Identification of Polyphenols from Chilean Brown Seaweeds Extracts by LC-DAD-ESI-MS/MS

ABSTRACT Polyphenol profiles were characterized in extracts of three Chilean brown seaweeds, Durvillaea antarctica (Chamisso) Hariot, Lessonia spicata (Suhr) Santelices, and Macrocystis integrifolia (Bory) by liquid chromatography with mass spectrometry detection (LC-MS/MS). Phlorotannins with different degrees of polymerization were identified in D. antarctica (trimers to octamers) and L. spicata (trimers to tetramers). No signals related to phlorotannins compounds were detected in M. integrifolia. L. spicata and M. integrifolia showed a great variety of flavonoid compounds in comparison with D. antarctica, mainly identified as glycoside forms in all the extracts. The antioxidant activity of brown seaweed extracts measured by ferric reducing power (FRAP) and oxygen radical absorbance capacity (ORAC) was significantly higher in D. antarctica, followed by M. integrifolia and L. spicata, in line with the total phenolic (TP) content. However, D. antarctica and M. integrifolia showed similar activity for free radical 2,2-diphenyl-1-picrylhydrazyl (DPPH) in spite of the differences found in TP content. D. antarctica as well as L. spicata would represent a potential source of phlorotannins, whereas M. integrifolia could be considered as an alternative source of flavonoids. The identification of polyphenols in extracts of Chilean brown seaweeds opens innovative opportunities for their use in the food and pharmaceutical industries.

Over the last two decades, the extraction and characterization of compounds with biological activity for the design of healthy or functional foods have become important areas of research in food science. In this context, brown seaweeds have been reported to be a rich source of bioactive compounds associated with several biological activities in both in vitro and in vivo models. Among the bioactive compounds in brown algae, polyphenols have been found to be responsible for antioxidant, antidiabetic, anticancer, antihypertensive, and anti-inflammatory activities (Holdt and Kraan, 2011). The study of phenolic compounds in brown seaweeds has been mainly focused on the identification of phlorotannins as a class of tannins exclusively synthesized in these seaweeds. Phlorotannins are phloroglucinol-based compounds biosynthesized by the acetate-malonate or polyketide pathway, are highly hydrophilic, and have a wide range of molecular sizes (Ferreres et al., 2012). Phlorotannins have been grouped according to interphloroglucinol linkages into four primary types as follows: fucols (with only phenyl linkages), phlorethols (with only aryl ether bonds), fucophlorethols (with phenyl and aryl ether linkages), and eckols (with dibenzodioxin linkages) (Isaza Martínez and Torres Castañeda, 2013). Phlorotannin compounds such as fucophloroethol, fucodiphloroethol, fucotriphloroethol, 7-phloroeckol, phlorofucofuroeckol, and bieckol/dieckol have been reported in the brown seaweeds Pelvetia canaliculata, Fucus spiralis, Fucus vesiculosus, Ascophyllum nodosum, Saccharina longicruris, Cystoseira nodicaulis, Cystoseira tamariscifolia, and Cystoseira usneoides (Ferreres et al., 2012;Steevensz et al., 2012). Different polymerization degrees of phlorotannins have been described in Fucus vesiculosus and Sargassum muticum (Montero et al., 2016;Steevensz et al., 2012). However, studies dealing with the extraction and characterization of polyphenols in Chilean seaweeds are scarce (Leyton et al., 2016;Tala et al., 2013) but essential to give added value to these marine natural products and to explore new potential uses in the food and pharmaceutical industries.
The objective of this study was to determine the content of total phenolic compounds and antioxidant activity of extracts of three Chilean brown seaweeds species (D. antarctica, L. spicata, and M. integrifolia), as well as to identify their polyphenol profiles by high performance liquid chromatography with mass spectrometry detection (HPLC-MS/MS).

Sample collection
Durvillaea antarctica, Lessonia spicata, and Macrocystis integrifolia were collected in January 2013 from intertidal rocks on "El Tabo" beach (33°27ʹ 31" S, 71°39ʹ 43ʹ W), which is located in the Valparaiso region (Chile). Fresh seaweeds were washed with cold water spurts to remove sand, stored at room temperature for 24 h, and then dried at 40°C for 7 days in a forced-air oven (WTE, Germany). The dried seaweeds (10% moisture) were stored in sealed plastic bags at −20°C until extract preparation.

Preparation of seaweed extracts
Dry seaweeds (40 g) were ground in a windmill (Fuchs-Müllen, Masch. Kom. Nº 18791, Kriens, Switzerland), macerated in ethanol:water (1:1 v/v) (100 mL) for 24 h at room temperature and then filtered (Whatman Nº1). This procedure was carried out twice with 100 mL of ethanol:water (1:1 v/v). The extracts were combined, and the volume was reduced in a rotary evaporator (Buchi R-3, Vacuum Pump v-700, Switzerland) at 40°C to a final volume of 150 mL. Finally, the seaweed extract was centrifuged (Z206A, Hermle, Germany) at 3500 rpm for 10 min. The resulting seaweed extracts were frozen at −20°C.

Identification of polyphenols by liquid chromatography with mass spectrometry
Phenolic compounds were obtained from D. antarctica, L. spicata, and M. integrifolia extracts using liquid-liquid extraction according to Peña-Neira et al. (2007). Briefly, 25 mL of the extract was treated with ethyl ether (3 x 10 mL for each sample) and with ethyl acetate (3 x 10 mL for each sample). The organic fractions were combined and then evaporated in a rotary evaporator (Buchi R-3, Vacuum Pump v-700) at 35°C. The residue was dissolved in 2 mL ethanol:water (1:1 v/v), filtered (0.45 µm filter, Millipore), and injected into the LC-ESI-MS/MS system, which consisted of an HPLC (HP1100, Agilent Technologies Inc., Santa Clara, CA, USA) connected with a mass spectrometer (Esquire 4000 ESI-Ion Trap LC/MS(n) system, Bruker Daltoniks GmbH, Germany). A C18 column (5 µm, 4.6 mm i.d. x 25 cm, Waters) was used in the analysis. The mobile phase was formic acid in deionized water (1% v/v, solvent A) and acetonitrile (solvent B), at a flow rate of 1 mL/min, with the following elution gradient: 0-5 min, 5% B; 5-60 min, 5-30% B (linear); 60-70 min, 30-60% B (linear); 70-80 min, 60% B; and 80-90 min, 60-5% B (linear). Phenolic compounds were detected at 280 nm. The mass spectral data were acquired in positive and negative modes. Ionization was performed at 3000 V with nitrogen as nebulizing gas at 40 psi, drying gas at 365°C, and at a flow rate of 10 L/min. All the scans were performed in the range of 20-2200 m/z. The collision-induced dissociation (CID) was performed by collisions with the helium background gas present in the trap and automatically controlled through SmartFrag option. The analysis of chromatograms and mass spectra was performed using DataAnalysis 3.2 (Bruker Daltonik GmbH). Identification of phenolic compounds was performed by m/z values of the molecular ions, their fragmentation pattern, and comparison with published data.

Statistical analyses
All the statistical analyses were calculated using Statgraphics Centurion XVI (Manugistics Inc., Statistical Graphics Corp., The Plains, VA, USA).

Results and discussion
Total phenolic compounds and antioxidant activity of D. antarctica, L. spicata, and M. integrifolia extracts Table 1 shows the content of total phenolic compounds (TPC) and antioxidant activity of the D. antarctica, L. spicata, and M. integrifolia extracts. The TPC content of the extract from D. antarctica was significantly higher than those of M. integrifolia and L. spicata. Higher TP contents have been reported in other brown seaweeds (Fucus vesiculosus, Laminaria ochroleuca, Undaria pinnatifida, Dictyota dichotoma, and Sargassum polycistum), ranging from 6.6 to 109 g PGE·kg-1 DW (Jiménez-Escrig et al., 2001;Koivikko et al., 2005;Matanjun et al., 2008).
Generally, various methods have been employed to measure the antioxidant activity of extracts, which include different mechanisms and/or action modes. FRAP and DPPH are based on single electron transfer, whereas ORAC is based on hydrogen atom transfer (Shahidi and Zhong, 2015). The antioxidant activity of brown seaweed extracts measured by FRAP and ORAC (Table 1) was significantly higher in D. antarctica, followed by M. integrifolia and L. spicata, in accordance with the TP content. However, no significant differences were found between D. antarctica and M. integrifolia in the antioxidant activity measured by DPPH, in spite of the lower TPC content of the latter, and the extract with the lowest TPC content showed the lowest DPPH activity. Similarly, several studies have reported a correlation between the content of total phenolic compounds in algae and their antioxidant activity, and the higher the TPC content, the higher the antioxidant activity (López et al., 2011;Rajauria et al., 2016). However, a lack of correlation between TP content and antioxidant activity have also been reported, suggesting that other components, such as chlorophyll and carotenoids, together with differences in the polyphenols profiles may affect the antioxidant activity (Belda et al., 2016). Furthermore, the predominant mode of antioxidant activity of each polyphenol should be taken into account also, since different antioxidant mechanisms are involved in the different methods used to evaluate the antioxidant activity. The comparison of these results with the previously reported values for both TPC content and antioxidant activity of brown seaweed is a difficult task, since these parameters may be affected by several factors such as the geographical location (Montero et al., 2016), seasonal variation (Conan et al., 2004;Maréchal et al., 2004), reproductive stage (Pansch et al., 2008), genus and species (Shibata et al., 2004). Furthermore, the extracting solvent used may play a key role in both TPC content and antioxidant activity of seaweed extracts (Sabeena Farvin and Jacobsen, 2013). In this study, a mixture of ethanol:water (50:50) was used as extracting solvent, since ethanol is a permitted solvent in the food industry, and these extracts are intended to be applied as food ingredients in the design of healthy or functional foods. In previous studies, ethanol was more efficient than water for phenolic extraction in the case of several brown seaweeds (Sabeena Farvin and Jacobsen, 2013). Mixtures of acetone:water have also showed high efficiency for the extraction of TPC and antioxidant activity in D. antarctica (70:30) (Tala et al., 2013) and H. elongate (60:40) (Belda et al., 2016). However, acetone is not permitted as extracting solvent for food applications. Figure 1(a-c) shows the UV chromatograms at 280 nm for D. antarctica, L. spicata, and M. integrifolia extracts, respectively. The identification of the marked peaks is detailed in Tables 2-4, which contain the positive and negative polarity precursor m/z signals for each peak, as well as their corresponding fragmentations. The fragmentations are arranged in order of decreasing intensity from left (base peak) to the right.

Phlorotannins profile
Usually, the mass spectral data are acquired in negative or positive ionization mode in most of the studies dealing with the identification of phlorotannins by HPLC/MS-MS in brown seaweeds (Rajauria et al., 2016;Steevensz et al., 2012;Wang et al., 2012). However, both negative and positive ionization modes were considered in this study. [M-H]-ions corresponding to trimers (m/z 373) were detected in both D. antarctica (peaks 1, 2, and 3; Table 2) and L. spicata (peak 1, Table 3 Eyles et al., 2007 N ([M-18-126 + H]+, m/z 231). According to MS data and fragmentation pattern, the trimers identified would be triphloroethol y/o fucophloroethol (Ferreres et al., 2012;Wang et al., 2012). The fragmentation patterns for trimers, together with their proposed chemical structures, are shown as supplementary material (supplementary material 1 and 2). Isomers with [M-H]-at m/z 497 were also detected in both D. antarctica (Table 2) and L. spicata (Table 3), which were considered to be phlorotannins composed by four phloroglucinol (phlorotannin tetramers). The product ions detected may be due to the loss of bifuhalol (m/z 233), one phloroglucinol unit (m/z 371), two phloroglucinol units (m/z 245), and successive loss of water (m/z 189) . Tetramers were also detected in the positive ionization mode for the D. antarctica extract, with protonated molecular ions ([M + H]+) at m/z 499 (peak 9, Table 2), whose fragmentation pattern also showed the combined loss of phloroglucinol and water (m/z 355), bifuhalol (m/z 233), and successive losses of phloroglucinol (m/z 373, m/z 247/248). Loss of phloroglucinol units suggest the breaking of C-O-C bonds, which would indicate that most of the detected phlorotannins corresponded to phloroethols. Similar phenolic compounds have been reported in other brown algae such as F. spiralis, C. usneoides, or F. vesiculosus, where these tetramers were tentatively identified as tetrafucol and/or fucodiphoroethol (Ferreres et al., 2012;Wang et al., 2012). In the D. antarctica extract, other compounds with m/z in the range 621-993 in the negative ion mode and m/z 623-995 in the positive ion mode were considered polymers with several phloroglucinol units. Thus, m/z of 621/623 were tentatively identified as pentamers (pentaphoroethol/trifucophloroethol) (Ferreres et al., 2012). The fragmentation patterns for pentamers, together with their proposed chemical structures, are shown as supplementary material (supplementary material 3 and 4). Molecular ions detected at m/z 745/747, 869/871, and 993/995 may be phenolic compounds composed by six, seven, or eight phloroglucinol units as described in other brown algae such as F. vesiculosus and F. spiralis (Ferreres et al., 2012;Wang et al., 2012). The product ions detected from these polymers showed the successive loss of phloroglucinol units. The phlorotannins profile of D. antarctica extract showed a higher degree of polymerization (trimers to octamers) than L. spicata extract, where only trimers and tetramers were identified. Geographic location, in addition to genus and species, has been reported to affect the degree of polymerization of phlorotannins in brown seaweeds (Steevensz et al., 2012). In contrast, signals related to phlorotannin compounds were not detected in M. integrifolia extract at positive or negative polarity (Figure 1c, Table 4).

Other polyphenols in brown seaweeds
Although phlorotannins are the major polyphenols found in brown algae (Ferreres et al., 2012;Isaza Martínez and Torres Castañeda, 2013;Rajauria et al., 2018;Steevensz et al., 2012;Wang et al., 2012), other polyphenols such as several flavonoid derivatives and phenolic acids have been reported also (Belda et al., 2016;López et al., 2011;Rajauria et al., 2016;Sabeena Farvin and Jacobsen, 2013). Most of the phenolic compounds were detected in negative polarity mode, and for some peaks, sodium ([M+ Na]+) or potassium ([M + K]+) adducts were detected in positive polarity mode. In a few cases, the identification required further analysis of the fragmentation, as was the case for the differentiation between luteolin and kaempferol derivatives (data not shown). Here, characteristic fragments, such as m/z 199 and m/z 175, were found in the negative fragmentation of luteolin (Sánchez-Rabaneda et al., 2003 and m/z 165 and m/z 121 were found in the positive fragmentation of kaempferol (Cuyckens and Claeys, 2004;Justino et al., 2009). Analyses of the fragmentation patterns were also pivotal for distinguishing between quercetin and hesperetin (data not shown); quercetin shows characteristic fragments in negative polarity mode (m/z 179 and m/z 151) (Gates and Lopes, 2012;Yang et al., 2012), and hesperetin shows a characteristic fragment of m/z 286 in negative polarity mode (Gates and Lopes, 2012;Zhao et al., 2013).
Flavonoid derivatives such as flavonols, flavones, flavanones, flavanonols, and flavanols were detected in the extracts of the three brown algae. In the case of flavonols, aglycone or glycoside forms of myricetin, quercetin, and kaempferol were identified (Figure 1a (Parejo et al., 2004). Furthermore, isorhamnetin-O-dihexoside as well as isorhamnetin-O-glucoside and isorhamnetin-O-rutinoside were tentatively identified in the case of L. spicata and M. integrifolia extracts, respectively. Two types of flavones were tentatively identified in the three algae extracts: aglycone or glycoside forms of apigenin and glycoside forms of luteolin. Two luteolin derivatives were proposed in the negative and positive ionization modes, luteolin-O-hexoside and luteolin-O-rutinoside, which showed losses corresponding to one unit of glucose or galactose ([M-162-H]-) and one unit of rhamnoglucose ([M-308-H]-) (Parejo et al., 2004). The flavonone heperitin was found as heperitin-O-rutinoside in the three extracts, with the loss of one unit of rhamnoglucose ([M-308-H]-) (Parejo et al., 2004). Furthermore, catechin/epicatechin was tentatively identified in both D. antarctica and M. integrifolia in the negative ionization mode. The peak at m/z 285 was the result of the molecular ion fragmentation. According to literature, fragmented ions at m/z 245 (which results from losing a CO 2 group) and 205 (probably due to the loss of the A-ring) were obtained after further fragmentation (Savic et al., 2014). The flavononol taxifolin was found in both L. spicata and M. integrifolia extracts as taxifolin-O-rhamnoside, and fragmented ions m/z 303 and m/z 285 were obtained, probably due to the successive loss of one unit of rhamnose ([M-146-H]-) and water ([M-18-H]-). Other phenolic compounds such as dimeric procyanidins were tentatively identified in both D. antarctica and M. integrifolia. Furthermore, in the case of M. integrifolia, other tannins such as gallotannins (digallloylglucose and trigalloylglucose) and the ellagitannins HHDP-galloylglucose were also proposed. Phenolic acid derivatives of dicaffeoylquinic acid and caffeic acid were also detected in the case of L. spicata and M. integrifolia.

Conclusions
D. antartica is a source of phlorotannins with higher degree of polymerization (trimer to octamer) than L. spicata (trimer to tetramer), whereas phlorotannins were not detected in M. integrifolia. A greater diversity of flavonoid compounds was identified in M. integrifolia, mainly as glycoside forms in all the extracts. The identification of polyphenols profile in both negative and positive ionization mode, together with the quantification of total phenolic compounds as well as the antioxidant activity of these Chilean marine resources opens new opportunities for their use in the food and pharmaceutical industries.