G-protein-coupled receptors (GPCRs) are integral membrane proteins involved in signal transduction, and constitute major drug targets for disease therapy. Sequence analysis of the human genome suggests that there are up to 1000 different GPCRs, each interacting specifically with ligands ranging from odorant molecules to neurotransmitters to hormones. However, despite their striking clinical relevance, high-resolution structural information on GPCRs is only available for the visual pigment rhodopsin (Palczewski et al., Science 289, 739-745, 2000; Okada et al., PNAS 99, 5982-5987, 2002), which can be purified in large amounts from natural tissue. Structure determination of any other GPCR requires a suitable heterologous expression system and a robust large-scale purification scheme to obtain milligram quantities of functional receptor protein for 3D crystallization studies. We focus on the three-dimensional structure determination of GPCRs by x-ray crystallography. This includes heterologous expression, large-scale purification of receptors in functional form, pharmacological characterization, receptor crystallization, and use of antibody fragments for co-crystallization experiments. We explore currently also the structure of a neuropeptide ligand bound to its GPCR. Since my arrival at the NIH in May 2001, and access to empty laboratory space in July 2001, I accomplished the following towards these goals: (1) We have established a fully functional laboratory suitable to carry out all aspects needed for structure determination of membrane proteins. (2) Automated large-scale receptor purification: Progress towards structure determination of integral membrane proteins is given by obtaining milligram quantities of functional receptors on a regular basis, i.e. every week. With Jim White (technical support), a large-scale, automated purification scheme for the GPCR neurotensin receptor was developed as a model system, using an Akta purifier machine (White & Grisshammer, unpublished results). This chromatography system allows purification at large scale of GPCRs on a weekly basis with extreme reliability, exploiting airsensors for automated loading of solubilized receptors and a fully automated two-column purification procedure (immobilized metal affinity chromatography, followed by a ligand affinity column). We can now generate every week 2-3 mg of functional receptor protein suitable for further studies. (3) Large-scale fermentation: Purification at large scale requires an adequate supply of receptor expressing cells as starting material. The GPCR for neurotensin can be expressed functionally in bacteria at levels sufficient for large-scale purification. With Dr. Joseph Shiloach and his co-workers (Biotechnology Unit NIDDK), we established conditions for fermentation at large scale resulting in receptor materials of high quality. This removes the current bottleneck to obtain enough cell paste as starting material for large-scale purification, allowing further scale-up of purification on a regular basis. (4) Identification of additional GPCRs suitable for crystallization: Although GPCRs seem to be similar, there are fundamental differences regarding expression levels and biochemical behavior, with profound consequences for successful crystal formation. We have screened 50 additional GPCRs in view of good expression in bacteria, and have identified at least one candidate suitable for further study. This is currently pursued by Christine Foster, a postdoctoral fellow in our laboratory. (5) Co-crystallization of bacterially expressed Fv antibody fragments with integral membrane proteins has successfully led to high-resolution 3D structures of a bacterial cytochrome oxidase (Iwata et al., Nature 376, 660-669, 1995; Ostermeier et al., Nat. Struct. Biol. 2, 842-846, 1995) and the yeast cytochrome bc1 complex (Hunte et al., Structure 8, 669-684, 2000). In addition, the high-resolution structure of a bacterial potassium channel in complex with a monoclonal Fab antibody fragment has been reported by MacKinnon's group (Zhou et al., Nature 414, 43-48, 2001). Co-crystallization with antibody fragments may also be beneficial for the structure determination of GPCRs, allowing better crystal contacts and constraining receptor flexibility. We have established a protocol for functional reconstitution of receptors into lipid vesicles. These proteoliposomes are currently being used to select for antibody fragments by the phage display system in collaboration with a group in Germany. This will provide the basis for obtaining suitable antibody Fv fragments for co-crystallization. (6) In collaboration with Marc Baldus at the Max-Planck-Institute for Biophysical Chemistry, Gottingen, Germany, we have almost completed the structure determination of isotopically labeled neurotensin, bound to its receptor, by solid state NMR techniques. Our ability to produce receptors on a regular basis, and to efficiently reconstitute receptors into lipid has allowed remarkable progress towards this goal. (7) I have established the Membrane Protein Interest Group (MPIG), which provides a forum for discussion relating to all aspects of membrane proteins (functional and structural). The MPIG was well received on campus and has been very active (see www.nih.gov/sigs/mpig/).
