Long-term growth data of Escherichia coli at a single-cell level

Long-term, single-cell measurement of bacterial growth is extremely valuable information, particularly in the study of homeostatic aspects such as cell-size and growth rate control. Such measurement has recently become possible due to the development of microfluidic technology. Here we present data from single-cell measurements of Escherichia coli growth over 70 generations obtained for three different growth conditions. The data were recorded every minute, and contain time course data of cell length and fluorescent intensity of constitutively expressed yellow fluorescent protein.

Deep Learning Automates the Quantitative Analysis of Individual Cells in Live-Cell Imaging Experiments

Live-cell imaging has opened an exciting window into the role cellular heterogeneity plays in dynamic, living systems. A major critical challenge for this class of experiments is the problem of image segmentation, or determining which parts of a microscope image correspond to which individual cells. Current approaches require many hours of manual curation and depend on approaches that are difficult to share between labs. They are also unable to robustly segment the cytoplasms of mammalian cells. Here, we show that deep convolutional neural networks, a supervised machine learning method, can solve this challenge for multiple cell types across the domains of life. We demonstrate that this approach can robustly segment fluorescent images of cell nuclei as well as phase images of the cytoplasms of individual bacterial and mammalian cells from phase contrast images without the need for a fluorescent cytoplasmic marker. These networks also enable the simultaneous segmentation and identification of different mammalian cell types grown in co-culture. A quantitative comparison with prior methods demonstrates that convolutional neural networks have improved accuracy and lead to a significant reduction in curation time. We relay our experience in designing and optimizing deep convolutional neural networks for this task and outline several design rules that we found led to robust performance. We conclude that deep convolutional neural networks are an accurate method that require less curation time, are generalizable to a multiplicity of cell types, from bacteria to mammalian cells, and expand live-cell imaging capabilities to include multi-cell type systems.

A noisy linear map underlies oscillations in cell size and gene expression in bacteria

During bacterial growth, a cell approximately doubles in size before division, after which it splits into two daughter cells. This process is subjected to the inherent perturbations of cellular noise and thus requires regulation for cell-size homeostasis. The mechanisms underlying the control and dynamics of cell size remain poorly understood owing to the difficulty in sizing individual bacteria over long periods of time in a high-throughput manner. Here we measure and analyse long-term, single-cell growth and division across different Escherichia coli strains and growth conditions. We show that a subset of cells in a population exhibit transient oscillations in cell size with periods that stretch across several (more than ten) generations. Our analysis reveals that a simple law governing cell-size control-a noisy linear map-explains the origins of these cell-size oscillations across all strains. This noisy linear map implements a negative feedback on cell-size control: a cell with a larger initial size tends to divide earlier, whereas one with a smaller initial size tends to divide later. Combining simulations of cell growth and division with experimental data, we demonstrate that this noisy linear map generates transient oscillations, not just in cell size, but also in constitutive gene expression. Our work provides new insights into the dynamics of bacterial cell-size regulation with implications for the physiological processes involved.

Linear population allocation by bistable switches in response to transient stimulation

Many cellular decision processes, including proliferation, differentiation, and phenotypic switching, are controlled by bistable signaling networks. In response to transient or intermediate input signals, these networks allocate a population fraction to each of two distinct states (e.g. OFF and ON). While extensive studies have been carried out to analyze various bistable networks, they are primarily focused on responses of bistable networks to sustained input signals. In this work, we investigate the response characteristics of bistable networks to transient signals, using both theoretical analysis and numerical simulation. We find that bistable systems exhibit a common property: for input signals with short durations, the fraction of switching cells increases linearly with the signal duration, allowing the population to integrate transient signals to tune its response. We propose that this allocation algorithm can be an optimal response strategy for certain cellular decisions in which excessive switching results in lower population fitness.

Generic metric to quantify quorum sensing activation dynamics

Quorum sensing (QS) enables bacteria to sense and respond to changes in their population density. It plays a critical role in controlling different biological functions, including bioluminescence and bacterial virulence. It has also been widely adapted to program robust dynamics in one or multiple cellular populations. While QS systems across bacteria all appear to function similarly-as density-dependent control systems-there is tremendous diversity among these systems in terms of signaling components and network architectures. This diversity hampers efforts to quantify the general control properties of QS. For a specific QS module, it remains unclear how to most effectively characterize its regulatory properties in a manner that allows quantitative predictions of the activation dynamics of the target gene. Using simple kinetic models, here we show that the dominant temporal dynamics of QS-controlled target activation can be captured by a generic metric, 'sensing potential', defined at a single time point. We validate these predictions using synthetic QS circuits in Escherichia coli. Our work provides a computational framework and experimental methodology to characterize diverse natural QS systems and provides a concise yet quantitative criterion for selecting or optimizing a QS system for synthetic biology applications.

Programmed cell death in bacteria and implications for antibiotic therapy

It is now well appreciated that programmed cell death (PCD) plays critical roles in the life cycle of diverse bacterial species. It is an apparently paradoxical behavior as it does not benefit the cells undergoing PCD. However, growing evidence suggests that PCD can be 'altruistic': the dead cells may directly or indirectly benefit survivors through generation of public goods. This property provides a potential explanation on how PCD can evolve as an extreme form of cooperation, although many questions remain to be addressed. From another perspective, as PCD plays a critical role in bacterial pathogenesis, it has been proposed as a potential target for new antibacterial therapy. To this end, understanding the population and evolutionary dynamics resulting from PCD and public goods production may be a key to the success of designing effective antibiotic treatment.

Engineering microbial systems to explore ecological and evolutionary dynamics

A major goal of biological research is to provide a mechanistic understanding of diverse biological processes. To this end, synthetic biology offers a powerful approach, whereby biological questions can be addressed in a well-defined framework. By constructing simple gene circuits, such studies have generated new insights into the design principles of gene regulatory networks. Recently, this strategy has been applied to analyze ecological and evolutionary questions, where population-level interactions are critical. Here, we highlight recent development of such systems and discuss how they were used to address problems in ecology and evolutionary biology. As illustrated by these examples, synthetic ecosystems provide a unique platform to study ecological and evolutionary phenomena that are challenging to study in their natural contexts.

Programming stress-induced altruistic death in engineered bacteria

Programmed death is often associated with a bacterial stress response. This behavior appears paradoxical, as it offers no benefit to the individual. This paradox can be explained if the death is 'altruistic': the killing of some cells can benefit the survivors through release of 'public goods'. However, the conditions where bacterial programmed death becomes advantageous have not been unambiguously demonstrated experimentally. Here, we determined such conditions by engineering tunable, stress-induced altruistic death in the bacterium Escherichia coli. Using a mathematical model, we predicted the existence of an optimal programmed death rate that maximizes population growth under stress. We further predicted that altruistic death could generate the 'Eagle effect', a counter-intuitive phenomenon where bacteria appear to grow better when treated with higher antibiotic concentrations. In support of these modeling insights, we experimentally demonstrated both the optimality in programmed death rate and the Eagle effect using our engineered system. Our findings fill a critical conceptual gap in the analysis of the evolution of bacterial programmed death, and have implications for a design of antibiotic treatment.

Oscillations by minimal bacterial suicide circuits reveal hidden facets of host-circuit physiology

Synthetic biology seeks to enable programmed control of cellular behavior though engineered biological systems. These systems typically consist of synthetic circuits that function inside, and interact with, complex host cells possessing pre-existing metabolic and regulatory networks. Nevertheless, while designing systems, a simple well-defined interface between the synthetic gene circuit and the host is frequently assumed. We describe the generation of robust but unexpected oscillations in the densities of bacterium Escherichia coli populations by simple synthetic suicide circuits containing quorum components and a lysis gene. Contrary to design expectations, oscillations required neither the quorum sensing genes (luxR and luxI) nor known regulatory elements in the P_luxI promoter. Instead, oscillations were likely due to density-dependent plasmid amplification that established a population-level negative feedback. A mathematical model based on this mechanism captures the key characteristics of oscillations, and model predictions regarding perturbations to plasmid amplification were experimentally validated. Our results underscore the importance of plasmid copy number and potential impact of “hidden interactions” on the behavior of engineered gene circuits - a major challenge for standardizing biological parts. As synthetic biology grows as a discipline, increasing value may be derived from tools that enable the assessment of parts in their final context.

Decoding biological principles using gene circuits

A major flavor of synthetic biology is the creation of artificial gene circuits to perform user-defined tasks. One aspect of this area is to realize ever-increasingly more complicated circuit behavior. Such efforts have led to the identification and evaluation of design strategies that enable robust control of dynamics in single cells and in cell populations. On the other hand, there is increasing emphasis on using artificial systems programmed by simple circuits to explore fundamental biological questions of broad significance.

Noise reduction by diffusional dissipation in a minimal quorum sensing motif

Cellular interactions are subject to random fluctuations (noise) in quantities of interacting molecules. Noise presents a major challenge for the robust function of natural and engineered cellular networks. Past studies have analyzed how noise is regulated at the intracellular level. Cell-cell communication, however, may provide a complementary strategy to achieve robust gene expression by enabling the coupling of a cell with its environment and other cells. To gain insight into this issue, we have examined noise regulation by quorum sensing (QS), a mechanism by which many bacteria communicate through production and sensing of small diffusible signals. Using a stochastic model, we analyze a minimal QS motif in Gram-negative bacteria. Our analysis shows that diffusion of the QS signal, together with fast turnover of its transcriptional regulator, attenuates low-frequency components of extrinsic noise. We term this unique mechanism "diffusional dissipation" to emphasize the importance of fast signal turnover (or dissipation) by diffusion. We further show that this noise attenuation is a property of a more generic regulatory motif, of which QS is an implementation. Our results suggest that, in a QS system, an unstable transcriptional regulator may be favored for regulating expression of costly proteins that generate public goods.

Graves' disease in HTLV-I carriers

Three carriers of human T-lymphotropic virus type I (HTLV-I) with Graves' disease are reported. All three cases were complicated with uveitis, and one also showed chronic arthropathy. Anti-HLTV-I antibody was found in the serum by the particle agglutination method and western blotting, and HTLV-I proviral DNA was detected in peripheral lymphocytes by the polymerase chain reaction and Southern blotting. HTLV-I is a causal agent of adult T-cell leukemia and HTLV-I associated myelopathy/tropical spastic paraparesis, and is believed to be related to the pathogenesis of diseases such as chronic arthropathy, uveitis, chronic bronchoalveolitis, and Sjögren's syndrome. On the other hand, retrovirus infection is considered to cause autoimmune diseases. Thus, the pathogenesis of Graves' disease in the present patients might be associated with HTLV-I infection.

Endotoxin-induced uveitis in mice. 1. Induction of uveitis and role of T lymphocytes

Endotoxin-induced uveitis (EIU) with a high frequency of posterior iris synechiae was induced by the systemic injection of 200 micrograms of endotoxin into C3H/HeN mice, an endotoxin-responsive strain. The cell number and the protein concentration within the aqueous humor began to increase 6 hours after the injection, achieving a peak at 24 hours, and decreased gradually thereafter. Inflammatory cells were observed in the anterior chamber, the vitreous body and near the iris-ciliary body histologically. Most of the inflammatory cells were polymorphonuclear cells. On the other hand, C3H/HeJ mice, an endotoxin-unresponsive strain, showed no increase in either cell number or protein concentration in the aqueous humor after endotoxin administration. Pretreatment of C3H/HeN mice with anti-Thy-1.2 antibody significantly decreased both the cell number and the protein concentration in the aqueous humor and the incidence of the posterior synechiae, as compared with the control group. Anti-CD4 antibody also significantly reduced the severity of EIU, while anti-CD8 antibody had no influence on the disease. Anti-IFN-gamma antibody increased the cell number in the aqueous humor. These observations indicate that T lymphocytes, especially CD4+ T lymphocytes, have an extremely important role in the development of EIU in mice.

[Endogenous uveitis in disseminated intravascular coagulation induced by endotoxin]

A study was made of endogenous uveitis in experimental disseminated intravascular coagulation (DIC) in rabbits. Endotoxin was injected intravenously twice with a 24-hour interval. The time courses of the following were examined: 1) aqueous flare using a laser flare-cell meter 2) the number of leukocytes in the peripheral blood 3) the tumor necrosis factor (TNF) activity in the serum and 4) histopathological changes in the eye, lung, liver and kidney. Aqueous flare increased at 1 hour and was maximal at 6 hours, accompanied by a rapid increase in TNF activity at 1 hour following the first endotoxin administration. The number of leukocytes decreased to 963 +/- 266 cells/mm3 at 1.5 hours with subsequent leukocytosis within 12 hours. After the second injection of endotoxin, the aqueous flare peaked in 30 minutes and was twice as high as the first peak. Leukocyte number and TNF activity showed the same behavior. However, TNF activity was 20% that of the first peak. Histopathological examination indicated fibrin formation in the small vessels of systemic organs within 3 hours following the second administration of endotoxin. Endotoxin induced uveitis was induced in experimental DIC, and leukocytes and TNF activity may thus perform important roles.

Production of prostaglandin E2 rather than E1 in experimental ocular inflammation of rabbit

The chopped anterior uvea of rabbit was allowed to react with exogenous [1-14C] arachidonic acid or prostaglandin (PG)H2. Several cyclooxygenase products such as PGD2, PGE2 and PGF2 alpha were produced. In the presence of glutathione PGE2 was a major product. When uveitis was induced by injection of bovine serum albumin into the vitreous body, there was a marked invasion of leukocytes. PGE2 in the aqueous humor increased about 3-fold as determined by radioimmunoassay using anti-PGE2 antibody cross-reacting with PGE1. Extracts from the inflamed aqueous humor and the incubation medium of chopped anterior uvea or peripheral polymorphonuclear leukocytes were analyzed by reverse-phase high-performance liquid chromatography, which allowed separation of PGE2 and PGE1. In all these preparations PGE2 was an almost sole immunoreactive PG, and PGE1 was hardly detectable in sharp contrast to an earlier report (Eakins et al.: Nature 239: 248 (1972].

[The dynamics of leucocytes and complements in endotoxin induced uveitis]

The dynamics of leukocytes, complement and tumor necrosis factor (TNF) were studied in endotoxin induced uveitis (EIU) in rats. Endotoxin was administered to footpads and the time courses of the following were examined: 1) the number of leukocytes in peripheral blood 2) the number of leukocytes and protein concentration of aqueous humor 3) complement (CH50, C1, C3, C5) in serum and aqueous humor 4) TNF in serum 5) histological changes in the eyes, lungs, liver, and skin. The number of leukocytes in peripheral blood decreased to one third that of controls during three to six hours after endotoxin administration, but increased thereafter along with the number of leukocytes and protein concentration in aqueous humor. More leukocytes were observed than in controls in the small vessels of the iris, lung, liver and skin, some of which attached to the vascular endothelium. Serum C3 increased and serum C5 transiently decreased after endotoxin administration. Serum TNF reached a peak (3500 IU/ml) at 1.5 hours and rapidly decreased subsequently. It is suggested that the adhesion of leukocytes to vascular endothelium and the activation of complement system may play important roles in the development of EIU.

Tetranactin, a macrotetrolide antibiotic, suppresses in vitro proliferation of human lymphocytes and generation of cytotoxicity

Tetranactin, a hydrophobic cyclic antibiotic produced by Streptomyces aureus, has previously been shown to suppress in vitro activation of rat lymphocytes by concanavalin A as well as the onset of experimental autoimmune uveoretinitis in Lewis rats. Here we report the effects of tetranactin on human T and NK lymphocytes in vitro. Tetranactin, at concentrations up to 100 ng/ml, was not toxic to human lymphocytes but completely abrogated the proliferation of human T lymphocytes in response to allogeneic cells in mixed lymphocyte cultures. Tetranactin also blocked the initiation of proliferation in response to interleukin-2, but did not block proliferation of interleukin-2-activated cells. Tetranactin also blocked generation of cytotoxic T lymphocytes and activated killer cells in the mixed lymphocyte culture. However, up to 100 ng/ml tetranactin did not alter the lytic activity of cytotoxic T or NK lymphocytes generated in its absence. The ability of low doses of tetranactin to block the induction of lymphoproliferation is similar to the action of cyclosporin A. Since cyclosporin A is also a cyclic hydrophobic molecule, the immunosuppressive actions of these two agents may involve a similar mechanism.

Immunosuppressive and anti-proliferative effects of a macrotetrolide antibiotic, tetranactin

The macrotetrolide tetranactin suppressed the appearance of experimental autoimmune uveoretinitis induced in Lewis rats with a soluble retinal antigen (S-antigen). The drug at 1 ng/ml inhibited the mitogen activation of unfractionated lymphocytes; incorporation of radiolabelled precursors such as thymidine, uridine and leucine into the cells was markedly reduced. The synthesis and release of IL-2 by mitogen-activated lymphocytes was significantly suppressed in the presence of tetranactin. Incorporation of 45Ca was also inhibited, while intracellular Na+ levels were increased. In view of the ionophore property of tetranactin, it was suggested that the drug might demonstrate its immunosuppressive effect by affecting intracellular cation concentrations.

Superoxide in ocular inflammation: human and experimental uveitis

A possible involvement of superoxide in the pathogenesis of uveal inflammation in man and experimental animals was investigated. Superoxide production by the leukocytes of Behcet patients was significantly higher in the attack phase than in the remission phase. Leukocyte superoxide generation was also enhanced in guinea pigs with S-antigen-induced experimental autoimmune uveoretinitis (EAU). If the animals were treated with superoxide dismutase (SOD) at the onset of EAU, aqueous humor cell count was significantly lower than that of control (i.e., without SOD treatment). Infiltration of the inflammatory cells in the anterior retina was markedly reduced in SOD-treated animals. A similar protective effect of SOD against tissue damage was also observed in a bovine serum albumin-induced passive Arthus type uveitis in rabbits. These results suggest that superoxide may play a role in causing tissue damage in animal models of ocular inflammation and possibly in Behcet disease.