Evolutionary origins of ion channels.

Ionic fluxes across cellular membranes mediate a tremendous variety of biological processes that are essential for the viability of most life-forms. These fluxes are important for cell volume regulation and swimming behavior in unicellular organisms, movements of stomatal pores in plants, and muscle contraction, exocrine cell secretion, and the generation of neuronal excitability in higher animals. In many cases, these biological processes are dependent on fluxes of specific ions tightly regulated by intracellular and extracellular signals. The permeation pathway responsible for moving ions across the hydrophobic environment of cellular membranes is provided by a diverse family of integral membrane proteins called ion channels, whose structural and functional properties have been the subject of decades of research. Members of this family have evolved unique and beautiful properties of ionic selectivity, activation gating (regulated channel opening), and inactivation gating (regulated channel closure) which make them ideally suited for each biologic task. Expression studies have suggested that these proteins are present in essentially all eukaryotes from protists and fungi to plants and animals (1). How has this rich functional diversity been created throughout evolution, and what did the primordial channels look like? Recent studies have provided compelling models for the generation of ion channel diversity (2, 3), and it has been suggested that these proteins arose in very early eukaryotes possibly through the modification of prokaryotic porins, colicins, or ABC (ATP-binding cassette) transporters (1). An important further step in understanding this evolutionary process comes with the report in this issue of the sequence analysis of a prokaryotic gene whose predicted protein product shares extensive topological and structural similarity with eukaryotic K+ channels (4). This protein most likely represents the most primitive homolog in the spectrum ofmodern ion channels and may therefore provide the closest look yet at the ancestral ion channel protein. Evolutionary studies are in general complicated retrospective studies attempting to infer the structural origins of modern proteins with information only from the successful survivors of the evolutionary process. Although it would be useful, knowledge of the structure of proteins from extinct organisms is lacking. What then can we infer about ion channel evolution from currently available information? Traditional biophysical methods have provided detailed functional analyses of ion channel properties, including the conformational changes underlying activation and inactivation, the mechanisms of ionic selectivity and permeation, and the specific interactions with pharmacological agents (1). In addition, recent molecular genetic techniques have facilitated the isolation, cloning, and expression of genes encoding ion channels in a variety of organisms (1, 5). These studies have shown that ion channels can be divided into distinct families and subfamilies, each with increasingly conserved functional and structural similarity across phyla. Indeed, a fundamental feature ofthese families is that there is more structural similarity among members of a given family from different species than among ion channels in a given species belonging to several families. Distinct ion channel families include the voltage/second messenger-gated channels, the inward-rectifier K+ channels, the ligand-gated channels, the internal Ca2+ release channels, and the ABC transporters. The second messengergated channels are curious in that they are operated by intracellular ligands such as Ca2+, cyclic nucleotides, and perhaps inositol phosphates but are structurally similar to voltage-gated channels. Interestingly, the ligand-gated channel family may have evolved relatively recently, since there are as yet no direct precursors identified for this family in lower eukaryotes. In contrast, all of the other families are represented in even the simplest eukaryotes. The voltage/second messenger-gated channel family shares a common overall topology of four repeated membranespannig domains that associate to form the ion conduction pathway (Fig. 1). Each domain consists of six hydrophobic membrane-spanning segments (S1-S6) and a membrane-inserted P segment in the S5-S6 linker that has been shown to be part of the pore-forming region and that contains the selectivity filter for ions. The hydrophilic N and C termini contain the inactivation gate and ligand-binding sites,