Mapping the functional surfaces of two signal transduction proteins

Abstract

The modern methodology to characterize protein-ligand interactions at the atomic level encompasses structural, mutagenesis, and functional studies. This multidisciplinary approach has been employed to characterize the functional surfaces of two signal transduction molecules, protein tyrosine phosphatase 1B (PTP1B) and gelsolin, for the divergent goals of inhibitor design and characterizing macromolecular assemblies.;The structure of the catalytically inactive mutant (C215S) of PTP1B has been solved in complex with bis-(para-phosphophenyl)methane (BPPM), a synthetic high-affinity substrate, and in the presence of saturating phosphotyrosine (pTyr). BPPM binds PTP1B in two modes, one in which it occupies the active site, and another in which a phosphophenyl moiety interacts with residues not previously observed to bind aryl phosphates. A second pTyr also binds the same site in the pTyr complex, confirming that these residues constitute a low-affinity aryl phosphate-binding site. Identification of this site provides a paradigm for the design of tight-binding, specific PTP1B inhibitors.;PTP1B/C215S has also been solved in complexes with two peptides, Ac-DAD(Bpa)pYLIPQQG and Ac-ELEFpYMDYE-NH2. These structures demonstrate how PTP1B accommodates both aromatic and acidic residues at a position one residue N-terminal to the pY in a bound peptide. These interactions provide a structural basis for binding peptides of diverse sequence, and may be utilized in ligand design studies.;Another signal transduction molecules whose interaction surface was characterized is the actin regulatory protein gelsolin. Gelsolin family members possess 3--6 homologous domains, of which the second domain interacts with F-actin through an unknown binding mode. The structure of the F-actin binding domain 2 of severin, the gelsolin homologue from Dictyostelium discoideum , has been solved by MIR, revealing an alpha-helix/beta-sheet sandwich similar to domains of gelsolin and villin. Comparison of the structures of several gelsolin family domains has identified residues which may constitute the F-actin binding surface of gelsolin domain 2. To test this hypothesis, three mutants of human gelsolin domain 2 anticipated to have decreased filament-binding activity were constructed, two of which had significantly reduced affinity for F-actin. The structural and mutagenesis studies have been integrated with the existing literature to construct an atomic model of the gelsolin domain 2/F-actin complex.