Final Report Summary - REGULATING FLIES (Physical principles of molecular signal integration in early fly development) During differentiation cells acquire specificity to fulfill a given function by expressing genes in particular molecular pathways. In the regulatory processes that control these decisions, the stochastic interaction of the molecular elements of the cell (proteins, genes) results in deterministic, reproducible outcomes -- functioning organisms. Living systems, both whole organisms and regulatory pathways, amaze us by the precision of their performance. How is this precision achieved under the physical constraints that biology must obey?In fruit fly development, proteins sequentially turn on genes, the expression of which lays out the blueprint for the segments of the insect's body. Genes code for proteins, which carry out specific functions. How are the experimentally observed cellular phenotypic states, which are directly linked to function, determined by the molecular interactions between proteins and genes. Experiments tell us that not all possible molecular strategies are realized in nature. Could certain molecular solutions be better suited for particular functions, such as noise reduction, information processing or phenotypic variability within a population? These problems are not limited to the study of fruit fly development and in some cases other systems provide interesting insight. However, the genetic toolbox developed in Drosophila allows for detailed measurements to be carried out in fly embryos, as well as the large community interested in studying this model organism experimentally and theoretically, makes early fly development a good system to test theoretical ideas.Development is an intrinsically dynamical process. However due to experimental limitations for a long time it remained studied on fixed embryos that offered only a static snapshot of the regulatory interactions that control development. Along with my experimental collaborators at the Curie Institute, N. Dostatni and M. Coppey, we developed a method to investigate the dynamics of transcription in the early stages of development in fruit flies. By analyzing this data we showed that after an initial reproducibly long period of inactivation after mitosis, transcription at the hunchback locus occurs in bursts, with clear states of reduced and enhanced expression. The precision of this dynamic process is quite low, even in the anterior region of the embryo. We also quantitatively described how transcription is initiated after mitosis, showing a spatial wave with the mean expression at the boundary increasing more slowly than in the anterior end. We also identified weak signatures of memory of the transcriptional state between subsequent cell cycles. The signal is weak due to strong external driving by the Bicoid gradient. Using a much more noisy system in fly development, a modified snail promoter, we identified to the first time strong transcriptional memory in development. Concurrently, I found the optimal network properties that allow for delayed information transmission given constraints on dissipation, showed how noise induced heterogeneity in bacterial populations can increase their fitness by exploiting a dynamical response and described a molecular integration mode that is useful when responding to opposing signals. Quantifying live imaging of fly embryo development. The simultaneous expression of the hunchback gene in the numerous nuclei of the developing fly embryo gives us a unique opportunity to study how transcription is regulated in living organisms. My experimental collaborators in the group of Nathalie Dostatni developed an MS2-MCP technique for imaging nascent messenger RNA in living Drosophila embryos, which allows us to quantify the dynamics of the developmental transcription process. We developed an efficient tracking and segmentation algorithm that further allowed for quantification of the observed microscopy movies. By doing this quantification, we discovered that Bicoid - the transcription factor traditionally considered to be sufficient along with hunchback self-regulation to generate sharp boundaries, is not essential for hunchback expression and does not provide sharp boundaries in early cell cycles, suggesting the role of an additional early regulator. The analysis of experiments also described the distributions of initiation times and characterized the expression of the hunchback promoter as a function of time and the embryo axis. We found that transcription initiation is independent of Bicoid concentrations, however Bicoid is needed to generate a sharp boundary between regions of high and low hunchback expression at later cell cycles.The initial measurement of the morphogens by the hunchback promoter takes place during very short cell cycles, not only giving each nucleus little time for a precise readout, but also resulting in short time traces of transcription. Additionally, the relationship between the measured signal and the promoter state depends on the molecular design of the reporting probe. We developed an analysis approach based on tailor made autocorrelation functions that overcomes the short trace problems and quantifies the dynamics of transcription initiation. Based on live imaging data, we identified signatures of bursty transcription initiation from the hunchback promoter. We showed that the precision of the expression of the hunchback gene to measure its position along the anterior-posterior axis is low both at the boundary and in the anterior even at cycle 13, suggesting additional post-transcriptional averaging mechanisms to provide the precision observed in fixed embryos.Life imaging techniques call for the development of new analysis techniques that depend on the details of the experimental design. We pointed out the trade-offs between signal-to-noise ratios and avoidance of problems coming from background fluorescence levels depending on the position of the fluorescent probes in the transgene construct. We discussed how an autocorrelation analysis can help in interpreting noisy data depending on the construct used.Transmission of active transcriptional states from mother to daughter cells has the potential to foster precision in the gene expression programs underlying development. Such transcriptional memory has been specifically proposed to promote rapid reactivation of gene expression profiles following cell division in Drosophila development. Given the short duration of the cell cycles, we considered this problem in the Bicoid-hunchback regulatory network. We identified weak signatures of memory at the boundary, but elsewhere in the embryo the signal is overridden by strong Bicoid regulation. Analyzing the data of my experimental collaborators in Montpellier, Berkeley and Paris, who monitored transcription from a modified noisy snail promoter in living Drosophila embryos, we showed the existence of transcriptional memory in fly development. They monitored the activities of stochastically expressed transgenes (using the MS2-MCP system described above) in order to distinguish active and inactive mother cells and the behaviors of their daughter nuclei following cell division. Our quantitative analyses revealed that there is a 4-fold higher probability for rapid reactivation when the mother experienced transcription. Memory nuclei activate transcription twice as fast as neighboring inactive mothers, leading to augmented levels of gene expression. We used dynamical traces to uncover the dynamical nature of how the hunchback expression pattern is formed in time. We found that the pattern formation is wave like: mRNAs are quickly made in the anterior end and form an expression boundary, which later moves to the more central region of the embryo. The increase in expression is not proportional along the anterior-posterior axis. The hunchback expression boundary reaches its final position within 4 minutes from the first observed transcription event. Surprisingly, upon reaching the final position, the boundary is already sharp and is not further refined in the remaining time of interphase.Transcriptional regulation. A crucial step in the regulation of gene expression in fly development is binding of transcription factor proteins such as Bicoid to regulatory sites along the DNA. But transcription factors act at nanomolar concentrations, and noise due to random arrival of these molecules at their binding sites can severely limit the precision of regulation. Recent work on the optimization of information flow through regulatory networks indicates that the lower end of the dynamic range of concentrations is simply inaccessible, overwhelmed by the impact of this noise. Motivated by the behavior of homeodomain proteins, such as the maternal morphogen Bicoid in the fruit fly embryo, with Sokolowski, Bialek and Tkacik we suggested a scheme in which transcription factors also act as indirect translational regulators, binding to the mRNA of other regulatory proteins. Intuitively, each mRNA molecule acts as an independent sensor of the input concentration, and averaging over these multiple sensors reduces the noise. We analyze information flow through this scheme and identify conditions under which it outperforms direct transcriptional regulation. Our results suggest that the dual role of homeodomain proteins is not just a historical accident, but a solution to a crucial physics problem in the regulation of gene expression.Information transmission under constraints in small gene networks. Regulatory response in gene circuits, such as the ones in early fly development, is often at a delay relative to input signaling, e.g. due to transcription and translation processes. Additionally, these circuits may function out of steady state, responding to inputs that are changing in time. With my collaborators, we found that topologies of maximally informative networks, in which the response is measured at a later time and the input signal, correspond to commonly occurring biological circuits with positive and negative feedback loops irrespective of the temporal delay specified. Most interestingly, circuits functioning out of steady state may exploit non-equilibrium absorbing states to transmit information optimally and feedback can additionally increase information transmission. We find that the there are many degenerate topologies that transmit similar information equally optimally.In order to transmit biochemical signals, biological regulatory systems dissipate energy with concomitant entropy production. For example, one of the likely modes of transcriptional regulation of hunchback - an irreversible cycle promoter model, is inherently nonequilbrium. Additionally, signaling often takes place in challenging environmental conditions. In a simple model regulatory circuit given by an input and a delayed output, I explored the trade-offs between information transmission and the system’s energetic efficiency. I determined the maximally informative network, given a fixed amount of entropy production and delayed response, exploring both the case with and without feedback. With my collaborators, we found that feedback allows the circuit to overcome energy constraints and transmit close to the maximum available information even in the dissipationless limit. Negative feedback loops, characteristic of shock responses, are optimal at high dissipation. Close to equilibrium positive feedback loops, known for their stability, become more informative. Asking how the signaling network should be constructed to best function in the worst possible environment, rather than an optimally tuned one or in steady state, we discovered that at large dissipation the same universal motif is optimal in all of these conditions. Systems that function of out of steady state transmit most information when they have slow relaxation dynamics. In this case feedback guarantees the best gain to cost ratio: these systems transmit the most information thanks to slow relaxation by dissipating less than systems without feedback.Noise increases the fitness of Bacillus subtilis. Fluctuations, or "noise", in the response of a biological system are usually thought to be harmful. However, it is becoming increasingly clear that in single-celled organisms, noise can sometimes help cells survive. This is because noise can enhance the diversity of responses within a cell population. Looking at the competence response of a population of Bacillus subtilis bacteria (competence is the ability of bacteria to take in DNA from their environment when under stress), I identified a novel benefit of molecular noise. Using computational modeling and experiments along with my collaborators in the Suel and Mugler groups I showed that noise increases the range of stress levels for which these bacteria exhibit a highly dynamic response, meaning that they are neither unresponsive, nor permanently in the competent state. Since a dynamic response is thought to be optimal for survival, this study suggests that noise is exploited to increase the fitness of the bacterial population.Phenotypic selection and heritability. Genetically identical cells in the same population can take on phenotypically variable states, leading to differentiated responses to external signals, such as nutrients and drug-induced stress. In order to understand the consequences for the population of having phenotypically variable cells when faced with external signals, in collaboration with T. Mora we studied a model to understand the structure of the phenotypic states, their heritability and stability under selective pressure. We considered the effects of selection acting on a single trait, which we explicitly linked to the variable number of proteins expressed by a gene. We showed how the population adapts the expression of this gene to enhance its fitness. We quantitatively related the overall fitness of the population to the heritability of expression levels and their diversity within the population. We showed how selection can influence both the variability in a population and the stability of the particular phenotypic states. We described conditions under which cells can transmit their phenotypic states to their offspring. This study opens the possibility to consider conditions for heritability of phenotypic states in more complex regulatory systems.Integration of regulatory signals in the innate immune system. Combining my interest in decision making and signal integration, driven by locally performed experiments, I looked at the integration of regulatory signals in the innate immune system, which functions as the first line of defense against pathogenic attacks. Specifically, I was interested in the role of feedback in signal integration in dendritic cells. In order to initiate an inflammatory response, cells of the innate immune system integrate environmental signals to decide whether bacterial signals are present in the blood. If bacterial signals are sensed, the response of the cells is to both produce and secrete TNF (Tumor necrosis factor) into the environment, which favours an inflammatory response, and to produce an antagonistic feedback in the form of the anti-inflammatory cytokine IL-10 (Interleukin 10). These signals are then sensed again by cells in the population leading to a final decision. Working in collaboration with the groups of V. Soumelis and T. Mora, I proposed a molecular mathematical model based on the idea of a bottleneck for opposing signal integration to understand how the opposing signals are integrated in order for the cell to make a decision about the concentration of the bacterial signal in the environment and as a result whether to initiate an inflammatory response or not. The model predicted the response of the cells in the case of a blocking both TNF and IL-10 receptors and was used distinguish autocrine and paracrine signaling in this system, showing that both pro and inflammatory signaling relies on paracrine signaling.
