Manfred S. Weiss

Structure of porin refined at 1.8 Å resolution

The crystal structure of porin from Rhodobacter capsulatus has been refined using the simulated annealing method. The final model consists of all 301 amino acid residues well obeying standard geometry, three calcium ions, 274 solvent molecules, three detergent molecules and one unknown ligand modeled as a detergent molecule. The final crystallographic R-factor is 18.6% based on 42,851 independent reflections in the resolution range 10 to 1.8 A. The model is described in detail.

Crystals of an integral membrane protein diffracting to 1.8 A resolution.

A new crystal form of porin from Rhodobacter capsulatus has been obtained. The crystals are rhombohedral, space group R3, with hexagonal axes a = b = 92.3 A, c = 146.2 A. They contain one monomer in the asymmetric unit and diffract to a resolution of at least 1.8 A.

Prediction of the general structure of OmpF and PhoE from the sequence and structure of porin from Rhodobacter capsulatus. Orientation of porin in the membrane

By comparing the hydrophilicity profiles and sequences of porin from Rhodobacter capsulatus with those of OmpF and PhoE from Escherichia coli, a set of insertions and deletions for alignment of the sequences has been deduced. With this alignment a similar folding of OmpF and PhoE has been predicted as found by X-ray structure analysis of porin from Rhodobacter capsulates. Furthermore, the orientation of the porin trimer in the outer membrane was inferred from topological data on PhoE. According to this result a single channel of approx. 30 A diameter starts at the outer surface. Near the middle of the outer membrane bilayer this channel branches out into three separate channels, each running within a single porin monomer to the periplasmic surface.

A common channel-forming motif in evolutionarily distant porins.

Four new crystal packings of Escherichia coli porins are presented (phosphoporin, maltoporin, and two crystal forms of matrix porin). These were determined by molecular replacement methods using a polyalanine trial model acquired from the refined coordinates of porin from Rhodobacter capsulatus. The successful molecular replacement shows that the dominant motif found in R. capsulatus porin (a 16-stranded antiparallel beta-barrel) also applies to the E. coli porins, despite the lack of significant amino acid sequence homology. A 30 degrees-40 degrees tilt of the beta-strands with respect to the membrane normal was derived from the intensity distributions in the X-ray diffraction patterns for each porin studied, stressing their similarity. In view of the evolutionary distance between enteric and photosynthetic bacteria, the antiparallel beta-barrel may have significance as a basic structural motif for the formation of bacterial membrane channel structures.

Crystallization, structure solution and refinement of hen egg-white lysozyme at pH 8.0 in the presence of MPD.

Hen egg-white lysozyme has been crystallized at slightly alkaline pH using 2-methyl-2,4-pentanediol (MPD) as the precipitant. The crystals are nearly isomorphous to crystals grown at acidic pH using sodium chloride as the precipitant. However, the growth kinetics differ markedly between the two conditions. The major reason for this is a molecule of MPD that binds tightly in between two lysozyme molecules and favors the growth of the crystals along the crystallographic c direction over growth perpendicular to it.

Dehydration leads to a phase transition in monoclinic factor XIII crystals.

Monoclinic factor XIII crystals have been transferred to a solution containing increasing amounts of the precipitant PEG 6000. At a concentration of about 36%(w/v) PEG 6000, a phase transition was observed. The space group of the crystals was preserved on the transition, but half of the 2(1) screw axes were lost, which meant that the unit-cell volume and the content of the asymmetric unit were doubled. The structure of factor XIII in the new crystal form was solved by molecular replacement. About 80% of the changes accompanying the transition can be explained by a rigid-body rotation of half of the factor XIII dimers in the lattice by about 5 degrees. The remaining changes are mostly small interdomain movements of the four domains which constitute one factor XIII monomer.

Online automated structure solution from multiple data sets

With the advent of the high brilliance and small spot sizes beam available from the modern synchrotron sources, the fast detector (i.e PILATUS, EIGER) and automated sample changer has accelerated the collection of X-ray data from protein crystals significantly. This provides possibility to collect large number of datasets from multiple protein crystals and to improve multiplicity and anomalous signal using appropriate scaling, which is beneficial not only for SAD phasing but also for MAD experiments. Keeping this in mind, two multiple datasets evaluation modes have been developed for various experimental phasing.

Environment and susceptibility of disulfide bonds towards radiation damage

Radiation induced damage to protein crystals during data collection is a major impediment for obtaining accurate structural information on macromolecules. Some of the specific impairments inflicted upon highly brilliant X-ray irradiation are disulfide bond cleavage and the loss of the integrity of carboxyl groups of acidic residues. There are results to indicate that not all disulfide bridges are equally susceptible to damage. Consideration of the disulfide bonds in the structures of elastase, lysozyme and acetylcholinesterase indicate that the S-S bonds which have a close contact with a carbonyl oxygen atom along the extension of the bond are more susceptible to breakage. Such an arrangement predisposes electron transfer to occur from the oxygen atom to the disulfide bond leading to its reduction. The energetics and charge transfer in S···O interactions were quantified based on ab initio calculations, using 6-31G(d,p)++ basis set on models generated mimicking protein structures. The charge transfer not only depends on distance but also on geometric orientation of carbonyl O atom with respect to the S atom involved in disulphide bond. Interaction between a nucleophile and electrophile, akin to hydrogen bonding, stabilizes protein structures; but this also provides a pathway of electron transfer to the S-S bond leading to its reduction during exposure of the protein crystal to intense X-ray beam.

Validation of a 96-fragment library for crystallographic screening

Andreas Heine1, Christof Siefker1, Namir Abazi1, Steffen Glöckner1, Nicole Bertoletti1, Engi Hassaan1, Francesca Magari1, Alexander Metz1, Manfred S. Weiss2, Gerhard Klebe1 1Institute Of Pharmaceutical Chemistry, Philipps-University Marburg, Marburg, Germany, 2Macromolecular Crystallography, HelmholtzZentrum Berlin für Materialien und Energie, Berlin, Germany E-mail: heinea@mailer.uni-marburg.de

Sulfur-SAD phasing and UV-RIP analysis on a single Cysteine bridge protein

The advantage of using the anomalous signal of sulfur for phase determination is that only a single, well-diffracting crystal is needed and that a native structure will be obtained. Using long-wavelength S-SAD to a resolution of 1.9 Å we have determined the novel structure of an 89 residue protein with only 2 Cysteines fixed in a disulfide bridge. To the best of our knowledge, the Bijvoet ratio in our example is one of the smallest for which a successful structure solution by S-SAD has been reported. Data were collected on 3 different volumes of a single crystal at beamline 14.1. at BESSY II, Berlin [1], at a wavelength of 1.8 Å. At this wavelength the maximum resolution obtainable was 1.9 Å. The data were processed in space group I222 with a low resolution R-factor of 3.2% and a multiplicity of 17. Based on an anomalous correlation coefficient cut off at 30% the signal extends to 2.6 Å. The sulfur substructure was determined using AutoSol/HYSS [2] showing a total of four clear sulfur positions in the asymmetric unit with a resulting FOM of 0.27 and a BAYES Coefficient of 0.36. The crystal has a solvent content of 62% and the structure reveals a dimer and large solvent channels. Density modification lead to well-defined electron density maps for the protein and associated solvent molecules. This example demonstrates that S-SAD phase determination can work with as little as one S-atom per 45 amino acid residues. Additionally, we performed a UV-RIP (ultraviolet radiation damage-induced phasing) experiment in which a dataset was collected before and after irradiating the crystal with a hard UV laser. An isomorphous difference map shows the clear disruption of both disulfide bridges and we are currently working on combined phasing using both anomalous and isomorphous differences based on the S-SAD and UV-RIP data.