Home   |  2000   |  2001   |  2002   |  2003   |  2004

SSRL Structural Molecular Biology Summer School 2003

September 16-19, 2003

Summary

Contents

Introduction

The fourth annual structural molecular biology summer school was held at SSRL from September 16-19th. The school focused on three synchrotron based techniques: small angle x-ray scattering, x-ray absorption spectroscopy, and protein crystallography, and the application of these techniques to biological problems. This year's summer school was attended by 19 students (representing 6 U.S. states, Canada, and the U.K.) and was led by a team of 15 tutors (who are internationally recognized experts in their field). It consisted of two days of lectures, followed by a day and a half of rotating practical sessions.

The summer school was opened by Keith Hodgson (director of SSRL), who gave a general introduction to synchrotron radiation. He outlined the development of synchrotron sources over the last 30 years and highlighted the future possibilities with fourth generation light sources.

Session 1: Small Angle X-ray Scattering

The first session of the SMB summer school was dedicated to biological small angle scattering. In his introductory talk, Hiro Tsuruta (SSRL) outlined the information content of solution x-ray scattering in comparison to single crystal diffraction and emphasized the role of small angle scattering in covering the characteristic length scale from sub-nanometers to sub-micrometers.

Patrice Vachette (LURE) summarized modern data interpretation tools, primarily developed by the group led by Dmitri Svergun (EMBL-Hamburg), who unfortunately had been forced to cancel his trip to this summer school. Typical solution scattering data from proteins give 5-15 Shannon channels, which give an order of magnitude of the number of independent structural parameters one can obtain from this type of data. In line with that notion, the use of spherical harmonics to obtain an ab initio low-resolution structure (program SASHA) was discussed. The limitations of this approach are overcome by a different shape determination method in the program DAMMIN, using a large number of small uniform-density spheres and simulated annealing optimization under looseness and disconnection penalties. Recent developments resulted in the program GASBOR which describes a protein shape as a chain of dummy residues whose distribution simulate typical neighboring characteristics in polypeptides. The importance of hydration shell surrounding the protein was discussed through the use of program CRYSOL in comparing crystal structures with those of solution structures using x-ray scattering. The last part of Vachette's talk dealt with the use of program ASSA or MASSHA to obtain a protein complex structure from two component structures by adjusting their relative positions by automated or interactive rigid-body refinement.

Hiro Tsuruta discussed experimental aspects of solution x-ray scattering studies. He emphasized the very low level of scattering signal from proteins in solution which require many precautions to be taken at many different levels of experiments. He summarized experimental geometries, sample cells and detectors commonly used in biological solution x-ray scattering. A significant length of time was spent discussing interparticle interference and sample requirements. Possible reasons for unsatisfactory data quality were given, as well as some helpful tips. He outlined some of time-resolved techniques for kinetic studies on protein conformation change and ended the talk by summarizing current instrumental developments to take full advantage of the high brilliance of SPEAR3 beam. Patrice Vachette was back on the podium to talk about applications of the techniques in solving biological problems, using the interpretation tools by Svergun. Several protein systems were used to demonstrate several different classes of protein structure problems, ranging from allosteric transition to virus maturation kinetics. A systematic study of human ceruloplasmin was most illustrative. Using a combination of all the interpretation tools, he showed that the trinuclear copper active site plays a key role in stabilizing the protein 3-D structure, and proposed a model for the conformation of the apoprotein. As a whole, his talk pointed out the unique information in solution x-ray scattering that complements crystal structure studies. Jack Johnson (Scripps) discussed the large-scale conformational change that takes place in the pH-induced maturation of HK97 bacteriophage capsid through the eyes of complementary techniques. The highly intricate cross-linked chainmail topology of the Head II HK97 particle had been revealed by single crystal diffraction, which led to the subatomic resolution model of Prohead II particle by rigid model refinement based on its cryoEM structure. The two-phase kinetics of local structural transition upon maturation were probed by UV fluorescence. The faster phase coincided very well with the kinetics of global structural events, monitored by time-resolved x-ray solution scattering, indicating highly cooperative nature of the phage particle maturation.

Session 2: X-ray Absorption Spectroscopy

The second session of the SMB summer school was dedicated to x-ray absorption spectroscopy (XAS). The first talk was given by Bob Scott (University of Georgia) and provided the students with a general introduction to XAS terminology, the utility of the technique, and elementary aspects of the theory behind it. He focused on what information could be obtained from the XANES region vs. the EXAFS region and what type of questions XAS can be used to address.

The second lecture in the session was given by Jim Penner-Hahn (University of Michigan). He discussed experimental aspects of XAS. He began with an overview of different detection methods and discussed their strengths and weaknesses. He then discussed how energy selection is done for an XAS experiment, including determination of energy resolution, selection of monochromator crystals, energy calibration, and rejection of harmonics. The next part of his talk emphasized the artifacts that can be introduced into XAS data due to sample heterogeneity, self-absorption effects, or monochromator crystal glitches.

The third talk in this session was given by Matthew Newville (University of Chicago and APS) and focused on XAS theory. He went through the development of the EXAFS equation using a simple physical picture of a scattered photoelectron. He explained how scattering amplitude and phase-shift depend on atomic number and how the program FEFF can be used to calculate these parameters. He then briefly discussed the origin of multiple-scattering contributions. In the next part of his talk, he went through XAS data reduction, including background-subtraction, normalization, k-weighting of EXAFS, and Fourier-transformation; and then went through an example of modeling EXAFS data. Finally, he concluded his talk by discussing the interpretation of XANES data and pointed out that though quantitative XANES analyses using first-principles are still rare, these are becoming possible.

Ninian Blackburn (Oregon Health and Sciences University) was the final speaker in this session and discussed specific examples of the application of XAS in structural molecular biology. He first discussed XAS studies of peptidylglycine monooxygenase, which consists of the PHM and PAL domains. At the time these XAS studies were conducted, no crystal structure was available, and the XAS results were important in establishing that the PHM domain has 2 copper centers that are chemically distinct. XAS studies of the PAL domain provided the first evidence that this enzyme contains a Zn-Fe binuclear site. The second example he discussed, involved the use of selenium substitution in cytochrome c oxidase and azurin to obtain additional structural information about the Cu active site in these proteins. The third example showed how combined XANES and EXAFS studies have been used to define the structure of metallochaperones. Finally, he concluded by discussing future applications that may be made possible with SPEAR3.

Session 3: Macromolecular Crystallography

The lectures in the Macromolecular Crystallography session covered all the steps required to go from a protein in solution to the 3-dimensional model.

The first speakers were George de Titta and John Luft. They talked about protein crystallization. George de Titta described the high throughput microbatch crystallization trial facility at the Hauptman-Woodward institute. John Luft explained how to proceed from successful crystallization leads and grow crystals suitable for crystallographic data collection. He explained how to change conditions to convert rationally between batch and vapor diffusion crystallization conditions. This talk was special because neither of the speakers was physically present at the School. Instead, they webcast their talk live from Buffalo university, using a network camera and a phone link.

The second lecture in the session was given by Bill Weiss, from Stanford University, on the subject of crystalline diffraction. He introduced most of the basic concepts needed to understand and exploit crystallographic data for structure determination: X-ray scattering, structure factors, the relationship between the crystal and reciprocal lattices and crystal symmetry.

Paul Adams went on to describe how to extract the structure factor phases from the amplitudes derived from the diffraction data. He gave an overview to the most commonly used phasing methods used by macromolecular crystallographer: Anomalous Dispersion, (SAD and MAD), Isomorphous Replacement and Molecular Replacement. He showed how density modification procedures can improve the initial phases and provide clearer electron density maps of the structure.

Paul Adams also talked about structural refinement. He discussed the optimization algorithms most commonly used for refinement, with a more detailed description of the simulated annealing method. He explained what parameters can be refined for the model and how to prevent over-fitting by introducing suitable restrains and cross-validation of the model.

The last talk in the session was given by Irimpan Mathews, from SSRL. He introduced the different level of protein structure, described the characteristics of the peptide bond. He talked about homology based models and how they can be of help to the biologists at different stages of a project, from crystallization to structure interpretation and elucidation of the biological function of the protein. He explained how to build and assess the quality of homology and crystallography models and finished by talking about the biologically relevant information it is possible to extract from these structure models.

Session 4: Advanced Applications

The fourth lecture session focused on advanced applications of protein crystallography, XAS, and SAXS. The session opened with a talk by Ashley Deacon (SSRL) on recent work done by the Joint Center for Structural Genomics. This multi-institution collaboration is funded by the protein structure initiative, seeking ultimately to provide structures for all gene products and identify their function. The lecture explained how the JCSG is applying high-throughput methods at several points in the pipeline from target selection to structure publication, the protocols used to optimize efficiency and future plans to increase automation of the whole process.

This was followed by a talk on single crystal XAS, given by Britt Hedman (SSRL). She emphasized that polarized single crystal XAS can be a powerful tool for enhancing the information content of both edge and EXAFS data. She then described the instrumentation that SSRL has developed for combined XAS and diffraction measurements. Her talk concluded with details of some preliminary single crystal XAS studies on putidaredoxin and nitrogenase.

Solution x-ray scattering has become an important structural tool in recent years in protein and RNA folding studies. Sebastian Doniach (Stanford) outlined the use of this technique in the folding kinetics of lysozyme. His group studied the folding event in time scale ranging from submilli- to hundreds of second using stopped-flow and rapid mixing techniques. He showed that the folding mechanism of lysozyme involves initial collapse at the beginning of two separate pathways that eventually lead to the fully folded structure. The folding mechanism of tetrahymena ribozyme was summarized in the second half of his talk. RNA folding takes place much differently than that of protein due to more substantial electrostatic role of counter ions in stabilizing secondary and tertiary structures of RNA.

The session concluded with a tour of SSRL beam lines led by Britt Hedman, Matthew Latimer, and Serena DeBeer George. The tour gave the students the opportunity to see a typical hard x-ray XAS beam line, a soft x-ray experimental setup, and the polarized XAS instrumentation at BL9. The students were also given a tour of the inside the SPEAR ring led by Richard Boyce.

Practical Sessions

The lectures were followed by three rotating hands-on practical sessions for each of the three techniques.

The small angle scattering part was led by Hiro Tsuruta, with additional support by Kazuki Ito (SSRL) and Patrice Vachette. A group of students was first given an outline on the optics and experimental setup at beam line 4-2, a dedicated biological small angle scattering station. Through the use of the EMBL gas chamber detector data collection system, the students learned the basic protocol in static data collection. The group then gathered in a conference room and examined some of the scattering data recorded earlier in the year. The students were guided through the use of EMBL programs GNOM to calculate electron pair distance distribution function and CRYSOL to compare crystal structures against solution structures. A quick trial of structure rendering by spherical harmonics (SASHA) and demonstrations of DAMMIN and GASBOR followed. Vachette warned against a blind and automatic use of this powerful new software, underlining the importance of keeping in mind the limitations and assumptions of each approach and of a critical evaluation of the results.

The XAS practical sessions were led by Serena DeBeer George, with additional support from Deanne Jackson Rudd (Stanford). The students were given a general introduction to XAS data analysis using the program EXAFSPAK (developed by Graham N. George). They learned the basics of calibrating, averaging, and background subtracting data. They were then guided through the use EXAFSPAK as an interface to calculate theoretical phase and amplitude parameters from the program FEFF. The students were given worksheets, which took them through 4 examples of EXAFS data. The tutors emphasized the importance of exploring the strengths and weaknesses of different XAS analysis software.

The macromolecular crystallography tutorial consisted in screening several crystals using the high throughput screening tools developed by JCSG and the SSRL SMB group. The participants learned how to determine the optimal strategy for data collection. They analyzed the data and learned about the most important data set quality indicators. The students also had the chance to handle lysozyme crystals and freeze them in liquid nitrogen. SMB staff members Ana González, Paul Ellis, Aina Cohen, Irina Tsyba, Dan Harrington and Irimpan Mathews assisted the students during this tutorial. Paul Adams also gave a demonstration of the program CNS, teaching the students what practical criteria and tools to use to monitor the results of structural refinement.

Both the lecture and practical sessions were enthusiastically received by the participants. The organizers thank all the tutors for providing a valuable learning experience for everyone involved. Thanks are also extended to the SSRL staff, particularly the user administration, for their attentive support throughout the summer school. This annual event is supported by NIH NCRR.

The SSRL SMB program is funded by NIH (NCRR, NIGMS) and DOE (BER).