SophosQM: Accurate Binding Affinity Prediction in Compound Optimization.

The optimization of compounds' binding affinity for a biological target is a crucial aspect of the drug development process. Being able to accurately predict binding energies in advance of synthesizing compounds would have a massive impact on the speed of the drug discovery process. The ideal binding affinity prediction method should combine accuracy, reliability, and speed. In this paper, we present SophosQM, a quantum mechanics (QM)-based approach, which can accurately predict the binding affinities of compounds to proteins. The binding affinity predictive models generated by SophosQM are based on the fragment molecular orbital (FMO) method to estimate the enthalpic component of the binding free energy, and a macroscopic descriptor, clog P, is used as an approximation of the entropic component. The affinity prediction is performed using multilinear regression, fitting the experimental values against the FMO-computed enthalpic term and clog P. The quality of the prediction can be assessed in terms of the correlation coefficient between experimental and predicted values. In this work, the method's reliability and accuracy are exemplified by applying SophosQM to 70 compounds binding to six different targets of pharmaceutical relevance. Overall, the results show a very satisfactory performance with a global correlation coefficient in the order of 0.9. Our predictions also show a satisfactory performance compared to data based on free energy perturbation. Finally, SophosQM can also be applied in high-throughput mode by using semiempirical QM methods to evaluate large portions of chemical space, while retaining a good level of accuracy, but decreasing the computing time to just a few seconds per compound.

Evaluation of 3- and 4-Phenoxybenzamides as Selective Inhibitors of the Mono-ADP-Ribosyltransferase PARP10.

Intracellular ADP-ribosyltransferases catalyze mono- and poly-ADP-ribosylation and affect a broad range of biological processes. The mono-ADP-ribosyltransferase PARP10 is involved in signaling and DNA repair. Previous studies identified OUL35 as a selective, cell permeable inhibitor of PARP10. We have further explored the chemical space of OUL35 by synthesizing and investigating structurally related analogs. Key synthetic steps were metal-catalyzed cross-couplings and functional group modifications. We identified 4-(4-cyanophenoxy)benzamide and 3-(4-carbamoylphenoxy)benzamide as PARP10 inhibitors with distinct selectivities. Both compounds were cell permeable and interfered with PARP10 toxicity. Moreover, both revealed some inhibition of PARP2 but not PARP1, unlike clinically used PARP inhibitors, which typically inhibit both enzymes. Using crystallography and molecular modeling the binding of the compounds to different ADP-ribosyltransferases was explored regarding selectivity. Together, these studies define additional compounds that interfere with PARP10 function and thus expand our repertoire of inhibitors to further optimize selectivity and potency.

Nintedanib targets KIT D816V neoplastic cells derived from induced pluripotent stem cells of systemic mastocytosis.

The KIT D816V mutation is found in >80% of patients with systemic mastocytosis (SM) and is key to neoplastic mast cell (MC) expansion and accumulation in affected organs. Therefore, KIT D816V represents a prime therapeutic target for SM. Here, we generated a panel of patient-specific KIT D816V induced pluripotent stem cells (iPSCs) from patients with aggressive SM and mast cell leukemia to develop a patient-specific SM disease model for mechanistic and drug-discovery studies. KIT D816V iPSCs differentiated into neoplastic hematopoietic progenitor cells and MCs with patient-specific phenotypic features, thereby reflecting the heterogeneity of the disease. CRISPR/Cas9n-engineered KIT D816V human embryonic stem cells (ESCs), when differentiated into hematopoietic cells, recapitulated the phenotype observed for KIT D816V iPSC hematopoiesis. KIT D816V causes constitutive activation of the KIT tyrosine kinase receptor, and we exploited our iPSCs and ESCs to investigate new tyrosine kinase inhibitors targeting KIT D816V. Our study identified nintedanib, a US Food and Drug Administration-approved angiokinase inhibitor that targets vascular endothelial growth factor receptor, platelet-derived growth factor receptor, and fibroblast growth factor receptor, as a novel KIT D816V inhibitor. Nintedanib selectively reduced the viability of iPSC-derived KIT D816V hematopoietic progenitor cells and MCs in the nanomolar range. Nintedanib was also active on primary samples of KIT D816V SM patients. Molecular docking studies show that nintedanib binds to the adenosine triphosphate binding pocket of inactive KIT D816V. Our results suggest nintedanib as a new drug candidate for KIT D816V-targeted therapy of advanced SM.

Structural Prediction of the Dimeric Form of the Mammalian Translocator Membrane Protein TSPO: A Key Target for Brain Diagnostics.

Positron emission tomography (PET) radioligands targeting the human translocator membrane protein (TSPO) are broadly used for the investigations of neuroinflammatory conditions associated with neurological disorders. Structural information on the mammalian protein homodimers-the suggested functional state of the protein-is limited to a solid-state nuclear magnetic resonance (NMR) study and to a model based on the previously-deposited solution NMR structure of the monomeric mouse protein. Computational studies performed here suggest that the NMR-solved structure in the presence of detergents is not prone to dimer formation and is furthermore unstable in its native membrane environment. We, therefore, propose a new model of the functionally-relevant dimeric form of the mouse protein, based on a prokaryotic homologue. The model, fully consistent with solid-state NMR data, is very different from the previous predictions. Hence, it provides, for the first time, structural insights into this pharmaceutically-important target which are fully consistent with experimental data.

Electrostatic versus Resonance Interactions in Photoreceptor Proteins: The Case of Rhodopsin.

Light sensing in photoreceptor proteins is subtly modulated by the multiple interactions between the chromophoric unit and its binding pocket. Many theoretical and experimental studies have tried to uncover the fundamental origin of these interactions but reached contradictory conclusions as to whether electrostatics, polarization, or intrinsically quantum effects prevail. Here, we select rhodopsin as a prototypical photoreceptor system to reveal the molecular mechanism underlying these interactions and regulating the spectral tuning. Combining a multireference perturbation method and density functional theory with a classical but atomistic and polarizable embedding scheme, we show that accounting for electrostatics only leads to a qualitatively wrong picture, while a responsive environment can successfully capture both the classical and quantum dominant effects. Several residues are found to tune the excitation by both differentially stabilizing ground and excited states and through nonclassical "inductive resonance" interactions. The results obtained with such a quantum-in-classical model are validated against both experimental data and fully quantum calculations.

Introducing QMC/MMpol: Quantum Monte Carlo in Polarizable Force Fields for Excited States.

We present for the first time a quantum mechanics/molecular mechanics scheme which combines quantum Monte Carlo with the reaction field of classical polarizable dipoles (QMC/MMpol). In our approach, the optimal dipoles are self-consistently generated at the variational Monte Carlo level and then used to include environmental effects in diffusion Monte Carlo. We investigate the performance of this hybrid model in describing the vertical excitation energies of prototypical small molecules solvated in water, namely, methylenecyclopropene and s-trans acrolein. Two polarization regimes are explored where either the dipoles are optimized with respect to the ground-state solute density (polGS) or different sets of dipoles are separately brought to equilibrium with the states involved in the electronic transition (polSS). By comparing with reference supermolecular calculations where both solute and solvent are treated quantum mechanically, we find that the inclusion of the response of the environment to the excitation of the solute leads to superior results than the use of a frozen environment (point charges or polGS), in particular, when the solute-solvent coupling is dominated by electrostatic effects which are well recovered in the polSS condition. QMC/MMpol represents therefore a robust scheme to treat important environmental effects beyond static point charges, combining the accuracy of QMC with the simplicity of a classical approach.

Solvent Effects on Excited-State Structures: A Quantum Monte Carlo and Density Functional Study.

We present the first application of quantum Monte Carlo (QMC) in its variational flavor combined with the polarizable continuum model (PCM) to perform excited-state geometry optimization in solution. Our implementation of the PCM model is based on a reaction field that includes both volume and surface polarization charges and is determined self-consistently with the molecular wave function during the QMC optimization of the solute geometry. For acrolein, acetone, methylenecyclopropene, and the propenoic acid anion, we compute the optimal exited-state geometries in water and compare our results with the structures obtained with second-order perturbation theory (CASPT2) and other correlated methods, and with time-dependent density functional theory (TDDFT). We find that QMC predicts a structural response to solvation in good agreement with CASPT2 with the only exception of the π → π* state of acrolein where the robustness of the QMC geometry must be contrasted to the sensitivity of the perturbation result to the details of the calculation. As regards TDDFT, we show that all investigated functionals systematically overestimate the geometrical changes from the gas phase to solution, sometimes giving bond variations opposite in trend to QMC.

Ground- and Excited-State Geometry Optimization of Small Organic Molecules with Quantum Monte Carlo.

We present a comparative study of the geometry optimization in the gas phase of acrolein, acetone, methylenecyclopropene, and the propenoic acid anion with special emphasis on their excited-state structures, using quantum Monte Carlo (QMC), multireference perturbation theory (CASPT2 and NEVPT2), second-order approximate coupled cluster (CC2), and time-dependent density functional theory (TDDFT). We find that, for all molecules, the geometries optimized with QMC in its simplest variational (VMC) flavor are in very good agreement with the perturbation results both in the ground and the excited states of either n → π* or π → π* character. Furthermore, the quality of the QMC structures is superior to those obtained with the CC2 method, which overestimates the CO bond in all n → π* excitations, or to the symmetry-adapted-cluster configuration interaction (SAC-CI) approach, which gives a poorer description of the CC bonds in the excited states. Finally, the spread in the TDDFT structures obtained with several current exchange-correlation functionals is large and does not reveal a clear relation between the defining features of the functionals and the quality of the optimized structures. In summary, our findings demonstrate the good performance of QMC in optimizing the geometries of these molecules, also in cases where other correlated or TDDFT approaches are inaccurate, and indicate that the method represents a robust reference approach for future structural studies also of larger systems.