Sofie Vincents Al-Saoudi:
Ubiquitin in signaling and protein quality control

Date: 15-10-2016    Supervisor: Rasmus Hartmann-Petersen



Protein ubiquitylation is an important post-translational modification that holds a variety of cellular functions. This Ph.D. thesis is comprised of two studies, of which one focused on ubiquitylation related to inflammatory signaling, and the other on the role of the ubiquitin-proteasome system in degradation of misfolded proteins.

In the first study, the interaction between a ubiquitin-protein ligase and a deubiquitylating enzyme is described. The linear ubiquitin chain assembly complex (LUBAC) is a ubiquitin-protein ligase complex that exclusively assemble polyubiquitin chains linked by the N-terminal methionine (M1), and recently, the deubiquitylating enzyme, OTULIN, was discovered to counter LUBAC activity by exclusively cleaving M1-linked ubiquitin chains. We provide the molecular detail of the interaction between the LUBAC subunit, HOIP, and OTULIN. The interaction was mapped to the PUB-domain of HOIP and a fragment termed the PUB-interacting motif (PIM) in OTULIN. The crystal structure of the complex was solved and explained the exclusive specificity of the OTULIN PIM for the HOIP PUB domain. Functionally, HOIP mutants inept in OTULIN binding display increased autoubiquitylation to the same degree as is observed when depleting cells of OTULIN. Also, OTULIN-mediated inhibition of NF-κB activation was less pronounced when signaling was induced by a HOIP PUB-mutant compared to wild type HOIP. Thus, the specific interaction that we characterize structurally also appears to be important for the cellular functions of LUBAC and OTULIN.

In the second study, a combined in silico and experimental approach was undertaken to investigate the use of structural stability calculations in predicting the metabolic stability of a protein. As a model protein we selected the human mismatch repair protein MSH2 which is related to the cancer-predisposition disease, Lynch syndrome. Of 24 different MSH2 variants, some of which have been linked to Lynch syndrome, we show that there is a strong correlation between the predicted structural stability and the protein half-life. We show that a predicted destabilization of 3 kcal/mol is sufficient to cause proteasomal degradation of MSH2 variants. Importantly our calculations can, in addition to protein turnover, also predict pathogenicity of MSH2 variants, suggesting that this approach can be applied for Lynch syndrome diagnosis, and perhaps for other hereditary diseases.