Biology subjects

Sztain, T.

Publications and source records attributed to Sztain, T..

8 recordsLinked to original sources

bioRxiv2026

Mechanism of molecular recognition revealed through dynamic drug binding pathways to SARS-CoV-2 main protease

Characterization of drug-binding pathways remains experimentally limited by transient intermediates and computationally challenging due to long timescales intractable for conventional molecular dynamics. To address these challenges, we combined solution NMR titrations with weighted ensemble (WE) enhanced sampling simulations to resolve atomistic pathways of nirmatrelvir binding to the SARS-CoV-2 main protease. NMR titration revealed residue-dependent heterogeneity spanning fast, intermediate, and slow exchange regimes. WE simulations complement the NMR by providing insights into unassigned residues and adding time-resolved and three-dimensional structural context. We map key interactions along two distinct binding pathways, provide dynamic explanations for residues involved in resistance, and capture unique backbone conformations compared to those sampled in unbound or bound states. Our comprehensive binding model is consistent with a combined conformational selection and induced fit mechanism in which early transient contacts are made with residues E47 and L50 and allosteric motions are centered around residue V204 of the distal domain. This synergistic application of WE and titration NMR enables a more comprehensive characterization of drug binding than either method alone, providing an integrated framework that may have broader applicability to defining structure-kinetic relationships and guiding design of next-generation inhibitors.

biophysics↗

bioRxiv2026

Parameterization of the PA endonuclease bimetallic center reveals the dynamics of clinically relevant mutations

Influenza A virus continues to impose a major global health and economic burden through seasonal epidemics and occasional pandemics, highlighting the critical need for continued antiviral development. As the latest addition to anti-influenza therapy, baloxavir marboxil (BXM) targets the highly conserved PA N-terminal endonuclease domain (PAN), blocking the cap-snatching process essential for viral transcription initiation. However, the rapid emergence of resistance mutations significantly reduces BXM susceptibility and compromises its clinical efficacy. Understanding the dynamics underlying resistance through computational modeling has been hindered by the complex electronic properties of the bimetallic catalytic center within the PAN active site, posing a challenge for accurate parameterization. Therefore, in this study, we systematically benchmarked metal-parameterization strategies for molecular dynamics (MD) simulations, including non-bonded, bonded, and hybrid models, using wild-type PAN in both apo and drug-bound states. Identification of reliable parameterization schemes enabled MD simulations of five clinically relevant mutants, I38T/F/M, A36V, and E23K, revealing how each reshapes the conformational landscape to modulate drug binding modes. Together, our results provide a path toward modeling complex sites in metalloenzymes and a mechanistic foundation for vulnerabilities in PAN to guide structure-based optimization of next-generation inhibitors. TOC O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=122 SRC="FIGDIR/small/731895v1_ufig1.gif" ALT="Figure 1"> View larger version (40K): org.highwire.dtl.DTLVardef@1488315org.highwire.dtl.DTLVardef@196057eorg.highwire.dtl.DTLVardef@42a77forg.highwire.dtl.DTLVardef@2b959_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗

bioRxiv2026

Assessing the Generalizability of Machine Learning and Physics Methods for DNA-Encoded Libraries

Predicting protein-ligand binding is a central challenge in computational drug discovery, and while machine learning (ML) and co-folding methods have advanced rapidly, their ability to generalize beyond training or parameterization regimes remains insufficiently understood. DNA-encoded libraries (DELs) enable ultra-large screening of billions of molecules simultaneously, providing a useful testbed for evaluating these approaches at scale. A recent NeurIPS competition revealed that even top performing ML models trained on DEL data failed at generalizing to out-of-distribution (OOD) chemical space. We investigated whether integrating structural modeling could bridge this generalization gap. We systematically assessed state-of-the-art ML, docking, and co-folding methods including Schrodinger Glide, Rosetta GALigandDock, and Boltz-2 with three biologically diverse protein targets screened against libraries containing multiple DEL synthesis formats. While ML excels in-distribution, OOD hit discrimination is dependent on both the target and ligand context, with no single method consistently dominating. These findings demonstrate that benchmark performance alone is insufficient to predict OOD performance, highlighting the need for system-dependent evaluation of binding prediction methods. We provide an open-source package for assessing protein-ligand prediction methods and analyzing high-throughput screening data: DEL-iver. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=118 SRC="FIGDIR/small/719394v2_ufig1.gif" ALT="Figure 1"> View larger version (25K): org.highwire.dtl.DTLVardef@cd8a17org.highwire.dtl.DTLVardef@252d0aorg.highwire.dtl.DTLVardef@b020f0org.highwire.dtl.DTLVardef@1429e27_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗

bioRxiv2024

Adaptation to high pressure; insights from the genome of an evolved Escherichia coli strain with increased piezotolerance

Pressure is a key environmental parameter that influences the activity and distribution of microbial life on our planet. Despite its key role there is still not a definitive list of essential genes for microbial adaptation to life under increasing pressure. In this study we used a previously characterized Escherichia coli strain (AN62) evolved to grow at pressure (60 MPa) non-permissive to the parental strain and performed comparative genomics in order to identify the genome-level adaptations that might allowed the observed pressure-adapted phenotype. We identified 18 mutations in total of which 3 mutations were present in both the parental and evolved strain, 3 mutations were only present in the parental strain, and 12 mutations were observed only in the evolved AN62 strain. Among the characterized mutations we observed a point mutation in the acyl carrier protein (acpPV43G). Complementation experiments revealed that the observed V43G mutation in AcpP is responsible for increased levels of cis-vaccenic acid but is not alone responsible for the pressure adapted phenotype. Further molecular dynamics and docking simulations suggested that the V43G mutation promoted stronger binding of the AcpP protein to partner enzymes of the fatty acid biosynthesis pathway involved in fatty acid unsaturation. Data SummaryEscherichia coli reads from the parental and evolved strain have been deposited in the Sequence Read Archive (SRA) under accession number RJNA600359. All software used in the bioinformatic analysis is publicly available. Impact StatementPressure is a key environmental parameter. Two-thirds of our planet is covered by oceans with an average depth of 3800m, which means that the majority of the marine life experiences deep sea conditions. Our results offer a list of gene mutations that could contribute to an improved pressure growth phenotype in Escherichia coli, offering a unique insight on the genome level adaptations that might contribute to high pressure adaptation.

microbiology↗

bioRxiv2024

Machine Learning Driven Simulations of SARS-CoV-2 Fitness Landscape

Predicting protein variant effects is a key challenge in preparing for pathogenic viral strains, understanding mutation-linked diseases, and designing new proteins. Protein sequence-structure-function relationships are difficult to model due to complex allosteric and epistatic effects. To investigate efficient modeling strategies, we trained supervised machine learning (ML) models with deep mutational scanning (DMS) libraries of SARS-CoV-2 receptor binding domain (RBD) sequences labeled with angiotensin converting enzyme 2 (ACE2) binding affinity. These models demonstrate superior performance predicting combinatorial mutation effects compared to adding or averaging the effects of point mutations and exhibit strong extrapolative performance ranking omicron variants when training only on wild type (WT) variants. We characterize the RBD fitness landscape combining ML with Markov Chain Monte Carlo simulations to predict evolutionary patterns from the WT sequence, and generate comparable sequence profiles to high fitness sequences in DMS data predicting mutations in unseen omicron variants. These models provide insight into the relationship between RBD sequence elements, and offer a new perspective on the use of DMS to predict emerging viral strains, which we anticipate will be applicable to other evolutionary prediction tasks. To facilitate application and future development of this strategy, we introduce Mavenets: https://github.com/SztainLab/mavenets.

bioinformatics↗

bioRxiv2021

#COVIDisAirborne: AI-Enabled Multiscale Computational Microscopy of Delta SARS-CoV-2 in a Respiratory Aerosol

We seek to completely revise current models of airborne transmission of respiratory viruses by providing never-before-seen atomic-level views of the SARS-CoV-2 virus within a respiratory aerosol. Our work dramatically extends the capabilities of multiscale computational microscopy to address the significant gaps that exist in current experimental methods, which are limited in their ability to interrogate aerosols at the atomic/molecular level and thus ob-scure our understanding of airborne transmission. We demonstrate how our integrated data-driven platform provides a new way of exploring the composition, structure, and dynamics of aerosols and aerosolized viruses, while driving simulation method development along several important axes. We present a series of initial scientific discoveries for the SARS-CoV-2 Delta variant, noting that the full scientific impact of this work has yet to be realized. ACM Reference FormatAbigail Dommer1{dagger}, Lorenzo Casalino1{dagger}, Fiona Kearns1{dagger}, Mia Rosenfeld1, Nicholas Wauer1, Surl-Hee Ahn1, John Russo,2 Sofia Oliveira3, Clare Morris1, AnthonyBogetti4, AndaTrifan5,6, Alexander Brace5,7, TerraSztain1,8, Austin Clyde5,7, Heng Ma5, Chakra Chennubhotla4, Hyungro Lee9, Matteo Turilli9, Syma Khalid10, Teresa Tamayo-Mendoza11, Matthew Welborn11, Anders Christensen11, Daniel G. A. Smith11, Zhuoran Qiao12, Sai Krishna Sirumalla11, Michael OConnor11, Frederick Manby11, Anima Anandkumar12,13, David Hardy6, James Phillips6, Abraham Stern13, Josh Romero13, David Clark13, Mitchell Dorrell14, Tom Maiden14, Lei Huang15, John McCalpin15, Christo- pherWoods3, Alan Gray13, MattWilliams3, Bryan Barker16, HarindaRajapaksha16, Richard Pitts16, Tom Gibbs13, John Stone6, Daniel Zuckerman2*, Adrian Mulholland3*, Thomas MillerIII11,12*, ShantenuJha9*, Arvind Ramanathan5*, Lillian Chong4*, Rommie Amaro1*. 2021. #COVIDisAirborne: AI-Enabled Multiscale Computational Microscopy ofDeltaSARS-CoV-2 in a Respiratory Aerosol. In Supercomputing 21: International Conference for High Perfor-mance Computing, Networking, Storage, and Analysis. ACM, New York, NY, USA, 14 pages. https://doi.org/finalDOI

biophysics↗

bioRxiv2020

AI-Driven Multiscale Simulations Illuminate Mechanisms of SARS-CoV-2 Spike Dynamics

We develop a generalizable AI-driven workflow that leverages heterogeneous HPC resources to explore the time-dependent dynamics of molecular systems. We use this workflow to investigate the mechanisms of infectivity of the SARS-CoV-2 spike protein, the main viral infection machinery. Our workflow enables more efficient investigation of spike dynamics in a variety of complex environments, including within a complete SARS-CoV-2 viral envelope simulation, which contains 305 million atoms and shows strong scaling on ORNL Summit using NAMD. We present several novel scientific discoveries, including the elucidation of the spikes full glycan shield, the role of spike glycans in modulating the infectivity of the virus, and the characterization of the flexible interactions between the spike and the human ACE2 receptor. We also demonstrate how AI can accelerate conformational sampling across different systems and pave the way for the future application of such methods to additional studies in SARS-CoV-2 and other molecular systems. ACM Reference FormatLorenzo Casalino1{dagger}, Abigail Dommer1{dagger}, Zied Gaieb1{dagger}, Emilia P. Barros1, Terra Sztain1, Surl-Hee Ahn1, Anda Trifan2,3, Alexander Brace2, Anthony Bogetti4, Heng Ma2, Hyungro Lee5, Matteo Turilli5, Syma Khalid6, Lillian Chong4, Carlos Simmerling7, David J. Hardy3, Julio D. C. Maia3, James C. Phillips3, Thorsten Kurth8, Abraham Stern8, Lei Huang9, John McCalpin9, Mahidhar Tatineni10, Tom Gibbs8, John E. Stone3, Shantenu Jha5, Arvind Ramanathan2*, Rommie E. Amaro1*. 2020. AI-Driven Multiscale Simulations Illuminate Mechanisms of SARS-CoV-2 Spike Dynamics. In Supercomputing 20: International Conference for High Performance Computing, Networking, Storage, and Analysis. ACM, New York, NY, USA, 14 pages. https://doi.org/finalDOI

biophysics↗

bioRxiv2020

Elucidation of cryptic and allosteric pockets within the SARS-CoV-2 protease

The SARS-CoV-2 pandemic has rapidly spread across the globe, posing an urgent health concern. Many quests to computationally identify treatments against the virus rely on in silico small molecule docking to experimentally determined structures of viral proteins. One limit to these approaches is that protein dynamics are often unaccounted for, leading to overlooking transient, druggable conformational states. Using Gaussian accelerated molecular dynamics to enhance sampling of conformational space, we identified cryptic pockets within the SARS-CoV-2 main protease, including some within regions far from the active site and assed their druggability. These pockets can aid in virtual screening efforts to identify a protease inhibitor for the treatment of COVID-19. O_FIG_DISPLAY_L [Figure 1] M_FIG_DISPLAY C_FIG_DISPLAY

biophysics↗