Analysis of Glucosinolate Content, Composition and Expression Level of Biosynthesis Pathway Genes in Different Chinese Kale Varieties

: The content and component of glucosinolates in edible stems and leaves of eight Chinese kale varieties from Japan and eight varieties from China were determined by HPLC-MS. Simultane-ously, the expression levels of glucosinolate biosynthesis pathway genes from four varieties with high and low total glucosinolate contents were analyzed by the qRT-PCR method. Four types of aliphatic glucosinolates (A-GLSs: GRA, SIN, GNA and GER) and indole glucosinolates (I-GLSs: 4-HGBS, GBS, 4-MGBS and NGBS) were detected in the stems and leaves of 16 varieties, and no aromatic glucosinolates (R-GLSs) were detected. A-GLSs account for more than 80.69% of the total content of total glucosinolates (T-GLSs), in which GNA and GRA are the main components of stems and leaves. Among Japanese varieties, QB1 has higher content of A- and T-GLSs, while that of XLB was lower; however, the corresponding varieties were ZH and DSHH in Chinese varieties. Among the above four varieties, the expression levels of SOT16 , CYP83B1 , SOT17 , CYP83A1 and MAM1 genes were signiﬁcantly higher in the varieties with higher GLSs; the expression levels of SOT16 and CYP83B1 were consistent with the content of I-GLSs; and SOT17 , CYP83A1 and MAM1 expression levels were consistent with A-GLSs content. At the same time, the expression levels of SOT16 and CYP83B1 in the leaves were higher than those in the stems. CYP83A1 and MAM1 genes were less expressed in the leaves than in the stems of lower content varieties. It is speculated that these genes may be the key genes regulating GLS biosynthesis in Chinese kale.


Introduction
Chinese kale (Brassica oleracea var. alboglabra) is a native Chinese vegetable belonging to the family of Brassicaceae and its edible organ includes tender stems, young leaves and bolting stems which are rich in vitamin C, carotene, calcium, magnesium, potassium, dietary fiber and glucosinolates [1]. Numerous epidemiological studies have demonstrated that the breakdown products of glucosinolates, such as isothiocyanate (ITCs), sulforaphane (SF) and indole-3-carbinol (I3C), are not only helpful to reduce the risk of cardiovascular disease, but also contribute to decreasing the incidence rate of carcinogenesis including breast, prostate, lung and pancreatic cancer [2][3][4]. Furthermore, GLSs and its derivatives have an important role against the invasion of pathogens and herbivores [5,6], and are also responsible for the specific flavor of Brassica plants [7].
So far, more than 200 kinds of natural glucosinolates have been found in 16 families of dicotyledons, and more than 120 different glucosinolates have been confirmed in glucosinolate, and lay the foundation for further improvements in beneficial glucosinolate components in Chinese kale by genetic engineering methods.

Plant Materials
The sixteen cultivated Chinese kale varieties used in this study are shown in Table 1. The seeds of each variety were sterilized for 10 min in 20% bleach (KAO, Japan), and washed with distilled water 5-6 times. Then, seeds were sown in a pot on 23 October 2019 and transplanted to the experimental base of the College of Horticulture in Hainan University on 20 November 2019. The experiment was arranged in random blocks and repeated 3 times. The area of each plot was 6 square meters and the row spacing was 15 × 25 cm. The cultivation and management methods are the same for all varieties.

GLSs Extraction and Purification
When the flower stalks and basal leaves of kale were roughly at the same height, and the flower balls were harvested but did not open, the edible stem and leaf parts of Chinese kale were separated quickly and cut into 1 cm sections, quickly mixed and stored in −80 °C . After that, the samples were placed in a freeze dryer for 24 h to remove moisture from the tissue. The freeze-dried samples were ground into powder by a powder grinder and put into a plastic sealing bag and blue desiccant bag to prevent the samples from being affected with moisture. Then, the samples were stored at −20 °C for future use.
Desulfo-glucosinolates (DS-GLSs) were extracted as previously described [39,40]. About 0.200 g powder was put into a 10 mL centrifuge tube, and 8 mL 70% methanol solution preheated at 80 °C was added into the centrifuge tube. After mixing, the tube was put into a water bath at 80 °C for 10 min, and mixed with intermittent oscillation. After 10 min, 1 mL barium acetate (0.5 mol/L) was added to remove the soluble protein. After centrifugation at 5000 rpm for 10 min, the supernatant was removed and stored at −20 °C for later use. For internal standard treatment, 100 μL of 2-propenyl glucosinolate (5 mmol/L, Sinigrin) were added to 0.200 g of dry sample powder. After mixing, 8 mL 70% methanol

Plant Materials
The sixteen cultivated Chinese kale varieties used in this study are shown in Table 1. The seeds of each variety were sterilized for 10 min in 20% bleach (KAO, Japan), and washed with distilled water 5-6 times. Then, seeds were sown in a pot on 23 October 2019 and transplanted to the experimental base of the College of Horticulture in Hainan University on 20 November 2019. The experiment was arranged in random blocks and repeated 3 times. The area of each plot was 6 square meters and the row spacing was 15 × 25 cm. The cultivation and management methods are the same for all varieties.

GLSs Extraction and Purification
When the flower stalks and basal leaves of kale were roughly at the same height, and the flower balls were harvested but did not open, the edible stem and leaf parts of Chinese kale were separated quickly and cut into 1 cm sections, quickly mixed and stored in −80 • C. After that, the samples were placed in a freeze dryer for 24 h to remove moisture from the tissue. The freeze-dried samples were ground into powder by a powder grinder and put into a plastic sealing bag and blue desiccant bag to prevent the samples from being affected with moisture. Then, the samples were stored at −20 • C for future use.
Desulfo-glucosinolates (DS-GLSs) were extracted as previously described [39,40]. About 0.200 g powder was put into a 10 mL centrifuge tube, and 8 mL 70% methanol solution preheated at 80 • C was added into the centrifuge tube. After mixing, the tube was put into a water bath at 80 • C for 10 min, and mixed with intermittent oscillation. After 10 min, 1 mL barium acetate (0.5 mol/L) was added to remove the soluble protein.
After centrifugation at 5000 rpm for 10 min, the supernatant was removed and stored at −20 • C for later use. For internal standard treatment, 100 µL of 2-propenyl glucosinolate (5 mmol/L, Sinigrin) were added to 0.200 g of dry sample powder. After mixing, 8 mL 70% methanol solution was preheated at 80 • C immediately, and then extracted and further prepared according to the above method.
Next, 4 mL of the extract was passed through the column filled with 1 cm DEAE Sephadex A25 (Sigma, St. Louis, MO, USA) four times. After all the filtrate was drained, 1 × 4 mL of sodium acetate buffer (0.1 mol/L, pH5) was added to clean the column. Finally, 200 µL of sulfate esterase (Sigma) was added to the column. After reaction at 37 • C for 18 h, the eluent was eluted with 1 × 4 mL ultra-pure water. The eluent was filtered by 0.22 µm PTFE needle filter and stored at −20 • C for detection and analysis.

HPLC-MS Analysis of GLSs
Determination: GLS separation was performed using Agilent Poroshell 120 EC-C18 reversed phase column (inner diameter 2.1 × 100 mm, particle diameter = 2.7 µm). Acidified water (0.1% formic acid, v/v) and acetonitrile (0.1% formic acid, v/v) were used as mobile phases A and B, respectively. The gradient program was carried out as follows: 0-3 min 97% A and 3% B; 10 min 80% A and 20% B; 13 min 20% A and 80% B; 13.5-14 min 50% A and 50% B; 15-16 min 97% A and 3% B. The flow rate was set at 0.30 mL/min throughout the gradient. The injection volume was 5 µL. The program was repeated three times for each sample, and the column temperature was maintained at 30 • C. The content of the glucosinolate components was calculated according to the international standardized response factor.
Mass spectrometry conditions: The Orbitrap high-resolution mass spectrometry was equipped with a suitable HESI source that was utilized with the following optimized parameter set: scan mode-full MS/AIF scan in positive ion mode; collision energy-35 eV; spray voltage-3.7 kV; capillary temperature-320 • C; auxiliary gas heater temperature-320 • C; sheath gas flow rate-40 arb; auxiliary gas flow rate-15 arb; sweep gas flow rate-5 arb; the auxiliary gas, sheath gas and sweep gas were all high-purity nitrogen.
Quantitative and qualitative of glucosinolates: Separation and identification of different desulfo-glucosinolates of Chinese kale were performed by using HPLC according to their retention times and confirmed by LC-ESI-MS analysis. Identification of GLSs was carried out at the positive ion mode. The DS-GLSs were confirmed according to their [M − G + H] + , [M + H] + , [M + Na] + and [M + K] + quasi-molecular ions and characteristic absorption peak [41]. The glucosinolate content was determined on the basis of retention time and peak area using sinigrin as an internal standard. The glucosinolate content was calculated using internal standard and response factors [42].

Expression Analysis of Genes Involving in the Glucosinolate Biosynthesis
According to instructions of the plant RNA extraction kit of Tiangen Biotech (Beijing, China) Co., LTD, the total RNA was extracted from the stems and leaves (the same growth material as the measurement of kale glucosinolate index) of the kale cultivars QB1 and ZH with high total glucosinolate content, and the low-glucosinolate content cultivars XLB and DSHH. Referring to cDNA sequences of Brassica MAMA1, SOT16 and other genes published in the NCBI database, the primers designed by Primer 5 for the expression detection of key genes in the glucosinolate biosynthesis pathway of Chinese kale are shown in Table 2, and the reference gene is Actin1 (XM_009117825.3). The expression levels of genes were detected by the real-time fluorescence quantitative method.

Statistical Analysis
The SPSS 23.0 software package was used to process the data, and analysis of variance and Duncan multiple range tests were carried out at a 95% confidence level.

Confirmation of Chinese Kale Desulfo-Glucosinolates
As shown in Table 3

The Content and Component Analysis of GLSs in Edible Stems and Leaves of Sixteen Chinese Kale Varieties
As shown in Figure 3a, the contents of aliphatic glucosinolate (A-GLSs) in the stems of Chinese kale varieties ranged from 238.27 to 1079.65 μmol/100 g DW (the average was 490.03 μmol/100 g DW). The contents of A-GLSs in the stems of Japanese kale varieties QB1 and Chinese kale varieties ZH were 1079.65 μmol/100 g DW and 780.29 μmol/100 g DW, respectively, which were significantly higher than those of other varieties, and there was a very significant difference between them. The A-GLSs content in leaves of different Chinese kale varieties ranged from 47.12 to 1179.2 μmol/100 g DW (the average content was 427.77 μmol/100 g DW), and the A-GLSs content in leaves of Japanese variety QB1 was significantly higher than 985.56 μmol/100 g DW of QB2. The A-GLSs content of QB2 was significantly higher than that of XLB and ZH, which ranked third and fourth and were significantly higher than other varieties. The content of A-GLSs in stems and leaves of the Japanese variety XLB was the lowest. In addition, except for the content of A-GLSs in stems of varieties SB, XL, XLB, DB and SJT being significantly higher than that in leaves, there was little difference in the content of A-GLSs in stems and leaves of most varieties.
The accumulation of indole glucosinolate (I-GLSs) in the stems of various kale varieties ranged from 22.5 to 67.09 μmol/100 g DW (average 35.61 μmol/100 g DW); the content of I-GLSs in the stem of variety KDSJ was 67.09 μmol/100 g DW, significantly higher than other varieties (Figure 3b). In leaves, the accumulation of I-GLSs was 10.73 to 87.13 μmol/100 g DW (average 45.95 μmol/100 g DW), the content of I-GLSs in the leaves of variety KDSJ was as high as 87.13 μmol/100 g DW, significantly higher than that of other varieties, and the content of varieties DB and XLB was lower. Except CB, XB, DB and XLB, the accumulation of I-GLSs in the leaves was higher than that in the stems or had little difference. Overall, the content of total I-GLSs in the stems and leaves of variety KDSJ was the highest, and the accumulation of total I-GLSs in variety DB was the lowest (Figure 3b).
The content of total glucosinolates (T-GLSs) ranged from 399.3 to 2346.03 μmol/100 g DW (with an average of 1005.41 μmol/100 g DW), and the highest content was found in Japanese variety QB1. Japanese variety QB2 and Chinese variety ZH had T-GLS content of 1665.54 μmol/100 g DW and 1649.49 μmol/100 g DW, respectively, and there was no significant difference in contents. The content of QB1 T-GLSs was significantly higher than

The Content and Component Analysis of GLSs in Edible Stems and Leaves of Sixteen Chinese Kale Varieties
As shown in Figure 3a, the contents of aliphatic glucosinolate (A-GLSs) in the stems of Chinese kale varieties ranged from 238.27 to 1079.65 µmol/100 g DW (the average was 490.03 µmol/100 g DW). The contents of A-GLSs in the stems of Japanese kale varieties QB1 and Chinese kale varieties ZH were 1079.65 µmol/100 g DW and 780.29 µmol/100 g DW, respectively, which were significantly higher than those of other varieties, and there was a very significant difference between them. The A-GLSs content in leaves of different Chinese kale varieties ranged from 47.12 to 1179.2 µmol/100 g DW (the average content was 427.77 µmol/100 g DW), and the A-GLSs content in leaves of Japanese variety QB1 was significantly higher than 985.56 µmol/100 g DW of QB2. The A-GLSs content of QB2 was significantly higher than that of XLB and ZH, which ranked third and fourth and were significantly higher than other varieties. The content of A-GLSs in stems and leaves of the Japanese variety XLB was the lowest. In addition, except for the content of A-GLSs in stems of varieties SB, XL, XLB, DB and SJT being significantly higher than that in leaves, there was little difference in the content of A-GLSs in stems and leaves of most varieties.
The accumulation of indole glucosinolate (I-GLSs) in the stems of various kale varieties ranged from 22.5 to 67.09 µmol/100 g DW (average 35.61 µmol/100 g DW); the content of I-GLSs in the stem of variety KDSJ was 67.09 µmol/100 g DW, significantly higher than other varieties (Figure 3b). In leaves, the accumulation of I-GLSs was 10.73 to 87.13 µmol/100 g DW (average 45.95 µmol/100 g DW), the content of I-GLSs in the leaves of variety KDSJ was as high as 87.13 µmol/100 g DW, significantly higher than that of other varieties, and the content of varieties DB and XLB was lower. Except CB, XB, DB and XLB, the accumulation of I-GLSs in the leaves was higher than that in the stems or had little difference. Overall, the content of total I-GLSs in the stems and leaves of variety KDSJ was the highest, and the accumulation of total I-GLSs in variety DB was the lowest (Figure 3b).
The content of total glucosinolates (T-GLSs) ranged from 399.3 to 2346.03 µmol/100 g DW (with an average of 1005.41 µmol/100 g DW), and the highest content was found in Japanese variety QB1. Japanese variety QB2 and Chinese variety ZH had T-GLS content of 1665.54 µmol/100 g DW and 1649.49 µmol/100 g DW, respectively, and there was no significant difference in contents. The content of QB1 T-GLSs was significantly higher than that of ZH and QB2, and the latter was significantly higher than that of other varieties, while the content of XLB T-GLSs was the lowest. Further analysis showed that the content of A-GLSs in T-GLSs accounted for 80.69% to 96.28% (average 90.50%) and was the main component. The total glucosinolates content in stem ranged from 267.38 to 1104.32 µmol/100 g DW, and in leaf it ranged from 57.87 to 1241.71 µmol/100 g DW. The T-GLSs content in stem was higher than that in leaf except for varieties LB, QB1, QB2, YC, DSHH and ZH (Figure 3c).
that of ZH and QB2, and the latter was significantly higher than that of other varieties, while the content of XLB T-GLSs was the lowest. Further analysis showed that the content of A-GLSs in T-GLSs accounted for 80.69% to 96.28% (average 90.50%) and was the main component. The total glucosinolates content in stem ranged from 267.38 to 1104.32 μmol/100 g DW, and in leaf it ranged from 57.87 to 1241.71 μmol/100 g DW. The T-GLSs content in stem was higher than that in leaf except for varieties LB, QB1, QB2, YC, DSHH and ZH (Figure 3c).

Analysis of GLSs Components and Contents in Edible Leaves of Chinese Kale Varieties
Three aliphatic glucosinolates, GRA, SIN and GNA, and four indole glucosinolates, 4-HGBS, GBS, 4-MGBS and NGBS, appeared in the edible leaves of 16 Chinese kale varieties, and the contents of GER in the leaves of different varieties were different. GER was not detected in half of the leaves from Japanese kale, but existed in Chinese kale leaves ( Table 4). The contents of GRA and GNA in A-GLSs ranged from 1.63% to 82.82% (33.08% on average) and 15.76% to 94.28% (61.86% on average), respectively, representing the main components of A-GLSs in the edible leaves of various Chinese kale varieties. Except DSHH, the content of GRA exceeds 100 μmol/100 g DW in the leaves of Chinese kale varieties from China, and DSCH and SJT are 211.83 and 201.88 μmol/100 g DW, respectively, significantly higher than that of other tested Chinese varieties. The GRA content in edible leaves of Japanese varieties XB and QB2 was higher than 100 μmol/100 g DW, but lower than 70 μmol/100 g DW in other varieties. On the contrary, high content of GNA mostly

Analysis of GLSs Components and Contents in Edible Leaves of Chinese Kale Varieties
Three aliphatic glucosinolates, GRA, SIN and GNA, and four indole glucosinolates, 4-HGBS, GBS, 4-MGBS and NGBS, appeared in the edible leaves of 16 Chinese kale varieties, and the contents of GER in the leaves of different varieties were different. GER was not detected in half of the leaves from Japanese kale, but existed in Chinese kale leaves ( Table 4). The contents of GRA and GNA in A-GLSs ranged from 1.63% to 82.82% (33.08% on average) and 15.76% to 94.28% (61.86% on average), respectively, representing the main components of A-GLSs in the edible leaves of various Chinese kale varieties. Except DSHH, the content of GRA exceeds 100 µmol/100 g DW in the leaves of Chinese kale varieties from China, and DSCH and SJT are 211.83 and 201.88 µmol/100 g DW, respectively, significantly higher than that of other tested Chinese varieties. The GRA content in edible leaves of Japanese varieties XB and QB2 was higher than 100 µmol/100 g DW, but lower than 70 µmol/100 g DW in other varieties. On the contrary, high content of GNA mostly appeared in the leaves of Japanese varieties. For example, the GNA contents of QB1 and QB2 were 1077.68 and 809.65 µmol/100 g DW, respectively, which was significantly higher than that of other varieties. The GNA content in the leaves of the Chinese kale variety ZH was the highest, reaching 629.1 µmol/100 g DW. The content of SIN in Chinese kale edible leaves ranged from 1.59 to 39.6 µmol/100 g DW, with the highest content in DSHH varieties and the lowest in TZSJDS varieties. The content of GER in all detected varieties was less than 1 µmol/100 g DW. Among indole glucosinolate, the content of GBS in leaves of Chinese kale varieties ranged from 8.76 to 77.36 µmol/100 g DW, and the GBS content of KDSJ was significantly higher than that of other varieties, while the contents of SB, TWDX, DSCH and ZH were not significantly different. The content of 4-MGBS ranged from 1.47 to 5.57 µmol/100 g DW. The contents of other components 4-HGBS and NGBS were all low.

Analysis of GLSs Components and Contents in Edible Stems of Chinese Kale Varieties
Different from the edible leaves, eight kinds of glucosinolates were detected in the edible stems of sixteen Chinese kale varieties. Similar to the leaves, GRA and GNA were also the main components of A-GLSs in the stems, with GRA accounting for 3.20% to 21.49%% (average 17.67%) and GNA accounting for 31.42% to 90.08% (average 58.45%) ( Table 5). The content of GRA in the stems of all Chinese kale varieties ranged from 63.49 to 201.05 µmol/100 g DW. Among them, the Chinese variety SJT and the Japanese variety SB had the highest GRA content, and the difference between the two was not significant, but their GRA contents were significantly higher than that of other varieties. The GRA contents were the lowest in Japanese variety LB and QB1 and the GRA content of Chinese varieties was generally higher than that of Japanese varieties in the stems. The GNA contents in the stems of each varieties were significantly different, and the average accumulation was 324.33 µmol/100 g DW, in which the GNA contents of LB, QB1 and ZH varieties were higher than 500 µmol/100 g DW, significantly higher than other varieties. The content of SIN ranged from 1.28 to 43.18 µmol/100 g DW, and the content of Japanese varieties was mostly higher than that of Chinese varieties in the stems. The content of GER in stem ranged from 2.28 µmol/100 g DW to 27.47 µmol/100 g DW, and the average content was about 80 times higher than that in leaf, but GER accounted for a small proportion of A-GLSs in stems and leaves. The content of GBS in the stems of various varieties of Chinese kale ranged from 9.76 to 52.78 µmol/100 g DW, accounting for a relatively high proportion in I-GLSs, and the content ranges of 4-HGBS and 4-MGBS are 1.47-7.20 and 0.50-6.58 µmol/100 g DW, respectively. The 4-MGBS content of the ZH variety was significantly higher than that of other varieties. The content of NGBS in stems was higher than that in leaves, but the content was very low.

Expression Analysis of Key Genes in GLSs Biosynthesis Pathway
The QB1 and ZH varieties with higher total glucosinolate content and the low-content XLB and DSHH varieties were selected from Japanese and Chinese varieties, and the GLSs biosynthetic gene expressions were analyzed. As shown in Figure 4d-g, the expression levels of aliphatic glucosinolate biosynthesis genes MAM1, CYP83A1, SOT17 and SOT18 were basically consistent with the changes in glucosinolate content; that is, the expression levels of these genes in high glucosinolate varieties were significantly higher than those in low glucosinolate varieties. In the same stem and leaf of the same variety, the relative expression level of CYP83A1 was more than 10 times that of MAM1, while the relative expression of SOT17 and SOT18 was low.

Discussion
Chinese kale is a specialty vegetable of the cruciferous family. It is closely related to common vegetables such as cabbage, mustard and the model plant Arabidopsis thaliana. Chinese kale, Brussels sprout, broccoli, red cabbage and cauliflower all belong to the vegetables of Brassica oleraca, but the glucosinolate content and main component of different varieties and organs are different in Brassica oleracea. The content of glucosinolates in Brussels sprout and broccoli is higher than that in cauliflower and red cabbage. PRO, SIN and GIB are the main components of A-GLSs in Brussels sprout [43]. The representative glucosinolates in Brassica napus, Brassica juncea and Brassica rape are PRO, SIN and GNA, respectively [44]. Although the research indicated that the same glucosinolate component existed in Chinese kale varieties and parts and showed that aliphatic glucosinolate accounted for more than 70% of the total glucosinolates in all Chinese kale varieties, GNA, GRA and SIN were also main glucosinolate in most Chinese kale varieties. However, there are still great differences in the types, main components and contents of glucosinolates detected by different varieties and detection parts of Chinese kale (Table 6). In this study, eight kinds of glucosinolates were detected from the stems and leaves of sixteen Chinese kale varieties from Japan and China, including four kinds of A-GLSs-GRA, SIN, GNA and GER, and four kinds of I-GLSs-4-HGBS, GBS, 4-MGBS and NGBS, but no aromatic glucosinolates were detected. These results were slightly different from previous studies that recorded 7, 9-11 and 13 kinds of components (Table 6). This may be related to the genotype of the variety studied, the planting location, environmental conditions and the method of determination. There were significant differences in the expression of each gene in different parts of Chinese kale (Figure 4d-g). The expression level of MAM1 in the stems is greater than that of leaves in the varieties with high glucosinolate content, but the expression level of CYP83A1 and SOT18 in leaves was greater, and the expression level of SOT17 in the leaves of the QB1 variety is significantly greater than that in the stems. The expression levels of MAM1 and CYP83A1 were higher in stems than in leaves in the low glucosinolate varieties, except the expression levels of SOT17 were more significant in leaves.
Compared with DSHH and XLB varieties with low glucosinolate content, the expression levels of indole glucosinolate biosynthesis genes CYP83B1, UTG74B1 and SOT16 were higher in the stems and leaves of Chinese kale varieties with high glucosinolate content varieties QB1 and ZH. On the contrary, the expression level of CYP79A2 in the leaves was significantly higher than in the stems (Figure 4a,b,h). The expression levels of CYP79A2, CYP83B1, UTG74B1 and SOT16 genes were higher in leaves than in stems in the same varieties (Figure 4a-c,h).
In the high glucosinolate varieties, the expression level of CYP83B1 in the stems of the ZH variety was significantly higher than that of QB1, but there was no significant difference in the leaves (Figure 4b). The expressions of UTG74B1 and SOT16 were not significantly different in the stems, while the expressions were significantly higher in the leaves of ZH than QB1 (Figure 4a,c). The expression of CYP79A2 in QB1 was also higher than in ZH (Figure 4h). A comparison of two low glucosinolate varieties DSHH and XLB indicates that the expression levels of SOT16, CYP83B1 and UTG74B1 had no significant differences in the stems and leaves, and the expression of CYP79A2 had no significant difference in the stems, but was significantly higher in the leaves of XLB than DSHH (Figure 4a-c,h).

Discussion
Chinese kale is a specialty vegetable of the cruciferous family. It is closely related to common vegetables such as cabbage, mustard and the model plant Arabidopsis thaliana. Chinese kale, Brussels sprout, broccoli, red cabbage and cauliflower all belong to the vegetables of Brassica oleraca, but the glucosinolate content and main component of different varieties and organs are different in Brassica oleracea. The content of glucosinolates in Brussels sprout and broccoli is higher than that in cauliflower and red cabbage. PRO, SIN and GIB are the main components of A-GLSs in Brussels sprout [43]. The representative glucosinolates in Brassica napus, Brassica juncea and Brassica rape are PRO, SIN and GNA, respectively [44]. Although the research indicated that the same glucosinolate component existed in Chinese kale varieties and parts and showed that aliphatic glucosinolate accounted for more than 70% of the total glucosinolates in all Chinese kale varieties, GNA, GRA and SIN were also main glucosinolate in most Chinese kale varieties. However, there are still great differences in the types, main components and contents of glucosinolates detected by different varieties and detection parts of Chinese kale ( Table 6). In this study, eight kinds of glucosinolates were detected from the stems and leaves of sixteen Chinese kale varieties from Japan and China, including four kinds of A-GLSs-GRA, SIN, GNA and GER, and four kinds of I-GLSs-4-HGBS, GBS, 4-MGBS and NGBS, but no aromatic glucosinolates were detected. These results were slightly different from previous studies that recorded 7, 9-11 and 13 kinds of components (Table 6). This may be related to the genotype of the variety studied, the planting location, environmental conditions and the method of determination.
Like other Brassica oleracea, A-GLSs accounted for a large proportion of the T-GLSs in Chinese kale, while the content of I-GLSs was low and the components were basically the same. However, GNA and GRA were the main glucosinolate components in Chinese kale, and the contents of SIN and GER were relatively low [45][46][47]. In our results, the proportion of aliphatic glucosinolates in edible stems and leaves of Chinese kale varieties from China and Japan was higher than that of other groups, reaching more than 80.69%. Although the content and component of glucosinolates were different, they all showed high GRA and GNA contents, and the GRA content of Chinese kale stems and leaves was higher than that of Japanese varieties, while the GNA content of Japanese varieties was relatively higher. Although the content of SIN in the Chinese kale varieties is lower than that of GRA and GNA, the total content of stems and leaves in most varieties exceeds 30 µmol/100 g DW. In addition, PRO was not detected in this study, which is different from previous results. Because the GNA hydrolysate 3-butenyl ITC can induce the death of prostate cancer cells through apoptosis, and GRA and its derived isothiocyanate sulforaphane have anti-inflammatory, anticancer and neuroprotective effects [48][49][50][51][52], the Chinese kale varieties with high GRA and GNA content, such as ZH and QB1 involved in this study, have important nutritional and health functions. Obtaining varieties with high content of GRA and GNA is also the goal of breeding Chinese kale in the future.
A total of 105 genes are associated with the contents and compositions of glucosinolates in Brassica oleracea L. genome, including MAM CYP79, CYP83, ST, AOP, IGMT and other biosynthetic pathway genes, MYB28, MYB29, MYB51 and other transcription factors and 22 catabolic genes [53], but not all genes were expressed in different subspecies of Brassica oleracea [54]. The higher expression of ST5, CYP81F1 and CYP84F1 in subspecies of cabbage, broccoli, and kale in MeJA treatment resulted in the accumulation of higher levels of GBS, while the increase in CYP81F4 gene expression resulted in the production of higher NGBS content. The higher levels of expression of ST5, AOP and GLS-OH genes give cabbage a higher aliphatic glucosinolate content than broccoli and kale [55]. All these results indicate that gene expression of different subspecies and different parts of B. oleracea affect the glucosinolate biosynthesis. Table 6. The types, main components and content ranges of glucosinolates determined by different methods.

Types of GLSs Main Components of GLSs A-GLSs
The BCAT4, MAM1 and CYP79F1 are upstream genes of glucosinolate synthesis in Chinese kale, which is also one of the subspecies of B. oleracea. They are highly expressed in Chinese kale genotypes with high GRA content, which may be related to the biosynthesis of aliphatic glucosinolate GRA [37]. The decreased expression levels of CYP83B1, SUR1 and UTG74B1 under NaCl treatment led to the decreased content of A-and I-GLSs, while the increased expression levels of CYP83B1, SUR1 and UTG74B1 under MeJA treatment led to the increased content of glucosinolates [59]. Further research found that the significantly down-regulated expression levels of MAM1/2, IPMDH1, CYP79F1, CYP83A1, SUR1, UGT74B1, UGT74C1, GSTU20, GSTF9, SOT18, CYP81F3, CYP81F4 and other genes involved in glucosinolates biosynthesis result in the decrease in the contents of A-GLSs, I-GLSs, T-GLSs, GNA, GRA, SIN, 4-MGBS and NGBS in Chinese kale under the treatment of far-red light [34]. In the transgenic Chinese kale overexpressing BoaUGT74B1, the contents of indole and aliphatic GLSs were 1.30-7.11 and 2.13-4.44 times higher than those in the wild type, respectively, and the highest 4-MGBS content in the transgenic "Sijicutiao" was 12.14 times higher than in the control. The highest GBS in "Fuzhouhuanghua" overexpressed plants was 22.76 times that of the control [60]. In our experiments, genes MAM1, CYP83A1, SOT17 and SOT18 involved in the biosynthesis of A-GLSs and CYP83B1, UTG74B1 and SOT16 involved in the synthesis of I-GLSs were significantly more expressed in stems and leaves of QB1 and ZH varieties with high glucosinolate content, and in stems and leaves of varieties with high GNA content. The expression levels of MAM1 and CYP83A1 in the stems and leaves of the Chinese kale variety ZH were significantly higher than those of QB1 varieties, which was consistent with the level of GRA content in the stems and leaves. The above results show that the content and components of glucosinolates in Chinese kale are affected by the expression of related genes, as in other subspecies of B. oleracea.
At the same time, we found that there was a certain relationship between the flow of glucosinolate components in the leaves and stems of Chinese kale, and that it was also related to the expression of genes. The expression levels of SOT17 and SOT18 in the A-GLSs synthesis pathway were significantly higher in QB1 than ZH varieties, and the GNA content was correspondingly higher in QB1 varieties, while the content of GRA was higher in ZH varieties and the MAM1 expression was higher in ZH varieties. The expression level of UGT74B1 in leaves of Chinese kale varieties with high GBS and 4-HGBS content was significantly higher than that of low-content varieties, while the content of GBS and 4-HGBS in stems was higher in low-content varieties. The expression levels of SOT16 and CYP83B1 genes were similar to the contents of GBS and 4-HGBS in Chinese kale leaves, but opposite to the contents of GBS and 4-HGBS in Chinese kale stems. CYP79A2 regulates the synthesis of aromatic glucosinolate, which is mainly expressed in Chinese kale leaves, but the expression level is very low. No targeted aromatic glucosinolate was detected in corresponding stems and leaves, which may be related to the competition of A-, I-and R-GLSs metabolic flow. Therefore, the influence of these genes on the composition and content of Chinese kale glucosinolate remains to be further verified.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of NAME.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.