© 1994 Nature Publishing Group http://www.nature.com/nsmb • editorial Mediating intracellular communication Being able to respond to the G-protein because of its requirement for environment is vital for all GTP. Indeed, the nature of the interaction cells, be they free-living uni­ between the rod-cell G-protein transducin cellular organisms or the and its cytoplasmic effector is explored in constituents of tissues that a paper from Margolskee and colleagues in make up multicellular plants and animals. this issue 1• Indeed, there are suggestions that the 2.5 billion years it took for multicellular organ­ G-proteins isms to appear may have been related to the There are three characterized classes of cell­ need to develop the complex signalling surface receptor proteins: those linked to mechanisms that allowed intercellular com­ ion channels, enzymes and G-proteins. The munication and the coordination of cellu­ last group of receptors, which typically have lar behaviour. A mere twenty years ago the a seven-helix transmembrane structure, is nature of these cellular information high­ by far the largest, its members being num­ ways was, for the most part, a black box. bered in the hundreds. Cellular responses Since then a vast effort has been directed to an extensive range of stimuli - light, towards shedding light on this area of bi­ smell, taste, hormones, pheromones, neu­ ology. The importance of the endeavour has rotransmitters and the like - are all medi­ now been recognized by the award of the ated through G-protein-coupled receptors. 1994 Nobel Prize for Medicine to the dis­ G-proteins are heterotrimers, made coverers of G-protein-mediated signal from a, ~ and y subunits (G ); each transduction; Martin Rodbell, recently re­ heterotrimer is coupled to a speci!lc recep­ tired from the National Institute of Envi­ tor, relaying the signals from the activated ronmental Health Sciences, North Carolina, receptor to effectors (enzymes or ion chan­ and Alfred G. Gilman, of the University of nels) that then generate one or more intra­ Texas Southwestern. cellular second messengers, such as cyclic Before the discovery of G-proteins it had nucleotide, Ca2+, inositol phosphates or been assumed that when a signal, such as a diacylglycerol. The specificity is clearly hormone, was bound by its cell-surface re­ important in maintaining the fidelity of the ceptor, the activated receptor signalled di­ signal transduction process. The a -subunit rectly to the effector in the cytoplasm. This determines the specificity of the G-protein/ view was dramatically revised in the early receptor interaction whereas interaction 1970s when Rodbell, then working at the with downstream effectors may be specified National Institute of Health in Bethesda, by G and/ or G subunits2• The various com­ demonstrated that a third component, a binations of a~yheterotrimer subunits allow transducer, passed the incoming signal for both subtle regulation and complex feed­ from receptor to effector. But it was not back control of receptors and effectors. until 1980 that Gilman and colleagues, then Cloning, expression and mutational at the University of Virginia in analysis of a number of G-proteins has ad­ Charlottesville (and independently, Tho­ vanced understanding of the intricacies of mas Pfeufer, University of Wurzburg, Ger­ the interactions between G-proteins and re­ many) managed to isolate and characterize ceptors, G-proteins and effectors, the the transducer, christening the molecule a mechanism and regulation of GTP hydroly- structural biology volume 1 number 11 november 1994 747 © 1994 Nature Publishing Group http://www.nature.com/nsmb editorial •

sis, the regulation of the membrane attach­ amino acid residues are particularly impor­ ment of G-protein components, and so on tant for PDE ac;:tivation. Three of these resi­ (for example, ref. 3 and references therein). dues are within helix a4 of the transducin Nonetheless, the problem of how this class of a-subunit and all have their side chains ex­ proteins function is far from solved. What, posed on the same, solvent exposed, face for example, is the molecular basis of the of the transducin molecule suggesting that specificity of the G-proteins? The carboxyter­ this helix defines the primary site of inter­ minus of the a-subunit of the G-protein is action between a ,-rod and PDEy. That result one of the major determinants of receptor is at odds with previous suggestions that the recognition, while the cytoplasmic loops that adjacent a4/~6 loop is the site of interac­ link the helices of the seven transmembrane tion3·5. There are potential problems associ­ receptors are the main determinants of G­ ated with using peptides to mimic the activ­ protein interaction. Sequences adjacent to the ity of much larger proteins, as discussed by C-terminus of the a-subunit have been im­ Artemyve and Hamm\ nonetheless the sol­ plicated in effector interactions. vent-exposed location of all five residues and In rod receptor cells of the vertebrate eye modelling of the various substitutions of the light responsive receptor, rhodopsin, is these residues argue against such artifacts. coupled to its effector, the enzyme cGMP­ While the authors have focused on one specific phosphodiesterase (PDE), by the particular region involved in the interac­ heterotrimeric G-protein transducin tion between G-protein and effector, the (G "). Inactive G "•GDP binds to the fact that this region is not seen to undergo ta1.,y tapy photoactivated rhoctopsin and releases a conformational change on exchange in GDP. GTP then binds to the a-subunit of the crystal structures of Gia·GTPyS (ref. 5) the receptor-bound transducin causing and G,a·GDP (ref. 6) suggests that there Gw.·GTP to separate from both the recep­ must be other sites of interaction between

tor and the ~y heterodimer. The G1a·GTP the two proteins. Clearly, the structure of complex then activates PDE by binding to the complex between Gta·GTP and PDEy that protein's inhibitoryy-subunit. Activation would help resolve this question. But the of the phosphodiesterase lowers the concen­ demands of cell biologists do not stop there: tration of cGMP, which closes cGMP-gated what of the interactions between Ga and cation channels and hyperpolarizes the reti­ G~y? how does Ga interact with receptor? nal rod cell, thereby generating the nerve im­ what are the parameters determining GTP pulse. The intrinsic GTPase activity of the a­ hydrolysis (see, for example, refs 3,5-7)? subunit hydrolyses the bound GTP to GDP, and so on. Certainly, there are no shortage inactivating Gta' releasing the inhibitory of challenges for structural biologists with PDEy-subunit which in turn switches off an interest in G-protein-mediated signal the phosphodiesterase. transduction. What is the structural basis of the Gia specificity for the PDEy subunit? I. Spickofsky, N. et al. Nature struct. Margolskee and colleagues have addressed Biol. 1, 771-780 (1994). 2. Clapham , D.E. & Neer, E.J. Nature this question by using varients of a syn­ 365, 403-406 (1993). thetic peptide from a region of G,a which 3. Conklin, B.R. & Bourne, H.R. Cell on its own is able to interact with and in­ 73, 631- 641 (1993). 4. Artemyve, N .0 . & Hamm, H.E. hibit the PDEy-subunit. By mutating indi­ Nature struct. Biol. 1, 752- 754 (1994). vidual residues in a short region of the Gia, 5. Noel, J.P., Hamm, H.E. & Sigler, and by using peptides from similar regions P.B. Nature 366, 654- 662 ( 1993). 6. Lambright, D.G., Noel, J.P., of other closely related G-protein a-sub­ Hamm, H .E. & Sigler, P.B. Nature unit, the authors are able to show that five 369, 621-628 (1994).

748 structural biology volume 1 number 11 november 1994