Molecular Interactions in the Voltage Sensor Controlling Gating Properties of CaV Calcium Channels

Abstract

Voltage-gated calcium channels (CaV) regulate numerous vital functions in nerve and muscle cells. To fulfill their diverse functions, the multiple members of the CaV channel family are activated over a wide range of voltages. Voltage sensing in potassium and sodium channels involves the sequential transition of positively charged amino acids across a ring of residues comprising the charge transfer center. In CaV channels, the precise molecular mechanism underlying voltage sensing remains elusive. Here we combined Rosetta structural modeling with site-directed mutagenesis to identify the molecular mechanism responsible for the specific gating properties of two CaV1.1 splice variants. Our data reveal previously unnoticed interactions of S4 arginines with an aspartate (D1196) outside the charge transfer center of the fourth voltage-sensing domain that are regulated by alternative splicing of the S3-S4 linker. These interactions facilitate the final transition into the activated state and critically determine the voltage sensitivity and current amplitude of these CaV channels.

Keywords: Ca(V); calcium channel; voltage gating; voltage-sensing domain.

Figure 1. Structural Modeling of the IVS3-S4 Loop in Different Conformational States Depicts a Higher Variability of Ca_V1.1a Loop Compared with Ca_V1.1e

(A) Transmembrane topology of the α₁ subunit of the voltage-gated calcium channels consisting of four repeats with six transmembrane domains each. The red mark indicates the position of exon 29 in the extracellular loop connecting IVS3 and IVS4. The sequence alignment of IVS3, the IVS3-S4 extracellular loop, and IVS4 of the two Ca_V1.1 splice variants shows the position of the 19 amino acids encoded by exon 29 in the IVS3-S4 linker, the positions of the positive charges in IVS4 (blue) and potential negative counter charges in IVS3 (red). Note that the colors of the bar above the sequences match the ribbon models below (green, S3; gray, exon 29; orange, S4). (B) Predicted conformations of both Ca_V1.1a and Ca_V1.1e IVS3-S4 loops in the resting state, intermediate states 1 and 2, and in the activated state. Gray highlights the extra 19 amino acids encoded by exon 29 present in the Ca_V1.1a loop, while the short IVS3-S4 linker of Ca_V1.1e is depicted in yellow. Note the substantial conformational changes of the Ca_V1.1a IVS3-S4 loop during gating and between different clusters, but less for the Ca_V1.1e IVS3-S4 loop. The images were generated using UCSF Chimera (Pettersen et al., 2004).

Figure 2. Homology Rosetta Modeling of the Ca_V1.1 Splice Variants Fourth Repeat VSD Gating States

(A) Superimposed structures of Ca_V1.1a and Ca_V1.1e VSD at the resting state, intermediate states 1 and 2, and the activated state. S1–S3 are shown as ribbons while the S4 transmembrane domain is shown as licorice to facilitate observation of similarities and differences in the structures of Ca_V1.1a and Ca_V1.1e. The phenylalanine “cap” is shown to indicate the position of the CTC. In the fourth VSD of Ca_V1.1, the CTC comprises a phenylalanine (F1161 = F) and a glutamate (E1164 = E2) in IVS2, and an aspartate (D1186 = D3) in IVS3. S4 positively charged arginines are highlighted in blue, while the negatively charged amino acids that could participate to the charge transfer are highlighted in red. The extra 19 amino acids encoded by exon 29 in Ca_V1.1a are shown in gray. (B) Distance measurements between the C^α atoms of IVS4-charged amino acids and the C^α atoms of F and E2 in IVS2, D3, and D4 (D1196) in IVS3. Only S2, S3, and S4 of Ca_V1.1e are shown for clarity. Black dotted lines indicate that the distances are the same in Ca_V1.1a and Ca_V1.1e in a given state, green dotted lines indicate that the distance is larger in Ca_V1.1e compared with Ca_V1.1a, while red dotted lines indicate that the distance is larger in Ca_V1.1a compared with Ca_V1.1e by at least 0.2 Å. Distances in the CTC differ in the intermediate state 2 and the activated state, whereas S3-S4 distances near the IVS3-S4 linker (as indicated by D4 to the closest arginine in S4 substantially differ in all states. Distances are measured using UCSF Chimera (Pettersen et al., 2004) and the values are presented in Table 1.

Figure 3. Structural Differences in the Fourth VSD of Ca_V1.1e and Ca_V1.1a Affect the Number and Strengths of H-bonds Sequentially Formed During Gating

(A) In the resting state, R1 and R2 of Ca_V1.1e form several H-bonds with D3 with distances ≤2 Å, while Ca_V1.1a R1 and R2 interaction with D3 results in less and weaker H-bonds. An additional arginine on the extracellular side of IVS4 (R0) forms one H-bond with an IVS3 aspartate (D4) in Ca_V1.1e but not in Ca_V1.1a. (B) H-bonds in intermediate state 1 are similar in both splice variants. (C) In intermediate state 2, R1 of Ca_V1.1e interacts with D4, and R2 with E2 and T, while in Ca_V1.1a, both R1 and R2 interact exclusively with E2 and T. (D) The activated state of Ca_V1.1e is characterized by multiple interactions of R1 and R2 with D4, while R0, R1, and R2 of Ca_V1.1a form interactions with different H-bond acceptors from IVS3 (D4, D4, S2). The H-bonds and their distances are illustrated using UCSF Chimera (Pettersen et al., 2004). Relevant H-bonds <2.0 Å are colored in red while distances >2.0 Å are shown in gray.

Figure 4. Aspartate at Position 1196 Is Critical in Determining the Voltage Sensitivity of Ca_V1.1 but Not Ca_V1.1a

(A–C) Representative calcium currents recorded from myotubes expressing Ca_V1.1a (blue), Ca_V1.1e (red), Ca_V1.1a-D4N (green), and Ca_V1.1e-D4N (wine) during a 200-ms step depolarization to the maximum current amplitude (A). The Ca_V1.1e splice variant has a ∼6-fold higher current amplitude and ∼26 mV left shift in voltage dependence compared with Ca_V1.1a. Mutating the negatively charged aspartate (D4) at position 1196 to the neutral asparagine (Ca_V1.1e-D4N, wine) reverts Ca_V1.1e calcium current properties to those of Ca_V1.1a, as illustrated by (B) the I/V curve and (C) voltage dependence of current activation. Mutating the same amino acid in Ca_V1.1a (Ca_V1.1a-D4N, green) has no effect on voltage sensitivity but slightly reduces the current amplitude. (D–F) Mutating the Ca_V1.1e aspartate (D4) to a glutamate results in the same reduction in amplitude (Ca_V1.1e-D4E, purple) and shift in the voltage dependence as observed for the aspartate to asparagine mutation. All data are presented as means ± SEM. Currents were analyzed as previously described (Tuluc et al., 2007) and the calcium current parameters and statistics are given in Table 2.

Figure 5. R1 and R2 Are Functionally Relevant Interaction Partners for D4 in Ca_V1.1e but Not in Ca_V1.1a

(A) Representative calcium currents recorded during a 200-ms step depolarization to the maximum current amplitude from myotubes expressing Ca_V1.1a (blue), Ca_V1.1e (red), or Ca_V1.1e with arginine-to-alanine mutations of R0 (Ca_V1.1e-R0A, gray), R1 (Ca_V1.1e-R1A, black), and R2 (Ca_V1.1e-R2A, orange). (B and C) Neutralizing R2 of Ca_V1.1e (Ca_V1.1e-R2A) converts the amplitude and voltage-dependence of Ca_V1.1e to that of Ca_V1.1a. R1 neutralization in Ca_V1.1e (Ca_V1.1e-R1A) elicits an even further right-shifted voltage dependence compared with Ca_V1.1a. In contrast, R0 neutralization of Ca_V1.1e (Ca_V1.1e-R0A) shifts the voltage dependence of Ca_V1.1e by only ∼7 mV to the right and slightly increases the current amplitude. (D–F) Neutralizing R1 and R2 charges in Ca_V1.1a (Ca_V1.1a-R1A, black and Ca_V1.1a-R2A, orange) does not affect the voltage dependence, but results in reduced calcium current amplitude. All data are presented as means ± SEM. The recording conditions and analysis were identical to those described in the legend of Figure 4. The calcium current parameters and statistics are given in Table 2.