MaxiK channel partners: physiological impact
- Rong Lu1,
- Abderrahmane Alioua1,
- Yogesh Kumar1,
- Mansoureh Eghbali1,
- Enrico Stefani134 and
- Ligia Toro124
- 1Department of Anaesthesiology, Division of Molecular Medicine2Department of Molecular and Medical Pharmacology3Department of Physiology4Cardiovascular Research Laboratory, University of California Los Angeles, Los Angeles, CA 90095-7115, USA
- Corresponding author L. Toro: Dept. Anesthesiology, UCLA, BH-509A CHS, Box 957115, Los Angeles, CA 90095-7115, USA. Email: ltoro{at}ucla.edu
Abstract
The basic functional unit of the large-conductance, voltage- and Ca2+-activated K+ (MaxiK, BK, BKCa) channel is a tetramer of the pore-forming α-subunit (MaxiKα) encoded by a single gene, Slo, holding multiple alternative exons. Depending on the tissue, MaxiKα can associate with modulatory β-subunits (β1–β4) increasing its functional diversity. As MaxiK senses and regulates membrane voltage and intracellular Ca2+, it links cell excitability with cell signalling and metabolism. Thus, MaxiK is a key regulator of vital body functions, like blood flow, uresis, immunity and neurotransmission. Epilepsy with paroxysmal dyskinesia syndrome has been recognized as a MaxiKα-related disorder caused by a gain-of-function C-terminus mutation. This channel region is also emerging as a key recognition module containing sequences for MaxiKα interaction with its surrounding signalling partners, and its targeting to cell-specific microdomains. The growing list of interacting proteins highlights the possibility that associations with the C-terminus of MaxiKα are dynamic and depending on each cellular environment. We speculate that the molecular multiplicity of the C-terminus (and intracellular loops) dictated by alternative exons may modulate or create additional interacting sites in a tissue-specific manner. A challenge is the dissection of MaxiK macromolecular signalling complexes in different tissues and their temporal association/dissociation according to the stimulus.
General properties of MaxiK channels
The pore-forming α-subunit of the large-conductance, Ca2+- and voltage-activated potassium channels (MaxiK, BK, BKCa) is encoded by a single gene (Slo, KCNMA1) with 27 constitutive exons and multiple alternative exons spanning about 750 kb of the human genome. The MaxiK gene has been highly selected through evolution and in mammals its constitutive exons predict a protein with ∼98% amino acid sequence identity. Each of the constitutive exons seem to be designed for a specific function as reflected by well-characterized regions of the channel, like the conduction pore, voltage sensor, ‘Ca2+ bowl’ and the first transmembrane segment, S0, that is necessary for β1-subunit interaction. Four α-subunits (MaxiKα) are necessary for MaxiK basic conduction function but alternative splicing and association with β-subunits (β1–β4) contribute to MaxiK molecular diversity (Fig. 1). Splice variation can regulate MaxiK targeting to different subcellular compartments (Zarei et al. 2001, 2004) suggesting the possibility that splice variant sequences recognize differential partners for traffic. The functional significance of vectorial targeting was revealed in myometrium in late pregnancy where overall transcript and protein levels are increased but protein is retained in intracellular organelles resulting in reduced functional expression of the MaxiK channel at the plasmalemma (Benkusky et al. 2000; Eghbali et al. 2003). Consistent with hormonal variations during pregnancy, MaxiK transcript and splice variant expression are regulated by hormones. Inclusion of STREX (STRess axis-regulated EXon) alternative exon can be inhibited by glucocorticoids and oestrogen, and stimulated by adrenocorticotropic hormone, testosterone and progesterone (Xie & McCobb, 1998; Lai & McCobb, 2002; Zhu et al. 2005). STREX adds to MaxiKα a protein kinase A (PKA) phosphorylation site that in heterologous systems switches the homomeric channel from being activated to being inhibited by PKA (Tian et al. 2001). In heteromeric channels, the equation is not simple: only one of the four subunits requires STREX inclusion for the channel to be inhibited by PKA (when all four constitutive –RQPS*– PKA phosphorylation sites are dephosphorylated), while all four α-subunits (with or without STREX) need to be constitutively PKA phosphorylated for channel activation (Tian et al. 2004). Nevertheless, the switch in channel phenotype resembles the physiological behaviour of MaxiK in the uterus where non-gravid myometrium expresses a major MaxiK population that is inhibited by PKA while during pregnancy most of the channels are positively PKA regulated (Pérez & Toro, 1994; Zhou et al. 2000). Thus, STREX can be considered a signalling–transducing exon as it may switch β-adrenergic stimulation linked to PKA activity from evoking MaxiKα activation to its inhibition in the uterus.
Physiology and pathophysiology of MaxiK channel α and β1 complex
The phenotype of MaxiK channels varies widely among tissues, cells and under different hormonal environments probably due to alternative splicing, associations with distinct regulatory subunits (e.g. β-subunits), and phosphorylation status. MaxiK channels sense and regulate membrane voltage and intracellular Ca2+; thus, it is not surprising that they play important roles in a variety of body functions such as neurotransmission, blood flow, uresis and immunity (Brayden & Nelson, 1992; Robitaille et al. 1993; Ahluwalia et al. 2004; Burdyga & Wray, 2005). In models of hypertension MaxiKβ1 subunit expression is reduced (Amberg et al. 2003) while in ageing coronary arteries the expression of both MaxiKα and MaxiKβ1 are drastically decreased supporting an increased coronary reactivity in the elderly (Marijic et al. 2001; Nishimaru et al. 2004). The first MaxiKα channelopathy, a human syndrome of concurrent generalized epilepsy with paroxysmal dyskinesia, is caused by a single missense mutation in the regulator of conductance for K+ (RCK) domain of the channel protein (Du et al. 2005) (Fig. 1). Gene deletions have also underscored the physiological relevance of MaxiK channel subunits with emphasis in smooth muscle functions where this channel is particularly abundant. Silencing MaxiKα produces incontinency, bladder overactivity, and erectile dysfunction (Meredith et al. 2004; Werner et al. 2005), while deletion of the smooth muscle MaxiKβ1 subunit causes slight hypertension and increased contractile response to vasoactive agonists (Brenner et al. 2000; Pluger et al. 2000). Deletion of the MaxiKα also affects cerebellar function and produces progressive deafness (Sausbier et al. 2004; Ruttiger et al. 2004). Thus, the MaxiK channel is emerging as an additional target for the treatment of a variety of smooth muscle and brain diseases.
MaxiK and signalling
MaxiK channels can work as cell rheostats favouring or limiting depolarization and contraction depending on the cell status or stimulus. Smooth muscle dilators such as β2-adrenergic agonists, nitric oxide, prostaglandin I2, arachidonic acid, calcitonin gene-related peptide, and carbon monoxide (CO) activate MaxiK channels (Marijic & Toro, 2001; Schubert & Nelson, 2001), while potent vasoconstrictors like angiotensin II and thromboxane A2 can inhibit MaxiK activity via Src tyrosine kinase-dependent phosphorylation (Alioua et al. 2002). Most agonistic actions involve intracellular signalling pathways like: (i) Ca2+ sparks that appear to be involved in the action of vasodilators via a negative feedback mechanism; (ii) kinases that phosphorylate the channel protein and modify its electrical activity; and (iii) G proteins that regulate channel activity independent of downstream kinases (Jaggar et al. 2000; Marijic & Toro, 2001; Schubert & Nelson, 2001). Accumulating evidence indicates that MaxiK modulation and signalling involves close associations with a variety of protein partners (Table 1).
MaxiK molecular complexes, partners and sites of interaction
In smooth muscle, the basic MaxiK working unit is probably that formed by α- and β1-subunits (Knaus et al. 1994; Giangiacomo et al. 1995; Wallner et al. 1996). A growing list of interacting proteins point to the view that MaxiKα intracellular domains, especially its long carboxyl terminus, may serve as an anchor port where several proteins come together to exert their functions (Fig. 1). Whether the interactions require tetramer formation and/or are influenced by α subunit assembly with β or other accessory subunit(s) are still open questions. Initial evidence that MaxiK channels may form strong complexes with signalling cascades comes from reconstitution experiments where MaxiK channel partners did not dissociate freely in the immense lipid bilayer but remained attached to the channel modulating its function. For example, channels reconstituted from uterine and coronary smooth muscle could be activated by β2-adrenergic agonists or GTP-γ-S indicating that β-adrenergic receptors and G-proteins remained attached to the channel (Toro et al. 1990; Scornik et al. 1993).
β2-adrenergic receptor (β2AR) and Ca2+ channels.
In human and rat smooth muscles (myometrium, lung, bladder and aorta) β2AR and MaxiKα associate in a macromolecular complex via the third intracellular loop of β2AR. MaxiKα also associates with α13.2 T-type calcium channels from brain and with L-type Ca2+ channels from bladder and brain tissues. MaxiKα and L-type Ca2+ channel association seems to be mediated by β2AR; as in heterologous expression, β2AR was necessary for this association (Chen et al. 2003; Liu et al. 2004; Chanrachakul et al. 2004; Grunnet & Kaufmann, 2004). Further, in brain, MaxiK can also associate with protein kinase A (PKA) (see below) and A-kinase anchoring protein (AKAP150), and in heterologous coexpression experiments, β2AR expression is necessary for association of MaxiKα and AKAP79 (human homologue of rat AKAP150), and the tripartite complex β2AR–AKAP–MaxiKα shows maximal activation of MaxiKα by β2-agonist stimulation (Liu et al. 2004). Consistent with β2AR requirement for MaxiKα and AKAP79 association, an AKAP-competing peptide was unable to prevent cAMP-dependent regulation of a MaxiKα splice variant expressed in HEK293 cells (Tian et al. 2003). One possibility is that AKAP and PKA are indirectly associated with MaxiK via β2AR; however, the precise mechanism of PKA targeting to MaxiK facilitating its phosphorylation remains elusive and may depend on the tissue or signalling cascade involved. Yet, the signalling complex comprised of MaxiKα, β2AR, AKAP, PKA and L-type Ca2+ channel would facilitate β2AR receptor signalling to MaxiK and its regulation by phosphorylation and Ca2+ in brain and most probably in smooth muscle.
Protein kinases.
Several protein kinases can associate to MaxiKα. Drosophila MaxiKα (dSlo) associates with Src tyrosine kinase and binds directly to PKA via a C-terminus region containing a PKA substrate site, RRGS* (Wang et al. 1999; Zhou et al. 2002). In contrast, the mammalian MaxiKα does not bind directly to the catalytic subunit of PKA (PKAc) in vitro; instead, brain PKAc associates with a putative leucine zipper at the MaxiKα C-terminus via an unknown mechanism (Tian et al. 2003). Proline-rich tyrosine kinase 2 (PyK2) can also associate with Maxiα when both are transiently coexpressed in a catalytically active (wild type)-dependent manner (Ling et al. 2004) suggesting that their association is phosphorylation dependent. In osteoblasts, prostaglandin E2 induces spleen tyrosine kinase (Syk) recruitment by MaxiKα C-terminus to its immunoreceptor tyrosine-based activation motif (ITAM), and hypotonic shock increases MaxiKα association with focal adhesion kinase (FAK) suggesting a role of MaxiKα in osteoblast mechanotransduction (Rezzonico et al. 2002, 2003). In native smooth muscles, however, information is lacking about MaxiKα associations with protein kinases. Nevertheless, the functional coupling of c-Src and MaxiK channels in vascular smooth muscle (Alioua et al. 2002) already indicates that these proteins may associate in a functional complex (Fig. 2).
Caveolin-1 and caveolae.
Caveolin, an integral membrane protein that acts as the coat protein of caveolae (plasma membrane invaginations enriched in cholesterol), associates with MaxiKα in cultured vascular endothelium and in human myometrium. Disruption of caveolae by both cholesterol depletion and caveolin-1 siRNA increases MaxiK currents, while the caveolin-1 scaffolding domain prevents MaxiK stimulation by isoproterenol (isoprenaline) in vitro. These data suggest a negative modulatory effect of caveolin on MaxiK, although the exact mechanism is still unknown. In myometrium, MaxiKα associates and colocalizes with caveolin-1 and caveolin-2, but not caveolin-3. Interestingly, actin also seems to be part of the MaxiKα complex. Similar to caveolae disruption, disassembly of the actin cytoskeleton in cultured smooth muscle increased the open-state probability of the channel (Wang et al. 2005; Brainard et al. 2005). Since MaxiKα possesses two consensus caveolin binding sites it is likely that caveolin directly binds to MaxiKα targeting MaxiK to caveolae to join other signalling partners (Fig. 2).
Oxygen and CO sensing.
Both haem and haemoxygenase 2 (HO-2) members of the O2-sensing and CO-generating cascade associate to MaxiKα. Haem seems to bind directly to the MaxiKα C-terminus at a haem binding pocket, CXXCH, since its mutagenesis prevented haem-induced channel inhibition (Tang et al. 2003). HO-2, which in the presence of oxygen uses haem and electrons from NADPH cytochrome-P450 reductase to generate CO, associates with MaxiKαβ complex in vitro. Further, HO-2 mediated regulation of MaxiKαβ by haem (+ NADPH) under hypoxia/normoxia is mimicked in carotid body glomus cells indicating that HO-2 acts as an O2 sensor of MaxiK channels (Williams et al. 2004a). Thus, MaxiK association with haem and HO-2 facilitates production of CO in situ, which enhances MaxiK activity, providing an additional gaseous mechanism for the maintenance of vascular tone. Interestingly, in vitro studies show that the MaxiKαβ complex can be coimmunoprecipitated with γ-glutamyl transpeptidase, a potential O2 sensor that metabolizes glutathione. However, siRNA experiments demonstrated that this enzyme is not involved in the acute response of MaxiK to hypoxia (Williams et al. 2004b).
Synaptic proteins and other partners from brain.
Slob, dSlo interacting protein 1 (dSLIP1) and the zeta isoform of 14-3-3 protein were the first brain MaxiKα partners discovered. Slob increases dSlo activity, dSLIP1 controls dSlo surface expression in vitro probably binding to its C-terminus, and 14-3-3 interacts with dSlo indirectly via Slob-controlling channel activity. This protein complex is localized presynaptically at Drosophila neuromuscular junctions and the binding is regulated by calcium/calmodulin-dependent kinase II phosphorylation (Xia et al. 1998; Schopperle et al. 1998; Zhou et al. 1999). Syntaxin 1A, an integral component of the neurotransmitter release machinery, can associate with MaxiKα (Ling et al. 2003) either via its C-terminal tail or the S0–S1 loop (Cibulsky et al. 2005). This interaction reduces MaxiK channel activity and perhaps decreases neurotransmitter release. β-Catenin is associated with MaxiKα in chicken cochlear hair cells via the hydrophobic S10 segment of MaxiKα suggesting that β-catenin may provide a physical link for MaxiK channel with other functional proteins at the presynaptic area (Lesage et al. 2004). Cereblon, a rat homolog of the human ATP-dependent Lon protease, binds to MaxiKα via its carboxyl-terminus (encompassing S6–S9 just prior to the Ca2+ bowl) preventing its surface expression in COS-7 cells, although both proteins colocalized to some degree in the cell body and dendrites of cultured hippocampal neurones (Jo et al. 2005). Microtubule-associated protein 1A and ankyrin-repeat family protein, ANKRA, also bind to the carboxyl-terminus of MaxiKα, with ANKRA binding towards the extreme carboxyl-end region and modifying channel kinetics (Park et al. 2004; Lim & Park, 2005). Thus, in neurones as in myometrium, MaxiKα seems to interact with scaffolding proteins and the cytoskeleton organization.
In summary, the MaxiKα subunit associates with multiple signalling proteins via its C-terminus. MaxiKα partners in specific smooth muscles or cell types, the plasticity of the interactions, the subcellular localization and their specific function is an emerging field that will help us understand the intricate relationships of MaxiK with signalling cascades and its possible role as signal transducer.
MaxiK channel topological characteristics A, protein topology of MaxiKα subunit. Arrows mark boundaries of translated constitutive exons (1–27). Asterisks, sites of splice variation. Green asterisk, site of STREX splicing. Dashed blue line, string of asparate residues forming the ‘Ca2+ bowl’. Dashed square, RCK domain. B and C, model (top view) of tetrameric assembly showing K+ in the pore with and without β-subunits and proteins interacting at the C-terminus.
MaxiK–GPCR functional coupling via Src tyrosine kinase and possible partners in smooth muscle microdomains A, aortic contraction by 5-hydroxytryptamine (5-HT) is reversed by an inhibitor of Src tyrosine kinase (PP2) in a dose-dependent manner demonstrating Src role in 5-HT-induced contraction. Bar, 10 milli Newtons (mNt). B, MaxiK blockade with Iberiotoxin (IbTx) hampers PP2-induced relaxation highlighting MaxiK–GPCR coupling via Src. C, hypothetical diagram of possible MaxiK interactions with partner proteins in caveolae domains where signals are integrated and transduced for a contractile response. GPCR, G-protein coupled receptor; Cav-1, caveolin-1; Gα,β,γ, G-protein α-, β- and γ-subunits. Panels A and B reproduced with permission from Alioua et al. PNAS 99, 14560–14565, Copyright 2002 National Academy of Sciences, USA (Alioua et al. 2002).
Table 1. MaxiKα partners
Acknowledgements
This work was supported by AHA (A.A. and M.E.), and NIH (HD046510, E.S.; HL54970, HL47382, L.T.).
Footnotes
-
- Accepted October 20, 2005.
- Received September 16, 2005.
- Revision received October 20, 2005.
- The Physiological society 2006
