PDZ (PSD-95/Disc large/Zonula occludens-1) protein interaction domains bind to cytoplasmic protein C-termini of transmembrane proteins. In order to identify new interaction partners of the voltage-gated L-type Ca2+ channel 1.2 and the plasma membrane Ca2+ ATPase 4b (PMCA4b), we used PDZ domain arrays probing for 124 PDZ domains. We confirmed this by GST pull-downs and immunoprecipitations. In PDZ arrays, strongest interactions with 1.2 and PMCA4b were found for the PDZ domains of SAP-102, MAST-205, MAGI-1, MAGI-2, MAGI-3, and ZO-1. We observed binding of the 1.2 C-terminus to PDZ domains of NHERF1/2, Mint-2, and CASK. PMCA4b was observed to interact with Mint-2 and its known interactions with Chapsyn-110 and CASK were confirmed. Furthermore, we validated interaction of 1.2 and PMCA4b with NHERF1/2, CASK, MAST-205 and MAGI-3 via immunoprecipitation. We also verified the interaction of 1.2 and nNOS and hypothesized that nNOS overexpression might reduce Ca2+ influx through 1.2. To address this, we measured Ca2+ currents in HEK 293 cells co-expressing 1.2 and nNOS and observed reduced voltage-dependent 1.2 activation. Taken together, we conclude that 1.2 and PMCA4b bind promiscuously to various PDZ domains, and that our data provides the basis for further investigation of the physiological consequences of these interactions.

1. Introduction

PDZ domains are protein interaction motifs that play a crucial role in cellular signaling and bind specifically to cytoplasmatic carboxyl (C-) terminal sequences of their interacting proteins, which often belong to transmembrane receptor and ion channel families. This motif typically spans 90–100 amino acids and was first found in three polypeptides: the mammalian protein postsynaptic density-95 (PSD-95), the Drosophila melanogaster epithelial tumor suppressor protein Discs Large (Dlg), and the mammalian epithelial tight junction protein Zonula occludens-1 (ZO-1) [17]. Typical ligands for PDZ domains are the high voltage-activated L-type calcium channel (LTCC) Cav1.2 and the plasma membrane calcium ATPase (PMCA) 4b, both essential for calcium homeostasis in excitable and nonexcitable cells.

Voltage-gated calcium channels (Cav) allow the cellular entry of calcium ( ) ions and initiate muscle excitation-contraction coupling, neurotransmitter release, gene expression, or hormone secretion. Cav1.2 channels play a major role for the voltage-dependent influx. Its overexpression augmented influx into cardiomyocytes, thereby increasing cardiac contractile force [8, 9]. Further studies demonstrated the importance of Cav1.2 in the heart. Homozygous Cav1.2 knockout mice die before day 14.5 p.c., presumably because L-type channels are indispensable during heart development [10, 11]. Furthermore, mutations in the CACNA1C gene, which codes for the Cav1.2 subunit, are causative for the Timothy syndrome. This disease is characterized by a multiorgan disorder with serious cardiac defects, sudden death, and other comorbidities [12, 13].

PMCAs extrude from the cytoplasm into the extracellular space, and they, as primary ion pumps, are regarded as essential for long term maintenance of low intracellular , which is a prerequisite for subsequent -dependent intracellular signaling and spatial changes in concentrations [1416]. Mammalian PMCAs are products of four genes (ATP2B1–ATP2B4), which share 80–90% sequence homology at the amino acid level in human, rat, and mouse [17]. In many cell types, PMCAs are concentrated in caveolae [1820] that are rich in receptors, signal transducers, effectors, and structural proteins, all of which are important for many signaling pathways and the maintenance of cytoskeletal networks [21, 22]. Differential splicing of PMCA RNA transcripts results in a variety of subtypes of these isoforms. More than 20 splice variants have been identified [17]. The C-termini of the b-splice variants of all PMCA isoforms is supposed to bind preferentially to type 1 PDZ domains. The human motif of the PMCA4b interacts with members of the membrane-associated guanylate kinase (MAGUK) family [17, 18], such as PSD-95/SAP90, SAP97/hDlg, SAP-102, PSD-93/Chapsyn-110 [18], and calcium/calmodulin-dependent serine protein kinase (CASK) [23], which contain one to six PDZ domains [24, 25]. MAGUKs are essential for organizing signaling complexes, and membrane protein trafficking [26]. PMCA4b interacts with PMCA-interacting single PDZ protein (PISP) [27], Na+/H+ exchanger regulatory factor 2 (NHERF2) [28], and neuronal nitric oxide synthase (nNOS), thereby regulating enzyme activity [29]. Besides C-terminal mediated interactions, PMCA4b also interacts with other proteins via other domains, for example, via the second intracellular loop with the tumor suppressor Ras-associated factor 1, calcineurin, and α1-syntrophin [3032]. So far, there are only few studies about the proteins that may possibly interact with Cav1.2. To gain more insight into the participating pathways, and the physiological roles of Cav1.2 and PMCA4b, we searched for new interaction partners. In this study, we show that the Cav1.2 motif ( ) interacts with the MAGUK protein CASK, and the membrane-associated guanylate kinase-inverted proteins 1, 2, 3 (MAGI-1, MAGI-2, MAGI-3), a subfamily of MAGUK proteins. MAGI-1, MAGI-2, MAGI-3 are also new interaction partners of PMCA4b. Interestingly, we discovered additional proteins that interact with Cav1.2 as well as with PMCA4b. These are the microtubule-associated testis specific serine/threonine kinase (MAST-205), the tight junction protein Zonula occludens 1 (ZO-1), Rho guanine exchange factor 11 and 12 (GEF11, GEF12), and Mint-2. In addition, we confirmed the lack of PMCA4b interactions with NHERF1 and NHERF2, and in contrast, the binding of Cav1.2 to the PDZ domain of NHERF1/2. We further showed that the nNOS interacts with Cav1.2, thereby influencing transmembrane calcium currents.

Our results provide the basis for more detailed analyses of PMCA4b and Cav channel regulation, specifically in dependence of the associated PDZ domain containing proteins.

2. Materials and Methods

2.1. Plasmid Constructs

Plasmids encoding Histidine (His)-tagged carboxyl termini of Cav1.2 (LTCC) and PMCA4b were made by standard molecular biology techniques. Codons for the final 10 (Cav1.2 ) or 15 (PMCA4b) amino acids were cloned as PstI-XbaI fragments into the pEXP vector (containing a 6xHis tag, Panomics, Fremont, CA, USA) to produce His-tagged fusion proteins (pEXP-Cav1.2 and pEXP-PMCA4b) for PDZ Domain Arrays (Panomics). The same codons of LTCC were inserted into the pGEX-4T-3 (GE Healthcare Bio-Sciences AB, Uppsala) vector to produce Glutathion-S-Transferase (GST) fusion proteins (pGEX-4T-3-LTCC). The expression constructs pGEX-4T-1-nNOS-PDZ and pcDNA3-nNOS were kind gifts from David Bredt (University of California, San Francisco, CA). The plasmids pRK5-kinase-MAST205 and pRK5-kinase-PDZ-MAST205 were kind gifts from Rafael Pulido (Centro de Investigacion Principe Felipe, Valencia, Spain) and the plasmid pcDNA3-Cav1.2 was a gift from Sebastian Maier (University of Wuerzburg, Germany).

2.2. TranSignal PDZ Domain Array

Plasmid constructs pEXP-Cav1.2 and pEXP-PMCA4b were transformed into E. coli BL21 (DE3) bacteria, which were then inoculated in 3 mL of LB/Amp (100 μg/mL) and shaken for an hour at 37°C at 300 rpm. After reaching an OD600 of 0.5–0.8, 1 mM isopropyl-1-thio-β-D-galactopyranoside (IPTG) was added to the bacteria before further growth for 3-4 h at 37°C to induce target protein expression. Cells were collected by centrifugation (4,000 ×g, 10 min 4°C), resuspended in 2 mL resuspension buffer (Panomics), and lysed with a sonicator. After clearing at 14,000 rpm for 5 min at 4°C, protein content of the supernatant was quantified via bicinchoninic acid (BCA) protein assay. The PDZ Domain Arrays were prepared according to the manufacturer’s instructions, incubated with diluted bacterial extract (5 μg/mL in blocking buffer) containing the His-tagged fusion proteins for 1-2 h at room temperature, and washed thrice with wash buffer for 5 min. They were then incubated with 1x Anti-Histidine horse radish peroxidise (HRP) conjugate (Panomics), diluted in wash buffer, for 1-2 h at room temperature. Antibody complexes were detected by enhanced chemiluminescence, using ECL Western blotting substrate (ECL Plus, GE Healthcare). X-rays were scanned and the signals were quantified with ImageJ. The values were standardized against the GST negative control.

2.3. Purification of GST Fusion Proteins

GST and GST fusion proteins were expressed in E. coli BL21(DE3) by induction with 1 mM IPTG for 6 h. Bacterial cells were pelleted, resuspended in PBS (137 mM NaCl, 2.7 mM KCl, 100 mM Na2HPO4, 2 mM KH2PO4, and pH 7.4) containing protease inhibitor (complete EDTA-free Protease Inhibitor Cocktail Tablets, Roche), and lysed by addition of lysozyme (1 mg/mL) and sonication. The lysate was cleared by centrifugation at 30,000 g for 20 min at 4°C. The pellet was resuspended in PBS and the resulting lysate was incubated with Glutathione-Sepharose (GE Healthcare) at constant rotation for 2 h at 4°C. The quantity of the washed and bound fusion proteins was estimated by Coomassie Blue staining of  SDS-polyacrylamide gels.

2.4. GST Pull-Down Assays

To prepare tissue lysates, organs were removed from mice and immediately homogenized in cold RIPA buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% Na-Deoxycholate, protease inhibitor, and optional 0.1% SDS). The homogenate was centrifuged at 4,000 ×g for 3 min. For cell lysates, the same RIPA buffer was used. 500 μg of each supernatant and 3 μg of bound GST or GST fusion proteins on agarose beads were rotated overnight at 4°C. The beads were pelleted and washed three times in PBS with protease inhibitors. Bound proteins were eluated in 2x Laemmli buffer (4% SDS, 20% glycerol, 10% 2-mercaptoethanol, 0,004% bromphenol blue, 0,125 M Tris HCl, and pH 6.8) [33] and separated on polyacrylamide gels followed by transfer onto nitrocellulose. Nitrocellulose membranes were blocked in TBST (TBS + 0.1% Tween) with 5% milk and incubated with primary and secondary antibodies. HRP-coupled secondary antibodies were detected using ECL Plus.

2.5. Coimmunoprecipitations

HEK 293 cells (DMEM supplemented with 10% FCS), ECV cells (DMEM supplemented with 10% FCS, 4.5 g/L glucose), and HEK 293 cells stably expressing (Cav1.2b) and the Cavβ2a subunit of the smooth muscle L-type calcium channel [34, 35] (a gift from Andrea Welling, TU Munich, Germany) (DMEM supplemented with 10% FCS, 200 μg/mL G418 plus 100 μg/mL hygromycin B) were grown to ~80% confluence. Cells were transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. After ~48 h, cells were rinsed with PBS and lysed in RIPA buffer without SDS. Subsequent to 10 min incubation on ice, cells were scraped from the plates and pelleted at 13,000 ×g for 10 min at 4°C. 300 μg of the lysate was used for each immunoprecipitation. 1–5 μg of antibodies: anti-Cav1.2 (Alomone Labs), anti-MAGI-3 (Abcam), anti-HA (Covance), and anti-PMCA4-JA9 (Sigma) were added, accordingly. After 2 h rocking at 4°C, 50 μL of protein A/G-agarose was added to the mixtures and rotated overnight at 4°C. In all immunoprecipitations, the positive controls were inputs of the respective protein and the negative controls were the relevant protein samples incubated with protein A/G-agarose beads and irrelevant antibodies (anti-AT2). Protein A/G-agarose was pelleted at 4,000 ×g for 30 s and washed twice with RIPA buffer containing protease inhibitors. Bound proteins were eluated in 2x Laemmli buffer and separated on polyacrylamide gels followed by transfer onto nitrocellulose for Western blotting as described above.

2.6. Antibodies for Immunoblotting

The following antibodies were used for immunoblotting: anti-ZO-1 (BD Transduction Laboratories) used at 1 : 1000 dilution, anti-nNOS (Zymed Laboratories) used at 1 : 2000 dilution, anti-Cav1.2 (Alomone Labs) diluted 1 : 200, anti-CASK (BD Transduction) diluted 1 : 1000, anti-NHERF1 (Cell Signaling) diluted 1 : 1000, anti-MAGI-3 (Abcam) diluted 1 : 1000, and anti-HA (Covance) diluted 1 : 1000. Secondary goat anti-mouse antibodies were purchased from Jackson Immuno Research and used at 1 : 5000 dilution, goat anti-rabbit (Jackson Immuno Research) used at 1 : 10000 dilution. From eBioscience we used rabbit IgG TrueBlot (diluted 1 : 1000) and mouse IgG TrueBlot (diluted 1 : 1000).

2.7. Immunohistochemistry

Rat hearts embedded in Tissue-Tek (Sakura) were frozen in liquid nitrogen, and cryosections (20 μm) were prepared. Cryosections were placed on glass slides, fixed in 4% paraformaldehyde/PBS for 10 min, permeabilized with 0.2% TritonX-100/PBS for 20 min, and blocked with 5% goat serum in PBS for 1 h to reduce nonspecific binding. Sections were incubated with primary antibodies overnight at 4°C, washed thrice in PBS, followed by incubation with the appropriate secondary antibodies. Stained sections were washed three times in PBS and mounted in Mowiol. The following antibodies were used: polyclonal rabbit anti-Cav1.2-ATTO 488 (Alomone Labs), polyclonal rabbit anti-NHERF1 (Cell Signaling), polyclonal rabbit anti-MAGI-3 (Abcam), monoclonal mouse anti-PMCA 5F10 (Sigma) and Alexa Fluor 488 goat anti-mouse, and Alexa Fluor 594 goat anti-rabbit (Invitrogen). As controls, the secondary Alexa Fluor-labelled antibodies were used alone. Confocal micrographs were obtained with an Eclipse E600 Nikon microscope using a C1 confocal scanning head and a 60-fold oil immersion objective.

2.8. Tricine-SDS-PAGE

To separate low molecular weight proteins, we used the following reagents: anode buffer (0.2 M Tris, pH 8.0), cathode buffer (0.1 M Tris, 0.1 M Tricine (Sigma Aldrich), 0.1% SDS, and pH 8.25), gel buffer (3.0 M Tris, 0.3% SDS, and pH 8.45), separating gel monomer 16,5 : 1 (49.5% T 6% C) and stacking gel monomer 33 : 1 (49.5% T 3% C), whereas T denotes the total percentage of acrylamide and bisacrylamide (Roth) and C the percentage of the crosslinker relative to the total concentration T [36]. Separating gel solution (16.5% T 6% C) was first prepared by mixing 10 mL separating gel monomer, 10 mL gel buffer, and 3.2 mL glycerol (Merck). For the stacking gel (4% T 3% C), we used 1 mL stacking gel monomer, 3.1 mL gel buffer, and 8.4 mL H2O. We then added 100 μL APS (Sigma Aldrich) and 10 μL TEMED (Sigma Aldrich) to both gel mixtures. After addition of cathode and anode buffer, electrophoresis was performed at 4°C at 30 V and 200 mA. After entering the stacking gel, the conditions were set to 90 V and 300 mA for 5 h. Gels were subsequently stained with Coomassie Brilliant Blue G250 (Merck).

2.9. Patch-Clamp Recordings

For patch-clamp experiments, HEK 293 cells were transiently transfected with either GFP alone or nNOS and GFP. Transfections were carried out as described above. Cells were analyzed 36 h after transfection. Membrane currents were measured in the whole-cell configuration of the patch-clamp technique. During the recordings, transfected cells were superfused by a bath solution containing NaCl 82 mM, TEA-Cl 20 mM, BaCl2 30 mM, CsCl 5.4 mM, MgCl2 1 mM, EGTA 0.1 mM, glucose 10 mM, HEPES 5 mM, pH 7.4 adjusted with NaOH; 302.9 mOsm. The perfusion chamber was mounted onto a stage of an inverted fluorescence microscope. Transfected cells were selected by GFP fluorescence. For whole-cell recording, patch pipettes of 3–5 MΩ were made from borosilicate tubes using a DMZ-Universal Puller (Zeitz). Pipettes were filled with a pipette solution containing CsCl 102 mM, TEA-Cl 10 mM, EGTA 10 mM, MgCl2 1 mM, Na2ATP 3 mM, HEPES 5 mM, pH 7.4, adjusted with CsOH; 248 mOsm. Membrane currents were recorded using an EPC-10 computer-controlled patch-clamp amplifier in conjunction with the software TIDA for data acquisition and analysis. The access resistance was compensated for values lower than 10 MΩ. For analysis of voltage-dependent activation, steady-state currents were plotted against the membrane potentials of the electrical stimulation. Plots of each individual cell were fitted using the Boltzmann equation. Statistical significance was tested using one-way analysis of variance (ANOVA). All data were given as mean ± SEM. n = number of independent experiments; * = statistical significance with . Mean values of data obtained from Boltzmann fits were calculated for each individual cell.

3. Results

3.1. Expression of PDZ Array Ligands and PDZ Domain Arrays I-IV

After comparison of C-terminal ends of mammalian Cav1.2 and PMCA4b (Figure 1(a)), the nucleotide sequences coding for the C-termini were cloned into pEXP bacterial expression vectors (Figure 1(b)). Verification of expression and size of His-tagged recombinant proteins in bacteria via Tricine gel analysis confirmed high expression levels and expected sizes of proteins, that is, for the pEXP read-through, pEXP-Cav1.2 and pEXP-PMCA4b, 9 kDa, 8.47 kDa, and 8.97 kDa, as calculated, respectively (Figure 1(c)). Probing the PDZ Domain Arrays with these bacterial lysates and subsequent detection of interactions with anti-6xHis antibodies revealed a series of positive spots on all PDZ arrays tested. As an example, typical results are displayed in Figure 2. Figure 2(a) gives an overview of the arrangement of the PDZ Domain Array III, and the corresponding results for Cav1.2 and PMCA4b are shown in Figures 2(b) and 2(c), respectively. Subsequent ImageJ quantifications of all four PDZ arrays tested are listed in the Supplemental Tables 1–4 in Supplementary Material available online at http://dx.doi.org/10.1155/2013/265182. The exemplary PDZ Domain Array III was spotted with PDZ domains of scaffolding proteins, especially MAGUKs. ImageJ quantification (Supplemental Table 1) revealed that both ligands interacted strongly with PDZ domains of MAGI-1, MAGI-2, and MAGI-3, and PDZ domains of SCRIB1, and TIP1.

Incubation of the PDZ Domain Array I, on which mainly PDZ domains of synaptic proteins were spotted (overview in Supplemental Figure 1(A)), with the C-terminal PDZ ligands of Cav1.2 and PMCA4b, revealed a panel of additional positive PDZ spots, representing possible interaction partners of the Cav1.2 and the PMCA4b (Supplemental Figures and ). ImageJ analysis of these signal intensities is listed in Supplemental Table 2. In this case, the Cav1.2 and PMCA4b C-termini interacted strongly with the PDZ domains of Mint-2-D1, OMP25, and Dlg-D1, and weakly with CASK-PDZ. In addition to these bindings, a promiscuous binding of the Cav1.2 C-terminus to HtrA2, hCLIM1, hPTP1E-D1, RIL, and ZO-2-D3 was detected. Besides the binding of the PMCA4b C-terminus to the PDZ domains of Mint-2-D1, OMP25, Dlg-D1, and CASK, the interaction with the Dlg PDZ domain 2 was very prominent and the signal strength was far above that of the positive controls.

The PDZ Domain Array II is arranged as shown in Supplemental Figure 2(a) and includes some tight junction proteins, sodium/hydrogen exchanger proteins, and other PDZ domains. Strong interactions for both PDZ ligands were observed for ZO-1-D1 and MAST-205 and for the PDZ positive controls (SAP-102). Cav1.2 C-terminus also had a high affinity for 4 other PDZ domains: ZO-1-D2, NHERF1-D1, NHERF2-D1/D2, some KIAA proteins, and nNOS (Supplemental Figures 2(B) and , analyses of signal intensities summarized in Supplemental Table 3).

Probing the PDZ Domain Array IV, mainly consisting of an assortment of scaffolding proteins, MAGUKs, Lin-7 proteins, nucleotide exchange factors, and syntrophins (Supplemental Figure ), revealed strong interaction of the PMCA4b C-terminus with PDZ domains of MUPP1, Dlg2 (Chapsyn-110), Dlg3 (SAP-102), LIN7A, LIN7B, LIN7C, SNA1, and SNB1. Both PDZ ligands interact with the PDZ domains of GEF11, GEF12, and SHK1, and the Cav1.2 C-terminus binds to that of PIST (results in Supplemental Figures 3(B) and and initial quantification of dot intensities summarized in Supplemental Table 4).

For further analyses, some of already established and some of newly identified interaction partners were selected (listed in Table 1), and the interactions were verified in coimmunoprecipitation and colocalization experiments.

3.2. Coimmunoprecipitations of 1.2 with Putative Interaction Partners

The interaction of Cav1.2 with different members of MAGUKs (CASK, MAGI-3, and ZO-1) and the proteins NHERF1 and MAST-205 was verified by coimmunoprecipitations. CASK, a 112 kDa protein, is expressed at neuronal synapses, where it interacts with neurexin, and in renal epithelial cells [37]. Therefore, we examined the putative interaction between Cav1.2 and CASK in lysates of HEK 293 cells, which were transfected with Cav1.2 . Coimmunoprecipitation of CASK with an anti-Cav1.2 antibody was further indicative for an interaction of the proteins (Figure 3(a)).

NHERF1, also known as Ezrin binding protein 50, is a 55 kDa phosphoprotein containing two PDZ domains [38]. To test for interaction of full-length proteins, HEK 293 cells, stably expressing the and the subunits of Cav1.2, were subsequently transfected with NHERF1 and coimmunoprecipitations using lysates of these cells were performed. NHERF1 was also coprecipitated with the Cav1.2-specific antibody in transfected cells (Figure 3(b)).

MAGI-3 (160 kDa) is predominantly expressed in brain but also in other organs [3941]. Similar to CASK and NHERF1, coimmunoprecipitations revealed interaction of Cav1.2α and MAGI-3 in mouse brain lysates (Figure 3(c)).

The tight junction protein ZO-1 is expressed in epithelial and, to a lesser extent, also in endothelial cells [42, 43]. Therefore, we searched for an interaction between Cav1.2 and ZO-1 in Cav1.2-transfected endothelial ECV cells. As depicted in Figure 3(d), the prominent band suggested a robust interaction of Cav1.2 and ZO-1 in ECV cells.

The serine/threonine kinase (Ser/Thr kinase) MAST-205 is expressed in testis, brain, and kidney tissues [44, 45]. To test for protein interactions between Cav1.2 and MAST-205, we used the HA-tag constructs pRK5-kinase-MAST-205 (36 kDa) and pRK5-kinase-PDZ-MAST-205 (77 kDa), precipitated with the Cav1.2 specific antibody, and detected the signal with a HA-specific antibody. In this case, the kinase domain alone (third lane) and also the PDZ-containing construct were coprecipitated (Figure 3(e)).

3.3. Coimmunoprecipitations of PMCA4b with Putative Interaction Partners

To confirm our data from the PDZ arrays, we tested for interaction between PMCA4b and CASK by coimmunoprecipitations from kidney and brain lysates, and transfected HEK 293 cells (Figure 4(a)). The results confirmed interaction of PMCA4b and CASK, as previously demonstrated [23]. Binding of the proteins PMCA4b and ZO-1 was demonstrated in extracts from various sources. The PMCA-specific antibody coprecipitated the 220 kDa protein ZO-1 in all cell lysates tested (Figure 4(b)). In both immunoblots, the positive controls were inputs of PMCA4b-transfected HEK 293 cells and the negative controls were the positive controls incubated with A/G-agarose and an irrelevant antibody (anti- ).

3.4. Colocalization of Cav1.2 and NHERF1, Cav1.2 and MAGI-3, and PMCA4b and MAGI-3 in Rat Cardiac Myocytes

Confocal laser scanning microscopy studies of rat heart sections revealed localization of Cav1.2, NHERF1, and MAGI-3 at intercalated discs (Figures 5(a)5(f)). Cav1.2 and MAGI-3 were additionally expressed at the transverse tubules (Figures 5(d)5(f)). As shown in Figures 5(g)5(i), PMCA4b and MAGI-3 were partially but not completely colocalized at the plasma membrane and the transverse tubules of rat cardiac myocytes.

3.5. Confirmation of Interaction of Cav1.2 with the PDZ Domain Containing Protein nNOS by GST Pull-Downs and Coimmunoprecipitation

It has been previously reported that the PDZ domain of nNOS interacts with the C-terminal end of PMCA4b [29]. The Domain Array II (Supplemental Figure 2) indicated that the PDZ domain of nNOS may also interact with Cav1.2 . To assess this interaction in more detail, we used affinity-purified GST fusion proteins containing the C-terminus of Cav1.2 (encoded by the plasmid pGEX-4T-3-LTCC), and the PDZ domain of nNOS (plasmid pGEX-4T-1-nNOS-PDZ). We performed pull-down assays from mouse brain and aorta extracts, and HEK 293 cell lysates, either nontransfected, or transfected with pcDNA3-Cav1.2 or pcDNA3-nNOS. As shown in Figure 6(b), the resulting GST pull-down clearly demonstrated binding between the C-terminus of Cav1.2 and the nNOS. This interaction was ascertained in reverse, confirming the interaction of nNOS and Cav1.2 (Figure 6(a)). Additionally, using an independent assay, we could prove this protein-protein interaction by conventional coimmunoprecipitations. In this assay, we cotransfected HEK 293 cells with the α and β subunits of Cav1.2 and nNOS. Subsequent precipitation with Cav1.2 -specific antibodies pulled down nNOS as well (Figure 6(c)).

3.6. Electrophysiological Properties

To examine the electrophysiological properties with Cav1.2 subunits, currents were measured in the whole-cell configuration of the patch-clamp technique. To activate voltage-dependent channels, membrane potentials of cells were clamped at a holding potential of −70 mV. From this holding potential the cells were depolarized stepwise in 9 voltage steps of +10 mV increment and a duration of 50 ms to depolarize the cell (Figure 7(a)). HEK cells, which stably express the Cav1.2 subunit, showed voltage-dependent inwardly directed currents with a fast time-dependent activation without inactivation (Figure 7(b)). The kinetic behavior of the Cav1.2 currents did not change when the cells were additionally transfected with pcDNA3-nNOS (Figure 7(c)). No differences in the overall activity of the currents were detected; Cav1.2 currents under control conditions but in the presence of nNOS showed no statistically different current densities (Figure 7(d)). However, the presence of nNOS led to a change in the voltage dependence of Cav1.2 currents (Figure 7(e)). To analyze the voltage dependence, normalized currents were plotted against their corresponding voltages of electrical stimulation, and the curve was fitted by the Boltzmann function to calculate basic parameters of voltage dependence, such as potential of half-maximal activation ( ), and the steepness of the curve ( ) (Figures 7(f)7(i)). nNOS induced a shift of the voltage-dependent activation and the potential of maximal current amplitude towards more positive voltages. Statistical analysis of parameters of voltage dependence of wild-type Cav1.2 currents showed that the activation threshold of the currents was not changed (Figure 7(f)) but the potential of half-maximal activation was shifted from −3 to +1.5 mV (Figure 7(g)), which was due to a shift of the slope of the Boltzmann fitted curve (Figure 7(h)) and not due to a shift of the activation threshold. This shift in the voltage dependence resulted in a different potential of maximal current amplitude (Figure 7(i)).

4. Discussion

The voltage-gated L-type calcium channel, Cav1.2, and the plasma membrane calcium ATPase, PMCA4b, play major roles in excitable and nonexcitable cells. Cav1.2 regulates the calcium entry into cells upon depolarization, while PMCA4b controls cellular calcium homeostasis by calcium extrusion. Both are important functional proteins in the heart and brain, but the specific tasks and the precise mechanisms are still investigated. The present studies were initiated to understand the regulatory consequence and the physiological background of the interactions from the C-terminal ligands Cav1.2 and PMCA4b with PDZ domain containing proteins. Using three independent assays (PDZ Domain Array, GST pull-down, and immunoprecipitation) and colocalization studies, we could show the interaction of a multiplicity of PDZ domain containing proteins and their ligands, Cav1.2 and PMCA4b.

PDZ domain proteins regulate the traffic and targeting of proteins, to assembly of signaling complexes and networks designed for efficient and specific signal transduction [6]. Presently, some of the described interaction partners of PMCA4b belong to the family of MAGUKs [18, 23, 46] but, in addition, nNOS and NHERF2 have been identified as interacting partners of PMCA C-termini [28, 29]. The C-terminal end of the PMCA splice variant 4b (ETSV) differs from other b variants [16, 17, 28], suggesting that the C-termini determine the specificity of interactions with other proteins. We identified new PDZ protein interaction partners of PMCA4b, whereby ZO-1, MAGI-1-3, Mint-2, and MAST-205 are of primary importance. For Cav1.2, we detected the same combination of proteins, with the addition of CASK, NHERF1, NHERF2 and nNOS.

Zonula occludens proteins are regulators of tight junction (TJ) assembly, and recent investigations have shown that these proteins also promote adherens junction (AJ) assembly [47]. The protein binding between the PDZ domain 1 of ZO-1 and the C-terminus of PMCA4b is insofar likely because both proteins are located at the membrane (Figure 4(b), Supplemental Figure ). A similar interaction is observed between the C-terminus of Cav1.2 and the PDZ domains 1 and 2 of ZO-1 (Figure 3(d) and Supplemental Figure 2(B)). We suggest that this protein interaction is probably important for the regulation of calcium ions and cytoskeletal dynamics at cell junctions and the plasma membrane.

In our experiments we detected that the three MAGI proteins (MAGI-1 domain 3; MAGI-2 and MAGI-3 domain 6) bind to the C-terminus of PMCA4b and Cav1.2 (Figures 2(b)-2(c) and Figure 3(c)). MAGI-1 and MAGI-3 are widely expressed in tissues like brain, heart, lung, and colon, but tend to localize to tight junctions of epithelial cells [48, 49]. MAGI-2 is exclusively widespread in neuronal tissue [50]. The group of Hall [51, 52] reported that the β1 adrenergic receptor (β1AR) binds MAGI-2 and MAGI-3. MAGI-2 enhances the receptor’s association with β-Catenin and its internalization, while MAGI-3 inhibits Gi-mediated ERK activation by β1AR. MAGI-2 and MAGI-3 also bind to the tumor suppressor gene product of PTEN (a tumor suppressor phosphatase). These MAGI proteins support PTEN suppression of Akt, which is involved in apoptosis suppression and growth induction [41, 53]. However, MAGI-3 is localized to diverse cellular compartments including the nucleus, cytoplasm, and junctional complexes at the cell surface [54], making it a central modulator of its function as scaffold protein. It is interesting that both of our investigated ligands, PMCA4b and Cav1.2, interact with all three MAGI proteins since the scaffolds are components of signaling complexes implicated in processes that require calcium. PMCA4b and Cav1.2 may play key roles in the arrangement of calcium-dependent AJs, and thus be involved in regulation of cell growth, cell morphology, and cell differentiation.

The Mint protein family (muc18-1-interacting protein) has three members, Mint-1, Mint-2, and Mint-3 [55, 56]. The Mint family plays a role in the arrangement of multiprotein complexes, and their ability to control the signaling and trafficking of membrane proteins [57]. Mints bind to Munc-18, a protein necessary for synaptic vesicle exocytosis, and to CASK, which is involved in targeting and localization of synaptic membrane proteins [55, 5863]. The presence of PDZ domains in Mints indicates a potential involvement of these proteins in connecting synaptic vesicles to the sites of synaptic intercellular junctions [24]. The multiprotein complex between our investigated ligands and Mint proteins could play a role in the exocytosis of synaptic vesicles, as the process requires a trigger and the resultant release of neurotransmitters is a -dependent reaction (Supplemental data).

MAST-205 (microtubule-associated serine/threonine kinase) is highly expressed in testis [44] and in kidney, adrenal glands, hindbrain, small intestine, and colon tissues [45]. This protein possesses a Ser/Thr kinase and one PDZ domain. Few protein interactions with MAST-205 have been identified. For example, the PDZ domain of MAST-205 additionally binds to PTEN, which regulates the cell growth and apoptosis [64]. The phosphorylation of PTEN by the kinase domain of MAST-205 suggests that PTEN could be a physiological substrate. The group of Yun demonstrated that MAST-205 modulates the transport activity of Na+/H+ exchanger (NHE3) in the renal proximal tubule, and this regulation was dependent on the presence of the kinase motif in MAST-205 [45]. Our studies suggest that the C-terminal tail of PMCA4b and Cav1.2 may act as component for specific and efficient PDZ domain recognition, which could be important in the control of PMCA4b and Cav1.2 protein phosphorylation, stability, and function (Figure 3(e), Supplemental Figures 2(B)– ).

We identified the PDZ domain containing protein CASK as a functional interaction partner of Cav1.2 (Supplemental Figures 1(B) and ). The MAGUK protein CASK consists of a -calmodulin kinase, a PDZ domain, an SH3 domain, and a guanylate kinase domain. It is mainly expressed at the neuronal presynaptic membrane, interacting with neuroligin-associated neurexin [6567], and additionally expressed in epithelial cells [37]. Mutation or deletion of CASK results in unusual synaptic function and perinatal death in mice [65], verifying its importance for brain development and function. CASK controls synapse formation and synaptic strength, and mutation or deletion in the gene leads to mental retardation [68]. All these studies indicate that Cav1.2 and PMCA4b in conjunction with CASK may play vital roles in the targeting of protein complexes in brain and epithelial cells, and in the modulation of synaptic transmission.

Other interesting interaction partners are the sodium-hydrogen exchanger regulatory factors, NHERF1 (also called EBP50) and NHERF2 (called E3KARP). In coexistence, they possess overlapping function as regulators of transmembrane receptors, transporters, and other proteins localized at or near the plasma membrane. The ERM (Ezrin, Radixin, Moesin, and Merlin) family of membrane cytoskeletal adapters is a crucial cellular target of NHERF [69, 70]. To regulate NHE3 signaling with cAMP, NHERF1 (or NHERF2), Ezrin, and protein kinase A form a multiprotein signal complex connecting NHE3 to the actin cytoskeleton. This complex is proposed to facilitate the phosphorylation and downregulation of NHE3 [7072], playing a crucial role in the proximal tubule, because H+ is secreted into the lumen by NHE3, essentially maintaining the acid base balance of the kidney. Another important aspect is the relationship between NHERF and CFTR (cystic fibrosis transmembrane regulator). The interaction between CFTR and NHERF may explain CFTRs ability to regulate other transport proteins, including the epithelial sodium channel, the renal outer medullary potassium channel, and NHE3 [73, 74]. Demarco et al. 2002 [28] suggested that the PDZ domains of NHERF1/2 bind the D-(S/T)-X-L motif (X represents any amino acid) at C-termini. Therefore, with the exception of PMCA4b, which has an ETSV motif, PMCAs 1b–3b (motif ETSL) interact with NHERF1/2. We confirmed these results with the PDZ Domain Array (Supplemental Figure ). Additionally, we identified a new interaction between NHERF1/2 with Cav1.2 (motif VSNL) in Supplemental Figure 2(B). Our findings and previous studies from other groups emphasize the importance of a terminal leucine residue for high affinity peptide interaction with NHERF [28, 75, 76]. The complex of Cav1.2 and NHERF1/2 may provide an indirect link between the channel and the actin cytoskeletal network, especially to stabilize the channel along the membrane and to allow its regulation by coassembled cAMP-dependent protein kinases. The PDZ domain 1 of NHERF1 is associated with SOCs (store operated calcium channel), Trp4, Trp5, and the phospholipases Cβ1 and Cβ2 [77], suggesting that NHERF can link the functions of SOCs to PLCβ to organize calcium and phosphoinositide metabolism, and control cell metabolism and growth. Our new results suggest an involvement of NHERF1/2 in the regulation of transport as well.

The colocalization of Cav1.2 and NHERF1 at intercalated discs and of Cav1.2 and MAGI-3 at the intercalated discs and the transverse tubules is noteworthy (Figures 5(a)5(f)). The same applies for PMCA4b and MAGI-3 at transverse tubules (Figures 5(g)5(i)). The intercalated discs of cardiomyocytes are crucial for pulse transmission and cell structure stabilization. The transverse tubules are invaginations of the plasma membrane of muscle cells that facilitate faster transfer of a depolarisation from the plasma membrane to the core of the cell. PMCA4b and Cav1.2 are both expressed at the caveolae of the plasma membrane, as mentioned above.

nNOS plays a crucial role in cardiomyocytes [78], regulating excitation-contraction coupling [79, 80], β-adrenergic inotropic response [79], and the development of heart failure [81, 82]. nNOS has been localized to the sarcolemma and the sarcoplasmic reticulum (SR) [29, 83, 84]. We have established that the C-terminus of Cav1.2 interacts with the PDZ domain of nNOS by means of a PDZ Domain Array, IP, and GST pull-down (Figures 6(a)6(c), Supplemental Figure 2(B)). Burkard et al. presented data from nNOS overexpressing mice that showed the negative inotropic effect of myocardial nNOS, which can reduce currents of Cav1.2 in their model system [85]. Furthermore, Sears et al. postulate that nNOS regulates influx by means of negative feedback, because increases in excite nNOS synthesis of NO, which, on the other hand, inhibits influx [80]. We now confirmed these studies by measurements of   currents in stable HEK 293 cells, transfected with nNOS.

The cells, which express Cav1.2 subunits showed Ba2+ currents with properties of L-type channels. The activation threshold and voltage of half-maximal activation are corresponding to that, which is commonly known for the cardiac subtype of L-type channels that corresponds with the Cav1.2 subunit. The currents from the cells, which stably express Cav1.2 subunits showed no inactivation. This can be due to the fact that was used instead of , which normally leads to loss of inactivation. Furthermore, the presence of inactivation is dependent on the coexpressed β-subunit. Upon nNOS expression only the voltage-dependent activation was changed. In the presence of nNOS the voltage-dependent activation was shifted towards more positive potentials and a change of the steepness of the curve of voltage-dependent activation resulted in shift in the voltage of maximal current amplitude while the activation threshold was unchanged. Thus, in the presence of nNOS, Cav1.2 subunits can be activated by the same depolarizing voltage difference but due to the shift of the half-maximal activation the currents amplitudes were smaller. With this effect, the channel current is reduced, resulting in a smaller voltage-dependent increase. In cells which express Cav1.2 subunits -dependent function would be reduced. For example, smooth muscle cells expressing active nNOS would display smaller depolarization-induced contractions and would support NOS-dependent relaxation.

In summary, we showed that Cav1.2 interacts physically with nNOS-, MAST-205-, MAGI-3-, NHERF1-, and ZO-1-PDZ domains; and PMCA4b with ZO-1-, MAST-205-, and MAGI-3-PDZ domains, all demonstrated via different assays (PDZ array, GST pull-down, and IP). The partial colocalization of Cav1.2 and MAGI-3, Cav1.2 and NHERF1, and PMCA4b and MAGI-3 in rat cardiomyocytes indicates that an interaction of these proteins is highly possible. From our results, we conclude that Cav1.2 and PMCA4b bind promiscuously to a variety of PDZ domains. The physiological consequence of some of these interactions remains to be investigated.


PSD-95:Postsynaptic density protein 95
SAP 90:Synapse-associated protein 90
PSD-93 (Chapsyn-110): Channel-associated protein of synapse-110
SH3 domain: Src homology 3domain
RIL: Reversion-induced LIM protein
SCRIB1: Scribble domain 1
TIP1: Tax interaction protein 1
OMP25:Mitochondrial outer membrane protein 25
MUPP:Multiple PDZ domain protein
hCLIM1: Human 36 kDa carboxyl terminal LIM domain protein
HtrA2: High temperature requirement protein A2
hPTP1E-D1: Protein tyrosine phosphatase 1E-domain 1.


The authors thank Dr. Andrea Welling (Institute of Pharmacology and Toxicology, TU Munich) for providing the stably transfected HEK 293 cells ( and β2a subunit of Cav1.2). The authors greatly appreciate the gift of the plasmid constructs, pGEX-4T-1-nNOS-PDZ, and pcDNA3-nNOS from Dr. Bredt (University of California, San Francisco, USA) and Rafael Pulido (Centro de Investigacion Principe Felipe, Valencia, Spain) for pRK5-kinase-MAST-205 and pRK5-kinase-PDZ-MAST-205. The work was supported by the IZKF Wuerzburg, TP E-33 to KS.

Supplementary Materials

Supplemental Figures 1–3 (a): Schematic representations of PDZ Domain array I, II, and IV organisation. Supplemental Figures 1–3 (b): PDZ Domain Arrays I, II, and IV showing interaction of the Cav1.2 C-terminus with spotted PDZ domains of proteins represented in corresponding Figures 1–3 (a). Supplemental Figures 1–3 (c): PDZ Domain Arrays I, II, and IV showing interaction of the PMCA 4b C-terminus with spotted PDZ domains of proteins represented in corresponding Figures 1–3 (a). In the Supplemental Tables 1–4, the abbreviations, full names and GeneBank accesion numbers of corresponding proteins, and mean grey values of scanned X ray exposures are given. Supplemental Table 5 provides measured values of electrophysiological properties.

  1. Supplementary Material