Understanding the effect of the environment on human health has benefited from progress made in measuring the exposome. High resolution mass spectrometry (HRMS) has made it possible to measure small molecules across a large dynamic range, allowing researchers to study the role of low abundance environmental toxicants in causing human disease, including examining their effects on biochemistry. Alzheimer’s disease is the most prevalent neurodegenerative disease in the world. While aging is the largest risk factor of the disease, evidence of risk factors for dementias show that lifestyle choices and the environment may modify disease onset and alter the projected prevalence. Observational epidemiological studies have linked exposure to the persistent pesticide dichlorodiphenytrichloroethane (DDT) with increased risk of Alzheimer’s disease (AD). In Chapter 2, using an aging cohort based in Washington Heights and Inwood in Northern Manhattan, I investigated systemic biochemical changes associated with Alzheimer’s disease (AD). Small molecules in plasma were measured in 59 AD cases and 60 healthy participants of African American, Caribbean Hispanic, and non-Hispanic white ancestry using untargeted liquid-chromatography–based ultra-high-resolution mass spectrometry. Metabolite differences between AD and healthy, the different ethnic groups and apolipoprotein E gene (APOE) ε allele status were analyzed. Untargeted network analysis identified pathways enriched by AD-associated metabolites. Then, in Chapter 3, using the genetically tractable nematode model Caenorhabditis elegans, I investigated whether DDT can exacerbate AD-related pathology. DDT is a persistent organic pollutant which, despite its ban in 1972, can be detected in the blood of most people in the U.S. I investigated whether DDT can exacerbate AD-related pathology using a transgenic C. elegans strain that expresses a mutant tau protein fragment that is prone to aggregation, as well as a mutant strain expressing a non-aggregating form of tau protein. DDT restricted the growth in all strains; however, the restriction was more severe in the aggregating tau transgenic strain. Further, I found that DDT exacerbates the inhibitory effects of aggregating tau protein on basal mitochondrial respiration, and increases the amount of time the worms spent curled/coiled. High-resolution metabolomics in the whole worm suggests that DDT reduces levels of several amino acids but increases levels of uric acid and adenosylselenohomocysteine. Surprisingly, developmental exposure to DDT blunts the lifespan reduction caused by aggregating tau protein suggesting a mitohormetic effect of the “double-hit” from DDT and aggregating tau protein or an antagonistic effect which could ultimately turn on lifespan extension pathways. Our data suggest that exposure to DDT likely exacerbates the mitochondrial inhibitory effects of aggregating tau protein in C. elegans. DDT may mimic some of the mitochondrial inhibitory effects induced by increased tau protein aggregation, suggesting that the genetic and environmental insult converge on a common mitochondrial inhibitory pathway, which has been associated with AD in several other studies. Finally, in Chapter 4, I determined changes in global metabolism associated with aggregating tau protein in both C. elegans and humans. We performed high-resolution metabolomic analysis on cerebrospinal fluid (CSF) and plasma obtained from patients of AD and mild cognitive impairment, and cognitively normal controls. Using a transgenic strain of C. elegans which expresses aggregating tau protein in all neurons, I studied the effect of aggregating tau protein on metabolism using high-resolution metabolomic analysis in the whole worm. In the population study, I found >300 features associated (p < 0.05) with phosphorylated tau levels in CSF. Metabolic pathway enrichment identified alterations in fatty acid and amino acid metabolism. Worms expressing aggregating tau showed >900 features altered. Pathway enrichment suggested alterations in glycerophospholipid, fatty acid and amino acid metabolism pathways. To determine which metabolic features are altered in both species, I analyzed annotated features for overlap. Five metabolites were concordant between human plasma and C. elegans, and four concordant between human CSF and C. elegans. Thus, in this analysis I provide evidence in support of using C. elegans to study changes in global metabolism associated with Alzheimer’s disease. In conclusion, using liquid and gas-based chromatography coupled with high-resolution mass spectrometry, we can measure levels of endogenous and exogenously derived small molecules in different biological matrices. By using the appropriate study design, we can identify candidate molecules and biochemical pathways associated with environmental exposures or disease in human populations. These candidates can be followed up by exposing an appropriate C. elegans strain: transgenic strains, mutant strains, or strains that are susceptible to RNAi based knockdown. Given their short life cycle and being amenable to high-throughput behavioral assays, they can readily provide functional and molecular readouts of the perturbation. The findings can provide leads for relevant policy around environmental exposures, understanding mechanisms of toxicity and disease, and identifying potential therapeutic targets.

The goal of this project was to elucidate the role of the homeodomain transcription factor family in terminal fate specification of an entire nervous system. In pursuit of this, I systematically identified the expression patterns of all 102 homeobox genes in the L4/adult stage C. elegans nervous system at a single-neuron resolution. This involved acquiring and/or generating high-quality fluorescent reporter reagents to tag all 102 homeodomain transcription factor proteins. Then, analyzing the expression of those reagents using a novel tool for whole nervous system identification in C. elegans, called NeuroPAL. The resulting expression atlas is the first complete picture of the homeobox family in any nervous system. It allowed the identification of new terminal selector proteins, including ceh-8, ceh-32, and ceh-31, in various neuron types and will continue to serve as a guide for future terminal selector identification across the nervous system. We discovered that every neuron type, and many subtypes, of the C. elegans nervous system express a completely unique set of homeodomain transcription factors. This unique expression code, along with the scores of homeodomain terminal selectors, suggests the possibility that every C. elegans neuron type is specified by a homeodomain terminal selector. We further probed the importance of the homeobox family in neuron specification by comparing its expression pattern with other transcription factor families. This necessitated high-quality data for the other major transcription factor families, bHLH, ZF, AtHook, bZip, and NHR, in every C. elegans neuron type. In collaboration with the labs of David Miller III at Vanderbilt and Marc Hammarlund at Yale, we used single cell RNA sequencing (scRNA-Seq) to molecularly profile all neuron types of the L4 stage C. elegans nervous system. We found that our homeodomain protein atlas was recapitulated fairly well in the scRNA-Seq data when thresholded and determined that the homeodomain transcription factor family is not alone in generating unique expression profiles for every neuron type. Two larger transcription factor families, ZF and NHR, are also uniquely expressed in each neuron identity. Instead, we found that the homeodomain transcription factor family is set apart from other families by their distinctly sparse expression across the nervous system at comparatively high levels. These expression patterns along with the numerous examples of functional homeodomain terminal selectors suggest that the family is an underlying theme in neuronal specification. We further extended this analysis to available scRNA-Seq datasets in the mouse nervous system and noted select commonalities in homeobox family expression across organisms. In all, this study shows yet again that analyzing homeodomain transcription factors leads to fruitful insights on organismal development. We found that the complexity of the C. elegans nervous system can be categorized and largely specified by a single family of transcription factors, building on previous studies of their importance in neuronal function.

Members of the Argonaute family of proteins, which interact with small RNAs, are the key players of RNAi and other related pathways. The C. elegans genome encodes 27 members of the Argonaute family. During this thesis research, we sought to understand the functions of the members of this gene family in C. elegans. Among the Argonaute family members, rde-1 and alg-1/2have previously been shown to be essential for RNAi and development, respectively. In this work, we wanted to assign functions to the remaining members of this large family of proteins. Here, we describe the phenotype of 31 deletion alleles representing all of the previously uncharacterized Argonaute members. In addition to rde-1, our analysis revealed that two other Argonaute members csr-1 and prg-1 are also essential for development. csr-1 is partially required for RNAi, and essential for proper chromosome segregation. prg-1, a member of PIWI subfamily of Argonaute genes, exhibits reduced brood size and temperature-sensitive sterile phenotype, implicating that it is required for germline maintenance. Additionally, we showed that RDE-1 interacts with trigger-derived sense and antisense siRNAs (primary siRNAs) to initiate RNAi, while several other Argonaute proteins, SAGO-1, SAGO-2, and perhaps others, functioning redundantly, interact with amplified siRNAs (secondary siRNAs) to mediate downstream silencing. Moreover, our analysis uncovered that another member of Argonaute gene family, ergo-1, is essential for the endogenous RNAi pathway. Furthermore, we built an eight-fold Argonaute mutant, MAGO8, and analyzed its developmental phenotype and sensitivity to RNAi. Our analysis revealed that the genes deleted in the MAGO8 mutant function redundantly with each other, and are required for RNAi and the maintenance of the stem cell totipotency.

Sexual reproduction is an evolutionary innovation that arose 1.2 billion years ago, and in that time, has allowed a rapid diversification of species outpacing that of asexually reproducing organisms. Successful sexual reproduction in animals requires the incredible coordination of complex genetic and behavioral factors; from the most fundamental levels of ensuring correct chromosome segregation and ploidy to the most complex of behavioral mating rituals, any failure can result in a complete loss of evolutionary fitness. In this thesis, I have explored the developmental programs that function to ensure somatic sex determination, sexual differentiation, and mating behaviors in C. elegans. C. elegans is an androdiecious nematode species that has been extensively characterized in regard to the sexual dimorphism of its development, nervous system, and behavioral outputs. Sex determination pathways are widely diverged across phyla, and C. elegans has coopted a Gli family transcription factor to serve as a cell autonomous global regulator of somatic sex determination. I investigated the expression of this transcription factor, tra-1, with cellular, subcellular, sex-specific, and temporal resolution in both sexes of C. elegans and found that it is dynamically regulated to control sex determination. In contrast to the upstream sex determination pathway, genes that control downstream sexual differentiation in animals display much higher functional conservation, and many of the regulators of sexual differentiation belong to a family of transcription factors known as the DMRT family. Downstream of the tra-1 global regulator, I found that the highly conserved DMRT family gene dmd-4 acts much more specifically in adult hermaphrodites to generate sexual dimorphism at the level of the phasmid sensory neurons PHA and PHB. Furthermore, the sexual dimorphism of DMD-4 is regulated post-translationally by a ubiquitin-binding domain that I also found to be functionally conserved in the human ortholog, Dmrt3. Although these transcription factors both demonstrate the high degree of genetic control that contributes to sex determination and sexual differentiation, I also described male-specific effects of early life stress on sexual dimorphic synaptic connectivity and behavior generated by the phasmid sensory neurons, indicating that sexual differentiation is also plastic to environmental cues encountered during the life of an organism. This thesis provides insight into how genetic pathways function at multiple levels to give rise to extensive sexual dimorphism in the soma of an animal, both globally and in regard to the development on individual cells, in addition to the ways in which these genetic pathways can be modified by environmental factors and organismal life history.

Cytokinesis is the physical division of one cell into two driven by an actomyosin contractile ring and positioned by signals from microtubules. This process is highly regulated spatially and temporally to ensure accurate division into two daughter cells. Here, I present work that builds upon our understanding of cytokinesis, focusing on the spatial requirements for protein function during cell division and midbody assembly. In Chapter 1, I present an introduction to cytokinesis and the cell and molecular mechanisms that govern the process. In Chapter 2, I present work I contributed to on the use of Upconverting nanoparticles for co-alignment of visible and infrared light on a light microscope. In Chapter 3, I present work developing a new microscopy technology called FLIRT (Fast Local Infrared Thermogenetics) that uses infrared light to inactivate fast-acting temperature sensitive protein function with subcellular precision and validate its use to study cytokinesis and cell fate signaling in the nematode Caenorhabditis elegans. In Chapter 4, I improve upon FLIRT technology by increasing its precision and demonstrate its use in studying the spatial regulation of key cytokinesis proteins including the actomyosin cytoskeleton in contractile ring constriction. The central spindle is an array of antiparallel overlapping microtubules that forms between the separating chromosomes in anaphase and is thought to serve as a signaling hub for cytokinesis. The central spindle is thought to become compacted during contractile ring constriction to form the dense midbody at the end of cell division. In Chapter 5, I investigate the requirements for central spindle microtubules in assembling midbodies in the C. elegans one-cell embryo. I present evidence that the CENP-F-like protein HCP-1 plays a primary role relative to its paralog HCP-2 in assembling the central spindle, and that the midbody can form independently of central spindle assembly. In Chapter 6, I discuss future directions for my work on both technology development and the mechanisms of cytokinesis. Through this work, I develop new technologies and hypotheses for how cytokinesis is spatially regulated within a cell, adding new complexity to our understanding of cell division.

In my thesis I aimed to contribute to the understanding of the mechanism of cytokinesis in C. elegans embryos. I wanted to analyze the relative contributions of different spindle parts – microtubule asters and the midzone - to cytokinesis furrow positioning. I developed a UV laser-based severing assay that allows the spatial separation of the region midway between the asters and the spindle midzone. The spindle is severed asymmetrically between one aster and the midzone. I found that the spindle provides two consecutive signals that can each position a cytokinesis furrow: microtubule asters provide a first signal, and the spindle midzone provides a second signal. The use of mutants that do not form a midzone suggested that the aster-positioned furrow is able to divide the cell alone without a spindle midzone. Analysis of cytokinesis in hypercontracile mutants suggests that the aster-positioned cytokinesis furrow and the midzone positioned furrow inhibit each other by competing for cortical contractile elements. I then wanted to identify the molecular pathway responsible for cytokinesis furrow positioning in response to the microtubule asters. To this end, I performed an RNAi screen, which identified a role for LET-99 in cytokinesis: LET-99 appeared to be required for aster-positioned cytokinesis but not midzone-positioned cytokinesis. LET-99 localizes as a cortical band that overlaps with the cytokinesis furrow. Mechanical displacement of the spindle demonstrated that the spindle positions cortical LET-99 at the site of furrow formation. The furrow localization of LET-99 depended on G proteins, and consistent with this finding, G proteins are also required for aster-positioned cytokinesis. (Anlage: Quick time movies, 466, 67 MB)

Intercellular communication is crucial during animal development and tissue maintenance to ensure that correct patterns of cell types are generated to meet the needs of the organism. During lateral specification, intercellular communication resolves cell fate decisions between equipotent cells, creating fate patterns that are biased by external factors in some contexts, but appear stochastic in others. The Notch signaling pathway mediates lateral specification; small differences in Notch activity are amplified by regulatory feedback loops to robustly differentiate cell fates based on relative levels of Notch activity. It is often unclear how noise in the environment is processed by cells to generate differences in Notch activity that can be translated into stochastic, but robust, cell fate outcomes. The nematode Caenorhabditis elegans contains a simple, Notch-mediated, stochastic lateral specification event; a small, random difference in Notch activity between two cells, the α cells, is amplified so that one α cell assumes Anchor Cell (AC) fate and the other assumes Ventral Uterine precursor cell (VU) fate. Two upstream factors bias the outcome of the AC/VU decision depending on the length of the time interval between the births of the α cells: the relative birth order of the α cells and the onset of expression of the transcription factor HLH-2. It is unknown how these factors create a difference in the relative Notch activity level between the two α cells, and limitations of existing Notch reporters have prevented the direct observation of Notch activity levels required for determining the relationships. In this thesis, I describe a genetically-encoded Sensor Able to detect Lateral Signaling Activity, or SALSA, which uses changes in nuclear Red:Green fluorescence to indicate Notch activity. I demonstrated that SALSA captures expected Notch activity patterns in four paradigms in C. elegans, encompassing both Notch homologs, and reports low levels of Notch activity that were predicted but undetectable with other Notch activity reporters. Using SALSA, I showed that the first-born α cell is able to develop an advantage in Notch activity prior to the birth of the other α cell when the time interval between α cell births is long, but the α cell that gains the Notch activity advantage is random with respect to birth order when the time interval between α cell births is short. These results agree with the current model of the AC/VU decision. I also describe Flexon, a method for the conditional activation of strong gene expression in specific cell lineages using a lox-stop-lox cassette encoded into an artificial exon flanked by two artificial introns. A flexon can be placed into the coding region of a gene to prevent translation of a functional gene product; gene expression is restored to specific lineages through expression of a tissue-specific Cre driver that excises the flexon. I show that flexon can be used to make bright, long-lasting, tissue-specific fluorescent lineage markers. I also showed that the flexon could be used for conditional activation of an endogenous gene by inserting a flexon into rde-1 to severely reduce RNAi activity and restore gene function in specific tissues using Cre drivers.

Elemental stoichiometry defines a critical understanding of the relationship between nutrient availability and usage throughout different levels of the biological community. We found there is a link between available phosphorus (P), cellular phosphorus, and nematode development as postulated by the growth rate hypothesis (GRH). I predicted that in a P-poor environment, cellular RNA concentrations would be lower than they are in P-rich environment, and thus the 18s rRNA expression level will have reduced. To most efficiently regulate the uptake of limited P, I predicted that nematodes in P-poor environments would decrease the number of copies of the 18s rRNA gene in their genome. I measured life history traits as well as rRNA gene expression and gene copy number. We found that elemental stoichiometry predicts evolutionary changes consistent with the Growth Rate Hypothesis. We sequenced and assembled a draft genome of P. murrayi. Although we expected to find genes responsible for stress tolerance, we hypothesized that in response to strong selection pressure associated with living in a simplified ecosystem, over time the genome of P. murrayi should have undergone significant decay (gene loss) relative to species in ecosystems structured more strongly by biotic interactions. We found significantly fewer genes in P. murrayi. To compare patterns of gene expression between two highly divergent Antarctic nematode species, we sequenced and assembled the transcriptomes of S. lindsayae and P. murrayi. Under laboratory conditions at 4˚C, S. lindsayae had significantly lower rates of gene expression but expressed a significantly larger number of genes. We speculate that the differences in gene expression are correlated with life history traits (developmental rates) while the differences in the number of genes expressed can be explained by their different genetic systems (S. lindsayae is amphimictic, P. murrayi is parthenogenic) and the soil environments to which they are adapted. Since we previously showed that differences in available P content can influence the evolution of gene expression via gene copy number, and that this ultimately influences growth rate, we wondered how much of this response is driven by genetics versus how strongly these patterns are driven by temperature. To better understand this, we maintained wild type populations of P. murrayi in P-rich and P-poor conditions at 5˚C, 10˚C and 15˚C in the laboratory for over 40 generations and sequenced the transcriptomes prepared from each treatment group. We found that nutrient levels played an important role in gene expression when the temperature is optimal for P. murrayi culturing and that temperature is more important in gene expression when the available P is limited. This work underscores the utility of using principles of elemental stoichiometry coupled with genomic and transcriptomics research tools to make and test predictions about life history evolution. The results of my work also inform inferences about the ways in which nutrient availability also drives the organization of trophic interactions and ultimately ecosystems.

Organisms are comprised of many cells with multiple distinct cell types, each of which must be decided precisely to ensure proper formation of a functional organism. In C. elegans, the basic helix-loop helix transcription factor HLH-2 is required for the specification of the anchor cell, or AC. The AC arises from a group of four somatic gonad cells, all of which initially express HLH-2. Two of the four cells, which we call β cells, lose AC competence early and instead become ventral uterine precursor cells, or VUs. We call the remaining two cells α cells. One α cell becomes the AC, while the other becomes a VU. Which α cell becomes the AC is random—50% of the time one α cell becomes the AC, while the other 50% of the time the other α cell becomes the AC. The choice of which cell becomes the AC and which becomes the VU is called the AC/VU decision, and occurs through reciprocal signaling by LIN-12/Notch and its ligand LAG-2/DSL. At first, both α cells express similar levels of lin-12 and lag-2. As the AC/VU decision progresses, the AC expresses higher levels of lag-2, and the VU expresses higher levels of lin-12. By this time, HLH-2 is only present in the specified AC, while it is post-translationally degraded in VUs. The mechanism by which HLH-2 is degraded and the consequences of disrupting its degradation on AC specification are unknown. In this work, we studied the function and regulation of HLH-2 during two stages of somatic gonad development. First, we used long-term fluorescence microscopy to visualize HLH-2 over the course of somatic gonad development. We found that HLH-2 expression begins in the parents of the α and β cells a consistent amount of time after their birth, and that the parent cell that first expresses HLH-2 almost always gives rise to the α cell that becomes the VU, while the second cell to express HLH-2 gives rise to the AC. This led us to study the effect of a loss of hlh-2 activity in the α and β cells. We generated an α and β cell-specific hlh-2(0) allele using genome editing tools and found that LIN-12 protein is not present in the absence of hlh-2 activity. Based on this discovery, we conceived a model where HLH-2 expression biases the first-expressing cell towards the VU fate by endowing it with an edge in lin-12 activity. Next, we focused on restriction of HLH-2 to the AC. Typically, HLH-2 protein is degraded in VUs, which we hypothesized was a crucial step in restriction of the AC fate to a single cell. We found that in a lin-12(0) background, HLH-2 is stabilized in VUs even when the resulting cell does not become an AC, indicating that lin-12 directly promotes HLH-2 degradation. This led us to search for a lin-12-regulated factor that targets HLH-2 for degradation in VUs. We identified seven ubiquitin-related genes whose depletion resulted in stabilized HLH-2 in VUs, but surprisingly did not cause an AC/VU defect. We suspect that HLH-2 degradation in VUs is one of multiple negative regulatory mechanisms that ensure the robustness of the AC/VU decision. The following research contributes new insights into how stochastic cell fate decisions amplify noise to ensure a consistent and reproducible outcome.

Candida albicans is the most common human fungal pathogen. The relationship between C. albicans and Enterobacter bacteria have yet to be explored. The hypothesis of this study is that C. albicans and both E. aerogenes and E. cloacae have a positive relationship and work together to infect the host. In this study, the physical cell-to-cell interaction, molecular components of said interaction, and the impact of the interaction on a live organism were explored. Results indicate that Enterobacter adheres to C. albicans and inhibits growth with unidentified secreted molecules. Als1p has potential involvement in the attachment of E. cloacae to C. albicans. Out of 480 E. cloacae mutants, 6 showed reductions in attachment to C. albicans. The presence of C. albicans in C. elegans may lead to less Enterobacter colonization. Future work involving this interaction should strive to identify the Enterobacter secreted molecules and genes necessary for their production.