Calcium channels: when is a subunit not a subunit?

The textbook view is that a calcium channel contains four subunits: α₁, α₂-δ, β and γ. This was initially based on biochemical purification of the L-type channel from skeletal muscle, as all four subunits remained associated during the harsh conditions needed for protein purification. The core of the channel is the α₁ subunit, which contains the voltage sensors, ion pore, and binding sites for drugs such as dihydropyridines. So what are the roles of the other subunits?

In Xenopus oocytes, the α₁ subunit alone produces a low level of channel activity, but Tarelius et al. (1997) found that antisense oligonucleotides against β subunits prevented functional expression of α₁. Furthermore, coexpression of α₁ with additional β both increased current amplitudes and modified channel gating. Thus, β subunits must have two separable effects: they not only facilitate initial channel expression, but also modify the kinetics of existing channels (Tareilus et al. 1997).

How can β subunits have two distinct actions? Figure 1 shows two possibilities. Perhaps there is only one β binding site, and β interacts with α₁ at two distinct stages of the calcium channel lifecycle (1-site model). Reversible binding of β to α₁ facilitates expression of α₁ in the plasma membrane, perhaps by masking an endoplasmic reticulum retention signal (Bichet et al. 2000). While in the plasma membrane, α₁ can reversibly associate with β, changing a minimally functional α₁-only channel to a mature channel. These α₁-β interactions are sequential, and thus could involve a single binding site. In the alternative view (2-site model), stoichiometric assembly of the α₁-β complex is necessary for expression of α₁ in the cell membrane, while occupancy of a second (low affinity?) site modulates channel kinetics. Both models seem consistent with existing data, although there is evidence for a second binding site (Birnbaumer et al. 1998; Gerster et al. 1999).

Figure 1. Two models for interactions of calcium channel α₁ and β subunits.

ER, endoplasmic (or sarcoplasmic) reticulum; PM, plasma membrane. α*₁ indicates the high-activity form of the calcium channel.

Put bluntly, are β subunits really stoichiometric subunits of the Ca²⁺ channel, or are they just modulatory proteins that can be associated with Ca²⁺ channels under some circumstances (like G protein βγ subunits, syntaxin, ryanodine receptors, etc.)? In the 1-site model, the β subunit appears as a modulatory protein, with two distinct effects at different stages. In the 2-site model, β is both a true subunit and a modulatory protein (at distinct binding sites).

One key question is how β modulates preexisting α₁ subunits in the plasma membrane. How rapid is that action? Is it reversible? Are α₁ subunits normally saturated with β? These questions have been difficult to address, in part because most studies have examined effects of β on α₁ by coexpression, which does not allow direct determination of whether the interaction is dynamic. Previous evidence for reversible association of β with α₁ has mostly been indirect (Canti et al. 2001; Restituito et al. 2001), although injection of purified β subunits into Xenopus oocytes can modulate calcium currents within 1 h (Yamaguchi et al. 1998). Furthermore, a peptide corresponding to the β-binding site in the I-II loop of the α₁ subunit can can modify single channel gating (Hohaus et al. 2000), possibly by causing β to dissociate from α₁.

In a paper in this issue of The Journal of PhysiologyGarcía et al. (2002) now show that acute application of purified β subunits affects currents through calcium channels, using whole cell dialysis of membrane vesicles from adult skeletal muscle. Over ≈20 min (the expected time for delivery of a ≈60 kDa protein from a patch pipette), the current doubled in amplitude. There was no obvious effect on activation kinetics, but a ≈16 mV shift of inactivation towards more negative voltages and an increased slow tail current were seen. Three lines of evidence suggest that the β subunits were modulating preexisting α₁ subunits, rather than enhancing membrane expression of α₁: (1) the time course was too fast for expression of new α₁ subunits; (2) the vesicles appeared to have no sarcoplasmic reticulum or other organelles; and (3) the gating currents did not increase, implying no change in the number of α₁ proteins that could move their voltage sensors. One implication of this study is that α₁ subunits are not normally fully saturated with β.

It seems that β subunits act (at least in part) by rapidly reversibly binding, in stark contrast to the traditional picture of an invariant stoichiometric subunit. Could this also be true for the other calcium channel ‘subunits’, notably the enigmatic γ?