Conserved Kv4 N-terminal Domain Critical for Effects of Kv Channel-interacting Protein 2.2 on Channel Expression and Gating*

Association of Kv channel-interacting proteins (KChIPs) with Kv4 channels leads to modulation of these A-type potassium channels (An, W. F., Bowlby, M. R., Betty, M., Cao, J., Ling, H. P., Mendoza, G., Hinson, J. W., Mattsson, K. I., Strassle, B. W., Trimmer, J. S., and Rhodes, K. J. (2000) Nature 403, 553–556). We cloned a KChIP2 splice variant (KChIP2.2) from human ventricle. In comparison with KChIP2.1, coexpression of KChIP2.2 with human Kv4 channels in mammalian cells slowed the onset of Kv4 current inactivation (2–3-fold), accelerated the recovery from inactivation (5–7-fold), and shifted Kv4 steady-state inactivation curves by 8–29 mV to more positive potentials. The features of Kv4.2/KChIP2.2 currents closely resemble those of cardiac rapidly inactivating transient outward currents. KChIP2.2 stimulated the Kv4 current density in Chinese hamster ovary cells by ∼55-fold. This correlated with a redistribution of immunoreactivity from perinuclear areas to the plasma membrane. Increased Kv4 cell-surface expression and current density were also obtained in the absence of KChIP2.2 when the highly conserved proximal Kv4 N terminus was deleted. The same domain is required for association of KChIP2.2 with Kv4 α-subunits. We propose that an efficient transport of Kv4 channels to the cell surface depends on KChIP binding to the Kv4 N-terminal domain. Our data suggest that the binding is necessary, but not sufficient, for the functional activity of KChIPs.


EXPERIMENTAL PROCEDURES
Cloning of KChIP2.2 cDNA-A DNA fragment encoding an N-terminal splice variant of human KChIP2 (GenBank TM /EBI accession number AF199598) was amplified by standard reverse transcription-PCR protocols from total RNA of human ventricle (see below) using primers CGCGGATCCACCATGCGGGGCCAGGGCCGCAAG (sense) * This work was supported by the Deutsche Forschungsgemeinschaft. 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 U.S.C. Section 1734 solely to indicate this fact. ‡ Present address: Bayer AG, Alfred-Nobel-Straße 50, 40789 Monheim, Germany.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF347114. and ATAAGAATGCGGCCGCTAGATGACATTGTCAAAGAGC (antisense), which contained BamHI and NotI restriction sites, respectively, at their 5Ј-ends. In addition, the sense primer contained a Kozak site (16) preceding the start codon. A single PCR product was obtained. After digestion with BamHI and NotI, it was subcloned into the mammalian expression vector pcDNA3 (Invitrogen). Sequencing revealed that the subcloned PCR fragment differed from the published KChIP2 sequence and represents a splice variant, which we named KChIP2.2 (GenBank TM /EBI accession number AF347114). A neuronal splice variant of KChIP2, KChIP2.1, which corresponds to the published KChIP2 sequence (1), was cloned from the total RNA of human cerebral cortex as described above.
In Vitro Translation and Immunoprecipitation Experiments-35 S-Labeled proteins were synthesized from the respective cDNAs (tagged KChIP2.2 and amino termini of Kv4.2 were cotranslated) by in vitro translation in the TNT coupled reticulocyte lysate system (Promega). Aliquots of the translation reaction were diluted 20-fold into immunoprecipitation buffer (20 mM Tris (pH 7.9), 150 mM NaCl, 0.5% Triton X-100, and 0.1% bovine hemoglobin) in precoated reaction tubes (immunoprecipitation buffer with 1% bovine hemoglobin). Polypeptides were immunoprecipitated by addition of anti-Myc antibody (9E10, Roche Molecular Biochemicals, Mannheim, Germany) and by protein A-Sepharose (Amersham Pharmacia Biotech) for 60 min at room temperature. Immunoprecipitates were washed twice in immunoprecipitation buffer, followed by wash buffer (20 mM Tris (pH 7.9) and 150 mM NaCl). To test the Ca 2ϩ dependence of KChIP2.2 binding to the Kv4.2 N terminus, the buffers contained either 1 mM CaCl 2 or 2 mM EGTA. Samples were boiled in SDS sample buffer for analysis on 15% SDSpolyacrylamide gels. Labeled proteins were visualized by autoradiography.
Functional Channel Expression-Recombinant Kv4␣ cDNA, either alone or in combination with KChIP2.2 cDNA, was transiently expressed after LipofectAMINE transfection of CHO or human embryonic kidney (HEK) 293 cells plated on poly-L-lysine (50 g/ml)-coated 35-mm plastic dishes at densities between 1 and 1.5 ϫ 10 4 cells/dish. For CHO cells, we used 0.1 g of Kv4␣ and 1 g of KChIP2.2 cDNA per dish. Since the expression of Kv4 channels was very low in CHO cells, this expression system was suitable only for studying the stimulatory effect of KChIP2.2 on Kv4 current densities. However, for a reliable analysis of the KChIP2.2 effect on Kv4 gating parameters, reasonably sized Kv4 current amplitudes (I Ն 1 nA at ϩ60 mV) are essential. These were obtained in HEK 293 cells when we used either 1 g of Kv4␣ cDNA alone or 0.1 g of Kv4␣ in combination with 1 g of KChIP2.2 cDNA. In all cases, 0.5 g of enhanced green fluorescent protein cDNA (CLON-TECH) was used to identify cells for electrophysiological recordings 12 h after transfection.
Electrophysiological Recordings and Data Analysis-During experiments, the cells were constantly superfused with an extracellular solution containing 135 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 2 mM MgCl 2 , 5 mM HEPES , 10 mM sucrose, and 0.01 mg/ml phenol red (pH 7.4) (NaOH). All recordings were performed in the whole-cell configuration of the patch-clamp technique at room temperature. Recording pipettes, pulled from thin-walled borosilicate glass capillaries using a DMZ puller (Zeitz, Augsburg, Germany), had bath resistances of between 2.5 and 3 megaohm when filled with an intracellular solution containing 125 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 11 mM EGTA, 10 mM HEPES, 2 mM K 2 ATP, and 2 mM glutathione (pH 7.2) (KOH). Currents were recorded with an EPC9 patch-clamp amplifier (HEKA, Lambrecht, Germany), and the program package PULSE ϩ PULSEFIT (HEKA) was used for data acquisition and analysis. Series resistance compensation was maximal (usually between 80 and 90%). Whole-cell configurations with series resistance values above 6 were not used for recordings. Activation and inactivation kinetics of Kv4-mediated currents were fitted simultaneously with Hodgkin-Huxley-related formalism (PULSE-FIT) as described previously (18). With this fitting procedure, the Kv4 current inactivation time course was approximated by two time constants, 1 and 2 . The voltage dependences of steady-state activation and inactivation were fitted with a Boltzmann function of the form G/G max ϭ (1/1 ϩ exp((V m Ϫ V1 ⁄2 )/k)), where V1 ⁄2 is the potential for half-maximal activation (V1 ⁄2,act ) or inactivation (V1 ⁄2,inact ), and the steepness of the voltage dependence is defined by the slope factor k. The kinetics of the recovery from inactivation were fitted with a single exponential function. Pooled data are given as means Ϯ S.E., and statistic analysis was done with GraphPAD InStat using unpaired t tests with Welch correction for unequal S.D. values when necessary.  (2)(3)(4)(5)17), Kv4.2 channels mediated typical A-type potassium currents (Fig. 2, A and B). Current densities (pA/pF) were estimated from the peak amplitudes of Kv4.2 currents that were recorded from each cell. The average Kv4.2 current density was very low (38 Ϯ 9 pA/pF, n ϭ 6) (Fig. 2C). In the presence of KChIP2.2, however, the Kv4.2 current densities were dramatically increased (2045 Ϯ 207 pA/pF, n ϭ 10) (Fig. 2, B and C). This represents a 55-fold stimulation, in comparison with a 9-fold increase in current density previously reported for coexpression of rat Kv4.2 with KChIP2.1 in CHO cells (1).
To further support our notion that the Kv4.2 N terminus is critical for association of KChIP2.2 with full-length Kv4.2 ␣-subunits, Kv4.2 and KChIP2.2-Myc were transiently expressed alone or together in CHO cells. We tagged Kv4.2 with a HA epitope to monitor the expression of Kv4.2 by immunofluorescent staining with anti-HA antibodies. KChIP2.2-Myc was stained with an anti-Myc antibody. When expressed alone, Kv4.2 was concentrated within the perinuclear endoplasmic reticulum and Golgi compartments, with some immunoreactivity apparent in the outer margins of the cell (Fig. 6A) as previously reported (1). Kv4.2⌬2-40-HA was markedly less concentrated in the perinuclear compartments than Kv4.2, and more immunoreactivity was apparent in the outer margins of the cell (Fig. 6B). KChIP2.2, on the other hand, was diffusely distributed throughout the cytoplasm of CHO cells (Fig. 6C), consistent with the diffuse distribution of KChIP1 in COS-1 cells (1). When KChIP2.2-Myc was coexpressed with Kv4.2-HA, the diffuse pattern of KChIP2.2-Myc immunofluorescence changed to such a degree that Kv4.2-HA and KChIP2.2-Myc immunofluorescence markedly overlapped, with a relatively high concentration of immunostain in the outer margins of the cell (Fig. 6, D-F). In contrast, when KChIP2.2-Myc was coexpressed with Kv4.2⌬2-40-HA, the patterns of both KChIP2.2-Myc and Kv4.2⌬2-40-HA immunofluorescence remained unchanged and did not overlap. No significant anti-KChIP2.2-Myc immunostain was detectable together with Kv4.2⌬2-40-HA in the outer margins of the cell (Fig. 6, G-I) (Fig. 7, A and B). Kv4.1⌬2-40 and Kv4.3⌬2-39 current densities in transiently transfected CHO cells were comparable to those observed for the wild type in the presence of KChIP2.2.
Next, we determined the kinetics of Kv4.1 and Kv4.3 currents Ϯ KChIP2.2. The results are summarized in Table II. As for Kv4.2 currents, KChIP2.2 did not markedly affect the activation time courses of Kv4.1 and Kv4.3 currents at ϩ40 mV, but slowed the inactivation time course. The respective increase in the inactivation time constant 1 was comparable to that observed upon coexpression of Kv4.2 with KChIP2.2 (Table I). In addition, the slower inactivation time constant 2 was increased, as well as the weighted amplitudes of 2 . Furthermore, KChIP2.2 accelerated the recovery of Kv4.1 and Kv4.3 channels from inactivation and shifted the midpoints of voltage activation (V1 ⁄2,act ) and of steady-state inactivation (V1 ⁄2,inact ) to more depolarized membrane potentials (Table II). In summary, the effects of KChIP2.2 on Kv4.1 and Kv4.3 channel expression and gating parameters were very similar, but not identical to those of KChIP2.2 on Kv4.2. DISCUSSION We have identified a KChIP2 splice variant in mRNA isolated from human ventricle. In comparison with the KChIP2.1 protein sequence (Ref. 1 and Fig. 1 Table I).
The mean time constant for the recovery from inactivation ( rec ) at Ϫ80 mV was 52.3 Ϯ 5.2 ms (n ϭ 7) for Kv4.2 ϩ KChIP2.1, faster than that for Kv4.2 alone (p Ͻ 0.0001), but not significantly different from that for Kv4.2 ϩ KChIP2.2 (p ϭ 0.5046) (see Table I). We also found that the voltage for half-maximal steady-state inactivation of Kv4.2 currents was shifted to more positive membrane potentials in the presence of KChIP2.1 (V1 ⁄2,inact ϭ Ϫ50.3 Ϯ 1.2 mV, n ϭ 6; p ϭ 0.0032), however, by a different degree as compared with KChIP2.2 (p Ͻ 0.0001) (see Table I). Thus, KChIP2.1 and KChIP2.2 distinctly altered Kv4 current properties. Nevertheless, the KChIP effects have in common an increased availability of active Kv4 channels by a rise in Kv4 channel density in the plasma membrane, slowed inactivation, an acceleration of the recovery from inactivation, and/or alteration of the operational range due to the change in the voltage dependence of steady-state inactivation.

Kv4.2/KChIP2.2 Current Properties Resemble Cardiac I to -
The results of Northern blot (6,19) and immunohistochemical (20) studies suggest that Kv4.2 and Kv4.3 subunits are expressed in cardiac tissue, in particular in ventricle (5, 15). Since KChIP2.2, which we isolated from cardiac tissue, physically   (22). Kv4 channels also inactivate rapidly, but the mechanism of inactivation is quite different and does not depend on the presence of an N-terminal Shaker-like inactivating domain (23,24). Yet, the Kv4 N termini also contain a highly conserved ϳ20-amino acid-long domain (Fig. 8) (15). The domain is characteristic for Kv4 N termini and is not found in other Kv subunit sequences. Our results indicate that the conserved Kv4 N-terminal domain may have important functions correlated with cell-surface expression of Kv4 channels and with binding KChIPs. The observed effects of KChIPs on Kv4 channel surface expression may be due to facilitated trafficking to the plasma membrane, delayed Kv channel turnover, altered interactions with the cytoskeleton, and/or alteration in intrinsic functional properties of the ion channel complex. We favor the notion that the Kv4 N terminus participates in regulation of Kv4 channel trafficking to the cell surface for the following reasons. In agreement with previous data (1), the Kv4.2 immunostaining pattern in the transiently transfected CHO cells showed an intense perinuclear Kv4.2 immunoreactivity in the absence of KChIP2.2. This is consistent with a retention of Kv4.2 protein in the endoplasmic reticulum or Golgi compartments. Deletion of the Kv4.2 N terminus or coexpression of Kv4.2 with KChIP2.2 produced a subcellular redistribution of Kv4 immunoreactivity to the cell membrane. Furthermore, our results showed that deletion of the Kv4.2 N terminus decreased the functional sensitivity of Kv4.2 channels to coexpressed KChIP2.2, most likely due to a loss of the KChIP-binding site on the Kv4 N terminus. Deletion of the conserved N terminus ablated KChIP2.2 binding to the Kv4.2 N terminus as well as KChIP2.2-mediated increases in current densities. At the same time, the deletions of the Kv4 N termini yielded a KChIPindependent increase in Kv4 channel density at the cell surface. Taken together, our observations strengthen the possibility that the Kv4 N terminus may contain an endoplasmic reticulum retention signal. Then, deletion of the Kv4 N terminus might attenuate the endoplasmic reticulum retention signal and accordingly stimulate a redistribution of Kv4 channels to the cell surface. Similarly, binding of KChIPs to the Kv4 N terminus may facilitate Kv4 channel trafficking to the cell surface by masking the endoplasmic reticulum retention signal.
A comparison of the expression data for Kv4.2 deletion mutants with Kv4.2/KChIP2.2 coexpression data shows some significant qualitative and quantitative differences with respect to current density increase. The deletions of the Kv4.2 N terminus gave rise to an up to ϳ30-fold KChIP-independent increase in Kv4.2 current density as compared with ϳ14and ϳ55-fold increases in Kv4.2 current density with coexpressed KChIP2.1 and KChIP2.2, respectively. Thus, binding of KChIPs to the Kv4.2 N terminus, as well as Kv4.2 N-terminal deletion, may cause different increases in current density. The data indicate, as discussed above, that the functional activity of KChIPs may be modulated by their variable amino termini. Furthermore, KChIP binding to the Kv4.2 N terminus is not Ca 2ϩ -dependent in vitro, but the functional activity of KChIPs is most likely Ca 2ϩ -sensitive because it requires intact EF-hands (1). Obviously, binding of KChIP to the Kv4 N terminus is necessary, but not sufficient, for functional KChIP activity. Future studies have to show whether the various mechanisms by which KChIPs regulate Kv4 channel activity involve distinct structures of the KChIP molecule, e.g. amino terminus, conserved core region, and Ca 2ϩ -binding sites.