Nervous system KV7 disorders: breakdown of a subthreshold brake

Authors


  • This report was presented at a symposium on Kv7 (KCNQ) potassium channels that are mutated in human diseases, held at a joint meeting of The Slovak Physiological Society, The Physiological Society and The Federation of European Physiological Societies in Bratislava, Slovakia on 14 September 2007. It was commissioned by the Editorial Board and reflects the views of the authors.

  • S. Maljevic and T. V. Wuttke contributed equally and are listed in alphabetical order.

Corresponding author H. Lerche: Neurologische Klinik und Institut für Angewandte Physiologie, Universität Ulm, Zentrum Klinische Forschung, Helmholtzstr. 8/1, D-89081 Ulm, Germany. Email: holger.lerche@uni-ulm.de

Abstract

Voltage-gated K+ channels of the KV7 (KCNQ) family have been identified in the last 10–15 years by discovering the causative genes for three autosomal dominant diseases: cardiac arrhythmia (long QT syndrome) with or without congenital deafness (KCNQ1), a neonatal epilepsy (KCNQ2 and KCNQ3) and progressive deafness alone (KCNQ4). A fifth member of this gene family (KCNQ5) is not affected in a disease so far. Four genes (KCNQ2–5) are expressed in the nervous system. This review is focused on recent findings on the neuronal KV7 channelopathies, in particular on benign familial neonatal seizures (BFNS) and peripheral nerve hyperexcitability (PNH, neuromyotonia, myokymia) caused by KCNQ2 mutations. The phenotypic spectrum associated with KCNQ2 mutations is probably broader than initially thought, as patients with severe epilepsies and developmental delay, or with Rolando epilepsy have been described. With regard to the underlying molecular pathophysiology, it has been shown that mutations with very subtle changes restricted to subthreshold voltages can cause BFNS thereby proving in a human disease model that this is the relevant voltage range for these channels to modulate neuronal firing. The two mutations associated with PNH induce much more severe channel dysfunction with a dominant negative effect on wild type (WT) channels. Finally, KV7 channels present interesting targets for new therapeutic approaches to diseases caused by neuronal hyperexcitability, such as epilepsy, neuropathic pain, and migraine. The molecular mechanism of KV7 activation by retigabine, which is in phase III clinical testing to treat pharmacoresistant focal epilepsies, has been recently elucidated as a stabilization of the open conformation by binding to the pore region.

The KCNQ gene family encodes five voltage-gated delayed rectifier K+ channels (KV7.1–5), which are mainly expressed in heart muscle (KV7.1), the central nervous system (KV7.2–5) and the inner ear (KV7.1, KV7.4) (Jentsch, 2000). According to their differential expression patterns, mutations in four of the five genes lead to distinct inherited diseases. KCNQ1 mutations cause the Romano-Ward or the Jervell and Lange-Nielson syndrome which are characterized by long QT syndrome (Wang et al. 1996) without or with congenital deafness, respectively. KCNQ2 and KCNQ3 mutations cause benign familial neonatal seizures (Biervert et al. 1998; Singh et al. 1998; Charlier et al. 1998) and mutations in KCNQ4 underlie a slowly progressive deafness (Kubisch et al. 1999). Functional expression of the known mutations revealed a consistent reduction of the resulting K+ current for KV7.1–4 (Chouabe et al. 1997; Wollnik et al. 1997; Biervert et al. 1998; Schroeder et al. 1998; Lerche et al. 1999; Kubisch et al. 1999), which can depolarize the surface membrane of the respective cells rendering them hyperexcitable. In the following, we will focus on the clinical pictures, genetics and pathophysiology of KV7.2/7.3-associated channelopathies and the therapeutic potential of KV7 channel activation.

Clinical features and genetics of KV7.2/3 channelopathies

Benign familial neonatal seizures (BFNS) Seizures in BFNS occur in the first days of life, often in clusters, and usually disappear spontaneously after weeks to months, so that treatment is only required for a short period. Seizures have a partial onset often with hemi-tonic or -clonic symptoms or with apnoeic spells, or can clinically appear as generalized. Interictal electroencephalograms (EEG) are usually normal. The recorded ictal EEGs approve a focal onset but can also show bilateral synchrony. The risk of recurring seizures later in life is about 15%. Although the psychomotor development is described to be normal in most cases, an increasing number of patients with mental retardation and difficult-to-treat epilepsies has been described (Alfonso et al. 1997; Dedek et al. 2003; Borgatti et al. 2004; Schmitt et al. 2005; Steinlein et al. 2007). Furthermore, there is increasing evidence, that benign epilepsy of childhood with centrotemporal spikes (Rolando epilepsy) may be associated with KCNQ2/3 mutations (Coppola et al. 2003; Neubauer et al. 2007). BFNS is autosomal dominantly inherited with a penetrance of 85%. More than 30 mutations in KCNQ2 and 3 mutations in KCNQ3 have been described to cause BFNS (Fig. 1). Recent data revealed that deletions or duplications of KCNQ2 are found in a significant proportion of BFNS families, so that conventional sequencing methods should be complemented by copy number variation analysis (Heron et al. 2007).

Figure 1.

BFNS- and PNH-causing mutations in KV7.2 and KV7.3 subunits
Schematic view of KV7.2 and KV7.3 subunits depicting all mutations that have been described so far as white symbols (Borgatti et al. 2004; Richards et al. 2004; Lerche et al. 2005; Hunter et al. 2006; Soldovieri et al. 2007). E119G (Wuttke et al. 2008) and S122L (Hunter et al. 2006), two recently identified BFNS mutations, are marked in magenta; the voltage sensor mutations R207W/Q identified in patients with PNH (Dedek et al. 2001; Wuttke et al. 2007) are marked in blue.

Peripheral nerve hyperexcitability (PNH) PNH (myokymia, neuromyotonia) is clinically characterized by a spontaneous and continuous muscle overactivity, which has been described as myokymia (undulating movements of distal skeletal muscle), fasciculations, cramps, or other symptoms, which are due to a hyperexcitability of peripheral motor neurons (Hart et al. 2002). A common cause is antibodies directed against voltage-gated K+ channels (autoimmune-mediated PNH) (Hart et al. 2002). However, mutations in two different K+ channel genes, KCNA1 and KCNQ2, can also lead to this syndrome, either in combination with episodic ataxia type 1 (KCNA1; Browne et al. 1994) or with BFNS (KCNQ2; Dedek et al. 2001). Recently, we described a sporadic case with PNH associated with a KCNQ2 mutation (Wuttke et al. 2007).

Pathophysiology of KV7.2/3 channelopathies

BFNS and other epileptic syndromes KV7.2 and KV7.3 emerged as the first voltage-gated K+ channels associated with an idiopathic form of epilepsy (Biervert et al. 1998; Charlier et al. 1998; Singh et al. 1998). RNA information for both channels has been found widespread throughout the brain including cortex, cerebellum, basal ganglia, hippocampus (Biervert et al. 1998; Schroeder et al. 1998; Tinel et al. 1998; Cooper et al. 2000), and rat cervical superior ganglion (SCG) cells (Wang et al. 1998). First functional studies in Xenopus oocytes showed that both KV7.2 and KV7.3 could form homomeric K+ channels when expressed alone, yielding slowly activating, non-inactivating K+ currents, not much above background level for KV7.3. Injection of KV7.2 and KV7.3 mRNA in an equimolar ratio produced currents that were at least 10-fold larger than the currents obtained by injection of KV7.2 cRNA alone, suggesting that these channels form heteromers (Schroeder et al. 1998; Wang et al. 1998; Yang et al. 1998). Differential sensitivity to TEA, with KV7.2 being more sensitive than KV7.3, was used to confirm the formation of heteromers with an intermediate TEA sensitivity in heterologous systems and suggested a 1: 1 stoichiometry of KV7.2/KV7.3 channels in SCG (Wang et al. 1998; Hadley et al. 2003). Homomeric KV7.2–KV7.5, as well as their heteromeric associations with KV7.3, have been recognized as molecular correlates of the M current, a slow K+ current that can be suppressed upon activation of muscarinic acetylcholine (mACh) receptors (Brown & Adams, 1980; Wang et al. 1998; Delmas & Brown, 2005). The M current is activated and controls the membrane potential in the subthreshold range of an action potential serving as a brake for neuronal firing.

Epilepsy-causing mutations have been found in the long cytoplasmatic C-terminus of KV7.2, in the pore regions (S5–S6 segments) of both KV7.2 and KV7.3 channels, in the voltage sensor S4 and in the S1–S2 region of KV7.2 (Fig. 1, Table 1). The common functional consequence of all mutations examined so far is a reduction of the resulting K+ current. It can be a complete loss of function of KV7.2 or KV7.3; however, a co-expression of wild type (WT) and mutant KV7.2 or KV7.3 with the WT of the other subunit in Xenopus oocytes in a 1: 1: 2 ratio, mimicking the situation in a patient, revealed a reduction in current size of merely 20–25% compared with co-expression of both WTs. Thus, relatively small changes of the M current appear to be sufficient to cause epileptic seizures in the neonatal period (Schroeder et al. 1998; Lerche et al. 1999, 2005; Jentsch, 2000; Singh et al. 2003; Bassi et al. 2005).

Table 1.  Examples of KV7.2/7.3 channelopathies
GeneProteinAffected channel regionMutationMolecular defectClinical phenotype
  1. Overview of the KCNQ2/3 mutations discussed in this review, summarizing their positions within the channel protein (see also Fig. 1), major molecular defects and associated disorders (BFNS, benign familial neonatal seizures; PNH, peripheral nerve hyperexcitability). For details and references see text.

KCNQ2 KV7.2S1–S2 regionE119GSlight depolarizing shift of the voltage dependence of activation and slowing of the activation kinetics; effects restricted to the subthreshold rangeBFNS
S122L 
S4 voltage sensorNon-charged residues:Depolarizing shift of the activation curve accompanied by a slowing of activation kinetics upon stronger depolarizationsBFNS
A196V 
L197P 
Charged residues 
R207WDramatic depolarizing shift of activation and slowing of activation kinetics; dominant-negative effect on WT currentsPNH, BFNS
R207Q 
R214WSlowed activation and faster deactivation; decreased voltage sensitivityBFNS
Pore regionMissense mutationReduced currents, may be due to reduced single channel conductance; no dominant-negative effectBFNS
Y284C 
C-terminusFrameshift and truncation 534ins(5 bp)Reduced currents and surface expression presumably due to impaired assemblyBFNS
Frameshift with prolonged nonsense sequenceReduced currents due to impaired protein stability 681del(1 bp)BFNS, centrotemporal spikes
Missense mutationsImpaired calmodulin binding possibly affecting channel traffickingBFNS
R353G 
L619R 
KCNQ3 KV7.3Pore regionMissense mutationsReduced currents, may be due to reduced single channel conductance; no dominant-negative effectBFNS
D305G 
W309R 
G310V 

After Schmitt et al. (2000) had shown that a mutation in the C-terminal part of KV7.1 causing the Jervell and Lange-Nielson syndrome disrupts assembly of KV7.1 channels, experiments using chimeras between KV7.1, KV7.2 and KV7.3 channels demonstrated that the assembly of KV7.2 and KV7.3 channels is also mediated by this region (Maljevic et al. 2003; Schwake et al. 2003). We may therefore presume that improper tetramerization affecting channel insertion in the surface membrane could explain how C-terminal mutations reduce KV7.2 currents. Indeed, in contrast to pore mutations in KV7.2 and KV7.3, reduced surface expression was observed for a KV7.2 mutant truncating the C-terminus (Schwake et al. 2000). Further pathomechanisms of C-terminal mutations could be reduced protein stability (Soldovieri et al. 2006) or impaired interaction with calmodulin (Richards et al. 2004), shown to regulate trafficking of KV7.2 subunits (Etxeberria et al. 2007). Crystallization of the KV7.4 assembly domain (Howard et al. 2007) and the biochemical analysis of the KV7.2 subunit interaction domain (Wehling et al. 2007) will expand our understanding of the complex role of the C-terminus for the function of KV7 channels and the pathophysiology of BFNS.

Despite presence of an intact C-terminal assembly domain, a clear dominant-negative effect on WT currents has not been reported for any of the missense mutations affecting the pore region of KV7.2 (Singh et al. 1998) or KV7.3 (Charlier et al. 1998; Schroeder et al. 1998; Hirose et al. 2000; Singh et al. 2003), which probably reduce K+ currents by affecting ion channel conductance and a haploinsufficiency mechanism. In fact, the only two KV7.2 mutations exibiting a pronounced dominant-negative effect are the S4 segment mutations associated with PNH (Dedek et al. 2001; Wuttke et al. 2007; see below). The effects of other epilepsy-associated variations within the voltage-sensing S4 segment of KV7.2 (Miraglia Del Giudice et al. 2000; Singh et al. 2003; Soldovieri et al. 2007) were described as slowed activation with faster deactivation and decreased voltage sensitivity for R214W (Castaldo et al. 2002), or atypical gating (rightward shift of the activation curve accompanied with a slowing of activation kinetics upon stronger depolarizing prepulses) caused by mutations of non-charged residues (Soldovieri et al. 2007).

Recently, the S1–S2 region emerged as an unexpected spot for BFNS mutations in KV7.2. For two mutations, S122L (Hunter et al. 2006) and E119G (Wuttke et al. 2008), we detected a significant reduction of the relative current amplitudes exclusively at subthreshold voltages (Fig. 2A and B). In reconstitution experiments supporting the presumed in vivo constellation (KV7.2-WT: KV7.2-mut: KV7.3-WT in a 1: 1: 2 ratio), the observed changes were rather subtle, even smaller than the 25% observed for other mutations (see above and Fig. 2B). However, even these minor changes were sufficient to produce prolonged bursts of action potentials and induced a lower threshold for infinitive firing in a one-compartment neuronal model cell (Fig. 2D). A structural model, calculated by using the coordinates of the crystallized KV1.2 channel (Long et al. 2005), proposed an interaction of residues E119 and S122 with a positive charge, R201, of the voltage sensor S4 (Fig. 2C). Taken together, these data show in a human disease model, that (i) very subtle changes in channel gating can be sufficient to cause BFNS, (ii) the subthreshold voltage range is most relevant for M channels to modulate neuronal firing and (iii) the S1–S2 region is critical for voltage sensing and may directly interact with S4.

Figure 2.

Functional analysis of homomeric and heteromeric KV7.2 WT and mutant channels harbouring the E119G or the S122L mutation
A, representative normalized raw current traces for KV7.2 WT, E119G and S122L mutant channels. Currents were elicited from a holding potential of −80 mV by depolarizations ranging from −80 to +10 mV in 10 mV steps, followed by a pulse to −30 mV to obtain tail currents. Current traces at subthreshold potentials of an action potential between −50 and −30 mV (marked by arrows) reveal current reductions for mutant compared to WT channels. B, conductance–voltage curves were constructed by plotting the normalized tail current amplitude recorded at −30 mV against the membrane potential. Lines represent standard Boltzmann functions fit to the data points. Comparison of WT KV7.2 and both mutated channels revealed a subtle but significant shift in the voltage dependence of activation towards depolarized potentials predominantly occurring at subthreshold potentials, predicting a current reduction of 44% for E119G and 75% for S122L at −50 mV. Upon co-expression of mutant and WT KV7.2 channels in a 1: 1 ratio or in a 1: 1: 2 ratio with KV7.3, the effects of the mutation were less pronounced, but still remained statistically significant for co-expressions with E119G mutant channels. These electrophysiological alterations suggest the involvement of the S1–S2 region of the KV7.2 channel in voltage-dependent gating. C, structural model exploring the role of the S1–S2 region in voltage-dependent gating of the KV7.2 channel. The three-dimensional structural model of a homomeric KV7.2 channel was generated based on sequence homology for segments S4–S6 and assuming a structural similarity of segments S1–S3 of KV7.2 with the published coordinates of KV1.2 (Long et al. 2005). The upper panel gives an overview of all 4 subunits with the transmembrane segments S1–S4 of one subunit (S1 (yellow), S2 (orange), S3 (red), S4 (pink) and the pore domain of the adjacent subunit S5 (green), selectivity filter/pore helix (light blue), S6 (blue)) being coloured for more clarity. The lower panel depicts a magnification of the region of interest in top view. The putative positions of the residues E119 and S122 in the outer S1 segment and their possible interaction partner R201 in S4 are shown in stick representation. Putative electrostatic interactions or formation of hydrogen bonds between E119 and S122 with R201 are depicted as green dashed lines. D, to investigate the impact of the E119G mutation on firing properties of neurons, a one-compartment computer cell model has been implemented according to Golomb et al. (2006). Model cells are continuously stimulated by the current Istim from time t= 0 ms. The M current, IM, is represented by KV7.2/KV7.3 (upper panel) or E119G/KV7.2/KV7.3 (middle panel) channels. The duration of the evoked action potential burst is relatively prolonged in cells in which IM is mediated by E119G/KV7.2/KV7.3 channels. Istim= 1 μA cm−2. Burst duration as a function of Istim (lower panel). The increase in burst duration with increasing Istim is steeper in cells that are simulated using the mutant channel. There is a threshold above which the M current is insufficient to terminate the burst. This threshold is lower in simulations with parameters for mutant channels (continuous line KV7.2/KV7.3, dotted line E119G/KV7.2/KV7.3). Figure modified from Hunter et al. (2006) and Wuttke et al. (2008).

PNH Unlike disease-causing mutations in KV7.1 or KV7.4, which often cause a dominant-negative effect on WT channels, the major pathophysiological mechanism underlying BFNS is a haploinsufficiency of KCNQ2. The only two mutations in KV7.2 with a reported prominent dominant-negative effect on WT currents affect the same arginine at position 207, within the S4 segment, the voltage sensor of KV7.2 channels (Fig. 1). Interestingly, both R207W (Dedek et al. 2001) and R207Q (Wuttke et al. 2007) are associated with a hyperexcitability of peripheral motor neurons (PNH) which has not been described for any other KCNQ2 or KCNQ3 mutation.

Functional expression of both PNH-associated mutations revealed large depolarizing shifts of the conductance–voltage relationships combined with prominent slowing of the time course of activation compared to WT channels (Fig. 3). These effects were more pronounced for R207W, which apart from peripheral hyperexcitability (myokymia) also causes neonatal seizures in all but one affected individual (Dedek et al. 2001). When the mutants were co-expressed with WT channels, and the currents analysed 200 ms after the onset of depolarizing voltage steps, a dominant-negative effect reducing the relative current amplitudes by > 70% was observed (Fig. 3C). This dramatic effect on WT currents is what distinguishes PNH-causing mutations in KV7.2 from the other mutations causing BFNS alone. Therefore, peripheral nerves might be more resistant to get hyperexcitable upon a reduction of the M current when compared to central neurons in the neonatal period.

Figure 3.

Functional characterization of WT (KV7.2) and mutant (R207Q, R207W) channels
A, representative raw current traces for homomeric KV7.2, R207Q and R207W channels. Currents were elicited from a holding potential of −80 mV by depolarizations ranging from −80 to +60 mV in 10 mV steps, and tail currents were recorded at −30 mV. B, time constants of activation (τact) for KV7.2, R207Q, R207W and the co-expressions R207Q/KV7.2 and R207W/KV7.2 were obtained by fitting a first order exponential function to the rising part of each current trace. Means for τact±s.e.m. were plotted against voltage, revealing a pronounced slowing of activation kinetics for both mutants. C, to evaluate the current reduction after short-term depolarizations, mutant current amplitudes were determined after 200 ms and normalized to the current amplitude at +60 mV after a 5 s depolarization and compared to the respective WT amplitudes normalized to the maximum amplitude reached after 2 s at +60 mV. Relative current amplitudes of the co-expressions R207Q/KV7.2 and R207W/KV7.2 were reduced in comparison to KV7.2 for a broad range of potentials. Comparison to the assumed 50% reduction of current in case of a haploinsufficiency of the KV7.2 WT (dashed line) predicted a strong dominant-negative effect for both R207Q and R207W when co-expressed with KV7.2 with a > 70% reduction of the relative current amplitude for potentials between −40 and +10 mV. Parameters are given as means ±s.e.m. Figure modified from Wuttke et al. (2007).

Expression pattern of KV7.2/KV7.3 and possible pathophy-siolological significance Expression of KV7.2 and KV7.3 channels in the brain overlap extensively, although (i) in situ hybridization indicates that they are not always expressed in the same ratio (Schroeder et al. 1998; Maljevic et al. 2007), (ii) immunohistochemistry reveals that some neurons stain only for one or the other subunit (Cooper et al. 2000) and (iii) that the expression pattern changes with development (Geiger et al. 2006; Weber et al. 2006; Maljevic et al. 2007). Recent studies have revealed that the main and functionally probably most important localization of KV7.2 and KV7.3 channels is at axon initial segments (AIS) (Devaux et al. 2004; Pan et al. 2006; Maljevic et al. 2007; Fig. 4).

Figure 4.

Localization of KV7.3 at the axon initial segment (AIS)
As an example to show M channel localization at the AIS, primary hippocampal neurons obtained from E18 mouse embryos were stained with antibodies directed against MAP-2, a somato-dendritic marker, KV7.3, and 4,6-diamidino-2-phenylindole (DAPI) (nucleus). Prominent staining of KV7.3 not overlapping with the somato-dendritic signal of MAP-2 reveals localization of this channel at the AIS (Maljevi S, Blazevic D & Lerche H, unpublished data).

Since the reduced K+ current mediated by heteromeric KV7.2/KV7.3 channels results in seizures preferentially during the neonatal period, the developmental changes in expression patterns could provide an explanation for the transient phenotype of BFNS. The available data suggest an up-regulation of both channels during development, both in unmyelinated and myelinated axons, as the major change within the first three postnatal weeks in mouse brains (Geiger et al. 2006; Weber et al. 2006; Maljevic et al. 2007). Therefore, the M channel-mediated K+ current might still be on a low level and a small reduction by mutations sufficient to cause seizures in neonates, since a critical amount of KV7 channels is probably needed for an adequate control of the subthreshold membrane potential. By contrast, this happens only rarely in adulthood, when M channels are abundantly available. Up-regulation of other K+ channels might also help to compensate for the deficit. In addition, the exclusive expression of a shorter, non-functional splice variant of KV7.2 in fetal brain, which can attenuate KV7.2/KV7.3-mediated currents (Smith et al. 2001), might play a role. Finally Okada et al. (2003) hypothesized a link to the excitatory action of γ-aminobutyric acid (GABA) in immature brain. In the first few weeks of life, when GABA is depolarizing due to a high intracellular Cl concentration, the M current might have an even more important role as an inhibitor of neuronal firing, which would remit parallel to the inhibitory switch of the GABAergic system.

Retigabine, a KV7 channel opener and novel anticonvulsant compound

Retigabine (RTG), initially derived from flupirtine, a substance used for therapy of acute and chronic pain, enhances the activity of KV7.2–5 channels (Rundfeldt & Netzer, 2000; Tatulian et al. 2001; Dupuis et al. 2002). Since none of the clinically used anticonvulsants exhibits a similar mechanism of action, RTG represents a new class of anticonvulsant compounds the effectiveness of which has been demonstrated in many seizure models (summarized in Wuttke & Lerche, 2006). It is currently undergoing phase III clinical testing for pharmacoresistent focal epilepsies. Since augmentation of the M current mediated by RTG-activated KV7.2 and KV7.3 channels will lead to a stabilization of the resting and subthreshold membrane potential towards the K+ equilibrium potential, thus generally reducing membrane excitability, M channels are an attractive pharmacological target to treat any disease going along with neuronal hyperexcitability, such as epilepsy, PNH, neuropathic pain, migraine and stroke. It is important to note that RTG does not enhance the activity of the cardiac KV7.1 channel (Fig. 5A) rendering cardiac side-effects unlikely.

Figure 5.

Putative binding site and molecular mechanism of the KV7 channel opener retigabine (RTG)
A, conductance–voltage curves were constructed by plotting the normalized tail current amplitude against the membrane potential. The lines are fits to a Boltzmann equation showing a left shift of the KV7.2 activation curve after application of 100 μm RTG, while the KV7.1 activation curve remains unaffected. B, knock-out of the RTG effect in KV7.2. Conductance–voltage plots for Q2[W236L]Q1 (W236 in KV7.2 has been replaced by the corresponding leucine of KV7.1) and Q2[S6]Q1 (transmembrane segment S6 of KV7.2 has been replaced with the S6 segment of KV7.1) before and after application of 100 μm RTG reveal a loss of the RTG-induced left shift of the activation curves. C, knock-in of RTG effects in KV7.1. Conductance–voltage plots of Q2[S6]Q1[A336G] (transmembrane segment S6 of KV7.2 has been replaced by the S6 segment of KV7.1, additionally A336 of KV7.1 has been replaced with the corresponding G301 of KV7.2) reveal an induction of a RTG-mediated left shift of the activation curve. In contrast, no shift was observed for Q2[S6]Q1 (compare Fig. 5B). D, putative model for the binding of the RTG molecule within the pore region of the KV7.2 channel. The panel depicts energy-optimized homology models of the KV7.2 pore domain based on the crystal structures of KscA in the closed (left) and MthK in the open conformation (right). The glycine at position 301, serving as a putative gating hinge, is shown in green, the tryptophan at position 236 in orange. The RTG molecule was manually docked in the open channel. The model proposes that RTG binds to a hydrophobic pocket formed upon opening of the KV7.2 channel at the intracellular part of the channel pore between S5 and S6, involving W236 and the flexibility of S6 which is dependent on G301. Figure modified from Wuttke et al. (2005).

Using recombinant channels it has been demonstrated that the main mechanism by which RTG enhances KV7.2 and KV7.3 channel activity is a hyperpolarizing shift of the activation curve (Main et al. 2000; Rundfeldt & Netzer, 2000; Wickenden et al. 2000; Fig. 5). The molecular correlates of this M channel activation and the probable binding site of RTG in KV7.2 and KV7.3 channels have been recently identified (Schenzer et al. 2005; Wuttke et al. 2005). We constructed a set of chimeric channels by interchanging the whole or parts of the pore region, or single amino acids between the RTG-sensitive KV7.2 and the RTG-insensitive KV7.1 channel. Functional expression in Xenopus oocytes revealed that the effect of RTG on KV7.2 channels could be completely abolished by either a replacement of a tryptophane in the cytoplasmic end of S5 (W236) in KV7.2 with the corresponding leucine in KV7.1 or by a replacement of the entire S6 transmembrane segment of KV7.2 for the corresponding one of KV7.1 (Fig. 5B). The importance of the tryptophane in S5 for RTG sensitivity could be shown similarly in KV7.2–5 channels by Schenzer et al. (2005). Further experiments revealed that a glycine in S6, often referred to as the so-called ‘gating hinge’ critically involved in opening of cation channels (Jiang et al. 2002), is also critical for the RTG effect (Fig. 5C). These data suggested a lipophilic interaction between the fluorophenyl ring of RTG and the aromatic W236 and pointed towards the importance of the flexibility of the S6 helix at position of G301, as illustrated in a structural computer model based on the crystal structures of KscA and MthK (Fig. 5D). By this mechanism RTG could stabilize the channel in its open conformation thereby explaining the RTG-induced hyperpolarizing shift of the activation curve.

In addition to its effects on KV7 channels, RTG mediates a facilitation of GABAergic inhibition (Otto et al. 2002) and exerts an unspecific inhibitory effect on sodium-, calcium- and kainate-induced currents when applied in higher concentrations (Rundfeldt & Netzer, 2000).

Appendix

Acknowledgements

This work was supported by grants from the European Union (Epicure: LSH 037315), the state Baden-Wuerttemberg (Landesforschungsschwerpunkt 1423/74), Bundesministerium fuer Bildung und Forschung (BMBF/NGFN2, 01GS0478), the Deutsche Forschungsgemeinschaft (DFG, Le1030/9-1) and the Thyssen-Stiftung. H.L. is a Heisenberg fellow of the DFG.

Ancillary