Essential, Completely Conserved Glycine Residue in the Domain III S2–S3 Linker of Voltage-gated Calcium Channel α1 Subunits in Yeast and Mammals*

Voltage-gated Ca2+ channels (VGCCs) mediate the influx of Ca2+ that regulates many cellular events, and mutations in VGCC genes cause serious hereditary diseases in mammals. The yeast Saccharomyces cerevisiae has only one gene encoding the putative pore-forming α1 subunit of VGCC, CCH1. Here, we identify a cch1 allele producing a completely nonfunctional Cch1 protein with a Gly1265 to Glu substitution present in the domain III S2–S3 cytoplasmic linker. Comparison of amino acid sequences of this linker among 58 VGCC α1 subunits from 17 species reveals that a Gly residue whose position corresponds to that of the Cch1 Gly1265 is completely conserved from yeasts to humans. Systematic amino acid substitution analysis using 10 amino acids with different chemical and structural properties indicates that the Gly1265 is essential for Cch1 function because of the smallest residue volume. Replacement of the Gly959 residue of a rat brain Cav1.2 α1 subunit (rbCII), positionally corresponding to the yeast Cch1 Gly1265, with Glu, Ser, Lys, or Ala results in the loss of Ba2+ currents, as revealed by the patch clamp method. These results suggest that the Gly residue in the domain III S2–S3 linker is functionally indispensable from yeasts to mammals. Because the Gly residue has never been studied in any VGCC, these findings provide new insights into the structure-function relationships of VGCCs.

Voltage-gated Ca 2ϩ channels (VGCCs) 5 in the plasma membrane mediate the influx of Ca 2ϩ that serves as the second messenger of electrical signals to initiate many cellular events, including muscle contraction, neurotransmitter release, and gene expression (1). The pore-forming component of VGCCs is provided by the ␣ 1 subunit, a protein of about 2000 amino acid residues (2). This subunit contains four structurally conserved domains (I-IV), each of which contains six transmembrane segments (S1-S6) and a membrane-associated loop between S5 and S6 (called the pore loop or P-loop). The voltage sensitivity of VGCCs and structurally related cation channels is conveyed by the S4 segments, which contain several positively charged residues. S2 and S3 contain conserved negative charges that are likely to interact electrostatically with the positively charged residues of S4. The S5-P-loop-S6 region forms the pore domain (see Fig. 1).
Mutations in the ␣ 1 subunits resulting in structural aberrations cause hereditary diseases (called Ca 2ϩ channelopathies) in mammals, such as incomplete congenital stationary night blindness, hypokalemic periodic paralysis, episodic ataxia type 2 and familial hemiplegic migraine in humans, and ataxia and seizures in mice (3)(4)(5)(6). Most of the mutations are predicted to produce truncated ␣ 1 subunits with no significant channel activity by introducing a premature stop codon or leading to aberrant splicing. Single missense mutations can also result in the production of complete or partial loss-of-function subunits and these mutation sites are likely to be restricted to amino acid residues at transmembrane segments or loops known to be important for channel activity (7)(8)(9).
When exposed to the mating pheromone ␣-factor, mutants lacking CCH1 die because of a deficiency in Ca 2ϩ uptake (10,11). This phenotype, as well as the Ca 2ϩ uptake ability, was used to assess the activity of wild-type and mutant Cch1 proteins in this study. To gain channel activity, Cch1 requires another protein, Mid1, as revealed by genetic analyses (10,11,13). Mid1 has no structural homologue in higher eukaryotes (17), but has stretch-activated Ca 2ϩ channel activity when expressed in mammalian cells (18,19). There is no orthologue of genes encoding the auxiliary subunits ␣ 2 ␦, ␤, and ␥ of mammalian VGCCs in the S. cerevisiae genome.
rbCII is a member of an L-type VGCC ␣ 1 subunit subfamily, Ca v 1.2, isolated from the rat brain (20,21), and has been used in this study as a mammalian counterpart of yeast Cch1 to attempt to generalize findings with Cch1. The activity of rbCII has been successfully analyzed when it is transfected in baby hamster kidney BHK6 cells that stably express rabbit ␤ 1a and ␣ 2 /␦ necessary for efficient targeting of ␣ 1 to the plasma membrane (22,23).
Here, we report that a mutant allele of yeast CCH1, formerly designated mid3-1 (17), has a missense mutation causing a Gly 1265 to Glu substitution that results in a complete loss-offunction. The glycine residue is located in the cytoplasmic linker between the S2 and S3 segments in domain III, completely conserved during evolution, and proven to be important as being the smallest residue volume. In addition, we show that the Gly residue is not only important for Cch1 function, but also for rbCII channel activity. Because attention has never been focused on the Gly residue in this linker of any VGCC, our finding should help elucidate the structure-function relationships of VGCCs.
DNA Sequencing-The CCH1 gene of various strains was amplified by PCR with LA Taq polymerase (Takara Bio Inc.) using the primers 5Ј-200F-EcoKpn and 3Ј-305R-Sph, listed in Table 1. The resulting 6.6-kb DNA fragments were directly sequenced with BigDye Terminator version 3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA).
Construction of Plasmids-The plasmids used in this study were listed in Table 2. The promoter of CCH1 (P CCH1 , 200 bp), its ORF (CCH1 ORF, 6,120 bp), and the terminator (T CCH1 , 305 bp) were amplified by PCR with LA Taq polymerase using the genomic DNA of H207 or X2180-1A cells as a template. The PCR primers and an adaptor are listed in Table 1.
To construct pCCH1D, a pUC18-derived plasmid carrying P CCH1 and T CCH1 , the PCR-amplified promoter and terminator fragments of the CCH1 gene of H207 were cut with appropriate restriction enzymes and inserted into the multicloning site of pUC18, then an adaptor, which contains a stop codon and an XhoI site, was inserted between P CCH1 and T CCH1 . The CCH1 knock-out plasmid pCCH1DH was constructed by inserting the HIS3-containing 1.8-kb BamHI-XhoI fragment of pJJ215 (27) between P CCH1 and T CCH1 of pCCH1D.
To clone the CCH1 gene in Escherichia coli, we first constructed a new S. cerevisiae-E. coli shuttle vector (named pBC111), which replicates with a low copy number in E. coli, by replacing the ColE1 ori of YCplac111 (28) with the ColE1 ori and the rop gene of pBR322, as follows. pBR322 was cut with SalI and MscI and the overhangs were filled-in using a DNA blunting kit (Takara Bio Inc.). The larger fragment was selfligated and the SalI site that remained was destroyed by a series of SalI-cut, fill-in, and self-ligation. The 2.9-kb AatII-SphI fragment of the resulting plasmid was ligated with the 3.9-kb SphI-AatII fragment of YCplac111 to produce pBC111. pBCS111 and pBCT111 are derivatives of pBC111, each of which carries P CCH1 -T CCH1 or P TDH3 -T ADH1 , respectively. pBCS111 was constructed by inserting the 0.5-kb EcoRI-SphI fragment of pCCH1D into the multicloning site of pBC111. To construct pBCT111, P CCH1 and T CCH1 of pBCS111 were replaced with the P TDH3 -containing 0.7-kb EcoRI-BamHI fragment of pUGPD (29) and the T ADH1 -containing 0.3-kb PstI-SphI fragment of pGBKT7 (Clontech, Palo Alto, CA), respectively. The PCR-amplified CCH1 ORF derived from H207 or H3031 was cut with BamHI and SalI and inserted between the BamHI and SalI sites of pBCS111 or pBCT111. The resulting plasmids were designated pBCS-CCH1H or pBCS-CCH1Hm1 and pBCT-CCH1H or pBCT-CCH1Hm1, respectively. Recombinant Cch1 proteins expressed from these plasmids have three additional amino acid residues (Val, Asp, and Thr) at the carboxyl terminus. The plasmids, pBCT-CCH1H-EGFP or pBCT-CCH1Hm1-EGFP, used to express EGFP-tagged Cch1 proteins were constructed by inserting the 0.6-kb NcoI (blunted)-NotI fragment of pEGFP (Clontech) between the SalI (blunted) and NotI sites of pBCT-CCH1H or pBCT-CCH1Hm1, respectively.
pBCMS-EGFP, a low-copy derivative of the mammalian expression plasmid pCMS-EGFP (Clontech) was constructed as follows. The restriction sites spanning from SphI to XbaI of the multicloning site of pBC111 were deleted, and the 2.9-kb AatII (blunted)-BamHI fragment of the resulting plasmid was inserted between the BglII and ApaLI (blunted) sites of pCMS-EGFP. rbCII cDNA (kindly gifted by Dr. T. P. Snutch; see Refs. 20 and 21) was inserted between the MluI and SalI sites of the multicloning site of pBCMS-EGFP. The resulting plasmid, pBCMS-EGFP-rbCII, was used to transiently express rbCII under the control of the immediateearly promoter/enhancer of cytomegalovirus (CMV promoter) in BHK6 cells.
In Vitro Site-directed Mutagenesis-In vitro site-directed mutagenesis was performed using the two-step PCR method reported by Higuchi et al. (30). The PCR primers are listed in Table 1. pBCT111 plasmids bearing various cch1 mutant genes were constructed by replacing the 0.4-kb KpnI-ApaI fragment of pBCT-CCH1H with the corresponding fragment that had been produced by PCR-based, in vitro sitedirected mutagenesis, and the nucleotide sequence was verified. pBCS111 plasmids bearing the same cch1 mutant genes were constructed by replacing the 4.1-kb SacI-SalI fragment of pBCS-CCH1H with the corresponding fragment of the pBCT111 bearing the cch1 mutant genes constructed as described above.
In the case of rbCII, QuikChange II XL Site-directed Mutagenesis Kit (Stratagene, La Jolla, CA) was used to mutagenize the rbCII cDNA subcloned in pBluescript II SK(Ϫ). The 1.4-kb SpeI-EcoRV fragment of the wild type cDNA was replaced with each mutated fragment, and each

Oligonucleotides used in this study
Underlines represent the entirety or a part of restriction sites. Boxed triplets show the initiation or termination codon. a Asterisk, The G1265S-F and G1265S-R were used to substitute Ser for the Gly 1265 of Cch1, and the G959E-F and G959E-R were used to substitute Glu for the Gly 959 of rbCII, for example. The doubly underlined triplets in these oligonucleotides were changed to appropriate nucleotides to substitute the corresponding amino acid for Gly 1265 or Gly 959 . Boldfaces indicates the changed nucleotides.
This study pBCMS-EGFP a ColE1-ori-rop P CMV EGFP marker This study pBCMS-EGFP-rbCII rbCII cDNA in pBCMS-EGFP This study a The ColE1 ori of YCplac111 and pCMS-EGFP have been replaced with a DNA fragment containing the ColE1 ori and the rop gene derived from pBR322. b The suffixes of CCH1, i.e. H and Hm1, represent the CCH1 and cch1-1 ORFs derived from the strains H207 (CCH1) and H3031 (cch1-1), respectively. AUGUST 31, 2007 • VOLUME 282 • NUMBER 35

Essential Glycine Residue in Voltage-gated Calcium Channels
JOURNAL OF BIOLOGICAL CHEMISTRY 25661 mutated full-length cDNA was introduced into pBCMS-EGFP after the nucleotide sequence was verified.
Preparation of Polyclonal Anti-Cch1 Antibodies-A PCR fragment encoding the Cch1 carboxyl-terminal peptide spanning from amino acid residue 1,949 to 2,039 was cut with BamHI and SalI and inserted into pQE30 (Qiagen Inc., Valencia, CA) to be conjugated with a His 6 tag at the amino terminus. The His 6 -tagged carboxyl-terminal peptide was purified from an E. coli transformant (strain JM109) carrying this plasmid using nickel-nitrilotriacetic acid-agarose beads (Qiagen Inc.) under denaturing conditions. Affinity purified rabbit antibodies against this peptide were prepared by Promega Co. (Madison, WI).
Western Blotting and Fluorescence Microscopy-Western blotting was carried out according to the method described previously (17) except that SDS-PAGE samples were heated for 3 min at 70°C. The affinity purified rabbit polyclonal antibodies against the carboxyl-terminal peptide described above were used at a concentration of 0.07 g/ml to detect the wild-type and various mutant forms of the Cch1 protein.
Fluorescence microscopy on cells expressing the Cch1G1265E-EGFP and Cch1-EGFP was performed as described previously (31).
Determination of the Viability of Yeast Cells-The viability of cells exposed to 6 M ␣-factor in SD.Ca100 medium was determined using the methylene blue liquid method (24).
Determination of Ca 2ϩ Accumulation in Yeast Cells-Exponentially growing cells in SD.Ca100 medium were incubated for 2 h with 6 M ␣-factor and 185 kBq/ml (1.85 kBq/nmol) of 45 CaCl 2 (PerkinElmer Life Sciences). Samples were taken, filtered through Millipore filters (type HA; 0.45 m) that had been presoaked in 5 mM CaCl 2 , and washed five times with the same solution. The radioactivity retained on the filters was counted with a scintillation mixture, ReadyProtein (Beckman Japan, Tokyo), in a liquid scintillation counter.
Electrophysiological Recordings-Electrophysiological recordings were performed in the whole cell patch clamp configuration using Patch/Whole Cell Clamp Amplifier Axopatch 200B (Axon Instruments, Foster City, CA) and A/D converter (Digidata 1200, Axon Instruments). Data acquisition was performed by using pCLAMP7 software (Axon Instruments). Capacitative components were electrically compensated. Leak subtraction was performed with P/N method (N ϭ 4 -6). All experiments were performed at room temperature.
Statistical Analysis-Statistical significance was determined using the unpaired Student's t test with a p value Ͻ0.05 required for significance.

RESULTS
Genetic Characterization of a cch1 Allele of S. cerevisiae-The phenotypes of the previously reported mid3-1 recessive mutant are essentially the same as those of the cch1⌬ mutant: these include mating pheromone-induced death, low Ca 2ϩ uptake activity, and low viability in the stationary phase. In addition, the corresponding genes have not been cloned despite numerous trials (11,13). These features prompted us to perform genetic analysis on them. The two mutants were crossed (MATa mid3-1 ϫ MAT␣ cch1⌬) and the resulting diploids showed low viability in the stationary phase, like mid3-1 and cch1⌬ haploids, suggesting that the two alleles belong to the same complementation group. The diploids were sporulated and subjected to tetrad analysis (32). All cells germinated from 72 spores dissected from 18 asci again showed low viability in the stationary phase. In addition, all the MATa cells from each spore (a total of 36 spores) showed the mating pheromone-induced death phenotype. These results indicate that mid3-1 is allelic to cch1⌬. We therefore renamed mid3-1 to cch1-1.
PCR products, synthesized using genomic DNA as a template of the cch1-1 mutant (strain H3031) as well as of the parental strain H207 (CCH1), were directly sequenced. Three independent PCR products for each strain were sequenced to avoid misinterpretation due to possible PCR errors. Consequently, we identified a missense mutation (G3794A) in cch1-1, which results in a substitution of glutamate (Glu) for glycine (Gly) at codon 1,265 (G1265E) ( Table 3).
We noted a strain-dependent polymorphism in CCH1. Seven nucleotides and six encoded amino acids in CCH1 of the strain

Essential Glycine Residue in Voltage-gated Calcium Channels
H207 we determined were different from those of CCH1 of the strain S288C listed in the Saccharomyces Genome Data base (www.yeastgenome.org) ( Table 3). Because S288C is isogenic to a standard laboratory yeast strain, X2180-1A, we determined the CCH1 nucleotide sequence of X2180-1A and four other standard laboratory strains and found that H207 shares the same nucleotide sequence with one strain and that X2180-1A and S288C share the same sequence with three other strains (Table 3). It is notable that the viability of X2180-1A cells was as high as that of H207 cells after exposure to ␣-factor (data not shown), suggesting that the six amino acid variations do not affect Cch1 function.
To confirm that the G1265E mutation is the only determinant for the phenotypes of the cch1-1 mutant, we first developed a new vector of low copy number in E. coli (designated pBC111) because cloning of CCH1 has long been impossible due to the probable toxicity of this gene or protein for E. coli (11,13). pBC111 is an E. coli-yeast shuttle vector having the ColE1-ori-rop sequence ensuring 15-20 copies per E. coli cell and an S. cerevisiae centromere ensuring 1-3 copies per S. cerevisiae cell. Using this, we generated the G1265E mutation in CCH1 cloned from X2180-1A and the resulting mutant gene was tested for its ability to rescue the mating pheromone-induced death phenotype and low Ca 2ϩ uptake activity of the cch1⌬ mutant. The results showed that this mutant gene did not rescue the two phenotypes at all, just like the G1265E mutant gene of H207 (data not shown), indicating that the two phenotypes are attributed to the G1265E mutation only. We therefore used the CCH1 alleles based on the H207 background hereafter and designated the mutant Cch1 protein with the G1265E mutation Cch1G1265E.
Gly 1265 Is Completely Conserved-Based on a conventional transmembrane domain prediction (33), Gly 1265 was predicted to be located in the cytoplasmic linker connecting the putative transmembrane segments S2 and S3 of domain III (Fig. 1), although this residue is alternatively predicted to be included in the S2 segment by other methods, such as SGD (www.yeastgenome.org) and UniProt (www.expasy.uniprot.org). Here, we followed the conventional prediction (33). To explore the importance of the glycine residue in the domain III S2-S3 linker, we compared the amino acid sequences of the corresponding linker of the VGCC family. Fig. 2 shows a multiple amino acid sequence alignment of the linker and its neighbors of a total of 58 VGCCs and candidates from 17 species, including humans, yeasts, and representative model organisms, such as mice, rats, zebrafish, fruit flies, and nematodes. This alignment clearly indicates that the glycine residue is completely conserved from yeasts to humans. The finding prompted us to investigate the importance of the Gly 1265 of Cch1 because the glycine residue has never been investigated in any VGCC or candidate, although it is remarkably conserved during evolution.
The G1265E Mutation Results in a Complete Loss of Function-There are three possibilities for the inability of the Cch1G1265E protein in cellular function: loss-of-function, instability, and lack of plasma membrane targeting. To examine these possibilities, we first investigated the amount of the Cch1 and Cch1G1265E proteins by Western blotting and found that the amount of Cch1G1265E produced from the cch1-1 gene was slightly smaller than that of the wildtype Cch1 protein (Fig. 3A, left panel). However, this decrease is unlikely to be the cause of the defect of Cch1G1265E because an ϳ20-fold overproduction of Cch1G1265E under the control of a strong TDH3 promoter (Fig. 3A, left panel) did not rescue the decreased viability and Ca 2ϩ uptake activity of the cch1⌬ mutant at all (Fig. 3, B and  C). Thus, a putative instability of the Cch1G1265E protein does not account for its inability in cellular function.
We then examined the subcellular localization of Cch1G1265E, tagged with green fluorescence protein (Cch1G1265E-EGFP) produced under the TDH3 promoter on a centromere plasmid, using fluorescence microscopy and found that Cch1G1265E-EGFP was normally localized to the plasma membrane and the endoplasmic reticulum membranelike structure (Fig. 3D). Note that the amount of Cch1-EGFP and Cch1G1265E-EGFP produced under the self-promoter on a centromere plasmid was too small to visualize their fluorescence images (13).
We also found that the 20-fold overproduction of Cch1G1265E in the CCH1 strain (Fig. 3A, right panel) resulted in only a small decrease in both the viability and Ca 2ϩ uptake activity (Fig. 3, E and F). This result is consistent with the genetic data that the cch1-1 (mid3-1) mutation is recessive (17) and indicates that the interferential effect of Cch1G1265E on wild-type Cch1 is small. Taken together, we conclude that the G1265E mutation results in a complete loss-of-function of Cch1, not its instability and mislocalization.
Importance of Gly 1265 as the Smallest Amino Acid Residue-We next investigated the cause of the complete loss-of-function of the Cch1G1265E protein. One can speculate three possibilities for the cause: a negative charge, hydrophilicity, or a large residue volume of glutamate compared with glycine. To examine these possibilities, we systematically substituted nine amino acids with different chemical and physical properties for Gly 1265 , introduced the resulting mutant genes on a centromere plasmid into the cch1⌬ mutant, and examined their activity by cell viability and Ca 2ϩ uptake assays (Fig. 4).  Table 3. Roman numerals I-IV represent putative domains and the Arabic numerals 1-6 putative transmembrane segments (S1-S6). The putative S4 segments in domains I-III, containing repeated motifs of a positively charged residue followed by two hydrophobic residues, are colored yellow. The putative S4 segment of domain IV lacks the motifs. P represents the putative pore loop. The amino and carboxyl termini are located in the cytoplasm.
We have confirmed, with Western blotting, that the amount of the mutant proteins was comparable with that of the wildtype Cch1 when they were produced from the self-promoter, and that it was at least 15-fold greater than that of the wild-type Cch1 when produced from the TDH3 promoter (data not shown).
When the activities of the mutant proteins produced under the self-promoter were compared, the Gly-Ser and Gly-Ala substitutions reduced the Ca 2ϩ uptake activity by ϳ30%, but did not affect cell viability (Fig. 4, A and  D), indicating that the remaining ϳ70% Ca 2ϩ uptake activity is sufficiently high to support the viability of cells. The Gly-Asp substitution slightly reduced both the cell viability and Ca 2ϩ uptake activity. The Gly-Asn substitution inactivated the Ca 2ϩ uptake activity, whereas it still maintained slight viability. Finally, the Gly-Thr, Gly-Val, Gly-Gln, Gly-Leu, and Gly-Lys substitutions resulted in the complete loss of Cch1 function. It is notable that when these mutant proteins were produced under the TDH3 promoter, the viability and Ca 2ϩ uptake activity were only slightly increased in the partially active mutant proteins and were not increased at all in the completely inactivated mutant proteins (Fig. 4, A and D), suggesting that a decrease in the amount of the mutant proteins is not a factor for the decrease in the viability and Ca 2ϩ uptake activity.
When those results obtained with the self-promoter were re-plotted in terms of the residue volume or hydrophobicity of the amino acids, we noted a clear relationship between the residue volume and activity of the mutant proteins: Fig. 4, B and E, show that when the residue volume is ϳ115 cubic Å or greater at the position of Gly 1265 , the Cch1 protein completely loses its activity. By contrast, Fig. 4, C and F, show that there is no relationship between hydrophobicity and Cch1 function. We therefore conclude that Gly 1265 is important for Cch1 activity because of the smallest residue volume. Finally, we note that among the Gly-Asp, Gly-Asn, and Gly-Thr substitutions, three of which result in almost the same residue volume, the more the residue is hydrophilic, the higher the activity of the Cch1 product (Fig.  4, C and F).
Gly 959 of Rat Ca v 1.2 Is Essential-On the basis of the above findings that the Gly 1265 of Cch1 is essential and that the Gly residue positionally corresponding to Gly 1265 is completely conserved during evolution, we examined the possibility that the Gly residue of mammalian VGCC ␣ 1 subunits is essential for their channel function. To do this, the Gly 959 residue of the rat Ca v 1.2 (rbCII), which is the positional counterpart of the Gly 1265 of Cch1, was replaced with Glu, Ser, Lys, or Ala. The resulting mutant rbCII proteins were expressed under the control of the CMV promoter in BHK6 cells expressing rabbit ␤ 1a and ␣ 2 /␦ subunits and examined for their function and subcellular localization.
Whole cell patch clamp recordings showed that the rbCIIG959E protein had no channel activity at all and that the rbCIIG959S protein had little activity (Fig. 5). The rbCIIG959K and rbCIIG959A proteins had no activity at all (data not shown).
Confocal fluorescent microscopy with antibodies specific to rbCII showed that a significant fraction of the wild-type rbCII protein, used as a control, was localized to the plasma mem-brane in BHK6 cells (supplemental materials Fig. S1). However, some fraction appeared to remain in the cytoplasm. Almost all fractions of the rbCIIG959E, rbCIIG959K, and rbCIIG959A proteins were mainly present in the cytoplasm. By contrast, the rbCIIG959S protein was completely localized to the plasma membrane. These results indicate that the Gly-Ser substitution in rbCII results in marked attenuation of channel activity without altering the localization efficiency of the protein. The Gly-Glu, Gly-Lys, and Gly-Ala substitutions bring about a complete loss of channel activity, and this loss may be due to a decreased efficiency of protein transport to the plasma membrane.

DISCUSSION
Here we have shown the importance of the Gly residue present in the domain III S2-S3 linker of the yeast putative VGCC ␣ 1 subunit Cch1 and of the rat VGCC ␣ 1 subunit rbCII, a member of Ca v 1.2. Based upon a conventional transmembrane prediction (33), the domain III S2-S3 linker of VGCCs is composed of 13 amino acid residues and the Gly residue is the only one amino acid residue that is completely conserved in the linker (Fig. 2). This suggests that the Gly residue is indispensable for the function of VGCCs, and we have proven this suggestion.
Our study on the Cch1G1265E protein have indicated that Gly 1265 is required for the function of Cch1, but probably not for the stability and subcellular localization of this protein because the amount of Cch1G1265E and the localization of Cch1G1265E-EGFP are essentially the same as those of the wild-type counterparts (Fig. 3, A and D). The reason for the importance of the Gly residue is that it has the smallest side chain, as revealed by systematic amino acid substitution experiments (Fig. 4).
As for rat rbCII, on the other hand, Gly 959 may be important for not only function but also localization. The rbCIIG959E, rbCIIG959S, rbCIIG959K, and rbCIIG959A proteins have no or very little Ba 2ϩ current, as revealed by patch clamp analysis. Indirect fluorescent immunomicroscopy has shown that, among these mutant proteins, the majority of the rbCIIG959E, rbCIIG959K, and rbCIIG959A proteins seem to be localized in the cytoplasm, instead of localization to the plasma membrane (supplemental materials Fig. S1). In addition, many of even the wild-type rbCII protein appears to remain in the cytoplasm. We speculate FIGURE 3. Characterization of the cch1-1 mutation. A, Western blotting of the Cch1G1265E protein produced from the cch1-1 gene on a plasmid. Left panel, the H317 strain (cch1⌬) was transformed with various plasmids and the extracts of the resulting transformants were subjected to Western blotting. Plasmids used here: vector, pBC111; CCH1, pBCS-CCH1H; G1265E, pBCS-CCH1Hm1; G1265Eox, pBCT-CCH1Hm1. Right panel, the parental strain H207 (CCH1) was transformed and treated as above. Plasmids used here: vector, pBC111; G1265Eox, pBCT-CCH1Hm1; CCH1ox, pBCT-CCH1H. Note that the samples were appropriately diluted as indicated under the panels. B, the viability of cells incubated with 6 M ␣-factor for 8 h in SD.Ca100 medium. The transformants used here were the same as the left panel of A. C, Ca 2ϩ accumulation of cells during a 2-h incubation with 6 M ␣-factor in SD.Ca100 medium. The transformants used here were the same as in B. D, subcellular localization of Cch1G1265E-EGFP. Exponentially growing cells of cch1⌬/pBCT-CCH1H-EGFP (left panels) and cch1⌬/pBCT-CCH1Hm1-EGFP (right panels) in SD.Ca100 medium were observed by confocal fluorescence microscopy (upper panels) or differential interference contrast microscopy (lower panels). Essentially the same results were obtained for two additional experiments. E, viability of cells incubated with 6 M ␣-factor for 8 h in SD.Ca100 medium. The transformants used here were the same as the right panel of A. Vector versus G1265E (p Ͻ 0.05). F, Ca 2ϩ accumulation of cells during a 2-h incubation with 6 M ␣-factor in SD.Ca100 medium. The transformants used were the same as those in E. Vector versus G1265E (p Ͻ 0.05). In B, C, E, and F, all the data are mean Ϯ S.D. from at least three independent experiments. that overexpression of rbCII mRNA driven by the strong CMV promoter might have resulted in the mislocalization of these proteins because of excess in the amount of the ␣ 1 subunit over that of the ␤ and ␣ 2 /␦ subunits. It is known that the correct targeting of the ␣ 1 subunit to the plasma membrane requires the ␤ subunit (23). By contrast, rbCIIG959S seems to be efficiently transported to the plasma membrane. Taken together, it is likely that Gly 959 is necessary for the function and localization of rbCII.
From a structural viewpoint of voltage-gated ion channels, we speculate that the Gly residue has a role for the formation of a voltage-sensing structure composed of the S1-S4 segments. Among six transmembrane segments, the S1-S4 segments are considered to function as the voltage sensor of structurally related, voltage-gated Ca 2ϩ , K ϩ , and Na ϩ channels (2). In the case of K ϩ channels whose structure is best elucidated, the four most extracellularly located basic residues of the S4 segment and the most intracellular acidic residue in the S2 segment are the major contributors to the gating charge movement that is coupled with the opening or closing of the pore of the channel (34,35). In addition, according to the helixpacking model, the interaction of conserved negative charges in S2 and S3 with the conserved positive charges in S4 makes S2 and S3 line one side of the gating canal (36,37). Thus, it is likely that the structure of the S2-S3 linker is also important for channel gating and the replacement of the Gly residue with any larger amino acid perturbs the spatial arrangement of S2 and S3. This perturbation may also affect the localization efficiency of VGCC ␣ 1 subunits, as seen in rbCII.
Finally, we would like to point out that the Gly residue of the S2-S3 linker is present not only in domain III but also in domains I, II, and IV with a slight exception (data not FIGURE 4. Effect of the substitution of various amino acids for Gly 1265 on Cch1 function. Cells of the cch1⌬ mutant expressing a Cch1 protein having a mutation in Gly 1265 from the self-promoter or TDH3 promoter were grown in SD.Ca100 medium, exposed to ␣-factor, and examined for viability and Ca 2ϩ accumulation. A, the viability of cells was determined 8 h after exposure to ␣-factor. Open and closed bars represent the viability of cells expressing a mutant Cch1 protein produced from the self-promoter and the TDH3 promoter, respectively. Mean Ϯ S.D. from at least three independent experiments. Plasmids used: vector, pBC111; CCH1 (G), pBCS-CCH1H; single letters from Glu to Lys represent the substituted amino acids at Gly 1265 . Amino acid abbreviation: G, Gly; E, Glu; A, Ala; S, Ser; D, Asp; N, Asn; T, Thr; V, Val; Q, Gln; L, Leu; K, Lys. B, replot of the viability data shown in A as a function of the volume of amino acid residues (38). The data based on the self-promoter are shown. C, replot of the viability data shown in A as a function of the hydrophobicity of amino acid residues (39). The data based on the self-promoter are shown. D, Ca 2ϩ accumulation was determined 2 h after the cells were exposed to ␣-factor. Open and closed bars represent the data obtained based on the self-promoter and the TDH3 promoter, respectively. Mean Ϯ S.D. from at least three independent experiments. Plasmids used and amino acid abbreviations were the same as those in A. E, replot of the Ca 2ϩ accumulation data shown in D as a function of the volume of amino acid residues. The data based on the selfpromoter are shown. F, replot of the Ca 2ϩ accumulation data shown in D as a function of the hydrophobicity of amino acid residues. The data based on the self-promoter are shown. FIGURE 5. Effects of substitution of various amino acids for Gly 959 on the Ca 2؉ channel activity of rbCII. Whole cell recordings of Ba 2ϩ currents through the L-type Ca 2ϩ channel rbCII and its mutant proteins expressed in BHK6 cells were shown. A, wild-type rbCII; B, the rbCIIG959E protein; C, the rbCIIG959S protein. The dysfunction of G959E could be due to the impairment of plasma membrane targeting (see supplemental materials Fig. S1). G959S showed markedly small Ba 2ϩ currents, although it was localized to the plasma membrane (see supplemental Fig. S1). D, current-voltage relationships (I-V curve) of wild-type and mutant rbCIIs (G959E and G959S).
shown). Therefore, the Gly residue should have an essential function in all of the four domains.