CNG and HCN channels: a structure-function analysis

In recent years, growing importance has been given to the studies of genomics and proteomics, the disciplines that collect available knowledge about genomes and proteins and whose main aim is to exploit the usefulness of this information to understand how organisms work and possibly foresee how eventual diseases can be cured. The research on proteins, meant as the essential molecular tools for cellular functioning, has a key role, and can be conducted at various levels, corresponding- to the degrees of structural complexity; but the ultimate clue to the comprehension of the molecular mechanisms of polypeptides, from enzymes to receptors and ion channels, resides in their three-dimensional structure. Until recently, for ion channels, due to the difficulties of obtaining crystals, only indirect investigations were possible (e.g., mutagenesis and electrophysiological characterization). Then, in 1998, the structure of the K+ channel KcsA was determined by x-ray chrystallography (Doyle et al., 1998), and the molecular principles of its conduction and selectivity were explained. The present research focuses on channels that share relevant features with K+ channels like KcsA, but differ from them for some important characteristics. Cyclic nucleotide-gated (CNG) channels display a pore region that is homologous to that of K+ channels, but they are not as selective in the permeation of cations, and their gating is ipsensitive to voltage, while it requires the direct binding of a ligand (cAMP or cGMP). Hyperpolarizationactivated and cyclic nucleotide-gated (HCN) channels, on the other hand, have an even higher affinity in their pore region with K+ channels, and their selectivity towards cations is intermediate between that of K+ and CNG channels; like the former, they are voltage-dependent, and their gating is modulated by cyclic nucleotides. CNG channels are fundamental in phototransduction and olfactory transduction where simple electrical stimuli are transferred as flow of information; HCN channels are at the basis of rhythmic activities including the pacemaker mechanisms of cardiac cells, the repetitive firing of neurons and the swimming behaviour of sperm. Within the more general investigation of ion channels structure and function, the aim of my thesis is to elucidate the structure of the pore regions of these proteins, to establish whether these pore regions are similar or not to the one of KcsA, and to gain insights about the molecular characteristics at the origin of the differences with K+ channels. In more detail, I have investigated the pore regions of the a-subunit of the CNG channel from bovine retinal rods (brCNGCa) and the HCN channel from the sperm of the sea urchin (SpHCN). I have used tools of molecular biology (site-directed mutagenesis) and electrophysiology (patch-clamp), and I have heterologously expressed the above mentioned channels in Xenopus laevis oocytes. My investigation aJ!d the results obtained can be summarized as follows. 1) The topology of the pore region of brCNGCahas bee.ii studied by testing the accessibility to Cd2 + of serially substituted cysteine residues, thus extending and completing a previous investigation (Becchetti et al., 1999) and gaining a clearer understanding of the position of important residues close to the selectivity filter possibly involved in gating, and whose localization within the pore loop was not precisely defined yet. My results allow to sketch a threedimensional model of the brCNGCa pore region which is different from a previously proposed one (Sun et al., 1996). 2) In brCNGCa, the substitution of the glutamate residue in position 363 with a neuter alanine induces desensitization of the channel in presence of a steady ligand concentration (Bucossi et al., 1996). In view of understanding whether channel gating in ligand-gated channels occurs through both local interactions between amino acid residues and global channel rearrangement, this artificial CNG channel desensitization offers useful general clues to study the energetic interactions occurring between residues involved in channel gating. Accordingly, I have investigated eventual major rearrangements in the pore region topology of the mutant channel E363A, by means of the substituted cysteine accessibility method (SCAM). I found that the pore region topology was not disrupted by the mutation, and that the nature of the residues immediately close to position 363 influences the structural rearrangements leading to desensitization. 3) In the pore region of HCN channels, some resid{ies in key positions are different from the homologous ones in most K+ channels. In particular, the GYG triplet which is the signature sequence of K+selective channels is followed by a basic or neuter residue, in place of an acidic one. Also, N-terminal to the same triplet, all HCN channels bear a cysteine residue, in place of a larger threonine or serine. In view of understanding whether these differences may account for the poor cation selectivity of HCN channels and for their regulation by external chloride, the topology of residues C428 and K433 in the SpHCN channel has been investigated. My results indicate that the pore loop topology resembles the one of KcsA, with residue C428 facing the intracellular side of the membrane and thus being responsible for the block by internal Cd^2+, and residue K433 located extracellularly, but with its side chain pointing towards the lipid phase, and being alone not responsible for ion selectivity and for regulation by external chloride.