Thesis (Ph. D.)--University of Toronto, 2008. / Includes bibliographical references.

<p> Behavior in higher eukaryotes is a complex process which integrates signals in the environment, the genetic makeup of the organism, and connectivity in the nervous system to produce extremely diverse adaptations to the phenomenon of existence. Unraveling the subcellular components that contribute to behavioral output is important for both understanding how behavior occurs in an unperturbed state, as well as understanding how behavior changes when the underlying systems that generate it are altered. Of the numerous molecular species that make up a cell, the regulation of messenger RNAs (mRNAs), the coding template of all proteins, is of key importance to the proper maintenance and functioning of cells of the brain, and thus the synaptic signals and information integration which underlie behavior. RNA binding proteins, a class of regulatory molecules, associate with mRNAs and facilitate their maturation from pre-spliced nascent transcripts, their stabilization and degradation ensuring appropriate levels are maintained, as well as their translation and subcellular compartmentalization, which ensures that proteins are translated at the appropriate level and in the places where they are required to fulfill their cellular functions. Our laboratory identified polymorphisms in the gene coding for the CUGBP and ELAV-like Factor 6 (CELF6) RNA binding protein to be associated with Autism Spectrum Disorder risk in humans. ASD is a spectrum of disorders of early neurodevelopment which present with lowered sociability and communication skills as well as restricted patterns of interests. When expression of the <i>Celf6</i> gene was ablated in mice, we found that they exhibited reductions to early communication as well as altered aspects of their exploratory behavior. In this dissertation, I explore the communication changes in young mouse pups with loss of CELF6 protein and identify that despite being able to produce vocalization patterns similar to their wild-type littermates, they nevertheless exhibit reduced response to maternal separation. Despite a history of literature on other CELF family proteins, the functions of the CELF6 protein in the brain have not been previously described. I provide characterization of the mRNA binding targets of CELF6 in the brain, and show that they share common UGU-containing sequence motifs which has been noted for other CELF proteins, and that CELF6 binding occurs primarily in the 3' untranslated regions (3' UTR) of mRNA. I hypothesized that this mode of interaction would result in regulation of mRNA degradation or translation efficiency as 3' UTR regions are known for providing binding sites for numerous regulators of such processes. In order to answer this question, I cloned sequence elements from the 3' UTRs of target mRNAs into a massively parallel reporter assay which has enabled me to test the effect of CELF6 expression on hundreds of binding targets simultaneously. When expressed in vitro, I found that CELF6 induced reduction to reporter library levels but exhibited few effects on translation efficiency, and I was able to rescue effects to reporter abundance mutation of binding motifs. Intriguingly, like CELF6, CELF3, CELF4, and CELF5 were all able to produce the same effect. CELF5 and CELF6 both showed similar, intermediate repression of reporter library mRNAs, while CELF3 and CELF4 exerted the strongest levels of repression. The level of repression under these conditions was somewhat predicted by number of motifs present per element, however a large amount of the variance in reporter levels is still unexplained and a mechanism for CELF6's action is unknown. Nevertheless, the work I present in this dissertation shows that CELF6 and other members of its family are key regulators of mRNA abundance levels which has direct implications to downstream consequence in the cell. As several of CELF6 binding target mRNAs are known regulators of neuronal signaling and synaptic function, the information I present is crucial for future experimentation. This work well help lead us to understand how behavior is altered when this protein is absent, along the way uncovering important mechanistic steps connecting the molecular landscape of cells to the behavior of organisms.</p><p>

The budding yeast (S. cerevisiae) centromere component, Cse4p is an evolutionarily conserved histone H3-like protein, with homologues identified in fission yeast, worm, fly and human. All histone H3-like proteins have C-terminal histone fold domains (HFD) that are highly similar to the HFD of H3, but carry very different N-termini with unknown functions. The Cse4p N-terminus contains 135 residues, with a large portion of charged amino acids and a high concentration of serines within the first 22 residues. Based on the current model that suggests that Cse4p replaces H3 in a specialized centromeric nucleosome, the Cse4p N-terminus would extend out from the putative Cse4p-nucleosome and may play a variety of roles in centromere function. To elucidate the function of the Cse4p N-terminus, we conducted two comprehensive and systematic mutagenesis studies involving alanine scanning and sequence deletions, and we defined a 33-amino acid domain that is essential for cell viability and chromosome segregation. This essential N-terminal domain (END) has functions distinct from that of the HFD as demonstrated by interallelic complementation between cse4 END and HFD mutant alleles and heterodimer formation of END-HFD mutant proteins. Mutating all the potential posttranslational sites in the END indicates that the END function does not require posttranslational phosphorylation or acetylation. Genetic studies involving dosage suppression, synthetic lethality and two-hybrid analysis reveal that the END interacts with the Ctf19p/Mcm21p/Okp1p kinetochore complex. These results are consistent with the current Cse4p-nucleosome model. Although Cse4p has an HFD resembling that of H3, unlike H3, Cse4p exclusively localizes at the centromere. An important question is whether the N-terminus of Cse4p is responsible for the specific centromere targeting of the protein. Lethal Cse4p proteins lacking regions of the N-terminus can localize to the centromere in the presence or absence of wildtype Cse4p as determined by chromatin immunoprecipitation. In contrast, some lethal Cse4p HFD mutant proteins as well as chimeric proteins consisting of the Cse4p N-terminus fused to the HFD of either H3 or the Cse4p human homologue, CENP-A, fail to localize to the centromere. We conclude that the N-terminus of Cse4p is not required for centromere targeting of the protein and that the Cse4p HFD is necessary and sufficient to confer centromere localization.

Hedgehog (Hh) signaling is important for patterning and cell differentiation within many embryonic tissues. Hh is expressed in the notochord and the floor plate and acts as a morphogen in patterning of the ventral central nervous system (CNS), pituitary gland, somites, pancreas, and many other tissues. Gli transcription factors act as main mediators of Hh signaling in vertebrates. Their unique and overlapping functions lead to differential outcome of Hh signaling in different cells and tissues. The goal of my dissertation is to determine how zebrafish Gli proteins work together to transduce Hh signals and to activate or repress the transcription of Hh target genes. I have shown that Gli1 acts only as an activator of Hh signaling similar to what is known from other species, while Gli2 and Gli3 act as both activators and repressors. Gli2 functions as a co-activator of Gli1 in subset of ventral cells in the posterior diencephalon and in the adaxial cells in the embryonic trunk. In contrast, activator role of Gli3 overlaps with Gli1 throughout the ventral CNS during early development of the embryo. Later, GO represses Hh targets in the dorsal spinal cord, but not in the forebrain. In contrast, Gli2 represses genes in both, the dorsal telencephalon and the spinal cord. I also demonstrate that Gli3 repressor function is temporally regulated by active Hh signaling, unlike Hh independent Gli2 repressor function. Thus, my detailed analysis of zebrafish Gli functions reveals complex interactions between Gli proteins in embryonic patterning. I also studied a novel zebrafish mutation umleitung (uml) that was identified because of defects in axon guidance and neural patterning in the ventral forebrain. My detailed phenotypic analysis showed that the uml mutation disrupts Hh signaling and forebrain patterning. I have genetically mapped uml on zebrafish chromosome 24 near the Zmarker z10372. My linkage analysis indicates there are no known components of the Hh signaling cascade in the uml genetic region. This leads to the exciting possibility that uml may encode a previously undefined regulator of Hh signaling. I have initiated a genomic walk toward finding a gene that encodes uml.

Removal of protein activity by genetic mutation or pharmacological inhibition has been used extensively to understand the normal function of a protein. However, null mutations eliminate gene function in all cells and pharmacological agents can diffuse through tissues to have similar global effects that can obscure the physiological function of a protein. This is a particular problem when studying proteins that function in many cell types or that have different cell-specific activities. The most direct strategy to study the function of a protein is to reduce or eliminate its activity only in specific cell types, rather than in all cells of an organism. The idea of targeting gene knockdown to specific cell types or to individual cells is not new and many strategies aim to do just this. However, these strategies result in variable knockdown efficiencies and can have silencing effects in neighboring cells and therefore knockdown is never cell-specific. We developed a novel method to knock down the expression of any gene and to restrict this knockdown to specific cell types in C. elegans. In this method we replaced endogenous genes with single copy integrated transgenes containing an engineered sequence tag that introduces premature stop codons (PTCs) into transgene mRNA. This tag causes the natural stop codon to be recognized as a PTC by the host's nonsense-mediated decay (NMD) machinery and does not disrupt gene function. In NMD-competent animals, a PTC-containing transgene is degraded and in NMD-defective animals, a PTC-containing transgene is expressed. Therefore, the expression of PTC-containing transgenes can be controlled by cell-specific activation of NMD. Using this technique, we replaced two endogenous genes with PTC-containing transgenes and directed degradation of their mRNA to specific cell types by restoring NMD activity in these cells. The single copy transgenes were expressed at levels comparable to the endogenous genes and were knocked down to ∼10% of endogenous by NMD, resulting in both global and cell-specific null-like phenotypes. This knockdown strategy can be used to cell-specifically knock down essentially any gene in the C. elegans genome and should provide new insights into understanding protein function.

This dissertation focuses on dioecious angiosperms with a genetic system of sex determination based on a pair of heteromorphic sex chromosomes. Silene latifolia and S. dioica (Caryophyllaceae), with their X/Y mechanism and heterogametic males, have become model organisms for the study of genetic sex determination in angiosperms. Random amplified polymorphic DNA (RAPD) markers provide a valuable tool for the study of the genetic nature of the male determining Y chromosome. We first embarked on a search for Y chromosome RAPD markers using the breeding technique of bulked segregant analysis, obtaining 101 Y chromosome markers which together, were shown to characterize the two species. Genetic mapping placed one of these Y chromosome molecular markers in the pseudoautosomal region of the sex chromosomes. This finding provides a tool for the study of recombination rates among sex chromosomes and relative rates of evolution of X and Y chromosomes. Finally, we investigated the molecular nature of a highly conserved Y chromosome marker and looked for homologous sequences in other related dioecious and hermaphroditic species. The marker of choice had been found to be conserved across males from four species of Silene, section Elisanthe. It was cloned, sequenced and analyzed by Southern hybridization. This Y chromosome marker is a high copy sequence that shares homology to female DNA. Forward and reverse primers were designed to amplify the male specific band only. The amplification pattern of the resulting sequence characterized amplified region (SCAR) in related taxa provide evidence for: (1) a common ancestry of the Y chromosomes of dioecious Silene, section Elisanthe, (2) a different branch of the evolution of dioecy in section Otites, (3) the distant relationship of the hermaphrodite S. noctiflora to the dioecious members of section Elisanthe and (4) homology between the Y chromosome of dioecious Silene dioica and autosomes of hermaphroditic S. flos-cuculi.

The regulation of genes involved in proliferation and cell cycle control plays a critical role in normal development and differentiation. Aberrant expression of genes promoting proliferation (oncogenes) or loss of genes involved in restraining cell growth (tumor suppressors) can result in cancer. The p53 tumor suppressor protein has been demonstrated to play a critical role in both tumorigenesis and normal developmental processes. The mdm2 proto-oncogene can regulate the activity and stability of p53 protein. This suggested that mdm2 functions in development and tumorigenesis through p53-dependent mechanisms. However, mdm2 has also been shown to interact with factors involved in the regulation of cell cycle control, transcription and ribosome biosynthesis. The aim of this dissertation was to examine whether mdm2 expression during breast tumorigenesis and during normal mammary gland development in the mouse was regulated by p53-dependent or -independent mechanisms. The first segment of this work involved the analysis of alternative splicing of mdm2 mRNA during breast tumorigenesis in mice and humans. The second element of this dissertation examined the role of p53 in regulating the expression of mdm2 mRNA in adult tissues from the mouse and during normal mammary gland development in the mouse. Results from these experiments demonstrated that truncated mdm2 mRNA are expressed in mouse and human breast tissues. These truncated transcripts are predicted to code for mdm2 proteins that have lost C-terminal sequences involved in regulating proteolytic cleavage of mdm2 itself and in targeting p53 for ubiquitin mediated proteosomal degradation.

Synaptic transmission between a neuron and its target is crucially dependent upon the precise spatial arrangement of proteins in the pre- and postsynaptic apparatus. PDZ domain-containing proteins such as the Drosophila tumor suppressor Discs-Large (DLG) play critical roles in synapse maturation by regulating the assembly of synaptic protein complexes. DLG is composed of a number of modular domains, including three PDZ-domains, an SH3 domain and an enzymatically inactive Guanylate Kinase-like (GUK) domain. Previous studies have shown that the PDZ domains of DLG mediate clustering of Shaker K+ channels and of the cell adhesion molecule Fasciclin II. However, the function of the GUK domain has been unclear. To understand the role of the GUK domain, we carried out a yeast-two hybrid screen for interacting partners of the DLG GUK domain. This screen lead to the identification of a novel synapse-associated protein, GUKHolder (GUKH). GUKH is a 1044 amino acid protein with a molecular weight of 110 kDa. Its sequence includes a GUK-holding domain, a region homologous to the C-terminal of the long isoform of Kelch, a WH1-like domain, and a PDZ-domain binding motif. These latter two features suggest that GUKH may interact not only with DLG, but also with other proteins, including proteins containing PDZ domains. GUKH is expressed at the larval neuromuscular junction and at epithelial cell borders in partial colocalization with DLG. Further, DLG can be co-immunoprecipitated with GUKH from Drosophila extracts, indicating an in vivo interaction between the two proteins. GUKH has also been shown to interact directly with SCRIBBLE (SCRIB), another synaptically expressed PDZ-domain protein known to have a genetic interaction with dlg in epithelial tissues. Synaptic SCRIB immunoreactivity is mislocalized in both gukh and dlg mutants. gukh, scrib, and dlg mutants all exhibit synaptic bouton defects at the ultrastructural level. These data indicate that all three proteins are required for proper synapse maturation, and support a model that the three proteins exist in a tripartite complex, with GUKH forming a link between the other two proteins, and further, between the protein scaffolds organized by the two proteins.

An investigation of the mechanism responsible for genetic hypervariability in Halobacterium halobium gas vesicle synthesis was conducted. Four partially vacuolated mutants (Vac$\sp{\delta-})$ H. halobium mutants were analyzed by Southern hybridization, cloning, and DNA sequence analysis. In each mutant a different halobacterial insertion element was responsible for the observed phenotype. The insertions mapped upstream of the H. halobium gvpA gene. DNA sequence analysis of the 5$\sp\prime$ and 3$\sp\prime$ regions of gvpA revealed 10 open reading frames; gvpD, E, F, G, H, I, J, K, L, and M; 5$\sp\prime$ to the gvpA gene in the opposite strand and two open reading frames, gvpC and N, in the region 3$\sp\prime$ to gvpA with the same transcriptional orientation as gvpA. A study was conducted to determine if the products of the gvpA, gvpC, gvpD, gvpE, gvpF, gvpJ and gvpM genes could be detected in purified H. halobium gas vesicles or whole cell lysates using immunological techniques. To do so, LacZ-Gvp fusion proteins were produced in E. coli and used to immunize rabbits. The antisera and protein-A column purified antibodies were used in immunoblot analysis of purified gas vesicles and cell lysates. The antiserum directed against the LacZ-GvpC fusion protein was successful in identifying a protein present in both purified gas vesicles and whole cell lysates, and this indicates that the gvpC gene product is a structural gas vesicle protein. Techniques were developed to allow for genetic analysis of gas vesicle synthesis in H. halobium. An H. halobium/E. coli shuttle vector, pJHGV3, which contains the gvpA gene cluster was constructed. Transformation of Vac$\sp-$ H. halobium strains, in which the gvpA gene cluster is deleted, with pJHGV3 resulted in complementation of gas vesicle synthesis. Methods were developed to allow non-polar mutations to be introduced into gvp genes present on pJHGV3. A plasmid containing a disruption of the region 3$\sp\prime$ to the gvpN gene, pJHGV33$\sp\prime$::$\kappa,$ was constructed and used to transform Vac$\sp-$ deletion mutants. Resulting transformants were Vac$\sp+$ indicating that there are not additional contiguous gvp genes downstream from gvpN. Together these techniques will provide useful tools in further analysis of the gas vesicle structure and its assembly.

No description available.