The M3/M4 Cytoplasmic Loop of thea1 Subunit Restricts GABAARs LateralMobility: A Study Using Fluorescence

Recovery After Photobleaching

Macarena Peran,1* Helen Hooper,2 Houria Boulaiz,3 Juan A. Marchal,3

Antonia Aranega,1 and Ramiro Salas4

1Department of Anatomy and Human Embryology, Faculty of Medicine,University of Granada, Granada, Spain

2School of Applied and Molecular Sciences, University of Northumbria atNewcastle, Newcastle, Great Britain

3Department of Health Sciences, University of Jaen, Jaen, Spain4Department of Neuroscience, Baylor College of Medicine, Houston, Texas

A crucial problem in neurobiology is how neurons are able to maintain neuro-transmitter receptors at specific membrane domains. The large structural heteroge-neity of gamma aminobutyric acid receptors (GABAARs) led to the hypothesisthat there could be a link between GABAAR gene diversity and the targeting prop-erties of the receptor complex. Previous studies using Fluorescence RecoveryAfter Photobleaching (FRAP) have shown a restricted mobility in GABAARs con-taining the a1 subunit. The M3/M4 cytoplasmic loop is the region of the a1 subu-nit with the lowest sequence homology to other subunits. Therefore, we askedwhether the M3/M4 loop is involved in cytoskeletal anchoring and GABAARclustering. A series of a1 chimeric subunits was constructed: a1CH (control sub-unit), a1CD (Cytoplasmic loop deleted), a1CD2, and a1CD3 (a1 with the M3/M4 loop from the a2 and a3 subunits, respectively). Our results using FRAP indi-cate an involvement of the M3/M4 cytoplasmic loop of the a1 subunit in control-ling receptor lateral mobility. On the other hand, inmunocytochemical approachesshowed that this domain is not involved in subunit targeting to the cell surface,subunit-subunit assembly, or receptor aggregation. Cell Motil. Cytoskeleton63:747–757, 2006. ' 2006 Wiley-Liss, Inc.

Key words: GABAARs lateral mobility; FRAP; receptor clustering; anchoring

INTRODUCTION

It is broadly believed that GABAAR molecular heter-ogeneity determines biophysical and pharmacologicalproperties of the receptor [for review see Sieghart, 2000].In addition, there are strong evidences of the localizationof GABAARs not only at synaptic sites but also at presyn-aptic and axonal sites [Sakatani et al., 1991; Nusser et al.,1998; Kullmann et al., 2005]. However, it is not clearwhether these neurotransmitter receptors are edited androuted to their final postsynaptic domains or if a freely mo-bile pool of receptors is maintained on the cell surface.Neuronal demands could recruit these free receptors to

This work was performed at Department of Biological Sciences,

Durham University, UK.

Contract grant sponsors: European community; University of Durham.

*Correspondence to: Dr. Macarena Peran, Departamento de Anatomıa

y Embriologıa Humana, Facultad de Medicina, Universidad de

Granada, E-18012 Granada, Spain. E-mail: mperan@ugr.es

Received 18 January 2006; Accepted 26 July 2006

Published online 6 October 2006 in Wiley InterScience (www.

interscience.wiley.com).

DOI: 10.1002/cm.20156

' 2006 Wiley-Liss, Inc.

Cell Motility and the Cytoskeleton 63:747–757 (2006)

specific regions of the neuronal membrane. Lateral move-ment in the plasma membrane represents a mechanism bywhich the local GABAAR concentration might be regu-lated to adjust the efficacy of synaptic inhibitory transmis-sion. In fact, we have previously shown that the diversityof subunit isoforms determines distribution and mobilityof GABAAR on the cell surface [Peran et al., 2001, 2004].

The use of quantum dot labelling of single gly-cine receptor molecules [Dahan et al., 2003] or single-molecule fluorescence microscopy of AMPA receptors[Tardin et al., 2003] have confirmed as well, that individ-ual neurotransmitter receptor molecules can be observedmoving in and out of synapses independent of endocyto-sis. Several groups are involved in the study of howGABAAR are anchored and maintained at specific mem-brane domains [review in Luscher and Keller, 2004].Their approaches, however, depend on biochemicalreconstitution of the system. The study presented herehas approached the problem of receptor anchoring andidentification of the molecular interactions in a differentmanner. Thus, attention was focused on the membranedynamics of the receptor, measuring its lateral mobilityby using the Fluorescence Recovery After Photobleach-ing (FRAP) method. FRAP is a powerful tool that per-mits the lateral motions of molecules in the membranesof single, living cells to be examined under physiologicalconditions [Reits and Neefjes, 2001]. In FRAP experi-ments fluorescently labelled molecules are irreversiblyphoto bleached by a high-intensity laser beam thatbriefly illuminates a small area of the cell. Diffusion ofsurrounding non-bleached molecules leads to recoveryof the fluorescence. A higher mobility of the moleculesresults in a shorter time of recovery. From the recoverycurve, it is possible to obtain estimates of the diffusioncoefficient and immobile fraction [Axerold et al., 1976].

We have already provided evidence, using FRAP,that GABAARs containing the a1 subunit have a restrictedlateral mobility [Peran et al., 2001, 2004]. Because themajor amino acid sequence divergence between a subunitisoforms is found in the M3/M4 cytoplasmic domain, wehypothesised that this region might potentially mediate a1subunit-specific association with the cytoskeleton and/orconfine receptors to specific neuronal domains. The intra-cellular loop M3/M4 contributes most of the cytoplasmicdomain of these receptors and includes multiple interactionsites for putative trafficking and postsynaptic scaffold pro-teins as well as phosphorylation sites for diverse serine/threonine and tyrosine kinases [Luscher and Keller, 2004].In addition, this region of the a1 subunit might contain aretention motif responsible for the retention of recombi-nant a1 subunit homo-oligomers in an intracellular com-partment [Peran et al., 2001].

The objective of the present work was to study thepossible implication of the a1 cytoplasmic loop in

GABAARs trafficking and anchoring. To do so severalchimeric a1 subunits were constructed: an a1 subunit inwhich the M3/M4 domain had been excised (a1CD) anda1 subunits where the M3/M4 loop was replaced by theloops of the a2 subunit (a1CD2) and the a3 subunit(a1CD3). The a2 and a3 were chosen for construction ofthe chimeric a subunits because there is evidence thatthese subunits form co-localised clusters with gephyrin[Koksma et al., 2005].

These chimeric subunits were transiently expressedwith wild-type b3 and g2s subunits to form recombinantGABAA receptors. FRAP approaches and immunocyto-chemical studies were performed on these transientlyexpressed receptors. We discovered that the M3/M4cytoplasmic loop of the a1 subunit regulates receptormobility but not receptor trafficking.

The advantages of using FRAP in the study ofGABAARs allowed us to investigate in a non-invasiveway the behaviour of these receptors in living cells.

These observations shed more light on the role ofthe M3/M4 cytoplasmic loop of the a1 subunit in recep-tor sorting, and we believe will be very useful for otherstudies aimed at explaining the role of the differentGABAAR subunits in physiology.

MATERIALS AND METHODS

Subcloning Strategy in the Production ofGABAAR aCH, aCD and GABAAR aCDax

The following chimeric GABAAR a1 subunitswere constructed (see Fig. 1).

1. Chimeric GABAAR a1, a1CH (control subunit).To create the a1CH a 0.9 kb Bsp120I - EcoNI 50restriction fragment (encoding the N0 terminalbd24 epitope) of the bovine GABAAR a1 wasligated to a 0.6 kb 30 EcoNI- EcoRI restrictionfragment of the rat into the mammalian expres-sion vector pCDNAIamp.

2. Cytoplasmic domain deleted GABAAR a1,a1CD. To do so the 30 end of rat GABAAR a1

was amplified from codon 414 (in effect de-leting codons 362–413) using an universal re-verse primer (sp6) and a forward primer (50aggatcctctctcgagcgtcagcaaaatcgaccg 30) whichencodes two unique restriction sites for BamHI(in bold) and XhoI (italic) at its 50 end. The30 end of GABAAR aCH was removed by diges-tion with BamHI and EcoRI and replaced withthe BamHI - EcoRI double digested rat PCRproduct.

3. GABAAR a1 in which the cytoplasmic domainof the a1 was swapped with the ones of the a2

748 Peran et al.

and a3 to give GABAAR a1CD2 and a1CD3. Tocreate GABAAR aCD(2-3). GABAAR aCD wasdouble digested with BamHI and XhoI. The CDsfrom GABAAR a2-3 were individually amplifiedusing forward primers encoding a BamHI restric-tion site and reverse primers encoding a XhoIrestriction site (Primers sequences used for aCD2:Forward: 50aggatccaggctccgtcatgata 30; Reverse:50gctcgagttgaaagttttcttgg 30. Primers sequences usedfor aCD3: Forward: 5

0aggatccagcagcccaaccaa 30;Reverse: 50gctcgagttgtaggtcttggtct 30). AmplifiedGABAAR a2-3 CDs were double digested withBamHI and XhoI and individually ligated intoBamHI and XhoI digested GABAAR aCD.

Cell Culture. HEK293 cells were cultured inEagle’s MEM supplemented with 10% fetal bovine se-rum (FBS). COS7 cells were grown in Dulbecco’s MEMwith 10% (FBS).

Transfection. Cells were plated on 35 mm poly-D-lysine coated dishes and transfected with Lipofectamine(GIBCO-BRL) following manufacturers instructions.Cells were analysed 48 h post-transfection. Mock trans-fections were performed with vector only.

Immunocytochemistry

Antibodies. Monoclonal anti-a1 (bd24, Boehringer,Germany) was used at a 1:50 dilution and polyclonal anti-

body anti-b102/103 for the b3 subunit was used at a 1:200dilution. Secondary antibodies used were TRITC-conju-gated goat anti-mouse and TRITC-conjugated goat anti-rabbit (Calbiochem, Nottingham, UK) or Cascade blue-conjugated goat anti-mouse antibody (Molecular Probes,Leiden, The Nertherlands) at 1:200 dilution.

Labelling of Fixed and Permeabilized Cells.

Cells were fixed with 4% paraformaldehyde in phosphatebuffered saline (PBS) for 15 min, washed twice with PBS,blocked in 10% heat inactivated serum (HIS) for 15 minand then incubated for 1 h in primary antibody diluted inbuffer A (1 mg/ml BSA, 10% HIS, 0.5% Triton X-100,PBS) at room temperature. Cells were washed three timesin PBS and then incubated for 30 min with secondary anti-bodies in buffer A. Controls were performed with mocktransfected cells and using secondary antibody only.

Labelling of Live Cells. The bottom of culturedishes were replaced by coverslips which permitteddirect viewing of live cells using a high numerical aper-ture objective in an inverted microscope. Transfectedcells plated in these dishes were washed twice in PBSthen incubated at 48C in primary antibody diluted ingrowth media for 45 min and secondary antibody for 20min at room temperature.

Cells visualized with Bodipy-Ro-1986 (a fluores-cent benzodiazepine that binds g subunit-containingGABAARs, Molecular Probes, Leiden, The Netherlands)and with the primary antibody bd24 were labelled for30 min with the bd24 diluted in PBS and with 100 nM of

Fig. 1. Schematic representation of the a1 subunit

and mutations.

a1 M3/M4 Cytoplasmic Loop Restricts GABAARs Lateral Mobility 749

the fluorescent benzodiazepine together with the second-ary antibody diluted in PBS for 30 min.

Image Capture and Analysis. Images were ob-tained using a Nikon FX-35A camera adapted to a Nikonfluorescence microscope. Photographs were recordedthrough a Zeiss 633 water immersion 1.2 NA objectiveon Ektachrome film pushed to ASA 3200. Images weretransferred to Adobe Photoshop 7.0.

Production of Fluorescently Labelled Anti-a1

Fab0 Antibody Fragment FRAP Probes

The monoclonal antibody bd24 (Boehringer) rec-ognizes both the bovine and human a1 subunit ofGABAAR. The bd24 antibody (10 lg) was digested with1 lg of papain in 0.2 mM Na-acetate, pH 5, 1 M cysteineand 20 mM EDTA for 15 h at 378C. Iodoacetamide wasthen added to a final concentration of 75 mM and incu-bated for 30 min at room temperature. Fab0 fragmentswere purified by Protein A column chromatography(Pierce, Northumberland, UK). For direct labelling,Fab0 fragments were conjugated with the fluorophoreBODIPY 493/503, SE (Molecular Probes, Leiden, TheNetherlands). Fab0 fragments were diluted to 900 ll in0.1 M sodium carbonate, pH 9.2 and 50 ll of BODIPY493/503, SE (1 mg/ml) were added in 5 ll aliquots, withgentle, but continuous stirring. The reaction was incu-bated in the dark for 8 h at 48C after which unbound dyewas removed by Sephadex G-10 column chromatography(Sigma, St. Louis, MO).

Fluorescence Photobleach Recovery

Cells for FRAP were washed in phenol red-freeMEM prior to experimentation. GABAARs on living cellswere labelled with 100–200 ll of Fab0a1-BODIPY (0.01lg/ll) diluted in 500 ll red-free MEM per 35 mm plate ofcells at room temperature for 20 min in dark. The cellswere gently washed three times with PBS and transferredto a microscope stage maintained at 22–248C.

For FRAP experiments performed with Bodipy-Ro-1986 as a probe cells were labelled with Bodipy-Ro-1986 at 40 or 100 nM in PBS-sucrose. Although atthese concentrations not all receptors are labelled, theseconcentrations were chosen based upon the Kd of thefluorescent benzodiazepine to help minimise any non-specific binding or lipid partitioning of the fluorescentprobe. The signal from the unbound fluorophore is negli-gible because the fluorescence of the Bodipy conjugatesis enhanced upon binding to the receptor. Non-specificbinding of Bodipy-Ro-1986 was determined by includ-ing chlorazepate (1 mM) in the FRAP assay. The non-specific labelling, based upon photon counts obtainedunder identical experimental conditions, was found to beless than 10% of total.

For FRAP experiments, a 633 and 1.2 NA waterimmersion objective was used directly in the culture dish.A Zeiss Universal fluorescence microscope was used withan argon-ion monitoring laser (k ¼ 498 nm, 5 mW).The beam was focused to a Gaussian radius of 1.2 lm anda �2 lm2 region of membrane was illuminated by a brief(10–200 ms) laser pulse (5 mW) photobleaching � 75% ofthe initial fluorescence. Bleaching during the initial moni-toring phase is negligible within the time frame of theexperiment. The time course of fluorescence recovery wasfollowed with an attenuated monitoring beam. DC and Fwere determined by curve-fitting procedures based on theo-retical models described previously [Axelrod et al., 1976].Single FRAP measurements were taken from each cell. Toavoid long exposures of the cells preparations and possiblelost of cell vitality, only 10 measurements were carried outfor each experiment. Standard mean errors were calculatedfrom the repeated measurements on cultures for all experi-ments. FRAP experiments performed with the samerecombinant receptor combinations were repeated on dif-ferent days and little variation in the results was found.

Statistical analyses were routinely performed usingthe statistical package Statistica version 6. A Kolmogorov-Smirnov test of normality was performed and the signifi-cance of the difference between data was tested using aStudent’s t-test. Data are presented throughout the study asmean +/- standard errors of the mean. In all cases, a proba-bility of less than 0.05 was regarded as statistically signifi-cant. Pearson-moment correlation was carried out amongmeasured variables.

RESULTS

Localization of the Recombinant ReceptorIncluding the Chimeric Subuniton the Cell Membrane

The cellular distribution of recombinant a1CHb3g2sand a1CDb3g2s receptors was studied by immunocyto-chemistry in HEK293 cells (data not shown) and COS7cells (Fig. 2). Both a1CHb3g2s and a1CDb3g2s wereexpressed in clusters on the surface of live HEK293 andCOS7 cells.

These results imply that the transfection of thegenetic material containing the truncated subunit were ableto form stable protein complexes. These chimeric a1 sub-units were able to assemble and be expressed on the cell sur-face in conjunction with b3 and g2 ‘‘wild type’’ subunits.

Lateral Mobility of Truncated Receptors

COS7 and HEK293 cells were transfected with b3and g2s subunit cDNAs together with each of the a1constructs: a1CH, a1CD, a1CD2 and a1CD3 and the mobil-ity of the expressed receptors were measured by FRAP.

750 Peran et al.

Recombinant a1Chimeric(b3g2s) complexes were labelledwith fluorescently labelled Fab0 antibody fragment of thea1 subunit specific bd24 antibody (a1- Fab

0 BODIPY).For a1CH(b3g2s) and a1CD(b3g2s)- expressed re-

ceptors parallel experiments were performed using fluo-rescently BODIPY FL Ro-1986, a fluorescent benzodiaze-pine that labels GABAARs containing the g2s subunit.

The non-specific labelling, based upon photoncounts obtained in experiments using non-transfectedcells, under identical experimental conditions, was foundto be less than 10% of total.

No statistically significant difference in the mobilefraction (% Recovery) of recombinant GABAARsexpressed in HEK293 was found. The recombinantreceptors containing the different a constructs were ableto freely move on the cell membrane. The mobile frac-tion of receptors comprising the a1 subunit chimerasranged from 65 to 71%. These results are in concordancewith those obtained before with the wild types of a subu-nits expressed on HEK293 cells [Peran et al., 2004].These cells seem to lack those elements necessary toanchor GABAARs since the receptors were mobilewhichever subunit combination was expressed.

In contrast, when the same receptors where ex-pressed in COS7 cells, a statistically significant differ-

ence in the relative mobile fractions of the expressedreceptors was found. Using Fab0 a1- BODIPY, the mobilefraction of a1CHb3g2s expressed receptors was 27.7% 66% (n ¼ 10) whilst 76.7% 6 7% of a1CDb3g2s expressedreceptors were freely mobile (n ¼ 10) (Fig. 3a). Recep-tors containing the chimeric a1 with the M3/M4 loopwere almost immobile, after 60 s only 27% of the fluo-rescence was recovery. The deletion of the cytoplasmicloop changed the behaviour of GABAAR containing thea1 subunit. The mobile fraction of a1CDb3g2s expressedreceptors went up to 70% in the same time.

When FRAP experiments were performed usingBODIPY FL Ro-1986 as a probe the results were verysimilar: 29% 6 8% for a1CHb3g2s expressed receptors(n ¼ 12) and 75% 6 10% for a1CDb3g2s expressedreceptors (n ¼ 10).

Despite the clear difference in the percentage of rela-tive mobile fractions of a1CHb3g2s and a1CDb3g2sreceptors expressed in COS7 cells, both receptor typeswere clustered at the cell surface (Fig. 2). This implies thatreceptor aggregation does not govern receptor mobility.

For COS7 cells (Fig. 3a) the recovery of the fluores-cence signal was 73.6%6 9% for a1CD2b3g2s and 71.6%6 8% for a1CD3b3g2s receptors, compared to 27.7% 66% for a1CHb3g2s. These results suggest that exchanging

Fig. 2. Cell surface expression of a1CHb3g2s and a1CDb3g2sGABAARs expressed in COS7 cells. COS7 cells were co-transfected with

the a1CH cDNA together with b3 and g2s subunit cDNAs (a and b); Cellsco-transfected with a1CD cDNA together with b3 and g2s subunit cDNAsare shown in (c) and (d). Cells were labelled live for the a1CH/CD subunit

with the a1 subunit-specific antibody bd24 (a and c) and visualised with

TRITC-conjugated anti-mouse. In addition, to visualise expressed com-

plexes containing g2s the benzodiazepine Bodipy-Ro-1986 was used (b

and d). Scale bars for all panels: 20 lm. [Color figure can be viewed in

the online issue, which is available at www.interscience.wiley.com.]

a1 M3/M4 Cytoplasmic Loop Restricts GABAARs Lateral Mobility 751

the cytoplasmic loop of the a1 subunit with the corre-sponding domain of the a2 or a3 subunits released the re-ceptor complex from the constraints that tethered the re-ceptor on the cell surface. Thus, although a1CD2 anda1CD3 chimeras have an a1 subunit amino-acid backbonethe replacement of the cytoplasmic loop transformed themobility pattern of wild type a1 subunits.

Figure 3b shows the rate of movement of a1CD,a1CH, a1CD2 and a1CD3-containing receptors was typicalof most membrane glycoproteins with diffusioncoefficients of the range of 0.97–1.44 3 10–10 cm2 s�1

[Jacobson et al., 1983]. It should be noted, however, thatalthough these values are expected using FRAP, when

single particle tracking is used, the absolute value of dif-fusion coefficients can be quite different [Simson et al.,1998]. It is therefore important to qualify these results inthe light of the technique used.

To determine whether receptor cluster density hasany effect on diffusion parameters, the diffusion coeffi-cient and mobile fractions were plotted against the pre-bleach fluorescence intensity. As can be seen in Fig. 4there is no evidence for a correlation between pre-bleachintensity and the mobile fraction.

In addition, statistical analysis of the data showedthat there were no significant correlations between pre-bleach intensity or bleach intensity versus recovery frac-tion or half-time of recovery (data not shown). Thismeans that the results of lateral mobility were not influ-enced by the experimental conditions.

Localisation of the a1 Subunit Chimerasin Transfected COS7 and HEK293 Cells

We have provided evidence that a1 subunit homo-oligomers expressed in HEK293 or COS7 cells are notdirected to the cell surface but remain intracellular,unable to exit the ER [Peran et al., 2001].

On that report, we showed that when COS7 cellstransfected with a1 are live-stained, no signal is appa-rent. In contrast, after fixation and permeabilization, a1signal is clearly intracellularly located. To test whetheran ER-retention signal is contained within the M3/M4cytoplasmic loop the truncated a1 subunit constructlacking this domain, a1CD, was transiently transfectedinto COS7 and HEK293 cells. Immunocytochemistrywas performed to determine where in these cells thissubunit was expressed and compared with the distribu-tion of bovine/rat chimera, a1CH. When COS7 andHEK293 cells transiently transfected with the a1CD sub-unit were stained live with the monoclonal antibodybd24, no surface fluorescence was observed. The sameresults were obtained with the chimeric subunit, a1CH(data not shown). Immunoreactive product was only visi-ble if the cells were fixed and permeabilised. Hence,a1CH and a1CD chimeric subunits were retained in an in-tracellular compartment. Figure 5a, shows that the intra-cellular distribution of a1CH-containing receptors inHEK293 cells, is similar to the pattern shown by a1CD-containing receptors (Fig. 5b) and to the one obtained fora1 subunit homo-oligomers [Peran et al., 2001]. Match-ing results were obtained in transfected COS7 cells(Fig. 5), the intracellular retention of expressed a1CHand a1CD-containing receptors are shown in Figs. 5c and5d, respectively.

To test whether the M3/M4 cytoplasmic loops of a2or a3 subunits could direct a1 subunit homo-oligomers tothe cell surface, the distribution of transiently expresseda1CD2 (Fig. 5e) and a1CD3 (Fig. 5f) subunit constructs

Fig. 3. FRAP results showing the mobile fraction or percentage re-

covery (a) and the Diffusion Coefficient (b) of recombinant

GABAARs containing the chimeric a1-subunit (a1CH, a1CD, a1CD2and a1CD3) together with the wild type subunits b3 and g2s in trans-

fected COS7 cells. Complexes were measured using the Fab0 a1-

BODIPY probe.

752 Peran et al.

were analysed by immunocytochemistry in COS7 cells.All the a1 subunit constructs were found retained intra-cellularly.

b3 Subunits Re-route a1 Subunits Chimerasto the Cell Surface

We have shown before [Peran et al., 2001] inagreement with [Connor et al., 1998] that the b3 subunitrescues the a1 subunit from its intracellular retention siteand re-route it to the cell surface. To determine if theM3/M4 cytoplasmic loop is necessary for a1-b3 subunitassembly, the a1 subunit constructs used in the previoussection were transfected into COS7 cells together withthe b3 subunit. The lack of a cytoplasmic loop (a1CD)did not prevent a1 subunits from been rescued by b3subunits. When a1CD subunit cDNA and b3 subunitcDNA were cotransfected into COS7 cells (Figs. 6c and6d), these subunits were found clustered at the cell sur-face in a pattern that paralleled that found for a1b3[Peran et al., 2001] and a1CHb3 (Figs. 6a and 6b) com-plexes expressed in these cells. Identical results wereobtained with the a1 subunit constructs a1CD2, and

a1CD3 when they were cotransfected into COS7 cellswith the b3 subunit (Figs. 6e and 6f and Figs. 6g and 6hrespectively).

Together these results demonstrate that the a1 sub-unit constructs (a1CD, a1CD2 and a1CD3) are restricted tothe ER when expressed alone as for the wild-type a1 subu-nit. Only when these subunits were co-expressed with theb3 subunit, they did acquire the signal to leave the ER andform co-localised clusters on the cell membrane.

DISCUSSION

Deletion of the M3/M4 Cytoplasmic Domain of a1Subunit Does Not Prevent Receptor Assembly orthe Rescue of the a1 Subunit From the ER

Functional receptor complexes are formed bydiverse subunits. Proper folding and assembly of the sub-units that constitute a receptor is required for the deliveryof multimeric proteins to the membrane. Therefore, het-erodimerization is a mechanism that enables the cells torestrict traffic to the cell membrane only to those recep-tor properly assembled. In some receptors, a specific ER

Fig. 4. Scatter plot of %Recovery of the fluorescence against pre-bleach intensity. (a) a1CHb3g2sGABAARs; (b) a1CDb3g2s GABAARs, (c) a1CD2b3g2s GABAARs and (d) a1CD3b3g2s GABAARs.

a1 M3/M4 Cytoplasmic Loop Restricts GABAARs Lateral Mobility 753

retention signal contained in one subunit is maskedby other type of subunit during receptor assembly[Zerangue et al., 1999; Bichet et al., 2000; Margeta-Mitrovic et al., 2000].

There are many studies reporting the localizationof ER retention signal in other neurotransmitter receptorchannels. For instance, deletion of a retention signalwithin the I-II loop of the Ca2+ channel a1 subunit facili-tates the expression of this subunit in absence of beta[Bichet et al., 2000]. In NMDA receptors, an RXR-typeER retention/retrieval motif in the C-terminal tail of theNR1 subunit has been described [Scott et al., 2001]. In

kainate receptors intracellular retention of the KA2 sub-unit is mediated through discrete protein trafficking sig-nals, including an arginine-rich ER retention/retrievalmotif and a di-leucine endocytic sequence in the C ter-minus of the KA2 subunit A [Ren et al., 2003].

In addition, studies have been carried out in thetype B GABA receptors. Thus, a RXR motif present inGABAB receptor GB1 subunits is masked by assembly

Fig. 5. Intracellular localisation of the a1 subunits chimeras

expressed in HEK293 and COS7 cells. HEK293 cells were transfected

with cDNAs encoding the a1 subunit chimeras a1CH (a) and a1CD(b). Cells were fixed and permeabilised in order to visualise the intra-

cellulary expressed homo-oligomers. Cells were labelled with the

anti-a1 subunit-specific monoclonal antibody (bd24) and visualised

with TRITC-conjugated anti-mouse secondary antibody COS7 cells

were transfected with the cDNAs encoding the a1 subunit chimeras:

a1CH (c), a1CD (d), a1CD2 (e) and a1CD3 (f). Cells were fixed and per-

meabilised in order to visualise the intracellular homo-oligomers.

Cells were labelled with the anti-a1 subunit-specific monoclonal anti-

body (bd24) and visualised with TRITC-conjugated anti-mouse sec-

ondary antibody. Scale bars: (a, b) ¼ 10 lm; (c–f) ¼ 20 lm.

Fig. 6. Cell surface expression of a1chimerab3 receptors in COS7

cells. COS7 cells were cotransfected with cDNAs of the a1 subunit

chimeras together with the b3 subunit cDNA. a1chimerab3 complexes

were labelled live with the a1 subunit-specific monoclonal antibody,

bd24 and visualised with cascade blue conjugated anti-mouse second-

ary antibody and with the polyclonal antibody anti-b102/103 for the

b3 subunit and visualised with TRITC-conjugated anti-rabbit second-

ary antibody; (a and b): a1CHb3 complexes; (c and d): a1CDb3 com-

plexes; (e and f): a1CD2b3 complexes; (g and h): a1CD3b3 complexes.

Scale bars for all panels: 20 lm.

754 Peran et al.

with GB2, ensuring heterodimerization [Margeta-Mitrovic et al., 2000; Calver et al., 2001; Pagano et al.,2001; Villemure et al., 2005].

In contrast with these findings, much less is knownabout the mechanisms that restrict the exit of homo-oli-gomeric a1 subunit GABAA receptor complex from theER. In this study we have showed new evidence that theER retention signal of a1 subunits is not containedwithin the M3/M4 cytoplasmic domain of the subunit. Itappears that this region of the protein does not containthe information that dictates whether the subunit isretained or programmed to leave the ER. Deletion of thiscytoplasmic domain did not allow the truncated a1 sub-unit to be expressed at the cell surface, and when by co-expressed with the b3 subunit the a1 subunit chimeraswere able to leave the ER.

Thus, the M3/M4 cytoplasmic loop of the a1 sub-unit appears not to be involved in subunit targeting to thecell surface, subunit-subunit assembly nor receptoraggregation.

We decided to perform our studies in COS7 cellsfor a number of reasons. First, COS7 cells have success-fully been used to study traffic of neuronal membraneproteins by several groups. For example, the membranesorting properties of neurotensin receptors [Perron et al.,2006], ionotropic glutamate receptors [Jaskolski et al.,2005], glycine receptors [Hanus et al., 2004], ATP-sensi-tive potassium channels [Hu et al., 2003], G protein acti-vated inwardly rectified potassium channels [Ma et al.,2002], and metabotropic glutamate receptors [Chanet al., 2001], among others, were studied in COS7 cells.In several of those reports, it was demonstrated that thecellular signals important for traffic in neurons were alsofunctional in COS7 cells. Second, the flat nature ofCOS7 cells makes them specially suited for our experi-ments. Finally, we have previously shown that the a1subunit restricts the lateral mobility of the receptor com-plex in both COS7 and cerebellar granule cells, whichsuggests that the anchoring mechanisms are sharedbetween these two types of cells [Peran et al., 2004].

Removal of the M3/M4 Cytoplasmic Domain of thea1 Subunit Releases the Lateral ConstraintsImposed on GABAAR Mobility

Accumulation or clustering of neurotransmitterreceptors at postsynaptic sites is believed to be crucialfor efficient synaptic transmission. However, the mecha-nism by which GABAARs are delivered to and main-tained at synapses is still poorly understood. In a reviewpresented by Luscher and Keller [2004], GABAAR syn-aptic localization seems to be in part encoded by struc-tural determinants present on a1, a2, a3 and g2 sub-units. Nevertheless, the same receptor subunits are alsoabundant at extrasynaptic sites, suggesting that receptors

move in and out of the postsynaptic membrane. Data oflateral diffusion of GABAAR is still scarce; in this workwe presented novel results from FRAP experiments sup-porting the role of a specific domain within the a1 sub-units in controlling lateral mobility of the receptors. InCOS7 cells transfected with the a1CD subunit lackingthe cytoplasmic loop M3/M4, receptor complexes werealmost freely mobile on the cell surface. In contrast, con-trol chimeric a1CH subunit-containing receptors showedrestricted mobility. This data suggest that deletion of thissequence prevents the formation of links with the cytos-keletal elements that tether the receptors at the cell sur-face. This finding agree with Fritschy et al. [2003] whohypothesized that the sorting and synaptic targeting ofGABAAR is determined by interactions between specificassociated proteins and appropriate sequence motifspresent in particular subunits. The limitation of the lat-eral diffusion of GABAAR could be due to interactionswith the underlying cytoskeletal network, interactionswith others membrane proteins, or interactions with theextracellular matrix.

The fact that not 100% of fluorescence is recoveredafter photobleach even when the M3/M4 region is deletedmight be explained by the mosaicism of the membrane.Single particle tracking data suggested that there are mem-brane microdomains that are not readily accessible for pro-teins to migrate into, therefore the effective area or recov-ery after photobleach might be smaller that the originalbleached area [Simson et al., 1998].

The lateral diffusion of glutamate receptors in andout of synaptic sites has been reported [Borgdorff andChoquet, 2002; Serge et al., 2002; Tardin et al., 2003].However, the mechanisms that control GABAARs mobilityhave not yet been described. Many groups are involved inthe study of proteins that may interact with GABAARs andcluster them to the membrane at precise postsynapticregions. Several candidates such us GABARAP [Chenet al., 2000], Raft1 and Collibistin, Plic1 or dynein lightchain 1 and 2 have been identified [revised in Triller andChoquet, 2003]. Yeast two hybrid and immunoprecipita-tion experiments showed that those proteins directly inter-act with gephyrin or inhibitory receptors.

Gephyrin is the core protein of the scaffold at inhibi-tory synapses and responsible for anchoring of the othermain inhibitory neurotransmitter receptors, the glycinereceptors, [Kirsch et al., 1991; Kirsch and Betz, 1993].This led to the hypothesis that this protein may beinvolved in the anchoring of GABAARs. Although studiessupport the hypothesis that gephyrin might be involved inpostsynaptic positioning of GABAA [Craig et al., 1994;Essrich et al., 1998], there is as yet no clear ‘‘in vivo’’ evi-dence. In fact, the postsynaptic localization of a1subunit-containing GABAARs in cultured hippocampal neuronsand the clustering of these receptors in spinal cord sections

a1 M3/M4 Cytoplasmic Loop Restricts GABAARs Lateral Mobility 755

from gephyrin knockout mice appear unaffected by loss ofgephyrin [Kneussel et al., 2001; Levi et al., 2004]. In addi-tion, GABAA receptors can form clusters in immatureneurons before being detectably colocalized with gephyrin[Dumoulin et al., 2000; Danglot et al., 2003]. Thus, incontrast to glycine receptors, gephyrin does not appear tointeract directly with GABAA receptors. Gephyrin mayfunction by trapping receptors, anchoring receptors orotherwise maintaining synaptic receptor clusters, ratherthan by inserting receptors in the plasma membrane.Experiments showing that disruption of microtubules withcolchicine do not affect gephyrin clustering in mature neu-rons in vitro [Allison et al., 1988; Van Zundert et al.,2002] led to the hypothesis that aggregation of GABAARand gephyrin at synaptic sites reflects a dynamic equilib-rium and is not dependent on stable interaction with thetubulin cytoskeleton.

Receptor clustering by scaffold proteins couldexplain receptor immobility. However, movement ofindividual clusters of receptors with or without interac-tion with scaffolding proteins has been shown in realtime using single particle tracking in cultured neurons[for review see Triller and Choquet, 2003]. In addition,we have found that the lateral mobility of the receptorwas not dependent on receptor clustering. Although thisdata agree with Serge et al. [2002], we did not find thelower diffusion coefficient for clustered receptors thatthey described. It seems then, that the binding of clus-tered receptors to rigid elements such as the cytoskeletonis necessary for immobilization. Taking the data pre-sented in this work into account, we conclude that thisinteraction is mediated by a specific region within the a1subunit of the GABAARs, the cytoplasmic loop M3/M4.

ACKNOWLEDGMENTS

GABAAR cDNA clones were kind gifts from P.Seeburg and H. Luddens.

REFERENCES

Allison DW, Gelfand VI, Spector I, Craig AM. 1988. Role of actin in

anchoring postsynaptic receptors in cultured hippocampal neu-

rons: Differential attachment of NMDA versus AMPA recep-

tors. J Neurosci 18:2423–2436.

Axerold D, Koppe DE, Schlessinger J, Elson EL, Webb WW. 1976.

Mobility measurements and analysis of fluorescence photo-

bleaching recovery kinetics. Biophys J 16:1055–1069.

Bichet D, Cornet V, Geib S, Carlier E, Volsen S, Hoshi T, Mori Y, De

Waard M. 2000. The I-II loop of the Ca2+ channel a1 subunit

contains an endoplasmic reticulum retention signal antagonized

by the b subunit. Neuron 25(1):177–190.

Borgdorff A, Choquet D. 2002. Regulation of AMPA receptor lateral

movement. Nature 417:649–653.

Calver AR, Robbins MJ, Cosio C, Rice SQ, Babbs AJ, Hirst WD,

Boyfield I, Wood MD, Russell RB, Price GW. 2001. The C-ter-

minal domains of the GABA(b) receptor subunits mediate in-

tracellular trafficking but are not required for receptor signal-

ling. J Neurosci 21:1203–1210.

Chan WY, Soloviev MM, Ciruela F, McIlhinney RA. 2001. Molecular

determinants of metabotropic glutamate receptor 1B traffick-

ing. Mol Cell Neurosci 17(3):577–588.

Chen L, Wang H, Vicini S, Olsen RS. 2000. The aminobutyric acid

type A (GABAA) receptor-associated protein (GABARAP)

promotes GABAA receptor clustering and modulates the chan-

nel kinetics. Proc Natl Acad Sci USA 97(21):11557–11562.

Connor JL, Boileau AJ, Czajkowski C. 1998. A GABAA receptor asubunit tagged with green fluorescent protein requires a b sub-

unit for functional surface expression. J Biol Chem 273:

28906–28911.

Craig AM, Blackstone CD, Huganir RL, Banker G. 1994. Selective

clustering of glutamate and g-aminobutyric acid receptors op-

posite terminals releasing the corresponding neurotransmitters.

Proc Natl Acad Sci USA 91:12373–12377.

Dahan M, Levi S, Luccardini C, Rostaing P, Riveau B, Triller A. 2003.

Diffusion dynamics of glycine receptors revealed by single-quan-

tum dot tracking. Science 302:442–445.

Danglot L, Triller A, Bessis A. 2003. Association of gephyrin with

synaptic and extrasynaptic GABAA receptors varies during de-

velopment in cultured hippocampal neurons. Mol Cell Neuro-

sci 23:264–278.

Dumoulin A, Levi S, Riveau B, Gasnier B, Triller A. 2000. Formation

of mixed glycine and GABAergic synapses in cultured spinal

cord neurons. Eur J Neurosci 12:3883–3892.

Essrich C, Lorez M, Benson JA, Fritschyt J-M, Luscher B. 1998. Post-

synaptic clustering of major GABAA receptor subtypes

requires the g2 subunit and gephyrin. Nat Neurosci 1:563–571.Fritschy JM, Schweizer C, Brunig I, Luscher B. 2003. Pre- and post-

synaptic mechanisms regulating the clustering of type A g-ami-

nobutyric acid receptors (GABAA receptors). Biochem Soc

Trans 31:889–892.

Hanus C, Vannier C, Triller A. 2004. Intracellular association of gly-

cine receptor with gephyrin increases its plasma membrane

accumulation rate. J Neurosci 24(5):1119–1128.

Hu K, Huang CS, Jan YN, Jan LY. 2003. ATP-sensitive potassium

channel traffic regulation by adenosine and protein kinase C.

Neuron 38(3):417–432.

Jacobson K, Elson E, Koppel DE, Webb WW. 1983. International

workshop on the application of fluorescence photobleaching

techniques to problems in cell biology. Fed Proc 42:72–79.

Jaskolski F, Normand E, Mulle C, Coussen F. 2005. Differential

trafficking of GluR7 kainate receptor subunit splice variants.

J Biol Chem 280(24):22968–22976.

Kirsch J, Betz H. 1993. Widespread expression of gephyrin, a puta-

tive glycine receptor tubulin linker protein, in rat brain. Brain

Res 621:301–310.

Kirsch J, Langoshch D, Prior P, Littauer UZ, Schmitt B, Betz H.

1991. The 93kDa glycine receptor-associated protein binds to

tubulin. J Biol Chem 266:2242–2245.

Kneussel M, Brandstatter JH, Gasnier B, Feng G, Sanes JR, Betz H.

2001. Gephyrin-independent clustering of postsynaptic GABAA

receptor subtypes. Mol Cell Neurosci 17:973–982.

Koksma VJ, Fritschy JM, Mack V, Kestere REV, Brussaard AB. 2005.

Differential GABAA receptor clustering determines GABA syn-

apse plasticity in rat oxytocin neurons around parturition and the

onset of lactation. Mol Cell Neurosci 28:129–140.

Kullman DM, Ruiz A, Rusakov DM, Scott R, Semyanov A, Walker

CW. 2005. Presynaptic, extrasynaptic and axonal GABAA

receptors in the CNS: Where and why? Prog Biophys Mol Biol

87:33–46.

756 Peran et al.

Levi S, Logan SM, Tovar KR, Craig AM. 2004. Gephyrin is critical

for glycine receptor clustering but not for the formation of

functional GABAergic synapses in hippocampal neurons.

J Neurosci 24:207–217.

Luscher B, Keller CA. 2004. Regulation of GABAA receptors traffick-

ing, channel activity, and functional plasticity of inhibitory

synapses. Pharmacol Ther 102(3):195–221.

Ma D, Zerangue N, Raab-Graham K, Fried SR, Jan YN, Jan LY.

2002. Diverse trafficking patterns due to multiple traffic motifs

in G protein-activated inwardly rectifying potassium channels

from brain and heart. Neuron 33(5):715–729.

Margeta-Mitrovic M, Jan YN, Jan LY. 2000. A trafficking checkpoint

controls GABA(B) receptor heterodimerization. Neuron 27(1):

97–106.

Nusser Z, Hajos N, Somogyi P, Mody I. 1998. Increased number of

synaptic GABAA receptors underlies potentiation at hippocam-

pal inhibitory synapses. Nature 395:172–177.

Pagano A, Rovelli G, Mosbacher J, Lohmann T, Duthey B, Stauffer

D, Risting D, Schuler V, Meigel I, Lampert C, Stein T, Prezeau

L, Blahos J, Pin JP, Froestl W, Kuhn R, Heid J, Kaupmann K,

Bettler B. 2001. C-terminal interaction is essential for surface

trafficking but not for heteromeric assembly of GABAB recep-

tors. J Neurosci 21(4):1189–1202.

Peran M, Hicks BW, Peterson NL, Hooper HT, Salas R. 2001. Lateral

mobility and anchoring of recombinant GABAA receptors depend

on subunit composition. Cell Motil Cytoskeleton 50: 89–100.

Peran M, Hooper HT, Rayner SH, Stephenson FA, Salas R. 2004.

Subunit specificity in anchoring of recombinant GABAA

receptors. Neurosci Lett 364:67–70.

Perron A, Sharif N, Gendron L, Lavallee M, Stroh T, Mazella J,

Beaudet A. 2006. Sustained neurotensin exposure promotes

cell surface recruitment of NTS2 receptors. Biochem Biophys

Res Commun 343(3):799–808.

Reits EAJ, Neefjes JJ. 2001. From fixed to FRAP: Measuring protein

mobility and activity in living cells. Nat Cell Biol 3(6):145–147.

Ren Z, Riley NJ, Garcia EP, Sanders JM, Swanson GT, Marshall J.

2003. Multiple trafficking signals regulate kainate receptor

KA2 subunit surface expression. J Neurosci 23(16):6608–

6616.

Sakatani K, Chesler H, Hassan AZ. 1991. GABAA receptors modulate

axonal conduction in dorsal columns of neonatal rat spinal

cord. Brain Res 542:273–279.

Scott DB, Blanpied TA, Swanson GT, Zhang C, Ehlers MD. 2001. An

NMDA receptor ER retention signal regulated by phosphoryla-

tion and alternative splicing. J Neurosci 21(9):3063–3072.

Serge A, Fourgeaud L, Hemar A, Choquet D. 2002. Receptor activa-

tion and homer differentially control the lateral mobility of

metabotropic glutamate receptor 5 in the neuronal membrane.

J Neurosci 22(10):3910–3920.

Sieghart W. 2000. Unraveling the function of GABA(A) receptors

subtypes. Trends Pharmacol Sci 21:411–413.

Simson R, Yang B, Moore SE, Doherty P, Walsh FS, Jacobson KA.

1998. Structural mosaicism on the submicron scale in the

plasma membrane. Biophys J 74:297–308.

Tardin C, Cognet L, Bats C, Lounis B, Choquet D. 2003. Direct imag-

ing of lateral movements of AMPA receptors inside synapses.

EMBO J 22:4656–4665.

Triller A, Choquet D. 2003. Synaptic structure and diffusion dynamics

of synaptic receptors. Biol Cell 95:465–467.

Van Zundert B, Alvarez FJ, Yevenes GE, Carcamo JG, Vera JC,

Aguayo LG. 2002. Glycine receptors involved in synaptic

transmission are selectively regulated by the cytoskeleton in

mouse spinal neurons. J Neurophysiol 87:640–644.

Villemure JF, Adam L, Bevan NJ, Gearing K, Chenier S, Bouvier M.

2005. Sub-cellular distribution of GABAB receptors homo-

and heterodimers. Biochem J 388:47–55.

Zerangue N, Schawappach B, Jan LY. 1999. A new ER trafficking sig-

nal regulates the subunit stoichiometry of plasma membrane

(ATP) channels. Neuron 22:537–548.

a1 M3/M4 Cytoplasmic Loop Restricts GABAARs Lateral Mobility 757