Voltage-gated calcium channels (CaVs or VGCCs) are transmembrane ion channel proteins that act as major regulator of calcium-related cell functions. Their primary role is to mediate transmembrane calcium influx in response to membrane depolarization. Depending on their sensitivity to membrane depolarization, the activation of CaVs requires either a high or low threshold of membrane potential, dividing CaVs in High- and Low-voltage activated channels (HVA and LVA respectively) (Carbone and Lux, 1984; Fedulova et al., 1985). The crucial component of the channel pore is CaVα₁, for which ten variants have been identified and classified based on their pharmacological properties and pore-opening kinetic: CaV1 and CaV2 for HVA, and CaV3 for LVA (Tsien et al., 1988; Catterall, 2011; Zamponi et al., 2015).

The CaV of skeletal muscle, also called dihydropyridine receptor (DHPR), was the first to be purified and cloned (Curtis and Catterall, 1984; Tanabe et al., 1987). In skeletal muscle fibers, CaV has a dual function of calcium channel and of voltage sensor of excitation-contraction coupling that controls, through a direct interaction, the opening of RyR1 (Ryanodine Receptor type 1), the Ca²⁺ release channel of the sarcoplasmic reticulum (Allard, 2018). Such a voltage sensor function for CaV and a direct interaction between CaV and RyR have also been described in neurons (Allard, 2018).

The main subunits of CaV, the αa1 subunits, are associated with auxiliary subunits that modulate expression and/or functional properties of the channel. In skeletal muscle, CaV1 is composed of five subunits: α1S (or CaV1.1), β1, α2δ, and γ (Hagiwara and Naka, 1964).

The function of both CaV1 and CaV2, members of HVA channels, needs the association of the auxiliary CaVβ subunit for their plasma membrane docking and proper gating (Schredelseker et al., 2005; Dayal et al., 2013), while the function of LVA class of channel (CaV3) is independent of this subunit (Zhang et al., 2013).

CaVβ are intracellular proteins that can either interact with the channel or be in their soluble form (Buraei and Yang, 2013).

Four different proteins, namely CaVβ1, CaVβ2, CaVβ3, and CaVβ4 encoded by four genes, exist in mammalians, each of them having several splice variants (Buraei and Yang, 2010). All CaVβ subunits are membrane-associated guanylate kinase (MAGUK) family members, with SH3 and GK domains as conserved domains whereas hook region, N- and C-terminal sequences are variables (Buraei and Yang, 2013). Hence, splice variants originate from alternative exon splicing and harbor different amino acidic compositions of variable regions, leading to specificities in protein interaction (Subramanyam et al., 2009; Obermair et al., 2010), subcellular targeting properties, and cellular localization, all influencing channel complexes stability, and activity (Campiglio et al., 2013).

Intracellular Ca²⁺ changes account for eukaryotic cell adaptation to external stimuli by modifying gene expression. By controlling Ca²⁺ influx into the cell, CaVs are therefore at the key position to mediate excitation-transcription (E-T) coupling. In point of fact, the mechanisms leading to CREB (cyclic AMP response element-binding) or NFAT (nuclear factor of activated T-cells) activation require the CaVs-dependent Ca²⁺ signaling (Dolmetsch et al., 2001; Hernández-Ochoa et al., 2007; Zhao et al., 2007). Noteworthy, if the role of CaVs in E-T coupling is mainly restricted to the initiation of the subsequent transcriptional activity of Ca²⁺, other mechanisms have been demonstrated to initiate gene regulation by generating a shorter isoform of the CaV1.2 pore forming subunit, which relocalized to the nucleus and held a transcription factor activity, (Gomez-Ospina et al., 2006; Gomez-Ospina et al., 2013). or by mobilizing intracellular CaVβ subunit after conformational changes upon membrane depolarization (Servili et al., 2018). Indeed, CaVβ2 subunit was recently demonstrated to be the mediator of CaV1.2-dependent E-T coupling, by interacting with H-Ras, which in turn activated MAPK (Mitogen Activated Protein Kinase)/ERK (Extracellular Signal-Regulated Kinase) pathway to induce CREB-directed gene expression in human neuronal SH-SY5Y cells (Servili et al., 2018). Nevertheless, the possibility that CaVs auxiliary subunits may be directly implicated as transcription factors has become an emerging hypothesis in the last 2 decades (Barbado et al., 2009).

This review will focus on CaVβs newly and less broadly described insights by illustrating their nuclei tracking in line with their role as regulators of gene expression.

After the initial cloning of CaVβ1 in skeletal muscle [CaVβ1D formerly known as CaVβ1A (Traoré et al., 2019)] (Ruth et al., 1989), further works described several variants of CaVβ1 expressed in muscle and other tissues CaVβ1 (Hibino et al., 2003; Hullin et al., 2003; Foell et al., 2004; Harry et al., 2004; Cohen et al., 2005), lacking the domain required for its interaction with CaVα _1S at the α-Interaction Domain (AID domain). These reports suggested for the first time that some CaVβ subunits isoforms may have a CaV-independent function. Subsequently, several studies have demonstrated the capability of CaVβs to translocate to the nucleus, giving additional indications toward a role for CaVβ distinct from its well-known function as modulator of CaV channels.

If not all, at least some isoforms of CaVβ1, CaVβ2, CaVβ3 and CaVβ4 proteins display nuclear localization properties upon appropriate conditions (Buraei and Yang, 2010). For both CaVβ1, in skeletal muscle (Traoré et al., 2019), and CaVβ4, in neurons (Subramanyam et al., 2009; Etemad et al., 2014), this nuclear localization has been demonstrated to be linked to electrical activity. Indeed, our recent study showed that in adult skeletal muscle, after nerve damage, the embryonic isoform CaVβ1E was expressed and localized to the nuclei and near the Z-lines, while the constitutive adult muscle variant, CaVβ1D, remained associated with CaV1.1 at the sarcolemma (Traoré et al., 2019). Similarly, it has been reported that in neurons, CaVβ4A and CaVβ4B translocated to the nuclei when electrical activity was aborted, whereas CaVβ4E did not display nuclear localization. Additionally, a decrease in nuclear targeting of CaVβ1E (Traoré et al., 2019) and CaVβ4 (isoforms A and B) (Etemad et al., 2014) has been correlated with the onset of electrical activity throughout development in muscle fibers and neurons. Interestingly, the proportion of CaVβ4 isoforms targeted to nuclei has been associated with their activity as gene regulators (CaVβ4B > CaVβ4A > CaVβ4E) (Etemad et al., 2014). The mechanism originating the nuclear localization of CaVβ2 and CaVβ3 has not been clearly characterized, however, it can be hypothesized that they follow the same depolarization-sensitive process.

As previously mentioned, the capacity to get to the nucleus is not held by all the CaVβ1, CaVβ2, CaVβ3 and CaVβ4 splicing variants. The mechanisms underlying the specificities of the nucleus-targeted proteins could be passive diffusion through nuclear membrane for small proteins, while large ones need a Nuclear Localization Sequence (NLS), allowing their binding to Importins, or require the association with nuclear proteins as a shuttle. The molecular aspects behind CaVβs translocation to the nucleus are still not fully understood. For CaVβ1 (Buraei and Yang, 2010; Taylor et al., 2014) and CaVβ4 (Tadmouri et al. 2012) the SH3 domain of the protein has been described to exhibit the functional features leading to nuclear shuttling. An additional aspect was highlighted for CaVβ1E which have been described to display a putative NLS signal in its sequence (Taylor et al., 2014; Traoré et al., 2019), suggesting that its nuclear targeting was occurring through its binding to Importin proteins. However, modified genetic constructions lacking the putative NLS sequence did not prevent CaVβ1 to enter the nucleus. As an example, Subramanyam and colleagues showed that a specific double-arginine motif at the N-terminal was necessary and sufficient to induce the recruitment of the CaVβ4B variant toward the nuclei in mouse brain (Subramanyam et al., 2009). Nevertheless, this domain was subsequently demonstrated to be only partially involved in CaVβ4B docking to the nucleus, and that SH3/GK protein interaction domain was required to control its nuclear targeting. If these data were confirmed by several studies, a supplemental and non-exclusive mechanism came up with the demonstration of a PxxP binding motif in the SH3 domain of CaVβ1, raising the possibility that CaVβ proteins might also bind to proteins that themselves shuttle to the nucleus (Buraei and Yang, 2010). An instance supporting this hypothesis is CaVβ4C, which interaction with HP1γ has been shown as mandatory to localize to the nucleus in mammalian cells (Hibino et al., 2003), while the truncation of a large part of its GK domain cut off the exclusive requirement of SH3/GK interaction for nuclear docking of CaVβs proteins. The molecular aspects of nuclear targeting were less studied for CaVβ2 and CaVβ3, for which some studies mentioning their binding with chaperone proteins may be relevant in supporting their tackling to the nuclei (Zhang et al., 2010; Pickel et al., 2021).

The evidence of the nuclear localization/translocation of several CaVβ variants spotted these proteins with an undeniable CaVα ₁ independent function and pinpointed their putative role in the modulation of gene expression. The next part of this review will summarize the mechanisms described for the CaVβ auxiliary subunits in the regulation of gene transcription.

Although isoforms of CaVβ1, CaVβ2, CaVβ3, and CaVβ4 are expressed in several regions of the brain (Buraei and Yang, 2013), ablation of CaVβ1, CaVβ2 and CaVβ3 have no major impact on neuronal function (Ball et al., 2002). On the contrary, the relevance of CaVβ4 in the nervous system physiology was shown in the lethargic (lh) mouse model, having ataxic and epileptic phenotype (Burgess et al., 1997). Its implication in the pathophysiology of neuronal disorders was confirmed in Humans, after the discovery that missense and coding mutations, affecting the N-terminal region of the protein, were associated with epilepsy and ataxia, respectively (Escayg et al., 2000; Tadmouri et al., 2012). In the cerebellum, CaVβ4 is the most expressed and the major auxiliary subunit, together with α2δ-2, of the CaV2.1 calcium channel. Interestingly, mutations in all three proteins have been reported to lead to an epileptic and ataxic phenotype. Therefore, CaVβ4 involvement in such pathological conditions was first linked with its CaV-associated role (Escayg et al., 2000).

However, an additional mechanism to further elucidate such pathologies came from the CaV-independent role of CaVβ4. Indeed, at the molecular level, human epilepsy and ataxia-associated mutations were found to prevent CaVβ4 to shuttle toward nuclei by disrupting the SH3/GK domains interaction and indicated that mis-regulated CaVβ4-dependent gene transcription may have a key relevance in the pathophysiology of these neurological disorders (Tadmouri et al., 2012) (Table 1).

In 2017, the importance of CACNB2 (gene coding for CaVβ2) as a genetic modifier of a Hypertrophic CardioMyopathy (HCM), in which the causal gene was MYBPC3 (Myosin-Binding Protein C), has been described for the first time (Zhang et al., 2017). The authors hypothesized that the potential mechanism modifying disease phenotype was based on the attenuation of CaV-dependent Ca²⁺ current associated with CACNB2 mutations. However, an additional possibility came out a few years later with a study that correlated the reduced cardiomyocyte hypertrophy to a CaV-independent CaVβ2 function (Pickel et al., 2021). CaVβ2 localization and activity in cardiomyocyte nuclei were shown to significantly regulate Calpain activity and Calpastatin expression (Pickel et al., 2021), a pro-hypertrophic protease and its inhibitor, respectively. Even though these events have not been explicitly linked to the reduction of cardiomyocyte hypertrophy, a correlation between the two might be hypothesized and would need to be further studied (Table 1).

Our recent work demonstrated the key role played by CaVβ1E in the context of age-related muscle wasting. We showed that CaVβ1E/GDF5 pathway counteracted the loss of muscle mass after denervation and that this signaling was defective in aged muscle fibers. Importantly, we overexpressed CaVβ1E in aged mouse muscles leading to increased GDF5 expression and activation of its signaling and therefore enabling the prevention of muscle mass loss and force decline during aging (Traoré et al., 2019). Importantly, the expression of an analogous of CaVβ1E has been also discovered in human muscle, decreasing in an age-related manner, indicating that the defective hCaVβ1E signaling might also be impaired in sarcopenic patients and suggesting the CaVβ1E/GDF5 axis having a therapeutic potential in muscle aging and linked pathologies (Traoré et al., 2019) (Table 1).

Nevertheless, little is known about the causes behind the decrease of CaVβ1E expression in aged muscles. One hypothesis we assessed, was a damaged neuro-muscular junction (NMJ), but we did not detect any NMJ changes in the 78-week old mice involved in the study which could have testified toward changes in electrical activity and modifications in basal CaVβ1E levels (Traoré et al., 2019). Chromatin methylation/demethylation events or other epigenetic alterations could further explain the unbalanced CaVβ1E/GDF5 axis in old muscle and should be investigated in future works.

To summarize, mutations in CaVβs encoding genes have currently been associated with disorders and, even if the pathological mechanisms have not always been fully characterized, both CaVβs roles linked or unlinked to CaVs can be argued. Brugada Syndrome type 4 and Episodic ataxia type 5 present causative mutations in genes coding for CaVβs proteins and other CaV subunits (CACNA2D1 and CACNA1A, and CACNA1A respectively–MalaCard database), corroborating CaVβ involvement in pathological mechanisms in a CaV-linked way (Table 1). Nevertheless, and as described above, CaVβ may also originate or modify pathological features independently of CaVs and an underestimation of these situations can be hypothesized.

While CaVβs have long been considered to present exclusively CaVs’ linked functions, we depicted in this review the elements pointing towards a far more extensive significance in organ and cell homeostasis. CaVβs proteins have been described as efficient modulators of gene expression, either through their effect on Ca²⁺ signaling or through their DNA-related activities enabled by their nuclear localization. Regardless of the molecular mechanisms in which CaVβs are implicated, these proteins have been shown to influence cell and tissue capability to respond to different stimuli and to adapt following environmental changes. In the light of the findings of these two last decades, the impact of CaVβ1 and CaVβ4 on gene expression has been demonstrated several times, whereas CaVβ2 and CaVβ3 are less deeply characterized. A better establishment and appreciation of CaVβs relevance in cell biology will probably strengthen our understanding in a plethora of physiological and pathological mechanisms.