Synchrotron radiation is used for a number of non-crystallographic applications. X-ray absorption spectroscopy gives detailed characterization of metal centers in biological macromolecules. Small-angle X-ray scattering is used to probe the dynamic and static size and shape of macromolecules in solution. X-ray microscopy is showing promise as an imaging tool. These biophysical methods may be applied to macromolecules in solution and do not require growth of single crystals to produce structural data. They account for a substantially smaller proportion of synchrotron beam time at present than does crystallography (Table A2). However, they are frequently applied to samples not amenable to crystallization and to properties inaccessible to the crystalline state. The complementary information they provide can be key to full understanding of biomolecular functions. This section describes the current status and future prospects of X-ray absorption spectroscopy, small-angle X-ray scattering, and X-ray microscopy.
X-ray Absorption Spectroscopy
Current status. X-ray absorption spectroscopy (XAS) is the study of spectral details on the high-energy side of an X-ray resonant frequency or absorption edge of a metal atom. X-ray absorption spectra are frequently divided into the regions of X-ray absorption near edge structure (XANES) within ~50 eV of the edge, and extended X-ray absorption fine structure (EXAFS) from ~50 eV to ~1000 eV above the edge. The XANES region contains information about the oxidation state and local geometry of the absorbing metal atom. The EXAFS region provides structural information about the absorbing atom, and can be analyzed to give bond lengths to ±0.02 Å, coordination numbers to ±1, and the chemical identity of ligands to within one row of the periodic table, e.g. distinguishing N from Cl but not from O.
One of the principal attractions of XAS for structural biology is to provide element-specific structural information about macromolecules in solution. Many samples of interest, such as reaction intermediates, are not amenable to study as single crystals or are difficult to crystallize. A second important application of XAS is as a powerful complement to crystallography in the study of metalloproteins. Metal-ligand distances available from XAS are typically an order of magnitude more accurate than those from macromolecular crystallography. Thus XAS data can be essential for defining the details of an active site and can aid in the refinement of a crystallographic model. For example, connectivity within the unusual Mo/Fe/S cluster in nitrogenase was obtained with crystallography, from which starting point EXAFS provided the complementary high-resolution structural information required to understand the mechanism. In addition, EXAFS data for a variety of metalloproteins, e.g. rubredoxin, A. vinelandii ferredoxin and lipoxygenase, have helped to correct initial errors in metal site structures obtained from crystallography. XANES measurements can often provide a direct and unambiguous determination of metal-ion oxidation state, e.g. the presence of Cu(II) and not Cu(III) in galactose oxidase. This information is difficult to extract with other spectroscopic methods and impossible to obtain from crystallography.
Presently two beamlines at SSRL (7-3 and 9-3) and one at NSLS (X-9) are semi-dedicated to biological XAS. One biological XAS line (BioCAT, split with SAXS) is under construction at APS. In addition, perhaps 6 other US beamlines are used with some frequency for biological XAS studies. Finally, several special-purpose beamlines (circularly polarized X-rays, low-energy X-rays, etc.) are currently being designed or constructed and will be used in part for biological studies.
Approximately 20-30 US research groups make frequent use of XAS for structural biology and perhaps 20-30 other groups use XAS occasionally for studies of biological samples. A large fraction of XAS studies are collaborative ventures, involving a group with X-ray expertise interacting with a group having biochemical expertise. Biological XAS is thus used, at least occasionally, by 100-200 research groups in bioinorganic and biophysical chemistry.
Emerging Trends. As the number of crystallized proteins has increased, so too has the number of uncrystallized proteins that have been isolated and purified. XAS will remain an important tool for characterizing the metal sites in these proteins. Realistic projections of the capabilities of the next generation of beamlines (APS, SPEAR-3) and of future detectors (see below) should allow XAS to be extended both to lower concentrations and to smaller sample volumes (see section on X-ray microscopy). These will allow XAS to be applied to samples that are either too dilute or too small for present technology, and thus will keep the demand for XAS beam time at least as high as present levels of two- to threefold oversubscription.
One particularly important growth area may be the use of polarized XAS to provide angle-resolved bond-length information to complement diffraction studies. For example, with the greatly enhanced crystal lifetimes that come with cryogenic experiments, a single crystal can be used for both crystallography and XAS. The three-dimensional structure of a metalloprotein can be determined by crystallography, the metal-ligand bond lengths to much greater precision by EXAFS, and the metal-ion oxidation state by XANES, all from one sample crystal. Improved X-ray fluorescence detectors on MAD beamlines are the only instrumental requirement for these hybrid experiments, because XANES is part of every MAD experiment. Polarized XANES data from single crystals could provide orientation-dependent anomalous scattering factors f' and f'' for use in metal-based MAD phasing. The combination of XAS with crystallography would provide a better description of the molecular structure than is possible from either technique alone. Such applications would increase significantly the demand for XAS beam time.
XAS is limited by its relatively low information content. In the 1991 BioSync report several new developments were identified, which had the potential to enhance XAS by providing more information: Circularly polarized XAS, soft X-ray XAS, energy-resolved X-ray fluorescence, and time resolved XAS measurements. Over the last 6 years, all of these have been demonstrated for biological samples. Circularly polarized XAS measurements distinguish different spin-states of a sample. Soft X-ray measurements provide direct access to electronic-structure information that can only be inferred from hard X-ray XAS measurements. High-resolution energy-resolved X-ray fluorescence offers the possibility of site-selective XAS, thus obviating the limitation that XAS is a bulk measurement giving only average structural information. Temporal resolution makes possible the use of XAS to determine the oxidation state and perhaps the molecular structure of transient reaction intermediates. Stopped-flow time-resolved XAS in combination with principal component analysis has been used in studies of supported catalysts to identify the reactive intermediate in a complex mixture. This approach should be readily applicable to metalloprotein studies.
Recommendations. The 1991 BioSync report noted that the most pressing limitation to biological XAS was the lack of suitable X-ray fluorescence detectors. Although the available detector technology has improved since then, this conclusion remains true. It is routine on many beamlines for users to decrease the available flux in order to avoid detector saturation. This problem will be even greater at the third-generation synchrotron radiation sources. Piecemeal efforts to develop new detectors have been made at the different synchrotron sources and these are beginning to show results. Nevertheless, the XAS community would benefit tremendously from a concerted effort to develop optimized detectors which could fully utilize the available X-ray flux. Optimized detectors would result in an immediate increase in sample throughput by a factor of two or more. This would make a substantial impact on the current beam time shortfall, and would facilitate the projected expansion in XAS applications.
In addition to optimizing present detectors, it is important to explore other detector technologies. Promising recent developments include superconducting tunnel junctions, microcalorimeters, and "transition edge" detectors. These offer the promise of both very high energy resolution (2-10 eV) and high count rates (20 kHz). Arrays of such detectors could revolutionize the practice of XAS. It is not clear at this point which of these detector technologies will be the most useful. It is clear that the XAS community needs to invest more effort in the development of the next generation of detectors.
The biological support facilities available for XAS have improved significantly since 1991, to the point that these are typically not a limiting feature in experimental progress. The availability of beam time remains a limiting factor for many studies. Based on the current over-subscription rates, a two-fold increase in the available beamlines would be completely utilized, even without the increase in demand that is anticipated. Although no increase in the number of beamlines is likely in the short term, it should be possible to increase the available XAS time through improved detectors, which will effectively increase throughput, and perhaps eventually through development of new beamlines. Any decrease in the operational time available at either SSRL or NSLS would have a serious and immediate negative impact on the field.
Many of the new XAS techniques (circular dichroism, soft X-ray, time-resolved) require unique, dedicated beamlines. These are presently under construction at ALS and APS. For the next few years, it appears that the availability of experimental time for these experiments should be sufficient to permit exploration of the range of applications of these methods in structural biology. If one of these methods proves to be widely applicable, additional beamline construction may be indicated 3-5 years from now.
In summary, the most pressing need for XAS is improved detectors. These would help alleviate the present shortfall in beam time. This is, however, a double edged sword, since improved detectors will make possible the application of XAS to smaller and more dilute samples, thus increasing the demand for beam time. In the 3-5 year time frame, some increase in XAS beamlines is likely to be necessary.
Small Angle X-ray Scattering
Current Status. X-ray scattering can be used to characterize solutes in a variety of solutions and can be performed over a range of scattering vectors. The most useful region for structural biology is small angle X-ray scattering (SAXS). SAXS is unique in providing information on macromolecular shape, e.g. radius of gyration, Rg, with high temporal resolution, although atomic-resolution structures cannot be retrieved. The dynamics of many important biological processes, such as protein folding, are studied by SAXS. For example, several distinct folding intermediates during renaturation of apo-myoglobin and lysozyme have been detected recently. A detailed description of protein folding is important for understanding prion virulence, e.g. bovine spongiform encephalopathy or "mad cow disease". Time-resolved SAXS is one of the few techniques that can provide direct information about the rate and mechanism of protein compaction during folding. Small-angle scattering is a powerful method to study molecular interactions and conformational flexibility in solution, which are key to understanding molecular communication in biological systems. Solution scattering in conjunction with high resolution structural data can provide invaluable insights into the interaction of individual components in molecular assemblies and complexes. This can be done in the absence of crystals and for the widest range of molecular weights and dimensions (~10 - 1000 Å), thus providing an important adjunct to structural data from NMR and crystallography. SAXS has been particularly useful in contributing to our understanding of biochemical regulation by providing insights into domain reorientation and protein-protein interactions important in signaling. Advances in molecular biology techniques combined with brilliant synchrotron X-ray sources have permitted studies of the interactions of regulatory proteins with active enzymes whose solubilities and stabilities are limited.
Current facilities in the US include beamline 4-2 at SSRL (semi-dedicated to SAXS) , and NSLS beamlines X9B and X12B. Beam time available for SAXS at NSLS is decreasing due to increased crystallographic demand. The BioCAT sector under development at the APS will include a significant SAXS resource. These beamlines are used for structural biology by approximately 20 US groups.
Emerging Trends. Historically, SAXS has been used by a small number of laboratories because fabrication of clean small-angle cameras is a delicate art that few have mastered; and SAXS measurements on dilute solutions require lengthy exposures, and consequently have low throughput. Synchrotron radiation has allowed SAXS to be carried out much more quickly and easily. There is a great opportunity for the development of well-equipped user facilities dedicated to SAXS. In parallel with this, we anticipate several important improvements in analytical and experimental techniques, listed below. These will improve both the ease of using SAXS and the scope of problems to which SAXS can be applied. The anticipated developments suggest that the potential for growth in the use of SAXS for biological samples is large.
Analytical improvements include:
Experimental improvements include:
Recommendations. Planning for synchrotron radiation resource allocation should allow for a significant growth in the use of SAXS. The technique is now poised for expansion into a much wider range of laboratories. Investment in user support, such as software development, scattering cell fabrication, hardware for kinetics, and so on, is crucial to this expansion since these improvements are necessary for the entry of new groups into the field. At the present time, improvements in user support are more important than the construction of new beamlines for SAXS.
X-ray Microscopy
It has long been recognized that the ability to visualize an object of interest is one of the cornerstones of advancement in science. For this reason, X-ray imaging holds special promise as a technique in structural biology. Much of the promise comes from the possibility to do spectroscopy with the imaging X-ray beam. The method is under development. The primary challenge for biological X-ray microscopy is to focus an intense enough X-ray beam on a small enough spot to image interesting biological objects, yet minimize the effects of radiation damage. A variety of technologies can be used to focus X-rays, including zone plates, Kumakov lenses, waveguides, refractive optics, capillary optics, grazing incidence mirrors, and normal incidence mirrors. At present, zone plates and grazing incidence mirrors are the technologies that are most highly developed in synchrotron X-ray microscopes.
The zone-plate approach provides superb spatial resolution -- as low as 30 nm in some cases. However, zone-plates have smaller apertures than are possible with mirrors and thus typically give lower flux. The lower flux, together with the fact that extremely precise alignment is required to scan energy with a zone-plate, has meant that zone-plate beamlines are most often used for fixed energy imaging, or possibly for dual wavelength differential imaging. In contrast, mirror-based beamlines have been used for a range of spectroscopic studies. Although the currently available X-ray fluxes have limited these to XANES studies, EXAFS and possibly SAXS measurements should be possible in the future. Most spectroscopic work has been done at a spatial resolution of 10-30 µm, although this has recently been extended to ~1 µm. The diffraction limit for Pt-coated Kirkpatrick-Baez mirrors is 40 nm at 7 keV (wavelength lambda = 0.2 nm). Thus it should be possible eventually to extend mirror-based beamlines to a resolution limit close to that of zone-plates. However, considerable improvements in mirror technology will be required to reach this limit. Zone-plate beamlines are sometimes referred to as X-ray microscopes while mirror-based beamlines are more often described as X-ray microprobes, with the names reflecting both the differences in spatial resolution and the tendency to use the former primarily for single wavelength imaging and the latter for energy-scanning spectroscopies. In terms of the experiments for which they can be used, however, there is considerable overlap between the different types of beamlines.
Current Status. Zone-plate beamlines are presently in use both as soft and hard X-ray sources. The soft X-ray sources, NSLS beamline X1A and ALS beamlines 7.0 and 6.1, typically operate in the 0.2-1.2 keV range (lambda = 7-1 nm) and produce spot sizes as small as 30 nm. This energy region is important as it includes the so-called water window, where the absorption due to oxygen is low, and thus where biological samples can be studied. The hard X-ray zone-plate beamline at APS operates at 8 keV (lambda = 0.2 nm) and produces a spot size of ca. 0.25 µm. Mirror-based beamlines are presently operated at NSLS (X26A), APS (GeoCARS), and ALS (10.3.1 and 10.3.2). These typically operate in the hard X-ray region (5-12 keV, lambda = 0.2-0.1 nm) and produce spot sizes in the range of 1 µm (APS and ALS) or 30 µm (NSLS). An additional microprobe beamline is presently being constructed at the ALS.
Most of the current applications utilize inorganic samples, addressing geological, semiconductor, and/or materials science problems. Extension to biological samples is straightforward, although some investment in sample preparation and handling will be required, since biological samples are generally less robust than inorganic samples. The technology for this is well established. Using an X-ray microprobe beamline, it should be possible to map the oxidation state and ligation of metal ions in biological samples at 1 µm resolution. This will permit studies of metal speciation in some large cells, and can be used for histochemical studies of metal ion distribution in tissues.
Currently the most prominent X-ray microscope in the US is the scanning transmission X-ray microscope (STXM) at the NSLS (beamline X1A). The operating parameters of the Brookhaven STXM and its experimental capabilities have both improved substantially in the last two years. In particular:
The Brookhaven STXM is usually operated in a wavelength window where oxygen has low absorption, so that water is transparent. Variation of lambda within this window can be used to give elemental contrast between C and N, or between different oxidation states of a single element. Contrast based on other elements (e.g. Ca2+) is also possible. A new version of the microscope that is much easier to operate is also ready. Existing facilities should permit a three-dimensional reconstruction of a single cell at ca. 50 nm resolution, with elemental contrast, so that protein-rich and nucleic acid-rich regions will show up differentially.
Emerging Trends. The existing X-ray microprobe beamline at NSLS is heavily oversubscribed. This situation is expected to ease somewhat as the new APS facility comes on line. However, with the higher flux and smaller spot size of the new beamline, new classes of experiments will become possible. At a 1 µm spot size, it is realistic to perform in situ XAS measurements on intact biological samples. This, together with the ability to image each of the elements in a sample at m resolution, is likely to attract a large group of new users. It is anticipated that this expansion in user base will rapidly lead to oversubscription at least as severe as that presently experienced at NSLS.
The recent dramatic improvements in the Brookhaven STXM have expanded significantly the range of samples that can be studied. This again suggests that demand from outside users will rise rapidly in the next few years. Many of these experiments, particularly those involving 3D data sets, will place heavy demands on beam time. At a flux of 106 photons/second, each pixel needs about 10 ms of exposure to give good statistics. For an object of typical size 5 µm several million pixels are required, which translates into several hours for a single 3D data set.
Several beamlines at the ALS provide facilities for X-ray microscopy that can be used for biological samples. Beamline 7.0.1 houses a scanning transmission and scanning photoelectron microscope and beamline 6.1 will provide a full-field transmission X-ray microscope with zone-plate lenses. The overall impact of the new ALS facilities will be to significantly increase the available time and available modalities for X-ray microscopy. If past developments in imaging are any guide, these new facilities will generate a significant increase in the users of and applications for X-ray microscopy in structural biology. Possible future developments include new techniques for dark-field and holographic imaging and new algorithms for "super-resolution", in which the diffraction pattern rather than the transmission is recorded for each pixel.
Recommendations. It appears likely that the future demand for access to both X-ray microprobe and X-ray microscopy beamlines will grow dramatically in the next 5 years. This is likely to lead to the need for additional beamlines dedicated to these experiments. In addition, the development of improved support facilities for microscopy will be important for facilitating the entry of new users into the field.
This Brookhaven STXM represents the current state of the art in X-ray microscopy. However other technologies, such as the imaging microscopy used at ALS beamline 6.1, may end up being the methods of choice for many problems. Given the recent developments in the field, substantial increase in demand for STXM access is anticipated. When the potential impact of several very exciting technical developments is considered, one should prepare for a hitherto unanticipated demand for beamline resources on the part of the X-ray microscopy community.