Machine learning-based prediction of enzyme substrate scope: Application to bacterial nitrilases.

Predicting the range of substrates accepted by an enzyme from its amino acid sequence is challenging. Although sequence- and structure-based annotation approaches are often accurate for predicting broad categories of substrate specificity, they generally cannot predict which specific molecules will be accepted as substrates for a given enzyme, particularly within a class of closely related molecules. Combining targeted experimental activity data with structural modeling, ligand docking, and physicochemical properties of proteins and ligands with various machine learning models provides complementary information that can lead to accurate predictions of substrate scope for related enzymes. Here we describe such an approach that can predict the substrate scope of bacterial nitrilases, which catalyze the hydrolysis of nitrile compounds to the corresponding carboxylic acids and ammonia. Each of the four machine learning models (logistic regression, random forest, gradient-boosted decision trees, and support vector machines) performed similarly (average ROC = 0.9, average accuracy = ~82%) for predicting substrate scope for this dataset, although random forest offers some advantages. This approach is intended to be highly modular with respect to physicochemical property calculations and software used for structural modeling and docking.

The taxonomic distribution of histamine-secreting bacteria in the human gut microbiome.

BACKGROUND: Biogenic histamine plays an important role in immune response, neurotransmission, and allergic response. Although endogenous histamine production has been extensively studied, the contributions of histamine produced by the human gut microbiota have not been explored due to the absence of a systematic annotation of histamine-secreting bacteria. RESULTS: To identify the histamine-secreting bacteria from in the human gut microbiome, we conducted a systematic search for putative histamine-secreting bacteria in 36,554 genomes from the Genome Taxonomy Database and Unified Human Gastrointestinal Genome catalog. Using bioinformatic approaches, we identified 117 putative histamine-secreting bacteria species. A new three-component decarboxylation system including two colocalized decarboxylases and one transporter was observed in histamine-secreting bacteria among three different phyla. We found significant enrichment of histamine-secreting bacteria in patients with inflammatory bowel disease but not in patients with colorectal cancer suggesting a possible association between histamine-secreting bacteria and inflammatory bowel disease. CONCLUSIONS: The findings of this study expand our knowledge of the taxonomic distribution of putative histamine-secreting bacteria in the human gut.

Mechanistic Investigation of Dimethylmercury Formation Mediated by a Sulfide Mineral Surface.

Mercury (Hg) pollution is a global environmental problem. The abiotic formation of dimethylmercury (DMeHg) from monomethylmercury (MMeHg) may account for a large portion of DMeHg in oceans. Previous experimental work has shown that abiotic formation of DMeHg from MMeHg can be facilitated by reduced sulfur groups on sulfide mineral surfaces. In that work, a mechanism was proposed in which neighboring MMeHg moieties bound to sulfide sites on a mineral surface react through an SN2-type mechanism to form DMeHg and incorporate the remaining Hg atoms into the mineral surface. Here, we perform density functional theory calculations to explore the mechanisms of DMeHg formation on the 110 surface of a CdS(s) (hawleyite) nanoparticle. We show that coordination of MMeHg substituents to adjacent reduced sulfur groups protruding from the surface indeed facilitates DMeHg formation and that the reaction proceeds through direct transmethylation from one MMeHg substituent to another. Coordination of Hg by multiple S atoms provides a transition-state stabilization and activates a C-Hg bond for methyl transfer. In addition, solvation effects play an important role in the surface reconstruction of the nanoparticle and in decreasing the energetic barrier for DMeHg formation relative to the corresponding reaction in vacuo.

Origin of the Isomer Stability of Polymethylated DOTA Chelates Complexed with Ln3+ ions.

DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid)-based chelates that give only a single isomer in solution when complexed with lanthanide (Ln3+) ions is of value for studying protein dynamics and interactions via NMR. Herein, we have investigated the geometries, energetics, and electrostatic potentials of Lu complexed with DOTA (1), ring methylated M4DOTA (2), and arm methylated R-DOTMA (3) and S-DOTMA (4), as well as, both ring and arm methylated 4S-4S-M4DOTMA (5) and 4S-4R-M4DOTMA (6) at the level of M06-L/6-31+G(d)-SDD, to elucidate the origin of the isomer stability. These analyses indicate that the electrostatic repulsion between the arm methyl and the neighboring carboxylate significantly destabilizes the square antiprism (SAP) isomer of Lu-5 and the twisted square antiprism (TSAP) isomer of Lu-6, while the steric repulsion between the ring and arm methyl groups attenuates the stability of both TSAP of Lu-5 and SAP of Lu-6. To rationalize the variable temperature proton NMR spectra, the energy barriers for the inter-conversion in Lu-5 and Lu-6 via arm rotation were also calculated. The modulation of the stability and rigidity of Ln complexes via a modification of DOTA is also discussed. Our investigation will aid to design better chelates for the Ln3+ ions for its use in molecular medicine.

Sigma- versus Pi-Dimerization Modes of Triangulene.

We show that the diradicaloid triangulene, a graphene nano-flake molecule, can aggregate in a variety of dimerization modes. We found by density functional theory modeling a number of triangulene dimers including six doubly bonded σ-dimers in addition to the previously reported six pancake bonded π-dimer isomers. The σ-dimers display a wide range of stabilities: the interaction energy of the most stable σ-dimer is -25.17 kcal mol-1 . Besides the doubly bonded σ-dimers with closed shell ground states, we also found an open-shell singly σ-bonded diradicaloid dimer. We found an interesting isomerization route between a doubly bonded σ-dimer, a singly bonded σ-dimer with a low-lying triplet state and two π-bonded dimers with low-lying quintet states. Derivatives of triangulene, trioxo-triangulenes (TOTs) have been previously characterized experimentally. Here, we show the reasons why so far only the π-dimer but not the σ-dimer was experimentally observed for all TOTs.

Pancake Bonding in π-Stacked Trimers in a Salt of Tetrachloroquinone Anion.

The crystal structure of [4-damp])2 [Cl4 Q]3 (4-damp=4-dimethylamino-N-methylpyridinium, Cl4 Q=tetrachloroquinone) salt is built up from slipped columnar stacks of quinoid rings composed of closely bound trimers with the intra-trimer separation distance of 2.84 Šand total charge of -2 whereas the inter-trimer distance is 3.59 Å. The individual rings exhibit partial negative charges that are distributed unevenly among the three Cl4 Qs in the trimer. The strong interactions within a trimer (Cl4 Q)32- have a partially covalent character with two-electron/multicentered bonding, that is extended over three rings, plausibly termed as "pancake bonding". The electron pairing within this multicentre bond leads to the fact that the crystals are diamagnetic and act as insulators. The studies of the structure and nature of bonding are based on X-ray charge density analysis and density functional theory.

Cethrene: The Chameleon of Woodward-Hoffmann Rules.

We demonstrate that the electrocyclic (EC) ring-closure of cethrene in solution proceeds in a conrotatory mode both thermally and photochemically. The facile photochemical EC process promises that cethrene can serve as an efficient chiroptical switch operated solely by light. As for the thermally activated EC reaction, a low reaction barrier and a solvation effect on the EC rate indicate that the C2-symmetric pathway predicted by DFT calculations might not be the correct mechanism. Instead, we argue that the molecular symmetry decreases along the reaction coordinate as a consequence of the low-energy singlet excited state in this diradicaloid molecule, which might lead to a lower activation energy in accord with that determined through kinetic studies. Cethrene, therefore, represents a thought-provoking molecular chameleon of the Woodward-Hoffmann rules that puts our chemical concepts and intuition to test.

Validation of density functionals for pancake-bonded π-dimers; dispersion is not enough.

π-Stacking pancake bonding between radicals poses special challenges to density functional theories (DFTs) due to their shorter than van der Waals contact distances, their multireference singlet ground states and the concurrently important dispersion interactions. We examined over 50 DFTs including 22 with dispersion corrections on four different π-dimerized pancake-bonded systems exploring the performances of these DFTs in the very short intermolecular contact regime. We examined crucial energetic as well as geometric parameters against available high-level multireference average quadratic coupled cluster (MR-AQCC) results. Overall we did not find an omnipotent DFT applicable for all four pancake-bonded π-dimers. However, some DFTs were found to perform well for each individual system: M05-2X and PBE0-MBD are the only DFTs that work well for the phenalenyl π-dimer; BLYP is the only appropriate DFT for the 1,2,4,6-thiatriazine π-dimer; O3LYP works best on the double pancake-bonded 1,3,2,4,6-dithiatriazine π-dimer with a few acceptable ones and MN15L is the best method for K+TCNE- π-dimer in addition to a few acceptable ones. We believe our findings can deliver insights towards the design and characterization of the pancake-bond based materials and the development of new DFTs.

Pancake Bond Orders of a Series of π-Stacked Triangulene Radicals.

Conjugated radicals are capable of forming π-stacking "pancake-bonded" dimers. Members of the family of triangulene hydrocarbons, non-Kekulé neutral multiradicals, can utilize more than one singly occupied molecular orbital (SOMO) to form multiple pancake-bonded dimers with formal bond orders of up to five. The resulting dimer binding energies can be quite high and the intermolecular contacts rather small compared to the respective van der Waals values. The preferred configurations are driven by the large stabilization energy of overlapping SOMOs.

Fluxional σ-Bonds of the 2,5,8-Trimethylphenalenyl Dimer: Direct Observation of the Sixfold σ-Bond Shift via a π-Dimer.

Direct evidence for σ-bond fluxionality in a phenalenyl σ-dimer was successfully obtained by a detailed investigation of the solution-state dynamics of 2,5,8-trimethylphenalenyl (TMPLY) using both experimental and theoretical approaches. TMPLY formed three diamagnetic dimers, namely, the σ-dimer (RR/SS), σ-dimer (RS), and π-dimer, which were fully characterized by (1)H NMR spectroscopy and electronic absorption measurements. The experimental findings gave the first quantitative insights into the essential preference of these competitive and unusual dimerization modes. The spectroscopic analyses suggested that the σ-dimer (RR/SS) is the most stable in terms of energy, whereas the others are metastable; the energy differences between these three isomers are less than 1 kcal mol(-1). Furthermore, the intriguing dynamics of the TMPLY dimers in the solution state were fully revealed by means of (1)H-(1)H exchange spectroscopy (EXSY) measurements and variable-temperature (1)H NMR studies. Surprisingly, the σ-dimer (RR/SS) demonstrated a sixfold σ-bond shift between the six sets of α-carbon pairs. This unusual σ-bond fluxionality is ascribed to the presence of a direct interconversion pathway between the σ-dimer (RR/SS) and the π-dimer, which was unambiguously corroborated by the EXSY measurements. The proposed mechanism of the sixfold σ-bond shift based on the experimental findings was well-supported by theoretical calculations.

Hetero-π-Dimers of Phenalenyls.

Homogeneous π-stacking dimers of phenalenyl and its derivatives have gained tremendous interest as components of conducting organic materials. For the first time, we investigate theoretically heterogeneous phenalenyl π-dimers. Key parameters, including charge transfer, interaction energy, singly occupied molecular orbital (SOMO) energy, and spin density, are studied with the help of density functional theory. We find that the amount of charge transfer between the two monomers in phenalenyl π-dimers correlates with the difference in the SOMO energies of the constituent monomers, where the SOMO energy plays the role of a monomer (group) electronegativity index. Charge transfer plays an important role in stabilizing the heterodimers while maintaining a significant diradicaloid character. For five heterodimers the interaction energy is found to be as large as -30 to -50 kcal mol(-1) . The presented correlation between the monomer SOMO energy levels and their stability can provide a simple predictive tool to design new highly stable π-stacking heterodimers.

Evidence of σ- and π-dimerization in a series of phenalenyls.

Phenalenyl and a wide variety of its derivatives form stable radicals, which often associate in various aggregates with interesting properties that include magnetism and high electrical conductivity. The two main modes of aggregation involve π-stacking pancake multicenter bond formation and σ-bond formation. We explore the energetics of the various σ- and π-dimers for six phenalenyl derivatives with both computational and experimental methods. A modern density functional theory (M05-2X) is used to survey the potential energy surface revealing the mechanism of the aggregation. In order to enrich experimental data, the triphenyl and trimethyl derivatives are newly prepared and their aggregation behaviors are investigated by various analytical methods including ESR, (1)H NMR, UV-vis, and single-crystal X-ray diffraction. The agreement between computations and experiments are very good forming the basis of describing trends in this series. We find that π-dimer formation can proceed via an asynchronous concerted path from the monomers or in a stepwise process via σ-dimers. The strength of the π-stacking pancake interaction depends strongly on substituents and covers a wide range both in terms of binding energies and contact distances. The spin densities in the π-stacking dimers reflect these trends and display a wide range of diradicaloid characters. Many σ-dimer configurations compete some of which are separated by small barriers leading to fluxional structures between σ-bonded configurations or σ- and π-bonded configurations.

3D graphene-Fe3O4 nanocomposites with high-performance microwave absorption.

3D Fe3O4-graphene nanocomposites were conveniently prepared via a direct hydrothermal grafting method. On the basis of the unique properties of both single-crystalline Fe3O4 and 3D chemically reduced graphene oxide, with characteristics such as ultralow density and high surface area, the as-prepared graphene-Fe3O4 nanocomposites showed high-performance microwave absorption ability and have the potential for application as advanced microwave absorbers.

Machine Learning Reveals the Critical Interactions for SARS-CoV-2 Spike Protein Binding to ACE2.

SARS-CoV and SARS-CoV-2 bind to the human ACE2 receptor in practically identical conformations, although several residues of the receptor-binding domain (RBD) differ between them. Herein, we have used molecular dynamics (MD) simulations, machine learning (ML), and free-energy perturbation (FEP) calculations to elucidate the differences in binding by the two viruses. Although only subtle differences were observed from the initial MD simulations of the two RBD-ACE2 complexes, ML identified the individual residues with the most distinctive ACE2 interactions, many of which have been highlighted in previous experimental studies. FEP calculations quantified the corresponding differences in binding free energies to ACE2, and examination of MD trajectories provided structural explanations for these differences. Lastly, the energetics of emerging SARS-CoV-2 mutations were studied, showing that the affinity of the RBD for ACE2 is increased by N501Y and E484K mutations but is slightly decreased by K417N.