Inhibitors of poly-ADP-ribose polymerase (PARP) family proteins are currently in clinical trials as cancer therapeutics, yet the specificity of many of these compounds is unknown. Here we evaluated a series of 185 small-molecule inhibitors, including research reagents and compounds being tested clinically, for the ability to bind to the catalytic domains of 13 of the 17 human PARP family members including the tankyrases, TNKS1 and TNKS2. Many of the best-known inhibitors, including TIQ-A, 6(5H)-phenanthridinone, olaparib, ABT-888 and rucaparib, bound to several PARP family members, suggesting that these molecules lack specificity and have promiscuous inhibitory activity. We also determined X-ray crystal structures for five TNKS2 ligand complexes and four PARP14 ligand complexes. In addition to showing that the majority of PARP inhibitors bind multiple targets, these results provide insight into the design of new inhibitors.

No abstract

The importance of the N-terminal region of human cystatin C or chicken cystatin for the kinetics of interactions of the inhibitors with four cysteine proteinases was characterized. The association rate constants for the binding of recombinant human cystatin C to papain, ficin, actinidin and recombinant rat cathepsin B were 1.1 x 10(7), 7.0 x 10(6), 2.4 x 10(6) and 1.4 x 10(6) M-1.s-1, whereas the corresponding dissociation rate constants were 1.3 x 10(-7), 9.2 x 10(-6), 4.6 x 10(-2) and 3.5 x 10(-4) s-1. N-Terminal truncation of the first ten residues of the inhibitor negligibly affected the association rate constant with papain or ficin, but increased the dissociation rate constant approx. 3 x 10(4)- to 2 x 10(6)-fold. In contrast, such truncation decreased the association rate constant with cathepsin B approx. 60-fold, while minimally affecting the dissociation rate constant. With actinidin, the truncated cystatin C had both an approx. 15-fold lower association rate constant and an approx. 15-fold higher dissociation rate constant than the intact inhibitor. Similar results were obtained for intact and N-terminally truncated chicken cystatin. The decreased affinity of human cystatin C or chicken cystatin for cysteine proteinases after removal of the N-terminal region is thus due to either a decreased association rate constant or an increased dissociation rate constant, or both, depending on the enzyme. This behaviour indicates that the contribution of the N-terminal segment of the two inhibitors to the interaction mechanism varies with the target proteinase as a result of structural differences in the active-site region of the enzyme.

The near-UV spectroscopic changes induced by the binding of recombinant human cystatin A to papain were appreciably different from those induced by cystatin C, reflecting mainly interactions involving the single tryptophan of cystatin C, Trp-106. Cystatin A bound tightly and rapidly to papain and cathepsin L, with dissociation equilibrium constants of approximately 10(-11)-10(-13) M and association rate constants of 3 x 10(6)-5 x 10(6) M-1.s-1. These affinities are at least 50-100-fold higher than previously reported values. The kinetics of binding to papain were consistent with a simple reversible bimolecular reaction mechanism, indicating that cystatin A, like chicken cystatin and cystatin C, binds to papain with no appreciable conformational adaptation of either reacting protein. Cystatin A bound more weakly to actinidin and cathepsins B, C and H, with dissociation equilibrium constants of 10(-8)-10(-9) M. The weaker binding to cathepsin B was largely due to a considerably reduced association rate constant (approximately 4 x 10(4) M-1.s-1), consistent with the 'occluding loop' of cathepsin B markedly restricting the access of cystatin A to the active site. The lower affinities for actinidin and cathepsins C and H were due partly to lower association rate constants (2 x 10(5)-6 x 10(5) M-1.s-1) but primarily to higher dissociation rate constants. The mode of binding of cystatin A to inactivated papains indicated that there is appreciably less space around the active-site cysteine of papain in the complex with cystatin A than in the complexes with chicken cystatin and cystatin C. An N-terminally truncated form of cystatin A, lacking the first six residues, had considerably lower affinity for papain than the full-length inhibitor, consistent with an intact N-terminal region being of importance for proteinase binding.

The interaction between five N-terminally truncated forms of chicken cystatin (starting at Leu-7, Leu-8, Gly-9, Ala-10 and Asp-15) and the cysteine proteinases papain and actinidin was studied by spectroscopic, kinetic and equilibrium methods. The u.v. absorption, near-u.v. c.d. and fluorescence emission difference spectra for the interactions with papain were all similar to the corresponding spectra for intact cystatin. The second-order association rate constants at 25 degrees C, pH 7.4, I 0.15, for the binding of the truncated forms to papain varied about 2-fold, from 6 x 10(6) to 1.5 x 10(7) M-1.s-1, and were comparable to the value of 9.9 x 10(6) M-1.s-1 for intact cystatin. In contrast, the rate constants for the dissociation of the complexes with papain increased markedly with increasing extent of truncation, from 7.5 x 10(-6)s-1 for Leu7 cystatin (a truncated form of cystatin having Leu-7 as its N-terminal amino acid) to 1.6s-1 for Ala10-cystatin, whereas the dissociation rate constants for the latter form and Asp15-cystatin were similar. Consequently, the binding affinities between the truncated cystatins and papain decreased in an analogous manner, as was also shown for the interaction with actinidin by equilibrium measurements. Studies of the binding of the truncated cystatins to inactivated papains indicated that small substituents on the active-site cysteine of the enzyme can be accommodated in the complex without any loss of affinity when the N-terminal segment of the inhibitor is removed. Taken together, the results suggest that in the N-terminal region of chicken cystatin only residues preceding Ala-10 participate in the interaction with proteinases. Of these residues, Leu-7 and Leu-8 together account for about two-thirds of the unitary free energy of binding contributed by the N-terminal region, the relative importance of the two residues being dependent on the target proteinase. Both Gly-9 and residues N-terminal of Leu-7 further stabilize the interaction but contribute substantially smaller binding energies than do the two leucine residues.

Insights into the thermodynamic and kinetic signature of the transient opening of a protein-binding pocket resulting from accommodation of suitable substituents attached to a given parent ligand scaffold are presented. As a target, we selected human aldose reductase, an enzyme involved in the development of late-stage diabetic complications. To recognize a large scope of substrate molecules, this reductase opens a transient specificity pocket. The pocket-opening step was studied by X-ray crystallography, microcalorimetry, and surface plasmon resonance using a narrow series of 2-carbamoyl-phenoxy-acetic acid derivatives. Molecular dynamics simulations suggest that pocket opening occurs only once an appropriate substituent is attached to the parent scaffold. Transient pocket opening of the uncomplexed protein is hardly recorded. Hydration-site analysis suggests that up to five water molecules entering the opened pocket cannot stabilize this state. Sole substitution with a benzyl group stabilizes the opened state, and the energetic barrier for opening is estimated to be ∼5 kJ/mol. Additional decoration of the pocket-opening benzyl substituent with a nitro group results in a huge enthalpy-driven potency increase; on the other hand, an isosteric carboxylic acid group reduces the potency 1000-fold, and binding occurs without pocket opening. We suggest a ligand induced-fit mechanism for the pocket-opening step, which, however, does not represent the rate-determining step in binding kinetics.

No abstract

To explore the variability in biosensor studies, 150 participants from 20 countries were given the same protein samples and asked to determine kinetic rate constants for the interaction. We chose a protein system that was amenable to analysis using different biosensor platforms as well as by users of different expertise levels. The two proteins (a 50-kDa Fab and a 60-kDa glutathione S-transferase [GST] antigen) form a relatively high-affinity complex, so participants needed to optimize several experimental parameters, including ligand immobilization and regeneration conditions as well as analyte concentrations and injection/dissociation times. Although most participants collected binding responses that could be fit to yield kinetic parameters, the quality of a few data sets could have been improved by optimizing the assay design. Once these outliers were removed, the average reported affinity across the remaining panel of participants was 620 pM with a standard deviation of 980 pM. These results demonstrate that when this biosensor assay was designed and executed appropriately, the reported rate constants were consistent, and independent of which protein was immobilized and which biosensor was used.