Kv7 (KCNQ) potassium channels that are mutated in human diseases

Kv7 channels are a small family of voltage-gated potassium channel subunits (Kv7.1–Kv7.5) that are encoded by the KCNQ genes (KCNQ1–5) (reviewed by Jentsch, 2000; and by Robbins, 2001). Structurally, they resemble other Kv channels, with six transmembrane domains, a voltage sensor in the fourth (S4) domain, and a pore loop between the S5 and S6 domains, but with an unusually long and complex C-terminal domain. Each subunit can form a homo-tetramer, but can also coassemble with other subunits (either Kv or non-Kv). Kv7.1 can coassemble with the single transmembrane domain β-subunits KCNE1–3, with consequent slowing of kinetics (KCNE1) or loss of voltage dependence (KCNE2,3). Kv7.2, 7.4 and 7.5 can coassemble with Kv7.3, but not with each other, and Kv7.1 cannot coassemble with any Kv7.2–7.5 subunit.

Collectively, they form the primary molecular basis of a number of long-established, diverse and apparently unrelated K⁺ currents (Table 1). These include the cardiac slow delayed rectifier current I_Ks/I_K2 (originally described by Noble & Tsien, 1968), the slow subthreshold neuronal M-current, I_K(M) (Brown & Adams, 1980), the slow delayed rectifier at nodes of Ranvier, I_Ks (Dubois, 1981), and a low threshold K⁺ current in cochlear outer hair cells, I_Kn (Housley & Ashmore, 1992). Kv7.1, with KCNE2 or 3 subunits, also contributes to K⁺ fluxes in some epithelial cells, and Kv7.5 and/or Kv7.4 appear to contribute to K⁺ currents in some vascular smooth muscle cells (see Table 1). Thus, Kv7 channels may be of interest to physiologists of many persuasions. The channels are also interesting and unusual in several other respects, many of which are discussed in the accompanying articles.

Table 1.

Kv7/KCNQ channels

Gene	Protein	Functional channel	Genetic disease
KCNQ1	Kv7.1	Cardiac delayed rectifier (+ KCNE1)1,2	Cardiac LQT1 (RWS)1,2
(KvLQT1, KQT1)		K+ transport in inner ear stria vascularis3,4, K+ transport in epithelia, e.g. colonic crypts (+ KCNE3)5, gastric parietal cells (+ KCNE2)6,7,8	LQT + deafness (JLNS)3
KCNQ2	Kv7.2	Neuronal M-current9 (+ Kv7.3)	BFNC10,11
(KQT2)		Nodal IKs12,13 (± Kv7.3)	BFNC + PNH (myokymia)14
KCNQ3	Kv7.3	Neuronal M-current (+ Kv7.2)9	BFNC15
KCNQ4	Kv7.4	Cochlear OHC K+ current16,17	DFNA2 (deafness)
		Vestibular utricle K+ current18
KCNQ5	Kv7.5	Neuronal M-current19,20 (± Kv7.3)
		Vascular smooth muscle K+ current 21,22	—

Abbreviations: LQT = Long-QT syndrome; RWS = Romano-Ward Syndrome; JLNS = Jervell and Lange-Nielsen Syndrome; BFNC = Benign Familial Neonatal Convulsions; OHC = outer hair cell; DFNA2 = autosomal dominant hearing loss (see Hereditory Hearing Loss Homepage http://webh01.ua.ac.be/hhh/).

References (selected): ¹Barhanin et al. 1996; ²Sanguinetti et al. 1996; ³Neyroud et al. 1997; ⁴Jentsch; ⁵Schroeder et al. 2000b; ⁶Dedek & Waldegger 2001; ⁷Grahammer et al. 2001; ⁸Heitzmann et al. 2004; ⁹Wang et al. 1998; ¹⁰Singh et al. 1998; ¹¹Bievert et al. 1998; ¹²Devaux et al. 2004; ¹³Schwarz et al. 2006; ¹⁴Dedek et al. 2001; ¹⁵Charlier et al. 1998; ¹⁶Kubisch et al. 1999; ¹⁷Kharkovets, 2006; ¹⁸Holt et al. 2007; ¹⁹Lerche et al. 2001; ²⁰Schroeder et al. 2000a; ²¹Brueggemann et al. 2007; ²²Yeung et al. 2007.

Genetic mutations

As implied by the title of the symposium, an exceptional proportion (4 out of 5) of the KCNQ genes are subject to mutations that give rise to human genetic diseases (Jentsch, 2000; see Table 1). Indeed, it was through the discovery of these mutations that this gene family, and the molecular composition of the Kv7 channels, was identified, and analysis of these mutations has given important clues regarding both the physiological and the biophysical function of these channels.

In their article, Peroz et al. (2008) discuss the functional consequences of three groups of KCNQ1 mutations. One group of mutations, in the N-terminus, affects trafficking; a second group, in the C-terminus, affects assembly of the tetramer (see further below); and a third group, also in the C-terminus, alters the responsiveness of the channels to membrane phosphatidylinositol-4,5-bisphosphate (PIP₂), which, for Kv7.1 channels, regulates their voltage sensitivity (again see further below).

Mutations of KCNQ2 (and, more rarely, KCNQ3) cause a form of infant epilepsy termed ‘benign familial neonatal convulsions’ (BFNC). Maljevic et al. (2008) describes this syndrome and some of the many mutations that induce it, including two mutations in the S1–S2 domains of Kv7.2 that produce only very small changes in channel gating (primarily a shift in threshold) but which nonetheless can clearly modify neuronal excitability – thus serving to indicate how crucial and how subtle the control of neuronal excitability by these channels can be. This results from the high Kv7 channel density in the axon initial segment (as shown by others and illustrated by Maljevic et al.), where they are colocalized with the Na⁺ channels and tightly regulate the action potential threshold at the site of its initiation (Shah et al. 2007). Maljevic et al. also describe two KCNQ2 mutations in the S4 domain that produce large shifts in Kv7.2/7.3 voltage sensitivity; unlike most BFNC mutants, these exert a dominant-negative effect leading to large current deficits and peripheral nerve hyper-excitability, in addition to BFNC. Finally, they address the issue of why BFNC only lasts a few weeks and does not (usually) extend into adulthood; they suggest (from their own work on mice) that it may partly relate to the increasing expression of Kv7.2 and Kv7.3 channels during postnatal development.

The C-terminal domain

For many functions (other than channel gating and permeation), the long C-terminus forms the ‘business centre’ of the Kv7 molecule. Haitin & Attali (2008) provide a comprehensive survey of many of these C-terminal functions. These include the role of calmodulin binding in membrane trafficking and Ca²⁺ sensing; the binding of PIP₂ and its facilitatory effect on channel opening; the way in which subunit tetramerization is accomplished through interaction of helix D in the ‘A-domain’ to form a coiled-coil complex; the facilitation of channel phosphorylation by binding of AKAP family proteins (yotiao and AKAP150); the regulation of channel expression by interaction with Nedd4–2 ubiquitin ligase; and the way in which the binding of ankyrin-G to the C-terminal of Kv7.2 and 7.3 promotes the localization of these subunits to axon initial segments and nodes of Ranvier. They also present new information showing that the C-terminal domain of Kv7.1 binds the C-terminus of the KCNE1 auxillary subunit, and suggest that the tetramerization coiled-coil structure may form a stable complex with this subunit.

The C-terminus is also the primary site through which neurotransmitters modulate Kv7 channel activity (Delmas & Brown, 2005). Thus, one of the key features of the original M-current (and the initial reason for its designation) is that it is inhibited by acetylcholine, through stimulation of muscarinic receptors (Brown & Adams, 1980). However, this response is not confined to muscarinic receptors, but extends to many other G-protein-coupled receptors that activate Gq. Hernandez et al. (2008) present a comprehensive analysis of the mechanisms responsible for this effect. Although all such receptors induce hydrolysis of PIP₂, from their own elegant experiments on sympathetic neurons they are able to draw a clear distinction between those receptors (muscarinic M1 and angiotensin ATII) that induce channel closure through depletion of membrane PIP₂ (a requirement for normal Kv7 channel function) and others (bradykinin B₂ and purinergic) through which closure results not from PIP₂ depletion but from the release of intracellular Ca²⁺ by inositol-1,4,5-trisphosphate (IP₃) and subsequent activation of channel-attached calmodulin. The reasons for this difference are that the latter receptors are more efficient at generating IP₃ and Ca²⁺ release, and (probably consequentially) accelerate PIP₂ synthesis sufficiently to prevent depletion. However, the precise mechanism for this effect on synthesis is not yet clear and requires further work.

Although provisionally located at the proximal end of the C-terminus (Zhang et al. 2003), the PIP₂-binding sites in the Kv7 channel have not yet been fully defined. By making chimaeras from Kv7 subunits with high (Kv7.3) and low (Kv7.4) ‘affinities’ for PIP₂, Hernandez et al. (2008) locate an alternative region (residues 428–484 of Kv7.2) further downstream from the immediate proximal region. This contains a number of basic residues which they show to line up well against a structural model of the PIP₂ binding site of the inward rectifier Kir2.1 channel. The next step is clearly to make point-mutations.

Pharmacology and function of neural Kv7/M channels

Initially the M-current was characterized functionally as a ‘braking current’, that served to limit neuronal discharge frequency. Though confirmed by subsequent work, immunocytochemical localization by Kv7 antibodies, and the development and use of Kv7-specific drugs, have highlighted important additional or consequential functions.

Drugs

One group of drugs (illustrated by linopirdine and XE991) block Kv7 channels and hence inhibit currents. Another group (represented by retigabine) produces a hyperpolarizing shift in the current–voltage (or P_open–voltage) curve and hence enhances currents. This latter effect is restricted to neural (Kv7.2–7.5) channels and does not extend to Kv7.1 channels. In their article, Maljevic et al. (2008) describe how their group used this property to identify the binding site for retigabine at a hydrophobic pocket between the intracellular parts of the S5–S6 domains. This pocket is revealed on channel opening, so that retigabine effectively stabilizes the open state, hence explaining the voltage shift.

Transmitter release

Immunocytochemical work shows that Kv7 channels are not restricted to the neuron soma but extend along axons and (possibly) to nerve terminals. If so, these channels might regulate transmitter release independently of their role in governing neuronal discharge frequency – and indeed, there is some antecedent evidence for this. Hernandez et al. (2008) address this question further with respect to sympathetic nerve endings, by amperometric measurement of noradrenaline release from neurites of cultured sympathetic neuron clusters following trains of electrical shocks applied to the connecting neurite bundle. They report that the Kv7 blocker XE991 enhances release whereas the M-current enhancer retigabine reduces release. This accords with previous observations concerning the effect of bradykinin on noradrenaline release, though, as the authors point out, the effects of other transmitters may be complicated by concurrent and opposing actions on the terminal Ca²⁺ currents.

Functions in the basal ganglia

Kv7.2 and (interestingly) Kv7.4 subunits are strongly expressed in the basal ganglia, including the substantia nigra, thus suggesting a role for these channels in regulating dopaminergic function. Hansen et al. (2008) discuss this possibility. They remind us that their (and others’) previous work has shown that retigabine reduces the release of dopamine from dopaminergic neurons, reduces the spntaneous discharge activity of these neurons and can also inhibit locomotor activity. Additionally, they confirm the presence of Kv7.4 transcripts in dopaminergic neurons of the substantia nigra and ventral tegmental area. They go on to show the localization of Kv7.4 protein in a subset of serotonergic (5-htdroxytryptamine-containing) neurons in the dorsal raphe nucleus, and an inhibition of discharge activity in some (but not all) serotonergic neurons in this region by retigabine. As they acknowledge, establishing a normal function for Kv7 channels in regulatiung serotonergic activity will require additional experiments using XE991. Nevertheless, these are provocative findings in that, although retigabine was initially developed simply as an anticonvulsant, they imply a potentially wider use for retigabine (or, better perhaps, a Kv7.4-specific enhancer) in such affective disorders as schizophrenia and anxiety.

Thus, work on this interesting small family of K⁺ channels continues to provide new insights into physiology, pharmacology, pathology and potential therapeutics, as these articles illustrate.