Membrane-localized β-subunits alter the PIP₂ regulation of high-voltage activated Ca²⁺ channels

Abstract

The β-subunits of voltage-gated Ca²⁺ (Ca_V) channels regulate the functional expression and several biophysical properties of high-voltage–activated Ca_V channels. We find that Ca_V β-subunits also determine channel regulation by the membrane phospholipid phosphatidylinositol 4,5-bisphosphate (PIP₂). When Ca_V1.3, -2.1, or -2.2 channels are cotransfected with the β3-subunit, a cytosolic protein, they can be inhibited by activating a voltage-sensitive lipid phosphatase to deplete PIP₂. When these channels are coexpressed with a β2a-subunit, a palmitoylated peripheral membrane protein, the inhibition is much smaller. PIP₂ sensitivity could be increased by disabling the two palmitoylation sites in the β2a-subunit. To further test effects of membrane targeting of Ca_V β-subunits on PIP₂ regulation, the N terminus of Lyn was ligated onto the cytosolic β3-subunit to confer lipidation. This chimera, like the Ca_V β2a-subunit, displayed plasma membrane localization, slowed the inactivation of Ca_V2.2 channels, and increased the current density. In addition, the Lyn-β3 subunit significantly decreased Ca_V channel inhibition by PIP₂ depletion. Evidently lipidation and membrane anchoring of Ca_V β-subunits compete with the PIP₂ regulation of high-voltage–activated Ca_V channels. Compared with expression with Ca_V β3-subunits alone, inhibition of Ca_V2.2 channels by PIP₂ depletion could be significantly attenuated when β2a was coexpressed with β3. Our data suggest that the Ca_V currents in neurons would be regulated by membrane PIP₂ to a degree that depends on their endogenous β-subunit combinations.

Voltage-gated Ca²⁺ (Ca_V) channels define excitable cells. These channels contribute to electrical excitability and Ca²⁺-dependent processes, such as neurotransmitter release, motility, and gene expression (1). An intriguing aspect of Ca_V channels is their slow, voltage-independent modulation by G_q protein-coupled receptors (G_qPCRs) (2–5). The several proposed mechanisms for this slow modulation of Ca_V channels include depletion of phosphatidylinositol 4,5-bisphosphate (PIP₂) via activation of phospholipase C (3–10), phosphorylation by protein kinases (11, 12), generation of arachidonic acid by phospholipase A₂ (13, 14), and other unspecified pathways involving Gα_q or Gβγ-subunits (15, 16). In the nervous system, the dissection of G_qPCR modulation of Ca²⁺ channels is made difficult by the generation of several second messengers and by the existence of multiple subtypes of Ca_V channels.

Ca_V channels are a complex of three protein subunits. The α1-subunit forms the voltage-sensitive, Ca²⁺-selective pore and is the target of selective antagonists. The other auxiliary subunits shape the physiological properties of the α1-subunit. Here we emphasize β-subunits. These subunits associate tightly with the cytoplasmic domain of the α1-subunit, imparting increased peak currents, altered rates of activation and inactivation, and increased density of functional channels (17–19). The β-subunits have four known genes (β1–β4) and several splice variants. Single neurons express multiple types in a tissue-specific manner (20, 21). Accordingly, different combinations of α1- and β-isoforms result in heterogeneous Ca_V channels susceptible to differential regulation.

Some Ca_V β-subunits are subject to posttranslational modification. The β2a-subunit is doubly palmitoylated (22), which targets it to the plasma membrane even in the absence of a coexpressed α1-subunit (23, 24). Palmitoylation confers on nonmembrane proteins membrane-targeting to specific areas, such as to caveolae for endothelial nitric oxide synthase (25) or the plasma membrane for src-related tyrosine kinases (26). In heterologous expression systems, the β2a-subunit slows Ca_V current inactivation significantly more than any other β-subunit isoform (27, 28). Studies with chimeric β-subunits demonstrated that cysteine residues in positions 3 and 4 of the N terminus are the sites for palmitoylation and are determinants of Ca_V channel inactivation (24, 27).

Membrane lipids are implicated in voltage-independent regulation of Ca_V channels (see reviews in refs. 4, 9, and 29). Wu et al. (7) and Gamper et al. (8) reported that membrane PIP₂ dramatically regulates Ca_V2.1 P/Q-type and Ca_V2.2 N-type channel activity. The authors found spontaneous rundown of current was opposed or reversed by the application of PIP₂ or Mg-ATP and accelerated by PIP₂ antibody applied to the cytoplasmic side of excised membrane patches. The authors proposed that depletion of PIP₂ on the plasma membrane is the cause of the G_q receptor-mediated slow inhibition of Ca_V currents in neurons. In contrast, Rittenhouse and colleagues (13, 14) reported that arachidonic acid mediates the muscarinic inhibition and an associated enhancement of N-type Ca²⁺ current in rat sympathetic neurons. The muscarinic modulation was attenuated when arachidonic acid production by PLA₂ was blocked. The authors also showed that the Ca_V β-subunit is important in determining the channel modulation by G_qPCRs (21). In their hypothesis, the palmitoyl chains of the Ca_V β2a-subunit interact directly with Ca_V2.2 channels and competitively block receptor-stimulated current suppression by arachidonic acid (30).

In recent studies, we characterized the ability of a voltage-sensing lipid phosphatase from zebrafish (Dr-VSP) (31–33) to deplete membrane PIP₂ (34). Activation of Dr-VSP using a +120 mV-depolarizing pulse depletes plasma membrane PI(4,5)P₂ within a second by removing the 5-phosphate and leaving PI(4)P (10). This tool is useful for experimental protocols that require reversible PIP₂ depletion. In cells transfected with various α1-subunits coexpressed with β3- and α2δ1-subunits, we demonstrated that Ca_V1.2, -1.3, -2.1, and -2.2 channels are partially inhibited when membrane PIP₂ is depleted. We report here that (i) the β-subunit affects sensitivity of Ca_V channels to PIP₂ and (ii) it is the palmitoylation of rat β2a that reduces the functional effects of PIP₂ in Ca_V channels. Our data are consistent with previous studies showing that, of several β-subunits tested (β1b, β2a, and β3), β2a is the most effective at suppressing G_q protein-mediated Ca²⁺ channel inhibition in neurons (35).

Results

PIP₂ Regulation of Ca_V Currents Depends on the Ca_V β-Subunit.

In our previous study, we found that four Ca_V α1-subunits coexpressed with β3- and α2δ1-subunits gave currents sensitive to PIP₂ depletion (10). In this study we show that the extent of PIP₂ modulation depends on which β-subunit is used. Our standardized protocol involved a test pulse (pulse a) to record baseline channel current, then a 1-s depolarization to +120 mV to activate Dr-VSP and deplete PIP₂, next a hyperpolarizing pulse for 0.4 s to remove any Ca²⁺ channel inactivation, followed finally by a second test pulse (pulse b) (Fig. 1A). We compared the a and b currents. Without the large depolarizing pulse to +120 mV, there was no significant difference between currents a and b (10). Similarly in control cells not expressing Dr-VSP, the Ca_V2.2 current amplitudes were almost the same before (trace a) and after (trace b) the 1-s depolarization in cells cotransfected with either β3- or β2a-subunits (Fig. 1A, Top, and B). In contrast, with expression of Dr-VSP, the 1-s depolarizing pulse significantly attenuated the current in pulse b in cells with β3-subunits but hardly had any effect in cells with β2a-subunits (Fig. 1A, Bottom, and B). The prominent tail currents of Ca_V2.2 channels were inhibited to the same degree as the peak inward currents during test pulses (Fig. 1A). Dr-VSP activation did not change the activation or deactivation kinetics of expressed currents. Our findings with PIP₂ modulation are reminiscent of those of others that coexpression of β2a-subunits greatly attenuates M₁ muscarinic inhibition of Ca_V currents (21).

Fig. 1.

Ca_V β-subunits determine Ca_V2.2 channel sensitivity to PIP₂ depletion. Whole-cell Ba²⁺ current in tsA-201 cells. (A) Cells transfected with Ca_V α1B (Ca_V2.2), α2δ1, and β3 (Left) or β2a (Right) subunits received a test pulse (a), then a depolarization to +120 mV for 1 s, a hyperpolarization to −150 mV for 0.4 s, and a second test pulse (b). The currents before (a) and after (b) the depolarizing pulse are presented in control (Upper) and Dr-VSP–expressing (Lower) cells. Peak tail current is indicated by arrowheads (trace a, open head; trace b, closed head). (B) Summary of Ca_V2.2 current inhibition (%) by the strong depolarizing pulse in control and Dr-VSP–expressing cells cotransfected with Ca_V β3- or β2a-subunits. Control, n = 5 for β3 and for β2a; Dr-VSP, n = 11 for b3 and n = 15 for β2a. (C) Ca_V2.2 current modulation by M₂ muscarinic receptor activation with 10 μM Oxo-M in cells expressing β3 (Upper) or β2a (Lower). (Insets) Current traces before (a) and during (b) Oxo-M application. (D) Summary of M₂ muscarinic suppression of Ca_V2.2 currents with β3 (n = 6) or β2a (n = 5).

In addition to slow G_q-coupled modulation by the M₁ muscarinic receptor, Ca_V2.2 channels are subject to a fast G_i/o-coupled modulation by M₂ muscarinic receptors (36). The Gβγ-subunits released from G_i/o inhibit Ca_V2.2 channels by binding directly to the α1-channel subunit (37, 38). We tested the effects of Ca_V β-subunits on modulation of Ca_V2.2 channels by M₂ muscarinic receptors. Fig. 1C shows that the current modulation by M₂ receptor was not significantly different between channels with β3- and β2a-subunits. We also considered whether the N-terminal Gβγ binding site on the Ca_V2.2 α1-subunit governs PIP₂ regulation. For the test we took advantage of the Gβγ-insensitivity of a chimeric channel construct α1C-1B (10, 39). In this chimera, the N terminus of Ca_V2.2 (N-type, α1B-subunit), which includes one of the Gβγ binding sites, is replaced by the N terminus of Ca_V1.2 (L-type, α1C-subunit). As shown in Fig. S1, PIP₂ regulation of chimeric Ca_V2.2 channels was still dramatically reduced in cells expressing the Ca_V β2a-subunit. Thus, the β-subunit–specific differential regulation of Ca_V channels by PIP₂ depletion does not generalize to regulation by Gβγ-subunits and does not depend on the N terminus of Ca_V2.2 channels where there is a Gβγ binding site.

We next tested whether regulation by PIP₂ of two other Ca_V channels (1.3 and 2.1) depends on the expressed β-subunit. For both channels, activation of Dr-VSP inhibited current much better in the presence of Ca_V β3-subunits than in the presence of Ca_V β2a-subunits (Fig. 2), echoing the findings for Ca_V 2.2 channels in Fig. 1.

Fig. 2.

Ca_V β-subunits set sensitivity of Ca_V2.1 and Ca_V1.3 channels to Dr-VSP. (A) Differential regulation of Ca_V2.1 channels by Dr-VSP. Typical Ca_V2.1 current with β3 (Left) or β2a (Right) before (a) and after (b) Dr-VSP activation in control (Upper) and Dr-VSP–expressing (Lower) cells. Arrowheads are peak tail currents. (B) Summary of the percent inhibition of Ca_V2.1 currents by Dr-VSP activation. For control, β3 (n = 7), β2a (n = 6); for Dr-VSP, β3 (n = 6), β2a (n = 6). (C) Inhibition of L–type Ca_V1.3 currents by PIP₂ depletion. (Left) Superimposed currents before (a) and after (b) the depolarizing pulse. (Right) Summary of the current inhibition by the +120-mV depolarizing pulse with β3 (n = 3) or β2a (n = 4).

Palmitoylation Determines PIP₂ Modulation of Ca_V2.2 Channels.

We considered the effect of β-subunits on Ca_V2.2 channels in more depth. The Ca_V β-subunits are key determinants of channel inactivation gating kinetics. Fig. 3A shows representative Ca_V2.2 currents normalized to peak current in cells with β2a, β2b, or β3-subunits. The current inactivation in cells transfected with β2a was slow, whereas for cells transfected with β3- and β2b-subunits it was fast (Fig. 3 A and B). These data are consistent with the previously reported ability of β2a to slow voltage-gated inactivation of Ca_V2.3 (α1E) (27). We then tested whether the palmitoylation of the β2a-subunit affects modulation by PIP₂ and is required for slowing inactivation. We used the palmitoylation-resistant mutant β2a(C3S,C4S), where both palmitoylated cysteine residues are replaced with serine (28). Prevention of lipidation changed the effects of the β2a-subunit so they were much closer to those of β2b- or β3-subunits in three ways. With the β2a(C3S,C4S) subunit: (i) Current inactivation was speeded to resemble β2b- or β3-subunits (Fig. 3B); (ii) the current density fell from the high value seen with β2a to the lower value seen with β2b and β3 (Fig. S2A); and (iii) the inhibition by PIP₂ depletion increased (Fig. 3C). Current inhibition with Dr-VSP was 9 ± 2% (n = 6) with β2a, 41 ± 3% (n = 6) with β2b, and 27 ± 4% (n = 8) with β2a(C3S,C4S). We also found that, with the expression of another nonpalmitoylated subunit, β1b, the activation of Dr-VSP significantly attenuated the current (36 ± 3%, n = 4) (Fig. S3).

Fig. 3.

Palmitoylation of Ca_V β-subunit attenuates inactivation gating and inhibition by PIP₂ depletion. (A) Differential effects of Ca_V β-subunits on the inactivation of Ca_V2.2 currents. Currents with different Ca_V β-subunits were measured during 500-ms test pulses to +10 mV or +30 mV in control cells. Overlaid current traces with different Ca_V β-subunits are scaled to the peak amplitude (I₀). Dashed line indicates zero current. (B) Ca_V2.2 current inactivation during the 500-ms depolarizing pulse calculated as final current amplitude (I) divided by initial peak (I₀) in cells expressing different β-subunits. Currents measured at +10 mV or +30 mV in control cells with no Dr-VSP. (C) Inhibition of Ca_V2.2 currents by Dr-VSP with various β-subunits. Currents in test pulses to +10 mV for 10 ms before (a) and after (b) the depolarizing pulse are superimposed. (D) Summary of the Ca_V2.2 current inhibition (%) by depolarization in control and Dr-VSP-expressing cells transfected with different Ca_V β-subunits; for control, β2a (n = 4), β2a(C3S,C4S) (n = 4), or β2b (n = 4); for Dr-VSP, β2a (n = 6), β2a(C3S,C4S) (n = 8), or β2b (n = 6). *P < 0.01, compared with β2a.

Targeting of the β3-Subunit to the Membrane.

The β3-subunits are soluble cytosolic proteins that bind to α1-subunits of Ca²⁺ channels. We examined the effects of anchoring β3-subunits to the membrane by adding an N-terminal 13 residue sequence of Lyn to the N terminus of β3. This chimera starts with GCIKS, where the G and C become myristoylated and palmitoylated, respectively. Confocal imaging shows that wild-type β3-subunits are distributed throughout the cytoplasm in the presence or absence of other subunits of Ca_V2.2 channels (Fig. 4A, Left). In contrast, with the added N terminus of Lyn, chimeric β3-proteins were mostly associated with the plasma membrane, even without expression of α1-subunits (Fig. 4A, Right, and Fig. S4A). The distribution of Lyn-β3-YFP was comparable that of the PIP₂ binding protein CFP-PH domain, which targets the cytoplasmic leaflet of the plasma membrane (Fig. S4B). Addition of lipidation sites changed the effects of the β3-subunit so they were more like those of palmitoylated β2a-subunits in three ways. With the Lyn-β3 subunit: (i) inactivation of Ca_V2.2 currents was slowed approximately twofold (Fig. 4C) compared with currents with the nonlipidated wild-type β3-subunit (Fig. 4B); (ii) the current density of the Ca_V2.2 channels was significantly enhanced (Fig. 4D); and (iii) inhibition by PIP₂ was blunted compared with cells with β3 (Fig. 4E). With the β3-subunits, currents were inhibited by 52 ± 2% (n = 6), whereas with Lyn-β3, currents were inhibited by 24 ± 3% (n = 6) (Fig. 4F).

Fig. 4.

Membrane targeting of Ca_V β3-subunit attenuates current inactivation and Dr-VSP–induced inhibition. (A) Confocal images of cells expressing cytosolic β3-YFP or membrane-targeted Lyn-β3-YFP without (Upper) or with (Lower) other Ca_V2.2 channel subunits. (Scale bars, 10 μm.) (B) Effect of membrane-targeted Ca_V Lyn-β3 on the inactivation of Ca_V2.2 currents. Currents were measured during 500-ms test pulses to +10 mV or +30 mV in control cells. Blue lines are current traces with Lyn-β3-YFP scaled to the peak amplitude of currents with β3-YFP. (C) Summary of the time constants for current inactivation. Data are mean ± SEM. (D) Ca_V2.2 current density (pA/pF) measured in cells expressing Ca_V2.2 channels with β3, β3-YFP, or Lyn-β3-YFP. Cells were transfected with the same amounts of cDNA. Average membrane capacitances for cells are 20 ± 3 pF for β3 (n = 5), 24 ± 3 for β3-YFP (n = 6), and 25 ± 7 for Lyn-β3-YFP (n = 6). (E) Currents in Ca_V2.2 channels with β3-YFP (Left) or Lyn-β3-YFP (Right) before (a) and after (b) a strong depolarizing pulse in Dr-VSP-expressing cells. The peak tail currents are indicated by arrowheads. (F) Summary of inhibition by PIP₂ depletion in cells with β3-YFP or Lyn-β3-YFP. Data are mean ± SEM. *P < 0.01, compared with β3-YFP.

β-Subunits Can Grade the Modulation of Ca_V Channels.

In physiological conditions, a single excitable cell expresses multiple types of Ca_V β-subunits and other subunits (20, 21). To mimic this situation, tsA cells transfected with both β2a- plus β3-subunits were compared with cells transfected with only one of them. The Ca_V2.2 current inactivation was measured at +10 and +30 mV. Fig. 5A shows that current inactivation in cells expressing both β-subunits looked like a mixture of some channels complexed with β3-subunits and a second population complexed with β2a-subunits. On average in cells transfected with equal quantities of β2a- and β3-subunit cDNAs, a little more than half of the current showed β2a inactivation (I/I₀) kinetics (Fig. 5A, Lower). This overrepresentation of β2a-like actions might be expected if β2a-subunits are more efficient at bringing functional channels to the surface. Next, the PIP₂ sensitivity of the currents was measured. Considering again that cells expressing both the β2a- and β3-subunits have a mixture of two channel complexes, the Dr-VSP–mediated current inhibition suggested that a little more than half the current comes from channels complexed with β2a-subunits that reduce their sensitivity to PIP₂ depletion (Fig. 5 B and C).

Fig. 5.

Ca_V β-subunits determine channel modulation by receptors. (A) Coexpression of Ca_V β2a with β3-subunits slows the inactivation of Ca_V2.2 currents. Current inactivation was measured during a 500-ms test pulse in cells expressing β3 (0.5 μg), or β2a (0.5 μg), or β3 (0.25 μg) plus β2a (0.25 μg). Current traces with β3 (blue) or β3 plus β2a (red) are scaled to the peak amplitude of current with β2a (black). (Lower) Summary of current inactivation [final current amplitude (I) divided by the initial current (I₀)] at +10 mV (n = 7) and +30 mV (n = 6). (B) Current inhibition by PIP₂ depletion in cells coexpressing β2a- and β3-subunits. Cells received a test pulse (a), then were depolarized to +120 mV for 1 s and hyperpolarized to −150 mV for 0.4 s, followed by the second test pulse (b). Currents before (a) and after (b) the depolarizing pulse are superimposed. Peak tail currents are indicated by arrowheads. (C) Summary of the Ca_V2.2 current inhibition (%) by depolarization in Dr-VSP–expressing cells transfected with various Ca_V β-subunits. (D) Current inhibition by M₁ muscarinic receptor activation with Oxo-M (10 μM) in cells expressing β3 or β2a. (Insets) Current traces for before (a) and during (b) Oxo-M application are overlaid. (E) Summary of the M₁ muscarinic suppression of the Ca_V2.2 currents in cells expressing various β-subunits. (n = 5 for control, n = 7 for β2a).

Finally, the effect of β-subunits on slow muscarinic modulation of Ca_V2.2 channels was tested in cells expressing M₁Rs. In cells expressing β3-subunits, stimulation with Oxo-M strongly decreased the Ca_V2.2 current (Fig. 5 D and E). The effect was somewhat weaker in cells with β2a-subunits and possibly intermediate with β2a(C3S,C4S), β2b, or a mixture of β2a and β3. The muscarinic modulation was stronger than the PIP₂ modulation, and the effects of changing subunits were weaker on muscarinic compared with PIP₂ modulation.

Discussion

The Ca_V β2a isoform is a β2 splice variant that differs from β2b only in the N terminus where it contains the determinants of palmitoylation at Cys3 and 4. Among Ca_V β-subunits, β2a is functionally unique. We have extended early findings on its ability, in comparison with the other β-subunits: (i) to localize to the plasma membrane by itself (23, 24), (ii) to enhance functional Ca_V channel expression (22, 40), (iii) to slow and minimize inactivation gating during membrane depolarization (27, 28), and (iv) to reduce slow suppression of current by G_qPCR signaling (21, 30). Another unique property is that (v) the β2a-subunit prevents inhibition of Ca_V currents by direct application of arachidonic acid (21). All of the well-documented properties listed above depend on palmitoylation of the β-subunit. When the palmitoylation sites are removed or the palmitoylation reaction is blocked, the β2a-subunit loses these unique properties and behaves more like the others. We have added another unique property: the β2a-subunit (vi) greatly reduces the sensitivity of several Ca_V channels to PIP₂ depletion. This property too is lost when the palmitoylation sites are defective and can be conferred on another β-subunit simply by appending a lipidation motif. In our experiments with the β3-subunit, N-type channels (Ca_V2.2) are the most sensitive to PIP₂ depletion (∼50–60%) and other channels, such as L-type (Ca_V1.3) and P/Q-type (Ca_V2.1), are inhibited by 20–30%. However, with β2a-subunits, all of the types of Ca_V channels are inhibited only <10–15% by PIP₂ depletion. Thus, the PIP₂ regulation of Ca_V channels is dependent on both the α1- (10) and β-subunits.

What is the situation in a really excitable cell that expresses several subtypes of Ca_V β-subunit? The channel properties will represent a mixture of channel complexes. Different cells could tune their Ca_V functional properties and expression by adjusting their mix of β-subunits and the degree of palmitoylation, which would affect the amplitude, inactivation behavior, and modes of modulation of channel current. Apparently some β-subunits can be exchanged from the channel complex within a few minutes (41). Previous studies reported that superior cervical ganglion neurons express multiple types of β-subunit with β2a and β3 being the major isoforms (21). In the superior cervical ganglion, channel inactivation is very slow compared with that in an expression system with β3 subunits (21, 42). That finding might suggest that most of the endogenous Ca_V2.2 channels are complexed with β2a-subunits. However, that conclusion would be inconsistent with many findings that the native channels are relatively sensitive to modulation by G_qPCRs, PIP₂ depletion, and arachidonic acid, as if complexed with β3-subunits. Future work will need to examine the possible involvement of β3-subunits in a more quantitative way. Bovine chromaffin cells express several subtypes of β-subunits including β2a and β3 and their N-type currents are almost noninactivating as in the superior cervical ganglion (43). The currents are also significantly inhibited by GPCRs through voltage-independent pathways (44). In contrast, mouse aorta expresses both β2a- and β3-subunits, and the calcium current is strongly inactivating. In the aorta of β3-null mice; however, the L-type current inactivation became slower by ∼twofold (45). The β2a-subunit is broadly expressed in brain, heart, and aorta, but its expression is lower than other β-subunits (46, 47), with β3 being predominant and the most abundant β-subunit in neuronal N-type channel complexes (47, 48).

Palmitoylation adds a 16-carbon fatty acid chain covalently to cysteine residues (49). The reversibility of palmitoylation introduces the possibility that the constellation of effects of β2a is under dynamic control through physiological signals. Indeed, in the experiments of Hurley et al. (28) it was suggested from observing different current time courses that some individual cells had more palmitoylation on their β2a-subunits than others. In some manner the palmitoylation confers numerous special functional properties on the β-subunit. The addition of hydrophobic palmitates should anchor the N terminus of the β2a-subunit more tightly to the lipid bilayer, increasing the probability that channels would encounter and complex with a β2a-subunit compared with a nonlipidated soluble β3-subunit. Partnering with an anchored subunit may also favor a special stabilizing conformational transition of the channel that affects its probability of being open.

As suggested extensively by the Rittenhouse laboratory (4, 21, 30), the β2a-subunit palmitates may be ligands that interact specifically with the channel as well. In their view, which is plausible, the closely tethered palmitoyl chains interact with a binding site on the channel that can otherwise interact with other fatty acid tails, including the arachidonic acid that is essential in their favored model and those of the PIP₂ that is essential in our model. If this competitive concept is correct, then an active form of the channel is stabilized by β-subunits with the palmitate or by PIP₂, and an inactive form is stabilized by the arachidonate. The two active forms would have somewhat different properties. The palmitate-occupied form would be noninactivating or poorly inactivating, whereas the PIP₂-occupied form would be a channel that can inactivate “normally.” It is tempting to think that modulation by G_qPCRs, and by arachidonic acid addition, and by PIP₂ depletion show the same sensitivity to the specific β-subunit expressed because they all share this same competitive interaction at a hydrophobic site on the channel.

We had tended to emphasize the PIP₂ headgroup and to de-emphasize the fatty-chain hypothesis both because in many channels the function of long-chain native PIP₂ can be mimicked by the soluble, short-chain version diC8-PIP₂ (suggesting little lipid specificity) and because, as our VSP experiments confirm, PI(4)P cannot substitute for PI(4,5)P₂ in maintaining channel function. A recent crystal structure of an inward rectifier Kir2.2 channel with phosphoinositides and our finding that channels with lipidated β2a subunits do not need PIP₂, strengthen the fatty chain hypothesis (50). The cytoplasmic domain of Kir2.2 has a stereo-specific binding site for the inositol phosphate ring, and the transmembrane domain has a nonspecific phospholipid binding site that will interact even with the short-chains on diC8-PIP₂ and with other phospholipids. Binding of diC8-PIP₂ across both domains draws them together, stabilizing an active open state. Thus, by analogy there may be two regions on the channels we have studied that are drawn together by the bidentate PIP₂ but not by binding of arachidonic acid, which lacks an interacting headgroup. Nonlipidated β-subunits that act only on cytoplasmic domains of the channel would not interfere with either site, but lipidated ones would be bidentate. These subunits might compete for the phospholipid site by displacing PIP₂ or arachidonic acid and thus “cross-link” the transmembrane domain to their normal binding site on the cytoplasmic domain. Virtually the same concept was advanced for Ca_V channels by Roberts-Crowley et al. (4) who proposed that “the phospholipid acts as Velcro between two regions.”

In short, there seem to be competitive interactions of fatty acid tails on a transmembrane part of the channel, and, for activating ligands, there may be simultaneous binding to a second cytoplasmic domain of the channel complex that physically locks this cytoplasmic domain to the transmembrane component in an activating position. Thus, PIP₂ and lipidated β subunits would be adaptor molecules that activate channels by bridging two domains.

Materials and Methods

All methods and materials including the clones are presented in the SI Materials and Methods. In brief, all experiments used cultured HEK tsA-201 cells transiently transfected with various cDNAs, plated onto poly-l-lysine-coated coverslip chips, and studied within 1–2 d. For Ca²⁺ channel expression, cells were transfected with the Ca_V α1-subunit, β-subunit, and α2δ1 in a 1:1:1 molar ratio. Whole-cell currents were measured at room temperature (22–25 °C) using a HEKA EPC-9 amplifier (HEKA Elektronik). Quantitative data are expressed as the mean ± SEM, and comparison between two groups used Student t test. Ba²⁺ currents were recorded during test pulses to –10 mV, 0 mV, or +10 mV to measure Ca_V1.3, Ca_V2.1, or Ca_V2.2 currents, respectively. The external Ringer's solution for electrophysiology contained: 150 mM NaCl, 10 mM BaCl₂, 1 mM MgCl₂, 10 mM Hepes, and 8 mM glucose, adjusted to pH 7.4 with NaOH. The pipette solution contained: 175 mM CsCl, 5 mM MgCl₂, 5 mM Hepes, 0.1 mM 1,2-bis(2-aminophenoxy)ethane N,N,N',N'-tetraacetic acid (BAPTA), 3 mM Na₂ATP, and 0.1 mM Na₃GTP, titrated to pH 7.4 with CsOH. For confocal imaging, tsA or HEK293 cells were imaged 24–48 h after transfection on poly-l-lysine coated coverslips. Primers are listed in Table S1.