The nanoscale morphology of amyloid fibrils can be visualized with atomic force microscopy (AFM) or electron microscopy (see, for example, (9,10)). The aggregation of α-synuclein into mature fibrils has been described by models based on AFM images (11) and is supposed to proceed through various stages, each with characteristic morphological features such as typical fibril heights and periodicities.

Although there are many studies on the morphology of fibrillar aggregates of various amyloidogenic proteins, there has been little uniformity in the analysis and description of fibril characteristics (11). In this Letter, we describe a quantitative analysis procedure that is applicable to atomic force micrographs of amyloid fibrils originating from any amyloidogenic protein. The procedure (Fig. 1) takes into account the limitations of the physical process of imaging a biological sample with atomic force microscopy, such as tip-sample convolution.

FIGURE 1 Amyloid fibril characterization procedure applied to an E46K α-synuclein fibril. The top panel shows an AFM topography image of an α-synuclein fibril (left panel) with a manually drawn (ImageJ (12)) segmented line profile (right panel (more ...)

We applied the quantitative image analysis procedure to AFM images (made in tapping mode and in liquid) of fibrils of wild-type and mutant α-synuclein formed after 72 h of aggregation (13). The analysis reveals grossly similar morphology but varying heights and periodicities among the disease-related mutants (Table 1). Fibrils from A30P and E46K α-synuclein appear smaller in height and correspondingly lower in peak height, trough height, and modulation depth than wild-type fibrils. Height measurements of wild-type α-synuclein fibrils in this study are consistent with those measured by others. We measure 10.0 ± 1.1 nm average height for the fibrils formed by wild-type α-synuclein, as compared to 11 ± 2 nm (14) and 9.8 ± 1.2 nm (11).

TABLE 1 Quantitative comparison of fibril morphology of disease-related α-synuclein mutants

Most E46K fibrils display a periodicity of around 55 nm (as in Fig. 2 A), but much larger periodicities (Fig. 2 B) or unperiodic fibrils (Fig. 2 C) are also observed. The periodicities of E46K and wild-type α-synuclein fibrils have modes at 55 and 65 nm, respectively, and fibrils formed by A30P α-synuclein display a wider variety of periodicities (Fig. 3).

FIGURE 2 Tapping mode atomic force micrograph of E46K α-synuclein in liquid after 72 h of aggregation. The color scale represents 17.6 nm in height. The fibril fragments marked A, B, and C are profiled along their length in the bottom panel indicating (more ...)

FIGURE 3 Distribution of periodicities in fibrils of disease-related α-synuclein mutants. Wild-type and E46K α-synuclein fibrils have modes at 65 and 55 nm, respectively. A30P shows a wider variety of periodicities. Histogram bin size is 10 nm. (more ...)

Quantitative discussion of morphological features of objects studied by AFM requires consideration of the influence of scanning system properties on the resulting image. One of the main limiting factors in deriving fibril properties from AFM images is tip-sample convolution, as evidenced by the large difference between the apparent width and the height of fibrils that are expected to be approximately round in cross section.

To assess the reliability of the measured parameters, we use a simple one-dimensional model for tip-sample convolution (Fig. 4). From Fig. 4 A, it follows that the tip radius is r_t = w²/16r_s, and the lowest measurable trough height can be calculated based on Fig. 4 B (see Supplementary Material). In the case of our α-synuclein fibrils, the trough height is overestimated in the images since the tip radius is large relative to the periodicity. This complicates attributing the observed periodicity to a supposed helical structure of the fibrils. We consider the measurements of the periodicity to be reliable since they were found to be equivalent for fibrils in any orientation with respect to the scan direction, and because the tip radius does not affect the periodicity observed but only the modulation depth, as shown in Fig. 4 B.

FIGURE 4 Simple one-dimensional models for tip-sample convolution for helical fibrils. Fig. 4 A allows calculation of the tip radius assuming the tip (green) is spherical and the sample (red) has a spherical cross section (see Supplementary Material). Fig. 4 (more ...)

In the hierarchical assembly model of α-synuclein fibril formation, as proposed by Khurana et al. (11), the trough height would be y_t = (3/2)d, where d is the protofibril diameter (equal to the radius of the mature fibril r_s). We can now compare the trough height values as modeled with the tip-sample convolution model, as modeled with the hierarchical assembly model, and as actually measured, for two fibrils with a different periodicity (Table 2). Based on this comparison, we conclude that the morphological features we observe are those one would expect of fibrils formed by twisted protofibrils, but we note that the lateral resolution of our images is not high enough to prove unambiguously that this model is indeed correct.

TABLE 2 Comparison of measured versus modeled trough heights

This study introduces a quantitative morphological analysis procedure of fibrillar aggregates of amyloidogenic proteins. We show that high-resolution atomic force microscopy under sample-favorable imaging conditions (imaging under buffer, low tapping amplitude (~4 nm), and low interaction force), combined with detailed image analysis, allows nanoscale comparison of fibrils formed by human wild-type α-synuclein and disease-related α-synuclein mutants. The fibrils formed by the disease-related mutants have a similar overall morphology, but a smaller diameter and a different periodicity than fibrils formed by wild-type α-synuclein. The observed fibril morphology is compatible with the hierarchical assembly model of α-synuclein fibrillization described in Khurana et al. (11). We emphasize that great care should be taken in deriving structural models from AFM data. We believe that the careful quantitative approach to determining fibril morphological characteristics reported here should be used more widely, and will be of increasing importance considering the growing number of studies that use advanced scanning probe microscopies for nanometer scale characterization of fibrils.

An online supplement to this article can be found by visiting BJ Online at http://www.biophysj.org.

The authors thank Kirsten van Leijenhorst for protein expression and purification, and Kees van der Werf for expert advice on atomic force microscopy.

This work is part of the research programme of the Stichting voor Fundamenteel Onderzoek der Materie, which is financially supported by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek.