Fibroblast Growth Factors (FGFs) are secreted molecules that activate the RAS/mitogen-activated protein kinase (MAPK) signaling pathway to establish dorsal polarity, maintain the isthmic organizer, and assure proper ventricle formation in the zebrafish. The mechanism of FGF regulation of these processes and the transcription factors involved are still unclear. Expression of the zebrafish PEA3 family of ETS transcription factors, Etv5, Erm, and Pea3, is responsive to FGF signaling, and these factors are likely transcriptional effectors of this pathway. I have determined the role of PEA3 ETS factors in FGF signaling and gene regulation through gain- and loss-of-function studies. Ectopic expression of a constitutively activated form of Etv5 induced FGF target transcripts, dual specificity phosphatase 6 (dusp6) and similar expression to fgfs (sef). The simultaneous knock-down of Etv5, Erm, and Pea3 produced phenotypes reminiscent of the fgf8 mutant, including the disruption of the mid-hindbrain boundary, diminished cardiac progenitors, and left/right patterning defects. Furthermore, the expression of FGF target genes was abolished in Etv5/Erm/Pea3 depleted embryos. To understand how FGF signaling and PEA3 ETS factors control gene expression, the transcriptional regulation of dusp6 was studied in mouse and zebrafish. Conserved Pea3/ETS binding sites were identified within the dusp6 promoter, and reporter assays show that one of these sites is required for dusp6 induction by FGFs in both species. In addition, I demonstrated the interaction of PEA3 ETS factors with the dusp6 promoter both in vitro and in vivo. These results revealed the requirement of ETS factors in transducing FGF signals in developmental processes.

Tbx6 is a member of the T-box family of transcription factors that is expressed within the primitive streak and presomitic mesoderm of the developing mouse embryo and is critical for the patterning and specification of the paraxial mesoderm. This work has investigated how Tbx6 interacts with other T-box transcription factors both in its endogenous expression domain and when it is ectopically expressed. To this end, we have used a combined genetic, transcriptional, and biochemical approach to investigate potential competition between Tbx6 and T, a T-box transcription factor co-expressed with Tbx6 in the primitive streak and tailbud. Additionally, we have developed a 3-component transgenic system to ectopically express Tbx6 outside of its endogenous domain, driving expression in the formed somites and limb buds. Three-component embryos displayed vertebral, rib, and limb anomalies resembling those in Tbx15 and Tbx18 null embryos. We showed that ectopic Tbx6 in these embryos could compete with Tbx15 and Tbx18 at the level of DNA binding. This suggested that the dynamic interplay of co-expressed T-box transcription factors might contribute to observed congenital birth defects resulting from heterozygous loss of a particular family member. While this work focused on one specific family of transcription factors, it implies that transcription factor family members that are related via a conserved DNA binding domain may also compete for downstream targets in vivo and that this then contributes to the overall developmental dynamics of the organism.

The existence of distinct species of life is generally explained by the genetic process of reproduction without recombination between populations and/or the ecological process of adaptation to different environments. Both processes affect prokaryotes, and have shaped existing genomes. Here, we use comparative genomic techniques to evaluate the dynamics of divergence among species of the Enterobacteriaceae. Bacteria such as Escherichia coli preferentially acquire allelic variants from closely related organisms (i.e. other E. coli) rather than from more diverged bacteria. Ecological differences between donor and recipient affect the probability of allelic variants becoming fixed across the recombining population. We examine the history of recombination among groups of genomes that no longer recombine with each other, but retain sufficient conservation of ancestral nucleotide sequences to allow recombination to be inferred. From these analyses, we conclude that substantial levels of recombination occurred between E. coli and diverging lineages even after some regions of the genomes had acquired many nucleotide differences. We identify two evolutionary radiations leading to E. coli where the disparity among loci confounds the phylogenetic relationships among species, as evidenced by topological incongruence among gene trees. The forces affecting recombination, reflected in both pairwise divergence and topologically informative sites, vary across regions of the genome measuring tens of kilobases. To examine the relationship between ecological differentiation and genetic recombination, we characterize differences that could be responsible for ecological differentiation among these species. Some of the loci with the most apparent functional differences (i.e. the gain and loss of genes) are associated with the greatest levels of sequence divergence between species, consistent with the hypothesis that ecological divergence interferes with homologous recombination, and therefore drives sequence divergence and genetic isolation. To investigate the role of more subtle ecological differentiation, we develop a statistical framework to evaluate codon usage bias of each protein-coding gene, taking into account the stochastic balance between codon selection, which is driven by the need for high expression, and mutational biases. This tool will be useful in future studies examining codon selection as contribution to diversification among the ecologically diverse species of Enterobacteriaceae.

Signal transduction pathways rely on proper protein activation and regulation to elicit appropriate responses to extracellular cues. Kinases within these pathways are commonly activated through a multi-step mechanism, which involves phosphorylation. Additionally, regulatory modifications can occur to modulate signals, ensuring that the proper intensity and duration of signaling is achieved in the correct context. I have identified two modes of regulation of the Drosophila JNKKK, Slipper. First, investigation into kinase domain phosphorylation reveals that conserved putative phosphorylation sites are required for Slpr function. Site-directed mutagenesis converting three serine/threonine residues to alanines (SlprAAA) renders the protein inactive, and SlprAAA behaves as a dominant negative during several Slpr-mediated processes. Transgenic flies expressing SlprAAA display phenotypes associated with a loss of JNK signaling such as dorsal open embryos that lose JNK pathway marker gene expression, cleft thorax indicative of a loss of JNK signaling during thorax closure, and a failure to rescue slpr mutants. Importantly, double alanine mutant analysis (TST mutated to AAT, ASA, and TAA) reveals that T295 is crucial for Slpr signaling as the two double mutants that contain alanine mutations at that residue are nonfunctional while SlprAAT retains some Slpr function. An additional phosphorylation site outside of the kinase domain was identified at a conserved MAPK consensus site, PXSP. Though flies expressing a non-phosphorylatable PXAP develop normally and have no defects in Slpr-dependent functions in morphogenesis, both embryos and adults are sensitive to heat shock. Conversely, a phospho-mimetic version, PXEP, confers thermoresistance. Biochemical assays implicate the Jun kinase, Bsk, as the MAPK required for this signaling to PXSP. These results suggest that Bsk phosphorylates Slpr in a positive feedback loop during heat stress to maintain homeostasis. Together, these experiments demonstrate the need for Slpr phosphorylation in two circumstances. Not only is phosphorylation required to maintain Slpr-dependent JNK signaling in morphological processes through proper protein activation, but it is also essential in a context-specific manner for a previously unidentified role of Slpr and the Drosophila JNK pathway in heat shock response.

Cancer cells typically have unstable genomes and ploidy, and these changes play an important part of malignant transformation in most tumorigenesis models. I observe that cancer cells can increase ploidy by failure of cell division producing a single daughter with a doubled genome. Cytokinesis failure usually is followed by the formation of multipolar spindles in the tested cell lines. Some of these cells are still capable of dividing and the daughters of multipolar division have an even greater chance of being multinucleated and multicentrosomal due to failure of cytokinesis and thus likely enter another round of multipolar division. However, some of the daughters inherit a single nucleus and centrosome, suggesting that they may escape a cycle of multipolarity and give rise to centrosomally-stable clones. I also show here that failure of cytokinesis in cancer cells is associated with cytoskeleton abnormalities. One type of defect is caused by decreased phosphorylation of the myosin regulatory light chain, a key regulatory element of cortical contraction during cell division. Reduced phosphorylation is often associated with high expression of myosin phosphatase and/or reduced myosin light chain kinase (MLCK) levels in a variety of cancer cell lines. When myosin light chain phosphorylation is restored to normal levels by phosphatase knockdown, the mitotic defects of malignant cells, including cytokinesis failure, multinucleation, and multipolar mitosis are all markedly reduced. Both overexpression of myosin phosphatase and inhibition of the MLCK can recapitulate the multinucleation in nonmalignant cells. These results show that ploidy defects in tumor cells can be caused by deficiencies in myosin light chain phosphorylation resulting from high expression of myosin phosphatase and low activity of MLCK. GpIb¦Á, a membrane glycoprotein, which is widely upregulated in cancer cells, has been shown here for the first time to play a role in cytokinesis, possibly also through regulating cytoskeleton remodeling. The overexpression of GpIb¦Á in p53-knockdown primary cells induces cytokinesis failure, tetraploidization, genomic instability and tumorigenesis. In addition, knockdown of GpIb¦Á reduces chromosome segregation defects in cancer cells. Together, these observations support a model that cytoskeleton-defect-mediated cytokinesis failure plays a major role influencing the chromosomal instability of cancer cells.

Salmonella enterica, a major causative agent of gastrointestinal illness, exhibits a host-specific pattern of infection, with certain serovars predominantly infecting particular hosts. Extensive variation is observed at the Salmonella rfb locus, which makes up serovar-defining O-antigen. Unlike other pathogens, this diversity cannot be explained by selective pressure from the host immune system. Here, I implicate the O-antigen to the physiological basis for escape from protozoan predators. These predators have differential feeding preferences on Salmonella and may be responsible for maintaining O-antigen diversity, controlling which serovars are able to survive predation to potentially cause disease. I demonstrated that the O-antigen plays a strong role in mediating predator escape and uncovered a trade-off that may exist between O-antigen identity and chain length regulation in response to the dual selective pressures of evading host intestinal predators and successful interaction with the host immune system. To complete these experiments, I developed two new techniques: a) genetic manipulation of non-Typhimurium Salmonella and b) multicolor flow cytometry for assessment of microbes in natural, complex environments. These results link variation at virulence loci to environmental selective pressures other than the host immune system and provide an explanation for the role of the rfb locus in the fragmented speciation process in Salmonella.

The conserved Paf1 complex is important for proper gene expression in both yeast and humans. The Paf1 complex has been shown to repress the transcription of numerous genes. However, the mechanism by which the Paf1 complex mediates transcriptional repression remains largely unstudied. Here I use ARG1 as a model gene to investigate transcriptional repression by the Paf1 complex in Saccharomyces cerevisiae. Interestingly, I found that Paf1 complex-dependent histone modifications that are normally associated with active transcription are enriched on the ARG1 coding region and contribute to repression. The Rtf1 subunit of the Paf1 complex appears to mediate ARG1 repression primarily through histone H2B ubiquitylation and histone H3 K4 methylation. However, Paf1 has repressive functions aside from these histone modifications. Interestingly, occupancy of the activator Gcn4 is increased at the ARG1 promoter in paf1Δ cells, resulting in ARG1 derepression that is dependent on the histone acetyltransferase Gcn5 and histone H3 acetylation sites. Together my results suggest that Paf1 mediates ARG1 repression by preventing Gcn4 recruitment to the ARG1 promoter and subsequent histone H3 acetylation. I found that Paf1 does not alter nucleosome occupancy at the ARG1 promoter. However, I detect antisense transcription in the ARG1 promoter that positively correlates with ARG1 sense transcription. Interestingly, Paf1 prevents antisense transcription from traversing the ARG1 promoter, representing a potential mechanism by which the Paf1 complex controls promoter accessibility and ultimately ARG1 expression. Given these results, I hypothesize that the Paf1 complex mediates ARG1 repression partially by facilitating histone modifications that are refractory to ARG1 transcription and partially by inhibiting antisense transcription which controls promoter accessibility. Importantly, events that I observed at my model gene, ARG1, are demonstrated at other Paf1 complex-repressed genes.

Polyomaviruses (PyVs) are ubiquitous DNA viruses that are not generally associated with pathogenicity. However, in immunosuppressed populations, PyVs can cause diseases, including BK Virus associated nephropathy, hemorrhagic cystitis and progressive multifocal leukoencephalopathy. There is currently no PyV specific inhibitor for these diseases. Because all PyVs express a conserved large T antigen (TAg) that is essential for viral replication, I hypothesized that inhibitors of the model TAg from Simian Virus 40 (SV40) would inhibit the replication of PyVs in general. TAg has multiple essential activities, but my work focused on the ATPase activity of TAg, which provides the helicase activity during viral replication. Two high throughput screens were performed on purified TAg to identify inhibitors of TAg ATPase activity. In the first screen, a quinaldine red screen of the MS2000 library of commercially available compounds identified the FDA approved compounds bithionol and hexachlorophene as inhibitors of the ATPase activity of TAg. The structure activity relationship of these bisphenols was refined, and the results suggested that inhibition of TAg was unique to these compounds, and is not a general feature of bisphenols. The compounds also inhibited SV40 DNA replication and infection, and are the most potent SV40 inhibitors reported. Surprisingly, bithionol and hexachlorophene inhibit replication of the clinically relevant PyV BK Virus (BKV), but not to the same extent. In the second screen, which employed the ADP Hunter assay and the NIH Molecular Libraries Probe Centers Network (MLPCN) compound library, three scaffolds of interest were identified. However, all three had limited potency when characterized in other in vitro assays. Chemical refinement of these scaffolds identified another bisphenol that inhibits TAg activity and TAg dependent DNA replication in vitro. Unfortunately, the known cytotoxicity of this compound limits its use as a therapeutic. Although I have not yet identified a PyV specific inhibitor that would be a suitable therapeutic, this work supports my overall hypothesis: small molecules that inhibit the ATPase activity of TAg reduce viral DNA replication, and viral infection. My work has also provided valuable insights to design future screens to identify inhibitors of PyV replication.

This work focuses on adducing general principles applicable to site-specific protein-DNA interactions by linking function to structural, thermodynamic and dynamic properties. The interaction of EcoRI endonuclease with specific, miscognate, and nonspecific DNA sequences is used as a model for protein-DNA interactions. We use four pulse Double Electron-Electron Resonance (DEER) Electron Spin Resonance (ESR) experiments to map distances and distance distributions between nitroxide spin labels placed at positions within the arms and the main domain of the EcoRI homodimer. These experiments show that the DNA occupies a similar binding cleft and is enfolded by the arms of the enzyme in all three classes of EcoRI-DNA complex. Additionally, changes in dynamics of main domain and arm residues within the three complexes were explored using Continuous Wave (CW) ESR spectroscopy. A position adjacent to a protein-phosphate contact shows decreased mobility relative to other arm residues that are not at the protein-DNA interface. Signal from this position shows the largest amount of an immobile component in the specific complex, progressively less immobile in the miscognate and nonspecific complexes. This fits with distribution breadths from DEER-ESR spectra and biochemical evidence that the nearby phosphate contact is made only in the specific complex. Residues at other positions show mobilities that are in agreement with our hypothesis that residues in the arms would be relatively more mobile than those in the main domain Using Electron Spin Echo Envelope Modulation (ESEEM) ESR we show that the paramagnetic Cu2+ ion is coordinated by an imidazole nitrogen. These experiments thus reveal a novel metal ion binding site. DEER measurements of distances between Cu2+ ions and Cu2+-nitroxide distances in the homodimeric EcoRI-DNA complex establish that the Cu2-coordinating residue is histidine 114, which is proximal to but not at the active site. This is consistent with our biochemical studies that show that Cu2+ cannot replace Mg2+ as a catalytic cofactor but instead completely inhibits EcoRI cleavage. We also use isothermal titration calorimetry (ITC) to directly determine a stoichiometry of two Cu2+ ions bound per homodimeric EcoRI-DNA complex; that is, each histidine 114 coordinates one Cu2+ ion.

Alternative pre-mRNA splicing is one of main contributors to protein diversity. A fundamental unanswered question is how splicing decision is altered by signal transduction. The C1 exon of the glutamate NMDA receptor NR1 subunit is alternatively spliced in a tissue-specific manner. It is predominantly included in the forebrain, and predominantly skipped in the hindbrain. Importantly, the C1 exon encodes the C1 protein cassette containing an ER retention signal that regulates the surface delivery rate of assembled receptors, and a calmodulin binding site that plays critical roles in learning and memory via regulating Ca2+ flux. Recently, it has been reported C1 exon skipping can be induced by cell depolarization. However, the molecular links mediating the depolarization on the cell membrane and alternative splicing in the nucleus largely remain unclear. Here I have investigated the roles of Ca2+ signaling, PKA and hnRNP A1 in mediating depolarization induced C1 splicing silencing. Specifically, using Ca2+ imaging, I have found the KCl induced sustained Ca2+ influx is associated with C1 splicing silencing. In addition, I have demonstrated that CaMK IV and PKA regulate C1 splicing in a hnRNP A1 dependent manner. Finally, I have shown that hnRNP A1 is recruited to the C1 splicing site upon depolarization in neurons. The results here have demonstrated signal transduction from depolarization to splicing factors, showing how splicing decision can be altered by external stimuli.