In Silico Predictions of Ecological Plasticity Mediated by Protein Family Expansions in Early-Diverging Fungi

Early-diverging fungi (EDF) are ubiquitous and versatile. Their diversity is reflected in their genome sizes and complexity. For instance, multiple protein families have been reported to expand or disappear either in particular genomes or even whole lineages. The most commonly mentioned are CAZymes (carbohydrate-active enzymes), peptidases and transporters that serve multiple biological roles connected to, e.g., metabolism and nutrients intake. In order to study the link between ecology and its genomic underpinnings in a more comprehensive manner, we carried out a systematic in silico survey of protein family expansions and losses among EDF with diverse lifestyles. We found that 86 protein families are represented differently according to EDF ecological features (assessed by median count differences). Among these there are 19 families of proteases, 43 CAZymes and 24 transporters. Some of these protein families have been recognized before as serine and metallopeptidases, cellulases and other nutrition-related enzymes. Other clearly pronounced differences refer to cell wall remodelling and glycosylation. We hypothesize that these protein families altogether define the preliminary fungal adaptasome. However, our findings need experimental validation. Many of the protein families have never been characterized in fungi and are discussed in the light of fungal ecology for the first time.

Coronavirus Genome Sequence Similarity and Protein Sequence Classification

The world currently is going through a serious pandemic due to the coronavirus disease (COVID-19). In this study, we investigate the gene structure similarity of coronavirus genomes isolated from COVID-19 patients, Severe Acute Respiratory Syndrome (SARS) patients and bats genes. We also explore the extent of similarity between their genome structures to find if the new coronavirus is similar to either of the other genome structures. Our experimental results show that there is 82.42% similarity between the CoV-2 genome structure and the bat genome structure. Moreover, we have used a bidirectional Gated Recurrent Unit (GRU) model as the deep learning technique and an improved variant of Recurrent Neural networks (i.e., Bidirectional Long Short Term Memory model) to classify the protein families of these genomes to isolate the prominent protein family accession. The accuracy of Gated Recurrent Unit (GRU) is 98% for labeled protein sequences against the protein families. By comparing the performance of the Gated Recurrent Unit (GRU) model with the Bidirectional Long Short Term Memory (Bi-LSTM) model results, we found that the GRU model is 1.6% more accurate than the Bi-LSTM model for our multiclass protein classification problem. Our experimental results would be further support medical research purposes in targeting the protein family similarity to better understand the coronavirus genomic structure.

Molecular Modeling Insights into Upadacitinib Selectivity upon Binding to JAK Protein Family

Rheumatoid arthritis (RA) is a chronic disease characterized by bone joint damage and incapacitation. The mechanism underlying RA pathogenesis is autoimmunity in the connective tissue. Cytokines play an important role in the human immune system for signal transduction and in the development of inflammatory responses. Janus kinases (JAK) participate in the JAK/STAT pathway, which mediates cytokine effects, in particular interleukin 6 and IFNγ. The discovery of small molecule inhibitors of the JAK protein family has led to a revolution in RA therapy. The novel JAK inhibitor upadacitinib (RinvoqTM) has a higher selectivity for JAK1 compared to JAK2 and JAK3 in vivo. Currently, details on the molecular recognition of JAK1 by upadacitinib are not available. We found that characteristics of hydrogen bond formation with the glycine loop and hinge in JAKs define the selectivity. Our molecular modeling study could provide insight into the drug action mechanism and pharmacophore model differences in JAK isoforms.

Exploring the multifunctionality of SR proteins

Members of the arginine–serine-rich protein family (SR proteins) are multifunctional RNA-binding proteins that have emerged as key determinants for mRNP formation, identity and fate. They bind to pre-mRNAs early during transcription in the nucleus and accompany bound transcripts until they are translated or degraded in the cytoplasm. SR proteins are mostly known for their essential roles in constitutive splicing and as regulators of alternative splicing. However, many additional activities of individual SR proteins, beyond splicing, have been reported in recent years. We will summarize the different functions of SR proteins and discuss how multifunctionality can be achieved. We will also highlight the difficulties of studying highly versatile SR proteins and propose approaches to disentangle their activities, which is transferrable to other multifunctional RBPs.

VDAC2 and the BCL-2 family of proteins

The BCL-2 protein family govern whether a cell dies or survives by controlling mitochondrial apoptosis. As dysregulation of mitochondrial apoptosis is a common feature of cancer cells, targeting protein–protein interactions within the BCL-2 protein family is a key strategy to seize control of apoptosis and provide favourable outcomes for cancer patients. Non-BCL-2 family proteins are emerging as novel regulators of apoptosis and are potential drug targets. Voltage dependent anion channel 2 (VDAC2) can regulate apoptosis. However, it is unclear how this occurs at the molecular level, with conflicting evidence in the literature for its role in regulating the BCL-2 effector proteins, BAK and BAX. Notably, VDAC2 is required for efficient BAX-mediated apoptosis, but conversely inhibits BAK-mediated apoptosis. This review focuses on the role of VDAC2 in apoptosis, discussing the current knowledge of the interaction between VDAC2 and BCL-2 family proteins and the recent development of an apoptosis inhibitor that targets the VDAC2–BAK interaction.

Improving ΔΔG predictions with a multi-task convolutional Siamese Network

The lead optimization phase of drug discovery refines an initial hit molecule for desired properties, especially potency. Synthesis and experimental testing of the small perturbations during this refinement can be quite costly and time consuming. Relative binding free energy (RBFE, also referred to as ∆∆G) methods allow the estimation of binding free energy changes after small changes to a ligand scaffold. Here we propose and evaluate a Convolutional Neural Network (CNN) Siamese network for the prediction of RBFE between two bound ligands. We show that our multi-task loss is able to improve on a previous state-of-the-art Siamese network for RBFE prediction via increased regularization of the latent space. The Siamese network architecture is well suited to the prediction of RBFE in comparison to a standard CNN trained on the same data (Pearson’s R of 0.553 and 0.5, respectively). When evaluated on a left-out protein family, our CNN Siamese network shows variability in its RBFE predictive performance depending on the protein family being evaluated (Pearson’s R ranging from-0.44 to 0.97). RBFE prediction performance can be improved during generalization by injecting only a few examples (few-shot learning) from the evaluation dataset during model training.

AlphaFold-Predicted Structures of KCTD Proteins Unravel Previously Undetected Relationships among the Members of the Family

One of the most striking features of KCTD proteins is their involvement in apparently unrelated yet fundamental physio-pathological processes. Unfortunately, comprehensive structure–function relationships for this protein family have been hampered by the scarcity of the structural data available. This scenario is rapidly changing due to the release of the protein three-dimensional models predicted by AlphaFold (AF). Here, we exploited the structural information contained in the AF database to gain insights into the relationships among the members of the KCTD family with the aim of facilitating the definition of the structural and molecular basis of key roles that these proteins play in many biological processes. The most important finding that emerged from this investigation is the discovery that, in addition to the BTB domain, the vast majority of these proteins also share a structurally similar domain in the C-terminal region despite the absence of general sequence similarities detectable in this region. Using this domain as reference, we generated a novel and comprehensive structure-based pseudo-phylogenetic tree that unraveled previously undetected similarities among the protein family. In particular, we generated a new clustering of the KCTD proteins that will represent a solid ground for interpreting their many functions.