Drug development Receptive receptors LIBRARY PHOTO SCIENCE MEDI-MATION/ One route to developing new drugs is to look at targeting the hundreds of G-protein-coupled receptors, found in cell membranes, that are not currently exploited clinically. Clare Sansom investigates

Nearly every medicine you take, be of around 60 per cent of drugs in it a painkiller, an asthma inhaler In short current clinical use. or an antihistamine, targets the  Cell-surface receptors, One group of receptor proteins protein products of the human such as G-protein-coupled in particular has risen to a position genome. Of course there are some receptors (GPCRs), are of pre-eminent importance to the exceptions to this rule, antibiotics the targets of over half the pharmaceutical industry – the and antivirals target the infecting drugs currently used G-protein-coupled receptors parasite, and some experimental  Developing medicines (GPCRs). These are membrane- drugs are now targeting the to target currently bound proteins that act as the cell’s machinery that creates those undrugged GPCRs may communicators. Richard Henderson proteins. help fulfil previously of the Medical Research Council Perhaps more surprisingly, 10 unmet medical needs (MRC) Laboratory of Molecular years into the human genome era  Recent developments Biology in Cambridge, UK, an the portion of the genome that is allowing the detailed 3D experienced researcher in membrane ‘pharmaceutically interesting’ structures of GPCRs to be protein structure, describes them as seems remarkably unchanged. For determined should push ‘signalling molecules that control the several decades, five groupings of this area forward whole of physiology’. proteins have been recognised as Exact statistics are difficult to comprising the largest proportion determine, but around half the of drug targets: hydrolases profits of the pharmaceutical (including proteases), kinases, industry derive from drugs that nuclear hormone receptors, target GPCRs. And the list of drugs ion channels and cell-surface that target these receptors includes receptors. Most important of all some of the best-known and most are undoubtedly the cell-surface GPCR in a lipid bilayer widely used drugs of recent years, receptors, which are the targets plasma membrane including triptans for migraine, 52 | Chemistry World | August 2010 www.chemistryworld.org beta blockers for hypertension, and Tagamet (cimetidine) for stomach ulcers. Other conditions that can be controlled by drugs targeting these receptors include psychiatric disorders, Parkinson’s disease and some forms of cancer. Yet until just a few years ago all the productive work in developing drugs to target this protein family had to be done without one of the most important weapons in the drug development armoury: detailed 3D structures of the proteins being targeted. Almost 800 genes in the human genome code for GPCRs, although slightly over half of these are the pharmaceutically uninteresting olfactory receptors that enable us to smell. Fiona Marshall, chief scientific officer of specialist drug discovery company Heptares Therapeutics, based in Welwyn Garden City, UK, explains that the industry has done little more than scratch the surface of the others. ‘Fifty six of the 370 non-olfactory GPCRs in the human genome are already targets of drugs in clinical use. However, several times that number are of interest to the pharma industry but have still not been “drugged”. There are candidate drugs targeted to some further GPCRs in clinical trials, but many of these are large, lipophilic molecules that traditionally have a high failure rate,’ she says. Scientists like Marshall and Henderson – who have devoted many decades between them to elucidating the structures and mechanisms of these proteins – believe that knowing what the GPCRs look like will help develop drugs to target the 300-odd currently undrugged GPCRs. In doing so, they hope to be able to fulfil previously unmet medical needs.

A clearer picture There are two main reasons why GPCR structures have proved so difficult to determine. The first is that they are not simple to crystallise – they are membrane-bound, and cannot be purified and crystallised without further steps. Instead, membrane proteins need to be solubilised by detergents before being purified and persuaded – with difficulty – to form crystals. While the first crystal structure of a soluble protein (myoglobin) was published in 1958, it took a further 27 years before researchers managed to crystallise and solve the structure of a membrane protein. The protein was the bacterial photosynthetic www.chemistryworld.org Chemistry World | August 2010 | 53 Drug development

reaction centre and in 1988, has, even among membrane proteins, Hartmut Michel, Johann proved particularly difficult. As Deisenhofer and Robert White explains, ‘GPCRs have to Huber were awarded the be able to flip easily from one Nobel prize in chemistry for conformational state to another determining its structure. –“off” to “on” – when a ligand The recognition of their binds. They are therefore work came over a quarter of a less stable than most trans- century after the same prize was membrane helical bundles, LAGUNA DESIGN / SCIENCE PHOTO LIBRARY LIBRARY PHOTO SCIENCE / DESIGN LAGUNA awarded to Max Perutz and John and this makes them almost Kendrew for their investigations impossible to crystallise with into the structures of the haem- detergent in their native form.’ containing globin proteins. Stephen White of the University Leading the pack of California at Irvine, US, During the last few years, the maintains a database containing the technical breakthroughs in this structures of membrane proteins field have been pioneered primarily and often compares its growth with by three groups – Chris Tate’s at that of the comprehensive protein the MRC Laboratory of Molecular structural database, the Protein Biology and two based in California, Data Bank (PDB). ‘The first version US, led respectively by Ray Stevens of our database was published in of the Scripps Research Institute and 1998, and contained fewer than a Brian Kobilka of Stanford University. dozen proteins,’ he says. ‘It now finally solved the structure of bovine The structure of The work of these teams has led has rather over 600 structures, rhodopsin to 2.8Å resolution, the rhodopsin was finally to a step change in the prospects equivalent to the PDB in 1991, and floodgates were opened. This solved in 2000 for GPCR structure determination. these represent only 241 distinctly was not, immediately, for further ‘Since the first DNA and protein different proteins.’ experimental structures, but for structures were solved in the 1950s, The first low resolution structure of structure prediction – ‘homology’ structural biology has progressed a transmembrane protein with seven or ‘comparative’ modelling – using through a marriage between biology helices was obtained in 1975. This a more reliable model than the and technical development,’ says was the prokaryotic light-sensitive unrelated bacteriorhodopsin. Stevens. proton pump, bacteriorhodopsin: The techniques and methodologies an analogue of the GPCR family Unraveling structures developed by the groups are very that captures light energy and uses The rhodopsin structure showed different. Stevens’ and Kobilka’s it to move protons across the cell that GPCRs consist of a bundle of work creates an extra hydrophilic membrane out of the cell. seven membrane-bound α-helices, domain by fusing a small protein During the next 15 years, many with a ligand binding site in the into a long loop between two of the developments in molecular centre of the bundle and a binding helices in the bundle. While Tate uses biology and x-ray crystallography site for a GTPase (the ‘G-protein’) site-directed mutagenesis, altering techniques laid the foundations on the side that binds to the cell each residue in turn, measuring the that have made today’s detailed membrane. When a ligand thermostability of each mutant and images possible. binds to the receptor it then combining the ‘best’ mutations The first ‘true’ GPCR to yield to induces the protein to to make an optimally stable receptor structural studies was the functional change conformation into locked in a pharmacologically equivalent of bacteriorhodopsin: the the active state, which defined conformation. All the groups light-sensitive protein rhodopsin. leads to the exchange of have benefited from advances in These proteins share a ligand, guanosine diphosphate microbeam x-ray diffraction retinal, and a mechanism in which (GDP) for guanosine that allow data collection the cis–trans isomerisation of triphosphate (GTP) from tiny crystals. ‘A lot retinal induces a conformational on the G-protein of membrane protein change in the protein that couples and the release of crystals are thin, flat to the G-proteins in rhodopsin the GTP-bound plates that are and to proton pumping in G-protein from difficult to work bacteriorhodopsin. the GPCR. This with,’ says White. A Russian group was the first then triggers a Kobilka and to obtain crystals of rhodopsin. chemical response in Stevens focused first ‘I remember a young PhD an effector molecule on the structure of the student showing me some pink inside the cell. GPCRs b2 adrenergic receptor, needles when I was attending a have evolved to bind to the best characterised FEBS [Federation of European a wide variety of natural GPCR after rhodopsin and Biochemical Societies] meeting in ligands and interact with the target of the well-known Moscow in 1983,’ says Henderson. many different effectors, and asthma drug Ventolin (salbutamol). ‘These were rhodopsin crystals.’ In so can give rise to an enormous Kobilka, who identified the gene for 2000, when Krzysztof Palczewski range of physiological responses. this receptor in the late 1980s as a of the University of Washington The fact that each GPCR can exist The β2 adrenergic postdoc in Robert Lefkowitz’ lab at in Seattle, US, Tetsuji Okada and in at least two conformations, ‘active’ receptor is the target Duke University, US, provided the Masashi Miyano from the Riken and ‘inactive’ is another reason why of the asthma drug biochemical insights and expertise

SCRIPPS RESEARCH INSTITUTE RESEARCH SCRIPPS institute in Japan, and colleagues determining the structure of GPCRs salbutamol in protein engineering, with Stevens, 54 | Chemistry World | August 2010 www.chemistryworld.org who has also worked on GPCR receptor subtypes had developed to gain structural structure determination for over for conditions data on GPCR targets and design HEPTARES 20 years, providing the necessary including drugs that target them. Malcolm technological know-how. inflammatory Weir, Heptares’s chief executive, The first structures of the bowel explains his long association with b2 adrenergic receptor GPCR disease.’ the project. ‘I was head of structural were published in 2007 after Receptos’ biology at GlaxoSmithKline in the the researchers had managed to first drug early 1990s when that company ‘lock’ the protein in an inactive candidate, an S1P1 first sponsored the Cambridge conformation. This involved binding allosteric receptor group to work on GPCR structures. a small protein – either a lysozyme agonist, is about to At that time it seemed an or an immunoglobulin domain enter Phase I trials for intractable problem, which makes – between the two helices, and a multiple sclerosis and our recent progress seem even partial inverse agonist to the ligand modulates the activity more remarkable.’ binding site. Stevens’ group has also of the GPCR by binding The company has been able to published structures of the human to a site adjacent to the raise £21 million in venture capital b2 adrenergic receptor bound to a active site. ‘Knowing funding even in the depths of number of different ligands and, in the structures recession. Its intellectual property 2008, the A2A adenosine receptor. of different is based on the patented ‘stabilised They have also determined structures receptor receptor’ (Star) technology and they for other medically important human subtypes have now stabilised about 20 Stars receptors including the chemokine allows us covering 12 different receptors in CXCR4 receptor, a proposed target to design specific active and inactive conformations. for anti-HIV drugs, and the D3 sub- allosteric agonists and In fact, White, with his wide class of dopamine receptors. antagonists that bind to sites other knowledge of the ‘structural than the natural ligand binding site. The β1 adrenergic universe’ of membrane proteins, Commercially viable These sites differ between receptor receptor is an important believes that scores of different Two spin out companies have been sub-types more than the ligand drug target for GPCR structures are currently being fomed to use the structural know- binding site does, and targeting hypertension worked on in industry. how they have gained to develop new them can help reduce side-effects,’ drugs that target GPCRs – San Diego, adds Stevens. Looking to the future US-based Receptos was started by Tate, Gebhard Schertler and their It may be useful to ask how these Stevens and Heptares was spun out colleagues solved the structure of the developments will affect GPCR from Tate’s group. b-1 adrenergic receptor, an important based drug discovery in the long Receptos was principally set up drug target for hypertension, using term: what will the industry be to solve the human structures of, site-directed mutagenesis to stabilise able to do with structures that it and design ligands for, a GPCR sub- it in an inactive conformation. couldn’t before? Weir explains family that binds the phospholipid ‘We were able to create a mutant that atomic coordinates are not sphingosine-1-phosphate (S1P). adrenoceptor that was 400 times the only useful output from their ‘Targeting receptors in this family as thermostable as the wild type, by new technological advances. can meet previously unmet combining just six thermostabilising ‘Having reliable techniques for medical needs,’ says Stevens. ‘Our mutations,’ explains Tate. obtaining 3D structures of GPCRs primary target is the S1P1 receptor In 2007, when they had obtained is a major development, but this sub-type, an important target for diffracting crystals, Tate’s team The asthma drug, is not all we can do with our multiple sclerosis, but we are also created Heptares to use the salbutamol, targets the stabilised receptors. We can look interested in targeting other S1P stabilisation technology they β2 adrenergic receptor at ligand binding directly using surface plasmon resonance and mutagenesis; measure binding kinetics directly; and use all this information to target drugs more precisely to allosteric binding sites. This should enable us to design more drugs to a wider variety of receptors, and make them more specific to receptor subtypes and so likely to have fewer side effects.’ Weir and Stevens believe that there is room enough in the field for both their companies and more. ‘The GPCR family is extremely large and there is enough space in this structure arena for both our companies. We are competitors, yes, but we are collegial, respectful and friendly competitors,’ says Stevens.

Clare Sansom is a freelance science

ALAMY writer based in Cambridge, UK

www.chemistryworld.org Chemistry World | August 2010 | 55