Scanning mutagenesis of the I-II loop of the Cav2.2 calcium channel identifies residues Arginine 376 and Valine 416 as molecular determinants of voltage dependent G protein inhibition

Direct interaction with the β subunit of the heterotrimeric G protein complex causes voltage-dependent inhibition of N-type calcium channels. To further characterize the molecular determinants of this interaction, we performed scanning mutagenesis of residues 372-387 and 410-428 of the N-type channel α1 subunit, in which individual residues were replaced by either alanine or cysteine. We coexpressed wild type Gβ1γ2 subunits with either wild type or point mutant N-type calcium channels, and voltage-dependent, G protein-mediated inhibition of the channels (VDI) was assessed using patch clamp recordings. The resulting data indicate that Arg376 and Val416 of the α1 subunit, residues which are surface-exposed in the presence of the calcium channel β subunit, contribute significantly to the functional inhibition by Gβ1. To further characterize the roles of Arg376 and Val416 in this interaction, we performed secondary mutagenesis of these residues, coexpressing the resulting mutants with wild type Gβ1γ2 subunits and with several isoforms of the auxiliary β subunit of the N-type channel, again assessing VDI using patch clamp recordings. The results confirm the importance of Arg376 for G protein-mediated inhibition and show that a single amino acid substitution to phenylalanine drastically alters the abilities of auxiliary calcium channel subunits to regulate G protein inhibition of the channel.


Background
The mammalian nervous system expresses nine different genes that encode different types of voltage-gated calcium channel (VGCC) α 1 subunits which interact with auxiliary subunits and form classes of VGCCs that are distinct in structure, pharmacology, and physiology [1]. VGCCs containing the α 1A and α 1B subunits (P/Q-and N-type channels respectively) are distinguished from other types by their localization to pre-synaptic nerve terminals, where they mediate calcium influx which contributes to evoked neurotransmitter release and overall synaptic function [2][3][4].
Inhibition of P/Q-and N-type channels resulting from activation of G-protein coupled receptors (GPCRs)-a crucial mode of regulation, notably illustrated in the relief of pain sensations in response to opioid drugs [5]has been studied for over 25 years [6][7][8][9][10][11][12][13][14][15][16]. This mode of regulation is complex and comprises multiple pathways that include direct and indirect actions of G proteins on the channel [17]. During membrane delimited G protein inhibition, GPCR activation releases Gβγ heterodimers which then bind directly to the α 1 subunits of P/Q-and N-type channels, and this interaction stabilizes closed channel conformations and culminates in channel inhibition [18,19].
A recent study suggests that interaction of Gβγ with N-type channels can slow the kinetics of channel transition to inactivated states [20]. However, most studies of the direct Gβγ-presynaptic channel interaction have investigated the slowing of transition to activated channel states, and have found this mode of inhibition to be more favored at hyperpolarized potentials, thus allowing for activity dependent dis-inhibition [21][22][23][24]. Hence, the term "voltage-dependent inhibition" (VDI) has been used to describe two experimental hallmarks of this Gβγ-mediated regulation: slowing of presynaptic channel activation, and relief of channel inhibition by a strong, depolarizing pre-pulse.
Gβγ-mediated VDI depends on a complex set of structural determinants that contribute to direct interaction between Gβγ and the presynaptic calcium channel. As such, the extent of VDI varies with the isoforms of the channel subunits and the G protein subunits in question [17]. Structure-activity relationship studies of the interaction have revealed roles for three cytosolic regions of the α 1 subunit: the N-terminus, the I-II linker domain, and the C-terminus [25][26][27][28][29][30]. While the C-terminal region of the channel is thought to play a supporting role as an enhancer of overall Gβγ-channel binding affinity [31], both the N-terminus and the I-II linker contribute directly to the inhibitory interaction with Gβγ. Furthermore, contact between the N-terminus and the I-II linker is demonstrated to be necessary for Gβγmediated VDI [25]. Efforts to resolve functionally important Gβγ-channel binding interactions have also revealed the direct involvement of two nearby sections of the I-II linker: amino acid residues 372-389 and 410-428 [27,30]. The first of these sections partially overlaps with the α 1 subunit alpha interaction domain (AID) and contains residues known to bind the calcium channel β subunit, presumably in a manner that precludes many of them from interaction with Gβγ [32]. However, the contribution of the other I-II linker residues in question-to direct Gβγ-binding and hence to channel inhibition-has remained unclear.
Here we aimed to further resolve the molecular determinants of Gβγ-mediated channel inhibition by testing the functional contribution of individual residues in the two above-mentioned sections of the α 1B I-II linker. Using a combination of alanine/cysteine scanning mutagenesis and whole-cell electrophysiological recordings from tsA-201 cells, we identify two residues of the I-II linker, Arg 376 and Val 416 , as key determinants of Gβγmediated, voltage-dependent modulation of N-type channels.
The construction of cDNAs encoding WT human Gβ 1 and Gγ 2 subunits have been described previously [33].
Ca v 2.2 a 1B mutants cDNAs encoding single-residue Ca v 2.2 α 1B mutants were constructed by overlap extension PCR [34], using WT α 1B as the DNA template and Pfu turbo DNA polymerase (Stratagene) according to manufacturer's instructions. AarI and BsiWI restriction sites, found in the native sequence at locations flanking the mutagenized I-II loop-encoding sequence, were included in the 5' ends of the non-mutagenic flanking primers. After mutagenizing and overlap extension reactions, mutant α 1B cDNA fragments were digested with AarI and BsiWI, and then sub-cloned into the (AarI-, BsiWI-digested) α 1B mammalian expression vector, pCMV30-14G. Codons for 27 residues within amino acid sequence regions 372-389 and 410-428 were individually substituted to alanine, and three naturally occurring alanine codons were substituted to cysteine. These , and a double alanine mutation, (both R376A and V416A), were also constructed using overlap/extension PCR as described above.

Tissue Culture and Transient Transfection
Human embryonic kidney tsA-201 cells were grown and transiently transfected using the calcium phosphate method as described previously (32). Transfection solutions for individual culture dishes contained a mixture of cDNA expression vectors, with the following quantities of each cDNA expression construct used: calcium channel α 1B subunit, 6 μg; Ca v β subunit (6 μg), Ca v α 2δ 1 subunit (6 μg), Gβ 1 subunit, 6 μg; Gγ 2 subunit, 6 μg; and 1 μg of a pEGFP marker vector (Clontech). Positive controls contained the WT α 1B subunit in place of mutant α 1B , and negative controls consisted of the WT α 1B subunit in the absence of exogenous Gβ 1 γ 2 . 12 hr post-transfection, cells were washed 1× with PBS pH 7.4, supplemented with fresh DMEM, and allowed to recover for an additional 12 hr. To prevent overgrowth, cells were subsequently transferred to a 29°C incubator and maintained for 24-72 hr prior to voltage-clamp recording.

Ca v 2.2 Voltage Clamp Recordings
Glass coverslips carrying cells expressing mutant or WT Ca v 2.2 channels were transferred to a 3.5-cm culture dish (Corning) containing external recording solution consisting of 20 mM BaCl 2 , 1 mM MgCl 2 , 10 mM HEPES, 40 mM TEA-Cl, 10 mM glucose, and 65 mM CsCl (pH 7.2 adjusted with TEA-OH). Micro-electrode patch pipettes were pulled using a Sutter P-87 microelectrode puller or a DMZ Universal puller, and manually fire-polished using a Narishige MF-830 Micro Forge to attain a typical resistance of 4-5 MÙ. Internal pipette solution consisted of 108 mM CsMeSO 4 , 4 mM MgCl 2 , 9 mM EGTA, and 9 mM HEPES (pH 7.2 adjusted with CsOH).
Whole cell patch clamp recordings were performed in voltage-clamp mode using an Axopatch 200B amplifier (Axon Instruments) linked to a personal computer with pCLAMP version 9.0 or 9.2 software. Series resistance was compensated by 85%, leak currents were negligible, and the data were filtered at 1 kHz. Individual pEGFPexpressing cells were held at -100 mV, and currents were evoked by stepping to a test potential of +20 mV. Only cells with current amplitudes greater than 50 pA and less than 1.5 nA were used for analysis.
Voltage-dependent G protein inhibition was assessed by application of a strong, depolarizing pre-pulse (PP) to +150 mV for 50 ms, immediately prior to the test potential-during alternating sweeps of an assay. Prepulse relief of inhibition was quantified as the ratio of peak current amplitudes observed in paired test pulses performed with (I +PP ) and without (I -PP ) the prepulse (i. e., I +PP /I -PP ).

Data Analysis
All data were analyzed using Clampfit version 9.2 (Axon Instruments) and fitted in Sigmaplot 2000 (SPSS Inc.). Statistical analyses were carried out using SigmaStat 2.03 (SPSS Inc.). All sample means are reported +/-SEM. Statistically significant differences between means were assessed using student's t-test, Mann-Whitney rank sum test, or one-way ANOVA at 95% confidence level as appropriate.

Results
Previously, two sections of the N-type channel I-II linker region, α 1B amino acid residues 372-389 and 410-428, were identified as functionally important binding sites for the Gβγ heterodimer [27,30]. To test the contribution of individual residues of these sections to Gβγmediated channel inhibition, alanine/cysteine scanning mutagensis was performed. Residues 372-389 include six amino acid residues that are predicted, on the basis of crystallographic data, to be unavailable for interaction with Gβγ, as their access is likely occluded by the calcium channel β subunit [32]. The remaining residues in this section, and in the second region (residues 410-428) were individually substituted to alanine or cysteine. The resulting mutant channels were coexpressed with human Gβ 1 γ 2 , and their respective susceptibilities to Gβγ-mediated VDI were quantified using a pre-pulse facilitation (PPF) paradigm (Figure 1). WT channels displayed the hallmark characteristics of Gβγ-mediated channel inhibition ( Figure 1A, left), including kinetic slowing of activation and relief of inhibition by a strong depolarizing pre-pulse (PPF ratio for WT channel assays: 2.36 +/-0.23). When examining PPF ratios obtained with mutant and WT channels, two of the 30 mutants examined, R376A and V416A, showed a significant loss of Gβγ-mediated inhibition when compared to wild type channels (PPF ratios 1.64 +/-0.18 and 1.44 +/-0.01, respectively; *P = 0.028 and 0.001, respectively; see Figure 1B). To test whether or not the effect of these mutations on Gβγ-mediated channel inhibition were additive, a double Ca v 2.2 α 1B mutant containing both the R376A and V416A substitutions was engineered. Co-expression of this double mutant with exogenous Gβ 1 γ 2 , and subsequent electrophysiological analysis using the PPF protocol, found the degree of Gβγ-mediated inhibition to be significantly less than that of WT channels, but similar to that observed in the presence of either one of the individual mutations alone (PPF ratio: 1.71 +/-0.15, *t-test, P = 0.023) (Figure 2A).
To further characterize the roles of Arg 376 and Val 416 , additional single mutant α 1B subunits, containing R376E, R376F, V416E, and V416K substitutions, were engineered and co-expressed with exogenous Gβ 1 γ 2 . Neither of the latter Val 416 substitutions resulted in significant changes in PPF ratio as compared to WT channels, Fig. 2A). However, the phenylalanine substitution at Arg 376 significantly increased the PPF ratio for channels coexpressed with Gβ1γ2 (mean values of 2.36 and 3.62 for WT and R376F, respectively, t-test p < 0.017, Fig. 2A), suggesting that this amino acid substitution actually increased Gβγ-mediated channel inhibition. These data thus further support the notion of residue 376 being an important determinant of VD G protein modulation of N-type channels.
The enhancement of Gβγ-mediated VDI by the R376F mutation, and the proximity of this mutation to critical sites of interaction between the α 1B and β subunits of the Ca v 2.2 channel, led us to ask whether the nature of the Ca v β subunit might affect this enhancement. To examine this issue, we coexpressed the α 2 -δ subunit and the R376F mutant with different isoforms of the Ca v β subunit in tsA-201 cells, and for each resulting population of Ca v 2.2 channels, we tested the effect of coexpression of heterologous Gβ 1 γ 2 on current densities and PPF ratios. As shown in Fig. 2B, little VD modulation was observed in the absence of exogenously coexpressed Gβ 1 γ 2 irrespective of the type of calcium channel β subunit that was present. For channels containing either Ca v β 2a or Ca v β 4 , coexpression with Gβ 1 γ 2 induced Tedford et al. Molecular Brain 2010, 3:6 http://www.molecularbrain.com/content/3/1/6 robust VD modulation of channel activity, whose magnitude was, however, smaller than that observed with channels containing Ca v β 1 . Strikingly, virtually no VD modulation was evident in R376F channels containing the Ca v β 3 subunit (i.e., there was no significant difference in PPF in the presence and the absence of G proteins). These data are in striking contrast to our previous findings showing that with WT Ca v 2.2, Gβ 1 γ 2 most strongly inhibited channels containing Ca v β 2a , followed by Ca v β 3 , Ca v β 4 and Ca v β 1B [35]. Hence, a single amino acid substitution in the Ca v 2.2 I-II linker drastically alters the Ca v β subunit dependence of Gβ 1 γ 2 inhibition of the channel.

Discussion
In this manuscript we have narrowly focused on the contribution of individual amino acids in the Ca v 2.2 I-II linker region to voltage dependent G protein inhibition of the channel. Among thirty amino acids in the I-II linker of the Ca v 2.2 channel, we have identified two, Arg 376 and Val 416 , that serve as determinants of G-protein mediated VDI of Ca v 2.2 channels, suggesting a highly localized interaction of Gβ 1 γ 2 with the I-II loop. The impact of single amino acid substitutions on G protein inhibition is reminiscent of our earlier findings showing that phosphorylation of a single I-II linker residue, Thr 422 , can disrupt modulation of Ca v 2.2 channels by Gβ 1 [36,37].
Arg 376 is particularly interesting because the R376F mutation drastically altered the impact of Ca v β subunit coexpression on the degree of VDI: whereas Gβ 1mediated inhibition of WT channels is reported to be strongest for channels containing Cavβ 2a , followed by Ca v β 3 , Ca v β 4 and Ca v β 1B , respectively [35], we report here that Gβ 1 -mediated regulation of R376F channels is strongest for channels containing Ca v β 1B , followed by Ca v β 2a , Ca v β 4 , Ca v β 3 , respectively, with no significant voltage-dependent inhibition observed for the latter. At 2 isoforms for each condition presented except for the negative control ("-Gβ 1 γ 2 "). Respective mutations in the Ca v 2.2 amino acid sequence are indicated beneath the corresponding columns (see METHODS for full description of the mutations used). Of the various conditions examined only the mutations R376A, R376F, V416A, and the double mutation R376A, V416A resulted in a significant loss of Gβγ-mediated channel inhibition, as measured by the degree of pre-pulse relief following a depolarizing pre-pulse, when compared to WT control (*p < 0.05, t-test, **p < 0.05 one-way ANOVA, Dunnett's method, or Kruskal-Wallis one-way ANOVA on ranks). Numbers in parentheses indicate numbers of cells tested for the respective condition. B: Histogram summarizing the results of PPF experiments using tsA-201 cells co-transfected to express α 1B mutant R376F with Ca v β isoforms β 1B , β 2 , β 3 , and β 4 , with or without heterologous human Gβ 1 γ 2 as indicated by labels beneath columns. Columns show mean PPF values with SE bars for each condition (see METHODS for full description of the mutations used). Of the conditions examined, coexpression of Ca v β 1B and Ca v β 2a , and Ca v β 4 resulted in statistically significant differences in mean current density for R376F channels expressed with and without heterologous Gβγ (*p < 0.05 using t-test, Mann-Whitney rank sum test, and t-test, respectively). this point we do not know how the mutation of residue 376 to phenylalanine increases VD G protein inhibition. When coexpressed with Ca v β 1B in the absence of heterologous Gβ 1 γ 2 , the R376F mutant had a half activation potential that did not differ significantly from that of the wild type channels (data not shown); moreover, at the majority of test potentials examined, the mutation yielded no significant changes in the rates of activation and inactivation (versus WT channels, data not shown), suggesting that the effects of the mutation on Gβγ modulation are not complicated by changes in biophysical properties of the channel. An increased degree of prepulse relief could occur as a result of several mechanisms. First, the mutation could destabilize the binding of Gβγ to the channel, thus resulting in more effective dissociation of Gβγ from the channel in response to membrane depolarization. This however seems unlikely, because the kinetics of the facilitated current were found to be similar for both the wild type and the mutant channels, indicating that both channels are completely dis-inhibited (and thus dissociated from Gβγ) following the application of a prepulse (data not shown). Agler and colleagues [25] reported that the N-terminus of Ca v 2.2 is capable of interacting with the domain I-II linker, and that this interaction contributes to G protein inhibition of the channel. It is thus conceivable that the nature of residue 376 could affect G protein inhibition indirectly by virtue of altering the binding of the N-terminus to the I-II linker.
Alternatively, it is possible that residue 376 is involved in transducing Gβγ binding to alter channel gating, such that a stronger voltage dependent inhibition is observed in the mutant channel. Residue 376 is three amino acid residues just upstream of an alpha helical structure (the AID, comprising residues 379-396) that is involved in binding of the Ca v β subunit [38], and could potentially serve as a hinge that links G protein binding to the gating machinery of the channel. However, it has also been proposed for Ca v 2.1 channels that I-II linker residues 357-378 are all part of a stable continuation of the alpha helical structure of the AID, and that stability and continuity of this helical structure is required for VD G protein inhibition of the channel [39]. In the latter case the R376F mutation may simply create a more stable binding pocket for Gβγ, perhaps in part by eliminating one of the eight positive charges carried by the side chains of I-II residues 357-396, which may render this section of the I-II linker less likely to move in response to a membrane depolarization event. Whatever the actual case, the proximity of Arg 376 to the Ca v β subunit interaction site also provides for a mechanism by which the nature of the Ca v β subunit could affect the extent of G protein inhibition that is observed.
Although alanine mutagenesis of residues 376 and 416 significantly reduced the effects of Gβγ, VDI was not completely eliminated, and the effects of the individual amino acid substitutions were not additive. This suggests that either other amino acid residues in the Ca v 2.2 α1 subunit might help stabilize the binding of Gβγ to the channel (such as for example, residue in the N-terminus), or that the Ca v β subunit may contribute directly to anchoring Gβγ to the channel. The latter would be consistent with recent findings showing that the presence of the Ca v β subunit is required to permit VDI of Ca v 2.1 channels [39].
Altogether, our data further implicate the domain I-II linker region as an important contributor to voltage dependent G protein modulation of N-type calcium channels. Furthermore, our results suggest that the regulation of N-type calcium channels by G proteins involves complex interactions between Gβγ, the Ca v 2.2 α1 subunit, and the auxiliary Ca v β subunit, and reveal that substitution of a single amino acid residue that is conserved in all HVA calcium channels is sufficient to significantly alter the interactions among these players. Although the precise molecular mechanism by which residue 376 couples Gβγ interactions to alterations in channel function remains unknown, the observation that highly localized alteration of a single amino acid residue increased G protein inhibition of the channel may offer potential avenues to enhance the efficacy of therapeutics acting on N-type channels via GPCRs.