Ile-177 and Ser-180 in the S1 Segment Are Critically Important in Kv1.1 Channel Function*

Ile-177 and Ser-180 are conserved residues in the first transmembrane segment (S1) of theShaker, Shab, Shaw, and Shalsubfamilies of voltage-gated K+ channels. Here we report that the mutation of these residues in Kv1.1 to leucine, proline, or arginine abolished the expression of outward potassium currents inXenopus oocytes. Co-injection of these mutant cRNAs and wild type Kv1.1 cRNA into Xenopus oocytes exerted a potent dominant negative effect resulting in the suppression of Kv1.1-encoded currents. Transient transfection experiments of COS-7 cells revealed that the S1 mutants directed the synthesis of Kv1.1 polypeptides. Quantitative co-immunoprecipitation assays revealed that most of the S1 mutants co-assembled and formed both homo- and heteromultimeric complexes. Furthermore, the mutated polypeptides could reach the plasma membranes of transfected Sol8 cells. We conclude that mutations of Ile-177 and Ser-180 do not interfere with either the assembly of multimeric channel complexes or the targeting of these complexes to the plasma membrane. It is likely that these residues are involved in helix-helix interactions that are critical to the proper functioning of voltage-gated potassium channels.

Ile-177 and Ser-180 are conserved residues in the first transmembrane segment (S1) of the Shaker, Shab, Shaw, and Shal subfamilies of voltage-gated K ؉ channels. Here we report that the mutation of these residues in Kv1.1 to leucine, proline, or arginine abolished the expression of outward potassium currents in Xenopus oocytes. Co-injection of these mutant cRNAs and wild type Kv1.1 cRNA into Xenopus oocytes exerted a potent dominant negative effect resulting in the suppression of Kv1.1-encoded currents. Transient transfection experiments of COS-7 cells revealed that the S1 mutants directed the synthesis of Kv1.1 polypeptides. Quantitative co-immunoprecipitation assays revealed that most of the S1 mutants coassembled and formed both homo-and heteromultimeric complexes. Furthermore, the mutated polypeptides could reach the plasma membranes of transfected Sol8 cells. We conclude that mutations of Ile-177 and Ser-180 do not interfere with either the assembly of multimeric channel complexes or the targeting of these complexes to the plasma membrane. It is likely that these residues are involved in helix-helix interactions that are critical to the proper functioning of voltage-gated potassium channels.
Voltage-gated potassium channels, the largest and the most diverse group of ion channels, play a central role in the propagation of signals and the determination of cellular excitability (1). At least four subfamilies of voltage-gated K ϩ channels have been identified. These subfamilies encode the Shaker (Kv1), Shab (Kv2), Shaw (Kv3), and Shal (Kv4) channel polypeptides and their mammalian homologues, which are highly conserved across species. Each channel is synthesized as a monomeric ␣ subunit, which assembles into a pore-forming tetrameric channel. These channels share several common architectural designs, such as a conserved core domain which is comprised of six transmembrane segments, and a teteramerization domain (T1 or NAB) at the amino terminus (2).
During the past several years, we have learned a great deal about the rules governing the assembly and multimerization of voltage-gated potassium channels. It is now established that the co-assembly of monomeric subunits occurs primarily within the same subfamily of Kv channels, resulting in the formation of either homo-or heterotetrameric complexes (3-7). The do-main that determines the specificity of subunit interactions consists of about 114 amino acids and is located at the NH 2 terminus. This domain is referred to as the tetramerization (T1) or NAB domain (3-7). Other regions in the central core may also be involved in channel co-assembly (8).
The role of the first transmembrane segment (S1) in the co-assembly of the Shaker-related and Aplysia K ϩ channels was first characterized in our report and the reports of others (9 -11). These reports were based on deletion analyses of these channels studied by several different approaches, including dominant negative expression assays, co-immunoprecipitation assays, and hydrodynamic analysis of co-assembled complexes (9 -11). It was hypothesized that the S1 segment played a direct allosteric role in stabilizing the assembly of the functional channel complexes (9,10). The S1 segments of all four subfamilies have two completely conserved amino acids, which in Kv1.1 correspond to Ile-177 and Ser-180. Here we examine the role of these residues in Kv1.1. Our results provide evidence that rather than playing a role in the assembly and targeting to the plasma membrane, these residues are directly involved in interactions that are critical to the proper functioning of Kv1.1 channels.

MATERIALS AND METHODS
Preparation and Quantitation of cRNA-Capped T7 run-off transcripts of gel-purified ApaI linearized templates were prepared using mMessage mMachine kit (Ambion). Accurate quantitation of the cRNAs and their integrity were examined using spectrophotometry, formaldehyde-agarose gel electrophoresis, and autoradiography for which [␣-32 P]UTP (NEN) was added as a tracer in the reactions.
Two-electrode Voltage Clamp of Xenopus Oocytes-The isolation and preparation of the oocytes was carried out as described by Stü hmer and Parekh (12). For co-injections, a fixed amount of the wild type cRNA (0.69 ng) was mixed with the mutant cRNA at 1:1 and 1:3 ratios. Whole oocyte currents were recorded 2-5 days post-injection by a two-microelectrode voltage clamp using a GeneClamp 500 amplifier (Axon Instruments, CA). The bath solution contained (in mM): NaCl 96, KCl 2, CaCl 2 0.5, MgCl 2 0.5, and HEPES 10, pH 7.5. The oocytes were held at Ϫ70 mV (in some experiments, Ϫ80 mV) and the currents were elicited by a series of 200-ms depolarization pulses from the holding potential to ϩ50 mV in 10 mV increments followed by a repolarization to 50 mV (in some experiments, Ϫ60 mV). A P/4 subtraction protocol was used to minimize the leakage and the capacitative currents (13). Data were expressed as mean Ϯ S.E. Student's t test was used to evaluate the statistical significance.
Binomial Distribution and Its Use to Determine the Potency of Dominant Negative Suppression-The fraction of the current carried by the homo-and heterotetrameric channel complexes (n ϭ 4) containing mutant subunits (i ϭ 0, 1, 2, 3, or 4) can be calculated using the binomial equation (14), The total amount of cRNA injected at 1:1 and 1:3 ratios of wild type to mutants is increased 2-and 4-fold, respectively; therefore, more channel complexes are expected to form. The amplitudes of the resultant currents would be expected to be greater than that of the wild type * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  alone, if the subunits do not exert dominant negative suppression. The percentage of the current will be the sum of the conducting complexes and can be predicted if one, two, or three subunits are required to block the channels (15,16).
Construction of FL-Kv1.1, FL-Kv1.1Ile177, FL-Kv1.1Ser180 Mutants, and TR-Kv1.1HA-Kv1.1 cDNA cloned from rat soleus muscle (17) was the starting template for creating the constructs. It was amplified by polymerase chain reaction using the T7 primer and a redundant reverse primer carrying Arg, Leu, and Pro codons at either Ile-177 or Ser-180 in the S1 segment and a silent mutation to introduce an XhoI site. The polymerase chain reaction product was subcloned and sequenced. An XhoI/KpnI fragment from Kv1.1 cDNA was added to each mutant to generate the full-length channel construct in Bluescript II vector (Stratagene) (FL-Kv1.1). Each of these Kv1.1Ile177 and Kv1.1Ser180 mutants was digested with KpnI enzyme, blunt-ended with T4 DNA polymerase, and then digested with BstEII enzyme. These fragments were then used to generate the FLAG epitope-tagged Kv1.1 (FL-Kv1.1) and the FLAG epitope-tagged mutants (FL-Kv1.1Ile177 or FL-Kv1.1Ser180) in pCDNA3 (11). We used the gene SOEing method to construct the truncated Kv1.1 containing the NH 2 -terminal 205 amino acid residues tagged with three copies of the HA epitope (TR-Kv1.1HA) (18).
Transient Transfections in COS-7 Cells-COS-7 cells were grown to about 70% confluency in Petri dishes (100 mm). LipofectAMINE reagent (30 l/10 g of DNA) was used in Opti-MEM medium (Life Technologies, Inc.) for 24 h and then changed to the complete medium for a total period of 66 -72 h (11). The cells were washed with 2 ϫ 5 ml of PBS 1 and scrapped in a total 1.5 ml of chilled PBS on ice. For cotranslation and co-assembly analysis, one-half of the cells transfected with either the FL-Kv1.1 or the TR-Kv1.1HA were mixed at this stage. The cells were pelleted and stored at Ϫ70°C until used.
Cell Lysis, Immunoprecipitation, and Immunoblotting-The cells were lysed for 5 min on ice in 1 ml of lysis buffer (mM final concentration): NaCl 150; Tris/HCl, pH 8.0, 10; EDTA 0.5; iodoacetamide 1; phenylmethylsulfonyl fluoride 0.5; 1% Triton X-100 and 4 g each of pepstatin A, chymostatin, leupeptin, antipain, and bestatin (Sigma). Equal amounts of protein (19) were used for co-immunoprecipitation in three sequential steps. In the first step, 4 g of anti-HA antibody (12CA5, Boehringher) was added (overnight at 4°C). Protein G-agarose slurry was added, mixed (4°C/4 h), and centrifuged. The pellets were then washed with lysis buffer. Supernatants from the first step were used for the second round of immunoprecipitation with 12CA5 antibody exactly as described above. In the final step, supernatants from the second step were used for immunoprecipitation with M2 anti-FLAG antibody (19 g) (Eastman). The precipitates were electrophoresed in 10% polyacrylamide-SDS gel, blotted onto polyvinylidene difluoride membranes, and probed with the anti-FLAG antibody. After washing with PBS, the blots were incubated with horseradish peroxidase-conjugated goat anti-mouse IgG (Zymed Laboratories Inc.) for 1 h at room temperature. The protein bands were detected by using chemiluminescence. After stripping, the blots were re-probed with anti-HA antibody (12CA5). The luminograms were scanned using a Hewlett-Packard Scan Jet 4C/T scanner at 600 disintegrations/min (sharp black and white photo). The images were analyzed and the bands of interest were quantitated as integrated densities by Scion Image program. The values obtained were linear at the exposure intervals we used.
Immunocytochemistry and Laser-scanning Confocal Microscopy-The cells were grown on coverslips in 6-well tissue culture dishes. For co-transfections, 1 g of each cDNA was used. The cells were washed (3 ϫ 1.5 ml of PBS with Ca 2ϩ and Mg 2ϩ ), fixed with 2% paraformaldehyde, permeabilized with 0.2% Triton X-100 in PBS and reacted with the anti-FLAG antibody (M2) (1:1000) (20). The secondary antibody used for staining was the fluorescein isothiocyanate-conjugated goat anti-mouse IgG (1:100) (Jackson ImmunoResearch). The cells were visualized by laser-scanning confocal microscopy (MRC1024, Bio-Rad). Fig. 1A shows the primary structure of the first transmembrane segment (S1) of Kv1.1 and the mammalian homologues of four voltage-gated K ϩ channels, Shaker, Shaw, Shab, and Shal. The wild type Kv1.1 polypeptide used in this study was tagged with a FLAG epitope at the NH 2 terminus (FL-Kv1.1) (Fig. 1B). Ile-177 and Ser-180, the two completely conserved amino acids in the S1 segment, were mutated to arginine, leucine, or proline (FL-Kv1.1I177R, FL-Kv1.1I177S, FL-Kv1.1I177P, or FL-Kv1.1S180R etc.) (Fig. 1B). Oocytes injected with FL-Kv1.1 cRNA showed a robust expression of a rapidly activating, non-inactivating outward potassium current (Fig. 2A). The inclusion of a FLAG epitope at the amino-terminal end did not produce any noticeable change in the properties of this current (10,17). In contrast, oocytes injected with each of the mutant cRNA showed no detectable outward potassium currents even after injection of higher doses of cRNAs or recording after longer post-injection periods (Fig. 2, B and C). Thus, the mutation of either Ile-177 or Ser-180 to Arg, Leu, or Pro abolished the Kv1.1-encoded out- 1 The abbreviation used is: PBS, phosphate-buffered saline.

Ile-177 and Ser-180 Mutants Did Not Express Outward Potassium Currents in Oocytes-
FIG. 1. The primary structure of the S1 segments of voltage-gated Kv channels. A, the sequence of amino acids 168 -186 in rat Kv1.1 (above) and the corresponding Shaker (below) K ϩ channels are shown. The amino acid sequence is shown in a single-letter code. Dashes indicate identity. The consensus sequence at the bottom contains all possible primary structures of the S1 segment of mammalian Kv1 through Kv4 and Drosophila channels. Mutations introduced at Ile-177 and Ser-180 are boxed. B, schematic representation of FL-Kv1.1 or the FL-Kv1.1Ile177 and FL-Kv1.1Ser180 constructs (above) and an HA-tagged truncated TR-Kv1.1HA (below) in pCDNA3 vector.
ward potassium currents.
Dominant Negative Suppression of Kv1.1-encoded Currents by Ile-177 and Ser-180 Mutants-To address the question of whether the mutations affected the ability of these subunits to co-assemble with the Kv1.1 polypeptide, we carried out cRNA co-injection experiments. In these experiments, the wild type and mutant subunits were expected to co-assemble into both homo-and heterotetrameric complexes, since they all have an intact T1 domain. The probability of each type of homo-and heterotetrameric complexes that would form is given by a binomial distribution (see "Materials and Methods"). The total current observed would be the sum of the currents carried by each of the conducting complexes. An accurate prediction by this equation rests upon several critical parameters. 1) The amplitudes of the expressed currents must be linear as a function of the amount of cRNA injected and the protein it yielded. 2) Both the wild type and the mutant subunits must be able to co-assemble with an equal probability. 3) Co-assembly with one or more mutant subunits may or may not completely suppress the assembled channel's conductance, but should not alter it.
In order to ensure the validity of the first consideration, we first measured the amplitude of outward potassium currents expressed in oocytes injected with increasing amounts of FL-Kv1.1 cRNA, using the same pulse protocol as above (Fig. 2D). As clearly shown in Fig. 2, E and F, the linearity of the current amplitudes was maintained fairly well at different voltages with the amounts of cRNA tested. Based on these results, we chose to use 0.69 ng of FL-Kv1.1 cRNA (expressing 10 to 30 A current at 0 mV) in all our co-injection experiments. Co-injection of FL-Kv1.1 with mutant cRNAs at a 1:1 ratio resulted in a significant suppression of Kv1.1-encoded currents (Fig. 3, A  and C-H, p Ͻ 0.01). The extent of current suppression was greatly enhanced at a 1:3 ratio of wild type to mutant cRNA (Fig. 3, C-H), except for the FL-Kv1.1I177P mutation, which had no noticeable effect on the amplitude of FL-Kv1.1 encoded currents (Fig. 3G). By contrast, we did not observe any suppression of these currents by co-injection of a 3-fold excess of control cRNA (pTRI-Xef1) (Fig. 3B). Co-injection of the S1 mutant cRNAs did not alter the slope conductance, voltagedependence or kinetics of activation of the Kv1.1-encoded currents (data not shown). Thus, suppression of the Kv1.1-encoded currents was specifically related to the presence of the mutant cRNAs. Using the binomial equation, the percentages of the predicted currents at both 1:1 and 1:3 ratios of co-injections are given if one or more mutant subunits are sufficient to block the formation of functional channels (Table I). The observed values were very close to the predicted values at the 1:3 ratio, assuming that co-assembly with one mutant subunit was sufficient to block the formation of functional channels. At a lower ratio of 1:1, a somewhat weaker suppression than predicted was observed. This phenomenon may have resulted from the lesser amounts of protein expressed by the mutant cRNAs as compared with the amount expressed by the wild type, thereby shifting the actual ratios of the subunits available for co-assembly. Indeed, we have consistently observed variations in the level of expression of mutant proteins in transfected COS-7 cells (see protein results below). The protein analysis also revealed that the level of the expressed FL-Kv1.1I177P protein was dramatically reduced, thus explaining its inability to suppress the currents in oocytes.
The principle of this method is based on a quantitative depletion of the pool of one of the interacting proteins (the fulllength FL-Kv1.1, the FL-Kv1.1Ile177, or FL-Kv1.1Ser180 mu- tants), which is in direct co-assembly with TR-Kv1.1HA. The heteromultimeric complexes were precipitated by the addition of non-limiting amounts of anti-HA antibody directed against TR-Kv1.1HA. Subsequently, anti-FLAG antibody was used to precipitate the remaining pool of homomultimeric complexes and unassembled FL-Kv1.1 subunits (or S1 mutants). The typical results obtained are presented in Fig. 4, A and B. When FL-Kv1.1 was transfected alone, anti-FLAG antibody could immunoprecipitate multiple polypeptides with apparent molecular masses of 56 to 59 kDa (Fig. 4A, lane 3). These bands most likely correspond to the differentially glycosylated and phosphorylated forms of Kv1.1 ␣ subunit (17,22,23). In contrast, two successive rounds of immunoprecipitation with anti-HA antibody failed to precipitate any Kv1.1 polypeptide (Fig. 4A,  lanes 1 and 2). The multiple, closely moving bands of FL-Kv1.1 and FL-Kv1.1Ile or FL-Kv1.1Ser mutants that we have observed here have similar molecular masses and therefore reflect different maturation stages. The transfection of TR-Kv1.1HA alone followed by its precipitation by either anti-HA antibody or anti-FLAG antibody showed that it could only be brought down by anti-HA antibody (Fig. 4B, lanes 4 and 5), not by anti-FLAG antibody (Fig. 4B, lane 6). TR-Kv1.1HA coded for two polypeptides with an apparent molecular mass of ϳ32 kDa. Their mobility corresponded to that of the in vitro translated TR-Kv1.1HA (not shown). When FL-Kv1.1 and Tr-Kv1.1HA were co-transfected, most of the multiple bands that correspond to FL-Kv1.1 polypeptides were immunoprecipitated by anti-HA antibody (Fig. 4A, lanes 7 and 8), whereas only a small fraction immunoprecipitated with the anti-FLAG antibody (Fig. 4A, lane 9). A corresponding depletion of the interacting TR-Kv1.1HA protein in two sequential rounds with anti-HA antibody can be seen in Fig. 4B (lanes 7 and 8), confirming their co-assembly with FL-Kv1.1. A third round of immunoprecipi-  tation with anti-FLAG antibody did not precipitate any TR-Kv1.1HA polypeptides (lane 9), validating the usefulness of the scheme. In contrast, in the mixed lysates of separately transfected COS-7 cells, the FL-Kv1.1 protein could be brought down only by anti-FLAG antibody and not by anti-HA antibody (Fig.  4A, lanes 10 -12). This observation confirms that the co-assembly of FL-Kv1.1 and TR-Kv1.1HA proteins depended upon their co-translation. Mock transfection using the vector alone did not yield any specific protein bands (Fig. 4, A and B, lanes 13-15).
All the FL-Kv1.1Ile177 and FL-Kv1.1Ser180 mutants were then tested for their ability to co-assemble by this sequential method. Fig. 4A shows the positions and the intensities of the polypeptides encoded by FL-Kv1.1Ile177 and FL-Kv1.1Ser180 mutants. Co-transfection of each of the FL-Kv1.1Ile177 or FL-Kv1.1Ser180 mutants with TR-Kv1.1HA resulted in the coprecipitation of multiple protein bands with apparent molecular masses similar to those of the FL-Kv1.1 polypeptides. However, we consistently observed variations in their expression levels (Fig. 4A, lanes 16 -33). Co-transfection of the FL-Kv1.1I177L followed by co-precipitation yielded somewhat elevated levels of this protein. In contrast, co-transfection of FL-Kv1.1I177P resulted in the co-precipitation of dramatically reduced levels of protein (Fig. 4A, lanes 31-33). Other mutants displayed lesser variations, and their co-expression levels were comparable to those achieved with co-transfected wild-type constructs (Fig. 4A, lanes 7-9). Despite these variations, the mutants were efficiently co-immunoprecipitated with the anti-HA antibody (Fig. 4A, lanes 16 -33). A corresponding depletion of the TR-Kv1.1HA with anti-HA antibody is not shown for the mutants, since it was exactly as obtained in the controls (Fig. 4B). Fig. 4C shows the ratios of the FL-Kv1.1 and mutant polypeptides that were immunoprecipitated with either the anti-HA or the anti-FLAG antibody. It is evident that larger fractions of the wild type (ϳ83%) and the mutant proteins (70 -75% for FL-Kv1.1Ile177 mutants and 93-96% for FL-Kv1.1Ser180 mutants) were co-immunoprecipitated with anti-HA antibody (heteromultimeric complexes), whereas much smaller fractions precipitated with the anti-FLAG antibody (homomultimeric complexes and unassembled subunits). The FL-Kv1.1I177P mutant was excluded from this analysis because of the barely detectable level of its protein.
FL-Kv1.1Ile177 and FL-Kv1.1Ser180 Mutatnts Are Targeted to the Plasma Membrane of Sol8 Cells-The failure of all of our mutants to form functional channels in oocytes led us to speculate that they were "folding mutants" which were trapped in the endoplasmic reticulum compartment and were therefore unable to reach the plasma membranes (11,21,24). Hence, we decided to study their subcellular localization using immunofluorescence imaging of transfected cells. This study could also shed some light on the hitherto unknown subcellular distribution of defective K ϩ channels in episodic ataxia (25) and in several types of long QT syndrome (26).
Since the FL-Kv1.1 cDNA was cloned from a rat soleus cDNA library (17), we decided to examine its subcellular expression in Sol8, a myogenic cell line derived from the mouse slow-twitch soleus muscle. Indeed, transiently transfected FL-Kv1.1 could be detected on the membranes of Sol8 cells as well as in the cytoplasm (Fig. 5B). Interestingly, we consistently observed that plasma membrane staining was more commonly found in regions of high confluency where some of the cells appeared to fuse, reminiscent of myotube formation. Untransfected Sol8 cells displayed some background fluorescence, but did not reveal any membrane staining (Fig. 5A). Thus, these results demonstrated that FL-Kv1.1 could be targeted to the plasma membrane of Sol8 cells. When we tested for membrane expression of FL-Kv1.1I177R, FL-Kv1.1I177L, and FL-Kv1.1Ser180 mutants, they could all be detected on the plasma membranes of transfected Sol8 cells. Two representative examples, FL-Kv1.1I177L and FL-Kv1.1S180P, are shown in Fig. 5, C and D. Thus, the mutation of either Ile-177 or Ser-180 to leucine, arginine, or proline did not prevent the targeting of the Kv1.1 polypeptides to the plasma membrane. Transient transfection experiments of FL-Kv1.1 and FL-Kv1.1 mutants into other cell lines (including COS-7, HEK293, Chinese hamster ovary, and Madin-Darby canine kidney) did not yield sufficient plasma membrane expression detectable by anti-FLAG antibodies. In contrast, the transfection of a control membrane protein, the cationic amino acid transporter (27), revealed clear membrane staining in these cells (data not shown). Taken together, these results indicate that Sol8 cells can express detectable levels of FL-Kv1.1 polypetides on the plasma membrane, with or without mutations in the S1 region.

Significance of the Results of Ile-177 and Ser-180
Mutations in the S1 Segment of Kv1.1-A significantly high level of conservation of amino acids in the core region of voltage-gated channels is indicative of functionally important sites and forms an important basis for their three-dimensional structural and functional modeling (28). The mutations of either Ile-177 or Ser-180, the two completely conserved amino acid residues in the S1 segments of the Kv1-4 voltage-gated K ϩ channels, abolished the expression of outward potassium currents in oocytes (Fig. 2, B and C). In transfected COS-7 cells, all mutant cDNA constructs (except FL-Kv1.1I177P) directed the synthesis of proteins, which were matured normally and formed products of the expected molecular mass, albeit with some variation in their steady-state levels (Fig. 4). It is conceivable that the presence of the potentially helix-disrupting proline residue at position 177 resulted in decreased protein stability.
Our analysis indicated that a block in the biosynthesis or maturation of Kv1.1 polypeptides could not explain the generation of non-functional channels. Indeed, the mutations did not alter the consensus sites for the processing and maturation of Kv1.1 protein in transfected cells (22,23,29,30). Moreover, most of the wild type or mutated Kv1.1 polypeptides formed heteromultimeric complexes with TR-Kv1.1HA (Fig. 4C). The smaller fraction, which was immunoprecipitated with anti-FLAG antibody in the last step, most likely contains both homomultimeric channel complexes and unassembled subunits. These results were in agreement with those from the electrophysiological measurements, in which co-injection of the mutants with the wild type cRNA in Xenopus oocytes suppressed most of the Kv1.1 encoded currents (Fig. 3). Collectively, we conclude that none of the Ile-177 or Ser-180 mutations obliterated the co-assembly of homo-or heteromultimeric complexes with either the wild type or the mutated subunits. Furthermore, the incorporation of a single mutant subunit was most likely sufficient to suppress Kv1.1-encoded currents.
The first transmembrane segment plays a critical role in initiating the insertion of newly translated polypeptides into the endoplasmic reticulum membrane and in promoting the stability and clustering of the other transmembrane segments of membrane proteins (31). Indeed, several studies (9, 32), including ours (10,11), have shown that the deletion of this segment abolished the assembly of Kv1 subunits. The distinct staining of each of the mutant FL-Kv1.1 channel protein on the plasma membranes of the Sol8 cells is an intriguing phenomenon (Fig. 5). We speculate that differentiated Sol8 myocytes express high levels of membrane-associated proteins that are important for the trafficking and membrane expression of Kv1.1 polypeptides. Assuming a co-translational co-assembly for all subunits (shown for the wild type in Fig. 4A, lanes  10 -12), the staining probably represents homomultimeric complexes and not a single subunit. These complexes could reach  Kv1.1, Kv1.1Ile177, and Kv1.1Ser180 mutants. A, confocal immunofluorescence imaging of untransfected Sol8 cells using anti-FLAG antibody. B-D, confocal immunofluorescence imaging of Sol8 cells transfected with FL-Kv1.1, FL-Kv1.1I177R, and FL-Kv1.1S180P, respectively. Laser-scanning confocal microscopy (MRC 1024, Bio-Rad) was used to examine the cells (10 -30% laser power). The captured images were saved in an Adobe PhotoShop format and adjusted equally by using the "brightness/contrast" from the menu. the plasma membrane and form non-functional channel complexes. In this context, it is worth noting that the W434F, a mutation in the pore region of the Shaker B K ϩ channel which rendered the channel non-conducting, apparently reached the membranes and "expressed" gating currents (33).
A Proposed Role for Ile-177 and Ser-180 in Critical Subunit Interactions-The S1 segment contains both the hydrophobic and hydrophilic residues (Fig. 1A). Current three-dimensional structural models of voltage-gated K ϩ channels predict that the S1 segment forms a transmembrane amphipathic ␣ helix (34). This model depicts the S1 segment closely packed with the S2 and S4 segments in a cylindrical bundle of ␣ helices in the outer half of the channel's "open" conformation. Direct experimental evidence for functionally critical interactions among the charged residues in the S2, S3, and S4 segments has emerged from recent elegant biochemical studies on the native Shaker K ϩ channel (35) and from studies investigating the interactions of synthetic transmembrane segments in phospholipid membranes (36). Our evidence for the existence of non-functional channels in the membranes suggests that both the Ile-177 and Ser-180 residues in the S1 segment of Kv1.1 might be critically important in helix-helix interactions, rather than in the assembly process per se, as we had previously hypothesized (10). It is likely that inclusion of one "bad" subunit in the tetrameric channel here could hinder critically important intra-or intersubunit interactions, which could in turn impede the sequential steps as the channel proceeds from the closed to the final open state. This is perhaps the best interpretation of all our results, which showed that all of the mutant proteins (except the I177P) were well expressed, could co-assemble to form multimeric complexes, exerted a dominant negative effect on the wild-type K ϩ currents, and could also reach the plasma membranes. In summary, our results highlight the important role that the Ile-177 and Ser-180 residues play in the function of Kv1.1 channels. These observations will help increase our understanding of the cellular mechanisms in long QT-syndrome and ataxia/myokymia, which have been envisaged to arise by dominant negative mechanisms of defective voltagegated K ϩ channels, including Kv1.1 (37).