Skip to main content

Development of the new generation of structural biology by coupling in vivo crystallography to intense x-ray sources

Periodic Reporting for period 1 - ivMX (Development of the new generation of structural biology by coupling in vivo crystallography to intense x-ray sources)

Reporting period: 2016-09-01 to 2018-08-31

Third generation synchrotron sources have revolutionized our understanding of the macromolecular machinery of the cell, with over 100,000 elucidated structures that provide atomic detail of macromolecular complexes, membrane proteins and viruses. The identification of protein crystals naturally occurring inside cells and organisms has opened a window for a new type of macromolecular crystallography (MX) and structural biology. The emergence of new approaches in MX, such as the in vivo crystallography (ivMX), coupled to upgraded light sources will allow us to greatly improve one of the bottlenecks of structural biology by removing the need for large crystals, which is achieved by mitigating radiation damages.

This proposal aimed at getting deeper insights into the yet uncontrollable events dictating in vivo crystal growth, by further developing sample handling and delivery procedures and applying these techniques to the structure determination and analysis of readily available ivMX systems. While deciphering these phenomena and applying them to external proteins recombinantly expressed in hosts where in vivo crystal growth could be identified, a small platform for ivMX would be initiated, with the aim of reducing the tedious and costly sample preparation steps currently used in MX. Additionally, the resulting technological developments for sample delivery would facilitate the implementation of serial crystallography approaches at synchrotrons and fourth generation light sources to optimize the use of these methods by MX scientists.
Overall, the ivMX project reached most of the initially defined objectives, further managing to catch the interest of a growing scientific community. Moreover, the methodology developed within the project allowed to setup a consortium for ivMX, regrouping most of the ivMX expert groups around the globe. Among the objectives of the project, the full instrumentation and implementation of the serial crystallography at synchrotron has been a success, specifically when coupled to innovative sample delivery methods.

The ivMX project was first delimited by four work packages (WP), each one containing well identified interconnected milestones extending through the lifetime of the project. WP1 concentrated on the characterization and preparation of several in vivo grown crystals, complemented by WP2 for their structure determination. WP3 and WP4 dealt with the design and engineering of microfluidic-based sample delivery tools, and their implementation within a pipeline for collecting serial crystallography data at synchrotrons. All the four WPs have been brought to completion, however sometimes requiring complementary tasks to answer difficulties in the initial choices of samples and approaches hard to fully handle and master.

Towards the end of the project, methods have been validated that allow identifying in vivo grown crystals directly within their host cells. X-ray data collection were collected, and macromolecular structures determined. Complementary to this, several approaches were proposed to facilitate structure determination of difficult protein crystals, notably through the identification of specific compounds that could additionally favour crystal formation within the cells. Finally, the microfluidic-based serial crystallography approach as installed at PROXIMA-1 of Synchrotron SOLEIL is a real benefit to the crystallography user community, fully robotised and automated for easier handling in an unassisted manner.
The natural occurence of in vivo grown crystals is not very well understood, mostly because of the difficulties to observe such rare events. Using known expression systems, it is now possible to influence this crystal formation within living cells and organisms, which in turns deepen the impact of ivMX on the production of macromolecular crystals directly from their natural environment. Numerous techniques exist for favouring the appearance of in vitro grown crystals, however the current developments are pushing further ivMX as a real alternative for growing macromolecular crystals, not amenable to more classical in vitro crystallisation pipelines.

The state- of -the -art in data collection strategies for macromolecular crystallography involves the collection of serial diffraction data from multiple crystals, and averaging these data in one single set to be as complete as possible. While in the generally accepted serial crystallography approach data are being collected in still images for each crystal, the current approach advocates for collecting data in small wedges rather than still images. Wedges of data have the advantage that they provide with an orientation matrix, in addition of a larger set of images for each crystal, which in turns becomes handy for collecting fully complete and redundant data from fewer crystals.

With the common goal of obtaining unbiased results, data collection strategies are numerous at synchrotron sources, although widely accepted techniques are preferred in the case of macromolecular crystallography experiments. These strategies are generally dictated by the lifetime of the sample, tightly linked to the manner by which it is presented to the x-ray beam. In the current developments, the microfluidic- based sample delivery approach adds one more strategy to the complex data recording of serial crystallography. When designed properly, the microfluidic-based sample sorting and trapping could be adapted at most of the diffraction instruments at synchrotron sources.

Various Serial Crystallography at Synchrotron (SXS) methodologies have been developed and implemented at different x-r-ay sources. Each of these approaches shows advantages over the others, mostly concentrating in rendering the experiments easier for users, as sample preparation in these techniques tend to become different than what it used to be. However, although diffraction SXS experiments proved to be working while rendering spectacular results at x-ray Free-Electron Lasers, all the currently implemented strategies suffer from the high consumption of samples, and as a consequence is sample costly. The main advantage of the currently microfluidic- based positioning of the samples at the x--ray beam interaction point lies in the fact that all the samples are efficiently exposed to the beam. Additionally, using the data collection in wedges of consecutive frames rather than still images, data merging does not rely on Monte Carlo integrations anymore; this strategy of providing with an initial orientation matrix for each sample position helps in speeding data processing, but also in minimizing the amount of necessary crystals to be collected prior to getting a fully complete data set. The amount of necessary sample is therefore strongly decreased, one answer to the concern of users curious in using SXS but reluctant to lose too many of their precious samples.