Page 58 - Macromolecular Crystallography

AU TOMATION OF CRYS TALLIZATION TECHNIQUES 47

Ducruix and Giegé, 1999; Chayen, 2005; Bergfors, Chayen, 2003). Uses of such phase diagrams are
1999). However, the ideal experiment does not often described in Sections 3.5.1 to 3.7.
happen and more often than not, excess nucleation
occurs, resulting in the formation of numerous,
low-quality crystals. The aim is therefore to devise 3.4 Automation and miniaturization
methods that will enable the experimenter to lead of screening procedures
the experiment from nucleation to growth, in order
There have never been any set rules or recipes
to ensure the desired results.
to determine where to start when attempting to
crystallize a new protein, hence the most common
means to get started is by using multifactorial
3.3 Practical considerations screens, in other words by exposing different
concentrations of the protein to be crystallized to
Although phase diagrams offer one of the basic and
numerous different crystallization agents, buffers,
most important pieces of knowledge necessary for
temperatures, etc., usually in combinations that
growing protein crystals in a rational way, they
have been successful with other proteins. Automa-
are not often employed because accurate quanti-
tion of screening procedures (e.g. Chayen et al., 1990;
tative phase diagrams require numerous solubility
Weber, 1990; Soriano and Fontecilla-Camps, 1993)
measurements. (The solubility is defined as the
began in the 1990s but only became wide spread
concentration of protein in the solute that is in
with the appearance of structural genomics, for
equilibrium with crystals.) Reaching equilibrium
which automation has become vital. Robotics are
can take several months because proteins diffuse
now available for most methods of crystallization
slowly. An additional limiting factor is that solubil-
(e.g. Luft et al., 2003; Mueller et al., 2001; Walter et
ity measurements require large amounts of sample
al., 2003; Santarsiero et al., 2002; Chayen et al., 1992).
(Ataka, 1993; Chayen et al., 1996).
The area of conditions called the ‘metastable
zone’ is situated between the solubility and super- 3.4.1 The microbatch method
solubility curves on the crystallization phase dia-
gram (Fig. 3.1). The supersolubility curve is defined The first semi-high-throughput automated system
as the line that separates the conditions where spon- to dispense crystallization trials of less than 1µl
taneous nucleation (or phase separation or precipi- was designed in 1990 to deliver batch trials under
tation) occurs, from those where the crystallization oil (Chayen et al., 1990). The method was named
solution remains clear if left undisturbed (Ducruix microbatch to define a microscale batch experiment.
and Giegé, 1992; Ducruix and Giegé, 1999). It was designed to obtain maximum information
The supersolubility curve is less well defined on the molecule to be crystallized while using
than the solubility curve but, experimentally, it is minimal amounts of sample. In order to prevent
found to a reasonable approximation much more the evaporation of such small volumes, the trials
easily. It has been reported that for practical pur- are dispensed and incubated under low density
3
poses, it is sufficient to obtain the supersolubility (0.87 g/cm ) paraffin oil (Fig. 3.2). The crystalliza-
curve. To construct it, one must set up crystalliza- tion drops remain under the oil since the aqueous
tion trials, varying at least two conditions (one of drops are denser than the paraffin oil.
which is typically the protein concentration) and Microbatch can be performed either manually or
plot their outcomes on a two-dimensional param- automatically (Chayen et al., 1992). It is the simplest
eter grid. The supersolubility curve can be obtained crystallization method and therefore can be easily
rapidly using robots and can aid in the separation performed in high-throughput trials. Current robots
of nucleation and growth using seeding and other can dispense microbatch trials down to 1 nl volumes.
means. A diagram containing the supersolubility Depending on the type of oils used to cover the trials,
curve (and not the solubility curve) is called a ‘work- this technique can be harnessed for both screening
ing phase diagram’ (Chayen, 2005; Saridakis and and optimization experiments.