Phasing Proteins at Low Resolution

Klas Andersson and Sven Hovmöller

An ab initio method which predicts the phases of the very low-order reflections for protein crystals has been developed. The method is a ‘direct method’ since no other data than the native data are used and can be an alternative when multiple isomorphous replacement and molecular replacement are insufficient.
    The method relies on three major assumptions at very low resolution:
· Many proteins are spherical or can at least be approximated as being globular.
· The phases of a point scatterer at the centre of gravity of a protein will be the same as those of the entire protein molecule, out to a well-defined d-value which is described by the G-function (the Fourier transform of a sphere).
· The electron density is rather smooth since carbon, nitrogen, and oxygen almost have the same number of electrons.
   The position and a crude shape of the protein molecule can be found, if the low-order reflections are accurately measured. The low-order reflections often have the highest |E|-values and are members of many triplet and quartet relations. Thus, they will play an important role in phase extension.

    References:

Andersson,K.M. and Hovmöller,S. Phasing Proteins at low Resolution Acta Cryst D52 (1996) 1174-1180.

Andersson, K.M. and Hovmöller, S. The average Atomic Volume and Density of Proteins. Zeitschrift für Kristallografie 213 (1998) 369-373.

Andersson,K.M. and Hovmöller,S. The protein content in crystals and packing coefficients in different space groups. Acta. Cryst D56 (2000) 789-790.