Conformational catalysis of cataract-associated aggregation in a human eye lens crystallin occurs via interface stealing

Abstract

Human γD-crystallin (HγD) is an abundant and highly stable two-domain protein in the core region of the eye lens. Destabilizing mutations and post-translational modifications in this protein are linked to onset of aggregation that causes cataract disease (lens turbidity). Wild-type HγD greatly accelerates aggregation of the cataract-related W42Q variant, without itself aggregating. The mechanism of this "inverse prion" catalysis of aggregation remained unknown. Here we provide evidence that an early unfolding intermediate with an opened domain interface enables transient dimerization of the C-terminal domains of wildtype and mutant, or mutant and mutant, HγD molecules, which deprives the mutant's Nterminal domain of intramolecular stabilization by the native domain interface and thus accelerates its misfolding to a distinct, aggregation-prone intermediate. A detailed kinetic model predicts universal power-law scaling relationships for lag time and rate of the resulting aggregation, which are in excellent agreement with the data. The mechanism reported here, which we term interface stealing, can be generalized to explain how common domain-domain interactions can have surprising consequences, such as conformational catalysis of unfolding, in multidomain proteins. SignificanceMost known proteins in nature consist of multiple domains. Interactions between domains may lead to unexpected folding and misfolding phenomena. This study of human γD-crystallin, a two-domain protein in the eye lens, revealed one such surprise: conformational catalysis of misfolding via intermolecular domain interface "stealing." An intermolecular interface between the more stable domains outcompetes the native intramolecular domain interface. Loss of the native interface in turn promotes misfolding and subsequent aggregation, especially in cataractrelated γD-crystallin variants. This phenomenon is likely a contributing factor in the development of cataract disease, the leading worldwide cause of blindness. However, interface stealing likely occurs in many proteins composed of two or more interacting domains.

A native chemical chaperone in the human eye lens

“…The N-terminal domain of HγD derives part of its stability from the domain interface (which may be disrupted over time, e.g., by deamidation or truncation 44,45 ), so the isolated domain forms aggregates at slightly elevated temperatures with oxidation. 23 Inositol suppressed aggregation of the wild-type N-terminal domain of HγD ("N-only") like that of the other variants. Aggregation of N-only and W130Q, like that of W42Q, was entirely redox-dependent (Figure1-supplement 3), so the aggregation precursor conformations adopted by all these variants are likely kinetically trapped by non-native disulfide bonds.…”

“…Note that we have previously shown the existence of a master curve that links the lag time, rate, and end-point extent of aggregation. 31…”

“…In this concentration range, myo-inositol already had a substantial effect in our assay, suppressing the rate of turbidity development by ~35% and increasing the lag time by a similar amount (Figure 1D), consistent with the "master curve" relationships between maximum rate, lag time, and endpoint turbidity. 23 It has been estimated that delaying the age of onset of cataract by ~14% would halve the global disease burden; 29 conversely, the near-complete loss of myo-inositol in cataractous lenses 39,40 could be a major factor increasing that burden. Less-abundant lens metabolites, including scyllo-inositol and glucose, likely contribute to aggregation suppression.…”

“…7 We have previously shown that the concentration-dependence of W42Q aggregation rate is quadratic, indicating a bimolecular rate-limiting step. 15,23 The proposed model in Figure 6C predicts that, by inhibiting this rate-limiting step, myo-inositol could alter the concentration dependence of W42Q aggregation in two ways. First, if it makes a dimer less likely to contribute to aggregation, then it should decrease the pre-exponential factor.…”

“…Fortunately, we have shown that solution turbidity is proportional to total mass for these aggregates 15 and that simple power laws relate the aggregation rate and lag time to protein concentration, such that all aggregation traces collapse onto a single master curve, where all points that map onto a given master-curve point have the same particle size distribution. 23 Despite intense research interest, [24][25][26] no preventative or therapeutic drugs against cataract have been approved to-date, leaving surgery as the only treatment option. Aggregate costs of cataract surgery are high in high-income countries, 27 while availability and outcomes are often poor in low-and middle-income ones.…”