Tables for
Volume F
Crystallography of biological macromolecules
Edited by E. Arnold, D. M. Himmel and M. G. Rossmann

International Tables for Crystallography (2012). Vol. F, ch. 4.1, pp. 99-100   | 1 | 2 |

Section 4.1.1. Introduction

C. Sauter,a B. Lorber,b A. McPhersonc and R. Giegéd*

aInstitut de Biologie Moléculaire et Cellulaire (IBMC), Centre National pour la Recherche Scientifique (CNRS), 15 rue René Descartes, Strasbourg, F-67084, France,bUPR 9002, IBMC–CNRS, 15 rue René Descartes Cedex, Strasbourg, 67084, France,cDepartment of Molecular Biology and Biochemistry, University of California, 560 Steinhaus, Irvine, CA 92697–3900, USA, and dMachineries Traductionnelles, ARN, UPR 9002, IBMC du CNRS, 15 rue René Descartes, Strasbourg, 67084, France
Correspondence e-mail:

4.1.1. Introduction

| top | pdf | Prologue

| top | pdf |

Macromolecular crystals are much softer than salt crystals. They contain [\sim]50% of solvent on average, with values ranging from as little as 25 up to 90%. A direct consequence is that, although morphologically indistinguishable, crystals of macromolecules differ in many respects from crystals of low-molecular-mass compounds. While the latter exhibit firm lattice forces, are highly ordered, generally physically hard and brittle, easy to manipulate, can usually be exposed to air, have strong optical properties and diffract X-rays intensely, crystals of macromolecules are, by comparison, smaller in size, they crush easily, disintegrate if allowed to dehydrate, exhibit weak optical properties and diffract X-rays poorly. They are temperature sensitive and undergo extensive damage after prolonged exposure to radiation.

Proteins or nucleic acids build up a crystalline scaffold, which may be imagined as an ordered gel with extensive interstitial spaces through which small molecules can diffuse freely. In proportion to molecular mass, large macromolecules establish far fewer packing interactions than do small molecules inside crystalline lattices. Since these contacts are responsible for the integrity of the crystals, this largely explains the differences in properties between the two types of crystals. Thus, liquid channels and solvent cavities are directly responsible for the generally poor diffraction properties of macromolecular crystals. Owing to the large spaces between adjacent molecules and the related weak lattice forces, every molecule in the crystal may not occupy exactly equivalent orientations and positions. Furthermore, because of their structural complexity and their conformational dynamics, macromolecules in a given crystal form may exhibit slight variations in their folding patterns or dispositions of side chains.

However, high solvent content is not as negative as it might appear at first glance. It allows maintenance of the macromolecular structures virtually unchanged from those in bulk solvent. As a consequence, ligand binding, enzymatic and spectroscopic characteristics, and other biochemical features are essentially the same as for the native molecule in solution. In addition, the dimensions of solvent channels are such that conventional chemical compounds, such as ions and heavy atoms, substrates or other ligands, may be freely diffused into and out of the crystals. Thus, many crystalline enzymes, though immobilized, are completely accessible for experimentation through alteration of the surrounding mother liquor.

The intrinsic instability of most macromolecules requires that conditions suitable for crystal growth are those that do not perturb their molecular properties. This explains why crystals must be grown from solutions compatible with the target macromolecules, i.e. within a narrow range of pH, temperature or ionic strength. Finally, because hydration is essential for the maintenance of structure, crystals of macromolecules must be always bathed in the mother liquor, even during data collection (except in the practice of cryocrystallography). Crystallization principles

| top | pdf |

The crystallization of biological macromolecules has often been considered unpredictable, although it obeys the same principles as that of small molecules (Giegé et al., 1995[link]; McPherson et al., 1995[link]; Rosenberger, 1996[link]; Chernov, 2003[link]; McPherson & Giegé, 2007[link]). It is, similarly, a multiparametric process. The difference compared with conventional crystal growth arises from the biochemical and biophysical properties of proteins or nucleic acids, and crystallization methods must take into account these features. The methods described below apply for most proteins, large RNAs, multimacromolecular complexes and viruses (for small oligonucleotides or peptides, crystallization by dialysis is not appropriate). For hydrophobic membrane proteins, special techniques are required (see Chapter 4.2[link] ).

Crystallization proceeds from macromolecules in solution that `aggregate' upon entering a supersaturated state and eventually undergo a phase transition. This leads to nuclei formation and ultimately to crystals that grow by different mechanisms. Supersaturation is the driving force of crystallization and is defined as the ratio [C]/[S], where [C] and [S] are the initial concentration of the macromolecule and its final concentration at saturation, i.e. its solubility. Nucleation is homogeneous when nuclei form in the bulk of the solution, but heterogeneous when they preferentially form on walls of crystallization vessels, on solid particles (dust, aggregates, seeds), or on the surface of existing crystals. Unlike most conventional crystals, protein crystals are, in general, not initiated from seeds, but are nucleated ab initio at high levels of supersaturation that can reach 200 to 1000% (in what follows and for simplicity, the generic name `protein' is used for macromolecule). It is this high degree of supersaturation that, in large part, distinguishes protein crystal formation from that of conventional crystals. That is, once a stable nucleus has formed, it subsequently grows under very unfavourable conditions of excessive supersaturation. Distant from the metastable zone, where ordered growth could occur, crystals rapidly accumulate nutrient molecules, as well as impurities. They also concomitantly accumulate statistical disorder and a high frequency of defects that exceeds those observed for most conventional crystals.

The different stages of crystallization (i.e. pre-nucleation, nucleation, growth, cessation of growth) can be visualized in a phase diagram (Fig.[link]). In short, phase diagrams are divided into undersaturated regions (where proteins are soluble) and supersaturated regions (where protein crystals nucleate and grow) delimited by the solubility curve. The supersaturated region is thermodynamically out of equilibrium and can be divided into three kinetically dependent domains: a precipitation domain (at extreme supersaturation) where macromolecules rapidly separate from solution in a solid state either amorphous or microcrystalline, a domain (at lower supersaturation) where nucleation occurs spontaneously and a metastable domain (at low supersaturation) where nucleation does not occur spontaneously but where crystals grow. This domain is favourable for seeding. The wisdom of the crystal grower will be to take advantage of an overall understanding of phase diagrams for designing crystallization strategies and selecting favourable solvent conditions (Sauter, Lorber et al., 1999[link]; Asherie, 2004[link]).


Figure | top | pdf |

Crystallization is a multiparametric process under the control of a great variety of biochemical, chemical and physical parameters. The crystal grower can play with them to drive a macromolecule from the soluble towards the crystalline state. In the top left, a typical two-dimensional phase diagram illustrates how this can be achieved. (a) By modifying two parameters (concentrations of macromolecule and crystallant), the system will move from the undersaturated region where the macromolecule is soluble into the supersaturated region beyond the solubility curve where it will try to escape from the solution. The prenucleation trajectory essentially depends on the crystallization method (see Fig.[link]). (b) As soon as they enter the supersaturation region, macromolecules will tend to aggregate. However, the system needs to cross an energy barrier to produce stable molecular assemblies, the nuclei, and this only happens in the so-called nucleation zone. (c) When stable nuclei are formed, they will capture more macromolecular entities from the mother liquor and produce three-dimensional crystals. (d) Crystals will grow until the system comes back to the solubility curve, crossing the metastable zone. Growth stops and crystals are in dynamic equilibrium with the mother liquor. When the system is driven to high supersaturation, macromolecules may rather produce amorphous or microcrystalline precipitates than useful monocrystals. Note that the limits between zones of the supersaturated zone (dashed curve) move with time.


Asherie, N. (2004). Protein crystallization and phase diagrams. Methods, 34, 266–272.
Chernov, A. A. (2003). Protein crystals and their growth. J. Struct. Biol. 142, 3–21.
Giegé, R., Drenth, J., Ducruix, A., McPherson, A. & Saenger, W. (1995). Crystallogenesis of biological macromolecules. Biological, microgravity, and other physico-chemical aspects. Prog. Cryst. Growth Charact. 30, 237–281.
McPherson, A. & Giegé, R. (2007). Crystallogenesis research for biology in the last two decades as seen from the international conferences on the crystallization of biological macromolecules. Cryst. Growth Des. 7, 2126–2133.
McPherson, A., Malkin, A. J. & Kuznetsov, Y. G. (1995). The science of macromolecular crystallization. Structure, 3, 759–768.
Rosenberger, F. (1996). Protein crystallization. J. Cryst. Growth, 166, 40–54.
Sauter, C., Lorber, B., Kern, D., Cavarelli, J., Moras, D. & Giegé, R. (1999). Crystallogenesis studies on yeast aspartyl-tRNA synthetase: use of phase diagram to improve crystal quality. Acta Cryst. D55, 149–156.

to end of page
to top of page