Biochimica et Biophysica Acta 1783 (2008) 1884–1892

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r.com/ locate /bbamcr

Identification of a functional interaction between Kv4.3 channelsand c-Src tyrosine kinase

Pedro Gomes a,⁎, Tomoaki Saito a, Cris del Corsso a, Abderrahmane Alioua a, Mansoureh Eghbali a,Ligia Toro a,b,d,e, Enrico Stefani a,c,d,e

a Department of Anesthesiology, Division of Molecular Medicine, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA 90095-1778, USAb Department of Molecular and Medical Pharmacology, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA 90095-1778, USAc Department of Physiology, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA 90095-1778, USAd Brain Research Institute, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA 90095-1778, USAe Cardiovascular Research Laboratory, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA 90095-1778, USA

Abbreviations: Kv, voltage-dependent K+; HEK293-Kcell line stably expressing Kv4.3 channels; SH2 and SH3,GST, glutathione S-transferase; PAGE, polyacrylamide ge⁎ Corresponding author. Present address: Institute of P

Faculty of Medicine, University of Porto, Alameda ProfessPorto, Portugal. Tel.: +351 225 513 642; fax: +351 225 513

E-mail address: pgomes@med.up.pt (P. Gomes).

0167-4889/$ – see front matter © 2008 Elsevier B.V. Aldoi:10.1016/j.bbamcr.2008.06.011

a b s t r a c t

a r t i c l e i n f o

Article history:

Voltage-gated K+ (Kv) chann Received 2 February 2008Received in revised form 2 June 2008Accepted 2 June 2008Available online 20 June 2008

Keywords:Potassium channelsKv4.3Protein tyrosine kinasesc-SrcPhosphorylationProtein-protein interactions

els are key determinants of cardiac and neuronal excitability. A substantial bodyof evidence has accumulated in support of a role for Src family tyrosine kinases in the regulation of Kvchannels. In this study, we examined the possibility that c-Src tyrosine kinase participates in the modulationof the transient voltage-dependent K+ channel Kv4.3. Supporting a mechanistic link between Kv4.3 and c-Src,confocal microscopy analysis of HEK293 cells stably transfected with Kv4.3 showed high degree of co-localization of the two proteins at the plasma membrane. Our results further demonstrate an associationbetween Kv4.3 and c-Src by co-immunoprecipitation and GST pull-down assays, this interaction beingmediated by the SH2 and SH3 domains of c-Src. Furthermore, we show that Kv4.3 is tyrosine phosphorylatedunder basal conditions. The functional relevance of the observed interaction between Kv4.3 and c-Src wasestablished in patch-clamp experiments, where application of the Src inhibitor PP2 caused a decrease in Kv4.3peak current amplitude, but not the inactive structural analogue PP3. Conversely, intracellular application ofrecombinant c-Src kinase or the protein tyrosine phosphatase inhibitor bpV(phen) increased Kv4.3 peakcurrent amplitude. In conclusion, our findings provide evidence that c-Src-induced Kv4.3 channel activationinvolves their association in a macromolecular complex and suggest a role for c-Src-Kv4.3 pathway inregulating cardiac and neuronal excitability.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

Kv channels are a complex and heterogeneous family of membraneproteins which are able to sense and respond to changes in membranepotential by undergoing conformational changes leading to poreopening or closing thereby regulating K+ ion flow. Kv4.3, a member ofthe Shal subfamily of Kv channels with predominant proteindistribution in brain and heart tissues, has emerged as the leadingcandidate K+ channel gene encoding the fast transient outwardcurrent referred to as A-type current in the nervous system andtransient outward K+ current (Ito) in the heart [1,2]. Evidence gatheredin the last few years suggests that transient outward K+ currentsencoded by the Kv4.x gene family in both neuronal and cardiac cells

v4.3, human embryonic kidneySrc homology domains 2 and 3;l electrophoresisharmacology and Therapeutics,or Hernâni Monteiro, 4200-319643.

l rights reserved.

can be modulated by protein kinases [3–8], to fine-tune Kv channelactivity. Tyrosine phosphorylation along with serine/threonine phos-phorylation has been gradually accepted as an important regulatorymechanism in the control of membrane excitability and ion channelfunction [9–11]. The non-receptor protein tyrosine kinase (PTK) c-Srcis a prototypical member of a family of nine closely relatedmembrane-bound kinases defined by a common structure that, inaddition to a catalytic kinase domain, includes amino-terminalregulatory regions termed Src homology 2 (SH2) and 3 (SH3) domains(for a review see [12]). These modular domains mediate intramole-cular and intermolecular interactions that are important in signaltransduction. Voltage- and ligand-gated ion channels have beenreported to interact with the SH2 and SH3 domains of Src kinase andto cause biophysical changes in channel activity due to the interactionor to tyrosine phosphorylation by Src [13–17]. Kv4.3 sequencecontains putative SH2 and SH3 domain binding motifs, makingKv4.3 a strong candidate for direct interaction with and/or phosphor-ylation by c-Src. It should be stressed that both gating properties andsurface expression of several ion channels can be regulated viaprotein-protein interactions with a variety of associated partners

1885P. Gomes et al. / Biochimica et Biophysica Acta 1783 (2008) 1884–1892

[18,19]. In addition, it is expected that the signal transduction complexinvolving an ion channel and a protein kinase will lead to a higherprobability of channel phosphorylation after kinase activation [15,17].

In the present study we investigated whether c-Src tyrosine kinasemight play a role in the signaling mechanism underlying Kv4.3regulation using a stable expression strategy. We show, through GSTpull-down assays and co-immunoprecipitation, that Kv4.3 proteinassociates with c-Src and that the SH2 and SH3 domains of the kinasemediate this interaction. This association has important functionalimplications, as it may result in enhanced efficiency of Kv4.3phosphorylation by c-Src leading to rapid modulation of Kv4.3 channelactivity in response to signaling pathways linked to c-Src activation.

2. Materials and methods

2.1. Reagents

Primary antibodies used in this study included rabbit polyclonalantibodies against Kv4.3 (Chemicon, Temecula, CA and Alomone Labs,Jerusalem, Israel), mouse monoclonal anti-Kv4.3 (provided by Dr.James S. Trimmer), monoclonal anti-c-Src (clone GD11, UpstateBiotechnology, Lake Placid, NY), polyclonal anti-Src (SC-18, SantaCruz Biotechnology, Santa Cruz, CA), monoclonal anti-phosphotyr-osine (pY) (clone 4G10, Upstate), polyclonal anti-Src-pY416 (CellSignaling, Beverly, MA), polyclonal anti-GST (Santa Cruz), andpolyclonal anti-GAPDH (Novus Biologicals, Littleton, CO). Primaryantibodies were tested for specificity and cross immunoreactivity.Secondary antibodies for Western blot were Alexa Fluor 680 anti-rabbit (Invitrogen, Carlsbad, CA) and IRDye 800 anti-mouse (RocklandImmunochemicals, Gilbertsville, PA) and immunocytochemistry wererhodamine-conjugated donkey anti-rabbit IgG and FITC-conjugateddonkey anti-mouse IgG (Jackson ImmunoResearch Laboratories, WestGrove, PA). PP2, PP3 and bpV(phen) were obtained from Calbiochem(San Diego, CA). Purified recombinant c-Src kinase was from Upstate.GST, GST-Src-SH2 and GST-Src-SH3 fusion proteins were purchasedfrom Panomics (Redwood City, CA). All other reagents were purchasedfrom Sigma (St. Louis, MO) unless otherwise noted.

2.2. HEK293 cells stably expressing Kv4.3 channels

Because HEK293 cells express endogenous c-Src but not Kv4.3channels, a stable HEK293 cell line with robust expression of Kv4.3channel long form (HEK293-Kv4.3) was generated as previouslydescribed [20] and used in the present work. HEK293-Kv4.3 cells werecultured in 100-mm tissue culture dishes in Dulbecco's ModifiedEagle's Medium with L-Glutamine (supplemented with 10% heat-inactivated fetal bovine serum, 400 μg/ml G418, 100 units/mlpenicillin, 100 μg/ml streptomycin) and kept in a humidified 5% CO2

incubator at 37 °C. Cells were passaged every 6–7 days. For patch-clamp and immunocytochemistry experiments, cells were plated onglass coverslips precoated with collagen/poly-D-lysine and used 1–2 days later.

2.3. Western blot analysis

HEK293-Kv4.3 cells were washed twice in ice-cold PBS and thenscraped into 1 ml of cold lysis buffer (1% NP-40, 1% Na-deoxycholate,10% glycerol, 150 mM NaCl, 50 mM Tris, 5 mM EDTA) supplementedwith 50 mM NaF, 5 mM Na3VO4, 2 mM PMSF, 1 mM iodoacetamide,0.15 mM aprotinin, 0.15 mM pepstatin and complete proteaseinhibitor cocktail (Roche). The homogenate was cleared by centrifuga-tion at 16,000 g for 20 min. The supernatant represents the whole-celllysate. In some experiments, cytosol and crude membrane fractionswere prepared. Briefly, cells were homogenized in M1 solution(0.25 M sucrose, 20 mM HEPES, 1 mM EDTA). The homogenate wascentrifuged at 1000 g for 10 min and the resulting supernatant at

100,000 g for 60 min. The supernatant was collected (cytosolicfraction). The pellet, representing the crude membrane fraction, wassolubilized in lysis buffer (see composition above). Equal amounts oftotal protein were denatured in SDS sample buffer, resolved on 10%SDS-PAGE gels and transferred to nitrocellulose membranes. Blotswere blocked with Tris-buffered saline (TBS: 50mM Tris–HCl, 150 mMNaCl, at pH 7.4) containing 5% nonfat dry milk for 1 h at roomtemperature. Thereafter, blots were incubated with indicated primaryantibodies in 1% nonfat milk/TBS with 0.5% Triton X-100 and 0.1%Tween 20 overnight at 4 °C, washed with TBS three times for 10 mineach, and then incubated with fluorescence-conjugated secondaryantibodies for 1 h. After washing, Western blots were imaged with anOdyssey infrared imaging system (LI-COR Biosciences, Lincoln, NE).The specificity of the antibodies was tested by preadsorbing with theantigenic peptide. For c-Src, antigenic peptide was unavailable, andnegative controls were obtained by omitting the primary antibody.Bands corresponding to the immunoreactive proteins were quantifiedusing the Odyssey software.

2.4. Immunocytochemistry

HEK293-Kv4.3 cells grown on glass coverslips were washed twicewith PBS, fixed in 4% paraformaldehyde in PBS for 20 min,permeabilized in 0.2% Triton X-100 in PBS for 30 min, and blockedin 5% normal donkey serum/0.2% Triton X-100 in PBS for 30 min atroom temperature. The cells were stained with specific anti-rabbitpolyclonal against Kv4.3 or anti-mousemonoclonal against pYor c-Srcovernight at 4 °C. After washing with 0.2% Triton X-100 in PBS threetimes for 5 min each, the coverslips were stained with rhodamine-conjugated donkey anti-rabbit IgG and FITC-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch Laboratories) for 1 h at roomtemperature. Cells were washed in PBS three times for 5 min each andmounted using Prolong antifade (Molecular Probes) and observedwith an Olympus IX-70 confocal fluorescence microscope. Confocalimages were acquired every 0.1 μm in the z plane. Images were 3Dblind deconvoluted using Auto Quant software.

2.5. GST pull-down assay

HEK293-Kv4.3 cells were washed twice in ice-cold PBS and celllysates were prepared by adding 1 ml of lysis buffer (see compositionabove). GST fusion proteins (3 μg) coupled to glutathione sepharose 4Bbeads (Amersham Biosciences, Uppsala, Sweden) were incubated with2 mg of cell lysates overnight at 4 °C with constant shaking. The beadswere washed four times in 1 ml of ice-cold lysis buffer and immunecomplexes were released from the beads by boiling in SDS samplebuffer for 5min. Pull-downproducts were separated on 10% SDS-PAGEgels and then analyzed byWestern blotting using anti-Kv4.3 and anti-GST specific antibodies.

2.6. Co-immunoprecipitation

For co-immunoprecipitation experiments, HEK293-Kv4.3 cells orHEK293-Kv4.3 cells transfected with c-Src were harvested 48 h post-transfection, washed twice with ice-cold PBS and scraped into lysisbuffer (see composition above) on ice and clarified by centrifugation.Total lysate (1 mg) was added to 50 μl of protein A agarose (PierceBiotechnology, Rockford, IL) and either 5 μg of monoclonal anti-Kv4.3 or 5 μg of monoclonal anti-c-Src. A control for nonspecificimmunoprecipitation was carried out in the absence of primaryantibody. After overnight incubation at 4 °C with constant shaking,immune complexes were washed twice with lysis buffer and twicewith wash buffer (lysis buffer with 0.1% NP-40) and then extractedwith SDS sample buffer or resuspended in kinase assay buffer for invitro kinase experiments. Extracts were resolved on 10% SDS-PAGEgels, transferred to nitrocellulose membranes, and immunoblotted

Fig. 1. Kv4.3 and c-Src share the same subcellular localization in HEK293 cells stablytransfected with Kv4.3 (HEK293-Kv4.3). (A–C) Co-localization of Kv4.3 and c-Src byconfocal microscopy. HEK293 cells transfected with Kv4.3 grown on coverslips werefixed with 4% paraformaldehyde and then double-stained with anti-Kv4.3 rabbitpolyclonal antibody (A, red) and anti-c-Src mouse monoclonal antibody (B, green). (C)Overlay of A and B showing co-localization of Kv4.3 and c-Src (yellow). (D) Fluorescenceintensity profiles along the white dashed line showing that Kv4.3 and c-Src labeling arepredominantly co-localized at or near the cell membrane. (E, F) Kv4.3 and c-Src areenriched in the crude membrane (CM) fractionwhen comparedwith the cytosolic (CYT)fraction. (G, H) Kv4.3 and c-Src signals were absent when the anti-Kv4.3 antibody wasincubated with the antigenic peptide and the anti c-Src antibody was omitted. Equalamounts of protein (50 μg) from each fraction were loaded in each lane.

1886 P. Gomes et al. / Biochimica et Biophysica Acta 1783 (2008) 1884–1892

with monoclonal antibody against pY and either polyclonal anti-bodies against Kv4.3 or Src.

2.7. Biotinylation

For biotinylation studies, cells were treated with vehicle, PP2, PP3,bpV(phen) or transfected with c-Src, and washed twice in PBS2+ (PBSsupplemented with 1 mM MgCl2 and 1 mM CaCl2). Live cells werebiotinylated with 1 mg/ml EZ-Link sulfo-NHS-SS-biotin (PierceBiotechnology) in PBS2+ for 30 min at 4 °C. Cells were then washedthree times with 100 mM glycine in PBS2+ to quench unreacted biotinfollowed by lysis (see composition above). Total biotinylated mem-brane proteins (1 mg) were isolated by incubating lysates with 50 μl ofNeutrAvidin agarose beads (Pierce) overnight at 4 °C. The beads wererinsed three times with lysis buffer (see composition above) beforeproteins were eluted with SDS sample buffer.

2.8. Electrophysiology

Whole-cell voltage-clamp recordings were obtained using patch-clamp amplifier (AxoPatch 200A; Axon Instruments, Union City, CA) atroom temperature. Pipettes were made from glass (PG52150–4; WPI,Sarasota, FL) and had resistance of 2.5–3.5 MΩ. Voltage-clamp bathsolution was (in mM): 140 Na-MES, 2 KCl, 2 CaCl2, 10 glucose and 10HEPES, pH 7.4. The pipette solution contained (in mM): 134 K-Methanesulphonate (K-MES), 6 KCl, 10 HEPES and 10 EGTA, pH 7.2.Analog signals were filtered at one-fifth of the sampling frequency (1to 10 kHz). Series resistance were electronically compensated (∼85%);voltage errors resulting from the uncompensated series resistancewere b10 mV. Current-voltage relationship for the Kv4.3 current wasconstructed from the current changes produced by a 500 ms voltage-clamp step applied in 10 mV increment from the holding potential of−70 mV. Current recordings were all obtained at a frequency rate of0.1 Hz. Data acquisition and analysis were performed using custommade software.

2.9. Statistics

Data were expressed as mean±SEM. Comparisons between twogroups were analyzed by Student's t-test. Statistical significance wasset at Pb0.05.

3. Results

3.1. Co-localization of Kv4.3 with c-Src at the plasma membrane inheterologously expressed cells

We propose that the Kv4.3 channel might be a new target for c-Srcmodulation. Accumulating evidence suggests that many ion channelsreside within a multiprotein complex that contains kinases and othersignaling molecules [21,22]. To address whether Kv4.3 and c-Src sharea common subcellular distribution, we used an HEK293 cell line stablyexpressing the long form of Kv4.3 (HEK293-Kv4.3) [20] and examinedthe subcellular localization of Kv4.3 and c-Src proteins by confocalmicroscopy. An antibody specific to Kv4.3 revealed a continuouspattern of immunofluorescence staining suggesting a predominantlocalization of this protein at the plasma membrane (Fig. 1A).Similarly, anti-c-Src antibody labeling revealed a major pool of c-Srcprotein present at the plasma membrane as well as a minor pooldiffused intracellularly (Fig. 1B). Interestingly, overlapping bothimages showed that both proteins displayed co-localization (indicatedby the yellow color) at the plasma membrane (Fig. 1C) that isconfirmed by the fluorescence intensity plot (Fig. 1D). In addition, cellfractionation experiments were carried out to further confirm thatKv4.3 and c-Src proteins share the same subcellular microenviron-ment. As expected from the immunocytochemistry data, Kv4.3 protein

is detected exclusively in crude membrane but not in cytosolicfractions (Fig. 1E). A Kv4.3 channel specific antibody labeled a band at∼68 kDa, which was completely blocked when the antibody waspreincubated with its corresponding antigenic peptide (Fig. 1G). Onthe other hand, c-Src antibody labeled a ∼60 kDa band in bothmembrane and cytosolic fractions being more abundant in membranefractions (Fig. 1F). Antigenic peptide for c-Src antibody was notavailable; alternatively, we omitted the primary antibody for negativecontrol. Under these conditions, the ∼60 kDa band was undetectable(Fig. 1H). These results suggest that Kv4.3 and c-Src reside on the samecellular compartment at the plasma membrane, raising the possibilityof a functional interaction between the two proteins.

1887P. Gomes et al. / Biochimica et Biophysica Acta 1783 (2008) 1884–1892

3.2. Physical interaction between Kv4.3 and c-Src

Having determined that Kv4.3 can be targeted together with c-Srcto the plasma membrane (Fig. 1), we investigated the physicalinteraction between these proteins. Kv4.3 sequence has a total of 24tyrosine residues, three of which lie within consensus motifs fortyrosine-specific phosphorylation (YXX ψ, where ψ can be leucine,isoleucine or valine [13,23] that could serve upon their phosphoryla-tion as Kv4.3 docking sites for Src kinase SH2 domain (Fig. 2A, toppanel). Furthermore, three proline-rich sequences in Kv4.3 bearingthe PXXP motif, one in the amino-terminal (26–31) and two in thecarboxy-terminal (618–621; 630–636) of the channel protein couldalso interact with Src kinase SH3 domain (Fig. 2A, bottom panel)[13,24]. We first examined whether Kv4.3 could interact with c-Src byperforming an in vitro binding assay using GST fusion proteins withthe Src SH2 and SH3 domains (GST-Src-SH2 and GST-Src-SH3). Pull-down experiments performed on lysates prepared from HEK293-Kv4.3 cells showed that Kv4.3 is efficiently precipitated by both GST-Src-SH2 and GST-Src-SH3 but not by GST alone (Fig. 2B) stronglysupporting an interaction between Kv4.3 and Src-SH2 and Src-SH3domains.

To determine whether Kv4.3-c-Src complexes also occur in intactcells, we performed co-immunoprecipitation studies in lysates fromHEK293-Kv4.3 cells with endogenous or overexpressed c-Src (Fig. 2C,left panel). As evidenced by theWestern blot analysis, cells transfectedwith Kv4.3 alone or Kv4.3 together with c-Src express similar levels ofKv4.3 protein (Fig. 2C) indicating that overexpression of c-Src does notalter expression of Kv4.3. Cell lysates from both cell types were

Fig. 2. In vitro and in vivo association between Kv4.3 and c-Src. (A) Kv4.3 consensus sequenceSH3 fusion proteins. Lanes are: cell lysate alone (50 μg protein) or cell lysate incubated with ebeads). Bound Kv4.3 to the GST fusion proteins was precipitated and immunodetected with aas a loading control for GST fusion proteins. (C) Co-immunoprecipitation of Kv4.3 with c-Src. Knative or transfected c-Src. Cell lysates (1 mg protein) were incubated with 5 μg of anti-Kv4.3agarose beads, separated on 10% SDS-PAGE, and immunoblotted with polyclonal anti-Kv4.3extracts (input) used for immunoprecipitation. Middle panels are the immunoprecipitated prpanels show control reactions carried out in the absence of anti-Kv4.3 (No Ab). (D) Thedensitometric analysis and the ratios plotted to determine the degree of associated c-Src no

subjected to immunoprecipitation with anti-Kv4.3 antibody (IP:Kv4.3; middle panel) or without antibody as control (IP: No Ab;right panel). The cell lysates and immunoprecipitates were subjectedto immunoblot analysis with specific antibodies to c-Src (Fig. 2C, toppanel) or Kv4.3 (Fig. 2C, bottom panel). As shown in Fig. 2C, anti-Kv4.3antibody efficiently precipitated Kv4.3 from lysates derived from cellsexpressing Kv4.3. A small but detectable amount of endogenous c-Srcwas also co-immunoprecipitated together with Kv4.3 which becamemore evident after co-expressing c-Src (8-fold increase; Fig. 2C and D)suggesting that c-Src is a molecular partner of Kv4.3. These resultsprovide strong evidence supporting the close association betweenKv4.3 and native or co-expressed c-Src kinase.

3.3. Kv4.3 is tyrosine phosphorylated

There is growing evidence supporting a role for Src family PTKs inthe regulation of several ion channel activities by direct tyrosinephosphorylation [25–28]. Having confirmed an interaction betweenKv4.3 and c-Src (Fig. 2), we next sought to determine whether Kv4.3channels could be tyrosine phosphorylated. As an initial approach totest this hypothesis, we examined the subcellular localization of Kv4.3and phosphotyrosine-containing proteins (pY) by immunofluores-cence confocal microscopy. We found that immunoreactivities of bothKv4.3 and pY proteins were present mainly at the plasma membranedisplaying a high degree of co-localization (Fig. 3A) raising thepossibility that Kv4.3 may serve as a potential substrate for plasmamembrane resident tyrosine kinases (e.g. c-Src). To further corrobo-rate the immunocytochemistry data (Fig. 3A), cells lysates were

s for interactionwith c-Src SH2 and SH3 domains. (B) Kv4.3 pull-down by GST-Src-SH2/ither GST as a control or GST-Src-SH2 or SH3 (GST was coupled to glutathione sepharosenti-Kv4.3 (top panel). The same blot was probed with anti-GST antibody (bottom panel)v4.3 was immunoprecipitated (IP) from lysates of stable HEK293-Kv4.3 cells expressingmouse monoclonal antibody. The immunocomplexes were then captured on protein Aand monoclonal anti-c-Src antibodies. Left panels show immunoblots of the same celloteins with anti-Kv4.3 antibody probed with anti-c-Src and anti-Kv4.3 antibodies. Rightc-Src and Kv4.3 blotting signals of the immunoprecipitates were quantified usingrmalized to the amount of immunoprecipitated Kv4.3.

Fig. 3. Kv4.3 is tyrosine phosphorylated. (A) HEK293-Kv4.3 cells grown on coverslipswere fixed and double-stained with anti-Kv4.3 (red) and anti-phosphotyrosine (pY,green) antibodies. An overlay of the two images is shown. Fluorescence intensityprofiles measured along the white dashed line illustrate a positive correlation betweenKv4.3 and pY fluorescent signals. (B) Kv4.3 and (C) c-Src immunoprecipitates (IP) fromstable HEK293-Kv4.3 cells expressing endogenous c-Src. Each cell lysate (1 mg) wasincubatedwith 5 μg of either anti-Kv4.3 or anti-c-Srcmousemonoclonal antibodies. Theimmunocomplexes bound to protein A agarose beads were separated on 10% SDS-PAGE,and immunoblotted with either polyclonal anti-Kv4.3 or polyclonal anti-c-Src andmonoclonal anti-pY antibodies. Right panels show control reactions carried out in theabsence of antibody (No Ab). Asterisks indicate nonspecific bands containing the heavychain of IgG.

1888 P. Gomes et al. / Biochimica et Biophysica Acta 1783 (2008) 1884–1892

immunoprecipitated with monoclonal anti-Kv4.3 antibody andimmunoblotted with both anti-Kv4.3 and anti-pY (Fig. 3B). Interest-ingly, immunoprecipitated Kv4.3 protein is efficiently tyrosinephosphorylated. As a positive control for tyrosine phosphorylation,immunoprecipitation using anti-c-Src antibody showed that thekinase is efficiently autophosphorylated (Fig. 3C).

3.4. Up-regulation of Kv4.3 currents by constitutively active c-Srctyrosine kinase

To establish a role for tyrosine phosphorylation on Kv4.3channel activity, we examined the effect of Src inhibition onKv4.3 currents using whole-cell patch-clamp recordings fromHEK293-Kv4.3 cells. As reported in our previous work [20], thiscell line expresses robust Kv4.3 currents (Fig. 4A1). Application ofthe membrane-permeable, Src tyrosine kinase inhibitor PP2(10 μM) [29] to HEK293-Kv4.3 cells resulted in a significantreduction of Kv4.3 peak currents (Fig. 4A2). In addition to thereduction in current amplitude, PP2 also altered the inactivationproperties of the channel as can be seen in the superimposedtraces shown in Fig. 4A3 from control (a) and after PP2 treatment(b) at +50 mV. The trace after PP2 treatment was scaled up tomatch the peak of the current in control to better show the changein the inactivation kinetics (Fig. 4A4). The amplitude and the decayof the outward K+ currents recorded at +50 mV were well fittedwith the sum of two exponential functions (fast and slowcomponents) and a steady state current. Both time constants ofthe fast and slow components of inactivation were significantlydecreased by approximately 50% after PP2 application (tfast:control=27±4 ms vs PP2=14±3 ms; tslow: control=139±22 ms vsPP2=81±8 ms; n=9 cells). The integrated area of the fastcomponent at +50 mV was also decreased by PP2 (Afast:control=3.0±0.4 vs PP2=1.7±0.3; n=9 cells), whereas the inte-grated area of the slow component was not affected by PP2 (Aslow:control=1.0±0.1 vs PP2=0.9±0.2). The steady state current was alsosignificantly reduced from 0.28±0.04 in control to 0.15±0.02 inPP2-treated cells. Fig. 4B shows the peak current density as afunction of voltage in control and after application of PP2. The peakcurrent density in PP2-treated patches reduced by ∼25–30%suggesting the existence of a PP2-insensitive pool of channelswhich may be uncoupled from c-Src. In support of this view,application of recombinant c-Src to the whole-cell patches furtherincreased the current density possibly by stimulating theuncoupled Kv4.3 channels (see Fig. 5C). The peak current densityin both control and after PP2 treatment were normalized to peakcurrent density in control condition at +50 mV (Imax) to measurethe channel fractional open probability (Po). The plot of I/Imaxcontrolshowed much lower I/Imaxcontrol in the presence of PP2 (Fig. 4C).Interestingly, when the same peak currents were normalized totheir corresponding Imax, the curves were superimposed indicatingthat the voltage dependence of channel opening remainedunchanged (Fig. 4C, inset). These findings suggest that Kv4.3channels that are not functionally coupled to c-Src have the samevoltage dependence of activation and the current reduction can beexplained by the opening of fewer channels or by channels with asmaller limiting Po value having a faster inactivation rate.

To further emphasize that the effects shown in Fig. 4A–C are due toPP2, we used PP3, an inactive analogue of PP2, as negative control.Application of 10 μMPP3 did not alter themagnitude or time course ofmembrane currents (Fig. 4D and E).We nextmeasured c-Src activity inthe HEK293-Kv4.3 cell line under control (Ct), PP2 or PP3 treatmentsby immunoblotting using an antibody against tyrosine phosphory-lated proteins (pY) or c-Src autophosphorylation at pY-416, a processindicative of c-Src activation state [30]. The Src inhibitor PP2 (10 μM;30 min) almost completely abolished tyrosine phosphorylation ofmultiple proteins (Fig. 4F) as well as autophosphorylation of c-Src atTyr-416 residue (Fig. 4G), whereas the inactive analogue PP3 (10 μM;30 min) had no effect (Fig. 4F and G). This result is in agreement withprevious reports showing that PP2, but not PP3, inhibits c-Src activityanalyzed by Western blotting with anti-c-Src-pY416 [31]. Thedecrease in c-Src autophosphorylation is not due to differentialexpression of c-Src because total c-Src protein levels were not affectedby treatment with either PP2 or PP3 (Fig. 4G).

Fig. 4. PP2, a selective inhibitor of Src family tyrosine kinases, reduces Kv4.3 currents. (A) Kv4.3 currents recorded from the same cell before (A1) and after (A2) exposure to 10 μMPP2in a HEK293 cell expressing Kv4.3. Cells were voltage-clamped in the whole-cell configurationwith holding potential (Vh) of −70 mV and stepped from −60 mV to +50 mV in 10-mVincrements with a pulse duration of 500 ms. The voltage step protocol is shown above the traces. Note that outward K+ peak currents become smaller after exposure to PP2. (A3) Thecurrent traces at +50 mV were superimposed for illustration purposes to show the change in both current amplitude and inactivation kinetics before (a) and after PP2 (b) treatment.(A4) The current trace after PP2 treatment (b) was scaled up to control (a) to better illustrate the change in the inactivation kinetics. (B) Current-voltage relationship of Kv4.3 currentdensity (pA/pF) before (○) and after (●) PP2 application. (C) Peak current was normalized to the maximum peak current recorded in control at +50 mV (I/Imax), before and after PP2.Peak currents were also normalized to their corresponding maximum peak currents at +50 mV (Inset). (D) PP3, an inactive structural analogue of PP2, does not significantly alter thepeak current of Kv4.3. Current traces to +50 mV showing indistinguishable current in the same cell, before (Control) and after PP3 (10 μM) treatment. (E) Time course of PP2-inducedinhibition and lack of effect of PP3 on Kv4.3 currents. Peak current amplitudes were normalized to the current amplitude before drug treatment (time 0) and plotted against time. (F)Inhibition by PP2 of protein tyrosine phosphorylation in cell lysates incubated for 30 min in the absence (Ct) or the presence of PP2 or PP3 (both at 10 μM). Immunoblot shows theprotein tyrosine phosphorylation status (pY). (G) c-Src activity was assessed with a phospho-specific antibody raised against the region of c-Src containing tyrosine-416, thepresumed site of autophosphorylation. The bar graph shows the ratio of phosphorylated c-Src (c-Src-pY416, top panel) to the amount of c-Src protein (c-Src, bottom panel) on eachcondition, and normalized to control values. ⁎Pb0.05, PP2 vs control and PP3.

1889P. Gomes et al. / Biochimica et Biophysica Acta 1783 (2008) 1884–1892

3.5. Basal c-Src activity in HEK293-Kv4.3 cell line

The previous results strongly suggest that the HEK293-Kv4.3 cellline has a certain level of basal constitutive c-Src activity, which issufficient to phosphorylate either directly or indirectly the Kv4.3channel causing its activation. In support of this view, application ofthe tyrosine phosphatase inhibitor bpV(phen) (10 μM) stimulatedKv4.3 peak currents by 26% in a time-dependent manner (Fig. 5A andB), suggesting that the channel has tyrosine residues that are readilyavailable to be phosphorylated by the constitutive Src kinases. In

addition, to test more directly the involvement of c-Src, wedetermined the effect of intracellular application of the activerecombinant c-Src kinase through the patch-clamp recording pipetteinto Kv4.3-expressing HEK293 cells. Application of purified c-Src(7.5 U/ml) into the patch pipette induced a significant increase by 35%of the Kv4.3 whole-cell current amplitude (Fig. 5C). bpV(phen) isknown to potently inhibit protein tyrosine phosphatases [32] andthereby allowing tyrosine residues to remain in a phosphorylatedstate. Therefore, we hypothesized that if PP2 acts by blocking c-Src-dependent phosphorylation of a critical tyrosine residue in Kv4.3

Fig. 5. Potentiation of Kv4.3 currents by a tyrosine phosphatase inhibitor and by active recombinant c-Src. (A) The tyrosine phosphatase inhibitor bpV(phen) significantly increasedKv4.3 peak current. Representative traces show currents recorded at +50 mV from the same cell, before (Control) and after bpV(phen) (10 μM) treatment. (B) Time course of Kv4.3peak current recorded before (○) and after exposure to 10 μM bpV(phen) (●). Peak current amplitudes were normalized to the current amplitude before treatment at time 0 andplotted against time. (C) Recombinant c-Src tyrosine kinase (7.5 U/ml) introduced into the patch pipette markedly increased Kv4.3 currents. Recordings are from the same cell before(Control), and 3 min after perfusion with recombinant c-Src. (D) Current traces from the same cell recorded before (Control, time 0), after PP2 treatment (3 min) and the subsequentapplication of PP2+bpV(phen) (6 min) at +50 mV. (E, F) Whole-cell lysates prepared from HEK293-Kv4.3 cells incubated for 30 min in the absence (Ct) or presence of 10 μM bpV(phen) or co-transfected with c-Src with antibodies against (E) pYor (F) c-Src and c-Src-pY416. (E) Immunosignals show the increased number of tyrosine phosphorylated proteins incells treated with bpV(phen) or overexpressing c-Src. Likewise, immunosignals detected with anti-pY indicated that the band at ∼60 kDa, presumably c-Src, increased by bpV(phen)treatment. (F) c-Src activity was assessed by immunoblotting of cell lysates with a phospho-specific antibody raised against the region of c-Src containing tyrosine-416, the presumedsite of autophosphorylation. The bar graph shows the ratio of phosphorylated c-Src (c-Src-pY416, top panel) to the amount of c-Src protein (c-Src, bottompanel) under each conditionafter normalization to control values. ⁎ Pb0.05, control vs bpV(phen) and c-Src.

1890 P. Gomes et al. / Biochimica et Biophysica Acta 1783 (2008) 1884–1892

protein, its action should be opposed by bpV(phen). Treatment ofpatches with PP2 (10 μM) reduced Kv4.3 peak currents andsubsequent application of bpV(phen) (10 μM) failed to induce anystimulation of Kv4.3 currents (Fig. 5D). These results suggest that themagnitude of the opposing activities of c-Src and a tyrosinephosphatase could therefore dictate the degree of tyrosine phosphor-ylation of Kv4.3 and its activity. The possibility that the inhibitoryeffect of PP2 on Kv4.3 currents may be due to a partial inhibition ofother protein tyrosine kinases cannot be ruled out. A positivecorrelation between the phosphorylation state of c-Src and Kv4.3channel activity was confirmed by measuring the level of autopho-sphorylated c-Src in lysates fromHEK293-Kv4.3 cells treated with bpV(phen) (10 μM; 30min) or transfected with c-Src (Fig. 5F). Under theseconditions, multiple proteins were phosphorylated on tyrosine

residues (Fig. 5E). The discrepancy between c-Src fold inductionfollowing bpV(phen) treatment or c-Src overexpression (Fig. 5F) andcorresponding effects on Kv4.3 currents might be due to the fact thatonly a small fraction of endogenous c-Src associates with the channel(Fig. 2C, middle panel). Taken together, these data demonstrate thatthe transient outward Kv4.3 current can be constitutively upregulatedby c-Src tyrosine kinase activity.

3.6. c-Src activity does not alter the abundance of Kv4.3 at the plasmamembrane

There is evidence suggesting that signaling pathways can mod-ulate ion channel activity by affecting the intracellular trafficking ofchannel proteins [33–35]. To address whether c-Src modulation of

Fig. 6. Lack of involvement of a Src-dependent synthesis/traffickingmechanism in Kv4.3regulation. Treatment with PP2, PP3, bpV(phen) or overexpression of c-Src does notaffect the number of Kv4.3 channels in either the total or biotinylated (surface)fractions. (A) After cell treatment, plasma membrane was biotinylated followed by celllysis and incubation of lysates with NeutrAvidin agarose beads to recover surfacebiotinylated proteins. Western blots were performed on both total (50 μg protein; left)and surface (recovered from 1 mg protein; right) fractions with anti-Kv4.3 and anti-GAPDH (used as a marker of cytosolic proteins) antibodies. The absence of detectableavidin-bound GAPDH (bottom, right) suggests no contamination by biotinylatedcytosolic proteins. (B) Reactive bands quantified by densitometric analysis arepresented as ratios of surface over total Kv4.3.

1891P. Gomes et al. / Biochimica et Biophysica Acta 1783 (2008) 1884–1892

Kv4.3 macroscopic currents involves remodeling of Kv4.3 channels atthe plasmamembrane, cells were treated with the Src kinase inhibitorPP2 (10 μM; 30 min), the inactive analogue PP3 (10 μM; 30 min), theprotein tyrosine phosphatase inhibitor bpV(phen) (10 μM; 30 min) ortransfected with c-Src, and followed by cell surface biotinylationexperiments. Surface expressed proteins were biotinylated on ice for30 min and then biotinylated proteins were separated by NeutrAvidinbeads, and subsequently analyzed by Western blot. All treatments aswell as c-Src transfection produced no significant changes in theprotein levels of Kv4.3 in either the total (Fig. 6A, left panel) orbiotinylated (Fig. 6A, right panel) fractions as compared withuntreated controls (Fig. 6B). Our results suggest that c-Src does notaffect cell membrane targeting and surface distribution patterns ofKv4.3.

4. Discussion

In the present study we used an HEK293 cell line expressing Kv4.3and provide evidence, for the first time, that Kv4.3 physically interactswith c-Src and that this interaction has the functional consequence ofincreasing channel current amplitude. In addition, we also demon-strate that Kv4.3 undergoes tyrosine phosphorylation.

There is a large body of evidence indicating that Shaker family K+

channels are regulated by tyrosine kinases. When Kv1.3 is co-expressed with v-Src in HEK293 cells, the channel becomes tyrosinephosphorylated and current is suppressed [36]. Similar findings areobserved upon the coexpression of Kv1.5 and Src [15]. In nativemouse Schwann cells, Fyn-mediated tyrosine phosphorylation ofKv1.5 leads to channel activation [37]. Therefore, protein tyrosinephosphorylation appears to be a common important regulatorymechanism for Shaker superfamily channels. In contrast, noinformation is available for Shal-type K+ channels (Kv4.x) regardingmodulation by tyrosine kinases. This superfamily of ion channels hasbeen reported to be a substrate of several protein kinases, namely

PKC [6], PKA [4], calmodulin-dependent protein kinase II [5,7,8] andmitogen-activated protein kinase ERK [3]. In this study, wehypothesized that c-Src may modulate the activity of Kv4.3 via twodistinct mechanisms: i) a direct or indirect protein-protein interac-tion, which is supported by the presence of potential SH3-bindingdomains in Kv4.3; ii) direct phosphorylation of the channel protein,which is consistent with the presence of three tyrosine phosphoryla-tion consensus sites in Kv4.3 sequence. Our work presents evidencesupporting a direct physical interaction between c-Src and Kv4.3channels. The following observations support this conclusion. Firstly,Kv4.3 was found to cluster at the plasma membrane and co-localizewith c-Src. Secondly, endogenous and transfected c-Src co-immuno-precipitate with Kv4.3. Finally, GST pull-down assays showed that theSH2 and SH3 domains of c-Src mediate the interaction between Kv4.3and the kinase. Kv4.3 might also recruit other SH2- and SH3-containing signaling molecules into the macromolecular signalingcomplex to maintain proper Kv4.3 channel regulation. Becausephysical interactions between Src kinases and their substrates arecritical for efficient target phosphorylation [38,39], the interactionbetween Kv4.3 and c-Src may also contribute to the rapid andefficient tyrosine phosphorylation of Kv4.3 channel, thereby regulat-ing the biophysical properties of the channel. In terms of function ourresults indicate that tyrosine phosphorylation of Kv4.3 contributes toan increase in Kv4.3 channel activity. Recombinant active c-Src andthe tyrosine phosphatase inhibitor bpV(phen) caused an upregula-tion of Kv4.3 current amplitude in transfected HEK293 cells, whereasthe Src kinase inhibitor PP2 decreased the current. In vitro studiessuggest that PP2 is far more potent at blocking Lck and Fyn than c-Src[29]. The application of 10 nM PP2 did not affect Kv4.3 currents (notshown). Thus, the lack of effect of nanomolar PP2 on Kv4.3 channelactivity is compatible with an action of the drug on c-Src, but not Lckor Fyn. Because the Kv4.3-c-Src complex resides in the plasmamembrane it is appropriately positioned to receive, and respond to,extracellular signals. This feature explains the fast response of Kv4.3currents to either Src inhibition or stimulation. At present, it isunclear how c-Src activates Kv4.3. The results presented here suggestthat c-Src interaction and/or phosphorylation of Kv4.3 channelstranslates into an effect on current amplitude, and does not seem toaffect voltage dependence and kinetics of activation of Kv4.3. Onepossible scenario is that c-Src kinase controls the number of activechannels or, alternatively, that tyrosine phosphorylation increases thechannel open probability (Po). However, biotinylation experimentssuggest that the abundance of Kv4.3 channels on the plasmamembrane is not regulated by a c-Src-dependent mechanism.Additional studies will be needed to clarify the molecular detailsleading from c-Src to Kv4.3 activation.

Kv4.3 channel activity has been shown to be downregulated inboth physiological and pathological cardiac hypertrophy [40,41].Interestingly, recent work has shown that an increase in Kv4.3 currentdensity by in vivo overexpression of Kv4.3 gene abrogates cardiachypertrophy [42]. In addition, it has also been shown that cardiac c-Srcactivity increases in both physiological [40] and pathological hearthypertrophy [43]. We speculated that the level of cardiac c-Src activityis a key factor in controlling the transition from reversible toirreversible heart hypertrophy [44]. Here, we demonstrate thatKv4.3 channel activity is increased by c-Src. In the context of bothcardiac physiological and pathological hypertrophy, an increase inexcitability due to tyrosine phosphorylation of Kv4.3 may play acompensatory role in which cardiac excitability has beencompromised.

Acknowledgments

This work was supported by NIH grants HD046510, HL088640 (ES);HL077705 (LT); HL080111 (ES, LT) and AHA National Center grants0435116N (ME) and 0435084N (AA). PG was a recipient of a

1892 P. Gomes et al. / Biochimica et Biophysica Acta 1783 (2008) 1884–1892

postdoctoral fellowship funded by FCT (Fundação para a Ciência e aTecnologia, Portugal).

References

[1] J.E. Dixon, W. Shi, H.S. Wang, C. McDonald, H. Yu, R.S. Wymore, I.S. Cohen, D.McKinnon, Role of the Kv4.3 K+ channel in ventricular muscle. A molecularcorrelate for the transient outward current, Circ. Res. 79 (1996) 659–668.

[2] P. Serodio, E. Vega-Saenz de Miera, B. Rudy, Cloning of a novel component of A-type K+ channels operating at subthreshold potentials with unique expression inheart and brain, J. Neurophysiology 75 (1996) 2174–2179.

[3] J.P. Adams, A.E. Anderson, A.W. Varga, K.T. Dineley, R.G. Cook, P.J. Pfaffinger, J.D.Sweatt, The A-type potassium channel Kv4.2 is a substrate for the mitogen-activated protein kinase ERK, J. Neurochem. 75 (2000) 2277–2287.

[4] A.E. Anderson, J.P. Adams, Y. Qian, R.G. Cook, P.J. Pfaffinger, J.D. Sweatt, Kv4.2phosphorylation by cyclic AMP-dependent protein kinase, J. Biol. Chem. 275(2000) 5337–5346.

[5] O. Colinas, M. Gallego, R. Setien, J.R. Lopez-Lopez, M.T. Perez-Garcia, O. Casis,Differential modulation of Kv4.2 and Kv4.3 channels by calmodulin-dependentprotein kinase II in rat cardiac myocytes, Am. J. Physiol Heart Circ. Physiol. 291(2006) H1978–H1987.

[6] T.Y. Nakamura, W.A. Coetzee, E. Vega-Saenz de Miera, M. Artman, B. Rudy,Modulation of Kv4 channels, key components of rat ventricular transient outwardK+ current, by PKC, Am. J. Physiol. 273 (1997) H1775–H1786.

[7] G.P. Sergeant, S. Ohya, J.A. Reihill, B.A. Perrino, G.C. Amberg, Y. Imaizumi, B.Horowitz, K.M. Sanders, S.D. Koh, Regulation of Kv4.3 currents by Ca2+/calmodulin-dependent protein kinase II, Am. J. Physiol. Cell Physiol 288 (2005)C304–C313.

[8] A.W. Varga, L.L. Yuan, A.E. Anderson, L.A. Schrader, G.Y. Wu, J.R. Gatchel, D.Johnston, J.D. Sweatt, Calcium-calmodulin-dependent kinase II modulates Kv4.2channel expression and upregulates neuronal A-type potassium currents, J.Neurosci. 24 (2004) 3643–3654.

[9] S.A. Siegelbaum, Channel regulation. Ion channel control by tyrosine phosphor-ylation, Curr. Biol. 4 (1994) 242–245.

[10] I.B. Levitan, Modulation of ion channels by protein phosphorylation and depho-sphorylation, Annu. Rev. Physiol 56 (1994) 193–212.

[11] E.A. Jonas, L.K. Kaczmarek, Regulation of potassium channels by protein kinases,Curr. Opin. Neurobiol. 6 (1996) 318–323.

[12] G.S. Martin, The hunting of the Src, Nat. Rev., Mol. Cell Biol. 2 (2001) 467–475.[13] T. Pawson, J.D. Scott, Signaling through scaffold, anchoring, and adaptor proteins,

Science 278 (1997) 2075–2080.[14] N.S. Magoski, G.F.Wilson, L.K. Kaczmarek, Protein kinasemodulation of a neuronal

cation channel requires protein-protein interactionsmediated by an Src homology3 domain, J. Neurosci. 22 (2002) 1–9.

[15] T.C. Holmes, D.A. Fadool, R. Ren, I.B. Levitan, Association of Src tyrosine kinasewitha human potassium channel mediated by SH3 domain, Science 274 (1996)2089–2091.

[16] H. Xu, H. Zhao, W. Tian, K. Yoshida, J.B. Roullet, D.M. Cohen, Regulation of atransient receptor potential (TRP) channel by tyrosine phosphorylation. SRC familykinase-dependent tyrosine phosphorylation of TRPV4 on TYR-253 mediates itsresponse to hypotonic stress, J. Biol. Chem. 278 (2003) 11520–11527.

[17] M.N. Nitabach, D.A. Llamas, I.J. Thompson, K.A. Collins, T.C. Holmes, Phosphoryla-tion-dependent and phosphorylation-independent modes of modulation ofshaker family voltage-gated potassium channels by SRC family protein tyrosinekinases, J. Neurosci. 22 (2002) 7913–7922.

[18] G. Michels, F. Er, I.F. Khan, J. Endres-Becker, M.C. Brandt, N. Gassanov, D.C. Johns, U.C. Hoppe, K+ channel regulator KCR1 suppresses heart rhythm by modulating thepacemaker current If, PLoS ONE 3 (2008) e1511.

[19] E. Zagha, A. Ozaita, S.Y. Chang, M.S. Nadal, U. Lin, M.J. Saganich, T. McCormack, K.O.Akinsanya, S.Y. Qi, B. Rudy, DPP10 modulates Kv4-mediated A-type potassiumchannels, J. Biol. Chem. 280 (2005) 18853–18861.

[20] M. Eghbali, R. Olcese, M.M. Zarei, L. Toro, E. Stefani, External pore collapse as aninactivation mechanism for Kv4.3 K+ channels, J. Membr. Biol. 188 (2002) 73–86.

[21] M.A. Davare, V. Avdonin, D.D. Hall, E.M. Peden, A. Burette, R.J. Weinberg, M.C.Horne, T. Hoshi, J.W. Hell, A beta2 adrenergic receptor signaling complexassembled with the Ca2+ channel Cav1.2, Science 293 (2001) 98–101.

[22] S.O. Marx, S. Reiken, Y. Hisamatsu, T. Jayaraman, D. Burkhoff, N. Rosemblit, A.R.Marks, PKA phosphorylation dissociates FKBP12.6 from the calcium releasechannel (ryanodine receptor): defective regulation in failing hearts, Cell 101(2000) 365–376.

[23] Z. Songyang, L.C. Cantley, Recognition and specificity in protein tyrosine kinase-mediated signalling, Trends Biochem. Sci. 20 (1995) 470–475.

[24] R. Ren, B.J. Mayer, P. Cicchetti, D. Baltimore, Identification of a ten-amino acidproline-rich SH3 binding site, Science 259 (1993) 1157–1161.

[25] F.S. Cayabyab, L.C. Schlichter, Regulation of an ERG K+ current by Src tyrosinekinase, J. Biol. Chem. 277 (2002) 13673–13681.

[26] C. Hisatsune, Y. Kuroda, K. Nakamura, T. Inoue, T. Nakamura, T. Michikawa, A.Mizutani, K. Mikoshiba, Regulation of TRPC6 channel activity by tyrosinephosphorylation, J. Biol. Chem. 279 (2004) 18887–18894.

[27] X.Y. Huang, A.D. Morielli, E.G. Peralta, Tyrosine kinase-dependent suppression of apotassium channel by the G protein-coupled m1 muscarinic acetylcholinereceptor, Cell 75 (1993) 1145–1156.

[28] Y.T. Wang, M.W. Salter, Regulation of NMDA receptors by tyrosine kinases andphosphatases, Nature 369 (1994) 233–235.

[29] J.H. Hanke, J.P. Gardner, R.L. Dow, P.S. Changelian,W.H. Brissette, E.J. Weringer, B.A.Pollok, P.A. Connelly, Discovery of a novel, potent, and Src family-selective tyrosinekinase inhibitor. Study of Lck- and FynT-dependent T cell activation, J. Biol. Chem.271 (1996) 695–701.

[30] S.M. Thomas, J.S. Brugge, Cellular functions regulated by Src family kinases, Annu.Rev. Cell Dev. Biol. 13 (1997) 513–609.

[31] M.C. Chen, T.E. Solomon, S.E. Perez, R. Kui, E. Rozengurt, A.H. Soll, Secretinregulates paracellular permeability in canine gastric monolayers by a Src kinase-dependent pathway, Am. J. Physiol Gastrointest. Liver Physiol 283 (2002)G893–G899.

[32] B.I. Posner, R. Faure, J.W. Burgess, A.P. Bevan, D. Lachance, G. Zhang-Sun, I.G.Fantus, J.B. Ng, D.A. Hall, B.S. Lum, A. Shaver, Peroxovanadium compounds. A newclass of potent phosphotyrosine phosphatase inhibitors which are insulinmimetics, J. Biol. Chem. 269 (1994) 4596–4604.

[33] S. Giovannardi, G. Forlani, M. Balestrini, E. Bossi, R. Tonini, E. Sturani, A. Peres, R.Zippel, Modulation of the inward rectifier potassium channel IRK1 by the Rassignaling pathway, J. Biol. Chem. 277 (2002) 12158–12163.

[34] G. Levin, T. Keren, T. Peretz, D. Chikvashvili, W.B. Thornhill, I. Lotan, Regulation ofRCK1 currents with a cAMP analog via enhanced protein synthesis and directchannel phosphorylation, J. Biol. Chem. 270 (1995) 14611–14618.

[35] E. Nesti, B. Everill, A.D. Morielli, Endocytosis as a mechanism for tyrosine kinase-dependent suppression of a voltage-gated potassium channel, Mol. Biol. Cell 15(2004) 4073–4088.

[36] T.C. Holmes, D.A. Fadool, I.B. Levitan, Tyrosine phosphorylation of the Kv1.3potassium channel, J. Neurosci. 16 (1996) 1581–1590.

[37] A. Sobko, A. Peretz, B. Attali, Constitutive activation of delayed-rectifier potassiumchannels by a src family tyrosine kinase in Schwann cells, EMBO J. 17 (1998)4723–4734.

[38] H. Liu, T. Nakazawa, T. Tezuka, T. Yamamoto, Physical and functional interaction ofFyn tyrosine kinase with a brain-enriched Rho GTPase-activating protein TCGAP, J.Biol. Chem. 281 (2006) 23611–23619.

[39] W. Yi, T.H. Lee, J.D. Tompkins, F. Zhu, X. Wu, C. Her, Physical and functionalinteraction between hMSH5 and c-Abl, Cancer Res. 66 (2006) 151–158.

[40] M. Eghbali, R. Deva, A. Alioua, T.Y. Minosyan, H. Ruan, Y. Wang, L. Toro, E. Stefani,Molecular and functional signature of heart hypertrophy during pregnancy, Circ.Res. 96 (2005) 1208–1216.

[41] S. Kaab, J. Dixon, J. Duc, D. Ashen, M. Nabauer, D.J. Beuckelmann, G. Steinbeck, D.McKinnon, G.F. Tomaselli, Molecular basis of transient outward potassium currentdownregulation in human heart failure: a decrease in Kv4.3 mRNA correlates witha reduction in current density, Circulation 98 (1998) 1383–1393.

[42] D. Lebeche, R. Kaprielian, F. del Monte, G. Tomaselli, J.K. Gwathmey, A. Schwartz, R.J. Hajjar, In vivo cardiac gene transfer of Kv4.3 abrogates the hypertrophicresponse in rats after aortic stenosis, Circulation 110 (2004) 3435–3443.

[43] Y. Takeishi, Q. Huang, J. Abe, M. Glassman, W. Che, J.D. Lee, H. Kawakatsu, E.G.Lawrence, B.D. Hoit, B.C. Berk, R.A. Walsh, Src and multiple MAP kinase activationin cardiac hypertrophy and congestive heart failure under chronic pressure-overload: comparisonwith acutemechanical stretch, J. Mol. Cell Cardiol. 33 (2001)1637–1648.

[44] M. Eghbali, Y. Wang, L. Toro, E. Stefani, Heart hypertrophy during pregnancy: abetter functioning heart? Trends Cardiovasc. Med. 16 (2006) 285–291.