International Journal of Molecular Sciences

Six small interference RNAs (siRNAs) have been approved as therapeutics since 2018 making them promising nanosystems due to selective gene knockdown activity. siRNA design is complex due to various factors, where the chemical modifications are crucial to improve its half-life and stability. Machine learning (ML) enabled more efficient analysis of siRNA data, moreover predicting efficacy and off-target effects. This work proposes a novel pipeline for predicting gene knockdown activity of chemically modified siRNAs across the whole range of activities leveraging both descriptors of siRNA chemical composition-aware property matrices and large language model (LLM) embeddings for target gene encoding. Several general-purpose and domain-specific fine-tuned LLMs were benchmarked on the target task, where the Mistral 7B general-purpose model slightly outperformed even the models pre-trained on genomic data. Proposed two-stage probability-enhanced model successfully mitigates data imbalance towards moderate-to-high active constructs and achieves state-of-the-art (SOTA) quality with R² = 0.84 and a RMSE = 12.27% on unseen data, where the probabilistic outputs of classifiers trained with F-scores up to 0.92 were used for regression model supervision. Moreover, leave-one-gene-out (LOGO) experiments show that the model is able to extrapolate on unseen genes, which further shows representativeness of siRNA features and gene embeddings. By filling the gap in the field of advanced chemical composition-aware siRNA design, our model aims to improve the efficacy of developed siRNA-based therapies.

Proteins play an important role in living organisms, and, for most of them, the function depends on their structure. There are some proteins that have similar properties but different structures. An example of this is ice-binding proteins (IBPs), which have different structures but share the ability to bind to ice. Many organisms have evolved such proteins to help them survive in cold environments. Therefore, it is important to study the oligomeric state of the active form in solutions. The activity of IBP is related to the area of their ice-binding site. We have demonstrated the presence of oligomeric forms of protein in solution using multiple techniques, such as mass spectrometry, native gel electrophoresis, atomic force microscopy (AFM), isothermal titration calorimetry (ITC) and small-angle X-ray scattering (SAXS). It is noteworthy that, to date, there have been no reports of the oligomerization of ice-binding protein from Longhorn sculpin. Additionally, our findings suggest that larger molecules may influence the ability of proteins to bind to ice. In our study, the ice-binding protein forms elongated assemblies with limited intermonomer interfaces. The combination of SAXS and AFM data indicates a structure that combines compactness and flexibility and probably consists of four monomeric units. The employment of molecular modelling methodologies resulted in the attainment of a tetrameric complex that is in alignment with AFM data. Details of oligomers observed using the methods in our study emphasize the importance of different techniques that complement each other in resolving structural features. Additionally, we suggest that the protein particles, which were dispersed on the surface, exhibit softness or the form planar complexes with loose quaternary structures. It is conceivable that, depending on ionic strength and/or temperature, the various oligomeric forms of the ice-binding protein form thermodynamically more favorable complexes than their monomeric forms.

Plant protease inhibitors (PPI) play a significant role against microbes, insects, and, to a considerable extent, human pathogens. PPIs inactivate hydrolase enzymes or depolarize the plasma membrane of the pathogens, thereby inhibiting their growth, replication, and invasion. Here, an active serine protease inhibitor was isolated and purified from the seeds of Cleome viscosa. The purified inhibitor was homogenous and exhibited a molecular weight of around 12 kDa as a monomer. The secondary structure analysis indicated that the inhibitor was predominantly composed of α-helical content. The kinetics experiments demonstrated a noncompetitive mode of inhibition towards serine protease when casein was used as the enzyme substrate. The inhibitor formed a stable complex with serine protease, having a likely 1:1 stoichiometry, as inferred from ITC, and the dissociation constant was examined to be K_d = 1.9 × 10⁻⁶ M with a Gibbs free energy of ΔG = −8.079 (kcal/mol). The inhibitor exhibits stable protease inhibition up to 90 °C. Further, in vitro preliminary studies revealed its inhibitory effects against HSV-2 function, evidence that it may have a role in the treatment of viral infections.