Why Would a Monomeric Ig Domain Form Amyloid Fibers?

Abstract

Antibody light chains are composed of two domains, variable and constant, that associate with heavy chains to form functional antibodies. In a collection of Bcell diseases, an excess of light chains is produced. These light chains form various types of aggregates, amongst which are amyloid fibers. In this case a lethal multi-organ disease called light chain amyloidosis, or AL, ensues. Recombinant variable light chain domains, corresponding to proteins isolated from patients with AL, are structurally indistinguishable from those of healthy donors. Many happen to be less stable thermodynamically, leading to a hypothesis linking poor stability to amyloidogenesis. The native state will not aggregate into fibers, so an unfolded intermediate of unknown conformation has been proposed as the fiber building block. We propose that monomeric amyloidogenic Ig domains may form amyloid fibers by combining any of the following non-exclusive attributes: a) low thermodynamic stability, which favors local or global denaturation; b) low structural diversity in the conformational ensembles sampled upon denaturation, which allows for more frequent population of aggregation-prone monomers; c) sets of non-native interactions that selectively stabilize aggregationprone monomers; d) multiple viable polymerization surfaces, to maximize the probability for seeding and extension of fibers; and e) stable protofibers that can live long enough to progress to fibers. We test the relative relevance of these five contributions for a set of four light chain variable Ig domains belonging to class D6a, one of the most common in AL, by performing MD simulations under native and high temperature conditions, docking of putative unfolding intermediates, and MD of the protofibers. For the most amyloidogenic variant derived form a patient (WIL), preliminary data suggest that it uses all five attributes. Acknowledgements: computing at KanBalam (UNAM) and Centro Nacional de Supercómputo (IPICyT); funding from CONACyT 133294.