Structural basis for K_V7.1–KCNE_x interactions in the I_Ks channel complex

Abstract

The cardiac I_Ks current is involved in action potential repolarization, where its primary function is to limit action potential prolongation during sympathetic stimulation. The I_Ks channel is mainly composed of K_V7.1 ion channels associated with KCNE1 auxiliary subunits. The availability of KCNE1 solution structure by nuclear magnetic resonance spectroscopy in conjunction with biochemical assays addressing K_V7.1–KCNE1 residue interactions has provided new insights into the structural basis for K_V7.1 modulation by KCNE1. Recent evidence further suggests that KCNE2 may associate with the K_V7.1–KCNE1 channel complex and modulate its current amplitude. Here we review recent studies in this area and discuss potential roles for multiple KCNE_x subunits in I_Ks generation and modulation as well as the clinical relevance of the new information.

Introduction

In human heart, the cardiac delayed rectifier current comprising I_Ks (cardiac slow delayed rectifier current) and I_Kr (cardiac rapid delayed rectifier current) is an important determinant of action potential duration. With its slow rate of activation, I_Ks primarily contributes to action potential repolarization during β-adrenergic stimulation, when its current amplitude is increased and rate of activation accelerated via the protein kinase A pathway. A number of studies have identified different signaling molecules, such as calmodulin and phosphatidylinositol 4,5-bisphosphate (PIP₂), which contribute to regulation of I_Ks and I_Kr in the heart (for an excellent review on these topics, see Charpen-tier et al¹). The α-subunit that mediates I_Ks is K_V7.1 (also known as KCNQ1 or KvLQT1; see “The I_Ks Babylon” in the Online Supplemental Data). K_V7.1 channels are tetramers, with each subunit containing six transmembrane segments forming peripheral voltage-sensing domains (S1–S4) and a central pore domain (S5–S6) (Figure 1A). The significance of K_V7.1 in normal heart function is highlighted by more than 240 identified KCNQ1 mutations associated with arrhythmias such as long QT syndrome, short QT syndrome, and atrial fibrillation (http://www.fsm.it/cardmoc/). In the majority of cases, K_V7.1 mutations associated with loss of function of the I_Ks current appear to result in long QT syndrome, whereas gain-of-function mutations lead to short QT syndrome or atrial fibrillation. However, K_V7.1 mutations simultaneously linked to long QT syndrome and atrial fibrillation have been reported.² As K_V7.1 properties are differentially modulated by the KCNE accessory subunits (Figure 1), this complexity may be at least partially due to a heterogeneous pattern of K_V7.1 association with different KCNE subunits in the heart.

Figure 1.

K_V7.1–KCNE1–KCNE2 current characteristics. A: Topology of K_V7.1 and KCNE subunits. K_V7.1 subunits encompass six transmembrane segments with intracellular N- and C- termini, where S1–S4 encode the voltage-sensing domain (VSD) and S5–S6 encode the pore domain (PD). The KCNE proteins contain a single transmembrane segment flanked by an extracellular N-terminus and a cytosolic C-terminus. B: Normalized amplitudes of currents elicited from Xenopus oocytes upon stimulation with the indicated voltage clamp protocol 48 hours after injection of K_V7.1 and KCNE1–KCNE2 cRNA as a function of clamp potential reveals that KCNE1 shifts the voltage-dependence of K_V7.1 activation in the depolarizing direction to a much greater extent than does KCNE2. C: Representative current traces from the experiments analyzed in B illustrate the diverse effects of KCNE proteins on K_V7.1 channel properties. Homomeric K_V7.1 channels activate relatively rapidly, exhibit sustained currents at maintained depolarization, and slowly deactivate upon repolarization. Co-expression with KCNE1 greatly increases K_V7.1 current amplitude and slows activation, in addition to its effect on channel voltage-dependency. Co-expression with KCNE2 reduces current amplitudes and induces a constitutively active current component. Activation of the time-dependent current component is slower than activation in the absence of KCNE2.²⁸ Co-expression with KCNE1 and KCNE2 generate currents with a mixture of the features, where KCNE1 dictates kinetics, whereas both KCNE1 and KCNE2 contribute to determination of current amplitude.^2,29D: Effects exerted by KCNE subunits on K_V7.1 currents. Strong effects are highlighted in bold.

Compelling evidence has established that KCNE1 is the major accessory subunit of the I_Ks channel. KCNE1 increases K_V7.1 channel conductance, shifts its activation to a more positive voltage range, and, importantly, confers the unique slow activation rate of I_Ks. Computational work suggests that KCNE1 resides in a cleft between voltage-sensing domains in the K_V7.1 channel structure. In support of this model, three K_V7.1 mutations associated with cardiac arrhythmia that reveal their phenotype only upon co-expression with KCNE1 all localize to a voltage-sensing domain–pore domain interface that is part of the open-state cleft where KCNE1 resides.³ Since 1999, other members of the KCNE family have been cloned and characterized⁴ (note that KCNE1 is equivalent to minK, and KCNE2–KCNE5 corresponds to minK-related peptides or MiRP1-4 in previous nomenclature; see “The I_Ks Babylon” in the Online Supplemental Data). All KCNE genes are reportedly transcribed into mRNA in the human heart,⁵ and expression of KCNE1–KCNE4 proteins has been detected. Emerging evidence suggests a role for KCNE2 in regulating I_Ks as well. KCNE2 reduces the K_V7.1 current amplitude and confers a constitutively active current component (Figures 1B–1D).⁶ If I_Ks in some cardiac myocytes is mediated by a K_V7.1 channel complex encompassing KCNE2, it would have major functional consequences. Understanding the structural requirements for KCNE modulation of K_V7.1 is important for delineating the role of KCNE subunits in I_Ks generation and regulation. Here we review current knowledge of the structural basis for K_V7.1–KCNE1 interactions (with focus on the membrane-spanning regions), describe evidence for the presence of additional KCNE proteins in the channel complex (with focus on KCNE2), and discuss the clinical relevance of these recent findings.

K_V7.1–KCNE1 channel stoichiometry

To resolve the structural basis for K_V7.1–KCNE_x interactions, it is essential to know the stoichiometry of the channel complex. The number of KCNE subunits in the I_Ks complex has been a matter of debate. Recently, an elegant approach was used: iterative rounds of channel blocking/modification by a chemically releasable channel inhibitor were used to bind the channel pore and simultaneously covalently modify a cysteine-bearing KCNE1 subunit. The study convincingly showed that K_V7.1 and KCNE1 associate in a 4:2 stoichiometry.⁷ There are at least two possible mechanisms for such a stoichiometry in K_V7.1–KCNE1 channel complexes. First, K_V channel assembly may involve first dimerization of α-subunits, followed by dimerization of dimers to form a tetramer. For K_V7.1–KCNE1 channels, KCNE1 has been speculated to be involved in such a dimeric assembly process.⁷ Recent reports support this mechanism. The carboxy end of KCNE1 (aa 109–129) binds to a double-helix bundle (helices C) in the cytoplasmic carboxy-terminus of K_V7.1,⁸ and these helix C bundles may be involved in dimerization of K_V7.1 dimers during channel biogenesis.⁹ Second, the KCNE1 solution nuclear magnetic resonance (NMR) structure implies that binding of additional KCNE1 subunits in the K_V7.1–KCNE1 channel complex may face steric hindrance due to helical secondary structures present in the KCNE1 extracellular and intracellular domains, which form an interface large enough to span multiple K_V7.1 subunits.³

KCNE1 location in K_V7.1 channels

Experimental observations suggest that KCNE1 is located in close proximity to the K_V7.1 pore domain so that the transmembrane segment of KCNE1 interacts directly with S6 of K_V7.1.^10–12 For example, it has been shown that KCNE1 position 42 can come close to K_V7.1 position 324 in S6 in the open state so that cysteine side chains engineered into these two positions can form a disulfide bond and lock the channel in the open state.¹³ In addition to interacting with the pore domain, KCNE1 interacts with the K_V7.1 voltage-sensing domain. For example, K_V7.1 position 226 in S4 can come close to KCNE1 position 44 in the open state so that cysteine side chains engineered into these positions can form a disulfide bond and stabilize S4 in its activated state.¹⁴ These interactions may be related to the ability of KCNE1 to right-shift the voltage-dependency of K_V7.1 channel activation. Although the association of KCNE1 with the K_V7.1 channel is mediated mainly by the transmembrane regions, secondary association between the carboxyl end of KCNE1 and bundles of helices C of K_V7.1 also contribute.⁸

Based on the K_V1.2 crystal structure and the KCNE1 solution NMR structure, models for the K_V7.1–KCNE1 complex in the open and the closed states have been proposed (Figure 2).³ Computational analysis of the channel complex localizes KCNE1 to a cleft between pore and voltage-sensing domains but with different interaction interfaces in the two states.³ In the closed state, the extracellular end of KCNE1 transmembrane segment forms contacts with S3 of one subunit and with S5 and S6 of another subunit, whereas in the open state, it forms an interface with S1, S5, and S6 from three different subunits.³ In general, these predictions are in agreement with functional studies using cysteine scanning mutagenesis and disulfide trapping approaches.^12,13 Hence, although still not completely resolved, a picture emerges that KCNE proteins are in contact with several helices from distinct subunits of the K_V7.1 channel.

Figure 2.

Model of K_V7.1–KCNE1 channel structure. K_V7.1–KCNE1 closed- and open-state models as generated by Kang et al³ and graphed using UCSF-Chimera. Top: Docking of the nuclear magnetic resonance (NMR) structure of one KCNE1 transmembrane segment (red) onto a homology model of K_V7.1 in the open (left) and closed (right) state, respectively, shows that KCNE1 resides in different clefts in the two states. The channel complex is visualized from the extracellular side. K_V7.1 subunits are colored gray, pink, blue, and purple, with S1–S6 and P helices labeled for one subunit. The voltage-sensing domain (VSD) and pore domain (PD) are also marked. Yellow spheres indicate the extracellular ends of S1 (I145), S4 (A226), and S6 (V324). In the open state, the extracellular end of the KCNE1 transmembrane segment forms an interface with S1, S5, and S6 from three different subunits; in the closed state, it forms contact with S3 of one subunit and S5 and S6 of another subunit. Bottom: Enlarged and rotated view of the boxed regions in the top panels, with addition of 10 NMR structures of KCNE1 residues 40–45 (color coded differently for individual NMR structures). This was achieved by aligning the 10 NMR KCNE1 structures (residues 40–70) to the KCNE1 transmembrane segment of the K_V7.1–KCNE1 model⁶ using the MatchMaker function of UCSF-Chimera (for clarity, only residues 40–45 are shown). Assuming (1) a wide swing of S1–S4 upon transition between closed and open states and (2) changing KCNE1 position and orientation relative to K_V7.1 in these states, the model suggests a wide range of motions in the extracellular end of the KCNE1 transmembrane segment (highlighted by the red circle) and possibilities of contacts with K_V7.1. This model is partially supported by the proximity of –S–S– partners as delineated in cysteine mutagenesis, although further experimental confirmation is needed.

KCNE1 interaction with K_V7.1 voltage sensor

Computational model simulations of K_V7.1–KCNE1 interactions suggest that the intracellular end of the KCNE1 transmembrane segment is close to the S5 end of the S4–S5 linker in K_V7.1. This may be the mechanism by which KCNE1 opposes S4–S5 linker movement and accordingly slows K_V7.1 channel opening.^3,15 This scenario is supported by functional studies reporting differential effects of S4–S5 linker mutations dependent on channel assembly with KCNE1.^16,17 Furthermore, KCNE1 is proposed to impede salt bridge formation between charged residues in S2 and S4 of the K_V7.1 voltage sensor.¹⁸ The pattern of salt bridge formation appears to differ between the channel's open and closed states, which correlates with axial rotation of KCNE1 transmembrane segments during I_Ks gating transitions between the two states.¹³ Compelling evidence thus supports a strategic location of KCNE1 between the K_V7.1 voltage sensor and pore domains, which allows KCNE1 to effectively modulate the K_V7.1 gating transition between the closed and open states.

How does KCNE1 slow K_V7.1 activation?

KCNE1 can transform K_V7.1-mediated currents into the characteristic slowly activating I_Ks by two possible mechanisms. KCNE1 either slows voltage sensor movement or delays coupling of voltage sensor movement to the activation gate. A careful study combining cysteine scanning mutagenesis and varying voltage-clamp pulse durations suggests that K_V7.1–KCNE1 voltage sensors reach their equilibrium position in less than 100 ms.¹⁹ Together with KCNE1's restriction of S4–S5 linker movement, it seems likely that the rate-limiting step in K_V7.1–KCNE1 channel activation is in the opening of the activation gate. In K⁺ channels, conformational changes allowing gating are attributed to a so-called gating hinge located deep within the membrane. In most K⁺ channels, the hinge is a conserved glycine residue, but K_V7.1 channels carry an alanine at the homologous position (336). K_V7.1 comprises a Pro-Ala-Gly sequence further downstream on S6, which corresponds to the K_V channel Pro-Val-Pro motif that adds to the flexibility of the region. In the absence of KCNE1, the ability of K_V7.1 to kink at Ala 336 facilitates gating, whereas the integrity of the Pro-Ala-Gly motif is required for channel activation.²⁰ Thus, it can be speculated that an additional mechanism by which KCNE1 slows K_V7.1 activation is by restraining the tendency of S6 to bend at either of these regions.

Structural determinants for diverse KCNE effects on K_V7.1 channels

Additional members of the KCNE family may contribute to I_Ks generation. The five KCNE proteins all have diverse effects on K_V7.1 conduction properties in heterologous systems, and their importance for cardiac function is evidenced by co-segregation of mutations in each of them and cardiac disorders (for reviews on these subjects, see Charpentier et al¹ and McCrossan and Abbott⁴). A cysteine scanning mutagenesis study suggested a similar position and orientation of the KCNE2 transmembrane segment with respect to K_V7.1 as found for KCNE1, with one face of KCNE2 transmembrane segment making intimate contact with the K_V7.1 pore domain and another face making more dynamic contacts with the voltage-sensing domain.⁶ However, it appears that minor structural differences determine the specific modulatory effects of KCNEs on K_V7.1. For instance, the drastic effect of KCNE3 in converting K_V7.1 into a constitutively active channel by shifting the equilibrium position of its voltage sensors to the activated state at all potentials^19,21 can be altered into that of KCNE1 by swapping just one amino acid within the transmembrane segment.²² Likewise, exchange of a single amino acid within the KCNE2 transmembrane segment converts the K_V7.1 modulatory properties of KCNE2 to those of KCNE1.⁶ It seems that discrete contact points between K_V7.1 and the KCNEs are determinants of the distinct modulatory effects of the different KCNE subunits on the K_V7.1 channel.

KCNE2 as a dynamic I_Ks regulator

As KCNE2–KCNE4 protein expression has been confirmed in human heart,^23–25 I_Ks may not be a simple channel complex encompassing four K_V7.1 subunits and two KCNE1 subunits. Experiments using heterologous expression systems have shown that different members of the KCNE family can associate with the same K_V7.1 channel simultaneously,^26–29 and exchange of KCNE1–KCNE2 subunits can occur even during the lifespan of K_V7.1–KCNE_x channel complexes in cells.²⁹ Because only KCNE1 can confer the unique slow activation rate of I_Ks, it is an obligate part of the I_Ks channel complex. Given the constraint that K_V7.1 and KCNE proteins associate in a 4:2 stoichiometry, it is conceivable that a portion of I_Ks channels in human myocytes may be mediated by a channel complex consisting of four K_V7.1 subunits, one KCNE1 subunit, and one KCNE subunit other than KCNE1. Thus, spatial heterogeneity in the KCNE expression profile is a potential mechanism for cellular regulation of I_Ks density.

Experiments performed on rat cardiomyocytes suggest a role for KCNE2 in I_Ks.^27,28 As KCNE2 decreases the amplitude of K_V7.1–KCNE1 currents (Figure 1C),^2,28 it is tempting to speculate that KCNE2 functions as a down-regulator of K_V7.1–KCNE1-mediated I_Ks (Figure 3).^27–29 The feasibility of this proposition is supported by experiments on adult guinea pig ventricular myocytes, where expression of KCNE2 protein by adenovirus-mediated gene transfer reduces I_Ks current density but not its gating kinetics or K_V7.1 protein level.²⁹

Figure 3.

Possible heterogeneity in I_Ks channel components. I_Ks density may be dynamically regulated by the composition of KCNE proteins in the channel complex. Expression of K_V7.1–KCNE1 channels gives rise to greater current amplitude than do K_V7.1–KCNE1–KCNE2 channels. As KCNE2 can modulate current mediated by K_V7.1–KCNE1 channels in the plasma membrane, it can be speculated that KCNE2 may function as a dynamical down-regulator of K_V7.1–KCNE1 currents, where KCNE2 may be able to exchange position with KCNE1 in the channel complex.²⁹ Such regulation of I_Ks density by KCNE2 could have a temporal and/or a spatial component, but this remains speculative based on the current evidence.

Clinical relevance

The clinical aspects of the information presented in this review are twofold. First, recent studies suggest that the subunit composition of I_Ks channels in cardiac myocytes likely is more complex and dynamic than previously envisioned. Not only KCNE1 but also KCNE2–KCNE4 may be involved in I_Ks generation. Because these other KCNE subunits can confer distinct phenotypes to K_V7.1 channels, changes in the subunit composition can have profound impact on I_Ks current amplitude and/or gating kinetics and thus action potential shape and duration. Concurrent atrial fibrillation and QT prolongation have been reported for individuals carrying mutations in K_V7.1. How K_V7.1 can form the molecular substrate for both conditions is perplexing, as QT prolongation is caused by a reduced repolarization reserve, whereas atrial fibrillation can be caused by increased K⁺ currents that shorten the repolarization phase and increase the risk of multiple reentrant wavelets. The complexity conveyed by the KCNE subunits may help explain some of the paradoxical phenotypes of K_V7.1 mutations. As a case in point, a carrier of the K_V7.1 mutation Q147R presented with concomitant atrial fibrillation and prolonged QT interval.² Functional studies revealed that the mutation causes a gain of function for channels encompassing KCNE2 and a loss of function for channels encompassing KCNE1. That functional consequences of the mutation are manifested only in presence of KCNE subunits underscores the importance of KCNE subunits for I_Ks generation. However, perhaps more importantly, it offers a possible molecular explanation for the patient's phenotype. Assuming a heterogenous distribution of KCNE1 and KCNE2 subunits in the heart, where KCNE1 plays a dominant role in ventricular I_Ks generation and KCNE2 is important for atrial I_Ks generation, the Q147R mutation may simultaneously lead to loss of function for I_Ks in the ventricles (QT prolongation) and gain of function for I_Ks in the atria (atrial fibrillation).

The second clinical implication is that the dynamic nature of I_Ks channel subunit composition can influence the channel's sensitivity to antiarrhythmic agents. It has long been recognized that KCNE1 association with K_V7.1 can dramatically increase its sensitivity to experimental I_Ks channel blocker (azimilide) or activator (mefenamic acid). On the other hand, KCNE2 association with K_V7.1 does not affect its sensitivity to either azimilide (unpublished observation) or mefenamic acid.²⁷

Conclusion

Reasoning that specific current densities may be regulated by dynamically altering the composition of accessory subunits in a particular ion channel complex rather than by regulating plasma membrane expression of the entire complex correlates with the finding that most protein complexes contain entities that are both periodically and constitutively expressed.³⁰ The task of delineating in detail how changes in I_Ks channel subunit composition influence its sensitivity to clinically used antiarrhythmic agents is important and, just like resolving the heterogeneity in cardiac KCNE protein expression profiles, requires future investigations.

ABBREVIATIONS

I_Kr

cardiac rapid delayed rectifier current

I_Ks

cardiac slow delayed rectifier current

K_V

voltage-gated potassium channel

NMR

nuclear magnetic resonance

PIP₂

phosphatidylinositol 4,5-bisphosphate