Inactivation of Ca_V1 and Ca_V2 channels

Voltage-gated calcium (Ca²⁺) channels (VGCCs) are crucial conduits for Ca²⁺ entry in both excitable and select non-excitable cells (Nowycky et al., 1985; Nilius et al., 1986; Pitt et al., 2021). They are integral to various physiological processes, including gene expression, hormone secretion, neurotransmitter release, synaptic function, and muscle contraction (Bean, 1989; Bers, 1991; Wu et al., 1999; Ma et al., 2011). Given the critical roles of VGCCs in both normal and pathological cellular functions, precise regulation and fine-tuning of their properties are essential (Lee et al., 1985). To this end, these channels employ two distinct forms of negative feedback regulation—voltage-dependent inactivation (VDI) and Ca²⁺-dependent inactivation (CDI). These regulatory processes are critical for the physiological function of the channel, enabling channels to modulate Ca²⁺ entry due to sustained activity, thus preventing the accumulation of excessive intracellular Ca²⁺ and contributing to dynamic physiological processes.

VGCCs are composed of a main α₁ subunit and auxiliary subunits that can modulate channel properties, including β, α_2δ, and γ. The α₁ subunit has four homologous domains, each containing six transmembrane α helices, designated S1–S6. The S1–S4 segments compose the voltage-sensing domain, with the S4 containing positive gating charges, which enable the channel to respond to changes in membrane voltage (Flucher, 2016). The S5 and S6 helices form the pore of the channel while the S5–S6 linker forms a loop containing the selectivity filter (Sather et al., 1994; Stephens et al., 2015). Lastly, S6 helices form the activation gate intracellularly.

There are three families of VGCCs based on their α₁ subunits. These include Ca_V1 (Ca_V1.1–1.4), Ca_V2 (Ca_V2.1–2.3), and Ca_V3 (Ca_V3.1–3.3), also known as L- (Ca_V1), P/Q- (Ca_V2.1), N- (Ca_V2.2), R-type (Ca_V2.3), and T-type (Ca_V3) (Catterall et al., 2005). Ca_V1 and Ca_V2 channel families are often referred to as high voltage-activated channels; however, the term may be somewhat of a misnomer for Ca_V1.3 channels, which activate at more hyperpolarized voltages as compared with other L-type channels (Avery and Johnston, 1996; Xu and Lipscombe, 2001). Ca_V3 channels are part of the low voltage-activated Ca²⁺ channel subfamily and are activated at more hyperpolarized membrane potentials, have more rapid inactivation kinetics, and smaller unitary conductance as compared with Ca_V1 and Ca_V2 channels (Nowycky et al., 1985; Nilius et al., 1986; Zamponi et al., 2015; Nanou and Catterall, 2018). The inactivation of Ca_V3 channels appears to differ from that of Ca_V1 and Ca_V2 channel families in terms of structural elements, modulation by auxiliary subunits, and presence of CDI. We therefore focus this review on the Ca_V1 and Ca_V2 channel families.

Ca_V1.2 and Ca_V1.3 channels are widely expressed across excitable cells and are responsible for critical cellular functions such as excitation–contraction coupling, excitation-transcription coupling, synaptic regulation, and hormone release (Bers, 2002; Helton et al., 2005; Striessnig et al., 2006; Zhang et al., 2006). Conversely, Ca_V1.1 and Ca_V1.4 are more selectively expressed in skeletal muscle and retinal cells, respectively (Catterall et al., 2005). Ca_V2.1 channels are highly expressed in presynaptic terminals throughout the brain and are enriched in cerebellar Purkinje cells, where they play a critical role in balance and movement (Indriati et al., 2013). Ca_V2.2 channels are expressed throughout the nervous system and contribute to synaptic maturation, gene expression, neuronal migration and survival, and fast neurotransmitter release (Komuro and Rakic, 1992; Simms and Zamponi, 2014; Gorman et al., 2019). Ca_V2.3 channels are highly expressed in the central and peripheral nervous systems (Wormuth et al., 2016) and have been identified in the endocrine and cardiovascular systems (Lu et al., 2004). Channel regulation constitutes a major factor enabling the modulation of channel gating relevant to each physiological context. Here, we review the mechanisms and physiological implications of CDI and VDI in Ca_V1 and Ca_V2 channels.

20 years after Hodgkin and Huxley postulated the existence of gating change (Hodgkin et al., 1952), two seminal papers provided experimental evidence for the movement of a voltage sensor within a membrane protein (Armstrong and Bezanilla, 1973; Schneider and Chandler, 1973). Schneider and Chandler were the first to measure the charge movement of Ca_V1.1 channels associated with excitation–contraction in skeletal muscle (Schneider and Chandler, 1973), a pivotal component in understanding how changes in membrane voltage led to conformational changes in the channel. This voltage-sensor movement was critical not only for channel activation but for VDI (Armstrong and Bezanilla, 1977), a critical feedback mechanism that can control the precise amount of Ca²⁺ influx through VGCCs (see Catacuzzeno et al. [2023] for a comprehensive review of the role of gating change in channel inactivation). Disruption of this process can lead to various pathological conditions, including cardiac arrhythmia, epilepsy, and chronic pain (Splawski et al., 2004; Gambeta et al., 2021; Herold et al., 2023; Hussey et al., 2023a).

Multiple mechanisms for VDI have been considered. The first involves a cytoplasmic segment of the channel that binds to and blocks the channel pore (Armstrong and Bezanilla, 1977; Bezanilla and Armstrong, 1977). This type of mechanism has been described as a “ball and chain” mechanism for the N-type inactivation of the K⁺ channels (Hoshi et al., 1990) and a “hinged lid” (IFM motif in the domain III–IV loop) for voltage-gated Na channels (West et al., 1992). However, structural evidence has demonstrated that for Na channels, the block of the intracellular gate does not occur via direct block by the IFM motif, but rather by insertion of the IFM into a cavity outside the S6 bundle. The effect of this is predicted to decouple the pore domain and the S4–S5 ring, causing the S6 helices to close (Huang et al., 2024). Interestingly, this “door wedge” model of inactivation shares many features of the hinged lid, making them difficult to distinguish experimentally (Fig. 1 A). The next mechanism involves a collapse of the selectivity filter or the P-loop in the S5–S6 linker as observed in the C-type inactivation (also known as P-type inactivation) of the K⁺ channels (Hoshi et al., 1990; Choi et al., 1991). Finally, VDI could potentially proceed from an allosteric mechanism in which the gating of the channel is modulated (Tadross et al., 2010).

The kinetics of VDI varies based on α₁ subunit identity and channel-binding partners. VDI in Ca_V1.1 is kinetically slow compared with other Ca_V1 channels (Bannister and Beam, 2013), and the kinetics of VDI of Ca_V1.2 and Ca_V1.3 are faster than that of Ca_V1.4 (Koschak et al., 2003) and have been shown to display two kinetically distinct components (Ferreira et al., 2003). In Ca_V1 channels, VDI primarily occurs following the initial opening of the channel. Upon depolarization, the α₁ subunit undergoes a conformational change, initiated by the movement of the VSD, and results in the opening of the intracellular activation gate, comprised of the S6 segments (Xie et al., 2005). With continued depolarization, the channel rapidly enters an inactivated state. Experimental evidence provides support for either a hinged lid or door wedge mechanism underlying the fast N-type-like inactivation seen in Ca_V1.2, Ca_V1.3, and Ca_V1.4 channels (Fig. 1 A) (Stotz et al., 2004). The “lid” or “wedge” in these channels appears to be comprised of the loop between homologous domains I and II, which has been proposed to interact with the distal S6 regions. A few lines of evidence support such a mechanism for Ca_V1 channels. First, VDI seen at the single-channel level is demonstrated as a complete loss of channel opening during sustained depolarization (Fig. 1 B) (Reuter et al., 1982; Yue et al., 1990), indicating complete closure or block of the channel. Next, multiple chimeric and point mutation studies demonstrate the importance of the S6 region in the initiation of VDI (Kraus et al., 1998; Stotz et al., 2000, 2004; Stotz and Zamponi, 2001a, 2001b; Splawski et al., 2004, 2005; Hoda et al., 2005; Hohaus et al., 2005; Raybaud et al., 2006; Barrett and Tsien, 2008; Dick et al., 2016; Bamgboye et al., 2022a). For example, the slower VDI of Ca_V1.2 can be sped up by replacing the S6 region of domain II with a homologous region from a more rapidly inactivating Ca_V2.3 channel (Stotz et al., 2000), while individual point mutations within the same region significantly modulate VDI (Stotz and Zamponi, 2001a). Overall, these experiments point to the importance of the distal S6 region in Ca_V1 VDI but do not distinguish between the hinged lid and the door wedge mechanisms. Next, mutations and chimeric studies aimed at the I–II loop of the channel result in similar alterations of VDI (Bourinet et al., 1999; Stotz et al., 2000, 2004; Bernatchez et al., 2001; Berrou et al., 2001; Stotz and Zamponi, 2001a, 2001b; Dafi et al., 2004). Interestingly, overexpression of the I–II loop region by itself was sufficient to increase VDI (Cens et al., 1999), indicating a role of the I–II loop in Ca_V1 VDI.

The I–II loop of VGCCs also contains an α-interaction domain (AID), the known binding site of the auxiliary β subunit (De Waard et al., 1995). As such, both the kinetics and magnitude of VDI can be modulated by the interaction with the β subunit, further supporting the role of the I–II loop in Ca_V1 inactivation. Transferring the AID segment from α_1C to the α_1E backbone was shown to slow VDI (Bernatchez et al., 2001), while mutations introduced within the AID of Ca_V1.2 (α_1c) (Dafi et al., 2004), Ca_V1.3 (α_1D) (Tadross et al., 2010), or Ca_V2.1 (α_1A) led to significant alterations of VDI (Kraus et al., 1998). Furthermore, modulation of VDI is dependent on the β subunit variant. In particular, β_2a is palmitoylated, such that the subunit is anchored to the membrane, thus restricting the mobility of the I–II loop and impairing VDI (Fig. 1 C) (Olcese et al., 1994; Herlitze et al., 1997; Stephens et al., 2000). On the other hand, β₁ subunits which lack this palmitoylation allow the free movement of the I–II loop, permitting VDI. Interestingly, this regulatory process appears to be further fine-tuned in select channels. Ca_V1.3 has been proposed to contain a “shield” that prevents the complete closure of the hinged lid, leading to minimal VDI regardless of the type of β subunit. As evidence, mutagenesis of the shield in Ca_V1.3 channels renders them sensitive to β subunit isoforms (Tadross et al., 2010). However, these experiments could be explained by either a hinged lid or door wedge mechanism, where the shield mutation similarly obstructs the action of the I–II loop.

More recent structural studies have further demonstrated the complex nature of Ca_V1 VDI and call into question the hinged lid hypothesis. While each Ca_V1 structure has validated the distal S6 regions as the intracellular gate (Wu et al., 2015; Yao et al., 2022; Chen et al., 2023; Gao et al., 2023), they fail to show the I–II loop acting as a blocking particle. While this may point toward a door wedge mechanism, this feature has not yet been described. Thus, the detailed mechanism underlying the VDI of Ca_V1 channels remains to be fully elucidated.

In addition to fast N-type-like inactivation, Ca_V1 channels have also been shown to exhibit slower C-type-like inactivation (Catacuzzeno et al., 2023). For Ca_V1.1, this form of inactivation represents the dominant mechanism of channel inactivation and is associated with a collapse of the selectivity filter as described for C-type-like inactivation (Hoshi et al., 1990; Choi et al., 1991; Catacuzzeno et al., 2023). Likewise, Ca_V1.2 channels have been shown to undergo two kinetically distinct components of VDI (Ferreira et al., 2003). The slow component of Ca_V1.2 VDI may be explained by a similar C-type inactivation mechanism. In support of such a mechanism, the VDI of both Ca_V1.1. and Ca_V1.2 is accompanied by a characteristic negative shift in the voltage dependence of the gating charge (Brum et al., 1988; Pizarro et al., 1989; Ferreira et al., 2003; Villalba-Galea et al., 2008). Moreover, mutagenesis of the selectivity filter of Ca_V1.3 channels demonstrates a potential role for the selectivity filter in modulating VDI (Del Rivero Morfin et al., 2024). Mutation of a highly conserved domain IV tryptophan within the selectivity filter significantly enhances VDI in a β subunit–independent manner. Moreover, this form of VDI exhibited a distinct U-shaped dependence on voltage, which was insensitive to the identity of the charge carrier. Thus, it appears that Ca_V1 channels are capable of undergoing a distinct form of VDI resulting from a mechanism of selectivity filter collapse, pointing to more than one VDI mechanism at play. Moreover, the VDI of Ca_V1 channels may be further modulated by interaction with other proteins, including the adaptor protein STAC3, which has been shown to reduce the VDI of Ca_V1.1 (Tuinte et al., 2022) and Ca_V1.2 (Wong King Yuen et al., 2017) channels.

Compared with Ca_V1 channels, Ca_V2 channels have much faster kinetics (onset/recovery) of VDI (Stotz et al., 2000). Both Ca_V2.3 and Ca_V2.1 channels possess fast N-type-like inactivation, although Ca_V2.3 inactivation is shown to be faster than Ca_V2.1 (Zhang et al., 1994; Herlitze et al., 1997). Ca_V2.2 possesses both N-type-like and relatively slower inactivation similar to C-type inactivation (McDavid and Currie, 2006; Zhu et al., 2015). Similar to Ca_V1, the magnitude and kinetics of VDI in Ca_V2 are also modulated by types of β subunits (Buraei and Yang, 2013). However, unlike Ca_V1, inactivation of Ca_V2.2 and Ca_V2.1 can be further modulated by G protein βγ subunits (Herlitze et al., 1997; McDavid and Currie, 2006). Additionally, Ca_V2 channels also display a U-shaped dependence on voltage, independent of the ionic current (Patil et al., 1998). For these channels, VDI has been shown to proceed preferentially from an intermediate closed-state channel conformation (Patil et al., 1998; Jones et al., 1999). Structural evidence for Ca_V2.2 channels indicates that the channels contain a W-helix within the domain II–III linker, which locks the intracellular gate of the channel in the closed configuration (Dong et al., 2021). Within this closed-inactivated state, the side chain of residue W768 of the II–III linker extends into and blocks the intracellular gate, while multiple residues within the W-helix form hydrophobic interactions with residues residing in the distal S6 regions. Thus, the closed-state inactivation of Ca_V2 channels partially mirrors a hinged lid mechanism.

Modulation of Ca_V channels by intracellular Ca²⁺ was first reported in Paramecium by Brehm and Eckert (1978). They noticed that the inward Ca²⁺ current “relaxed within 10 ms” of initiation and that this inactivation disappeared when extracellular Ca²⁺ was replaced by strontium or barium. Since this remarkable discovery, a tremendous amount of cumulative work further sheds light on the Ca²⁺-dependent regulation of VGCCs, which takes the form of either CDI or Ca²⁺-dependent facilitation (CDF), where the channel open probability decreases or increases in response to a rise in cytosolic Ca²⁺, respectively.

Although there are several channel regions with Ca²⁺-binding capacity, including the EF hand-like region in the C-terminus of the α₁ subunit itself, calmodulin (CaM), a bilobal Ca²⁺-binding protein, is shown to be the central driver of both CDI and CDF in Ca_V1 and Ca_V2 channels (Lee et al., 1999; Zühlke et al., 1999; Lee et al., 2000; DeMaria et al., 2001). As evidence, overexpression of a CaM-binding inhibitor peptide and deletion of the CaM-binding region in the α_1A subunit of Ca_V2.1 channels render the channel unresponsive to changes in intracellular Ca²⁺ concentrations (Lee et al., 1999, 2000). Moreover, mutagenesis of the four EF hands, the Ca²⁺-binding motifs, across the N- and C-lobes of CaM (CaM₁₂₃₄) eliminates CDI in L-type channels and both CDI and CDF of Ca_V2.1 (Peterson et al., 1999; Zühlke et al., 1999; DeMaria et al., 2001). Of note, this striking dominant negative effect of CaM₁₂₃₄ in eliminating Ca²⁺-dependent regulation hints at the fact that CaM can interact with the α₁ subunit in a Ca²⁺-independent manner. ApoCaM (Ca²⁺-free CaM) pre-associates with the IQ domain of the α₁ subunit, as seen via FRET live-cell imaging (Erickson et al., 2001, 2003) and dansyl–CaM binding (Zühlke et al., 1999; Pitt et al., 2001). This pre-associated apoCaM acts as a resident Ca²⁺ sensor for the VGCCs. At resting or low levels of cytosolic Ca²⁺, the C-lobe of apoCaM engages the IQ domain in the C-terminus of the α₁ subunit, while the N-lobe appears to associate weakly with dual vestigial EF hands upstream from the IQ domain (Fig. 2 A) (Erickson et al., 2001, 2003; Pitt et al., 2001; Ben Johny et al., 2013). Upon the rise of intracellular Ca²⁺, both lobes of CaM are calcified and mobilized to various segments of the N- and C-termini of the α₁ subunit, initiating CDI. The IQ region has been identified as a primary locus for calcified C-lobe CaM binding (Lee et al., 2003; Chaudhuri et al., 2004; Van Petegem et al., 2005), possibly forming a tripartite complex with the two vestigial EF hands within the proximal C-terminus of the channel (Bazzazi et al., 2013; Ben Johny et al., 2013). The locus for the calcified N-lobe, on the other hand, is less well-defined. While the calcified N-lobe of CaM is capable of binding the IQ domain (Van Petegem et al., 2005; Mori et al., 2008; Minor and Findeisen, 2010), mutagenesis studies indicate that this interaction may not be functionally relevant for CDI within the holo-channel (Mori et al., 2008). For Ca_V1.2 and Ca_V1.3 channels, the calcified N-lobe of CaM has been shown to bind to an N-terminal spatial Ca²⁺-transforming element (NSCaTE) within the amino terminus of the channel (Ivanina et al., 2000; Dick et al., 2008; Tadross et al., 2008). However, a similar motif has not been identified for other channels.

Unlike VDI, where channel opening is completely blocked, CDI proceeds via an allosteric mechanism, whereby Ca²⁺ binding to CaM results in a transition of the channel from a high open probability mode (mode 1) to a lower open probability mode (mode Ca) (Fig. 2 B). At the single-channel level, this allosteric mechanism can be visualized as persistence in channel openings through the duration of a depolarization (Fig. 2 C) with a reduced open probability as compared with the initial mode 1 openings seen at the start of the depolarization in Ca²⁺ (see bottom averaged trace) or throughout the step in the absence of Ca²⁺ (Fig. 1 B) (Yue et al., 1990; Imredy and Yue, 1994). The structural determinants of this reduced channel opening remain largely unknown; however, domain II of the selectivity filter has been identified as a critical element, implicating the collapse of the selectivity filter as a potential endpoint for CDI (Abderemane-Ali et al., 2019). Thus, VGCCs may utilize a mechanism similar to the C-type inactivation of K⁺ channels for CDI.

CaM regulation of VGCCs displays a remarkable bipartition of regulation, such that each lobe of CaM can impart a distinct and independent form of channel regulation. The initial observation of this functional bipartition again emerged from observations in Paramecium, where mutations to residues within the N-lobe of CaM rendered Paramecium under reactive to stimuli, while mutations to residues in the C-lobe of CaM led to over reactive Paramecium, reflecting loss of either a Ca²⁺-dependent Na⁺ current or a Ca²⁺-dependent K⁺ current, respectively (Kink et al., 1990). Invaluable tools with targeted ablation of Ca²⁺-binding to the EF hands of N- or C-lobe CaM were developed to further dissect the Ca²⁺-dependent regulatory processes of ion channels (Xia et al., 1998). Mutagenesis of the two EF hands in the N-lobe (CaM₁₂) enables evaluation of C-lobe Ca²⁺ regulation, while mutagenesis of the EF hands in the C-lobe (CaM₃₄) provides evaluation of N-lobe CaM regulation (Liang et al., 2003). Overexpression of either CaM₁₂ or CaM₃₄ results in robust CDI in Ca_V1.3, visualized as the stronger decay of the Ca²⁺ current as compared with when Ba²⁺ is used as the charge carrier (Fig. 3 A), demonstrating the ability of either lobe to impart CDI. However, the kinetics of CDI differ significantly between the two lobes, demonstrating that each lobe is capable of imparting distinct regulatory features. For Ca_V2.1 channels, known to undergo both CDI and CDF, this bipartition of function is even more dramatic, with the N-lobe of CaM supporting CDI, while the C-lobe is responsible for CDF (Zühlke et al., 1999; DeMaria et al., 2001). On the other hand, Ca_V2.2 and Ca_V2.3 channels appear to only exhibit CDI due to N-lobe calcification (Liang et al., 2003). Thus, each lobe of CaM can impart an independent form of Ca²⁺ regulation, differing across channel subtypes (Fig. 3).

In addition to functional bipartition, CaM also exhibits spatial selectivity in Ca²⁺ sensing. This phenomenon is unveiled by the utilization of various types of Ca²⁺ buffers. In the presence of low or physiological Ca²⁺ buffering, Ca²⁺ entry consists of two major components. The local Ca²⁺ signal is comprised of rapid spikes of Ca²⁺ resulting from individual channel openings and closings (Fig. 3 B). A second, smaller global Ca²⁺ signal arises from the accumulation of cytosolic Ca²⁺ due to the aggregate opening of multiple channels and Ca²⁺ sources throughout the cell. Strong buffers may be used to restrict Ca²⁺ to the nanodomain, preventing the global accumulation of Ca²⁺ and enabling evaluation of the functional effect of local Ca²⁺ signaling (Stern, 1992). Interestingly, each lobe of CaM can distinguish between these two kinetically distinct Ca²⁺ signals. The C-lobe of CaM invariably responds to local Ca²⁺, such that regulation persists even in the presence of high intracellular Ca²⁺ buffering (Fig. 3 C). On the other hand, the N-lobe of CaM often requires a global elevation in Ca²⁺, which is only present under low intracellular Ca²⁺ buffering (DeMaria et al., 2001; Lee et al., 2003; Dick et al., 2008; Tadross et al., 2008). The mechanism by which this spatial Ca²⁺ selectivity comes about stems from the kinetics of Ca²⁺ binding to each lobe of CaM, where the C-lobe exhibits relatively slow Ca²⁺ (un)binding kinetics (slow CaM mechanism) ensuring a response to the large local Ca²⁺ spikes (Tadross et al., 2008). The N-lobe, on the other hand, can rapidly (un)bind Ca²⁺ (slow-quick-slow mechanism), enabling the lobe to respond to either the slow global Ca²⁺ signal or the rapid Ca²⁺ spikes of the local signal, dependent on the affinity of the calcified lobe for the channel. Thus, N-lobe spatial selectivity can be tuned for each channel subtype (Fig. 3 C). While the majority of VGCCs exhibit global Ca²⁺ selectivity for N-lobe CaM-mediated regulation, Ca_V1.2 and Ca_V1.3 channels contain the NSCaTE Ca²⁺/CaM-binding motif, which binds to the N-lobe of CaM, modifying the spatial selectivity of Ca_V1.2 and Ca_V1.3 channels from a global to a local selectivity (Dick et al., 2008; Tadross et al., 2008).

As a critical feedback mechanism, CDI can be modified by a variety of mechanisms, enabling fine-tuning of channel regulation suited to distinct biological needs. First, alternative splicing of the genes encoding the α₁ subunits may modify the elements responsible for CDI within select channel variants, resulting in significant changes in CDI. The distal C-terminus of Ca_V1.3 and Ca_V1.4 channels are known to contain a region known as an inhibitor of CDI (ICDI) or C-terminal modulatory (CTM) domain (Singh et al., 2006; Wahl-Schott et al., 2006; Sang et al., 2021) that can compete with apoCaM for binding to the IQ domain. This interaction results in a decreased open probability of the channel (Adams et al., 2014) and loss of CDI and can be modulated by cytosolic expression levels of CaM (Liu et al., 2010). The inclusion of ICDI/CTM is highly regulated by alternative splicing. In Ca_V1.3 channels, exon 42 gives rise to a long form of the channel containing ICDI/CTM, which exhibits minimal CDI. The inclusion of alternative exon 42A or utilization of an alternative 3′ splice-acceptor site in exon 43 produces a truncated channel bereft of the motif, thus restoring CDI (Singh et al., 2008; Bock et al., 2011). Additionally, the splicing of exons 44 and 48 modulates CDI by altering ICDI/CTM interaction with the IQ domain (Tan et al., 2011). Similar features were identified in Ca_V1.4, which does not display CDI due to the presence of ICDI/CTM in the full-length channel, but exhibits significant CDI upon either deletion of exon 47 (Williams et al., 2018) or inclusion of exon 43 containing a premature stop codon (Tan et al., 2012). Moreover, phosphorylation of this domain in Ca_V1.4 can further modulate the interaction, enabling dynamic regulation of CDI within select cell types (Sang et al., 2016). Thus, modification of the ICDI/CTM interaction with the IQ domain represents a highly tunable CDI modulatory mechanism in Ca_V1.3 and Ca_V1.4 channels.

Another example of alternative splicing as a method to diversify channel properties is found in Ca_V2.1. Splicing at the EF hands in (exon 37) and C-terminus distal to the CBD domain (exon 47) modulates the magnitude of CDF without altering CDI (Chaudhuri et al., 2004). The distribution of these splice variants is heterogeneous in the human brain, showing both spatial variation across brain regions and temporal variation across stages of brain development. This added layer of complexity likely allows cells to fine-tune the properties of Ca_V2.1 channels to best suit their functional requirement across time and space.

In addition to splice variation, RNA editing represents another mechanism that enables the modulation of CDI. For Ca_V1.3, the IQ motif is subject to significant RNA editing, resulting in channels with altered affinity between the channel and apoCaM (Bazzazi et al., 2013; Huang et al., 2012). By modulating the interaction between the channel IQ region and ICDI/CTM, RNA editing tunes the magnitude of CDI in Ca_V1.3. Notably, the pattern of RNA editing varies across different regions of the brain, rendering a landscape of Ca_V1.3 channels with varying degrees of CDI that best serves their functional requirements (Huang et al., 2012).

Further fine-tuning of CDI can occur through the interaction of the channel with multiple proteins. The CaM-like calcium-binding protein (CaBP) has been shown to bind to Ca_V1.2 and Ca_V1.3, both competitively against CaM and allosterically (at a binding site distinct from the IQ region), resulting in significant changes in the magnitude of CDI (Yang et al., 2006; Findeisen and Minor, 2010; Yang et al., 2014; Hardie and Lee, 2016). CaBP4 is highly expressed in inner hair cells and impedes CDI of Ca_V1.3 (Yang et al., 2006, 2014), a modulation necessary for continued perception of sustained auditory stimuli (Lewis and Hudspeth, 1983). CaBP1, found primarily in neurons, inhibits CDI and unveils CDF of Ca_V1.2 (Zhou et al., 2005; Findeisen and Minor, 2010), further demonstrating the complexity of channel regulation in the brain. Additionally, SH3 and cysteine-rich domain (STAC) can bind to and modulate the gating of VGCCs (Polster et al., 2015, 2018; Campiglio et al., 2018; Niu et al., 2018b; Flucher and Campiglio, 2019). STAC1 and STAC2 are expressed primarily in the brain (Suzuki et al., 1996; Nelson et al., 2013), peripheral nervous system (Legha et al., 2010), retina (Wilhelm et al., 2014), and inner ear (Cai et al., 2015), while STAC3 is expressed in the skeletal muscle (Nelson et al., 2013). All three STAC isoforms are shown to impede CDI of Ca_V1.2 and Ca_V1.3 without alteration of CDI in Ca_V2.1 channels (Niu et al., 2018a; Polster et al., 2018). These “third-party” binding partners add yet another layer of sophistication to the Ca²⁺-dependent regulatory processes of VGCCs.

VGCCs are present in all excitable cells such that channel inactivation profiles have the potential to impact a myriad of physiological functions. One of the most well-recognized roles for VGCC inactivation occurs in the heart, where Ca_V1.2 plays a critical role in shaping action potential (AP) morphology. The channel opens primarily during the plateau phase of the AP, and inactivation occurs during the sustained depolarization, reducing the depolarizing effect of the channel. The importance of this inactivation process has long been recognized, yet it is not always trivial to identify the relative impact of VDI versus CDI. In fact, computational models of the cardiac AP demonstrate that either or both VDI and CDI can support the inactivation required to produce a physiological AP (Livshitz and Rudy, 2007; Morotti et al., 2012). Reduction of CDI via overexpression of either CaM₁₂₃₄ or CaM₃₄ in adult guinea pig ventricular myocytes results in a profound prolongation of the AP (Fig. 4 A) (Alseikhan et al., 2002), demonstrating the importance of CaM in controlling the AP duration and implicating CDI as a critical regulator of cardiac AP morphology. Yet, the conclusion is complicated by the myriad of CaM effectors in the heart and does not address the potential impact of VDI. To address this, Morales et al. (2019) utilized a combination of CaM mutants and overexpressed β subunit variants to demonstrate distinct effects of CDI and VDI on cardiomyocyte function, where disruption of either process modulated the AP at baseline, with a larger effect due to CDI disruption. In addition, under conditions of β-adrenergic stimulation, CDI was shown to be of paramount importance as compared with VDI. Thus, both processes contribute significantly to the AP morphology, yet CDI appears to exert a greater impact.

In the brain, VGCCs play important roles in synaptic transmission, excitation–transcription coupling, and long-term potentiation. Altered Ca²⁺ entry through multiple VGCCs can disrupt numerous neurological functions, yet the specific role of inactivation in these processes is often difficult to determine. While the rapid time course of a neuronal AP reduces the importance of inactivation in shaping a single AP, as seen in cardiac myocytes, trains or bursts of APs may be modulated by VGCC inactivation properties. For Ca_V2.1, CDI and CDF have been shown to play an important role in synaptic plasticity (Cuttle et al., 1998; Tsujimoto et al., 2002; Xu and Wu, 2005; Mochida et al., 2008; Adams et al., 2010) and have been implicated in long-term potentiation and memory (Nanou et al., 2016). Moreover, numerous processes in the brain have been shown to modulate the extent of VGCC inactivation in a tissue-specific manner. RNA editing of the Ca_V1.3 channel has been shown to reduce CDI and occurs in as many as half the transcripts in the brain (Huang et al., 2012) and is controlled in a neuron-specific manner by ADAR (Huang et al., 2018). This editing is thought to play a role in the rhythmicity of the suprachiasmatic nucleus (Huang et al., 2012) and in synaptic plasticity, learning, and memory (Zhai et al., 2022). In addition, splice variation has been shown to modify the inactivation of multiple VGCCs, demonstrating a role for inactivation modulation in neurophysiology. Ca_V2.1 channels contain alternative exons within the C-terminus of the channel known to significantly modulate CDI and CDF (Soong et al., 2002). These variants are differentially expressed (Bunda and Andrade, 2022) and are known to modulate the pathogenesis of select Ca_V2.1 variants (Adams et al., 2009, 2010) and factor into protecting the brain from disease (Aikawa et al., 2017), demonstrating the physiological importance of CDI and CDF. For Ca_V1.3 and Ca_V1.4, the presence of ICDI/CTM within the long channel variant significantly suppresses CDI (Scharinger et al., 2015; Williams et al., 2018), enabling tissue-specific suppression of inactivation in the retinal ganglion, chromaffin, and inner hair cells, and modulation of CDI across different tissues and in development (Lieb et al., 2012; Scharinger et al., 2015; Williams et al., 2018).

In addition, multiple subunits are known to modify VGCC regulation, demonstrating the importance of tuning VDI and CDI in distinct contexts. For example, the α₂δ subunit has been shown to accelerate channel inactivation, and co-expression of α₂δ with Ca_V2.1 and Ca_V2.2 resulted in increased coupling to exocytosis and enhanced synaptic transmission (Hoppa et al., 2012). On the other hand, the β_2A subunit strongly diminishes the VDI of most VGCCs as compared with other β subunits and can have significant effects on synaptic function (Wemhöner et al., 2015). Physiologically, this phenomenon has been associated with asynchronous transmission resulting from prolonged openings of Ca_V2.1 at the synapse (Müller et al., 2010). RIM proteins modulate channels both through interaction with the β subunit and G-protein receptors, causing a decrease in VDI of Ca_V2.2 (Weiss et al., 2011). Likewise, RIM2α was shown to slow the inactivation of Ca_V1.3 in inner hair cells (Gebhart et al., 2010). Such specialized modulation of inactivation appears to be of particular importance in the auditory and visual systems, where slow inactivation of Ca_V1.3 and Ca_V1.4 may be critical to enable the graded responses of these synapses (McRory et al., 2004; Inagaki and Lee, 2013). For Ca_V1.3 in cochlear inner hair cells, VDI is suppressed in a splice-dependent manner by a combination of β subunit identity and association of the channel with presynaptic scaffolding proteins, including RIM2α and RBP2 (Ortner et al., 2020b). CDI in these cells is likewise suppressed by the presence of CaBP4 (Yang et al., 2006), as well as the expression of the long ICDI/CTM-containing channel variant (Scharinger et al., 2015). Overall, numerous mechanisms impart exquisite control over inactivation across different tissues and throughout development, demonstrating the importance of CDI and VDI in physiology.

Numerous genetic mutations involved in the disruption of channel inactivation have been associated with severe clinical manifestations. Timothy syndrome (TS), also known as long-QT syndrome type 8 (LQT8), is a multisystem disorder often including profound long-QT syndrome (LQTS) and neurodevelopmental disorders (NDD), resulting from a genetic mutation within Ca_V1.2. The canonical TS mutation, G406R, causes a profound reduction in both CDI and VDI (Fig. 4 B) (Splawski et al., 2004, 2005; Barrett and Tsien, 2008; Dick et al., 2016). Like G406R, many TS mutations impact CDI, VDI, and activation (Splawski et al., 2004, 2005; Barrett and Tsien, 2008; Boczek et al., 2015; Wemhöner et al., 2015; Dick et al., 2016; Landstrom et al., 2016; Bamgboye et al., 2022a), making it difficult to determine the relative impact of each form of channel regulation. However, selected mutations associated with LQTS have been shown to exhibit more selective effects on CDI or VDI (Bamgboye et al., 2022a), demonstrating that each form of channel regulation can contribute significantly to the cardiac features of the disorder. Importantly, even a modest alteration of CDI (Bamgboye et al., 2022a) may correlate with LQTS and sudden cardiac death (Wemhöner et al., 2015). Conversely, one mutation (L762F), which appears to have a more selective effect on VDI (Bamgboye et al., 2022a), correlated with a somewhat less severe presentation of LQTS across multiple members of a single family (Landstrom et al., 2016). However, it is difficult to draw direct conclusions due to limited known occurrences of the mutations and the existence of other cardiac variants within the family members harboring the L762F mutation (Landstrom et al., 2016). Nonetheless, the pattern of biophysical effects (Bamgboye et al., 2022a) thus far appears to point to more severe cardiac phenotypes in patients harboring mutations involving a loss of CDI. Consistent with the idea that CDI plays a dominant role in determining the cardiac AP, genetic mutations in CaM (calmodulinopathies), which disrupt Ca²⁺ binding to CaM and thus severely diminish CDI of Ca_V1.2, also cause a profound prolongation of the cardiac AP and severe LQTS (Limpitikul et al., 2014; George, 2015; Jensen et al., 2018; Crotti et al., 2019; Nyegaard and Overgaard, 2019). Thus, in the heart, disruptions of either VDI or CDI can dramatically prolong the AP and cause severe pathologies, although decreases in CDI appear likely to produce more significant phenotypes.

In the brain, TS mutations have been linked to NDD, pointing to a causative role for overt increases in Ca²⁺ entry (Splawski et al., 2004; Limpitikul et al., 2016; Pinggera and Striessnig, 2016; Pinggera et al., 2017). However, unlike in the heart, a severe NDD phenotype associated with many TS patients (often presenting as autism spectrum disorder) appears to correlate more consistently with hyperpolarizing shifts in channel activation rather than inactivation (Marcantoni et al., 2020; Bamgboye et al., 2022a). Similar mutations within the Ca_V1.3 channel appear to follow an analogous pattern, where mutations alter multiple channel biophysical properties, but a severe NDD phenotype appears to correlate largely with activation changes (Pinggera et al., 2015; Ortner et al., 2023; Dannenberg et al., 2024). Nonetheless, this result represents only the severe neurodevelopmental deficits historically described in TS. As the recognition of patients harboring Ca_V1.2 mutations has increased, so too has the variety of reported symptoms and phenotypes, with a particular increase in recognized neurological symptoms (Gillis et al., 2012; Bozarth et al., 2018; Rodan et al., 2021; Levy et al., 2023; Timothy et al., 2024). Moreover, patients harboring mutations in CaM (calmodulinopathies) often exhibit neurological features, including seizures and neurodevelopmental delay (Crotti et al., 2013; Retterer et al., 2016; Takahashi et al., 2016; Crotti et al., 2019; Hussey et al., 2023b). Thus, it is likely that deficits in CDI play a critical role in neuropathogenesis.

While it is difficult to rule out hypoxic brain injury secondary to cardiac arrest in cases involving Ca_V1.2 mutations, multiple neurological phenotypes have been associated with inactivation deficits in VGCCs primarily expressed in the brain. Inactivation changes have been identified in multiple Ca_V1.4 mutations, resulting in congenital stationary night blindness type 2 (Hemara-Wahanui et al., 2005; Hoda et al., 2005, 2006). While many of these mutations do not represent selective effects on inactivation, truncation mutations that remove the CTM/ICDI region from the channel cause a predominant increase in CDI and congenital stationary night blindness type 2, demonstrating the importance of sustained (non-inactivating) current to support the tonic release of neurotransmitters at sensory cell ribbon synapses (Singh et al., 2006; Wahl-Schott et al., 2006; Griessmeier et al., 2009). Mutations in Ca_V2.1 associated with episodic ataxia type 2 have been shown to alter multiple channel properties, including modulation of inactivation (Spacey et al., 2004; Cuenca-León et al., 2009). In many cases, the distinction between CDI and VDI effects was not made, although the use of Ba²⁺ implicates increased VDI in the pathogenesis of select mutations (Spacey et al., 2004). Likewise, Ca_V2.1 mutations associated with familial hemiplegic migraine type 1 (FHM1) have also been shown to alter inactivation, in some cases increasing inactivation (Serra et al., 2009). Kinetic changes in VDI due to the S218L mutation in Ca_V2.1 have been implicated in the pathogenic mechanism underlying cortical spreading depression in FHM1 (Tottene et al., 2005). Alternative splicing of Ca_V2.1 channels, which alters multiple channel properties, including recovery from inactivation, modulates the pathogenesis of FHM1 (Adams et al., 2009). Likewise, mutations in Ca_V2.3, which are causative of developmental and epileptic encephalopathies have been identified within the distal S6 regions and shown to modulate channel activation and inactivation (Helbig et al., 2018; Sampedro-Castañeda et al., 2023). In addition, amyotrophic lateral sclerosis has been associated with a reduction in RNA editing of the frontal cortex resulting in alterations in Ca_V1.2. Karagianni et al. (2024), implicating altered CDI of Ca_V1.2 in the pathogenesis of the disease. Finally, calmodulinopathy mutations not only reduce CDI in Ca_V1.2 channels, but decrease CDI in Ca_V1.3 and diminish CDF in Ca_V2.1, potentially contributing to the neurological effects identified in some patients (Hussey et al., 2024, Preprint). Thus, the pathogenic effects of disrupted VGCC inactivation mechanisms appear to be a recurrent theme, with the identification of the specific impact of each form of channel regulation within different tissues representing an important question still to be resolved.

Many Ca²⁺ channel blockers (CCBs) are known to be state dependent, such that the drug interacts preferentially with the inactivated channel. The state-dependent block of Ca_V1 channels by multiple CCBs has been well characterized (Kanaya et al., 1983; Hering et al., 1997; Berjukow et al., 2000). CCBs are divided into three main classes: dihydropyridines, phenylalkylamines, and benzothiazepines. Dihydropyridines are FDA-approved Ca_V1 channel blockers, often used in the treatment of hypertension. These drugs bind to a hydrophobic pocket at the interface between domains III and IV of the pore domain, accessible via fenestrations, which are state dependent (Zhao et al., 2019; Gao and Yan, 2021). As a result, conformational changes modulate access of the drug to the binding site, resulting in an enhanced block of the inactivated state of the channel. Phenylalkylaminess and benzothiazepines bind to an overlapping site within the central cavity of the channel (Zhao et al., 2019) and also show preferential binding to the inactivated state of the channel. This state-dependent block of CCBs provides a use-dependent feature, where sustained or repeated depolarization increases channel block (Hering et al., 1997). Such state-dependent block can also be observed in Ca_V2 channels. Multiple state-dependent inhibitors of Ca_V2.2 have been identified as potential analgesics (Abbadie et al., 2010; Pajouhesh et al., 2010; Swensen et al., 2012), with a block of the channel biased toward open and inactivated states.

This inactivation-dependent block can further modulate the efficacy of these drugs in the context of inactivation-modulating channel mutations. Genetic mutations in VGCCs which reduce inactivation result in significant increases in Ca²⁺ current. Although CCBs appear to present a useful therapeutic option, multiple reports indicate a lack of efficacy in these patients. In particular, verapamil has been tried in the treatment of both TS and calmodulinopathies but failed to reduce the QT interval (Jacobs et al., 2006; Shah et al., 2012; Crotti et al., 2013; Webster et al., 2017). This lack of efficacy may be attributed to the state-dependent properties of CCBs, including verapamil (Johnson et al., 1996; Catterall, 2000), where the drug binds preferentially to the inactivated state of the channel. Thus, the drug does not inhibit the mutant channels lacking inactivation to the same degree as wild-type channels (Sheng et al., 2012; Bamgboye et al., 2022b), failing to rescue the prolonged AP produced by the TS mutation (Fig. 4 C). While this state-dependent effect may be a disadvantage for mutations causing deficits in inactivation, it could also represent an opportunity for channels where inactivation is enhanced, resulting in increased CCB efficacy on the mutant channel as has been shown for select Ca_V1.3 mutations (Hofer et al., 2020; Ortner et al., 2020a). Thus, channel inactivation stands as a critical mechanism in both normal physiology and pathophysiology, capable of impacting the progression of disease and options available for treatment. However, the role of inactivation in the pathogenesis of multiple disorders also points to a potential therapeutic target with significant advantages compared with a complete block of the channel. For example, roscovitine has been shown to enhance the VDI of Ca_V1.2. (Yarotskyy and Elmslie, 2007; Yarotskyy et al., 2010), making it a promising candidate for the treatment of LQTS and TS (Yarotskyy et al., 2009; Karagueuzian et al., 2017; Song et al., 2017; Angelini et al., 2021), potentially bypassing the limitations of CCBs.

Eduardo Ríos served as editor.

We thank all members of Ivy Dick’s lab at the University of Maryland School of Medicine for their support. We also thank Dr. Manu Ben-Johny from Columbia University for insightful comments and discussion.

This project was supported by a National Institutes of Health/National Heart, Lung, and Blood Institute grant (1R01HL149926).

Author contributions: W.B. Limpitikul: writing—original draft, review, and editing. I.E. Dick: conceptualization, funding acquisition, visualization, and writing—original draft, review, and editing.