An introduction to the

crystallogenesis of biological
unacrounolecules
R. GIEGE and A. DUCRUIX

'II Ya 13. des mysteres, qui pn!parent a l'avenir d'immenses travaux ct appellent des
aujourd'hui les plus serieuses meditations de la science'
Pasteur, 1860, in Le~ons de Chimie.

1. Introduction
The word 'crystal ' is derived from the Greek root ' krustallos' meaning 'clear
ice'. Like ice, crystals are chemically well defined, and many among of them
are of transparent and glittering appearance, like quartz, which was for a long
time the archetype. Often they are beautiful geometrical solids with regular
faces and sharp edges, which probably explains why crystallinity, even in the
figurative meaning, is taken as a symbol of perfection and purity. From the
physical point of view, crystals are regular three-dimensional arrays of atoms,
ions, molecules, or molecular assemblies. Ideal crystals can be imagined as
infinite and perfect arrays in which the building blocks (the asymmetric units)
are arranged according to well-defined symmetries (forming the 230 space
groups) into unit cells that are repeated in the three-dimensions by trans-
lalions. Experimental crystals, however, have finite dimensions. An implicil
Consequence is that a macroscopic fragment from a crystal is still a crystal,
because the orderly arrangement of molecules within such a fragment still
extends at long distances. The practical consequence is that crystal fragments
can be used as seeds (Chapter 7). In laboratory-grown crystals the periodicity
IS never perfect, due to different kinds of local disorders or long-range
Imperfections like dislocations. Also, these crystals are often of polycrystal-
hne nature. The external forms of crystals are always manifestations of their
Internal structures and symmetries, even if in some cases these symmetries
~ay be hidden at the macroscopic level, due to differential growth kinetics of
t e Crystal faces. Periodicity in crystal architecture is also reflected in their
macroscopic physical properties. The most straightforward example is given
 1: An introduction to the crystallogenesis
R. Giege and A. Ducruix

by the ability of crystals to diffract X-rays, neutrons, or electrons, the

phenomenon underlying structural chemistry and biology (for introdu~tory
texts see refs 1 and 2), and the major aim of this book is to present the
methods employed to produce three-dimensional crystals of biological macro-
molecules, but also two-dimensional crystals (Chapter 12), needed for
diffraction studies. Other properties of invaluable practical applications
should not be overlooked either, as is the case of optical and electronic
properties which are at the basis of non-linear optics and modern electronics
(for an introduction to physical properties of molecular crystals see ref. 3).
Crystals furnish one of the most beautiful examples of order and symmetry in
nature and it is not surprising that their study fascinates scientists (4).
What characterizes biological macromolecular crystals from small molecule
crystals? In terms of morphology, one finds with macromolecular crystals the
same diversity as for small molecule crystals (Figure 1). In terms of crystal
size, however, macromolecular crystals are rather smail, with volumes rarely
exceeding 10 mm" and thus they have to be examined under a binocular
microscope. Except for special usages, such as neutron diffraction , this is not
too severe a limitation. Among the most striking differences between the two
families of crystals are the poor mechanical properties and the high content of
solvent of macromolecular crystals. These crystals are always extremely
fragile and are sensitive to external conditions. T his property can be used as a
preliminary identification test: protein crystals are brittle or will crush when
touched with the tip of a needle, while salt crystals that can sometimes
develop in macromolecule crystallization experiments will resist this treat-
ment. This fragility is a consequence both of the weak interactions between
macromolecules within crystal lattices and of the high solvent content (from
20% to more than 80%) in these crystals (Chapter 14). For that reason,
macromolecular crystals have to be kept in a solvent-saturated environment, Figure
tate of 1.
HEW From 1 I at es t 0 pe rfeet crystals of biological macromolecules. (a) Precipi-
I prec'p't
otherwise dehydration will lead to crystal cracking and destruction. The high of the com le~s~zyme; (b) yeast aspartyl-tRNA synthetase microcrystals; (3) spherulites
solvent content, however, has useful consequences because solvent channels tiSSue inh .,bPt etw~en yeast aspartyl-tRNA synthetase and tRNAAsp; (d) short needles of
permit diffusion of small molecules, a property used for the preparation of Crystals' If II ye
or protetn
t . . . of met a II opro t ease; (e) aspartyl-tRNA synthetase long needle-like
M et thO I .
isomorphous heavy-atom derivatives needed to solve the structures (Chapters crystals , of Hypasd InitIator
' tRNA tn pates WIth growth defects; (g) plate-shaped
13 and 14). Further, crystal structures can be considered as native structures, aynthetase sho ~ erma Itneatum collagenase; (h) tetragonal crystals of aspartyl -tRNA
of 'akin' of den~tng ~rowth defects together with brush-like needle bunches' (i) example
as is indeed directly verified in some cases by the occurrence of enzymatic re-
hollow extre~~~~ (~r~tein. around a HEW lysozyme crystal; (j) crystal of ~ellitin with
actions within crystallatlices upon diffusion of the appropriate ligands (5, 6). HEw lysozyme " (rn ' , twtnned and twinned-embedded crystals of collagenase and
Other characteristic properties of macromolecular crystals are their rather rne crystalli z~tio~ n, 0, p)hPe~ect three-dimensional crystals; (m) polymorphism in a
d rop sowing c b·
weak optical birefringence under polarized light: colours may be intense for NA syntheta se/tR NAAsp com "u IC an d ort h orhomblc . crystal-habits of aspartyl-
large crystals but less bright than for salt crystals (isotropic cubic crystals or crystals of HEW I plex, (n) crystals of tRNAAsp with cracks due to ageing"
ysozyme with fourfold symmetry; (p) crystals of H. lineatu';
amorphous material will not be birefringent). Also, because the building blockS
composing macromolecules are enantiomers (L-amino acids in protein,s-
except in the case of some natural peptides-and D-sugars in nucleic aCids)
macromolecules will not crystallize in space groups with inversion syfll~
me tries. Accordingly, out of the 230 possible space groups, macromolecul~: l
do only crystallize in the 65 space groups without such inversions (7). Wh "
3
 1: An introduction to the erystallagenesis
R. Giege and A. Ducruix
that for technical difficulties, the X-ray structure of urease has only be deter-
small organic molecules prefer to crystallize in space groups in which it is mmed very recently (20). Besides being a method of purification, crystalliza-
easiest to fill space, proteins crystallize primarily in space groups in which it is tIon expenments established the view that pure enzymes are proteins, a fact
easiest to achieve connectivity (8). Macromolecular crystals are also charac- not obvIOUS to all at that time (21). Another scientific achievement arose in
terized by large unit cells with dimensions that can reach up to 1000 A for 1935 when Stanley crystallized tobacco mosaic virus. Influenced by Northrop's
virus crystals (9). From a practical point of view, it is important to remember conceptIon of enzymes, and usmg methods developed for proteins, he prepared
that crystal morphology is not synonymous with crystal quality. Therefore, the the VIruS 10 pure crystallme state. (he believed the virus was an autocatalytic
final diagnostic of the suitability of a crystal for structural studies will always protem) and showed that II retams Its mfectivity after several recrystalliza-
be the quality of the diffraction pattern which reveals its internal order, as is tIons (reVIewed m ref. 10). The importance and the implications for biology of
reflected at first glance by the so-called 'resolution' parameter (Chapter 14). these dlscovenes was recogmzed rapIdly and in 1946 the Nobel Prize for
Crystal growth, which is a very old activity that has always intrigued Chemistry was awarded to Sumner, Northrop, and Stanley.
mankind, and many philosophers and scientists have compared it with the
biological process of reproduction, and it has even been speculated that the
duplication of genetic material would occur through crystallization-like 2.2 Crystallogenesis and structural biology
mechanisms (10). Nowadays, the theoretical and practical frames of crystallo- The use of crystals in structural biology goes back to 1934, when Bernal and
genesis are well established for small molecules, but less advanced for Crowfoot (D. Hodgkin) produced the first X-ray diffraction pattern of a
macromolecules, although it can be anticipated that many principles under- protem, that of crystalline pepsin (22). Since then representatives of most
lying the growth of small molecule crystals will apply for that of macro- famlltes of macromolecules have been crystallized, but using mainly empirical
molecules (11, 12). Until recently, crystallization of macromolecules was methods and WIthout ratIOnal control of the growth mechanisms. This can be
rather empirical, and because of its unpredictability and frequent irrepro- well understood because in the early days of structural biology the interest of
ducibility, it has long been considered as an 'art' rather than a science. It is SCIentIsts was mainly directed at the development of the X-ray methods
only in the last 15 years that a real need has emerged to better understand and needed to resolve the structures rather than in that of a rationalization of the
to rationalize the crystallization of biological macromolecules. It can be stated ~r:stallization procedures. Therefore , only limited efforts were expended in
at present that the small molecule and macromolecular fields are converging, derstandmg or Improvmg macromolecular crystallization procedures. At
with an increasing number of behaviours or features known for small pr:sent X-ray methods are well established (2,7,23). The overall scheme of
molecules that are now found for macromolecules (13). th~ ~arlO~s steps of the resolution of a three-dimensional structure is sum-
m,lrt7.ed m Figure 2. But production of suitable crystals diffracting at high
resolu
W' h tton often
. . t h e b ott1eneck 10
remams . structure determination projects
2. Crystallization and biology: a historical background m It the rapId development of biotechnologies and the unlimited potential of
acromolecular e
directed '
ngmeermg. (h'
w Ich requires structural knowledge for site-
2.1 Before the X-rays ther' '. mutagenesIs expenments or for design of factitious macromolecules)
It is often forgotten that some advances in biochemistry and molecular C IS now an mcrea slOg
resolutio . nee d f or macromolecular crystals diffracting at high
biology have their origin in crystallization data and that empirical crystal assemhl" n,,. not
Th only proteins b u tal so nuc 1" .
eiC aCids and multI-macromolecular
growth of biological materials is as old as biochemistry. The first reports on les. Ih
rather easit anks to the b'10 t echno i ' 1 tools ..
oglca .
It IS now pOSSIble to obtain
protein crystals were published more than a century ago when Funke, IIlilhgrams)Y, e admounts of pure macromolecules (in most cases several
Hiinefeld , Lehman , Teichman, and others crystallized. haemoglobin
. from the he troductory. I ee d e
ch to start a crysta II"IzatlOn project . (Chapter 3). In this
blood of various invertebrates and vertebrates, and It was propheSIed that t h in biolo aPtMer we present the general trends of the science of crystal
study of crystals would shed light on the exact nature of proteinic substanC~~ nted tn foilgy. . ore extensl ve d"ISCUSSlOns and the practICal . details will be
(14-17). This was followed by the crystallIzatIOn of hen egg whIte albumtn a I . owmg chapters.
a series of plant proteins (reviewed in refs 16 and 17). The beauty of crysta s fi rst tnaJor breakth
. h rough towards better and easier crystallizations was
certainly fascinated the physiological chemists in these early days, since a~ d'fft e ' 1960s' of crys t a 11'Izatton
, In .
cryst~~
micromethods (e.g. dialysis
atlas with extensive descriptions of the morphologies of haemoglobin ma'Crf>ln()le"I'du" SolOn) (Chapter 5). It was promoted by structural pro-
was published in 1909 (15). In 1926 Sumner reported the crystallIzatIOn d
. N h ystal ltze (24). Further . reluctant to crystallize easily and available in limited
slgmficant im provements came from the discovery of
urease from Jack beans (18), soon followed by orthrop w 0 cr te
pepsin and a series of other proteoJytic enzymes (19). It is interesting to no
5
 1: An introduction to the crystallogenesis
R. Giege and A. Ducruix
the development of genetic engineering technologies for overexpression of
1- PURIFICATION OF MACROMOLECULES protems (Chapter 3), explains the increasing number of crystallized nucleic
from wild·type. engineered or overproducing organisms acids or rare proteins.
(possibility of in vitro synthesis for nucleic acids and small peptides) . Today: crystal growth research is stimulated by macromolecular engineer-
mg reqUIrements, but also to some extent by space-science projects (crystal-
2- CRYSTALLIZATION lization under microgravity conditions, Chapter 6) and its descriptive stage is
by de novo crystallization or seeding techniques moving towards a new: more. quantitative discipline, biocrystallogenesis,
whIch mcludes bIology, bIOchemIstry, physics, and engineering related aspects.
For speCIalIzed lIterature see refs 35-42, and for general reviews 43-46.
3- DATA MEASUREMENTS
characterization of space group and diffraction resolution;
measurements of diffraction intensities on an electronic area detector
(possible use of neutron and frequenUy of tunable X-ray synchrotron radiation); 3. General principles
frequent data acquisition by cry<KrystaJlographic methods. 3.1 A multiparametric process
Biocrystallization, like any crystallization, is a multiparametric process involv-
ing the three classical steps of nucleation, growth, and cessation of growth.
4- PHASE DETERMINATION
using methods based on isomorphous replacement (preparation of heavy atom derivatives),
What makes crystal growth of biological macromolecules different is, first, the
anomalous scattering. molecular replacement and non-crystallographic symmetry. much larger number of parameters than those involved in small molecule
or direct calculations (e.g. from maximum entropy) crystal growth (Table 1) and, secondly, the peculiar physico-chemical pro-
perlIes of the compounds. For instance, their optimal stability in aqueous
media is restricted to a rather narrow temperature and pH range. But the
5- ELECTRON DENSITY MAP COMPUTATION AND INTERPRETATION malO .dlfference from small molecule crystal growth is the conformational
interpretation of ntini-maps; model building on computer graphic displays fle~l bIlIlY and chemical versatility of macromolecules, and their consequent
greater sensltlvIty to external conditions. This complexity is the main reason
~h y systematic investigations were not un4ertaken earlier. Furthermore, the
6- MODEL REFINEMENT IInportance of some parameters, such as the geometry of crystallization
least-square refinements; restrained refinements; . vessels or the biological origin of macromolecules, had not been recognized. It
IS only recently that the hierarchy of parameters has been perceived. A prac-
Figure 2. Steps involved in the resolution of the 30 structure of a biological macro-
tIcal conseque nce 0 f thOIS new perceptIOn . was the development of statistical
molecule (for more details see Chapter 14).
methods
f to screen crys·ta II"IzatlOn con d'ItlOns
. (Chapter 3). For a rational design
o growth
contr II d condition s, h owever, p h YSlcal
. ..
and bIologIcal parameters have to be
specific properties of additives to be included in crystallization solvents, such bioi 0. e . One of the alms of thIS book is to give to crystal growers of
as the polyamines (25, 26) and the non-ionic detergents (27-30) whIch g~ve oglcal macromol eeu Ies t h e conceptual and methodological tools needed
to ach
eJOs
the clue fo r crystallizing nucleic acids (Chapter 8) or membrane prot ICvC such control.
(Chapter 9). Also, a more systematic use of organic cosmotropes (compounds
promoting 'order'), like polyhydric alcohols that are stabilizers of proteIDe 3.2 Purity
structure when at high concentration, may facilitate crystallization of fleJO bl Because macromolecul
fication pia san es are ~xtracted from complex biological mixtures,
proteins (31). . ' 32) !y. howey: . extremely Important role in crystallogenesis (Chapter 2).
T he perception of the importance of punty for growmg better crystals (
It!~ can sr, IS not' an abs0 I
uet '
reqUIrement .
SInce crystals of macro-
was an important achievement in the field (Chapter 2), and the adequate
mall 0 omellmes be obt alOe ' d from mIxtures.
.
choice of the biological material was an important determinant for the succe~
But such crystals are
.'1
... ...·.iifI~.."i,,::;r qgrow
r as poly crys tIl'
of many crystallizations. With the ribosome for instance, the preparatI~~o~
a me masses, are not weJl shaped and are of
~.:rystalli,,",'t;uc>na ,ty , and thus cannot be used for diffraction st~dies. How-
homogeneous particles from halophIlIc or thermophIlIc bactena mstead h< f
mesophilic bacteria, considerably improved crystal quality (33 , 34) . AIS~')~ or
..lIlticalinn (47)0 macromole . cuI f mIxtures
es rom ' may be used as a tool
, espeCIally in industry (48). For the purpose of X-ray
ease of synthesizing oligonucleotides with automated methods (Chapter
7
 R. Giege and A. Ducruix
Table 1. Paramete rs affecting the crystallization (and/or the so lu bil ity) of macro-
molecules a
Intrinsic physico-chemical parameters
• Supe rsaturat ion (concentration of macromolecules and precipitants)
• Temperature, pH (fluctuations ofthese parameters)
• Time (rates of equ ili brati on and of growth)
• Ionic strength and purity of chemicals (nature of precip itant, butter, additives)
• Diffusion and convection (gels, microgravity)
Vo lume and geometry of samples and set-ups (surface of crystallization chambe rs)
Solid particles, wa ll and interface effects (e.g. homogeneous versus heterogeneous
nucleation, ep itaxy)
• Dens ity and viscos ity effects (d ifferences between crystal and mother liquor)
• Pressure, electric and magnetic fie lds
• Vibratio ns and sound (acoustic waves)
• Sequence of events (experimenta list versus robot)
Biochemical and biophysical parameters
• Se nsiti vity of conformations to phys ical pa rameters (e.g. tempe rature, pH, ionic
strengt h, solvents)
Binding of ligands (e.g. substrates, cofactors, metal ions, other ions)
• Specific addit ives (e.g. reducing agents, non-ion ic dete rgents, polyamines) re lated with
properties of macromolecu les (e.g. oxidation, hydrophilicity versus hydrophobicity,
polye lect rolyte nature of nuc leic acids)
Ageing of samp les (redox effects, denatu rati on, o r degradation)
Biological parameters
• Rarity of most biologica l macromolecules
• Biological sources and physiological state of organ isms or ce ll s (e.g. thermophi les
versus halophiles or mesophiles, grow ing ve rsus stationary phase)
Bacterial contam in ants
Purity of macromolecules
Macromolecular contaminants (odd macromolecules or small molecules)
• Sequence (micro) heterogeneities (e.g. fragmentation by proteases or nucleases-
fragmented macromo lecules may better crystal li ze - , partial or heterogeneous post-
translational modifications)
• Conformationa l (micro) heterogeneities (e.g. flexible domains, oligomer and conformer
equilibria, aggregation, denaturation)
• Batch effects (two batches are not identical)
a Although all these parameters have not been screened system~tical'y, especially for the crysta l-
lization of a given macromolecule, all of them have been evaluated Individually in isolated cases.
crystallography, high-quality monocrystals of appreciable size (0.1 mm at least
for the dimension of a face) are needed. It is our belief that poor punty IS the
most common cause of unsuccessful crystallization, and for crystallogeneSlS
the purity requirements of macromolecules have to be higher than in other
fields of molecular biology. Purity has to be of ' crystallography grade': the
Q
1: An introduction to the crystallogenesis
macromolecules not only have to be pure in terms of lack of contaminants,
they have also to be conformationally 'pure' (32). Denatured macro-
molecules, or macromolecules with structural micro heterogeneities, adversely
affect crystal growth more than do unrelated molecules, especially when
structural heterogeneities concern domains involved in crystal packing. On
the other hand, the presence of microquantities of proteases (or nucleases)
can alter the structure of the macromolecules during storage or the rather
long time needed for crystallization. As a consequence, when starting a
crystallization project one has to be primarily concerned with purification
methodologies and to take all precautions against protease and nuclease
action. To have reproducible results, the physiological state of cells should be
controlled, because protease (or nudease) levels may vary as well as the
balance between cellular components. As a general rule, batches of macro-
molecules should not be mixed and crystallization experiments should be
conducted on fresh material so that ageing phenomena are limited . For more
details see Chapter 2.
In summary, we emphasize the importance of macromolecular purity in
biological crystallogenesis, and in cases of unsuccessful experiments we rec-
ommend first improvement or reconsideration of the purification procedure
of the molecules of interest.
3.3 Solubility and supersaturation
To grow crystals of any compound, molecules have to be brought in a super-
saturated, thermodynamically unstable state, which may develop in a crystal-
line or amorphous phase when it returns to equilibrium. Supersaturation can
be achieved by slow evaporation of the solvent or by varying parameters
(hsted in Table 1). The recent use of pressure as a parameter is of note (49).
From this it follows that knowledge of macromolecular solubility is a pre-
reqUIsIte for controlling crystallization conditions. However, the theoretical
background underlying solubility is still controversial , especially regarding
salt effects (50), so that solubility data almost always originate from experi-
mental determinations. Specific quantitative methods permitting such deter-
nllnahons on small protein samples are available (51-53). The main output
was the experimental demonstration of the complexity of solubility behaviours,
~IllPhaSizing the importance of phase diagram determinations for a rational
eSlgn of crystal growth (Chapter 10).
As to the nature of the salt used to reach supersaturation, one can wonder
why a mmomum' . so frequently chosen by crystal growers (54). This
sulfate IS
Usage. IS m fact mCI
. 'd ental and results from the practices of biochemists for
SaI hng- t .
eff . ou protems. Indeed many other salts can be employed, but their
COn'eChveness fo r m. d ucmg
. crystaIIzatlOn
I·· ··
IS vanable
(52.
) The practical
. th
larg, equenc e IS ' supersaturatIOn
at protem . can be reached (or changed) in a
Used.C ConcentrafIon range 0 f ' and salt, proVIded
protem . that adequate salts are
z
1: An introduction to the crystallogenesis
R. Giege and A. Ducruix
crystallization
d df process
. . are not known WI'th certamty,. the theoretical frame
nee e or explammg their role is well established (11- 13 46 56) Al h
3.4 Nucleation, growth, and cessation of growth expenmental tools exist . that
. are needed for measunng . t h'e contnbutlOns
, : so, t of e
Because proteins and nucleic acids require defined pH and ionic strength for t hese parameters. ThIS IS the case for solubilit a .
stability and function , biomacromolecule crystals have to be grown from measurements as well as for monitoring pH d Y nd aggregatIOn state
d 11) an temperature (Chapters 10
chemically rather complex aqueous solutions. Crystallization starts by a an ., or
t' evenf the effect of microgravity (Chapter 6) . Corre IatlOns
' between
nucleation phase (i.e. the formation of the first ordered aggregates) which is th e vana
. IOn 0 a parameter and the ab'I't I I Y 0 f a gIven
. macromol I t
followed by a growth phase. Nucleation conditions are sometimes difficult to crystalhze
. h are expected
.. to be found. Finally, an d per haps the most Important
. ecu e 0
reproduce, and thus seeding procedures with preformed crystalline material was t .e recogmtIOn of the importance of purity. Th ' we emphaSIze
us agam . that'
should not be overlooked as in many cases they represent the only method to expen mentehrs should primarily devote their efforts to starting crystailization
obtain reproducible results (55) (Chapter 7). It should be noticed that nucleation attempts WIt molecules of the highest biochemical qu aI't I y.
requires a greater supersaturation than growth, and that crystallization rates
increase when supersaturation increases. Thus nucleation and growth should
be uncoupled, which is almost never done consciously but occurs sometimes 4.2 Towards an active control of crystal growth
under uncontrolled laboratory conditions. From a practical point of view, inter, Only few ' "attempts have been published d escn'b'mg t he aclIve . control of
face or wall effects as well as shape and volume of drops can affect nucleation crysta
k II IzatIOn
b expenments. In general the h'ISt ory 0 f expenments
. . not well
IS
or growth, and consequently the geometry of crystallization chambers or drops t:own, fecause crystal growers do not monitor parameters. This is especially
has to be defined. For additional information see Chapter 1l. e ca.se or temperature, which is almost never known with accurac even 'f
Cessation of growth can have several causes. Apart from trivial ones, like expenments are conducted in thermostated cabinets (this may bY' d I
depletion of the macromolecules from the crystallizing media, it can result
from growth defects, poisoning of the faces, or ageing of the molecules. Better ~~::~:~~~etUh~~e~f~I~ltl;;~;:I:I'h~:ec~~~~~\~~S07%::t:c:;:n;!c~~~~~~g~e:~!~:~~~~
. will b d0 owmg
control of growth conditions, in particular of the flow of molecules around growth 'b c apters methods . . t 0 contro I macromolecular crystal
the crystals, may in some cases overcome the drawbacks as was shown in crystallizatio: s:s~r~:: (e.g. for the kmelIcs of evaporation in vapour diffusion
microgravity experiments (refs 56, 57, and Chapter 6). Chapter 11). 'Althoug/:~:s~' !~~ aclIve VIdeo and temperature control, see
required I'n t . erent aspects are all m theIr mfancy and
s rumentatlOn often d t' " '
3.5 Packing believe that user-friendly meth:;~I:;ie~x~:ilo~:x~sts °lnlydas prototypes, we
With biological macromolecules, crystal quality may be correlated with the laboratories b .. eve ope soon, and that
packing of the molecules within the crystalline lattices, and external crystal may e eqUIpped WIth the adequate instrumentation.
morphology with internal structure. As shown by the periodic bond chain
(PBe) method, direct protein-protein contacts are essential in determining !.3 The future of biological crystallogenesis
packing and morphology (Chapter 11). Forces involved in packing of macro- s mentIOned
present at ver before
low ' m 0 d ' methods give access to . molecules
er.n genetic
molecules may be considered as weak as compared to those maintaining the
structural or y f amounts m cells and help to solve problems linked to
cohesion of small molecule crystals. They involve salt bridges, hydrogen bonds, r"c. In particul
con ormallonal heteroge'f
" . nel les 0 f protems
. reluctant to crystal-
Van der Waals, dipole-dipole, and stacking interactions (58-60). It must also permit easie ar, engl~ee~mg of actIve van ants containing compact cores will
be borne in mind that the weak cohesion of macromolecular crystals results
molecular e r c rystalhzatIons of proteins with flexible domains Macro
from the fact that only a small part of macromolecular surfaces participate in
World, Wherengmeenng
man ' will ce r t am . Iy fi nd many applications in the
. RNA-
intermolecular contacts (61), the remaining being in contact with the solvent deserve structu I y nbozymes, pseudo-knots, and other RNA constructs
(exceptions may be found for small proteins). This explains the commonly crystal perfecf ra mveslIgations (Chapter 8) . On the other hand search of
observed polymorphism of biological macromolecular crystals. mcnt of data c~~~ ' preparation of large size crystals, together with improve-
~ihallenges, notab~ct~~~ methods at, synchrotron or neutron sources, are other
4. From empiricism to rationality mpc-resolved crYS;all the resolutIOn of structures at highest resolution and
Or the b' I . ography (62) .
4.1 Towards a better understanding of parameters lugi cal problems
10 agIst
a~dstud ymg
. crysta I growth should be correlated with bio-
T o date, the major parameters underlying crystallization of macromolecules , crystalhzatlOn projects on macromolecular complexes,
have been recognized (Table 1), and if their respective contributions in the
11
 R. Giege and A. Ducruix 1: An introduction to the crystallogenesis
on membrane proteins (Chapter 9) , and especially on engineered proteins, 5. Is it a fusion protein? If yes, which protease is used to cleave the
are bcing developed. For the physicist, growing large monocrystals can be a protein?
goal in itself, and one might speculate that exploration of optical, electrical,
mechanical, and other physical properties of crystalline arrays made from 6. How many mg/litre of cu lture can you produce?
biological macromolecules or assemblies can lead to novel frontiers in the 7. How many mg of protein can you obtain per standard purification?
material science of tomorrow. Finally, for the chemist, usage of chemical and
molecular biology tools could lead in future to the design of molecular devices B. Biochemistry
and other nanostructures mimicking macromolecular crystals, as was discussed
for artificial self-assembling nucleic acids (63). 1. How long does it take to purify one batch of protein?
In conclusion, the rational approach for prompt crystallizations will de- 2. How do you assess the purity of the protein?
mand a synergy between biochemically and physically directed research and (a) Electrophoresis (native/denaturating conditions).
usage of automated methods for the control of nucleation and growth, as well (b) HPLC (which phase).
as the rapid preparation of high-quality monocrystals.
(c) Ion spray (mass spectrometry).
(d) Checking of N-terminus.
5. Search for crystallization strategies (e) Activity assay.
No universal answer(s) can be given to the obvious questions about how to 2. What are the principal characteristics of the protein?
start a crystallization project and what kind of strategy would be the most (a) Molecular weight.
appropriate to be successful. Because of the multiparametric nature of the
(b) Isoelectric point (ca lculated or measured).
crystallization process and the diversity of the individual properties of pro-
teins, the advice would be to collect all possible information on the protein (c) Glycosylation (yes/no).
one intends to crystallize, so that to hierarchize the variables of the process (d) Number of free cysteine(s).
and if possible to rcstrict their number. To this aim the questionnaire in (e) Number of disulfide bridges.
Protocol 1 addresses a number of basic questions. (f) Hydrophobicity or hyd rophilicity.
(g) What are the ligands?
Protocol 1. Before starting crystallizations, what are the 3. What are the friendly (or unfriendly) solvents?
biological and biochemical characteristics of your
protein? 4. Is the protein monomeric of oligomeric? How did you check it
(chromatography, light scattering, others)? Does your protein has a
tendency to aggregate?
If your project concerns nucleic acids or their complexes with proteins
see Chapter 8; for membrane proteins see Chapter 9. 5. What is the stability of the protein versus time, temperature, or pH?

A. Biology and production
1. What is the biological origin of your protein? From a micro-organism
(mesophi le or thermoph ile), a plant, an anima l (which tissue)? etc.
2. Is the gene of the protein sequenced?
3. Is the protein cloned? In what expression vector and in what cells was
it overexpressed (E. coli, yeast, bacu lovirus, others)?
4. Is it a his-tagged protein? If yes do you plan to crystallize it with or
without the tag?
a The better these questions can be answered, the easier will be the design of
crystalhzatIOn strategy. Good knowledge of the characteristics of the pro-
tem, of its availability, will guide the experimenter. Quite often, it helps
people to be aware that the premise of crystallization may be as important as
the crystallization itself.
References
1. Pi~kworth Glusker, 1. and Trueblood , K. N. (1985). Crystal structure analysis, a
pruner. Oxford University Press, New York.
 R. Giege and A. Dueruix 1: An introduction to the erystallogenesis
2. Drenth, J. (1995). Principles of protein X-ray crystallography. Springer-Verlag,
leruzalmi, D. and Steitz, T. A. (1997) . 1. Mol. Bioi., 274, 748.
Berlin. 31 . Giege, R., Dock, A.-C., Kern, D., Lorber, B., Thierry, J.-c., and Moras, D. (1986).
3. Wright, J. D. (1987). Molecular crystals. Cambridge University Press, Cambridge. 32 6 554 .
1. Cryst. Growth,7, .
4. Lima-de-Faria, J. (ed.) (1990). HislOrical atlas of crysrallography. Kluwer, 31 Yonath, A., Frolow, F., Shoham, M., Mtissig, l., Makowski, I., Glotz, c., et al.
Dordrecht. _.' (1988).1. Cryst. Growth, 90, 231. .
5. Hajdu, J. , Acharaya, K. R., Stuart, D. I., Barford, D., and Johnson, L. N. (1988). 34. Trakhanov, S., Yusupov, M., Shirikov, V., Garber, M., Mltschler, A., Ruff, M., et
Trends Biochem. Sci., 13,104. al. (1989).1. Mol. Bioi., 209, 327. .
6. Mozzarelli, A. and Rossi, G. L. (1996). Annu. Rev. Biophys. Bioma!. StruCI., 25,
35. Feigelson, R. S. (ed.) (1986). Proc. 1st Int. ConL Protem Crystal Growth,
3430. Stanford, CA, USA, 1985. J. Cryst. Growth, 76, 529.
7. Blundell, T. L. and Johnson, L. M. (1976). Protein crystallography. Academic 36. Giege, R., Ducruix, A., Fontecilla-Camps, J., Feigelson, R. S., Kern , R., and
Press, New York. McPherson, A. (ed.) (1988). Proc. 2nd Int. ConL Crystal Growth of Biological
8. Wukovitz, W. and Yeates, T. O. (1995). Nature Struct. Bioi., 2, 1060. Macromolecules, Bischenberg, France, 1987. J. Cryst. Growth, 90, 1.
9. Usha, R., Johnson, J. E., Moras, D., T hierry, J.-c., Feurme, R., and Kahn, R. 37. Carter, C. W., lr. (ed.) (1990). Methods: a companion to methods in enzymology,
(1984). J. Appl. Crystallogr., 17, 147.
Vol. 1, pp. 1- 127.
10. Kay, L. L. (1986). ISISI1. Hist. Sci. Soc., 77, 450. 38. Ward, K. and Gilliland, G. (ed.) (1990). Proc. 3rd Int. ConL Crystal Growth of
11. Feigelson, R. S. (1988).1. Cryst. Growth, 90, I. Biological Macromolecules, Washington, DC, USA, 1989. J. Cryst. Growth, 110,1.
12. Boistelle, R. and Astier, J.-P. (1988). J. Cryst. Growth, 90, 14. 39. Stezowsky, J. J. and Littke, W. (ed.) (1992). Proc. 4th Int. Conf. Crystal Growth of
13. Rosenberger, F., Vekilov, P. G., Muschol, M. , and Thomas, B. R. (1996). J. Cryst. Biological Macromolecules, Freiburg, Germany, 1991. J. Cryst. Growth, 122, 1.
Growth, 168, 1. 40. Glusker, J. P. (ed.) (1994). Proc. 5th Int. ConL Crystal Growth of Biological
14. Lehman, C. G. (1853). Lehrbuch der physiologische Chemie. Leipzig. Macromolecules, San Diego, CA, USA, 1993. Acta Cryst., D50, 337.
15. Reichert, E . T. and Brown, A. P. (1909). The differentiation and specificity of 41. Miki, K., Ataka, M., Fukuyama, K., Higuchi, Y., and Miyashita, T. (ed.) (1996).
corresponding proteins and other vital substances in relation to biol?gical ~/as~ifi­ Proc. 6th Int. Conf. Crystal Growth of Biological Macromolecules, Hiroshima,
calion and evolulion: the crystallography of hemoglobins. Carnegie InstItutIOn,
Japan, 1995. J. Cryst. Growth, 168, 1.
Washington DC. 42. Drenth, J., and Garcia-Ruiz, l. M. (ed.) (1999). Proc. 7th Int. ConL Crystal
16. Debru, C. (1983). L'esprit des prottines: histoire et philosophie biochimiques. Growth of Biological Macromolecules, Granada, Spain, 1998.1. Cryst. Growth, 196,
Hermann, Paris. pp. 185-720.
17. McPherson, A. (1990).1. Cryst. Growth, 110, 1. 43. Wood, S. P. (1990). 1n Protein purification applications: a practical approach (ed.
18. Sumner, J. B. (1926). J. Bioi. Chem., 69, 435. E. L. V. Harris and S. Angal), pp. 45-58. IRL Press, Oxford.
19. Northrop, 1. H., Kunitz, M., and Herriot, R. M. (1948). Crystalline enzymes. 44. Weber, P. C. (1991). Adv. Protein Chern ., 41,1.
Columbia University Press, New York. 45. McPherson, A., Malkin, A. J., and Kuznetsov, Y. G. (1995). Structure , 3, 759.
20. labri, E. , Carr, M. B., Hausinger, R. P., and Karplus, P. A. (1995). Science, 268, 46. Durbin, S. D. and Feher, G. (1996). Annu. Rev. Phys. Chern., 47,171.
998. 47. Jakoby, W. B. (1971). In Methods in enzymology (ed. W. B. Jakoby), Vol. 22,
21. Dounce, A. L. and Allen, P. Z. (1988). Trends Biochem. Sci., 13,317. pp. 248-52. Academic Press, London .
22. Bernal, J. D. and Crowfoot, D. (1934). Natllre, 133, 794. 48. Judge, R. A., Johns, M. R., and White, E. T. (1995). Biotechnol. Bioeng., 48, 316.
23. Jones, c., Mulloy, B., and Sanderson, M. R. (ed.) (1996). In Methods in molecular 49. Lorber, B., Jenner, G., and Giege, R. (1996). J. Cryst. Growth, 103, 117.
biology. Crystallographic methods and protocols, Vol. 114, pp. 1-394. Humana 50. Von Hippel, P. H. and Schleich, T. (1969) . In Structure and stability of biological
Press, Totowa, Nl, USA. macromolecules (ed. S. N. Timasheff and G. D. Fasman), Vol. 2, pp. 417-574.
24. McPherson, A. (1982). Preparation and analYSis of protein crystals. Wiley, New Dekker.
York. 51. Mikol, V. and Giege, R. (1989). J. Cryst. Growth , 97, 324.
25. Kim, S. H . and Rich, A. (1968). Science, 162,1381. 52. Ries-Kautt, M. and Ducruix, A. (1989). J. Bioi. Chern. , 264, 745.
26. Dock, A.-c., Lorber, B., Moras, D ., Pixa, G., Thierry, J. -c., and Giege, R. (1984). 53. Cacioppo, E., Munson, S., and Pusey, M. L. (1991). J. Cryst. Growth, 110, 66.
Biochimie, 66, 179. 54. Gilliland, G. L., Tung, M., Blakeslee, D. M., and Ladner, l. E. (1994). Acta Cryst.,
27. Michel, H. (1982).J. Mol. Bioi., 158, 567. D50,408.
28. Ktihlbrandt, W. (1988). Q. Rev. Biophys., 21, 429. 55. Thaller, C, Weaver, L. H., Eicheie, G., Wilson, E., Karlson, R., and Jansonius, 1.
29. Arnoux, B., Ducruix, A., Reiss-Husson, F., Lutz, M., Norris, J., Schiffer, M., et al. N. (1981). J. Mol. Bioi., 147, 465.
(1989). PEBS Lett., 258, 47. 56. Giege, R., Drenth, J., Ducruix, A., McPherson, A., and Saenger, W. (1995). Prog.
30. Michel, H. (ed.) (1991). Crystallization of membrane proteins. CRC Press, Boca Cryst. Growth Charact. , 30, 237.
Raton, FL, USA. 57. MCPherson, A. (1997). Trends Biotechnol. 15, 197.
14
 R. Giege and A. Ducruix
58. Bergdoll, M. and Moras, D. (1988). j. Cryst. Growth , 90, 283.
59. Salemme, F. R. , Genieser, L., Finzel, B. c., Hilmer, R. M., and Wendolosky, J. J.
(1988). j . Cryst. Growth, 90, 273. .
60. Wang, A. H. J. and Teng, M. K. (1988). j. Cryst. Growth, 90, 295.
61. Carugo, O . and Argos, P. (1997). Protein Sci., 6, 2261.
62. Chayen, N. E., Boggon, T. J., Cassetta, A., Deacon, A., Gleichmann, T., Habash,
J., et al. (1996). Q. Rev. Biophys. , 29, 227.
63. Seeman, N. C. (1991). Curro Opin. Strllct. Bioi. , 1, 653.
Biochemical aspects and handling of
macromolecular solutions and
crystals
B . LORBER and R. GlEGE

1. Introduction
The quality and quantity of the macromolecular samples are important pre-
requisites for successful crystallizations. Proteins and nucleic acids extracted
from living cells or synthesized in vitro differ from small molecules by
additional properties intrinsic to their chemical nature and their larger size.
They are frequently difficult to prepare at a high degree of purity and
homogeneity. Besides traces of impurities, harsh treatments may decrease their
stability and activity through different kinds of alterations. Consequently, the
quality of biomacromolecules depends on the way they are prepared and
handled. As a general rule purity and homogeneity are regarded as conditions
sine qua non. Accordingly, purification, stabilization, storage, and handling of
macromolecules are essential steps prior to crystallization attempts. Other
difficulties in crystal growth may come from the source of the biological
material. It is advisable to have at disposal a few milligrams of material when
starting first crystallization trials although structures were solved with sub-
milligram quantities of protein (1). Once crystals suitable for X-ray analysis
can be produced, additional material is often needed to improve their quality
and size and to prepare heavy-atom derivatives. It is thus essential that
isolation procedures are able to supply enough fresh material of reproducible
quality. Similar situations are encountered with multi-macromolecular
assemblies (e.g. viruses, nucleosomes, ribosomal particles, or their subunits).
. This chapter discusses biochemical methods used to prepare and character-
Ize macromolecules intended for crystallization assays. Practical aspects
concerning manipulation and qualitative analyses of soluble proteins will be
emphasized. The cases of nucleic acids and membrane proteins are described
In more detail in Chapters 8 and 9. Peculiar aspects of molecular biology that
are important for crystallogenesis are presented in Chapter 3. They include
the design of engineered macromolecules with new physical properties or