An integrative model of AMPA receptor trafficking reveals the central contribution of local translation in subtype-specific kinetics

An integrative model of AMPA receptor trafficking reveals the central contribution of local translation in subtype-specific kinetics | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results An integrative model of AMPA receptor trafficking reveals the central contribution of local translation in subtype-specific kinetics View ORCID Profile Surbhit Wagle , Maximilian K. Kracht , Anne Bührke , Amparo Acker-Palmer , View ORCID Profile Nataliya Kraynyukova , View ORCID Profile Anne-Sophie Hafner , View ORCID Profile Erin M. Schuman , View ORCID Profile Tatjana Tchumatchenko doi: https://doi.org/10.1101/2025.02.08.637220 Surbhit Wagle 1 Institute of Experimental Epileptology and Cognition Research, University of Bonn Medical centre , Venusberg-Campus 1, 53127 Bonn, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Surbhit Wagle Maximilian K. Kracht 2 Buchmann Institute of Molecular Life Sciences (BMLS), Goethe Universität , Frankfurt Am Main, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site Anne Bührke 4 Max Planck Institute for Brain Research , Synaptic Plasticity, Frankfurt Am Main, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site Amparo Acker-Palmer 2 Buchmann Institute of Molecular Life Sciences (BMLS), Goethe Universität , Frankfurt Am Main, Germany 5 Institute of Cell Biology and Neuroscience, Goethe Universität , Frankfurt Am Main, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site Nataliya Kraynyukova 1 Institute of Experimental Epileptology and Cognition Research, University of Bonn Medical centre , Venusberg-Campus 1, 53127 Bonn, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Nataliya Kraynyukova For correspondence: nkrayn{at}uni-bonn.de anne-sophie.hafner{at}donders.ru.nl tatjana.tchumatchenko{at}uni-bonn.de Anne-Sophie Hafner 3 Donders Institute for Brain, Cognition and Behavior, Radboud University , Nijmegen, Netherlands Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Anne-Sophie Hafner For correspondence: nkrayn{at}uni-bonn.de anne-sophie.hafner{at}donders.ru.nl tatjana.tchumatchenko{at}uni-bonn.de Erin M. Schuman 4 Max Planck Institute for Brain Research , Synaptic Plasticity, Frankfurt Am Main, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Erin M. Schuman Tatjana Tchumatchenko 1 Institute of Experimental Epileptology and Cognition Research, University of Bonn Medical centre , Venusberg-Campus 1, 53127 Bonn, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Tatjana Tchumatchenko For correspondence: nkrayn{at}uni-bonn.de anne-sophie.hafner{at}donders.ru.nl tatjana.tchumatchenko{at}uni-bonn.de Abstract Full Text Info/History Metrics Preview PDF Abstract AMPA-type glutamate receptors (AMPARs) underlie most of the excitatory synaptic transmission in the brain and are crucial for implementing long-term synaptic plasticity. AMPARs are multi-protein complexes composed of two types of subunits: pore-forming subunits GluA1-4 that assemble in the endoplasmic reticulum and form the glutamate-gated ion channel, and auxiliary subunits that modulate receptor bio-physical properties and mediate their forward trafficking to the plasma membrane. Here, using a combination of theoretical and experimental approaches, we elucidate the kinetics of essential trafficking steps and the protein sources necessary to explain the experimentally observed distribution of AMPARs and the response of different AMPAR subtypes to LTP induction. Our data indicate that the mRNA coding for one of the most abundant AMPAR auxiliary subunits, CNIH-2, is abundant in dendrites. Consistent with this mRNA distribution, CNIH-2 is locally synthesized. In contrast, the pore-forming subunits GluA1 and GluA2 are mostly synthesized in the cell body. CNIH-2 synthesis increases after the (chemical) induction of long-term potentiation. Strikingly, the translation of CNIH-2 is required for the plasma membrane insertion of GluA2-containing receptors and not GluA1-homomeric AMPARs. Using the selective trafficking of GluA2-containing AMPARs by CNIH-2, our computational model can explain the distinct temporal profiles in response to plasticity induction of two major subtypes of AMPARs, the slow-response of the calciumimpermeable (GluA2-containing) and fast kinetics of the calcium-permeable (GluA2-lacking) AMPARs. Introduction α -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid-type glutamate receptors (AMPARs) mediate most of the excitatory synaptic communication in the brain and play a central role in the expression of long-term potentiation (LTP). Controlling the number and activity of AMPARs in the postsynaptic density (PSD) effectively regulates the strength of excitatory synaptic transmission [ 1 ]. Many neurons in the brain, including pyramidal neurons in the cortex and hippocampus, have dendritic arbors that extend several hundreds of micrometers, allowing them to receive inputs from thousands of other neurons. These dendrites typically contain dendritic spines at a density of 1-2 spines per micron [ 2 ]. To detect excitatory inputs (i.e., glutamate release from the axons of presynaptic neurons), the postsynaptic neurons must ensure that their dendritic spines are populated with AMPARs. Which mechanisms do neurons employ to deliver AMPARs to distal sites and maintain their availability throughout the dendritic arbors? Previous experimental studies have shown that having abundant dendritic mRNA is a powerful strategy neurons employ to maintain a dendritic protein pool [ 3 , 4 ]. Later theoretical work on plasticity-related protein Calcium-Calmodulin Kinase 2 alpha (CamKIIa) showed that local dendritic protein synthesis significantly contributes to the homogenous distribution and high concentration of the protein throughout the dendritic arbors [ 5 ]. For the reasons outlined below, however, we still lack an understanding of the spatial and temporal mechanisms affecting AMPAR availability at synapses. AMPARs are macromolecular complexes comprising a tetrameric core of GluA1-4 poreforming subunits [ 6 – 8 ] as well as auxiliary subunits that are divided in three main families: the transmembrane AMPAR regulatory proteins (TARPs,γ-2,γ-3,γ-4,γ-5,γ-7, and γ-8) [ 9 , 10 ], the cornichon homologs (CNIH-2 and CNIH-3) [ 11 ], GSG1L [ 12 ] and the CKAMPs (CKAMP44 and CKAMP44-like proteins) [ 13 , 14 ]. There are a variety of long-range trafficking possibilities for AMPARs, such as cytoplasmic or membrane diffusion and active transport, as well as local dynamics, including endo- and exocytosis and trafficking between the PSD and an adjacent part of the dendrite. These processes have different time scales and can be affected by the receptor subunit composition and the history of the preceding events at different dendritic locations. Additionally, the specific subunit composition can influence the functional properties of AMPARs, such as their conductance, gating kinetics, and localization density. How do long-range (i.e., a full dendrite) and local (i.e., a spine attached to a piece of dendrite) dynamics interfere or interact with one another? How does the specific subunit composition of AMPARs affect their functional and trafficking properties? A major divergence in functional properties of AMPARs arises from the presence of the GluA2 subunit. GluA2-containing AMPARs are impermeable to Ca 2+ [ 15 ] and are constitutively recruited during plasticity [ 16 ]. Conversely, GluA2-lacking AMPARs are Ca 2+ permeable and can exhibit activity-regulated synaptic recruitment [ 1 , 17 ]. Recent work [ 18 ] has shown that individual subunits of AMPARs have higher mobility when not assembled into functional homo/heterotetramers. However, since a majority of AMPARs in the neocortex and hippocampus are heterotetramers of Glua1-2 subunits [ 1 ] this study will concentrate on these two types of AMPARs. For clarity, from now on, we refer to GluA1 homomers as GluA1-homomeric AMPARs and GluA1/2 and GluA2/3 heterotetramers as GluA2-containing AMPARs. Several studies have also suggested that synaptic plasticity demands are rapidly met by fast exocytosis of calcium-permeable GluA1 homotetramers [ 17 ], which are later replaced by calcium-impermeable, GluA2-containing AMPARs. Interestingly, recent studies show that the incorporation of calcium-permeable AMPARs is crucial for the full expression of LTP [ 19 ]. In contrast, following chemically induced long-term potentiation (LTP), the exo-cytosis rate of GluA2-containing AMPARs at stimulated spines and the nearby dendrite increases, resulting in a higher surface concentration of those receptors. Specifically, following LTP induction, the concentration of GluA2-containing receptors slowly increases within the next 1-2 hours, saturates, and remains elevated for several hours [ 1 ]. Additionally, GluA1-homomeric AMPARs typically exhibit higher single-channel conductance compared to GluA2-containing receptors [ 20 ]. This difference in conductance can influence the strength and amplitude of synaptic responses mediated by these receptors. Also, because the GluA1 subunit possesses the majority of the phosphorylation sites, GluA2-containing receptors may be less dynamically regulated by phosphorylation than GluA1-homomeric receptors [ 21 ]. At the cell surface, GluA1-homomeric AMPARs show a tighter synaptic localization compared to GluA2-containing ones that appear more diffused and constitute a reserve pool of extrasynaptic AMPARs [ 22 ]. In most neurons, different types of AMPAR auxiliary subunits co-exist. Such heterogeneity impacts the functional properties of AMPARs, such as their trafficking kinetics, stability in the plasma membrane and the PSD, and channel conductance [ 23 ]. All families of auxiliary proteins promote AMPAR surface expression, but each family modulates gating and pore properties differently. For example, CNIH-2 has been shown to slow receptor deactivation significantly - by up to fivefold - thereby extending the excitatory postsynaptic current duration, regardless of the GluA subunit combination [ 24 ]. As a result, synapses expressing CNIH-2-containing AMPARs exhibit a larger net depolarizing ion influx into the postsynaptic neuron. However, the precise role of CNIH-2 is still debated, with some seemingly conflicting findings. Initially, CNIH-2 was believed to assist AMPAR transport between the ER and Golgi apparatus (GA) in hippocampal CA1 neurons [ 25 ]. However, some studies suggest that CNIH-2 can bypass this early anterograde pathway and directly transport AMPARs to the plasma membrane [ 26 ]. Auxiliary subunit interactions can be dependent on the pore-forming subunit composition of AMPARs. For instance, CNIH-2 shows a stronger preference for GluA1/2 heteromers over GluA2/3 receptors [ 27 ]. Nonetheless, there are conflicting findings on whether CNIH-2 binds preferentially to GluA2-lacking or GluA2-containing AM-PARs [ 28 , 29 ]. Hence, it is still unclear how auxiliary subunits affect AMPAR availability at synapses and response to LTP. Here, we have developed a mathematical model to understand how diverse transport mechanisms and their biologically plausible temporal changes impact the availability of receptors at different dendritic locations. At the start, we used experimental data of fluorescently labeled endogenous GluA2-containing AMPARs to get a well-characterized baseline model for long-range and local AMPAR trafficking. Using a series of assays that preserve AMPAR naive stoichiometry and our model, we analyze which trafficking mechanisms are most critical for the availability of AMPARs at distal dendritic sites and how their exocytosis-to-endocytosis ratio is regulated across different dendritic locations. Finally, we examine the mechanisms that can explain a rapid decrease of GluA1-homomeric receptors and a sustained increase in GluA2-containing AMPAR concentration following cLTP induction. Our model suggests that a prolonged increase in GluA2-containing AMPAR exocytosis can account for the differences in the concentration dynamics of GluA1-homomers and GluA2-containing AMPARs. Interestingly, we show that the auxiliary subunit CNIH-2 is abundantly synthesized in dendrites, and chemical-LTP induction increases its synthesis. Moreover, interfering with CNIH-2 synthesis strikingly reduces GluA2-containing AMPAR at the cell surface but not that of the GluA1-homomers. This observation suggests that CNIH-2—an auxiliary subunit that specifically promotes GluA2 (but not GluA1) exocytosis—may be a critical factor underlying the differences in GluA1 and GluA2 temporal concentration profiles following cLTP induction. Taken together, our experimental and modeling results reveal a novel role the CNIH-2 local synthesis plays in diversifying glutamate receptor complexes in dendrites. Results Active transport of AMPARs can explain their homogeneous distribution along the dendrites Diffusion and active transport are known trafficking mechanisms that shape AMPAR distribution along the neuronal dendrites. However, the local transport of AMPAR subunits is not very well understood. To understand the role of local synthesis in AMPAR availability in dendrites, we first investigated the distribution of the mRNAs coding for AMPAR subunits. We analyzed fluorescent in-situ hybridization data to detect Gria1 and Gria2 mRNA-coding, respectively, for GluA1 and GluA2 AMPAR subunits ( Fig 1 and Fig S1 ). First, we quantified the endogenous levels of Gria1 and Gria2 mRNA molecules in hippocampal cultured neurons using fluorescent in-situ hybridization ( Fig 1 and Fig S1 ). We calculated the fractions of Gria1 and Gria2 localized in the cell body and in dendrites and compared them to CamK2a mRNAs [ 3 , 5 ] detected in the same neuron. While a majority of CamK2a mRNA was detected in dendrites (over 50% in both experimental datasets), both Gria1 and 2 were localized primarily in neuronal somata (74% and 82%, respectively), with a smaller dendritic mRNA fraction of 26% and 18%, respectively ( Fig 1A-B and Fig S1A-C ). Our previous work shows that although CamK2a mRNA distribution declines towards the distal tips of dendrites, it still provides a sufficient pool for local synthesis [ 5 ]. To understand if the Gria1 and 2 mRNA distribution behave similarly, we fitted the mRNA dendritic distributions with the exponential functions ( Fig 1C and Fig S1C , Methods) and compared their relative abundance at 100 µm dendritic distance (1C and Fig S1C , inset). At a dendritic distance of 100 µm , the distribution of CamK2a mRNA molecules showed a moderate decrease of 42%, while the Gria1 and GriA2 mRNA concentrations dropped by 96% and 98%, respectively( Fig 1C , inset). The rapid decline and relatively low levels of both Gria1 and Gria2 mRNA with distance to soma challenge the possibility that local synthesis is at the core of maintaining AMPA receptors availability at distal sites as it does for CamK2a . We note that previous studies reported the local synthesis of AMPAR subunit GluA1 upon BDNF stimulation [ 30 ] and secreted Amyloid precursor protein-alpha mediated LTP [ 31 ] in proximal dendrites up to a few tens of microns. These results are consistent with the presence of Gria1 and Gria2 in dendrites in the vicinity of the soma and do not contradict the lack of local GluA1 and GluA2 synthesis in more distal locations observed in our analyses. Download figure Open in new tab Figure 1: Active transport of AMPAR GluA2 subunits compensates for the lack of their local translation. (A) Top : Cultured rat hippocampal neurons at 18-21 DIV labeled using fluorescence in situ hybridization to detect Gria2 (left, green) and CamK2a mRNAs (right, green) and immunostained for MAP2 (magenta). Scale bar = 20 µ m. Bottom : Zooming into a representative dendrite shows a smaller number of Gria2 mRNA (left) compared to CamK2a mRNA (right). Scale bar = 5 µ m. (B) Somatic (hollow) and dendritic (filled) fraction of total mRNA for Gria2 computed as a fraction in a compartment with respect to the total mRNA count in a neuron (18 cells, p-value: 6.9*10 −5 ) and CamK2a (23 cells, p-value: 0.23), Gria2 -dendrite vs CamK2a -dendrite (p-value:0.02). (C) Exponential fit of mRNA puncta density distribution for Gria2 ( n = 12 dendrites, exponent Gria2 = -0.04 ± 0.009) compared to the CamK2a distribution ( n = 33 dendrites, exponent CamK2a = -0.006 ± 0.001), see Methods for mRNA exponential fit. Inset : Normalized fitted Gria2 mRNA distribution compared to CamK2a . (D) Top : Cultured rat hippocampal neurons at DIV 18-21 processed for antibody labeling of GluA2 (white) and FI MAP2 (magenta). Bottom : Zoom into a representative dendrite shows a homogeneous GluA2 protein distribution along the first 100 µ m. Scale bar = 20 µ m. (E) MAP2 normalized GluA2 intensity (5 µ m bins, median: green squares with IQR). (F) - (H) GluA2 distributions based on molecular dynamics model (Methods). (F), (G) Active transport V P = 0. (F) The half-life ( T 1/2 ) of GluA2 protein increased over the reported maximum of 4.35 days cannot provide constant GluA2 distribution. (G) Increasing the reported diffusion constant ( D P ) of GluA2 by a factor of 10 leads to a 9% GluA2 decrease at 100 µ m but does not support the increase at 500 µ m (inset). (H) Active transport of V P = 1.4 * 10 −3 µ m/s leads to an observed 7% decrease in GluA2 protein concentration and increases concentration at 500 µ m (inset). We investigated whether the density of GluA2-containing receptors in cultured hippocampal neurons (21 days in vitro) aligned with previously reported trends. Using protein antibody labeling, we visualized total GluA2-containing AMPARs after fixation and permeabilization ( Fig 1D ). Receptor density was analyzed in the first 100 µm of primary dendrites by measuring fluorescence intensity over 5 µm segments. Our data did not show any major trend across the initial 100 µm of dendritic length. To explore the molecular mechanisms underlying the observed AMPAR distribution, we developed a computational model incorporating intracellular vesicle velocity ( V P ), diffusion rate ( D P ), and half-life ( λ P ) ( Eq 4 ), using the biologically plausible parameter ranges ( Table 1 ). Fitting the experimentally observed GluA2 distribution ( Fig 1E ) to the model ( Eq 4 ) (Methods) revealed a marginal 7% decrease at 100 µm distance from the soma ( Fig 1E ). Previous experimental studies examining AMPAR distribution at a longer distance (100 − 250 µm ) [ 32 ] reported up to a two-fold increase in AMPAR-mediated currents along CA1 apical dendrites and attributed them to a proportional increase in AMPAR concentration [ 32 ]. This phenomenon, known as distance-dependent scaling, suggests an accumulation of AMPARs toward the distal dendritic location. A subsequent study [ 33 ] confirmed distance-dependent scaling specifically in AMPARs containing the GluA2 subunit. Building on these findings, we next examined how variations in half-life ( T 1/2) , diffusion ( D s ), and active transport ( V p ) within biologically reported ranges ( Table 1 ) might shape the AMPARs distribution. Specifically, we investigated parameter combinations that could reconcile our experimentally observed near-uniform AMPAR distribution close to the soma with the increasing AMPAR concentration trends reported in previous studies [ 32 , 33 ]. View this table: View inline View popup Download powerpoint Table 1: Experimentally reported parameters measured in dendrites of hippocampal cultured neurons; a) includes values that are comparable for both GluA1/2 subunits or if the value is reported only for one of the subunits. View this table: View inline View popup Download powerpoint Table 2: Parameter values used for steady-state, baseline model of GluA1-homomeric AMPARs. Previous modeling studies have shown that the active transport velocity strongly affects the distribution of molecules along the dendrite [ 5 , 34 ]. AMPARs are composed of transmembrane proteins; thus, they are transported in vesicles in the cytoplasm. AMPAR-containing vesicles are actively moved inside the cytoplasm using microtubule-mediated transport. Those vesicles containing AMPARs move bidirectionally at a speed in the range of 1.46 µm / s in both anterograde and retrograde directions [ 35 ]. However, experimental studies do not report a net anterograde active transport velocity range for non-modified receptors. Therefore, we first considered the active transport velocity an unknown parameter of our mathematical model (Supplemental Information). In the next two steps, we assumed that the net active transport was equal to zero ( V P = 0) (anterograde and retrograde transport fully compensate for one another) and analyzed whether the half-life and diffusion parameter variations can explain the experimentally observed distribution of GluA2 receptors, which was approximately constant near the soma and should increase to-wards the dendritic end. First, we analyzed how different values of the protein half-life parameter ( T 1/2 ) impact the shape of the GluA2 distribution. Since the range of experimentally reported diffusion parameters spans two orders of magnitude, varying from 0.01 to 1 µm 2 / s ( Table 1 ), we first set an intermediate value of D P = 0.1 µm 2 / s , from [ 36 ] in our simulations. We then assumed that the half-life of the GluA2-containing AMPARs was in the range from 1.95 to 4.35 days, as reported in the previous experimental studies [ 37 , 38 ]. When we set the half-life to 1.95 days, GluA2 concentration dropped by 47% at 100 µm dendritic distance. This drop was significantly reduced to 39% when setting the GluA2 half-life to its maximum reported value of 3.12 days. The larger half-life values lead to a longer lifetime of the corresponding molecules, allowing a larger number of them to reach further dendritic locations via diffusion. Nonetheless, a decrease of 39% in GluA2 concentration at 100 µm is much higher than our experimentally observed decrease of 7%. Further increasing the GluA2 half-life to 4.35 days (half-life reported for GluA1, Table 1 ) still resulted in a 33% drop at 100 µm dendritic distance ( Fig 1F , Fig S2A ) which further decreased towards the end of model dendritic length ( Fig 1F , inset, Fig S2A , inset). While increasing the GluA2 half-life did raise its concentration in dendrites, it could not reproduce the experimentally measured GluA2 distribution. We next set GluA2 half-life to the maximally reported value of 3.12 days [ 37 ] and explored how modulating the receptor diffusion in cellular membranes could shape its distribution. Interestingly, while diffusion is a rather fast parameter (see Table 1 ), because it is bidirectional, it is relatively slow at moving receptors through the long distances of dendritic arbors [ 39 ]. In addition, at the cell surface, dendritic spines further slow down receptor trafficking [ 40 ]. Indeed, AMPARs can diffuse in the plasma membrane at a rate up to 1 µm 2 / s [ 41 , 42 ]. However, the diffusion drops to 10 −2 µm 2 / s in synaptic compartments [ 43 ]. In addition, receptors are also highly mobile in intracellular organelle membranes and particularly in the endoplasmic reticulum that forms a continuous organelle throughout the neuronal dendritic arbor [ 44 ]. To understand the impact of diffusion on the distribution of GluA2, we analyzed the diffusion constants in the range from 10 −2 to 1 µm 2 / s starting from the experimentally reported diffusion constant of D P = 0.1 µm 2 / s [ 36 ]. For the experimentally reported diffusion constant of D P = 0.1, we observed a 39% GluA2 decrease at a 100 µm dendritic distance ( Fig 1G , Fig S2B ). We divided the diffusion constant by a factor of 10, as reported by [ 43 ], which led to a large drop of 79% in GluA2 concentration ( Fig 1G ). Increasing the diffusion constant by a factor of two led to a 28% concentration reduction. Multiplication of the diffusion constant D P by a factor of 10 led to a 9% decrease at 100 µm dendritic distance. This decrease is comparable with our experimentally observed reduction in GluA2 concentration ( Fig 1G ). However, this value corresponds to the upper bound of the experimentally reported value of diffusion rate in the dendritic shaft, and the presence of spines leads to a lower net diffusion rate [ 45 ]. In addition, similar to the effect observed for half-life ( T 1/2 ) values at a larger dendritic distance, different diffusion constants also resulted in a decrease towards the distal tip in our model ( Fig 1G , inset, Fig S2B , inset). However, these model predictions differ from previous experimental reports on AMPAR distribution at longer distances (100 − 350 µm ). Since increasing neither the GluA2 half-life nor diffusion coefficient was sufficient to reproduce distance-dependent scaling of AMPARs, we next considered the possibility that the active transport velocity of GluA2 along microtubules plays a role. We had previously set the active transport velocity to be zero as transport along microtubules in dendrites is known to be bidirectional [ 46 ]. In contrast, in the next step, we set the active transport velocity V P to a positive value, resulting in a biased anterograde movement of intracellular vesicles toward the dendritic tips. To test this idea, we set the receptor half-life to 3.12 days and fitted the active velocity constant V P and the diffusion constant D P to first match our experimentally measured GluA2 fluorescence distribution. The best fit active velocity value was V P = 1.4 × 10 −3 µm / s , and the diffusion constant was D P = 0.22 µm 2 / s . With these physiologically plausible values for our parameters, we obtained a distribution of GluA2-containing receptors that displayed a 7% drop at 100 µm dendritic distance ( Fig 1H , Fig S2C ). Strikingly, the best-fitted model also predicted a steady increase in receptor concentration that reached 180% at 500 µm dendritic distance ( Fig 1H , inset). To understand how a lower value of the active transport velocity might affect the global distribution of GluA2 in our model branch, we divided its value by a factor of 2 and 10. As the active transport velocity was lowered, we observed an accumulation of receptors near the somatic regions ( Fig 1H ). When dividing the active transport velocity by a factor of 10, we observed that the gradual increase of GluA2 concentration in distal dendrites was lost ( Fig 1H , inset). Interestingly, multiplying the fitted active transport parameter by a small factor of 3/2 led to a 500% larger GluA2 concentration towards a dendritic tip, showing that even minimal increases in the active transport velocity can lead to a large accumulation of GluA2 molecules at distal dendritic sites ( Fig 1H , inset and Fig S3 ). In conclusion, our experimental data show that Gria1 and Gria2 mRNAs, coding for the two most abundant AMPAR pore-forming subunits GluA1 and GluA2, are predominantly translated in the somatic compartment under basal conditions. This observation reduces the likelihood that local translation underlies the reported increase in GluA2 concentration at distal dendritic locations. In our simulations, only by modulating the active transport velocity of GluA2 intracellular vesicles could we recapitulate the experimentally observed distribution of GluA2 along dendrites. Since changes in AMPAR numbers close to and at synapses are a crucial step during the expression of synaptic plasticity, we then explored how exo- and endocytosis shaped AMPAR distribution in the plasma membrane and cytoplasm. Endogenous concentrations of AMPARs in the cytoplasm, cell membrane, and postsynaptic density The simulations presented in the previous section indicate that active transport is crucial for delivering AMPARs to distal dendritic sites. Since the number of AMPARs in the postsynaptic density (PSD) correlates with the synaptic strength, our next goal was to understand the mechanisms underlying AMPAR insertion in the PSD under basal and plasticity conditions. To this end, we analyzed AMPAR trafficking between the cytoplasm (i.e., intracellular compartment), the plasma membrane (i.e., neuronal surface), and the PSD. AMPAR exocytosis occurs near the synaptic site (spine head or local dendritic shaft), where the AMPARs can enter the PSD via lateral diffusion in the plasma membrane [ 47 ]. Thus, modeling the rate and location for endo- and exocytosis and synaptic trapping can provide a better understanding of the impact of these processes on synaptic plasticity. Notably, the trafficking speed and the residence time (lifetime in a compartment) of transmembrane proteins differ when they are in the plasma membrane or the cytoplasm [ 40 ]. Thus, modeling intracellular and surface proteins separately is crucial to estimate their redistribution accurately. Here, we extended the previous version of the model to examine AMPAR distribution in different dendritic compartments: the cytoplasm, the plasma membrane, and the PSD. With this model, we analyzed how endocytosis, exocytosis, and diffusional trapping at PSD shape the AMPAR distribution between the three compartments. We distinguish between mobile AMPARs in the cytoplasm ( P c ), mobile receptors in the plasma membrane in extrasynaptic regions ( P s ), and the immobile receptors at PSDs ( P psd ) In this model, the intracellular diffusion ( D c ), motor-based transport ( V p ), and receptor degradation rate ( λ c ) shape the distribution in the cytoplasm along the dendrites. Furthermore, we assumed that the AM-PARs in the plasma membrane move exclusively via lateral diffusion ( D s ) - Brownian movement. For local trafficking between surface and cytoplasm, we introduced coupling terms of exocytosis and endocytosis rates ( β and α , respectively). Synaptic uptake can impact the availability of the AMPARs at different dendritic locations. Surface receptors can get trapped by binding to anchoring proteins (e.g., PSD-95) in the PSD with an incorporation rate of η . Trapped receptors will eventually unbind from anchoring proteins and exit the PSD at the rate of γ . Finally, we assumed that each synapse has a maximum capacity ( ω ) for holding AMPARs, determined by the number of the so-called “slots” that are postsynaptic density structures consisting of proteins such as PSD-95 that anchor AMPARs in the PSD and restrict their movement [ 29 , 48 , 49 ]. For simplicity, we did not consider the nanoscale organization of these slots, which can influence the diffusional trapping of AMPARs inside the PSD (see [ 50 ]). To set accurate parameters for our model, we performed immunolabeling of endogenous surface and intracellular GluA2 ( Fig 2A ). We used secondary antibodies carrying different fluorophores before and after permeabilization. We then determined the trafficking parameters under basal conditions by fitting the distribution of the GluA2-containing AMPARs at the surface and in the cytoplasm ( Fig S4 ) to the steady state distribution Eqs. 9 - 11 of the model in Eqs. 1 - 3 . We first computed the exocytosis and endocytosis rates, which ensure continuous cycling of the GluA2-containing AMPARs between the plasma membrane and the cytoplasm. To determine the exocytosis rate, we used our theoretical prediction in Eq. 18 showing that the ratio between the exo- and endocytosis rates ( β / α ) is given by the ratio between the total surface to intracellular GluA2 fluorescence intensity (sGluA2/iGluA2). Under basal conditions, the levels of cytoplasmic GluA2 were higher than on the surface GluA2 in both somata and dendrites ( Fig 2B ), as previously observed [ 51 , 52 ]. Also, the fit of this ratio remained constant along the dendrite (up to 100 µm ; Fig 2C . Using the measurements of fluorescence intensity in the cytoplasm and at the surface ( Fig S4 ) and the formula in Eq. 18 , we computed the exocytosis rate to be β = 2.4 × 10 −4 s −1 , which was smaller than the endocytosis rate α = 6 × 10 −4 s −1 that we adopted from experiments reported by Rosendale and colleagues [ 53 ]. Download figure Open in new tab Figure 2: The endogenous distribution of surface and intracellular GluA2-containing AMPARs. (A) Schematic of antibody labeling (left, top) with example images which were used to determine intracellular and surface protein concentration in dendrites (top, right) and somata (bottom). The GFP signal (green) represents the intracellular area. iGluA2 (cyan) represents GluA2 subunits in the cell cytoplasm, while sGluA2 (magenta) corresponds to GluA2 subunits confined to the cell surface (scale bars 10 µ m). (B) Fluorescent intensity ratio between surface and intracellular GluA2 was equal to 0.37 ± 0.062 in somata (mean ± std, N = 15) and 0.40 ± 0.063 in dendrites (mean ± std, N = 69, p = 0.35, z -stat=0.86). (C) Dendritic distribution of surface to intracellular GluA2 fluorescent intensity ratio fitted with an exponential function shows an insignificant decrease of 3% at 100 µ m (exponent 2.9 * 10 −4 ). (D) The schematic of the surface (top, left, magenta) and intracellular areas (top, right, cyan) of a synapse. Example images (bottom) show dendritic segments with spine and corresponding shaft ROIs (scale bar 1 µ m). (E) The balance between the exocytosis and endocytosis rates in spine (0.56 ± 0.12) and dendritic shaft (0.5 ± 0.09) ( N = 55, p = 0.0034, z -stat=0.83); (F) The surface (top, left) and intracellular (top, right) synaptic enrichment was computed as the corresponding spine area divided by the shaft area (magenta for the surface and cyan for the cytoplasm area, respectively SE surf = 1.01 ± 0.84, SE int = 0.88 ± 0.68) ( N = 55, p = 0.45, z -stat = 0.55). We note that the exact sites of AMPAR exo- and endocytosis will likely impact the speed of the AMPAR insertion and removal from the PSD, thus affecting the AMPAR responses to plasticity cues. However, the precise localization of those sites (i.e., near the PSD or on the dendritic shaft) remains an active source of debate. Exocytosis of AMPARs has been reported to occur primarily at peri-synaptic sites [ 54 ], on the dendritic shaft [ 55 ], or both [ 56 ]. Importantly, the closer the exocytosis sites to the PSD, the shorter the distance AMPARs need to diffuse to reach the PSD and participate in synaptic transmission. Furthermore, spine necks act as a bottleneck that limits the surface diffusion of AMPARs [ 57 ] and, hence, slower response time. Similarly, the exact site of the AMPAR endocytosis is still a topic of active research, with recent evidence suggesting AMPAR internalization at the peri-synaptic endocytic zone [ 53 ]. To establish accurate parameters for our model, we conducted a detailed analysis of our experimental data to estimate the ratio between the exo- and endocytosis rates near the spines ( Fig 2D , bottom). We also evaluated whether this ratio varies along the dendrite ( Fig 2C ). To analyze the GluA2-containing AMPAR concentration near spines, we used fluorescence images. We manually created two rectangular ROIs, one to mark the spine head region and the second to constrain the dendritic shaft directly adjacent to the corresponding spine ( Fig 2D ). Having marked the corresponding areas, we asked if the surface-to-intracellular (or exocytosis-to-endocytosis, β / α ) ratio of GluA2-containing AMPARs in spine heads and dendritic shafts under the spines deviated from the average ratio we computed ( Fig 2B, C ). We found that the exocytosis-to-endocytosis ( β / α ) ratio was slightly increased in the vicinity of spines ( Fig 2E ) compared to the average ratio along the dendrite ( Fig 2B, C ). Specifically, we found that the mean exocytosis-to-endocytosis ratio was 0.5 in the dendritic shafts adjacent to the spine heads and 0.55 in the spine heads ( Fig 2E ). Notably, the exocytosis-to-endocytosis ratio near and inside spines exceeded the average ratio of 0.4 previously computed for dendrites ( Fig 2B ). Finally, we used the ratio observed in the dendrite (i.e. 0.4) in our model as it is a better representative value. AMPARs enter the PSD via lateral diffusion following exocytosis to the cell surface. Once AMPARs reach a PSD, they can be trapped owing to interactions with scaffolding proteins. Several intracellular signaling pathways, such as phosphorylation, are known to regulate their interaction with the scaffolding proteins. To leave the PSD, AMPARS unbind from the scaffolding slots and diffuse out of the PSD. To determine the net incorporation ( η ) and exit ( γ ) rates of GluA2-containing AMPARs, we once again analyzed our experimental data (2F). We calculated a factor we named “synaptic enrichment” as the ratio of integrated density between spine heads and extrasynaptic ROIs for receptors at the cell surface (2F, left) and in the cytoplasm (2F, right). The mean of both ratios was approximately equal to 1, suggesting that the fraction of synaptic and extrasynaptic surface receptors was roughly 50% each. We used this ratio to fit the incorporation and exit parameters in our model η = 7 × 10 −4 s −1 , γ = 2.3 × 10 −2 s −1 . Interestingly, we also observed a systematic increase in the synaptic enrichment from proximal to distal dendritic sites. This was due to decreased shaft fluorescence while the synaptic fluorescence remained constant (see supplementary Fig S5 ). To test if our method incorporating 2D areas occupied by the spine head and shaft provided accurate results, we re-computed synaptic enrichment using the 1D method described in [ 58 ]. In the 1D method, the fluorescence intensity was evaluated along the straight line passing through the spine head (most likely PSD region) and the dendritic shaft area. The corresponding graph of the fluorescence intensity contained two peaks - the first corresponding to the intensity in the spine head and the second in the dendritic shaft. The synaptic enrichment was calculated as the ratio of the peak values (see supplementary Fig S6 ). Using both methods, we arrived at a synaptic enrichment ratio of approximately 1 ( Fig 2F ). Finally, we fitted the steady state equations in Eqs. 13 to estimate global transport parameters (diffusion constants D s , D c , and net drift V p ) from the normalized GluA2 fluorescent intensity ( Fig S4 ). We obtained the best-fit values for diffusion coefficients D s = 0.7 µm 2 / s, D c = 0.7 µm 2 / s , and the active transport velocity of V p = 10 −4 µm / s . We note that these values were computed using a different dataset from the mouse hippocampal neurons, which can explain their difference from the results obtained for the experimental data in Fig 1 . Overall, using experimentally measured AMPAR distribution and extended molecular dynamics model, we characterized the local and global trafficking parameters of endogenous AMPARs. Strikingly, we found that the AMPAR endocytosis rate exceeds its exocytosis rate, and the exocytosis- to-endocytosis ratio is constant along the dendrite. The coexistence of two regimes “live longer - travel slower” in the neuron plasma membrane, especially near synapses, and “live shorter - travel faster” in the cytoplasm - seem to balance each other such that it enables GluA2-containing AMPAR availability at distal dendritic sites and their stability near synapses. Next, we used the molecular dynamics model to study temporal changes in AMPAR concentration due to the regulation of local trafficking upon plasticity induction. GluA1-homomeric and GluA2-containing AMPAR elicit distinct temporal response upon LTP Modeling AMPAR temporal dynamics is crucial for understanding the time scales of synaptic plasticity. One of the challenges in understanding AMPAR temporal dynamics is that the onset and duration of LTP-related processes depend on the specific subunit composition of AMPARs involved. So far, we have focused our experimental investigations mainly on GluA2-containing AMPARs. However, recent studies report that the GluA2-lacking or calcium-permeable AMPARs are essential for LTP [ 59 , 60 ]. Calcium permeable AMPARs get trapped rapidly during a short and transient window after the start of LTP induction - we note that the most abundant calcium-permeable AMPAR pore-forming complex is a GluA1 homotetramer [ 1 ]. To predict and compare the temporal response of the GluA2-containing AMPARs and GluA1 homomers, we compared our model predictions with experimental studies. We used data from two independent studies to isolate the unique behavior of GluA1 homomers and GluA2-containing receptors. First, we used our model to simulate the experimental data from Tanaka and Hirano [ 61 ] in which the over-expression of GluA1 promoted the formation of GluA1-homomeric AMPARs. During cLTP induction, these receptors increased in the plasma membrane and, to a greater extent, in the PSD, peaking at the end of the cLTP stimulation using glycine. Following the chemical stimulation, the concentration of GluA1-homomeric receptors gradually decreased. Incorporating the cLTP protocol from [ 61 ] in our model, we tested the model’s ability to recapitulate the temporal response of the GluA1 homomers upon cLTP induction (see Fig 3A and Materials and Methods). According to Tanaka and Hirano, the GluA1 exocytosis frequency increases immediately after cLTP induction but returns to basal level when glycine is washed out ( Fig 3A , S7 ) [ 61 ]. Importantly, changes in GluA1 exocytosis rate are consistent with other independent studies (e.g., [ 56 ]), indicating that the homomeric GluA1 AMPAR exocytosis rate increases immediately upon LTP induction and lasts only during the stimulation phase. We simulated an instantaneous 3.5-fold increase of the GluA1 exocytosis rate ( β ) homogeneously throughout the dendrite during the 10 minutes of cLTP induction ( Fig. S7 ). Our simulation showed that a 3.5-fold increase in exocytosis rate for 10 minutes leads to an increase in GluA1 concentration in the plasma membrane and in PSDs ( Fig. S8A ). However, when modifying the exocytosis rate alone, the model predicted a transiently higher concentration of GluA1 in the dendrite compared to the PSDs, which contradicted experimental observations of Tanaka and Hirano study [ 61 ]. To address this, we adjusted the model to promote GluA1 incorporation in PSDs by increasing both the exocytosis rate and the PSD incorporation rate ( η ) upon cLTP induction. The prolonged increase in incorporation rate reflects increased receptor binding to scaffold proteins, associated with phosphorylation of GluA1 and stargazin C-terminal domains and increased synapse size [ 1 , 62 , 63 ]. We simulated an instantaneous 1.3-fold increase in the incorporation rate together with a 3.5-fold transient increase in the exocytosis rate ( Fig 3A ). Using these parameters, our model was able to replicate the experimentally observed magnitude and temporal profile of GluA1 homomeric AMPARs ( Fig 3A and C ). Notably, neither exocytosis nor the incorporation rate up-regulation alone could reproduce the experimentally observed temporal profile of GluA1 surface concentration ( Fig. S8A and B ). Interestingly, our model predicted that 30 minutes following cLTP induction, the concentration of GluA1 in the plasma membrane in extrasynaptic areas dropped back to the baseline level and decreased below the baseline after ∼ one hour while the synaptic GluA1 concentration reached an equilibrium concentration of 15% above the baseline ( Fig 3A ). Download figure Open in new tab Figure 3: A constant increase in exocytosis rate upon LTP can explain temporal dynamics of GluA1-homomeric AMPARs but not GluA2-containing AMPAR. (A) cLTP induction led to 3.5-fold increase in exocytosis rate ( β ) ([ 61 ]) and 30% increase in synaptic hopping-in rate ( η ) of GluA1-homomeric AMPARs. The mathematical model matches the amplitude and subsequent decrease in response of GluA1-homomeric AMPARs reported in [ 61 ]. (B) Concentration of GluA2-containing AMPARs increases within two hours following cLTP induction [ 63 ]. Long lasting increase of constant magnitude in the exocytosis rate cannot explain experimental data. (C) Pairwise relation between parameter change upon LTP induction and model performance against data, (see Fig S8 and S9 ). GluA1-homomeric AMPA data is from [ 61 ], GluA2-containing AMPAR data is from [ 63 ]. Second, changes in synaptic efficacy induced by cLTP are known to last for several hours. As described above, these long-lasting changes have been attributed to an increase in the concentration in GluA2-containing AMPARs that, according to a recent report, can last for at least two hours [ 63 ]. Clavet-Fournier and colleagues used super-resolution microscopy and antibodies against endogenous GluA2 to monitor the concentration of GluA2 in PSDs after cLTP induction. Their experiments showed increased synaptic GluA2-containing receptors that persisted beyond the stimulation period and reached roughly a 150% increase two hours after cLTP induction. To replicate these findings, we incorporated the experimental data for GluA2-containing receptors into our model [ 63 ] ( Fig 3B ). However, as we observed in the case of GluA1, increasing the exocytosis and PSD incorporation rates enhanced receptor concentration in PSDs during stimulation but was insufficient to sustain the increased concentration beyond the stimulation period ( Fig 3B , Fig S9 ). In fact, our model failed to match both the kinetics and large concentration increase of GluA2-containing receptors ( Fig 3B and C ). This suggests that additional mechanisms lasting beyond the stimulation phase were required for the prolonged GluA2 accumulation in PSDs. As an initial attempt to adjust our model, we investigated if increasing the exocytosis rate beyond the stimulation phase can lead to a long-lasting increase in GluA2-containing AMPAR concentration in PSD ( Fig S9 ). Our model predicted that the increase in the receptor concentration lasted roughly as long as the increase in the exocytosis rate. Moreover, a direct change in exocytosis rate showed faster kinetics of GluA2-AMPAR change compared to the data. This suggests that in the case of GluA2-containing receptors, the increase in exocytosis rate is, on the one hand, limited or slow and, on the other hand, lasting for several hours despite the arrest of the stimulation. Strikingly, delayed, and long-lasting responses are typically expected when protein synthesis is recruited upon stimulation [ 64 ]. However, we showed previously that GluA2 mRNA is not present at high levels in dendrites ( Fig 1 ). Auxiliary subunit CNIH-2 is locally translated and preferentially mediates GluA2-containing receptor exocytosis We shifted our focus to other components of the AMPAR macromolecular complexes, namely, auxiliary subunits. Leveraging the translatome dataset published by Glock and colleagues [ 4 ], we confirmed the low levels of AMPAR pore-forming subunits GluA1-2 mRNAs in dendrites but identified one of the AM-PAR auxiliary subunits Cornichon-2 (Cnih2) as an abundant dendritic mRNA. To validate this finding, we analyzed Cnih2 mRNA localization in cultured rat hippocampal neurons using in situ hybridization and compared it to CamK2a mRNA. Remarkably, the distribution of Cnih2 mRNA closely resembled that of CamK2a ( Fig 4A ). Approximately 50% of Cnih2 mRNA was localized in the dendrites, similar to the 60% observed for CamK2a mRNA ( Fig 4B ). Fitting Cnih2 and CamK2a mRNA distribution with exponential decay functions ( Fig 4C , Materials and Methods) revealed comparable reductions at 100 µm from the soma: 59% for Cnih2 and 52% for CamK2a . Next, we investigated whether CNIH-2 protein was indeed synthesized locally. To assess de novo CHIH-2 protein synthesis, we performed fluorescence non-canonical amino acid tagging coupled with proximity ligation assay (FUNCAT-PLA) [ 65 ]. FUNCAT-PLA couples the incorporation of a non-canonical amino acid here, AHA, a methionine derivative, during protein synthesis with an antibody-based detection of a protein-of-interest; the coincidence of the signals is detected with proximity ligation assay. Neurons were incubated for 1 hour with AHA (4 mM) or Methionine as a control and then fixed for further processing (see Materials and Methods). FUNCAT-PLA revealed widespread CNIH-2 synthesis in somata and throughout dendrites, with approximately 70% of the nascent protein localized to dendrites ( Fig S10A-D ). In the Methionine control condition, the FUNCAT-PLA signal was essentially absent ( Fig S10A-C ). Together, these results indicate that CNIH-2 is locally synthesized in the dendrites. Download figure Open in new tab Figure 4: Local translation of Cornichon 2 regulates the trafficking of GluA2-containing AMPA receptors to the neuronal surface (A) Top : Cultured rat hippocampal neurons (18-21 DIV) processed for FISH against Cnih2 (left, green) and CamK2a mRNA (right, green) and fluorescently immunostained (FI) MAP2 (magenta). Scale bar = 20 µ m. Bottom : Zooming into a representative dendrite shows a smaller number of Cnih2 mRNA (left) compared to CamK2a mRNA (right). Scale bar = 5 µ m. (B) Somatic (hollow) and dendritic (filled) fraction of total mRNA for Cnih2 (27 cells, p-value: 1.0) and CamK2a (19 cells, p-value: 6e-4), Cnih2 -dendrite vs CamK2a -dendrite (p-value:0.08). (C) Exponential fit of mRNA puncta density distribution for Cnih2 ( n = 36 dendrites, exponent Cnih2 = -0.009 ± 0.002) compared to the CamK2a distribution ( n = 38 dendrites, exponent CamK2a = -0.007 ± 0.002), see Methods for mRNA exponential fit. Inset : Normalized fitted Cnih2 mRNA distribution compared to CamK2a . (D) left column: Representative images of hippocampal cultured neurons showing the distribution of the newly synthesized CNIH-2 after 1 hr of metabolic labeling with AHA and an increase of CNIH-2 new proteins in dendrites after cLTP induction ( right column: ). Scale bar 20 µ m in neuron images, 5 µ m in dendrite images. (E) Box plots showing the density of newly synthesized CNIH-2 in the soma. (F) Same as (E) but in dendrites ( n = 30 for basal, n = 22 for cLTP neurons). (G) Cultured hippocampal neurons showing newly synthesized surface GluA1 and GluA2 subunits in neurons expressing an shRNA against CNIH-2 or a control shRNA. Without CNIH-2 proteins, newly synthesized GluA2-containing AMPARs are absent at the somatic and dendritic surface. Scale bar 20 µ m. (H) Box plots show the abundance of newly synthesized GluA1 (top) and (I) GluA2 (bottom) containing receptors at the cell surface in soma and dendrite (GluA1 Cntl n = 39, GluA1 without CNIH-2 n = 28; GluA2 Cntl n = 29, GluA2 without CNIH-2 n = 25 neurons analyzed per conditions). * p ≤ 0.05 *** p ≤ 0.001. Circles represent individual neuron values, box lines for median, mean shown as black triangles. To examine whether CNIH-2 contributes to the activity-dependent trafficking of AMPARs to the neuronal plasma membrane, we applied a cLTP protocol (see Materials and Methods) and tracked newly synthesized CNIH-2 proteins using FUNCAT-PLA ( Fig 4D ). cLTP induction resulted in a significant, approximately twofold increase in the dendritic CNIH-2 protein levels within an hour ( Fig 4F ), while no significant changes were observed in neuronal somata ( Fig 4E ). This experiment shows the direct regulation of an AMPAR auxiliary subunit protein level by activity within the dendrites. We next asked how dendritic CNIH-2 synthesis regulates AMPAR trafficking to the neuronal surface. To this end, we optimized the FUNCAT-PLA pipeline [ 65 ] to label nascent GluA1 and GluA2-containing receptors localized in the plasma membrane (see Materials and Methods). To investigate whether CNIH-2 synthesis promotes the trafficking of AMPAR to the cell surface, we used a shRNA strategy to reduce Cnih2 transcript levels. We transfected hippocampal cultured neurons with CNIH-2 shRNA or a control shRNA at 8 days in vitro (DIV). After 11 days of expression of the shRNA, neurons were incubated for 2 hours with AHA (4 mM or methionine as a control), followed by surface labeling with GluA1 or GluA2 antibodies prior to fixation. The efficiency of CNIH-2 expression knockdown after 11 days of expression was assessed by both in situ hybridization against Cnih2 mRNA and FUNCAT-PLA to detect nascent CNIH-2 protein ( Fig S11 ). Reduced CNIH-2 expression had no effect on the surface expression of the GluA1 subunit ( Fig 4G and H ) but dramatically reduced the surface expression of GluA2 subunits in both the cell soma and the dendritic arbor ( Fig 4G and I ). These findings suggest that CNIH-2 is critical in driving the surface expression of GluA2-containing AMPARs. CNIH-2 local synthesis-driven exocytosis can explain the slow and persistent response of GluA2-containing AMPARs upon LTP Building on the results from the previous section, we then incorporated CNIH-2 in our model as a regula-tor, under the control of local protein synthesis, of GluA2-containing AMPAR exocytosis rate. Using our previously developed model [ 5 ], we examined steady-state distributions and concentrations of CNIH-2 mRNA and protein under basal and plasticity conditions (see also Supplemental Information). In the model, a prolonged 2-hour up-regulation of CNIH-2 local synthesis led to an approximately 10-fold increase in CNIH-2 protein concentration within dendrites ( Fig 5A ). While the increase in CNIH-2 concentration was only observed during the up-regulation of its local protein synthesis, a 5-fold elevated CNIH-2 concentration was still present 5 hours after the start of the simulation ( Fig 5A ). In contrast, enhanced CNIH-2 synthesis confined to the somatic region led to a marginal increase in CNIH-2 dendritic concentration of roughly 1.15-fold in proximal dendrite after 2 hours of up-regulation but, importantly, this increase was reduced to 1.05-fold after 5 hours ( Fig 5B ), left). Finally, a brief 10-minute up-regulation of CNIH-2 translation led to a moderate 2-fold increase in CNIH-2 concentration after 2 hours in the proximal dendritic region. This increase dropped to 1.5-fold after 5 hours ( Fig 5C ). Download figure Open in new tab Figure 5: Sustained local synthesis-driven exocytosis can explain the time course of cLTP-induced changes in GluA2-containing AMPARs. (A) Distribution of CNIH-2 proteins following brief up-regulation (for 10 mins) of local synthesis rate ( β p ). (B) : The time and space-dependent increase in CNIH-2 concentration was used to model an increase in the GluA2 exocytosis rate. The resulting GluA2-containing AM-PAR distribution matched the temporal response kinetics but not the amplitude reported in [ 63 ]. (C) Distribution of CNIH-2 proteins following prolonged up-regulation (for 2 hrs) of local synthesis rate ( β p ). (D) : The time and space-dependent increase in CNIH-2 concentration was used to model an increase in the exocytosis rate. The resulting GluA2-containing AMPAR distribution matches the amplitude and temporal response reported in [ 63 ]. (E) The change in CNIH-2 concentration following prolonged up-regulation (for 2 hrs) of somatic flux ( J pin ). (F) The mathematical model predicted only a small increase in surface and synaptic GluA2-containing AMPAR concentration as a response to somatic synthesis up-regulation. (G) Pairwise relation between CNIH-2 trafficking kinetics upon chemical LTP induction and model performance against data from [ 63 ]. Next, we incorporated time-dependent changes in CNIH-2 distribution in our mathematical model of AM-PAR trafficking. Specifically, we simulated an increase in the exocytosis rate of GluA2-containing AM-PARs driven by the CNIH-2 local synthesis combined with a 1.3-fold increase in receptors PSD incorporation rate ( η ) after cLTP induction ( Fig 5D ). Importantly, we assumed that the exocytosis rates ( β ) varied in time and space proportionally to the modeled time-dependent changes in the CNIH-2 distribution ( Fig 5A ), Eq 32 . With these parameters, our model reproduced the experimental results reported by Clavet-Fournier and colleagues ([ 63 ]), showing a gradual increase in GluA2 subunits in PSDs that reached 150% 2 hours after cLTP induction and remained elevated after 5 hours ( Fig 5D, G ). Does de-novo CNIH-2 synthesis need to occur in dendrites to obtain a long-lasting increase in GluA2-containing AMPAR concentration in PSDs? To determine whether dendritic CNIH-2 synthesis was necessary for the sustained GluA2 increase, we restricted increased CNIH-2 synthesis to the somatic region in our next simulation and used changes in CNIH-2 dendritic concentration as the change in exocytosis rate. Additionally, we also increased the incorporation rate by a factor of 1.3. Strikingly, our model predicted a marginal GluA2 increase of 15% in PSDs after 5 hours, far below the experimentally reported concentration [ 63 ]( Fig 5E, G ). This 15% increase was due to the increased synaptic incorporation. These results highlight the critical role of sustained local protein synthesis, specifically of CNIH-2, in controlling the kinetic of GluA2-containing receptors in PSDs during long-term potentiation. Finally, we tested whether a brief 10-minute CNIH-2 local synthesis increase could recapitulate experimental results ( Fig 5F, G ). This brief upregulation failed to match the amplitude of changes in synaptic GluA2, however it showed a similar kinetics as observed in the data. These findings suggest local synthesis of CNIH-2 as the mechanism behind activity-dependent GluA2 changes and highlight the importance of the long-lasting increase in CNIH-2 local synthesis and a concomitant increase in the receptor incorporation rate. Based on our experimental measurements and mathematical modeling, we replicated the fast and transient synaptic incorporation of GluA1 homomers as a result of fast and transient changes in their exocytosis coupled with a modest rise in synaptic receptor incorporation rate. We also showed that the slow but persistent synaptic changes in GluA2-containing AMPARs can be explained by a long-lasting up-regulation of the local synthesis of its auxiliary subunit CNIH-2 (2 hours), again coupled with a modest rise in the receptor synaptic incorporation rate ( η ). Thus, our integrative approach suggests a novel model of AMPAR delivery to synapses upon chemical LTP induction, where GluA1-homomers are exocytosed immediately after LTP induction while GluA2-containing AMPARs exocytosis slowly picks up relying on CNIH-2-local translation and lasts for hours ( Fig. 6 ). Download figure Open in new tab Figure 6: Proposed model of trafficking of AMPARs subtypes upon LTP induction (A) Synapses distributed along the dendritic arbors require GluA1- and GluA2-containing AMPARs for LTP. The time scales of different AMPAR trafficking mechanisms range from seconds to hours. (B) On the global scale of a neuron, both AMPARs move via fast bidirectional active transport and slow passive diffusion on the cell surface or inside the cytoplasm. (C) The local trafficking modes, including slow endocytosis and exocytosis of AMPARs, help to maintain a sufficient pool of extrasynaptic receptors. However, the balance between these two process differs for the receptor subtypes with GluA1-homomeric AMPARs having larger surface pool and GluA2-containing AMPAR having larger intracellular pool. The synaptic concentration of AMPARs is further tuned by their diffusion towards the synapse and binding to PDZ domain scaffold proteins at the PSD. (D) Our results indicate a novel mechanism specific to GluA2-containing AMPARs that requires local translation of auxiliary subunit, CNIH2, that can in turn modify the exocytosis rate of the AM-PARs, bringing them closer to synaptic sites. Discussion Our study employed an integrative approach combining experiments and computational modeling to unravel the role of the fast and slow kinetics in global and local trafficking of AMPARs. This approach provides novel insights into the molecular mechanisms underlying the long-recognized but poorly understood temporal differences in synaptic insertion dynamics of the two primary AMPAR subtypes - GluA1 homomers and GluA2-containing heteromers. Notably, we present the first evidence of how the local dendritic translation of the AMPAR auxiliary subunit CNIH-2 regulates the dynamics of GluA2-containing receptors at synapses, unveiling novel perspectives on synaptic function and plasticity ( Fig 6 ). Synthesis of AMPAR subunits from the corresponding mRNA can occur in various neuronal compartments. Previous studies [ 3 , 66 ] have identified Gria1 and Gria2 mRNAs in dendrites, suggesting that local protein synthesis can provide a plausible mechanism for rapid protein availability in dendrites. Additionally, previous studies found that the translation rate of AMPARs is location-specific and can differ under basal conditions and during plasticity. In fact, [ 31 , 66 , 67 ] reported the dynamic regulation of AMPAR mRNA subunits in synapses. However, these prior analyses have predominantly focused on proximal parts of dendritic arbors near the cell somata. In our earlier work, we also observed the dynamic regulation of the newly synthesized GluA1 subunits in dendrites during homeostatic scaling [ 17 ]. These experiments quantified the total newly synthesized GluA1 subunits after 3 hours of metabolic labeling using AHA [ 65 ]. However, the rapid processes of diffusion and intracellular trafficking made it challenging to distinguish whether the newly synthesized subunits originated from the soma or the dendrites [ 5 ]. In this study, we quantified the relative distribution of AMPAR subunits mRNA in the somatic and dendritic compartments of neurons under basal conditions. Our results revealed that mRNA coding for the pore-forming subunits GluA1 and GluA2 were predominantly somata-enriched in pyramidal neurons ( Fig 1A-C and Fig S1 ). Our results are in line with those of Glock and colleagues [ 4 ], who investigated the translatome of neurons by comparing the somata layer and neuropile (containing mainly dendrites, axons, and glia) from the CA1 region of the hippocampus. Their RNA footprint data show very small coverage for Gria1 and Gria2 in the neuropil but a large coverage in the somata layer [ 4 ]. Interestingly, despite the mRNA concentration rapidly decreasing with the distance from the soma, the GluA2 subunit exhibits an almost constant distribution along the dendrites ( Fig 1E ). Using our model, we demonstrate that these seemingly contradictory observations can be explained by a biased anterograde active transport of GluA2-containing AMPARs. Specifically, our results indicate that only biased active transport can lead to AMPAR density increase with the distance from the soma - a phenomenon reported in prior AMPAR studies as distance-dependent scaling of AMPARs [ 32 , 33 ]. We also explored the impact of other mechanisms, including passive diffusion rates and protein half-life. However, we found that these factors could not account for the observed distance-dependent scaling of AMPARs. This distance-dependent increase has been reported for other ion channels, such as HCN channels and Kv4.2 potassium channels [ 68 , 69 ]. While the increase in concentration of Kv4.2 along the dendrite is attributed to its active transport, no evidence of active transport was reported for HCN channels. We hypothesize that biased active transport can be a common mechanism for ion channels and other transmembrane proteins to achieve a distance-dependent concentration increase along dendritic arbors. Previous experimental studies have shown that endocytosis and exocytosis of AMPARs take place throughout the whole dendritic tree [ 53 , 55 ]. However, the precise balance between these two processes remains poorly characterized. In this study, we employed distinct antibodies to label surface and intracellular GluA2-containing AMPARs before and after neuronal membrane permeabilization, enabling us to visualize the two receptor populations independently ( Fig 2 ). By defining the surface-to-intracellular GluA2 fluorescent intensity ratio, we quantified the net balance between exocytosis and endocytosis rates. Our computational model demonstrates that, under basal conditions, the rate of exocytosis is lower than that of endocytosis, leading to a substantial pool of intracellular GluA2-containing AMPARs. Quantitatively, we estimate that approximately 70% of these receptors are located intracellularly, consistent with the upper bound of the experimentally reported range (30% - 70%) [ 70 – 72 ]. Furthermore, we found that this balance between exocytosis and endocytosis is consistently maintained throughout the dendritic tree ( Fig 2C ). Interestingly, we observed that the surface-to-intracellular GluA2 fluorescence ratio is higher in dendritic spines than in adjacent dendritic shafts, suggesting that GluA2-containing AMPARs may have a slight preference for direct exocytosis at perisynaptic sites rather than in nearby extrasynaptic dendritic regions ( Fig 2E ). Additionally, we predicted that under basal conditions, half of the surface AMPARs are immobilized at PSDs while the other half remain mobile in extrasynaptic regions of the plasma membrane. This is in line with the previous reports that utilized single particle tracking and fluorescence recovery after photobleaching techniques that estimated the fraction of immobile AMPARs to range between 30 and 50% [ 73 , 74 ]. Plasticity responses of AMPARs can differ depending on their subunit compositions [ 1 ]. GluA1 homomers exhibit fast surface insertion and accumulation in PSDs [ 56 , 61 ], whereas GluA2-containing, heteromeric AMPARs exhibit slower synaptic accumulation [ 61 , 63 ]. Using our model, we investigated the temporal dynamics of synaptic AMPAR density changes for these two receptor subtypes ( Fig 3 ). For GluA1 homomers, an immediate increase in exocytosis combined with a modest rise in the synaptic incorporation rate was sufficient to replicate the observed rapid and transient accumulation. These results are encouraging as our model could replicate the GluA1-homomeric dynamics without an explicit fitting to the experimental data. In contrast, GluA2-containing heteromers required a time-dependent change in the exocytosis rate to match the experimentally reported concentration, while a short-term increase alone was insufficient. We propose that this time-dependent exocytosis may be driven by local changes in the CNIH-2 concentration - a protein known to regulate AMPAR trafficking and synaptic accumulation. Interestingly, Cnih2 mRNA is highly abundant throughout the entire length of neuronal dendrites ( Fig 4A-C ), and a previous study using ribosome profiling confirmed that this dendritic mRNA undergoes local translation[ 4 ]. Using metabolic labeling with AHA for 1 hour, we observed that most newly synthesized CNIH-2 proteins localized to dendrites with their synthesis increasing after cLTP induction ( Fig 4D-F ). To explore the role of CNIH-2 in AMPAR trafficking, we used shRNAs targeting Cnih2 mRNA. Unexpectedly, this led to a specific decrease in GluA2-containing receptors at the plasma membrane ( Fig 4G-I ), confirming that CNIH-2 synthesis is crucial for GluA2-containing AMPAR surface trafficking. Our computational model further showed that a local up-regulation of CNIH-2 synthesis could explain the amplitude and temporal dynamics of GluA2-containing AMPARs ( Fig 5 ). In contrast, increasing the somatic influx of CNIH-2 resulted in only minimal effects on AMPAR concentration at the cell surface and in synapses, highlighting the critical role of the local CNIH-2 synthesis in regulating AMPAR trafficking under activity-dependent conditions. The molecular mechanisms driving the increase in local CNIH-2 synthesis upon LTP induction remain an active area of research. One key molecule implicated in memory maintenance via regulation of GluA2-containing AMPARs trafficking is protein kinase M ζ (PKM ζ ) [ 75 ]. Notably, experimental studies have shown that PKM ζ controls dendritic protein synthesis through phosphorylation-mediated inhibition of Pin1 (Peptidyl-prolyl cis-trans isomerase NIMA-interacting 1), a suppressor of protein synthesis [ 76 ]. Future experimental work can explore whether PKM ζ regulates GluA2-containing AMPARs trafficking by modulating dendritic translation of CNIH2, providing new insights into molecular mechanisms of memory maintenance. Finally, our findings elucidate alternative routes for AMPAR delivery to synapses ( Fig 6 ). GluA1 homomers appear to utilize the conventional secretory pathway, involving somatic Golgi apparatus (GA) processing, producing AMPARs with mature glycosylation profiles sensitive to Endonuclease H treatment. In contrast, GluA2-containing heteromers, synthesized at the surface of the somatic ER, may bypass GA membranes and reach the neuronal plasma membrane directly in dendrites. This process appears to be mediated by the local synthesis of CNIH-2, resulting in AMPARs with Endonuclease H-resistant glycans, as shown in previous work [ 77 ]. While a complete understanding of the molecular interplay between AMPAR auxiliary subunits remains to be uncovered, our findings provide insights into how distinct AMPAR populations are selectively recruited to synapses under different conditions. This selective recruitment may have profound implications for therapeutic strategies aimed at restoring or modulating AMPAR synaptic transmission and plasticity - processes that are often disrupted in neurological and neurodegenerative disorders. Author Contributions SW performed the simulations and analyzed the simulations and experimental data. SW built the model with theoretical and conceptual input from NK and TT and experimental feedback from EMS, ASH, and TT. The experimental data shown in Figures 1 and 4 were acquired by ASH in EMS’s lab at the Max Planck Institute for Brain Research. The experimental data shown in Figure 2 were acquired by MKK in AAP’s Lab at Goethe University and were supervised by AAP. AB performed the shRNA validation experiments S11 in EMS’s lab at the Max Planck Institute for Brain Research under the supervision of ASH and EMS. CNIH-2 protein data S10 were acquired by ASH at Radboud University. SW, NK, ASH, and TT wrote the original draft of the manuscript, and all authors edited the manuscript. Methods Experimental methods View this table: View inline View popup Download powerpoint Experimental protocols Mouse Experiments Isolation of embryonic hippocampal neurons Dissociated hippocampal neuron cultures were prepared from embryonic day E16.5-18.5 mouse embryos (C57Bl/6N; Charles River Laboratories). The hippocampi were dissected in pre-chilled dissection medium (Hank’s balanced salt solution (HBSS) with 1% GlutaMax, 1% HEPES, 1% Pen/Strep), digested in 1ml Papain for 15min at 37°C, washed twice in pre-warmed DMEM medium (Dulbecco’s modified Eagle medium (DMEM) with 10% fetal bovine serum (FBS)) and twice in pre-warmed NB+ medium (Neurobasal medium supplemented with 2mM Glutamax, 77.7mM D-Glucose). Neurons were then gently dissociated with a fire-polished Pasteur pipette, centrifuged for 5min at 71 g, and finally plated at 30 − 40 × 10 3 cells / cm 2 on poly-D-lysine coated coverslips placed in a 24-well plate. Neurons were kept in NB++ medium (Neurobasal medium supplemented with 2mM Glutamax, 77.7mM D-Glucose and 1:50 B27) at 37°C and 5% CO2 for 14 days until fixation. Animal experiments were approved by the Hessian authorities. Transfection For subsequent staining, neurons were transfected with a plasmid expressing GFP at 11 days in vitro (DIV) using calcium phosphate transfection. Neurons were transfected by adding 15µl 2x HeBs buffer (274mM NaCl, 10mM KCl, 1.4mM Na 2 HPO 4 x 7 H 2 O , 15mM D-Glucose, 42mM HEPES) with 15µl DNA and calcium solution (2 µg of DNA in 250 mM CaCl 2 solution) per well at DIV11 after plating. The conditioned culture medium was replaced with 400 µl of NB+ medium before adding the DNA mixture. After incubation for 10 min at 37°C under 5% CO 2 , the neurons were washed three times with NB+, which had been pre-incubated at 37°C under 5% CO 2 , and the conditioned NB++ medium was added back to the cells. Immunostaining, image acquisition and analysis Coverslips with attached neurons (DIV 14) were washed and fixed once at room temperature DPBS containing calcium and magnesium before being fixed in 4% paraformaldehyde in PBS (pH = 7.4) containing 4% of sucrose for 10 minutes on ice. Afterward, neurons were washed with NH4Cl for 10 min on ice, blocked in blocking buffer (2% BSA/4% ND-S/PBS) for 30 min at room temperature, and incubated over three days with mouse anti-GluA2 antibody (1:500, Millipore) at 4°C and labeled with donkey-anti-mouse Cy3 (1:500, Jackson Immuno Research) overnight at 4°C. For the visualization of extra- and intracellular GluA2, neurons were subsequently permeabilized for 5 min in 0.4% Triton X-100 in PBS, blocked for 30 min, incubated overnight with the same mouse anti-GluA2 (1:500) and chicken anti-GFP (1:1000, Abcam), and labeled with donkey anti-mouse Cy5 (1:500, Jackson Immuno Research) and donkey anti-chicken Alexa 488 (1:500, Jackson Immuno Research) overnight at 4°C. Hippocampal neurons were imaged using Leica TCS SP5 confocal microscopes and 63x oil objectives (NA 1.4). Z stacks spanning the entire volume of neurons were obtained, and channels were separated and collapsed to a sum intensity projection in ImageJ. Manual annotation of dendritic tree using segmented line tool and somata using polygonal ROIs was performed. Signal intensity profile in dendritic ROIs the was measured using a constant width and integrated density was calculated for somatic ROIs. Rat Experiments Cultured neurons Dissociated rat hippocampal cultures were prepared and maintained as described previously [ 85 ]. Briefly, we dissected hippocampi from postnatal day 0-1 rat pups of either sex (Sprague-Dawley strain; Charles River Laboratories), dissociated them with papain (Sigma) and plated them at a density of 40 × 10 3 cells/ cm 2 on poly-D-lysine coated glass-bottom Petri dishes (MatTek). Neurons were maintained and matured in a humidified atmosphere at 37°C and 5% CO 2 in growth medium (Neurobasal-A supplemented with B27 and GlutaMAX-I, life technologies) for 18-21 days in vitro (DIV) to ensure synapse maturation. All experiments complied with national animal care guidelines and the guidelines issued by the Max Planck Society, and were approved by local authorities. For transfection, DIV7-11 neurons were transfected using Effectene (Qiagen), as previously described. Transfected cells were maintained until DIV19 for experiments. In situ hybridization in cultured neurons In situ hybridization was performed using QuantiGene ViewRNA kit from Affymetrix (now Thermo Fisher Scientific) mostly following the provider instructions. In brief, cells (DIV 18-24) were fixed for 20 min at room temperature using a 4% paraformaldehyde (PFA) solution (4% paraformaldehyde, 2.5% Sucrose, in lysine-phosphate buffer). The Proteinase K treatment was omitted in order to preserve the integrity of the dendrites. After permeabilization using the provider’s detergent buffer for 2 min, cells were directly incubated for 3h at 40°C with detection probes. Incubations with pre-amplification, amplification and detection probes were reduced to 40 min each. After completion of in situ hybridization, cells were washed with PBS and incubated in blocking buffer (4% goat serum in PBS) for 1hr. Neurons were subsequently processed for immunofluorescence using standard methods as described below. Immunofluorescence in cultured neurons All steps were performed at room temperature, unless stated otherwise. Glass bottom dishes with attached neurons (DIV 18-21) were fixed in paraformalde-hyde 4% in lysine phosphate buffer pH7.4 containing 2.5% sucrose for 15-20 min. For simple immunofluorescence, neurons were permeabilized for 10 min in PBS + 0.5% Triton-X 100. After 2 quick washes in PBS, neurons were then incubated in blocking buffer (4% goat serum in PBS) for 30 min. Then neurons were incubated for 1 hr or overnight with primary antibodies, and after 3 washes of 5 min, they were incubated for 1.5 hr with secondary antibodies. We used the following antibodies: guinea pig anti-MAP2 (Synaptic Systems, 1:2000), mouse anti-biotin (Sigma, 1:1000), rabbit anti-biotin (Cell Signaling, 1:1000), mouse anti-Puromycin (Kerafast, 1:2500), homemade rabbit anti-Puromycin (1:250), homemade anti-GluA1 N-terminal domain (targeted sequence QWRTSDSRDHTRVDWKRPKC; KO validated by western blot) (fixed sample 1/500; live-labeling 1/200), mouse anti-GluA2 N-terminal domain from Eric Gouaux previously used in Nair et al., 2013 (fixed sample 1/100; live-labeling 1/500), rabbit anti-CNIH-2 (Synaptic Systems, 1/200), homemade rabbit anti-TARP Gamma-8 C-terminal domain (targeted sequence PGTLSKEAAASNTNT, 1/200; the antibody staining co-localizes with synaptic protein Bassoon and shows strong extrasynaptic labeling). Chemical long-term potentiation (cLTP) For cLTP experiments, the original medium of cultured neurons was replaced for 5 min by artificial cerebrospinal fluid (ACSF) at 37°C. In control condition, the ACSF contained Ca2+ (2 mM) and Mg2+ (2 mM) and supplemented with B27 and MEM Amino Acids (50X) (Thermo Fisher Scientific). In the cLTP-induction condition, the ACSF differed from the control-ACSF as followed: Ca2+ (3 mM), Mg2+ (0 mM), glycine (200 µ M), picrotoxin (100 µ M). After the 5 min induction the cells were placed in growth medium containing either methionine (Met) or L-Azidohomoalanine (AHA) at 4 mM for 1 hr. Fluorescence non-canonical amino acid tagging To investigate the distribution of newly synthesized proteins we used a recently developed method combining fluorescence non-canonical amino acid tagging with the proximity ligation assay (FUNCAT-PLA) [ 65 ]. Cultured neurons were incubated for 1 hr with growth medium were Met was replaced by AHA (4 mM). In control conditions the growth medium contained Met (4 mM) and no AHA. Puromycylation To determine the sites of synthesis of AMPAR components we used a recently developed method combining puromycylation and proximity ligation assay (Puro-PLA) [ 65 ]. Cultured neurons were labeled with 10 µ M puromycin (Sigma-Aldrich) for 2 min. In control experiments cells were treated with 40 µ M of the protein synthesis inhibitor anisomycin (Sigma-Aldrich) for 30 min prior to and during puromycin labeling. Proximity ligation assay Detection of newly synthesized proteins by proximity ligation was carried out using anti-puromycin or anti-biotin antibodies (in case of AHA metabolic labeling) in combination with protein-specific antibodies against our proteins of interest (tom Dieck et al., 2015)(complete list of anti-bodies in the Immunofluorescence section). We used Duolink reagents (Sigma) and followed the protocol provide by the manufacturer with some modifications described below. We routinely used rabbit PLAplus and mouse PLAminus probes amplification and label probe binding. Briefly, after washing the primary antibodies 3 times in PBS, PLA probes were applied in 1:10 dilution in PBS with 4% goat serum for 1 h at 37 °C, washed several times with wash buffer A (0.01 M Tris, 0.15 M NaCl, 0.05% Tween 20) and incubated for 30 min with the ligation reaction containing the circularization oligos and T4 ligase prepared according to the manufacturer’s recommendations (Duolink Detection reagents Red, Sigma) in a prewarmed humidified chamber at 37°C. Amplification and label probe binding was performed after further washes with wash buffer A with the amplification reaction mixture containing Phi29 polymerase and the fluorophore-labeled detection oligo prepared according to the manufacturer’s recommendations (Duolink Detection reagents Red, Sigma) in a prewarmed humidified chamber at 37 °C for 100 min. Amplification was stopped by three washes in wash buffer B (0.2 M Tris, 0.1 M NaCl, pH 7.5). For better signal stability, cells were kept in wash buffer B at 4°C until imaging. Image acquisition and analysis Cultured cells were imaged using Zeiss LSM780/880 confocal microscopes using a 63x oil objective NA 1.4 and a 40x oil objective NA 1.3. Z-stacks spanning the entire volume of imaged neurons were obtained. Image Analysis Puncta detection Cell culture images were analyzed to access mRNA and newly synthesized protein distributions in somata and dendrites using a custom built Python tool called SpyDen [ 84 ]. The code and instruction manual for SpyDen is freely-available for download online https://github.com/meggl23/SpyDen . Statistical analysis All data are shown as the mean ± s.e.m. unless stated otherwise. Box plots show the median, quartile and whisker using lines and mean using triangle. Swarm plot shows the individual values. We performed all the data analysis using Python programming language. We did not assume that the data were normally distributed, hence we used Mann-Whitney (rank-sum) test for single comparisons unless stated otherwise. We did not use any statistical methods to predetermine sample sizes. Model for the combined distribution of surface and cytoplasmic proteins in dendrites (related to section 2.1 and Fig 1 ) Under basal conditions, AMPAR are primarily produced in the soma and move along the dendrites via active intracellular transport as well as diffusion within both the plasma membrane and intracellular compartments. The following boundary value problem describes the dynamics of a protein ( P ) along the dendrites shaped by the diffusion ( D P ), active transport ( V P ) and degradation ( λ P ) We assume that the concentration of the protein ( P ) satisfies the boundary conditions whereby J influx > 0 is the rate of protein influx in the dendrite from the soma and Eq (6) means that no proteins come out of the dendritic tip located at the distance L from the soma. The steady state solution of the Eq (4) (i.e., solution under the assumption ) is a sum of two exponential functions whereby . We compute the constants C 1 , C 2 using the boundary conditions in Eqs (5), (6) and obtain the steady state dendritic distribution of the protein Model for GluA2-containing AMPARs concentration in surface, cytoplasm and PSD (related to section 2.2 and Fig 2 ) To compute the steady-state solutions of the model in Eq (1) - (3), we assume that the time derivatives of the functions P s , P c and P psd are equal to 0 and obtain the following system of ODE The relation in Eq (11) implies that steady state of the synaptic proteins ( P psd ) is related to the steady state concentration of the proteins in the cell membrane Combining Eqs (10) and (11) , we reduce the system in Eqs 9 - 11 to two equations We assume that the dendrite has the length L and the protein molecules can not escape the dendrite at the point x = L , i.e., the functions P c and P s fulfill the following boundary conditions Also, we assume that the proteins P c produced in the soma are released in the cytoplasm at x = 0 with a constant influx rate J cin and reach the plasma membrane P s only via exocytosis. Hence we assume noflux from the soma into the dendrite, i.e., We solve Eqs 13 numerically using Python library solve_bvp which computes solution of the ODE system subjected to two-point boundary conditions. Fitting model to the data Fitting mRNA data To estimate the decrease in Gria1, Gria2, Cnih2 and CamK2a mRNA distributions along a dendritic branch, we performed automatic optimization of the binned empirical data (bin size = 7.5 µ m) and an exponential function of the form f ( x ) = A * e − Bx , where f represents the distribution of the mRNA along a line dendrite. Independent variable x corresponds to the location on the dendrite. Parameters A ( > 0) and B ( > 0) were fitted using minimize function of lmfit python package for nonlinear least square minimization between the data and the model. Fitting D P and V P for total GluA2 model To estimate the global transport parameters for the model, we numerically minimized the error between normalized distribution of the first 100 µ m dendritic distribution from the model and experimental data over the two-dimensional parameter space to fit the experimental data. For both, experimental data and model, a binning (bin size = 5 µ m) and normalization by the first bin value was done before the optimization. The fitting was done using the minimize function of lmfit python package for nonlinear least square minimization between the data and the model. We note that in all of our simulations, we assumed the length of the model dendrite to be L = 500 µm and then performed the fitting. Exocytosis rate for the extended GluA2 model To calculate net exocytosis rate at a steady state we computed the ratio of total surface GluA2 fluorescence to the total intracellular GluA2 fluorescence . The following proof shows that this ratio is indeed equal to the ratio between the exocytosis and endocytosis rates ( β / α ). Integrating Eqs 13 from 0 to L , we obtain Using the boundary conditions from Eqs. (14 - 15), we obtain Hence, we compute the ratio between the exocytosis and endocytosis rates Here we used the endocytosis rate as described in 1. For the GluA1-homomer version of the model, we used a ratio of 1.7 to obtain the exocytosis rate. This ratio was reported in [ 86 ]. Fitting D s , D c and V p in the extended model The exocytosis rate was calculated using the above mentioned method before fitting the global transport parameters, D s , D c and V p . To estimate the parameters we numerically minimized the error between normalized distribution of the first 100 µ m dendritic distribution computed as a model solution and experimental data. Fitting η in the extended model All parameters of the extended model were either computed as described in previous sections or set to an experimentally reported value before we estimated the incorporation rate η (see, Table 1 ). To fit η in the extended model ( Eqs 1 -3), we calculated the mean ratio of P psd and P s over the first 60 µm length of the dendrite using minimize function to reduce the difference between this ratio and the mean surface synaptic enrichment. Baseline model of GluA1-homomeric AMPARs To generate the steady-state trafficking dynamics of GluA1-homomeric AMPARs, we used the same model as for GluA2-containing AMPARs described in Eqs (1)-(3). For the GluA1 model, and used the following parameters which were either previously reported or estimated in this study Supplemental information Time-dependent solution of the extended AMPAR trafficking model To study plasticity-dependent changes in AMPAR concentration in different compartments (for both GluA1 homomeric and GluA2-containing heteromeric receptors), we use the original time-dependent model in Eqs 1 - 3 and compute its solution numerically. To this end, we discretize the spatial variable of the PDE (Δ x = 1 µ m) and transform the system of PDE into a system of ODEs We determine P s , P c at all intermediate spatial nodes ( i = 1, …, L − 1). For the boundary nodes ( i = 0, L ), we use the boundary conditions Eqs (14)-(15) at x = 0 and x = L and obtain We use the scipy.solve_ivp function to compute P c and P s with the explicit Runge-Kutta method of order 5(4). Trafficking dynamics and steady state distribution of CNIH-2 mRNA The production, degradation and transport dynamics of CNIH-2 mRNA molecules can be described by the model proposed in [ 5 ] First, we determine the steady state distribution for the equation 19 . Setting the time-derivative to 0 we obtain The general solution, , of the equation 20 reads where . We calculate the constants C 1 , C 2 using the following boundary conditions: and derive the steady state distribution . Here J Rin is the constant rate of influx of mRNA from soma into the dendrite. The second boundary condition describes the no-flux boundary condition. We obtain Trafficking dynamics and steady state distribution of CNIH-2 protein The CNIH-2 protein model considers the CNIH-2 somatic translation rate, local dendritic translation rate, diffusion, and active transport. Hence, we can use the protein trafficking model introduced in [ 5 ] for CamK2a protein to model the CNIH-2 protein dynamics. We obtain First, we determine the steady state of the equation 24 by setting the time-derivative to 0 The general solution of the equation 25 reads whereby We use the following boundary conditions to obtain the constants C 1 and C 2 in Eq. 26 The final steady state CNIH2 protein distribution reads Time-dependent solution of CNIH-2 subunit trafficking model How does the concentration of CNIH-2 evolve in our model dendrite over time? To calculate the spatiotemporal distribution of CNIH-2 protein ( P CNIH −2 ( x, t )), we return to the original partial differential equation of CNIH-2 24. In order to simulate the spatio-temporal dynamics of CNIH-2 during basal state and upon cLTP induction, we first discretized the spatial variable ( x ) of the model and converted PDE to a time-dependent ODE We computed coefficients corresponding to P i , P i −1 and P i +1 and rewrote the last equation as Now we could determine P at all intermediate nodes ( i = 1, …, L −1) except the boundary nodes i = 0, L . For the boundary nodes, we used boundary conditions at x = 0 and x = L and determined the temporal evolution of the CNIH-2 protein at the boundary points and obtained the differential equations Finally, we used the solve _ ivp function of the scipy Python library to integrate the ODE in Eqs. (29 - 31) and calculated P CNIH −2 as a function of dendritic location and time. CNIH-2 local translation upregulation upon plasticity and exocytosis of GluA2-containing AMPARs In case of cLTP-driven translation up-regulation ( Fig 5A ), we multiplied the basal translation rate ( β P ) at each location of the dendrite by the factor of 500 for 2 hours following cLTP induction. This is in-line with previous reports of translation up-regulation which peaks at 30 minutes and sustains for 2 hours following cLTP induction [ 88 , 89 ]. After 2 hours of simulation time, the local translation rate was reset to the basal level. For the increased somatic influx simulation (in Fig 5B, E ), we increased the in-flux rate J pin by a factor of 500. After 2 hours of simulation time, we reset the influx rate to a baseline value. Next, we considered the local concentration of CNIH-2 to be a direct regulator of GluA2-containing AM-PAR exocytosis. Hence, for plasticity simulations, we changed the AMPARs exocytosis rate proportionally to the modification in CNIH-2 concentration. More precisely, we normalized the CNIH-2 concentration after plasticity induction by the baseline concentration of CNIH-2 and multiplied it with the basal exocytosis rate. Here, β ( x, t ) represents the exocytosis rate as a function of dendritic location and time. Estimation of Endocytosis rates from published live imaging Usually, studies investigating receptor internalization, endocytosis rate is measured in terms of events per µm 2 per minute. For the consistency with trafficking parameters in our models, we converted endocytosis rate to represent an event per µm 2 per minute. To this end, we multiplied the reported rate with the number/concentration of receptors per vesicle and the area of the ROIs. For example, [ 53 ] report the endocytosis rate to be (2.1 ± 1.1) * 10 −3 events per µm −2 per minute for GluA1 and (2.5 ± 0.4) * 10 −3 events per µm −2 per minute for GluA2 vesicles. These rates were measured in a circular ROI of radius 300 nm . The estimated number of AMPARs per vesicle is 50 [ 90 , 91 ]. Using these values we obtain the mean endocytosis rates: for GluA1 and GluA2 AMPA receptors, respectively. Extended Figures Download figure Open in new tab Figure S1: GluA1 mRNA is predominantly localised in soma. (A) Top : Cultured rat hippocampal neurons at 18-21 DIV processed for FISH against Gria1 (left, green) and CamK2a mRNA (right, green) and fluorescently immunostained (FI) MAP2 (magenta). Scale bar = 20 µ m. Bottom : Zooming into a representative dendrite shows a smaller number of Gria1 mRNA (left) in comparison to CamK2a mRNA (right). Scale bar = 5 µ m. (B) Somatic (hollow) and dendritic (filled) fraction of total mRNA for Gria1 (19 cells, p-value: 1e-06) and CamK2a (19 cells, p-value: 1), Gria1 -dendrite vs CamK2a -dendrite (p-value:0.019). (C) Exponential fit of mRNA puncta density distribution for Gria1 ( n = 12 dendrites, exponent Gria1 = -0.03 ± 0.007) compared to the CamK2a distribution ( n = 33 dendrites, exponent CamK2a = -0.001 ± 0.001), see Methods for mRNA exponential fit. Inset : Normalized fitted Gria1 mRNA distribution compared to CamK2a . Download figure Open in new tab Figure S2: Protein distributions for different velocity, diffusion and life-time constants, related to Figure 2 F-H . (A)-(C) Parameters as in Figure 2 in all of the panels, we used the value of at x = 0 to normalized the distributions. Download figure Open in new tab Figure S3: Impact of the active transport velocity on dendritic distribution of AMPARs (A) Binned distribution of Glua2 subunit of AMPAR, slow and fast transport distribution and theory fit distribution. (B)-(D)) steady-state, dendritic distribution of Glua2-containing AMPAR (surface+intracellular) for three different sets of transport parameters. The degradation and the somatic flux parameters were set to λ P = 2.57 * 10 −6 s −1 , J Pin = 0.01 proteins/s. The insets contain the dendritic concentration normalized at the starting point of primary dendrite ( x = 0). Download figure Open in new tab Figure S4: Fitting the global parameters of the full model to surface and intracellular GluA2 fluorescent signaling data. (A) Dendritic distribution of soma-normalized GluA2 confined to the surface (sGluA2 data and model fit,) and B intracellular GluA2 (iGluA2 data and model fit, cyan, right). N Dendrites = 11 Download figure Open in new tab Figure S5: Synaptic enrichment increases with distance (A) scatter plot of Synaptic enrichment vs dendritic distance with regression line (magenta) and 95% CI (shaded region), spearman correlation coefficient =0.32,p-value=0.017. (B) scatter plot of spine areas vs dendritic distance with regression line (magenta) and 95% CI (shaded region), spearman correlation coefficient =0.0.05,p-value=0.73. (C) scatter plot of integrated fluorescence intensity from spine (in magenta) and shaft (in cyan) vs dendritic distance with regression lines for spine in magenta and 95% CI (shaded region) and shaft in cyan with 95 % CI, spearman correlation coefficients = 0.14,p-value=0.31 (for spines) and spearman correlation coefficients = -0.27,p-value=0.045 (for shaft). Shaft shows a decrease and spines remain constant. Download figure Open in new tab Figure S6: Synaptic enrichment calculated using two different methods are comparable (A) Representative confocal image of hippocampal neuron cultures from mice transfected with GFP, anti-GluA2 labeled before and after permeabilization (sGluA2 and iGluA2 respectively. scale bar, 5 µm (B-C) Representative line scan of sGluA2 across synaptic and dendritic ROIs (inset) used to calculate synaptic enrichment factor as described in [ 58 ]. (C) for iGluA2. Two Gaussian kernels were fitted to the normalized intensity values for calculating the synaptic enrichment factors. (D-E) Correlation of synaptic enrichment factor calculated using the two methods described in [ 92 ] and [ 58 ], n spines = 55 (from 8 neurons) for sGluA2 (D) and iGluA2 (E) (sGluA2: r= 0.43, p = 0.001, iGluA2: = 0.47, p = 3E-4). (F-G) Median and inter-quartile interval synaptic enrichment factor gives comparable results using the two methods for both sGluA2 (F) and iGluA2 in ( G ). (mean is shown as black triangle). Download figure Open in new tab Figure S7: Simulating plasticity response of GluA1-homomeric AMPARs required changes in exocytosis rate β and synaptic incorporation rate η . A) We used a step function where the exocytosis rate ( β ) was increased 3.5 fold to the basal rate. This matches well with the change in exocytosis frequency reported in ([ 61 ] FigS3C ). B) an additional change in the synaptic incorporation rate ( η ) was necessary which reflects the average change in spine head size measured in ([ 63 ] Fig2F ). Download figure Open in new tab Figure S8: GluA1-homomeric AMPAR requires fast increase in exocytosis and enhanced diffusional trapping for LTP dependent changes (related to Fig 4A ) In both subpanels, a 10 min chemi-calLTP stimulation was applied which led to different changes in trafficking parameters. (A) The exocytosis rate ( β ) was step increased 3.5 folds for the duration of stimulus followed by a step decrease back to basal level. The synaptic incorporation ( η ) rate was left unaltered (B) The exocytosis rate ( β ) was kept at basal level, however the The synaptic incorporation rate ( η ) was increased by 30%. None of the simulations could replicate the experimental observations. Download figure Open in new tab Figure S9: Constant increase in GluA2-AMPARs exocytosis cannot explain their temporal profile upon cLTP induction. We simulated the GluA2-containing AMPARs model with a constant increase in exocytosis rate by 3.5 folds for A) 10 minutes, B) 30 minutes. C) 1 hour. In addition to the synaptic incorporation rate, η was increased by 1.5 folds for the entire duration of simulation. None of the the simulation matched the experimental profile of synaptic GluA2 increase from [ 63 ]. Download figure Open in new tab Figure S10: Auxiliary subunit CNIH-2 gets translated locally in the dendrite (A) Top Representative images of cultured hippocampal neurons showing the distribution of the newly synthesized AMPAR auxiliary subunit CNIH-2 after 1 hr of metabolic labeling with AHA or, in control conditions, with methionine (Met). Scale bar = 20 µ m. bottom Representative dendrites showing that CNIH-2 nascent protein is abundant in dendrite after 1h of metabolic labeling. Scale bar = 10 µ m. (B-C) Box and strip plot Bar density of new CNIH-2 in soma (B) and in dendrite (C) ( n met = 33, n AHA = 29 neurons analyzed per condition) D Fraction of total nascent protein in AHA treatment found in dendritve vs somatic compartment. ***, p ≤ 0.001. Circles represent individual neuron values, box lines for median, mean shown as black triangles. (E) Top : Cultured rat hippocampal neurons at DIV 18-21 processed for antibody labeling of CNIH-2 protein (gray) and FI MAP2 (magenta). Bottom : Zoom into a representative dendrite shows a homogeneous GluA2 distribution along the first 100 µ m. Scale bar = 20 µ m. (F) MAP2 normalized CNIH-2 intensity, (5 µ m bins, median: green squares with IQR). Download figure Open in new tab Figure S11: Validation of CNIH-2 knockdown using an shRNA strategy. A Bar graphs and strip plot showing the density of Cnih2 mRNA in soma of neurons expressing a control shRNA or an shRNA against Cnih2 ( n Cntl = 8, n CNIH −2 = 18 neurons analyzed per condition). *** p ≤ 0.001. B Bar graphs and stripplot showing the density of nascent CNIH-2 proteins in soma of neurons expressing a control shRNA or an shRNA against Cnih2 ( n Cntl = 49, n CNIH −2 = 23 neurons analyzed per condition). ***, p ≤ 0.001. All data are shown as individual neurons. Circles represent individual neuron values, box lines for median, mean shown as black triangles. Acknowledgments We would like to thank all our lab members for fruitful discussions and helpful feedback. This study was supported by the University of Bonn Medical Center (SW, NK, TT), the University of Mainz Medical Center (SW, TT), the German Research Foundation via CRC1080 (SW, TT, MKK, AAP), the Donders Institute for Brain, Cognition and Behaviour and Faculty of Science, Radboud University Nijmegen Netherlands (AH). This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (‘MolDynForSyn’, grant agreement No. 945700) (TT) & (‘MemCode’, grant agreement No. 101076961) (AH). AH also received support from the EMBO long-term postdoctoral fellowship (ALTF 1095-2015) and the Alexander von Humboldt Foundation (FRA-1184902-HFST-P). Footnotes ↵ * Co-senior authors References 1. ↵ Diering , G. H. & Huganir , R. L. The AMPA receptor code of synaptic plasticity . Neuron 100 , 314 – 329 ( 2018 ). OpenUrl CrossRef PubMed 2. ↵ Ofer , N. , Berger , D. R. , Kasthuri , N. , Lichtman , J. W. & Yuste , R. Ultrastructural analysis of dendritic spine necks reveals a continuum of spine morphologies . Developmental neurobiology 81 , 746 – 757 ( 2021 ). OpenUrl CrossRef PubMed 3. ↵ Cajigas , I. J. et al. The local transcriptome in the synaptic neuropil revealed by deep sequencing and high-resolution imaging . Neuron 74 , 453 – 466 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 4. ↵ Glock , C. et al. The translatome of neuronal cell bodies, dendrites, and axons . Proceedings of the National Academy of Sciences 118 , e2113929118 ( 2021 ). OpenUrl Abstract / FREE Full Text 5. ↵ Fonkeu , Y. et al. How mRNA localization and protein synthesis sites influence dendritic protein distribution and dynamics . Neuron 103 , 1109 – 1122 ( 2019 ). OpenUrl CrossRef PubMed 6. ↵ Seeburg , P. H. The TiPS/TINS lecture: the molecular biology of mammalian glutamate receptor channels. en . Trends Pharmacol. Sci . 14 , 297 – 303 ( Aug . 1993 ). OpenUrl CrossRef PubMed 7. Hollmann , M. & Heinemann , S. Cloned glutamate receptors. en . Annu. Rev. Neurosci . 17 , 31 – 108 ( 1994 ). OpenUrl CrossRef PubMed Web of Science 8. ↵ Sobolevsky , A. I. , Rosconi , M. P. & Gouaux , E. X-ray structure, symmetry and mechanism of an AMPA-subtype glutamate receptor. en . Nature 462 , 745 – 756 ( Dec . 2009 ). OpenUrl CrossRef PubMed Web of Science 9. ↵ Tomita , S. et al. Functional studies and distribution define a family of transmembrane AMPA receptor regulatory proteins. en . J. Cell Biol . 161 , 805 – 816 ( May 2003 ). OpenUrl Abstract / FREE Full Text 10. ↵ Milstein , A. D. , Zhou , W. , Karimzadegan , S. , Bredt , D. S. & Nicoll , R. A. TARP subtypes differentially and dose-dependently control synaptic AMPA receptor gating. en . Neuron 55 , 905 – 918 ( Sept . 2007 ). OpenUrl CrossRef PubMed Web of Science 11. ↵ Schwenk , J. et al. Functional proteomics identify cornichon proteins as auxiliary subunits of AMPA receptors . Science 323 , 1313 – 1319 ( 2009 ). OpenUrl Abstract / FREE Full Text 12. ↵ Schwenk , J. et al. High-resolution proteomics unravel architecture and molecular diversity of native AMPA receptor complexes . Neuron 74 , 621 – 633 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 13. ↵ Khodosevich , K. et al. Coexpressed auxiliary subunits exhibit distinct modulatory profiles on AMPA receptor function. en . Neuron 83 , 601 – 615 ( Aug . 2014 ). OpenUrl CrossRef PubMed 14. ↵ Von Engelhardt , J. et al. CKAMP44: a brain-specific protein attenuating short-term synaptic plasticity in the dentate gyrus. en . Science 327 , 1518 – 1522 ( Mar . 2010 ). OpenUrl Abstract / FREE Full Text 15. ↵ Burnashev , N. , Monyer , H. , Seeburg , P. H. & Sakmann , B. Divalent ion permeability of AMPA receptor channels is dominated by the edited form of a single subunit . Neuron 8 , 189 – 198 ( 1992 ). OpenUrl CrossRef PubMed Web of Science 16. ↵ Shi , S.-H. , Hayashi , Y. , Esteban , J. A. & Malinow , R. Subunit-specific rules governing AMPA receptor trafficking to synapses in hippocampal pyramidal neurons . Cell 105 , 331 – 343 ( 2001 ). OpenUrl CrossRef PubMed Web of Science 17. ↵ Sutton , M. A. et al. Miniature neurotransmission stabilizes synaptic function via tonic suppression of local dendritic protein synthesis . Cell 125 , 785 – 799 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 18. ↵ Morise , J. et al. AMPA receptors in the synapse turnover by monomer diffusion . Nature communications 10 , 5245 ( 2019 ). OpenUrl CrossRef PubMed 19. ↵ Wakazono , Y. , Midorikawa , R. & Takamiya , K. Temporal and quantitative analysis of the functional expression of Ca2+-permeable AMPA receptors during LTP. en . Neurosci. Res . 198 , 21 – 29 ( Jan . 2024 ). OpenUrl CrossRef PubMed 20. ↵ Coombs , I. D. & Cull-Candy , S.G. Single-channel mechanisms underlying the function, diversity and plasticity of AMPA receptors . en. Neuropharmacology 198 , 108781 ( Oct . 2021 ). OpenUrl CrossRef PubMed 21. ↵ Park , J. Phosphorylation of the AMPAR-TARP complex in synaptic plasticity. en . Proteomes 6 , 40 ( Oct . 2018 ). OpenUrl CrossRef PubMed 22. ↵ Chater , T. E. & Goda , Y. The shaping of AMPA receptor surface distribution by neuronal activity. en . Front. Synaptic Neurosci . 14 , 833782 ( Mar . 2022 ). OpenUrl CrossRef PubMed 23. ↵ Jacobi , E. & von Engelhardt , J. Modulation of information processing by AMPA receptor auxiliary subunits. en . J. Physiol . 599 , 471 – 483 ( Jan . 2021 ). OpenUrl CrossRef PubMed 24. ↵ Boudkkazi , S. , Brechet , A. , Schwenk , J. & Fakler , B. Cornichon2 dictates the time course of excitatory transmission at individual hippocampal synapses . Neuron 82 , 848 – 858 ( 2014 ). OpenUrl CrossRef PubMed Web of Science 25. ↵ Harmel , N. et al. AMPA receptors commandeer an ancient cargo exporter for use as an auxiliary subunit for signaling . PloS one 7 , e30681 ( 2012 ). OpenUrl CrossRef PubMed 26. ↵ Brockie , P. J. et al. Cornichons control ER export of AMPA receptors to regulate synaptic excitability . Neuron 80 , 129 – 142 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 27. ↵ Herring , B. E. et al. Cornichon proteins determine the subunit composition of synaptic AMPA receptors . Neuron 77 , 1083 – 1096 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 28. ↵ Coombs , I. D. et al. Cornichons modify channel properties of recombinant and glial AMPA receptors . Journal of Neuroscience 32 , 9796 – 9804 ( 2012 ). OpenUrl Abstract / FREE Full Text 29. ↵ Bissen , D. , Foss , F. & Acker-Palmer , A. AMPA receptors and their minions: auxiliary proteins in AMPA receptor trafficking . Cellular and Molecular Life Sciences 76 , 2133 – 2169 ( 2019 ). OpenUrl CrossRef PubMed 30. ↵ Jung , Y. et al. BDNF-induced local translation of GluA1 is regulated by HNRNP A2/B1 . Science Advances 6 , eabd2163 ( 2020 ). OpenUrl FREE Full Text 31. ↵ Livingstone , R. W. et al. Secreted amyloid precursor protein-alpha enhances LTP through the synthesis and trafficking of Ca2+-permeable AMPA receptors . Frontiers in Molecular Neuroscience 14 , 660208 ( 2021 ). OpenUrl CrossRef PubMed 32. ↵ Andrasfalvy , B. K. & Magee , J. C. Distance-dependent increase in AMPA receptor number in the dendrites of adult hippocampal CA1 pyramidal neurons. en . J. Neurosci . 21 , 9151 – 9159 ( Dec . 2001 ). OpenUrl Abstract / FREE Full Text 33. ↵ Shipman , S. L. , Herring , B. E. , Suh , Y. H. , Roche , K. W. & Nicoll , R. A. Distance-dependent scaling of AMPARs is cell-autonomous and GluA2 dependent . Journal of Neuroscience 33 , 13312 – 13319 ( 2013 ). OpenUrl Abstract / FREE Full Text 34. ↵ Bressloff , P. C. & Newby , J. M. Stochastic models of intracellular transport . Reviews of Modern Physics 85 , 135 ( 2013 ). OpenUrl CrossRef 35. ↵ Hangen , E. , Cordelieres , F. P. , Petersen , J. D. , Choquet , D. & Coussen , F. Neuronal activity and intracellular calcium levels regulate intracellular transport of newly synthesized AMPAR . Cell reports 24 , 1001 – 1012 ( 2018 ). OpenUrl CrossRef PubMed 36. ↵ Borgdorff , A. J. & Choquet , D. Regulation of AMPA receptor lateral movements . Nature 417 , 649 – 653 ( 2002 ). OpenUrl CrossRef PubMed Web of Science 37. ↵ Dörrbaum , A. R. , Kochen , L. , Langer , J. D. & Schuman , E. M. Local and global influences on protein turnover in neurons and glia . Elife 7 , e34202 ( 2018 ). OpenUrl CrossRef PubMed 38. ↵ Cohen , L. D. et al. Metabolic turnover of synaptic proteins: kinetics, interdependencies and implications for synaptic maintenance . PloS one 8 , e63191 ( 2013 ). OpenUrl CrossRef PubMed 39. ↵ Sartori , F. et al. Statistical laws of protein motion in neuronal dendritic trees. en . Cell Rep . 33 , 108391 ( Nov . 2020 ). OpenUrl CrossRef PubMed 40. ↵ Groc , L. & Choquet , D. Linking glutamate receptor movements and synapse function . Science 368 , eaay4631 ( 2020 ). OpenUrl Abstract / FREE Full Text 41. ↵ Opazo , P. & Choquet , D. A three-step model for the synaptic recruitment of AMPA receptors. en . Mol. Cell. Neurosci . 46 , 1 – 8 ( Jan . 2011 ). OpenUrl CrossRef PubMed 42. ↵ Inavalli , V. K. et al. A super-resolution platform for correlative live single-molecule imaging and STED microscopy . Nature methods 16 , 1263 – 1268 ( 2019 ). OpenUrl CrossRef PubMed 43. ↵ Youn , Y. et al. Quantitative DNA-PAINT imaging of AMPA receptors in live neurons . Cell Reports Methods 3 ( 2023 ). 44. ↵ Wang , T. et al. Local zones of endoplasmic reticulum complexity confine cargo in neuronal dendrites . Cell 148 , 309 – 321 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 45. ↵ Santamaria , F. , Wils , S. , De Schutter , E. & Augustine , G. J. Anomalous diffusion in Purkinje cell dendrites caused by spines . Neuron 52 , 635 – 648 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 46. ↵ Tas , R. P. et al. Differentiation between oppositely oriented microtubules controls polarized neuronal transport. en . Neuron 96 , 1264 – 1271 .e5 ( Dec . 2017 ). OpenUrl CrossRef PubMed 47. ↵ Opazo , P. et al. CaMKII triggers the diffusional trapping of surface AMPARs through phosphorylation of stargazin . Neuron 67 , 239 – 252 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 48. ↵ Hafner , A.-S. et al. Lengthening of the stargazin cytoplasmic tail increases synaptic transmission by promoting interaction to deeper domains of PSD-95 . Neuron 86 , 475 – 489 ( 2015 ). OpenUrl CrossRef PubMed 49. ↵ Sheng , M. & Hoogenraad , C. C. The postsynaptic architecture of excitatory synapses: a more quantitative view . Annu. Rev. Biochem . 76 , 823 – 847 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 50. ↵ Kokolaki , M. L. , Fauquier , A. & Renner , M. Molecular crowding and diffusion-capture in synapses . IScience 23 ( 2020 ). 51. ↵ Bissen , D. , Kracht , M. K. , Foss , F. , Hofmann , J. & Acker-Palmer , A. EphrinB2 and GRIP1 stabilize mushroom spines during denervation-induced homeostatic plasticity . Cell Reports 34 , 108923 . ISSN: 2211-1247 . https://www.sciencedirect.com/science/article/pii/S2211124721002370 ( 2021 ). OpenUrl CrossRef PubMed 52. ↵ Passafaro , M. , Piëch , V. & Sheng , M. Subunit-specific temporal and spatial patterns of AMPA receptor exocytosis in hippocampal neurons . Nature neuroscience 4 , 917 – 926 ( 2001 ). OpenUrl CrossRef PubMed Web of Science 53. ↵ Rosendale , M. , Jullié , D. , Choquet , D. & Perrais , D. Spatial and temporal regulation of receptor endocytosis in neuronal dendrites revealed by imaging of single vesicle formation . Cell reports 18 , 1840 – 1847 ( 2017 ). OpenUrl CrossRef PubMed 54. ↵ Yang , Y. , Wang , X.-b. , Frerking , M. & Zhou , Q. Delivery of AMPA receptors to perisynaptic sites precedes the full expression of long-term potentiation . Proceedings of the National Academy of Sciences 105 , 11388 – 11393 ( 2008 ). OpenUrl Abstract / FREE Full Text 55. ↵ Yudowski , G. A. et al. Real-time imaging of discrete exocytic events mediating surface delivery of AMPA receptors . Journal of Neuroscience 27 , 11112 – 11121 ( 2007 ). OpenUrl Abstract / FREE Full Text 56. ↵ Patterson , M. A. , Szatmari , E. M. & Yasuda , R. AMPA receptors are exocytosed in stimulated spines and adjacent dendrites in a Ras-ERK–dependent manner during long-term potentiation . Proceedings of the National Academy of Sciences 107 , 15951 – 15956 ( 2010 ). OpenUrl Abstract / FREE Full Text 57. ↵ Gupta , R. & O’Donnell , C. Dendritic spine neck plasticity controls synaptic expression of long-term potentiation . bioRxiv , 2023 – 01 ( 2023 ). 58. ↵ Getz , A. M. et al. High-resolution imaging and manipulation of endogenous AMPA receptor surface mobility during synaptic plasticity and learning . Science Advances 8 , eabm5298 ( 2022 ). OpenUrl CrossRef PubMed 59. ↵ Plant , K. et al. Transient incorporation of native GluR2-lacking AMPA receptors during hippocampal long-term potentiation . Nature neuroscience 9 , 602 – 604 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 60. ↵ Park , P. et al. The role of calcium-permeable AMPARs in long-term potentiation at principal neurons in the rodent hippocampus . Frontiers in synaptic neuroscience 10 , 42 ( 2018 ). OpenUrl CrossRef PubMed 61. ↵ Tanaka , H. & Hirano , T. Visualization of subunit-specific delivery of glutamate receptors to postsynaptic membrane during hippocampal long-term potentiation . Cell reports 1 , 291 – 298 ( 2012 ). OpenUrl CrossRef PubMed 62. ↵ Henley , J. M. , Barker , E. A. & Glebov , O. O. Routes, destinations and delays: recent advances in AMPA receptor trafficking . Trends in neurosciences 34 , 258 – 268 ( 2011 ). OpenUrl CrossRef PubMed Web of Science 63. ↵ Clavet-Fournier , V. et al. Pre-and postsynaptic nanostructures increase in size and complexity after induction of long-term potentiation . iScience 27 ( 2024 ). 64. ↵ Wagle , S. , Kraynyukova , N. Hafner , A.-S. & Tchumatchenko , T. Computational insights into mRNA and protein dynamics underlying synaptic plasticity rules . Molecular and Cellular Neuroscience, 103846 ( 2023 ). 65. ↵ Tom Dieck , S. et al. Direct visualization of newly synthesized target proteins in situ . Nature methods 12 , 411 – 414 ( 2015 ). OpenUrl CrossRef PubMed 66. ↵ Grooms , S. Y. et al. Activity bidirectionally regulates AMPA receptor mRNA abundance in dendrites of hippocampal neurons . Journal of Neuroscience 26 , 8339 – 8351 ( 2006 ). OpenUrl Abstract / FREE Full Text 67. ↵ Ju , W. et al. Activity-dependent regulation of dendritic synthesis and trafficking of AMPA receptors . Nature neuroscience 7 , 244 – 253 ( 2004 ). OpenUrl CrossRef PubMed Web of Science 68. ↵ Magee , J. C. Dendritic hyperpolarization-activated currents modify the integrative properties of hippocampal CA1 pyramidal neurons . Journal of Neuroscience 18 , 7613 – 7624 ( 1998 ). OpenUrl Abstract / FREE Full Text 69. ↵ Bellotti , A. et al. Paradoxical relationships between active transport and global protein distributions in neurons . Biophysical Journal 120 , 2085 – 2101 ( 2021 ). OpenUrl CrossRef PubMed 70. ↵ Hall , R. & Soderling , T. Quantitation of AMPA receptor surface expression in cultured hippocampal neurons . Neuroscience 78 , 361 – 371 ( 1997 ). OpenUrl CrossRef PubMed Web of Science 71. Rubio , M. E. & Wenthold , R. J. Differential distribution of intracellular glutamate receptors in dendrites . Journal of Neuroscience 19 , 5549 – 5562 ( 1999 ). OpenUrl Abstract / FREE Full Text 72. ↵ Perestenko , P. V. & Henley , J. M. Characterization of the intracellular transport of GluR1 and GluR2 α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor subunits in hippocampal neurons . Journal of Biological Chemistry 278 , 43525 – 43532 ( 2003 ). OpenUrl Abstract / FREE Full Text 73. ↵ Tardin , C. , Cognet , L. , Bats , C. , Lounis , B. & Choquet , D. Direct imaging of lateral movements of AMPA receptors inside synapses . The EMBO journal ( 2003 ). 74. ↵ Heine , M. et al. Surface mobility of postsynaptic AMPARs tunes synaptic transmission . Science 320 , 201 – 205 ( 2008 ). OpenUrl Abstract / FREE Full Text 75. ↵ Migues , P. V. et al. PKMζ maintains memories by regulating GluR2-dependent AMPA receptor trafficking . Nature neuroscience 13 , 630 – 634 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 76. ↵ Westmark , P. R. et al. Pin1 and PKMζ sequentially control dendritic protein synthesis . Science signaling 3 , ra18 – ra18 ( 2010 ). OpenUrl Abstract / FREE Full Text 77. ↵ Hanus , C. et al. Unconventional secretory processing diversifies neuronal ion channel properties . elife 5 , e20609 ( 2016 ). OpenUrl CrossRef PubMed 78. Ashby , M. C. , Maier , S. R. , Nishimune , A. & Henley , J. M. Lateral diffusion drives constitutive exchange of AMPA receptors at dendritic spines and is regulated by spine morphology . Journal of Neuroscience 26 , 7046 – 7055 ( 2006 ). OpenUrl Abstract / FREE Full Text 79. Joensuu , M. et al. Subdiffractional tracking of internalized molecules reveals heterogeneous motion states of synaptic vesicles . Journal of Cell Biology 215 , 277 – 292 ( 2016 ). OpenUrl Abstract / FREE Full Text 80. Nair , D. et al. Super-resolution imaging reveals that AMPA receptors inside synapses are dynamically organized in nanodomains regulated by PSD95 . Journal of Neuroscience 33 , 13204 – 13224 ( 2013 ). OpenUrl Abstract / FREE Full Text 81. Ehlers , M. D. , Heine , M. , Groc , L. Lee , M.-C. & Choquet , D. Diffusional trapping of GluR1 AMPA receptors by input-specific synaptic activity . Neuron 54 , 447 – 460 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 82. Lee , S. H. et al. Super-resolution imaging of synaptic and Extra-synaptic AMPA receptors with different-sized fluorescent probes . Elife 6 , e27744 ( 2017 ). OpenUrl CrossRef PubMed 83. Schindelin , J. et al. Fiji: an open-source platform for biological-image analysis . Nature methods 9 , 676 – 682 ( 2012 ). OpenUrl CrossRef PubMed 84. ↵ Wagle , S. et al. SpyDen: Automating molecular and structural analysis across spines and dendrites . bioRxiv , 2024 – 06 ( 2024 ). 85. ↵ Aakalu , G. , Smith , W. B. , Nguyen , N. , Jiang , C. & Schuman , E. M. Dynamic visualization of local protein synthesis in hippocampal neurons . Neuron 30 , 489 – 502 ( 2001 ). OpenUrl CrossRef PubMed Web of Science 86. ↵ Fujii , S. , Tanaka , H. & Hirano , T. Suppression of AMPA receptor exocytosis contributes to hippocampal LTD . Journal of Neuroscience 38 , 5523 – 5537 ( 2018 ). OpenUrl Abstract / FREE Full Text 87. Triesch , J. , Vo , A. D. & Hafner , A.-S. Competition for synaptic building blocks shapes synaptic plasticity . Elife 7 , e37836 ( 2018 ). OpenUrl CrossRef PubMed 88. ↵ Chirillo , M. A. , Waters , M. S. , Lindsey , L. F. , Bourne , J. N. & Harris , K. M. Local resources of polyribosomes and SER promote synapse enlargement and spine clustering after long-term potentiation in adult rat hippocampus . Scientific reports 9 , 3861 ( 2019 ). OpenUrl CrossRef PubMed 89. ↵ Pirbhoy , P. S. , Farris , S. & Steward , O. Synaptic activation of ribosomal protein S6 phosphorylation occurs locally in activated dendritic domains . Learning & Memory 23 , 255 – 269 ( 2016 ). OpenUrl Abstract / FREE Full Text 90. ↵ Masugi-Tokita , M. et al. Number and density of AMPA receptors in individual synapses in the rat cerebellum as revealed by SDS-digested freeze-fracture replica labeling . Journal of Neuroscience 27 , 2135 – 2144 ( 2007 ). OpenUrl Abstract / FREE Full Text 91. ↵ Antal , M. et al. Numbers, densities, and colocalization of AMPA-and NMDA-type glutamate receptors at individual synapses in the superficial spinal dorsal horn of rats . Journal of Neuroscience 28 , 9692 – 9701 ( 2008 ). OpenUrl Abstract / FREE Full Text 92. ↵ Helm , M. S. et al. A large-scale nanoscopy and biochemistry analysis of postsynaptic dendritic spines . Nature Neuroscience 24 , 1151 – 1162 ( 2021 ). OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted February 08, 2025. Download PDF Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following An integrative model of AMPA receptor trafficking reveals the central contribution of local translation in subtype-specific kinetics Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share An integrative model of AMPA receptor trafficking reveals the central contribution of local translation in subtype-specific kinetics Surbhit Wagle , Maximilian K. Kracht , Anne Bührke , Amparo Acker-Palmer , Nataliya Kraynyukova , Anne-Sophie Hafner , Erin M. Schuman , Tatjana Tchumatchenko bioRxiv 2025.02.08.637220; doi: https://doi.org/10.1101/2025.02.08.637220 Share This Article: Copy Citation Tools An integrative model of AMPA receptor trafficking reveals the central contribution of local translation in subtype-specific kinetics Surbhit Wagle , Maximilian K. Kracht , Anne Bührke , Amparo Acker-Palmer , Nataliya Kraynyukova , Anne-Sophie Hafner , Erin M. Schuman , Tatjana Tchumatchenko bioRxiv 2025.02.08.637220; doi: https://doi.org/10.1101/2025.02.08.637220 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Neuroscience Subject Areas All Articles Animal Behavior and Cognition (7628) Biochemistry (17659) Bioengineering (13881) Bioinformatics (41909) Biophysics (21435) Cancer Biology (18576) Cell Biology (25478) Clinical Trials (138) Developmental Biology (13366) Ecology (19885) Epidemiology (2067) Evolutionary Biology (24299) Genetics (15597) Genomics (22482) Immunology (17725) Microbiology (40356) Molecular Biology (17160) Neuroscience (88523) Paleontology (666) Pathology (2830) Pharmacology and Toxicology (4820) Physiology (7636) Plant Biology (15123) Scientific Communication and Education (2044) Synthetic Biology (4290) Systems Biology (9817) Zoology (2269)