Aβ Oligomer Dissociation Is Catalyzed by Fibril Surfaces

Abstract

Oligomeric assemblies consisting of only a few protein subunits are key species in the cytotoxicity of neurodegenerative disorders, such as Alzheimer's and Parkinson's diseases. Their lifetime in solution and abundance, governed by the balance of their sources and sinks, are thus important determinants of disease. While significant advances have been made in elucidating the processes that govern oligomer production, the mechanisms behind their dissociation are still poorly understood. Here, we use chemical kinetic modeling to determine the fate of oligomers formed in vitro and discuss the implications for their abundance in vivo. We discover that oligomeric species formed predominantly on fibril surfaces, a broad class which includes the bulk of oligomers formed by the key Alzheimer's disease-associated Aβ peptides, also dissociate overwhelmingly on fibril surfaces, not in solution as had previously been assumed. We monitor this "secondary nucleation in reverse" by measuring the dissociation of Aβ42 oligomers in the presence and absence of fibrils via two distinct experimental methods. Our findings imply that drugs that bind fibril surfaces to inhibit oligomer formation may also inhibit their dissociation, with important implications for rational design of therapeutic strategies for Alzheimer's and other amyloid diseases.

Keywords: Alzheimer’s; dissociation; fibrils; inhibitor; kinetics; oligomer; therapeutic.

Figure 1

(a) Coarse-grained reaction steps in amyloid oligomer and fibril assembly kinetics (rate constants in brackets). Nonfibrillar oligomers are formed by free association of monomers (“primary association”, k_o1) and subsequently undergo conformational conversion into fibrils (k_conv) that rapidly elongate via addition of monomers to their ends (k₊). However, the majority of oligomers dissociate back to monomers (k_d1), rather than convert to fibrils. In Aβ aggregation, nonfibrillar oligomers are also formed through association on the surfaces of fibrils (k_o2). For all but the lowest fibril concentrations, this occurs much more rapidly than the primary association of oligomers. However, the extent to which oligomers also dissociate at the surface of fibrils (k_d2) has not hitherto been determined. (b) Catalytic effect of fibrils on hypothetical single step and multistep oligomerization reactions. In the absence of fibrils, Aβ oligomer formation and dissociation are both slow. Oligomers are less thermodynamically stable than monomers. In the presence of fibrils, the relative thermodynamic stabilities of monomers and oligomers (ΔG) are unchanged, but oligomerization is greatly accelerated. In a single-step reaction, the only energy barrier present is reduced in height by the fibril surface catalyst. Therefore, both oligomer formation and dissociation are accelerated equally. In a multistep reaction, the forward and backward rates for the step that is catalyzed increased equally. However, if enough catalysis occurs, then the greatest energy barrier is now different. The free energy barrier ΔE^‡ dividing the species is reduced equally, but the Arrhenius-type prefactor will be affected differently in the forward and reverse directions. Thus, formation and dissociation of oligomers are still both accelerated by fibrils but to slightly different extents.

Figure 2

Two independent experiments with different methodologies confirm that fibrils catalyze oligomer dissociation. Fluorescently labeled and unlabeled monomeric protein with a molar ratio of 1:1.5 (a) or 1:3.5 (b) was incubated until c. 50% (a) or c. 95% (b) of monomers have formed fibrils. (a) For fluorescence correlation spectroscopy (FCS), all samples were then centrifuged to remove most fibrils, and unlabeled fibril seeds were added back in to some of the samples. (b) For microfluidic free-flow electrophoresis (μFFE), only some samples were centrifuged, with the uncentrifuged samples thus effectively containing labeled seeds. (c) FCS measurements of the mass concentration of oligomers and monomers were taken after approximately 5 min of incubating oligomers ± fibrils. (d) μFFE measurements of the number concentration of oligomers and monomers were taken after approximately 3 h of incubating oligomers ± fibrils. In both cases, relative labeled oligomer concentrations are much lower when fibrils are present (due to both a decrease in absolute oligomer concentrations and an increase in absolute monomer concentrations), showing that fibrils cause many more oligomers to dissociate in the first 5 min and that this effect holds under very different labeling conditions. The bars represent means over the experiment repeats. t-tests confirmed the significance of this decrease (p < 0.001 in both experiments, see Materials and Methods).

Figure 3

Fibril-mediated oligomer dissociation is consistent with the available kinetic data on amyloid proteins that undergo rapid secondary nucleation. Top: oligomer concentration and bottom: fibril concentration. Global fits are to kinetic models in which the great majority of oligomers dissociate on fibrils. (a) Aβ42 oligomer-mediated fibril formation kinetics (total monomer concentration m_tot = 5 μM) both in the presence and absence of a known inhibitor of oligomer formation through secondary nucleation (data taken from ref (24); most rate constants unchanged). (b) Aβ40 oligomer-mediated fibril formation kinetics (m_tot = 10 μM; data again taken from ref (24); most rate constants unchanged). (c) α-Synuclein oligomer-mediated fibril formation kinetics in both the presence and absence of 1 μM fibril seed (m_tot = 100 μM; data and rate constants taken from ref (31)).

Figure 4

Effect of oligomer-mediated dissociation on Aβ42 aggregation kinetics under constant-monomer conditions, simulated using rate parameters from ref (24). (a,b) Constant m = 5 μM and no removal of oligomers. In the absence of fibril-mediated dissociation, both would rise exponentially indefinitely (red, dashed). However, we have found that dissociation is actually fibril-mediated. As a result, oligomer concentration (green) instead plateaus, whereas fibril mass concentration (blue) rises initially exponentially but subsequently as t² once oligomer concentration plateaus. (c,d) Conditions closer to those expected in vivo with fast oligomer clearance mechanisms, i.e., m = 0.1 μM and k_cl = 25κ. No plateau is reached for several months. (e,f) Fibril-binding inhibitors must be added sufficiently early to reduce Aβ42 oligomer concentrations under constant-monomer conditions. Solid lines: no inhibition. Dashed lines: mild inhibition, k_Ic_d = 1. Dotted lines: strong inhibition, k_Ic_d = 4. (e) If the inhibitor is added at t = 0, it can greatly delay the time at which the plateau is reached but cannot affect the plateau concentration. (f) If the inhibitor is added at t = 1.14 h (approximately, the half-time for oligomer formation), it has little effect on the oligomer kinetics. If added after the plateau, inhibitors would have no effect at all, even at very high concentrations.