Recent progress in sequencing of various genomes has uncovered thousands of proteins with little homology to characterized proteins, and hence, proteins with unknown function. The prospect of uncovering function through the acquisition of structural information on a genome wide scale (structural genomics) is appealing. NMR has a role to play in massive structure determination efforts, both as a screening tool for choice of structure determination protocol, and as a structure determination tool for non-crystallizable proteins. Our efforts in this area are divided into two segments, NMR screening and database management, and rapid determination of backbone structures.

NMR Screening. Initial screens are conducted on non-isotopically labeled samples of soluble (>0.1mM) proteins less than 60kDa in molecular weight. The screens utilize automated flow probe technology to collect normal one dimensional proton spectra, one dimensional amide proton exchange spectra, and one dimensional diffusion spectra. These spectra are analyzed to give estimates of secondary structure types, a percentage of stable fold, and a level of aggregation. Partially folded and heterogeneously aggregated proteins are subjected to buffer variation to optimize homogeneity and the percentage of stable fold. Well folded proteins and proteins greater than 25kDa in molecular weight are set aside to await the results of crystallization trials. For smaller proteins showing high levels of homogeniety and moderate levels of well formed structure requests for 15N labeling are sent to the expression core. Please see the Data Status page for current screening results.

Backbone fold determination. 15N labeled samples are screened based on two dimensional HSQC spectra to optimize resolution and homogeneity. 1H-15N residual dipolar couplings and amide proton exchange data are collected in one of three field oriented liquid crystal media to provide a basis for classification as to fold family. Proteins that do not clearly fall into a known fold family are subjected to more detailed analysis. Current examples rely on the assignment of backbone resonances and identification of secondary structure elements using normal scalar coupling and NOE based experiments. Dipolar data are used to orient predetermined elements in a three dimensional fold. Strategies that accomplish simultaneous assignment of resonances and determination of structure using a wider range of dipolar measurements are also being developed and employed. Below we compare a structure of acyl carrier protein determined by the secondary element fold protocol to a structure determined by traditional NMR methods.

[ View Status Report ]