Research Article: High-throughput in situ experimental phasing

Date Published: August 01, 2020

Publisher: International Union of Crystallography

Author(s): Joshua M. Lawrence, Julien Orlans, Gwyndaf Evans, Allen M. Orville, James Foadi, Pierre Aller.


A new procedure for experimental phasing at room temperature using diffraction data collected directly from the crystallization plates is described, demonstrated and tested.

Partial Text

The exploitation of multiple crystals is becoming common in data collection and analysis at synchrotrons and X-ray free-electron lasers (XFELs); the encompassing methods are often termed serial macromolecular crystallography (SMX). Indeed, new scenarios have first been envisaged and then implemented to increase the probability of obtaining complete and good-quality sets of reflections by combining data from several to thousands of crystals in creative ways. These strategies have been applied in merging partial data sets from membrane proteins and large complexes (Arakawa et al., 2015 ▸; Axford et al., 2012 ▸, 2015 ▸; Huang et al., 2015 ▸, 2016 ▸; Mylona et al., 2017 ▸), in methodologies adopted to reduce radiation-induced global damage (Garman, 2010 ▸; Garman & Owen, 2006 ▸; Owen et al., 2006 ▸, 2014 ▸) and in the enhancement of anomalous signal for SAD phasing (Giordano et al., 2012 ▸; Liu et al., 2011 ▸, 2012 ▸, 2013 ▸; Rose et al., 2015 ▸; Terwilliger et al., 2016 ▸). It has been particularly enlightening and encouraging to discover that the impact of radiation-induced alterations is minimized by distributing the dose required to solve high-resolution structures over many samples, even without the need for crystal cryocooling. While a clear advantage of working at 100 K versus room temperature (RT) is to slow the global radiation-induced alterations in the sample, flash-cooling often introduces stress and deformations into the crystals. If, on the one hand, a higher X-ray dose can be applied to crystals at 100 K, thus making it possible to collect complete data sets, then, on the other hand, serious non-isomorphism between samples can limit their utility for phasing with isomorphous derivatives. Typically, RT crystallography is used in the early stage of the crystallization process to screen potential hits in situ directly in the crystallization plate (Aller et al., 2015 ▸; Axford et al., 2012 ▸; Bingel-Erlenmeyer et al., 2011 ▸; Douangamath et al., 2013 ▸; le Maire et al., 2011 ▸; Sanchez-Weatherby et al., 2019 ▸). However, the technique has been successful in facilitating high-throughput ligand-binding studies (Gelin et al., 2015 ▸), in solving viral structures (Axford et al., 2012 ▸) and those of membrane proteins (Axford et al., 2015 ▸), and in providing more physiologically relevant RT structures of immuno­globulins (Davies et al., 2017 ▸). In the present work, we demonstrate that the benefits of RT crystallography can be extended to experimental phasing by exploiting tens to hundreds of isomorphous crystals.

The main goal of the present work is to demonstrate experimental phasing and structure solution with data from multiple RT crystals, as described above. In this article, the focus is on de novo phasing with a variety of popular heavy-atom derivatives and serial crystal data-collection strategies. Successful experimental phasing was achieved with four derivatives for lysozyme and one derivative for proteinase K.

The procedure for experimental phasing tested and described in this article extends the emphasis on isomorphism for native data described in previous studies (Axford et al., 2015 ▸; Foadi et al., 2013 ▸). Classic derivatization procedures are lengthy and require sustained and substantial manual handling. This affects individual samples in a unique way, thus making the production of isomorphous crystals very difficult. The automated pipeline detailed above reduces human intervention to a minimum, making the production of similar crystals more likely. An important feature characterizing our method is the ready availability of a ‘derivatization kit’ made up of HA compounds which are nontoxic and are regularly used by the community. More hazardous compounds can be added to the kit if the liquid handler is exclusively dedicated to experimental phasing. In fact, the use of a liquid handler for HA soaking is, in such instances, even safer for the operator. For example, in this study, manual sample handling is minimized; the focus is on evaluating derivatization through a systematic analysis of the easy-to-assess indicators described in Section 3.1. To this end, the effect of different HAs on the crystals can be explored using the mean and standard deviation of the aLCV.

The following reference is cited in the supporting information for this article: Zeldin et al. (2013 ▸).




Leave a Reply

Your email address will not be published.