X-ray Vision in Structural Genomics

Updated! Suppliers of Tools for X-ray Crystallography Courtesy of Amersham Pharmacia BiotechDetail of the electron density map of deacetoxycephalosporin C synthase from Streptomyces clavuligerus. Two important approaches can be used to determine the three-dimensional structure of macromolecules. Nuclear magnetic resonance (NMR) spectroscopy yields information on the structure of proteins in solution, but it has a size limitation of approximately 150 amino acid residues (about 16,500 daltons), thereby limiting its usefulness to smaller proteins. In contrast, X-ray crystallography has no apparent size limit for generation of structural data, but it requires crystallization of the target molecule. Owing to recent technical advances, X-ray crystallography is now the preferred method for precise structural determination of proteins.1 In the last five years, researchers have published reports on the nucleotide sequences of several eukaryotic genomes. However, the functions of the majority of gene products from these genomes, including those from the human genome, are currently unknown. In 1999, a group of structural biologists called for a structural genomics initiative to determine the three-dimensional structure of all proteins expressed in the human genome using high-throughput crystallization techniques.2 In September 2000, the National Institute of General Medical Sciences (NIGMS) of the National Institutes of Health sponsored a protein structure initiative to determine the three-dimensional structures of thousands of proteins within the decade.2 Like all such initiatives, the goal of the structural genomics initiative is information. Scientists engaged in structural genomics seek to extract useful information about the biological function of a macromolecule from its three-dimensional structure, which may yield important clues about the protein's biochemical function when compared to other proteins whose structure and function are known. For instance, structural studies may identify highly conserved amino acid residues within proteins that might not be apparent at the amino acid sequence level. These structural motifs could then be used for identification of unknown proteins. Ultimately, the functional identification of unknown proteins by X-ray crystallography could dramatically increase scientists' understanding of human disease processes and aid the development of protein-based and drug-based therapies. For example, X-ray crystallographic studies of the mouse Tubby protein (the product of the tub gene responsible for the obese phenotype in the Tubby strain of mice) demonstrated that the core domain of this protein contained a positively charged groove that likely functioned as a DNA-binding protein.3 Subsequent experiments demonstrated that Tubby protein was, in fact, a bipartite transcription factor that could regulate gene expression in cellular assays.3 X-ray crystallographic studies are also useful for evaluating ligand-binding proteins. During crystallization, protein molecules loosely pack together, and this organization permits the diffusion of small molecules to the active sites of proteins. Both natural ligands and drugs that bind proteins at their catalytic sites with high affinities can therefore be added to the crystallization reactions. These ligands become highly specific probes for the active sites of enzymes or ribozymes, yielding insights into how these macromolecules recognize and bind ligands and drugs. An extension of this type of research is structure-based drug design.4 The three-dimensional structure of a molecule can be used to aid the design of new ligands, or to identify new protein targets for previously characterized ligands. Information on the suitability of proteins as receptors for newly developed or currently approved drugs can also be obtained. Researchers have successfully applied this rational approach to drug design in the development of HIV protease inhibitors whose functional activity was based on the three-dimensional structure of HIV-1 protease protein.5 These efforts in rational drug design could be more efficiently carried out if the crystal structures of more pharmacologically important target molecules were available. Finally, structural studies may yield critical insights into the mechanisms of protein folding. A newly translated protein must adopt its native conformation to carry out its biological function. For a protein to fold properly, interactions with solvent water molecules must be broken and replaced with protein-protein interactions. Many researchers believe that a limited number of protein folds exist in nature; current data places the number between 1,000 and 5,000 distinct protein folds.6 Knowledge of protein folding motifs may prove to be an essential component for understanding and eventually improving the expression of proteins, for increasing enzyme stability, or as clues to the function of unknown proteins. An understanding of protein folding has clinical implications as well, as improper protein folding may be an underlying cause of disease. For example, prion proteins in sheep may direct the misfolding of native proteins, leading to scrapie.7 Likewise, Alzheimer's disease and Parkinson's disease may be caused in part by protein misfolding, and the subsequent formation of intracellular deposits that contain misfolded and insoluble proteins.8 Technical Primer In 1912, scientists first observed that a beam of X-rays could be diffracted or scattered in an orderly pattern by electrons of individual atoms within a crystal lattice. Thus, the atomic structure of a crystal can be precisely determined by the way it diffracts a beam of X-rays in different directions. Researchers direct a beam of X-rays with a specific wavelength at a crystal so that a majority of deflected X-ray waves are scattered in various directions. If the diffracted electromagnetic waves are in phase such that the maximal and minimal point of each wave are synchronous, then the waves will reinforce each other and produce a bright light. However, if the diffracted waves are out of phase, the waves will interfere and cancel each other out to produce a darkened signal. Originally, the deflected X-rays were detected using film, requiring manual measurement of the diffraction pattern. The development of charge coupled device (CCD) X-ray detectors and imaging plates has facilitated the rapid acquisition and analysis of structural data from crystallized macromolecules, eliminating the laborious and error-prone measurement step. The atomic resolution of a crystal structure can be enhanced by the use of high-intensity X-ray sources. One source is a synchrotron, which functions as a particle accelerator that circulates electrically charged particles at close to the speed of light. Synchrotrons are the most powerful source for the generation of X-rays, and can be used to supply X-rays for characterization of large protein structures that diffract X-rays poorly. Unfortunately, obtaining a reflection pattern is not enough to solve a crystal structure. Researchers also need to know the phase angle of each reflection. The problem is akin to looking through a microscope without a focusing lens. Since the phase angles are not recorded in the diffraction pattern, structure determination is impossible unless the phase problem is solved. Scientists can employ a number of techniques to solve the phase problem. A common approach is multiple isomorphous replacement (MIR). Atoms of heavy elements such as selenium can be incorporated into specific amino acid residues of recombinant proteins when these gene products are produced in bacteria.9 These seleno-methionine or seleno-cysteine residues act as landmarks that aid in determining the crystallographic phases so that the three-dimensional structure of a molecule can be obtained.9 The diffraction of seleno-methionine "phasing atoms," combined with the knowledge of their ordered positions in the molecule, are then used to calculate the phases for subsequent orientation of all atoms in the crystal.9 Alternatively, heavy metal atoms, such as osmium, iridium, mercury, or uranium, can be introduced into the crystal by soaking the crystal in a solution of the metal. Some molecules contain naturally occurring heavy atoms, which can also be used for phase calculations.9 Tuneable synchrotron radiation sources permit wavelength variation and have enabled the development of multiple wavelength anomalous dispersion (MAD) techniques to rapidly solve phasing problems.10 Like MIR, MAD requires the analysis of heavy metal derivatives. A third approach to solving the phasing problem is molecular replacement. In this technique, the phase of an unsolved crystal structure is solved using a model from a previously solved, related protein. Sir William Lawrence Bragg and his father, William Henry Bragg, devised methods for examining these X-ray diffraction patterns, and developed the basic theory of X-ray crystallography. They observed that if two waves of electromagnetic radiation arrive at the same point in phase and produce maximal intensity, then the difference between the distances they have traveled is an integral multiple of their wavelengths. They defined Bragg's Law, which describes the relationship between the angle of deflection and the periodic spacing of atoms in a crystal lattice. Modern crystallographers use a Fourier transform to refocus the diffraction pattern into an electron density map. In practical terms, computers analyze the diffraction pattern and phase angles to produce electron density maps of the crystallized molecule, and then fit the molecular structure into this calculated electron density map. In the late 1950s, the structural determination of myoglobin by John Kendrew and of hemoglobin by Max Perutz showed that X-ray crystallography could be used to determine the three-dimensional structure of large protein molecules. Likewise, X-ray diffraction patterns obtained by Rosalind Franklin were instrumental in the elucidation of the three-dimensional structure of DNA by Francis Crick, Maurice Wilkins, and James Watson. In 1962, Kendrew and Perutz shared the Nobel Prize in Chemistry, while Crick, Wilkins, and Watson shared the Prize in Medicine. In recent years, X-ray crystallography has been used successfully to examine the structure of even larger cellular complexes--for example, the proteasome11 and both prokaryotic ribosomal subunits.12-15 There are a number of ways to crystallize a protein. In one method, scientists place droplets of the purified protein solution on a cover slip and position it above an enclosed solution of high salt concentration. The salt in the solution extracts water vapor away from the hanging droplet, decreasing droplet volume and increasing protein concentration. Under optimal conditions, this approach results in the formation of highly ordered crystalline structures. The Structure of the Prokaryotic Ribosome The enhanced capability of X-ray crystallography to characterize macromolecular structures was recently demonstrated by structural determination of both subunits of the prokaryotic ribosome. Ribosomes are ribonucleoprotein complexes that catalyze peptide bond formation during protein synthesis. These cytoplasmic structures may also assist in the maturation (folding) of proteins after their translation. Determination of the crystal structure of the ribosome has been a sort of "holy grail" for X-ray crystallographers, due to its central role in translation and its large size, and a number of research groups have been working on this problem.12-15 Courtesy of NASA Marshall Space Flight CenterDetail from a three-dimensional structure of space-grown insulin crystals. In 2000, Thomas Steitz, Peter Moore, and coworkers at Yale University identified the structure of the large ribosomal subunit of the archaebacterium, Haloarcula marismortui, at 2.4-angstrom (Å) resolution.12 Their structural studies indicated that ribosomal proteins are intertwined with ribosomal RNA (rRNA). The 23S rRNA forms a single mass of tightly packed helices, and globular domains of the ribosomal proteins form the exterior of the large subunit. These ribosomal proteins are thought to stabilize and orient the rRNA into a conformation that allows catalytic activity by the 23S rRNA. Researchers hypothesize that this interdependence of rRNA and ribosomal proteins maintains both the structure and function of the large ribosomal subunit. These studies elegantly demonstrated that 23S rRNA catalyzes the peptidyl transferase reaction of protein synthesis, and thus functions as a ribozyme.12 This "all-RNA active site" model was supported by several lines of evidence. First, the active site cleft of the large ribosomal subunit, to which tRNAs bind, is essentially devoid of proteins. Second, this region contains the most highly conserved domain of 23S rRNA sequences. Finally, two ancillary ribosome crystals contained bound tRNA substrate analogs that further identified this cleft as the active site for peptide bond formation. X-ray crystallographic studies further suggested that a highly conserved adenine base functioned in acid-base catalysis during peptide bond formation.12 This study dramatically demonstrates the tremendous potential of X-ray crystallography to yield functional information from structural data, but researchers continue to probe this interesting question. Indeed, researchers in the labs of J.H.D. Cate and Harry Noller recently published a 5.5 Å-resolution crystal structure of the complete 70S ribosome from Thermus thermophilus.15 Going High-Throughput The April 2, 2001, issue of The Scientist reported on the intense interest that has arisen in identifying the function of proteins whose genes have only recently been identified.16 Structural genomics will likely require technological advances in the expression of cloned genes, purification of proteins, and crystallization of purified proteins. For X-ray crystallography studies, a gene must be expressed at sufficient levels, purified to homogeneity, and then crystallized so that it diffracts X-rays for the generation of high-resolution structural data.17 The formation of automated screens for protein expression and affinity tags, such as poly-histidine, for protein purification may provide pure protein in sufficient quantities for X-ray crystallographic studies. At present, small nonmembrane proteins of less than 50 kDa are the most suitable for structural genomics studies. These proteins are generally highly expressed in prokaryotic cells, can be easily purified, and should be amenable to screens for rapid identification of optimal crystallization conditions. Unfortunately, many pharmacologically important proteins are membrane-bound. At present, X-ray crystallographic studies of transmembrane proteins is technically difficult because of the amphipathic nature of these proteins, the difficulty in obtaining sufficient amounts of highly purified protein, and the difficulty in obtaining suitable crystals. As a result, high-resolution structures have been described for only a limited number of membrane proteins. Recently the NIGMS named seven centers to begin structural genomics work on a $150 million, five-year pilot program.18 This work could lead to the solution of up to 10,000 crystal structures in the next 10 years. In contrast, only about 2,000 unique structures have been determined in the past 40 years of research.18 One of the funding recipients is the Joint Center for Structural Genomics, a consortium of research institutions based in La Jolla, Calif., which is currently developing high-throughput robotics techniques for protein production, crystallization, and structure determination. These studies will initially focus on novel macromolecules from Caenorhabditis elegans, and on human proteins thought to be involved in cell signaling. A number of companies have developed kits and reagents to ease the preparation of high-quality crystals. Other companies supply crystal growth chambers and matrices. For a complete listing of companies offering products related to crystallization, see the table on page 31. Improvements in protein isolation, protein purification, crystallization, and X-ray diffraction, could turn the vision of the structural genomics initiative into a reality. One thing at least is certain: the marriage of proteomics, genomics, and structural biology will be a fruitful one, and the number of genes listed as "function unknown" will only decrease. Gregory Smutzer (smutzer@hotmail.com) is a freelance writer in Lansdowne, Pa. David Christianson at the University of Pennsylvania and Helen Pace at Thomas Jefferson University provided technical expertise on this article. References 1.G.T. Montelione, S. Anderson, "Structural genomics: Keystone for a Human Proteome Project," Nature Structural Biology, 6:11-2, 1999. 2.S.K. Burley, "An overview of structural genomics," Nature Structural Biology, Structural Genomics, Suppl.:932-4, Nov. 2000. 3.T.J. Boggon et al., "Implication of tubby proteins as transcription factors by structure-based functional analysis," Science, 286:2119-25, 1999. 4.G. Klebe, "Recent developments in structure-based drug design," Journal of Molecular Medicine, 78:269-81, 2000. 5.R.S. Goody, "Rational drug design and HIV: Hopes and limitations," Nature Medicine, 1:519-20, 1995. 6.S.E. Brenner et al., "Population statistics of protein structures: Lessons from structural classifications," Current Opinions in Structural Biology, 7:369-76, 1997. 7.H. Rezaei et al., "High yield purification and physico-chemical properties of full-length recombinant allelic variants of sheep prion protein linked to scrapie susceptibility," European Journal of Biochemistry, 267:2833-9, May 2000. 8.W. P. Esler et al., "Point substitution in the central hydrophobic cluster of a human beta-amyloid congener disrupts peptide folding and abolishes plaque competence," Biochemistry, 35:13914-21, 1996. 9.T.J. Boggon, L. Shapiro, "Screening for phasing atoms in protein crystallography," Structure with Folding & Design, 8:R143-9, July 15, 2000. 10.W.A. Hendrickson et al., "Selenomethionyl proteins produced for analysis by multiwavelength anomalous diffraction (MAD): A vehicle for direct determination of three-dimensional structure," European Molecular Biology Journal, 9:1665-72, 1990. 11.A. Wlodawer, "Proteasome: A complex protease with a new fold and a distinct mechanism," Structure, 3:417-20, 1995. 12.N. Ban, et al., "The complete atomic structure of the large ribosomal subunit at 2.4 Å resolution," Science, 289:905-20, Aug. 11, 2000. 13.F. Schluenzen et al., "Structure of functionally activated small ribosomal subunit at 3.3 angstroms resolution," Cell, 102:615-23, Sept. 1, 2000. 14.B.T. Wimberly et al., "Structure of the 30S ribosomal subunit," Nature, 407:327-39, Sept. 21, 2000. 15.M.M. Yusupov et al., "Crystal structure of the ribosome at 5.5 Å resolution," Science, 292:883-96, May 4, 2001. 16.D. Steinberg, "Is a human proteome project next?," The Scientist, 15[7]:1, April 2, 2001. 17.A.M. Edwards et al., "Protein production: Feeding the crystallographers and NMR spectroscopists," Nature Structural Biology, Structural Genomics, Suppl.: 970-2, Nov., 2000. 18.R.F. Service, "Structural biology gets a $150 million boost," Science, 289:2254-5, Sept. 29, 2000.

X-ray Vision in Structural Genomics

Interested in reading more?

Become a Member of