Preparation of High Quality Protein Crystals by

1
Preparation of High Quality Protein Crystals by
Surface Entropy Reduction - Tyrosine as a
Crystal Contact Catalyst. Tomek Boczek, David R.
Cooper, Kasia Grelewska, Gosia Pinkowska, Gosia
Sikorska, Michal Zawadzki, and Zygmunt
DerewendaPSI Center for Structure and Function
Innovation and Department of Molecular
Physiology and Biological Physics University of
Virginia. Charlottesville, VA, 22908
The Results
The Screens
Abstract
Crystallization is a limiting step in
macromolecular crystallography. In cases where
proteins are particularly recalcitrant to
crystallization efforts, mutational modification
of surface properties may be essential. We
previously suggested that targeted replacement of
clusters of residues with high-conformational
entropy (lysines, glutamates and/or glutamines)
with alanines leads to formation of epitopes that
are capable of mediating crystal contacts. This
is because the entropic cost of immobilizing
large side chains at the intermolecular contact
regions has been reduced and crystal contacts can
be formed by the mutated epitopes. This Surface
Entropy Reduction (SER) method has facilitated
the crystallization and structure determination
of a number of novel proteins and has also led to
the discovery of crystal forms that diffract to
significantly higher resolution than the
wild-type form. However, it has not been
conclusively demonstrated if alanine constitutes
the best choice for replacement of high-entropy
residues. Here we present a systematic study of
the replacement of nine Lys/Glu-rich patches in
RhoGDI with four target residues Ser, Thr, His
and Tyr. All four amino acids are known to occur
at interfaces with significantly higher incidence
than Lys or Glu / Gln, and may mediate weak
protein-protein interactions leading to crystal
formation. Our results show that tyrosine is a
particularly good choice for the target amino
acid, with threonines and histidines also
performing quite well. The mutated residues
often participate in crystal contacts, in both
homotypic (symmetric) or heterotypic
(head-to-tail) intermolecular interactions. We
also examined a crystallization method proposed
by Janet Newman that replaces the normal
crystallization reservoir solutions with 1.5 M
NaCl. The results are very promising with more
than half of the mutants in this series yielding
more crystals when salt was used as the reservoir
solution. Moreover, this method greatly
increased the variety of conditions that yielded
crystals, with little overlap of the conditions
that yielded crystals for the two types of
screens. This suggests a crystallization
strategy for proteins for which crystallization
is the major bottleneck. Creating several
mutants by replacing patches glutamates and
lysines with two or more target residues and
conducting screens with normal and alternate
reservoir solutions greatly increases the chances
of obtaining diffraction quality crystals

• The Super Screen
• We use a custom 96 well screen we had Nextal
  generate for us.
• For details see,
• http//ginsberg.med.virginia.edu/dcoop/Superscree
  n
• We use sitting drop setups. For all of our
  screens, the protein was concentrated to 15
  mg/ml.
• Standard Screens
• Drops are mixed 11 with protein solution and
  Super Screen reagent.
• Reservoir is 100 ?l of Super Screen reagent.
• Salt Screens
• As above, drops are mixed 11 with protein
  solution and Super Screen reagent.
• Reservoir is 100 ?l of 1.5 M NaCl
• This has the advantage of significantly reducing
  the cost of crystallization setups. 10 ml screen
  reagent 10,000 drops.

• Things To Note
• Standard Screens
• Tyrosine, by far, scored the most hits. Three
  times as many as the next highest, threonine.
• Almost half of tyrosines hits were for the DY
  mutant, and more than half of threonines hits
  were the IT mutant.
• Histidines were the most consistent, with 7 of 8
  mutants yielding hits.
• Salt Screens
• Once again, tyrosine produces the most hits.
• In 20 of 32 mutants, the salt screen produced as
  many or more hits than the standard screen.
• Threonines were the most consistent, with 7 of 8
  mutants yielding hits.
• The two screens yielded crystals in different
  conditions. There are very few conditions that
  produce hits in both screens.
• Overall
• Of the 16 screens performed for each target
  residue, threonine and histidine had the most
  screens (12) yielding crystals. Tyrosine had 11.
• Threonine was the only target residue that
  produced a hit for every mutant when both screens
  were performed. The other target residues
  yielded crystals for 7 of 8 mutants.
• There seems to be a preference for a particular
  target residue for some mutants. The D mutant
  worked best with tyrosine and the I mutant worked
  best with threonine.
• For each mutant, 3 of the 4 target residues
  yielded crystals when both screens were
  performed. Therefore, if we had used two target
  residues with any one mutant, we would have found
  at least one crystal producing condition.

The Crystals
Experimental Design
Old and New RhoGDI Crystal Forms
Previous Xtals
Model Protein RhoGDI (66-204) RhoGDI posses all
of the characteristics of proteins that lend
themselves to crystallization by the SER method.
It expresses and purifies well and can be
concentrated easily, but wild-type RhoGDI is
difficult to crystallize. Additionally RhoGDI is
rich in lysines (10.1 -- average frequency is
7.2 ) and glutamates (7.9 -- average frequency
is 3.7), giving us many potential mutation sites.
New Xtals
References
DY K138Y, K141Y

• Janet Newman
• Rebecca Page
• Other Crystal Screen Paper
• Zygmunts Methods
• Plug Luki

The Mutant Series Nine mutants were chosen, with
the mutations designed to reduce or eliminate
clusters of high-entropy residues. The mutants
were designated by two letters the first letter
indicates which mutations the mutant contains and
the second letter designates the target amino
acid. Thus, the CY mutant is K135Y, K138Y,
K141Y. The B mutant was discontinued do to low
expression or solubility.
32 PEG 8000, 0.22M (NH4)2SO4, 0.1M Cacodylate
pH6.5 Resolution 2.1Å Spacegroup P21
(a32.0,b55.1,c38.9,?107.5)
Conclusions
CY K135Y, K138Y, K141Y
We have previously shown that replacing large,
highly-entropic residues with alanines can
facilitate crystallization. The experiment
described here demonstrates that other residues
can be used to replace surface exposed lysines
and glutamates. This study reveals that
tyrosines, threonines and histidines are suitable
target amino acids for crystallization by the SER
method. Overall, tyrosines produced the most
impressive statistics for number of hits, but
threonines and histidines appear to be slightly
more consistent. In several cases, one target
residue produced many more hits for a particular
mutant than did the other target residues. This
mutant / target residue correlation leads us to
conclude that a wise strategy for crystallization
by the SER method would be to choose a cluster of
highly entropic residues to mutate, and mutate
them to two or more target residues. Additionally
, this experiment examined a recent approach to
crystal screening that uses an alternate
reservoir solution, instead of the traditional
mother liquor. The alternate reservoir screens
produced even more hits than the traditional
screening method. Additionally, this type of
screening gave rise to crystals in different
conditions than the standard screen did. The
efficacy and efficiency of this type of screening
makes it a tool that any crystallographer should
employ.
All RhoGDI mutants were expressed as a GST-fusion
protein. After cleavage of the fusion protein
with rTEV, GST was removed by size exclusion
chromatography.
GST-RhoGDI
4.0 M Sodium Formate Resolution 2.50Å Spacegroup
C21 (a69.2,b78.0,c50.9,?91.5)
RhoGDI
Initial Purification
Size Exclusion Purification