Structurele Biologie van de steroïde Receptoren met als doelstelling rationeel medicijn en proteïne ontwerp

Steroid receptors (SRs) are natural genetic switches that can turn genes "on" and "off". They function as transcription factors (TFs) that are activated through binding of steroid hormones. All the SRs share the same architecture with an amino-terminal domain (NTD), a ligand binding domain (LBD) and a DNA binding domain (DBD). Because of their many roles in normal physiology and diverse pathologies, SRs are important drug targets. Because of their well-known structure, they can also be further engineered into tunable synthetic biology tools. The androgen receptor (AR) is an important research topic because it is a pivotal target for prostate cancer (PCa) treatments, due to its role in promoting cancer development and progression. The LBD is the most characterized domain, with plenty of experimental information on its structure, how it binds ligands and how it recruits coregulators. All current AR antagonists that are used in the clinic target the LBD. Unfortunately, they induce resistant forms of PCa which is still dependent on AR, but unaffected by these antagonists. We contributed to three strategies to come to alternative AR antagonists. Targeting the DBD is a first option considering its function in DNA recognition and DNA-dependent dimerization, which are believed to be essential in the androgen regulation of transcription. A deep understanding on the structure-function relationships within the DBD is a prerequisite for the development of drugs that would target it. We have obtained high resolution structures based on biophysical, biochemical and X-ray crystallography studies of the human AR-DBD bound to two natural androgen response elements, MTV and C3 (1) ARE. We confirmed that the AR-DBD favors binding in a head-to-head binding mode. Moreover, the first α-helix of the DBD, containing the P-box, recognizes and binds to the DNA major groove. Superimposition of the two novel structures on the previously published direct repeat structure revealed a bending of the second DBD monomer towards the second hexanucleotides which is much more prominent on these natural AREs. The structural evidence suggested that the AR adopts a special recognition modality in which the first DBD monomer binds to the first hexanucleotide with high affinity, whereas the second DBD monomer binds cooperatively to the second hexanucleotides. Our structural studies also provide insight in the effects of the Androgen Insensitivity Syndrome (AIS) mutations on the structural integrity, specific protein-DNA interactions, and DNA dependent dimerization of the DBD. The second option to come to better antagonists is the LBD, which remains an important site for the design of next-generation antagonists. Unfortunately, no antagonistic form of the AR-LBD has been reported, which obstructs a more refined search for novel antagonists. We hypothesized that engineering the crystal packing of the AR-LBD with crystal adjuvants, such as POMs and calixarenes may help to space up the crystal packing framework and allow for example the repositioning of the helix 12 to antagonistic conformations. Using Biolayer Interferometry (BLI), we assessed the binding of these adjuvants to the AR-LBD. Next, we attempted co-crystallization of the POM or calixarenes to the AR-LBD. We obtained many AR-LBD crystals that diffracted well. Unfortunately, no protein-POMs or calixarenes hybrid crystals were found. As a third strategy, we attempted crystal soaking and co-crystallization of novel in-house developed compounds targeting the dimerization site of the AR-LBD. These compounds bind the AR-LBD in BLI. Unfortunately, we did not obtain AR-LBD-DIMs crystals, most likely because of the limited access of the DIMs in the agonistic crystal packing of AR-LBD. Finally, apart from the work on the AR protein related to drug discovery, we also performed biophysical characterization of the human estrogen receptor (ER) protein which was engineered into tunable switches responsive to a non-natural agonist (chemical ligand) of choice. By using protein engineering, new ER variants were created, such as preTERRA and TERRA engineered towards tamoxifen agonism, as well as preBASE and BASE, engineered towards bisphenol A (BPA) [1]. By using differential scanning fluorimetry, we assessed the thermal stability of the WT and ER variants in presence of different ligands. In general, the addition of the ligands improved the structural stability of the LBD, which is reflected in their melting temperature (Tm) by thermal shift assay. We confirmed that the TERRA is more stabilized by tamoxifen then by estradiol. The same trend was observed for the BASE, with a higher stabilization by the BPA analogue BPZ. In conclusion, by obtaining the high-resolution human AR DBD crystal structures bound to different natural response elements, we have gained critical insights into the roles of this domain in AR DNA binding, dimerization, cooperative binding mechanisms, and the effects of AIS mutations on these. Furthermore, these structural insights into the DBD can provide valuable knowledge for developing potential therapeutic targets for prostate cancer to overcome AR resistance by mutations in the LBD. Moreover, we have provided technical and feasibility insights on the introduction of adjuvant compounds into native structures. These can aid in risk mitigation for future crystallography applications in similar studies and experimental setups for challenging projects. Lastly, the structural studies of new receptors developed through protein engineering enhances our understanding of the action mechanisms and ligand effects in these engineered proteins. In the absence of structures, this provides valuable insights into protein folding, stability, and functional dynamics, supporting the further design of stable therapeutic proteins for future applications.