XDL2011 Workshop 2
Biomolecular Structure from Nanocrystals and Diffuse Scattering
Monday, June 13th - Tuesday, June 14th, 2011
Organizers: Ed Lattman (Hauptmann-Woodward Medical Research Inst.), Mavis Agbandje-McKenna (University of Florida), Keith Moffat (University of Chicago), & Sol Gruner (Cornell University) Workshop Agenda (html)
Workshop Poster (pdf)
Purpose: The purpose of the workshop is to assess the state-of-the-art in the use of hard x-ray nanobeams (1) determine biomolecular structures from nanocrystals, and (2) obtain bimolecular dynamical information from analysis of diffuse scattering by crystals, including nanocrystals. We are especially interested in exploring what might be most feasible with Energy Recovery Linac (ERL) or Ultimate Storage Ring (USR) x-ray sources. Description: (1) Protein crystallography is limited mainly by protein crystallization and crystal growth. The large-scale NIGMS Protein Structure Initiative centers (where failure as well as success is tracked) typically manage to crystallize less than 10% of the targets for which protein expression is achieved. Intensive effort - sometimes years of effort - on the part of laboratories committed to the study of a particular target raises this percentage somewhat, but often at the expense of altering the targets by, for example, proteolysis or mutagenesis. A number of initiatives are underway to deal with the problem of protein crystal growth. A more direct and imaginative approach is to carry out X ray diffraction (XRD) imaging experiments on single molecules using next generation X-ray sources, such as the Linac Coherent Light source (LCLS) being developed at SSRL, the Energy Recovery Linac (ERL) envisioned at CHESS, or the X ray Free-Electron Laser (XFEL) being developed at DESY. These sources are designed to produce beams of brilliant, coherent hard X ray pulses. In the single molecule approach, the data set is built of large number of Poisson-noise limited images. This XRD molecular imaging approach may allow structures to be determined without a crystal, but many technical difficulties must be resolved if this is to become feasible. An intermediate approach, which may be easier and accomplish the same goals, is to make use of nanocrystals. Microcrystalline precipitates occur frequently and can provide 3-d XRD data. Every protein crystallization experiment is actually a nonequilibrium experiment involving two sequential steps: First homogenous nucleation conditions must be created to induce nanocrystals to form, and then conditions must change to induce a few of the crystals to grow to a size suitable for x-ray analysis. Evidence is accumulating that the first step may well be much easier than the second. Eight years of experience of the Hauptmann-Woodward high-throughput crystallization laboratory (HTCS) are archived with images of more than 16 million microliter-scale crystallization trials comprising almost 11 thousand proteins. Visual analysis of the archived data show that precipitates obtained in apparently unsuccessful protein crystallization experiments are, more often than not, microcrystalline or contain microcrystalline components. Kim and Gillilan at CHESS have quantified this through a collaborative effort analyzing crystal size distributions for images of 68 of these proteins associated with the Northeast Structural Genomics Consortium. They studied 104,448 of the HTC laboratory images and found that tiny visible crystals occurred approximately 50% of the time. These optical observations have been complemented by XRD powder observations by Robert Von Dreele at the Advanced Photon Source. Protein powder patterns were recorded from HTCS laboratory protein crystallization precipitates in which it was not possible to distinguish any vertex-edge-face microcrystal morphology by light microscopy. These crystals are in the sub-micron to micron size range, and we refer to them as nanocrystals in the present context, these nanocrystals are common. Moreover, even apparently successful crystallization experiments that yield nice-looking crystals that prove to diffract poorly, more often than not, yield a quantity of nanocrystalline precipitate. May questions pertaining to structure from tiny crystals remain to be answered: How difficult is it to grow nanocrystals of membrane proteins and complexes? Are nanocrystals generally more perfect than larger crystals? What are the implications of the larger surface to volume ratio of tiny crystals? How does the time-dependent radiation damage component of room temperature crystals scale with dose rate? What are the best procedures to acquire and analyze nanocrystalline diffraction data? Etc. The premise of the workshop is that it is technologically feasible to acquire complete data sets from a large number of nanocrystals, and the challenges are simpler than acquiring complete data sets from single molecules. The workshop will examine this premise from the standpoint of what might be needed to use nanocrystals to determine protein structures. Workshop contributions: Presentations will focus on the issues of nanocrystallography, and on the advantages of ERL & USR, as well as other novel hard x-ray sources, for such experiments. (2) Diffuse scattering is important. Past about 2.5Å resolution more than half the photons scattered from a typical crystal do not end up in the Bragg peaks, but rather in the background. Non-Bragg scattering arises from electronic and atomic motion or disorder, and has major two components: scattering from the Debye-Waller model of independent atom motions produces isotropic and circularly symmetric background, while correlated atom motions produces complex, often highly structured patterns. Indeed, it is roughly true that the diffuse scattering is the Fourier-Transform of the pair-correlation function of the deviation of the positions of the atoms in the crystal from their average locations. At the lowest size scale, nanocrystalline diffraction starts to manifest characteristics intermediate between that of large crystals and quasi-crystalline order. Extracting information from diffuse scattering is difficult. The analysis required to separate the pair correlation function from the generally coherent Bragg scatter is difficult and enormously time consuming. It is difficult to measure the diffuse scattering, and under typical conditions motions of all time regimes are conflated. In principle time-resolved diffuse scattering measurements can be made, to unscramble motions associated with different time regimes. Yet this is a moment when theory, x-ray sources and detectors are making such rapid progress that tremendous opportunities are present. The premise to be explored during this part of the workshop is that these new technologies will allow static and dynamic structural information to be extracted from protein diffuse scatter. Diffuse scattering data from macromolecules in solution is of increasing importance as we strive to understand the role of dynamics in enzyme function. What is the state of the art? What additional information is available from intense microbeams consisting of fast pulses? What time-dependent information may be extracted from novel sample chambers (e.g., lamellar flow cells, converging micro-droplets, etc.) with ERL and USR sources? Etc. Workshop contributions: Presentations will focus on data collection and analysis, with emphasis on the advantages and disadvantages of the time structure, coherence and other characteristics of the ERL and USR beams. Method: The workshop will occur at Cornell's Robert Purcell Center on June 13-14, 2011 and is open to anyone who may be interested. The workshop will start with a short overview of ERL and USR specifications and capabilities for frontier science with nanometer hard x-ray beams (see accompanying attachment for some ERL/USR background). The program, outlined below, will then continue with a core group of invited world-leaders to give short talks and lead the discussion in the relevant science areas. In order to allow time for discussion after each talk, speakers are asked to rigorously adhere to the 20 minute time limit. There will be ample additional time for open discussion and poster sessions. Workshop participants are especially encouraged to be inventive and explore unorthodox ideas. Poster Session: There will be space to put up poster of up to 4'(height) x 6'(width) [1.2 x 1.8 meters]. If you plan to have a poster, please submit a ½ page abstract to Laura Houghton at lab49@cornell.edu and she will try to compile them into the program. There is a specific poster session time shown in the program, but they can remain up for the entire duration of the workshop. Workshop Results: Our goal is to communicate the possibilities for science with an ERL or USR, as well as with other sources, and engage the community in developing ideas for the science case. Imaginative thinking will be required -- that's why we are assembling this workshop group! Documenting the results of the workshop is essential. We hope to explore compelling science that is uniquely enabled by an ERL or USR. Accordingly, each invited participant will be asked to present at least one novel experiment of interest (described by a 1-3 paragraphs of text, an appropriate graphic, and relevant references) that would be very difficult to perform without the capabilities of an ERL or USR. Time will be provided towards the end of the workshop to summarize results and conclusions. Of particular importance is the group discussion at the end of the meeting. It is therefore very important that all invited workshop attendees stay through to the end of the workshop. It is in this period that our discussion leaders will try to pull all the good ideas generated into a summary and we need your very active participation to do this well. As a motivator to stay through to the end and to further the development of "community" for this brief several day period, we have arranged a series of excursions. Depending on the workshop these include, for example, Cayuga Lake Boat dinner cruises, local winery tours, visits to Ithaca's local spectacular gorge parks, etc.