(Georgia Institute of Technology, 2010-11-10)
Fernandez, Ariel
We shall present on research endeavors focusing on controlling specificity in molecularly targeted anticancer therapy. A basic goal is to reduce toxic side effects by structure-based drug design exploiting our understanding of the evolutionary basis of specificity. Emphasis will be placed on engineering kinase inhibitors (KIs) with minimal clinical uncertainty. A radical innovation is on its way in terms of rationally redesigning KIs to reduce their toxicity. As it turns out, we can control specificity to an unprecedented degree and hence decisively contribute to test the limits of therapeutic efficacy. To support this claim, we start by noting that cross reactivities of KIs arise because of the structural similarity and amino acid conservation across evolutionarily related (paralog) kinases. Yet, while paralogs share a similar structure, they are "wrapped" differently. As we compare the microenvironments of intramolecular hydrogen bonds aligned across paralog structures we notice crucial differences: some hydrogen bonds are exposed to solvent in one kinase and shielded from water in another, or, rather, one hydrogen bond may be deficiently wrapped in one kinase but well wrapped in another. Taking into account such local differences, we are able to redesign KIs because deficiently wrapped hydrogen bonds -the so-called dehydrons - not only distinguish paralogs but are also inherently sticky. Hence, a new design strategy to achieve higher specificity emerges as we turn KIs into exogenous wrappers of dehydrons that are not conserved across structures of common ancestry. Such wrapping designs enable paralog discrimination and hence yield drugs capable of curbing side effects.
(Georgia Institute of Technology, 2010-10-13)
Roberts, Richard J.
The discovery of new restriction enzymes has progressed from the slow and laborious process of examining individual bacterial strains experimentally to the much higher through-put that can be obtained using bioinformatics to analyze genome sequences. Nevertheless, both the traditional biochemical approach and the use of biochemistry to test bioinformatics predictions are an integral part of current practices. The analysis of bacterial genomes for their content of genes related to restriction-modification systems faces many of the challenges found more generally when annotating genomes. The methodology that we have developed will be described and examples will be drawn from the analysis of both individual microorganisms and the environmental projects that broadly sample various environments. In the final part of the talk, the more general problem of genome annotation will be discussed and a new project, COMBREX, that aims to accelerate the experimental assignment of function to genes will be described.
(Georgia Institute of Technology, 2010-09-08)
Nurse, Paul
The growth and reproduction of all living organisms are dependent on the cell cycle, the process which leads to cell division. Uncontrolled division of cells is important for disease particularly cancer. Two events, S-phase and mitosis, are common to all cell cycles and are necessary for the two newly divided cells to receive a full complement of genes. In fission yeast the onset of S-phase and mitosis are controlled by a single cyclin dependent kinase with different levels of CDK activity bringing about progression through the cell cycle in an ordered fashion. Activation of CDK activity is determined by growth rate and cell size, with cell size determined by a gradient mechanism with controlling molecules diffusing from the periphery of the cell to be sensed in the middle of the cell.