR IT nevertheless remain largely unknown. AMPAR IT is difficult to study in vertebrates due to the lack of reliable labeling methods and current limitations of imaging systems for detecting fast-moving, low-contrast small vesicles. In cultured rat hippocampal neurons, we have overcome these two hurdles using (1) a molecular tool allowing the retention and on-demand release of the newly synthesized AMPAR from the ER/Golgi and (2) the photobleaching of a portion of a dendrite followed by fast video acquisition (Hangen et al., 2018; Rivera et al., 2000). This allowed the characterization of IT of GluA1 AMPAR subunit, which forms homomeric calcium-permeable receptors, and can be inserted at synapses during synaptic plasticity (Plant et al., 2006; Sanderson et al., 2012). We found that during chemically induced Long Term Potentiation of synaptic transmission (cLTP), the number and velocities of GluA1-containing vesicles are increased compared to the basal state (Hangen et al., 2018). These changes in vesicle velocities may be due to the diversity of molecular motors associated with AMPAR, although the exact motors involved are unknown. Molecular motors associate with their cargo through intermediate components, such as adaptors, scaffolds, and transmembrane proteins (Klopfenstein et al., 2000). AMPAR is part of a macromolecular complex composed of the receptor per se surrounded by a set of associated auxiliary (Bissen et al., 2019) and cytosolic proteins. Some of these intracellular partners have been shown to be associated with motor proteins and can modulate AMPAR surface expression. Of particular interest in this regard, 4.1 N and SAP97 intracellular proteins are directly and specifically associated with GluA1 C-terminal (C-ter.) domain, the most variable domain between the different AMPAR subunits (Diering and Huganir, 2018; Sans et al., 2001; Shen et al., 2000). The C-ter. domain of GluA1 is particularly interesting for the regulation of IT as mutations on this domain modulate its transport and could be responsible for its upregulation during cLTP (Hangen et al., 2018). However, in the recent years, the C-ter. domain has been under intense scrutiny and its role in mediating synaptic plasticity has been debated. On the one hand, the C-ter. domain of native GluA1 and GluA2 has been suggested to be necessary and sufficient to drive NMDA receptor-dependent LTP and LTD, respectively (Zhou et al., 2018). On the other hand, the expression of heteromeric receptors containing the GluA1 subunit lacks the C-ter. domain maintains a normal basal trafficking and LTP at CA1 synapses in acute hippocampal slices (Díaz-Alonso et al., 2020). The method we developed (Hangen et al., 2018) and the results reported here will fuel this debate and allow to determine the exact contribution of GluA1 C-ter. domain for the regulation of IT properties of newly synthesized GluA1 subunit. In red blood cells, the protein 4.1 (4.1R) is critical for the organization and maintenance of the spectrin–actin cytoskeleton and for the attachment of the cytoskeleton to the cell membrane. 4.1 N, the neuronal form of 4.1, may function to confer stability and plasticity to the neuronal membrane via interactions with multiple binding partners such as spectrin-actin–based cytoskeleton, integral membrane channels, and receptors. In neurons, 4.1 N associates specifically with GluA1 and colocalizes with AMPAR at excitatory synapses (Walensky et al., 1999). The C-ter. domain of 4.1 N mediates the interaction with the membrane-proximal region of GluA1. It has been suggested that 4.1 N regulates AMPAR trafficking by providing a critical link between the actin cytoskeleton and AMPAR (Shen et al., 2000). Phosphorylation of S816 and S818 residues in GluA1 regulates activity-dependent GluA1 insertion at the PM by enhancing the interaction between 4.1 N and GluA1. This suggests that 4.1 N is important for the expression of LTP, but doesn’t affect basal synaptic transmission (Lin et al., 2009). However, while the regulation of GluA1 exocytosis by binding to 4.1 N has been established, its potential involvement in AMPAR IT still remains unknown. SAP97, another important GluA1 C-ter. domain interactor is a member of the MAGUK family of proteins that play a major role in the trafficking and targeting of membrane ion channels and cytosolic structural proteins in multiple cell types (Fourie et al., 2014). Within neurons, SAP97 is localized throughout the secretory trafficking pathway and at the postsynaptic density (PSD). The role of SAP97 in the control of synaptic function is still unclear despite the fact that the PDZ2 domain of SAP97 interacts directly with the last four amino acids of GluA1 (Cai et al., 2002). The interaction between SAP97 and GluA1 occurs early in the secretory pathway, while the receptors are in the ER or cis-Golgi, and participates in its forward trafficking from the Golgi to the PM, suggesting that SAP97 acts on GluA1 solely before its synaptic insertion and that it does not play a major role in anchoring AMPAR at synapses (Sans et al., 2001; Fourie et al., 2014). SAP97 is a protein known for its involvement in NMDAR (Jeyifous et al., 2009) and AMPAR IT, thanks to its role as an adaptor protein between GluA1 and the actin-based motor MyoVI (Wu et al., 2002). However, the role of SAP97 in the trafficking and synaptic localization of AMPAR is still debated with conflicting results have been reported (Fourie et al., 2014; Kay et al., 2022; Zhou et al., 2008). Moreover, its role in the induction and maintenance of LTP is yet not well characterized. Here, we have a unique experimental pipeline that allows us to differentiate IT from exocytosis of a given protein by measuring them independently. We report the role of the interactions between 4.1 N and SAP97 with the C-ter. domain of GluA1 by analyzing IT and exocytosis of newly synthesized GluA1 deleted for this domain under basal conditions and during synaptic activity. We identify different roles of the interactions between 4.1N-GluA1 and SAP97-GluA1 during basal transmission and after induction of cLTP in hippocampal cultured neurons. Results To study the properties of AMPAR IT, we used cDNA constructs to express GluA1 and its different mutants subcloned in the ARIAD system and tagged at its N-terminus (ARIAD-Tag-GluA1) (Hangen et al., 2018). With this technology, receptors are retained in the ER in the basal state thanks to a conditional aggregation domain. Receptor release from the ER and follow-up secretion is tightly controlled with a cell-permeant drug (D/D solubilizer or ARIAD ligand: AL) that disrupts aggregation (Rivera et al., 2000). The synchronized release of receptors triggered by the addition of AL allows expressed proteins to progress through the secretory pathway in a synchronous manner, particularly adapted to monitor IT. Important features of this system include (1) no or low basal secretion and (2) a rapid and high level of secretion in response to the addition of AL (Hangen et al., 2018). This allowed us to measure three main parameters of AMPAR intracellular trafficking. First, the total number of GluA1-containing vesicles after the synchronized release of receptors was used as a measure of ER/Golgi export efficiency. Second, GluA1 vesicle transport properties (speed, fraction of time spent moving or pausing) were measured. Third, we determined the kinetic and extent of GluA1 appearance on the cell surface by live immunolabeling at various times after release. The comparative measurement of these different parameters allowed us to decipher finely the regulatory steps of GluA1 intracellular transport. 4.1 N and SAP97 are important proteins implicated in the regulation of AMPAR PM localization. Among all AMPAR subunits, these two proteins are specifically associated with the GluA1 C-ter. domain (Lin et al., 2009; Rouach et al., 2005; Schwenk et al., 2014), (Figure 1—figure supplement 1A). We analyzed how interactions between these associated proteins and GluA1 regulate AMPAR IT in basal conditions and after induction of cLTP in cultured rat hippocampal neurons. GluA1 intracellular transport and exocytosis are dependent on the expression of 4.1N or SAP97 4.1 N and SAP97 participate in the biosynthesis and processing of AMPAR in the hippocampus (Sans et al., 2001; Shen et al., 2000). Previous studies established that knocking down 4.1 N by expression of a specific sh-RNA substantially reduced the frequency of GluA1 exocytosis, indicating that 4.1 N is critical for GluA1 insertion at the PM (Lin et al., 2009). On the other hand, SAP97 has been shown to associate with GluA1 containing AMPAR while they are in the ER, with SAP97 dissociating from the receptor at the PM (Sans et al., 2001). We decided to knock down each of these two proteins independently and analyze IT and externalization of newly synthesized GluA1 in basal condition. We expressed sh-RNAs against 4.1 N or SAP97 or their corresponding control (scramble) and analyzed the trafficking of ARIAD-Tag-GluA1 after the addition of the ligand to release the protein from the ER (Figure 1). Figure 1 with 1 supplement see all Download asset Open asset Intracellular transport and exocytosis of GluA1 are dependent on the expression of 4.1N and SAP97. ( A) Top: Western blots of 4.1N and SAP97 expression in cultured rat hippocampal neurons after virus infection with scramble-RNA (scr.) or sh-RNA against 4.1N and SAP97; bottom: quantification of proteins normalized with actin on the scr. condition (sh-4.1N: 44.4 +/−11.2%,n=5, sh-SAP97: 58.3 +/−7.3%, n=4). (B) Top: Western blots showing the expression of 4.1N and SAP97 WT and rescue after transfection of the proteins in COS-7 cells; bottom: quantifications normalized with actin on the scr. condition (sh-4.1N: 26.2 +/−10.1%, sh-SAP97: 47.1 +/−10.7%, n=4; for rescue proteins; scr.: 105.9 +/−7.9%, sh-4.1N: 141.2 +/−3.1%, scr.: 87.0 +/−1.5%, sh-SAP97: 93.0 +/−2.2%, n=4). (C to G) Parameters of intracellular transport of ARIAD-TdTom-GluA1 expressed with scramble-RNA (scr.) or sh-RNA against 4.1N and SAP97. (C) Vesicle number (vesicles/20 µm2/min; scr. 4.1 N: 17.2 +/−2.7, sh-4.1N: 9.3 +/−1.7, n=3 scr. SAP97: 19.5 +/−2.5; sh-SAP97: 4.2 +/−0.7, n=4), (D) Representative kymographs of the routes of the vesicles in the function of the time in the video. (E) Mean speeds of the vesicles in control (expression of scr.) and when 4.1 N or SAP97 are decreased (expression of sh) (µm/s; scr. 4.1 N: 1.56 +/−0.07, sh-4.1N: 1.35 +/−0.05, n=3; scr. SAP97: 1.60 +/- 0.05, sh-SAP97: 1.21 +/- 0.07, n=4), (F–G) Time spent by a vesicle in a moving state (Move) or in pausing state (Pause) (% Move: scr. 4.1N: 73.13 +/−0.83, sh-4.1N: 67.77 +/−1.25; % pause: scr. 4.1N: 26.87 +/−0.83, sh-4.1N: 32.22 +/−1.25) and (% Move: scr. SAP97: 75.57 +/−5.89, sh-SAP97: 60.68 +/−1.74; % pause: scr. SAP97: 24.43 +/−5.89, sh-SAP97: 39.32 +/−1.75) (n=3) (H) Representative image of live extracellular labeling of ARIAD-GFP-GluA1 after 45 and 60 min. of incubation with AL expressed with sh-RNA for 4.1N with or without the corresponding rescue proteins and quantifications (% of cle. 45 min. after AL; sh-4.1N: 57.36 +/- 6.27, sh-4.1N on rescue: 105.00 +/- 10.34; 60 min after AL; sh-4.1N: 37.29 +/−4.16, sh-4.1N on rescue: 89.73 +/−7.81) (n=3). (I) Representative image of live extracellular labeling of ARIAD -GFP-GluA1 after 45 and 60 min of incubation with AL expressed with sh-RNA for SAP97 with or without the corresponding rescue proteins and quantifications (% of cle. 45 min after AL; sh-SAP97: 49.64 +/−3.41, sh-SAP97 on rescue: 72.91+/−5.58, 60; 60 min after AL; sh-SAP97: 33.33 +/−3.29, sh-SAP97 on rescue: 61.69 +/−4.32) (n=3). The 100% values for H and I correspond to the extracellular labeling of the control (Scramble: Scr.) for the same times of incubation with AL. Scale bar: 25 µm. Figure 1—source data 1 Individual data values for the bar graphs in panels A, B, C, E, F, H and I. Raw images and summary of the western blots from panels A and B. https://cdn.elifesciences.org/articles/85609/elife-85609-fig1-data1-v1.zip Download elife-85609-fig1-data1-v1.zip We first controlled the efficacy of the sh-RNA in rat-cultured hippocampal neurons by expressing viruses containing respectively scramble RNA (scr.), sh-RNA against 4.1 N (sh-4.1N) or against SAP97 (sh-SAP97) (Figure 1A). Expression of the endogenous proteins were significantly decreased by expression of the corresponding sh-RNA. To test the specificity of the sh-RNA, we expressed 4.1 N or SAP97 or the corresponding rescue proteins in COS-7 cells together with the scr.-RNA or the sh-RNA and quantified expression of 4.1 N and SAP97 by western blot analysis (Figure 1B). As in neurons, expression of sh-RNA decreased the expression of the corresponding wild type proteins without affecting the expression of the corresponding rescue proteins showing the specificity of our sh-RNA. We then analyzed the parameters of GluA1 IT when 4.1 N or SAP97 sh-RNA or corresponding scr.-RNA were expressed (Figure 1C–G). We expressed GluA1 subcloned in the ARIAD vector, induced the transport of the protein by the addition of the AL, and analyzed the transport 30–60 min after the addition of the AL. During this time window, the vesicles are traveling in dendrites with almost no background signal coming from the PM (Hangen et al., 2018). In both cases, the total number of vesicles transporting GluA1 was decreased, although less drastically when 4.1 N was knocked down than when SAP97 was knocked down (Figure 1C). For each cell, we traced the corresponding kymographs and calculated the mean speed of the vesicles (Figure 1D–E, Sup. Figure 1B). We found similar values for speed for the OUT (from the cell body to the dendrite) and for the IN (from the dendrite to the cell body) directions (Figure 1—figure supplement 1C). We thus decided to pool the speeds of transport of the OUT and the IN directions. The mean speed of the vesicles was only decreased by 13% by expression of the sh-4.1N compared to its scr.-RNA (Figure 1E, Figure 1—figure supplement 1C). When sh-SAP97 was expressed, the mean speed was decreased by 25% compared to the corresponding scr.-RNA. Thanks to the kymographs, we calculated the percentage of time spent in the moving and in the pausing states for each vesicle (Figure 1F–G, Figure 1—figure supplement 1D–E). The time spent moving was decreased by 8% when the sh-4.1N was expressed to the benefit of the pausing time. Indeed, when the expression of SAP97 was decreased, the time spent moving by a vesicle was decreased by 20% to the benefit of the time spent in pause. We then analyzed the kinetics of externalization of GluA1 in the same conditions as for the IT experiments. We performed live extracellular labeling of GluA1 at 45 and 60 min after the addition of the AL on hippocampal rat cultured neurons (Figure 1H–I). Expression of sh-4.1N decreased massively the externalization of GluA1 compared to its externalization with the expression of the corresponding scr. (Figure 1H). The 4.1 N rescue protein could prevent this decrease in the rate of externalization when expressed together with the sh-4.1N. Expression of sh-SAP97 also decreased the rate of externalization of GluA1 to the same extent as sh-4.1N did (Figure 1I). Expression of SAP97 rescue protein partially restored the externalization of GluA1. In conclusion, reducing the expression of 4.1 N and SAP97 both diminished GluA1 IT in rat-cultured hippocampal neurons. However, the effects were all less drastic when 4.1 N expression was decreased than when SAP97 was. This was the case for the decrease in the number of vesicles released upon the addition of AL, for the decrease in the vesicle speed, and for the increase in pausing time. However, we found that the externalization of GluA1 at the PM was equally inhibited by the absence of 4.1 N or SAP97. Because the down-regulation of 4.1 N or SAP97 could have indirect effects on GluA1 transport properties, we then studied the impact of GluA1 mutations that inhibit its interaction with these proteins. AMPAR traffic is regulated by the interaction between GluA1 C-ter. domain and 4.1N or SAP97 GluA1 IT and PM localization is dependent on the expression of 4.1 N and SAP97. Knocking down either 4.1 N or SAP97 decreases massively the exit of GluA1 from the ER-Golgi and impacts IT and exocytosis of the receptor to the PM. However, the overall impact of SAP97 is more drastic on IT whereas we found the same effect for both conditions on the PM localization of GluA1. This may be because the interaction between 4.1 N and GluA1 might be necessary mainly for the exocytosis of the receptor at the PM. We thus decided to analyze if the interaction between GluA1 and 4.1 N or GluA1 and SAP97 is important for the intracellular transport and exocytosis of the newly synthesized receptor in basal synaptic transmission (Figure 2). Figure 2 with 1 supplement see all Download asset Open asset 4.1N/GluA1 and SAP97/GluA1 interactions differently regulate GluA1 traffic in basal transmission. (A) Co-immunoprecipitation of endogenous GluA1 with 4.1N and SAP97 in cultured rat hippocampal neurons. Control (Ctl.) is performed without an antibody. Western blot of GluA1, 4.1N and SAP97 as indicated. (B) Diagram of the different truncated mutants on the C-terminal (C-ter.) domain of GluA1. (C) Representative images of live labeling of ARIAD-GFP-GluA1 after the addition of AL during different times as indicated. Scale bar: 25 µm. (D) Quantification of the exit of ARIAD-GFP-GluA1-WT (WT) and ARIAD-GFP-GluA1-Δ78 (Δ78) over time after the addition of AL. For the WT, 100% of exit is taken after 120 min of addition of AL. (WT versus Δ78, arbitrary unit (a. (u).): 30 min, 20.43 +/−2.14 vs 45.02 +/−7.75; 60 min, 49.39 +/−3.42 vs 9.54 +/−1.95; 90 min, 102.68 +/−5.83 vs 13.64 +/−1.70; 120 min, 101.95 +/−5.37 vs 22.02 +/−4.59) (E) Quantification of the exit of ARIAD-GFP-GluA1-Δ4.1N (Δ4.1N) and ARIAD-GFP-GluA1-Δ7 (Δ7) over time after addition of AL induction (Δ4.1N: 30 min, 93.22 +/−10.00; 60 min, 68.75 +/−4.43; 90 min, 47.94 +/−4.01; 120 min, 41.19 +/− 2.70; Δ7: 30 min, 90.07 +/−7.42; 60 min, 77.98 +/−7.26; 90 min, 67.35 +/−5.68; 120 min, 72.04 +/−6.60). The 100% values correspond to the value of the WT for the same time, shown by a dotted line. (F) Traced kymographs for the different mutants. (G) Number of vesicles detected for the ARIAD-TdTom-GluA1-WT (WT) and the ARIAD-TdTom-GluA1-Δ78 (Δ78) (vesicles / 20 µm2/min; GluA1-WT: 12 +/−2.3, Δ78: 0.86 +/−0.2, n=4). (H) Parameters of intracellular transport for ARIAD-TdTom-GluA1-Δ4.1N (Δ4.1). Vesicle number (vesicles/20 µm2/min; GluA1-WT: 18.68 +/−2.05, Δ4.1N: 10.52 +/−1.40, n=5), mean speeds (µm/s; WT: 1.55 +/−0.03, Δ4.1N: 1.64 +/−0.05) and percentage of time in each state (% Move WT: 76.75 +/−0.53 %, Δ4.1N: 78.52 +/−0.78%; % pause: WT: 23.10 +/−0.53 %, Δ4.1N: 21.31 +/−0.78 %) (n=5). (I) Parameters of intracellular transport for the ARIAD-TdTom-GluA1-Δ7 (Δ7). Vesicle number (vesicles/20 µm2/min; GluA1-WT: 14.45 +/−1.82, Δ7: 8.16 +/−1.21), mean speeds (µm/s; WT: 1.48+/−0.03, Δ7: 1.30+/−0.05) and percentage of time in each state (% Move: WT: 75.18+/−0.70%, Δ7: 69.67+/−1.04%; % pause: WT: 24.82+/−0.69%, Δ7: 30.33+/−1.04%) (n=4). Figure 2—source data 1 Individual data values for the bar graphs in panels D, E, G, H and I. Raw images and summary of the western blots from panel A. https://cdn.elifesciences.org/articles/85609/elife-85609-fig2-data1-v1.zip Download elife-85609-fig2-data1-v1.zip We first checked by co-immunoprecipitation experiments if endogenous GluA1 and 4.1 N and GluA1 and SAP97 are interacting in our model of rat-cultured hippocampal neurons (Figure 2A). Indeed, we found that immunoprecipitation of 4.1 N or SAP97 co-immunoprecipitated GluA1 and immunoprecipitation of GluA1 co-immunoprecipitated 4.1 N and SAP97. It has been shown that GluA1 binds SAP97 by its last four amino acids (Leonard et al., 1998) whereas it binds 4.1 N on a peptide domain localized just after its fourth transmembrane domain (Shen et al., 2000). We designed different GluA1 mutants in order to study the impact of these interactions on GluA1 IT (Figure 2B). We deleted the entire C-ter. domain of GluA1 (deletion of the last 78 amino acids of GluA1 leaving only four amino acids after the last transmembrane domain, Δ78) or each of the interaction sites for 4.1 N (Δ4.1N: deletion of 14 amino acids and five amino acids after the last transmembrane domain) and for SAP97 (Δ7: deletion of the last seven amino acids of GluA1). For each mutant, we studied their PM localization as a function of the time of incubation with the AL and the characteristic of their IT. The externalization of newly synthesized GFP-GluA1-Δ78 or GFP-GluA1-WT was monitored by live immunolabeling with an antibody directed against GFP (Figure 2C and D). Quantification of the GFP staining revealed an almost complete disappearance of GluA1-Δ78 externalizations. This experiment demonstrates that the C-ter. domain is necessary for newly synthesized GluA1 to be externalized at the PM, even 2 hr after triggering GluA1 ER exit. We then studied if the interaction between GluA1 and 4.1 N or SAP97 is necessary for the localization of GluA1 at the PM (Figure 2C and E, Figure 2—figure supplement 1A). We performed extracellular labeling of GFP-GluA1-WT and the mutants, GFP-GluA1-Δ4.1N, and GFP-GluA1-Δ7 respectively deleted for their binding site for 4.1 N or SAP97, after induction of IT by addition of the AL during different times. For analysis of these experiments, we normalized the externalization values of the mutants to that of GFP-GluA1-WT at the corresponding times of incubation with AL in paired experiments. Both mutants were less exocytosed than GluA1-WT from 30 min to 2 hr after the addition of the AL. At each time point, the extracellular labeling of GluA1 lacking the 4.1 N binding site (Δ4.1N) was inferior to the one lacking the SAP97 binding site (Δ7). Indeed, when we quantified at 90 and 120 min. the difference in exocytosis between the two mutants, we found that this difference between Δ4.1N and Δ7 proteins was highly significant (Figure 2—figure supplement 1A). Exocytosis of the Δ4.1N mutant is less important than for the Δ7 mutant (Supplementary file 1). This result shows that the interaction of GluA1 with 4.1 N or SAP97 plays a role in the surface expression of newly synthesized GluA1. This lack of normal exocytosis of the mutants can be due either to an inhibition of externalization or to a decrease in their IT. We thus analyzed the IT parameters of the different mutants taking GluA1-WT as a control (Figure 2F–I, Sup. Figure 2—figure supplement 1B and C). For these experiments, we expressed ARIAD-TdTom-GluA1, WT, or mutants, in order to be in the best conditions to detect the transport vesicles. We first analyzed the IT of the Δ78 mutants (Figure 2F–G). The number of vesicles transporting this mutant was very low compared to GluA1-WT and this prevented the analysis of their IT parameters. The C-ter. domain of GluA1 is thus mandatory for the exit of newly synthesized GluA1 from the ER and the Golgi, likely explaining its requirement for GluA1 surface expression. We then characterized the IT of the mutant deleted for the 4.1 N binding site (Δ4.1N) (Lin et al., 2009; Figure 2F and H, Figure 2—figure supplement 1B and C). The number of vesicles transporting the protein was decreased compared to GluA1-WT. This is in accordance with a decrease in the number of vesicles that we found when 4.1 N was knocked down by the expression of the sh-4.1N. In contrast, the mean speeds of the vesicles were the same for GluA1-WT and GluA1-Δ4.1N. The time spent in each state was similar for the two proteins. These results indicate that binding of GluA1 to 4.1 N is important for its ER or Golgi export and exocytosis at the PM but does not affect its IT once the vesicles are released from the Golgi apparatus. We then analyzed the IT of the mutant deleted for the last seven amino acids corresponding to the binding site of SAP97 (Δ7) (Zhou et al., 2008; Figure 2F1, Figure 2—figure supplement 1B and C). For this mutant, the number of vesicles released was decreased compared to GluA1-WT, as for GluA1-Δ4.1N. We also found a highly significant effect of the Δ7 mutant deletions of the PDZ binding domain on all the GluA1 IT parameters. First, the mean speed was decreased by 12% for the Δ7 mutants. Moreover, the time spent in a moving state was decreased by 7% and, conversely, the percentage of time in pause was increased by 22% compared to the WT. All these changes, although relatively modest, are significant and can have important functional impacts when cumulated over time. This corresponds to what we found for IT properties when the expression of SAP97 was decreased: modified ER/Golgi export and PM exocytosis, reduced speed, and increased time in pause. For these two mutants, the vesicle number is similarly decreased but if the mean speed is also decreased for the Δ7 this is not the case for the Δ4.1N which has a completely normal IT speed. On the contrary, the exocytosis is largely decreased for the Δ4.1N protein (Supplementary file 1). AMPAR traffic is regulated by the specific interaction between GluA1 C-ter. domain with 4.1N Since we observed an impact on GluA1 IT when 4.1 N was knocked down, we were surprised by the absence of impact on IT of the Δ4.1N deletion on GluA1 IT. We thus decided to analyze the characteristics of IT with a GluA1 mutant that does not bind 4.1 N and has the same C-ter. domain length (Figure 3, Figure 3—figure supplement 1). During LTP, protein kinase C (PKC) phosphorylates the serine 81