Expression, localization, and functional properties of Bestrophin 3 channel isolated from mouse heart

Abstract

Bestrophins are a novel family of proteins that encode calcium-activated chloride channels. In this study we establish that Bestrophin transcripts are expressed in the mouse and human heart. Native mBest3 protein expression and localization in heart was demonstrated by using a specific polyclonal mBest3 antibody. Immunostaining of isolated cardiac myocytes indicates that mBest3 is present at the membrane. Using the patch-clamp technique, we characterized the biophysical and pharmacological properties of mBest3 cloned from heart. Whole cell chloride currents were evoked in both HEK293 and COS-7 cells expressing mBest3 by elevation of intracellular calcium. mBest3 currents displayed a K_D for Ca²⁺ of ∼175 nM. The calcium-activated chloride current was found to be time and voltage independent and displayed slight outward rectification. The anion permeability sequence of the channel was SCN⁻>I⁻>Cl⁻, and the current was inhibited by niflumic acid and DIDS in the micromolar range. In addition, we generated a site-specific mutation (F80L) in the putative pore region of mBest3 that significantly altered the ion conduction and pharmacology of this channel. Our functional and mutational studies examining the biophysical properties of mBest3 indicate that it functions as a pore-forming chloride channel that is activated by physiological levels of calcium. This study reports novel findings regarding the molecular expression, tissue localization, and functional properties of mBest3 cloned from heart.

ca²⁺-activated cl⁻ channels (Cl_Ca) are expressed in a wide range of cell types where they play important functional roles (reviewed in Ref. 16). Despite the significance of their conductance in many tissues, the molecular identity for the encoding channel remains unknown. Bestrophins have been identified as a novel family of Cl_Ca (46) with multiple members expressed in different species (17). Several Bestrophin members have been shown to produce a calcium-activated chloride current (I_ClCa) when expressed heterologously (30, 33–36, 41, 46, 48, 49). Since Bestrophins are proposed as candidates to encode Cl_Ca, we examined whether Bestrophins are present in the heart, a tissue from which I_ClCa has been recorded (16). In addition, only a few studies have investigated Bestrophin protein expression. Immunohistochemisty studies have localized human Bestrophin 1 to the basolateral membrane of the retinal pigment epithelium (23); murine Bestrophin 1 to tracheal and colonic epithelial cells and kidney (4, 5); and murine Bestrophin 2 to the basolateral membranes of olfactory sensory neurons, colonic, tracheal, and nonpigmented cilliary epithelial cells (3, 5, 30). There have been no studies on the tissue and cellular localization of native Bestrophin 3 in any tissue including the heart.

The murine Bestrophin family (mBest1-3) has been previously identified (20). A detailed examination of the biophysical properties of mBest2 has been reported (30, 32, 34, 35). Whole cell patch-clamp analysis has also been reported for a truncated form of mBest3 (33, 37). These authors were unsuccessful in inducing functional chloride channels when they overexpressed the full-length mBest3 in HEK293 cells even though this protein was expressed at the plasma membrane (33, 37). A truncated form of mBest3 was reported to produce a Ca²⁺-insensitive chloride conductance when expressed heterologously (33). More recently it has been demonstrated that the full-length mBest3 encodes a functional Cl_Ca when expressed heterologously (41); however, a detailed characterization of the biophysical properties of this murine Bestrophin family member is lacking.

In the present study we examined whether Bestrophin transcripts are expressed in both the mouse and human heart using RT-PCR and real-time PCR. mBest3 is the most predominant isoform expressed in the mouse heart. Therefore, we investigated mBest3 protein expression and tissue and cellular localization in the heart using a specific mBest3 polyclonal antibody. Using the whole cell patch-clamp technique, we characterized the biophysical and pharmacological properties of heterologously expressed mBest3, which was cloned from the heart. We generated a mutation in the putative pore of mBest3 that significantly changed the ion conduction and pharmacology of this channel. We report novel findings regarding the molecular expression, localization, and functional properties of mBest3 cloned from the heart.

MATERIALS AND METHODS

Total RNA isolation and RT-PCR.

Adult BALB/c mice were sedated by exposure to AErrane isoflurane (Baxter Healthcare, Deerfield, IL) and killed by cervical dislocation. The mouse chest cavity was opened, and the heart was quickly excised. This protocol has been approved by the University of Nevada Institutional Animal Care and Use Committee and is compliant with the Guide for the Care and Use of Laboratory Animals (1996, NIH, Bethesda, MD). Total RNA was isolated from mice hearts using TRIzol reagent (Invitrogen, Carlsbad, CA), following the manufacturers instructions. Human total RNA was purchased from Stratagene (Cedar Creek, TX). First-strand cDNA was prepared from 1 μg of RNA using Oligo(dT)_(12–18) primer and Superscript II reverse transcriptase (Invitrogen). The resulting cDNA was directly used for the PCR of murine and human Bestrophin isoforms using AmpliTaq Gold PCR Master Mix (Applied Biosystems, Foster City, CA). The specific primer pairs are described in Table 1. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control. The amplification profile for these primer pairs were as follows: 95°C for 10 min, then 35 cycles at 95°C for 15 s, annealing temperature (T_A) for 20 s, and 72°C for 30 s, followed by a final step at 72°C for 7 min. PCR reactions were performed in a GeneAmp 2700 thermal cycler (Applied Biosystems).

Table 1.

Oligonucleotide primers used in this study

Primer	Gene Accession No.	Oligonucleotide Sequences (5′-3′)	Amplicon Size
GAPDH	NM_008084	GTCTTCACCACCATGGAGA (sense)	170 bp
		AAGCAGTTGGTGGTGCAG (antisense)
Best1	NM_011913	GCTTGATCGGGAGGCAGTTTCTGAACCCA (sense)	341 bp
		GAAGGTGGAGCCCATGAAGGAATGCCG (antisense)
Best2	NM_145388	CGACAACAGTGCCCTAAAGTTGC (sense)	463 bp
		CTGAAGCAGGAAAGCGGTGG (antisense)
Best3	NM_001007583	TCCACGGGAGCGACCAGCATGGG (sense)	363 bp
		CCAGCGGAATGCCAACCCAGTCGTAACC (antisense)
BEST1	NM_004183	GCTGGCGGGGCAGCATCTACAAGC (sense)	330 bp
		TGGCGTAGCGGATGAGCGTGCGC (antisense)
BEST2	NM_017682	GCAGAAGCGCTACTTCGAGAAGC (sense)	319 bp
		CATAAACCCAGCCTCCACCACG (antisense)
BEST3	NM_032735	GATGAAGAAGGACATTTACTGGGACG (sense)	284 bp
		GGGGTAAGAACATGGAGCTGTC (antisense)
BEST4	NM_153274	TGGTGGTCCCAGTACACAAGCA (sense)	346 bp
		AGAGAGCGATATCGTCACGTATTCGC (antisense)
Best1-q	NM_011913	ACCCAGCAGCCACCCTACTGAGCAGTCAGCA (sense)	178 bp
		GGTTGACATCTGGTCCAAACGGCGCTGTTG (antisense)
Best2-q	NM_145388	TCTATACTGGGATGCAGCAGAAGCTCGC (sense)	182 bp
		GCTGCAGAAAGTCCCCGTGAACCTCT (antisense)
Best3-q	NM_001007583	TCTTCTACGCAGGATGGCTCAAGGTTGCAGAG (sense)	200 bp
		TGTGTACGGTGGCCGGGCAGCAGAATC (antisense)

GAPDH, glyceraldehyde-3-phosphate dehydrogenase. Best1, Best2, Best3 correspond to mus musculus Bestrophin isoforms. BEST1, BEST2, BEST3, and BEST4 are Homo sapiens Bestrophin isoforms. Best1-q, Best2-q, and Best3-q are specific primers utilized for quantitative PCR.

Quantitative RT-PCR.

Real-time quantitative PCR was performed on an ABI Prism 7000 sequence detector (Applied Biosystems) using SYBR Green chemistry with gene-specific primers, as described previously (7). The primers used for quantitative RT-PCR (qBest1-3) are shown in Table 1. GAPDH was used as a reference gene. Standard curves were generated for each set of primers using serially diluted solutions of cDNA. The slopes of our standard curves for mBest1, mBest2, mBest3, and GAPDH were similar; thus the efficiency of our primer sets were considered equal and permitted relative quantitation of Bestrophin transcripts. Unknown quantities relative to the standard curve for the Bestrophin primers were calculated, and these values were normalized to endogenous GAPDH RNA within the same sample. cDNA was obtained from four different mice, and each cDNA sample was tested in triplicate. Data are presented as means ± SE. Significance among groups was tested by ANOVA.

Cloning of mBest3 from mouse hearts.

A 2.7-kb cDNA fragment containing the entire protein coding sequence for mBest3 was amplified by PCR from the mouse heart with specific primers designed for mBest3 (GenBank accession no. NM_001007583). The PCR amplification product was purified and inserted into the pcDNA3.1 vector (Invitrogen), and recombinant plasmids were sequenced at the Nevada Genomics Center. mBest3 nucleotide sequences were analyzed with Vector NTI software (Invitrogen) and BLASTN software at the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/BLAST/). For expression studies, the mBest3 transcript was ligated into pcDNA4/TO/c-myc-HIS vector (Invitrogen), which contains a tetracycline operon (TO) for tetracycline-inducible expression. The mBest3 stop codon was removed, and XhoI and EcoRI restriction sites were placed at the termini by PCR so that the mBest3 transcript was inserted in frame with the COOH-terminal c-myc epitope tag.

Site-directed mutagenesis of mBest3.

A F80L mutation of mBest3 was generated using the Quickchange site-directed mutagenesis kit (Stratagene). mBest3 was PCR amplified with Pfu DNA polymerase and primers that introduced the specific mutation. The mutation was confirmed by DNA sequencing.

Heterologous expression of mBest3 in mammalian cell lines.

Functional expression of mBest3 was carried out using the Tetracycline-Regulated Expression system (T-REx, Invitrogen). TRex-293 cells [stably transfected with pcDNA6/TR, containing the tetracycline repressor (TR) protein], were transfected with 10 μg of the mBest3-pcDNA4/TO/c-myc-HIS plasmid using Polyfect reagent (QIAGEN, Valencia, CA). Stable cell lines expressing mBest3 were selected on the basis of resistance to the antibiotics zeocin and blasticidin, conferred by the pcDNA/TO and pcDNA/TR plasmids, respectively. TRex-293 cell culture media was Dulbecco's Modified Eagle Medium (DMEM) High Glucose (Invitrogen) with 10% (vol/vol) FBS (Hyclone Laboratories, Logan, UT), 1% (vol/vol) penicillin-streptomycin, 5 μg/ml blasticidin, and 250 μg/ml zeocin. Three out of seven cell cultures were tested for tetracycline-inducible mBest3 channel expression. Transfected and untransfected TRex-293 cells were seeded onto glass coverslips 24–48 h before the recordings. Tetracycline (1 μg/ml) was added to each dish of cells 3–6 h before the recording. In some instances, mBest3 or mBest3-F80L plasmids were transiently transfected into HEK-tsA201 or COS-7 cells (obtained through the American Type Culture Collection, Bethesda, MD) for patch-clamp analysis or immunohistochemistry. Cells in 35-mm culture dishes were transfected with 2 μg plasmid DNA using Polyfect reagent (QIAGEN) according to the manufacturer's protocol. Twenty-four hours after transfection, cells were seeded at a low density on glass coverslips. Patch-clamp recordings were performed 24 h later.

Production of anti-mBest3 antibody.

A polyclonal mBest3 antibody was developed using a commercial provider (Open Biosystems, Huntsville, AL). The antibody was generated in rabbits immunized with a 15-amino acid peptide derived from the COOH-terminus of mBest3 (residues ⁶⁵⁵TGESPKGTPQRPRTWF⁶⁶⁹).

Protein extraction and Western blot analysis.

Western blot experiments were performed similarly to those described in Ref. 6. Tissues were pulverized under liquid nitrogen and homogenized with a glass Dounce homogenizer in ice-cold T-PER Tissue Protein Extraction Reagent (Pierce, Rockford, IL) containing 1 mM EDTA and 1× Complete Protease Inhibitor cocktail (Roche, Nutley, NJ). Protein concentrations were determined by the Bradford assay (Bio-Rad Laboratories, Hercules, CA). Protein lysates (50 μg) were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) through 10% gels under reducing conditions. A broad-range SDS-PAGE standard (Bio-Rad Laboratories) was included to determine molecular mass. Proteins were electrophoretically transferred onto nitrocellulose membranes for 1.5 h at 24 V (Genie blotter, Idea Scientific, Minneapolis, MN). Membranes were blocked for 2 h with 5% skimmed milk in Tris-NaCl-Tween 20 (TNT) buffer (containing 100 mM Tris, pH 7.5, 150 mM NaCl, 0.1% Tween-20) and incubated with mBest3 antisera (1:5,000 in 5% milk-TNT) at 4°C overnight. Excess antibody was removed by 3 × 5 min washes with TNT buffer, and membranes were incubated with goat anti-rabbit alkaline phosphatase-conjugated antibody (1:7,500 dilution in 5% milk-TNT; Promega, Madison, WI) for 2 h at 4°C. Excess secondary antibody was removed by 3 × 5-min washes with TNT buffer. Color was developed with the alkaline phosphatase substrate, 5-bromo-4-chloro-3-indolylphosphate/nitro blue tetrazolium (BCIP/NBT). Negative controls of protein blots incubated with 1) rabbit preimmune serum, and 2) the mBest3 antibody preabsorbed with the immunization peptide at 4°C overnight were also performed. Immunoblots were scanned to obtain images of mBest3 immunoreactive bands using Quantity One software (Bio-Rad Laboratories).

Immunohistochemistry and confocal analysis.

For cryosections of the mouse heart, the heart was removed and flushed with Krebs-Ringer-bicarbonate solution. The tissues were fixed with 4% paraformaldehyde in PBS (pH 7.4) for 20 min. The fixed tissue was cut into sections and cryoprotected in increasing gradients of sucrose in PBS (5, 10, and 15% for 30 min each and in 20% overnight). Tissues were embedded in Tissue Tek (Miles Scientific) and rapidly frozen in isopentane precooled in liquid nitrogen. Cryosections were cut at 10 μm, using a cryostat (Leica CM3050, Leica Microsystems, Bannockburn, IL) and mounted onto Vectabond-coated glass microscope slides (Vector Laboratories, Burlingame, CA). Embedding medium was removed by 3 × 15-min washes with PBS, and nonspecific antibody binding was reduced by blocking with 10% bovine serum albumin in PBS containing 0.03% Triton X-100 for 1 h and then incubated with polyclonal anti-Best3 antibody (1:500 dilution, raised in rabbits) for 24 h at 4°C. Sections were then washed with PBS (2 × 10 min) and incubated in Alexa 488 conjugated goat anti-rabbit IgG secondary antibody (Molecular Probes, Invitrogen) at 5 μg/ml for 1 h at room temperature. After incubation with the secondary antibody, tissues were washed in PBS (3 × 15 min) and mounted using Vectashield aqueous mounting medium (Vector labs). Sections incubated without mBest3 antibody were used as negative controls.

For immunohistochemical analysis of cardiac myocytes, single atrial and ventricular myocytes from a mouse heart were isolated using a Langendorff-type apparatus followed by enzymatic dispersion technique as previously described (6). Isolated cardiac myocytes were fixed with 4% paraformaldehyde in PBS (pH 7.4) for 20 min at 4°C. Single immunofluorescence labeling was carried out as described above except that the mBest3 antibody was diluted in PBS containing 0.03% Triton X-100 to aid with antibody penetration of cell during incubation.

For dual immunofluorescence labeling of transfected HEK-tsA201 cells expressing Best3-c-myc fusion constructs, cells were plated onto glass coverslips and fixed in 100% acetone for 10 min and then washed 3 × 10 min with PBS. Fluorescent immunohistochemical staining with mBest3 antibody was carried out as described above. After the final wash, slides were not mounted but incubated with a second primary antibody, mouse monoclonal c-myc (AbCam, Cambridge, MA) 1:2,000 dilution overnight at 4°C. Cells were washed with PBS (2 × 10 min) and incubated with Alexa 594 conjugated chicken anti-mouse IgG secondary antibody (Molecular Probes, Invitrogen) at 5 μg/ml for 1 h at room temperature. Control sections for double-labeling immunohistochemistry were prepared by omitting the first primary antibody in one preparation, the second primary antibody in a second preparation, and both antibodies in a third preparation. Immunofluorescently stained preparations were examined and documented on a Radiance 2100 laser scanning confocal microscope (Bio-Rad Laboratories) with appropriate excitation wavelengths for Alexa 488 and Alexa 594. The confocal microscope is also equipped with a transmitted light detector allowing simultaneous acquisition of DIC and fluorescent images. Digital z-series and composites were obtained using LaserSharp software (Bio-Rad). Further image manipulation or analysis was carried out using Simple PCI image analysis software (Compix, Cranberry Township, PA), and final images were prepared using Adobe Photoshop software.

Electrophysiological methods.

Macroscopic currents were recorded at room temperature in the whole cell configuration using an Axopatch 200A amplifier. Patch pipettes were made of borosilicate glass (Sutter Instrument, Novato, CA), pulled with a Sutter P-57 puller (Sutter Instrument) and fire-polished. Voltage-clamp protocols were generated and analyzed using pCLAMP 9.2 software (Axon Instruments, Union City, CA). In all experiments, cells were held at −50 mV (V_h), and currents were evoked by pipette solutions containing 10 mM BAPTA and free [Ca²⁺] set to values ranging from 10 to 1,000 nM by the addition of 0.84–8.7 mM CaCl₂ as determined by the calcium chelator program EQCAL (Biosoft, Cambridge, UK). Free [Ca²⁺] was verified independently using a Ca²⁺-sensitive electrode (model 93-20, Thermo Orion) using calibrated solutions (CALBUF-2; World Precision Instruments, Sarasota, FL). In initial experiments, currents were recorded with either <1 nM or 500 nM Ca²⁺ using the pipette solution containing 10 mM BAPTA described above. When the current was found to be Ca²⁺ sensitive, subsequent experiments on transfected cells were recorded using a pipette solution containing a range of Ca²⁺ concentrations (<1 nM, 100 nM, 250 nM, 500 nM, and 1 μM) to determine the K_D for Ca²⁺. For recording steady-state current voltage (I-V) relationships, cells were held at −50 mV and stepped from −100 to +100 mV in 10-mV increments for 1 s. The ionic nature of the charge carrier was determined from reversal potential (E_rev) values elicited by a ramp protocol. Cells were stepped initially from V_h to a depolarized potential (+60 mV) and then stepped to voltages between −100 mV and +100 mV (10-mV increments) followed by a repolarizing step to −60 mV (total duration 1 s). The shift in reversal potential produced by replacement of the normal external solution with one containing either sodium thiocyanate (NaSCN), sodium iodide (NaI), or sodium-d-gluconate (Na-D-GA) instead of NaCl allowed us to determine the relative anion permeability of mBest3 using the Goldman-Hodgkin-Katz (GHK) equation (see Data analysis). Changes in junction potential were minimized by the use of a 3 M KCl agar bridge. All data are reported as the mean of n cells ± SE.

Solutions.

Bath and pipette solutions were chosen to facilitate the recording of Cl⁻ current in isolation. The external solution used to superfuse the cells contained (in mM) 126 NaCl, 10 HEPES, 20 glucose, 1.8 CaCl₂, 1.2 MgCl₂, and 10 TEA-Cl, and pH was set to 7.35 with 10 M NaOH. The pipette solution used to activate Cl⁻ currents contained (in mM) 20 TEA-Cl, 106 CsCl, 5 HEPES, 10 BAPTA, 3 ATPdiNa, 0.2 GTP.diNa, and 0.5 MgCl₂, and pH was set to 7.2 by addition of CsOH. Chloride currents were studied with intracellular Ca²⁺ concentration ([Ca²⁺]_i) clamped at known concentrations in the pipette solution as described above. In experiments where the external Cl⁻ was reduced, Cl⁻ was replaced by an equimolar concentration (126 mM) of the monovalent anions thiocyanate (SCN⁻), iodide (I⁻), or d-gluconate⁻. DIDS and niflumic acid (NFA) were prepared as stock solutions of 50 and 100 mM, respectively, in dimethyl sulfoxide before working solutions were made. All chemicals were obtained from Sigma-Aldrich (St. Louis, MO).

Data analysis.

All data were accrued from n cells with error bars representing the means ± SE. Reversal potentials and pharmacological percentage block were determined by curve fitting of individual current traces using Clampfit (PClamp, version 9.2; Molecular Devices) All data were first pooled in Excel and means exported to Origin 7.5 (OriginLab, Northampton, MA) or PRISM 3.0 (Graphpad Software) software for plotting and curve fitting. PRISM 3.0 was used to determine statistical significance between groups with one-way or two-way ANOVA tests. P < 0.05 was considered to be statistically significant. Graphs and current traces were exported to CorelDraw 12 (Corel, Fremont, CA) for final processing of the figures. Relative permeability of mBest3 was determined by measuring the shift in reversal potential (E_rev) upon changing the bath solution from one containing 126 mM Cl⁻ to another with 126 mM of another monovalent anion, where X is the substitute anion. The permeability ratio was estimated using the Goldman-Hodgkin-Katz equation: P_X/P_Cl = [Cl⁻]_i/[X]_o * exp (z ΔE_revF/RT), where ΔE_rev is the difference between the reversal potential with the test anion X and that observed with symmetrical Cl⁻, z = −1 and F, R, and T have their normal thermodynamic meanings. The calcium dependence of I_ClCa elicited by mBest3 expression was determined at each test potential by plotting the mean conductance at −80 mV and +80 mV of the ramp from n cells against pipette [Ca²⁺] and fitting the data with the Hill equation lacking fitting constraints on upper and lower asymptotes. For this equation, Y is the Cl⁻ conductance (nS/pF), Y_max is the maximal conductance, K_d is the apparent binding affinity constant, η is the Hill coefficient, and c is a constant: Y = Y_max*[Ca²⁺]^η/(K_d^η + [Ca²⁺]^η) + c.

RESULTS

Tissue expression of Bestrophin family in mouse and human hearts.

Bestrophins have recently been proposed to constitute a new family of Ca²⁺-activated Cl⁻ channels (5, 30, 33–36, 41, 46, 48, 49). We investigated whether human Bestrophin transcripts [hBest1-4; (24, 29, 42)] and murine Bestrophin transcripts [mBest1-3; (20)] are expressed in the heart. It should be noted that a recent correction in the genome database renamed mBest4 (mVMD2-L3) as mBest3 (33, 37), presumably to simplify the nomenclature for the Bestrophin family by adopting a similar denomination for orthologous sequences. Indeed, the previously named mBest4 has more overall homology with hBest3 (VMD2-L2) than hBest4 (VMD2-L3), therefore, mBest4 has now been renamed as mBest3 and the mBest3 pseudogene is renamed as mBest4.

We performed RT-PCR on RNA prepared from the mouse and human heart. Gene-specific primers for mouse mBest1-3 and human Best1-4 were designed to be nonhomologous to other members of the Bestrophin family (Table 1). Figure 1 shows the gel analysis of our results. Detectable transcripts for murine Bestrophins (Best1-3) (Fig. 1A) and human Bestrophins (BEST1-4) (Fig. 1B) were detected. The amplification products were sequenced to confirm that PCR produced specific amplification for each of the Bestrophins. A nontemplate control reaction was included to control for primer contamination. In addition, RT-PCR with GAPDH primers was performed on both the mouse and human heart cDNA to control for the integrity of the cDNA and for genomic DNA contamination, since these primers were designed to span an intron (data not shown).

Fig. 1.

Bestrophin transcripts are expressed in the mouse and human heart. Bestrophin expression was determined by RT-PCR on RNA prepared from mouse and human hearts using specific primers designed against each of the mouse and human Bestrophins. RT-PCR products were resolved alongside a 100-bp marker. Amplicons for murine Bestrophins Best1 (341 bp), Best2 (462 bp), Best3 (363 bp) (A) and human Bestrophins BEST1 (330 bp), BEST2 (319 bp), BEST4 (346 bp), and BEST3 (284 bp) (B) were observed. C: real-time quantitative PCR analysis of mBest1-3 expression in mouse heart. Representative amplification plots for GAPDH and each mBest subtype is indicated. Cycle threshold (C_T) values were determined from the threshold intercept marked on the figure. D: expression of murine Bestrophin transcripts was calculated from C_T values using the standard curve analysis method and reported relative to GAPDH expression within the same tissue (arbitrary units). Results are expressed as means ± SE (n = 4 heart preparations). The expression of mBest3 is significantly greater than mBest1 and mBest2 (P < 0.05).

Real-time quantitative PCR analysis was performed to quantify Bestrophin transcript expression in mouse heart tissue. Bestrophin expression was calculated relative to the expression of an endogenous control housekeeping gene (GAPDH) in the same tissue. The relative transcriptional expression of the murine Bestrophins in the mouse heart (expressed as mBest/GAPDH) was determined. Representative amplification plots for GAPDH and each mBest subtype is indicated in Fig. 1C. mBest mRNA expression relative to GAPDH (arbitrary units) was calculated to be 0.000044 ± 0.000023 for mBest1, 0.000023 ± 0.000011 for mBest2, and 0.002115 ± 0.001355 for mBest3 (means ± SE, n = 4 heart preparations). These data are represented by a bar graph in Fig. 1D. ANOVA statistical analysis was performed using cDNA from at least four different isolations. The expression levels of mBest1 and mBest2 were significantly less than those of mBest3 (P < 0.05).

Cloning of mBest3 from mouse hearts.

mBest3 was cloned from mouse hearts by PCR amplification with gene-specific primers designed from identified Genbank sequences, which spanned the open reading frame of the mBest3 gene. The mBest3 cDNA was cloned into the mammalian expression vector pcDNA3.1/TOPO, and several recombinant clones were sequenced. The nucleotide sequence of mBest3 cloned from mouse hearts is 100% identical to the Genbank accession no. NM_001007583. The deduced amino acid sequence of mBest3 is shown in Fig. 2 aligned with hBest3 (GenBank accession no. NM_032735). Vector NTI software was used to compare the protein sequences of mouse and human Bestrophin 3 orthologs, and the percentage amino acid identity was calculated as 98% in the NH₂-terminus and 66% in the COOH-terminus (overall identity is 83%). The division between NH₂- and COOH-terminal domains corresponds to amino acid 360 (49). Putative transmembrane domains (TMD) were determined from hydropathy analysis using the TMHMM Server version 2.0 software program (http://www.cbs.dtu.dk/services/TMHMM/). The putative TMD regions are underlined in Fig. 2. These TMDs are highly conserved among the Bestrophin family members.

Fig. 2.

Alignment of Bestrophin 3 sequences. Amino acid sequence alignment of mBest3 cloned from the mouse heart with hBest3 (Genbank accession no. NM_032735). Identical amino acids are shaded in grey. The putative hydrophobic transmembrane regions are underlined.

Characterization of mBest3 antibody.

To examine mBest3 protein expression in the heart, we generated a polyclonal antibody raised against a COOH-terminal peptide of mBest3. The COOH-terminus is optimal for antibody design as it displays poor conservation across Bestrophin family members. Indeed the epitope sequence (residues ⁶⁵⁵TGESPKGTPQRPRTWF⁶⁶⁹) is unique to mBest3 and is not present in the other murine Bestrophin homologs (mBest1 and mBest2) or human Best3 (Fig. 2). mBest3 protein expression in a number of mouse tissues was investigated by Western blot analysis. Figure 3A shows a strong mBest3 immunoreactive signal in all mouse tissues examined, including the mouse heart. mBest3 protein migrates at 78 kDa, which is close to the predicted molecular weight of 76.4 kDa. Figure 3B shows that no signal was obtained when rabbit prebleed sera is used instead of the mBest3 antibody confirming that the band obtained is specific to mBest3. Further controls for mBest3-specific labeling was performed by preincubating the primary antibody with the control peptide at varying concentrations (1:1, 10:1, 100:1 vol/vol; antigen:antibody) (Fig. 3C, i-iii). The 78-kDa mBest3 band disappeared in a concentration-dependent manner. Treating with excess control peptide saturates the binding capacity of the mBest3 antibody, which results in unlabeled Bestrophin on the Western blot. The reactivity and specificity of mBest3 antibody was also assessed by immunofluorescence of mBest3 expression in transfected HEK cells (Fig. 4). HEK-tsA201 cells were transiently transfected with mBest3-pcDNA/TO/c-myc-His plasmid, resulting in the expression of mBest3-c-myc fusion protein. Cells were subsequently dual labeled with mBest3 antibody (Fig. 3A, green fluorescence) and a commercially available c-myc antibody (Fig. 3B, red fluorescence). The colocalization of mBest3 and c-myc reactivity confirms the specificity of the mBest3 polyclonal antibody (Fig. 3C). The merged differential phase-contrast (DIC) and confocal images indicate that HEK-tsA201 cells not expressing mBest3-c-myc are unreactive for both antibodies (Fig. 3D).

Fig. 3.

Western blot analysis of mBest3 protein expression. A: protein lysates (50 μg) from mouse tissues were fractionated on 10% SDS gels. Standard Western blotting procedure was performed using mBest3 rabbit polyclonal antiserum. mBest3 immunoreactivity was detected as a single band that migrated at 78 kDa. B: similar protein blot incubated with rabbit preimmune serum. C: Western blots were performed mouse protein lysates with mBest3 antibody that was preincubated with the immunization peptide at varying concentrations (1:1, 10:1, 100:1 vol/vol; antigen:antibody). The 78-kDa mBest3 band disappeared in a concentration-dependent manner.

Fig. 4.

Analysis of mBest3 protein expression. Immunofluorescence analysis of mBest3-c-myc fusion protein expression in transfected HEK-tsA201 cells. Cells were dual labeled with mBest3 antibody (A, green fluorescence) and a commercially available c-myc antibody (B, red fluorescence). C: colocalization of mBest3 and c-myc. D: merged DIC and confocal images indicate that only HEK cells expressing mBest3-c-myc are reactive for both antibodies. Scale bar = 10 μm.

mBest3 protein expression and localization in mouse hearts.

mBest3 localization in the heart was examined in longitudinal sections of the mouse heart. Figure 5 shows confocal micrographs of the mouse heart immunolabeled with mBest3 antibody. Intense staining was observed within all regions of the atrium (Fig. 5A) and ventricle (Fig. 5B), including the epicardium, myocardium, and endocardium. Labeling with secondary antibody alone was included as a control (Fig. 5C). Single atrial and ventricular myocytes were isolated from mouse hearts by enzymatic dispersion, and immunofluorescence labeling with mBest3 antibody was carried out. The DIC images identify cardiac myocytes by their oblong shape and clear cross striations. Confocal imaging shows a ubiquitous distribution of mBest3 in both atrial and ventricular myocytes (Fig. 6, A and B, respectively). Higher magnification confocal micrographs reveal mBest3 immunoreactivity as a punctate staining pattern with distinct localization of Best3 near the membrane of ventricular myocytes (arrows in Fig. 6C).

Fig. 5.

mBest3 immunoreactivity in mouse heart. Confocal micrographs taken from 10-μm cryosections of a mouse heart showing immunolabeling with mBest3 antibody (Alexa 488, green fluorescence). mBest3-like immunoreactivity shows intense staining within the epicardium (EP), myocardium (MC), and endocardium (EN) of the heart. The control is secondary antibody alone. Scale bar = 100 μm.

Fig. 6.

mBest3 distribution in mouse isolated atrial and ventricular myocytes. Differential phase-contrast (DIC) and confocal z-stack images showing mBest3 immunoreactivity in isolated murine atrial (A) and ventricular myocytes (B). Propidium iodide (red fluorescence) was used as a nuclear counterstain. The boxes in B indicate the regions of the ventricular myocyte from which higher magnification confocal micrographs were obtained (C), showing that mBest3 is present at the membrane of cardiac myocytes (arrows).

mBest3 induces I_ClCa when expressed heterologously.

To confirm the role of mBest3 as a Ca²⁺-activated Cl⁻ channel, we used whole cell patch-clamp technique to examine the biophysical properties of mBest3 cloned from the heart. mBest3 was heterologously expressed in mammalian HEK293 cells because this cell line has been reported to have little or no endogenous I_ClCa (12, 28, 34–36, 46). Previously, we experienced problems when stably expressing chloride channels, such as mCLCA1 in HEK cells (7) over long periods of time. To minimize this effect, we used a tetracycline-regulated expression (T-REx) system to characterize our cloned mBest3 channel. This system allows for rapid and efficient tetracycline-inducible expression of proteins in mammalian cell lines and has been successfully used for the functional expression of K⁺ channels (47). mBest3 channel expression is therefore under the control of a competition between tetracycline repressor (TR) protein constitutively expressed in HEK293 cells (TRex-293) and exogenously added tetracycline. mBest3 channel expression was induced by addition of tetracycline (1 μg/ml) to the culture media. Whole cell patch-clamp experiments were performed 3–6 h later.

Under conditions prohibiting the activity of K⁺ channels, chloride currents were studied with free [Ca²⁺]_i clamped at known concentrations in the pipette solution, a technique previously used by several groups in different cell types, such as studies to activate I_ClCa in smooth muscle (15), endothelial (26), and parotid acinar cells (2). Expression of mBest3 yielded a Cl⁻ current in the presence of 500 nM Ca²⁺ in the pipette solution, which was time and voltage independent and displayed very slight outward rectification (Fig. 7B). These biophysical properties are similar to those of mBest2 and mBest3 shown previously (30, 33, 34, 41). Eight of ten cells from the same stable line expressed currents of similar size and this is in agreement with the high efficiency of tetracycline induction (47). Control experiments were carried out on both untransfected TRex-293 cells and TRex-293 cells transfected with Bestrophin but not induced with tetracycline (Fig. 7A). Neither of these control experiments displayed significant I_ClCa. To examine the absolute requirement for the presence of intracellular Ca²⁺ for activation of this current, cells were treated with tetracycline and dialyzed with 10 mM BAPTA with no added Ca²⁺ (free [Ca²⁺] <1 nM). These cells failed to exhibit current and resembled the recordings made from untransfected cells. The amplitude of the current evoked by 500 nM Ca²⁺ in transfected cells was significantly larger at all potentials (P < 0.001) than currents evoked by the same pipette solution in untransfected cells, transfected cells without tetracycline induction, or in transfected cells dialyzed with a pipette solution containing <1 nM Ca²⁺. This is reflected in Fig. 7C, which shows the mean I-V relationships for currents evoked with 500 nM Ca²⁺ in the pipette, TRex-293 untransfected cells (n = 8), TRex-293 transfected with mBest3 in the absence of tet induction (n = 4), TRex-293 transfected with mBest3 after tet induction (n = 13) and currents evoked with <1 nM Ca²⁺ in TRex-293 transfected with mBest3 (n = 11). The I-V relationship of our expressed mBest3 channel is almost linear and resembles that seen for hBest1, mBest2, and both the full-length and a truncated form of mBest3 (30, 33, 34, 41, 49) but contrasts with that of hBest3 (the mBest3 human ortholog), which strongly rectifies in the inward direction (49). Figure 7D shows a summary of the mean current densities measured at −80 and +80 mV. These data show that mBest3 current displays slight voltage dependence between −80 mV (−34.5 ± 5.5 pA/pF) and +80 mV (42.2 ± 6.3 pA/pF) (n = 13, P < 0.0001).

Fig. 7.

Whole cell currents of mBest3 expressed in TRex-293 cells. Families of currents evoked by pipette solutions containing 500 nM Ca²⁺ from TRex-293 cells transfected with mBest3 in the presence (B) and absence of tetracycline (A) at voltages between −100 and +100 mV. The voltage protocol is shown as inset in A. C: mean current-voltage relationship for currents evoked by 500 nM Ca²⁺ in untransfected TRex-293 cells (n = 8), TRex-293 transfected with mBest3 in the absence of tet induction (n = 4), TRex-293 transfected with mBest3 after tet induction (n = 13), and currents evoked with <1 nM Ca²⁺ in TRex-293 transfected with mBest3 (n = 11). Measurements were taken immediately before the end of each test potential. Each point is the mean of 4–13 cells, error bars represent means ± SE. Currents were recorded 3–6 h after tetracycline induction. mBest3 conductance is time independent and displayed slight outward rectification. The current reversed at 0 mV. A summary of the mean current densities recorded at −80 and +80 mV is shown in D.

Expression of mBest3 in TRex-293 cells led to the induction of I_ClCa that was never detected in untransfected cells. Although this suggests that mBest3 carries this chloride current, it is also possible that mBest3 may upregulate the expression of an endogenous protein that results in I_ClCa. To rule out this possibility, we transiently transfected COS-7 cells with our mBest3 cDNA and performed patch-clamp experiments identical to those shown in Fig. 7. We believed that if the biophysical properties of the mBest3 current were the same when mBest3 was expressed in different mammalian cell line, this would reinforce the likelihood that mBest3 encodes the current. Expression of mBest3 in COS-7 cells (Fig. 8B) evoked a I_Cl similar to that expressed in TRex-293 cells transfected with mBest3. Similar to untransfected TRex-293 cells, untransfected COS-7 cells (Fig. 8A) failed to display this current. Figure 8C shows the mean I-V relationships for transfected and untransfected cells (n = 5). As for TRex-293 cells, expression of mBest3 in COS-7 cells led to the appearance of a time-independent Ca²⁺-activated Cl⁻ conductance, which displayed slight voltage dependence and whose linear I-V relationship reversed near the predicted equilibrium potential for Cl⁻.

Fig. 8.

Whole cell currents of mBest3 expressed in COS cells. Representative families of currents evoked by pipette solutions containing 500 nM Ca²⁺ from COS-7 cells transiently transfected with mBest3 (B) and untransfected COS-7 cells (A). The voltage-clamp (I-V) protocol is shown as inset in A. C: mean I-V relationship for currents evoked by 500 nM Ca²⁺ in untransfected and mBest3-transfected COS-7 cells. Each point is the mean of 5 cells. mBest3 conductance is time independent and displayed slight outward rectification. The current reversed at 0 mV.

Ca²⁺ sensitivity of mBest3-induced currents.

Typical families of Cl_Ca currents generated by expression of mBest3 in TRex-293 cells were recorded from different cells with one of four [Ca²⁺]_i are shown in Fig. 9, A–D. [Ca²⁺]_i was set using the Ca²⁺ chelator BAPTA (free [Mg²⁺] set to 0.5 mM) and using the same voltage protocol as in Fig. 7. When [Ca²⁺]_i was <1 nM, the currents recorded were very small (Fig. 9A) consistent with the data presented in Fig. 7. Pipette solutions containing 100 (Fig. 9B), 250 (Fig. 9C), or 1,000 nM (Fig. 9D) [Ca²⁺]_i led to increasingly larger currents. Figure 9E shows a summary of the mean I-V relationships of mBest3 currents evoked with <1 nM, 100 nM, 250 nM, 500 nM, and 1 μM Ca²⁺. All currents generated with different [Ca²⁺] in the pipette displayed a linear I-V relationship and reversed close to zero as expected for a Cl⁻-selective channel. To determine the Ca²⁺ sensitivity of mBest3 currents, dose-response relationships were obtained by plotting current density measured at +80 and −80 mV as a function of [Ca²⁺], and these were fitted to the Hill equation (solid lines) in Fig. 9F. At −80 mV, K_D was 161.9 ± 49.2 nM Ca²⁺, and Hill coefficient (n_H) was 2.6 ± 1.5, whereas at +80 mV K_D was 174.9 ± 51.8 nM Ca²⁺, and n_H was 2.1 ± 0.9. These K_D values are significantly different (P = 0.03) implying that the Ca²⁺ sensitivity displays slight voltage dependence.

Fig. 9.

Intracellular Ca²⁺ sensitivity of mBest3 currents expressed in TRex-293 cells. Whole cell voltage-clamp recordings were obtained from TRex-293 cells stably expressing mBest3 with pipette solutions containing various free [Ca²⁺]. Representative currents evoked with <1 nM free Ca²⁺ (nominally 0 Ca²⁺) (A), 100 nM (B), 250 nM (C), or 1 μM free Ca²⁺ (D). Voltage steps of 1-s duration were applied from a holding potential of −50 mV to voltages between −100 and +100 mV in 10-mV steps. E: mean current-voltage relationships of currents obtained with [Ca²⁺]_i of <1 nM (squares, n = 11), 100 nM (circles, n = 4), 250 nM (triangles, n = 8), 500 nM (inverted triangles, n = 13), and 1 μM (side triangle, n = 9). F: dependence of mBest3 currents on intracellular [Ca²⁺], at −80 mV (squares), and +80 mV (circles). Amplitude of the average current densities were plotted versus [Ca²⁺] and fitted to the Hill equation. Each point is the mean of between 4 and 13 cells with error bars representing means ± SE.

Anion selectivity of mBest3-induced currents.

To confirm that the expressed mBest3 current is indeed carried by chloride, ion replacement studies were carried out. The measured reversal potential (E_rev) of mBest3 current (E_rev = 0.83 ± 0.9, n = 12) agrees with the calculated Nernst equilibrium potential for Cl⁻ of 0 mV with 126 mM Cl⁻ on both sides of the membrane. In the majority of experiments the major extracellular cation was Na⁺ and the major intracellular cation was Cs⁺. Therefore, if the mBest3 channel was permeable to these cations, the measured E_rev would be significantly different from zero. mBest3 current exhibited a lyotropic sequence of anion permeability, SCN⁻>I⁻>Cl⁻ (Fig. 10) similar to many other Cl⁻ channels. The relative permeability of external ions (X) with respect to Cl⁻ (P_x/P_Cl) was estimated by the shift in E_rev of the current under bionic conditions and calculated using the Goldman-Hodgkin-Katz equation. The bar graph in Fig. 10B shows P_x/P_Cl for Cl⁻, SCN⁻ (n = 10), I⁻ (n = 10), and d-gluconate⁻ (n = 7). The P_x/P_Cl was 3.5:1.3:1:0.3 for SCN⁻, I⁻, Cl⁻, d-gluconate⁻, respectively. The bar graph in Fig. 10C depicts the average relative slope conductance ratios (G_X/G_Cl) obtained from the measured slopes of the I-V relationships between −50 and +50 mV for experiments similar to that displayed in Fig. 10A. Substitution of extracellular Cl⁻ with SCN⁻ produced a significant increase in conductance (1.7 ± 0.1), whereas substitution with I⁻ only slightly decreased the conductance (0.9 ± 0.01) (Fig. 10C). Partial replacement of extracellular Cl⁻ with d-gluconate⁻ decreased the channel conductance as would be expected. With the use of a one-way ANOVA Dunnett's comparison test, P values were calculated for the following: G_SCN/G_Cl, P < 0.001; G_I/G_Cl, P > 0.05; and G_D-Gluconate/G_Cl, P < 0.001. Therefore, only the average slope relative conductance ratios of G_I/G_Cl were not significantly different from each other.

Fig. 10.

Ionic nature of the current evoked in TRex-293 cells expressing mBest3. Whole cell mBest3 currents activated by 500 nM Ca²⁺ in the pipette solution were measured with bath solutions containing either 126 mM NaCl, NaSCN, NaI, or Na-D-gluconate. Inset, ramp protocol used to determine the voltage dependence of the activated current under different external anionic conditions. A: effect of external anions on the I-V relationship show an anion permeability of SCN>I>Cl. B: relative permeability ratios (P_x/P_Cl) were calculated with the GHK equation from measured differences in reversal potentials between Cl⁻ and other substituted anions and show an anion permeability of SCN⁻>I⁻>Cl⁻. C: average slope relative conductance ratios (G_x/G_Cl) were obtained from the measurement of the slope of the I-V relationship between −50 and +50 mV from reversal potentials. Error bars represent means ± SE.

Pharmacology of mBest3-induced currents.

We investigated the pharmacological profile of mBest3-expressed currents by investigating the effects of niflumic acid (NFA) and DIDS, both of which have previously been shown to block I_ClCa in a variety of cell types (13, 14, 31, 50), including murine (51), rabbit (19, 45, 54), and canine (9, 53) cardiac muscle cells. Figure 11 shows the effect of NFA (A) and DIDS (C) on typical ramp currents recorded from TRex-293 cells stably expressing mBest3 in the presence of 500 nM Ca²⁺ in the pipette solution. The ramp protocol is shown as an inset. Mean data for such similar experiments are reported in Fig. 11, B and D. NFA (100 μM) blocked the current by 64.7 ± 4.3% at −80 mV and 71.2 ± 4.6% at +80 mV (B; n = 4), whereas 300 μM NFA exerted a slightly higher percentage block at both voltages, 76.3 ± 6.4% at −80 mV and 79.9 ± 6.9% at +80 mV. DIDS (100 μM) blocked the mBest3 current by 70.7 ± 5.7% and 82.9 ± 4.7% at −80 mV and +80 mV, respectively (D; n = 6). Block of mBest3 currents by NFA was not significantly voltage dependent (100 μM, P = 0.1888, 300 μM, P = 0.1407). However, the block by 100 μM DIDS displayed a slight voltage dependence (P = 0.0017).

Fig. 11.

Niflumic acid and DIDS inhibit mBest3 currents. Effect of NFA (A) and DIDS (C) on whole cell currents recorded from TRex-293 cells stably expressing mBest3. Cells were voltage clamped with 500 nM intracellular free Ca²⁺, and the I-V relationships were obtained with the protocol shown. Bar graphs show the mean percentage block of mBest3 current by 100 μM (n = 4) and 300 μM (n = 4) NFA (B), and 100 μM DIDs (n = 6) (D) at −80 and +80 mV. Error bars represent means ± SE.

Mutagenesis alters mBest3 permeability and pharmacology.

To strengthen our observation that mBest3 is a Ca²⁺-activated Cl⁻ channel and represents the pore-forming subunit responsible for the current evoked in mBest-transfected cells, we performed site-directed mutagenesis and patch-clamp analysis of the mutated channel. The conserved putative second transmembrane domain (TMD2) of Bestrophin proteins is thought to play an important role in channel permeation (32, 34, 35, 49). Similar to the mutagenesis studies of this region in mBest2 (32, 34, 35, 49), we hoped to introduce a mutation that could alter the anion permeability or the conductance of the channel. We aligned the highly conserved putative TMD2 of all the murine and human Bestrophins (amino acids 75–94) (Fig. 12A). A number of mutations in this region have been identified in patients with vitelliform macular dystrophy (VMD; Best disease), (43), the human disease from which Bestrophin was originally identified and positionally cloned. However, expression of some of these mutants in heterologous cell systems fail to generate Cl_Ca (36). We decided to mutate mBest3 phenylalanine at position 80 (F80) to a lysine (L), since F80L is a naturally occurring single nucleotide polymorphism (SNP) in some Best disease families (43), and therefore is a physiologically relevant mutant. In addition, the serine at position 79 in mBest2 has been reported as a critical residue of the pore region as this residue has previously been shown to affect the conductance selectivity of mBest2 (34, 35). However, threonine, a structurally similar amino acid occupies this position in mBest3 and hBest3 (see Fig. 12A).

Fig. 12.

The mBest3 F80L mutation alters the ionic nature and pharmacology of mBest3 currents. A: alignment of the conserved putative second TMD region of murine and human Bestrophin proteins. Identical amino acids are shaded in black, numbers correspond to mBest3 amino acid sequence. The F80L mutation is indicated. B: F80L mutant significantly decreases the permeability of mBest3 to SCN⁻, wild-type (n = 8), and F80L (n = 10). Experimental protocols are indicated. C: effect of 100 μM DIDS on currents recorded from HEK-TsA201 cells expressing mBest3 wild-type (n = 6) or mBest3 F80L mutant (n = 4). D: DIDS (100 μM) was applied to the bath solution and the percent inhibition of currents at −80 mV and +80 mV was measured. The block of mBest3 F80L current by DIDS is significantly voltage dependent.

mBest3-F80L mutant was transiently expressed in HEK-tsA201 cells and assessed by the whole cell patch-clamp technique. We found that this mutation had a major impact on the shift in E_Rev and this in turn resulted in a decrease in the P_SCN/P_Cl compared with wild-type current (Fig. 12B). P_SCN/P_Cl of mBest3 wild-type was 3.5 ± 0.24 (n = 8) and mBest3-F80L mutant was 1.6 ± 0.04 (n = 10). The permeability to I⁻ was tested but this did not significantly change (P_I/P_Cl of mBest3 wild-type was 1.33 ± 0.03, n = 10 and mBest3-F80L mutant was 1.34 ± 0.03, n = 7). The average slope conductances measured between −50 and +50 mV also did not change significantly G_SCN/G_Cl of mBest3 wild-type: 1.7 ± 0.05 and F80L: 1.8 ± 0.28 (data not shown). A mutation of mBest2 at position 80 (phenylalanine to arginine, F80R) was reported to influence the pharmacology of the channel (35). We thus investigated whether the F80L mutation altered the inhibition of mBest3-induced current by DIDS and NFA. The inhibition of mBest3-F80L current by 100 μM NFA was not significantly different from wild-type (data not shown; n = 4). Interestingly, F80L had a very prominent effect on the block by DIDS. Figure 12C shows the effect of 100 μM DIDS on currents recorded from HEK cells expressing mBest3-F80L with 500 nM Ca²⁺ in the pipette solution. The average percentage block by 100 μM DIDS at −80 mV and +80 mV for both mutant and wild-type is plotted in Fig. 12D. DIDS blocked the mBest3-F80L current 12.9 ± 2.3% at −80 mV and 67.2 ± 3.7% at +80 mV (n = 4), suggesting that the F80L mutation converted the slightly voltage-dependent block of wild-type mBest3 by DIDS to one that is significantly voltage dependent (P < 0.05).

DISCUSSION

In summary our data have shown that mBest3 cloned from mouse heart produces a DIDS- and niflumic acid-sensitive Ca²⁺-dependent Cl⁻ current when expressed heterologously in different cell types. mBest3 protein is expressed in both murine cardiac myocytes and at the plasma membrane of mBest3-transfected HEK cells. The currents are anion selective with a permeability profile of SCN⁻>I⁻>Cl⁻ and are activated by physiological concentration of Ca²⁺ (K_D = ∼ 175 nM). Finally, our data supports the concept that mBest3 is the pore-forming subunit of the anion channel as a single point mutation within the speculated pore region of Bestrophins reduced the permeability of the channel to SCN⁻ and converted the relatively voltage-independent block produced by DIDS to one that is strongly voltage dependent.

Bestrophins are expressed in the heart.

Recent electrophysiological studies have provided evidence that Cl_Ca plays a significant role in the regulation of cardiac action potential duration under normal and abnormal conditions (for reviews see Refs. 18 and 40). The molecular nature of the channel underlying this current remains unknown. Bestrophins are proposed as candidates to encode Cl_Ca (17). We examined whether this family is present in the heart, a tissue known to have I_ClCa (18). Here we report the positive identification of Bestrophin transcripts in both the mouse and the human heart. We used real-time quantitative RT-PCR to determine the levels of murine Bestrophin expression relative to the housekeeping gene GAPDH in the mouse heart. Although all three murine Bestrophin genes were shown to be expressed at the mRNA level, the transcriptional levels of mBest3 were found to be much higher than those of mBest1 (48.1-fold) and mBest2 (92.0-fold), which explains why we focused this study on mBest3.

After the identification and cloning of this predominant isoform from the heart, our next goal was to determine the localization of mBest3 protein in cardiac myocytes. To accomplish this, we generated a rabbit anti-mouse Best3 polyclonal antibody. The reactivity and specificity of the mBest3 antibody was confirmed by colocalization of Best3-c-myc fusion protein expression in HEK cells and Western analysis incorporating preimmune and preabsorbed serum controls. With the use of this antibody, a unique 78-kDa band, close to the predicted molecular weight for mBest3, was recognized by Western blot analysis of several mouse tissue lysates, including the heart. Using RT-PCR, it has been reported that mBest3 has a restricted tissue expression to the heart and testis (20). We find that mBest3 is expressed in several tissues at both the transcript (data not shown) and protein level (Fig. 3). The discrepancy between our results and those of Kramer et al. (20) could be due to their RT-PCR primers binding within splice regions. We have confirmed the presence of mBest3 transcripts in tissues other than the heart by performing RT-PCR using various primers that anneal to multiple regions of the mBest3 transcript. The present study represents the first demonstration of immunolabeling of a Bestrophin protein in cardiac cells. Immunofluorescence indicates that Best3 has a ubiquitous distribution in mouse heart. Confocal analysis reveals robust staining and localization of mBest3 protein near the plasma membrane of isolated cardiac myocytes. The uniform pattern of punctate staining within the myocyte warrants further investigation. Determining the precise localization of mBest3 and its possible proximity to other channels and receptors involved in excitation contraction coupling may shed light on the physiological function of this channel in the cardiac myocyte. mBest3 membrane staining pattern was also observed in HEK cells transfected with mBest3-c-myc cDNA, a result consistent with the appearance of a plasma membrane Ca²⁺-dependent Cl⁻ conductance in our electrophysiological studies.

Biophysical and pharmacological properties of mBest3-induced I_ClCa.

Based on the substantial evidence reported to date, Bestrophins are now considered genuine Cl⁻ channels (5, 11, 30, 32–36, 41, 46, 48, 49). Their reported anion selectivity, small unitary conductance, Ca²⁺ dependence, and sensitivity to Cl⁻ channel blockers such as DIDS and niflumic acid make this protein a likely molecular candidate for the small conductance Cl_Ca expressed in a wide range of cell types including mammalian cardiac myocytes. Three murine Bestrophin homologues (mBest1-3) have been previously identified (20), and the biophysical properties of the current induced by heterologous expression of mBest2 has been studied extensively by two independent groups (30, 32, 34, 35). We successfully recorded Ca²⁺-activated Cl⁻ currents after expressing the full-length mBest3 clone in two different mammalian cell lines by two distinct methods: 1) transient transfection in HEK-tsA201 and COS-7 cells; and 2) stable expression induced by a tetracycline-inducible method. Qu et al. (33) reported that the heterologous expression of mBest3 in HEK293 cells fails to induce significant Cl⁻ currents with <1-s voltage pulses within a physiological range. This group performed mutagenesis studies that enabled them to isolate a specific region in the COOH-terminus of mBest3 (33, 37), which they showed inhibited the expressed chloride current (I_Cl). When they expressed truncated versions of this protein, with the inhibitory motif removed, they were able to induce large I_Cl, which were found to be Ca²⁺ insensitive. Based on our study, it is possible that the mutated region may be responsible for the Ca²⁺ dependence of this channel and this will require investigation. A more recent study by Melvin's group investigating mBest3 function in exocrine gland, indicated that HEK293 cells overexpressing mBest3 elicited a Ca²⁺-sensitive conductance (41). This study is interesting in that it highlights that full-length mBest3 encodes a Cl_Ca; however, unlike our study, Srivastava et al. (41) did not provide a detailed examination of the properties of mBest3 in terms of Ca²⁺ sensitivity, anion selectivity, pharmacology using the “classical” I_ClCa inhibitor NFA, or mutagenesis to investigate the conduction pore.

In this study, the Cl⁻ currents elicited by overexpression of mBest3 were time and voltage independent and displayed only slight outward rectification at the more positive potentials, which is similar to heterologously expressed hBest1, hBest2, dmBest1 (46, 49), xBest2a and b (36), and mBest2 (30, 34), which exhibited no signs of voltage-dependent activation and inactivation kinetics. This contrasts with ceBest1 (46), hBest4 (49), and in particular, hBest3 (48, 49), the human ortholog sharing 83% identity with mBest3, which rectifies inwardly and displays very slow activation kinetics (over a period of many seconds) when elicited by long hyperpolarizing steps. It should be mentioned that time-dependent I_ClCa could only be detected following the expression of mBest3 (33) and hBest3 (49) when elicited by very hyperpolarizing pulses lasting several seconds. Again, besides differences in recording methods, the potential reasons for such a discrepancy between our study and that of Qu et al. (33) remain unclear.

We examined the anion selectivity sequence of mBest3-induced current and found SCN⁻>I⁻>Cl⁻, which follows the Eisenman “weak field strength” lyotrophic series (10). This permeability sequence is identical to that described for cardiac I_ClCa and Cl⁻ currents elicited by expression of mBest2 (34) and truncated mBest3 (33). To establish that the I_ClCa that appeared when we overexpressed the full-length mBest3 cDNA was indeed attributed to this protein and not due to upregulation of some endogenous channel protein, we felt it was necessary to express an mBest3 mutant that would somehow change the electrophysiological properties of the channel. Recently, mutagenesis studies have defined specific regions of the putative second TMD of Bestrophins as important determinants for the permeability of the channel (32, 34, 35, 49). A number of these mutations inhibited channel function or altered the selectivity and conductance of the channel so it was concluded that these residues were located in the channel pore. Previously a number of different mutations of a phenylalanine amino acid at position 80 (F80) of mBest2 have been reported to effect the SCN⁻ permeability of the channel, and F80R was shown to exhibit voltage-dependent block by DIDS (35). It has previously been reported that F80L is a mutation in some Best's disease patients (44) and so we overexpressed a variation of our mBest3 protein that incorporated this mutation. We found that this amino acid substitution (F80L) reduced P_SCN/P_Cl to less than half of that measured for wild-type currents and altered the pharmacology of this channel. In the present study we used both niflumic acid and DIDS as tools to investigate the pharmacological properties of our heterologously expressed mBest3. Neither of these are specific blockers for the Ca²⁺-activated Cl⁻ conductance; however, they have been commonly used to identify and isolate I_ClCa in native cells and have been shown to inhibit Bestrophin-induced currents (17, 30). DIDS and NFA inhibited mBest3-mediated I_ClCa at concentrations similar to those used to inhibit I_ClCa. in other tissues (13). Furthermore, the F80L mutation converted the slightly voltage-dependent block exerted by DIDS to one that is significantly more voltage dependent with current inhibition enhanced by membrane hyperpolarization. Taken together, these results strongly suggest that mBest3 is the pore-forming subunit responsible for the I_ClCa evoked by mBest3 expression and is not the product of an endogenous channel being indirectly upregulated by mBest3 transfection.

Finally, our data provide convincing evidence that the mBest3-induced Cl⁻ conductance is activated within a narrow range of physiological Ca²⁺ concentrations as found for other heterologously expressed Bestrophins. Other studies reported a K_D value for Ca²⁺ in the range of 210 to 400 nM [hBest4: 230 nM (48); xBest2a: 210 nM; xBest2b: 228 nM (36), and mBest2: 230 nM (34) and 400 nM (30)]. Our K_D values for Ca²⁺ for mBest3 fall slightly below the above-mentioned range, and this could be related in part to the Ca²⁺ chelator used in our study (27). We used 10 mM BAPTA to buffer [Ca²⁺] up to the 1 μM range, whereas other studies, including a recent report on a Ca²⁺-insensitive truncated form of mBest3 (33) have all used EGTA, which may have contributed to this difference. The estimated K_D for our heterologously expressed mBest3 at +80 mV was 175 nM Ca²⁺, and there was very slight voltage dependence of Ca²⁺ activation. Also, our Hill coefficient n_H was ∼2. Hill coefficients higher than 1 suggest the presence of multiple cooperating Ca²⁺ binding sites on the channel. Interestingly, Tsunenari et al. (48) reported a n_H of less than unity for hBest4 in inside-out patches, which is quite unusual. Our estimated n_H value certainly falls along the range of n_H values estimated for I_ClCa recorded from many cell types, including Xenopus oocytes (n_H = ∼2–3) (21, 22), pulmonary arterial smooth muscle cells [n_H = ∼2–4; (1)], and pancreatic acinar cells [n_H = ∼1–2, (2)]. Whether Ca²⁺ activates the channel directly by binding to the Bestrophin protein or indirectly through some regulatory protein(s) remains to be determined.

Physiological role of bestrophin 3 in hearts.

The Bestrophin 3 chloride channel has not being assigned a physiological role. This study of mBest3 expression and cellular location in cardiac myocytes, together with biophysical properties as a calcium-dependent chloride channel, may provide clues to its function. A recent study by Matchkov et al. (25) assigned a role for Best3 as a cGMP-dependent Ca²⁺-sensitive Cl⁻ current with unique characteristics in certain smooth muscle cells. To propose a similar physiological role for Best3 in the heart is difficult because a cGMP-dependent, Ca²⁺-sensitive Cl⁻ conductance has not been reported in this tissue. Furthermore, the cGMP dependence of recombinant Best3-induced I_ClCa has not been examined.

In the heart, Ca²⁺-activated Cl⁻ channels contribute significantly to the early transient outward current responsible for phase 1 repolarization, which in turn modulates L-type Ca²⁺ current dynamics and excitation-contraction coupling (18, 38). Indeed, many of the electrophysiological properties of native cardiac I_ClCa and of mBest3 current expressed in HEK cells are similar. Both currents display the same anion permeability sequence, linear or near linear, I-V relationship with symmetrical Cl⁻ and sensitivity to chloride channel antagonists such as NFA and DIDS. Similar to mBest3 I_ClCa, cardiac I_ClCa activated by clamped Ca²⁺ is voltage and time independent. However, it is difficult to accurately characterize the kinetics of activation and inactivation of native cardiac I_ClCa due to the difficulty of studying this current in isolation, in the absence of I_Ca and Ca²⁺-induced Ca²⁺ release. When intracellular Ca²⁺ is not clamped, native I_ClCa appears voltage dependent and displays strong inactivation kinetics. Any apparent voltage dependence of I_ClCa can be attributed to the voltage-dependent Ca²⁺ channel, and any apparent time dependence is most likely due to changes in [Ca²⁺]_i in close proximity to the channel. In this study mBest3 currents displayed a K_D for Ca²⁺ of 175 nM; however, the Ca²⁺ sensitivity of the native current has not been reported. Cardiac I_ClCa is dependent on Ca²⁺ influx via voltage-dependent Ca²⁺ channels and Ca²⁺ release from the SR stores (18, 40, 51). A more precise notion about the physiological role of mBest3 in the heart awaits the characterization of the phenotype of Best3 null mice. Although we found mBest3 to be the most abundant isoform in the mouse heart, mBest1 and mBest2 were also present. The assignment of a physiological role of mBest3 in the heart may be complicated by the presence of other Bestrophin family members, since it is quite possible that cardiac Cl_Ca are heteromultimers composed of different Bestrophin subunits. It is unknown whether Cl_Ca is due to the expression of a single gene, expression of multiple members of a gene family, or the expression of different genes from multiple gene families. Indeed, three independent groups have identified recently that the TMEM16 family of genes encode novel proteins associated with Ca²⁺-dependent Cl⁻ channel activity in epithelial cells (8, 39, 52), further adding to the complexity of identifying the molecular entity of Cl_Ca. In conclusion, our combined molecular and electrophysiological data provides novel findings regarding mBest3 localization in the heart and provides a detailed characterization of recombinant mBest3 biophysical properties for future comparisons with native Cl_Ca activities.

GRANTS

This publication was made possible by grant number P 20RR 15581 (to F. C. Britton) from the National Centre for Research Resources (NCRR), a component of the National Institutes of Health (NIH) and National Heart, Lung, and Blood Institute Grant RO1 HL-091238-01A1 (to F. C. Britton) and R01 HL-075477-01 (to N. LeBlanc).