High antioxidant activity and low toxicity of gold nanoparticles synthesized with Virola oleífera : an approach using design of experiments

Metallic nanoparticles synthesized by plant extracts and biomolecules represent one of the most promising frontiers in the antioxidants field. Plant compounds apart from act as reducing agents, also allow functionalizing these nanoparticles having multiple applications in different industries. However, the low reproducibility and the difficulty in size, shape and stability control have made it difficult to produce on a large scale. Here we report an optimization process to obtain gold nanoparticles (AuNPs) from green synthesis using a factorial design aiming to present a controlled synthesis using Virola oleifera as a reductant agent. We evaluate the influence of the reaction time, temperature, stirring rate, pH and extract concentration monitoring the Localized surface-plasmon resonance (LSPR) band shift (Δλ). The nanoparticles were characterized using Zeta potential, UV-vis, DLS, TEM, Raman spectroscopy, FTIR and XRD. The as-synthesized AuNPs were stable and homogenous, octahedral, monodisperse, and in shape and size and showed great antioxidant activity determined with ABTS+ and DPPH. Furthermore, these nanoparticles presented low cytotoxicity. Finally, using the factorial design, we were able to develop an optimal path for green gold nanoparticles with high antioxidant activity, low toxicity and good morphological characteristic.

AuNPs are typically synthesized via physical and chemical, which are usually costly and environmentally hazardous. Therefore, the green biosynthesis has attracted increasing attention, increasing the interest eco-friendly synthesis and sustainability.
The trend in recent years has focused on the methods under the concept of "Green Chemistry", aiming the utilization of nontoxic capping agents, less hazardous reducing agents, and selection of environmentally benign solvents [6][7][8][9]. However, these methods are subject to many fluctuating aspects such as operational conditions (temperature, time, concentration, pH, stirring rate, among others) and the wide variety of organic compounds present in plant extracts, causing some drawbacks, such as polydispersity, poor stability and even weak biocompatibility [10]. Thereby, the standardization that allows a monodisperse, stable, non-toxic and bioactive nanoparticles synthesis remain a challenge.
The traditional technique of 'one-factor-at-a-time' to optimize the factors that influence nanoparticles growth requires a lot of work and time without an accurate mathematical model equation to describe the whole process. Therefore, the aim of this work was to standardize AuNPs biosynthesis using a complex resin with a plethora of biomolecules, Virola oleifera, evaluating the reaction time, temperature, stirring rate, pH and extract concentration on the size and shape of nanoparticles using complete factorial design 3 2 .
Then, some physico-chemical parameters were characterized and finally, the AuNPs showed good antioxidant activity in radical scavenging ABTS and DPPH analysis, low toxicity and good stability.
In summary, data presented here demonstrated that as synthesized AuNPs based on green synthesis needs to be controlled and optimized to AuNPs present and preserve its biological activity. This study might provide a good candidate and offer new insights for future translational study on AuNPs applications.

Plant material
V. oleifera power was obtained according to Pereira et al (2015) [11]. The resin was obtained by making cuts in the bark of the V. oleifera tree and collected in an amber glass container, keeping it in the dark and at 4°C during the transportation to the laboratory. The resin was dried at 50°C and then ground to obtain 10g of dry resin.

Synthesis of gold nanoparticles
The resin of the plant was added to a gold precursor solution in a ratio of three to one (m/m). The gold precursor was tetrachloroauric acid (Sigma-Aldrich, St. Louis, MO, USA) in a concentration of 2.5 x 10 -4 mmol/L. For the synthesis, a factorial design 32 was carried out with 3 levels and 2 variables. The synthesis was performed according to the experimental design.

Experimental design and statistical analysis
The choice of variables and each of their levels was previously determined. Details on these experiments can be found in the Supporting Information (Table S1). Then, as the next step, a fractional factorial design 2  , with 2 levels and 5 variables, was assembled to know which is the main variable that influences the nanoparticle synthesis process ( Table 1). The reaction time, temperature, stirring rate, pH and extract concentration were also evaluated in this step. In the final step, a 3 2 factorial design was made with 3 levels and 2 variables to determine the optimum conditions for the nanoparticle synthesis ( Table 2). For this step, the pH and reaction time were evaluated as independent variable. The effects of each selected variable were analyzed using STATISTICA version 10.0. An analysis of variance (ANOVA) of the data was also performed, and the values were considered significant at values of p < 0.05. The pareto diagram that identifying the significant variables are show in supplementary material (S1). The optimal values of the independent variables were determined using a three-dimensional analysis of the response surface of the independent and dependent variables.

Gold nanoparticle characterization. (UV-Vis, TEM, DLS, FTIR, RAMAN and X-rays)
The optical properties of the solutions were analyzed using the UV-vis absorption spectrophotometer (UV-1800 Shimadzu, Japan). Image J software was used to analyze the mean diameter and aspect ratio of 1000 particles from transmission electron microscopy images ((TEM-JEM-1400, JEOL, USA Inc. operated at 120kv). Diffraction Xray analysis was performed using the Phillips PW 1710 diffractometer (Cuk radiation) to evaluate the crystallographic structure of nanoparticles. For the analysis of stability and hydrodynamic size, the Microtac Zetatrac particle analyzer was used. Infrared spectrometry and Raman scattering were performed in the FTIR (FT-MIR FTLA 2000 Bomem) and Raman (ALPHA 300 R Confocal Raman Spectrometer). The total concentration of AuNPs was determined using ICP-MS (Perkin Elmer, Optima 7000, USA). The flocculation parameter were used for the stability analysis of the synthesized nanoparticles according to the Weisbecker [12] description, later modified by Mayya [13], taking and analyzing specific UV-vis measurements.

Statistical and full factorial design analysis
The applied experimental design allowed the identification of the experimental factors fluencing the physicochemical properties of the AuNPs that directly affect size and consequently their monodispersion.
Previous fractional factorial design (2 5-1 ) (Supplementary material) studies was applied to determined the variables that most influenced monodispersion of as synthesized AuNPs. Delta lambda (Δλ) was the factor with the most significant variables, pH and time.
UV-vis absorption spectra show three regions well defined were according to the pH ranges ( Fig. 1). When pH changes from acid to basic, Δλ decrease. The pH variation influences directly the Δλ, and consequently the size variation of AuNPs. The presence of deprotonated (at basic pH) and protonated (at acid pH) form of carboxylic acid in the V. oleifera extract can be relationed with this phenomenum. When COOgroup interact with gold atoms in a basic pH there has a decrease in the size of AuNPs. On the other hand the interaction between COOH results on increase AuNPs size [14].
To investigate influence of the time on Δλ variation, we vary the time on three levels (1, 6 and 11 min) within each pH range (4.7, 7.7 and 10.7) ( Table 3). No significant variation was identified that would interfere in the Δλ.
Through ANOVA analysis the pH was the most influential variable (Table 4). It should be highlighted that complete factorial design showed that none combinatorial effects was significant for Δλ.    The Response Surface Graph (Fig. 2) was applied to determine the best region where the optimal response occurs. The reduction, nucleation, growth and stabilization of nanoparticles occurs very fast.

Characterization of Gold Nanoparticles
At the end of full factorial design, three assays were determined: AuNPs synthesized at values of pH 4.7, 7.7 and 10.7, which can be interpreted as groups of AuNPs of different sizes 35nm, 22nm and 9nm respectively ( Fig. 3A and Fig. 3G).
X-ray diffraction showed a crystal structure of AuNPs (Fig. 3B) TEM images mainly showed spherical and non-agglomerated AuNPs obtained from the three conditions with different pH (4.4, 7.7 and 10.7) (Fig. 3C, Fig. 3D y Fig. 3E). The presence of a cloud around the nanoparticles confirmed the existence of organic compounds formed by molecules that come from the plant extract.
The surface potential in all groups of AuNPs was also measured using the zeta potential.
AuNPs -pH 4.7 and AuNPs -pH 7.7 show moderate stability with a positive charge of 21.59 and 20.37 mV respectively, finally the zeta potential for AuNPs -pH 10.7 was 4.27 mV showing rapid aggregation [20][21][22][23] (Fig. 4A). It can be inferred that the molecules present on the surface of the AuNPs are mainly composed of positively charged groups that are responsible for the good stability of the nanoparticles. This can be attributed to the presence of polyphenols, flavonoids and proteins from the plant extract [24,25].
The hydrodynamic size of the AuNPs was measured by DLS to determine the size of the AuNPs coated with all the phytochemicals present in V. oleifera. AuNPs -pH 4.7 and AuNPs -pH 7.7 had hydrodynamic sizes close to 100nm, while AuNPs -pH10.7 had hydrodynamic sizes close to 1000nm (Fig. 4B, Fig. 4C and Fig. 4D). When comparing these sizes with those obtained by TEM they are 35 nm, 24 nm and 8nm respectively   SC / Final volume of colloidal solution [26][27][28]. Table 5 shows the concentration of AuNPs obtained for each run, observing a high concentration of nanoparticles in the prepared colloids.

Surface Analysis
In order to identify the functional groups of the formed AuNPs, Fourier transform infrared spectroscopy (FT-IR) and RAMAN spectroscopy were used. FT-IR bands revealed for both extract and as synthesized AuNPs are found to be located as following, 1601 cm-1 which is attributed to the C = O stretch of the carbonyl group (Fig. 5A), at 1433 cm-1 attributable to axial deformation of CH aliphatic groups ; at 1510 cm-1 and 1513 cm-1 due to the presence of aromatic structures C = C, at 1280 cm-1 it is related to the vibratory stretch C-O of the alcohol bonds, at 1090 cm-1 it is related to the ester group, and finally at 1034 cm-1 it is related to C-H bending vibrations [29].
In the Raman spectrum, the following peaks were observed at 1347 cm-1 and 1579 cm-

Stability of AuNPs
Stability tests were performed only in the midpoint (pH 7.7). The formulation prepared at pH 7.7 proved to be stable between pH 7 to 11 and presented a high degree of flocculation at an acidic pH and a very basic pH ( Fig. 6A and Fig. 6B). The sample prepared at pH 7.7 shows to be stable at a salt concentration of less than 10mM NaCl (Fig. 6C and   Fig. 6D).

Antioxidant activity of AuNPs
In this study, two antioxidant evaluation methods, ABTS + and DPPH, were used to investigate the antioxidant activity of AuNPs -pH7.7. The radical scavenging activity of AuNPs was estimated by comparing the percentage of ABTS + radical inhibition with Trolox and V. oleifera. The results revealed the existence of an effective radical scavenging activity in the AuNP solution compared to Trolox (Fig. 7A). The radical scavenging activity was 73.4%, for AuNPs -pH 7.7. The V. oleifera extract showed a greater inhibitory activity with 71.6% greater than the Trolox which was 52.7%. Similarly, the DPPH assay showed a similar pattern of reducing activity for AuNPs -pH 7.7, V.
oleifera and Trolox (Fig. 7B) being 69%, 66.6% and 53% respectively. As can be seen, the antioxidant activity of the AuNPs was greater than the antioxidant activity of the V.
oleifera extract, possibly due to modifications in the electrostatic density of the surface molecules [32,33].

Cytotoxicity of the AuNPs
An Alamar Blue assay was performed to assess the cytotoxicity of AuNPs by determining the survival percentage of the J774A.1 cell line (Fig. 8). The AuNPs presented low cytotoxicity at low concentrations, with an IC50 value of 130 ug / ml. It has been reported that the cytotoxicity of the AuNPs depends on its size, shape, surface charge and the exposed cell line [34,35], so that the smaller the nanoparticle size, the greater the distribution in the cell and the greater the toxicity, which could explain why the V.
oleifera extract is more cytotoxic in the same concentrations as the AuNPs. The surface charge and the size of the AuNPs (around 100nm) would cause less affinity for the cell [36][37][38] and consequently higher survival percentage.

CONCLUSIONS
We describe a controlled and reproducible process of green synthesis of gold nanoparticles, which was achieved through a fractional factorial design (2 5-1 ) to evaluate the variables that interfere in the synthesis of gold nanoparticles using V. oleifera extract. Through the complete factorial design (3 2 ) the synthesis process was optimized, identifying the reaction time and pH as the variables that most influenced during the synthesis to obtain nanoparticles of different sizes.
The synthesized nanomaterials were monodisperse, almost spherical and showed good stability, characteristics determined by UV-vis, TEM, DLS and Zeta Potential.
Furthermore, the similarity of the spectra showed in FTIR and Raman between the V. oleifera extract and the AuNPs, indicates the adsorption of the extract on the surface of the AuNPs. An evaluation of the flocculation parameter was carried out, with the aim of applying these NPs in biological systems, showing that the NPs were stable in a pH range (7 to 11) and ionic strength (0.05 to 10 mM). L -1 NaCl).
Finally, the AuNPs showed even greater antioxidant activity than the V. oleifera and trolox extract with low cytotoxicity at low concentrations. This study might provide a good candidate and offer new insights for use of green gold nanoparticles as antioxidant on the market.