(588f) Allosteric Communication in Transthyretin: Identification Using Statistical Coupling Analysis
Transthyretin (TTR) is a 55 kDa homotetrameric protein that serves as a transport protein of thyroxine and retinol in blood and cerebrospinal fluid. Thyroxine, a thyroid hormone, binds in a central hydrophobic channel of the TTR tetramer. Retinol binds to the 21 kDa retinol binding protein RBP, and then the retinol-RBP complex interacts with residues in the solvent-exposed EF helix of TTR.
TTR binds to the Alzheimer’s related peptide beta-amyloid (Aβ) and protects against Aβ neurotoxicity, both in vivo and in vitro. Previously, we identified two binding domains on TTR where Aβ interacts: strand G, and the EF helix. Curiously, these two binding domains overlap strongly with the two natural TTR ligand-binding domains. We showed specific involvement of two residues, L82 on the EF helix and L110 on strand G, in TTR- Aβ binding. Mutation of either of these residues to alanine led to both a loss of binding and a loss of neuroprotection.
TTR is normally a very stable tetramer, and the strand G binding site lies in the interior of the tetramer and is relatively inaccessible to larger Aβ oligomers. Thus, we wondered how Aβ oligomers gain access to this binding site. Our collected data supports a model wherein binding of Aβ to the external EF helix induces destabilization of the TTR tetramer, and exposure of the strand G binding site. Such a model suggests the hypothesis that there is allosteric communication between the exterior EF helix and interior residues involved with TTR tetramer formation and stability.
To test this hypothesis, we used a method called statistical coupling analysis (SCA) to search for protein sectors in TTR that could serve as a conduit for allosteric communication. Briefly, all known sequences of transthyretin and transthyretin-like proteins were aligned, and conservation of specific amino acids at each position was calculated. Then we looked for correlations in amino acid frequency between all pairs of positions in the entire sequence. The hypothesis is that a strong correlation in a particular pair of positions suggests evolutionarily conserved interactions between the residues in that pair. Once the entire matrix of positional correlations is found, spectral decomposition is used to identify sets of positions that constitute a protein sector. This analysis yielded evidence of a protein sector that indeed links L82 on the EF helix to a ‘stripe’ of residues on adjacent β-strands lining the hydrophobic interior channel of TTR. We experimentally tested this statistical analysis by site-directed mutagenesis of residues in the putative protein sector, and investigation of the effect of these mutations on TTR tetramer stability, and the ability of Aβ to induce destabilization. Finally, we provide direct evidence that RBP and Aβ compete for binding to the EF helix, and that RBP binding to TTR prevents both Aβ-mediated destabilization of TTR tetramers, and TTR-mediated inhibition of Aβ aggregation.