Search

Approximately 30% of Candida genus isolates are resistant to all currently available antifungal drugs and it is highly important to develop new treatments. Additionally, many current drugs are toxic and cause unwanted side effects. 1,3-beta-glucan synthase is an essential enzyme that builds the cell walls of Candida.Targeting CaFKS1, a subunit of the synthase, could be used to fight Candida.In the present study, a machine-learning model based on chemical descriptors was trained to recognise drugs that inhibit CaFKS1. The model attained 96.72% accuracy for classifying between active and inactive drug compounds. Descriptors for FDA-approved and other drugs were calculated, and the model was used to predict the potential activity of these drugs against CaFKS1.Several drugs, including goserelin and icatibant, were detected as active with high confidence. Many of the drugs, interestingly, were gonadotrophin-releasing hormone (GnRH) antagonists or agonists. A literature search found that five of the predicted drugs inhibit Candida experimentally.This study yields promising drugs to be repurposed to combat Candida albicans infection. Future steps include testing the drugs on fungal cells in vitro.

The objective of this research is, with the analysis of existing data of thyroid cancer (TC) metabolites, to develop a machine-learning model that can diagnose TC using metabolite biomarkers. Through data mining, pathway analysis, and machine learning (ML), the model was developed. We identified seven metabolic pathways related to TC: Pyrimidine metabolism, Tyrosine metabolism, Glycine, serine, and threonine metabolism, Pantothenate and CoA biosynthesis, Arginine biosynthesis, Phenylalanine Metabolism, and Phenylalanine, tyrosine, and tryptophan biosynthesis. The ML classifications’ accuracies were confirmed through 10-fold cross validation, and the most accurate classification being 87.30%. The metabolic pathways identified in relation to TC and the changes within such pathways can contribute to more pattern recognition for diagnostics of TC patients and assistance to TC screening. The results also point to a possibility for the development of using ML methods for TC diagnostics and further applications of ML in general cancer-related metabolite analysis.

Context: SARS CoV 2 induced cytokine storm is the major cause of COVID 19 related deaths. Patients have been treated with drugs that work by inhibiting a specific protein partly responsible for the cytokines production. This approach provided very limited success, since there are multiple proteins involved in the complex cell signaling disease mechanisms. Objective: To elucidate using machine learning (ML) the set of drugs targeting a group of proteins involved in the mechanism of cytokine storm. Methods: We selected for targeting five proteins: Angiotensin II receptor type 1 (AT1R), A disintegrin and metalloprotease 17 (ADAM17), Nuclear Factor Kappa B (NF 𝜅B), Janus kinase 1 (JAK1) and Signal Transducer and Activator of Transcription 3 (STAT3) that are involved in the SARS CoV 2 induced cytokine storm pathway. We developed ML models for these five proteins, using known active inhibitors. After developing the model for each of these proteins, FDA-approved drugs were screened to find novel therapeutics for COVID 19. Results: We identified twenty drugs that are active for four proteins and eight drugs active for five proteins. Mitomycin C is the most active drug across all five proteins with an average prediction score of 0.886. For further validation of these results, we used the PyRx software to conduct protein–ligand docking experiments and calculated the binding affinity. The docking results support findings by the ML model. Conclusions: It is possible to elucidate the drugs, targeting simultaneously several proteins related to cytokine production to treat the cytokine storm in COVID 19 patients.

Purpose: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused millions of deaths worldwide, pushing the urgent need for an efficient treatment. Nonstructural protein 15 (NSP15) is a promising target due to its importance for SARS-CoV-2’s evasion of the host’s innate immune response. Methods: Using the crystal structure of SARS-CoV-2 NSP15 endoribonuclease, we developed a pharmacophore model of the functional centers in the NSP15 inhibitor’s binding pocket. With this model, we conducted data mining of the conformational database of FDA-approved drugs. The conformations of these compounds underwent 3D fingerprint similarity clustering, and possible conformers were docked to the NSP15 binding pocket. We also simulated docking of random compounds to the NSP15 binding pocket for comparison. Results: This search identified 170 compounds as potential inhibitors of SARS-CoV-2 NSP15. The mean free energy of docking for the group of potential inhibitors were significantly lower than for the group of random compounds. Twenty-one of the compounds identified as potential NSP15 inhibitors were antiviral compounds used in the inhibition of a range of viruses, including MERS, SARS-CoV, and even SARS-CoV-2. Eight of the selected antiviral compounds in cluster A are pyrimidine analogues, six of which are currently used in a clinical setting. Four tyrosine kinase inhibitors were identified with potential SARS-CoV-2 inhibition, which is consistent with previous studies showing some kinase inhibitors acting as antiviral drugs. Conclusions: We recommended testing of these 21 selected antiviral compounds for the treatment of COVID-19.

P38-alpha (MAPK14) is a protein kinase that is implicated in the pathological mechanisms of BAG3 P209L myofibrillar myopathy, cancers, Alzheimer’s disease and other diseases like rheumatoid arthritis. Inhibition of p38 has shown promise as treatment for these diseases. Traditional drug discovery methods were unable to create both effective and safe small molecule inhibitors, so we used machine learning to elucidate potential p38 blockers from existing FDA-approved drugs. Using available bioactivity data, we determined the best existing p38 inhibitors and applied fingerprint clustering to isolate the compounds with similar structures. Descriptors were calculated for these clustered compounds and the most important of these descriptors were determined through a machine-learning based feature selection algorithm. This data served as the training set for a deep neural network that was fine-tuned to a 92% validation accuracy. The neural network model was applied to a database of FDA-approved drugs, revealing 149 potential p38 inhibitors, whose efficacy were confirmed by docking simulations to be statistically significantly higher than random FDA drugs and slightly higher than known inhibitors. Our study not only reveals potential treatments for p38-mediated diseases but also demonstrates the capability of integrating various machine-learning techniques and computational algorithms to predict novel functions of existing pharmaceuticals.

Neuroligins are postsynaptic cell-adhesion proteins that associate with their presynaptic partners, the neurexins. Using small-angle X-ray scattering, we determined the shapes of the extracellular region of several neuroligin isoforms in solution. We conclude that the neuroligins dimerize via the characteristic four-helix bundle observed in cholinesterases, and that the connecting sequence between the globular lobes of the dimer and the cell membrane is elongated, projecting away from the dimer interface. X-ray scattering and neutron contrast variation data show that two neurexin monomers, separated by 107 A, bind at symmetric locations on opposite sides of the long axis of the neuroligin dimer. Using these data, we developed structural models that delineate the spatial arrangements of different neuroligin domains and their partnering molecules. As mutations of neurexin and neuroligin genes appear to be linked to autism, these models provide a structural framework for understanding altered recognition by these proteins in neurodevelopmental disorders.

To learn more about the evolution of the cholinesterases (ChEs), acetylcholinesterase (AChE) and butyrylcholinesterase in the vertebrates, we investigated the AChE activity of a deuterostome invertebrate, the urochordate Ciona intestinalis, by expressing in vitro a synthetic recombinant cDNA for the enzyme in COS-7 cells. Evidence from kinetics, pharmacology, molecular biology, and molecular modeling confirms that the enzyme is AChE. Sequence analysis and molecular modeling also indicate that the cDNA codes for the AChE(T) subunit, which should be able to produce all three globular forms of AChE: monomers (G(1)), dimers (G(2)), and tetramers (G(4)), and assemble into asymmetric forms in association with the collagenic subunit collagen Q. Using velocity sedimentation on sucrose gradients, we found that all three of the globular forms are either expressed in cells or secreted into the medium. In cell extracts, amphiphilic monomers (G(1)(a)) and non-amphiphilic tetramers (G(4)(na)) are found. Amphiphilic dimers (G(2)(a)) and non-amphiphilic tetramers (G(4)(na)) are secreted into the medium. Co-expression of the catalytic subunit with Rattus norvegicus collagen Q produces the asymmetric A(12) form of the enzyme. Collagenase digestion of the A(12) AChE produces a lytic G(4) form. Notably, only globular forms are present in vivo. This is the first demonstration that an invertebrate AChE is capable of assembling into asymmetric forms. We also performed a phylogenetic analysis of the sequence. We discuss the relevance of our results with respect to the evolution of the ChEs in general, in deuterostome invertebrates, and in chordates including vertebrates.

To study synapse formation by neuroligins, we co-cultured hippocampal neurons with COS cells expressing wild type and mutant neuroligins. The large size of COS cells makes it possible to test the effect of neuroligins presented over an extended surface area. We found that a uniform lawn of wild type neuroligins displayed on the cell surface triggers the formation of hundreds of uniformly sized, individual synaptic contacts that are labeled with neurexin antibodies. Electron microscopy revealed that these artificial synapses contain a presynaptic active zone with docked vesicles and often feature a postsynaptic density. Neuroligins 1, 2, and 3 were active in this assay. Mutations in two surface loops of neuroligin 1 abolished neuroligin binding to neurexin 1beta, a presumptive presynaptic binding partner for postsynaptic neuroligins, and blocked synapse formation. An analysis of mutant neuroligins with an amino acid substitution that corresponds to a mutation described in patients with an autistic syndrome confirmed previous reports that these mutant neuroligins have a compromised capacity to be transported to the cell surface. Nevertheless, the small percentage of mutant neuroligins that reached the cell surface still induced synapse formation. Viewed together, our data suggest that neuroligins generally promote artificial synapse formation in a manner that is associated with beta-neurexin binding and results in morphologically well differentiated synapses and that a neuroligin mutation found in autism spectrum disorders impairs cell-surface transport but does not completely abolish synapse formation activity.

Waglerin-1 (Wtx-1) is a 22-amino acid peptide that competitively antagonizes muscle nicotinic acetylcholine receptors (nAChRs). Previous work demonstrated that Wtx-1 binds to mouse nAChRs with higher affinity than receptors from rats or humans, and distinguished residues in alpha and epsilon subunits that govern the species selectivity. These studies also showed that Wtx-1 binds selectively to the alpha-epsilon binding site with significantly higher affinity than to the alpha-delta binding site. Here we identify residues at equivalent positions in the epsilon, gamma, and delta subunits that govern Wtx-1 selectivity for one of the two binding sites on the nAChR pentamer. Using a series of chimeric and point mutant subunits, we show that residues Gly-57, Asp-59, Tyr-111, Tyr-115, and Asp-173 of the epsilon subunit account predominantly for the 3700-fold higher affinity of the alpha-epsilon site relative to that of the alpha-gamma site. Similarly, we find that residues Lys-34, Gly-57, Asp-59, and Asp-173 account predominantly for the high affinity of the alpha-epsilon site relative to that of the alpha-delta site. Analysis of combinations of point mutations reveals that Asp-173 in the epsilon subunit is required together with the remaining determinants in the epsilon subunit to achieve Wtx-1 selectivity. In particular, Lys-34 interacts with Asp-173 to confer high affinity, resulting in a DeltaDeltaG(INT) of -2.3 kcal/mol in the epsilon subunit and a DeltaDeltaG(INT) of -1.3 kcal/mol in the delta subunit. Asp-173 is part of a nonhomologous insertion not found in the acetylcholine binding protein structure. The key role of this insertion in Wtx-1 selectivity indicates that it is proximal to the ligand binding site. We use the binding and interaction energies for Wtx-1 to generate structural models of the alpha-epsilon, alpha-gamma, and alpha-delta binding sites containing the nonhomologous insertion.