Voltage-gated calcium channels (VGCCs), also referred to as voltage-dependent calcium channels (VDCCs), represent a category of voltage-gated ion channels characterized by their permeability to calcium ions (Ca²⁺). ^[1] These channels are multimeric proteins composed of five co-assembled subunits, specifically (α1, α2δ, β1-4, and γ). VGCCs can be broadly categorized into high-voltage-activated channels, which are further subdivided into L-, R-, P/Q-, and N-types, and low-voltage-activated channels (T-type). ^[2] The α subunits are large, single-chain polypeptides consisting of approximately 2000 amino acid residues, organized into four homologous domains. Each domain contains six transmembrane segments (S1-S6) and a pore-forming loop. The positively charged fourth transmembrane segment (S4) functions as a voltage sensor and plays a critical role in channel gating. Recent studies on the crystal structure of the bacterial NavAb channel have unveiled several novel structural features when compared to earlier potassium channel structures, including a short selectivity filter, with ion selectivity determined by interactions with glutamate side chains. ^[3]

The pore-forming α1 subunits of Cav channels exhibit nearly identical topologies, comprising approximately 2000 residues organized into four homologous repeats, designated I-IV. Each repeat contains six transmembrane helices (S1-S6). The S5 and S6 segments from all four repeats collectively form the central pore domain (PD), creating a specific pathway for the selective passage of Ca ions across the cell membrane. Concurrently, the S1-S4 segments within each repeat constitute the voltage-sensing domain (VSD), wherein the S4 segment is characterized by repetitive positively charged residues, such as arginine or lysine, which serve as gating charges. The four sets of VSDs encircle the PD in a domain-swapped configuration, thereby cooperatively linking fluctuations in membrane potential to the gating of ion conduction through the pore. ^{[4, 5, 6]} The core α1 subunit is sufficient for the autonomous function of Cav3 channels; however, the Cav1 and Cav2 subfamilies require auxiliary extracellular α2δ and intracellular β subunits to ensure proper membrane trafficking and physiological functionality. ^{[7, 8, 9]} Notably, Cav1.1 channels, which are specialized for skeletal muscle, also associate with a transmembrane γ subunit that shares a similar folding pattern with claudins. ^[10] Within the Cav1 and Cav2 subfamilies, the α1 subunits interact with the α2δ subunits through their extracellular segments, while engaging with the β subunits on the cytosolic side. ^[11] Four mammalian α2δ genes, specifically CACNA2D1 to D4, encode the extracellular subunits α2δ-1 to -4, respectively. ^[12] These gene products initially exist as preproteins for the α2δ subunits, which subsequently undergo post-translational proteolysis to yield the mature α2 and δ proteins. Importantly, these two proteins remain interconnected through inter-subunit disulfide bond formation prior to proteolytic cleavage. ^[13] The mature α2δ subunits are highly glycosylated proteins that facilitate the trafficking of α1 subunits to the cell membrane, thereby enhancing channel expression and modulating channel properties. Additionally, four subtypes of β subunits (β1-β4), which possess multiple splice isoforms, contribute to channel trafficking, modulate gating properties, and interact with intracellular signaling molecules. The specific composition of subunit constituents varies according to the particular type of Cav channel and the specific tissue or cell type in which it is expressed. ^[14]

the N1 amine is hydrogen-bonded (H-bonded) to the hydroxyl group of Ser1011 on P1III, the two oxygen groups of the C3-ester are each H-bonded to Thr935 and Gln939 (Black sphere in the 3D structure viewer) on S5III, the nitrophenyl ring is situated within a hydrophobic pocket formed by Val932 on S5III as well as Met1057 and Phe1060 on S6III, the DHP backbone and the methyl groups are surrounded by Phe1008 on P1III as well as Tyr1365 and Met1366 on S6IV. ^[15] In addition to polar interactions, the nitrophenyl ring is accommodated in a hydrophobic pocket formed by Val932, on S5III, and Met1057 and Phe1060, on S6III, while the carbon backbone of the dihydropyridine ring and the methyl groups are surrounded by Phe1008, on P1III, and Tyr1365 and Met1366, on S6IV. Despite the lack of direct interaction between Tyr1048 and nifedipine, single point mutations corresponding to Y1048F and Y1048A resulted in reduced affinity with DHP by ∼10- and ∼1,000-fold, respectively. While the aromatic ring of Tyr1048 is surrounded by hydrophobic residues on S5III, P1III, and S6III, its hydroxyl group is H-bonded to Gln939. The requirement for Tyr1048 on DHP antagonist binding may be achieved indirectly through stabilization of Gln939 and local structures. There is extra space in the fenestration, close to the C3 and C5 ester groups of the dihydropyridine ring. Such unoccupied space may provide accommodation for additional modifications on the two groups, thereby providing a potential explanation for the compatibility of L-type Cav channels with multiple DHP compounds. ^[16]

The molecular recognition of the gabapentinoid drugs, gabapentin and mirogabalin, by the α2δ-1 subunit have been investigated, but the structural observations are somewhat ambiguous. Gabapentinoid drugs were supposed to occupy a pocket identical to the L-leucine binding site at the Cache1 domain. No discernible conformational change was observed in the α2δ-1 subunit upon binding to gabapentin or mirogabalin. However, our group ever refrained from conclusively assigning densities in the same site to gabapentinoid (pregabalin) due to its similar size and shape to the endogenous L-leucine ligand. ^[17]

Diltiazem is coordinated mainly by a large number of hydrophobic residues on S6I, P1III, S6III, and S6IV in Cav1.1. Among these residues, replacement of three amino acids, Tyr1365, Ala1369, and Ile1372, (Red sphere in the 3D structure viewer) on S6IV with non-L-type Cav channel residues was shown to significantly decrease the sensitivity to diltiazem. Sequence alignment shows that these three residues are highly conserved among L-type, but not in the other, Cav channels. These residues may, thus, collectively define the specificity for diltiazem by the L-type Cav channels. ^[18]

Mutational analysis identified Tyr1365, Ala1369, and Ile1372 on S6IV to be essential for verapamil binding. ^[19]

Ziconotide, derived from ω-conotoxin MVIIA, acts as a selective blocker for Cav2.2 channels and received FDA approval in 2004 for the treatment of severe chronic pain. ^[20] Structural studies have unveiled the molecular underpinnings of the specific pore blockage of ziconotide on human Cav2.2. ^[21] Ziconotide is situated in the electronegative pocket surrounding the entrance to SF. To accommodate ziconotide, ECLIII must move upward together with the α2δ-1 subunit. The positively charged residues in ziconotide effectively neutralize the negatively charged pocket that attracts ions into the ion entrance, creating a spatial obstruction to prevent Ca2+ influx into the pore. Further sequence comparison revealed that half of the crucial residues required for ziconotide coordination were not conserved among Cav channels, thus elucidating the molecular basis for the selective pore blockade of ziconotide. ^[22]

Calciseptine, a member of the three-finger toxin isolated from the venom of black mamba, consists of 60 amino acids and features four pairs of disulfide bonds. With the ability to selectively block Cav1.2 and Cav1.3 channels, calciseptine effectively inhibits the contraction of smooth and cardiac muscles. In contrast to other VSD or PD-binding toxins, calciseptine is positioned on the PD shoulder and thoroughly interacts with the P2 pore helices and the ECLs in repeats III and IV. ^[23]

Phenylalkylamines are intracellular pore blockers, which are thought to enter the pore from the cytoplasmic side of the channel and block it. Their receptor site is formed by amino acid residues in the S6 segments in domains III and IV, in close analogy to the local anaesthetic receptor site on sodium channels. ^[24]

DHP agonists occupy the same binding site as antagonists. ^[16] The -NO₂ group of (R)-(+)-Bay K 8644 can be coordinated by the polar residues on S5III, whereas the C3 ester group, similarly to that of nifedipine, is accommodated in a hydrophobic pocket, allowing for stable association with the inactivated channel (Figure S7). The structure, thereby, provides an important clue to the observation that the presence of an electron withdrawing group, at the C5 position of the pyridine ring, is associated with agonistic activity.

Dihydropyridines can be channel activators or inhibitors, and therefore are thought to act allosterically to shift the channel toward the open or closed state, rather than by occluding the pore. Their receptor site includes amino acid residues in the S6 segments of domains III and IV and the S5 segment of domain III. The dihydropyridine receptor site is closely apposed to the phenylalkylamine receptor site and shares some common amino acid residues. ^[25]

Z944 exhibits characteristics of both a pore blocker and an allosteric antagonist ^[26]. A recent study has further unveiled analogous binding modes observed in Cav3.3 channels complexed with mibefradil, pimozide, and otilonium bromide.^[27] PD173212 was observed to block Cav2.2 channels through a dual mechanism involving both pore blockage and allosteric modulation. ^[28]