Evaluation of bioactive compounds and antioxidant capacity of edible feijoa (Acca sellowiana (O. Berg) Burret) flower extracts

Abstract

Abstract

The research aimed to evaluate the polyphenolic composition and the antioxidant capacity of edible extracts of feijoa (Acca sellowiana (O. Berg) Burret) flowers. Phenolic compounds of whole feijoa flower (FM), feijoa petals (PM) and feijoa petals juice (PJ) were identified by high-performance liquid chromatography coupled with electrospray mass spectrometry and quantified by liquid chromatography coupled with ultraviolet/visible detection. Moreover, the total polyphenol (TP) content was measured spectrophotometrically and the antioxidant capacities of the extracts were evaluated by FRAP, CUPRAC, DPPH^·, and ABTS^·+ assays. The FM showed TP content (395.14 ± 7.91 mg GAE/L) higher than PM and PJ, and exhibited better antioxidant capacities. FM extracts were characterized by the high content of anthocyanins (115.3 ± 3.6 mg/L), flavonols (42.9 ± 3.3 mg/L) and the presence of ellagic acid (7.9 ± 0.2 mg/L) and other galloyltannins. In addition, cyanidin-3-O-glucoside and apigenin were detected in all the three extracts. The present study provided an overview on particular bioactive compounds that characterise different parts of edible feijoa flowers. Among the latter, FM proved to be the most suitable for exploitation in the food and health manufactory.

Graphic abstract

Introduction

There is a rising consumer demand for healthy natural products, from herbal drugs to single active pharmaceutical ingredients, and new vegetable sources are highly appreciated (Oga et al. 2016). Flowers represent a valuable source of compounds which affect target functions of the organism in a positive way (Fernandes et al. 2017) and could be considered as a new frontier in vegetable sustainable exploitation to obtain products with nutraceutical interest. Currently, some flowers are used either unprocessed or processed in different formulations for the health and food industry (Ammar et al. 2015; Fernandes et al. 2017). Feijoa (Acca sellowiana (O. Berg) Burret) is a plant which belongs to the Myrtaceae family and is typical from Brazil, but it also widely grows in parts of the USA, in the Mediterranean area, in New Zealand and in Australia (Weston 2010). It is well known for its edible fruits, which have a smooth, green skin and soft, white, sweet flesh with a very aromatic flavour (Weston 2010). In addition to the fruits, the petals of the flowers are being eaten, usually in salads, sweets and as dish decorations (De Souza et al. 2016), since they possess a pleasant taste and intense colour. These flowers are formed by four to six fleshy petals, rose-coloured inside and white outside, with red stamens reaching up to 2 cm above the flower, and slightly thickened stigma (Belous et al. 2014).

Considering the massive flower production of the feijoa plant, which fits well with a possible exploitation by food supplements, confectionery industry and pharmaceutical manufactory, the aim of this research was to evaluate the potential nutritional interest of edible feijoa flowers, evaluating their antioxidant capacity and the presence of bioactive compounds. To date, only few data on these compounds of feijoa flower buds and their health properties have been published (Aoyama et al. 2018). In that study, acetone extract showed the presence of pedunculagin, gallic acid, cyanidin glucoside, flavone, ellagic acid and gossypetin arabinofuranoside. In this new study, the content in phenolic compounds with potential biological activities were intensively evaluated in different edible extracts of feijoa flowers (whole macerate flowers, petal macerate and petal juice), by assessing a quali-quantitative analysis by high resolution mass spectrometry ((HR) LC-ESI-Orbitrap-MS and (HR) LC-ESI-Orbitrap-MS/MS) and HPLC–DAD. In addition, the total polyphenol (TP) and reducing sugars content was measured, and the antioxidant and antiradical capacities of the extracts were evaluated by FRAP, CUPRAC, DPPH^·, and ABTS^·+ assays.

Materials and methods

Chemicals

All the chemicals used were of analytical grade. Standard of phenolic compounds were purchased from Extrasynthese (Genay Cedex, France). Methanol, acetonitrile, ferrous sulphate, phosphoric acid 85% w/w, sodium carbonate, ammonium acetate, ferric chloride, CuSO₄·5H₂O, CuCl₂·2H₂O, potassium persulphate, 1,1-diphenyl-2-picrylhydrazyl radical (DPPH^·), (±)-6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox), 2,4,6-tris(2-pyridyl)-1,3,5-triazine (TPTZ), 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonate radical cation (ABTS^·+), neocuproine (2,9-dimethyl-1,10-phenanthroline) hydrochloride, Folin–Ciocalteu’s phenol reagent and gallic acid were obtained from Sigma-Aldrich (Milan, Italy). Acetonitrile, water and formic acid of LC–MS grade were purchased from Merck (Darmstadt, Germany). Food grade ethanol (95% v/v) was purchased from Bellini srl (Paratico, BS, Italy). Ultrapure water (18 MΩ cm) was obtained with a Milli-Q Advantage A10 System apparatus (Millipore, Milan, Italy).

Feijoa flower samples

Acca sellowiana (O.Berg) Burret mature flowers (3.0 kg in triplicate) from cultivated growing plants were randomly harvested in May 2017 in an experimental field in Cagliari (Sardinia, Italy; 39°13′49.0″N 9°08′03.1″E) by professional pickers. The specimens were identified by Prof. Gianluigi Bacchetta (University of Cagliari, Italy) and voucher number DISVA.ALI.01.2017 was deposited at the Department of Life and Environmental Sciences of the University of Cagliari (Italy). After the collection, the flowers were cleaned and separated to obtain whole feijoa flowers (FM), feijoa petals (PM) and feijoa petals juice (PJ) extracts. Feijoa petals were manually separated by cutting the petal at the junction with the receptacle. FM and PM were obtained by adding an EtOH:H₂O mixture (80:20 v/v) to the plant material (solvent:plant ratio 9:1 v/w). The macerates were extracted using ultrasonic bath (3 cycles of 30 min, t = 10 ± 2 °C) and the liquid phase was separated from the plant material. The extracting procedure was performed once again by adding fresh solvent and the two-liquid phase were reunited. PJ extracts were obtained squeezing the feijoa petals using a Polsinelli Enologia mod. ALU2530 stainless steel manual press (Isola del Liri, FR, Italy). The FM, PM and PJ extracts were freeze-dried and stored at − 20 °C in dark glass vials.

Determination of total polyphenol content (Folin–Ciocalteu’s assay), total reducing power (FRAP and CUPRAC assays) and free radical scavenging activity (DPPH^· and ABTS^·+ assays)

All assays were performed spectrophotometrically on the samples diluted with H₂O and spectrophotometric readings were carried out with a Cary 50 spectrophotometer (Varian, Leinì, TO, Italy) using 10 mm Kartell® plastic cuvette. TP content was determined with a modified Folin–Ciocalteu’s method (Tuberoso et al. 2013) by adding 100 μL of sample to 0.5 mL of Folin–Ciocalteu’s phenol reagent. After 5 min, 3 mL of 10% Na₂CO₃ (w/v) was added, the mixture was shaken and then diluted with water to a final volume of 10 mL. After a 90 min incubation period at room temperature, the absorbance was read at 725 nm against a blank. The total polyphenols content results, expressed as mg/L of gallic acid equivalent (GAE), were obtained using a calibration curve of a freshly prepared gallic acid standard solution (10–200 mg/L). The FRAP assay was assessed preparing a ferric complex of 2,4,6-tris(pyridin-2-yl)-1,3,5-triazine (TPTZ) and Fe³⁺ (0.3123 g TPTZ, 0.5406 g FeCl₃·6H₂O in 100 mL acetate buffer pH 3.6) according to Tuberoso et al. (2013). Twenty μL of sample were dissolved in 2 mL of ferric complex and, after an incubation period of 4 min in the dark, absorbance at 593 nm was measured. CUPRAC assay was performed according to Bektaşǒglu et al. (2006) with some modifications. Briefly, 100 μL of sample were dissolved in a mixture of 500 μL of 10 mM CuCl₂ solution in water, 500 μL of 7.5 mM neocuproine solution in methanol, and 500 μL of 1.0 M CH₃COONH₄ buffer at pH = 7.0. After an incubation period of 30 min in the dark, absorbance at 450 nm was measured. For both FRAP and CUPRAC assays results were expressed as mmol/L of Fe²⁺ because quantitative analysis was performed according to the external standard method using 0.1–2 mmol/L FeSO₄.

The DPPH^· assay is based on the ability of the antioxidant to scavenge the radical cation 1,1-diphenyl-2-picrylhydrazyl radical (Tuberoso et al. 2013). Fifty microliters of sample were dissolved in 2 mL of 0.06 mmol/L DPPH^· in methanol. Then, spectrophotometric readings were carried out at 517 nm after an incubation period of 60 min in the dark.

The ABTS^·+ assay was performed according to Re et al. (1999) with some modifications. The ABTS^·+ cation radical was produced by the reaction between 10 mL of 2 mM ABTS in H₂O and 100 μL of 70 mM potassium persulfate, stored in the dark at room temperature for 24 h. The ABTS^·+ solution was then diluted with methanol to obtain an absorbance of 0.70 ± 0.02 at λ = 734 nm, and was equilibrated at 30 °C. Samples were prepared in triplicate by diluting 20 μL of extracts (same dilution of DPPH^·+ assay) in 2 mL of the ABTS^·+ solution diluted with methanol. After 1 min of reaction, absorbances were recorded at 734 nm. DPPH^· and ABTS^·+ data were reported as Trolox equivalent antioxidant capacity (TEAC, mmol/L) because quantification was carried out by using a Trolox calibration curve in the range of 0.05–1.0 mmol/L.

LC–MS/MS and HPLC–DAD analyses

The electrospray ionisation (ESI) source of a Thermo Scientific LTQ-Orbitrap XL mass spectrometer (Thermo Scientific, Dreieich, Germany) was tuned in negative ion mode with a standard solution of kaempferol-3-O-glucoside (1 µg/mL) introduced at a flow rate of 5 μL/min by a syringe pump. Calibration of the Orbitrap analyser was performed using the standard LTQ calibration mixture composed by caffeine and the peptide MRFA (calibration solution purchased from the manufacturer), dissolved in 50:50% v/v water/acetonitrile solution. Resolution for the Orbitrap mass analyser was set at 30,000. The mass spectrometric spectra were acquired by full range acquisition covering m/z 180–1600 in LC–MS. The data recorded were processed with Xcalibur 2.0 software (Thermo Fisher Scientific).

LC-ESI-LIT Orbitrap MS was performed using a Finnigan Surveyor HPLC (Thermo Fisher, San Jose, CA, USA) coupled to a hybrid Linear Ion Trap (IT) Orbitrap mass spectrometer (Thermo Scientific). Analyses were performed using a Kinetex Evo C18 (150 mm × 2.1 mm particle size 5 µm) column, eluted with water containing 0.1% formic acid (solvent A) and acetonitrile containing 0.1% formic acid (solvent B). A linear gradient program at a flow rate of 0.200 mL/min was used: 0 to 20 min, from 5 to 15% (B); 20 to 40 min, from 15 to 35% (B); 40 to 50 min from 35 to 100% (B) then to 100% (B) for 5 min and back to 5% (B) for other 5 min. 10 µL of a solution 1 mg/mL of feijoa extracts in water was injected. The ESI source and MS parameters were the same used by D’Urso et al. (2017).

Detection and quantitative analysis of phenolic compounds were carried out using an HPLC–DAD method as described by Tuberoso et al. (2017). The extracts diluted with ultrapure water were filtered through Econofilter RC membrane (0.45 μm, Ø 25 mm, Agilent Technologies, Milan, Italy) and injected into the HPLC column without any further purification. The calibration curves were built with the method of external standard, correlating the area of the peaks versus the concentration. Standards of ellagic acid, quercetin-3-O-galactoside, quercetin-3-O-glucoside, quercetin, kaempferol, apigenin, delphinidin-3-O-glucoside, and cyanidin-3-O-glucoside were used. The full validation procedure, in agreement with the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use guidance note (ICH 2005), is reported by Tuberoso et al. (2017).

Determination of dry residue and reducing sugars

The dry matter of feijoa flowers extracts was evaluated by drying 1000 µL of solution for 2 h in a thermostatic oven at 105 ± 1 °C and weighing until a constant weight value was attained. Glucose and fructose contents were determined using the ENZYTEC™ d-Glucose/d-Fructose/Sucrose E1247 assay kit according to the producer procedure (R-Biopharm AG, Darmstadt, Germany).

Statistical analysis

The GraphPad Prism 5.0 software (GraphPad software, San Diego, CA, USA) was used to calculate the means and standard deviations of three independent experiments, involving triplicate analysis for each sample. Evaluation of statistical significance of differences was performed by one-way analysis of variance (One-way ANOVA). The statistical analysis was done using 2-way analysis of variance, followed by Bonferroni post hoc test. All data are expressed as means ± standard error of the means. p < 0.05 was considered statistically significant.

Results and discussion

LC-ESI-Orbitrap MS phenolic identification in FM, PM and PJ feijoa products

Edible feijoa flowers extracts were intensively investigated for the first time, and the profile is substantially different from feijoa fruits polyphenolic extracts and feijoa leaves polyphenolic extracts, which were more investigated (El-Shenawy et al. 2008; Ruberto and Tringali, 2004). (HR) LC-ESI-Orbitrap MS and (HR) LC-ESI-Orbitrap-MS/MS analysis of extracts from different parts of feijoa flowers were conducted in negative and positive ion mode. Individual components were identified by comparison of their m/z values extracted in the Total Ion Courrent (TIC) profile with those of the selected compounds described in literature characterised by Orbitrap MS/MS (Table 1) for feijoa fruits, feijoa leaves and other plant species. LC-ESI-Orbitrap MS/MS experiments were run by using Dependent Data Scan, in order to submit the major ions in TIC profiles to fragmentation experiments using the source and trap parameters previously selected by ESI–MS and ESI–MS/MS direct introduction experiments. Table 1 reports identification of compounds, based on High Resolution Mass Spectrometric Data, chemical formula derived by accurate mass mensuration, retention times, MS/MS results and references used for identification. High resolution mass values are all not differing more than 5 ppm respect to the exact mass calculated for the same molecule.

Table 1.

Compounds identification by LC-ESI-Orbitrap-MS in feijoa flowers macerate (FM), feijoa petals macerate (PM), and feijoa petals juice (PJ)

	RT		Molecular formula	ppm	MSMS	Identity	References	FM	PM	PJ
[M–H]−
1	5.19	783.0662	C34H23O22	− 0.59	764/481/301	Pedunculagin	Aoyama et al. (2018), D’Urso et al. (2017), de Lira Teixeira et al. (2015)	x	x	-
2	7.90	933.0621	C41H25O26	− 0.72	–	Castalagin	El-Shenawy et al. (2008)	x	-	-
3	8.77	783.0662	C34H23O22	− 0.59	764/481/301	Pedunculagin isomer	Aoyama et al. (2018), D’Urso et al. (2017), de Lira Teixeira et al. (2015)	x	x	-
4	12.61	633.0720	C27H22O18	− 0.22	513/469/330	Strictinin	Fukuda et al. (2003)	x	-	-
5	16.38	785.0831	C34H25O22	− 0.13	683/433/301	Tellimagrandin I	Baxter et al. (1999)	x	-	-
6	16.71	483.1012	C20H19O11	1.10	301/257	Nilocitin	Mahmoud et al. (2001)	x	-	-
7	20.44	321.1553	C14H25O8	2.80	179/143/113/89	Glucitol-tetramethyl-diacetate	Wang et al. (2013)	x	x	x
8	21.28	935.0783	C41H28O26	− 0.20	765/545	Casuarinin	Baxter et al. (1999)	x	-	-
9	23.98	433.0402	C19H13O12	0.09	301	Ellagic acid pentoside isomer I	Ferreres et al. (2013)	x	-	-
10	24.19	477.0662	C21H17O13	− 0.26	315	Methyl ellagic acid glucoside	Wu et al. (2008)	x	x	-
11	24.96	433.0402	C19H13O12	0.09	301	Ellagic acid pentoside isomer II	Ferreres et al. (2013)	x	-	-
12	25.16	300.9987	C14H5O8	0.10	257/229/185	Ellagic acid	El-Shenawy et al. (2008)	x	x	-
13	25.58	463.0871	C21H19O12	− 0.01	301	Quercetin galactoside (hyperin)	El-Shenawy et al. (2008)	x	x	x
14	25.76	447.0957	C21H19O11	− 0.94	285	Kaempferol hexoside	Doi et al. (2001)	x	x	x
15	26.15	463.0871	C21H19O12	− 0.01	301	Quercetin glucoside (isoquercitrin)	El-Shenawy et al. (2008)	x	x	x
16	26.16	455.2121	C19H35O12	− 0.45	409/240	Unknown		x	x	x
17	26.50	579.1348	C26H27O15	0.50	285/429/447/255	Kaempferol-di-rhamnoside (kaempferitrin)	Jaiswal et al. (2014)	x	x	x
18	27.19	433.0761	C20H17O11	− 0.91	301	Quercetin xyloside	Aoyama et al. (2018), El-Shenawy et al. (2008)	x	x	-
19	27.63	433.0761	C20H17O11	− 0.98	301	Quercetin arabinoside	Aoyama et al. (2018), El-Shenawy et al. (2008)	x	x	x
20	28.51	431.0967	C21H19O10	− 1.16	269	Apigenin-hexoside	Zhu et al. (2015)	x	x	x
21	28.98	447.0918	C21H19O11	− 0.73	285	Kaempferol galactoside	Wu et al. (2008)	x	x	x
22	30.17	315.0139	C15H7O8	0.15	301	Methyl ellagic acid	Wu et al. (2008)	x	-	-
23	30.38	447.0918	C21H19O11	− 0.73	285	Kaempferol glucoside	Doi et al. (2001)	x	x	x
24	30.57	551.1034	C24H23O15	0.40	343	Phosphatidylcholine	Wu et al. (2008)	x	x	-
25	31.14	301.0349	C15H9O7	0.01	179/ 151	Quercetin	El-Shenawy et al. (2008)	x	x	x
26	34.73	394.9716	C15H7O11S	0.01	315	Methyl ellagic acid sulphate isomer I	Horai et al. (2010)	x	-	-
27	35.02	285.0452	C15H9O6	1.85	187/ 159	Kaempferol	El-Shenawy et al. (2008)	x	x	x
28	37.53	394.9705	C15H7O11S	0.16	315	Methyl ellagic acid sulphate isomer II	Horai et al. (2010)	x	-	-
29	38.87	269.1234	C15H9O5	2.90	183/ 151	Apigenin	Zhu et al. (2015)	x	x	x
[M]+/[M–H]+
30	18.38	627.1533	C27H31O17	2.10	465/303	Delphinidin di-glucoside	Sarais et al. (2016)	x	-	-
31	20.97	465.1008	C21H21O12	0.90	303	Delphinidin-glucoside	Sarais et al. (2016)	x	x	-
32	22.20	435.0909	C20H19O11	− 1.27	303	Delphinidin-pentoside	Sarais et al. (2016)	x	x	-
33	23.43	449.1055	C21H21O11	− 2.3	287	Cyanidin-3-O-glucoside	Sarais et al. (2016)	x	x	x
34	30.56	614.2832	C37H42O8	2.35	438/420	Unknown		x	x	-
35	33.79	343.2945	C20H39O4	− 5.02	240	Unknown		x	x	x
36	37.26	327.2171	C18H33O5	0.53	Non fragmented	Unknown		-	-	x
37	39.38	329.2328	C18H31O5	0.53	Non fragmented	Unknown		-	-	x

x, detected; -, not detected

The negative LC–MS metabolic profiles highlighted the presence of a large group of compounds, of which 26 were phenolic compounds, corresponding to flavone glycosides and ellagic acid derivatives (Table 1, Fig. 1). The positive LC–MS metabolic profiles highlighted instead the presence of several anthocyanins (Table 1).

Fig. 1.

LC-ESI-Orbitrap MS total ion chromatogram acquired in negative ion mode for whole feijoa flower (FM), feijoa petals (PM) and feijoa petals juice (PJ). LC and MS conditions are reported in the experimental section. Compound number as reported in Table 1

Compounds 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 22, 26, 28 were identified as compounds pertaining to the galloyltannins class, and most of them are derivatives of ellagic acid. It is interesting to note that these molecules were found only in FM and in PM, but not in PJ extracts. Compounds 2, castalagin and 12, ellagic acid, were previously identified in A. sellowiana leaves (El-Shenawy et al. 2008; Ruberto and Tringali 2004). Compound 1 and 3 showed a pseudomolecular ion at m/z 783.0679 that produced in MS/MS a similar fragmentation pattern. After MS/MS analysis, fragment ions at m/z 301 (ellagic acid; [M − 482]⁻, loss of HHDP hexose) and at m/z 481 (deprotonated HHDP hexose; [M − 302]⁻, loss of HHDP), were evident, thus compounds were identified as di-HHDP hexose, presumably pedunculagin and its isomer; the fragmentation pattern was in accordance with pedunculagin (D’Urso et al. 2017) in Myrtus communis. Moreover, pedunculagin was previously reported in flower buds of Feijoa sellowiana (Aoyama et al. 2018). Compound 4, strictinin, showing a HR-MS value at m/z 785.0831 corresponding to a formula C₂₇H₂₂O₁₈, was identified by comparing data with literature, the compounds was previously found in walnuts (the seeds of Juglans regia L.) (Fukuda et al. 2003). Compound 5, tellimagrandin I, m/z 785.0831, C₃₄H₂₅O₂₂, and compound 8, casuarinin, accurate mass value 935.0783 m/z, with formula C₄₁H₂₈O₂₆ were reported in Phytochemical Dictionary published (Baxter et al. 1999). Compound 6, nilocitin, 483.1012, corresponding to the molecular formula C₂₀H₁₉O₁₁, was putatively identified by comparison with data published (Mahmoud et al. 2001). Compounds 10 and 22 were identified as respectively methyl ellagic acid glucoside and methyl ellagic acid (Wu et al. 2008). Compounds 9 and 11 were identified as isomers of ellagic acid pentoside derivative isomers, with m/z value of 433.0402 and molecular formula of C₁₉H₁₃O₁₂. Compounds 26 and 28 were putatively identified by using Mass Bank by comparing their m/z values and their MS/MS fragmentation (Horai et al. 2010). Compounds 13, 14, 15, 17, 18, 19, 20, 21, 23, 25, 27, 29 were identified in all the three different extracts as flavonoids and their glycosylated forms. In particular, aglycones occurrent in feijoa flowers are quercetin, kaempferol, and apigenin. Glycosylation are possible both with hexose and pentose sugar. The occurrence of the single metabolites is different among the three extracts under identification and it is reported in Table 1. In addition, LC-ESI-Orbitrap-MS was conducted in positive ion mode with the aim to identify the presence metabolites pertaining to the group of anthocyanins. Different peaks were observed in the positive TIC profiles, and four of them were attributed to anthocyanins and specifically to delphinidin glycosylated derivatives and cyanidin-3-O-glucoside (compounds 30, 31, 32 and 33). Their identification was confirmed by comparison with literature data (Sarais et al. 2016).

HPLC–DAD analysis of major phenolic classes in feijoa flowers

Targeted phenolic compounds were dosed by HPLC–DAD analysis (Table 2). Among them, anthocyanins, flavonols and ellagitannins were the most represented, while hydroxycinnamic acids and benzoic acids were not detected. Although FM and PM showed similar composition, quantitative differences were observed. For instance, quercetin-3-O-galactoside was more concentrated in FM (12.4 ± 0.4 mg/L) and kaempferol-3-O-hexoside in PM (25.6 ± 1.6 mg/L). The most concentrated flavonol in PJ was apigenin (3.4 ± 0.2 mg/L). All these flavonols exert important activities in improving health, as has often been demonstrated through different studies (Lu et al. 2016). Anthocyanins were largely detected in FM (115.3 ± 3.6 mg/L), probably because of the presence of numerous intense red stamens (Ramirez and Kallarackal 2017) which are known to be a source of anthocyanins in flowers (Mori et al. 2009). Among anthocyanins, cyanidin-3-O-glucoside resulted to be the most represented on all the three extracts (115.3 ± 3.6, 31.7 ± 1.0 and 10.5 ± 0.4 mg/L in FM, PM and PJ, respectively). Anthocyanins are known to strongly contribute to antioxidant capacity and anti-inflammatory properties of fruit and vegetables (Święciło et al. 2018), which can reduce the risk of chronic diseases involving oxidative stress and inflammation. Specifically, cyanidin-3-O-glucoside is well-known for being one of the major anthocyanins in berries (Veberic et al. 2015) and for positive effects on counteracting oxidative damage and inflammation. Furthermore, it proved to have a significantly higher relative bioavailability (Diaconeasa et al. 2017). Ellagitannins were mainly dosed in FM and ellagic acid resulted to be the most concentrated (7.9 ± 0.2 mg/L). These compounds together with condensed tannins proved to have much greater free radical scavenging ability than ascorbic acid, tocopherols and low-molecular weight polyphenols (Szajdek and Borowska 2008). This class of compounds is also well-known to exert several biological effects as anti-inflammatory and anti-microbial agents and may be protective against different chronic diseases (Landete 2011).

Table 2.

Targeted phenolic compounds determined by HPLC–DAD analysis (mg/L, mean ± SD)

Compound	Identificationa	LOD	LOQ	Feijoa flowers extract
		mg/L	mg/L	Flower macerate (FM)	Petal macerate (PM)	Petal juice (PJ)
Galloyltannins
Pedunculaginb	UV–VIS, MS	0.4	1.1	2.0a ± 0.0	nd	nd
Pedunculaginisomerb	UV–VIS, MS	0.4	1.1	1.2a ± 0.1	nd	nd
Ellagic acid pentoside isomer Ib	UV–VIS, MS	0.4	1.1	1.8a ± 0.1	nd	nd
Methyl ellagic acid glucosideb	UV–VIS, MS	0.4	1.1	2.0a ± 0.0	nd	nd
Ellagic acid pentoside isomer IIb	UV–VIS, MS	0.4	1.1	tr	nd	nd
Ellagic acid	Rt, UV–VIS, MS	0.4	1.1	7.9a ± 0.2	0.4b ± 0.0	nd
Total				15.0a ± 0.9	0.4b ± 0.0	nd
FLAVONOLS
Quercetin-3-O-galactoside	Rt, UV–VIS, MS	0.6	1.8	12.4a ± 0.4	2.2 ± 0.2	nd
Kaempferol-3-O-hexosidec	UV–VIS, MS	0.4	1.2	5.9a ± 0.1	25.6b ± 1.6	1.2c ± 0.0
Quercetin-3-O-glucoside	Rt, UV–VIS, MS	0.6	1.8	3.4a ± 0.2	tr	nd
Kaempferol-di-rhamnosidec	UV–VIS, MS	0.4	1.2	3.2a ± 0.1	tr	nd
Apigenin hexosided	UV–VIS, MS	0.6	1.8	3.9a ± 0.1	4.6b ± 0.1	nd
Kaempferol-3-O-hexosidec	UV–VIS, MS	0.4	1.2	5.5a ± 0.2	6.8b ± 0.2	nd
Quercetin	Rt, UV–VIS, MS	0.4	1.2	2.6a ± 0.0	2.5a ± 0.2	nd
Kaempferol	Rt, UV–VIS, MS	0.4	1.3	4.3a ± 0.1	1.8b ± 0.2	nd
Apigenin	Rt, UV–VIS, MS	0.6	1.7	1.7a ± 0.0	1.7a ± 0.1	3.4b ± 0.2
Total				42.9a ± 3.3	45.1a ± 2.1	4.7b ± 0.5
Anthocyanins
Delphinidin di-glucosidee	UV–VIS, MS	0.4	1.2	tr	tr	nd
Delphinidin-3-O-glucoside	Rt, UV–VIS, MS	0.4	1.2	tr	tr	nd
Delphinidin-pentosidee	UV–VIS, MS	0.4	1.2	tr	tr	nd
Cyanidin-3-O-glucoside	Rt, UV–VIS, MS	0.5	1.6	115.3a ± 3.6	31.7b ± 1.0	10.5c ± 0.4
Total				115.3a ± 3.6	31.7b ± 1.0	10.5c ± 0.4

Results are reported as the mean value ± standard deviation; n = 3

Different superscript letters in each row indicate significant differences (p ≤ 0.05)

nd, not detected (< LOD); tr, traces (< LOQ)

^aRt, comparison with retention time of pure standard; UV–VIS, comparison with UV–VIS spectra of pure compound or similar pure standards; MS, comparison with LC/ESI/QqQ/MS spectra of pure compound or literature data (see references in Table 1);

^bDosed with the calibration curve of ellagic acid

^cDosed with the calibration curve of kaempferol-3-O-glucoside

^dDosed with the calibration curve of apigenin

^eDosed with the calibration curve of delphinidin-3-O-glucoside

Antioxidant capacity and reducing sugars content of feijoa flowers

Results of the antioxidant capacity assessed by FRAP, CUPRAC, DPPH^·, and ABTS^·+ assays and the TP content dosed by Folin–Ciocalteu’s assay are reported in Table 3. It can be seen that FM had far apart the best antioxidant activity among the flower parts studied (20.99 ± 1.34 mmol Fe²⁺/L vs. 3.59 ± 0.03 and 0.97 ± 0.08 mmol Fe²⁺/L of PM and PJ, respectively). This outcome could be explained as FM proved to be three to fourfold more concentrated in polyphenols (395.14 ± 7.91 mg GAE/L) than PM (98.59 ± 7.89 mg GAE/L) and PJ (114.53 ± 9.46 mg GAE/L). This high content in polyphenol contained in FM confirmed the data obtained by the HPLC–DAD analysis. The FM total polyphenol content was comparable to that of a decoction/infusion of other edible flowers such as Opuntia ficus-indica (Ammar et al. 2015) and Hibiscus sabdariffa (Prenesti et al. 2007) ones, which as well as feijoa flowers have been considered for human consumption. Thus, it is conceivable that FM could be, among feijoa flowers products, better exploitable in the perspective of a novel nutraceutical product design. On the other hand, PJ contains detectable amounts of fructose and glucose (6.40 ± 0.58 and 3.14 ± 0.58% w/v respectively, data not shown), which are present in traces in PM and absent in FM. This could make the juice from feijoa petals particularly appreciated for preparing sweet edible products without adding exogenous sweeteners.

Table 3.

Antioxidant capacity and total polyphenol content of different extracts of feijoa flowers

Parameter	Extract
	Flower macerate (FM)	Petal macerate (PM)	Petal juice (PJ)
FRAPa (mmol Fe2+/L)	20.99a ± 1.34	3.59b ± 0.03	0.97c ± 0.08
CUPRACa (mmol Fe2+/L)	7.62a ± 0.46	1.46b ± 0.11	1.09c ± 0.07
DPPHb (mmol TEAC/L)	2.30a ± 0.14	0.40b ± 0.00	0.16c ± 0.01
ABTS·+b (mmol TEAC/L)	9.35a ± 0.42	1.53b ± 0.12	1.72b ± 0.14
Total polyphenolsc (mg GAE/L)	395.14a ± 7.91	98.59b ± 7.89	114.53b ± 9.46

Results are reported as the mean value ± standard deviation; n = 3

Different superscript letters in each row indicate significant differences (p ≤ 0.05)

^aFRAP and CUPRAC values are expressed as Fe²⁺ millimolar concentration, obtained from a FeSO₄ solution having an antioxidant capacity equivalent to that of the dilution of the feijoa macerate/juice

^bDPPH^· and ABTS^·+ values are expressed as TEAC millimolar concentration, obtained from a Trolox solution having an antiradical capacity equivalent to that of the dilution of the feijoa macerate/juice

^cGAE, gallic acid equivalent

Conclusion

This research represents the first comprehensive study on the chemical composition of edible A. sellowiana flowers, whose massive production makes them suitable to exploitation as source of antioxidant compounds. The preliminary analyses of different extracts of this product, assessed by LC–MS/MS and HPLC–DAD, revealed an interesting content of ellagitannins, anthocyanins and flavonoids. FM macerate showed the highest concentration of these compounds, which was reflected in the hugest antioxidant activity among the three extracts. However, also the two other extracts which were studied, namely PM and PJ, possessed good anthocyanins and reducing sugar concentration, respectively. In particular, PJ could be used directly as a food, as it is currently done in different world regions, or added to beverages and smoothies for improving their nutraceutical properties. Therefore, feijoa flowers, which can be found widespread on all continents, could be recommended for the extraction of bioactive compounds and the preparation of nutraceuticals and dietary supplements, whose request in the international marketplace is increasing.