Evolutionary and structural determinants of transthyretin’s cryptic protease function

Transthyretin (TTR) is a β-rich tetrameric protein best known for thyroid hormone distribution and for its involvement in amyloid diseases. Beyond these roles, TTR has been reported to act as a cryptic protease in vitro, cleaving substrates such as neuropeptide Y, apolipoprotein A-I, and amyloid-β (Aβ). This activity is described as metal-sensitive, since chelators reduce or abolish proteolysis, but its structural basis remains unresolved. Evolutionarily, TTR originated from 5-hydroxyisourate hydrolase (HIUase), an enzyme of purine catabolism. Thus, TTR provides a model for testing how an ancestral hydrolase scaffold was repurposed into a hormone-binding protein while potentially retaining or reconfiguring catalytic microenvironments towards a proteolytic behavior. Here, we investigate whether TTR’s cryptic protease function reflects ancestral catalytic retention or a secondary lineage-specific emergence from structural changes affecting metal responsiveness, channel accessibility, substrate binding, and tetramer stability. Vertebrate TTR sequences were analyzed using codon-based models, ancestral sequence reconstruction, and structural mapping. Extant and reconstructed proteins were structurally evaluated by probe-based hotspot mapping, normal-mode analysis, Aβ docking, and molecular dynamics to test how channel motions orient candidate residues and stabilize Aβ-bound poses. These simulations guide QM/MM modeling of candidate cleavage pathways. In parallel, recombinant EptatretusTTR, AncTTR, AncTetrapoda, and mutants are being produced for biochemical validation. Native Aβ cleavage will be assessed by Tricine gel electrophoresis, while proteolysis of a fluorogenic Aβ-derived peptide will be monitored by spectrofluorimetry under untreated, zinc, EDTA, protease-inhibitor, and heat-treated conditions. Molecular evolution analyses indicate strong structural constraint across TTR, with heterogeneous selective pressure but no evidence for pervasive diversifying selection across the TTR coding sequence, including proteolytic behavior. Branch-level shifts suggest limited lineage-specific tuning, while ancestral reconstructions do not recover the proposed metallopeptidase-like arrangement. Zinc-binding configurations likely represent lineage-specific and allosteric solutions, while the actual catalytic mechanism is yet to be characterized. Structural analyses instead support a distributed model in which ligandable channel regions, hydration, residue orientation, and conformational accessibility may govern Aβ engagement and possible peptide-bond cleavage, rather than solely the proposed protease motif region. Overall, ancient TTR protease potential, if present, is best framed as a staged, structurally constrained property of an evolutionarily remodeled HIUase scaffold.