X-RAY CRYSTALLOGRAPHY 93
synchrotron radiation. Anomalous scattering occurs when the electrons
in atoms can no longer be considered free electrons, but rather electrons
affected by the X rays themselves. The effect is negligible for light atoms
C, N, and O but becomes useful for heavier atoms (S onwards). Anoma-
lous scattering can be exploited for protein phase angle determination
by the multiple wavelength anomalous dispersion (MAD) method.
Determination of the absolute confi guration of a protein structure is also
possible.
- The molecular replacement method assumes similarity of the unknown
structure to a known one. This is the most rapid method but requires the
availability of a homologous (similar) protein ’ s structure. The method
relies on the observation that proteins that are similar in their amino
acid sequence (homologous) will have very similar folding of their poly-
peptide chains. This method also relies on the use of Patterson functions.
As the number of protein structure determinations has been increasing
rapidly, the molecular replacement method becomes extremely useful
for determining protein phase angles. - The so - called direct methods rely on the principles that phase informa-
tion is included in the intensities, that electron density is always positive,
and that the crystal contains atoms that are or may be considered equal.
Phase relations based on probability theory have been formulated and
applied to clusters of refl ections. Direct methods are standard techniques
for determining phase angles in smaller molecules ( < 100 atoms). Recent
advances in theory by H. A. Hauptman of the Hauptman – Woodward
Medical Research Institute and Department of Structural Biology,
School of Medicine and Biomedical Sciences, State University of New
York at Buffalo have allowed the structures of small proteins to be solved
using the so - called Shake and Bake statistical method. In essence, the
Shake and Bake method executes phase angle determination in small
protein molecules by applying it to the strongest refl ections only. More
information on this developing technique can be found in the reference
18 articles and on the website http://www.psc.edu/science/Hauptman/
Hauptman.html.
A preliminary structural model of a protein is arrived at using one of the
methods described above. Calculated structure factors, | Fcalc |, based on the
model generally are in poor agreement with the observed structure factors
|Fobs |. The agreement is represented by an R - factor defi ned as found in equa-
tion 3.11 , where k is a scale factor.
R
FkF
F
hkl
hkl
=
−
×
∑
∑
obs calc
obs
100 (3.11)
Refi nement takes place by adjusting the model to fi nd closer agreement
between the calculated and observed structure factors. For proteins the