Search

The recent introduction of the SEA-TROSY experiment (Pellecchia et al. (2001) J. Am. Chem. Soc., 123, 4633-4634) can alleviate the problem of resonance overlap in 15N/2H labeled proteins. This method selectively observes solvent exposed amide protons with a SEA element. However, SEA-TROSY spectra may be contaminated with exchange-relayed NOE contributions from fast exchanging hydroxyl or amine protons and longitudinal relaxation contributions. Furthermore, for non-deuterated proteins or protein-ligand complexes, SEA-TROSY spectra may contain NOE contributions from aliphatic protons. In this communication, a modified version of the SEA element, a Clean SEA element, is introduced to eliminate these artifacts.

Recent studies have indicated that the relaxation rate of the (1)H-(13)C multiple-quantum coherence is much slower than that of the (1)H-(13)C single-quantum coherence for non-aromatic methine sites in (13) C labeled proteins and in nucleic acids at the slow tumbling limit. Several heteronuclear experiments have been designed to use a multiple-quantum coherence transfer scheme instead of the single-quantum transfer method, thereby increasing the sensitivity and resolution of the spectra. Here, we report a constant time, gradient and sensitivity enhanced HMQC experiment (CT-g/s-HMQC) and demonstrate that it has a significant sensitivity enhancement over constant time HMQC and constant time gradient and sensitivity enhanced HSQC experiments (CT-g/s-HSQC) when applied to a (13)C and (15) N labeled calmodulin sample in D(2)O. We also apply this approach to 3D NOESY-HMQC and doubly sensitivity enhanced TOCSY-HMQC experiments, and demonstrate that they are more sensitive than their HSQC counterparts.

Previously we demonstrated a sensitivity enhancement of the original TROSY experiment by a factor of [Formula: see text] by the use of the sensitivity enhanced TROSY (en-TROSY) scheme. Here, we develop a gradient and sensitivity enhanced TROSY experiment (gs-TROSY), which is designed to select magnetization transfer pathways that suppress spectral artifacts and reduce the number of required phase cycles while having minimal loss of sensitivity. Both of these experimental methods (en-TROSY and gs- TROSY) have been combined with a water flip-back scheme which provides a further increase in sensitivity for labile NH groups by avoiding water saturation. We also apply these TROSY schemes to 3D NOESY-TROSY and 3D TOCSY-TROSY experiments.

The FMN-binding domain of human NADPH-cytochrome P450 reductase, corresponding to exons 3-7, has been expressed at high level in an active form and labelled with 13C and 15N. Most of the backbone and aliphatic side-chain 1H, 15N and 13C resonances have been assigned using heteronuclear double- and triple-resonance methods, together with a semiautomatic assignment strategy. The secondary structure as estimated from the chemical shift index and NOE connectivities consists of six alpha-helices and five beta-strands. The global fold was deduced from the long-range NOEs unambiguously assigned in a 4D 13C-resolved HMQC-NOESY-HMQC spectrum. The fold is of the alternating alpha/beta type, with the five beta-strands arranged into a parallel beta-sheet. The secondary structure and global fold are very similar to those of the bacterial flavodoxins, but the FMN-binding domain has an extra short helix in place of a loop, and an extra helix at the N-terminus (leading to the membrane anchor domain in the intact P450 reductase). The experimental constraints were combined with homology modelling to obtain a structure of the FMN-binding domain satisfying the observed NOE constraints. Chemical shift comparisons showed that the effects of FMN binding and of FMN reduction are largely localised at the binding site.