Contents lists available at ScienceDirect

Food Chemistry

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

Characterization of glucosinolates in 80 broccoli genotypes and differentorgans using UHPLC-Triple-TOF-MS method

Zhansheng Lia,b,c,1, Shuning Zhenga,b,c,1, Yumei Liua,b, Zhiyuan Fanga,b, Limei Yanga,b,Mu Zhuanga,b, Yangyong Zhanga,b, Honghao Lva,b, Yong Wanga,b, Donghui Xua,b,c,⁎

a Institute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences, Beijing, Chinab Key Laboratory of Biology and Genetic Improvement of Horticultural Crops, Ministry of Agriculture, P. R. China, Beijing 100081, Chinac Key Laboratory of Vegetables Quality and Safety Control, Ministry of Agriculture and Rural Affairs of China, Beijing 100081, China

A R T I C L E I N F O

Keywords:BroccoliGenotypeOrgansUHPLC-Triple-TOF-MSGlucosinolate

A B S T R A C T

We aimed to characterize and quantify glucosinolate compounds and contents in broccoli, and a total of 80genotypes and eight developmental organs were analyzed with UHPLC-Triple-TOF-MS. The method was vali-dated in terms of performance, and the coefficients of determination (R2) were 0.97 and 0.99 for glucoraphaninand gluconapin, respectively. In 80 genotypes, twelve glucosinolates were found in broccoli florets ranging from0.467 to 57.156 µmol/g DW, with the highest glucosinolate content being approximately 122-fold higher thanthe lowest value. The principal component of glucobrassicin, neoglucobrassicin and glucoraphanin explained60.53% of the total variance. There were positive correlations among hydroxyglucobrassicin, methoxygluco-brassicin, glucobrassicin, glucoerucin, gluconasturtiin, glucoraphanin, and glucotropaeolin (P < 0.05). Theroot contained 43% of total glucosinolates in 80 genotypes, and glucoraphanin represented 29% of the totalglucosinolate content in different organs. The mutant broccoli genotypes were found by analysis of gluconapincontents in different organs.

1. Introduction

Glucosinolates (GLS) are a group of plant secondary metabolitescomprising at least 120 known structures mainly found in cruciferousplants, including Chinese cabbage, broccoli, cabbage, Chinese kale andso on (Brown, Yousef, Reid, Chebrolu, Thomas, Krueger, et al., 2015;Fahey, Zalcmann, & Talalay, 2001; Halkier & Gershenzon, 2006).Glucosinolates can be divided into three groups according to theiramino acid precursors: aliphatic glucosinolates are derived from me-thionine (Met), alanine (Ala), leucine (Leu), isoleucine (Ile) or valine(Val); benzenic glucosinolates are derived from phenylalanine (Phe) ortyrosine (Tyr); and indolic glucosinolates are derived from tryptophane(Trp) (Grubb & Abel, 2006; Sønderby, Geu-Flores, & Halkier, 2010).The glucosinolates and their degradation products have been widelyhelpful as anticancer agents in human health, for defense against insectsand disease in plants, and for flavor regulation in cruciferous vegetables(Abdull Razis, Iori, & Ioannides, 2011; de Oliveira, Brasil, & Furstenau,2018; Halkier & Gershenzon, 2006). Therefore, glucosinolate and itsdegradation products (sulforaphane, benzyl isothiocyanate, indole-3-carbinol, and 3, 3′-diindolylmethane) have become a research hotspot

in food science, medicine, and botany (Alumkal, Slottke, Schwartzman,Cherala, Munar, Graff, et al., 2015; Liang, Li, Yuan, & Vriesekoop,2008).

At present, the qualitative and quantitative analysis methods ofglucosinolates focus on high-performance liquid chromatography(HPLC), ultrahigh-performance liquid chromatography (UPLC), and li-quid chromatography-mass spectrometry (HPLC-MS). Since liquidchromatography-mass spectrometry can perform qualitative andquantitative analysis based on the characteristic mass spectrometryinformation of different glucosinolates, which is better than HPLC, li-quid chromatography-mass spectrometry has become a popular tool inthe analysis of glucosinolates (Alumkal, et al., 2015; Bello, Maldini,Baima, Scaccini, & Natella, 2018; Liang, Li, Yuan, & Vriesekoop, 2008).Additionally, time-of-flight high-resolution mass spectrometry (TOF-MS) has the advantages of high resolution, high sensitivity and fastanalysis speed, so it is more effective in the qualitative analysis ofglucosinolates than are more simple MS methods (Bell, Oruna-Concha,& Wagstaff, 2015). Mass spectrometry and secondary mass spectro-metry complete the structural identification of compounds in plantsamples in the absence of standards.

https://doi.org/10.1016/j.foodchem.2020.127519Received 10 September 2019; Received in revised form 5 June 2020; Accepted 5 July 2020

⁎ Corresponding author at: Key Laboratory of Vegetables Quality and Safety Control, Ministry of Agriculture and Rural Affairs of China, Beijing 100081, China.E-mail address: xudonghui@caas.cn (D. Xu).

1 Co-first authors: Zhansheng Li and Shuning Zheng.

Food Chemistry 334 (2021) 127519

Available online 10 July 20200308-8146/ © 2020 Published by Elsevier Ltd.

T

The reports of glucosinolate biosynthesis have been extensivelystudied in Arabidopsis, and the regulation genes have been compre-hensively described in previous research (Grubb & Abel, 2006;Sønderby, Geu-Flores, & Halkier, 2010). Recently, increasing researchhas focused on glucosinolates in Brassica plants, especially glucor-aphanin and its hydrolysis product, sulforaphane, in broccoli (Z. S. Li,Liu, Li, Fang, Yang, Zhuang, et al., 2019; Liang, Li, Yuan, & Vriesekoop,2008; Sønderby, Geu-Flores, & Halkier, 2010). Variation in glucosino-late accumulation among different organs and developmental stages ofArabidopsis thaliana has been studied over the past 40 years (P. D.Brown, Tokuhisa, Reichelt, & Gershenzon, 2003). However, there is nocomprehensive report on glucosinolate variation among additionalgenotypes and developmental organs, though some research points outthe change in total glucosinolates or several glucosinolates in broccoliflorets after postharvest based on a small number of materials (Baik,Juvik, Jeffery, Wallig, Kushad, & Klein, 2003; Volker, Freeman,Banuelos, & Jeffery, 2010; Wang, Gu, Yu, Zhao, Sheng, & Zhang, 2012),representing a lack of evidence of additional genotypes and classicaldevelopmental organs. At the same time, past research in glucosinolatesmostly used HPLC and lacked accurate structure calculations based onmass spectrometry information. Therefore, a comprehensive under-standing of the glucosinolate profile and content during broccoli de-velopment is still needed for genetic breeding or basic research onBrassica plants.

In this study, ultrahigh-performance liquid chromatography-time-of-flight mass spectrometry (UHPLC-Triple-TOF-MS) is established andused for the analysis of glucosinolates in different broccoli genotypesand developmental organs, which provides a reliable analytical tech-nology platform for the analysis of glucosinolates in cruciferous plantsin the future. Glucosinolates and their degradation products have beenextensively reported but with little use of ultrahigh-performance liquidchromatography with TOF-MS scan-IDA-Product ion scan (UHPLC-Triple-TOF-MS). Our work demonstrates that glucosinolates can be si-multaneously detected and quantified by using UHPLC-Triple-TOF-MS.The new method also provides necessary MS and MS/MS informationfrom one injection based on high-resolution TOF-MS. At the same time,accurate masses of molecular ions and fragment ions were obtained andstated in this study.

2. Materials and methods

2.1. Chemical and reagents

The chemicals used for sample preparations were HPLC-gradeacetonitrile (JT Baker, USA) and formic acid (Fluka, Buchs,Switzerland). Deionized water was purified through a Milli-Q system(Millipore, Bedford, MA, USA). Standards of 4-methylsulfinylbutylglucosinolate (glucoraphanin, GRA) and 3-butenyl glucosinolate (glu-conapin, NAP) were > 98% pure, purchased from LKT Laboratories,Inc. (St. Paul, Minnesota, USA).

2.2. Plant materials and pretreatment

Broccoli (Brassica oleracea L. var. italic) samples were planted andcollected from the Institute of Vegetables and Flowers, ChineseAcademy of Agricultural Sciences (IVF-CAAS). The collected samples,including different organs of roots, leaves, stalks, florets and develop-mental buds from 6 inbred lines (Fig. 1) and florets from the other 80genotypes (inbred lines) genetically separated from Japan, America, theNetherlands, Switzerland and Egypt, were freeze-dried, ground intopowder and stored at low temperature.

The extraction procedure of glucosinolates was performed ac-cording to previously described methods (Zheng, Zhang, Liu, Lv, Lw,Xu, et al., 2017). The dried and ground broccoli samples (0.5 g) wereextracted using methanol (70% v/v), vortexed for 1 min at room tem-perature and extracted for 15 min in a water bath at 70 °C. Then, the

mixture was extracted for 15 min in a water bath at 70 °C and sonicatedfor 30 min at room temperature. After centrifugation for 5 min at13,000 × g, the supernatant was collected into a 2 mL centrifugal tube.The supernatant was dried by nitrogen flow, and the dry residue wasdissolved in ultrapure water. The solution was filtered through a0.22 μm syringe filter and stored at −20 °C until detection analysis.Meanwhile, different concentrations of 1, 10, 100 and 1000 ng/mg of 4-methylsulfinylbutyl and 3-butenyl glucosinolate (standard) were addedto the dry samples, and 5 duplicates for each treatment were set tocalculate the extraction recovery (n = 5).

2.3. Simultaneous identification and characterization of glucosinolates

The Shimadzu LC-30A HPLC system (Japan) was equipped with anSPD-20 UV detector with a Waters ACQUITY UPLC BEH C18 column(2.1 × 100 mm, 1.7 μm) (Milford, MA, USA). Acidified water (0.1%formic acid, v/v) and acetonitrile (0.1% formic acid, v/v) were used asmobile phases A and B, respectively. The gradient program was carriedout as follows: 0.0 min, 95% A and 5% B; 6.0 min, 60% A and 40% B;6.5 min, 100% B; 8.0 min, 100% B; 8.1 min, 95% A and 5% B; and10.0 min, 95% A and 5% B. The flow rate was set at 0.40 mL/minthroughout the gradient. The injection volume was 1 μL, and thecolumn temperature was maintained at 40 °C.

The HPLC system was coupled to a quadrupole-time-of-flight (ABSciex, USA) orthogonal accelerated Q-TOF mass spectrometer equippedwith an electrospray ionization source (ESI). Parameters for analysiswere set using negative and positive ion modes, with spectra acquiredover a mass range from m/z 80 to 1000. The optimal values of theESI–MS parameters were capillary voltage, −4.5 kV; drying gas tem-perature, 550 °C; drying gas flow, 10.0 μL/ min; nebulizing gas pres-sure, 0.34 MPa; air curtain gas pressure, 0.24 MPa; auxiliary atomizinggas pressure, 0.34 MPa; collision and collision RF voltage, −35 eV and15 eV; transfer time 70 μs; and prepulse storage, 5 μs. Moreover, au-tomatic MS/MS experiments were performed using nitrogen as collisiongas, with the collision energy values adjusted as follows: m/z 100,20 eV; m/z 500, 30 eV; and m/z 1000, 35 eV.

The MS data acquisition was processed with Analyst® TF 1.7.1software (AB Sciex Pte. Ltd., Singapore), and PeakView® 2.1 (AB SciexPte. Ltd., Singapore) and MasterView™ 1.0 (AB Sciex Pte. Ltd.,Singapore) software was used to identify and analyze the glucosino-lates. The MS/MS fragments were inferred through the FormularFindermodel in PeakView®, as well as based on the first mass spectrographs,isotope-ratio mass spectrometry, molecular unsaturation, etc. In addi-tion, the model of the Fragment Pane in MasterView® was used forstructural analysis to further verify the molecular structure of the glu-cosinolates.

Fig. 1. Different organs of broccoli were collected in the bolting stage.

Z. Li, et al. Food Chemistry 334 (2021) 127519

2

2.4. Statistical analysis

All assays were conducted in triplicate, and the results are presentedas the mean ± standard deviation (SD). The data were processed usingthe IBM SPSS® ver. 19.0 (SPSS, Chicago, IL, USA). One-way ANOVA andTukey’s multiple-range test were used to evaluate the significant dif-ferences (P < 0.05). Principal component analysis (PCA), factor ana-lysis (FA), correlation analysis, cluster analysis and regression analysiswere performed to evaluate the difference and relevance of glucosino-lates among genotypes and organs in broccoli (P < 0.05).

3. Results and discussion

3.1. Method validation and extraction development

Prevention of myrosinase activity is necessary during the glucosi-nolate extraction procedure in vegetative tissues. For this reason, sev-eral extraction methods have been used specifically to prevent the ac-tivation of myrosinase. Generally, extractions are conducted attemperatures of 65–100 °C, close to the water solvent or 70% aqueousmethanol boiling point. Currently, a well-known analytical methodused for glucosinolates is desulfonation of glucosinolates with sulfatasewith subsequent analysis using a RP-HPLC gradient system (Baik, Juvik,Jeffery, Wallig, Kushad, & Klein, 2003; Wang, Gu, Yu, Zhao, Sheng, &Zhang, 2012). These methods provide reliable quantitative data andinformation related to glucosinolate variations, but they require muchtime and labor for analyses and require special equipment. Therefore, asimple method for the quantitative analysis of glucosinolates is neces-sary for the fields of plant breeding and food processing.

In this study, Table S1 presents the validation results, and thecoefficients of determination (R2) were 0.97 and 0.99 for glucoraphaninand gluconapin, respectively. Linear fitting was preferred to avoidpossible overfitting by higher order curves. The limits of detection(LODs) and limits of quantification (LOQs) were indicative of thepractical suitability for determining the two compounds. In terms ofaccuracy, most points appear to fit between the 75.5–83.2% ranges and80.1–89.1% ranges for glucoraphanin and gluconapin, respectively.Finally, the RSD (%) value for stability was lower than 5% in the de-termination. According to the determination of florets in broccoli from80 genotypes, twelve glucosinolate compounds were detected byUHPLC-MS/MS (Table S2).

In previous studies, the determination methods of glucosinolatesinclude HPLC, HPLC-MS or LC-MS based on analysis of thiosine (de-sulfo-GSLs) compounds after desulfurization (B. G. Hansen,Kliebenstein, & Halkier, 2007; Stewart, Nho, & Jeffery, 2004; Tian,Yang, Avila, Fish, Yuan, Hui, et al., 2018). In the pretreatment, gluco-sinolates need to be desulfurized by a DEAE ion exchange column andsulfatase, so the whole process is complicated and time-consuming. Inthis study, a high-temperature water bath and ultrasonic methods wereused to directly extract intact glucosinolates from plant samples (Vo,Trenerry, Rochfort, White, & Hughes, 2014), greatly simplifying thesteps and time of sample preparation, and the extraction method istime-saving with good recovery and linearity (Table S1).

3.2. Spectrometry cracking and characteristics of glucosinolates based onTriple-TOF

The UHPLC-MS/MS technique enabled the precise and concurrentidentification of glucosinolates in broccoli byproducts that previouslyneeded separate sample treatments and chromatographic conditions(Fig. 2). The structure of glucosinolates is composed of β-D-glucose (Glc)linked to a sulfonate aldoxime group and a side-chain R derived from anamino acid. In view of the hydrophilic anionic nature of glucosinolates,the compounds have a good mass spectrometric response in the nega-tive ion mode ESI. Under CID collision energy, the R group of the aminoacid side chain is broken, and it will be further broken to produce

characteristic product ions and characteristic neutral missing fragmentions. The bond between the sulfur atom in the sulfonate aldehyde groupand adjacent carbon atom is broken, and the hydrogen atom is rear-ranged to produce the characteristic product ion of glucosinolate. Them/z values are 290.9844, 274.9895, 259.0124, 241.0018, 195.0327,96.9596 and 79.9568; in addition, the characteristic neutral loss ofglucosinolates can also occur, and the neutral loss groups are SO (379.9568 u), Glc (162.0528 u), SGlc (195.0327 u), 'SGlc-OH' (178.03 uand 'Glc + SO3′ (242.0096 u). Using these characteristic product ionsand neutral lost groups, we can not only analyze the molecular struc-ture of glucosinolate compounds in the MS/MS spectra but also filterfragments and neutral losses by using the Fragment and Neutral LossFilter function in PeakView software. Glucosinolates are screened andextracted accurately and quickly in each information association-ionscanning channel.

In addition to the above-described characteristic product ions andneutral loss fragments from the glucosinolate consensus group, the side-chain groups also produce important fragment ions. Methyl sulfinylglucosinolate, which has a neutral loss of a side-chain methylsulfonyl(CH3SO) under CID collision energy, produces high-intensity fragmentions. For example, 4-methylsulfinyl butyl glucosinolate, 3-methylsulfi-nylpropyl glucosinolate, and 5-methylsulfinyl pentyl glucosylate allhave a neutral loss of the side-chain methylsulfinyl group. The corre-sponding fragment ions m/z 372.0431, m/z 358.0275 and m/z386.0581 were generated, respectively.

Two isomers of guanidine glucosinolates were found in this study: 1-methoxy-3-mercaptomethyl glucosinolate and 4-methoxy-3-mercapto-methyl glucosinolate. These compounds have the same molecular for-mula (C17H22N2O10S2) and molecular ion peaks ([MH]-, m/z 477.0643)but have different molecular structures in which the methoxy group hasdifferent substituent positions in the anthracene ring. The extracted ionchromatograms of the two isomers are shown in Fig. 2, and it is seenthat complete separation of the chromatogram is achieved with thismethod. Due to the difference in the molecular structure of the twoisomers, the mass spectrometry cleavage pathways of the two isomersare different, resulting in different fragment ions. 1-Methoxy-3′-mer-captomethyl sulfate is cleavable at the N-atom of the methoxy groupand indole ring; the NO bond is broken by CID collision energy, re-sulting in the loss of methoxy group, and the fragment ion m/z446.0479 is obtained. The main fragment ion of glycoside representsthe common group of glucosinolates and their fragment ions, which arecharacteristic of the glucosides, with values of m/z 274.9920,259.0145, 195.0333, 96.9603, and 74.9915 and a retention time of3.8 min. The distinction between two steroidal glucosinolate isomers isaccomplished by observing the TOF-MS/MS fragment ions (Zheng,et al., 2017).

3.3. Evaluation of glucosinolate compounds in florets based on 80 broccoligenotypes

Glucosinolates and their secondary products are popularly re-commended as anticancer agents in humans, flavor control in cruci-ferous plants, and chemical barriers in pest control (Gorissen, Kraut, deVisser, de Vries, Roelofsen, & Vonk, 2011; Grubb & Abel, 2006; Mithen,Bennett, & Marquez, 2010). In this study, the florets of a total of 80broccoli genotypes from around the world were characterized, andtwelve glucosinolate compounds were detected in the mature broccoliflorets (Table 1). Of these twelve glucosinolates, six belonged to ali-phatic glucosinolates, namely, glucoraphanin, glucoerucin, glucoiberin,gluconapin, progoitrin and sinigrin; four belonged to indole glucosi-nolates, namely, 4-methoxyglucobrassicin, glucobrassicin, neogluco-brassicin and 4-hydroxyglucobrassicin; and only two were phenyl glu-cosinolates, namely, gluconasturtiin and glucotropaeolin (Table 1). Ofthe twelve glucosinolates detected in florets of broccoli, some com-pounds were consistent with previous reports, and a unique glucosi-nolate, glucotropaeolin, was also detected in this work (L. P. Guo, Yang,

Z. Li, et al. Food Chemistry 334 (2021) 127519

3

& Gu, 2016; R. F. Guo, Yuan, & Wang, 2011; Wang, Gu, Yu, Zhao,Sheng, & Zhang, 2012).

In Brassica plants, glucosinolate varieties and contents are closelyrelated to A, B and C genomes (Farnham, Wilson, Stephenson, & Fahey,2004; Liu, Liu, Yang, Tong, Edwards, Parkin, et al., 2014). It has beenreported that Brassica plants commonly contain 18 to 27 glucosinolates(L. P. Guo, Yang, & Gu, 2016; R. F. Guo, Yuan, & Wang, 2011; M.Hansen, Moller, Sorensen, & Detrejo, 1995; Wang, Gu, Yu, Zhao, Sheng,& Zhang, 2012), with approximately twelve aliphatic glucosinolates(Cartea, Velasco, Obregon, Padilla, & de Haro, 2008). Glucoiberin, si-nigrin, and glucoerucin are usually present in B. oleracea vegetablessuch as cabbage, broccoli, kohlrabi and cauliflower, and sinigrin is alsoproduced in mustard green (B. juncea). Glucoraphanin, a functionalcomponent, is found in B. oleracea vegetables but is abundant inbroccoli. Gluconapin and progoitrin are commonly present in B. rapavegetables such as Chinese cabbage, mustard and turnip, B. oleraceavegetables such as cabbage, broccoli and cauliflower, B. juncea vege-tables, mainly in mustard green, and B. napus (rapeseed). Dehydroer-ucin is the dominant aliphatic glucosinolate in radish, accounting forover 80% of the total glucosinolates, and glucobrassicanapin is themain glucosinolate in B. rapa vegetables (L. P. Guo, Yang, & Gu, 2016;Ishida, Hara, Fukino, Kakizaki, & Morimitsu, 2014; Ishida, Nagata,Ohara, Kakizaki, Hatakeyama, & Nishio, 2012). However, some glu-cosinolates, such as gluconasturtiin and glucotropaeolin, are rarelydetected in B. oleracea vegetables, but we have detected both of thoseglucosinolate compounds, although they were in a relatively low con-tent in broccoli florets shown in Table 1.

The glucosinolate concentrations in Brassicaceae are influenced bygenotypes, plant tissues, growth and harvest time, environmental fac-tors such as climate, and cultivation conditions including fertilizationand soil, and genetic variations in the composition and contents ofglucosinolates have been reported (Farnham, Wilson, Stephenson, &Fahey, 2004; Ishida, Nagata, Ohara, Kakizaki, Hatakeyama, & Nishio,2012; G. Li & Quiros, 2003). However, the main factor influencing theamount of glucosinolates is usually dependent on genotype and as-sessed by investigating glucosinolate changes in different species, en-vironments, or genotype-environment interactions (Fahey, Zalcmann, &Talalay, 2001; Farnham, Wilson, Stephenson, & Fahey, 2004; L. P. Guo,Yang, & Gu, 2016). In our work, twelve glucosinolate compounds weredetected in the florets of broccoli, and seven glucosinolates, glucor-aphanin, glucoerucin, 4-methoxyglucobrassicin, glucobrassicin, neo-glucobrassicin, 4-hydroxyglucobrassicin and gluconasturtiin, were allfound in mature florets. Two glucosinolates, glucoiberin and gluco-tropaeolin, were found in most materials, and three glucosinolates,gluconapin and progoitrin and sinigrin, were just found in a fewbroccoli genotypes. In contrast to previous studies, gluconasturtiin andespecially glucotropaeolin were detected in florets, and both glucosi-nolate compounds were present at relatively low levels, which might bethe reason that they were not detected in broccoli (Farnham, Wilson,

Stephenson, & Fahey, 2004; L. P. Guo, Yang, & Gu, 2016; Wang, Gu, Yu,Zhao, Sheng, & Zhang, 2012).

As shown in Table 1, we found that the total content of glucosino-lates in florets ranged from 0.467 to 57.156 µmol/g DW in the 80genotypes, with the highest content being approximately 122-foldhigher than the lowest: glucoraphanin ranging from 0.136 to14.973 µmol/g DW, glucoerucin ranging from 0.008 to 6.273 µmol/gDW, glucoiberin ranging from 0 to 0.878 µmol/g DW, gluconapinranging from 0 to 2.728 µmol/g DW, progoitrin ranging from 0 to4.537 µmol/g DW, sinigrin ranging from 0 to 3.161 µmol/g DW, 4-methoxyglucobrassicin ranging from 0.014 to 3.915 µmol/g DW, glu-cobrassicin ranging from 0.103 to 27.690 µmol/g DW, neogluco-brassicin ranging from 0.018 to 45.954 µmol/g DW, 4-hydro-xyglucobrassicin ranging from 0.014 to 3.289 µmol/g DW,gluconasturtiin ranging from 0.004 to 0.441 µmol/g DW, and gluco-tropaeolin ranging from 0 to 0.040 µmol/g DW. According to the re-sults, < 22% of the genotypes contained the compounds gluconapin,progoitrin or sinigrin, and the corresponding contents of the threeglucosinolates were all at low levels (< 4.537 µmol/g DW), which co-incides with some reports but differs from several others (high contentin broccoli) (Gorissen, Kraut, de Visser, de Vries, Roelofsen, & Vonk,2011; L. P. Guo, Yang, & Gu, 2016; Mithen, Bennett, & Marquez, 2010;Wang, Gu, Yu, Zhao, Sheng, & Zhang, 2012), suggesting that somemutant plants exist in these materials.

Table 1 shows the presence of neoglucobrassicin and gluco-tropaeolin in some genotypes of broccoli, which not only provides newinformation but also validates previous research in B. oleracea broccoli(Wang, Gu, Yu, Zhao, Sheng, & Zhang, 2012). Few reports state thatgluconapoleiferin might be detected in broccoli. However, the com-pound is widely and commonly detected in B. rapa Chinese cabbage(Wiesner, Zrenner, Krumbein, Glatt, & Schreiner, 2013). Meanwhile, wefound a large variation in glucosinolate amounts in different genotypesof broccoli, enriching previous research.

A dendrogram using ward linkage was established based on a re-scaled distance cluster combined with SPSS 19.0 software (Fig. S1).From Figure S1, the cluster analysis presented 80 genotypes that can bedivided into three subgroups; bred lines B100 and B125 were clusteredtogether and contained all twelve glucosinolates, which was differentfrom the other two large clusters containing only some glucosinolates.Therefore, the results suggested that Brassica plants with differentcompounds and contents of glucosinolates, including broccoli, cabbage,cauliflower, Chinese cabbage and so on, should be cultivated andchosen by humans. Nilsson et al. (2006) detected glucobrassicin andglucoiberin as the main glucosinolates in kale, while sinigrin was foundas the major glucosinolate in white cabbage (Nilsson, Olsson, Engqvist,Ekvall, Olsson, Nyman, et al., 2006). Glucoiberin was reported as theonly important glucosinolate in kales, while glucoraphanin, sinigrin,glucobrassicin, and glucoiberin were found to be the most abundant incabbage (Kushad, Brown, Kurilich, Juvik, Klein, Wallig, et al., 1999).

Fig. 2. The total glucosinolate chromatogram (TIC)corresponding to the analysis of broccoli sampleswith glucosinolates by UHPLC/TOF-MS/MS and se-parate MS/MS spectra of 4-hydroxyglucobrassicin, 4-methoxyglucobrassicin, glucobrassicin, glucor-aphanin, glucotropaeolin, neoglucobrassicin, glu-coerucin, glucoiberin, gluconapin, gluconasturtiin,progoitrin, and sinigrin. The retention time of in-dividual glucosinolate was shown in Table S2.

Z. Li, et al. Food Chemistry 334 (2021) 127519

4

Table1

Gluco

sino

late

conten

tsin

floretsba

sedon

differen

tge

notype

sof

broc

coli(umol/g

DW).

No

Alip

haticsgluc

osinolates

Indo

lesgluco

sino

lates

Phen

ylgluc

osinolates

GRAa

GER

GIB

NAP

PRO

SIN

4MGBS

GBS

NGBS

4HGBS

GNS

GTL

B57

10.401

±0.94

6b0.07

0±

0.00

70.01

4±

0.00

2ND

ND

ND

0.23

2±

0.02

56.00

7±

0.64

63.36

6±

0.37

00.85

3±

0.07

80.12

3±

0.01

5ND

B58

6.38

0±

0.58

00.11

7±

0.01

20.21

5±

0.02

7ND

ND

ND

0.96

0±

0.10

419

.816

±2.13

113

.059

±1.43

50.38

3±

0.03

50.44

1±

0.05

40.01

3±

0.00

2B5

93.11

8±

0.28

30.48

5±

0.05

10.00

9±

0.00

1ND

ND

ND

1.70

1±

0.18

57.38

6±

0.79

410

.783

±1.18

51.17

6±

0.10

70.11

7±

0.01

40.01

1±

0.00

1B6

00.60

2±

0.05

50.12

3±

0.01

3ND

0.03

8±

0.00

40.46

7±

0.05

5ND

0.17

4±

0.01

91.57

7±

0.17

01.20

4±

0.13

20.09

0±

0.00

80.03

3±

0.00

4ND

B61

2.44

2±

0.22

20.77

3±

0.08

10.29

5±

0.03

7ND

ND

ND

0.74

3±

0.08

113

.552

±1.45

73.09

4±

0.34

01.62

9±

0.14

80.06

9±

0.00

90.00

9±

0.00

1B6

20.98

4±

0.08

90.08

0±

0.00

8ND

0.39

0±

0.04

40.44

6±

0.05

2ND

0.25

8±

0.02

85.32

4±

0.57

30.93

4±

0.10

30.17

3±

0.01

60.04

2±

0.00

50.01

1±

0.00

1B6

31.42

5±

0.13

01.30

0±

0.13

7ND

ND

ND

ND

0.07

6±

0.00

83.07

2±

0.33

00.42

9±

0.04

70.27

4±

0.02

50.09

7±

0.01

2ND

B64

0.52

9±

0.04

80.30

8±

0.03

20.07

6±

0.01

0ND

ND

ND

0.84

3±

0.09

25.66

6±

0.60

93.12

9±

0.34

41.67

2±

0.15

20.02

8±

0.00

30.01

6±

0.00

2B6

52.66

2±

0.24

26.27

3±

0.66

00.16

4±

0.02

00.43

8±

0.04

90.88

9±

0.10

50.29

8±

0.03

81.05

3±

0.11

411

.514

±1.23

813

.740

±1.51

00.99

9±

0.09

10.17

4±

0.02

10.01

1±

0.00

1B6

60.35

1±

0.03

20.01

3±

0.00

1ND

ND

ND

ND

0.02

7±

0.00

30.69

3±

0.07

40.01

8±

0.00

20.04

3±

0.00

40.00

9±

0.00

1ND

B67

1.07

7±

0.09

80.05

2±

0.00

5ND

ND

ND

ND

0.02

9±

0.00

30.20

1±

0.02

20.12

9±

0.01

40.04

6±

0.00

40.01

7±

0.00

2ND

B68

0.13

6±

0.01

20.00

8±

0.00

1ND

0.00

8±

0.00

1ND

ND

0.01

7±

0.00

20.34

3±

0.03

70.10

8±

0.01

20.01

4±

0.00

10.01

0±

0.00

1ND

B69

1.25

7±

0.11

40.31

1±

0.03

30.00

8±

0.00

1ND

ND

ND

0.73

3±

0.08

00.74

6±

0.08

03.32

5±

0.36

51.23

5±

0.11

20.04

7±

0.00

60.00

8±

0.00

1B7

09.29

0±

0.84

50.29

7±

0.03

10.02

1±

0.00

3ND

ND

ND

0.78

7±

0.08

68.65

2±

0.93

04.50

5±

0.49

51.46

7±

0.13

30.18

8±

0.02

30.02

1±

0.00

3B7

11.85

0±

0.16

80.17

6±

0.01

9ND

ND

ND

ND

0.13

2±

0.01

43.75

1±

0.40

31.98

0±

0.21

80.40

9±

0.03

70.03

2±

0.00

40.00

9±

0.00

1B7

21.27

9±

0.11

60.12

6±

0.01

3ND

ND

ND

ND

0.35

4±

0.03

84.77

2±

0.51

33.74

3±

0.41

10.87

0±

0.07

90.03

8±

0.00

50.00

8±

0.00

1B7

30.16

9±

0.01

50.00

8±

0.00

10.00

7±

0.00

1ND

ND

ND

0.01

4±

0.00

20.10

3±

0.01

10.08

7±

0.01

00.07

0±

0.00

60.00

8±

0.00

1ND

B74

2.15

5±

0.19

60.03

0±

0.00

30.11

8±

0.01

50.28

8±

0.03

20.92

1±

0.10

80.32

1±

0.04

10.28

2±

0.03

13.36

4±

0.36

21.60

9±

0.17

70.46

5±

0.04

20.02

7±

0.00

30.01

9±

0.00

2B7

53.34

4±

0.30

40.41

2±

0.04

30.14

3±

0.01

8ND

ND

ND

0.36

4±

0.04

08.13

2±

0.87

47.66

3±

0.84

22.72

8±

0.24

80.11

2±

0.01

40.02

4±

0.00

3B7

60.60

4±

0.05

50.48

4±

0.05

10.01

3±

0.00

2ND

ND

ND

0.32

6±

0.03

50.33

7±

0.03

60.44

9±

0.04

90.49

2±

0.04

50.02

5±

0.00

3ND

B77

5.05

9±

0.46

01.35

3±

0.14

20.00

9±

0.00

1ND

ND

ND

1.36

9±

0.14

912

.279

±1.32

02.54

3±

0.27

90.90

3±

0.08

20.11

1±

0.01

40.00

7±

0.00

1B7

811

.346

±1.03

10.26

0±

0.02

70.04

3±

0.00

50.58

1±

0.06

52.09

3±

0.24

60.03

8±

0.00

50.54

0±

0.05

97.12

8±

0.76

61.48

2±

0.16

31.46

3±

0.13

30.26

2±

0.03

2ND

B79

3.80

2±

0.34

60.87

8±

0.09

2ND

ND

ND

ND

0.66

5±

0.07

227

.690

±2.97

716

.730

±1.83

92.27

4±

0.20

70.15

5±

0.01

90.04

0±

0.00

5B8

02.38

9±

0.21

70.04

6±

0.00

5ND

ND

ND

ND

0.21

8±

0.02

40.93

6±

0.10

15.00

2±

0.55

00.18

2±

0.01

70.02

0±

0.00

20.00

9±

0.00

1B8

114

.973

±1.36

11.11

2±

0.11

70.02

5±

0.00

3ND

ND

ND

0.43

9±

0.04

87.22

7±

0.77

70.63

2±

0.06

91.18

1±

0.10

70.23

6±

0.02

90.01

6±

0.00

2B8

20.96

1±

0.08

70.20

0±

0.02

1ND

0.05

9±

0.00

70.19

7±

0.02

3ND

0.38

4±

0.04

23.80

3±

0.40

93.00

8±

0.33

10.16

0±

0.01

50.05

5±

0.00

7ND

B83

0.33

0±

0.03

00.07

2±

0.00

80.03

4±

0.00

40.04

8±

0.00

50.10

4±

0.01

20.04

6±

0.00

60.79

3±

0.08

64.51

0±

0.48

57.79

7±

0.85

70.47

2±

0.04

30.02

2±

0.00

30.01

8±

0.00

2B8

41.44

2±

0.13

10.13

9±

0.01

50.07

5±

0.00

9ND

ND

ND

0.70

4±

0.07

77.05

1±

0.75

84.92

6±

0.54

10.25

2±

0.02

30.22

6±

0.02

80.01

0±

0.00

1B8

51.37

1±

0.12

50.17

3±

0.01

8ND

ND

ND

ND

0.68

0±

0.07

41.91

3±

0.20

64.41

0±

0.48

51.28

4±

0.11

70.08

3±

0.01

0ND

B86

1.73

5±

0.15

80.17

1±

0.01

8ND

ND

ND

ND

0.24

5±

0.02

72.42

4±

0.26

10.52

0±

0.05

71.20

5±

0.11

00.12

2±

0.01

5ND

B87

4.17

0±

0.37

90.12

1±

0.01

3ND

ND

ND

ND

0.48

9±

0.05

37.85

0±

0.84

42.77

0±

0.30

40.71

1±

0.06

50.04

9±

0.00

60.01

4±

0.00

2B8

81.50

2±

0.13

70.11

0±

0.01

2ND

ND

ND

ND

0.37

9±

0.04

113

.021

±1.40

00.65

5±

0.07

20.80

1±

0.07

30.02

2±

0.00

30.01

1±

0.00

1B8

91.95

3±

0.17

80.25

5±

0.02

70.34

9±

0.04

4ND

ND

ND

0.08

2±

0.00

90.99

9±

0.10

71.39

5±

0.15

30.76

4±

0.06

90.03

9±

0.00

50.01

6±

0.00

2B9

00.92

0±

0.08

40.81

8±

0.08

60.11

3±

0.01

4ND

ND

ND

0.20

3±

0.02

26.46

1±

0.69

53.29

1±

0.36

20.78

8±

0.07

20.09

2±

0.01

10.02

0±

0.00

3B9

12.57

2±

0.23

40.30

8±

0.03

20.24

9±

0.03

1ND

ND

ND

1.04

5±

0.11

47.70

3±

0.82

816

.589

±1.82

31.42

8±

0.13

00.10

9±

0.01

3ND

B92

6.34

2±

0.57

70.34

5±

0.03

6ND

2.41

9±

0.27

24.53

7±

0.53

4ND

1.02

4±

0.11

19.59

4±

1.03

22.42

5±

0.26

61.53

1±

0.13

90.18

9±

0.02

3ND

B93

6.73

7±

0.61

20.47

3±

0.05

00.01

4±

0.00

2ND

ND

ND

0.54

1±

0.05

910

.165

±1.09

331

.315

±3.44

10.67

1±

0.06

10.30

7±

0.03

80.01

0±

0.00

1B9

45.94

6±

0.54

10.07

1±

0.00

7ND

ND

ND

ND

0.41

7±

0.04

52.85

3±

0.30

70.93

7±

0.10

30.32

1±

0.02

90.05

0±

0.00

6ND

B95

0.89

6±

0.08

10.09

1±

0.01

00.10

9±

0.01

4ND

ND

ND

0.05

9±

0.00

60.28

4±

0.03

10.32

9±

0.03

60.08

5±

0.00

80.05

6±

0.00

7ND

B96

11.928

±1.08

40.28

1±

0.03

00.02

9±

0.00

4ND

ND

ND

0.51

2±

0.05

68.61

3±

0.92

615

.111

±1.66

01.20

0±

0.10

90.12

1±

0.01

5ND

B97

2.34

8±

0.21

30.50

0±

0.05

30.01

3±

0.00

2ND

ND

ND

1.35

3±

0.14

79.35

8±

1.00

69.15

9±

1.00

61.19

6±

0.10

90.06

2±

0.00

80.01

2±

0.00

1B9

81.21

9±

0.11

10.34

1±

0.03

6ND

0.15

1±

0.01

70.44

3±

0.05

2ND

1.40

5±

0.15

36.86

7±

0.73

88.79

9±

0.96

71.12

5±

0.10

20.03

4±

0.00

40.01

2±

0.00

1B9

911

.864

±1.07

91.43

2±

0.15

10.01

3±

0.00

2ND

ND

ND

1.45

3±

0.15

89.67

2±

1.04

08.38

9±

0.92

21.75

4±

0.15

90.05

1±

0.00

60.00

9±

0.00

1B1

008.54

8±

0.77

71.07

9±

0.11

40.55

7±

0.07

02.72

8±

0.30

73.10

8±

0.36

63.16

1±

0.40

01.15

0±

0.12

524

.015

±2.58

24.06

3±

0.44

71.12

0±

0.10

20.31

9±

0.03

90.03

4±

0.00

4B1

011.57

8±

0.14

30.75

8±

0.08

00.01

6±

0.00

2ND

ND

ND

1.31

1±

0.14

319

.692

±2.11

78.01

5±

0.88

10.71

0±

0.06

50.05

9±

0.00

70.01

8±

0.00

2B1

022.94

2±

0.26

70.52

5±

0.05

50.12

8±

0.01

6ND

ND

ND

0.17

2±

0.01

99.04

2±

0.97

21.14

0±

0.12

50.61

8±

0.05

60.08

7±

0.01

10.01

1±

0.00

1B1

030.40

5±

0.03

70.03

9±

0.00

4ND

ND

ND

ND

0.60

2±

0.06

51.60

4±

0.17

212

.884

±1.41

60.04

1±

0.00

40.02

6±

0.00

3ND

B104

5.88

4±

0.53

50.53

7±

0.05

7ND

0.57

9±

0.06

51.99

3±

0.23

4ND

0.45

6±

0.05

08.64

6±

0.93

03.02

2±

0.33

20.50

4±

0.04

60.05

6±

0.00

70.02

3±

0.00

3B1

050.45

9±

0.04

20.21

0±

0.02

20.87

8±

0.11

0ND

ND

0.01

4±

0.00

20.50

1±

0.05

48.10

2±

0.87

12.66

2±

0.29

20.22

8±

0.02

10.12

3±

0.01

50.00

9±

0.00

1B1

065.81

0±

0.52

80.16

5±

0.01

70.01

8±

0.00

2ND

ND

ND

0.32

1±

0.03

58.89

6±

0.95

73.26

0±

0.35

80.33

7±

0.03

10.04

3±

0.00

50.00

9±

0.00

1B1

0710

.773

±0.97

90.69

4±

0.07

30.02

2±

0.00

31.66

5±

0.18

73.42

9±

0.40

30.03

4±

0.00

40.25

1±

0.02

77.77

4±

0.83

65.53

0±

0.60

81.23

8±

0.11

30.09

4±

0.01

20.02

2±

0.00

3B1

081.35

4±

0.12

30.09

2±

0.01

0ND

ND

ND

ND

0.13

7±

0.01

50.60

0±

0.06

40.24

9±

0.02

70.60

3±

0.05

50.02

5±

0.00

3ND

B109

0.22

9±

0.02

10.02

4±

0.00

3ND

ND

ND

ND

0.04

0±

0.00

40.26

8±

0.02

90.24

5±

0.02

70.09

5±

0.00

90.01

4±

0.00

2ND

(con

tinuedon

next

page)

Z. Li, et al. Food Chemistry 334 (2021) 127519

5

Table1(con

tinued)

No

Alip

haticsgluc

osinolates

Indo

lesgluco

sino

lates

Phen

ylgluc

osinolates

GRAa

GER

GIB

NAP

PRO

SIN

4MGBS

GBS

NGBS

4HGBS

GNS

GTL

B110

1.65

3±

0.15

00.02

6±

0.00

30.01

6±

0.00

2ND

ND

ND

0.05

1±

0.00

61.56

4±

0.16

80.22

3±

0.02

50.24

4±

0.02

20.02

8±

0.00

3ND

B111

7.05

6±

0.64

16.18

6±

0.65

1ND

ND

ND

ND

1.54

0±

0.16

77.74

0±

0.83

23.30

6±

0.36

32.51

7±

0.22

90.20

3±

0.02

50.01

0±

0.00

1B1

123.69

2±

0.33

60.14

5±

0.01

50.00

9±

0.00

1ND

ND

ND

0.18

4±

0.02

08.76

2±

0.94

213

.444

±1.47

70.78

8±

0.07

20.04

5±

0.00

6ND

B113

14.123

±1.28

40.46

9±

0.04

90.52

2±

0.06

5ND

ND

ND

0.96

0±

0.10

45.11

7±

0.55

013

.644

±1.49

91.55

5±

0.14

10.26

8±

0.03

30.01

8±

0.00

2B1

143.01

5±

0.27

40.19

4±

0.02

0ND

ND

ND

ND

0.30

7±

0.03

32.90

5±

0.31

20.31

4±

0.03

50.28

1±

0.02

60.05

3±

0.00

7ND

B115

2.11

8±

0.19

30.12

9±

0.01

40.12

1±

0.01

5ND

ND

ND

0.18

7±

0.02

01.11

3±

0.12

01.16

3±

0.12

80.27

2±

0.02

50.01

7±

0.00

20.01

1±

0.00

1B1

165.40

8±

0.49

20.33

7±

0.03

6ND

ND

ND

ND

0.65

2±

0.07

14.11

8±

0.44

36.23

9±

0.68

60.73

9±

0.06

70.20

7±

0.02

60.01

3±

0.00

2B1

179.31

0±

0.84

61.05

4±

0.11

1ND

0.41

8±

0.04

70.82

4±

0.09

7ND

0.39

8±

0.04

313

.881

±1.49

35.78

8±

0.63

63.28

9±

0.29

90.02

4±

0.00

30.03

4±

0.00

4B1

184.08

0±

0.37

10.05

2±

0.00

50.01

0±

0.00

1ND

ND

ND

0.28

9±

0.03

13.79

3±

0.40

80.65

2±

0.07

20.26

2±

0.02

40.01

9±

0.00

20.00

7±

0.00

1B1

197.54

0±

0.68

51.73

7±

0.18

30.01

7±

0.00

2ND

ND

ND

0.10

5±

0.01

113

.938

±1.49

92.96

7±

0.32

61.99

8±

0.18

20.06

9±

0.00

9ND

B120

1.24

7±

0.11

30.35

1±

0.03

70.04

4±

0.00

60.15

1±

0.01

70.50

3±

0.05

90.15

4±

0.01

90.21

4±

0.02

37.40

6±

0.79

65.33

1±

0.58

60.68

5±

0.06

20.05

8±

0.00

70.01

0±

0.00

1B1

211.20

3±

0.10

90.24

4±

0.02

60.00

8±

0.00

1ND

ND

ND

0.62

8±

0.06

812

.955

±1.39

36.97

2±

0.76

60.19

2±

0.01

70.05

4±

0.00

70.00

9±

0.00

1B1

221.28

0±

0.11

60.07

0±

0.00

70.12

2±

0.01

5ND

ND

ND

0.14

8±

0.01

61.71

6±

0.18

52.65

9±

0.29

20.36

6±

0.03

30.03

0±

0.00

4ND

B123

4.07

0±

0.37

00.04

7±

0.00

50.02

1±

0.00

3ND

ND

ND

0.20

6±

0.02

28.37

8±

0.90

11.31

0±

0.14

40.16

8±

0.01

50.04

4±

0.00

50.02

1±

0.00

3B1

246.27

8±

0.57

10.16

3±

0.01

70.00

9±

0.00

1ND

ND

ND

0.46

3±

0.05

08.50

1±

0.91

49.06

0±

0.99

61.13

3±

0.10

30.07

4±

0.00

90.00

7±

0.00

1B1

257.18

4±

0.65

30.88

0±

0.09

30.43

9±

0.05

52.06

9±

0.23

23.23

5±

0.38

12.23

8±

0.28

31.34

2±

0.14

618

.092

±1.94

54.63

8±

0.51

00.98

8±

0.09

00.34

8±

0.04

30.03

3±

0.00

4B1

2611

.409

±1.03

70.08

1±

0.00

90.03

0±

0.00

4ND

ND

ND

0.26

3±

0.02

94.18

6±

0.45

014

.485

±1.59

20.43

8±

0.04

00.04

0±

0.00

50.01

0±

0.00

1B1

272.99

9±

0.27

30.04

4±

0.00

50.23

3±

0.02

90.31

9±

0.03

61.43

8±

0.16

90.42

2±

0.05

30.19

7±

0.02

12.44

6±

0.26

311

.431

±1.25

60.24

4±

0.02

20.04

3±

0.00

50.01

3±

0.00

2B1

280.17

5±

0.01

60.03

3±

0.00

40.01

2±

0.00

2ND

ND

ND

0.06

0±

0.00

70.56

0±

0.06

00.09

5±

0.01

00.06

7±

0.00

60.00

4±

0.00

00.00

1±

0.00

0B1

294.53

3±

0.41

20.34

7±

0.03

7ND

ND

ND

ND

2.47

4±

0.26

95.54

9±

0.59

78.79

1±

0.96

60.77

2±

0.07

00.09

8±

0.01

20.02

1±

0.00

3B1

304.58

3±

0.19

1±

0.02

00.01

2±

0.00

2ND

ND

ND

1.15

5±

0.12

64.62

4±

0.49

745

.954

±5.05

00.59

3±

0.05

40.02

9±

0.00

40.01

5±

0.00

2B1

317.35

6±

0.66

90.22

8±

0.02

40.01

0±

0.00

1ND

ND

ND

3.91

5±

0.42

612

.814

±1.37

83.36

2±

0.36

92.52

2±

0.22

90.20

8±

0.02

60.00

9±

0.00

1B1

3211

.891

±1.08

10.49

6±

0.05

20.01

6±

0.00

2ND

ND

ND

0.60

2±

0.06

518

.573

±1.99

70.30

9±

0.03

42.92

7±

0.26

60.21

7±

0.02

70.01

1±

0.00

1B1

335.19

7±

0.47

22.20

8±

0.23

20.01

1±

0.00

1ND

ND

ND

1.39

5±

0.15

223

.937

±2.57

42.47

8±

0.27

22.42

8±

0.22

10.16

7±

0.02

10.01

3±

0.00

2B1

344.09

9±

0.37

30.18

2±

0.01

9ND

ND

ND

ND

0.75

9±

0.08

30.92

9±

0.10

07.72

3±

0.84

90.05

8±

0.00

50.07

6±

0.00

90.01

3±

0.00

2B1

355.10

5±

0.46

40.17

8±

0.01

9ND

0.97

3±

0.10

92.87

2±

0.33

8ND

0.22

0±

0.02

43.69

5±

0.39

711

.122

±1.22

21.18

1±

0.10

70.02

8±

0.00

30.01

9±

0.00

2B1

368.87

8±

0.80

71.02

0±

0.10

70.77

9±

0.09

7ND

ND

ND

0.16

9±

0.01

80.94

0±

0.10

19.15

9±

1.00

60.82

5±

0.07

50.03

9±

0.00

50.01

0±

0.00

1Meanc

4.19

2±

3.74

20.52

8±

1.01

40.07

8±

0.16

70.16

7±

0.50

90.34

4±

0.90

70.08

4±

0.43

30.59

9±

0.61

37.00

3±

6.04

35.69

1±

7.08

90.87

3±

0.74

90.09

3±

0.09

00.01

0±

0.00

9Ran

ge0.13

6–14

.973

0.00

8–6.27

30–

0.87

80–

2.72

80–

4.53

70–

3.16

10.01

4–3.91

50.10

3–27

.690

0.01

8–45

.954

0.01

4–3.28

90.00

4–0.44

10–

0.04

0

aGluco

sino

latesab

brev

iation

s:GS:

gluc

osinolates,GRA:gluc

orap

hanin(4-m

ethy

lsulph

inylbu

tylGS);GER

:gluc

oerucin(4-m

ethy

lthiob

utyl

GS);GIB:gluc

oibe

rin(3-m

ethy

lsulfiny

lpropy

lGS);NAP:

gluc

onap

in(3-

buteny

lGS);PR

O:prog

oitrin

(2-R-hyd

roxy

-3-buten

ylGS);SIN:sinigrin

(2-prope

nylGS);4M

GBS

:4-metho

xygluc

obrassicin

(4-m

etho

xy-3-ind

olylmethy

lGS);GBS

:gluc

obrassicin

(3-ind

olylmethy

lGS);NGBS

:ne

oglu-

cobrassicin(1-m

etho

xy-3-ind

olylmethy

lGS);4

HGBS

:4-hyd

roxy

gluc

obrassicin

(4-hyd

roxy

-3-ind

olylmethy

lGS);G

NS:

Gluco

nasturtiin

(2-phe

nylethyl

GS);G

TL:g

luco

trop

aeolin

(Ben

zylgluc

osinolate).

bValue

sareexpressedas

mean

±stan

dard

deviation(SD)(n

=3),a

nd“N

D”stan

dsforno

tde

tected

.cMeanpresen

tedas

theav

erag

e±

SD.

Z. Li, et al. Food Chemistry 334 (2021) 127519

6

In our work, the principal component analysis presented three fac-tors that explained 60.53% of the variance, and the correspondingcontributions were 34.28%, 17.17% and 9.08% (Table S3), and theglucosinolate concentration varied depending on the genotype. FromTable S4, component 1 presented a positive correlation among thetwelve factors, directly corresponding to total glucosinolate amounts.Component 2 showed four factors with a negative correlation amongglucoiberin, gluconapin, progoitrin and sinigrin of the downstreamproducts of C3 and C4 aliphatic glucosinolates, which corresponded tothe up- and downstream regulation of glucosinolates. In component 3,most of the factors were negatively correlated, and a complex re-lationship was observed among them. From Figure S2, we could clearlyfind the principal glucosinolates in 80 broccoli genotypes, and as wellas the positive and negative relationships of total glucosinolate andindependent component.

From the correlation based on the Pearson test, we found positivecorrelations among hydroxyglucobrassicin, methoxyglucobrassicin,glucobrassicin, glucoerucin, gluconasturtiin, glucoraphanin, and glu-cotropaeolin (P < 0.05). The aliphatic glucosinolates, glucoiberin,gluconapin, gluconasturtiin, glucotropaeolin, progoitrin, and sinigrin,were related correspondingly (P < 0.05), providing validation andevidence regarding the regulatory pathways of aliphatic glucosinolates(Table S5). However, it must be noted that there was still a problemregarding glucoraphanin and gluconapin in this study. As shown inprevious reports, AOP2 (BoGSL-ALK) could play a role in B. rapa andcollard, producing gluconapin from glucoraphanin, but in broccoli,there is the presence of a nonfunctional allele BoGSL-ALK (GenBank no.AY044424) originating from a 2-bp deletion in exon 2 (G. Li & Quiros,2003; Sønderby, Geu-Flores, & Halkier, 2010); thus, gluconapin shouldbe absent in broccoli florets, but in some previous reports and from 18genotypes in this study, gluconapin was detected in florets (B. G.Hansen, Kliebenstein, & Halkier, 2007; G. Li & Quiros, 2003; Stewart,Nho, & Jeffery, 2004), which is an obvious contradiction in broccoli(Fig. 3). As mutant materials, 18 broccoli genotypes have been studiedregarding AOP2 regulatory function.

In addition, glucosinolates are usually influenced by environmental

factors such as soil, climate and cultivation conditions including ferti-lization, harvest time, plant organ, and physical damage. However,wide genetic variations in the glucosinolate contents and compoundshave been reported from previous studies, and glucosinolates are af-fected by genotype and environment (Carlson, Daxenbichler, VanEtten,Tookey, & Williams, 1981; Ishida, Nagata, Ohara, Kakizaki,Hatakeyama, & Nishio, 2012; G. Li & Quiros, 2003). The accumulationof aliphatic glucosinolates in B. rapa is enhanced by low nitrogen andhigh sulfur supply, and high temperature can enhance the accumulationof aliphatic glucosinolates (Volker, Freeman, Banuelos, & Jeffery,2010). However, the change and induction of molecular and transportmechanisms of glucosinolates remain to be further studied in the future.

3.4. Evaluation of glucosinolates in different broccoli organs

Among the many varieties of vegetables, Brassicaceae vegetableshave received the most attention because their unique constituents,glucosinolates, are abundant in edible parts and are regarded as mostlikely to support human health through continuous consumption. Forthe future, the breeding of Brassicaceae vegetables by particularly ad-dressing beneficial glucosinolates is expected to grow in importance(Gorissen, Kraut, de Visser, de Vries, Roelofsen, & Vonk, 2011; Mithen,Bennett, & Marquez, 2010).

To better define the relationship between glucosinolate content andleaf development, six genotypes and their eight organs were plantedand treated in this study. From the distribution of glucosinolate com-pounds in different organs of broccoli B1-B6 (Fig. 4), we found that thetotal glucosinolate content ranged from 0.214 to 73.527 µmol/g DW indifferent organs, with the highest content being approximately 343-foldthan the lowest value and much higher than that in florets. In thesematerials, a total of eight glucosinolate compounds were found exceptin B4, which had nine compounds in roots. Among the organs, sevenglucosinolates, 4-hydroxyglucobrassicin, 4-methoxyglucobrassicin,glucobrassicin, glucoerucin, gluconasturtiin, glucoraphanin and neo-glucobrassicin, were found, but with the following exceptions: all werefound in roots except glucotropaeolin, which was just found in B4; all

Fig. 3. Glucosinolate core structure and side-chain modification pathway for 3C, 4C and 5C aliphatic glucosinolates. The red frame presents glucosinolates detectedin broccoli florets, and the dotted green line indicates the potential query depending on the AOP2± gene (G. Li & Quiros, 2003; Mithen, Bennett, & Marquez, 2010).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Z. Li, et al. Food Chemistry 334 (2021) 127519

7

were found in stalks, except in B4 and B5; all were found in leaves,except glucoerucin; all florets presented glucoiberin, except in B1;young buds also presented glucoiberin; mature buds also presentedglucoiberin, except in B1-B3; buds before flowering also presentedglucoiberin, except in B1 and B5; and flowers also presented glucoi-berin. The glucosinolate content ranged from 8.482 to 73.527 µmol/gDW in roots with an average value of 41.518 µmol/g DW, 0.486 to3.475 µmol/g DW in stalks with an average value of 1.503 µmol/g DW,

0.214 to 1.563 µmol/g DW in leaves with an average value of0.682 µmol/g DW, 5.859 to 24.948 µmol/g DW in florets with anaverage value of 13.383 µmol/g DW, 8.144 to 17.009 µmol/g DW inflorets with an average value of 12.664 µmol/g DW, 7.262 to12.122 µmol/g DW in young buds with an average value of10.647 µmol/g DW, 1.363 to 15.093 µmol/g DW in buds before flow-ering with an average value of 6.790 µmol/g DW, and 4.568 to11.878 µmol/g DW in buds before flowering with an average value of

Fig. 4. The contents of nine glucosinolates in eight organs of six broccoli genotypes (B1-B6), the organs were roots (R), stalks (S), leaves (L), florets (Ft), young buds(Y), mature buds (M), buds before flowering (B) and flowers (Fr), and nine individual glucosinolate was glucoraphanin (GRA), glucoerucin (GER), glucoiberin (GIB),glucobrassicin (GBS), neoglucobrassicin (NGBS), 4hydroxyglucobrassicin (4MGBS), 4-Hydroxyglucobrassicin (4HGBS), glucotropaeolin (BGS), and gluconasturtiin(GNS).

Z. Li, et al. Food Chemistry 334 (2021) 127519

8

8.806 µmol/g DW.In Arabidopsis, the highest glucosinolate concentrations are found in

reproductive organs, including seeds, siliques, flowers and developinginflorescences, followed by young leaves, the root system and fullyexpanded leaves (Grubb & Abel, 2006; B. G. Hansen, Kliebenstein, &Halkier, 2007; Sønderby, Geu-Flores, & Halkier, 2010). Our datashowed that the roots contained almost 43% of the total glucosinolates,which was different from that in Arabidopsis. According to severalprevious reports, under deficiency of nitrogen or sulfur, the levels ofseveral glucosinolates decreased in leaves but increased in roots, whichsuggests that some glucosinolates are initially generated and accumu-lated in roots first (B. G. Hansen, Kliebenstein, & Halkier, 2007; Kushad,et al., 1999; Z. S. Li, et al., 2019). Meanwhile, some genes such asMYB68, MYB34/ATR1, MYB51/HIG1 and MYB122 can regulate theproduction of indolic glucosinolates in roots and late-stage rosetteleaves, which usually cause changes in indolic glucosinolate levels(Gigolashvili, Yatusevich, Rollwitz, Humphry, Gershenzon, & Flugge,2009). The florets and developmental buds contained 7% to 14% of thetotal glucosinolates, respectively, and the leaves and stalks contained1% and 2% of the total glucosinolates (Fig. 5A). Tracer studies havedemonstrated de novo synthesis in siliques and phloem transport ofglucosinolates from mature and senescing leaves to inflorescences anddeveloping fruits in Arabidopsis (Chen, Glawischnig, Jorgensen, Naur,Jorgensen, Olsen, et al., 2003; Mikkelsen, Petersen, Olsen, & Halkier,2002). Active and specific glucosinolate uptake into Brassica leaf pro-toplasts is mediated by a proton-coupled symporter (Chen, et al., 2003).

Glucosinolate profiles have been systematically monitored inArabidopsis during plant development and vary significantly betweentissues and organs (P. D. Brown, Tokuhisa, Reichelt, & Gershenzon,2003; Halkier & Gershenzon, 2006). However, previous research hasnot verified this finding in broccoli, so the result might provide newadditional evidence for glucosinolate metabolism research in Brassicaplants (Liang, Li, Yuan, & Vriesekoop, 2008). The distribution of eachglucosinolate in all the materials suggested some differences among thecompounds (Fig. 5B): glucoraphanin represented 29% of the total glu-cosinolates, followed by gluconasturtiin (24%), glucoerucin (15%),glucobrassicin (12%), neoglucobrassicin (8%), 4-hydroxyglucobrassicin(7%), and 4-methoxyglucobrassicin (5%). Glucoiberin and gluco-tropaeolin constituted 2% and 1% of the total glucosinolates, respec-tively, much less than that of glucoraphanin. The results were obviouslydifferent from those in florets of the 80 genotypes, and there are nosimilar reports in previous research that could provide useful evidenceand a basis for studying the metabolism and diversity of glucosinolatesin Brassica developmental organs.

In addition, in this study, the dendrogram, component plot (rotatedspace) and regression of glucosinolates in different broccoli organs werecarried out based on statistical analysis. According to comparison of

glucosinolate contents in different organs, some similar conclusionswere revealed in this study based on representative broccoli genotypeB1. From Figure S3, the related distance cluster combine of glucosi-nolates based on cluster analysis presented the root as an individualbranch, the stalks and leaves were close branches, and the florets, de-velopmental buds and flowers were relatively close branches. Thecluster result clearly revealed that the vegetative and reproductive or-gans of broccoli should be divided into three parts to analyze glucosi-nolate concentrations and structures. The roots, as a single branch ofvegetative organs, are different from other organs in glucosinolateproducts. All the reproductive organs exhibited one large branch var-iation based on genotype, including florets, buds, flowers, siliques andseeds, which usually contain more glucoraphanin in terms of totalglucosinolates, which were consistent with the previous reports (L. P.Guo, Yang, & Gu, 2016; Z. S. Li, et al., 2019). More information aboutbuds was also shown in this study, which were young buds with higherglucoraphanin content good for development of cruciferous tea. At thesame time, the regression analysis of glucosinolates indicated that therewere linear relationships between hydroxyglucobrassicin and gluco-brassicin, glucoraphanin, glucobrassicin and glucoiberin, gluco-tropaeolin. This result was consistent with the correlation analysis ofglucosinolates contents based on 80 genotypes, providing new insightinto glucosinolate compounds in different organs and genotypes, whichmight also help us evaluate the concentration of glucosinolates basedon a linear regression equation (Fig. S3). So far, few reports state si-milar results based on regression analysis. The principal componentanalysis was also carried out to reveal the characteristics of glucosi-nolates in different organs, and it was concluded that aliphatic gluco-sinolates were domain glucosinolates in broccoli, consistent with mostprevious reports; three components were basically described in thecomponent plot in rotated space (Fig. S3), but small differences werealso observed in the concentrations and compounds based on the B1-B6genotypes (Nilsson, et al., 2006).

4. Conclusion

Our work firstly provide a fast and reliable method for the de-termination of glucosinolates in broccoli based on UHPLC-Triple-TOF-MS, and the correlations among glucosinolates, genotypes and organsare all thoroughly determined by investigating variations in glucosi-nolate profiles in 80 genotypes and eight developmental organs. Twelveglucosinolates constitute three major variation components frombroccoli inbred lines, and the vegetative and reproductive organs con-tain different glucosinolate concentrations and components with re-spect to domain aliphatic glucosinolates. The statistical analysis ofglucosinolates indicates that the root is different from the other vege-tative organs with respect to glucosinolate metabolism, and the

Fig. 5. The percentage of total glucosinolate in different organs, including roots, stalks, leaves, florets, young buds, mature buds, buds before flowering and flowers(A), and the average percentage of nine glucosinolate distributed in all the materials (B).

Z. Li, et al. Food Chemistry 334 (2021) 127519

9

reproductive organs might be clustered together with a higher level ofglucoraphanin, which is helpful for humans. The study also mightprovide some insight into broccoli mutations for deep research on theAOP2 gene, as well as some methods to evaluate the major componentsand concentrations of glucosinolates in Brassica plants based on statis-tical analysis.

Author contributions

Authors Y. Liu, Z. Fang, L. Yang, and S. Zheng conducted the ex-periments. Authors Z. Li, D. Xu, M. Zhuang, Y. Zhang, Y. Wang, Y. Wangand H. Lv designed the studies and interpreted the results. Author Z. Liwrote the manuscript. Yumei Liu; Zhiyuan Fang; Limei Yang:Programming, software development; designing computer programs;implementation of the computer code and supporting algorithms;testing of existing code components.

Declaration of Competing Interest

The authors declare that they have no known competing financialinterests or personal relationships that could have appeared to influ-ence the work reported in this paper.

Acknowledgments

The present study was supported by grants from the National KeyResearch and Development Program of China (2017YFD0101805), theNational Nature Science Foundation (31872947, 31501761), the ChinaAgriculture Research System (CARS-23-A8, CARS-23-E03), the Scienceand Technology Innovation Program of the Chinese Academy ofAgricultural Sciences (CAAS-ASTIP-IVFCAAS), and the State KeyLaboratory of Vegetable Germplasm Innovation (SKL-VGI), CentralPublic-interest Scientific Institution Basal Research Fund (Y2020PT35).

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.foodchem.2020.127519.

References

Abdull Razis, A. F., Iori, R., & Ioannides, C. (2011). The natural chemopreventive phy-tochemical R-sulforaphane is a far more potent inducer of the carcinogen-detoxifyingenzyme systems in rat liver and lung than the S-isomer. Int J Cancer, 128(12),2775–2782.

Alumkal, J. J., Slottke, R., Schwartzman, J., Cherala, G., Munar, M., Graff, J. N., ... Mori,M. (2015). A phase II study of sulforaphane-rich broccoli sprout extracts in men withrecurrent prostate cancer. Invest New Drugs, 33(2), 480–489.

Baik, H. Y., Juvik, J., Jeffery, E. H., Wallig, M. A., Kushad, M., & Klein, B. P. (2003).Relating glucosinolate content and flavor of broccoli cultivars. J Food Sci, 68(3),1043–1050.

Bell, L., Oruna-Concha, M. J., & Wagstaff, C. (2015). Identification and quantification ofglucosinolate and flavonol compounds in rocket salad (Eruca sativa, Eruca vesicariaand Diplotaxis tenuifolia) by LC-MS: Highlighting the potential for improving nu-tritional value of rocket crops. Food Chem, 172, 852–861.

Bello, C., Maldini, M., Baima, S., Scaccini, C., & Natella, F. (2018). Glucoraphanin andsulforaphane evolution during juice preparation from broccoli sprouts. Food Chem,268, 249–256.

Brown, A. F., Yousef, G. G., Reid, R. W., Chebrolu, K. K., Thomas, A., Krueger, C., ... Juvik,J. A. (2015). Genetic analysis of glucosinolate variability in broccoli florets usinggenome-anchored single nucleotide polymorphisms. Theoretical and Applied Genetics,128(7), 1431–1447.

Brown, P. D., Tokuhisa, J. G., Reichelt, M., & Gershenzon, J. (2003). Variation of glu-cosinolate accumulation among different organs and developmental stages ofArabidopsis thaliana. Phytochemistry, 62(3), 471–481.

Carlson, D. G., Daxenbichler, M. E., VanEtten, C. H., Tookey, H. L., & Williams, P. H.(1981). Glucosinolates in crucifer vegetables: Turnips and rutabagas. J Agric FoodChem, 29(6), 1235–1239.

Cartea, M. E., Velasco, P., Obregon, S., Padilla, G., & de Haro, A. (2008). Seasonal var-iation in glucosinolate content in Brassica oleracea crops grown in northwesternSpain. Phytochemistry, 69(2), 403–410.

Chen, S. X., Glawischnig, E., Jorgensen, K., Naur, P., Jorgensen, B., Olsen, C. E., ...

Halkier, B. A. (2003). CYP79F1 and CYP79F2 have distinct functions in the bio-synthesis of aliphatic glucosinolates in Arabidopsis. Plant Journal, 33(5), 923–937.

de Oliveira, M. R., Brasil, F. D., & Furstenau, C. R. (2018). Sulforaphane PromotesMitochondrial Protection in SH-SY5Y Cells Exposed to Hydrogen Peroxide by anNrf2-Dependent Mechanism. Molecular Neurobiology, 55(6), 4777–4787.

Fahey, J. W., Zalcmann, A. T., & Talalay, P. (2001). The chemical diversity and dis-tribution of glucosinolates and isothiocyanates among plants. Phytochemistry, 56(1),5–51.

Farnham, M. W., Wilson, P. E., Stephenson, K. K., & Fahey, J. W. (2004). Genetic andenvironmental effects on glucosinolate content and chemoprotective potency ofbroccoli. Plant Breeding, 123(1), 60–65.

Gigolashvili, T., Yatusevich, R., Rollwitz, I., Humphry, M., Gershenzon, J., & Flugge, U. I.(2009). The Plastidic Bile Acid Transporter 5 Is Required for the Biosynthesis ofMethionine-Derived Glucosinolates in Arabidopsis thaliana. Plant Cell, 21(6),1813–1829.

Gorissen, A., Kraut, N. U., de Visser, R., de Vries, M., Roelofsen, H., & Vonk, R. J. (2011).No de novo sulforaphane biosynthesis in broccoli seedlings. Food Chem, 127(1),192–196.

Grubb, C. D., & Abel, S. (2006). Glucosinolate metabolism and its control. Trends in PlantScience, 11(2), 89–100.

Guo, L. P., Yang, R. Q., & Gu, Z. X. (2016). Cloning of genes related to aliphatic gluco-sinolate metabolism and the mechanism of sulforaphane accumulation in broccolisprouts under jasmonic acid treatment. J Sci Food Agric, 96(13), 4329–4336.

Guo, R. F., Yuan, G. F., & Wang, Q. M. (2011). Sucrose enhances the accumulation ofanthocyanins and glucosinolates in broccoli sprouts. Food Chem, 129(3), 1080–1087.

Halkier, B. A., & Gershenzon, J. (2006). Biology and biochemistry of glucosinolates.Annual Review of Plant Biology, 57, 303–333.

Hansen, B. G., Kliebenstein, D. J., & Halkier, B. A. (2007). Identification of a flavin-monooxygenase as the S-oxygenating enzyme in aliphatic glucosinolate biosynthesisin Arabidopsis. Plant Journal, 50(5), 902–910.

Hansen, M., Moller, P., Sorensen, H., & Detrejo, M. C. (1995). Glucosinolates in BroccoliStored under Controlled-Atmosphere. Journal of the American Society for HorticulturalScience, 120(6), 1069–1074.

Ishida, M., Hara, M., Fukino, N., Kakizaki, T., & Morimitsu, Y. (2014). Glucosinolatemetabolism, functionality and breeding for the improvement of Brassicaceae vege-tables. Breed Sci, 64(1), 48–59.

Ishida, M., Nagata, M., Ohara, T., Kakizaki, T., Hatakeyama, K., & Nishio, T. (2012). Smallvariation of glucosinolate composition in Japanese cultivars of radish (Raphanussativus L.) requires simple quantitative analysis for breeding of glucosinolate com-ponent. Breed Sci, 62(1), 63–70.

Kushad, M. M., Brown, A. F., Kurilich, A. C., Juvik, J. A., Klein, B. P., Wallig, M. A., &Jeffery, E. H. (1999). Variation of glucosinolates in vegetable crops of Brassicaoleracea. J Agric Food Chem, 47(4), 1541–1548.

Li, G., & Quiros, C. F. (2003). In planta side-chain glucosinolate modification inArabidopsis by introduction of dioxygenase Brassica homolog BoGSL-ALK. Theoreticaland Applied Genetics, 106(6), 1116–1121.

Li, Z. S., Liu, Y. M., Li, L. Y., Fang, Z. Y., Yang, L. M., Zhuang, M., ... Lv, H. H. (2019).Transcriptome reveals the gene expression patterns of sulforaphane metabolism inbroccoli florets. PLoS One, 14(3).

Liang, H., Li, C., Yuan, Q., & Vriesekoop, F. (2008). Application of high-speed counter-current chromatography for the isolation of sulforaphane from broccoli seed meal. JAgric Food Chem, 56(17), 7746–7749.

Liu, S., Liu, Y., Yang, X., Tong, C., Edwards, D., Parkin, I. A., ... Paterson, A. H. (2014).The Brassica oleracea genome reveals the asymmetrical evolution of polyploid gen-omes. Nature Communications, 5, 3930.

Mikkelsen, M. D., Petersen, B. L., Olsen, C. E., & Halkier, B. A. (2002). Biosynthesis andmetabolic engineering of glucosinolates. Amino Acids, 22(3), 279–295.

Mithen, R., Bennett, R., & Marquez, J. (2010). Glucosinolate biochemical diversity andinnovation in the Brassicales. Phytochemistry, 71(17–18), 2074–2086.

Nilsson, J., Olsson, K., Engqvist, G., Ekvall, J., Olsson, M., Nyman, M., & Akesson, B.(2006). Variation in the content of glucosinolates, hydroxycinnamic acids, car-otenoids, total antioxidant capacity and low-molecular-weight carbohydrates inBrassica vegetables. J Sci Food Agric, 86(4), 528–538.

Sønderby, I. E., Geu-Flores, F., & Halkier, B. A. (2010). Biosynthesis of glucosinolates –gene discovery and beyond. Trends in Plant Science, 15(5), 283–290.

Stewart, K. E., Nho, C. W., & Jeffery, E. H. (2004). Gene regulation by glucosinolate hy-drolysis products from broccoli. Phytochemicals: Mechanisms of Action107–119.

Tian, M., Yang, Y., Avila, F. W., Fish, T., Yuan, H., Hui, M., ... Li, L. (2018). Effects ofSelenium Supplementation on Glucosinolate Biosynthesis in Broccoli. J Agric FoodChem, 66(30), 8036–8044.

Vo, Q. V., Trenerry, C., Rochfort, S., White, J., & Hughes, A. B. (2014). Preparation and X-ray analysis of potassium (2,3-dichlorophenyl)glucosinolate. Acta Crystallogr C StructChem, 70(Pt 6), 588–594.

Volker, S. E., Freeman, J., Banuelos, G., & Jeffery, E. H. (2010). Impact of selenatetreatment on broccoli glucosinolate profile and bioactivity. Faseb Journal, 24.

Wang, J. S., Gu, H. H., Yu, H. F., Zhao, Z. Q., Sheng, X. G., & Zhang, X. H. (2012).Genotypic variation of glucosinolates in broccoli (Brassica oleracea var. italica)florets from China. Food Chem, 133(3), 735–741.

Wiesner, M., Zrenner, R., Krumbein, A., Glatt, H., & Schreiner, M. (2013). GenotypicVariation of the Glucosinolate Profile in Pak Choi (Brassica rapa ssp chinensis). JAgric Food Chem, 61(8), 1943–1953.

Zheng, S. N., Zhang, Y. G., Liu, G. Y., Lv, J., Lw, Z. X., Xu, X. M., ... Xu, D. H. (2017).Glycerolipid Profiling of Yellow Sarson Seeds Using Ultra High Performance LiquidChromatography Coupled to Triple Time-of-Flight Mass Spectrometry. HorticulturalPlant Journal, 3(4), 141–150.

Z. Li, et al. Food Chemistry 334 (2021) 127519

10