Tyrosine Phosphorylation of Kv1.2 Modulates Its Interaction with the Actin-binding Protein Cortactin
2002; Elsevier BV; Volume: 277; Issue: 41 Linguagem: Inglês
10.1074/jbc.m205005200
ISSN1083-351X
AutoresDavid G. Hattan, Edmund Nesti, Teresa G. Cachero, Anthony D. Morielli,
Tópico(s)Neuroscience and Neuropharmacology Research
ResumoTyrosine phosphorylation evokes functional changes in a variety of ion channels. Modulation of the actin cytoskeleton also affects the function of some channels. Little is known about how these avenues of ion channel regulation may interact. We report that the potassium channel Kv1.2 associates with the actin-binding protein cortactin and that the binding is modulated by tyrosine phosphorylation. Immunocytochemical and biochemical analyses show that Kv1.2 and cortactin co-localize to the cortical actin cytoskeleton at the leading edges of the cell. Binding assays using purified recombinant proteins reveal a 19-amino acid span within the carboxyl terminus of Kv1.2 that is necessary for direct cortactin binding. Phosphorylation of specific tyrosines within the C terminus of Kv1.2 attenuates that binding. In HEK293 cells, activation of the M1 muscarinic acetylcholine receptor evokes tyrosine phosphorylation-dependent suppression of Kv1.2 ionic current. We show that M1 receptor activation also reduces the interaction of cortactin with Kv1.2 and that mutant Kv1.2 channels deficient for cortactin binding exhibit strongly attenuated ionic current. These results demonstrate a dynamic, phosphorylation-dependent interaction between Kv1.2 and the actin cytoskeleton-binding protein cortactin and suggest a role for that interaction in the regulation of Kv1.2 ionic current. Tyrosine phosphorylation evokes functional changes in a variety of ion channels. Modulation of the actin cytoskeleton also affects the function of some channels. Little is known about how these avenues of ion channel regulation may interact. We report that the potassium channel Kv1.2 associates with the actin-binding protein cortactin and that the binding is modulated by tyrosine phosphorylation. Immunocytochemical and biochemical analyses show that Kv1.2 and cortactin co-localize to the cortical actin cytoskeleton at the leading edges of the cell. Binding assays using purified recombinant proteins reveal a 19-amino acid span within the carboxyl terminus of Kv1.2 that is necessary for direct cortactin binding. Phosphorylation of specific tyrosines within the C terminus of Kv1.2 attenuates that binding. In HEK293 cells, activation of the M1 muscarinic acetylcholine receptor evokes tyrosine phosphorylation-dependent suppression of Kv1.2 ionic current. We show that M1 receptor activation also reduces the interaction of cortactin with Kv1.2 and that mutant Kv1.2 channels deficient for cortactin binding exhibit strongly attenuated ionic current. These results demonstrate a dynamic, phosphorylation-dependent interaction between Kv1.2 and the actin cytoskeleton-binding protein cortactin and suggest a role for that interaction in the regulation of Kv1.2 ionic current. muscarinic acetylcholine receptor HEK293 cells stably expressing with M1 muscarinic acetylcholine receptors HEK293 cells stably expressing both M1 muscarinic acetylcholine receptors and Kv1.2 potassium channels high speed pellet Src homology 2 and 3, respectively amino acid(s) glutathioneS-transferase Ion channels regulate a wide range of cellular processes, including development (1Moody W.J. J. Neurobiol. 1998; 37: 97-109Crossref PubMed Scopus (56) Google Scholar, 2Ribera A.B. Spitzer N.C. J. Neurobiol. 1998; 37: 190-197Crossref PubMed Scopus (3) Google Scholar), and neuronal plasticity (3Alkon D.L. Nelson T.J. Zhao W. Cavallaro S. Trends Neurosci. 1998; 21: 529-537Abstract Full Text Full Text PDF PubMed Scopus (133) Google Scholar, 4Stevens C.F. Sullivan J. Curr. Biol. 1998; 26: 151-153Abstract Full Text Full Text PDF Google Scholar). Accordingly, the activity of nearly all ion channels is under some form of post-translational control, most commonly direct phosphorylation of the ion channel protein (5Chien A.J. Hosey M.M. J. Bioenerg. Biomembr. 1998; 30: 377-386Crossref PubMed Scopus (34) Google Scholar, 6Ismailov I.I. Benos D.J. Kidney Int. 1995; 48: 1167-1179Abstract Full Text PDF PubMed Scopus (64) Google Scholar, 7Levitan I.B. Annu. Rev. Physiol. 1994; 56: 193-212Crossref PubMed Scopus (484) Google Scholar). The earliest and most widely reported examples involve serine/threonine phosphorylation; however, it is also now recognized that tyrosine phosphorylation is an important means of ion channel regulation (8Siegelbaum S.A. Curr. Biol. 1994; 4: 242-245Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar). The nicotinic acetylcholine receptor was the first ligand-gated ion channel found to be regulated by tyrosine phosphorylation (9Hopfield J.F. Tank D.W. Greengard P. Huganir R.L. Nature. 1988; 336: 677-680Crossref PubMed Scopus (201) Google Scholar). The shaker family potassium channel Kv1.2, which is expressed abundantly in cardiac muscle (10Paulmichi M. Nasmith P. Helmiss R. Reed K. Boyle W.A. Nerbonne J.M. Peralta E.G. Clapham D.E. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 7892-7895Crossref PubMed Scopus (64) Google Scholar) and neurons (11Sheng M. Tsaur M.L. Jan Y.N. Jan L.Y. J. Neurosci. 1994; 14: 2408-2417Crossref PubMed Google Scholar, 12Tsaur M.L. Sheng M. Lowenstein D.H. Jan Y.N. Jan L.Y. Neuron. 1992; 8: 1055-1067Abstract Full Text PDF PubMed Scopus (165) Google Scholar, 13Wang H. Kunkel D.D. Schwartzkroin P.A. Tempel B.L. J. Neurosci. 1994; 14: 4588-4599Crossref PubMed Google Scholar), was the first voltage-gated channel found to be so regulated. Kv1.2 α-subunit protein becomes tyrosine-phosphorylated upon the activation of M1 muscarinic acetylcholine receptors (mAChRs),1 leading to a profound suppression of Kv1.2 ionic current (14Huang X.Y. Morielli A.D. Peralta E.G. Cell. 1993; 75: 1145-1156Abstract Full Text PDF PubMed Scopus (243) Google Scholar). Part of the mechanism for such suppression involves phosphorylation of the N-terminal tyrosine Tyr132. However, phosphorylation of additional tyrosines within Kv1.2 is also likely to be required, since Y132F mutant channels are only partially resistant to tyrosine phosphorylation and ionic current suppression. Little is known about the signaling proteins involved in tyrosine phosphorylation-dependent suppression of Kv1.2 ionic current; however, binding of the activated form of the guanine nucleotide-binding protein RhoA to Kv1.2 is required (15Cachero T.G. Morielli A.D. Peralta E.G. Cell. 1998; 93: 1077-1085Abstract Full Text Full Text PDF PubMed Scopus (84) Google Scholar). RhoA is an important regulator of actin dynamics (16Tapon N. Hall A. Curr. Biol. 1997; 9: 86-92Crossref Scopus (698) Google Scholar, 17O'Connell C.B. Wheatley S.P. Ahmed S. Wang Y. J. Cell Biol. 1999; 144: 305-313Crossref PubMed Scopus (115) Google Scholar, 18Schafer D.A. Welch M.D. Machesky L.M. Bridgman P.C. Meyer S.M. Cooper J.A. J. Cell Biol. 1998; 143: 1919-1930Crossref PubMed Scopus (152) Google Scholar), and the physical association between RhoA and Kv1.2 suggests a link between the cytoskeleton and tyrosine phosphorylation-dependent regulation of Kv1.2. This is consistent with a growing body of data indicating that the cytoskeleton has an important role in ion channel function and regulation. In addition to affecting channel localization (19Burke N.A. Takioto K., Li, D. Han W. Watkins S.C. Levitan E.S. J. Gen. Physiol. 1999; 113: 71-80Crossref PubMed Scopus (59) Google Scholar, 20Sheng M. Wyszynski M. Bioessays. 1997; 19: 847-853Crossref PubMed Scopus (133) Google Scholar), proteins that compose or associate with the cytoskeleton can influence ion channel activity per se. That conclusion is based primarily on reports showing that pharmacological disruption of the actin cytoskeleton profoundly affects the physiology of a number of ion channels (21Rosenmund C. Westbrook G.L. Neuron. 1993; 10: 805-814Abstract Full Text PDF PubMed Scopus (458) Google Scholar, 22Undrovinas A.I. Shander G.S. Makielski J.C. Am. J. Physiol. 1995; 269: H203-H214PubMed Google Scholar, 23Cantiello H.F. Exp. Physiol. 1996; 81: 505-514Crossref PubMed Scopus (54) Google Scholar, 24Prat A.G. Xiao Y.F. Ausiello D.A. Cantiello H.F. Am. J. Physiol. 1995; 6: C1152-C1161Google Scholar). Such pharmacological data still comprise the bulk of the evidence linking the cytoskeleton to channel physiology, yet several recent reports identify specific cytoskeleton-associated proteins that affect ion channel function. They include ankyrin binding to the ryanodine receptors to influence its ability to release calcium (25Bourguignon L.Y. Chu A. Jin H. Brandt N.R. J. Biol. Chem. 1995; 270: 17919-17922Google Scholar), SAP90 binding to kainate receptors to affect desensitization (26Garcia E.P. Mehta S. Blair L.A. Wells D.G. Shang J. Fukushima T. Fallon J.R. Garner C.C. Marshall J. Neuron. 1998; 21: 727-739Abstract Full Text Full Text PDF PubMed Scopus (228) Google Scholar), α-actinin coupling with Kv1.5 to affect current density (27Maruoka N.D. Steele D.F., Au, B.P. Dan P. Zhang X. Moore E.D. Fedida D. FEBS Lett. 2000; 473: 188-194Crossref PubMed Scopus (90) Google Scholar), and filamin interacting with Kv4.1 to affect current density (28Petrecca K. Miller D.M. Shrier A. J. Neurosci. 2000; 20: 8736-8744Crossref PubMed Google Scholar). Together, these findings show that the cytoskeleton has an important but still poorly understood role in regulating the activity of a variety of ion channels. Here we report that Kv1.2 associates with the F-actin-binding protein cortactin in a tyrosine phosphorylation-dependent manner and that such binding may have a role in phosphorylation-dependent channel regulation. Cortactin is a substrate for tyrosine phosphorylation within the cortical cytoskeleton, and its modular structure suggests a multifaceted role within the cell. An N-terminal region binds to the Arp2/3 complex, thereby providing a site for actin filament nucleation (29Uruno T. Liu J. Zhang P. Fan Y. Egile C., Li, R. Mueller S.C. Zhan X. Nat. Cell Biol. 2001; 3: 259-266Crossref PubMed Scopus (464) Google Scholar, 30Weed S.A. Karginov A.V. Schafer D.A. Weaver A.M. Kinley A.W. Cooper J.A. Parsons J.T. J. Cell Biol. 2000; 151: 29-40Crossref PubMed Scopus (338) Google Scholar). Central cortactin repeat regions bind to and cross-link actin filaments (31Wu H. Parsons J.T. J. Cell Biol. 1993; 120: 1417-1426Crossref PubMed Scopus (452) Google Scholar), an activity that is regulated by cortactin tyrosine phosphorylation (32Huang C., Ni, Y. Wang T. Gao Y. Haudenschild C.C. Zhan X. J. Biol. Chem. 1997; 272: 13911-13915Abstract Full Text Full Text PDF PubMed Scopus (221) Google Scholar). The C terminus of cortactin contains tyrosines that can serve as SH2 domain docking site for Src kinase (33Okamura H. Resh M.D. J. Biol. Chem. 1995; 270: 26613-26618Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar) and an SH3 domain that binds to a number of proteins, including Shank (34Naisbitt S. Kim E., Tu, J.C. Xiao B. Sala C. Valtschanoff J. Weinberg R.J. Worley P.F. Sheng M. Neuron. 1999; 23: 569-582Abstract Full Text Full Text PDF PubMed Scopus (810) Google Scholar), ZO-1 (35Katsube T. Takahisa M. Ueda R. Hashimoto N. Kobayashi M. Togashi S. J. Biol. Chem. 1998; 273: 29672-29677Abstract Full Text Full Text PDF PubMed Scopus (127) Google Scholar), and dynamin (36McNiven M.A. Kim L. Krueger E.W. Orth J.D. Cao H. Wong T.W. J. Cell Biol. 2000; 151: 187-198Crossref PubMed Scopus (343) Google Scholar). Collectively, these features make cortactin well suited to couple tyrosine kinase signaling between membrane proteins and the cytoskeleton (37Weed S.A. Parsons J.T. Oncogene. 2001; 20: 6418-6434Crossref PubMed Scopus (363) Google Scholar). Despite these findings, the role of cortactin remains obscure in many cases. We have found that cortactin co-localizes with Kv1.2 within F-actin-rich leading edge structures. Further, we show that cortactin binds directly to the carboxyl terminus of Kv1.2 and that the interaction is reduced by phosphorylation of specific tyrosines within the Kv1.2 C terminus. The same mutations within Kv1.2 that reduce cortactin binding also result in channels that generate significantly less ionic current than do wild type channels. These data reveal a tyrosine kinase-dependent interaction between Kv1.2 and the actin cytoskeleton via cortactin and suggest a role for cortactin in Kv1.2 function. Protein grade Triton X-100 was purchased fromCalbiochem; glutathione, and protein G-conjugated Sepharose and glutathione-Sepharose beads were purchased from Amersham Biosciences; and all other reagents were purchased from Sigma or as indicated. GST-C was generated using a pGEX-2T vector into which the C terminus of Kv1.2 (aa 413–499) had been subcloned between the BamH1 and EcoR1 multicloning sites. GST-wild type cortactin was a gift from C. Huang and X. Zhan (American Red Cross, Rockville, MD). GST-SH3 encodes the SH3-domain of cortactin (aa 490–546) subcloned into pGEX-2T between the BamHI and EcoRI sites and was a gift from Brian Kay (University of North Carolina, Chapel Hill, NC). Truncation mutants were generated by introducing stop codons using the QuikChange mutagenesis kit (Stratagene); GST-cortactn-ΔSH3 was generated by introducing a stop codon at aa 497 within the GST-wild type cortactin construct. Approximately 5 μg of purified GST fusion protein bound to Sepharose beads were incubated for 1 h at 4 °C with 293m cell lysates or with 100 nm soluble recombinant cortactin (derived by thrombin cleavage of cortactin from GST) in Triton lysis buffer. Bound proteins were eluted from the beads and then separated by SDS-PAGE. GST fusion proteins were detected with anti-GST antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Stable 293 cell lines were created as described (15Cachero T.G. Morielli A.D. Peralta E.G. Cell. 1998; 93: 1077-1085Abstract Full Text Full Text PDF PubMed Scopus (84) Google Scholar). All stable cell lines expressing Kv1.2wt and mutant channels also stably expressed Kvβ2 subunits. For all experiments, the 293 cells were plated to low confluence (∼3.5 × 106 cells/100-mm plate) 12 h before use. Ionic current generated by Kv1.2 protein expressed in Xenopus oocytes was measured as described (15Cachero T.G. Morielli A.D. Peralta E.G. Cell. 1998; 93: 1077-1085Abstract Full Text Full Text PDF PubMed Scopus (84) Google Scholar). Briefly, Xenopus oocytes were injected with 18 ng of cRNA encoding Kv1.2wt, Kv1.2–458stop, or Kv1.2-Y417F. Recordings were made 24–48 h after injection using standard two-electrode voltage clamp at 25 °C in OR-2 saline (82.5 mm NaCl, 2 mmKCl, 1 mm MgCl2, 1.8 mmCaCl2, 5 mm HEPES, pH 7.5). 293 cells were treated with a control saline or carbachol (10 μm) for 20 min at 37 °C. The cells were washed with ice-cold phosphate-buffered saline and the lysed in 1 ml of Triton lysis buffer (100 mm KCl, 2 mmATP, 3.94 μm MgCl2 (250 nm free magnesium), 1 mm Na4BAPTA, 1 mmsodium orthovanadate, 1 mm dithiothreitol, 5 mmHEPES, 0.5% Triton X-100, protease inhibitors (aprotinin, leupeptin, phenylmethylsulfonyl fluoride), pH 7.4, at 4 °C). The lysate was transferred to a 1.5-ml microcentrifuge tube and centrifuged at 16,000 × g for 4 min. The supernatant was transferred to an ultracentrifuge tube and centrifuged at 100,000 ×g for 1 h. The soluble fraction supernatant was transferred to a fresh microcentrifuge tube. The remaining high speed pellet (HSP) was washed and resuspended in a modified radioimmune precipitation assay buffer (50 mm Tris, 150 mmNaCl, 1 mm EDTA, 1% Nonidet P-40, 0.25% deoxycholate, 0.1% SDS, 10% glycerol, 1 mm dithiothreitol, 1 mm NaF, 1 mm sodium orthovanadate, and protease inhibitors (aprotinin, leupeptin, phenylmethylsulfonyl fluoride), pH 7.4, at 4 °C). Three distinct anti-Kv1.2 antibodies were used: 1) αKv-p, an affinity-purified rabbit anti-Kv1.2 polyclonal antibody against a peptide region (aa 468–481) within the carboxyl terminus of Kv1.2; 2) αKv-e, a rabbit polyclonal antibody directed against a peptide encoding a portion of the first extracellular loop (aa 192–208) of Kv1.2 (BIOSOURCE International); 3) αKv-i, a mouse monoclonal antibody raised against the nearly complete C terminus (aa 428–499) of Kv1.2 (Upstate Biotechnology, Inc., Lake Placid, NY). Monoclonal anti-cortactin antibodies were from Upstate Biotechnology, and rabbit polyclonal anti-cortactin antibodies were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). HEK293 cells were plated onto noncoated glass coverslips and grown for 12 h. To detect surface Kv1.2, αKv-e (100 μg/ml) was added to the growth medium overlaying living cells for 20 min at 37 °C and then washed 3 × 5 min with ice-cold phosphate-buffered saline. All other antibodies were applied after fixation. Primary antibodies are as indicated. Secondary antibodies were AlexaFlour-488-, AlexaFlour-658-, or AlexaFlour-647-conjugated goat anti-mouse or anti-rabbit. Imaging was with a Bio-Rad MRC 1024ES confocal scanning laser microscope system or with an Olympus BX50 epifluorescence microscope as indicated. Images were collected and stored digitally. To determine whether Kv1.2 and cortactin can co-associate, we performed immunoprecipitation studies with native proteins expressed in rat brain. Western blot with an anti-cortactin monoclonal antibody revealed that cortactin co-immunoprecipitated with a polyclonal antibody directed against an epitope within the C terminus of Kv1.2 (αKv-p) or with polyclonal antibody directed against the first extracellular loop of Kv1.2 (αKv-e) but not with protein G beads alone (Fig.1 a). Some cortactin breakdown is typically evident in our co-immunoprecipitation of Kv1.2 from brain tissue but not from HEK293 cells. It is possible that a protease present in brain remains part of the Kv1.2 immunocomplex, where it acts on cortactin. Indeed, cortactin proteolysis may be a means of cortactin regulation in platelets (38Huang C. Tandon N.N. Greco N.J., Ni, Y. Wang T. Zhan X. J. Biol. Chem. 1997; 272: 19248-19252Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar). Since cortactin co-immunoprecipitated with two antibodies directed against different epitopes within Kv1.2, it is unlikely that the observed positive signals occurred through nonspecific antibody cross-reactivity. To examine the interactions between Kv1.2 and cortactin in more detail, we chose HEK293 cells transfected with and under stable selection for a plasmid encoding the M1 muscarinic acetylcholine receptor, and the potassium channel Kv1.2-α and Kv1.2β2 subunits (293mK cells) (15Cachero T.G. Morielli A.D. Peralta E.G. Cell. 1998; 93: 1077-1085Abstract Full Text Full Text PDF PubMed Scopus (84) Google Scholar,39Tsai W. Morielli A.D. Peralta E.G. EMBO J. 1997; 15: 4597-4605Crossref Scopus (188) Google Scholar). This system has the added advantage of allowing the use of a HEK293 cell line expressing M1 muscarinic receptors but not Kv1.2-α or Kv1.2β2 (293m cells) as a rigorous control for nonspecific antibody binding. Consistent with our findings with rat brain, we found that cortactin co-immunoprecipitated with Kv1.2 in the 293 cell system. The cortactin signal was specific to the presence of Kv1.2 and does not represent antibody cross-reactivity because it did not appear in the 293m parental cell line. Data from three independent experiments are summarized in Fig. 1 c. Actin dynamics within the cortical cytoskeleton are regulated in part by the actin-binding protein cortactin. The central region of cortactin contains actin-binding domains that serve to cross-link actin filaments (31Wu H. Parsons J.T. J. Cell Biol. 1993; 120: 1417-1426Crossref PubMed Scopus (452) Google Scholar), whereas an N-terminal domain binds to the Arp2/3 complex, which in turn mediates nucleation of actin filaments (30Weed S.A. Karginov A.V. Schafer D.A. Weaver A.M. Kinley A.W. Cooper J.A. Parsons J.T. J. Cell Biol. 2000; 151: 29-40Crossref PubMed Scopus (338) Google Scholar). Functionality of cortactin within the cortical cytoskeleton is expanded by the presence of a variety of proteins that bind to cortactin via SH2 and SH3 interactions. Thus, cortactin can dynamically influence physiology associated with the cortical cytoskeleton either directly through cross-linking or indirectly through its binding partners. In particular, cortactin is thought to serve as a nexus for tyrosine kinase signaling within the cortical cytoskeleton (37Weed S.A. Parsons J.T. Oncogene. 2001; 20: 6418-6434Crossref PubMed Scopus (363) Google Scholar). Since Kv1.2 is regulated by tyrosine phosphorylation, and since several other ion channels have been shown to associate with components of the cortical cytoskeleton (40Molday L.L. Cook N.J. Kaupp U.B. Molday R.S. J. Biol. Chem. 1990; 265: 18690-18695Abstract Full Text PDF PubMed Google Scholar, 41Smith P.R. Saccomani G. Joe E.H. Angelides K.J. Benos D.J. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 6971-6975Crossref PubMed Scopus (148) Google Scholar, 42Srinivasan Y. Elmer L. Davis J. Bennett V. Angelides K. Nature. 1988; 333: 177-180Crossref PubMed Scopus (307) Google Scholar, 43Wechsler A. Teichberg V.I. EMBO J. 1998; 17: 3931-3939Crossref PubMed Scopus (168) Google Scholar), we hypothesized that the interaction between Kv1.2 and cortactin occurs within the cortical cytoskeleton. To test this idea, we used differential centrifugation to obtain a cellular fraction enriched in the cortical cytoskeleton. 293mK cells were lysed and subjected to sequential centrifugation in Triton lysis buffer to produce a Triton soluble fraction and a cortical cytoskeleton enriched fraction (HSP) (44Sheetz M.P. The Bilayer and Shell Model of Erythrocyte Membrane. A. R. Liss, New York1980Google Scholar). Western analysis revealed that both Kv1.2 and cortactin are present in the HSP and soluble fractions (Fig.2). Since equal volumes of HSP or soluble fraction lysate were loaded into each lane, the stronger signals evident in the soluble fraction indicate a higher molar concentration of both Kv1.2 and cortactin relative to HSP lysate. Nevertheless, the ability of Kv1.2 to immunoprecipitate cortactin was limited to the HSP fraction (Fig. 2). Identical results were obtained in three independent experiments. To confirm our biochemical findings that Kv1.2 and cortactin associate with each other and that the association occurs within the cortical cytoskeleton, we used confocal and wide field immunofluorescence microscopy to detect co-localization within fixed 293mK cells. Since they conduct potassium ions across the plasma membrane, Kv1.2-α subunits that compose functional Kv1.2 channels reside at the surface of the cell rather than within intracellular compartments. We therefore first sought to differentially detect surface versusinternal Kv1.2 protein and to then ask whether cortactin co-localizes preferentially with either of those subpopulations. To accomplish this, we used rabbit polyclonal antibody directed against an extracellular portion of Kv1.2 (αKv-e), in conjunction with a mouse monoclonal antibody directed against the intracellular C terminus of Kv1.2 (αKv-i). Confocal microscopy revealed that the αKv-e (Fig.3 b) localizes to the surface of the cell, primarily within leading edge structures, whereas the αKv-i signal localizes primarily within and throughout the central region of the cell (Fig. 3 a). Fig. 3 c is a merge of Fig. 3 a (green) and Fig. 3 b(red), showing that a distinct population of surface Kv1.2 can be readily distinguished from the intracellular population. Presumably, the surface population represents channels capable of conducting ionic current, whereas the internal population may represent immature channel protein (45Li D. Takimoto K. Levitan E.S. J. Biol. Chem. 2000; 275: 11597-11602Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar). An indication of the possible physiological role of the Kv1.2-cortactin interaction may come from knowing whether the interaction occurs within the surface or the internal population of Kv1.2. To determine which Kv1.2 population cortactin co-localizes with, we used immunofluorescence labeling of either surface or total Kv1.2 and simultaneous labeling of cortactin in fixed 293mK cells. Fig.4 a shows a confocal micrograph of total Kv1.2 labeled with αKv-e antibody. The same cells were labeled with a monoclonal antibody directed against cortactin (Fig.4 b). Fig. 4 c shows a merge of Fig. 4,a (red) and b (green). Little if any co-localization of total Kv1.2 and cortactin can be observed. In contrast, confocal images of surface Kv1.2 (Fig. 4,d and e (red)) and cortactin (Fig. 4,e and i (green)) clearly indicate that the proteins co-localize. That this co-localization occurs primarily in the leading edges of the cell was confirmed with lower magnification wide field fluorescent images of cells surface-labeled with αKv-e (Fig. 4 g) and monoclonal anti-cortactin (Fig.4 h). Co-localization of surface Kv1.2 and cortactin is readily detectable in the leading edge structures of nearly all of the cells around the circumference of the cell cluster. Cortactin is enriched within lamellipodia (30Weed S.A. Karginov A.V. Schafer D.A. Weaver A.M. Kinley A.W. Cooper J.A. Parsons J.T. J. Cell Biol. 2000; 151: 29-40Crossref PubMed Scopus (338) Google Scholar, 31Wu H. Parsons J.T. J. Cell Biol. 1993; 120: 1417-1426Crossref PubMed Scopus (452) Google Scholar, 46Kaksonen M. Peng H.B. Rauvala H. J. Cell Sci. 2000; 113: 4421-4426Crossref PubMed Google Scholar, 47Weed S.A., Du, Y. Parsons J.T. J. Cell Sci. 1998; 111: 2433-2443PubMed Google Scholar), where it contributes to the highly dynamic reorganization of the membrane-associated actin cytoskeleton. Our finding that Kv1.2 and cortactin co-localize within that region is consistent with our biochemical data showing co-association in a cortical cytoskeleton enriched fraction. To confirm that the leading edge structures in which we observe Kv1.2 and cortactin co-localization are enriched with F-actin, we triple-labeled 293mK cells with αKv-e primary antibody (Fig. 4, i (blue) and l), α-cortactin primary antibody (Fig. 4, i (red) and j), and fluorescence-conjugated phalloidin to label F-actin (Fig. 4, i (green) and k). F-actin staining in the leading edges was pronounced, and co-localization of all three proteins is evident (Fig. 4, i(white and arrows)). Taken together, the data shown in Figs. Figure 1, Figure 2, Figure 3, Figure 4 indicate that Kv1.2 and cortactin co-associate and that they do so within the membrane-associated actin cytoskeleton. Having determined that cortactin co-associates with Kv1.2 in immunoprecipitation and immunofluorescence assays, we next determined whether or not the co-association is direct. To map the domains of the Kv1.2/cortactin interaction, we performed pull-down assays using purified recombinant GST fusion proteins of regions within Kv1.2 and purified recombinant cortactin. Soluble recombinant cortactin and immobilized GST fusion proteins of the amino (G-N) or the carboxyl (G-C) terminal regions of Kv1.2 were generated as described under "Experimental Procedures." Western analysis using a monoclonal anti-cortactin antibody indicates that purified recombinant cortactin binds directly to G-N or G-C but not to GST alone (Fig. 5 a,top). The bottom half of the same gel was stained for total protein to confirm equal loading of G-N, G-CP, and GST (Fig.5 a, middle). Data from three independent experiments are quantified in Fig. 5 a (bottom). Cortactin binds to both the N- and C-terminal regions of Kv1.2, and binding to either may be important to Kv1.2 physiology. However, cortactin binding to the C terminus of Kv1.2 is significantly more robust than to the N terminus, and we therefore chose to proceed by examining only the interaction between Kv1.2 C terminus and cortactin. To gain insight into the regions within the Kv1.2 C terminus required for cortactin binding, we measured the effect of serial truncations of GST-C on cortactin binding. Fig. 5 b indicates binding of 100 nm recombinant cortactin to truncated forms of GST-C normalized to its binding to full-length GST-C. GST-489stop, GST-474stop, GST-470stop, and GST-464stop all bind to cortactin to the same degree as does full-length GST-C. In contrast, GST-462stop bound significantly more weakly. This reduced but detectable binding was maintained with GST-458stop and GST-448stop. Introduction of a stop codon at sites 445 or beyond, however, completely eliminated cortactin binding. Thus, the 19-amino acid span between Lys446 and Glu463, with a small critical zone at each end, is required for cortactin binding to the C terminus of Kv1.2 (Fig.5 c). It is important to note that the cortactin binding span may not be the site of physical interaction between Kv1.2 and cortactin. Instead, it represents a region within the C terminus that is required for such binding. The actual cortactin binding region may lie within that region or within in a portion of the C terminus between aa 413 and 445. Interestingly, a potential SH3 binding ligand region exists in the Kv1.2 C terminus between aa 436 and 442 (Fig.6 a). The Kv1.2 sequence, CPKIPSSP, is not identical to a consensus cortactin SH3 ligand sequence, +PPΨPXKP (48Sparks A.B. Rider J.E. Hoffman N.G. Fowlkes D.M. Quillam L.A. Kay B.K. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 1540-1544Crossref PubMed Scopus (331) Google Scholar); however, considerable similarity exists between them. Since an SH3 interaction is important for tyrosine phosphorylation-dependent suppression of a related channel, Kv1.5 (49Holmes T.C. Fadool D.A. Ren R. Levitan I.B. Science. 1996; 274: 2089-2094Crossref PubMed Scopus (231) Google Scholar), we asked whether cortactin binds to Kv1.2 via an SH3 interaction. We first asked whether soluble recombinant Kv1.2 C terminus could bind to full-length GST-cortactin (G-FL in Fig. 6), to a GST fusion protein containing only the cortactin SH3 domain (G-SH3), or to GST immobilized on glutathione-Sepharose beads. Only full-length cortactin and not the isolated cortactin SH3 domain bound to the soluble Kv1.2 C terminus in a pull-d
Referência(s)