A Neutron Diffraction Analysis of Myoglobin. III. Hydrogen-Deuterium Bonding in Side Chains

A structural analysis of proteins that uses neutron (Schoenborn, 1969a) instead of X-ray diffraction should enable the positioning of hydro-gen atoms and determine hydrogen-deuterium exchange. This is possible due to the relatively large but negative neutron-scattering factor for hydrogen (Table 1). The magnitude of neutron-scattering factors does not vary significantly enough to use the heavy atom phasing technique. A small number of isotopes (11aCd, 149Sm, 15'~Gd, 157Gd) do, however, exhibit relatively large anomalous dispersion that can be used for phase determination. Since the structure as determined by X-ray diffraction is often known before a neutron study is begun, it should be possible to use that information to determine an accurate set of neutron phases. This would certainly simplify data collection and would remove the difficulty of making an anomalously scattering protein deriva-tive. There are a number of ways in which neutron phases could possibly be determined from X-ray structural data. First, approximate neutron phases can be calculated from the known structure and would provide the starting point for a classical Fourier (Lipson and Cochran, 1966a) refinement. The resulting "neutron structure " could then be further efined by Diamond's real space refinement technique adapted to Fourier maps with positive and negative density. "Tangent refinement " (Lipson and Cochran, 1966b) is a possible alternative and would again use calculated neutron phases as a starting point. This process, however, would have to utilize ~,2 due to the presence of negative scattering factors. For completely unknown structures there also is the possibility of determining phases by extending