New technical developments in Grenoble:

A new study at the HTX facility in Grenoble shows how a simple thermal stability assay can help predict the crystallization likelihood of biological samples.

The identification of crystallization conditions for biological molecules relies largely in a trial and error process in which a number of parameters are explored in large screening experiments. Currently construct design and sample formulation are recognized as critical variables in this process and often a number of protein variants are assayed for crystallization either sequentially or in parallel, which adds complexity to the screening process. Significant effort is dedicated to sample characterization and quality control experiments in order to identify at an early stage and prioritize those samples which would be more likely to crystallize. However, large datasets relating crystallization success to sample properties are generally lacking.

Thanks to a new, large scale study carried out at the HTX facility of the EMBL Grenoble outstation (Dupeux et al., Acta Cryst. D67, 2011), a method to estimate the crystallization likelihood of a particular sample based on a simplified thermofluor assay has been developed. This method is easy to implement and consumes very small amounts of sample. The results of this assay can be used to determine optimal incubation temperature for the crystallization experiments or to prioritize those constructs which are mole likely to produce crystals. More generally this work provides an objective test that can contribute to make decisions in both focused and structural genomics crystallography projects.

You can access the full article by clicking here or at the Acta Crystallographica, Section D online resources.


Additional service available at the HTX facility in Hamburg!

In the future, samples from P-CUBE users will be tested by SDS-PAGE, MALDI-TOF MS and a 3-point thermofluor experiment. This serves as a minimal feedback to both, users and the facility manager, of the sample quality immediately prior to the crystallization experiment. Sample degradation due to shipping or insufficient sample quality, due to user inexperience will be monitored and recorded. The data provide the basis for an assessment of whether the sample quality could be improved. The service itself does not require additional sample and will be performed from the technically necessary surplus amount.

In an extended access mode, users with high-grade samples (as defined by the criteria above and evaluated by EMBL-Hamburg experts) that did not succeed in crystallization experiments, will be given the opportunity to visit the EMBL Hamburg facilities to further characterize their sample(s) with SAXS, DLS, Thermofluor, ITC, methylation, limited proteolysis MS and further sample purification. After discussion between the user and the EMBL Hamburg responsible staff, an experimental plan will be designed that includes the appropriate methods. This plan will then be submitted for approval to the TNA board (TBD). The expected outcome is a sample or samples with properties that will enable SAXS experiments and new rounds of crystallization. Users will have at least a much better characterized sample and likely a low resolution structure from the SAXS measurement. We expect that one visit will last on average 5 working days. The requested funds will cover travel reimbursement, accommodation in the local campus guesthouse and all materials for the experiments.


Speeding up crystallization refinement experiments with the OptiLink software module

The identification of initial crystallization conditions has been facilitated by the introduction of automated laboratory equipment and nanovolume crystallization robots. However, initial crystallization conditions often produce crystals with poor x-ray diffraction properties and have to be optimized. Different strategies have been developed to assist in this process in a high throughput scale, however the traditional approach consisting in the elaboration of multi-well optimization screens where the concentration of one or several components of the crystallization solution are varied systematically remains a common approach. Commercial liquid-handling robots can be used to generate multi-well customized optimisation screens from a limited number of component stock solutions and this type of equipment is often present in crystallization laboratories. However the varied nature of the optimization screens results in the need to reprogram the liquid-handling robot for each specific experiment, which requires a significant amount of time and manpower. A commonly used approach to overcome this problem is to introduce constrains in the design of the optimisation screens so that a limited number of programs can be used to satisfy many conditions.

Above: The OptiCrys software for the design of optimization experiments (J.  appl Crys 39, 446-453.).

Software to assist crystallographers in the design and reporting of optimisation screens, such as OptiCrys developed as part of the E.C.-funded BioxHit project by the Institut de Biologie Structurale Jean-Pierre Ebel (IBS, Grenoble, France) with the support of CEA/DSV (Céline Charavay, GIPSE, Grenoble) exist. OptiCrys provides a graphical user interface for the generation of crystallization screens without any limitation in any plate format and provides convenient tools like, for example, the automatic calculation of component gradients. OptiCrys outputs the volumes of reagents to be delivered at each well position as well as a table of individual component concentrations to facilitate reporting. OptiCrys, is freely available for academic researchers and is in use in many crystallography laboratories. However, its use in a high throughput context was limited by the lack of an output format that could be directly read by liquid-handling robots.  In order to address this problem the GIPSE group of the IBS in collaboration with the High Throughput Crystallization Laboratory (HTX lab) of the EMBL Grenoble outstation has developed OptiLink; a new software module that takes the output from OptiCrys and generates files that can be directly read by a Tecan Robot, ridding of the need to reprogram the liquid-handling system. OptiLink can be adapted to a variety of robot configurations through a simple configuration file and allows to specify different liquid classes. The combined use of OptiCrys and OptiLink can contribute to save time and manpower facilitating significantly the optimisation of crystal diffraction properties. Unlike other existing programs OptiCyrs and OptiLink are freely available to academic users under a LGPL license. For further information visit the IBS ( or HTX lab web sites (

Delphine Blot (IBS) and José A. Márquez (EMBL Grenoble)


Mammalian Expression Platform at Oxford: Implementation of robot

The P-CUBE Oxford Mammalian Expression platform has successfully implemented a CompacT SelecT cell culture robot for large-scale expression of multiple target proteins in a transient format. Successful protocols for automated cell culture and transient transfection of human embryonic kidney (HEK), 293T and 293S GnTI- cells in various flask formats have been designed. The ability of the robot to handle 10 layers of HYPERFlasks (High Yield PERformance Flask, surface area 1720 cm2) further enhanced the capacity for high throughput protein production.

Protein yields obtained by this method were similar to those produced manually, with the added benefit of reproducibility. Automating cell maintenance and transient transfection allows the expression of high quality recombinant protein in a completely sterile environment with limited support from a cell culture scientist. This automated method for large scale transient transfection is offered as a Europe-wide service via the P-CUBE initiative.

The CompacT SelecT robot consists of two compartments (Fig 1a). The first is a humidified 5% CO2, 37°C incubator/flask hotel unit (Fig 1b) with space for 40 new input flasks and 90 production flasks for cell maintenance and transfection.  The second is a laminar flow compartment containing a Stäubli robotic arm (Stäubli Robotics, Faverges, France) (Fig 1c), a pipette head (Fig 1d), a waste receptacle and cocktail bar (Fig 1e).  The laminar flow compartment also has a barcode reader to track flasks and a flask de-capper (Fig 1f). The unit also incorporates a Cedex automated cell counting module (Fig 1g). Cells are automatically maintained by regular harvesting and seeding of T175 flasks, while production is performed in Triple flasks and HYPERFlasks, following the flowchart described in Fig 1h.

Data from HYPERFlask transient transfections performed on the CompacT SelecT robot over a period of four months, using both HEK 293T and HEK 293S GnTI- cells and using only constructs confirmed to be secreted in small-scale experiments, revealed a broad range of pure, crystallization grade, protein yields (Fig. 2). An overall trend of higher expression levels in HEK 293T (Fig. 2a) versus HEK 293S GnTI- cells (Fig. 2b) was observed, but in all cases the protein quantities obtained were sufficient to set extensive crystallization trials using nanolitre scale technologies.

Yuguang Zhao (University Oxford)