Accurately Predicting Barrier Heights for Radical Reactions in Solution Using Deep Graph Networks.

Quantitative estimates of reaction barriers and solvent effects are essential for developing kinetic mechanisms and predicting reaction outcomes. Here, we create a new data set of 5,600 unique elementary radical reactions calculated using the M06-2X/def2-QZVP//B3LYP-D3(BJ)/def2-TZVP level of theory. A conformer search is done for each species using TPSS/def2-TZVP. Gibbs free energies of activation and of reaction for these radical reactions in 40 common solvents are obtained using COSMO-RS for solvation effects. These balanced reactions involve the elements H, C, N, O, and S, contain up to 19 heavy atoms, and have atom-mapped SMILES. All transition states are verified by an intrinsic reaction coordinate calculation. We next train a deep graph network to directly estimate the Gibbs free energy of activation and of reaction in both gas and solution phases using only the atom-mapped SMILES of the reactant and product and the SMILES of the solvent. This simple input representation avoids computationally expensive optimizations for the reactant, transition state, and product structures during inference, making our model well-suited for high-throughput predictive chemistry and quickly providing information for (retro-)synthesis planning tools. To properly measure model performance, we report results on both interpolative and extrapolative data splits and also compare to several baseline models. During training and testing, the data set is augmented by including the reverse direction of each reaction and variants with different resonance structures. After data augmentation, we have around 2 million entries to train the model, which achieves a testing set mean absolute error of 1.16 kcal mol-1 for the Gibbs free energy of activation in solution. We anticipate this model will accelerate predictions for high-throughput screening to quickly identify relevant reactions in solution, and our data set will serve as a benchmark for future studies.

Automated input structure generation for single-ended reaction path optimizations.

The exploration of a reaction network requires highly automated workflows to avoid error-prone and time-consuming manual steps. In this respect, a major bottleneck is the search for transition-state (TS) structures, which frequently fails and, therefore, makes (manual) revision necessary. In this work, we present a technique for obtaining suitable input structures for automated TS searches based on single-ended reaction path optimization algorithms, which makes subsequent TS searches via this method significantly more robust. First, possible input structures are generated based on the spatial alignment of the reactants. The appropriate orientation of reacting groups is achieved via stepwise rotations along selected torsional degrees of freedom. Second, a ranking of the obtained structures is performed according to selected geometric criteria. The main goals are to properly align the reactive atoms, to avoid hindrance within the reaction channel and to resolve steric clashes between the reactants. The developed procedure has been carefully tested on a variety of examples and provides suitable input structures for TS searches within seconds. The method is in daily use in an industrial setting.

Efficient switching of mCherry fluorescence using chemical caging.

Fluorophores with dynamic or controllable fluorescence emission have become essential tools for advanced imaging, such as superresolution imaging. These applications have driven the continuing development of photoactivatable or photoconvertible labels, including genetically encoded fluorescent proteins. These new probes work well but require the introduction of new labels that may interfere with the proper functioning of existing constructs and therefore require extensive functional characterization. In this work we show that the widely used red fluorescent protein mCherry can be brought to a purely chemically induced blue-fluorescent state by incubation with ß-mercaptoethanol (ßME). The molecules can be recovered to the red fluorescent state by washing out the ßME or through irradiation with violet light, with up to 80% total recovery. We show that this can be used to perform single-molecule localization microscopy (SMLM) on cells expressing mCherry, which renders this approach applicable to a very wide range of existing constructs. We performed a detailed investigation of the mechanism underlying these dynamics, using X-ray crystallography, NMR spectroscopy, and ab initio quantum-mechanical calculations. We find that the ßME-induced fluorescence quenching of mCherry occurs both via the direct addition of ßME to the chromophore and through ßME-mediated reduction of the chromophore. These results not only offer a strategy to expand SMLM imaging to a broad range of available biological models, but also present unique insights into the chemistry and functioning of a highly important class of fluorophores.

QM/MM-Based Calculations of Absorption and Emission Spectra of LSSmOrange Variants.

The goal of this computational work is to gain new insight into the photochemistry of the fluorescent protein (FP) LSSmOrange. This FP is of interest because besides exhibiting the eponymous large spectral shift (LSS) between the absorption and emission energies, it has been experimentally observed that it can also undergo a photoconversion process, which leads to a change in the absorption wavelength of the chromophore (from 437 to 553 nm). There is strong experimental evidence that this photoconversion is caused by decarboxylation of a glutamate located in the close vicinity of the chromophore. Still, the exact chemical mechanism of the decarboxylation process as well as the precise understanding of structure-property relations in the measured absorption and emission spectra is not yet fully understood. Therefore, hybrid quantum mechanics/molecular mechanics (QM/MM) calculations are performed to model the absorption and emission spectra of the original and photoconverted forms of LSSmOrange. The necessary force-field parameters of the chromophore are optimized with CGenFF and the FFToolkit. A thorough analysis of QM methods to study the excitation energies of this specific FP chromophore has been carried out. Furthermore, the influence of the size of the QM region has been investigated. We found that QM/MM calculations performed with time-dependent density functional theory (CAM-B3LYP/D3/6-31G*) and QM calculations performed with the semiempirical ZIndo/S method including a polarizable continuum model can describe the excitation energies reasonably well. Moreover, already a small QM region size seems to be sufficient for the study of the photochemistry in LSSmOrange. Especially, the calculated ZIndo spectra are in very good agreement with the experimental ones. On the basis of the spectra obtained, we could verify the experimentally assigned structures.

Heuristics-Guided Exploration of Reaction Mechanisms.

For the investigation of chemical reaction networks, the efficient and accurate determination of all relevant intermediates and elementary reactions is mandatory. The complexity of such a network may grow rapidly, in particular if reactive species are involved that might cause a myriad of side reactions. Without automation, a complete investigation of complex reaction mechanisms is tedious and possibly unfeasible. Therefore, only the expected dominant reaction paths of a chemical reaction network (e.g., a catalytic cycle or an enzymatic cascade) are usually explored in practice. Here, we present a computational protocol that constructs such networks in a parallelized and automated manner. Molecular structures of reactive complexes are generated based on heuristic rules derived from conceptual electronic-structure theory and subsequently optimized by quantum-chemical methods to produce stable intermediates of an emerging reaction network. Pairs of intermediates in this network that might be related by an elementary reaction according to some structural similarity measure are then automatically detected and subjected to an automated search for the connecting transition state. The results are visualized as an automatically generated network graph, from which a comprehensive picture of the mechanism of a complex chemical process can be obtained that greatly facilitates the analysis of the whole network. We apply our protocol to the Schrock dinitrogen-fixation catalyst to study alternative pathways of catalytic ammonia production.

Mode-tracking based stationary-point optimization.

In this work, we present a transition-state optimization protocol based on the Mode-Tracking algorithm [Reiher and Neugebauer, J. Chem. Phys., 2003, 118, 1634]. By calculating only the eigenvector of interest instead of diagonalizing the full Hessian matrix and performing an eigenvector following search based on the selectively calculated vector, we can efficiently optimize transition-state structures. The initial guess structures and eigenvectors are either chosen from a linear interpolation between the reactant and product structures, from a nudged-elastic band search, from a constrained-optimization scan, or from the minimum-energy structures. Alternatively, initial guess vectors based on chemical intuition may be defined. We then iteratively refine the selected vectors by the Davidson subspace iteration technique. This procedure accelerates finding transition states for large molecules of a few hundred atoms. It is also beneficial in cases where the starting structure is very different from the transition-state structure or where the desired vector to follow is not the one with lowest eigenvalue. Explorative studies of reaction pathways are feasible by following manually constructed molecular distortions.

A stable phosphanyl phosphaketene and its reactivity.

Sodium phosphaethynolate, Na(OCP), reacts with the bulky P-chloro-diazaphosphole yielding a phosphanyl phosphaketene, which is stable for weeks under an inert atmosphere in the solid state. This compound is best described as a tight ion pair with a remarkably long P-P bond distance (2.44 Å). In solution, this phosphaketene dimerizes under loss of CO to give 1,2,3-triphosphabicyclobutane identified by an X-ray diffraction study. As an intermediate, a five-membered heterocyclic diphosphene was trapped in a Diels-Alder reaction with 2,3-dimethylbutadiene. The formation of this intermediate in a hetero-Cope-rearrangement as well as dimerization/CO loss were computed with various DFT methods which allowed us to understand the reaction mechanisms.

Facile synthesis and theoretical conformation analysis of a triazine-based double-decker rotor molecule with three anthracene blades.

The facile synthesis of a rotor-shaped compound with two stacked triazine units, which are symmetrically connected by three anthracene blades through oxygen linkers, is presented. This new double-decker, which is a potential monomer for two-dimensional polymerization, was synthesized by using readily available, cheap building blocks, exploiting the known selectivity difference for the nucleophilic substitution of cyanuric chloride. The crystal structure of a C3h symmetric rotor-shaped compound with 9,10-dihydroanthracene blades, which is a direct precursor to the targeted monomer, and the crystal structure of the new double-decker with the desired C3h symmetry, are also reported. The synthetic efforts were preceded by a computational analysis, which was triggered by the question of conformational stability of the potential monomer. Two stable conformers could be found, and the barrier for the transition path in the gas phase between these conformers was determined by quantum chemical calculations. Exploratory Born-Oppenheimer molecular-dynamics simulations revealed a strong influence of solvent-solute interactions on the stability of the conformers, which resulted in an energetic preference of the C3h symmetric conformation of the double-decker.

Structure-Property Relationships of Fe4 S4 Clusters.

In this theoretical study, the sensitivity of Fe4 S4 cluster properties, such as potential energy, spin coupling, adiabatic detachment energy, inner-sphere reorganization energy, and reactivity, to structural distortions is investigated. [Fe4 S4 (SH)4 ]3-/2-/1- model clusters anchored by fixed hydrogen atoms are compared with Fe4 S4 clusters coordinated by ethyl thiolates with fixations according to cysteine residues in crystal structures. For the model system, a dependence of the ground-state spin-coupling scheme on the hydrogen-hydrogen distances is observed. The minima of the potential energy surface of [Fe4 S4 (SH)4 ]2-/1- clusters are located at slightly smaller hydrogen-hydrogen distances than those of the [Fe4 S4 (SH)4 ]3- cluster. For inner-sphere reorganization energies the spin-coupling scheme adopted by the broken-symmetry wave function plays an important role, since it can change the reorganization energies by up to 13 kcal mol-1 . For most structures, [Fe4 S4 (SR)4 ]2- and [Fe4 S4 (SR)4 ]1- (R=H or ethyl, derived from cysteine) favor the same coupling scheme. Therefore, the reorganization energies for this redox couple are relatively low (6-12 kcal mol-1 ) compared with the 2-/3- redox couple favoring different spin-coupling schemes before and after electron transfer (14-18 kcal mol-1 ). However, one may argue that more reliable reorganization energies are obtained if always the same spin-coupling pattern is enforced. All theoretical observations and insights are discussed in the light of experimental results distilled from the literature.

Current computer modeling cannot explain why two highly similar sequences fold into different structures.

The remarkable recent creation of two proteins that fold into two completely different and stable structures, exhibit different functions, yet differ by only a few amino acids poses a conundrum to those hoping to understand how sequence encodes structure. Here, computer modeling uniquely allows the characterization of not only the native structure of each minimally different sequence but also systems in which each sequence was modeled onto the fold of the alternate sequence. The reasons for the different structural preferences of two pairs of highly similar sequences are explored by a combination of structure analyses, comparison of potential energies calculated from energy-minimized single structures and trajectories produced from molecular dynamics simulations, and application of a novel method for calculating free energy differences. The sensitivity of such analyses to the choice of force field is also explored. Many of the hypotheses proposed on the basis of the nuclear magnetic resonance model structures of the proteins with 95% identical sequences are supported. However, each level of analysis provides different predictions regarding which sequence-structure combination should be most favored, highlighting the fact that protein structure and stability result from a complex combination of interdependent factors.