Methods 34 (2004) 254–265 www.elsevier.com/locate/ymeth Introduction to protein crystallization Alexander McPherson¤ Department of Molecular Biology and Biochemistry, University of Califonia, Irvine 560 Steinhaus Hall, Irvine, CA 92697-3900, USA Accepted 24 March 2004 Abstract Biological macromolecules can be crystallized by a variety of techniques, and using a wide range of reagents which produce super- saturated mother liquors. These may, in turn, be applied under diVerent physical conditions such as temperature. The fundamental approaches to devising successful crystallization conditions and the factors that inXuence them are summarized here. For the novice, it is hoped that this brief review might serve as a useful introduction and a stepping-stone to a successful X-ray strucutre determina- tion. In addition, it may provide a framework in which to place the articles that follow. 2004 Elsevier Inc. All rights reserved. 1. Some history [50–52,43] proved an important tool in establishing the properties and nature of catalytic macromolecules. In Protein crystallization developed in the latter half of the late 1930s, however, a new application for protein the 19th century for three reasons, (a) it provided a crystallization appeared as a result of the studies, using means for the puriWcation of speciWc proteins from an X-ray diVraction analysis, of Bernal and Crowfoot [2], otherwise impure mixture at a time when few other Perutz [45], and others. Today, though crystallization is means existed, (b) it served as a demonstration that a still an admired and respected procedure by enzymolo- protein had been puriWed (which even now is taken as a gists and protein chemists, X-ray crystallography, and pretty good measure), and (c) it was an interesting labo- the structure determination of macromolecules and their ratory curiosity. Initially, the crystallization of hemoglo- complexes stand as the principal objective of those bin from a variety of sources was really nothing more involved in crystallization. than that [36], though a thoughtful attempt to relate A fundamental change in protein crystallization, its hemoglobin crystal form to evolution was made around investigation and its application, occurred in the 1980s. 1900 [48]. The Wrst two reasons, however, were dominant This was due to the development of recombinant DNA in the last quarter of the 19th century and biochemists technology which permitted researchers, for the Wrst such as Osborne used it extensively to isolate and char- time, to prepare ample amounts of otherwise rare and acterize proteins, particularly those from seeds. The elusive proteins. Currently, the majority of proteins meaning of this early research for us today is to show addressed by X-ray diVraction are derived from recom- Wrst of all that speciWc proteins can often be isolated binant sources. Among the other consequences is that from quite impure preparations, but more important, today we are generally working with more homogeneous these pioneers deduced many of the approaches for the protein samples which exhibit greater reproducibility growth of protein crystals that are still in use. than ever before. Needless to say, this has both acceler- Between 1900 and 1940, the emphasis was on ated the progress of X-ray crystallography, and greatly enzymes, and again, crystallization in the hands of Sum- expanded both its applicability and its appeal to bio- mer, Northrup, Kunitz, Herriott, and their colleagues chemists and molecular biologists [37,38]. Ultimately, structural biologists would like to describe all living systems, and the materials they produce, in ¤ Fax: 1-949-824-5992. molecular and even atomic terms. This Wrst requires a E-mail address: [email protected]. precise knowledge of the building block molecules, the 1046-2023/$ - see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ymeth.2004.03.019 A. McPherson / Methods 34 (2004) 254–265 255 proteins, nucleic acids, lipids, and polysaccharides. It In addition to the dramatic impact that knowledge of further requires information of a somewhat diVerent three-dimensional structures of proteins has had on fun- nature, rules or guidelines to specify how the building damental research in biochemistry and biology, macro- block molecules are joined, organized, and assembled molecular structure, is of formidable value in into higher order structures. Those greater structures biotechnology as well. Here, it provides the essential include macromolecular complexes, assemblies, organ- knowledge required to apply the technique of rational elles, cell walls, membranes, cytoskeletons, etc. From drug design in the creation and discovery of new drugs these assemblies we can Wnally reach a molecular and and pharmaceutical products [37,38]. It serves as the atomic level description of the living cell, and from that basis of powerful approaches now being applied in understand, in terms of classical chemistry and physics, emerging biotechnology enterprises, as well as major the architecture and mechanics of living matter. pharmaceutical companies, to identify lead compounds The dynamics within the cell, the mechanisms respon- to treat a host of human ailments, veterinary problems, sible for the dynamics, and the cell's interactions with and crop diseases in agriculture. The underlying hypoth- exterior inXuences are equally important. To understand esis is that if the structure of the active site of a salient those, however, we Wrst need to delineate how the build- enzyme in a metabolic or regulatory pathway is known, ing block molecules respond to chemical and physical then chemical compounds, such as drugs, can be ratio- forces, how the responses are regulated, and how the nally designed to inhibit or otherwise aVect the behavior responses are transmitted through the hierarchy of of that enzyme. assemblies and higher structures. This in turn means A second approach, of equal importance to biotech- visualizing the building block molecules, not in a single nology that also requires knowledge of three-dimen- state, but in all of those available as a consequence of sional macromolecular structure is the genetic their molecular interactions. engineering of proteins. Although recombinant DNA The salient elements of this more detailed and com- techniques provide the essential synthetic role that per- prehensive understanding of life's design and processes mits modiWcation of proteins, structure determination are the structures of the building block molecules, and by X-ray crystallography provides the analytical func- the principles of how they assemble and interact. To the tion. It serves as a structural guide for rational and pur- precision required, these properties can only be poseful changes, in place of random and chance amino addressed by X-ray crystallography. The atomic struc- acid substitutions. Direct visualization of the structural tures of the building block molecules, the proteins and alterations that are introduced by mutation oVers new other macromolecules, must be elucidated. This includes directions for chemical and physical enhancements. not only those easily solubilized and crystallized, but Presently, and in the foreseeable future, the only tech- also those that resist current techniques. nique that can yield atomic level structural images of Progress in molecular biology and its application to biological macromolecules is X-ray diVraction analysis human medicine, agriculture, and industrial processes as applied to single crystals. While other methods may have for the past two decades been crucially dependent produce important structural and dynamic data, for the on a detailed knowledge of macromolecular structure at purposes described above, only X-ray crystallography is the atomic level. This has included proteins, nucleic adequate. As its name suggests, application of X-ray acids, viruses, and other large macromolecular com- crystallography is absolutely dependent on crystals of plexes and assemblies. Redundancies in structural ele- the macromolecule, and not simply crystals, but crystals ments emerging from the now constant Xow of newly of suYcient size and quality to permit accurate data col- determined molecular structures suggest that the num- lection. The quality of the Wnal structural image is ber of naturally occurring structural motifs and sub- directly determined by the perfection, size, and physical structures (domains) may be Wnite. Ultimately then, all properties of the crystalline specimen, hence the crystal macromolecular structures may be classiWed and cata- becomes the keystone element of the entire process, and logued according to polypeptide folds. Once all, or most the ultimate determinant of its success [35]. of the folds which are utilized by nature, are known, then When crystallizing proteins for X-ray diVraction this will provide predictive insight, based primarily on analysis, one is usually dealing with homogeneous, often amino acid sequence, into the structures and functions of exceptionally pure, macromolecules, and the objective unknown proteins. The sequences of most proteins, it is may be to grow only a few large, perfect crystals. It is important to note, are currently being elucidated by a important to emphasize that while the number of crys- broad array of sequencing eVorts, such as the human tals needed may be few, often the amount of protein genome project, carried out both by government and the available may be severely limited. This in turn places private sector. Extension of these genome projects