HPLC-DAD and HPLC-ESI-MS Analysis of Polyphenol-rich Extracts from Mango (Mangifera indica L.), Tommy Atkins and Haden Varieties, Cultivated in Dominican Republic

Background: Antioxidant, anti-inflammatory, and analgesic effects of mango stem bark extracts (MSBE) have been reported. A previous published report described the identification of several components (polyphenols, polyols, and sugars) by HPLC, MS, and NMR. Mangiferin (2-β-D-glucopyranosyl-1,3,6,7-tetrahydroxyl-9Hxanthen-9-one) was identified as the major bioactive component of MSBE. However, MSBE has shown more potent effects than model mangiferin solutions in several in vitro experiments at different concentrations. Therefore, there are other extract components which contribute to the observed pharmacological effects. Objective: The identification of other polyphenolic components in MSBE, which may contribute to the observed pharmacological effects through a synergic pathway. Method: Polyphenol-rich extracts from mango stem bark and branch trees from two varieties (Haden and Tommy Atkins), cultivated in Dominican Republic, were analyzed by HPLC-DAD. Butanol MSBE extract from Haden mango stem bark was analyzed by HPLC-MS-ESI. Results: The mangiferin content in Haden mango extracts by HPLC-DAD was considerably high as compared to previous reports in the literature from other mango varieties. The identification of Haden mango stem bark butanolic extract components by HPLCESI-MS led to the unambiguous identification of 20 components: 2 benzoic acid derivatives, gallic acid and 8 gallate derivatives, 2 benzophenones (maclaurin and iriflophenone types), 3 flavonoids (catechin, epicatechin, and quercetin), mangiferin, isomangiferin, homomangiferin, and noratyriol. Conclusion: HPLC-ESI-MS analysis of the Haden mango SB extract showed the presence of many polyphenolic components, not previously reported, which may correlate to the antioxidant, anti-inflammatory, and/or analgesic effects of this extract mostly through a synergistic effect of its components. The potential exploitation of mango by-products from the Haden variety would be the best option for obtaining polyphenolrich extracts from mango agricultural by-products to be used as bioactive ingredients in nutraceutical, cosmeceutical, and/or pharmaceutical formulations.


Introduction
Mango pulp and juice, besides the fresh fruit commercialization, are the main commercialized products from the mango fruit industry; worldwide fruit production is above 40 million tons being India, China, Pakistan, and Brazil the main producers [1]. Uses of mango (Mangifera indica L.) extracts with antioxidant properties have been reported extensively in the last two decades and claims about its antioxidant properties have been published [2][3][4]. Mango fruit processing technologies produce a large amount of wastes, mainly peel and seed. The amount of (Mangifera indica L.), Tommy Atkins and Haden Varieties, Cultivated in Dominican Republic peel and seeds wasted from fruit collection has been estimated between 35% and 60%, respectively, depending on the region and the variety [5]. In order to assess the potential uses of these wastes, several procedures have been developed to obtain polyphenol-rich extracts from peel and seed kernel [6][7][8][9], with an impact on the industrial diversification of mango production, and the reduction of the environmental impact of industrial residues, within the concept of a circular economy. However, studies about the utilization of mango agricultural by-products (leaves, stem bark, and branch trees) for similar goals have not been performed in the same extent. Uses of leaf extracts, mainly aqueous decoctions for ethnomedical uses, have been reported as well as their antioxidant and anti-inflammatory properties [10] Leaves have been studied by several extraction procedures to obtain antioxidant-rich extracts [11,12]. However, little attention has been devoted to the potential exploitation of stem bark (SB), and branch trees (BT) in commercial mango plantations [2].
Our work on mango-derived products has focused mainly on SB, and BT, as agricultural by-products, which are not commonly used as starting raw materials for a possible industrial application. About 40 thousand tons of mangoes are collected annually in Dominican Republic along more than 3,000 ha from 58 varieties [1]. Trees are pruned annually in the autumn season (September-November) and these wastes are left in the field. Moreover, SB may be collected from mango trees throughout the year, and this collection may be repeated in the same tree every 3-4 years [13]. It has been estimated that more than 1,500 tons of mango BT are left in the field, and around 300 tons of mango SB may be collected annually in Dominican Republic. These agricultural by-products have a large content of antioxidant flavonoids and xanthones. Therefore, if these products are processed, leading to standardized polyphenol-rich extracts, with a defined content of flavonoids and xanthones, it may lead to a new source of bioactive ingredients to be used in food, cosmetic, and/or pharmaceutical formulations.
A previous work on the chemical characterization of mango stem bark-butanolic extract [14] led to the identification of mangiferin, 2-β-D-glucopyranosyl-1,3,6,7tetrahydroxy-9H-xanthen-9-one (MF), as the main component in the polyphenol fraction of mango stem barkbutanolic extract (7%). That extract has been studied extensively, both pre-clinical and clinical, and MF described as one of the most promising candidates to be developed as a bioactive ingredient to be used in cosmeceutic, nutraceutic or pharmaceutical formulations [15]. However, further studies demonstrated that isolated MF from the extract might be acting as antioxidant through a mechanism involving metal complexes [16]. We studied the major and trace element concentrations in the stem bark of 16 Cuban mango varieties [17], and observed that Se concentration was significantly higher for the Haden variety (Code H01). That result correlated with the antioxidant effect of mango stem bark extract not only to MF, but to Se (and possibly Cu and Zn).
The study of mango SB extract and MF on lipid peroxidation on red blood cells concluded that the extract had a higher protective effect than MF, which could explain that other extract components might be acting through a synergistic effect [18]. Therefore, further studies about chemical composition of the mango SB aqueous extract are needed in the attempts to correlate health effects to chemical components, where polyphenols play a significant role.
We report the polyphenolic characterization of agricultural by-products extracts by HPLC-DAD (stem bark and branch tree) from two mango varieties (Haden, and Tommy Atkins) cultivated in Dominican Republic, in order to demonstrate the potential uses of these by-products as sources of polyphenol-rich ingredients for industrial uses. The HPLC-ESI-MS analysis of polyphenols from the stem bark of Haden mango butanolic extract, as the most promising variety for industrial exploitation, is also reported for the determination of other phenolic constituents that may contribute to the observed biological activities.

Chemicals
PA and HPLC grade solvents (methanol, n-butanol, acetonitrile, and DMSO) and reagents (acetic acid, formic acid, sodium carbonate, and sodium benzoate) were purchased from J. T. Baker (USA). Folin-Ciocalteau reagent was purchased from Merck (Darmstadt, Germany). Milli-Q system from Millipore (Milford, MA) was used for producing ultrapure water. The following standards were purchased from Sigma-Aldrich Co. (

Plant Material
The stem bark (SB) was collected, free of biological contamination (fungi), from trees grown in farms located in the Bani (Haden), and Puerto Plata (Tommy Atkins) regions, Dominican Republic, in the autumn season (2018 and 2019), without affecting the ecosystem, according to a Standardized Operational Procedure (SOP). Briefly, bark was marked, with not more than 2 cm depth, making a rectangle of around 10 cm width and up to 50 cm height, depending on the tree size. The bark was separated without damaging the inner part of the stem, with specially designed tools. Bark pieces were cleaned manually from dust and residues, and collected in sealed polystyrene bags, approximately 50 kg per bag. Bags were stored protected from light at room temperature, and transported to the lab within the next 7 days after collection. Voucher specimens were deposited at the Biological Products Archive, National Evangelic University, Santo Domingo, Dominican Republic (Codes: 10228B-Haden-, and 10229B-Tommy Atkins). The branch trees (BT) were collected, free of biological contamination (fungi), from trees grown in farms located in Azua (Haden), and Bani (Tommy Atkins) regions, Dominican Republic, in the 2018 autumn season, according to a SOP. Briefly, BT were cut and left to dry in the field for 1 week by solar radiation. Dried branches were milled on-site with a crusher and collected in polystyrene sealed bags, approximately 50 kg per bag. Bags were stored protected from light at room temperature, and transported to the laboratory within the next 7 days after collection. Voucher specimens were deposited at the Biological Products Archive, National Evangelic University, Santo Domingo, Dominican Republic (Codes: 10228T-Haden-, and 10229T-Tommy Atkins).

Sample Preparation
SB and BT were dried at sunlight for 7 days and subsequently with a hot-air dryer (Thermo Scientific, USA) at 60°C for 2 hours. Both samples were milled on mortar grinder (Retsch, Germany, Model RM 200), and packed in vacuum-sealed bags (500 g) until extraction. SB, and BT yields were determined for both dried plant materials.

Water Content
The Water Content (WC) of dried SB and BT plant materials, and dried plant extracts were determined with a moisture analyzer (Radwag, Poland, Model PMR 50) by triplicate, and results expressed as mean value±standard error (p > 0.05).

Extraction Methods
100 g of milled SB, or BT (maximum 10% WC) were extracted with a Soxhlet equipment (1 L) with a mixture of 70:30 methanol: deionized water for 1 hour. Solvent was evaporated under vacuum with a rotary evaporator at 60°C (Rotavapor E100, MCR, Israel). The residue was brought to dryness with a dry flow of nitrogen, and extract yield was determined. Five batches for each mango by-product were extracted and yield results expressed as mean value±standard error (p > 0.05). The solid extract from the Haden mango stem bark (1 g) was suspended in 50 mL distilled water and extracted twice with 25 mL water-saturated n-butanol. The nbutanol layer was brought into a rotary evaporator, and nbutanol was evaporated under vacuum until a brown solid residue was obtained. Samples for HPLC-ESI-MS analysis were prepared by dissolving 10 mg of the solid extract in DMSO into a 10-mL volumetric flask (1 mg/mL).

Polyphenol Content
The Polyphenol Content (PPC) of extracts was determined by a modified Folin-Ciocalteu method using a catechin-equivalent standard [14]. Mango extract (1.15 mg) was dissolved in methanol (2 mL), and the solution was diluted 10-fold with distilled water. Folin-Ciocalteau reagent (0.5 mL) was added to the diluted solution followed by 0.5 mL of sodium carbonate, 100 g/L solution. The absorbance was measured at 700 nm (Thermo Scientific, Genesys 10 spectrophotometer) with a blank sample (water plus reagents) in the reference cell (1 cm-depth quartz). Quantification was performed by plotting the absorbance value in a calibration curve of (+) catechin used as standard phenol.

Quantitative Determination of Polyphenols
Single polyphenols in mango SB, and BT extracts were determined by HPLC-DAD analyses with a Young Lin HPLC System (South Korea) equipped with a YL-9110 quaternary pump, YL-9150 autosampler (fitted with a 20 µL loop), a YL-9160 diode-array detector (DAD) coupled to a data acquisition and processing system (Clarity software). Column (RP-18, 5 µm, 250 x 4 mm i.d., Merck, Germany) was placed in a YL-9131 column oven at 30ºC. Solvents were degassed (YL-9101), and injection volume was 20 µL. Separation was carried out by gradient elution with two solvents [A=acetic acid (0.1%) in water; B=acetic acid (0.1%) in methanol]. The ratio of A:B was increasing from 9:1 to 1:9 in 35 min at a flow rate of 1 mL/min. Detection wavelength was fixed at 278 nm. Gallic acid, methyl gallate, propyl gallate, (+) catechin, (-) epicatechin, mangiferin, and quercetin were identified by the retention times of the chromatograms of the available pure standards with retention times of 4.5, 7.2, 8.4, 11.8, 13.5, 14.6, and 19.6, respectively. Isomangiferin (RT=15.0) was identified by comparing its retention time and UV spectra with the mangiferin standard, which had isomangiferin as impurity.

Quantitation
Quantitation was performed by plotting the integrated peak area into the calibration curve of the corresponding pure standard, except for the unknown (peak 2) and isomangiferin (peak 8), which were calculated by an Internal Normalization Method. All polyphenols were assayed as (+) catechin equivalent.

Linearity
Linearity was determined for the seven available pure standards. Linearity of responses was determined with five levels of concentrations (5, 10, 15, 20, and 25 µg/mL), and three injections for each concentration. A linear correlation coefficient (r=0.9998) was observed, and the minimun detection limit was 0.18 ng, which resulted in a signal-tonoise ratio of 3.5:1.

HPLC-ESI-MS Analysis of Polyphenols
Polyphenolic composition of Haden mango SB butanol extract was determined through a Thermo Finnigan LCQ-Ion Trap (Thermo Separation, USA) with electrospray ionization. Mass spectra were recorded on negative mode between 50 and 1,500 Da; colision chamber temperature 375°C; pressure 4.1 bar; dry nitrogen flow as nebulizer, 10 mL/min, and helium as colision gas. First and second orders fragmentation patterns were recorded at 1.2, and 1.5 V, respectively. Chromatographic separation was performed on a Waters 600E-HPLC with a RP-18 column, 250 x 4.5 mm (Kromasil, Sweden) at 30°C. Injection volume was 5 µL. Separation was carried out by gradient elution with two solvents [A=formic acid (0.1%) in water; B=formic acid (0.1%) in acetonitrile]. The ratio of A:B was increasing from 9:1 to 10:0 in 90 min at a flow rate of 1 mL/min. Data acquisition and peak integration analysis was performed using XCalibur software.

Statistic Analysis
Data from both varieties, Haden and Tommy Atkins, were analyzed using the statistical programs SPSS 9.0. Nonparametric Friedman test using Wilconxon's pair test was assessed. Mann-Whitney U test was used to estimate statistical differences (p > 0.05) between varieties. All data in the tables are expressed as the mean value±standard error.

Agricultural and Extraction Yields
Water content from SB was down out from around 50% to a range between 6.5% and 8% with subsequent solar drying and rotary hot air drying. BT, which was allowed to solar drying in the field for a week, was directly submitted to hot air drying due to its low water content (around 10% WC). BT yields were considerably higher as compared to SB. A significant lower yield of SB was obtained in the Bani farm as compared to the Puerto Plata farm (Table 1). The best extract yield (12.4%) was obtained from Haden SB, Bani farm, being significant higher than extracts of other mango by-products. BT extracts from the same variety had the lowest extraction yield, but it seems to have the highest content of polyphenols and MF, although it was non-significant statiscally. MF content in the polyphenol fraction ranged from 41,6% (BT Tommy Atkins) to 55.4% (SB Haden). Results of extraction procedures and chemical analysis of mango by-products extracts from both varieties are shown in Table 2.
The collection and storage of mango agricultural by-products following a SOP is crucial for extract standardization. Variables as land geographic site, soil type, parts to be collected (i.e. fruit, peel, kernel, SB or BT), and procedures for collection must be considered for the adequate elaboration of a SOP in order to obtain the best collection yield. Although the SB collection yields (46%-66%) were lower than those from BT (88%-96%), PPC and MF concentrations were significantly higher in the SB extracts. Nevertheless, these values were similar in BT extracts from the Haden variety (Table 2). Mango variety may affect chemical composition of by-products extracts as it has been demonstrated i.e. for its mineral content [17]. The high PPC and MF concentrations from the Haden mango extracts should be considered as evidences on the selection of this variety for future commercial exploitation of polyphenol-rich extracts from mango by-products. However, other mango varieties should be studied to determine options for the valorization of mango agricultural by-products for the production of polyphenol-rich extracts. Antioxidant properties of mango extracts, mainly from SB, have been extensively studied, both pre-clinical and clinical [15]. Phenolic antioxidants of the Haden mango SB have been studied being MF the main component of the polyphenol fraction (7 g/100 g d. w.) [14]. We found that MF content were 13.3, and 9.4 g/100 g d.w for Tommy Atkins; 12.5 and 13.5 g/100 g d.w. for Haden SB and BT by-products, respectively, which means that MF content in the polyphenolic extracts from these varieties cultivated in Dominican Republic were almost two times higher than that previous study with Cuban mangoes. Haden accounted for 79% and 83%, and Tommy Atkins for 82% and 68% of the MF content in the PPC values for SB, and BT extracts, respectively. Probably, the lower latitude of Dominican Republic as compared to Cuba (higher sun radiation), and the difference on soil types (ferralytic in Cuba vs clay in Dominican Republic) might explain these results. A comparison of antioxidant activities of mango peel liqueurs, for the same varieties reported in the present work (Haden, and Tommy Atkins), concluded that Haden had the highest antioxidant activity, which was correlated to its higher PPC by Principal Component Analysis [19]. The chemopreventive effects of leaf extracts from Haden and Ataulfo mango varieties have been reported to be higher than Kent, Francis, and Tommy Atkins varieties on SW-480 colon carcinoma cells, and that effect was attributed to its antioxidant capacity, and the increase of apoptopic biomarkers [20]. Thus, the selection of the Haden variety as a source for obtaining polyphenol-rich extracts seems to be appropriated for further product development.
Valorization of agricultural by-products for both the obtainment of high value-added products, and the reduction of green house-gas emissions to the atmosphere has received considerable attention from the scientific and engineering community [23]. Concepts such as "biorefinery" and "circular economy" have emerged as a consequence, and many reports have been published in this field [24,25]. Our results may be considered as a contribution to the approach to apply the concept of "circular economy" in mango plantations considering the use of BT, and SB for obtaining polyphenol-rich extracts from these green raw materials. Chemical characterization of polyphenols but also other significant components, which may contribute to the observed ethnomedical and in vitro, or in vivo biological effects, is a challenge that is always facing natural products researchers. Nevertheless, the main effort on mango-derived products has been focused on the potential of mango seeds and peels, as by-products of mango fruit industry, for their use as raw material for the production of polyphenolic-rich bioactive compounds [26].
A procedure using microwave irradiation and subsequent solvent extraction yield approximately 8 g/100 g d.w. of polyphenols in seed kernel, Keitt variety [27]. A solvent extraction procedure on dehydrated kernel powder and peel (Tommy Atkins variety) yield a PPC between 2 and 15 g/100 g d.w., respectively, which means that probably the highest PPC in mango by-products for obtaining polyphenol-rich extracts would be found in the seed kernel [7]. These previous reports did not determine the MF content in the polyphenol fraction. MF yield from mango leaves, Osteen variety (more than 1 g/100 g d.w.) has been reported using supercritical fluids and sub-critical water (100°C, 40 bar) [14]. However, MF content reported for leave extracts from mango Ataulfo variety was extremely low (1.2 mg/100g), probably due to the use of acetone as the extracting solvent (Mangifera indica L.), Tommy Atkins and Haden Varieties, Cultivated in Dominican Republic [28]. Similar results were obtained when a mixture of hexane: acetone was employed for polyphenols extraction from peel and seed kernel, Tommy Atkins variety [9].
Generally, the polyphenol concentration is higher in fruit biowastes than in fruit pulp [29]; several polyphenol-rich extracts from these sources have been increasingly marketed as grape seed extract from wine production waste hydrotyrosol from olive oil factories wastes, and resveratrol from pomegranate peel waste, just to mention few examples [30][31][32]. Mango SB extract has shown higher protective effects against oxidative stress than vitamins C and E [33]. Its antioxidant profile has been reviewed, and it has been assumed that its major component (MF) accounts for this biological effect [34]. Although MF has a poor solubility in water, it may be bioavailable in human plasma, and therefore it would be possible to use on cosmetic and pharmaceutical formulations, besides its benefits as antioxidant in nutraceutic supplements [35,36]. MF has shown to have five polymorphs, depending on the extraction conditions; and the amorphous crystalline form has the highest water-solubility in acid medium [37]. The potential benefits of MF for human health have been reviewed and it has the potential as an anti-inflammatory or immunomodulatory agent in several neurodegenerative diseases [38,39].

Polyphenols by HPLC-ESI-MS
According to previous described results on polyphenolic composition, we decided to study more in depth the Haden mango SB extract, in order to complete the identification of other polyphenols, which may contribute to its observed biological effects. Figure 1 shows the chromatogram of HPLC-ESI-MS analysis of that extract. 52 peaks were recorded, from which 20 (38%) could be identified by comparison to the mass spectra of pure standard compounds when available, or by comparison of their mass spectra and fragmentation patterns to published mass spectral data.  Table 4.   Table 4 shows the results of the unambigous identification of extract components, and the structure estimation of other components not fully identified, in the butanolic extract of the Haden mango SB by HPLC-ESI-MS. Figure 2 shows the structures of the identified components.
Several gallotannins and benzophenone derivatives have been identified in mango peels, pulp, and kernel (Tommy Atkins) by HPLC/ESI/MS as galloylated maclaurin and iriflophenone glucosides [40]. Gallates, gallotannins, flavonoids, ellagic acid, xanthones, benzophenones, and maclaurin derivatives have been identified in microwaveassisted extracts from peel and seeds of 3 mango varieties (Keith, Sensation, and Gomera) cultivated in Spain [8]. The identification of 32 gallates and gallotannins, 7 hydroxybenzophenone (maclurin and iriflophenone) derivatives, 6 xanthonoids (including isomangiferin and MF derivatives); 11 phenolic acids, and 8 flavonoids (rhamnetin and quercetin derivatives) was performed by UPLC-ESI-MS in peel and pulp of Keith and Tommy Atkins mango varieties in Costa Rica [40]. Our results are similar to those previous results in terms of qualitative polyphenols composition, but this is the first report from the Haden mango stem bark extract. Tommy Atkins, Keith, and other mango varieties have been the most studied in terms of polyphenolics composition; however, the Haden mango seems to be more attractive in terms of higher PPC than other mango varieties. Figure 1. The presence of a large diversity of xanthones, flavonoids, gallic acid and its derivatives, gallotannins, together with a large amount of terpenoids and steroids in mango fruit, peel, leaves, seeds, and stem bark is a complex matter for any researcher in the attempts to correlate chemical composition to the biological activities. It is commonly accepted that the contribution of single components to biological effects of mango extracts depends on variety, cultivation soil, climate conditions, and part of the plant used as green raw material. Therefore, our results should be considered as a report from a typical variety (Haden), in clay soils on tropical conditions, using SB or BT as agricultural by-products sources from Dominican mango plantations, as the best option for the production of polyphenol-rich extracts to be used as bioactive ingredients in nutraceutic, cosmeceutic and/or pharmaceutic formulations. Fourteen MF derivatives (peaks 4, 12 to 14, 16 to 21, 23 to 25, and 36) could be identified in the butanolic extract of the Haden mango SB extract in the present work, which led to the question if MF alone or the pool of MF derivatives are the responsible of the extract's biological effects, which needs to be investigated in the future. HPLC-ESI-MS analysis of the Haden mango SB extract showed the presence of many polyphenolic components, not previously reported, which may correlate to the antioxidant, antiinflammatory, and/or analgesic effects of this extract mostly through a synergic effect of its components. Significantly, the content of (+)-catechin in the SB extract was almost twice as compared to the BT extract (Table 2), which has to be considered in future studies about the contribution of this polyphenol to the extract biological effects. Further studies about polyphenol composition of other mango varieties in different locations would add a more comprehensive approach to this topic.

Conclusions
The potential exploitation of mango agricultural byproducts from the Haden variety would be the best option for obtaining polyphenol-rich extracts from these green raw materials to be used as bioactive ingredients in nutraceutical, cosmeceutical, and/or pharmaceutical formulations. HPLC-ESI-MS analysis of the Haden mango SB extract showed the presence of many polyphenolic components, not previously reported, which may correlate to the antioxidant, antiinflammatory, and/or analgesic effects of this extract mostly through a synergistic effect of its components. The presence of 14 MF-related derivatives led to the question if MF alone or the combination of MF derivatives is the responsible of the extract's biological effects, which needs to be investigated in the future.

Supplementary Information
File with chromatographic, UV, and MS data supporting compounds identification.

Conflicts of Interest
The authors declare that they have no competing interests.