Transfer of L-type Calcium Channel IVS6 Segment Increases Phenylalkylamine Sensitivity of

Abstract Conditioned (“use-dependent”) inhibition by phenylalkylamines (PAAs) is a characteristic property of L-type calcium (Ca) channels. To determine the structural elements of the PAA binding domain we transferred sequence stretches of the pore-forming regions of repeat III and/or IV from the skeletal muscle α subunit (α) to the class A α subunit (α) and expressed these chimeras together with β and α/ subunits in Xenopus oocytes. The corresponding barium currents (I) were tested for PAA sensitivity during trains of depolarizing test pulses (conditioned block). I of oocytes expressing the α subunit were only weakly inhibited by PAAs (less than 10% conditioned block of I during a 100-ms pulse train of 0.1 Hz). Transfer of the transmembrane segment IVS6 from α to α produced an enhancement of PAA sensitivity of the resulting α/α chimera comparable to L-type α subunits (about 35% conditioned block of I during a 100-ms pulse train of 0.1 Hz). Our results demonstrate that substitution of 11 amino acids within the segment IVS6 of α with the corresponding residues of α is sufficient to transfer L-type PAA sensitivity into the low sensitive class A Ca channel.

Conditioned ("use-dependent") inhibition by phenylalkylamines (PAAs) is a characteristic property of Ltype calcium (Ca 2؉ ) channels. To determine the structural elements of the PAA binding domain we transferred sequence stretches of the pore-forming regions of repeat III and/or IV from the skeletal muscle ␣ 1 subunit (␣ 1S ) to the class A ␣ 1 subunit (␣ 1A ) and expressed these chimeras together with ␤ 1a and ␣ 2 /␦ subunits in Xenopus oocytes. The corresponding barium currents (I Ba ) were tested for PAA sensitivity during trains of depolarizing test pulses (conditioned block). I Ba of oocytes expressing the ␣ 1A subunit were only weakly inhibited by PAAs (less than 10% conditioned block of I Ba during a 100-ms pulse train of 0.1 Hz). Transfer of the transmembrane segment IVS6 from ␣ 1S to ␣ 1A produced an enhancement of PAA sensitivity of the resulting ␣ 1A /␣ 1S chimera comparable to L-type ␣ 1 subunits (about 35% conditioned block of I Ba during a 100-ms pulse train of 0.1 Hz). Our results demonstrate that substitution of 11 amino acids within the segment IVS6 of ␣ 1A with the corresponding residues of ␣ 1S is sufficient to transfer L-type PAA sensitivity into the low sensitive class A Ca 2؉ channel.
To localize Ca 2ϩ antagonist interaction domains within Ltype ␣ 1 subunits we have recently shown (8) that sensitivity for DHP Ca 2ϩ channel blockers and activators can be transferred to class A ␣ 1 subunits (␣ 1A ) by substituting regions close to the channel pore in repeats III and IV (segments IIIS5, IIIS6, IVS6, and the respective S5-S6 linkers) with the corresponding L-type ␣ 1 sequences (from ␣ 1S or ␣ 1C ). The DHP sensitivity was lost after replacement of short sequence stretches within these regions by the ␣ 1A sequence. For example, when segment IIIS5 was replaced by ␣ 1A sequence, DHP sensitivity disappeared. The same effect was observed after replacing the IVS5-IVS6 linker. These results suggest that the DHP molecules interact with multiple amino acid residues located within distant regions of the primary structure.
Hockerman et al. (9) recently identified three amino acid residues within segment IVS6 of a L-type Ca 2ϩ channel ␣ 1C subunit (Tyr-1463, Ala-1467, and Ile-1470, numbering according to ␣ 1C-c ) (10) as critical determinants for high affinity block by PAAs. Mutation of these residues within ␣ 1C-c to non-L-type resulted in a decrease of PAA sensitivity.
In our present work we studied the importance of the IVS6 segment for the formation of PAA interaction domains by investigating whether this region also supports PAA sensitivity in a non-L-type sequence environment. We addressed this question by testing whether the characteristics of L-type channel block by PAAs can be transferred to ␣ 1A that forms a non-L-type Ca 2ϩ channel. We demonstrate that currents through ␣ 1A expressed in Xenopus oocytes are less sensitive to PAAs than L-type currents. Transfer of the skeletal muscle IVS6 segment into ␣ 1A resulted in a chimeric ␣ 1 construct that displayed PAA sensitivity comparable to L-type currents. We therefore conclude that L-type IVS6 also supports PAA sensitivity in a non-L-type sequence environment. Construction of Chimeric ␣ 1 cDNAs-␣ 1 chimeras (AL21, AL22, and AL23; for nomenclature, see Fig. 1A) consisting of ␣ 1A from rabbit brain (BI-2) (11) and ␣ 1S from carp skeletal muscle (12) were constructed and inserted into the polyadenylating transcription plasmid pNKS2 (provided by O. Pongs). Polymerase chain reaction (PCR) was used to create common restriction sites by introducing silent cDNA mutations. Mutations were introduced into forward and/or reverse PCR primers or by the "gene SOEing" technique (13). Amplification of cDNA by PCR (Thermocycler 60, Biomed) was performed with 35 cycles at low stringency (1 min at 94°C, 30 s at 42°C, 1.5 min at 72°C) using proofreading Pfu-polymerase (Stratagene). Chimeras AL21, AL22, and AL23 were constructed as follows (PCR-generated restriction sites are indicated by asterisks). AL21 (amino acid numbers in parentheses): A(1-1723), S(1311-1437), A(1856 -2424). The SfiI-ClaI* fragment (nucleotide numbers in parenthesis) of A(4296 -4925) was ligated into the SfiI (4296A) and ClaI* (4925A) sites of the chimeric construct AL9-pNKS2 (8). AL22: A(1-1723), S(1311-1402), A(1821-2424). A BamHI* site at position 4303 (S) was created by "gene SOEing" in the KpnI*-BglII (5467A-6185A) fragment of chimeric construct AL12s (8). This PCR product was ligated into the KpnI* (5467A) and BglII (6185A) sites of AL21. AL23: A(1-1791), S(1374 -1402), A(1821-2424). KpnI*-BamHI* fragment of A(5467-5667) was ligated into the KpnI* (5467A) and BamHI* (4303S) sites of AL22. The construction of chimeras L h , L s , AL1, and AL4 was described previously (8,14). The correct nucleotide composition of the chimeras was verified by extensive restriction enzyme mapping and by cDNA sequencing with the dideoxy chain termination method (15).
Voltage Clamp Measurements-Ba 2ϩ inward currents (I Ba ) through voltage-gated Ca 2ϩ channels were measured between 2 and 7 days after injection of the oocytes (19) using the two-microelectrode voltage-clamp technique (Turbo Tec 01C, NPI-Electronic, Germany). Endogenous Ca 2ϩ channel currents were studied after injection of ␤ 1a and ␣ 2 /␦ subunit cRNAs. Endogenous currents were present only in a minority of the tested oocyte batches and reached levels of expression between 5 and 80 nA. Only oocytes displaying I Ba that where at least 10 times larger than endogenous currents were used for further analysis. Voltage-recording and current-injecting microelectrodes were filled with 2. The recording chamber (150-l total volume) was continuously perfused at a flow rate of 1 ml/min with control-or drug-containing solutions. Leakage current correction was performed by using average values of scaled leakage currents elicited by a 10-mV hyperpolarizing voltage step. The pClamp software package (version 5.51, Axon Instruments, Inc.) was used for data acquisition and analysis. Data were filtered at 1 kHz, digitized at 1 kHz, and stored on a computer hard disk.
Estimation of Conditioned I Ba Block by PAAs-Conditioned ("use-dependent") block of I Ba by the PAAs (Ϫ)-D888 and (Ϯ)-D600 was measured during trains of either 100-or 800-ms test pulses. Test pulses were applied after a 3-min equilibrium period in drug-containing solution at a standard frequency of 0.1 Hz. The conditioned channel block by PAAs was estimated as the inhibition of peak I Ba after 12 depolarizing test pulses (Figs. 1B and 2). The holding potential was Ϫ80 mV and test potentials were applied to the peak potential of the current voltage relationship of the Ca 2ϩ channel constructs. Because of incomplete recovery from inactivation of I Ba , some chimeras displayed a decay in peak I Ba during the pulse trains in the absence of drug. To estimate the peak I Ba decay under control conditions we applied similar test pulses in the absence of drug, which were preceded by a 3-min rest period (see Figs. 1B and 2). The peak I Ba inhibition during the first pulse after a 3-min equilibration in the drug-containing solution was defined as "resting-state dependent block." Data are given as ranges or mean Ϯ S.E. Statistical significance of I Ba block by PAA compared to current decay under control conditions was calculated according to unpaired Student's t test.

PAA Effects on L-type Ca 2ϩ
Channel Currents-To investigate if Xenopus oocytes are an appropriate expression system for studying PAA effects on Ca 2ϩ channels, we compared the PAA sensitivity of wild type ␣ 1A with the previously described L-type ␣ 1 chimeras L h and L s (8,14) after expression in oocytes. The injection of ␣ 1 subunit cRNAs, together with ␤ 1a and ␣ 2 /␦ subunit cRNAs, resulted in expression of calcium channels with barium current (I Ba ) amplitudes exceeding those of endogenous currents at least 10-fold (see "Experimental Procedures").
Both L-type chimeras are sensitive to DHP Ca 2ϩ channel agonists and antagonists and have been characterized previously (8,14,21). Chimera L h (Fig. 1A) corresponds to ␣ 1C (22) but with its NH 2 terminus replaced by the respective sequence from carp skeletal muscle (␣ 1S ) (12) to increase the yield of expression (21). Chimera L s (Fig. 1A) corresponds to L h with repeats III and IV replaced by sequence of the carp ␣ 1S . L s was tested for PAA sensitivity because carp ␣ 1S sequence stretches were used for the construction of ␣ 1A /␣ 1S chimeras (see below). As shown in Fig. 1B L Fig. 2A) which was most prominent for the slowly inactivating chimera L s . In the presence of drug I Ba recovered by 61 Ϯ 7% (mean for L h , n ϭ 6) from conditioned block during a 3-min rest at Ϫ80 mV.
The PAA (Ϯ)-emopamil exhibits 1-2 orders of magnitude lower affinity for the PAA binding domain of L-type Ca 2ϩ channels than (Ϫ)-D888 and (Ϯ)-D600 (24). Unlike these PAAs, (Ϯ)-emopamil (100 M) did not induce conditioned block of I Ba (shown for chimera L s in Fig. 1B). This suggests that the observed PAA effects are mediated by specific interaction with the PAA binding domain.
PAA Effects on Ca 2ϩ Channel Currents through ␣ 1A Subunits-In contrast to the L-type Ca 2ϩ channel ␣ 1 chimeras, PAA-induced block of I Ba for ␣ 1A was much less pronounced. Fig. 2B illustrates I Ba recordings from an oocyte expressing ␣ 1A during trains of test pulses in the absence and presence of 50 M (Ϫ)-D888. In the absence of drug I Ba decreased by 4 Ϯ 1% (n ϭ 7) during a 100-ms pulse train. This decrease in peak current amplitude was more pronounced (10 Ϯ 4%, n ϭ 5) after increasing the pulse length to 800 ms (Fig. 1B) and presumably resulted from incomplete recovery of I Ba from inactivation. 50 M (Ϫ)-D888 (Figs. 1B and 2B) or 100 M (Ϯ)-D600 (Fig. 1B) caused a small but significant additional conditioned block of ␣ 1A current during 100-ms pulse trains (corresponding to about 10% of the peak current value). I Ba block by 50 M (Ϫ)-D888 during a 800-ms pulse train was enhanced to 21 Ϯ 4% (n ϭ 5) (Fig. 1B). Taken together, these data demonstrate that Ca 2ϩ channels formed by ␣ 1A subunits are only weakly sensitive to PAAs compared to the L-type chimeras L s and L h .
PAA Effects on Ca 2ϩ Channel Currents through ␣ 1A /␣ 1S Subunits-To determine if the PAA sensitivity of L-type Ca 2ϩ channels can be transferred from an L-type Ca 2ϩ channel to ␣ 1A , we constructed a series of chimeras between ␣ 1A and L-type sequence (Fig. 1A). When repeat IV and the adjacent carboxyl terminus of the PAA-sensitive chimera L s were replaced by ␣ 1A sequence PAA sensitivity of the resulting chimera AL4 (Fig. 1A) was reduced to the level of the ␣ 1A subunit (Fig. 1B).
AL1 represents the first of four chimeras in which ␣ 1S sequence was introduced into ␣ 1A within repeats III and IV. It contains L-type sequences in the S5-S6 linkers and adjacent segments S6 in repeats III and IV (8) (Fig. 1A). A substantial fraction of I Ba from chimera AL1 did not recover from inactivation between the 10-s interpulse interval of the train in the absence of PAAs. As shown for ␣ 1A (Fig. 2B) this resulted in a decrease in I Ba amplitude during frequent depolarizations and was more pronounced if prolonged test pulses were applied (data not shown). This prevented the analysis of conditioned block during trains of pulses longer than 100 ms. During 100 ms pulse trains the PAA sensitivity of AL1 was comparable to constructs L h and L s : 50 M (Ϫ)-D888 induced a conditioned block of 24% (n ϭ 4) beyond the peak current decay of I Ba observed in the absence of drug (Fig. 1B). In chimera AL21 repeat III completely consisted of ␣ 1A sequence. The observed PAA sensitivity still resembled that of AL1 (Fig. 1B). Furthermore, neither the removal of L-type sequence on the cytoplasmic side of IVS6 (generating chimera AL22) nor of the IVS5-IVS6 linker (leading to chimera AL23, see Figs. 1 and 2B) decreased the PAA sensitivity. I Ba through chimera AL23 exhibited less than 2% (n ϭ 12) run-down and displayed the characteristic features of PAA sensitivity. (i) Resting state-dependent I Ba block of chimera AL23 was less than 5% (n ϭ 12), (ii) the fraction of Ca 2ϩ channels blocked by PAAs was dependent on the application of depolarizing test pulses (I Ba was inhibited during a train of 100-ms pulses by 57 Ϯ 6% (n ϭ 7) in the presence of 50 M (Ϫ)-D888 compared to 21 Ϯ 4% (n ϭ 12) under control conditions; Fig. 1B), and (iii) I Ba recovered from conditioned block in the presence of 50 M (Ϫ)-D888 by 95 Ϯ 3% (n ϭ 12) during a 3-min rest at Ϫ80 mV. As was the case with L h and L s , no frequency-dependent effect of (Ϯ)-emopamil was observed during trains of 100-ms test pulses in chimera AL23 (Fig. 1B).
Biophysical Properties of Chimera AL23-In 40 mM Ba 2ϩ solution I Ba of chimera AL23 activated at a threshold of Ϫ10 mV and reached peak current values between 10 and 20 mV (16 Ϯ 1.2 mV, n ϭ 26). The I Ba of ␣ 1A had a similar threshold as AL23 (Ϫ10 mV) and reached a peak current at 13 Ϯ 1.4 mV (n FIG. 2. I Ba recordings illustrating PAA block of ␣ 1A , L-type ␣ 1 chimeras and ␣ 1A /␣ 1S chimeras. A, I Ba through chimeras L h and L s during trains of 12 consecutive depolarizing voltage steps applied at 0.1 Hz in absence (control, left column) and presence (right column) of 50 M (Ϫ)-D888. The pulse lengths were 100 and 800 ms for L h and 800 ms for L s . B, I Ba through ␣ 1A and chimeric ␣ 1A /␣ 1S constructs AL22 and AL23 during 100-ms test pulse trains applied at 0.1 Hz. Recordings in the absence (control, left column) and presence (right column) of 50 M of (Ϫ)-D888 are illustrated. The decrease in I Ba in chimeras AL22 and AL23 in the absence of drug is caused by incomplete recovery from inactivation between the applied test pulses (compare Fig. 1B). Currents were recorded during depolarizations to 10 mV (A) and 15 mV (B) from a holding potential of Ϫ80 mV. ϭ 15). However, AL23 (as well as AL21 and AL22) inactivated faster than ␣ 1A (Fig. 2B). I Ba of AL23 decayed by 68 Ϯ 6% (n ϭ 12) during a 100-ms test pulse to 10 mV, whereas only 14 Ϯ 4% (n ϭ 10) of ␣ 1A current inactivated under identical experimental conditions. These data suggest a possible role of structural elements of segment IVS6 in inactivation of ␣ 1A . DISCUSSION To gain further insight into the molecular organization of the high affinity PAA binding domain, we investigated if the transmembrane domain IVS6 of a skeletal muscle ␣ 1S subunit (25) transfers L-type PAA sensitivity to a non-L-type channel ␣ 1 subunit. The ␣ 1A subunit, as a putative PAA-insensitive ␣ 1 , was selected for the following reasons. (i) With the exception of ␣ 1A and ␣ 1E , all other ␣ 1 subunits cloned so far (including ␣ 1B ) (9) have been shown or are considered to be (3) PAA-sensitive to various degrees. (ii) We recently demonstrated that DHP sensitivity can be transferred to ␣ 1A by introducing L-type channel (␣ 1C or ␣ 1S ) sequences into regions surrounding the channel pore (8). (iii) The segment IVS6 of ␣ 1A does not contain the high affinity PAA binding motif described by Hockerman et al. (9).
We found that ␣ 1A , when coexpressed in Xenopus oocytes together with ␤ 1a and ␣ 2 /␦, exhibited low sensitivity to PAAs. This is in contrast to DHP sensitivity, which is absent in ␣ 1A (8,26,27). The PAA sensitivity was, however, much smaller than that for L-type chimeras L h and L s (Fig. 1B). Analysis of PAA effects on several ␣ 1A /␣ 1S chimeras revealed that introduction of L-type (␣ 1S ) IVS6 was sufficient to increase PAA sensitivity to L-type levels, whereas replacement of the entire repeat IV in the chimera L s by ␣ 1A sequence decreased PAA sensitivity. Fig.  3 illustrates that segments IVS6 are largely conserved between L-type ␣ 1S and ␣ 1A and differ only in 11 positions. These include positions 1386, 1390, and 1393 of ␣ 1S (numbering according to Grabner et al. (12)) that are known to comprise critical determinants of PAA sensitivity in L-type channels (9). Our data demonstrate that segment IVS6 can participate in the formation of a PAA binding pocket in a non-L-type channel environment. The finding that the exchange of 11 amino acids residues within a single putative transmembrane helix (Fig. 3) is sufficient to increase ␣ 1A PAA sensitivity, differs from our previous observations with DHPs. To obtain a DHP agonist and antagonist sensitive chimera, L-type sequences had to be transferred into transmembrane segments IIIS5, IIIS6, and IVS6 as well as into the respective S5-S6 linkers (8). In contrast, the transfer of PAA sensitivity from L-type Ca 2ϩ channels into ␣ 1A did not require the introduction of other than the IVS6 se-quence. The low PAA sensitivity of ␣ 1A suggests that additional interaction sites are provided by ␣ 1A . Low PAA sensitivity was also observed for N-type Ca 2ϩ channels (␣ 1B ) and for L-type channels (␣ 1C ) lacking the high affinity determinants for PAA sensitivity in segment IVS6 (9). Therefore, additional regions of PAA interaction may be localized in sequence stretches conserved among these ␣ 1 subunits. Future studies will concentrate on the possible involvement of these regions in the interaction of Ca 2ϩ channels with PAAs.
As previously shown in mammalian cells, PAA block of I Ba in Xenopus oocytes is also dependent on the application of depolarizing test pulses (Fig. 2). PAAs are believed to interact selectively with the open Ca 2ϩ channel conformation (28,29) which complicates an estimation of drug association and dissociation rate constants (30). Inhibition of open Ca 2ϩ channels by PAAs is also supported by the observed acceleration of I Ba decay in chimeras L h and L s in the presence of PAAs (Fig. 2A). The extent of block induced by (Ϫ)-D888 (50 M) and (Ϯ)-D600 (100 M) during 12 pulses (100 ms) in chimera AL23 is comparable to the inhibition of the L-type chimeras L h and L s . Resting state-dependent block was almost absent, but pronounced inhibition developed during a train of depolarizing pulses. Furthermore, blocked I Ba recovered almost completely during a 3-min pulse-free interval in the presence of drug. Thus, the mechanism of Ca 2ϩ channel block in Xenopus oocytes appeared to be similar to PAA block of L-type channels in various mammalian cells (23,28,31). As previously observed, e.g. for DHPs (8) and Ca 2ϩ antagonist Ro 40-5967 (32), the effective drug concentrations for channel block after expression in Xenopus oocytes were higher than required in electrophysiological studies using mammalian cells (see Refs. 9 and 28).
An additional finding of our study was that introducing L-type sequence into ␣ 1A did not only enhance PAA sensitivity but also changed the inactivation kinetics of I Ba . Interestingly, the implantation of the transmembrane segment IVS6 from the slowly inactivating L-type chimera L S into ␣ 1A did not result in a transfer of the slower L-type inactivation kinetics into the faster inactivating ␣ 1A . Unexpectedly, this sequence substitution accelerated the inactivation kinetics compared to that of ␣ 1A . This finding gives an example where kinetic properties of Ca 2ϩ channels are not simply transposed by swapping corresponding sequences between different ␣ 1 subunits as was previously shown for structural elements of repeat I as well as III and IV (14,33,34). Our observation, that inactivation of ␣ 1A Ca 2ϩ channels is accelerated by changes in the amino acid sequence of segment IVS6 indicates a possible involvement of this region in inactivation gating in addition to its role in forming the PAA interaction domain.