Alternative splicing of a beta4 subunit proline-rich motif regulates voltage-dependent gating and toxin block of Cav2.1 Ca2+ channels

Abstract

Ca2+ channel beta subunits modify alpha1 subunit gating properties through direct interactions with intracellular linker domains. In a previous report (Helton and Horne, 2002), we showed that alternative splicing of the beta4 subunit had alpha1 subunit subtype-specific effects on Ca2+ channel activation and fast inactivation. We extend these findings in the present report to include effects on slow inactivation and block by the peptide toxin omega-conotoxin (CTx)-MVIIC. N-terminal deletion and site-directed mutagenesis experiments revealed that the effects of alternative splicing on toxin block and all aspects of gating could be attributed to a proline-rich motif found within N-terminal beta4b amino acids 10-20. Interestingly, this motif is conserved within the third postsynaptic density-95 (PSD-95)/Discs large/zona occludens-1 domain of the distantly related membrane-associated guanylate kinase homolog, PSD-95. Sequence identity of approximately 30% made possible the building of beta4a and beta4b three-dimensional structural models using PSD-95 as the target sequence. The models (1) reveal that alternative splicing of the beta4 N terminus results in dramatic differences in surface charge distribution and (2) localize the proline-rich motif of beta4b to an extended arm structure that flanks what would be the equivalent of a highly modified PSD-95 carboxylate binding loop. Northern blot analysis revealed a markedly different pattern of distribution for beta4a versus beta4b in the human CNS. Whereas beta4a is distributed throughout evolutionarily older regions of the CNS, beta4b is concentrated heavily in the forebrain. These results raise interesting questions about the functional role that alternative splicing of the beta4 subunit has played in the evolution of complex neural networks.

Fig. 1.

Effects of β_4a and β_4bon slow inactivation and recovery from slow inactivation of Ca_v2.1 Ca²⁺ channels. Studies were performed with Xenopus oocytes expressing α_1A, α₂/δ-1, and either β_4a or β_4b. Reference (I_R) and test current (I_T) traces were generated by 300 msec step depolarizations from various holding potentials to either 0 mV (β_4b) or +10 mV (β_4a). Maximum values from 300 msec I_R andI_T current traces were used to calculateI_T/I_Rwhere indicated. Barium (5 mm) was used as the charge carrier. A, Influence of β_4a and β_4b on the development of slow inactivation at a conditioning potential (CP) of −40 mV. After a reference pulse (I_R) measured from a holding potential of −80 mV, oocytes were held at −40 mV for 5 min. During this time, 300 msec test pulses (I_T) were applied every 15 sec. Eachpoint represents the mean value ofI_T/I_R from 11 (β_4a) or 10 (β_4b) different recordings. The SEM is shown for each pointunless the values were smaller than the symbol. Thesolid lines represent double-exponential fits to the data. B, Representative reference (I_R) and 5 min (I₅) current traces from Ca_v2.1 complexes containing either β_4a(top) or β_4b (bottom) generated as described in A. C, Voltage dependence of slow inactivation. The ratio ofI₅ to I_R, generated as in A, plotted as a function of conditioning potential for Ca_v2.1 complexes containing either β_4a or β_4b. Data pointsrepresent the means of at least six determinations at a given membrane potential. Lines serve only to connect the data points. D, Influence of β_4a and β_4b on the time course of recovery from slow inactivation. After a 300 msec reference pulse (I_R) measured from a holding potential of either −80 or −100 mV, oocytes were held at −30 mV for 5 min. The membrane potential was then returned to either −80 or −100 mV, and sequential test pulses (I_T) were applied at 15 sec intervals for a total of 3 min. Eachpoint represents the mean of at least seven different recordings. Solid lines represent the single-exponential fits of the data.

Fig. 2.

Effects of β_4a and β_4bon the blockade of Ca_v2.1 channels by ω-CTx-MVIIC. Studies were performed with Xenopus oocytes expressing α_1A, α₂/δ-1, and either β_4a or β_4b.A, Onset and degree of block by a 10 min exposure to 2 μmω-CTx-MVIIC for Ca_v2.1 subunit combinations at a holding potential (HP) of −80 mV. Each pointrepresents the mean of seven (β_4a) or eight (β_4b) different recordings. The SEM is shown for each data point unless smaller than thesymbol. Onset of block for both subunit combinations was fit (line) to a single-exponential time course plus a constant. B, The rate constants for the time course of the onset of toxin block were determined from steady-state degree of block from single exponential fits at four different toxin concentrations (0.2, 0.6, 2, and 6 μm) for Ca_v2.1 complexes containing either β_4a or β_4b. The averaged rate constants were plotted as a function of toxin concentration (minimum of n = 7, ±SEM). The line represents a linear fit to the data.

Fig. 3.

Localization of differential effects on Ca_v2.1 gating and pharmacology to β_4bN-terminal amino acids 10–20. The first 10 (β_4bΔ1–10), first 20 (β_4b Δ1–20), or second 10 (β_4b Δ10–20) aa of the N terminus of the β_4b subunit were removed using PCR. The deletion mutants as well as β_4a or β_4b were expressed with α_1A and α₂δ-1 in Xenopusoocytes. A, Effects of the N-terminal deletion mutants on the voltage dependency of activation of Ca_v2.1 channels. Plots were derived from averaged I–V data up to +10 mV for each β₄ subunit combination. Data pointsrepresent the means of the normalized data at a given membrane potential for a minimum of nine different recordings. Smooth lines represent single Boltzmann fits to the averaged data.B, Normalized, averaged isochronal inactivation curves for Ca_v2.1 complexes containing the various β₄ subunits. Points represent the means of the normalized data at a given membrane potential for a minimum of nine different recordings. Smooth lines represent single Boltzmann fits to the averaged data. C, Effects of β₄ N-terminal deletion mutants on the development of slow inactivation at a conditioning potential (CP) of −40 mV. Reference (I_R) and test (I_T) currents were generated as in Figure 1A. Each point represents the mean value ofI_T/I_R from 13 (β_4b Δ1–10) or nine (β_4b Δ1–20) different recordings. The solid lines represent double-exponential fits to the data. D, Onset and degree of block by a 10 min exposure to 2 μm ω-CTx-MVIIC for Ca_v2.1 complexes containing β_4b Δ1–10 or β_4b Δ1–20. Data were generated as in Figure2A. Each point represents the average of a minimum of seven recordings. The solid lines represent single-exponential fits to the data.

Fig. 4.

The β₄ subunit is a distant homolog of PSD-95. Identification of a conserved GXXDXPXXP motif critical to Ca_v2.1 gating. A, Amino acid alignment of the A domains of the human spinal cord β_4a (amino acids 1–63) and β_4b (amino acids 1–97) subunits with the third PDZ domain (amino acids 294–391) of PSD-95. Vertical bars denote identical amino acids between β_4b and PSD-95. Important amino acids involved in modulating the leftward shift in the voltage dependence of activation and inactivation of β_4b (GXXDXPXXP) are highlighted.Arrows and hatched bars represent predicted α-helices and β-strands of the third PDZ domain of PSD-95, respectively. B, Differences in theV_1/2 values of activation and inactivation of β_4a, β_4b, and β_4b N-terminal amino acid mutants. Solid bars represent average V_1/2 values of a minimum of nine different recordings for each β₄subunit variant. Positive or negative shifts in theV_1/2 values (in millivolts) of activation and inactivation of β_4a and β_4b mutants are relative to the V_1/2 values of activation and inactivation of β_4b. Currents were generated as described in Figure 3, A and B. The SEM for each bar is shown. Asterisks denote statistical significance (p < 0.05) by a Student's two-sample equal variance t test.

Fig. 5.

Real-space optimization structural models of the A domains of β_4a (A) and β_4b (B) based on sequence identities with the third PDZ domain of PSD-95. Ribbon (left) and electrostatic surface potential (right) diagrams were created using MOLMOL (Koradi et al., 1996). For ribbon diagrams, arrows indicate β-strands, and helices indicate α-helices. For surface potential diagrams, red, white, and blue regions indicate negatively charged, hydrophobic, and positively charged amino acids, respectively.

Fig. 6.

Differential distribution of β_4a and β_4b mRNA in the human CNS. A, Electrophoresis of full-length β_4a (left) and β_4b (right) cRNAs (includes 5′ and 3′ untranslated) in a 1% agarose formaldehyde denaturing gel. RNA markers (in kilobases) are indicated on the left.B, Northern blot analysis performed with human multiple tissue blot (Human Brain II; Clontech) and a ³²P-labeled β₄ subunit probe (coding nucleotides 215–1628 plus ∼300 bp 3′ untranslated sequence). Molecular masses on theright correspond to labeled blot markers.C, A human β₄ subunit genome map depicting the lengths of intron sequences (b, bases) between alternatively spliced β_4a and β_4b N-terminal exons and the beginning of exon 2. Solid lines represent exons, and dashed lines represent introns.Numbers in parentheses below solid lines indicate position on chromosome 2. Boxesindicate protein sequence (β_4a inparentheses). The first and last two amino acids of each sequence are indicated above each box.