During the 1990s, there was explosive growth in the number and complexity of macromolecular structures being determined by X-ray crystallography, as evidenced by the exponential increase in the number of structures published and submitted to the Protein Data Bank. This growth has been made possible by the convergence of a large number of new technologies, including the following:
Improved systems for cloning and expressing wild-type and mutant proteins;
Improved protein and nucleic acid purification techniques;
Immortalization of crystals by cryogenic freezing;
Very high brilliance X-ray synchrotron sources;
Fast, accurate area detectors with high dynamic range;
Superfast, inexpensive computers; and
Readily available software packages for data acquisition and reduction, phasing, and refinement.
For the most part, however, protein crystallization is done in much the same trial-and-error manner it was a decade ago, albeit with somewhat less tedium since the introduction of reagent kits and the growing use of automated systems. It is still more art than science. NASA, to its credit, has sponsored a large number of ground-based research projects aimed at understanding the fundamentals of the crystallization process. These included investigations of depletion zones around growing crystals (McPherson et al., 1999), studies of defect formation during protein crystal growth and the effects of these defects on diffraction resolution (Dobrianov et al., 1998, 1999), and analyses of predictors for protein crystallization using light-scattering measurements (Kao et al., 1998; Ansari et al., 1997). 1
These studies of the crystallization process have occasionally included flight components (McPherson et al., 1999). However, one of the main goals of NASA's program on crystallization in the microgravity environment has been the growth of crystals in space that are of better quality than those available on the ground. In this report, the task group focuses on evaluating the results of the program's effort in this area to date, commenting on the hardware available and in development for future work on the ISS and offering suggestions for improving the project selection process and NASA's outreach to the scientific community.
The determination of macromolecular structures by X-ray crystallography at a level of detail sufficient for the construction of reliable atomic models requires crystals that diffract X rays to Bragg spacings of 3.5 Å or better. The minimal Bragg spacing to which diffraction measurements can be obtained, loosely referred to as the resolution of the crystallographic analysis, limits the accuracy of the resultant structure in two ways. First, the resolution of the analysis places a limit on the structural features that can be directly visualized in electron density maps calculated using the X-ray data. A resolution of at least 3.5 Å is required to see structural elements in proteins, such as alpha-helices or beta-sheets. Second, once an initial atomic model has been constructed, the resolution of the analysis determines the accuracy with which the parameters of the atomic model can be refined. Positional coordinates in a refined macromolecular structure are determined much more precisely than the resolution of the analysis would indicate. When stereochemical constraints are used in the refinement, such as information about the bond lengths between atoms, the precision of a protein structure typically is approximately 0.5 Å for an analysis carried out at 3 Å resolution and is better than 0.1 Å for a 1.5 Å resolution analysis. In the relatively rare cases where data to better than 1 Å are obtained, individual hydrogen atoms can often be distinguished and the disorder within the protein structure can be described in detail.
These are a few examples of the projects under way; a complete list and description of NASA-funded projects in protein crystal growth can be obtained on the Web at <http://peer1.idi.usra.edu/peer_review/taskbook/micro/mg99/mtb.cfm >.