Emerging Computational Methods for the Rational Discovery of Allosteric Drugs

Abstract

Allosteric drug development holds promise for delivering medicines that are more selective and less toxic than those that target orthosteric sites. To date, the discovery of allosteric binding sites and lead compounds has been mostly serendipitous, achieved through high-throughput screening. Over the past decade, structural data has become more readily available for larger protein systems and more membrane protein classes (e.g., GPCRs and ion channels), which are common allosteric drug targets. In parallel, improved simulation methods now provide better atomistic understanding of the protein dynamics and cooperative motions that are critical to allosteric mechanisms. As a result of these advances, the field of predictive allosteric drug development is now on the cusp of a new era of rational structure-based computational methods. Here, we review algorithms that predict allosteric sites based on sequence data and molecular dynamics simulations, describe tools that assess the druggability of these pockets, and discuss how Markov state models and topology analyses provide insight into the relationship between protein dynamics and allosteric drug binding. In each section, we first provide an overview of the various method classes before describing relevant algorithms and software packages.

Figure 1

When a small molecule binds to the allosteric site of a protein, information is transferred through the protein molecule to its active site. Two different methods of transmission can be defined. The first mechanism, here defined as the “domino model”, is a sequential set of events propagating linearly from the allosteric site to the active site. Binding of the effector triggers local structural changes that sequentially propagate via a single pathway to the active site. It was suggested that this mechanism is applicable for the PDZ domain family, G protein-coupled receptors, the chymotrypsin class of serine proteases, and hemoglobin. The second mechanism, defined here conceptually as a “violin model”, is based on vibration pattern changes inside the protein. In a violin its pitch can be changed by a slight movement of the violin player’s finger on the fingerboard. Information about the finger movement is, thus, transferred throughout the whole body of the violin with no specific pathway for the signal transduction. By analogy, protein allosteric site is a fingerboard of the protein and a small signaling molecule is the player’s finger. If a protein is in a particular vibration mode, it is possible to suggest that binding a small effector molecule to a specific site can change this mode. The signal, thus, will be spread throughout the whole protein including its active site. The “domino model” is a reliable way to transfer information in a macro world, but on a molecular level, with significant thermal motions of the protein, this mechanism will be prone to random triggering of the domino chain reaction, creating noise in the signaling system. Thermal motions in the case of the “violin model” do not hinder the transduction. In fact, the permanent motion of the molecule is a prerequisite for this mechanism. Reproduced with permission from ref (45). Copyright 2015 Cell Trends in Biological Sciences.

Figure 2

Allosteric protein fructose 1,6-bisphosphatase, shown for illustration. Orthosteric and allosteric pockets (yellow and red, respectively) are bound to an endogenous ligand and an allosteric effector, respectively. Note that the allosteric site is distant from the orthosteric site such that there is no overlap between the bound poses of the allosteric and orthosteric ligands. Despite the distance between them, the allosteric effector measurably modifies the enzymatic activity at the orthosteric site. Illustration derived from PDB IDs 2Y5K(53) and 3IFC.