Issue 100
EMBL Grenoble technology teams provide a sneak peek into their latest collaborative project in structural biology services
Have you ever wondered how groundbreaking new scientific technologies come into being?
Or how scientists and engineers work together to push the frontiers of innovation?
We caught up with several members of EMBL Grenoble’s technology-related teams, who opened their doors to give us a sneak peek into this process of innovation. Their skills and collaborative strengths are demonstrated aptly by one of their latest projects: the complete automation of an integral step in X-ray crystallography, a technique used by scientists worldwide to determine the 3D structures of proteins and other macromolecules.
EMBL Grenoble has been at the forefront of technological developments in the field of structural biology for three decades. This expertise has been built over time thanks to collaborative work between structural biologists and engineers.
Franck Felisaz, a mechanical engineer in the team of Gergely Papp, joined EMBL in the mid-nineties to work on instrumentation projects. “The early days of instrumentation at EMBL Grenoble involved what were pretty much craft projects, based on progressive iterations,” explained Felisaz. “The teams were not structured like they are today, but there was a family-type work environment, where engineers would talk with the scientists over a cup of coffee about the challenges they faced.”
According to Felisaz, such informal interactions often led to creative solutions to issues scientists face in their work. He referred to this interactive and problem-solving spirit as a “virtuous circle creating a trustful relationship between engineers and scientists” – something that still persists at EMBL Grenoble.
With the opening of the European Synchrotron Radiation Facility (ESRF) next to EMBL Grenoble in 1994, the use of X-ray-based techniques for determining the structure of macromolecules became popular with structural biologists. EMBL Grenoble started to jointly operate crystallography beamlines dedicated to structural biology with the ESRF, through the Joint Structural Biology Group (JSBG).
Macromolecular crystallography allows scientists to examine the structure of crystallised macromolecules by bombarding them with X-rays. However, the process involved many challenges.
“Scientists realised in the late nineties that manual operations were a real bottleneck,” said Felisaz, who was already thinking of automating processes back then. “I had been discussing with them for some time about how we would reduce the time wasted and inaccuracies that result from manual manipulations.”
This need for technological improvements came at the same time as Florent Cipriani taking over the leadership of engineering activities at EMBL Grenoble, and the decision to create separate instrumentation and synchrotron teams. “Benefitting from experiences from the industrial world, the newly structured instrumentation team at EMBL Grenoble started to offer solid and appropriate solutions for structure determination based on macromolecular crystallography,” said Felisaz.
The team started to develop prototype instruments allowing reliable and reproducible automation of processes. These innovations included, among other things, sample holder standards (to hold crystals and facilitate robotic handling), sample changers (to handle crystals), and goniometers (to present the crystals to the X-ray beam), and were real game-changers for scientists who rapidly adopted them. Many were also commercialised for use all around the world.
Felisaz explains that the innovation cycle doesn’t stop here: once a technology has been adopted by scientists, they will eventually find new ways of using it for their research and, together with engineers, arrive at opportunities to take it to the next level.
One of the recent additions to the portfolio of EMBL inventions is the CrystalDirectTM technology, originally developed by the Marquez and Cipriani teams in EMBL Grenoble. This technology streamlines the preparation of crystals for diffraction experiments and makes it possible to integrate crystallisation and synchrotron data collection into a continuous experimental workflow than can be controlled and operated over the internet.
The CrystalDirectTM technology is currently implemented at the crystallisation facilities at EMBL Grenoble and EMBL Hamburg, and has been utilised by hundreds of users from academia and industry around the world. The integration of a CrystalDirect™robot at the MASSIF-1 beam line, which offers new and exciting experimental opportunities, is a good example of the synergies between scientists and engineers at EMBL.
This massive project involved the three technology-related teams at EMBL Grenoble: the Marquez Team, running the High-throughput crystallography platform (HTX), the McCarthy Team, operating the EMBL-ESRF beamlines, and the Papp Team, leading the instruments development.
It also included the Structural Biology group at the ESRF, in line with the long-standing and fruitful collaboration between EMBL Grenoble and ESRF over the last few decades through the JSBG. The project benefitted from the latest upgrade to the synchrotron in 2020 (EBS – the ‘Extremely Brilliant Source’) – which made ESRF the most powerful source of X-ray worldwide, allowing unprecedented possibilities in terms of research and technological developments.
This project is therefore the culmination of several innovations developed in parallel at EMBL Grenoble, as it combines CrystalDirect™ – an automated crystal harvester developed in 2008 by the Marquez and Cipriani teams – and MASSIF-1 – a unique automated beamline dedicated to macromolecular crystallography, jointly developed and operated by EMBL and the ESRF.
Macromolecular crystallography requires a high intensity X-ray beam which is produced by a specific ring-shaped infrastructure, called a synchrotron, composed of ‘beamlines’ that use X-rays generated by accelerating electrons to near-light speeds in the ring.
To carry out diffraction experiments at a synchrotron, large numbers of crystals have to be produced and prepared for data collection. This typically happens at dedicated robotic facilities like the HTX lab in EMBL Grenoble.
The CrystalDirect™ technology consists of an extremely precise robot that streamlines the step of crystal sample harvesting and preparation for X-ray diffraction at the beamline. It automatically harvests the biomolecular crystal from the crystallisation plate where it has been grown, and preserves the crystal sample by cryo-cooling it.
“CrystalDirect and MASSIF-1 were a perfect match at the perfect time,” said Matthew Bowler, EMBL beamline scientist in the McCarthy team, in charge of operating MASSIF-1. Bowler joined EMBL in 2012, after spending several years at the ESRF where he started developing ideas for a fully automated beamline – MASSIF-1 – which became operational in 2012 and received a major upgrade in 2020.
“When MASSIF-1 was coming online, CrystalDirect™ was being developed. The latter started to generate a large number of crystals, and we actually had the capacity and automation to run them with MASSIF-1 without anybody having to do it manually. So the obvious next step was to marry these two technologies to have CrystalDirect™ on the beamline,” added Bowler.
This idea was first tested in 2018 as a proof of concept on another beamline, ID30B, putting together three technologies developed at EMBL Grenoble: the ‘Flex’ – a robotic arm in charge of moving crystal samples from one place to another, the goniometer – an instrument presenting the crystal to the beam, and CrystalDirect™ – a machine that harvests the crystal samples. They identified key issues that were likely to stall the project and would need to be resolved before integrating the CrystalDirect™ machine in the beamline.
After a successful proof of concept, the construction and integration phase started in 2021 on MASSIF-1, led by the Papp team – with Frank Felisaz in charge of building the machine, mechatronic engineer Marcos Lopez Marrero in charge of the electronics, and software engineer Jeremy Sinoir in charge of preparing the machine software for its integration in a beamline environment. The engineers kept the user firmly in mind during the process.
“The experimental room of MASSIF-1 is very small, so it has been very challenging to fit everything into this space within the imposed constraints. We wanted to make sure that the ergonomy of the installation would be optimal for scientists so that they could easily handle their samples, even if it would be more complicated for the maintenance,” said Felisaz.
“Once the CrystalDirect™ instrument was installed in the beamline, the project then became like a football game, requiring a lot of coordination and where everyone had to work together,” said Bowler. All the scientists and engineers involved in the project had weekly meetings, as well as brainstorming and technical planning meetings, to plan the work carefully.
“The main challenge was to get about five different pieces of hardware and five different pieces of software to work together reliably to be able to provide the service to users,” explained Bowler. “It’s generally the case with every beamline, but we added an extra level of complexity with the harvester; it was a huge challenge.”
Serena Rocchio, postdoctoral fellow in the McCarthy and Marquez teams, was in charge of coordinating the project. “The complexity also came with having to coordinate this new installation together with running the normal operation of the beamline,” said Rocchio. “This wouldn’t have been possible without the great teamwork and expertise from different backgrounds we had access to.”
Rocchio worked in close collaboration with software engineer Jeremy Sinoir from the Papp team, who made sure that all the machines’ software could work together. “There was a specific part requiring a lot of attention and meticulousness – error case management,” said Sinoir. “This beamline has to work completely autonomously, so we couldn’t afford any malfunctions. For each potential problem encountered, we needed to develop a recovery procedure.”
This was achieved by integrating the harvester into MXCuBE, an open-source beamline control software developed by an international consortium coordinated by the ESRF and with the CRIMS software, used to operate the HTX facility. This adaptation was done by Jean-Baptiste Florial, a Full Stack Software engineer in the McCarthy team, Peter Murphy and Raphael Bourgeas, CRIMS Software Developers in the Marquez Team, and Jeremy Sinoir in the Papp Team. This integration enables scientists to design experiments over the internet from their labs, which later on will be automatically executed at the HTX lab and at the beam line.
Finally, the scientific validation of this new pipeline was a crucial component of all the project coordination. It offered the team the chance to integrate new experimental modalities, like data collection at room temperature – an opportunity the scientists identified before the project started and has now been implemented.
Usually, crystal samples are cryo-cooled to a temperature of 100 Kelvin (-173°C) to allow them to resist radiation damage from powerful X-rays that normally destroy biological material. Data collection at room temperature is thus extremely difficult to do and until now involved a laborious manual process. Since the new automated pipeline allows the crystal sample to be quickly harvested and immediately mounted on the beamline, it provides scientists an accessible and reliable way of collecting data at room temperature that can now be automated for the first time.
After running all the commissioning phases, the new automated ‘protein to structure pipeline’ was used for the first time on one of Bowler’s and Marquez collaborative research projects with Mohamed-Ali Hakimi, a researcher at the Institute of Advanced Biosciences. The scientists are currently working on a protein from the parasite Toxoplasma gondii with the objective of developing new drugs against Toxoplasmosis. The combination of low- and room-temperature automated modalities makes this beamline unique and particularly useful when a lot of screening is required, like in the case of drug development.
“When we got our first sample, it worked perfectly! It was a fantastic feeling,” Bowler said. “We know that this is going to support amazing research and hundreds of people across Europe will find it useful. It’s very exciting”. This rapid integration was made possible by years of preliminary work of Marquez, McCarthy and Papp Teams in collaboration with the ESRF Structural Biology group, which provided a solid and reliable basis for the implementation of this fully automated pipeline.
This enthusiasm was shared by the engineers involved in the project. “The first time that several of our most successful machines started working together was truly an incredible moment; we’ve never done that before!” said Sinoir.
“We made a unique and super-optimised beamline,” added Felisaz. “We were able to add a lot of value to a service that was working well and make something that works even better in terms of high-performance automation. It showcases what is possible to do with such a system and could push the performance of other beamlines too.”
This project was made possible by the synergy between the capabilities of HTX lab and the MASSIF-1 beamline. MASSIF-1 has offered the new CrystalDirect™ modalities to external academic users since October last year. “It is quite impressive and fulfilling to see such a project taking shape and to share each step of this experience with great scientists and engineers,” said Rocchio. “Now, we want to give the possibility to perform challenging experiments to the entire community and move to high-throughput while working with multiple projects.”
Together with the engineers, technicians and scientists involved in the project, Rocchio is also identifying opportunities for improvement, like automating and streamlining even more processes. They are, for instance, working on improving the washing and supply chain of ‘pins’ – the crystal sample holders.
This is just one of many matters that scientists and engineers are discussing over a cup of coffee at EMBL Grenoble – just like old times. In Felisaz’s words, “This is how continuous innovation works; we always have in the back of our minds the search for something better.”
EMBL Grenoble isn’t alone in exploring synergies between biologists and engineers. At EMBL Hamburg, structural research relies on cooperation between experts in different disciplines, including those in the Instrumentation Team, which consists of physicists, mechanical and software engineers, and robotics experts. Together, they construct diverse instruments for X-ray-based structural biology tailored to the needs of users. Last year, a new transfer system has joined the MARVIN (MultiAxesRobotic-VersatileINstaller) robot family. The MARVIN system enables the quick and safe handling of fragile crystals, which are used by crystallographers to determine the structure of proteins. The new MARVIN will accompany the CrystalDirect™ Harvester, which is already in operation at EMBL Hamburg.
Over the years, the CrystalDirect™ project has been, in part, funded by EC infrastructure grants like Bioxhit, iNEXT, iNEXT discovery, as well as by ANR grants and to some limited extent by Instruct-ERIC. This helped fund the construction of the CrystalDirect-2 machine, which is now installed at MASSIF-1. The MASSIF-1 upgrade project is a collaboration between EMBL and ESRF. User access to the new pipelines exploiting the integration of HTX lab services and the new capabilities of CrystalDirect-2 at MASSIF-1 are funded by Instruct-ERIC and the EC-funded iNEXT Discovery, via the HTX lab with beam time provided by the ESRF, as per the EMBL-ESRF agreement.
What was it like to work in structural biology back in 1999?
We asked Matthew Bowler, beamline scientist co-responsible for the MASSIF-1 beamline. In 1999, Bowler was a Master’s student working on membrane proteins and had obtained beamtime at a brand new beamline at the ESRF. This beamline had just opened and was run by an EMBL scientist, Andy Thomson – now working at the SOLEIL synchrotron. “I was incredibly excited to travel from the United Kingdom to Grenoble to use this new beamline. Membrane proteins are very difficult to crystallise and need a lot of screening. We managed to screen 10 crystals in 24 hours. It seemed amazing at that time because you could only get a very limited number of crystals, and you had to do everything manually. But if you compare it with what it’s possible to do today, it’s just completely different! Now you can get through 200 to 500 crystals in a day – without anyone having to be physically present and manipulate things. ”
EMBLetc., the online magazine of Europe’s life sciences laboratory, celebrates its 24th birthday with its 100th issue. We took a walk through the past issues of this dynamic publication, and here are 10 recurring themes that emerged.
Scientists at EMBL Rome are developing new paradigms to study the impact of diverse environmental factors on reproduction in mammals and disease risk in their progeny.
Click here to download a quick overview of all the articles included in this digital issue of EMBLetc.