SPring-8 New Apparatus, Upgrades & Methodology Research Frontiers 2021 Research Frontiers 2021 110 Guidelines for high-efficiency/accuracy data collection with multiple small-wedge data collection using the automatic data collection ZOO system In protein crystallography, the data collection m e t h o d i s m o d i f i e d i n a c c o r d a n c e w i t h t h e combination of crystal and X-ray beam sizes, such as in rotation data collection, helical data collection, small-wedge synchrotron crystallography (SWSX), and serial synchrotron rotation crystallography (SSROX) experiments [1]. In SWSX, a cryoloop with multiple crystals is raster-scanned with X-rays, 5 – 20° small-wedge data are collected from each crystal, and a large number of small-wedge datasets are merged to produce a highly complete dataset after individual data reduction [2]. SPring-8 BL45XU was designed to enable fully automatic measurements. All beamline components are controlled by the beamline control software BSS [3]. The ZOO system [4] has achieved automated data collection for all goniometer- based data collections in protein crystallography by communicating with BSS. Small crystals such as LCP crystals of membrane proteins (membrane protein crystals) are measured by SWSX combined with microbeams using the ZOO system at SPring-8. However, there has been no systematic study of the effect of the absorbed dose on the final merged dataset in SWSX. We performed experiments to evaluate the optimal dose for SWSX using microcrystals of the same size and microbeams in order to efficiently collect highly accurate data. The optimal dose for obtaining highly accurate data was investigated by sulfur-SAD (S-SAD) phasing. The evaluation samples used were lysozyme crystals with a controlled size of approximately 20 m m. All SWSX datasets were automatically collected using the ZOO system at BL45XU. The beam size for data measurement was 18 (H) × 20 (V) m m 2 . 10° small-wedge data were collected from each crystal with 18 combinations of a wavelength set of 1.0, 1.4, and 1.7 Å, and doses of 1, 2, 5, 10, 20, and 40 MGy (Fig. 1). More than 400 small-wedge datasets were collected under each condition. The dose per crystal of the data measured using the ZOO system was estimated by RADDOSE 3D [5]. Data processing and merging for each small-wedge dataset were performed by XDS using KAMO [6]. The clustering process for data merging was BLEND [7]. For each merged dataset, S-SAD phasing was performed using SHELXC, SHELXD, and SHELXE in SHELX [8]. Among the six doses at a wavelength of 1.0 Å, only the dose of 5 MGy resulted in successful phasing by S-SAD (Table 1). At a wavelength of 1.4 Å, the doses of 1, 2, 5, 10, and 20 MGy resulted in successful phasing by S-SAD, but not the dose of 40 MGy. Experimental phases were successfully determined for all doses at a wavelength of 1.7 Å. Furthermore, to determine the contribution of the number of merged datasets at wavelengths of 1.4 and 1.7 Å in phase determination, CCmap for each dose was plotted (Fig. 2). The number of merged datasets was set to eight patterns (25, 50, 75, 100, 125, 150, 175, and 200 sets), and 10 repetitions of random dataset extraction and merging were independently conducted for each number of merged datasets for two different wavelengths, 1.4 and 1.7 Å. For both wavelengths, CCmap increased with the number of merged datasets, except at a dose of 40 MGy at a wavelength of 1.4 Å. It was clearly shown that the higher the number of merged datasets, the easier the phase determination becomes at any dose. Fig. 1. Details of SWSX data collection. (a) Multiple crystals in cryoloop. (b) Results of an X-ray raster scan of multiple crystals in a cryoloop. (c) 5–20° small-wedge data are collected from each crystal. (a) (c) (b) Research Frontiers 2021 Research Frontiers 2021 111 For both wavelengths of 1.4 and 1.7 Å, the dose of 5 MGy produced the least number of datasets and increased CCmap, and the second-best doses were 2 and 10 MGy. The number of merged datasets that increased CCmap was higher at a wavelength of 1.7 Å than at a wavelength of 1.4 Å. This is because the Bijvoet ratio of lysozyme in S-SAD is 1.76% (f” = 0.67) at a wavelength of 1.7 Å, which is larger than 1.23% (f”= 0.47) at a wavelength of 1.4 Å. These results indicate that a dose of approximately 5 MGy at any wavelength enables the phase determination of S-SAD at the lowest number of datasets, which can be measured in a shorter time. The higher the number of merged datasets, the easier the phase determination becomes. This shows that data accuracy improves with the number of merged datasets in SWSX. For high doses of 20 MGy at a wavelength of 1.4 Å and 20 and 40 MGy at a wavelength of 1.7 Å radiation damage was severe and the data accuracy decreased. S-SAD phasing was made possible by increasing the number of datasets to be merged and improving the accuracy. However, at a higher dose of 40 MGy at a wavelength of 1.4 Å, the degraded data accuracy due to radiation damage could not be recovered by the merging process. On the other hand, at the lower doses of 1 and 2 MGy at both wavelengths of 1.7 and 1.4 Å, the signal was insufficient, but it was recovered by increasing the number of datasets to be merged. The general dose slicing strategy is to add the sliced data in a merging process to ensure accurate measurements. However, this is not efficient because it requires a long measurement time for a number of repetitions. In SWSX, increasing the number of crystals in the merging process makes the phase determination easier. The dose of 5 MGy at both wavelengths of 1.7 and 1.4 Å was the most efficient in degrading the accuracy through radiation damage and in recovering the accuracy in accordance with the number of merges. In this study, we suggested that the optimal condition for high-efficiency/accuracy data collection condition with SWSX is to acquire one data set at a dose of 5 MGy per crystal. Seiki Baba Japan Synchrotron Radiation Research Institute (JASRI) Email: baba@spring8.or.jp References [1] K. Hasegawa et al. : J. Synchrotron Rad. 24 (2017) 29. [2] V. Cherezov et al. : J. R. Soc. Interface. 6 (2009) S587. [3] G. Ueno et al. : J. Synchrotron Rad. 12 (2005) 380. [4] K. Hirata et al. : Acta Cryst. D75 (2019) 138. [5] O. B. Zeldin et al. : J. Appl. Crystallogr. 46 (2013) 1225. [6] K. Yamashita et al. : Acta Cryst. D74 (2018) 441. [7] J. Foadi et al. : Acta Cryst. D69 (2013) 1617. [8] G. M. Sheldrick: Acta Cryst. D66 (2010) 479. Table 1. Numbers of merged crystals and S-SAD phasing statistics of lysozyme crystals at different doses at a wavelengths of 1.0, 1.4, and 1.7 Å. Fig. 2. Correlation between the number of datasets merged for each dose and CCmap for wavelength of (a) 1.4 Å and (b) 1.7 Å. The mean CCmap values derived from the phase determination for 10 randomly selected datasets are plotted against the number of datasets merged. Dose (MGy) 1 2 5 10 20 40 1.0 Å Number of crystals 366 357 384 243 428 399 S-SAD Phasing Failure Failure Success Failure Failure Failure Wavelength Wavelength 1.4 Å Number of crystals 335 400 404 219 316 348 S-SAD Phasing Success Success Success Success Success Failure Wavelength 1.7 Å Number of crystals 353 453 358 395 218 298 S-SAD Phasing Success Success Success Success Success Success (a) CCmap Number of Sub-datasets Number of Sub-datasets CCmap (b) 25 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 1 MGy 2 MGy 5 MGy 10 MGy 20 MGy 40 MGy 1 MGy 2 MGy 5 MGy 10 MGy 20 MGy 40 MGy 50 75 100 125 150 175 200 25 50 75 100 125 150 175 200
home
- JP
- EN