Regulation of Gene Transcription by Voltage-gated L-type Calcium Channel, Cav1.3*

Background: Cav1.3 Ca2+ channel is highly expressed in atria and pacemaking cells in the heart. Results: The C terminus of the Cav1.3 Ca2+ channel can translocate into the nucleus. Conclusion: The C terminus of Cav1.3 can function as a transcriptional regulator to regulate Ca2+-activated K+ channels. Significance: New insights into the cross-talk between ion channels may have broad therapeutic ramifications beyond cardiac myocytes. Cav1.3 L-type Ca2+ channel is known to be highly expressed in neurons and neuroendocrine cells. However, we have previously demonstrated that the Cav1.3 channel is also expressed in atria and pacemaking cells in the heart. The significance of the tissue-specific expression of the channel is underpinned by our previous demonstration of atrial fibrillation in a Cav1.3 null mutant mouse model. Indeed, a recent study has confirmed the critical roles of Cav1.3 in the human heart (Baig, S. M., Koschak, A., Lieb, A., Gebhart, M., Dafinger, C., Nürnberg, G., Ali, A., Ahmad, I., Sinnegger-Brauns, M. J., Brandt, N., Engel, J., Mangoni, M. E., Farooq, M., Khan, H. U., Nürnberg, P., Striessnig, J., and Bolz, H. J. (2011) Nat. Neurosci. 14, 77–84). These studies suggest that detailed knowledge of Cav1.3 may have broad therapeutic ramifications in the treatment of cardiac arrhythmias. Here, we tested the hypothesis that there is a functional cross-talk between the Cav1.3 channel and a small conductance Ca2+-activated K+ channel (SK2), which we have documented to be highly expressed in human and mouse atrial myocytes. Specifically, we tested the hypothesis that the C terminus of Cav1.3 may translocate to the nucleus where it functions as a transcriptional factor. Here, we reported for the first time that the C terminus of Cav1.3 translocates to the nucleus where it functions as a transcriptional regulator to modulate the function of Ca2+-activated K+ channels in atrial myocytes. Nuclear translocation of the C-terminal domain of Cav1.3 is directly regulated by intracellular Ca2+. Utilizing a Cav1.3 null mutant mouse model, we demonstrate that ablation of Cav1.3 results in a decrease in the protein expression of myosin light chain 2, which interacts and increases the membrane localization of SK2 channels.

atrial compared with ventricular tissues, and null mutation of the SK2 channel results in atrial arrhythmias and atrioventricular node dysfunction (11,12). Moreover, different SK channel isoforms exist in human and mouse cardiac myocytes that can heteromultimerize via the coiled-coil domains in the C termini of the channels (10). Of clinical importance, I K,Ca contributes significantly to the repolarization process in human atria (8). Moreover, a recent study using genome-wide association analysis has provided evidence for a possible genetic link between SK channel polymorphisms and lone atrial fibrillation in humans (16).
To investigate the underlying mechanisms for the increase in atrial arrhythmias in Ca v 1.3 null mutant mice, the possible functional coupling between Ca v 1.3 and SK channels was tested (17). As it turns out, Ca v 1.3 colocalizes with SK2 channels in atrial myocytes. Moreover, the null mutation of Ca v 1.3 results in a significant decrease in I K,Ca and prolongation in atrial repolarization, which may underlie the increase in atrial arrhythmias (17).
In the present study, we tested the hypothesis that the decrease in I K,Ca in Ca v 1.3 Ϫ/Ϫ mice may result directly from impaired expression of SK2 channel interacting proteins leading to abnormal membrane localization of the SK2 channel protein in Ca v 1.3 Ϫ/Ϫ mice. Using a combination of techniques including yeast two-hybrid (Y2H) 4 assays, biochemical analyses, immunofluorescence confocal microscopic imaging, microarray analyses, and in vivo siRNA-mediated gene silencing, we report for the first time that the C terminus of Ca v 1.3 translocates to the nucleus where it functions as a transcription factor to regulate the expression of SK2 channel interacting proteins.

EXPERIMENTAL PROCEDURES
All animal care and procedures were approved by the University of California, Davis, Institutional Animal Care and Use Committee. Animal use was in accordance with the National Institutes of Health and institutional guidelines.
Plasmid Construction-pGBKT7 vector containing the GAL4 binding domain (BD) and pGADKT7 vector containing the GAL4 activation domain (AD) were used for Y2H assays (Clontech, Palo Alto, CA). A bait construct was generated in pGBKT7 vector containing a human cardiac SK2 C-terminal fragment encoding 380 -487 amino acid residues (GenBank TM accession number AY258141) (8) and used to screen a human heart cDNA library (MATCHMAKER, Clontech) as previously described (17). The construct was generated using PCR in-frame with the GAL4 BD and subcloned into the pGBKT7 vector. Constructs for the C-terminal domains of Ca v 1.3 and Ca v 1.2 Ca 2ϩ channels in pGBKT7 were Ca v 1.3-Ct, 1505-2203; and Ca v 1.2-Ct, 1505-2171. The numbers refer to amino acid sequences according to the rat Ca v 1.3 Ca 2ϩ channel (GenBank accession number D38101) and rabbit cardiac Ca v 1.2 Ca 2ϩ channel (GenBank accession number X15539), respectively. The cDNA fragments were placed in-frame with the DNA binding domain of GAL4 using polymerase chain reaction (PCR) and subcloned into the pGBKT7 vector. All clones were sequence verified. Rat Ca v 1.3 (␣ 1D ) Ca 2ϩ channel in pCMV6b vector was a kind gift from Dr. S. Seino (Chiba University, Chiba, Japan) and rabbit cardiac Ca v 1.2 Ca 2ϩ channel was a kind gift from Dr. Timothy Kamp (University of Wisconsin). pEYFP-N1-Ca v 1.3-ct was generated using the Ca v 1.3 cDNA fragment encoding 1505-2203 amino acid residues cloned in pEYFP-N1 (Clontech). The numbers refer to amino acid sequences according to the rat Ca v 1.3 Ca 2ϩ channel (GenBank accession number D38101). Human myosin light chain 2 (MYL2) clone with c-myc tag in pReceiver mammalian expression vector was purchased from GeneCopoeia (catalog number EX-T0572-M09, Rockville, MD).
Y2H Assays-Y2H assays were performed using the GAL4 system (MATCHMAKER GAL4 Two-hybrid System 3, Clontech) as previously described (17,18). Both bait and prey constructs in pGBKT7 and pGADKT7 were transfected to the AH109 yeast host strain. None of bait and prey plasmids had detectable autoactivation. For experiments in Fig. 5, pGADKT7 (containing the Gal4 activating domain) was used as a vector transformation marker because both pGBKT7 and pGADKT7 included the nutrition marker for Leu/Trp separately. The reporter genes in our experiments were histidine, adenine prototrophy, and ␤-galactosidase activity.
For liquid ␤-galactosidase assays, each fresh culture in the medium appropriate for the system and plasmids was spun, and the pellet was re-suspended in Z buffer (60 mM Na 2 HPO 4 , 40 mM NaH 2 PO 4 , 10 mM KCl, 1 mM MgSO 4 , pH 7.0), and then placed in liquid nitrogen for 1 min to freeze the cells and then the cells thawed in a 37°C water bath for 1 min. The freeze/ thaw cycles were repeated twice to ensure that the cells have been cracked open. Cells were spun to pellet the cell debris and the supernatant was transferred to a new microcentrifuge tube with Z buffer containing ␤-mercaptoethanol. Ortho-Nitrophenyl-␤-galactoside (2.5 mg/ml, Sigma) was added to the sample, and the mixture was incubated at 37°C for 6 h. The absorbance was read at 420 nm and represented the amount of protein present in each sample. For strong enzymatic reactions, diluted yeast lysate was used.
Y2H Screens-Y2H screening was performed with the GAL4 system using a protocol of MATCHMAKER GAL4 Two-hybrid System 3 (Clontech) as previously described (17). AH109 was used as the yeast host strain. A bait construct containing the SK2 C terminus in pGBKT7 vector was used to screen the human heart cDNA library of 3.5 ϫ 10 6 clones in the pACT2 prey vector (human heart MATCHMAKER cDNA library catalog number HL4042AH). Positive clones were selected by histidine and adenine prototrophy and assayed for ␣-galactosidase activity. Positive clones were isolated and characterized by sequencing.
Subcellular Fractionation and Western Blotting-Mouse heart tissue was rapidly removed and homogenized with a Dounce homogenizer for 20 up and down strokes with a loose pestle followed by 20 up and down strokes with a tight pestle and then again 10 up and down strokes with the loose pestle in buffer containing 250 mM sucrose, 20 mM HEPES (pH 7.4), 2 mM EGTA, 3 mM sodium azide, and Complete protease inhibitors (Roche Applied Science). The homogenate was centrifuged for 5 min at 700 ϫ g to remove the cell debris. The supernatant was then centrifuged for 60 min at 31,000 ϫ g at 4°C in a Beckman SW41 rotor. The pellet was resuspended in the homogenization buffer and subjected to ultracentrifugation in a discontinuous sucrose density gradient (32,40, and 50% (w/v) in 20 mM HEPES-NaOH, pH 7.4) for 2 h at 210,000 ϫ g in Beckman SW41 rotor. The interfaces of the homogenize buffer 32, 32/40, and 40/50% are the plasma membrane fractions, whereas the vesicles and mitochondrial fraction are pelleted at the bottom of tube. Protein concentration was measured by the BCA method (Thermo Fisher Scientific, Rockford, IL) and analyzed by Western blot as described previously (17,18).
RNA target preparation for microarray expression analysis was performed as per the instructions of the GeneChip 3Ј IVT Express Kit and GeneChip Hybridization, Wash, and Stain Kit (Affymetrix Inc., Broderick, CA). RNA was hybridized onto GeneChip mouse genome 430A 2.0 array (Affymetrix Inc.). Expression data were analyzed using dChip (DNA-Chip analyzer) software. Atrial tissues were isolated from three animals each for Ca v 1.3 Ϫ/Ϫ and WT littermates. A total of 6 samples were hybridized onto 6 different chips. Percentages for P call were 71. 5 Immunofluorescence Confocal Microscopy-Immunofluorescence labeling was performed as described previously (17,18). Cells were fixed by 4% paraformadehyde in phosphate-buffered saline (PBS) for 30 min at room temperature, washed three times with PBS, treated with 0.4% Triton X-100 in PBS for 15 min, then washed and treated with Ϫ20°C methanol for 10 min. Finally, cells were washed and treated with antibodies. Immunofluorescence-labeled samples were examined using a Pascal Zeiss confocal laser scanning microscopy. For double staining, Alexa Fluor 488-conjugated secondary antibody was excited at 488 nm with an Argon laser and detected with a 505-530-nm band pass filter, whereas Alexa Fluor 546-conjugated secondary antibody was excited at 543 nm with a HeNe1 laser and detected using a 560-nm long pass filter.
For atrial myocytes, which have been transduced with recombinant lentiviral vectors containing shRNA directed against Myl2, cells were immunofluorescence labeled using anti-SK2 (1:100) and anti-MLC2 (1:200) antibodies. Both antibodies were rabbit polyclonal antibodies and were pre-treated with the Zenon Alexa Fluor 555 and 647 rabbit IgG labeling kit (Invitrogen) according to the manufacturer's protocol.
To ascertain that there were no overlaps between the detection, singly labeled cells were imaged under identical conditions as those used for dual-labeled probes to confirm proper signal isolation of each channel. Imaris Bitplane software (Bitplane Scientific Software, Zurich, Switzerland) was used to generate three-dimensional reconstructions of confocal Z-stack images.
Cell Culture and Transfection-To generate primary cardiomyocyte cultures, hearts were dissected from 1-2-day-old mouse pups as previously described (20). Protein and RNA were isolated from the cardiomyocytes at different time points and subjected to analysis by Western blotting, RT-PCR, and immunofluorescence confocal microscopy. Neonatal cardiomyocytes and tsA 201 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 units/ml of penicillin, and 100 g/ml of streptomycin. Cells were maintained at 37°C in a humidified atmosphere containing 5% CO 2 . All cell culture reagents were purchased from Invitrogen.
Assessment of [Ca 2ϩ ] i -Neonatal mouse cardiomyocytes were isolated and cultured for 24 h. Cells were then incubated with the Ca 2ϩ indicator, Fluo 4-AM (10 M, Invitrogen) for 20 min at room temperature, and washed with DMEM. Three groups of cells were used including 1) cells incubated with DMEM alone for 1 h, 2) cells incubated with DMEM plus 2 mM EGTA for 1 h, and 3) cells incubated with DMEM plus 100 mM KCl for 1 h. Confocal line scan imaging was performed using a LSM 700 Zeiss confocal laser scanning microscope. Images were acquired at sampling rates of 0.7 ms/line and 0.07 m/pixel, with radial and axial resolutions of 0.4 and 1.0 m, respectively, as previously described (8,22). Ca 2ϩ transients were expressed as the normalized local fluorescence (F/F o ), where F o refers to the fluorescence level before depolarization using OriginPro 7 software (OriginLab Corp., Northampton, MA).
Lentiviral Vectors-Four shRNA directed against Myl2 were subcloned into the Ligase-free shRNA Cloning and Expression vector (pPS-H1-LCS-GFP, System Bioscience, Mountain View, CA) according to the manufacturer's protocol. The shRNA sequences were verified by sequencing.
The sequences for siRNAs directed against Myl2 (Qiagen) used in our study are as follow: Mm_Myl2_1 SI01321467 CAC-CGGCAATCTTGATTATAA; Mm_Myl2_2 SI01321474 CAG-AGACGGCTTCATCGACAA; Mm_Myl2_3 SI01321481 CTCAGACACCATGGCACCAAA; Mm_Myl2_4 SI01321488 CCCAGGGCTGTGCGCAAATAA. The lentivirus was produced using HEK293 cells. The pPS-H1-LCS (2 g), pMD.G (1 g), and pCMVD8.9 (4 g) plasmids were diluted in 750 ml of Opti-MEM (Sigma) separately for 5 min at room temperature. 20 l of Lipofectamine 2000 (Invitrogen) was diluted in 750 ml of Opti-MEM and incubated for 5 min at room temperature. The plasmids and Lipofectamine were mixed gently and incubated for 20 min at room temperature. HEK293 cells were washed with Opti-MEM without serum and antibiotics before adding the plasmids and incubated for 6 h at 37°C with 5% CO 2 . Fresh DMEM with 10% fetal bovine serum and 1% penicillin/ streptomycin was added at 6 and 24 h after transfection. The media was collected on day 3, centrifuged at 1500 ϫ g, and the lentivirus was precipitated using PEG-it (System Bioscience) according to the manufacturer's protocol. Transducing HEK293 cells with serial dilutions of the lentivirus was performed to determine the titer of the lentivirus. The concentration was determined using the Ultra-rapid lentiviral global titering Kit (System Bioscience) according to the manufacturer's protocol.
Targeted siRNA-mediated Gene Silencing Using Homogenous Transmural Lentiviral-mediated Atrial Gene Transfer-Homogenous transmural viral-mediated gene transfer in 10-week-old male C57Bl/6J mice was performed using techniques previously described in large animal models (23) and the technique was adapted for gene delivery for mouse atrial tissues as follow. Right thoracotomy was performed to expose the pericardial sac. Solutions used for gene delivery were made by adding 1% trypsin (w/v) to the PBS, chilled to 4°C, and 20% (w/v) of poloxamer (BASF Corp.) was then added slowly to obtain a total volume of 45 l (per animal). After the poloxamer had dissolved into the solution, the mixture was warmed to 37°C to achieve a gel-like consistency. Immediately before use, 30 l of lentivirus containing siRNA directed against Myl2 were added to the poloxamer/trypsin/PBS solution for a final virus concentration of 2 ϫ 10 9 pfu/ml to obtain a total volume of 75 l (per animal). Lentiviral mixture (75 l) was applied over the epicardial surface of the right atrium using a pipettor and incubated for 15 min before chest closure. Mice were maintained for 14 days when single atrial myocytes were isolated for immunofluorescence confocal microscopy as described above. To directly test the efficiency of lentiviral transduction using the above technique, atrial myocytes were isolated from animals that were transduced using a control lentivirus containing only green fluorescence protein (GFP) as a reporter gene. Atrial myocytes were sorted using fluorescence-activated cell sorting on a Cytomation MoFlo Cell sorter (University of California Davis Optical Biology CORE) to quantify for GFP positive cells compared with sham operated animals that did not receive the lentivirus. Data were acquired using Summit software (Cytomation) and analyzed using FlowJo software (version 9.0.1 Treestar Inc., San Carlos, CA).
Human Embryonic Kidney (HEK) 293 Cells and Plasmids Transfection-HEK293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml of penicillin, 100 g/ml of streptomycin. Cell lines were maintained at 37°C in a humidified atmosphere containing 5% CO 2 . All cell culture reagents were purchased from Invitrogen. HEK293 cells were transfected using the following plasmid compositions: 1) pSK2-IRES-EGFP alone or in combination with 2) pReceiver-hMLC2-c-Myc (1 g for each plasmid from GeneCopoeia) using Lipofectamine TM 2000 (Invitrogen) according to the manufacturer's protocol.

Ablation of Ca v 1.3 Ca 2ϩ Channels in Mouse Atrial Myocytes Results in Abnormal Surface Membrane Localization of SK2
Channels-Our previous work has demonstrated that null deletion of the Ca v 1.3 channel results in abnormal function of SK2 channels and prolongation of atrial repolarization and atrial arrhythmias (1,17). To directly determine the molecular mechanisms for abnormal function of SK2 in Ca v 1.3 Ϫ/Ϫ mice, subcellular distribution of the SK2 channel in isolated mouse atrial myocytes was examined by immunofluorescence confocal microscopy. Surprisingly, the subcellular distribution of SK2 channels in Ca v 1.3 Ϫ/Ϫ mice (Fig. 1a) was highly abnormal compared with that of WT animals (Fig. 1b) in both single staining using anti-SK2 antibodies and double staining probed by anti-SK2 and anti-␣-actinin antibodies. ␣-Actinin is an F-actin cytoskeletal protein located along the Z-lines in cardiac myocytes. As shown in Fig. 1b, in atrial myocytes isolated from WT mice, the SK2 channel showed the highest staining along the Z-line and co-localized with ␣-actinin protein, which is consistent with our previously published data (17,18). In contrast, in Ca v 1.3 Ϫ/Ϫ mice, SK2 showed a diffuse distribution pattern (Fig.  1a).
The above findings raised the question whether the expression level of SK2 channels may be altered in Ca v 1.3 Ϫ/Ϫ mice. To address this, we used western blot analyses to compare the expression level of SK2 channel protein in whole heart lysates from mutant (KO) and WT animals. There were no significant differences in SK2 channel protein expression, suggesting that ablation of the Ca v 1.3 channel did not affect protein expression of the SK2 channel (Fig. 2, a and b). The abnormal subcellular localization of SK2 channels in atrial myocytes in Ca v 1.3 Ϫ/Ϫ mice was further assessed using discontinuous sucrose density gradient ultracentrifugation to obtain subcellular fractionation from atrial tissues (Fig. 2c). The purity of the sarcolemmal membrane fractionation was tested and normalized using plasma membrane Na ϩ /K ϩ -ATPase. Notably, the subcellular fractionation revealed limited existence of SK2 channel protein at the plasma membrane in Ca v 1.3 Ϫ/Ϫ (KO) null mutant mice. Instead, the microsomal fraction (Fig. 2c, lane 5) was found to contain a high density of SK2 channel protein in Ca v 1.3 Ϫ/Ϫ atrial tissues. In contrast, SK2 channel protein was quite apparent in the plasma membrane fraction, which was confirmed by the presence of plasma membrane Na ϩ /K ϩ -ATPase in WT tissues. This data support the notion that trafficking of the SK2 channel proteins to the plasma membrane may be impaired in The C Terminus of Ca v 1.3 (Ca v 1.3-ct) Is Found in the Nucleus of Cardiac Myocytes-Changes in subcellular distribution of channel proteins can be influenced by a number of factors, such as the production of constituent subunits, the proper assembly of these subunits, the trafficking of assembled proteins to the plasma membrane, and recycling or degradation pathways. To further investigate how the Ca v 1.3 Ca 2ϩ channel may affect subcellular distribution of the SK2 channel protein, we hypothesize that the Ca v 1.3 channel may regulate transcription of endogenous genes critical for SK2 channel trafficking. Previous experiments in neurons have suggested that the C terminus of the Ca v 1.2 channel encodes a transcription factor (25). More recent study further demonstrates that the C terminus of the Ca v 1.2 channel is produced by activation of a cryptic promoter in exon 46 of CACNA1C, the gene that encodes Ca v 1.2 (26).
To determine whether the C terminus of the Ca v 1.3 channel may regulate transcription in cardiac myocytes, we directly tested for possible nuclear localization of Ca v 1.3 channels using an antibody that recognizes the C terminus of Ca v 1.3 (amino acids 1661-1990). Interestingly, instead of demonstrating only plasma membrane localization as shown in Fig. 3a, Ca v 1.3 was also observed to be highly localized in the nucleus in neonatal cardiomyocyte when the C-terminal antibody was used (Fig. 3a,  lower panels). In contrast, no nuclear fluorescence was observed when using an antibody recognizing an epitope in the II-III cytoplasmic loop of the Ca V 1.3 channel (amino acids 859 -875, Fig. 3a, upper panels). To provide further evidence that the C terminus of Ca V 1.3 was highly localized in the nucleus, we isolated nuclei from adult atrial myocytes from WT mice to identify the localization pattern of Ca v 1.3 by using antibodies recognizing different epitopes (Fig. 3b). Consistently, the C-terminal antibody of the Ca v 1.3 channel detected a high nuclear fluorescence (lower panels) in contrast to the antibody directed against the II-III cytoplasmic loop (upper panels), suggesting that the C terminus of Ca v 1.3 was also enriched in nuclei in adult atrial myocytes. Fig. 3c further demonstrated the localization of Ca v 1.3 along the Z-lines in adult atrial myocytes  FEBRUARY 20, 2015 • VOLUME 290 • NUMBER 8 co-stained with anti-Ca v 1.3 (loop antibody) and ␣-actinin2 antibodies. A different set of antibodies directed against the C-(C-term) and N-terminal (N-term) domains were used and similar results were obtained (Fig. 3d). These monoclonal antibodies were obtained from the University of California Davis/ National Institutes of Health NeuroMab Facility.

Gene Transcription by Ca v 1.3 Calcium Channel
To directly rule out the possibility that the observed nuclear staining by C-terminal antibodies was nonspecific, we fused yellow fluorescent protein (YFP) to the C terminus of the Ca v 1.3 channel (Ca v 1.3-ct-YFP, Fig. 4, a and b). We observed distinct nuclear fluorescence with prominent punctate patterns when Ca v 1.3-ct-YFP was transfected in tsA 201 cells (Fig. 4c, top row) and native neonatal cardiomyocytes (Fig. 4d, top row). In comparison, there was a very low YFP signal in the cytosol of the cells. Z-stack images and three-dimensional reconstruction were further obtained to demonstrate the localization of the C-terminal domain within the nuclei of cells (Fig. 4, h and i). [Ca 2ϩ ] i , we assessed distribution of Ca v 1.3-ct-YFP using immunofluorescence confocal microscopy in both tsA 201 cells and neonatal cardiomyocytes following treatment with agents that affect [Ca 2ϩ ] i levels. Decreasing the free extracellular Ca 2ϩ using 2 mM EGTA for 1 h did not affect nuclear localization of Ca V 1.3-ct-YFP (Fig. 4, c and d, middle panels). Conversely, treatment with 100 mM KCl for 1 h to increase the activity of Ca 2ϩ channels resulted in a significant decrease in nuclear fluorescence (Fig. 4, c and d, lower panels). The corresponding changes in [Ca 2ϩ ] i and spontaneous electrical activities after the addition of 2 mM EGTA or 100 mM KCl are depicted using Fluo-4 AM and confocal line scan imaging (Fig. 4f). A control experiment is shown in Fig. 4e in tsA 201 cells transfected with the pEYFP-N1 control vector illustrating diffuse nuclear and cytoplasmic fluorescence.

The C Terminus of Ca v 1.3 (Ca v 1.3-ct) Activates Transcription as Assessed by Y2H-To investigate whether
Ca v 1.3-ct can activate transcription when recruited to a promoter by a heterologous DNA binding domain, we constructed an N-terminal fusion of Ca V 1.3-ct with the Gal4-DNA binding domain from yeast (pGBKT7-Ca v 1.3-ct, Fig. 5a). The GAL4-DNA binding domain recognizes the UAS-DNA sequence but requires a transcriptional activation domain to activate transcription (pGADKT7). We introduced pGBKT7-Ca v 1.3-ct into yeast cells along with three reporter genes (histidine, adenine prototrophy, and ␣-galactosidase activity). All reporter genes could be activated. Thus, the yeasts transfected by pGBKT7-Ca V 1.3-ct showed a robust growth in high stringency medium (Fig. 5a, left panel, second sector). pGBKT7-Ca v 1.3-ct could significantly activate transcription compared with GAL4 alone or channel lacking the GAL4-DNA binding domain. Taken together, these results suggest that the C terminus of Ca v 1.3 may be produced as a soluble protein in cells and translocates to the nucleus where it activates transcription when recruited to promoters of the genes. In comparison, we generated another construct including the N-terminal fusion of Ca V 1.3-ct with the GAL4activation domain (pGADKT7-Ca V 1.3-ct). The GAL4-DNAactivation domain could activate transcription but requires a DNA binding domain to recognize the UAS-DNA sequence to activate transcription. Consequently, none of the reporter genes could be activated (Fig. 5a, left panel, third sector). Quantification data obtained using ␤-galactosidase activity further support the results of Y2H assays (Fig. 5b), suggesting that the C terminus of Ca v 1.3 can act as a transcription factor. Similar data were obtained using the C terminus of Ca v 1.2 (Fig. 5b). Nonetheless, the findings using the Y2H assay are only suggestive and additional experiments were performed as detailed below to further test the roles of Ca v 1.3 C-terminal domain as transcriptional regulators. GAPDH was used as a loading control. b, summary data for SK2 protein expression levels normalized to GAPDH. c, abnormal distribution of SK2 channels was observed in Ca v 1.3 Ϫ/Ϫ mice using a discontinuous sucrose density gradient ultracentrifugation. The purity of membrane fractionation in mouse atrial tissues was tested and normalized by Na,K-ATPase as the plasma membrane marker. Each sample of atrial myocytes was isolated from 5 animals and the experiments were repeated independently three times.

Differential Transcript Expression in Atria from Ca v 1.3 Ϫ/Ϫ Compared with WT Littermates-To further test whether
Ca v 1.3 may regulate transcription of endogenous genes in vivo, we used microarray expression analysis to identify differences in transcript expression levels between Ca v 1.3 Ϫ/Ϫ and WT littermates (Fig. 6). Using three independent atrial tissues from each group, we observed 24 transcripts that were up-regulated more than 2-fold (p Ͻ 0.005) and 11 transcripts that were upregulated 3 or more than 3-fold (p Ͻ 0.005) in Ca v 1.3 Ϫ/Ϫ compared with WT littermates. Among the up-regulated mRNAs, several were related to immunoglobulin. Furthermore, there were 12 transcripts that were down-regulated more than 3-fold (p Ͻ 0.005) in Ca v 1.3 Ϫ/Ϫ compared with WT animals. Among the down-regulated mRNAs, three were related to myosin light chain (MLC, 10 different transcripts are shown in Fig. 6). To verify the results of the microarray analyses, semi-quantitative RT-PCR was performed (Fig. 7a). The mRNA expression of Myl2 encoding for MLC2 was significantly decreased in Ca v 1.3 Ϫ/Ϫ mice, in accordance with the results from microarray analyses (n ϭ 3). Furthermore, we measured MLC2 protein levels by western blot analysis. Consistently, the MLC2 protein level was reduced significantly in Ca v 1.3 Ϫ/Ϫ atrial myocytes compared with that of WT littermates (Fig. 7b). Taken together, the findings are consistent with possible transcriptional regulation of the Myl2 gene by Ca v 1.3 channels.
C Terminus of SK2 Channel Directly Interact with MLC2 Using Y2H Assays-The data so far suggest that Ca v 1.3 may affect trafficking of the SK2 channel even though ablation of Ca v 1.3 did not significantly alter the expression level of the SK2 protein. We hypothesize that abnormal subcellular localization of the SK2 channel protein in Ca v 1.3 Ϫ/Ϫ mice may be a direct result of the impaired expression of MLC2. To directly test this hypothesis, we assessed whether MLC2 interacted with the SK2 channel. A bait construct was generated in pGBKT7 plasmid containing GAL4-DNA BD using human cardiac SK2 C-terminal domain (Fig. 7c, upper panel). The bait construct was used to screen the human heart cDNA library (MATCHMAKER, Clontech). Consequently, three different independent positive clones (C2, C4, and C24) were found to belong to Myl2. None of the bait and prey plasmids had detectable autoactivation. To further confirm this interaction, we switched the C4 clone, which contains full-length Myl2 cDNA, to pGBKT7-BD vector and moved SK2-C1 to the pGADKT7-AD vector, and obtained a strong interaction between SK2-C1 and MLC2 showing a similar white color phenotype to the positive colonies of pGADT7-T with p53, suggesting that SK2 physically interacts with MLC2 in Y2H assays (Fig. 7c, lower panels).
Targeted siRNA-mediated Gene Silencing by Homogenous Transmural Atrial Gene Transfer-To directly test the roles of MLC2 in atrial myocytes, we used the previously published technique of homogenous transmural atrial gene transfer in large animal models (23) for gene delivery to mouse atrial tis- sues as described (22). Single atrial myocytes were isolated 48 h after in vivo viral transduction.
Normal Expression of MLC2 Is Required for Membrane Localization of SK2 Channels-Treatment of atrial myocytes with MLC2-specific siRNAs resulted in a significant decrease in MLC2 expression. Moreover, knockdown of MLC2 resulted in an abnormal localization of SK2 channels with a complete lack of staining along the Z line (Fig. 7d, right panels) compared with the normal SK2 staining pattern in Fig. 7d, left panels. The staining pattern of MLC2 in control atrial myocytes was shown in the left panel illustrating subcellular localization of MLC2 (Fig. 7d). Of note, because both anti-SK2 and anti-MLC2 antibodies were rabbit polyclonal antibodies. Both antibodies were pre-treated with the Zenon Alexa Fluor 555 and 647 rabbit IgG labeling kit (Invitrogen) in Fig. 7d. The differences in the techniques used may result in a slight discrepancy in the intensity of SK2 staining between Figs. 7d and 1a when Ca v 1.3 was ablated.   FEBRUARY 20, 2015 • VOLUME 290 • NUMBER 8

JOURNAL OF BIOLOGICAL CHEMISTRY 4671
Co-expression of MLC2 Increases the Functional Expression of SK2 Channel-Finally, we performed patch-clamp recordings of apamin-sensitive currents in HEK293 cells expressing the SK2 channel alone compared with SK2 co-expressed with MLC2 (Fig. 7e). There was a significant increase in apaminsensitive currents in cells co-expressing SK2 and MLC2 consistent with the notion that MLC2 facilitates and increases membrane localization of SK2 channels. Fig. 8A shows a schematic representation of the recombinant lentiviral constructs containing GFP as the reporter gene and four shRNA directed against MLC2 (LV-MLC2) and scrambled sequence (LV-Scramble). We further demonstrated the successful isolation of GFP-positive atrial myocytes in Fig. 8B using fluorescence-activated cell sorting (FACS). Isotype-matched antibodies were used to control for background fluorescence (FL) and is shown in the left panel labeled as Background FL compared with the Cell-Specific FL shown in the right panel.
Taken together, our data suggests that the C-terminal domain of the Ca v 1.3 channel acts as a transcriptional regulator for Myl2 expression. MLC2 protein physically interacts with SK2 channels and facilitates the membrane localization of SK2 channels. A schematic is shown in Fig. 7f.

Ablation of Ca V 1.3 Results in a Decreased Expression of MLC2 and Abnormal Surface Membrane Localization of SK2
Channels-We have previously demonstrated the molecular coupling between Ca v 1.3 and SK2 channels in cardiac myocytes (17). Moreover, null deletion of Ca v 1.3 results in a significant decrease in I K,Ca in atrial myocytes leading to prolongation of repolarization and atrial arrhythmias (17). Indeed, we have previously shown that direct protein-protein interaction between SK2 channels and two different cytoskeletal proteins, ␣-ac-tinin2 and filamin A, is required for the surface membrane localization of the SK2 channels (18,27). Specifically, we demonstrate that filamin A (FLNA) and ␣-actinin2 (Actn2) directly interact with the N and C termini of SK2 channels, respectively (18,27) (Fig. 7f). Here, using a combination of techniques including biochemical analyses, Y2H assays, and confocal imaging, we directly tested the mechanisms underlying the observed functional cross-talk between Ca v 1.3 and SK2 channels. We demonstrate that the C terminus of the Ca v 1.3 Ca 2ϩ channel can be found to be localized in the nucleus and acts as a transcription factor in atrial myocytes. Our data further suggest that nuclear translocation of the C-terminal domain of Ca v 1.3 Ca 2ϩ channels is directly regulated by [Ca 2ϩ ] i . Finally, using microarray analyses and siRNA-mediated gene silencing, we provide new insights into the underlying mechanisms for the observed functional coupling between Ca v 1.3 and SK2 channels. Ca v 1.3 directly regulates the expression of MLC2, which interacts with and aids in surface membrane localization of SK2 channels.
L-type Ca 2ϩ channels are highly expressed in skeletal and cardiac muscle, where they play important roles in excitationcontraction coupling. Several laboratories as well as ours have documented expression of the Ca v 1.3 Ca 2ϩ channel, in addition to the Ca v 1.2 isoform in pacemaking tissues and atria (2,4,5,34). Physiologically, the Ca v 1.3 channel is less sensitive to dihydropyridines compared with the Ca v 1.2 channel. Moreover, the Ca v 1.3 channel exhibits a low threshold of activation compared with the Ca v 1.2 isoform (35)(36)(37). This distinct biophysical feature of Ca v 1.3 may contribute to its role  ). b, Western blot analysis showing a significant decrease in MLC2 protein in Ca v 1.3 Ϫ/Ϫ mice compared with WT animals (n ϭ 3 animals for each group were used for Western blot analyses). c, MLC2 interacted directly with the SK2 channel as assessed using Y2H assays. d, left panels show the SK2 staining pattern in atrial myocytes treated with control siRNA. Right panels: treatment of atrial myocytes with siRNA specific to MLC2 resulted in a significant decrease in MLC2 expression. Moreover, there was a significant decrease in SK2 membrane expression. Lower panels compare the SK2 staining of the regions outlined in white boxes at higher magnification. A total of three animals were used for each group. e, apamin-sensitive currents recorded from HEK293 cells transfected with human SK2 channel alone compared with SK2 channel co-expressed with hMLC2. Data from nontransfected cells are shown in the lower left panel. Right lower panel compared the current density in pA/pF of apamin-sensitive currents. A total of 6 cells were performed for each group with similar results. f, a schematic model illustrating the C terminus of Ca v 1.3 as a transcriptional regulator for MLC2 expression. MLC2 physically interacts with SK2 channels and facilitates the targeting of SK2 channels to the plasma membrane. SK2 channels are shown to interact with filamin A (FLNA) and ␣-actinin2 (Actn2) cytoskeletal proteins via the N and C termini, respectively. CaMBD refers to calmodulin binding domain within the C terminus of SK2 channels. FEBRUARY 20, 2015 • VOLUME 290 • NUMBER 8 in pacemaking activities in sinoatrial and atrioventricular nodes.

Gene Transcription by Ca v 1.3 Calcium Channel
The C Terminus of Ca v 1.2 Ca 2ϩ Channel Encodes a Transcription Factor-[Ca 2ϩ ] i is tightly regulated and plays critical roles in multiple cellular processes including synaptic transmission, cell division, differentiation, and contraction. Moreover, previous studies have demonstrated the involvement of voltage-gated Ca 2ϩ channels in the regulation of gene expression (38 -40). Ca 2ϩ influx through L-type Ca 2ϩ channels can activate transcription factors such as cAMP-response elementbinding protein, MEF, and NFAT (41)(42)(43). Two possible mechanisms that have been proposed include activation of nuclear Ca 2ϩ -dependent enzymes (44) and activation of Ca 2ϩ -dependent factors around the inner pore of the Ca 2ϩ channel with propagation of the signal to the nucleus (39).
More recent studies have provided evidence for a new mechanism whereby the C terminus of Ca v 1.2 can directly enter the nucleus and act as a transcriptional regulator (25). The C-terminal domain of Ca v 1.2 has been shown to autoregulate Ca v 1.2 expression in cardiac myocytes (45). Previous studies have documented that L-type Ca 2ϩ channels may be cleaved at their C terminus and the C-terminal domain can bind to the truncated channel to exert its inhibitory effect on the channel function (46,47). In addition to Ca v 1.1 and Ca v 1.2, previous studies have reported that C-terminal domains of Ca v 1.3 (48), Ca v 2.1 (49), and Ca v 2.2 (50) may be cleaved in neurons. The cleaved product of Ca v 2.1 has been shown to be localized in the nucleus of neurons (51) suggesting that other members of Ca v may also act as transcriptional regulators in different tissues. On the other hand, one recent study demonstrates that the C terminus of the Ca v 1.2 channel is produced by activation of a cryptic promoter in exon 46 of CACNA1C, the gene that encodes Ca v 1.2 and does not represent the cleaved product (26). Additional studies will need to be conducted to determine the mechanisms for generation of the Ca v 1.3 C-terminal domain as well as domains within the C terminus that are required for the observed nuclear translocation.
We have provided intriguing data that translocation of the Ca v 1.3 C terminus is inversely dependent on [Ca 2ϩ ] i . Additional evidence is provided that the interventions used in our experiments altered the spontaneous firing frequency and [Ca 2ϩ ] i . These findings are somewhat unexpected. An increase in [Ca 2ϩ ] i would be predicted to lead to a mechanism, which would abbreviate the APD to avoid Ca 2ϩ overload. However, our data suggest that the C terminus translocates from the nucleus when [Ca] i is high, which would be predicted to lead to a decrease in membrane expression of SK2 channels. On the other hand, the findings likely represent one of the many factors that are altered by an increase in [Ca 2ϩ ] i . Additional studies are needed to further decipher additional genes regulated by the C terminus of Ca v 1.3.
SK Channels-SK channels are encoded by at least three distinct genes, KCNN1, KCNN2, and KCNN3 (52)(53)(54). The channels have been described in a wide range of tissues (55,56). Pharmacologically, SK channels can be distinguished by their sensitivity toward the bee venom apamin (57,58). SK channels are highly unique in that they are gated solely by [Ca 2ϩ ] i . The Ca 2ϩ -binding protein calmodulin (CaM) binds to the SK channel through the CaM-binding domain (CaMBD), which is located in an intracellular region of the ␣ subunit immediately carboxyl-terminal to the pore (21).
Physiological Significance-Recent data from our laboratory as well as others have provided evidence for critical roles of both the Ca v 1.3 Ca 2ϩ channel (1)(2)(3)(4)(5) and SK channels in hearts (8, 9, 13-15, 17, 18). Both channels are highly expressed in atrial myocytes and pacemaking tissues. Because SK channels are gated solely by changes in [Ca 2ϩ ] i , the channels are predicted to play critical roles in pathological conditions in which there is a significant increase in [Ca 2ϩ ] i , for example, during heart failure (15) or atrial fibrillation, leading to a decrease in the refractory period. The data presented here support a novel mechanism for the additional cross-talk between Ca v 1.3 and SK channels. Indeed, Ca 2ϩ channels are not only required for activation of the SK channels. Ca v 1.3 may influence membrane localization of the SK channels via transcriptional regulation of SK channel interacting proteins, which are required for the proper trafficking of the SK channels.
Future Studies-Data in Fig. 6 provide evidence for differential transcript expression in Ca v 1.3 null mutant mice compared with WT littermates, however, the findings do not directly demonstrate transcriptional regulation. Future studies are required to test direct regulation of the Myl2 gene by the C terminus of Ca v 1.3.