Binding MOAD (Mother of All Databases).

摘要

Binding MOAD (Mother of All Databases) is the largest collection of high-quality, protein-ligand complexes available from the Protein Data Bank. At this time, Binding MOAD contains 11,368 protein-ligand complexes composed of 3583 unique protein families and 5363 unique ligands. We have searched the crystallography papers for all structures and compiled binding data for 3543 (31%) of the protein-ligand complexes. The binding-affinity data ranges 13 orders of magnitude. This is the largest collection of structural binding data to date in the literature.;This database of protein-ligand complexes is proving very useful in exploring biophysical patterns of molecular recognition and enzymatic regulation. Mining Binding MOAD has revealed physical differences in how enzymes and nonenzymes bind small molecules. High-affinity ligands of enzymes are much larger than those with low affinity, but high- and low-affinity ligands of nonenzymes are the same size. This suggests that different approaches may be appropriate for improving the affinity of ligands. While the addition of complementary functional groups is likely to improve the affinity of an enzyme inhibitor, it may not be as fruitful for ligands of nonenzymes. For nonenzymes, small changes and isosteric replacements might be more productive. Furthermore, nonenzymes were found to have higher ligand efficiencies. The different efficiencies are not due to differences in the physicochemical properties of the ligands; instead, the amino-acid composition of the pockets are very different despite very similar distributions of amino acids in the overall protein sequences.;This study aims to address the issue of protein flexibility upon ligand binding. The influence of ligand binding on protein flexibility is examined by analyzing a large number of proteins crystallized with and without ligands. A baseline comparison of the natural variation of protein structure with and without ligands is first established, and then differences between the apo and holo are analyzed. It is shown that, in general, ligand binding stabilizes the protein and results in a smaller backbone root mean square deviation (RMSD) among holo-protein structures, compared the backbone RMSD of the apo-protein structures. Furthermore, the holo structures appear to sample a smaller subset of the space inhabited by apo structures, because the difference between apo and holo structures is smaller than variation seen among apo structures themselves. The size of the bound ligand does not appear to matter in determining the rigidification. While ligand binding generally does not induce large changes in the backbone, they are significant. Ligand binding does have distinct impact on the active site, as revealed by all-atom, active-site RMSD and the range of chi1 variation. Apo structures are observed to have a certain range of flexibility in their active sites, just as holo structures have a similar, but smaller, degree of variation among their active sites. However, greater variation has been found between these two groups as opposed to within either group by themselves. This suggests that ligand binding induces active-site side chains to occupy a different conformational space before and after binding. The influence on the active site could not be easily attributed to features such as ligand size, resolution, protein function, or catalytic composition.;The studies above illustrate the usefulness of large carefully annotated datasets for studying protein-ligand interactions. Binding MOAD has almost doubled in size since it was originally introduced in 2004, demonstrating steady growth with each annual update. Several technologies are described, such as natural language processing, that help drive this constant expansion.;In summary, Binding MOAD is a valuable resource. It has helped to illuminate fundamental differences between enzymes and nonenzymes and allowed for examination of the influence ligand binding has in protein flexibility. It has great potential to further advance our understanding of protein-ligand interactions.