Kvβ Subunits Increase Expression of Kv4.3 Channels by Interacting with Their C Termini*

Auxiliary Kvβ subunits form complexes with Kv1 family voltage-gated K+ channels by binding to a part of the N terminus of channel polypeptide. This association influences expression and gating of these channels. Here we show that Kv4.3 proteins are associated with Kvβ2 subunits in the brain. Expression of Kvβ1 or Kvβ2 subunits does not affect Kv4.3 channel gating but increases current density and protein expression. The increase in Kv4.3 protein is larger at longer times after transfection, suggesting that Kvβ-associated channel proteins are more stable than those without the auxiliary subunits. This association between Kv4.3 and Kvβ subunits requires the C terminus but not the N terminus of the channel polypeptide. Thus, Kvβ subunits utilize diverse molecular interactions to stimulate the expression of Kv channels from different families.

KAT1 channels (11). Hence, auxiliary Kv␤ subunits are structurally well characterized, yet the specificity and mechanism of interaction between Kv␣ and ␤ subunits remain obscure.
Kv␤ subunits influence expression and function of K ϩ channels. Specifically, distinct Kv␤ subunits differentially affect heterologously expressed Kv1 family channels. A longstretched N-terminal peptide in Kv␤1 and Kv␤3 gene products produces rapid inactivation on most of Kv1 family channels by a mechanism similar to the action of a ball peptide present at the N terminus of some channel ␣ subunits (12). Furthermore, Kv␤2 subunits have been shown to increase stability and cell surface expression of Kv1 family channels (13) without producing rapid inactivation. Thus, expression of distinct Kv␤ subunits controls excitability by differentially affecting the expression and gating of Kv1 family channels.
Although many studies have shown that Kv␤1 and Kv␤2 subunits can associate with heterologously expressed channels from diverse families, it remains unclear whether these auxiliary subunits are present as complexes with non-Kv1 family channels in native cells. Furthermore, structural features of the interaction between Kv␤ subunits and non-Kv1 family channel polypeptides remain unknown. To address these questions, we examined complexes consisting of Kv␤ and Kv4.3 channel subunits. We show here that Kv4.3 proteins are associated with Kv␤2 subunits in the brain and that this interaction requires the C terminus of the channel polypeptide.
Cell Culture and Transfection-HEK 293 cells (American Type Culture Collection, Manassas, VA) were maintained at 37°C under 5% CO 2 atmosphere in high glucose Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Transient transfection was carried out by the calcium phosphate-DNA coprecipitation method (Transfinity, Life Technologies, Inc.).
For patch clamp recording, pCMV-Kv4.3 alone (0.3 g for 60-mm dish) or in combination with 5ϫ-excess pCMV-Kv␤1.1 or pCMV-Kv␤2.1 were used. In addition, EGFP-C1 plasmid (50 ng/60-mm dish) was cotransfected to aid in the identification of transfected cells. Transfected cells in 60-mm plates were split into 35-mm dishes 5 h after transfection and used for whole-cell recordings 48 -72 h after transfection.
For immunoblot analysis, cells on 100-mm plates were transfected with expression constructs at the same ratio as for patch clamp recording (0.9 g of pCMV-Kv4.3/dish). Transfected cells were divided into five 60-mm plates at various densities 5 h after transfection and used for immunoblot analysis at various days after transfection. Cell extracts were prepared by suspending the collected cell pellet in 100 l of lysis buffer (20 mM Tris-HCl (pH 7.4), 0.15 M NaCl, 1% Triton X-100, 1 mM iodoacetamide, 0.2 mM phenylmethylsulfoxide, and 1 mM EDTA). The suspension was kept on ice for 10 min and centrifuged at 10,000 ϫ g for 5 min to remove nuclear debris.
Electrophysiological Recording-Whole-cell voltage-clamp recording (17) was performed with an EPC-9 patch-clamp amplifier using the Pulse program (HEKA Electronik, Lambrecht, Germany) on a Power Macintosh computer. Patch pipettes were filled with a solution containing 140 mM KCl, 1 mM MgCl 2 , 1 mM EGTA, and 10 mM HEPES (pH 7.4). Bath solution contained 155 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 20 mM glucose, 10 mM HEPES (pH 7.4). Series resistance compensation was set at 70%. Peak currents were converted into conductance (G) by the formula G ϭ I/(V m Ϫ V rev ) assuming a reversal potential V rev of Ϫ84 mV, where V m is the membrane voltage of depolarization pulses. Using the first-order Boltzmann equation G/G max ϭ 1/(1 ϩ exp[(V 1 ⁄2 Ϫ V)/slope factor]), the half-maximal voltages (V 1 ⁄2) and the slope factors were acquired. Statistical analysis was carried out using the Mann-Whitney two-tailed test. All the data in the text are presented as means Ϯ S.E.
Biochemical Association Assays-Two days after transfection, transfected HEK 293 cells on 100-mm dishes were harvested with ice-cold phosphate-buffered saline. Triton extract was prepared by suspending the pelleted cells in 0.4 ml of solution containing 1% Triton X-100, 20 mM Tris-HCl (pH 7.9), 50 mM NaCl, and 5 mM imidazole. The extract was mixed with 100 l (50% slurry) of preactivated His-bind resin (Novagen, Milwaukee WI) for 2 h with gentle shaking. The resin was washed 5ϫ with the same solution, except that the imidazole concentration was 40 mM. The bound materials were then eluted with 0.1 M EDTA.
Immunoprecipitation-Immunoprecipitation was performed with polyclonal anti-panKv␤ antibody (18) or monoclonal anti-Kv4.3 antibody. The latter antibody was generated against a synthetic peptide corresponding to a part of the N terminus of rat Kv4.3 polypeptide (amino acids 25-40) CPMPLAPADKNKRQDE. 2 Whole rat brain tissue was homogenized in 0.32 M sucrose solution supplemented with 1 mM iodoacetamide, 0.2 mM phenylmethyl sulfonate, and 1 mM EDTA. The homogenate was centrifuged at 1,000 ϫ g for 10 min to remove nuclear debris. The supernatant was transferred to a new tube and centrifuged at 100,000 ϫ g for 1 h. Protein concentration was determined using Bio-Rad protein assay reagent with bovine serum albumin as a standard. The nuclei-free membrane fraction was suspended in a solution containing 2% Triton X-100, 20 mM Tris-HCl (pH 7.5), 0.15 M NaCl, 1 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride and 1 mM iodoacetamide at a protein concentration of ϳ10 mg/ml. Triton extract was then obtained by centrifugation of the suspension at 100,000 ϫ g for 30 min. After preclearing with fixed protein A-containing Staphylococcus aureus cells (Pansorbin, Calbiochem), Triton extract was incubated overnight with monoclonal anti-Kv4.3 or polyclonal antipan Kv␤ (20) antibody and Pansorbin. The bound materials were collected by centrifugation and washed 4ϫ with the same Triton-containing solution. The bound materials were eluted by heating in 2ϫ SDS sample buffer and subjected to immunoblot analysis.
Immunoblot Analysis-Proteins were separated on a 7.5% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. The membrane was coated with 5% nonfat dry milk in phosphate-buffered saline containing 0.1% Tween 20 and probed with primary antibody followed by incubation with horseradish peroxidase-conjugated secondary antibody. Primary antibodies against Kv4.3, GFP, and Express tag were purchased from Alomone labs (Jerusalem, Israel), MBL International Corp. (Watertown MA), and Invitrogen, respectively. Anti-Kv1.4 (18), anti-Kv2.1 (19), and polyclonal anti-panKv␤ (20) antibodies were previously generated. Bound antibody was detected by chemiluminescence method (PerkinElmer Life Sciences). Immunoreactivity was quantified using densitometry of the developed films.
Confocal Microscopy-Confocal Images of GFP fluorescence were taken on a Molecular Dynamic 2001 scanning laser confocal microscope with a 60ϫ oil immersion objective lens (1.4 NA) using 488-nm excitation and 510-nm emission filters with 3% maximal laser intensity. Cell surface localization was evaluated by comparing the location of fluorescence with bright field images of cells.

Kv4.3 Proteins Are Present in Association with Kv␤2
Subunits in the Brain-To test for the presence of Kv4.3⅐Kv␤ subunit complexes, we first used anti-Kv4.3 monoclonal antibody for immunoprecipitation from rat brain extract. Anti-Kv4.3 antibody effectively and specifically precipitated its targeted channel proteins but not Kv1.4 proteins (Fig. 1A). Importantly, the immunoprecipitated material was found to contain significant immunoreactive Kv␤ subunit proteins detected with polyclonal anti-panKv␤ antibody (20). This antibody detects two bands on the blot with distinct sizes. The larger and smaller bands are known to correspond to Kv␤1 and Kv␤2, respectively (13,19,20). In addition, the larger band may also contain Kv␤3 subunits. We found that only smaller molecular weight Kv␤2 subunits were significant in the precipitated material. To further obtain evidence for the presence of Kv4.3⅐Kv␤ complexes in the brain, anti-panKv␤ antibody was used for immunoprecipitation from the brain extract (Fig. 1B). The antibody precipitated significant Kv4.3 proteins in addition to its targeted proteins. In contrast, no detectable Kv2.1 proteins were found in the precipitated material. Hence, brain Kv4.3 channel proteins are present in association with Kv␤2 subunits.  (13,19,20). Note that a fraction of immunoreactive Kv␤ subunits (Kv␤2) is recovered in the anti-Kv4.3 antibody-precipitated material. Anti-panKv␤ antibody also precipitated a portion of Kv4.3 proteins.
To test if this elevation in current density was correlated with an increase in Kv4.3 protein level, we measured channel proteins by immunoblot analysis (Fig. 3A). The Kv4.3 protein level significantly increased when coexpressed with Kv␤2.1 subunits 3 or 4 days after transfection (Fig. 3B, n ϭ 6, p Ͻ 0.05). In contrast, coexpression of these auxiliary subunits did not produce changes in Kv2.1 protein levels. Similar increases in Kv4.3 proteins were also produced by Kv␤1.1 at 3 days after transfection (Fig. 3C). Hence, Kv␤ subunits increase total cellular Kv4.3 protein level.
Effects of Kv␤ Subunits on Kv4.3 Channel Gating-Next we examined the gating properties of Kv4.3 channels with or without coexpression of Kv␤ subunits. Excess Kv␤ subunits (␣:␤ ϭ 1:5) were used to enhance formation of Kv4.3⅐Kv␤ complexes. Kv␤ subunits have been shown to shift voltage dependence of activation of Kv1 family channels to the left. However, we found that Kv␤1.1 or Kv␤2.1 subunits produce no significant change in voltage dependence of activation of Kv4.3 channels (Fig. 4A). The voltage for half-maximal activation and the slope factor were Ϫ7.00 Ϯ 1. The most profound effect produced by Kv␤ subunits on Kv1 family channels is acceleration of inactivation, due to the ball and chain mechanism. Thus, the effect of Kv␤1.1, which contains a ball peptide, as well as Kv␤2.1 on inactivation properties of Kv4.3 channels was examined. Coexpression of these auxiliary subunits did not significantly affect time constant of inactivation (Fig. 4B) We also measured the steady-state inactivation using a test pulse to ϩ40 mV after a conditioning prepulse at various voltages (Fig. 4, C and D). The voltage for half-maximal inactivation and the slope factor were Ϫ48. 4  We also determined whether association with Kv␤1.1 or Kv␤2.1 might influence the recovery from inactivation (Fig. 4,  E and F). A protocol of two consecutive depolarizing test pulses interrupted by variable interpulse intervals at Ϫ70 mV was used to determine the time course of recovery from inactivation (Fig. 4E). The recovery from inactivation was fitted by a single exponential function. Time constant for recovery from inactivation was 181 Ϯ 24 ms for Kv4.3 alone (n ϭ 6), 169 Ϯ 11 ms for Kv4.3 ϩ Kv␤1.1 (n ϭ 5), and 176 Ϯ 38 ms for Kv4.3 ϩ Kv␤2.1 (Fig. 4F, n ϭ 5). Thus, Kv␤ subunits do not affect recovery from inactivation. Taken together, Kv␤1.1 and Kv␤2.1 subunits produce no marked effects on Kv4.3 channel gating.
The C Terminus of Kv4.3 Proteins Is Required for Association with Kv␤ Subunits-To assess association of Kv channel ␣ subunits with Kv␤ subunits, we first examined localization of GFP-tagged Kv␤ subunits upon coexpression of various channel proteins. Confocal microscopy revealed that GFP-Kv␤2.1 (Fig. 5A) or GFP-Kv␤1.1 (data not shown) were predominantly present in the cytosol in the absence of channel ␣ subunits. Coexpression of Kv4.2 or Kv4.3 proteins as well as Kv1.4 proteins, but not Kv2.1 proteins, localized the fluorescence to plasma membrane (Fig. 5A). Similarly, a splicing variant of Kv4.3, which contains a 19-amino acid insertion at the C terminus (15), targeted GFP-Kv␤ fusion proteins to plasma membrane. Thus, Kv4 family channel proteins regardless of the presence or absence of the insertion can associate with Kv␤ subunits.
Kv1 family channels interact with Kv␤ subunits via a highly conserved region of the N terminus. Although the corresponding region of Kv4 family polypeptides exhibits significant sequence homology, this peptide itself was insufficient for association with Kv␤ subunits (5,6). To identify the region important for association, we generated chimeric channel proteins consisting of Kv4.3 and Kv2.1 polypeptides. If chimeric proteins are capable of interacting with Kv␤ subunits, the fluorescence would be expected in plasma membrane or other membrane-associated compartments. Replacing the N terminus of Kv4.3 protein with that of Kv2.1 polypeptide (Kv4.3-Kv2.1N) did not affect the ability to localize Kv␤2.1 (Fig. 5B) and Kv␤1.1 (data not shown) subunits to plasma membrane and other membrane-associated regions. In contrast, substituting the C terminus of Kv4.3 protein with that of Kv2.1 protein (Kv4.3-Kv2.1C) eliminated plasma membrane localization of the fluorescence. This chimeric channel (Kv4.3-Kv2.1C) was functional as confirmed by patch clamp recording (data not shown). Thus, the C terminus, but not the N terminus, of Kv4.3 polypeptide is required for localizing Kv␤ subunits at plasma membrane.
We also used protein biochemical assays to test association. Histidine (His 6 )-tagged wild type and chimeric channel proteins were expressed with Kv␤1.1 or Kv␤2.1 subunits. After purification with His-binding beads, copurified Kv␤ subunit proteins were examined by immunoblot analysis (Fig. 6). Significantly higher levels of immunoreactive Kv␤1.1 or Kv␤2.1 proteins were recovered from cells coexpressed with Kv4.3 or Kv4.3-Kv2.1N than those with Kv2.1 or Kv4.3-Kv2.1C. These results demonstrate that the C terminus, but not the N terminus, of Kv4.3 channels is necessary for association with Kv␤ subunits. DISCUSSION Kv␤ subunits have been suggested to interact with various K ϩ channels including Kv2.2 (10), Kv4.2 (7,8), and several EAG family (9) and plant KAT1 (11) channels in addition to Kv1 family channels. In particular, Kv4.2 channels were found to interact with Kv␤1 and -2 subunits in heterologous expression systems (7,8), suggesting that the same Kv␤ subunits might form complexes with K ϩ channels from the two different families. However, previous studies had not addressed whether such association actually occurs with native channels and did not identify a physiological effect of the association. In this study, we have shown that Kv4.3 proteins are associated with Kv␤2 subunits in the brain. We also found that coexpression of Kv␤ subunits leads to increases in Kv4.3 current density and protein level without altering gating properties. Finally, this association requires the C terminus, but not the N terminus, of the channel polypeptide. Thus, the same Kv␤2 subunits influence expression and function of channels from the two different families by distinct interaction mechanisms.
Association of Kv␤ subunits appeared to produce different effects on interacting channels. Our results indicate that the ball peptide of Kv␤1.1 does not alter inactivation of Kv4.3 channels. Similarly, it has been shown that Kv␤1 does not markedly influence inactivation kinetics of Kv4.1 (21), Kv4.2, (8) and Drosophila Shal (5) channels. Likewise, Drosophila Kv␤ subunit HK was unable to produce rapid inactivation on  Fig. 2. Currents at 50-mV depolarization pulses were normalized, and the averages of these traces (n Ն 4) are shown in B. Steady-state inactivation (D) and recovery from inactivation (F) were determined using pulse paradigms shown in C and E, respectively. n Ն 4 for each point. Points and error bars represent mean and S.E., respectively. No significant differences in any of the three gating properties between any two conditions were detected. EAG or ERG channels (9). Thus, the effect of Kv␤ subunit ball peptide depends on associating channels. The lack of the ball peptide effect may be due to differences in the ball peptideaccepting structure of channels. Recent work indicated that a difference in the amino acid sequence in a small region of the linker between S4 and S5 affects the interaction with the Kv␤1 ball peptide (22). To support this possibility, one of the important amino acids identified in the above-mentioned study, Arg-324 in Kv1.1, is altered to serine at the corresponding position (Ser-421) in Kv4.3. Another possibility is that Kv␤ subunits are positioned in such a way that the ball peptide cannot access the internal pore region of Kv4.3 channels. Since we found that structural requirement for association with Kv␤ subunits differs between Kv1 family and Kv4.3 channels, it is possible that the relation of a Kv␤ tetramer to a Kv4.3␣ tetramer does not allow the ball peptide to act on its receptor region of the channel. In addition to the difference in rapid inactivation, a recent study revealed that sensitivity to O 2 tension differs between Kv␤ subunit-associated Kv4.2 and Shaker channels (8). In analogy to the difference in rapid inactivation, this difference in hypoxia response may arise from distinct ability of channels to respond O 2 tension signals. Alternatively, specific interaction of Kv4.2 channels with Kv␤ subunits may be essen-tial for the regulation. Further structural and functional information of channel complexes may resolve these issues.
In contrast to specific alterations in channel gating, association of Kv␤ subunits commonly increases current amplitude or density of various channels. This has been observed with Kv1 family (13), Kv2.2 (10), and EAG family (9) channels. Our results also revealed that Kv␤ subunits increase Kv4.3 current density and proteins. In addition to Kv␤ subunits, other channel auxiliary subunits for K ϩ channels as well as Na ϩ and Ca 2ϩ channels have been shown to increase associating channel current density. It is assumed that the exit from endoplasmic reticulum is the rate-limiting step for plasma membrane protein targeting. Therefore, the generally observed increase in current density by various auxiliary subunits may be due to masking of endoplasmic reticulum retention signals present in channel proteins. This mechanism has been implicated for controlling selective cell surface expression of heteromeric ATP-sensitive K ϩ channel complexes (23,24) and voltage-gated Ca 2ϩ channel (25). Thus, it is possible that some of the Kv␤ subunit effect on Kv4.3 channel current density and proteins may be due to masking potential endoplasmic reticulum retention signals in the channel polypeptide. However, our previous study found that GFP-tagged Kv1.4 and Kv1.5 channels are efficiently transported to plasma membrane in the absence of Kv␤ subunits (26,27). Similarly, we found efficient plasma membrane localization of GFP-tagged Kv4.3 (data not shown). Furthermore, coexpression of Kv␤ subunits produced no apparent changes in localization of these GFP-tagged channel proteins. Thus, it is likely that Kv␤ subunits increase Kv4.3 as well as Kv1 family proteins in endoplasmic reticulum and at the plasma membrane. This stabilization effect is further supported by our finding that the Kv␤ effect on Kv4.3 protein level is larger at longer times after transfection. Hence, Kv4.3⅐Kv␤ complexes are likely more stable than those without these auxiliary subunits.
Despite the similarity between Kv4 and Kv1 family polypeptides, our data indicate that the two family proteins exhibit distinct requirements for interaction with Kv␤ subunits. A part of the N terminus of Kv1 family polypeptide is sufficient for association (5,6). In contrast, our results demonstrated that the corresponding region of Kv4 family peptide is not neces-  sary. Instead, the association requires the C terminus of Kv4.3 polypeptide. The importance of the C terminus for interaction with Kv␤ subunits was also suggested in Kv2.2⅐Kv␤4 complex formation; a part of the C terminus of Kv2.2 protein is required for the increase in current density produced by Kv␤4 coexpression in Xenopus oocytes (10). Thus, the C terminus of Kv2.2 and Kv4 family polypeptides is likely to be involved in association with Kv␤ subunits. The apparent lack of sequence similarity between the N terminus of Kv1 family and the C termini of Kv2.2 or Kv4.3 polypeptides suggest that the interaction between Kv␣ and -␤ subunits may be more complex than previously assumed. To further elucidate interaction mechanisms, we generated a chimeric Kv2.1 channel containing the C terminus of Kv4.3 polypeptide. We found that this chimera does not efficiently associate with Kv␤ subunits (data not shown), suggesting that the Kv4.3 C terminus may not be sufficient for association. However, this chimera was found to be nonfunctional. Therefore, misfolding of this chimeric channel protein might be responsible for the observed lack of interaction. Thus, a simple explanation for the requirement of the C terminus is that this peptide interacts with a site of Kv␤ polypeptide that is distinct from one for the Kv1 family N terminus. Alternatively, the C terminus may indirectly participate in interaction. For example, the C-terminal peptide interacts with other part of the channel protein to place an association site in a position for efficient interaction with these auxiliary subunits. More detailed analyses are required to differentiate these possibilities.
Recently identified Ca 2ϩ -binding subunits (KChIP) are likely to play important roles in controlling the expression and function of Kv4 family channels (28). In addition, our results indicate that Kv4.3 channels are present, at least in part, in association with Kv␤2 subunits in brain. The association of non-Kv1 family channels with Kv␤ subunits may have pronounced effects under physiological conditions. Although mutations in Drosophila Kv␤ subunit HK and Kv1 channel Shaker resulted in almost identical electrophysiological changes in some giant neurons, alterations in other types of neurons caused by these HK mutations were distinct from those by Shaker mutations (29). Thus, the interaction of Kv␤ subunits with non-Kv1 channels is likely to play important roles in controlling neuronal excitability. Our results indicate that the major effect produced by Kv␤ subunits on Kv4 family channels is to increase the number of functional channels. The association of Kv␤ subunits may also have other regulatory functions on these channels, such as sensing redox state of the cell. A novel interaction mechanism between Kv4 family channels and Kv␤ subunits may be important for these regulatory functions. Hence, Kv␤ subunits control neuronal excitability by influencing expression and function of Kv1 and Kv4 family channels.