Brookhaven symposia in biology最新文献

Both imaging and simple converging collimator systems can provide a focused neutron beam. Each system has been tested and appears suitable for exploitation. Imaging systems employing bent multilayer monochromators can be made simply by elastically bending a flat multilayer. The periodic spacing of the multilayer can be made to match that of the sample. The converging Soller slit provides a purely geometric means of prouducing a focused beam. Wavelength spread and distribution across the beam can be controlled separately. Soller-slit technology is further advanced than that of multilayers. The experiments sketched above demonstrate that converging or focused neutron beams suitable for low-angle scattering work can be produced with existing technology. Focusing incurs little or no loss of intensity and does not seriously distort the resolution function of a low-angel instrument. It permits a small spectrometer to make use of both advanced detector technology and large samples and thus to approach the capability of a much larger machine.

The use of a pyrolytic graphite analyzer is shown to contribute to the cleanliness of a neutron protein crystallographic study. If neutrons of approximately 1.5-A wavelength are used, higher orders are reduced by nearly an order of magnitude, and background (arising largely from incoherent inelastic neutron-proton scattering) is reduced by nearly a factor of five. These advantages are gained at the expense of approximately 50% of measured integrated intensity and a distortion of integrated intensity with scattering angle. Because background scatter is generally large compared with peak reflectivity of a protein, the large background reduction by the analyzer more than compensates for reduced peak intensity to improve the statistics of most peak reflectivity measurements. The luminance function distortion of intensity data is not large, produces a slight smearing of atomic scattering density, and can be calculated and adjusted for in the data.

The tangent refinement technique has been shown to work marginally for protein neutron data, at least to the extent that the images of all atoms except hydrogen are sharpened. However, the process is very sensitive to over-refinement, and all the maps were sharpened excessively. Refinements in the squared structure were more stable, and the maps were somewhat better than those from the original structure refinement. This technique might be helpful in certain instances, but the poor quality of the maps prevents it from being generally useful. Perhaps the tangent refinement technique will be more fruitful in application to neutron protein data when experimental phase information is available.

We have described here a direct method for the solution of the phase problem for lamellar neutron diffraction from hydrated oriented multilayers of model membranes. As a result, the neutron-scattering profiles of such model membranes can be uniquely determined. We note that such methods can easily be extended to cases in which the thickness of the water layer w is somewhat greater than d/2 as long as it is substantially less than d [e.g., w less than 3d/4 would suffice, since deltaP(-d/4 less than or equal to x less than or equal to d/4) would contain only contributions from see article without overlapping contributions from see articles and to cases in which isotopic labeling is used instead of partial H2O-D2O exchange.

The refinement results in optimum positioning of atoms in the density in most cases, and the juxtaposition of positive and negative density leads to high contrast in the maps and rapid convergence of the refinement. The real-space technique is particularly useful in handling the problem of peak overlap, which is difficult in the neutron case. All refinements to date have been done on the original map, which was phased without hydrogen and deuterium atoms, but the next series of refinements will contain H and D atoms, hydrogen exchange information, and structured water molecules. Comparison of the refinements of CO, Mb and met Mb have shown the iron atom to move nearly into the heme plane. No deuterium atom exists to bond the distal histidine to the CO ligand. Atomic coordinates of all atoms and water molecules will be published on completion of the refinements.