Editorial

Macromolecular crystal growth: prospects for the future

Perhaps the most significant development of the last decade, so far as macromolecular crystal growth is concerned, has been its expanded accessibility, not only to X-ray diffractionists, but to molecular and structural biologists. Growing protein, nucleic acid, and virus crystals is no longer a family business confined to the community of crystallographers. Commercial screening matrices, databases, and a growing instructive literature have made it available as a tool to anyone with a little imagination and a willingness to explore new technologies. Structure-function correlations are now de rigor in enzymology, drug design based on the crystal structures of complexes a la mode. Exploitation of crystallographic results to explain and describe natural phenomena will continue to grow, extension of the methodologies to more difficult structural problems will further drive attempts to crystallize even more macromolecules, and this will, in turn, inspire a need for more effective and expeditious techniques to grow those crystals.

New challenges to our abilities to grow crystals appear, however, with increased frequency. We can crystallize most macromolecules if we can only purify, stabilize, and solubilize them. We can even crystallize them in association with a variety of coenzymes, inhibitors, ligands and effectors. But there are still membrane proteins, large nucleic acids, fibrous proteins, dynamic multidomain proteins, enveloped viruses, glycoproteins, lipoproteins and others too numerous to name or as yet unrecognized. Structural biology is increasingly turning its attention to higher levels of organization, away from single macromolecules, and toward large assemblies. Intricate complexes have already been crystallized; ribosomes, viruses, nucleosomes, arrangements of multiple macromolecules hundreds of angstroms in size. Molecular structure analysis is merging with nanotechnology. Indeed, it is this area of investigation, large asymmetric complexes, that may provide the focus for development of the next generation of crystallographers, and crystal growers.

One might argue that crystallization will be less of a problem, not more. With intense synchrotron radiation, cryocrystallography, and direct methods for phasing, we may need but a single crystal, and a small one at that. But it must be a very good crystal, even though small. Intense radiation can destroy a protein crystal even if it's frozen; freezing itself can crack a crystal to bits or increase it's mosaicity; and if it doesn't diffract to near atomic resolution, then how can we use direct methods? So problems remain, even for those less demanding, less perverse structures, and important questions still seek answers; why do X-rays destroy crystals, why does freezing increase intensity width, why don't crystals diffract to atomic resolution?

Future research on macromolecular crystal growth will address those problems. It has begun to do so now. New methods are being developed using static and quasi elastic light scattering to predict and monitor the process of nucleation. Theories regarding phase transitions and colloid theory are beginning to have an impact on our thinking of how macromolecules interact in solution. Atomic force microscopy and interferometry are yielding quantitative measures for the underlying thermodynamic and kinetic parameters of macromolecular crystallization. The mechanisms of crystal growth are being identified and described. Application of new techniques and revival of old, such as X-ray topography, are being utilized to illuminate the defect structures of macromolecular crystals, and the impact of impurities on the crystal growth process.

Are these investigations important? Without doubt. Definitive evidence is rapidly accumulating that a crystal's mechanical properties, perfection, terminal size, mosaicity, perhaps even the ultimate resolution and quality of its diffraction pattern are at least as much a function of its defect character as to any failings of its constituent molecules. A future trend will be the use of even more emerging technologies to better understand how and why these things are true, and how better to control them.

To this point, we have relied on refinements and variations of reagents, procedures, and approaches which evolved over the past 150 years to grow protein crystals. Will that be the future as well? Perhaps not. Of course we will continue to develop more perceptive, insightful, efficient screens, more intricate mother liquors, and more sophisticated apparatus, but other developments are on the horizon. Surfaces that promote nucleation, organized membrane arrays to inspire two or three dimensional nucleation, use of Fabs and chimeric molecules to alter solubility, designer crystals built with self-assembling units. Genetic engineering, mutations, truncations, clever constructs will increasingly play a role in crystallization. Just as genetic attachment of His tags aided in protein purification, crystallization chaperones may be next. Application of the synthetic power of recombinant DNA methods with the analytic insight of X-ray diffraction will, in the next decades, provide not only the sequence of the human genome, but the structures of its products, and their assemblies. Systematic and rational approaches to crystal growth will provide the bridge.

Crystal growth, in a sense, plays a service role to the greater enterprise of structure determination. Macromolecular crystals are only intermediates in a larger process, they otherwise lack material value. On the other hand, crystal growth and its principles are instructive, and of value for other reasons. Macromolecular crystal growth is currently, because of the size of molecules and their kinetics of growth, the best experimental model we have for crystal growth of any molecules from solution. It serves as well, both in practice and in principle, as a model for self-assembling systems, both biological and otherwise, and contributes to an understanding of their fundamental processes. More recently a confluence of ideas associated with macromolecular crystal growth, colloid physics, and theories of phase transitions has begun emerging with potentially profound consequences for our understanding of events in concentrated macromolecular solutions.

As computer and electronics based technologies continue to increase in power, as mathematical approaches are further refined, macromolecular crystal growth will assume a role as the keystone of X-ray crystallography. Fortunately for all of us, its ideas and applications are evolving and strengthening, and our confidence is increasing as we better understand, both quantitatively and qualitatively, the phenomena involved. Macromolecular crystallization will not be a rate limiting problem, a barrier to be surmounted, it will be one of our most powerful implements for opening new avenues of research into structural biology and physical biochemistry.

Alexander McPherson