Use of AlphaFold 2 to predict stabilizing mutations for the R337H variant in the tetramerization domain of TP53.

Abstract

e15118 Background: TP53 mutations occur in nearly 50% of human cancers, profoundly affecting protein structure and function. The R337H variant is the most common germline mutation in TP53, associated with Li-Fraumeni syndrome. Unlike most TP53 mutations, which occur in the DNA-binding domain (DBD, residues 102–292), R337H is uniquely located in the alpha helix of the tetramerization (TET) domain (residues 325–355). This mutation destabilizing the protein tetramer and impairing p53’s tumor-suppressor function. Leveraging AlphaFold 2 (AF2), we aimed to identify stabilizing mutations that could restore function to the R337H variant. Methods: AF2 was used to generate structural predictions of wild-type and mutated TP53 variants as implemented in ColabFold. RMSD values (Å) were calculated using PyMOL version 2.6.1. PyMOL's alignment algorithms (align, super, and cealign) were applied to quantify deviations in both the DBD and TET domains. Mutations were considered stabilizing if they showed agreement across all alignment algorithms and achieved ≥40% reduction in RMSD. To validate AF2's predictive capability, we analyzed the pathological mutation R249S in the DBD and its known compensatory mutation H168R, which has been experimentally demonstrated to stabilize p53 function. For R337H, we systematically introduced amino acid substitutions across all positions in the TET domain (residues 325–355), generating 589 double-mutant variants (R337H plus one additional substitution). Results: AF2 accurately predicted stabilization of R249S by H168R, with RMSD reductions utilizing the super alignment method in the DBD from 1.3 Å to 0.4 Å (68.2%) and in the TET domain from 10.8 Å to 5.2 Å (51.5%), with similar results with the align and CEalign methods. For the R337H germline variant, the RMSD of the TET domain using the super, align, and cealign methods was 10.6 Å, 11.8 Å, and 12.5 Å, respectively. Among 589 double-mutant variants, five substitutions: A347W, E336C, E339I, E343T, and E349D, reduced TET RMSD by ≥40% across all alignment methods. A347W reduced RMSD from 10.6 Å to 6.2 Å (41.9%), from 11.8 Å to 5.0 Å (57.7%), and from 12.5 Å to 6.1 Å (51.1%) in the super, align, and cealign methods, respectively. Similarly, E336C reduced RMSD from 10.6 Å to 5.4 Å (48.9%), from 11.8 Å to 5.5 Å (53.7%), and from 12.5 Å to 7.5 Å (40.2%). Comparable reductions were observed for E339I, E343T, and E349D. RMSD deviations in the DBD domain remained low ( < 1.5 Å across methods) for all analyzed variants, with no destabilizing effects observed Conclusions: AlphaFold 2 predicts five stabilizing mutations for the R337H germline variant which stabilize the TET domain while preserving the integrity of the DNA-binding domain. These findings provide candidates for experimental validation and clinical application. Gene editing techniques such as CRISPR-CAS9 are especially favored due to lack of functional TP53 interference.