Protein Design Using Physics Informed Neural Networks.

The inverse protein folding problem, also known as protein sequence design, seeks to predict an amino acid sequence that folds into a specific structure and performs a specific function. Recent advancements in machine learning techniques have been successful in generating functional sequences, outperforming previous energy function-based methods. However, these machine learning methods are limited in their interoperability and robustness, especially when designing proteins that must function under non-ambient conditions, such as high temperature, extreme pH, or in various ionic solvents. To address this issue, we propose a new Physics-Informed Neural Networks (PINNs)-based protein sequence design approach. Our approach combines all-atom molecular dynamics simulations, a PINNs MD surrogate model, and a relaxation of binary programming to solve the protein design task while optimizing both energy and the structural stability of proteins. We demonstrate the effectiveness of our design framework in designing proteins that can function under non-ambient conditions.

Modeling the structure of the frameshift-stimulatory pseudoknot in SARS-CoV-2 reveals multiple possible conformers.

The coronavirus causing the COVID-19 pandemic, SARS-CoV-2, uses -1 programmed ribosomal frameshifting (-1 PRF) to control the relative expression of viral proteins. As modulating -1 PRF can inhibit viral replication, the RNA pseudoknot stimulating -1 PRF may be a fruitful target for therapeutics treating COVID-19. We modeled the unusual 3-stem structure of the stimulatory pseudoknot of SARS-CoV-2 computationally, using multiple blind structural prediction tools followed by µs-long molecular dynamics simulations. The results were compared for consistency with nuclease-protection assays and single-molecule force spectroscopy measurements of the SARS-CoV-1 pseudoknot, to determine the most likely conformations. We found several possible conformations for the SARS-CoV-2 pseudoknot, all having an extended stem 3 but with different packing of stems 1 and 2. Several conformations featured rarely-seen threading of a single strand through junctions formed between two helices. These structural models may help interpret future experiments and support efforts to discover ligands inhibiting -1 PRF in SARS-CoV-2.

GLUT1 and TUBB4 in Glioblastoma Could be Efficacious Targets.

Glioblastoma multiforme (GBM) is the most aggressive and deadly brain tumor, portending a median 13-month survival even following gross total resection with adjuvant chemotherapy and radiotherapy. This prognosis necessitates improved therapies for the disease. A target of interest for novel chemotherapies is the Warburg Effect, which describes the tumor's shift away from oxidative phosphorylation towards glycolysis. Here, we elucidate GLUT1 (Glucose transporter 1) and one of its associated binding partners, TUBB4 (Tubulin 4), as potentially druggable targets in GBM. Using data mining approach, we demonstrate that GLUT1 is overexpressed as a function of tumor grade in astrocytoma's and that its overexpression is associated with poorer prognosis. Using both mass spectrometry performed on hGBM (human glioblastoma patient specimen) and in silico modeling, we show that GLUT1 interacts with TUBB4, and more accurately demonstrates GLUT1's binding with fasentin. Proximity ligation assay (PLA) and immunoprecipitation studies confirm GLUT1 interaction with TUBB4. Treatment of GSC33 and GSC28 cells with TUBB4 inhibitor, CR-42-24, reduces the expression of GLUT1 however, TUBB4 expression is unaltered upon fasentin treatment. Using human pluripotent stem cell antibody array, we demonstrate reduced levels of Oct3/4, Nanog, Sox2, Sox17, Snail and VEGFR2 (Vascular endothelial growth factor receptor 2) upon CR-42-24 treatment. Overall, our data confirm that silencing TUBB4 or GLUT1 reduce GSC tumorsphere formation, self-renewal and proliferation in vitro. These findings suggest GLUT1 and its binding partner TUBB4 as druggable targets that warrant further investigation in GBM.

Virtual screening using covalent docking to find activators for G245S mutant p53.

TP53 is the most mutated gene in all cancers. The mutant protein also accumulates in cells. The high frequency of p53 mutations makes the protein a promising target for anti-cancer therapy. Only a few molecules have been found, using in vitro screening, to reactivate the mutant protein. APR-246 is currently the most successful mutant p53 activator, which reactivates the transcriptional activity of p53 by covalently binding to C124 of the protein. We have recently created in silico models of G245S-mp53 in its apo and DNA-bound forms. In this paper we further report on our in silico screening for potential activators of G245S-mp53. We filtered the ZINC15 database (13 million compounds) to only include drug-like molecules with moderate to standard reactivity. Our filtered database of 130,000 compounds was screened using the DOCKTITE protocol in the Molecular Operating Environment software. We performed covalent docking at C124 of G245S-mp53 to identify potential activators of the mutant protein. The docked compounds were ranked using a consensus scoring approach. We also used ADMET Predictor™ to predict pharmacokinetics and the possible toxicities of the compounds. Our screening procedure has identified compounds, mostly thiosemicarbazones and halo-carbonyls, with the best potential as G245S-mp53 activators, which are described in this work. Based on its binding scores and ADMET risk score, compound 2 is likely to have the best potential as a G245S-mp53 activator compared to the other top hits.

The molecular mechanism of action of methylene quinuclidinone and its effects on the structure of p53 mutants.

One of the most important tumor suppressor proteins in eukaryotic cells is the transcription factor called p53. The importance of this protein in cells comes from the fact that it regulates a wide variety of cellular processes including the cell cycle, metabolism, DNA repair, senescence and apoptosis. In cancer cells, p53 is a major target as the most mutated protein, which has led to the search for potential activators of the mutant protein. Currently, the only mutated-p53 activator in clinical trials is a small molecule called APR-246. There is evidence that the active metabolite of APR-246 binds covalently to mutant p53 and restores its wild-type (wt) activity. In this work, we created atomistic in silico models of the wt, mutant and drugged mutant p53 proteins each in complex with DNA. Using molecular dynamics simulations we generated equilibrated models of the complexes. Detailed analysis revealed that the binding of the APR-246 active metabolite to the mutant proteins alters their interaction with DNA. In particular, the binding of the molecule at loop L1 of the protein allows the loop to anchor the protein to DNA similarly to wt p53. Several important p53-DNA interactions lost due to mutation were also restored in the drugged mutants. These findings, not only provide a possible mechanism of action of this drug, but also criteria to use in virtual screening campaigns for other p53 activators.

Insights into the Effect of the G245S Single Point Mutation on the Structure of p53 and the Binding of the Protein to DNA.

The transcription factor p53 is a potent tumor suppressor dubbed as the "guardian of the genome" because of its ability to orchestrate protective biological outputs in response to a variety of oncogenic stresses. Mutation and thus inactivation of p53 can be found in 50% of human tumors. The majority are missense mutations located in the DNA binding region. Among them, G245S is known to be a structural hotspot mutation. To understand the behaviors and differences between the wild-type and mutant, both a dimer of the wild type p53 (wt-p53) and its G245S mutant (G245S-mp53), complexed with DNA, were simulated using molecular dynamics for more than 1 µs. wt-p53 and G245S-mp53 apo monomers were simulated for 1 µs as well. Conformational analyses and binding energy evaluations performed underline important differences and therefore provide insights to understand the G245S-mp53 loss of function. Our results indicate that the G245S mutation destabilizes several structural regions in the protein that are crucial for DNA binding when found in its apo form and highlight differences in the mutant-DNA complex structure compared to the wt protein. These findings not only provide means that can be applied to other p53 mutants but also serve as structural basis for further studies aimed at the development of cancer therapies based on restoring the function of p53.

A computational method for selecting short peptide sequences for inorganic material binding.

Discovering or designing biofunctionalized materials with improved quality highly depends on the ability to manipulate and control the peptide-inorganic interaction. Various peptides can be used as assemblers, synthesizers, and linkers in the material syntheses. In another context, specific and selective material-binding peptides can be used as recognition blocks in mining applications. In this study, we propose a new in silico method to select short 4-mer peptides with high affinity and selectivity for a given target material. This method is illustrated with the calcite (104) surface as an example, which has been experimentally validated. A calcite binding peptide can play an important role in our understanding of biomineralization. A practical aspect of calcite is a need for it to be selectively depressed in mining sites.

Protein Structural Analysis of Calbindin D28k Function and Dysregulation: Potential Competition Between Ca(2+) and Zn(2.).

Calcium homeostasis is an essential physiological process requiring tight control in the normal cell. The dysregulation of calcium homeostasis may play a key role in the onset of Alzheimer's disease (AD) and other disorders, whether through the loss of calcium binding or calcium sensing capacity. Calbindin D28k (CB-D28k), a calcium binding protein composed of six EF-hands, four of which can bind Ca(2+), has been implicated in AD-related calcium dysregulation. In this study, docking and molecular dynamics calculations were employed to refine the protein data base model in order to understand the underlying structural variations between functional and non-functional EF-hands. Molecular modeling calculations improved the modelled protein structure: helix-loop-helix sequences were formed in all hands and most canonical interactions were formed in the four functional hands. The protein can also bind Zn(2+), potentially altering the Ca(2+) binding capability. Analysis of calculated structures of Zn(2+) bound protein showed that only half of the correct EF-hand canonical interactions of Ca(2+) were formed with loop residues. These results have important implications for the understanding of calcium dysregulation as well as for the development of novel therapeutic strategies in AD and other central nervous system disease processes, or in conditions of brain injury where calcium homeostasis is compromised.

Ranking the Binding Energies of p53 Mutant Activators and Their ADMET Properties.

The guardian of the genome, p53, is the most mutated protein found in all cancer cells. Restoration of wild-type activity to mutant p53 offers promise to eradicate cancer cells using novel pharmacological agents. Several molecules have already been found to activate mutant p53. While the exact mechanism of action of these compounds has not been fully understood, a transiently open pocket has been identified in some mutants. In our study, we docked twelve known activators to p53 into the open pocket to further understand their mechanism of action and rank the best binders. In addition, we predicted the absorption, distribution, metabolism, excretion and toxicity properties of these compounds to assess their pharmaceutical usefulness. Our studies showed that alkylating ligands do not all bind at the same position, probably due to their varying sizes. In addition, we found that non-alkylating ligands are capable of binding at the same pocket and directly interacting with Cys124. The comparison of the different ligands demonstrates that stictic acid has a great potential as a p53 activator in terms of less adverse effects although it has poorer pharmacokinetic properties.

Algebraic and topological indices of molecular pathway networks in human cancers.

Protein-protein interaction networks associated with diseases have gained prominence as an area of research. We investigate algebraic and topological indices for protein-protein interaction networks of 11 human cancers derived from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. We find a strong correlation between relative automorphism group sizes and topological network complexities on the one hand and five year survival probabilities on the other hand. Moreover, we identify several protein families (e.g. PIK, ITG, AKT families) that are repeated motifs in many of the cancer pathways. Interestingly, these sources of symmetry are often central rather than peripheral. Our results can aide in identification of promising targets for anti-cancer drugs. Beyond that, we provide a unifying framework to study protein-protein interaction networks of families of related diseases (e.g. neurodegenerative diseases, viral diseases, substance abuse disorders).

Mathematical and computational modeling in biology at multiple scales.

A variety of topics are reviewed in the area of mathematical and computational modeling in biology, covering the range of scales from populations of organisms to electrons in atoms. The use of maximum entropy as an inference tool in the fields of biology and drug discovery is discussed. Mathematical and computational methods and models in the areas of epidemiology, cell physiology and cancer are surveyed. The technique of molecular dynamics is covered, with special attention to force fields for protein simulations and methods for the calculation of solvation free energies. The utility of quantum mechanical methods in biophysical and biochemical modeling is explored. The field of computational enzymology is examined.