JBC Origene Your Gene Company

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M600679200 on June 22, 2006

J. Biol. Chem., Vol. 281, Issue 33, 23908-23921, August 18, 2006
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
281/33/23908    most recent
M600679200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bedoukian, M. A.
Right arrow Articles by Partin, K. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bedoukian, M. A.
Right arrow Articles by Partin, K. M.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Different Domains of the AMPA Receptor Direct Stargazin-mediated Trafficking and Stargazin-mediated Modulation of Kinetics*Formula

Matthew A. Bedoukian, Autumn M. Weeks, and Kathryn M. Partin1

From the Department of Biomedical Sciences, Colorado State University, Fort Collins, Colorado 80523

Received for publication, January 23, 2006 , and in revised form, June 14, 2006.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Stargazin is an accessory protein of AMPA receptors that enhances surface expression and also affects the biophysical properties of the receptor. AMPA receptor domains necessary for either of these two processes have not yet been identified. Here, we used confocal imaging and electrophysiology of heterologously expressed, fluorophore-tagged GluR1, GluR2, and stargazin to study surface expression and desensitization kinetics. Stargazin-mediated trafficking was sensitive to the nature of the AMPA receptor cytoplasmic domain. The insertion of YFP after residue 15 of the truncated cytoplasmic tail of GluR1i perturbed stargazin-mediated trafficking of the receptor but not its modulation of desensitization kinetics. This construct also failed to permit fluorescence resonance energy transfer (FRET) with stargazin in the endoplasmic reticulum (ER), whereas FRET between fluorophore-tagged stargazin and non-truncated AMPA receptors demonstrated a specific interaction between these proteins, both in the ER and the plasma membrane. Rather than encoding a specific binding site, the fluorophore-tagged C terminus may restrict access to one or more ER retention sites. Although perturbations of the C terminus impeded stargazin-mediated trafficking to the plasma membrane, the effects of stargazin on the biophysical properties of AMPA receptors (i.e. modulation of desensitization) remained intact. These data provide strong evidence that the AMPA receptor domains required for stargazin modulation of gating and trafficking are separable.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
AMPA2 ({alpha}-amino-3-hydroxy-5-methyl-isoxazole-4-propionate) receptors, a subtype of ionotropic glutamate receptors, are expressed at the postsynaptic membrane of neurons where they mediate rapid excitatory synaptic transmission (1-4). Native AMPA receptors are hetero-oligomers composed of four subunits (GluR1-4) that are either a flip (i) or flop (o) isoform (5). AMPA receptors play a critical role in neuronal signal transduction that is necessary for memory and learning. AMPA receptors cycle rapidly in and out of the plasma membrane in an activity-dependent manner (6-9) that requires assembly with auxiliary proteins such as stargazin (10).

Stargazin, also known as {gamma}-2 or CACNG2, is a member of the transmembrane AMPA receptor regulatory protein (TARP) family (11-14). It was initially identified from the stargazer mouse, an inbred mouse strain with a phenotype of an unsteady gait, persistent head-raising ("stargazing"), and frequent spike-wave discharges (12, 15). Granule cells from the cerebellum of stargazer mice are missing functional AMPA receptors. A biochemical interaction between stargazin and both AMPA receptors and PSD-95 exists (16). The interaction between stargazin and AMPA receptors is essential for efficient delivery of receptors to the surface of cerebellar granule cells, whereas its interaction with PSD-95 is essential for clustering receptors to the postsynaptic membrane.

Stargazin enhances the total current of AMPA receptors, consistent with its ability to traffic more receptors to the plasma membrane where current is measured (17-21). Recombinant AMPA receptors are poorly trafficked to the cell surface in the absence of stargazin and remain trapped in intracellular pools (22). Stargazin may facilitate AMPA receptor export from the ER (11) by masking ER retention signals of the tetrameric receptor in vivo (10). In addition to enhancing AMPA receptor trafficking, stargazin also slows AMPA receptor desensitization and deactivation (19-21) and increases channel opening (20). The domains of stargazin essential for modulating trafficking versus biophysical properties are partially separable. Stargazin is a four-pass transmembrane protein: the cytoplasmic C-terminal domain is required for receptor trafficking but the first extracellular domain controls stargazin modulation of AMPA receptor biophysical properties (20, 21). The stargazin extracellular domain may allosterically modulate the AMPA receptor extracellular ligand binding core, altering AMPA receptor subunit interactions (14). Single particle electron microscopy indicates that stargazin associates primarily with the AMPA receptor transmembrane domains (23, 24). However, there is little information about AMPA receptor domains involved in stargazin-mediated trafficking or modulation.

The present study uses functional deletions of AMPA receptors to identify domains necessary for effective trafficking and modulation of desensitization by stargazin. Because the intracellular C terminus of stargazin is necessary for targeting AMPA receptors, we hypothesized that an intracellular region of the AMPA receptor such as the C terminus might directly interact with stargazin. Consistent with this idea, the C termini of AMPA receptors are known to bind a number of different proteins that affect trafficking and the stabilization of the channel at synapses (supplemental Fig. S1). Our data suggest that stargazin requires access to a cytoplasmic binding site for effective trafficking to the surface membrane but does not require this same interaction for modulation of desensitization.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Transient Transfections for Electrophysiology—Human embryonic kidney 293 (HEK-293) fibroblasts (CRL 1573; American Type Culture Collection) were cultured as described previously (25). Cells were transiently transfected using FuGENE 6 reagent (Roche Applied Science) or Lipofectamine 2000 (Invitrogen, Carlsbad, CA) with AMPA receptor cDNA (0.5-2 µg/35-mm dish) and if the channel was not tagged, soluble yellow fluorescent protein cDNA (0.1-0.15 µg/35-mm dish). When used, stargazin was always added in a 1:2 stargazin ratio (with total amounts of cDNA transfected ranging from 0.1 to 3 µg). After transfection, 10-40 µM NBQX was added to the medium to prevent cell toxicity.

Transfections for Confocal Microscopy—Collagen- or poly-D-lysine-coated 14-mm glass bottom culture dishes (MatTek Corp., Ashland, MA) were incubated with ECL Attachment Matrix (Upstate Cell Signaling Solutions, Lake Placid, NY) for 1 h at 37 °C then washed with cMEM before plating cells. Cells were transfected using Lipofectamine 2000 (Invitrogen) when 60-90% confluent and incubated under identical conditions as cells used for standard electrophysiology. For each transfection 70 µl of MEM was incubated with 3 µl of Lipofectamine and in another tube 30 µl of MEM was incubated with 0.1-3 µg of total cDNA and thoroughly mixed. Contents of the tubes were combined, and after 20-30 min the solution was added to the cells along with 0-10 µM NBQX. 4.5-24 h later, the solution was exchanged with fresh cMEM with 20-40 µM NBQX. All cells were imaged at room temperature 2-3 days after transfection. Immediately before imaging the solution was exchanged with cMEM containing no phenol red.

Outside-out Patch Recordings—Currents were recorded 2-3 days after transfection, as described previously (25). Extracellular solutions (ECS) contained the following: 20 mM sucrose, 145 mM NaCl, 5.4 mM KCl, 5 mM HEPES, 1 mM MgCl2, 1.8 mM CaCl2·H2O, and 0.01 mg/ml phenol red, pH 7.3. Outside-out membrane patches were voltage-clamped at -60 mV using an Axopatch 200B amplifier (Molecular Devices, Union City, CA). Synapse (version 3.6d; Synergy Research, Silver Spring, MD) controlled piezoelectric movement, data acquisition, and trace analysis. Responses were filtered at 5 kHz, digitized at 10-500 µs/point, and stored on a Power Macintosh computer (Apple Computers, Cupertino, CA) using an ITC-16 interface (InstruTech, Port Washington, NY). Micropipettes (TW150F; 2-5 M{Omega}; World Precision Instruments, Sarasota, FL) contained the following (in mM): 135 CsCl, 10 CsF, 10 HEPES, 5 Cs-BAPTA, 1 MgCl2, and 0.5 CaCl2, pH 7.2 (292 mOsm). Patches were perfused at 0.2 ml/min with solutions emitted from a two-barrel flow pipe made with {theta} tubing (BT150-10; Sutter Instruments, Novato, CA). One barrel contained vehicle (control) composed of the following: 145 mM NaCl, 5.4 mM KCl, 5mM HEPES, 1 mM MgCl2, 1.8 mM CaCl2·H2O, with 0.01 mg/ml phenol red, pH 7.3. The other barrel had this solution plus L-glutamate (10 mM). After going into voltage clamp, an outside-out patch was pulled, lifted up to the flow pipe, positioned near the interface between the glutamate-free and glutamate-containing solution and jumped rapidly from the vehicle control into glutamate. Rapid solution exchanges of 1 or 500 ms were driven by a piezoelectric device (Burleigh Instruments, Fishers, NY). Solution exchange rates were determined at the end of each experiment by open-tip junction currents and excluded if rise times exceeded 0.5 ms.

Analysis of Rapid Responses—Desensitization rates were estimated by fitting a single exponential function ({tau}des) to the 500-ms response decay (from 95% of peak to steady state). Deactivation rates were estimated by fitting a single exponential function ({tau}deact) to the 1-ms response decay (from 95% of peak to steady state). 3-20 responses per patch were averaged for analysis. Current traces and graphs were plotted using KaleidaGraph 3.5 (Synergy Software, Reading, PA).

Generation of Constructs—The CMV expression plasmids (pRK) for GluR1, GluR2 (R607Q) and GluR6 (R621Q) were provided by Dr. Peter Seeburg (Max Planck Institute for Medical Research, Heidelberg, Germany). R181YFP was generated using overlapping PCR to make an in-frame fusion protein with CFP or YFP, using pECFP or pEYFP (Clontech, Palo Alto, CA) as templates. The first residue from the fluorescent protein followed immediately after the last amino acid in the sequence ATGL. R146YFP was made by inserting the restriction site Mlu1 (ACGCGT) between amino acids GGG and SGE of the C-terminal domain using QuikChange II XL site-directed mutagenesis kit (Stratagene, La Jolla, CA). The final sequence at the fusion was GGGTR(YFP). R136YFP, R115YFP, R17YFP, R12YFP, R246YFP, and R216YFP were generated similarly (see supplemental Fig. S1 for the location of insertion sites). All point mutations were made using the QuikChange II XL Site-directed mutagenesis kit as was the insertion of a stop codon for R1i14 after amino acids KRMK of the cytoplasmic tail. The insertion sequence of the R1i15+38YFP, R1i7+38YFP, and R1i2+38YFP constructs had a 38-amino acid linker before fluorophore attachment with the sequence: TRGGSEQKLISEEDLSQFRVSPLDRTWNLGETVELKTR. GluR2i and GluR2o {Delta}ATD were constructed to delete 380 amino acids of the mature protein. The N terminus thus began with LPS preceded by the Kozak sequence (ACC) and the GluR6 signal sequence. Stargazin constructs: StgEGFP/Gw1-CMV plasmid (British Biotechnology; Ref. 26) was a generous gift from Dr. David Bredt (UCSF, San Francisco), as was pcDNA3-stg. StgCFP and stgYFP were inserted in frame at the BglII site in pcDNA3 (Stratagene) homologous to stgGFP but truncated after amino acid 269. GluR4i {Delta}ATD and FLAG-GluR4i were kind gifts from Dr. Kari Keinanen (University of Helsinki; Helsinki, Finland; Ref. 27) FLAG-GluR4i {Delta}CTD was made by inserting a stop codon after EF, the first two amino acids of the cytoplasmic tail. Kv2.1 was a kind gift from Dr. Michael Tamkun (Colorado State University; Fort Collins, CO, Ref. 28). The CD8 plasmid (MGC-34614) was purchased from American Type Culture Collection, and CD3CFP was a generous gift from Dr. Nicholas Gascoigne (The Scripps Research Institute; La Jolla, CA, Ref. 29). DNA mutations were confirmed by sequencing (Macromolecular Resources, Fort Collins, CO).

Measurements of FRET—Transfected cells were imaged using an LSM 510 META laser scanning confocal microscope (Zeiss, Thornwood, NY) with a Plan-Achromat x63/1.4 oil DIC objective (30). CFP and YFP were excited with separate sweeps of the 458- and 514-nm lines, respectively, of an argon laser operated at 6.3-6.7 A, attenuated to 5 and 2% respectively, and directed to the cell via a 458/514-nm dual dichroic mirror. Airy units for imaging were between 1 and 2 for CFP and always 1 for YFP. The emitted cyan fluorescence was directed to a photomultiplier with a 460-500-nm bandpass filter and yellow fluorescence was directed to a photomultiplier equipped with a 530-nm long-pass filter. Confocal fluorescence intensity data (ICFPpre and IYFPpre) were recorded as planar line scans digitized at 8 or 12-bits. Repeated scans (20-50) with unattenuated 514-nm illumination to photobleach YFP, which required ~15-45 s at maximal scan rates, were used to photobleach YFP. After completion of YFP bleaching, fluorescence intensity (ICFPpost and IYFPpost) was measured using the identical parameters as before bleaching. FRET efficiency (E) was calculated as 1-(ICFPpre/ICFPpost), where ICFPpre and ICFPpost are the background-corrected CFP fluorescence intensities before and after photobleaching YFP, respectively. Membrane FRET measurements were taken from cells illustrating an obvious increase in fluorescence intensity around the cell perimeter relative to the cytosol. Whole cell measurements were taken from visible fluorescence from the entire cell (minus the nucleus), whereas cytosolic measurements included the whole cell without obvious plasma membrane.

Confocal Imaging of GFP and Fluorescent Microspheres—Cells transfected with stgGFP were excited using the 488-nm line of an argon laser attenuated to 6%. Emission was collected using the meta filter set to capture between 508 and 551 nm. Yellow-green biotin-labeled 0.04-µM beads (Fluospheres, Molecular Probes) were imaged at the bottom of an ECL-coated cell dish at identical settings. Determination of protein at the cell surface was calculated as described previously (31).

Immunofluorescence for Confocal Imaging—After transfection as described above for 3 days with 0.3 µg of FLAG-GluR4i cDNA with or without 0.6 µg of stg, cells were rinsed with phosphate-buffered saline then fixed with 4% paraformaldehyde for 15-20 min. Cells were blocked for 1 h with 1% bovine serum albumin and 3% normal goat serum in phosphate-buffered saline. A 1:9000 dilution of anti-FLAG M2 monoclonal antibody (Sigma) was applied for 1 h then washed 3 x 10 min in blocking solution before applying 1:700 goat anti-mouse Alexa Fluor 568 (Molecular Probes) for 1 h. Cells were again washed 3 x 10 min in blocking solution then rinsed several times in phosphate-buffered saline before imaging with an Olympus FV1000 confocal microscope. All staining and imaging were done at room temperature.

Quantitation of Surface Expression—Transfected HEK-293 cells were visualized using a confocal microscope and divided into 3 categories (where a "ring" is defined as markedly enhanced surface expression, which if imaged as an optical slice with a confocal microscope, appears as enhanced fluorescent intensity co-localizing with the plasma membrane): definitely rings, definitely no rings, and unclassifiable. Unclassifiable cells were represented by, for example, cells with very low expression or hints of enhanced surface expression in only a very small portion of the surface membrane. Only 4% of over a hundred cells co-transfected with R1i15YFP and stargazin were deemed unclassifiable, whereas 14% of over two hundred cells co-transfected with R1i81YFP and stargazin were not classifiable. To ensure non-biased quantification, a blind observer was trained to detect rings from non-rings and did not identify a single cell co-transfected with R1i15YFP and stargazin that had a surface expression ring. Statistical significance was determined with an unpaired Student's t test. Data are reported as mean ± S.E.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Stargazin Alters GluR1i Distribution in the Cytosol and the Plasma Membrane—To assay stargazin-mediated trafficking of AMPA receptors, in-frame chimeric proteins were constructed by fusing a fluorescent protein (ECFP or EYFP) to the C-terminal domain of either GluR1 (R1) or GluR2 (R2). Functional channels were expressed heterologously in HEK-293 cells then imaged using a confocal microscope. Effects on trafficking were assayed by determining subcellular localization and, in particular, whether or not the protein could be detected at the plasma membrane surface. Confocal imaging through live cells that expressed fluorophore-tagged AMPA receptors allowed us to distinguish between accumulation of protein in reticular, perinuclear networks (presumably colocalized with the ER) or within the plasma membrane, such that the labeled protein formed prominent surface rings.

Full-length GluR1flip (R1i) with YFP fused to its C terminus (R1i81YFP) was heterologously expressed with and without stargazin. This construct yielded a fully functional ion channel (Fig. 1). Although currents may be recorded from this construct, confocal imaging of R1i81YFP demonstrated uniformly distributed fluorescence, typical of ER-associated expression (Fig. 1b). In contrast, expression of fluorophore-tagged stargazin (stgYFP) alone resulted in the formation of pronounced surface rings in virtually every cell. When R1i81YFP was co-transfected with stargazin, the plasma membrane intensity relative to the cytosolic intensity was markedly increased in a population of cells (Fig. 1b and Table 1), because of decreased cytosolic expression of R1i81YFP. The population of fluorescent cells with pronounced surface expression was quite variable (~5-35%) across more than ten independent transfections (using a 1:2, receptor:stg ratio in all cases). Nevertheless, increased amounts of stargazin always resulted in an increased percentage of R1i81YFP-expressing cells with rings (Fig. 1c, inset). However, in the absence of stargazin, surface rings of R1i81YFP were never observed at any cDNA concentration used (0.1-2 µg). The ability to generate pronounced surface rings in all cells was not because of poor transfection efficiencies, in that ≥96% of all cells (n>100) that contained R1i81YFP also expressed the fluorophore-tagged stargazin.


View this table:
[in this window]
[in a new window]
  		TABLE 1
C-terminal mutations of GluR1 and GluR2 alter stargazin-mediated trafficking

The constructs imaged with confocal microscopy after co-expression with excess stargazin are shown in the left column, the corresponding percentage of cells forming stargazin-mediated rings is shown in the right column. The raw data (actual number of cells with rings/total cells counted) for each construct are shown in parentheses.

 
Stargazin has a dual effect on AMPA receptors: it traffics them to the surface and also alters their biophysical properties. Specifically, stargazin slows receptor deactivation and desensitization, as well as increasing open probability and current amplitude (17-21). We therefore assessed whether stargazin could modulate the biophysical properties of both populations of co-transfected cells (with and without ER retention). Outside-out membrane patches from cells co-transfected with R1i81YFP and stargazin, either with or without pronounced surface expression, demonstrated a significant slowing of receptor desensitization compared with R1i81YFP in the absence of stargazin (Fig. 1c). As expected, cells co-transfected with R1i81YFP and stargazin with surface rings had current amplitudes about 6-fold greater than without stargazin (1507 ± 426 pA).


Figure 1
View larger version (44K):
[in this window]
[in a new window]
  		FIGURE 1.
Stargazin traffics fluorophore-tagged GluR1 to the plasma membrane. a, topology of the AMPA receptor and stg proteins showing the site of fluorophore insertions at the C termini. Dashed line indicates C-terminal truncation of fluorophore-tagged stargazin. b, confocal images of yellow fluorescence in HEK-293 cells expressing (R1i81YFP, left) (stgYFP, center), and R1i81YFP co-expressed with stg (right). Profile intensities (in arbitrary units) along the red line demonstrate that surface expression of GluR1 is markedly enhanced by stargazin. c, time constants of desensitization (black bars) measured as the decay in response to a 500-ms pulse of 10 mM glutamate, or deactivation (gray bars) measured as the decay in response to a 1-ms pulse of 10 mM glutamate, for R1i81YFP in the absence or presence of stargazin (left). Cells co-expressing R1i81YFP and stargazin were visually scored as either not having pronounced surface expression (- rings) or having pronounced surface expression (+ rings). Mean current amplitude measured in response to a 500-ms pulse of 10 mM glutamate in the absence or presence of stargazin, with and without surface expression rings (right). (*, p < .05; **, p < .01 comparing stargazin + or - rings to without stargazin; ##, p < .02 comparing R1i81YFP deactivation with stargazin + or - visible surface expression rings). Inset shows ring formation as a function of increasing concentrations of stargazin cDNA (filled circles), co-transfected with a constant amount (0.2 µg) of R1i81YFP. The solid line represents a curve fit with to a logarithmic function extrapolated to 0, R2 = 0.91.

 
In addition, the desensitization kinetics were significantly different between these two cell populations. In the absence of stargazin, GluR1i desensitized with a decay time constant of 3.7 ± 0.4 ms, whereas stargazin-transfected cells with rings had desensitization kinetics (8.5 ± 0.8 ms) that were only slightly slower than stargazin-transfected cells without rings (7.1 ± 1.1 ms). Deactivation kinetics followed this same trend, with decay kinetics in the absence of stargazin (0.7 ± 0.1 ms) slowed to 1.0 ± 0.2 ms versus 2.6 ± 0.3 ms in cells with pronounced stargazin-mediated surface expression (Fig. 1c).

An Intracellular Interaction between Stargazin and AMPA Receptors Is Necessary for Up-regulation of Surface Expression—To map the domains of AMPA receptors essential for stargazin-mediated trafficking, a series of receptor deletion proteins (fused to fluorophores) was constructed. To determine the role of the AMPA receptor C terminus, R1i and R2i C-terminal truncations were made (Figs. 2a and 4a). As seen for R1i81YFP, R1i46YFP, and R1i36YFP expression was mostly cytosolic in the absence of stargazin. However, these receptors could form fluorescent membrane surface rings when co-expressed with stargazin. No decrease in the frequency of surface rings or differences in the ratio of membrane to cytosolic intensity were found for these truncated proteins compared with the wild-type C terminus. However, a third fusion protein that deleted all but the first 15 amino acids of the C terminus, R1i15YFP, resulted in a receptor that never formed surface rings (10 independent transfections with >1000 counted fluorescent cells). Rings also did not form when R1i15YFP was co-expressed with stgCFP (Fig. 2b). Rings were not seen even at a 1:8 stargazin ratio.

The homologous deletion made in GluR2 (R2i16) tagged with CFP or YFP was also tested for stargazin-mediated trafficking. Similar to R1i15YFP, the expression of this protein was cytosolic in either the absence or presence of stargazin. In contrast, the C-terminal deletion mutation R2i46YFP also formed surface expression rings (Fig. 4a). Thus, the cytoplasmic requirements for trafficking of GluR1 or GluR2 by stargazin appear to be similar. All flop isoforms tested followed the same trend.

The Intracellular Interaction Necessary for Effective Stargazin-mediated Trafficking Is Not Necessary for Stargazin-mediated Modulation of Desensitization—The inability of stargazin to traffic R115YFP or R216YFP to the surface membrane suggests that stargazin cannot interact with these proteins. To determine whether trunctions that impair trafficking also impair modulation of AMPA receptor function, we examined the electrophysiological properties of the mutated receptors. Outside-out patches of cells transfected with the C-terminal truncations, R1i15YFP or R1i36YFP, demonstrated no significant difference in stargazin-mediated slowing of desensitization compared with the non-truncated R1i81YFP (Fig. 2, c and e). Stargazin slowed the kinetics of R1i81YFP from 3.2 ± 0.1 to 5.8 ± 0.7 ms, R1i36YFP from 3.4 ± 0.5 to 4.8 ± 0.3 ms, and R1i15YFP from 2.9 ± 0.1 to 5.8 ± 0.3 ms. It is important to note that cells with surface expression rings were not selected for in these experiments and are not paired with the data used for Fig. 1.

As expected if the formation of surface expression (and reduced ER retention) is related to overall current density, R1i15YFP showed only a 2-fold increase in current amplitude when co-expressed with stargazin (370 ± 113 versus 733 ± 175 pA), whereas both R1i81YFP (198 ± 69 versus, 694 ± 301 pA) and R1i36YFP (172 ± 82 versus, 630 ± 140 pA) showed a ~3.5-fold increase in current amplitude (Fig. 2d). R1i14, which lacked a fluorophore and had one less amino acid than R1i15YFP, produced currents that were significantly smaller 30 ± 3 pA, but when co-transfected with stargazin the mean current increased ~11-fold to 364 ± 118 pA. Together, the data presented in Figs. 1 and 2 suggest that the domains of AMPA receptors necessary for trafficking of or modulation by stargazin are separable.

The Cytoplasmic Tail of AMPA Receptors Does Not Contain a Specific Stargazin Binding Site—Stargazin did not traffic R115YFP to the surface. To determine if this was caused by steric hindrance when the fluorophore was attached immediately after residue 15, we inserted a 38 amino acid linker between the last AMPA receptor residue of R1i15YFP and the fluorophore (R1i15+38YFP). Co-transfection of R1i15+38YFP with stargazin permitted ring formation, demonstrating that no more than 15 amino acids of the cytoplasmic tail of AMPA receptors were necessary for stargazin-mediated trafficking. To determine if there is a specific interaction between stargazin and any of the residues within the conserved proximal 14 amino acids in the tail of AMPA receptors, a series of point mutations was constructed. These mutations either disrupt a known protein-interacting site (supplemental Fig. S1) or replace critical differences in the proximal cytoplasmic tail between AMPA receptors and GluR6 (Fig. 3a), which does not associate with stargazin (18). All point mutants were constructed within R1i81YFP or R1i36YFP. Each of these mutant receptors permitted stargazin-mediated surface expression rings (Fig. 3a) suggesting that specific C-terminal residues are not necessary for stargazin trafficking. Because previous studies have shown that an AMPA receptor without a C terminus has virtually no surface expression (32), we constructed an extreme C-terminal truncation, R1i2YFP as a negative control. As expected, R1i2YFP did not form surface expression rings either in the absence or presence of stargazin. However, when the same 38-amino acid linker was added between the receptor and the fluorophore (R1i2+38YFP), surface expression rings formed for a small number of cells (<1%) co-transfected with stargazin (Table 1). The failure to see more surface expression rings may be caused by the already diminished surface expression that results from this impaired receptor.

The experiments with fluorophore-tagged receptor mutants led us to question whether the cytoplasmic tail of AMPA receptors contains a specific stargazin binding site or whether the tail can in some cases interfere with another stargazin binding site present at the cytoplasmic face of the receptor. To demonstrate that stargazin does not require specific determinants on the cytoplasmic tail, we used another assay for surface expression, namely immunofluorescence of unpermeabilized cells transfected with an extracellularly FLAG-tagged receptor. FLAG-R4i2 (with only the first 2 amino acids of the cytoplasmic tail of GluR4i (32)) was co-transfected with stargazin. Although surface expression levels of this construct without stargazin were almost undetectable, stargazin did increase the surface expression (Fig. 3b) but not to the levels of full-length FLAG-R4i (with or without stargazin). This experiment verified through an independent method that nearly all of the C terminus is dispensable for stargazin-mediated trafficking of AMPA receptors, as long as there is no fluorophore attached.


Figure 2
View larger version (43K):
[in this window]
[in a new window]
  		FIGURE 2.
A cytoplasmic interaction between stargazin and GluR1 promotes stargazin modulation of trafficking but is not essential for stargazin modulation of desensitization. a, fluorophore-tagged C-terminal deletions of GluR1 (left); + or - indicate whether co-expression with stargazin increased surface expression (right). Two constructs (R1i2+38YFP and R1i15+38YFP) have an insertion of a 38-amino acid linker between the receptor and the fluorophore (thin line). Note that R1i14 does not contain a fluorophore tag. b, confocal images and corresponding profile intensities of yellow fluorescence from R1i truncations co-expressed with stgCFP. c, electrophysiological response of R1i truncations to a 500-ms pulse of 10 mM glutamate (black line is response without stargazin, gray line is response with stargazin. Mean current amplitudes plotted on a log scale (d) and desensitization rates (e) of tagged R1i81YFP, R1i36YFP, and R1i15YFP, and R1i14 expressed with (gray bars) and without (black bars) stargazin. (*, p < .05; **, p < .01; ***, p < .005.)

 
Mutations of ER Retention Signal Residues Enhance Trafficking without and with Stargazin—Stargazin may enhance AMPA receptor trafficking by masking an ER retention signal (10). Although we found that stargazin could enhance the surface expression of both the R4i2 and R1i2+38YFP constructs, the fact that stargazin did not fully rescue surface expression is likely attributed to the importance of the proximal cytoplasmic tail residues for AMPA receptor expression. To rule out that this lack of complete rescue was not because of an impaired ability of stargazin association with specific residues of the cytoplasmic tail, we tested R1i7+38YFP (with the third of seven amino acids being Leu instead of Cys, so all seven residues would be identical to the first seven amino acids of the untraffickable GluR6 tail) (Fig. 3a). This construct has the 38-amino acid linker between the seventh amino acid and the fluorophore. Although this receptor should not interact with protein 4.1, shown to be important for AMPA receptor surface expression (32, 33), in one experiment we found that 28% of all fluorescent cells had pronounced stargazin-mediated surface expression rings, comparable to wild-type R1i81YFP. In addition, residues 4-6 of the cytoplasmic tail (implicated in ER retention, Ref. 34) were mutated from YKS to FQA (Fig. 3a), using R1i81YFP as a template. Even without stargazin, surface expression rings were formed in ~2% of the cells, something we had never observed for R1i81YFP. Additionally, 36% of fluorescent cells formed stargazin-mediated surface expression rings (Table 1).

Because stargazin may increase surface expression by blocking an intracellular ER retention signal in the pore, we tested R246YFP (Arg607). Wild-type GluR2 contains a residue that undergoes RNA editing, converting the glutamine for an arginine in the pore region (35). This channel would be expected to be retained largely in the ER (34), have difficulty forming tetramers (36), and presumably not get to the surface membrane even with stargazin unless stargazin could somehow block the ER retention signal. The number of cells co-transfected with stargazin and R246YFP that had rings was about the same regardless of whether residue 607 was an Arg or a Gln (Table 1). Together, these data suggest that stargazin occludes one or more ER retention signals at the cytoplasmic face of the receptor.


Figure 3
View larger version (66K):
[in this window]
[in a new window]
  		FIGURE 3.
The GluR1 cytoplasmic tail does not contain a specific stargazin-binding site. a, alignment of the first 14 amino acids of the cytoplasmic tails of the AMPA receptors GluR1-4, and the kainate receptor GluR6, top. Shaded areas show identity to GluR1. Below are a series of mutant R1i81YFP constructs with the mutations as indicated by the unshaded residues. All constructs formed stargazin-mediated surface expression rings, as indicated by +. b, confocal images of fixed, unpermeabilized HEK-293 cells expressing FLAG-R4i or FLAG-R4i{Delta}CTD without or with stargazin. Cells were labeled with a FLAG antibody and visualized with Alexa Fluor 568. All fluorescent images were acquired at the same gain; red pseudo-color represents intensity saturation. A smaller DIC image is shown for each field of cells.

 


Figure 4
View larger version (41K):
[in this window]
[in a new window]
  		FIGURE 4.
The N-terminal domain (ATD) is not essential for stargazin modulation of trafficking or desensitization. a, fluorophore-tagged N- and/or C-terminal deletions of GluR2 (left); + or - indicate whether co-expression with stargazin increased surface expression as determined by confocal microscopy (right). b, confocal images of R2i{Delta}ATD46YFP/stg (left) and R2o{Delta}ATD46YFP/stg (right) demonstrating the formation of surface rings and aggresomes (arrows). c, bar graphs show mean time constant of desensitization (left) and mean current amplitudes on a log scale (right) of R2o{Delta}ATD46YFP and R2i{Delta}ATD without (black) or with (gray) stargazin. Inset showing glutamate-evoked currents; red line is without stargazin and black line is with, vertical scale bar is 200 pA, horizontal scale bar is 60 ms. (***, p < .001). d, representative images of cells co-transfected with ATD-deleted receptors and stargazin. Panels 1, 3, and 5 are confocal images acquired at a 652 amplifier gain setting and 2, 4, and 6 are the same field acquired at a 237 amplifier gain setting to permit visualization of only the aggresomes. Images 1 and 2 are cells transfected with R2i{Delta}ATD46YFP (no aggresomes), images 3 and 4 are cells co-transfected with R2i{Delta}ATD46YFP/stg (large aggresomes), and images 5 and 6 are cells co-transfected R2o{Delta}ATD46YFP/stg (no aggresomes). Red pseudo-color represents intensity saturation.

 
Stargazin Mediates Trafficking of AMPA Receptor N-terminal Deletions with Isoform Differences Related to Differential Protein Stability—The N-terminal domain (ATD) of glutamate receptors has been implicated in assembly, trafficking, and allosteric modulation (37-39). To test whether or not the ATD plays a role in stargazin-mediated trafficking or modulation of the biophysical properties of AMPA receptors, the flip and flop isoforms of GluR2 (R2i and R2o) lacking the ATD, similar to R4i{Delta}ATD (27), were made.

R2o{Delta}ATD and R2i{Delta}ATD were tagged with YFP at the C terminus, after amino acid 46 of the cytoplasmic tail (R2o{Delta}ATD46YFP, R2i{Delta}ATD46YFP). Without stargazin, these proteins demonstrated a cytosolic expression pattern, and with stargazin, R2o{Delta}ATD46YFP formed pronounced surface expression rings (Fig. 4b). In contrast to R2o{Delta}ATD46YFP, R2i{Delta}ATD46YFP did not form surface expression rings. A large number of cells expressing R2o{Delta}ATD46YFP or R2i{Delta}ATD46YFP formed aggresomes (40) with or without stargazin (Table 2), but this was most pronounced for R2i{Delta}ATD46YFP with stargazin (Fig. 4, b and d). All R1 and R2 fluorescently tagged, C-terminal deletions with an intact ATD, however, were virtually free of aggresomes regardless of the length of the C-terminal tail (with the exception of GluR2 Arg607, but not R607Q). Differential aggresomal accumulation in the flip isoform may account for the inability of R2i{Delta}ATD46YFP to permit efficient stargazin-mediated surface expression. Aggresome formation may also explain the reduction in current amplitude seen for R2i{Delta}ATD (382 ± 129 versus 166 ± 67 pA with stargazin) as well as R4i{Delta}ATD (166 ± 43 versus 23 ± 4 pA with stargazin, p < .01). Consistent with this interpretation, stargazin increased the current amplitude of R2o{Delta}ATD46YFP (41 ± 17 versus 597 ± 146 pA with stargazin).


View this table:
[in this window]
[in a new window]
  		TABLE 2
ATD deletions promote aggresome formation in the absence or presence of stargazin

The constructs tested are indicated in the left column, and the percentage of cells forming aggresomes in the right column. The raw data (actual number of cells with aggresomes/total cells counted) for each condition is shown in parentheses. Stargazin was co-expressed in a 1:2 stg ratio (in µg) and the total concentration of AMPA receptor DNA was not changed.

 


Figure 5
View larger version (46K):
[in this window]
[in a new window]
  		FIGURE 5.
FRET analysis suggests stargazin self-assembly. a, photobleaching of stgCFP and stgYFP co-transfected into HEK-293 cells (1CFP:4YFP). Upper two panels demonstrate CFP and YFP emission prior to photobleaching; lower two panels show that after selective YFP photobleaching, stgCFP emission is enhanced. b, confocal images of green fluorescence from 1 µg of stgGFP transfected into HEK-293 cells (left) and green fluorescence arising from a 0.4 µm fluorescent bead (right) measured by the same settings as the cells (the image of the bead was then digitally magnified for clarity). CFP and YFP have a Forster radius of about 50 Å, so the maximum distance these fluorophores could be apart and still transfer energy is 100 Å. By this reasoning, 104 molecules/µm2 would be necessary to get FRET from overcrowding of the membrane. Imaging the beads at the lowest setting possible yielded a maximum detection limit before saturation of 820 stargazin molecules per µm2. Approximately 50% of the imaged cells were not saturated. If the brightest cells were even twice the detection limit at 1640/µm2, this would still fall short of the 10,000/µm2 needed. c, mean membrane FRET efficiencies (E) = 1 - (ICFPpre/ICFPpost) arising from stargazin self-assembly in HEK-293 cells with varying input ratios and total DNA concentrations (0.1-1.0 µg); in all cases, the FRET interaction between stargazin molecules is significantly greater than control (p < .03).

 
The ATD Is Not Necessary for Stargazin Modulation of AMPA Receptor Kinetics—Whereas previous studies using R4i{Delta}ATD showed only a modest change in desensitization from wild-type R4i (32), R2i{Delta}ATD desensitization kinetics ({tau}des = 13.7 ± 0.6 ms) were much slower than wild-type R2i ({tau}des = 6.9 ± 0.5 ms). {tau}des for R2o{Delta}ATD46YFP was 2.9 ± 0.3 versus 1.4 ± 0.1 ms for wild-type R2o. Stargazin also modulated desensitization of R4i{Delta}ATD ({tau}des = 5.6 ± 0.4 versus 8.4 ± 0.7 ms with stargazin) and R2i{Delta}ATD ({tau}des = 13.7 ± 0.6 versus 32.3 ± 4.1 ms with stargazin) (Fig. 4c). In contrast, the effects of stargazin on modulation of R2o{Delta}ATD46YFP were more modest ({tau}des = 2.9 ± 0.3 versus 4.1 ± 0.9 ms with stargazin). This flip/flop difference has been previously reported for full-length GluR2 (21) and suggests that the ATD does not play a significant role in stargazin-mediated modulation of desensitization.

We next looked at deactivation kinetics and found no detectable slowing of deactivation for R2i{Delta}ATD ({tau}deact = 1.9 ± 0.6 versus 1.4 ± 0.3 ms with stargazin), or for R2o{Delta}ATD46YFP (1.8 ± 0.3 versus 2.4 ± 0.8 ms with stargazin). The rate of deactivation, however, was slower without an ATD for both R2i{Delta}ATD ({tau}deact = 1.9 ± 0.6 ms compared with wt R2i = 0.9 ± 0.1 ms (41) and R2o{Delta}ATD46YFP ({tau}deact = 1.8 ± 0.3 versus 0.7 ± 0.1 ms for wild-type R2o).

FRET Analysis Suggests Stargazin Self-assembly—Our results from the N and C termini deletions strongly suggest that stargazin acts upon an AMPA receptor intracellular site to direct trafficking to the surface. Whereas biochemical methods such as pull-down assays would not distinguish between a direct protein-protein interaction or participation in a protein complex, FRET between fluorophores on two proteins is strong evidence for a tight (≤100 Å) intermolecular interaction. Therefore, we studied the interaction using FRET between the fluorophore-tagged GluR1 and stargazin proteins. Initially, we measured the ability of stargazin molecules to undergo FRET in the absence of receptor. Varying ratios and total concentrations of stgCFP and stgYFP were transfected, and FRET in the membrane was measured using a photobleaching protocol (Fig. 5, a and c). At all DNA concentrations and ratios tested, FRET between two stargazin molecules (7.8-16.2% efficiency) was significantly greater than a membrane control, Kv2.1CFP: R1i81YFP: stargazin at a 1:2:2 ratio (0.2 ± 0.1% FRET efficiency, p < .005). Because of the profound stargazin-stargazin membrane fluorescence compared with that of cells co-transfected with fluorescent AMPA receptors and stargazin, it was necessary to rule out that FRET occurred from overcrowding of the plasma membrane.

Yellow-green fluorescent beads were used to estimate how much stargazin protein was in the membrane (31). A confocal image of stgGFP-transfected cells is shown in Fig. 5b, with a fluorescent bead in a different dish taken at an identical setting. Analysis of the comparative intensity of the bead and the cell (see figure legend) suggests that the membrane density of stargazin was ~820 molecules per µm2, and therefore overcrowding by fluorophore-tagged stargazin could not explain the FRET data shown in Fig. 5c, unless certain regions of the membrane have greatly increased protein density. These data are consistent with an interpretation that stargazin:stargazin FRET arises from specific homo-oligomerization rather than nonspecific membrane crowding.

FRET Occurs between Stargazin and GluR1—To determine whether stargazin and AMPA receptors interact closely enough to permit energy transfer that can be measured by FRET, we used a competition assay. The first goal was to determine whether the stargazin complex known to FRET (stgCFP:stgYFP) could be disrupted by co-expression with a non-fluorescent AMPA receptor. The FRET efficiency of a 1:1 ratio of stgCFP and stgYFP (0.2 µg of total cDNA) was 10.6 ± 1.4%, which was competed by overexpressing R1i (2 µg of cDNA), reducing the FRET efficiency to 3.9 ± 1.1%, p = .003 (Fig. 6, a and c). However, neither 2.0 µg of cDNA encoding CD8 (11.8 ± 2.6%), Kv2.1 (8.0 ± 0.9%) nor R6 (8.5 ± 1.6%) significantly decreased membrane FRET (Fig. 6a). A dose dependence of R1i competition was determined, in that concentrations of 0.5, 1.0, and 2.0 µg of cDNA significantly reduced membrane FRET efficiency from the basal level of 16.2 ± 2.8% to 7.7 ± 1.4, 9.1 ± 1.6, and 5.8 ± 1.1%, respectively (Fig. 6, b and d). The reduction of FRET between stargazin molecules suggests that when there is an excess amount of R1i, but not other membrane proteins, the stargazin homo-oligomer population declines.

Based upon the mutational analysis of the C terminus of GluR1, one would predict that co-expression of R1i81CFP with stgYFP would permit FRET, whereas co-expression of R1i15CFP with stgYFP would not. Indeed, R1i81CFP:stgYFP produced robust FRET (20.2 ± 2.1%) in the plasma membrane, but also produced FRET (7.6 ± 1.0%) in the cytosolic, reticular network of cells with AMPA receptor rings (Fig. 6e, left). However, cells that did not contain R1i81CFP rings had a cytosolic FRET efficiency (3.4 ± 1.0%) that was significantly lower and not significantly different than the negative soluble CFP and YFP control (2.4 ± 0.5%). As predicted, co-expression of R1i15CFP with stgYFP showed no membrane fluorescence (Fig. 6e, right) and no significant FRET efficiency in the cytosol (4.1 ± 1.0%).

As a putative negative control for membrane FRET, the T cell receptor CD3CFP (29) was co-expressed with stgYFP (0.3 µg: 1 µg of stgYFP), but also showed robust membrane FRET efficiency (30.4 ± 2.1%) that was competed to 15.4 ± 2.0% by stargazin. Thus, we cannot rule out that the high expression levels of stargazin may contribute to some background FRET from overcrowding. The FRET between stargazin and both negative controls, CD3 and Kv2.1CFP (~20%), was much higher than would be expected of a negative control. Nevertheless, a control FRET experiment using Kv2.1CFP with excess R181YFP and non-fluorescent stargazin (to get R181YFP to the membrane) demonstrated ~0% membrane FRET efficiency.

AMPA Receptor Hetero-oligomerization Can Rescue C-terminal Deletions with Trafficking Defects—The previous experiments focused on AMPA receptor homo-oligomer trafficking by stargazin. Because our data support the idea that stargazin and AMPA receptors associate in the ER where subunit assembly is also occurring, we tested whether stargazin-mediated trafficking requires access to each of the subunits in the tetrad. R115- and R216-tagged channels (shorter C termini) that did not form surface expression rings with stargazin were co-transfected with both stargazin and an R1 or R2 channel that did form surface expression rings when co-transfected with stargazin (medium or "longer" C termini). R1i81YFP, R1i46CFP, and R1i36YFP could rescue R1i15CFP trafficking (Fig. 7a). R1i46CFP was able to rescue R2o16YFP as well as R2o{Delta}ATD16YFP, suggesting that stargazin does not need to bind all four subunits in a tetramer.

We further analyzed the data to assess the trafficking of subunits with different C termini within a tetramer. Two channels that could be trafficked independently by stargazin had virtually the same membrane to cytosolic fluorescence intensity ratio, (Imembrane/Icytosol)A/(Imembrane/Icytosol)B, where A = R1i46CFP and B = R2o46YFP (.99 ± .06) or A = R1i46YFP and B = R1i81CFP (0.94 ± .06). In contrast, short channels co-expressed with long channels had significantly reduced membrane to cytosolic ratios compared with the long forms, for example, where A = R2o16YFP and B = R1i46CFP (0.78 ± .06); where A = R1i15YFP and B = R1i46CFP (0.77 ± .04); and, where A = R1i15CFP and B = R1i36YFP (0.71 ± .08) (Fig. 7b).

Shorter forms are rescued to the point that they are only ~25% less effectively trafficked than the longer forms. This suggests that perhaps only 1 long form per tetramer is needed for stargazin-mediated trafficking. We therefore compared the membrane FRET, in the presence of stargazin, between tetramers composed of short and medium length subunits, R1i15 and R1i46, at different ratios of co-expression (Fig. 7c). As a control for differences in expression levels we compared the FRET between R1i15CFP and R1i46YFP (1:1; 0.5 CFP: 0.5 YFP: 2 stg) or (1:4; 0.2 CFP:0.8 YFP: 2 stg) and R1i46CFP and R1i15YFP (1:1) or (1:4) in the cytosol. We found equal cytosolic FRET between both combinations of subunits at 1:1 (~7%) and 1:4 (~11%). As expected the FRET efficiency was greater at the 1CFP:4YFP ratio because more tetramers would be composed of excess YFP subunits. We next measured the FRET efficiency at the membrane for R1i15CFP and R1i46YFP at 1:4 (18.5 ± 2.9%) to determine the maximum FRET efficiency and what value we could expect if stargazin was trafficking tetramers predominantly in a 1R1i15:3R1i46 ratio. The membrane FRET efficiency of the 1:1 ratio (16.5 ± 2.8%) was significantly greater than the 1:1 cytosolic ratio, but not different from the 1:4 cytosolic or membrane FRET. This suggests that the preferred heteromeric stoichiometry for stargazin mediated trafficking was 1R1i15:3R1i46. In support of this conclusion, membrane FRET efficiency between R1i46CFP and R1i15YFP at 1:1 was significantly less (6.5 ± 1.8%) and no greater than the cytosolic 1:1 ratio, suggesting that tetramers composed of fewer R1i15YFP and more R1i46CFP were preferentially trafficked.


Figure 6
View larger version (62K):
[in this window]
[in a new window]
  		FIGURE 6.
Stargazin assembly with GluR1i is a specific interaction. a, FRET efficiency of self-assembled stargazin molecules at the membrane (1 stgCFP:1 stgYFP) is competed when co-transfected with excess R1i in HEK-293 cells (1:1:20 or 0.1:0.1:2 in µg), but not by CD8, Kv2.1 or GluR6 (1:1:20 or 0.1:0.1:2). b, membrane FRET efficiency between stargazin molecules (0.07 µg of stgCFP:0.13 µg stgYFP) co-transfected with 0.5, 1.0, or 2.0 µg of R1i. c, confocal images of (0.07 µg stgCFP:0.13 µg stgYFP) before (red) and after stgYFP photobleach (blue) with corresponding profile intensities. d, same interaction as c but competed with 2 µg of R1i. (*, p < .05; **, p < .005; ***, p < .001). e, confocal images of a HEK-293 cell co-transfected with R1i81CFP and stgYFP, pre- and post-photobleaching of YFP (left). Confocal images of HEK-293 cells co-transfected with R1i15CFP and stgYFP pre- and post-photobleaching of YFP (right).

 
As an attempt to rule out that other combinations of heteromeric receptors besides 1R1i15:3R1i46 could not be trafficked by stargazin we looked at membrane FRET using a 1R1i46CFP:4R1i15YFP ratio, to force the majority of tetramers with a CFP subunit into a 3R1i15:1R1i46 ratio. The membrane FRET efficiency (13.3 ± 1.6%) was significantly greater than the 1:1 membrane FRET efficiency (p < .03). This suggests that stargazin can traffic hetero-oligomers containing two short subunits, and does not rule out the possibility that stargazin may traffic tetramers containing a single long subunit.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
In this study, we found that stargazin-mediated trafficking of GluR1 and GluR2 is hindered when CFP or YFP is inserted at the proximal cytoplasmic tail. We also determined that the first 380 amino acids of AMPA receptors (the ATD) are not necessary for stargazin trafficking but this domain has an important role in tetrameric stability of AMPA receptors. FRET analysis demonstrated that a homo-oligomeric population of stargazin exists in the plasma membrane, which could be specifically out-competed by GluR1 protein, but not by high concentrations of CD8, GluR6, or Kv2.1. AMPA receptors with traffickable C termini interact with stargazin in a close association that permits FRET in both the plasma membrane and the ER network of transfected cells.


Figure 7
View larger version (83K):
[in this window]
[in a new window]
  		FIGURE 7.
Hetero-oligomerization can rescue C-terminal deletions with trafficking defects. a, confocal images of cyan and yellow fluorescence from HEK-293 cells expressing R1i15 or R2o16, with either longer form of R1i or R2o (1 shorter:1 longer:2 stg). In all cases, R115 and R216 co-transfected with stargazin without a longer form had failed to form surface expression rings (Fig. 2 and data not shown). b, ratio of mean membrane fluorescence intensity/mean cytosol fluorescence intensity was compared between two different constructs (A, constructs in the left column;or B, constructs in the right column) that could either be independently trafficked by stargazin or required co-expression of a longer form to form rings. Two constructs that could independently form surface expression rings with stargazin had approximately equal amounts of membrane: cytosolic intensity, whereas short forms co-expressed with long forms were significantly less and were trafficked ~75% as well as the long forms. (*, p < .05 for both controls; {blacksquare}, p < .05 compared with R1i46C + R2o46Y). c, membrane FRET efficiency between R1i15 and R1i46 was compared with cytosolic FRET efficiency using a 1CFP:1YFP:2stg or 0.2CFP:0.8YFP:2stg ratio between tagged R1i15 and R1i46.

 
The AMPA Receptor-Stargazin Binding Site—We propose that stargazin interacts with AMPA receptors via a binding site that is comprised of the AMPA receptor domain 2 of the ligand binding core (including the flip/flop region), the transmembrane domains, and the cytoplasmic face including access to the pore. An extracellular interaction with the flip/flop domain is consistent with single particle electron microscopy experiments that show the primary interaction being near the transmembrane domains (23, 24). Swapping experiments using {gamma}-5 (an inactive, structurally similar protein to stargazin) suggest that only the second stargazin transmembrane domain is an important AMPA receptor contact, and was necessary for maintaining kainate responses (20). Consistent with our results, it was not a required domain for trafficking. Because our experiments suggest that the primary, specific, intracellular site of interaction with stargazin is not the C terminus of AMPA receptors, other intracellular sites are implicated. The part of the cytoplasmic tail of stargazin that was found to be essential for AMPA receptor trafficking (up to residue 269, (26)) contains 16 basic residues (and 4 negative residues). Interestingly, the 26 amino acid intracellular domain after M1 of GluR1 contains 8 acidic residues (and 3 positive residues). This cytoplasmic region between M1 and the pore loop may contribute to the stargazin-AMPA receptor interaction necessary for trafficking receptors to the surface. A stargazin interaction with this site may explain how GluR2 Arg607 homo-oligomers, typically retained in the ER (34), can form surface expression rings when co-transfected with stargazin.

Our results are consistent with previous studies suggesting that the first extracellular domain of stargazin plays a role in AMPA receptor trafficking (21, 26). The swap of {gamma}-5 in this region still results in a stargazin hybrid that enhances AMPA receptor surface expression, though not as robustly (26). If there are at least two distinct sites of interaction between stargazin and AMPA receptors (one extracellular, and one intracellular) the removal of an extracellular interaction, though not intrinsically necessary for trafficking, would reduce association between the two proteins.

Stargazin Enhances Trafficking by Blocking ER Retention—Our data suggest that in order for GluR1 and GluR2 to be trafficked by stargazin, it must have intracellular access to the cytoplasmic face of AMPA receptors (Fig. 8). In contrast, no such cytoplasmic interaction is necessary for subsequent stargazin-mediated modulation of desensitization. The intracellular interaction responsible for stargazin-mediated trafficking may block one or multiple ER retention signals; multiple ER retention signals are consistent with the graded response in current density seen with progressive C-terminal deletions of stargazin (21). The receptor mutant that exemplifies the different moieties of the stargazin-receptor interaction is R1i15CFP. Although the kinetics of this channel were modulated by stargazin as well as the kinetics of R1i81CFP, stargazin was unable to force the channel to form surface rings. This difference could be explained if R1i81CFP interacts with stargazin in the ER while R1i15CFP does not (or has a reduced affinity of interaction), or if association of stargazin in the ER with R1i15CFP homo-oligomers does occur, but fails to block any ER retention signals.


Figure 8
View larger version (37K):
[in this window]
[in a new window]
  		FIGURE 8.
Model of stargazin modulation of trafficking versus desensitization of R1i15YFP and R1i81YFP. Shown is a model that assumes that stargazin binding occurs at both dimer interfaces of the tetrameric channel; our data would also be consistent with a model in which stargazin binds individual subunits within the tetramer. In the ER stargazin (red) associates with higher affinity to the dimer interface of the R1i81YFP AMPA receptor tetramer (magenta, with an extended tail showing the yellow YFP attachment) than to the dimer interface of the R1i15YFP AMPA receptor (magenta, with a minimal tail showing the yellow YFP). The thickness of the magenta arrows correlates with the hypothesized relative affinities; the lower affinity arises because of some steric hindrance by the fluorophore. Known ER retention signals are shown in black; one is located in the proximal cytoplasmic tail and the other within the pore loop. Association between stargazin and R1i81YFP in the ER blocks at least one retention signal, hypothesized to be the pore loop. When more than one stargazin associates with AMPA receptors in the ER more retention signals are blocked, permitting greater exodus from the ER to the surface membrane (blue). This pathway represents stargazin modulation of trafficking, which is sensitive to the C terminus. The few R1i15YFP channels that reach the surface membrane without the aid of stargazin can associate with stargazin (red arrow) despite the lowered affinity because of the excess amount of stargazin in the surface membrane. This pathway represents stargazin modulation of desensitization and deactivation, which is relatively independent of the nature of the C terminus.

 
Support for the hypothesis that there is an initial interaction between AMPA receptors and stargazin in the ER is lent not only by measuring FRET between stargazin and AMPA receptors in the cytosol but also by experiments in which different glutamate receptor constructs were overexpressed in the presence of stargazin. If there were an initial protein-protein interaction in the ER, we should be able to sequester stargazin in the ER and prevent it from forming surface expression rings. An interpretation of results of this experiment (Table 3) is that stargazin can be sequestered in the ER by overexpression of R181CFP but not R115CFP, suggesting not only that association between stargazin and R115CFP occurs preferentially at the surface membrane, but also that multiple stargazin molecules per tetramer traffic AMPA receptors more efficiently.


View this table:
[in this window]
[in a new window]
  		TABLE 3
Stargazin surface expression is reduced by R1i81CFP

Quantification of the cells forming stgYFP expression rings when stgYFP was expressed alone or in the presence of channel (0.15 stg: 2 µg channel). Channel combinations are indicated in the left column, with the percentage of cells with pronounced stgYFP rings in the right column. The raw data (actual number of cells with rings/total cells counted) for each condition are shown in parentheses.

 
Stargazin is found primarily in the surface membrane and could forego its interaction with AMPA receptors until both proteins reached the plasma membrane. The lower affinity association may be similar to the stargazin association with calcium channels, which alters the biophysical properties of this voltage-gated channel without influencing its trafficking (42). Our hypothesis would predict that stargazin overexpression would result in more association with R115YFP in the ER. The failure of R115YFP to form surface expression rings even when co-expressed with a 1:8 stg ratio is consistent with there being a transmembrane/extracellular association between the two proteins in the ER that does not associate intracellularly to block any retention sites.

Stargazin Modulation Demonstrates Flip/Flop Isoform Differences in Channels without the ATD—Although this study did not focus on splice-isoform differences in stargazin trafficking and/or modulation of AMPA receptors, there is evidence that stargazin affects AMPA receptors in an isoform-selective manner (21). Flip and flop isoforms differ in their kinetic properties and allosteric modulation (43-45), so if stargazin is modulating deactivation and/or desensitization through the ligand-binding core, one might predict that there could be splice isoform differences. Isoform differences may also explain the discrepancy between our data and those of Arai and co-workers (46), who found that C-terminal deletions of GluR1o resulted in increased rates of deactivation and desensitization. The desensitization kinetics of the GluR1i C-terminal deletions we studied were not significantly different from wild type. Additionally, deletion of 52 amino acids of the C terminus (equivalent to R129) for GluR1o was shown to prevent stargazin-mediated effects on desensitization and deactivation (47), which is significantly different than what we found with GluR1i.

Our results indicate that the ATD is not necessary for stargazin modulation of desensitization for either the flip or flop isoform. However, stargazin association may alter the stability of both GluR2i{Delta}ATD and GluR4i{Delta}ATD homo-tetramers. GluR2o{Delta}ATD was not only trafficked effectively to the surface membrane by stargazin, aggresome frequency was decreased with stargazin (Table 2). In contrast, the aggresomes increased in size for GluR2i{Delta}ATD co-transfected with stargazin (Fig. 4d) without changing significantly in frequency. Because GluR2 modulation of desensitization by stargazin is strongly influenced by the flip-flop isoform, this suggests that there could be an interaction between stargazin and the flip/flop region.

Aggresome formation and density may be related to the amount of monomeric and dimeric subunits in the ER. This correlation is in agreement with the aggresome formation observed when we expressed wt GluR2 Arg607 homomers but not R607Q homomers (Table 2). Arg607 homomers are mostly in the monomeric or dimeric state in the ER while R607Q homomers have an enhanced proclivity to form tetramers (36). It is interesting that stargazin did not alter the number of aggresomes for Arg607. This result is in agreement with previous work that provides evidence that stargazin binds only to tetramers in the ER and would not affect the dimer-dimer interaction.

Insight into the Stoichiometry of the Stargazin-AMPA Receptor Interaction—The finding that GluR1 subunits with differing C termini could form hetero-oligomers but were different in their ability to be trafficked by stargazin enabled us to rule out that four stargazin binding sites (1 per subunit) were necessary for stargazin-mediated trafficking. The ability of multiple stargazin molecules to associate with a receptor was confirmed by different stoichiometry-dependent FRET efficiencies (supplemental Fig. S2). Our study of hetero-oligomers leaves open the possibility that stargazin binds to a dimer interface composed of two sufficient subunits (2 short:2 long) or binds to a single "sufficient" subunit (3 short:1 long). We hypothesize that only one stargazin per tetramer needs to bind to enhance trafficking of the AMPA receptor although more than one association leads to the blockade of more ER retention signals and thus more effective trafficking. The question still remains as to whether there are two stargazin binding sites per tetramer or one for each subunit.

Summary—Our studies bring a new level of resolution to investigate the nature of the interaction between AMPA receptors and stargazin. However, the contribution of startgazin to the synapse is complex, in part because of the many unpredicted activities of this protein (see for example Ref. 48), and the fact that its actions are activity-dependent (49). An intriguing property of stargazin is that it interacts with both calcium channels and AMPA receptors at the plasma membrane (50). It is unclear whether stargazin trafficking of AMPA receptors from the ER, modulation of AMPA receptor biophysical properties in the plasma membrane, complex formation with AMPA receptors and calcium channels, or ability to mediate cell-cell adhesion are all regulated by activity and contribute to synaptic plasticity. Additional experiments will be needed to address these important issues.


   FOOTNOTES
 
* This work was supported by National Institutes of Health Predoctoral Training Grant NS43115-02 (to M. A. B.) and Grant R01MH64700 (to K. M. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

Formula The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1 and S2. Back

1 To whom correspondence should be addressed: Dept. of Biomedical Sciences, Div. of Neuroscience, CO State University, Fort Collins, CO 80523-1617. Tel.: 970-491-2263; Fax: 970-491-7907; E-mail: kpartin{at}lamar.colostate.edu.

2 The abbreviations used are: AMPA, {alpha}-amino-3-hydroxy-5-methylisoxazole-4-propionate; cMEM complete minimal essential medium; FRET, fluorescence resonance energy transfer; ECFP, enhanced cyan fluorescent protein; EYFP, enhanced yellow fluorescent protein; EGFP, enhanced green fluorescent protein; GluR, glutamate receptor; HEK, human embryonic kidney; ATD, N-terminal domain; stg, stargazin; i, flip; o, flop; R1, GluR1; R2, GluR2; CMV, cytomegalovirus; wt, wild type; ER, endoplasmic reticulum. Back


   ACKNOWLEDGMENTS
 
We thank John Gieser for his help in the production of this manuscript. We thank Drs. Kurt Beam, Nancy Lorenzon, and Michael Tamkun for useful discussions and their critique of an earlier version of the manuscript.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION