Structural Basis for Aβ1–42 Toxicity Inhibition by Aβ C-Terminal Fragments: Discrete Molecular Dynamics Study

Abstract

Amyloid β-protein (Aβ) is central to the pathology of Alzheimer’s disease (AD). Of the two predominant Aβ alloforms, Aβ1–40 and Aβ1–42, the latter forms more toxic oligomers. C-terminal fragments (CTFs) of Aβ were recently shown to inhibit Aβ1–42 toxicity in vitro. Here we studied Aβ1–42 assembly in the presence of three effective CTF inhibitors and an ineffective fragment, Aβ21–30. Using a discrete molecular dynamics approach that recently was shown to capture key differences between Aβ1–40 and Aβ1–42 oligomerization, we compared Aβ1–42 oligomer formation in the absence and presence of CTFs or Aβ21–30 and identified structural elements of Aβ1–42 that correlated with Aβ1–42 toxicity. CTFs co-assembled with Aβ1–42 into large heterooligomers containing multiple Aβ1–42 and inhibitor fragments. In contrast, Aβ21–30 co-assembled with Aβ1–42 into heterooligomers containing mostly a single Aβ1–42 and multiple Aβ21–30 fragments. The CTFs, but not Aβ21–30, decreased the β-strand propensity of Aβ1–42 in a concentration-dependent manner. CTFs and Aβ21–30 had a high binding propensity to the hydrophobic regions of Aβ1–42 but only CTFs were found to bind the Aβ1–42 region A2-F4. Consequently, only CTFs but not Aβ21–30 reduced the solvent accessibility of Aβ1–42 in the region D1-R5. The reduced solvent accessibility of Aβ1–42 in the presence of CTFs was comparable to the solvent accessibility of Aβ1–40 oligomers formed in the absence of Aβ fragments. These findings suggest that the region D1-R5, which was more exposed to the solvent in Aβ1–42 than in Aβ1–40 oligomers, is involved in mediating Aβ1–42 oligomer neurotoxicity.

Amino acid substitutions [K16A] and [K28A] distinctly affect amyloid β-protein oligomerization

“…The first DMD4B-HYDRA study of Aβ 40 and Aβ 42 oligomer formation showed that Aβ 42 oligomers are characterized by increased plasticity in the N-terminal region relative to Aβ 40 oligomers [44], the feature that was hypothesized to be linked to increased toxicity of Aβ 42 relative to Aβ 40 oligomers [48]. This result was unexpected because the sequences of Aβ 40 and Aβ 42 differ by two additional amino acids, IA, at the C-terminus of the latter, so that structural differences would be expected to occur in the C-terminal region.…”

“…In the present study, Aβ 40 is the only variant that forms oligomers, which lack the conformations with extended N-to-C distances, whereby both substitutions, [K16A] and [K28A], cause changes in the N-terminal secondary structure of Aβ 40 oligomers that are similar to those induced by previously reported Aβ 40 sequence modifications described above. The question that remains to be explored is whether or not increased plasticity at the N-termini of Aβ 42 oligomers plays a role in Aβ 42 -mediated toxicity [48].…”

“…7c, lower left map). The dominance of the C-terminal region in Aβ 42 but not Aβ 40 assembly is structurally linked to increased flexibility and solvent exposure of the N-terminal region in Aβ 42 relative to Aβ 40 oligomers, which was postulated to underlie distinct abilities of the two peptides to mediate toxicity [44,48]. As the C-terminal region dominates quaternary contacts of Aβ 42 …”

“…Similarly, the physiologically relevant oligomers formed by both Aβ 3−40 and Aβ 3−42 , which have been shown to be more toxic than their WT counterparts, have even more flexible and solvent-accessible N-termini than those formed by WT peptides [46]. The predictions of DMD4B-HYDRA simulations, including the effect of toxicity inhibitors of the Aβ 42 assembly structure [48], provide a basis for a plausible structuretoxicity relationship that can be further examined. To elucidate the role of lysines at early stages of Aβ 40 and Aβ 42 assembly and complement the existing experimental findings [42], we here apply the DMD4B-HYDRA approach to investigate the effect of lysine substitutions with alanine, K16A and K28A, on Aβ oligomer assembly dynamics and resulting structures.…”

Folding of pig gastric mucin non-glycosylated domains: a discrete molecular dynamics study

“…For most proteins in an aqueous solution, E HP is expected to be of a similar value. We thus use here E HP = 0.3, which corresponds to E HP = 1.4 kcal/mol or 2.3 k B T, the value that was used in the Aβ studies of folding and oligomerization [16,19,23]. The solvent parameter associated with effective electrostatic interactions, E CH , was varied between 0 and 1 (see Section 3.1) to explore the effect of ionic strengths on folding.…”