Bioavailability and in vivo metabolism of intact glucosinolates

Journal of Functional Foods 24 (2016) 450–460

Available online at www.sciencedirect.com

ScienceDirect j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / j ff

Bioavailability and in vivo metabolism of intact glucosinolates Jens Christian Sørensen a,*, Heidi Blok Frandsen a, Søren Krogh Jensen b, Niels Bastian Kristensen c, Susanne Sørensen a, Hilmer Sørensen a a

Department of Food Science, Biochemistry and Bioprocessing, Faculty of Science, University of Copenhagen, Rolighedsvej 26, DK-1958 Frederiksberg C, Denmark b Department of Animal Science, Aarhus University, Blichers Alle 20, P.O. Box 50, DK-8830 Tjele, Denmark c Knowledge Centre for Agriculture, Agro Food Park 15, Skejby, DK 8200 Aarhus N, Denmark

A R T I C L E

I N F O

A B S T R A C T

Article history:

Health benefits associated with consumption of cruciferous vegetables have received con-

Received 10 November 2015

siderable attention with a hitherto focus on the role and bioactivity of glucosinolate degradation

Received in revised form 19 April

products. We investigated the in vivo metabolism of intact glucosinolates by following their

2016

fate in digesta and in the endogenous metabolism in pigs. This is the first study to show

Accepted 25 April 2016

an intact glucosinolate, sinalbin, being absorbed and transformed to a sinalbin metabolite

Available online

in the liver by glucuronidation expectedly performed by liver phase II enzymes with subsequent excretion to the urine. From LC–MS/MS data we propose a structure of the sinalbin

Keywords:

metabolite as containing two 4-oxybenzyl groups. Sinalbin and the metabolite were de-

Glucosinolate

tected in plasma from the hepatic vein with a ratio of metabolite to sinalbin of approximately

Absorption

12:1 after 2–4 hours. Induction of liver phase II enzymes by intact glucosinolates indicates

Digestive system

that these also themselves are bioactive compounds with potential health risks or benefi-

Hepatic vein

cial effects.

Phase II enzymes

1.

Introduction

Intact glucosinolates and their transformation products are bioactive compounds responsible for smell, taste and metabolic effects when they are present in food or feed consumed by monogastrics. In particular, glucosinolate degradation products have been investigated for their possible physiological effects with a specific food perspective on the effects of the degradation products as reviewed by Holst and Williamson (2004) and Jeffery and Araya (2009). Glucosinolates are as intact compounds present in all plants of the order Capparales and only in few other plants (Bellostas, Sørensen, Sørensen, &

© 2016 Elsevier Ltd. All rights reserved.

Sørensen, 2007; Bones & Rossiter, 2006; Fahey, Zalcmann, & Talalay, 2001). Glucosinolates comprise a group of more than 140 structurally different compounds with well-defined structure, chemotaxonomic occurrence and co-occurring in the plant with myrosinase isoenzymes (EC 3.2.1.147). However, only few types of glucosinolates are generally present in individual plant species and plant parts. The well-defined structures of glucosinolates consist of alkyl-N-hydroximine-O-sulphate esters with a β- D -thioglucopyranoside group attached to the hydroximine carbon in Z-configuration to the sulphate group (Bellostas et al., 2007; Bones & Rossiter, 2006; Ettlinger & Kjaer, 1968; Ettlinger & Lundeen, 1956, 1957; Fahey et al., 2001; Kjaer, Thomsen, & Hansen, 1960; Sørensen, 1990) (Fig. 1).

* Corresponding author. Department of Food Science, Biochemistry and Bioprocessing, Faculty of Science, University of Copenhagen, Rolighedsvej 26, DK-1958 Frederiksberg C, Denmark. Tel.: +4535332435; fax: +4535332398. E-mail address: [email protected] (J.C. Sørensen). http://dx.doi.org/10.1016/j.jff.2016.04.023 1756-4646/© 2016 Elsevier Ltd. All rights reserved.

Journal of Functional Foods 24 (2016) 450–460

451

But-3-enylglucosinolate [Gluconapin] (2R)-2-Hydroxybut-3-enylglucosinolate [Progoitrin] Benzylglucosinolate [Glucotropaeolin] Phenethylglucosinolate [Gluconasturtiin] 4-hydroxybenzylglucosinolate [Sinalbin] 4-Hydroxyindol-3-ylmethylglucosinolate [4-Hydroxyglucobrassicin] Fig. 1 – Structure and names of selected glucosinolates included in pig diets: semisystematic- and [trivial] names.

In the genus Sinapis, sinalbin (4-hydroxybenzylglucosinolate) is the dominating glucosinolate (Buskov et al., 2000a; Griffiths, Birch, & Hillman, 1998), whereas quantitatively dominating glucosinolates in double low rapeseed are progoitrin, gluconapin and 4-hydroxyglucobrassicin (Bellostas et al., 2007; Fahey et al., 2001). The interest in sinalbin and other food and feed relevant glucosinolates is closely linked to the physico-chemicalbiochemical properties of both intact glucosinolates and especially the complex group of glucosinolate degradation products produced in autolysis-, non-enzymatic- and myrosinasecatalysed reactions. The composition of the degradation products is to a great extent defined by reaction conditions and by the glucosinolate structures (Agerbirk, Olsen, & Sørensen, 1998; Bellostas, Petersen, Sørensen, & Sørensen, 2008a; Bellostas et al., 2007; Bellostas, Sørensen, Sørensen, & Sørensen, 2008b, 2009; Bones & Rossiter, 2006; Buskov et al., 2000a; Buskov, Olsen, Sørensen, & Sørensen, 2000b; Griffiths et al., 1998). Sinalbin has electrophilic properties at the methylene carbon in the benzyl group corresponding to that of the indol-3-ylmethyl carbon in indol-3-ylmethylglucosinolates, resulting in easy release of the thiocyanate ion and the 4-hydroxybenzylgroup from the thiohydroxamate or 4-hydroxybenzylisothiocyanate (Agerbirk et al., 1998; Buskov et al., 2000a,b). The bioactivity and specific biological effects resulting from use of glucosinolate containing plant materials in feed and food depend on the actual structures and concentrations both of the intact glucosinolates and the various transformation products reaching the digestive system, internal organs and the xenobiotic systems and metabolism (Andersen et al., 2010; Bille, Eggum, Jacobsen, Olsen, & Sørensen, 1983; Bjerg, Eggum, Jacobsen, Otte, & Sørensen, 1989; Jeffery & Araya, 2009; Kloss et al., 1996; Lai, Miller, & Jeffery, 2010; Song, Morrison, Botting,

& Thornalley, 2005). So far a great deal of attention has been devoted to the bioavailability of isothiocyanates, and especially to the isothiocyanate sulphoraphane (Holst & Williamson, 2004; Jeffery & Araya, 2009), with a particular focus on the link between reduction of cancer risk and consumption of broccoli (Jeffery & Araya, 2009). Several studies have confirmed that sulphoraphane is absorbed or transported across the intestinal wall, since sulphoraphane metabolites have been detected in the urine and plasma (Bheemreddy & Jeffery, 2007; Clarke et al., 2011; Song et al., 2005; Vermeulen, Klopping-Ketelaars, van den Berg, & Vaes, 2008; Vermeulen, van den Berg, Freidig, van Bladeren, & Vaes, 2006). Previous investigations have thus mainly focused on biologic effects from the glucosinolate transformation products, whereas possible effects from the intact glucosinolates in the diet/digesta rarely have been studied. Previous experiments with rats have shown that intake of too high levels of intact glucosinolates has negative effects on protein utilization and on internal organs (Andersen et al., 2010). Several intact glucosinolates were tested in these studies, and the intact glucosinolates as well as myrosinase catalysed reaction products gave rise to significant biologic effects in the animals (Andersen et al., 2010; Bille et al., 1983; Bjerg et al., 1989). The effects varied as a response to the glucosinolate structure and concentration (Andersen et al., 2010), and only in some cases the presence of myrosinase led to increased biological effects. For sinalbin, increasing the dietary levels reduced the biological value of dietary protein, whereas inclusion of myrosinase in the diet did not further reduce the biological value (Bille et al., 1983). We need, therefore, more specific knowledge on the actual structure and concentration of the compounds causing the

452

Journal of Functional Foods 24 (2016) 450–460

effects on internal organs and influencing the xenobiotica metabolism, thus introducing possible health risks or benefits (Bellostas et al., 2008b; Bille et al., 1983; Bjerg et al., 1989; Jeffery & Araya, 2009; Kloss et al., 1996; Lai et al., 2010). Several studies reveal that glucosinolates decrease in concentration when, as part of the digesta, they pass through the stomach to the small intestine (Elfoul et al., 2001; Freig, Campbell, & Stanger, 1987; Michaelsen, Otte, Simonsen, & Sørensen, 1994; Rowan, Lawrence, & Kershaw, 1991; Vallejo, Gil-Izquierdo, Perez-Vicente, & Garcia-Viguera, 2004). It remains to be confirmed whether the glucosinolate loss is due to transformation reactions or due to absorption of the glucosinolates. In vitro investigations indicate that the glucosinolates can be transported across the intestinal wall in a facilitated transport, and that structurally different glucosinolates have different transportation rates (Michaelsen et al., 1994). Only a limited number of studies have addressed the absorption of intact glucosinolates in vivo, where much lower concentrations of glucosinolates were detected in urine or plasma than that present in the consumed diets (Bheemreddy & Jeffery, 2007; Cwik, Wu, Muzzio, McCormick, & Kapetanovic, 2010; Song et al., 2005). In contrast, a recent study reports around 30% recovery of the intact glucosinolates in the urine when glucosinolates were administered intragastrically to mice (Budnowski et al., 2015). Furthermore, recent in vitro studies have shown that intact glucosinolates can modulate phase II enzymes from rat liver (Razis, Bagatta, De Nicola, Iori, & Ioannides, 2010, 2011a; Razis, Iori, & Ioannides, 2011b). Interestingly, the same authors found that the R-sulphoraphane was a potential inducer of phase II enzymes, but the S-enantiomer was not (Razis et al., 2011b). The same relation was found for goitrin, where the S-enantiomer showed physiological effects different from that of the R-enantiomer (Bille et al., 1983; Bjerg et al., 1989). These findings illustrate the highly specific relationship between structure and biological effects. The aim of the present study was to investigate the absorption of intact glucosinolates and the endogenous metabolism in pigs in order to examine the biological effects of intact glucosinolates in vivo and the possible role of intact glucosinolates as health risks or beneficial compounds present in vegetables from the genus Brassica and cruciferous crops.

2.

Materials and methods

2.1.

Chemicals

Methanol, 1-propanol, sodium cholate, disodium hydrogen phosphate, taurine, FeSO4, TNA and citric acid were purchased from Sigma-Aldrich. Boric acid and sodium acetate were purchased from Merck. Other chemicals, glucosinolates and myrosinases were from the laboratory collection (Bellostas et al., 2007, 2008a; Sørensen, Sørensen, Bjergegaard, & Michaelsen, 1999).

2.2.

Crude extract analysis with capillary electrophoresis

Freeze dried samples from the pigs’ digesta, tissues, blood and urine were extracted by standard procedures (Sørensen et al.,

1999) using boiling methanol (100%) as extraction solvent followed by centrifugation, evaporation of the supernatant and re-dissolution of the residues in water or 5–20% (V/V) 1-propanol. The resulting crude extracts were analysed by micellar electrokinetic capillary chromatography (MECC) in a Hewlett-Packard HP3D CE capillary electrophoresis system (Agilent, Waldbronn, Germany) (Sørensen et al., 1999). The separation buffer consisted of 35 mM sodium cholate, 100 mM disodium hydrogen phosphate, 50 mM taurine, 2% 1-propanol, and pH 7.3 (Bellostas, Sørensen, & Sørensen, 2006). Fused silica capillaries (Agilent, USA) were used in the dimension of 50 µm × 64.5 cm. Samples were injected from the anodic end of the capillary (vacuum injection 50 mbar, 1 s). The electrophoresis was performed at 30 °C for 40 min and the applied voltage was 18 kV. Detection was performed on-column with diode array detector at 214, 230, 280 nm (Sørensen et al., 1999).

2.3.

Desulphoglucosinolate determination

Intact glucosinolates from digesta, plasma and urine crude extracts were isolated by anionic exchange (Sephadex DEAE A-25) and on-column desulphated with sulphatase (Bjergegaard, Michaelsen, Møller, & Sørensen, 1995; Sørensen, 1990; Sørensen et al., 1999). The desulphoglucosinolates were analysed by MECC in a Hewlett-Packard HP3D CE capillary electrophoresis system (Agilent, Waldbronn, Germany). The separation buffer consisted of 200 mM sodium cholate and 250 mM boric acid, pH 8.5. A fused silica capillary with the dimension of 75 µm × 64.5 cm was used, and the electrophoresis was performed at 60 °C for 40 min with 12 kV and detection at 230 and 280 nm. Samples were injected from the anionic end of the capillary (vacuum injection 50 mbar, 1 s) (Bjergegaard et al., 1995).

2.4.

Isolation of the sinalbin metabolite

Preparative HPLC was used to isolate the sinalbin metabolite from the desulphated urine sample. A Gilson HPLC system (Gilson, Middleton, WI, USA) equipped with a Gilson UV/Visdetector was used. The mobile phases were A: 0.1% trifluoroacetic acid in MilliQ-water and B: 0.1% trifluoroacetic acid and 90% acetonitrile in MilliQ-water. The HPLC programme was as follows: 0–5 min 100% A, 5–70 min 60% A, 70– 75 min 0% A, 75–90 min 0% A and 90–100 min 100% A. The flow rate of the mobile phase was 1 mL/min, the temperature was 40 °C and UV detection was performed at 230 nm and 280 nm. The column was a Phenomenex Gemini 5 µm C18 110 A with the dimension 250 × 10 mm. Fractions were collected every 5 min. Unipoint system software was used for instrument programming and data processing.

2.5.

1

H NMR of sinalbin metabolite

The desulphated urine sample and fractions from urine isolated by HPLC were carefully dried and dissolved in D2O prior to recording the 1H NMR spectra at 400 MHz on a Brucker 450 NMR spectrometer.

2.6.

Urine sample analysis on LC–MS/MS

The urine sample (Pig B) collected during the day was separated by HPLC and Hewlett Packard HP system with auto-

Journal of Functional Foods 24 (2016) 450–460

453

sampler, diode-array and fraction collector. The method was as described for the developed standard procedure (Petersen et al., 2011). Briefly, the mobile phases were 99% solvent A (H2O) and 1% solvent B (CH3CN with 0.5% HCOOH). The flow was 0.325 mL/min, and UV detection was read at 230 and 280 nm. The mass spectrometer was operated in TOF scan mode (range 79–1000 m/Z) and the ionization source was electrospray.

animals and care of experimental animals and was under the supervision of the Danish Animal Experiments Inspectorate.

2.7.

Benzylglucosinolate was detected in digesta (collected 24 hours after feeding) from the small intestine in pig A, which had received benzyl- and phenethylglucosinolate. No phenethylglucosinolate was detected in any of the digesta samples reflecting that this glucosinolate was added to the feed in half the concentration of benzylglucosinolate. It is thus present in a concentration considered to be below the detection level in the applied in vivo system, or it might have been transformed, absorbed or already left the digestive system. The reducing conditions and presence of ferro ions in the stomach have been found to give chemical conditions, which make non-enzymatic transformation of intact glucosinolates feasible during digestion in the stomach (Bellostas et al., 2008b; Buskov et al., 2000a). This could explain why only a small amount of intact glucosinolate could be detected in the digesta. Nitriles derived from benzyl- and phenethylglucosinolate were however detected in digesta (results not shown), which shows that the intact glucosinolates were transformed nonenzymatically under the acidic and reducing conditions in the stomach. Nitriles of benzyl- and phenethylglucosinolate are lipophilic and volatile with poor solubility in aqueous systems, making their quantification a challenge when using MECC. In pig B (fed a diet containing sinalbin and rapeseed glucosinolates) neither intact glucosinolates nor nitriles or other transformation products could be detected in the digesta collected at 24 h after feeding glucosinolates. This is not unexpected with respect to the total digestive tract passage time in pigs. The quantitatively dominating glucosinolate in rapeseed is progoitrin, which in non-enzymatic reactions (corresponding to stomach conditions) are transformed into a nitrile ((2R)-1cyano-2-hydroxybut-3-ene) and a thionamide (1-thionamide2-hydroxybut-3-ene) (Bellostas et al., 2008b). The nitrile has only weak chromophores and poor solubility in water; hence, it is difficult to detect in the MECC–DAD system used. The thionamide has a characteristic UV-spectrum with max absorbance at 267 nm, which gives basis for its determination in the MECC–DAD system, but it was not detected, possibly because only low levels would be expected (Bellostas et al., 2008b). The conditions in the stomach provide the possibility for sinalbin to be transformed into several compounds, due to the chemical properties from the p-hydroxybenzyl group (Bellostas et al., 2008; Buskov et al., 2000a). Sinalbin would thus, under reducing conditions, be transformed into the p-hydroxybenzylcyanide, but it would also easily release the thiocyanate ion and 4-hydroxybenzylcarbonium ion derived products (Agerbirk et al., 1998; Buskov et al., 2000b). This means that a variety of compounds could be produced in the stomach, where the 4-hydroxybenzylcarbonium ion can react with nucleophilic compounds, as thiols and free amines in the feed. The hydroxy group in sinalbin makes the transformation prod-

Choice of glucosinolates

The metabolism of dietary intact glucosinolates in monogastrics was followed in two pigs fed standard diets, which were added intact glucosinolates. For this purpose, sinalbin was chosen as a model glucosinolate owing to the special properties of the 4-hydroxybenzyl side chain in this glucosinolate. The chromophore system facilitates the simultaneous detection of the glucosinolate and its transformation products by use of recently developed MECC procedure (Bellostas et al., 2006). Benzyl, phenethyl and rapeseed glucosinolates were also used in the trial, and in the present study, focus has been on the intact glucosinolates, which are hydrophilic compounds, while most of the glucosinolate metabolites from benzyl-, phenethyl- and rapeseed glucosinolates are volatile and lipophilic.

2.8. pigs

Absorption and transformation of glucosinolates in

Two 70 kg pigs (LYxD) were added permanent catheters in the portal vein, hepatic vein and hepatic artery; for details see elsewhere (Jørgensen, Serena, Theil, & Engberg, 2010; Yde, Jansen, Theil, Bertram, & Knudsen, 2012). The pigs were fed a standard diet added intact glucosinolates. A single oral treatment for pig A was based on the diet added benzyl glucosinolate (10 mmol) and phenethyl glucosinolate (3 mmol), and for pig B, the diet was added sinalbin (10 mmol) and a purified glucosinolate mixture from rapeseed (5 mmol). The pig diets used were standard feeds and did not contain any plant material from plants of the Capparales order apart from the added glucosinolates, and consequently no myrosinase was present. The amounts of isolated glucosinolates added to the pig diets correspond to the glucosinolate levels in a standard pig diet with the protein part originating from double low rapeseed (rapeseed containing less than 20 µmol glucosinolate/g seed DM and less than 2% erucic acid in the oil (Bellostas et al., 2007)). During the day, blood samples were collected from permanent catheters at the portal vein, hepatic vein and the hepatic artery. Blood samples were collected in Na-heparinized vacutainers, and plasma was harvested. Samples were taken simultaneously from the hepatic artery, the portal, and the hepatic veins continuously before and after feeding (−30, 30, 60, 90, 150, 210, 270, 330, 390 min), and one urine sample was collected during the day and during sacrifice. Next day liver, kidney and digesta from the stomach, small intestine, caecum and colon were collected in order to detect glucosinolates and glucosinolate derived transformation products in the GI (gastrointestinal) tract. Crude extracts from the digesta were analysed in a cholate based MECC system (Bellostas et al., 2006; Petersen et al., 2011). The study complied with the Danish Ministry of Justice Law no. 1306 (23 November 2007) concerning experiments with

3.

Results and discussion

3.1.

Glucosinolates in the GI system

454

Journal of Functional Foods 24 (2016) 450–460

mAU 30

mAU 80

80

4

300

* D A D 1 , 1 9 .1 7 4 (3 3 7 m A U , - ) R e f= 1 8 .9 7 4 & 1 9 .4 1 7 o f 1 2 0 5 1 0 0 0 0 0 1 0 .D m AU

20*

300

250

200

150

100

70

50

70

250

0 200

225

250

275

300

325

350

375

25

nm

* D A D 1 , 1 6 .5 0 4 (1 0 0 m A U ,A p x ) R e f= 1 6 .3 7 0 & 1 6 .7 2 0 o f 1 2 0 5 1 0 0 0 0 0 1 0 .D m AU 100

60

60

80

60

40

200

2

20

20

50

50

0 200

225

250

275

300

325

350

375

nm

20 20

40

15

40

150

* D A D 1 , 1 5 .2 5 4 (1 .3 m A U , - ) R e f= 1 5 .1 6 4 & 1 5 .3 9 4 o f 1 2 0 5 1 0 0 0 0 0 1 0 .D m AU 1

0 .8

0 .6

30

0 .4

30

0 .2

10

0

100

-0 .2 200

225

250

275

300

325

350

375

nm

20

20 * D A D 1 , 1 3 .9 3 7 (1 .9 m A U ,A p x ) R e f= 1 3 .8 3 4 & 1 4 .0 8 7 o f 1 2 0 5 1 0 0 0 0 0 1 0 .D

3

m AU 1 .5

1 .2 5

1

10

10

2

0 .5

0 .2 5

0

-0 .2 5 200

0

5

50

0 .7 5

225

250

275

300

325

350

375

nm

4

20'

5 6

0

5 10

10

6 15

15

20

20

min.

Fig. 2 – MECC separation and individual absorbance spectra of desulphated compounds present in urine from pig B fed rapeseed glucosinolates and sinalbin. 4: Progoitrin, 5: Epiprogoitrin, 6: Napoleiferin, 2: Gluconapin, 20’: Sinalbin metabolite, 20: Sinalbin, 20*: Sinalbin metabolite. Inserted diode array UV scans range from 200–400 nm.

ucts more soluble in water than those found for the benzyl and phenethyl derivatives, but in the MECC system it was still difficult to obtain reliable quantitation of the nitrile and carbinol, since many other compounds from the feed matrix also appeared with partly overlapping migration times in the analytical system.

3.2.

Retrieval of glucosinolates in urine

The absorption of glucosinolates was investigated in the collected urine samples. Urine samples were collected during the sampling day and upon sacrifice the next day. Methanol– water extracts of the freeze-dried urine samples were oncolumn desulphated and analysed by MECC–DAD according to the method described elsewhere (Bjergegaard et al., 1995). In urine from pig B, which was fed rapeseed glucosinolates and sinalbin, it was possible to detect the pattern of peaks corresponding to the intact glucosinolates fed to the pig (Fig. 2), whereas no glucosinolates could be detected in urine from pig A fed benzyl glucosinolate and phenethyl glucosinolate (data not shown). The DAD-UV scan of each of these peaks from pig B urine had the characteristic glucosinolate pattern (Fig. 2) with high absorbance at 230 nm, and spiking with reference compounds (Fig. 3) confirmed that progoitrin, epiprogoitrin,

0

0

10

10

15

15

20

20

min.

Fig. 3 – MECC separation of reference compounds of rapeseed glucosinolates and sinalbin as desulphoglucosinolates by MECC: 4: Progoitrin, 5: Epiprogoitrin, 6: Napoleiferin, 2: Gluconapin, 20: Sinalbin, 3: Glucobrassicanapin. napoleiferin, gluconapin, and sinalbin were excreted with the urine. Interestingly, the UV-scan of the major peak (peak 20*; Fig. 2) is similar to that of sinalbin, although with a relatively higher absorbance at 280 nm compared to 230 nm. Since this compound appeared in the urine of pig B, it was considered to be a metabolite of sinalbin produced in the body. The late migration (19 min) of this desulphosinalbin metabolite indicated that the compound was negatively charged, and the similarity of the UV scans of sinalbin and the compound (peak 20*) indicated that no additional chromophores had been added to the compound. An additional smaller peak with similar UVscan to that of sinalbin also appeared (peak 20’, Fig. 2). The urine samples from pig B collected a few hours after feeding had a relatively high content of progoitrin (0.86 µmol urine sample), gluconapin (0.5 µmol/ urine sample), sinalbin (1.4 µmol/urine sample) and sinalbin metabolite (3.5 µmol/ urine sample) determined by MECC. Epiprogoitrin and napoleiferin also appear as small peaks, but they occur in concentrations below the level of quantification. These levels correspond to recovery of the glucosinolates in urine of 0.027% for rapeseed glucosinolates (progoitrin and gluconapin), 0.014% of sinalbin, plus 0.035% of the added sinalbin being recovered as a glucuronic acid metabolite. The urine samples collected immediately after sacrifice also contained the dominating sinalbin metabolite, but in lower concentration, and none of the other glucosinolates were detected. As the urine was not collected during the entire day, it is not possible to give a value for the total recovery of intact glucosinolates being absorbed and excreted in the urine.

455

Journal of Functional Foods 24 (2016) 450–460

0.16 Relative Area mAU/g DM

0.14 0.12 0.1 0.08 0.06

0.04 0.02 0 -30

30

90

150

210

270

330

390

Time(min)

Fig. 4 – Relative contents of sinalbin and sinalbin metabolite (20*, Fig. 2) in plasma from the hepatic vein of pig B as function of time from feeding, determined by MECC of desulphated samples. Grey columns: Intact glucosinolate; black columns: Glucosinolate metabolite.

3.3.

Retrieval of glucosinolates in blood

Plasma samples were collected from the portal and hepatic veins as well as from the hepatic artery of both pigs continuously after feeding time. Collection and analysis of plasma from the portal and hepatic veins taken as a function of time have been found to be a valuable method to gain knowledge of absorption and biotransformation of the absorbed compounds (Jørgensen et al., 2010; Yde et al., 2012). In plasma of the portal vein collected from pig B (fed sinalbin and rapeseed glucosinolates) a peak appeared after 30 min with a migration time similar to that of sinalbin, but no metabolites were found. In plasma samples from the hepatic vein the concentration of sinalbin varied as a function of time, and both sinalbin and the sinalbin metabolite (20* Fig. 2) were also detected in these samples. This suggests production of the metabolite (20* Fig. 2) by action of xenobiotic enzymes in the liver. The metabolite and sinalbin appeared in the plasma sample from the hepatic vein taken 90 min after feeding, where they were detected in nearly equal concentrations (Fig. 4). After 4.5 hours, the sinalbin metabolite present in plasma of the hepatic vein from pig B increased in concentration and in samples taken after 5.5 and 6.5 hours after feeding, respectively, the concentration of the sinalbin metabolite was 10 and 13 times higher than that of the sinalbin concentration. As was also the case for urine, no glucosinolates could be detected in plasma samples from pig A, which was fed benzyl glucosinolate and phenethylglucosinolate (data not shown).

3.4.

Identification of a sinalbin metabolite

The structure of the sinalbin metabolites produced in vivo was further investigated by group separations of low molecular weight compounds in digesta, blood and urine from pigs fed with intact glucosinolates added to the diet. The group separations comprised on-column transformations of glucosinolates into desulphoglucosinolates followed by MECC, preparative LC, DAD-UV, LC–MS/MS and 1H NMR. As described below, the quan-

titatively dominating in vivo sinalbin metabolite produced by O-glucuronidation in the liver phase II xenobiotica reactions has all of the analytical data consistent with 4-(β- D glucuronyloxy)benzyl-4-O-benzylglucosinolate (Fig. 5).

3.4.1.

LC–MS/MS analysis of glucosinolates

The quantitatively dominating peak from MECC of urine desulphoglucosinolates from pig B (peak 20*, Fig. 2) was further investigated by LC–MS/MS. The MS fragmentation pattern (Fig. 6, Table 1) is in agreement with the suggested structure (Fig. 5), following the stability of bonds in the parent ion (Mw 628 g/ mol) with decarboxylation, resulting in the first dominating peak at m/z 584 g/mol in negative mode. The easy loss of the carboxylic acid group from the glucuronic acid moiety is considered to be the first step in the fragmentation pattern. It seems thus reasonable that the actual molecular mass of the parent molecule is 627 g/mol in negative mode corresponding to C27H33O14NS (Fig. 6). Table 1 shows the fragmentation with the most abundant ion having a mass of 452 (100%), where a fragment of 177 m/z (the glucuronic acid part) has been lost. The second most abundant ion has an m/z value of 346, corresponding to desulphosinalbin, which is the parent ion found by others when analysing desulphosinalbin by MS (Tolra, Alonso, Poschenrieder, Barcelo, & Barcelo, 2000; Zhang et al., 2011). The fragmented ion from m/z 452 and m/z 346 has a mass of 106 m/ z, and this corresponds to a para-oxy-benzyl group attached to desulphosinalbin. The third most abundant ion has a mass of m/z 240 (346-106) corresponding to the thioglucose-aldoxime. There are similarities with the pattern generally found for desulphated glucosinolates, although there is no ion representing the typical loss of the glucose/thioglucose moiety (Christensen, Kjær, Madsen, Olsen, & Sørensen, 1982; Tolra et al., 2000; Zhang et al., 2011). The fragmentation pattern in plot 3 (Fig. 6) and Table 1 agrees well with the one described for plot one (Table 1). In this plot the parent ion has a mass of 497 m/ z. This fragment is considered to be a result of cyclic reactions with transfer of the carboxylic group from the glucuronic acid

456

Journal of Functional Foods 24 (2016) 450–460

A

B

Fig. 5 – Proposed structures of (A) the desulphated metabolite from biotransformation of sinalbin in the liver and (B) the corresponding intact glucosinolate metabolite from urine of pig B.

part of the metabolite to the phenolic group leading to a release of the rest of the glucuronic acid part (m/z 132).

3.4.2.

Spectrophotometric analysis of glucosinolate products

The desulphated urine sample was also analysed by protonNMR (Fig. 7), showing characteristic signals of para-oxybenzyl moieties with doublets at 7.2 ppm (H2 and H6) and 6.8 ppm (H3 and H5) with the typical ortho coupling at 8.6 Hz, and singlets for the CH2— group at 3.7–3.8 ppm (Fig. 7). The urine sample contained other compounds with corresponding doublets at 6.8–7.6 and singlets at 3.7–3.8 ppm (Fig. 7). However, the HPLC and 1H NMR spectra revealed the presence of additional compounds containing a p-oxybenzyl group. Attempts to isolate these metabolites with use of preparative HPLC failed, and it was not possible to obtain additional 1H NMR structure information of the metabolite found by LC–MS/MS. Additional investigations required for this are considered to be outside the present work. The DAD-UV-scan (Fig. 2) confirms that the only chromophores in the quantitatively dominating metabolite are from the two p-oxybenzyl groups, with the characteristic relation between absorbance at 230 and 280 nm. The metabolite has

a relatively higher absorbance at 280 nm compared to 230 nm than that found for sinalbin. The observed migration times in the applied MECC system correspond to what is expected from a negatively charged compound, e.g. the glucuronic acid part of the metabolite with a pKa’ value of the glucuronic acid around 3 (Wang, Loganathan, & Linhardt, 1991). The sinalbin metabolites were captured on the anion exchanger as was the case for the sulphate group at the aldoxime in intact glucosinolates. This means that the glucuronic acid derivatives are slightly retarded in their release from the column during their isolation as desulphoglucosinolates. The rapeseed glucosinolates progoitrin and gluconapin do not contain a phenolic group (Fig. 1). They were, therefore, not transformed into glucuronides in the liver as found for sinalbin, but even low concentrations of these compounds were detected in the urine (Fig. 2) together with sinalbin and its metabolites, dominated by the glucuronide (20*).

3.5.

Fate of sinalbin in the GI and the xenobiotic system

The identified metabolite (Fig. 5) is thought to be created from addition of a p-hydroxybenzyl group to sinalbin followed by

Table 1 – Summary of LC–MS/MS data for the desulphosinalbin metabolite from urine. PLOT 1

PLOT 3

m/z(%)

Ion

m/z(%)

Ion

584(0) 452(100) 346(65) 240(12)

[M- COOH]− [M-H-glucuronic acid]− [Desulphosinalbin-H]− [Thioglucose-hydroxime-H]−

497(0) 452(10) 346(12) 321(100) 240(5) 215(50) 132(10)

[M-H-glucuronic acid + carboxylic acid]− [M-H-glucuronic acid]− [Desulphosinalbin-H]− [Thioglucose-hydroxime-H]− [Glucuronic acid-COOH]−

Journal of Functional Foods 24 (2016) 450–460

457

Fig. 6 – LC–MS/MS spectrum of desulphated urine sample from one pig (pig B). Plot 1: Fragmentation of parent ion with mass 497 (rearrangement through cyclic formation); plot 2: UV-scan of peak; and plot 3: Fragmentation of parent ion with mass 584 (after decarboxylation).

glucuronidation. In this study it has not been confirmed whether the addition of an extra p-hydroxybenzyl group to sinalbin has happened under the acidic and reducing conditions in the stomach. From a chemical point of view it seems likely that

the carbonium ions generated from the non-enzymatic sinalbin degradation in the stomach could react with the phenol group on the intact glucosinolate (Buskov et al., 2000b). The released carbonium ions have an electrophilic group (—CH2+) and a nucleophilic group (OH), and the proposed reaction between the intact sinalbin and the carbonium ion is feasible, but this type of compound has not been described previously.

3.5.1.

Fig. 7 – 1H NMR spectrum of a methanol extracted and desulphated urine sample from pig B.

Conversion in the liver

Glucuronidation is one of the major pathways of the phase II xenobiotic metabolism, where phenols are O-glucuronidated to promote their excretion by making them polar and water soluble, facilitating their excretion through the kidneys (Parkinson et al., 1995). The glucuronidation takes place in the liver where a β-D-glucuronyl moiety is transferred from uridine 5′-diphospho-β-D-glucuronic acid by UDP-glucuronosyl transferase (EC 2.4.1.17) to the xenobiotic compound (Parkinson et al., 1995). Glucuronide metabolites are excreted in the urine or the bile, and there seems to be a molecular weight dependent relation, where larger molecules are excreted into the bile rather than the urine. The relation between molecular weight and the preferred route of excretion varies among animal species. In the present study, where focus is on hydrophilic small molecules, bile was not examined, but if glucosinolates or the

458

Journal of Functional Foods 24 (2016) 450–460

glucosinolate metabolites had been excreted as bile-micelles, they would have been recognized as the intact compound in the analysed MeOH–Water extract of the digesta. However, as the digesta was collected the next day, then it is not likely that any of such excretion would still be there. The reabsorption of metabolites from bile is often seen as a small increase in concentration in the hepatic vein; however, we did not observe this. It is often found that compounds, which are glucuronidated by liver phase II enzymes, are also sulphated. Since the urine and plasma samples were desulphated as part of the sample treatment, recovery of a sulphated metabolite could not be accomplished. The LC–MS/MS and MECC showed the formation of a metabolite containing two oxybenzyl-groups. However, the proton NMR data did not give a clear picture of the metabolite. There is a need for a final identification of the structures of the metabolites involved in the sinalbin transformation. This will be the subject for additional metabolism studies dealing with the reactions and transformations of glucosinolates occurring in the stomach upon feeding.

3.5.2.

Absorption of glucosinolates to blood and urine

The results obtained indicate that minor amounts of intact glucosinolates are absorbed via the gut epithelium to the blood, as it appears in the portal vein and urine, where low levels (below 1% of ingested amount) of the intact glucosinolates were found. The xenobiotic compounds are, however, only accumulated at low concentrations in the blood and for a short time. Then the compounds are metabolized in the liver and/or excreted with the urine. The results in Fig. 4 clearly show a time related production of the sinalbin metabolite. This delay may be ascribed both to the time required to upregulate phase II liver enzymes and to the transport time for the release of the compound from the liver to the blood. The relatively low concentration detected in plasma and urine are in agreement with results obtained in other studies (Bheemreddy & Jeffery, 2007; Cwik et al., 2010; Song et al., 2005), and detection of the compounds in plasma and urine is seen as a balance between absorption rate and excretion, dilution in blood, and the xenobiotica metabolism (Holst & Williamson, 2004). The amount of intact glucosinolates absorbed does not account for the decrease in digesta concentrations compared to feed seen in the present and in several other studies. The present study has not focused on the nitriles and their possible absorption from the intestine to the blood. But it seems likely that these transformation reactions are responsible for some of the glucosinolate loss noticed in the digestive system. The fact that intact glucosinolates can reduce the biological value when they are fed in high levels to animals may thus both be correlated to the uptake of intact glucosinolates and to their transformation in the digestive tract. For the pig A fed benzyl- and phenethylglucosinolate it was not possible to detect any intact glucosinolates in plasma or urine, in contrast to the feeding trial involving sinalbin and rapeseed glucosinolates (pig B). This could be explained by a high conversion in the stomach of intact glucosinolates into their corresponding nitriles, or that the absorbed glucosinolates are in too low concentration to be detected. Nitriles derived from benzyl and phenethyl glucosinolates were detected in digesta

in the intestines, indicating that the intact glucosinolates were transformed under the reducing conditions in the stomach. However, attempts to quantify the non-enzymatic transformation products were not made in this study.

4.

Conclusion

The present work is the first in vivo study to show the metabolism of intact sinalbin in mammals with transformation/ glucuronidation of sinalbin in the liver and excretion of glucosinolate and products thereof to the urine. The proposed structure of the sinalbin metabolite is a 4-(β- D glucuronyloxy)benzyl-4-O-benzylglucosinolate. The glucosinolate levels added to the diets in the present study have been included at levels relevant to food and feed based on cruciferous crops, double low rapeseed and cruciferous vegetables. Sinalbin, benzyl-, phenethyl- and rapeseed glucosinolates were investigated for their fate in digesta and in the endogenous metabolism in pigs. A sinalbin metabolite was found to be glucuronidated in the liver expectedly by liver phase II enzymes, and the transformation products were excreted to the urine. From LC–MS/MS data the metabolite is proposed to be a product with two 4-oxy-benzyl groups. The presence of intact glucosinolates and metabolites of intact glucosinolates in blood as well as liver phase II transformations followed by excretion to the urine suggests that the intact glucosinolates may contribute to the biological effects and potential health risks and benefits including potential effects on internal organs and the often discussed effects on cancer and other biological effects.

Acknowledgements Financial support for the project Rapeseed Protein for Piglets from Danish Pig Production, Axelborg (Copenhagen, Denmark), from Scanola A/S (Aarhus, Denmark) and for the project HITFOOD from Danish National Advanced Technology Foundation (Project code: 058-2012-01) is gratefully acknowledged. The LC–MS/MS and 1H NMR support from Jan H. Christensen (Department of Plant and Environmental Sciences, University of Copenhagen) and Carl Erik Olsen (Department of Plant and Environmental Sciences, University of Copenhagen) is also gratefully acknowledged. REFERENCES

Agerbirk, N., Olsen, C. E., & Sørensen, H. (1998). Initial and final products, nitriles, and ascorbigens produced in myrosinasecatalyzed hydrolysis of indole glucosinolates. Journal of Agricultural and Food Chemistry, 46, 1563–1571. Andersen, K. E., Frandsen, H. B., Jensen, S. K., Bellostas, N. M., Sørensen, A. D., Sørensen, J. C., & Sørensen, H. (2010). Glucosinolates in Brassica – Health risks, but also benefits. Norwegian Academy of Science and Letter, 104–124. Bellostas, N., Petersen, I. L., Sørensen, J. C., & Sørensen, H. (2008a). A fast and gentle method for the isolation of myrosinase complexes from brassicaceous seeds. Journal of Biochemical and Biophysical Methods, 70, 918–925.

Journal of Functional Foods 24 (2016) 450–460

Bellostas, N., Sørensen, A. D., Sørensen, J. C., & Sørensen, H. (2007). Genetic variation and metabolism of glucosinolates. Advances in Botanical Research: Incorporating Advances in Plant Pathology, 45, 369–415. Bellostas, N., Sørensen, A. D., Sørensen, J. C., & Sørensen, H. (2008b). Fe2+-catalyzed formation of nitriles and thionamides from intact glucosinolates. Journal of Natural Products, 71, 76– 80. Bellostas, N., Sørensen, A. D., Sørensen, J. C., & Sørensen, H. (2009). Type and concentration of redox reagents influencing nitrile formation upon myrosinase (Brassica carinata)catalyzed hydrolysis of glucosibarin. Journal of Molecular Catalysis B: Enzymatic, 57, 229–236. Bellostas, N., Sørensen, J. C., & Sørensen, H. (2006). Micellar electrokinetic capillary chromatography – Synchronous monitoring of substrate and products in the myrosinase catalysed hydrolysis of glucosinolates. Journal of Chromatography. A, 1130, 246–252. Bheemreddy, R. M., & Jeffery, E. H. (2007). The metabolic fate of purified glucoraphanin in F344 rats. Journal of Agricultural and Food Chemistry, 55, 2861–2866. Bille, N., Eggum, B. O., Jacobsen, I., Olsen, O., & Sørensen, H. (1983). Antinutritional and toxic effects in rats of individual glucosinolates (+/−myrosinases) added to a standard diet. 1. Effects on protein-utilization and organ weights. Zeitschrift für Tierphysiologie, Tierernährung und Futtermittelkunde – Journal of Animal Physiology and Animal Nutrition, 49, 195–210. Bjerg, B., Eggum, B. O., Jacobsen, I., Otte, J., & Sørensen, H. (1989). Antinutritional and toxic effects in rats of individual glucosinolates (+/− myrosinases) added to a standard diet .2. Zeitschrift für Tierphysiologie, Tierernährung und Futtermittelkunde – Journal of Animal Physiology and Animal Nutrition, 61, 227–244. Bjergegaard, C., Michaelsen, S., Møller, P., & Sørensen, H. (1995). Desulphoglucosinolates by micellar electrokinetic capillary chromatography based on a bile-salt. Journal of Chromatography. A, 717, 325–333. Bones, A. M., & Rossiter, J. T. (2006). The enzymic and chemically induced decomposition of glucosinolates. Phytochemistry, 67, 1053–1067. Budnowski, J., Hanske, L., Schumacher, F., Glatt, H., Platz, S., Rohn, S., & Blaut, M. (2015). Glucosinolates are mainly absorbed intact in germfree and human microbiotaassociated mice. Journal of Agricultural and Food Chemistry, 63, 8418–8428. Buskov, S., Hasselstrøm, J., Olsen, C. E., Sørensen, H., Sørensen, J. C., & Sørensen, S. (2000a). Supercritical fluid chromatography as a method of analysis for the determination of 4-hydroxybenzylglucosinolate degradation products. Journal of Biochemical and Biophysical Methods, 43, 157–174. Buskov, S., Olsen, C. E., Sørensen, H., & Sørensen, S. (2000b). Supercritical fluid chromatography as basis for identification and quantitative determination of indol-3-ylmethyl oligomers and ascorbigens. Journal of Biochemical and Biophysical Methods, 43, 175–195. Christensen, B. W., Kjær, A., Madsen, J. Ø., Olsen, C. E., & Sørensen, H. (1982). Mass spectrometric characteristics of some pertrimethylsilylated desulfoglucosinolates. Tetrahedron, 38, 353–357. Clarke, J. D., Hsu, A., Riedl, K., Bella, D., Schwartz, S. J., Stevens, J. F., & Ho, E. (2011). Bioavailability and inter-conversion of sulforaphane and erucin in human subjects consuming broccoli sprouts or broccoli supplement in a cross-over study design. Pharmacological Research, 64, 456–463. Cwik, M. J., Wu, H. P., Muzzio, M., McCormick, D. L., & Kapetanovic, I. (2010). Direct quantitation of glucoraphanin in dog and rat plasma by LC-MS/MS. Journal of Pharmaceutical and Biomedical Analysis, 52, 544–549.

459

Elfoul, L., Rabot, S., Khelifa, N., Quinsac, A., Duguay, A., & Rimbault, A. (2001). Formation of allyl isothiocyanate from sinigrin in the digestive tract of rats monoassociated with a human colonic strain of bacteroides thetaiotaomicron. FEMS Microbiology Letters, 197, 99–103. Ettlinger, M. G., & Kjaer, A. (1968). Sulphur compounds in plants. In T. J. Mabry (Ed.), Recent advances in phytochemistry (99 59– 144). Amsterdam: North-Holland Publishing Company. Ettlinger, M. G., & Lundeen, A. J. (1956). The structures of sinigrin and sinalbin – An enzymatic rearrangement. Journal of the American Chemical Society, 78, 4172–4173. Ettlinger, M. G., & Lundeen, A. J. (1957). First synthesis of a mustard oil glucoside – The enzymatic Lossen rearrangement. Journal of the American Chemical Society, 79, 1764–1765. Fahey, J. W., Zalcmann, A. T., & Talalay, P. (2001). The chemical diversity and distribution of glucosinolates and isothiocyanates among plant. Phytochemical, 56, 5–51. Freig, A. A. H., Campbell, L. D., & Stanger, N. E. (1987). Fate of ingested glucosinolates in poultry. Nutrition Reports International, 36, 1337–1345. Griffiths, D. W., Birch, A. N. E., & Hillman, J. R. (1998). Antinutritional compounds in the brassicaceae: Analysis, biosynthesis, chemistry and dietary effects. Journal of Horticultural Science and Biotechnology, 73, 1–18. Holst, B., & Williamson, G. (2004). A critical review of the bioavailability of glucosinolates and related compounds. Natural Product Reports, 21, 425–447. Jeffery, E. H., & Araya, M. (2009). Physiological effects of broccoli consumption. Phytochemistry Reviews, 8, 283–298. Jørgensen, H., Serena, A., Theil, P. K., & Engberg, R. M. (2010). Surgical techniques for quantitative nutrient digestion and absorption studies in the pig. Livestock Science, 133, 57–60. Kjaer, A., Thomsen, H., & Hansen, S. E. (1960). Isothiocyanates .3. Glucocapangulin, a novel isothiocyanate-producing glucoside. Acta Chemica Scandinavica, 14, 1226–1227. Kloss, P., Jeffery, E., Tumbleson, M., Zhang, Y., Parsons, C., & Wallig, M. (1996). Studies on the toxic effects of crambe meal and two of its constituents, 1-cyano-2-hydroxy-3-butene (CHB) and epi-progoitrin, in broiler chick diets. British Poultry Science, 37, 971–986. Lai, R. H., Miller, M. J., & Jeffery, E. (2010). Glucoraphanin hydrolysis by microbiota in the rat cecum results in sulforaphane absorption. Food & Function, 1, 161–166. Michaelsen, S., Otte, J., Simonsen, L. O., & Sørensen, H. (1994). Absorption and degradation of individual intact glucosinolates in the digestive-tract of rodents. Acta Agriculturae Scandinavica, Section A – Animal Science, 44, 25–37. Parkinson, A., Ogilvie, B. W., Buckley, D. B., Kazmi, F., Czerwinski, M., & Parkinson, O. (1995). Biotransformation of xenobiotics. In C. D. Klaassen (Ed.), Casarett and Doull’s toxicology: The basic science of poisons (pp. 185–366). New York: The McGraw-Hill Companies. Petersen, I. L., Tomasi, G., Sørensen, H., Boll, E. S., Hansen, H. C. B., & Christensen, J. H. (2011). The use of environmental metabolomics to determine glyphosate level of exposure in rapeseed (Brassica Napus L.) seedlings. Environmental Pollution, 159, 3071–3077. Razis, A. F. A., Bagatta, M., De Nicola, G. R., Iori, R., & Ioannides, C. (2010). Intact glucosinolates modulate hepatic cytochrome P450 and phase II conjugation activities and may contribute directly to the chemopreventive activity of cruciferous vegetables. Toxicology, 277, 74–85. Razis, A. F. A., Bagatta, M., De Nicola, G. R., Iori, R., & Ioannides, C. (2011a). Up-regulation of cytochrome p450 and phase ii enzyme systems in rat precision-cut rat lung slices by the intact glucosinolates, glucoraphanin and glucoerucin. Lung Cancer (Amsterdam, Netherlands), 71, 298–305.

460

Journal of Functional Foods 24 (2016) 450–460

Razis, A. F. A., Iori, R., & Ioannides, C. (2011b). The natural chemopreventive phytochemical r-sulforaphane is a far more potent inducer of the carcinogen-detoxifying enzyme systems in rat liver and lung than the s-isomer. International Journal of Cancer. Journal International du Cancer, 128, 2775–2782. Rowan, T. G., Lawrence, T. L. J., & Kershaw, S. J. (1991). Effects of dietary copper and a probiotic on glucosinolate concentrations in ileal digesta and in feces of growing pigs given diets based on rapeseed meals. Animal Feed Science and Technology, 35, 247–258. Song, L. J., Morrison, J. J., Botting, N. P., & Thornalley, P. J. (2005). Analysis of glucosinolates, isothiocyanates, and amine degradation products in vegetable extracts and blood plasma by LC-MS/MS. Analytical Biochemistry, 347, 234–243. Sørensen, H. (1990). Glucosinolates: Structure-propertiesfunction. In F. Shahidi (Ed.), Canola and rapeseed: Production, chemistry, nutrition and processing technology (pp. 149–172). New York: Van Nostrand Reinhold Publisher. Sørensen, H., Sørensen, S., Bjergegaard, C., & Michaelsen, S. (1999). Chromatography and capillary electrophoresis in food analysis. UK: Royal Society of Chemistry. Tolra, R. P., Alonso, R., Poschenrieder, C., Barcelo, D., & Barcelo, J. (2000). Determination of glucosinolates in rapeseed and thlaspi caerulescens plants by liquid chromatographyatmospheric pressure chemical ionization mass spectrometry. Journal of Chromatography. A, 889, 75–81. Vallejo, F., Gil-Izquierdo, A., Perez-Vicente, A., & Garcia-Viguera, C. (2004). In vitro gastrointestinal digestion study of broccoli

inflorescence phenolic compounds, glucosinolates, and vitamin C. Journal of Agricultural and Food Chemistry, 52, 135– 138. Vermeulen, M., Klopping-Ketelaars, I. W. A. A., van den Berg, R., & Vaes, W. H. J. (2008). Bioavailability and kinetics of sulforaphane in humans after consumption of cooked versus raw broccoli. Journal of Agricultural and Food Chemistry, 56, 10505–10509. Vermeulen, M., van den Berg, R., Freidig, A. P., van Bladeren, P. J., & Vaes, W. H. J. (2006). Association between consumption of cruciferous vegetables and condiments and excretion in urine of isothiocyanate mercapturic acids. Journal of Agricultural and Food Chemistry, 54, 5350–5358. Wang, H. M., Loganathan, D., & Linhardt, R. J. (1991). Determination of the pka of glucuronic-acid and the carboxy groups of heparin by C-13-nuclear-magnetic-resonance spectroscopy. The Biochemical Journal, 278, 689–695. Yde, C. C., Jansen, J. J., Theil, P. K., Bertram, H. C., & Knudsen, K. E. B. (2012). Different metabolic and absorption patterns of betaine in response to dietary intake of whole-wheat grain, wheat aleurone or rye aleurone in catheterized pigs. European Food Research and Technology = Zeitschrift für LebensmittelUntersuchung und -Forschung. A, 235, 939–949. Zhang, J. X., Sun, X. M., Zhang, Z. P., Ni, Y. W., Zhang, Q., Liang, X. M., Xiao, H. B., Chen, J. P., & Tokushisa, J. G. (2011). Metabolite profiling of arabidopsis seedlings in response to exogenous sinalbin and sulfur deficiency. Phytochemical, 72, 1767–1778.