Plant Breeding 123, 60—65 (2004) Ó 2004 Blackwell Verlag, Berlin ISSN 0179-9541 Genetic and environmental effects on glucosinolate content and chemoprotective potency of broccoli M. W. Farnham1 ,P.E.Wilson2 ,K.K.Stephenson3 and J.W. Fahey3,4 1 US Department of Agriculture, Agricultural Research Service, US Vegetable Laboratory, 2700 Savannah Hwy., Charleston, SC 29414, USA, E-mail: [email protected]; 2 Phytochemicals and Health Group, New Zealand Institute for Crop and Food Research Limited, Private Bag 4704, Christchurch, New Zealand; 3 Lewis B. and DorothyCullman Cancer Chemoprotection Center, Department of Pharmacologyand Molecular Sciences, Johns Hopkins School of Medicine, Baltimore, MD 21205, USA; 4 Center for Human Nutrition, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD 21205, USA With 1 figure and 3 tables Received February 28, 2003/Accepted July 11, 2003 Communicated by G. Wricke Abstract gut as a result of the activityof its microflora (Shapiro et al. Broccoli is well recognized as a source of glucosinolates and their 1998, 2001). Certain isothiocyanates are known to inhibit isothiocyanate breakdown products. Glucoraphanin is one of the most tumour formation in mammals initiated bya varietyof abundant glucosinolates present in broccoli and its cognate isothio- chemical carcinogens (Zhang and Talalay1994). Sulphora- cyanate is sulphoraphane, a potent inducer of mammalian detoxica- phane, the cognate isothiocyanate of glucoraphanin (4-meth- tion (phase 2) enzyme activity and anti-cancer agent. This study was ylsulphinylbutyl glucosinolate), has been the focus of designed to measure: glucosinolate levels in broccoli florets from an numerous studies because it is a potent inducer of mammalian arrayof genotypes grown in several environments; the elevation of a detoxication and antioxidant (phase 2) enzyme activity that keyphase 2 enzyme,quinone reductase, in mammalian cells exposed to protects against tumourigenesis in a rodent mammarytumour floret extracts; and total broccoli head content. There were significant model (Zhang et al. 1994, Faheyet al. 1997). Robust evidence environmental and genotype-by-environment effects on levels of glucoraphanin and quinone reductase induction potential of broccoli points to a pivotal role played by phase 2 enzymes in the heads; however, the effect of genotype was greater than that of detoxication of electrophiles and in the suppression of carcin- environmental factors. The relative rankings among genotypes for ogenesis and mutagenesis (Kensler 1997, Talalayand Fahey glucoraphanin and quinone reductase induction potential changed, 2001). when expressed on a per head basis, rather than on a concentration Glucoraphanin is a relativelyabundant aliphatic glucosino- basis. Correlations of trait means in one environment vs. means from a late present in harvested florets of cultivated broccoli; second were stronger for glucoraphanin and quinone reductase however, other aliphatic glucosinolates, such as glucoiberin induction potential on a per head basis than on a fresh weight (3-methylsulphinylpropyl glucosinolate) and glucoerucin concentration basis. Results of this studyindicate that development of (4-methylthiobutyl glucosinolate), have also been identified a broccoli phenotype with a dense head and a high concentration of in broccoli tissues. In general, glucoiberin, glucoerucin, and glucoraphanin to deliver maximum chemoprotective potential (high enzyme induction potential/glucoraphanin content) is a feasible goal. other aliphatic glucosinolates occur at levels <10% of those of glucoraphanin (Carlson et al. 1987, Shelp et al. 1993, Kushad Key words: Brassica oleracea — doubled haploids — gluco- et al. 1999), and their cognate isothiocyanates typically raphanin — isothiocyanates — Italica group — nutrition — account for <2% of total glucosinolate-derived phase 2 sulphoraphane induction potencyof broccoli (Zhang et al. 1992, Prestera et al. 1993). Other prominent glucosinolates in broccoli florets Epidemiological evidence that relates broccoli vegetable con- are indole glucosinolates such as glucobrassicin (indole- sumption to a reduction in the risk of certain cancers in humans 3-ylmethyl glucosinolate), neoglucobrassicin (1-methoxyin- dates back 25 years. Such evidence has gained momentum and dole-3-ylmethyl glucosinolate), and 4-hydroxyglucobrassicin. validityover the ensuing quarter-century,and it is particularly Results of in vitro assays indicate that the primary hydrolysis strong for cancers of the colon/rectum (Graham et al. 1978, products of these glucosinolates have verylow phase 2 inducer Kohlmeier and Su 1997, Verhoeven et al. 1997) and prostate potential which is estimated to be <5% of the total for (Jain et al. 1999, Kolonel et al. 2000). This chemoprotective broccoli (Faheyet al. 1997, 1998). effect is thought to be due in large part to the glucosinolates Faulkner et al. (1998) have suggested that genetic factors present in these cruciferous vegetables (Beecher 1994, Zhang which induce high levels of methylsulphinylalkyl glucosino- and Talalay1994, Hecht 2000, Talalayand Fahey2001). lates in wild relatives of broccoli could be transferred to Glucosinolates are b-thioglucoside N-hydroxysulphates with cultivated broccoli. This would require numerous generations an aglycone (or R-group) that is an alkyl, alkenyl, thioalkyl, of selection to improve the horticultural phenotype. Alternat- thioalkenyl, aryl, arylalkyl or indolyl moiety (reviewed by ively, the genetic diversity of glucoraphanin concentration Rosa et al. 1997, Faheyet al. 2001). These compounds are extant in relativelyelite broccoli germplasm could be more hydrolysed by myrosinase to their cognate isothiocyanates quicklycaptured byusing the elite lines to breed new cultivars when plant cells are damaged (e.g. chewed), or in the human (Farnham et al. 2000). U. S. Copyright Clearance Center Code Statement: 0179–9541/2004/2301–0060 $ 15.00/0 www.blackwell-synergy.com Genetic and environmental effects on glucosinolate content and chemoprotective potencyof broccoli 61 Broccoli genotype has a significant effect on the glucosinolate produced a 10–12 cm size head. Two heads per plot were sampled at profile in broccoli florets as well as on the plant levels of random for laboratoryanalysis,and subtending stalks were cut to a glucoraphanin (Giamoustaris and Mithen 1996, Faulkner 15 cm length. Heads were immediatelyplaced on ice in a cooler and et al. 1998, Kushad et al. 1999, Li et al. 2001). However, only within 30 min of field harvest, fresh weights (FWs) were recorded a few studies (Farnham et al. 2000, Rosa and Rodrigues 2001, (excepting 1996), approximatelyhalf of the florets were cut from the stem, placed in individual sealable freezer bags, and frozen at )80°C. Brown et al. 2002) have examined and identified the interac- Boiling 80% methanol (aqueous) extracts were prepared from the tion between genotype and environment on levels of gluco- fresh-frozen florets and stored at )20°C until needed for the bioassay raphanin in broccoli tissues. Brown et al. (2002) estimated a of and for direct quantitation of quinone reductase induction moderatelyhigh broad sense heritabilityof 54% for this trait. potential and of glucosinolates byhigh-performance liquid chroma- Farnham et al. (2000) and Rosa and Rodrigues (2001) found tography(HPLC), as described in the previous work (Farnham et al. high correlations between glucoraphanin expression in one 2000). environment and a second. Farnham et al. (2000) also showed that glucoraphanin concentration and phase 2 enzyme induc- Bioassay of quinone reductase induction potential: Bioassayof the tion potential of genotypes were highly correlated. representative phase 2 enzyme quinone reductase, was performed using In a previous study(Farnham et al. 2000), the focus was on Hepa 1c1c7 cells as described originallybyProchaska et al. (1992) and genotypic differences for floret concentrations of glucorapha- modified byFaheyet al. (1997). Excess myrosinase [0.0003 units/ml of nin and phase 2 induction potential. In the current study, a cell culture medium; purified as described byShikita et al. (1999)], and 500 lM ascorbate, was added at the time microtitre plates were dosed subset of the previouslyexamined elite broccoli genotypes with broccoli extracts, to achieve complete hydrolysis of glucosinolates grown in three environments was evaluated, and concentration during a 48-h incubation at 37°C. With this method, conversion of of glucoraphanin and other glucosinolates and for phase 2 extracted glucosinolates to their cognate isothiocyanates is essentially enzyme induction potential were estimated. Another goal was quantitative. One unit of inducer activityis the concentration that to examine the importance of broccoli head weight as a factor doubles quinone reductase activityin a microtitre well containing that governs the ultimate quantityof glucoraphanin (and 150 ll of medium. Hence, a compound with a CD (the Concentration hence the chemoprotective potential) of a harvested broccoli of a compound required to Double the quinone reductase specific head. This is as critical in breeding for enhanced chemopro- activityin Hepa 1c1c7 murine hepatoma cells) of 1.0 lM has 6667 units tective broccoli as is tissue concentration. of inducer activityper lmol. Inducer potencyof extracts is expressed as units/g FW. Materials and Methods Paired-ion chromatography of glucosinolates: Plant extracts were chromatographed isocraticallyin acetonitrile/water (1 : 1, v/v) con- Plant materials: Plant