Changes in the Secondary Structure and Assembly of Proteins on Fluoride Ceramic (CeF3) Nanoparticle Surfaces.

Fluoride nanoparticles (NPs) are materials utilized in the biomedical field for applications including imaging of the brain. Their interactions with biological systems and molecules are being investigated, but the mechanism underlying these interactions remains unclear. We focused on possible changes in the secondary structure and aggregation state of proteins on the surface of NPs and investigated the principle underlying the changes using the amyloid β peptide (Aβ16-20) based on infrared spectrometry. CeF3 NPs (diameter 80 nm) were synthesized via thermal decomposition. Infrared spectrometry showed that the presence of CeF3 NPs promotes the formation of the β-sheet structure of Aβ16-20. This phenomenon was attributed to the hydrophobic interaction between NPs and Aβ peptides in aqueous environments, which causes the Aβ peptides to approach each other on the NP surface and form ordered hydrogen bonds. Because of the coexisting salts on the secondary structure and assembly of Aβ peptides, the formation of the β-sheet structure of Aβ peptides on the NP surface was suppressed in the presence of NH4+ and NO3- ions, suggesting the possibility that Aβ peptides were adsorbed and bound to the NP surface. The formation of the β-sheet structure of Aβ peptides was promoted in the presence of NH4+, whereas it was suppressed in the presence of NO3- because of the electrostatic interaction between the lysine residue of the Aβ peptide and the ions. Our findings will contribute to comparative studies on the effect of different NPs with different physicochemical properties on the molecular state of proteins.

Introduction

In the field of material sciences, the self-assembly of molecules is a major area of interest for researchers to develop soft matter.[1] Furthermore, in biology, the secondary structure and assembly state of a protein are important for regulating the function and activity of the protein. The formation of fibrillar aggregates of peptides and proteins is associated with various diseases including neurogenerative disorders.[2] The dysregulation of the structure and state of proteins can lead to diseases in biological organs.[3,4] For example, fibril formation via the self-assembly of denatured proteins causes amyloid diseases, such as Alzheimer’s disease,[5] which is the most common neurodegenerative disease and is associated with cognitive and physical decline.[6] Pathologically, in patients with Alzheimer’s diseases, senile (neuritic) plaques of the amyloid beta (Aβ) protein are observed in the brain tissue.[7]

Amino acid residues in proteins interact with each other through hydrogen bonds, disulfide bonds, electrostatic interactions, and hydrophobic interactions; these interactions affect protein conformation. As the distance between peptides reduces due to protein aggregation, their secondary structure can change due to an increase in interactions among peptides, for example, an increase in the β-sheet structure via enhanced hydrogen bonds. Although Aβ generally aggregates more at higher concentrations, even at low concentrations, it forms a β-sheet structure that can promote self-assembly via aromatic interactions between phenylalanine residues.[8,9] Especially, the hydrophobic domain of Aβ—the region around residues 17–20, LVFF—is important for β-sheet formation.[10]

In general, changes in the secondary structure and protein aggregation can be enhanced at liquid–liquid and liquid–air interfaces, as observed through denaturation by surfactants. In addition to these interfaces, liquid–solid interfaces are also likely to be the sites for protein denaturation because the domains with high affinity for the solid are exposed to the molecular surface. Molecular dynamics (MD) simulations have suggested that the structural arrangement of Aβ attached to the surface of carbon nanotubes—an inorganic nanomaterial—changes with the radius of the nanotubes.[11] The rate of peptide aggregation on the solid–liquid interface is determined by the affinity (binding force) between the peptide and the solid material and the shape (roughness) of the solid surface.[12] While a high affinity between peptides and solid surfaces enhances their adsorption and inhibits the self-assembly (aggregation) of peptides, solid materials with middle affinities and a rough surface (with a nanoscale morphology) accelerate the aggregation of peptides. In contrast, in a study in which all-atom MD simulations of the conformation of Aβ16–22 peptides were performed, hydrophobic interactions were found to prevent the formation of β-sheet structures in the presence of gold nanoparticles (NPs).[13] The charge on the surface of NPs is also likely to be important for Aβ fibrillation. Negatively charged NPs inhibit the formation of Aβ fibrillation, whereas positively charged NPs have no effect on fibril formation.[14] The aggregation dynamics of Aβ peptides (Aβ16–21) on fullerene NP models was also investigated using MD simulation along with the effect of the coexistence of ionic salts.[15] However, there has still been no experimental evidence of how Aβ peptides behave on NP surfaces in physiological environments containing salts.

In the present study, fluoride ceramic NPs, which are expected to be applied to the biomedical field, were used as a target material. Fluoride has a moderate phonon energy of 350 cm–1 and has been widely utilized in the biomedical field for applications such as brain imaging[16−18] as a fluorescent contrast agent containing rare-earth ions.[19−22] This is because it has both high chemical durability, as a lower phonon energy reduces the chemical stability similarly to chlorides, and high luminescence efficiency when doped with rare earths because higher phonon energy causes quenching via enhanced thermal relaxation.[23] Fluoride nanomaterials labeled within the long-wavelength (>1000 nm) near-infrared (NIR) region, called the second and third NIR (NIR-II/III) biological windows,[24] have been developed for NIR fluorescence computed tomography,[25] photodynamic therapy,[26] and fluorescence nanothermometers[27,28] for in vivo investigations of deep tissues such as tissues in the peritoneal cavity.[29] Rare-earth-doped fluoride crystals have also been developed for application in lifetime-based NIR fluorescence thermometers.[30] Fluoride crystals containing Gd3+ (e.g., NaGdF4) and luminescent rare earths have been developed for bimodal imaging in fluorescence and magnetic resonance modalities.[31−36] CeF3 NPs were used in this study as a fluoride ceramic that can show a surface reactivity similar to that of the fluorides, as mentioned above. The aim of this study was to investigate the effect of fluoride ceramic NPs, which are expected to have further biomedical applications, on the conformation and assembly of Aβ molecules using an in vitro experimental system and MD simulation.

Results and Discussion

Fluoride NPs synthesized in this study were characterized using X-ray diffraction (XRD) and dynamic light scattering (DLS). As shown in Figure a, the XRD patterns showed that the NPs were majorly CeF3 with small amounts of NaCeF4. Data from DLS showed that the CeF3 NPs showed a major peak at a diameter of 80 nm with a low polydispersity index (0.116), although it contained a minor fraction peak at 20 nm. We considered the representative CeF3 particle size to be 80 nm of the peak in the size distribution in our following investigations.

Figure 1

Characterization of fluoride NPs used in this study. (a) XRD pattern of the samples and references of CeF3 (JCPDS no. 08-0045) and NaCeF4 (JCPSD no. 75-1294). (b) DLS spectra showing the hydrodynamic diameter of the fluoride NPs dispersed in cyclohexane.

In this study, fragment peptides with small molecular weights, Aβ16–20, which allow MD simulations to be performed easily, were used to investigate the secondary conformational changes and aggregation of Aβ using both Fourier transform infrared (FT-IR) spectroscopy and the MD simulations. CeF3 NPs were dispersed in an aqueous solution to interact with Aβ16–20 (KLVFF) in the aqueous solution. FT-IR spectroscopy was used for analyzing the secondary structure of proteins. Especially, the amide I vibration peak (appearing around 1650 cm–1 and mainly attributed to C=O stretching) was focused because it is hardly affected by the nature of the side chains but depends on the secondary structure of the backbone.[37] Thus, it is commonly used for the secondary structure analysis[38] that can also be applied to in situ analyses under microscopy.[39] Not only the secondary structure but also the aggregation of the Aβ peptide in solvents can be analyzed using FT-IR spectroscopy.[40] Hydrochloric acid (1 mmol/L) was used to maintain the dispersibility of the CeF3 NPs. However, the water molecule showed peaks not only at 3300 cm–1 but also at 1650 cm–1 in the IR region; these peaks interfere with those of the amide I band at 1650 cm–1,9 which is the target of analysis in this study. Therefore, deuterium chloride and deuterium oxide were used, instead of hydrochloric acid and water, as the dispersion media for CeF3 and solution of Aβ16–20. Deuterium chloride did not affect the FT-IR spectra of Aβ16–20 solution at this concentration (final 0.25 mmol/L) (data not shown). FT-IR spectra of the samples in which Aβ16–20 was interacted with different concentrations of CeF3 NPs were analyzed. As shown in Figure a, Aβ16–20 showed two major peaks at 1674 and 1640 cm–1, which correspond to aggregates and monomers, respectively.[9] Deconvolution analysis using Gaussian fitting showed that the FT-IR absorption spectra of Aβ16–20 also included, in addition to the major peaks, a minor peak at 1618 cm–1 corresponding to β-sheet formation of Aβ16–20[9,38] (Figure b). The β-sheet formation of Aβ16–20 (6 mg/mL) increased in the presence of 3 mg/mL CeF3 NPs, and this increase was not observed in the presence of 6 mg/mL CeF3 NPs as the ratios of the β-sheet peak in the total amide I absorption were 6.8, 9.3, 6.9, and 6.3% in Aβ16–20 that interacted with 0, 3, 6, and 9 mg/mL of CeF3 NPs, respectively (Figure b–e and Table ). This may be due to the difference in the number of Aβ molecules per surface area of the NPs, which is the site of the NP–protein interaction in the system. Because the shape of CeF3 NPs (density: 6.16 g/cm3) is approximated to be a sphere with a diameter of 80 nm (2.7 × 105 nm3/particle), the mass and surface area are 1.7 × 10–15 g and 2.0 × 104 nm,[2] respectively, leading to a specific surface area per mass of 1.2 × 1019 nm2/g. The surface area of the CeF3 particles in a dispersion of 3 mg/mL was 3.7 × 1016 nm2/mL, while the total particle surface area in the dispersion was proportional to the particle concentration. Although the ratio of Aβ molecules that were attracted to the NP surface, that is, their local enrichment rate on the surface, in the dispersion was unknown, our results suggest that a certain enrichment of Aβ molecules promotes intermolecular bonding, thereby promoting β-sheet formation.

Figure 2

Amide I band in the FT-IR spectra of Aβ16–20 interacted with CeF3 NPs. (a) FT-IR spectra of Aβ16–20 (6 mg/mL) that interacted with different concentrations of CeF3 NPs (3, 6, and 9 mg/mL). (b–e) Deconvolution results via Gaussian fitting for the amide I band in FT-IR spectra of Aβ16–20 (6 mg/mL) with CeF3 NP concentrations of (b) 0, (c) 3, (d) 6, and (e) 9 mg/mL.

Table 1

Ratio of Each Component in the Amide I Band of the FT-IR Spectra of Aβ16–20 (6 mg/mL) That Interacted with Different Concentrations of CeF3 NPsa

	Aβ	Aβ + NPs 3 mg/mL	Aβ + NPs 6 mg/mL	Aβ + NPs 9 mg/mL
monomer	0.20	0.20	0.20	0.20
aggregate	0.74	0.70	0.74	0.73
β sheet	0.068	0.093	0.069	0.063

The ratios were obtained via deconvolution of the amide I band observed in each sample.

The effect of coexisting ions on the behavior of the Aβ16–20 peptide on the surface of CeF3 NPs was studied using Aβ16–20 in D2O with dissolved NaCl, NH4Cl, and NaNO3 (0.15 M). The effects of these ions at the same concentration were investigated in this experiment to compare the principle of action of each ion. Even without CeF3 NPs, NH4+ promoted β-sheet formation of Aβ16–20, whereas NO3– enhanced the monomer retention of Aβ16–20 (Table ). Na+ and Cl– did not affect the amide I band of Aβ16–20 solution in D2O (Table ). β-Sheet formation of Aβ16–20 (6 mg/mL) was increased (14%) by CeF3 NPs (3 mg/mL) in the presence of NaCl as well as in the absence of salts, whereas no elevation of β-sheet formation by NPs was observed in the presence of NH4+ and NO3– (Figure ). The results suggest that NH4+ and NO3– suppressed the β-sheet formation of Aβ promoted on CeF3. MD simulations were further performed on four monomers of Aβ16–20 in the presence of these salts. The findings showed that the NO3– was strongly bound to the peptide as compared to chloride in the absence of NPs. The average distance between the peptide and NO3– was 0.67 nm, whereas the distance with Cl– was 0.99 nm (Table ). Elevated peaks were observed on the radial distribution function (rdf) plots between peptide residues and NO3– and Na+ of NaNO3 with maximum peak values of ∼9.5 at a 0.42 nm distance, as shown in Figure a, and ∼0.88 at a 0.97 nm distance, as shown in Figure b). In contrast, comparatively low peak values were observed between peptides and ions of NaCl and NH4Cl (with maximum peak values of ∼0.96 at a 0.95 nm distance, as shown in Figure a, and ∼0.38 at a 1.0 nm distance, as shown in Figure b), indicating their weak interactions. Among the different residues of Aβ16–20 (KLVFF), NO3– strongly interacted with the lysine (K-16) residue, with the average distance between lysine and NO3– being 0.4 nm (Table ). The possible reasons for lysine and NO3– interactions are as follows: 1) the strong electrostatic interactions between positively charged lysine and anions and 2) the formation of hydrogen bonds between lysine’s sidechain and NO3–.[41] Consequently, these strong interactions suppressed the formation of β-sheets in the secondary structures of Aβ16–20 peptides in the NaNO3 environment. Representative snapshots of the systems under study are shown in Figure . In addition to NaNO3, as shown in Figure b and Table , the cation of NH4Cl interacted with the phenylalanine (F) residues of Aβ16–20 via cation−π interactions, which might have enhanced the β-sheet formation.[42]

Table 2

Ratio of Each Component in the Amide I Band of the FT-IR Spectra of Aβ16–20 (6 mg/mL) without and with Salts (0.15 M)

	Aβ	Aβ + NaCl	Aβ + NH4Cl	Aβ + NaNO3
monomer	0.21	0.21	0.19	0.23
aggregate	0.72	0.73	0.74	0.71
β sheet	0.067	0.062	0.075	0.057

Figure 3

Increase in each separated peak (fold-change) in the amide I band of Aβ16–20 dissolved with salts due to CeF3 NPs. The fold changes of the peaks of the (a) monomer (1674 cm–1), (b) aggregate (1640 cm–1), and (c) β-sheet (1618 cm–1) of Aβ16–20 (6 mg/mL) dissolved in D2O containing each salt (0.15 M) due to the coexistence of CeF3 NPs (1 and 3 mg/mL) are shown.

Table 3

Average Distances (in nm) between the Centers of Mass of the Salt Ions and Aβ16–20 Peptide Residues at the End of the Simulation

	NaCl		NH4Cl		NaNO3
peptide residue	Na+	Cl–	NH4+	Cl–	Na+	NO3–
K-16	1.21 ± 0.30	0.77 ± 0.20	1.25 ± 0.32	0.76 ± 0.23	0.95 ± 0.28	0.40 ± 0.03
L-17	1.13 ± 0.29	0.93 ± 0.21	1.30 ± 0.34	1.01 ± 0.24	1.06 ± 0.22	0.67 ± 0.08
V-18	1.30 ± 0.30	1.05 ± 0.23	1.31 ± 0.33	1.11 ± 0.22	1.09 ± 0.23	0.58 ± 0.04
F-19	1.31 ± 0.30	1.02 ± 0.23	1.20 ± 0.32	1.16 ± 0.26	1.08 ± 0.26	0.88 ± 0.14
F-20	1.14 ± 0.29	0.96 ± 0.22	1.11 ± 0.35	1.06 ± 0.30	0.97 ± 0.23	0.82 ± 0.17
average	1.22 ± 0.30	0.95 ± 0.24	1.23 ± 0.34	1.02 ± 0.29	1.03 ± 0.20	0.67 ± 0.20

Figure 4

rdf plots of the interactions between Aβ16–20 peptides and the (a) anions and (b) cations. The results were averaged among four peptides present in the systems at the end of the MD simulations, when the systems were stabilized (last 10 ns of the MD run).

Figure 5

Representative snapshots of the peptide aggregates and ions within 0.1 nm of peptides in the systems under study: (a) no salt, (b) 0.15 M NaCl, (c) 0.15 M NH4Cl, and (d) 0.15 M NaNO3. Coloring methods in VMD: 1. Secondary structure of the peptide: beta sheet = yellow, beta bridge = tan, bend = cyan, turn = cyan, and coil = white. 2. Ions: Na+ = green, NH4+ = blue and white, Cl– = purple, and NO3– = blue and red.

The solvent accessible surface areas (SASAs) of the peptides were further studied to compare the aggregation kinetics under different environments (Figure ). The total SASA (SASA0) of the four peptides in the beginning of the simulation was 39 nm2, which during the 100 ns of the simulations, decreased to ∼22 nm2 (SASA100), indicating peptide aggregation (Figure a). The initial aggregation kinetics was quantified by estimating the time when the total SASA of peptides reached 34 nm2 (SASA34) during the first 30 ns of the simulations (Figure b). According to SASA plots (Figure b), enhanced aggregation kinetics was observed in the presence of 0.15 M NH4Cl (SASA34 was reached in 8 ns), which was related to the enhanced formation of beta sheets, observed from IR spectra (as shown previously in Table ). The slowest aggregation kinetics was observed in the system with 0.15 M NaNO3 (SASA34 was reached in 19 ns), which corresponded to the retention of monomers in this environment, consistent with the results of IR absorption (shown previously on Table ).

Figure 6

Time evolution of the total SASA of the Aβ16–20 peptides in the systems under study: (a) within 100 ns of the MD run and (b) within 30 ns of the MD run. The average values of each 200 ps of the simulation run were plotted, which corresponded to 50 frames of the run.

Conclusions

We evaluated the changes in the secondary structure and assembly of the Aβ peptide (Aβ16–20) due to CeF3 NPs using liquid film FT-IR measurements and MD simulations. CeF3 NPs were found to locally concentrate Aβ16–20 on their surfaces, possibly due to the hydrophobic interaction between NPs and Aβ16–20 in aqueous environments, and promote the β-sheet formation of Aβ16–20. The concentrated Aβ16–20 on the NP surface formed ordered hydrogen bonds to form a β-sheet. This increase in the β-sheet formation of Aβ16–20 on the NP surfaces was suppressed in the presence of NH4+ and NO3– ions. Hydrogen bonding between Aβ peptides were dominant when concentrated on CeF3 NP surfaces in the absence of NH4+ or NO3–. In the presence of NH4+ or NO3–, the hydrogen bonding was suppressed due to dominant bonding between the NPs and Aβ peptides. The formation of the β-sheet structure of Aβ peptides was promoted in the presence of NH4+ ions, whereas it was suppressed in the presence of NO3– ions regardless of the presence/absence of CeF3 NPs, which can be explained by the electrostatic interaction between the lysine residue (amino group) of Aβ peptides and the ions. Although this study was performed using the Aβ16–20 peptide, future research will be conducted using full-length Aβ (Aβ1–42) to reveal more realistic in vivo phenomena. The analysis technique using FT-IR spectroscopy and MD will contribute to comparative studies of the effect of NPs on the molecular state of proteins under various physicochemical conditions.

Materials and Methods

Materials

Cerium (III) chloride heptahydrate (CeCl3·7H2O), oleic acid, deuterium oxide (D2O), and amyloid beta (16–20) peptide [Aβ16–20; Ac-Lys-(Me)Leu-Val-(Me)Phe-Phe-NH2] were purchased from Sigma-Aldrich Co. (St Louis, MO, USA), and 1-octadecene was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). The N-methylated form of Aβ16–20 is a commercially available good model for the investigation with increased stability; however, possible changes in the peptide conformation and aggregation state associated with a possible increase in hydrophobicity should be noted. Sodium hydroxide, ammonium chloride (NH4Cl), sodium nitrate (NaNO3), sodium chloride (NaCl), methanol, ethanol, hexane, cyclohexane, and 20% deuterium chloride solution (DCl) (5.34 mol/L) in D2O were purchased from Fujifilm Wako Pure Chemical Co. (Osaka, Japan). Ammonium fluoride was purchased from Kanto Chemical Co., Inc. (Tokyo, Japan). All reagents were used without further purification.

Synthesis and Characterization of CeF3 NPs

Fluoride NPs were synthesized via thermal decomposition.[43] CeCl3·7H2O (1 mmol) was dissolved in distilled water (3 mL), mixed with oleic acid (12 mL) and 1-octadecene (30 mL), and stirred at 100 °C for 20 min and at 160 °C for 40 min in a nitrogen atmosphere, giving cerium oleate. After cooling to 50 °C, sodium hydroxide (2.5 mmol) and ammonium fluoride (4 mmol) dissolved in methanol were slowly added to the cerium oleate sample, and the sample was heated at 100 °C for 20 min and further at 310 °C for 50 min in a nitrogen atmosphere. The NPs collected via precipitation were purified using centrifugal washing (20 000 g, 10 min, ×3) with a hexane–ethanol mixed solvent and dispersed in cyclohexane. The NPs were characterized using XRD (Rint-Ultima 3, Rigaku Co., Tokyo, Japan) and DLS (ELSZ-2000ZS, Otsuka Electronics Co., Ltd., Osaka, Japan).

FT-IR Spectroscopy for Samples in Solution

Fluoride NPs (27 mg/mL) in cyclohexane (750 μL) were slowly added dropwise into 1 mmol/L DCl solution in D2O (750 μL) and stirred for 16 h to remove cyclohexane via evaporation and to exchange the dispersion media with DCl/D2O. Aβ16–20 was dissolved in D2O at 8 mg/mL. The Aβ16–20 solution (8 mg/mL) in D2O with and without NH4Cl, NaNO3, or NaCl (0.2 M) was mixed with different concentrations (3–27 mg/mL) of the NP dispersion in 1 mmol/L DCl solution at a 3:1 volume ratio (thus, the final concentration of Aβ16–20 in the mixed samples was 6 mg/mL). The final concentrations of the NPs were set at 1–9 mg/mL because the concentration order of milligrams per milliliter is the dose commonly used for imaging contrast agents for visualizing blood flow.[27,36,44] FT-IR spectra including amide bands were recorded using an FT/IR-6200 spectrometer (Shimadzu Co., Kyoto, Japan) for the mixed samples sandwiched between two CaF2 plate windows (spacer 0.025 mm). The analysis was performed for each sample within 30 min after mixing Aβ16–20 with the NP dispersion.

MD Simulations

MD simulations were performed using GROMACS 2019.6 software with a GROMOS 54A7 force field. Four Aβ16–20 peptide monomers with a concentration of 6 mg/mL were inserted in a 9 × 9 × 9 nm3 box. The simulations were performed in the absence of salts and in the presence of 0.15 M NaCl, NH4Cl, and NaNO3 solutions. The MD run was performed for 100 ns for each system, following the methodology described in a previous study.[15]

Analysis of the MD Simulations

Formation of peptide aggregates and kinetics of aggregation were studied via SASA analysis. The interactions between ions and peptide residues were studied in the last 10 ns of the simulations, when the peptide aggregates were produced. The rdf and intermolecular distance analyses were performed using the centers of mass of the peptides, averaged among four peptides. Visual molecular dynamics (VMD) software was used for the visualization of the systems under the study.