Abstract:
Half a century ago thalidomide caused one of the biggest pharmaceutical tragedies known by now. It was widely prescribed to pregnant women as a sedative, but displayed teratogenic properties, causing limb malformations and other developmental defects in more than 10,000 babies worldwide. Nevertheless, thalidomide and its derivatives are used nowadays in treatment of leprosy and multiple myeloma. The protein cereblon was identified as a primary target of thalidomide in the cell. As a substrate receptor, cereblon (CRBN) is linked via the damaged DNA binding protein 1 (DDB1) and cullin 4A to the E3 ubiquitin ligase machinery. The drug binds to the C-terminal region of cereblon, also referred to as CULT domain (cereblon domain of unknown activity, binding cellular ligands and thalidomide). This domain represents the most conserved part of cereblon and is also found solely in single-domain proteins in bacteria and, as a secreted form, in eukaryotes. Based on its ligand binding properties, its high degree of conservation and its intracellular as well as extracellular localization, a common interest arose in understanding the functional role of the CULT domain in vivo.
The CULT domain carries a number of highly conserved cysteine and tryptophan residues within its amino acid sequence. In the solved crystal structure of the bacterial CULT domain, four conserved cysteines stabilize the protein fold by coordinating a zinc ion. Three invariant tryptophan residues build an aromatic cage, to which the ligand binds. Considering the structural similarity of uridine and thalidomide, we tested uridine binding to the hydrophobic pocket and could show an identical mode of binding. So far, uridine represents the only natural ligand, for which an interaction with the CULT domain has been shown. Further studies demonstrated that parts of the CULT domain fold upon ligand binding, thus stabilizing the protein. The pocket is highly similar to the aromatic cages found in histone readers that recognize methylated lysine or arginine residues in chromatin. This structural similarity suggests analogous ligands for cereblon, which include a distinct type of post-translational modifications.
We developed a FRET-based in vitro assay for testing and characterizing ligand binding to cereblon. The determination and comparison of the substrate affinities for three CULT domains, from Homo sapiens, Magnetospirillum gryphiswaldense, and the secreted Caenorhabditis elegans, revealed similar values for the same ligands on a relative scale. This study convincingly confirmed the bacterial protein as a robust model system for both: (i) testing the specificity of the CULT domain to its ligands; and (ii) deciphering the function of cereblon proteins. Using the FRET-assay, we showed that various therapeutically relevant pharmaceuticals display affinities to the CULT domain with binding constants in the micromolar range. These off-target effects were further validated by applying an in vivo assay in zebrafish embryos to test the teratogenic potential of these compounds mediated via their interaction with cereblon.
Searching for proteins interacting with cereblon in vivo, we identified transcription termination factor Rho to bind to the bacterial CULT domain. The assumption of a possible role of cereblon in transcriptional regulation is further supported by the fact that hCRBN interacts with LSD1, a demethylase which is essential for transcriptional regulation in multicellular organisms. Taken together, our data imply a potential function of cereblon in transcriptional repression and/or activation, particularly during limb formation.
Print-on-Demand