universiteit antwerpen - IAP-P6/31

1. Molecular architecture of K⁺ channels and their role in excitability (WP1)

Snyders (P4) - Tytgat (P2) - Seutin (P3) - Heidbuchel, Willems, Sipido (P1) - M. Vos (EU1)

Previous findings and rationale

In this work package we will investigate the molecular architecture of K⁺ channels in relation to their function (and malfunction) in specific excitable cells.

Previous work in the group of P4 has identified several novel subunits and interaction sites of K⁺ channels. Screening of the human genome identified 3 α-subunits and a splice variant of a β-subunit (Ottschytsch et al, 2002, Van Hoorick et al, 2003). The α-subunits are electrically ‘silent’ in heterologous expression systems but interact with Kv2.1 and Kv2.2 resulting in currents with altered biophysical properties. Since the Kv2 subunits are ubiquitously expressed and the functional properties of Kv2.1 and Kv2.2 are very similar we thus hypothesize that heteromultimerization with silent subunits underlies the functional diversity that is observed in native tissue. The KchIP1b β-subunit interacts with Kv4.x subunits but recent work also identified an interaction with Kv1.5. In a Y2H screen using the intracellular N-terminal segments of members from the Kv1 through Kv9 subfamilies, interaction with the corresponding segment of Kv6.3, Kv10.1 and Kv11.1 was examined. These experiments revealed interactions of the novel subunits with Kv2, Kv3 and Kv5 subfamily (Ottschytsch et al, 2002). Additional experiments revealed interaction of the silent subunits with Kv3.1 and Kv5.1. These data strongly suggest that these interactions might be widespread and functionally important. This hypothesis will be investigated further. Analysis of K⁺ channel expression and diversity of subunit composition in native tissues might provide further clues to test this hypothesis. Furthermore, Kv2 and Kv4 subunit have been identified in DRG neurons but the presence of silent or KChIP subunits has not been investigated. As a consequence, heteromultimerization of Kv2 with silent subunits and interactions between Kv4 and KChIP might occur in these cells. As such, multimerization of subunits is likely to contribute to the functional diversity in K⁺ channels that is observed in native tissue.

P3 has been investigating the role of various channels in the regulation of the firing rate and pattern of midbrain dopaminergic (DA) neurons. These neurons are important because their dysfunction leads to devastating illnesses such as Parkinson’s disease and schizophrenia. In vitro, it was shown that I_h channels contribute to the slow pacemaker activity of a subgroup of DA neurons (Seutin et al, 2001). P3 and others showed that SK channels (probably mostly made up of SK3 subunits) underlie the medium duration afterhyperpolarization (mAHP) that follows each action potential in these neurons. However, the neurophysiological consequences of SK channel blockade in DA neurons in vivo were unknown. Using several recent SK channel blockers – some of which were characterized by the same group in collaboration with Dr. JF Liégeois (ULg) –, partner 4 demonstrated that SK channel blockade potently facilitates burst firing in these neurons in vivo (Waroux 2005). This is a significant finding because burst firing induces a large increase in dopamine release at the level of the terminals. More recently, this lab demonstrated a very robust effect of the KCNQ opener retigabine in these cells. This study was performed in collaboration with scientists at Neurosearch (Denmark) who had shown that DA neurons mainly express KCNQ4 channels. In the slice preparation, retigabine inhibited the firing and hyperpolarized all DA neurons; this effect was potently inhibited by the KCNQ blocker XE991. Retigabine also inhibited the firing of these cells in vivo (Hansen et al 2006). Surprisingly, however, administration of XE991 had little effect by itself in vitro (the drug is not suitable for in vivo experiments). As a consequence, the level of activity of DA neurons could be modulated by SK and/or KCNQ channels.

The first short chain scorpion venom peptide (noxiustoxin) was isolated from Centruroides noxius and affected potassium permeability in squid giant axon. The discovery of charybdotoxin significantly increased research in this area, since it was an excellent ligand model for studying K⁺ channel function and structure. It has also become clear that numerous other venomous creatures (snakes, spiders, snails,...) can contain valuable bioactive substances or toxins that interact specifically and with high affinitity with different classes of K⁺ channels (e.g. Tytgat et al., 2001; Fan et al., 2003). The K⁺ channel acting peptides (KTx’s) were classified into three families (the α-, β- and γ-scorpion toxins) and a systematic numbering system was proposed (for review see Tytgat et al, 1999). Since then, the number of known peptides, extracted from scorpion venoms that block or modify potassium permeability in excitable and non-excitable cells have increased drastically to the point that now about 120 peptides are known. Recently, so-called evolutionary trace analyses were also carried out to understand the evolution of the pharmacologically important epitopes (i.e. mutation rate of particular amino acids), responsible for the selectivity and affinity of the toxins towards diverse K⁺ channel isoforms (Zhu et al., 2004). The laboratory of P2 has developed a large library of toxins from animal venoms. Purification and screening approaches have identified peptides to block ion channels with high specificity. To date multiple toxins have been identified that block a variety of ion channels including Na⁺ and K⁺ channels. The availability of all these, will allow us to use them in this WP for a better understanding of the function, structure and pharmacology of the different types of K⁺ channels that will be studied. The binding site for a biological toxin is very specific and is clearly related to the amino acid sequence/protein structure. As a consequence, the characterization of the binding site might enable to identify other targets on a more hypothesis driven basis. Furthermore, in the process of drug development the optimization of the drug is an important factor. As toxins might be therapeutically valuable, a detailed knowledge of the molecular nature of the binding site as well as the drug or toxin binding properties will lead to a more hypothesis driven approach in the optimization.

Within the group of P1, several lines of research have focused on the role of K⁺ channels in membrane stability in cardiac cells. In collaboration with Partner EU1, the importance of alterations in K⁺ channel function for enhanced susceptibility in compensated cardiac hypertrophy was reported (Sipido et al, 2002). More recent work from the group of EU1 has focused on expression of subunits and found downregulation of expression of KCNQ1 as well as KCNE1. Currently within the group of P1 two large animal models are under study for electrical remodeling. Willems and Heidbuchel are studying a sheep model of atrial fibrillation. While in this model K⁺ channel alterations may not be at the center of the arrhythmic event, block of K⁺ channels, more specifically the HERG channel, has been shown to be a valid approach for treatment. Within the project this model will be used to test and validate peptide channel blockers. A second large animal model is the pig with ischemic cardiomyopathy. In these animals we observe action potential prolongation, but the underlying channel remodeling still needs to be explored. Characterization of K⁺ channels and subunits, as well as the potential for pharmacological block will be explored in within this project.

The Center for Inherited Cardiac Disease is directed by Hein Heidbuchel (P1). The patient population includes several hundred families (from both Flanders and Wallony) with congenital arrhythmic disease. Screening of these patients for mutations or polymorphisms has already identified a number of novel mutations, e.g. in the Na⁺ channel gene SCN5A and the K⁺ channel gene KCNH2, but several remain to be investigated (Rossenbacker et al, 2004). P4 has performed such analysis regularly for hERG and KCNQ1 (e.g. Paulussen et al, 2002, Boulet et al, 2006) and thus collaboration of P4 and P1 will allow in depth characterization of novel mutations.

Aims:

explore the functional consequences of specific subunit assemblies of K⁺ channels
evaluate the pharmacological spectrum of novel toxins
identify relevant binding sites through mutational analysis
explore the architecture of K⁺ channels in dopaminergic neurons
identify alterations in expression of K⁺ channel subunits in congenital and acquired cardiac disease

Methodology

Screening of these patients for mutations or polymorphisms has already identified a number of novel mutations in the Na⁺ channel gene SCN5A and the K⁺ channel gene KCNH2, but several remain to be investigated (Rossenbacker et al, 2004). P4 has performed such analysis regularly for hERG and KCNQ1 (e.g. Paulussen et al, 2002, Boulet et al, 2006). The screening for the interactions of proteins by protein-protein interaction assays will be performed with full length or parts of α- and/or β-subunits of ion channels. We can use the commercial systems (Y2H or Cytotrap) or the MAPPIT system developed by J. Tavernier (University of Gent, Eyckerman et al, 2001); the latter can be used with full length membrane proteins. Identified interactions will be confirmed by co-IP by the introduction of tags (HA, c-myc, 6-His) for expression studies and with antibodies for native tissue. Alternatively we envisage the use of FRET to confirm interactions in living cells. For this purpose we will construct fusion proteins with CFP, YFP; the technology for FRET is available within the consortium. Using these approaches we expect to find not only α- and β-subunits, but also interactions with e.g. cytoskeleton proteins. While not discounting their value we intend to focus initially on α- and β-subunits. Functional electrophysiological studies will be performed in expression systems; both the mammalian and oocyte systems are available. The α- and β-subunits under investigation are subcloned in a mammalian or oocyte expression vector. In the case of mammalian expression the cells can be transiently or stably transfected. In the case of oocytes, the RNA is synthesized and injected. The mammalian cells are investigated electrophysiologically by patch clamp measurements in the whole-cell or excised patch (for single-channel measurements) configuration.

Thanks to the ‘toxinology’ research over the past 10 years, P2 has a large library of unique venoms (from scorpions, spiders, snakes, sea anemones, cone snails, amphibians, cnidarians)> these are available for the IAP partners (not commercially available). The active components present in these venoms, will be purified and identified by means of 2 state-of-the art machines that we have: i) LC/DAD/ UV/VIS with automated fraction collection (Gilson) and ii) LC/MS (Finnigan, Deca XP). To this end, we will use a combination of different HPLC columns: C4, C8, C18 and/or phenyl for reversed phase chromatography, together with ion exchange columns. In a ‘bioassay-guided manner’, venoms will be screened via electrophysiology (two-electrode voltage clamp (Axon Instruments) in Xenopus laevis oocytes expressing one of the following cDNA clones: Kv1.1, 1.2, 1.3, 1.4, 1.5, Shaker IR, Kir2.1, Kir3.1, Kir3.2, hERG, minK, KvLQT1) and the peptide/toxin of interest will be identified in an iterative way. Concomitantly, the full primary sequence of the peptide/toxin will be determined via one or more strategies mentioned hereafter: i) Edman degradation (i.e. the biochemical way), ii) mass spectrometry (i.e. the newest LC/MS/MS way) or iii) cDNA cloning (i.e. the molecular biology way). It should be mentioned that we have in our laboratory all the necessary cDNA clones for the bioassay-testing, as well as the expertise in molecular biology for cDNA cloning of the peptides/toxins. For the Edman degradation, a long-standing collaboration with Prof. E. Waelkens (Laboratory of Biochemistry, KUL) and Prof. L. Schoofs (Laboratory for Developmental Physiology and Molecular Biology, KUL) exists. After the primary bio-assays in the lab of P2, further application of peptides to study K⁺ channel architecture will be done in the lab of P4. Here the different subunits as identified can be expressed in mammalian cell lines (HEK293, CHO, Ltk, COS). In a next step they are used to probe the presence of subunits in native cells (P3, 1, see below).

Mutational analysis of K⁺ channel subunits and toxin interactions will further characterize the binding site for specific peptides. To test for the different factors that determine toxin binding a two-fold approach will be used. First, by combining mutagenesis of the putative binding site with characterization of several blocking parameters (affinity, kinetics and voltage dependence), we will identify the residues that are important for toxin binding to the ion channel; such analysis has already been performed by P4. Second, for peptidyl toxins, we will also modify the toxin by mutagenesis of its sequence. As a result the toxin binding to the binding site can be analyzed in detail. Furthermore, by combining mutagenesis of the channel with several modifications to the toxin, we will be able to determine all aspects of drug binding and binding site at the greatest detail, in part through mutant cycle analysis (Imredy and MacKinnon, 2000). Comparing the sequence of the binding site with the corresponding one of other ion channels should allow to identify putative other targets or to predict potencies. This should allow verifying if the interpretation of the interaction of the toxin with its binding site is correct.

Previously both brain slice recordings and in vivo recordings in the anaesthetized rat (coupled to drug iontophoresis) were used. To investigate the SK3 and KCNQ4 channels as potential new ion channel targets to modulate the level of activity of DA neurons, organotypic slice cultures will be developed (P). DA neurons are known to retain most of their physiological properties in these organotypic slice cultures of DA neurons. At first the expression of these channels will be confirmed on the RNA level (P4). Furthermore, the siRNA strategy will be developed for these cultures (with P4 for siRNA and P1 for adenoviral transfections). This strategy will be validated by combining electrophysiological (P3) and immunocytochemistry (Danish collaborator of P3) methods. Slice cultures of these cells will also be used for the experiments of P3 in WP5 (see below). As a second tempting but high risk approach, we envisage to downregulate the density of these channels in vivo by local application of siRNA. The ventral tegmental area will be chosen because it is an oblong area which appears to be well suited to obtain a relatively homogenous diffusion. This allows to address the firing rate and pattern of rat DA neurons after chronic downregulation of their SK3 channels. In that case the presence of bursting observed after acute blockade of the channels (Waroux 2005) will be investigated. Conversely, an adaptation of the properties of the DA neurons or the network that they are part of could occur. This will also be addressed in mice using conditional SK3 knock-out mice received from Dr. J. Adelman (Vollum Institute, Portland, USA).

For cardiac remodeling in the case of atrial fibrillation and of ischemic cardiomyopathy, basic functional characterization of K⁺ channels is done by the group of P1 (Willems, Sipido). ‘Classic’ channels are studied as drug-sensitive components (low-dose 4AP-sensitive I_Kur, dofetilide-sensitive I_Kr, chromanol-sensitive I_Ks). These data are compared to directed expression studies. More importantly, in the lab of P4 a global screening for non-classical channels and subunits will be performed with the unbiased microarray approach. Until recently DNA chips for such microarray experiments were based on commercially available oligo’s and incomplete with respect to α- and β-subunits of ion channel subunits. The custom arrays of the VIB micro-array facility contains 21000 spots which correspond to about 10000 genes ( http://www.microarrays.be/ ) and has been updated by the Snyders group with respect to ion α- and β-channels subunits. These microarrays have been used to investigate the expression in the HEK293, COS, Ltk and CHO cell lines. Moreover, specific microarrays for channelopathies have been developed and can be accessed through collaboration (Escande, Nantes, Lamirault et al. 2006). When specific toxin peptides are available (P2), screening for expression by microarray will be complemented by functional studies in the native cells.

Identification of mutations and/or polymorphisms of K⁺ channels relevant for human arrhythmic disease is done in the Center for Human Inherited Cardiac Disease (Heidbuchel, Partner 1) with the genetic analysis/sequencing performed at the Center for Human Genetics (KUL). Investigation of the functional consequences of mutations/deletions will be performed by partner 4 by mutagenesis, expression, electrophysiological analysis and by confocal microscopy of tagged subunits.