Page 103 - Macromolecular Crystallography

92 MACROMOLECULAR CRYS TALLOGRAPHY

Soaking times can be as little as a few hours or as long crystals, by observing deterioration (cracking or
as several weeks, but usually on the order of 1 to 4 dissolution) of the crystals. Concentrations of the
days. Soaking times are dependent on temperature heavy-atom reagent and soaking times should be
and heavy-atom compound concentration; at lower adjusted to insure that the crystals do not show
temperatures and heavy-atom concentrations it may serious cracks; minor surface cracks may not be
be necessary to soak for longer periods of time. The detrimental to some crystals. However, some crys-
concentration of the heavy-atom reagent used for tals are very sensitive to most heavy-atom reagents,
derivative preparation will depend on its solubility they tend to shatter and lose their ability to diffract,
in the mother liquor. A good starting value is 1 mM, evenifthereagentsareverydiluteandsoaktimesare
but concentrations as low as 0.05 mM and as high short. This can be overcome by crosslinking the crys-
as 100 mM have been reported. The ideal derivative tals with glutaraldehyde before soaking. A colour
is arrived at by varying soaking time and heavy- change of the crystal during soaking does not always
atom compound concentration. The latter variable mean that the heavy atom has specifically bound,
is more useful since mass action can force the forma- since non-specific binding may also cause the crystal
tion of a derivative even in the case of weak binding to change colour. Non-specific binding can be min-
functional groups. Soaking times as short as 1 h com- imized by back-soaking in the stabilizing solution,
bined with concentrations of 0.3 mM were reported the solution that does not contain any heavy atom.
to produce good mercury derivatives of iron super- Although the soaking method for heavy-atom
oxide dismutase (Ringe et al., 1983). In addition derivative preparation is by far the simplest and
to temperature and heavy-atom compound concen- most common, it is not the only method used. One
tration, the composition of the mother liquor and can first derivatize the macromolecule, and then
pH should be considered. Many of the buffers, crystallize. This procedure is less frequently used
additives, and precipitants used in mother liquors, because of drawbacks such as the inability to pro-
such as tris, phosphate, citrate, β-mercaptoethanol, duce isomorphous crystals due to the disruption of
dithiothreitol, and NH 3 derived from ammonium intermolecular contacts by the heavy atoms. Other
sulphate at high pH, may compete with the pro- frequent problems are the introduction of additional
tein for heavy-atom binding. It may be necessary heavy-atom sites (a potential complicating factor in
at times to transfer crystals into more appropriate phasing) by exposing sites hidden by crystal con-
mother liquor before derivative preparation. For tacts, and changing the solubility of the derivatized
example crystals grown out of ammonium sulphate macromolecule.
may be transferred to lithium sulphate to avoid the Theabovetwomethodsforderivativepreparation
formation of metal–ammonia complexes, and salts have been successfully used in the phasing of both
may possibly be replaced by polyethylene glycol nucleic acids and proteins. However, an additional
(PEG). Such changes in mother liquor are best done methodhasbeenusedforphasingofnucleicacids, in
incrementally and slowly to avoid shocking the crys- which the heavy atom is synthetically incorporated
tals. Also, one should recognize that the solubility into the molecule. For example, Drew et al. (1980)
of heavy atom reagents in the mother liquor and determined the structure of d(CGCG) 2 by incorpo-
their binding to functional groups on the protein, rating 5-bromocytosine into the synthesis of their
are pH dependent. The ideal pH range is 6 to 8; nucleic acid and then using the bromine atoms for
lower pH may result in protonation of glutamic and phasing.
aspartic acids of proteins, while at higher pH many
heavy-atom reagents are labile and form insoluble
6.6 Assessment of derivative formation
hydroxides.
As indicated above, the search for suitable heavy- As might be expected, not every crystal soaked in
atom derivatives is as empirical as searching for a solution containing a heavy-atom reagent will be
crystallization conditions. To speed up the process a derivative. Recently, Garman and Murray (2003)
and to save time, initial scanning for suitable heavy- have summarized many of the techniques used in
atom derivatives can be done visually, using small evaluating derivative formation; these include mass