Prediction of allosteric sites and signaling: Insights from benchmarking datasets

Abstract

Allostery is a pervasive mechanism that regulates protein activity through ligand binding at a site different from the orthosteric site. The universality of allosteric regulation complemented by the benefits of highly specific and potentially non-toxic allosteric drugs makes uncovering allosteric sites invaluable. However, there are few computational methods to effectively predict them. Bond-to-bond propensity analysis has successfully predicted allosteric sites in 19 of 20 cases using an energy-weighted atomistic graph. We here extended the analysis onto 432 structures of 146 proteins from two benchmarking datasets for allosteric proteins: ASBench and CASBench. We further introduced two statistical measures to account for the cumulative effect of high-propensity residues and the crucial residues in a given site. The allosteric site is recovered for 127 of 146 proteins (407 of 432 structures) knowing only the orthosteric sites or ligands. The quantitative analysis using a range of statistical measures enables better characterization of potential allosteric sites and mechanisms involved.

Keywords: allosteric site detection; benchmarking; graph theory measures.

Graphical abstract

Figure 1

Atomistic graph construction Main steps of the atomistic protein graph construction package, BagPype, using the structure of bovine seminal ribonuclease (PDB: 11BG) as an example.

Figure 2

Bond-to-bond propensity analysis on the atomistic graph of bovine seminal ribonuclease (PDB: 11BG) where the orthosteric residues (green) are used as the perturbation source (A) All residues are colored by residue QS (see legend) obtained from bond-to-bond propensity analysis. (B) Surface representation of the protein structure colored by QS. Relevant sites are highlighted and labeled accordingly.

Figure 3

Summary of allosteric site detection results for 118 structures in the ASBench database Propensity analysis was conducted for all 118 structures under two conditions: (1) with allosteric ligand in the protein structure (blue) and (2) without allosteric ligand in the protein structure (orange). The x-axis represents the number of statistical measures that successfully identify the allosteric site. Each bar indicates the number of protein structures of which the allosteric sites can be detected by a certain number of statistical measures shown on the x-axis. Take the last two bars as an example: the allosteric site(s) can be detected using all six statistical measures for 26 proteins structures with the presence of allosteric ligand (blue bar). When the allosteric ligand is removed from the structures, allosteric site(s) of 19 structures can be identified with all six measures (orange bar). Detailed data can be found in Tables S3 and S4.

Figure 4

Structure of human muscle glycogen phosphorylase (PDB: 1Z8D) The orthosteric (green) and two allosteric (circled in blue and red) site residues are highlighted as spheres.

Figure 5

Summary of allosteric site detection results for 314 structures in the CASBench database Propensity analysis was conducted for all 314 structures under two conditions: (1) using orthosteric ligand(s) as the perturbation source (blue) and (2) using orthosteric residues (removed orthosteric ligands) as the perturbation source (orange). The x-axis represents the number of statistical measures that successfully identify the allosteric site. Each bar indicates the number of protein structures of which the allosteric sites can be detected by a certain number of statistical measures shown on the x-axis. Take the last two bars as an example: the allosteric site(s) can be detected using all six statistical measures for 56 proteins structures when using the orthosteric ligand(s) as the perturbation source. When using the orthosteric site residues as the perturbation source, allosteric site(s) of 58 structures can be identified with all six measures. Detailed data can be found in Tables S6 and S7.