Understanding the binding specificity of transcription factors is an important step towards accurate computational prediction of regulatory sequences governing gene expression. Higher-throughput binding site characterization methods have long been available in the laboratory for the study of protein-DNA interactions in solution or upon a surface. In this thesis a new method is introduced for characterization of inducible regulatory sequences in living cells based on construction and analysis of promoter-reporter gene plasmids. Spiked oligonucleotides are used to generate libraries of regulatory sequences with subtle variations from a known regulatory element. Screening of the library in cell culture for the capacity of the mutated sequences to mediate expression provides a diverse collection of responsive and non-responsive sequences to aid in understanding the sequence requirement for an inducible transcription factor binding site. We apply the methodology to the study of antioxidant responsive elements, the target sites of the Nfe212 transcription factor. These target sequences commonly found in the promoters of detoxification genes modulate gene expression in response to a diverse array of chemicals. The variants serve as a primary screen for future targeted mutational analysis to further characterize context-specific sequence requirement in the ARE and/or interdependence between positions. Moreover, a transcription factor binding profile can be generated from functional ARE variants in the library screen. Such an ARE profile performs as well as standard profiles based on bona fide ARE sequences drawn from the scientific literature. / Medicine, Faculty of / Medical Genetics, Department of / Graduate

No description available.

The homeodomain transcription factor family of Cdx genes, Cdx1, Cdx2 and Cdx4, are known to play essential roles in many developmental processes including neural tube closure, axial elongation, hematopoiesis and gastrointestinal patterning. In the adult, Cdx1 and Cdx2 are both expressed strictly in the adult intestinal epithelium, but their functions and mechanisms of action at this stage are poorly understood. Cdx transcription factors have also been reported to be lost in intestinal cancers. To circumvent early lethality, a conditional loss of function strategy was used to inactivate Cdx2 in the adult intestinal epithelium. These conditional mutants were crossed to Cdx1-/- mice to examine potential functional compensation between these family members as well as into APC(min/+) mice to study their role in tumorigenesis. Using these models, I have found that Cdx2 regulates adult intestinal homeostasis and differentiation in the small intestinal epithelium, while both Cdx1 and Cdx2 contribute to colon homeostasis. Furthermore, Cdx transcription factors are tumor suppressors in the development of Wnt-induced colorectal cancer, and impact several pathways including TGF-β and Eph-ephrin signaling. Finally, Cdx2 regulates Eph-ephrin signaling through direct activation of the Notch pathway. Altogether, this study underscores critical roles and mechanisms of action for Cdx members in the adult intestine and in intestinal tumorigenesis.

ERM est un facteur de transcription de la famille ETS appartenant au groupe PEA3 qui joue un rôle important dans divers processus biologiques dont la tumorigenèse mammaire. La régulation de son activité transcriptionnelle nécessite des modifications post-traductionnelles et des interactions avec des partenaires. Nous avons mis au point différentes techniques de chromatographie d’affinité dans le but d’identifier de nouveaux partenaires d’ERM. Parmi ceux-ci, trois ont été confirmés comme partenaires directs d’ERM : la protéine CoAA (CoActivator Activator) et les protéines MED23 et MED25. MED23 et MED25 sont des sous-unités du médiateur, complexe qui régule la transcription en intégrant les signaux entre des facteurs de transcription et la machinerie transcriptionnelle. Nous avons démontré que ces protéines interagissent in vitro et in vivo avec ERM et sont nécessaires à l’activation transcriptionnelle induite par ce facteur de transcription. Toutefois, ces sous-unités ont une capacité différente de recrutement du médiateur sur ERM in vitro.La ribonucléoprotéine CoAA régule l’expression de certains gènes et l’épissage alternatif des ARN messagers. CoAA interagit in vitro et in vivo avec ERM. L’activité d’ERM est augmentée par la surexpression de CoAA tandis qu’elle est diminuée par sa sous-expression. Cette activation médiée par CoAA est liée à une modulation du taux de sumoylation d’ERM. Ces travaux ont permis de définir de nouvelles voies de régulation de l’activité transcriptionnelle d’ERM et des autres membres du groupe PEA3. Il reste maintenant à préciser les mécanismes impliqués dans la modulation de l’activité des membres du groupe PEA3 par ces nouveaux interacteurs. / ERM is an ETS transcription factor which belongs to the PEA3 group and is involved in several processes such as migration and dissemination during organogenesis and cancer development. Regulation of its transcriptional activity requires post-translational modifications and interactions with partner proteins. In order to identify new ERM partners, we have developed various affinity chromatography techniques to isolate new potential partners. Among these candidates, CoAA (CoActivator Activator), MED23 and MED25 directly interact with ERM.MED23 and MED25 are subunits of the mediator. The mediator is a 30 sub-units multi-protein complex which mediates signals from transcription factors bound at upstream promoter elements or enhancers to RNA polymerase II and the general initiation factors bound at the core promoter. We found that MED23 and MED25 interact with ERM in vitro and in vivo and are required for transcriptional activation induced by ERM. However, these sub-units display various ability to recruit the mediator on ERM in vitro. The heterogeneous nuclear ribonucleoprotein-like protein CoAA regulates gene expression and RNA splicing. We demonstrated that ERM interacts in vitro and in vivo with CoAA. ERM transcriptional activity is enhanced upon CoAA overexpression and is decreased by CoAA knock-down. We demonstrated that CoAA modulates ERM transcriptional activity by decreasing sumoylated ERM levels. This work demonstrated new ways to regulate the activity of ERM and the two other PEA3 group members. The molecular mechanisms involved in the modulation of PEA3 member activity by these partners remain to be clarified.

Effects of oestrogen on prolactin transcription patterns in living pituitary tissue Amanda Louise Patist PhD in the Faculty of Medical and Human Sciences, September 2013, The University of ManchesterOestrogen is a well-known modulator of the transcription and secretion of prolactin as well as having a role in the physiological proliferation and possibly also in pathological hyperplasia of lactotrophs. Our group has studied prolactin gene promoter regulation in single living pituitary cells in Fischer 344 (d2eGFP-hPRL 455) rats, that express a destabilised eGFP reporter gene under the control of a human prolactin genomic fragment, and identified prolactin transcription cycles that occur in a non-circadian and non-synchronised fashion. Pulsatile transcription has been identified in fetal tissue, stabilising during development. Here we assess the effects of physiological and supraphysiological oestrogen exposure in vivo on prolactin transcription in the adult rat. We have established and validated models of both states by evaluating the expression of the hPRL-d2eGFP transgene during the oestrous cycle and in males with long-term oestradiol-releasing implants, respectively. The oestrous cycles of adult female rats were synchronised by IP LHRH injection. A 1.8-fold increase in the number of cells expressing the prolactin transgene at oestrus (n=7) as compared to diestrus (n=5) and a 10.6 fold increase in mean fluorescence per cell was identified by flow cytometry. In males, chronic oestrogen stimulation induced a 2.5-fold increase in pituitary weight, a 5.2-fold increase in number of cells expressing the transgene and 4.4-fold increase in fluorescence per cell, as indicated by flow cytometry (n=3). Immunofluorescence, qPCR and serum analysis confirmed the high-production state of lactotroph cells in both female and male models. 250µm pituitary slices were imaged for 48 hours using time-lapse confocal microscopy and pulsatile fluorescent reporter activity was evident in both female and male models. Interestingly, no significant difference was seen between transcription cycle patterns (timing or amplitude of transcriptional pulses) in individual cells between high and low oestrogen states. Using a novel mathematical model, that calculates transcription rate from fluorescence data, we have been able to study the transcription rates displayed by single cells and the estimated points at which a cell switches from one rate to another. Patterns in switches in transcription rates were similar between high and low oestrogen states, suggesting that individual pituitary cells within the context of tissue, continue to display cyclical patterns of gene expression, in states of both high and low prolactin production. This implies that cyclical transcription is a fundamental property of prolactin gene expression that persists in different physiological states, and that modulation of cycle characteristics is not the mechanism for increased prolactin synthesis in these circumstances.

Speech recognition is now a key topic in computer science with the proliferation of voice-activated assistants, and voice-enabled devices. Many companies over a speech recognition service for developers to use to enable smart devices and services. These speech-to-text systems, however, have significant room for improvement, especially in domain specific speech. IBM's Watson speech-to-text service attempts to support domain specific uses by allowing users to upload their own training data for making custom models that augment Watson's general model. This requires deciding a strategy for picking the training model. This thesis experiments with different training choices for custom language models that augment Watson's speech to text service. The results show that using recent utterances is the best choice of training data in our use case of Digital Democracy. We are able to improve speech recognition accuracy by 2.3% percent over the control with no custom model. However, choosing training utterances most specific to the use case is better when large enough volumes of such training data is available.

Bibliography: pages 133-155. / Insulin regulates cellular gene expression and modulates specific mRNA levels in liver cells. As yet, the mechanism of this control is still unclear. The effects are initiated following the binding of insulin to the plasma membrane receptor. Although several mediators of the signal from the plasma membrane to the nucleus have been proposed, none has proved capable of eliciting all of the effects of insulin on gene expression. Therefore, the possibility that insulin itself may directly regulate transcription at the level of the nucleus, was investigated.

Promoters are key regions that are involved in differential transcription regulation of protein-coding and RNA genes. The gene-specific architecture of promoter sequences makes it extremely difficult to devise a general strategy for their computational identification. Accurate prediction of promoters is fundamental for interpreting gene expression patterns, and for constructing and understanding genetic regulatory networks. In the last decade, genomes of many organisms have been sequenced and their gene content was mostly identified. Promoters and transcriptional start sites (TSS), however, are still left largely undetermined and efficient software able to accurately predict promoters in newly sequenced genomes is not yet available in the public domain. While there are many attempts to develop computational promoter identification methods, reliable tools to analyze long genomic sequences are still lacking. In this dissertation, I present the methods I have developed for prediction of promoters for different organisms. The first two methods, TSSPlant and PromCNN, achieved state-of-the-art performance for discriminating promoter and non-promoter sequences for plant and eukaryotic promoters respectively. For TSSPlant, a large number of features were crafted and evaluated to train an optimal classifier. Prom- CNN was built using a deep learning approach that extracts features from the data automatically. The trained model demonstrated the ability of a deep learning approach to grasp complex promoter sequence characteristics. For the latest method, DeeReCT-PromID, I focus on prediction of the exact positions of the TSSs inside the eukaryotic genomic sequences, testing every possible location. This is a more difficult task, requiring not only an accurate classifier, but also appropriate selection of unique predictions among multiple overlapping high scoring genomic segments. The new method significantly outperform the previous promoter prediction programs by considerably reducing the number of false positive predictions. Specifically, to reduce the false positive rate, the models are adaptively and iteratively trained by changing the distribution of samples in the training set based on the false positive errors made in the previous iteration. The new methods are used to gain insights into the design principles of the core promoters. Using model analysis, I have identified the most important core promoter elements and their effect on the promoter activity. Furthermore, the importance of each position inside the core promoter was analyzed and validated using a large single nucleotide polymorphisms data set. I have developed a novel general approach to detect long range interactions in the input of a deep learning model, which was used to find related positions inside the promoter region. The final model was applied to the genomes of different species without a significant drop in the performance, demonstrating a high generality of the developed method.

The literal meaning of "transcription" is "written across"; in freer terminology it is "the musical counter-part of literary translation." The terms "arrangement" and "transcription: are used synonymously by nearly all musical dictionaries. However, even then , there is often a distinction made, an arrangement being considered to adhere strictly to the original text while a transcription admits of creative participation on part of the arranger, ranging from the simple transference of musical medium to the complete reworking of a composition with additions and modifications.

Transcription factors (TFs) are proteins that respond to a specific chemical signal and bind to DNA. In some bacteria, TFs control transition metal ion homeostasis by specifically binding with a particular metal ion or ions and then interacting with DNA. Although most first row metal ions are required as micronutrients for life, many of them can cause cellular damage or death if their concentrations are too high; this makes these TFs and their biological interactions excellent targets for drug development. The bacteria, Streptococcus pneumoniae is a pathogenic microorganism responsible for a range of diseases that target the young and the old, including pneumonia, meningitis, and bacteremia. Herein, we describe our efforts to study the Streptococcus pneumoniae TF responsible for copper(II) ion homeostasis. This thesis describes the classical biochemical techniques used to over-express and purify CopR. It also describes a series of preliminary characterization data associated with this novel copper(II)-dependent TF.