Comparative Mutational Analysis of AI-Predicted CIRBP Structures Reveals Conserved and Species-Specific Stability Landscapes in Humans and Long-Lived Whales
Year: 2026
Type: conference-paper
Abstract: AI and bioinformatics offer new tools for identifying proteins that support healthy aging. Cold-inducible RNA-binding protein (CIRBP) is a stress-responsive protein involved in RNA regulation and DNA double-strand break repair, recently linked to enhanced ge-nome maintenance in the long-lived bowhead whale. We combined comparative sequence analysis, Al-phaFold structure prediction, and computational satu-ration mutagenesis to investigate conserved and spe-cies-specific stability determinants of CIRBP. Multi-ple sequence alignment revealed strong conservation across six mammalian CIRBP orthologs, with limited species-specific variation. FoldX saturation mutagen-esis revealed broadly similar ΔΔG profiles across all four species, indicating conserved energetic con-straints. The mean ΔΔG profiles of human and bow-head whale CIRBP were positively correlated (r = 0.7251), supporting overall conservation of muta-tional stability patterns. G30 was particularly sensi-tive to mutation across all four species, with G30W among the strongest destabilizing substitutions. In contrast, G122, G124, and residue 126 exhibited spe-cies-dependent mutational effects. Structural compar-ison of human and bowhead whale AlphaFold models showed high overall similarity (RMSD = 0.332 Å), with the N-terminal RNA-recognition motif more con-served than the flexible C-terminal region. These findings show that combining AI-predicted structures with mutational stability analysis can reveal energetic constraints not evident from sequence conservation alone, identifying candidate residues for experimental follow-up.
Authors: Christina Teng, Malcolm Anthony
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: diamond
OpenAlex ID: W7225432851
