N Terminus Is Key to the Dominant Negative Suppression of CaV 2 Calcium Channels IMPLICATIONS FOR EPISODIC ATAXIA TYPE 2 *

Expression of the calcium channels CaV2.1 and CaV2.2 is markedly suppressed by co-expression with truncated constructs containing Domain I. This is the basis for the phenomenon of dominant negative suppression observed for many of the episodic ataxia type 2 mutations in CaV2.1 that predict truncated channels. The process of dominant negative suppression has been shownpreviously to stem from interaction between the full-length and truncated channels and to result in downstream consequences of the unfolded protein response and endoplasmic reticulum-associated protein degradation. We have now identified the specific domain that triggers this effect. For both CaV2.1 andCaV2.2, theminimum construct producing suppression was the cytoplasmic N terminus. Suppression was enhanced by tethering the N terminus to the membrane with a CAAX motif. The 11-amino acid motif (including Arg52 and Arg54) within the N terminus, which we have previously shown to be required for G protein modulation, is also essential for dominant negative suppression. Suppression is prevented by addition of an N-terminal tag (XFP) to the full-length and truncated constructs. We further show that suppression of CaV2.2 currents by the N terminus-CAAX construct is accompanied by a reduction in CaV2.2 protein level, and this is also prevented by mutation of Arg52 and Arg54 to Ala in the truncated construct. Taken together, our evidence indicates that both the extreme N terminus and the Arg52, Arg54 motif are involved in the processes underlying dominant negative suppression.

meric complexes consisting of the pore-forming Ca V ␣1 subunit together (except in the case of the Ca V 3 channels) with an accessory ␤ and ␣ 2 ␦ subunit. The Ca V ␣1 subunit consists of four homologous domains (I-IV), each consisting of six transmembrane (TM) segments (see Fig. 1A). The domains are linked by intracellular loops and have intracellular N and C termini. Ten mammalian ␣1 subunit genes have been cloned and divided into three subfamilies Ca V 1-3 (2).
Mutations of calcium channel ␣1 subunits can contribute to a number of pathological states (3). In particular, mutations in the CACAN1A gene encoding Ca V 2.1 result in familial hemiplegic migraine and episodic ataxia type 2 (4). Many of the episodic ataxia type 2 mutations in Ca V 2.1 predict truncated forms of this channel, although missense mutations are also found (4 -7). This disease is dominant, and thus there is one wild-type (WT) allele and one mutant allele, both of which are likely to be expressed, although nonsense-mediated decay would reduce the expression of some mutant alleles (8). In many cases, the mutant channels, as well as either being nonfunctional or having reduced functionality, are dominant negative, in that they also suppress the function of the WT channel (9 -11).
In our initial study on truncated Ca V ␣1 subunits, we found that truncated constructs containing Domain I suppressed Ca V 2.2 currents and reduced the level of full-length Ca V 2.2 protein (12). We then showed that for both Ca V 2.2 and Ca V 2.1, this suppression required interaction between the full-length and the mutant construct (9). In this study, we also examined the effect of a two-domain construct predicted by an episodic ataxia type 2 mutation (9). We and others have also identified previously that the suppressive mechanism involves a reduction in protein synthesis resulting from the unfolded protein response (9) and an acceleration of proteasome-mediated decay (10).
Here, we have dissected the determinants required for suppression, which has increased our understanding of the mechanisms involved in the pathophysiology of episodic ataxia type 2. We find that the interaction between a truncated construct and a related full-length channel, identified previously (9), requires the presence of the N terminus on either or both of the full-length or the truncated channels. We also show that the N terminus of Ca V 2.2 or Ca V 2.1 alone is sufficient to suppress expression of the full-length channel. Suppression can be prevented by incorporation of a bulky tag on the N terminus or by removal of part of the N terminus. We further identify the motifs within the N terminus that are essential for suppression to occur and show that suppression can also be induced of endogenous channels in neurons.
Cell Culture and Heterologous Expression-COS-7 cells were cultured as described previously (18). The tsA-201 cells were cultured in a medium consisting of Dulbecco's modified Eagle's medium (DMEM), 10% fetal bovine serum, 1% Glutamax, 100 units/ml penicillin, and 100 g/ml streptomycin (Invitrogen). Cells were transfected using FuGENE 6 (Roche Diagnostics). The cDNAs (all at 1 g/l) for Ca V ␣1 subunits, truncated domain constructs, ␣ 2 ␦-1 or ␣ 2 ␦-2, ␤1b, and GFP, when used as a reporter of transfected cells, were mixed in a ratio of 3:1.5:2:1:0.2, unless stated otherwise. When particular subunits were not used, the volume was made up with water or blank vector, or the volume of transfection reagent was reduced, all with equivalent results.
Dorsal root ganglion (DRG) neurons isolated from Sprague-Dawley rats (175-250 g) in ice-cold Hanks' balanced salt solution (Invitrogen) were transferred to DMEM nutrient mixture F-12 (DMEM/F12) containing 0.4 mg/ml trypsin, 0.6 mg/ ml collagenase type 1 (both from Worthington Biochemical Corp.), and 100 units/ml DNase (Invitrogen), saturated with 95% O 2 /5% CO 2 , and incubated for 1 h at 37°C. The neurons were washed in DMEM/F12 supplemented with 10% fetal bovine serum, 1% penicillin-streptomycin, and 1% Glutamax (10% DMEM). Neurons were dissociated by vigorous shaking, centrifuged twice for 9 min at 800 ϫ g, and the pellet was resuspended in 200 l of Amaxa rat neuron nucleofector solution (Lonza Cologne AG, Cologne, Germany). Suspended neurons were mixed with cDNA for the truncated domain constructs (40 ng/l) and YFP (20 ng/l) DNA and transfected with nucleofector program O-003 following the manufacturer's instructions. The effect of the constructs was compared with DRG neurons expressing only YFP cDNA. The transfection reagent was neutralized with 500 l of 10% DMEM supplemented with 50 ng/ml nerve growth factor. Neurons from each group were plated on poly-L-lysine (0.5 mg/ml)-coated 22-mm coverslips (BDH), placed in 35-mm polystyrene tissue culture dishes for 2 h to settle, flooded with 10% DMEM supplemented with 50 ng/ml nerve growth factor, and cultured for 3-4 days at 37°C. Prior to experiments, the numerous neurite processes were eliminated by replating to improve voltage-clamp recording. Culture medium was removed, and cells were incubated for 5 min at 37°C in 1 ml of 10% DMEM containing 0.2% type 1 collagenase. The enzymatic reaction was stopped with 1 ml of 10% DMEM, and the neurons were triturated and spun for 9 min at 800 ϫ g. The pellet was resuspended in 300 l of 10% DMEM, and the cells from each group were plated on poly-Llysine-coated coverslips and left to recover for at least 2 h at 37°C before recording.
Xenopus oocytes were prepared, injected, and utilized for electrophysiology as described previously (17), with the following exceptions. Plasmid cDNAs for the different calcium channel subunits ␣1, ␣ 2 ␦, ␤1b, and truncated or mutated domains and other constructs were mixed in 2:1:2:2 ratios at 1 g/l, unless stated otherwise, and 9 nl was injected intranuclearly, after 2-fold dilution of the cDNA mixes. When the truncated domain was not included it was replaced by an equivalent volume of empty vector, water, or a cDNA for a nonfunctional transmembrane protein, Kir-AAA (15) with equivalent results.
Electrophysiology-For tsA-201 cells, the patch pipette solution contained 140 mM cesium aspartate, 5 mM EGTA, 2 mM MgCl 2 , 0.1 mM CaCl 2 , 2 mM K 2 ATP, 10 mM Hepes, pH 7.2, 310 mOsm with sucrose. The external solution contained 150 mM tetraethylammonium bromide, 3 mM KCl, 1.0 mM NaHCO 3 , 1.0 mM MgCl 2 , 10 mM Hepes, 4 mM glucose, 1 mM BaCl 2 , pH 7.4, 320 mOsM with sucrose. For DRGs, the patch pipette solution contained 140 mM cesium aspartate, 10 mM EGTA, 2 mM MgCl 2 , 5 mM K 2 ATP, 10 mM Hepes, pH 7.2, 310 mOsm with sucrose. The external solution was identical to that described above, except 10 mM BaCl 2 was used, and 1 M tetrodotoxin was included in the medium to suppress voltage-gated Na ϩ currents. I Ba was recorded using an Axopatch 1D amplifier (Axon Instruments, Molecular Devices, Sunnyvale CA), and data were filtered at 2 kHz and digitized at 10 kHz. Analysis was performed using pClamp9 (Axon) and Origin 7 (Microcal Origin, Northampton, MA). Current records are shown following leak and residual capacitance current subtraction (P/4 protocol). Incompletely subtracted capacitative transients have been truncated in traces shown. Recordings in Xenopus oocytes were performed as described (19), and all recordings were performed 48 -60 h after injection for Ca V 2.2 and 72-80 h after injection for Ca V 2.1. The Ba 2ϩ concentration was 10 mM, unless stated otherwise. When stated, current-voltage (I-V) plots were fit with a modified Boltzmann equation as described, for determination of the voltage for 50% activation (19).
Western Blotting and Calcium Channel Subunit Quantification-COS-7 cells were processed for SDS-PAGE as described (12). Samples (50 g of cell lysate protein/lane) were separated using Novex 4 -12% Tris-glycine or 4 -12% BisTris NuPAGE gels (Invitrogen) and transferred electrophoretically to polyvinylidene fluoride membranes. The membranes were blocked with 3% bovine serum albumin and 0.02% Tween 20 and then incubated overnight at room temperature with the relevant primary antibody: 1:1000 dilution of anti-Ca v 2.2 (12). Detection was performed either with a 1:1000 dilution of goat anti-rabbit (or anti-mouse) IgG-horseradish peroxidase conjugate (Bio-Rad) and ECL Plus (Amersham Biosciences), or with a 1:1000 dilution of goat anti-rabbit IgG-Cy5 conjugate (Amersham Pharmacia Biotech), all in conjunction with a Typhoon 9410 Variable Mode Imager (Amersham Pharmacia Biotech), set in chemiluminescence or fluorescence mode, respectively. Protein bands were quantified using ImageQuant 5.2. The same amount of total protein was loaded for all samples on a gel for accurate comparison between lanes.

RESULTS
To extend our studies on the suppression of Ca V 2.x channels by truncated domains containing Domain I ( Fig. 1A) (9,12), in terms of the specific truncated domain involved, we first mutated various structural motifs in Domain I and expressed the resultant constructs to narrow down the element(s) responsible for the suppression. We have used a number of different expression systems and methods to ensure that our results are able to generalize beyond a single system. Key experiments have been reproduced by more than one method.
Which Structural Elements within Domain I of Ca V 2 Channels Are Required for Suppression?-Ca V 2.2-Dom I alone produced ϳ90% suppression of Ca V 2.2 currents in Xenopus oocytes ( Fig. 1, B and C). Ca V 2.2-Dom I was previously found to be more effective than Ca V 2.2-Dom I-II to inhibit Ca V 2.2 currents (12). We then found that a construct consisting of the N terminus and the first four TM segments (S1-S4) of Ca V 2.2 (Ca V 2.2-Dom I-4TMs) was as effective as Ca V 2.2-Dom I, Ca V 2.2 I Ba being reduced to 13% of control ( Fig. 1, B and C). A conserved set of structural motifs in this region of Ca V 2.2-Dom I is the charged amino acids in TM segments S1-S4, which might mediate inappropriate interaction between the fulllength and truncated channel. However, mutation of all charged amino acids in the TM segments S1, S2, S3, and S4 to hydrophobic residues, within the truncated Ca V 2.2-Dom I-4TMs construct, did not significantly affect the ability of this construct to suppress Ca V 2.2 currents (Fig. 1, B and C). A second potential source of interaction is the conserved cysteines (in S1 and S2) which might form disulfide bonds with the fulllength WT channel. However, when the cysteine in S1 (Cys 110 ) was also mutated to serine to form Ca V 2.2-Dom I-4TMs (no charges, C110S), suppression of I Ba was again undiminished (Fig. 1, B and C).
Role of the N Terminus of Ca V 2.2 in Dominant Negative Suppression-We then surmised that the cytoplasmic N terminus might contain structural elements involved in suppression. To examine the role of the N terminus, we utilized truncated and full-length constructs of Ca V 2.2, in which either or both were engineered to contain N-terminal deletions. We have previously shown that ⌬1-55 Ca V 2.2 produced functional channels (17).
We first utilized expression in tsA-201 cells and compared the ability of two N-terminally deleted, truncated constructs of I) for their ability to suppress expression of Ca V 2.2 and ⌬1-55 Ca V 2.2 I Ba ( Fig. 2A). We found that the two N-terminally truncated Domain I constructs showed consistently less suppression of ⌬1-55 Ca V 2.2 than of Ca V 2.2 itself (Fig. 2), and there was no suppression of ⌬1-55 Ca V 2.2 by the construct with the longer N-terminal deletion, ⌬2-91 Ca V 2.2-Dom I (Fig. 2).
We then examined whether ⌬2-91 Ca V 2.2 was functional and found that, unlike truncations up to residue 55, its expression did not result in any discernible calcium channel currents (Fig. 3A). We next examined whether this N-terminally truncated channel would suppress expression of the full-length WT Ca V 2.2 I Ba and found substantial inhibition of Ca V 2.2 currents in Xenopus oocytes (Fig. 3A). The peak Ca V 2.2 I Ba at ϩ5 mV was reduced to 41.4 Ϯ 10.1% of control (p ϭ 0.016) in the additional presence of ⌬2-91 Ca V 2.2. However, there was no effect on any other properties of the currents, including steady-state inactivation (Fig. 3B). A similar result was observed when the same constructs were expressed in tsA-201 cells (Fig. 3C). Taken together, these data suggest that at least one intact N terminus, on either the full-length or the truncated channel construct, is required for suppression; and in agreement with this hypothesis, we found no suppression of ⌬1-55 Ca V 2.2 currents by the ⌬2-91 Ca V 2.2 construct in the same experiment (Fig. 3C) (12). This result was confirmed in the present study, the peak I Ba resulting from expression of GFP-Ca V 2.2 was nonsignificantly reduced in the presence of the N terminus of Ca V 2.2 (residues 1-95), by 8.9 Ϯ 19.5% (n ϭ 13; Fig. 4A).
In the light of the results described above, we then examined whether a construct consisting of the Ca V 2.2 cytoplasmic N terminus alone would be capable of inhibiting WT Ca V 2.2 I Ba . We found a 40% reduction in WT Ca V 2.2 currents when the Ca V 2.2 N terminus was co-expressed (Fig. 4, B and C).
To examine how the Ca V 2.2 N terminus was inhibiting Ca V 2.2 currents, we attached an extended CAAX motif to its C terminus, consisting of the last 10 amino acids of H-Ras. This would promote both prenylation and palmitoylation of the polypeptide and thus enhance the concentration associated with both plasma and internal membranes (20). To confirm the localization of constructs to which a CAAX motif was attached, we examined the distribution of GFP-CAAX, which was found to be associated both at the plasma membrane and also with cytoplasmic organelles (Fig. 4D), in contrast to free GFP, which was observed uniformly throughout the cytoplasm and also in the nucleus (Fig. 4D). The membrane-tethered Ca V 2.2 N terminus-CAAX produced a very strong inhibition of Ca V 2.2 currents, by 70% at 0 mV, whereas a control prenylated protein (GFP-CAAX) produced no significant inhibition (Fig. 4, B, C, and E). We then examined whether truncation of the Ca V 2.2 N terminus would prevent this inhibition and found that no significant inhibition was produced by Ca V 2.2 N terminus (⌬2-42)-CAAX (Fig. 4B), indicating that the extreme N terminus is involved in the process of suppression of Ca V 2.2 currents. We previously identified an 11-amino acid motif (residues 45-55) in the N terminus of Ca V 2.x channels, YKQSxAQRART, which was essential for G protein-mediated inhibition of these chan-nels (17,21). Two key amino acids involved in this process were found to be the two arginine residues in this motif. We therefore examined whether the same motif was involved in suppression of Ca V 2.2 current expression by mutating these two amino acids to alanine in the Ca V 2.2 N terminus-CAAX construct. We found that these mutations prevented the effect of the N terminus on Ca V 2.2 currents, even producing a small increase compared with Ca V 2.2 alone (Fig. 4, B and E). Together with the previous result, this suggests that docking of the free N terminus via a motif including Arg 52 and Arg 54 is involved in its ability to suppress Ca V 2.2 currents but that the extreme N terminus of Ca V 2.2 must also play a role.
Further studies showed that co-expression of the Ca V 2.2 N-terminal construct also produced a significant inhibition (ϳ50%) of ⌬1-55 Ca V 2.2 I Ba (Fig. 4F), and co-expression of Ca V 2.2 N terminus-CAAX strongly inhibited (by 75%) ⌬1-55 Ca V 2.2 I Ba (Fig. 4F), indicating that the suppressive effect of the Ca V 2.2 N terminus does not require the presence of the same motif on the full-length channel. This provides evidence that the suppressive effect is probably not via dimerization of the N termini.
One potential explanation for the reduction of Ca V 2.2 I Ba by the Ca V 2.2 N-terminal construct, given the involvement of the amino acids 45-55 (17,21), was that there could be tonic modulation of the calcium channel currents (22). However, co-expression of the Ca V 2.2 N-terminal constructs was not associated with slowed activation of the Ca V 2.2 currents (Fig. 4G). This was therefore discarded as a possibility.
Inhibition of Ca V 2.1 by the N Terminus of Ca V 2.1-We next examined the effect of the N terminus of Ca V 2.1 on the expression of the P/Q-type channel Ca V 2.1. Co-expression of the Ca V 2.1 N terminus with Ca V 2.1/␣ 2 ␦-2/␤4 produced 37.9% inhibition of the peak I Ba currents (Fig. 5, A and B). The relevance of this combination of channel subunits is that it is likely to be one of the main channel complexes in cerebellar Purkinje neurons (13), which mediates many of the effects of the mutations in episodic ataxia type 2. The effect of co-expression of the N terminus of Ca V 2.1 attached to a CAAX motif was a 43.8% reduction in peak I Ba , and this was also reversed by the corresponding mutations in the key arginine residues, Arg 57 and Arg 59 (Fig. 5B).
Furthermore, the CAAX motif-linked N terminus of Ca V 2.2, which has a strong homology to Ca V 2.1, also inhibited Ca V 2.1 currents by 63.5%, whereas the mutant R52A/R54A Ca V 2.2 N terminus-CAAX construct produced no inhibition, rather resulting in an increase in the peak I Ba compared with Ca V 2.1/ ␣ 2 ␦-2/␤4 alone (Fig. 5C).
Biochemical Basis for Suppression by the Ca V 2.2 N Terminus of Ca V 2.2 Channel Expression-We previously established that dominant negative suppression by truncated constructs involves a reduction of Ca V ␣1 subunit protein expression (9,12). In the present study, we found that co-expression of the Ca V 2.2 N terminus-CAAX with Ca V 2.2/␤1b/␣ 2 ␦-1 in tsA-201 cells resulted in a consistent decrease in the level of Ca V 2.2 protein (Fig. 6A). This was normalized to the total amount of protein in each sample and quantified as a 53% reduction, from six separate transfections (Fig. 6B). We also confirmed that expression of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was not altered by these manipulations (supplemental Fig. 1).
In our previous study implicating the unfolded protein response in this process, we also examined the level of co-expressed ␣ 2 ␦ protein and found it also to be reduced (9). The level of ␣ 2 ␦-1 was also reduced by co-expression of the Ca V 2.2 N terminus in the present study, by 51% (n ϭ 6; Fig. 6, A and B). To confirm whether the RAR motif in the N terminus was involved in this response, we examined the effect of the R52A/ R54A Ca V 2.2 N terminus-CAAX construct. In agreement with our electrophysiological results, the R52A/R54A Ca V 2.2 N terminus-CAAX construct produced no suppression of expression of Ca V 2.2 and ␣ 2 ␦-1 protein (Fig. 6, A and B).
Is There an Interaction between the N Terminus of Ca V 2.2 and Other Domains of Ca V 2.2?-We were unable to demonstrate any positive interactions between the N terminus of Ca V 2.2 and the Ca V 2.2 I-II loop, Ca V 2.2 Dom I or a number of other Ca V 2.2 sequences, using the yeast two-hybrid assay (supplemental data), in contrast to a previous study (22). Therefore we cannot identify a high affinity interaction of the Ca V 2.2 N terminus with a particular peptide domain of Ca V 2.2 using this system, indicating that it is perhaps more likely to interact in a complex binding pocket, made up of multiple elements.
Effect of Truncated Calcium Channels on Endogenous Calcium Channel Currents in DRG Neurons-We wished to examine whether the truncated constructs would also affect endogenous calcium channel currents, and we therefore expressed these constructs in DRG neurons using Amaxa transfection. We performed experiments 4 days after transfection, to allow synthesis of endogenous channels to occur, and in the presence of 10 M nifedipine, to isolate native N-type calcium channel currents (Fig. 7, A-C). Expression of Ca V 2.2 N terminus-CAAX produced a statistically significant reduction in DRG I Ba (Fig.  7B), whereas R52A/R54A Ca V 2.2 N terminus-CAAX produced no reduction in DRG I Ba (Fig. 7D).

DISCUSSION
Here, we have examined the process of dominant negative suppression of Ca V 2.1 and Ca V 2.2 currents by truncated Ca V ␣1 constructs, which was identified by Raghib et al. (12). Our previous conclusion was that if Ca V 2.2 is co-expressed with truncated constructs of Ca V 2.2 containing Domain I, expression of the full-length Ca V 2.2 channel protein is almost completely prevented. We subsequently identified that there was a requirement for interaction between the full-length and truncated constructs (9), which has been confirmed and extended by others (10).
We previously observed cross-suppression between the different subclasses of Ca V 2 channels (Ca V 2.1, 2.2, 2.3), where conservation both within the TM segments and in the cytoplasmic N and C termini and loops is fairly high. However, there was no significant cross-suppression between full-length Ca V 3.1 and truncated constructs of Ca V 2.2, and vice versa (9). This suggests that the response is induced by the association of part of the truncated domain with segments of a cognate fulllength channel with which it shows an affinity. Our present results identify specific motifs involved in the interaction.
We have found that one of the main regions involved in interaction is the N terminus of these channels. We observed that the optimum requirements for this to occur are that the N terminus should be both full-length and have a free N terminus (i.e. not tagged with XFP). When these conditions are met, in either or both of the truncated and the full-length Ca V 2.2 channels, suppression occurs, but when the extreme N terminus is missing from both constructs, or both are tagged with XFP, suppression is markedly reduced or absent.
Our finding that an N-terminal XFP tag hinders the dominant negative suppression process may also explain some anomalies in the literature, where such tagged constructs have been used in the study of this phenomenon (7,23).
Amino acids 1-55 of Ca V 2.2 contain the N-terminal motif MVRFGDEL attached to a highly flexible region GGRYGGTG-GGERARGGGAGGAGGPGQGGLPPG, representing amino acids 9 -40 and identified as being glycine-rich (56%) and hence of low complexity. This region is followed by YKQSIAQRART, which is the 11-amino acid motif that we have previously identified to be essential for G protein modulation in Ca V 2.x channels. This motif is highly conserved in the Ca V 2 family (17,21) and predicted to form an ␣-helix (PSIPRED 2.6). An additional  , n ϭ 14). The statistical significances of the differences were determined by one-way ANOVA followed by post-hoc Dunnett's test. *, p Ͻ 0.05. Error bars indicate S.E. C, example of current traces for voltage steps between Ϫ40 mV and ϩ65 mV for neurons expressing YFP only (control) or with Ca V 2.2 Dom I or Ca V 2.2 N terminus-CAAX (left to right). Recordings were made with 10 mM Ba 2ϩ in the presence of 10 M nifedipine. D, I Ba (recorded in the presence of 10 M nifedipine at ϩ10 mV) for DRG neurons expressing R52A/R54A Ca V 2.2 N terminus-CAAX (white bar, n ϭ 10), normalized as a percentage of control (black bar, n ϭ 9). motif residing within amino acids 56 -95 (MALYNPIPVKQN-CFTVNRSLFVFSEDNVVRKYAKRITEWPPFE) was also identified as being involved in the role of the N terminus of Ca V 2.2 to mediate G protein modulation by G␤␥ (22). We have now found that this region is essential for functional expression of the channel.
A possible scenario is that the high mobility of the low complexity region in the N terminus will allow the N terminus to interact with a distant binding pocket on the channel and that the motif containing the key amino acids Arg 52 and Arg 54 is involved in this process. This intramolecular interaction might be required as a quality check point for correct folding during channel synthesis. RXR motifs have previously been identified to regulate endoplasmic reticulum retention and retrieval in other channels (24) and may be acting by a similar mechanism here.
In agreement with this hypothesis, our results show that the Ca V 2.x N terminus, particularly when attached C-terminally to a CAAX motif, results in strong suppression of Ca V 2.2 expression, both in terms of functional currents and at the level of Ca V 2.2 protein. This suppression by the N terminus is prevented when the N terminus is truncated and when the amino acids Arg 52 and Arg 54 are mutated to Ala. We hypothesize that once the N terminus, either as a free domain or attached to a truncated channel, has interacted intermolecularly with the full-length channel, the misfolded aggregate may both be directed to the proteasomal pathway as suggested by others (10,25) and may also trigger the unfolded protein response to suppress further translation (9,26).
The relevance to episodic ataxia type 2 is that many of the mutations found in this dominant disease result in premature protein truncation, but in all the mutations to date, the N terminus is intact, the first known mutation being in Domain I (6). A number of studies have shown previously that the truncated channels predicted by episodic ataxia type 2 mutations interact with the full-length channel (9 -11, 23). Here, we have identified that an intact N terminus is essential for interaction between the truncated domain and the full-length channel. Future work to discover the site of interaction may now allow the development of therapeutic agents that hinder this process.