Spin water-protein NOEs

Primary screens were performed in duplicate on mixtures of 80 compounds (1 pM per compound, 5 pM protein), where samples in each mixture were orthogonally pooled so that no two compounds that are in one well are also together in another well. Primary hits were achieved when the same compound in two wells were observed. The GPC spin-column eluates were partially resolved using a reversed-phase C18 HPLC column (Waters Xterra 2.1 x 20.0 mm) with a 1-min ballistic gradient and total cycle time of 2 min. The HPLC eluates were analyzed with a Micromass eight-way MUX interfaced to an LCT Tof ESI-MS system. Achieved throughput was 100 000 compounds per day. 

The temperature dependence of the MRD profile for the protein-water systems where the protein is magnetically a solid, is remarkably weak. The relaxation rate is proportional to IjT, which is consistent with Eq. (4) that was derived on the assumption that the relaxation process is a direct spin-phonon coupling rather than an indirect or Raman process. If it were a Raman process, there would be no magnetic field dependence of the relaxation rate therefore, the temperature dependence provides good evidence in support of the theoretical foundations of Eq. (6). 

Since water protons are not bound to or nuclei, the water signal is also suppressed by the spin-lock purge pulse. In practice, the suppression of the water signal is sufficient to record HSQC spectra of protein samples dissolved in mixtures of 95% H20/5% D2O without any further water suppression scheme [12]. For optimum water suppression the carrier frequency must be at the frequency of the water resonance. On resonance, the phase of the water magnetization is not affected by imperfections of the first 180°(ff) pulse, so that no solvent magnetization ends up along the axis of the spin-lock purge pulse. 

Measurements of the dynamic properties of the surface water, particularly NMR measurements, have shown that the characteristic time of the water motion is slower than the bulk water value by a factor of less than 100. The motion is anisotropic. There is litde or no irrotadonally bound water. Study of a protein labeled covalently with a nitroxide spin probe (Polnaszek and Bryant, 1984a,b) has shown that the diffusion constant of the surface water is about 5-fold below the bulk water value. The NMR results are in agreement with measurements of dielectric relaxation of water in protein powders (Harvey and Hoekstra, 1972). 

Relaxations of spins of atoms are free of these complications. A comparison of the relaxations in two different proteins (18), one with no buried water molecule and another with buried water molecules, showed a great difference at low frequencies. This difference was interpreted as due to these few buried molecules, offering NMR spectroscopy a way to observe them through their specific dynamics. The water molecules at the surface of proteins... 

---