Biology subjects

Rauscher, S.

Publications and source records attributed to Rauscher, S..

12 recordsLinked to original sources

bioRxiv2026

Generating Structural Ensembles of Disordered Proteins with Diffusion Models

Generative models of protein structures have accelerated scientific discovery. However, the leading structure-prediction models are trained primarily on crystal structures, limiting their ability to sample the conformations of intrinsically disordered proteins (IDPs). All-atom molecular dynamics (MD) simulations can probe the diversity of IDP conformational ensembles, but these simulations are slow and costly. Leveraging the speed of generative models and the accuracy of MD, we develop Zephyr, an SE(3) equivariant diffusion model trained on simulated IDP ensembles. The model generates the structure of a single protein sequence in two phases: a coarse-grained structure is first sampled and heavy atoms are subsequently denoised onto this structure. We evaluate our method on two fronts: ensemble accuracy and sampling efficiency. Across five test systems, the model reproduces key geometric and energetic metrics of the MD ensembles, while sampling at least 63 times faster.

biophysics↗

bioRxiv2025

Hierarchically ordered multi-timescale structural dynamics of the intrinsically disordered p53 transactivation domain

Intrinsically disordered proteins (IDPs) exhibit pronounced structural dynamics, which is crucial for their functional versatility. Yet their dynamics slower than nanoseconds remain largely elusive. We combined high-power relaxation dispersion nuclear magnetic resonance spectroscopy with molecular dynamics simulations to characterize these kinetics and the underlying structural interconversions of a prototypical IDP, the N-terminal transactivation domain of the tumor suppressor p53 (p53-TAD). We find a complex hierarchy of structural dynamics on timescales covering over seven orders of magnitude, ranging from fast nanoseconds backbone re-orientations, via sub-microsecond helix-formation dynamics involving many structural sub-states and transition times, to transient tertiary structure formation slower than 25 microseconds. These rich structural dynamics of p53-TAD, and likely those of other IDPs, parallel the timescale hierarchy of the conformational dynamics of folded proteins. One-Sentence summaryA hierarchical energy landscape governs kinetics and structural dynamics of the disordered p53 transactivation domain.

biophysics↗

bioRxiv2024

An all-atom view into the disordered interaction interface of the TRIM5α PRYSPRY domain and the HIV capsid

Tripartite motif-containing protein 5, alpha isoform (TRIM5) is an innate immune factor that provides rhesus macaques with immunity to HIV. Despite high sequence similarity to rhesus TRIM5, human TRIM5 weakly restricts HIV without the introduction of mutations. The structural underpinnings of this functional difference are poorly understood because the interaction interface between TRIM5 and its target, the HIV capsid, involves intrinsically disordered regions. Here, we use all-atom molecular dynamics simulations to study several TRIM5 variants: rhesus TRIM5, human TRIM5, and human TRIM5 with an R332P mutation (a mutation known to enhance HIV restriction). Our data reveal differences in the conformational ensembles of wild-type and R332P human TRIM5, including a significant increase in the formation of a turn that includes the mutation site residue. We also carried out simulations of rhesus TRIM5 in complex with the HIV capsid protein. Our results indicate that the variable loops of TRIM5 are highly flexible, both in solution and in the complex with the capsid, indicative of a fuzzy interaction interface. Simulations of the complex, as well as experimental infections, indicate the basis for weak HIV restriction by human TRIM5 that is improved by an R332P mutation: replacing the positively charged arginine with a neutral residue decreases electrostatic repulsion with residue R82 of the capsid protein, which can be phenocopied by neutralizing mutations at capsid residue R82. Overall, our simulations provide a first view of the atomistic details of the HIV capsid-TRIM5 binding interface and a molecular mechanism for the observed functional differences between variants of the TRIM5 protein. Significance StatementHumans and viruses are engaged in an evolutionary arms race, which involves evolutionary cycles in which viruses escape host defense and hosts evolve to chase after them. The protein TRIM5 is a crucial part of the immune systems defense against viruses, and the region of TRIM5 that interacts with viral capsids is evolving rapidly. While a wealth of structural information is separately available for the HIV capsid protein and TRIM5, the interaction interface between these two proteins remains poorly understood because it involves multiple disordered regions. We present a first atomistic view of the TRIM5-capsid interaction interface using all-atom molecular dynamics simulations. We use our model of the interface and the probabilistic description of interactions to explain the improved HIV restriction of a mutation in the human TRIM5 protein, which has been proposed for human gene therapy, in terms of specific electrostatic interactions. Finally, we validate our findings with experimental infections. Our data provide detailed structural insight into how TRIM5 evolves improved restriction against its retroviral targets.

biophysics↗

bioRxiv2024

The Conformational Space of the SARS-CoV-2 Main Protease Active Site Loops is Determined by Ligand Binding and Interprotomer Allostery

The main protease (Mpro) of SARS-CoV-2 is essential for viral replication and is, therefore, an important drug target. Here, we investigate two flexible loops in Mpro that play a role in catalysis. Using all-atom molecular dynamics simulations, we analyze the structural ensemble of Mpro in an apo state and substrate-bound state. We find that the flexible loops can adopt open, intermediate (partly open) and closed conformations in solution, which differs from the partially closed state observed in crystal structures of Mpro. When the loops are in closed or intermediate states, the catalytic residues are more likely to be in close proximity, which is crucial for catalysis. Additionally, we find that substrate binding to one protomer of the homodimer increases the frequency of intermediate states in the bound protomer, while also affecting the structural propensity of the apo protomers flexible loops. Using dynamic network analysis, we identify multiple allosteric pathways connecting the two active sites of the homodimer. Common to these pathways is an allosteric hotspot involving the N-terminus, a critical region that comprises part of the binding pocket. Taken together, the results of our simulation study provide detailed insight into the relationships between the flexible loops and substrate binding in a prime drug target for COVID-19.

biophysics↗

bioRxiv2024

The structural influence of the oncogenic driver mutation N642H in the STAT5B SH2 domain

The point mutation N642H of the signal transducer and activator of transcription 5B (STAT5B) protein is associated with aggressive and drug-resistant forms of leukemia. This mutation is thought to promote cancer due to hyperactivation of STAT5B caused by increased stability of the active, parallel dimer state. However, the molecular mechanism leading to this stabilization is not well understood as there is currently no structure of the parallel dimer. To investigate the mutations mechanism of action, we conducted extensive all-atom molecular dynamics simula-tions of multiple oligomeric forms of both STAT5B and STAT5BN642H, including a model for the parallel dimer. The N642H mutation directly affects the hydrogen bonding network within the phosphotyrosine (pY)-binding pocket of the parallel dimer, enhancing the pY-binding in-teraction. The simulations indicate that apo STAT5B is highly flexible, exploring a diverse conformational space. In contrast, apo STAT5BN642H accesses two distinct conformational states, one of which resembles the conformation of the parallel dimer. The simulation predic-tions of the effects of the mutation on structure and dynamics are supported by the results of hydrogen-deuterium exchange (HDX) mass spectrometry measurements carried out on STAT5B and STAT5BN642H in which a phosphopeptide was used to mimic the effects of parallel dimer-ization on the SH2 domain. The molecular-level information uncovered in this work contributes to our understanding of STAT5B hyperactivation by the N642H mutation and could help pave the way for novel therapeutic strategies targeting this mutation.

biophysics↗

bioRxiv2024

PEG-mCherry interactions beyond classical macromolecular crowding

The dense cellular environment influences bio-macromolecular structure, dynamics, interactions and function. Despite advancements in understanding protein-crowder interactions, predicting their precise effects on protein structure and function remains challenging. Here, we elucidate the effects of PEG-induced crowding on the fluorescent protein mCherry using molecular dynamics simulations and fluorescence-based experiments. We identify and characterize specific PEG-induced structural and dynamical changes in mCherry. Importantly, we find interactions in which PEG molecules wrap around specific surface-exposed residues in a binding mode previously observed in protein crystal structures. Fluorescence correlation spectroscopy experiments capture PEG-induced changes, including aggregation, suggesting a potential role for the specific PEG-mCherry interactions identified in simulations. Additionally, mCherry fluorescence lifetimes are influenced by PEG and not by the bulkier crowder dextran or by another linear polymer, polyvinyl alcohol, highlighting the importance of crowder-protein soft interactions. This work augments our understanding of macromolecular crowding effects on protein structure and dynamics.

biophysics↗

bioRxiv2023

Functional Protein Dynamics in a Crystal

Proteins are molecular machines and to understand how they work, we need to understand how they move. New pump-probe time-resolved X-ray diffraction methods open up ways to initiate and observe protein motions with atomistic detail in crystals on biologically relevant timescales. However, practical limitations of these experiments demands parallel development of effective molecular dynamics approaches to accelerate progress and extract meaning. Here, we establish robust and accurate methods for simulating dynamics in protein crystals, a nontrivial process requiring careful attention to equilibration, environmental composition, and choice of force fields. With more than seven milliseconds of sampling of a single chain, we identify critical factors controlling agreement between simulation and experiments and show that simulated motions recapitulate ligand-induced conformational changes. This work enables a virtuous cycle between simulation and experiments for visualizing and understanding the basic functional motions of proteins.

biophysics↗

bioRxiv2023

Fluorescent protein lifetimes report increased local densities and phases of nuclear condensates during embryonic stem cell differentiation

Fluorescent proteins (FP) are frequently used for studying proteins inside cells. In advanced fluorescence microscopy, FPs can report on additional intracellular variables. One variable is the local density near FPs, which can be useful in studying densities within cellular bio-condensates. Here, we show that a reduction in fluorescence lifetimes of common monomeric FPs reports increased levels of local densities. We demonstrate the use of this fluorescence-based variable to report the distribution of local densities within heterochromatin protein 1 (HP1) in mouse embryonic stem cells (ESCs), before and after early differentiation. We find that local densities within HP1 condensates in pluripotent ESCs are heterogeneous and cannot be explained by a single liquid phase. Early differentiation, however, induces a change towards a more homogeneous distribution of local densities, which can be explained as a liquid-like phase. In conclusion, we provide a fluorescence-based method to report increased local densities and apply it to distinguish between homogeneous and heterogeneous local densities within bio-condensates.

biophysics↗

bioRxiv2022

LAWS: Local Alignment for Water Sites -- tracking ordered water in simulations

Protein solvation plays a crucial role in protein function; accurate modeling of protein-water interactions is important for understanding the molecular basis of protein function. In addition to protein structure, x-ray crystallography provides positions of crystallographic water sites (CWS) coordinated by the protein, which can serve as a strong benchmark for modeling accuracy. Such comparison requires special method-ological considerations that take into account the dynamic nature of proteins. However, existing methods for analyzing CWS in MD simulations rely on global alignment of the protein onto the crystal structure, which introduces substantial errors in the case of significant structural deviations. Here, we propose a method called local alignment for water sites (LAWS), which is based on multilateration -- an algorithm widely used in GPS tracking. LAWS considers the contacts formed by CWS and protein atoms in the crystal structure and uses these interaction distances to track the CWS in a simulation. LAWS provides a framework to quantify how perturbations of the local protein environment affect the preservation of crystallographic waters in simulations. We applied our method to a 1 {micro}s simulation of a protein crystal and demonstrate that 76% of CWS are preserved. Compared to existing methods, LAWS identifies more high-confidence (low B-factor) CWS, which are characterized by more prominent water density peaks and a less-perturbed protein environment.

biophysics↗

bioRxiv2022

Mesenchymal stem cells support human vascular endothelial cells to form vascular sprouts in human platelet lysate-based matrices.

During tissue regeneration, mesenchymal stem cells can support endothelial cells in the process of new vessel formation. For a functional interaction of endothelial cells with mesenchymal stem cells a vascular inductive microenvironment is required. Using a cellular model for neo-vessel formation, we could show that newly formed vascular structures emanated from the embedded aggregates, consisting of mesenchymal stem cells co-cultured with human umbilical vein endothelial cells, into the avascular HPL-based matrices bridging distances up to 5 mm to join with adjacent aggregates with the same morphology forming an interconnected network. These newly formed vascular sprouts showed branch points and generated a lumen as sign of mature vascular development. Mesenchymal stem cells in fluid phase could bind to and interact with adherent human umbilical vein endothelial cells when a shear-force of 2 dyne/mm2 was applied in the flow chamber using a Bioflux(R)200 device. Under these conditions, mesenchymal stem cells bind to human umbilical vein endothelial cells previously damaged by laser beam irradiation to interact and shed particles. In conclusion, we observed that mesenchymal stem cells support human umbilical vein endothelial cells to form new vessels in HPL-based matrices when co-cultured in spherical aggregates to generate complex vascular networks in a primarily avascular scaffold and bind to the damaged cells under shear-force to eventually aid in their regeneration.

bioengineering↗

bioRxiv2020

Distance-based metrics for comparing conformational ensembles of intrinsically disordered proteins

Intrinsically disordered proteins (IDPs) are proteins whose native functional states represent ensembles of highly diverse conformations. Such ensembles are a challenge for quantitative structure comparisons as their conformational diversity precludes optimal superimposition of the atomic coordinates, necessary for deriving common similarity measures such as the root-mean-square deviation (RMSD) of these coordinates. Here we introduce superimposition-free metrics, which are based on computing matrices of C-C distance distributions within ensembles and comparing these matrices between ensembles. Differences between two matrices yield information on the similarity between specific regions of the polypeptide, whereas the global structural similarity is captured by the ens_dRMS, defined as the root-mean-square difference between the medians of the C-Cdistance distributions of two ensembles. Together, our metrics enable rigorous investigations of structure-function relationships in conformational ensembles of IDPs derived using experimental restraints or by molecular simulations, and for proteins containing both structured and disordered regions. Statement of SignificanceImportant biological insight is obtained from comparing the high-resolution structures of proteins. Such comparisons commonly involve superimposing two protein structures and computing the residual root-mean-square deviation of the atomic positions. This approach cannot be applied to intrinsically disordered proteins (IDPs) because IDPs do not adopt well-defined 3D structures, rather, their native functional state is defined by ensembles of heterogeneous conformations that cannot be meaningfully superimposed. We report new measures that quantify the local and global similarity between different conformational ensembles by evaluating differences between the distributions of residue-residue distances and their statistical significance. Applying these measures to IDP ensembles and to a protein containing both structured and intrinsically disordered domains provides deeper insights into how structural features relate to function.

bioinformatics↗

bioRxiv2016

The Liquid Structure of Elastin

The protein elastin imparts extensibility, elastic recoil, and resilience to tissues including arterial walls, skin, lung alveoli, and the uterus. Elastin and elastin-like peptides are intrinsically disordered hydrophobic proteins that undergo liquid-liquid phase separation upon self-assembly. Despite extensive study for over eighty years, the structure of elastin remains controversial. We use molecular dynamics simulations on a massive scale to elucidate the structural ensemble of aggregated elastin-like peptides. Consistent with the entropic nature of elastic recoil, the aggregated state is stabilized both by the hydrophobic effect and by conformational entropy. The polypeptide backbone forms transient, sparse hydrogen-bonded turns and remains significantly hydrated even as self-assembly triples the extent of nonpolar side-chain contacts. The assembly approaches a maximally-disordered, melt-like state, which may be called the liquid state of proteins. These findings resolve long-standing controversies regarding elastin structure and function and afford insight of broad relevance to the phase separation of disordered proteins.

biophysics↗