Almost all cellular processes depend on protein-protein interactions (PPI) which therefore are promising targets for drug design or tool compound development. PPIs as targets have the advantage that the interaction surfaces are highly specific and unique, meaning potent and selective compounds are less likely to cause side effects compared to targeting enzyme activities. However, targeting PPIs is usually challenging due to their large interaction surface and shallow pockets. It is important for such surfaces to find hotspots where compounds engage in strong interactions with the protein. Finding hotspots can be achieved by screening fragments (organic compounds < 300 Da), which form specific interactions with the target. These fragments can be used as starting points for compound development. In order to identify fragments bound to a protein, crystallographic fragment screening [1] is a great tool as it not only informs about the presence of a fragment, but also about its 3D position towards the protein. Thus, enabling structure-guided drug design or tool compound development for any target that can be crystallized reproducibly and gives diffraction data to sufficient resolution.

We chose the spliceosomal protein-protein complex of yeast Aar2 and Prp8^RNaseH (AR) [2], to screen our novel over 100-membered F2X-Universal Library [3]. The spliceosome is a large cellular machine that depends on dynamically changing PPIs throughout its functional cycle and thus is an ideal object of study for PPI compound development. Prp8 is highly conserved in the spliceosome and acts as a scaffolding protein. Prp8 is shuttled into the nucleus via a shuttling factor Aar2. Utilizing the complex for our screen has the advantage to identify fragments bound close to the interface of AR while also probing the surfaces of Prp8^RNaseH and Aar2 at the same time. That means information can be gathered to either develop an enhancer for the interaction of AR or inhibitors for other interactions Prp8^RNaseH is involved in or to find potentially interesting sites on Aar2. In previous work a representative subset of the F2X‑Universal Library, the F2X‑Entry Screen (96 fragments) was screened against AR, which already resulted in 20 unique hits that hinted towards possible hot spots [3]. However, in order to verify the found hotspots as such, to potentially find more hotspots and to confirm the found binding motifs of the fragments the full F2X‑Universal Library, i.e. 1013 compounds in total, were screened. For the automatic data analysis FragMAXapp [4] in combination with cluster4x [5] was utilized and this resulted in 278 unique hits scattered across the two proteins highlighting hotspots on their surfaces. This translates into a hit rate of 27,5%, which is the highest hit rate for such a large screen to the best of our knowledge and successfully validated the library. The additional hits found by screening the complete F2X‑Universal Library validated certain binding modes presented by the F2X‑Entry Screen and provided novel hotspots. This way we gained a large number of interesting starting points to develop potential modulators for several PPIs inside the spliceosome.

The hypoxia mediated prolyl-hydroxylase isoforms 1-3 (PHD1-3) are members of the Fe(II)-/2-oxoglutarate (2OG)-dependent oxygenase superfamily of enzymes. PHD1-3 catalyse the trans-4-prolyl-hydroxylation of the labile hypoxia-inducible factor-α subunit (HIFα) in the presence of cofactors; Fe(II), 2OG, ascorbic acid, and molecular dioxygen. The hydroxylation of the oxygen degradation domain (ODD) of the HIF1-3α substrates marks the transcription factor for degradation via the ubiquitin-E3 ligase-28S proteasome pathway. In hypoxic conditions, PHD1-3 are less active and the HIFα subunits translocate into the nucleus and form an α,β-heterodimer with the stabile HIFβ subunit that subsequently leads to the genetic cascade in response to hypoxia, e.g. upregulation of erythropoietin (EPO) and vascular endothelial growth factor (VEGF). The inhibition of the PHDs and specifically PHD2, the most commonly expressed isoform, have been the focus for small-molecule based therapies to treat individuals with blood disorders, such as anaemia and ischaemia-related disorders. Recently, several PHD inhibitors have been approved; Roxadustat (Japan, Chile, and China), Daprodustat (Japan), Molidustat (Japan), Enarodustat (Japan), and Vadadustat (Japan). Co-crystal structures of some of the clinical inhibitors have been challenging to obtain due to heterogeneity induced through ligand complexation. A clinical inhibitor complex structure with PHD2 has been obtained using a novel crystallisation system utilising succinate, co-product of the PHDs reaction, as a crystallisation tool to stabilise the active site during incubation with Molidustat, the related compound IOX4, and monodentate binding inhibitor Takeda-17. The novel crystal form was then further used for inhibitor soaking and scale-up of crystal growth (~500 crystals) for an XChem fragment-based screen (I04-1, Diamond Light Source) in search for allosteric binding sites. To date, the crystallisation system yielded 3 inhibitor co-crystal structures, 8 complex structures through soaking, and 5 hits from the fragment screen.

Fragment-based lead discovery (FBLD) is by now an established drug development strategy, which has so far delivered four novel drugs and more than 40 additional molecules in clinical trials. Starting points for FBLD are usually found by biophysical screening of fragment libraries with several hundred and up to a few thousand compounds. Crystal-based fragment screening has become increasingly popular over the last few years, facilitated by automation, improvements in beamline instrumentation and software development. The FragMAX facility provides a new user platform for crystallographic fragment screening at BioMAX, the first operational beamline for macromolecular crystallography at MAX IV Laboratory. This presentation will provide an overview of the different components of the FragMAX facility and describe the screening process. It will highlight different modes of access and outline planned developments.
The FragMAX platform started serving external users in 2019 and has since established an international user program that is open to academic and industrial research groups. The platform consists of three main components: (i) a crystal preparation facility, (ii) diffraction data collection at BioMAX and (iii) FragMAXapp, an intuitive web-based tool for large-scale data processing. The facility provides free access to several fragment libraries, notably, the in-house developed FragMAXlib. The screening processes have been adjusted throughout the COVID-19 pandemic and our aim is to further develop the platform so that users with different levels of experience can routinely achieve actionable screening hits for their targets.
While the facility was conceptualized for crystal-based fragment screening, its design is entirely generic. It can therefore be used for other applications that require large-scale crystal preparation and potential new applications will be discussed at the end of the presentation.