subunit-specific surface mobility of differentially labeled ampa receptor subunits

16
European Journal of Cell Biology 87 (2008) 763–778 Subunit-specific surface mobility of differentially labeled AMPA receptor subunits Michel Kropf a , Guillaume Rey a , Liliane Glauser a , Karina Kulangara a , Kai Johnsson b , Harald Hirling a, a Brain Mind Institute, Faculte´des Sciences de la Vie, Ecole Polytechnique Fe´de´rale de Lausanne EPFL, CH-1015 Lausanne, Switzerland b Faculte´des Sciences de Base, Ecole Polytechnique Fe´de´rale de Lausanne, CH-1015 Lausanne, Switzerland Received 20 December 2007; received in revised form 27 February 2008; accepted 28 February 2008 Abstract Lateral mobility of AMPA-type glutamate receptors as well as their trafficking between plasma membrane and intracellular compartments are major mechanisms for the regulation of synaptic plasticity. Here we applied a recently established labeling technique in combination with lentiviral expression in hippocampal neurons to label individual ACP-tagged AMPA receptor subunits specifically at the surface of neurons. We show that this technique allows the differential labeling of two receptor subunits on the same cell. Moreover, these subunits are integrated into heteromeric receptors together with endogenous subunits, and these labeled receptors are targeted to active synapses. Sequential labeling experiments indicate that there is basal surface insertion of GluR1, GluR2 and GluR3, and that this insertion is strongly increased following potassium depolarization. Moreover, we found that ACP-labeled GluR3 shows the highest surface mobility among GluR1, GluR2, and GluR3. In double-infected neurons the diffusion coefficient of labeled GluR2 at the surface of living neurons is significantly higher in GluR2/GluR3-infected neurons compared to GluR1/GluR2-infected neurons suggesting a higher mobility of GluR2/3 receptors compared to GluR1/2 receptors. These results indicate that surface mobility is regulated by different subunit compositions of AMPA receptors. r 2008 Elsevier GmbH. All rights reserved. Keywords: Trafficking; Lateral movement; Endocytosis; Receptor trafficking Introduction Fast synaptic transmission at excitatory glutamatergic synapses is principally mediated by AMPA-type gluta- mate receptors (AMPAR). AMPAR rapidly cycle between internal compartments and the neuronal sur- face by exocytosis/endocytosis. In addition, there is shuttling of receptors between extrasynaptic plasma membrane sites and the post-synaptic density (PSD). These two dynamic events are believed to be major components that determine efficacy of synaptic trans- mission. During synaptic potentiation additional recep- tors are recruited to the synapse, while upon synaptic depression receptors are removed from the PSD (Beattie et al., 2000; Lu et al., 2001; Luscher et al., 1999; Man et al., 2000; Matsuda et al., 2000; Shi et al., 2001; Xia et al., 2000). AMPAR are heterotetramers com- posed of the four subunits GluR1-4 (or GluR-A-D), ARTICLE IN PRESS www.elsevier.de/ejcb 0171-9335/$ - see front matter r 2008 Elsevier GmbH. All rights reserved. doi:10.1016/j.ejcb.2008.02.014 Corresponding author. Tel.: +41 21 693 5363; fax: +41 21 693 5350. E-mail address: harald.hirling@epfl.ch (H. Hirling).

Upload: michel-kropf

Post on 04-Sep-2016

214 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESS

European Journal of Cell Biology 87 (2008) 763–778

0171-9335/$ - se

doi:10.1016/j.ej

�Correspondfax: +4121 693

E-mail addr

www.elsevier.de/ejcb

Subunit-specific surface mobility of differentially labeled

AMPA receptor subunits

Michel Kropfa, Guillaume Reya, Liliane Glausera, Karina Kulangaraa,Kai Johnssonb, Harald Hirlinga,�

aBrain Mind Institute, Faculte des Sciences de la Vie, Ecole Polytechnique Federale de Lausanne EPFL,

CH-1015 Lausanne, SwitzerlandbFaculte des Sciences de Base, Ecole Polytechnique Federale de Lausanne, CH-1015 Lausanne, Switzerland

Received 20 December 2007; received in revised form 27 February 2008; accepted 28 February 2008

Abstract

Lateral mobility of AMPA-type glutamate receptors as well as their trafficking between plasma membrane andintracellular compartments are major mechanisms for the regulation of synaptic plasticity. Here we applied a recentlyestablished labeling technique in combination with lentiviral expression in hippocampal neurons to label individualACP-tagged AMPA receptor subunits specifically at the surface of neurons. We show that this technique allows thedifferential labeling of two receptor subunits on the same cell. Moreover, these subunits are integrated into heteromericreceptors together with endogenous subunits, and these labeled receptors are targeted to active synapses. Sequentiallabeling experiments indicate that there is basal surface insertion of GluR1, GluR2 and GluR3, and that this insertionis strongly increased following potassium depolarization. Moreover, we found that ACP-labeled GluR3 shows thehighest surface mobility among GluR1, GluR2, and GluR3. In double-infected neurons the diffusion coefficient oflabeled GluR2 at the surface of living neurons is significantly higher in GluR2/GluR3-infected neurons compared toGluR1/GluR2-infected neurons suggesting a higher mobility of GluR2/3 receptors compared to GluR1/2 receptors.These results indicate that surface mobility is regulated by different subunit compositions of AMPA receptors.r 2008 Elsevier GmbH. All rights reserved.

Keywords: Trafficking; Lateral movement; Endocytosis; Receptor trafficking

Introduction

Fast synaptic transmission at excitatory glutamatergicsynapses is principally mediated by AMPA-type gluta-mate receptors (AMPAR). AMPAR rapidly cyclebetween internal compartments and the neuronal sur-face by exocytosis/endocytosis. In addition, there is

e front matter r 2008 Elsevier GmbH. All rights reserved.

cb.2008.02.014

ing author. Tel.: +4121 693 5363;

5350.

ess: [email protected] (H. Hirling).

shuttling of receptors between extrasynaptic plasmamembrane sites and the post-synaptic density (PSD).These two dynamic events are believed to be majorcomponents that determine efficacy of synaptic trans-mission. During synaptic potentiation additional recep-tors are recruited to the synapse, while upon synapticdepression receptors are removed from the PSD (Beattieet al., 2000; Lu et al., 2001; Luscher et al., 1999;Man et al., 2000; Matsuda et al., 2000; Shi et al., 2001;Xia et al., 2000). AMPAR are heterotetramers com-posed of the four subunits GluR1-4 (or GluR-A-D),

Page 2: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778764

with GluR1/2 and GluR2/3 as the major forms in thehippocampus (Wenthold et al., 1996). These differentsubunit compositions show differential trafficking be-havior. During long-term potentiation (LTP) GluR1/2receptors have been proposed to be inserted intosynapses in an activity-dependent fashion via calcium-calmodulin-dependent kinase II-dependent phosphory-lation (CaMKII) (Hayashi et al., 2000; Shi et al., 1999).This regulation depends on the GluR1 subunit (Passa-faro et al., 2001; Shi et al., 2001). In contrast to LTP,induction of long-term depression (LTD) acts on GluR2in order to internalize receptors by endocytosis intoendosomes (Lee et al., 2004). GluR2/3 receptors cyclealso in an activity-independent manner between theneuronal surface and endosomal compartments (Passa-faro et al., 2001; Shi et al., 2001). These endosomescomprise important sorting stations between endosomalrecycling and degradation pathways (Ehlers, 2000;Kulangara et al., 2007; Lee et al., 2004; Park et al.,2004; Steiner et al., 2005).

Although a recent study on the implication of theexocyst complex in AMPAR recruitment showed inser-tion of AMPAR directly at the PSD (Gerges et al.,2006), physiological recordings using a photoactivatableAMPAR antagonist suggested fast insertion of receptorsat distant sites of synapses (Adesnik et al., 2005). Inaccordance with the latter possibility, GluR2 taggedwith the pH-sensitive variant of GFP is inserted alongthe dendritic shaft (Ashby et al., 2006; Yudowskiet al., 2007). Such an insertion distant to the synapserequires lateral movement along the plasma membrane.

Fig. 1. Labeling of surface AMPAR subunits with the ACP techniqu

acyl carrier protein (ACP) tag is fused to the extracellular aminoterm

The coenzyme A (CoA) derivative is covalently linked to serine 3

(b) Principle of sequential two-color labeling with two carrier prote

step the PPTase AcpS links CoA to the ACP tag, but not to the

specifically to the ACP9 tag, because ACP was saturated in the firs

Shuttling of antibody-labeled surface GluR2 in and outof synapses has been shown in cultured hippocampalneurons (Bats et al., 2007; Borgdorff and Choquet,2002), but little information is available on the subunit-specific characteristics of receptor movement at thesurface.

In the present study we analyzed surface mobility ofdifferent AMPAR subunits by a recently developedtechnique that allows to specifically label a given surfacereceptor with synthetic probes. The technique involvesexpressing the surface protein of interest as a fusionprotein with an acyl carrier protein (ACP; 77 residues)or a mutant thereof (ACP9) at its extracellular domain(George et al., 2004; Sielaff et al., 2006). These tags canthen be specifically derivatized with synthetic probesusing the bacterial phosphopantetheine transferases(PPTases) AcpS (which labels only ACP) or Sfp (whichlabels both ACP and ACP9), and synthetic CoAderivatives (Fig. 1). As the PPTases and the CoAderivatives are membrane impermeable, the approachallows to label only the cell-surface subpopulation of agiven receptor, whereas internal receptors are notmodified. In addition, two receptors or subunits canbe labeled simultaneously, and the tags are of muchsmaller size than antibody conjugates.

Here we single- or double-infected hippocampalneurons with lentiviral vectors expressing ACP- orACP9-tagged AMPAR subunits. We found that thissystem allows for differential labeling of two subunits ofGluR1/2- or GluR2/3-infected cells. These exogenouslyexpressed GluRs form heteromeric receptors with

e. (a) Scheme of the labeling reaction. The 77 amino acid long

inus of the GluR subunit (downstream of the signal peptide).

6 of the ACP by a phosphopantethein transferase (PPTase).

in tags, ACP and its mutant ACP9, on the same cell. In a first

ACP9 tag. In a second step another PPTase, Sfp, links CoA

t step by AcpS.

Page 3: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778 765

exogenous as well as endogenous subunits, which areintegrated into active synapses. In sequential labelingexperiments we detected newly inserted GluR1 andGluR2, which was increased in both cases by potassiumdepolarization, suggesting basal and stimulus-inducedcycling of these subunits. Interestingly, surface mobilityof ACP-labeled GluR2/3 receptors is significantly higherthan that of GluR1/2 receptors.

Materials and methods

Primary neuron culture

Hippocampal neurons were prepared from P0 Spra-gue-Dawley rats (Charles River/Iffa Credo, L’ArbresleCedex, France) as previously described (Steiner et al.,2002). The hippocampal neurons were plated in growthmedium (Earle’s MEM medium, 20mM glucose,0.5mM glutamine, 100U/ml penicillin, 100 mg/ml strep-tomycin, 10% horse serum) either on 35-mm plasticdishes (Falcon, BD Biosciences, Bedford, USA) coatedwith poly-D-lysine (5 mg/cm2) (BD Biosciences, FranklinLakes, USA) and laminin 0.7 mg/cm2 (Invitrogen,Carlsbad, USA) at a density of 23,000 cells/cm2 forbiochemical experiments, or on poly-D-lysine/laminin-coated borosilicate glass coverslips (Marienfeld GmbH& Co. KG, Lauda-Konigshofen, Germany) at a densityof 12,500 cells/cm2 for immunocytochemistry and live-cell imaging. After 2 h the growth medium was changedto Neurobasal/B27 medium (Invitrogen) containing0.5mM glutamine, 100U/ml penicillin, 100 mg/ml strep-tomycin. The neurons were kept in culture at 37 1C and5% CO2. Three days after plating, the desired concen-tration of lentivirus for infection was added to thecultures.

Antibodies

Antibodies against the following proteins wereapplied for Western blotting (W) or immunofluores-cence (IF): monoclonal antibodies: GluR2 (MAB397;W, 1:200; IF, 1:100; BD Biosciences, Franklin Lakes,USA), GluR3 (3B3; W, 1:500, Zymed; San Francisco,USA), HA (W, 1:1000; Covence, Berkeley, USA), PSD-95 (IF, 1:200; Affinity BioReagent, Golden, USA),Synaptophysin (IF, 1:2000; Synaptic Systems, Gottin-gen, Germany); polyclonal antibodies: GluR1 (W,1:300; Chemicon, Temecula, USA), GluR2/3 (W,1:1000; Chemicon). For Western blots we used second-ary Alexa Fluors 680 nm-coupled goat anti-rabbit IgGand secondary Alexa Fluors 680 nm-coupled goat anti-mouse IgG (W: 1:2000; Molecular Probes Inc., Eugene,USA) or IRDyeTM800 conjugated goat anti-rabbit IgGor IRDyeTM800 conjugated goat anti-mouse IgG (W,

1:2000; Rockland, Gilbertsville, USA). For immunocy-tochemistry we used Cy3- or Cy5-coupled (IF 1:200;Jackson ImmunoResearch, West Grove, USA) andAlexa Fluors 405- or 488-coupled (IF 1:200; MolecularProbes Inc.) secondary anti-mouse and anti-rabbit IgG.

DNA constructs

For the expression of ACP- or ACP9-tagged GluRsubunits, the cDNAs encoding ACP or ACP9 (Georgeet al., 2004; Sielaff et al., 2006) were amplified by PCRand subcloned in-frame into an XbaI site locatedbetween the sequence encoding the signal peptide andan HA-tag in the cDNAs of rat GluR1, GluR2 orGluR3 which had been cloned in the vector GW1(kindly provided by Dr. M. Passafaro, Milano, Italy).To transfer the cDNAs into lentiviral expressionvectors, the ACP- or ACP9-tagged cDNAs wereamplified by PCR and first cloned into the vectorpENTR/D-TOPO using the pENTR Directional TOPOCloning Kit (Invitrogen). Procedures were carried outaccording to the manufacturer’s protocol. Transfer intothe lentivector SIN-PGK-CassRFA-WHV (kindly pro-vided by Dr. R. Luthi-Carter, Lausanne, Switzerland)was performed with the Gateway LR Clonase IIEnzyme Mix (Invitrogen) according to the manufac-turer’s protocol.

Lentivirus production

For the lentiviral production, HEK293T cells, cul-tured in DMEM containing 10% fetal calf serum and1% penicillin-streptomycin, were transfected by thecalcium-phosphate technique with the following plas-mids (quantities of DNA per 9-cm dish): pCMVD8.92(13 mg), pRSV-ReV (3.75 mg), pMD2G (3 mg), and SIN-PGK-WHV vector containing the transgene (13 mg).Seventy-two hours after transfection the medium wascollected and spun in a SW32Ti ultracentrifuge rotor at19,000 rpm and 4 1C for 90min. The pellet wasresuspended in a final volume of 3ml of phosphate-buffered saline (PBS), 0.5% bovine serum albumin for a50� concentrated virus stock. The viral titer wasdetermined using the HIV-1 P24 ELISA kit (Zeptome-trix Corporation, Buffalo, USA).

Immunocytochemistry

Neurons on glass coverslips were washed in PBS andfixed in 4% (w/v) paraformaldehyde/4% (w/v) sucroseduring 12min. For PSD-95 staining, cells were fixed20min with methanol at �20 1C. Immunolabeling wascarried out as described (Kulangara et al., 2007).Coverslips were mounted in 50% glycerol containing

Page 4: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778766

Mowiol (Fluka) and DABCO (Sigma-Aldrich) to retardphotobleaching.

ACP labeling

For the labeling of the ACP-tagged AMPARsubunits, we used CoA-Cy3, CoA-Alexa488, CoA-Cy5and CoA-biotin (George et al., 2004) or CoenzymeA disodium (Fluka). Recombinant AcpS (George et al.,2004) and Sfp (Sielaff et al., 2006; Stachelhaus et al.,1998) were expressed as described, using the Escherichia

coli strain BL21 and induction of expression by IPTGfor 3.5 h at 24 1C. Neurons were washed with PBS andthen incubated 15min at room temperature in thereaction mixture (5 mM CoA-derivative, 1 mM AcpS orSfp, 10mM MgCl2 in PBS). Cells were then washedthree times in PBS. For ACP-ACP9 dual-labeling,sequential reactions were performed, starting with anAcpS reaction for 15min, followed by a PBS wash, andthen by another Sfp reaction for 15min. Cells wereeither fixed, or kept in the incubator or under amicroscope for live-cell imaging. To determine newlyinserted receptors, sequential AcpS-labelings were car-ried out (see Fig. 5(a)). For unstimulated conditions allincubation steps included TTX (2 mM) to block sponta-neous stimulation. To verify for receptor internalization,ACP-GluR2-infected neurons were labeled with Cy3-CoA, followed by incubation for 30min at 37 1C with amouse anti-GluR2 antibody in conditioned medium,which binds to an extracellular epitope. After washing,cells were stimulated with 100 mM AMPA for 2min,followed by further incubation for 13min in conditionedmedium. Neurons were rinsed with PBS/glycine (pH 2.5)to remove residual surface antibody, fixed, and im-munolabeled with an Alexa488-conjugated secondaryantibody.

FM1-43 uptake

Neurons infected with SIN-PGK-ACP constructswere stained at DIV18-20 with FM1-43FX (MolecularProbes Inc.), which is a fixable analog of FM1-43.Coverslips containing cultured hippocampal neuronswere first washed for 10min with K5 buffer (128mMNaCl, 5mM KCl, 2.7mM CaCl2, 10mM glucose,20mM HEPES, 1mM MgCl2, pH 7.4). Then K5 bufferwas removed, and 10 mM FM1-43 FX in K80 buffer(53mM NaCl, 80mM KCl, 2.7mM CaCl2, 10mMglucose, 20mM HEPES, 1mM MgCl2, pH 7.4) wasadded for 60 s. Cells were washed several times with K5buffer and then the ACP-tagged proteins were labeledwith a CoA-Cy5 reaction as described above. Neuronswere washed with PBS, fixed and mounted in Mowiol.

Surface biotinylation of ACP-tagged subunits

Three 3.5-cm dishes of neuronal cultures were infectedwith viruses expressing ACP-tagged AMPAR subunits.At DIV18-20 neurons were washed with PBS and thenbiotinylated with a reaction containing 5 mM CoA-biotin (procedure described above). Cells were thenwashed once with PBS and once with buffer B (20mMHEPES, pH 7.4, 2mM EDTA, 2mM EGTA, 0.1mMDTT, 0.1M KCl), and lysed in 200 ml lysis buffer(20mM HEPES, pH 7.4, 2mM EDTA, 2mM EGTA,0.1mM DTT, 0.1M KCl, 1% Triton X-100) containing0.3mM PMSF, 0.7 mg/ml pepstatin, 2 mg/ml aprotinin,2 mg/ml leupeptin, and incubated 5min on ice. Lysatewas then spun 10min at 16,000g and 4 1C. The super-natant was incubated 1 h at 4 1C with streptavidin-Sepharose beads (GE Healthcare, Uppsala, Sweden).Then the beads were washed three times in buffer B(20mM HEPES, pH 7.4, 2mM EDTA, 2mM EGTA,0.1mM DTT, 0.1M KCl) containing 0.5% Triton X-100. Finally, sample-buffer (1M Tris–HCl, pH 6.8, 10%glycerol, 2% SDS, 5% b-mercaptoethanol, 0.2 g/mlbromophenol blue) was added to the beads for separa-tion by SDS–PAGE. Proteins were transferred ontoProtran BA83 nitrocellulose membranes (Schleicher &Schuell, Dassel, Germany), blocked in Odyssey blockingsolution diluted in PBS according to the manufacturer’sprotocol (Li-CorTM Biosciences, Lincoln, USA). Themembranes were probed with the indicated primaryantibodies and Alexa Fluors-coupled secondary anti-body. Fluorescent signals on blots were visualized anddirectly quantified by the CCD camera of the Odysseysystem. For receptor insertion experiments, neuronswere first labeled by a reaction with 20 mM unconjugatedCoA, then incubated for the indicated periods of time,and finally biotinylated using biotin-CoA in a secondlabeling reaction.

Quantitative image analysis

For the quantification of the new insertion of GluR1 orGluR2 (CoA-Cy3, 570 nm), all confocal images weretaken on a Leica TCS-SP2 AOBS microscope (LeicaMicrosystems, Wetzlar, Germany) with the same param-eters for the HeNe laser channel. The neuron wasoutlined using a graphic tablet. To quantify thefluorescence, a threshold of 80 was applied, and theaverage fluorescence intensity was measured usingMetamorph software (Molecular Device Corporation,Sunnyvale, USA). The intensities given for the differentincubation times (15, 30, 45, or 60min), either with K5/TTX or with K80, were normalized by subtracting thebackground value at time 0min. The following numberof cells n were analyzed in Figs. 6 and 7: for GluR1:30min K80, 25; 30min K5/TTX, 30; 60min K80, 30;

Page 5: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778 767

60min K5/TTX, 23; for GluR2: 15min K80, 26; 15minK5/TTX, 43; 30min K80, 18; 30min K5/TTX, 22; 45minK80, 37; 45min K5/TTX, 40; 60min K80, 41; 30min K5/TTX, 19.

Diffusion analysis

Image sequences were acquired with a Spinning-DiskMicroscope UltraView ERS (Perkin Elmer, Waltham,USA) at 30Hz, and were first filtered with a KalmanStack Filter (plug-in for ImageJ Software). The first6–10 images from the filtered stack of images wereremoved because the Kalman Filter is an iterative filter.Fluorescent particles were selected manually and theirtrajectories were reconstructed with Metamorph soft-ware, which generates a text file with the coordinates ofthe particles in function of the time. Data processing forthe calculation of the mean squared distance (MSD) anddiffusion coefficient D was implemented on Matlabsoftware (MathWorks, Natick, USA). MSD is definedas the average on several trajectories of the squareddistance of a particle:

MSDðtÞ ¼ h½rðtÞ � rð0Þ�2iðtÞ ¼ 4Dt (1)

Most of our particles were confined at synaptic sites. Wethus applied the model for a confined Brownian particle(Kusumi et al., 1993):

MSDðtÞ ¼L2

31� exp

�12Dt

L2

� �� �(2)

This equation tends toward a constant value when thetime goes to infinity, and is approximately linear forsmall t. Assuming that the process is stationary, we usedthe ergodic theorem (Qian et al., 1991):

MSDðtÞ ¼ h½rðtÞ � rð0Þ�2iðtÞ ¼

Z½rðtþ sÞ � rðsÞ�2 ds (3)

With only a finite number of frames, we applied thisdiscrete form of Eq. (3):

MSDðtnÞ ¼MSDðndtÞ ¼1

N � n

XN�n

j¼1

½xððnþ jÞdtÞ

� xðjdtÞ�2 þ ½yððnþ jÞdtÞ � yðjdtÞ�2 (4)

The linear term of a polynomial least square fit of theMSD plot was used to calculate D values. TetraSpeckTM

fluorescent beads of 0.2 mm (Molecular Probes Inc.)immobilized in mounting medium were used as acontrol. Analyses were done on three independentexperiments. The number of cells and particles analyzedwere: ACP-GluR1 (5/98); ACP-GluR2 (8/198); ACP-GluR3 (5/87); GluR1/2 (14/317) and GluR2/3 (11/392);for beads 125 particles were analyzed.

Results

ACP-labeling technique to differentially label

AMPAR subunits at the neuronal surface

Trafficking characteristics of AMPAR are determinedby their subunit composition (Shi et al., 2001). ACP-tagging of surface receptors allows for simultaneouslabeling of different polypeptide chains with syntheticprobes using PPTases and synthetic CoA derivatives(Fig. 1(a)) (George et al., 2004; Johnsson et al., 2005;Meyer et al., 2006; Prummer et al., 2006; Vivero-Polet al., 2005). We applied here this approach to GluR1,GluR2 and GluR3. The ACP tag (77 residues) can belabeled by the PPTases AcpS and Sfp, and the ACP9 tag(a mutant of ACP) only by Sfp (Sielaff et al., 2006).These enzymes can therefore link CoA-derivatives toS36 in ACP or ACP9 (Fig. 1). Here we used thefluorophores Alexa488-, Cy3- and Cy5-conjugated CoAas well as biotin-conjugated CoA to follow the surfaceinsertion and mobility of ACP- or ACP9-tagged GluR1,GluR2 and GluR3 on single- or double-infected neuroncultures (Fig. 1(b)).

In order to verify that the ACP-tagged AMPARsubunits localize to synaptic sites, we individuallyinfected primary hippocampal neurons at DIV3 withlentiviruses expressing ACP- or ACP9-tagged GluR1,GluR2, or GluR3. At DIV20 we performed ACP-labeling using the AcpS or Sfp enzymes, respect-ively, followed by immunolabeling for synaptophysin(Fig. 2(a); presynaptic marker) or PSD-95 (Fig. 2(b);post-synaptic marker). All three subunits showed a clearcolocalization with PSD-95 and close juxtaposition withsynaptophysin. This is particularly visible in the imagesof the enlarged areas in Fig. 2 (arrows), indicatingcorrect targeting to synaptic sites.

Next we verified whether ACP-tagged AMPARsubunits are internalized as is known for endogenousAMPAR. To this end we performed a labeling reactionby Cy3-CoA on neurons infected to express ACP-GluR2, followed by incubation with an antibody againstan extracellular domain of GluR2 to label all exogenousand endogenous surface AMPAR (antibody feeding).After stimulation with AMPA (2min) and furtherincubation (13min), acid-wash, fixation, and stainingwith secondary antibody, we identified a fraction ofACP-tagged receptors in intracellular compartments(Fig. 2(c)). This suggests that ACP-tagged GluR2 isfunctional with respect to receptor trafficking.

We then asked whether synapses containing theseexogenous subunits are responsive to stimulation.To this end we infected neurons with lentivirusesencoding ACP-GluR1, ACP-GluR2, or ACP9-GluR3,and incubated them with AcpS or Sfp and syntheticCoA derivatives. This was followed by depolarizationwith buffer containing 80mM potassium in the presence

Page 6: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESS

Fig. 2. Synaptic localization of ACP-tagged GluR1, GluR2, or GluR3 in single-infected neurons. Hippocampal neurons were

infected at DIV3 with either ACP-GluR1, ACP-GluR2, or ACP9-GluR3. At DIV20 hippocampal neurons were labeled with CoA-

Cy3, fixed and then (a) immunolabeled with an antibody against synaptophysin to label presynaptic boutons, or (b) with an

antibody against PSD-95 to label postsynaptic densities. There is a clear colocalization or juxtapositioning between the synaptic

markers and the different GluR subunits. This is particularly evident in the enlarged areas. (c) Evoked internalization of ACP-

tagged GluR2. Neurons infected to express ACP-GluR2 were labeled with CoA-Cy3 and then incubated with extracellularly binding

anti-GluR2 antibody (GluR2 Ab). After AMPA stimulation (100 mM, 2min) and further incubation in conditioned medium

(13min), neurons were acid-washed to remove remaining surface antibody, fixed and stained with a secondary antibody. Bar: 10 and

2 mm.

M. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778768

of the styryl dye FM1-43FX to load presynaptic sites ofactive synapses (Betz et al., 1992) (Fig. 3). Indeed,ACP- or ACP9-GluR-positive sites are mostly positivefor FM1-43FX uptake (Fig. 3; see enlarged areas).This shows that synapses containing ACP-taggedAMPAR can actively take up FM1-43FX in livingneurons.

Dual-infected ACP-tagged GluRs form heteromeric

receptors with exogenous and endogenous subunits

The different specificities of AcpS and Sfp allow forselective labeling of ACP- and ACP9-tagged AMPARwith different fluorophores. To analyze whether suchtwo exogenous AMPAR subunits form heteromeric

Page 7: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESS

Fig. 3. ACP-tagged AMPAR subunits are targeted to active synapses. Hippocampal neurons were infected at DIV3 with

lentiviruses encoding ACP-GluR1, ACP-GluR2 or ACP9-GluR3. At DIV20, neurons were labeled with CoA-Cy5 and then loaded

with FM1-43 by stimulation with buffer containing 80mM potassium for 1min. ACP-tagged subunits are juxtapositioned to FM1-

43-loaded presynaptic boutons. Bars: 10 and 2mm.

M. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778 769

receptors like naturally occurring AMPAR, we infectedhippocampal neurons with either ACP-GluR1 andACP9-GluR2 (Fig. 4(a)), or ACP-GluR2 and ACP9-GluR3 (Fig. 4(b)). These two combinations representthe major subunit composition in the hippocampus(Wenthold et al., 1996). We then performed sequentiallabeling with AcpS and CoA-Cy3, and with Sfp andCoA-Alexa488, resulting in Cy3-labeled ACP-taggedsubunit and Alexa488-labeled ACP9-tagged subunit.Preliminary experiments verified that the first reactionby AcpS saturated the ACP tag ensuring that the secondreaction by Sfp specifically labels the ACP9 tag. Thelabelings were followed by fixation and immunolabelingfor synaptophysin or PSD-95. With both subunitcombinations there was a strong colocalization betweenthe two subunits (Fig. 4(a) and (b), dual overlays inforth rows). This suggests that the two subunits aretargeted to the same synapses. Moreover, these colocal-izing AMPAR subunits localized to PSD-95-positivesynapses (Fig. 4(a) and (b), right column, triple overlayin fifth rows), and are in the majority of cases juxta-positioned to synaptophysin-positive presynaptic bou-tons (Fig. 4(a) and (b), left column, triple overlays infifth rows). These data indicates that synapses containedboth exogenously expressed AMPAR subunits.

Naturally occurring AMPAR are hetero-tetramericcomplexes, with the combinations GluR1/2 or GluR2/3

predominant in the hippocampus. Therefore, we ana-lyzed whether in these dual-infected neurons the ACP/ACP9-tagged subunits form homomeric receptors, orwhether they can rather be integrated into heteromericreceptors. To this end, we did ACP-biotinylationexperiments on neurons double-infected with ACP-GluR2/ACP9-GluR1 (Fig. 4c1), or ACP-GluR3/ACP9-GluR1 (Fig. 4c2), or ACP-GluR3/ACP9-GluR2(Fig. 4c3). We carried out an AcpS-labeling reactionusing biotin-conjugated CoA, which is specificallylinked to the ACP-tagged subunit. Following streptavi-din-bead precipitation (Fig. 4(c), elution) from neuronextracts, we found that in each case the ACP9-coinfected subunit was coprecipitated with the ACP-tagged subunit (Fig. 4(c), arrows). Interestingly, wecould also detect the lower migrating bands correspond-ing to the coprecipitating endogenous subunits(Fig. 4(c), arrow heads). This strongly suggests thatnot only did the two exogenously expressed,tagged subunits form heteromeric receptors witheach other, but that these tagged subunits integratedinto the normal, endogenous pool of AMPAR formingheteromeric receptors with endogenous subunits. As canbe judged from the intensities of the bands, theexpression level of the tagged subunits was inferior tothe endogenous expression level at the applied virustiter.

Page 8: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESS

Fig. 4. ACP-tagged AMPAR subunits form heteromeric receptors in double-infected hippocampal neurons. (a, b) Hippocampal

neurons were infected at DIV3 with lentiviruses encoding either (a) ACP-GluR1 and ACP9-GluR2 or (b) ACP-GluR2 and ACP9-

GluR3. At DIV20, neurons were first labeled with CoA-Cy3 by AcpS (for ACP tags), and subsequently with CoA-Alexa488 by Sfp

(for ACP9 tags). Neurons were fixed and immunolabeled for synaptophysin (left columns in a and b) or PSD-95 (right columns in a

and b). Colocalization between double-infected subunits and synaptic markers is indicated by arrowheads. Bars: 2 mm. (c)

Hippocampal neurons were double-infected at DIV3 for (c1) ACP-GluR2 and ACP9-GluR1, or GFP and ACP9-GluR1 as control,

(c2) ACP-GluR3 and ACP9-GluR1, or GFP and ACP9-GluR1, (c3) ACP-GluR3 and ACP9-GluR2, or GFP and ACP9-GluR2. At

DIV20 an AcpS reaction with CoA-biotin was performed which in each case labels the ACP-tagged subunit. Western blots after

streptavidin-precipitations show that the coinfected subunit (arrows) as well as endogenous subunits (arrowheads) are

coprecipitated.

Fig. 5. Depolarization increases the surface insertion rate of GluR2-containing receptors. (a) Scheme of the pulse-chase experiment.

Newly inserted receptors are determined by a CoA-Cy3 labeling reaction which was preceded by CoA-488- and unconjugated CoA-

labeling reactions to block initially present receptors. (b) Hippocampal neurons, infected at DIV3 with ACP-GluR2, were labeled at

DIV20 with CoA-Alexa488 (images 1, 4, 7), then further blocked with unconjugated CoA. Neurons were then either immediately

labeled with CoA-Cy3 (0min, image 2) to check for complete blocking, or incubated before the CoA-Cy3 reaction in the following

manner: either for 2min in K5 (buffer containing 5mM potassium) and 28min in K5 (non-stimulated; image 5), or for 2min in K80

(buffer containing 80mM potassium) and 28min in K5 (stimulated; image 8). Cells were fixed, and immunolabeled for the

presynaptic marker synaptophysin (images 3, 6, 9). Potassium depolarization increases labeling for newly inserted receptors. Inserts

are zooms of the boxed region showing clusters of surface receptors colocalizing with synaptophysin. All the reactions and

incubations of the K5 condition were done in presence of TTX (2 mM) to avoid varying stimulations due to spontaneous activity.

(c) Quantification of the Cy3 fluorescent signals from experiments illustrated in (a), with the indicated times of incubation. Average

pixel intensity of fluorescence was normalized on the background control fluorescence intensity at 0min, and the value of 0 on the

y-axes corresponds to average background level (*po0.05; **po0.01; for n please refer to the Materials and methods section).

(d) ACP-GluR2-infected neurons at DIV20 were either directly labeled with CoA-biotin (lane 1), or blocked with unconjugated CoA

and then immediately labeled with CoA-biotin (lane 2), or blocked with unconjugated CoA followed by incubation with K80 for

2min and with K5 for 28min (lane 3; 300) or 58min (lane 4; 600), or blocked with unconjugated CoA and incubated with K5/TTX

for 30min (lane 5; 300) or 60min (lane 6; 600). Finally, after labeling with CoA-biotin, biotinylated receptors were determined by

streptavidin-precipitation and Western blot from cell extracts. (e) Quantification of signals of precipitated receptors. The newly

inserted ACP-GluR2 are normalized to their amount in cell extract (*po0.05; n, 4).

M. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778770

Page 9: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778 771

Page 10: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778772

Detection of newly inserted AMPAR subunits at the

surface of hippocampal neurons

In order to characterize further the trafficking ofdifferent ACP-labeled AMPAR subunits, we analyzedthe turnover of surface receptors by a pulse-chasesequential labeling of ACP-tagged AMPAR at differenttime-points with different fluorophores (Fig. 5(a)).Receptors at time zero were prelabeled by a first AcpSreaction with Alexa488-conjugated CoA (CoA-488),and a further blocking reaction using unconjugatedCoA to ensure complete saturation of surface receptorsat time zero. The neurons were then incubated fordifferent times, and the newly inserted receptors weredetermined by a second labeling reaction with CoA-Cy3.We either kept the neurons unstimulated (K5; 5mMpotassium buffer containing TTX to suppress sponta-neous activity), or depolarized them for 2min with K80(80mM potassium buffer). Labeling at time zero withAlexa488 yielded the regular surface pattern of ACP-GluR2 (Fig. 5(b), images 1, 4, 7). When the Cy3 labelingwas performed immediately after the Alexa488 labeling,no Cy3-labeled receptor could be detected, verifying fora complete labeling of receptors at time zero (Fig. 5(b),image 2). When neurons were incubated for 30minbetween the first and second labeling, little reinsertionwas observed under control conditions (Fig. 5(b); K5/TTX; image 5). In contrast, when neurons had beenbriefly depolarized, we detected newly appeared ACP-GluR2 (Fig. 5(b); image 8). A time-course of differentincubation periods indicated that this appearanceoccurred mainly between 15min and 30min (Fig. 5(c)).Subsequent immunolabeling for synaptophysin indi-cated that these new receptors become integrated intosynapses (Fig. 5(b); images 3, 6, 9; see also insert).

To substantiate these results, we performed thissequential labeling protocol using biotin-conjugatedCoA instead of fluorescent CoA. Neurons, either brieflystimulated with K80 (Fig. 5(b), lane 3–4), or kept in K5/TTX (Fig. 5(d), lane 5–6), as above, were first blockedwith simple unconjugated CoA, followed by no incuba-

Fig. 6. GluR1 is inserted at the neuronal surface at a higher rate tha

DIV3 to express ACP-GluR1 or ACP-GluR3. At DIV20, surface

unconjugated CoA to ensure complete saturation. The negative cont

Cy3. The other cells were treated for 2min with K5 (unstimulated) o

in K5. Then neurons were AcpS-labeled with CoA-Cy3 and fixed. A

in presence of TTX (2 mM) to avoid varying stimulations due to spon

were quantified on confocal images. Fluorescent signal was norm

Relative insertion of GluR1 after 30min incubation. ACP-GluR1-in

with Cy3-CoA (0min), or blocked by AcpS reactions using Alexa488

K80, and incubation for 28min in conditioned medium, and finally C

CoA-Cy3-labeling of GluR1 (upper row), or after preblocking, stim

containing receptors after K80 depolarization and 30min incuba

receptors. n ¼ 35 (0min); n ¼ 48 (30min).

tion (lane 2; corresponds to background labeling), or30min incubation (lane 3, 5), or 60min incubation (lane4, 6). Then newly inserted receptors were biotinylated bybiotin-CoA, cells were lysed, and the resulting extractsused in streptavidin-bead precipitations. Western blot-ting and signal quantification indicated also an increasedinsertion upon K80 depolarization at 30 and 60mincompared to unstimulated TTX/K5 conditions(Fig. 5(e)). Although not significantly different fromthe background level, we detected signals from unsti-mulated TTX/K5 conditions. This might be due to ahigher sensitivity of the biochemical approach. Quanti-fication of the background signal (Fig. 5(d); lane 2)showed that it corresponds to only 0.13% of the signalwithout previous reaction with unconjugated CoA (lane1). This verified for the efficient blocking by the firstreaction of the surface receptors present at time 0.

We then carried out the same sequential labelingprotocol for ACP-GluR1 and ACP-GluR3. As beforefor GluR2, appearance of new GluR1- or GluR3-containing receptors after 30min was enhanced by80mM potassium depolarization (Fig. 6(a), third row,and Fig. 6(b), black bars). While the insertion of GluR3was comparable to the one of GluR2, the stimulus-induced increase in fluorescence was more robust forGluR1 than for GluR2 or GluR3. In the case of GluR1,a significant difference between 0 and 30min could alsobe detected for K5-treated neurons (Fig. 6(b), secondrow, and Fig. 6(c), white bars). These experimentsindicate that potassium depolarization enhances theinsertion rates of all three tested AMPAR subunits whenindividually infected into hippocampal neurons.

In order to obtain information about the relativedegree of receptor insertion compared to the steady-state level of receptors, ACP-GluR1-expressing neuronswere labeled with CoA-Cy3, either with precedingblocking reaction, K80 depolarization and 30minincubation (Fig. 6(c–e); 30min), or without any of thesetreatments (Fig. 6(c–e); 0min). We found that the pixelintensity in the first case corresponds to 15.13% of thatin the second case (Fig. 6(e)).

n GluR2 or GluR3. (a) Hippocampal neurons were infected at

receptors were AcpS-labeled with CoA-Alexa488 and with

rol neurons (0min) were then directly AcpS-labeled with CoA-

r K80 (stimulated), followed by incubation at 37 1C for 28min

ll the reactions and incubations of the K5 condition were done

taneous activity. (b) Fluorescence signals from CoA-Cy3 in (a)

alized on the background control fluorescence intensity. (c)

fected hippocampal neurons were either directly AcpS-labeled

-CoA and unconjugated CoA, followed by 2min stimulation by

oA-Cy3/AcpS labeling (30min). (d) Typical images from direct

ulation and incubation (lower row). (e) Reinsertion of GluR1-

tion corresponds to 15.13% of directly labeled total GluR1

Page 11: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778 773

Page 12: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778774

Surface mobility of ACP-labeled AMPAR

Movements of AMPAR along the neuron surfacecontribute to changes in the number of synaptic receptors(Choquet and Triller, 2003). Therefore, we analyzed themobility of the labeled clusters of ACP-tagged AMPARsubunits by spinning-disk confocal microscopy. We usedmature living hippocampal neurons which had beeninfected at DIV3 to express either ACP-GluR1, ACP-GluR2 or ACP-GluR3, and the observations were doneat DIV18. We determined a diffusion coefficient forACP-GluR2 of 0.00670.0003mm2/s (Fig. 7(c); gray bar).This value is in the same range as the diffusiondetermined previously for antibody-labeled endogenoussurface GluR2 (Tardin et al., 2003) or by fluorescencerecovery after photobleaching (FRAP) of neuronsexpressing GluR2 which was tagged with a pH-sensitiveGFP variant (Ashby et al., 2006). The coefficient ofACP-GluR1 was similar to the one of GluR2(0.006470.003mm2/s; Fig. 7(c); white bar). In contrast,ACP-GluR3 moved with a significantly higher diffusioncoefficient (0.008670.0004mm2/s, Fig. 7(c), black bar)than the other two subunits.

In the hippocampus GluR2 is contained in mostAMPAR either in the combination GluR1/2 or GluR2/3.Therefore, we then analyzed the mobility of GluR2-containing receptors on neurons that had been double-infected with viruses expressing either ACP9-GluR1/ACP-GluR2 or ACP-GluR2/ACP9-GluR3 (Fig. 7(d),histogram of diffusion coefficients in e). Interestingly,GluR2 has a significantly higher diffusion coefficientwhen coinfected with GluR3 (0.006970.001 mm2/s) thanwhen coinfected with GluR1 (0.004870.0003 mm2/s).This indicates that clusters of GluR2/3-AMPAR aremore mobile than the ones of GluR1/2-AMPAR. AsACP-tagged subunits colocalized strongly with synapticmarkers in these mature cultures of hippocampalneurons at DIV18 (see Fig. 4), this result suggests thatsubunit composition determines the mobility of recep-tors that can be integrated into synapses.

Discussion

AMPAR exocytosis and endocytosis as well as lateralmovements along the surface are two important meansof controlling tightly the number of receptors present at

Fig. 7. GluR1/2 and GluR2/3 AMPAR show different surface diffus

the indicated subunits, AcpS-labeled with CoA-488 at DIV20, and

images during 4 s at 30Hz was acquired. (a) Raw image for ACP-G

particles outlined. (c) Diffusion coefficient for the individual subuni

200 nm diameter embedded in solid mounting medium (beads). (d

Neurons were infected either for ACP-GluR2 and ACP9-GluR1, or

and imaged. (e) Histogram of results in (d) showing the distribution o

with either GluR1 (open bars) or GluR3 (closed bars) (**po0.001)

the synapse (Groc and Choquet, 2006). This receptornumber strongly contributes to the regulation of theefficacy of synaptic transmission. However the molecu-lar mechanisms that control these movements of thedifferent subunit variants of AMPAR remain unclear.

In the present study we used the recently developedACP-labeling technique (Gronemeyer et al., 2005;Johnsson and Johnsson, 2007) to specifically labelsurface AMPAR in cultures of dissociated neurons,and to analyze the timing of exocytosis of the AMPARsubunits GluR1, GluR2, and GluR3 following depolar-ization. This approach offers a number of importantadvantages over more traditional labeling approaches:Firstly, the labeling is restricted to only cell-surfaceproteins. This is different from GFP-tagged receptors,which are visible throughout the cell, including plasmamembrane and intracellular compartments of thesecretory and endosomal pathways. AMPAR subunitstagged with pH-sensitive GFP have previously beenused to study surface mobility (Ashby et al., 2006;Kopec et al., 2006; Yudowski et al., 2007), but they canonly be used on living cells as, upon fixation, the pH-dependent specificity for surface receptors is lost.Secondly, the ACP tag is rather small compared toGFP or to much larger extracellularly binding anti-bodies. Compared to antibodies, this small size mini-mizes problems of lateral receptor migration into thesynaptic cleft. Thirdly, a single construct can be labeledwith a variety of different fluorophores and othersynthetic probes and the time-point of labeling can befreely chosen. Ju and colleagues have used the relatedtetracysteine technique to label AMPAR with biarseni-cal dyes (Ju et al., 2004). In these experiments thetetracysteine tag needed to be located on the intracel-lular side of the plasma membrane as the tetracysteinetag is sensitive to oxidation. The approach did thereforenot allow distinguishing receptors present at the cellsurface from the intracellular receptor pool. Finally, theACP-labeling technique allows to label several subunitswith different fluorophores on the same cell (Vivero-Polet al., 2005).

When individually infected into neuron cultures, theACP-tagged AMPAR subunits localized at synapticsites, as shown by strong colocalization with PSD-95and juxtaposition with synaptophysin. This is compar-able to previous studies using GFP-tagged AMPARsubunits (Perestenko and Henley, 2003). Moreover

ion. Neurons were infected at DIV3 with lentiviruses encoding

imaged using a confocal spinning-disk microscope. A series of

luR2. (b) Zoom of the area indicated in (a) with the analyzed

ts. As a control of the system, we imaged fluorescent beads of

) GluR2 mobility is influenced by the subunit composition.

for ACP-GluR2 and ACP9-GluR3. ACP-GluR2 was labeled

f the diffusion coefficient D of GluR2 when expressed together

.

Page 13: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778 775

Page 14: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778776

ACP-tagged receptors are present at synapses capable ofFM1-43 uptake upon depolarization (Betz et al., 1992).Also, ACP-tagged subunits appear to be functional withrespect to trafficking since ACP-GluR2 is internalizedinto intracellular compartments. In neurons double-infected with tagged GluR1 and GluR2, or with taggedGluR2 and GluR3, both subunits were mostly localizingto the same cluster. In addition, ACP-biotinylation andstreptavidin precipitation indicated that these subunitsform heteromeric receptors. Interestingly, coprecipita-tion of endogenous subunits shows that these tagged,infected subunits form also receptors with the endogen-ous pool of AMPAR subunits.

Lateral mobility of AMPAR was shown in differentstudies and is thought to play an important role in themodulation of receptor numbers at the synapses(Adesnik et al., 2005; Ashby et al., 2006; Borgdorffand Choquet, 2002). Surface mobility, as represented bythe diffusion coefficient D, differs for synaptic andextrasynaptic AMPAR. While the latter are highlymobile (Tardin et al., 2003), the synaptic receptors areanchored at the PSD by various scaffolding proteinsthrough PDZ interactions (Garner et al., 2000). Thediffusion observed here for ACP-GluR2 is in the rangeof the value previously observed for synaptic receptors(Ashby et al., 2006; Tardin et al., 2003). This isconsistent with the fact that we performed the experi-ments on mature cultures of approximately 3 weeks invitro. Moreover, the applied lentiviral infection ex-pressed the subunits at levels lower than the endogenoussubunits, which reduces the risk of artefacts due tooverexpression. In general, the diffusion of single-infected subunits was slightly higher than that of thereceptors following double-infection. Interestingly,when the mobility of GluR2 was determined in neuronscoinfected with either GluR1 or GluR3, we found that itwas significantly higher in the latter case. Although wecannot exclude that some ACP-GluR2 associated withendogenous GluR3 (in the case of coexpression withACP9-GluR1), or with endogenous GluR1 (in the caseof coexpression of ACP9-GluR3), our interpretation isthat the peak of the distribution of diffusion coefficientsamong all the observed structures is shifted due to thealtered ratio of available subunits to form heteromericreceptors. This higher mobility suggests a lower degreeof anchoring of GluR2/3 receptors compared to GluR1/2 receptors. Recent findings showed that activity reducesmobility of GluR1, presumably due to synaptic anchor-ing (Ehlers et al., 2007). In light of these experiments,the weaker diffusion of GluR1/2 receptors that wefound here in mature cultures under conditions ofspontaneous activity might suggest that GluR1/2receptors possess a lower exchange rate betweensynaptic and extrasynaptic sites than GluR2/3 receptors.

It is important to note that we observed here themobility of several labeled receptors per fluorescent

signal, and not of single-labeled molecules, as we did notobserve the one-step photobleaching typical for singlefluorescently labeled molecules (data not shown). This isnot surprising, given our saturating labeling conditions.However, even in studies analyzing clearly single-labeledreceptors by antibodies against the extracellular part ofGluR2 the authors had no means to discriminate singlemoving receptors from microaggregates with non-labeled receptors (Tardin et al., 2003).

Using sequential ACP reactions that included asaturating labeling at the beginning, we visualized theappearance of new surface receptors. It is important tonote that the site where we detect new surface receptorsdoes not necessarily correspond to the insertion site: atthose sites only one or a few receptor molecules might beinserted, which could fall below our detection limits.Therefore, we detect newly inserted receptor, whichmight have migrated laterally to form receptor assem-blies. The origin of the inserted receptors might be newlysynthesized receptors, or endocytosed and recycledreceptors, which located to endosomes at the time ofthe first labeling. This might be due to different ratios ofavailable subunits in single- or double-infected neurons.In initial experiments we used for control cells only K5which yielded strong variations among different culturedishes, presumably due to different degrees of sponta-neous activity. Therefore, in later experiments weapplied K5/TTX in order to avoid this activity and thusreduce variability between cells. When we used thissequential labeling to compare the insertion rate ofGluR1, GluR2, and GluR3, we found that potassiumdepolarization increased the new insertion of all threesubunits, with the strongest effect for GluR1. We foundbasal recruitment of GluR1 after 30min also withoutdepolarization, and no significant insertion of GluR3 atresting conditions. While comparison between thedifferent data sets for GluR1, GluR2 and GluR3 isdifficult because of the separate, individually infectedneurons, these results suggest that under resting condi-tions GluR1 can also cycle at significant rates. Since thenon-detection of GluR2 and GluR3 at resting condi-tions is a negative result, it is difficult to judge whetherthis truly reflects non-cycling of GluR3 in the absence ofstimulation, or a lack of sensitivity to detect a lowdegree of insertion. Interestingly, potassium depolariza-tion induced GluR3 insertion. This suggests that ACP-GluR3-containing AMPAR are also prone to stimulus-evoked insertion. Although transfected GFP-GluR1 hasbeen detected at the surface in dissociated neurons, inslice cultures it was mostly intracellular in the absence ofstimulation (Shi et al., 1999). Passafaro et al. (2001)observed under control conditions a stronger exocytosisrate for GluR2 than for GluR1. This different resultcompared to our data might arise from the differentexpression systems. Lentiviral infection expresses theexogenous subunit at lower levels than the endogenous

Page 15: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778 777

subunits, and thereby favors the formation of hetero-meric receptors. However, the massive surface recruit-ment of GluR1 that we observed followingdepolarization with 80mM potassium, is comparableto the results upon glycine, NMDA, or insulintreatments by Passafaro et al. (2001). In conclusion,using this new surface labeling technique we could showsubunit-specific differences in surface insertion andmobility of AMPAR subunits.

Acknowledgments

We acknowledge the important help of NathalieGeorge in the preparations of CoA substrates andenzymes. We want to thank Maria Passafaro for makingavailable to us the GluR cDNAs, and William Pralongfor critical reading of the manuscript. This work wassupported by grants from the Swiss National ScienceFoundation (3100A0-111935/1).

References

Adesnik, H., Nicoll, R.A., England, P.M., 2005. Photoinacti-

vation of native AMPA receptors reveals their real-time

trafficking. Neuron 48, 977–985.

Ashby, M.C., Maier, S.R., Nishimune, A., Henley, J.M., 2006.

Lateral diffusion drives constitutive exchange of AMPA

receptors at dendritic spines and is regulated by spine

morphology. J. Neurosci. 26, 7046–7055.

Bats, C., Groc, L., Choquet, D., 2007. The interaction between

Stargazin and PSD-95 regulates AMPA receptor surface

trafficking. Neuron 53, 719–734.

Beattie, E.C., Carroll, R.C., Yu, X., Morishita, W., Yasuda,

H., von Zastrow, M., Malenka, R.C., 2000. Regulation of

AMPA receptor endocytosis by a signaling mechanism

shared with LTD. Nat. Neurosci. 3, 1291–1300.

Betz, W.J., Mao, F., Bewick, G.S., 1992. Activity-dependent

fluorescent staining and destaining of living vertebrate

motor nerve terminals. J. Neurosci. 12, 363–375.

Borgdorff, A.J., Choquet, D., 2002. Regulation of AMPA

receptor lateral movements. Nature 417, 649–653.

Choquet, D., Triller, A., 2003. The role of receptor diffusion in

the organization of the postsynaptic membrane. Nat. Rev.

Neurosci. 4, 251–265.

Ehlers, M.D., 2000. Reinsertion or degradation of AMPA

receptors determined by activity-dependent endocytic sort-

ing. Neuron 28, 511–525.

Ehlers, M.D., Heine, M., Groc, L., Lee, M.C., Choquet, D.,

2007. Diffusional trapping of GluR1 AMPA receptors by

input-specific synaptic activity. Neuron 54, 447–460.

Garner, C.C., Nash, J., Huganir, R.L., 2000. PDZ domains in

synapse assembly and signalling. Trends Cell Biol. 10,

274–280.

George, N., Pick, H., Vogel, H., Johnsson, N., Johnsson, K.,

2004. Specific labeling of cell surface proteins with

chemically diverse compounds. J. Am. Chem. Soc. 126,

8896–8897.

Gerges, N.Z., Backos, D.S., Rupasinghe, C.N., Spaller, M.R.,

Esteban, J.A., 2006. Dual role of the exocyst in AMPA

receptor targeting and insertion into the postsynaptic

membrane. EMBO J. 25, 1623–1634.

Groc, L., Choquet, D., 2006. AMPA and NMDA glutamate

receptor trafficking: multiple roads for reaching and leaving

the synapse. Cell Tissue Res 326, 423–438.

Gronemeyer, T., Godin, G., Johnsson, K., 2005. Adding value

to fusion proteins through covalent labelling. Curr. Opin.

Biotechnol. 16, 453–458.

Hayashi, Y., Shi, S.H., Esteban, J.A., Piccini, A., Poncer, J.C.,

Malinow, R., 2000. Driving AMPA receptors into synapses

by LTP and CaMKII: requirement for GluR1 and PDZ

domain interaction. Science 287, 2262–2267.

Johnsson, N., Johnsson, K., 2007. Chemical tools for

biomolecular imaging. ACS Chem. Biol. 2, 31–38.

Johnsson, N., George, N., Johnsson, K., 2005. Protein

chemistry on the surface of living cells. Chembiochem 6,

47–52.

Ju, W., Morishita, W., Tsui, J., Gaietta, G., Deerinck, T.J.,

Adams, S.R., Garner, C.C., Tsien, R.Y., Ellisman, M.H.,

Malenka, R.C., 2004. Activity-dependent regulation of

dendritic synthesis and trafficking of AMPA receptors.

Nat. Neurosci. 7, 244–253.

Kopec, C.D., Li, B., Wei, W., Boehm, J., Malinow, R., 2006.

Glutamate receptor exocytosis and spine enlargement

during chemically induced long-term potentiation. J.

Neurosci. 26, 2000–2009.

Kulangara, K., Kropf, M., Glauser, L., Magnin, S., Alberi, S.,

Yersin, A., Hirling, H., 2007. Phosphorylation of glutamate

receptor interacting protein 1 regulates surface expression

of glutamate receptors. J. Biol. Chem. 282, 2395–2404.

Kusumi, A., Sako, Y., Yamamoto, M., 1993. Confined lateral

diffusion of membrane receptors as studied by single

particle tracking (nanovid microscopy). Effects of cal-

cium-induced differentiation in cultured epithelial cells.

Biophys. J. 65, 2021–2040.

Lee, S.H., Simonetta, A., Sheng, M., 2004. Subunit rules

governing the sorting of internalized AMPA receptors in

hippocampal neurons. Neuron 43, 221–236.

Lu, W., Man, H., Ju, W., Trimble, W.S., MacDonald, J.F.,

Wang, Y.T., 2001. Activation of synaptic NMDA receptors

induces membrane insertion of new AMPA receptors and

LTP in cultured hippocampal neurons. Neuron 29,

243–254.

Luscher, C., Xia, H., Beattie, E.C., Carroll, R.C., von

Zastrow, M., Malenka, R.C., Nicoll, R.A., 1999. Role of

AMPA receptor cycling in synaptic transmission and

plasticity. Neuron 24, 649–658.

Man, H.Y., Lin, J.W., Ju, W.H., Ahmadian, G., Liu, L.,

Becker, L.E., Sheng, M., Wang, Y.T., 2000. Regulation of

AMPA receptor-mediated synaptic transmission by cla-

thrin-dependent receptor internalization. Neuron 25,

649–662.

Matsuda, S., Launey, T., Mikawa, S., Hirai, H., 2000.

Disruption of AMPA receptor GluR2 clusters following

long-term depression induction in cerebellar Purkinje

neurons. EMBO J. 19, 2765–2774.

Meyer, B.H., Martinez, K.L., Segura, J.M., Pascoal, P.,

Hovius, R., George, N., Johnsson, K., Vogel, H., 2006.

Page 16: Subunit-specific surface mobility of differentially labeled AMPA receptor subunits

ARTICLE IN PRESSM. Kropf et al. / European Journal of Cell Biology 87 (2008) 763–778778

Covalent labeling of cell-surface proteins for in-vivo FRET

studies. FEBS Lett. 580, 1654–1658.

Park, M., Penick, E.C., Edwards, J.G., Kauer, J.A., Ehlers,

M.D., 2004. Recycling endosomes supply AMPA receptors

for LTP. Science 305, 1972–1975.

Passafaro, M., Piech, V., Sheng, M., 2001. Subunit-specific

temporal and spatial patterns of AMPA receptor exocytosis

in hippocampal neurons. Nat. Neurosci. 4, 917–926.

Perestenko, P.V., Henley, J.M., 2003. Characterization of the

intracellular transport of GluR1 and GluR2 alpha-amino-3-

hydroxy-5-methyl-4-isoxazole propionic acid receptor subunits

in hippocampal neurons. J. Biol. Chem. 278, 43525–43532.

Prummer, M., Meyer, B.H., Franzini, R., Segura, J.M.,

George, N., Johnsson, K., Vogel, H., 2006. Post-transla-

tional covalent labeling reveals heterogeneous mobility of

individual G protein-coupled receptors in living cells.

Chembiochem 7, 908–911.

Qian, H., Sheetz, M.P., Elson, E.L., 1991. Single particle

tracking. Analysis of diffusion and flow in two-dimensional

systems. Biophys. J. 60, 910–921.

Shi, S.H., Hayashi, Y., Petralia, R.S., Zaman, S.H., Wenthold,

R.J., Svoboda, K., Malinow, R., 1999. Rapid spine delivery

and redistribution of AMPA receptors after synaptic

NMDA receptor activation. Science 284, 1811–1816.

Shi, S., Hayashi, Y., Esteban, J.A., Malinow, R., 2001.

Subunit-specific rules governing AMPA receptor trafficking

to synapses in hippocampal pyramidal neurons. Cell 105,

331–343.

Sielaff, I., Arnold, A., Godin, G., Tugulu, S., Klok, H.A.,

Johnsson, K., 2006. Protein function microarrays based on

self-immobilizing and self-labeling fusion proteins. Chem-

biochem 7, 194–202.

Stachelhaus, T., Mootz, H.D., Bergendahl, V., Marahiel,

M.A., 1998. Peptide bond formation in nonribosomal

peptide biosynthesis. Catalytic role of the condensation

domain. J. Biol. Chem. 273, 22773–22781.

Steiner, P., Sarria, J.C., Glauser, L., Magnin, S., Catsicas, S.,

Hirling, H., 2002. Modulation of receptor cycling by

neuron-enriched endosomal protein of 21 kD. J. Cell Biol.

157, 1197–1209.

Steiner, P., Alberi, S., Kulangara, K., Yersin, A., Sarria, J.C.,

Regulier, E., Kasas, S., Dietler, G., Muller, D., Catsicas, S.,

Hirling, H., 2005. Interactions between NEEP21, GRIP1

and GluR2 regulate sorting and recycling of the glutamate

receptor subunit GluR2. EMBO J. 24, 2873–2884.

Tardin, C., Cognet, L., Bats, C., Lounis, B., Choquet, D.,

2003. Direct imaging of lateral movements of AMPA

receptors inside synapses. EMBO J. 22, 4656–4665.

Vivero-Pol, L., George, N., Krumm, H., Johnsson, K.,

Johnsson, N., 2005. Multicolor imaging of cell surface

proteins. J. Am. Chem. Soc. 127, 12770–12771.

Wenthold, R.J., Petralia, R.S., Blahos II, J., Niedzielski, A.S.,

1996. Evidence for multiple AMPA receptor complexes in

hippocampal CA1/CA2 neurons. J. Neurosci. 16,

1982–1989.

Xia, J., Chung, H.J., Wihler, C., Huganir, R.L., Linden, D.J.,

2000. Cerebellar long-term depression requires PKC-

regulated interactions between GluR2/3 and PDZ do-

main-containing proteins. Neuron 28, 499–510.

Yudowski, G.A., Puthenveedu, M.A., Leonoudakis, D.,

Panicker, S., Thorn, K.S., Beattie, E.C., von Zastrow, M.,

2007. Real-time imaging of discrete exocytic events

mediating surface delivery of AMPA receptors. J. Neurosci.

27, 11112–11121.