Molecular determinants of inactivation and G protein modulation in the intracellular loop connecting domains I and II of the calcium channel α_1A subunit

Abstract

Synaptic transmission is regulated by G protein-coupled receptors whose activation releases G protein βγ subunits that modulate presynaptic Ca²⁺ channels. The sequence motif QXXER has been proposed to be involved in the interaction between G protein βγ subunits and target proteins including adenylyl cyclase 2. This motif is present in the intracellular loop connecting domains I and II (L_I-II) of Ca²⁺ channel α_1A subunits, which are modulated by G proteins, but not in α_1C subunits, which are not modulated. Peptides containing the QXXER motif from adenylate cyclase 2 or from α_1A block G protein modulation but a mutant peptide containing the sequence AXXAA does not, suggesting that the QXXER-containing peptide from α_1A can competitively inhibit Gβγ modulation. Conversion of the R in the QQIER sequence of α_1A to E as in α_1C slows channel inactivation and shifts the voltage dependence of steady-state inactivation to more positive membrane potentials. Conversion of the final E in the QQLEE sequence of α_1C to R has opposite effects on voltage-dependent inactivation, although the changes are not as large as those for α_1A. Mutation of the QQIER sequence in α_1A to QQIEE enhanced G protein modulation, and mutation to QQLEE as in α_1C greatly reduced G protein modulation and increased the rate of reversal of G protein effects. These results indicate that the QXXER motif in L_I-II is an important determinant of both voltage-dependent inactivation and G protein modulation, and that the amino acid in the third position of this motif has an unexpectedly large influence on modulation by Gβγ. Overlap of this motif with the consensus sequence for binding of Ca²⁺ channel β subunits suggests that this region of L_I-II is important for three different modulatory influences on Ca²⁺ channel activity.

Neuronal voltage-gated Ca²⁺ channels are involved in multiple cellular functions including neurotranstransmitter release, Ca²⁺-mediated regulatory processes, and generation of dendritic action potentials. They consist of complexes of a pore-forming α₁ subunit of 190–250 kDa in association with α₂δ and β subunits (1, 2). Electrophysiological and pharmacological studies distinguish at least six classes of Ca²⁺ channel currents designated L-, N-, P-, Q-, R-, and T-type (2, 3). Ca²⁺ channels containing α_1C or α_1D subunits are thought to be responsible for L-type currents, α_1B for N-type currents, and α_1A for both P- and Q-type calcium currents (2, 4). A major goal of current research is to correlate the observed differences among the properties of these channel types with the structures of their α₁ subunits.

P/Q-type and N-type Ca²⁺ channels containing α_1A or α_1B subunits differ from L-type Ca²⁺ channels containing α_1C or α_1D subunits in voltage-dependent inactivation (5–8) and in modulation by G proteins (9). Voltage-dependent inactivation of L-type Ca²⁺ channels containing cloned α_1C is slower than inactivation of P/Q-type Ca²⁺ channels containing cloned α_1A expressed with the same auxiliary subunits in Xenopus oocytes (5, 6). In addition, N-type and P/Q-type Ca²⁺ channels, but not L-type Ca²⁺ channels, are modulated by neurotransmitter receptors acting through pertussis toxin-sensitive G proteins via membrane-delimited pathways that cause a positive shift in the voltage dependence of channel activation which is reversed by strong depolarization (9–12). This modulatory effect is mediated by G protein βγ subunits (13, 14), possibly through direct binding to the Ca²⁺ channel.

A consensus sequence (QXXER) has been proposed to be involved in the interaction of Gβγ subunits with adenylyl cyclase type 2, inwardly rectifying K⁺ channels, and phospholipase Cβ (15). A related consensus sequence is within or adjacent to the region of the COOH terminus of inwardly rectifying K⁺ channels which is implicated in binding and regulation by Gβγ subunits (16, 17). It is also found in the intracellular loops connecting domains I and II of α_1A and α_1B, which are modulated by Gβγ, but not in the corresponding segment of α_1C, which is not modulated by this pathway (18–20). Thus, this consensus sequence could be involved in G protein modulation of Ca²⁺ channels. In these experiments, we probed the role of this consensus sequence in Ca²⁺ channel function and modulation with synthetic peptides containing this sequence and with mutations of this motif in Ca²⁺ channel α₁ subunits. The results point to an important role for this region of the Ca²⁺ channel in both voltage-dependent inactivation and G protein modulation.

EXPERIMENTAL PROCEDURES

cDNAs encoding Ca²⁺ channel subunits α_1A [rbA isoform, (19)] and β_1b were cloned in pMT2XS (8), α_1C [rbC-II isoform (18)] in pZem229, α₂δ (21) in pZEM228, and CD8 in EBP-pcD. tsA-201 cells were transfected with the α₁, α₂δ, and β cDNAs in 1:1:1 molar ratios plus CD8 in either calcium phosphate or lipofectamine (Stratagene) and incubated for at least 48 h. Positively transfected cells were identified by labeling with anti-CD8 antibody tagged with fluorophore and analyzed by whole-cell patch clamp as described (14, 22). Currents were recorded and filtered at 10 kHz (α_1A) or 4 kHz (α_1C) with an eight-pole Bessel filter. Leak and capacitative currents were measured using hyperpolarizing pulses and subtracted with the p/4 method. Cells were bathed in an external solution containing 100 mM Tris, 4 mM MgCl₂, and 10 mM BaCl₂ with pH adjusted to 7.3 with methanesulfonic acid. The internal pipette solution consisted of 120 mM aspartic acid, 5 mM CaCl₂, 2 mM MgCl₂, 10 mM Hepes, 10 mM EGTA, and 2 mM Mg·ATP with pH adjusted to 7.3 with CsOH. When indicated in the figure legends, guanosine 5′-[γ-thio]triphosphate (GTP[γS]) was added to the internal solution at a concentration of 0.6 mM.

The mutants in the α_1A subunit were constructed using the method of Ho et al. (23). A XhoI/KpnI restriction fragment containing the 5′ region of the α_1A cDNA was subcloned into pBluescript II SK(+) (Stratagene) and used for PCR with Pfu polymerase (Stratagene). Amplified mutant cDNA was subcloned into the BspHI and StuI sites of the XhoI/KpnI construct. This construct was subcloned into full-length α_1A in pMT2XS and used for transfection. The mutations in α_1C were constructed using a single reaction PCR strategy with Pfu polymerase and the rbC-II cDNA in the expression vector pZem229 as template. The PCR products were subcloned into rbC-II in pZem229 in which two BglII sites had been previously destroyed using NgoMI and BglII. Mutations were verified by cDNA sequencing.

RESULTS

Effects of Peptide Inhibitors on G Protein Modulation.

P/Q-type Ca²⁺ channels composed of α_1A, α₂δ, and β_1b were expressed in the human embryonic kidney cell line tsA-201, a derivative of HEK293 (22). Ba²⁺ currents through Ca²⁺ channels were measured in the whole-cell voltage-clamp configuration. In the presence of GTP[γS] in the intracellular solution, the voltage dependence of activation was shifted toward positive membrane potentials and was less steep than in control intracellular solution (Fig. 1A, compare ▵ and ○), as described (14). This shift reflects modulation of the Ca²⁺ channel by Gβγ (14). Strong depolarizing prepulses increase Ba²⁺ currents and shift the activation curve to the left (Fig. 1A) by reversing the voltage-sensitive modulation (14) (Fig. 1A, •). This facilitation of the Ba²⁺ currents and negative shift of the voltage dependence of activation by depolarizing prepulses provides an index of G protein modulation. This experimental protocol was used to test the ability of synthetic peptides from adenylyl cyclase type 2 (AC2) and from the α_1A subunits of Ca²⁺ channels (rbA) containing the putative QXXER G protein binding sequence (Scheme I) to inhibit G protein modulation competitively following diffusion into the cell from the recording pipette.

Figure 1.

Effects of synthetic peptide inhibitors on G protein modulation of P/Q-type Ca²⁺ channels. (A) Effect of GTP[γS] on the facilitation of Ba²⁺ currents of the P/Q-type channel. Tail currents with Ba²⁺ as current carrier were recorded with (•, ○) or without (▵) GTP[γS] in the intracellular solution. A 4-ms test pulse (test 1) to the indicated test potential was applied from the holding potential of −60 mV. After 1 s, a 10-ms conditioning prepulse to +100 mV was applied, the cell was repolarized to −60 mV for 10 ms, and a second 4-ms test pulse (test 2) to the same potential as test pulse 1 was applied. Tail currents were normalized to the largest tail current in each series of the test pulses, and means ± SEM were plotted against test pulse potential. An increase in tail current after test 2 in comparison to test 1 indicates facilitation. Control intracellular solution, test 1▵; ○, GTP[γS], test 1 •; GTP[γS], test 2. The current traces shown in each panel were recorded during and after test 1 and test 2 at +25 mV. (B–D) Effects of synthetic peptide inhibitors on the facilitation of tail currents of the P/Q-type channel. Peptides (see text for sequence) were applied intracellularly at 100 μM concentration. Tail currents were recorded and analyzed as described in A. (B) AC2 peptide. (C) AC2/AXXAA peptide. (D) rbA peptide.

G protein modulation was blocked by a peptide from adenylyl cyclase type 2 (AC2) which contains the QXXER sequence (Fig. 1B), but not by an identical peptide in which the sequence was changed to AXXAA (Fig. 1C). These results suggest that the AC2 peptide can bind to Gβγ subunits and prevent their interaction with Ca²⁺ channels. An analogous peptide containing the QXXER sequence from L_I-II of α_1A also blocks G protein modulation (Fig. 1D). These results suggest a role for the QXXER sequence of α_1A in modulation of Ca²⁺ channel gating by Gβγ.

Comparison of Voltage-Dependent Gating of α_1A and α_1C.

The Ba²⁺ currents mediated by α_1A inactivate significantly more rapidly than those mediated by α_1C (Fig. 2A; τ = 127 ± 14 ms for α_1A vs. 915 ± 279 ms for α_1C), and the time constant of inactivation for α_1C is substantially larger than for α_1A across a wide range of test voltages (Fig. 2B, solid symbols). α_1A also activates at more negative membrane potentials than α_1C (Fig. 2C, solid symbols). The midpoint of the activation curve was 23.6 ± 1.4 mV (n = 15) for α_1A compared with 35.4 ± 2.6 mV for α_1C, and the voltage dependence was less steep for α_1C. Similarly, the voltage dependence of steady-state inactivation at negative membrane potentials where channels inactivate primarily from the closed state is 32 mV more negative for α_1A than for α_1C (Fig. 1D, solid symbols).

Figure 2.

Effects of mutations in the QXXER motif on voltage-dependent gating of P/Q-type and L-type Ca²⁺ channels. Control intracellular solutions without GTP[γS] were used in all experiments. (A) Ba²⁺ currents elicited by a 1000-ms test pulse to +30 mV from a holding potential of −60 mV for the indicated α₁ subunits and mutants. (B) Voltage dependence of the time constant for inactivation. Ba²⁺ currents were recorded during 1000-ms test pulses to the indicated potentials from a holding potential of −60 mV. Currents were fitted with one exponential and the time constants were plotted against the voltage of the test pulse. •, α_1A; ○, α_1A/QQIEE; ▴, α_1C; and ▵, α_1C/QQLER. (C) Voltage dependence of activation. Tail currents were recorded at a holding potenial of −60 mV following a 4-ms test pulse to the indicated potential. Tail currents were normalized to the largest tail currents in each series of test pulses, and means ± SEM were plotted as a function of the test voltage. •, α_1A; ○, α_1A/QQIEE; ▴, α_1C; and ▵ α_1C/QQLER. (D) Steady-state inactivation. Tail currents were recorded following a 4000-ms prepulse to the indicated potential and a 4-ms test pulse to +30 mV. Tail currents were normalized to the largest tail currents in each series of test pulses, and means ± SEM were plotted as a function of the prepulse voltage. •, α_1A; ○, α_1A/QQIEE; ▴, α_1C; and ▵, α_1C/QQLER.

Effects of Mutations in the QXXER Sequence on Voltage-Dependent Gating.

Because Ca²⁺ channels containing α_1A and α_1C differ in their gating behavior, we examined whether the amino acid difference in the fifth position of the QXXER motif between α_1C (QQLEE) and α_1A (QQIER) might alter the gating properties of the channel. The voltage dependence of activation is only slightly shifted to more negative potentials for the mutant channel α_1A/QQIEE compared with wild-type (WT) α_1A (Fig. 2C). In contrast, large differences were found for the kinetics of inactivation and for the voltage dependence of steady-state inactivation. The time course of inactivation of the mutant is much slower than WT α_1A (Fig. 2A) over a broad voltage range (Fig. 2B). The effect is most striking at +15 mV where the mutant is slowed 26-fold (Fig. 2B). The midpoint of the voltage dependence of steady-state inactivation is shifted from −25.4 ± 2.2 mV for α_1A to −6.9 ± 1.6 mV for the mutant channel, and the voltage dependence is steeper (Fig. 2D). Thus, the inactivation properties of the mutant α_1A containing the QQIEE sequence in L_I-II resemble those of α_1C.

Complementary effects on inactivation were observed for mutation of the QQLEE sequence in α_1C to QQLER. The voltage dependence of activation is shifted slightly to more negative potentials for the mutant channel compared with WT α_1C, as observed for the converse mutation in α_1A (Fig. 2C). The midpoint for the voltage dependence of steady-state inactivation changed from 7.1 ± 2.3 mV for WT α_1C to 1.1 ± 2.0 mV for the mutant (Fig. 2D). The mutant channel inactivated with a time constant of 238 ± 10 ms at +30 mV compared with a time constant of 915 ± 279 ms for WT (Fig. 2A), and more rapid inactivation for the mutant was observed over the whole voltage range tested (Fig. 2B). Thus, conversion of QXXEE to QXXER makes inactivation of Ca²⁺ channels containing α_1C substantially more similar to those containing α_1A, although the effects are not as large as the effects of the converse mutation in α_1A.

Effects of Mutations in the QXXER Sequence on G Protein Modulation.

Our results with intracellular peptides implicate the QXXER sequence in L_I-II in G protein modulation of Ca²⁺ channels (Fig. 1). As shown in Fig. 3 A and B, activation of α_1A is inhibited by GTP[γS] and facilitated by depolarizing prepulses but α_1C is not. To test whether mutations in this sequence alter G protein regulation, we analyzed the modulation of Ca²⁺ channels containing α_1A subunits in which the QQIER sequence was mutated to QQIEE and AQIAA. Neither of these mutations in α_1A block G protein modulation of the voltage dependence of Ca²⁺ channel activation (Fig. 3 C and D). In fact, prepulse facilitation is significantly enhanced for these mutations relative to α_1A (Fig. 3D). The α_1A and α_1C subunits also differ in the third position in the QXXER consensus sequence. Conversion of this residue from I in α_1A/QQIEE to L as in α_1C to yield the double mutant α_1A/QQLEE had relatively little effect on inactivation but had a dramatic effect on modulation by Gβγ (Fig. 3E). The voltage dependence of activation in the presence of GTP[γS] was much less shifted to positive membrane potentials, and prepulses to positive voltages caused a correspondingly small facilitation of the Ba²⁺ current. These results show that the amino acid residue in the third position within the QXXER motif is a crucial determinant of G protein modulation since the subtle change of I to L can nearly completely block facilitation.

Figure 3.

Effects of mutations in the QXXER motif on G protein modulation. Tail currents were recorded following paired test pulses. A 4-ms test pulse (test 1) to the indicated test potential was applied from the holding potential of −60 mV. After 1s, a 10-ms conditioning prepulse to +100 mV was applied, the cell was repolarized to −60 mV for 10 ms, and a second 4-ms test pulse (test 2) to the same potential as test 1 was applied. Tail currents were normalized to the largest tail current in each series of test pulses, and means ± SEM were plotted against test pulse potential. An increase in tail current after test 2 compared with test 1 indicates facilitation. The current traces shown in each panel were recorded during and after test 1 and test 2 at +30 mV. ▵, Control intracellular solution, test 1; ○, GTP[γS], test 1; •, GTP[γS], test 2. (A) α_1A. (B) α_1C. (C) α_1A/AQIAA. (D) α_1A/QQIEE. (E) α_1A/QQLEE. (F) α_1C/QQLER.

We examined whether G protein modulation could be transferred to α_1C. Introduction of the QXXER motif into α_1C by changing QQLEE to QQLER did not induce facilitation in the presence of GTP[γS] (Fig. 3F). Moreover, further mutation to QQIER as in α_1A also did not induce facilitation in the presence of GTP[γS] (data not shown). Thus, the QXXER sequence is not sufficient to confer Ca²⁺ channel modulation by Gβγ.

To detect quantitative changes in the rates and extents of G protein modulation in these mutants, we investigated the rates of onset and reversal of G protein action in intracellular solution containing GTP[γS]. To measure the rate of onset of G protein action, a test pulse was applied to measure the baseline Ba²⁺ current, positive prepulses were applied to reverse G protein modulation, an interval of varying duration was interposed to allow G protein modulation to recover, and a second test pulse was applied to measure the extent of recovery of G protein modulation at each time point (Fig. 4A). The facilitation ratio (I_tail[test 2]/I_tail[test 1]) was largest for α_1A/QQIEE, intermediate for WT and α_1A/AQIAA, and lowest for α_1A/QQLEE. The time constants for the onset of G protein action to reduce I_Ba under these conditions were similar for α_1A with all four sequences (Fig. 4B). To measure the rate of reversal of G protein action, a test pulse was applied to measure the baseline Ba²⁺ current, a positive prepulse of varying duration was interposed to reverse G protein modulation, and a second test pulse was applied to measure the extent of the G protein effect at each time point (Fig. 4B). The time constant for recovery from G protein modulation was increased from 4.8 ± 0.4 ms (n = 15) for WT α_1A to 9.1 ± 1.3 ms (n = 5) for α_1A/QQIEE and was reduced to 2.5 ± 0.6 ms (n = 4) for α_1A/QQLEE. Because the voltage-dependent reversal of G protein modulation is thought to reflect dissociation of Gβγ from the Ca²⁺ channel upon depolarization, these results indicate that the change from QQIER to QQIEE stabilizes G protein binding and slows Gβγ dissociation while the further mutation to QQLEE substantially destabilizes G protein binding and accelerates Gβγ dissociation. Evidently, changes in the stability of the bound G protein are largely responsible for the differences in the extent of facilitation for these different α_1A subunits.

Figure 4.

Effects of mutations in the QXXER motif of the α_1A subunit on the rates of onset and reversal of G protein modulation in the presence of GTP[γS]. (A) Rate of onset. A 4-ms test pulse (test 1) to +30 mV was applied from the holding potential of −60 mV. After 1s, a 10-ms conditioning prepulse to +100 mV was applied to completely relieve G protein inhibition, the cell was repolarized to −60 mV for a period of 1–500 ms, and a second 4-ms test pulse (test 2) to +30 mV was applied. Tail current of the second test pulse was divided by the first test pulse, and means ± SEM were plotted versus the time interval at −60 mV between prepulse and test pulse 2. •, α_1A/QQIER; ○, α_1A/QQIEE; ▴, α_1A/AQIAA; and ▵, α_1A/QQLEE. (B) The time courses of G protein action (A) were fitted with one exponential, and the corresponding time constants were plotted. (C) Rate of reversal. A 4-ms test pulse (test 1) to +30 mV was applied from the holding potential of −60 mV. After 1s, a conditioning prepulse to +100 mV for a period of 1 to 18 ms was applied, the cell was repolarized to −60 mV for 10 ms, and a second 4-ms test pulse (test 2) to +30 mV was applied. The tail current of the second test pulse was divided by the first test pulse, and means ± SEM were plotted against time interval at +100 mV. •, α_1A/QQIER; ○, α_1A/QQIEE; ▴, α_1A/AQIAA; and ▵, α_1A/QQLEE. (D) Time courses for reversal of G protein action were fitted with one exponential and the corresponding time constants were plotted.

Effects of the Ca²⁺ Channel β Subunit on Expression of WT and Mutants.

Coexpression of Ca²⁺ channel β subunits with the α₁ subunits increases the current amplitude, shifts the voltage dependence of activation and steady-state inactivation to more negative potentials, and accelerates inactivation (2). The QXXER sequence overlaps the β subunit interaction domain on Ca²⁺ channel α₁ subunits (24). However, as both α_1A and α_1C bind β subunits, little effect of our mutations on binding of β subunits is expected. The principal effect of coexpression of β subunits with α_1A in tsA-201 cells is a substantial increase in current amplitude. For WT α_1A, inclusion of the β subunit increased Ba²⁺ currents from 101 ± 21 pA (n = 11) to 660 ± 622 (n = 28). For α_1A/QQIEE, coexpression of β subunits increased Ba²⁺ currents from 99 ± 26 pA (n = 10) to 1170 ± 250 pA (n = 35) and for α_1A/QQLEE the increase was from 120 ± 47 pA (n = 5) to 592 ± 391 pA (n = 9), indicating that the binding and action of the β subunit were retained in these mutants. Similar results were observed with the α_1A/AQIAA mutation. Expression of α_1C WT and mutants without β subunits did not yield detectable currents, so the α_1C mutants that yield normal levels of Ba²⁺ current must also interact effectively with β subunits.

DISCUSSION

L_I-II Is Involved in Binding of Caβ Subunits, Voltage-Dependent Inactivation, and G Protein Modulation.

The intracellular loops that connect the four homologous domains of the Ca²⁺ channel α₁ subunit and the NH₂ and COOH termini are the largest domains of the protein and are the most divergent domains among the different isoforms of the α₁ subunits. It is likely that these large domains are important in interaction of Ca²⁺ channels with intracellular proteins that serve as intracellular effectors or modulators. Previous studies have implicated L_II-III of the Ca²⁺ channel α₁ subunits in interaction with two different intracellular effectors, the ryanodine-sensitive Ca²⁺ release channel in the sarcoplasmic reticulum of skeletal muscle (25) and the SNARE (SNAP receptor) proteins involved in docking and exocytosis of neurotransmitter vesicles in neurons (26). These results suggest that L_II-III may be specialized as an effector interaction domain. L_I-II of the α₁ subunit is known to be involved in binding of the β subunits of Ca²⁺ channels, which have important modulatory effects on the voltage dependence of activation and the kinetics and voltage dependence of inactivation (24). The consensus sequence for binding of Caβ subunits overlaps the QXXER sequence. Our present results implicate L_I-II directly in both voltage-dependent inactivation and G protein modulation, suggesting that L_I-II may be specialized to serve as a modulatory domain for Ca²⁺ channel gating. Voltage-dependent conformational changes in L_I-II involving the QXXER motif could alter the affinity for Gβγ binding as well as the ease of transition to the inactivated state. Complex regulatory interactions among L_I-II, Caβ, and Gβγ may be expected to provide subtle regulation of channel properties. Interestingly, the corresponding intracellular loop of Na⁺ channels contains a family of phosphorylation sites for cAMP-dependent protein kinase and protein kinase C which also modulate channel function (27), indicating that this loop serves a modulatory function in the structurally related Na⁺ channel α subunit as well.

Arginine in the QXXER Sequence Is an Important Determinant of Voltage-Dependent Inactivation.

Mutation of R to E in the fifth position of the QXXER sequence has a striking effect on voltage-dependent inactivation. The R/E mutation in α_1A slows the rate of inactivation from the open state and also inhibits steady-state inactivation from the closed state by shifting the voltage dependence to more positive membrane potentials. These effects create a channel with inactivation properties very similar to α_1C. The converse mutation in α_1C accelerates the rate of inactivation and shifts the voltage dependence of steady-state inactivation toward more negative membrane potentials, but its effects are not as pronounced. Our results point to this arginine residue as a key determinant of the rapid, voltage-dependent inactivation and relatively negative steady-state inactivation of Ca²⁺ channels containing α_1A. This single amino acid difference accounts for much of the difference in inactivation properties between these channel types.

In addition to this amino acid residue in L_I-II, analysis of chimeras formed from α_1A and α_1E showed that the membrane-spanning segment S6 in the first domain (IS6) and its flanking regions are also important determinants of the kinetics of voltage-dependent inactivation (28). It seems most likely that the transmembrane segments of the Ca²⁺ channels contain the main elements required for voltage-dependent inactivation, in analogy with C-type inactivation of K⁺ channels (29), and that the inactivation process can be further modulated by sequences in L_I-II and by interactions with Caβ subunits and Gβγ subunits with this intracellular loop.

Peptides Containing the QXXER Sequence Block G Protein Modulation.

Our results show that synthetic peptides containing the QXXER sequence derived from either adenylyl cyclase type 2 or from α_1A can effectively prevent G protein modulation. Similarly the AC2 peptide inhibited regulation of adenylyl cyclase type 2, phospholipase Cβ, and voltage-gated K⁺ channels (15). It is thought that this peptide binds to Gβγ subunits and prevents their interaction with regulatory targets. This mechanism may also explain the effects of the rbA peptide from L_I-II of the α_1A subunit that we have used in our experiments. However, the Ca²⁺ channel peptide may also affect interactions with Caβ subunits that bind to an overlapping segment of α_1A (24). Experiments in which the binding of Gβγ is measured directly will be needed to define the mechanism of action of these peptides on Ca²⁺ channel modulation.

Mutations in the QXXER Sequence Affect G Protein Modulation.

Mutation of QQIER in α_1A to QQIEE, AQIAA, or QQLEE substantially alters the extent and the rate of reversal of Gβγ regulation. These results clearly implicate the QXXER sequence in L_I-II in G protein modulation of α_1A, and indicate that it plays a role in determining the affinity for Gβγ binding and action on α_1A. However, mutations in the QXXER sequence do not completely abolish G protein modulation, and introduction of a QXXER sequence into α_1C does not transfer G protein modulation. These results are most consistent with the conclusion that this sequence is important for G protein modulation but does not constitute the entire interaction site for Gβγ. Consistent with this idea, Gβγ apparently binds to multiple regions of inwardly rectifying K⁺ channels, only one of which contains the QXXER sequence in the COOH-terminal domain (17), and x-ray crystallography shows that the interaction between Gβγ and phosducin involves numerous amino acid residues spread through both the NH₂ and COOH termini of phosducin (30). Further experiments will be necessary to define all of the requirements for binding and modulation of Ca²⁺ channels by Gβγ.