Meeting report

XIX Congress and General Assembly

[IUCr Congress logo]

'...The original venue for IUCr XIX was Jerusalem - often connoted the 'City of Peace'. Throughout its history, however, Jerusalem, the meeting point of three continents and the focus of three religions has witnessed periods of conflict. Unfortunately, we are living through one of these periods. So, in May of 2001 we moved the venue of IUCr XIX from the City of Peace here to Geneva, a city which prides itself on being a place where in the past peace has been made among nations and making peace for the future is a business of the present.

We are reaching the end of another rather long and varied scientific meeting, during which we have enjoyed the sharing of ideas and the collegeality of fellow scientists. They came from more than fifty countries and a variety of cultures and backgrounds. Some might argue that a scientific meeting is a luxury of peace; others, that it serves to promote the peace we all long for. I would like to toast all the participants in IUCr XIX who came to Geneva to participate in this meeting and wish them godspeed in their return in peace to their homes and laboratories with renewed scientific enthusiasm and deepened friendship, to meet again in this forum in Florence in 2005. Bon Appetit, a pleasant evening - LaChaim.'

Joel Bernstein at the Congress Banquet

Keynote reports

New superconductor MgB2 and related compounds

[Akimitsu, Antipov] Jun Akimitsu and Evgeny Antipov
J. Akimitsu (Aoyama-Gakuin U., Japan) described the structure, effects of substitution and physical properties of this superconductor exhibiting the highest Tc (39K) among metal and alloys. The discovery of superconductivity in MgB2 by J. Akimitsu and co-workers in January 2001 was one of the most important achievements in Solid State Physics and attracted great attention in the scientific community resulting in a large number of publications (about 1.5 citations per day).

Evgeny Antipov

The good, the bad and the ugly; experiences at the synchrotron

[Garman, Dauter] E. Garman and Z. Dauter
Z. Dauter (NCI, Brookhaven Nat’l Lab, USA) described the use of synchrotron radiation for structural biology. He outlined the essential components of a protein crystallography beamline, showing examples from beamline X9B at BNL, where he is coordinating an effort to integrate diffraction data collection and processing into a single control programme. He showed some potential pitfalls of synchrotron data collection, including the undesirable effects of radiation damage, but much to the relief of the audience, he did not categorise his many beamline users into `Good, Bad or Ugly’!

The questions and comments at the end of the lecture focussed on issues related to worldwide efforts to automate protein crystallography synchrotron beamlines, and how data quality might be affected by reduced experimenter intervention.

Elspeth Garman

Structural mechanisms of self-assembly and polymorphic supercoiling of the bacterial flagellum

[Stuart, Branden] David Stuart and Carl Branden
K. Namba (Osaka U., Japan) reported on the simulataneous application of X-ray crystallographic, electron microscopic and other biophysical techniques in conjunction with biochemical and molecular biology approaches to unravel the structure and function of a nanomachine in the form of the bacterial flagellum which enables bacteria to swim. Although the flagellar filament is made up of a single protein flagellin, the tubular organisation of the protein molecules within it is such as to permit different modes of swimming. The motor at the base of the filament is connected to the latter by a hook which serves as a universal joint. Detailed structure analysis of the component proteins and their organisation in the flagellum have provided valuable insights into the complex movements that contribute to the mobility of bacteria.

M. Vijayan

Virus structure: the simple, the complex and the greasy

[Tanaka, Schenk, Woolfson] Michiyoshi Tanaka and Henk Shenk present Michael Woolfsonwith the Ewald Prize at the Opening Ceremony
D. Stuart (Wellcome Trust Center for Human Genetics, UK) focused on spherical viruses with emphasis on those that he and his group have studied in Oxford. The importance of quasi-equivalence for building up the icosahedral viral coat was exemplified by the simple T=3 plant viruses which contain three identical subunits in the icosahedral asymmetric unit. Stuart pointed out that the same principle applies to the foot-and-mouth disease virus,FMDV, where the shell is built up from three chemically different but structurally similar polypeptide chains.

He also pointed out that the receptor binding site is quite different in FMDV from that in the related common cold virus.
Complex viruses were exemplified by the structure of the core of the bluetongue virus which his group determined a few years ago in a crystallographic 'tour de force' operation. The shell comprises 900 subunits of two different types. One type forms an inner scaffold in a T=2 arrangement which forms a symmetry mismatch to the T=13 outer layer formed by the second type of subunit. Nature overcomes the 'forbidden' T=2 arrangement by forming dimers with the subunits in different conformations. An animated display illustrated beautifully this arrangement. The transcription complex is inside the shell together with a high concentration of mRNA molecules, 400mg/ml. This RNA is highly organized by the scaffold.

The greasy viruses were exemplified by the bacteriophage PRD1 in which the core is enveloped by a lipid bilayer containing spike proteins. He also gave a progress report on his structural studies of a modified virus without the spike proteins which are done in collaboration with D.H. Bamford in Finland and R. Burnett in US. The crystals, which contain about 66MDa in the asymmetric unit, are unstable, diffract to 4Å resolution and give only one image per crystal. So far, 2,317,561 reflexions have been measured, phases were obtained from a cryoEM model combined with the known structure of one subunit and chain tracing is under way aided by the presence of 3.600 Se-met residues. He described features of the preliminary model and pointed out 150Å long thin strings of additional electron density running between the building blocks of the shell which corresponds to a 9kDa cementing particle called P30. He also showed electron density corresponding to the lipid bilayer where some structural organization could be seen. The structural studies of PRD1 which provide proof of principle that it is possible to obtain structures of enveloped viruses left the audience stunned in admiration.

Carl Branden

Contribution of direct methods to macromolecular structure determination

G. Sheldrick’s (Göttingen, Germany) lecture, which played to a near capacity crowd on the last morning of the conference, displayed a neat symmetry with the opening lecture of the Congress given by M. Woolfson. In his Ewald Prize oration Prof. Woolfson, traced the development of direct methods from their early beginnings to their present use, which has extended to small proteins and more importantly as a combined approach with other methods to phase macromolecular structures. George Sheldrick took the liberty of redefining the title of his lecture to include not only ab initio methods but also the use of weak phasing information such as that which results from the anomalous signal from sulfur atoms in proteins or phosphorus atoms in nucleic acids. He outlined the dual real space/reciprocal space recycling methods implemented in SHELXD for both substructure solution and MAD phasing. An example included the location of 160 selenium atoms. SHELXE is the latest development in the SHELX family. It implements 'high throughput phasing' as a means to simply and quickly obtain robust phase information with a view to providing information at the time of data collection which can be used to further plan the conduct of the experiment. Despite the fact that SHELXE was not aimed at providing the final phases, in several cases it appears to be competitive with more sophisticated methods. It makes some simplifying assumptions including acceptance of the heavy atom coordinates from SHELXD without further refinement and the presence of only one type of anomalous scatterer. SHELXE also implements 'sphere of influence' density modification for more rapid calculation. In summary Prof. Sheldrick demonstrated that, despite his protestations that SHELX is only a hobby, he is playing a vital role in developing new methods of structure solution and making them available to the community in a useable form. The lecture concluded with a discussion period.

Mitchell Guss

Chemical bonding as electronic coherence

W. Weyrich (U. of Konstanz, Germany) demonstrated with examples how the information from diffraction and Compton scattering studies can be combined using the Density Matrix (DM) representation of the electronic wave-functions in solids. He presented diagrams showing the relations of the observables in position and momentum space emphasizing that the chemical bonding cannot be understood by the charge density only, which corresponds to the diagonal part of the DM. The off-diagonal part carries essential information about the wave-function coherence between atoms, and this can be derived from the electron momentum density measured by Compton scattering (see abstract p. C191, Acta Cryst. A58 (suppl.)).

There were several questions and comments from the audience:

Q: The negative parts in the Wigner function are not due to the Pauli principle, but rather due to the fact that the Wigner function is not a true phase-space density.

A: It is the Pauli principle that forces the electrons of the fluorine atom and molecule into p-orbitals, which give a negative contribution to the Wigner function. We agree about the non-classical nature of the Wigner function.

Q: Can it be expected that the approach of utilizing position and momentum densities for a more complete understanding of chemical bonding will be applicable to complex systems such as macromolecules, or will we have to extrapolate from findings for simple systems?

A: Since the momentum space data are vector-space data, they behave in a way similar to the Patterson function, though at the wave-function level rather than at the density level. I am therefore afraid that for very complicated systems we will have to extrapolate from simpler systems. Remember, however, that the essential information of chemical bonding in the diffraction data is included in a small number of low-order reflections. It can be envisaged that a comparable number of Compton profiles can be measured.

Q: Why are you so puristic and do not impose additional constraints such as a model on your reconstruction of density matrices?

A: By choosing a particular discrete representation, i.e. a basis set, we also have a model aspect. However, by a prudent choice we can come close to our goal of not biasing the result and rather following a clear, quantum-mechanically well-defined method.

P. Suortti

The GTPase switch: a familiar, conserved module with unexpected variations

A. Wittinghofer (MPI, Germany) began with an overview of the ever-growing number of branches of the Ras superfamily of GTP-binding proteins and their common functional cycle, which takes them from a GDP-bound or 'off' state to a GTP-bound or 'on' state and back again. These transitions are accelerated by guanine nucleotide exchange factors (GEFs) and GTPase activating proteins (GAPs), respectively, and only the triphosphate complexes interact with downstream effector molecules. The related structural characteristics were described, followed by examples of crystal structures of complexes between small G-proteins with their respective GEF or GAP. Although the core features, most of which can even be found in motor proteins, are repeated in all G-proteins, there still exists a wide variety of molecule-specific variations that contribute to the unique interactions typical for each G-protein. In contrast to the conservation shown by the G-proteins, their GEFs and GAPs are completely unrelated; e.g. the RanGEF (RCC1) adopts a beta-propeller fold whereas the RhoGEF (Sos1) consists of mostly alpha helices; similarly, RanGAP, which contains 11 leucine-rich repeats that adopt the nonglobular shape of a crescent, bears no resemblance to RhoGAP or RasGAP. Despite their dissimilarity, GEFs and GAPs bind close to the 'switch regions' of the G- proteins. The interaction sites of the various effector molecules, however, cover almost the complete surface of the different G-proteins. Then, Dr. Wittinghofer discussed the roles of the members of the Rho branch of the family tree in the organization of the actin cytoskeleton, followed by striking examples of how various pathogens make use of this system in their quest to invade their host cells and exploit them to their advantage. The bacterial proteins involved are excellent mimics of their eukaryotic counterparts even at the mechanistic level but, again, structurally completely unrelated, which points to their independent origin and convergent evolution.

Emil F. Pai

COMCIFS open meeting

How to provide a seamless flow of computer-based information between crystallography and its neighboring disciplines was the question discussed during the open COMCIFS meeting held during the IUCr Congress in Geneva. The talks addressed the need for compatibility at the level of the definition of scientific concepts and at the level of file structures.

The session opened with a talk by B. McMahon and ended with a talk by J. Westbrook, both of whom pointed out the need for compatibility between the scientific definitions (the ontologies) provided by the CIF (Crystallographic Information File) dictionaries and those provided by the dictionaries of the disciplines of chemistry and biology. The Protein Data Bank has achieved a seamless connection with molecular biology by ensuring that the neighboring fields use STAR dictionaries that are fully compatible with CIF. Chemical databases are still in an experimental state and there is a danger that a variety of different and mutually incompatible dictionaries will be developed making it difficult to transfer information from one chemical data file to another and complicate the interchange of information between chemical and crystallographic databases.

The remaining talks of the session were devoted to compatibility of the file structures. While CIF has the best developed set of dictionaries of any discipline, it is currently not well provided with the software required to manipulate the files.

Disciplines now developing file structures are attracted by XML (eXtended Markup Language) because it is well provided with software written by the information technology community. However, few disciplines have developed the dictionaries needed for serious XML applications. Both J. Westbrook and N. Spadaccini described programs that can convert CIFs to XML files thereby allowing crystallographers to exploit the XML software, H. Bernstein showed that although CIF and XML have many similarities, there is more than one way in which a CIF can be mapped into XML. Emphasizing that CIF has always remained one step ahead of XML in functionality, S. Hall demonstrated an advanced dictionary language, StarDDL (Star Dictionary Definition Language), that will allow derived information to be calculated directly using algorithms stored in the dictionary. Because of the stable yet flexible design of CIF, the large archive of CIFs built up over the last decade will be able to exploit this feature, allowing the user to retrieve derived information that may not be explicitly stored in the CIF.

The session described a field expanding so rapidly that, even though software and dictionary developments may have difficulty keeping up with each other, the CIF project remains on the cutting edge of information technology and the opportunities for developing innovative software have never been greater.

David Brown