In Silico drug repurposing pipeline using deep learning and structure based approaches in epilepsy.

Due to considerable global prevalence and high recurrence rate, the pursuit of effective new medication for epilepsy treatment remains an urgent and significant challenge. Drug repurposing emerges as a cost-effective and efficient strategy to combat this disorder. This study leverages the transformer-based deep learning methods coupled with molecular binding affinity calculation to develop a novel in-silico drug repurposing pipeline for epilepsy. The number of candidate inhibitors against 24 target proteins encoded by gain-of-function genes implicated in epileptogenesis ranged from zero to several hundreds. Our pipeline has repurposed the medications with most anti-epileptic drugs and nearly half psychiatric medications, highlighting the effectiveness of our pipeline. Furthermore, Lomitapide, a cholesterol-lowering drug, first emerged as particularly noteworthy, exhibiting high binding affinity for 10 targets and verified by molecular dynamics simulation and mechanism analysis. These findings provided a novel perspective on therapeutic strategies for other central nervous system disease.

Drug Repositioning for Amyloid Transthyretin Amyloidosis by Interactome Network Corrected by Graph Neural Networks and Transcriptome Analysis.

Amyloid transthyretin (ATTR) amyloidosis caused by transthyretin misfolded into amyloid deposits in nerve and heart is a progressive rare disease. The unknown pathogenesis and the lack of therapy make the 5-year survival prognosis extremely poor. Currently available ATTR drugs can only relieve symptoms and slow down progression, but no drug has demonstrated curable effect for this disease. The growing volume of pharmacological data and large-scale genome and transcriptome data bring new opportunities to find potential new ATTR drugs through computational drug repositioning. We collected the ATTR-related in the disease pathogenesis and differentially expressed (DE) genes from five public databases and Gene Expression Omnibus expression profiles, respectively, then screened drug candidates by a corrected protein-protein network analysis of the ATTR-related genes as well as the drug targets from DrugBank database, and then filtered the drug candidates on the basis of gene expression data perturbed by compounds. We collected 139 and 56 ATTR-related genes from five public databases and transcriptome data, respectively, and performed functional enrichment analysis. We screened out 355 drug candidates based on the proximity to ATTR-related genes in the corrected interactome network, refined by graph neural networks. An Inverted Gene Set Enrichment analysis was further applied to estimate the effect of perturbations on ATTR-related and DE genes. High probability drug candidates were discussed. Drug repositioning using systematic computational processes on an interactome network with transcriptome data were performed to screen out several potential new drug candidates for ATTR.

Structures of the Insecticidal Toxin Complex Subunit XptA2 Highlight Roles for Flexible Domains.

The Toxin Complex (Tc) superfamily consists of toxin translocases that contribute to the targeting, delivery, and cytotoxicity of certain pathogenic Gram-negative bacteria. Membrane receptor targeting is driven by the A-subunit (TcA), which comprises IgG-like receptor binding domains (RBDs) at the surface. To better understand XptA2, an insect specific TcA secreted by the symbiont X. nematophilus from the intestine of entomopathogenic nematodes, we determined structures by X-ray crystallography and cryo-EM. Contrary to a previous report, XptA2 is pentameric. RBD-B exhibits an indentation from crystal packing that indicates loose association with the shell and a hotspot for possible receptor binding or a trigger for conformational dynamics. A two-fragment XptA2 lacking an intact linker achieved the folded pre-pore state like wild type (wt), revealing no requirement of the linker for protein folding. The linker is disordered in all structures, and we propose it plays a role in dynamics downstream of the initial pre-pore state.

Human Plasma IgG N-Glycome Profiles Reveal a Proinflammatory Phenotype in Chronic Thromboembolic Pulmonary Hypertension.

BACKGROUND: The pathological mechanism of chronic thromboembolic pulmonary hypertension (CTEPH) is not fully understood, and inflammation has been reported to be one of its etiological factors. IgG regulates systemic inflammatory homeostasis, primarily through its N-glycans. Little is known about IgG N-glycosylation in CTEPH. We aimed to map the IgG N-glycome of CTEPH to provide new insights into its pathogenesis and discover novel markers and therapies. METHODS: We characterized the plasma IgG N-glycome of patients with CTEPH in a discovery cohort and validated our results in an independent validation cohort using matrix-assisted laser desorption/ionization time of flight mass spectrometry. Thereafter, we correlated IgG N-glycans with clinical parameters and circulating inflammatory cytokines in patients with CTEPH. Furthermore, we determined IgG N-glycan quantitative trait loci in CTEPH to reveal partial mechanisms underlying glycan changes. RESULTS: Decreased IgG galactosylation representing a proinflammatory phenotype was found in CTEPH. The distribution of IgG galactosylation showed a strong association with NT-proBNP (N-terminal pro-B-type natriuretic peptide) in CTEPH. In line with the glycomic findings, IgG pro-/anti-inflammatory N-glycans correlated well with a series of inflammatory markers and gene loci that have been reported to be involved in the regulation of these glycans or inflammatory immune responses. CONCLUSIONS: This is the first study to reveal the full signature of the IgG N-glycome of a proinflammatory phenotype and the genes involved in its regulation in CTEPH. Plasma IgG galactosylation may be useful for evaluating the inflammatory state in patients with CTEPH; however, this requires further validation. This study improves our understanding of the mechanisms underlying CTEPH inflammation from the perspective of glycomics.

Data-independent acquisition mass spectrometry reveals comprehensive plasma protein profiles in the natural history of patients with hereditary transthyretin amyloidosis (ATTRv).

OBJECTIVES: Hereditary transthyretin amyloidosis (ATTRv) is a rare, fatal, autosomal dominant disease with more than 140 mutations discovered. Three phenotypes of amyloid infiltration are neuropathy (ATTRv-PN), cardiopathy (ATTRv-CM), and neuropathy + cardiopathy (ATTRv-MIX). The lack of ATTR-specific biomarkers, difficulties in biopsy evidence, and limited known pathogenic mechanisms have made diagnosis difficult. Newly emerging noninvasive measures for monitoring progression and disease-modifying therapies have improved early diagnosis and patient management. METHODS: Our research applies the latest technology, Data-Independent Acquisition-Based Quantitative Proteomics (DIA), to reveal comprehensive plasma protein profiles in the natural history of Chinese patients with hereditary transthyretin amyloidosis (ATTRv). We analyzed differentially expressed proteins (DEPs) in three phenotypes (ATTRv-PN, ATTRv-CM, and ATTRv-MIX). RESULTS: Serum samples were collected from a total of 18 patients (6 ATTRv-PN, 5 ATTRv-CM, and 7 ATTRv-MIX patients) and 20 healthy participants as a control group. Combined with the results of the proteomic and bioinformatic analyses, we found 30 DEPs and protein interaction networks clustered in KRT family proteins and DSC3 between ATTRv-PN and the control, which were rich in the estrogen signaling pathway and the cell adhesion molecule (CAM) pathway. CONCLUSION: This study demonstrates a global and significant proteomic profile in different stages of ATTRv.

Knowledge-based BERT: a method to extract molecular features like computational chemists.

Molecular property prediction models based on machine learning algorithms have become important tools to triage unpromising lead molecules in the early stages of drug discovery. Compared with the mainstream descriptor- and graph-based methods for molecular property predictions, SMILES-based methods can directly extract molecular features from SMILES without human expert knowledge, but they require more powerful algorithms for feature extraction and a larger amount of data for training, which makes SMILES-based methods less popular. Here, we show the great potential of pre-training in promoting the predictions of important pharmaceutical properties. By utilizing three pre-training tasks based on atom feature prediction, molecular feature prediction and contrastive learning, a new pre-training method K-BERT, which can extract chemical information from SMILES like chemists, was developed. The calculation results on 15 pharmaceutical datasets show that K-BERT outperforms well-established descriptor-based (XGBoost) and graph-based (Attentive FP and HRGCN+) models. In addition, we found that the contrastive learning pre-training task enables K-BERT to 'understand' SMILES not limited to canonical SMILES. Moreover, the general fingerprints K-BERT-FP generated by K-BERT exhibit comparative predictive power to MACCS on 15 pharmaceutical datasets and can also capture molecular size and chirality information that traditional binary fingerprints cannot capture. Our results illustrate the great potential of K-BERT in the practical applications of molecular property predictions in drug discovery.

Allosteric Role of Substrate Occupancy Toward the Alignment of P-glycoprotein Nucleotide Binding Domains.

P-glycoprotein (Pgp) is an ATP-binding cassette transporter that eliminates toxins from the cell but causes multidrug resistance in chemotherapies. The crystal structures of Pgp revealed drug-like compounds bound to an inward-facing conformation in which the energy-harnessing nucleotide binding domains (NBDs) were widely separated with no interfacial interaction. Following drug binding, inward-facing Pgp must transition to an NBD dimer conformation to achieve ATP binding and hydrolysis at canonical sites defined by both halves of the interface. However, given the high degree of flexibility shown for this transporter, it is difficult to envision how NBDs overcome entropic considerations for achieving proper alignment in order to form the canonical ATP binding site. We explored the hypothesis that substrate occupancy of the polyspecific drug-binding cavity plays a role in the proper alignment of NBDs using computational approaches. We conducted twelve atomistic molecular dynamics (MD) simulations (100-300 ns) on inward-facing Pgp in a lipid bilayer with and without small molecule substrates to ascertain effects of drug occupancy on NBD dimerization. Both apo- and drug-occupied simulations showed NBDs approaching each other compared to the crystal structures. Apo-Pgp reached a pseudo-dimerization in which NBD signature motifs for ATP binding exhibited a significant misalignment during closure. In contrast, occupancy of three established substrates positioned by molecular docking achieved NBD alignment that was much more compatible with a canonical NBD dimerization trajectory. Additionally, aromatic amino acids, known to confer the polyspecific drug-binding characteristic of the internal pocket, may also govern polyspecific drug access to the cavity. The enrichment of aromatics comprising the TM4-TM6 portal suggested a preferential pathway over the aromatic-poor TM10-TM12 for lateral drug entry from the lipid bilayer. Our study also suggested that drug polyspecificity is enhanced due to a synergism between multiple drug-domain interactions involving 36 residues identified in TM1, 5, 6, 7, 11 and 12.

Computational studies of plasma lipoprotein lipids.

Plasma lipoproteins are macromolecular assemblies of proteins and lipids found in the blood. The lipid components of lipoproteins are amphipathic lipids such as phospholipids (PLs), and unesterified cholesterols (UCs) and hydrophobic lipids such as cholesteryl esters (CEs) and triglycerides (TGs). Since lipoproteins are soft matter supramolecular assemblies easily deformable by thermal fluctuations and they also exist in varying densities and protein/lipid components, a detailed understanding of their structure/function is experimentally difficult. Molecular dynamics (MD) simulation has emerged as a particularly promising way to explore the structure and dynamics of lipoproteins. The purpose of this review is to survey the current status of computational studies of the lipid components of the lipoproteins. Computational studies aim to explore three levels of complexity for the 3-dimensional structural dynamics of lipoproteins at various metabolic stages: (i) lipoprotein particles consist of protein with minimal lipid; (ii) lipoprotein particles consist of PL-rich discoidal bilayer-like lipid particles; (iii) mature circulating lipoprotein particles consist of CE-rich or TG-rich spheroidal lipid-droplet-like particles. Due to energy barriers involved in conversion between these species, other biomolecules also participate in lipoprotein biological assembly. For example: (i) lipid-poor apolipoprotein A-I (apoA-I) interacts with ATP-binding cassette transporter A1 (ABCA1) to produce nascent discoidal high density lipoprotein (dHDL) particles; (ii) lecithin-cholesterol acyltransferase (LCAT) mediates the conversion of UC to CE in dHDL, driving spheroidal HDL (sHDL) formation; (iii) transfer proteins, cholesterol ester transfer protein (CETP) and phospholipid transfer protein (PLTP), transfer both CE and TG and PL, respectively, between lipoprotein particles. Computational studies have the potential to explore different lipoprotein particles at each metabolic stage in atomistic detail. This review discusses the current status of computational methods including all-atom MD (AAMD), coarse-grain MD (CGMD), and MD-simulated annealing (MDSA) and their applications in lipoprotein structural dynamics and biological assemblies. Results from MD simulations are discussed and compared across studies in order to identify key findings, controversies, issues and future directions. This article is part of a Special Issue entitled: Biosimulations edited by Ilpo Vattulainen and Tomasz Róg.

Tools and procedures for visualization of proteins and other biomolecules.

Protein, peptides, and nucleic acids are biomolecules that drive biological processes in living organisms. An enormous amount of structural data for a large number of these biomolecules has been described with atomic precision in the form of structural "snapshots" that are freely available in public repositories. These snapshots can help explain how the biomolecules function, the nature of interactions between multi-molecular complexes, and even how small-molecule drugs can modulate the biomolecules for clinical benefits. Furthermore, these structural snapshots serve as inputs for sophisticated computer simulations to turn the biomolecules into moving, "breathing" molecular machines for understanding their dynamic properties in real-time computer simulations. In order for the researcher to take advantage of such a wealth of structural data, it is necessary to gain competency in the use of computer molecular visualization tools for exploring the structures and visualizing three-dimensional spatial representations. Here, we present protocols for using two common visualization tools--the Web-based Jmol and the stand-alone PyMOL package--as well as a few examples of other popular tools.

Equilibrated atomic models of outward-facing P-glycoprotein and effect of ATP binding on structural dynamics.

P-glycoprotein (Pgp) is an ATP-binding cassette (ABC) transporter that alternates between inward- and outward-facing conformations to capture and force substrates out of cells like a peristaltic pump. The high degree of similarity in outward-facing structures across evolution of ABC transporters allowed construction of a high-confidence outward-facing Pgp atomic model based on crystal structures of outward-facing Sav1866 and inward-facing Pgp. The model adhered to previous experimentally determined secondary- and tertiary- configurations during all-atom molecular dynamics simulations in the presence or absence of MgATP. Three long lasting (>100 ns) meta-stable states were apparent in the presence of MgATP revealing new insights into alternating access. The two ATP-binding pockets are highly asymmetric resulting in differential control of overall structural dynamics and allosteric regulation of the drug-binding pocket. Equilibrated Pgp has a considerably different electrostatic profile compared to Sav1866 that implicates significant kinetic and thermodynamic differences in transport mechanisms.

Molecular dynamics study of Zn(aß) and Zn(aß)2.

The aggregation of Aß-peptide (Aß) is widely considered to be the critical step in the pathology of Alzheimer's disease. Small, soluble Aß oligomers have been shown to be more neurotoxic than large, insoluble aggregates and fibrils. Recent studies suggest that biometal ions, including Zn(II), may play an important role in the aggregation process. Experimentally determining the details of the binding process is complicated by the kinetic lability of zinc. To study the dynamic nature of the zinc-bound Aß complexes and the potential mechanisms by which Zn(II) affects Aß oligomerization we have performed atomistic molecular dynamics (MD) simulations of Zn(Aß) and Zn(Aß)2. The models were based on NMR data and predicted coordination environments from previous density functional theory calculations. When modeled as 4-coordinate covalently bound Zn(Aß) n complexes (where n = 1 or 2), zinc imposes conformational changes in the surrounding Aß residues. Moreover, zinc reduces the helix content and increases the random coil content of the full peptide. Although zinc binds at the N-terminus of Aß, ß-sheet formation is observed exclusively at the C-terminus in the Zn(Aß) and most of the Zn(Aß)2 complexes. Furthermore, initial binding to zinc promotes the formation of intra-chain salt-bridges, while subsequent dissociation promotes the formation of inter-chain salt-bridges. These results suggest that Zn-binding to Aß accelerates the aggregation of Aß by unfolding the helical structure in Aß peptide and stabilizing the formation of vital salt-bridges within and between Aß peptides.

Sequential removal of photolabile protecting groups for carbonyls with controlled wavelength.

A group of robust and easy-to-make photolabile protecting groups (PPGs) for carbonyl compounds has been developed. Sequential removal of different PPGs is achieved via control of irradiation wavelength.