cellPLATO - an unsupervised method for identifying cell behaviour in heterogeneous cell trajectory data.

Advances in imaging, segmentation and tracking have led to the routine generation of large and complex microscopy datasets. New tools are required to process this 'phenomics' type data. Here, we present 'Cell PLasticity Analysis Tool' (cellPLATO), a Python-based analysis software designed for measurement and classification of cell behaviours based on clustering features of cell morphology and motility. Used after segmentation and tracking, the tool extracts features from each cell per timepoint, using them to segregate cells into dimensionally reduced behavioural subtypes. Resultant cell tracks describe a 'behavioural ID' at each timepoint, and similarity analysis allows the grouping of behavioural sequences into discrete trajectories with assigned IDs. Here, we use cellPLATO to investigate the role of IL-15 in modulating human natural killer (NK) cell migration on ICAM-1 or VCAM-1. We find eight behavioural subsets of NK cells based on their shape and migration dynamics between single timepoints, and four trajectories based on sequences of these behaviours over time. Therefore, by using cellPLATO, we show that IL-15 increases plasticity between cell migration behaviours and that different integrin ligands induce different forms of NK cell migration.

cellPLATO: an unsupervised method for identifying cell behaviour in heterogeneous cell trajectory data.

Advances in imaging, cell segmentation, and cell tracking now routinely produce microscopy datasets of a size and complexity comparable to transcriptomics or proteomics. New tools are required to process this 'phenomics' type data. Cell PLasticity Analysis TOol (cellPLATO) is a Python-based analysis software designed for measurement and classification of diverse cell behaviours based on clustering of parameters of cell morphology and motility. cellPLATO is used after segmentation and tracking of cells from live cell microscopy data. The tool extracts morphological and motility metrics from each cell per timepoint, before being using them to segregate cells into behavioural subtypes with dimensionality reduction. Resultant cell tracks have a 'behavioural ID' for each cell per timepoint corresponding to their changing behaviour over time in a sequence. Similarity analysis allows the grouping of behavioural sequences into discrete trajectories with assigned IDs. Trajectories and underlying behaviours generate a phenotypic fingerprint for each experimental condition, and representative cells are mathematically identified and graphically displayed for human understanding of each subtype. Here, we use cellPLATO to investigate the role of IL-15 in modulating NK cell migration on ICAM-1 or VCAM-1. We find 8 behavioural subsets of NK cells based on their shape and migration dynamics, and 4 trajectories of behaviour. Therefore, using cellPLATO we show that IL-15 increases plasticity between cell migration behaviours and that different integrin ligands induce different forms of NK cell migration.

Machine learning enhanced cell tracking.

Quantifying cell biology in space and time requires computational methods to detect cells, measure their properties, and assemble these into meaningful trajectories. In this aspect, machine learning (ML) is having a transformational effect on bioimage analysis, now enabling robust cell detection in multidimensional image data. However, the task of cell tracking, or constructing accurate multi-generational lineages from imaging data, remains an open challenge. Most cell tracking algorithms are largely based on our prior knowledge of cell behaviors, and as such, are difficult to generalize to new and unseen cell types or datasets. Here, we propose that ML provides the framework to learn aspects of cell behavior using cell tracking as the task to be learned. We suggest that advances in representation learning, cell tracking datasets, metrics, and methods for constructing and evaluating tracking solutions can all form part of an end-to-end ML-enhanced pipeline. These developments will lead the way to new computational methods that can be used to understand complex, time-evolving biological systems.

Convolutional Neural Networks for Classifying Chromatin Morphology in Live-Cell Imaging.

Chromatin is highly structured, and changes in its organization are essential in many cellular processes, including cell division. Recently, advances in machine learning have enabled researchers to automatically classify chromatin morphology in fluorescence microscopy images. In this protocol, we develop user-friendly tools to perform this task. We provide an open-source annotation tool, and a cloud-based computational framework to train and utilize a convolutional neural network to automatically classify chromatin morphology. Using cloud compute enables users without significant resources or computational experience to use a machine learning approach to analyze their own microscopy data.

Cell-scale biophysical determinants of cell competition in epithelia.

How cells with different genetic makeups compete in tissues is an outstanding question in developmental biology and cancer research. Studies in recent years have revealed that cell competition can either be driven by short-range biochemical signalling or by long-range mechanical stresses in the tissue. To date, cell competition has generally been characterised at the population scale, leaving the single-cell-level mechanisms of competition elusive. Here, we use high time-resolution experimental data to construct a multi-scale agent-based model for epithelial cell competition and use it to gain a conceptual understanding of the cellular factors that governs competition in cell populations within tissues. We find that a key determinant of mechanical competition is the difference in homeostatic density between winners and losers, while differences in growth rates and tissue organisation do not affect competition end result. In contrast, the outcome and kinetics of biochemical competition is strongly influenced by local tissue organisation. Indeed, when loser cells are homogenously mixed with winners at the onset of competition, they are eradicated; however, when they are spatially separated, winner and loser cells coexist for long times. These findings suggest distinct biophysical origins for mechanical and biochemical modes of cell competition.

Single-cell approaches to cell competition: High-throughput imaging, machine learning and simulations.

Cell competition is a quality control mechanism in tissues that results in the elimination of less fit cells. Over the past decade, the phenomenon of cell competition has been identified in many physiological and pathological contexts, driven either by biochemical signaling or by mechanical forces within the tissue. In both cases, competition has generally been characterized based on the elimination of loser cells at the population level, but significantly less attention has been focused on determining how single-cell dynamics and interactions regulate population-wide changes. In this review, we describe quantitative strategies and outline the outstanding challenges in understanding the single cell rules governing tissue-scale competition dynamics. We propose quantitative metrics to characterize single cell behaviors in competition and use them to distinguish the types and outcomes of competition. We describe how such metrics can be measured experimentally using a novel combination of high-throughput imaging and machine learning algorithms. We outline the experimental challenges to quantify cell fate dynamics with high-statistical precision, and describe the utility of computational modeling in testing hypotheses not easily accessible in experiments. In particular, cell-based modeling approaches that combine mechanical interaction of cells with decision-making rules for cell fate choices provide a powerful framework to understand and reverse-engineer the diverse rules of cell competition.

Miro clusters regulate ER-mitochondria contact sites and link cristae organization to the mitochondrial transport machinery.

Mitochondrial Rho (Miro) GTPases localize to the outer mitochondrial membrane and are essential machinery for the regulated trafficking of mitochondria to defined subcellular locations. However, their sub-mitochondrial localization and relationship with other critical mitochondrial complexes remains poorly understood. Here, using super-resolution fluorescence microscopy, we report that Miro proteins form nanometer-sized clusters along the mitochondrial outer membrane in association with the Mitochondrial Contact Site and Cristae Organizing System (MICOS). Using knockout mouse embryonic fibroblasts we show that Miro1 and Miro2 are required for normal mitochondrial cristae architecture and Endoplasmic Reticulum-Mitochondria Contacts Sites (ERMCS). Further, we show that Miro couples MICOS to TRAK motor protein adaptors to ensure the concerted transport of the two mitochondrial membranes and the correct distribution of cristae on the mitochondrial membrane. The Miro nanoscale organization, association with MICOS complex and regulation of ERMCS reveal new levels of control of the Miro GTPases on mitochondrial functionality.

Context-Dependent Energetics of Loop Extensions in a Family of Tandem-Repeat Proteins.

Consensus-designed tetratricopeptide repeat proteins are highly stable, modular proteins that are strikingly amenable to rational engineering. They therefore have tremendous potential as building blocks for biomaterials and biomedicine. Here, we explore the possibility of extending the loops between repeats to enable further diversification, and we investigate how this modification affects stability and folding cooperativity. We find that extending a single loop by up to 25 residues does not disrupt the overall protein structure, but, strikingly, the effect on stability is highly context-dependent: in a two-repeat array, destabilization is relatively small and can be accounted for purely in entropic terms, whereas extending a loop in the middle of a large array is much more costly because of weakening of the interaction between the repeats. Our findings provide important and, to our knowledge, new insights that increase our understanding of the structure, folding, and function of natural repeat proteins and the design of artificial repeat proteins in biotechnology.

A Liquid to Solid Phase Transition Underlying Pathological Huntingtin Exon1 Aggregation.

Huntington's disease is caused by an abnormally long polyglutamine tract in the huntingtin protein. This leads to the generation and deposition of N-terminal exon1 fragments of the protein in intracellular aggregates. We combined electron tomography and quantitative fluorescence microscopy to analyze the structural and material properties of huntingtin exon1 assemblies in mammalian cells, in yeast, and in vitro. We found that huntingtin exon1 proteins can form reversible liquid-like assemblies, a process driven by huntingtin's polyQ tract and proline-rich region. In cells and in vitro, the liquid-like assemblies converted to solid-like assemblies with a fibrillar structure. Intracellular phase transitions of polyglutamine proteins could play a role in initiating irreversible pathological aggregation.

PyFolding: Open-Source Graphing, Simulation, and Analysis of the Biophysical Properties of Proteins.

For many years, curve-fitting software has been heavily utilized to fit simple models to various types of biophysical data. Although such software packages are easy to use for simple functions, they are often expensive and present substantial impediments to applying more complex models or for the analysis of large data sets. One field that is reliant on such data analysis is the thermodynamics and kinetics of protein folding. Over the past decade, increasingly sophisticated analytical models have been generated, but without simple tools to enable routine analysis. Consequently, users have needed to generate their own tools or otherwise find willing collaborators. Here we present PyFolding, a free, open-source, and extensible Python framework for graphing, analysis, and simulation of the biophysical properties of proteins. To demonstrate the utility of PyFolding, we have used it to analyze and model experimental protein folding and thermodynamic data. Examples include: 1) multiphase kinetic folding fitted to linked equations, 2) global fitting of multiple data sets, and 3) analysis of repeat protein thermodynamics with Ising model variants. Moreover, we demonstrate how PyFolding is easily extensible to novel functionality beyond applications in protein folding via the addition of new models. Example scripts to perform these and other operations are supplied with the software, and we encourage users to contribute notebooks and models to create a community resource. Finally, we show that PyFolding can be used in conjunction with Jupyter notebooks as an easy way to share methods and analysis for publication and among research teams.

Local cellular neighborhood controls proliferation in cell competition.

Cell competition is a quality-control mechanism through which tissues eliminate unfit cells. Cell competition can result from short-range biochemical inductions or long-range mechanical cues. However, little is known about how cell-scale interactions give rise to population shifts in tissues, due to the lack of experimental and computational tools to efficiently characterize interactions at the single-cell level. Here, we address these challenges by combining long-term automated microscopy with deep-learning image analysis to decipher how single-cell behavior determines tissue makeup during competition. Using our high-throughput analysis pipeline, we show that competitive interactions between MDCK wild-type cells and cells depleted of the polarity protein scribble are governed by differential sensitivity to local density and the cell type of each cell's neighbors. We find that local density has a dramatic effect on the rate of division and apoptosis under competitive conditions. Strikingly, our analysis reveals that proliferation of the winner cells is up-regulated in neighborhoods mostly populated by loser cells. These data suggest that tissue-scale population shifts are strongly affected by cellular-scale tissue organization. We present a quantitative mathematical model that demonstrates the effect of neighbor cell-type dependence of apoptosis and division in determining the fitness of competing cell lines.

Conformational transition of FGFR kinase activation revealed by site-specific unnatural amino acid reporter and single molecule FRET.

Protein kinases share significant structural similarity; however, structural features alone are insufficient to explain their diverse functions. Thus, bridging the gap between static structure and function requires a more detailed understanding of their dynamic properties. For example, kinase activation may occur via a switch-like mechanism or by shifting a dynamic equilibrium between inactive and active states. Here, we utilize a combination of FRET and molecular dynamics (MD) simulations to probe the activation mechanism of the kinase domain of Fibroblast Growth Factor Receptor (FGFR). Using genetically-encoded, site-specific incorporation of unnatural amino acids in regions essential for activation, followed by specific labeling with fluorescent moieties, we generated a novel class of FRET-based reporter to monitor conformational differences corresponding to states sampled by non phosphorylated/inactive and phosphorylated/active forms of the kinase. Single molecule FRET analysis in vitro, combined with MD simulations, shows that for FGFR kinase, there are populations of inactive and active states separated by a high free energy barrier resulting in switch-like activation. Compared to recent studies, these findings support diversity in features of kinases that impact on their activation mechanisms. The properties of these FRET-based constructs will also allow further studies of kinase dynamics as well as applications in vivo.

Single-Molecule Imaging to Characterize the Transport Mechanism of the Nuclear Pore Complex.

In the eukaryotic cell, a large macromolecular channel, known as the Nuclear Pore Complex (NPC), mediates all molecular transport between the nucleus and cytoplasm. In recent years, single-molecule fluorescence (SMF) imaging has emerged as a powerful tool to study the molecular mechanism of transport through the NPC. More recently, techniques such as single-molecule localization microscopy (SMLM) have enabled the spatial and temporal distribution of cargos, transport receptors and even structural components of the NPC to be determined with nanometre accuracy. In this protocol, we describe a method to study the position and/or motion of individual molecules transiting through the NPC with high spatial and temporal precision.

Cyanobacteria use micro-optics to sense light direction.

Bacterial phototaxis was first recognized over a century ago, but the method by which such small cells can sense the direction of illumination has remained puzzling. The unicellular cyanobacterium Synechocystis sp. PCC 6803 moves with Type IV pili and measures light intensity and color with a range of photoreceptors. Here, we show that individual Synechocystis cells do not respond to a spatiotemporal gradient in light intensity, but rather they directly and accurately sense the position of a light source. We show that directional light sensing is possible because Synechocystis cells act as spherical microlenses, allowing the cell to see a light source and move towards it. A high-resolution image of the light source is focused on the edge of the cell opposite to the source, triggering movement away from the focused spot. Spherical cyanobacteria are probably the world's smallest and oldest example of a camera eye.

Cyanobacteria are blue-green bacteria that are abundant in the environment. Cyanobacteria in the oceans are among the world's most important oxygen producers and carbon dioxide consumers. Synechocystis is a spherical single-celled cyanobacterium that measures about three thousandths of a millimetre across. Because Synechocystis needs sunlight to produce energy, it is important for it to find places where the light is neither too weak nor too strong. Unlike some bacteria, Synechocystis can't swim, but it can crawl across surfaces. It uses this ability to move to places where the light conditions are better. It was already known that Synechocystis cells move towards a light source that is shone at them from one side, which implies that the cyanobacteria can "see" where the light is. But how can such a tiny cell accurately detect where light is coming from? Schuergers et al. tracked how Synechocystis moved in response to different light conditions, and found that the secret of "vision" in these cyanobacteria is that the cells act as tiny spherical lenses. When a light is shone at the cell, an image of the light source is focused at the opposite edge of the cell. Light-detecting molecules called photoreceptors respond to the focused image of the light source, and this provides the information needed to steer the cell towards the light. Although the details are different, and although a Synechocystis cell is in terms of volume about 500 billion times smaller than a human eyeball, vision in Synechocystis actually works by principles similar to vision in humans. Schuergers et al.'s findings open plenty of further questions, as other types of bacteria may also act as tiny lenses. More also remains to be learnt about how the cyanobacteria process visual information.

Mapping the Topography of a Protein Energy Landscape.

Protein energy landscapes are highly complex, yet the vast majority of states within them tend to be invisible to experimentalists. Here, using site-directed mutagenesis and exploiting the simplicity of tandem-repeat protein structures, we delineate a network of these states and the routes between them. We show that our target, gankyrin, a 226-residue 7-ankyrin-repeat protein, can access two alternative (un)folding pathways. We resolve intermediates as well as transition states, constituting a comprehensive series of snapshots that map early and late stages of the two pathways and show both to be polarized such that the repeat array progressively unravels from one end of the molecule or the other. Strikingly, we find that the protein folds via one pathway but unfolds via a different one. The origins of this behavior can be rationalized using the numerical results of a simple statistical mechanics model that allows us to visualize the equilibrium behavior as well as single-molecule folding/unfolding trajectories, thereby filling in the gaps that are not accessible to direct experimental observation. Our study highlights the complexity of repeat-protein folding arising from their symmetrical structures; at the same time, however, this structural simplicity enables us to dissect the complexity and thereby map the precise topography of the energy landscape in full breadth and remarkable detail. That we can recapitulate the key features of the folding mechanism by computational analysis of the native structure alone will help toward the ultimate goal of designed amino-acid sequences with made-to-measure folding mechanisms-the Holy Grail of protein folding.

Importin-ß modulates the permeability of the nuclear pore complex in a Ran-dependent manner.

Soluble karyopherins of the importin-ß (impß) family use RanGTP to transport cargos directionally through the nuclear pore complex (NPC). Whether impß or RanGTP regulate the permeability of the NPC itself has been unknown. In this study, we identify a stable pool of impß at the NPC. A subpopulation of this pool is rapidly turned-over by RanGTP, likely at Nup153. Impß, but not transportin-1 (TRN1), alters the pore's permeability in a Ran-dependent manner, suggesting that impß is a functional component of the NPC. Upon reduction of Nup153 levels, inert cargos more readily equilibrate across the NPC yet active transport is impaired. When purified impß or TRN1 are mixed with Nup153 in vitro, higher-order, multivalent complexes form. RanGTP dissolves the impß•Nup153 complexes but not those of TRN1•Nup153. We propose that impß and Nup153 interact at the NPC's nuclear face to form a Ran-regulated mesh that modulates NPC permeability.

From artificial antibodies to nanosprings: the biophysical properties of repeat proteins.

In this chapter we review recent studies of repeat proteins, a class of proteins consisting of tandem arrays of small structural motifs that stack approximately linearly to produce elongated structures. We discuss the observation that, despite lacking the long-range tertiary interactions that are thought to be the hallmark of globular protein stability, repeat proteins can be as stable and as co-orperatively folded as their globular counterparts. The symmetry inherent in the structures of repeat arrays, however, means there can be many partly folded species (whether it be intermediates or transition states) that have similar stabilities. Consequently they do have distinct folding properties compared with globular proteins and these are manifest in their behaviour both at equilibrium and under kinetic conditions. Thus, when studying repeat proteins one appears to be probing a moving target: a relatively small perturbation, by mutation for example, can result in a shift to a different intermediate or transition state. The growing literature on these proteins illustrates how their modular architecture can be adapted to a remarkable array of biological and physical roles, both in vivo and in vitro. Further, their simple architecture makes them uniquely amenable to redesign-of their stability, folding and function-promising exciting possibilities for future research.

Selectivity mechanism of the nuclear pore complex characterized by single cargo tracking.

The nuclear pore complex (NPC) mediates all exchange between the cytoplasm and the nucleus. Small molecules can passively diffuse through the NPC, whereas larger cargos require transport receptors to translocate. How the NPC facilitates the translocation of transport receptor/cargo complexes remains unclear. To investigate this process, we tracked single protein-functionalized quantum dot cargos as they moved through human NPCs. Here we show that import proceeds by successive substeps comprising cargo capture, filtering and translocation, and release into the nucleus. Most quantum dots are rejected at one of these steps and return to the cytoplasm, including very large cargos that abort at a size-selective barrier. Cargo movement in the central channel is subdiffusive and cargos that can bind more transport receptors diffuse more freely. Without Ran GTPase, a critical regulator of transport directionality, cargos still explore the entire NPC, but have a markedly reduced probability of exit into the nucleus, suggesting that NPC entry and exit steps are not equivalent and that the pore is functionally asymmetric to importing cargos. The overall selectivity of the NPC seems to arise from the cumulative action of multiple reversible substeps and a final irreversible exit step.

ATPase site architecture and helicase mechanism of an archaeal MCM.

The subunits of the presumptive replicative helicase of archaea and eukaryotes, the MCM complex, are members of the AAA+ (ATPase-associated with various cellular activities) family of ATPases. Proteins within this family harness the chemical energy of ATP hydrolysis to perform a broad range of cellular processes. Here, we investigate the function of the AAA+ site in the mini-chromosome maintenance (MCM) complex of the archaeon Sulfolobus solfataricus (SsoMCM). We find that SsoMCM has an unusual active-site architecture, with a unique blend of features previously found only in distinct families of AAA+ proteins. We additionally describe a series of mutant doping experiments to investigate the mechanistic basis of intersubunit coordination in the generation of helicase activity. Our results indicate that MCM can tolerate catalytically inactive subunits and still function as a helicase, leading us to propose a semisequential model for helicase activity of this complex.