RICERCA ITALIANA     

IONIC CHANNELS ACTIVATED BY MEMBRANE HYPERPOLARIZATION AND REGULATED BY CYCLIC NUCLEOTIDES (HCN CHANNELS)

Scuola Internazionale Superiore di Studi Avanzati di Trieste

Abstract

The present proposal aims at clarifying functional properties and at determining molecular mechanisms of ionic channels gated by membrane hyperpolarization and regulated by cyclic nucleotides (CN). In order to understand the physiological properties of HCN channels and the molecular basis of their functions, it is necessary to propose a highly interdisciplinary approach that in Italy could be made exclusively by joining together laboratories of different scientific areas but sharing the same interests. The present proposal will combine the tools more commonly used for ion channel analysis, i.e. biochemistry, structural and molecular biology, electrophysiology, with more theoretical approaches from Bioinformatics and Computational Physics and Chemistry. Our specific aims in the present proposal are:

1 - to study at a molecular level, the composition in subunits and the conformational changes associated with the opening and closure of HCN channels
2 - to build molecular models of HCN channels, to study the changes associated to the opening and the closing of the channels and the possible ways for the interaction among different HCN isoforms
3 - to study the cellular and subcellular localization of the different isoforms in order to try to understand how they are distributed and how the targeting and turnover processes in the membrane are regulated. At this stage, we will try to establish the nature of possible molecular interactors and their functional effects on the distribution of the above mentioned channels.
4 - to identify and study from a functional point of view the role and the modulation of these channels in several nervous structures
a) Olfactory neurons from the main and accessory olfactory bulbs and epitheliums
b) Retinal neurons and in particular in bipolar cells
c) Hippocampal and cerebellar neurons
d) Oscillatory neurons. <<<

Principal Investigator

Vincent Aldo TORRE Scuola Internazionale Superiore di Studi Avanzati di TRIESTE

Research Objectives

The present proposal aims at clarifying molecular mechanisms and functional properties of ionic channels activated by hyperpolarization and sensitive to cyclic nucleotides(CN).
Ion channels sensitive to CNs fall into different subfamilies of related proteins. Members of one subfamily, designated CNG channels, require cAMP and cGMP for opening but are largely voltage-independent. Members of a second subfamily, the so-called „pacemaker" channels, are activated by hyperpolarization and by CNs (HCN channels).

HCN channels are present in cardiac as well as in many neuronal tissues (Santoro et al, 1997, 1998; Moosmang et al., 2001) and play a fundamental role in visual signal processing in the retina of mammals (Gargini et al., 1999; Demontis et al., 1999,2002). CNG and HCN channels mediate sensory transduction and processing in vision and olfaction. These channels, although belonging to the same super-family of voltage-gated channels, differ significantly from usual K+, Na+ and Ca2+ channels, by their specific gating and selectivity.
We propose a highly interdisciplinary approach to understanding the physiological properties of HCN channels and the molecular basis of their functions. This proposal will combine the usual tools used for ion channel analysis, i.e. biochemistry, structural and molecular biology, electrophysiology, with more theoretical approaches from bioinformatics, computational physics and chemistry.
Such an multidisciplinary approach is necessary for studying complex molecular assemblies or structures such as ionic channels and their modulation mechanisms or interactions with other proteins.
This approach is not only based on the four UdRs composing the present proposal, but benefits also from the international collaborations of the individual UdRs. Indeed, the UdR SISSA -Torre has a close collaboration with Kristina Djinovic and Doriano Lamba of the Structural Biology Unit of ELETTRA, the synchrotron radiation facility at Basovizza near Trieste with which SISSA is running a joint PhD program in Structural and Functional Genomics. The UdR SISSA-Torre collaborates also with Benjamin Kaupp (FZJ Julich, Germany) and Gebhard Shertler (MRC Cambridge, UK) within an HFSP project on the understanding of the relation between structure and function in CNG and HCN channels. The UdR at SISSA-Menini has a close collaboration on the investigation of olfactory coding with Peter Mombaerts (Rockefeller Univ., New York, USA) and with Stuart Firestein (Columbia Univ., New York, USA). The UdR of Milan is currently collaborating with W. Kühlbrand (Max Plank Inst. for Biophysics, Frankfurt) and Daniel Minor (U. of California, San Francisco) concerning direct and indirect approaches (cristallography and directed evolution, respectively) to the structural studies of the miniature K+ channel Kcv. More specifically, on HCN channels, the UdR of Milano collaborates with R. Robinson (Columbia University, New York) on the assembly of heterotetramers in HCN channels expressed in mammalian heart. Also, the UdR of Milano has received from W. Chung (Columbia University, New York) a mouse strain carrying a spontaneous mutation in the HCN2 gene, which results in a neurological (atassic) phenotype of the homozigote mice. The UdR in Pisa has a standing collaboration with J. Stone of the RSB, University of Canberra, AU, on immunohistochemistry and high resolution confocal microscopy of mammalian retina. Therefore, resources and competence available extend well beyond the four UdR, as requested by COFIN grant. Within the COFIN grant we plan to make substantial progress in three major topics of the biology of HCN channels:
- the intracellular regulation of their synthesis and the associated trafficking and interaction with other intracellular proteins (Phase 1 of the present project )
- the understanding of molecolar events underlying channel gating, i.e. the conformational changes associated to channel opening and closing (Phase 2 of the present project )
-the understanding of the physiological and functional role of HCN channels in the peripheral and central nervous system (Phase 3 of the present project )
The present project has the following specific aims:
1 - to identify the HCN isoforms (HCN1,HCN2,HCN3 ed HCN4) in olfactory neurons and to characterize their functional properties.
2 - to characterize HCN channels in retinal neurons and in particular in bipolar cells and to understand their functional and physiologcal role in the visual processing.
3 - to characterize biochemical and phisiological properties of HCN channels, in the hippocampus, in the cerebellum and in other pacemaker neurons of the central nervous system.
4 - to study the intracellular trafficking and the interactions of HCN channels in the cytooplasm and in other intracellular compartments. The role of HCN channels in synaptic plasticity will also be analysed and in particular their role in long-term potentiation and/or depression.
5 - to obtain a detailed molecular description of the conformational changes underlying channel gating, by electrophsyiological experiments with wild type and mutant channels.
6 - to build molecular models of HCN channels and of the conformational changes underlying gating and cyclic nucleotide regulation. We wiill also analyse the assembly of different isoforms in tetrameric structures.
The UdRs of Milan and SISSA-Torre will gather molecular information on the structural events underlying channel gating in HCN channels by appropriate electrophysiological experiments with genetically modified channels. These experimental data will be combined in preliminary models of channel gating of HCN and CNG channels. The UdR of Milan will also gather data on the intracellular trafficking of HCN channels and of the related molecular and cellular biology. The UdRs of Pisa and of SISSA-Menini will analyse functional properties of HCN channels in the retina and in the diverse olfactory neurons. The UdR of Pisa will analyse how HCN channels shape the graded voltage response in bipolar neurons. Together with the UdR of Milano the possible interaction of HCN and EAG channels in native (bipolar neurons) and recombinant channels (in oocytes and mammalian cells) will be analysed. The UdR of Milano will provide a mouse strain carrying a natural mutation in the CNB domain of HCN2, which results in a truncated protein lacking the CN binding site. The UdRs of Pisa and SISSA-Menini will study the properties of the current in bipolar neurons and olfactory neurons, respectively and will analyse its role in sensory processing. The present proposal is based on several collaborations planned among the proposers:
1 - The UdR of Pisa and of SISSA-Menini will collaborate and exchange experimental protocols of immunocitochemistry for labelling in the retina and in the olfactory systems. The UdR will make available to the other UdR their antibodies of the specific HCN isoforms.
2 - the UdR of Pisa and the UdR of Milan have planned joint experiments in order to analyse the role and properties of the CNB domain in HCN2 channels.
3- the UdR of Pisa and the UdR of SISSA-Torre will analyse together the high pass filtering properties originating by specific HCN channels in rods and in bipolar cells.
4 - the UdR of SISSA-Torre and the UdR-Milan will collaborate for the development of models of gating in HCN channels.
These collaborations are explained in more details in the Modulo B.
The proposal extends from establishing new basic functional and physiological properties to relate these properties to specific molecular mechanisms. The four UdRs of the present proposal will meet every 3 months to discuss the results obtained and plan new joint experiments. <<<

First Results

-Analysis of channels localization in brain tissue and heart
-Analysis of their subcellular localization and identification of subcellular compartments involved in trafficking
-Analysis of channel turnover
-Analysis of channel insertion in me <<<

Timescale

24 months

National and international background

STATE OF THE ART ON THE RESEARCH ON ION CHANNELS
Ionic channels are membrane proteins playing a fundamental role in cell physiology and in signal transduction and transmission (Hille 1992). Potassium, sodium, calcium and chloride ions are used ingeniously by living systems in the performance of fundamental cellular tasks. Through the action of ion pumps, a large fraction of a cell's metabolic energy is spent establishing transmembrane ion gradients. Ion channels have been extensively investigated from an electrophysiological and biochemical point of view and their amino acid sequence determined by molecular biology. The determination of the three dimensional (3D) structure of the K channel (Doyle et al 1998) from the bacteria Streptomices lividans - i.e. the KcsA channel - has opened a new era, allowing us to understand the relation between structure and function of ionic channels at atomic level.

HCN CHANNELS
The so-called Ih current (Di Francesco 1993) flows through HCN channels, discovered in sinoatrial node cells (Brown et al. 1979) and found in hippocampal pyramidal neurons (Halliwell and Adams 1982) and in photoreceptors (Bader et al 1979). HCN channels underlie the rhythmic electric activity of many neurons and of myocytes in the heart. Besides this pace-making function, HCN channels subserve other functions as well. In many neurons HCN channels co-determine resting potential and thereby play an important role in the integrative behaviour of neurons and the sensitivity of synaptic input (Pape 1996). HCN channels display some unusual properties. Unlike most other voltage-dependent channels, HCN channels are opened by hyperpolarisation rather than by depolarisation. The voltage of half-maximal activation, V1/2 varies between different channels and cell types, but values between -25 mV and -95 mV seem to be typical. Channels are selective for K+ ions, yet they are lacking the exquisite K+ selectivity of other K+ channels. As a consequence, HCN channels carry a Na+ inward current, which strongly depolarises the membrane. The activity of channels is directly enhanced by cAMP and low extracellular Cs+ concentrations block the current. HCN channels are cousins of the family of voltage-gated K+ channels and CNG channels. So far, 4 different mammalian genes and several invertebrate genes have been identified (HCN1-HCN4; Gauss et al. 1998; Santoro et al. 1998; Ludwig et al. 1998, Vaccari et al., 1999). The channels probably embody six transmembrane segments (S1-S6), a pore loop between S5 and S6 - much as K+ channels-, and a CNB domain in the C-terminus region - similar to CNG channels. The pore region is highly conserved among different HCN channels. They share with K+ channels a common GYG sequence motif, which has been recognised as the signature sequence of K+-selective channels. HCN channels have in the outer vestibule of the pore a postively charged or neutral residue (Lys in position 433 of the spHCN channel from sea urchin sperm and Arg in the HCN2 channel from human tissue) unlike K+ and CNG channels which all have a negatively charged residue (Asp in K+ channels and Glu in CNG channels). The fourth transmembrane segment (S4) of HCN channels has a typical voltage-sensor motif with 8-10 regularly spaced Arg or Lys residues at every third position. Similar sequence motifs have previously been identified in many voltage-dependent channels. HCN channels carry in the C-terminus region a domain of about 80-100 aa residues that is highly homologous to the CNB domains of CNG channels, protein kinases A and G and the catabolite gene activator protein (CAP). The CNB domain of HCN channels differs from that of CNG channels in a few key positions, suggesting that those differences might underly the profoundly higher selectivity for cAMP compared to CNG channels, which are more sensitive to cGMP. Recent studies indicate also that native HCN channels are, in most cases, heteromeric. In pacing cells of the hearth, both HCN1 and HCN4 have been identified, and activation kinetics, sensitivity to CN and voltage-dependence of the Ih current are intermediate between those of HCN1 and HCN4. Ih currents in CA1 hippocampal neurons have activation kinetics and sensitivity to cAMP intermediate between HCN1 and HCN2, suggesting that native channels are heteromeric. Additional complexity is added by the possible interaction of accessory proteins, such as MinK-related peptide isoforms, with HCN or EAG channels to form native Ih or specific macroscopic currents, such as Ikx in retinal rods. Interaction with other proteins may explain the selective targeting of HCN channels to specific subcellular compartments. The gating, i.e. the molecular mechanisms underlying the transition between the open and closed state, in CNG and HCN channels will be studied by the UdR of SISSA-Torre and by the UdR-Milan.

HCN CHANNELS IN THE RETINA
There is also evidence to show that HCN1 channels are involved in the processing of visual information. It has been observed that administration of organic blockers of hyperpolarisation-activated currents (Ih) in human subjects causes a variety of subjective visual disturbances which often take the form of phosphenes, smeared images or stroboscopic effects, that may be interpreted as a reduction of the "frame rate" of vision. ERG recordings conducted in animal models indicate that Ih blockers affect the frequency dependent gain of signal processing between rod photoreceptors and rod-bipolar cells (Gargini et al., 1999) suggesting that suppression of Ih in rods and/or rod-bipolar cells impairs the temporal acuity of vision.
The expression of HCN1 in mammalian retinal rods (Moosmang et al., 2001; Demontis et al., 2002) and the similarity of native Ih in rods with that recorded from HEK 293 expressing homomeric HCN1 subunits suggest that gating of homomeric HCN1 channels is involved in the control of the temporal resolution of vision. Investigations of the band-pass amplification properties of mammalian rods (Demontis et al., 1999) have shown that Ih gating in these cells improves the temporal resolution of the visual response, but they have also suggested that other steps of band-pass amplification are required to account for the final temporal performance of the visual system. It seems reasonable to assume that HCN channels with isoform compositions to be defined are expressed in post-synaptic neurones as a further step in the processing of temporal resolution of visual responses. Specifically, certain features of the visual temporal acuity would require the expression of HCN isoforms with specific gating and modulatory properties. In humans, psychophysical data obtained from both normal subjects and rod-achromats (individuals that lack color vision) show that an increase in the ambient light enhances the temporal resolution of vision (Conner and McLeod, 1977; Hess and Nordby, 1986). The role of HCN channels in visual processing in the retina will be addressed by the UdR of Pisa.

HCN CHANNELS IN NEURONAL NETWORKS
Activity in neural networks spreads by the conduction of action potentials along nerve fibres, and then by the chemical process of synaptic transmission at synapses. Among the channels contributing to modulation of neuronal excitability and synaptic strength, our interest is focussed on a recently cloned channel family, the HCN channels.
These channels are indeed involved in controlling cell excitability and adaptation, but also synaptic strength (Beaumont. & Zucker 2000; Mellor et al 2002) and rhythmic functions (Clapham 1998). Despite their important role, detailed information on the subcellular distribution in dendrites, cell soma, and nerve terminals and on their functional relevance in the process of synaptic transmission and synaptic integration is largely incomplete. This project is aimed at studying the distribution of ionic channels in hippocampal synapses and dendrites, identifying their functional role for synaptic activity and activity dependent changes such as LTP (Bliss & Collingridge1993; Mellor et al 2002) including recruitment of silent synapses.

STRUCTURE OF IONIC CHANNELS
Many properties of ion channels are difficult to discern in absence of high-resolution structural data. The structure of one of its members, a prokariotic K+ channel KcsA, reveals cation-selective architectural features that are probably shared by all family members (Doyle at al., 1998, MacKinnon et al., 1998). The recently published structure of a eukariotic Cl- channel ClC (Dutzler et al., 2002) suggests a basis for understanding the ion selectivity principles that apply to the entire family of anion channels. Recently, the structures of two related channels of the aquaporin family (AQPs) - the bacterial glycerol conducting channel GlpF and human aquaporin-1 have been determined to high resolution using X-ray crystallography. Basic mechanisms of mechanosensation have been elucidated by structure determination of a prokaryotic mechanosensitive ion channel MscL. Water soluble domains of ion channels are more amenable to X-ray diffraction studies; this is different from integral membrane proteins, which are intrinsically difficult to crystalise. The recently solved structure of the C-linker and of cyclic nucleotide binding domain of the mHCN2 channel (Zagotta et al 2003 ) is particularly relevant to the present proposal.


MOLECULAR MODELLING AND COMPUTATIONAL CHEMISTRY
Model-building on the basis of the known 3D structure of homologous proteins is at present the most reliable method to obtain structural information on proteins for which no direct experimental data are available (Chothia et al. 1986; Sander et al. 1991). At the basis of homology modelling lies the observation that 3D structures are better conserved during evolution than protein primary sequences (Vriend et al. 1991; Russel et al. 1992). Model building by homology is a multi step process which can be summarised by the following steps: i -template recognition; ii -alignment; iii -backbone generation; iv - loops generation; v - side chain generation; vi - overall model optimization; vii - model verification with optional repetition of previous steps.
The core of computational chemistry useful to understand the relation between structure and function in proteins is molecular dynamics (MD). This new emerging field usually distinguishes between classical MD simulations, based on classical mechanics, and ab initio simulations, based on approximate solutions of the Schroedinger equations. Classical MD simulations require the a priori knowledge of appropriate potentials describing the different atomic interactions, while ab initio simulations do not. In a classical MD simulation, one obtains the evolution of the system, i.e. of the considered protein, using Newton's second law of motion. The combined development of supercomputer technology, algorithm parallelisation (Lim et al. 1997), force fields (Cornell et al. 1995; Jorgensen et al. 1996) and time-saving techniques permits nowadays the simulation of complex biological systems in explicit solvent up to few ns and with higher accuracy, The major limitation of classical molecular dynamics is the exact computation of the potential energy V, which usually depends on a set of parameters typical of individual atoms. The ensemble of potentials and parameters is known as the force field (FF). Several FFs have been developed so far, such as CHARMM, AMBER, GROMOS and SYBYL. However, as polarisation effects are not implemented yet, they perhaps do not describe correctly processes for which polarisation plays a major role; furthermore FFs do not permit to follow the chemical evolution of the system e.g. bond formation and breakup, proton transfer phenomena. These types of phenomena can be reliably studied with the Car-Parrinello method (Car et al. 1985). This approach allows the researcher to perform parameter-free MD simulations in which all the interactions are calculated directly via an electronic structure method, ab initio density functional theory. The CP method has been applied to a variety of biochemical and pharmaceutical problems. <<<