GPCRs: an update on structural approaches to drug discovery

G-protein-coupled receptors (GPCRs) are a major opportunity for drug discovery in the post-genomic era. There are thought to be more than 500 therapeutically relevant GPCRs out of a total of over 700 identified to date, although only one, rhodopsin, has been the subject of a full 3D X-ray crystallography study. Two structurally related proteins, bacteriorhodopsin and sensory rhodopsin, which are not GPCRs but are part of the seven-helix membrane receptor family, have also been the subject of X-ray crystallographic studies and have been used in GPCR modeling studies. The significant differences between these rhodopsin structures, the relatively low sequence homology between individual GPCRs, and some difficulties in rationalizing point-mutation data suggests that homology-based molecular modeling alone will not provide the accurate structural information on individual receptors required for ligand design and in silico screening. In the absence of such structural information, several approaches can be used to assist in the discovery of ligands.     

Relevance of rhodopsin studies for GPCR activation

Rhodopsin, the dim-light photoreceptor present in the rod cells of the retina, is both a retinal-binding protein and a G protein-coupled receptor (GPCR). Due to this conjunction, it benefits from an arsenal of spectroscopy techniques that can be used for its characterization, while being a model system for the important family of Class A (also referred to as “rhodopsin-like”) GPCRs. For instance, rhodopsin has been a crucial player in the field of GPCR structural biology. Until 2007, it was the only GPCR for which a high-resolution crystal structure was available, so all structure–activity analyses on GPCRs, from structure-based drug discovery to studies of structural changes upon activation, were based on rhodopsin. At present, about a third of currently available GPCR structures are still from rhodopsin. In this review, I show some examples of how these structures can still be used to gain insight into general aspects of GPCR activation. First, the analysis of the third intracellular loop in rhodopsin structures allows us to gain an understanding of the structural and dynamic properties of this region, which is absent (due to protein engineering or poor electron density) in most of the currently available GPCR structures. Second, a detailed analysis of the structure of the transmembrane domains in inactive, intermediate and active rhodopsin structures allows us to detect early conformational changes in the process of ligand-induced GPCR activation. Finally, the analysis of a conserved ligand-activated transmission switch in the transmembrane bundle of GPCRs in the context of the rhodopsin activation cycle, allows us to suggest that the structures of many of the currently available agonist-bound GPCRs may correspond to intermediate active states. While the focus in GPCR structural biology is inevitably moving away from rhodopsin, in other aspects rhodopsin is still at the forefront. For instance, the first studies of the structural basis of disease mutants in GPCRs, or the most detailed analysis of cellular GPCR signal transduction networks using a systems biology approach, have been carried out in rhodopsin. Finally, due again to its unique properties among GPCRs, rhodopsin will likely play an important role in the application of X-ray free electron laser crystallography to time-resolved structural biology in membrane proteins. Rhodopsin, thus, still remains relevant as a model system to study the molecular mechanisms of GPCR activation. This article is part of a Special Issue entitled: Retinal Proteins—You can teach an old dog new tricks.     

Structure of rhodopsin and the metarhodopsin I photointermediate

The structure of the visual pigment rhodopsin in the dark state was first investigated by electron microscopy (EM). More recently, rhodopsin has been crystallised in two different space groups--a tetragonal P4(1) crystal form and a trigonal P3(1) packing arrangement. The structures of the pigment, determined by X-ray crystallography from these two crystal forms, show many similarities, but also significant differences. These differences are most extensive in the G-protein-binding region of the cytoplasmic surface, where the location of the loop between helices 5 and 6 is highly variable. A combination of EM and spin labelling suggests that this loop adopts the native conformation in the P3(1) crystal form. The X-ray structures also show the location of structural water molecules that are important for colour tuning, stabilisation of the ground state and receptor activation, and act as a template for modelling other G-protein-coupled receptors. A major current focus of structural work on rhodopsin is investigation of the activated state of the receptor. After careful spectroscopic characterisation of light activation in two-dimensional crystals, a map of the metarhodopsin I intermediate was obtained by EM from two-dimensional crystals. In addition, NMR studies are providing information about the structure of activated states of rhodopsin. In the future, structural information will show how rhodopsin becomes activated and how it couples to downstream signalling pathways.     

Molecular simulations and solid-state NMR investigate dynamical structure in rhodopsin activation

Rhodopsin has served as the primary model for studying G protein-coupled receptors (GPCRs)-the largest group in the human genome, and consequently a primary target for pharmaceutical development. Understanding the functions and activation mechanisms of GPCRs has proven to be extraordinarily difficult, as they are part of a complex signaling cascade and reside within the cell membrane. Although X-ray crystallography has recently solved several GPCR structures that may resemble the activated conformation, the dynamics and mechanism of rhodopsin activation continue to remain elusive. Notably solid-state ((2))H NMR spectroscopy provides key information pertinent to how local dynamics of the retinal ligand change during rhodopsin activation. When combined with molecular mechanics simulations of proteolipid membranes, a new paradigm for the rhodopsin activation process emerges. Experiment and simulation both suggest that retinal isomerization initiates the rhodopsin photocascade to yield not a single activated structure, but rather an ensemble of activated conformational states. This article is part of a Special Issue entitled: Membrane protein structure and function.     

G-protein coupled receptor structure

Because of their central role in regulation of cellular function, structure/function relationships for G-protein coupled receptors (GPCR) are of vital importance, yet only recently have sufficient data been obtained to begin mapping those relationships. GPCRs regulate a wide range of cellular processes, including the senses of taste, smell, and vision, and control a myriad of intracellular signaling systems in response to external stimuli. Many diseases are linked to GPCRs. A critical need exists for structural information to inform studies on mechanism of receptor action and regulation. X-ray crystal structures of only one GPCR, in an inactive state, have been obtained to date. However considerable structural information for a variety of GPCRs has been obtained using non-crystallographic approaches. This review begins with a review of the very earliest GPCR structural information, mostly derived from rhodopsin. Because of the difficulty in crystallizing GPCRs for X-ray crystallography, the extensive published work utilizing alternative approaches to GPCR structure is reviewed, including determination of three-dimensional structure from sparse constraints. The available X-ray crystallographic analyses on bovine rhodopsin are reviewed as the only available high-resolution structures for any GPCR. Structural information available on ligand binding to several receptors is included. The limited information on excited states of receptors is also reviewed. It is concluded that while considerable basic structural information has been obtained, more data are needed to describe the molecular mechanism of activation of a GPCR.     

Integrated analysis of the conformation of a protein-linked spin label by crystallography,EPR and NMR spectroscopy

Long-range structural information derived from paramagnetic relaxation enhancement observed in the presence of a paramagnetic nitroxide radical is highly useful for structural characterization of globular, modular and intrinsically disordered proteins, as well as protein–protein and protein-DNA complexes. Here we characterized the conformation of a spin-label attached to the homodimeric protein CylR2 using a combination of X-ray crystallography, electron paramagnetic resonance (EPR) and NMR spectroscopy. Close agreement was found between the conformation of the spin label observed in the crystal structure with interspin distances measured by EPR and signal broadening in NMR spectra, suggesting that the conformation seen in the crystal structure is also preferred in solution. In contrast, conformations of the spin label observed in crystal structures of T4 lysozyme are not in agreement with the paramagnetic relaxation enhancement observed for spin-labeled CylR2 in solution. Our data demonstrate that accurate positioning of the paramagnetic center is essential for high-resolution structure determination.     

X-ray diffraction of heavy-atom labelled two-dimensional crystals of rhodopsin identifies the position of cysteine 140 in helix 3 and cysteine 316 in helix 8

We have used site-specific heavy-atom labelling and X-ray diffraction to localize single amino acid residues in the cytoplasmic domain of the integral membrane protein rhodopsin, the dim-light photoreceptor of retinal vertebrate rod cells. Two-dimensional orthorhombic crystals of the space group p22(1)2(1) (a=59.5(+/-1) A and b=82.7(+/-1.5) A) were produced from detergent-solubilized, partially delipidated rhodopsin. To obtain milligram amounts of two-dimensional crystals, which are required for X-ray diffraction, the yield of the crystalline material was significantly increased by reconstitution of rhodopsin in the presence of cholesterol (1:2 to 1:10 mol/mol) and by adding polar organic solvents to the dialysis buffer. The native cysteine residues C140 and C316 were then selectively labelled with mercury using the sulphydryl-specific reagent p-chloromercuribenzoate (1.6-2.1 mol Hg per mol rhodopsin). The labelling did not affect the unit cell dimensions. Optical absorption spectra of labelled and native two-dimensional rhodopsin crystals showed the characteristic 11-cis-retinal peak at 498 nm, which corresponds to the dark state of rhodopsin. The in-plane position of the mercury label was calculated at 9.5 A resolution from the intensity differences in the X-ray diffraction patterns of labelled and native crystals using Fourier difference methods and the phase information from electron crystallography. The label positions were in excellent agreement with the positions of C140 at the cytoplasmic end of helix 3 and of C316 in the cytoplasmic helix 8 recently obtained from three-dimensional rhodopsin crystals. Whereas these high-resolution diffraction studies were performed under cryogenic conditions (100 K), our results were obtained at room temperature with fully hydrated membranes and in the absence of loop-loop crystal contacts. To study the structural changes of the cytoplasmic loops involved in activation and signal transduction, our more physiological conditions offer important advantages. Furthermore, the localization of C316 is the first direct proof that the electron density on top of helix 1 observed by cryo-electron microscopy is a part of the C-terminal loop. Our approach is of particular interest for investigations of other membrane proteins, for which 3D crystals are not available. Structural constraints from heavy-atom labels at strategic sites enable the assignment of a position in the amino acid sequence to features visible in a low-resolution density map and the study of conformational changes associated with different functional states of the membrane protein.     

The synthesis and high-level expression of a beta2-adrenergic receptor gene in a tetracycline-inducible stable mammalian cell line

High-level expression of G-protein-coupled receptors (GPCRs) in functional form is required for structure-function studies. The main goal of the present work was to improve expression levels of beta2-adrenergic receptor (beta2-AR) so that biophysical studies involving EPR, NMR, and crystallography can be pursued. Toward this objective, the total synthesis of a codon-optimized hamster beta2-AR gene suitable for high-level expression in mammalian systems has been accomplished. Transient expression of the gene in COS-1 cells resulted in 18 +/- 3 pmol beta2-AR/mg of membrane protein, as measured by saturation binding assay using the beta2-AR antagonist [3H] dihydroalprenolol. Previously, we reported the development of an HEK293S tetracycline-inducible system for high-level expression of rhodopsin. Here, we describe construction of beta2-AR stable cell lines using the HEK293S-TetR-inducible system, which, after induction, express wild-type beta2-AR at levels of 220 +/- 40 pmol/mg of membrane protein corresponding to 50 +/- 8 microg/15-cm plate. This level of expression is the highest reported so far for any wild-type GPCR, other than rhodopsin. The yield of functional receptor using the single-step affinity purification is 12 +/- 3 microg/15-cm plate. This level of expression now makes it feasible to pursue structure-function studies using EPR. Furthermore, scale-up of beta2-AR expression using suspension cultures in a bioreactor should now allow production of enough beta2-AR for the application of biophysical techniques such as NMR spectroscopy and crystallography.     

Use of detergents in two-dimensional crystallization of membrane proteins

Structure determination at high resolution is actually a difficult challenge for membrane proteins and the number of membrane proteins that have been crystallized is still small and far behind that of soluble proteins. Because of their amphiphilic character, membrane proteins need to be isolated, purified and crystallized in detergent solutions. This makes it difficult to grow the well-ordered three-dimensional crystals that are required for high resolution structure analysis by X-ray crystallography. In this difficult context, growing crystals confined to two dimensions (2D crystals) and their structural analysis by electron crystallography has opened a new way to solve the structure of membrane proteins. However, 2D crystallization is one of the major bottlenecks in the structural studies of membrane proteins. Advances in our understanding of the interaction between proteins, lipids and detergents as well as development and improvement of new strategies will facilitate the success rate of 2D crystallization. This review deals with the various available strategies for obtaining 2D crystals from detergent-solubilized intrinsic membrane proteins. It gives an overview of the methods that have been applied and gives details and suggestions of the physical processes leading to the formation of the ordered arrays which may be of help for getting more proteins crystallized in a form suitable for high resolution structural analysis by electron crystallography.     

Investigations of Photosensitive Proteins by Serial Crystallography

This review contains recent data on serial femtosecond X-ray crystallography (SFX), based on a femtosecond X-ray free electron laser, as well as on the possibilities of its application for studying photosensitive proteins. Development of this method began rather recently, and it has already shown its effectiveness and some unique advantages over conventional X-ray structural analysis. This technology is especially promising for structural studies of membrane proteins and for kinetic studies. The main principle of the method, the possibility of its application in structural biology, its advantages and disadvantages, as well as its prospects for further development are analyzed in this review. Special attention is given to publications in which the SFX method has been used to study photosensitive proteins.     

Methyl-accepting protein associated with bacterial sensory rhodopsin I.

In vivo radiolabeling of Halobacterium halobium phototaxis mutants and revertants with L-[methyl-3H] methionine implicated seven methyl-accepting protein bands with apparent molecular masses from 65 to 150 kilodaltons (kDa) in adaptation of the organism to chemo and photo stimuli, and one of these (94 kDa) was specifically implicated in phototaxis. The lability of the radiolabeled bands to mild base treatment indicated that the methyl linkages are carboxylmethylesters, as is the case in the eubacterial chemotaxis receptor-transducers. The 94-kDa protein was present in increased amounts in an overproducer of the apoprotein of sensory rhodopsin I, one of two retinal-containing phototaxis receptors in H. halobium. It was absent in a strain that contained sensory rhodopsin II and that lacked sensory rhodopsin I and was also absent in a mutant that lacked both photoreceptors. Based on the role of methyl-accepting proteins in chemotaxis in other bacteria, we suggest that the 94-kDa protein is the signal transducer for sensory rhodopsin I. By [3H]retinal labeling studies, we previously identified a 25-kDa retinal-binding polypeptide that was derived from photochemically reactive sensory rhodopsin I. When H. halobium membranes containing sensory rhodopsin I were treated by a procedure that stably reduced [3H]retinal onto the 25-kDa apoprotein, a 94-kDa protein was also found to be radiolabeled. Protease digestion confirmed that the 94-kDa retinal-labeled protein was the same as the methyl-accepting protein that was suggested above to be the signal transducer for sensory rhodopsin I. Possible models are that the 25- and 94-kDa proteins are tightly interacting components of the photosensory signaling machinery or that both are forms of sensory rhodopsin I.     

Atomic models by cryo-EM and site-directed spin labeling: application to the N-terminal region of Hsp16.5

We report an approach for determining the structure of macromolecular assemblies by the combined application of cryo-electron microscopy (cryo-EM) and site-directed spin labeling electron paramagnetic resonance spectroscopy (EPR). This approach is illustrated for Hsp16.5, a small heat shock protein that prevents the aggregation of nonnative proteins. The structure of Hsp16.5 has been previously studied by both cryo-EM and X-ray crystallography. The crystal structure revealed a roughly spherical protein shell with dodecameric symmetry; however, residues 1-32 were found to be disordered. The cryo-EM reconstruction at 13 A resolution appeared similar to the crystal structure but with additional internal density corresponding to the N-terminal regions of the 24 subunits. In this study, a systematic application of site-directed spin labeling and EPR spectroscopy was carried out. By combining the EPR constraints from spin label accessibilities and proximities with the cryo-EM density, we obtained an atomic model for a portion of the Hsp16.5 N-terminal region in the context of the oligomeric complex.     

Structural characterization of proton-pumping rhodopsin lacking a cytoplasmic proton donor residue by X-ray crystallography

DTG/DTS rhodopsin, which was named based on a three-residue motif (DTG or DTS) that is important for its function, is a light-driven proton-pumping microbial rhodopsin using a retinal chromophore. In contrast to other light-driven ion-pumping rhodopsins, DTG/DTS rhodopsin does not have a cytoplasmic proton donor residue, such as Asp, Glu, or Lys. Because of the lack of cytoplasmic proton donor residue, proton directly binds to the retinal chromophore from the cytoplasmic solvent. However, mutational experiments that showed the complicated effects of mutations were not able to clarify the roles played by each residue, and the detail of proton uptake pathway is unclear because of the lack of structural information. To understand the proton transport mechanism of DTG/DTS rhodopsin, here we report the three-dimensional structure of one of the DTG/DTS rhodopsins, PspR from Pseudomonas putida, by X-ray crystallography. We show that the structure of the cytoplasmic side of the protein is significantly different from that of bacteriorhodopsin, the best-characterized proton-pumping rhodopsin, and large cytoplasmic cavities were observed. We propose that these hydrophilic cytoplasmic cavities enable direct proton uptake from the cytoplasmic solvent without the need for a specialized cytoplasmic donor residue. The introduction of carboxylic residues homologous to the cytoplasmic donors in other proton-pumping rhodopsins resulted in higher pumping activity with less pH dependence, suggesting that DTG/DTS rhodopsins are advantageous for producing energy and avoiding intracellular alkalization in soil and plant-associated bacteria.     

Structure and dynamics of dark-state bovine rhodopsin revealed by chemical cross-linking and high-resolution mass spectrometry

Recent work using chemical cross-linking to define interresidue distance constraints in proteins has shown that these constraints are useful for testing tertiary structural models. We applied this approach to the G-protein-coupled receptor bovine rhodopsin in its native membrane using lysine- and cysteine-targeted bifunctional cross-linking reagents. Cross-linked proteolytic peptides of rhodopsin were identified by combined liquid chromatography and FT-ICR mass spectrometry with automated data-reduction and assignment software. Tandem mass spectrometry was used to verify cross-link assignments and locate the exact sites of cross-link attachment. Cross-links were observed to form between 10 pairs of residues in dark-state rhodopsin. For each pair, cross-linkers with a range of linker lengths were tested to determine an experimental distance-of-closest-approach (DCA) between reactive side-chain atoms. In all, 28 cross-links were identified using seven different cross-linking reagents. Molecular mechanics procedures were applied to published crystal structure data to calculate energetically achievable theoretical DCAs between reactive atoms without altering the position of the protein backbone. Experimentally measured DCAs are generally in good agreement with the theoretical DCAs. However, a cross-link between C316 and K325 in the C-terminal region cannot be rationalized by DCA simulations and suggests that backbone reorientation relative to the crystal coordinates occurs on the timescale of cross-linking reactions. Biochemical and spectroscopic data from other studies have found that the C-terminal region is highly mobile in solution and not fully represented by X-ray crystallography data. Our results show that chemical cross-linking can provide reliable three-dimensional structural information and insight into local conformational dynamics in a membrane protein.     

NMR studies on fully hydrated membrane proteins, with emphasis on bacteriorhodopsin as a typical and prototype membrane protein

The 3D structures or dynamic feature of fully hydrated membrane proteins are very important at ambient temperature, in relation to understanding their biological activities, although their data, especially from the flexible portions such as surface regions, are unavailable from X-ray diffraction or cryoelectron microscope at low temperature. In contrast, high-resolution solid-state NMR spectroscopy has proved to be a very convenient alternative means to be able to reveal their dynamic structures. To clarify this problem, we describe here how we are able to reveal such structures and dynamic features, based on intrinsic probes from high-resolution solid-state NMR studies on bacteriorhodopsin (bR) as a typical membrane protein in 2D crystal, regenerated preparation in lipid bilayer and detergents. It turned out that their dynamic features are substantially altered upon their environments where bR is present. We further review NMR applications to study structure and dynamics of a variety of membrane proteins, including sensory rhodopsin, rhodopsin, photoreaction centers, diacylglycerol kinases, etc.     

NMR studies on fully hydrated membrane proteins, with emphasis on bacteriorhodopsin as a typical and prototype membrane protein

The 3D structures or dynamic feature of fully hydrated membrane proteins are very important at ambient temperature, in relation to understanding their biological activities, although their data, especially from the flexible portions such as surface regions, are unavailable from X-ray diffraction or cryoelectron microscope at low temperature. In contrast, high-resolution solid-state NMR spectroscopy has proved to be a very convenient alternative means to be able to reveal their dynamic structures. To clarify this problem, we describe here how we are able to reveal such structures and dynamic features, based on intrinsic probes from high-resolution solid-state NMR studies on bacteriorhodopsin (bR) as a typical membrane protein in 2D crystal, regenerated preparation in lipid bilayer and detergents. It turned out that their dynamic features are substantially altered upon their environments where bR is present. We further review NMR applications to study structure and dynamics of a variety of membrane proteins, including sensory rhodopsin, rhodopsin, photoreaction centers, diacylglycerol kinases, etc.     

Structural basis for sensory rhodopsin function

The crystal structure of sensory rhodopsin II from Natronobacterium pharaonis was recently solved at 2.1 A resolution from lipidic cubic phase-grown crystals. A critical analysis of previous structure-function studies is possible within the framework of the high-resolution structure of this photoreceptor. Based on the structure, a molecular understanding emerges of the efficiency and selectivity of the photoisomerization reaction, of the interaction of the sensory receptor and its cognate transducer protein HtrII, and of the mechanism of spectral tuning in photoreceptors. The architecture of the retinal binding pocket is compact, representing a major determinant for the selective binding of the chromophore, all-trans retinal to the apoprotein, opsin. Several chromophore-protein interactions revealed by the structure were not predicted by previous mutagenesis and spectroscopic analyses. The structure suggests likely mechanisms by which photoisomerization triggers the activation of sensory rhodopsin II, and highlights the possibility of a unified mechanism of signaling mediated by sensory receptors, including visual rhodopsins. Future investigations using time-resolved crystallography, structural dynamics, and computational studies will provide the basis to unveil the molecular mechanisms of sensory receptors-mediated transmembrane signaling.     

Oxidative DNA cleavage induced by an iron(III) flavonoid complex: synthesis, crystal structure and characterization of chlorobis(flavonolato)(methanol) iron(III) complex

A flavonol iron(III) complex, [Fe(flavonolato)(2)Cl(MeOH)], has been prepared. The compound has been characterized by X-ray crystallography, spectroscopy, magnetism and electronic paramagnetic resonance (EPR) at X- and Q-band. The geometrical environment around the metal is best described as rhombic distorted octahedral. This distortion has also been inferred from the magnetic measurements and from the EPR spectra at different temperatures, E/D(rhombicity parameter) approximately 0.06. The DNA cleavage activity of the iron(III) complex with and without ascorbate/hydrogen peroxide is reported. Mechanisms of the oxidative cleavage have been proposed when DNA strand scission is performed both with and without ascorbate/hydrogen peroxide activation.     

Photoactive yellow protein, bacteriophytochrome, and sensory rhodopsin in purple phototrophic bacteria.

The purple photosynthetic bacteria contain a large variety of sensory and regulatory proteins, and those responding to light are among the most interesting. These currently include bacteriophytochrome (Bph), sensory rhodopsin (SR), and photoactive yellow protein (PYP), which all appear to function as light sensors. We herein interpret new findings within the context of current knowledge. For greater detail, the reader is referred to comprehensive reviews on these topics. Of the three proteins, only PYP has been well-characterized in terms of structure and physical-chemical properties in the purple bacteria, although none have well-defined functions. New findings include a cluster of six genes in the Thermochromatium tepidum genome that encodes presumed sensory rhodopsin and phototaxis proteins. T. tepidum also has a gene for PYP fused to bacteriophytochrome and diguanylate cyclase domains. The genes for PYP and its biosynthetic enzymes are associated with those for gas vesicle formation in Rhodobacter species, suggesting that one function of PYP is to regulate cell buoyancy. The association of bacteriophytochrome genes with those for reaction centers and light-harvesting proteins in Rhodopseudomonas palustris suggests that the photosynthetic antenna as well as the reaction center are regulated by Bphs. Furthermore, Rc. centenum PPR is reversibly photobleached at 702 nm rather than red-shifted as in other phytochromes, suggesting that PPR senses the intensity of white light rather than light quality. PYP from Halorhodospira(aka Ectothiorhodospira)halophila is of special interest because it has become the structural prototype for the PAS domain, a motif that is found throughout the phylogenetic tree and which plays important roles in many signaling pathways. Thus, the structural and photochemical characterization of PYP, utilizing site-directed mutagenesis, provides insights into the mechanism of signal transduction.     

A structure-based strategy for discovery of small ligands binding to functionally unknown proteins: combination of in silico screening and surface plasmon resonance measurements

In the postgenomic era, many researchers and organizations have been engaged in structural and functional analyses of proteins. As a part of these efforts, searching for small organic compounds that bind specifically to target proteins is quite important. In this study, we have developed a rational strategy for ligand discovery based on the three-dimensional structures of target proteins, which were elucidated by X-ray crystallography and nuclear magnetic resonance spectroscopy. The strategy has three features: (i) rapid selection of candidate compounds by in silico screening, (ii) automated preparation of sample solutions with robotics, and (iii) reliable evaluation of the candidates with surface plasmon resonance. Applying the strategy to a protein, At2g24940 from Arabidopsis thaliana, we discovered four small ligands out of a commercially available library of about 150 000 compounds. Although these compounds had only weak affinities to the target protein, with dissociation constants ranging from 68 to 120 microM, they apparently possess common structural features. They would be leads for the development of specific inhibitors/drugs for At2g24940, and provide important clues toward elucidation of the protein function.