Biology subjects

Papoian, G. A.

Publications and source records attributed to Papoian, G. A..

9 recordsLinked to original sources

bioRxiv2026

Overcoming the accuracy-generalization tradeoff in docking and scoring for prospective virtual screening

Virtual screening promises access to tens of billions of synthetically accessible, diverse compounds, yet it is rarely used as a primary hit-discovery strategy in contemporary drug-discovery campaigns. We argue that this gap reflects the real-world underperformance of the underlying docking and scoring methods: classical docking is generalizable but limited in accuracy by simple functional forms and parsimonious parameterization, whereas recent machine-learning approaches are highly expressive but do not generalize well to novel molecules and pockets, their reported accuracy often inflated by train-test leakage. To address these challenges, we introduce DODock and DOScore, docking and scoring ML/physics hybrid frameworks that are also highly expressive, yet generalize much better out of distribution compared with the prior ML approaches. This generalization has been prospectively tested in several ways. First, DODocks blind prediction of a drug candidate binding to PCSK9 was compared to the crystal structure that was subsequently solved, recovering the binding pose to 1.2 [A] RMSD. We also used DODock and DOScore in prospective virtual screening campaigns against four therapeutic targets, spanning an ectoenzyme (CD73), a kinase (IRAK4), an extended-substrate protease (FXI), and an allosteric inhibition of protein-protein interface (IL17). These screens yielded many chemically novel, biochemically and cellularly active inhibitors. In the case of CD73, which is a historically challenging target for virtual screening, our screen resulted in a roughly hundredfold improvement in hit rate over a recent machine-learning screen. Our results indicate that the apparent ceiling in the accuracy of virtual screening that seemed to have somewhat plateaued in the last two decades is not fundamental, and that structure-based interrogation of ultralarge chemical space may eventually become a credible primary route to novel chemical matter.

biophysics↗

bioRxiv2024

SMART DATA FACTORY: VOLUNTEER COMPUTING PLATFORM FOR ACTIVE LEARNING-DRIVEN MOLECULAR DATA ACQUISITION

This paper presents the Smart Distributed Data Factory (SDDF), an AI-driven distributed computing platform designed to address challenges in drug discovery by creating comprehensive datasets of molecular conformations and their properties. SDDF uses volunteer computing, leveraging the processing power of personal computers worldwide to accelerate quantum chemistry (DFT) calculations. To tackle the vast chemical space and limited high-quality data, SDDF employs an ensemble of machine learning models to predict molecular properties and selectively choose the most challenging data points for further DFT calculations. The platform also generates new molecular conformations using molecular dynamics with the forces derived from these models. SDDF makes several contributions: the volunteer computing platform for DFT calculations; an active learning framework for constructing a dataset of molecular conformations; a large public dataset of diverse ENAMINE molecules with calculated energies; an ensemble of state-of-the-art ML models for accurate energy prediction. The energy dataset was generated to validate the SDDF approach of reducing the need for extensive calculations. With its strict scaffold split, the dataset can be used for training and benchmarking energy models. By combining active learning, distributed computing, and quantum chemistry, SDDF offers a scalable, cost-effective solution for developing accurate molecular models and ultimately accelerating drug discovery.

biophysics↗

bioRxiv2022

Enabled primarily controls filopodial morphology, not actin organization, in the TSM1 growth cone in Drosophila

Ena/VASP proteins are processive actin polymerases that are required throughout animal phylogeny for many morphogenetic processes, including axon growth and guidance. Here we use live imaging of morphology and actin organization in the TSM1 axon of the Drosophila wing to dissect the mechanism of Ena action. We find that altering Ena activity has a substantial impact on filopodial morphology in this growth cone, but exerts only modest effects on actin organization. This is in contrast to the main regulator of Ena, Abl tyrosine kinase, which has profound effects on actin and only mild effects on TSM1 growth cone morphology. These data suggest that the primary role of Ena in this axon may be to link actin to morphogenetic processes of the plasma membrane, rather than regulating actin organization itself. These data also suggest that a key role of Ena, acting downstream of Abl, may be to maintain a constant filopodial organization of the growth cone, even as Abl activity varies in response to guidance cues in the environment. Summary statementWe dissect the function of the actin polymerase, Enabled, in axon growth by live-imaging of actin dynamics and axon morphology of the TSM1 neuron in its native environment in vivo.

neuroscience↗

bioRxiv2021

Computational simulations reveal that Abl activity controls cohesiveness of actin networks in growth cones

Extensive studies of growing axons have revealed many individual components and protein interactions that guide neuronal morphogenesis. Despite this, however, we lack any clear picture of the emergent mechanism by which this nanometer-scale biochemistry generates the multi-micron scale morphology and cell biology of axon growth and guidance in vivo. To address this, we studied the downstream effects of the Abl signaling pathway using a computer simulation software (MEDYAN) that accounts for mechanochemical dynamics of active polymers. Previous studies implicate two Abl effectors, Arp2/3 and Enabled, in Abl-dependent axon guidance decisions. We now find that Abl alters actin architecture primarily by activating Arp2/3, while Enabled plays a more limited role. Our simulations show that simulations mimicking modest levels of Abl activity bear striking similarity to actin profiles obtained experimentally from live-imaging of actin in wild type axons in vivo. Using a graph-theoretical filament-filament contact analysis, moreover, we find that networks mimicking hyperactivity of Abl (enhanced Arp2/3) are fragmented into smaller domains of actin that interact weakly with each other, consistent with the pattern of actin fragmentation observed upon Abl overexpression in vivo. Two perturbative simulations further confirm that high Arp2/3 actin networks are mechanically disconnected and fail to mount a cohesive response to perturbation. Taken together, these data provide a molecular-level picture of how the large-scale organization of the axonal cytoskeleton arises from the biophysics of actin networks. Highlight summaryHow do single-molecule dynamics produce multi-micron scale changes in actin organization in an extending axon? Comparison of computational simulations to in vivo data suggests that Abl kinase and Arp2/3 expand actomyosin networks by fragmenting into multiple domains, thus toggling the axon between states of local vs global internal connectivity.

cell biology↗

bioRxiv2021

Different translation dynamics of β- and γ-actin regulates cell migration

{beta}- and {gamma}-cytoplasmic actins are ubiquitously expressed in every cell type and are nearly identical at the amino acid level but play vastly different roles in vivo. Their essential roles in embryogenesis and cell migration critically depend on the nucleotide sequences of their genes, rather than their amino acid sequence. However it is unclear which gene elements underlie this effect. Here we address the specific role of the coding sequence in {beta}- and {gamma}-cytoplasmic actins intracellular functions, using stable cell lines with exogenously expressed actin isoforms and their "codon-switched" variants. When targeted to the cell periphery using the {beta}-actin 3'UTR, {beta}-actin and {gamma}-actin have differential effects on cell migration. These effects directly depend on the coding sequence. Single molecule measurements of actin isoform translation, combined with fluorescence recovery after photobleaching, demonstrate a pronounced difference in {beta}- and {gamma}-actins translation elongation rates, leading to changes in their dynamics at focal adhesions, impairments in actin bundle formation, and reduced cell anchoring to the substrate during migration. Our results demonstrate that coding sequence-mediated differences in actin translation play a key role in cell migration.

cell biology↗

bioRxiv2020

Binding Dynamics of Disordered Linker Histone H1 with a Nucleosomal Particle

Linker histone H1 is an essential regulatory protein for many critical biological processes, such as eukaryotic chromatin packaging and gene expression. Mis-regulation of H1s is commonly observed in tumor cells, where the balance between different H1 subtypes has been shown to alter the cancer phenotype. Consisting of a rigid globular domain and two highly charged terminal domains, H1 can bind to multiple sites on a nucleosomal particle to alter chromatin hierarchical condensation levels. In particular, the disordered H1 amino- and carboxyl-terminal domains (NTD/CTD) are believed to enhance this binding affinity, but their detailed dynamics and functions remain unclear. In this work, we used a coarse-grained computational model AWSEM-DNA to simulate the H1.0b-nucleosome complex, namely chromatosome. Our results demonstrate that H1 disordered domains restrict the dynamics of both globular H1 and linker DNA arms, resulting in a more compact and rigid chromatosome particle. Furthermore, we identified regions of H1 disordered domains that are tightly tethered to DNA near the entry-exit site. Overall, our study elucidates at near atomic resolution the way the disordered linker histone H1 modulates nucleosomes structural preferences and conformational dynamics.

biophysics↗

bioRxiv2020

Tensile Force Induced Cytoskeletal Reorganization: Mechanics Before Chemistry

Understanding cellular remodeling in response to mechanical stimuli is a critical step in elucidating mechano-activation of biochemical signaling pathways. Experimental evidence indicates that external stress-induced subcellular adaptation is accomplished through dynamic cytoskeletal reorganization. To study the interactions between subcellular structures involved in transducing mechanical signals, we combined experimental and computational simulations to evaluate real-time mechanical adaptation of the actin cytoskeletal network. Actin cytoskeleton was imaged at the same time as an external tensile force was applied to live vascular smooth muscle cells using a fibronectin-functionalized atomic force microscope probe. In addition, we performed computational simulations of active cytoskeletal networks under a tensile external force. The experimental data and simulation results suggest that mechanical structural adaptation occurs before chemical adaptation during filament bundle formation: actin filaments first align in the direction of the external force, initializing anisotropic filament orientations, then the chemical evolution of the network follows the anisotropic structures to further develop the bundle-like geometry. This finding presents an alternative, novel explanation for the stress fiber formation and provides new insight into the mechanism of mechanotransduction. Author SummaryRemodeling the cytoskeletal network in response to external force is key to mechanosensing and locomotion. Despite much focus on cytoskeletal remodeling in recent years, a comprehensive understanding of actin remodeling in real-time in cells under mechanical stimuli is still lacking. We integrated stress-induced 3D actin imaging and 3D computational simulations of actin cytoskeleton to study how the actin cytoskeleton form bundles and how these bundles evolve over time upon external tensile stress. We found a rapid actin alignment and a slower bundle evolution leading to denser bundles. Based on these results, we propose a "mechanics before chemistry" model of actin cytoskeleton remodeling under external force.

cell biology↗

bioRxiv2018

Elastic and Rigidified CENP-A Nucleosomes govern Centromeric Chromatin Plasticity

Histone variants fine-tune transcription, replication, DNA damage repair, and faithful chromosome segregation. Whether and how nucleosome variants encode unique mechanical properties to their cognate chromatin structures remains elusive. Here, using novel in silico and in vitro nanoindentation methods, extending to in vivo dissections, we report that histone variant nucleosomes are intrinsically more elastic than their canonical counterparts. Furthermore, binding proteins which discriminate between histone variant nucleosomes suppress this innate elasticity and also compact chromatin. Interestingly, when we overexpress the binding proteins in vivo, we also observe increased compaction of chromatin enriched for histone variant nucleosomes, correlating with diminished access. Together, these data suggest a plausible link between innate mechanical properties possessed by histone variant nucleosomes, the adaptability of chromatin states in vivo, and the epigenetic plasticity of the underlying locus.\n\nSignificanceNucleosomes are the base unit which organize eukaryotic genomes. Besides the canonical histone, histone variants create unique local chromatin domains that fine-tune transcription, replication, DNA damage repair, and faithful chromosome segregation. We developed computational and single-molecule nanoindentation tools to determine mechanical properties of histone variant nucleosomes. We found that the CENP-A nucleosome variant is more elastic than the canonical H3 nucleosome but becomes stiffer when bound to its partner CENP-C. In addition, CENP-C induces cross-array clustering, creating a chromatin state that less accessible. These data suggest that innate material properties of nucleosomes can influence the ultimate chromatin state, thereby influence biological outcomes.

biochemistry↗

bioRxiv2017

Internal modifications in the CENP-A nucleosome modulate centromeric dynamics

Post-translational modifications (PTMs) of core histones have studied for over 2 decades, and are correlated with changes in transcriptional status, chromatin fiber folding, and nucleosome dynamics. However, within the centromere-specific histone H3 variant CENP-A, few modifications have been reported, and their functions remain largely unexplored. In this multidisciplinary report, we utilize in silico computational and in vivo approaches to dissect lysine 124 of human CENP-A, which was previously reported to be acetylated in advance of replication. Computational modeling demonstrates that acetylation of K124 causes tightening of the histone core, and hinders accessibility to its C-terminus, which in turn diminishes CENP-C binding. Additionally, CENP-A K124ac/H4 K79ac containing nucleosomes are prone to DNA sliding. In vivo experiments using an acetyl or unacetylatable mimic (CENP-A K124Q and K124A respectively) reveal alterations in CENP-C levels, and a modest increase in mitotic errors. Furthermore, mutation of K124 results in alterations in centromeric replication timing, with the permanently acetylated form replicating centromeres early, and the unacetylable form replicating centromeres late. Purification of native CENP-A proteins followed by mass spectrometry analysis reveal that while CENP-A K124 is acetylated at G1/S, it switches to monomethylation during early and mid-S phase. Finally, we provide evidence that the HAT p300 is involved in this cycle. Taken together, our data suggest that cyclical modifications within the CENP-A nucleosome can influence the binding of key kinetochore proteins, the integrity of mitosis and centromeric replication. These data support the emerging paradigm that core modifications in histone variant nucleosomes transduce defined changes to key biological processes.

biochemistry↗