<<

. 26
( 61 .)



>>

shatter if a small force 1sapplied. A destructive, but quick test 1stherefore to
poke the crystal with a hair or finely drawn glass fiber. Protein crystals ˜111
readily shatter with such a force, whereas salt crystals will remam intact or
make a discernible crunching noise. Another quick, but destructive test is to
stain with filtered Coomassie dye solution. Protein crystals should take up the
dye and become much darker than the background.
Although crystals may appear to be of good-quality externally by optical
examination, they may not necessarily have satisfactory crystallme order, To
test the quality of their Internal order, a diffraction pattern must be measured.
Most crystals are very sensitive to changes in humidity. The crystals therefore
Lu˜si, Anderson, and Hope
766
need to be maintained under the same humidity as that under which they were
grown. This can be achieved by mountmg the crystal in a capillary with reservon-
solution at either end and sealing tt to maintain the crystal in an atmosphere with
the appropriate humidity. This was the great technique that permitted Bernal and
Hodgkins to obtain the first protein diffraction pattern 60 yr ago.
Flash-freezing has become a recent and very popular method to preserve
crystals for X-ray data collection (24,2.5) The crystals are first briefly trans-
ferred to a cryoprotectant solution (usually a mixture of reservoir solution diluted
with antifreezing solute, such as glycerol) and then quick-frozen m a nitrogen
stream at 100 K. The low temperature effectively immortahzes the crystals
against the damaging effects of X-rays
3.9. Preparation of Derivatives
In order to solve the crystal structure, it is necessary to obtain mformation
about the phase of each reflection that arises from diffraction This can be
achieved by doping the crystal with a heavy atom, which will change the mten-
stty distribution of the diffracted X-rays. Details of the optimal conditions for
soaking heavy metals into protein crystals are described elsewhere (12) Extreme
caution is required m handling these deadly compounds, since they bmd pro-
teins avidly, mcludmg those that comprtse the protein crystallographer. Like
growing crystals themselves, these soaking methods suffer from a caprrcrous
element. Fortunately, procedures have been designed to rationalize the prepa-
ration of heavy atom derivatives (16) Using site-directed mutagenesis, one
can mtroduce cysteme residues as target sites for modification by mercurial
compounds. The best strategy 1sto choose ammo acids that are hydrophrhc,
since these will most likely be on the surface of the protein and therefore more
likely to be exposed for reaction.
3.10. Methods for Growing and Evaluating Two-Dimensiona/
Crystals for Electron Microscopy
Two-dimensional crystallography offers a convenient method for obtaining
structural mformation. Although the resolution of the method is limited, there
are a number of advantages for using this approach. Only microgram quantities
of protein are required at concentrations as little as 250 pg/mL, and the results
can be known within hours rather than weeks, since the crystals can be imme-
diately visualized with the electron microscope.
3.707. Crystalhzation of Protein on Lip/d Layers
Crystalhzation depends on the binding of the macromolecules to a charged
water-exposed surface of lipid layers (17) Association with this surface ori-
ents and concentrates the macromolecule in two dimensions, whereas lateral
167
Crystallization of Macromolecules

diffusion facilitates crystallization. A number of macromolecules have been
successfully crystalhzed using thts method, including yeast RNA polymerase II
(5,. Crystals were formed from polymerase II at a concentration of Xl-250 l.tg/mL.
The droplet of protein was coated with a lipid mixture containmg 0.05 mg/mL of
octadecylamine and 0.45 mg/mL of egg L-a-aphosphattdylcholine m 1: 1 chlo-
roform/hexane. The drops were then incubated under nitrogen or argon for 30-
60 mm; during this time, crystals formed on the droplet surface. The crystals
were then transferred to a carbon-coated grid, using a wire loop. This loop was
placed over the drop, and a thin layer of protein lipid was carefully lifted and
passed onto the electron microscope grid. The grid was allowed to dry and then
washed with 2 mL of distilled water. After washing, the sample was stamed
using uranyl acetate and examined by electron microscopy.
Recently, the lipid surface method has bean made more general through the
development of nickel-chelating lipids for histidme-tagged proteins (18).
3.10.2. Crystallization of Protein on Mica Surfaces
Another method of crystalhzmg proteins in two dimensions is to use freshly
cleaved mica. A few microliters of protein solution are spread over the top of
the mica and allowed to dry. The protein-coated mica 1sthen covered m a very
tine layer of carbon, and using a razor blade, the surface of the protem mica is
scored very gently in a crisscross fashion. The sample is then placed m a shal-
low dish containing distilled water, and the coated specimen floats from the
surface of the mica. The electron microscope grid is then placed under the
protein sample and lifted from the water. The sample is then dried and stained
with uranyl acetate or other suitable negative stain.
4. Notes
I. Preparatton of protein-requirement for purity and quantity. Perhaps the most
important factors affecting the success of crystalhzatton trials are the purity and
abundance of the sample. Because there are no rules dictating the conditions
under which a macromolecule will crystallize, many different conditions must be
examined before a successful result might be obtained. The more material that is
avarlable, the easier the task of surveying a wide range of condmons Even tf
crystals are readily found in an initial screen, it may be necessary to consume
more materral m refining conditions to optrmtze crystal quality and m the search
for suitable heavy-atom derivatives Experience has demonstrated repeatedly that
specimens that are free of substantial heterogeneity, which include small-mol-
ecule contaminants, are most likely to crystallize. Heterogeneity IS likely to ruin
the umform interactron of proteins, which is required for generating a crystallme
lattice. Protein samples must be free of traces of protease, which can degrade the
materral into a useless heterogeneous mrxture with time. Elsewhere m this book,
chapters are dedicated to detailed methods of purtficatlon of proteins (e g ,
Luisi, Anderson, and Hope

ammomum sulfate prectprtation, ton-exchange, and hydrophobic mteractton
chromatography), and these should be referred to for advtce on designing purifi-
cation protocols (see also ref. 19) We descrtbe here some procedures for analy-
sts of the purtty of the specimen.
The first goal of the hopeful crystallographer IS to ensure that the specimen for
crystalhzatton is as pure as possible by analysts with denaturmg gel electrophoresrs
(sodmm dodecyl sulfate-polyacrylamide gel electrophorests, or SDS-PAGE),
which is the easiest and most common method of evaluatmg the sample for con-
taminating protems Differences m molecular wetght of about 1000 Dalton can be
readily detected by SDS-PAGE Coomassie blue staining of SDS-PAGE gels can
detect about 0.5 pg of protein, whereas silver-stammg increases the level of detec-
tion about lo-fold. The response of the silver stain IS nonlmear with protem quan-
tity, and rt may be difficult to estimate the relative abundance of contammants
A homogeneous band on SDS-PAGE may, however, not be a suffictently rigor-
ous test of purity For mstance, nonprotemactous substanceswill not be detected, and
highly baste protems may sometimes carry nucleic actd contaminants. UV spec-
trophotometry IS a useful method to check protein samples for nonprotem-
acious contammants. Proteins have characteristtc adsorption maxtma at 280 nm
owmg to the presence of aromatic rmgs m tryptophan, phenyalanme, and tyrosme,
whereas nucleic acids absorb maxtmally at 260 nm Contammatmg nucletc actds
will affect crystal growth of protems, but these can be readrly removed by
hydroxyapatite chromatography.
Microheterogenetty wtthm a protein may also go undetected by SDS-PAGE
This heterogeneity may arrse from protein modifications, e.g., glycosylatton or
phosphorylation, which are posttranslational processes, or from effects, such as
oxrdatton and deammatton Microheterogeneity may have a strong effect on crys-
talhzatrons or no effect whatsoever It depends on the case, and every case IS
different. However, the safest strategy IS to ensure that the protein suffers from as
little microheterogeneity as possible
A number of sensitive procedures are available to detect mtcroheterogeneity.
One method is two-dimenstonal PAGE, where the fn-st dtmension separates
according to the isoelectrtc point of the protein and the second separates by stze,
as m regular SDS-PAGE. Microheterogeneity owing to charge differences can
also be detected conveniently by isoelectric focusing (IEF) gels Any
microheterogenetty detected by IEF under nondenaturmg conditions may be owing
to differences m primary structure or in the type and number of prosthettc groups. For
thus reason, it is helpful to examine the protein sample by denaturing IEF.
Electrospray mass spectrometry 1s an extremely senstttve method that can
reveal specimen heterogeneity and gave a clue regarding Its nature (20) This
method provides accurate mass determmatton to less than the mass of a single
ammo acid residue of proteins up to about 100 kDa. Since the protein sequence 1s
usually known, the expected mass can be very accurately calculated. Mass devia-
tions can usually be accounted for by proteolysis or covalent modificatron A
related technique IS pepttde fingerprmtmg by mass spectrometry Here, the pro-
Crystallization of Macromolecules 169
tein could be treated with sequencmg grade protease to generate a characteristtc
pattern of fragments m the mass spectrometer. This fingerprint aids in corrobo-
rating the identity of the protein and defining possible sites of modtficatton
Capillary electrophoresis has proven to be a useful method for detecting mtcro-
heterogeneity. A fused sihca capillary is filled with electrolyte support buffer
and placed between two buffer reservoirs containing high-voltage electrodes
Protein is introduced at one end of the capillary and mtgrates under the influence
of the electrrc field. Because the capillary can be made to very long dunenstons,
extremely high-resolution separatton can be attamed, capable of resolving mol-
ecules differing by single charges
Deammatron of glutamme and asparigme residues may occur at low pH and
lead to charge change This can be prevented by avoiding acidic condtttons at
all stages m protein preparation. Another source of microheterogeneity 1s OXI-
dation of cysteme and methtonme residues. Oxtdatton can be avoided If the
specimen is carefully kept under reducing condtttons. One should include gentle
reducing agents m all buffers, such as dithtothreitol, dithtoerythrettol, or
mercaptoethanol Degassmg buffers and flushing with nitrogen may also pro-
long the reducing environment
Microheterogenetty in recombinant protetns can arise from ammo actd
mtsincorporatton When foreign genes are expressed m Escherzchza ˜011,
mtsfoldmg of the expressed protein and mtsmcorporation of ammo acids may
occur tf the codons used by the foreign genes are relatively rare m the bacteria
(reviewed m ref 21). One possible way to overcome the problem with ammo
acid mtsincorporatton in recombinant proteins would be to synthesize the gene
chemically using codons optimal for expression m E cob
Glycosylatton inherently produces high levels of mtcroheterogenetty owing to
the variation in branching of the carbohydrates. There are also several examples
where removal of part or the entire carbohydrate may have assisted crystal growth
(22˜23) It should be pointed out, however, that there are other examples where
crystals have been grown from proteins that have a mixture of different carbohy-
drate moieties. Indeed the presence of carbohydrate may be essential for stability
and solubility of some proteins, so complete removal of the carbohydrate may not
be helpful in preparing specimens for crystallization m certain cases. Potential prob-
lems caused by glycosylation have to be mvestigated for each case
If the carbohydrate is suspected as bemg a possible culprit retarding crystal
growth, then there are several ways of removing the offending party. For instance,
the siahc acid residues of the carbohydrates bear charged carboxylates and are a
source of charge heterogeneity. These sugars can be removed using neurammi-
dase This treatment should result m the protein being less heterogeneous on IEF
gels. More drastically, hydrogen fluoride treatment will remove the carbohydrate
moiety entirely (24) This procedure 1s extremely dangerous and requires special
safety apparatus. It 1scritical that the protein be lyophtltzed to complete dryness
for this procedure. As a safer alternative, commercially available cocktails of
endoglycosidases may also be used to remove carbohydrate. Smce the cocktail
Luisr, Anderson, and Hope
170
may not be very pure, It may be necessary to m&de protease inhibitors to pre-
vent degradation of the sample Site-dtrected mutagenests can also be used to
alter known glycosylation sites For Instance, Bentley et al. (23) mutated aspar-
agines to glutammes m the P-chain of the T-cell antigen receptor to stop ammo-
linked glycosylatlon The aglycosylated material yielded crystals, whereas the
native glycosylated maternal failed to do so
DNA and RNA purtty™ Purity of nucleic acids may be readily tested by denatur-
ing gel electrophoresis. Mass spectrometry is also becoming a very useful method
for evaluating the nucleic acid specimen. A convenient matrix has been recently
developed that is used in matrix-assisted laser desorption iomzation time-of-flight
mass spectrometry (MALDI-TOF MS) (25)
Useful procedures to increase the chances of growing crystals. If the recombi-
nant proteins are msoluble, one can purtfy the material under denaturmg condi-
tions and refold the specrmen (21) Some proteins aggregate extenstvely under
low salt condmons. In this case, one can use htgh salt, a volatile salt, such as
ammomum bicarbonate, or ammomum acetate to improve the solubihty (26) It
is also a good strategy to try homologs of the protein from related species This
worked very well for structural studies of the RING finger domain from equine
herpesvirus (27) Here, the correspondmg segment of the protein from the viral
homologs (human herpes simplex virus) yielded only very poorly behaved pro-
teins. Mutations that change surface residues sometimes provtde another useful
variation that has successfully resulted in crystallization, which failed for the
native protein (28,29)
We can suggest that if the protein of interest is poorly behaved, one might trim
it down to stable structural cores for analysts by NMR or to be used for crystalh-
zatlon trrals. Detimtion of structural subdomams by proteolysis has been quite
successful m a number of cases for structural studies, and it is parttcularly useful
for preparing NMR specimens. Use sequencing-grade trypsm or TPCK-treated
chymotrypsm
Cocrystalhzatlon of Fab/protem complexes may facilitate crystal growth. If
good monoclonal antibodies (MAb) are available that recognize the protein under
nondenaturmg condmons (and therefore most likely recognize exposed surface
epitope), they could be used for generation of Fab fragments. The Fab complexes
appear to have been helpful for growing crystals m several cases (3&33)

References
1 Savva, R., McAuley-Hecht, K , Brown, T., and Pearl, L. (1995) The structural
basis of specific base-excision repair by uracil-DNA glycosylase. Nature 373,
487-493
2. Pley, H W , Flaherty, K M , and McKay, D B (1994) Three-dimensional struc-
ture of a hammerhead ribozyme Nature 372, 68-74
3 Scott, W. G., Finch, J. T., and Klug, A. (1995a) The crystal structure of an all-
RNA hammerhead ribozyme. a proposed mechanism for RNA catalytic cleavage
Cell 81,991-1002
Crystallization of Macromolecules 171
4. Joachimiak, A. and Sigler, P B. (1991) Crystallization of protem-DNA com-
plexes Methods Enzymol 208, 82-99
5 Darst, S. A , Kubalek, E. W , Edwards, A. M , and Kornberg, R D. (199 1) Two-
dimensional and epuaxial crystallization of a mutant form of yeast RNA poly-
merase II J Mol Biol. 221, 347-357.
6. Ducnux, A and Glege, R. (1992) Crystalltzatton of Nucleic Actds and Protews
A Practical Approach IRL, Oxford University, Oxford, UK.
7. Carter, C W., Jr and Carter, C W (1979) Protem crystallization using mcom-
plete factorial expertments J Bzol Chem 254, 12,219-12,223
8 Abergel, C , Moulard, M , Moreau, H., Loret, E., Cambillau, C., and Fontecllla-
Camps, J. C. (1991) Systematic use of the incomplete factorial approach m the
design of protein crystallization experiments J. Btol Chem 266, 20,13 l-20,138
9. Jancarik, J. and Kim, S -H. (1991) Sparse matrix samplmg. a screening method
for crystallization of proteins J Appl Crystallogr 21, 67-7 1.
10 Scott, W. G , Finch, J T , Grenfell, R., Fogg, J., Smith, T , Gait, M J , and Klug,
A. (1995) Rapid crystalhzation of chemically synthesized hammerhead RNAs
using a double screening procedure J Mol Bzol 250,327-332.
11. Doudna, J A., Grosshans, C , Goodmg, A , and Kundrot, C. E. (1993) Crystalh-
zation of rrbozymes and small RNA motifs by a sparse matrix approach Proc
Nat1 Acad Sci USA 90,7829-7833
12 Blundell, T L and Johnson, L N. (1976) Protean Crystallography Academic,
London.
13 Hegde, R S , Grossman, S R., Lalmms, L. A., and Sigler, P B (1992) Crystal
structure at 1.7 A of the bovie papillomavu-us-1 E2 DNA-binding domain bound
to its DNA target. Nature 359, 505-5 12
14. Gamblm, S. J. and Rogers, D. W (1993) Some practical details of data collection
at 100 K, m Data Collectton and Processtng Proceedings of the CCP4 Study
Weekend (Sawyer, L , Isaacs, N , and Bailey, S., eds.), SERC Daresbury Labora-
tory, Warrington, UK
15. Hope, H. (1988) Cryocrystallography of biological macromolecules: a generally
apphcable method Acta Cryst B44, 22-26
16 Nagai, K., Oubridge, C., Jessen, T.-H , Li, J., and Evans, P R. (1990) Crystal
structure of the RNA-binding domain of the Ul small ribonucleoprotein A Nature
348,5 15-520.
17 Edwards, A. M , Darst, S A, Hemming, S. A, Li, Y , and Kornberg, R. D.
(1994) Epitaxial growth of protein crystals on lipid layers. Nature Struct Bzol
l(3), 195-197
18. Kubalek, E W , LeGrice, S F J , and Brown, P. 0. (1994) Two-dimensional
crystallization of histidme-tagged, HIV- 1 reverse transcriptase promoted by a

<<

. 26
( 61 .)



>>