Food-grade amorphous SiO₂ NPs were purchased from Evonik Industries AG (Essen, Germany). Materials used were as follow: acacia honey (Dongsuh Food Co., Ltd., Seoul, Republic of Korea), D-(+)-glucose (Sigma-Aldrich, St. Louis, MO, USA), D-(−)-fructose (Sigma-Aldrich), sucrose (Sigma-Aldrich), D-(+)-maltose monohydrate (Sigma-Aldrich), skim milk powder (Seoul Milk, Seoul, Republic of Korea), casein sodium salt from bovine milk (Sigma-Aldrich), extra virgin olive oil (imported from Spain, Beksul, CJ CheilJedang, Seoul, Republic of Korea), 37-component fatty acid methyl esters (FAMEs; Sigma-Aldrich), hexane (Sigma-Aldrich), sodium hydroxide (Sigma-Aldrich), biphenyl (Sigma-Aldrich), methanol (Samchun Chemical Co., Ltd., Gyeonggi-do, Republic of Korea), acetonitrile (HPLC grade, Samchun Chemical Co., Ltd.), water (HPLC grade, Samchun Chemical Co., Ltd.), and PBS buffer (NaCl 137 mM, KCl 2.7 mM, Na₂HPO₄ 10 mM, KH₂PO₄ 1.8 mM; Dongin Biotech. Co., Republic of Korea).

Particle size and morphology were determined by scanning electron microscopy (SEM; FEIQUANTA 250 FEG, Hillsboro, OR, USA). Zeta potentials and hydrodynamic radii of NPs in aqueous suspension and food matrices were measured with Zetasizer Nano System (Malvern Instruments, Worcestershire, UK). Stock solution of SiO₂ NPs [50 mg/ml in distilled water (DW)] was prepared, stirred for 30 min, and diluted to designated concentrations just prior to experiments.

Particles (5 mg/ml) were dispersed in 10% honey, 1 mg/ml of skim milk solution, olive oil, and PBS buffer, respectively. After different incubation times at 25°C, supernatants were collected by ultracentrifugation (16,000 × g) for 15 min, and analysis of Si in the supernatants was performed as described previously (Paek et al., 2014).

Different concentrations (1, 2, 5, and 10%) of acacia honey were prepared in DW and incubated with 5 mg/ml NPs with shaking at different temperatures (4, 25, and 40°C). 10% concentration of honey, containing ~42.4% fructose, ~29.6% glucose, ~0.2% sucrose, and ~0.1% maltose, was used as the highest concentration due to high viscosity over this level. Fructose, glucose, sucrose, and maltose at each concentration of 1, 2, and 5% were mixed and used as sugar mixtures to mimic honey matrix containing only equivalent amounts of each sugar, without trace nutrients. After designated incubation times (1, 24, 48 h, and 7 d), the samples were centrifuged at 23,000 × g for 1 h. The supernatants were analyzed after washing with distilled and deionized water (DDW) or without washing, and then, filtered through syringe filter (Agela Technologies, Wilmington, DE, USA). Saccharide concentrations were quantified by high performance liquid chromatography (HPLC) using a Shimadzu HPLC system (Kyoto, Japan), equipped with RID-10A refractive index detector, on a Hypersil APS-2 column (250 × 4.6 mm i.d., 5 μm, 120 Å, Thermo Fisher Scientific, MA, USA). The mobile phase was acetonitrile:water (80:20, v/v) and flow rate was set at 1 ml/min. Column temperature was maintained at a constant 40°C and injection volume of sample was 20 μl. Each experiment was repeated three times on separate days.

Different concentrations of SiO₂ NPs (1, 2.5, and 5 mg/ml) were suspended in 1 mg/ml of skim milk solution (in DW) or 0.35 mg/ml of casein solution (in DW), and incubated with shaking at different temperatures (4, 25, and 40°C). After designated incubation times (1, 24, 48 h, and 7 d), the suspensions were subjected to protein fluorescence quenching analysis using a luminescence spectrometer (SpectraMax® M3, Molecular Devices, CA, USA). Excitation wavelength was set at 280 nm and fluorescence emission intensity was measured at wavelength from 300 to 420 nm. Quenching ratio was calculated as (I0-I)/I0, where I0 and I stand for basal fluorescence emission intensity in controls (NPs-untreated proteins) and experimental groups (NPs-treated proteins), respectively. Each experiment was repeated three times on separate days.

SiO₂ NPs (50 mg/ml) were dispersed in olive oil for 30 min and diluted to 20 mg/ml prior to experiments. NPs suspended in olive oil were incubated with shaking for 1, 24, and 48 h at different temperatures (4, 25, and 40°C). After centrifugation at 23,000 × g for 1 h, fatty acid methyl esters were prepared by alkaline transmethylation (Liang et al., 2011). The supernatants (1 μl) were spiked with internal standard solution (final concentration of 100 μg/ml), and 1 ml of 0.4 M NaOH-CH₃OH was added and reacted for 5 min with ultrasonification. The methyl esters were extracted with 5 ml of hexane, diluted to 3-folds, and analyzed by gas chromatography-mass spectrometry (GC-MS). A standard mixture (FAMEs, 1 ml of 10 mg/ml) was dissolved in 9 ml of hexane and the standard solution was diluted to 10, 20, 50, 100, 200, and 500 μg/ml concentrations, and then, spiked with internal standard biphenyl (100 μg/ml). The adsorbed fatty acids on NPs were calculated after subtraction of reduced fatty acids in the supernatants from those in olive oil control. Each experiment was repeated three times on separate days. All GC-MS analyses were performed with an Agilent 5977E GC-MSD system (Agilent Technologies, CA, USA), including 7820A GC instrument coupled with a 5977E MS detector. A StabilWax® capillary column (30 m × 0.25 mm, 0.25 μm thickness, Restek, PA, USA) was used, and column temperature was hold for 1 min at 50°C, programmed from 50 to 200°C at the rate of 25°C/min, from 200 to 230°C at the rate of 3°C/min, and then hold for 23 min at 230°C. The injection temperature was kept at 250°C and the carrier gas was helium. The column flow was 1 ml/min and a sample of 3 μl was injected with a split ratio of 5:1. The ion source temperature was 230°C and the samples were ionized by electron impact ionization at 70 eV. Selected-ion monitoring was performed at m/z 55, 67, 74, 79, and 87.

Two different concentrations (5 and 10 mg/ml) of SiO₂ NPs were dispersed in 1 ml PBS and reacted with shaking for various times (1, 24, 48 h, and 7 d) at different temperatures (4, 25, and 40°C). To remove unbound minerals from NPs, the samples were centrifuged at 23,000 × g for 1 h and washed with DDW. This procedure was repeated three times. The aliquots were digested with 10 ml of ultrapure nitric acid at ~180°C and diluted with 3 ml of DDW. After filtering through syringe filter (0.22 μm, Agela Technolgies), total Na concentrations were determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, JY2000 Ultrace, HORIBA Jobin Yvon, longjumeau, France). Each experiment was repeated three times on separate days.

Results were expressed as means ± standard deviations. Experimental values were compared with corresponding untreated control values. Statistical analysis was performed using the Student's test for unpaired data and one-way analysis of variance (ANOVA) with Tukey's Test in SAS Version 9.4 (SAS Institute Inc., Cary, NC, USA) was carried out to determine the significances of intergroup differences. Statistical significance was accepted for p-values of <0.05.

Particle morphology, primary particle size, and size distribution were examined by SEM. Figure 1 shows that food-grade SiO₂ NPs dispersed in DW had irregular particle morphology with an average particle size of 24.1 ± 3.5 nm. Zeta potential and hydrodynamic size of SiO₂ NPs were determined to be −28.2 ± 1.0 mV and 287.6 ± 1.7 nm, respectively. Solubility of SiO₂ NPs in DW was not detected and did not increase in the presence of food matrices.

Zeta potential values of SiO₂ NPs in honey changed to less negative charges as honey concentration and incubation time increased (Table 1), regardless of temperature (Table 2). Hydrodynamic radii of SiO₂ NPs in honey also significantly increased as honey concentration increased (Table 1), without effect of incubation time and temperature (Tables 1, 2). However, zeta potentials and hydrodynamic radii of NPs could not be detected at concentrations of more than 2% honey after incubation for 7 d due to the formation of high aggregation (Supplementary Figure 1).

Since honey contains different amounts of saccharides and other trace nutrients, the interactions between SiO₂ NPs and saccharides were further evaluated using sugar mixtures composed of equivalent amounts of fructose, glucose, sucrose, and maltose, in order to investigate the effects of only saccharides on interactions. It is worth noting that these are major four saccharide components in acacia honey. The highest concentration of each saccharide was set at 5%, based on maximum solubility of both fructose and maltose in DW. Zeta potentials of SiO₂ NPs in sugar mixtures did not significantly change, except at 5%, and were not affected by incubation time and temperature (Tables 2, 3). Interestingly, hydrodynamic sizes of NPs decreased in the presence of sugar mixtures in a time-dependent manner (Table 3). No aggregation tendency was observed in sugar mixtures (Supplementary Figure 1).

When the interactions between SiO₂ NPs and saccharides in honey were quantified by HPLC, Figure 2A shows that fructose and glucose, the most abundant two saccharides in honey, interacted with SiO₂ NPs in a concentration-dependent manner at 25°C, while no interactions between SiO₂ NPs and sucrose or maltose were found. NP interaction with fructose was slightly higher than that with glucose at more than 5% honey (p < 0.05). However, the interactions between SiO₂ NPs and glucose or fructose in honey were not affected by incubation time (Figure 2B) for 48 h. Meanwhile, the interactions between SiO₂ NPs and glucose or fructose increased as temperature increased, especially, low interaction levels were detected at 4°C, compared to 25 and 40°C (Figure 2C).

Similar tendency was found in sugar mixtures containing equivalent amounts of fructose, glucose, sucrose, and maltose. Concentration-, but not time-dependent interactions between SiO₂ NPs and saccharides were found (Figures 3A,B), and these interactions significantly increased when incubation temperature increased at 40°C (Figure 3C). In particular, all different types of saccharide interacted with NPs in a similar manner. Significant differences in the interactions between saccharide types were not remarkably found (p > 0.05). Overall, < 4.6 and 3.1% of saccharides in honey and sugar mixtures, respectively, were determined to interact with SiO₂ NPs. It should be noted that quantitative analysis was performed after ultracentrifugation of reacted NPs with saccharides, collecting precipitated NPs, and detaching adsorbed saccharides on NPs with HPLC eluent. Whereas, no adsorbed saccharide was detected after washing the centrifuged and precipitated NPs with DDW.

Skim milk and its main protein component, casein were used to determine the interactions between SiO₂ NPs and proteins. Skim milk solution at 1 mg/ml (in DW) was applied for protein fluorescence measurement, because this concentration exhibited the highest fluorescence intensity without precipitation (data not shown). Based on casein content in skim milk, 0.35 mg/ml of casein solution was used for comparative study. Zeta potential values of SiO₂ NPs significantly changed to less negative direction in skim milk solution, whereas more negative zeta potentials were found in casein solution (Table 4). Effect of temperature on the interactions was not found (Table 2; p > 0.05). Aggregation was observed after incubation for 7 d at 25 and 40°C in both skim milk and casein solutions treated with NPs (Supplementary Figure 1). Meanwhile, hydrodynamic radii of NPs were not remarkably influenced by incubation time, temperature, and the presence of skim milk or casein (Tables 2, 4).

Interactions between SiO₂ NPs and protein were further estimated by measuring protein fluorescence quenching ratio in the presence of NPs. Figure 4 shows that fluorescence quenching ratio immediately increased just after adding NPs in skim milk solution at all temperatures tested in a NP concentration-dependent manner. Remarkably high fluorescence quenching ratio (more than 60% fluorescence quenching) was induced at 25°C after incubation for 7 d (Figure 4B), where aggregation was observed (Supplementary Figure 1). Meanwhile, fluorescence quenching ratio significantly increased at 25 and 40°C compared to 4°C. However, overall fluorescence quenching ratios were < 40% and blue or red shift was not observed.

When the interactions between SiO₂ NPs and casein, a major protein in skim milk, was assessed, similar tendency was obtained (Figure 5). NP concentration-dependent increase in fluorescence quenching ratio was found, and 60–70% quenching was induced by NPs at 25°C after 7 d of incubation. Interactions between NPs and casein at 25°C were comparable to those at 40°C, except at 7 d post-incubation, while significantly reduced fluorescence quenching was observed at 4°C. Blue shift was only observed at 40°C after 48 h of incubation (Supplementary Figure 2). On the other hand, fluorescence quenching ratios of NPs in 0.35 mg/ml of casein solution (Figure 5) were similar to those in 1 mg/ml of skim milk solution (Figure 4).

Zeta potentials and hydrodynamic radii of SiO₂ NPs in the presence of olive oil could not be detected because of intense yellow color and high viscosity of olive oil. GC-MS analysis reveals that olive oil used in the presence study contains ~59.3% oleic acid, ~11.4% palmitic acid, and ~5.7% linoleic acid. Because SiO₂ NPs have hydrophilic surface characteristics (Jesionowski and Krysztafkiewicz, 2002) and slightly bound fatty acids on the surface of NPs are likely to be easily detached during washing with organic solvents, the interactions between SiO₂ NPs and lipids were estimated by measuring fatty acid composition in the supernatant after reaction with NPs, followed by subtraction of reduced fatty acids from those in olive oil control. The results show that < 1.8% oleic acid and 0.4% palmitic and linoleic acids interacted with SiO₂ NPs, and the effects of incubation time and temperature on the interactions were not found (Figure 6; p > 0.05). However, interactions at 4°C after 48 h and at all temperatures tested after 7 d could not be detected, resulted from gelation of SiO₂ NPs in olive oil (Supplementary Figure 1).

PBS buffer was used to investigate the interaction effects of minerals on the surface of SiO₂ NPs. Zeta potentials of NPs significantly and rapidly changed to more negative charge in PBS without effects of incubation time, NP concentration, and temperature (Tables 2, 5). DLS data reveal that significantly reduced size distribution of NPs was found just after adding NPs in PBS and their reduced hydrodynamic radii were maintained for 7 d. Remarkable aggregation was not observed even after incubation for 7 d (Supplementary Figure 1).

Quantitative analysis was also performed by measuring the amount of bound Na⁺ to NPs, the most abundant mineral in PBS, by ICP-AES. As shown in Figure 7A, the interactions between SiO₂ NPs and Na⁺ increased as NP concentration and incubation time increased. However, no effect of temperature on the interactions was found (Figure 7B; p > 0.05). Overall, < 1.7% of Na⁺ was determine to interact with SiO₂ NPs (Figure 7B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.