Amino acids in segment IVS6 and beta-subunit interaction support distinct conformational changes during Ca(v)2.1 inactivation.

Ca(v)2.1 mediates voltage-gated Ca2+ entry into neurons and the release of neurotransmitters at synapses of the central nervous system. An inactivation process that is modulated by the auxiliary beta-subunits regulates Ca2+ entry through Ca(v)2.1. However, the molecular mechanism of this alpha1-beta-subunit interaction remains unknown. Herein we report the identification of new determinants within segment IVS6 of the alpha(1)2.1-subunit that markedly influence channel inactivation. Systematic substitution of residues within IVS6 with amino acids of different size, charge, and polarity resulted in mutant channels with rates of fast inactivation (k(inact)) ranging from a 1.5-fold slowing in V1818I (k(inact) = 0.98 +/- 0.09 s(-1) compared with wild type alpha(1)2.1/alpha2-delta/beta1a k(inact) = 1.35 +/- 0.25 s(-1) to a 75-fold acceleration in mutant M1811Q (k(inact) = 102 +/- 3 s(-1). Coexpression of mutant alpha(1)2.1-subunits with beta(2a) resulted in two different phenotypes of current inactivation: 1) a pronounced reduction in the rate of channel inactivation or 2) an attenuation of a slow component in I(Ba) inactivation. Simulations revealed that these two distinct inactivation phenotypes arise from a beta2a-subunit-induced destabilization of the fast-inactivated state. The IVS6- and beta2a-subunit-mediated effects on Ca(v)2.1 inactivation are likely to occur via independent mechanisms.


Institut fü r Biochemische Pharmakologie, Peter-Mayr-Strasse 1, A-6020 Innsbruck, Austria
Ca v 2.1 mediates voltage-gated Ca 2؉ entry into neurons and the release of neurotransmitters at synapses of the central nervous system. An inactivation process that is modulated by the auxiliary ␤-subunits regulates Ca 2؉ entry through Ca v 2.1. However, the molecular mechanism of this ␣ 1 -␤-subunit interaction remains unknown. Herein we report the identification of new determinants within segment IVS6 of the ␣ 1 2.1-subunit that markedly influence channel inactivation. Systematic substitution of residues within IVS6 with amino acids of different size, charge, and polarity resulted in mutant channels with rates of fast inactivation (k inact ) ranging from a 1.5-fold slowing in V1818I (k inact ‫؍‬ 0.98 ؎ 0.09 s ؊1 compared with wild type ␣ 1 2.1/␣ 2 -␦/␤ 1a k inact ‫؍‬ 1.35 ؎ 0.25 s ؊1 ) to a 75-fold acceleration in mutant M1811Q (k inact ‫؍‬ 102 ؎ 3 s ؊1 ). Coexpression of mutant ␣ 1 2.1-subunits with ␤ 2a resulted in two different phenotypes of current inactivation: 1) a pronounced reduction in the rate of channel inactivation or 2) an attenuation of a slow component in I Ba inactivation. Simulations revealed that these two distinct inactivation phenotypes arise from a ␤ 2a -subunit-induced destabilization of the fast-inactivated state. The IVS6-and ␤ 2a -subunit-mediated effects on Ca v 2.1 inactivation are likely to occur via independent mechanisms.
Calcium (Ca 2ϩ ) entry through Ca v 2.1, also known as class A or P/Q-type Ca 2ϩ channels (1), plays a central role in triggering the release of neurotransmitters from presynaptic nerve terminals as well as influencing other critical neuronal functions (2). The pore-forming ␣ 1 2.1-subunit is encoded by the CACNA1A gene, which is highly expressed in the central nervous system (3)(4)(5)(6). Mutations and deletions in the ␣ 1 2.1-subunit result in certain neurological disorders such as the familiar hemiplegic migraine, episodic ataxia type 2, and spinocerebellar ataxia type 6 (Ref. 5 and 7-11; see Ref. 12 for review). The importance of this channel type for neurological disorders is further emphasized by the recent finding that Ca v 2.1-deficient mice develop hallmark characteristics of severe ataxia and dystonia and subsequently die 3-4 weeks after birth (13).
The rate of voltage-dependent channel closure during depolarization, a process termed inactivation, is an important de-terminant of Ca 2ϩ entry during a neuronal action potential. Three different types of Ca 2ϩ channel inactivation processes have been identified in Ca v 2.1 channels: fast and slow voltagedependent inactivation (14,15) as well as a Ca 2ϩ -dependent inactivation mechanism (16). Ca v 2.1 inactivation is strongly influenced by ␣ 1 2.1-␤-subunit interaction. For example, coexpression of ␤ 2a results in substantially slower I Ba inactivation compared with channels composed of either ␤ 1a -or ␤ 3 -subunits (14,17,18). Sequence stretches responsible for ␣ 1 2.1 interaction with different ␤-subunits have been identified on intracellular linkers between domains I-II (I-II linker) and the amino and carboxyl termini (19 -21). Moreover, a soluble NSF (Nethylmaleimide factor) attachment protein receptor (SNARE)protein interaction domain within the II-III linker also influences inactivation gating of Ca v 2.1 (22)(23)(24).
The structural determinants of Ca v 2.1 inactivation are widely spread over the ␣ 1 2.1-subunit (see Ref. 25 for review). The first molecular determinants involved in the process of Ca v 2.1 inactivation were disclosed by Zhang et al. (26). This study demonstrated a key role of segment IS6 and adjacent intra-and extracellular stretches in the rate of channel inactivation. An important role of segment IVS6 was subsequently demonstrated by Döring et al. (27). Mutations within IVS6 of Ca v 2.1 produce profound effects on the pharmacological properties of Ca v 2.1 as well as on inactivation (28). However, the precise relationship (e.g. common or independent mechanisms) between IVS6-and ␤-subunit-mediated alterations in Ca V 2.1 inactivation remains to be elucidated.
Here we evaluate previously observed IVS6-mediated changes in Ca v 2.1 inactivation (27) at the level of single amino acids. The results identify a number of residues within IVS6 that represent critical determinants of voltage-dependent inactivation of Ca V 2.1. Systematic substitution of the important inactivation determinant Met-1811 by amino acid residues of different side chain length, charge, or polarity resulted in an array of channel constructs with inactivation properties that varied from being comparable with that of wild-type (e.g. M1811S) to an acceleration of inactivation by nearly 2 orders of magnitude (M1811Q). Thus, Met-1811 obviously imparts a strong influence on the rate of fast voltage-dependent inactivation of Ca V 2.1.

EXPERIMENTAL PROCEDURES
Amino acids in segment IVS6 that are conserved in all Ca 2ϩ channels were substituted by alanine residues (except for the alanine at position 1817, which was replaced by a serine residue). Only 3 out of the 8 mutant cDNAs (N1813A, A1817S, and M1820A) resulted in functional channels when the cRNAs were injected into Xenopus oocytes. However, substitution of the other amino acids by a methionine residue resulted in functional Ca v 2.1 mutants: Y1799M, F1800M, S1802M, F1803M, F1809M. Ala-1817 was additionally mutated to a methionine, resulting in functional A1817M construct. All constructs were inserted into the polyadenylating transcription plasmids pNKS2 (a kind gift of Dr. O. Pongs) and verified by sequence analysis.
Electrophysiology-Preparation of stage V-VI oocytes from Xenopus laevis, synthesis of capped off run-off poly(A ϩ ) cRNA transcripts from linearized cDNA templates, and injection of cRNA were performed as previously described in detail by Grabner et al. (31). Barium currents through calcium channels (I Ba ) were studied 2-7 days after microinjection of approximately equimolar cRNA mixtures of wild type ␣ 1 2.1 IVS6 mutants (0.3 ng/50 nl) with ␤ 1a , ␤ 2a , or ␤ 3 (0.1 ng/50 nl) and ␣ 2 -␦ (0.1 ng/50 nl) using the two microelectrode voltage clamp technique. The bath solution contained 40 mM Ba(OH) 2 , 50 mM NaOH, 5 mM HEPES, 2 mM CsOH adjusted to pH 7.4 with methanesulfonic acid as previously described (32). I Ba of Ca v 3.1 (33) was measured using the same extracellular solution after injection of equimolar cRNA mixtures of ␣ The rate of inactivation (k decay ) was determined by fitting the initial phase of the current decay to I Ba ϭ C exp(-k decay t). Voltage steps were applied from Ϫ80 mV to the peak current potential of the current voltage relationship (I-V) (i.e. voltage of the maximal inward current of the different ␣ 1 2.1/␣ 2 -␦/␤ 1a constructs ranged from 7 Ϯ 3 mV in L1812I to 20 Ϯ 3 mV in mutant Y1797V). Coexpression of ␤ 2a did not significantly shift the peak of the I-V curve compared with other ␤-subunit compositions.
The voltage of half-maximal inactivation (V 0.5,inact ) under quasi steady-state conditions was measured using a multi-step protocol. A control test pulse (50 ms to the peak potential of the I-V curve) was followed by a 1.5-s step to Ϫ100 mV followed by a 3-s conditioning step, a 4-ms step to Ϫ100 mV, and a subsequent test pulse to the peak potential. Inactivation during the 3-s conditioning pulse was calculated as I Ba,inact ϭ 1 Ϫ I Ba test/I Ba control .
The pulse sequence was applied every 3 min from a holding potential of Ϫ100 mV, and the estimated inactivation curves were fitted to a Boltzmann equation: where V is the membrane potential, V 0.5,inact is the midpoint voltage of the inactivation curve, k is a slope factor, and I ss represents the fraction of a noninactivating current. The voltage of half-maximal activation (V 0.5,act ) was estimated from g peak /g peak,max ϭ 0.5, where g peak ϭ I peak / (V Ϫ E rev ), g peak,max is the maximum value of g peak measured at the descending part of the I-V curve, and E rev is the reversal potential.
Recovery from inactivation was studied using a conventional doublepulse protocol. After depolarizing the channels from a holding potential of Ϫ80 mV for 3 s to the peak current potential of the I-V curve, 30-ms test pulses were applied at various time intervals to the same voltage. Peak I Ba values were normalized to the peak current measured during the pre-pulse, and the time course of I Ba recovery from inactivation was fitted to a biexponential function (I Ba,recovery ϭ A exp(Ϫt/ fast ) ϩ B exp(Ϫt/ slow ) ϩC).
The pClamp software package (Version 6.0 Axon Instruments, Inc.) was used for data acquisition and preliminary analysis. Microcal Origin 5.0 was employed for analysis and curve fitting. Data are given as the mean Ϯ S.E. Statistical significance was calculated according to Student's unpaired t test (p Ͻ 0.05 for n Ն 4).

Role of Ca v 2.1-specific IVS6 Amino Acids in Inactivation
Gating-We have previously reported that replacing ␣ 1 2.1 segment IVS6 by ␣ 1 1.1 sequence results in a pronounced acceleration of the inactivation kinetics of the resulting Fig. 1A illustrates the putative folding structure of a Ca v 2.1 channel (segment IVS6 is highlighted). Of the 25 amino acids predicted to form transmembrane segment IVS6, 8 residues are different between ␣ 1 2.1and ␣ 1 1.1-subunits (shaded residues shown in Fig. 1B). To assess their individual impact in inactivation gating, we substituted each of these residues by the corresponding ␣ 1 1.1 amino acids. The resulting 8 mutant channels, Y1797V, V1801L, I1804Y, F1805M, S1808A, M1811I, L1812I, V1818L ( Fig. 1C), were expressed together with ␣ 2 -␦ and ␤ 1a auxiliary subunits in Xenopus oocytes, and their inactivation properties were analyzed using the two-microelectrode voltage clamp technique (see "Experimental Procedures").
Wild type ␣ 1 2.1/␣ 2 -␦/␤ 1a channels inactivated at a rate (k inact ) of 1.35 Ϯ 0.25 s Ϫ1 . Only two of the point mutations (M1811I and L1812I) induced a pronounced acceleration of inactivation kinetics. Mutants V1801L, F1805M, and S1808A exhibited a small but significant acceleration, whereas V1818I inactivated at a significantly slower rate compared with wild type Ca v 2.1 ( Fig. 2A). An 8-fold acceleration in the rate of I Ba inactivation observed for M1811I (k inact ϭ 10.8 Ϯ 1.5 s Ϫ1 ) was not significantly different from that of AL23 (k inact ϭ 11 Ϯ 0.8 s Ϫ1 , see inset of Fig. 2A). The second fastest inactivation rate (k inact ϭ 5.26 Ϯ 0.39 s Ϫ1 ) was observed for mutation L1812I, located adjacent to M1811. Compared with wild type, the double mutant ML1811/1812II (ML/II) 1 exhibited an accelerated rate of current inactivation (not significantly different from M1811I, p Ͼ 0.05, Fig. 2A). However, only ML/II, and not M1811I, exhibited a V 0.5,inact value comparable with that of AL23 (Fig. 2, A and B). Thus, these data indicate that both Met-1811 and Leu-1812 significantly contribute to the AL23 inactivation phenotype.
The effects of the amino acid substitutions on the voltage dependence of channel activation and inactivation are illustrated in Fig. 2B (V 0.5,inact ) ranged between Ϫ32 Ϯ 2 mV in ML/II to Ϫ6 Ϯ 3 mV in Y1797V. A comparison of the mean half-activation voltages revealed that only L1812I and V1801L induced a significant shift to more negative voltages (V 0.5,act ϭ Ϫ3 Ϯ 2 mV (L1812I) and V 0.5,act ϭ Ϫ7 Ϯ 2 mV (V1801L) compared with Ca v 2.1 V 0.5,act ϭ 2 Ϯ 2, Fig. 2B). Fig. 2 demonstrate that Met-1811 plays a pivotal role in the rate of voltage-dependent inactivation of Ca v 2.1. To more thoroughly evaluate the role of residue 1811, we systematically replaced this amino acid by residues of different size, polarity, and charge. Surprisingly, all substitutions of Met-1811 accelerated the time course of current inactivation. Replacements of Met-1811 by amino acids with side chains of different charge or polarity (Gln, Glu, Asn, Lys) induced a substantial acceleration of the I Ba decay ranging from about 12-fold M1811K (k inact ϭ 16.3 Ϯ 1.3 s Ϫ1 ) to about 75-fold for M1811Q (k inact ϭ 102 Ϯ 3 s Ϫ1 ). Interestingly, M1811Q inactivated with similar kinetics as Ca v 3.1 (k inact ϭ 104 Ϯ 3 s Ϫ1 ), a channel known to display one of the fastest inactivation kinetics among all Ca 2ϩ channel subtypes (Fig. 3). Substitutions by hydrophobic residues of different size resulted in less pronounced effects on I Ba decay (range between 2-fold in M1811A (k inact ϭ 2.23 Ϯ 0.24 s Ϫ1 ) and 8-fold in M1811I (k inact ϭ 10.8 Ϯ 1.5 s Ϫ1 ); Fig. 3, A and B).
To investigate inactivation properties of methionine 1811 substitutions in more detail, we also analyzed the rate of recovery from inactivation. The rapidly inactivating construct M1811Q/␣ 2 -␦/␤ 1a recovered from fast inactivation with surprisingly similar kinetics as that of wild type ␣ 1 2.1/␣ 2 -␦/␤ 1a channels (see Fig. 4A). Analogous observations were made for other Met-1811 mutants (Fig. 4B). Thus, these data suggest that despite pronounced acceleration in the rate of fast inactivation (ranging from about 12-fold in M1811K to almost 75-fold in . The most dramatic acceleration was induced by mutation M1811Q with a rate of I Ba decay (102 Ϯ 3 s Ϫ1 , n ϭ 6). The rate of I Ba inactivation of Ca v 3.1 (104 Ϯ 3 s Ϫ1 , gray column) estimated at the peak current potential of Ϫ20 mV is given for comparison (n ϭ 4). Mutant ␣ 1 2.1-subunits were coexpressed together with ␣ 2 -␦-and ␤ 1a -subunits. The ␣ 1 3.1-subunit was coexpressed with the ␣ 2 -␦-subunit. Significant differences (p Ͻ 0.05) compared with ␣ 1 2.1 are indicated by asterisks. Inset, scaled superimposed peak I Ba during 100-ms depolarizations from Ϫ80 to Ϫ20 mV (Ca v 3.1) and to 20 mV (Ca v 2.1 and M1811Q). B, scaled superimposed I Ba of the indicated Ca v 2.1 mutants (same voltage protocol as in Fig. 2). C, voltages of halfmaximal activation (V 0.5,act , open circles) and half-maximal inactivation potential (V 0.5,inact , filled circles) of the Met-1811 mutants (n Ն 3). Fig. 3A), these mutations facilitate entry into the fast-inactivated state without affecting stability of the inactivated state.

IVS6 Mutations Produce Different Kinetic Phenotypes of ␤ 2a -Subunit Modulation-It is well established that different
␤-subunits differentially modulate the inactivation kinetics of Ca v 2.1 (14,15,17,18,34). To elucidate the role of structural changes in different parts of segment IVS6 on ␤-subunit modulation, we systematically analyzed I Ba inactivation kinetics of our Ca v 2.1 mutants when expressed in combination with either ␤ 1a -subunit or ␤ 2a -subunits.
As illustrated in Fig. 6, we observed two principal patterns of ␤ 2a -subunit modulation. In most of the IVS6 mutants, coexpression of the ␤ 2a -subunit dramatically slowed the fast component in I Ba decay (called herein type I modulation, Fig. 6, A  and B, left panels). This pattern of ␤ 2a modulation was previously documented for wild type ␣ 1 2.1 (14,15,17,18). A different type of modulation was observed for M1811Q, M1811E, M1811N, and M1811K. In those fast-inactivating ␣ 1 2.1(mutant)/␣ 2 -␦/␤ 1a constructs, coexpression of ␤ 2a had much less effect on the transient current decay but, instead, induced a slowly inactivating I Ba component (type II modulation, Fig. 6, A  and B, right panels). Thus, our data clearly demonstrate that depending on the initial rate of I Ba inactivation, coexpression of the ␤ 2a -subunit results in distinct inactivation phenotypes. 25 for review). Fast and slow voltage-dependent inactivation of this channel type are regulated by auxiliary ␤-subunits and other intracellular regulator proteins (14,15,17,18,(22)(23)(24).
The molecular mechanism of Ca 2ϩ channel inactivation and its modulation by various ␤-subunits are still incompletely understood. In particular, the question of whether ␤-subunitinduced changes in Ca 2ϩ channel gating are dependent on inactivation determinants in segment IVS6 remains unanswered.
In the present study we created 25 ␣ 1 2.1 point mutants with substantially different inactivation properties by replacing residues in segment IVS6 that are conserved in all other Ca 2ϩ channel classes by either alanine or methionine residues. In addition, we also mutated specific ␣ 1 2.1 IVS6 residues to the corresponding residues found in ␣ 1 1.2. The impact of inactivation determinants in segment IVS6 in ␤-subunit modulation was subsequently analyzed by expressing the mutant ␣ 1 2.1 with either fast-inactivating ␤ 1a -and ␤ 3 -subunits or the slowinactivating ␤ 2a -subunit.
Hot Spots of Inactivation Determinants in Segment IVS6 -The individual replacement of eight non-conserved IVS6 amino acids in ␣ 1 2.1 by the corresponding ␣ 1 1.1 residues enabled the identification of two amino acids that strongly influence Ca v 2.1 inactivation. Our data demonstrate that mutations of M1811I result in an acceleration in the time course of I Ba decay to nearly the same extent as the replacement of the entire IVS6 segment by ␣ 1 1.1 sequence (Fig. 2A). The second strongest effect was observed for substitution L1812I located adjacent to Met-1811 (Fig. 1A). Simultaneous replacement of both ␣ 1 2.1 residues by their ␣ 1 1.1 counterparts almost perfectly reproduced the phenotype of chimera AL23 (27), a chimera in which the entire IVS6 segment is of ␣ 1 1.1 sequence. The similarity between ML/II and AL23 is emphasized by the fact that these two constructs exhibit not only similar rates of I Ba inactivation ( Fig. 2A) but also similar midpoint voltages of activation and steady-state inactivation (Fig. 2B). Significant but less pronounced effects on channel inactivation were also observed for mutations V1801L, F1805M, and S1808A ( Fig. 2A). Other amino acid substitutions had minor effects on the time course of I Ba inactivation. However, their contribution to Ca v 2.1 inactivation is illustrated by the significant shifts of the availability curve (Fig. 2B).
Most dramatic changes in Ca v 2.1 inactivation occurred upon substitution of Met-1811 by glutamine or other charged/polar amino acids (M1811Q Ͼ M1811N Ͼ M1811E Ͼ M1811K). For example, M1811Q induced a 75-fold acceleration in the initial rate of I Ba inactivation compared with wild type, resulting in a Ca v 2.1 mutant inactivating with similar kinetics as Ca v 3.1 (Fig. 3).
We mutated each of the amino acids that are conserved between all Ca 2ϩ channel ␣ 1 -subunits in order to probe the role of these residues in Ca 2ϩ channel inactivation. Out of the eight alanine substitution mutants, only N1813A, M1820A, and A1817S formed functional Ca 2ϩ channels. It is unclear if the cRNA of the non-expressing mutants is translated and the resulting ␣ 1 -subunits represent non-conducting channels or if expression is stopped on the level of translation. However, substitution of the residues concerned by methionine (Y1899M, F1800M, S1802M, F1803M, F1809M, and A1817M) resulted in functional ␣ 1 2.1 mutants. Of these mutants, the most significant acceleration in channel inactivation was observed for mutations N1813A and A1817M. However, compared with up to a 75-fold acceleration in I Ba inactivation observed upon mutation of Met-1811, substitutions of the non-conserved IVS6 residues clearly exhibit a less significant influence on the rate of Ca v 2.1 inactivation.
Thus, we hypothesize that Met-1811 along with the closely located Leu-1812, Asn-1813, and Ala-1817 residues plays a critical role in helix packing within this putative bundle-crossing region of Ca 2ϩ channel ␣ 1 -subunits (in analogy to the orientation of S6 segments in KcsA channels (40); see Ref. 41 for review). Consistent with this notion, strong effects of amino acid substitutions within the inner pore region of Ca 2ϩ channel ␣ 1 -subunits have previously been reported for segment IVS6 of ␣ 1 1.2 (42) and segment IIIS6 of ␣ 1 2.1 (15).
The Role of IVS6 Mutations in ␤ 2a -Subunit Modulation of Ca v 2.1 Inactivation-Multiple ␤-subunits appear to be associated to various extents in different parts of the mammalian brain with the ␣ 1 2.1-subunit (43)(44)(45). Therefore, inactivation of Ca v 2.1 is expected to be modulated by tissue-specific ␤-subunit expression. To gain a deeper understanding of the molecular mechanism of ␤-subunit modulation of ␣ 1 2.1, we investigated the role of inactivation determinants within segment IVS6 on Ca v 2.1 inactivation observed in the presence of either ␤ 1a -or  3-7). B, I Ba of Ca V 2.1, M1811I (type I, left panels) and M1811Q, M1811N (type II, right panels) illustrate the two patterns of ␤ 2a -subunit modulation. Type I (evident as a significant decrease in the rate of fast inactivation k inact ) was found to be characteristic for channel mutants with only moderately changed inactivation kinetics. Type II modulation barely affected the transient current component but, instead, attenuated a slow phase in I Ba inactivation. Type II modulation was distinctive for most rapidly inactivating mutants. Note the different time scales in left and right panels. ␤ 2a -subunits. Thus, we compared the kinetic properties of fastinactivating ␣ 1 2.1(mutant)/␣ 2 -␦/␤ 1a channels with the corresponding channel construct formed using ␤ 2a -subunits (␣ 1 2.1(mutant)/␣ 2 -␦/␤ 2a ). As illustrated in Fig. 6, A and B, we observed two principle phenotypes of ␤ 2a modulation. One pattern of modulation (type I, Fig. 6A) was exhibited by wild type Ca v 2.1 channels and all IVS6 mutations that induced only moderate changes in the rate of fast inactivation ( Figs. 2A and  3A). Coexpressing these constructs with ␤ 2a resulted in a dramatic decrease of fast inactivation (illustrated in the left panels of Fig. 6B for wild type ␣ 1 2.1/␣ 2 -␦/␤ 2a and M1811I/␣ 2 -␦/␤ 2a channels). Type II modulation was found only in mutants that exhibited very rapid inactivation kinetics (such as M1811Q/ ␣ 1 2.1/␣ 2 -␦/␤ 1a and M1811N/␣ 1 2.1/␣ 2 -␦/␤ 1a , Fig. 6, A and B, right panels). For these mutants, coexpression of ␤ 2a did not markedly slow the transient component in I Ba inactivation but, instead, attenuated a slow component of current decay.
One hypothesis that can be put forward to explain the different kinetic phenotypes is that structural changes in segment IVS6 affects the interaction of ␤ 2a with ␣ 1 2.1(mutant)-subunits. Under this scenario, IVS6 residues would form part of a receptor for a ␤ 2a -modulated inactivation gate or lid interacting from the intracellular side of the channel pore (see Ref. 46 for a proposed hinged lid mechanism of Ca v 2.3 involving the I-II linker and segments IIS6 and IIIS6; see also Ref. 47). In the frame of such a hypothesis, structural changes in segment IVS6 would alter inactivation by changing the affinity of such a receptor site.
Alternatively, the ␣ 1 2.1-␤ interaction might be unaffected by structural changes in segment IVS6, and type I and type II modulation could simply result from an interplay of changes in microscopic inactivation rate constants induced by the different constructs. We have, therefore, analyzed the ␤ 2a -induced changes in inactivation gating of ␣ 1 2.1 point mutants with type I and type II modulation in terms of a simple Ca v 2.1 inactivation model that accounts for state transitions between open (O), fast-inactivated (Fast-I), and slow-inactivated states (Slow-I) during a membrane depolarization ( Fig. 7A; see also Ref. 15). Each of the ␣ 1 2.1 point mutations caused individual effects on the microscopic rate constants of fast (␣, ␤) and slow inactivation (␥, ␦). Fig. 7B illustrates four typical examples for type I or type II modulation simulated by means of domestically written software: wild type ␣ 1 2.1/␣ 2 -␦/␤ 1a channels, mutation M1811I/ ␣ 2 -␦/␤ 1a inducing an 8-fold acceleration of I Ba decay, and two mutants that were typical for type II modulation (M1811Q/␣ 2 -␦/␤ 1a with 75-and M1811N/␣ 2 -␦/␤ 1a with about 20-fold faster I Ba decay than wild type).
An ϳ1000-fold acceleration of a single microscopic transition rate (␤) from the fast-inactivated to the open state reproduced the ␤ 2a -subunit-induced changes in inactivation in all cases. A similar acceleration in the rate constant ␤ reproduced type I and type II modulation of other ␣ 1 2.1 mutants illustrated in Fig. 6A (data not shown).
Taken together, the simulations suggest that the ␤ 2a -subunit destabilizes the fast-inactivated channel conformation similarly for each of the ␣ 1 2.1 mutants. Therefore, the different inactivation patterns of wild type Ca v 2.1 and ␣ 1 2.1(mutant)/␣ 2 -␦/␤ 2a constructs (type I and type II) are likely to arise from the interplay of microscopic transition rates between the open and fast-inactivated states (see the legend of Fig. 7).
In summary, we identified new hot spots of determinants of Ca 2ϩ channel inactivation within the IVS6 segment of ␣ 1 2.1. Substitution of two Ca v 2.1 amino acids by their Ca v 1.1 counterparts (ML1811/1812II) reproduced the inactivation properties (i.e. rate of inactivation, V 0.5,act , and V 0.5,inact ) of a previously studied chimeric channel (AL23, Ref. 27), suggesting that these residues markedly contribute to voltage-dependent inactivation of Ca v 2.1. Nevertheless, the structural details of how the IVS6 segment within Ca v 2.1 influences channel inactiva- In wild type (␣ 1 2.1/␣ 2 -␦/␤ 1a ) and other channel constructs with a comparable rate of fast inactivation, the impact of the ␤ 2a -subunit-induced increase in the backward rate constant ␤ (100 s Ϫ1 ) is much larger than the impact of the forward rate ␣ (1-6 s Ϫ1 ). Consequently, only a negligible amount of channels reside in Fast-I, and the kinetics of I Ba decay of ␣ 1 2.1/␣ 2 -␦/␤ 2a channels are mainly determined by the rates of slow inactivation (␥ and ␦). In faster-inactivating mutants, the forward rate ␣ (23 s Ϫ1 in M1811N and 90 s Ϫ1 in M1811Q) is comparable with the backward rate ␤ (about 30 s Ϫ1 ), resulting in biphasic inactivation kinetics with a pronounced transient component (determined by the rates ␣ and ␤) and a slowly decaying second phase in I Ba inactivation (rates ␥ and ␦). tion remain to be clarified. Noncovalent interactions of specific IVS6 residues with neighboring transmembrane segments (i.e. IS6 or IVS5) are feasible. It will, therefore, be interesting to study if similar changes in Ca v 2.1 inactivation occur upon amino acid substitutions on other S6 segments by charged or polar residues.
The most intriguing result of this study, perhaps is that IVS6-and ␤ 2a -subunit-mediated conformational changes during Ca v 2.1 inactivation apparently occur in an independent manner (Fig. 7). Hence, type I and type II kinetics of the different ␣ 1 2.1 IVS6 mutants can be simulated by assuming a uniform ␤ 2a -mediated destabilization of the fast-inactivated channel state (Fig. 7).