The inositol (1,4,5)-trisphosphate receptor (InsP3R) mediates Ca2+ release from intracellular stores in response to generation of second messenger InsP3. InsP3R was biochemically purified and cloned, and functional properties of native InsP3-gated Ca2+ channels were extensively studied. However, further studies of InsP3R are obstructed by the lack of a convenient functional assay of expressed InsP3R activity. To establish a functional assay of recombinant InsP3R activity, transient heterologous expression of neuronal rat InsP3R cDNA (InsP3R-I, SI− SII+ splice variant) in HEK-293 cells was combined with the planar lipid bilayer reconstitution experiments. Recombinant InsP3R retained specific InsP3 binding properties (Kd = 60 nM InsP3) and were specifically recognized by anti–InsP3R-I rabbit polyclonal antibody. Density of expressed InsP3R-I was at least 20-fold above endogenous InsP3R background and only 2–3-fold lower than InsP3R density in rat cerebellar microsomes. When incorporated into planar lipid bilayers, the recombinant InsP3R formed a functional InsP3-gated Ca2+ channel with 80 pS conductance using 50 mM Ba2+ as a current carrier. Mean open time of recombinant InsP3-gated channels was 3.0 ms; closed dwell time distribution was double exponential and characterized by short (18 ms) and long (130 ms) time constants. Overall, gating and conductance properties of recombinant neuronal rat InsP3R-I were very similar to properties of native rat cerebellar InsP3R recorded in identical experimental conditions. Recombinant InsP3R also retained bell-shaped dependence on cytosolic Ca2+ concentration and allosteric modulation by ATP, similar to native cerebellar InsP3R. The following conclusions are drawn from these results. (a) Rat neuronal InsP3R-I cDNA encodes a protein that is either sufficient to produce InsP3-gated channel with functional properties identical to the properties of native rat cerebellar InsP3R, or it is able to form a functional InsP3-gated channel by forming a complex with proteins endogenously expressed in HEK-293 cells. (b) Successful functional expression of InsP3R in a heterologous expression system provides an opportunity for future detailed structure–function characterization of this vital protein.
Release of Ca2+ from intracellular stores in response to generation of the second messenger inositol (1,4,5)- trisphosphate (InsP3)1 is a common mechanism used by many cell types to raise cytosolic Ca2+ concentration (Berridge, 1993). InsP3-induced Ca2+ release is supported by a highly specialized ion channel, the inositol (1,4,5)-trisphosphate receptor (InsP3R). InsP3Rs were biochemically purified and cloned, and functional properties of these channels were extensively studied (reviewed by Taylor and Richardson, 1991; Berridge, 1993; Furuichi et al., 1994; Bezprozvanny and Ehrlich, 1995). Type I InsP3R (InsP3R-I) was initially purified (Supattapone et al., 1988) and cloned (Furuichi et al., 1989; Mignery et al., 1990) from rodent cerebellum. InsP3R-I is a complex of four 2,749 amino acid subunits (Mignery et al., 1989; Maeda et al., 1991). Different splice variants of InsP3R-I are expressed in specific regions of the brain, at different stages of neuronal development, and in nonneuronal tissues (Nakagawa et al., 1991a, 1991b). After initial characterization of InsP3R-I, multiple isoforms of InsP3R cDNA have been identified (reviewed by Furuichi et al., 1994).
Functional properties of the InsP3R have been characterized using Ca2+ flux measurements or planar lipid bilayer recordings (reviewed by Bezprozvanny and Ehrlich, 1995). The majority of functional studies were performed with cerebellar or vascular smooth muscle InsP3R isoforms, corresponding to splice variants of InsP3R-I (Islam et al., 1996). InsP3R is a relatively nonselective cation channel (Bezprozvanny and Ehrlich, 1994), with conduction properties similar to that of ryanodine receptor (RyanR), another intracellular Ca2+ release channel. The central signaling role of InsP3R is highlighted by multiple levels of its modulation. Activity of InsP3R is affected by phosphorylation (reviewed by Ferris and Snyder, 1992), allosterically modulated by ATP (Ferris et al., 1990; Iino, 1991; Bezprozvanny and Ehrlich, 1993), and display bell-shaped Ca2+ dependence (Iino, 1990; Bezprozvanny et al., 1991; Finch et al., 1991). Putative structural determinants responsible for the InsP3R modulation have been identified using biochemical methods (reviewed by Furuichi et al., 1994).
Until now, structural and functional studies of the InsP3R largely proceed in parallel, with most of the functional work performed with the native InsP3R and recombinant InsP3R characterized mainly by biochemical methods. The main obstacle to further understanding InsP3R structure–function has been the lack of a convenient functional assay of expressed InsP3R activity. Recently, skeletal isoform of another intracellular Ca2+ release channel, RyanR, has been successfully expressed in HEK-293 and Chinese hamster ovary cell lines (Bhat et al., 1997; Chen et al., 1997). Basic single channel properties of recombinant skeletal RyanR were characterized using the planar lipid bilayer reconstitution assay (Bhat et al., 1997; Chen et al., 1997). We took a similar approach and set out to establish an assay of recombinant InsP3R function by combining planar lipid bilayer reconstitution technique with transient expression of neuronal rat InsP3R-I cDNA in HEK-293 cells. We achieved high level functional expression of InsP3R-I in HEK-293 cells and compared most basic functional properties of recombinant InsP3R-I with the analogous properties of native rat cerebellar InsP3R. Successful functional expression of InsP3R in heterologous expression systems provides an opportunity for future detailed structure–function characterization of this vital protein.
Materials And Methods
InsP3R-I Expression Methods
The full length neuronal rat InsP3R-I cDNA clone in pCMV expression vector (pCMVI-9) (Mignery et al., 1990) was generously provided by Dr. Thomas Südhof (HHMI, Department of Molecular Genetics, UT Southwestern Medical Center at Dallas). This clone corresponds to SI− SII+ splice variant of InsP3R-I (Mignery et al., 1990). We confirmed the fact that our working InsP3R-I clone is indeed SI− SII+ by sequencing corresponding regions. To increase efficiency of InsP3R-I expression in HEK-293 cells, InsP3R-I coding sequence was subcloned into pcDNA3 (Invitrogen Corp., San Diego, CA) expression vector and 5′ untranslated region of the original pCMVI-9 clone was substituted with the Kozak consensus sequence (Kozak, 1987). To achieve this, we (a) subcloned a 7.9-kb Acc65I/XbaI fragment of pCMVI-9 into Acc65I/XbaI-digested pcDNA3 vector (InsP3R-pcDNA3-A/X); (b) amplified the 5′ end of the rat InsP3R-I clone by 15 cycles of PCR using high fidelity enzyme PfuI (Stratagene Inc., La Jolla, CA) with CGG GGT ACC GCC ACC ATG TCT GAC AAA ATG TCT A and CCG AAC CTC AGC AGG AGA AAC pair of primers. The resulting 1.3-kb PCR fragment was isolated, digested with KpnI, and cloned into KpnI-digested and -dephosphorylated InsP3R-pcDNA3-A/X clone. The correct orientation of PCR fragment insertion into the KpnI site was verified by PCR using the same pair of primers, and the 5′ end of the clone with correct orientation of the fragment was sequenced from ATG to the KpnI site (1,255) in both directions to rule out the possibility of mutations introduced at the PCR step. The resulting construct (InsP3R-pcDNA3) was used in HEK-293 cell transfection experiments. Some of the experiments were performed with InsP3R-pCEP4 clone, constructed in a similar way except that pCEP4 expression vector (Invitrogen Corp.) was used instead of pcDNA3.
Human embryonic kidney cells (HEK-293, ATCC accession No. CRL-1573) were chosen for heterologous expression of InsP3R. HEK-293 cells could be transfected by standard calcium-phosphate method with high efficiency and were used previously for functional expression of RyanR (Chen et al., 1997). HEK-293 cells were maintained in high glucose DMEM supplemented with 10% heat-inactivated fetal bovine serum and penicillin:streptomycin mixture (all from GIBCO BRL, Gaithersburg, MD). Cells were maintained in a tissue culture incubator at 37°C under 5% CO2. 1 d before transfection, HEK-293 cells in an exponential growth phase were subcultured into large (75 cm2) culture flasks using trypsin-EDTA treatment. On the next day, HEK-293 cells at ∼50% confluence were transfected with InsP3R-pcDNA3 DNA using the calcium phosphate method (MBS kit; Stratagene Inc.) or Lipofectamine reagent (GIBCO BRL) according to the manufacturer's suggestions and returned to the tissue culture incubator for several days to allow InsP3R-I expression. Control HEK-293 cells were transfected by pcDNA3 DNA following identical protocols.
Rat cerebellar microsomes were isolated essentially as described previously for canine preparation (Bezprozvanny et al., 1991; Bezprozvanny and Ehrlich, 1993, 1994). Briefly, cerebellar microsomes were excised from 12 Sprague Dawley rats (4–5-wk old) that were killed by carbon dioxide inhalation and decapitated. Cerebellar microsomes were minced and manually homogenized on ice using Teflon/glass tissue homogenizer in 15 ml of homogenization buffer (5 mM NaN3, 1 mM EDTA, 20 mM HEPES, pH 7.4) supplemented with the protease inhibitors cocktail (2 μg/ml aprotinin, 1 μg/ml leupeptin, 1 mM benzamidine, 0.2 mM γ-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF), 10 μg/ml pepstatin, 0.1 mg/ml PMSF). Another 15 ml of homogenization buffer was added to the homogenate, and the suspension was centrifuged for 15 min at 4,000 g (J 25.50 rotor; Beckman Instruments, Inc., Fullerton, CA). The supernatant fluid was filtered through cheesecloth, and the filtrate was centrifuged for 30 min at 90,000 g (Ti 50.2 rotor; Beckman Instruments, Inc.). The pellet from the latter spin was resuspended in 30 ml of high salt buffer B (0.6 M KCl, 5 mM NaN3, 20 mM Na4P2O7, 1 mM EDTA, 10 mM HEPES, pH 7.2) using Teflon/glass manual homogenizer and centrifuged for 15 min at 4,000 g (J 25.50 rotor). The resulting supernatant fluid was centrifuged for 30 min at 90,000 g (Ti 50.2 rotor). The pellet from the last spin was resuspended in 0.5 ml of storage buffer (10% sucrose, 10 mM MOPS, pH 7.0), aliquoted, snap frozen in liquid nitrogen, and stored at −80°C for future experiments. Total microsomal protein concentration (Bradford assay, BioRad kit; Bio-Rad Laboratories, Richmond, CA) in rat cerebellar microsomal preparation was typically close to 5 mg/ml.
Microsomes from HEK-293 cells were prepared by modification of the procedure described above for rat cerebellar microsomes. Briefly, 48 h after transfection with DNA, HEK-293 cells were loaded with 10 μM BAPTA-AM following standard protocol (Molecular Probes, Inc., Eugene, OR) and kept in serum-free DMEM overnight. On the next day, the HEK-293 cells were collected from two large (75 cm2) culture flasks using trypsin-EDTA treatment, washed with PBS, and pelleted by centrifugation at 4°C for 5 min at 3,000 rpm (GH 3.8 rotor; Beckman Instruments). The cellular pellet was resuspended in 2.5 ml homogenization buffer A (5 mM NaN3, 1 mM EDTA, 50 mM Tris-HCl, pH 8.0) supplemented with protease inhibitors cocktail and incubated on ice for 20 min. After incubation, cells were manually homogenized on ice with teflon-glass homogenizer, and 2.5 ml of homogenization buffer B (0.5 M sucrose, 1 mM EDTA, 20 mM HEPES, pH 7.5) was added. Cells were rehomogenized with teflon-glass homogenizer and centrifuged for 15 min at 3,000 rpm (GH 3.8 rotor). Concentrated high salt buffer (5×) was added to the supernatant to yield 0.6 M KCl and 20 mM Na4P2O7. Microsomes were incubated on ice for 30 min and centrifuged for 15 min at 3,000 rpm (GH 3.8 rotor). The resulting supernatant was centrifuged for 30 min at 100,000 g (52,000 rpm, Ti 100.3 rotor; Beckman Instruments). The pellet from the last spin was resuspended in 0.15 ml storage buffer (10% sucrose, 10 mM MOPS, pH 7.0), aliquoted, quickly frozen in liquid nitrogen, and stored at −80°C. The microsomal protein concentration in HEK-293 microsomal preparation was typically close to 6 mg/ml.
Immunological Detection of InsP3R-I
Peptide (RIGLLGHPPHMNVNPQQPA) corresponding to the carboxy terminus of rat InsP3R-I (Mignery et al., 1990) was synthesized (Biopolymers Facility, HHMI, UT Southwestern Medical Center at Dallas) and coupled to keyhole limpet hemocyanin by glutaraldehyde treatment (Harlow and Lane, 1988). Two New Zealand White rabbits were immunized with the resulting antigenic conjugate mixed with Freund's complete adjuvant and boosted with the same antigen mixed with Freund's incomplete adjuvant 3 and 6 wk after initial immunization. Serum from the rabbit that had a higher titer of specific antibody was collected to yield a stock of anti–InsP3R-I antibody (T443). For analysis, microsomal proteins were solubilized in SDS (sodiumdodecyl sulfate) gel loading buffer, resolved by SDS-PAGE electrophoresis (6% polyacrylamide), and transferred to PVDF (polyvinylidene difluoride) membrane. Western blotting with T443 antibody was performed using the Western Max detection kit (Amresco Inc., Solon, OH) according to manufacturer's protocol. Titer and specificity of T443 antibody were determined to be similar to that of previously reported T210 anti–InsP3R-I antibody (Mignery et al., 1989), generated against the same COOH-terminal fragment of rat InsP3R-I. T210 antibody was a kind gift of Dr. Thomas Südhof.
Immunostaining of transfected HEK-293 cells with T443 antibody was performed according to standard protocol (Harlow and Lane, 1988). Briefly, transfected HEK-293 cells were grown on glass coverslips in tissue culture incubator (37°C, 5% CO2) for 24 h. Cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. Nonspecific antibody binding sites were blocked by Fraction V bovine serum albumin and stained with T443 antibody diluted 1:1,000 in the blocking solution. Bound T443 antibodies were visualized with rhodamine-conjugated anti– rabbit IgG (Jackson Immunoresearch Laboratories, Inc., West Grove, PA) and mounted on microscopic slides with Aqua-Mount (Lerner Labs) media for fluorescent microscopy (Axiovert 135, 20× objective; Carl Zeiss, Inc., Thornwood, NY) or laser (Argon/ Krypton) confocal immunofluorescence microscopy (MRC-1024; Bio-Rad Laboratories; attached to an Axiovert 135, 60× oil-immersed objective; Carl Zeiss, Inc.).
Specific [3H]InsP3 (Amersham Corp., Arlington Heights, IL) binding to microsomes isolated from HEK-293 cells or rat cerebellar microsomes was measured with minor modifications of procedure described previously (Chadwick et al., 1990). Briefly, microsomes (5 μg protein) were incubated on ice with 10 nM [3H]InsP3 in the binding buffer (50 mM Tris-HCl, pH 9.0, 1 mM EDTA, 1 mM DTT, 100 mM NaCl) and precipitated with 12.5% PEG and 1.2 mg/ml γ-globulin at 14,000 g. Precipitates were quickly washed with the binding buffer, dissolved in Soluene, and their [3H] content was determined by liquid scintillation counting. Nonspecific counts, determined in the presence of 25 μM nonlabeled InsP3, were subtracted from the total to yield specific binding. The density of specific InsP3 binding sites was determined by normalization of obtained results for specific [3H]InsP3 radioactivity and the amount of microsomal protein in the assay. Scatchard analysis of specific InsP3 binding sites in microsomal preparations was done by adding variable amounts of InsP3 (from 10 nM to 8 μM concentrations) to [3H]InsP3 binding experiments performed as described.
Planar Lipid Bilayer Recordings and Single-Channel Data Analysis
Planar lipid bilayers were formed from PE (phosphatidylethanolamine):PS (phosphatidylserine) (3:1) synthetic lipid (Avanti Polar Lipids, Alabaster, AL) mixture in decane on the small (100– 200-μm diameter) hole in Teflon film separating two chambers 3 ml each (cis and trans). Before formation of the bilayer, the hole was prepainted with PC (phosphatidylcholine):PS mixture (3:1). Recombinant InsP3R-I from HEK-293 cells or rat cerebellar InsP3R were incorporated into the bilayer by microsomal vesicle fusion as described previously for canine preparation (Bezprozvanny et al., 1991; Bezprozvanny and Ehrlich, 1993, 1994). Single channel currents were recorded at 0-omV transmembrane potential using 50 mM Ba2+ dissolved in HEPES, pH 7.35, in the trans (intraluminal) side as a charge carrier (Bezprozvanny and Ehrlich, 1994). In most experiments (standard recording conditions of InsP3R activity), the cis (cytosolic) chamber contained 110 mM Tris dissolved in HEPES, pH 7.35, 0.2 μM free Ca2+ (Bezprozvanny et al., 1991) buffered with 1 mM EGTA and 0.7 mM CaCl2, 1 mM Na2ATP (Bezprozvanny and Ehrlich, 1993), and 2 μM InsP3. We found that 2 μM of ruthenium red in the cis chamber increases native and recombinant InsP3R single channel open probability (Po) 1.5–2-fold (Lupu and Bezprozvanny, manuscript in preparation). Thus, in most experiments, channels were recorded in the presence of 2 μM of ruthenium red in the cis chamber to stimulate InsP3R activity and inhibit cerebellar RyanR (Bezprozvanny et al., 1991). All additions (InsP3, ATP, CaCl2, heparin) were to the cis chamber from the concentrated stocks with at least 30 s stirring of solutions in both chambers.
InsP3R single channel currents were amplified (OC-725; Warner Instruments, Hamden, CT), filtered at 1 kHz by a low pass eight-pole Bessel filter, digitized at 5 kHz (Digidata 1200; Axon Instruments, Foster City, CA) and stored on computer hard drive and recordable optical discs. Single channel data for off-line computer analysis (pClamp 6.0.3; Axon Instruments) were filtered digitally at 500 Hz and, for presentation of the current traces data, were filtered at 200 Hz. Po was determined using half-threshold crossing criteria (t ≥ 2 ms) from records lasting at least 2.5 min.
Transient Expression of InsP3R in HEK-293 Cell Line
Transfection of HEK-293 cell line with InsP3R-pcDNA3 clone resulted in transient expression of InsP3R-I in 20– 30% of transfected cells as determined by immunocytochemical staining with T443 anti–InsP3R-I polyclonal antibody (Fig. 1,A). No detectable staining with T443 antibody was observed in untransfected HEK-293 cells (data not shown) or cells transfected with pcDNA3 vector alone (Fig. 1 B). At the subcellular level, expressed InsP3R was localized to cytoplasmic (presumably endoplasmic reticulum) and perinuclear regions of transfected HEK-293 cells as determined using laser confocal fluorescent microscopy (data not shown). Observed subcellular distribution of InsP3R expressed in HEK-293 cells was similar to that reported previously for InsP3R expressed heterologously in COS cells (Takei et al., 1994).
Quantification of InsP3R Expression in HEK-293 Cells
Endoplasmic reticulum–enriched microsomal preparation was isolated from HEK-293 cells 72 h after transfection with InsP3R-pcDNA3 or pcDNA3 as described in materials and methods. A major band was recognized on Western blots with T443 anti–InsP3R-I antibody in microsomes extracted from InsP3R-pcDNA3–transfected cells (Fig. 2), whereas the signal was very faint in membranes obtained from pcDNA3-transfected HEK-293 cells (Fig. 2) or untransfected HEK-293 cells (data not shown). The position of the band detected by T443 antibody in microsomes from InsP3R-pcDNA3–transfected HEK-293 cells corresponds to the expected position of full-length rat InsP3R-I (∼260 kD; Fig. 2, arrow) and to the position of the band detected by T443 antibody in rat cerebellar microsomes loaded on the same gel. Shorter immunoreactive products apparent in both lanes most likely result from limited proteolysis of InsP3R during microsomal extraction from HEK-293 cells and rat cerebellum. Based on relative density of major immunoreactive bands (Fig. 2), we estimated that InsP3R-I in microsomes from InsP3R-pcDNA3–transfected HEK-293 cells are at least 10-fold more abundant than in HEK-293 cells transfected with pcDNA3 alone. The density of InsP3R-I in rat cerebellar microsomes is approximately twofold higher than in microsomes from InsP3R-pcDNA3–transfected HEK-293 cells (Fig. 2).
InsP3R-III is the major isoform expressed in most cultured cell lines (De Smedt et al., 1997). T443 antibody, raised against carboxy-terminal peptide fragment of InsP3R-I (see materials and methods), may not recognize most of the endogenous InsP3R present in HEK-293 cells. For better comparison of InsP3R-I expression level with endogenous InsP3R, we used quantitative [3H]InsP3 binding assay as described in materials and methods. At 10 nM [3H]InsP3, the density of specific InsP3 binding sites was determined to be no more than 0.2 pM/mg in untransfected HEK-293 cells or HEK-293 cells transfected with pcDNA3 vector (Fig. 3,A). The density of specific [3H]InsP3 binding sites varied from 1.5 to 8 pmol/mg between different microsomal preparations from InsP3R-pcDNA3–transfected HEK-293 cells, yielding an average value of 4.4 ± 1.1 pmol/mg (n = 6) (Fig. 3,A). Thus, transfection of HEK-293 cells with InsP3R-pcDNA3 results in at least a 20-fold increase in InsP3R density. For comparison, we determined the density of InsP3R in rat cerebellar microsomes to be equal to 14 ± 2 pmol/mg (n = 10) (Fig. 3,A), 70-fold above endogenous background in HEK-293 cells and 3-fold higher than in microsomes from InsP3R-pcDNA3–transfected HEK-293 cells. These data agree with results from T443 Western blotting (Fig. 2). When [3H]InsP3 binding to microsomes from InsP3R-pcDNA3–transfected HEK-293 cells was characterized by Scatchard analysis (Fig. 3 B), affinity of specific InsP3 binding sites equal to 60 nM was determined, similar to previously reported Kd values for the rat cerebellar InsP3R (Worley et al., 1987) and the rat InsP3R-I expressed in COS cells (Mignery et al., 1990).
We concluded that transient transfection of HEK-293 cells with the InsP3R-pcDNA3 clone results in high level heterologous expression of rat neuronal InsP3R. Expressed InsP3R appears to be full length, has normal immunoreactive properties, and has the expected subcellular localization. Expressed InsP3R retains its normal InsP3 binding properties. The density of expressed InsP3R is at least 20-fold higher than the density of endogenous InsP3R present in HEK-293 cells, and only 2–3-fold lower than InsP3R density in rat cerebellar microsomes.
Single-Channel Recording of Recombinant InsP3R Expressed in HEK-293 Cells
High level heterologous expression of InsP3R in HEK-293 cells provides an opportunity for characterization of single channel properties of recombinant InsP3R. When microsomes isolated from InsP3R-pcDNA3–transfected HEK-293 cells were fused to planar lipid bilayers, no channel activity was observed in control conditions (Fig. 4,A, top). Addition of 2 μM InsP3 to the cytosolic (cis) compartment activated InsP3-gated channels in the bilayers (Fig. 4 A). InsP3-gated channels from InsP3R-pcDNA3–transfected HEK-293 cells were inhibited by heparin (data not shown), in agreement with known InsP3R pharmacological properties. InsP3-gated channels were observed frequently (in 30 of 76 experiments) with microsomes from InsP3R-pcDNA3–transfected HEK-293 cells, but did not appear in experiments with microsomes from untransfected HEK-293 cells (n = 4) or pcDNA3-transfected HEK-293 cells (n = 5). We did not observe InsP3-gated channels in experiments with microsomes from pCMVI-9-transfected HEK-293 cells (Mignery et al., 1990; n = 5), presumably due to low InsP3R expression levels with this construct (we detected only ∼0.3 pmol/mg specific [3H]InsP3 binding sites in microsomes isolated from pCMVI-9-transfected HEK-293 cells). We also observed strong correlation between the efficiency of InsP3R-pcDNA3 transfections of HEK-293 cells (as judged by the density of specific [3H]InsP3 binding sites) and the frequency of InsP3-gated channels' appearance in planar lipid bilayer experiments. Indeed, occurrence of InsP3-gated channels in bilayers varied from 25% (6 of 24) for less optimal transfection experiments (2 pmol/mg specific [3H]InsP3 binding sites in microsomal preparation) to 51% (25 of 49) in more successful transfections (8 pmol/mg specific [3H]InsP3 binding sites), comparable to the success rate of InsP3R incorporation in experiments with rat cerebellar microsomes (typically ∼60% for most cerebellar microsomal preparations). No channel activity was observed in experiments with microsomes that had a density of [3H]InsP3 binding sites of <2 pmol/mg. All these data lead to the conclusion that endogenous InsP3R background (no more than 0.2 pmol/mg [3H]InsP3 binding sites) is negligible in our planar lipid bilayer assay, and InsP3-gated channels observed in these experiments correspond to the activity of recombinant InsP3R-I expressed in HEK-293 cells.
Rodent cerebellar InsP3R corresponds to InsP3R-I SII+ splice isoform, which is 85% SI− and 15% SI+ (Nakagawa et al., 1991a). Thus, our working clone InsP3R-pcDNA3 (InsP3R-I, SI− SII+, see materials and methods) corresponds to predominant InsP3R isoform expressed in rodent cerebellum. It is interesting to compare single channel properties of recombinant rat InsP3R-I with characteristics of native rat cerebellar InsP3R. To perform this comparison, we isolated rat cerebellar microsomes (see materials and methods) and fused them to planar lipid bilayers. The addition of 2 μM InsP3 to the cytosolic compartment resulted in InsP3-gated channel activity (Fig. 4 B), which was inhibited by heparin (data not shown). Thus, rat cerebellar InsP3R activity could be recorded in planar lipid bilayers by using an experimental protocol similar to one used previously for canine cerebellar InsP3R single-channel recordings (Bezprozvanny et al., 1991; Bezprozvanny and Ehrlich, 1993, 1994).
Detailed comparison of functional properties of recombinant InsP3R-I and native rat cerebellar InsP3R in the standard recording conditions (0.2 μM free Ca2+, 1 mM ATP, 2 μM InsP3 on the cis side of the membrane; 50 mM Ba2+ on the trans side as a current carrier) was performed. Fig. 5 compares the results of single channel data analysis obtained in representative experiments with recombinant InsP3R-I (top) and native rat cerebellar InsP3R (bottom). Similar data from three independent experiments with recombinant InsP3R-I (rec InsP3R-I) and native rat cerebellar InsP3R (cer InsP3R) were averaged together to generate Table I. Open dwell time distribution of recombinant and native InsP3R could be fit with a single exponent yielding the mean open time τo (Fig. 5, Table I). Closed dwell time distribution for expressed InsP3R-I and native cerebellar InsP3R could be fit better with a sum of two exponential functions: W1 exp(−t/τc1) + W2 exp(−t/τc2) (Fig. 5). Average values of time constants, τc1 and τc2, and relative weight factors, W1 and W2, are in Table I. Single channel current amplitude histograms for both types of channels were fit with a single Gaussian function that yielded mean size of the single channel current at 0 mV transmembrane voltage i (Fig. 5, Table I). From current measurement at different transmembrane potentials, single channel conductance γ of recombinant InsP3R-I and native rat cerebellar InsP3R in the standard recording conditions was close to 80 pS (Fig. 6, Table I). Thus, we concluded that most fundamental gating and conduction properties of recombinant neuronal rat InsP3R-I heterologously expressed in HEK-293 cells are not significantly different from analogous characteristics of native rat cerebellar InsP3R.
Bell-shaped Ca2+ Dependence of Recombinant InsP3R Expressed in HEK-293 Cells
Bell-shaped dependence of InsP3R on cytosolic Ca2+ is one of the most fundamental InsP3R properties responsible for complex spatiotemporal aspects of Ca2+ signaling (Berridge, 1993). It is not known if Ca2+ interacts directly with the InsP3R or if some auxiliary Ca2+-binding protein is required to confer Ca2+ sensitivity to InsP3R. To address this question, we analyzed modulation of recombinant InsP3R by cytosolic Ca2+ in planar lipid bilayer reconstitution experiments. We monitored recombinant InsP3R activity in bilayers in the presence of 2 μM InsP3 and 1 mM Na2ATP at cis (cytosolic) Ca2+ concentrations in the range between 10 nM and 5 μM Ca2+. Ca2+ concentration in the cis chamber, calculated according to Fabiato (1988), was adjusted by using calibrated 20 mM CaCl2 stock solution and 1 mM mixture of HEDTA and EGTA. We determined that recombinant InsP3R expressed in HEK-293 cells retained bell-shaped dependence on cytosolic Ca2+ concentration (Fig. 7, n = 3), which was similar to Ca2+ dependence of native canine (Bezprozvanny et al., 1991) and rat (data not shown) cerebellar InsP3R. Recombinant InsP3R was also allosterically modulated by ATP (data not shown), similar to native cerebellar InsP3R (Bezprozvanny and Ehrlich, 1993).
In this paper, we report transient expression of neuronal rat InsP3R (InsP3R-I, SI− SII+ splice isoform) (Mignery et al., 1990) in HEK-293 cell line and characterize basic functional properties of recombinant InsP3R-I at the single channel level. We compared properties of recombinant InsP3R expressed in HEK-293 cells with rat cerebellar InsP3R. Rodent cerebellar InsP3R corresponds to InsP3R-I SII+ splice isoform, which is 85% SI− and 15% SI+ (Nakagawa et al., 1991a). InsP3R expressed in HEK-293 cells were specifically recognized by rabbit polyclonal antibody raised against rat InsP3R-I carboxy terminus. Recombinant InsP3R had normal (∼260 kD) mobility on SDS-PAGE gel, and expected mobility for intracellular Ca2+ release channel subcellular localization to endoplasmic reticulum and perinuclear region. Expressed InsP3R retained specific InsP3 binding activity with a Kd of 60 nM, similar to the native cerebellar InsP3R. Based on immunological assays and quantitative [3H]InsP3 binding experiments, the estimated density of recombinant InsP3R-I in microsomes extracted from InsP3R-pcDNA3–transfected HEK-293 cells was at least 20-fold above endogenous InsP3R background and 2–3-fold lower than the InsP3R density in rat cerebellar microsomes. When incorporated into planar lipid bilayers, the recombinant rat InsP3R-I formed functional InsP3-gated Ca2+ channels very similar in their gating and conduction properties to native rat cerebellar InsP3R. Channel properties of recombinant rat InsP3R-I were also very similar to the properties of canine cerebellar InsP3R, previously characterized in detail (Bezprozvanny and Ehrlich, 1994). Similar to native cerebellar InsP3R, recombinant InsP3R-I displayed bell-shaped dependence on cytosolic Ca2+ concentration (Bezprozvanny et al., 1991) and was allosterically modulated by ATP (Bezprozvanny and Ehrlich, 1993).
The background from endogenous InsP3R present in HEK-293 was insignificant in planar lipid bilayer assay, most likely due to 20-fold molar excess of recombinant InsP3R. This is a main advantage of the planar lipid bilayer reconstitution method as it enabled us to treat HEK-293 cells as “zero background” host. From our experience, the microsomal density of InsP3R of at least 2 pmol/mg is required for successful InsP3R reconstitution into bilayers, 10-fold higher than the density of endogenous InsP3R in microsomes isolated from HEK-293 cells. In comparison, Ca2+ flux assay is affected by endogenous InsP3R background to a far greater extent, and endogenous InsP3R present in L-type mouse fibroblasts obscured previously published measurements of recombinant InsP3R activity using this method (Mi-yawaki at al., 1990). Thus, we concluded that InsP3-gated channels recorded in our experiments with microsomes from InsP3R-pcDNA3–transfected HEK-293 cells correspond to the activity of recombinant rat InsP3R, which possess basic functional properties of native cerebellar InsP3R. However, we cannot rule out the possibility that functional InsP3-gated channel is formed from the complex of heterologously expressed neuronal rat InsP3R-I with the auxiliary components recruited from the set of proteins endogenously present in HEK-293 cells.
Transient expression of rat and mouse InsP3R-I cDNA has been previously reported using COS cells (Mignery et al., 1990) and the NG-108 cell line (Furuichi et al., 1989). Although recombinant InsP3R-I obtained in these studies retain normal immunoreactivity and specific InsP3 binding properties, it has not been demonstrated whether they form a functional InsP3-gated Ca2+ channel. InsP3-induced Ca2+ release is facilitated in the L-fibroblast stable cell line overexpressing mouse InsP3R-I (Miyawaki et al., 1990). This result indicates that the InsP3R-I expressed in L-fibroblasts are functional channels. Detailed characterization of recombinant InsP3R-I functional properties has not been previously reported due to substantial background signal from endogenous InsP3R in InsP3-induced Ca2+ flux measurements. Thus, the present report constitutes the first description of basic functional properties of recombinant InsP3R-I expressed in a heterologous system. We found that heterologous expression of InsP3R-I cDNA in HEK-293 cells is sufficient to confer most basic functional properties of native cerebellar InsP3R. Gating and conductance properties of recombinant InsP3R-I are very similar to native rat cerebellar InsP3R (Table I).
Similar to native cerebellar (Bezprozvanny et al., 1991; Finch et al., 1991) and smooth muscle (Iino, 1990) InsP3R, recombinant InsP3R-I displayed bell-shaped dependence on cytosolic Ca2+ concentration (Fig. 7). Supporting biphasic regulation of InsP3R-I requires interaction of Ca2+ with both activating and inhibitory sites of InsP3R. Ca2+ may bind directly to InsP3R or act via auxiliary Ca2+ binding protein. If it exists, this auxiliary protein must be ubiquitously expressed to cause similar Ca2+ sensitivity of native cerebellar InsP3R (Bezprozvanny et al., 1991) and recombinant InsP3R expressed in HEK-293 cells (Fig. 7). One obvious candidate for this role is calmodulin (CaM). Indeed, biochemical analysis identified a unique CaM-binding site in the coupling domain of InsP3R-I (Yamada et al., 1995). We hypothesize that CaM plays a role in biphasic modulation of InsP3R by Ca2+. Purified InsP3R is no longer inhibited by Ca2+ (Callamaras and Parker, 1994), probably due to loss of CaM during the purification procedure. Thus, it is possible that CaM is responsible for the inhibitory part of the bell-shaped curve. This hypothesis, as well as other ideas related to structural determinants responsible for InsP3R conductance and modulation by Ca2+ and ATP, can now be tested using methodology described in the present report.
In conclusion, functional expression of recombinant InsP3R in HEK-293 cells provides a useful model for studies of functional differences between InsP3R-I splice variants and InsP3R isoforms, and in combination with site-directed mutagenesis of InsP3R cDNA will lead to a better understanding of structural determinants responsible for most fundamental InsP3R functional properties.
We are grateful to G. Mignery and T.C. Südhof for rat InsP3R-I cDNA. We thank G. DeMartino and X. Lin for advice with rabbit polyclonal antibody production, K. Luby-Phelps and H.R. Payne for help with immunofluorescent microscopy and confocal imaging, and J. Ma and S.R.W. Chen for advice on transfection procedures and communicating their results before publication. I. Bezprozvanny is thankful to S. Bezprozvannaya for tremendous support and encouragement of his work. E. Kaznacheyeva is on leave from the Institute of Cytology, Russian Academy of Sciences (St. Petersburg, Russia).
Supported by the American Heart Association, Robert Welch Foundation, and start-up funds from UT Southwestern (to I. Bezprozvanny).
Abbreviations used in this paper
Drs. Kaznacheyeva and Lupu contributed equally to this work and should be considered co-first authors.
Address correspondence to Dr. Ilya Bezprozvanny, Department of Physiology, K4.112, UT Southwestern Medical Center at Dallas, Dallas, TX 75235-9040. Fax: 214-648-8685; E-mail: firstname.lastname@example.org