Corresponding author N. Decher: Department of Physiology, University of Utah, 95 N 2000 E, Salt Lake City, UT 84112, USA. Email: firstname.lastname@example.org
Kv4.3 channels conduct transient outward K+ currents in the human heart and brain where they mediate the early phase of action potential repolarization. KChIP2 proteins are members of a new class of calcium sensors that modulate the surface expression and biophysical properties of Kv4 K+ channels. Here we describe three novel isoforms of KChIP2 with an alternatively spliced C-terminus (KChIP2e, KChIP2f) or N-terminus (KChIP2g). KChIP2e and KChIP2f are expressed in the human atrium, whereas KChIP2g is predominantly expressed in the brain. The KChIP2 isoforms were coexpressed with Kv4.3 channels in Xenopus oocytes and currents recorded with two-microelectrode voltage-clamp techniques. KChIP2e caused a reduction in current amplitude, an acceleration of inactivation and a slowing of the recovery from inactivation of Kv4.3 currents. KChIP2f increased the current amplitude and slowed the rate of inactivation, but did not alter the recovery from inactivation or the voltage of half-maximal inactivation of Kv4.3 channels. KChIP2g increased current amplitudes, slowed the rate of inactivation and shifted the voltage of half-maximal inactivation to more negative potentials. The biophysical changes induced by these alternatively spliced KChIP2 proteins differ markedly from previously described KChIP2 proteins and would be expected to increase the diversity of native transient outward K+ currents.
KChIP proteins were first described as neuronal calcium sensors that bind to the N-terminus of Kv4 channels and modulate the surface expression and the rate of onset of IA inactivation (An et al. 2000). KChIP proteins also induce fast recovery from inactivation, a characteristic feature of native A-type K+ currents (An et al. 2000). Two KChIP2 isoforms were subsequently cloned from human heart, a 252 aa isoform (KChIP2b) lacking exon 3 and a 220 aa isoform (KChIP2c) lacking exons 2 and 3 (Bahring et al. 2001; Decher et al. 2001). Both isoforms altered the kinetics of Kv4.3 current in a manner that more closely resembled the native Ito,f current in epicardial tissues compared to Kv4.3 alone (Bahring et al. 2001; Decher et al. 2001). Ohya et al. (2001) reported that KChIP2a was present in the human heart. KChIP2c is the predominat isoform expressed in the human heart (Decher et al. 2001; Ohya et al. 2001) and the shift in the voltage dependence of inactivation by this isoform is more pronounced compared to KChIP2b (Decher et al. 2001). In humans and dogs the Ito,f gradient is paralleled by a KChIP2 mRNA-gradient, whereas in rodents the Ito,f gradient is paralleled by a Kv4 mRNA-gradient (Rosati et al. 2001). Subsequently, two additional KChIP2 isoforms, KChIP2t and KChIP2d, were described. KChIP2t (Deschenes et al. 2002) is identical to KChIP2c except for a short seven amino acid insert between exons 4 and 5. KChIP2d (Patel et al. 2002b) is the shortest of all the KChIPs and contains only the last three C-terminal exons. Surprisingly, this minimal KChIP2 isoform retains function, accelerating the recovery from inactivation and slowing the rate of inactivation of Kv4 currents (Patel et al. 2002b).
In this study we report three novel KChIP2 isoforms with alternative sequences in either the C-terminus (KChIP2e, KChIP2f) or the N-terminus (KChIP2g). These novel KChIP2 isoforms affect Kv4 properties differently from the previously described KChIP isoforms, extend the potential diversity of native transient outward potassium currents and provide new insights into KChIP domains involved in Kv4 channel regulation.
Cloning of alternatively spliced KChIP2 variants
The three KChIP2 isoforms cloned in this study were amplified from a cDNA that was obtained with primers annealing to the KChIP2 gene (5′-AAT TCC TAC AGG AGG GGC ACT C-3′ and 5′-GAA AGA ACT GGG AGT AAA TCT GCT TG-3′). PCR of human heart cDNA was used to amplify a KChIP2 isoform previously registered in the NCBI database as KChIP2.5 (AF367019). In this study, this isoform is called KChIP2f, in accordance with the most frequently used nomenclature for KChIP2 proteins (Table 1, Fig. 1A). Similar to KChIP2c (AF347114), KChIP2f lacks exons 2 and 3 of KChIP2a. However, in addition KChIP2f is missing exon 7 and has an alternative exon 10, such that the last 15 amino acids of the C-terminus differ from the sequence of KChIP2c. The regular exon 10 of KChIP2f is replaced by the alternative exon 10 at a non-canonical splicing site, as the alternative exon 10 is adjacent to exon 9 on genomic DNA. PCR amplification of this isoform generated two PCR products: KChIP2f and a novel isoform (KChIP2e) that retained exon 7.
Table 1. KChIP2 isoforms with the different nomenclature used
Cloning or expression
Given are the KChIP2 names, the authors, accession numbers and the species the KChIP2 isoforms were cloned from (n.y.a. = not yet assigned).
We also cloned a third KChIP2 splice variant with an alternative N-terminus that we named KChIP2g (Fig. 1A). The nucleotide sequence (AK027347) and protein sequence (BAB55052.1) of this isoform were previously submitted to the NCBI database, but not characterized in a functional assay. It is similar to KChIP2c as it lacks exons 2 and 3, but its N-terminus also differs in the first 29 amino acids (Fig. 1A and B). The clone was obtained by PCR amplification of human brain cDNA using primers binding to the alternative exon 1 and the regular exon 10. Table 1 summarizes the accession numbers and the names used for KChIP2 isoforms, including the three novel isoforms characterized in this study.
Amplified products were automatically sequenced on both strands (ABI 310, Perkin Elmer) and cloned into pSGEM vector (Villmann et al. 1997). For expression studies in Xenopus laevis oocytes, cDNA was linearized and capped cRNA was synthesized using CapScribe Buffer (Roche) and T7 polymerase (Roche).
PCR on cDNA from human heart and brain
Expression of the novel KChIP2 isoforms was analysed using PCR on cDNA from human heart tissue that was obtained from explanted hearts of patients with dilated cardiomyopathy (DCM). Stratascript (Stratagene) was used for reverse transcription reactions. Studies were performed according to the Declaration of Helsinki. Informed consent was obtained before organs were explanted and the procedure was approved by the Ethical Review Board of the University of Munich. In addition, commercial cDNA from human brain (Clontech) was analysed.
KChIP2e and KChIP2f were detected with primers binding to exon 1 and the alternative exon 10: 5′-CCG CAA GGA GAG TTT GTC CG-3′ and 5′-GGG TAA TGT AGA GGG CAG GGA GC-3′. KChIP2g was amplified with primers binding to the alternative exon 1 and regular exon 10: 5′-CAG CGC GAT CCC TCT ACC AGC-3′ and 5′-GGA GGC GTA GGA TGA GGA TAG ACC-3′. The C-termini of KChIP2 with regular exon 10 were amplified by primers binding to exon 6 and the 3′ UTR of the KChIP2 gene: 5′-CCT ATG CCA CTT TTC TCT TCA ATG C-3′ and 5′-GGA GGC GTA GGA TGA GGA TAG ACC-3′. The C-termini of KChIP2e and of KChIP2f were amplified with primers binding to exon 6 and alternative exon 10: 5′-CCT ATG CCA CTT TTC TCT TCA ATG C-3′ and 5′-GGG TAA TGT AGA GGG CAG GGA GC-3′. The annealing temperature for all primer combinations was 55°C. PCR products were verified by sequencing after subcloning into the pCRII TOPO vector (Clontech) using ABI-Sequencer 3700 (Kit Version 1.1).
Injection and voltage clamp of oocytes
Stage IV and V Xenopus laevis oocytes were isolated and injected with cRNA encoding Kv4.3 and/or KChIP2 isoforms using standard techniques (Goldin, 1991; Goldin & Sumikawa, 1992). Oocytes were isolated by dissection from adult Xenopus laevis. The frogs were anaesthetized by immersion in 0.2% tricaine for 15–20 min then placed on ice during dissection and removal of ovarian lobes. The incision was sutured closed and the frogs allowed to recover for about 1 month before removal of a second set of oocytes. Frogs were killed by pithing after anaesthetization with tricaine. RNA quality was judged by gel electrophoresis and its concentration quantified by UV spectroscopy using Ribogreen® (Molecular Probes, Eugene, OR, USA). Injected oocytes were cultured in Barth's solution supplemented with 50 μg ml−1 gentamycin and 1 mm pyruvate at 18°C for 1–3 days before use. Barth's solution contained (mm): 88 NaCl, 1 KCl, 0.4 CaCl2, 0.33 Ca(NO3)2, 1 MgSO4, 2.4 NaHCO3, 10 Hepes (pH 7.4).
For voltage-clamp experiments, oocytes were bathed in a modified ND96 solution containing (mm): 96 NaCl, 4 KCl, 1 MgC12, 1 CaC12, 5 Hepes (pH 7.6). Currents were recorded at room temperature (21–23°C) using standard two-microelectrode voltage-clamp techniques (Stuehmer, 1992) and a Geneclamp 500 amplifier, Digidata 1322a and pCLAMP 8 software (all from Axon Instruments, Union City, CA, USA).
The long isoform of human Kv4.3 (GenBank accession number NM_004980), the principal isoform in the heart (Dilks et al. 1999), was expressed alone or together with a single KChIP2 isoform. The voltage dependence of Kv4.3 channel inactivation was determined with a standard dual pulse protocol. From a holding potential of −80 mV, a 1.5 s conditioning pre-pulse to a variable voltage ranging from −100 to +10 mV was applied in 10 mV increments, followed by a 500 ms pulse to a test potential of +40 mV. Peak Kv4.3 currents recorded during the test pulse were normalized to the peak current measured following the most negative conditioning pre-pulse for each oocyte and the resulting relationship fitted with a Boltzmann function to estimate the half-point (V1/2) and the slope factor (k) for the voltage dependence of channel inactivation. Recovery from inactivation was analysed by a different dual pulse protocol. From a holding potential of −90 mV, channels were activated and inactivated by a 200 ms depolarization to +40 mV, then allowed to recover for a variable time at −90 mV before a test pulse to +40 mV was reapplied. The time constant for recovery from inactivation was obtained by a mono-exponential fit of the peak current amplitudes measured during the test pulse as a function of the recovery time at −90 mV.
To monitor the surface expression of channels, a haemagglutinin (HA) protein epitope was inserted between Ala-291 and Phe-292 located within the S3–S4 extracellular loop of Kv4.3 short (Acc. no. AF205856). The amino acid sequence of the HA-tagged region was 291YPYDVPDYA292. The chemiluminescence assay was performed as described (Zerangue et al. 1999). Briefly, oocytes were blocked with ND96/1% IgG-free bovine serum albumin, labelled with anti-HA antibody and horseradish peroxidase-conjugated secondary antibody sequentially, then washed with ND96 solution. Oocytes were immersed in 100 μl of SuperSignal enzyme-linked immunosorbent assay substrate (Pierce Biotechnology Inc., Rockford, IL, USA). Relative light units (RLU) were counted with a Spectra Max Gemini EM microtiter plate luminometer (Molecular Devices, Sunnyvale, CA, USA).
Clampfit 8 (Axon Instruments) and Origin 7 (OriginLab Corp ., Northampton, MA, USA) software were used for data analyses. All curve fitting procedures were based on the simplex algorithm. Data are reported as the mean ±s.e.m. (n= number of oocytes). Statistical differences of averaged data were evaluated by Student's unpaired t test. Significance was assumed if P < 0.05.
Different expression profiles of the KChIP2 isoforms with alternative N- or C-termini
Expression of the three novel KChIP2 isoforms was analysed by PCR on cDNA from human heart and brain. The C-terminal isoforms KChIP2e and KChIP2f are expressed in the heart, whereas KChIP2g is predominantly found in the brain (Fig. 2). Expression of KChIP2e and KChIP2f is restricted to the atrium, as no transcripts were detected in ventricles or in the brain (Fig. 2A). KChIP2g is predominantly expressed in the brain and barely detectable in the atria or ventricles (Fig. 2B). PCR amplification of the C-termini of KChIP2 isoforms with alternative exon 10 or regular exon 10 was in accordance with these findings, and it also appears that KChIP2e expression in the atria is much greater than KChIP2f (Fig. 2C and D). The C-termini of KChIP2 isoforms were amplified from exon 6 to the 3′ UTR after exon 10 (Fig. 2C) or from exon 6 to alternative exon 10 (Fig. 2D). Transcripts were detected for C-termini with regular (Fig. 2C) or alternative (Fig. 2D) exon 10, either containing exon 7 (upper arrow) or lacking this exon (lower arrow). PCR products of the C-termini with alternative exon 10, representing the isoforms KChIP2e and KChIP2f, were present in atria but not the brain (Fig. 2A and D).
KChIP2e and KChIP2f are expressed in atria from both non-failing (NF) and failing hearts (DCM). PCR was used to amplify the complete isoforms from the regular exon 1 to the alternative exon 10 (Fig. 2A). Both isoforms were detected in 1 out of 6 NF left atrial and 2 out of 9 NF left ventricular tissue samples (data not shown). Expression levels for both isoforms were higher in failing tissue (DCM): 4 out of 8 DCM left atria and 4 out of 9 DCM left ventricles (data not shown). The integrity of the cDNA in experiments where no KChIP2 isoforms were detected was confirmed by verification of GAPDH expression. Thus, these isoforms appear to be differentially expressed/regulated during dilated cardiomyophathy or heart failure in general; however, verification will require more samples.
KChIP2e and KChIP2f alter properties of Kv4.3 currents
We first describe the effects of the two KChIP2 isoforms (e and f) that were cloned from human heart cDNA. Expression of Kv4.3 cRNA in Xenopus oocytes induced a fast activating and inactivating A-type K+ current (Fig. 3A). Expression of KChIP2 isoforms alone did not generate currents in Xenopus oocytes (data not shown). Coinjection of Kv4.3 with an equal amount of KChIP2c or KChIP2f cRNA (15 ng each) significantly increased peak current amplitudes (Fig. 3B and D), whereas coinjection with KChIP2e decreased current (Fig. 3C). The relative changes in Kv4.3 current at +40 mV induced by coexpression of the KChIP isoforms is plotted in Fig. 3E. Current amplitudes were increased by factors of 2.28 ± 0.23 for KChIP2c (n= 10) and 1.42 ± 0.12 (n= 8) for KChIP2f. In contrast, coexpression with KChIP2e decreased relative current amplitude to 0.42 ± 0.12 (n= 8).
We next determined if KChIP-mediated changes in Kv4.3 magnitude were altered by decreasing the relative amounts of injected Kv4.3 to a Kv4.3/KChIP cRNA ratio of 1 : 10. Coexpression of 1.5 ng Kv4.3 cRNA with 15 ng of KChIP2c or KChIP2f cRNA induced the same changes in Kv4.3 magnitude as was observed when oocytes were injected with 15 ng of both cRNAs (compare Fig. 3E and G). In contrast, KChIP2e further decreased Kv4.3 at the 1 : 10 cRNA ratio (Fig. 3F) to a relative amplitude value of 0.11 ± 0.02 (n= 37; Fig. 3G). The concentration-dependent interaction between Kv4.3 and KChIP2e was further explored in two additional experiments using a cRNA ratio of 1 : 6. Coexpression of 2.5 ng Kv4.3 with 15 ng KChIP2e reduced relative current amplitude from 1.0 ± 0.09 (n= 9) for Kv4.3 alone to 0.57 ± 0.08 (n= 9) with KChIP2e. Reducing the total amount of injected cRNA by 50%, but maintaining the same 1 : 6 ratio (1.25 ng Kv4.3 plus 7.5 ng KChIP2e cRNA) reduced relative current amplitude to the same extent, namely from 1.0 ± 0.13 (n= 5) for Kv4.3 alone to 0.56 ± 0.11 (n= 5) with KChIP2e. These findings indicate that the reduction in Kv4.3 current amplitude by KChIP2e was not limited by translational capacity of the oocytes. Finally, we examined the effects of coexpressing equimolar (w/w ratio of 3 : 1) concentration of Kv4.3 and the individual KChIP2 isoforms (Fig. 3H). KChIP2c and KChIP2f increased current, whereas KChIP2e alone or in combination with KChIP2f decreased Kv4.3 current. As expected, these results indicate that the net effects of KChIP2 on Kv4.3 properties will depend on the relative cell-dependent expression levels of the different KChIP2 isoforms.
KChIP2c alters the properties of Kv4.3 channel inactivation (Bahring et al. 2001; Decher et al. 2001). We compared the effects of KChIP2 isoforms c, e and f on the voltage dependence and kinetics of Kv4.3 inactivation when the two cRNAs were expressed at a 1 : 1 ratio (15 ng each). The time constants for the onset of inactivation (τ) were determined for currents activated by depolarizing pulses to potentials ranging from +10 to +50 mV (Fig. 4A). Since the fast component of inactivation dominates Kv4 inactivation, time constants were determined from the first 200 ms of inactivating currents. For Kv4.3 alone τ was 69 ± 3 ms (n= 35) at +40 mV. Coexpression with KChIP2c and KChIP2f increased τ to 138 ± 9 ms (n= 10) and 91 ± 8 ms (n= 26), respectively. In contrast, KChIP2e decreased τ to 54 ± 2 ms (n= 18), an effect opposite to all previously described KChIP proteins. The effect of the KChIP2 isoforms c, e and f on the recovery from inactivation (τrecov) of Kv4.3 channels at −90 mV are shown in Fig. 4B. For Kv4.3 alone, τrecov was 220 ± 13 ms (n= 30). KChIP2c accelerated the recovery from inactivation as previously described (Bahring et al. 2001; Decher et al. 2001). KChIP2e slowed and KChIP2f had no effect on τrecov (Fig. 4B). Finally, we compared the voltage dependence of inactivation for Kv4.3 alone and in the presence of the KChIP2 isoforms c, e and f (Fig. 4C). For Kv4.3 alone the V1/2 of the relationship was −49.8 ± 0.6 mV (n= 17) with a slope factor (k) of 5.5 ± 0.3 mV. KChIP2c shifted the relationship to more positive potentials; however, the V1/2 and k for Kv4.3 inactivation were not significantly altered when Kv4.3 was coexpressed with KChIP2e (V=−51.0 ± 0.4 mV, k= 4.9 ± 0.1 mV, n= 14) or KChIP2f (V=−49.9 ± 0.7 mV, k= 5.2 ± 0.2 mV, n= 10). Expression of a 10-fold excess of KChIP2e (1.5 ng Kv4.3, 15 ng KChIP2e cRNA) that resulted in a strong reduction in current amplitudes (Fig. 3F) also did not cause a shift in the voltage dependence of Kv4.3 inactivation (Fig. 4D). Thus, unlike previously described KChIP2 isoforms , KChIP2 isoforms e and f had no effect on the voltage dependence of Kv4.3 inactivation. The effects of KChIP2 isoforms on the properties of Kv4.3 channel inactivation are summarized in Table 2.
Table 2. Biophysical properties of currents for Kv4.3 and Kv4.3/KChIP channels expressed in Xenopus oocytes
Kv4.3 + KChIP2c
Kv4.3 + KChIP2e
Kv4.3 + KChIP2f
Kv4.3 + KChIP2g
Values are mean ±s.e.m.*P < 0.05 versus Kv4.3.
Relative current amplitude
1.00 ± 0.08
2.28 ± 0.23*
0.42 ± 0.12*
1.42 ± 0.12*
1.96 ± 0.30*
at +40 mV
τinact at +40 mV (ms)
69 ± 3
138 ± 9*
54 ± 2*
91 ± 8*
127 ± 7*
τ, recovery from inactivation
220 ± 13
* 48 ± 9
* 411 ± 25
235 ± 17
92 ± 8*
at −90 mV (ms)
V 1/2inact (mV)
−49.8 ± 0.6
−33.9 ± 1.4*
−51.0 ± 0.4
−49.9 ± 0.7
−57.9 ± 0.4*
Slope factor, k (mV)
5.5 ± 0.3
4.6 ± 0.3*
4.9 ± 0.1*
5.2 ± 0.2
3.5 ± 0.2*
KChIP2g shifts voltage dependence and slows inactivation of Kv4.3 channel
Kv4.3 amplitude was enhanced by coexpression with KChIP2g (Fig. 5A and B). At +40 mV, peak outward current was increased by a factor of 1.96 ± 0.30 (n= 12) compared to Kv4.3 alone (Fig. 5C). KChIP2g also slowed the onset of Kv4.3 inactivation in a voltage-independent manner (Fig. 5D). Fitting the first 200 ms of the inactivating currents to a mono-exponential function, the time constant of Kv4.3 inactivation at +40 mV was 69 ± 3 ms (n= 35) and was increased by coexpression with KChIP2g to 127 ± 7 ms (n= 13). Because slowing of inactivation was pronounced for KChIP2g, we also analysed the influence of this isoform on the slow time component of inactivation during the 1.5 s voltage steps. KChIP2g slowed the fast component of inactivation but had only a minor effect on the rate of slow inactivation (Fig. 5E). KChIP2g also increased the relative contribution of the fast component of Kv4.3 channel inactivation (Fig. 5F). In contrast to the effects on the onset of Kv4.3 inactivation, KChIP2g accelerated the recovery from inactivation (Fig. 5G). At −90 mV, τrecov was 220 ± 13 ms (n= 30) for Kv4.3 alone compared to 92 ± 8 ms (n= 8) for Kv4.3/KChIP2g. Thus, like KChIP2c, coexpression of KChIP2g increased the magnitude, slowed the onset of inactivation and accelerated the recovery from inactivation of Kv4.3 current. However, whereas KChIP2c shifted the voltage dependence of Kv4.3 inactivation to more positive potentials (Fig. 4C), KChIP2g had the opposite effect (Fig. 5H, Table 2) and shifted the V1/2 for inactivation by −24 mV to −57.9 ± 0.4 mV (n= 8).
KChIP2e reduces surface expression
Reduced current amplitudes of Kv4.3 channels by KChIP2e could result from altered gating properties (lack of channel opening, reduction in single channel conductance, or open probability) or reduced surface expression. To test whether the reduction in Kv4.3 current amplitudes by KChIP2e is paralleled by a reduction in surface expression, an HA epitope was inserted in the S3–S4 linker of the short Kv4.3 isoform (see Methods). This channel (Kv4.3-HA) inactivated at a more negative potential (V=−118.0 ± 1.1 mV; slope factor, k= 8.36 ± 0.75; n= 8) than WT Kv4.3 channels, but when currents were elicited from a more negative holding potential, currents were of similar magnitude to WT channels. Coexpression of Kv4.3-HA with KChIP2e in a 1 : 10 ratio (w/w) reduced relative current amplitudes from 1.0 ± 0.15 to 0.3 ± 0.08 (n= 9). The chemiluminescence signal (in RLU) of oocytes expressing Kv4.3-HA with KChIP2e (1 : 10 ratio w/w) was also reduced when compared with Kv4.3-HA (Fig. 6). The relative surface expression of protein was increased 1.90 ± 0.15 fold (n= 12) for coexpression with KChIP2c and it decreased by the factor 0.25 ± 0.06 (n= 12) for coexpression with KChIP2e. The reduction of surface expression of HA-tagged Kv4.3 short channels is consistent with the current reduction observed by this KChIP2 isoform. These data suggest that the reduction of current amplitude caused by coexpression of Kv4.3 and KChIP2e was primarily caused by reduction of surface expression and not altered gating or single channel properties.
Coexpression of human KChIP2c with Kv4.3 reconstitutes several features of native ventricular epicardial Ito,f (Bahring et al. 2001; Decher et al. 2001) and native A-type currents of the brain (An et al. 2000), including the voltage dependence of inactivation and rate of recovery from inactivation. Most KChIP2 isoforms slow the rate of inactivation of Kv4.3 currents. In isolated human cardiomyocytes, Ito,f inactivates faster than heterologously expressed Kv4.3/KChIP2 channels. The cause of this apparent discrepancy is confused, however, because the use of nifedipine to block Ca2+ currents and higher temperature used to record Ito,f can accelerate the rate of inactivation (Wettwer et al. 1994). Our discovery of a human cardiac KChIP2 isoform (KChIP2e) that accelerates the rate of Kv4.3 inactivation could potentially account for some of the apparent discrepancy in kinetics between KChIP2/Kv4.3 currents and native Ito,f. Although KChIP2e and KChIP2f are present in the human heart, their functional role in vivo is unknown. To evaluate the physiological relevance of these and other isoforms it will be necessary to determine their expression patterns relative to one another in different regions in the heart. This type of analysis is further complicated by our finding that KChIP2e has concentration-dependent effects on the biophysical properties of Kv4.3.
Regardless of the possible physiological relevance of the various KChIP2 isoforms to native Ito,f, characterization of these isoforms can provide insights into the structural basis of KChIP2 modulation of Kv4.3. All the alternatively spliced KChIP2 isoforms studied here are missing exons 2 and 3. KChIP2c has no other structural changes relative to KChIP2a or 2b (Fig. 1). In addition, isoforms 2e and 2f have an alternative exon 10 and isoform 2g has an alternative exon 1. Finally, isoform 2f is the same as isoform 2e, except that it also lacks exon 7. Thus, for purposes of comparison, KChIP2c is a useful standard basis of reference. The absence or alternative splicing of specific exons can be correlated with some, but not all, of the differential effects of a KChIP2 isoform on Kv4.3. First, KChIP2e reduced the magnitude of Kv4.3, accelerated the onset of inactivation and slowed the recovery from inactivation, effects opposite to all the other isoforms. The only structural difference between KChIP2c and KChIP2e is the alternative exon 10 of KChIP2e. This alternative splicing in the C-terminus strongly reduces the surface expession of Kv4.3. Alternative splicing of exon 10 cannot exclusively explain these findings because KChIP2f has the same exon 10 as KChIP2e but increased Kv4.3 amplitude, albeit to a lesser extent than the KChIP2c or g. However, KChIP2f had no effect on the recovery from inactivation and only slightly slowed its onset, implicating a potentially important role for exon 7 in these effects. Second, the presence of an alternative exon 10 in isoforms 2e and 2f was associated with a loss of the typical effect (positive shift in V1/2) on channel gating associated with inactivation, whereas an alternative exon 1 (KChIP2g) caused the voltage dependence of Kv4.3 inactivation to be shifted in the negative direction. Patel et al. (2002b) demonstrated the importance of the C-terminus in mediating KChIP2 effects on the rates for onset and recovery from inactivation. The minimal KChIP2d isoform they reported contained only exons 8–10 but slowed the rate of inactivation and accelerated the recovery from inactivation of Kv4.3 (Patel et al. 2002b). Mutations in the EF-hand of exon 9 did not change the ability of KChIP2d to speed the recovery from inactivation (Patel et al. 2002b). Although findings with KChIP2d raise questions about the functional importance of the N-terminus in modulating Kv4 currents (Patel et al. 2002b), other findings support such a role. These include (1) the shift in the voltage dependence of inactivation by KChIP2c being more pronounced compared to KChIP2b, (2) the reversed shift in voltage dependence of inactivation induced by KChIP2g, and (3) the report of an N-terminal alternatively spliced KChIP4 isoform that prevents A-type inactivation of Kv4 (Holmqvist et al. 2002). These findings highlight the relevance of the N-terminus in modulating the rate of Kv4 inactivation and the voltage dependence of inactivation.
The novel KChIP2 isoforms reported here increase the potential functional diversity of transient outward (A-type) currents. Moreover, KChIP2e and KChIP2f demonstrate the importance of the C-terminus in influencing current density, and the rates of inactivation and recovery from inactivation of Kv4. Results with KChIP2g demonstrate that regions other than the C-terminus are required for functionality and that the N-terminus plays an important role in modulating the voltage dependence of inactivation. Evaluation of the physiological relevance of these isoforms in determining the properties of Ito,f in different regions of the brain and heart under normal conditions and in disease is the main challenge for future studies.
We thank Christina Kaufmann, Peter Westenskow and Krista Kinard for excellent technical assistance.
Requests for reagents
Requests for reagents should be directed to Klaus Steinmeyer, email@example.com