Back

Perilipin 4 repetitive region forms amyloid fibrils promoted by a genetic expansion found in myopathy patients

Choufani, C.; Fuggetta, N.; Sabbagh, B.; Moulin, C.; Franckhauser, C.; Lyonnais, S.; Perme, T.; Arteni, A.-A.; Pagnotta, S.; Rouquier, S.; Ruggieri, A.; Bousset, L.; Copic, A.

2025-06-04 biochemistry
10.1101/2025.06.03.657256 bioRxiv
Show abstract

Lipid-binding amphipathic proteins play key roles in cellular organization, yet their conformational flexibility makes them prone to aggregation. Here we study the aggregation of perilipin 4 (PLIN4), a protein of the lipid droplet (LD)-binding perilipin family. PLIN4 contains an unusually long ([~]1000 amino acid, aa) amphipathic helix composed of tandem 33-aa repeats, which folds on the LD surface. A genetic expansion that increases the number of repeats is associated with skeletal muscle dysfunction in patients suffering from a late-onset vacuolar myopathy (MRUPAV). We show that the genetic expansion and the precise sequence of the 33 aa repeats do not significantly alter PLIN4 interaction with LDs in cells. By contrast, purified fragments of the PLIN4 repetitive region form amyloid fibrils as characterized by cryo-EM and atomic force microscopy. Repeat expansion strongly accelerates fibril formation, consistent with enhanced self-assembly driven by the accumulation of identical repeats. PLIN4 aggregation is attenuated by the presence of LDs, indicating that lipid binding modulates the balance between membrane association and self-assembly. Together, these results establish PLIN4 as a previously unrecognized amyloid-forming protein and suggest that repeat expansion enhances amyloidogenic self-assembly, providing a foundation for future studies of its role in degenerative muscle disease.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.