Isolation, identification and activity of natural antioxidants from horehound (Marrubium vulgare L.) cultivated in Lithuania

Food Chemistry 130 (2012) 695–701

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Isolation, identification and activity of natural antioxidants from horehound (Marrubium vulgare L.) cultivated in Lithuania Audrius Pukalskas a, Petras Rimantas Venskutonis a, Sofia Salido b, Pieter de Waard c, Teris A. van Beek d,⇑ a

Department of Food Technology, Kaunas University of Technology, Radvile˙nuz pl. 19, LT-50254 Kaunas, Lithuania Dpto. Quimica Inorganica y Organica, Facultad de Ciencias Experimentales, Universidad de Jaen, 23071 Jaen, Spain c Wageningen NMR Centre, Wageningen University, Dreijenlaan 3, 6703 HA Wageningen, The Netherlands d Laboratory of Organic Chemistry, Natural Products Chemistry and Microfluidics Group, Wageningen University, Dreijenplein 8, 6703 HB Wageningen, The Netherlands b

a r t i c l e

i n f o

Article history: Received 14 December 2010 Received in revised form 4 May 2011 Accepted 26 July 2011 Available online 2 August 2011 Keywords: Horehound Marrubium vulgare Antioxidants Flavonoids

a b s t r a c t In an earlier screening of Lithuanian plants, horehound (Marrubium vulgare) showed good antioxidant activity and as this species is used in herbal teas and cough pastilles it was selected for further investigation. Some fractions of the aerial parts were strong scavengers of the model free radicals DPPH and ABTS+. Activity in the b-carotene bleaching assay and the rapeseed oil oxidation assay was lower. Several active compounds were observed in the crude methanol–water extract, and in butanol and methyl tert-butyl ether fractions using HPLC with on-line radical scavenging detection. After multi-step fractionation of these fractions five compounds possessing radical scavenging activity were purified and their structures were elucidated by NMR and MS as 5,6-dihydroxy-7,40 -dimethoxyflavone (syn. ladanein), 7-O-b-glucopyranosyl luteolin, 7-O-b-glucuronyl luteolin, verbascoside and forsythoside B. Their activities were tested in off-line DPPH and ABTS+ free radical scavenging assays, and compared with the antioxidants rosmarinic acid and Trolox. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Antioxidants are widely used in foods, cosmetics and pharmaceuticals (Cuvelier, Richard, & Berset, 1996). Since the 1980’s there has been a growing interest in research and application of natural antioxidants instead of synthetic ones, caused by a consumer demand for natural food additives. Additionally, for the latter the burden of proof of safety may be less rigorous than that required for synthetic antioxidants. The antioxidant activity of many plants has been investigated, including species growing in Lithuania (Miliauskas, Venskutonis & van Beek, 2004; Pukalskas et al., 2002). However, to date, only rosemary and sage extracts are commercially available as flavourless, odourless, and colourless antioxidant extracts. Preliminary assessment of horehound antioxidant properties in rapeseed oil at 80 °C showed that an acetone leaf extract was as active as that of sage and this prompted us to investigate this plant in greater detail. Horehound (Marrubium vulgare L.), a member of the Labiatae, is native to North Africa, Central and Western Asia, and Southern Europe. It grows wild in dry sandy soils and wastelands. The species can be cultivated successfully in Lithuania and is harvested twice a year as a medicinal raw material (Puziene & Gudanavicius,

⇑ Corresponding author. Tel.: +31 317 482376; fax: +31 317 484914. E-mail address: [email protected] (T.A. van Beek). 0308-8146/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodchem.2011.07.112

1978). Horehound also serves as a raw material for herbal extracts and beverage industries. For instance, the plant has been used as a substitute for hops in beer-breweries and is currently used to make herbal teas. It is an ingredient of cough pastilles and of European candies, e.g. RicolaÓ. Phytochemical investigations of horehound resulted in the isolation of the flavonoids apigenin and luteolin and their 7-glucosides together with quercetin and its 3-glucoside and 3-rhamnoglucoside (Kovalewski & Matlwaska, 1978). Nawwar et al. reported on the isolation and structural elucidation of the flavonoids luteolin, and apigenin 7-lactates together with their 200 -Ob-glucuronides and 200 -O-b-glucosides (Nawwar, El-Mousalamy, Barakat, Buddrus, & Linscheid, 1989). Very recently ladanein (5,6dihydroxy-7-40 -dimethoxyflavone) was isolated from M. vulgare. This compound has possible antileukemic activity (Alkhatib et al., 2010). Several diterpenoids were also found, the main one being marrubiin (Knoess, 1994; Popa & Pasechnick, 1975; Rey, Levesque, & Pousset, 1992). The essential oil of horehound has been relatively well studied by several different groups (Khanavi, Ghasemian, Motlagh, Hadjiakhoonid, & Shafiee, 2005; Morteza-Semnani, Saeedi, & Babanezhad, 2008; Weel, Venskutonis, Pukalskas, Gruzdiene, & Linssen, 1999). The aim of this study was the evaluation of antioxidant properties of horehound extracts and their fractions obtained by using different solvents and to characterise antioxidative active components. This study was inspired by the following considerations:

696

A. Pukalskas et al. / Food Chemistry 130 (2012) 695–701

(i) botanicals are one of the main sources in the development of functional foods and food supplements; (ii) phytochemical composition in terms of secondary metabolites depends on many factors, including cultivation conditions; (iii) developments in analytical techniques enable to expand our knowledge on the composition of plant metabolites, particularly of less investigated species.

2. Materials and methods 2.1. Materials The following solvents were used for the extraction and fractionation: methanol, hexane, methyl tert-butyl ether, and butanol. All solvents were distilled prior to use. Solvents used for preparative chromatography and antioxidant activity testing were of analytical grade (Sigma Chemical, St. Louis, MO). The solvents used for HPLC were of HPLC grade (Lab-Scan Analytical Sciences, Dublin, Ireland). The following reagents were used in the antioxidant activity experiments: Tween 40, trans-b-carotene, linoleic acid (purity ca. 99%), 2,2-diphenyl-1-picrylhydrazyl hydrate (DPPH) (95%, Sigma–Aldrich Chemie, Steinheim, Germany), rosmarinic acid (Extrasynthese, Genay, France), 2,6-di-tert-butyl-hydroxytoluene (BHT) (Sigma–Aldrich Chemie, Steinheim, Germany), 2,20 -azino-bis(3ethylbenzothiazoline-6-sulphonic acid) diammonium salt (ABTS) (Fluka Chemie, Buchs, Switzerland) and Trolox 97% (Sigma–Aldrich Chemie, Steinheim, Germany). Freshly manufactured rapeseed oil obtained from low erucic acid bearing seeds of Brassica napus L. was donated by the company Obeliu aliejus, Lithuania. Deuterated methanol, deuterated chloroform and deuterated dimethyl sulfoxide (Acros Organics, Geel, Belgium) were used to prepare solutions of compounds for NMR analysis. 2.2. Preparation of plant extracts Aerial parts of Marrubium vulgare were obtained from the collection of Kaunas Botanical Garden, air dried in a Vasara ventilated oven (Utenos krosnys, Utena, Lithuania) at 30 °C for about 48 h and ground to fine particles before use. Acetone extracts (AE) were obtained by extracting approximately 15 g of freshly ground leaves with 900 ml of acetone, during 4 h in a Soxhlet extraction apparatus. 50 g of dried plant material was also extracted (2  1 l) with methanol–water–acetic acid (79:20:1) at room temperature for 24 h. Solvent and plant material was constantly mixed on an Ikamag ‘‘RTC basic’’ magnetic stirrer (IKA Labortechnik, Staufen, Germany). The extracts obtained were concentrated in a rotary evaporator at 40 °C. The extracts were suspended in 500 ml ultra pure water and then extracted with hexane, methyl tert-butyl ether and finally butanol. Several successive extractions with every solvent were made, every time using 100 ml of solvent. In total 500–600 ml of each solvent were used. Organic solvents were removed with a rotary evaporator at 40 °C. The remaining aqueous phase was freeze-dried. 2.3. b-Carotene bleaching test The antioxidant activities (AA) of herb extracts and reference antioxidants (BHT and rosmarinic acid) were determined using a previously developed (Marco, 1968) and later modified (Dapkevicius, Venskutonis, van Beek, & Linssen, 1998) method with slight changes. trans-b-Carotene (1 mg) was dissolved in 5 ml of chloroform and 1.0 ml of this solution was transferred with a pipette into a 100 ml round bottom flask. Linoleic acid (25 ll) and Tween 40 (200 mg) were added to the b-carotene solution and chloroform was evaporated under vacuum at 40 °C. Oxygenated ultra pure

water (50 ml), obtained by sparging with air during 15 min, was added and the mixture was vigorously shaken by hand. This emulsion was freshly prepared prior to each experiment. Stock solutions of reference antioxidants (BHT and rosmarinic acid, each 0.01%) and herb extracts (0.1%) were prepared in methanol. The b-carotene – linoleic acid emulsion (250 ll) was dosed into every well of 96-well microtiter plates (Greiner Labortech, The Netherlands) and 30 ll of methanolic solutions of the antioxidants were added. An equal amount of methanol was used for the blank sample. Four replicates were prepared for every tested sample. The microtiter plates were incubated at 55 °C during 120 min. The absorbance of the samples was measured in an EAR 400 Microtiter reader (SLT instruments, Austria) at 490 nm. Readings of all samples were performed immediately after preparation of the samples (t = 0 min) and at 15 min intervals during 120 min. The antioxidant activity coefficient (AAC) was calculated from the obtained data by the following formula (Chevolleau, Mallet, Ucciani, Gamisans, & Gruber, 1992)

AAC ¼ ½ðAA;120  AB;120 Þ=ðAB;0  AB;120 Þ  1000; where AA,120 and AB,120 are the absorbances of the sample with added antioxidant and the blank sample, respectively, at t = 120 min; AB,0 is the absorbance of a blank sample at t = 0 min. 2.4. Evaluation of antioxidant activity in rapeseed oil The oxidative deterioration of oil was monitored under Schaal Oven Test conditions. The extracts (0.01%, w/w) were added directly to 25.00 ml of rapeseed oil and mixed on a magnetic stirrer for 10 min at 50 °C. Three replicates were prepared for each sample. The samples were placed in open 150 ml beakers in a ventilated oven HS122A (ZPA, Hungary) and kept at 55 °C. A blank sample was prepared under the same conditions without addition of antioxidant. To evaluate the antioxidant activity of the plant extracts, the rate of autoxidation of rapeseed oil was assessed by the increase of the peroxide value (PV), which was measured by using method Cd 8-53 of the American Oil Chemists’ Society and by measuring the UV absorption at 232 and 268 nm after appropriate dilution to obtain a reading between 0.2 and 0.8 AU. 2.5. DPPH assay Radical scavenging activities of isolated pure compounds against the stable radical DPPH were measured using the method of Von Gadow, Joubert, & Hansmann (1997). The concentrations of compounds tested and DPPH were taken on a molar basis, and the concentrations of compounds (EC50, (mol/l antiox)/(mol/l DPPH)) needed to scavenge 50% of the DPPH (Yen & Duh, 1994) were determined. For reasons of clarity the antiradical scavenging power (ARP) of antioxidants, calculated as 1/EC50, are reported. The higher the ARP, the more efficient the antioxidant. 2.6. ABTS assay The ABTS+ radical cation was produced by oxidising ABTS with potassium persulfate (Ahmad, Hakim, & Shehata, 1983). The assay was performed as described earlier (Pukalskas et al., 2002). The TEAC values at the specific time points were calculated from the comparison of activities of isolated compounds with Trolox. 2.7. HPLC–DPPH conditions and instrumentation The on-line DPPH scavenging test was similar to the one those used earlier by our group (Dapkevicius, van Beek, & Niederländer, 2001; Niederländer, van Beek, Bartasiute, & Koleva, 2008) on an HPLC system equipped with a Waters 600E multisolvent delivery

A. Pukalskas et al. / Food Chemistry 130 (2012) 695–701

system (Millipore Corp., Waters Chromatography Division, Milford, MA), and an autosampler Model 231 (Gilson Medical Electronics, Middleton, WI). Separation of compounds was carried out on a 25 cm  4.6 mm i.d. end-capped Alltima C18 analytical column (Alltech Associates, Deerfield, IL). The linear binary gradient for horehound was formed at a constant flow rate of 0.8 ml/min. Solvent A was water-acetonitrile (98:2), and solvent B was 100% acetonitrile. Separation of the methyl tert-butyl ether fraction was performed as follows: initial isocratic conditions of 80% of solvent A for 5 min were followed by an increase to 50% of solvent B during 20 min, then isocratic conditions for 5 min, followed by an increase of solvent B to 100% during 10 min and holding the isocratic conditions for 5 min. Finally the gradient was returned to its initial conditions in 5 min and the column was re-equilibrated during 5 min. Separation of the butanol fraction was performed as follows: initial isocratic conditions of 85% of solvent A for 35 min were followed by an increase to 80% of solvent B during 10 min, then isocratic conditions for 5 min. Finally the gradient was returned to its initial conditions in 5 min and the column was re-equilibrated during 5 min. Compounds eluting from the column were detected with a Waters 990 series Photodiode Array Detector (Millipore Corp. Waters Chromatography Division, Milford, MA) over the range 210–450 nm. Data were processed with Waters software, version LCA-6.22a. After the separation and detection a 104 M solution of DPPH in methanol was added with a 45 ml laboratory-made syringe pump (Free University, Amsterdam, The Netherlands) at a flow rate of 0.70 ml/min. The mixture was continuously introduced into a 15 m  0.25 mm i.d. reaction coil and the decrease in absorbance of a DPPH solution was measured at 517 nm with a 759A model absorbance detector (Applied Biosystems, Foster City, CA) equipped with a tungsten lamp. 2.8. Isolation of active compounds from horehound 2.8.1. Methyl tert-butyl ether fraction One active compound was detected by on-line DPPH radical scavenging in the methyl tert-butyl ether fraction of the methanol–water–acetic acid extract of horehound. The fractionation of this material was carried out on a 15  4 cm i.d. column packed with Sephadex LH-20. Fraction (0.48 g) was loaded on the column and pure methanol was used as the mobile phase. A total of 50 fractions (4–5 ml each) were collected. All fractions were checked for purity on TLC afterwards by spraying them with a 0.2% DPPH solution in methanol. The active fractions 29–37 appeared to be identical, and they were combined and checked for purity with HPLC-DAD using the same gradient as for the methyl tert-butyl ether fraction. According to its HPLC profile and UV spectrum, fraction 29–37 appeared to be pure. It was evaporated until dryness with a rotary evaporator yielding 15 mg of compound 1. 2.8.2. Butanol fraction This fraction was separated on a silica gel column. Two grams were loaded on 100 g of silica gel and eluted with a mixture of chloroform–methanol–water (60:22:4) till fraction 112; with a mixture of chloroform–methanol–water (6:4:1) till fraction 144; with methanol till fraction 163, and finally with a mixture of methanol–water (1:1). In total 192 tubes (15–20 ml each) were collected. After checking all fractions for purity by silica gel TLC and detecting active compounds by spraying with DPPH, most of the fractions were found to have radical scavenging activity. Similar active fractions were combined and collected. Fractions obtained after combining were evaporated till dryness in a rotary evaporator.

697

Tubes 25–28 (59.7 mg) were further chromatographed on a 20  2 cm i.d. column loaded with Sephadex LH-20. The material was eluted with methanol–water (1:1). 26 fractions were collected in total. All the fractions were checked by HPLC (analysis conditions described in Section 2.7) for purity and presence of the wanted compound. Fractions 18 and 19 were combined, and the solvent was evaporated with a rotary evaporator at a temperature of 45 °C yielding 5 mg of compound 2. Tubes 36–47 (123 mg) were re-chromatographed on a 20  2 cm i.d. column loaded with Sephadex LH-20. The material was eluted with methanol–water (1:1) and 18 fractions (6–8 ml each) were collected. After checking on HPLC (analysis conditions described in Section 2.7) fractions 11 and 12 were combined and evaporated to dryness in a rotary evaporator yielding 27 mg of compound 3. Tubes 60–80 (286 mg) were re-chromatographed on a 20  2 cm i.d. column loaded with Sephadex LH-20. The material was eluted with methanol–water (2:3) and 23 fractions were collected in total. After checking on HPLC (analysis conditions described in Section 2.7) fractions 5–6 and 8–9 were combined and evaporated to dryness in a rotary evaporator. From the combined fraction 5–6 compound 5 (9 mg) was obtained. Fraction 8–9 (88 mg) consisted of two compounds. One of the compounds was the previously isolated 3. To separate the unknown compound the material was again loaded on a Sephadex column and eluted with methanol–water (1:4). Sixteen fractions were collected and checked with HPLC. Fraction 12 was evaporated till dryness yielding 24 mg of compound 4. 2.9. Spectral data of isolated compounds See Supplementary data. 3. Results and discussion 3.1. Antioxidant activity of horehound extracts and their fractions There are several fast and simple methods to preliminary screen plant extracts for their antioxidant activity. The results of such a screening may be used to select the most effective extracts for a more comprehensive examination. Such frequently used methods as determination of peroxide value, spectrophotometric measurement of conjugated dienes, b-carotene–linoleate cooxidation and DPPH radical scavenging assays were selected in this study for the assessment of horehound extracts and their fractions in rapeseed oil. None of these methods on its own is ideal for evaluating antioxidant activity but together, while realising their pros and cons, they nonetheless provide useful information on extracts and individual compounds. The rapeseed oil oxidation test comes closest to real foods but is slow, is less suitable for polar antioxidants and the intrinsic instability of peroxides may cause problems. The fast and easy DPPH radical scavenging assay is fully artificial (lacks even a substrate) but nonetheless provides a useful and reproducible screen. All powerful natural antioxidants are detected with this assay. It is less suitable for very hindered phenolics (e.g. BHT), the solvent plays a significant role and the activity is not always linear with concentration due to different reaction mechanisms. The b-carotene-linoleate oxidation test is in between the other two tests in terms of speed and easiness. The polyenic substrate and antioxidants compete for the various radicals formed during the oxidation of a lipid. The assay is not always reproducible due to the presence of an emulsion system, i.e. partitioning effects play a role. Koleva et al. compared methods for screening of plant extracts for antioxidant activity (Koleva, van Beek, Linssen, de Groot, & Evstatieva, 2002).

698

A. Pukalskas et al. / Food Chemistry 130 (2012) 695–701

Crude extracts usually contain a great number of components, which may possess different antioxidant or in some cases prooxidant activities as well as be neutral in terms of their effects on oxidation and/or radical scavenging processes. Therefore, simple fractionation of horehound extracts by liquid–liquid extraction was performed to separate different classes of compounds and find the most effective fractions. The main characteristics of the extracts isolated by two different extraction solvents and their fractions can be found in Table 1. The results of the b-carotene bleaching assay (Table 1) revealed that the highest activity in the linolenic acid model system was exhibited by the methanol–water extract of horehound (AAC = 810) and its butanol (AAC = 818) and water (AAC = 810) fractions. The acetone extract and its fractions possessed lower activities. Also Erdogan Orhan et al. reported a relatively weak activity of the acetone extract of M. vulgare in the DPPH assay (Erdogan Orhan et al., 2010). The results indicate that the fractions of higher polarity were the most active ones in this system. In the DPPH assay the most effective radical scavengers were butanol fractions obtained from acetone and methanol–water extracts. Butanol fractions of aromatic and medicinal plants are often rich in flavonoid glycosides. For instance, the radical scavenging capacity of the butanol fraction of the methanol–water extract at the applied concentration was similar to that of rosmarinic acid, which is a powerful natural antioxidant. It also should be noted that the results of these two assays are quite different, showing once again that the method selected for the assessment of plant extracts plays an important role in antioxidant activity measurements. In addition, all fractions of the methanol–water extract were tested in the accelerated rapeseed oil oxidation system at 55 °C (Table 2). It was found that this extract and its butanol fraction had no effect in this system. Other fractions showed only a slight effect on the oxidation process, while only the water fraction significantly retarded oil oxidation. Hexane and methyl tert-butyl ether fractions of horehound methanol–water extract had negative effects on the oil oxidation process, i.e. they increased peroxide values in the course of oil storage. There was a good correlation between the results obtained by the peroxide value and conjugated dienes measured by UV absorbance methods (results not shown). It may be concluded that the extracts of horehound and their fractions did not possess remarkable activity in oil oxidation tests. All tested fractions were considerably less active than the synthetic antioxidant BHT. Most likely, the extracts isolated by using polar solvents were not evenly distributed in the hydrophobic environment. This is in agreement with previously obtained results (Bandoniene, Pukalskas, Venskutonis, & Gruzdiene, 2000; Miliauskas, van Beek, Venskutonis, Linssen & de Waard, 2004).

Table 2 Evaluation of the antioxidant activity of horehound methanol–water extract and its fractions through the formation of peroxides in rapeseed oil stored at 55 °C. Additive

Induction period (IP), h

Protection factor (PF)

Ranking of PF

Blank (no additives) BHT 0.02% Crude extract 0.1% Hexane fraction 0.1% Methyl tert-butyl ether fraction 0.1% Butanol fraction 0.1% Water fraction 0.1%

173 222 175 112 132

– 1.28 1.01 0.65 0.76

– Very low Very low Prooxidation Prooxidation

175 189

1.01 1.09

Very low Very low

3.2. Identification of horehound constituents Horehound fractions were also analysed using on-line HPLCradical scavenging detection (Niederländer et al., 2008). In the most active butanol fraction four compounds (compounds 2, 3, 4, 5) possessing DPPH radical scavenging activity were detected (Fig. 1). In addition, one additional radical scavenger was detected in the methyl tert-butyl ether fraction of the methanol–water– acetic acid extract of Marrubium vulgare (compound 1). The (+)-ESI-MS of compound 1 showed pseudomolecular ions at m/z 315 [M+H]+ and m/z 337 [M+Na]+, while a peak at m/z 313 [MH] was observed in negative mode. These data clearly indicated a molecular weight of 314 Da. The UV spectrum of 1 had three maxima at 217, 285 and 332 nm. Two singlets in the 1HNMR spectrum at 3.91 and 3.99 ppm, each integrating for 3H, were assigned to methoxyl groups. Two singlets at 6.69 ppm and 6.88 ppm integrating for 1H each correspond to two non-coupled aromatic protons. Two doublets (J = 8.9 Hz) at 7.13 ppm and 8.04 ppm, integrating for 2H each, are typical of a 1,4-disubstituted benzene ring and one singlet at 12.65 ppm, integrating for 1H, corresponds to a hydroxyl proton involved in hydrogen bonding. These signals agree with a flavone skeleton with a 1,4-disubstituted B ring and a trisubstituted A ring (Markham & Geiger, 1993). After analyzing the HMBC spectrum of this compound and taking into account the 13C-NMR assignments found in the literature (Horie et al., 1998), it was concluded that 1 is 5,6-dihydroxy-7,40 -dimethoxyflavone (Fig. 2). 1H and 13C NMR spectral assignments of 1 can be found in the Supplementary material. This compound (syn. ladanein) has been recently reported by Alkhatib et al. in M. vulgare. Possibly it has antileukemic activity (Alkhatib et al., 2010). Compound 2 was isolated from the butanol fraction of the methanol–water–acetic acid extract of horehound. The UV spectrum of 2 showed maxima at 252, 266 (sh) and 342 nm and was characteristic for a flavone. The ()-ESI mass spectrum of

Table 1 Yields and antioxidant activities of horehound methanol–water and acetone extracts and their fractions. Extraction solvent

Fraction/compound

Yield, %

AAC

DPPH scavenging, %

MeOH 80%water 19%acetic acid 1%

Crude Hexane Methyl tert-butyl ether Butanol Water

20.1 1.0 2.9 2.1 10.5

810 774 678 818 810

58.2 ± 0.7 6.1 ± 0.2 30.0 ± 0.7 87.9 ± 0.3 22.2 ± 0.8

Acetone

Crude Hexane Methyl tert-butyl ether Butanol Water

10.9 2.1 1.2 1.9 0.3

657 647 778 548 643

19.5 ± 0.2 5.7 ± 0.1 48.5 ± 0.5 71.3 ± 0.3 32.7 ± 0.3

882 870

88.6 ± 0.2 68.9 ± 0.6

Rosmarinic acid (reference) BHT (reference)

699

A. Pukalskas et al. / Food Chemistry 130 (2012) 695–701

Fig. 1. Combined on-line HPLC-UV and DPPH chromatograms of butanol fraction of horehound methanol–water extract. Upper chromatogram: UV absorbance, lower chromatogram: DPPH radical scavenging activity (RSA).

3' 2' 8

H3CO

9

O

6

OH

2'

1' 2

5

3'

1

7

HO

OH

OCH3 4' 5'

6'

3

10

R

1

8

O

O

5'

1' 6'

6

1

2

7

4

O

OH 4'

4

3

5

OH

O

2 R = Glucose 5 R=Glucuronic acid Fig. 2. Flavonoids isolated from horehound: 5,6-dihydroxy-7,40 -dimethoxyflavone (1); 7-O-b-glucopyranosyl luteolin (2); 7-O-b-glucuronyl luteolin (5).

compound 2 showed a pseudomolecular ion peak [MH] at m/z 447. The 1H-NMR spectrum showed two singlets at 6.67 and 7.39 ppm integrating for 1H each, corresponding to non-coupled aromatic protons, two doublets at 6.43 and 6.76 ppm with J = 2 Hz, integrating for 1H each, probably corresponding to two protons in meta position, and two doublets at 6.89 and 7.40 ppm (J = 8.6 Hz), integrating for 1H each and corresponding to two protons in ortho position. A doublet at 5.02 ppm suggested the anomeric proton (H-1) of a sugar unit. According to the 1H and 13 C spectral data, the sugar is glucopyranose. The large coupling (7.3 Hz) indicated the b configuration of the sugar moiety. To determine the exact structure of the aglycone part of the molecule an HMBC spectrum was recorded. The HMBC spectral data in combination with the 1H and 13C spectra revealed a luteolin moiety. An interaction between the sugar H-1 and the flavone C-7 showed that the sugar is attached to C-7. On the basis of the above data 7-O-b-glucopyranosyl luteolin was proposed as the structure for 2 (Fig. 2). This compound has been reported from numerous plants. The 13C NMR spectral data obtained for 2 are in agreement with the literature (Markham, Whitehouse, & Webby, 1987). 1H and 13C NMR spectral assignments of 2 can be found in the Supplementary material. Compound 5, isolated from the butanol fraction of horehound, had a similar UV spectrum as 2, with maxima at 255, 266 (sh)

and 349 nm. The ()-ESI-MS spectrum recorded in negative mode showed a pseudomolecular ion peak [MH] at m/z 461, i.e. 14 Da higher than that of 2. This can be explained by the presence of an additional CH2 group or by the replacement of CH2OH by COOH. The aglycone part of its 1H NMR spectrum was identical to that of 2. The peak at 175.5 ppm in the 13C NMR spectrum indicated that the sugar must be an uronic acid. Through a closer study of the NMR data the sugar was identified as glucuronic acid. As there was a cross peak between H-1 of glucuronic acid and C-7 of the aglycone part, 7-O-b-glucuronyl luteolin was proposed as the final structure (Fig. 2). This compound has been reported earlier from thyme and Tanacetum species. 1H and 13C NMR spectral assignments of 5 can be found in the Supplementary material. Compound 3, isolated from the butanol fraction of horehound had a UV spectrum similar to that of caffeic acid with maxima at 219, 242 (sh), 302 (sh) and 331 nm. The ()-ESI mass spectrum showed a pseudomolecular ion peak [MH] at m/z 623. The 1H NMR showed signals in the low field part of the spectrum in close agreement with those of caffeic acid and its derivatives. Another part of the spectrum was attributable to 2-phenylethanol. A doublet with a small coupling constant at 4.78 ppm was recognized as an anomeric proton and a doublet at 0.96 ppm as the methyl group of rhamnose. The small coupling constant of the anomeric proton indicated the a configuration for rhamnopyranose. The

700

A. Pukalskas et al. / Food Chemistry 130 (2012) 695–701

OR

O 7

2

HO

9 8

4

O

5

1

3

HO

6 4

O

1

2' 8'

2

O

6

O

3

1'

7'

3'

OH

4' 6'

5

1' 6'

H3C

5'

O

OH

OH

5'

2'

3'

HO

3 4

4'

R= H R = Apiose

OH

OH

Verbascoside Forsythoside B

Fig. 3. Structures of verbascoside (3) and forsythoside B (4).

Table 3 Radical scavenging activities of compounds isolated from horehound in DPPH and ABTS+ assays. No.

Compound

TEAC1

1 2 3 4 5 6

5,6-Dihydroxy-7,40 -dimethoxyflavone (1) 7-O-b-Glucopyranosyl luteolin (2) Verbascoside (3) Forsythoside B (4) 7-O-b-Glucuronyl luteolin (5) Rosmarinic acid (reference)

1.07 0.94 1.43 1.47 0.95 1.49

anomeric proton at 4.37 belongs to glucose. The large coupling constant (J = 7.9 Hz) is in accordance with the b configuration. In the spectrum of compound 3, a triplet at 4.65 ppm, attributed to H-4 of glucose, was observed. For glucobioses normally only the anomeric proton can be observed above 4 ppm. Thus the signal at 4.65 ppm reflects the esterification with caffeic acid of the 4OH of glucose. To confirm all links between the sugars and the aglycone part of the molecule an HMBC spectrum was recorded. Cross peaks between glucose H-4 and caffeic acid C-9, glucose H3 and rhamnose C-1, and H-80 of the aglycone and C-1 of the glucose confirm that compound 3 is verbascoside (Fig. 3). The spectral data are in agreement with those presented in the literature (Andary, Wylde, Laffite, Privat, & Winternitz, 1982). 1H and 13C NMR spectral assignments of 3 can be found in the Supplementary material. The UV spectrum of compound 4 was similar to that of caffeic acid with maxima at 221, 245 (sh), 301 (sh) and 330 nm. The ESI mass spectrum, recorded in negative mode, showed a pseudomolecular ion peak at m/z 755. The NMR spectral data were very similar to those recorded for verbascoside. An anomeric proton signal at 4.94 ppm must belong to an additional sugar unit. Taking into account all the 1H and 13C NMR data, the additional sugar was identified as b-apiose. In the HMBC spectrum a cross peak between H-1 of apiose and C-6 of glucose was present. This confirms that compound 4 is forsythoside B. The structure of the compound is presented in Fig. 3. All NMR spectral data are in agreement with the literature data reported for this compound (Calis, Hosny, Khalifa, & Ruedi, 1992). 1H and 13C NMR spectral assignments of 4 can be found in the Supplementary material. 3.3. Radical scavenging activities of isolated compounds All isolated compounds were tested in DPPH and ABTS+ assays (Table 3) and their radical scavenging activity was compared with that of the well known natural antioxidant rosmarinic acid. It was

min

TEAC6 1.31 1.48 1.65 1.66 1.25 1.54

min

ARP30

min

2.45 1.85 2.76 2.47 5.22 6.9

found that in polar media (ABTS+ assay) the activity of verbascoside (3) and forsythoside B (4) is similar to that of rosmarinic acid, which is due to the similarities in the aglycone parts of these compounds. However, in the DPPH assay rosmarinic acid was approximately a three times stronger radical scavenger than horehound compounds 3 and 4. The activities of luteolin glycosides in the ABTS+ assay were quite similar, which was expected because the aglycone is identical. Surprisingly, in the DPPH assay the activity of 7-O-b-glucuronyl luteolin 5 was about twice higher than that of its glucopyranosyl analogue. It suggests that the sugar substituent can play an important role in defining the ability of compound to scavenge free radicals. No literature data about this phenomenon was found. 4. Conclusions Horehound extracts exhibited antioxidant activity in oils and radical scavenging capacity in model reaction systems. The activity mostly depends on the applied extraction solvent. Methanol–water extraction was found to be a good method for the isolation of a broad range of polar compounds possessing antioxidant properties, however polar extracts are less suitable for their application in lipophilic media. Five compounds possessing radical scavenging activity were isolated from horehound. All isolated compounds were identified using 1D and 2D NMR and MS techniques. The compounds were mainly glycosides and as such they were rather polar. Verbascoside and forsythoside B were not previously reported in Marrubium vulgare. All identified compounds were comparatively strong DPPH and ABTS+ radical scavengers. Direct application of these polar glycosides in oils would be rather ineffective due to their low solubility. However, extracts containing these compounds might find an application in more polar heterogeneous systems, such as emulsions.

A. Pukalskas et al. / Food Chemistry 130 (2012) 695–701

Acknowledgements This work was partially supported by Lithuanian State Science and Studies Foundation and Agency for International Science and Technology Development Programmes in Lithuania (COST Action B35). Another part of this work was financed by Wageningen International Agricultural Centre. Also the authors are very grateful to Elbert J. C. van der Klift for help with the HPLC equipment. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.foodchem.2011.07.112. References Ahmad, M. M., Hakim, S. H., & Shehata, A. A. Y. (1983). The behaviour of phenolic antioxidants, synergists and their mixtures in two vegetable oils. Fette Seifen Anstrichmittel, 85(21), 479–485. Alkhatib, R., Joha, S., Cheok, M., Roumy, V., Idziorek, T., Preudhomme, C., et al. (2010). Activity of ladanein on leukemia cell lines and its occurrence in Marrubium vulgare. Planta Medica, 76, 86–87. Andary, C., Wylde, R., Laffite, C., Privat, G., & Winternitz, F. (1982). Structures of verbascoside and orobanchoside, caffeic acid sugar esters from Orobanche rapum-genistae. Phytochemistry, 21(5), 1123–1127. Bandoniene, D., Pukalskas, A., Venskutonis, P. R., & Gruzdiene, D. (2000). Preliminary screening of antioxidant activity of some plant extracts in rapeseed oil. Food Research International, 33(9), 785–791. Calis, I., Hosny, M., Khalifa, T., & Ruedi, P. (1992). Phenylpropanoid glycosides from Marrubium alysson. Phytochemistry, 31(10), 3624–3626. Chevolleau, S., Mallet, J. F., Ucciani, E., Gamisans, J., & Gruber, M. (1992). Antioxidant activity in leaves of some Mediterranean plants. Journal of the American Oil Chemists Society, 69, 1269–1271. Cuvelier, M. E., Richard, H., & Berset, C. (1996). Antioxidative activity and phenolic composition of pilot-plant and commercial extracts of sage and rosemary. Journal of the American Oil Chemists Society, 73(5), 645–652. Dapkevicius, A., van Beek, T. A., & Niederländer, H. A. G. (2001). Evaluation and comparison of two improved techniques for the on-line detection of antioxidants in liquid chromatography eluates. Journal of Chromatography A, 912(1), 73–82. Dapkevicius, A., Venskutonis, R., van Beek, T. A., & Linssen, J. P. H. (1998). Antioxidant activity of extracts obtained by different isolation procedures from some aromatic herbs grown in Lithuania. Journal of the Science of Food and Agriculture, 77(1), 140–146. Erdogan Orhan, I., Belhattab, R., S enol, F. S., Gülpinar, A. R., Hosbas, S., & Kartal, M. (2010). Profiling of cholinesterase inhibitory and antioxidant activities of Artemisia absinthium, A. herba-alba, A. fragrans, Marrubium vulgare, M. astranicum, Origanum vulgare subsp. glandolossum and essential oil analysis of two Artemisia species. Industrial Crops and Products, 32, 566–571. Horie, T., Ohtsuru, Y., Shibata, K., Yamashita, K., Tsukayama, M., & Kawamura, Y. (1998). 13C NMR spectral assignment of the A-ring of polyoxygenated flavones. Phytochemistry, 47, 865–874.

701

Khanavi, M., Ghasemian, L., Motlagh, E. H., Hadjiakhoonid, A., & Shafiee, A. (2005). Chemical composition of the essential oils of Marrubium parviflorum Fisch & C.A. and Marrubium vulgare L. from Iran. Flavour and Fragrance Journal, 20(3), 324–326. Knoess, W. (1994). Furanic labdane diterpenes in differentiated and undifferentiated cultures of Marrubium vulgare and Leonurus cardiaca. Plant Physiology and Biochemistry, 32, 785–789. Koleva, I. I., van Beek, T. A., Linssen, J. P. H., de Groot, Ae., & Evstatieva, L. N. (2002). A comparison of three methods for screening of plant extracts for antioxidant activity. Phytochemical Analysis, 13, 8–17. Kovalewski, Z., & Matlwaska, L. (1978). Flavonoids in Marrubium vulgare herbage. Herba Polonica, 24, 183–186. Marco, G. J. (1968). A rapid method for evaluation of antioxidants. Journal of the American Oil Chemists Society, 45, 594–598. Markham, K. R., & Geiger, H. (1993). 1H nuclear magneic resonance spectroscopy of flavonoids and their glycosides. In J. B. Harborne (Ed.), Flavonoids, advances in research since 1986 (pp. 410). London: Chapman and Hall. Markham, K. R., Whitehouse, L. A., & Webby, R. F. (1987). Luteolin 7-O-[6-O-a-Larabinofuranosyl]b-D-glucopyranoside and other new flavonoid glycosides from New Zealand Dacrydzum species. Journal of Natural Products, 50(4), 660–663. Miliauskas, G., van Beek, T. A., Venskutonis, P. R., Linssen, J. P. H., & de Waard, P. (2004). Antioxidative activity of Geranium macrorrhizum. European Food Research and Technology, 218(3), 253–261. Miliauskas, G., Venskutonis, P. R., & van Beek, T. A. (2004). Screening of radical scavenging activity of some medicinal and aromatic plant extracts. Food Chemistry, 85(2), 231–237. Morteza-Semnani, K., Saeedi, M., & Babanezhad, E. (2008). The essential composition of Marrubium vulgare L. from Iran. Journal of Essential Oil Research, 20(6), 488–490. Nawwar, M. A. M., El-Mousalamy, A. M. D., Barakat, H. H., Buddrus, J., & Linscheid, M. (1989). Flavonoid lactates from leaves of Marrubium vulgare. Phytochemistry, 28, 3201–3206. Niederländer, H. A. G., van Beek, T. A., Bartasiute, A., & Koleva, I. I. (2008). Antioxidant activity assays on-line with liquid chromatography. Journal of Chromatography A, 1210(2), 121–134. Popa, D. P., & Pasechnick, G. S. (1975). The structure of vulgarol – A new diterpenoid from Marrubium vulgare. Khimiya Prirodnykh Soedinenii, 11, 752–756. Pukalskas, A., van Beek, T. A., Venskutonis, R. P., Linssen, J. P. H., van Veldhuizen, A., & de Groot, A. (2002). Identification of radical scavengers in sweet grass (Hierochloe odorata). Journal of Agricultural and Food Chemistry, 50(10), 2914–2929. Puziene, G., & Gudanavicius, S. (1978). Data on the phenological investigations of common horehound. Medicina, 16, 130–136. Rey, J. P., Levesque, J., & Pousset, J. L. (1992). Extraction and high performance liquid chromatographic methods for the c-lactones parthenolide (Chrysanthemum parthenium Bernh., Marrubium vulgare L.) and artemisinin (Artemisia annua L.). Journal of Chromatography, 605, 124–128. Von Gadow, A., Joubert, E., & Hansmann, C. F. (1997). Comparison of the antioxidant activity of aspalathin with that of other plant phenols of rooibos tea (Aspalathus linearis), alpha-tocopherol, BHT, and BHA. Journal of Agricultural and Food Chemistry, 45(3), 632–638. Weel, K. G. C., Venskutonis, P. R., Pukalskas, A., Gruzdiene, D., & Linssen, J. P. H. (1999). Antioxidant activity of horehound (Marrubium vulgare L.) grown in Lithuania. Fett Lipid, 101(10), 395–400. Yen, G. C., & Duh, P. D. (1994). Scavenging effect of methanolic extracts of peanut hulls on free-radical and active-oxygen species. Journal of Agricultural and Food Chemistry, 42, 629–632.