Potential of Annona muricata L. seed oil: phytochemical and nutritional characterization associated with non-toxicity

L.C. Pinto^a, C.O. de Souza^b, S.A. de Souza^c, H.B. da Silva^d, R.R. da Silva^h, A.T. Cerqueira-Lima^d, T.O. Teixeira^d, T.M.S. da Silva^c, K.C.P. Medeiros^e, M. Bittencourt^f, H.R. Brandão^f, J.I. Druzian^b, A.S. Conceição^g, M.V. Lopes^h and C.A. Figueiredo^d,*

^aDepartment of Food Science, School of Nutrition, Federal University of Bahia.

^bFaculty of Pharmacy, Federal University of Bahia, Brazil.

^cLaboratory of Phytochemistry Bioprospecting, Federal Rural University of Pernambuco, Brazil.

^dInstitute of Health Sciences, Federal University of Bahia.Av. Reitor Miguel Calmon, s/n, Canela, CEP: 41100-110 - Salvador, Bahia, Brasil.

^eFederal University of Rio Grande do Norte, Brazil.

^fVeterinary Hospital UNIME (HOSVET). Metropolitan Union of Education and Culture (UNIME), Lauro de Freitas, Bahia, Brazil.

^gUniversity of Bahia State, Department of Education, Herbarium of the State University of Bahia (HUNEB) Paulo Afonso, Brazil.

^hDepartment of Life Sciences, State University of Bahia, Brazil.

*Corresponding author: cavfigueiredo@gmail.com

1. INTRODUCTIONTOP

Annona muricata L., known as Soursop fruit, is native to America and has recently been aclimatized and established in many continents such as America, Africa and Asia (Nogueira et al., 2005). A. muricata has a history of popular use over the past few years and has been extensively described in the scientific literature as possessing biological properties such as analgesic and anti-inflammatory (Ishola et al., 2014), antidiabetic and antioxidant (Florence et al., 2014). Due to the presence of acetogenins, Annona has anti-tumor and apoptosis-inducing properties (Moghadamtousi et al., 2015).

The soursop fruit and other parts of the plant are considered to be underutilized. One fruit may contain more than 200 seeds, which means the disposal of a large amount of seeds. The main countries which produce soursop are Brazil, Venezuela and Colombia.b Bahia, Brazil is considered the world’s largest producer of soursop and produces about 8,000 tons of fruit per year. It is used in processing industries in the production of fruit juices and pulps (ADAB, 2015), which results in large amount of seeds that are discarded and non-utilized commercially. The unused structures of the plant, seeds for example, can be a source of herbal products, bio-compounds and pharmaceutical ingredients. However, information on the composition, nutritional value and medicinal uses of the A. muricata seeds is still limited. Anuragi et al., (2016) report that the graviola seed is rich in oil and can be exploited in the oil industy. Elagbar et al., (2016) identified important fatty acids in the oil of A. muricata seeds and emphasized that this oil can be exploited for industrial, cosmetic, and medicinal purposes.

The characterization of the functional quality of A. muricata seeds may favor their use in the human diet, especially as a functional food. Such products are commonly known to contain bioactive compounds with properties that prevent degenerative diseases and other morbidities. Typically, anti-inflammatory and antioxidant properties exerted by medicinal plants are attributed different fatty acids, phenolic compounds (Dimitrios, 2006), alkaloids or other bioactive compounds, which are present in plant extract and oil. However, natural compounds can have toxic effects and interfere in the vegetable’s consumption.

Given the above, the aim of this study was to evaluate the nutritional quality of A. muricata seeds and its oil through the chemical composition, quantification of phenolic compounds, antioxidant activity in vitro and toxicological evaluation in vivo.

2. MATERIALS AND METHODSTOP

2.1. Sample preparationTOP

The seeds were kindly provided in two batches (sample 1 and 2), by a of fruit pulp industry located in Ilhéus, Southern Bahia, Brazil. The seeds were identified and registered in the Herbarium of the State University of Bahia - Paulo Afonso HUNEB-Collection (n° 28720).

The seeds were separated for grinding manually, subjected to crushing, and stored at -20 °C until analysis of composition. Before crushing, the seeds used for oil extraction were previously lyophilized to facilitate crushing and better extraction of ethereal components,

2.2. Oil ExtractionTOP

Oil extraction was performed according to Blig-Dyer (1959) using methanol and chloroform as solvents. The decanted phase in the extraction process was collected and concentrated in a rota-evaporator equipment at 30 rpm under vacuum (695 mmHg) at 50-55 °C for 45 minutes, and was then subjected to a final drying by means of direct contact with nitrogen gas to enable the volatilization of residual solvent and the precipitation of the crude oil (AmPtO). The oil was centrifuged at 2300g/5min (4 °C) to separate precipitated lipids and for obtaining the supernatant oil (AmSO). The patent for this extraction process was deposited at the Intellectual Property Institute (INPI) (BR102016029177-1).

2.3. Centesimal compositionTOP

Centesimal composition was performed on fresh samples, in triplicate, and the determination of lipids, proteins, dietary fiber, moisture and ash was performed according to the analytical method of the Adolfo Lutz Institute (2008). The quantification of carbohydrates was performed by percentage differences in comparison to other components.

2.4. Quantification of phenolic compoundsTOP

The Folin and Ciocalteu assay was used for the quantification of total phenolic compounds using gallic acid as a standard in lyophilized seeds (Instituto Adolfo Lutz, 2008). Aqueous and methanolic extracts in the sample were analyzed in a concentration of 0.5 mg·mL⁻¹ in a spectrophotometer VIS/UV at 760 nm. A standard curve was plotted using gallic acid (0 to 8 mg·mL⁻¹). The results were expressed as milligram of tannic acid (ETA) and gallic acid (EGA) equivalents per gram of the dry weight sample.

2.5. In vitro antioxidant activityTOP

An AmSO extract was made by adding the oil (0.5g) to a methanol solution (50%), and after one hour of incubation at room temperature (28±1 ºC), acetonic solution (70%) was added, followed by distilled water to a final volume of 100 mL (Rufino et al., 2007). DPPH (2,2-diphenyl-1-picrylhydrazyl) free radical scavenging assay was made as per Roesler et al. (2007) to obtain the final result, which is expressed in IC₅₀ (g sample/g DPPH). The ABTS [2,2’-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid)] method was used with standard Trolox and UV/VIS at 734nm (Rufino et al., 2006). A ferric reducing antioxidant potential (FRAP) assay was made using the standard ferrous sulphate and spectrophotometer VIS/UV at 595nm (Rufino et al., 2006).

2.6. Fatty acid profileTOP

Fatty acids profile was determined by the capillary column gas chromatographic method applied to the oil methyl esters. The amount of total fatty acids (sum of free and bound fatty acids) in the oil was obtained by transesterification into the corresponding methyl esters (fatty acid methyl esters (FAME)), through saponification with NaOH (0.5N) in methanol, followed by methylation with a solution of boron trifluoride (BF₃) catalyst (12% in methanol). The FAME were extracted with iso-octane and stored in an inert atmosphere (N₂) in the freezer at −18 °C. The FAME separation was performed on a gas chromatograph (Varian^® 3800) equipped with a flame ionization detector and a fused silica capillary column Elite-WAX (30m×0.32 mm×0.25 mm). The analysis parameters were: injector temperature of 250 °C and detector temperature of 280 °C. The following thermal program was used: 150 °C for 16 min, then increased by 2 ºC·min⁻¹ up to 180 °C, held for 25 min, followed by an increase of 5 ºC·min⁻¹ up to 210 °C, held for 25 min. Helium was used as carrier gas at 1.3 mL·min⁻¹. Nitrogen gas was used as make up gas (30 mL·min⁻¹); the flow of hydrogen gas and air were provided at 30 and at 300 mL·min⁻¹, respectively. The injections were performed in duplicate for each extraction in a volume of 1 μL.

The FAME were identified by comparing retention times with known mixed standards (189–19-1AMP FAME Mix C4–C24; Sigma-Aldrich®). The quantification of fatty acids, expressed in milligrams per gram of lipids, was performed by adding internal standard (C23:0 Sigma®, USA) and calcuated using equation 1. Reported yields were averaged from duplicate extractions (Nascimento et al., 2003):

Where: A_x Area of methyl ester fatty acid peek in the chromatogram of the sample.

W_is Weight (in milligrams) of internal standard added to the sample.

CF_sConversion factor of fatty acid methyl ester to fatty acid.

A_is Area of internal standard methyl ester of fatty acid peek in the chromatogram of the sample.

W_s Sample weight (in milligrams).

CF_x Correction factor response of each fatty acid methyl ester ionization detector, relative to 23:0

2.7. Physiochemical screening of crude oilTOP

The AmPtO sample was separated as precipitate (solid residue in low temperature) and AmSO (supernatant oil). AmPtO (50 mg) was dissolved in methanol and acetonitrile (1:1) and extracted in solid phase extraction (SPE). A Sep-Pak silica cartridge (Waters) was sequentially conditioned with 5 mL of hexane and 5 mL dichloromethane to prevent the cartridge from drying. AmPtO was passed through the cartridge and eluted with 10 mL of dichloromethane, ethyl acetate and methanol. The eluates were dried under reduced pressure in a rotatory evaporator at 40 °C to yield 14.0 mg dichloromethane, 13.5 mg ethyl acetate and 8.0 mg of methanol fractions. After evaporating to dryness by rotary evaporator, the residues were dissolved in methanol and acetonitrile, filtered through a 0.45-µm nylon syringe filter (Whatman) and injected into the HPLC system. The standard stock and working solutions were stored at 4 °C.

All reagents used were of analytical grade. HPLC-grade methanol and acetonitrile were purchased from Tedia and Merck, respectively. The deionized water was purified using a Milli-Q system (Millipore, Milford, MA, USA).

The thin layer chromatographic profiles were made with AmSO, AmPtO and their fractions obtained by SPE extraction according to Grzybowski et al., (2012). Chromatoplates were made with hexane: chloroform: nitroethane: ethyl acetate: acetone: methanol: acetonitrile: water (12:2:4:4:1:2:1.6:0.1, v/v) as mobile phase. The spray for acetogenin constituents was specifically Kedde´s reagent (alkalinized 3,5-dinitrobenzoic acid).

HPLC was performed in a Shimadzu Prominence LC-20AT equipped with a (SPD-M20) diode array detector (HPLC-DAD) (Shimadzu Corp. Kyoto, Japan). The samples were injected into a Rheodyne 7125i injector with a 50 μL loop. The column heater was set at 30 °C. The chromatographic separation was performed with a Luna Phenomenex C-18 column (250 mm x 4.6 mm x 5 μm, Phenomenex). The compounds were separated using a mobile phase consisting of water (A) and acetonitrile:methanol (1:1, B). The separation gradient was: 0-20 min 70-100% B, 20-40 min 100%, 45 min stop at a flow rate of 1.0 mL·min⁻¹. A 50 μL sample was injected and the detection of compounds was performed using light with a wavelength of 215 nm. The infrared absorption spectra were recorded in KBr pellets using a Varian 640 FT-IR spectrophotometer with a PIKE ATR accessory operating in the 4000-400 cm⁻¹ range. Liquid chromatograph Electrospray ion source mass spectrometer (LC-ESI-MS) was obtained in negative electrospray mode using an Esquire 3000 Plus instrument (Bruker). Thin-layer chromatography plates were run using 60 F₂₅₄ silica gel (Macherey-Nagel).

2.8. In vivo toxicity evaluationTOP

2.8.1. Animals and ethical considerationsTOP

BALB/C mice of 8-10 weeks old were used in this study. The animals were housed at 22 °C and received water and food ad libitum until the time of the experiment. Animal procedures were performed in accordance with the recommendations of the Ethics Committee of the Institute of Health- Federal University of Bahia Sciences (CEUA/ICS 029/2012).

2.8.2. Experimental protocolsTOP

The experimental protocol aimed to verify the toxicity of repeated doses (ANVISA, 2010) of 14 days-exposure daily to AmPtO and AmSO doses was 0.5 mL·Kg⁻¹ animal weight and 1.0 mL·Kg⁻¹ animal weight, respectively. The oil samples were diluted in saline and emulsified using DMSO (dimethylsulfoxide, 0.5%, Sigma). The animals were divided into three groups, each group with 6 animals. On day 14 (D14), the animals were euthanized (150 mg·Kg⁻¹) thiopental associated with 10 mg·Kg⁻¹ lidocaine) and histopathological and biochemical samples were taken for the analysis of toxicity.

AmPtO protocol groups were defined as follows: the CtrlAmPtO group (control): received saline orally (200 µL of saline + DMSO 0.5%); the AmPtO0.5 group received crude oil orally at 0.5 mL·kg⁻¹ (oil diluted in 200 μL of saline + DMSO 0.5%); the AmPtO1.0 group received crude oil orally at 1.0 mL/kg (oil diluted in 200 μL of saline + DMSO 0.5%).

The AmSO protocol was followed with the same oil doses (0.5 mL·kg⁻¹ and 1.0 mL·kg⁻¹) and the groups were defined as CtrlAmSO group; AmSO0.5 group; AmSO1.0 group. AmSO was dissolved in saline followed by filtration (0.2 μm). The filtration process allowed separating the precipitate in the crude oil of the supernatant liquid fraction (AmSO). The waste precipitates were retained in the filter, thus, a quantity of 40% of AmSO weight was added to correct the average percentage weight loss during filtration.

2.8.3. Histopathological, physical, biochemical and haematological analysisTOP

The animals were weighed on D1 and D14 of the protocol. On D14, before euthanasia, three independent observers trained in animal experiments evaluated the overall appearance of the animals, according to the score of toxicity signs: (0) normal appearance and coat; (1) discrete change in coat and discrete cachexia; (2) coat change and cachexia (Lerco et al., 2003). Liver and spleen were collected for weight evaluation. The ratio of liver weight/body weight of the animal was calculated to identify liver relative weight (Ritter et al., 2012).

Liver and kidney samples were paraffin embedded twice after fixing with 10% formalin using a microtome; the paraffin blocks were cut into serial 3 μM sections. In the staining step, the slides were immersed in hematoxylin-eosin (HE). Histopathological changes were evaluated by two different pathologists, blindly.

Blood glucose levels, total cholesterol, urea and creatinine were performed in semi-automatic Bioplus 2000 equipment using commercial kits (Doles) according to the manufacturer. The quantitation of leukocytes was performed by manual counting, with dilution in Turk liquid (200μL liquid Turk + 20 μL of blood with Ethylenedeamine Tetra Acetic Acid (EDTA)).

2.9. Statistical analysisTOP

One-way analysis (ANOVA) and Tukey Test correction to post-test (for data with normal distribution) and Kruskal Wallis and Dunn as post-test (for data without normal distribution) were used to determine statistical significance among groups using GraphPad V5 (GraphPad Software Inc., San Diego, CA, USA) software. The results were considered significant when p < 0.05.

3. RESULTSTOP

3.1. Centesimal composition, phenolic compounds and in vitro antioxidant activityTOP

The results of centesimal composition (wt %) of A. muricata seeds in natura were: moisture 34.7±0.02; ash 1.3±0.01; proteins 9.85±0.31; lipids 18.3±0.28; fibers 24.7±0.52 and carbohydrates 11.15. The polyphenols of aqueous extracts in the A. muricata lyophilized seeds showed a variation of 310-320mg%, which is 3.2±0.02 mg TAE/g and 3.1 ±0.02 mg GAE/g (tannic acid and gallic acid equivalents, respectively). Methanol extracts showed turbidity, making them impossible to read in the spectrophotometer.

The In vitro antioxidant capacity, AmSO showed 370133.0 μg·mL⁻¹ DPPH (DPPH assay); 40.2 μmol trolox/g (ABTS assay) and 50.7 μmol Fe₂SO₄·g⁻¹ (FRAP assay). Regarding the DPPH assay, the lowest value of the sample, able to reduce DPPH by 50% (IC₅₀), had better antioxidant activity. The assessment of the ABTS and FRAP methods in a better antioxidant activity was based on the higher amount of equivalent in trolox or ferrous sulfate per gram of sample, respectively. Overall, in comparison with standard gallic acid (1.30 μg·mL⁻¹ DPPH; 18.917.0 μmol trolox/g; 41946.7 μmol Fe₂SO₄·g⁻¹), AmSO had low antioxidant activity.

3.2. Fatty acid profileTOP

A. muricata seed oil had a total of 75% unsaturated fatty acids (Table 1). The fatty acid profile of the two oil samples analyzed, each in duplicate, were similar. Considering the percentage of unsaturated fatty acids, 40% was monounsaturated, represented by 39.2% oleic acid (n-9). The second highest detected percentage (33%) of fatty acid was linoleic (n-6). The oil contained a low content of α-linolenic acid (n-3). Regarding the saturated content (24.5%), the most representative fatty acid was palmitic (19.3%).

3.3. Phytochemical screeningTOP

Figure 1-A shows the chromatographic profile in chromatoplate of the oil, precipitate and the fractions obtained by SPE extraction. This clearly revealed predominant rose-wine coloration spots (positive Kedde’s reaction) on the acetogenin family members (Figure 1-A). The absorption in the UV spectrum at 215 nm was observed in HPLC-DAD analyses (Figure 1-B and 1-C). The presence of acetogenins was confirmed by the LC-ESI-MS analysis, which detected several peaks with molecular weights (m/z values), which are typical of these compounds. In all the samples, the most abundant peaks corresponded to the molecular weight corresponding to annonacin (C₃₅H₆₄O₇; mol. wt. 595 [M-H]^-). Fractionation of the extract (precipitate residue) by SPE allowed the observation of the major acetogenin (Figure 1-B and 1-C), which was confirmed to be annonacin (m/z 595, [M-H]^-, mainly in the methanol fraction (Figure 1-D).

3.4. In vivo toxicityTOP

3.4.1. SurvivalTOP

In the AmPtO protocol, the AmPtO1.0 group showed an important reduction in survival in the early days of the Protocol and reached a percentage lower than 50%. In the AmSO protocol, all the animals showed 100% survival.

3.4.2. Physical, biochemical and haematological analysisTOP

Table 2 shows the biochemical and physical parameters of in vivo experiments. The AmPtO1.0 group showed a statistical difference in blood glucose level compared to the CtrlAmPtO group and AmPtO0.5 group (p < 0.001 and p < 0.01, respectively). The AmPtO1.0 group also showed weight loss (variable average -1.3g) compared to the other experimental groups. The AmPtO0.5 group was characterized by moderate signs of toxicity (score 1) and the AmPtO1.0 group showed more intense toxicity signs (score 2). In the AmSO protocol, liver weight was inversely proportional to oil doses (p < 0.01 and p < 0.001, respectively for AmSO0.5 and AmSO1.0) in comparison to the control group (CtrlAmSO) and was evidenced by liver relative weight (ratio of 0.027) in relation to body weight.

3.4.3. Histopathological analysisTOP

Figures 2 (A, B, C) and 3 (A, B, C) show the impact of toxicity in the liver tissue. In the AmPtO protocol, the CtrlAmPtO group showed preserved morphology with isomorphic hepatocytes arranged in cords, directed to the centrilobular hepatic veins and hepatic sinusoid, with thin walls with the frequent presence of Kupffer´s cell. Portal spaces were regularly distributed and the typical polygonal morphology of hepatocytes and evident cell nucleus were observed. The hepatocytes of the AmPtO0.5 group and AmPtO1.0 group showed focal and minimal histological changes. The AmPtO0.5 and AmPtO1.0 groups showed discrete hepatic steatosis (Figure 2-B and 2-C, black circle). The liver section of the AmPtO0.5 group showed some hemorrhagic areas with necrosis (Figure 2-B, red arrow). Histopathological changes were not identified in the case of the AmSO protocol.

Figures 2 (D, E, F) and 3 (D, E, F) show the kidney histopathological analysis. The CtrlAmPtO group showed normal characteristics of the renal parenchyma. Magnification at 100x revealed an eosinophilic cortex with the abundance of proximal convoluted tubules along with well distributed and uniform renal corpuscles. Magnification at 400x allowed for the observation of glomeruli with Bowman´s capsule and capillary tube supported by delicate mesangium, discarding features of glomerulonephritis (Figure 2-D). The abundant proximal convoluted tubules had preserved cubic epithelium and strongly eosinophilic cytoplasm which differed from the distal convoluted tubules. The AmPtO0.5 group had similar characteristics, however, the glomeruli showed discrete atrophy and restricted Bowman´s space (Figure 2-E, white arrow), and the presence of capillary hyperemia in the medullary area between the collecting tubules. The AmPtO1.0 group had some tubules with more eosinophilic cytoplasm, the presence of precipitates, more basophilic cell nucleus associated with karyolysis and coagulative necrosis (Figure 2-F, asterisk). In the AmSO protocol (Figure 3) similar morphological characteristics in the cortical area, glomeruli with proportional size, the presence of a capsule and Bowmn´s space, simple cubic and upright tubular epithelium were observed. In the medullary zone, the tubules’ normal collectors between the various blood capillaries, with low columnar cubic epithelium with visible lumen were also observed.

4. DISCUSSIONTOP

A. muricata L. seeds have potential for use in human and animal diets, provided they are not toxic. Fasakin et al., (2008) also evaluated the composition of the seeds of soursop fruit from Nigeria, and identified a high content of oil and protein, which were low in toxins (tannins, phytate, and cyanide). These authors identified 22.57% lipids in the seed, and a total of 27.34% protein, which is higher than in the current study. These variations may due to the climatic and edaphic conditions in which the plants were grown (Marineli et al., 2004). Therefore, it is important to characterize the fruits in different regions and countries. Elagbar et al., (2016) identified 21.5% of lipids in A. muricata seeds’ crude fixed oil on a dry weight basis. In the crude oil of the present study, a total of 18.3% lipids were identified in the in natura samples. Considering the presence of moisture (34.7%), it is probable that our seed sample had a higher yield of lipids and greater potential for oil extraction. Santos et al., (2014) evaluated total phenolic compounds in four varieties of Annona cherimola, of Portuguese origin, and found values lower than 0.05 mg·g^₋₁ of seed sample. In the current study, A. muricata seeds had a total of approximately 0.31 mg·g⁻¹ of phenolics, hence, they can be considered a moderate source of phenolic compounds.

The Seed oil showed low antioxidant activity in vitro. Vegetable oils have a considerable content in tocopherols, which are potential antioxidants (Guinazi et al., 2009). It is probable that A. muricata seeds oil can naturally have low antioxidant compounds or such compounds may be lost in the oil extraction process.

Despite the low antioxidant activity identified in the seeds, leaf extracts of the A. muricata are well characterized in the literature as having antioxidant properties. Lee et al., (2014) evaluated the antioxidant activity and components in dry extracts of the A. muricata leaves and reported DPPH IC₅₀ at 98.9±9.1 mg/mg extract, and the total content of phenolics to be 86.5±14.8 mg GAE/g extract. Leaf extracts of A. muricata were identified as a rich source of antioxidant components such as carotenoids.

The high lipid content also enhances the use of A. muricata seeds or the production of pharmaceutical and industrial oils. The most representative fatty acids in the oil analyzed were oleic and linoleic, which are well characterized and imply benefits for cardiovascular health and a reduced risk of metabolic syndrome (Mayneris-Perxachs et al., 2014). Elagbar et al. (2016) identified fourteen fatty acids in the fixed oil of A. muricata seeds extracted using hexane or diethylether and found a total of approximately 23% saturated, 40% monounsaturated and 36% polyunsaturated. In this study, oleic (39%) and linoleic (35%) acids were the main fatty acids identified. The results were similar to our study, despite the different methods of oil extraction.

Some studies suggest that in addition to the benefits of the fatty acids present in vegetable oils, the presence of minor components which are contained in the natural source of the oil act synergistically, favoring the inherent functional properties, for example in Chilean chia oil (Marineli et al., 2014). Some alkaloids have been identified in A. muricata such as anonina, muricina and muricinina (Franzão and Melo, 2007) and acetogenins (Moghadamtousi et al., 2015). Acetogenins are derived from long chain fatty acids (C35 and C39) and are found specifically in the Annonaceae family and have antitumor activity and apoptosis-inducing activity (Moghadamtousi et al., 2015). Thus, it is important to identify the bioactive compounds in vegetable oils which have functional and pharmacological properties.

Alkaloids were not found in AmSO or AmPtO (data not shown). The principal spot was present in all the samples analyzed in the present study, mainly in the methanol fraction, and may be associated to the presence of annonacin, considering that acetogenin is the major compound in A. muricata seeds (Moghadamtousi et al., 2015; Grzybowski et al., 2012; Champy et al., 2002). The strong absorption at 1745 cm⁻¹ (lactone C=O), a strong aliphatic C-H absorption below 3000 cm⁻¹ in the IR spectrum suggested the presence of an unsaturated lactone moiety in an annonacin-type acetogenin (Champy et al., 2002). Le Ven et al., (2012) identified anonacin as the main acetogenin in extracts of the manufactured A. muricata nectar. In this study, the authors identified nine compounds of Fragmentation of m/z 603.4797 ([C35H64O7Li] +), which corresponds to the isomers of annonacin.

Champy et al., (2002) analyzed annonacin in extracts obtained from A. muricata. In all preparations, fruit pulp, nectar, infusions and decoctions of leaves, the quantification showed that annonacin represented approximately 70% of all acetogenins. The analysis in the present study showed that A. muricata seed extracts and oil contained acetogenins and annonacin was the major compound. Annonacin was characterized by its neurotoxic properties and as an inhibitor of mitochondrial complex I (Yu et al., 1992; De Pedro et al., 2013). However, acetogenins can be related to the etiology of neurodegenerative diseases, such as Guadeloupean atypical Parkinsonism and perturbations in the protein Tau due to depletion ATP supply and interrupt the transport of mitochondria to the sum of neural cells (Yu et al., 1992). The pharmacokinetic mechanisms of annonacin are still to be known.

The mechanical filtration process enabled a reduction in the presence of acetogenin, as shown in Figure 1-A. Other methods may be implemented to reduce toxic components, and thereby facilitate the use of A. muricata seed oil in the human diet.

Acute exposure to different doses of seed oil (AmPtO and AmSO) was performed for assessing toxicological effects. The process of oil filtration was important to differentiate the protocols, with crude oil (with precipitate) and liquid oil (oil supernatant). The survival rate of less than 50% in the AmPtO1.0 group was high and implies considerable toxic effects of crude oil. In AmSO protocols, the AmSO0.5 and AmSO1.0 groups received filtered oil and remained alive throughout the experiment, suggesting that the mortality identified in the AmPtO1.0 group was linked to the precipitated consumption in the oil, which was depleted after filtration.

Some biochemical and physical parameters evaluated confirmed the toxic effect of the crude oil (AmPtO). The AmPtO1.0 group showed increased blood glucose, a negative weight variable, toxicity signs and histopathological changes in hepatic and renal tissue (Table 2 and Figures 2 and 3). Animal groups that received AmSO showed no toxicity signs or histopathological changes, which indicates that the precipitate in AmPtO may be related to the toxicity observed in this group. This can probably be associated to the presence of acetogenins, more concentrated in the crude oil and its fractions (Figure 2).

Acetogenins, especially annonacin, act as mitochondrial complex I inhibitors and are implicated in neurological disturbances, arising from the decreased ATP available to neural cells (Yu et al., 1992; De Pedro et al., 2013). The mitochondrial respiratory chain can be separated into 4 enzyme complexes, with complex I representing the NADH-ubiquinone reductase. This enzyme plays a crucial role in the generation of ATP and its role in the performance of neurodegenerative diseases has been considered (Schapira, 2010). The hepatotoxic and nephrotoxic effects in the AmPtO consumption identified in this study may also be associated with the inhibitory action of the mitochondrial enzyme complex in these specific organs.

AmSO showed nutritional quality and no toxicity. However, a reduction in the liver weight of the AmSO0.5 and AmSO1.0 groups compared to CtrlAmSO group was observed (Table 2). The hepatoprotective effect has been identified in coconut oil, the study of Zakaria et al., (2011). Treatment with virgin coconut oil reduced the hepatotoxicity induced by acetaminophen, reduced the dosages of liver enzymes and improved liver morphology. A reduction in the levels of creatinine also was observed in the AmSO1.0 group in compared to CtrlAmSO, but without statistical significance. Hepatoprotective effects and renal metabolism can be exploited in other in vivo studies on the pharmacological potential of AmSO.

Furthermore, the identification of acetogenins in the oil suggests that this product may have pharmacological potential since these compounds mainly have anticancer activity. Annona muricata possess a wide spectrum of biological activities, and the use of their by-products or derived products, such as oil, as pharmaceutical or food ingredients is promising.

5. CONCLUSIONSTOP

A. muricata seed oil (liquid fraction) (AmSO) has potential for use in food, especially due to its contents in oleic and linoleic fatty acids. AmSO showed no relevant antioxidant activity and moderate phenolic compounds.

The oil precipitate (AmPtO) showed a great amount of acetogenins and was considered toxic according to the parameters evaluated in the experimental protocol in vivo. The acetogenin identified in the oil was anonacin. The toxic property of the precipitate can be associated with the presence of the acetogenin. The liquid fraction of oil (AmSO) showed no toxic properties.

Further studies on toxic action mechanisms, as well as studies concerning biological activities and pharmacological properties can be conducted to better characterize the seed oil of soursop and encourage its use in medicine.

ACKNOWLEDGMENTSTOP

The authors thank the Coordination of Improvement of Higher Level Personnel (CAPES), National Research Council (CNPq) and Foundation of Science and Technology Support of the State of Pernambuco (FACEPE).

REFERENCESTOP