Cloning and expression of a novel K+ channel regulatory protein, KChAP.

Voltage-gated K+ (Kv) channels are important in the physiology of both excitable and nonexcitable cells. The diversity in Kv currents is reflected in multiple Kv channel genes whose products may assemble as multisubunit heteromeric complexes. Given the fundamental importance and diversity of Kv channels, surprisingly little is known regarding the cellular mechanisms regulating their synthesis, assembly, and metabolism. To begin to dissect these processes, we have used the yeast two-hybrid system to identify cytoplasmic regulatory molecules that interact with Kv channel proteins. Here we report the cloning of a novel gene encoding a Kv channel binding protein (KChAP, for K+ channel-associated protein), which modulates the expression of Kv2 channels in heterologous expression system assays. KChAP interacts with the N termini of Kvalpha2 subunits, as well as the N termini of Kvalpha1 and the C termini of Kvbeta subunits. Kv2.1 and KChAP were coimmunoprecipitated from in vitro translation reactions supporting a direct interaction between the two proteins. The amplitudes of Kv2. 1 and Kv2.2 currents are enhanced dramatically in Xenopus oocytes coexpressing KChAP, but channel kinetics and gating are unaffected. Although KChAP binds to Kv1.5, it has no effect on Kv1.5 currents. We suggest that KChAP may act as a novel type of chaperone protein to facilitate the cell surface expression of Kv2 channels.

The electrical properties of excitable cells are determined in large part by the voltage-gated K ϩ channels (Kv) 1 they possess. Multiple Kv channels control the falling phase of the action potential in excitable cells. Kv channels are also important in many nonexcitable cells, where they may contribute to diverse processes such as volume regulation, hormone secretion, and activation by mitogens. The extensive diversity in Kv currents is matched by the multiplicity of genes encoding the poreforming or ␣-subunit of Kv channels. About 20 mammalian Kv␣ genes have been cloned, and most have been assigned to one of four major subfamilies based on sequence similarities: Kv1, Kv2, Kv3, and Kv4 (1). Each K ϩ channel gene encodes a single subunit, and functional channels are formed by the tetrameric association of individual subunits apparently mediated by specific binding between the N-terminal domains of subunits within individual subfamilies (2,3). With multiple Kv channel genes whose products may assemble as multisubunit heteromeric complexes (4 -6), there may be hundreds of functionally distinct K ϩ channels. Given the great diversity and fundamental importance of K ϩ channels, the cellular mechanisms regulating their synthesis, assembly, and metabolism are of prime interest but remain largely unknown.
The identification and characterization of accessory or modulatory subunits for Kv channels is a new and rapidly expanding area of research. One family of modulatory proteins that interact with Kv␣1 channels, Kv␤ subunits, has been cloned and characterized in the past several years. Kv␤ subunit genes, cloned from heart (7-10) and brain (11)(12)(13), encode cytoplasmic proteins that form stable complexes with Kv␣1 subunits and exert multiple effects on Kv␣1 currents. The three Kv␤1 isoforms and Kv␤3 introduce inactivation into Kv␣1 subunit currents but with variable potency (12)(13)(14). A second effect of Kv␤ subunits is to increase the surface expression of certain Kv␣1 channels. This has been demonstrated both as an increase in the number of dendrotoxin-binding sites (for Kv1.2 transient expression) (15), as well as an increase in the number of functional channels (16). Complexes between Kv␣1 and Kv␤ subunits have been found to form in the endoplasmic reticulum (15,17), suggesting that Kv␤ subunits assist in the folding and assembly of at least some Kv␣1 subunits. The association of Kv1.2 with Kv␤ subunits produces more efficient glycosylation of Kv1.2, increases the stability of Kv1.2 through Kv1.2⅐Kv␤ complex formation and results in an increase in cell surface expression (15).
To gain more information about the synthesis, assembly, and metabolism of K ϩ channels, we have used the yeast two-hybrid system to identify novel cytoplasmic molecules that interact with Kv subunits. Using Kv␤1.2 as bait, we screened a rat brain cDNA library in the GAL4 activation domain vector and isolated a novel gene that encodes a K ϩ channel-binding protein that we have termed KChAP (for K ϩ channel-associated protein). In addition to Kv␤ subunits, KChAP also binds to the N termini of Kv␣1 and Kv␣2 subunits. Coexpression of KChAP with Kv␣2 subunits results in a dramatic enhancement of both total Kv2.1 protein and surface expression of functional Kv2 channels. The unique sequence and properties of KChAP suggest that it may belong to a novel class of proteins with "chaperone-like" properties.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Library Screen-The entire Kv␤1.2 coding sequence (amino acids 1-408) was subcloned in frame into the GAL4 DNA binding domain vector, pGBT9 (CLONTECH) after polymerase chain reaction-mediated addition of a 5Ј EcoRI site and a 3Ј SalI site and used to screen a pGAD10 rat brain cDNA library (CLONTECH). Transformants in the yeast Y190 strain were plated on synthetic medium lacking tryptophan, leucine, and histidine but containing 3-aminotriazole (25 mm). After incubation for 8 days at 30°C, His ϩ colonies were screened for ␤-galactosidase activity by a filter lift assay. Individual pGAD10 recombinant plasmids were screened for interaction with Kv␤1.2 by repeating the yeast two-hybrid assay.
Cloning of Full-length cDNA-The KChAP-Y/pGAD10 plasmid contained a 1.78-kb insert with an open reading frame of 264 amino acids. To obtain the full-length clone, the 32 P-labeled KChAP-Y insert was used to screen a rat brain cDNA library in gt10 (CLONTECH). One of the hybridizing clones contained an insert of 3.2 kb with a single open reading frame of 574 amino acids. KChAP cDNA without 5Ј-or 3Јuntranslated sequences for subcloning in frame into the yeast twohybrid vector pGBT9 was prepared by polymerase chain reaction to include a 5Ј EcoRI site and a 3Ј SalI site.
Analysis of Protein-Protein Interactions by the Yeast Two-hybrid System-Protein-protein interactions were monitored with the yeast Matchmaker two-hybrid system (CLONTECH  . Human Gu-binding protein (GBP) cDNA encoding the region from residues Met 49 to Asp 645 was obtained by reverse transcriptionpolymerase chain reaction from human brain poly(A) ϩ RNA. Met 49 corresponds to the start methionine residue in KChAP. Protein-protein interactions were tested by cotransformation of plasmid pairs into the yeast host strain Y190 as described previously (16). The appearance of a blue color within 8 h was scored as a positive interaction.
Northern Blot Analysis-A rat multiple tissue Northern blot (CLON-TECH) was probed with a 32 P-riboprobe spanning the region encoding the C-terminal 167 amino acids of KChAP. A T7 promoter sequence was engineered directly onto the end of the 501-base pair coding fragment using the Lig'nScribe kit from Ambion and the riboprobe synthesized with the Maxiscript T7 kit (Ambion). The blot was hybridized with probe (10 6 cpm/ml) overnight at 68°C in NorthernMax hybridization buffer (Ambion). Two room temperature washes in 2 ϫ SSC, 0.1% SDS (15 min each) were followed by two washes at 70°C in 0.1 ϫ SSC, 0.1% SDS (20 min each).
In Vitro Translation and Immunoprecipitation-Full-length KChAP cDNA was removed from pGBT9 with EcoRI and SalI and subcloned into a pCR3 vector, which we modified to allow the cloning of EcoRI/SalI fragments in frame behind a c-myc tag. cRNA for c-myc-KChAP was prepared with the T7 mMESSAGE mMACHINE kit (Ambion). cRNAs for c-myc-KChAP and Kv2.1 were translated in vitro either separately or together in rabbit reticulocyte lysates in the presence of [ 35 S]methionine using the Retic Lysate IVT kit (Ambion). A maximum of 500 ng of cRNA was used in each 25-l translation reaction. Canine pancreatic microsomes (Boehringer Mannheim) (1 l/25 l of translation reaction) were included in reactions in which Kv2.1 was translated. For immunoprecipitation (IP), 10-l aliquots of each translation were diluted into 1 ml of IP buffer (1% Triton X-100, 150 mM NaCl, 50 mM Tris, pH 7.5, 1 mM EDTA). To monitor the ability of the two proteins to associate after translation, 10-l aliquots of individual translates of Kv2.1 and c-myc-KChAP were mixed in 1 ml of IP buffer prior to the addition of antibody. IP was performed with two primary antibodies: anti-Kv2.1 polyclonal (1:100; Upstate Biotechnology, Inc) or anti-c-myc monoclonal (1:400; Boehringer Mannheim). After the addition of the primary antibody, the reactions were mixed gently overnight at 4°C. Immune complexes were collected on magnetic beads coupled to either anti-rabbit or anti-mouse secondary antibodies (Dynal, Inc). After four washes in IP buffer, bound protein was eluted by boiling in SDS sample buffer and analyzed on 10% polyacrylamide/SDS gels. The gel was fixed, soaked in Amplify (Amersham Pharmacia Biotech), and radiolabeled protein detected by fluorography.
Immunofluorescence Microscopy-Five days postinjection, two microelectrode voltage clamp recordings were made from Xenopus oocytes injected with either Kv2.1 cRNA alone or coinjected with Kv2.1 plus c-myc-KChAP cRNAs. Several hours after recording, the same oocytes were fixed and sectioned as described previously (18). After incubation for 2 h in 1% bovine serum albumin/phosphate-buffered saline to block nonspecific binding sites, the oocyte sections were incubated at 4°C overnight with primary antibody (anti-Kv2.1 polyclonal, 1:100) in 1% bovine serum albumin/phosphate-buffered saline. The secondary antibody (fluorescein isothiocyanate-conjugated anti-rabbit, 1:100; Cappel Labs) was added for 2 h at room temperature. For oocytes coinjected with c-myc-KChAP, the same protocol was followed with anti-c-myc monoclonal antibody (1:400) and tetramethylrhodamine B isothiocyanate-conjugated anti-mouse secondary antibody (1:125; Sigma). The sections were examined with an Olympus BH-2 microscope.
Oocyte Fractionation and Western Blotting-To isolate membranes, oocytes were homogenized in 0.3 M sucrose, 10 mM NaPO 4 , pH 7.4 (20 l/oocyte) containing a protease inhibitor mixture (Complete, Boehringer Mannheim). After removal of nuclei and debris by centrifugation at 3000 ϫ g for 10 min, the supernatant was spun at 48,000 ϫ g for 1 h to pellet membranes. Membranes from adult rat brain were prepared using the same protocol. In some experiments, oocyte nuclei were removed manually (19). Enucleated oocytes were processed as described above while the nuclei were extracted for 1 h in homogenization buffer plus 1% Triton X-100. Following a 20-min spin at 10,000 ϫ g, the supernatant containing solubilized nuclear proteins was collected. Protein concentrations were determined by the BCA method (Pierce). For Western blotting, proteins were separated on SDS-polyacrylamide gel electrophoresis and blotted to polyvinylidene difluoride membranes. After blocking with 5% nonfat dry milk in phosphate-buffered saline plus 0.1% Tween 20, blots were incubated with primary antibodies, either a monoclonal Kv2.1 antibody (Upstate Biotechnology, Inc; 1:1000) or monoclonal anti-c-myc antibody (1:400), for 1 h at room temperature. The blots were then incubated with secondary antibody (anti-mouse HRP conjugate, Amersham Pharmacia Biotech; 1:3000) and developed with the ECL-Plus detection system (Amersham Pharmacia Biotech).
Expression in Xenopus Oocytes and Electrophysiology-For cRNA synthesis and expression in Xenopus oocytes, full-length KChAP coding sequence was subcloned into the vector, pCR3 (Invitrogen). KChAP cRNA was prepared using the T7 mMESSAGE mMACHINE kit (Ambion) following linearization of the construct with NotI. GBP cDNA was subcloned into a modified pSP64 vector (NruI site for linearization incorporated past the poly(A) ϩ tail) for in vitro transcription with SP6 polymerase. cRNAs for Kv1␣ subunits were prepared as described previously (7,14). Rat Kv2.1 in pBluescript was linearized with NotI, and cRNA was prepared with T7 polymerase. cRNAs were mixed and injected into Xenopus oocytes as described previously (7). HERG cDNA was kindly provided by Dr. M. Keating. Kv2.2 was a gift from Drs. S. Snyder and J. Trimmer. Kir2.2 cRNA was prepared as described previously (20). We also used a cRNA encoding Kv2.1⌬N in which the N-terminal 139 amino acids had been deleted (21).
Measurement of Xenopus oocyte whole cell currents was performed using the standard two-microelectrode voltage clamp technique. Bath solution contained (in mmol/liter): 5 KOH, 100 NaOH, 0.5 CaCl 2 , 2 MgCl 2 , 100 methanesulfonic acid, and 10 HEPES (pH 7.4). Solution containing 50 K ϩ was prepared by replacing an equivalent concentration of Na ϩ . Electrodes were filled with 3 M KCl and had a resistance of 0.3-0.6 megohms. All recordings were made at room temperature. Linear leakage and capacity transient currents were subtracted (P/4 prepulse protocol) unless specified, and data were low pass-filtered at 1 kHz. pClamp software (Axon Instruments) was used for generation of the voltage-pulse protocols and data acquisition. Data are reported as means Ϯ S.E. Comparisons among multiple groups of oocytes were performed by one-way analysis of variance test, Student's t test, and Student-Newman-Keuls post hoc test. Means are considered to be significantly different when p Ͻ 0.05.

Isolation of a Novel Kv␤ and Kv␣ Subunit-binding Protein
with the Yeast Two-hybrid System-Full-length Kv␤1.2 was used as bait to screen a rat brain cDNA library in the GAL4 activation domain vector, pGAD10. We isolated one clone that exhibited a strong positive signal in the ␤-galactosidase assay. pGAD10 plasmid DNA containing a 1.78-kb insert was isolated from this clone and tested positive for interaction with Kv␤1.2. Sequence analysis of the clone, which we termed KChAP-Y (for K channel-associated protein; Y refers to the fragment isolated in the yeast two-hybrid screen), revealed a novel clone with no similarity to Kv␣ or Kv␤ subunits.
We tested the specificity of interaction of KChAP-Y with a panel of Kv␤, Kv␣, and other K ϩ channel subunit fragments with the yeast two-hybrid assay. As shown in Fig. 1, KChAP-Y interacted with both Kv␤1.2 and Kv␤2 subunits. KChAP-Y interacted with the conserved Kv␤1 C terminus but not the unique N terminus of Kv␤1.2, suggesting that the protein may recognize conserved sequences among Kv␤ subunits. Kv␤ subunits interact specifically with the N terminus of Kv1␣ subunits, so we tested these fragments for binding to KChAP-Y as well. Surprisingly, a positive signal was observed between the N termini of Kv1.1, Kv1.2, Kv1.4, Kv1.5, and KChAP-Y. Just as with the Kv␤ subunits, however, no interaction was evident between the Kv1.4 C terminus and KChAP-Y. KChAP-Y also interacted with the N termini of Kv2.1 and Kv2.2 but not with the N terminus of the electrically silent Kv2 partner, Kv6.1 (22). Further specificity for a subset of Kv channels was apparent from the lack of interaction with the N terminus of the inward rectifier K ϩ channel, Kir2.2, and the N terminus of the delayed rectifier K ϩ channel, HERG. Thus, KChAP-Y apparently interacts with both the C terminus of Kv␤ subunits and the N termini of Kv1 and Kv2 ␣-subunits.
Cloning and Sequence Analysis of Full-length KChAP-Screening of a rat brain cDNA library with the KChAP-Y coding sequence produced a 3.2-kb insert, which overlapped KChAP-Y and contained a single open reading frame of 574 amino acids. The initiating methionine was assigned as the first ATG downstream from three in frame stop codons. Hydropathy analysis indicated no potential membrane spanning domains in KChAP, suggesting that the protein was cytoplasmic (not shown).
Search of the GenBank TM nonredundant data base revealed significant homology with the mammalian gene encoding GBP (23). GBP was isolated originally in a yeast two-hybrid screen as a protein that binds to the Gu/RNA helicase II subunit. Alignment of KChAP with GBP is presented in Fig. 2. GBP has an N-terminal extension of 55-57 amino acids compared with KChAP, but over the 574-amino acid open reading frame of KChAP, the two proteins are 50% identical. We tested the binding of both full-length KChAP (amino acids 1-574) and human GBP (amino acids 49 -645) with K ϩ channel fragments in the yeast two-hybrid assay as was described for KChAP-Y in Fig. 1. Full-length KChAP was identical to KChAP-Y in its interaction with protein partners in the yeast two-hybrid assay, while GBP did not interact with any of the tested fragments including Kv␤ and Kv␣ subunits (data not shown). Thus, although KChAP shares significant homology with GBP, interaction with Kv␤ and Kv␣ subunits appears to be a unique feature of KChAP.
Northern Blot Analysis-The expression of KChAP mRNA was examined in a panel of rat tissues. The blot was probed with a fragment of KChAP encoding amino acids Lys 407 -Asp 574 , a region with minimal homology to GBP, to avoid detecting GBP transcripts as well. As shown in Fig. 3, a single band of ϳ3.2 kb was detected in a variety of tissues including heart and brain with especially high levels in lung and kidney.
Functional Characterization of KChAP-Kv Interactions-The surprising finding that KChAP associated with Kv␣1 and Kv␣2 subunits as well as Kv␤ subunits led us to examine the functional consequences of KChAP-K ϩ channel interaction upon heterologous expression in Xenopus oocytes. Whole oocyte currents were recorded by two-electrode voltage clamp from eggs injected with cRNAs encoding different Kv␣-subunits alone or with saturating concentrations of KChAP. Coexpression with KChAP produced a dramatic 3-fold increase in the amplitude of Kv2.1 currents (Fig. 4A). No change in Kv2.1 currents was apparent when the channel was coexpressed with GBP (data not shown). At more depolarized potentials, Kv2.1 has an opening probability of about 0.9 (24), suggesting that the increased currents recorded when KChAP was coexpressed were probably due to an increase in the number of functional channels. KChAP also interacted with the N terminus of Kv1.5 but, in contrast to Kv2.1, produced no significant change in Kv1.5 currents when coexpressed in oocytes (Fig. 4B). The experiments were done so that Kv1.5 expressed whole cell currents at ϩ70 mV in the range of 0.5-5 A. This greatly reduced the possibility that amplitude changes might be missed as a result of voltage clamp difficulties. Thus, while KChAP interacted with the N termini of both Kv2.1 and Kv1.5 in the yeast two-hybrid assay, KChAP only produced amplitude increases in Kv2.1 currents in oocyte expression assays. Fig. 4C summarizes the effects of KChAP on the current amplitudes of a variety of K ϩ channels. In 13 batches of oocytes coinjected with both Kv2.1 and KChAP, we observed an average increase in whole oocyte currents of about 2.5-fold compared with oocytes expressing Kv2.1 alone. KChAP also produced comparable increases in Kv2.2 currents. We also examined the functional expression of a deleted Kv2.1 in which the N-terminal 139 residues were removed (21). As shown in Fig. 4C, coexpression with KChAP did not significantly alter current amplitudes in oocytes expressing Kv2.1⌬N. This suggests that binding between the Kv2.1 N terminus and KChAP is critical for current enhancement.
Two K ϩ channels that did not exhibit N-terminal binding to KChAP, Kir2.2 and HERG, were also tested. For each channel, experiments were done with whole cell control inward currents not exceeding 10 A at Ϫ100 mV. As shown in Fig. 4C, neither channel exhibited altered current amplitudes in the presence of KChAP.
KChAP Increased Functional Expression of Kv2.1 without Altering Channel Kinetics or Gating-The expression enhancement of Kv2.1 currents in the presence of KChAP could be due to an increase in the number of functional channels at the cell surface or an alteration in the kinetics or gating of individual channels. To distinguish between these mechanisms, we used both immunocytochemical and electrophysiological methods. We examined the surface expression of Kv2.1 protein in oocytes   FIG. 1. Yeast two-hybrid assay of the interaction of a novel protein, KChAP, with K ؉ channel fragments. The KChAP fragment isolated from the yeast two-hybrid screen (KChAP-Y in the GAL4 activation domain (AD) vector pGAD10) was tested for interaction with a panel of K ϩ channel fragments cloned in the GAL4 DNA binding domain (BD) vector, pGBT9. Cotransformants of KChAP-Y/pGAD10 with individual binding domain plasmids in the yeast strain Y190 were selected by plating on medium lacking both tryptophan and leucine (ϪTrp/ϪLeu). A representative colony from each cotransformation was spotted onto another ϪTrp/ϪLeu plate for assay of the lacZ reporter gene. After growth for 2.5 days at 30°C, yeast colonies were lifted to paper filters and assayed for ␤-galactosidase activity. Interaction between fusion proteins was revealed by color development in the X-gal column resulting from transcriptional activation of the lacZ reporter gene. Note that KChAP-Y does not produce autonomous transcriptional activation when cotransformed with the pGBT9 vector only. All constructs expressed as binding domain fusions were shown to be negative for autonomous transcriptional activation with the activation domain plasmid, pGAD424, prior to cotransformation with KChAP-Y (not shown). expressing either Kv2.1 alone or Kv2.1 plus c-myc-KChAP by immunocytochemistry. Fig. 5 shows the whole cell currents recorded from a single oocyte injected with Kv2.1 alone (panel A) or Kv2.1 plus c-myc-KChAP (panel B). Currents were increased about 3-fold in the c-myc-KChAP-coinjected egg. The same two oocytes were fixed, sectioned, and co-stained with Kv2.1 and c-myc antibodies. Kv2.1 at the cell surface was visualized by indirect immunofluorescence with a fluorescein isothiocyanate-conjugated secondary antibody. Fluorescence at the oocyte surface was much brighter in the egg expressing both Kv2.1 and KChAP (panel D) compared with the one expressing Kv2.1 alone (panel C), suggesting that the amount of Kv2.1 protein at the cell surface was increased when the channel was coexpressed with KChAP. No staining with the c-myc antibody was seen, suggesting that KChAP is not present at the cell surface with Kv2.1 (panel E).
To determine whether KChAP increased the total expression of Kv2.1 protein or only altered the subcellular distribution of the channel, we examined membrane fractions from oocytes injected with Kv2.1 alone or Kv2.1 plus KChAP by Western blotting. As shown in Fig. 6A, the amount of Kv2.1 protein in oocyte membranes was increased in oocytes coinjected with KChAP ( compare lanes 1 and 2). Densitometry of the blots indicated about a 2.5-fold increase in Kv2.1 protein in the presence of KChAP. Similar results were obtained when Kv2.1 was immunoprecipitated from homogenates of total oocyte protein (data not shown). This value is comparable with the increase observed in Kv2.1 currents with KChAP.
Since no KChAP was detected at the cell surface of coexpressing oocytes, we examined the cellular distribution of KChAP in oocytes by Western blotting. We manually removed the nuclei from pools of oocytes expressing only Kv2.1 or Kv2.1 plus c-myc-KChAP and probed the nuclear fraction as well as the soluble and membrane fractions prepared from the enucleated oocytes with an anti-c-myc antibody to detect tagged KC-hAP. Most of the KChAP protein was present in the nuclear fraction with smaller amounts detectable in the soluble as well as the membrane fractions (Fig. 6B). The signal was so strong in the nuclear material compared with the other two fractions,  however, that we were not able to estimate accurately the relative amounts in each fraction. No Kv2.1 was detected in the nuclear fraction, indicating that contamination with nonnuclear membranes was minimal (data not shown).
The kinetics and gating of Kv2.1 channels were not altered in the presence of KChAP. As shown in Fig. 7A, the voltagedependence of activation and the kinetics of activation and deactivation of Kv2.1 channels were not changed. Coexpression with KChAP did not alter the sensitivity of Kv2.1 channels to  Ϫ80 mV, and pulses were from Ϫ70 mV to ϩ80 mV in 10-mV steps with 50 mM K ϩ in bath solution. B, averaged macroscopic currents from 10 oocytes in one injection series measured 5 days postinjection from oocytes injected with Kv1.5 cRNA (50 ng/l) alone (left) or Kv1.5 plus KChAP cRNAs (50 and 500 ng/l, respectively) (right). Holding potential was Ϫ80 mV, and pulses were from Ϫ70 to ϩ70 mV in 10-mV steps with 5 mM K ϩ in bath solution. C, bar plot showing averaged current levels in the presence of KChAP as fractions of currents in the absence of KChAP (control current). The numbers above each bar indicate the number of batches of oocytes examined for each K ϩ channel. Oocytes were injected with either K ϩ channel cRNAs or K ϩ channel plus KChAP cRNAs, and currents were recorded from 8 -12 oocytes in each batch. Whole oocyte currents were measured 2 days after injection (Kir2.2 and Kv2.1⌬N) or 5-6 days after injection (Kv2.1, Kv2.2, Kv1.5, and HERG), and the ratio of means (I coinjected /I control ) was calculated. For Kv2.1, Kv2.2, and Kv1.5, the holding potential was Ϫ80 mV. Steady-state currents were measured at a test potential of ϩ70 mV (5 or 50 K ϩ in bath). Kir2.2 steady-state and HERG tail currents were recorded with 50 K ϩ in the bath at test potentials to Ϫ100 mV with a prepulse to ϩ20 mV. Asterisks indicate that in all injection series, current amplitudes in oocytes coinjected with KChAP were significantly higher than in oocytes without KChAP (t test, p Ͻ 0.05).  Fig. 7C. In all of the electrophysiological experiments presented thus far, KChAP and Kv2.1 cRNAs were mixed prior to injection into oocytes in the same pipette. Importantly, when we injected the two cRNAs separately into the same oocyte with different pipettes, no enhancement of Kv2.1 currents was observed (data not shown). This result suggests that KChAP exerts its effect on Kv2.1 in oocytes by a direct physical association with the channel, which is facilitated by the coinjection of both cRNAs into the same space inside the oocyte.
In Vitro Association of KChAP and KV2.1-The electrophysiological data suggest that a direct interaction between Kv2.1 and KChAP occurs and is responsible for the enhancement in Kv2.1 currents observed in oocytes. We used an in vitro binding assay to demonstrate the ability of the two proteins to associate. Kv2.1 and KChAP cRNAs were translated in vitro either separately or together in a rabbit reticulocyte lysate in the presence of [ 35 S]methionine. We used a commercially available anti-Kv2.1 polyclonal antiserum to immunoprecipitate Kv2.1 and analyzed the immunoprecipitated material with SDS-polyacrylamide gel electrophoresis and fluorography to detect the presence of associated KChAP. Since an antibody to KChAP was not available, we used an epitope tag fused to the N terminus of KChAP (c-myc) to allow detection. As we had previously shown in Fig. 5, the c-myc tag did not interfere with the functional interaction of KChAP and Kv2.1. Control reactions with Kv2.1 translated alone showed that in vitro translated Kv2.1 was immunoprecipitated with anti-Kv2.1 antibody but not anti-c-myc antisera (Fig. 8, lanes 1 and 2). Similarly, anti-c-myc antisera immunoprecipitated c-myc-KChAP but not Kv2.1 (Fig. 8, lanes 3 and 4). Kv2.1 antibody coimmunoprecipitated complexes of Kv2.1 and KChAP when the two cRNAs were cotranslated (Fig. 8, lane 5) but not when the two cRNAs were translated in separate reactions and mixed together prior to the addition of primary antibody (Fig. 8, lane 6). This result suggests that the association of KChAP with Kv2.1 occurs cotranslationally, since the mature proteins added after translation did not coimmunoprecipitate. All reactions involving translation of Kv2.1 cRNA shown here included canine pancreatic microsomes to allow the channel to insert into membrane as it was synthesized. When microsomes were omitted from cotranslation reactions, no coimmunoprecipitation of the two proteins was observed (data not shown). DISCUSSION We have cloned a novel gene, kchap, which encodes a K ϩ channel regulatory protein. All of our data point to a direct interaction between KChAP and Kv2.1 as responsible for the current enhancement observed in Xenopus oocytes. Yeast twohybrid assays showed an interaction between KChAP and the N terminus of Kv2.1. When an N-terminally truncated Kv2.1 was expressed in oocytes, no expression enhancement was observed. Also, Kv2.1 and KChAP were co-immunoprecipitated from in vitro translations in which the two proteins were translated in the same reaction, but not when they were translated separately and mixed together before immunoprecipitation. This suggests a cotranslational association of the two proteins, which is also supported by oocyte data. In order to observe Kv2.1 current enhancement in oocytes, the KChAP and Kv2.1 cRNAs had to be mixed together and coinjected with the same pipette. No changes in Kv2.1 currents were noted when the two cRNAs were injected from two separate pipettes, suggesting that the two proteins must be translated in close proximity. However, in the oocyte, KChAP might bind only transiently to Kv2 channels and not remain stably attached to the mature channel as evidenced by the lack of KChAP staining at the oocyte surface.
In Xenopus oocytes, KChAP does not appear to alter the time course of Kv2.1 current expression. Without KChAP, the amplitude of whole oocyte Kv2.1 currents continues to increase slowly from the time of injection through about 6 days, at which time the current levels plateau and remain fairly constant through 9 days postinjection. This would suggest that transit of Kv2.1 channels to the oocyte surface is a rather slow process. KChAP does not change this profile but increases the amplitude of Kv2.1 currents recorded at each time point.
One possible explanation for the effect of KChAP on Kv2.1 expression is that KChAP may act as a chaperone to facilitate either the translation, assembly, or stability of Kv2.1 channels. An increase in the amount of Kv2.1 protein in oocyte membranes in the presence of KChAP as seen by Western blotting is consistent with this interpretation. Kv␤ subunits also exhibit chaperone-like effects on Kv channels. However, Kv␤ subunits (Kv␤1 and Kv␤2) form stable complexes with Kv␣ subunits, and although they are cytoplasmic proteins, travel to the cell surface with pore-forming Kv␣ subunits and remain tightly attached there as part of mature channel complexes (28 -30). KChAP defines a different structural gene from Kv␤ subunits, however, and there are significant differences between Kv␤ and KChAP action as well. Unlike Kv␤ subunits 1-3, KChAP modulates both Kv2.2 and Kv2.1 currents but has no effect on Kv1.5 currents.
The significance of KChAP binding to Kv␤ or Kv␣1 subunits is currently unknown. KChAP binds to the conserved C terminus of Kv␤ subunits. This is the same domain to which the N termini of Kv␣1 subunits have been shown to bind (14,25,26). One could envision that KChAP might modulate the association of Kv␣1 and Kv␤ subunits and, thus, indirectly affect the expression and/or kinetic properties of Kv1 currents, but this is speculative at present.
Although KChAP and Kv␤ subunits share no sequence homology, KChAP is 50% identical to the human GBP (23). GBP is a novel protein that was cloned by interaction with the Gu/RH-II helicase in the yeast two-hybrid system. GBP did not interact with either Kv␣ or Kv␤ subunits in yeast two-hybrid assays and did not modulate Kv2.1 currents in coexpression assays in oocytes. GBP is a nuclear protein that has been shown to produce proteolytic cleavage of Gu/RH-II (23). The significance of the sequence similarity between the two proteins is unclear at present. Interestingly, however, KChAP protein was primarily detected in the nuclei of injected oocytes. The relationship of the nuclear localization of KChAP with its effects on Kv2 channels is presently unclear. If a direct interaction between KChAP and Kv2.1 is required for expression enhancement, then this may occur transiently prior to KChAP moving to the nucleus. We cannot rule out, however, that KChAP may play an as yet unknown role in the nucleus, which results in enhanced Kv2.1 protein and current levels.
To summarize, we have described a novel gene product, KChAP, which binds to Kv␣ and Kv␤ subunits. KChAP increases currents expressed by Kv2.1 and Kv2.2 but not Kv1.5, and the expression enhancement is due to an increase in total protein as well as in the number of functional Kv2 channels at the cell surface. We speculate that KChAP may act as a novel type of chaperone protein for Kv2 channels.