HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 23, 20423-20430, June 7, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/23/20423    most recent M109412200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wong, W. Articles by Schlichter, L. C. Search for Related Content PubMed PubMed Citation Articles by Wong, W. Articles by Schlichter, L. C. Social Bookmarking                      What's this?

Cell Surface Targeting and Clustering Interactions between Heterologously Expressed PSD-95 and the Shal Voltage-gated Potassium Channel, Kv4.2^*

Wei Wong^§¶, Evan W. Newell^**, Denis G. M. Jugloff^§, Owen T. Jones^§, and Lyanne C. Schlichter^§§§

From the  Division of Cellular and Molecular Biology, Toronto Western Research Institute, University Health Network, Toronto, Ontario M5T 2S8, and the ^§ Department of Pharmacology and the  Department of Physiology, University of Toronto, Toronto, Ontario M5S 1A8, Canada

Received for publication, September 28, 2001, and in revised form, March 28, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Kv4.2 is a voltage-gated potassium channel that is critical in controlling the excitability of myocytes and neurons. Processes that influence trafficking and surface distribution patterns of Kv4.2 will affect its ability to contribute to cellular functions. The scaffolding/clustering protein PSD-95 regulates trafficking and distribution of several receptors and Shaker family Kv channels. We therefore investigated whether the C-terminal valine-serine-alanine-leucine (VSAL) of Kv4.2 is a novel binding motif for PSD-95. By using co-immunoprecipitation assays, we determined that full-length Kv4.2 and PSD-95 interact when co-expressed in mammalian cell lines. Mutation analysis in this heterologous expression system showed that the VSAL motif of Kv4.2 is necessary for PSD-95 binding. PSD-95 increased the surface expression of Kv4.2 protein and caused it to cluster, as shown by deconvolution microscopy and biotinylation assays. Deleting the C-terminal VSAL motif of Kv4.2 eliminated these effects, as did substituting a palmitoylation-deficient PSD-95 mutant. In addition to these effects of PSD-95 on Kv4.2 distribution, the channel itself promoted redistribution of PSD-95 to the cell surface in the heterologous expression system. This work represents the first evidence that a member of the Shal subfamily of Kv channels can bind to PSD-95, with functional consequences.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The influence that voltage-gated ion channels exert on the integrative physiology of excitable cells is determined by the inherent biophysical properties of the channels and their cell surface distribution. Kv4.2 is a fast transient (A-type) voltage-gated potassium (Kv)¹ channel of the Shal subfamily that is found in heart and neurons (1). In myocytes, Kv4.2 is a major component of the I_to current, which is crucial for recovery from the QT interval and for repolarization (2). In neurons of the central nervous system, Kv4.2 is expressed primarily somatodendritically (3, 4), where its concentration at postsynaptic sites may affect the back propagation of action potentials (5) and may therefore modulate long term potentiation at synapses (6). In contrast, Kv1.4, a Kv channel of the Shaker subfamily, is primarily expressed on axons (3, 4). Its positioning along the axon and at presynaptic sites (7-9) is thought to regulate the efficiency of action potential propagation (10) and to control neurotransmitter release (11). It is therefore crucial to understand the basis for the specialized distributions of Kv channels in order to fully understand their roles in neuronal excitability. The mechanisms that specify ion channel distributions are not well known and likely differ between those that determine localization in sub-domains of cells (e.g. axon, soma, and dendrite) versus lateral organization at the membrane, such as within channel clusters.

Post-synaptic density 95 (PSD-95) is a membrane-associated guanylate kinase that helps localize and cluster numerous proteins. Kv1.4 and other Kv1 members are among the best known examples of in vitro ion channel clustering by PSD-95, but this is the only subfamily of Kv channels known to interact with PSD-95 (12, 13). PSD-95-mediated clustering has profound effects on the trafficking of Kv1 channels: it promotes surface expression (14) and suppresses endocytosis (15). Because clusters of PSD-95 with Kv1 channels are apparently immobile (16), this interaction may contribute to the polarized surface distribution exhibited by many Kv1 channels. Although several PSD-95-binding motifs begin with a glutamate residue, such as Kv1.4 (ETDV) and Kv1.5 (ETDL), several Kv1 channels (e.g. Kv1.1, Kv1.2, and Kv1.3) associate with PSD-95 through an alternate binding motif in which the glutamate residue is replaced by a hydrophobic amino acid (17).

Kv4.2 has a C-terminal Val-Ser-Ala-Leu (VSAL) that is very similar to the hydrophobic amino acid (T/S)XV motif; thus, we asked whether Kv4.2 associates with PSD-95 and whether VSAL represents a binding motif. We show, through co-immunoprecipitation and mutational analysis, that Kv4.2 protein does bind to PSD-95 in a heterologous expression system and that this interaction requires the VSAL motif. The interaction had physiological consequences for both Kv4.2 and PSD-95 protein distributions. PSD-95 increased the surface expression of Kv4.2 protein and caused it to cluster, as shown by deconvolution microscopy and biotinylation assays, and these effects required both the VSAL motif in Kv4.2 and the palmitoylation motif in PSD-95. Conversely, interaction with Kv4.2 increased the surface expression of PSD-95. Thus, by affecting the amount and location of channel protein, the PSD-95 interaction with Kv4.2 could affect the integrated electrophysiological properties of neurons.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Myc-tagged PSD-95 in the expression vector, pGW1-CMV, was a generous gift from Dr. Morgan Sheng (Harvard University, Boston, MA). C3,5S-PSD-95, the palmitoylation-deficient PSD-95 mutant in which the cysteine residues at positions 3 and 5 are replaced by serine residues, was described previously (15). Oligonucleotide primers were synthesized by ACGT Inc. (Toronto, Ontario, Canada). All chemicals were obtained from Sigma unless otherwise indicated.

Preparing the Kv4.2 Mammalian Expression Constructs-- Kv4.2 constructs in mammalian expression vectors were prepared using PCR with first-strand cDNA synthesized from rat hippocampus as template (Fast Track, Stratagene, La Jolla, CA). We first PCR-amplified a FLAG-tagged, wild-type cDNA (which we call Kv4.2) that included residues 1-630 (18), using the primers 5' ATGGACTACAAGGACGACGATGACAAGGCAGCCGGTGTTGCAGCATGG 3' (forward; also encodes the FLAG epitope after the start codon) and 5' TTACAAAGCAGACACCCTGAC 3' (reverse). This was ligated into the multiple cloning site of pCRII (Invitrogen, Carlsbad, CA) and then excised with NsiI/XhoI for ligation into the PstI/XhoI sites of pSVK3 (Amersham Biosciences) and pBK-CMV (Stratagene). Expression of Kv4.2 in pBK-CMV was examined by immunoblotting. Then, to facilitate studies of channel localization using fluorescence microscopy, an N-terminal fusion of the wild-type Kv4.2 construct to enhanced cyan fluorescent protein, ECFP (C-Kv4.2), was prepared by excising the coding sequence from the pSKV3 clone with EagI and ligating it into the Bsp120 I site of pECFP-C1 (CLONTECH, Palo Alto, CA). By using C-Kv4.2 cDNA as the template, a Kv4.2 deletion mutant lacking the four C-terminal amino acids (VSAL) was prepared by Pfu (Stratagene) PCR using the primers 5' TTTGAATTCTATGGCAGCCGGTGTTGCAGCATGG 3' (forward, encodes an EcoRI site before the start codon) and 5' TTTGGGCCCTTTTTACCTGACGATGTTTCCTCC 3' (reverse, encodes an ApaI site after the stop codon). This cDNA was ligated into the EcoRI and ApaI sites of pcDNA3.1 or pECFP-C1 to produce the Kv4.2 deletion mutant (Kv4.2VSAL) and the ECFP-tagged Kv4.2 deletion mutant (C-Kv4.2VSAL). For patch clamp analysis, we used a full-length Kv4.2 clone in pBiG (CLONTECH) containing a Kozak consensus sequence before the start codon (generously provided by Dr. P. Backx, University of Toronto). From this clone, a Kv4.2 mutant lacking the VSAL motif was generated with Pfu PCR, using the forward primer 5' TTTGAATTCTCCGCCATGGCAGCCGGTG 3' (encodes an EcoRI site and Kozak consensus sequence before the start codon), and the same reverse primer was used to create the deletion mutant (above).

The identities of all constructs used were verified through dideoxy sequencing (York University Sequencing Facility, and the University Health Network Research DNA Sequencing Facility, Toronto, Ontario, Canada).

Cell Culture and Transfection-- We used three cell lines for heterologous expression. HEK293T cells were formerly called tsA201 and are HEK293 cells that constitutively express the SV40 large T antigen to allow plasmid replication using the SV40 origin. Although they are apparently the same cells, we will follow the nomenclature used by the laboratories from which we obtained them. HEK293T cells (from Dr. D. Papazian, UCLA, CA) were maintained and transfected as described previously (19), but using only 0.5 µg of total DNA. This low amount of DNA was transfected because HEK293T cells were used for indirect immunofluorescence, and it was easier to discern surface versus subcellular staining when less protein was expressed. tsA201 cells (from Dr. P. Backx) were mainly used for biochemical studies, where it was advantageous to increase the amount of protein expressed. They were grown at 37 °C, 95% O₂, 5% CO₂ in Dulbecco's modified essential medium (University Health Network Sera and Media Services, Toronto) supplemented with 10% fetal bovine serum, 10 µg/ml penicillin, and 10 units/ml streptomycin (all supplements obtained from Invitrogen). The tsA201 cells were seeded at 4.5 × 10⁵ cells/cm² and then transfected the next day with 2.5-10 µg of total DNA, using the calcium phosphate precipitation method (20). For fluorescence imaging, the cells were replated the following day on collagen- and poly-L-lysine-coated coverslips and used 36-48 h after transfection.

For heterologous expression for electrophysiology experiments, we used CHO cells, which were maintained as described previously (21). CHO cells were seeded at 2.07 × 10⁵ cells/cm² and then transfected on the following day with 2 µg of total DNA plus 0.2 µg of pEGFP-C1 (CLONTECH) using FuGENE 6 (Roche Molecular Biochemicals). The cells were replated the day after transfection on poly-L-lysine-coated coverslips, and EGFP-positive cells were patch clamped 36-48 h after transfection, as described below.

Electrophysiology-- Whole-cell patch clamp recordings (15, 22) were used to assess the functional expression of Kv4.2 (with Kozak sequence), Kv4.2VSAL (with Kozak sequence), and C-Kv4.2 (ECFP-tagged) in transfected CHO cells. The pipette solution (300-310 mosM) contained (in mM) 140 KCl, 5 EGTA (potassium salt), 4 MgATP, 1 MgCl₂, and 10 HEPES, adjusted to pH 7.2 with KOH. The bath solution (310-320 mosM) contained 140 NaCl, 10 glucose, 5.4 KCl, 2 CaCl₂, 1 MgCl₂, and 10 HEPES, adjusted with NaOH to pH 7.4. All recordings were made at room temperature.

Co-immunoprecipitation-- tsA201 cells were co-transfected with Myc-tagged PSD-95 and one of the Kv4.2 channel constructs, so that we could immunoprecipitate PSD-95-containing complexes from these cells with an antibody directed against Myc. The Kv4.2 antibody was not effective for immunoprecipitation; thus, the reverse experiment could not be done. The cells were lysed with 500 µl of Nonidet P-40 co-IP buffer (1% Nonidet P-40, 100 mM NaCl, 20 mM Tris, pH 8.5, and 1 mM EDTA) supplemented with protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 1 µg/ml aprotinin, 1 µg/ml leupeptin, and 1 µM pepstatin A). The lysates were incubated with co-IP buffer for 20 min on ice and then clarified by centrifugation, and a small aliquot was reserved for direct immunoblotting. The remaining sample was pre-cleared by a 3-h incubation with protein A/G-agarose (Oncogene Science, Cambridge, MA), which was also used to capture antibody-protein complexes. The immunoprecipitates were boiled in 2× SDS gel-loading buffer prior to SDS-PAGE.

SDS-PAGE and Immunoblot Analysis-- A polyclonal antibody was prepared against a peptide (Vetrogen, London, Ontario, Canada) corresponding with residues 23-43 of Kv4.2 (GenBank^TM accession number S64320) with an additional C-terminal cysteine for coupling to keyhole limpet hemocyanin (Pierce). After confirming the identity of the peptide by amino acid analysis and mass spectroscopy, and cross-linking it to keyhole limpet hemocyanin with m-maleimidobenzoyl-N-hydroxysulfosuccinimide ester (Pierce), the conjugated peptide was purified by dialysis against phosphate-buffered saline (PBS) and injected into New Zealand White rabbits (Division of Comparative Medicine, University of Toronto, Toronto, Ontario, Canada). Next, antisera were collected from the immunized rabbits and the IgG fraction enriched by protein A affinity chromatography (MAPS kit, Bio-Rad). This antibody has been extensively used to detect Kv4.2 in immunoblots of heart tissue (23, 24), brain (23, 25), and transfected cells (25), as well as in immunohistochemistry of heart tissue (24).

Protein samples were prepared, electrophoresed, and transferred to nitrocellulose as described previously (26). The nitrocellulose membranes were incubated overnight at 4 °C with the Kv4.2 polyclonal antibody at 1:1000. Immunoreactivity was detected by incubating the membrane with goat anti-rabbit conjugated to horseradish peroxidase (Cedarlane Laboratories, Hornby, Ontario, Canada; 1:4000), followed by enhanced chemiluminescence (ECL; Amersham Biosciences). Quantification was done with a GS-670 imaging densitometer and Molecular Analyst 1.1 (Bio-Rad).

Biotinylation of Cell-surface Proteins-- Kv4.2 contains two putative extracellular lysine residues (18); thus, surface and intracellular channel protein can be differentiated using cell-impermeant biotinylating reagents that target lysine residues. Before transfection, tsA201 cells were plated on poly-L-lysine-treated (Sigma) 6-well plates. Biotinylation of cell-surface proteins was performed essentially as described previously (27), except that cells were biotinylated with 1.5 mg of sulfo-NHS-LC-biotin per 35-mm well (Pierce). Biotinylated cell membranes were incubated in 50 µl of SDS lysis buffer (1% SDS, 20 mM Tris, pH 8.0, 50 mM NaCl, 5 mM EDTA) for 10 min at 65 °C. The cell extracts were then diluted with 4 volumes of Triton lysis buffer (1% Triton, 20 mM Tris, pH 8.0, 50 mM NaCl, 5 mM EDTA) and incubated for a further 60 min on ice before pelleting insoluble materials. Biotinylated proteins were isolated and immunoblotted as described previously (28), except that Neutravidin Ultralink (Pierce) was used to affinity purify the biotinylated proteins.

Fluorescence Imaging-- To assess Kv4.2 protein localization and clustering, live tsA201 cells were washed with Dulbecco's PBS (139 mM NaCl, 2.7 mM KCl, 8.8 mM Na₂HPO₄, 1.48 mM KH₂PO₄) and mounted on glass slides. Cells were imaged using a Zeiss Axioplan 2 Imaging microscope (Carl Zeiss, Inc., Thornwood, NY) equipped with a digital camera (Zeiss AxioCam) and a 100× oil-immersion objective. Image stacks spanning the full z axis of the cell (with images spaced 200 nm apart) or spanning 0.52 µm (with images spaced 25 nm apart) were collected for each transfection condition. All z image stacks were deconvolved in AxioVision 4.0 (Carl Zeiss, Inc.) using the constrained iterative algorithm.

Localization of Kv4.2 and PSD-95 proteins was assessed by indirect immunofluorescent detection of ECFP-tagged channels and Myc-tagged PSD-95 (29). Cells were washed with PBS, fixed with 4% paraformaldehyde, and permeabilized with 0.1% Triton. Nonspecific binding was blocked with 3% heat-inactivated goat serum and 0.1% bovine serum albumin in PBS. Cells were then incubated with 1:2000 anti-GFP (Molecular Probes) (to detect ECFP-tagged channels) and/or 1:200 anti-Myc (to detect the tagged PSD-95) for 1 h at room temperature. This was followed by a 1-h incubation at room temperature with goat anti-rabbit conjugated to Alexa Fluor 488 (1:1000) and goat anti-mouse conjugated to Alexa Fluor 594 (1:200) in 0.1% bovine serum albumin in PBS. The coverslips were mounted in Slowfade (Molecular Probes) and imaged as above. Images were overlaid in AxioVision 4.0.

Statistical Tests-- Student's t test was used to analyze electrophysiological and immunoblotting data. Percentages (from the biotinylation cell surface assay) were subjected to arcsin transformation to produce data with a normal distribution (30) prior to statistical analysis using the Student's t test. Values of p < 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Heterologously Expressed Full-length, Deletion Mutant, and ECFP-tagged Kv4.2 Constructs Form Functional Channels-- Because tsA201 cells express an endogenous potassium current (31) that is not Kv4.2 (see Refs. 32 and 33; see also Fig. 2A), we used CHO cells for electrophysiological analysis. Consistent with previous reports (31, 32), no voltage-gated currents were detected in whole-cell recordings from cells transfected with pEGFP-C1 alone (mock-transfected; n = 6) (Fig. 1A). However, within 48 h after Kv4.2 cDNA was transiently transfected, K⁺ currents appeared (Fig. 1B), which were large, depolarization-activated, rapidly inactivating and similar to previously described (32, 33) Kv4.2 currents expressed in heterologous systems. Current densities ranged from 12 to 105 pA/pF (n = 5). As demonstrated previously (34, 35), inactivation was best described as the sum of two inactivation time constants: a fast component accounting for 83 ± 1% of inactivation (_fast = 17.7 ± 1.5 ms at +60 mV) and a slow component (_slow = 435 ± 114 ms at +60 mV) (all error values are S.E. and all tests are Student's t test, unless otherwise indicated). Cells co-transfected with Kv4.2 and PSD-95 produced currents with similar characteristics to cells expressing Kv4.2 alone (data not shown), with a comparable range of current densities (21-81 pA/pF, n = 7; p > 0.5). PSD-95 did not alter the inactivation kinetics (_fast = 19.2 ± 3.1 ms at +60 mV, p > 0.5; _slow = 296 ± 104 ms at +60 mV, p > 0.2) or the contribution of the fast component to inactivation (81 ± 2%, p > 0.2).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Heterologously expressed Kv4.2 forms functional channels. K+ currents in CHO cells were elicited by voltage clamp steps delivered at 10-mV increments from a holding potential of 70 mV (A, inset). Each voltage step was 300 ms long, with an inter-pulse interval of 20 s to allow recovery from inactivation. Representative traces are from mock-transfected cells (A), or cells transfected with Kv4.2 (with Kozak sequence) (B), Kv4.2VSAL (with Kozak sequence) (C), or C-Kv4.2 (D). Kv4.2 and C-Kv4.2 are full-length and contain the VSAL motif but differ in the presence of an N-terminal ECFP tag on C-Kv4.2. This tag was added to assist in localizing Kv4.2 channels in living cells (see Fig. 5). Kv4.2VSAL lacks the last four amino acids at the C terminus. The inset to B shows a Kv4.2-expressing cell before (pre-treatment) and 10 min after perfusing the bath with 20 µM quinidine.

Cells transfected with Kv4.2VSAL expressed a current that resembled the wild-type current (Fig. 1C) with a current density range of 4-56 pA/pF (n = 4), indicating that Kv4.2 channels lacking the C-terminal VSAL motif form functional channels on the cell surface. Deletion of the VSAL motif did not alter the time constants of inactivation (_fast = 24.0 ± 2.6 ms at +60 mV, p > 0.2; _slow = 441 ± 145 ms at +60 mV, p > 0.5) or the contribution of the fast component of inactivation (81 ± 9%; p > 0.5). Transfection of C-Kv4.2 (ECFP-tagged) into CHO cells produced K⁺ currents (Fig. 1D) ranging from 6 to 90 pA/pF (n = 4); thus, adding the ECFP tag to the N terminus of Kv4.2 does not prevent formation of functional channels. One clear difference was that ECFP-tagged Kv4.2 inactivated more slowly than Kv4.2 current: only one inactivation time constant could be fit ( = 224 ± 35 ms at +60 mV). This effect of ECFP is not surprising because the fast component of inactivation of Shal family K⁺ channels involves concerted actions of the N- and C-terminal domains (34), interactions that could be impeded by a bulky N-terminal tag. Similarly two proteins (KChIP and frequenin) were recently shown to bind to the N terminus of Kv4.2 and enhance the slow component of inactivation (35, 36).

Kv4.2 channels are blocked by quinidine, which reduces the peak current and accelerates the apparent inactivation rate (37). An application of 20 µM quinidine reduced the peak Kv4.2 current from 1200 to 295 pA at +60 mV (75% reduction) and reduced the fast time constant of inactivation to 4.6 ms (Fig. 1B, inset).

Kv4.2 Binds to PSD-95; the VSAL Motif Is Necessary-- The polyclonal anti-Kv4.2 antibody we made has been used in other studies (23-25); nevertheless, we confirmed its performance in our system. The antibody recognized a single ~70-kDa band in tsA201 cells that were transfected with the full-length channel (Fig. 2A, lane 5). The size is consistent with both the predicted (71 kDa) and previously observed molecular weight of Kv4.2 (18, 36). Further evidence for antibody specificity is the lack of immunoreactive bands in lysates from mock-transfected cells or from cells expressing vector alone (lanes 1-4), which also means that the endogenous K⁺ channel (mentioned above) is not Kv4.2. Pre-incubating the antibody with the antigenic peptide eliminated all immunoreactive bands (lanes 9-12). We created several cDNA constructs to assess the role of the C-terminal VSAL motif in its interaction with PSD-95. The deletion constructs, Kv4.2VSAL and C-Kv4.2VSAL, lack the VSAL motif at the C terminus. Lysates from tsA201 cells that had been transfected with the different constructs were immunoblotted with the Kv4.2 antibody to verify that the different proteins were well expressed. The antibody recognized one major band for each construct: at about 96, 68, and 94 kDa for C-Kv4.2, Kv4.2VSAL, and C-Kv4.2VSAL, respectively (Fig. 3A, lanes 5-8). Again, the product sizes are close to those expected and observed for an ECFP-tagged Kv4.2 fusion protein (38) and the slightly reduced size of the VSAL deletion mutant. As a further control, we determined the linear detection range of the antibody before quantifying differences in channel expression at the cell surface (below). Western analysis showed that the band density was a linear function of protein loaded between 2.5 and 20 µg for both wild-type and ECFP-tagged constructs (Fig. 2B).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   The Kv4.2 antibody recognizes all the Kv4.2 constructs used. A, in lysates from transfected tsA201 cells, the Kv4.2 antibody recognized bands of about 70 kDa for Kv4.2 (lane 5), 68 kDa for Kv4.2VSAL (lane 6), 96 kDa for C-Kv4.2 (lane 7), and 94 kDa for C-Kv4.2VSAL (lane 8). Pre-incubating the antibody with the Kv4.2 antigenic peptide eliminated these bands (lanes 9-12). Immunoreactive bands were also absent in mock-transfected cells (lane 1) or cells transfected only with the vectors, pBK-CMV (lane 2), pcDNA3.1 (lane 3), or pBK-CMV (lane 4). B, determining the linear range of labeling and detection by the Kv4.2 antibody. The insets show Western blots (WB) for a range of protein amounts loaded onto the gel (2.5, 5, 10, and 20 µg) for both the wild-type (Kv4.2) and ECFP-tagged constructs (C-Kv4.2). The graph shows that densitometry readings were a linear function of protein loaded between 2.5 and 20 µg (95% confidence intervals are indicated by broken lines).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Kv4.2 co-immunoprecipitates with PSD-95; deleting VSAL eliminates the interaction. Proteins were immunoprecipitated (IP) with Myc antibody from lysates of cells co-expressing Myc-tagged PSD-95 and either C-Kv4.2 or C-Kv4.2VSAL and were probed with the Kv4.2 antibody through Western blotting (WB). Full-length C-Kv4.2 was immunoprecipitated in the presence of Myc antibody (2nd lane) but not in its absence (3rd lane). In contrast, little or no deletion mutant (C-Kv4.2VSAL) was immunoprecipitated by the Myc antibody (5th lane). L, lysate.

To assess whether Kv4.2 co-immunoprecipitates with PSD-95, full-length constructs of C-Kv4.2, or the VSAL-lacking C-Kv4.2VSAL were co-expressed with Myc-tagged PSD-95 in tsA201 cells. Lysates prepared from these cells were incubated with an antibody directed against Myc, and the precipitates were probed with Kv4.2 antibody. Anti-Myc immunoprecipitated C-Kv4.2 (Fig. 3, lane 2), although omitting the Myc antibody eliminated the Kv4.2 band (lane 3), thus confirming that the channel did not bind nonspecifically to the agarose beads. The Kv4.2 mutant lacking the C-terminal VSAL motif (C-Kv4.2VSAL) did not immunoprecipitate with Myc antibody (lane 5). The differential ability of PSD-95 to immunoprecipitate C-Kv4.2 and C-Kv4.2VSAL was not because of differences in protein expression, since direct immunoblotting of lysates from transfected cells revealed robust expression (lanes 1 and 4). Collectively, these results indicate that the VSAL motif is necessary for Kv4.2 binding to PSD-95.

PSD-95 Increases the Cell Surface Expression of Kv4.2 Protein-- tsA201 cells were transfected with Kv4.2 or C-Kv4.2, with or without PSD-95 co-transfection. A cell-impermeant biotinylating reagent that modifies extracellular lysine residues was used to label cell-surface proteins, which were then affinity-purified on Neutravidin beads. Western blots of cell lysates were probed with the Kv4.2 antibody to monitor the total amount of Kv4.2 construct expressed. Precipitates were probed with the Kv4.2 antibody on the same blot as the lysates to obtain the biotinylated fraction of Kv4.2 (Fig. 4A). The proportion of biotinylated Kv channel was calculated by expressing each band density in the biotinylated fraction as a percentage of the band density in the corresponding lysate (i.e. total protein). When the amounts of biotinylated Kv4.2 and C-Kv4.2 in the presence or absence of PSD-95 were compared (Fig. 4B), PSD-95 significantly increased the biotinylated fraction by 2.4-fold for Kv4.2 (n = 3) and 1.4-fold for C-Kv4.2 (p < 0.05). Because only surface channels become biotinylated, PSD-95 increased the surface expression of Kv4.2. Patch clamp recordings did not reveal a significant increase in Kv4.2 current when co-transfected with PSD-95; however, this is not surprising because variability in transiently transfected Kv4.2 currents is large.² Biochemical assays, such as cell surface biotinylation, performed on populations of cells provide a better indication of average changes in localization.

View larger version (41K): [in this window] [in a new window]   Fig. 4.   PSD-95 increases the cell surface expression of Kv4.2 protein. A biotinylation assay was used to measure the surface fraction of channel protein (see "Experimental Procedures"), and representative Western blots are shown (A and C). For each sample, labeled proteins were precipitated with Neutravidin beads from 20 µg of total protein in cell lysates and then immunoblotted with the Kv4.2 antibody to measure the biotinylated protein. Total cellular expression of Kv4.2 protein was assessed by directly immunoblotting 2.5 µg of each lysate, which was one-eighth of the amount of protein added to the beads. These protein amounts were chosen because they were within the linear range of the Kv4.2 antibody (see Fig. 2B). A, upper panel, representative Western blots from biotinylation assays using tsA201 cells transfected with full-length Kv4.2 alone (shaded bars) or together with PSD-95 (open bars). Lower panel, representative Western blots from biotinylation assays using cells transfected with ECFP-tagged Kv4.2 (C-Kv4.2) alone (shaded bars) or with PSD-95 (open bars). B, summary of experiments like those in A, using the same open and shaded bar markings. The biotinylated proportion was calculated by expressing the intensities of the Kv4.2 bands in the precipitates as a fraction of those in the corresponding cell lysates. Each value represents the mean ± S.E. of the number of experiments from different transfected cell batches indicated on each bar. Student's t test (after arcsin transformation) indicates values that differ at the p < 0.05 level. C, representative Western blots from biotinylation assays using cells co-transfected with the ECFP-tagged Kv4.2 (C-Kv4.2) and the palmitoylation-deficient PSD-95 mutant (C3,5S-PSD-95) (hatched bars), or with the ECFP-tagged Kv4.2 deletion mutant (C-Kv4.2VSAL) and PSD-95 (shaded bars). D, summary of experiments in part C. Each value represents the mean ± S.E. of three experiments from different transfected cell batches. None of the values differ from each other at the p < 0.05 level (Student's t test after arcsin transformation).

We next investigated regions of Kv4.2 and PSD-95 that are likely to mediate the observed increase in Kv4.2 surface expression. Cell surface biotinylation was examined for cells co-expressing the deletion mutant (C-Kv4.2VSAL) plus wild-type PSD-95 or full-length Kv4.2 (C-Kv4.2) plus a palmitoylation-deficient PSD-95 mutant, C3,5S-PSD-95 (Fig. 4C). This PSD-95 mutant can bind to, but not cluster, other Kv channels (15, 29). Unlike wild-type PSD-95, the palmitoylation-deficient mutant did not significantly increase the proportion of full-length Kv4.2 on the cell surface (Fig. 4D) (p > 0.05, Student's t test). Wild-type PSD-95 failed to increase the surface fraction of the Kv4.2 mutant that lacked the putative PSD-95-binding motif (C-Kv4.2VSAL), when compared with full-length Kv4.2 alone (Fig. 4D) (p > 0.2). The surface expression of Kv4.2 full-length channel in the presence of PSD-95 was significantly higher than the amount of Kv4.2VSAL (C-Kv4.2 plus PSD-95, 13.5 ± 1.0%; C-Kv4.2VSAL plus PSD-95, 8.3 ± 0.7%; p < 0.05). Thus, PSD-95 binding and clustering appear to be required for the increased surface expression of Kv4.2. Total channel protein levels did not differ between the constructs or transfection combinations, as indicated by the intensities of Kv4.2 immunoreactive bands in cell lysates (bands marked "total"; p > 0.05).

The specificity of the biotin-Neutravidin interaction was confirmed by incubating lysates of cells transfected with Kv4.2 constructs but not exposed to sulfo-NHS-LC-biotin; immunoreactive bands were absent after affinity precipitation with Neutravidin beads (data not shown). The high proportion of non-biotinylated channel protein we observed is consistent with the large intracellular fraction seen from our imaging data (below) and in earlier studies (22, 36). This distribution might result from examining transient channel overexpression at a relatively short time after transfection (48 h in our study).

Protein Domains Involved in Kv4.2 Surface Expression and Clustering by PSD-95-- Live, transfected tsA201 cells were imaged by exploiting the fluorescent ECFP-tagged Kv4.2 protein in order to monitor channel protein distribution without potential fixation artifacts. In cells expressing C-Kv4.2 alone, there was considerable channel protein in an internal reticular network with some fluorescence present at the outer cell margins (Fig. 5A; representative of 34 cells). Some cells had highly fluorescent patches in the perinuclear region, as in Fig. 5D. Kv4.2 was reported previously (22, 36, 39) to be mainly on internal membranes, with a circle of strong fluorescence around the nucleus, in contrast to the discrete "hot spots" that we observed. PSD-95 co-expression (Fig. 5B; representative of 40 cells) resulted in clustering of the channel protein and good expression on the cell surface in about 50% of the cells imaged. This clustering did not occur when C-Kv4.2 was co-expressed with the palmitoylation-deficient PSD-95 mutant (Fig. 5C; representative of 24 cells); instead, considerable fluorescence was present on intracellular compartments, similar to cells expressing C-Kv4.2 alone. Therefore, eliminating the palmitoylation sites on PSD-95 prevented it from clustering and recruiting Kv4.2 to the cell surface. PSD-95 did not cluster the VSAL-lacking Kv4.2 protein (Fig. 5D; representative of 26 cells), and the channel proteins were largely expressed on internal membranes. Thus, binding of Kv4.2 to PSD-95 through the VSAL motif appears to facilitate both clustering and recruitment of the channel to the cell surface.

View larger version (80K): [in this window] [in a new window]   Fig. 5.   Protein domains involved in Kv4.2 surface expression and clustering by PSD-95. Fluorescence microscopy of live, transfected tsA201 cells using the ECFP tag on the Kv4.2 protein and deconvolution deblurring. Imaged z stacks were acquired from cells transfected with C-Kv4.2 (A), C-Kv4.2 plus PSD-95 (B), C-Kv4.2 plus C3,5S-PSD-95 (C), C-Kv4.2VSAL plus PSD-95 (D), ECFP (E), or salmon sperm DNA (mock-transfected) (F). The z stacks were deconvolved and a single image was selected for presentation (scale bars, 5 µm). ECFP-tagged full-length Kv4.2 alone, or in the presence of the palmitoylation-deficient PSD-95, displayed significant localization on internal membranes, as did the VSAL deletion mutant, even in the presence of PSD-95. In contrast, C-Kv4.2 exhibited considerable surface expression in the presence of PSD-95 and was clustered (some clusters marked by arrowheads). When ECFP was expressed alone (E) the fluorescence was distributed throughout the cell, whereas mock-transfected cells (F) did not fluoresce.

Kv4.2 Promotes PSD-95 Redistribution to the Cell Surface-- Because the localization of PSD-95 can be affected by other interacting proteins (see "Discussion"), we asked whether Kv4.2 could serve this function. Immunocytochemistry was conducted on fixed transfected HEK293T cells using anti-Myc to detect PSD-95 and anti-GFP to detect C-Kv4.2, C-Kv4.2VSAL, or ECFP. We amplified the fluorescent signal from the ECFP-tagged channels by using anti-GFP and a fluorescent secondary antibody in order to match the intensity of the PSD-95 signal. Formaldehyde fixation and permeabilization were required in order to label the intracellular epitopes on PSD-95 and Kv4.2, and were performed according to standard procedures (29). PSD-95 was diffusely distributed throughout the cell when transfected alone (Fig. 6A). Cells expressing only C-Kv4.2 exhibited extensive labeling of internal membranes (Fig. 6B). When PSD-95 and C-Kv4.2 were co-transfected, both proteins were well expressed at the cell surface (Fig. 6C, arrowheads). This redistribution did not occur when the VSAL motif was removed from Kv4.2 (Fig. 6D). Recruitment of PSD-95 to the surface membrane did not result from the ECFP used to tag the Kv4.2 protein because, when co-transfected with the pECFP vector (Fig. 6E), PSD-95 displayed a diffuse distribution similar to that seen for PSD-95 alone (Fig. 6A).

View larger version (79K): [in this window] [in a new window]   Fig. 6.   Kv4.2 promotes PSD-95 redistribution to the cell surface. Immunofluorescent microscopy of fixed, transfected HEK293T cells using deconvolution deblurring, as in Fig 5. Anti-Myc and an Alexa Fluor 594-conjugated secondary antibody were used to detect PSD-95 (top panels). Anti-GFP and an Alexa Fluor 488-conjugated secondary antibody were used to detect C-Kv4.2, C-Kv4.2VSAL, or ECFP (bottom panels). Imaged z stacks were acquired from cells transfected with PSD-95 (A) or C-Kv4.2 alone (B), PSD-95 plus C-Kv4.2 (C), PSD-95 plus C-Kv4.2VSAL (D), and PSD-95 plus ECFP (E). The z stacks were deconvolved, and a single image was selected for presentation (scale bars, 5 µm). PSD-95 displayed a diffuse cytoplasmic distribution when transfected alone (A) or when co-transfected with the Kv4.2 VSAL deletion mutant (D) or with ECFP (E). ECFP-tagged Kv4.2 (C-Kv.42) exhibited significant localization on internal membranes when expressed alone (B), as did the VSAL deletion mutant, even in the presence of PSD-95 (D). In contrast, both PSD-95 and C-Kv4.2 exhibited considerable surface expression when co-expressed (C, arrowheads).

We noted some differences between cells imaged "live" and those fixed and permeabilized. The Kv4.2 clustering observed in live cells was not apparent in fixed cells (compare Fig. 5B with 6C, lower panel). One possible reason is that the fluorescence signal from live cells was low, making the small clusters easier to discriminate. In contrast, when co-labeling fixed cells, the fluorescence signal was amplified using antibodies with high intensity Alexa fluorophores, which will produce more overlap in the fluorescence signal. Also, in some fixed cells the ECFP-tagged channel protein was found in a perinuclear ring (Fig. 6, C and D), which was not observed in live cells. The fluorescence amplification used in the cells may have improved the resolution needed to see the perinuclear localization or fixation and permeabilization may have altered the channel protein distribution. This observation reinforces the usefulness of fluorescent probes that can be monitored in vivo, such as ECFP.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have demonstrated a physical interaction between Kv4.2 channels and PSD-95 heterologously expressed in mammalian cells, and we have presented biochemical evidence that this interaction has functional consequences: Kv4.2 becomes clustered and its surface expression increases. Through mutation analysis, we showed that this association requires the VSAL motif at the C terminus of the channel and palmitoylation of PSD-95. Thus, VSAL can be added to the list of PSD-95-binding motifs in proteins. VSAL is similar to the hydrophobic PSD-95-binding motifs of some other Kv channels (Kv1.1, Kv1.2, and Kv1.3) but with the terminal valine replaced with leucine. A terminal leucine is important for PDZ binding in another ion channel: binding of the cystic fibrosis transmembrane conductance regulator to ezrin-radixin-moesin phosphoprotein 50 (40).

The increased cell surface expression of Kv4.2 in the presence of PSD-95 could result from enhanced trafficking to the plasma membrane or stabilization of Kv4.2 at the surface. Previous studies indicate that PSD-95 suppresses internalization of ₁-adrenergic receptors (41) and of Kv1.4 (15), another neuronal Kv channel, but one that is mainly axonal in distribution. Unfortunately, the endocytosis rate of Kv4.2 could not be determined using the methods applied to Kv1.4 (biotinylation of an N-linked glycosylated residue) or to the ₁-adrenergic receptor (internalization of an antibody bound to an external epitope). All currently available Kv4.2 antibodies label internal epitopes (this paper and Refs. 4, 6, 18, 38, and 42) and Kv4.2 is not glycosylated (18). Nevertheless, the same protein regions were critical for suppressing Kv1.4 endocytosis (15) and for the PSD-95-mediated increase in Kv4.2 surface expression seen in the present study, a C-terminal motif on the channel and palmitoylation of PSD-95. Interestingly, some other proteins that interact with Kv4.2 increase the current density, i.e. the K⁺ channel interacting protein, KChIP (36), and filamin, a cytoskeletal protein (43), raising the intriguing possibility that all these interacting proteins increase expression at the cell surface. Alternatively, binding of the PDZ-containing PSD-95 protein might mask an endoplasmic reticulum retention signal, as recently shown for some NMDA receptors (44); however, this seems unlikely because channels lacking the VSAL motif were expressed on the cell surface.

It is notable that PSD-95 can undergo redistribution in the presence of interacting proteins in heterologous expression systems. PSD-95 is diffusely distributed throughout the cytoplasm when expressed alone, as demonstrated in this and several other studies (12, 14). However, when co-expressed with Kv1.4, neuroligin, or ₁-adrenergic receptors, PSD-95 localization on the cell surface is greatly increased (41, 45, 46). In our study a similar redistribution was also apparent when PSD-95 and Kv4.2 were co-expressed, which suggests that each protein can influence the distribution pattern of its interacting partner. Interestingly, fixation of transfected cells, which was required for immunofluorescent detection of PSD-95, abolished the clustering observed in the live cells. This might explain why a previous study (13) that examined Kv4.2 interactions with PSD-95 failed to detect an association between these two proteins.

What physiological roles might there be for interaction between PSD-95 and Kv4.2? An obvious role is to establish or maintain specialized channel distributions. In supraoptic neurons, Kv4.2 is clustered at the post-synaptic membrane (47), particularly at sites of synaptic contact, and in hippocampal neurons, where Kv4.2 is also present (3, 4, 18), PSD-95 is only present in post-synaptic regions (48). Because Kv4.2 is somatodendritic and we found that clustering of Kv4.2 was induced by PSD-95 in transfected cells, it might also do so at the post-synaptic membrane. PSD-95 clusters other proteins, including NMDA receptors (49) and Sema4C (50) in native tissues, and NMDA receptors (49, 51), Kir2.1 (52), Kir2.3 (52), and Kv1 channels (12) in heterologous expression systems. The diverse types and distribution patterns of proposed binding partners have been puzzling because PSD-95 was previously thought to be concentrated at post-synaptic sites. However, a recent examination of PSD-95 distribution using electron microscopy revealed a surprisingly high amount in non-synaptic sites, for instance in axons (53). Thus, in principle, PSD-95 might interact in vivo with both somatodendritically located proteins, such as Kv4.2, and axonally located proteins, such as some Kv1 channels.

Kv4.2 is increasingly implicated (3, 4, 18, 54, 55) as one molecular correlate of the dendritic A-type K⁺ channel in pyramidal hippocampal neurons, which regulates spike frequency through its activity at sub-threshold membrane potentials (5, 56, 57) and controls the amplitude and back propagation of action potentials (5). The dendritic A-type K⁺ channel strongly influences the induction of long term potentiation by NMDA receptors (5, 56, 57); thus, it is significant that PSD-95 may position Kv4.2 close to NMDA receptors where it can facilitate these functions.

Another important possibility is that PSD-95 influences the effect of Kv4.2 on neuronal activity by forming complexes between Kv4.2 and signaling molecules. Like many Kv channels, the biophysical properties and activity of Kv4.2 are modified by diverse protein kinases (6, 38, 58). PSD-95 nucleates a macromolecular complex containing one of these kinases, cAMP-dependent protein kinase, because it contains a binding site for the protein kinase A-anchoring protein, AKAP79/150 (59). This interaction targets cAMP-dependent protein kinase to excitatory synapses (60) to facilitate modulation of NR2B-containing NMDA receptors (59). Multiprotein signaling complexes containing Kv4.2 are likely to be key players in the modulation of neuronal function. Although we did not detect a significant change in current, PSD-95 affects Kv4.2 localization (increased cell surface expression and clustering) and may thereby affect its post-insertional modulation (by forming complexes with signaling molecules).

	ACKNOWLEDGEMENTS

We are very grateful X.-P. Zhu for expert technical assistance and to Dr. E. F. Stanley for advice and use of the deconvolution microscope. We thank Dr. R. Khanna for helpful comments on the manuscript and for earlier electrophysiology experiments that are not included here.

	FOOTNOTES

^* This work was supported in part by Canadian Institutes of Health Grant MT13657, Heart and Stroke Foundation of Canada Grant T-3726 (to L. C. S.), and a Natural Sciences and Engineering Research Council grant (to O. T. J.). Parts of this work were previously published as abstracts (61, 62).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^¶ Supported by an Ontario Graduate Scholarship and a University of Toronto Connaught Scholarship.

^** Supported by a National Science Foundation Graduate Student Fellowship.

Current address: Division of Neuroscience, School of Biological Sciences, University of Manchester, Manchester M13 9PT, UK.

^§§ To whom correspondence should be addressed: MC9-415, Toronto Western Hospital, 399 Bathurst St., Toronto, Ontario M5T 2S8, Canada. Tel.: 416-603-5800, ext. 2052; Fax: 416-603-5745; E-mail: schlicht@uhnres.utoronto.ca.

Published, JBC Papers in Press, March 28, 2002, DOI 10.1074/jbc.M109412200

² P. Backx, personal communication.

	ABBREVIATIONS

The abbreviations used are: Kv, voltage-gated K⁺ channel; C-Kv4.2, ECFP-tagged Kv4.2; HEK, human embryonic kidney cells; tsA201, human embryonic kidney cells expressing the large T antigen; IP, immunoprecipitation; PBS, phosphate-buffered saline; CHO, Chinese hamster ovary; GFP, green fluorescent protein; VSAL, Val-Ser-Ala-Leu; ECFP, enhanced cyan fluorescent protein; NMDA, N-methyl-D-aspartic acid.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES