REVIEWS

TIBS 25 – DECEMBER 2000 45 Baird, G.S. et al. (1999) Circular permutation and receptor insertion within green fluorescent proteins. Proc. Natl. Acad. Sci. U. S. A. 96, 11241–11246 46 Gordon, G.W. et al. (1998) Quantitative fluorescence resonance energy transfer measurements using fluorescence microscopy. Biophys. J. 74, 2702–2713 47 Bastiaens, P. I.H. et al. (1996) Imaging the intracellular trafficking and state of the AB(5) quaternary structure of cholera-toxin. EMBO J. 15, 4246–4253 48 Bastiaens, P. I.H. and Jovin, T.M. (1998) FRET microscopy. In Cell Biology: A Laboratory Handbook (Vol. 3) (Celis, J.E., ed.), pp. 136–146, Academic Press 49 Wouters, F.S. and Bastiaens, P. I.H. (1999) Fluorescence lifetime imaging of receptor tyrosine kinase activity in cells. Curr. Biol. 9, 1127–1130 50 Bastiaens, P. I.H. and Squire, A. (1999) Fluorescence lifetime imaging microscopy: spatial resolution of biochemical processes in the cell. Trends Cell Biol. 9, 48–52

51 Verveer, P. J. et al. (2000) Global analysis of fluorescence lifetime imaging microscopy data. Biophys. J. 78, 2127–2137 52 Klar, T.A. et al. (2000) Fluorescence microscopy with diffraction resolution barrier broken by stimulated emission. Proc. Natl. Acad. Sci. U. S. A. 97, 8206–8210 53 Nagorni, M. and Hell, S. W. (1998) 4Pi-confocal microscopy provides three-dimensional images of the microtubule network with 100- to 150-nm resolution. J. Struct. Biol. 123, 236–247 54 Hell, S.W. et al. (1997) Far-field fluorescence microscopy with three-dimensional resolution in the 100-nm range. J. Microsc. 187, 1–7 55 Soumpasis, D.M. (1983) Theoretical analysis of fluorescence photobleaching recovery experiments. Biophys. J. 41, 95–97 56 Kubitscheck, U. et al. (1994) Lateral diffusion measurement at high spatial resolution by scanning

Synchrotron crystallography Wayne A. Hendrickson The past decade has seen an explosive growth in atomic-level structures determined by X-ray crystallography. Synchrotron radiation and a number of technical advances related quite directly to its development have fueled this growth. With the most recent advances coming to be used collectively and new resources being built, the foundation is laid for a dramatic further expansion of synchrotron crystallography in the next decade. Both the highthroughput applications of structural genomics and also the challenging studies of macromolecular machinery are expected to flourish.

57

58 59

60

microphotolysis in a confocal microscope. Biophys. J. 67, 948–956 Clegg, R.M. (1996) Fluorescence resonance energy transfer spectroscopy and microscopy. In Fluorescence Imaging Spectroscopy and Microscopy (Wang, X.F. and Herman, B., eds), pp. 179–251, Wiley Adams, S.R. et al. (1991) Fluorescence ratio imaging of cyclic AMP in single cells. Nature 349, 694–697 Wouters, F.S. et al. (1998) FRET microscopy demonstrates molecular association of non-specific lipid transfer protein (nsL-TP) with fatty acid oxidation enzymes in peroxisomes. EMBO J. 17, 7179–7189 Gadella, T.W.J., Jr and Jovin, T.M. (1995) Oligomerization of epidermal growth-factor receptors on A431 cells studied by time-resolved fluorescence imaging microscopy – a stereochemical model for tyrosine kinase receptor activation. J. Cell Biol. 129, 1543–1558

pated growth in macromolecular crystallography. Moreover, advances in nuclear magnetic resonance (NMR) spectroscopy, electron microscopy and computational biology have also contributed to the newly found prominence of structural biology. Yet, crystallography overwhelmingly dominates structure determination, and two developments predominate in its expansion to increasingly complex subjects at evergreater speed. One is the production of appropriate samples from recombinant DNA, which pervades all of molecular biology, and the other develops from the widespread use of synchrotron radiation, which derives from physics.

Synchrotron radiation THE POWER OF X-ray crystallography was evident already 25 years ago when TiBS was launched. Nevertheless, although compelling, the evidence of crystallographic prowess was limited enough that a review article then could tabulate details for each protein crystal structure ever published (78 in all, 15 in 1975)1, and atomic coordinate depositions into the Protein Data Bank (PDB) through 1975 numbered only 18. Now, on average, more new structures are published every week than appeared in a year then, and PDB holdings currently exceed 13 500. These structures cogently address diverse biochemical processes of central importance, and many of them are amazingly complex. Once the richness of X-ray diffraction patterns from protein crystals was revealed in 1934 by Bernal and Crowfoot (later Hodgkin)2, it was immediately clear that atomic-level structures of W.A. Hendrickson is at the Howard Hughes Medical Institute, Dept of Biochemistry and Molecular Biophysics, Columbia University, New York, NY 10032, USA. Email: [email protected]

biological macromolecules were in prospect. ‘... [I]t can be inferred that the arrangement of atoms within the protein molecule is...of a perfectly definite kind,’ they said. Knowing that all atoms contribute to every spot in a diffraction pattern and that the number of unique diffraction spots can greatly exceed the number of atomic parameters, one has a proof that the diffraction data should suffice to define the atomic structure of a crystallized macromolecule. Nevertheless, it took decades before the pioneering efforts of Perutz and Kendrew provided the means to solve the problem of turning diffraction patterns into atomic coordinate sets. Their solution was sufficiently general to support application to a broad set of problems. Improvements were made in the course of succeeding applications, and by the time that TiBS was established it appeared that protein crystallography was a science of textbook maturity3. Applications of this exciting technology continued apace for several years; more recently, however, the expansion has been altogether explosive. Several elements have conspired in this unantici-

0968 – 0004/00/$ – See front matter © 2000, Elsevier Science Ltd. All rights reserved.

Electromagnetic radiation (light) is produced as electrons are accelerated centripetally through bending magnets used to maintain the closed orbit of a synchrotron. This synchrotron radiation was seen initially as a costly waste product in the design of accelerators for high-energy physics experiments, but it was realized later that synchrotron radiation could, in principle, compete favorably with conventional X-ray tubes for X-ray studies4. This opportunity came to the attention of biological crystallographers with the report, published in 1971, of diffraction experiments conducted on muscle at the DESY synchrotron in Hamburg5. The interest then focused on the prospect of greater X-ray fluxes; later on, other properties of synchrotron radiation have been exploited as well, including its continuous spectrum and definite time structure. Although uses for synchrotron radiation in crystallography were already foreseen when TiBS was launched1,3, the first tests were only just being reported6. To many, the benefits of such experimentation at a remote site were not obvious at the outset. Compelling demonstrations

PII: S0968-0004(00)01721-7

637

REVIEWS

TIBS 25 – DECEMBER 2000

Box 1. Glossary of synchrotron sources for macromolecular crystallography Acronym

Facility

Location

ALS ANKA APS BESSY II BSRF CAMD CHESSa CLS DESYb

Advanced Light Source ANKA Synchrotron Radiation Facilityd Advanced Photon Source Berliner Electronenspeicherring Synchrotrond Beijing Synchrotron Radiation Facilityd Center for Advanced Microstructures and Devicesd Cornell High Energy Synchrotron Source Canadian Light Sourced Deutsches Elektronen-Synchrotron Diamondd Elettra European Synchrotron Radiation Facility Light Source of Barcelonad National Synchrotron Light Laboratory Laboratoire pour l’Utilisation du Rayonnement Electromagnétique MAX-laboratory National Synchrotron Light Source Photon Factory Pohang Light Source Synchrotron Experimental Science Applications in the Middle Eastd Swiss Light Sourced Source Optimisee de Lumiere d’Energie Intermediare de Lured Super Photon Ring Shanghai Synchrotron Storage Facilityd Stanford Synchrotron Radiation Laboratory Synchrotron Radiation Research Center Synchrotron Radiation Source

Berkeley, CA, USA Karlsruhe, Germany Argonne, IL, USA Berlin, Germany Beijing, China Baton Rouge, Louisiana, USA Ithaca, NY, USA Saskatoon, Canada Hamburg, Germany Rutherford, UK Trieste, Italy Grenoble, France Barcelona, Spain Campinas, Brazil Orsay, France Lund, Sweden Brookhaven, NY, USA Tsukuba, Japan Pohang, Korea Amman, Jordan Villigen, Switzerland Saclay, France Harima, Japan Shanghai, China Stanford, CA, USA Hsinchu City, Taiwan Daresbury, UK

ESRF LLS LNLS LURE MAX II NSLS PLS SESAME SLS Soleil SPring8 SSRF SSRLc SRRC SRS

aCHESS was built on the Cornell Electron Storage Ring (CESR); bAn European Molecular Biology Laboratory (EMBL) outstation was built on this site; cSSRL was built on the Stanford Positron Electron Accelerator Ring (SPEAR); dNot yet operational for biological crystallography.

took time and expensive equipment, and only in the last decade has the impact been felt strongly. Synchrotron development came in stages. First-generation facilities were parasitic, built around the ‘waste’ from machines designed for physics. These included an EMBL outstation on DESY in Hamburg, SSRL on SPEAR at Stanford and CHESS on CESR at Cornell (see glossary of synchrotron sources in Box 1). Although the accelerator rings were already in place at these sites, a substantial buildout in specialized instrumentation was required to extract X-ray beams. In particular, the continuous spectral output made monochromator devices essential for the first time, and the radiation levels set a new standard for experimental enclosures. Once the utility of synchrotron radiation was established for several areas of research, a second generation of facilities developed that were dedicated to the production of synchrotron radiation. These included machines converted from their origins in physics (e.g. SRS at Daresbury, and SSRL eventually; see Box 1) and those designed expressly to produce X-rays (notably, LURE at Orsay, the Photon Factory in Tsukuba and NSLS at Brookhaven; see Box 1). In optimizing for synchrotron radiation, it was found that magnetic array devices, called wig-

638

glers and undulators, inserted into straight sections between bending magnets, could produce markedly enhanced X-ray beams. Most recently, a third generation of high-energy machines was built to exploit insertion devices, undulators in particular, to the fullest (ESRF in Grenoble, APS at Argonne and SPring8 in Harima; see Box 1) (Fig. 1). Initially, the exploitation of synchrotron radiation in biological crystallography lagged behind that in other research areas, such as materials science. Because crystallography is inherently so powerful, the need for improvement was not always pressing and could be seen as diversionary – ‘the better is the enemy of the good’. New technology would be adopted rapidly once thoroughly proven, but the expense and long lead-time to develop the instrumentation of synchrotron radiation, accelerators and beamlines alike, provided a barrier. The first advantage of synchrotron radiation to be appreciated in crystallography, its intensity, could be provided by single-crystal monochromators on side stations. Several excellent beamlines of this kind were developed and these convinced the crystallographic community of the utility of synchrotron radiation. Subsequently, the value in exploiting the continuous spectral distribution of syn-

chrotron radiation became apparent, but these initial beamlines were inappropriate for multiwavelength anomalous diffraction (MAD) experiments or for timeresolved crystallography based on Laue diffraction. Fortunately, with the inception of third-generation machines, there also came the opportunity to design new beamlines. Concomitantly, the crystallography community has gained sophistication and, albeit belated, it is now a leader in synchrotron development.

Technical advances Macromolecular crystallography has been transformed remarkably in the past quarter century. Notable benefits have come from advances in instrumentation, analytical methods and recombinant DNA technology. There was steady technological progress from the mid-1970s through the 1980s. This was consolidated subsequently and augmented during the 1990s by additional advances that were consequent to synchrotron radiation. Although the fruitfulness of crystallography in the year 2000 is manifest, the full impact of the collective implementation of these most recent developments is yet to be felt. Several innovations happened before 1990. There were important advances in instrumentation: area detector systems proliferated, replacing photographic film

TIBS 25 – DECEMBER 2000

REVIEWS

almost completely; computing (a) (b) power improved dramatically, including the introduction of graphical display units; and synchrotron beamlines were develBending magnet oped, as discussed above. In adUndulator dition, a number of new methods were devised: stereochemically restrained refinement was introduced, becoming BM U generally adopted through the (c) use of computer programs such as PROLSQ and X-PLOR; molecular graphics procedures were developed, with programs such Synchrotron as FRODO changing the way models were fitted into electron-density maps; hangingTi BS drop vapor diffusion and sparse-matrix screens became Figure 1 the standards for crystallizaThird-generation synchrotron radiation sources. Examples are taken from the Advanced Photon Source tion; and molecular replace(APS) at Argonne National Laboratory. (a) Schematic illustration of radiation producing magnetic devices ment grew into a routine proin one synchrotron sector. A synchrotron storage ring with 40 sectors is shown at the bottom, an excerpt cedure for solving unknown containing three straight sections is shown in the middle, and details from a bending magnet and an unstructures from homologs in the dulator are shown at the top. BM, dipole bending magnet; U, undulator array. As electrons (blue) are accelerated through a magnet at high energy (E) they emit X-rays (red) instantaneously into a narrow cone growing collection of knowns, with an opening angle of 1/g = mc2/E, which for APS corresponds to ~70 microradians (~0.004°). For a particularly with program sysbending magnet, the electrons continue to emit X-rays tangential to the entire arc as they pass through tems such as MERLOT and the magnet; thus, the resulting beam is a vertically narrow (1/g) but arbitrarily wide fan from which up to AMoRe. Finally, and perhaps 2 milliradians (~0.1°) is usually captured horizontally. For an undulator, the constructive interference that most importantly, whereas bedefines undulator radiation constrains the X-ray beam both horizontally and vertically to a very narrow fore 1976 only natural products cone of approximately 1/(g √N) where N is the number of undulator periods7. In practice, horizontal diveravailable in adequate abungence is somewhat greater than vertical divergence, and measured values from APS undulator A are found to be 42 and 12 microradians (~0.0007°), respectively. (b) Photograph of an APS undulator A bedance were suitable for crystalfore placement into an accelerator straight section. This 2.4-m-long device is a linear array of 72 pairs of lography – afterwards, technolalternately oriented permanent magnets spaced with a period of 3.3 cm. The particle beam passes beogy developed for producing tween the magnetic poles through a gap that can be reduced to 8.5 mm. The magnetic field ranges from arbitrary molecules at high ~0.3 Tesla at 20-mm gap to ~1.0 Tesla at 8-mm gap. (c) Photograph of the synchrotron storage ring at the yield, irrespective of natural APS. This accelerator ring has a circumference of 1.1 km. It maintains electron (or positron) beams at a abundance. Recombinant DNA current of 100 mA and an energy of 7 GeV. It is divided into 40 quasi-symmetric segments, and each of technology made this possible 35 of these segments can provide X-ray beams from insertion devices in its straight section as well as from its bending magnet. Photographs in (b) and (c) were supplied by Rick Fenner of the APS. for proteins, initially in Escherichia coli and, subsequently, in diverse animal cells cultures, synchrotron radiation through construc- demands on detector characteristics. and chemical synthesis of oligonu- tive interference7. The electron beam is As exposure times were reduced from cleotides rejuvenated the study of DNA deflected back and forth in a slalom-like hours to minutes to seconds, improved and ultimately opened up opportunities path as it passes through a small gap be- detector speed was needed for efficient for DNA–protein complexes. tween poles of the magnetic array. operation. The first kinds of diffraction Most advances in the technological Synchrotron radiation from these slight area detectors, multiwire proportional underpinnings for crystallography since excursions is directed entirely into the chambers and phosphor imaging plates, 1990 relate quite directly to synchrotron forward direction along the line of the are limited, respectively, by count rate radiation. Six prime developments have straight section, and X-ray waves emitted performance and read-out time. CCD deevolved into maturity, and they com- from successive excursions interfere con- tectors are fast by virtue of both elecpletely change the scene for rapid struc- structively as in X-ray diffraction, which tronic recording and read-out of images, ture determination. These are undula- greatly amplifies the beam intensity. The and they are now standard equipment on tors, charge-coupled device (CCD) spectral positions (wavelengths) of the synchrotron beamlines8. Detector readdetectors, cryopreservation, MAD phas- resulting energy harmonics is controlled out performance remains limiting at unduing, selenomethionyl proteins and struc- by the magnetic field strength, which can lator beamlines. The development of pixel ture-solving automation. be adjusted by changing the undulator array detectors, whereby each picture Foremost among these is the undula- gap. X-ray beams from undulators are element carries its own read-out circuitry, tor, made possible by the third-generation laser-like and exceptionally brilliant. is expected to address this problem. synchrotron facilities. An undulator is a Brilliance is characterized by a high flux Another critical advance has come magnetic array of approximately 100 al- of highly parallel X-rays focused into a with techniques for cryopreservation of ternating magnetic poles that is inserted small cross-sectional area. macromolecular crystals through rapid into a straight section of a high-energy The strength of X-ray beams from freezing. Crystal freezing to liquid nitrosynchrotron ring to enhance the output of synchrotron sources placed increasing gen temperatures has earlier origins, but

639

REVIEWS

TIBS 25 – DECEMBER 2000

MAD phased structures

2000

Year of onset

1998 1996 1994 1992 1990 1988 Synchrotron beamline Ti BS

Figure 2 Evolution of synchrotron beamlines producing multiwavelength anomalous diffraction (MAD)phased structures. Each bar corresponds to an individual beamline, and the lower boundary of each bar establishes the year of onset for publications of MAD-phased structures by that beamline. Each segment within a bar corresponds to the number of novel atomic-level structures published in a given year from MAD data measured at that beamline. The smallest unit is one and the largest number from one beamline in one year is 23. The uppermost segment in a bar corresponds to publications from 1999 when colored pink. Criteria for entry are those used for inclusion in the annual Macromolecular Structures compendium (Hendrickson, W.A. and Wüthrich, K. eds, Current Biology Publications). It is noteworthy that the number of newly productive beamlines increases with time in recent years and that those beamlines that ultimately become highly productive usually require a few years to reach maturity.

more recent developments9,10 have led to widespread adoption of this simple but potent technology11. The resistance to radiation damage that freezing affords is important with conventional X-ray sources, but it becomes a crucial enabling factor in synchrotron studies of large biological assemblages (which are highly sensitive to radiation), small crystals (which must endure large specific doses), and to facilitate MAD phasing experiments (which are optimal when all data can be measured from one single crystal). MAD analysis of diffraction measurements made at multiple wavelengths provides a highly advantageous solution to the phase problem in X-ray crystallography12. The continuously tunable character of synchrotron radiation makes it possible to exploit sharp changes in scattering that come from resonance as X-ray energies approach electronic transition energies of elements in a crystal. Appropriate anomalous scatterers include elements that occur naturally in proteins (e.g. Fe, Zn, Mo), elements used in conventional heavy-atom derivatives (e.g. Pt, Hg, U), elements that are replacements for biological ions (e.g.

640

Lu31 for Mg21), and elements that can be incorporated systematically into macromolecules (e.g. Se, Br). MAD phasing was introduced in the mid-1980s, but it has caught fire only in the past few years13. This has come after the example of a few facilities was adopted elsewhere and appropriate synchrotron beamlines proliferated (Fig. 2). The growth of MAD phasing has been dramatic; nearly as many MAD-phased structures were published in 1999 alone as in all previous years and the increase over 1998 was 135%. A particularly important contributor to the acceptance of MAD phasing has been provided by biology itself. Many cell cultures, E. coli in particular, are able to incorporate selenomethionine systematically into proteins in place of methionine14,15. The Se sites in selenomethionyl proteins then prove to be highly effective as phasing centers for MAD experiments. The abundance of these sites is such as to give adequate phasing signals, while being sparse enough that the Se sub-structure can be determined. Moreover, as evidenced by the viability of 100% selenomethionyl bacteria, selenomethionyl proteins are

essentially isostructural and isofunctional with their native counterparts. MAD phasing based on selenomethionine has been highly successful. Selenomethionyl proteins account for approximately 65% of all MAD-phased structures. Indeed, this success has motivated methods to extend its application to mega-site problems, larger proteins where there are dozens of Se sites16. Finally, procedures for structure determination from diffraction data are becoming increasingly automated and effective. The quality of information made possible by the data from synchrotron experiments at modern facilities with current analytical methods is such that these diffraction patterns can often be converted into structural models with a minimum of human intervention. A rich repertoire of software facilitates these computations. Among the more recent additions are procedures for solving mega-site Se sub-structures such as SnB (Ref. 17), comprehensively general programs for phase determination such as SOLVE (Ref. 18) and SHARP (Ref. 19), and the ARP/wARP programs20 for automated chain tracing and model building into the resulting electron-density maps.

Recent triumphs Synchrotron crystallography dominates in current structural biology. Data from synchrotrons contribute to the majority of structure publications, and the fraction based on synchrotron experiments is .80% for publications in the most prominent of journals, where exciting structural pictures grace the covers of many issues. Similarly, the fraction of PDB depositions of novel crystal structures that report the use of synchrotron radiation has risen sharply and now exceeds 80% (Ref. 21). A sampling of recent structures22–30 is displayed in Fig. 3 to give an impression of the range and character of present-day synchrotron crystallography. Although these particular structures are among the most significant, many others are comparably complex and rich in biological impact. These remarkable triumphs include viruses and viral components; ribosomes and nucleosomes; RNA structures; membrane pumps, channels and receptors; signaling molecules; cellular transport machinery; transcription and replication factors and their complexes with DNA; and a multitude of fundamental enzymes. In addition to their role in providing structural pictures of key components of biological processes, new synchrotron capabilities are opening completely new

REVIEWS

TIBS 25 – DECEMBER 2000

Figure 3 (a) (b) (c) Collection of some recent structural triumphs. (a) Bacterial 50S ribosomal subunit22. The RNA portion is rendered in a gray pseudo-spacefilling model and the protein subunits are drawn as gold backbone worms. An inhibitor shown in red marks the active site for peptidyl transfer. (b) Bovine rhodopsin23. This first structure of a G-protein-coupled receptor is shown as a ribbon drawing for helices (in various colors) and as a pink worm for the b strands, loops N (e) (f) and extensions. The retinal chromo- (d) N phore is in yellow and carbohydrate units are in blue. The molecule is C C viewed from within the plane of the membrane with the cytoplasmic side above the seven transmembrane helices. (c) Reovirus core particle24. This 700-Å-diameter viral particle consists of an inner l1 shell that packages the RNA genome, a C C stabilizing outer s2 shell, and turretlike l2 structures (blue) through N N which nascent mRNA transcripts emerge and are modified with (g) (h) (i) methylated guanosine caps. The protein subunits are represented by Ca traces: l1, 120 copies in red; s2, 60 copies each in green and white and 30 copies in yellow; and l2, 60 copies in blue around the fivefold axes. (d) Human HMG-CoA reductase25. This ribbon diagram shows each of the subunits of the tetrameric molecule in a different color. Catalytic sites are formed at the interfaces between pairs of subunits. Statin inhibitors of this enzyme are Ti BS effective in lowering serum cholesterol levels. Reproduced, with permission, from Ref. 25. (e) Yeast RNA polymerase II (Ref. 26). Backbone models are shown for the ten protein subunits in the crystal, each in a different color. Zinc ions are in green and the active-site magnesium ion is in pink. (f) Yeast ATP synthase complex27. The contours for one molecular complex have been isolated from an electron-density map of the crystal at 3.9-Å resolution. Ca models for the a, b and g chains of F1 are superimposed above, additional middle density corresponds to the d and e subunits of F1, and the bottom portion corresponds to ten c subunits of the transmembrane FO portion. Reproduced, with permission, from Ref. 27 © 1999 American Association for the Advancement of Science. (g) Nucleosome core particle28. A 146-bp fragment of DNA is wound around a histone octamer. The backbone model for one DNA strand is turquoise and the other is brown. Ribbon models of the histone subunits, two of each, are colored yellow for H2A, red for H2B, blue for H3 and green for H4. (h) Mismatch-repair protein MutS (Ref. 29). The ribbon diagram for one protein subunit is green and the other is blue; DNA, sharply kinked at the base mismatch, is drawn as a space-filling model in red and pink. (i) HIV envelope glycoprotein gp120 (Ref. 30). The structure is a complex of the core portion of gp120 with binding portions of the cellular receptor CD4 and a neutralizing antibody that marks the chemokine co-receptor binding site. Core gp120 is shown as an all-atom model in red, the D1D2 fragment of CD4 is shown as a ribbon diagram in yellow, and the antibody Fab fragment is shown as a fuzzy molecular envelope in blue. Figures (c) and (g)–(i) were reproduced, with permission, from Refs 24 and 28–30, respectively, © Macmillan Magazines Limited. Figures were kindly provided by authors.

opportunities. Time-resolved crystallography, usually exploiting the continuous spectrum and special time structure of synchrotron radiation, is coming into its own to provide insight into reaction mechanisms31. Using short wavelengths, structures are being determined at true atomic resolution where hydrogen positions and bonding electrons are seen32. Taking advantage of brightness, structures can be determined from crystals with dimensions under 10 mm (Ref. 33). Exploiting beam fluxes, high-throughput methods are being implemented.

Electron-density maps from MAD data collected at undulator sources in less than 1 hour are sufficiently accurate for automatic chain tracing34. Batteries of mutational variants provide mechanistic insights. Series of complexes between pharmaceutical targets and inhibitory compounds are being used in structurebased drug development.

Prospects and challenges The future of synchrotron crystallography is bright. What seems to be in store goes well beyond expansion con-

tinuing at its recent pace. Most currentday structural studies do not take full advantage of the technical advances described above. In particular, the first dedicated undulator beamlines have become operational only recently. Capacity will increase dramatically as these and other new beamlines under construction or in planning are brought online. Other technical developments, most likely including ones not now anticipated, will also appear. On the application side, spurred by the new capabilities, completely fresh approaches are

641

REVIEWS being initiated; programs in structural genomics are under way, the emphasis on macromolecular assemblages and molecular machines is ever increasing, and structure determination in support of pharmaceutical development is moving to an unprecedented scale. X-ray technology is expanding both quantitatively and qualitatively. New beamlines are in development, many at undulator sources, and a new wave of lower-energy synchrotron facilities (SLS in Switzerland, Soleil in France, Diamond in England, CLS in Canada; see Box 1) are being built with low-emittance particle beams. These can accommodate narrow-gap undulators that function into the hard X-ray regime needed for biological crystallography. Alternatively, as at the ALS (Box 1) superconducting dipole magnets (superbends) can be introduced to produce hard X-rays. For all of these, biology is now a driving force. For the more distant future, free electron lasers are being developed that promise to create coherent X-ray beams orders of magnitude more brilliant than those from undulators, and prospective experiments are being modeled35. Effective utilization of X-ray beams from modern synchrotrons is a challenge. To harness the power of these beams, appropriate high heat-load optics are needed36. To handle diffraction acquisition rates, pixel array detectors are being developed that will reduce readout times by more than an order of magnitude, while improving other performance characteristics. To utilize powerful beams with greater efficiency, robotic devices and software automation are being developed for sample changing, crystal alignment and data collection. Adequate computing power is available, but practical implementation for handling the enormous data acquisition rates from large format detectors is ever daunting. Radiation damage remains as a nemesis, and efforts are in progress to understand and cope with this reality. Synchrotron capabilities provide for exciting opportunities in structural biology. The exploitation of undulator brilliance in crystallography is only in its infancy. Manifold benefits come from the laser-like properties of undulator radiation: high-energy resolution can enhance MAD phasing signals, low divergence can permit ready measurement of diffraction from large unit cells, fine focus can make microcrystals useful, and abundant flux can support high-

642

TIBS 25 – DECEMBER 2000 throughput crystallography. Wigglers and bending magnets also produce very potent X-ray beams when configured with appropriate instrumentation. The technological advances now being implemented and in prospect offer a strong basis for structure determination efforts that can outstrip current rates. Structural genomics projects aim to determine structures on a pan-genomic scale by which representative structures can be generated for a substantial fraction of the sequence families in all living organisms. This obviously requires the production of crystallizable molecules at high throughput37, but measurements of diffraction and structure analysis must also be highly expeditious as they now can be34. In addition to the reservoir of background information expected to come from a pan-genomic picture of atomic structure, speed in structure determination also permits structure to drive biological discovery through early hypotheses for tests of cellular activity as well as biochemical mechanism, as in the Tubby example38. Speed might not be limiting in applications to molecular machines and other macromolecular assemblages. Such studies also require strong ties to the corresponding biology and will benefit greatly from improved means for production of samples, notably in the case of recombinant membrane proteins or systems of obligatorily cosynthetic partners. Nevertheless, even when finally coaxed to crystallize, these forefront problems in structural biology also present challenging diffraction problems. Such complicated systems typically produce weakly diffracting samples that are highly sensitive to radiation damage, the crystals are often quite small and unit cells are usually quite large. Enhanced crystallographic technology is essential for the analysis of ever-larger and more complex systems at increasingly refined technical limits. The results can be glorious, as in the recent atomic-level structures of ribosomal subunits22,39,40. The high-resolution data for each of these spellbinding structures were measured at undulator beamlines. In sum, we can expect in the coming years to see synchrotron crystallography pursued with both speed and finesse. These structural trends in biochemical sciences will advance cell and molecular biology and the results will also have important repercussions for biotechnology and medicine.

Acknowledgements I thank the many colleagues from my own laboratory who have participated with me in synchrotron experiments and those at synchrotron facilities around the world who developed the instruments that we have used. The contributions of Craig Ogata at the National Synchrotron Light Source, Kevin D’Amico at the Advanced Photon Source and Paul Phizackerley at the Stanford Synchrotron Radiation Laboratory have especially impacted my own research. This work has also benefited from NIH support through grant GM34102, from HHMI support for the X4 beamlines at Brookhaven, and from DOE and NSF support of the US synchrotron facilities.

References 1 Matthews, B.W. (1976) X-ray crystallographic studies of proteins. Annu. Rev. Phys. Chem. 27, 493–523 2 Bernal, J.D. and Crowfoot, D. (1934) X-ray photographs of crystalline pepsin. Nature 133, 794–795 3 Blundell, T.L. and Johnson L.N. (1976) Protein Crystallography, Academic Press 4 Parratt, L.G. (1959) Use of synchrotron orbit-radiation in x-ray physics. Rev. Sci. Instrum. 30, 297–299 5 Rosenbaum, G. et al. (1971) Synchrotron radiation as a source for X-ray diffraction. Nature 230, 434–437 6 Phillips, J.C. et al. (1976) Applications of synchrotron radiation to protein crystallography: preliminary results. Proc. Natl. Acad. Sci. U. S. A. 73, 128–132 7 Als-Nielsen, J. and McMorrow, D. (2000) Elements of Modern X-ray Physics, Wiley 8 Gruner, S.M. and Ealick, S.E. (1995) Charge coupled device X-ray detectors for macromolecular crystallography. Structure 3, 13–15 9 Hope, H. (1988) Cryocrystallography of biological macromolecules: a generally applicable method. Acta Crystallogr. B44, 22–26 10 Teng, T.Y. (1990) Mounting of crystals for macromolecular crystallography in a free-standing thin film. J. Appl. Crystallogr. 23, 387–391 11 Rodgers, D.W. (1994) Cryocrystallography. Structure 2, 1135–1140 12 Hendrickson, W.A. (1991) Determination of macromolecular structures from anomalous diffraction of synchrotron radiation. Science 254, 51–58 13 Hendrickson, W.A. (1999) Maturation of MAD phasing for the determination of macromolecular structures. J. Synchrotron Rad. 6, 845–851 14 Hendrickson, W.A. et al. (1990) Selenomethionyl proteins produced for analysis by multiwavelength anomalous diffraction (MAD): a vehicle for direct determination of three-dimensional structure. EMBO J. 9, 1665–1672 15 Bellizzi, J.J., III et al. (1999) Producing selenomethionine-labeled proteins with a baculovirus expression vector system. Structure 7, R263–R267 16 Deacon, A.M. and Ealick, S.E. (1999) Selenium-based MAD phasing: setting the sites on larger structures. Structure 7, R161–R166 17 Smith, G.D. et al. (1998) The use of SNB to determine an anomalous scattering substructure. Acta Crystallogr. D54, 799–804 18 Terwilliger, T.C. and Berendzen, J. (1999) Automated structure solution for MAD and MIR. Acta Crystallogr. D55, 849–861 19 de la Fortelle, E. and Bricogne, G. (1997) Maximumlikelihood heavy-atom parameter refinement for multiple isomorphous replacement and multiwavelength anomalous diffraction methods. Methods Enzymol. 277, 472–494 20 Perrakis, A. et al. (1999). Automated protein model building combined with structure refinement. Nat. Struct. Biol. 6, 458–463 21 Minor, W. et al. (2000) Strategies for macromolecular synchrotron crystallography. Structure 8, R105–R110 22 Ban, N. et al. (2000) The complete atomic structure of the large ribosomal subunit at 2.4 Å resolution. Science 289, 905–920 23 Palczewski, K. et al. (2000) Crystal structure of rhodopsin: a G protein-coupled receptor. Science 289, 739–745

J0URNAL CLUB

TIBS 25 – DECEMBER 2000 24 Reinisch, K.M. et al. (2000) Structure of the reovirus core at 3.6Å resolution. Nature 404, 960–967 25 Istvan, E.S. et al. (2000) Crystal structure of the catalytic portion of human HMG-CoA reductase: insights into regulation of activity and catalysis. EMBO J. 19, 819–830 26 Cramer, P . et al. (2000) Architecture of RNA polymerase II and implications for the transcription mechanism. Science 288, 640–649 27 Stock, D. et al. (1999) Molecular architecture of the rotary motor in ATP synthase. Science 286, 1700–1705 28 Luger, K., et al. (1997) Crystal structure of the nucleosome core particle at 2.8Å resolution. Nature 389, 251–260 29 Obmolova, G. et al. (2000) Crystal structures of

Reducing (oxidative) stress: structure and mechanism Oxidative damage to biological macromolecules has been linked to aging, cancer and Alzheimer’s disease, and is suggested to play a role in numerous other pathophysiological states. The amino acid methionine, either in isolation or within proteins, can act as an oxidant scavenger by undergoing oxidation to methionine sulfoxide. Subsequent reduction of the sulfoxide is achieved by the regulatory enzyme peptide methionine sulfoxide reductase (MsrA), which has been shown to be in the minimal required gene set. Recent studies have determined high-resolution structures for two conformations of the bovine enzyme1 and proposed a mechanism for catalysis of the homologous Escherichia coli enzyme2. Previous results demonstrated the importance of a universally conserved cysteine residue in the enzyme active site and suggested that two other cysteine residues act in a concerted thiol-disulfide exchange mechanism. Crystal structures of two different bovine MsrA truncations were determined to 1.6- and 1.7-Å resolution by Lowther and colleagues1. The observed structure was of the mixed a/b type, with a core containing a twolayer sandwich, a-b plaits motif. The first form (residues 21–219; full-length bMsrA comprises 233 residues) crystallized in the presence of DTT, the in vitro source of enzyme-reducing thiols contributed in vivo by thioredoxin. In this structure, DTT was linked by disulfide bonds to two of the cysteine residues expected to be catalytically active. The second form of bMsrA was extended in the C terminus by ten residues, and this extension consisted of a glycine-rich tail that may bring the third cysteine residue near the active site. All the cysteines proposed to be catalytically active are surface residues. In complementary work,

30

31 32 33

34

mismatch repair protein MutS and its complex with a substrate DNA. Nature 407, 703–710 Kwong, P. D. et al. (1998) Structure of an HIV gp120 envelope glycoprotein in complex with the CD4 receptor and a neutralizing human antibody. Nature 393, 648–659 Ren, Z. et al. (1999) Laue crystallography: coming of age. J. Synchrotron Rad. 6, 891–917 Dauter, Z. et al. (1997) The benefits of atomic resolution. Curr. Opin. Struct. Biol. 7, 681–688 Pebay-Peyroula, E. et al. (1997) X-ray structure of bacteriorhodopsin at 2.5 Ångstroms from microcrystals grown in lipidic cubic phases. Science 277, 1676–1681 Walsh, M.A. et al. (1999) Taking MAD to the extreme: ultra fast protein crystal structure determination. Acta Crystallogr. D55, 1168–1173

Boschi-Muller and co-workers characterized mutants of E. coli MsrA with one or more cysteine-to-serine substitutions2. The quantitation of free thiols at different steps in the reduction pathway led to the determination that enzyme-catalysed reduction proceeds through a sulfenic acid intermediate. This sulfenic acid is reduced through thioldisulfide exchange carried out by the remaining cysteine residues. In addition to the protective functions of MsrA, the protein has been shown to be important for infection of cells by pathogenic bacteria. These invaders rely on adhesins to bind host tissues, and bacterial MsrA mutants show decreased binding to endothelial cell receptors, indicating that MsrA is required for proper adhesin presentation. Therefore, a more thorough understanding of the structure and mechanism of MsrA could have implications both for limiting damage due to oxidative stress and for development of novel antibacterial therapeutics. 1 Lowther, W.T. et al. (2000) Structure and mechanism of peptide methionine sulfoxide reductase, an ‘antioxidation’ enzyme. Biochemistry 39, 13307–13312 2 Boschi-Muller, S. et al. (2000) A sulfenic acid enzyme intermediate is involved in the catalytic mechanism of peptide methionine sulfoxide reductase from Escherichia coli. J. Biol. Chem. 275, 35908–35913

REBECCA W. ALEXANDER Email: [email protected]

Extracellular iron: a signal that can be sensed and transduced in bacteria Iron is a vital biocatalyst required for numerous metabolic functions. However, iron can be extremely toxic when present in excess because it promotes the formation of highly reactive oxygen species that are able to damage biological macromolecules. Cells have evolved

35 Neutze, N. et al. (2000) Potential for biomolecular imaging with femtosecond X-ray pulses. Nature 406, 752–757 36 Freund, A.K. (1996) Third-generation synchrotron radiation X-ray optics. Structure 4, 121–125 37 Stevens, R.C. (2000) Design of high-throughput methods of protein production for structural biology. Structure 8, R177–R185 38 Boggon, T.J. et al. (2000) Implication of Tubby proteins as transcription factors by structure-based functional analysis. Science 286, 2119–2125 39 Schluenzen, F. et al. (2000) Structure of functionally activated small ribosomal subunit at 3.3Å resolution. Cell 102, 615–623 40 Wimberly, B.T. et al. (2000) Structure of the 30S ribosomal subunit. Nature 407, 327–339

several regulatory mechanisms to ensure sufficient iron supply and to manage the toxicity of reactive iron. These mechanisms rely on cytosolic factors that sense intracellular iron and adapt the expression of proteins involved in iron uptake, storage and utilization. This is the case with the bacterial protein Fur (ferric uptake regulator) that, when associated with Fe21, represses the transcription of genes involved in iron acquisition (and activates genes involved in the defense against oxidative insult), allowing the cell to cope with high-iron conditions. In the September issue of Cell, Wösten et al.1 describe an alternative mechanism of adaptation to high-iron environment based on the PmrA–PmrB system in Salmonella enterica. In such a twocomponent system, PmrB is the sensor protein that is autophosphorylated and then phosphorylates and activates the transcriptional regulator PmrA. The signal sensed by PmrB was unknown. Analysis of the periplasmic domain of PmrB revealed two conserved ExxE motifs found in two iron permeases, an iron transporter and a chelator, suggesting that the PmrA–PmrB system could respond to iron. Indeed, the authors have shown that the PmrA regulon was activated specifically by extracytoplasmic Fe31 (and to a lesser extent by Al31) in a PmrB-dependent manner. Mutagenesis of the periplasmic domain of PmrB confirmed the crucial role of the ExxE motif in the response to extracellular Fe31. The responsiveness of the PmrA–PmrB system suggested that it could be implicated in resistance to high-iron conditions that enteric bacteria can encounter in the gastrointestinal tract of animal hosts. This hypothesis was confirmed by the inability of pmrA and pmrB mutants to grow in high iron. However, in contrast to the Fur regulator that modulates the level of cytoplasmic iron and the antioxidant defenses of the cell, the PmrA–PmrB system did not seem to mediate resistance to the pro-oxidant properties of iron. Indeed, the regulation of the Fur protein appeared to be normal in a pmrA mutant. Furthermore, the same mutant exhibited normal sensitivity to oxidative insult, whereas an Escherichia

643