INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Voltage-dependent Ca²⁺ channels (VDCCs) are composed of a pore-forming 1-subunit, associated with accessory subunits, including a cytoplasmic -subunit and largely extracellular 2-subunit (Dolphin, 1998, for review). -Subunits regulate a number of properties of VDCCs, increasing current density, in part by recruitment of channels into the plasma membrane (Brice et al., 1997; Bichet et al., 2000) and hyperpolarizing the voltage dependence of activation and also steady-state inactivation (except in the case of 2a) (De Waard and Campbell, 1995; Stephens et al., 1997). A role for -subunits in G-protein inhibition of calcium channels has also been reported (Campbell et al., 1995; Bourinet et al., 1996; Qin et al., 1997; Roche and Treistman, 1998; Meir et al., 2000). This has been interpreted in terms of an interaction at an overlapping binding site (Bourinet et al., 1996). However, we have shown that there is not a simple competition between -subunits and G dimers (Meir et al., 2000; Canti et al., 2000). Indeed, in a system (COS-7 cells) in which no endogenous -subunit protein was detected by immunocytochemistry, the presence of heterologously expressed -subunits was essential for the relief of G-mediated inhibition by prepulse facilitation (Meir et al., 2000).

The point has recently been made that we do not know how many -subunits bind physiologically to a functional calcium channel (Birnbaumer et al., 1998). Three G and -subunit interaction sites have been identified on various 1-subunits, and a high-affinity site within the I-II loop (Zamponi et al., 1997; De Waard et al., 1997) and other interaction sites of lower affinity were measured on the C terminus (Qin et al., 1997; Walker et al., 1998) and the N terminus (Walker et al., 1999; Canti et al., 1999; Stephens et al., 2000). However, none of these studies has been able to address whether in an intact channel the same -subunit binds with high affinity to the I-II linker and interacts via different domains with lower affinity at the N and C termini. It is certainly possible that these three intracellular domains of 1-subunits all form part of a complex binding pocket for both a single -subunit and, when present, a G dimer. Alternatively, several -subunits might bind to different sites on the same channel.

The aim of the present work was to examine the dependence of a number of the effects of VDCC -subunits on the concentration of 3-subunit expressed in Xenopus oocytes. We wished to determine whether we could distinguish different concentration dependencies, which would provide evidence for more than one binding process for -subunits. We have examined the effect of 3-subunit concentration on the maximum conductance of 1B, as a measure of expression level, on the voltage dependence of steady-state inactivation, and on the rate of prepulse facilitation during G protein inhibition. We have not examined the effects of 3 on inactivation kinetics, because in a previous study, we found little effect of 3 on this parameter for 1B (Stephens et al., 2000).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Expression of constructs

The following cDNAs were used: rabbit 1B (GenBank L15453) and the I49A mutant of 1B (Canti et al., 2000), rat 3 (M88751), rat 2-1 (M86621), and rat D2_long dopamine receptor (X17458, N5G) in the vector pMT2. Xenopus oocytes were prepared, injected, and maintained as described previously (Canti et al., 2000). The 1B, 3, 2-1, and D2 receptor cDNAs (1 ng nl¹, except for the -subunit) were mixed in a ratio of 3:4:1:3, respectively, and 4 nl was injected (except when otherwise stated) into the nuclei of stage V and VI oocytes. The 3 cDNA was diluted up to 1:500 before mixing, and when 3 cDNA was not used, it was replaced by buffer. The entire range of 3 cDNA concentrations was examined in each experiment to minimize batch-to-batch variation in oocyte expression levels.

Antisense oligonucleotides

The 25-mer antisense oligonucleotide DNA (ODN) sequence used was GCA CTC CTC ATC CAG CGC TCC ACA G (Tareilus et al., 1997). The scrambled nonsense ODN sequence was CTC GTA GCG CAC CAC CTA CCT CAG C (Gibco, Paisley, UK). The nucleotides were phosphodiester linked, except the first and last three nucleotides, in bold, which were phosphorothioate linked, to reduce degradation. In these experiments 1B cDNA, 2-1 cDNA, and ODN (4 or 40 µM) were mixed in a ratio of 3:1:3 before injection of 9 nl per oocyte.

Expression and purification of H6C3

A full-length 3 with C-terminal hexahistidine tag (H6C3) was produced by two-stage polymerase chain reaction (PCR) to remove an internal AflIII site (10 cycles each stage) using Pfu polymerase (Stratagene, Amsterdam, The Netherlands), 3 in pMT2 as template, and the following primers for stage one: forward, 5'CCCACATGTATGACGACTCC3'; reverse, 5'CGGGGGGACATGCTCCGCCTGCTTTT3'.

The resulting PCR product was purified and used as the forward primer in the second stage reaction with 5'GGGAATTCTCAATGATGATGATGATGATGGTAGCTGTCCTTAGGCCA3' as the reverse primer. The resulting PCR product (~1.5 kb) was purified from agarose gel, digested with AflIII and EcoRI, and sub-cloned into NcoI- and EcoRI-digested pET28b (Novagen, Nottingham, UK) to give H6C3-pET28b. BL21 Codon Plus (IRL) Escherichia coli (Stratagene) were transformed with H6C3-pET28b, and cultures were grown overnight to saturation at 37°C in LB (pH 5.5) supplemented with kanamycin, chloramphenicol, and 1% w/v glucose, diluted 1:10 with the same medium, and grown for an additional 3 h before cooling to room temperature and induction with 0.5 mM isopropylthio--D-galactoside. The cultures were grown for 3 h after induction and harvested by centrifugation; pellets were then stored at 70°C until required.

E. coli pellets containing expressed H6C3 protein were lysed at 4°C by sonication in 20 mM phosphate buffer (pH 7.4), containing one protease inhibitor tablet (Complete EDTA-free, Roche Diagnostics, Lewes, UK) per liter of pelleted culture. Solid NaCl was added to the lysate to a final concentration of 1 M NaCl before the lysate was cleared at 20,000 × g at 4°C for 15 min. Imidazole solution (pH 7.4) was then added to the resulting supernatant to give a final concentration of 20 mM before loading onto a Ni²⁺-primed 5-ml HiTrap chelating column (Amersham Pharmacia, Uppsala, Sweden) equilibrated with load buffer (20 mM phosphate buffer, pH 7.4, 1 M NaCl, 20 mM imidazole, 0.15% w/v octylglucoside, and one protease inhibitor tablet per 100 ml). The column was washed thoroughly with wash buffer (same as load buffer but with 40 mM imidazole) before H6C3 was eluted from the column in elution buffer (same as load buffer but with 200 mM imidazole).

Peak UV₂₈₀ absorbance fractions were rapidly buffer exchanged on a Sephadex G-25 (Amersham Pharmacia) column into ion-exchange (IEX) buffer (20 mM 2-[N-morpholino] ethanesulfonic acid, pH 6.0, one protease inhibitor tablet per 200 ml) supplemented with 500 mM NaCl, before dilution 1:10 with IEX buffer. The diluted sample was loaded onto a 1-ml SP-Sepharose HP column (Pharmacia), and the column was washed with IEX buffer before H6C3 proteins were eluted in a linear gradient of 0-1 M NaCl in IEX buffer. Fractions containing H6C3 were identified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis, with Coomassie blue staining, before concentrating to 0.5 mg ml¹ using Centriplus concentrators.

Surface plasmon resonance binding assay

All assays were performed on a Biacore 2000 (Biacore, Uppsala, Sweden) at 25°C in 10 mM Hepes, 500 mM NaCl, 3 mM EDTA, 0.005% polysorbate-20, pH 7.4. Glutathione S-transferase (GST) and the GST_1BI-II linker fusion proteins were purified as previously described (Bell et al., 2001) and immobilized on individual flow cells of a CM5 dextran chip using an anti-GST polyclonal antibody kit (Biacore) according to the manufacturer's instructions. To obtain identical molar loadings of the different molecular mass proteins the following resonance unit (RU) correction factors were used during immobilization: GST = 1; GST_1BI-II linker = 1.57. H6C3 protein was diluted as stated, and H6C3 injections were performed using a flow rate of 50 µl min¹ for 5 min.

Determination of the amount of 3-subunit in Xenopus oocytes

Oocytes were injected intranuclearly, as described previously (Canti et al., 2000), with 1B/2-1 subunits and either 3, 45, 720, or 1440 pg of 3 cDNA. After 5 days of incubation at 18°C, individual oocytes (following brief electrophysiological recording to verify expression of 1B I_Ba) were lysed in hypotonic buffer (10 mM Tris, pH 7.4, containing protease inhibitors (Roche) plus 1 mM EDTA), solubilized in 2% SDS, assayed for protein content, diluted as necessary to remain within the linear range (see Fig. 2), and separated by SDS-PAGE, followed by immunoblotting for 3-subunits. H6C3 subunit standards (0.2-3 ng) were run in parallel. The polyvinylidene fluoride membranes were blocked with 3% bovine serum albumin for 5 h at 55°C and incubated overnight at 20°C with a 1:500 dilution of anti-3 monoclonal antibody raised against residues 418-484 of human 3 (Day et al., 1998; Bogdanov et al., 2000). This region is 97% and 80% identical to the corresponding rat and Xenopus 3 sequence, respectively. Because the antibody is a monoclonal, it is highly likely to bind to an epitope that is well conserved between the three species. The primary antibody was followed by a 1:1000 dilution of goat anti-mouse IgG-horseradish peroxidase conjugate (BioRad Laboratories, Richmond, CA) for 1 h at 20°C. Detection was performed using ECL (Amersham Pharmacia Biotech), the films were subsequently scanned, and the amount of 3 subunit in each sample was determined using Imagequant (Molecular Dynamics, Sunnyvale, CA) from the standard curve of purified H6C3 protein on the same blots. To estimate the -subunit content of plasma membrane and internal (cytosolic) fractions, oocytes were placed in hypotonic buffer (5 mM Hepes, pH 7.4, with protease inhibitors) for 30 min at room temperature. The plasma membrane was then isolated with fine forceps from the cell interior contents (cytosol). The two fractions from a given number of oocytes were pooled and each homogenized in hypotonic buffer. The plasma membrane fractions were washed three times in hypotonic buffer by centrifugation (100,000 × g for 30 min at 4°C). Both fractions were solubilized in SDS-PAGE buffer and then assayed for 3 and total protein content.

Electrophysiological recording of I_Ba

Two-electrode voltage-clamp recordings from oocytes were performed as described previously (Canti et al., 2000). Oocytes were held at 100 mV, and currents were evoked every 15 s using 5 mM Ba²⁺ as charge carrier, unless otherwise stated. The 1B I_Ba was always recorded using the same sequence and timing of protocols. For every cell, a maximal concentration of quinpirole (100 nM) was applied, after which marked and prolonged over-recovery (>30 min) was usually observed due to the removal of tonic inhibition (Canti et al., 2000). This provided a stable baseline from which all experimental measurements were then made. All experimental data regarding the effect of quinpirole application were obtained during a second application of the drug. Current amplitude measurements were taken at 20 ms after the start of test pulses except in the steady-state inactivation protocol, where currents were measured at their peak. All values are mean ± SEM, and statistical significances were determined by Student's t-test.

The observed 1B I_Ba currents were not contaminated with endogenous oocyte currents, as these were measured in non-injected oocytes from every batch and were less than 10 nA for the maximum I_Ba in 5-10 mM Ba²⁺. All currents were leak-subtracted on-line using a P/4 protocol.

Data analysis

Data were analyzed using Clampfit (Axon Instruments, Foster City, CA) and ORIGIN 5.0 (Microcal, Northampton, MA).

Current-voltage (IV) curves were fit to a combined Boltzmann and linear function:
where V_t represents the test potential, G_max the maximum conductance, V_rev the apparent reversal potential, V_50,act the potential for half-activation, and k the slope factor.

The steady-state inactivation curves were either fit by a single Boltzmann function:
where I_max is the peak current value, V_50,inact is the potential for half-inactivation, or to a double Boltzmann function:

where V_50,inact,a and k_a are the parameters associated with the component A of amplitude a, and V_50,inact,b and k_b are the corresponding parameters associated with the second component B of amplitude b = (1  a). The ² values associated with each fit were used to assess whether a single or double function best fit each individual dataset.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

To examine how the biophysical effects of -subunits are dependent on expressed -subunit concentration, we have expressed a range of concentrations of the VDCC 3-subunit by injecting increasing amounts of 3 cDNA, from 0 to 720 pg, together with a fixed concentration of 1B and 2-1 cDNA into Xenopus oocytes. The rat 3-subunit was used because it is one of the main -subunits associated with 1B in native tissues (Witcher et al., 1993; Burgess et al., 1999). Furthermore, using a PCR strategy with primers in regions of homology between all mammalian -subunits, only 3 cDNA was isolated from Xenopus oocytes, and it is therefore reasonable to conclude that it is the only -subunit present in these cells (Tareilus et al., 1997). In agreement with this, comparison of results obtained with the 3 antibody used in this study, and a pan--subunit antibody (Campbell et al., 1995), revealed no additional -subunit immunoreactive bands on immunoblots (results not shown).

Determination of the endogenous 3-subunit concentration in Xenopus oocytes and that resulting from injection of increasing 3 cDNA

The relationship between the concentration of 3 cDNA injected into the Xenopus oocytes and the amount of 3 protein expressed was examined by constructing a standard curve with H6C3 (Fig. 1 a) and using this to determine the amount of 3 protein in non-injected and 3 cDNA-injected oocytes. The expression of 3 protein was linear up to the highest amount of injected 3 cDNA examined (1.44 ng), which is twice the maximum amount used in the electrophysiological experiments (Fig. 1 b). The endogenous 3 level in 1B/2-1-injected oocytes was very similar to that in non-injected oocytes (112 ± 2% of the non-injected level; n = 10). Furthermore, it was reduced by 47 ± 4% (n = 10) after 40 µM A/S ODN injection (Fig. 1 b). The amount of endogenous 3 per oocyte was 0.56 ± 0.02 ng (n = 15), which corresponds to an average concentration of ~17 nM, assuming an oocyte radius of 0.5 mm. The ratio of 3 protein associated with the plasma membrane and internal fraction was examined at the different concentrations of 3 cDNA injected into Xenopus oocytes (Fig. 1 c). Although the amount of 3-subunit was linearly dependent on 3 cDNA concentration in both fractions (Fig. 1 c, inset), that in the internal (cytosolic) fraction increased to a much greater extent, as expected, because 3 is unlikely to be plasma membrane associated unless bound to a calcium channel (Bogdanov et al., 2000). The ratio (per oocyte) of 3 protein in the cytosol/membrane rises to over 30 at the highest concentration of 3 cDNA used.

View larger version (29K): [in this window] [in a new window]   FIGURE 1   Dependence of Xenopus oocyte 3 protein level on 3 cDNA injected. (a) Standard curve for 3 protein from 0.2-3 ng (mean ± SEM; n = 5). H6C3 was purified and detected using a 3 monoclonal antibody as described. The same protein band was also detected with an anti-hexahistidine antibody (results not shown). The straight line is a linear regression fit (y = 12.6x; R = 0.993), from which the amount of 3 in oocyte lysates was determined. The inset at the top shows a representative immunoblot of the standard curve, using the H6C3 protein whose purity is shown in Fig. 3 a. (b) Immunodetection of 3 subunits in total oocyte lysates using the 3 monoclonal antibody. (Inset) Representative immunoblot. Oocytes were injected as follows: lane A, no injected DNA; lane B, 1B/2-1 + A/S ODN 40 µM; lane C, 1B/2-1 alone; lanes DG, 1B/2-1 + 3 pg, 45 pg, 720 pg, and 1.44 ng of 3 cDNA, respectively. The dilutions used are indicated above the blots. For the graph, the total optical density of the 3 bands was determined and the 3 level calculated from the standard curve and plotted against 3 cDNA injected (). Data are mean ± SEM of 10-15 values each. The line is a linear regression fit to the data (y = 1.79 + 0.061x; R = 0.995). (c) 3 subunits were quantified by immunodetection in isolated oocyte plasma membrane and internal protein fractions using the 3 monoclonal antibody. The ratio of 3 subunit/oocyte in the internal/oocyte membrane fraction is shown as a function of increasing amounts of injected 3 cDNA (3, 45, 720, and 1440 pg; ) in addition to 1B and 2-1 subunits. Data are the ratio of mean values for of 8-20 determinations each. The line is the linear regression fit to the data (R = 0.987). The inset plot shows separately the increase in 3 subunit in the internal () and oocyte membrane fractions () from which the ratios were determined.

Dependence of the 1B G_max on the concentration of 3-subunit resulting from co-expression of increasing amounts of 3 cDNA

We first examined the influence of increasing the 3-subunit concentration on the G_max obtained from the 1B IV relationships (Fig. 2). Examples of I_Ba at 0 mV are shown in Fig. 2 a, i, and the mean IV relationships in Fig. 2 b. In the absence of expressed 3-subunit, the G_max was 0.014 ± 0.002 µS (n = 13). The G_max showed strong dependence on the expressed 3-subunit, with 15 pg of injected 3 cDNA giving a plateau G_max of 0.029 ± 0.004 µS (n = 19). The V_50,act showed a more gradual hyperpolarizing shift with increasing 3-subunit to a maximum of 15.3 mV (see Fig. 2 e), resulting in an increased current amplitude at 0 mV, after the conductance increase had saturated, because of the increased driving force at the more hyperpolarized potentials (Fig. 2, a and b). There was no effect of 3-subunit expression on the V_rev (see Fig. 2).

View larger version (34K): [in this window] [in a new window]   FIGURE 2   Modulation of 1B biophysical properties by co-expression of VDCC  subunits or 3 antisense ODNs. The 1B channel was expressed with 2-1 and the dopamine D2 receptor and either without or with 3-720 pg of 3 cDNA or with A/S or N/S ODNs. (a) Example currents at 0 mV: (i) for the different concentrations of 3 cDNA as given; (ii) for A/S ODN (4 and 40 µM), compared with N/S ODN (40 µM). (b) IV relationships for the different concentrations of 3 cDNA (symbols given in key) were obtained from voltage steps between 40 and +50 mV, and IBa at 20 ms was fit to a modified Boltzmann function (see Materials and Methods). The V50,act values were (mV) 2.8 ± 0.9 (0 pg of 3 cDNA, n = 13); 1.4 ± 1.6 (3 pg, n = 13); 1.2 ± 0.9 (6 pg, n = 19); 4.3 ± 0.9 (15 pg, n = 19); 8.5 ± 1.0 (45 pg, n = 10); 9.8 ± 0.8 (90 pg, n = 12); 11.9 ± 1.0 (360 pg, n = 9), and 12.5 ± 0.9 (720 pg, n = 9). The k (slope factor) values were between 3.8 and 5.7 mV. The Vrev values in all groups lay between 44.3 mV and 50.3 mV, with no statistically significant differences between them and no effect of 3 subunit co-expression. (c) IV relationships comparing the A/S and N/S conditions: , 40 µM N/S ODN, n = 9; , mean of 4 and 40 µM A/S ODN, n = 8. (d) Normalized IV relationships for the N/S and A/S ODN conditions (same symbols as in c), showing the lack of effect on the V50,act. In comparison, the normalized IV relationship for the 45 pg of 3 cDNA condition is also given (, n = 10). (e) The Gmax was determined as the slope conductance from the linear region of the IV relationships, normalized to the value for 1B/2-1 and plotted against the amount of 3 protein calculated from the linear regression fit to Fig. 1 b (left axis, ). The open circle represents the Gmax determined following the A/S ODN relative to the N/S ODN data from c. The data are fit by a logistic function with a midpoint of 0.54 ng of 3protein and a power coefficient of 1.9. The arrow represents the measured endogenous 3 protein level. The amount of 3 cDNA injected is indicated above each point. The V50,act values were determined from the IV relationships in b and are plotted against the amount of 3 protein (right axis, ). The data are fit by a logistic function with a midpoint at 2.71 ng of 3 protein and a power coefficient of 1.7.

Effect of a 3 antisense ODN on 1B calcium channel expression in Xenopus oocytes

To determine to what extent the expression of 1B currents, in the absence of co-expressed  subunits, relies on the presence of the endogenous Xenopus oocyte -subunits, a 3 A/S ODN was injected, together with 2-1 cDNA and a higher concentration of 1B cDNA (see Materials and Methods). A scrambled N/S ODN was used as a control. Two different A/S ODN concentrations were used (4 and 40 µM), which gave very similar results. For the oocytes injected with the N/S ODN, 11/12 oocytes expressed I_Ba (Fig. 2 a, ii), and from the IV relationship (Fig. 2 c) the G_max was 0.036 ± 0.006 µS (n = 9). For 4 and 40 µM 3 A/S ODN, 5/15 and 5/20 oocytes, respectively, expressed I_Ba above background noise (Fig. 2 a, ii), and from these oocytes the G_max was 0.018 ± 0.004 µS (n = 4) and 0.019 ± 0.007 µS (n = 4), respectively (see Fig. 2 c for a comparison of the N/S and pooled A/S data). It is clear from the normalized IV relationships that although the 3 A/S ODN reduced G_max by ~50%, it did not significantly depolarize the V_50,act compared with either the N/S ODN (Fig. 2 d) or the 0-pg 3 cDNA control. The V_50,act was +2.8 ± 0.9 mV for 1B/2-1 (n = 13), +1.4 ± 1.7 mV for 1B/2-1 + 40 µM N/S ODN (n = 9), +2.5 ± 2.1 mV for 1B/2-1 + 4 µM A/S ODN (n = 4), and +4.5 ± 0.7 mV for 1B/2-1 + 40 µM A/S ODN (n = 4). In comparison, 45 pg of 3 cDNA, a concentration that produced an approximate doubling of the G_max compared with the 1B/2-1 control (Fig. 2 b) and an ~10-fold increase in 3 protein (Fig. 1 b), also caused a marked hyperpolarizing shift in the V_50,act to 8.5 ± 1.0 mV (n = 10; Fig. 2d).

By estimating the 3 protein level for each 3 cDNA concentration used, from linear regression of the data in Fig. 1 b and taking into account the dilution factor used, we then determined the dependence of 1B G_max on amount of 3 protein expressed per oocyte (Fig. 2 e, ). The data, including the point generated from the A/S experiment (), are well fit by a sigmoid concentration-response curve. The midpoint of this curve (0.54 ng of 3 protein per oocyte, approximately equivalent to 16.3 nM) occurs at about the endogenous level of 3 protein in the Xenopus oocytes, and the plateau is reached at ~2.3 ng of 3 protein per oocyte (69 nM).

We also examined the concentration dependence of the V_50,act (Fig. 2 e, ) and found it to be fit by a sigmoid concentration-response curve with a higher midpoint at 2.7 ng of 3 protein.

Binding affinity of the purified 3 protein for the I-II linker of 1B

The same purified H6C3 protein used for the quantification of endogenous and expressed 3 levels (Fig. 3 a) was also used to examine its binding to the 1B I-II linker immobilized as a GST fusion protein (Fig. 3 a) on a Biacore 2000. The binding of H6C3 was concentration dependent and reversible, between 5 and 40 nM H6C3, the highest concentration examined (n = 3-6 for each concentration; example sensorgrams given in Fig. 3 b). The mean sensorgram for 20 nM H6C3 is shown in Fig. 3 c (n = 6). Both the on- and off-rates were well fit by a single-exponential function to the mean data, and the K_D was determined to be 26 nM (see legend to Fig. 3 c). Similarly, for the 10 nM 3 binding data, the mean K_D was calculated to be 19.6 nM, from fits of the on- and off-rates for the mean data (n = 3; results not shown). These estimates of the affinity of H6C3 binding to the 1B I-II linker are in very good agreement with the functional data. The specific k_on for H6C3 binding to the 1B I-II linker in this system was estimated to be 2.0 × 10⁵ M¹ s¹ (see Fig. 3 c).

View larger version (25K): [in this window] [in a new window]   FIGURE 3   Binding of 3 protein to GST fusion proteins of 1B I-II linker. (a) Coomassie-blue-stained 12.5% SDS-PAGE gel (Brabet et al., 1988) of the proteins used in the surface plasmon resonance binding assay. Approximately 0.5 µg of the following proteins was loaded: GST, GST1BI-II loop, and H6C3. The positions of molecular mass markers (Sigma) are shown for comparison. (b) Examples of Biacore 2000 sensorgrams. Approximately 4 fmol of the fusion protein or GST was immobilized via the anti-GST antibody on an individual flow cell of a CM5 dextran sensor chip. The VDCC 3 protein was diluted to the concentrations stated (5, 10, 20, and 40 nM 3) and injected over all flow cells at a flow rate of 50 µl min1 for 5 min. The resulting sensorgram from the flow cell containing GST was subtracted from those containing the GST1BI-II loop as a correction for bulk refractive index changes during 3 perfusion and for nonspecific binding of the 3 analyte to the GST moieties of the fusion protein. (c) Mean sensorgram for 20 nM 3 subunit. The data are themean ± SEM of six separate experiments. The single-exponential fits to the on- and off-rates of the mean sensorgram are shown in bold lines (on = 107 s, and off = 190.7 s). Assuming 1:1 binding, the KD was obtained from koff/kon, where koff is determined directly from off, and kon is calculated from (1/on) = kon × [] + koff. From the mean data, koff was 5.2 × 103 s1. However, this may be an underestimate as the koff derived from Biacore data may be contaminated by rebinding of the ligand. The calculated kon was 2.0 × 105 M1 s1.

Dependence of steady-state inactivation of 1B on the concentration of 3-subunit resulting from co-expression of increasing amounts of 3 cDNA

We next examined the effect of heterologous expression of the 3-subunit on the steady-state inactivation of 1B currents. For the two extreme conditions, no 3 cDNA (and therefore no heterologously expressed 3-subunit) and the maximal amount (720 pg) of 3 cDNA injected, the data could be fit by a single Boltzmann function (Fig. 4 a). The V_50,inact was hyperpolarized by the co-injection of 720 pg of 3 cDNA from 38.7 ± 1.0 mV to 67.6 ± 1.0 mV (Fig. 4 a). For intermediate concentrations of 3 cDNA (6-90 pg), both the individual (data not shown) and mean steady-state inactivation curves could be well fit only by a double Boltzmann function (Fig. 4 a). The two components of steady-state inactivation, A and B, had V_50,inact values of ~40 and 70 mV, respectively, which were relatively invariant (Fig. 4 b). However, the proportion of A decreased systematically, with a corresponding increase in B, as the amount of 3 cDNA, and therefore the concentration of expressed 3 protein, was increased (Fig. 4 c). On a log concentration plot, the percentage of A and B could be fit by reciprocal logistic functions with a midpoint of ~4 ng of 3 protein/oocyte (Fig. 4 c), corresponding to an average concentration of ~120 nM 3 protein.

View larger version (34K): [in this window] [in a new window]   FIGURE 4   Effect of 3 subunit co-expression on steady-state inactivation of 1B. Steady-state inactivation curves were obtained from 100-ms-duration test pulses to 0 mV, preceded by a 25-s prepulse to the conditioning potential given (100-0 mV). The fits are to a single or double Boltzmann function (see Materials and Methods). (a) Mean steady-state inactivation curves for 1B co-expressed with 0-720 pg of 3 cDNA. Peak IBa values from individual experiments were normalized and fit to either a single or a double Boltzmann function, as appropriate, judged by goodness-of-fit criteria. The symbols used are the same as in Fig. 2 b. (b) The mean values for V50,inact,a () and V50,inact,b () obtained for the two components of the individual steady-state inactivation curves, plotted against nanograms of 3 protein expressed, determined from Fig. 1 b, for each concentration of 3 cDNA injected (0-720 pg). (c) The mean percentage of the two components A (with V50,inact,a) and B (with V50,inact,b) from the individual normalized single or double Boltzmann fits are given for each concentration of 3 cDNA injected. The data are fit to logistic functions with midpoints at 4.25 and 4.43 ng of 3 protein and power coefficients (analogous to a Hill coefficient) of 1.28 and 1.20 for components A and B, respectively.

Dependence of the facilitation rate of G-protein-modulated 1B currents on 3 cDNA concentration

For the entire range of 3 cDNA injected, application of 100 nM quinpirole produced a significant inhibition of 1B I_Ba, being maximal between 64.4% and 73.1% (see traces in Fig. 5). This inhibition was voltage dependent, in that it could be overcome by a large depolarizing prepulse to +100 mV (voltage protocol in Fig. 5 a), a process termed facilitation. However, the duration of the prepulse required to overcome the inhibition was greater the lower the -subunit concentration. The facilitation rate was studied by increasing the duration of the prepulse in successive sweeps (Fig. 5, b-d). Injection of the highest amount of the VDCC 3 cDNA (720 pg) caused a very marked increase in the facilitation rate of the G-protein-modulated I_Ba during a +100-mV prepulse, compared with that for the 1B/2-1 currents in the absence of co-expressed 3-subunits (see example overlaid traces in Fig. 5 , b and d, and corresponding individual and mean facilitation rates in Fig. 5, e and f). At both these extremes, the individual and mean facilitation rates could be fit by a single exponential (Fig. 5, e and f). In the absence of co-expressed 3-subunits, the time constant for facilitation (_facil) was 94.7 ± 7.2 ms (n = 12; _slow). With 720 pg of 3 cDNA, _facil was 4.7 ± 0.4 ms (n = 7; _fast). We next examined whether the two extremes depicted in Fig. 5, b and d, represent facilitation of two separate populations, as suggested from the steady-state inactivation data. If this were the case, then injection of intermediate amounts of 3 cDNA would result in facilitation that could be fit by the sum of the two exponentials, represented by the extremes in Fig. 5, e and f. An example for an intermediate amount of 3 cDNA (45 pg) is given in Fig. 5 c (overlaid traces), and an example of the facilitation rate from an individual experiment is given in Fig. 5 e to show that it was well fit only by a double exponential. The double-exponential fits for each individual experiment at intermediate amounts of 3 cDNA resolved two components, A' and B', having time constants _slow and _fast. Their values, and their respective proportions, were determined from all the individual data whose mean facilitation curves are shown in Fig. 5 f. The percentage of A' and B' varied systematically and reciprocally with 3 cDNA injected, and therefore with concentration of 3 protein expressed (Fig. 5 g), in an almost identical manner to components A and B from the steady-state inactivation data (Fig. 4 c). The percentage of A' and B' could be fit with logistic functions, with half-maximal values at ~4 ng of 3 protein expressed/oocyte (Fig. 5 g).

View larger version (41K): [in this window] [in a new window]   FIGURE 5   Effect of 3 subunit concentration on facilitation rates during inhibition of 1B currents by the dopamine D2 agonist quinpirole. (a) Voltage waveform for measurement of facilitation rate in the presence of quinpirole. The duration of the prepulse (t) was increased in 5- or 10-ms steps as appropriate for the time course. (b-d) Family of traces for the 1B/2-1, 1B/2-1 + 45 pg of 3 cDNA, and 1B/2-1 + 720 pg of 3 cDNA subunit combinations. Tail currents have been clipped for better visualization of the currents. Calibrations in b refer to all traces. (e) Representative individual data for the facilitation rate are given for the intermediate concentration of 45 pg of 3 cDNA (), compared with 0 pg of 3 cDNA () or 720 pg of 3 cDNA (), as used in b-d above. The difference between the amplitude of the P1 and P2 currents (P2-P1) at each t was normalized to the plateau (P2P1) value and fit to a single- or double-exponential function. The 0- and 720-pg 3 cDNA facilitation rates are well fit by a single exponential (; slow = 106 ms, 2 = 0.00011 for the 0-pg 3 cDNA data; fast = 3.72 ms, 2 = 0.00009), whereas the 45-pg 3 cDNA data are not well fit by a single exponential (· · ·;  = 16.7 ms, 2 = 0.0187). In contrast, the points are well fit to a double exponential (; fast = 4.2 ms, 56%; slow = 35.5 ms, 44%; 2 = 0.00011). (f) For the individual data at all 3 cDNA concentrations, the facilitation rates were determined as in e. The corresponding mean facilitation rates are shown, with single-exponential fits to the two 3 cDNA concentrations at each extreme (0 and 3 pg of 3 cDNA for slow and 360 and 720 pg of 3 cDNA for fast) and a double-exponential fit (slow and fast) for all the intermediate 3 cDNA concentrations. (g) The mean percentage of the two components A' () and B' () (with slow and fast, respectively), for the single- or double-exponential functions determined for the individual data, are given for each level of 3 protein expressed. The data are fit to logistic functions, both with midpoints of 3.7 ng of 3 protein expressed/oocyte and power coefficients 1.8 and 1.5, for components A' and B', respectively. (h) The values for slow and fast obtained for the two exponential components (A' and B') of the individual facilitation curves, for each concentration of 3 cDNA injected. The slow data are fit with a logistic function with midpoint of 2.7 ng of 3 protein expressed/oocyte and power coefficient of 1.2. (Inset) Plot of 1/slow against estimated concentration of 3 protein expressed; slope (kslow) = 6.86 × 107 M1 s1.

Whereas the mean value of _fast was always similar to that observed with the maximal amount of injected 3 cDNA, being between 4.5 and 5.6 ms (Fig. 5 h, ), the value of _slow was not constant but systematically decreased as the amount of 3 cDNA injected was increased, from over 90 ms in the absence of co-injected 3 cDNA to 27.0 ± 2.5 ms (n = 10) at 90 pg of injected 3 cDNA (Fig. 5 h, ). The variation of _slow with the concentration of 3 protein expressed was fit by a logistic function, with a midpoint at 2.7 ng of 3 protein/oocyte. A plot of 1/_slow versus 3 protein concentration is linear (Fig. 5 h, inset; R = 0.9991), with a slope of 6.9 × 10⁷ M¹ s¹.

For the A/S ODN-injected oocytes, quinpirole-mediated voltage-dependent inhibition remained present, in those oocytes expressing I_Ba, with a very similar maximal degree of inhibition of I_Ba being observed for A/S and N/S ODN-injected oocytes (Fig. 6 a). The corresponding _facil was ~89 ms for currents in the N/S ODN-injected oocytes, almost identical to that obtained above for 1B/2-1 expressed in the absence of exogenous 3 cDNA (Fig. 6, b and inset). _facil remained a single exponential but was significantly slower for I_Ba recorded from the 3 A/S ODN-injected oocytes, being ~116 ms for both 4 and 40 µM A/S ODN (Fig. 6, b and inset).

View larger version (29K): [in this window] [in a new window]   FIGURE 6   The effect of the 3 antisense ODN on kinetics of facilitation of 1B currents in the presence of quinpirole. (a) Examples of maximal inhibition of 1B IBa (C) by 100 nM quinpirole (Q) for 40 µM N/S and 40 µM A/S ODN injected oocytes. The test potential is 5mV. (b) Mean normalized facilitation time course (as in Fig. 5 f) for the N/S ODN and the combined 4 and 40 µM A/S ODN conditions in the presence of quinpirole, using the protocol in Fig. 5 a, both of which were fit to a single exponential. (Inset) Mean facil (slow) for the N/S and A/S conditions (n = 6 and 7, respectively). *p < 0.05, compared with the N/S ODN.

The effect of 3-subunit on the expression and facilitation rate of the N-terminal 1B mutant I49A

We have previously shown that the N-terminal amino acid sequence 45-55 of 1B contains critical determinants for G modulation and that mutation of I49 to A causes partial loss of G protein modulation (Canti et al., 1999), as confirmed here (Fig. 7, a and b, insets). We have subsequently shown that this motif is involved in VDCC -subunit-regulated inactivation of 1B (Stephens et al., 2000). We have therefore compared the amount of expression and the facilitation rate of wild-type 1B with that of the I49A mutant of 1B, both in the absence and in the presence of co-injected 3 cDNA. In the presence of heterologously expressed 3-subunit, the expression of I49A mutant and wild-type 1B was identical, with the G_max values being 0.024 ± 0.003 (n = 13) and 0.022 ± 0.005 (n = 12) µS, respectively (Fig. 7 a). In contrast, in the absence of co-expressed 3-subunit, the expression of I49A 1B was more than double that of the wild-type 1B, with the G_max values being 0.033 ± 0.004 (n = 20) and 0.015 ± 0.002 (n = 20) µS, respectively (p < 0.01; Fig. 7 b). This suggests the possibility that the I49A 1B has an increased affinity for the endogenous 3 responsible for trafficking it to the membrane. In agreement with this hypothesis, the facilitation rate for I49A 1B in the absence of co-expressed 3-subunit could be fit by a single exponential (Fig. 7 c), giving a _facil of 28.3 ± 1.7 ms (n = 10), which is nearly threefold faster than for wild-type 1B (84.3 ± 3.1 ms; n = 9; p < 0.0001; Fig. 7 c). For 1B I49A in the presence of a maximal concentration of 3-subunit protein (720 pg of 3 cDNA injected), the facilitation rate for this mutant was slightly faster (4.7 ± 0.5 ms; n = 8), than for wild-type 1B (7.4 ± 0.5 ms; n = 5; p < 0.01; Fig. 7 c). This small increase in the presence of 3 may represent an enhancement of the off-rate for G in the presence of 3, as suggested previously (Canti et al., 1999).

View larger version (25K): [in this window] [in a new window]   FIGURE 7   The effect of the I49A mutation of 1B on the expression and kinetics of facilitation of 1B currents in the absence and presence of co-expressed 3. (a) Mean IV relationships for 1B () and I49A 1B () IBa in the presence of a maximal concentration of co-expressed 3 (720 pg of 3 cDNA). (Inset) histogram of percent inhibition by 100 nM quinpirole at 0 mV, for the two conditions, with the n values given on the bars. (b) Mean IV relationships for 1B () and I49A 1B () IBa in the absence of co-expressed 3. (Inset) Histogram of percent inhibition by 100 nM quinpirole at 0 mV, for the two conditions, with the n values given on the bars. (c) Mean normalized facilitation time course (as in Fig. 5 f) for IBa in the presence of quinpirole, using the protocol in Fig. 5 a for the I49A (, n = 13; , n = 20) and wild-type 1B (, n = 12; , n = 20), in the presence ( and ) or absence ( and ) of co-expressed 3 cDNA. Ba2+ (10 mM) was used as the charge carrier in these experiments. All time courses were fit to a single exponential, and mean data are given in the text.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of 3-subunit concentration on G_max of 1B currents

The endogenous oocyte concentration of 3 protein is here estimated to be ~17 nM, and the 1B G_max was strongly dependent on the concentration of heterologously expressed 3, with a half-maximal value at approximately the endogenous level of 3 protein. This correlates well with the affinity of the I-II linker of 1B for the 3-subunit, determined from Biacore experiments to be ~20 nM. This region has recently been suggested to harbor an endoplasmic reticulum retention signal, masked by a -subunit (Bichet et al., 2000). We may therefore express this interaction as:

(1)

Furthermore, the 3 A/S ODN reduced the number of cells expressing I_Ba and, in those oocytes that did express currents, reduced the mean G_max by ~50%. This is in close correlation with the estimated 47% reduction of endogenous 3 protein by the A/S ODN found in those cells expressing I_Ba. It also agrees with previous results on 1E expression, where it was suggested that the endogenous Xenopus oocyte 3-subunit plays an essential trafficking role (Tareilus et al., 1997). It was unclear from that study whether the endogenous chaperoning 3-subunits remain associated with each expressed 1-subunit in the plasma membrane in a high-affinity complex, because they do not affect the biophysical properties of the channel (Tareilus et al., 1997). Thus, the endogenous level of 3 protein appears to be insufficient to produce the effects of -subunits on biophysical properties that are observed on heterologous expression of -subunits. This is reinforced by the finding in the present study that there was no effect of partial depletion of the endogenous -subunit by the A/S ODN on the V_50,act of the residual I_Ba. In apparent contradiction to this, if the affinity for 3 remains at 17 nM, because the endogenous concentration of 3-subunit is estimated to be ~17 nM, when 1B is expressed alone it should be at least 50% bound in the membrane to a single endogenous 3-subunit via the I-II linker in the control or N/S ODN condition (to form 1B · 3; Eq. 1).

We can put forward two alternative hypotheses to account for the present results. Either a second -subunit is bound, which accounts for the biophysical effects of -subunits on the voltage dependence of activation and inactivation and on the facilitation rate, or there is a reduction in affinity of the mature 1B-subunit in the polarized plasma membrane for the 3-subunit responsible for its trafficking, so that the two species dissociate, and mature 1B-subunits are largely non-complexed to -subunits at the endogenous level of Xenopus oocyte -subunits. In agreement with either hypothesis, injection of 3 protein into Xenopus oocytes expressing 1C-subunits alone had acute effects on their biophysical properties (Yamaguchi et al., 1998), although the concentration dependence was not studied.

Effect of -subunit concentration on voltage dependence of activation and steady-state inactivation

The effect of increasing the concentration of the 3-subunit on the V_50,act revealed that the concentration dependence was right-shifted, compared with that for the G_max. However, we did not attempt in this study to resolve these data into two populations, as was done for the V_50,inact. The effect of increasing the concentration of the 3-subunit on the V_50,inact fits the hypothesis that there are two populations of 1B channel, A and B, with independent behaviors, corresponding to 1B associated or not with a -subunit, bound with the K_D of ~120 nM for 3. Both of the hypotheses outlined above are compatible with this result: either a second -subunit is bound or the affinity of the 3-subunit for the mature 1B channel is markedly reduced, compared with its affinity for the nascent channel. Thus, population A would represent either 1B bound to a single -subunit or free 1B in the plasma membrane (and for both cases, it is denoted 1B*). Population B would be 1B* · 3:

(2)

The equilibrium relationship between the populations A and B implies the presence of a pool of free -subunits (Eq. 2), a notion that is supported by the fractionation experiment. Injection of 30 pg of 3 cDNA (producing ~120 nM 3 protein) results in an ~7-fold increase in 3 protein expression compared with the endogenous level and a proportionately greater increase in cytoplasmic 3. How