Pehr A.B. Harbury
Molecular Structure and Molecular Engineering
We are studying the molecular mechanisms that confer specific shapes on proteins, and which determine how proteins recognize small molecules. The goal is to elucidate predictive principles by which novel structures and catalytic properties can be conferred accurately on designed polypeptides, and to achieve the rational design of ligands for proteins of known conformation. The lab relies primarily on three tools: (a) the computational engineering of structures at atomic resolution, made possible by the advent of classical molecular mechanics potentials (b) biophysical characterization of peptide proteins composed from an expanded amino acid alphabet (c) the generation and screening of combinatorial libraries, both in vivo using bacterial screens and sexual PCR and in vitro by synthesis of compounds on solid support.
The lab is divided into two main areas of research. The first group is focused on the computational design and experimental verification of novel a/b barrel proteins. The a/b barrel fold was chosen because of the structural regularity of its naturallyoccurring examples, because its tetrameric point symmetry reduces the asymmetric unit that must be designed, because it supports a wide variety of catalytic functions that are amenable to selection, and because its constituent secondary structures can be subjected to backbone parameterization. Computational engineering of fourfold symmetric a/b barrels requires the exploration of eight backbone degrees of freedom and of sidechain identity at ten core positions, which lies within the reach of current algorithms and hardware. The designed barrel sequences are made by peptide synthesis and are characterized structurally. Designed a/b barrels will be outcrossed against naturally occurring catalytic a/b barrels to determine a minimal set of mutations required to confer catalytic activity on the designed scaffolds.
The second group is focused on the synthesis, screening and crystallographicallyguided optimization of small-molecule libraries based on athio bamino acid polymers. These polymers are particularly suitable pharmaceutical substrates because of the diversity they afford per unit molecular weight, because their chemical structure is expected to confer favorable pharmacokinetic properties, and because their synthesis is both modular and expedient. Methods are being developed to express athio bamino acid polymers biosynthetically using a modified in vitro translation extract, and to select for binding to macromolecular targets using polyribosome display. Cocrystal structures of lead compounds will be used to computationally screen the enormous number of possible generating monomers present in the Fine Chemical Directory, and to thereby develop predictions for ligand improvement that may be rapidly tested.