Scholarly article on topic 'The chemical composition on fingerprint of Glandora diffusa and its biological properties'

The chemical composition on fingerprint of Glandora diffusa and its biological properties Academic research paper on "Chemical sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Arabian Journal of Chemistry
OECD Field of science
Keywords
{" Glandora diffusa (Lag.) D.C. Thomas" / "Metabolic profile" / Cholinesterase / Cytotoxicity}

Abstract of research paper on Chemical sciences, author of scientific article — Fátima Fernandes, Paula B. Andrade, Federico Ferreres, Angel Gil-Izquierdo, Isabel Sousa-Pinto, et al.

Abstract Glandora diffusa (Lag.) D.C. Thomas is a medicinal species widely consumed as herbal tea. Despite being commercialized by several herbs distributors, the genuineness of the marketed product is unknown. Among secondary metabolites with proven taxonomic interest, the phenolics profile was herein used as “fingerprint” of three commercial G. diffusa samples. Furthermore, the knowledge on the composition of this species was extended and its amino acids, fatty acids, sterols and triterpenes profiles were studied for the first time. The phenolics profile was characterized by HPLC-DAD. All other metabolites were determined by GC–MS. Despite similar qualitative profiles, significant quantitative differences were observed among the three samples. Their potential as antioxidant and anti-Alzheimer and cytotoxicity was evaluated and relationship between chemical composition and activities was considered. Ethanolic extracts showed a potent dose-dependent response against DPPH , a mild inhibitory effect on both acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) and to be toxic to both human colorectal (Caco-2) and gastric adenocarcinoma (AGS) cells only for high concentrations. This work covers interdisciplinary aspects in the pharmaceutical and biomedical sciences, by focusing on metabolic profiling and quality assurance of a medicinal species used in folk medicine. The results improve the knowledge on G. diffusa and encourage its use, not only as herbal tea, as it is traditionally consumed, but also in pharmaceutical preparations, ethanol being a cheap and feasible solvent to recover its bioactive components.

Academic research paper on topic "The chemical composition on fingerprint of Glandora diffusa and its biological properties"

Accepted Manuscript

The chemical composition on fingerprint of Glandora diffusa and its biological properties

Arabian Journal of Chemistry

PII: DOI:

Reference:

Fatima Fernandes, Paula B. Andrade, Federico Ferreres, Angel Gil-Izquierdo, Isabel Sousa-Pinto, Patricia Valentâo

S1878-5352(15)00031-3 http://dx.doi.org/10.1016/j.arabjc.2015.01.012 ARABJC 1567

To appear in:

Arabian Journal of Chemistry

Received Date: 26 October 2014

Accepted Date: 25 January 2015

Please cite this article as: F. Fernandes, P.B. Andrade, F. Ferreres, A. Gil-Izquierdo, I. Sousa-Pinto, P. Valentâo, The chemical composition on fingerprint of Glandora diffusa and its biological properties, Arabian Journal of Chemistry (2015), doi: http://dx.doi.org/10.1016/j.arabjc.2015.01.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The chemical composition on fingerprint of Glandora diffusa and its biological properties

Fatima Fernandesa, Paula B. Andradea, Federico Ferreresb, Angel Gil-Izquierdob,

Isabel Sousa-Pintoc d, Patricia Valentao3'* Short title: Glandora diffusa fingerprinting and bioactivity

• V^

a REQUIMTE/LAQV, Laboratorio de Farmacognosia, Departamento de Química, Faculdade de Farmácia, Universidade do Porto, R. Jorge Viterbo Ferreira, n° 228, 4050-313 Porto, Portugal b Research Group on Quality, Safety and Bioactivity of Plant Foods, Department of Food Science and Technology, CEBAS (CSIC), P.O. Box 164, 30100 Campus University Espinardo, Murcia, Spain

c Interdisciplinary Centre for Marine and Environmental Research (CIIMAR/CIMAR), Rua dos

Bragas n° 289, 4050-123 Porto, Portugal d Faculty of Sciences, University of Porto (FCUP), Rua do Campo Alegre s/n, 4169-007 Porto,

Portugal

?s, Umvt

* Corresponding author. Tel.: +351 220428653; fax: +351 226093390. E-mail address: valentao@ff.up.pt (P. Valentao).

Abstract Glandora diffusa (Lag.) D. C. Thomas is a medicinal species widely consumed as herbal tea. Despite being commercialized by several herbs distributors, the genuineness of the marketed product is unknown. Among secondary metabolites with proven taxonomic interest, the phenolics profile was herein used as "fingerprint" of three commercial G. diffusa samples. Furthermore, the knowledge on the composition of this species was extended and its amino acids, fatty acids, sterols and triterpenes profiles were studied for the first time. The phenolics profile was characterized by HPLC-DAD. All other metabolites were determined by GC-MS. Despite similar qualitative profiles, significant quantitative differences were observed among the three s amples. Their potential as antioxidant and anti-Alzheimer's and cytotoxicity was evaluated and relationship between chemical composition and activities was considered. Ethanolic extracts showed a potent dose-dependent response against DPPH', a mild inhibitory effect on both acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) and to be toxic to both human colorectal (Caco-2) and gastric adenocarcinoma (AGS) cells only for high concentrations. This work covers interdisciplinary aspects in the pharmaceutical and biomedical sciences, by focusing on metabolic profiling and quality assurance of a medicinal species used in folk medicine. The results improve the knowledge on G. diffusa and encourage its use, not only as herbal tea, as it is traditionally consumed, but also in pharmaceutical preparations, ethanol being a cheap and feasible solvent to recover its bioactive components.

KEYWOI

Glandora diffusa (Lag.) D. C. Thomas; Metabolic profile; Cholinesterase; Cytotoxicity

1. Introduction

Medicinal plants have been used for centuries throughout the world and many people still rely on indigenous medicinal plants for their safe or primary health care needs (Daur, 2013). The reasons for the application of specific medicinal plants in the treatment of certain diseases are being discovered and the use of those species gradually abandons the empiric framework and becomes reasoned on explicatory facts (Petrovska, 2012). Nowadays, the potential of a panoply of phytochemicals, like phenolic compounds, fatty acids, sterols, triterpenes and glucosinolates, in the prevention or treatment of several pathologies, such as cardiovascular diseases, diabetes, neurodegenerative disorders and cancer, is known (Burda and Oleszek, 2001; Fan et al., 2010; Ferreres et al., 2014; Hooper and Cassidy, 2006; Lléo, 2007; Ramadan and El-Shamy; 2013Surh et al., 2003). In fact, the plant kingdom is an inexhaustible source of health-promoters metabolites (El-Chaghaby, 2014). In the recent years, the quest for natural food additives has become an increasing concern. Consumers' demand for healthier foods has been the initiative for many researchers seeking for natural alternatives (El-Chaghaby, 2014). Several herbs with recognized therapeutic applications and used as medicinal plants are often undervalued by unawareness of their metabolic profile, as they are often important sources of bioactive compounds.

Glandora diffusa (Lag.) D. C. Thomas (Boraginaceae) (synonyms: Lithospermum diffusum Lag. and Lithodora diffusa (Lag.) I. M. Johnst.) is one of the six species of the genus Glandora, being commonly known as "scrambling-gromwell". This herbal species is a medicinal plant spontaneously found in the Mediterranean area (Ferreres et al., 2013; Thomas et al., 2008). Its pollen is present in some honeys and the herbal tea, as it is traditionally consumed, is used as diuretic, depurative and antihypertensive (Sá-Otero et al., 2006). These properties sparked the attention of researchers and pharmacological studies have been conducted in order to explore other potentialities of G. diffusa. In a recent work, Ferreres and collaborators reported the presence of high amounts of several kinds of phenolics in the aqueous extract, its good antiradical activity and strong ability to inhibit a-glucosidase, an enzyme related with diabetes mellitus (Ferreres et al., 2013). Nevertheless, the interest

in these metabolites can go further. As secondary metabolism products, phenolic compounds are specific of certain families, constituting important chemotaxonomic markers. According to this, phenolic profiles can be successfully used in the determination of authenticity of different products from vegetal origin (Ferreres et al., 2014).

In Portugal, G. diffusa aerial parts are widely commercialized by several plants distributors, being sold in herbal shops. The marketed product (the one that is consumed) corresponds to the mixture of all vegetal tissues above soil. Nevertheless, there are no studies about the authenticity of what is sold. As far as we are aware, the only study available is related with the characterization of its phenolics and was performed by our group (Ferreres et al., 2013). Thus, the aim of this work was to use phenolics profile as "fingerprint" for authenticity control of commercialized G. diffusa material

commercial samples of G. diffusa were analysed regarding primary and secondary metabolites. Since

and to improve the knowledge on its metabolic composition. For this purpose, three distinct

arding prima

the chemical profile can greatly influence the biological effects (Milella et al., 2011), the samples were also tested as inhibitor of acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE), enzymes associated with the aetiology of Alzheimer's disease (AD) (Lléo, 2007), for antiradical capacity, since reactive species are in the origin of several pathologies, and for cytotoxicity on gastric and intestinal carcinoma cells lines, as gastrointestinal cancers rank second in overall cancer-related deaths (Surh, 2003).

The comparison of the samples is interesting because the differences observed will be representative of the procedure followed by each supplier, from production to selection, transformation and conservation of the vegetal materials. This information may be important for consumers' choice.

2. Materials and methods

2.1. Standards and reagents

Quercetin-3-O-rutinoside, kaempferol-3-O-rutinoside, isorhamnetin-3-O-glucoside and rosmarinic acid

were from Extrasynthese (Genay, France). Caffeic and p-coumaric acids, alanine, glycine, valine,

isoleucine, serine, threonine, irans-4-hydroxyproline, norvaline, pelargonic acid, capric acid, lauric acid, myristic acid, palmitic acid, margaric acid, linoleic acid, a-linolenic acid, oleic acid, stearic acid, docosahexaenoic acid (DHA), methyl linolelaidate, as well as lanosterol, cholesterol, campesterol, betulin, stigmasterol, ^-sitosterol, lupeol, lupeol acetate, desmosterol, N-methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA), 1,1-diphenyl-2-picrylhydrazyl radical (DPPH'), ^-nicotinamide adenine dinucleotide reduced form (NADH), sodium pyruvate, AChE from electric eel (type Vl-s, lyophilized powder), acetylthiocholine iodide (ATCI), BuChE from equine serum (lyophilized powder), S-butyrylthiocholine chloride (BTCC), 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB), dimethyl sulfoxide (DMSO), triton X-100 and (4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) were obtained from Sigma-Aldrich (St. Louis, MO, USA). The n-alkane series (C8-C40) was purchased from

Supelco (Bellefonte, PA, USA). Methanol, potassium dihydrogen phosphate, acetonitrile and

rom Merck ( ell cultur

phosphoric, formic and acetic acids were obtained from Merck (Darmstadt, Germany) and sulphuric acid from Pronalab (Lisboa, Portugal). Reagents for cell culture were obtained from Invitrogen (Gibco Laboratories; Lenexa, KS): Dulbecco's modified Eagle's medium (4.5 g/L glucose, with l-glutamine and pyruvate; DMEM), phosphate-buffered saline (PBS), hepes buffered saline (HBS), non-essential amino acids (NEAA), foetal bovine serum (FBS), antibiotic (10,000 U/mL penicillin, 10,000 ^g/mL streptomycin), fungizone (250 ^g/mL amphotericin B), human transferrin (4 mg/mL) and trypsin-EDTA.

The water was treated in a Milli-Q water purification system (Millipore, Bedford, MA, USA).

zone (2

........

The drie

2.2. Plant material

dried aerial parts of G. diffusa were purchased in the local market, from three different medicinal plants distributors. To try to minimize differences related with growth stage and harvesting time, all samples were purchased in the same period ( September, 2012). Voucher specimens were deposited at Laboratorio de Farmacognosia, Faculdade de Farmácia, Universidade do Porto, under the

identification GD-AC-010113 (sample A), GD-MN-010113 (sample B) and GD-MC-010113 (sample C).

The plant material was powdered in an appliance mill (model A327R1, Moulinex, Spain) and the mean particle size was lower than 910 pm.

2.3. Extracts preparation

Extracts for phenolics characterization and biological assays were prepared as follows: each sample (ca. 1 g) was thoroughly mixed with ethanol (20 mL) for 30 min at 40°C, under magnetic stirring (300 rpm), and then filtered through a Büchner funnel. To avoid differences of samples treatment the extraction procedure was performed simultaneously, in an agitation plate with six stirring vacancies, temperature and speed control. The extracts were concentrated to dryness under reduced pressure (40°C) and stored at -20°C protected from light.

2.4. HPLC-DAD phenolic compounds analysis

ds analysi

redissolve

The dried ethanolic extracts were redissolved in methanol and filtered through a 0.45 pm size pore membrane before HPLC-DAD analysis.

Chromatographic analyses were performed as previously described (Ferreres et al., 2013), using a Gilson HPLC-DAD unit with a Luna C18 column (250 x 4.6 mm, 5 pm particle size; Phenomenex, Macclesfield, UK). Elution was performed with water (1% formic acid) (A) and methanol (B), starting with 30% B and using a gradient to obtain 50% B at 50 min and 80% B at 60 min. The flow rate was 0.8 mL/min and the injection volume 20 pL.

Detection was achieved with a Gilson diode array detector. Spectral data from all peaks were collected in the range of 200-400 nm and chromatograms were recorded at 330 nm. The data were processed on Unipoint System software (Gilson Medical Electronics, Villiers le Bel, France). The different phenolic compounds were identified by comparing their elution order and UV-vis spectra

with authentic standards and with data previously obtained by our group, using the same analytical conditions (Ferreres et al., 2013). Peak purity was checked by the software contrast facilities.

Quantification was performed by external standard method, at 330 nm. Since standards of all identified compounds are not commercially available, their quantification was achieved as follows: compounds 2, 6, 7, 11, 13-18, 20 and 22-25 were quantified as rosmarinic acid, compounds 4, 5 and 12 as quercetin-3-O-rutinoside, compounds 8-10 and 19 as kaempferol-3-O-rutinoside and compound 21 as isorhamnetin-3-O-rutinoside. The other compounds were quantified as themselves.

„ .... __

2.5.1. GC-MS apparatus

Analysis was performed with a Varian CP-3800 gas chromatographer coupled to a Varian Saturn 4000 mass selective ion trap detector (USA), a Saturn GC-MS workstation software version 6.8 and a VF-5 ms (30 m x 0.25 mm x 0.25 ^m) column (VARIAN). A CombiPAL autosampler (Varian, Palo Alto, CA) was used in all experiments. The injector port was heated to 250°C. Injections were performed in split mode, with a ratio of 1/40. The carrier gas was helium C-60 (Gasin, Portugal), at a constant flow of 1 mL/min. The ion trap detector was set as follows: transfer line, manifold and trap temperatures

were 280, 50 and 180°C, respectively. The mass ranged from 50 to 600 m/z, with a scan rate of 6 scan/s. The emission current was 50 ^A and the electron multiplier was set in relative mode to auto tune procedure. The maximum ionization time was 25,000 ^s, with an ionization storage level of 35 m/z. The injection volume was 2 ^L.

2.5.2. GC-MS conditions for trimethylsilyl (TMS) derivatives analysis

Stock solutions of each amino acid, fatty acid and sterol, as well as of the internal standards (IS) norvaline, methyl linolelaidate and desmosterol were prepared in ethanol and kept at -20°C until analysis.

TMS derivatives analysis by GC-MS was performed as previously reported (Pereira et al., 2012). The oven temperature was set at 100°C for 1 min, then increasing 20°C/min to 250°C and held for 2 min, 10°C/min to 300°C and held for 10 min. All mass spectra were acquired in electron impact (EI) mode. Ionization was maintained off during the first 3 min to avoid solvent overloading. Identification of compounds was achieved by comparison of their retention time and mass spectra with those of pure TMS standards derivatives, prepared and injected under the same conditions, and of NIST 05 MS Library Database. In addition, the retention index (RI) was experimentally calculated and the values were compared with those reported in NIST WebBook and in the literature (Andrade et al., 2013; Pereira et al., 2012) for GC columns with 5%-phenyl-95%-dimethylpolysiloxane. For the RI determination, an n-alkanes series (C8-C40) was used.

2.5.3. Derivatization

For GC-MS analysis ethanolic extracts were prepared according to the above mentioned conditions, but with addition of norvaline, methyl linolelaidate and desmosterol as internal standards, at final concentrations of 2.4, 20 and 16 pg/mL, respectively. The extracts were then filtered through a 0.45 pm membrane (Millipore). An aliquot of 500 pL of extract was transferred to a glass vial, the solvent was evaporated under nitrogen stream and 100 pL of the derivatizing reagent (MSTFA) was added to the dried residue. The vial was capped, vortexed and heated for 30 min in a dry block heater maintained at 60

2.5.4. Quantification

For quantification purposes, each sample was injected in triplicate and the amount of metabolites was determined by using the calibration curves of the respective TMS derivatives previously described (Pereira et al., 2012).

All compounds were quantified in Full Scan mode, with the exception of linoleic (m/z 262, 337 and 352), a-linolenic (m/z 191, 335, and 350) and oleic (m/z 264, 339 and 354) acids derivatives that

were quantified by the area obtained from the reprocessed chromatogram, using the characteristic m/z fragments.

2.6. Antiradical activity

DPPH' scavenging was determined spectrophotometrically at 515 nm on a Multiskan Ascent plate reader (Thermo, Electron Corporation), by monitoring the disappearance of DPPH', as before (Andrade et al., 2013). The reaction mixtures in the sample wells consisted of dried extract dissolved in methanol and DPPH' methanolic solution. The plate was incubated for 30 min at room temperature after addition of DPPH', the absorbance being then measured at 515 nm. The scavenging activity (SA) was calculated as percentage of DPPH' discolouration using the following equation: % SA = 100 * (1 -As/Ac), where As is the absorbance of the resulting solution when the extract has been added at a particular concentration, and Ac is the absorbance of the DPPH' solution. The extract concentration providing 50% inhibition (IC50) was calculated from the graph of SA percentage vs extract concentration.

2.7. Cell-free assay

2.7.1. Buffers

The following buffers were used. Buffer A: 50 mM Tris-HCl, pH 8; buffer B: 50 mM Tris-HCl, pH 8, containing 0.1% bovine serum albumin (BSA); buffer C: 50 mM Tris-HCl, pH 8, containing 0.1 M NaCl and 0.02 M MgCl.6H2O.

2.7.2. Enzymes

AChE and BuChE were dissolved in buffer A to make 1000 U/mL (AChE ) and 25 U/mL (BuChE) stock solutions and further diluted with buffer B to get 0.44 U/mL enzyme (AChE ) and 0.1 U/mL (BuChE) for the microplate assay.

2.7.2. AChE and BuChE inhibitory activity

The inhibition of AChE and BuChE was assessed following the method described by Rhee et al. (2001), with modifications. AChE and BuChE activities were measured using a 96-well Multiskan Ascent microplate reader (Thermo, Electron Corporation), based on Ellman's method. In each well the mixture consisted of 25 pL of 15 mM ATCI/ BTCCI in water, 125 pL of 3 mM DTNB in buffer C, 50 pL of buffer B and 25 pL of dried ethanolic extract dissolved in a solution of 10% methanol in buffer A and the absorbance was measured at 405 nm every 13 s for five times. After this step 25 pL of AChE (0.44 U/mL) or BuChE (25 U/mL) were added and the absorbance was read again every 13 s for eight times. The rates of reactions were calculated by Ascent Software version 2.6 (Thermo Labsystems Oy). The rate of the reaction before adding the enzymes was subtracted from that obtained after enzyme addition in order to correct eventual spontaneous hydrolysis of substrate. The inhibitory activity (IA) was calculated as a percentage of the 5-thio-2-nitrobenzoate produced using the equation: % IA = 100 * (1 - As/Ac), where As corresponds to the rate of reaction when the extract has been added at a particular concentration, and Ac is the rate of reaction without sample (control).

2.8. Cellular assays

nd Ac is £

2.8.1. Cell culture conditions and treatments

Human colorectal adenocarcinoma cell line (Caco-2) from the American Type Culture Collection (LGC Standards S.L.U., Spain) was routinely cultured using DMEM supplemented with 10% FBS, 1% non-essential amino acids, 1% penicillin (5,000 U/mL), 1% fungizone and 6 pg/mL transferrin. Human gastric adenocarcinoma cells line (AGS) from the American Type Culture Collection (LGC Standards S.L.U., Spain) were grown in DMEM supplemented with 10% FBS, 2% of penicillin/streptomycin and 1% of NEAA. Both cell lines were cultured in an incubator at 37°C, with a humidified atmosphere of 95% air and 5% CO2. When cells reached 80-85% confluence they were washed with HBS, trypsinized and sub-cultured in 96-wells plates at a density of 22,500 cells/cm (for AGS cells) and 21,000 cells/cm2 (for Caco-2 cells).

The dried ethanolic extracts of G. diffusa were dissolved and diluted in medium containing 0.5% (v/v) DMSO. The final concentration of DMSO did not affect cellular viability. To determine the effect of the extracts on AGS and Caco-2 cells, viability was assessed 24 h after exposure, by MTT reduction and lactate dehydrogenase (LDH) assays.

2.8.2. Cell effects

2.8.2.1. MTT reduction assay MTT reduction to formazan was measured as described before (Sousa et al., 2012). Briefly, after cells treatment, the medium was removed and the cells were incubated for 30 min, at 37°C, with culture medium containing 0.5 mg/mL MTT. Afterwards, the solution was removed and formazan crystals

were solubilized with 200 ^L DMSO. The absorbance of the resulting purple solution was measured

d as the perc r independent assa

spectrophotometrically at 570 nm. Data are presented as the percentage of MTT reduction of treated cells relative to control (untreated ones). Four independent assays were conducted, each one of them in quadruplicate.

2.8.2.2. LDH leakage assay LDH activity was measured as previously described (Sousa et al., 2012). Briefly, the release of the cytosolic enzyme LDH into the culture medium was evaluated as follows: after 24 h exposure, an aliquot of the culture medium was taken and mixed with NADH and pyruvate buffered solution. LDH activity was evaluated spectrophotometrically by following the oxidation of NADH to NAD+, at 340 nm. Results are expressed in percentage of treated cells relative to untreated ones from three independent experiments. Four independent assays were conducted.

2.9. Statistical analysis

All statistical calculations were performed with the GraphPad Prism version 6.00 for Windows (GraphPad Software, San Diego California, USA). One-way analysis of variance (ANOVA), using the

Turkey's multiple comparison test, was carried out on data obtained for chemical analysis. One-way analysis of variance (ANOVA), using Dunnett's comparison with the control test, was carried out on data obtained for biological analysis. In all cases, p values lower than 0.05 were considered statistically significant.

3. Results and discussion

3.1. Phenolic compounds

As referred above, phenolic compounds are secondary metabolites with proven taxonomic interest. As so, in order to confirm the authenticity of three commercial G. diffusa samples, their phenolic profile was determined by HPLC-DAD and compared with the previous one established for this species by HPLC-DAD-ESI/MSn (Ferreres et al., 2013). In general, using the same analytical conditions, similar chromatograms were obtained, allowing the identification of the compounds (Fig. 1).

The qualitative composition of the three analysed samples revealed to be similar (Fig. 1, Table 1), comprising twenty-five compounds from three classes: phenolic acid monomers (compounds 1 and 3), polymers of caffeic acid, namely dimers (compounds 15, 17 and 23), trimers (compounds 2, 7, 11, 13, 14, 16, 20 and 25), tetramers (compounds 6, 18, 24) and a pentamer (compound 22), and flavonoids (compounds 4, 5, 8-10, 12, 19 and 21). Compounds 13 and 14 were found only in sample C. This qualitative profile is identical to that of an aqueous one previously described (Ferreres et al., 2013), except for lithospermic acid isomer and salvianolic acid E, which were not found herein. This compositional similarity (around 93%) allows confirming the authenticity of the three commercial G. diffusa samples.

Concerning to the quantitative composition, the three samples revealed some differences (Table 1). As some compounds were coeluting, the pairs 8+9, 13+14 and 23+24 were not quantified. Thus, although they are indicated in the chromatogram (Fig. 1) to provide a more complete fingerprint, their quantification was not included in Table 1, which contains only the quantification results of well separated compounds.

Sample C showed the highest phenolics content (ca. 1008 mg/kg dried sample), which was significantly higher than that of samples A and B (p < 0.001) (Table 1).

Caffeic acid polymers were the main compounds in all samples, being responsible for ca. 58, 72 and 81% of the phenolics amounts in samples A, B and C, respectively, while the flavonoids fraction accounted for ca. 41, 27 and 18% of the determined phenolic compounds, respectively (Ta Caffeic and p-coumaric acids represented only ca. 1% of the total phenolics in the three sa

Individually, with the exception of compound 3, which was found in trace amounts, the content of each phenolic compound on the three G. diffusa samples was also significantly different (Table 1). Except salvianolic acid I (7) and compounds 11 and 20, which content was significantly higher in sample B, and for compound 22, quantified only in sample A, the amount of the several phenolic acids was significantly higher in sample C.

Rosmarinic acid (15), salvianolic acid B (18) and the flavonoid quercetin-3-0-(6-rhamnosyl)glucoside (12) were the major compounds in the three G. diffusa samples. Rosmarinic acid (15) content was around 4 and 2 times higher in sample C than in samples A and B, respectively. Rosmarinic acid (15) and salvianolic acid B (18) are reported to be common constituents of different species of Boraginaceae family (Ferreres et al., 2013; Petersen and Simmonds, 2003). Rosmarinic acid has been previously considered as a preformed defence compound that is constitutively accumulated (Park et al., 2008). Salvianolic acid B, also known as lithospermic acid, has been involved in Salvia miltiorrhiza Bunge cultures defence mechanisms, as it is accumulated by this species after its exposure to different elicitors (Dreger et al., 2010).

Quercetin-3-0-(6-rhamnosyl)glucoside (12) was the major flavonoid in the three samples, representing ca. 19% of total compounds in sample A and 12 and 11% in samples B and C, respectively. Its content was also significantly higher in sample C (p < 0.001) (Table 1). All samples presented considerable amounts of the others quercetin derivatives (compounds 4 and 5), corresponding to ca. 25, 19 and 21% of total of flavonoids in samples A, B and C, respectively (Table

1). In absolute terms, sample B was richer in flavonoids (ca. 185.8 mg/kg, dried sample) than samples C (ca. 181.8 mg/kg, dried sample) and A (ca. 175.6 mg/kg, dried sample) (Table 1).

Besides the referred major flavonoids, the richness of flavonoids of sample B is mainly due to its kaempferol-3-O-(6-rhamnosyl)hexoside (19) and isorhamnetin-3-O-(6-rhamnosyl)hexoside (21) contents, which are significantly higher than those found in samples A and C (Table 1).

Quantitative differences were also observed when comparing the three ethanolic extracts studied herein and the aqueous extract previously reported (Ferreres et al., 2013). As the samples were purchased in the local market from three different medicinal plants distributors, no information about several factors that can influence their qualitative, as well as quantitative, biochemical composition was available. Taking into account the lack of knowledge about the origin of the samples, the observed differences may be related to their growth conditions. It is well known that the development of a plant in different places can reflect a series of variations, particularly climatic conditions, soil characteristics, humidity and pests, among others, which will certainly lead to a distinct chemical composition. Furthermore, also the physiological stage cannot be ignored. In fact, the maturation stage of the plant is an important factor for its chemical composition (Lee et al., 2010). Additionally, procedures like harvesting, storage and drying can also lead to quantitative differences (FAO, 1981).

Despite the presence of rosmarinic and salvianolic acids in different genus of Boraginaceae, these compounds cannot be proposed as chemical markers for G. diffusa since they are common constituents of diverse species of Salvia genus (Lu and Foo, 2002). On the other hand, the occurrence of highly glycosylated quercetin and kaempferol derivatives is common in plant kingdom (Pereira et al., 2010).

3.2. Other metabolites

The ethanolic extraction combined with a gas chromatography-mass spectrometry analysis, constitutes a multi-target method for the simultaneous analysis of several classes of metabolites in natural matrices, such as amino acids, fatty acids, sterols and triterpenes (Pereira et al., 2012).

triterpe

To extend the knowledge on G. diffusa, amino acids, fatty acids, sterols and triterpenes composition of ethanolic extracts was studied by GC-MS. As far as we know, there are no reports about these compounds in G. diffusa.

A representative chromatographic profile of G. diffusa is shown in Fig. 2. The derivatization method allowed determining eight amino acids, fifteen fatty acids, four sterols and three triterpenes the three samples of G. diffusa (Table 2).

In a general way, except the amino acids composition, the different samples showed a similar chemical profile (Fig. 2, Table 2). Sample A presented the highest variety and sample B was the one with the poorest qualitative profile.

Glycine (3) was found only in sample A. Isoleucine (6), serine (7), threonine (9) and trans-4-hydroxyproline (12) were not detected in sample B and homocysteine (10) was absent in sample C. As can be observed in Table 2, the amino acids content was significantly different among the three samples (p < 0.001), varying from 46.5 to 1595.5 mg/kg of dried G. diffusa (Table 2). Their content was 34.3 and 3.29 times higher in sample A than in samples B and C, respectively (Table 2).

Individually, glycine (3) was clearly the major amino acid of sample A, representing ca. 47% of its total amino acids content. Alanine (2) and trans-4-hydroxyproline (12) were also important compounds in samples A and C, their content being significantly higher in the first (p < 0.001) (Table 2).

In young plants, amino acids biosynthesis is regulated by a metabolic network that links nitrogen assimilation with carbon metabolism. This network is strongly controlled by the metabolism of four central amino acids, namely glutamine, glutamate, aspartate, and asparagine, which are then converted into all of the other amino acids by various biochemical processes. They also serve as major molecules of nitrogen transport, including transport from vegetative to reproductive tissues. Its metabolism is subjected to a concerted regulation by physiological, developmental and hormonal signals (Galili et al., 2008). Thus, amino acids are important to life and also have many functions in plant metabolism (Galili et al., 2008). Besides constituting building blocks of proteins, these compounds also participate

acting as s

in many metabolic networks that control growth and adaptation to the environment. They are also involved in secondary plant metabolism, namely in the biosynthesis of several classes of compounds, such as glucosinolates, alkaloids, cyanogenic heterosides and phenolic compounds (Galili et al., 2008).

Fatty acids (FA) constituted the most represented class, both in diversity and quantity (Fig. 2, Table 2). Eleven saturated fatty acids (caproic (1), enanthic (4), pelargonic (8), capric (11), lauric (13), myristic (14), palmitic (16), margaric (17), stearic (21), docosanoic (22) and lignoceric (23)), two monounsaturated fatty acids (MUFA) (palmitelaidic (15) and oleic (20) acids) and two polyunsaturated fatty acids (PUFA) (linoleic (18) and a-linolenic (19) acids) were determined (Table 2). Linoleic (18), a-linolenic (19) and oleic acids (20) were already reported in several genus of Boraginaceae family (Yunusova et al., 2012). FA corresponded to 48, 75 and 82% of the non-phenolic metabolites of samples A, B and C, respectively (Table 2), significant differences among samples being observed (Table 2). The highest fatty acids content was found in sample C (ca. 5677.5 mg/kg, dried sample) (Table 2).

Fatty acids are important compounds, acting as signalling molecules, as energy storage and as surface layer, protecting the plant from environmental and biological stress (Ohlrogge, 1997). They act as hormones or their precursors in the human organism, help the digestive process and are a source of metabolic energy (Burtis and Ashwood, 1996).

The three samples essentially contain saturated fatty acids, which correspond to ca. 89, 93 and 95% of total fatty acids in samples A, B and C, respectively. Palmitic (16) and stearic (21) acids were the major fatty acids in all samples (Table 2), their content being significantly higher in sample C (Table 2) (p < 0.001).

Concerning to unsaturated fatty acids, a-linolenic (19) and oleic (20) acids were the most abundant ones, representing ca. 86, 88 and 74% of total unsaturated fatty acids in samples A, B and C, respectively (Table 2). Linoleic (18) and palmitelaidic acids (15) were the less representative compounds (Table 2). As they cannot be synthesized by the human organism, due to the lack of desaturase enzymes required for their production, a-linolenic and linoleic acids are essential

; to imi

metabolites. They must be obtained from the diet and represent two families of PUFA (omega-3 and omega-6, respectively), being precursors of their higher molecular weight and more unsaturated counterparts (Yunusova et al., 2012). Linoleic acid originates the omega-6 fatty acids series, which includes y-linolenic and arachidonic acids. Linoleic acid can be converted to hormone-like substances called eicosanoids, which affect physiological reactions ranging from blood clotting to immune response (Jeong and Lachance, 2001).

Oleic acid is included in the omega-9 family and is not essential for humans, once they possess all the enzymes required for the synthesis of this MUFA. It has been described in several species of Boraginaceae family and is recognized for its effectiveness in health promotion (Yunusova et al., 2012). Several authors have referred the contribution of oleic acid to the reduction of cholesterol levels, promoting the decrease of cardiovascular diseases, as well as its anti-diabetic activity (by inhibiting a-glucosidase) (Andrade et al., 2013) and anti-inflammatory and antioxidant properties (Andrade et al., 2013; Vassiliou et al., 2009).

Four sterols were also identified in G. diffusa ethanolic extracts (Table 2), all of them being already reported in different genus of Boraginaceae family (Yunusova et al., 2012). The three analysed samples exhibited similar sterols content (Table 2). ^-Sitosterol (28) and cholesterol (25) were the main compounds in the three samples and campesterol (26) the minor one (Table 2).

Plant-derived sterols are structurally similar and functionally analogous to cholesterol, but, unlike cholesterol, they are not synthesized by humans (Hooper and Cassidy, 2006). Plant materials can contain free and esterified sterols, ^-sitosterol, campesterol and stigmasterol being the most abundant in nature. These compounds are products of the isoprenoid biosynthetic pathway, being derived from squalene (Piironen et al., 2000). They are involved in important cellular processes, such as in the adaptation of membranes to temperature and regulation of their fluidity (Bouvier et al., 2005), also participating in cellular differentiation and proliferation (Piironen et al., 2000). ^-Sitosterol (28), the main sterol in G. diffusa, is the principal A5-sterol in several plant materials, being the most

efficient compound acting in membranes to restrict the motion of fatty acyl chains (Bouvier et al., 2005).

Phytosterols have been playing major roles in several areas, namely in pharmaceuticals, nutrition and cosmetics. In pharmaceutical terms, they are used in the production of therapeutic steroids (Fernandes and Cabral, 2007). Nutritionally, the intake of phytosterols may cause a decrease of plasma cholesterol levels, which can be due not only to the inhibition of intestinal cholesterol absorption, but also to other effects on hepatic / intestinal cholesterol metabolism (Hooper and Cassidy, 2006). Additionally, several pharmacological activities have been reported for these compounds, like anti-carcinogenic, anti-inflammatory and antidiabetic (Fernandes and Cabral, 2007). In the cosmetics industry this kind of compounds is usually incorporated in creams and lipsticks (Fernandes and Cabral, 2007). Thus, considering its high phytosterols content, G. diffusa may be incorporated in products to

19) and lu]

nti-cholesterol additives), or topically

be used orally, such as pharmaceutics and functional like cosmetics.

Concerning to triterpenes, j#-amyrin (29) and lupeol acetate (30) were found. Additionally, a compound whose mass spectrum closely resembled that of lanosterol (compound 24) was also present in all samples (Fig. 2, Table 2). The mass fragmentation of this compound was compared with that of some triterpenes, but no match was found. The higher content of triterpenes was observed in sample B (Table 2). Compound (24) was the major triterpene in samples B and C (corresponding to ca. 43 and 42% of total triterpenes, respectively), while lupeol acetate (30) was the most abundant in sample A (ca. 45% of total triterpenes).

Despite being found in minor amounts, j#-amyrin (29) is an important compound, as it possesses a wide spectrum of activities, including anti-inflammatory, anti-ulcer, anti-hyperlipidemic, antitumor and hepatoprotective (Oliveira et al., 2005).

3.3. Biological activities

3.3.1. Antiradical

The antiradical potential of the extracts was assessed by the DPPH' assay, a simple chemical test, commonly used as screening method to give basic information about the ability of extracts/compounds to scavenge free radicals. All G. diffusa ethanolic extracts showed a good concentration-dependent scavenging capacity against DPPH' (Fig. 3), IC50 values ranging from 0.070 mg of dried extract/mL (sample C) to 0.094 mg of dried extract/mL (sample A).

The good antiradical activity observed seems to be closely related with the chemical composition of the three samples of G. diffusa. Sample C, which exhibited higher phenolics and fatty acids contents, also showed the best scavenging capacity. In fact, several compounds found in the ethanolic extracts of the three samples, namely caffeic, rosmarinic and salvianolic acids and flavonoids, such as quercetin and kaempferol derivatives, have already been reported as scavengers of DPPH' (Burda and Oleszek, 2001; Maurya and Devasagayam, 2010; She et al., 2010).

Quercetin is one of the most potent antioxidants among polyphenols. The principal role of quercetin and its derivatives in plants is related with its antioxidant capacity, by the ability to transfer hydrogen or an electron, as well as by the chelation of metal ions and inhibition of oxidases (Materska, 2008). Despite the presence of quercetin, kaempferol and isorhamentin linked to sugars in these extracts, they are hydrolysed after ingestion, mostly in the gastrointestinal tract, and then absorbed and metabolized into its aglycone form that is more effective than the glycosylated one (Walle, 2004).

Some fatty acids like linoleic, linolenic and oleic acids have already been reported as strong scavengers of DPPH' (Andrade et al., 2013; Ramadan and El-Shamy, 2013). On the other hand, despite phytosterols are known for their antioxidant activity at the membrane level, by inhibiting lipids oxidation, their radical scavenging capacity is lower than that reported for phenolic compounds. The antioxidant activity of phytosterols has been attributed to the formation of an allylic free radical, which is then isomerized, leading to a stable tertiary radical that causes an interruption of the oxidative cascade (Kochhar, 2000). Concerning triterpenes, previous reports indicated that a-amyrin (a j#-amyrin

isomer) and lupeol acetate do not show DPPH' scavenging activity (Biskup et al., 2012; Lucetti et al., 2010). This is not surprising, as triterpenes lack structural features mostly related to the antioxidant activity of a compound, such as o-dihydroxybenzene (catechol) structure (Burda and Oleszek, 2001). These data are consistent with the obtained results. In general, the contents of sterols and triterpenes in samples A and B were similar to that of sample C; nevertheless, the first samples exhibited a phenolic composition, the DPPH' scavenging activity being lower.

3.3.2. Acetyl- and butyrylcholinesterase inhibition

AChE and BuChE are the principal enzymes associated with the aetiology of AD. One of the therapeutic strategies in AD is the use of inhibitors of cholinesterase in order to increase the levels of

enes in poorer

acetylcholine in the synaptic cleft.

Concerning to AChE inhibition, dose-response behaviour was observed for all samples, sample B being the most effective one (IC20 at 0.977 mg of dried extract/mL) (Fig. 3). Sample C was the less active, inhibiting only ca. 19% at the highest concentration tested (1.29 mg of dried extract/mL). It was not possible to test higher concentrations due to low solubility of the extract: the formation of a precipitate was noticed, invalidating the assay.

Regarding BuChE inhibitory activity, the same behaviour was found. Sample B showed the strongest capacity (IC20 at 0.452 mg of dried extract/mL). Sample C was the least effective, only ca. 17% inhibition being observed at the highest tested concentration (1.28 mg of dried extract/mL) (Fig. 3). Due to solubility issues it was not possible to test higher concentrations.

Both G. diffusa ethanolic extracts were less effective than galantamine (positive control tested under same conditions), which displayed IC20 values of 1 ^g/mL against both AChE and BChE (data not shown).

Recent studies point to phenolic compounds ability to inhibit cholinesterase (Fan et al., 2010; Ferreres et al., 2014). AChE inhibition capacity of fatty acids like oleic, linolenic and linoleic acids was also previously described (Ren et al., 2006). Additionally, palmitic acid, the major fatty acid in G.

diffusa, was already reported to have moderate AChE inhibitory activity, while linolenic acid exhibited mild activity against BuChE (Fang et al., 2010). Hereupon, G. diffusa ethanolic extracts exhibited low inhibitory activity, contradicting what could be expected. Sample C, the material with highest content of the mentioned bioactive compounds, was the less effective against AChE and BChE. It is known that the activity of an extract is due to the joint action of all its constituents, which may not correspond to the sum of the activities shown by each compound individually. Thus, the low AChE and BuChE inhibitory capacity may be due to interactions among the several components of the tested ethanolic extracts.

3.3.3. Effects on AGS and Caco-2 cells

f the teste<

lavonoids

A variety of compounds from natural sources, for example, flavonoids, phenolic acids, sterols and polyunsaturated fatty acids, has proven benefits against cancer (Fernandes and Cabral, 2007; Ferreres

ist cancer 1003). Ta

et al., 2014; Hooper and Cassidy, 2006; Surh, 2003). Taking into account the rich chemical composition of G. diffusa, we considered to be relevant to evaluate its effect on tumour-related cell lines, such as human colorectal adenocarcinoma cell line (Caco-2) and human gastric adenocarcinoma cells line (AGS). After exposure, cells treated with the highest extracts concentrations exhibited modifications of size, shape and confluence. Viability, namely concerning to mitochondrial function (evaluated by the MTT reduction) and membrane integrity (evaluated by the LDH release), was assessed. The obtained results point to a low cytotoxic effect (Fig. 4). The three G. diffusa extracts interfered with the mitochondrial function (p < 0.001) of both AGS and Caco-2 cells at the highest concentrations tested (Fig. 4).

Concerning to AGS cells, as ascertained by the results for MTT reduction, a significant decrease of cell viability of about 75, 45 and 33% was noticed when exposed to 3 mg of sample A dried extract/mL, 2 mg of sample B dried extract/mL and 4 mg of sample C dried extract/mL, respectively. Regarding Caco-2 cells, the same extracts concentrations of samples A, B and C lead to 60, 41 and 27% of cell death, respectively.

Sample B was cytotoxic for AGS cells at 1 mg of dried extract/mL (p < 0.05). Nevertheless, sample A was the most effective, being cytotoxic for both AGS (p < 0.01) and Caco-2 (p < 0.05) cells at 1.5 mg of dried extract/mL (Fig. 4). Although sample C also revealed a toxic effect on Caco-2 cells at the second highest concentration tested (2 mg of dried extract/mL) (p < 0.01), it was the less cytotoxic (Fig. 4).

In a general way, G. diffusa ethanolic extracts did not affect the membrane integrity, as ascertained by the results of the LDH leakage assay (Fig. 4). Nevertheless, as significant LDH leakage was detected only for the highest tested concentration of sample A, and just in Caco-2 cell s (3 mg/mL), only in this case cell membrane was disrupted; under those conditions osmolality issues related with the high extract's concentration may have affect the results. Apoptosis and necrosis are two major forms of cell death observed in normal processes and disease. A key signature for necrotic cells is the permeabilization of the plasma membrane. This event can be quantified in tissue culture settings by measuring the release of the intracellular enzyme LDH. Nevertheless, despite the necessity of being combined with other methods, quantification of LDH release is a useful method for the detection of necrosis (Chan et al., 2013). Thus, as the loss of cell viability reflected by the MTT assay was not accompanied by elevated levels of LDH in the medium, it can be supposed that the extracts do not cause cell death by a necrotic process.

These results may point to an apoptotic effect exerted by the extracts. However, other assays are required for the detection of apoptosis and further mechanistic and clinical studies are needed to establish whether these extracts can be exploited to achieve, not only preventive, but also therapeutic effects on carcinogenic process (Surh, 2003).

4. Conclusion

The phenolics profile was successfully used to guarantee the authenticity of three commercial samples of G. diffusa: the similar phenolics composition of the ethanolic extract and of the water one previously reported (Ferreres et al., 2013) helped us to ascertain the genuineness of the marketed

product. G. diffusa amino acids, fatty acids, sterols and triterpenes were studied for the first time, improving the knowledge on the chemical composition of this species. Despite the similarity of the qualitative profiles, significant quantitative differences were observed among the three samples.

G. diffusa ethanolic extracts revealed a good antiradical activity, a mild effect on cholinesterase and were cytotoxic for AGS and Caco-2 cells at high concentrations. These activities are related with the chemical composition. Nevertheless, it should be highlighted that the effect of an extract cannot be simply extrapolated from the activities of their isolated compounds. Synergic and/or antagonic effects between all the constituents of an extract have to be considered. As herbal teas of marketed products are usually prepared, a lot of constituents, instead of an isolated compound or a fraction of compounds, is consumed. As so, the study of extracts seems to be more realistic due to the occurrence of possible interactions between components, which determine the overall effect.

Acknowledgements

ie Europea

This work received financial support from the European Union (FEDER funds through COMPETE) and National Funds (FCT, Fundagäo para a Ciencia e a Tecnologia) through project Pest-C/EQB/LA0006/2013. The work also received financial support from the European Union (FEDER funds) under the framework of QREN through Project N0RTE-07-0124-FEDER-000069, from CYTED Programme (Ref. 112RT0460) CORNUCOPIA Thematic Network and from the project AGL2011-23690 (CICYT). To all financing sources the authors are greatly indebted. F.F. is indebted to FCT for the grant (SFRH/BPD/98732/2013).

References

enoid a-

entals o 4.J., 2013.

Andrade, P.B., Barbosa, M., Matos, R.P., Lopes, G., Vinholes, J., Mouga, T., Valentao, P., 2013. Valuable compounds in macroalgae extracts. Food Chem. 138,1819-1828.

Biskup, E., Gol^biowski, M., Gniadecki, R., Stepnowski, P., Lojkowska, E., 2012. Triterpenoic amyrin stimulates proliferation of human keratinocytes but does not protect them against UVB damage. Acta Biochim. Pol. 59, 255-260.

Bouvier, F., Rahier, A., Camara, B., 2005. Biogenesis, molecular regulation and function of plant isoprenoids. Prog. Lipid Res. 44, 357-429.

Burda, S., Oleszek, W., 2001. Antioxidant and antiradical activities of flavonoids. J. Agric. Food Chem. 49, 2774-2779.

Burtis, C.A., Ashwood, E.R., 1996. Tietz fundamentals of clinical chemistry, fourth ed. W. B. Saunders Company, Philadelphia.

Chan, F.K.-M., Moriwaki, K., De Rosa, M.J., 2013. Detection of necrosis by release of lactate dehydrogenase activity, in: Snow, A.L., Lenardo, M.J. (Eds.) Immune Homeostasis - Methods in Molecular Biology. Springer Science, New York, pp. 65-70.

Daur, I., 2013. Chemical composition of selected Saudi medicinal plants. Arab. J. Chem. doi:10.1016/j.arabjc.2013.10.015.

Dreger, M., Krajewska-Patan, A., Gorska-Paukszta, M., Pieszak, M., Buchwald, W., Mikolajczak, P., 2010. Production of the secondary metabolites in Salvia miltiorrhiza in vitro cultures. Herba Polonica 56, 78-90.

El-Chaghaby, G.A., Ahmad, A.F., Ramis, E.S., 2014. Evaluation of the antioxidant and antibacterial properties of various solvents extracts of Annona squamosa L. leaves. Arab. J. Chem. 7, 227-233.

Fan, P., Terrier, L., Hay, A.-E., Marston, A., Hostettmann, K., 2010. Antioxidant and enzyme inhibition activities and chemical profiles of Polygonum sachalinensis F. Shmidt ex Maxim (Polygonaceae). Fitoterapia 81, 124-131.

Fang, Z., Jeong, S.Y., Jung, H.A., Choi, J.S., Min, B.S., Woo, M.H., 2010. Anticholinesterase and

antioxidant constituents from Gloiopeltis furcate. Chem. Pharm. Bull (Tokyo) 58, 1236-1239. FAO (FAO Corporate Document Repository). FAO agricultural services bulletin no.43, 1981. Available online:

http://cd3wd.com/data/1005_core_agric_2_/_ag_postharvest_1012_inpho_en_lp_114580_.]

(accessed on 14th January 2014).

irade, P

)ds. Biore;

0 .pdf

Fernandes, P., Cabral, J.M.S., 2007. Phytosterols: applications and recovery methods. Biores. Technol. 98, 2335-2350.

Ferreres, F., Grosso, C., Gil-Izquierdo, A., Valentao, P., Azevedo, C., Andrade, P.B., 2014. HPLC-DAD-ESI/MSn analysis of phenolic compounds for quality control of Grindelia robusta Nutt and bioactivities. J. Pharm. Biomed. Anal. 94, 163-172. Ferreres, F., Vinholes, J., Gil-Izquierdo, A., Valentao, P., Goncalves, R.F., Andrade, P.B., 2013. In vitro studies of a-glucosidase inhibitors and antiradical constituents of Glandora diffusa (Lag.) D.C. Thomas infusion. Food Chem. 136, 1390-1398. Galili, S., Amir, R., Galili, G., 2008. Genetic engineering of amino acid metabolism in plants. Adv.

Cell Mol. Biol. Plants 1, 49-80. Hooper, L., Cassidy, A., 2006. A review of the health care potential of bioactive compounds. J. Sci.

Food Agric. 86, 1805-1813. Jeong, W.S., Lachance, P.A., 2001. Phytosterols and fatty acids in fig (Ficus carica var. Mission) fruit

and tree components. J. Food Sci. 66, 278-281. Kochhar, S.P., 2000. Stabilisation of frying oils with natural antioxidative components. Eur. J. Lipid

Sci. Technol. 102, 552-559. Lee, J.E., Lee, B.J., Chung, J.O., Hwang, J.A., Lee, S.J., Lee, C.H., Hong, Y.S., 2010. Geographical and climatic dependencies of green tea (Camellia sinensis) metabolites: a 1H NMR-based metabolomics study. J. Agric. Food Chem. 58, 10582-10589. Lléo, A., 2007. Current therapeutic options for Alzheimer's disease. Curr. Genomics 8, 550-558.

-Review. I tire of hydrox

mel. J.

Lu, Y., Foo, L.Y., 2002. Polyphenolics of Salvia - a review. Phytochemistry 59, 117-140. Lucetti, D.L., Lucetti, E.C., Bandeira, M.A., Veras, H.N., Silva, A.H., Leal, L.K., Lope, A.A., Alves, V.C., Silva, G.S., Brito, G.A., Viana, G.B., 2010. Anti-inflammatory effects and possible mechanism of action of lupeol acetate isolated from Himatanthus drasticus (Mart.) Plumel. J. Inflamm. (Lond) 7, 60-71. Materska, M., 2008. Quercetin and its derivatives: chemical structure and bioactivity-Review. Pol. J.

Food Nutr. Sci. 58, 407-413. Maurya, D.K., Devasagayam, T.P.A., 2010. Antioxidant and prooxidant nature of hydroxycinnamic

acid derivatives ferulic and caffeic acids. Food Chem. Toxicol. 48, 3369-3373. Milella, L., Caruso, M., Galgano, F., Favati, F., Padula, M.C., Martelli, G., 2011. Role of the cultivar in choosing clementine fruits with a high level of health-promoting compounds. J. Agric. Food Chem. 59, 5293-5298.

NIST Chemistry WebBook, Available onlin e: http )://webbook.nist.gov/chemistry/name-ser.html

(accessed on 3m February 2014). Ohlrogge, J.B., 1997. Regulation of fatty acids synthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 48, 109-36.

Oliveira, F.A., Chaves, M.H., Almeida, F.R., Jr. Lima, R.C., Silva, R.M., Maia, J.L., Brito, G.A., Santos, F.A., Rao , V.S., 2005. Protective effect of a- and ß-amyrin, a triterpene mixture from Protium heptaphyllum (Aubl.) March. trunk wood resin, against acetaminophen-induced liver injury in mice. J. Ethnopharmacol. 98, 103-108. Park, S.U., Uddin, M.R., Xu, H., Kim, Y.K., Lee, S.Y., 2008. Biotechnological applications for

rosmarinic acid production in plant. Afric. J. Biotechnol. 7, 4959-4965. Pereira, D.M., Valentäo, P., Ferreres, F., Andrade, P.B., 2010. Metabolomic analysis of natural products, in: Zacharis, C.K., Tzanavaras, P.D. (Eds.), Reviews in pharmaceutical & biomedical analysis. Bentham eBooks, Thessaloniki, pp. 1-19.

ttp://we

.H., Al

n of fatty

Pereira, D.M., Vinholes, J., Guedes de Pinho, P., Valentao, P., Mouga, T., Teixeira, N., Andrade, P.B., 2012. A gas chromatography-mass spectrometry multi-target method for the simultaneous analysis of three classes of metabolites in marine organisms. Talanta 100, 391-400.

Petersen, M., Simmonds, M.S.J., 2003. Rosmarinic acid. Phytochemistry 62, 121-125.

Petrovska, B.B., 2012. Historical review of medicinal plants' usage. Pharmacogn Rev. 6, 1-5.

Piironen, V., Lindsay, D.G., Miettinen, T.A., Toivo, J., Lampi, A.-M., 2000. Plant sterols: biosynthesis, biological function and their importance to human nutrition. J. Sci. Food Agric. 80, 939-966.

Ramadan, M.F., El-Shamy, H., 2013. Snapdragon (Antirrhinum majus) seed oil: characterization of fatty acids, bioactive lipids and radical scavenging potential. Ind. Crop Prod. 42, 373-379.

Ren, Y., Houghton, P., Hider, R.C., 2006. Relevant activities of extracts and constituents of animals used in traditional Chinese medicine for central nervous system effects associated with Alzheimer's disease. J. Pharm. Pharmacol. 58, 989-996.

Rhee, K., van de Meent, M., Ingkaninan, K., Verpoorte, R., 2001. Screening for acetylcholinesterase inhibitors from Amaryllidaceae using silica gel thin-layer chromatography in combination with bioactivity staining. J. Chromatogr. A 915, 217-223.

Sá-Otero, M.P., Armesto-Baztan, S., Díaz-Losada, E., 2006. A study of variation in the pollen spectra of honeys sampled from the Baixa Limia-Serra do Xurés Nature Reserve in north-west Spain. Grana 45,137-145.

She, G.M., Xu, C., Liu, B., Shi, R.B., 2010. Polyphenolic acids from mint (the aerial of Mentha haplocalyx Briq.) with DPPH radical scavenging activity. J. Food Sci. 75, C359-C362.

Sousa, C., Fernandes, F., Valentao, P., Rodrigues, S., Coelho, M., Teixeira, J.P., Silva, S., Ferreres, F., Guedes de Pinho, P., Andrade, P.B., 2012. Brassica oleracea L. var. costata DC and Pieris brassicae L. aqueous extracts reduce methyl methanesulfonate-induced DNA damage in V79 hamster lung fibroblasts. J. Agric. Food Chem. 60, 5380-5387.

Surh, Y.-J., 2003. Cancer chemoprevention with dietary phytochemicals. Nat. Rev. Cancer 3, 768-778.

Thomas, D.C., Weigend, M., Hilger, H.H., 2008. Phylogeny and systematics of Lithodora (Boraginaceae-Lithospermeae) and its affinities to the monotypic genera Mairetis, Halacsya and Paramoltkia based on ITS1 and trnLUAA-sequence data and morphology. Taxon. 57, 79-97.

Vassiliou, E.K., Gonzalez, A., Garcia, C., Tadros, J.H., Chakraborty, G., Toney, J.H., 2009. Oleic acid and peanut oil high in oleic acid reverse the inhibitory effect of insulin production of the inflammatory cytokine TNF- both in vitro and in vivo systems. Lipids Health Dis. 26, 8-25.

Walle, T., 2004. Absorption and metabolism of flavonoids. Free Radic. Biol. Med. 36, 829-837.

Yunusova, S.G., Khatmulina, L.I., Fedorov, N.I., Ermolaeva, N.A., Galkin, E.G., Yunusov, M.S., 2012. Polyunsaturated fatty acids from several plant species of the family Boraginaceae. Chem. Nat. Compd. 48, 361-366.

Figure 1 HPLC-DAD phenolic profile of the ethanolic extract from G. diffusa aerial parts (sample C). Detection at 330 nm. (1) Caffeic acid; (2) Salvianolic acid H; (3) p-Coumaric acid; (4) Quercetin-3-0-(2,6-di-rhamnosyl)galactoside; (5) Quercetin-3-0-(2,6-di-rhamnosyl)glucoside; (6) Salvianolic acid E isomer; (7) Salvianolic acid I; (8) Kaempferol-3-0-(2,6-di-rhamnosyl)hexoside; (9) Kaempferol-3-0-(2-rhamnosyl)galactoside; (10) Kaempferol-3-0-(2-rhamnosyl)glucoside; (12) Quercetin-3-0-(6-rhamnosyl)glucoside; (13) Salvianolic acid A isomer; (15) Rosmarinic acid; (18) Salvianolic acid B; (19) Kaempferol-3-0-(6-rhamnosyl)hexoside; (21) Isorhamnetin-3-0-(6-rhamnosyl)hexoside; (23) Methyl rosmarinic acid; (24) Salvianolic acid E isomer; (25) Salvianolic acid C isomer. Compounds 11, 14, 16, 17, 20 and 22 are unknown caffeic acid derivatives.

Figure 2 GC-MS profile of G. diffusa ethanolic extract (sample A). Identity of compounds as in Table 2. Internal standard: A-norvaline; B-methyllinolelaidate; C-desmosterol. All compounds

correspond to their trimethylsilyl (TMS) derivativ

Figure 3 Antiradical activity of ethanolic extracts of G. diffusa against DPPH and their enzymatic inhibition towards acetylcholinesterase and butyrylcholinesterase. Results show mean ±

SEM of three experiments performed in triplicate. &

Figure 4 Cytotoxicity evaluation in AGS and Caco-2 cells for different concentrations of extracts from G. diffusa samples. Results are presented as mean ± SEM of 4 independent experiments (triplicates were performed in each experiment). Mean values were significantly different compared with the respective control (a p < 0.05, aa p < 0.01 and aaa p < 0.001); (bbb p < 0.001). For MTT, the control corresponds to untreated cells (100% of viability). For LDH, control corresponds to the basal LDH release by untreated cells (100%).

Figure 1

10 20 30 40 50

Minutes

Figure 2

Minutes

Figure 3

Figure 4

Table 1 Polyphenolic composition of ethanolic extracts of aerial parts of G. diffused.

Sample A Sample B Sample C

Average mg/Kg dried sample

Phenolic acid monomers

1 Caffeic acid 3.1 (0.0)aa 4.0 (0.3)bbb 9.2 (0.2)ccc

3 p-Coumaric acid n.q. n.q. n.q.

Caffeic acid polymers

Dimers

15 Rosmarinic acid 140.7 (8.4)aaa 313.0 (3.0)bbb 590.8 (2.3)ccc

17 Unknown 15.6 (2.5)aaa 33.2 (0.1)bbb 42.1 (0.3)ccc

Trimers i

2 Salvianolic acid H 5.5 (0.0) 5.6 (0.1)bbb 16.8 (0.1)ccc

7 Salvianolic acid I aaa n.q. 4.7 (0.3)bbb n.q.

11 Unknown 1.6 (0.0)aaa 3.5 (0.0)bbb 2.7 (0.2)ccc

16 Unknown 14.1 (0.3)aaa 32.2 (1.3) 34.6 (1.0)ccc

20 Unknown 4.7 (0.2)aaa 16.1 (0.5)bbb 2.8 (0.3)cc

25 Salvianolic acid C isomer 4.8 (0.0)a 5.7 (0.4)bbb 14.9 (0.4)ccc

Tetramers

6 Salvianolic acid E isomer 19.4 (0.1)aa 11.5 (1.3)bbb 29.5 (2.3)ccc

18 Salvianolic acid B 43.6 (2.4)aaa 73.1 (1.4)bbb 82.7 (0.7)ccc

Pentamer

22 Unknown 0.8 (0.0) n.q. n.q.

Flavonoids

4 Quercetin-3-0-(2,6-di-rhamnosyl)galactoside 16.3 (0.6) 19.0 (1.1) 22.2 (2.7)c

5 Quercetin-3-0-(2,6-di-rhamnosyl)glucoside 28.0 (1.3)aa 15.8 (3.5) 15.9 (0.3)cc

10 Kaempferol-3-0-(2-rhamnosyl)glucoside 3.7 (0.4)aa 4.7 (0.1)bbb 2.0 (0.1)ccc

12 Quercetin-3-0-(6-rhamnosyl)glucoside 82.2 (1.2) 81.1 (0.9)bbb 110.3 (7.0)ccc

19 Kaempferol-3- 0-(6-rhamnosyl)hexoside 23.0 (2.8)a 30.6 (0.4)bb 16.5 (3.3)c

21 Isorhamnetin-3-O-(6-rhamnosyl)hexoside 22.4 (0.5)aaa 34.6 (1.1)bbb 14.9 (3.0)cc

I 429.5 (20.7)aaa 688.4 (15.8)bbb 1007.9 (24.2)ccc

A Results expressed as mean (standard deviation) of three determinations; n.q.: not quantified; n.d.: not detected. a Comparison between sample A and B, ap < 0.05, aa p < 0.01, aaa p < 0.001. b Comparison between sample B and C, bbp < 0.01, bbb p < 0.001. c Comparison between sample C and A, cp < 0.05, ccp < 0.01, ccc p < 0.001.

Table 2

Amino acids, fatty acids, sterols and triterpenes composition of ethanolic extracts of

aerial parts of G. diffusa

Peak RI

Compound

Sample A

Sample B

Sample C

Average mg/Kg dried sample

Amino acids

2 1067 10951 Alanine 228.0 (3.2)aaa

3 1083 11151 Glycine 746.0 (7.9)aaa

5 1175 12101 Valine 19.6 (0.46)aaa

6 1252 12741 Isoleucine 123.7 (11.6)aaa

7 1314 13432 Serine 151.9 (8.6)aaa

9 1341 13461 Threonine 21.0 (2.4)aaa

10 1363 Homocysteine 65.9 (6.3)aa

12 1497 15001 irans-4-Hydroxyproline 239.5 (10.5)aaa

n.q. n.d.

j b n.d.

46.5 (4.3)'

Total 1595.5 (50.9)a

46.5 (4.3)

>UUU ^ )bbb 4

248.2 (5.0) ccc

20.7 (1.8)ccc

48.8 (6.6)ccc

(1.3)ccc

135.5 (3.6)ccc

485.1 (18.3)°

Fatty acids

1 1044 Caproic (C6:0) 47.0 (4.1)aaa 93.1 (4.0)bbb 127.8 (7.5)ccc

4 1130 Enanthic (C7:0) 32.5 (1.7) 32.0 (2.5)bbb 158.5 (12.1)ccc

8 1324 13552 Pelargonic (C9:0) 50.4 (5.0)a 75.6 (1.1)bbb 193.7 (16.0)ccc

11 1426 14502 Capric (C10:0) 69.8 (6.2)a 152.6 (9.4)bbb 434.5 (38.1)ccc

13 1628 16432 Lauric (C12:0) 47.1 (2.4)aa 85.4 (8.0)bbb 164.0 (12.1)ccc

14 1822 18432 Myristic (C14:0) 17.6 (1.7)a 96.7 (9.6)bbb 644.6 (52.9)ccc

15 2002 Palmitelaidic (C16:1) 14.4 (1.4)aa bbb n.q. 57.1 (5.2)ccc

16 2021 20402 Palmitic (C16:0) 1116.6 (1.1)a 1404.8 (132.8)bbb 2256.1 (36.7)ccc

17 2118 21471 Margaric (C17:0) 33.6 (3.0) 26.8 (2.42)bbb 74.0 (4.6)ccc

18 2192 22141 Linoleic (C18:2) 21.0 (0.6) 20.5 (1.2) 21.4 (1.5)

19 2201 22251 a-Linolenic (C18:3) 115.3 (11.1)a 91.8 (4.8) 100.7 (9.9)

20 2203 22162 Oleic (C18:1) 106.2 (5.4)aaa 66.7 (5.7)bbb 125.6 (1.1)cc

21 2220 22342 Stearic (C18:0) 438.8 (34.4) 380.9 (31.1)bbb 1125.9 (127.4)ccc

22 2619 26322 Docosanoic (C22:0) 78.4 (6.9) 60.2 (6.1)bbb 106.7 (10.5)c

23 2818 28502 Lignoceric (C24:0) 70.0 (5.1) 59.1 (5.4)bb 87.0 (7.6)c

Sterols

Total 2258.4 (90.1) 2646.2 (224.3)bbb 5677.5 (343.1 )c

25 3147 31971 Cholesterol 194.3 (18.2) 181.0 (15.7) 181.4 (11.5)

26 3252 Campesterol 28.8 (2.9) 28.1 (2.9) 29.8 (2.9)

27 3275 Stigmasterol 59.0 (5.7) 50.7 (2.9) 51.9 (1.4)

28 3335 33553 r ß-Sitosterol 233.7 (11.3) 236.6 (13.1) 239.5 (15.2)

y Total 515.7 (38.0) 496.5 (34.6) 502.6 (31.0)

Triterpenes

24 29 30 3118 3368 3414 M+ 467.9 m/z (467.9 (100), 75 (70), 73 (45), 469 (35), 57 (30) ß-amyrin Lupeol acetate 114.0 (11.2)a 45.6 (1.9)aaa 132.1 (8.5) 142.6 (8.1) 66.3 (3.2)bb 124.5 (9.6)bb 117.6 (11.2) 79.2 (3.7)ccc 81.0 (8.1)ccc

Total 291.7 (21.7) 333.4 (20.9)b 277.9 (23.0)

2 4661.3 (200.68)a 3522.6 (284.1)bbb 6943.1 (415.4)ccc

A Results expressed as mean (standard deviation) of three determinations; n.q.: not quantified; n.d.: not detected. 1 and 4 quantified as nonanoic acid; 10 quantified as cysteine; 22 and 23 quantified as docosahexaenoic acid; 24 quantified as ergosterol; 29 quantified as lupeol. RIExp - Retention Index obtained in this experiments; RILit - Retention Index described in literature: !Pereira et al., 2012; 2NIST Chemistry WebBook (http://webbook.nist.gov/chemistry/name-ser.html);

3Andrade et al., 2013.

a Comparison between sample A and B, ap < 0.05, aap < 0.01, aaap < 0.001. b Comparison between sample B and C, b p < 0.05, bbp < 0.01, bbb p < 0.001. c Comparison between sample C and A, cp < 0.05, ccp < 0.01, cccp < 0.001.