Determination of Polyphenolic Compounds by Ultra-Performance Liquid Chromatography Coupled to Tandem Mass Spectrometry and Antioxidant Capacity of Spanish Subtropical Fruits

In an analysis of the seven types of subtropical fruits most consumed and produced in southern Spain, UPLC-ESI-MS/MS was used to quantify 14 phenolic species: five hydroxycinnamic acids, seven hydroxybenzoic acids and two flavonoids (quercetin and naringenin). In each case, in addition, antioxidant capacity was determined by FRAP, ABTS and DPPH. Of these fruits, carambola (or starfruit) presented the highest levels of phenolic compounds, cherimoya (custard apple) and kiwi were the richest in non-flavonoid phenolic compounds and papaya had the highest levels of the flavonoids studied. Higher mean values were recorded in home-grown fruits than in imported varieties by ABTS and DPPH methods. Persimmon’s antioxidant capacity was well above that of the other fruits, according to our analyses.


Introduction
In recent years, there has been increasing interest in dietary phytochemicals or How to cite this paper: Muñoz sensory properties, such as flavour and colour, and contribute to the aroma and taste of many foodstuffs of vegetable origin. Consequently, they are of great importance in the food industry.
Spain is one of the main suppliers of tropical fruit to other European countries, and exported 115,129 tonnes in 2012 (latest available data), corresponding to 0.95% of all its fruit and vegetable exports [3] (MAPAMA, 2017). The Spanish Mediterranean coast presents environmental characteristics typical of a Mediterranean subtropical climate, which makes the coastal zones of Granada and Málaga provinces the only areas suitable in Andalusia for this type of crop [4] (CMAOT, 2015).
The present study was conducted to evaluate the bioactive compounds of Spanish-grown tropical fruits, especially their antioxidant capacity. The data obtained on these tropical fruits can be usefully incorporated into tables of nutritional composition.
In addition to the above study goals, ultra-high performance liquid chromatography (UPLC), coupled with mass spectrometry, was used to quantify individual phenolic compounds of nutritional importance in these home-grown subtropical fruits.

Samples
The following tropical fruits are the types most consumed in Spain and have been analysed in this study: carambola (starfruit) (Averrhoa carambola, n = 10) cherimoya (custard apple) (Annona cherimolla, Mill., n = 10), kiwi (Actinidiadeliciosa cv Hayward, n = 10) (Mangifera indica L., n = 10), papaya (Carica papaya L., n = 10), persimmon (Diospyros kaki L.) and avocado (Persea Americana Mill., Hass and Bacon varieties). Samples of Spanish-grown and imported fruits were analysed (except for some phenolic species of which, due to limited seasonal availability, only Spanish-grown samples were analysed). The samples were obtained from retailers in the cities of Granada and Málaga and were immediately taken to the laboratory to be processed. All non-edible parts-skin, seeds and/or pits-were removed; the separated edible part was then triturated in a blender and the juice extracted from each fruit. The processed samples were stored at 4˚C until needed for analysis, which was performed as soon as possible.
Each sample consisted of two randomly-chosen pieces of fruit from the items contained in the sales unit, usually a basket or mesh bag. All analyses were conducted in triplicate.

Reagents and Standards
The following reagents were obtained from Panreac Química SL (Barcelona,

Phenolic Compounds
Individual phenolic compounds. The UPLC-ESI-MS/MS method was used to identify the different phenolic compounds [5]  Treatment of the sample. A liquid-liquid extraction with final concentration was performed. Then, 20 mL of diethyl ether was added and the resulting solution was frozen at −20˚C for 24 hours, after which it was centrifuged for 10 minutes at 9000 rpm. The supernatant was transferred to a separatory funnel and three extractions made with 20 mL diethyl ether. A spatula tip of anhydrous sodium sulphate was added to the organic extract; this was then filtered and the filtrate was passed to a heart-shaped flask for rotary evaporation at 30˚C to the smallest possible volume. The extracts were collected with 1 mL methanol/water mixture (1:1), filtered through a 0.20 μm membrane filter and passed to a chromatography vial for analysis. Analytical validation. Table 1 and Table 2 show the main analytical validation parameters of the method used. Figure 1 shows a chromatogram obtained for a sample of mango.

FRAP Assay
The ferric-reducing ability of each sample solution was estimated using the procedure described previously [6] by Benzie and Strain (1996), adapted to a microplate reader as reported by Pastoriza et al. [7]. Calibration curves were performed using Trolox stock solutions. The results are expressed as mmol equivalents of Trolox per L of sample.

ABTS + Assay
The antioxidant capacity was estimated as the radical scavenging activity, following the procedure described previously [8] by Roberta et al. (1999) as adapted to a microplate reader by Pastoriza et al. [7]. Calibration was performed with a Trolox stock solution as described above. Results are expressed as mmol equivalents of Trolox per L of sample.

DPPH Assay
Radical scavenging activity was determined according to the method reported Pastoriza et al. [7]. Calibration was performed with a Trolox stock solution as described above. Results are expressed as mmol equivalents of Trolox per L of sample.

Statistical Analyses
The statistical analyses were conducted using the Statgraphics Centurion XVI.II53 (2009) software package. The normality of the data set was confirmed by the Shapiro-Wilk test, and the homogeneity of the variances was confirmed by Levene's test, for a significance level of 5% (p > 0.05). One-way analysis of variance (ANOVA) was performed to compare more than two groups, and in cases where significant differences were obtained (p < 0.05), the multiple LSD test or Tukey's rank test was applied. If a non-parametric test was needed, the Kruskal-Wallis test was used. The degree of significance for all tests was p ≤ 0.05.

Individual Phenolic Compounds
Our  of fruit, respectively). This phenolic compound is used as a substrate to determine the activity of the cholesterol esterase enzyme [15]. (Planutis et al., 1986).
The fruits richest in the phenolic compounds analysed were carambola and cherimoya, with 0.821 and 0.791 mg/100g of fresh fruit, respectively. According to Macheix et al. [11], 5 -100 mg/100g of fresh fruit is the normal range, but in our analysis of these hydroxycinnamic acids, none of the fruit samples fell within this interval.

Hydrobenzoic Acids
As observed above, gallic acid was the only phenolic compound quantified in all the fruits analysed, with a presence ranging from 3.41E−01 ± 0.008 mg/100g of fruit in papaya to 7.99E + 00 ± 0.585 mg/100g of fruit in kiwi. This phenolic compound is found in a wide range of fruits and vegetables, at a concentration of 0.5 -15 mg/100g of fresh fruit [11] (Macheix et al., 1990), and therefore our samples, while heterogeneous, all contain a significant quantity within this interval.
Gallic acid is the main phenolic compound in mango [16]. (Kim et al. 2007 Macheix et al. [11], all the fruits in our samples except papaya fell within this range.

Flavonols
Except in cherimoya, quercetin was quantified in all the fruits analysed., with a content ranging from 1.86E−03 ± 0.001 mg/100g of fruit in mango to 1.45E−02 ± 0.0012 mg/100g in papaya.
According to Macheix and Fleuriet [11], the usual range of flavonols is 0.

Flavanones
Naringenin was identified and quantified in four of the fruits analysed, with a content ranging from 2.01E−03 ± 0.002 mg/100g of fruit in mango, to 4.20E−03 ± 0.012 mg/100g in paraguayo. The other two fruits in which it was present were cherimoya and carambola (2.06E−03 ± 0.018 and 3.75E−03 ± 0.002 mg/100g of fruit, respectively).
Of the fourteen phenolic compounds and seven hydroxybenzoic acids analysed, gallic acid was found in all the fruits, and was present in the highest quantities. The mean values were significantly higher (p < 0.01) for kiwi and paraguayo than for the other fruits, with a content ranging from3.41E−01 ± 0.008 mg/100g of fruit in papaya to 7.99E+00 ± 0.585 mg/100g in kiwi. Gallic acid provides important health benefits through its anti-cancer and cardioprotective  [22]. Analysis of the variance among the phenolic compounds quantified in each of the fruits analysed revealed statistically significant differences (p < 0.001) for gallic acid, vanillic acid, p-coumaric acid and quercetin (p < 0.005).
With respect to the total quantity of phenolic compounds (as determined by liquid chromatography), the fruit with the highest quantity was kiwi, followed by paraguayo and carambola (kiwi > paraguayo > carambola > cherimoya > mango > papaya).
Carambola had the highest concentrations of six individual phenolic species: p-coumaric acid (p = 0.0135), vanillic acid (p = 0.038), ferulic acid, ellagic acid, pOH benzoic acid and 3,5 dimethoxybenzoic acid, and in another threequercetin (p = 0.0383), caffeine and naringenin -it was the second-most abundant (difference non-significant). Papaya had the highest content of quercetin and sinapic acid. Kiwi had more gallic acid (p < 0.01) than any other fruit, and cherimoya had the highest caffeine content (p < 0.01).

Antioxidant Capacity
Due to the hydrophilic and lipophilic nature of the antioxidant components found in tropical fruits, diverse methods have been proposed to evaluate their antioxidant capacity. We performed three of these-ABTS, DDPH and FRAP.

ABTS Method
The results obtained for Spanish-grown and imported tropical fruits, expressed as mmol of Trolox equivalent antioxidant capacity (TEAC) per g of fruit, are shown in Table 4.
In our samples, the mean antioxidant activity ranged from 2.226 ± 0.761 mmol/g for avocado (var. Bacon) to 31.532 ± 16.801 mmol/g for persimmon.
Another fruit presenting high antioxidant capacity was cherimoya, with 22.094 ± 11.101 mmol/g. After persimmon and cherimoya, the next highest values were obtained for carambola, kiwi and papaya (Table 4) (8.734 ± 4.471, 5.406 ± 0.805 and 3.769 ± 0.834 mmol/g, respectively). Like persimmon, these fruits have abundant vitamin C; on the other hand, the levels of beta-carotene are lower than in persimmon, which may explain their lower antioxidant capacity [23] (Morillas-Ruíz and Delgado-Alarcón 2012).

Kiwi and avocado (var. Hass) had similar values in imported and Span-
ish-grown samples (5.65 ± 0.723 and 2.293 ± 0.592 mmol/g, respectively) but in papaya and carambola, the values in the imported fruits were lower than in the Spanish-grown ones (3.13 ± 0.414 and 7.18 ± 4.952 mmol/g, respectively).
Our values for the antioxidant capacity of carambola were lower than those obtained by previous studies [23]  tively, which were also lower than those reported elsewhere for kiwi [26] (Lim et al., 2007) and papaya [27]  For persimmon, the values observed in our samples were lower than those reported by Fei et al. [29], for both Spanish-grown and imported fruit.
For mango and cherimoya, too, our values were somewhat lower than those obtained previously [30] (Vasco et al., 2008), for Spanish-grown and imported fruit.

DPPH Method
The results obtained for the Spanish-grown fruit samples, expressed as mmol of TEAC per g of fruit, are shown in Table 4. The mean values ranged from 191.657 ± 114.168 mmol/g of fruit for persimmon to 0.859 ± 0.442 mmol/g for mango. As with the ABTS method, the values ranged widely among samples.
With the DPPH method (see Table 4), although persimmon continued to present the greatest antioxidant capacity, the values were lower, both overall (2.432 ± 1.814 mmol/g) and in Spanish-grown and imported fruit (3.898 ± 1.427 and 0.966 ± 0.391 mmol/g, respectively). The values for Spanish-grown carambola (14.828 ± 4.732 mmol/g) were the second highest of all the samples, but were followed very closely bypapaya (14.167 ± 2.510 mmol/g). The Spanish-grown kiwi samples presented values of 8.486 ± 1.150 mmol/g, while lower ones were obtained for avocado (var. Hass) (3.820 ± 0.827 mmol/g), cherimoya (3.898 ± 1.427 mmol/g) and mango, which had the lowest value of all the fruits sampled (0.738 ± 0.443 mmol/g).
In the imported fruits, as with the Spanish-grown samples, the highest antioxidant value was found in persimmon (154.611 ± 99.870 mmol/g) and the lowest one in mango (0.979 ± 0.426 mmol/g). Cherimoya also presented a very low value (0.966 ± 0.391 mmol/g), well below that found in Spanish-grown fruits. In avocado and papaya, too, the values observed were lower than those of the home-grown varieties (3.563 ± 0.581 and 10.65 ± 1.788 mmol/g, respectively).However, imported kiwi and carambola had higher values than the Spanish-grown equivalents (10.010 ± 2.172 and 15.690 ± 4.273 mmol/g, respectively).
Our values for avocado and papaya were lower than those reported previously for fruit pulp [23]  but for papaya they were higher than those observed in a previous paper [27], by Zuhair et al. (2013). For persimmon, both Spanish-grown and imported, the values we obtained using the DPPH method were higher than those of Fei et al. [29].
The values for kiwi were lower than those reported before [31] [23] by Du et al. (2009) and Morillas-Ruiz et al. (2012) while for carambola, they were lower than those of Clerici et al. [24] but higher than those found by Morillas-Ruiz et al. [23]. The latter author [23] reported higher values than ours for mango, which is in line with the findings of Vasco et al. [30]. Table 4 shows our results for Spanish-grown and imported fruit samples, expressed as mmol of TEAC per g of fruit. The antioxidant activity of the Span- As with the methods described above, FRAP showed persimmon to have a greater antioxidant activity than the other fruits. Avocado had the lowest antioxidant capacity (as in the ABTS method). Among the varieties of avocado, the antioxidant capacity of var. Bacon (6.096 ± 7.422 mmol/g) was nearly three times stronger than that afforded by var. Hass (2.284 ± 1.198 mmol/g).
Among the imported samples, persimmon still gave the highest value (44.127 ± 48.661 mmol/g), and this was higher than that obtained for the Spanish-grown equivalent. Avocado (2.595 ± 1.289 mmol/g), mango (2.872 ± 1.950 mmol/g), carambola (6.264 ± 2.722 mmol/g) and papaya (4.388 ± 1.465 mmol/g) all had lower values than those obtained for home-grown fruits. The value for kiwi was slightly above that found in Spain-grown fruit (5.970 ± 1.391 mmol/g).
The values obtained for avocado and papaya were lower than those reported for fruit pulp [28] by Morais et al. (2015). However, by the ABTS method, the value for our papaya samples was higher than that reported [27] by Zuhair et al.

Nutritional Impact of Subtropical Fruits as Antioxidants in the Diet
In recent years, there has been increasing interest in dietary phytochemicalsincluding isothiocyanates, phenolic compounds, flavonoids, isoflavones, lignans, saponins and cumestrol-due to their possible protective action against cardio-  [36].
The contribution made by the consumption of Spanish-grown tropical fruit to antioxidant capacity and phenolic compound intake is shown in Table 5. As observed above, the consumption of tropical fruits in Spain is low, ranging from 0.4 to 0.9 kg/year, except for that of kiwi, which rises to 3.0 kg/year. The daily consumption of one kiwi (Spanish grown) provides over 20% of total antioxidant needs, and so this tropical fruit, the most commonly consumed in Spain, contributes significantly not only to mineral intake but also to antioxidant capacity.
The contribution to real personal diets is best calculated using the edible portion size of each fruit, which ranges from 90 g for avocado to 130 g for kiwi. Persimmon makes the highest daily contribution in the Spanish diet to antioxidant capacity, exceeding 700% [37]. The consumption of a portion of Spanish-grown cherimoya or papaya contributes over 50% of antioxidant capacity needs. Mango is the tropical fruit that makes the lowest contribution to total antioxidant capacity (16.14%) in the Spanish diet.
In order to enhance antioxidant capacity and to increase the intake of phenolic compounds, the consumption of tropical fruit, cultivated in the coastal areas of Granada and Málaga provinces, could be included in nutritional recommendations.   The presence of other nutrients in these foods (especially sugars, which have well-known implications for health) should also be taken into account. It is also important to consider the real bioavailability of phenolic compounds in these fruits.

Conclusions
According to our liquid-liquid extraction technique and UPLC-ESI-MS/MS analysis, carambola had the highest total content of the 14 individual phenolic compounds studied. Cherimoya and kiwi were the richest in non-flavonoid phenolic compounds, and papaya was richest in flavonoids. Gallic acid was the only phenolic compound quantified in all seven fruits.
The mean values found for antioxidant capacity, by ABTS, DPPH, were higher in Spanish-grown than in imported fruits. Persimmon show dean antioxidant capacity that was notably superior to that of the other fruits, which nevertheless also presented useful antioxidant capacity and, therefore, health benefits.
Carambola and cherimoya were also found to be very rich in these bioactive compounds.
Finally, our study contributes local data that can be incorporated into tables of nutritional composition in reference to flavonoid and non-flavonoid phenolic species. We also supply data on antioxidant capacity obtained from the consumption of a portion of these fruits, and relate these values to the Spanish daily intake. In this respect, we note, in particular, that the consumption of a portion of persimmon or cherimoya supplies over 100% or 50%, respectively, of the daily antioxidant requirements of a healthy adult.