Although locomotion may seem effortless, it relies on spinal circuits producing complex patterns of muscle contractions. Spinal cord central pattern generator circuits (CPGs) produce motor output autonomously as a result of descending commands from the brain and modulations caused by internal physiological cues. The cerebrospinal fluid (CSF) constitutes a fascinating interface through which chemical cues produced in the brain can reach neurons within the entire central nervous system. Located along the central canal, CSF-contacting neurons (CSFns) project a ciliary tuft into the CSF and an axon into the spinal cord. These mysterious cells were identified in over 200 species of vertebrates by Kolmer and Agduhr who hypothesized that they could act as a proprioceptive sensory neuron. The difficulties of labeling and accessing these cells have prevented the characterization of their sensory physiology or specification. However newly developed tools and experimental approaches in zebrafish provide us with a unique opportunity to overcome these technical challenges. The transparency of zebrafish at early stages makes it a powerful system for identifying neurons by their morphology and for investigating and manipulating neuronal activity of specific cells in the intact animal. Using an optogenetic approach, Dr. Wyart and colleagues discovered that CSFns could strongly modulate the activity of spinal CPGs. In this proposed project, the Delmas, Lewis and Wyart labs will combine expression profiling, channel physiology, behavioural analysis and loss-of-function assays to elucidate how CSFns are specified and how they sense chemical and mechanical cues in the CSF. The Lewis lab has pioneered the use of FAC sorting and transcriptome analysis to identify genes expressed in distinct subtypes of spinal neurons. The Delmas lab has unraveled the functions of channels underlying chemo- and mechano-sensation in mammals. The Wyart lab has combined optical and physiological approaches to identify the functions of CSFns. Together we will establish how genetic networks specify these cells and unravel the molecular mechanisms by which CSFns are physiologically activated in vivo and contribute to behaviour.

Autophagy, the process whereby cells degrade their own components, is essential for normal cell function and mammalian development, and also has been implicated in numerous human diseases. Autophagy involves a wide array of protein-protein interactions, including a large number involving ubiquitin, ubiquitin-like proteins and various adaptor modules. The ability to target and disrupt these protein-protein interactions would enable dissection of the molecular details of autophagy, which would greatly enhance our basic understanding of the process and could lead to the identification of targets for therapeutic intervention. This project will develop and apply a toolkit of protein-based modulators to explore the roles of distinct molecular networks in autophagy. We will first identify protein interaction interfaces, and subsequently, will use computational design and combinatorial protein engineering to develop high affinity sensors or inhibitors that can function in live cells. Structural studies will be performed to determine the molecular basis of their action. These tools will then be used in vivo settings, using variety of cell and mouse models, to monitor and control physiological and pathophysiological autophagy processes.

The processing and encoding of external stimuli is essential for all life forms to react appropriately to environmental cues. When animals move, however, they activate their own sensory systems. Expectations about self-generated sensory input and its subtraction from the total input are essential for accurate sensory processing. Organisms have optimized a solution for distinguishing between self-generated and external stimulations by involving a descending (efferent) neural pathway that modulates ascending (afferent) activity. Mechanosensory hair cells, for example, are subject to feedback regulation by cholinergic efferent neurons that originate from several nuclei in the brainstem. The present research proposal aims at understanding how efferent neurons control mechanosensory processing.
Fishes and amphibians have a superficial and accessible mechanosensory system, the lateral line, which circumvents many of these problems. The lateral line serves to detect hydromechanical variations around the animal’s body and controls two characteristic behaviors: the stereotypic fast escape responses to sudden mechanical input, and complex and highly variable obstacle avoidance and rheotaxis. The simplicity of the lateral line in the zebrafish larva makes it a particularly powerful experimental model system.
Versatile recent technological developments suggest that efferent/afferent interactions in the zebrafish lateral line can be interrogated using a combination of pharmacological and genetic manipulations of neural circuits in vivo, intravital structural and functional imaging of neuronal ensembles, and fictive behavioral paradigms. Combining the knowledge and strength of each group, we will be able to clarify efferent/afferent interactions and lay the foundations to further ask how the nervous system adjusts its sensitivity to discriminate between external- and self-stimulation. This collaborative project is based on molecular studies of efferent neurons, in toto visualization and manipulation of the zebrafish lateral line, and zebrafish neurophysiology, functional imaging and behavior.

Continuous location estimation is a difficult computational problem. Roboticists grappling with the problem of noisy sensor measurements, changing features in the external world, and novel environments, find that probabilistic representation and inference of locations yields good performance, but storing and updating a single best-guess about location does not. Animals are capable of spatial navigation performance that equals or bests that of mobile robots. Yet our understanding of animal navigation is predicated on the framework of single-hypothesis tracking; there is little exploration of the compelling possibility that the brain’s spatial representations enable probabilistic coding and inference.
In this proposal, we will directly test the general hypothesis that mammals use probabilistic representation and inference for location estimation. We will address this possibility by combining investigation, in mice and humans, with computational models of probabilistic computation on spatial variables. Applying state-of-the art techniques in computational simulation and theory, in vivo electrophysiology and genetic manipulations in mice, and ultra-high resolution neuroimaging and behavior in humans, we will determine if (i) performance of spatial behaviors, (ii) representation of spatial information by place cells, and (iii) high resolution signatures of activity obtained with fMRI, are each consistent with implementation of probabilistic computations. We will in addition seek to obtain the first evidence for molecular mechanisms that control probabilistic computation. Similar experimental paradigms for animals and humans will ensure comparability of our results between species. Implementation of close reciprocal interactions between the three participating laboratories will ensure that the computational modelling studies generate theoretical predictions for the experimental studies, the results of which will in turn help refine the theoretical models.
Taken together, our multidimensional, interactive approach has the potential to provide the first account of the neural mechanisms that allow our brains to perform accurate spatial computations in the face of sensory uncertainty. This will provide a new conceptual framework for investigation of neural circuits underlying navigation and for integration of experimental results obtained in mice and humans.

One essential element very often missing to explain cell-fate control concerns the spatiotemporal regulation of translation. How does translation at synapses regulate the growth and connection strength of individual synapses during learning process? How does localized mRNA translation contribute to maintaining the pluripotency of stem cells? To address these fundamental biological questions, we propose a multi-disciplinary approach combining tools and concepts from synthetic biology, biophysics and neurobiology. Our strategy is to rationally design functionally controllable RNA-protein (RNP) complexes to spatiotemporally control mRNA translation in living cells. One of the main objectives will be to build synthetic RNPs that can traffic, translate or degrade a mRNA molecule. Our design will integrate magnetic nanoparticles and biological molecules to combine both the functional properties of native RNPs and the capability of being manipulated in space and time using magnetic forces. We hypothesize that local translation in cells often play critical roles in cell-fate decisions, such as the establishment of polarity and regulation of synapse formation and plasticity. However, currently available techniques are limited with their capacity in regulating local translation with spatiotemporal precision. Our new engineering approaches to develop synthetic RNPs in target cells will compensate pre-existing techniques by adding precise spatiotemporal controls. With synthetic RNP-mediated spatiotemporal translational regulatory systems, we will ask fundamental biological questions that are not accessible using current cell biology and ‘omics’ approaches, including: (1) What is the specific role of local translation at a synapse in terms of synaptic plasticity and memory formation and consolidation? (2) Can we control cellular behaviors of stem cells or differentiated cells by providing spatiotemporal gene expression dimensions? The new apporach will allow us to examine the effect of localized mRNA translation in controlling cell-fate behavior for two general cell states including pluripotency of stem cells and differentiated state of neurons. Such approaches will open new directions to better understand how local translation contributes to cell-fate conversion.

The mammalian cerebral cortex commands all higher order brain functions, including perception, language, emotion and cognition. Our understanding of how a functional cortex is constructed from a set of specific neural stem cells remains however largely fragmentary. The long-term goal of this application is to fill this knowledge gap. As a working hypothesis we propose that the lineage relationship of nascent neurons fundamentally influences the structural and functional assembly of the cortex. To test our hypothesis, we will integrate state-of-the-art experimental, bioinformatics and mathematical modeling approaches to quantitatively analyze the genesis of excitatory neurons and their integration in the developing mouse cortex at the clonal level. First we will systematically trace the lineage trees of cortical excitatory neurons in vivo and at clonal density using the MADM (Mosaic Analysis with Double Markers) technique. Using computational approaches we perform 3D image analysis to map the spatiotemporal progression of individual lineages. Utilizing the unprecedented resolution of experimental MADM datasets, we will delineate the exact cellular events responsible for the step-wise generation of discrete lineages, and develop mathematical models that quantitatively describe progenitor cell behavior at both clonal and population levels. Next, we aim to identify molecular signatures that define individual lineages. Guided by knowledge based on systematic characterization of cortical lineages, we will perform transcriptome analysis to compare and contrast gene expression within individual lineages and across distinct lineages. Comprehensive bioinformatics and statistical algorithms will be employed to uncover potential lineage-specific molecular signatures. Finally, we will holistically determine the structural and functional organization of clonally-related cortical neurons. To this end we will reconstruct individual lineages in the mature cortex, extract their spatial organization principles, and systematically map synaptic connectivity within and across lineages using multi-electrode whole-cell recordings and optical stimulation to reveal functional organization principles. The innovative integration of unique, powerful interdisciplinary approaches promises in-depth insights for understanding mammalian cortical development and function.

In the left hemisphere of the human brain, damage to the ventrolateral frontal (VLF) cortex, which includes a specialized region known as Broca’s area and adjacent precentral motor cortex, leads to speech problems. At the same time, there is an older system playing some as yet unspecified role in vocalization and speech in the dorsomedial frontal (DMF) region, which includes the supplementary motor area, presupplementary motor cortex, and the adjacent cingulate motor areas. Little is known of how these two systems interact anatomically and functionally. In primates that lack speech, such as chimpanzees and macaque monkeys, these regions may be involved in voluntary vocal production. Critically, there is nothing known of how these two systems may or may not have changed during primate brain evolution with increasing selection for language and speech functions in the human lineage. The aim of this interdisciplinary group of scientists is to adopt an evolutionary framework to study the role of these two brain regions in vocalization/speech, their cytoarchitectonic characteristics, and their anatomical and functional connectivity. The multimodal approach proposed here is particularly innovative because it combines comparative cytoarchitectonic analysis in three critical primate species to establish comparable areas. Further, invasive anatomical connection tracing studies that can only be carried out in the macaque monkey will provide totally unambiguous connection data that can then be employed to test predictions about functional and anatomical connectivity in the human and chimpanzee brains using non-invasive procedures including diffusion tensor imaging and resting state functional magnetic resonance imaging. Finally, functional imaging techniques will be used to characterize the VLF and DMF networks in relation to vocal production. This is the first systematic comparative study undertaken in the three most critical primate species for understanding DMF and VLF regions. Each research team brings a unique set of scientific skills and pragmatic strengths that provides for a novel, interdisciplinary effort to advance our understanding of the origins of vocal production and eventually human speech.

During the past 40 years, behavioral studies have confirmed the presence of magnetoreception in all major groups of Vertebrates, including fish, amphibians, reptiles, birds, and mammals, as well as in insects, crustaceans, and in a variety of microorganisms including the magnetotactic bacteria and protists. Electrophysiological studies in birds and fish indicate that the ophthalmic branch of the trigeminal nerve is one of the conduits of magnetic field information to the brain, implying that this pathway evolved before the last common ancestor of Vertebrates; more recent work suggests the vestibular system in birds might also be involved. Other work suggests magnetoreception might also arise in the retina, perhaps through optical effects on a protein like cryptochrome.
Despite this widespread presence in animals, the question of whether or not humans possess a remnant of these ancient magnetosensory systems has remained an extraordinarily controversial issue. Humans possess both biological magnetite, as well as cryptochromes, arguing that magnetoreception is a physical possibility. However, reports of magnetic fields influencing human orientation, made by a British group 30 years ago, failed the ‘acid test’ of Science when the initial orientation effects could not be repeated, thereby precluding tests of the magnetic influence. In contrast, more recent and as yet preliminary results from psychological experiments by the Kirschvink group support the hypothesis that some human subjects may have a subconscious remnant of this magnetic sensory system. Using our combined geophysical, psychological, and engineering skills, our goals for this HFSP project are to attempt verification, improve the power of the procedures, map the neural circuits underlying the sensitivity, and ultimately develop behavioral-neural techniques to bring human magnetoreception into conscious awareness. If successful, these tools could also be applied to test biophysical hypotheses for the underlying transduction mechanism(s). As a significant byproduct, the psychophysical and engineering techniques developed will lead to more interactive BMI (Brain Machine Interface). It will be applicable to any other cases where a human brain has potentially plastic sensory abilities but not revealed yet.

Pharmaceutical drugs represent one of the most powerful methods available for manipulating important aspects of our physiology. As such, their development has been a big driver for the advances in healthcare over the last century. Drugs have been used not only to treat diseases directly, but also as an experimental tool for researchers trying to understand how the body works.
Although the many benefits of pharmaceutical drugs are well known, so are their limitations. Drugs commonly have ‘side effects’ on aspects of our physiology that were previously healthy and, over repeated application of a drug, people often become desensitized, meaning that they require higher and higher doses to achieve the desired effects. One can view both of these limitations as a consequence of the failure of drugs to recapitulate the ‘natural’ modulation of physiology that generally keeps us healthy. Natural checks and balances on cell physiology act by regulating the activity of specific proteins in specific cells in a particular part of the body, while drugs often are much less selective in the nature and location of their activity. Similarly, while the activity of a natural process is modulated in ‘real time’, drug concentrations build up slowly and hang around for a long time. It is, of course, not time to throw the pharmaceutical baby out with the bath water, nevertheless alternative methods of manipulating physiology would be most welcome.
As many as a third of currently prescribed drugs target a specific mode of communication between cells that originates with a large family of proteins called G protein-coupled receptors (GPCRs). We propose developing a group of modified GPCRs that will be light sensitive. These proteins will turn ordinary cells into photoreceptors, allowing us to control their physiology simply by shining light on them. Because only those cells containing these engineered proteins will respond to light, and because its easy to switch light on and off rapidly (or to leave the light on for hours/days), this approach will allow scientists to approach a more ‘natural’ modulation of cell physiology. In the first instance these proteins will be available for researchers to use in cell cultures and/or animals to gain a better understanding of how the body works. Ultimately, we hope that they will be adapted for treating diseases in people.

Complex, multicellular organisms are made up of many cell types, each performing different functions. Understanding the evolution of cells' social interactions is hard: multicellularity evolved long ago and dissecting individual interactions is difficult. We therefore evolve social behavior in two microbes, E. coli and budding yeast. Comparing a prokaryote and a eukaryote will suggest hypotheses for how social interactions between cells evolved in plants and animals.
We will use sucrose utilization as a model to study social evolution. Sucrose hydrolysis releases one molecule of glucose and one of fructose, highly preferred carbon sources for many microbes. In yeast, the enzymes that hydrolyze sucrose are extracellular public goods, whereas in E. coli they are plasmid-encoded, private goods.
We will use genetic engineering, experimental evolution, theory, and simulation to investigate three aspects of social evolution: cooperation, the origin of multicellularity, and the division of labor. We have four aims:
1) Engineer and evolve differentiation within multicellular yeast aggregates to test the idea that undifferentiated, multicellular aggregates evolved before the division of labor.
2) Study the role of plasmid transfer in bacterial cooperation. Does the production of secreted public goods by plasmids select for the faster transfer of plasmids to neighboring cells?
3) Ask how prior history and initial conditions determine the trajectory of evolution. We will engineer yeast and bacteria so they can evolve from similar starting points. We will vary whether sucrose-harvesting systems are public or private, whether they are transmissible or chromosomal, and whether cells can differentiate.
4) We will compare our experiments with three different methods to model social evolution: analytical theory, numerical simulation, and evolving digital organisms within computers.
The proposed work will benefit from a) combining engineering and evolution as experimental approaches, b) comparing different approaches to model these problems, c) strong interaction between theory and experiment, d) the interplay between expertise in yeast and bacterial cell biology, and e) regular meetings to exchange results, critiques and ideas that will influence future research.

We propose to elucidate the basis for positional information by hormones during plant morphogenesis. While it is known that cell fate decisions require simultaneous input from multiple hormones, to-date a precise understanding of how these signals are coordinated and act together to drive morphogenesis does not exist. Our limited mechanistic understanding is largely due to the difficulty to quantify the distribution of these small molecules in space and time. To explore this fundamental question, we will exploit recent advances in synthetic biology to engineer an RNA-based biosensor platform applicable to a broad range of small molecules and in particular to hormones. Using live-imaging technologies, we will use the sensors to obtain quantitative dynamic 3D maps of hormone distributions and relate these maps to the spatio-temporal distribution of cell identities, both during normal morphogenesis and upon perturbations of hormone levels. This analysis will be done on the shoot apical meristem, one of the best-characterized developmental systems in higher plants. In this context, mathematical approaches will be essential to analyze and establish a predictive model for how multiple hormones influence cell fate in a spatio-temporal manner.

A bird that flies rapidly through dense foliage to land on a branch– as birds often do – engages in a veritable three-dimensional slalom, in which it has to continually dodge branches and leaves, and plan a collision-free path to the goal in real time. Here we aim to study how birds control the speed of their cruising flight, fly safely through narrow gaps, decelerate to a smooth landing, and how they transition seamlessly through these different flight modes. Research with insects indicates that there are neural programs for controlling specific features of flight, such as regulation of velocity and altitude, measuring distance flown, and homing. However, we do not know how these programs are organized to generate the complex flight trajectories that we observe in nature. There is an analogous deficit in our understanding of flight mechanics because most studies concern single modes of flight, such as take-off, cruising, hovering, or turning. We know little about how animals make transitions between flight modes. Our research will address these fundamental gaps in our understanding of animal flight by studying visual control of flight and wing kinematics in three bird species: two parrots, which are specialized for cruising flight (Budgerigars: Srinivasan; Lovebirds: Lentink) and a hummingbird, specialized for hovering flight (Altshuler). We focus on birds because they are visually agile, and can be trained to perform complex behaviors. The project will begin by characterizing the visual control of key flight modes: take-off, cruise, obstacle avoidance, and landing. We will then study transitions between these modes. The most effective way to study how vision is used to control flight involves the use of virtual reality environments. We will develop tools for instantaneously tracking the position and orientation of a bird’s head, and also tracking its wing position although not in real time. The head tracker will be combined with a video presentation system to create an arena in which what the bird sees can be manipulated relative to how it flies. The results from these experiments will have implications for the design of bio-inspired aircraft and of bird-friendly man-made structures, such as wind farms.

This project combines our collective expertise in developmental biology, neuroscience, engineering, sleep biology and in-vivo molecular imaging to create an innovative chick embryo model for advancing scientific understanding of the developmental origins of sleep, waking and brain rhythms in higher vertebrate animals. Using a novel extension of embryo microsurgical techniques, an accessible port to the embryo head through the eggshell and respiratory membranes will be made at 1/10 of embryonic development (day 2 of a 21-day incubation period), and microscopic devices engineered to measure and wirelessly report brain temperature will be surgically implanted in the brain ventricles of individual embryos, to learn when rhythmic changes in brain temperature first emerge. Eggs are then sealed until 16 days of incubation, when the embryo head will be accessed, and skull electrodes implanted. The electrodes are connected to a miniature electroencephalography (EEG) system specifically engineered for continuous wireless recording from the eggs. Individual eggs will also be attached to a device for recording embryo heartbeat and behavior from vibrations on the eggshell. During continuous EEG and behavioral recording, eggs will have periodic PET/CT images taken of their brains to relate EEG patterns to brain metabolic activity patterns using a novel dynamic imaging procedure with sub-millimeter spatial resolution and a temporal resolution of between 30 seconds and 5 minutes (CT images will be used to developmentally stage individual embryos). This information will allow us to identify characteristics reliably associated with different embryo brain states. This “phenotyping” procedure permits embryo brains to be collected when they either have been in a particular brain state for a certain period of time, or at the transition between brain states, greatly facilitating the histological and molecular characterization of the role of changes in immediate-early gene expression and neurotransmitter system function that previous work has implicated in the control of adult sleep states and brain state changes, and permitting manipulative experiments. This system will significantly accelerate the identification of the neural systems organizing the initial emergence of these important systems brain functions, and of the mechanisms through which they operate.

Oxygen, that supports all aerobic life, is abundant in the atmosphere because of its constant regeneration by photosynthetic water oxidation by green plants, algae, and cyanobacteria. This reaction is catalyzed by a Mn4CaO5 cluster associated with photosystem II (PSII), a multi-subunit membrane protein complex. Given the role of PSII in maintaining life in the biosphere and the future vision of a renewable energy economy, understanding the mechanism of the photosynthetic water oxidation is considered to be one of science’s grand challenges. Although the structure of PSII and the chemistry at the catalytic site have been studied intensively, an understanding of the atomic-scale chemistry from light absorption to water-oxidation requires a new approach beyond the conventional steady state X-ray crystallography and X-ray spectroscopy at cryogenic temperatures. Following the dynamic changes in the geometric and electronic structure of the Mn4CaO5 cluster and PSII at ambient conditions, while overcoming the severe X-ray damage to the redox active catalytic center, is key for deriving the reaction mechanism. The intense and ultra-short femtosecond (fs) X-ray pulses of the LCLS (Linac Coherent Light Source) X-ray free electron laser provide an opportunity to overcome the current limitations of room temperature data collection for biological samples at regular X-ray sources. The fs X-ray pulses make it possible to acquire the signal before the sample is destroyed, which realizes the light-induced snap-shot study proposed here. The objective of this proposal is to study the protein structure and dynamics of PSII, as well as the chemical structure and dynamics of the Mn4CaO5 cluster (charge, spin, and covalency) during the light-driven process of PSII to elucidate the mechanism by which water is oxidized to dioxygen. We will design and apply a full suite of time-resolved X-ray diffraction and X-ray absorption/emission spectroscopy methods to follow the reaction at room temperature, that will provide an unprecedented combination of correlated data between the protein, the co-factors, and the Mn4CaO5 cluster, all of which are necessary for a complete understanding of photosynthetic water oxidation.

This project aims to artificially assemble a functional ‘switch complex’ of the bacterial flagellar motor (BFM) using DNA self-assembly to build a structural scaffold. The switch complex is a ~4 mega Dalton protein superstructure that converts a flux of ions across a membrane into mechanical rotation at speeds of up to 1700 Hz and consists of a reverse gear that switches rotation direction in a handful of milliseconds, making the BFM the most sophisticated nanoscale rotary motor known.
Building biology on a molecular scale is a fundamentally new approach to the Biological Sciences, which adopts a ‘top-down’ approach where the subcomponents are teased out of complete systems rather than pieced together from the bottom up. This suits the study of biological systems, which conveniently self assemble in vivo, but also limits our ability to manipulate systems to understand them. The BFM is one such example which can only be assembled and functionally characterised in vivo. As a result, elucidating a molecular picture of its structure and function is extremely difficult or impossible, and proposed molecular models remain controversial and not conclusively tested. However, the BFM may also present an important opportunity to study biology from the bottom-up. It is one of the best-characterised large protein complexes and decades of study have given us a blue print outlining the requirements for building the switch complex, which include a circular nanoscale structural scaffold. Recent advances in DNA self-assembly make it possible to build such a scaffold, allowing the programmed construction of nanoscale 3D objects at near-atomic precision.
We will design and build DNA nanostructures as structural scaffolds to artificially assemble the BFM switch complex ex vivo. By controlling and manipulating the assembly process, we will exploit remarkable new opportunities to address long-standing questions about the BFM and large, dynamic protein complexes in general, including the elucidation of a complete atomic-scale picture of this ~4 mega Dalton protein superstructure.

Photosynthesis is essential for all life on Earth and has enormous environmental and agricultural impacts. The most advanced photosynthetic machinery is that of higher-plant chloroplasts, where it is hosted by one of the most complex membranous systems found in cells. Notably, this massive and elaborate system forms essentially from scratch, commencing with undifferentiated plastids that contain no or very little photosynthetic proteins or internal membranes. Equally remarkable, upon aging of leaves, seasonal changes, or the plant’s transition to the reproductive phase, this system, along with other components of the chloroplast, breaks down and is recycled in a highly regulated process reminiscent of programmed cell death. In addition to being of fundamental scientific interest, these processes bear an immediate agricultural relevance as they have a direct impact on biomass production, plant longevity, flowering, fruiting and seed production. Despite their importance, the processes underlying the biogenesis and breakdown of the photosynthetic machinery in higher-plant chloroplasts are largely elusive. This is because their study necessitates an integrative approach that combines structural and biophysical studies with high-throughput transcriptomic, proteomic and metabolomic analyses, all of which must be performed along the entire developmental pathway of the plastids, from incipient chloroplast differentiation to senescence. Here we propose to undertake such an interdisciplinary approach to comprehensively characterize the buildup and disintegration of the photosynthetic machinery in the chloroplasts of dicots – the largest group of flowering plants, which includes many agriculturally important plant species as well as most trees. These studies, along with subsequent systems analysis and reverse genetic approaches, are expected to significantly advance our knowledge of the elementary processes that underlie differentiation and senescence of higher-plant chloroplasts and, in this way, the construction and dismantling of the most efficient energy conversion device that exists in nature. In addition, they will aid future in silico modeling approaches, as well as efforts to improve photosynthesis by breeding or genetic engineering.

This project takes a multidisciplinary quantitative approach to investigate the consequences of maximal growth rates. It draws on a range of cutting-edge approaches, from in vitro biochemistry to evolution and high-throughput characterization. It will also be guided by mathematics at every step. In addition to the insights into cell physiology, we expect to set a new world record in rapid growth of biomass.

Sophisticated models exist for the evolutionary pathways by which proteins have evolved in Nature over billions of years to form an impressive diversity of structure and to carry out many functions with unrivalled efficiency. Directed protein evolution in the test tube can emulate natural evolution, but is often limited by low hit rates and small improvements during evolution cycles. Burdened by their evolutionary history, proteins often show low evolvability, due to loss-of-function and/or structure or epistatic ratchets (ie deleterious or permissive mutations that only become beneficial upon occurrence of further mutations). Given the danger of getting lost or stuck in sequence space, the question arises how natural and forced evolution avoids dead ends. We will create a mathematical model to simulate the effects of mutational robustness on evolvability using parameters that can be adjusted to experimentally determined values for neutrality and accessibility of novel phenotypes. Based on experimental measurements of the dynamics of fitness landscapes of a protein population during evolution cycles we aim at insight into the fundamental as well as an effective basis for protein engineering by evolution. Investigations of duplicated specialized proteins (paralogs) and their ancestral precursors showed the latter often to be promiscuous (i.e. to have multiple functions) and more highly evolvable, suggesting that evolution may proceed by shifting the relative balance of two functions along paths of successive point mutations. We probe the evolution dynamics of promiscuous members of the metallo-ß-lactamase and alkaline phosphatase superfamilies. Ancestral sequences will be computed from alignments and databases as starting points for evolution to give a dynamic experimental picture of fitness of a protein population in response to the ancestor sequence and to selective pressure. Such distribution functions will be gathered by ultrahigh-throughput screening of >10^6 clones in picolitre droplet compartments during in vitro and in vivo evolution cycles, followed by systematic characterization of multiple fitness traits such as expression level, stability or reaction kinetics for multiple activities. These approaches will enable a dynamic picture of protein evolution at various levels from transition state recognition, protein biophysics to population genetics.

We address a key question in developmental biology: how does an organism reach its final size and shape, in the face of stochastic variation at the cellular level? Previous research has focused on mutants and conditions that affect the global size and shape of organs, enabling the discovery of a large number of regulators. However, most analyses have considered only average cell behaviors, overlooking local heterogeneity and stochastic variation. In this context, how the final size and shape of an organism are determined remains poorly understood.
Here we adopt an orthogonal strategy by taking advantage of the observation that many organisms have a remarkably consistent shape, yet at the cellular level, cell growth and shape can be highly variable. Our goal is to resolve this apparent contradiction between the cellular and organismal levels. We propose that cell-cell communication mechanisms coordinate cell stochastic variability so as to yield consistent organs. Since growth remains locally heterogeneous, these mechanisms may, counter-intuitively, maintain or even enhance local heterogeneity.
We will test this hypothesis in Arabidopsis thaliana, as it produces a large number of almost identical flowers with stereotyped floral organs. We choose to work on the sepal, which is accessible for live imaging and mechanical measurements. Using a combination of experimental and theoretical approaches, we will identify mutants with variable sepal size and shape. Because of their role in coordination, the corresponding genes are likely to include the hormonal/nutrient pathways and mechanical sensing and response. As plant growth is driven by the internal turgor pressure and restrained by cell walls, we will assess the activity of these growth regulators by quantifying the mechanics of cells in these mutants and in wild type to causally link gene activity and shape. We will use data-driven approach and information theory to extract the underlying patterns of correlations between cells, notably to qualify the most plausible coordinating mechanisms that regulate cell stochasticity. These hypotheses will be tested by combining a mechanical model of sepal growth that incorporates identified signals and local experimental perturbations, such as cell ablation, force application, and genetic mosaics.

A challenging issue for bacterial pathogens is the need to modulate host cells behavior to allow their own successful propagation. Intracellular pathogens secrete a large number of proteins (“effectors”) into their host cells to either trigger entry, escape cellular defense, modulate cellular responses or interact with intracellular components that are critical for infection. We hypothesize that pathogens also secrete small RNAs that interact with host cellular mRNAs and non-coding RNAs, thus modulating cellular activity for their own benefit. Discovery of such RNAs among the vast amount of cellular RNAs is highly challenging due to lack of proper techniques; however, the emergence of ultra-highthroughput sequencing and computational inference (Sorek’s expertise), in combination with single-cell fluorescence Imaging techniques (Palmer) and extensive expertise in pathogenesis and cell biology (Cossart) will allow us to address this challenge. Our aim is to identify secreted RNAs using the model intracellular pathogen Listeria monocytogenes, and understand their roles in pathogenicity. Although scientifically 'risky', if successful these studies could lead to a major breakthrough in the fields of bacterial pathogenicity and RNA biology.