NMR Spectroscopy of Large Biological Macromolecules in Solution

Abstract

of 400, 500, 600, 750 and 800 MHz. Each of these advances in magnet design benefitted biomolecular NMR through improved intrinsic sensitivity and peak separation [5,9,12], but for commonly used heteronuclear experiments (e.g. [10,11,15]), the advantages of using higher magnetic fields were partly offset by field-dependent line broadening, which again causes loss of sensitivity and spectral resolution. Using the TROSY technique [13,16 • ,17 • ], line broadening at higher magnetic fields, which is a manifestation of increased transverse relaxation rates that also cause deterioration of the sensitivity in complex NMR experiments, has been largely suppressed. This results in improved spectral resolution (Figure 2), improved effective sensitivity or a combination of both, depending on the experiment used (see below). At high magnetic fields, chemical shift anisotropy (CSA) of 1 H, 15 N and 13 C nuclei can be a significant source of transverse relaxation, in addition to the omnipresent relaxation as a result of dipole-dipole coupling. TROSY exploits constructive interference between dipole-dipole coupling and CSA relaxation, and actually uses CSA relaxation at higher fields to cancel field-independent dipolar relaxation. Technically, the TROSY approach is based on the following: in heteronuclear two-spin systems , such as 15 N-1 H in the amide groups of proteins or in nucleic acid bases, the NMR signal of each nucleus is split into two components by the scalar spin-spin coupling. In 2D correlation experiments, one therefore observes a four-line fine structure (Figure 3b). With the advent of modern multidimensional NMR, this four-line pattern routinely collapses into a single, centrally located line by broad-band decoupling techniques (Figure 3a) [12,15,18], with the expectation of obtaining a simplified spectrum and improved sensitivity. As has long been known (see [13] for a literature survey), however, the individual multiplet components have different transverse relaxation times and, hence, different line widths (Figure 3b), which are mixed by the aforementioned decoupling. For large molecules studied at higher magnetic fields, the differential line broadening of the individual multiplet components is very pronounced and results in the deterioration of the averaged signal (Figure 3a,b). Using the TROSY technique, the multi-plet structure is not decoupled (Figure 4) and only the narrowest, most slowly relaxing line of each multiplet is retained (Figure 3c). The sensitivity loss as a result of the use of only one out of the four multiplet components is partially recovered both by using a new polarization transfer element that retains 50% (not only 25%) of the original proton polarization [17 • ] and by the fact that the absence of decoupling allows the use of the steady-state heteronuclear magnetization, in addition to the proton polarization [16 • ,17 • ]. Overall, when working with molecular sizes above 20 kDa, a superior ratio of peak height to noise is readily achieved with TROSY compared with corresponding conventional experiments. The absence of any mixing of the individual NMR transitions that correspond to the four fine structure Figure 2 115 125 120 10 8 ω 2 (1 H) (ppm) ω 1 (15 N) (ppm) (a) COSY 9 1 0 8 ω 2 (1 H) (ppm) 9 115 125 120 ω 1 (15 N) (ppm) (b) TROSY Current Opinion in Structural Biology A comparison of the 15 N-1 H correlation spectra of a protein with a molecular weight of 45 kDa recorded using (a) conventional procedures ('COSY') and (b) TROSY. Both spectra were measured at a proton resonance frequency of 750 MHz, using a 0.8 mM sample of uniformly 15 N-and 2 H-labeled gyrase-45 from Staphylococcus aureus in water at 25°C and pH 8.6.