The Skaggs Institute
for Chemical Biology

Structure-Based Design of Bioactive Agents

M.R. Ghadiri, J.M. Beierle, A. Chavochi, L. Leman, A. Montero, C.A. Olsen

We are interested in advancing rational structure-based strategies for the design of bioactive agents.

Generation Of Turn Mimetics

Many protein-protein interactions are mediated through the recognition of β-turn secondary structures. Consequently, small-molecule β-turn mimetics are valuable probes for assessing bioactive ligand conformations, establishing pharmacophoric requirements, and pursuing rational drug designs. Although effective drug scaffolds have been developed to precisely position up to 4 functional groups primarily in 2 dimensions, an analogous rigid scaffold capable of predictably juxtaposing 4 amino acid side chains in 3 dimensions has not been readily available. In order to meet this deficiency, diverse approaches have been taken to constrain peptides or peptidelike structures into turn conformations.

One strategy for generating turn mimics is the use of cyclic tetrapeptides. Because of their appropriate size, shape, and useful synthetic modularity, cyclic tetrapeptides in principle offer an attractive platform to mimic β-turn regions. However, these tetrapeptides remain largely unexplored because of poor synthetic efficiency in constructing the strained 12-membered ring, an inability to control cis-trans backbone geometry, and the apparent requirement to sacrifice 1 of 4 amino acid residues to incorporate a proline or other turn-forming residue.

To confront the limitations associated with cyclic tetrapeptides as β-turn mimics, we have designed and structurally analyzed 2 alternative classes of 13- or 14-membered ring pseudotetrapeptides containing either 1 or 2 triazole moieties, respectively (Fig. 1). Moreover, we have completed the design, syntheses, structural analyses, and determination of the somatostatin receptor binding activities of a library of all 16 possible strereoisomeric pseudotetramers incorporating the somatostatin pharmacophore. In these studies, we exploited the 1,4-disubstituted 1,2,3-triazole as a trans peptide-bond surrogate. Structural analysis of the diastereomeric library with nuclear magnetic resonance (NMR) spectroscopy indicated that each peptide scaffold adopts a distinct, rigid, conformationally homogeneous turnlike structure in solution. The 3-dimensional pharmacophoric display of the pseudotetrapeptides is systematically altered by varying the stereochemistry around the otherwise constitutionally identical scaffolds, yielding both compounds with broad-spectrum activity against the 5 human somatostatin receptor subtypes and compounds with receptor selectivity. Our studies provide a basic set of scaffolds with subtle but predictable differences in the spatial display of amino acid side chains that are useful for rational, structure-based drug design.

Fig. 1. Chemical and molecular structures of representative members of 2 classes of cyclic pseudotetrapeptide scaffolds. All 4 compounds had 1H NMR spectra consistent with a single conformational species in solution. Structures were determined by using multidimensional NMR (1–3) or x-ray crystallography (4). Dmb = dimethoxybenzyl.

Inhibitors Of Histone Deacetylases

A fundamental strategy in rationally designing synthetic compounds to bind a protein of interest is to use a known ligand as a structural model to specify the precise conformational and pharmacophoric requirements for binding. Despite the remarkable success of this approach, a major difficulty is that compared to the receptor-bound structure, free ligands (in the absence of their cognate receptors) often adopt multiple conformations in solution or in the solid state. These occurrences can make design models based on the free ligand structure difficult to obtain or even misleading.

Using the rigid scaffold strategy described earlier, we have gathered evidence that the more potent conformation of apicidin, an archetypal member of a family of naturally occurring cyclic tetrapeptide inhibitors of histone deacetylases (HDACs), is not the previously believed all-trans (t-t-t-t) structure that predominates in solution, but rather a cis-trans-trans-trans (c-t-t-t) conformation (Fig. 2). Our approach relies on the design, synthesis, structural characterization, and functional analysis of a series of cyclic pseudotetrapeptides bearing 1,4- or 1,5-disubstituted 1,2,3-triazole amino acids that serve as trans- or cis-amide bond surrogates, respectively. We have shown that by replacing an amide bond with a triazole, we can fix the bond in question in either a trans- or a cis-like configuration, allowing us to individually probe the binding affinity of distinct peptide conformations. The heterocyclic compounds adopt conformations that overlay closely with the targeted conformations of apicidin and have potent HDAC inhibitory activities, in some instances equivalent to or better than those of the natural product. This study highlights the usefulness of triazole-modified cyclic peptides in constructing useful bioactive probe molecules, supports the c-t-t-t conformation as the bioactive conformation of the cyclic tetrapeptide HDAC inhibitors, and provides a useful 3-dimensional pharmacophoric model for use in advancing design principles for more selective HDAC inhibitors.

Fig. 2. NMR structures for the triazole-modified apicidin analogs (side chains are omitted for clarity). A, Peptide 2 adopted a t-t-t-t conformation (top) that overlays well (bottom) on the lowest energy calculated conformation of apicidin (yellow; which reportedly closely matches the predominant conformation of apicidin in dimethylsulfoxide). B, Peptide 11 adopted a c-t-t-t conformation; 2 families of structures were observed that differ in the rotation of the Trp/Aoda amide relative to the backbone (top). The structures overlay well on the crystal structure of apicidin (bottom). C, Peptide 12 adopted a c-t-c-t conformation (top), which overlays well on the crystal structure of the natural product dihydrotentoxin (yellow; bottom). D, Overlay of Ca atoms for compounds 2 (magenta), 11 (2 structural families, green and cyan), and 12 (yellow). Because of the c-t-c-t conformation of 12, the Aoda, Trp, and Leu side chains of this peptide project in the same plane as the backbone ring rather than projecting upward and out of the plane as for the other peptides. E, Overlay of Ca and Cb atoms for compounds 2 (magenta) and 11 (2 structural families, green and cyan). The Ca atoms of the Ala and Ile/Leu residues are farther apart in 2, and the Ca-Cb vector of 2 directs the Ile side chain outward away from the ring rather than directly above the ring in 11.