Leahy - Research Interests

Molecular Aspects of Cell Adhesion and Signaling
Multicellular organisms require mechanisms for intercellular communication and adhesion. These functions are often intertwined and invariably involve molecular interactions at the cell surface. Our laboratory is interested in the nature of these interactions and how they generate signals that influence cell fate. We hope to generate insight into the way many individual interactions are integrated to create and maintain viable organisms. As an essential first step in understanding the nature of specific interactions, we are pursuing X-ray crystallographic studies of targeted receptors, their ligands, and ultimately complexes of these receptors and ligands. Our work focuses on members of the Hedgehog, Wnt, and epidermal growth factor (EGF) signaling pathways.
The EGF Receptor Family
The EGF receptor (EGFR, ErbB1, HER1) consists of a ~630 amino acid extracellular domain, a single transmembrane spanning region, and a cytoplasmic tyrosine kinase. Humans possess three EGFR homologs, HER2 (ErbB2, Neu), HER3 (ErbB3), and HER4 (ErbB4), which share 40-45% sequence with EGFR and one another. These receptors mediate the growth and differentiation of many cell types during normal animal development. Overexpression and activation of these receptors is found in many human cancers and is associated with the development and severity of these tumors. Ligand-induced receptor oligomerization is thought to activate the cytoplasmic kinase domains, which leads to phosphorylation and activation of downstream targets, but the molecular details of activation have been poorly understood. At least 11 different ligands activate EGFR family members with most exhibiting specificity for more than one receptor and many appearing to signal best through heteromeric receptor complexes. We have expressed the extracellular domains of the four human EGFR homologs and determined the crystal structures of HER3, HER2, and HER2 complexed with a monoclonal antibody, Herceptin, used in the therapy of human breast cancers. We have also collaborated with the groups of Mark Lemmon and Kate Ferguson (University of Pennsylvania) to determine the structure of the EGFR extracellular domain. By comparing our structures with the recently determined structures of EGFR complexed with either EGF or TGF_, it has become apparent that ErbB receptors adopt an autoinhibited conformation in the absence of ligand binding and that ligand binding induces a significant conformational change in the receptors, which promotes receptor dimerization and activation. Intriguingly, HER2 appears to be fixed in an active-like conformation even in the absence of ligand, which explains both why no HER2 ligand has been found and the readiness of HER2 to serve as a partner with other ErbB receptors in heteromeric signaling complexes. The failure of HER2 to adopt an autoinhibited conformation may also explain in part why it is unique among ErbB receptors in transforming cells when overexpressed in the absence of ligand. Overexpression of HER2 is found in 20-30% of human breast cancers and is associated with more aggressive tumors and a poorer prognosis. Elucidation of the molecular details of ErbB receptors and their signaling mechanisms has stimulated development of new inhibitors of their function that may prove useful cancer therapies. These inhibitors include antibodies targeting regions of ErbB receptors critical for function and mutated ligands that are likely to bind the receptor but not induce the conformational change in the receptor that is required for signaling.
We are currently attempting to grow crystals of larger fragments of these receptors to gain more insight into how conformational changes in the extracellular regions of these ligands are transmitted across the cell membrane. We are also studying the properties of chimeric ErbB receptors to identify features responsible for mediating specific functions and are.
return to top

Hedgehog and Wnt
Many secreted factors have been shown to participate in the induction and patterning of various tissues during the development of multicellular organisms. Prominent among these factors are members of the Hedgehog and Wnt families of signaling molecules. To provide a molecular picture of how these factors mediate their effects we are expressing components of the Hedgehog and Wnt signaling pathways for biochemical and structural studies.
Hedgehog proteins undergo an intramolecular autoprocessing that results in an amino-terminal ~19-kDa fragment (Hh-N) that is attached to the cell membrane via a covalently attached cholesterol moiety and a soluble carboxyl-terminal ~25-kDa domain (Hh-C). All signaling activities of Hedgehog proteins have been shown to reside in Hh-N; Hh-C is responsible for both the peptide cleavage and cholesterol transfer components of the autoprocessing reaction. In collaboration with Philip Beachy (HHMI, Johns Hopkins University), we determined the crystal structures of the two major domains of the Hedgehog protein. Hh-C proved homologous to the intein region of self-splicing proteins, and the mechanisms of Hedgehog autoprocessing and protein self-splicing share many features. The structure of Hh-N revealed an unexpected homology to zinc hydrolases, suggesting that an enzymatic activity may participate in Hedgehog signaling. Subsequent mutagenesis studies have indicated that a hydrolytic activity is not required for Hedgehog signaling, however, and the relationship between Hh-N and zinc hydrolases appears to be one of form, not function. We are now expressing downstream effectors of Hedgehog signaling, including fragments of the Patched, Smoothened, and CG9211 integral membrane receptors and the cytoplasmic proteins Fused, Suppressor of Fused, Costal-2, and Cubitus Interruptus. Our goal is to identify the regions of each protein that interact with other pathway members and understand how each interaction in the signaling cascade is regulated. It is expected that determination of crystal structures of interaction complexes will form a large part of our efforts and contribute significantly to our understanding of these interaction events. In collaboration with Jeremy Nathans (HHMI, Johns Hopkins University), we have recently determined the crystal structures of ligand-binding fragments of two members of the Frizzled protein superfamily. Frizzleds serve as receptors for Wnt proteins, and these structures enabled identification by site-directed mutagenesis of a Frizzled surface important for mediating interactions with Wnt proteins. In addition, both structures revealed a homologous dimer interface despite both proteins existing as monomers in solution at concentrations up to 100 _M. The presence of a weak yet conserved dimer interface suggests that dimerization may be of biological significance. In particular, dimerization in response to ligand binding may be a feature of Frizzled-mediated signaling. Attempts to test this hypothesis are currently underway but have been hampered by difficulties obtaining biochemically-defined preparations of Wnts. The recent discovery by Nusse and colleagues of the site of a lipid modification of Wnts is being exploited to isolate a mutated Wnt with better solubility properties. Many other proteins, including the LDL receptor related protein 6, Wnt Inhibitory Factor, Dickkopf, and Kremen have been shown to act as either inhibitors or co-receptors for Wnt signaling, and studies of the molecular roles played by these molecules in Wnt signaling are also underway.

	return to top

Leahy Laboratory

The EGF Receptor Family

Hedgehog and Wnt