Hydroethanolic extractions were performed on each fruit, A. unedo and R. idaeus, and the yield was determined. The extract yield was higher for A. unedo (79%, w/w) than for R. idaeus (47%, w/w). The total content in phenols was determined (Table 1). A. unedo extract yielded a similar content of total phenolics (16.46 ± 3.66 mg GAE. g^-1 dw) to R. idaeus (13.23 ± 0.94 mg GAE. g^-1 dw). Content in total phenols are in the range of values described in the literature for A. unedo and R. idaeus [9,18,19]. Phenolic compounds previously identified in strawberry tree fruits were mainly gallic acid derivatives, anthocyanins and proanthocyanidins, which are highly active as antioxidants [20]. Within these groups of antioxidant compounds, proanthocyanidins were the most abundant, representing more than 80% of the total flavonoid content [10]. The phenolic composition obtained for strawberry tree fruits used in this work (see supplementary data, Table 1 and Figure 1) was similar to the reported bibliography. We also detected some quercetin and ellagic acid derivatives as minor components. The raspberry extracts gave LC-MS profiles similar to previous work [14] and were dominated by anthocyanins and ellagitannins (see supplementary data, Table 3 and Table 4 and Figure 2), with a number of other minor components.

Anthocyanins are a group of phenolic compounds of great interest in nutrition and medicine because of their potent antioxidant capacity and possible protective effects on human health [21]. The total anthocyanin content of raspberry was much higher than strawberry tree fruits (Table 1). Three main anthocyanins (cyanidin-3-O-β-D-galactopyranoside, delphinidin-3-O-β-D-glucopyranoside and cyanidin-3-O-β-D-arabinopyranoside; see supplementary data, Table 2 and Figure 1C) were seen in A. unedo fruits which agrees with previous reports [12]. On the other hand, the most abundant R. idaeus anthocyanins are usually cyanidin-3-O-sophoroside, cyanidin-3-O-(2^G)-glucosylrutinoside, cyanidin-3-O-glucoside, pelargonidin-3-O-sophoroside, cyanidin-3-O-rutinoside and pelargonidin-3-O-(2^G)-glucosylrutinoside. Lower amounts of pelargonidin-3-O-glucoside and pelargonidin-3-O-rutinoside were also detected [14]. The raspberry extracts used in this study differed from reported anthocyanin compositions (e.g. [14]) with the dominant anthocyanins being cyanidin-3-O-sophoroside and cyanidin-3-O-rutinoside, probably as a result of varietal differences (see supplementary data, Table 4 and Figure 2C).

Based on their high phenol content, it was possible to reasonably anticipate a high antioxidant activity for both A. unedo and R. idaeus berries. The antioxidant capacity, assessed by ORAC, was similar for both fruits (11.66 ± 2.01 mmol TE.100 g^-1 dw for strawberry tree fruit and 15.37 ± 2.73 mmol TE. 100 g^-1 dw for raspberry). Antioxidant capacities reported in the literature are difficult to compare due to differences in the methodologies used. However, García-Alonso and co-workers determined the antioxidant activity of 27 fruits, including strawberry tree and raspberry, and established a ranking (TEAC method) where strawberry tree antioxidant capacity was ranked slightly higher than that of raspberry [22].

Both A. unedo and R. idaeus berries have been characterized as good sources of antioxidants, with their antioxidant properties associated to different groups of compounds. For A. unedo, this activity is attributed to the high flavonoid content, (mainly comprised by proanthocyanidins, cyanidin and delphinidin glycosides), ellagic acid and its diglucoside derivative, vitamin C and E and carotenoids [10]. R. idaeus studies performed with an ellagitannin-rich fraction show that this fraction had considerably higher antioxidant capacity than the original raspberry extract or an anthocyanin-rich fraction [14], a observation already noted by others [23]. Over 50% of the total raspberry antioxidant capacity is conferred by Sanguiin H6 and Lambertianin C [23].

Knowledge of the chemically-determined antioxidant characterization of A. unedo berries is certainly important. However, its biological significance remains to be evaluated. It is therefore crucial to validate these bioactivities on different cell-based human-disease models assays and, at the end, in the overall organism.

Based on promising in vitro antioxidant results, a neuroprotective test was performed using an SK-N-MC human neuroblastoma cell line. Firstly, a non-toxic range of concentrations of raspberry and strawberry tree fruit extracts were defined. Toxicity tests involved the determination of cell viability after a 4 h incubation in the range 0 to 500 μg GAE. mL^-1 medium. No toxic effects were observed in cell viability until 125 μg GAE. mL^-1 for both fruits, as shown in Figure 1. For concentrations higher than 125 μg GAE. mL^-1, the viability decreased, particularly with raspberry, probably because the compounds become toxic for cells. This phenomenon should happen since many plant secondary products seem to have a paradoxical (hormetic) effect on diseases that depend on their concentration and thus level of consumption [24]; at low concentrations a compound could exert a benefic effect, but when its concentration is too high for the organism/cells it starts to be toxic [25,26,27].

To evaluate the neuroprotective effect of the fruit extracts, the neuroblastoma cell line SK-N-MC was submitted to an oxidative stress after a pre-incubation with the dried fruit extracts solubilized in medium. Peroxides are often used as models to induce oxidative damage in cells cultured under in vitro conditions [17]. To this end, the ability of hydrogen peroxide (H₂O₂) to induce oxidative stress in SK-N-MC cells was assessed by the cell viability test already described above. Non-toxic concentrations of the fruits extracts were used in the SK-N-MC cells challenged with 1 mM H₂O₂ to evaluate the neuroprotection capacity of the phytochemicals. This test evaluates the cytoprotective activities of the compounds, by protecting cell viability after pre-conditioning the cells with the extracts before the oxidative injury.

Although the chemically-measured antioxidant activity exhibited by strawberry tree fruit was slightly lower from that of raspberry, the raspberry extract rescued cell survival while no significant effect in neuroblastoma viability was detected with the strawberry tree berry extract, using the same amount of polyphenols, estimated by Folin- Ciocaulteau method (Figure 2).

In fact, the incubation of the human cells with 1 mM H₂O₂ for 1.5 hours reduced viability to 25%, whereas pre-incubation with R. idaeus extract enhanced cell survival to 33.8%. In other words, the overall increase in cell viability with R.idaeus was estimated as 36.6% at 50 µg GAE. mL^-1 and 27.8% at 125 µg GAE. mL^-1, but decreased for higher concentrations (8.9% at 175 µg GAE. mL^-1).

Both R. idaeus and A. unedo fruits are potential sources of bioactive compounds like vitamin C, phenolic acids, proanthocyanidins and flavonol derivatives, however the absence of cell protection detected for A. unedo suggests that its phytochemical composition is not effective against H₂O₂ stress induced in this neuroblastoma cell models, but does not exclude biological effectiveness in other systems. These fruits are qualitatively quite different in their chemical composition which could explain the different bioefficacy in protecting cells from oxidative stress. In fact, the biological effects presented by R. idaeus fruits could be attributed to ellagitannin content (not present in A. unedo) and the higher levels of anthocyanins.

This study is a good example of the importance in evaluating the biological function of phenolics to validate their in vitro antioxidant activity. Further assays in different biosystems will be performed to validate the biological effects of A. unedo phytochemicals.

Arbutus unedo L. fruits were collected in November 2007, by random sampling in an extensive area of Arrábida Natural Park (southern region of Portugal). The samples were immediately ground and freeze dried and stored at -80°C prior to extraction. Rubus idaeus cv. Polka was grown in Herdade Experimental da Fataca, Odemira, Portugal.

To each 1 g of lyophilized powder, 12 mL of hydroethanolic solvent (ethanol 50% (v/v)) were added and the mixture was shaken for 30 min, at room temperature in the dark. The mixture was then centrifuged at 12,400 g during 10 min at room temperature. The supernatant was filtered through paper filter and then through 0.20 µm cellulose acetate membrane filters. The resulting extracts were stored frozen at –80 °C, no longer than one month. For each fruit extract, an aliquot was freeze dried under vacuum conditions and the yield was determined.

Determination of total phenolic compounds was performed by the Folin-Ciocalteau method [28]. Briefly, to each well of a microplate, 235 µL water, 5 µL sample (or solvent, in the control), 15 µL Folin-Ciocalteau’s reagent (Fluka®) and 45 µL saturated Na₂CO₃ were added. The microplate was incubated for 30 min at 40 °C and the absorbance at 765 nm measured. Gallic acid was used as the standard and the results were expressed in mg of gallic acid equivalents per g of dry weight of plant material (mg GAE. g^-1 dw).

The total anthocyanin content of the fruit extracts was determined using a pH differential absorbance method [18]. Absorbance readings were related to anthocyanin content using the molar extinction coefficient of 12100 calculated for cyanidin-3-O-glucoside. Results were expressed as mg of cyanidin 3-glucoside equivalents per 100 g of dry weight of plant material (mg cy-3-glucoside. 100 g^-1 dw).

Peroxyl radical scavenging capacity was determined by the ORAC (Oxygen Radical Absorbance Capacity) method [19,29]. Briefly, the reaction mixture contained 150 µL of sodium fluorescein (0.2 nM) (Uranine, Fluorescein Sodium Salt® TCI Europe), 25 µL sample and 25 µL of 2,2’-azobis(2-amidopropane)dihydrochloride (41.4 g.L^-1). The blank contained 25 µL 75 mM phosphate buffer (pH 7.4) instaed of sample, whereas the standards contained 25 µL of 10 to 50 µM 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox®) in place of the sample. The fluorescent emission at 515 nm was then monitored kinetically during 30 min at 37 °C, after excitation at 493 nm using a FLx800 Fluorescence Microplate Reader (Biotek). The final results were calculated using the area differences under the fluorescence decay curves between the blank and the sample, and were expressed as mM Trolox equivalents per g of dry weight of plant material (mM TE. 100 g^-1 dw).

Phenolic extracts were dried by rotary evaporation, ressuspended in 5% (v/v) acetonitrile in water and was analyzed on a LCQ-DECA system controlled by the XCALIBUR software (2.0, ThermoFinnigan). The LCQ-Deca system comprised a Surveyor autosampler, pump and photo diode array detector (PDAD) and a Thermo Finnigan mass spectrometer iontrap. The PDA collected spectral data from 200-600 nm and scanned three discrete channels (at 280, 365 and 510 nm). The samples were applied to a C-18 column (Synergi Hydro C18 column with polar end capping, 4.6 mm x 150 mm, Phenomonex Ltd.) and eluted over a gradient of 95:5 solvent A:B at time=0 minutes to 60:40 A:B at time = 60 minutes at a flow rate of 400 μL/min. Solvent A was 0.1% (v/v) formic acid in ultra pure water and solvent B 0.1% (v/v) formic acid in acetonitrile. The LCQ-Deca LC-MS was fitted with an ESI (electrospray ionization) interface and analyzed the samples in positive and negative-ion mode. Two scan events, full scan analysis in mass range 80-2000 m/z followed by data dependent MS/MS of the most intense ions, were used for compounds detection and identification. The data-dependent MS/MS used collision energies (source voltage) of 45%. The capillary temperature was set at 275 °C with sheath gas at 60 psi and auxiliary gas at 10 psi. Before the analysis, the system was tuned by using known concentrations of cyanidin-3-glucoside (positive mode) and quercetin-3-glucoside (negative mode) in ultrapure water.

Human neuroblastoma SK-N-MC cells were obtained from the European Collection of Cell Cultures (ECACC) and cultured in EMEM supplemented with 2 mM Glutamine, 10% (v/v) heat-inactivated fetal bovine serum (Gibco), 1% (w/v) of non-essential amino acids and sodium pyruvate (1mM). The cells were maintained at 37 °C in 5% CO₂. All experiments were carried out 24-48 h after cells were seeded.

Hydroethanolic extracts were prepared as described above, dried under vacuum and solubilized in cell medium for the cytotoxicity tests. The test was performed by a 96-well plate cell viability assay on the neuroblastoma human cell line SK-N-MC, to identify the nontoxic range of extract concentrations. Toxicity tests involved a 4 h incubation in the range 0 to 500 μg GAE. mL^-1 medium. Cell viability was assessed using the CellTiter-Blue^® Cell Viability Assay, a fluorescent method based on the ability of living cells to convert a non-fluorescent redox dye (resazurin) into a fluorescent end product (resorufin), according to the manufacture instructions. Nonviable cells rapidly lose their metabolic capacity and thus do not generate the fluorescent signal.

Neuroblastoma cells were pre-incubated for 1 h with strawberry tree or raspberry fruit extracts and then were treated with H₂O₂ (1 mM) for 1.5 h. Fruit hydroethanolic extracts were prepared as described above, dried under vacuum and dissolved in cell medium. Cell viability was evaluated as described above.

The results reported in this work are the averages of at least three independent experiments and are represented as the mean ± SD. Differences among treatments were detected by analysis of variance [30] with Tukey HSD (Honest Significant Difference) multiple comparison test (α = 0.05) using SigmaStat 3.10 (Systat).