Identification of a New Co-factor, MOG1, Required for the Full Function of Cardiac Sodium Channel Nav1.5*

The cardiac sodium channel Nav1.5 is essential for the physiological function of the heart and contributes to lethal cardiac arrhythmias and sudden death when mutated. Here, we report that MOG1, a small protein that is highly conserved from yeast to humans, is a central component of the channel complex and modulates the physiological function of Nav1.5. The yeast two-hybrid screen identified MOG1 as a new protein that interacts with the cytoplasmic loop II (between transmembrane domains DII and DIII) of Nav1.5. The interaction was further demonstrated by both in vitro glutathione S-transferase pull-down and in vivo co-immunoprecipitation assays in both HEK293 cells with co-expression of MOG1 and Nav1.5 and native cardiac cells. Co-expression of MOG1 with Nav1.5 in HEK293 cells increased sodium current densities. In neonatal myocytes, overexpression of MOG1 increased current densities nearly 2-fold. Western blot analysis revealed that MOG1 increased cell surface expression of Nav1.5, which may be the underlying mechanism by which MOG1 increased sodium current densities. Immunostaining revealed that in the heart, MOG1 was expressed in both atrial and ventricular tissues with predominant localization at the intercalated discs. In cardiomyocytes, MOG1 is mostly localized in the cell membrane and co-localized with Nav1.5. These results indicate that MOG1 is a critical regulator of sodium channel function in the heart and reveal a new cellular function for MOG1. This study further demonstrates the functional diversity of Nav1.5-binding proteins, which serve important functions for Nav1.5 under different cellular conditions.

The cardiac sodium channel Na v 1.5 is essential for the physiological function of the heart and contributes to lethal cardiac arrhythmias and sudden death when mutated. Here, we report that MOG1, a small protein that is highly conserved from yeast to humans, is a central component of the channel complex and modulates the physiological function of Na v 1. 5 MOG1 and Na v 1.5 and native cardiac cells. Co-expression of MOG1 with Na v 1.5 in HEK293 cells increased sodium current densities. In neonatal myocytes, overexpression of MOG1 increased current densities nearly 2-fold. Western blot analysis revealed that MOG1 increased cell surface expression of Na v 1.5, which may be the underlying mechanism by which MOG1 increased sodium current densities. Immunostaining revealed that in the heart, MOG1 was expressed in both atrial and ventricular tissues with predominant localization at the intercalated discs. In cardiomyocytes, MOG1 is mostly localized in the cell membrane and co-localized with Na v 1.5. These results indicate that MOG1 is a critical regulator of sodium channel function in the heart and reveal a new cellular function for MOG1. This study further demonstrates the functional diversity of Na v 1.5-binding proteins, which serve important functions for Na v 1.5 under different cellular conditions.

. The yeast two-hybrid screen identified MOG1 as a new protein that interacts with the cytoplasmic loop II (between transmembrane domains DII and DIII) of Na v 1.5. The interaction was further demonstrated by both in vitro glutathione S-transferase pull-down and in vivo co-immunoprecipitation assays in both HEK293 cells with co-expression of
The SCN5A gene on chromosome 3p21 encodes Na v 1.5, the ␣-subunit of the cardiac sodium channel, which plays an important role in generating the cardiac action potential and mediating the conduction of cardiac electrical impulses (1,2). Numerous mutations in the SCN5A gene have been identified in long QT syndrome, Brugada syndrome, idiopathic ventricular fibrillation, cardiac conduction defects, and dilated cardiomyopathy associated with atrial fibrillation (3)(4)(5)(6)(7). Previous studies also suggest that alterations in the cardiac sodium channel function may be linked to two other major cardiac diseases, heart failure and atrial fibrillation. The expression of Na v 1.5 protein expression was reduced about 30% in a dog model of heart failure (8), whereas in a dog model of chronic atrial fibrillation, the sodium current density/function was decreased in atrial cells (9).
Due to its critical importance in cardiac physiology and human disease, studies to define the regulatory proteins and other components of the multiprotein Na v 1.5 complex have been of intense interest. The major component of the sodium channel complex is the pore-forming ␣-subunit Na v 1.5, which consists of four homologous domains (DI, DII, DIII, and DIV) with each domain containing six transmembrane segments (S1-S6) (1,2). The channel complex contains other subunits, for which there are at least four ␤-subunits identified thus far. Co-expression of the ␤1-subunit with Na v 1.5 causes a small but significant acceleration in the recovery from inactivation as well as an increase in current density that may be due to an increased targeting efficiency of the mature channel to the cell membrane (10,11). The ␤3-subunit causes a depolarizing shift in steady-state inactivation and a slower recovery from inactivation than the ␤1-subunit (12). The ␤2and ␤4-subunits have little effect on the channel kinetics of Na v 1.5 (13,14).
In addition to ␤-subunits, accessory proteins have been identified for Na v 1.5 whose interactions have been shown to form a multiprotein complex (reviewed in Refs. 15 and 16). Ankyrin-G was shown to interact with Na v 1.5, and this interaction involved a 9-amino acid motif in the cytoplasmic loop II between DII and DIII, the result of which promoted the localization of the channels to cell membrane in cardiomyocytes. The C terminus of Na v 1.5 contains 244 amino acid residues that have been shown to have important interactions with proteins such as FHF1B (fibroblast growth factor-homologous factor 1B), calmodulin, Nedd4-like ubiquitin-protein ligase, dystrophin and syntrophin, Fyn, and PTPH1 (protein-tyrosine phosphatase) (15,17,18). In a recent study, 14-3-3 was found to interact with the cytoplasmic loop I between DI and DII, and its dimerization was needed for current regulation (19). Despite this apparent diversity in accessory proteins, the complete composition of the cardiac sodium channel complex remains poorly understood. It is reasonable to expect that many more proteins are involved in the dynamic networks of protein-protein interactions with Na v 1.5 and underscores the significance of multiprotein complexes that are critical for normal cardiac function.
MOG1 was initially identified as a suppressor that was able to rescue the temperature-sensitive defect of Saccharomyces cerevisiae Ran, a protein involved in nucleocytoplasmic transport, microtubule, and nuclear assembly, and the spatial and temporal organization of the eukaryotic cell (20,21). In vitro studies showed that MOG1 can bind to Ran-GTP (20) and release GTP (22), but its in vivo function is not clear. MOG1 has been shown to be a highly conserved protein from yeast to humans (23). Human MOG1 gene contains five exons and four introns and encodes a protein of 187 amino acids with a calculated molecular mass of 20 kDa (23). The highest expression of MOG1 was detected in the heart by Northern blot analysis (23), but the exact physiological function of MOG1 in the heart is unknown.
To identify new proteins associated with the cardiac sodium channel complex, we performed a yeast two-hybrid screen with separate intracellular domains of Na v 1.5 as baits. When the cytoplasmic loop II was used as the bait, a candidate Na v 1.5interacting protein, MOG1, was identified. We further demonstrated the interaction between MOG1 and Na v 1.5 by both in vitro GST 3 pull-down and in vivo co-immunoprecipitation assays. We examined the physiological role of MOG1 by coexpression of MOG1 and Na v 1.5 in HEK293 cells and neonatal cardiomyocytes. Our results indicate that MOG1 is a co-factor for Na v 1.5 and modulates the appropriate expression and function of Na v 1.5.
Two anti-MOG1 antibodies were developed by GeneMed Synthesis, Inc. The first MOG1 antibody (number 2738) was developed with a MOG1-specific peptide, C-QPPPDNRSSLG-PENL at the N-terminal section, and the immunogene for the second antibody (number 3350) was a peptide, COOH-NQQVAKDVTLHQALLRLPQYQTDL at the C-terminal section. The rabbit polyclonal anti-Na v 1.5 antibody was as described previously (26).
Yeast Two-hybrid Screen-Yeast two-hybrid analysis was performed with a premade MATCHMAKER human heart cDNA library constructed in S. cerevisiae host strain, Y187 (Clontech). pACT2-derived constructs generate fusion proteins with the GAL4 activation domain fused to a library of other proteins. The baits for library screening were five different segments of Na v 1.5 fused to the GAL4 DNA binding domain in the pAS2 vector. Since Na v 1.5 is a membrane protein, the yeast two-hybrid screen with the entire Na v 1.5 protein is unlikely to be fruitful, because the protein may not enter the yeast nuclei. Thus, we selected five cytoplasmic segments of Na v 1.5 as baits. These include the N-terminal domain (amino acids 1-123), cytoplasmic LI between DI and DII (Na v 1.5-LI; amino acids 437-711), cytoplasmic LII between DII and DIII (Na v 1.5-LII; amino acids 940 -1200), the inactivation gate between DIII and DIV (amino acids 1471-1523), and the C-terminal domain (amino acids 1773-2016). A bait plasmid was transfected into yeast strain Y187 with a library of human heart cDNAs fused to GAL4 activation domain. Positive colonies were identified as instructed by the manufacturer (Clontech). Approximately 10 7 primary transformants were screened for each of the baits. DNA was isolated from each positive clone and used to transform Escherichia coli HB101 (Leu Ϫ ) to isolate only pACT2 derivative plasmids. The cDNA insert from each positive pACT2 derivative clone was amplified by PCR and sequenced by the BigDye Terminator version 1.1 cycle sequencing kit and an ABI PRISM 3100 genetic analyzer. The DNA sequences were then characterized by Blast analysis against the NCBI data base to determine the identity of the potential Na v 1.5-interacting proteins.
The 35 S-labeled MOG1 protein was prepared using a TNT Quick Coupled Transcription/Translation system (Promega). Briefly, 1 g of plasmid DNA was mixed with TNT Quick Master mix and [ 35 S]methionine and incubated for 90 min at 30°C. The 35 S-labeled MOG1 protein was mixed with GST, GST-Na v 1.5-LI, or GST-Na v 1.5-LII immobilized on glutathione-Sepharose 4B beads in binding buffer (1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM EDTA, 20 mM Tris-HCl, 150 mM NaCl, 0.1% Triton X-100) and incubated for 2-3 h at 4°C. After binding, the beads were washed with binding buffer, and bound proteins were eluted with 1ϫ SDS loading buffer, separated on a 12% SDS-polyacrylamide gel, dried, and visualized by exposing to x-ray film at Ϫ80°C for 12 h.
Co-Immunoprecipitation (Co-IP) Analysis-HEK293/Na v 1.5 cells were maintained in Dulbecco's minimum essential medium supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen) and transfected with 10 g of pcDNA3.1C-MOG1 DNA for the expression of His-tagged MOG1 with Lipofectamine 2000 (Invitrogen). Transfection was carried out with 80% confluent cells in a 10-cm plate. Cells were harvested and lysed in TNEN buffer (50 mM Tris/HCl, pH 7.5, 150 mM NaCl, 2.0 mM EDTA, 1.0% Nonidet P-40, protease inhibitor mixture) 24 -48 h after transfection. 500 g of total cell extracts was mixed with a rabbit polyclonal anti-Na v 1.5 antibody (26) and incubated on a rotator for 3 h at 4°C. Protein A/G-Sepharose 4B beads (Sigma) were added, and incubation was continued for another 2 h. The bound proteins were eluted by boiling the samples for 5 min in 1ϫ SDS loading buffer, separated by SDS-PAGE, and transferred onto a nitrocellulose membrane. The membrane was probed with an anti-His antibody (Sigma) recognizing His-MOG1 fusion protein, and the protein signal was visualized by enhanced chemiluminescence according to the manufacturer's instructions (Amersham Biosciences). For reverse co-IP, a rabbit polyclonal anti-MOG1 antibody was used for immunoprecipitation, and the anti-Na v 1.5 antibody was used for Western blot analysis. Similar co-IP studies were performed to study the interaction between the cytoplasmic loop II of Na v 1.5 and MOG1 by co-expressing His-tagged Na v 1.5-LII and FLAGtagged MOG1 in HEK293 cells and with an anti-His antibody and an anti-FLAG antibody.
For co-immunoprecipitation analysis of MOG1 and Na v 1.5 in cardiac cells, total protein lysates were isolated from adult mouse hearts. Mice were sacrificed, and the hearts were excised, washed in Hanks' buffer, cut into pieces, and lysed in the lysis buffer (20 mM Tris-HCl, pH 7.4, 1 mM EDTA, 150 mM NaCl, 1% Nonidet P-40, mixture of protease inhibitors). The lysates were sonicated on ice (Sonicator 3000; Misonic Inc., Farmingdale, NY), incubated on ice for 30 min, and then microcentrifuged at 4°C for 10 min. The supernatant was collected, and its concentration was measured using the Bio-Rad Protein Assay dye. The lysates were precleared on protein A/G beads at 4°C for 40 min prior to co-IP studies. Protein A/G beads (40 l) were incubated with 2 g of the anti-Na v 1.5 (or anti-MOG1) (1:500 dilution in PBST) and 500 g of cell lysates with gentle rocking for 2 h at 4°C. After washing five times with washing buffer (phosphate-buffered saline, protease inhibitors, and 0.5% Triton X-100), the immunoprecipitates were subjected to SDS-PAGE and immunoblot analysis using the anti-MOG1 antibody (or anti-Na v 1.5). Cell lysates incubated with protein A/G beads alone were used as a negative control. Co-IP of lysates with normal IgG was also used as a negative control in each experiment. Studies were repeated at least three times.
Isolation of Mouse Cardiomyocytes-For isolation of neonatal mouse cardiomyocytes, 10 -15 mouse hearts were collected from 3-day-old CBA/B6 mouse neonates. The ventricles were excised, and myocytes were isolated using the neonatal rat/ mouse cardiomyocyte isolation kit from CELLUTRON Life Technology. The cells were plated on uncoated 100-mm plates to reduce the contamination of cardiac fibroblasts. Myocytes were then cultured in the NS medium (CELLUTRON Life Technology) supplemented with 10% fetal bovine serum. Isolation of adult cardiomyocytes was performed as previously described by us (27)(28)(29)(30).
Immunohistochemistry-Immunostaining was performed on adult cardiomyocytes or frozen heart sections (6 m) with the polyclonal antibodies anti-Na v 1.5 and anti-MOG1, as described (31,32).
Electrophysiological Analysis-HEK293/Na v 1.5 cells or neonatal cardiomyocytes were transfected with 1 g of pcDNA3.1C-MOG1 using Lipofectamine 2000 (Invitrogen), and electrophysiological recordings were performed as described previously (27,29,30). Vector pIRE-GFP DNA expressing green fluorescent GFP protein (0.25 g) was cotransfected together with pcDNA3.1C-MOG1 to serve as an indicator (24,25). Only GFP-positive cells were selected for recording sodium currents. Pipettes were fabricated from borosilicate glass (FHC, Inc.), and electrode resistance ranged from 2 to 3 megaohms when filled with pipette solution with the following composition: 20 mM NaCl, 130 mM CsCl, 10 mM HEPES, 10 mM EGTA, pH 7.2, with CsOH. Voltage command pulses were generated using the Multipatch 700B amplifier (Axon Instruments) under the control of a desktop computer with pCLAMP software (9.0; Axon Instruments). Currents were filtered at 5 kHz (Ϫ3 dB, 4-pole Bessel filter) following series resistance compensation. The holding potential for all pulse protocols was Ϫ100 mV, and experiments were performed at room temperature (22°C). For HEK293 cells, the composition of the bath solution was 70 mM NaCl, 80 mM CsCl, 5.4 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES, 10 mM glucose, pH 7.3, with CsOH. For neonatal cells, the NaCl concentration was reduced to 20 mM, and CsCl was increased to 120 mM for better voltage control. To reduce contaminating Ca 2ϩ and K ϩ (transient outward) currents, 2 mM CdCl 2 , and 2 mM 4-aminopyridine, respectively, were added to the bath.
Currents were elicited using 100-ms step pulses in 10-mV increments from a holding potential of Ϫ120 mV. Currents were filtered at 5 kHz with a 4-pole Bessel filter and sampled at 20 kHz. Patches contained several channels, so multiple overlapping openings were generally observed near the start of the depolarization, but isolated single channel currents could be measured at later times. Events were detected using the halfamplitude threshold-crossing method, following digital Gaussian filtering to a final 2 kHz frequency and spline interpolation. Single channel current amplitudes were measured from open events lasting Ͼ0.18 ms (2 times the dead time of the filter). The single-channel conductance was calculated by linear regression for currents from Ϫ80 to Ϫ20 mV.
To determine the voltage dependence of gating and the maximal conductance, peak currents at each voltage were converted to chord conductance, G ϭ I/(V Ϫ V R ), where V R is the extrapolated reversal potential from a linear fit to currents from Ϫ20 to 0 mV. In each experiment, G-V relations were fitted to a Boltzmann function, G ϭ G max /(1 ϩ e Ϫ(V ϪVs)/k ), where G max is the maximal conductance, V s is the voltage where G ϭ G max /2, and k is a slope factor. No significant differences in G max , V s , or k were observed between controls and MOG1 cells.
Western Blot Analysis-To determine the expression level of the Na v 1.5 protein in the membrane fraction, HEK293/Nav1.5 cells were transiently transfected with pcDNA3.1C-MOG1 or empty vector pcDNA3.1C as control. Transfected cells were cultured for 48 h and washed with cold phosphate-buffered saline. Cells were then collected and suspended in the lysis buffer (1% Triton X-100, 150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, a protease inhibitor mixture). Cell lysates were centrifuged at 14,000 rpm for 10 min at 4°C to remove cell debris, nuclei, and large particulates. The supernatant portion that contains both the extracted membrane proteins and cytosolic proteins was collected and further centrifuged at 40,000 rpm to separate the membrane fraction from the cytosolic fraction. The pellet from the second centrifugation step contains the membrane protein fraction, and the supernatant contains the cytosolic protein fraction. Equal amounts of protein extracts were resolved by 2% SDS containing protein loading buffer and separated on SDS-PAGE. Western blot analysis was performed as described previously (22). A rabbit polyclonal anti-Na v 1.5 was diluted in 1:500 in 0.3% bovine serum albumin in PBST and used in Western blot analysis. The signal was detected using enhanced chemiluminescence (ECL kit; Amersham Biosciences). An anti-KCNQ1 antibody (Santa Cruz Biotechnology) was used as a loading control at 1:500 dilution in PBST.
We also used Pierce Mem-PER mammalian membrane protein extraction kit (Pierce) for separating the membrane fraction from the cytosolic fraction through phase partitioning. Cells (5 ϫ 10 Ϫ6 ) were transfected, cultured, and collected as described before and lysed with 150 l of Mem-Per Reagent A containing a mild proprietary detergent by incubation at room temperature for 10 min. 450 l of the second diluted Mem-Per Reagent (2 parts Reagent C and 1 part Reagent B) was added to cell lysates to solubilize the membrane proteins. The mixture was incubated on ice for 30 min and centrifuged at 10,000 ϫ g for 3 min at 4°C. The supernatant was then incubated at 37°C for 10 min and then centrifuged at 10,000 ϫ g for 2 min to separate the membrane fraction (bottom layer) from the hydrophilic fraction (top layer). The membrane and cytosolic fractions were used for Western blot analysis with anti-Na v 1.5 and anti-KCNQ1 antibodies as described above, and similar results were obtained.
Statistical Analysis-Data are represented as mean values Ϯ S.E. Statistical analysis was performed using an analysis of variance and two-tailed Student's t test to compare means, and significance was set at the indicated p values.

MOG1 Was Identified as a Candidate Protein That Interacts
with Na v 1.5 by a Yeast Two-hybrid Screen-To identify proteins that interact with Na v 1.5, we carried out a yeast two-hybrid screen. We screened a human cardiac pretransformed MATCHMAKER cDNA library (Clontech, Inc.). Five cytoplasmic segments of Na v 1.5 were used as baits for the screening, and these include the N terminus, loop I (between DI and DII), loop II (between DII and DIII), loop III (between DIII and DIV), and C terminus. The Na v 1.5 C terminus-GAL4 DNA binding domain fusion protein alone activated transcription of reporter genes; thus, no further screening was performed with this bait. No positive clones were obtained with the N-terminal domain and cytoplasmic loop I baits. There were 223 positive clones obtained with the cytoplasmic loop II bait, and 29 positive clones with the loop III bait. Most clones were excluded as false positives, since they were derived from mitochondrial DNA or because the encoded peptides were out of frame with the GAL4 activation domain. One positive clone from the screening was found to encode the portion of MOG1 spanning amino acid residues 65-186 (GenBank TM accession number AF265206) and was independently identified twice from the library screening.
Interaction between MOG1 and Na v 1.5 as Shown by GST Pull-down-To further demonstrate the interaction between MOG1 and Na v 1.5, we carried out a GST pull-down assay. The cytoplasmic loop II of Na v 1.5 (amino acids 940 -1200) was fused to GST (GST-Na v 1.5-LII), expressed in E. coli, and purified. Radioactively labeled MOG1 protein was prepared by in vitro transcription followed by translation using [ 35 S]methionine, and this appeared as a single band of 28 kDa (Fig. 1A, lane  1). The GST-Na v 1.5-LII fusion protein successfully pulled down 35 S-MOG1 in the assay (Fig. 1A, lane 2), whereas two negative controls, GST alone (Fig. 1A, lane 4) and the GST-Na v 1.5 cytoplasmic loop I fusion protein (GST-Na v 1.5-LI; Fig.  1A, lane 3), failed to interact with 35 S-MOG1. These results suggest that MOG1 interacts directly with the cytoplasmic loop II of Na v 1.5.
Interaction between MOG1 and Na v 1.5 as Shown by Co-Immunoprecipitation-HEK293/Na v 1.5 was transfected with pcDNA3.1C-His-MOG1. Cell extracts were immunoprecipitated using an antibody against MOG1 (Fig. 2A, lane 3) or with control rabbit preserum IgG (Fig. 2A, lane 2). The bound proteins were then detected by Western blot analysis with an anti-Na v 1.5 antibody. The anti-MOG1 antibody, but not the control IgG, precipitated a 250-kDa Na v 1.5 protein. Reciprocal co-immunoprecipitation was also performed. The anti-Na v 1.5 antibody was used for immunoprecipitation, and the anti-MOG1 antibody was used for Western blot analysis. MOG1 protein was successfully precipitated by the anti-Na v 1.5 antibody but not by the rabbit preserum control (Fig. 2B).
Since MOG1 was pulled out from the yeast two-hybrid library using Na v 1.5-LII, we next determined whether the loop II interacts with MOG1 using the co-immunoprecipitation assay. HEK293 cells were transiently co-transfected with expression constructs for FLAG-tagged MOG1 (pCMV10-MOG1) and His-tagged Na v 1.5-LII (pcDNA3.1A-Na v 1.5-LII). Cell extracts were immunoprecipitated using a monoclonal antibody against His or with control IgG. The bound proteins were then detected by Western blot analysis with a monoclonal anti-FLAG antibody. The anti-His antibody, but not the control IgG, precipitated FLAG-tagged MOG1 (Fig. 2C). A reciprocal co-immunoprecipitation assay was performed using an anti-FLAG antibody for immunoprecipitation. Na v 1.5-LII fusion protein was successfully precipitated by the anti-FLAG antibody but not by control mouse serum IgG (Fig. 2D). These immunoprecipitation results illustrate that the association of MOG1 to Na v 1.5 is mediated by the cytoplasmic loop II of Na v 1.5.
In Vivo Interaction between MOG1 and Na v 1.5 in Cardiac Cells-The interaction between MOG1 and Na v 1.5 was further confirmed by co-immunoprecipitation assays in vivo in mouse cardiac cells. Total protein extracts were prepared from mouse hearts. Both MOG1 and Na v 1.5 proteins are abundantly expressed in cardiac cells and can be easily detected by Western blot analysis (Fig. 2, E (lanes 1) and F (lane 1)). Two protein bands were detected for MOG1, which may represent the two MOG1 isoforms derived from alternatively spliced transcripts as reported previously by Marfatia et al. (23). Mouse cardiac protein extracts were precipitated either with a polyclonal antibody against MOG1 (Fig. 2E, lane 2) or with control IgG (Fig.  2E, lane 3). The bound proteins were then detected by immunoblot analysis with an anti-Na v 1.5 antibody. The MOG1 antibody easily precipitated an Na v 1.5 protein (Fig. 2E, lane 2). Similar experiments revealed that the anti-Na v 1.5 antibody could precipitate MOG1 proteins from mouse cardiac cell extracts (Fig. 2E, lane 2). These results indicate that MOG1 interacts with Na v 1.5 in vivo in cardiac cells.
MOG1 Increases Sodium Current Density in a Mammalian Expression System-We wished to determine whether the interaction of MOG1 with Na v 1.5 modified channel function. For this purpose, we expressed MOG1 in HEK293/Na v 1.5 cells and measured the whole-cell sodium currents. As shown in Fig.  3, A and B, the sodium current density (expressed as peak current normalized to cell capacitance, pA/pF) across the range of test potentials was significantly increased in cells co-expressed with MOG1. The maximum current density normally measured at Ϫ30 mV for vector cells was shifted to Ϫ35 mV and increased by 61 pA/pF when co-expressed with MOG1. This is despite the fact that co-expression with MOG1 did not alter cell capacitance (16.6 Ϯ 4.1 pF (n ϭ 12) versus 18.6 Ϯ 1.9 pF (n ϭ 9), p ϭ not significant). Steady-state activation and inactivation gating properties were evaluated using the pulse protocols shown in the insets of Fig. 3C. The data for channel activation are the mean normalized conductance plotted against the test potential. MOG1 significantly shifted the voltage dependence of activation to more negative potentials, although the shift was very small, only by 4 mV (V1 ⁄ 2act ϭ Ϫ49.1 Ϯ 0.1 mV (n ϭ 5) versus Ϫ45.1 Ϯ 0.2 mV (n ϭ 9), p Ͻ 0.05). A two-pulse protocol was used to estimate the membrane potential dependence of inactivation. Cells were stepped to conditioning potentials for 500 ms as shown on the abscissa before depolarization to Ϫ20 mV (50-ms step), and peak current from the test potential was normalized to peak sodium current in the absence of a conditioning step. Cells expressing MOG1 showed no difference in the inactivation kinetics of Na v 1.5 ( Fig. 3C; V1 ⁄ 2inact ϭ Ϫ84.1 Ϯ 0.1 mV (n ϭ 5) versus Ϫ85.5 Ϯ 0.1 mV (n ϭ 7), p ϭ not significant). Similarly, recovery from inactivation was not changed ( Fig. 3D; t1 ⁄ 2 ( 1 ) ϭ 4.2 Ϯ 0.2 ms (n ϭ 5) versus 3.8 Ϯ 0.1 ms (n ϭ 6), p ϭ not significant), as this was assessed using a two-pulse protocol, and the fractional current (P2/P1) was plotted against interpulse duration between P1 and P2. In summary, these results show that MOG1 increases whole-cell Na v 1.5 currents.
MOG1 and Sodium Current Density in Neonatal Cardiomyocytes-To examine the important role of MOG1 in native cardiomyocytes, MOG1 was overexpressed in 3-day neonatal myocytes. Our results show that sodium current density was increased 2-fold compared with vector-transfected cells (Fig. 4). These results suggest that MOG1 may play a critical role in the physiological function of sodium channels in native cardiomyocytes.
Strong Expression of MOG1 at Intercalated Discs in Cardiomyocytes and Heart Tissues-To further corroborate the interaction of MOG1 to Na v 1.5, we studied the expression patterns of MOG1 in the heart. A peptide competition experiment was used to evaluate the specificity of the MOG1 antibody. The immunofluorescent signal was significantly eliminated when the MOG1 antibody (number 2728) was preabsorbed with the antigen peptide (Fig. 5, compare A  and B), suggesting that the MOG1 antibody has a high specificity. As shown in Fig. 5C, MOG1 is widely expressed in both atrial and ventricular muscles, and interestingly, this expression was highly localized in the intercalated discs. Identical results were obtained with the second, independent anti-MOG1 antibody (number 3350; data not shown), confirming the finding with antibody number 2728. Immunostaining with connexin 43 antibody made it possible to easily distinguish the AV node from atrial and ventricular tissues (Fig. 5D). Similar to connexin 43, expression of MOG1 protein was higher in atrial and ventricular tissues compared with AV nodal tissues.
Immunostaining studies were also performed in isolated mouse ventricular myocytes, and the results show that MOG1 expression was also particularly strong in the intercalated discs regions. Furthermore, MOG1 is mostly localized outside of the nucleus in cardiomyocytes. More importantly, the expression pattern of MOG1 was shown to overlap with that of Na v 1.5 (Fig. 5E).

MOG1 and Sodium Currents in HEK293/Na v 1.5 Cell-attached
Patches-To determine whether MOG1 has any effect on single channel conductance, sodium currents were recorded using the cellattached configuration of the patch clamp technique from HEK293/ Na v 1.5 cells transiently transfected with MOG1 or empty vector (control). No significant effect on current amplitudes was observed (1.0 Ϯ Ͼ0.3 nanosiemens with control vector (n ϭ 5) versus 0.8 Ϯ 0.4 nanosiemens with MOG1 (n ϭ 4)) (Fig. 6A). The single-channel slope conductances were not significantly changed either (6.0 Ϯ 0.4 picosiemens control, n ϭ 3 versus 6.1 Ϯ 0.8 picosiemens with MOG1, n ϭ 3) (Fig. 6B). These results suggest that MOG1 does not affect the conductance of single sodium channel.
MOG1 Increases Cell Surface Expression of Na v 1.5-To investigate the potential mechanism by which MOG1 increases  the sodium channel density, Western blot analysis was performed with the membrane fraction and cytoplasmic fraction of protein extracts isolated from HEK293/Na v 1.5 cells transiently transfected with MOG1 or empty vector (control). As shown in Fig. 7A, Na v 1.5 expression in the membrane fraction was increased with overexpression of MOG1 in comparison with the control. Accordingly, Na v 1.5 expression in the cytoplasmic fraction was reduced with overexpression of MOG1 (Fig. 7B). These results suggest that MOG1 increases cell surface expression of Na v 1.5, which is consistent with the finding that co-expression of MOG1 and Na v 1.5 increased sodium channel densities (Figs. 3 and 4).

DISCUSSION
Here we report the identification of a novel Na v 1.5-interacting protein, MOG1. We showed that MOG1 was an important co-factor for Na v 1.5 and could modulate the function of the cardiac sodium channel. Specifically, we demonstrated that MOG1 interacted with Na v 1.5 and played a critical role in regulation of sodium current densities. Using a yeast two-hybrid screen, we identified MOG1 as a candidate protein that interacted with the cytoplasmic loop II of Na v 1.5. Both in vitro GST pulldown and in vivo co-immunoprecipitation analyses further demonstrated this interaction (Fig. 2). Most importantly, co-immunoprecipitation studies demonstrated the interaction between endogenous MOG1 and Na v 1.5 in cardiac cells (Fig. 2).
The interaction between MOG1 and Na v 1.5 is important for the physiological function of Na v 1.5. Coexpression of MOG1 in HEK293/ Na v 1.5 cells increased whole-cell sodium current density (Fig. 3). Similarly, overexpression of MOG1 in neonatal cardiomyocytes also resulted in increased whole-cell sodium current density (Fig. 4). The amplitude of the sodium current in neonatal myocytes (240 pA/pF) was slightly higher than that in HEK293/Na v 1.5 cells (200 pA/pF) (Figs. 3 and 4). These results suggest that MOG1 can modulate the physiological function of Na v 1.5.
Increased sodium current densities in HEK293/Na v 1.5 cells caused by the overexpression of MOG1 suggest that there is either an increase in the number of sodium channels on the cell surface or an enhancement in single channel conductance. . Overexpression of MOG1 in HEK293/Na v 1.5 cells increased sodium current density. A, raw traces for sodium currents with (right) and without (left) overexpression of MOG1 that were elicited with the current protocol depicted in the inset. B, for activation, cells were held at Ϫ100 mV and depolarized in 5-mV increments. The current-voltage relationship for both cell groups is summarized with current amplitudes normalized to cell capacitance (pA/pF, abscissa). Both steady-state activation and inactivation were determined by fitting the peak currents with a Boltzmann distribution, I/I max ϭ 1/(1 ϩ e((ϪV Ϫ V1 ⁄2 )/s)), where I is the current at test potential V, I max is the maximum current, and V1 ⁄2 is the potential giving the half-maximum current (see "Results" for fitting parameters). C, steady-state activation (right) and inactivation curves (left). Steady-state activation was plotted over the indicated voltage range and expressed as the current at the test potential over the maximum current (I/I max , abscissa). A two-pulse protocol was used to estimate the membrane potential dependence of inactivation. Cells were stepped to conditioning potentials for 500 ms as shown on the abscissa before depolarization to Ϫ20 mV (50-ms step), and peak sodium current from the test potential was normalized to peak sodium current in the absence of a conditioning step. D, recovery from inactivation was assessed for both cell groups utilizing a two-pulse protocol, and the fractional current (P2/P1) was plotted against interpulse duration between P1 and P2. The fraction of channels that had recovered following various time intervals was calculated by dividing the peak current measured during a test pulse to Ϫ20 mV. The average data were fitted with a biexponential function, I/I max ϭ A 1x (1 Ϫ e(-t/ 1 )) ϩ A 2x (1 Ϫ e(Ϫt/ 2 )). In cells overexpressing MOG1, no effects in the inactivation kinetics were observed, but a hyperpolarizing 4-mV shift was detected in channel activation. All studies were repeated at least three times, and the same results were obtained (data not shown).
Kinetic analysis of the steady-state activation and inactivation kinetics of Na v 1.5 in HEK293/Na v 1.5 cells overexpressed with MOG1 revealed a small 4-mV hyperpolarizing shift in activation with little or no effects on channel inactivation. This, by itself, did not reveal the underlying basis for the increase (ϳ60 pF/pA) in the current density in HEK293/Na v 1.5 cells with overexpression of MOG1. Following examination of cell-attached patches from HEK293/Na v 1.5 cells overexpressed with MOG1, we further revealed no effect on single channel conductance (Fig. 6B). These results suggest that an enhancement in single channel conductance cannot account for the effect of MOG1 on Na v 1.5. The above analysis prompted us to examine the possibility that an increase in the number of sodium channels on the cell surface may be the underlying mechanism for the finding of increased sodium current densities by MOG1. This hypothesis was tested by Western blot analysis, which revealed an increase in Na v 1.5 expression in the plasma membrane (Fig. 7A), and a decrease in Na v 1.5 expression in the cytoplasm of HEK293/ Na v 1.5 cells overexpressed with MOG1 (Fig. 7B). These results suggest that the channel number on the cell surface was increased. Taken together, the available evidence supports a model in which MOG1 increases the number of sodium channel and/or availability on the cell surface, which then results in an increase in sodium current density.
MOG1 is highly conserved from yeast to humans, suggesting that it is an essential protein for cellular functions. However, the specific physiological function(s) of MOG1 is obscure.
MOG1 was shown to interact with Ran GTPase, a protein required for the trafficking of proteins and RNA in and out of the nucleus, and mediates the release of GTP from Ran in vitro (20,21). Thus, MOG1 was proposed to play a regulatory role in nuclear import and export by maintaining the Ran-GTP gradient from the nuclei to the cytoplasm (22). During the nucleocytoplasmic transport, MOG1 was proposed to shuttle between the cytoplasm and the nucleus (22). Indeed, in HEK293 cells, human MOG1 is localized throughout the cell (23). Due to its cytoplasmic localization, the function of MOG1 may not be restricted to the nucleus; on the contrary, it may function outside of the nucleus. Lu et al. showed that yeast MOG1 may MOG1 Modulates Function of Na v 1.5 MARCH 14, 2008 • VOLUME 283 • NUMBER 11 interact with an osmotic stress sensor Sln1p and regulates the SLN1-SKN7 signal transduction in yeast (34). It is interesting to note that Sln1p is a plasma membrane protein, specifically a two-transmembrane domain sensor of the high osmolarity glycerol response pathway (35). These results suggest that MOG1 can interact with a plasma membrane protein. Thus, our finding of the interaction between MOG1 and membrane protein Na v 1.5 is no surprise. Oki et al. (36) recently showed that yeast MOG1 can also interact with Cid13 (a poly(A) polymerase for suc22 mRNA encoding a subunit of ribonucleotide reductase) with a potential role in regulation of cell cycle S-M transition. In addition, genetic suppression studies in yeast implicated that MOG1 could be required for membrane localization of Opi3p, a phospholipid methyltransferase required for membrane formation, and might play a role in the Ssp1-mediated stress response pathway (36 (23). Increased cell surface localization of Na v 1.5 by MOG1 may be achieved through two potential mechanisms, increased transport of Na v 1.5 to cell surface or reduced turnover of Na v 1.5 when localized on plasma membrane. Trafficking of membrane proteins like Na v 1.5 is a highly regulated process (reviewed by Herfst et al. (37)). The first key step of the process is the transition from the ER to Golgi compartment, which is governed by the ER retention motifs and ER export signals. The cytoplasmic loop I of Na v 1.5 contains several ER retention motifs (the RXR motif) that are involved in mediating the increase of sodium currents by protein kinase A (38). There is a potential ER export signal (DXE) in the C terminus of Na v 1.5 that may regulate the exit of fully folded and assembled Na v 1.5 out of the ER (37). Selective Golgi export or transport of Na v 1.5 from the Golgi to plasma membrane is another key step for trafficking of Na v 1.5 (37). The trafficking of Na v 1.5 may be regulated by its interacting proteins. The PDZ domain located at the C terminus of Na v 1.5 can associate with dystrophin and syntrophin proteins (18). In dystrophin-deficient mice, expression of Na v 1.5 was reduced by 50%, and the sodium current from cardiomyocytes was reduced by 29% (18). Since PDZ domain-interacting proteins can affect the membrane protein trafficking (39), dystrophin and syntrophin proteins may regulate the appropriate expression of Na v 1.5 by controlling the transport of Na v 1.5 to plasma membrane. Co-expression of the ␤1 subunit with Na v 1.5 caused a small acceleration in the recovery from inactivation and an increased density of sodium currents (10,11). Further studies showed an increase of Na v 1.5 localization to plasma membrane by the ␤1 subunit (10, 11); however, it remains to be determined how ␤1 increases trafficking of Na v 1.5 to plasma membrane. The function of MOG1 is similar to that of the ␤1 subunit and may regulate the trafficking of Na v 1.5 to plasma membrane. Co-expression of MOG1 with Na v 1.5 caused a small but significant shift of the activation curve to a more neg-  ative potential and an increased density of sodium currents (Figs. 3 and 4). Western blot analysis showed increased expression of Na v 1.5 on plasma membrane (Fig. 7). As with ␤1, the molecular mechanism by which MOG1 increases targeting of Na v 1.5 on plasma membrane is not clear. Since studies from yeast suggest that MOG1 can interact with the nucleocytoplasmic transport machinery, a hypothesis that can be tested in the future is that MOG1 may interact with the transport complex involved in the transport of Na v 1.5 from the Golgi to plasma membrane or from the ER to Golgi. In addition to increased trafficking of Na v 1.5 to the plasma membrane, MOG1 may reduce the turnover of Na v 1.5 localized on the plasma membrane. Any change in the internalization and degradation of Na v 1.5 will affect the expression level of Na v 1.5 on the cell surface. The ubiquitin-protein ligase Nedd4 is another Na v 1.5-interacting protein that may mediate the ubiquitin-dependent internalization and subsequent degradation of Na v 1.5. Accordingly, co-expression of Nedd4 and Na v 1.5 led to reduction of the sodium current density by 40% (40) and accelerated internalization of Na v 1.5 (33). It is unlikely that MOG1 is involved in the ubiquitin-dependent internalization and subsequent degradation of Na v 1.5. On the other hand, the direct interaction between MOG1 and Na v 1.5 may stabilize the Na v 1.5 protein and reduce its degradation, which may explain the finding of increased cell surface expression of Na v 1.5 by MOG1.
A previous study with Northern blot analysis revealed that the heart was the tissue with the highest expression level of MOG1 (23), suggesting that MOG1 plays an important role in cardiac function. The results from the present study further define the role of MOG1 in the heart. We provide strong evidence that MOG1 regulates the proper expression and function of Na v 1.5 in cardiomyocytes. Further studies with immunostaining revealed that the expression of MOG1 was concentrated in the intercalated discs of both atrial and ventricular cells (Fig. 5C). Moreover, the expression pattern of MOG1 was similar to the pattern shown by connexin 43, the primary gap junctional protein in cardiomyocytes, whereby the expression intensities were localized to the intercalated disc regions (Fig.  5D). Since intercalated discs are essential for regulating electrical coupling between cells in the myocardium, co-localization of MOG1 and Na v 1.5 in these regions suggests that MOG1 may play an important role in establishing and regulating electrical connections between cardiomyocytes.
We found that in cardiomyocytes, MOG1 was mostly localized on cell membrane and outside of the nucleus (Fig. 5E). In HEK293 cells, it was reported that MOG1 was expressed throughout the cell, including the nucleus and cytoplasm (23); however, the study could not distinguish whether MOG1 was also localized on plasma membrane due to complication of strong signal of MOG1 in cytoplasm. The cell-specific subcellular localization of MOG1 in cardiomyocytes versus HEK293 cells is an interesting observation, but the molecular mechanism warrants further investigations. Cell-or tissue-specific proteins that interact with MOG1 may be a rational explanation for the observation. The interaction with Na v 1.5 or other cardiac specific membrane proteins is expected to recruit MOG1 to the plasma membrane.
Marfatia et al. (23) identified two alternatively spliced MOG1 transcripts, hMOG1a and hMOG1b. Consistent with this finding, our Western blot analysis of cardiac cell extracts revealed two isoforms of the MOG1 protein. Since the hMOG1a was the MOG1 isoform identified in our yeast two-hybrid screen, functional studies for MOG1 in this study were carried out with hMOG1a only. Searches of data bases, including the UCSC Genome Browser data base (available on the World Wide Web) revealed three alternatively spliced MOG1 transcripts. The functional significance of the other alternatively spliced MOG1 transcript(s) or protein isoform(s) remains to be established in the future.
Na v 1.5 mutations that affect its cell surface expression (trafficking) have been associated with Brugada syndrome and other types of lethal arrhythmias. Thus, genetic mutations in MOG1 may affect the expression and function of Na v 1.5, leading to similar cardiac diseases. On the other hand, identification of proteins like MOG1 that regulate expression and function of Na v 1.5 may serve as interesting targets for developing interventional options to manage lethal arrhythmias associated with Na v 1.5 mutations or abnormalities. Furthermore, the present findings provide new insights into the physiological role of MOG1 in vivo in mammalian cells, and future studies of MOG1 will probably offer more insights into its role and its relevance in myocardial function.