Abstract
Allostery commonly refers to the mechanism that regulates protein activity through the binding of a molecule at a different, usually distal, site from the orthosteric site. The omnipresence of allosteric regulation in nature and its potential for drug design and screening render the study of allostery invaluable. Nevertheless, challenges remain as few computational methods are available to effectively predict allosteric sites, identify signalling pathways involved in allostery, or to aid with the design of suitable molecules targeting such sites. Recently, bond-to-bond propensity analysis has been shown successful at identifying allosteric sites for a large and diverse group of proteins from knowledge of the orthosteric sites and its ligands alone by using network analysis applied to energy-weighted atomistic protein graphs. To address the identification of signalling pathways, we propose here a method to compute and score paths of optimised propensity that link the orthosteric site with the identified allosteric sites, and identifies crucial residues that contribute to those paths. We showcase the approach with three well-studied allosteric proteins: h-Ras, caspase-1, and 3-phosphoinositide-dependent kinase-1 (PDK1). Key residues in both orthosteric and allosteric sites were identified and showed agreement with experimental results, and pivotal signalling residues along the pathway were also revealed, thus providing alternative targets for drug design. By using the computed path scores, we were also able to differentiate the activity of different allosteric modulators.
Introduction
Proteins are the primary functional units of the cell, and their regulation is often governed by allostery: a process where binding at a distal "allosteric" site triggers a functional change at the primary "orthosteric" or active site. Despite its biological importance, mapping the specific atomic pathways through which these signals travel remains a significant challenge. Traditional Molecular Dynamics (MD) simulations are often too computationally expensive for high-throughput screening, while coarse-grained models frequently lack the atomic resolution required for precise drug design. This paper introduces a graph-theoretical approach to bridge this gap, mapping the "hidden" communication lines within proteins using atomistic networks.
Methodology
The proposed method builds upon bond-to-bond propensity analysis to identify Propensity Optimised Paths (POPs):
- Atomistic Protein Graphs: Proteins are represented as networks where nodes are atoms and edges represent covalent and non-covalent interactions (hydrogen bonds, salt bridges, hydrophobic contacts), weighted by interaction energy.
- Propensity Scoring: Using a graph-theoretical transfer matrix, the researchers calculate how perturbations propagate through the network. Quantile regression is applied to identify "high-propensity" bonds that communicate more strongly than expected for their physical distance.
- Path Optimization: The core novelty lies in finding the "shortest" path using these high-propensity bonds. A biophysical constraint is imposed: paths cannot take more than one consecutive step along the protein backbone, forcing the signal to "jump" across non-covalent interactions, mimicking natural energy diffusion.
- Computational Mutagenesis: Residues are systematically "deleted" from the graph to measure their impact on communication (z-scores), identifying pivotal residues that act as relay stations.
Results and Discussion
Case Study: h-Ras
In the GTPase h-Ras, the analysis identified the catalytic residue Gln61 as a central hub. Crucially, the method revealed that specific, structurally conserved water molecules are essential nodes in the allosteric circuit rather than mere fillers. The identified POPs traverse the $\beta4$ and $\beta5$ strands, aligning with experimental hydrogen-bonding networks.
Case Study: Caspase-1
For Caspase-1, an enzyme involved in inflammation, the model identified Arg286 and Glu390 as critical for site-to-site coupling. Simulated mutations (e.g., Arg286Ala) resulted in significant drops in path scores, mirroring experimental data where these mutations disrupt allosteric inhibition.
Case Study: PDK1 Modulator Ranking
The method was applied to 11 different structures of the kinase PDK1. The resulting path scores successfully ranked allosteric activators by their experimental potency. Furthermore, the analysis distinguished between activators and inhibitors based on their distinct communication patterns to the regulatory residue Sep241.
Conclusion
By identifying "pivotal" residues that lie outside traditional binding pockets, this POPs approach reveals entirely new targets for drug discovery. While currently focused on single optimized paths from static crystal structures, future developments will move toward evaluating ensembles of pathways to better capture the dynamic and robust nature of protein communication.
Summary updated