insight review articles 186 NATURE | VOL 413 | 13 SEPTEMBER 2001 | www.nature.com P hototransduction, the process by which light energy is converted into a photoreceptor’s electrical response, has long been at the forefront of studies, not only of sensory transduction, but also cell signalling more generally. Pioneering studies in the 1970s and 80s unravelled the biochemical steps of excitation in vertebrate rods and, together with seminal studies of hormone-stimulated adenylate cyclase, led to the discovery and characterization of G-protein signalling 1 . These cascades, whereby heptahelical transmembrane receptors such as rhodopsin catalytically activate heterotrimeric G proteins, are widely found not only in many sensory receptors (see review in this issue by Firestein, pages 211–218), but also throughout the body, where they respond to all manner of chemical messengers, such as hormones, neurotransmitters, odorants and tastants. Photoreceptor performance One hallmark of such cascades is their capacity for amplifi- cation. Early psychophysical experiments indicating that photoreceptors were capable of responding to single photons 2 were confirmed, first in invertebrates, and later in vertebrate rods, by electrophysiological recordings showing that quantized events (quantum bumps) could be recorded in response to absorption of single photons of light 3,4 (Fig. 1). Other functional attributes shared by vertebrate and invertebrate photoreceptors include low ‘dark noise’ (spon- taneous thermal isomerizations of rhodopsin, which sets the ultimate limit on absolute sensitivity 5 ); efficient mecha- nisms for response termination; the coding of intensity by graded potentials; and the ability to light adapt — that is, to reduce amplification as background intensity increases. But there are also differences that hint at a dichotomy in the underlying molecular machinery. First, vertebrate photore- ceptors hyperpolarize, because the transduction channels close in response to light, whereas in most invertebrates the channels open, leading to depolarization. Second, in rods, the trade-off between amplification and response speed limits human temporal resolution to ~10 Hz under dim conditions. But fly photoreceptors possess the fastest known G-protein-signalling pathways, responding around 10 times more quickly than mammalian rods and ~100 times faster than toad rods recorded at similar temperatures (Fig. 1). Third, rods have only a limited ability to adapt, rapidly saturating as intensity increases; only the less sensitive cones can respond under daylight intensities. By contrast, despite their exquisite sensitivity to single photons, fly photoreceptors successfully light adapt over the entire environmental range, up to ~10 6 effectively absorbed photons per second (Fig. 1) 6–8 . The phototransduction cascade in vertebrate rods is understood in unparalleled detail 9,10 , and widely cited as the textbook example of G-protein signalling, but the molecu- lar strategies underlying invertebrate phototransduction are still being deduced. We focus here on recent studies in the fruitfly Drosophila, highlighting the similarities and differences with the well-established scheme in vertebrates. Key to our understanding is Drosophila’s unique genetic potential, which has been exploited to identify the elements of the cascade, while a powerful mix of molecular genetic and physiological analysis is providing insight into the molecular choreography by which these photoreceptors achieve their exceptional performance. Photoreceptor ultrastructure Vertebrate and invertebrate photoreceptors both sequester their transduction machinery in specialized subcellular compartments (Fig. 2). Their structure is dictated in the first instance by the need to maximize the amount of light-absorbing membrane. Vertebrate rods achieve this with stacks of membranous discs internalized in the rod outer segment, which is separated from the rest of the cell by a short ciliary stalk. By contrast, invertebrate photoreceptors have tightly packed microvilli, which together form a cylindrical rhabdomere (from the Greek rod). Like its vertebrate counterpart, this acts as a light- or waveguide, trapping axially directed light, and at the same time contains most of the molecules of the transduction cascade. Converging studies indicate that individual microvilli, each only 60 nm in diameter, may be semiautonomous units of excitation and adaptation 7,11,12 . Together with their molecular organization, this miniatur- ization may be the key to understanding the amplification, rapid kinetics and adaptational capacity of these remarkable receptors. The visual cycle Phototransduction begins with the absorption of light by rhodopsin, triggering the 11-cis to all-trans photoisomer- ization of the chromophore (retinal or 2-dehydro-retinal in vertebrates, 3-hydroxy-retinal in flies 13 ) and formation of the activated metarhodopsin state. In vertebrates, all-trans retinal subsequently dissociates and must be re-isomerized through a lengthy and time-consuming enzymatic pathway that dictates the time course of dark adaptation following bleaching illumination (~30 min for rods). Invertebrate Visual transduction in Drosophila Roger C. Hardie & Padinjat Raghu Department of Anatomy, University of Cambridge, Downing Street, Cambridge CB2 3DY, UK (e-mail: rch14@hermes.cam.ac.uk) The brain’s capacity to analyse and interpret information is limited ultimately by the input it receives. This sets a premium on information capacity of sensory receptors, which can be maximized by optimizing sensitivity, speed and reliability of response. Nowhere is selection pressure for information capacity stronger than in the visual system, where speed and sensitivity can mean the difference between life and death. Phototransduction in flies represents the fastest G-protein-signalling cascade known. Analysis in Drosophila has revealed many of the underlying molecular strategies, leading to the discovery and characterization of signalling molecules of widespread importance. © 2001 Macmillan Magazines Ltd