Atomic force microscopy for single molecule characterisation of protein aggregation

Abstract

The development of atomic force microscopy (AFM) has opened up a wide range of novel opportunities in nanoscience and new modalities of observation in complex biological systems. AFM imaging has been widely employed to resolve the complex and heterogeneous conformational states involved in protein aggregation at the single molecule scale and shed light onto the molecular basis of a variety of human pathologies, including neurodegenerative disorders. The study of individual macromolecules at nanoscale, however, remains challenging, especially when fully quantitative information is required. In this review, we first discuss the principles of AFM with a special emphasis on the fundamental factors defining its sensitivity and accuracy. We then review the fundamental parameters and approaches to work at the limit of AFM resolution in order to perform single molecule statistical analysis of biomolecules and nanoscale protein aggregates. This single molecule statistical approach has proved to be powerful to unravel the molecular and hierarchical assembly of the misfolded species present transiently during protein aggregation, to visualise their dynamics at the nanoscale, as well to study the structural properties of amyloid-inspired functional nanomaterials.

Keywords: Amyloid; Atomic force microscopy; Biophysics; Protein aggregation; Resolution; Single molecule imaging.

Fig. 1

Principles of atomic force microscopy. A) Simplified representation of an atomic force microscope [65]. B) Schematic representation of the cantilever oscillation damping as a function of z piezo displacement (tip-sample separation) (adapted from Ref. [66]). If the cantilever is excited exactly at one of its resonance frequencies f_res, the set point amplitude of oscillations A_Sp of the tip, falls into a bi-stable region; where the dominant forces are interchanging between attractive and repulsive regimes, which results in a non-linear response of the signal. C) Schematic representation of a change in the effective frequency f_eff of oscillation in tapping mode, non-contact mode and bi-stable regime. In free space, the cantilever oscillates at a free oscillation amplitude A₀ and resonance frequency f_res. As the cantilever is brought closer towards the sample surface, the tip-sample forces dampen the amplitude and modify the frequency of oscillation, until the set point amplitude A_Sp of oscillations is reached. At small tip-sample distances (left), repulsive interactions dominate and the resonance frequency f_res will shift to an effective frequency f_eff higher than f_res. At large tip-sample distances (centre) attractive forces dominate and the resonance frequency f_res will shift to an effective resonant frequency f_eff lower than the f_res. The change in the amplitude of oscillation ΔA will depend on the chosen set point amplitude A_sp or drive frequency f_drive. The cantilever falls into bi-stable regime (right) when its oscillations are driven at the maximum of the resonance frequency. In the bi-stable regime the same set point oscillation amplitude can be reached at two distinct effective resonance frequencies (f₁ and f₂), which represent different tip-sample distances causing instability.

Fig. 2

Schematic representation of distinct pathways and mechanisms of nucleation dependent amyloid fibril formation [29]. The black arrows represent the common conversion mechanisms in misfolding and amyloid formation, which can follow three possible pathways of nuclei formation [29]: nucleated polymerisation (brown arrows), nucleated conformation conversion (blue arrows) and native-like aggregation (green arrows). Besides primary nucleation, secondary processes, such as secondary nucleation (top) and fibril fragmentation (bottom), can occur during aggregation.

Fig. 3

Tip-sample interaction control by measuring at constant phase change [58]. a-e) AFM phase images of different protein aggregates for single molecule statistical analysis. f) Phase change in each image, which was constant and always smaller than Δ20°.

Fig. 4

Effects of pixel sizes on final AFM resolution [108]. Schematic of a sample imaging process (A). The dotted line and continuous blue line represent the object profile contoured with the AFM probe. The resolution of the profile (black line) is lost upon increasing the pixel size, I-V. Sampling an AFM morphology maps at pixel sizes 3, 5, 10, 20, 40 and 80 Å (B I-VI respectively) All morphology maps have a vertical full grey level range of 10 Å and a frame size of 300 Å.

Fig. 5

Convolution effect in AFM measurements. An overestimate of the lateral dimensions and of the volume of A) oligomeric and B) fibrillar species may be caused by the finite geometrical shape of the tip.

Fig. 6

AFM imaging of the α-synuclein (a-c) and Aβ (d-f) fibrillisation process [57]. Imaging of the amyloid fibril formation via AFM enabled to compare aggregation process of α-synuclein and Aβ. During the initial stages of aggregation Aβ formed oligomeric species (d), while α-synuclein remained mainly in the monomeric form (a). At later stages of aggregation fibrillary (b) and protofibrillar (e) species were observed for α-synuclein and Aβ respectively. Mature amyloid fibrils were detected for both proteins at the late stages of aggregation (c, f).

Fig. 7

Monitoring of α-synuclein aggregation time course by AFM and identification of single-strand protofilaments [6]. A) Visualisation of initial time point of α-synuclein aggregation and subsequent statistical characterisation of aggregate cross-sectional height (on the right) enabled identification of monomeric and dimeric species. After 1 day of incubation elongated protofilaments species were detected (B). Statistical analysis of cross-sectional height, of these species, allowed identification of subnanometre fibril-like aggregates, termed single strand protofilaments, After 10 days of incubation, protofibrillar and fibrilar species of α-synuclein were observed (C). Statistical analysis of cross-sectional height of all aggregate species enabled to establish hierarchical assembly of α-synuclein fibrils. D) Model of α-synuclein fibril formation.

Fig. 8

Monitoring of Aβ40 aggregation time course in the absence (A) and presence (B) of Zn²⁺ [17]. Detailed statistical analysis of species height and length (C) enabled to determine that Zn²⁺ inhibits formation of the mature fibrils and directs the Aβ40 aggregation process towards the formation of spheroidal aggregates.