ブックタイトル日本結晶学会誌Vol60No5-6

概要

日本結晶学会誌Vol60No5-6

The Long-wavelength Macromolecular Crystallography I23 at Diamond Light Sourcedata is presented in the same way as for all the other MXbeamlines, with the special detector geometry, correction fordetector shadowing automatically taken into account by theprocessing pipelines.3．ResultsSeveral novel structures have been solved based on thelong wavelength data collected at I23 over the past twoyears. 28-33）The first published structure was that of ThcOx,an oxidase protein from the cyanobactin pathway. 34）Crystalswere highly non-isomorphous and varied in diffractionquality. Molecular replacement failed, labelling with SeMetwas consistently unsuccessful and, despite various attemptsto bind heavy atoms, it was not possible to find a derivativeto allow phasing. The structure could be solved from datacollected at a wavelength ofλ＝3.096 A from a crystaldiffracting to 3.15 A, with anomalous data extending to only4.2 A. While for native phasing experiments at wavelengthsaroundλ＝2 A typically high multiplicity data is needed（～100）, for ThcOx 400°of diffraction data, resulting ina multiplicity of 26, was sufficient to solve the structure.Figure 3 a shows the ThxOc structure with the anomalousdifference Fourier map（5σ）.The salmonella-secreted effector SseK3（PMID:29449376）was also solved by SAD, 30）but in this case thedata showed some pathologies, such as pseudo-translationand anisotropy, potentially explaining the difficulty ofsolving the structure on standard beamlines. 1440°of datawere collected with 0.1 s exposure and 0.1°oscillationfrom a single crystal containing seleno-methionine labelledproteins, at a wavelength of 2.755 A. The position of theselenium and sulphur atoms could be found in SHELXDand confirmed the presence of four Ssek3 in the asymmetricunit. At the experiment wavelength, the Se anomalous signalis still stronger than the S one, hence the highest anomalouspeaks correspond to Se. After phase improvement, most ofthe structure could be built automatically. The anomalousdifference Fourier map（5σ）is displayed in Fig.3b.A third example, demonstrating the capabilities of thebeamline towards high-resolution data collections, is thestructure of the plastic degrading enzyme PETase. 29）Phasescould be obtained from a first crystal based on data collectedat a wavelength of 2.455 A and data to 1.7 A resolution. 3600images collected with 0.1 s exposures per 0.1°per imagewere of sufficiently high quality to solve the structure. Asecond crystal with different morphology turned out to bea different polymorph, with the space group being P2 12 12 1,rather than C222 1, the crystal used for phasing had belonged日本結晶学会誌第60巻第5･6号（2018）Fig.3 a）anomalous difference Fourier map（5σ）for ThcOx.b）anomalous difference Fourier map（5σ）for Ssek3.Electron density（2F o-F c, 1σ）around Trp257 ofPETase.to. As the crystal diffracted to the edge of the detector atlonger wavelengths, a wavelength of 1.240 A was chosenand diffraction data to a resolution of 0.92 A were collectedfrom 1800 images atκ,φ＝0°, followed by 800 additionalimages collected opening the kappa arm of the goniometertoκ＝－50°（φ＝70°）for complete diffraction data. Thecombination of low-background in-vacuum crystallography,the kappa goniometer and the curved detector allowed tocollect very high quality data without any signs of radiationdamage from this important enzyme. Part of the electrondensity（2F o-F c, 1σ）is shown in Fig.3c.4．Summary & OutlookThe long-wavelength MX beamline I23 at DiamondLight Source has been a very challenging project. The novelapproach to establish in-vacuum crystallography requiredseveral bespoke solutions which have been discussed above.237