A re¯ection intensity integration method is presented based upon ab initio calculation of three-dimensional (x, y, 3) re¯ection boundaries from a few physical crystal and instrument parameters. It is especially useful in challenging circumstances, such as the case of a crystal that is far from spherical, anisotropic mosaicity, 1 2 peak splitting, interference from close neighbours, twin lattices or satellite re¯ections, and the case of streaks from modulated structures, all of which may frustrate the customary pro®le-learning and -®tting procedures. The method, called EVAL-14, has been implemented and extensively tested on a Bruker Nonius KappaCCD diffractometer.

A novel diffraction data integration method is presented, EVAL15, based upon ab initio calculation of three‐dimensional (x, y, ω) reflection profiles from a few physical crystal and instrument parameters. Net intensities are obtained by least‐squares fitting the observed profile with the calculated standard using singular value decomposition. This paper shows that profiles can be predicted satisfactorily and that accurate intensities are obtained. The detailed profile analysis has the additional advantage that specific physical properties of the crystal are revealed. The EVAL15 method is particularly useful in circumstances where other programs fail, such as regions of reciprocal space with weak scattering, crystals with anisotropic shape or anisotropic mosaicity, Kα1/Kα2 peak splitting, interference from close neighbours, twin lattices, or satellite reflections of modulated structures, all of which may frustrate the customary profile learning and fitting procedures. EVAL15 allows the deconvolution of overlapping reflections.

A procedure is described, for an Eulerian or Kappa goniostat, to determine the exact spindle value 9 at the moment of impact of a re¯ection on an area detector; such information is essential for the calculation of accurate re¯ection vectors and, eventually, the unit-cell and lattice orientation. The method is based on a comparison of re¯ection impact coordinates on two related images. One image is obtained by spindle rotation over a few degrees (`9 scan') and the other by the same 9 rotation, but with a superposed rotation (1) perpendicular to the ®rst (`9/1 scan'). In both cases, the spindle is kept perpendicular to the primary beam, i.e. 3 = 0 or 180 . Therefore, on the second image, exactly the same re¯ections as on the ®rst will appear, but on a different spot on the detector. From the tangential separation between corresponding re¯ection impacts, the moment of impact (9 value) is calculated. The method is especially useful in small-molecule work, where rather wide scans are required for a reasonable number of re¯ections in one image. It is shown that for our purpose the Eulerian 9/1 scan can be simulated practically exactly by simultaneous uniform rotations of the axes 3 K , and 9 K of the Kappa goniostat.

The solvent behaviour of¯ash-cooled protein crystals was studied in the range 100±180 K by X-ray diffraction. If the solvent is within large channels it crystallizes at 155 K, as identi®ed by a sharp change in the increase of unit-cell volume upon temperature increase. In contrast, if a similar amount of solvent is con®ned to narrow channels and/or individual cavities it does not crystallize in the studied temperature range. It is concluded that the solvent in large channels behaves similarly to bulk water, whereas when con®ned to narrow channels it is mainly protein-associated. The analogy with the behaviour of pure bulk water provides circumstantial evidence that only solvent in large channels undergoes a glass transition in the 100±180 K temperature range. These studies reveal that¯ash-cooled protein crystals are arrested in a metastable state up to at least 155 K, thus providing an upper temperature limit for their storage and handling. The results are pertinent to the development of rational crystal annealing procedures and to the study of temperature-dependent radiation damage to proteins. Furthermore, they suggest an experimental paradigm for studying the correlation between solvent behaviour, protein dynamics and protein function.

Square-planar platinum() complexes containing N,C,N terdentate coordinating, anionic 'pincer' ligands reversibly bind gaseous SO 2 in the solid state by Pt-S bond formation and cleavage giving five-coordinate adducts. When the starting material is crystalline, exposure to this gas leads to quantitative adduct formation with the unique feature that the product is also crystalline, although the crystal structures of the adduct and the SO 2 free complex are significantly different from each other and are both non-porous. Remarkably, the reverse reaction, i.e. the release of SO 2 gas, modifies but does not destroy the crystalline ordering in the arylplatinum assemblies. These processes include repetitive expansion and reduction of the crystal lattice without any loss of crystallinity of the material. Variation of the ligand framework revealed that the presence of intermolecular interactions such as α-or β-type networks is not an essential prerequisite for these crystalline transformations. Consequently, this class of supermolecules provides access to sensitive crystalline switches with 'on' and 'off' positions which are a direct response to the gaseous environment.

The anticancer complexes cisplatin and carboplatin target the DNA major groove, forming intrastrand and interstrand cross-links between guanine bases through their N7 atoms, causing distortion of the DNA helix and apoptotic cell death. A major side effect of these drugs is toxicity, which is caused via binding to many proteins in the body. A range of crystallographic studies have been carried out involving the cocrystallization of hen egg-white lysozyme (HEWL) as a test protein with cisplatin and carboplatin in aqueous and dimethyl sulfoxide (DMSO) conditions. Different cryoprotectants, glycerol and Paratone, were used for each of the cisplatin and carboplatin cocrystallization cases, while silicone oil was used for studies involving N-acetylglucosamine (NAG). Both cisplatin and carboplatin do not bind to HEWL in aqueous media on the timescales of the conditions used here, but upon addition of DMSO two molecules of cisplatin or carboplatin bind either side of His15, which is the only His residue in lysozyme and is assumed to be an imidazolyl anion or a chemical resonance moiety, i.e. both imidazole N atoms are chemically reactive. To identify the platinum-peak positions in the 'with DMSO conditions', anomalous scattering maps were calculated as a cross-check with the F(o) - F(c) OMIT maps. Platinum-occupancy σ values were established using three different software programs in each case. The use of EVAL15 to process all of the diffraction data sets provided a consistent platform for a large ensemble of data sets for the various protein and platinum-compound model refinements with REFMAC5 and then SHELXTL. Overall, this extensive set of crystallization and cryoprotectant conditions allowed a systematic evaluation of cisplatin and carboplatin binding to lysozyme as a test protein via detailed X-ray crystal structure characterizations. DMSO is used as a super-solvent for drug delivery as it is deemed to cause no effect upon drug binding. However, these results show that addition of DMSO causes the platinum anticancer drugs to bind to HEWL. This effect should be considered in toxicity assessments of these drugs and perhaps more widely.

An X-ray crystal structure showing the binding of purely carboplatin to histidine in a model protein has finally been obtained. This required extensive crystallization trials and various novel crystal structure analyses.

No abstract