Search

In large molecular structures, the magnetization of all hydrogen atoms in the solute is strongly coupled to the water magnetization through chemical exchange between solvent water and labile protons of macromolecular components, and through dipole-dipole interactions and the associated "spin diffusion” due to slow molecular tumbling. In NMR experiments with such systems, the extent of the water polarization is thus of utmost importance. This paper presents a formalism that describes the propagation of the water polarization during the course of different NMR experiments, and then compares the results of model calculations for optimized water polarization with experimental data. It thus demonstrates that NMR spectra of large molecular structures can be improved with the use of paramagnetic spin relaxation agents which selectively enhance the relaxation of water protons, so that a substantial gain in signal-to-noise can be achieved. The presently proposed use of a relaxation agent can also replace the water flip-back pulses when working with structures larger than about 30kDa. This may be a valid alternative in situations where flip-back pulses are difficult to introduce into the overall experimental scheme, or where they would interfere with other requirements of the NMR experiment

A fully automated, NOE-based NMR structure determination of a uniformly 13C,15N-labeled protein was achieved in crude cell-extract, without purification of the overexpressed protein. Essentially complete sequence-specific assignments were obtained using triple resonance experiments, based on the high intensity of the resonances from the overexpressed protein relative to those of the background. For the collection of NOE distance constraints, efficient discrimination between NOE cross peaks from the target protein and background signals was achieved using the programs ATNOS and CANDID. In the iterative ATNOS/CANDID procedure, the identification of the desired protein NOEs is initially guided by the self-consistency of the protein NOE-network. Although the intensities of the signals in this network vary over a wide range, and are in many instances comparable to or smaller than those of the background, the first cycle of calculations resulted in the correct global polypeptide fold, and the structure was then refined in six subsequent cycles using the intermediate NMR structures for additional guidance. The experience gained with this work demonstrates that the ATNOS/CANDID procedure for automatic protein structure determination is highly robust and reliable in the presence of intense background signals, and might thus also represent a platform for future protein structure determinations in physiological fluids

In large molecular structures, the magnetization of all hydrogen atoms in the solute is strongly coupled to the water magnetization through chemical exchange between solvent water and labile protons of macromolecular components, and through dipole-dipole interactions and the associated "spin diffusion” due to slow molecular tumbling. In NMR experiments with such systems, the extent of the water polarization is thus of utmost importance. This paper presents a formalism that describes the propagation of the water polarization during the course of different NMR experiments, and then compares the results of model calculations for optimized water polarization with experimental data. It thus demonstrates that NMR spectra of large molecular structures can be improved with the use of paramagnetic spin relaxation agents which selectively enhance the relaxation of water protons, so that a substantial gain in signal-to-noise can be achieved. The presently proposed use of a relaxation agent can also replace the water flip-back pulses when working with structures larger than about 30kDa. This may be a valid alternative in situations where flip-back pulses are difficult to introduce into the overall experimental scheme, or where they would interfere with other requirements of the NMR experiment

A fully automated, NOE-based NMR structure determination of a uniformly 13C,15N-labeled protein was achieved in crude cell-extract, without purification of the overexpressed protein. Essentially complete sequence-specific assignments were obtained using triple resonance experiments, based on the high intensity of the resonances from the overexpressed protein relative to those of the background. For the collection of NOE distance constraints, efficient discrimination between NOE cross peaks from the target protein and background signals was achieved using the programs ATNOS and CANDID. In the iterative ATNOS/CANDID procedure, the identification of the desired protein NOEs is initially guided by the self-consistency of the protein NOE-network. Although the intensities of the signals in this network vary over a wide range, and are in many instances comparable to or smaller than those of the background, the first cycle of calculations resulted in the correct global polypeptide fold, and the structure was then refined in six subsequent cycles using the intermediate NMR structures for additional guidance. The experience gained with this work demonstrates that the ATNOS/CANDID procedure for automatic protein structure determination is highly robust and reliable in the presence of intense background signals, and might thus also represent a platform for future protein structure determinations in physiological fluids