Fast model systems for misfolding, binding and aggregation

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Fast (microsecond) experiments and molecular dynamics simulation are finally working hand-in-hand to provide validated atomistic pictures of protein folding dynamics. Larger proteins and millisecond folders are around the corner, although empirical force fields need more improvement before mechanistic predictions become as reliable as average rate coefficients or native structures. It is time to extend this simulation experiment interplay to misfolding, aggregation and binding processes. The problem is that such processes are usually slow - seconds to days. In order to compare experiment and simulation directly during the next 4 years, when simulations will not reach far beyond the 1 ms regime yet, the solution is simple: create small and fast experimental model systems for misfolding, aggregation and binding. These are analogous to fast two-state and downhill folding studied during the last 10 years: certainly not all proteins fold that way, but much useful was learned from our ability to directly compare such model proteins with simulation. Here, we propose development of: 1) ?6-85 as a model system for sheet-containing misfolded traps (T denaturation and T-jumps) as well as excess helix-containing traps (P denaturation, P-jumps). The key is that all folding and misfolding events are completed in ~ 1 ms or faster. 2a) Fast-folding tethered proteins (WW and ?6-85) to facilitate the interplay between transient aggregation and folding. Oligomeric aggregates include chimeras with elements swapped among monomers, as well as less structured aggregates. Tethering allows high effective concentrations without going to high protein concentration (leading to uncontrollable aggregation), and also helps keep MD simulation by keeping reactants in close proximity. 2b) U1A-RNA binding to multiple sites on U1A protein will be a fast model system for binding at multiple sites. 3) For the next generation of misfolding/aggregation/binding research, we will develop a single molecule instrument capable of high throughput (106 molecules/day) and studying 2 or more molecules interacting without cross-linking them or confining them in a micelle. Five simulation groups have been brought on board (Schulten, Shaw, Cheung, Luthey-Schulten and Pande), and their students/postdocs are already working closely with mine so simulation can be developed in parallel with the experiments. The goal is to provide data in the few microsecond to few millisecond range, amenable to full atom simulation over the next 4 years, so misfolding/aggregation/binding processes can be studied at the atomistic level, but on systems small and fast enough for simulation to work. We build on our knowledge of ?6-85, WW domain and U1A, so much is known experimentally and computationally about the monomeric building blocks of our misfolding/aggregation/binding models.