R-loops are nucleic acid hybrids which form when an RNA invades duplex DNA to pair with its template sequence. Although they are implicated in a growing number of gene regulatory processes, their mechanistic origins remain unclear. We here report real-time observations of cotranscriptional R-loop formation at single-molecule resolution and propose a mechanism for their formation. We show that the bacterial Mfd protein can simultaneously interact with both elongating RNA polymerase and upstream DNA, tethering the two together and partitioning the DNA into distinct supercoiled domains. A highly negatively supercoiled domain forms in between Mfd and RNA polymerase, and compensatory positive supercoiling appears in front of the RNA polymerase and behind Mfd. The nascent RNA invades the negatively supercoiled domain and forms a stable R-loop that can drive mutagenesis. This mechanism theoretically enables any protein that simultaneously binds an actively translocating RNA polymerase and upstream DNA to stimulate R-loop formation.

Bundles of stiff filaments are ubiquitous in the living world, found both in the cytoskeleton and in the extracellular medium. These bundles are typically held together by smaller cross-linking molecules. We demonstrate, analytically, numerically, and experimentally, that such bundles can be kinked, that is, have localized regions of high curvature that are long-lived metastable states. We propose three possible mechanisms of kink stabilization: a difference in trapped length of the filament segments between two cross-links, a dislocation where the endpoint of a filament occurs within the bundle, and the braiding of the filaments in the bundle. At a high concentration of cross-links, the last two effects lead to the topologically protected kinked states. Finally, we explore, numerically and analytically, the transition of the metastable kinked state to the stable straight bundle.

In the field of artificial intelligence, a combination of scale in data and model capacity enabled by unsupervised learning has led to major advances in representation learning and statistical generation. In the life sciences, the anticipated growth of sequencing promises unprecedented data on natural sequence diversity. Protein language modeling at the scale of evolution is a logical step toward predictive and generative artificial intelligence for biology. To this end, we use unsupervised learning to train a deep contextual language model on 86 billion amino acids across 250 million protein sequences spanning evolutionary diversity. The resulting model contains information about biological properties in its representations. The representations are learned from sequence data alone. The learned representation space has a multiscale organization reflecting structure from the level of biochemical properties of amino acids to remote homology of proteins. Information about secondary and tertiary structure is encoded in the representations and can be identified by linear projections. Representation learning produces features that generalize across a range of applications, enabling state-of-the-art supervised prediction of mutational effect and secondary structure and improving state-of-the-art features for long-range contact prediction.

Intracellular phase separation of proteins into biomolecular condensates is increasingly recognized as a process with a key role in cellular compartmentalization and regulation. Different hypotheses about the parameters that determine the tendency of proteins to form condensates have been proposed, with some of them probed experimentally through the use of constructs generated by sequence alterations. To broaden the scope of these observations, we established an in silico strategy for understanding on a global level the associations between protein sequence and phase behavior and further constructed machine-learning models for predicting protein liquid–liquid phase separation (LLPS). Our analysis highlighted that LLPS-prone proteins are more disordered, less hydrophobic, and of lower Shannon entropy than sequences in the Protein Data Bank or the Swiss-Prot database and that they show a fine balance in their relative content of polar and hydrophobic residues. To further learn in a hypothesis-free manner the sequence features underpinning LLPS, we trained a neural network-based language model and found that a classifier constructed on such embeddings learned the underlying principles of phase behavior at a comparable accuracy to a classifier that used knowledge-based features. By combining knowledge-based features with unsupervised embeddings, we generated an integrated model that distinguished LLPS-prone sequences both from structured proteins and from unstructured proteins with a lower LLPS propensity and further identified such sequences from the human proteome at a high accuracy. These results provide a platform rooted in molecular principles for understanding protein phase behavior. The predictor, termed DeePhase, is accessible from https://deephase.ch.cam.ac.uk/.

Subclasses of lymphocytes carry different functional roles to work together and produce an immune response and lasting immunity. Additionally to these functional roles, T and B cell lymphocytes rely on the diversity of their receptor chains to recognize different pathogens. The lymphocyte subclasses emerge from common ancestors generated with the same diversity of receptors during selection processes. Here, we leverage biophysical models of receptor generation with machine learning models of selection to identify specific sequence features characteristic of functional lymphocyte repertoires and subrepertoires. Specifically, using only repertoire-level sequence information, we classify CD4⁺ and CD8⁺ T cells, find correlations between receptor chains arising during selection, and identify T cell subsets that are targets of pathogenic epitopes. We also show examples of when simple linear classifiers do as well as more complex machine learning methods.

Understanding the mechanics of blood flow is necessary for developing insights into mechanisms of physiology and vascular diseases in microcirculation. Given the limitations of technologies available for assessing in vivo flow fields, in vitro methods based on traditional microfluidic platforms have been developed to mimic physiological conditions. However, existing methods lack the capability to provide accurate assessment of these flow fields, particularly in vessels with complex geometries. Conventional approaches to quantify flow fields rely either on analyzing only visual images or on enforcing underlying physics without considering visualization data, which could compromise accuracy of predictions. Here, we present artificial-intelligence velocimetry (AIV) to quantify velocity and stress fields of blood flow by integrating the imaging data with underlying physics using physics-informed neural networks. We demonstrate the capability of AIV by quantifying hemodynamics in microchannels designed to mimic saccular-shaped microaneurysms (microaneurysm-on-a-chip, or MAOAC), which signify common manifestations of diabetic retinopathy, a leading cause of vision loss from blood-vessel damage in the retina in diabetic patients. We show that AIV can, without any a priori knowledge of the inlet and outlet boundary conditions, infer the two-dimensional (2D) flow fields from a sequence of 2D images of blood flow in MAOAC, but also can infer three-dimensional (3D) flow fields using only 2D images, thanks to the encoded physics laws. AIV provides a unique paradigm that seamlessly integrates images, experimental data, and underlying physics using neural networks to automatically analyze experimental data and infer key hemodynamic indicators that assess vascular injury.

Chloride ion–pumping rhodopsin (ClR) in some marine bacteria utilizes light energy to actively transport Cl⁻ into cells. How the ClR initiates the transport is elusive. Here, we show the dynamics of ion transport observed with time-resolved serial femtosecond (fs) crystallography using the Linac Coherent Light Source. X-ray pulses captured structural changes in ClR upon flash illumination with a 550 nm fs-pumping laser. High-resolution structures for five time points (dark to 100 ps after flashing) reveal complex and coordinated dynamics comprising retinal isomerization, water molecule rearrangement, and conformational changes of various residues. Combining data from time-resolved spectroscopy experiments and molecular dynamics simulations, this study reveals that the chloride ion close to the Schiff base undergoes a dissociation–diffusion process upon light-triggered retinal isomerization.

Asymmetric cell division generates two daughter cells with distinct characteristics and fates. Positioning different regulatory and signaling proteins at the opposing ends of the predivisional cell produces molecularly distinct daughter cells. Here, we report a strategy deployed by the asymmetrically dividing bacterium Caulobacter crescentus where a regulatory protein is programmed to perform distinct functions at the opposing cell poles. We find that the CtrA proteolysis adaptor protein PopA assumes distinct oligomeric states at the two cell poles through asymmetrically distributed c-di-GMP: dimeric at the stalked pole and monomeric at the swarmer pole. Different polar organizing proteins at each cell pole recruit PopA where it interacts with and mediates the function of two molecular machines: the ClpXP degradation machinery at the stalked pole and the flagellar basal body at the swarmer pole. We discovered a binding partner of PopA at the swarmer cell pole that together with PopA regulates the length of the flagella filament. Our work demonstrates how a second messenger provides spatiotemporal cues to change the physical behavior of an effector protein, thereby facilitating asymmetry.

Collective motion of active matter is ubiquitously observed, ranging from propelled colloids to flocks of bird, and often features the formation of complex structures composed of agents moving coherently. However, it remains extremely challenging to predict emergent patterns from the binary interaction between agents, especially as only a limited number of interaction regimes have been experimentally observed so far. Here, we introduce an actin gliding assay coupled to a supported lipid bilayer, whose fluidity forces the interaction between self-propelled filaments to be dominated by steric repulsion. This results in filaments stopping upon binary collisions and eventually aligning nematically. Such a binary interaction rule results at high densities in the emergence of dynamic collectively moving structures including clusters, vortices, and streams of filaments. Despite the microscopic interaction having a nematic symmetry, the emergent structures are found to be polar, with filaments collectively moving in the same direction. This is due to polar biases introduced by the stopping upon collision, both on the individual filaments scale as well as on the scale of collective structures. In this context, positive half-charged topological defects turn out to be a most efficient trapping and polarity sorting conformation.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a beta-CoV that recently emerged as a human pathogen and is the causative agent of the COVID-19 pandemic. A molecular framework of how the virus manipulates host cellular machinery to facilitate infection remains unclear. Here, we focus on SARS-CoV-2 NSP1, which is proposed to be a virulence factor that inhibits protein synthesis by directly binding the human ribosome. We demonstrate biochemically that NSP1 inhibits translation of model human and SARS-CoV-2 messenger RNAs (mRNAs). NSP1 specifically binds to the small (40S) ribosomal subunit, which is required for translation inhibition. Using single-molecule fluorescence assays to monitor NSP1–40S subunit binding in real time, we determine that eukaryotic translation initiation factors (eIFs) allosterically modulate the interaction of NSP1 with ribosomal preinitiation complexes in the absence of mRNA. We further elucidate that NSP1 competes with RNA segments downstream of the start codon to bind the 40S subunit and that the protein is unable to associate rapidly with 80S ribosomes assembled on an mRNA. Collectively, our findings support a model where NSP1 proteins from viruses in at least two subgenera of beta-CoVs associate with the open head conformation of the 40S subunit to inhibit an early step of translation, by preventing accommodation of mRNA within the entry channel.

There has been much success recently in theoretically simulating parts of complex biological systems on the molecular level, with the goal of first-principles modeling of whole cells. However, there is the question of whether such simulations can be performed because of the enormous complexity of cells. We establish approximate equations to estimate computation times required to simulate highly simplified models of cells by either molecular dynamics calculations or by solving molecular kinetic equations. Our equations place limits on the complexity of cells that can be theoretically understood with these two methods and provide a first step in developing what can be considered biological uncertainty relations for molecular models of cells. While a molecular kinetics description of the genetically simplest bacterial cell may indeed soon be possible, neither theoretical description for a multicellular system, such as the human brain, will be possible for many decades and may never be possible even with quantum computing.

The OLE (ornate, large, and extremophilic) RNA class is one of the most complex and well-conserved bacterial noncoding RNAs known to exist. This RNA is known to be important for bacterial responses to stress caused by short-chain alcohols, cold, and elevated Mg²⁺ concentrations. These biological functions have been shown to require the formation of a ribonucleoprotein (RNP) complex including at least two protein partners: OLE-associated protein A (OapA) and OLE-associated protein B (OapB). OapB directly binds OLE RNA with high-affinity and specificity and is believed to assist in assembling the functional OLE RNP complex. To provide the atomic details of OapB–OLE RNA interaction and to potentially reveal previously uncharacterized protein–RNA interfaces, we determined the structure of OapB from Bacillus halodurans alone and in complex with an OLE RNA fragment at resolutions of 1.0 Å and 2.0 Å, respectively. The structure of OapB exhibits a K-shaped overall architecture wherein its conserved KOW motif and additional unique structural elements of OapB form a bipartite RNA-binding surface that docks to the P13 hairpin and P12.2 helix of OLE RNA. These high-resolution structures elucidate the molecular contacts used by OapB to form a stable RNP complex and explain the high conservation of sequences and structural features at the OapB–OLE RNA-binding interface. These findings provide insight into the role of OapB in the assembly and biological function of OLE RNP complex and can guide the exploration of additional possible OLE RNA-binding interactions present in OapB.

Control over cell growth by mobile regulators underlies much of eukaryotic morphogenesis. In plant roots, cell division and elongation are separated into distinct longitudinal zones and both division and elongation are influenced by the growth regulatory hormone gibberellin (GA). Previously, a multicellular mathematical model predicted a GA maximum at the border of the meristematic and elongation zones. However, GA in roots was recently measured using a genetically encoded fluorescent biosensor, nlsGPS1, and found to be low in the meristematic zone grading to a maximum at the end of the elongation zone. Furthermore, the accumulation rate of exogenous GA was also found to be higher in the elongation zone. It was still unknown which biochemical activities were responsible for these mobile small molecule gradients and whether the spatiotemporal correlation between GA levels and cell length is important for root cell division and elongation patterns. Using a mathematical modeling approach in combination with high-resolution GA measurements in vivo, we now show how differentials in several biosynthetic enzyme steps contribute to the endogenous GA gradient and how differential cellular permeability contributes to an accumulation gradient of exogenous GA. We also analyzed the effects of altered GA distribution in roots and did not find significant phenotypes resulting from increased GA levels or signaling. We did find a substantial temporal delay between complementation of GA distribution and cell division and elongation phenotypes in a GA deficient mutant. Together, our results provide models of how GA gradients are directed and in turn direct root growth.

Fast excitatory synaptic transmission in the central nervous system relies on the AMPA-type glutamate receptor (AMPAR). This receptor incorporates a nonselective cation channel, which is opened by the binding of glutamate. Although the open pore structure has recently became available from cryo-electron microscopy (Cryo-EM), the molecular mechanisms governing cation permeability in AMPA receptors are not understood. Here, we combined microsecond molecular dynamic (MD) simulations on a putative open-state structure of GluA2 with electrophysiology on cloned channels to elucidate ion permeation mechanisms. Na⁺, K⁺, and Cs⁺ permeated at physiological rates, consistent with a structure that represents a true open state. A single major ion binding site for Na⁺ and K⁺ in the pore represents the simplest selectivity filter (SF) structure for any tetrameric cation channel of known structure. The minimal SF comprised only Q586 and Q587, and other residues on the cytoplasmic side formed a water-filled cavity with a cone shape that lacked major interactions with ions. We observed that Cl⁻ readily enters the upper pore, explaining anion permeation in the RNA-edited (Q586R) form of GluA2. A permissive architecture of the SF accommodated different alkali metals in distinct solvation states to allow rapid, nonselective cation permeation and copermeation by water. Simulations suggested Cs⁺ uses two equally populated ion binding sites in the filter, and we confirmed with electrophysiology of GluA2 that Cs⁺ is slightly more permeant than Na⁺, consistent with serial binding sites preferentially driving selectivity.

Diauxie, or the sequential consumption of carbohydrates in bacteria such as Escherichia coli, has been hypothesized to be an evolutionary strategy which allows the organism to maximize its instantaneous specific growth—giving the bacterium a competitive advantage. Currently, the computational techniques used in industrial biotechnology fall short of explaining the intracellular dynamics underlying diauxic behavior. In particular, the understanding of the proteome dynamics in diauxie can be improved. We developed a robust iterative dynamic method based on expression- and thermodynamically enabled flux models to simulate the kinetic evolution of carbohydrate consumption and cellular growth. With minimal modeling assumptions, we couple kinetic uptakes, gene expression, and metabolic networks, at the genome scale, to produce dynamic simulations of cell cultures. The method successfully predicts the preferential uptake of glucose over lactose in E. coli cultures grown on a mixture of carbohydrates, a manifestation of diauxie. The simulated cellular states also show the reprogramming in the content of the proteome in response to fluctuations in the availability of carbon sources, and it captures the associated time lag during the diauxie phenotype. Our models suggest that the diauxic behavior of cells is the result of the evolutionary objective of maximization of the specific growth of the cell. We propose that genetic regulatory networks, such as the lac operon in E. coli, are the biological implementation of a robust control system to ensure optimal growth.

Artificial mechanical perturbations affect chromatin in animal cells in culture. Whether this is also relevant to growing tissues in living organisms remains debated. In plants, aerial organ emergence occurs through localized outgrowth at the periphery of the shoot apical meristem, which also contains a stem cell niche. Interestingly, organ outgrowth has been proposed to generate compression in the saddle-shaped organ–meristem boundary domain. Yet whether such growth-induced mechanical stress affects chromatin in plant tissues is unknown. Here, by imaging the nuclear envelope in vivo over time and quantifying nucleus deformation, we demonstrate the presence of active nuclear compression in that domain. We developed a quantitative pipeline amenable to identifying a subset of very deformed nuclei deep in the boundary and in which nuclei become gradually narrower and more elongated as the cell contracts transversely. In this domain, we find that the number of chromocenters is reduced, as shown by chromatin staining and labeling, and that the expression of linker histone H1.3 is induced. As further evidence of the role of forces on chromatin changes, artificial compression with a MicroVice could induce the ectopic expression of H1.3 in the rest of the meristem. Furthermore, while the methylation status of chromatin was correlated with nucleus deformation at the meristem boundary, such correlation was lost in the h1.3 mutant. Altogether, we reveal that organogenesis in plants generates compression that is able to have global effects on chromatin in individual cells.

The migratory dynamics of cells in physiological processes, ranging from wound healing to cancer metastasis, rely on contact-mediated cell–cell interactions. These interactions play a key role in shaping the stochastic trajectories of migrating cells. While data-driven physical formalisms for the stochastic migration dynamics of single cells have been developed, such a framework for the behavioral dynamics of interacting cells still remains elusive. Here, we monitor stochastic cell trajectories in a minimal experimental cell collider: a dumbbell-shaped micropattern on which pairs of cells perform repeated cellular collisions. We observe different characteristic behaviors, including cells reversing, following, and sliding past each other upon collision. Capitalizing on this large experimental dataset of coupled cell trajectories, we infer an interacting stochastic equation of motion that accurately predicts the observed interaction behaviors. Our approach reveals that interacting noncancerous MCF10A cells can be described by repulsion and friction interactions. In contrast, cancerous MDA-MB-231 cells exhibit attraction and antifriction interactions, promoting the predominant relative sliding behavior observed for these cells. Based on these experimentally inferred interactions, we show how this framework may generalize to provide a unifying theoretical description of the diverse cellular interaction behaviors of distinct cell types.

Turbulence is an important determinant of phytoplankton physiology, often leading to cell stress and damage. Turbulence affects phytoplankton migration both by transporting cells and by triggering switches in migratory behavior, whereby vertically migrating cells can actively invert their direction of migration upon exposure to turbulent cues. However, a mechanistic link between single-cell physiology and vertical migration of phytoplankton in turbulence is currently missing. Here, by combining physiological and behavioral experiments with a mathematical model of stress accumulation and dissipation, we show that the mechanism responsible for the switch in the direction of migration in the marine raphidophyte Heterosigma akashiwo is the integration of reactive oxygen species (ROS) signaling generated by turbulent cues. Within timescales as short as tens of seconds, the emergent downward-migrating subpopulation exhibited a twofold increase in ROS, an indicator of stress, 15% lower photosynthetic efficiency, and 35% lower growth rate over multiple generations compared to the upward-migrating subpopulation. The origin of the behavioral split as a result of a bistable oxidative stress response is corroborated by the observation that exposure of cells to exogenous stressors (H₂O₂, UV-A radiation, or high irradiance), in lieu of turbulence, caused comparable ROS accumulation and an equivalent split into the two subpopulations. By providing a mechanistic link between the single-cell mechanics of swimming and physiology on the one side and the emergent population-scale migratory response and impact on fitness on the other, the ROS-mediated early warning response we discovered contributes to our understanding of phytoplankton community composition in future ocean conditions.

Immune response to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection during the current pandemic remains a field of immense interest and active research worldwide. Although the severity of acute infection may depend on the intensity of innate and adaptive immunity, leading to higher morbidity and mortality, the longevity of IgG antibodies, including neutralizing activity to SARS-CoV-2, is viewed as a key correlate of immune protection. Amid reports and concern that there is a rapid decay of IgG antibody levels within 1 mo to 2 mo after acute infection, we set out to study the pattern and duration of IgG antibody response to various SARS-CoV-2 antigens in asymptomatic and symptomatic patients in a community setting. Herein, we show the correlation of IgG anti-spike protein S1 subunit, receptor binding domain, nucleocapsid, and virus neutralizing antibody titers with each other and with clinical features such as length and severity of COVID-19 illness. More importantly, using orthogonal measurements, we found the IgG titers to persist for more than 4 mo post symptom onset, implying that long-lasting immunity to COVID-19 from infection or vaccination might be observed, as seen with other coronaviruses such as SARS and Middle East respiratory syndrome.