Because of its ancient origins, little is known about the first steps in the transition to multicellularity. Perhaps the largest gap in our knowledge concerns the origin of cellular differentiation. Our proposed research will examine how aging (a nearly-universal feature of cellular life) can be co-opted by simple multicellular organisms to drive the emergence of novel multicellular development. The proposal combines the PIs’ complementary expertise in aging and protein homeostasis (Saarikangas) and evolutionary biology (Ratcliff). Our research uses a unique approach: we evolve multicellularity from a single-celled yeast in the lab, examining how, over thousands of generations of evolution, they become more complex. Following our preliminary results, we focus on the role of aging and the chaperone protein Hsp90, mechanistically examining their role in the origin of simple cellular differentiation in our model system. We contextualize these results with mathematical modeling, examining the potential for age-dependent development to arise in different lineages of multicellular organisms. By showing how multicellular development can arise de novo, our research illuminates a problem of profound biological importance, and removes one of the last conceptual hurdles remining in our understanding of the origin of multicellular life.

Biomolecules in cells can self-organize into immiscible fluid droplets via liquid-liquid phase separation (LLPS). These liquid-like domains act as organizational hubs and play important physiological functions, including facilitation of biochemical reactions, transport, and signaling. Alteration of their size, internal structure, and coalescence dynamics have been linked to neurodegenerative diseases, including Alzheimer’s, Parkinson’s, and ALS. Despite the continuous discoveries of new liquid-like compartments in cells, a biophysical understanding of their formation and resulting properties is still missing. Here, I will describe experiments designed to study: i) how intermolecular interactions control the assembly, material properties such as viscosity and surface tension, and dynamical arrest of liquid-like droplets; and ii) how non-equilibrium processes in cells, such as mechanical stresses exerted by the cytoskeleton, affect LLPS and regulate the size distribution and fluidity of the liquid domains. I will develop an in vitro system based on phase separating mixtures of DNA oligomers suspended in an active cytoskeletal network composed of purified filamentous proteins that will model the liquid condensates and cell interior. DNA oligomers are an ideal system to relate molecular structure to phase behavior because of the specificity and determinism of the base pairing interactions. Active cytoskeletal networks will mimic the non-equilibrium features of the cell interior. By exploiting the unique features of these systems, I will build a biophysical framework to decipher the roles of non-equilibrium pathways to LLPS and irreversible protein aggregation in living cells.

The ribosome which functionally establishes all proteomes across all life kingdoms, has historically been considered as a homogenous molecular machine that only passively contributes to gene expression. In contrast, there is growing evidence that ribosomes are heterogeneous and actively regulate gene expression through selective mRNA translation. In humans, there are hundreds of ribosomal DNA (rDNA) copies spread across multiple chromosomal loci, with recent studies showing copy number and sequence variations with unknown effects. In this project, I will characterize for the first time how variant rDNA alleles give rise to heterogeneous ribosomes that contribute to human physiology and disease. Using full genome sequencing data for 120,000 healthy and diseased individuals, I will (1) map population-wide rDNA variability in sequence and copy number variation and (2) identify rDNA variation that predicts human disorders. To get to a mechanistic understanding of rDNA variability I will (3) employ reporter assays to characterize the impact of ribosomal RNA (rRNA) variation on ribosome functionality and (4) use super resolution microscopy to test if rRNAs display expression specificity within a tissue and between different tissues including cancer samples. rDNA sequence variants reflect a hidden facet to our genomes and human health with the potential to change our understanding of cellular physiology at a global populations level. By studying the genetic variability of human rDNAs and its role in health and disease we aim to find how this diversity controls gene expression at the translation level and detect disease associated variants for early disease prognostics.

The regeneration of the optic nerve in the adult mammalian brain is limited [1,2]. During embryonic development different guidance molecules direct retinal ganglion cell (RGC) axons from the retina to the dorsal lateral geniculate nucleus (dLGN) [3]. Within the dLGN certain receptor/ligand gradients [4,5] and spontaneous patterned neural activity [6] assure correct retinotopic mapping. However, little is known on how these developmental mechanisms could be exploited to functionally reconnect RGC axons to the adult brain. Thus, we will ask: 1) How could axons of in vitro grown RGCs reinnervate and functionally reconnect to the dLGN in the adult rodent? 2) How does the axonal innervation pattern in the adult dLGN adapt in response to stimulation or genetic modification of the RGCs? To tackle these challenging questions, we will build a novel biohybrid platform in which RGCs are cultured on an implantable stretchable multi-chamber multielectrode array (MEA) [7] that is connected to a bioabsorbable axon-guiding nanofiber tube. In contrast to traditional organoid or explant based axonal rewiring approaches the new biohybrid system offers three unique advantages: First, we can modulate gene expression in a spatially defined subpopulation of RGCs. Second, we can spatiotemporally control the activity patterns of the RGCs. Third, we can target axons to deep brain regions and stimulate at an almost single cell resolution. This is significantly better than the mm3 size stimulation of any conventional deep brain stimulation electrode and has the potential to restore functional vision in blind patients without eyes.

A number of transformative molecular tools for biomedical research and therapeutics have been developed. However, a lack of safe, efficient delivery methods has precluded the widespread therapeutic application of these tools. One promising route to overcoming this limitation is to investigate the diverse biomolecule exchange systems found in nature and develop delivery tools based on these systems. Recently, a novel membrane vesicle (MV)-mediated horizontal gene transfer system was found in archaea. In this system, a relatively large 50 kbp plasmid is enclosed in a MV and delivered to other archaea. These plasmid vesicles (PVs) use a viral-like plasmid, pR1SE, that encodes the proteins required to assemble PVs and transfer the plasmid to other cells. As intercellular communication by MV is a universal process among the three domains of life, these systems may be exploited to deliver tools in human cells or the human microbiome. However, the mechanisms of archaeal MV and PV formation are largely unknown. I propose to identify the key proteins in the PV system. I will narrow down the candidates by remote homology detection among pR1SE and related viruses, and validate them by biochemical experiments. Furthermore, I will use genome-wide CRISPR screening to identify key proteins of host archaeal cells. Ultimately, I aim to reconstruct the PV system towards programmable synthetic vesicles production. This proposal will elucidate basic archaeal biology and provide a hint to the origin of viruses with a long-term objective of harnessing archaeal vesicle systems for delivery tools. Finally, this approach will serve as a paradigm for engineering archaea for molecular technologies.

Some animals can regenerate lost organs and tissues with high efficiency. In many such species, cellular dedifferentiation plays a major role in this process. How dedifferentiation is induced in vivo in the correct context is unknown and represents a major question in regenerative biology. The cnidarian Hydractinia can regenerate any lost body part in a tissue-specific manner. Regeneration of some body parts is mediated by adult, migratory stem cells called i-cells that are absent from the intact head. Surprisingly, host lab found that isolated heads can, nevertheless, regenerate a fully functional animal including i-cells that arise by dedifferentiation. Following amputation, isolated heads transiently display features consistent with cellular senescence. Here, I propose to study the mechanisms accompanying injury-induced senescence in Hydractinia dedifferentiation-dependent whole-body regeneration. My approach includes the mechanistic characterization of the molecular/cellular events that occur between decapitation and the appearance of new i-cells through dedifferentiation. For this, I will perform transcriptomic analyses, gene function experiments, and in vivo imaging. The strength of my proposal rests upon the usage of a unique animal model that is highly regenerative, displays natural dedifferentiation, and allows performing in vivo experiments that are impossible to conduct in other animals.

Many bacteria have evolved protein machineries to inject proteins into eukaryotic or prokaryotic cells. The type III secretion system (T3SS), widely distributed among gram-negative bacteria, is made up of a needle complex, an export apparatus and a sorting platform (SP). Although, the whole injectosome was characterized in a recent study, the assembly process of the SP is poorly understood. It has been shown that SpaO (a core component of the SP) localizes in two spatial populations, one at the bacterial membrane and another one located within the bacterial cytosol. This observation of the presence of SpaO-associated clusters away from the bacterial membrane suggests the presence of fully or partially assembled SPs in the bacterial cytoplasm. Furthermore, there are two existing versions of SpaO, a short (SpaOS) and a long (SpaOL) version. It has been postulated that SpaOS plays a role in chaperoning SpaOL during the assembly process of the SP but is not a structural part of the SP itself. If this holds true, SpaOS should only be localized in the (partly assembled) SPs located in the bacterial cytosol. In order to understand the assembly of the SP and the role of SpaOS in this process, I will image all proteins of the SP using recently developed 4Pi microscopy together with DNA-PAINT super-resolution microscopy. This novel combination, will enable me to image all proteins of the SP with molecular resolution (5 nm isotropic resolution). This will be an ambitious step towards understanding the sequence of events leading to the assembly of the SP in the T3SS. I anticipate that the tools developed for this work will bring new technologies to the field of bacterial physiology.

Identical genomes in social insects generate castes with distinct morphologies and behaviors. The Harpegnathos ants have a remarkable lifestyle, whereby the queen’s death triggers transition of a subset of workers to become her substitutes, called gamergates. This entails minor changes in morphology, yet a pronounced shift in behavior, and remarkably, a fivefold extension of lifespan. This process is reversible if the gamergate is placed in a colony with a real queen. The capability of such a dramatic transition in adult life provides a unique context to study neuronal plasticity and aging. Moreover, social insects challenge major aging theories as their calorie intake and fecundity positively correlate with lifespan, in contrast to most animals. I aim to provide an extensive mechanistic description of the worker-to-gamergate transition in H. saltator, and to characterize the mechanisms of the resultant aging deceleration. I will study the brain as it regulates caste transition and produces hormones implicated in aging, while also being affected by both processes. First, I will perform single-cell RNA-sequencing of the worker and nascent gamergate brains to describe changes in genes controlling synaptic wiring and cellular signaling. Next, I will test whether deceleration of aging in gamergates is accompanied by decreased rates of molecular damage, protein synthesis, transposon activity, and telomere shortening. Finally, I will combine both datasets to link transcriptional changes during the caste transition with aging-related molecular processes. Deciphering the paradoxical aging mechanism in social insects may uncover previously unknown strategies to prolong lifespan.

Glycerophospholipids are one of the major lipid species in cellular membranes of eukaryotic cells. Among glycerophospholipids, phosphatidylethanolamine (PE) is a highly conserved lipid, found in all organelles but mainly synthesized in the endoplasmic reticulum (ER) and in mitochondria. If it has been shown to play essential roles in membrane protein structure and function, its distribution and metabolism have not been fully elucidated. Indeed, little is known about lipid transport in terms of quantitative and qualitative contribution of each pathway, because methods to measure lipid transport between organelles in cells are lacking. Unlike proteins which are relatively easily manipulated by genetic techniques, individual lipid species cannot be readily modified in vivo. Here, I propose to develop and use chemical tools to study PE transport and metabolism in living cells with high spatiotemporal resolution. I will synthesize photocaged PE which can be localized with precision to mitochondria (or ER) within cells and then released with irradiation of light. Using isotope-labelling, I will trace their transport and quantify the metabolites inside living cells by lipidomics and mass spectrometry approaches. Moreover, mass spectrometry imaging could allow visualizing the de novo labeled phospholipids. This interdisciplinary approach will give us the spatiotemporal resolution that is needed to create for the first time a 4D map of PE metabolism in cells. It will allow us to elucidate its metabolism related to its dependence on intracellular localization and prove that PE can or cannot be transported, for instance, bidirectionally between ER and mitochondria.

Touch the sensitive plant Mimosa pudica and it rapidly changes shape, folding its leaves to deter herbivory and dislodge pests. This and other feats of rapid plant motion have fascinated scientists for centuries, but the physiological mechanisms underlying them remain poorly understood. I propose a collaborative, interdisciplinary project to elucidate two key aspects of rapid plant motion: sensation of mechanical stimuli and rapid generation of force. I hypothesize that knowledge of the sensory and motor strategies of motile animals can inform our approach to the study of rapid plant motion, and outline a novel experimental approach that draws tools and theory from the fields of biomechanics, neurophysiology, and materials science to explore rapid motion in Mimosa pudica, a model system for the study of rapid plant movement. I have identified an ideal host laboratory in which to carry out this work situated at a nexus of expertise in mechanotransduction, plant physiology, and mechanical modeling. To this environment I will bring expertise in skeletal muscle physiology and the biomechanics of force production. The proposed work forges a connection between seemingly disparate areas of physiological research, will provide unique training that increases my technical and theoretical breadth, and has great potential to elucidate basic principles underlying sensory and motor strategies in biological systems.

Protein degradation systems maintain protein homeostasis. A failure of these systems causes various diseases, such as neurodegenerative diseases and cancers. In eukaryotic cells, ubiquitin is a general marker for selective degradation and determines protein lifetime in vivo. In selective degradation, E3 ubiquitin ligases determine target proteins. Although there are ~800 E3 ligases in human, only a handful of them have already shown to recognize specific short peptide motifs called “degrons”. Moreover, the feature(s) for unstable or misfolded structures recognized by E3 ligases remain unclear. In part, this is because we lack a comprehensive approach to investigate the global relationship between protein structural stability, ubiquitination status, and lifetime of the protein. To reveal the effect of protein folding stability (and other features) on ubiquitination and biological lifetime in vivo, I propose to measure these parameters for thousands of designed mini-proteins, whose folding stability has been previously characterized in detail. First, I will measure biological lifetime for these mini-proteins by flow cytometry, and monitor their ubiquitination status by using top-down proteomics approach. Then, I will analyze these data by using in silico analysis and decipher what factor(s) determine ubiquitination states and biological lifetime. This highly innovative and comprehensive approach using thousands of designed proteins will allow me to uncover the fundamental principle for protein lifetime in vivo and provide a mechanistic basis for designing better tools to manipulate protein lifetime.

During gastrulation, embryonic stem cells are primed towards three germ layers: the ectoderm, the mesoderm and the endoderm. While acquiring their fate, the cells re-arrange massively and the embryo polarizes with the appearance of an antero-posterior axis. The signals that are necessary for this patterning and the role of the mechanical forces remain poorly understood, in particular for the human embryo. Synthetic biology has thus developed new models of embryos using human stem cell lines (hESCs). We propose to study the effect of differential mechanical tension on the appearance of a primitive streak in a synthetic human embryo. We will investigate the relationship between the WNT pathway activation and mechanical instability, following recent results on a biomechanical regulation of this pathway. We will first work on the correlation between the substrate induced curvature and WNT expression in 2D cultures of human stem cells. Secondly, we will map the resulting stress field of the cells to this expression pattern. Thirdly, we will study cellular self-organization under mechanical load in a 3D model synthetic embryo. We will eventually focus on the inducer role of a second cell population in the polarization of this synthetic embryo. In the four cases we will correlate the resulting dynamic features with the expression of markers relative to the three germ layers. We hope to provide a first quantitative understanding of the relationship between the force field and the differentiation of stem cells in a model synthetic embryo. This project will be co-supervised by Dr. Eric Siggia and Dr. Ali Brivanlou. The experiments will be performed in the Brivanlou lab.

Metabolic transformation is a key characteristic of cancer cells and considered a promising avenue for therapeutic intervention. Nevertheless, metabolic lethalities identified in laboratories often fail to translate to the clinic. This may be because the complexities of tumor metabolism cannot be recapitulated by studying cancer cell autonomous metabolism alone: tumors are not solid masses of cancer cells, but include host tissue, stroma, and immune cells that together shape the tumor metabolic environment. This insight highlights the need to observe and interrogate tumor metabolism in vivo. Metabolic studies rely heavily on mass spectrometry-based analysis of metabolites extracted from cells or tissues, but this approach lacks spatial resolution to resolve metabolic heterogeneity between cells. To address this unmet need, I will develop a transformative platform for single-cell tissue metabolomics using mass spectrometry imaging and use it to study redox metabolism in liver metastatic cancer to investigate a major unanswered question in cancer research: how does the host tissue support metastatic tumor growth? This work contributes a new strategy to incorporate spatial resolution into metabolomics and provides the first comprehensive analysis of the metabolic symbiosis between host and cancer cells in vivo. It aims to redefine the traditional approach to metabolic studies that can progress our understanding of metabolic heterogeneity. Focusing on liver metastatic cancer, I aim to provide novel mechanistic understanding of redox-metabolic crosstalk between cells in the tumor-host niche and explore strategies to interfere with redox metabolism to combat metastasis.

Living organisms constantly make important decisions in their dynamic environments in order to survive. Animals make effective decisions in response to sensory stimuli by recruiting both innate and learned information in their nervous systems, thereby guiding them to approach or avoid cues that are appetitive or aversive respectively. Importantly, in order to adapt to a changing environment, animals also need to update their memories over time. Understanding how the nervous system invokes previously stored information to dictate future decisions has been a central and historical goal in neuroscience. Recently, significant progress has been made in identifying the connectivity and functionality of specific circuits involved in distinct aspects of learning and memory. However, a challenging task has been to obtain a comprehensive and systems-level understanding of neural circuit dynamics by which memories are retained, extinguished or expanded and integrated with innate information to eventually drive various behavioural actions. In this proposal I aim to address this problem by using a novel setup to record the activities of multiple individual neurons that function at different levels downstream of the learning and memory centre in the insect brain in freely behaving animals subject to specific tasks over time. Importantly, by combining this data with synaptic-level circuit maps generated by electron microscopy, mathematical modelling to produce reliable circuit motifs, and precise optogenetic manipulations to test the models in vivo, we will obtain both a global and high-resolution understanding of memory-based and context-dependent action-selection.

Epigenetic modifications regulate development. In particular, genome-wide DNA methylation patterns undergo dynamic changes at cornerstones of mammalian development. While DNA methylation is essential for normal embryogenesis, how it regulates and maintains cell fate and function in vivo remains unclear. A key methylation-regulated developmental process is parental imprinting and perturbations to methylation imprints result in embryonic lethality. We discovered that paternal deletion of an essential imprinting control region in mammals results in a temporal-specific compensatory methylation switch onto the opposite maternal allele, minimising detrimental developmental defects. This suggests a novel paradigm of epigenetic plasticity with wider implications for the epigenetic control of genome function. Here, we will utilize cutting-edge DNA and RNA single-cell sequencing methods and in vivo knockout models to study the functional roles of parental-origin specific epigenetic plasticity during development. We will determine whether loss of imprinting drives compensating changes on the other allele and decipher the associated mechanism of regulation, maintenance and hierarchical interactions. We will investigate the genome-wide plasticity of imprinting by relating imprinted gene dosage dynamics with developmental time points and tissue expression, with specific focus on particular developmental niches. Our approach will elucidate a new avenue of epigenetic control during development and disease, contribute to our understanding of the evolution and flexibility of imprinting and provide novel insights into regenerative medicine, reproduction and epigenetic inheritance.

Sleep is widely acknowledged to be essential to brain function, including learning and memory. Although learning relies on synaptic plasticity, there is surprisingly no consensus on the nature and function of the synaptic changes sleep would mediate. The controversy in the field partly derives from the complexity of sleep structure, as sleep is composed of REM and non-REM stages suggested to hold different functions that are poorly investigated at the synaptic level. To address this issue, I propose to use a recently-developed 3D two-photon microscope in order to continuous monitor synaptic changes during learning and specific sleep stages. The mechanisms underlying sleep stage-dependent synaptic plasticity will be further investigated by combining Ca2+ imaging, light-induced CaMKII inhibitor and optogenetics. The data obtained will provide new insights into the role of sleep in synaptic plasticity, and might further reveal fundamental mechanism underlying learning and memory.

The regeneration of the adult skeletal muscle relies on the activity of resident muscle stem cells (satellite cells). These PAX7-positive (PAX7+) cells are predominantly quiescent in uninjured muscles, while being able to rapidly proliferate upon injury. Mouse satellite cells are known to be heterogeneous, with subpopulations exhibiting graded quiescent states and variable regenerative potential. However, the molecular identities and developmental trajectories of satellite cell subsets are not well established. Moreover, due to the limited access to human tissues, little is known about the biology of human PAX7+ cells. The host laboratory has recently developed robust differentiation protocols that enable efficient generation of myofibers and PAX7+ cells from pluripotent stem cells in vitro. In this project, I plan to use single-cell transcriptomics to chart the dynamic cellular landscape of human in vitro myogenic differentiation, as well as mouse muscle development in vivo, with the aim of revealing the identities of emerging cells during myogenesis, developmental roadmap for cellular heterogeneity and molecular signatures of quiescence. In addition, preliminary results implicate the retinoic acid (RA) signaling in controlling satellite cell quiescence. Assisted by single-cell profiling, the role of RA will be characterized in detail. To further uncover in an unbiased fashion the gene networks linked to quiescence, I will also conduct functional genomic screens based on the in vitro platform. Overall, this study will use diverse techniques and model systems to provide new insights into muscle stem cell biology and potentially benefit the treatment for muscle diseases.

Phenotypic evolution concurs with spatial and temporal gene expression changes. Most work have been focused on spatial gene expression variations, yet little has been known about whether the temporal alteration of gene expression, termed heterochrony, also contributes to morphological variations. In this proposed work, I aim to address this question using forelimb (FL) and hindlimb (HL) development in the mouse embryo as a model, due to the ease of experimental manipulation of this system. Although gene regulatory network underlying limb development is highly conserved, the morphology of FL and HL differs. An important distinction in their development is the relative duration of gene expression time – nearly all limb regulators express for a shorter period in HL, correlates with its curtailed differentiation time-scale comparing to FL. Whether and to what extent do such temporal gene expression differences relate to morphological divergence remains unclear. To this end, I will systematically and quantitatively characterize the kinetics of key limb regulators’ expression during the FL and HL development in the mouse embryo; then combining in vitro culture assays and sequencing methods to investigate the mechanisms causing temporal gene expression differences between the two appendages; finally, I will functionally assess the role of such heterochrony to limb morphology by altering the gene expression timing, first experimentally, in mouse and chick embryos; and then theoretically, by using mathematical modelling. Together, this proposed work can advance our understanding of the connection from genotype to phenotype and will also aid biomedical research on limb defects.

Keen observers of nature have often wondered why diverse species seem to behave similarly. For example, different species of Hawaiian spiders spin similar web architectures, diverse anoles lizard species bob their heads with the same styles and speeds, and distinct species of damselflies avoid predators using the same techniques. These and many other striking examples are thought to represent the convergent evolution of behaviors. Does this convergence reflect changes in the same genes or does evolution act through many genetic routes to create the same behaviors? Genetic differences clearly play a role in behavioral variation, but it remains challenging to identify the genes that underlie evolution of behaviors. However, convergence in the genetic basis of developmental or physiological traits has been discovered with many specific examples of genes and mechanisms, so it is possible to use studies of convergence to discover how behaviors evolve. Therefore, we will use a powerful comparative system to discover the genes and molecular mechanisms that underlie convergent evolution of behaviors for the first time across divergent animals.
The Caenorhabditis nematodes offer a unique experimental platform to connect behavioral differences to genetic differences. Starting with the keystone model organism, C. elegans, and existing data, we will characterize and classify genetic differences across wild isolates from three species of Caenorhabditis - C. briggsae, C. elegans, and C. tropicalis. Whole-genome genotype data combined with high-throughput, high-content imaging of behaviors from these same wild isolates will be input into unsupervised machine learning algorithms to create a high-resolution genotype-phenotype map for a range of natural behaviors and examples of convergence. This map will be queried for signatures of shared genetic changes at orthologous genes to identify which variants are most important evolutionarily. The result will provide the first systematic glimpse into the genomic “knobs” that control behaviors at single-variant resolution across species and insights into the repeatability of the evolution of behaviors.

Sea urchins are marine animals genetically close to the vertebrate lineage. Being eye-less and lacking a central nervous system (NS), these animals instead feature dermal photoreceptors dispersed over their spherical body surface and feeding into a decentralized NS. However, sea urchins can visually resolve objects and move towards them, and they can detect looming visual stimuli from any direction and accurately point their spines towards them. Such performance is normally associated with proper eyes feeding information into a brain. Sea urchins thus offer access to a unique visual system of a type that to date has not been studied in terms of its information processing. This alternative solution to vision may also have potential biomimetic applications for robotic miniaturization, smart probes, and intelligent materials where dispersed light detectors control the properties of the material.
The core of the proposed project is to investigate and model the neural mechanisms of information processing, which enables sea urchins to perform spherical vision by deploying an obviously very different mechanism from today's technology, and also very different from visually guided behavior in most other animals. Our study includes molecular and morphological identification of cell types, measurements of behavioral responses and electrophysiological photoreceptor responses, mapping the connectome of sea urchin photoreceptors and NS, and theoretical modelling of the information processing underlying visually guided behavior. We will map the connectomics of the NS and record the activity from key positions in the processing of visual information and generation of locomotory responses. The data will be used for computational modelling of the entire process from visual input to motor control. Special focus will be given to behavioral decisions where small changes in stimuli cause behavioral switches. We will also use genetic approaches to test the agreement between theoretical models and actual behavior.