No description available.

No description available.

No description available.

No description available.

The yeast Fus1p SH3 domain recognizes non-PXXP motif targets. This is contrary to most SH3 domains, which recognize PXXP-containing sequences. In this thesis, I characterized atypical binding of the Fus1p SH3 domain and provide insight into atypical SH3 domain interactions. I demonstrated that the Fus1p SH3 domain binds R(S/T)(S/T)SL-containing peptides derived from yeast proteins, Bnr1p and Ste5p. Through mutagenesis studies and comparisons to other SH3 domains, I showed that the Fus1p SH3 domain utilizes a portion of the same binding surface as typical SH3 domains. However, the PXXP-binding surface is debilitated in the WT domain by the substitution of unusual residues at three key conserved positions. By replacing these residues, I created a version of the Fus1p SH3 domain that binds a PXXP-containing peptide with extremely high affinity. Based on my data and analysis, I clearly delineated two distinct surfaces comprising the typical SH3 domain binding interface, and show that one of these surfaces is the primary mediator of almost every “non-canonical” SH3 domain-mediated interaction. I demonstrated that the Fus1p SH3 domain also binds Bni1p and Pea2p through yeast two-hybrid experiments, which do not contain either PXXP or R(S/T)(S/T)SL motifs. Through mutagenesis studies and two-hybrid experiments, I showed that the Fus1p SH3 domain utilizes a binding surface comprised of two sub-surfaces to interact with Bni1p and Pea2p. The sub-surfaces include the same portion of the binding surface in typical SH3 domains utilized in R(S/T)(S/T)SL binding of the Fus1p SH3 domain, and a novel binding site. I also showed that the protein binding surface of the Fus1p SH3 domain has a role(s) for efficient mating. Based on my findings in this thesis, I propose that dramatic alterations in SH3 domain specificity can be simply explained as a modulation of the binding strengths of sub-binding sites within the binding surface.

The effect of chain extender chemical structure on the physical and mechanical properties of thermoplastic polyurethane/urethaneurea elastomers was systematically investigated. Several series of materials were synthesised using 4,4' -diphenylmethane diisocyanate (MDI) and poly(tetramethylene oxide) glycol (PTMG), each series incorporating a particular class of chain extender compound. Elucidation of the influence of chain extender structural variations within each series was the principal objective of the investigation…

The tetratricopeptide repeat (TPR) motif contains 34 amino acids forming a helix-turn-helix structure. Different numbers of tandem TPR motifs assemble to form a TPR domain, thereby generating a polypeptide-binding interaction surface. The TPR domain provides a scaffold for mediating protein-protein interactions. Proteins that contain TPR domains exist in a broad range of organisms. These proteins have various functions. Cyclophilin 40 (Cyp40) and C-terminal Hsc70 interaction protein (CHIP) are two typical members of the family of TPR-containing proteins. Both proteins have the ability to bind the molecular chaperones Hsp70 and Hsp90. In most cases, TPR domains act as a scaffold to link chaperone and substrate or multi-protein complexes. Recent evidence suggests that Hsp90 binding to TPR domains can change the overall protein conformation but the allosteric mechanism triggered by ligand binding to the TPR domain remained unknown. This study focuses on using biophysical methods on the two TPR domain containing proteins Cyp40 and CHIP. In particular, this study reveals how the binding of the molecular chaperones Hsp70/90 to the TPR domains of Cyp40 and CHIP influences protein conformation and function. Here we show how conformational changes of the TPR domains affect structure and activity of Cyp40 and CHIP. By using biophysical methods, including thermal denaturation assay (TDA), differential scanning calorimetry (DSC), hydrogen deuterium exchange with mass spectrometry (HDX-MS) and small angle X-ray scattering (SAXS), together with enzymatic assays, we showed that (1) heat shock proteins allosterically affect the enzyme activity of both Cyp40 and CHIP, (2) heat shock proteins bind to the TPR domains of both Cyp40 and CHIP; (3) the binding increases the thermostability of both proteins. Further, by mutating an essential lysine in the TPR1 domain of both proteins (K30 for CHIP, and K227 for Cyp40) to alanine, the thermostability was significantly affected. The SAXS data showed in addition of the SRMEEVD peptide reduced the flexibility of CHIP. HDX-MS experiments suggest that the dynamic alteration due to binding with the Hsp90 peptide or the mutations further reduce the flexibility of the catalytic domains of both proteins. The results imply that the allosteric effects on the enzymatic activity are consequences of dynamic changes of the TPR domains. Hsp70 was also found to bind less tightly to CHIP-K30A than to wild-type CHIP, and thus showed less inhibition of enzymatic activity. These results further confirmed the discovery, that the dynamics of TPR domains allosterically affect enzymatic activity.

This thesis presents new domain-theoretic models for randomized algorithms. A randomized algorithm can use random bits from an oracle to control its computation. The possible random bits form a binary tree, where each random choice of a bit is a branching of the tree. The randomized algorithm then determines what the output should be for each branch. This idea forms the basis of our random choice functors. However, the functor only provides one half of the model. We must also show how multiple randomized algorithms can be combined or composed. This is where the monadic structure comes into play. If we wish to join multiple randomized algorithms to form one resulting algorithm, then we can run each algorithm in parallel, using the same random bits for each. Monads are used to add a computational effect to an existing semantic model. In order to work with models of the lambda calculus, it is important to work in a Cartesian closed category of domains, due to Lambek's theorem and Scott's corollary. Our first random choice monad is shown to be an endofunctor of the Cartesian closed category BCD. If we wish to add multiple computational effects, then we can compose monads as long as the monads enjoy a distributive law. It is shown that in the category BCD, our first random choice monad enjoys a distributive law with the lower powerdomain for nondeterminism. Two variations of the random choice monad are then given. The first variation has a distributive law with the convex powerdomain in the categories RB and FS, while the second variation has a distributive law with the upper powerdomain in BCD. We use the random choice monads to develop a new programming language, Randomized PCF. This extends the language PCF by adding in random choice, allowing for the programming of randomized algorithms. A full operational semantics is given for Randomized PCF, and a random choice monad is used to give it a mathematical model (denotational semantics). Finally, an implementation of Randomized PCF is developed, and the Miller-Rabin algorithm is implemented in Randomized PCF. / Tyler Barker

<p>Today, millions of purchased domain sites names are sitting unused with no real web designs or concrete purpose coupled with them. Why would not owners engage a web-hosting domain-parking hotel so they can earn money through eyeballs advertising or click revenue while their sites sits unused? Parking hotels claim access to passive domain monetization through advertising programs tailored to generate revenue via automatic web page generations containing tailored advertisement. When visitors access these web sites, revenue is generated for domain owners. This study sought to investigate if parking hotels’ advertising claim, that you can make money through them, carries substance. Hence, is it possible to generate sufficient revenue using a parking hotel’s advertising revenue model to pay for the cost of domain ownership and, if possible, generate excess revenue? The study’s epistemological approach, trying to distinguish the truth about the hotels’ claims, stems from a ontological discussion around web hosting hotels’ advertising vehicle existence and its ability to generates revenue. The study challenged a parking hotel’s claim through an inductive quantitative approach by watching advertising revenue for 59 domain names over 105 days. Quantitative data concluded, through a statistical approach, that there were insufficient advertising income, approximately a nickel, to cover the annual cost of approximately $6 USD. Therefore, the study concluded that the advertising claims were misleading and recommends that sites are not purchased, or renew, for the purpose of making money through web hotels’ advertising models.</p>

No description available.