Acta Cryst. (2005). D61, 1173–1180 doi:10.1107/S0907444905019451 1173
Acta Crystallographica Section D
Post-crystallization treatments for improving
diffraction quality of protein crystals
a Heras* and Jenni fer L.
Institute for Molecular Bioscience and ARC
Special Research Centre for Functional and
Applied Genomics, University of Queensland,
Brisbane QLD 4072, Australia
# 2005 International Union of Crystallography
Printed in Denmark – all rights reserved
X-ray crystallography is the most powerful method for
determining the three-dimensional structure of biological
macromolecules. One of the major obstacles in the process is
the production of high-quality crystals for structure determi-
nation. All too often, crystals are produced that are of poor
quality and are unsuitable for diffraction studies. This review
provides a compilation of post-crystallization methods that
can convert poorly diffracting crystals into data-quality
crystals. Protocols for annealing, dehydration, soaking and
cross-linking are outlined and examples of some spectacular
changes in crystal quality are provided. The protocols are
easily incorporated into the structure-determination pipeline
and a practical guide is provided that shows how and when to
use the different post-crystallization treatments for improving
Received 19 April 2005
Accepted 22 June 2005
Knowledge of the three-dim ensional structure of proteins is
essential in unde rstanding their biological role and underpins
downstream applications such as the design of new drugs.
More than 85% of the macromolecular structures currently
held in the Protein Data Bank have been determined by X-ray
crystallography, making X-ray diffraction by far the most
successful method for determining the structures of large
molecules. A major stumbling block in this approach is the
requirement for high-quality crystals. Despite technical and
methodological advances in the ﬁeld, including the develop-
ment of highly focused X-rays, high-throughput crystallization
techniques and automated X-ray data analysis, crystal growth
still remains an empirical and tedious process and a common
occurrence is the production of well formed crystals that are
not suitable for diffrac tion studies. Loose packing of molecules
and large solvent volume are common problems that result in
low-resolution and poor-quality diffraction.
What are the options if crystals form but do not diffract
well? Strategies that have been used in the past include
searching for new crystallization conditions to identify a new
crystal form, crystallizing the protein of interest from a
different organism, crystallizing a different form of the protein
by using proteases that produce smaller protein fragments,
generating new constructs encoding a truncated form of the
protein or mutating surface amino acids to enhance protein
crystallization (Longenecker et al., 2001). These methods all
have one thing in common: they each give up on the crystals
that have alrea dy been grown. However, before capitulating,
there are several quick and simple methods that could be
considered. Post-crystallization soaking, cross-linking, crystal
annealing and controlled dehydration have been reported to
a Heras obtained her PhD degree from
The University of Navarra (Spain), for her work
on computer-assisted antidepressant drug
design. She is currently a postdoctoral scientist at
the Institute for Molecular Biosicence at the
University of Queensland (Australia). Her
research interests include protein structure and
function particularly with regard to protein
folding, protein–protein interactions and drug
design. Jenny Martin is an Associate Professor at
the Institute for Molecular Bioscience, University
of Queensland. Her research interests include
protein structure, folding and evolution and
protein:ligand interactions. Her technological
interests include high throughput crystallography
and drug design. Jenny received a DPhil from
the University of Oxford, UK.
dramatically improve diffraction resolution of protein crystals.
This review provides an overview of the different methods
that can be used when faced with the all-too-frequent situation
of protein crystals with poor-quality diffraction and provides a
practical guide to implementing the procedures.
2. Crystal annealing
Protein crystals are sensitive to X-ray radiation and their
diffraction quality rapidly deteriorates after being exposed to
high-intensity X-rays. Data collection at cryogenic tempera-
tures (around 100 K) has become a vital method for protein
crystallography since it reduces radiation damage and
increases crystal lifetime, allowing whole data sets to be
collected from a single crystal (Hope, 1988, 1990; Rodgers,
1997; Garman, 1999). Protein crystals typically contain 50%
water by mass and if cooled too slowly the water undergoes a
large volume change associated with the transition from liquid
water to hexagonal ice, resulting in degradation of crystal
quality (Kriminski et al., 2002). Ice formation and consequent
crystal damage can be avoided by ﬂash-cooling the crystal to
produce an amorphous or vitriﬁed form of water. Never-
theless, ﬂash-cooling techniques can introduce lattice disorder,
resulting in increased mosaicity and reduced diffraction
resolution (Rodgers, 1994; Garman & Schneider, 1997). This
damage, which is especially acute for large crystals and crystals
with high solvent content, is thought to be a consequence of
uneven cooling and differential expansion of the solvent and
the crystal lattice (Juers & Matthews, 2001; Kriminski et al.,
Approaches for reducing mosaic spread caused by ﬂash-
cooling include a systematic exploration of cryoprotectants,
sequential soaking of crystals in increasing concentrations of
cryosolutions and control of the ﬂash-cooling rate (Garman,
1999). However, a rapid and easy method to reduce ﬂash-
cooling-induced disorder that can increase diffraction quality
of protein crystals has been described. This metho d, called
crystal annealing, involves warming the ﬂash-cooled crystal t o
room temperature and ﬂash-cooling it again (Harp et al.,
1998). Three different crystal-annealing protocols have been
reported. Macromolecular crystal annealing (MCA) consists
of removing a cryocooled crystal from the cold gas stream and
placing it in 300 ml cryosolution. After 3 min equilibration, the
crystal is recoo led in the cryostream (Fi g. 1) (Harp et al., 1998).
The ﬂash-annealing (FA) metho d involves blocking the cold-
stream for 1.5–2 s three times with intervals of 6 s between
each thawing step (Fig. 1) (Yeh & Hol, 1998). Annealing on
the loop (AL), a variation of the ﬂash-annealing method, also
involves blocking the cryostream, but in this case the length of
time varies from crystal to crystal: the cold-stream is blocked
until the drop becomes clear, which is an indication that it has
reached ambient temperature. When the cold nitrogen-gas
stream is blocke d, the ﬂash-cooled drop becomes covered in
ice, thus turning the drop opaque. As the drop warms to room
temperature, the drop becomes clear (Harp et al., 1999). AL
does not use multiple cycles of warming and ﬂash-cooling; a
single annealing step will do (Fig. 1). Harp and coworkers used
crystals from six different proteins to evalua te the three
annealing protocols and concluded that MCA treatment gives
better and more reproducible results, FA was inadequate for
most crystals and AL was successful with small crystals with
low solvent content (Harp et al., 1999). A more recent
annealing protocol, also carried out in the loop, restricts the
warming temperature to below the bulk-solvent melting point
(230–250 K) and this was reported to give more reproducible
improvement in the diffraction quality of protein crystals
(Kriminski et al., 2002). In this case, the temperature of the
cryostream is quickly increased to 230–250 K and maintained
until the crystal is equilibrated with the mix of cold and warm
gas (10 s) prior to recooling. Application of this approach
requires a system that permits regulation of the temperature in
A number of researchers have reported success in
extending the diffraction resolution and reducing mosaicity of
protein crystals by crystal annealing. Table 1 summarizes
several examples, providing information about the speciﬁc
methods used as well as the improvements achieved. In some
cases, this post-crystallization treatment has produced spec-
tacular results. For example, in the case of three different
proteins, N-acetylglucosamine 6-phosphate deacetylase
(GlcNAc6P), dmpFG-encoded 4-hydroxy-2-ketovalerate
aldolase-aldehyde dehydrogenase and arsenate reductase, the
diffraction resolution improved from 6A
tion in the case of GlcNAc6P) to 2A
upon crystal annealing
(Table 1). Furthermore, for copper nitrite reductase and
F1-ATPase the annealing procedure reduced the mosaicity
from more than 1
3. Crystal dehydration
Water plays a crucial role in maintainin g the structure and
activity of protein molecules both in solution and in crystalline
form (Frey, 1994; Timasheff, 1995). Crystallographers have
investigated in detail the water-mediated transformations in
protein crystals and it is well known that reduction of solvent
content can produce more closely packed and better ordered
1174 Heras & Martin
Post-crystallization treatments Acta Cryst. (2005). D61, 1173–1180
Annealing. Schematic outline of crystal-annealing procedures that can
improve crystal quality. Macromolecular crystal annealing (MCA):
remove cryocooled crystal from the cryostream, place it in 300 mlof
cryosolution for 3 min and then recool (Harp et al., 1998). Flash annealing
(FA): block the cold-stream for 1.5–2 s three times with intervals of 6 s
between thawing steps (Yeh & Hol, 1998). Annealing on the loop (AL):
block cold-stream until the crystal becomes clear and then ﬂash-cool
again (Harp et al., 1999).
crystals, extending the resolution of X-ray diffraction patterns
(Salunke et al., 1985; Frey, 1994). Indeed, after reviewing
different post-crystallization methods to improve crystal
diffraction, crystal de hydration emerges as the treatment that
has produced the most remarkable improvements in the
diffraction resolution of protein crystals (Heras et al., 2003;
Abergel, 2004). Some of these examples are summarized in
Table 2. It should be noted that while our literature search
yielded these 18 exampl es, this is by no means a comprehen-
sive list and there may be other cases that we have missed.
Several different protocols have been developed for crystal
dehydration. The ﬁrst dehydration experiments were carried
out by Perutz, who allowed crystals mounted in a capillary
tube to lose water to the atmosphere through a small hole
made in the tube (Perutz, 1946; Bragg & Perutz, 1952). This
method was further improved by connecting the glass capillary
to a salt reservoir so that by changing the salt concentration
the relative humidity in the capillary could be varied and
different crystal forms produced (Huxley & Kendrew, 1953).
In that paper, Huxley and Kendrew also described an appa-
ratus for controlled shrinkage of protein crystals. In another
example, Madhusudan and coworkers reduced the solvent
content in monoclinic lysozyme from 33 to 22% by placing a
few drops of K
solution in the glass capillary; this
improved the diffraction resolution from 2.5 to 1.75 A
(Madhusudan et al., 1993). Similarly, the response of tetra-
gonal lysozyme to dehydration was characterized by mounting
crystals in X-ray capillaries and equilibrating them against
different saturated salt solutions (various salts were used to
achieve relative humidities ranging from 97 to 75%;
Dobrianov et al., 2001).
Crystal dehydration can also be performed by transferring
the crystals into a dehydrating solution, which is usually the
original mother liquor either with a higher concentration of
precipitant or supplemented with cryoprotective agents such
as PEG 400, PEG 600, glycerol or MPD. It should be noted
that one mechanism of cryoprotection is to reduce protein
solvation, so that cryoprotectants can act as dehydratants and
vice versa. On the other hand, cryoprotectants such as glycerol
can also act as co-solvents, increasing protein solubility.
Further information about cryoprotectants and their
mechanisms can be found in Timasheff & Arakawa (1988),
Charron et al. (2002) and Mi et al. (2004).
The ﬁrst example of crystal dehydration performed by
moving the crystal into dehydrating solution was reported by
Schick & Jurnak (1994). They improved the diffraction reso-
lution of the guanidine nucleotide-exchange factor complex
from 4 to 2.5 A
(Table 2) by a serial transfer of the protein
crystal to droplets (50 ml) of cryoprotective agent of increasing
concentration, with incubations of 5 min in each condition
(method 1; Fig. 2). In contrast, Haebel et al. (2001) extended
the diffraction resolution of crystals of a DsbC–DsbD
complex from 7 to 2.6 A
resolution (2.3 A
resolution at a
synchrotron; Table 2) by the slow addition to the crystal
droplet of a total of eight times the crystallization drop volume
of dehydrating solution followed by equilibration against air
for a period of 30 min (method 2; Fig. 2). Alternatively, the
crystal can be transferred from the crystallization drop into a
5 ml hanging drop of dehydrating solution, which is then
equilibrated for 12–16 h against the same dehydrating solution
at 277 K (method 3; Fig. 2; Heras et al., 2003). Crystals of
DsbG dehydrated in this way exhibited a dramatic improve-
ment in diffraction quality, with the pattern improving from
to beyond 2 A
resolution (Table 2). A more gentle
dehydrating method consists of the serial transfer of the cover
slip holding the crystal droplet over reservoirs containing
increasing concentrations of dehydrating solution with incu-
bations of 8–12 h over each condition (method 4; Fig. 2).
Acta Cryst. (2005). D61, 1173–1180 Heras & Martin
Post-crystallization treatments 1175
Summary of the effect of crystal annealing on different protein crystals.
EG, ethylene glycol; MME, monomethyl ether; MPD, 2-methyl-2,4-pentanediol; NR, not reported; Paratone, Paratone-8277 (previously known as Paratone-N);
PEG, polyethylene glycol.
group Precipitant‡ Cryoprotectant
Nucleosome core particle P2
22.5% MPD/Paratone 51 NR/3.1 0.82/0.34 MCA a
Histone octamer P3
21 70% (NH
15% glycerol/Paratone 65 NR/3.0 0.34/0.22 MCA a
Glycerol kinase C2 35% PEG 4K/7.5% PEG 200 20% EG NR 3.7/2.8 >2/1FA b
Inorganic pyrophosphatase R32 1.4–2 M NaCl 27–30% glycerol 40 1.8/1.15 0.70/0.30 MCA [salt] c
2 1.85 M NaH
20% glycerol 46 Medium/2.0 NR MCA d
15% PEG 8K Paratone 48 >6/2.1 NR MCA e
20–25% PEG 4K 20% PEG 4K/5% glycerol 35 3/2 NR FA f
Arsenate reductase P2
30–35% PEG MME 2K 30–35% PEG MME 2K 47 6.0/2.3 NR FA g
dUTP pyrophosphatase P2
23.4% PEG 1.5K 40% PEG 1.5K 63 3/2.2 NR MCA h
Copper nitrite reductase P6
40–50% PEG MME 550 40–50% PEG MME 550 NR 2.5/1.0 1.5/0.3 AL i
Lipoprotein receptor LolB P2
30% PEG MME 2K 30% PEG MME 2K 44 NR/1.9 NR FA j
TAXI I endoxylanase inhibitor P2
23% PEG 4K 20% glycerol 47 NR/1.75 NR AL k
11% PEG 6K Paratone 55 NR/2.8 1/0.28 AL l
† GlcNAc6P, N-acetylglucosamine 6-phosphate deacetylase; DmpFG, dmpFG-encoded 4-hydroxy-2-ketovalerate aldolase-aldehyde dehydrogenase; dUTP pyrophosphatase,
deoxyuridine triphosphate nucleotidohydrolase. ‡ Does not include information about buffers or additives. § MCA, macromolecular crystal annealing; FA, ﬂash-annealing; AL,
annealing on the loop. } a, Harp et al. (1998); b, Yeh & Hol (1998); c, Samygina et al. (2000); d, Ferreira et al. (2000); e, Manjasetty et al. (2001); f,Ve
et al. (2001); g, Guan et al.
(2001); h, Han et al. (2001); i, Ellis et al. (2002); j, Takeda et al. (2003); k, Sansen et al. (2003); l, Mueller et al. (2004).
Combining dehydration with other post-crystallization
treatments such as annealing, soaking, cryocooling or re-
hydration has also resulted in spectacular improvements in the
diffraction quality and resolution of protein crystals (Table 2)
(Izard et al., 1997; Tong et al., 1997; Pang et al., 2002; Abergel,
2004). For example, a recent publication reported the dramatic
improvement in diffraction resolution of three protein crystals
upon annealing and dehydration [Escherichia coli YbgL from
12 to 2.6 A
, E. coli YggV (HAM1) from 12 to 2.6 A
Candida albicans 3-dehydroquinate dehydratase from no
diffraction to 3 A
]. The method involved removing the poorly
diffracting crystal from the cryostream and plac ing it in a 10 ml
droplet containing dehydrating solution (90% mother liquor
and 10% cryoprotectant such as glycerol or ethylene glycol)
followed by air drying from 15 min to 2 h (Table 2; Abergel,
Dehydration has also proven successful in improving the
diffraction resolution of membrane-protein crystals. The
diffraction limit and quality of prokaryotic CLC chloride
channel crystals improved (resolution from 8 to 4 A
from >5 to 1
) upon slow dehydration over a period of months
(Kuo et al., 2003).
Finally, in addition to these dehydration methods, a device
that controls the humidity surrounding the crystal has also
been described and successfully utilized to improve diffraction
of protein crystals (Kiefersauer et al., 2000).
4. Other methods
4.1. Post-crystallization soaking without dehydration
Post-crystallization soaking is similar to crystal dehydration
in that both processes involve soaking protein crystals in
solutions containing increased precipitant concentrations or
cryoprotectants. However, dehydration implies shrinking of
the crystal lattice and lowering of the solvent content of the
crystal, whereas post-crystallization soaking without dehy-
dration does not involve a change in unit-cell or solvent
content, yet still leads to a notable improvement of the
diffraction quality of the crystal.
Several examples have been described where soaking
crystals in higher ionic strength solutions, cryoprotectants or
heavy-atom-containing solutions results in improvement in
the quality of the crystals. For example, Fu and coworkers
transferred MTCP-1 protein crystals grown in ammonium
sulfate to fresh drops containing a higher salt concentration
and incubated the crystals for one to ﬁve months. This
extended the diffraction resolution from 3 to 2 A
(Fu et al.,
1998, 1999). They postulated that the improvement in
1176 Heras & Martin
Post-crystallization treatments Acta Cryst. (2005). D61, 1173–1180
Summary of the effect of crystal dehydration on different protein crystals.
AS, ammonium sulfate, cryst. drop, crystallization drop; EG, ethylene glycol; exp., exposure; HA, heavy atom; incub., incubation; incr., increment; MPD, 2-methyl-
2,4-pentanediol; PEG MME, PEG monomethylether; NR, not reported; PEG, polyethylene glycol; ppt, precipitant; satd, saturated; sol., solution.
group Precipitant‡ Dehydrating agent
20% PEG 4K 28–40% various PEGs Method 1 (5 min) 61/55 4.0}/2.7} a
NF-B P52–DNA I2
4–6% PEG 4K Ppt+ 30% PEG 400
Method 1 52/49 3.5††/2.0†† b
6% PEG 3.4K 46% PEG 3.4K Method 1 (5% incr.; 3 d) 56/48 3.7}/2.2} c
2 25% PEG MME 5K/
40% PEG MME 5K/
Method 2 (30 min) 55/41 7.0}/2.6} (2.3††) d
DsbG C2 20% PEG 4K 30% PEG 4K Method 3 (12 h) 90/53 10}/2.0} (1.7††) e
E. coli YbgL C20.8M sodium citrate Ppt + 10% EG Annealing/air dehydrate (2 h) NR/57 12}/2.6} (1.8††) f
E. coli YggV P 4
2 35% (NH
37.5% AS/10% glycerol Annealing/air dehydrate (30 min) NR/38 12}/2.6} (2††) f
3-Dehydro dehy P2 11% PEG 8K Ppt + 10% glycerol Annealing/air dehydrate (15 min) NR/88 None/3 f
Rv2002 product P3
21 20% PEG 3K Ppt + 10% MPD Annealing/air dehydrate (5 h) NR/35 2.1††/1.8†† g
Peptide deform P2
12% PEG 4K 20% PEG 4K/
10% PEG 400
Annealing/air dehydrate (30 min) NR/50 2.0††/1.85†† h
HCMV prot P4
2 16% PEG 4K 30% PEG 4K/Na
Method 1 (3–5 d) 58/56 3.0}/2.5} (2.0††) i
PDH R32 6% PEG 3K Ppt Dehydrate/rehydrate NR/73 7.0††/4.2†† j
FAD-indep ALS C2 6–8% PEG 8K/
Ppt/30% PEG 600 Method 3/cryocool (24 h) NR/52 2.9}/2.6} k
sol. Method 4 (20 h) 33/22 2.5}/1.75} l
2 0.48–0.75 M NaCl Satd salt sol. Method 4 (days/weeks) NR 1.6††/3.7†† m
15% PEG 3350 33% PEG 3350 Method 4 (2 h) 58/52 5.0††/2.85†† n
CLC Cl channel P 222 22–32% Jeffamine Ppt Incub. in cryst. drop (5 months) NR 7.5††/4.0†† o
2 14–16% PEG 2K 20% glycerol/20% EG Incub. under oil 2–4 h/
air exp. 10 min
NR/61.7 4.0††/2.3†† p
† EF-Tu-Ts, guanidine nucleotide-exchange factor complex EF-Tu-Ts; NF-B P52–DNA, transcription factor NF-B P52–DNA complex; HIV(RT)–inhib., HIV1 reverse transcriptase–
inhibitor complex; 3-Dehydro dehy, C. albicans 3-dehydroquinate dehydratase; Peptide deform, peptide deformylase; HCMV prot, human cytomegalovirus protease; PDH, pyruvate
dehydrogenase; FAD-indep ALS, FAD-independent acetolactate synthase; Lysozyme, hen egg-white lysozyme; RFC–PCNA, replic ator factor C–proliferating nuclear antigen
complex. ‡ Does not include information about buffers or additives. § a, Schick & Jurnak (1994); b, Cramer & Muller (1997); c, Esnouf et al. (1998); d, Haebel et al. (2001); e, Heras
et al. (2003); f, Abergel (2004); g, Yang et al. (2002); h,Kimet al. (2002); i, Tong et al. (1997); j, Izard et al. (1997); k, Pang et al. (2002); l, Madhusudan et al. (1993); m, Dobrianov et al.
(2001); n, Bowman et al. (2004); o,Kuoet al. (2003); p, Hunsicker-Wang et al. (2005). } X-ray diffraction resolution on a rotating-anode source. †† X-ray diffraction resolution at a
diffraction quality was a consequence of the rearrangement of
surface residues to form better packing interactions. The
method was further optimized by Petock and coworkers, who
varied the composition of the soaking solution, combining
ammonium sulfate with PEG 3400, which reduced the incu-
bation time of MTCP-1 protein crystals to 1–10 weeks (Petock
et al., 2001).
Post-crystallization soaking in solu-
tions containing cryoprotecting agents
such as glycerol can also improve the
quality of protein crystals (Sousa,
1995). Thus, Rould and coworkers
reported that the diffraction limit of
co-crystals could be increased by
soaking in 20% glycerol followed by
cooling to 265 K (Rould et al., 1991).
This treatment also improved the order
in disordered regions of the crystal.
Moreover, in a recent study on the
versatility of malonate as cryoprotec-
tant, it was found that soaking in 50%
sodium malonate solution can elim-
inate crystal disorder and improve the
resolution limit (Holyoak et al., 2003).
Preparation of heavy-atom deriva-
tives for phasing purposes (reviewed in
Garman & Murray, 2003) can involve
the immersion of protein crystals
into heavy-atom-containin g solutions.
Heavy-atom derivative s produced in
this way usually diffract less well than
native crystals, although in some cases
the resolution limit has been shown to
improve. For example, Cramer and
Muller reported that the anisotropic
diffraction exhibited by NF-B P52–
DNA cocrystals could be corrected by
soaking heavy-metal ions into the
crystal (Table 2; Cramer & Muller,
1997). This soaking process was
accompanied by unit-cell shrinkage,
suggesting that dehydration could also
have contributed to the change.
However, improvements in diffraction
resolution without apparent dehydra-
tion have also been reported
upon heavy-atom derivatization. For
example, the diffraction resolution of
MscL protein crystals, a mechano-
sensitive ion channel from Myco-
bacterium tuberculosis, improved from
7 to 3.5 A
resolution upon soaking in
heavy-atom compounds (Chang et al.,
Data collection at cryogenic
temperatures requires the addition of a
Acta Cryst. (2005). D61, 1173–1180 Heras & Martin
Post-crystallization treatments 1177
Dehydration methods. Method 1: serial transfer of crystals from the crystallization drop to 50 ml
drops containing increasing amounts of dehydrating solution. The dehydrating solution can consist of
mother liquor with increasing precipitant concentration or it can be supplemented with increasing
concentrations of cryoprotective agents such as PEG 400, glycerol or MPD. Depending on the
stability of the crystal, the concentration of dehydrating agent can be increased in steps of 5% up to
30%(w/v), steps of 0.5% up to5%(w/v) or as follows: 1, 2, 3, 4, 5, 10, 15, 20%. Soaking time can
also vary from 5 to 15 min (in this case crystals are air dehydrated) to days (in this case dehydrating
drops are equilibrated against a reservoir containing dehydrating solution) (Schick & Jurnak, 1994;
Esnouf et al., 1998). Method 2: add dehydrating solution slowly to the drop containing the crystal
(about eight times the crystallization drop volume) and air dehydrate for more than 30 min.
Dehydrating solution consists of crystallizing conditions with a 10–12% increase in precipitant
concentration (e.g. PEG, MPEG) and 5–10% addition of cryoprotective agent such as glycerol
(Haebel et al., 2001). Method 3: transfer the crystal from the crystallization drop into a 5 ml hanging
drop of dehydrating solution and equilibrate against a reservoir with the same dehydrating solution
(dehydrating solution: crystallization condition containing 5–10% more precipitating agent;
incubation time: 12–16 h; Heras et al., 2003). Method 4: After crystal growth, equilibrate the
crystallization drop against reservoirs containing increasing concentration of dehydrating agent
(dehydrating solution: mother liquor containing increasing concentrations of either precipitant or
low-molecular-weight PEG, glycerol or MPD. Concentration is increased in steps of 5%. Incubation
time: 8–12 h each). For very fragile crystals it is recommended that these dehydration procedures be
performed at 277 K. Crystal soaking without dehydration is performed similarly to method 1 by, for
example, transferring the crystal to 10 ml of soaking solution consisting of mother liquor containing a
cryoprotectant (20% glycerol or 40–50% malonate) and incubating for a few seconds to 5 min
(Holyoak et al., 2003).
Crystal cross-linking by vapour diffusion. Transfer cover slip containing the crystal to a new reservoir
with precipitating solution and a microbridge (Hampton Research) containing a sitting drop (2–5 ml)
of 25% glutaraldehyde pH 3. Equilibrate for 30–60 min and stop the process by placing the cover slip
over a new reservoir with fresh precipitant solution (Lusty, 1999).
suitable cryoprotectant to the protein
crystal. While some crystals can be
directly dipped in cryoprotective solu-
tions, other crystals do not tolerate this
procedure. In this situation, it may be
possible to chemically cross-link the
protein crystal using glutaraldehyde or
another cross-linking reagent before
proceeding with the cryoprotection
procedure (Quiocho & Richards,
1964). Crystal cross-linking increases
the robustness of the crystal against
mechanical stress, reduces its solubility
(Quiocho & Richards, 1964) and can
also improve diffraction quality.
Cross-linking involves the reaction
of lysi ne amines with aldehydes of the
cross-linking (usually glutaraldehyde)
molecule; improvement in diffraction
quality depends on the position of the lysines in the crystal and
the total number of lysines in the asymmetric unit (Lusty,
1999). The cross-linking reaction is pH-dependent: the more
basic the solution the more susceptible the aldehyde is to
condensation, which reduces the number of free aldehyde
groups (Monsan et al., 1975). Reproducibility and loss of
diffraction owing to the use of excessive cross-linking agent
are also common problems with this post-crystallization
treatment. A more gentle cross-linking method has been
described in which glutaraldehyde is introduced into the
crystal by vapour diffusion (Lusty, 1999). The process is
carried out by equilibrating a hanging droplet holding the
crystal over a reservoir containing precipitant solution and a
microbridge (Hampton Research) holding 2–5 ml of 25%
glutaraldehyde (Fig. 3). The process was evaluated on crys tals
of three different proteins. This showed not only that the
process was highly reproducible, but also that crystal cross-
linking can prevent lattice disorder caused by cryocooling.
Crystal diffraction of a selenomethionyl N-cadherin fragment
) and a DNA comple x of MMLV reverse transcriptase
) was similar before and after cross-linking, but for a
HIV-1 gp120 ternary complex diffraction improved from 2.7 to
. Strikingly, mosaicity was substantially reduced in all
three cases after cross-linking (from 2–5
1999). A similar cross-linkin g approach was used to improve
the diffraction properties of protein–DNA complex crystals
nyi et al., 2001). In this case the resolution improved
from 3.2 to 1.9 A
and the mosaicity decreased from 2to0.5
after glutaraldehyde cross-linking.
Another cross-linking method that has been described
involves equilibrating the protein crystal against a reservoir
containing mother liquor supplemented with 0.125% glutar-
aldehyde (2–3 h; Jacobson et al., 1996).
5. Practical suggestions
How can we best make use of these methods to improve
crystal quality? The purpose of this section is to provide a
protocol for selecting the most appropriate post-crystallization
treatment. We have classiﬁed protein crystals into four
different categories according to their diffraction quality at
cryo- and room temperature and for each category suggested
the most suitable treatment (Fig. 4).
5.1. Category 1
This is the ideal case; crystals have high-quality diffraction
patterns at both room and cryo-temperatures (Fig. 4). In this
situation no post-crystallization treatment is necessary,
although crystal dehydration might be worthwhile as a quick
option to further extend crystal diffraction. The choice of
dehydrating method will depend on the robustness of the
crystal. For delicate crystals, more gentle methods are
recommended such as methods 3 or 4 (Fig. 2). For robust
crystals, quicker dehydration methods can be attempted such
as methods 1 or 2 (Fig. 2). In all these methods the dehydrating
agent used is the crystallization condition with a higher
precipitant concentration or with added cryoprotective agents
(such as low-molecular-weight PEGs, glycerol or MPD). In
Fig. 2, speciﬁc concentrations of dehydrating agent are
suggested for each dehydrating method. These values are
based on published data as well as our own experience;
however, these concentrations should be used as a guide only
and may be adapted to the speciﬁc requirements of each
5.2. Category 2
In this case, crystals diffract to high resol ution at room
temperature but poorly at 100 K (Fig. 4). If the cooling process
signiﬁcantly affects the diffraction quality of the crystal, data
collection at room temperature could be an option, although
radiation damage to the protein crystal may limit the
diffraction resolution and the quality of the ﬁnal three-
dimensional structure. If cryo data measurement is preferred,
the post-crystallization treatment to try ﬁrst is crystal
annealing (Fig. 1). Annealing on the loop is the quickest
1178 Heras & Martin
Post-crystallization treatments Acta Cryst. (2005). D61, 1173–1180
Flow diagram depicting practical experiments to perform on crystals with various diffraction
method since it simply involves blocking the cryostream until
the drop becomes clear and then ﬂash-cooling the cry stal
again. Macromolecular crystal annealing is also recommended
since this method has yielded the greatest improvements in
diffraction quality of protein crystals upon annealing. If crystal
annealing does not yield improved results, serial transfer of
the crystal into increasing concentrations of cryoprotectant
can be attempted. This treatment reduces the osmotic shock
suffered by the crystal when exposed to the cryoprotecting
buffer and may improve crystal diffraction at cryo-tempera-
tures. Also, different types of cryoprotectants should be tried
(Garman & Schneider, 1997; Garman, 1999). Oils such as
Paratone-N can also be utilized as a cryoprotectant for
biological macromolecular crystals (Hope, 1988) and these
also can have a dehydra ting effect on protein crystals, which
may improve their diffraction resolution. If no suitable cryo-
protectant can be found, then cross-linking of protein crystals
prior to cryoprotection is recommended (Fig. 3).
5.3. Category 3
This category includes those crystals that diffract poorly at
both 298 and 100 K. Here, it is worthwhile investigating
different post-crystallization treatments. Crystal dehydration
should be attempted ﬁrst since, to the best of our knowledge,
this is the post-crystallization treatment that has resulted in
the most dramatic improvements in diffraction resolution of
protein crystals (Table 2, Fig. 2). On ce again, the selection of
the dehydration method will depend on the stability of the
crystal. For salt-grown crystals, soaking in higher salt
concentrations supplemented with 5–10% PEG 3350 should
be tried. Crystal soaking in malonate is also a good option
since this may not only improve the diffraction resolution, but
this particular salt can also act as a cryoprotectant. For crystals
in category 3, annealing only is unlikely to improve the
diffraction quality since their poor resolution is not entirely a
consequence of the cooling process. However, combinations of
one of the previous treatments (dehydration and soaking)
with crystal anneal ing are recommended (Fig. 4).
5.4. Category 4
In the unlikely case that crystals diffract well only at
cryo-temperature but not at room temperature, no post-
crystallization treatment is required. However, we suggest
investigating crystal annealing since the thawing and recooling
process may further increase resolution. Also, a combination
of annealing with either crystal soaking in cryoprotectant or
crystal dehydration might be worth trying to further extend
the diffraction resolution.
On a ﬁnal note, given the spectacular results obtained by
crystal dehydration and the simplicity of this particular
approach, we recommend using this procedure routinely as a
method to potentially improve the diffraction limit of protein
crystals. Crystals can be dehydrated in many different ways
and the selection of the method will depend on the stability of
the crystal. For robust crystals, methods 1 and 2 are recom -
mended, whereas methods 3 and 4 are more suitable for
delicate crystals. It should also be noted that methods 2 and 4
could easily be incorporated into a high-throughput structure
This work was funded by grants from the Australian
Research Council to BH and JLM and a University of
Queensland Early Career Researcher Grant to BH.
Abergel, C. (2004). Acta Cryst. D60, 1413–1416.
Bowman, G. D., O’Donnell, M. & Kuriyan, J. (2004). Nature
(London), 429, 724–730.
Bragg, W. L. & Perutz, M. F. (1952). Acta Cryst. 5, 323–328.
Chang, G., Spencer, R. H., Lee, A. T., Barclay, M. T. & Rees, D. C.
(1998). Science, 282, 2220–2226.
Charron, C., Kadri, A., Robert, M. C., Giege
, R. & Lorber, B. (2002).
Acta Cryst. D58, 2060–2065.
Cramer, P. & Muller, C. W. (1997). FEBS Lett. 405, 373–377.
Dobrianov, I., Kriminski, S., Caylor, C. L., Lemay, S. G., Kimmer, C.,
Kisselev, A., Finkelstein, K. D. & Thorne, R. E. (2001). Acta Cryst.
Ellis, M. J., Antonyuk, S. & Hasnain, S. S. (2002). Acta Cryst. D58,
Esnouf, R. M., Ren, J., Garman, E. F., Somers, D. O., Ross, C. K.,
Jones, E. Y., Stammers, D. K. & Stuart, D. I. (1998). Acta Cryst.
Ferreira, F. M., Mendoza-Hernandez, G., Calcagno, M. L., Minauro,
F., Delboni, L. F. & Oliva, G. (2000). Acta Cryst. D56, 670–672.
Frey, M. (1994). Acta Cryst. D50, 663–666.
Fu, Z. Q., Du Bois, G. C., Song, S. P., Harrison, R. W. & Weber, I. T.
(1999). Acta Cryst. D55, 5–7.
Fu, Z. Q., Du Bois, G. C., Song, S. P., Kulikovskaya, I., Virgilio, L.,
Rothstein, J. L., Croce, C. M., Weber, I. T. & Harrison, R. W. (1998).
Proc. Natl Acad. Sci. USA, 95, 3413–3418.
Garman, E. (1999). Acta Cryst. D55, 1641–1653.
Garman, E. & Murray, J. W. (2003). Acta Cryst. D59, 1903–1913.
Garman, E. F. & Schneider, T. R. (1997). J. Appl. Cryst. 30, 211–237.
Guan, Z., Hederstedt, L., Li, J. & Su, X.-D. (2001). Acta Cryst. D57,
Haebel, P. W., Wichman, S., Goldstone, D. & Metcalf, P. (2001). J.
Struct. Biol. 136, 162–166.
Han, B. W., Lee, J. Y., Yang, J. K., Lee, B. I. & Suh, S. W. (2001). Acta
Harp, J. M., Hanson, B. L., Timm, D. E. & Bunick, G. J. (1999). Acta
Cryst. D55, 1329–1334.
Harp, J. M., Timm, D. E. & Bunick, G. J. (1998). Acta Cryst. D54, 622–
Heras, B., Edeling, M. A., Byriel, K. A., Jones, A., Raina, S. & Martin,
J. L. (2003). Structure, 11, 139–145.
Holyoak, T., Fenn, T. D., Wilson, M. A., Moulin, A. G., Ringe, D. &
Petsko, G. A. (2003). Acta Cryst. D59, 2356–2358.
Hope, H. (1988). Acta Cryst. B44, 22–26.
Hope, H. (1990). Annu. Rev. Biophys. Biophys. Chem. 19, 107–126.
Hunsicker-Wang, L. M., Pacoma, R. L., Chen, Y., Fee, J. A. & Stout,
C. D. (2005). Acta Cryst. D61, 340–343.
Huxley, H. E. & Kendrew, J. C. (1953). Acta Cryst. 6, 76–80.
Izard, T., Sarfaty, S., Westphal, A., de Kok, A. & Hol, W. G. (1997).
Protein Sci. 6, 913–915.
Jacobson, E. M., Li, P., Rosenfeld, M. G. & Aggarwal, A. K. (1996).
Proteins, 24, 263–265.
Juers, D. H. & Matthews, B. W. (2001). J. Mol. Biol. 311, 851–862.
Acta Cryst. (2005). D61, 1173–1180 Heras & Martin
Post-crystallization treatments 1179
Kiefersauer, R., Than, M. E., Dobbek, H., Gremer, L., Melero, M.,
Strobl, S., Dias, J. M., Soulimane, T. & Huber, R. (2000). J. Appl.
Cryst. 33, 1223–1230.
Kim, H. W., Han, B. W., Yoon, H. J., Yang, J. K., Lee, B. I., Lee, H. H.,
Ahn, H. J. & Suh, S. W. (2002). Acta Cryst. D58, 1874–1875.
Kriminski, S., Caylor, C. L., Nonato, M. C., Finkelstein, K. D. &
Thorne, R. E. (2002). Acta Cryst. D58, 459–471.
Kuo, A., Bowler, M. W., Zimmer, J., Antcliff, J. F. & Doyle, D. A.
(2003). J. Struct. Biol. 141, 97–102.
Longenecker, K. L., Garrard, S. M., Shefﬁeld, P. J. & Derewenda, Z. S.
(2001). Acta Cryst. D57, 679–688.
Lusty, C. L. (1999). J. Appl. Cryst. 32, 106–112.
Madhusudan, Kodandapani, R. & Vijayan, M. (1993). Acta Cryst.
Manjasetty, B. A., Croteau, N., Powlowski, J. & Vrielink, A. (2001).
Acta Cryst. D57, 582–585.
Mi, Y., Wood, G. & Thoma, L. (2004). AAPS J. 6, 22.
Monsan, P., Puzo, G. & Mazarguil, H. (1975). Biochimie, 57, 1281–
Mueller, D. M., Puri, N., Kabaleeswaran, V., Terry, C., Leslie, A. G. &
Walker, J. E. (2004). Acta Cryst. D60, 1441–1444.
Pang, S. S., Guddat, L. W. & Duggleby, R. G. (2002). Acta Cryst. D58,
Perutz, M. F. (1946). Trans. Faraday Soc. B, 42, 187–195.
Petock, J. M., Wang, Y. F., DuBois, G. C., Harrison, R. W. & Weber,
I. T. (2001). Acta Cryst. D57, 763–765.
Quiocho, F. A. & Richards, F. M. (1964). Proc. Natl Acad. Sci. USA,
nyi, A., Pohl, E., Scholer, H. R. & Wilmanns, M. (2001). Acta
Cryst. D57, 1634–1638.
Rodgers, D. W. (1994). Structure, 2, 1135–1140.
Rodgers, D. W. (1997). Methods Enzymol. 276, 183–203.
Rould, M. A., Perona, J. J. & Steitz, T. A. (1991). Nature (London),
Salunke, D. M., Veerapandian, B., Kondandapani, R. & Vijayan, M.
(1985). Acta Cryst. B41, 431–436.
Samygina, V. R., Antonyuk, S. V., Lamzin, V. S. & Popov, A. N. (2000).
Acta Cryst. D56, 595–603.
Sansen, S., Verboven, C., De Ranter, C. J., Gebruers, K., Brijs, K.,
Courtin, C. M., Delcour, J. A. & Rabijns, A. (2003). Acta Cryst.
Schick, B. & Jurnak, F. (1994). Acta Cryst. D50, 563–568.
Sousa, R. (1995). Acta Cryst. D51, 271–277.
Takeda, K., Miyatake, H., Yokota, N., Matsuyama, S. I., Tokuda, H. &
Miki, K. (2003). Acta Cryst. D59, 1224–1226.
Timasheff, S. N. (1995). Methods Mol. Biol. 40, 253–269.
Timasheff, S. N. & Arakawa, T. (1988). J. Cryst. Growth, 90, 39–
Tong, L., Qian, C., Davidson, W., Massariol, M. J., Bonneau, P. R.,
Cordingley, M. G. & Lagace, L. (1997). Acta Cryst. D53, 682–
, J., Marek, J., Zouhar, J., Brzobohaty, B. & Su, X.-D. (2001).
Acta Cryst. D57, 140–142.
Yang, J. K., Yoon, H. J., Ahn, H. J., Lee, B. I., Cho, S. H., Waldo, G. S.,
Park, M. S. & Suh, S. W. (2002). Acta Cryst. D58, 303–305.
Yeh, J. I. & Hol, W. G. (1998). Acta Cryst. D54, 479–480.
1180 Heras & Martin
Post-crystallization treatments Acta Cryst. (2005). D61, 1173–1180