For good agreement with experimental solution scattering data from proteins, the calculation of the scattering profile from atomic coordinates must include an estimate of the contribution of the hydration layer. For X-rays, it has been shown that a hydration layer 3 Å thick, with an electron density ~10% greater than the bulk solvent yields good agreement between experimental data and profiles calculated from atomic coordinates for representative proteins [1]. The assumption for calculation of neutron scattering profiles is that the hydration layer has the same basic characteristics – 3 Å thick, with neutron scattering length density ~10% greater than the bulk solvent [2]. However, it is often reported that SANS is insensitive to the presence of a hydration layer around the protein, as better fits to the data are often found by ignoring the hydration layer.
A recent round robin study on a variety of well-behaved protein was carried out [3], in part to help benchmark methods used to calculate scattering profiles of proteins from atomic coordinates. Both SANS and SAXS data were collected, and a variety of different software packages were used to fit the experimental scattering data. Among the different packages, WAXSiS [4] employs an approach whereby the structure of the hydration layer of a protein is determined using molecular dynamics and then explicitly included in the calculation of the scattering profile. The WAXSiS derived scattering curves do an excellent job reproducing the features of the experimental data. The explicit treatment of the hydration layer is computationally expensive, at least in the context of rigid-body structural optimisations, where millions of calculations are typically carried out to yield a model consistent with the experimental data. However, it may give some insight into the apparent paradox where the hydration layer is seemingly visible in SAXS data, but not in a SANS data.
Results from the analysis of MD simulations, and calculation of SANS profiles for a number of well characterised proteins will be discussed, together with the influence of H/D exchange on the calculation.