The Skaggs Institute
for Chemical Biology

Scientific Report 2006

New Approaches in the Design of Anticancer Agents and Synthetic Catalysts

M.R. Ghadiri, J.M. Beierle, A. Chavochi, W.S. Horne, Z.-Z. Huang, L. Leman, A. Montero

We are interested in devising molecular and supramolecular processes to control the structure and function of de novo designed peptides. These efforts include the design of potent anticancer agents based on conformationally homogenous cyclic peptides and fabrication of self-assembling helical peptides that mimic the intermodular aminoacyl transferase activity of nonribosomal peptide synthetases.

Small-Molecule 3-Dimensional Rigid Scaffolds in the Design of Somatostatin Agonists and Histone-Deacetylase Inhibitors

In contrast to the considerable advances made in drug design based on 2-dimensional rigid scaffolds, little progress has been made in devising effective 3-dimensional scaffolds. Small-molecule scaffolds with predictable and conformationally homogeneous 3-dimensional structures are expected to have significant usefulness in drug design and discovery, because typical therapeutic agents, whether synthetic or derived from natural products, generally act on their biological targets (receptors) by mimicking the 3-dimensional structural features of their native biological ligands—most often peptides and proteins. Moreover, many protein-ligand interactions are thought to involve the recognition of β-turn protein secondary structures. Consequently, it is thought that constrained peptides designed to adopt specific turn structures can provide valuable probes for assessing the conformations of bioactive ligands and aid in the rational design of therapeutic agents.

We recently developed 12-, 13-, and 14-membered peptidomimetic scaffolds, incorporating 1,2,3-triazole ε^2,6-amino acids as novel cis or trans dipeptide surrogates, that mimic precisely side-chain juxtaposition and functional-group presentations of the 4 residues that make up β-turn structures (Fig. 1). Structural analyses based on x-ray crystallography, 2-dimensional ¹H nuclear magnetic resonance spectroscopy, and distance-geometry calculations confirmed that the heterocyclic peptide scaffolds adopt rigid and conformationally homogeneous β-turn structures in solution and in the solid state. We have used this approach to probe the bioactive conformations of natural cyclic tetrapeptide histone-deacetylase (HDAC) inhibitors.

Fig. 1. The chemical structures of the natural HDAC inhibitor apicidin V (top), the synthetic heterocyclic analog (middle), and the solution nuclear magnetic resonance structure indicating the β-turn– like rigid backbone conformation and side-chain display (bottom).

Because of their key roles in the transcriptional regulation of a number of genes involved in cell proliferation, cell-cycle progression, differentiation, and apoptosis, HDACs are important new targets in the design of novel anticancer agents. Our studies indicate that appropriately designed triazole-modified cyclic peptides can have potent HDAC inhibitory activities, rivaling those of many previously known natural and synthetic HDAC inhibitors. We also showed the usefulness of 3-dimensional scaffolds in the design of somatostatin agonists with high receptor subtype selectivity. Because the constrained heterocyclic scaffolds can be used to display a variety of amino acid side chains and backbone geometries in specific and predictable 3-dimensional arrangements, we think that our basic design concepts can have applications in a number of other structure-based rational design of drugs.

Design of Synthetic Peptide Catalysts

Ribosomes, nonribosomal peptide synthetases, and polyketide synthetases are biological machines that catalyze “instructed” chemical synthesis, a phenomenon not yet matched by any synthetic system. Although chemists have amassed a remarkable record of achievements in devising proficient and selective catalysts, the phenomenon of autonomous instructed synthesis has remained beyond reach. We have focused on the design of catalysts that can mimic the sequential aminoacyl transferase activity of nonribosomal peptide synthetases. We have designed several modular supramolecular peptide constructs that efficiently promote site-specific aminoacyl transfer between noncovalently associated α-helical subunits in neutral aqueous solutions (Fig. 2). Moreover, in studies based on homomeric and heteromeric assemblies, substitutions of amino acids in the active site, ¹⁵N nuclear magnetic resonance–based pK_a measurements, kinetic analysis, and reaction modeling indicate that the de novo designed catalysts have some of the basic hallmarks of natural enzymes, including precise positioning of active-site residues, covalent catalysis, general acid-base catalysis, pK_a modulation of active-site residues, and multiple product turnovers. We hope that these studies will help establish the basic blueprint for the future de novo design of programmable peptide synthetases.

Fig. 2. Top, X-ray structure of the de novo designed 4-helix bundle aminoacyl transferase peptide emphasizing the juxtaposition of active-site residues. Bottom, Schematic illustration of the sequential aminoacyl loading and intermodular aminoacyl transfer steps.