1 year ago

2004 Scientific Report

Laboratory of Structural

Laboratory of Structural Sciences H. Eric Xu, Ph.D. Dr. Xu went to Duke University and the University of Texas Southwestern Medical Center, where he earned his Ph.D. in molecular biology and biochemistry. Following a postdoctoral fellowship with Carl Pabo at MIT, he moved to GlaxoWellcome in 1996 as a research investigator in nuclear receptor drug discovery. Dr. Xu joined VARI as a Senior Scientific Investigator in July 2002. Staff Yong Li, Ph.D. David Tolbert, Ph.D. Phuzile Ludidi, Ph.D. Jennifer Daugherty, B.S. Laboratory Members Amanda Kovach, B.S. Jennifer Kretschman, B.S. Kelly Suino, B.S. Visiting scientist Ross Reynolds, Ph.D. Research Interests Our laboratory uses x-ray crystallography to study the structures and functions of key protein complexes that are important in basic biology, as well as in drug discovery relevant to human diseases such as cancer and diabetes. Currently we are focusing on nuclear hormone receptors and the Met tyrosine kinase receptor. The nuclear hormone receptor family comprises a large number of ligand-regulated and DNAbinding transcriptional factors, which include receptors for classic steroid hormones such as estrogen, progesterone, androgens, and glucocorticoids, as well as for proxisome proliferator activators, vitamin D, vitamin A, and thyroid hormones. One distinguishing fact about the classic receptors is that they are among the most successful targets in the history of drug discovery: every receptor has one or more cognate synthetic ligands currently being used as medicines. In the last two years, we have developed the following projects centering on the structural biology of nuclear receptors. Peroxisome proliferator–activated receptors The peroxisome proliferator–activated receptors (PPARα, δ, and γ) are the key regulators of glucose and fatty acid homeostasis and, as such, are promising therapeutic targets for cardiovascular disease, diabetes, and cancer. Millions of patients have benefited from treatment with PPARγ medicines for type II diabetes; 2 of the top 100 drugs in 2003 were PPARγ ligands, with combined sales of over .0 billion. To understand the molecular basis of ligandmediated signaling by PPARs, we have determined crystal structures of each PPAR’s ligandbinding domain (LBD) bound to diverse ligands, including fatty acids, the lipid-lowering drugs called fibrates, and a new generation of anti-diabetic drugs, the glitazones (Fig. 1). We have also determined crystal structures of these receptors bound to co-activators or co-repressors. These structures provide a framework for understanding the mechanisms of agonists and antagonists, as well as the recruitment of co-activators and co-repressors. These structures also have provided crucial insights for designing the next generation of PPAR medicines. Currently we are developing this project beyond the structures of the ligand-binding domains into the structures of large PPAR fragment/DNA complexes. The human glucocorticoid receptor The human glucocorticoid receptor (GR) is a key regulator of energy metabolism and of homeo- Figure 1. Crystal structure of the PPAR ligandbinding domains. Each PPAR molecule is shown bound to its respective ligand. Each contains 13 α helices and 4 small β strands. The helices are arranged into a three-layer sandwich fold to create a cavity for ligand binding. 62

stasis of the immune system. GR is also a classic target of drug discovery because of its association with numerous pathological conditions. There are more than 10 GR ligands (including dexamethasone) that are currently used for treatment of such diverse conditions as asthma, allergy, autoimmune diseases, and cancer. At the molecular level, GR can function either as a transcription activator or a transcription repressor. Both functions are tightly regulated by small ligands that bind to the LBD. We have determined a crystal structure of the GR LBD bound to dexamethasone and a co-activator motif from TIF2. The structure reveals a novel LBD-LBD dimer interface (Fig. 2), an unexpected charge clamp responsible for sequence-specific binding of co-activators, and a unique ligand-binding pocket to account for specific recognition of diverse GR ligands. Currently we are crystallizing GR with various steroid or nonsteroid ligands. The information from these structures should provide a rational basis for designing new GR ligands that would reduce side effects relative to current GR drugs. In a collaborative effort, we are extending our studies to the structure of a large GR fragment bound to DNA. The human androgen receptor The androgen receptor (AR) is the central molecule in the development and progression of prostate cancer, and as such it serves as the molecular target of anti-androgen therapy. However, the majority of prostate cancer patients develop resistance to such therapy, mostly due to mutations in AR that alter its three-dimensional structure, allowing it to escape repression by anti-androgen treatment. Such a hormone-independent prostate cancer is highly aggressive and is responsible for most deaths; currently there is no cure. The development of effective therapies requires detailed understanding of the structure and functions of the central molecule, i.e., the androgen receptor and its interactions with hormones and co-regulators. In this project, our goal is to determine the structures of the mutated AR proteins that alter the response to anti-hormone therapy. We are working on the crystal structure of the full-length AR/DNA complex. Structural genomics of nuclear receptor ligand-binding domains The ligand-binding domain (LBD) of nuclear receptors contains key structural elements that Figure 2. Two views of the crystal structure of the human glucocorticoid receptor LBD bound to dexamethasone and a TIF2 co-activator. mediate ligand-dependent receptor regulation; this domain has been the focus of intense structural studies. Crystal structures for more than half of the 48 human nuclear receptors have been determined. These structures have illustrated the details of ligand binding, the conformational changes induced by agonists and antagonists, the basis of dimerization, and the mechanism of coactivator and co-repressor binding. They also provide many surprises in terms of the identity of ligands, the size and shape of the ligand-binding pockets, and the structural implications of the receptor signaling pathways. There is now only one receptor, constitutive androstanol receptor (CAR), for which the LBD structure remains unsolved. Currently we are studying the mouse CAR molecule in a collaborative effort. The Met tyrosine kinase receptor Met is a tyrosine kinase receptor that is activated by hepatocyte growth factor/scatter factor (HGF/SF). Aberrant activation of the Met receptor has been linked to the development and metastasis of many types of solid tumors and has been correlated with poor clinical prognosis. HGF/SF has a modular structure with an N-terminal domain, four kringle domains, and an inactive serine protease domain. The structure of the N-terminal domain, which has a single kringle domain (NK1), has been determined; less is known about the structure of the Met extracellular domain. The molecular basis of the Met-HGF/SF interaction and the resulting activation of Met signaling remain poorly understood. In collaboration with George Vande Woude and Ermanno Gherardi, we are developing this project to solve the crystal structure of the Met receptor/HGF complex. 63

Publications by Year