Cell motility: Bioelectrical control of behavior without neurons.

Single cells are capable of remarkably sophisticated, sometimes animal-like, behaviors. New work demonstrates bioelectric control of motility through the differential regulation of appendage movements in a unicellular organism that walks across surfaces using leg-like bundles of cilia.

The nuclear transport factor CSE1 drives macronuclear volume increase and macronuclear node coalescence in Stentor coeruleus.

Stentor coeruleus provides a unique opportunity to study how cells regulate nuclear shape because its macronucleus undergoes a rapid, dramatic, and developmentally regulated shape change. We found that the volume of the macronucleus increases during coalescence, suggesting an inflation-based mechanism. When the nuclear transport factor, CSE1, is knocked down by RNAi, the shape and volume changes of the macronucleus are attenuated, and nuclear morphology is altered. CSE1 protein undergoes a dynamic relocalization correlated with nuclear shape changes, being mainly cytoplasmic prior to nuclear coalescence, and accumulating inside the macronucleus during coalescence. At the end of regeneration, CSE1 protein levels are reduced as the macronucleus returns to its pre-coalescence volume. We propose a model in which nuclear transport via CSE1 is required to increase the volume of the macronucleus, thereby decreasing the surface-to-volume ratio and driving coalescence of the nodes into a single mass.

Homologous chromosome recognition via nonspecific interactions.

In many organisms, most notably Drosophila, homologous chromosomes in somatic cells associate with each other, a phenomenon known as somatic homolog pairing. Unlike in meiosis, where homology is read out at the level of DNA sequence complementarity, somatic homolog pairing takes place without double strand breaks or strand invasion, thus requiring some other mechanism for homologs to recognize each other. Several studies have suggested a "specific button" model, in which a series of distinct regions in the genome, known as buttons, can associate with each other, presumably mediated by different proteins that bind to these different regions. Here we consider an alternative model, which we term the "button barcode" model, in which there is only one type of recognition site or adhesion button, present in many copies in the genome, each of which can associate with any of the others with equal affinity. An important component of this model is that the buttons are non-uniformly distributed, such that alignment of a chromosome with its correct homolog, compared with a non-homolog, is energetically favored; since to achieve nonhomologous alignment, chromosomes would be required to mechanically deform in order to bring their buttons into mutual register. We investigated several types of barcodes and examined their effect on pairing fidelity. We found that high fidelity homolog recognition can be achieved by arranging chromosome pairing buttons according to an actual industrial barcode used for warehouse sorting. By simulating randomly generated non-uniform button distributions, many highly effective button barcodes can be easily found, some of which achieve virtually perfect pairing fidelity. This model is consistent with existing literature on the effect of translocations of different sizes on homolog pairing. We conclude that a button barcode model can attain highly specific homolog recognition, comparable to that seen in actual cells undergoing somatic homolog pairing, without the need for specific interactions. This model may have implications for how meiotic pairing is achieved.

Mitochondrial networks through the lens of mathematics.

Mitochondria serve a wide range of functions within cells, most notably via their production of ATP. Although their morphology is commonly described as bean-like, mitochondria often form interconnected networks within cells that exhibit dynamic restructuring through a variety of physical changes. Further, though relationships between form and function in biology are well established, the extant toolkit for understanding mitochondrial morphology is limited. Here, we emphasize new and established methods for quantitatively describing mitochondrial networks, ranging from unweighted graph-theoretic representations to multi-scale approaches from applied topology, in particular persistent homology. We also show fundamental relationships between mitochondrial networks, mathematics, and physics, using ideas of graph planarity and statistical mechanics to better understand the full possible morphological space of mitochondrial network structures. Lastly, we provide suggestions for how examination of mitochondrial network form through the language of mathematics can inform biological understanding, and vice versa.

The flagellar length control system: exploring the physical biology of organelle size.

How cells build and maintain dynamic structures of defined size is currently an important unsolved problem in quantitative cell biology. The flagella of the unicellular green algaChlamydomonasprovide a highly tractable model system to investigate this general question, but while the powerful genetics of this organism have revealed numerous genes required for proper flagellar length, in most cases we do not understand their mechanistic role in length control. Flagellar length can be viewed as the steady state solution of a dynamical system involving assembly and disassembly of axonemal microtubules, with assembly depending on an active transport process known as intraflagellar transport (IFT). The inherent length dependence of IFT gives rise to a family of simple models for length regulation that can account for many previously described phenomena such as the ability of flagella to maintain equal lengths. But these models requires that the cell has a way to measure flagellar length in order to adjust IFT rates accordingly. Several models for length sensing have been modeled theoretically and evaluated experimentally, allowing them to be ruled out. Current data support a model in which the diffusive return of the kinesin motor driving IFT provides a length dependence that ultimately is the basis for length regulation. By combining models of length sensing with a more detailed representation of cargo transport and availability, it is now becoming possible to formulate concrete hypotheses to explain length altering mutants.

Testing the ion-current model for flagellar length sensing and IFT regulation.

Eukaryotic cilia and flagella are microtubule-based organelles whose relatively simple shape makes them ideal for investigating the fundamental question of organelle size regulation. Most of the flagellar materials are transported from the cell body via an active transport process called intraflagellar transport (IFT). The rate of IFT entry into flagella, known as IFT injection, has been shown to negatively correlate with flagellar length. However, it remains unknown how the cell measures the length of its flagella and controls IFT injection. One of the most-discussed theoretical models for length sensing to control IFT is the ion-current model, which posits that there is a uniform distribution of Ca2+ channels along the flagellum and that the Ca2+ current from the flagellum into the cell body increases linearly with flagellar length. In this model, the cell uses the Ca2+ current to negatively regulate IFT injection. The recent discovery that IFT entry into flagella is regulated by the phosphorylation of kinesin through a calcium-dependent protein kinase has provided further impetus for the ion-current model. To test this model, we measured and manipulated the levels of Ca2+ inside of Chlamydomonas flagella and quantified IFT injection. Although the concentration of Ca2+ inside of flagella was weakly correlated with the length of flagella, we found that IFT injection was reduced in calcium-deficient flagella, rather than increased as the model predicted, and that variation in IFT injection was uncorrelated with the occurrence of flagellar Ca2+ spikes. Thus, Ca2+ does not appear to function as a negative regulator of IFT injection, hence it cannot form the basis of a stable length control system.

Role of intraflagellar transport in transcriptional control during flagellar regeneration in Chlamydomonas.

Biosynthesis of organelle precursors is a central part of the organelle size control problem, but what systems are required to control precursor production? Genes encoding flagellar proteins are up-regulated during flagellar regeneration in Chlamydomonas, and this up-regulation is critical for flagella to reach their final length, but it not known how the cell triggers these genes during regeneration. We present two models based on transcriptional repressor that is produced either in the flagellum or in the cell body and sequestered in the growing flagellum. Both models lead to stable flagellar length control and can reproduce the observed dynamics of gene expression. The two models make opposite predictions regarding the effect of mutations that block intraflagellar transport (IFT). Using quantitative measurements of gene expression, we show that gene expression during flagellar regeneration is greatly reduced in mutations of the heterotrimeric kinesin-2 that drives IFT. This result is consistent with the predictions of the model in which a repressor is sequestered in the flagellum by IFT. Inhibiting axonemal assembly has a much smaller effect on gene expression. The repressor sequestration model allows precursor production to occur when flagella are growing rapidly, representing a form of derivative control.

Single-cell analysis of habituation in Stentor coeruleus.

Although learning is often viewed as a unique feature of organisms with complex nervous systems, single-celled organisms also demonstrate basic forms of learning. The giant ciliate Stentor coeruleus responds to mechanical stimuli by contracting into a compact shape, presumably as a defense mechanism. When a Stentor cell is repeatedly stimulated at a constant level of force, it will learn to ignore that stimulus but will still respond to stronger stimuli. Prior studies of habituation in Stentor reported a graded response, suggesting that cells transition through a continuous range of response probabilities. By analyzing single cells using an automated apparatus to deliver calibrated stimuli, we find that habituation occurs via a single step-like switch in contraction probability within each cell, with the graded response in a population arising from the random distribution of switching times in individual cells. This step-like response allows Stentor behavior to be represented by a simple two-state model whose parameters can be estimated from experimental measurements. We find that transition rates depend on stimulus force and also on the time between stimuli. The ability to measure the behavior of the same cell to the same stimulus allowed us to quantify the functional heterogeneity among single cells. Together, our results suggest that the behavior of Stentor is governed by a two-state stochastic machine whose transition rates are sensitive to the time series properties of the input stimuli.

Modular, cascade-like transcriptional program of regeneration in Stentor.

The giant ciliate Stentor coeruleus is a classical model system for studying regeneration and morphogenesis in a single cell. The anterior of the cell is marked by an array of cilia, known as the oral apparatus, which can be induced to shed and regenerate in a series of reproducible morphological steps, previously shown to require transcription. If a cell is cut in half, each half regenerates an intact cell. We used RNA sequencing (RNAseq) to assay the dynamic changes in Stentor's transcriptome during regeneration, after both oral apparatus shedding and bisection, allowing us to identify distinct temporal waves of gene expression including kinases, RNA -binding proteins, centriole biogenesis factors, and orthologs of human ciliopathy genes. By comparing transcriptional profiles of different regeneration events, we identified distinct modules of gene expression corresponding to oral apparatus regeneration, posterior holdfast regeneration, and recovery after wounding. By measuring gene expression after blocking translation, we show that the sequential waves of gene expression involve a cascade mechanism in which later waves of expression are triggered by translation products of early-expressed genes. Among the early-expressed genes, we identified an E2F transcription factor and the RNA-binding protein Pumilio as potential regulators of regeneration based on the expression pattern of their predicted target genes. RNAi-mediated knockdown experiments indicate that Pumilio is required for regenerating oral structures of the correct size. E2F is involved in the completion of regeneration but is dispensable for earlier steps. This work allows us to classify regeneration genes into groups based on their potential role for regeneration in distinct cell regeneration paradigms, and provides insight into how a single cell can coordinate complex morphogenetic pathways to regenerate missing structures.

A unicellular walker controlled by a microtubule-based finite-state machine.

Cells are complex biochemical systems whose behaviors emerge from interactions among myriad molecular components. Computation is often invoked as a general framework for navigating this cellular complexity. However, it is unclear how cells might embody computational processes such that the theories of computation, including finite-state machine models, could be productively applied. Here, we demonstrate finite-state-machine-like processing embodied in cells using the walking behavior of Euplotes eurystomus, a ciliate that walks across surfaces using fourteen motile appendages (cirri). We found that cellular walking entails regulated transitions among a discrete set of gait states. The set of observed transitions decomposes into a small group of high-probability, temporally irreversible transitions and a large group of low-probability, time-symmetric transitions, thus revealing stereotypy in the sequential patterns of state transitions. Simulations and experiments suggest that the sequential logic of the gait is functionally important. Taken together, these findings implicate a finite-state-machine-like process. Cirri are connected by microtubule bundles (fibers), and we found that the dynamics of cirri involved in different state transitions are associated with the structure of the fiber system. Perturbative experiments revealed that the fibers mediate gait coordination, suggesting a mechanical basis of gait control.

Modeling cell biological features of meiotic chromosome pairing to study interlock resolution.

During meiosis, homologous chromosomes become associated side by side in a process known as homologous chromosome pairing. Pairing requires long range chromosome motion through a nucleus that is full of other chromosomes. It remains unclear how the cell manages to align each pair of chromosomes quickly while mitigating and resolving interlocks. Here, we use a coarse-grained molecular dynamics model to investigate how specific features of meiosis, including motor-driven telomere motion, nuclear envelope interactions, and increased nuclear size, affect the rate of pairing and the mitigation/resolution of interlocks. By creating in silico versions of three yeast strains and comparing the results of our model to experimental data, we find that a more distributed placement of pairing sites along the chromosome is necessary to replicate experimental findings. Active motion of the telomeric ends speeds up pairing only if binding sites are spread along the chromosome length. Adding a meiotic bouquet significantly speeds up pairing but does not significantly change the number of interlocks. An increase in nuclear size slows down pairing while greatly reducing the number of interlocks. Interestingly, active forces increase the number of interlocks, which raises the question: How do these interlocks resolve? Our model gives us detailed movies of interlock resolution events which we then analyze to build a step-by-step recipe for interlock resolution. In our model, interlocks must first translocate to the ends, where they are held in a quasi-stable state by a large number of paired sites on one side. To completely resolve an interlock, the telomeres of the involved chromosomes must come in close proximity so that the cooperativity of pairing coupled with random motion causes the telomeres to unwind. Together our results indicate that computational modeling of homolog pairing provides insight into the specific cell biological changes that occur during meiosis.

Determining protein polarization proteome-wide using physical dissection of individual Stentor coeruleus cells.

Cellular components are non-randomly arranged with respect to the shape and polarity of the whole cell.1-4 Patterning within cells can extend down to the level of individual proteins and mRNA.5,6 But how much of the proteome is actually localized with respect to cell polarity axes? Proteomics combined with cellular fractionation7-11 has shown that most proteins localize to one or more organelles but does not tell us how many proteins have a polarized localization with respect to the large-scale polarity axes of the intact cell. Genome-wide localization studies in yeast12-15 found that only a few percent of proteins have a localized position relative to the cell polarity axis defined by sites of polarized cell growth. Here, we describe an approach for analyzing protein distribution within a cell with a visibly obvious global patterning-the giant ciliate Stentor coeruleus.16,17 Ciliates, including Stentor, have highly polarized cell shapes with visible surface patterning.1,18 A Stentor cell is roughly 2 mm long, allowing a "proteomic dissection" in which microsurgery is used to separate cellular fragments along the anterior-posterior axis, followed by comparative proteomic analysis. In our analysis, 25% of the proteome, including signaling proteins, centrin/SFI proteins, and GAS2 orthologs, shows a polarized location along the cell's anterior-posterior axis. We conclude that a large proportion of all proteins are polarized with respect to global cell polarity axes and that proteomic dissection provides a simple and effective approach for spatial proteomics.

Biosynthesis of Linear Protein Nanoarrays Using the Flagellar Axoneme.

Applications in biotechnology and synthetic biology often make use of soluble proteins, but there are many potential advantages of anchoring enzymes to a stable substrate, including stability and the possibility for substrate channeling. To avoid the necessity of protein purification and chemical immobilization, there has been growing interest in bio-assembly of protein-containing nanoparticles, exploiting the self-assembly of viral capsid proteins or other proteins that form polyhedral structures. However, these nanoparticles are limited in size, which constrains the packaging and the accessibility of the proteins. An axoneme, the insoluble protein core of the eukaryotic flagellum or cilium, is a highly ordered protein structure that can be several microns in length, orders of magnitude larger than other types of nanoparticles. We show that when proteins of interest are fused to specific axonemal proteins and expressed in living Chlamydomonas reinhardtii cells, they become incorporated into linear arrays, which have the advantages of high protein loading capacity and single-step purification with retention of biomass. The arrays can be isolated as membrane-enclosed vesicles or as exposed protein arrays. The approach is demonstrated for both a fluorescent protein and an enzyme (beta-lactamase), showing that incorporation into axonemes retains protein function in a stable, easily isolated array form.

A simple method to generate human airway epithelial organoids with externally orientated apical membranes.

Organoids, which are self-organizing three-dimensional cultures, provide models that replicate specific cellular components of native tissues or facets of organ complexity. We describe a simple method to generate organoid cultures using isolated human tracheobronchial epithelial cells grown in mixed matrix components and supplemented at day 14 with the Wnt pathway agonist R-spondin 2 (RSPO2) and the bone morphogenic protein antagonist Noggin. In contrast to previous reports, our method produces differentiated tracheobronchospheres with externally orientated apical membranes without pretreatments, providing an epithelial model to study cilia formation and function, disease pathogenesis, and interaction of pathogens with the respiratory mucosa. Starting from 3 × 105 cells, organoid yield at day 28 was 1,720 ± 302. Immunocytochemistry confirmed the cellular localization of airway epithelial markers, including CFTR, Na+/K+ ATPase, acetylated-α-tubulin, E-cadherin, and ZO-1. Compared to native tissues, expression of genes related to bronchial differentiation and ion transport were similar in organoid and air-liquid interface (ALI) cultures. In matched primary cultures, mean organoid cilia length was 6.1 ± 0.2 µm, similar to that of 5.7 ± 0.1 µm in ALI cultures, and ciliary beating was vigorous and coordinated with frequencies of 7.7 ± 0.3 Hz in organoid cultures and 5.3 ± 0.8 Hz in ALI cultures. Functional measurement of osmotically induced volume changes in organoids showed low water permeability. The generation of numerous single testable units from minimal starting material complements prior techniques. This culture system may be useful for studying airway biology and pathophysiology, aiding diagnosis of ciliopathies, and potentially for high-throughput drug screening.

Regeneration in Stentor coeruleus.

We often think about regeneration in terms of replacing missing structures, such as organs or tissues, with new structures generated via cell proliferation and differentiation. But at a smaller scale, single cells, themselves, are capable of regenerating when part of the cell has been removed. A classic model organism that facilitates the study of cellular regeneration in the giant ciliate Stentor coeruleus. These cells, which can grow to more than a millimeter in size, have the ability to survive after extensive wounding of their surface, and are able to regenerate missing structures. Even a small piece of a cell can regenerate a whole cell with normal geometry, in a matter of hours. Such regeneration requires cells to be able to trigger organelle biogenesis in response to loss of structures. But subcellular regeneration also relies on intracellular mechanisms to create and maintain global patterning within the cell. These mechanisms are not understood, but at a conceptual level they involve processes that resemble those seen in animal development and regeneration. Here we discuss single-celled regeneration in Stentor from the viewpoint of standard regeneration paradigms in animals. For example, there is evidence that regeneration of the oral apparatus in Stentor follows a sender-receiver model similar to crustacean eyestalk regeneration. By drawing these analogies, we find that many of the concepts already known from the study of animal-scale regeneration and development can be applied to the study of regeneration at the cellular level, such as the concepts of determination, induction, mosaic vs. regulative development, and epimorphosis vs. morphallaxis. We propose that the similarities may go beyond analogy, and that some aspects of animal development and regeneration may have evolved by exploiting pre-existing subcellular developmental strategies from unicellular ancestors.

Analysis of Motility Patterns of Stentor During and After Oral Apparatus Regeneration Using Cell Tracking.

Stentor coeruleus is a well-known model organism for the study of unicellular regeneration. Transcriptomic analysis of individual cells revealed hundreds of genes-many not associated with the oral apparatus (OA)-that are differentially regulated in phases throughout the regeneration process. It was hypothesized that this systemic reorganization and mobilization of cellular resources towards growth of a new OA will lead to observable changes in movement and behavior corresponding in time to the phases of differential gene expression. However, the morphological complexity of S. coeruleus necessitated the development of an assay to capture the statistics and timescale. A custom script was used to track cells in short videos, and statistics were compiled over a large population (N ~100). Upon loss of the OA, S. coeruleus initially loses the ability for directed motion; then starting at ~4 h, it exhibits a significant drop in speed until ~8 h. This assay provides a useful tool for the screening of motility phenotypes and can be adapted for the investigation of other organisms.

Analysis of biological noise in the flagellar length control system.

Any proposed mechanism for organelle size control should be able to account not only for average size but also for the variation in size. We analyzed cell-to-cell variation and within-cell variation of length for the two flagella in Chlamydomonas, finding that cell-to-cell variation is dominated by cell size, whereas within-cell variation results from dynamic fluctuations. Fluctuation analysis suggests tubulin assembly is not directly coupled with intraflagellar transport (IFT) and that the observed length fluctuations reflect tubulin assembly and disassembly events involving large numbers of tubulin dimers. Length variation is increased in long-flagella mutants, an effect consistent with theoretical models for flagellar length regulation. Cells with unequal flagellar lengths show impaired swimming but improved gliding, raising the possibility that cells have evolved mechanisms to tune biological noise in flagellar length. Analysis of noise at the level of organelle size provides a way to probe the mechanisms determining cell geometry.

Microfluidic guillotine reveals multiple timescales and mechanical modes of wound response in Stentor coeruleus.

BACKGROUND: Wound healing is one of the defining features of life and is seen not only in tissues but also within individual cells. Understanding wound response at the single-cell level is critical for determining fundamental cellular functions needed for cell repair and survival. This understanding could also enable the engineering of single-cell wound repair strategies in emerging synthetic cell research. One approach is to examine and adapt self-repair mechanisms from a living system that already demonstrates robust capacity to heal from large wounds. Towards this end, Stentor coeruleus, a single-celled free-living ciliate protozoan, is a unique model because of its robust wound healing capacity. This capacity allows one to perturb the wounding conditions and measure their effect on the repair process without immediately causing cell death, thereby providing a robust platform for probing the self-repair mechanism. RESULTS: Here we used a microfluidic guillotine and a fluorescence-based assay to probe the timescales of wound repair and of mechanical modes of wound response in Stentor. We found that Stentor requires ~ 100-1000 s to close bisection wounds, depending on the severity of the wound. This corresponds to a healing rate of ~ 8-80 µm2/s, faster than most other single cells reported in the literature. Further, we characterized three distinct mechanical modes of wound response in Stentor: contraction, cytoplasm retrieval, and twisting/pulling. Using chemical perturbations, active cilia were found to be important for only the twisting/pulling mode. Contraction of myonemes, a major contractile fiber in Stentor, was surprisingly not important for the contraction mode and was of low importance for the others. CONCLUSIONS: While events local to the wound site have been the focus of many single-cell wound repair studies, our results suggest that large-scale mechanical behaviors may be of greater importance to single-cell wound repair than previously thought. The work here advances our understanding of the wound response in Stentor and will lay the foundation for further investigations into the underlying components and molecular mechanisms involved.

Deep Convolutional and Recurrent Neural Networks for Cell Motility Discrimination and Prediction.

Cells in culture display diverse motility behaviors that may reflect differences in cell state and function, providing motivation to discriminate between different motility behaviors. Current methods to do so rely upon manual feature engineering. However, the types of features necessary to distinguish between motility behaviors can vary greatly depending on the biological context, and it is not always clear which features may be most predictive in each setting for distinguishing particular cell types or disease states. Convolutional neural networks (CNNs) are machine learning models allowing for relevant features to be learned directly from spatial data. Similarly, recurrent neural networks (RNNs) are a class of models capable of learning long term temporal dependencies. Given that cell motility is inherently spacio-temporal data, we present an approach utilizing both convolutional and long- short-term memory (LSTM) recurrent neural network units to analyze cell motility data. These RNN models provide accurate classification of simulated motility and experimentally measured motility from multiple cell types, comparable to results achieved with hand-engineered features. The variety of cell motility differences we can detect suggests that the algorithm is generally applicable to additional cell types not analyzed here. RNN autoencoders based on the same architecture are capable of learning motility features in an unsupervised manner and capturing variation between myogenic cells in the latent space. Adapting these RNN models to motility prediction, RNNs are capable of predicting muscle stem cell motility from past tracking data with performance superior to standard motion prediction models. This advance in cell motility prediction may be of practical utility in cell tracking applications.