Chemical shift optimization and ensemble averaging in protein NMR spectroscopy

摘要

A problem often encountered in multidimensional NMR spectroscopy is that an existingudchemical shift list of a protein has to be used to assign an experimental spectrum but does not fitudsufficiently well for a safe assignment. A similar problem occurs when temperature or pressure series ofudn-dimensional spectra are to be evaluated automatically. Two slightly different algorithms,udAUREMOL-SHIFTOPT1 and AUREMOL-SHIFTOPT2 have developed here that fulfill this task.udTheir performance is analyzed employing a set of simulated and experimental two-dimensional andudthree-dimensional spectra obtained from three different proteins. Peak probability and atom type basedudweighted averaging is introduced in order to reduce the influence of the wrong assignment during theudassignment process.udChemical shift prediction programs often use a single energy minimized structure as input, butudensemble averaging of chemical shifts gives better prediction values irrespective of the predictionudmethod. This is in agreement with the fact that proteins in solution occur in multiple conformationaludstates in fast exchange on the chemical shift time scale. However, in contrast to the real conditions inudsolution at ambient temperatures, the chemical shift prediction methods seems optimal to predict theudlowest energy ground state structure that is only weakly populated under these conditions. An analysisudof the data shows that a chemical shift prediction can be used as measure to define the minimum size ofudthe structural bundle required for a faithful description of the structural ensemble.udReliable homo- and heteronuclear chemical shift distributions are required for the automatedudassignment procedures. However, the statistics derived from the Biological Magnetic Resonance Bankud(BMRB) is not clean and is not structurally unbiased. Therefore, refined chemical shift statistics wasudcreated from a structural database of non-homologous proteins (Nh3D) that comprises 806 differentudthree-dimensional structures. The chemical shift data base was created by calculating the resultingudchemical shifts with the prediction programs SHIFTS and SHIFTX. Analysis of the obtained data setudshows that unbiased chemical shift statistics improves the a priori probability values for resonanceudassignment, removes ambiguities in assignment to certain level and helps to make stereochemicaludassignments.