The fatty acid composition of buriti oil was determined by gas chromatography (model GCMS-QP2010S; Shimadzu, Tokyo, Japan), coupled to an FID flame ionization detector according to AOCS method Ce 1-62 (10). Fatty acid methyl esters were separated according to Hartman and Lago's method (11) on Agilent DB–23 (Santa Clara, CA, USA) capillary column ((50% cyanopropyl)-methylpolysiloxane), l=60m, d_int=0.25 mm, δ_film=0.25 μm. The following oven temperature schedule was employed: 110 °C for 5min, 110-215 °C (5 °C/min), 215 °C for 24 min; detector temperature 280 °C; injector temperature 250 °C, column flow 1.0 mL/min, linear velocity 24 cm/s, entrainment gas helium, 1:50 split, injected volume 1.0 μL. The molar fatty acid composition was calculated using the reference standard for the mixture of fatty acid methyl esters (FAME) obtained from Sigma-Aldrich, Merck (Supelco 37 Component FAME mix, Darmstadt, Germany). FAME were quantified using relative peak areas. Analyses were carried out in duplicate, and the mean and standard deviations of each sample were calculated (12).

Lipidclass (triacylglycerols (TAG), diacylglycerols (DAG), monoacylglycerols (MAG), and free fatty acids) analysis of buriti oil and structured buriti oil was performed byhigh-performance sizeexclusion chromatography (HPSEC). The samples were diluted to 1:100 in tetrahydrofuran and analyzed in a liquid chromatograph equipped with a Rheodyne 7725i injector and a Waters 510 pump (Waters Associates, Milford, MA, USA) with two Ultrastyragel columns (100 and 500 Å; Waters Associates), each with 25 cm×0.77 cm i.d., packed with a styrene–divinylbenzene copolymer (d_internal=10 mm), connected in series, and a refractive index detector (Hewlett Packard, Palo Alto, CA, USA). Conditions were the following: mobile phase tetrahydrofuran (HPLC grade), flow 1 mL/min, and injection volume 20.0 µL (13). The classes of compounds were identified by comparing the elution times with triacylglycerols, diacylglycerols, monoacylglycerols and free fatty acid standards and the analysis was performed in duplicate.

The TAG composition of buriti oil was analyzed by reversed-phase HPLC, following the method described by Holčapek et al. (14) and adapted by Carvalho et al. (15). The chromatographic peaks of the samples were compared with the retention times of chromatographic peaks of soybean oil, which was used as a reference sample (16) and identified based on the expected elution order (considering that retention is increased), and on the fatty acid composition of the samples that were used as a reference. TAG composition of the samples was determined by internal normalization. Analysis was carried out in duplicate.

The levels of α-, β-, γ- and δ-tocopherols were determined according to AOCS method Ce 8-89 (17). Samples were diluted in hexane at a concentration of 0.1 mg/mL and then injected in UHPLC Ultimate® 3000 liquid chromatograph (Dionex, Sunnyvale, CA, USA) with a PerkinElmer (Waltham, MA, USA) Series 200a fluorescence detector (290 nm excitation and 330 nm emission).The microparticulate silica column l=250 mm, d_int=4 mm, with each particle measuring approx. 5 μ with mobile phase composed of HPLC grade hexane (99%) and isopropanol (1%) was used.

The qualitative composition of tocopherols was determined by comparison of peak retention times with those of tocopherol standards (α-, β-, γ- and δ-tocopherols), and the quantitative composition was determined by normalizing the area under the curve. Analysis was performed in triplicate, and mass fraction of tocopherol in oil samples was expressed in mg/kg. The conversion to α-tocopherol equivalent unit (α-TE) was obtained through coefficient 1 for α-tocopherol, 0.5 for β-tocopherol, 0.1 for γ-tocopherol and 0.03 for δ-tocopherol, according to Darnet et al. (18).

β-Carotene content was determined spectrophotometrically, according to the method of de França et al. (19). An aliquot of 0.1 g of oil was diluted in 25 mL of a 7:3. solvent mixture of n-hexane and acetone (>99%, Merck, Darmstadt, Germany) and absorbance was read at 453 nm using a UV-Vis Agilent (Agilent Technologies, Waldbronn, Germany) spectrophotometer. Carotenes in the extracts were calculated in terms of β-carotene, using a standard absorbance curve calibrated with β-carotene (>99%, Merck, Darmstadt, Germany). The standard curve was prepared with 0.1 g of β-carotene diluted up to 100 mL with the 7:3 solvent mixture of n-hexane and acetone. Aliquots were taken from this solution and diluted to five different concentrations. Absorbance was read at 453 nm. Results are given in μg of β-carotene per g of buriti oil. The analysis was performed in triplicate.

Phenolic compounds were determined by the Folin–Ciocalteu method, according to Hrncirik and Fritsche (20). This method is based on the reduction of phosphomolybdic and phosphotungstic acids by phenolic hydroxyls, which produces a blue colour. Phenolic compounds were extracted from the oil with a solution of water/methanol 60:40. Folin-Ciocalteu reagent (Sigma-Aldrich, Merck, St. Louis, MO, USA) was added to suitable aliquots of the extracts. After 3 min, a sodium carbonate solution (35%) was added to the mixture, which was diluted with water to a final volume of 1 mL. Absorbance was measured after 2 h at 725 nm on a Shimadzu spectrophotometer UV-1800 (Kyoto, Japan). The blank contained all the constituents of the reaction except phenolic solution, which was substituted by distilled water. A calibration curve was prepared using a standard solution of gallic acid (Sigma-Aldrich, Merck) at concentrations of 0.01–0.1 μg/mL. Analysis was performed in triplicate, and results are expressed in μg of gallic acid equivalents per g of buriti oil.

The droplet size distribution of NLCs was measured using dynamic light scattering (Mastersizer 2000; Malvern Instruments, Malvern, UK) (24). Results are presented as Z-average particle diameters. The refractive indices of the dispersed and continuous phases used in these calculations were 1.46 and 1.33, respectively. Analyses were performed in triplicate with all samples: NLC_BO, NLC_BO6h and NLC_BO24h; after 1, 15 and 30 days of storage at refrigeration (4 °C) and room (25 °C) temperatures.

Zetapotential (caused by electric charges around droplets) can provide information about interfacial properties. Zeta potential was determined using Zetasizer Nano-ZS equipment (Malvern Instruments) and samples were diluted (1:100) in distilled water (25). Analyses were performed in triplicate with NLC_BO, NLC_BO6h and NLC_BO24h after 1, 15 and 30 days of storage at refrigeration (4 °C) and room (25 °C) temperatures.

Confocal laser scanning microscopy was carried out using a TCS SP5 II microscope (Leica Microsystems, Heidelberg, Germany). The oil phase of NLCs (100 μL) was dyed with10 μL of Nile Red solution in ethanol (1 mg/mL). Samples were mounted on glass slides and examined under 10× and 63×magnification at an excitation and emission wavelength of 543 and 605 nm, respectively. Autofluorescence was analyzed using the built-in software LAS lite (Zeiss Inc., Toronto, Canada) to evaluate the morphology of fat crystals (26). Analyses were performed with NLC_BO, NLC_BO6h and NLC_BO24h after 1, 15 and 30 days of storage at room (25 °C) temperature.

The polymorphic forms of fat crystals were investigated by X-ray diffraction, according to AOCS method Cj 2-95 (27). Analyses were performed using a PW 1710 Philips diffractometer (Malvern Panalytical, Almelo, The Netherlands) with Bragg–Brentano geometry (θ:2θ) and Cu-Kα radiation (λ=1.54056 Å) at 40 kV and 30 mA. Step sizes of 0.02° (2θ scale) and an acquisition time of 2 s were used in the scanning range of 5–40° (2θ scale). Polymorphic forms were identified by the characteristic short spacings of each crystal. Analyses were performed with NLC_BO and NLC_BO24h after 30 and 120 days of storage at room (25 °C) temperature.

Oxygen radical absorbance capacity (ORAC) assay was performed in a 96-well plate (TPP Techno Plastic Products AG, Trasadingen, Switzerland), using fluorescein (Sigma-Aldrich, Merck) as a fluorescent probe, according to the procedures of Dávalos et al. (28). The measurements were performed using a microplate reader (FLUOstar OPTIMA; BMG Labtech, Offenburg, Germany) with fluorescence filters at 485 nm excitation and 520 nm emission. The reaction, performed at 37 °C, was started by thermal decomposition of 2,2’-azobis(2-methylpropionamidine) (AAPH; Sigma-Aldrich, Merck, Steinheim, Germany) in a 75-mM phosphate buffer (pH=7.4). ORAC values were calculated using the difference between the area under the fluorescein decay curve of the sample and the blank (net area under the curve). Standard curves were constructed using 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox, Sigma-Aldrich, Merck) at concentrations between 30 and 1500 μmol/mL. Regression equations between the net area under the curve and antioxidant concentration were calculated. Final values are expressed in μmol Trolox equivalents (TE) per mL of dry matter and analyses were performed in triplicate with NLC_BO and NLC_BO24h.

FRAP assay was carried out according to Benzie and Strain (29). In the dark environment, aliquots of 30 μL of standard or blank samples were added to 90 μL of distilled water and 900 μL of the FRAP reagent (2.25 mL of 0.3 M acetate buffer (Riedel-de Haen, Seelze, Germany), pH=3.6, 225 μL of 10 mM TPTZ (2,4,6-tripyridyl-s-triazine; Fluka Chemicals, Switzerland) in 40 mM HCl and 225 μL of 20 mM iron(III) chloride solution. The mixture was vortexed and 200-μL aliquots were transferred to a clear 96-well microplate. Absorbance was read using a microplate reader (FLUOstar OPTIMA; BMG Labtech) at 595 nm and 37 °C for 30 min of reaction. To achieve the standard curve, Trolox® solutions were prepared at concentrations between 15 and 1500 μmol/mL. Results are expressed in μmol Trolox equivalents (TE) per mL of dry matter and analyses were performed in triplicate with NLC_BO and NLC_BO24h.

NLCs have been shown to be carriers with great ability to disperse bioactive compounds. The addition of solid lipids to emulsions improves stability and allows greater control of the release of bioactive molecules (22).

In this study, NLCs prepared with interesterified buriti oil had small droplets, especially NLC_BO24h sample, which had droplets that were 46% smaller than those of NLC_BO (Table 1). This is a very interesting result, under the same process conditions but with the structured oil, we obtained smaller nanocarriers. The TAG modifications from the biotransformation reaction may have influenced the particle size. The interesterification formed more unsaturated TAG (OOO), we believe, exerting a positive effect on NLC droplet size, even after the addition of Compritol 888 ATO to the system. Nanocarriers prepared with non-interesterified oils have larger droplets and are less stable than those prepared with interesterified oils, as their components have different properties (such as viscosity and melting temperature). Interesterification can produce single-phase lipids that do not undergo phase separation, which results in more stable emulsions (32). Furthermore, interesterification can reduce the types of TAGs initially present in the starting oil, thereby forming a lipid phase with a narrower range of polymorphic forms, greater capacity for interaction with other components, and more organized packing (33).

Storage time and temperature did not influencedroplet size or stability. The properties of lipid nanoparticles are typically affected by many factors, including type and concentration of lipids and surfactants and viscosity of the lipid phase. It has been reported that particle size increases as the viscosity of the lipid phase increases (34) and that NLCs have smaller droplets when oil is added to the system, as liquid fats reduce lipid phase viscosity and surface tension. In contrast, increased surfactant concentration decreases particle size (35), because surfactants facilitate the fragmentation of nanoparticles during homogenization, reducing surface tension at the interface between solid and liquid phases.

In a study by Walker et al. (36), nanoemulsions were prepared using spontaneous emulsification with 10% total oil phase (50% fish oil and 50% lemon oil, by mass) and different surfactant/oil ratio. These nanoparticles were maintained at 5, 20 and 37 °C for 14 days to assess their physical stability. The surfactant concentration also had a large impact on the average particle diameter of the emulsions: as the surfactant/oil ratio increased, the mean particle diameter decreased. In addition, the mean particle size did not depend strongly on storage time or temperature, as occurred in our study. The fact that there was no significant change in the average particle size of the nanoemulsions after being stored at 5, 20 and 37 °C for 14 days suggests that they were stable against flocculation, coalescence and Ostwald ripening.

Yang et al. (37) investigated the effect of different liquid carrier oils on the crystallization and aggregation behaviour of tristearin NLC dispersions. The results demonstrated that NLC suspension stability was strongly affected by the type and amount of the carrier oil. Unsaturated fatty acids and short chain fatty acids result in lower melting points, affecting the stability of NLC dispersions. These results were confirmed by using olive oil (melting point of −6 °C) and palm oil (melting point of 35 °C) for the production of nanoparticles. Olive oil is rich in oleic acid, as is buriti oil; this property allowed for a much more stable dispersion than that obtained with palm oil.

Zeta potential is a parameter used to evaluate the stability and characteristics of particles. It indicates the electric charge on the surface of nanoparticles as well as the electrostatic repulsion between them. Table 2 shows the ζ-potential of NLCs prepared with buriti oil and structured buriti oil after 1, 15 and 30 days of storage at 4 or 25 °C.

A minimum ζ-potential value of ±30 mV is required to prevent coalescence of droplets (38). Interesterification of buriti oil influenced the interfacial properties of droplets under the evaluated conditions. However, although NLC_BO initially had higher ζ-potential values than NLC_BO6h and NLC_BO24h, differences among the ζ-potential values of NLCs became non-significant with storage time, indicating that interesterification helped maintain stability over time. The higher value of ζ-potential, positive or negative, the lower the probability of the droplets coalescing and better their stability (39). Thus, the physical stability of the nanostructured lipid carriers was maintained over time.

All dispersions exhibited negative ζ-potential values, showing that repulsion forces predominated over attraction forces. This property is desirable in NLCs, as it prevents the formation of aggregates and is an indicator of particle stability. NLCs, solid lipid nanoparticles and lipid emulsions also had a negative charge in a study by Fang et al. (40). The negative charge was caused by the anionic fractions of the lipophilic emulsifier (soybean phosphatidylcholine) and the fatty acid glycerides in the lipid core of Compritol and Precirol.

The stability of lipid nanoparticles is affected by the type and concentration of lipids. Niculae et al. (23) used different lipid matrices to prepare NLCs. Carnauba oil was added to the carrier system initially prepared with pomegranate seed oil only, and the ζ-potential increased from −21.7 to −30.9 mV. The components of carnauba oil might have disrupted the surfactant shell, causing a rearrangement of the surface charge and a consequent change in ζ-potential. In addition, the authors found that when oil concentration decreased, the physical stability of nanocarriers slightly increased. In a study by How et al. (41) three different formulations of NLC were prepared, containing olive oil/hydrogenated palm oil (HPO) ratios of 1:9, 2:8 and 3:7 and Polysorbate 80. These authors observed that increase in the oil to HPO ratio had caused increase in the ζ-potential of the NLC, while the increase in surfactant content decreased particle size.

Surfactants also play an important role in the stabilization of NLC lipid particles. Tween 80, the surfactant used in this study, is a non-ionic surfactant that provides steric repulsion through which the semi-solid particles of NLCs are stabilized. If the amount of surfactant is not sufficient, some lipid particles might be uncovered, which can lead to flocculation, aggregation and gelling (42). In a study by Niculae et al. (43), lipid nanoparticles prepared with Tween 80 as the main surfactant had better physical stability than those prepared with Tween 20, showing the importance of surfactants for nanoparticle stability. Thus, interesterification of buriti oil might have allowed for a better interaction between emulsifier and oil, providing greater stability.

As storage temperature did not affect particle stability, microscopic analysis was not performed with samples stored under refrigeration. Fig. 1 shows the morphology of NLCs after 1, 15 and 30 days of storage at room temperature.

The confocal microscopy images (Fig. 1) show individual droplets with a spherical shape. In general, NLCs had a relatively wide droplet size range and droplets were polydisperse, without uniformity. The distribution of NLC_BO droplets was more homogeneous up to 15 days of storage, but droplets were more agglomerated after 30 days than those of other NLCs. NLC_BO6h showed a similar behaviour, but droplet aggregation was lower than that of NLC_BO after 30 days of storage.In contrast, NLC_BO24h had a more heterogeneous structure, exhibiting droplet aggregation throughout the storage period (44). Wegmüller et al. (45) reported the occurrence of particle agglomerates and fat crystals in the microcapsules developed in their study. Agglomeration might have occurred because microparticles were not completely solidified when cooled. The authors observed that small particles showed fewer fat crystals on the surface than larger ones, as small particles require less time to be totally solidified.

Different NLC morphologies are observed when using different microscopes, and these morphological discrepancies might be attributed to the type of lipid and surfactant or the method used to develop NLCs. Many studies suggest that incorporation of bioactive compounds into a solid lipid matrix can improve controlled release, charge, and physical and chemical stability by disrupting the crystal packing structure (21).

TAGs occur in three crystalline forms, α, β′ and β, which differ in the arrangement of fatty acids. X-ray diffraction analysis was performed to investigate the arrangement of fatty acid chains and identify the polymorphism of crystals by determining the dimensions of the crystal unit and sub-cells. Polymorphs diffract X-rays at different angles as a result of their different geometric configurations, and a diffraction at wide angles corresponds to short spacings (distances between parallel acyl groups in TAG) of sub-cells, which allows the characterization of different polymorphs in fats (46).

The short spacings and polymorphic forms of NLC_BO and NLC_BO24h are presented in Table 3. NLCs show a complex polymorphism as a result of β′ fusion and solid lipid recrystallization. This behaviour can be explained by the nanosize of NLC particles, which cause slower transition to a stable state, and may also be related to the presence of DAG above 5% in crude and interesterifiedburiti oil (12), favoring the formation of β′ polymorph. In addition, interesterification reduced TAG melting range, forming more unsaturated fatty acids and causing changes to crystal morphology and polymorphic forms. TAGs usually crystallize first in α and β′ forms even though β form is more stable. Factors such as cooling rate, crystallization heat, agitation level, and nanocarrier composition can affect the number and type of crystals. However, if fats are complex mixtures of TAGs, different polymorphic forms and liquid oil can coexist at a certain temperature (47).

Solid lipids are predominantly found in the stable β-modification. When incorporated into lipid nanoparticles, the lipid transforms into the less ordered, metastable β′-modification, causing a distortion in the crystalline structure (48). Previous study demonstrated that the incorporation of lipids in NLCs leads to slower polymorphic transition and lower polydispersity index (49). In addition, NLC crystallization, melting temperatures and polymorphic content have been shown to be strongly dependent on the amount of oil. Jenning et al. (50) prepared NLCs with glyceryl behenate (Compritol 888 ATO) as the solid lipid and caprylic/capric triglycerides (Miglyol 812) as the liquid lipid. NLC dispersions were most stable in the absence of oil and at high concentrations of oil. Other studies showed that the addition of oil to solid lipid nanoparticles did not disrupt the structure of lipid crystals, contrary to the expected (49, 51). Given the contradictory results, the effect of oil type on NLC polymorphic behaviour and morphology is still poorly understood (21).

The ORAC assay measures the ability of an antioxidant to sequester free radicals through the donation of hydrogen atoms. Pulido et al. (52) described the FRAP method as an alternative to determine the ferric reducing capacity of biological fluids and aqueous solutions of pure compounds. These two assays are of great importance because they indicate the antioxidant action in foods and physiological systems.

Table 4 shows the results of the ORAC and FRAP assays. In both assays, NLC_BO24h had the highest antioxidant activity. NLC_BO24h had approx. 23% higher ORAC value and 16% higher FRAP value than NLC_BO, demonstrating the influence of interesterification on the antioxidant activity of nanocarriers. Although there was a significant loss of tocopherols after 24 h of interesterification, the concentration of β-carotene was not affected and the concentrations of OOO-TAG (12.5%) and phenolic compounds increased, which could explain the increase in antioxidant activity (12). In addition, it can be considered that larger particles, such as NLC_BO, tend to generate imperfect, amorphous crystals and can expel bioactive compounds, which causes a lower antioxidant capacity of these samples (53).

Haeiwa et al. (54) reported that unsaturated fatty acids, particularly oleic acid, increase the intracellular levels of lipid peroxidation products, indicating that oleic acid can promote adaptive response and protect cells against oxidative stress-related injury. Buriti oil contains considerable amounts of carotenoids, tocopherols, and phenolic compounds, which give it an important antioxidant power, favouring its own conservation and contributing to the treatment of oxidative stress-related diseases (18).

In a previous study carried out by our research group, the antioxidant capacity was also higher of the structured lipids produced with Lipozyme TL-IM and enzyme from Rhizopus sp., respectively, followed by pure buriti oil. This result indicates that enzymatic interesterification significantly increased the antioxidant capacity of the oil, independently of the content of minor compounds (55). In a study by Poyato et al. (56), the antioxidant capacity expressed as Trolox equivalents, measured by the L-ORAC (lipophilic radical absorption capacity of oxygen) assay, of linseed oil emulsions (8315.4 μmol/100 g) was higher than those of NLCs developed in the present study, whereas for olive oil emulsions (1978.9 μmol/100 g) was similar to those of NLCs. These data suggest that the antioxidant capacity of emulsions differs according to oil type and antioxidant compound.

The antioxidant activity of bioactive compounds is associated with the inhibition of free-radical chain initiation by oxygen binding or chelation of catalytic metal ions to retard oxidation and peroxide decomposition, prevent continuous hydrogen abstraction, and protect DNA, proteins and lipids against oxidative damage (3). The results of this study suggest that nanocarriers prepared with interesterified buriti oil can be used in several applications for the improvement of health, including cosmetic formulations, food products and pharmaceuticals.