Biology subjects

Gruebele, M.

Publications and source records attributed to Gruebele, M..

8 recordsLinked to original sources

bioRxiv2023

Rapid automated 3-D pose estimation of larval zebrafish using a physical model-trained neural network

Quantitative ethology requires an accurate estimation of an organisms postural dynamics in three dimensions plus time. Technological progress over the last decade has made animal pose estimation in challenging scenarios possible with unprecedented detail. Here, we present (i) a fast automated method to record and track the pose of individual larval zebrafish in a 3-D environment, applicable when accurate human labeling is not possible; (ii) a rich annotated dataset of 3-D larval poses for ethologists and the general zebrafish and machine learning community; and (iii) a technique to generate realistic, annotated larval images in novel behavioral contexts. Using a three-camera system calibrated with refraction correction, we record diverse larval swims under free swimming conditions and in response to acoustic and optical stimuli. We then employ a convolutional neural network to estimate 3-D larval poses from video images. The network is trained against a set of synthetic larval images rendered using a 3-D physical model of larvae. This 3-D model samples from a distribution of realistic larval poses that we estimate a priori using a template-based pose estimation of a small number of swim bouts. Our network model, trained without any human annotation, performs larval pose estimation with much higher speed and comparable accuracy to the template-based approach, capturing detailed kinematics of 3-D larval swims. Author SummaryLarval zebrafish swimming has been studied extensively in 2-D environments, which are restrictive compared to natural 3-D habitats. To enable rapid capture of 3-D poses, we collect three orthogonal video projections of swim behaviors in several behavioral settings and fit poses to a physical model. We then use the physical model to generate an auto-annotated stream of synthetic poses to train a convolutional neural network. The network model performs highly accurate pose predictions on over 600 real swim bouts much faster than a physical model fit. Our results show that larvae frequently exhibit motions inaccessible in a 2-D setup. The annotated dataset could be used by ethologists studying larval swimming dynamics, and by the machine learning community interested in multi-dimensional time series and 3-D reconstruction. Using the ability to render images with multiple synthetic poses, our method can be extended to collective behavior.

animal behavior and cognition↗

bioRxiv2022

Cytoplasmic folding, mis-folding, and early stages of aggregation

We examine how cellular interactions in an all-atom model of the Homo sapiens cytoplasm influence the early folding events of Protein B (PB), a three-helix bundle protein. While PB is known to fold during in vitro simulations in three microseconds, all three initially unfolded PB copies in our cytoplasm model never completely reached their native topology across our 31 microsecond simulation. We were able to capture initial formation of all three helices and a compact topology similar to the native state. Sticking interactions between PB and surrounding macromolecules, as well as other unfolded PBs, became competitive with PB folding. Interaction between PB copies seeded beta-strand formation, modeling initial events of protein aggregation. Finally, the fold-switching potential of PB related GA domains has been explored in previous studies, and the sticking and crowding in our model thus initiates sampling of helix/sheet structural plasticity of PB.

biophysics↗

bioRxiv2022

Metabolons, quinary structure, and domain motion: enzyme choreography in the cytoplasm

How do enzymes form metabolons inside cells? To answer that question, we created an all-atom model of a section of the human cytoplasm and simulated it for over 30 microseconds. Among other proteins, nucleic acids, and metabolites, the model contains three successive members of the glycolytic cycle: glyceraldehyde-3-phosphate dehydrogenase (GAPDH), phosphoglycerate kinase (PGK), and phosphoglycerate mutase (PGM). These enzymes interact to form transient, but long-lived, multi-enzyme complexes with characteristic lifetimes in the 1 to 5 s range, thus modeling the functional metabolon structures that facilitate compartmentalization of metabolic pathways and substrate channeling in cell. We analyze the quinary structure between enzymes down to the formation of specific hydrogen-bonded interactions between side chains, together with the movement, in concert, of water molecules in or out between interacting amino acids to mediate contact formation and dissolution. We also observed large-scale enzymatic domain motion that has been proposed to convert between substrate-accessible and catalytically functional states: a direct hinge-bending motion of up to 28{degrees} changes the relative orientation of the N- and C-terminal domains of PGK, causing the initially open, and presumably inactive, conformation of PGK to sample both "semi-closed" and "closed" conformations. Although classical molecular dynamics (MD) cannot simulate enzymatic activity, closed structures are the functionally active forms of PGK, and their equilibrium with open structures opens the door for future quantum mechanics/molecular mechanics (QM/MM) and other reactive simulations of the cytoplasm.

biophysics↗

bioRxiv2022

A Spatial Whole-Cell Model for HBV Viral Infection and Drug Interactions

Despite a vaccine, hepatitis B virus (HBV) remains a world-wide source of infections and deaths, and tackling the infection requires a multimodal approach against the virus. We develop a whole-cell computational platform combining spatial and kinetic models for the infection cycle of a virus host cell (hepatocyte) by HBV. We simulate a near complete viral infection cycle with this whole-cell platform stochastically for 10 minutes of biological time, to predict viral infection, map out virus-host as well as virus-drug interactions. We find that with an established infection, decreasing the copy number of the viral envelope proteins can shift the dominant infection pathways from secreting the capsids from the cell to re-importing the capsids back to the nucleus, resulting in higher viral DNA referred to as covalently closed circular DNA (cccDNA) copy number. This scenario can mimic the consequence of drugs designed to manipulate viral gene expression (such as siRNAs). Viral capsid mutants lead to their destabilization such that they disassemble at nuclear pore complexes, result in an increase in cccDNA copy number. However, excessive destabilization leading to cytoplasmic disassembly does not increase the cccDNA copy number. Finally, our simulations can predict the best drug dosage and timing of its administration to reduce the cccDNA copy number which is the hallmark of infection. Our adaptable computational platform can be utilized to study other viruses, more complex host-virus interactions, and identify the most central viral pathways that can be targeted by drugs or a combination of them.

systems biology↗

bioRxiv2021

Molecular Mechanism of Capsid Disassembly in Hepatitis B Virus

The disassembly of a viral capsid leading to the release of its genetic material into the host cell is a fundamental step in viral infection. Hepatitis B virus (HBV) in particular consists of identical capsid protein monomers that dimerize and also arrange themselves into pentamers or hexamers on the capsid surface. By applying atomistic molecular dynamics simulation to an entire solvated HBV capsid subject to a uniform mechanical stress protocol, we monitor the disassembly process down to the level of individual amino acids. The strain of an external force combined with thermal fluctuations causes structurally heterogeneous cracks to appear in the HBV capsid. Unlike the expectation from purely mechanical considerations, the cracks mainly occur within and between hexameric sites, whereas pentameric sites remain largely intact. Only a small subset of the capsid protein monomers governs disassembly. These monomers are distributed across the whole capsid, but belong to regions with a high degree of collective motion that we label communities. Cross-talk within these communities is a mechanism of crack propagation leading to destabilization of the entire capsid, and eventually its disassembly. We identify specific residues whose interactions are most readily lost during disassembly: R127, I139, Y132, N136, A137, and V149 are among the hotspots at the interfaces between dimers that lie within or span hexamers, leading to dissociation. The majority of these hotspots are evolutionary conserved, indicating that they are important for disassembly by avoiding over-stabilization of capsids. SignificanceHepatitis B virus (HBV) is a DNA virus that is 100 times more infectious than HIV. Despite the availability of a vaccine, the chronic infection rate of this virus is still about 300 million people globally. HBV chronic infection, for which no cure is currently available, can lead to liver cancer. Therefore, there is an unmet need to investigate the infection cycle of the virus. One of the most crucial steps in virus replication cycle is the release of its genetic material to the nucleus. During this step, the viral capsid enclosing the genetic material disassembles. However, its mechanism is unknown. Here, we utilize molecular simulations to shed light on the events leading to the capsid disassembly with atomistic detail.

biophysics↗

bioRxiv2018

Phase Separation of YAP Reprograms Cells for Long-term YAP Target Gene Expression

Yes-associated Protein (YAP) is a transcriptional co-activator that regulates cell proliferation and survival by binding to a selective set of enhancers for potent target gene activation, but how YAP coordinates these transcriptional responses is unknown. Here, we demonstrate that YAP forms liquid-like condensates in the nucleus in response to macromolecular crowding. Formed within seconds of hyperosmotic stress, YAP condensates compartmentalized YAPs DNA binding cofactor TEAD1 along with other YAP-related transcription co-activators, including TAZ, and subsequently induced transcription of YAP-specific proliferation genes. Super-resolution imaging using Assay for Transposase Accessible Chromatin with photoactivated localization microscopy (ATAC-PALM) revealed that YAP nuclear condensates were areas enriched in accessible chromatin domains organized as super-enhancers. Initially devoid of RNA Polymerase II (Pol II), the accessible chromatin domains later acquired Pol II, producing newly transcribed RNA. Removal of YAPs intrinsically-disordered transcription activation domain (TAD) prevented YAP condensate formation and diminished downstream YAP signaling. Thus, dynamic changes in genome organization and gene activation during YAP reprogramming is mediated by liquid-liquid phase separation.

cell biology↗

bioRxiv2018

An In-Silico Mammalian Whole-Cell Model Reveals the Influence of Spatial Organization on RNA Splicing Efficiency

Spatial organization is a characteristic of eukaryotic cells, achieved by utilizing both membrane-bound and non-bound organelles. We model the effects of this organization and of organelle heterogeneity on RNA splicing (the process of making translationally-ready messenger RNA) and on splicing particles (the building blocks of splicing machinery) in mammalian cells. We constructed a spatially-resolved whole HeLa cell model from various experimental data and developed reaction networks to describe the RNA splicing processes. We incorporated these networks into our whole-cell model and performed stochastic simulations for up to 15 minutes of biological time. We find that the number of nuclear pore complexes affects the number of assembled splicing particles; that a slight increase of splicing particle localization in nuclear speckles (non-membrane-bound or- ganelles) leads to disproportionate enhancement in the mRNA splicing and reduction in the transcript noise; and that compartmentalization is critical for a correctly-assembled particle yield. Our model also predicts that the distance between genes and speckles has a considerable effect on effective mRNA production rate, further emphasizing the importance of genome organization around speckles. The HeLa cell model, including organelles and subcompartments, provides an adaptable foundation to study other cellular processes which are strongly modulated by spatio-temporal heterogeneity.\n\nSignificance StatementThe spliceosome is one of the most complex cellular machineries that cuts and splices the RNA code in eukaryotic cells. It dynamically assembles, disassembles, and its components are formed in multiple compartments. The efficiency of splicing process depends on localization of its components in nuclear membrane-less organelles. Therefore, a computational model of spliceosomal function must contain a spatial model of the entire cell. However, building such a model is a challenging task, mainly due to the lack of homogeneous experimental data and a suitable computational framework. Here, we overcome these challenges and present a whole HeLa cell model, with nuclear, subnuclear, and extensive cytoplasmic structures. The three-dimensional model is supplemented by reaction-diffusion processes to shed light on the function of the spliceosome.

biophysics↗

bioRxiv2017

Blue light is a universal tactic signal for Escherichia coli

Blue light has been shown to elicit a tumbling response in E. coli, a non-phototrophic bacterium. The exact mechanism of this phototactic response is still unknown, and its biological significance remains unclear. Here, we quantify phototaxis in E. coli by analyzing single-cell trajectories in populations of free-swimming bacteria before and after light exposure. Bacterial strains expressing only one type of chemoreceptor reveal that all five E. coli receptors - Aer, Tar, Tsr, Tap and Trg - are capable of mediating a response to light. In particular, light exposure elicits a running response in Tap-only strain, the opposite of the tumbling response observed for all other strains. Light therefore emerges as a universal stimulus for all E. coli chemoreceptors. We also show that blue light exposure causes a reversible decrease in swimming velocity, a proxy for proton motive force. We hypothesize that rather than sensing light directly, chemoreceptors sense light-induced perturbations in proton motive force.\n\nImportanceOur findings provide new insights on the mechanism of E. coli phototaxis, showing that all five chemoreceptor types respond to light and that their interactions play an important role in cell behavior. Our results also open up new avenues for examining and manipulating E. coli taxis. Since light is a universal stimulus, it may provide a way to quantify interactions between different types of receptors. Since light is easier to control spatially and temporally than chemicals, it may be used to study swimming behavior in complex environments. Since phototaxis can cause migration of E. coli bacteria in light gradients, light may be used to control bacterial density for studying density-dependent processes in bacteria.

biophysics↗