Polymer-Peptide Conjugates Convert Amyloid into Protein Nanobundles through Fragmentation and Lateral Association

Abstract

The assembly of proteins into amyloid fibrils has become linked not only with the progression of myriad human diseases, but also important biological functions. Understanding and controlling the formation, structure, and stability of amyloid fibrils is therefore a major scientific goal. Here we utilize electron microscopy-based approaches combined with quantitative statistical analysis to show how recently developed kind of amyloid modulators-multivalent polymer-peptide conjugates (mPPCs)-can be applied to control the structure and stability of amyloid fibrils. In doing so, we demonstrate that mPPCs are able to convert 40-residue amyloid beta fibrils into ordered nanostructures through a combination of fragmentation and bundling. Fragmentation is shown to be consistent with a model where the rate constant of fibril breakage is independent of the fibril length, suggesting a local and specific interaction between fibrils and mPPCs. Subsequent bundling, which was previously not observed, leads to the formation of sheet-like nanostructures which are surprisingly much more uniform than the starting fibrils. These nanostructures have dimensions independent of the molecular weight of the mPPC and retain the molecular-level ordering of the starting amyloid fibrils. Collectively, we reveal quantitative and nanoscopic understanding of how mPPCs can be applied to control amyloid structure and stability, and demonstrate approaches to elucidate nanoscale amyloid phase behavior in the presence of functional macromolecules and other modulators.

Keywords: amyloid; functional polymers; polymer–peptide conjugates; protein electron microscopy; protein nanomaterials.

Figure 1.

Overview of mPPC-induced Aβ₄₀ disassembly. (a) Schematic representation of the joint effects of fragmentation and bundling, induced by the introduction of mPPCs, to produce bundles of amyloid segments. (b–d) Representative time-lapse TEM images of disassembly 0, 48, and 72 h after addition of mPPC. Additional examples and control samples without mPPC are presented in Figure S3. (e) Quantitative analysis of the fibril lengths (l_fibril) and bundle widths (w_bundle) over time. The shaded regions around the center line correspond to one standard deviation about the mean of the distribution. Scale bars: 200 nm.

Figure 2.

Fibril morphology evolution along the disassembly pathway. (a–c) From 24 h (a), 48 h (b), to 72 h (c) after adding mPPC to preassembled fibrils, the average fibril length l_fibril decreases from 130 nm to 67 nm at 48 h and 59 nm within 72 h. The inset in c shows a zoomed in view of the 72 h l_fibril distribution highlighting the smaller size regime. (d–f) In parallel, lateral association of fibril segments leads to the increase of the average bundle width w_bundle from 23 nm at 24 h (d) to 50 nm within 72 h (f) after adding mPPC. All curves denote a best-fit log–normal distribution as detailed in Figure 3.

Figure 3.

Fibril length dimensions converge to a log–normal size distribution over time, indicating uniform random fragmentation. (a–d) Comparison of fits to the 72 h l_fibril cumulative probability distribution (Figure 2c), including the normal distribution (a), Weibull distribution (b), log–normal distribution (c), and exponential distribution (d). For comparison, the scales of the x and y axis are such that an ideal fit to the distribution would correspond to a straight line. The same data are shown on a linear scale in Figure S4 (see also Tables S1–S2). (e) Schematic representation of uniform random fragmentation with a lower-bound size limit.

Figure 4.

mPPCs with different molecular weights produce qualitatively and quantitatively similar bundles. (a–b) Representative negative stain micrographs of bundles disassembled by (a) 45 kDa mPPC and (b) 166 kDa mPPC. The distribution of fibril lengths (c) and bundle widths (d) after 72 h of disassembly with increasing mPPC molecular weights are qualitatively and quantitatively consistent. See also, Figure S5–S7. Both distributions are of the logarithm of their respective bundle dimension, to highlight their log–normal nature. Sample sizes for each dataset are given in Tables S1–S2.

Figure 5.

Structural analysis of the products of disassembly. (a) Schematic illustrating the internal β-sheet spacing (0.47 nm) characteristic of Aβ₄₀ amyloid fibrils. Bundles coming from disassembly by 90 kDa mPPC after 72 h were chosen as representatives. (b) Radially integrated nanobeam electron diffraction pattern exhibiting higher order peaks (0.24 and 0.16 nm) corresponding to the β-sheet spacing, and to the graphene sheet used as a protective agent. (c) TEM images of a bundle collected from a wide range of tilt angles indicate that the thickness of bundles is relatively small and the overall structure is plate-like. (d) Magnified TEM image of a bundle resulting from disassembly of amyloid fibrils, highlighting the well-defined “stripes.” (e) Grayscale intensity integrated vertically along several different bundles, indicating a periodicity Δ of 6.2 ± 1.1 nm, consistent with protofibril length scales.