1906
DOI 10.1002/mnfr.201200225
Mol. Nutr. Food Res. 2012, 56, 1906–1916
RESEARCH ARTICLE
Isothiocyanate concentrations and interconversion of sulforaphane to erucin in human subjects after consumption of commercial frozen broccoli compared to fresh broccoli Shikha Saha1 , Wendy Hollands1 , Birgit Teucher1 , Paul W. Needs1 , Arjan Narbad1 , Catharine A. Ortori2 , David A. Barrett2 , John T. Rossiter3 , Richard F. Mithen1 and Paul A. Kroon1 1
Institute of Food Research, Norwich Research Park, Norwich, UK Centre for Analytical Bioscience, School of Pharmacy, University of Nottingham, Nottingham, UK 3 Department of Life Sciences, Division of Cell & Molecular Biology, Sir Alexander Fleming Building, Imperial College, London, UK 2
Scope: Sulforaphane (a potent anticarcinogenic isothiocyanate derived from glucoraphanin) is widely considered responsible for the protective effects of broccoli consumption. Broccoli is typically purchased fresh or frozen and cooked before consumption. We compared the bioavailability and metabolism of sulforaphane from portions of lightly cooked fresh or frozen broccoli, and investigated the bioconversion of sulforaphane to erucin. Methods and results: Eighteen healthy volunteers consumed broccoli soups produced from fresh or frozen broccoli florets that had been lightly cooked and sulforaphane thio-conjugates quantified in plasma and urine. Sulforaphane bioavailability was about tenfold higher for the soups made from fresh compared to frozen broccoli, and the reduction was shown to be due to destruction of myrosinase activity by the commercial blanching-freezing process. Sulforaphane appeared in plasma and urine in its free form and as several thio-conjugates forms. Erucin N-acetyl-cysteine conjugate was a significant urinary metabolite, and it was shown that human gut microflora can produce sulforaphane, erucin, and their nitriles from glucoraphanin. Conclusion: The short period of blanching used to produce commercial frozen broccoli destroys myrosinase and substantially reduces sulforaphane bioavailability. Sulforaphane was converted to erucin and excreted in urine, and it was shown that human colonic flora were capable of this conversion.
Received: April 20, 2012 Revised: August 22, 2012 Accepted: August 28, 2012
Keywords: Broccoli / Erucin / Sulforaphane
1
Introduction
Epidemiological studies indicate that the consumption of broccoli is associated with the reduction in the risk of degenerative disease such as cancer [1]. 4-Methylsulphinylbutyl glucosinolate (glucoraphanin) accumulates in the florets of broccoli. Following consumption, it is hydrolysed into the corresponding isothiocyanate (sulforaphane) either by the Correspondence: Dr. Paul Kroon, Institute of Food Research, Colney Lane, Norwich NR4 7UA, UK E-mail:
[email protected] Fax: +44-1603-507723 Abbreviations: ESP, epithiospecifier like proteins; ITC, isothiocyanate C 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
plant thioglucosidase myrosinase or by bacterial thioglucosidases in the colon, if the myrosinase has been denatured by cooking [2]. A nitrile derivative of the glucosinolate may also be produced due to the interaction of plant “epithiospecifier-like” proteins (ESP). Mild cooking can denature these proteins but leave myrosinase enzymes intact, maximising the amount of glucosinolate conversion to isothiocyanate. Following absorption, sulforaphane is conjugated with glutathione and metabolised via the mercapturic acid pathway, resulting in predominantly N-acetyl-cysteine conjugates that are excreted in the urine [3]. Several studies have quantified the pharmacokinetics of sulforaphane following the consumption of broccoli with either active or inactive myrosinase. Most of these studies have analysed sulforaphane and its thiol metabolites in plasma and urine through a www.mnf-journal.com
1907
Mol. Nutr. Food Res. 2012, 56, 1906–1916
cyclocondensation reaction [4, 5]. While this method is efficient at quantifying total thiols, it is not able to discriminate between free isothiocyanates and individual thiol metabolites, or between different isothiocyanates. LC-MS/MS methods have been developed that are able to distinguish between different metabolites, and have demonstrated that unconjugated sulforaphane is found in the plasma [6]. Analyses through both approaches have shown that between 30 and 80% of sulforaphane is absorbed into the blood stream or excreted in the urine compared to the level of the parent glucosinolate if broccoli is consumed with active myrosinase, but this drops to 10% or less if broccoli is consumed with inactive myrosinase, with a later Tmax associated with colonic absorption as opposed to absorption through the stomach or small intestine [7]. An important question is the fate of the isothiocyanate or glucosinolate that is not absorbed. One possibility is that it is not absorbed but excreted in the faeces. Alternatively, there may be alternative routes to metabolism of sulforaphane post absorption, and conjugation with proteins has been proposed [4]. A third possibility may be represented by the bioconversion of sulforaphane to other isothiocyanates. Recent studies carried out in humans and rats have reported the in vivo interconversion of sulforaphane to erucin [8]. The erucin is the enzymatic hydrolysis product of glucoerucin, mainly found in rocket salad species, and structurally represents the reduced analogue of sulforaphane [8]. It was suggested that the conversion of glucoraphanin/sulforaphane to the reduced glucosinolate/isothiocyanate occurred in the liver, and that there was enterohepatic circulation of glucosinolates so that hydrolysis occurred via gut microbes [9]. Another study has confirmed the in vivo interconversion of sulforaphane to erucin suggesting an inter-subject variability [10]. In the current study, we quantified sulforaphane metabolism following consumption of soups containing mildly cooked broccoli (i.e. with active myrosinase, but inactive ESP) and frozen broccoli with inactivated myrosinase, and analysed a subset of samples for the presence of erucin and its thiol conjugates. In addition, we investigated whether this conversion could be due to gut microbial activity.
2
Materials and methods
2.1 Materials Sinigrin was obtained from Sigma Chemicals (UK). Sulforaphane was obtained from LKT Laboratories, Inc. (Minnesota, USA). Methyl-, ethyl- and butyl-isothiocyanates, cysteine, cysteine-glycine, glutathione and N-acetyl-cysteine were purchased from Aldrich (UK). DEAE Sephadex A25 and SP Sephadex C25 were obtained from Amersham Biosciences (Sweden). The synthesis of cysteine, cysteine-glycine, glutathione and N-acetyl-cysteine conjugates of sulforaphane was done by an optimized protocol published elsewhere [3]. The synthetic products were purified by preparative HPLC C 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
using a C18 column and the structures of the synthesized materials were confirmed by 1 H NMR (JEOL, EX 270) and LC– MS–MS (Quattro Ultima tandem mass, Waters Micromass, Manchester, UK). Sulforaphane-nitrile was synthesized by the published protocol [11]. All solvents and other chemicals used were of HPLC grade and purity was assessed to be 95% or greater in all compounds. Sulfatase (Type H-1 from Helix pomatia) was obtained from Sigma Chemicals and purified before use. Sulfatase (300 mg) was dissolved in ice-cold water (12 mL), mixed with ice-cold ethanol (12 mL) and stirred. Following centrifugation (3000 × g, 6 min) the supernatant was mixed with ice-cold ethanol (1.5 × vol), stirred and centrifuged to obtain a pellet that was dissolved in water (8 mL). Samples (2 mL) were passed sequentially through columns of Sephadex A25 and Sephadex C25 to obtain the purified sulfatase that was stored at −20⬚C.
2.2 Subjects and study design Eighteen apparently healthy volunteers aged 20–65 years were recruited by study scientists to participate in this study, which was conducted at the Human Nutrition Unit at the Institute of Food Research. All study participants were assessed for eligibility on the basis of a health questionnaire and the results of clinical laboratory tests. All volunteers were screened for full blood count, fasting glucose, liver function and urea and electrolytes. The following exclusions applied: smokers; diagnosed with long-term medical conditions such as asthma (unless untreated within the past 2 years), heart disease, gastro-intestinal disease, diabetes, cancer; regular prescribed medication (except HRT and oral contraceptive); taking dietary supplements (unless judged not to affect study outcome) or antibiotics for greater than 4 weeks before the start of the study; pregnant; blood donation within 4 months prior to start of study; BMI less than 18.5 or greater than 35; clinical results at screening judged by the medical advisor to affect study outcome or be indicative of a health problem. Subjects were genotyped for the GSTM1 and GSTT1 alleles but remained blind to the test results. Genomic DNA was isolated from whole blood samples (200 L) using a Genelute genomic DNA mini-prep kit (Sigma-Aldrich). GSTM1 and GSTT1 genotypes were identified by quantitative realR ) on a 7500 real-time PCR system (Aptime PCR (Taqman plied Biosytems, Warrington, UK). Amplification reactions R Universal Master Mix, genomic (25 L) contained Taqman DNA (50 ng), forward and reverse primers (500 ng each), R probe with 3 TAMRA and 5 -FAM a dye-labelled TaqMan dye-labels (see Gasper et al. for primer and probe sequencess R DNA polymerase enzyme (Applied [16]) and Amplitaq Gold Biosystems). Following activation of the DNA polymerase (10 min at 95⬚C), reaction mixtures were subjected to 40 PCR cycles, each consisting of 95⬚C for 15 s and 60⬚C for 60 s. The study was explained to participants and written informed consent was subsequently obtained. The study protocol was approved by the Human Research Governance Committee at www.mnf-journal.com
1908
S. Saha et al.
the Institute of Food Research and Norwich Research Ethics Committee (LREC 2003/082). The study was a randomized two-phase crossover design investigating the bioavailability of phytochemicals from fresh and processed broccoli in subjects of known GSTM1 genotype. Each of the two test phases comprised a 5-day period of intervention separated by a washout period of at least 1 week. During each period of intervention, subjects followed a cruciferous vegetable-free diet. To aid compliance, a list of authorized and prohibited foods was provided to volunteers. On day 3 of the intervention, subjects arrived at the Human Nutrition Unit following an overnight fast and an intravenous catheter was inserted. A baseline blood sample was obtained. Subjects were given a standard breakfast consisting of two slices of white toast with spread prior to ingesting a cold broccoli soup that had been prepared either from lightly cooked fresh broccoli (100 g fresh weight) or frozen broccoli (100 g) (see below). The volunteers were blinded to the intervention meal. Blood samples (10 mL) were collected into lithium heparin tubes at 15, 30 and 45 min and 1, 1.5, 2, 3, 4, 6, 8, 24 and 48 h after broccoli consumption. Blood was immediately centrifuged at 2000 × g (10 min at 4⬚C) and samples of the plasma acidified with HCl. Urine was collected the day before consumption of broccoli (24-h collection) and between 0–2, 2–4, 4–6, 6–8, 8–24 and 24–48 h after consumption. The amount of urine from each period was measured and subsamples were acidified with HCl. All plasma and urine samples were subsequently stored at −80⬚C until analysis.
2.3 Broccoli cultivation and preparation Broccoli (cultivar Marathon) was grown at the ADAS Experimental Research Station (Terrington, Norfolk, UK) and harvested in September 2005. One half of the harvested broccoli was used fresh (within 24 h of harvesting from the field) while the other half was processed at a commercial vegetable processing factory (Christian Salvesen, Bourne, Lincolnshire, UK) as follows: blanching (91.2⬚C × 140 s), blast freezing (−33⬚C), storage (−28⬚C). For both fresh and frozen broccoli, individual soups were prepared by cooking 100 g broccoli florets (pre-thawed in case of frozen material) with 150 g water for 75 s in a microwave oven on full power (700 W), followed by homogenisation in a domestic food blender. For each of the fresh and frozen broccoli soups, the individual soup portions were combined, mixed and individual portions (235 g) frozen in bags. Broccoli soups were kept frozen (−18⬚C) until use. For the interventions, soups were thawed overnight in a fridge and consumed without further processing. Samples of each soup type were taken for analysis of glucosinolates, isothiocyanate and nitrile. To investigate the effects of cooking broccoli on the potential for sulforaphane production, a large batch (20 kg) of broccoli was purchased from a local supermarket, washed in water and cut into small florets (3–4 cm). Individual samples (200 g fresh weight) were subjected to steaming (florets C 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Mol. Nutr. Food Res. 2012, 56, 1906–1916
cooked in pre-heated domestic steamer), or boiling (florets added to pan of boiling water with enough water to just cover florets) or microwave cooking (in a bowl with 50 g water on 700 watts full power) for 0, 0.25, 0.50, 0.75, 1.0, 1.5, 2, 3, 4, 5 or 7 min. Immediately after the cooking period had ended, the broccoli tissue was cooled using dry ice and frozen (−20⬚C). After freeze-drying and milling, samples were analysed for glucosinolates, isothiocyanate and nitrile as described below.
2.4 Analysis of glucosinolates, isothiocyanate, and nitrile in broccoli Broccoli tissue and soups were freeze-dried and powdered (domestic food mixer) prior to analysis. All samples were extracted and analysed in triplicate. Broccoli glucosinolates were measured using a method that converts the glucosinolates to the equivalent desulfoglucosinolates [12]. Briefly, Samples (100–200 mg) of broccoli powder were extracted with hot aqueous methanol (70% vol/vol, 5 mL) following the addition of internal standard (sinigrin). Samples were mixed by vortexing and incubated at 70⬚C for 20–30 min with occasional mixing. Extracts were allowed to cool and a sample (3 mL) of the supernatant was applied to an ion exchange column that was subsequently washed with water (2 × 0.5 mL) and then 0.02 M sodium acetate (2 × 0.5 mL). The columns were then layered with purified sulfatase (75 L) and incubated at RTP overnight. The desulfoglucosinolates were eluted by sequential application of 0.5, 0.5, and 0.25 mL water and analysed using HPLC as described below. Glucosinolate breakdown products (isothiocyanate and nitrile) were measured in samples of broccoli powder that had been dissolved in buffer to facilitate the action of endogenous enzymes (myrosinase) and cofactors (ESP). Samples (40 mg) of freeze-dried broccoli powder were thoroughly mixed with phosphate-buffered saline (Invitrogen, Paisley, UK). The tubes were sealed and vortex mixed. The samples were incubated on the heating block at 37⬚C for 2 h, with vortex mixing every 15 min to ensure optimal hydrolysis. The tubes were centrifuged (13 000 × g, 4⬚C for 30 min) and the supernatants were removed and filtered (0.2 m) and analysed by HPLC as described below. Desulfoglucosinolates were analysed using a Waters Spherisorb ODS2 (250 × 4.6 mm id, 5 m particle size) column connected to a model 1100 HPLC system (Agilent Technologies, Waldbronn, Germany) comprising of a binary pump, degasser, cooled autosampler, column oven and diode array detector. Samples were eluted at 1.0 mL/min with a gradient of increasing ACN using water (solvent A) and ACN (solvent B). The gradient started at 0% solution B increasing over 25 min to 50% B and finally re-equilibrated to 0% B for 13 min. Desulfoglucosinolates were monitored at 229. Glucoraphanin was detected in the broccoli and soups but glucoerucin was not detected in the broccoli and soups. Glucoraphanin was quantified using absorbance at 229 nm by www.mnf-journal.com
Mol. Nutr. Food Res. 2012, 56, 1906–1916
comparison with the internal standard (sinigrin) peak area ratio and with the correction factor (1.13) of the desulfoglucoraphanin [12]. Glucosinolate breakdown products (isothiocyanate and nitrile) were analysed using a Phenomenex Luna C-18(2) (150 × 3 mm id, 3 m particle size) column connected to an Agilent 1100 HPLC system as described above but with an additional on-line mass spectrometric detector (Agilent Technologies). Samples were eluted at 0.3 mL/min with a gradient of increasing ACN using 0.1% aq. formic acid (solvent A) and 0.1% formic acid in ACN (solvent B). The gradient started at 0% solution B increasing over 30 min to 30% B and finally re-equilibrated to 0% B for 10 min. Sulforaphane was monitored using absorbance at 235 nm, and were quantified by comparison to external standard curves (linear regression coefficients >0.99). Sulforaphane-nitrile was monitored using MS in full scan positive ion mode with electrospray ionisation and quantified using selected ion monitoring (m/z = 146) and quantified by external standard curve of authentic sulforaphane-nitrile (R2 > 0.99).
2.5 Analysis of sulforaphane and sulforaphane conjugates in plasma and urine Plasma and urine were analysed for sulforaphane and its glutathione, cysteine-glycine, cysteine and N-acetyl-cysteine conjugates using a novel, validated method that has been described recently [6]. Each sulforaphane conjugate was synthesized and purified for use in this study as described previously [3].
2.6 Analyses of erucin conjugate in urine Erucin N-acetyl-cysteine conjugate was synthesized by analogy with the procedure of Kassanhun et al. [3] for sulforaphane–N-acetyl-cysteine. Hence erucin was added to a solution of N-acetyl-cysteine, pre-adjusted to pH 7.8, in aqueous ethanol, it was purified by reversed-phase HPLC. Urine samples (1 mL) were acidified to pH 4 with formic acid, centrifuged at 13 000 × g for 10 min, filtered by (0.45 m) polypropylene syringe filter and 100 L samples were injected directly onto the HPLC column. Erucin N-acetyl-cysteine was analysed using a Phenomenex Luna C-18(2) (150 × 3 mm id, 3 m particle size) column connected to an Agilent 1100 HPLC system as described above with an additional on-line mass spectrometric detector (LC/MSD SL, Agilent Technologies) was added. Samples were eluted at 0.3 mL/min with a gradient of increasing ACN using 0.1% aq. formic acid (solvent A) and 0.1% formic acid in ACN (solvent B). The gradient started at 0% solution B increasing over 12 min to 50% B and finally reequilibrated to 0% B for 8 min. Erucin N-acetyl-cysteine was monitored using MS in full scan positive ion mode with electrospray ionisation. Identification was performed on the basis of retention time and spectra matching respect to spiking C 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1909 synthetic standard in urine. Quantification was performed by selected ion monitoring (m/z = 325) mode and spiking standard calibration curve of authentic erucin N-acetyl-cysteine. (R2 > 0.99).
2.7 Bioconversion of glucoraphanin by gut bacteria The ability of the human gut microbiota to transform specific glucosinolates was examined in a simple batch fermentation model. The faecal inoculum was obtained from a healthy volunteer who had not taken any antibiotics or pre- or probiotics in the previous 2 months. The freshly voided faecal material was homogenised (10% w/v) in 0.1M PBS (pH 7.0). One millilitre of this slurry was then used to inoculate a vessel containing 9 mL of pre-sterilised and pre-reduced basal growth medium (peptone water 2 g/L, yeast extract 2 g/L, NaCl 0.1 g/L, K2 HPO4 0.04 g/L, MgSO4 .7H2 O 0.01 g/L, CaCl2 .6H2 O 0.01 g/L, NaHCO3 2 g/L, Tween-80 2 mL, hemin 0.02 g/L, vitamin K1 10 L, cysteine HCl 0.5 g/L, bile salts 0.5 g/L, pH 7.0). The incubation was performed at 37⬚C under oxygen-free atmosphere (10% H2 ; 10% CO2 ; 80% N2 ) using an anaerobic cabinet (Don Whitley Scientific, Shipley, West Yorkshire, UK). Glucosinolates were added to a final concentration of 3 g/mL. Samples (1 mL) were removed at intervals, centrifuged at 12 000 × g for 5 min and the supernatants were filter sterilised and stored at −20⬚C until further analysis. The glucosinolate concentrations were analysed by HPLC using pure glucosinolate standards glucoraphanin and glucoerucin. Glucosinolates were isolated and purified following the procedure of Thies [13]. Rocket seed (Eruca sativa) was ground to a fine powder (50 g) in a coffee grinder and then defatted with petroleum ether (40–60⬚C fraction, 7 × 200 mL). The defatted seed powder was extracted with boiling 80% methanol for 20 min, filtered and the extraction repeated. The combined filtrates were evaporated under reduced pressure using a rotary evaporator. The residue was taken up in water and treated with an equimolar mixture of barium and lead acetate (0.5 M). The precipitate was removed by centrifugation and added to a DEAE Sephadex A25 column (1.4 g) pre-equilibriated with 6M imadazole formate. Following repeated washes with water, the column was cleaned with a mixture of formic acid: isopropanol: water (3:2:5) and rinsed with water. Glucosinolates were eluted with 0.5 M potassium sulphate. Excess potassium sulphate was removed by mixing the aqueous eluate with ethanol in a 1:1 ratio and removing the precipitate by centrifugation. The supernatant was evaporated to near dryness and taken up in methanol and allowed to stand at −20⬚C for 20 min and then centrifuged to remove the precipitate and the supernatant evaporated to give a syrupy residue. The glucoerucin was then further purified by chromatography on a Sephadex G10 equilibriated with water. Glucoerucin elution was monitored at an absorbance of 230 nm and fractions freeze-dried to give the purified product (400 mg). Purity was assessed by HPLC using a pure sinigrin standard (Apin Chemicals UK). www.mnf-journal.com
1910
S. Saha et al.
Glucoraphanin was prepared following the procedure of Iori et al. [14]. Hydrolysis products (1 mL) were extracted (X2) with dichloromethane (1 mL) and the combined extracts dried with anhydrous magnesium sulphate. The dichloromethane extract was concentrated to 200 L and analysed by GC-MS as previously described [11].
2.8 Data analysis Statistical analyses were performed using the R data analysis software (R Development Core Team (2006). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3–900051-07–0, URL http://www.R-project.org.). Repeated Measures models were used to analyse the data. For urine, “Total sulforaphane” was the response and “Volunteer” was included as a “Random Effect”. Plasma data were treated similarly but using different response variables (Cmax, Tmax, AUC). For all models, regression diagnostics were checked to determine if data transformation, outlier omissions, or alternative non-parametric models were required. All results from the models were considered significant if p < 0.05.
3
Results
3.1 Subjects A total of 24 volunteers were genotyped in order to obtain almost equal numbers of GSTM1 null (n = 10) and positive (n = 8) subjects. In this rather limited sample (n = 24), the prevalence of the GSTM1 null genotype was 66.6%, which is within the range for Caucasians [15]. Eighteen subjects (16 females, two males) completed the entire study without reporting adverse events, and had the following anthropometric characteristics (mean ± SD: age 45.1 ± 11.0 years; height 1.70 ± 0.07 m; weight 70.5 ± 12.1 kg; body mass index 25.7 ± 3.9 kg m–2 ).
3.2 Glucosinolates, sulforaphane, and sulforaphane nitrile in soup Analysis of the soups for glucosinolates, sulforaphane and nitrile revealed that the soup made from lightly cooked fresh broccoli contained significant quantities of sulforaphane (4.16 mg per serving) but no glucosinolates; these data are consistent with lightly cooked fresh broccoli retaining sufficient myrosinase to hydrolyse all the glucoraphanin, part of which is converted to sulforaphane. In contrast, the soups made from commercially frozen broccoli contained significant amounts of glucosinolates (18.6 mg per serving) but no detectable sulforaphane, indicating complete loss of myrosinase activity. These data show that the blanching process completely destroyed the myrosinase. Both soups contained small C 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Mol. Nutr. Food Res. 2012, 56, 1906–1916
quantities of sulforaphane nitrile (0.08 mg and 0.196 mg for lightly cooked fresh and frozen broccoli, respectively).
3.3 Kinetics and bioavailability of sulforaphane conjugates in urine and plasma Before consumption of broccoli soups, sulforaphane and sulforaphane conjugates were not detected in samples of plasma or urine from any of the volunteers. Following consumption of the soups, sulforaphane and a range of thioconjugates representing the entire mercapturic acid metabolic/excretory pathway (Fig. 1) were detected and quantified. The order of prevalence of the conjugates in plasma was sulforaphane > sulforaphane-cysteine-glycine > sulforaphane-cysteine > sulforaphane-glutathione ∼ sulforaphane-N-acetyl-cysteine. The mean concentrations were as follows (lightly cooked from fresh, lightly cooked from frozen broccoli soups, respectively) sulforaphane (0.6194, 0.0336 M), sulforaphane-cysteineglycine (0.4037, 0.0084 M), sulforaphane-cysteine (0.1616, 0.0013 M), sulforaphane-N-acetyl-cysteine (0.1229, 0.0031 M) sulforaphane-glutathione (0.1151, 0.0004 M). The order of prevalence of the conjugates in urine was sulforaphaneN-acetyl-cysteine > sulforaphane-cysteine > sulforaphane > sulforaphane-cysteine-glycine > sulforaphane-glutathione. The mean concentrations were as follows (lightly cooked from fresh, lightly cooked from frozen broccoli soups, respectively) sulforaphane-N-acetyl-cysteine (10.103, 1.680 M), sulforaphane-cysteine (2.676, 0.445 M), sulforaphane (1.105, 0.136 M), sulforaphane-cysteine-glycine (0.020, 0.001 M) sulforaphane-glutathione (0.005, 0.000 M). When volunteers consumed a soup made from a standard portion of fresh broccoli that was lightly cooked, sulforaphane conjugates appeared in plasma within 15 min and total conjugate concentrations peaked at 0.21 M after 2 h (Fig. 2A). After peaking, plasma levels declined to very low levels (mean 0.5 nM) at 48 h. In contrast, when volunteers consumed a soup made from frozen broccoli that had been cooked in the same way, sulforaphane conjugates were not detected in plasma until 1 h after consumption of the soup, peaked much later at 6 h, and the maximal concentration achieved in plasma was substantially lower (0.020 versus 0.21 M, p for difference 60%) up to 2 min before they declined rapidly; nitrile yields remained above 30% up to 75 s before declining rapidly. These data indicate that the best cooking methods for time-dependent retention of sulforaphane production in broccoli are steaming and microwave cooking. Broccoli boiled for just 2 min generated less than 15% mole yield of sulforaphane from glucoraphanin.
3.4 Effects of cooking broccoli The complete lack of sulforaphane in the soups made from commercial frozen broccoli that had been cooked only briefly (75 s) is likely to be due to the blanching step used during commercial freezing of broccoli. We investigated the effects of the cooking process and the cooking time on the potential for sulforaphane production through the action of residual endogenous broccoli thioglucosidase activity. Samples of fresh broccoli yielded 50–60% of sulforaphane and 30–50% of sulforaphane-nitrile (molar yields) upon hydrolysis of freezedried tissue in buffer. When cooked by boiling, sulforaphane molar yields remained largely unaltered for the first 1.5 min and then declined, while sulforaphane nitrile molar yields were less than 5% for all of the cooked samples. Using a microwave, sulforaphane molar yields increased to 80% at 75 s and then declined to 30% at 3 min and less than 6% thereafter, while nitrile yields remained near 50% for up to C 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.5 Erucin N-acetyl-cysteine in urine To assess the possibility of sulforaphane and erucin interconversion, six urine samples were analysed for the presence of erucin N-acetyl-cysteine conjugate (the major metabolite in urine) from three volunteers following consumption of either the lightly cooked or frozen broccoli. In each case, this compound could be detected. The pharmacokinetics of erucin was not similar to that of sulforaphane. The rate of urinary excretion (mols) from the lightly cooked fresh soup was highest in the 4–6 h collection period and declined thereafter, whereas from the lightly cooked frozen broccoli soup, the rate of excretion peaked in the 8–24 h (see Fig. 3). The ratio of erucin N-acetyl-cysteine and sulforaphane Nacetyl-cysteine was similar in between subjects (three subjects) in lightly cooked fresh broccoli consumption group www.mnf-journal.com
1912
S. Saha et al.
Mol. Nutr. Food Res. 2012, 56, 1906–1916
the initial concentration at the 4-h sample point and then decreased dramatically in the 8 and 24 h samples. Glucoerucin was not detected at the 0- and 4-h sample points but appeared in appreciable concentrations in the 8 and 24 h samples. The hydrolysis products at each time point were analysed by GC-MS, and product formation was observed only in the 8 and 24 h samples. Erucin, erucin-nitrile, sulforaphane and sulforaphane-nitrile were all detected in the 24 h sample (Table 3). However, no hydrolysis products were detected in the 0 and 4 h samples while only erucin nitrile was detected in the 8 h sample. The peak abundances for erucin (28.06%) and erucin-nitrile (68.53%) were much higher than for sulforaphane and sulforaphane-nitrile where only trace amounts were detected (Table 3). No glucoerucin or glucoraphanin breakdown products (sulforaphane, sulforaphane nitrile, erucin and erucin nitrile) were detected in incubations conducted with autoclave-sterilised faecal samples indicating that the origins of the reduction of glucoerucin and hydrolysis products are entirely enzymatic.
4
Discussion
4.1 Bioavailability
Figure 2. Plasma pharmacokinetic profiles (A) and urinary excretion rates (B) following consumption of broccoli florets that were lightly cooked from fresh (filled circles and bars) or from frozen (open circles and bars). The lightly cooked fresh broccoli soup contained 23.5 moles of sulforaphane and no glucoraphanin; the lightly cooked frozen broccoli soup contained 42.5 moles of glucoraphanin and no sulforaphane.
but variable in frozen broccoli consumption group (Table 2). The sulforaphane N-acetyl-cysteine was approximately three times higher than erucin N-acetyl-cysteine in both groups (Table 2).
3.6 Bioconversion of glucoraphanin to glucoerucin in the human gut Glucoerucin is the major glucosinolate in Rocket seeds [14] providing a convenient source of material in good yields. In order to obtain glucoraphanin, the reduced form of glucoerucin was oxidised with hydrogen peroxide followed by purification on DEAE-A25 Sephadex and Sephadex G10. The bioconversion of glucoraphanin in a simple batch fermentation model of human gut microbiota was followed by HPLC and GC-MS analysis (Fig. 4). The time course curve (Fig. 4C) shows that the glucoraphanin concentration remained near C 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Epidemiological evidence consistently supports a link between increased cruciferous vegetable consumption and reduced risk of cancer at several sites [17]. Sulforaphane, the isothiocyanate released from glucoraphanin, has attracted particular interest for its anticarcinogenic effect. Sulforaphane was present predominantly as free sulforaphane in plasma (see results section 3.3, Fig. 2A) but also as a number of conjugates (glutathione, cysteine-glycine, cysteine, Nacetyl-cysteine) [6]. Here, we report that consumption of a standard portion of lightly cooked broccoli provides a reasonable dose of sulforaphane (100 M in soup; 23.5 mol), and leads to the rapid appearance in plasma of a number of thiolconjugates that are subsequently excreted via the urine. These data are mainly consistent with a report describing the plasma pharmacokinetics and urinary excretion of sulforaphane following consumption of a soup made from standard broccoli, and a soup made from broccoli with enhanced levels of sulforaphane [16]. Further, none of our subjects excreted all the ingested sulforaphane; mean urinary sulforaphane excretion was 58.5% of dose and the highest individual urinary yield was 79.6%. The reasons for the differences in urinary yield between this study and that reported previously [16] are not clear, but it is plausible that the different dosages are an important factor. Broccoli is consumed largely as a cooked vegetable, although a small proportion is eaten raw, for example when used in salads. It has been established that extended cooking of broccoli and other cruciferous vegetables inactivates endogenous myrosinase enzyme activity and that as a consequence, the glucosinolates enter the gastrointestinal tract intact [2]. Shorter cooking times can lead to enhanced www.mnf-journal.com
1913
Mol. Nutr. Food Res. 2012, 56, 1906–1916
Table 1. Comparison of plasma pharmacokinetic and urinary excretion data for the study population and stratified according to GSTM1 genotype
GSTM1-positive (n = 8)
Plasma sulforaphanec) Cmax (mol/L) Fresh broccoli 0.271 ± 0.127b) Frozen broccoli 0.031 ± 0.030 Tmax (h) Fresh broccoli 1.47 ± 0.74 Frozen broccoli 9.14 ± 6.72 AUC (mol h/L) Fresh broccoli 1.26 ± 0.80 Frozen broccoli 0.414 ± 0.576 Urinary sulforaphanec) Total urinary excretion (0–48 h) (mol) Fresh broccoli 13.39 ± 1.52 Frozen broccoli 1.90 ± 0.62
GSTM null (n = 10)
p-valuea) Broccoli type
Genotype
0.224 ± 0.140 0.021 ± 0.020
