A New Procedure: PARSE

"The ability to determine dynamic structures is not a solved problem yet," says James. "The obstacles, in some sense, are that if it's so difficult to determine the average structure for a molecule in solution, then how could I figure out the many interconverting structures for that same molecule in solution. If a molecule assumes one structure part of the time and another structure part of the time, how do I figure out which ones these are? And the spectral parameters used to determine structure are also influenced by how fast the exchange between structures occurs."

Conventional computational techniques, for instance, do not specify the minimum number of conformations -- different structural configurations -- possible and the probability that each one exists. Rather, these techniques average together the many conformations and develop an "average" structure that may never exist.

A solution, says James, is a computational approach developed by James' research associate Nikolai Ulyanov. Known as PARSE, it sifts through NMR data, creating hundreds of possible conformations. The viability of this approach depends on high quality experimental data, but with quality input PARSE is capable of determining which among many possible structures best fit the data.

Nine different structures of the same sequence of DNA as determined by PARSE. This result shows that the spatial orientation of the sugars (deoxyribose, shown as five-membered rings) in the backbone structure of DNA can shift under physiological conditions from the typical S-conformation (blue) to N-conformation (green). This could affect interactions between DNA and proteins such as RNA-polymerase.

go back to the main page