Modulation of human Kv4.3/KChIP2 channel inactivation kinetics by cytoplasmic Ca2+
Abstract
The transient outward current (Ito) in the human heart is generated by Kv4.3 channels that interact with the cytoplasmic EF-hand protein KChIP2, known to modulate Kv4.3 inactivation when co-expressed. This study investigates Kv4.3 channels co-expressed with either wild-type (wt) or EF-hand-mutated (ΔEF) KChIP2 in human embryonic kidney (HEK) 293 cells. The experiments were conducted with and without BAPTA-AM, and with varying intracellular Ca2+ concentrations via the patch-pipette in whole-cell patch-clamp recordings. The findings reveal that Ca2+ is not essential for Kv4.3/KChIP2 complex formation. However, current decay was faster and recovery from inactivation was slower in the presence of 50 μM Ca2+ compared to nominally Ca2+-free conditions using BAPTA. Mutation of EF-hand 4 (ΔEF4) did not abolish this Ca2+-dependent effect, while mutation of EF-hand 2 (ΔEF2) eliminated it. Interestingly, mutation of EF-hand 3 (ΔEF3) converted the Ca2+-dependent slowing into an acceleration of recovery. When CaMKII activity was inhibited by KN-93, Ca2+ accelerated recovery kinetics even in the ΔEF2 background, but not in ΔEF3 or ΔEF4. These results suggest that Ca2+ binding to specific KChIP2 EF-hands, along with CaMKII activation, acutely modulates Kv4.3/KChIP2 inactivation gating. The study supports a model in which elevated Ca2+ levels and CaMKII activity influence Ito kinetics and frequency-dependent availability in cardiac cells.
Introduction
The transient outward current (Ito) plays a critical role in cardiac electrophysiology, contributing to phase 1 repolarization of the ventricular action potential. This process influences the amplitude and duration of the plateau phase governed by Ca2+ influx, thereby affecting both electrical propagation and excitation-contraction coupling in the myocardium. Ito is mediated by voltage-gated potassium channels from the Kv4 family, predominantly Kv4.3 in human hearts, which associate with cytoplasmic Kv channel-interacting proteins (KChIPs). Among the three members of the Kv4 subfamily (Kv4.1, Kv4.2, and Kv4.3), Kv4.3 is the primary contributor to Ito in human cardiomyocytes. The KChIP family consists of four members (KChIP1 through KChIP4), each with multiple splice variants, and KChIP2 is the isoform predominantly expressed in the heart.
The interaction between Kv4 channels and KChIPs is well characterized structurally and functionally. Both the N-terminal and C-terminal cytoplasmic domains of Kv4 subunits are involved in KChIP binding. The resulting Kv4/KChIP complex is believed to have an octameric configuration, with a 4:4 stoichiometry and KChIPs arranged around the Kv4 tetramerization domains. Co-expression of KChIPs, except for KChIP4a, enhances surface expression and significantly alters the inactivation properties of Kv4 channels. KChIPs typically slow the early phase and accelerate the late phase of current decay, producing a characteristic crossover when current traces are normalized. A hallmark effect of KChIP co-expression is the pronounced acceleration of recovery from inactivation.
KChIPs are part of the recoverin superfamily of Ca2+-binding proteins and closely related to neuronal calcium sensor protein 1 (NCS1 or frequenin). These proteins typically contain four EF-hand motifs, although EF1 is degenerate and non-functional. The importance of Ca2+ binding to KChIPs in their modulation of Kv4 channel expression and gating has been debated. While some studies indicate that Ca2+ binding is required for Kv4-KChIP association, others suggest that the interaction is Ca2+-independent. Additionally, free KChIPs undergo conformational and oligomeric changes in response to Ca2+, similar to other EF-hand proteins.
However, it remains unclear whether Ca2+ binding to KChIPs can acutely alter the gating of Kv4 channels already embedded in the membrane. To address this, the study co-expressed Kv4.3 with wild-type and EF-hand-mutated KChIP2b in HEK293 cells and analyzed macroscopic current decay and recovery kinetics using whole-cell patch-clamp techniques under varying intracellular Ca2+ conditions. The data suggest that acute modulation of Kv4.3/KChIP2 channel gating by Ca2+ depends both on direct EF-hand binding and CaMKII activity, with specific EF-hands playing distinct roles in this regulation.
Materials and Methods
Constructs and Heterologous Channel Expression
The short splice variant of human Kv4.3 was co-expressed with human KChIP2b (also known as KChIP2.1, hereafter referred to as KChIP2) using the pcDNA3 mammalian expression vector. To inactivate individual Ca2+-binding motifs in KChIP2, site-directed mutagenesis was performed by substituting the first aspartate residue in each 12-amino acid EF-hand motif with alanine. This generated the mutants KChIP2ΔEF2 (D135A), KChIP2ΔEF3 (D171A), and KChIP2ΔEF4 (D219A).
Three phosphorylation sites in Kv4.3 (Thr53, Ser535, Ser569) were mutated to alanine (Kv4.3AAA) to prevent phosphorylation, or to aspartate (Kv4.3DDD) to mimic phosphorylation. HEK293 cells were transiently transfected with these constructs using Lipofectamine 2000. Wild-type Kv4.3 was used at 0.1 μg cDNA per dish, while mutant forms were used at 0.2 μg. KChIP2 was co-transfected at 1 μg per dish, and enhanced green fluorescent protein (EGFP) cDNA (0.5 μg per dish) was included to identify transfected cells. In some cases, 10 μM BAPTA-AM was added to the culture medium immediately after transfection to buffer intracellular Ca2+.
Recording Techniques and Data Analysis
Whole-cell patch-clamp recordings were performed at room temperature (20–22 °C). The bath solution contained (in mM): NaCl 135, KCl 5, CaCl2 2, MgCl2 2, Hepes 5, and sucrose 10 (pH 7.4, adjusted with NaOH). Patch pipettes were pulled from borosilicate glass, fire-polished, and filled with internal solutions containing 0.8 mM free Mg2+ and varying free Ca2+ concentrations. Total ion concentrations were calculated using WEBMAXC Extended.
The nominal Ca2+-free internal solution contained (in mM): KCl 125, MgCl2 4.8, BAPTA 2, EDTA 2, EGTA 2, HEPES 10, sucrose 13, K2ATP 2, and glutathione 2 (pH 7.2, KOH). Internal solutions with 10 μM or 50 μM free Ca2+ included CaCl2 and adjusted MgCl2 and sucrose levels accordingly. The solutions referred to as BAPTA (Ca2+-free) and 50 μM Ca2+ were most commonly used.
In some experiments, 10 μM BAPTA-AM was added to both the culture medium and recording solutions (referred to as BAPTA-AM). CaMKII activity was inhibited using 1 μM KN-93, or an ineffective analog KN-92, applied both intracellularly and extracellularly.
Macroscopic currents were recorded at a test voltage of +40 mV using an EPC9 amplifier and Pulse software. Signals were filtered at 2.9 kHz, digitized at intervals of 350–1000 μs, and analyzed using PulseFit and Kaleidagraph. Current decay was fitted with a triple-exponential function to extract time constants (τ1, τ2, τ3) and their relative amplitudes (amp1, amp2, amp3). Recovery from inactivation was assessed using a double-pulse protocol, and fitted with a single- or double-exponential function depending on the experimental condition. Statistical comparisons were made using Student’s t-test, and data were presented as mean ± SEM.
Results: Cytoplasmic Ca2+ influences Kv4.3/KChIP2 inactivation kinetics
To determine whether cytoplasmic Ca2+ affects the inactivation kinetics of pre-existing Kv4.3/KChIP2 channel complexes, whole-cell patch-clamp recordings were performed using internal solutions containing either BAPTA (nominally Ca2+-free) or 50 μM free Ca2+. After establishing the whole-cell configuration, cells were repeatedly stimulated every 15 seconds for five minutes, allowing time for intracellular solutions to equilibrate.
Both conditions produced typical A-type currents, with a modest (\~10–15%) rundown in peak current amplitude over time. Despite this rundown, the decay kinetics of the currents were stable, and the final traces were analyzed. Currents recorded with BAPTA and 50 μM Ca2+ had comparable peak amplitudes, but differed in their inactivation properties. In the presence of 50 μM Ca2+, current decay was faster, primarily due to changes in the relative amplitudes of the exponential components. Specifically, amp2 was reduced and amp3 was increased, while changes in τ1–τ3 were minimal.
Theoretical current traces generated from the averaged fit parameters showed that higher cytoplasmic Ca2+ slightly diminished the typical slowing of inactivation associated with KChIP2 co-expression.
Recovery from inactivation was also affected by cytoplasmic Ca2+. At −80 mV, recovery was significantly slower when 50 μM Ca2+ was present compared to BAPTA. However, this effect was less pronounced at −100 mV and absent at −120 mV, indicating a voltage-dependent attenuation of KChIP2’s modulatory effect by Ca2+. These results suggest that although Kv4.3/KChIP2 channels can form and function without Ca2+, elevated cytoplasmic Ca2+ can partially suppress KChIP2-mediated modulation of inactivation and recovery kinetics.
KChIP2 EF-hands are involved in the calcium sensitivity of Kv4.3 inactivation
To determine if the observed calcium effects depend on EF-hand-mediated calcium binding by KChIP2, Kv4.3 was co-expressed with KChIP2 variants carrying individual EF-hand mutations: ΔEF2, ΔEF3, and ΔEF4. Electrophysiological recordings with calcium chelator BAPTA showed that all mutants retained some ability to modulate Kv4.3 gating, but their effects were generally weaker than those of wild-type KChIP2. This reduced modulation is consistent with previous findings using BAPTA-AM, supporting a role for calcium in enabling proper KChIP2-mediated modulation during channel formation.
When the intracellular calcium concentration was increased to 50 μM, inactivation time constants remained largely unchanged for all mutants. However, for the ΔEF3 mutant, the amplitude of the late component of inactivation was significantly reduced in the presence of calcium compared to BAPTA, suggesting EF-hand 3 plays a role in mediating calcium’s influence on late inactivation kinetics.
All EF-hand mutants accelerated recovery from inactivation, indicating that this function does not strictly require intact EF-hands. However, calcium sensitivity of the recovery phase varied by mutant. The ΔEF2 mutation eliminated the calcium-induced slowing of recovery observed with wild-type KChIP2. Interestingly, the ΔEF3 mutant exhibited faster recovery in the presence of calcium than with BAPTA, suggesting EF-hand 3 mediates an inhibitory calcium effect on recovery. The ΔEF4 mutant showed a recovery pattern similar to wild-type, though the calcium effect was less pronounced.
These findings suggest that individual EF-hands differentially contribute to calcium sensitivity in Kv4.3/KChIP2 inactivation and recovery processes. This differential involvement could result from distinct roles in calcium binding or structural dynamics, or it may indicate the presence of additional mechanisms beyond EF-hands in calcium-dependent modulation.
Calcium-dependent modulation of inactivation via KChIP2 EF-hands is influenced by CaMKII activity
Calcium-dependent pathways involving kinases and phosphatases represent alternative routes for calcium-modulated gating. Here, attention was focused on calcium/calmodulin-dependent protein kinase II (CaMKII), a known regulator of the transient outward current (Ito). Emphasis was placed on recovery from inactivation, which appeared more sensitive to changes in intracellular calcium than the kinetics of macroscopic current decay. CaMKII may modulate Kv4.3 channels by phosphorylating three potential cytoplasmic sites.
To explore this, mutant versions of Kv4.3 in which these sites were altered to either alanine (Kv4.3AAA) or aspartate (Kv4.3DDD) were tested. Recovery kinetics in the presence of either BAPTA or 50 μM calcium were found to be virtually identical for these mutants, suggesting calcium-dependent modulation requires phosphorylation at these sites.
To further assess CaMKII dependence, wild-type Kv4.3 channels were co-expressed with either wild-type or EF-hand mutant KChIP2, and recovery from inactivation was measured in both BAPTA and 50 μM calcium, with the CaMKII inhibitor KN-93 present. Under these conditions, two kinetic components were observed. The fast component reflected actual recovery, while the slower one was attributed to a different mechanism. In the absence of KChIP2, no calcium-dependent changes in recovery were observed. In contrast, Kv4.3/KChIP2 complexes exhibited calcium-dependent recovery differences. However, with KN-93, recovery was accelerated by calcium, inverting the usual calcium-induced slowing. This acceleration was also seen in Kv4.3/KChIP2ΔEF2, but not in Kv4.3/KChIP2ΔEF3 or Kv4.3/KChIP2ΔEF4. Notably, all EF-hand mutants showed very slow recovery in the presence of KN-93, indicating that CaMKII and KChIP2 both influence Kv4.3 channel modulation.
Discussion
This study demonstrates that co-expression of KChIP2 alters Kv4.3 inactivation characteristics, particularly recovery from inactivation, even when intracellular calcium is heavily buffered with BAPTA-AM during protein assembly. Recovery kinetics of existing Kv4.3/KChIP2 channels in the membrane were affected by altering cytoplasmic calcium concentrations. Kv4.3 channels were co-expressed with both wild-type and EF-hand-mutated KChIP2 in low and elevated calcium conditions, and the role of CaMKII was examined using KN-93.
Results suggest that calcium binding to EF3 and EF4 of KChIP2 accelerates recovery from inactivation, while CaMKII activity exerts an opposing influence, slowing recovery through interaction with EF2 and EF3. The specific structural roles underlying this interaction remain unresolved.
In neuronal calcium sensor proteins like KChIPs, calcium binding to EF-hands is known to induce relevant conformational changes. Studies using fluorescence and ligand-binding assays confirm that KChIPs bind calcium (and magnesium) through EF-hands. EF2 binds calcium with low affinity and is usually occupied by magnesium under physiological conditions, whereas EF3 and EF4 have higher calcium affinity. Structural data support the presence of calcium ions in EF3 and EF4 only.
Changes in calcium concentration are known to trigger conformational changes that influence inactivation and recovery kinetics. This raises two key questions: whether calcium binding to KChIP EF-hands affects channel complex formation and trafficking, and whether calcium binding acutely modifies channel gating in the membrane. Prior work has sometimes investigated these questions separately, but rarely both together. The present study systematically combined EF-hand mutations with calcium variation in electrophysiological experiments.
Previous structural studies showed that multi-EF-hand KChIP1 mutants are incapable of modulating Kv4.3 inactivation. Pull-down assays revealed that these mutants have impaired binding to Kv4.3 N-terminal fragments, suggesting a calcium-dependent assembly process. However, our BAPTA-AM results suggest that calcium may not be strictly necessary for assembly. Mutating EF-hand motifs individually does not drastically alter the protein structure, but this may not apply to multiple mutations. It remains possible that the lack of Kv4.3 binding in multiple mutants is calcium-independent.
KChIP binding to Kv4 full-length proteins may be less sensitive to calcium than binding to N-terminal fragments. The N-terminal sequence differences between Kv4.2 and Kv4.3 could also contribute to varying calcium sensitivity. Our previous work showed that wild-type KChIP2c binds a Kv4.2 N-terminal fragment in both calcium conditions, while EF-hand mutants showed reduced calcium-enhanced binding. The ΔEF3 mutation in particular showed weakened binding in calcium.
These findings support the idea that calcium binding to EF-hands is important for Kv4/KChIP complex formation during assembly. Co-expression of EF-hand mutants resulted in faster recovery kinetics, but did not slow fast inactivation. This supports the idea that EF-hand-mediated calcium binding during assembly modulates recovery, but is less involved in initial inactivation.
Other studies have reported that calcium enhances the binding of wild-type KChIP2 to N-terminal Kv4.2 fragments, while EF-hand mutants fail to show this enhancement. Specifically, the ΔEF3 mutant displayed weaker binding in the presence of calcium. This suggests that BAPTA-AM, by preventing calcium binding during assembly, may alter the KChIP-induced suppression of N-type inactivation.
To investigate whether existing Kv4.3/KChIP2 channels are directly sensitive to calcium on a short timescale, pipette solutions with controlled free calcium concentrations were used. Previous studies in neurons indicated that A-type currents mediated by Kv4 channels can be modulated by calcium. Specifically, calcium was found to reduce A-type current amplitude in hippocampal neurons, while in cerebellar neurons, high calcium levels enhanced this current, potentially through a slower decay of the current.
In our experiments, free calcium concentrations were selected to fall within the high-affinity range for KChIP2 binding. At a concentration of 50 micromolar calcium, calmodulin is also activated, which can initiate endogenous CaMKII activity. To differentiate the effects of CaMKII activation from the direct binding of calcium to KChIP2, we tested Kv4.3 channel mutants that are either resistant to phosphorylation (Kv4.3AAA) or mimic a phosphorylated state (Kv4.3DDD). Neither mutant exhibited calcium-dependent changes in recovery kinetics, which may be attributed to their inherently fast recovery properties. This suggests that physiological modulation of these channels is more complex than the effects observed by simply mutating all phosphorylation sites.
Further tests examined calcium’s effect in the presence of the CaMKII inhibitor KN-93, focusing solely on the fast component of channel recovery. Under these conditions, calcium was found to accelerate recovery in Kv4.3/KChIP2 channels, implying that direct calcium binding to KChIP2 enhances recovery speed. This accelerating effect was absent in mutants lacking EF-hand 3 (ΔEF3) or EF-hand 4 (ΔEF4), but remained in the mutant lacking EF-hand 2 (ΔEF2). This indicates that calcium binding to EF3 and EF4 is necessary for this modulation to occur.
In the absence of KN-93, calcium actually slowed recovery kinetics. The contrasting effects of calcium on recovery depending on the presence or absence of KN-93 underscore the important role of CaMKII. Additional biochemical evidence shows that CaMKII forms complexes with Kv4 channels, reinforcing its key role in regulating recovery kinetics through phosphorylation-dependent mechanisms.
Our findings indicate that calcium occupancy of KChIP2 EF-hands governs the function of the entire Kv4.3/KChIP2/CaMKII signaling complex. Although the structural details remain unclear, EF3 appears to be particularly important. Under control conditions, the effects of elevated calcium on Kv4.3/KChIP2 channels lacking EF3 resemble the effects observed in wild-type channels treated with KN-93. Few studies have focused on the individual roles of KChIP EF-hands in regulating Kv4/KChIP channels. One study examined a minimal KChIP2 splice variant, KChIP2d, which is only 70 amino acids long and contains a single EF-hand. KChIP2d modulates Kv4.3 inactivation similarly to longer KChIP2 variants. However, recovery kinetics data comparing conditions with and without calcium chelation does not support calcium-dependent modulation of Kv4.3/KChIP2d recovery. These results suggest that a single EF-hand, as in KChIP2d, is insufficient for the type of modulation observed in Kv4.3 channels complexed with longer KChIP2 variants.
Many examples of calcium-mediated effects involving KChIP EF-hands relate to gene expression regulation. KChIP3 can bind to downstream regulatory elements (DREs) in various genes in both excitable and non-excitable cells when calcium concentration is low, functioning as a transcriptional repressor. In this role, KChIP3 is often called DREAM (downstream regulatory element antagonist modulator). Elevated calcium and the resulting increased calcium binding to EF-hands cause conformational changes that release DREAM from DNA, permitting gene transcription. This was initially demonstrated for the prodynorphin gene in spinal cord neurons. Similarly, the Na+/Ca2+ exchanger 3 gene in cerebellar neurons is regulated by DREAM in a calcium-dependent manner. DREAM can also interact with cAMP-responsive element binding protein (CREB) in a calcium-dependent way, repressing CRE-dependent transcription. In cardiomyocytes, cytosolic DREAM has been found to translocate to the nucleus in a calcium- and CaMKII-dependent manner. There, DREAM may repress transcription of the CACNA1c gene, which encodes the voltage-dependent calcium channel Cav1.2. CaMKII activity thus appears to invert the effect of elevated calcium on the transcriptional repressor DREAM, similar to the opposing calcium effects observed on Kv4.3/KChIP2 inactivation depending on the presence of KN-93.
A direct regulation of ion channel gene transcription by KChIP2 in cardiomyocytes, comparable to DREAM’s mechanism, has not been demonstrated. However, recent findings suggest that KChIP2 may act as a calcium-dependent transcriptional repressor for certain microRNAs, such as miR-34b and miR-34c. These microRNAs in turn target ion channel genes in cardiomyocytes, including SCN5A, SCN1B, and KCND3, which encode the voltage-dependent sodium channel alpha-subunit Nav1.5, the Navβ1 subunit, and Kv4.3, respectively.
Earlier work by Anderson and colleagues demonstrated a direct link in rat cerebellar stellate cells between calcium influx through Cav channels and KChIP3-dependent regulation of A-type current. KChIP3 serves as a calcium sensor for Kv4.2 channels within a macromolecular signaling complex that includes Cav3 channels. The observed negative shift of about 10 mV in the steady-state inactivation curve of the A-type current when Cav channels were blocked with mibefradil suggests that calcium normally enters the cell and binds to KChIP3. This binding keeps the voltage dependence of steady-state inactivation in a more depolarized range, ensuring high availability of Kv4.2 channels for firing regulation. The calcium-dependent positive shift in the inactivation curve implies that EF-hand calcium binding enhances KChIP3’s effect on Kv4.2 inactivation, which aligns with the acceleration of Kv4.3/KChIP2 recovery kinetics seen in the presence of a CaMKII inhibitor. Slowed recovery from inactivation, which reduces channel availability, could represent maladaptive changes when calcium and CaMKII levels are elevated.
By analogy to the putative Kv4.2/KChIP3/Cav3 complex in the cerebellum, a similar macromolecular signaling complex composed of Kv4.3, KChIP2, and Cav1.2 may exist in cardiomyocytes. It is notable that KChIP2 can bind directly to Cav1.2 channels. This suggests that KChIP2 might functionally link Kv4.3 and Cav1.2 channels, creating a direct calcium-sensing mechanism via occupancy of KChIP2 EF-hands. The existence of such a Kv4.3/KChIP2/Cav1.2 complex in the heart, and the possible involvement of CaMKII—especially under pathological conditions such as hypertrophy or heart failure—warrants further electrophysiological investigation, potentially supported by mass spectrometry and high-resolution structural studies.