Quantum-Chemical Quasi-Docking for Molecular Dynamics Calculations

The quantum quasi-docking procedure is used to compare the docking accuracies of two quantum-chemical semiempirical methods, namely, PM6-D3H4X and PM7. Quantum quasi-docking is an approximation to quantum docking. In quantum docking, it is necessary to search directly for the global minimum of the energy of the protein-ligand complex calculated by the quantum-chemical method. In quantum quasi-docking, firstly, we look for a wide spectrum of low-energy minima, calculated using the MMFF94 force field, and secondly, we recalculate the energies of all these minima using the quantum-chemical method, and among these recalculated energies we determine the lowest energy and the corresponding ligand position. Both PM6-D3H4X and PM7 are novel methods that describe well-dispersion interactions, hydrogen and halogen bonds. The PM6-D3H4X and PM7 methods are used with the COSMO implicit solvent model as it is implemented in the MOPAC program. The comparison is made for 25 high quality protein-ligand complexes. Firstly, the docking positioning accuracies have been compared, and we demonstrated that PM7+COSMO provides better positioning accuracy than PM6-D3H4X. Secondly, we found that PM7+COSMO demonstrates a much higher correlation between the calculated and measured protein–ligand binding enthalpies than PM6-D3H4X. For future quantum docking PM7+COSMO is preferable, but the COSMO model must be improved.

Theoretical Researches On Binding Modes and Stability of Hydrogen Bonds Between Uracil and Formic Acid

The hydrogen bond formation with formic acid would affect the complementary pair of bases between uracil and adenine, but the binding modes and spectral properties of hydrogen bonds are still obscure. Density functional theory and time-dependent density functional theory were applied to investigate the intermolecular hydrogen bonds between uracil and formic acid. The reduced density gradient (RDG), bond lengths and vibration absorption frequencies revealed that the most probable uracil-formic acid (U-FA) interaction mode formed in the position c of FA and the site 1 of U, that is, the mode 1c. The theoretical parameters in excited state complexes manifested that the variety of hydrogen bond configurations led to different degrees of strengthening or weakening of molecular interaction. In the implicit solvent (water), the formations of O-H∙∙∙O in the uracil-formic acid complexes were promoted obviously. These theoretical studies would positively affect the researches of life science and medicinal chemistry.

Computational Study of Catalytic Urethane Formation

Polyurethanes (PUs) are widely used in different applications, and thus various synthetic procedures including one or more catalysts are applied to prepare them. For PU foams, the most important catalysts are nitrogen-containing compounds. Therefore, in this work, the catalytic effect of eight different nitrogen-containing catalysts on urethane formation will be examined. The reactions of phenyl isocyanate (PhNCO) and methanol without and in the presence of catalysts have been studied and discussed using the G3MP2BHandHLYP composite method. The solvent effects have also been considered by applying the SMD implicit solvent model. A general urethane formation mechanism has been proposed without and in the presence of the studied catalysts. The proton affinities (PA) were also examined. The barrier height of the reaction significantly decreased (∆E0 > 100 kJ/mol) in the presence of the studied catalysts, which proves the important effect they have on urethane formation. The achieved results can be applied in catalyst design and development in the near future.

Ergodic Algorithmic Model (EAM), with Water as Implicit Solvent, in Chemical, Biochemical, and Biological Processes

For many years, we have devoted our research to the study of the thermodynamic properties of hydrophobic hydration processes in water, and we have proposed the Ergodic Algorithmic Model (EAM) for maintaining the thermodynamic properties of any hydrophobic hydration reaction at a constant pressure from the experimental determination of an equilibrium constant (or other potential functions) as a function of temperature. The model has been successfully validated by the statistical analysis of the information elements provided by the EAM model for about fifty compounds. The binding functions are convoluted functions, RlnKeq = {f(1/T)* g(T)} and RTlnKeq = {f(T)* g(lnT)}, where the primary linear functions f(1/T) and f(T) are modified and transformed into parabolic curves by the secondary functions g(T) and g(lnT), respectively. Convoluted functions are consistent with biphasic dual-structure partition function, {DS-PF} = {M-PF} ∙ {T-PF} ∙ {ζw}, composed by ({M-PF} (Density Entropy), {T-PF}) (Intensity Entropy), and {ζw} (implicit solvent). In the present paper, after recalling the essential aspects of the model, we outline the importance of considering the solvent as “implicit” in chemical and biochemical reactions. Moreover, we compare the information obtained by computer simulations using the models till now proposed with “explicit” solvent, showing the mess of information lost without considering the experimental approach of the EAM model.

Accurate Description of Protein–Protein Recognition and Protein Aggregation with the Implicit-Solvent-Based PACSAB Protein Model

We used the PACSAB protein model, based on the implicit solvation approach, to simulate protein–protein recognition and study the effect of helical structure on the association of aggregating peptides. After optimization, the PACSAB force field was able to reproduce correctly both the correct binding interface in ubiquitin dimerization and the conformational ensemble of the disordered protein activator for hormone and retinoid receptor (ACTR). The PACSAB model allowed us to predict the native binding of ACTR with its binding partner, reproducing the refolding upon binding mechanism of the disordered protein.

Implicit solvent models fail to reproduce secondary structures of de novo designed peptides

We test a range of standard implicit solvent models and protein forcefields for a set of 5 experimentally characterized, designed α-helical peptides. 65 combinations of forcefield and implicit solvent models are evaluated in >800 µs of molecular dynamics simulations. The data show that implicit solvent models generally fail to reproduce the experimentally observed secondary structure content, and none performs well for all 5 peptides. The results show that these models are not usefully predictive.

Mining the Protein Data Bank to improve prediction of changes in protein-protein binding

Predicting the effect of mutations on protein-protein interactions is important for relating structure to function, as well as for in silico affinity maturation. The effect of mutations on protein-protein binding energy (ΔΔG) can be predicted by a variety of atomic simulation methods involving full or limited flexibility, and explicit or implicit solvent. Methods which consider only limited flexibility are naturally more economical, and many of them are quite accurate, however results are dependent on the atomic coordinate set used. In this work we perform a sequence and structure based search of the Protein Data Bank to find additional coordinate sets and repeat the calculation on each. The method increases precision and Positive Predictive Value, and decreases Root Mean Square Error, compared to using single structures. Given the ongoing growth of near-redundant structures in the Protein Data Bank, our method will only increase in applicability and accuracy.

Bridging the experiment-calculation divide: machine learning corrections to redox potential calculations in implicit and explicit solvent models

Prediction of redox potentials is essential for catalysis and energy storage. Although density functional theory (DFT) calculations have enabled rapid redox potential predictions for numerous compounds, prominent errors persist compared to experimental measurements. In this work, we develop machine learning (ML) models to reduce the errors of redox potential calculations in both implicit and explicit solvent models. Training and testing of the ML correction models are based on the diverse ROP313 dataset with experimental redox potentials measured for organic and organometallic compounds in a variety of solvents. For the implicit solvent approach, our ML models can reduce both the systematic bias and the number of outliers. ML corrected redox potentials also demonstrate less sensitivity to DFT functional choice. For the explicit solvent approach, we significantly reduce the computational costs by embedding the microsolvated cluster in implicit bulk solvent, obtaining converged redox potential results with a smaller solvation shell. This combined implicit-explicit solvent model, together with GPU-accelerated quantum chemistry methods, enabled rapid generation of a large dataset of explicit-solvent-calculated redox potentials for 165 organic compounds, allowing detailed investigation of the error sources in explicit solvent redox potential calculations.

Bridging the experiment-calculation divide: machine learning corrections to redox potential calculations in implicit and explicit solvent models

Prediction of redox potentials is essential for catalysis and energy storage. Although density functional theory (DFT) calculations have enabled rapid redox potential predictions for numerous compounds, prominent errors persist compared to experimental measurements. In this work, we develop machine learning (ML) models to reduce the errors of redox potential calculations in both implicit and explicit solvent models. Training and testing of the ML correction models are based on the diverse ROP313 dataset with experimental redox potentials measured for organic and organometallic compounds in a variety of solvents. For the implicit solvent approach, our ML models can reduce both the systematic bias and the number of outliers. ML corrected redox potentials also demonstrate less sensitivity to DFT functional choice. For the explicit solvent approach, we significantly reduce the computational costs by embedding the microsolvated cluster in implicit bulk solvent, obtaining converged redox potential results with a smaller solvation shell. This combined implicit-explicit solvent model, together with GPU-accelerated quantum chemistry methods, enabled rapid generation of a large dataset of explicit-solvent-calculated redox potentials for 165 organic compounds, allowing detailed investigation of the error sources in explicit solvent redox potential calculations.