Biology subjects

Popowicz, G. M.

Publications and source records attributed to Popowicz, G. M..

11 recordsLinked to original sources

bioRxiv2026

Local structural preference maps encode transferable protein-interface energetics

Relating the drivers of binding affinity to structural features remains difficult because affinity emerges from many weak, context-dependent interactions, while experimental affinity and mutation data are limited. Here, we ask whether a machine-learning model can learn energetically relevant interaction preferences directly from native structures, without affinity or mutation labels. We decompose interfaces into local structural motifs and learn the compatibility of each motif with its surrounding molecular environment. Applied across a protein surface, these models yield target preference maps (TPMs), spatial fields of local interaction compatibility. Trained exclusively on native antibody-antigen structures without affinity or mutation labels, TPMs recover charged, aromatic and backbone-mediated recognition across peptide-protein and other non-antibody interfaces. On an out-of-domain SKEMPI benchmark, wild-type TPM scores discriminated mutation-sensitive positions (AUC 0.643, n=1323), while TPM-based substitution scores ranked alternative amino acids comparably to FoldX and Rosetta Flex {Delta}{Delta}G. These results show that native interface structures can supervise learning of transferable, energetically relevant molecular preferences without direct thermodynamic labels.

bioinformatics↗

bioRxiv2025

It is Not Always Crystal Clear: In-Pocket Analysis for Understanding Binding Site Microenvironment Properties

Structural biology enabled key advances in biochemistry, drug discovery, and AI-driven three-dimensional structure modelling. Despite its importance, reliable and general computational tools for chemical validation of structural models remain limited, especially when small-molecule ligands are involved. This challenge is exacerbated when experimental resolution is insufficient to define interaction chemistry, leaving critical interpretations ambiguous. We introduce In-Pocket, a quantum mechanical optimisation method biased by experimental data that evaluates the chemical nature of protein-ligand binding interfaces. In-Pocket improves model quality, corrects chemically inconsistent structures, and identifies the energetically most favourable ligand-conformers within biomacromolecular complexes. The approach is generally applicable to a wide range of systems, including those containing transition metals. Benchmarking demonstrates that In-Pocket enhances the reliability of legacy and newly solved structures. With automated, user-friendly routine chemical validation, In-Pocket offers unique opportunities to improve the integrity of structural databases and impact diverse domains: from molecular biology to rational drug design and application of generative AI.

bioinformatics↗

bioRxiv2025

A universal model for drug-receptor interactions

Modern AI models can decode the genomic landscape and protein structure world. Yet, they fail to generalize to one of the most important fields: small-molecule drug discovery. Since the late 1970s, the advent of macromolecular crystallography inspired the notion that structural knowledge alone could enable a "lock-and-key" approach to drug design. However, drug discovery continues to depend on costly, resource-intensive, and largely serendipitous screening campaigns that probe only an infinitesimal fraction of the drug-like chemical space. Despite some successful cases, our understanding of, and reasoning from, non-bonded interaction chemistry remains limited for general applicability. Furthermore, though structural databases contain hundreds of thousands of entries, a strong historical bias pervades protein-drug structures, hindering reliable advances through AI scaling. Here, we present a machine-learning framework that learns atom-type-specific spatial preference maps from local protein microenvironments in protein-ligand structures. By excluding ligand topology from the model input and learning from local atom-level environments, the framework is designed to reduce dependence on whole-ligand memorization and to capture transferable interaction preferences. The resulting maps recover chemically meaningful interaction patterns, including cases involving bridging waters and metal-dependent environments. The model was validated using retrospective and prospective real-world data in drug optimization when targeting a challenging protein-protein interface. This shows that the method can provide interpretable workflows to guide molecule optimization and provide input for downstream generative or docking workflows.

bioinformatics↗

bioRxiv2025

The essential co-chaperone Sgt1 regulates client dwell time in the Hsp90 chaperone cycle

The Hsp90 machinery is the most complex chaperone system in the eukaryotic cell. It is characterized by numerous co-chaperones that modulate the function of Hsp90. In S. cerevisiae, most of these cofactors can be deleted without affecting viability. Of the three essential ones, only the function of Sgt1 remained enigmatic. Our in vivo and in vitro experiments define key structural elements and determine the essential function of Sgt1 in the chaperoning of client proteins. We show that yeast Sgt1 exhibits a unique binding mode to Hsp90. The simultaneous interaction of Sgt1 with Hsp90 and client proteins enhances client maturation efficiency. Specifically, Sgt1 stabilizes Hsp90-client complexes and prevents their dissociation by the co-chaperone Aha1. Together, our findings reveal a distinct regulatory mechanism of the Hsp90 function, highlighting Sgt1 as a critical modulator of chaperone cycle progression.

biochemistry↗

bioRxiv2025

Revealing the unexpected interplay between the Proteasome Activator PA200 and the immunoproteasome

The proteasome activator PA200 binds to the catalytic core of the proteasome, the 20S, and activates its proteolytic activities. The cellular function of PA200 is poorly understood and appears to be cell type and differentiation specific. Recent evidence suggests that PA200 not only binds to the standard 20S (s20S) proteasome but also to the specialized immunoproteasome (i20S) which plays a key role in anti-viral and anti-tumor immunity. We here investigated the interaction of PA200 and the immunoproteasome in detail. We show the very first cryo-EM structures of the singly- and doubly-capped i20S-PA200 complexes that revealed no major difference regarding the first binding event of PA200 to the i20S vs. the s20S. However, first PA200 binding triggered a subtle and long range allosteric bending of the i20S barrel which was not seen in the s20S-PA200 complexes. This resulted in major structural rearrangements in the opposite unbound ring - the displacement of atoms up to 5.4 [A] and the increase in its outer diameter - thereby increasing the occupancy of the second PA200 binding site. Mass photometry confirmed higher occupancy of PA200 to the i20S versus the s20S. Binding of PA200 to the i20S enhanced proteasomal activation compared to the s20S. Co-expression of PA200 and the i20S in cells and tissues, however, is restricted but their interaction is favored upon co-expression. The expression of PA200 and the catalytic subunits of the i20S is differentially regulated depending on the cellular context. Our data also suggest that PA200 has the potential to regulate i20S gene expression whereas the i20S has no effects on PA200 expression. Overall, this work sheds new light on the interaction of PA200 with the i20S from a structural, mechanistic and cellular point of view. Importantly, we identify PA200 as a key regulator of the i20S whenever PA200 and the catalytic subunits of the i20S are co-expressed in the same cell.

biochemistry↗

bioRxiv2024

Curcumin Inhibition of DYRK Kinases

Curcumin is known as a dietary supplement with several health benefits, including antioxidant, blood sugar-lowering, anti-inflammatory, and anti-cancer properties. These benefits make it useful in the treatment of diabetes, neurodegenerative diseases, and cancer. This research focuses on how curcumin interacts with dual-specificity tyrosine-regulated kinases (DYRKs), particularly DYRK1A and DYRK2, which control important cellular processes such as protein breakdown, gene activity and DNA packaging. DYRK1A is crucial for the growth and survival of pancreatic {beta}-cells, which are key in the treatment of diabetes. Curcumin helps increase insulin release and sensitivity by affecting these cells. DYRK2, on the other hand, is involved in cancer, where curcumins ability to block its activity helps to reduce tumour growth and spread. These interactions demonstrate curcumins potential as a versatile treatment option. By affecting DYRK1A and DYRK2, curcumin promotes {beta}-cell health in diabetes and fights cancer by inhibiting cell growth and promoting cell death. This study underscores the promise of curcumin as a natural compound for treating complex diseases associated with oxidative stress and inflammation, highlighting the need for a better understanding of DYRK kinases to effectively exploit their role in disease processes and cell signalling.

molecular biology↗

bioRxiv2024

Quantum Chemistry in a Pocket: A Multifaceted Tool to Link Structure and Activity

We introduce In-Pocket Analysis, a simple and efficient protein-ligand complex structure optimization algorithm. It provides structural biology and structure-based drug discovery with a much-needed tool for data curation and quantification analysis. In-pocket analysis removes unphysical tension from experimental structures, yielding high-quality atomic models. This includes deformed bonds, structural clashes, proton refinement, and solvent envelope. The algorithm is compatible with quantum methods or force fields, delivering precise calculations of binding energies and QSAR. Its applications include refinement of experimental structures, calculation of inter- and intramolecular energetics, scoring of docked poses, scaffold decoration/hopping, or building of explainable QSAR. For validation and benchmarking we used high-resolution crystal structures of the N-terminal domain of PEX14 with small-molecule inhibitors. PEX14 is a target candidate against Trypanosomiasis and Leishmaniasis. The application of our algorithm allowed for an explanation of unexpected binding poses, rationalization of binding energies, and affinities as well as precise quantification of solvent-mediated interactions. TABLE OF CONTENTS O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=160 SRC="FIGDIR/small/590615v1_ufig1.gif" ALT="Figure 1"> View larger version (60K): org.highwire.dtl.DTLVardef@11f4173org.highwire.dtl.DTLVardef@1020358org.highwire.dtl.DTLVardef@12986ecorg.highwire.dtl.DTLVardef@1590617_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2023

Can Quantum Chemistry Improve the understanding of Protein-Ligand Interactions? Implications for Structure Based Drug Discovery

We introduce an Energy Decomposition Analysis suitable for understanding the nature of non-covalent binding in large chemical systems, like those of drug-protein complexes. The method is atom specific, thus allowing rationalization of the role that each atom or functional group plays for the interaction. Visual representations are constructed in the form of interaction maps, depicting the different contributions for electrostatics, polarization, dispersion (lipophilicity), etc. This marks the departure from atomistic models towards electronic interaction ones, that better correlate with experimental data. The maps provide a quick access to the driving forces behind the formation of intermolecular complexes, and the key contributors for each interaction. This allows constructing quantum mechanical models of binding. The presented method is validated against experimental binding data for the difficult to target protein-protein interface for PEX14-PEX5 and its inhibitors.

bioinformatics↗

bioRxiv2023

MISATO - Machine learning dataset for structure-based drug discovery

Large language models (LLMs) have greatly enhanced our ability to understand biology and chemistry. Yet, relatively few robust methods have been reported for structure-based drug discovery. Highly precise biomolecule-ligand interaction datasets are urgently needed in particular for LLMs, that require extensive training data. We present MISATO, the first dataset that combines quantum mechanics properties of small molecules and associated molecular dynamics simulations of about 20000 experimental protein-ligand complexes. Starting from the PDBbind dataset, semi-empirical quantum mechanics was used to systematically refine these structures. The largest collection to date of molecular dynamics traces of protein-ligand complexes in explicit water are included, accumulating to 170 s. We give ML baseline models and simple Python data loaders, and aim to foster a thriving community around MISATO (https://github.com/t7morgen/misato-dataset). An easy entry point for ML experts is provided without the need of deep domain expertise to enable the next generation of drug discovery AI models.

bioinformatics↗

bioRxiv2021

Acriflavine, a clinically aproved drug, inhibits SARS-CoV-2 and other betacoronaviruses

SO_SCPLOWUMMARYC_SCPLOWThe COVID-19 pandemic caused by SARS-CoV-2 has been socially and economically devastating. Despite an unprecedented research effort, effective therapeutics are still missing to limit severe disease and mortality. Using high-throughput screening, we identified acriflavine as a potent papain-like protease (PLpro) inhibitor. NMR titrations and a co-crystal structure confirm that acriflavine blocks the PLpro catalytic pocket in an unexpected binding mode. We show that the drug inhibits viral replication at nanomolar concentration in cellular models, in vivo in mice and ex vivo in human airway epithelia, with broad range activity against SARS-CoV-2 and other betacoronaviruses. Considering that acriflavine is an inexpensive drug approved in some countries, it may be immediately tested in clinical trials and play an important role during the current pandemic and future outbreaks.

microbiology↗

bioRxiv2020

Deep learning model can predict water binding sites on the surface of proteins using limited-resolution data

The surfaces of proteins are generally hydrophilic but there have been reports of sites that exhibit an exceptionally high affinity for individual water molecules. Not only do such molecules often fulfil critical biological functions, but also, they may alter the binding of newly designed drugs. In crystal structures, sites consistently occupied in each unit cell yield electron density clouds that represent water molecule presence. These are recorded in virtually all high-resolution structures obtained through X-ray diffraction. In this work, we utilized the wealth of data from the RCSB Protein Data Bank to train a residual deep learning model named hotWater to identify sites on the surface of proteins that are most likely to bind water, the so-called water hot spots. The model can be used to score existing water molecules from a PDB file to provide their ranking according to the predicted binding strength or to scan the surface of a protein to determine the most likely water hot-spots de novo. This is computationally much more efficient than currently used molecular dynamics simulations. Based on testing the model on three example proteins, which have been resolved using both high-resolution X-ray crystallography (providing accurate positions of trapped waters) as well as low-resolution X-ray diffraction, NMR or CryoEM (where structure refinement does not yield water positions), we were able to show that the hotWater method is able to recover in the "water-free" structures many water binding sites known from the high-resolution structures. A blind test on a newly solved protein structure with waters removed from the PDB also showed good prediction of the crystal water positions. This was compared to two known algorithms that use electron density and was shown to have higher recall at resolutions >2.6 [A]. We also show that the algorithm can be applied to novel proteins such as the RNA polymerase complex from SARS-CoV-2, which could be of use in drug discovery. The hotWater model is freely available at (https://pypi.org/project/hotWater/).

bioinformatics↗