Crystallographic Fragment Screening In Drug Discovery: Background & Advantages

Crystallographic fragment screening has emerged relatively recently as the most efficient biophysical screening method. Screening a fragment library against a drug target in fragment-based drug discovery (FBDD) is widely used at the early stages of discovery projects to identify new hits that can be developed into lead molecules. While lead-like compound libraries are expected to conform to Lipinski’s Rule of 5, fragments are thought to adhere to the Rule of Three. As discussed in an earlier post, this rule specifies that the molecules in the library should have a molecular weight of less than 300 Da, fewer than three hydrogen bond donors or acceptors, fewer than three rotatable bonds, and a lipophilicity value (logP) of less than 3.

The smaller size and fewer interactions with the target give the fragment an advantage of a higher probability of binding to the protein. However, this advantage also comes with a disadvantage: low binding affinity and specificity toward the drug target. Therefore, more sensitive biophysical screening techniques, such as NMR spectroscopy, surface plasmon resonance (SPR), or thermal-shift assays (TSA), were initially employed to detect fragment binding. Crystallographic fragment screening was also introduced as an alternative method. An X-ray structure enables direct observation of fragment binding while allowing for the analysis of the details of its interactions with the protein. This provides a clear advantage for the subsequent steps of hit expansion and lead generation. However, the low throughput of this method and technical difficulties have limited the application of crystallographic fragment screening at the initial stage. This changed with the arrival of high-flux synchrotron beamlines, fast X-ray detectors, and the automation of many experimental steps, which substantially increased the process throughput and enabled the collection of hundreds of X-ray data sets per day.

These developments have significantly lowered barriers for screening larger fragment libraries, leading to several synchrotrons now operating platforms for crystallographic fragment-based screening. Among these are the XChem facility at Diamond Light Source (UK), the Crystallographic Fragment Screening Center at Helmholtz-Zentrum Berlin at BESSY II (Germany), the screening centers at EMBL Grenoble (France), the platform at SLS (Switzerland), and, of course, the FragMAX facility at MAX IV Laboratory. Established in 2019, FragMAX is only a few kilometers from SARomics’ laboratory building. The platform offers a wide range of fragment screening services, including proprietary libraries, laboratory automation equipment, and diverse software solutions for crystal inspection, controlling the soaking process, data collection, data reduction, and initial structure refinement.

Case study: Comparison with Other Screening Techniques

In a paper by Gerhard Klebe’s group (Schiebel et al., 2016), titled “Six Biophysical Screening Methods Miss a Large Proportion of Crystallographically Discovered Fragment Hits: A Case Study,” the authors systematically compared six different biophysical fragment screening techniques with crystallographic fragment screening to evaluate how accurately biophysical methods predict fragment hits identified by crystallography. The methods analyzed included fluorescence-based high-concentration biochemical screens (HCS), saturation transfer difference NMR (STD-NMR), a reporter displacement assay (RDA), native mass spectrometry (MS), microscale thermophoresis (MST), and thermal shift assays (TSA). Of the 361 fragments in the screening library, crystallographic fragment screening against the aspartic proteinase endothiapepsin (EP) identified 71 hits. This corresponded to a success rate of 20%, higher than for any of the six alternative biophysical screening methods, which ranged from 2% to 17%. The results also demonstrated the superiority of crystallographic screening: for 44% of the crystallographic hits (31 of the 71), none of the six screening methods indicated binding. Moreover, 30% of the X-ray hits (21 of the 71 hits) were predicted by only one of all applied biophysical methods.

Although the use of focused libraries can increase success rates, the library used in this work was designed for general purposes and is not biased toward aspartic proteases. Interesting to note that the bound fragments sampled the complete EP binding cleft, comprising eight subsites S2′ through S6.

Practical Considerations

In contrast to typical protein crystallography work, which often requires only a few high-quality crystals, screening a fragment library requires several hundred crystals for fragment soaking. As stated by the FragMax team, “the crystallization system must ensure that crystals can be grown reliably and reproducibly in large quantities without needing to prepare tens of crystallization plates.” Additionally, because library compounds are typically dissolved in DMSO, the stability of the crystals in DMSO or another organic solvent, if required, must be tested before fragment soaking. Other requirements include using 3-lens SWISSCI sitting-drop crystallization plates, crystals must be 50 μm in at least one direction, be tolerant to handling, and have a packing that allows access to the ligand binding site. Crystals with different packing (e.g., different space groups) can be beneficial, as packing variation may enhance the accessibility of the ligand-binding site and reduce the incidence of so-called false negatives. The quality and resolution of the X-ray data must also be considered, as they determine the quality of the final X-ray structure. Clearly, considering the high value of synchrotron beamtime, all these requirements must be met before arriving at the screening facility. Kanchugal et al., 2025, recently published a more detailed account of the FragMax platform. The paper provides a thorough description of the platform components, including the instrumentation and software used at each stage of the process. The paper also includes practical advice for inexperienced users and information about beamtime applications. However, our clients don’t need to apply for beamtime.

Fragment Screening at SARomics Biostructure

The SARomics Biostructures CRO team has extensive experience in all forms of fragment screening services, including crystallographic fragment screening. At the recent (2025) Fragment-Based Lead Discovery Conference (FBLD Conference) in Cambridge, UK, SARomics Biostructures AB presented a new kinase-focused fragment library for crystallographic fragment screening. The library has been developed in collaboration with Jens Carlsson, Szymon Pach, and Philip Ullmann (Uppsala University), with the focus on generating a fragment library containing sociable and water-soluble fragments that can be rapidly expanded using analogues from the Enamine REAL library.

The designed kinase-focused fragment library, comprising 300 compounds, was tested against two kinases (CDK2 and an internal target) using the FragMax fragment screening platform, yielding hit rates of 13% to 16%. The hits were subsequently ranked using a combination of differential scanning fluorimetry (DSF) and free-energy calculations based on molecular dynamics (MD) simulations, followed by the selection and ranking of follow-up analogues using docking for hit expansion. Please download a poster for details.

You can also download a PDF presentation with more details about our company’s fragment screening services or contact us directly to discuss and plan your project. As a CRO, we operate within the framework of our agreement with MAX IV Laboratory, which means our clients do not need to apply for beamtime or for accessing crystallographic fragment screening facilities. For more information about our services, please visit our protein crystallography services page.