• Publié par : Christian DELAMARCHE
  • Imprimer cette page

Fibrillar protein aggregation

The misfolding of normal proteins and their fibrillar aggregation in diverse organs and tissues is associated with more than 20 human severe diseases, classified as amyloidosis or “conformational diseases”. It is becoming clear that short amino acid sequence stretches (hot spots) act as seeds for driving amyloid fibril formation. In our approach, we combine in silico and in vitro techniques to study the molecular mechanisms of the nucleation phase, correspondin g to the first phase of peptide assembly. Our studies includes the effect of chaperones and competitor peptides.

We have recently created AMYPdb, a specialized database dedicated to amyloid precursor proteins and to their molecular signatures (amypdb.univ-rennes1.fr). The 3,621 signatures stored in the database can be screened to highlight amino acids involved in the misfolding and aggregation steps of amyloid proteins. We will further develop the database toward two axes. First, in collaboration with J. Aguilar (University of Merida, Venezuela) we will define and develop a mathematical/computational method for the comparison of regular expressions. This constitutes an innovative approach that could have general implications in bioinformatics. The comparison of the 3,651 signatures of AMYPdb against themselves could reveal conserved patterns and common rules between unrelated amyloid families. The second axis is based on the development of methods to predict hot spots directly from the analysis of the protein sequences. We will implement six to ten of the best known algorithms to calculate a consensus score of aggregation supported by statistical analysis. We think that the development of a meta-predictor, directly associated with AMYPdb, will produce better results and will be more useful than individual methods accessible on distinct web sites.
For the in vitro experiments, we will focus on proteins involved in kidney diseases. Indeed, kidney is one of the most frequent targets which accumulate aggregates of insoluble proteins. Our major models will be Immunoglobulin light chain and fibrinogen A-alpha in order to benefit of the clinical expertise of collaborators in Rennes (CHU and “Centre de Compétence des Amyloses Bretagne”) and to study the effect of known pathological mutations in collaboration with the genotyping team of S. Valleix (Institute Cochin, U1016, Paris). The hot spot regions, as well as protective mutations will be predicted starting from the in silico experiments. Short synthetic peptides, or polypeptides produced in bacterial system, will be analyzed with a variety of methods commonly used to study amyloid fibril formation; e.g. Congo Red assay, sedimentation velocity profiles and transmission electron microscopy. These in vitro experiments could help to design and test peptide-based inhibitors of the nucleation phase.