Oligomerization of Ca_vβ subunits is an essential correlate of Ca²⁺ channel activity

Abstract

Voltage-gated calcium channels conduct Ca²⁺ ions in response to membrane depolarization. The resulting transient increase in cytoplasmic free calcium concentration is a critical trigger for the initiation of such vital responses as muscle contraction and transcription. L-type Ca_v1.2 calcium channels are complexes of the pore-forming α_1C subunit associated with cytosolic Ca_vβ subunits. All major Ca_vβs share a highly homologous membrane associated guanylate kinase-like (MAGUK) domain that binds to α_1C at the α-interaction domain (AID), a short motif in the linker between transmembrane repeats I and II. In this study we show that Ca_vβ subunits form multimolecular homo- and heterooligomeric complexes in human vascular smooth muscle cells expressing native calcium channels and in Cos7 cells expressing recombinant Ca_v1.2 channel subunits. Ca_vβs oligomerize at the α_1C subunits residing in the plasma membrane and bind to the AID. However, Ca_vβ oligomerization occurs independently on the association with α_1C. Molecular structures responsible for Ca_vβ oligomerization reside in 3 regions of the guanylate kinase subdomain of MAGUK. An augmentation of Ca_vβ homooligomerization significantly increases the calcium current density, while heterooligomerization may also change the voltage-dependence and inactivation kinetics of the channel. Thus, oligomerization of Ca_vβ subunits represents a novel and essential aspect of calcium channel regulation.—Lao, Q. Z., Kobrinsky, E., Liu, Z., Soldatov, N. M. Oligomerization of Ca_vβ subunits is an essential correlate of Ca²⁺ channel activity.

Functional calcium channels are clustered (1–4) complexes of the pore-forming α₁ subunits with auxiliary α₂δ and cytosolic Ca_vβ subunits (5). Each of the 4 known types of Ca_vβ, encoded by 4 different β_1–4 genes (6), share a highly homologous central membrane associated guanylate kinase-like (MAGUK) domain (7–10) that binds to the α₁ subunits at the conserved α-interaction domain (AID) situated in the linker between transmembrane repeats I and II (11). Association between α₁ and Ca_vβ is essential for voltage gating, calcium-induced inactivation, and plasma membrane targeting of the channels (12). Earlier experiments showed that the α₁/α₂δ/Ca_vβ complexes are stable in nonionic detergents, and a stoichiometric ratio between α₁, α₂δ, and Ca_vβ in a purified functional channel is ∼1:1:1.3 (13). However, the exact ratio between the subunits in the plasma membrane-bound channels remains unclear. Overexpression of Ca_vβ was shown to affect the electrophysiological properties of the Ca_v1.2 (14–16) and Ca_v2 channels (17–19), and may have pathophysiological consequences (20–22). The nature and mechanisms of these effects remain largely unexplained. A stimulating hypothesis suggesting involvement of higher-order regulatory complexes (α₁β)β_n has been suggested by Tareilus et al. (18). This hypothesis becomes particularly attractive in light of a later finding that the purified Ca_vβs exhibit tendency to reversible aggregation (23). Our report provides evidence that Ca_vβ subunits in naturally occurring and recombinant Ca_v1.2 channels form oligomeric complexes and that the Ca_vβ oligomerization is a new molecular correlate of calcium channel regulation.

MATERIALS AND METHODS

Molecular biology

The plasmids coding for Flag-, V5-His-, and Venus-tagged proteins were constructed using p3×FLAG-Myc-CMV^TM24 (Sigma, St. Louis, MO, USA), pCDNA3.1D/V5-His-TOPO (Invitrogen, Carlsbad, CA, USA), and monomeric mVenus- C1 vectors (24). In this study we used cDNAs coding for the following human Ca_vβ subunits: β_1b (M92302), β_2d (AF423191), and β₃ (X76555). Flag-α_1C, α_1CAID, α_1CAID/IQ, and α₂δ-1 as well as red fluorescent protein (RFP)-, Cerulean-, and Venus-tagged Ca_vβs were prepared as described previously (1, 25). Flag-tagged Ca_vβ constructs were produced by PCR amplification of each cDNA with the respective linkers followed by incorporation into p3×FLAG-Myc-CMV^TM24 at 5′-EcoRI/XbaI-3′ for β_2d and β₃, and at 5′-EcoRI/BamHI-3′ for β_1b. Flag-β₃ΔGK was produced by 2-step PCR as described previously (25). The inner primers were 5′-TATGACCGTGGTGCCCTCCCACCCAGCCCCTGGCCCCGGACTTCT-3′ (sense) and 5′-GGAGGGCACCACGTCATATGGGGG-3′ (antisense); the outer primers were 5′-atataaagcttATGTATGACGACTCCTACGTGCCC-3′ (sense) and 5′-atatatctagaGTAGCTATCCTTGGGCCAAGGCCG-3′ (antisense). The resulting PCR product was then incorporated into p3×FLAG-Myc-CMV^TM24 at 5′-HindIII/XbaI-3′. The Flag tag is on the N termini of all the Flag-tagged proteins. V5-His-β₃ and I-II linker were subcloned into the pCDNA3.1D/V5-His-TOPO vector according to standard PCR scheme, and the V5-His tag is on the C termini of the fusion proteins. The Flag-tagged GK fragments were constructed by PCR amplification of each fragment with linkers followed by subcloning into the vector at 5′-HindIII/EcoRI-3′ sites. The Venus-tagged GKC fragments were cloned by PCR amplification using 5′-BspEI/EcoRI-3′ sites of mVenus-C1 vector. QuickChange II Site-Directed Mutagenesis Kit and GeneMorph II EZClone Domain Mutagenesis Kit (Stratagene, La Jolla, CA, USA) were used for mutagenesis study. All recombinant plasmids were confirmed by DNA sequencing.

Cells and transfection

Primary human aortic smooth muscle cells were purchased from Lonza (Walkersville, MD, USA) and cultured for 5 d following the manufacturer's instructions. Cos7 cells were maintained in Dulbecco's modified Eagle medium (10% FBS, 4.5 g/L glucose) at 37°C in 5% CO₂. Early passage cells (≈10⁶/100-mm dish) were used for transfection by Effectene (Qiagen, Valencia, CA, USA). The cells were cultured for 72 h before use in experiments.

Coimmunoprecipitation (co-IP) and Western blot analysis

EZview Red FLAG M2 affinity gel was used for Flag IP with 3× Flag peptide (200 ng/μl) (Sigma) to elute the IP proteins. His pulldown was carried out with MagneHis (Promega, Madison, WI, USA) using 0.5–1 M imidazole in PBS for elution. A Cell Surface Protein Isolation Kit (Pierce Biotechnology, Rockford, IL, USA) was used for biotinylation of surface membrane proteins and isolation of the plasma membrane-bound calcium channel complexes. Standard heat denaturation at 95°C for 5 min in 1× Laemmli's sampling buffer was applied to the protein samples showing a “ladder” on Western blots. All other samples were treated by 50 mM DTT at 37°C for 30 min prior to heat denaturation. Tris-Acetate SDS-PAGE (3–8%) was used for the ladders, and 4–20% Tris-glycine SDS-PAGE was used for all other samples. Antibodies were obtained from Sigma (Flag, GFP), Invitrogen (V5; Invitrogen) and Millipore (α_1C, β₃; Millipore, Billerica, MA, USA). The dilution of primary antibodies for immunoblotting was as follows: 1:5000 for Flag (1:1000 for Flag-α_1C), 1:10,000 for GFP, 1:3000 for V5, 1:1000 for β₃, and 1:500 for α_1C. For Far-Western blotting, Flag- and V5-His-tagged β₃ were expressed separately in Cos7 cells and purified as described above. V5-His-β₃ was resolved by SDS-PAGE and transferred to nitrocellulose membrane, followed by blocking with 3% BSA in PBS for 1 h and incubation of purified Flag-β₃ (preheated at 65°C for 10 min) in 3% BSA at room temperature for 3 h. The membrane was then washed with TBS, blocked with 5% milk, and subjected to immunoblotting by an anti-Flag antibody. Blue native PAGE was performed as described previously (26) in 4–12% Bis-Tris gel, and Sigma MW-GF-1000 was used as molecular mass marker.

Colocalization of Venus-β₃ and RFP-β₃ by confocal microscopy

Cos7 cells were transfected with Venus-β₃ and RFP-β₃ (0.1 μg each vector/10⁴ cells/slide). Colocalization of differentially labeled β₃ was analyzed 72 h after transfection by the confocal laser scanning microscopy on a LSM510 confocal microscope (Carl Zeiss, Oberkochen, Germany) equipped with a ×63 1.4-numerical aperture (n.a.) oil DIC objective and multiple filter sets. The yellow fluorescent protein (YFP) cube (excitation 514 nm, emission 522–565 nm) was used to detect Venus-β₃ with an argon laser; the RFP cube (excitation 561 nm, emission > 575 nm) was used to detect RFP-β₃ with a diode-pumped solid-state (DPSS) 561-10 laser .

Colocalization of Venus-β₃ and Flag-β₃ by immunostaining and confocal microscopy

Cos7 cells were transfected with Venus-β₃ and Flag-β₃ (0.1 μg each vector/10⁴ cells/slide). After 72 h in culture, transfected cells were fixed with 2% formaldehyde in PBS and permeabilized with 2.5% Triton-X in PBS. Subsequently, the cells were blocked with 5% BSA in PBS for 2 h at room temperature and immunostained with a mouse anti-Flag antibody (Sigma, 1:250 dilution) for 1 h, followed by standard immunostaining with rhodamine Red X-conjugated donkey anti-mouse antibody (Jackson ImmunoResearch, West Grove, PA) as the secondary antibody (at 1:2000 dilution). Colocalization of Venus-β₃ and Flag-β₃ was analyzed by confocal microscopy using YFP and RFP cubes, respectively, as described above. The fluorescent-positive cells were confirmed by phase-contrast microscopy. Before image recording, the noise was adjusted on nontransfected cells and the cells immunostained only with the secondary antibody.

Plasma membrane fluorescence resonance energy transfer (FRET) measurements

Cells were transfected with α_1C, α₂δ, Venus-β_2d, and Cerulean-β_2d at 1:1:1:1 mol/mol ratio as described previously (1). Images were recorded with a pixel size of 200 nm using a 14-bit Hamamatsu C9100–12 digital camera (Hamamatsu City, Japan) mounted on a Nikon TE2000 epifluorescent microscope (Nikon, Tokyo, Japan) equipped with a ×60 1.45-n.a. oil objective and multiple filter sets (Chroma Technology, Rockingham, VT, USA). Excitation light was delivered by a 175-W xenon lamp. Excitation filter sets were changed by a high-speed filter wheel system (Lambda 10-2; Sutter Instrument, Novato, CA, USA). The Dual-View system (Optical Insights, Santa Fe, NM, USA) was used for the simultaneous acquisition of 2 fluorescence images (donor and FRET). Images were collected and analyzed using C-Imaging (Compix, Cranberry Township, PA, USA) and MATLAB v.7.0.4 (The Mathworks, Natick, MA, USA). FRET was measured from the plasma membrane area with 3 filter sets: for the YFP cube, excitation filter 500/20 nm, dichroic beam splitter 515 nm, emission filter 535/30 nm; for the cyan fluorescent protein (CFP) cube, excitation filter 436/20 nm, dichroic beam splitter 505 nm, emission filter 480/40 nm. The average plasma membrane FRET efficiency was calculated as described elsewhere (1).

Cellular electrophysiology

Whole-cell recordings were performed exactly as described previously (12) with 20 mM CaC1₂ in bath medium. Calcium currents were evoked by stepwise depolarization applied from the holding potential of −90 mV.

RESULTS

Ca_vβs form homo- and heterooligomeric complexes

In our experiments we used primary human aortic smooth muscle cells expressing predominantly β₃ calcium channel complexes (27). Cell lysate was immunoblotted with an anti-β₃ antibody after standard heat denaturation (5 min, 95°C) with or without prior DTT treatment (Fig. 1A). DTT treatment revealed predominantly the monomeric β₃ on the blot. However, in the absence of DTT treatment we observed ladder-like β₃ immunoactive bands on the blot (Fig. 1Aa), suggesting the presence of β₃ supramolecular complexes in naturally occurring human vascular calcium channels. The ladder-like bands are β₃ specific, because preincubation of the β₃ antibody with its immunogen peptide inhibited recognition of this pattern by the antibody (Fig. 1Ab).

Figure 1.

Oligomerization of Ca_vβ subunits. A) Immunoblotting (IB) of β₃ in human aortic smooth muscle cells with an anti-β₃ antibody after denaturation for 5 min at 95°C with or without prior treatment with DTT (a). To demonstrate the specificity of the antibody, in control (b), an anti-β₃ antibody was preincubated with its immunogen peptide at room temperature for 1 h prior to the same immunoblotting procedure. B) Oligomerization of recombinant Ca_vβs expressed in Cos7 cells. Flag-labeled β_1b, β_2d, or β₃ (0.5 μg each) was expressed in Cos7 cells and immunoprecipitated (IP) and immunoblotted with an anti-Flag antibody with or without prior DTT treatment. C) An aliquot of the Flag-β₃ immunoprecipitate was analyzed without denaturation by blue native PAGE, followed by immunoblotting with an anti-Flag antibody (left panel) or with a rabbit anti-mouse IgG antibody (m IgG; right panel). Arrows mark oligomers; asterisk indicates monomer. D) Colocalization of Venus-β₃ and RFP-β₃ (a) or Venus-β₃ and Flag-β₃ (b) coexpressed in Cos7 cells. Images are confocal laser-scanning micrographs of representative cells obtained with YFP (Venus-β₃) and RFP cubes (RFP-β₃ and Flag-β₃/Red-X). Merged images illustrate colocalization of differentially labeled β₃ subunits. Scale bars = 20 μm. E) FRET between Venus- and Cerulean-tagged β_2d subunits coexpressed in Cos1 cells with α₂δ-1 and α_1C (shaded bar) or its Ala-mutants α_1CAID (α_1C with disrupted AID region; open bar) or α_1CAID/IQ (the same plus deleted LA/IQ region; solid bar). Inset: sketch of known interactions of α_1C (gray) and β_2d (blue) (24). FRET efficiency was measured as described previously (1) in live cells in the plasma membrane region of interest at −10 mV, 22°C. Number of tested cells is shown in the bars. *P < 0.05. F) Ca_vβ oligomers contain ≥3 β subunits. Flag-β₃ and V5-His-β₃ (0.5 μg each) were coexpressed in the absence (lane C, control) or presence (lane 1) of Venus-β₃ (0.8 μg); the anti-Flag-IP fraction was subsequently pulled down on a His-binding matrix and immunoblotted with an anti-GFP antibody. G) Far-Western blotting. V5-His-β₃ was pulled down by His; aliquots were separated by SDS-PAGE and probed by Far-Western blotting (see Materials and Methods) using 3% BSA (top right, control) or purified Flag-β₃ in 3% BSA (top right). V5-His-β₃ was verified by immunoblotting with an anti-V5 antibody after the Far-Western procedure (bottom panel). All plasmid DNA amounts in this study are reported per transfection per 10⁶ cells/100-mm dish; 3 μg of Flag-β₃ plasmid was used for purification of Flag-β₃.

To exclude any possible effect of endogenous channel subunits, such as α_1C (28), on Ca_vβ supramolecular complex formation, we then expressed recombinant Ca_vβs in calcium channels-free Cos7 cells (12). Epitope tags (Flag, His-V5 and GFP variants) were added to the recombinant Ca_vβs to ease purification and detection. In control experiments, we found that both anti-Flag and anti-β₃ antibodies recognize identical ladder-like Flag-β₃ bands (Supplemental Fig. S1A). All antibodies to tags used in this study are specific to their respective tagged β₃ subunits, which are also well recognized by an anti-β₃ antibody in the immunoblotting and immunoprecipitation assays (Supplemental Fig. S1B). Moreover, a ladder-like β₃ pattern on the gel is not due to β₃ overexpression, since it was also observed when the expression level of β₃ was lowered to that of the mouse brain tissue level (Supplemental Fig. S2A, B). The efficiency and specificity of the Flag-IP/Western blot assay to study protein oligomerization sensitive to DDT treatment was demonstrated with the example of Flag-tagged bacterial alkaline phosphatase expressed in Cos7 cells (Supplemental Fig. S1C). In the absence of DTT treatment, a Flag-specific IP/Western blot analysis successfully detected a dimer and a trimer of the protein, a result consistent with previous reports (29, 30), while no reactivity was observed with the nontransfected cells (Supplemental Fig. S1C, panel NT).

To investigate whether the ladder-like pattern of β₃ is common in all major Ca_vβs, Flag-tagged human β_1b (31), β_2d (32) (splice variants of β₁ and β₂, respectively; ref. 33), and β₃ were expressed in Cos7 cells, immunoprecipitated as described in experimental procedures, and then immunoblotted with an anti-Flag antibody under the denaturing conditions described in Fig. 1A. A distinct ladder pattern was observed on the blots with all 3 types of Ca_vβ (Fig. 1B), suggesting their oligomerization.

The oligomerization of β₃ was independently demonstrated by blue native PAGE technology (34) in the absence of denaturing (Fig. 1C). Flag-β₃ was expressed in Cos7 cells, immunoprecipitated with an anti-Flag antibody, and subjected to blue native PAGE, followed by immunoblotting with an anti-Flag antibody. The result revealed that the Flag-β₃ immunoprecipitate is composed of a mixture of the monomer and oligomers, the latter being predominant, as one can judge from staining intensity of the bands. However, these bands were not detectable when an aliquot of the same sample was immunoblotted with the secondary anti-mouse antibody (mIgG) alone, pointing to the lack of contamination by the immunoprecipitating antibody in the sample that may cause a high-molecular-weight shift.

Several independent methods confirmed the fact of Ca_vβ-Ca_vβ interaction. Consistent with Ca_vβ oligomerization, Venus-β₃ colocalized with RFP-β₃ (revealed by laser scanning microscopy, Fig. 1Da) or with Flag-β₃ coexpressed in Cos7 cells (revealed by immunostaining, Fig. 1Db). An independent measurement of FRET between Venus-β_2d and Cerulean-β_2d associated with the plasma membrane-bound α_1C also points to Ca_vβ oligomerization (Fig. 1E). FRET between Venus-β_2d and Cerulean-β_2d coexpressed with α_1C (Fig. 1E, shaded bar) may be due to their oligomerization or the proximity of their binding sites in α_1C. Indeed, it was established previously (25, 35) that α_1C has 2 binding sites for Ca_vβs, the AID and IQ motifs (see Fig. 1D, inset). Using α_1C mutants with disrupted AID region (α_1CAID), we found that the IQ motif alone is sufficient to dock the oligomers of differentially labeled β_2d (Fig. 1E, open bar). FRET between Venus- and Cerulean-labeled β_2d docked at α_1C is significantly decreased only when both AID and LA/IQ of α_1C are disrupted (Fig. 1E, solid bar), preventing α_1C/Ca_vβ interaction. The plasma membrane localization of α_1C and its mutants is confirmed in Supplmental Fig. S3. Thus, 3 independent techniques, blue native PAGE, in vivo colocalization, and FRET, each confirmed our IP/Western blot data that Ca_vβs are subject to oligomerization, which is not affected by overexpression of proteins and the type of host cells.

We then determined that Ca_vβs form a complex of ≥3 individual molecules. Figure 1F illustrates the experiment where Flag-β₃, Venus-β₃, and V5-His-β₃ were coexpressed in Cos7 cells and a 2-step pulldown procedure was applied to find out whether these 3 differentially labeled β₃ molecules coprecipitate. The Flag-β₃ complexes were immunoprecipitated by an anti-Flag antibody and eluted under native conditions, and then the His-V5-β₃/Flag-β₃ complexes were precipitated by His pulldown. Immunoblotting of these complexes with an anti-GFP antibody (Fig. 1F) revealed that Venus-β₃ (lane 1), but not vector alone (lane C), survived every stage of this 2-stage pulldown procedure. Each of the steps of immunoprecipitation was β₃-specific, as it was revealed by the Western blot assay of the aliquots with an antibody to β₃ (Supplemental Fig. S1D). Taken together, this result provides evidence for the existence of β₃ complexes composed of ≥3 individual β₃ proteins. Thus, Ca_vβs show tendency to form multimolecular complexes.

To investigate whether the interactions between Ca_vβ subunits are direct rather than indirect, we performed a Far-Western blotting by detecting a purified V5-His-β₃ with purified Flag-β₃ as a probe (Fig. 1G). V5-His-β₃ was purified by His pulldown from the lysates of the expressing cells and the aliquots of the purified protein were separated by SDS-PAGE and probed by Far-Western blotting using purified Flag-β₃. Our results indicated that the interactions between different β₃ molecules are direct and may occur in vitro. Taking into account the existence of β₃ complexes containing ≥3 individual molecules (see Fig. 1F) and the apparent molecular masses of the oligomers (see Fig. 1C), these data suggest that the mixture of Ca_vβ oligomers contains trimers together with higher supramolecular complexes.

Homo- and heteromeric complexes of Ca_vβs

To examine whether Ca_vβs form both homo- and heteromeric complexes, we coexpressed Flag-β₃ with mVenus-labeled β₃, β_1b, or β_2d in Cos7 cells. Oligomeric complexes were immunoprecipitated with an anti-Flag antibody and analyzed on Western blots by anti-Flag and anti-GFP antibodies. In a typical experiment, shown in Fig. 2A, Flag-β₃ coimmunoprecipitated with Venus-β₃ (lane 1) as well as with mVenus-β_1b (lane 2) and mVenus-β_2d (lane 3). Because an anti-Flag antibody did not pull down the mVenus-labeled β subunits in the absence of Flag-β₃ (Fig. 2A, lanes C1–C3), and Flag-β₃ was not coprecipitated with mVenus (lane C4), this result clearly demonstrates that β₃ forms complexes with β₃, β_1b, and β_2d. Immunoblot analyses revealed similar results for Flag-β_2d (Fig. 2B) and Flag-β_1b (Fig. 2C), suggesting that all major Ca_vβ subunits have a tendency to form homo- and heteromeric complexes.

Figure 2.

Ca_vβs form both homo- and heterooligomers. A) Venus-labeled β₃ (lanes C1, 1) β_1b (lanes C2, 2), β_2d (lanes C3, 3) (0.8 μg each), or mVenus (control, lane C4; 0.4 μg) were expressed in the absence (lanes C1–C3) or presence (lanes 1–3, C4) of Flag-β₃ (0.5 μg) in Cos7 cells, immunoprecipitated by an anti-Flag antibody, and immunoblotted with anti-Flag and anti-GFP antibody. B, C) Venus-labeled Ca_vβs were also coimmunoprecipitated with Flag-β_2d (B) or Flag-β_1b (C). (Note that the presence of low-molecular-mass bands recognized by an anti-GFP antibody may represent degraded Ca_vβs).

Molecular structures responsible for Ca_vβ oligomerization

We next sought to determine whether β-subunit oligomerization originates from specific Ca_vβ structural domains (7, 36). Nine mVenus-labeled fragments of β₃ (Fig. 3A) were coexpressed with Flag-β₃ in Cos7 cells. Co-IP of β₃ with an anti-Flag antibody and immunoblot analysis with an anti-GFP antibody (Fig. 3B) showed that Ca_vβ oligomerization does not involve the Src homology 3 (SH3) subdomain and is entirely due to the guanylate kinase (GK) subdomain of MAGUK. Because both the N- and C-terminal halves of GK (GKN and GKC, respectively) bind to β₃ (Fig. 3B, lanes 8, 9), we further divided GK into 4 fragments (NN, NC, CN, and CC) for the analysis of binding activity (Fig. 3C). Results showed that NN and CC bind to GKN, while NN, CN, and CC bind to GKC. Thus, oligomerization of Ca_vβ subunits is mediated by multiple determinants residing in GK. Indeed, deletion of GK domain from the full length β₃ completely abolished its binding to the full-size β₃ (Fig. 3D). Mutagenesis analyses of the GK subdomain (see Supplemental Fig. S4) confirmed our conclusion that all 3 regions of GK, shown in Fig. 3C (red boxes), are responsible for the oligomerization. However, combined mutations of the residues that were found to be critical for the binding of GK fragments to β₃ did not eliminate oligomerization of Ca_vβ₃ (Supplemental Table S1). The β₃ subunit contains 4 cysteine residues. Two of them are in the SH3 domain, which is not responsible for Ca_vβ oligomerization. We found that combined mutation of C328 and C346 in GK into alanines did not abolish oligomerization of β₃ (Supplemental Table S1). Taken together, these data suggest that some yet unknown structural factors other than established residues and disulfide bonds contribute to the Ca_vβ oligomerization.

Figure 3.

Multiple determinants of Ca_vβ oligomerization. A) Domain structure of β₃ (485 aa) with regard to the fragments tested for binding to β₃. B) Co-IP of Flag-β₃ with indicated mVenus-labeled fragments of β₃ (lanes 1–9) and mVenus (lane C, control). C) Co-IP of Flag-labeled GKN and GKC with mVenus-labeled NN, NC, CN, and CC fragments or mVenus (lane C). D) Deletion of GK from the full-length β₃ subunit abolished its binding with other β₃ molecules. V5-His-β₃ and Flag-β₃ (left lanes) or Flag-β₃ΔGK (right lanes) were coexpressed in Cos7 cells (see lysates, right panel). Unlike Flag-β₃, Flag-β₃ΔGK was not bound to V5-His-β₃, as revealed by His pulldown (left panel). We used 0.5 μg of each plasmid for cotransfection.

Ca_vβ oligomers bind to the α_1C subunit at its I-II linker

To find out whether binding to α_1C affects oligomerization of Ca_vβ subunits, Flag-α_1C, α₂δ, V5-His-β₃, and Venus-β₃ were coexpressed in Cos7 cells, and the plasma-membrane proteins were labeled by extracellular biotinylation before cell lysis (Fig. 4A). Sequential His- and avidin-pulldown procedures identified differentially labeled β₃ associated with Flag-α_1C, thus confirming that β₃ oligomers exist in the plasma-membrane-bound α_1C/α₂δ/β₃ channel complexes. Because the AID region in α_1C is the major binding site for Ca_vβ, we focused on the interaction of oligomeric β subunits with I-II linker (Fig. 4B). Flag-β₃ and Venus-β₃ were coexpressed in the absence (Fig. 4B, lane 2) or presence of the increasing amounts of Venus-I-II linker (25) (Fig. 4B, lanes 3–5). Immunoblot analysis with an anti-GFP antibody showed that the complexes pulled down by an anti-Flag antibody contained Venus-β₃ independently on AID. The oligomerization of β₃ was not affected by the presence of AID. In fact, I-II linker binds to homo- and heterooligomeric Ca_vβ complexes, as shown in Fig. 4C. Flag-β₃ was coexpressed in Cos7 cells with Venus-β_1b, Venus-β₃ or Venus-β_2d, or mVenus (Fig. 4C, lane C) in the presence of V5-His-I-II linker. Two-step purification of the oligomeric Ca_vβ complexes bound to I-II linker by His pulldown followed by immunoprecipitation with an anti-Flag antibody revealed the presence of homo- and heterooligomeric complexes composed of I-II linker and ≥2 individual β subunits. Since I-II linker does not oligomerize (Fig. 4D), these results suggest that I-II linker may dock complexes of Ca_vβ subunits. Taken together, our data show that Ca_vβs oligomerize at α_1C subunits residing in the plasma membrane and bind to the AID.

Figure 4.

Ca_vβ oligomers bind to α_1C at the α_1C I-II linker. A) Oligomers of β₃ exist in the plasma-membrane-bound α_1C/α₂δ/β₃ channel complexes. Flag-α_1C, α₂δ, V5-His-β₃, and Venus-β₃ were coexpressed in Cos7 cells, and the plasma-membrane proteins were biotinylated. Anti-Flag IP of cell lysate isolated β₃ bound to α_1C. His pulldown confirmed the interaction of V5-His and Venus-β₃ associated with α_1C, while avidin pulldown confirmed it in the plasma-membrane fraction of α_1C. B) Oligomerization of β₃ is not affected by AID. Co-IP of homooligomers of Flag-β₃ and Venus-β₃ in the absence (lanes C, 2) or presence of increasing amounts of Venus-I-II linker containing AID (25) (0.1–0.4 μg/transfection, lanes 3–5). C) Evidence that I-II linker binds to Ca_vβ homo- and heterooligomers. Co-IP of Flag-β₃ and V5-His-I-II linker with Venus-labeled β₃, β_1b, or β_2d. Constructs were coexpressed in Cos7 cells. His pulldown was performed to isolate β subunits bound to AID, and the subsequent anti-Flag co-IP of the isolated β subunits was performed to check whether they were oligomers or monomers. D) Lack of oligomerization of I-II linker. His pulldown of V5-His-I-II linker and Venus-I-II linker coexpressed in Cos7 cells revealed the absence of Venus-I-II linker in the precipitate. Plasmids used for cotransfection were as following: Flag-α_1C (1.2 μg), α₂δ (1.0 μg), V5-His-β₃ (0.5 μg), Venus-Ca_Vβs (0.8 μg), V5-His-I-II linker (0.3 μg), and Venus-I-II linker (0.4 μg or as indicated).

Functional modulation of calcium channel associated with Ca_vβ oligomerization

It is interesting that all active fragments of GK (see Fig. 3C) increased the binding between Venus- and Flag-labeled β₃ as compared to control (Fig. 5A, lane 1), GKC showing the strongest enhancing activity. The non-GK fragments had no effect on the binding between different β₃ molecules (Supplemental Fig. S5). Effect of GKC on β₃ oligomerization was further confirmed by the enhancement of a ladder-like pattern on the blot (Fig. 5B, arrows), when Flag-β₃ was coexpressed with GKC (Fig. 5B, lanes 2, 3). The maximum augmentation of β₃ oligomerization was reached (Fig. 5C, red arrow) on cotransfection of β₃ with GKC at 2:1 ratio (plasmid, w/w). The presence of GK fragments did not affect the interaction of β₃ with I-II linker (Supplemental Fig. S6A), although some of these fragments exhibited direct binding activity to the I-II linker (Supplemental Fig. S6B). We then confirmed that coexpression of GKC stimulates β₃–α_1C association in the confines of the plasma membrane (Fig. 5D). Flag-α_1C, α₂δ, and Venus-β₃ were coexpressed in Cos7 cells in the presence of increasing amounts of GKC, and the membrane bound complexes were labeled by cell surface biotinylation and isolated by avidin pulldown under conditions described in Fig. 4A. In the membrane fraction, increased expression of GKC leads to a significant increase in the amount of β₃ coimmunoprecipitated with approximately the same amount of α_1C (P<0.05, n=5) suggesting an increase in the β₃/α_1C ratio in the plasma membrane complexes. In the cytosolic fraction, there was no such increase in the β₃/α_1C ratio (Fig. 5D, E). Given the overexpression of β₃ (as evident from the comparison of β₃ and α_1C ratio in the lysate with that in the coimmunoprecipitate), this result demonstrated that the coexpression of GKC with the calcium channel complex increases the efficiency of β₃ association with the plasma membrane α_1C molecules independently on the excess of β₃ but in a GKC dose-dependent manner.

Figure 5.

Functional significance of Ca_vβ oligomerization. A-C) GKC increases oligomerization of β₃. A) mVenus (lane 1) or indicated Venus-labeled fragments (lanes 2–9) were coexpressed with a mixture of Flag-β₃ and Venus-β₃. The β₃ complexes were immunoprecipitated by an anti-Flag antibody and immunoblotted by an anti-GFP antibody. B) Flag-β₃ (0.5 μg) was coexpressed with mVenus (lane 1) or increasing amounts of Venus-GKC (lanes 2, 3). Top panel: anti-Flag IP and immunoblotting without DTT treatment. Arrows mark oligomers; asterisk indicates monomer. C) Mixture of Flag-β₃ (0.3 μg) and Venus-β₃ (0.5 μg) plasmids was supplemented with mVenus (lane C, control) or increasing amounts of Venus-GKC vector (0.05–1.2 μg/transfection, lanes 1–6). Flag-β₃ was immunoprecipitated from Cos7 cells by an anti-Flag antibody and immunoblotted with Flag or GFP antibodies as indicated. Red arrow indicates an apparent maximum augmentation of β₃ oligomerization. D, E) Coexpression of GKC stimulates β₃-α_1C association in the plasma membrane. D) Flag-α_1C, α₂δ, and Venus-β₃ (1.2, 1.0, and 0.8 μg, respectively) were coexpressed in Cos7 cells in the presence of mVenus (control) or increasing amounts of Venus-GKC (0.01–0.4 μg/transfection). Plasma membrane fraction of the Flag coimmunoprecipitates was isolated by avidin pulldown, as described in Fig. 4A, and the flow-through fractions were analyzed for cytosolic complexes. Resulting immunoblot is representative of 5 independent experiments, which exhibited similar levels of band intensity. E) Band intensities of β₃ were normalized to the respective band intensities of α_1C to estimate the β₃/α_1C ratio in the channel complex. Histogram is average of 5 measurements with Image J to quantify the band intensities (mean±sd); Students' t test was used for statistical analysis in comparison to control. *P < 0.05; paired 2-tailed distribution. F, G) Augmentation of β₃ homooligomerization enhances the calcium current density. Cos7 cells were transfected with α_1C (1.2 μg), α₂δ (1 μg), and either 0.8 (black circles) or 1.6 μg (open circles) of β₃, GKC (0.4 μg, gray), or a mixture of 0.8 μg β₃ and 0.4 μg GKC (red). Shown are respective current-voltage relationships (F) and representative traces of the Ca²⁺ currents recorded in response to a stepwise shift of membrane potential from the holding potential of −90 to +20 mV and scaled to the peak current (G). H, I) Effect of heterooligomerization on current-voltage relationship (H) and kinetics of inactivation of the Ca²⁺ current (I). Cos7 cells were transfected as above except using 0.4 μg each β₃ and β_2d (blue). For comparison, respective β_2d channel data are shown in magenta (25). J) Lack of calcium channel activity (at +20 mV) in Cos7 cell expressing α_1C (1.2 μg) and α₂δ (1 μg).

Coexpression of the calcium channel with β₃ and GKC (2:1, plasmid w/w) in Cos7 cells resulted in a 2.3-fold increase of the calcium current density (red; n=10) as compared to β₃ (n=5, P<0.01; Fig. 5F, solid circles), whereas doubling the cDNA amount of β₃ in the α_1C/α₂δ/β₃ cotransfection in the absence of GKC lead to only a slight increase in the current density (n=5; Fig. 5F, open circles). In the control experiment, coexpression of α_1C and α₂δ in the absence of Ca_vβ produced silent channels (Fig. 5J) (12). Coexpression of α_1C and α₂δ with GKC in the absence of β₃ generated a low-amplitude calcium current (Fig. 5G, gray trace) characterized by the voltage dependence shifted by 20 mV to the hyperpolarizing direction and a notably slower inactivation kinetics of the calcium current as compared to β₃ or β₃ + GKC (Fig. 5F, G). No significant changes in reversal potential (104.5±3.4 mV), voltage at 50% of current activation (V_0.5=25.9±0.9 mV), slope factor k_I-V (−16.8±0.3), and kinetics of the current decay (Fig. 5G, black and red traces) were observed. In contrast, heterooligomerization of β₃+β_2d (1:1; Fig. 5H, blue) caused a 20-mV depolarization shift of the peak I-V curve (V_0.5=58.1±6.9 mV, k_I-V=−20.0±1.4, n=9) with regard to those of β₃ (Fig. 5H, black solid circles) and β_2d (Fig. 5H, magenta open circles; ref. 25). The rate of inactivation (Fig. 5I, blue trace) was not intermediate to that of β₃ (Fig. 5I, black trace) and β_2d (Fig. 5I, magenta trace) but notably reduced. Thus, Ca_vβ homo- and heterooligomerization significantly affect different functional properties of calcium channels.

DISCUSSION

The major conclusions of our study are that all major Ca_vβ subunits form oligomers in both naturally occurring and recombinant Ca_v1.2 calcium channels and that Ca_vβ homooligomerization affects the calcium current density, while heterooligomerization may also affect voltage-dependence and inactivation kinetics of the current. These findings support the hypothesis of Tareilus et al. (18) and add new dimensions to our understanding of the roles of Ca_vβs. First of all, Ca_vβs modulate the calcium channel functions by direct association with α₁ at multiple sites. In the Ca_v1.2 channel, those sites include AID (11) and IQ motifs (35), while β₂ subunits, in addition, bind with their unique C-terminal domain to the pre-IQ LA region (25). As we have shown previously (25), each of these 3 determinants is sufficient to support the calcium current in the presence of Ca_vβ (e.g., β_2d). Because Ca_vβ oligomerization occurs independently on association with α_1C (Fig. 4B), each of these sites probably interacts with higher order Ca_vβ complexes rather than monomers. Thus, multifaceted Ca_vβ modulation extends into a realm where homo- or heteromeric Ca_vβ complexes of variable size mediate the regulatory functions, but how many complexes are involved is less clear.

Another critical element of α_1C-Ca_vβ interaction seems to be their ratio. A stoichiometric 1:1:1.3 ratio of α_1B, α₂δ, and β₃ is sufficient for the purified and reconstituted N-type Ca_v2.2 calcium channel to retain such essential functional properties of the native channel as distinct voltage dependence and sensitivity to ω-conotoxin blockade (13). The finding of Ca_vβ oligomerization puts a new twist on the role of Ca_vβ while maintaining the basic structural principles of α_1C-Ca_vβ interaction. In a double-transgenic mouse model (22), an enhanced expression of β₂ over constitutive Ca_v1.2 caused increase of the single channel activity, suggesting a mechanistic link to the development of heart failure. In a different study (20), an adenovirus-assisted overexpression of β_2a in ventricular myocytes caused a 3-fold increase in the calcium current density. These results may well be explained by the homo- or heterooligomerization of Ca_vβ subunits. Indeed, an augmentation of β₃ homooligomerization increased the β₃/α_1C ratio on the plasma membrane as well as the calcium current density without significant changes in voltage-dependence and kinetics of the current (Fig. 5F, G). Thus, we have a case where Ca_vβ oligomerization may provide a mechanistic explanation for the relationship between the properties of the channel and levels of Ca_vβ expression. This basic mechanism may have a role in both the developmental-dependent and pathophysiological changes of calcium channel functioning such as increased frequency of Ca_v1.2 coupled gating events in hypertensive smooth muscle cells and Timothy syndrome (LQT8) (37). Moreover, Ca_vβ heterooligomerization may contribute to the remodeling of calcium channel properties in response to the increased expression of β₂ subunits in heart failure (22) and be a pathogenic factor in hypertension (38) and other pathophysiological conditions.

Binding studies (Fig. 1D) showed that ≥3 individual β₃ proteins coassemble to from oligomeric complex. However, the mechanism of Ca_vβ oligomerization remains unresolved, although 3 determinants appear to be identified in our study. The oligomerization of “classic” MAGUK (e.g., PSD-95) is mediated by swapping of complementary SH3 subdomains (39). Surprisingly, domain mapping experiments did not show an involvement of the SH3 subdomain of β₃ in intermolecular interaction, all 3 identified determinants of Ca_vβ oligomerization being localized to GK (Fig. 3). These results are consistent with the previous report that Ca_vβ GK fragments expressed in bacteria dimerize and may modulate Ca_v2.3 channel gating (40). Although alanine mutations of each of the identified determinants inhibited binding of the respective fragments to β₃ (Supplemental Fig. S4), the combined mutation of all 3 regions was not sufficient to eliminate oligomerization of β₃. Which unidentified yet structural factors contribute to assembly of Ca_vβ subunits, why some isolated GK fragments (e.g., GKC; Fig. 5C) promote Ca_vβ oligomerization, and how these properties contribute to Ca_vβ regulation of calcium signaling are questions remaining to be answered.

We have shown previously that the structural organization of Ca_v1.2 channels in the plasma membrane depends on the type of Ca_vβ subunits present (1). Although all major types of Ca_vβ tend to form oligomers, the extent of oligomerization may vary. Modulation of Ca²⁺ channel through oligomerization of Ca_vβs, described here for Ca_v1.2, may provide a common mechanism in muscle, secretory, and neuronal Ca_v1 and Ca_v2 channels, which also require Ca_vβs for function. The recent finding of interaction between β_1b subunits of Ca_v2 calcium channels revealed recently by the yeast 2 hybrid technique (see Supplemental Table 1 in ref. 41) supports this hypothesis.