Claude Desplan - US grants

Several morphogenetic gradients have been shown to pattern the early Drosophila embryo. Bicoid patterns head, thorax and abdomen through the specific activation of zygotic genes that direct proper development along the antero-posterior axis. However at least two other systems hunchback (hb) and caudal (cad), appear to also play important roles in the embryo. These two genes may represent the ancestral patterning system of insects whose function has been partially taken over by Bicoid in Drosophila.

One of the problems in uncovering the role of these genes is the high level of redundancy and of cross-interactions among the different maternal systems. We propose to analyze the morphogenetic properties of each system in an embryo that is fully responsive to patterning by these molecules, but where all potential interactions/redundancy with the other maternal systems have been eliminated. Construction of this 'embryonic test tube' requires that we genetically remove all maternal patterning systems (by generating flies whose germline is mutant for bicoid, hunchback, caudal, nanos, the terminal and dorso-ventral pathways). In this background, we will provide artificial gradients for each morphogen to be tested. We will first focus our attention on the role of Bicoid, Hunchback and Caudal. Because of the cross-interactions among these systems, it is still unclear whether Bcd has any role on his own, and what the morphogenetic roles of Hb and Cad are. We will also test the ability of Nanos and the terminal system to pattern the embryo. As an extension of this work, we will test the role other molecules that have been proposed to act as morphogens, in particular the gap genes products as well as secreted molecules such as Dpp, Wg and Hh. The function of these molecules has so far been analyzed in situations where an extensive pre-pattern already exists.

This work will provide insights into the direct role of critical patterning molecules on cellular events that lead to the determination of positional values in the embryo or, more generally, in other tissues.

DESCRIPTION (provided by applicant): The major task of the visual system is to build a spatial map of the environment: A precise retinotopic map projects an image of the retina onto the optic lobes. We are concerned with the qualitative map built by the Drosophila visual system for color vision and recognition of the vector of polarized light. The rhodopsins expressed in the inner photoreceptors (PR's) R7 and R8 defines functionally distinct ommatidia that are involved in color vision. To address how the visual system integrates information about the quality of light, we offer the following aims: 1. Mutual exclusion of rhodopsins: We will study how Rhodopsin proteins are involved in the process of excluding expression of other rh's in the same cell and what the downstream events are. This function represents a paradigm that might be of interest for other sensory systems that use similar types of receptors. 2. Signaling pathway between R7 and R8: We have isolated a series of mutants that affect the communication between R7 and R8 in opposite ways without affecting R7 rh expression. We will characterize this cascade of genes genetically and molecularly, both in the adult retina and in the larval eye that represents a highly simplified visual system. 3. A behavior assay for color vision: We will develop a behavior assay for color vision. We will use our collection of mutants and appropriate genetic tools to evaluate the function of the different subclasses of ommatidia and analyze new mutants that affect color processing in the medulla. Finally, we will test the role of the two classes of PR's in the larval eye to entrain the circadian clock at different times of day. 4. A new screen to identify mutations affecting rh expression: We have designed an insertional mutagenesis recessive mosaic screen based on a new universal transposable element piggyBac with a different insertional specificity than P-elements. This screen will aid us in cloning genes from the previous screen as well as for identifying new mutants. This system represents a simple paradigm to a basic problem of sensory perception using a very amenable genetic system that has so far not been used for the study of color vision. As the nocturnal mouse model is not appropriate for studying color vision, this places Drosophila in a unique position for such experimental manipulations.

DESCRIPTION (Adapted from applicant's abstract): Color discrimination is performed in the Drosophila retina by the coupling between the inner photoreceptors R7 and R8. Two distinct sets of ommatidia, which are distributed stochastically throughout the retina, are characterized by the specific expression of distinct rhodopsin genes (rh). The p ommatidia are characterized by the expression of rh3 in R7 and rh5 in R8, while the y ommatidia have rh4 in R7 and rh6 in R8. Expression of rh3 and rh4 in R7 is mutually exclusive, and the coupling between R7 and R8 rh's is absolute. The applicants are investigating the control of this exclusion and the factors that contribute to coordinate expressions of rh's. The specific aims of the proposed experiments are: (1) that Pax-6, a gene that plays a critical role in eye development in all species, is an essential regulator of rh gene expression and that the ancestral role of Pax-6 was as a regulator of late photoreceptor determination and the control of rh expression; (2) regulatory sequences will be sought that determine the correct expression of rh's in the particular subtypes of R7 and R; (3) factors that regulate rh expression will be identified using the yeast one-hybrid system, and the genes encoding these factors, in particular the orthodenticle gene, will be characterized genetically; and (4) an enhancer trap screen that relies on specific expression patterns in subsets of inner photoreceptors will be used to identify genes important for proper retinal patterning and color vision. This will allow us to understand how patterning of sensory receptor exclusive expression is achieved, addressing a problem that is common to other sensory systems.

Several morphogenetic gradients have been shown to pattern the early Drosophila embryo. Bicoid patterns head, thorax and abdomen through the specific activation of zygotic genes that direct proper development along the antero-posterior axis. However at least two other systems hunchback (hb) and caudal (cad), appear to also play important roles in the embryo. These two genes may represent the ancestral patterning system of insects whose function has been partially taken over by Bicoid in Drosophila.

One of the problems in uncovering the role of these genes is the high level of redundancy and of cross-interactions among the different maternal systems. We propose to analyze the morphogenetic properties of each system in an embryo that is fully responsive to patterning by these molecules, but where all potential interactions/redundancy with the other maternal systems have been eliminated. Construction of this 'embryonic test tube' requires that we genetically remove all maternal patterning systems (by generating flies whose germline is mutant for bicoid, hunchback, caudal, nanos, the terminal and dorso-ventral pathways). In this background, we will provide artificial gradients for each morphogen to be tested. We will first focus our attention on the role of Bicoid, Hunchback and Caudal. Because of the cross-interactions among these systems, it is still unclear whether Bcd has any role on his own, and what the morphogenetic roles of Hb and Cad are. We will also test the ability of Nanos and the terminal system to pattern the embryo. As an extension of this work, we will test the role other molecules that have been proposed to act as morphogens, in particular the gap genes products as well as secreted molecules such as Dpp, Wg and Hh. The function of these molecules has so far been analyzed in situations where an extensive pre-pattern already exists.

This work will provide insights into the direct role of critical patterning molecules on cellular events that lead to the determination of positional values in the embryo or, more generally, in other tissues.

DESCRIPTION (provided by applicant): This proposal is a renewal of 5R01EY013012-12 Patterning the retina for color vision. Color vision requires the comparison between photoreceptors with different sensitivity. In Drosophila, this is achieved by photoreceptors R7 and R8. 30% of R7s contain UV-sensitive Rh3 and are associated with R8 containing blue-Rh5. The remaining 70% of R7s contain UV-Rh4 associated with R8 containing green-Rh6. This stochastic choice is made in R7 where the transcription factor Spineless activates Rh4. In the absence of spineless, R7s express Rh3 and signal R8 to express Rh5. This signal is then interpreted by a bistable loop between the Hippo/Warts tumor suppressor pathway and the growth regulator Melted. After the R8 fate is established, the Rhodopsin molecules themselves contribute to its maintenance in order to avoid co-expression of other Rhodopsins. In aim 1, we will determine how the architecture of the Hippo/Warts pathway differs in tumor suppression where it is homeostatically regulated to control proliferation, and in R8 fate specification where t is bistable. We will determine when and how the pathway is activated and how positive feedback loops control the R8 decision. We will then analyze the cross-transcriptional repression of warts and melted. In aim 2, we will elucidate the pathway downstream of Rhodopsins that is required for the exclusion mechanisms ensuring that a single Rhodopsin is expressed in a photoreceptor. We will also address how the Rhodopsin activity pathway interacts with the Warts/Melted bistable loop. In aim 3, we will characterize novel regulators of the R8 subtype specification pathway that were identified in an RNAi screen. We will also take advantage of wild type natural variants that exhibit extensive co-expression of Rh5 and Rh6 to identify new factors involved in the exclusion mechanism. Finally, we will extend the RNAi screen to signaling molecules in order to target the pathway involved in the communication between R7 and R8. The results of this study will provide fundamental insights into the role of the Hippo/Warts tumor suppressor pathway in cell specification and will provide insights into sensory receptor cross- regulation. Overall, our findings will not only provide novel fundamental concepts for the field of sensory organ development, but will also contribute to cell fate specification and growth regulation.

[unreadable] DESCRIPTION (provided by applicant): Although the major task of the visual system is to build a spatial map of the environment, it is also able to detect the wavelength of light. To achieve color vision, animals compare the inputs of at least two spectral types of photoreceptors (PRs). In Drosophila, color vision is achieved by the two inner PRs R7 and R8 which express different rhodopsins (rhs) and project to close targets in the medulla part of the optic lobe. Three classes of ommatidia with different types of inner PRs exist. In p ommatidia, R7 contains the UV-Rh3 while R8 contains blue-Rh5. In y ommatidia, R7 contains UV-Rh4 and R8 green-Rh6. This results from a stochastic decision made in R7 to express rh4 and to exclude rh3. A signal is then sent from the rh3-expressing R7 to induce rh5 in R8. By default, the other R8 express rh6. A third class of ommatidia, located at the dorsal rim area (DRA), express Rh3 in both R7 and R8. These PRs measure the vector of light polarization, contributing to the "compass" of the fly. In the following aims, we propose to investigate how the distinction between the three classes of ommatidia is achieved: 1. Stochastic decision for the choice of color photoreceptors: After their initial recruitment, the inner PRs differentiate sequentially to acquire their specific subtype. We will study the steps that include control of inner PR differentiation by spalt, of R7 fate by prospero and of R8 fate by senseless. We will investigate how spineless controls the choice between p and y ommatidia by inducing rh4 expression in yR7, and how its expression is regulated. 2. Localized control of polarized light sensitive photoreceptors: We will continue our investigations of the signaling events that allow the expression of homothorax in the DRA. We will study how hth and other effectors control rh gene expression and adjust the morphogenesis of these PRs to the special task of polarized light detection. 3. Transcriptional expression of rhodopsin promoters: We will continue our dissection of the rh promoters as critical tools to understand the genetic regulations leading to PR subtype specification. We will focus on the repression mechanisms that allow rh6 to be restricted to yR8 and on the regulation ofrh3 in the DRA. We will also study how rh4 reflects the stochastic choice resulting from spineless expression, and how rh5 is controlled. The regulative network will then be reconstructed in cell culture. 4. Enhancer trap screen for genes expressed in subsets of photoreceptors: We will continue the investigation of genes that were identified in an enhancer trap screen. We will also utilize a new type of transposable element (piggyBac) to identify genes that have not been targeted by P-elements. An EP-type over-expression screen to identify further regulators will also be continued. Finally, we will use a bioinformatics approach to search for promoters sharing regulatory elements with PR specific genes. [unreadable] [unreadable] [unreadable]

DESCRIPTION (provided by applicant): The aim of this proposal is to obtain funding to support the participation of speakers as well as more junior students and postdocs to the third Visual System Development Gordon Research Conference to be held at Roger Williams University, Bristol, Rhode Island, June 6-11, 2004. The purpose of this Gordon Conference is to bring together investigators using Drosophila and vertebrate animal models to study the development and the evolution of the visual system. The goal is to generate an understanding of common principles that regulate visual system development. This is the third time that this conference is being held as a Gordon Conference. The first incarnation of this meeting was held as the Drosophila Visual System Development Workshop in 1992. The scope of the meeting was expanded to include vertebrate systems in 1994, when the seminal paper from Walter Gehring's lab showed that the same 'master control gene' (Pax6) specifies eye organ fate in both flies and vertebrates. Visual System Development Workshops continued to provide a stimulating venue for cross-fertilization of ideas between the Drosophila and vertebrate scientific communities. The meeting was accepted as a Gordon Conference in 2000. The Gordon Conference model is an ideal fit for this conference, with a selected set of talks and ample time for discussion as well as posters providing an excellent compromise between breadth and depth in a limited agenda. This format allows for, indeed requires, ample time for both formal and informal discussion, giving participants the opportunity to interact and exchange ideas in a highly stimulating and scientifically exciting setting. This type of scientific meeting provides individual investigators with new ideas and potential collaborations, and could also serve as a strategic planning tool for generating ideas for the upcoming NEI five-year plan for vision research. The field has grown and continues to expand rapidly. Having the capacity to accept 120 people has spread the benefit of The Visual System Development conference more widely.

PROJECT SUMMARY This proposal is a renewal of 5R01EY017916-04 Mapping the optic lobes for color vision. Visual information gathered by the Drosophila eye is first processed in the optic lobes, lamina, medulla and lobula complex. The medulla is the first relay in the neural network for color vision, but it is the second step in the motion detection pathway. In the previous granting period, we have defined over 70 cell types that form overlapping retinotopic maps before projecting to the lobula complex. Because the medulla contains a finite number of neurons and connections that can be studied in exquisite detail, we can address fundamental questions about the development and function of a sophisticated neural structure. We present experiments to understand how optic neuron diversity is generated and how these neurons establish their retinotopic connections to photoreceptors. This will produce knowledge and tools that we will then use to analyze the function of individual neurons in optic pathways through electrophysiology and behavior assays. Three specific aims will help us reach these goals. (1). Sequential neuroblast (NB) switching generates neuronal diversity. We have discovered that the NBs generated as a wave of differentiation from the medulla neuroepithelium express sequentially at least three transcription factors in a process similar to the sequence observed in embryonic NBs. The neurons emerging from these NBs maintain expression of these genes and become different cell types. We will identify new NB and neuronal markers and identify the adult neuronal cell types derived from larval neurons marked by combinations of TFs. (2). Regionalization of medulla neuroepithelium and specialization of neuroblasts. The types of neurons generated by the sequentially generated NBs differ in distinct regions of the crescent shaped medulla Outer Proliferation Center (OPC). In the central part, young NBs produce both local columnar cells that remain where they were generated, as well as a smaller number of non-columnar neurons that migrate to occupy their retinotopic position in the entire adult medulla. We will analyze how the lineage of NBs is modified by their position along the OPC. We will then manipulate the positional identity of NBs and analyze the consequence on the neuronal composition of the medulla. We will thus define the combinatorial transcription factor code specifying medulla neuron types and will relate it to their adult fates. (3). Function of individual medulla neurons. We will record electrophysiological activity of these neurons in response to various light stimuli. We will continue our analysis of neurons involved in motion detection and extend this analysis of medulla neurons involved in chromatic pathways. The tools generated in aims 1 and 2 will allow us to silence of stimulate these neurons through 'intersectional' expression to analyze the motion and color behavior of flies using our flight simulator.

DESCRIPTION (provided by applicant): In Drosophila, color vision is achieved by the two inner photoreceptors R7 and R8 which express different rhodopsins and compare their outputs in the medulla part of the optic lobe. Three classes of ommatidia with different functions form a mosaic in the retina. In p ommatidia, R7 contains the UV-Rh3 while R8 contains blue-Rh5. In y ommatidia, R7 contains UV-Rh4 and R8 green-Rh6. These two classes are randomly distributed in the retina with a 30:70 ratio. This results from a stochastic decision made in R7 to express rh4 and to exclude rh3. Coordinate expression in R8 allows the formation of ommatidia that specialize in the discrimination of shorter (p) or longer (y) wavelengths of light. Ommatidia located at the dorsal rim area (DRA) specialize in the detection of the vector of polarized light. We have identified many of the factors that contribute to the choice of rhodopsins in the various classes of photoreceptors. In the following aims, we propose to further investigate the mechanisms regulating the stochastic choice of spineless (ss) expression as well as the spatial control of specific subsets of photoreceptors. 1. Regulation of photoreceptor choice in the larval eye and adult brain eyelet: The larval eye contains two types of photoreceptors: four express blue-Rh5 while eight contain green-Rh6, reminiscent of the Rh5:Rh6 ratio in adult R8 cells. We will compare the regulation of these rhodopsins in these two systems. At metamorphosis, the larval eye becomes the extra-retinal eyelet that contains four green-Rh6 photoreceptors. We will investigate how Ecdysone signaling leads to the death of the eight green-Rh6 photoreceptors and to the trans-determination of the four blue-Rh5 photoreceptors into green-Rh6. 2. Regulation of the stochastic expression of spineless: We will pursue our attempts to understand how transcription of ss is activated in a subset of R7 cells. We will explore the possibility that lozenge represses ss and might thus be responsible for its stochastic activation. We will dissect the ss promoter and identify factors that control its expression in the retina. 3. Localized distribution of specialized ommatidia: We will also analyze the function of an effector of ss, dve, a homeobox gene that appears to be involved in the repression of rh3 in cells that express ss. We will also investigate its relationship to Iro-C genes that control the development of two subtypes of ommatidia (DRA and a novel type hat co-expresses Rh3 and Rh4 in yR7 in the dorsal part of the eye). 4. An FRT-piggyBac screen to identify novel genes regulating R7 subtype specification. We have devised a genetic screen using a collection of 4,000 FRT-piggybac insertion lines provided by the lab of Liqun Luo. We will generate whole mutant eyes for each insertion and will evaluate subtype ratio visually with rhodopsin GFP reporters using water immersion fluoroscopy. This unbiased approach will allow us to complete our understanding of the regulatory pathway that leads to the formation of the retinal mosaic. PUBLIC HEALTH RELEVANCE: The Drosophila eye has served as a very powerful model system to understand how different classes of photoreceptors develop to form the retinal mosaic that is responsible for the function of the eye: motion detection, dim light or color vision. We offer to pursue our investigations of the events that allow photoreceptors to take on specific fates, express one of the rhodopsin photopigments and exclude all others. The identification of regulators of these processes will lead to a better understanding of genes whose mutations cause blindness in humans.

DESCRIPTION (provided by applicant): The gene regulatory network that controls segmentation in insects has adapted to different developmental situations. For instance, the Bicoid morphogen plays a pivotal role in patterning the anterior of the long germ Drosophila embryo. However, no bicoid homologue has been found outside Diptera, even in other long germ insects. Here, we propose to continue our work on the long germ embryo of Nasonia (Nv), the wasp that is the focus of much of our studies. We have shown that four maternal functions are required to pattern the Nasonia embryo: hunchback (hb), orthodenticle (otd1), giant and caudal are the key maternal components of an ancestral patterning system whose function has been taken over by bicoid in Drosophila. We will ask how these genes act together to pattern the axis and achieve segmentation of the Nasonia embryo. Aim 1. We will continue our study of gap genes in Nasonia. We will test their cross-regulatory interactions and analyze their phenotypes using parental RNAi that works very well in Nasonia. We will pursue our analysis of pair-rule genes, in particular even-skipped (eve) and ftz whose expression patterns are dramatically different from those of flies, in particular in the posterior regions. We will test how they are controlled by upstream factors and will investigate the phenotype of pRNAi embryos. Aim 2 We will study the molecular regulation of pair-rule genes, focusing on eve whose control is understood in exquisite detail in Drosophila. We will identify the individual regulatory modules that control anterior pair-rule stripes of Nv eve, or posterior segmental expression. This will highlight the two modes of segmentation that appear to co-exist in insects: one for the anterior (and the only one in Drosophila), and one for the posterior segments that is reminiscent of the mode of segmentation of Tribolium. We will analyze the molecular mechanisms of Nv eve stripe 2 expression using genetics, bioinformatics and transgenics. Site-directed mutagenesis of the eve regulatory region will allow us to assess the roles of Otd1, Hb, gap gene as well as pair-rule gene products and how they differ from their roles in Drosophila. Aim 3 Our preliminary data indicate that localization of mRNA is a critical component of Nasonia's ability to pattern its embryo in the absence of bicoid. Nv otd1 mRNA is localized at both poles while giant is present only at the anterior and caudal form a mRNA gradient. We will evaluate the ability of the 3'UTR of these genes to direct mRNA localization in Drosophila and in Nasonia and assess whether the signals work across species. We will manipulate the cytoskeleton to define the requirements for mRNA localization machinery. The function of genes involved in this process in Drosophila will be tested by RNAi in Nasonia, and their importance for localizing each mRNA determined. PUBLIC HEALTH RELEVANCE: The Drosophila embryo is the best known complex biological system. However, it is a unique example that does not represent the diversity of life. By comparing Drosophila development with Nasonia, we will be able to distinguish general mechanisms that have broad implications for the development of animals from more specific processes.

Developmental Biology is at the center of the Life Sciences. Its goal is to understand how the fertilized egg grows and is transformed into an adult organism. Developmental biologists have uncovered the basic biological processes of embryogenesis and pattern formation that led to the Nobel Prize awarded in 1995. They also study the formation of organs, of the nervous system and how sex is determined. They have also discovered the process of aging and how cell death can be programmed to shape an organism (Nobel 2002). Developmental biologists were also the first to clone animals. They have discovered microRNAs (that led to the Lasker award in 2008) and have analyzed the major signaling pathways whose mis-regulation leads to cancer. Developmental biologists have also developed important technological advances, from genome manipulation (Nobel 2007) to RNAi (Nobel 2006) and in vivo imaging using fluorescent proteins (Nobel 2008). They have provided the foundations for stem cell biology and tissue engineering and have created the context to understand human birth defects and disease. The impact of the field is a result of the study of whole organisms (rather than isolated cells) using a wide variety of technological and intellectual approaches. As these studies expand into new areas such as tissue regeneration and systems biology, major contributions can be expected in the coming years. Developmental Biology is thus a core discipline of the Life Sciences and is transforming the Medical Sciences. Over 120 developmental biologists will meet on June 21-26 at the Gordon Research Conference to discuss recent progress in the field and exchange ideas. The conference will include minorities and a large proportion of women that reflects their high representation in the field. Emphasis was given to inviting a number of young speakers and to avoid repeats of previous conferences held in the same field.

DESCRIPTION (provided by applicant): This proposal is a renewal of 5R01EY013010-12 Exclusion & Coordination of rh genes. Visual information is gathered in the Drosophila compound eye by photoreceptors specialized in various tasks. We focus on the different types of photoreceptors involved in color vision. They express Rhodopsin photopigments that detect lights of different wavelengths: R7 photoreceptors express a UV sensitive Rhodopsin (Rh3 or Rh4), and R8 express either a Blue (Rh5) or a Green (Rh6) Rhodopsin. Color vision is achieved through comparison of light information received by the R7 and R8 of each individual eye (ommatidia). Since there are two types of ommatidia, Rh3 input is compared with Rh5, or Rh4 with Rh6. The two types of ommatidia are distributed stochastically in the retina with a 30:70 ratio. Their specification is controlled by the transcription factor Spineless that is it expressed stochastically in a subset of R7 cells. In aim 1, we propose to investigate the molecular mechanisms that control how spineless is intrinsically expressed in a stochastic manner. We will identify the different elements of the spineless promoters that control expression in all R7 cells, and those that lead to repression in a stochastic subset. We will use biochemical approaches such as 3C and Chromatin ImmunoPrecipitation to investigate the molecular mechanisms involved in the repression. We will also take advantage of natural variants mapping at spineless that have dramatically decreased ratio of Rh3:Rh4 to define the elements that control this ratio. In aim 2, we will generalize these observations to define the mechanisms by which the rhodopsin promoters integrate the information about cell specification to drive high level and yet exquisitely specific expression of different rhodopsins in distinct photoreceptor types that are otherwise highly related. We will investigate how elements common to all rhodopsin genes drive broad photoreceptor expression while specific elements restrict expression to their appropriate subset. We will identify cis-acting elements as well as the transcription factors that act on these sites. We will then use our deep knowledge of the system to reconstruct the rhodopsin promoters with defined elements reassembled as synthetic promoters. This achievement would represent a unique example where we understand the grammar of the transcription code and would be applicable to elucidating mechanisms of control of other terminal differentiation genes such as vertebrate opsins. PUBLIC HEALTH RELEVANCE: Drosophila, with its genetic amenability and its sophisticated visual behavior, has been very successfully developed as a model system to study retinal patterning. We investigate how the different subtypes of photoreceptor neurons are generated and distributed in the retina to achieve motion detection and color vision, in particular, studying the mechanisms by which stochastic choices increase diversity of photoreceptors and how the promoters of rhodopsin genes are regulated. The principles deduced from this project will be applicable to our understanding of patterning of the human retina.

? DESCRIPTION (provided by applicant): How the huge neuronal diversity observed in our brain is generated and how these neurons are assembled into functional networks are poorly understood. A unique model to study these complex questions is the Drosophila mushroom body (MB), which processes olfactory learning and memory, because its basic function and development are well characterized. Each MB consists of five cell types that are born sequentially from four identical neuroblasts: first, 'non-intrinsic' (ni) neurons that do not participate to the MB are born, then ?, ?'?', pioneer-?? and finally ?? neurons are generate. Vertebrate neural progenitors also produce neuron types sequentially, indicating that temporal patterning is conserved. Drosophila NBs have emerged as an excellent model to study the molecular mechanisms that regulate temporal patterning. We hypothesize that a temporal transcription factor cascade functions in MB neuroblasts to define each MB neuronal type during development. We also hypothesize that, unlike in most brain structures, MB intermediate precursors divide symmetrically to produce two identical neurons to generate the large number of identical neurons necessary to perform MB functions. We will determine the molecular mechanisms regulating the temporal and symmetrical production of MB cell types. This will provide novel insights into developmental programs generating the diverse neurons that make up our brains. Aim 1: Define and test the mechanisms that sequentially specify MB neuron types. The sequential production of neurons from MB NBs involves four transitions resulting in five distinct cell types. We hypothesize that MB neuroblasts sequentially express a series of transcription factors as they age to confer neuronal identity. We will identify these transcription factors and use mutant clonal analysis, RNAi and overexpression assays to address the function of these factors in producing MB neuron types. Aim 2: Study the regulation of asymmetric NB and symmetric GMC division in the MB. Although MB neuroblasts divide asymmetrically, MB intermediate precursors divide symmetrically to generate two identical neurons. In all other studied lineages, these cells divide asymmetrically to generate a NotchON and a NotchOFF neuron. We will test whether both identical MB neurons are NotchOFF or NotchON as a consequence of rotating the axis of the last cell division orthogonal to the MB neuroblast division plane. Aim 3: Determine the RNA profile of mature MB neuron types. We have Gal4 lines that mark individual MB neuron types. We will use them and others we identify to FACS and then transcriptionally profile adult ?, ?'?', pioneer-?? and ?? neurons to understand the specification of these neurons. Adult neurons may also maintain expression of the transcription factors that specify the temporal windows of MB neuroblasts. Expression and function of candidate genes differentially expressed between each MB neuron type will be tested.

? DESCRIPTION (provided by applicant): In the Drosophila eye, inner photoreceptors (PRs) R7 and R8 mediate color vision and belong to two subtypes, expressing UV-sensitive Rh3 in R7 and Blue-Rh5 in R8 (p subtype), or UV-Rh4 in R7 and Green-Rh6 in R8 (y subtype). Reminiscent of the stochastic distribution of L and M cones in the human retina, the two subtypes are distributed stochastically, with a p:y ratio of 35:65. We have deciphered the gene regulatory network that controls PR subtype specification. The random choice between p and y fate is controlled by stochastic expression of the transcription factor Spineless in 65% of R7s where it induces yR7/Rh4 fate. We have shown that each spineless allele makes an intrinsic stochastic choice to be expressed; however, inter-chromosomal communication synchronizes expression of the two alleles. We will address whether the mechanisms of stochastic choices can be generalized, and why Drosophila uses a stochastic distribution while other species exhibit highly deterministic PR patterning. We will use transcriptomics and cis- regulatory dissection to continue our investigations of PR subtype specification. Finally, we will investigate how PRs that are stochastically determined find their corresponding targets in the optic lobes to convey color information to the brain. Aim 1. Stochastic vs. deterministic choices in photoreceptor determination. Using the MS2 system, manipulation of the spineless proximal promoter and its epigenetic landscape, we will live image spineless transcription in PRs to understand whether spineless is initially stochastically repressed, or if noisy fluctuations are locked-in at a given time. We will analyze the role of spineless in other species, e.g. in butterflies that have two R7s that make independent stochastic choices, and use CRISPR to knock out spineless in these species. Another fly family, Dolichopodidae (Doli), has eyes with alternate stripes of Green and Blue R8s. We will address the regulation and function of spineless in this deterministic process. We will swap the Doli spineless gene into D. melanogaster to make it deterministic. Finally, we will analyze how the neural network is reconfigured in the visual system of males of many fly species to build an improved motion detector. Aim 2. Cis- and trans-regulatory logic of photoreceptor subtype specification. We have obtained a comprehensive knowledge of the gene regulatory network controlling PR determination and the cis-regulation of Rh promoters. However, PR specification precedes Rh expression by two days. We will therefore analyze how early genes (dpr11, dve, warts and melted) are expressed in subtypes of PRs and compare the logic of regulation at several stages for these different genes with functions in PR specification or axon pathfinding. We will compare the early and late transcriptomes of the four types of color as well as polarization PRs and identify potential regulators predicted by our cis-analysis of Rh promoters. Aim 3. How do stochastically determined PRs find their target in the optic lobes? We will investigate whether and how p and y R7 and R8 connect to dedicated subsets of neurons in the medulla and address how these neurons are specified to transmit color information. We will study the role of Ig proteins Dprs and their DIP receptors that allow the matching of PRs and their target in the medulla. Finally, we will study how stochastically determined PRs specify their target neurons or allow survival of pre-existing medulla neurons.

Project summary Visual information is processed through specialized channels: In each ommatidium of the Drosophila retina, motion is processed by the six ?outer photoreceptors? that project to the lamina. Color is detected by photoreceptors R7 and R8 that belong to two subtypes: In the p subtype, pR7 expresses UV-sensitive Rh3 and pR8 contains Blue-Rh5. In the y subtype, UV-Rh4 in yR7 is coupled to Green-Rh6 in yR8. We offer to study the signal that relays this information from R7 to R8 to coordinate the expression of Rhodopsins (Aim 1). We will extend this work to the signals originating from outer photoreceptors, which control the division of lamina precursor cells (LPCs) by releasing Hedgehog, and which induce the differentiation of the lamina neurons L1-L5 by releasing EGF/Spitz from their termini. In Aim 2, we will determine how the development of photoreceptors and of their lamina targets is coordinated. We also discovered that distinct glial cell types play specific roles in patterning each of the five lamina cell types. This has prompted us to study how glia originate in different regions of the optic lobes, how they migrate to take on their final positions, and how they mediate differentiation signals to L1-L5 (Aim 3). Aim 1. Specification of the R8 subtype fate by signaling from R7 photoreceptors. We will analyze the significance of multiple BMP/Activin ligands that participate in R7-R8 communication, and test whether they signal through their canonical receptors. The signal is reinforced in R8 by a bistable loop involving the Warts tumor suppressor and a growth regulator, Melted. We will determine whether Mad/Smad2 activate melted or repress warts, to generate and maintain the correct ommatidial R7 and R8 subtypes. The BMP/Activin signals are restricted to a single ommatidium, and do not diffuse to neighboring ommatidia. We have shown that the Hibris cell adhesion molecule limits this diffusion and will study how it affects R7-R8 pairing. Expanded signal diffusion leads to clusters of Rh5 R8s that are reminiscent of clusters of pigmentation in the eye of Eristalinus flies. We will use our evo-devo expertise to study how the signal may be propagated in specific species. Aim 2: How do signals from photoreceptors specify the five cell types of the lamina? We will explore how signals from the outer photoreceptors regulate proliferation and cell cycle exit to generate lamina precursors. We will then manipulate EGFR to ask how Spitz levels from photoreceptors affect lamina differentiation. We will also test how EGFR/Notch antagonism among lamina precursors contributes to L1-L5 diversification. Aim 3. The role of different glial populations for the specification of lamina neurons. We have evidence that the order and kinetics of lamina cell determination is coupled with glial development. We will manipulate glial cells to determine how they are involved in mediating the EGF signal from photoreceptors. Using genomics and live imaging, we will characterize the different populations of glia and determine their origin. We will then develop a quantitative and predictive model of the interactions that contribute to specifying lamina neuronal diversity.

? DESCRIPTION (provided by applicant): Four neural structures in the Drosophila optic lobes (lamina, medulla, lobula and lobula complex) sequentially process visual information through neural networks of specialized cell types organized in a retinotopic manner. We will continue our investigations to understand how neuronal diversity is generated in the developing optic lobes, which are comprised of more than 100 cell types. Our observations suggest that neuronal specification in the medulla results from the integration of three mechanisms: (i) 800 neuroblasts express a sequence of temporal transcription factors to generate distinct types of neurons as they age, each contacting one of the 800 columns innervated by photoreceptors. (ii) The temporal series is modified locally by regional transcription factors and produces neurons that innervate multiple columns. (iii) Binary fate choice via Notch further diversifies daughters of the terminal cell division. In the posterior-most region of the developing medulla and in the progenitor region of the lobula complex, neurogenesis differs significantly with a different set of transcription factors that act not only to specify neuronal fate but also to control the precursor mode of division and the death or survival of neurons. This illustrates how complex brain structures use different strategies to adapt and produce the correct number of specific cell types with the appropriate characteristics. We will investigate the mechanisms controlling this neurogenesis. Aim 1: Temporal progression of neuroblasts: Timing and transition mechanisms Temporal patterning is a general mechanism to generate neural diversity in flies and vertebrates. We will explore the molecular processes controlling the temporal progression of neuroblasts in the medulla. Aim 2. Regionalization of the medulla neuroepithelium and specialization of neuroblasts We will investigate the rules that modify the output of the temporal series in different regions of the medulla progenitor domain. This allows the local production of neurons that migrate to occupy the entire medulla. Aim 3. Correlation between transcription factor expression and neuronal characteristics To understand how transcription networks control the characteristics of neurons, we will use large-scale single cell transcriptomics to identify regulatory interactions and determine how these define the identity of each neuron. Aim 4. Regulation of the mode of neuroblast division and neuronal survival or death by temporal patterning We will investigate how temporal transcription factors act on the cell cycle and on pro-apoptotic genes to characterize the different strategies used by distinct parts of the optic lobes to produce specialized neurons. Aim 5. Temporal patterning-independent neurogenesis in lobula complex progenitors We will explore the molecular mechanisms that control a different mode of neurogenesis that produces 3 types of lobula neurons without a temporal series by controlling the rapid exit of neuroblasts from proliferation. This ambitious work will allow us to identify basic principles of neural patterning and diversity generation, which have broad implications for other neuronal systems in flies and vertebrates.

Project summary: Ants are social insects that can be developed as experimentally tractable organisms for probing the dynamic changes in the epigenetic programs that control an array of processes, including aging. Aging is a process of progressive decline in intrinsic physiological functions. There is a well-established trade-off between lifespan and reproduction as higher reproductive activity in females is associated with shorter lifespan. However, in social insects, the reproductive queen has up to 10X longer lifespan than non-reproductive workers. In the ant Harpegnathos saltator, adult individuals that are not exposed to queen pheromones can undergo a reversible switch from non-reproductive workers to reproductive pseudo-queens (gamergates) that exhibit fully developed ovary and, importantly, a 5X increase in lifespan, showing that aging is reversible. Lifespan is shortened again when gamergates are reverted to workers (revertants). Thus, Harpegnathos provides an effective system to study epigenetic regulation of aging and rejuvenation given the adult plasticity that allows switching between castes. We have performed transcriptome analysis of the longevity-regulatory tissues: the fat body, ovary and brain, in workers vs. gamergates vs. revertants. We have identified a group of differentially expressed genes (DEGs), some of which have been implicated in the regulation of longevity, e.g. IIS (insulin and IGF signaling) pathway components. To further analyze the genetic and epigenetic regulation of longevity in ants, we will first compare physiology, lifespan, transcriptome and histone modifications in gamergates derived from young vs. old workers to ascertain the epigenetic regulation of aging and, most intriguingly, rejuvenation. The cellular localization and functions of important DEGs will be further analyzed. Second, we will utilize our newly established genetic tool ? CRISPR in ants ? to generate knockout (KO) ants in the two DEGs in the ovary, Hs- IMPL2 and Hs-ALS, which likely act as inhibitors of IIS in ants. These KO ants will be used to characterize the role of IIS in the dramatically extended lifespan in gamergates. Third, Hs-ILP1 (insulin-like peptide 1) is differentially expressed in the brain and Hs-ILP2 in the ovary. While both Hs-ILP1 and Hs-ILP2 are strongly increased in gamergates compared to workers, IIS is decreased in the fat body and ovary. To address this paradox, we will analyze (a) the role of Hs-ILP1 and -ILP2 in regulating the two branches of IIS: AKT and MAPK; (b) the role of Hs-ImpL2 and Hs-ALS in regulating activity of Hs-ILPs; and (c) how two insulin receptors (InRs) mediate the role of Hs-ILPs in differentially regulating IIS.

Project Summary To investigate the generation of neural diversity, we use the simple brain of Drosophila that has only 100,000 neurons but can support complex behaviors and simple learning. The highly deterministic nature of Drosophila brain development allows us to define general rules that control the generation of neural diversity and are applicable to mammalian development, even when this is further modulated by activity-dependent plasticity. The genetic control of optic lobe development can be investigated in depth thanks to the repetitive nature of the system, where information from the 800 unit-eyes (ommatidia) projects to 800 parallel retinotopic columns that sequentially process the visual information through more than 200 cell types. The generation of neural diversity results from the integration of three mechanisms: (i) ~800 medulla neuroblasts (NBs) are patterned by the sequential expression of temporal transcription factors that generate distinct types of neurons at each temporal window. (ii) NBs produce different neurons depending on their location in the neuroepithelium: Spatial factors locally modify the outcome of the temporal series. (iii) Binary fate choice via Notch signaling further diversifies the two daughters of ganglion mother cells (GMCs) born from each NB division. In contrast, NB transitions in the mushroom body, a brain region involved in learning, are controlled by extrinsic factors, generating fewer neuron types through the use of extremely long lineages. The broad context of this proposal will address how basic principles of neurogenesis explain the vast diversity of neurons in the optic lobes and the restricted diversity in the mushroom body, and will help us understand more complex brain structures and instruct further studies in mammals. We will investigate the mechanisms controlling neurogenesis through 4 aims: Aim 1: Temporal progression of neuroblasts: Timing and transition mechanisms: Temporal patterning is a general mechanism to generate neural diversity in flies and vertebrates. We will identify all the temporal factors and investigate their mode of cross-regulation that controls the timing of transitions. Aim 2. Intrinsic specification of neuroblasts in culture: A given temporal transcription factor appears to control the expression of the next factor in the series and to repress the previous factor to form a transcriptional clock mechanism. We will use live imaging of transcription (MS2 system) and of protein expression in vivo and in cultured NBs to investigate the intrinsic molecular mechanisms controlling the timing of transitions. Aim 3. Specification of multi-columnar neurons: We will investigate how multicolumnar neurons are produced locally in response to spatial factors while innervating the entire retinotopic map. We will also investigate how their cell bodies move to distribute throughout the optic lobe. Aim 4. Extrinsic cues for neuroblast transitions in the mushroom body: The mushroom body NBs have very long lineages but produce a limited number of cell types. We will study how Ecdysone and Activin signaling mediate extrinsic transitions between cell types and how they control gradients of RNA binding proteins acting in neuroblasts.

Project Summary How neuronal progenitor cells produce an enormous diversity of neuronal and glial cell types is a fundamental topic that remains largely unresolved. Drosophila has been a pivotal model system to study these complex questions of neurogenesis and research in its nervous system has contributed to several important concepts that apply to mammals. These include temporal and spatial patterning, cell death, neural and/or glial specification and asymmetric cell division. Nevertheless, the full resolution of its neuronal lineages remains elusive. In this proposal we will take advantage of powerful genetic tools derived from synthetic biology to reconstruct the entire neuronal lineage of multiple brain structures at single cell resolution. We will develop transgenic flies in which lineages can be autonomously recorded and analyzed through single cell transcriptomics. We hypothesize that our knowledge on the structure and molecular nature of adult brain development will provide us with a unique advantage to not only reconstruct the entire lineage tree of the Drosophila brain, but also to form new hypotheses on how each of the brain structures form very faithfully and uniformly from precursor cells. Aim 1 Development of a progressive lineage recorder in Drosophila. There are currently no established CRISPR-based lineage methods in the Drosophila nervous system. We will adapt GESTALT in Drosophila to `scar' the DNA during lineage progression and progressively record this lineage through development. We will use genetic tools to gain spatio-temporal control over our lineage measurements and use in silico modeling of the barcode structure to optimize the activity of the system. We will generate flies with enough target sites to capture the entire neuronal diversity generated during neuronal development. We will empirically evaluate different versions of the technology and select the best one for single cell lineage tracing. Aim 2 Defining neuronal birth order and clonal relationships in the adult brain. We will lineage trace through the neuronal development while simultaneous sequencing the transcriptome of single cells. This should allow us to identify the different neural subtypes using our single cell atlas. We will characterize the lineage information per cell and combine this with the published methods capable of reconstructing multi-tree lineages to identify different lineage relationships in our data. We will build on the stereotypical mode of neuronal development to refine this structure and reconstruct the neuronal lineages Aim 3 Experimental validation of the reconstructed neuronal lineages. We will use our prior knowledge of neuronal development in combination with post hoc validation to benchmark our lineage reconstruction. We will first compare our lineage reconstruction of the local motion detectors in Drosophila to their known simple and well-defined lineage. We will use region-specific Gal4 lines to lineage trace different subregions of the neuroepithelium followed by FACS and single cell sequencing to identify the neurons born in these regions. Evaluation of those relationships in our reconstructed tree will provide further validation for our lineage trees.

Visual information detected by the retina is sent for further processing to deeper layers of the visual centers that feed into specialized behavior circuits. Much is known about how visual centers and local circuits function, but we do not understand how the many cell types that compose them are generated and how these circuits assemble and are coordinated between brain regions. We study the simplified, yet highly performing visual system of Drosophila for which we have obtained a deep understanding of neural diversity, mechanisms that also apply to mammalian neurogenesis. In a separately funded grant, we have generated a very large dataset where we have identified through single cell mRNA sequencing the individual transcriptome of most (169 neural types) neurons and glia in the four optic lobe ganglia, lamina, medulla, lobula and lobula plate, through six development stages starting when the neurons are first generated. This represents a huge resource that will allow us to identify the molecular pathways involved in the processes studied here. In the current proposal, we will define how the circuits formed by optic lobe neurons are assembled and how development of the different optic lobe neuropils is synchronized: Aim 1. Retinotopic projection of photoreceptors to the lamina and medulla. We will define the role of the lamina in the establishment of retinotopy in the medulla and what guides photoreceptors and lamina neurons to their retinotopic location. We will then identify the molecular guidance pathways involved in pathfinding in lamina and medulla, and determine the potential role of pioneer neurons that might guide retinotopy of the other neurons. Aim 2 Timing of differentiation and layer formation in the medulla neuropil. Medulla neurons are born from the same neural progenitors in a sequential order, a fundamental mechanism of 'temporal patterning'. We will test the model that temporal patterning allows medulla neurons to progressively innervate each layer of the lobula and of the medulla and we will define the molecular mechanisms synchronizing birth order and layer formation. Aim 3: Development of output neurons from the lobula: Visual features and optic glomeruli. Signals from the medulla are conveyed to the lobula and are then passed on to Lobula Columnar Neurons (LCNs) that connect to 'optic glomeruli' that control behavior. We will study how ~20 subtypes of LCNs connect to different layers of the lobula and to specific glomeruli in the central brain and will define the molecular mechanisms of their specific targeting. Aim 4. Building the broad-field motion pathway. Motion is computed by neurons that compare the outputs of upstream neurons in a specific orientation. We will investigate the developmental programs that instruct the dendrite orientation of the first orientation-selective neurons (T4 and T5) and how they each project to one of the 4 layers of the lobula plate that each detects motion in one of 4 cardinal directions. This study will provide fundamental insights into the coordination of various elements of a simple and amenable visual system. Our findings will not only provide novel fundamental concepts for the development of circuits for sensory processing, but will also contribute general concepts applicable to circuit formation in vertebrates.

Project Summary The establishment of neuronal connectivity requires axons to select the proper neurons to form synapses with, which can be achieved through guidance signals or through activity-dependent processes. We still have a poor understanding of the mechanisms and molecules that direct neuronal specific connectivity pattern. The Drosophila color vision circuit offers a powerful paradigm to study synaptic specificity because of the availability of a connectome, a deep knowledge of its development, and powerful genetic tools to manipulate the circuit. In the fly retina, color photoreceptors R7 and R8 are stochastically specified, whereas their neuronal medulla targets are produced through a highly deterministic program. We will study how this propagation is achieved. Activity-dependent neural patterning is another powerful mean to coordinate different brain regions. Neuronal activity can occur at early developmental stages, prior to sensory input and the onset of synaptogenesis in the form of calcium waves whose significance has not yet been fully elucidated. We will study how early waves of activity that we observe at a very specific time point during fly retinal development are generated and what role they play. Aim 1. How do stochastically determined photoreceptor subtypes find their targets in the optic lobes? Aim 1.1. Synaptic specificity downstream of color photoreceptors: We will identify medulla neurons specific to subsets of photoreceptors and will identify the factors involved in the recognition by their input neurons. Aim 1.2. Establishment of synaptic specificity downstream of R7 color photoreceptors: We will study how a family of cell adhesion molecules allows matching between R7 cells and their Dm8 targets in the brain Aim 1.3. How does information from p and yR7 propagate downstream of Dm8: Once a medulla cell has made contacts with its cognate photoreceptor, it must itself transmits this information to its downstream partners. We will investigate how these choices are propagated down the visual pathway. Aim 1.4. Apoptotic pathway regulating the culling of unconnected neurons: Target neurons that are not connected die. We will study how specific adhesion molecules regulate this death and the molecular pathways involved. Aim 2. Mechanisms and functions of waves of spontaneous activity in the retina Aim 2.1. Description and cellular substrates of retinal calcium waves: We will analyze which cells are involved in calcium waves and whether these waves propagate to downstream medulla regions. Aim 2.2. Molecular mechanisms of retinal calcium waves: We will study how ER calcium stores and gap junction proteins are required to generate and propagate the waves. Aim 2.3. Determine the developmental role of retinal calcium waves: By disrupting the calcium waves, we will study what role they play in patterning the retina, and/or medulla neurons that are targets of photoreceptors.