* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Caveolin-3 and SAP97 form a scaffolding protein complex that
Survey
Document related concepts
Tissue engineering wikipedia , lookup
Cell membrane wikipedia , lookup
Cell culture wikipedia , lookup
Mechanosensitive channels wikipedia , lookup
Cytokinesis wikipedia , lookup
Cell encapsulation wikipedia , lookup
Extracellular matrix wikipedia , lookup
Cellular differentiation wikipedia , lookup
Endomembrane system wikipedia , lookup
Organ-on-a-chip wikipedia , lookup
Signal transduction wikipedia , lookup
Transcript
Am J Physiol Heart Circ Physiol 287: H681–H690, 2004; 10.1152/ajpheart.00152.2004. Caveolin-3 and SAP97 form a scaffolding protein complex that regulates the voltage-gated potassium channel Kv1.5 Eduardo J. Folco, Gong-Xin Liu, and Gideon Koren Bioelectricity Laboratory, Cardiovascular Division, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115 Submitted 31 March 2004; accepted in final form 31 March 2004 lipid rafts; potassium channels; caveolae; heart TARGETING OF PROTEINS to specific membrane regions may have a profound effect on their function. Membrane-associated guanylate kinases (MAGUKs) are PDZ domain-containing proteins that have been proposed to play a role in the targeting of sodium channels, N-methyl-D-aspartate (NMDA) receptors, and voltage-gated potassium channels to discrete plasma membrane domains in neuronal cells (5, 8, 22). MAGUKs also contain other interaction modules, such as the Src homology-3 (SH3) and guanylate kinase (GUK)-like domains. Hence, MAGUKs serve as a scaffold for the assembly of different polypeptides into macromolecular signaling complexes (8). Lipid rafts are membrane microdomains rich in cholesterol and sphingolipids that are enriched in important signaling molecules (19). Caveolae represent a subpopulation of lipid rafts that contain the scaffolding protein caveolin, which can organize the assembly of macromolecular complexes and regulate protein function within the caveolae. Three caveolin genes (Cav-1, Cav-2, and Cav-3) have been identified. Cav-1 and Cav-2 have a wide distribution among different cell types, whereas Cav-3 is expressed exclusively in all types of muscle (21). Mutations in the human Cav-3 gene are associated with muscular dystrophy (16, 17), and inactivation of the Cav-3 gene in mice leads to muscle degeneration (7, 11), cardiac hypertrophy, and progressive cardiomyopathy (27). Address for reprint requests and other correspondence: G. Koren, Bioelectricity Laboratory, Cardiovascular Div., Brigham and Women’s Hospital, 75 Francis St., Boston, MA 02115 (E-mail: [email protected]). http://www.ajpheart.org Very little is known about scaffolding proteins that mediate the organization of “transducisomes” in the heart in general and those that contain potassium channels (channelosomes) in particular. SAP97 is expressed abundantly in the heart, and recent reports indicate that it can interact with voltage-gated and inward rectifier potassium channels (14, 26). We recently reported that Kv1.5 colocalizes with SAP97 in the heart and interacts with it in heterologous expression systems (18). Godreau et al. (10) have confirmed these interactions and reported that Kv1.5 and SAP97 associate in the human atrium. Martens and colleagues (15) recently reported that in heterologous expression systems, certain voltage-gated potassium channels are differentially targeted to distinct subpopulations of lipid rafts: in mouse L cells, Kv2.1 targets to noncaveolar lipid rafts, whereas Kv1.5 targets to caveolae (15). The mechanisms that regulate these targeting events are unknown. The observed association of Kv1.5 with SAP97 (18) and its localization in caveolae (15) led us to hypothesize that SAP97 associates with Cav-3 to form a scaffolding complex that can recruit ion channels and regulate their function. Here, we report that Cav-3 interacts with SAP97 in the heart, in heterologous expression systems, and in vitro. We define the sites of proteinprotein interactions involved in the assembly of this complex and use model cell systems to demonstrate that it can recruit Kv1.5 to form a tripartite complex. Furthermore, we show that the Cav-3/SAP97 complex selectively regulates the expression of currents encoded by a glycosylation-deficient Kv1.5 mutant. MATERIALS AND METHODS Materials. The following antibodies were used: anti-PDZ [previously characterized (18) mouse monoclonal antibody (MAb) that recognizes SAP97] from Upstate Biotechnology, anti-Cav-3 (mouse MAb) from BD Transduction Laboratories, anti-Cav-3 (goat pAb) from Santa Cruz Biotechnology, anti-FLAG epitope (M2, mouse MAb) from Sigma, and anti-c-Myc (mouse MAb) and anti-green fluorescent protein (GFP) [rabbit polyclonal antibody (PAb)] from Clontech. The secondary antibodies were from Zymed Laboratories, the glutatione-Sepharose 4B was from Amersham Pharmacia Biotech, and the protein A/G agarose was from Santa Cruz Biotechnology. DNA cloning and bacterial expression. The construction of rat SAP97, FL-Kv1.5, and FL-Kv1.5-A cDNAs was described previously (18). Myc-tagged Cav-3 and Kv2.1 cDNAs were gifts from Drs. Thomas Michel and Barbara Wible, respectively. The various SAP97 deletion mutants (S1–S5; see schemes in Fig. 2A) were amplified by PCR in-frame with the FLAG epitope coding sequence at the 5⬘-end and subcloned into pCDNA3 (Invitrogen). GFP-⌬SAP (see Fig. 2A) and GFP1.5C (see Fig. 4B) were amplified by PCR and subcloned into 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. 0363-6135/04 $5.00 Copyright © 2004 the American Physiological Society H681 Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 Folco, Eduardo J., Gong-Xin Liu, and Gideon Koren. Caveolin-3 and SAP97 form a scaffolding protein complex that regulates the voltage-gated potassium channel Kv1.5. Am J Physiol Heart Circ Physiol 287: H681–H690, 2004; 10.1152/ajpheart.00152.2004.—The targeting of ion channels to particular membrane microdomains and their organization in macromolecular complexes allow excitable cells to respond efficiently to extracellular signals. In this study, we describe the formation of a complex that contains two scaffolding proteins: caveolin-3 (Cav-3) and a membrane-associated guanylate kinase (MAGUK), SAP97. Complex formation involves the association of Cav-3 with a segment of SAP97 localized between its PDZ2 and PDZ3 domains. In heterologous expression systems, this scaffolding complex can recruit Kv1.5 to form a tripartite complex in which each of the three components interacts with the other two. These interactions regulate the expression of currents encoded by a glycosylation-deficient mutant of Kv1.5. We conclude that the association of Cav-3 with SAP97 may constitute the nucleation site for the assembly of macromolecular complexes containing potassium channels. H682 SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 RESULTS Cav-3 interacts directly with SAP97. Kv1.5 has been shown to interact with SAP97 (18) and to target to caveolae (15) in heterologous expression systems. These findings prompted us AJP-Heart Circ Physiol • VOL to hypothesize that SAP97 and Cav-3 form a scaffolding complex capable of recruiting Kv1.5 and regulating its function. We first carried out coimmunoprecipitation and pulldown experiments to test whether SAP97 and Cav-3 form a complex in the mouse heart. After SAP97 was immunoprecipitated from mouse ventricle lysates with the use of anti-PDZ antibodies, the immunopellets were size separated by SDSPAGE and immunoblotted with anti-Cav-3 antibody. Figure 1A, lane 2, shows that anti-PDZ antibodies could coprecipitate Cav-3. Control immunoprecipitations with IgG often yielded a weak background signal corresponding to Cav-3 polypeptide (Fig. 1A, lane 3). To provide additional evidence that this complex exists in the heart, we made a cDNA construct encoding a GST fusion polypeptide that contains the SAP97binding site of Kv1.5 (GST1.5C1; Fig. 1B). The fusion protein was expressed in E. coli, purified by glutathione-agarose chromatography, and reacted with total mouse heart lysates. Figure 1B shows that GST1.5C1 coprecipitated SAP97 and Cav-3 from whole heart mouse extracts (lane 2), whereas control GST did not precipitate either polypeptide (lane 3). A direct association of GST1.5C1 with Cav-3 seems unlikely because GST1.5C1 was unable to pull down Cav-3 in similar experiments with extracts from Cav-3-transfected COS-7 cells (not shown). Therefore, the Cav-3 that was pulled down from heart extracts by GST1.5C1 likely represents the fraction bound to SAP97. The interaction between SAP97 and Cav-3 was confirmed in cotransfection experiments and in vitro. Figure 1C shows that anti-PDZ antibody specifically coprecipitated Cav-3 (lane 8, bottom), and anti-Cav-3 antibody specifically coprecipitated SAP97 (lane 8, top) only from lysates of cotransfected cells but not from mock-transfected cells (lane 7) or cells transfected with either Cav-3 (lane 5) or SAP97 (lane 6) alone. We also tested for direct interaction in an in vitro translation assay: after T7-directed coexpression of SAP97 and Cav-3 cDNAs in reticulocyte lysate, both polypeptides were coprecipitated with anti-Cav-3 antibody (Fig. 1D, lane 4). Control precipitations showed that SAP97 was not pulled down by anti-Cav-3 antibody in the absence of Cav-3 (lane 5) or when control IgG was used in the precipitation (lane 6). Cav-3-binding site of SAP97 resides in a region located between PDZ2 and PDZ3. We next made a series of SAP97 deletion mutants to map the site of interaction with Cav-3. Initially, we made three FLAG-tagged cDNA constructs: S1SAP97(451–926), which contained the PDZ3, SH3, and GUKlike domains (Fig. 2A); S2-SAP97(201–504), which contained the PDZ1 and PDZ2 domains followed by a proline-rich region and part of PDZ3 (Fig. 2A); and SAP97(1–200), which contained the NH2-terminal multimerization domain (not shown). Each of these constructs was cotransfected with Cav-3 and tested for interaction with it in coimmunoprecipitation assays using anti-FLAG and anti-Cav-3 antibodies. Cav-3 specifically coprecipitated with S2 (see below) but not with S1 (not shown), indicating that the central segment of SAP97 is sufficient for binding to Cav-3. The expression of the NH2-terminal fragment of SAP97 (amino acid residues 1–200) could not be detected in this system. To further define the domain in S2 involved in its interaction with Cav-3, we made a series of deletion mutants in which we kept PDZ1 and PDZ2 but removed the NH2 terminus (S3), the COOH terminus (S4), or both (S5) (Fig. 2A). The mutants were 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 pEGFP-C3 in-frame with the GFP coding sequence. Kv1.5-N1-FL (see Fig. 4B) was amplified by PCR in-frame with the FLAG epitope coding sequence at the 3⬘-end and subcloned into pCDNA3. The SAP97 deletion mutants S3 and S5 were amplified by PCR and subcloned into pRSETB (Invitrogen) in-frame with the 6xHis tag coding sequence. GST1.5C1 (see Fig. 1B) and GST1.5C2 (see Fig. 4B) were amplified by PCR and subcloned into pGEX-4T-1 in-frame with the glutathione S-transferase (GST) coding sequence. Cell lines and transfections. COS-7 and Chinese hamster ovary (CHO) cells were transfected using Fugene 6 (Roche) and harvested 24 h posttransfection. We searched for a suitable expression system for our biochemical and electrophysiological analyses. We chose COS-7 cells for the biochemical analyses, because CHO cells express an endogenous SAP97 that interacts with Kv1.5 and interferes with coprecipitation assays (9). Thus the use of transfected COS-7 cells for the biochemical characterization of the tripartite complex simplified the interpretation of these experiments. Functional assays of transiently transfected COS-7 cells proved impractical, because variable fractions of the transfected COS-7 cells expressed Kv1.5-encoded potassium currents (18). By contrast, ⬎90% of the transfected CHO cells expressed Kv1.5-encoded currents. Thus all functional studies were done in CHO cells. Protein expression, pull-down assays, and immunoprecipitations. The expression of the 6xHis-tagged (S3 and S5) and GST fusion (GST1.5C1 and GST1.5C2) proteins was carried out in Escherichia coli strain BL21pLysS (Novagen) and DH5␣, respectively. After induction of expression by the addition of 0.5 mM isopropyl--Dthiogalactoside (Sigma), fusion proteins were affinity purified with agarose-bound glutathione (for GST1.5C1 and GST1.5C2) or antiPDZ antibody (for S3 and S5). For pull-down assays and immunoprecipitations, cells were lysed with buffer A [50 mM Tris (pH 7.5), 150 mM NaCl, 1% Igepal, 0.1% SDS, and a mixture of protease inhibitors] and centrifuged at 15,000 g for 10 min. For the pull-down assays, cell extracts were precleared for 1 h and incubated overnight with the corresponding agarose-bound affinity-purified fusion proteins. The beads were washed five times with buffer A, resuspended in equal volume of 2⫻ sample buffer, boiled for 5 min, and subjected to SDS-PAGE and immunoblotting. The immunoprecipitation experiments were carried out as described previously (18). For immunoblotting, samples were transferred to polyvinylidene difluoride (PVDF) membranes, and the blots were probed with the indicated antibodies. T7-directed cotranscription/cotranslation was carried out in the TNT reticulocyte lysate expression system (Promega) in the presence of L-[35S]methionine (Amersham Pharmacia Biotech). After incubation for 90 min at 30°C, the lysates were diluted 20-fold with buffer A, subjected to immunoprecipitation as described above, and analyzed by SDS-PAGE and fluorography. Electrophysiological studies. Recordings were made with an Axopatch-200B amplifier (Axon Instruments) using the standard whole cell configuration of the patch-clamp technique (3, 18, 29). Briefly, the pipette resistances were 2– 4 M⍀ when filled with (in mM) 50 KCl, 65 K-glutamate, 5 MgCl2, 5 EGTA, 10 HEPES, 5 K2-ATP, and 0.2 Tris-GTP (pH 7.2). The extracellular bath solution contained (in mM) 140 NaCl, 5.4 KCl, 1 CaCl2, 1 MgCl2, 0.33 NaH2PO4, 7.5 glucose, and 5 HEPES (pH 7.4). The currents were recorded at room temperature (21–23°C). The holding potential was ⫺70 mV; the test potential ranged from ⫺60 to ⫹60 mV, lasting 400 ms; and the tail currents were recorded at ⫺30 mV. Data are expressed as means ⫾ SE. ANOVA was applied for analyzing the multigroup data. Student’s t-test was used to compare unpaired data between two groups, and a two-tailed P ⬍ 0.05 was taken to indicate statistical significance. SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 H683 tested for interaction with Cav-3 in cotransfection experiments. Figure 2B shows that the deletion of the COOH terminus in S4 and S5 abrogated the interactions (lanes 7 and 8), whereas S2 and S3 interacted with Cav-3 with similar efficiency (lanes 5 and 6). These results indicate that the 99-amino acid segment COOH terminal to PDZ2 (residues 406 –504; ⌬SAP) is necesAJP-Heart Circ Physiol • VOL sary for the association of SAP97 and Cav-3. The sequence between the COOH-terminal end of PDZ2 and the NH2terminal end of PDZ3 contains 10 serine and 2 threonine residues, with at least 4 putative phosphorylation sites. Deletion of these sites (S4) eliminated the multiple bands observed (Fig. 2B, compare lanes 1 and 2 vs. lanes 3 and 4). 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 Fig. 1. SAP97 interacts with caveolin-3 (Cav-3) in vivo and in vitro. A: coimmunoprecipitation assays from heart lysates. Extracts from mouse ventricle (lane 1) were incubated with anti-PDZ (lane 2) or control IgG (lane 3), and the bound proteins were analyzed by immunoblotting with anti-PDZ (top) and anti-Cav-3 (bottom). IP, immunoprecipitation. B: pull down of Cav-3 and SAP97 from heart lysates using GST1.5C1, the GST-Kv1.5C-terminal fusion polypeptide shown in top scheme. Lysates from mouse ventricle (lane 1) were incubated with agarose-bound GST1.5C1 (lane 2) or control glutathione S-transferase (GST; lane 3), and the bound proteins were analyzed by immunoblotting with anti-PDZ (top blot) and anti-Cav-3 (middle blot). Bottom, Coomassie blue staining of the recombinant proteins used in the assay. TM, transmembrane segment; P, pore. C: coimmunoprecipitation assays in COS-7 cells. Cells were cotransfected with different combinations of Cav-3, SAP97, and pCDNA3 vector (⫺) cDNAs. One-half of the cell extracts were reacted with anti-Cav-3 (top right) and one-half with anti-PDZ (bottom right). The bound proteins were analyzed by immunoblot with anti-PDZ (top, lanes 5– 8) or anti-Cav-3 (bottom, lanes 5– 8). Lanes 1– 4 are total cell extracts. The immunoreactive endogenous polypeptide with a molecular weight similar to SAP97 (lanes 1 and 3, top) did not interact with Cav-3 (lane 5). D: coimmunoprecipitation assays from in vitro coexpressed Cav-3 and SAP97. Indicated cDNAs were transcribed and translated in vitro in the presence of [35S]methionine, and the resulting polypeptides were immunoprecipitated with anti-Cav-3 (lanes 4 and 5) or control IgG (lane 6). Bound proteins were separated by SDS-PAGE and visualized by fluorography. Lanes 1–3 depict total reticulocyte lysates programmed with the indicated cDNAs. H684 SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 We further confirmed the mapping of the Cav-3 binding site of SAP97 in pull-down experiments with bacterially expressed 6xHis-tagged S3 and S5 deletion mutants (Fig. 2A). The recombinant polypeptides were isolated from E. coli lysates by immunoprecipitation with anti-PDZ antibody (Fig. 2C, lanes 1 and 2) and then reacted with equal amounts of extracts from cells transfected with Cav-3. Consistent with our observations in transfected cells, only S3, but not S5, could pull down Cav-3 (Fig. 2C, lanes 4 and 3, respectively). To establish that ⌬SAP is necessary and sufficient for binding to Cav-3, we made a cDNA construct encoding a fusion protein in which ⌬SAP was linked to the COOH terminus of GFP (GFP-⌬SAP, see scheme in Fig. 2A). Cotransfection experiments showed that GFP-⌬SAP coprecipiAJP-Heart Circ Physiol • VOL tated with Cav-3 (Fig. 2D, lane 4), whereas GFP did not (lane 3), indicating that ⌬SAP is sufficient for binding to Cav-3. Both lysates contained equal amounts of Cav-3 and GFP polypeptides (Fig. 2D, lanes 1 and 2). Kv1.5 is recruited to the Cav-3/SAP97 scaffolding complex. Having established the formation of the Cav-3/SAP97 scaffolding complex, we next tested whether Kv1.5 could be recruited to the complex by examining the formation of a tripartite complex containing Kv1.5, Cav-3, and SAP97. After performing triple transfections, we used specific antibodies to immunoprecipitate each of the three polypeptides from total cell extracts. Analysis of the immunoprecipitates by SDSPAGE and Western blotting revealed that each of the three components of the putative tripartite complex coprecipitated 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 Fig. 2. Mapping of the Cav-3-binding site of SAP97. A: schematic representation of the various SAP97 mutants used. B: coimmunoprecipitation of Cav-3 and several SAP97 truncation mutants. Cells were cotransfected with Cav-3 and S2 (lanes 1 and 5), S3 (lanes 2 and 6), S4 (lanes 3 and 7), or S5 (lanes 4 and 8) cDNAs. One-half of the cell lysates were reacted with anti-Cav-3 (top, lanes 5– 8) and one-half with anti-PDZ (bottom, lanes 5– 8), and the bound proteins were analyzed by immunoblotting with anti-PDZ (top) or anti-Cav-3 (bottom). Lanes 1– 4 are total cell extracts. **Position of IgG light chain. C: pull down of Cav-3 from cell lysates by recombinant S3. S3 and S5 were isolated from E. coli lysates by immunoprecipitation with anti-PDZ and then incubated with equal portions of lysates from cells transfected with Cav-3 cDNA. Lanes 1 and 2 are recombinant polypeptides, precipitated and blotted with anti-PDZ. Lanes 3 and 4 are immunoblot analyses using anti-Cav-3 of the polypeptides pulled down with recombinant S5 and S3, respectively. D: coimmunoprecipitation of Cav-3 and GFP-⌬SAP. Cells were cotransfected with Cav-3 and green fluorescent protein (GFP; lanes 1 and 3) or Cav-3 and GFP-⌬SAP (lanes 2 and 4), and the cell lysates were immunoprecipitated with anti-GFP. Lanes 3 and 4 are immunoblot analyses using anti-Cav-3 of the polypeptides coprecipitated with GFP or GFP-⌬SAP, respectively. Lanes 1 and 2 represent the immunoblot analysis of total lysates using anti-Cav-3 (top) and anti-GFP (bottom). SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 with the other two polypeptides (Fig. 3A, lanes 1–3). No signal was detected in a control reaction with nonimmune mouse IgG (Fig. 3A, lane 4), indicating that the observed coimmunoprecipitations were specific. The T-D-L sequence located at the COOH terminus of Kv1.5 is critical for its association with SAP97, because this interaction was abolished in FL-Kv1.5-A, which has a mutation in the three COOH-terminal amino acid residues of Kv1.5 (T-D-L to A-A-A) (18). We reasoned that if SAP97, Cav-3, and FL-Kv1.5-A were all present in a tripartite complex, FL-Kv1.5-A might coimmunoprecipitate with SAP97 even though the direct interaction between the two proteins was AJP-Heart Circ Physiol • VOL abrogated. Triple transfection experiments showed that this was indeed the case (Fig. 3B): FL-Kv1.5-A coprecipitated with SAP97 in the presence (lane 7) but not in the absence (lane 6) of coexpressed Cav-3. Lanes 5 and 8 of Fig. 3B show the positive (FL-Kv1.5) and negative (no SAP97) controls of the experiment, respectively. This result further confirms the formation of the tripartite complex in this system and suggests that Kv1.5 may bind directly to both SAP97 and Cav-3. Kv1.5 interacts with Cav-3 through a region located at the NH2 terminus. We next designed a series of cotransfection experiments to confirm the specific association of Cav-3 with Fl-Kv1.5. Cotransfection of Cav-3 with FL-Kv1.5 (Fig. 4A, lanes 1–7) resulted in coprecipitation of Cav-3 and the channel using antibodies to both Cav-3 (lane 4) and FLAG epitope (lane 6). By contrast, cotransfection of Cav-3 with Myc-Kv2.1, a channel known to target to noncaveolar lipid rafts, did not result in coprecipitation using anti-Cav-3 antibodies (Fig. 4A, lane 10). We therefore concluded that Kv1.5 specifically associates with Cav-3. We made a series of Kv1.5 mutants to delineate the site of its interaction with Cav-3 (Fig. 4B). The long NH2- and COOH-terminal regions of Kv1.5 reside in the cytoplasmic side of the plasma membrane and are therefore candidates for the site of interaction with Cav-3, which localizes to the inner leaflet of the plasma membrane. Several lines of evidence indicated that the COOH-terminal cytoplasmic region of Kv1.5 is not involved in the interaction with Cav-3. First, in cotransfection experiments, a deletion mutant encoding a truncated FL-Kv1.5 polypeptide lacking the 86 COOH-terminal amino acids (TrFL-Kv1.5, Fig. 4B) was able to bind Cav-3 as efficiently as full-length FL-Kv1.5. Second, a cotransfected fusion protein containing the entire COOH-terminal region of Kv1.5 linked to GFP (GFP1.5C, Fig. 4B) failed to bind Cav-3. Finally, a GST fusion protein containing the entire COOHterminal region of Kv1.5 (GST1.5C2, Fig. 4B), expressed in E. coli and purified by glutathione-agarose affinity chromatography, failed to pull down Cav-3 from lysates of transfected cells (results not shown). To test for interaction of the NH2 terminus with Cav-3, we made a cDNA construct encoding a membrane-anchored polypeptide (1) containing the entire NH2-terminal cytoplasmic region of Kv1.5, followed by the first transmembrane segment (TM1), 18 amino acid residues corresponding to the first extracellular loop, and a COOH-terminal FLAG tag (Kv1.5N1-FL, Fig. 4B). Cotransfection of Cav-3 with Kv1.5N1-FL (Fig. 4C) resulted in coprecipitation of Cav-3 and the truncated channel with antibodies to both Cav-3 (Fig. 4C, top, lane 6) and FLAG epitope (Fig. 4C, bottom, lane 7), indicating that the Cav-3-binding site of Kv1.5 resides in the NH2-terminal portion of the molecule. Cav-3/SAP97 scaffolding complex selectively regulates the function of a glycosylation-deficient mutant of FL-Kv1.5. The above results indicate that multiple interactions are involved in the formation of the Kv1.5-SAP97-Cav-3 complex, in which each component binds the other two. We next designed a series of electrophysiological experiments to functionally validate our biochemical findings. An examination of the subcellular localization of the expressed polypeptides in triple transfection experiments in CHO cells revealed an extensive colocalization of Cav-3 with both Kv1.5 and SAP97, primarily at the level of the plasma membrane (not shown). Figure 5A shows that 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 Fig. 3. Formation of a tripartite complex containing SAP97, Cav-3, and FL-Kv1.5. A: coimmunoprecipitation of SAP97, FL-Kv1.5, and Cav-3. Lysates from two 100-mm dishes of triple-transfected cells were pooled, divided in 4 portions, and immunoprecipitated with anti-PDZ, anti-FLAG, anti-Cav-3, or nonimmune control IgG. Bound proteins were analyzed by immunoblotting with anti-PDZ (top), anti-FLAG (middle), and anti-Cav-3 (bottom). Lane 5 is a Western blot of the total cell extract. B: coimmunoprecipitation of SAP97 and FL-Kv1.5-A in the presence of Cav-3. Cells were cotransfected with different combinations of SAP97, FL-Kv1.5, FL-Kv1.5-A, and pCDNA3 vector (⫺) cDNAs. Cell extracts were reacted with anti-PDZ (lanes 5– 8), and the bound proteins were analyzed by immunoblotting with anti-PDZ (top), anti-FLAG (middle), and anti-Cav-3 (bottom). Lanes 1– 4 are immunoblots of total cell extracts. The immunoreactive endogenous polypeptide with a molecular weight similar to SAP97 (lanes 4 and 8, top) did not interact with FL-Kv1.5 or Cav-3 (lane 8). H685 H686 SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 FL-Kv1.5 transfected into CHO cells codes for rapidly activating outward currents. The cotransfection of Cav-3 with FL-Kv1.5 did not lead to changes in current density (not shown). During our studies on the effect of protein glycosylation on Kv1.5 function, we created a Ser292-to-Gly mutation that codes for a glycosylationdeficient mutant of the channel (FL-Kv1.5Glyc⫺). Transfection of FL-Kv1.5Glyc⫺ revealed that it coded for one-half the current density of FL-Kv1.5 (Fig. 5, A–C), with a shift to the Fig. 4. Kv1.5 interacts with Cav-3 through a region located at the NH2 terminus. A: coimmunoprecipitation of Cav-3 with FL-Kv1.5 coexpressed in transfected COS-7 cells. In lanes 1–7, cells were cotransfected with different combinations of Cav-3, FL-Kv1.5, and pCDNA3 vector (⫺) cDNAs. One-half of the cell extracts were reacted with anti-Cav-3 (lanes 4 and 5) and one-half with anti-FLAG (lanes 6 and 7). Bound proteins were analyzed by immunoblotting with anti-FLAG (top) or anti-Cav-3 (bottom). Lanes 1–3 are total cell extracts. In lanes 8 –11, cells were cotransfected with Myc-Kv2.1 and Cav-3 (lanes 8 and 10) or Myc-Kv2.1 and pCDNA3 vector (⫺) (lanes 9 and 11) cDNAs. Cell extracts were reacted with anti-Cav-3, and the bound proteins were analyzed by immunoblotting with anti-Myc (lanes 10 and 11). Lanes 8 and 9 are total cell extracts. B: schematic representation of the various Kv1.5 fusion and truncation mutants. C: coimmunoprecipitation of Kv1.5N1-FL and Cav-3. Cells were cotransfected with different combinations of Kv1.5N1-FL, Cav-3, FL-Kv1.5, and pCDNA3 vector (⫺) cDNAs. One-half of the cell extracts were reacted with anti-Cav-3 (top, lanes 5–7) and one-half with anti-FLAG (bottom, lanes 5–7). Bound proteins were analyzed by immunoblot with anti-FLAG (top) or anti-Cav-3 (bottom). Lanes 1– 4 are total cell extracts. AJP-Heart Circ Physiol • VOL 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 right of the voltage dependence of steady-state activation (Fig. 5D). These results suggest that FL-Kv1.5Glyc⫺, similar to the Shaker and Kv1.1 channels, may have lower membrane expression than FL-Kv1.5. This mutant was recruited to the Cav-3/SAP97 scaffolding complex similarly to FL-Kv1.5 (Fig. 5E). In striking contrast with the lack of effect of Cav-3 on the wild-type channel, cotransfection of increasing amounts of Cav-3 cDNA inhibited up to 90% of the FL-Kv1.5Glyc⫺encoded currents (from 312.3 ⫾ 22.6 to 33.1 ⫾ 6.1 pA/pF, P ⬍ 0.001; Fig. 6A). Moreover, Cav-3 shifted the voltage dependence of the steady-state activation curve of the channel to the left (Fig. 6B). These results suggest that the decrease in the macroscopic Kv1.5-encoded currents due to the coexpression with Cav-3 (Fig. 6A) was not due to a modification of the gating properties of Kv1.5, because such a modification would have led to an increase in current. Rather, the marked reduction of the outward currents most likely reflects a decrease in the number of functional Kv1.5 channels on the cell surface. We next tested whether cotransfection of exogenous SAP97 could abrogate the suppressive effect of Cav-3 on FLKv1.5Glyc⫺-encoded currents. Figure 6C shows that the cotransfection of SAP97 with FL-Kv1.5Glyc⫺ and Cav-3 abolished the inhibitory effect of Cav-3. We therefore hypothesized that SAP97, through its high-affinity interactions with the COOH terminus of Kv1.5 and Cav-3, competed with an endogenous PDZ domain-containing protein that mediated the inhibition and rescued the basal current level. To test the hypothesis that SAP97 interactions with Cav-3 are critical for the rescue, we coexpressed FL-Kv1.5Glyc⫺ with SAP97, Cav-3, and GFP-⌬SAP, the GFP fusion protein that contains the segment of SAP97 that interacts with Cav-3 (see Fig. 2A). GFP-⌬SAP abolished the rescue by SAP97 (Fig. 6D), providing functional proof that this segment is important for SAP97/ Cav-3 association and that these interactions are crucial for abrogating the inhibitory effect of Cav-3. We next mutated the COOH-terminal T-D-L of FLKv1.5Glyc⫺ to A-A-A to confirm that an endogenous MAGUK protein was involved in the inhibition of channel function by Cav-3. This mutant (FL-Kv1.5Glyc⫺A) binds normally to Cav-3 but does not interact with PDZ-containing proteins (not shown). In contrast to the observed inhibition of FLKv1.5Glyc⫺ activity by Cav-3, the cotransfection of Cav-3 with FL-Kv1.5Glyc⫺A did not cause any decrease in the FL- SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 H687 Kv1.5Glyc⫺A⫺-encoded currents (Fig. 6E). This result confirms that the observed inhibition of FL-Kv1.5Glyc⫺ by Cav-3 was mediated by an endogenous MAGUK protein, and that the interaction with Cav-3 alone was not sufficient to inhibit channel activity. This conclusion is further supported by the finding that the coexpression of FL-Kv1.5Glyc⫺ and Cav-3 with GFP1.5C, a GFP fusion protein containing the whole COOHterminal cytoplasmic tail of Kv1.5 (see scheme in Fig. 4B), also abolished the inhibitory effect of Cav-3 on FL-Kv1.5Glyc⫺ (Fig. 6F). This fusion protein likely competed with FLAJP-Heart Circ Physiol • VOL Kv1.5Glyc⫺ for the endogenous PDZ-containing protein that mediates the inhibition. It is important to note that the interaction of Kv1.5 with the endogenous MAGUK did not exert a direct effect on Kv1.5 activity, because the mutation of the COOH terminus of Kv1.5 to A-A-A had no apparent effect on the overall channel-encoded currents [Fig. 6, compare FLKv1.5Glyc⫺ (A) and FL-Kv1.5Glyc⫺A (E)]. Collectively, these results support the notion that the Cav-3/SAP97 complexes can recruit Kv1.5 to form tripartite complexes that may regulate channel function. 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 Fig. 5. Expression of FL-Kv1.5- and FL-Kv1.5Glyc⫺-encoded currents. A and B: representative current traces from Chinese hamster ovary cells transfected with either 0.1 g FL-Kv1.5 (A) or 0.1 g FL-Kv1.5Glyc⫺ (B). C: currentvoltage (I-V) relationships at the end of the test pulse. The glycosylation-deficient mutant encodes smaller currents (697.6 ⫾ 80.1 vs. 312.3 ⫾ 22.6 pA/pF at ⫹60 mV, P ⬍ 0.001). D: steady-state activation of FL-Kv1.5 and FL-Kv1.5Glyc⫺ determined by tail current analyses [E, voltage at half-maximal activation (V1/2) ⫺2.8 ⫾ 1.3 mV, n ⫽ 15; F, V1/2 3.18 ⫾ 0.6 mV, n ⫽ 10; P ⬍ 0.001]. The mutation shifted the voltage dependence of steady-state activation to the right. Data are expressed as means ⫾ SE. Student’s t-test and ANOVA were used to evaluate the statistical significance. E: coimmunoprecipitation assays from COS-7 cells cotransfected with Cav-3, SAP97, and FL-Kv1.5Glyc⫺. Anti-FLAG antibody precipitated the tripartite complex including SAP97 and Cav-3 only in the presence of the channel (lane 4) but not in its absence (lane 3). Immunoblots with anti-PDZ antibody (top), anti-FLAG antibody (middle), and anti-Cav-3 (bottom) are shown. H688 SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 DISCUSSION In neuronal and epithelial cells, the polarized expression of membrane proteins relies largely on scaffolding proteins, which dictate the rules for the organization of large complexes involving specific interactions. Scaffolding proteins often interact with one another, thus providing multiple binding sites for complex assembly. Examples of these interactions are the association of the PDZ-containing proteins CASK, Mint1, and Velis at the synaptic junction (4) and the recruitment of SAP97 to the lateral surface of epithelial cells through its interaction with CASK (13). Similar principles possibly govern the assembly of macromolecular complexes in the heart. We propose that the interaction of Cav-3 with SAP97 may form the nucleation site for the assembly of a macromolecular complex that contains ion channels, receptors, and signaling molecules involved in electrical impulse propagation. Our results provide evidence for the recruitment of Kv1.5 to the Cav-3/SAP97 AJP-Heart Circ Physiol • VOL scaffolding complex in heterologous expression systems and suggest that protein interactions within this complex can regulate channel function. We mapped the Cav-3 binding site of SAP97 to a segment located COOH terminal to PDZ2. The results of a BLAST search that used this fragment as query revealed that it contains a serine- and proline-rich region (amino acid residues 430 – 447) that has a significant degree of sequence similarity with two caveolar enzymes: ceramidase (24) and adenylyl cyclase type V (25). Thus this segment may represent a previously unrecognized caveolin-binding domain present in various proteins. In a molecular model of SAP97 created by Wu et al. (28), the Cav-3-binding site is predicted to be a -turn readily accessible from the solvent, facing the exterior of the molecule. It is interesting that this -turn forms a lid over the PDZ2 peptide-binding pocket, suggesting that the channel interaction with PDZ2 and the binding 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 Fig. 6. I-V curves of FL-Kv1.5Glyc⫺. A: Cav-3 inhibits FL-Kv1.5Glyc⫺-encoded currents in a dose-dependent manner. Control transfection of 0.1 g FL-Kv1.5Glyc⫺ (312.3 ⫾ 22.6 pA/pF, n ⫽ 18) and cotransfection of 0.1 g FL-Kv1.5Glyc⫺ with 0.01 g Cav-3 (356.9 ⫾ 32.6 pA/pF, n ⫽ 18), 0.1 g Cav-3 (106.5 ⫾ 17.4 pA/pF, n ⫽ 20), or 1.0 g Cav-3 (33.1 ⫾ 6.1 pA/pF, n ⫽ 21) (P ⬍ 0.001) are shown. B: Cav-3 shifts the voltage activation curve to the left. FLKv1.5Glyc⫺ alone (0.1 g) (n ⫽ 10) and 0.1 g FL-Kv1.5Glyc⫺ with 0.1 g Cav-3 (n ⫽ 12) (V1/2 of ⫺2.0 ⫾ 1.5 vs. 3.2 ⫾ 0.6 mV; P ⬍ 0.05) are shown. C: SAP97 abrogates the inhibitory effect of Cav-3. FLKv1.5Glyc⫺ alone (0.1 g) (312.3 ⫾ 22.6 pA/pF, n ⫽ 18), 0.1 g FL-Kv1.5Glyc⫺ cotransfected with 0.1 g Cav-3 (106.5 ⫾ 17.4 pA/pF, n ⫽ 20), and 0.1 g FL-Kv1.5Glyc⫺ cotransfected with 0.1 g Cav-3 and 1.0 g SAP97 [317.8 ⫾ 41.5 pA/pF, n ⫽ 24; P ⫽ not significant (NS) vs. control] are shown. D: SAP97 did not abolish Cav-3 inhibition in the presence of GFP-⌬SAP. FLKv1.5Glyc⫺ alone (0.1 g) (312.3 ⫾ 22.6 pA/pF, n ⫽ 18); 0.1 g FL-Kv1.5Glyc⫺ cotransfected with 2.0 g GFP-⌬SAP (261.7 ⫾ 41.1 pA/pF, n ⫽ 11); 0.1 g FL-Kv1.5Glyc⫺ cotransfected with 0.1 g Cav-3 and 2.0 g GFP-⌬SAP (132.4 ⫾ 19.0 pA/pF, n ⫽ 13); and 0.1 g FL-Kv1.5Glyc⫺ cotransfected with 1.0 g SAP97, 0.1 g Cav-3, and 2.0 g GFP-⌬SAP (170.2 ⫾ 29.5 pA/pF, n ⫽ 15; P ⫽ NS) are shown. E: 0.1 g Cav-3 did not inhibit the currents encoded by 0.1 g FL-Kv1.5Glyc⫺A (291.7 ⫾ 43.0 vs. 402.8 ⫾ 48.3 pA/pF; P ⫽ 0.09). F: GFP1.5C abrogates the Cav-3 inhibition. Cotransfection of 0.1 g FLKv1.5Glyc⫺ with 1.0 g GFP1.5C (293.0 ⫾ 65.1 pA/pF, n ⫽ 7) or with 1.0 g GFP1.5C and 0.1 g Cav-3 (279.4 ⫾ 45.1 pA/pF, n ⫽ 10; P ⫽ NS) are shown. Data are means ⫾ SE. Student’s t-test and ANOVA were used to evaluate the statistical significance. Current density is in pA/pF at ⫹60 mV. The amount of DNA in all transfections was kept constant with vector DNA (pcDNA3). SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 AJP-Heart Circ Physiol • VOL ACKNOWLEDGMENTS We thank Thomas Michel and Barbara Wible for Myc-tagged Cav-3 and Kv2.1 cDNAs, respectively. We also thank Giselle Martı́nez Nöel for contributing to the experiment shown in Fig. 1B. GRANTS G. Koren is a recipient of grants from the National Heart, Lung, and Blood Institute. REFERENCES 1. Babila T, Moscucci A, Wang H, Weaver FE, and Koren G. Assembly of mammalian voltage-gated potassium channels: evidence for an important role of the first transmembrane segment. Neuron 12: 615– 626, 1994. 2. Bennett E, Urcan MS, Tinkle SS, Koszowski AG, and Levinson SR. Contribution of sialic acid to the voltage dependence of sodium channel gating. A possible electrostatic mechanism. J Gen Physiol 109: 327–343, 1997. 3. Brunner M, Guo W, Mitchell GF, Buckett PD, Nerbonne JM, and Koren G. Characterization of mice with a combined suppression of I(to) and I(K,slow). Am J Physiol Heart Circ Physiol 281: H1201–H1209, 2001. 4. Butz S, Okamoto M, and Sudhof TC. A tripartite protein complex with the potential to couple synaptic vesicle exocytosis to cell adhesion in brain. Cell 94: 773–782, 1998. 5. Craven SE and Bredt DS. PDZ proteins organize synaptic signaling pathways. Cell 93: 495– 498, 1998. 6. Eldstrom J, Choi WS, Steele DF, and Fedida D. SAP97 increases Kv1.5 currents through an indirect N-terminal mechanism. FEBS Lett 547: 205–211, 2003. 7. Galbiati F, Engelman JA, Volonte D, Zhang XL, Minetti C, Li M, Hou H Jr, Kneitz B, Edelmann W, and Lisanti MP. Caveolin-3 null mice show a loss of caveolae, changes in the microdomain distribution of the dystrophin-glycoprotein complex, and t-tubule abnormalities. J Biol Chem 276: 21425–21433, 2001. 8. Garner CC, Nash J, and Huganir RL. PDZ domains in synapse assembly and signalling. Trends Cell Biol 10: 274 –280, 2000. 9. Godreau D, Vranckx R, Maguy A, Goyenvalle C, and Hatem SN. Different isoforms of synapse-associated protein, SAP97, are expressed in the heart and have distinct effects on the voltage-gated K⫹ channel Kv1.5. J Biol Chem 278: 47046 – 47052, 2003. 10. Godreau D, Vranckx R, Maguy A, Rucker-Martin C, Goyenvalle C, Abdelshafy S, Tessier S, Couetil JP, and Hatem SN. Expression, regulation and role of the MAGUK protein SAP-97 in human atrial myocardium. Cardiovasc Res 56: 433– 442, 2002. 11. Hagiwara Y, Sasaoka T, Araishi K, Imamura M, Yorifuji H, Nonaka I, Ozawa E, and Kikuchi T. Caveolin-3 deficiency causes muscle degeneration in mice. Hum Mol Genet 9: 3047–3054, 2000. 12. Khanna R, Myers MP, Laine M, and Papazian DM. Glycosylation increases potassium channel stability and surface expression in mammalian cells. J Biol Chem 276: 34028 –34034, 2001. 13. Lee S, Fan S, Makarova O, Straight S, and Margolis B. A novel and conserved protein-protein interaction domain of mammalian Lin-2/CASK binds and recruits SAP97 to the lateral surface of epithelia. Mol Cell Biol 22: 1778 –1791, 2002. 14. Leonoudakis D, Mailliard W, Wingerd K, Clegg D, and Vandenberg C. Inward rectifier potassium channel Kir2.2 is associated with synapseassociated protein SAP97. J Cell Sci 114: 987–998, 2001. 15. Martens JR, Sakamoto N, Sullivan SA, Grobaski TD, and Tamkun MM. Isoform-specific localization of voltage-gated K⫹ channels to distinct lipid raft populations. Targeting of Kv15 to caveolae. J Biol Chem 276: 8409 – 8414, 2001. 16. McNally EM, de Sa Moreira E, Duggan DJ, Bonnemann CG, Lisanti MP, Lidov HG, Vainzof M, Passos-Bueno MR, Hoffman EP, Zatz M, and Kunkel LM. Caveolin-3 in muscular dystrophy. Hum Mol Genet 7: 871– 877, 1998. 17. Minetti C, Sotgia F, Bruno C, Scartezzini P, Broda P, Bado M, Masetti E, Mazzocco M, Egeo A, Donati MA, Volonte D, Galbiati F, Cordone G, Bricarelli FD, Lisanti MP, and Zara F. Mutations in the caveolin-3 gene cause autosomal dominant limb-girdle muscular dystrophy. Nat Genet 18: 365–368, 1998. 18. Murata M, Buckett PD, Zhou J, Brunner M, Folco E, and Koren G. SAP97 interacts with Kv1.5 in heterologous expression systems. Am J Physiol Heart Circ Physiol 281: H2575–H2584, 2001. 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 of Cav-3 to the -turn may be mutually regulated. Furthermore, the occurrence of multiple potential phosphorylation sites in this segment raises the possibility that the interaction of SAP97 and Cav-3 is dynamic. In this study and in our previous work (18) we have demonstrated that the T-D-L sequence at the COOH terminus of Kv1.5 is essential for its interaction with SAP97. These results differ from those recently reported by Eldstrom et al. (6) and presumably reflect differences in the experimental systems used (COS-7 vs. human embryonic kidney cells, FLAG-tagged rKv1.5 vs. T7-tagged hKv1.5) that are likely related to the presence of distinct sets of endogenous PDZ proteins and SAP97 isoforms in different cell types. Interestingly, Eldstrom et al. show that the NH2 terminus of Kv1.5, which we show to interact with Cav-3, is necessary for the regulation of channel function by SAP97 and suggest that both Kv1.5 and SAP97 are in a complex with ␣-actinin. Those results and ours highlight the importance of the formation of complexes involving multiple protein interactions for channel regulation. Our observation that Cav-3 selectively inhibits the currents elicited by glycosylation-deficient Kv1.5 channels is puzzling. N-linked glycosylation is a ubiquitous modification of potassium channels. Previous studies suggested that this cotranslational modification may control folding, trafficking, and stability of these channels (2). However, the exact role may vary among different channels. For example, glycosylation reflects surface expression of HERG (23, 30), alters the voltage dependence of activation of Kv1.1 and KvLQT1/minK, and modulates the open probability of ROMK1 (20, 31). We and other investigators have shown that N-linked glycosylation is not essential for surface expression of Kv1.1 and Shaker channels in Xenopus oocytes (1). However, recently Khanna et al. (12) showed that unglycosylated Shaker mutants have a faster turnover rate and reduced surface expression due to rapid degradation by the proteasome pathway (12). We are currently investigating the mechanism underlying the selective downregulation of FL-Kv1.5Glyc⫺ by Cav-3. Our working hypothesis is that the Cav-3/endogenous MAGUK complex plays a role in quality control of channel assembly, mediating the selective retention of FL-Kv1.5Glyc⫺ in the secretory pathway and targeting it for degradation. Alternatively, the complex may promote the selective internalization of FL-Kv1.5Glyc⫺ from the plasma membrane. However, we cannot rule out the possibility that the high level of surface expression of wildtype Kv1.5 may somehow mask an eventual regulation of this channel by Cav-3 as well, an effect that could be unmasked in the case of the glycosylation-deficient mutant because it has a lower level of surface expression. In summary, our results show that a scaffolding complex containing Cav-3 and SAP97 can recruit Kv1.5 by means of multiple protein-protein interactions. Furthermore, we show that this scaffolding complex regulates the expression of currents encoded by a glycosylation-deficient mutant of Kv1.5. In excitable cells, the targeting of ion channels to appropriate membrane subdomains and their organization in multiprotein complexes are critical for cell excitation. Our results represent a first step towards the elucidation of the assembly of potassium channel-containing complexes (channelosomes) in the heart. H689 H690 SCAFFOLDING COMPLEX OF CAVEOLIN-3 AND SAP97 AJP-Heart Circ Physiol • VOL 26. Tiffany AM, Manganas LN, Kim E, Hsueh YP, Sheng M, and Trimmer JS. PSD-95 and SAP97 exhibit distinct mechanisms for regulating K(⫹) channel surface expression and clustering. J Cell Biol 148: 147–158, 2000. 27. Woodman SE, Park DS, Cohen AW, Cheung MW, Chandra M, Shirani J, Tang B, Jelicks LA, Kitsis RN, Christ GJ, Factor SM, Tanowitz HB, and Lisanti MP. Caveolin-3 knock-out mice develop a progressive cardiomyopathy and show hyperactivation of the p42/44 MAPK cascade. J Biol Chem 277: 38988 –38997, 2002. 28. Wu H, Reissner C, Kuhlendahl S, Coblentz B, Reuver S, Kindler S, Gundelfinger ED, and Garner CC. Intramolecular interactions regulate SAP97 binding to GKAP. EMBO J 19: 5740 –5751, 2000. 29. Zhou J, Jeron A, London B, Han X, and Koren G. Characterization of a slowly inactivating outward current in adult mouse ventricular myocytes. Circ Res 83: 806 – 814, 1998. 30. Zhou Z, Gong Q, Epstein ML, and January CT. HERG channel dysfunction in human long QT syndrome. Intracellular transport and functional defects. J Biol Chem 273: 21061–21066, 1998. 31. Zhu J, Watanabe I, Gomez B, and Thornhill WB. Determinants involved in Kv1 potassium channel folding in the endoplasmic reticulum, glycosylation in the Golgi, and cell surface expression. J Biol Chem 276: 39419 –39427, 2001. 287 • AUGUST 2004 • www.ajpheart.org Downloaded from http://ajpheart.physiology.org/ by 10.220.33.3 on August 3, 2017 19. Okamoto T, Schlegel A, Scherer PE, and Lisanti MP. Caveolins, a family of scaffolding proteins for organizing “preassembled signaling complexes” at the plasma membrane. J Biol Chem 273: 5419 –5422, 1998. 20. Pabon A, Chan KW, Sui JL, Wu X, Logothetis DE, and Thornhill WB. Glycosylation of GIRK1 at Asn119 and ROMK1 at Asn117 has different consequences in potassium channel function. J Biol Chem 275: 30677–30682, 2000. 21. Parton RG, Way M, Zorzi N, and Stang E. Caveolin-3 associates with developing T-tubules during muscle differentiation. J Cell Biol 136: 137–154, 1997. 22. Pawson T and Scott JD. Signaling through scaffold, anchoring, and adaptor proteins. Science 278: 2075–2080, 1997. 23. Petrecca K, Atanasiu R, Akhavan A, and Shrier A. N-linked glycosylation sites determine HERG channel surface membrane expression. J Physiol 515: 41– 48, 1999. 24. Romiti E, Meacci E, Tanzi G, Becciolini L, Mitsutake S, Farnararo M, Ito M, and Bruni P. Localization of neutral ceramidase in caveolinenriched light membranes of murine endothelial cells. FEBS Lett 506: 163–168, 2001. 25. Rybin VO, Xu X, Lisanti MP, and Steinberg SF. Differential targeting of beta-adrenergic receptor subtypes and adenylyl cyclase to cardiomyocyte caveolae. A mechanism to functionally regulate the cAMP signaling pathway. J Biol Chem 275: 41447– 41457, 2000.