One tactic for cysteine protease inhibition is to form a covalent bond between an electrophilic atom of the inhibitor and the thiol of the catalytic cysteine. In this study, we evaluate the reaction free energy obtained from a hybrid Quantum Mechanical/Molecular Mechanical (QM/MM) free energy profile as a predictor of affinity for reversible, covalent inhibitors of rhodesain. We demonstrate that the reaction free energy calculated with the PM6/MM potential is in agreement with the experimental data and suggest that the free energy profile for covalent bond formation in a protein environment may be a useful tool for the inhibitor design.

Protein Nucleoside Phosphorylase (PNP) is an enzyme, which catalyzes reversible conversion process (ribosylation and phosphorolysis) between nucleobases (purines) and their nucleosides. Experimental studies showed that calf PNP ribosylates purine analogs in specific positions – 2,6-diamino-8-azapurine in positions 7 or 8 and 8-azaguanine in position 9 of the triazole ring. The reason of this phenomena can be a result of different exposition of purine substrates to the channel leading to the binding site. This hypothesis was verified by application of molecular modelling techniques to two complexes of purine analogs 2,6-diamino-azapurine – calf PNP (pdb-code: 1LVU) and 8-azaguanine – calf PNP (pdb-code: 2AI1). The results obtained with combination of quantum chemistry, docking and molecular dynamics methods showed qualitative validity of our hypothesis. Binding free energies of protein-ligand systems showed that most probable binding poses expose N8 nitrogen for 2,6-diamino-8-azapurine and N9 nitrogen for 8-azaguanine into the binding channel and ruled out exposition of N9 for 2,6-diamino-8-azapurine and N7 for 8-azaguanine, partially in agreement with the experimental data. The other important result obtained in this study is a significantly higher population of protonated form of crucial residue Glu-201 present in the binding pocket, compared to the standard protonation of free glutamic acid in solution. This result combined with populations of tautomeric forms of both investigated systems strongly suggests that 2,6-diamino-8-azaguanine and 8-azaguanine is recognized by proteins with deprotonated and protonated Glu-201 residue, respectively. Comparison of computed binding poses of the investigated ligands to the inhibitors present in crystal structures suggests that modification of (S)-PMPDAP inhibitor, in which 2-(phosphonomethoxy)propyl chain is attached at position 8 instead of position 9, might increase its binding affinity.

Transcription factors (TFs) play the key role in the process of major regulators that control critical cellular processes and response to environmental conditions. Yeast is a single-cell fungal organism that is an important model organism in understanding fundamental biological pathways and transcriptional regulatory networks. The transcriptional regulation in yeast has been studied extensively both computationally and experimentally using traditional methods and high-throughput technologies. However, the identities of transcription factors that regulate major functional categories of genes remain unknown. Due to the avalanche of biological data in the post-genomic era, it is an urgent need to develop automated computational methods to enable accurate identification of efficient transcription factor binding sites from the large number of candidates. In this paper, we analyze high-resolution DNA-binding profiles and motifs for TFs, covering all possible contiguous 8-mers. First, we divide all 8-mers motifs into 16 various categories, and select all sorts of samples from each category by setting the threshold of E-score. Then, we employ five feature representation methods. Also, we adopt a total of three feature selection methods to filter out useless features. Finally, we use Extreme Gradient Boosting (XGBoost) as the base classifier, and adopt the one-vs-rest strategy to construct 16 binary classifiers to solve this multi-classification problem. In the experiment, our method achieves the best performance with the overall Acc of 79.72% and the MCC of 0.77. We find the similarity relationship among each category from different TF families and obtain sequence motif schematic diagram via multiple sequence alignment. The complexity of DNA recognition may act as an important role in the evolution of gene regulatory.

Apoptosis is a key cell death pathway in mammalian cells. Understanding this process and its regulation has been a subject of study in the last three decades. Members of the Bcl-2 family of proteins are involved in the regulation of apoptosis through mitochondrial poration with the subsequent initiation of apoptosis. Deregulation of pro-apoptotic proteins contributes to the progression of many tumour processes. Understanding how these pore-forming Bcl-2 proteins Bak and Bax are activated is key to find new anti-cancer treatments. Since no drug capable of activating Bak has been disclosed yet, the study of the structural features of BH3 peptides -known Bak activators- relevant for binding along with its binding energy decomposition analysis, results essential for designing novel small molecule mimics of BH3. Interestingly, a BH3 Bim analogue inactivating Bak has recently been discovered, opening a question on the molecular features that determine the function of BH3 peptides. Therefore, the present work is aimed at understanding the way BH3 peptides activate or inactivate Bak in order to identify differential structural features that can be used in drug design. For this purpose, complexes of Bak with an activator and an inhibitor have been subjected to a molecular dynamics study. Structural differences were assessed by means of the fluctuations of the corresponding Principal Components. Moreover, the MMPB/GBSA approach was used to compute the binding free energy of the diverse complexes to identify those residues of the BH3 peptide that exhibit the larger contributions to complex formation. The results obtained in this work show differences between activators and inhibitors, both in structural and energetic terms, which can be used in the design of new molecules that can activate or inactivate pro-apoptotic Bak.

Cucurbit[n]urils (CB[n], n = 5, 6, 7, 8, 10, 14) and their derivatives due to the hydrophobic cavities and polar carbonyl portals have been considerably explored for their potential uses as drug delivery systems. It is important to understand how these macrocyclic compounds interact with guests. Camptothecin (CPT), as a natural alkaloid, is a topoisomerase inhibitor with antitumor activity against breast, pancreas and lung cancers. The application of this drug in cancer therapy is restricted due to its low aqueous solubility and high toxicity. Recently, the complex formation between the cucurbit[7]uril (CB[7]) / acyclic cucurbit[4]uril (aCB[4]) nanocontainers and CPT have been evaluated to overcome the potential drawbacks of the related drug. Herein, using computational methods, we identified the interaction mechanism of CPT with CB[7]/aCB[4]s consist of benzene and naphthalene sidewalls (aCB[4]benzene and aCB[4]naphthalene, respectively), since the experimental approaches have not completely provided information at the molecular level. Our molecular docking and molecular dynamics (MD) simulations show that CB[7] and its two acyclic derivatives form stable inclusion complexes with CPT especially through hydrophobic interactions. We also found that aCB[4]s with the aromatic sidewalls can attach to CPT through π-π interactions. This investigation highlights aCB[4]s due to the structural properties and flexible nature as a better nanocontainers for controlled release delivery of pharmaceutical agents in comparison with CB[7] nanocontainer.

Previous studies demonstrated the efficiency of the Molecular Dynamics with excited Normal modes (MDeNM) method on the characterization of large structural changes at a low computational cost. We present here MDeNM-EMfit, an extension of the original method designed to the flexible fit of structures into cryo-EM maps. Here, instead of a uniform exploration of the collective motions described by normal modes, sampling is directed towards conformations with increased correlations with the experimental map. Future perspectives to improve the accuracy of fitting and speed of calculations are discussed in light of the results.

The melanocortin receptor 4 (MC4R), is expressed predominantly in the central nervous system and regulates food intake, sexual function and is also thought to be responsible for effects on mood and cognition. It belongs to the melanocortin receptors subfamily of G protein-coupled receptors (GPCRs). Here, we have synthesized and structurally characterized three peptides that bind to MC4R, producing different signaling events. AgRP, a naturally occurring antagonist, HLWNRS, the minimal sequence of the N-terminal with partial agonist activity and aMSH, a full agonistic peptide. By implementing molecular dynamics simulations on the different peptide-receptor complexes, we propose their molecular basis of binding, in order to investigate their differential molecular properties regarding the activation states of the receptor. Our analysis shows that the agonist and partially agonist induce a rotation in transmembrane helix 3, which is known to be involved in the key events occurring during GPCR activation, and this movement is impacted by certain aromatic residues and their positioning in the orthosteric binding site of the receptor.

Atomic neural networks (ANNs) constitute a class of machine learning methods for predicting potential energy surfaces and physico-chemical properties of molecules and materials. Despite many successes, developing interpretable ANN architectures and implementing existing ones efficiently are still challenging. This calls for reliable, general-purpose and open-source codes. Here, we present a python library named PiNN as a solution toward this goal. In PiNN, we designed a new interpretable and high-performing graph convolutional neural network variant, PiNet, as well as implemented the established Behler-Parrinello high-dimensional neural network. These implementations were tested using datasets of isolated small molecules, crystalline materials, liquid water and an aqueous alkaline electrolyte. PiNN comes with a visualizer called PiNNBoard to extract chemical insight ``learned'' by ANNs, provides analytical stress tensor calculations and interfaces to both the Atomic Simulation Environment and a development version of the Amsterdam Modeling Suite. Moreover, PiNN is highly modularized which makes it useful not only as a standalone package but also as a chain of tools to develop and to implement novel ANNs. The code is distributed under a permissive BSD license and is freely accessible at \href{https://github.com/Teoroo-CMC/PiNN/}{https://github.com/Teoroo-CMC/PiNN/} with full documentation and tutorials.

Methamphetamine (METH) is one of the highly addictive non-opioid psychostimulants, acting as a xenobiotic-associated molecular pattern to target TLR4 and activate microglia. However, the molecule recognition of METH by innate immune receptor TLR4/MD-2 is not well understood. METH exists in two enantiomeric forms and it is unclear whether the TLR4 innate immune recognition with METH is stereoselective. Herein molecular dynamics simulations were performed to dissect the recognition of (+)-METH and (-)-METH by TLR4/MD-2 at the atomic level. Amphetamine (AMPH), which is an analog of METH, was also investigated for comparison. Computational simulations indicate METH binds into the interaction interface between MD-2 as well as TLR4* that is from the adjacent copy of TLR4-MD-2, therefore stabilizing the active heterotetramer (TLR4/MD-2)2 complex. The calculated binding free energies and potential of mean force values show (−)-METH and (+)-METH have similar TLR4/MD-2 binding affinity. Further dynamics analyses of bindings with TLR4/MD-2 indicate that (−)-METH and (+)-METH behave similarly. Unlike the stereo-selective neuron stimulating activities of METH, no enantioselectivity was observed for METH interacting with TLR4/MD-2 complex as well as activating TLR4 signaling. Compared to METH, AMPH showed much weaker interactions with TLR4/MD-2, indicating that the substituted methyl group is critical in the molecular recognition of METH by TLR4/MD-2. In all, this study provides a molecular insight into the innate immune recognition of METH, which demonstrates that METH could be non-enantioselectively sensed by TLR4/MD-2.

Structural and biochemical studies of the Hsp70 chaperones have provided the molecular view of the chaperone biochemical cycle by revealing a complex interplay between allosteric conformational states that controls the feedback loop between stimulation of the ATPase activity and the substrate release. Allosteric regulation in the Hsp70 chaperones and efficient substrate targeting are mediated by J-domain cochaperones through a dynamic interaction network controlled by the regulatory hotspots. In the current work, we have simulated conformational landscapes and residue interaction networks in the open, closed and cochaperone-bound DnaK structures. The results of this work have shown that J-domain can selectively enhance direction-specific signal propagation from the substrate binding domain to the catalytic center and promote structural environment required for ATP hydrolysis. By employing different network-based approaches, we examined the role and contribution of post-translational modifications sites in allosteric regulation of human Hsp70. The central finding of this analysis indicated that conserved phosphorylation sites localized preferentially in the nucleotide-binding domain regions are often aligned with the allosteric control points and serve as effector centers in the Hsp70. We have found that cooperation of post-translational modifications sites is based on the governing role of phosphorylation sites in dictating regulatory switching functions, while the bulk of acetylation sites can be involved in sensing the long-range signals and executing allosteric changes during the ATPase cycle. The results of this study highlight the important role of phosphorylation sites in exerting control over allosteric changes in the Hsp70. The network-centric framework for analysis of conformational dynamics and chaperone landscapes can explain a range of structural and functional experiments, providing a robust dynamic model of Hsp70 regulation by cochaperones and sites of post-translational modifications.

As new generations of thin-film semiconductors are moving towards solution-based processing, the development of printing formulations will require information pertaining to the free energies of mixing of complex mixtures. From the standpoint of in silico materials design, this move necessitates the development of methods that can accurately and quickly evaluate these formulations in order to maximize processing speed and reproducibility. Here, we make use of molecular dynamics (MD) simulations in combination with the two-phase thermodynamic (2PT) model to explore the free energy of mixing surfaces for a series of halogenated solvents and high boiling point solvent additives used in the development of thin-film organic semiconductors. While the combined methods generally show good agreement with available experimental data, the computational cost to traverse the free-energy landscape is considerable. Hence, we demonstrate how a Bayesian optimization scheme, coupled with the MD and 2PT approaches, can drastically reduce the number of simulations required, in turn shrinking dramatically both the computational cost and time.

Graph neural networks (GNNs) have attracted significant attention from the chemical science community because molecules can be represented as the featured graph. In particular, graph convolutional network (GCN) and its variants have been widely used and have shown a state-of-the-art performance in analyzing molecules, such as molecular label classification, drug discovery, and molecular property prediction. However, in the molecular analysis, existing GCNs have two fundamental limitations: 1) information of the molecular scale is distorted; 2) global structures in a molecule are ignored. These limitations can seriously degrade the performance in the machine learning-based molecular analysis because the scale and global structure of a molecule occasionally have a significant effect on the molecular properties. To overcome the limitations of existing GCNs, we comprehensively analyzed the limitations of GCNs and developed a costless solution for the limitations of GCNs. To demonstrate the effectiveness of our solution, extensive experiments were conducted on various benchmark datasets

The catalytic mechanism of SalL chlorinase has been investigated by combining quantum mechanical/molecular mechanical (QM/MM) techniques and umbrella sampling simulations to compute free energy profiles. Our results shed light on the interesting fact that the substitution of chloride with fluorine in SalL chlorinase leads to a loss of halogenase activity. Potential of Mean Force based on DFTB3/MM analysis shows that fluorination corresponds to a barrier of 13.5 kcal.mol-1 higher than chlorination. Additionally, our results present a molecular description of SalL acting as a chlorinase instead of a methyl halide transferase.

Histidine Kinases (HK) of bacterial two component systems represent a hallmark of allosterism in proteins, being able to detect a signal through the sensor domain and transmit this information through the protein matrix to the kinase domain which, once active, autophosphorylates a specific histidine residue. Inactive to active transition results in a large conformational change that moves the kinase on top of the histidine. The precise mechanism leading to autophosphorylation and the coupling between conformational and chemical steps are still largely unknown. In the present work, we use several molecular simulation techniques (Molecular Dynamics, Hybrid QM/MM and constant ph MD) to study the activation and autophosphorylation reactions in L.plantarum WalK, a cis acting HK. In agreement with previous results, our results show that the chemical step, requires, a tight coupling with the conformational step in order to maintain the histidine phosphoacceptor in correct tautomeric state, with a reactive -nitrogen. During the conformational transition, the kinase domain is never released and walks along the HK helix axis, breaking and forming several conserved residue based contacts. The phosphate transfer reaction is concerted in the TS region and is catalyzed through the stabilization of the negative developing charge of transferring phosphate along the reaction.

Protein-peptide interactions play a crucial role in many cellular and biological functions, which justify the increasing interest in the development of peptide-based drugs. However, predicting experimental binding modes and affinities in protein-peptide docking remains a great challenge for most docking programs due to some particularities of this class of ligands, such as the high degree of flexibility. In this paper, we present the performance of the DockThor program on the LEADS-PEP dataset, a benchmarking set composed of 53 diverse protein-peptide complexes with peptides ranging from 3 to 12 residues and with up to 51 rotatable bonds. The DockThor performance for pose prediction on redocking studies was compared with some state-of-the-art docking programs that were also evaluated on LEADS-PEP dataset: AutoDock, AutoDock Vina, Surflex, GOLD, Glide, rDock, and DINC, as well as to the task-specific docking protocol HPepDock. Our results indicate that DockThor could dock 40% of the cases with an overall backbone RMSD below 2.5 Å when the top-scored docking pose was considered, exhibiting similar results to Glide and outperforming other protein-ligand docking programs, whereas rDock and HPepDock achieved superior results. Assessing the docking poses closest to the crystal structure (i.e., best-RMSD pose) DockThor achieved a success rate of 60% in pose prediction. Due to the great overall performance to handle with peptidic compounds, DockThor program can be considered as a suitable for docking highly flexible and challenging ligands, with up to 40 rotatable bonds. DockThor is freely available as a virtual screening web server at https://www.dockthor.lncc.br/v2/.

Due to its specific physicochemical properties (fluorescence, photosensitizing and redox reactions), the vitamin B2, also called riboflavin (RF), has been generating a lot of interest in nanotechnologies and bioengineering for the last decade. RF, by targeting its RFVT transporters overexpressed in some cancers, is particularly used to functionalize nanovectors for anti-cancer drug delivery. From a physiopathological point of view, a RF deficiency has been implicated in various pathologies, including mendelian diseases. RF deficiency is mainly due to natural variants of its RFVT transporters that make them inactive and therefore prevent RF transport. The lack of structural data about RFVT is a major drawback for a better understanding of the role of the mutations in the molecular mechanism of these transporters. In this context, this work was aimed at investigating the 3D structure of RFVT3 and its interactions with RF. For this purpose, we used an in silico procedure including protein threading, docking and molecular dynamics. Our results propose that the natural variant W17R, known to be responsible for BVVL syndrome, prevents the recognition of RF by RFVT3 and thus blocks its transport. This in silico procedure could be used for elucidating the impact of pathogenic mutations of other proteins. Moreover, the identification of RF binding site will be useful for the design of RF-functionalized nanovectors.

In drug design, the importance of molecular solvation and desolvation is increasingly appreciated and water molecules are recognized as active contributors to protein-ligand binding. However despite a number of theoretical approaches, computational tools are still far from routinely integrating solvation features into rational Structure-Affinity-Relationships (SAR). In this contribution, we present a set of solvent functional-based models which calculate the relative binding free energy contributions resulting from solvation for a diverse set of 53 thrombin protein-ligand complexes. These protein-ligand complexes were further matched into chemically similar pairs of ligand molecules. Our solvent functionals are based on molecular dynamics simulations in conjunction with GIST (Grid Inhomogeneous Solvation Theory) processing and they are calibrated using accurate experimental data from ITC measurements. We found, that excellent agreement with experimental measurements can be achieved by considering either the desolvation of the protein binding pocket or the ligand in solution prior to binding. The incorporation of contributions from the protein-ligand complexes generally result in good agreement with experimental measurements, but require additional adjustment of spatial cutoff parameters. In addition, we investigated the transfer of the trained solvent functionals to another protein target, which revealed deviating performance results indicating a target-specific treatment of solvation features within the model. Together with our tool Gips, we provide a way to automatically generate solvent functional parameters from GIST data and allow for the design of compounds with favorable solvation properties given the chemical similarity and affinity range of the matching pairs in the training set. Finally, we reflect on the resemblance with the popular 3D-QSAR method, as our study allows for (retrospective) insights about the high predictive power of this well-established method.

In this work, we report an ab initio investigation based on density functional theory calculations within van der Waals D3 corrections to investigate the activation of CO2 on transition-metal (TM) 13-atom clusters (Ru, Rh, Pd, Ag), which is a key step for the development of subnano catalysts for the CO2 reduction. From our analyses, which includes the calculations of several adsorption properties and employing the Spearman correlation analysis, we found that CO2 adopts two distint configurations on the selected TM13 clusters (lower symmetry putative global minimum structure and high symmetry icosahedron structure), namely, a bend CO2 configuration in which the OCO angle is about 125 to 150◦ , while a linear CO2 as in gas-phase is obtained for several configurations. The two CO2 configurations have very different adsorption energies, which can be explained by the interaction mechanism, i.e., chemisorption for the bend CO2 with a charge transfer towards CO upon adsorption and physisorption for the linear CO2. Thus, the CO2 activation occurs only in the first case and it is driven by charge transfer from the TM clusters to the CO molecule (i.e., CO2-delta ), which is confirmed by our Bader charge analysis and vibrational frequencies. Further details aligned with our conclusions are discussed within the paper.

The TATA-box binding protein (TBP) is an important element of the transcription machinery in archaea and eukaryotic organisms. TBP is expressed in organisms adapted to different temperatures, indicating a robust structure, and experimental studies have shown that the mid-unfolding temperature (Tm) of TBP is directly correlated with the optimal growth temperature (OGT) of the organism. To understand which are the relevant structural requirements for its stability, we present the first structural and dynamic computational study of TBPs, combining molecular dynamics (MD) simulations and a quantitative structure-property relationship (QSPR) over a set of TBPs of organisms adapted to different temperatures. We found that the main structural properties of TBP used to adapt to high temperature are: an increase in the ease of desolvation of charged residues at the surface, an increase in the local resiliency, the presence of Leu clusters in the protein core, and an increase in the loss of hydrophobic packing in the N-terminal subdomain. In view of our results, we consider that TBP is a good model to study thermal adaptation and our analysis opens the possibility of performing protein engineering on TBPs to study transcription at high or low temperatures.

5-HT1A receptor (5-HT1AR) agonists have been implicated in the treatment of a variety of central nervous system (CNS) diseases such as depression and anxiety et al. Based on our previously compound FW01 (Ki = 51 ± 16 nM) obtained by virtual screening, a series of FW01 derivatives were designed and synthesized by the modification of the amide tail group as well as indole head group of FW01. SAR exploration found that amide tail group and indole head group play pivotal roles in determining the binding affinity and selectivity on dopamine and serotonin receptor subtypes. Among all tested compounds, 9_24 has a Ki value of 5 ± 0.6 nM with a good selectivity towards 5-HT1AR. The [35S] GTPγS assay showed that 9_24 is a full agonist towards 5-HT1AR with an EC50 value of 0.059 nM, which shows 266.2 and 146.4-fold selectivity to 5-HT2A and D3 respectively. Molecular dynamics simulations and molecular docking studies with 5-HT1AR-9_24 were performed to disclose the mechanism of its high activity and selectivity. Finally, a detailed stepwise 9_24 induced signal transduction mechanism of 5-HT1AR is proposed.

This article introduces Web-ARM, a specialized, on-line available, tool designed to build quantum mechanical/molecular mechanical models of rhodopsins, a widely spread family of light-responsive proteins. Web-ARM allows to rapidly build models of rhodopsins with a documented quality and to predict trends in UV-Vis absorption maximum wavelengths, based on their excitation energies computed at the CASPT2//CASSCF/Amber level of theory. Web-ARM builds upon the recently reported, python-based a-ARM protocol [J. Chem. Theory Comput., 2019, 15, 3134-3152], and as such necessitates only a crystallographic structure or a comparative model in PDB format and a very basic knowledge of the studied rhodopsin system. The user-friendly web interface uses such input to generate congruous, gas-phase models of rhodopsins and, if requested, their mutants. We present two possible applications of Web-ARM, which showcase how the interface can be employed to assist both research and educational activities in fields at the interface between chemistry and biology. The first application shows how, through Web-ARM, research projects (e.g., rhodopsin and rhodopsin mutant screening) can be carried out in significantly less time with respect to using the required computational photochemistry tools via a command line. The second application documents the use of Web-ARM in a real-life educational/training activity, through a hands-on experience illustrating the concepts of rhodopsin color tuning.

Identification of correct protein-ligand binding poses is important in structure-based drug design and crucial for the evaluation of protein-ligand binding affinity. Protein-ligand coordinates are commonly obtained from crystallography experiments that provide a static model of an ensemble of conformations. Binding Pose Metadynamics (BPMD) is an enhanced-sampling method that allows an efficient assessment of ligand stability in solution. Ligand poses that are unstable under the bias of the metadynamics simulation are expected to be infrequently occupied in the energy landscape, thus making minimal contributions to the binding affinity. Here, the robustness of the method is studied using crystal structures with ligands known to be incorrectly modelled as well as 63 structurally diverse crystal structures with ligand fit to electron density from the Twilight database. Results show that BPMD can successfully discriminate compounds whose binding pose is not supported by the electron density from those with well-defined electron density.

Leveraging new data sources is a key step in accelerating the pace of materials design and discovery. To complement the strides in synthesis planning driven by historical, experimental, and computed data, we present an automated, unsupervised method for connecting scientific literature to inorganic synthesis insights. Starting from natural language text, we apply word embeddings from language models, which are fed into a named entity recognition model, upon which a conditional variational autoencoder is trained to generate syntheses for any inorganic materials of interest. We show the potential of this technique by predicting precursors for two perovskite materials, using only training data published over a decade prior to their first reported syntheses. We demonstrate that the model learns representations of materials corresponding to synthesis-related properties, and that the model's behavior complements existing thermodynamic knowledge. Finally, we apply the model to perform synthesizability screening for proposed novel perovskite compounds.