Glucosinolate profiles in the seeds, leaves and roots of rocket salad (Eruca sativa Mill.) and anti-oxidative activities of intact plant powder and purified 4-methoxyglucobrassicin

Abstract

Six desulfo-glucosinolates (DS-GSLs), DS-glucoraphanin, DS-4-(β-d-glucopyranosyldisulfanyl)butyl GSL, DS-glucoerucin, DS-glucobrassicin, DS-dimeric 4-mercaptobutyl GSL and DS-4-methoxyglucobrassicin, were isolated from rocket salad (Eruca sativa Mill.) By using high-pressure liquid chromatography (HPLC) and identified by their retention times on HPLC and electrospray ionization mass spectrometry (ESI-MS) techniques. Total GSL contents were 125, 11 and 22 µmol g⁻¹ dry weight (DW) in seeds, leaves and roots, respectively. Two DS-GSLs, namely DS-glucoraphanin and DS-glucoerucin, were the major GSLs in seeds and roots, while DS-glucoraphanin, DS-glucoerucin and DS-dimeric 4-mercaptobutyl GSL were the predominant GSLs in leaves (> 1 µmol g⁻¹ DW or > 10% of the total GSL content). In addition, the anti-oxidant activity (n = 3) of intact leaf and root powders, whole DS-GSL eluates and purified DS-4-methoxyglucobrassicin were measured using the XYZ-dish method. DS-4-methoxyglucobrassicin showed the greatest anti-oxidant activity value (3.8 unit g⁻¹ DW), and the value for whole DS-GSL eluates (1.6) was higher than that of intact leaves (1.5) and roots (0.87).

Key words:

• anti-oxidative activity
• Eruca sativa
• glucosinolates
• LC/MS
• rocket salad

INTRODUCTION

Rocket salad (Eruca sativa Mill.), commonly known as “rucola” in Italy and “kibana-suzushiro” in Japan, is an annual herb belonging to the Brassicaceae family. With its spicy hot flavor, rocket is widely consumed fresh as salad, prepared as a steamed vegetable or used as a spice or food ingredient in European countries. In Asia, particularly in India and Pakistan, the plant serves as an important source of oil and its seed meal is commonly used as animal feed because it is a rich source of protein despite its high glucosinolate (GSL; Fig. 1) and erucic acid contents (Ashraf and Noor 1992; Pita Villamil et al. 2002). Most studies have focused on the effects of nitrate and ammonium nutrition on nitrate content in rocket (Bianco et al. 1998; Santamaria and Gonnella 2001; Santamaria et al. 1998; Santamaria et al. 2002), as well as on its GSLs and their breakdown products, particularly the isothiocyanates (Bennett et al. 2002; Cerny et al. 1996; Jirovetz et al. 2002; Miyazawa et al. 2002; Weckerle et al. 2001). When crushed, rocket leaves release a strong smell of horseradish combined with sesame seeds. The characteristic pungent or bitter taste and flavor of leafy vegetables, particularly those belonging to the Brassicaceae (i.e. Cruciferae) family may be related to the presence of GSLs and their associated hydrolytic products: isothiocyanates, thiocyanates and nitriles.

Over 120 individual GSLs, a group of sulfur-containing secondary metabolites have been identified from 11 plant families of dicotyledonous angiosperms, particularly those from the Brassicaceae (Fahey et al. 2001). Approximately twenty of these GSLs, possessing the same basic structure (Fig. 1a), and differing only in their side chains (R), are commonly present in appreciable amounts in any particular plant (Olsen and Sørensen 1981). The GSLs are named according to their R-group and aglycone moieties, whose biosynthetic derivation arises from a few known amino acids. Aliphatic GSLs, bearing an

Figure 1  Enzymatic desulfation of glucosinolates. (a) Glucosinolate; (b) desulfated natural glucosinolate.

alkyl or alkenyl side chain, are derived from methionine, aromatic-GSLs from phenylalanine (and possibly tyrosine) and indolyl GSLs from tryptophan (Fahey et al. 2001).

In the plant, intact GSLs and their breakdown products constitute part of the general defense mechanisms against herbivores and microorganisms (Wallsgrove et al. 1999). The specific biological activity of different GSLs varies considerably, thus, it is important to precisely identify and quantify the specific GSLs in any tissue under investigation (Kiddle et al. 2001). In recent years, various GSLs have been isolated and identified from the seeds and leaves of rocket salad. (Bennett et al. 2002; Cerny et al. 1996; Iori et al. 1999; Jirovetz et al. 2002; Kim et al. 2004). However, in our experience, a number of minor GSLs consistently separated using high-pressure liquid chromatography (HPLC) have not yet been identified. Fortunately, two indolyl GSLs were detected from the concentrated extracts of rocket leaves according to their retention times, but not characterized in detail (Kim et al. 2004). Moreover, there have been several previous attempts to obtain GSL profiles from different tissues within one plant. Major differences in the relative amount of individual GSLs have been observed between plant parts, indicating that individual GSLs have defined distribution patterns (Macfarlane-Smith and Griffiths 1988; Sang et al. 1984).

The isothiocyanates are an important group of hydrolytic products of GSLs and are potentially considered to possess anti-carcinogenic properties. In particular, sulforaphane (the isothiocyanate of glucoraphanin) and indole-3-carbinol (the isothiocyanate of glucobrassicin) are believed to confer beneficial effects on human health through a variety of mechanisms (Bradfield and Bjeldanes 1984; Fahey et al. 2001; Gross et al. 2000; Mithen et al. 2000; Nho and Jeffery 2004). Anti-oxidative activity, particularly free radical scavenging activity, is important because of the putative role of free radicals as cancer-causing agents in both foods and biological systems. Therefore, in a biological system, anti-oxidative properties are frequently used as an index of the activity of a defense mechanism against active oxygen species or free radicals. At the present time, there is an increased interest in non-synthetic anti-oxidants of plant origin, including ascorbic acid, carotenoids, flavonoids, phenolic compounds and vitamins. To date, only the anti-oxidative properties of whole extracts and purified GSLs from brussels sprouts (Brassica oleracea L. var. gemmifera Zenker), broccoli (Brassica oleracea L. var. italica Plenck.), red cabbage, cabbage (Brassica oleracea L.) and cauliflower (Brassica oleracea L. var. botrytis L.) have been systematically studied by measuring their effect on lipid peroxidation using the deoxyribose assay, a radical cation 2,2′-azinobis-3-ethyl-benzothiazoline-6-sulfonate (ABTS), and the bleomycin assay (Plumb et al. 1996). The authors concluded that the total anti-oxidative activities of extracts from cooked and autolysed brussels sprouts were not significantly different because the phenolic compounds and purified GSLs expressed only weak anti-oxidative properties. By contrast, Kim et al. (2004) obtained relatively high anti-oxidative activities for two desulfo-glucosinolates (DS-GSLs), DS-4-(β-D-glucopyranosyldisulfanyl)butyl GSL and DS-dimeric 4-mercaptobutyl GSL, purified from rocket leaves using the XYZ-dish chemiluminescence (CL) method (H₂O₂-acetaldehyde system; Yoshiki et al. 1995). Each DS-GSL bore an asymmetrical S–S bond that could serve as an electron donor.

Consequently, the aim of the present study was to identify and characterize GSLs as their desulfated analogs, particularly two indolyl GSLs, present in rocket salad using HPLC and electrospray ionization mass spectrometry (ESI-MS). Different plant tissues, seed, leaf and root, were analyzed to determine their complete GSL profiles and to look for differences between the organs in the relative proportions of individual GSLs. In addition, the anti-oxidative properties of the intact plants and their desulfated analogs were assessed using the XYZ-dish method.

MATERIALS AND METHODS

Plant materials

Rocket (Eruca sativa Mill. cultivar “Odyssey”) seeds were purchased from Sakata Seed Company (Yokohama, Japan). Seeds were kept at 100°C for 30 min in the oven to deactivate endogenous myrosinase and the seeds were defatted using hexane several times and then dried. Seed powder was analyzed for GSL profiles with four replications.

For plant samples, seeds were sown in a plastic pot containing vermiculite on 24 September 2002 and grown in a greenhouse at the National Agricultural Research Center for Hokkaido Region (longitude 141°21′E; latitude 43°04′N). Two weeks later individual seedlings were removed and placed 0.1 m apart in a plastic container (300 mm × 130 mm × 120 mm, 1.8 L). The plantlets were cultivated hydroponically in the greenhouse and nourished with Hoagland-type solutions (Hoagland and Arnon 1938) as follows: (macronutrients: KCl 6 mmol L⁻¹, Ca[NO₃]₂ 2 mmol L⁻¹, NH₄NO₃ 5 mmol L⁻¹, NH₄H₂PO₄ 1 mmol L⁻¹, MgSO₄ 1 mmol L⁻¹; micronutrients: MnCl₂·4H₂O [1.81 g L⁻¹], H₃BO₃[2.86 g L⁻¹], ZnSO₄·7H₂O [0.22 g L⁻¹], [NH₄]₆Mo₇O₂₄·4H₂O [0.09 g L⁻¹], CuSO₄·5H₂O [0.08 g L⁻¹], Fe-ethylendiamine–N, N, N¹, N¹-tetraacetic acid (EDTA) [7.54 g L⁻¹]). The initial pH of the nutrient solution was adjusted to 6.0.

Plants were watered every 2–3 days in the morning with the nutrient solutions. All samples were harvested on 12 November (49 days after sowing) and separated into leaves and roots. Samples were then lyophilized, ground and stored in a plastic bottle until chemical analysis. Four plants per treatment were analyzed to determine GSL contents.

Chemicals

The HPLC-grade acetonitrile (CH₃CN), gallic acid, hydrogen peroxide (H₂O₂, 35%), acetaldehyde (CH₃CHO, approximately 90%) and ODS C18 matrix (Wakosil 40 C18, 30–50 µm particles) for column chromatography were purchased from Wako Pure Chemical Industries (Osaka, Japan). Aryl sulfatase (type H-1, EC 3.1.6.1) was obtained from Sigma-Aldrich Chemical Company (St Louis, MO, USA), and sinigrin (2-propenyl GSL) from Tokyo Kasei Kogyo Company (Tokyo, Japan) as an internal (extraction) standard. DEAE-Sephadex A−25 was supplied by Amersham Biosciences (Uppsala, Sweden).

Desulfo-glucosinolate analysis

Crude GSLs were extracted with 70% (v/v) boiling MeOH from defatted seed meal and ground plants and applied to DEAE-Sephadex A−25. The GSLs were then converted to their DS-GSLs by overnight treatment with aryl sulfatase (Fig. 1b). The eluate containing the desulfated products was analyzed using an HPLC-equipped Inersil ODS-2 column (4.6 mm × 250 mm, GL Science, Tokyo, Japan) at 227 nm as previously described (Kim et al. 2004). Individual DS-GSLs were identified by comparing their retention times using HPLC and MS spectra. DS-4-methoxyglucobrassicin was purified from the leaves of rocket salad using a 40C18 ODS open column (30 cm × 3.0 cm and 80 cm × 1.0 cm internal diameter) after desulfation (Kim et al. 2004).

Mass spectrometry

Analysis of DS-GSLs was carried out on an API-100 instrument (Perkin-Elmer Sciex Instruments, Pomona, CA, USA) equipped with an atmospheric pressure ESI source. The MS operating conditions were as follows: ionspray voltage, 4.8 kV (positive-ion mode); orifice voltage, 40 V; nebulizer gas, air; curtain gas, nitrogen. The purified DS-4-methoxyglucobrassicin directly collected from the HPLC effluent was directly injected in 0.1% AcOH solution into the ESI solvent flow (10 µL min⁻¹) using a loop injection valve. Spectra were scanned in the range m/z 100–1000. The API 100 instrument was used with TUNE software (C-Preliminary Release version) for data acquisition and evaluation (Kim et al. 2004).

Anti-oxidative activity

The anti-oxidative properties of plant powders and the purified GSL were detected by CL using the H₂O₂-acetaldehyde system as previously described in detail (Kim et al. 2004). The anti-oxidative activities were quantitatively measured in comparison with CL intensity (peak area for 1 min at 23°C) of gallic acid (100 µL of 1 mmol L⁻¹/1 mL, 100 µmol L⁻¹ = 0.10 unit g⁻¹) as the synthetic anti-oxidant, and the value (unit g⁻¹) was calculated based on 1 unit = CL intensity of a 1 mmol L⁻¹ gallic acid solution (Iwai et al. 2000).

RESULTS AND DISCUSSION

Separation and identification of DS-GSLs using liquid chromatograph mass spectrometry

The HPLC chromatograms of DS-GSLs separated from defatted seeds, leaves and roots of rocket salad are shown in Fig. 2. The DS-GSLs were identified based on their retention times compared with those in our database. For confirmation, mass spectra of the individual DS-GSLs were obtained using positive-ion ESI-MS analysis. The HPLC profiles were very similar to the HPLC chromatograms, except for indolyl GSLs, reported by Bennett et al. (2002), who used the same solvent mixture and the same DS-GSL extraction method. Their identities, including chemical structures, trivial (semisystematic) names and molecular weights are listed in Table 1. Peak numbers followed the elution order of HPLC profiles are shown in Fig. 2, and their retention times from the HPLC chromatograms shown in Fig. 2b. The DS-GSLs were identified from their protonated molecular ions [M+H]⁺ in the mass spectra of DS-GSLs. Peaks 1 and 3 were identified using ESI-MS as DS-glucoraphanin and DS-glucoerucin and had molecular weights of 357 (m/z 358, [M+H]⁺) and 341 (m/z 342, [M+H]⁺), respectively (data not shown). DS-sinigrin (SIN, peak 2) is commonly used as an external or internal standard. Interestingly, two indolyl GSLs in the previous study (Kim et al. 2004) were detected from the concentrated extracts of rocket leaves according to their retention times compared to our database (i.e. HPLC chromatograms of broccoli, Chinese cabbage, cauliflower, rapeseed and vegetable turnip rape). In the present study, the two indolyl GSLs (peaks 4 and 6) were characterized in detail using ESI-MS and identified as DS-glucobrassicin and DS-4-methoxyglucobrassicin (Fig. 3a,b, respectively). The protonated molecular ion [M+H]⁺ and a group-specific

Table 1 Structures of the desulfo-glucosinolates isolated from rocket salad

Peak no.	Retention time (min)^†	Structure of R-groups^‡	Trivial (semisysternatic) name	Molecular weight	Response factor^§
1	10.33	CH3-SO (CH2)4	DS-glucoraphanin (DS-4-methylsulfinylbutyl GSL)	357	1.07
2^¶	11.09	CH2 = CH-CH2−	DS-sinigrin (DS-2-propenyl GSL)	279	1.00
3	14.59	C6H11O5-S-S-(CH2)4−	DS-4-(β-D-glucopyranosyldisulfanyl)butyl GSL	521	1.00^††
4	19.78	CH3-S-(CH2)4−	DS-glucoerucin (DS-4-methylthiobutyl GSL)	341	1.00^††
5	21.36	Indole-3-CH2	DS-glucobrassicin (DS-3-indolylmethyl GSL)	368	0.29
6	23.10	{–(CH2)4-S-S-(CH2)4−}	DS-dimeric 4-mercaptobutyl GSL	652	1.00^††
7	23.53	Indole-(4-OCH3)-3-CH2−	DS-4-methoxyglucobrassicin (DS-4-methoxy-3-indolylmethyl GSL)	398	0.25
^†The retention times were based on Fig. 2b.
^‡General structure of the sides (R) are shown in Fig. 1.
^§International Organization for Stanardization (ISO 9167-11992).
^¶Peak no. 2, DS-sinigrin was used as an international standard.
^††Undecided by the ISO.

Figure 2  High-pressure liquid chromatography (HPLC) profiles of desulfo-glucosinolates (DS-GSL) isolated from rocket salad (a) seeds, (b) leaves and (c) roots. Peak numbers refer to the GSLs listed in Table 1. 1, DS-glucoraphanin; 2, DS-sinigrin, used as an internal standard; 3, DS-4-(β-D-glucopyranosyldisulfanyl)butyl GSL; 4, DS-glucoerucin; 5, DS-glucobrassicin; 6, DS-dimeric 4-mercaptobutyl GSL; 7, DS-4-methoxyglucobrassicin.

fragment ion of [M+Na]⁺ were the specific peaks used to determine the GSLs.

The identified DS-GSLs consisted of two aliphatic DS-GSLs (DS-glucoraphanin and DS-glucoerucin), two

Figure 3  Molecular ion region of electrospray ionization mass spectrometry of two indolyl desulfo-glucosinolates isolated from rocket salad. (a) Spectrum of DS-glucobrassicin [m/z 369, (M+H)+]; (b) spectrum of DS-4-methoxyglucobrassicin [m/z 399, (M+H)+]. Spectra were recorded between m/z 100 and 1000, but are shown between m/z 100 and 550.

indolyl DS-GSLs and two unique DS-GSLs (DS-4-[β-D-glucopyranosyldisulfanyl]butyl GSL and DS-dimeric 4-mercaptobutyl GSL). The two aliphatic DS-GSLs are, in general, found in significant amounts in rocket (Barillari et al. 2005; Bennett et al. 2002; Cerny et al. 1996). This was the first study to identify DS-glucobrassicin in the leaves of rocket, and DS-4-methoxyglucobrassicin has recently been identified by Bennett et al . (2002). The related compounds bis(4-isothiocyanatobutyl) disulfide and 4-mercaptobutyl GSL or 4-mercaptobutyl isothiocyanate of the dimeric GSL have been detected from leaves (Bennett et al. 2002; Cerny et al. 1996; Kim et al. 2004). The dimeric GSL was firstly identified by Bennett et al. (2002) as a structurally unique compound with two glucoses and an S–S bond and, thus, it was

Table 2 Variation in proportion (expressed as % of total glucosinolate [GSL]) of individual desulfo-GSLs (DS-GSLs) identified in the rocket salad cultivar “Odyssey”

DS-GSL name	Plant part (%)
	Seed	Leaf	Root
DS-glucoraphanin	5.6 ± 0.16	11.4 ± 1.11	13.7 ± 0.55
DS-4-(β-D-glucopyranosyldisulfanyl)butyl GSL	–	6.8 ± 0.74	0.95 ± 0.12
DS-glucoerucin	94.4 ± 0.16	29.8 ± 2.31	81.5 ± 0.97
DS-glucobrassicin	–	0.36 ± 0.05	–
DS-dimeric 4-mercaptobutyl GSL	–	51.6 ± 2.96	3.8 ± 0.39
DS-4-methoxyglucobrassicin	–	0.12 ± 0.01	–
Total DS-GSL content (µmol g^−1 dry weight)	125 ± 6.98	11 ± 0.78	22 ± 2.35
Values are mean ± standard deviation (n = 4).

questioned as to whether it is naturally present in plants or an artifact produced immediately during the extract process. As a DS-GSL, it has a relatively high molecular weight (M _r) of 652 because of its dimeric disulfide structure formed from DS-4-mercaptobutyl GSL. The response factors for the quantification were experimentally determined by the International Organization for Standardization (ISO) in 1992 to calculate the individual DS-GSL content or total GSL contents in rapeseed using HPLC (International Organization for Standardization 1992). Unfortunately, only the response factors of 14 DS-GSLs were listed by the ISO. Therefore, a response factor of 1.00 was assigned to the other DS-GSLs, such as DS-4-(β-D-glucopyranosyldisulfanylbutyl GSL, DS-glucoerucin and DS-dimeric 4-mercaptobutyl GSL.

Proportions of GSLs among different plant organs

The proportions of the individual DS-GSLs to total GSL content in the seeds, leaves and roots are shown in Table 2. The highest total GSL content was observed in the seeds (125 µmol g⁻¹ dry weight [DW]), and this value was approximately six-fold greater than that recorded in the roots (22 µmol g⁻¹ DW) and 12-fold that of the leaves (11 µmol g⁻¹ DW). Given its very low total GSL content, rocket salad is suitable for consumption as a leafy salad vegetable. The European Community fixed the total GSL level of defatted rapeseed meal at below 35 µmol g⁻¹ DW of seed and the recommendation for rapeseed was below 20 µmol g⁻¹ DW from 1991 because of undesirable effects on humans and livestock owing to pungency and goitrogenic activity (Milford and Evans 1991). Predicting the total or individual GSL content in edible parts of rocket, compared to their contents as determined by breeders, would reduce the time necessary for whole plant analysis. Moreover, total GSL contents vary widely in the different organs of any given plant (Kim et al. 2001, 2003; Sang et al. 1984). Therefore, it is difficult to determine which plant organs contain the majority of the total whole plant GSL content (Barillari et al. 2005). As the roles of GSLs in plants become clearer, it may be necessary to monitor the content or the relative proportion of individual GSLs rather than total GSL.

Two DS-GSLs, DS-glucoraphanin and DS-glucoerucin, were found to be the predominant GSLs in both seeds and roots, while three DS-GSLs (DS-glucoraphanin, DS-glucoerucin and DS-dimeric 4-mercaptobutyl GSL) were the main components in the leaves (> 1 µmol g⁻¹ DW or > 10% of the total GSL content). Thus, our study is in agreement with the finding that seeds generally present simpler HPLC profiles together with higher GSL levels compared to leaves (Kim et al. 2001; Sang et al. 1984). Interestingly, DS-dimeric 4-mercaptobutyl GSL, which presented the greatest proportion (52%) of total DS-GSLs in the leaves, was not found in the seeds and only detected in roots in trace amounts (4%). The two indolyl DS-GSLs were found in the leaves as trace amounts. DS-4-(β-D-glucopyranosyldisulfanyl)butyl GSL accounted for 7% of DS-GSLs in the leaves, but only trace amounts (< 1%) in the roots.

Anti-oxidative activity of the new GSL

Anti-oxidants are believed to play a very important role in the human body's defense system against reactive oxygen species such as superoxide radicals, singlet oxygen, H₂O₂ and hydroxyl radicals. The anti-oxidant activities of intact leaf/root powder, whole DS-GSL eluates and purified DS-4-methoxymglucobrassicin were measured using the XYZ-dish method (Table 3). The activity of intact leaf powder (1.5 unit g⁻¹) was greater than that of intact root powder (0.87 unit g⁻¹) even though the total GSL content (11 µmol g⁻¹ DW) of the leaves was half that of the roots (22 µmol g⁻¹ DW) (Table 2). This indicates that not only GSLs, but other phytochemicals such as flavonoids and polyphenols, present anti-oxidant activities (Plumb et al. 1996).

DS-4-methoxyglucobrassicin showed the greatest activity among the samples examined (3.8 unit g⁻¹), two-fold

Table 3 Anti-oxidative activity in rocket salad

Plant part	Activity (unit g^−1)^†
Intact leaf powder	1.5 ± 0.24
Intact root powder	0.87 ± 0.07
Purified desulfo-glucosinolate  (DS-GSL) from the leaves	–
Whole DS-GSL eluates	1.6 ± 0.20
DS-4-methoxyglucobrassicin	3.8 ± 0.54
^†The activity value (unit g^−1) of each sample was calculated as 1 unit = chemiluminescence intensity of a 1 mmol L^−1 gallic acid solution. Values are mean ± standard deviation (n = 3).

greater than that of whole DS-GSL eluates; by contrast, the lowest activity was found in the root powder. In addition, the value of whole DS-GSL eluates (1.6 unit g⁻¹) was significantly higher (two-fold higher) than that of the intact root powder, but similar to that of the intact leaf powder. We have recently reported the importance of the side chain of the parent GSL in conferring anti-oxidative activity (XYZ-dish method) to two purified DS-GSLs (DS-4-[β-D-glucopyranosyldisulfanyl]butyl GSL and DS-dimeric 4-mercaptobutyl GSL) (Kim et al. 2004). Moreover, direct anti-oxidant activity of purified glucoerucin (95%) has recently been investigated in rocket seeds and sprouts by Barillari et al. (2005), and their findings implied that glucoerucin shows good direct anti-oxidant activity because of its ability to decompose hydroperoxides and hydrogen peroxide. The 1,1-diphenyl-2-picrylhydrazyl (DPPH) method has been widely used to evaluate the free radical scavenging activities of various flavonoids and polyphenols in food systems. However, use of the XYZ-dish method has been very limited. Thus, no direct comparison can be made between XYZ-dish and DPPH method-derived assessments of anti-oxidant activity. Recently, the relationship between photon emission and chemopreventive activity in tea (green, oolong and black) was investigated by Yoshiki et al. (2004). The authors concluded that photon intensity was highly correlated with chemopreventive activity against hydrogen peroxide (H₂O₂, R ² = 0.75), superoxide (O⁻ ₂, R ² = 0.80) and the DPPH radical (R ² = 0.75).

ACKNOWLEDGMENT

The authors wish to thank the Japanese Society for the Promotion of Science (JSPS) for financial assistance.