Rossmann-toolbox: a deep learning-based protocol for the prediction and design of cofactor specificity in Rossmann fold proteins.

The Rossmann fold enzymes are involved in essential biochemical pathways such as nucleotide and amino acid metabolism. Their functioning relies on interaction with cofactors, small nucleoside-based compounds specifically recognized by a conserved ßαß motif shared by all Rossmann fold proteins. While Rossmann methyltransferases recognize only a single cofactor type, the S-adenosylmethionine, the oxidoreductases, depending on the family, bind nicotinamide (nicotinamide adenine dinucleotide, nicotinamide adenine dinucleotide phosphate) or flavin-based (flavin adenine dinucleotide) cofactors. In this study, we showed that despite its short length, the ßαß motif unambiguously defines the specificity towards the cofactor. Following this observation, we trained two complementary deep learning models for the prediction of the cofactor specificity based on the sequence and structural features of the ßαß motif. A benchmark on two independent test sets, one containing ßαß motifs bearing no resemblance to those of the training set, and the other comprising 38 experimentally confirmed cases of rational design of the cofactor specificity, revealed the nearly perfect performance of the two methods. The Rossmann-toolbox protocols can be accessed via the webserver at https://lbs.cent.uw.edu.pl/rossmann-toolbox and are available as a Python package at https://github.com/labstructbioinf/rossmann-toolbox.

MINT: software to identify motifs and short-range interactions in trajectories of nucleic acids.

Structural biology experiments and structure prediction tools have provided many high-resolution three-dimensional structures of nucleic acids. Also, molecular dynamics force field parameters have been adapted to simulating charged and flexible nucleic acid structures on microsecond time scales. Therefore, we can generate the dynamics of DNA or RNA molecules, but we still lack adequate tools for the analysis of the resulting huge amounts of data. We present MINT (Motif Identifier for Nucleic acids Trajectory) - an automatic tool for analyzing three-dimensional structures of RNA and DNA, and their full-atom molecular dynamics trajectories or other conformation sets (e.g. X-ray or nuclear magnetic resonance-derived structures). For each RNA or DNA conformation MINT determines the hydrogen bonding network resolving the base pairing patterns, identifies secondary structure motifs (helices, junctions, loops, etc.) and pseudoknots. MINT also estimates the energy of stacking and phosphate anion-base interactions. For many conformations, as in a molecular dynamics trajectory, MINT provides averages of the above structural and energetic features and their evolution. We show MINT functionality based on all-atom explicit solvent molecular dynamics trajectory of the 30S ribosomal subunit.

Electrostatic interactions in aminoglycoside-RNA complexes.

Electrostatic interactions often play key roles in the recognition of small molecules by nucleic acids. An example is aminoglycoside antibiotics, which by binding to ribosomal RNA (rRNA) affect bacterial protein synthesis. These antibiotics remain one of the few valid treatments against hospital-acquired infections by Gram-negative bacteria. It is necessary to understand the amplitude of electrostatic interactions between aminoglycosides and their rRNA targets to introduce aminoglycoside modifications that would enhance their binding or to design new scaffolds. Here, we calculated the electrostatic energy of interactions and its per-ring contributions between aminoglycosides and their primary rRNA binding site. We applied either the methodology based on the exact potential multipole moment (EPMM) or classical molecular mechanics force field single-point partial charges with Coulomb formula. For EPMM, we first reconstructed the aspherical electron density of 12 aminoglycoside-RNA complexes from the atomic parameters deposited in the University at Buffalo Databank. The University at Buffalo Databank concept assumes transferability of electron density between atoms in chemically equivalent vicinities and allows reconstruction of the electron densities from experimental structural data. From the electron density, we then calculated the electrostatic energy of interaction using EPMM. Finally, we compared the two approaches. The calculated electrostatic interaction energies between various aminoglycosides and their binding sites correlate with experimentally obtained binding free energies. Based on the calculated energetic contributions of water molecules mediating the interactions between the antibiotic and rRNA, we suggest possible modifications that could enhance aminoglycoside binding affinity.

Harnessing the power of AI in precision medicine: NGS-based therapeutic insights for colorectal cancer cohort.

Purpose: Developing innovative precision and personalized cancer therapeutics is essential to enhance cancer survivability, particularly for prevalent cancer types such as colorectal cancer. This study aims to demonstrate various approaches for discovering new targets for precision therapies using artificial intelligence (AI) on a Polish cohort of colorectal cancer patients. Methods: We analyzed 71 patients with histopathologically confirmed advanced resectional colorectal adenocarcinoma. Whole exome sequencing was performed on tumor and peripheral blood samples, while RNA sequencing (RNAseq) was conducted on tumor samples. We employed three approaches to identify potential targets for personalized and precision therapies. First, using our in-house neoantigen calling pipeline, ARDentify, combined with an AI-based model trained on immunopeptidomics mass spectrometry data (ARDisplay), we identified neoepitopes in the cohort. Second, based on recurrent mutations found in our patient cohort, we selected corresponding cancer cell lines and utilized knock-out gene dependency scores to identify synthetic lethality genes. Third, an AI-based model trained on cancer cell line data was employed to identify cell lines with genomic profiles similar to selected patients. Copy number variants and recurrent single nucleotide variants in these cell lines, along with gene dependency data, were used to find personalized synthetic lethality pairs. Results: We identified approximately 8,700 unique neoepitopes, but none were shared by more than two patients, indicating limited potential for shared neoantigenic targets across our cohort. Additionally, we identified three synthetic lethality pairs: the well-known APC-CTNNB1 and BRAF-DUSP4 pairs, along with the recently described APC-TCF7L2 pair, which could be significant for patients with APC and BRAF variants. Furthermore, by leveraging the identification of similar cancer cell lines, we uncovered a potential gene pair, VPS4A and VPS4B, with therapeutic implications. Conclusion: Our study highlights three distinct approaches for identifying potential therapeutic targets in cancer patients. Each approach yielded valuable insights into our cohort, underscoring the relevance and utility of these methodologies in the development of precision and personalized cancer therapies. Importantly, we developed a novel AI model that aligns tumors with representative cell lines using RNAseq and methylation data. This model enables us to identify cell lines closely resembling patient tumors, facilitating accurate selection of models needed for in vitro validation.

Thermodynamics of the fourU RNA thermal switch derived from molecular dynamics simulations and spectroscopic techniques.

Bacterial 5' untranslated regions of mRNA, termed thermal switches or thermometers, change their structure in response to temperature change. This structural change provides for the regulation of gene expression. One of such thermal switches, called fourU, is present in the Salmonella species. Mutations of fourU were found to abrogate its regulatory properties. We investigated the thermodynamics of the fourU fragment responsible for its structural changes. All-atom molecular dynamics simulations at various temperatures and spectroscopic experiments in solution were performed for the wild-type fourU and its mutants. We found that the U11C and A8C mutations stabilize the fourU structure in comparison to the wild-type fourU, and the double-point G14A/C25U mutant has the most destabilizing effect on the fourU hairpin 2 responsible for temperature sensing. The G14A/C25U mutant is also the easiest to strand-invade by a complementary oligonucleotide as indicated by fluorescence spectroscopy experiments.

Thermal Stability of Peptide Nucleic Acid Complexes.

Peptide nucleic acid (PNA) is a neutral nucleic acid analogue that base pairs with itself and natural nucleic acids. PNA-nucleic acid complexes are more thermally stable than the corresponding complexes of natural nucleic acids. In addition, PNA is biostable and thus used in many antisense and antigene applications to block functional RNA or DNA via sequence-specific interactions. We have recently developed force field parameters for molecular dynamics (MD) simulations of PNA and PNA-involving duplexes with natural nucleic acids. In this work, we provide the first application of this force field to biologically relevant PNA sequences and their complexes with RNA. We investigated thermal stabilities of short PNA-PNA, PNA-RNA, and RNA-RNA duplexes using UV-monitored thermal denaturation experiments and MD simulations at ambient and elevated temperatures. The simulations show a two-state melting transition and reproduce the thermal stability from melting experiments, with PNA-PNA being the most and RNA-RNA the least stable. The PNA-PNA duplex also displays the highest activation energy for melting. The atomistic details of unfolding of PNA duplexes suggest that all PNA-PNA bases melt concomitantly, whereas the RNA-RNA and PNA-RNA are destabilized from the termini toward the central part of the duplexes.

Improved Force Fields for Peptide Nucleic Acids with Optimized Backbone Torsion Parameters.

Peptide nucleic acids are promising nucleic acid analogs for antisense therapies as they can form stable duplex and triplex structures with DNA and RNA. Computational studies of PNA-containing duplexes and triplexes are an important component for guiding their design, yet existing force fields have not been well validated and parametrized with modern computational capabilities. We present updated CHARMM and Amber force fields for PNA that greatly improve the stability of simulated PNA-containing duplexes and triplexes in comparison with experimental structures and allow such systems to be studied on microsecond time scales. The force field modifications focus on reparametrized PNA backbone torsion angles to match high-level quantum mechanics reference energies for a model compound. The microsecond simulations of PNA-PNA, PNA-DNA, PNA-RNA, and PNA-DNA-PNA complexes also allowed a comprehensive analysis of hydration and ion interactions with such systems.

Interactions of 2'-O-methyl oligoribonucleotides with the RNA models of the 30S subunit A-site.

Synthetic oligonucleotides targeting functional regions of the prokaryotic rRNA could be promising antimicrobial agents. Indeed, such oligonucleotides were proven to inhibit bacterial growth. 2'-O-methylated (2'-O-Me) oligoribonucleotides with a sequence complementary to the decoding site in 16S rRNA were reported as inhibitors of bacterial translation. However, the binding mode and structures of the formed complexes, as well as the level of selectivity of the oligonucleotides between the prokaryotic and eukaryotic target, were not determined. We have analyzed three 2'-O-Me oligoribonucleotides designed to hybridize with the models of the prokaryotic rRNA containing two neighboring aminoglycoside binding pockets. One pocket is the paromomycin/kanamycin binding site corresponding to the decoding site in the small ribosomal subunit and the other one is the close-by hygromycin B binding site whose dynamics has not been previously reported. Molecular dynamics (MD) simulations, as well as isothermal titration calorimetry, gel electrophoresis and spectroscopic studies have shown that the eukaryotic rRNA model is less conformationally stable (in terms of hydrogen bonds and stacking interactions) than the corresponding prokaryotic one. In MD simulations of the eukaryotic construct, the nucleotide U1498, which plays an important role in correct positioning of mRNA during translation, is flexible and spontaneously flips out into the solvent. In solution studies, the 2'-O-Me oligoribonucleotides did not interact with the double stranded rRNA models but all formed stable complexes with the single-stranded prokaryotic target. 2'-O-Me oligoribonucleotides with one and two mismatches bound less tightly to the eukaryotic target. This shows that at least three mismatches between the 2'-O-Me oligoribonucleotide and eukaryotic rRNA are required to ensure target selectivity. The results also suggest that, in the ribosome environment, the strand invasion is the preferred binding mode of 2'-O-Me oligoribonucleotides targeting the aminoglycoside binding sites in 16S rRNA.