**1. Introduction**

The interface of a solid surface and a protein at the nanoscale level are of interests for many applications including material science, biomedical science (e.g., implantation of artificial bones, heart, organs or blood clotting), industry (e.g., the manufacturing of biosensors, bio separation processes, and drug delivery), and research in the development of new materials. The functionalities of peptide coated nanomaterials have remarkably broad applications in areas where nano-size has a very significant effect [1], including nano-light switching devices [2], disease controlling materials combined with DNA [3], DNA sensors [4], control of human cell activity [5], photo dynamic therapy [6], and optical biosensors that quantify heavy metal pollution in water [7]. However, very little is known regarding how the proteins adhere to nanoscale solid surfaces.

Amyloidogenic peptides, such as Aβ1–40 or Aβ1–42, α-synuclein (α-syn), and β2 microglobulin (β2m) are all regarded as hallmark peptides associated with key onset mechanisms of neurodegenerative diseases, such as Alzheimer's disease or Parkinson's disease. Because of this, the formation process and characterization of amyloid fibrils have been extensively investigated. Fibrils are usually several hundreds of micrometers in size and consist of pre-fibrils, which are built of unit oligomers. Therefore, the formation of unit oligomers from soluble and nontoxic monomers is regarded as a key intermediate process of fibrillogenesis and is considered to be a reversible process. On the other hand, the formation of fibrils or pre-fibrils are considered non-reversible processes. The nuclei-based pre-fibril formation mechanism is considered to be the most reasonable method for interpreting the fibril formation process. A key stage in this fibrillogenesis is the formation of an intermediate oligomer through a reversible process which then leads to one-direction pre-fibril formation. A major conformational change of the monomer to an intermediate oligomer requires significant protein folding, requiring a Gibbs energy of −10 kcal/mol [8–10] based on computational calculations. The most important on-set process of any fibrillogenesis is networking between peptides. Considering that fibrils are formed irreversibly, the networking between peptides must be "effective" and occur due to a strong interaction. However, a direct investigation of this networking process is still lacking.

The PI's group has been investigating the reversible self-assembly process (i.e., reversible networking process) on amyloidogenic peptide-coated gold colloidal nanoparticles. The peptides are relatively small, amphiphilic (i.e., consists of both hydrophilic and hydrophobic segments) peptides and the temperature/pH conditions for folded/unfolded conformations are well studied. The great advantage of this system is that a monomer peptide can be prepared on the nano-surface by orientating each peptide so that it may undergo the most effective networking process. Peptides are adsorbed over the nano-surface and are used to make connections between two nano-particle surfaces by making networks between peptides. Because of the networking between peptides, the assembly to the gold colloidal aggregates results in a drastic change in a spectroscopic feature, meaning that the networking process can thus be spectroscopically probed. Therefore, this work is viewed as the best prototype system to learn how nanoscale surface potentials interact with a peptide and if a specific structure can be selectively constructed [11–13]. Although these peptides eventually form irreversible insoluble amyloid fibrils, initial stages in fibrillogenesis are still reversible processes. We hypothesize that the peptide-peptide networking must be established by an unfolded conformation of each peptide and this unfolded conformation will be strongly enhanced at the nano-particle surface. As observed in negatively charged micelles and Teflon particles, β-sheet formation of Aβ on hydrophobic graphite surfaces [14] or at air–water interfaces [15] indicate an involvement of interfacial surface potential utilized for the conforming intermediate [16–20].

This study aims to clarify (1) an exact attaching sequence or portion of the peptide and (2) orientation of the peptide over the nano-scale surface, and (3) identify the probable conformation of peptides for successful networking. It is well known that the amyloidogenic peptides (e.g., amyloid beta: Aβ; beta 2 microglobulin: β2m; and alpha-synuclein: α-syn) adsorb onto a gold surface through a sulfur atom of a thiol (–SH) group. These amyloidogenic peptides undergo drastic structural changes (protein folding) to form many units of toxic polymers that eventually combine to create a few micron-sized fibers (i.e., amyloid fibrils), which are known to cause neurodegenerative diseases [21–35]. However, Aβ and α-syn do not possess any sequences which contain a thiol group (i.e., Cysteine (Cys, C)). Contradicting the lack of a Cys sequence, the existence of gold nano-colloids were reported to enhance peptide-peptide networking using Aβ adsorbed on a nano-gold colloid as a "core" of a fiber [36].

There is no clear explanation of why amyloidogenic peptides adsorb onto gold so effectively. Since the detailed structural information of adsorbed peptides at the "core" is not known, contributing factors to the peptide-peptide networking needs to be fully investigated. This study describes a systematic method used to extract both a plausible peptide orientation and which segments interact with the colloidal surface. Based on these data, in Section 4.3.3 we describe a novel systematic procedure to extract the coverage ratio of peptide (Θ) onto the nano-gold colloids. Quite unexpectedly, the surface coverage conditions appeared to depend somewhat on the nano-particle size. While simulation does not explain the nano-size dependence on surface coverage, the observed trend suggests a plausible "packing" formation of the peptide due to a physical surface area condition.

#### **2. Results**

#### *2.1. Extraction of* Θ

This hypothesis stating a linear relationship between dpH and ΔpHo was clearly proved to be true when *d*pH was plotted as a function of ΔpHo for each tested gold colloidal size and in all three amyloidogenic peptides (Aβ1–40, α-syn, and β2m) as shown in Figure 1, while ovalbumin-coated gold colloid did not show any sign of linear relationship [36]. Each data point shown in Figure 1 corresponds to different gold colloidal sizes for a given peptide. It shows that the coverage area is determined by the size of the nano-particle and there must be an equilibrium electrostatic shielding value for a given nano-gold metal surface. The average coverage ratio, Θavg., for each amyloidogenic peptide was extracted as: Θavg (Aβ1–40) = 0.6 ± 0.2, Θavg (α-syn) = 0.6 ± 0.2, and Θavg (β2m) = 0.7 ± 0.2. By using the reported structural data most suited to the conditions of our work, the axial length *a* and *b* (*a* < *b*) of an approximated prolate for Aβ1–40 [17] α-syn [37], and β2m [38] were initially estimated to be: Aβ1–40 (*a*, *b*) = (2.1 nm, 4.1 nm), α-syn (*a*, *b*) = (3.1 nm, 8.5 nm), and β2m (*a*, *b*) = (2.5 nm, 4.6 nm).

**Figure 1.** *Cont.*

**Figure 1.** A hypothesized linear relationship between *d*pH vs. ΔpHo was plotted for each amyloidogenic peptide-coated nano gold colloid, (**a**). Aβ1–40 in blue, (**b**). α-syn in red, and (**c**).β2m in green based on the values obtained from fitting sigmoidal plot with Equation (1). Each linear line was given as a guide for a correlation between *d*pH vs. ΔpHo. (*d*pH = m ΔpHo + b). Supporting information of this figure is available at Supplementary Materials.

#### *2.2. Distance between Colloidal Particles*

A representative TEM image of β2m-coated 30 nm gold colloids is shown in Figure 2a. As the magnified image clearly shows, distinct spacing noted as "Δ" between gold colloids were observed. (Figure 2b) The spaces between the gold particles was measured over multiple measurements per image for each size of gold particle and β2m. The morphology of the gold colloidal aggregates coated with Aβ1–40, α-syn, as well as albumin were also studied, and the aggregates formed were more extensive in size and number (size exceeded up to a few microns and the number of gold colloids far exceeded 500 or 1000) than those formed by β2m-coated gold colloid. Therefore, the density of the aggregates by Aβ1–40, as well as albumin coated gold colloid aggregates, extended in the longitudinal direction resulting in preventing the view the section of planar topology. In contrast, β2m formed relatively smaller aggregates, ranging within 500 nm with less than 100 colloids. This allowed us to visualize a two-dimensional view of the aggregate and made it possible to disclose the spacing between two adjacent gold colloids. While the peptide character of the studied peptides is not equivalent, we assume the networking character can be similar. Also, the physical dimension of the networking section compared to the diameter of each colloid can be approximated to be the same. Therefore, the information obtained for β2m will be shown as illustrative for the other two peptide coated gold colloids.

**Figure 2.** (**a**) A TEM (transmission electron microscopy) image of 30 nm gold colloidal particles coated with β2m at pH 4. The scale bar for 100 nm is shown as a guide. (**b**) A magnified section of the red box showing the diameter of a gold colloid *d* and the distance between the adjacent colloidal particles (Δ). (Inset): A typical histogram showing the Δ and the observed numbers of the distribution. The distribution was fit by a Gaussian profile and shown by a dotted curve. This histogram is for the β2m-coated 30 nm gold colloid.

The average distances of adjacent nano-gold colloids and the average number of gold nano-particles to form one aggregate for *d* = 10, 30, 60, and 80 nm particles are summarized in Table 1. We can conclude that the average distance of the adjacent nano-gold colloids was extracted to be 2.0 ± 0.8 nm, indicating insignificant size dependence of the gold colloid and the number of nano-gold colloids forming a cluster.


**Table 1.** The average distance, (Δ), of adjacent β2m-coated nano-gold colloids and the average number (η) of gold colloids consisting in an aggregate (gold colloid cluster) at pH 4.0 ± 0.3.

#### *2.3. Simulation of* Θ *and Orientation*

The procedures explained in Section 4.3.3 was applied to analyze the data points shown in Figure 3. The optimized axial lengths of prolate for three amyloidogenic peptides as a function of colloidal size are listed in Table 2. In order to reproduce extracted Θ, the spiking-out orientation need to be utilized for all cases except for *d* = 100 nm gold coated with Aβ1–40 and α-syn as well as *d* = 10, 20, 60 nm gold coated with β2m, which exhibited a lie-down orientation. For example, Aβ1–40 coated gold 20 nm diameter showed Θobs = 0.74 for *a* = 1.4 nm and *b* = 2.2 nm in order to reach maximum coverage ratio 0.37 under the first layer, Θ*f*,cal, with total number of attached peptides to be *n*total = 111. After the second layer was added, it is calculated to have a maximum of Θ*s*,cal = 0.72. In order to be consistent with the observed Θobs = 0.74, a contribution of the second layer, γ (Θ*s*,cal) should be 0.51, so that Θobs = 0.74 = Θ*f*,cal + γ (Θ*s*,cal) × Θ*s*,cal =0.37 + 0.51 × 0.72=0.37 + 0.37. There was almost no contribution of the second layer when Θobs < 0.5 for *d* = 100 nm of both Aβ1–40 and α-syn as well as *d* = 10, 20, and 60 nm for β2m.

**Figure 3.** A plot for experimentally obtained Θ for (**a**) Aβ1–40 in blue (**b**) α-syn in red, and (**c**) β2m in green as a function of reported gold diameter *d* (nm). All data points were extracted by using Equation (2). Each dashed curve line is simulated by a method described in Section 4.3.3 with the parameters tabulate in Table 2. The dotted line shows an upper limit of the Θ value obtained by a single layer model. (**d**) The over-laid plots of all (**a**–**c**) are shown, and indicating (**a**,**b**) almost overlay each other.


**Table 2.** In each box, optimized axial length of a prolate (*a* and *b*), the sketches of orientation of adsorption, n*f*,tot(see Equation (3)), and circular graph indicating % of occupied surface area by adsorption (first layer in red, second layer in blue, and unoccupied in gray) for Aβ1–40, α-syn, and β2m as a function of gold size d (and *d*) nm.

In summary, as shown in Figure 4, all studied amyloidogenic peptides interacted with nano-gold surface at the original pH where a sample was prepared (~pH 7) with unfolded condition with spiking out condition.

**Figure 4.** A picture of an illustrative model and sketch demonstrating a peptide aligning over the surface of a gold colloidal particle with a diameter, *d*. In the image on the **left**, the peptide and gold core are both shown with the same color, and a dotted oval indicates a prolate shaped peptide. In the sketch on the **right**, prolate shaped peptide is shown in blue and the core gold colloid with diameter d is shown in red sphere.
