Mutation effect on conformational sampling of Annexin A11

Several neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS), are associated with the formation of aberrant protein aggregates. Annexin A11 has attracted therapeutic interest because it contains an intrinsically disordered region (IDR) that modulates liquid–liquid phase separation. Mutations within this IDR promote molecular aggregation, primarily affecting neurons and skeletal muscle. We investigated the conformational profile of the protein before and after pathological mutations, aiming to clarify aggregation mechanisms.

BioEmu, a generative AI model, was used to build conformational ensembles for wild-type human Annexin A11 and for the P36R, G38R, D40G and D40Y mutants. BioEmu recovers fundamental properties such as residue flexibility, motion correlations and local residue contacts. Two independent replicate runs generated 5,000 frames each, with filtering disabled, using model v1.2. A pipeline enabling local, global and comparative analyses was built, including RMSF, PCA, and contact maps. Secondary structure was characterised with DSSP to assess helical, β-sheet and coil propensities.

Comparative PCA did not discriminate the global conformational profiles of wild-type and mutant ensembles, reflecting BioEmu's limitation in predicting large conformational changes arising from single point mutations. However, preliminary DSSP comparisons across replicas indicate a consistent ~15% increase in helical propensity over residues 30–40 in P36R relative to wild type. This local stabilisation was accompanied by a ~17% increase in contact frequency, both between IDR residues 39–46 and residues 240–252 of annexin repeat I, and between residues 189–193 and residues 465–468 of annexin repeat IV. The remaining three mutants also deviated from wild type in these same regions and at comparable magnitude, although the direction and extent of these changes are still being quantified.

The results suggest that P36R shifts the interaction equilibrium between the IDR and the structured domain by altering the helical pattern of residues 30–40. The convergence of all four ALS-associated mutations onto the same regions points to a shared perturbation mechanism, which we are currently characterising. Understanding these mechanisms is essential for engineering binders that modulate Annexin A11 dynamics and prevent pathological aggregation.

This work was supported by the São Paulo Research Foundation (FAPESP # 2026/00945-5).