Very fast prediction and rationalization of pK_a values for protein–ligand complexes

Abstract

The PROPKA method for the prediction of the pKa values of ionizable residues in proteins is extended to include the effect of non‐proteinaceous ligands on protein pKa values as well as predict the change in pKa values of ionizable groups on the ligand itself. This new version of PROPKA (PROPKA 2.0) is, as much as possible, developed by adapting the empirical rules underlying PROPKA 1.0 to ligand functional groups. Thus, the speed of PROPKA is retained, so that the pKa values of all ionizable groups are computed in a matter of seconds for most proteins. This adaptation is validated by comparing PROPKA 2.0 predictions to experimental data for 26 protein–ligand complexes including trypsin, thrombin, three pepsins, HIV‐1 protease, chymotrypsin, xylanase, hydroxynitrile lyase, and dihydrofolate reductase. For trypsin and thrombin, large protonation state changes (|n| > 0.5) have been observed experimentally for 4 out of 14 ligand complexes. PROPKA 2.0 and Klebe's PEOE approach (Czodrowski P et al. J Mol Biol 2007;367:1347–1356) both identify three of the four large protonation state changes. The protonation state changes due to plasmepsin II, cathepsin D and endothiapepsin binding to pepstatin are predicted to within 0.4 proton units at pH 6.5 and 7.0, respectively. The PROPKA 2.0 results indicate that structural changes due to ligand binding contribute significantly to the proton uptake/release, as do residues far away from the binding site, primarily due to the change in the local environment of a particular residue and hence the change in the local hydrogen bonding network. Overall the results suggest that PROPKA 2.0 provides a good description of the protein–ligand interactions that have an important effect on the pKa values of titratable groups, thereby permitting fast and accurate determination of the protonation states of key residues and ligand functional groups within the binding or active site of a protein. Proteins 2008. © 2008 Wiley‐Liss, Inc.

Identification of cation-binding sites on actin that drive polymerization and modulate bending stiffness

“…Mutations within this predicted site can render cytoplasmic actin nonpolymerizable (44) and can be lethal in S. cerevisiae (herein referred to as yeast) (45). We note that an acidic residue within 3 Å of the polymerization site (Asp288) is predicted to have a pK a shifted from approximately 4 to 7.1 [using PROPKA software (46)] when incorporated within a filament (SI Text). Consequently, protonation at this site may account for the filament stabilizing effects of decreasing the solution pH (47).…”

Enzyme Promiscuity in Enolase Superfamily. Theoretical Study of o-Succinylbenzoate Synthase Using QM/MM Methods

“…The protonation state of titratable residues at pH 7 was determined using the program PropKa 3.1. [26][27][28][29] All the His residues were protonated because they are located at the surface of our model and are accessible to the solvent.…”

Specific interactions between DNA and regulatory protein controlled by ligand-binding: Ab initio molecular simulation

“…We here evaluated the pK a values of His residues contained in this structure by using the PROPKA Web Interface 3.0 [11][12][13][14] and determined the His protonation based on the pK a values. His residues with pK a value larger than 6 have the Hip protonation, while those with pK a value smaller than 6 have the Hid or Hie protonation.…”