Leptosphaeria maculans Alters Glucosinolate Profiles in Blackleg Disease–Resistant and -Susceptible Cabbage Lines

Blackleg, a fungal disease caused by Leptosphaeria maculans, is one of the most devastating diseases of Brassica crops worldwide. Despite notable progress elucidating the roles of glucosinolates in pathogen defense, the complex interaction between B. oleracea (cabbage) and L. maculans infection that leads to the selective induction of genes involved in glucosinolate production and subsequent modulation of glucosinolate profiles remains to be fully understood. The current study was designed to identify glucosinolate-biosynthesis genes induced by L. maculans and any associated alterations in glucosinolate profiles to explore their roles in blackleg resistance in 3-month-old cabbage plants. The defense responses of four cabbage lines, two resistant and two susceptible, were investigated using two L. maculans isolates, 03–02 s and 00–100 s. A simultaneous increase in the aliphatic glucosinolates glucoiberverin (GIV) and glucoerucin (GER) and the indolic glucosinolates glucobrassicin (GBS) and neoglucobrassicin (NGBS) was associated with complete resistance. An increase in either aliphatic (GIV) or indolic (GBS and MGBS) glucosinolates was associated with moderate resistance. Indolic glucobrassicin (GBS) and neoglucobrassicin (NGBS) were increased in both resistant and susceptible interactions. Pearson correlation showed positive association between GER content with GSL-OH (Bol033373) expression. Expressions of MYB34 (Bol007760), ST5a (Bol026200), and CYP81F2 (Bol026044) were positively correlated with the contents of both GBS and MGBS. Our results confirm that L. maculans infection induces glucosinolate-biosynthesis genes in cabbage, with concomitant changes in individual glucosinolate contents. In resistant lines, both aliphatic and indolic glucosinolates are associated with resistance, with aliphatic GIV and GER and indolic MGBS glucosinolates particularly important. The association between the genes, the corresponding glucosinolates, and plant resistance broaden our molecular understanding of glucosinolate mediated defense against L. maculans in cabbage.

Blackleg, a fungal disease caused by Leptosphaeria maculans, is one of the most devastating diseases of Brassica crops worldwide. Despite notable progress elucidating the roles of glucosinolates in pathogen defense, the complex interaction between B. oleracea (cabbage) and L. maculans infection that leads to the selective induction of genes involved in glucosinolate production and subsequent modulation of glucosinolate profiles remains to be fully understood. The current study was designed to identify glucosinolate-biosynthesis genes induced by L. maculans and any associated alterations in glucosinolate profiles to explore their roles in blackleg resistance in 3-month-old cabbage plants. The defense responses of four cabbage lines, two resistant and two susceptible, were investigated using two L. maculans isolates, 03-02 s and 00-100 s. A simultaneous increase in the aliphatic glucosinolates glucoiberverin (GIV) and glucoerucin (GER) and the indolic glucosinolates glucobrassicin (GBS) and neoglucobrassicin (NGBS) was associated with complete resistance. An increase in either aliphatic (GIV) or indolic (GBS and MGBS) glucosinolates was associated with moderate resistance. Indolic glucobrassicin (GBS) and neoglucobrassicin (NGBS) were increased in both resistant and susceptible interactions. Pearson correlation showed positive association between GER content with GSL-OH (Bol033373) expression. Expressions of MYB34 (Bol007760), ST5a (Bol026200), and CYP81F2 (Bol026044) were positively correlated with the contents of both GBS and MGBS. Our results confirm that L. maculans infection induces glucosinolate-biosynthesis genes in cabbage, with concomitant changes in individual glucosinolate contents. In resistant lines, both aliphatic and indolic glucosinolates are associated with resistance, with aliphatic GIV and GER and indolic MGBS glucosinolates particularly important. The association between the genes, the corresponding glucosinolates, and plant resistance broaden our molecular understanding of glucosinolate mediated defense against L. maculans in cabbage.
Keywords: blackleg disease, Leptosphaeria maculans, glucosinolates, resistance, cabbage, expression analysis INTRODUCTION Blackleg is one of the most devastating diseases of Brassica crop species, causing an estimated $900 million of crop losses throughout the world every year (Howlett, 2004;Fitt et al., 2006Fitt et al., , 2008. Blackleg is caused by the hemi-biotrophic fungal pathogen Leptosphaeria maculans. Severe blackleg infection can result in complete loss of Brassica napus canola or oilseed rape crops (Li et al., 2003;Rouxel et al., 2003;Sprague et al., 2006). Blackleg is similarly devastating for vegetable varieties of B. oleracea, including cabbage (Humpherson-Jones, 1985;Rico et al., 2001;Dilmaghani et al., 2010Dilmaghani et al., , 2013Piliponyte-Dzikiene et al., 2015); in fact, the first report of a blackleg epidemic in cabbage (in Wisconsin, USA) is almost a century old (Henderson, 1918). Despite the availability of various chemical control measures (Del Rio and Ruud, 2013;Fraser et al., 2016;Koh et al., 2016), development of effective resistance through breeding remains the most widely accepted means of protecting Brassica germplasm.
The relationship between glucosinolate levels and the resistance of Brassicaceae family members to diverse fungal pathogens remains unclear. Upon L. maculans infection, overall levels of glucosinolates are not strongly correlated with pathogen resistance in different Brassica species (Mithen and Magrath, 1992;Sexton et al., 1999). In fact, a negative correlation between Alternaria infection and glucosinolate levels was reported in Brassica napus (Doughty et al., 1991;Giamoustaris and Mithen, 1997). By contrast, upon Sclerotinia sclerotiorum infection, the extent of pathogen-induced accumulation of indolic glucosinolates in Brassica napus is positively correlated with plant resistance (Li et al., 1999). These seeming contradictions could reflect the lifestyles of the individual fungal pathogens, e.g., necrotrophic versus biotrophic (Sanchez-Vallet et al., 2010), their host ranges, e.g., Brassicaceae-specialist versus broad-spectrum (Buxdorf et al., 2013), the genetic purity of the host plants, e.g., isogenic versus heterozygous lines, in addition to the range and quantity of glucosinolates and their break-down products produced by the host plants.
The association between total glucosinolate content and plant resistance to L. maculans infection has been studied since the 1990's. Nevertheless, no obvious relationship between glucosinolate profiles and L. maculans resistance has been established, possibly due to the fact that these studies focused on total glucosinolate levels rather than individual glucosinolate compounds (Mithen and Magrath, 1992;Giamoustaris and Mithen, 1997;Sexton et al., 1999;Kliebenstein et al., 2002). Previous studies concluded that the glucosinolate-myrosinase system is not a major determinant of blackleg resistance (Wretblad and Dixelius, 2000;Andreasson et al., 2001). This conclusion was supported by an earlier report in resynthesized B. napus F 2 progeny, in which disease resistance and glucosinolate profiles did not co-segregate (Mithen and Magrath, 1992). By contrast, infection with the hemibiotrophic fungus L. maculans was shown to trigger the accumulation of both aliphatic (gluconapin, i.e., 3-butenyl glucosinolate; progoitrin, i.e., 4hydroxybutyl glucosinolate; glucobrassicanapin, i.e., 4-pentenyl glucosinolate; and gluconapoleiferin, i.e., 2-hydroxy-4-pentenyl glucosinolate) and indolic (glucobrassicin, i.e., 3-indolylmethyl glucosinolate and 4-hydroxy-glucobrassicin) glucosinolates in Brassica rapa (Abdel-Farid et al., 2010). A few recent reports have shown that resistance to obligate biotrophs, hemibiotrophs, and necrotrophs might be associated with the production of indolic glucosinolates in Brassicaceae (Bednarek et al., 2009;Hiruma et al., 2013). In Arabidopsis, accumulation of indolic 4-methoxy-glucobrassicin in response to fungal attack was regulated by the expression of CYP81F2, which is activated by the myrosinase PEN2 (Bednarek et al., 2009). These contrasting reports on the association of glucosinolate profiles and resistance in Brassica species call for further investigation of plant-pathogen interactions at the molecular and biochemical levels.
Our previous studies have shown that the upregulation of glucosinolate biosynthesis genes is associated with the upregulation of individual glucosinolate compounds in B. oleracea inbred lines Yi et al., 2016). Here, we investigate changes in expression of genes associated with glucosinolate biosynthesis upon infection with L. maculans and correlate these changes with alterations in glucosinolate profiles. These experiments were performed in both resistant and susceptible genotypes against two different L. maculans isolates.

Plant Materials and Growth Conditions
Seeds of four cabbage inbred lines (B. oleracea var. capitata) were obtained from Asia Seeds Ltd. (Seoul, South Korea). Two of the lines were reported to be resistant to blackleg disease at the seedling stage and the other two were susceptible (Figure 1; Robin et al., 2017a). Seeds were sown in a gardensoil mixture composed of peat moss, coco peat, perlite, zeolite, and vermiculite in 32-celled trays in a plant culture room with a temperature of 20 ± 2 • C, 16-h day-length, and a light intensity of ca. 400 µmol m −2 s −1 at bench level (florescent light bulbs; Yi et al., 2015). Seedlings were grown for four weeks before transfer to a glasshouse, where they were grown for another two months before infection with L. maculans. FIGURE 1 | Cabbage lines (A) and scoring criteria (B) against blackleg disease. Three-month-old cabbage plants were inoculated to test their response to two Leptosphaeria maculans isolates 03-02 s and 00-100 s. Middle-aged leaves with yellow squares were infected. Infected leaves were scored two weeks after inoculation on a scale from 1−9 based on visual scoring criteria, where a score of "1" represents the least infected and "9" represents the most infected cabbage plants. In our previous assessment of seedling resistance, the lines BN4059 and BN4072 were susceptible while BN4098 and BN4303 were resistant (Robin et al., 2017a).

L. maculans Isolates and Inoculation
Details of the two fungal isolates, culture methods, and spore solution preparation have been previously published (Robin et al., 2017a). In brief, two isolates of L. maculans, 03-02s (AvrLm1-4-6-7-11-J1-S, AvrLep1-2-3) and 00-100s (AvrLm2-3-6-9-J1-S, AvrLep1-2), were provided by Agriculture and Agri-Foods (AAFC), Saskatoon, Canada. Each isolate was sub-cultured on 20% V8 agar plates and grown at 22 • C with a 16-h photoperiod under fluorescent light to produce a fungal inoculum. After 10 days, each plate was flooded with 10 mL of sterile distilled water and scraped with a sterile microscope slide to prepare a spore suspension. The spore suspension was filtered using sterile Miracloth (EMD Millipore Corporation, USA) to remove the mycelia and other debris. The clear suspension was centrifuged to concentrate the spores. The spore suspensions of 03-02 s and 00-100 s isolates were then diluted to 2.25 × 10 7 spores mL −1 and used to inoculate small wounds created on middleaged leaves of 3-month-old cabbage plants ( Figure 1A). The experiment was laid out using a randomized block design with six plants per isolate. Approximately four wounds were created per cm 2 leaf area. A set of six plants were kept as controls and another six plants were wounded but not infected (mock-treated). Diseased areas on the leaf surface were scored two weeks after inoculation on a scale of 1-9, based on the severity of the infection (Figure 1B), and categorized as resistant (R, score range 1.0-3.0), moderately resistant (MR, 3.1-6.0), susceptible (S, 6.1-7.0) and highly susceptible (HS, 7.1-9.0).

Collection and Preparation of Leaf Samples for HPLC and Expression Analysis
Leaf samples from three biological replicates, randomly chosen from six inoculated plants, were collected from control, mocktreated, and L. maculans inoculated plants four days after inoculation to estimate endogenous glucosinolate content and quantify the expression of selected genes from the glucosinolate biosynthesis pathway ( Figure S1). Leaf samples collected for HPLC analysis and real-time PCR were immediately stored at −80 • C after flash-freezing in liquid nitrogen.

Estimation of Glucosinolate Content
Leaf samples from three biological replicates from each of the control, mock, and L. maculans infected plants were used for extracting desulfo-glucosinolates through a modified HPLC protocol as previously described (Yi et al., 2015Robin et al., 2016). Frozen leaf tissue stored at −80 • C was treated with methanol and then ground to a fine powder. The powdered leaf samples were preserved at 70 • C for 10 min and then kept at room temperature for about an hour. The powdered samples were then centrifuged at 10,000 × g at 4 • C for 8 min. This centrifugation step removes structural components and protein molecules as sediments. The supernatant was then passed over an anion-exchange column. Centrifugation and anionexchange chromatography was repeated two more times. The supernatant collected from the final step of anion-exchange chromatography was considered the crude glucosinolate sample. The crude glucosinolates were then subjected to desulfation. Here, 0.5 mL 50 mM barium acetate and 0.5 mL 50 mM lead acetate were mixed with the crude glucosinolates followed by centrifugation at 2,000 × g for 10 min. The supernatant was then passed through a pre-equilibrated (with 0.5 M sodium acetate) DEAE-Sephadex column. Desulfation was initiated by the addition of 250 µL aryl sulfatase to the column and was allowed to run over-night for 16 h. The desulfated glucosinolates were then eluted with 1 mL distilled water. The eluted desulfoglucosinolates were purified by high-speed centrifugation at 20,000 × g for 4 min at 4 • C followed by filtering through a PTFE filter (13 mm, 0.2 µm, Advantec, Pleasanton, CA, USA). The purified glucosinolates were then subjected to HPLC analysis in a Waters 2695 HPLC system (Waters, Milford, MA, USA) equipped with a C 18 column (Zorbax Eclipse XBD C 18 , 4.6 mm × 150 mm, Agilent Technologies, Palo Alto, CA, USA). Both water and acetonitrile were used as mobile phase solvents. Individual glucosinolate compounds were detected using a PDA 996 UV-visible detector (Waters) at a wavelength of 229 nm. For quantification of the detected glucosinolates, a standard curve prepared from commercial sinigrin (SIN) was used. Mass spectrometry analysis (HPLC/MS, Agilent 1,200 series, Agilent Technologies) facilitated the identification of individual glucosinolates . Chemical names, common names, abbreviations, and chemical structures of the glucosinolates identified in this study are given in Figure S2.

Primer Design for Expression Analysis of Glucosinolate Biosynthesis Genes
Thirty-eight genes involved in glucosinolate biosynthesis were selected for transcription analysis to determine how transcript levels are affected by pathogen inoculation (Table S1, Figure  S1). Eleven of the genes encode transcription factors: five in the aliphatic biosynthesis pathway and six in the indolic pathway. Of the other 27 genes, 10 are aliphatic biosynthesis genes and 17 are indolic biosynthesis genes Yi et al., 2016). Primers were designed using primer3plus software (http:// primer3plus.com/) and primer efficiency was tested according to Robin et al. (2016).

cDNA Synthesis and Real-Time Quantitative PCR Analysis
Total RNA was extracted from the collected leaf samples using RNeasy mini kit, Catalog No. 74106, Qiagen, Valencia, CA, USA. cDNA synthesis was performed from total RNA using a PrimeScript-based kit (Takara Bio, Inc., Shiga, Japan). To conduct quantitative RT-PCR (qPCR), iTaqTM SYBR R Green Super-mix was used with ROX (Bio-Rad, Hercules, CA, USA). For each reaction, a total reaction volume of 20 µL was prepared containing 10 µL PCR master mix, 7 µL ultra-pure water, 2 µL forward and reverse primers and 1 µL cDNA template with a concentration of 60 ng µL −1 . PCR conditions were as follows: denaturation at 95 • C for 10 min, 40 cycles of amplification with denaturation at 95 • C for 20 s, annealing at 58 • C for 20 s, and amplification and signal acquisition at 72 • C for 30 s. Data were recorded as fluorescence at the end of each of 40 cycles for each sample. Each biological replicate was tested in three technical replicates. Quantification (Cq) analysis was done using LightCycler96 software (Roche, Mannheim, Germany). Livak's comparative 2 − Ct method was used to calculate the relative expression of each sample (Livak and Schmittgen, 2001). Three different actin genes selected from the NCBI database, GenBank Accession Nos. AF044573 (Zhang et al., 2012), JQ435879 (Nawaz et al., 2014), and XM_013753106 (Lee et al., 2015), were expressed in all inbred lines and were used as a reference.

Statistical Analysis
A one-way analysis of variance was conducted to test the statistical significance of different treatments on the four cabbage inbred lines using Minitab 18 statistical software (Minitab Inc., State College, PA, USA). A posthoc Tukey's pairwise comparison was conducted to visualize statistical significance of 16 treatment × genotype combinations. Test statistic, degrees of freedom, and p-values of statistical significance for glucosinolate contents and relative expression of biosynthesis genes are given in Table S2.
Separate heat maps were drawn to show association between blackleg scores and glucosinolate contents or expression of genes using conditional formatting option in Microsoft Excel. Pearson correlation coefficient was estimated between glucosinolate content and expression of biosynthesis genes in 16 treatment × genotype combinations (Supplementary Appendix 1).

Resistance of Adult Cabbage Lines to L. maculans Infection
Inoculation of cabbage leaves with two L. maculans isolates resulted in different responses in different lines, based on visual scoring of blackleg disease symptoms ( Table 1). Of the eight total combinations of four cabbage genotypes and two isolates (as shown in Table 1) only one combination, BN4303 × 00-100s, exhibited complete resistance, with the lowest visual score of 2 (range 1-3). The combinations of BN4303 × 03-02 s (score range 3-5) and BN4098 × 00-100s (score range 4-6) showed moderate resistance. BN4098, which exhibited moderate resistance to isolate 00-100 s, was susceptible to isolate 03-02 s (score range 6-8). Lines BN4072 and BN4059 were susceptible to both isolates, with the latter being highly susceptible to both isolates (Table 1). Overall, line BN4303 was the most resistant, as it showed complete resistance to isolate 00-100 s and moderate resistance to 03-02 s, whereas BN4098 showed moderate resistance only to isolate 00-100 s and the other two lines were susceptible.

Constitutive Glucosinolate Contents Vary in Cabbage Inbred Lines
In untreated control plants, levels of both aliphatic and indolic glucosinolates varied significantly between the four cabbage lines (Figure 2, Supplementary excel data file). Among the control plants, BN4059, the most susceptible cabbage line, had the 1 | Resistance scoring of four cabbage lines at three months of age (adult stage) against two L. maculans isolates "03-02 s" and "00-100 s" 15-days post-inoculation.

Line
Seedling resistance status against both isolates highest total levels of glucosinolates while BN4303, the most resistant line, had the lowest total levels ( Figure S3). PRO was found only in BN4059, which also had the highest levels of aliphatic glucoiberin (GIB), sinigrin (SIN), and glucoiberverin (GIV), but no glucoerucin (GER). BN4059 also had the highest levels of indolic GBS and MGBS, but no NGBS. Notably, BN4303 had the lowest levels of SIN between the four lines. Hydroxy-glucobrassicin (HGBS), another indolic glucosinolate, was detected only in untreated control plants of the BN4098 line ( Figure S3).

L. maculans Inoculation Alters the Glucosinolate Profile of Adult Cabbage Plants
Inoculation of 3-month-old cabbage plants with either of the two L. maculans fungal isolates remarkably changed the glucosinolate profiles in leaves of the fully and moderately resistant cabbage lines. The content of aliphatic GIV increased significantly in BN4303 in response to both fungal isolates while it decreased or was not produced in the susceptible genotypes BN4059 and BN4072 (Figure 2). In the resistant line BN4303, levels of GIV increased by 2.77-and 4.14-fold four days after inoculation with 03-02 s and 00-100 s, respectively compared to mock-treated plants (Figure 2). By contrast, in the most susceptible line BN4059, GIV levels decreased 3.72-and 6.28fold following fungal inoculation with 03-02s and 00-100s, respectively, compared to mock-treated plants (Figure 2). Infection with 00-100 s induced the accumulation of glucoerucin (GER) in the resistant line BN4303; no GER was detected in control plants (Figure 2). The most susceptible line, BN4059, did not produce any glucoerucin (GER). Other notable changes in aliphatic glucosinolates were (i) a 3-to 8-fold decrease in progoitrin (PRO) levels in mock-treated and fungal-inoculated plants in the susceptible cabbage line BN4059 compared to the untreated plants, indicating that PRO levels are sensitive to both leaf injury and fungal attack; (ii) accumulation of gluconapin (GNA) in BN4072 in response to 03-02 s infection and (iii) an increase in GER in mock-treated BN4303 plants compared to the untreated plants, indicating that GER levels are also sensitive to both leaf injury and fungal attack (as GER levels further increased in the same line upon 00-100 inoculation).
Overall, our results show that the complete resistance of the line BN4303 against 00-100s isolate was associated with simultaneously increased contents of both aliphatic (GIV and GER) and indolic (GBS and NGBS) glucosinolates upon inoculation, while the enhanced accumulation of an aliphatic glucosinolate (GIV) was associated with moderate resistance in this line against 03-02 s isolate ( Table 2). By contrast, the moderate resistance of the line BN4098 against the 00-100 s isolate was associated with enhanced accumulation of only indolic glucosinolates (GBS, NGBS and MGBS).

Expression Changes in Transcription Factor and Aliphatic Glucosinolate Biosynthesis Genes in Cabbage Lines With Blackleg Disease
The expression levels of majority of the transcription factors under untreated control conditions were comparatively higher in BN4098 compared to the other three lines ( Figure S4, Supplementary excel data file). Like the transcription factor genes, the expression levels of most of the aliphatic structural biosynthesis genes were comparatively higher in BN4098 than in the other three cabbage lines under untreated control conditions (Figure 3, Figure S5). However, FMOGS-OX2 (Bol010983) and FMOGS-OX5 (Bol029100) showed higher expression in BN4059 control plants compared to the other FIGURE 2 | Glucosinolate contents of leaf samples from the four cabbage lines (BN4059, BN4072, BN4098, and BN4303) under four different treatments (control, mock, inoculation with 03-02 s or 00-100 s) four days after inoculation. Detection limit of glucosinolates was 0.025 µmol g −1 DW. The mean of three biological replicates is presented. Vertical bars indicate standard error of mean. Different letters indicate statistically significant differences between genotype × treatment combinations. Upward-pointing blue arrows indicate increased glucosinolate content in response to L. maculans infection.
three lines, consistent with the highest content of total, and most individual, aliphatic glucosinolates in this genotype ( Figure S5).
None of the genes for aliphatic transcription factors were upregulated in any resistance combination; rather, expression of these five genes generally decreased in this genotype upon infection ( Figure S4). Expression of ST5b and GSL-OH, which are involved in the biosynthesis of aliphatic glucosinolates, increased in both of the resistant lines, BN4303 and BN4098, during blackleg infection (Figure 3). In the resistant line BN4303, expression of ST5b genes Bol026201 and Bol026202 increased, respectively, 4.31-and 2.73-fold after 00-100 s infection and 3.23-and 1.70-fold after 03-02 s infection compared to mocktreated plants (Figure 3). Expression of GSL-OH increased 3.33fold in BN4303 after 00-100s infection compared to mocktreated plants. In susceptible combinations, BN4072 × 03-02s and BN4098 × 03-02s, AOP2 (Bo2g102190) showed 24.4and 3.39-fold higher expression, respectively, compared to mock treated plants ( Figure S5). AOP2 also showed a 2.7-fold increase in BN4303 × 03-02 s compared to mock treated plants ( Figure S5).

Expression Changes in Transcription Factors Involved in the Biosynthesis of Indolic Glucosinolates in Cabbage Lines with Blackleg Disease
The expression levels of five B. oleracea genes encoding three transcription factors were measured under control, mock, and infection conditions. In untreated control plants, five out of six transcription factors involved in indolic glucosinolate biosynthesis, the exception being MYB34 (Bol036262), were highly expressed in BN4098 compared to the other three lines (Figure 4). The genes Bol007760 and Bol017062, encoding the transcription factor MYB34, showed increased expression in the resistant line BN4303 with blackleg disease. Bol007760 and Bol017062 showed a 9.35-and 2.13-fold increase against 00-100s isolates, respectively, compared to mock-treated plants GIV, Glucoiberverin, GER, glucoerucin, GBS, glucobrassicin, NGBS, neoglucobrassicin and MGBS, methoxy-glucobrassicin.

Association Between Pathogen-Induced Upregulation of Pathway Genes and Accumulation of Corresponding Glucosinolates and the Plant's Resistance Status
A heat map emphasized that the fold changes under pathogen inoculation compared to mock-treated samples in expression of genes involved in the biosynthesis of glucosinolates were often consistent with the blackleg resistance in the cabbage lines. For example, in the resistant line BN4303 under 00-100 s infection when ST5b (Bol026201 and Bol026202) and GSL-OH (Bol033373) expression was increased, the accumulation of two aliphatic glucosinolates GIV and GER was also increased (Figures 6A,B). Higher expression levels of the ST5b genes Bol026201 and Bol026202 were associated with an increased content of aliphatic GIV in the moderately resistant interaction of BN4303 × 03-02 s. However, Pearson correlation coefficients showed positive correlation only between GIV content and MYB28 (Bol007795) expression ( Figure 7A). GER content exhibited a positive correlation with both MYB28 (Bol007795) and GSL-OH (Bol033373). In the BN4303 × 00-100 s interaction, increased expression of MYB34 (Bol007760) and ST5a (Bol026200) was concomitant with higher levels of GBS, and upregulation of CYP81F4 (Bol032712, Bol032714) and CYP81F2 (Bol026044) was associated with increased NGBS content (Figures 6A,C). Pearson correlation tests indicated a positive association of GBS content with both MYB34 (Bol007760) and ST5a (Bol026200) expression along with that FIGURE 3 | Relative expression of aliphatic glucosinolate biosynthesis genes in leaf samples from the four cabbage lines (BN4059, BN4072, BN4098, and BN4303) under four different treatments (control, mock, inoculation with 03-02 s or 00-100 s) four days after inoculation. The mean of three biological replicates is presented. Vertical bars indicate standard error. Different letters indicate statistically significant differences between genotype × treatment combinations. Upward-pointing blue arrows indicate increased glucosinolate content in response to L. maculans infection.  of CYP81F4 (Bol032712, Bol032714), CYP81F4 (Bol028918), and CYP81F2 (Bol026044). NGBS content showed a positive correlation with CYP81F4 (Bol032712, Bol032714) expression along with CYP81F4 (Bol028918), CYP81F1, and IGMT1 (Bol020663) expression ( Figure 7B). In another moderately resistant interaction, BN4098 × 00-100 s, increased expression of ST5a (Bol026200) corresponded to an increased content of GBS (Figures 6A,C). Furthermore, the increased expression of CYP81F4 (Bol032712, Bol032714) and CYP81F2 (Bol026044) corresponded with an increased content of MGBS. Among these three genes, MGBS content showed positive correlation only with CYP81F2 (Bol026044). In the susceptible interaction BN4059 × 00-100 s with increased content of GBS, expression of CYP81F4 genes was also increased (Figures 6A,C, 7B). These data demonstrate a connection between these genes and changes in the glucosinolate profiles of resistant plants in response to L. maculans infection although a few deviations are often noted.

Resistance of Cabbage Inbred Lines Differs Between the Seedling and Adult Stages
It has long been known that blackleg resistance in adult plants could be different from that in seedlings, as resistance at these stages has been reported to be under differing genetic control (Ballinger and Salisbury, 1996;Larkan et al., 2016). It was also reported that adult plant resistance is not race-specific (Rimmer, 2006), a conclusion that was later debated by other studies (e.g., Marcroft et al., 2012;Raman et al., 2012). Despite our previous findings that line BN4098 is resistant to both 03-02 s and 00-100 s isolates at the seedling stage (Robin et al., 2017a), our current results show that is susceptible to 03-02 s at 3 months of age (Table 1). This difference in susceptibility-both at different growth stages and against different isolates of L. maculanscould be under genetic control, and will be a subject for further investigation.

The Aliphatic Glucosinolates GIV and GER and the Indolic Glucosinolates GBS and NGBS Are Associated with Resistance to Blackleg Disease
Associations between pathogen induced glucosinolate accumulation and plant resistance were detectable for certain combination of compounds only under particular line × pathovar combinations (Figure 6B). There were both positive and negative associations between glucosinolate abundance and resistance phenotypes. In addition, glucosinolate levels were often substantially altered under mock treatment compared to control plants. For instance, PRO and SIN contents were generally decreased in mock-treated plants compared to control plants in different line × pathovar combinations (Figure 2). However, the GIB, GIV, and GBS contents generally remained unaffected under mock treatment in different line × pathovar combinations compared to the control. We, therefore, considered glucosinolate level at the mock-treated samples as a benchmark to calculate pathogen-induced increase in accumulation. The contents of GBS and NGBS also exhibited upregulation in susceptible interactions in BN4059 and BN4072 (Figure 2). Aliphatic GIB and SIN contents were increased in BN4059 × 00-100 s combinations (Figure 2). Nevertheless, both aliphatic (glucoiberverin, glucoerucin) and indolic (glucobrassicin and neoglucobrassicin) glucosinolates increased to higher concentrations in the completely resistant line BN4303 upon blackleg infection (00-100 s isolate) ( Table 2), indicating that resistance in adult cabbage plants may be achieved through an accumulation of both aliphatic and indolic glucosinolate compounds. These findings are somewhat different from current data that suggest indolic glucosinolates alone may provide resistance to nectrotrophic and hemibiotrophic pathogens (Sanchez-Vallet et al., 2010;Hiruma et al., 2013; SIN, sinigrin;GIV, glucoiberverin;GER, glucoerucin;GBS, glucobrassicin;NGBS, neoglucobrassicin;and MGBS, methoxyglucobrassicin. et al., 2016;Wu et al., 2016). The absence of a pathogeninduced increase in either aliphatic or indolic glucosinolates was associated with only moderate resistance in the BN4098 × 00-100 s and BN4303 × 03-02 s combinations, respectively ( Table 2). By contrast, susceptible interactions exhibited an increase in one or only a few aliphatic or indolic glucosinolate compounds with a simultaneous decrease in other compounds. During infection of BN4059 for example, there was a significant increase of GIB, SIN, GBS, and NGBS but a decrease in aliphatic GIV and indolic MGBS, which may have resulted in this line's susceptibility to 00-100s (Figure 2). Furthermore, the marked upregulation of a single glucosinolate was also not associated with resistance. For example, GNA increased dramatically in BN4072 upon infection with 03-02 s, but neither aliphatic GIV nor any of the three indolic glucosinolates were induced, which may have resulted in this line's susceptibility to this isolate (Figure 2). These data suggest that specific glucosinolate profiles may confer resistance to cabbage to L. maculans infection in a genotype-and isolate-specific manner.

Increased Expression of St5b Genes Led to Increased Levels of GER and GIV in Resistant Cabbage
ST5b genes are involved in secondary modifications of desulfoglucosinolates to GIB and GIV and then to other aliphatic glucosinolate compounds ( Figure S1, Liu et al., 2014;Yi et al., 2015). In a recent study, increased expression of ST5b genes was found to be associated with higher levels of three aliphatic glucosinolates, glucoraphanin, SIN, and GNA, which accumulated via GIB and GIV biosynthesis . In this study, infection of BN4303 with either 00-100 s or 03-02 s induced the expression of two ST5b genes, Bol026201 and Bol026202; infected plants also showed a striking increase in GIV and GER levels ( Figure 6A). Thus, our results indicate that infection by the blackleg pathogen induced the upregulation of these genes, subsequently leading to the increase in GIV that was associated with the resistance of plants (Figures 2, 3). However, this correlation based association is required to be validated further through genetic studies.

GIV Is a Candidate to Play a Crucial Role in Blackleg Disease Resistance
The conspicuous contrasting patterns of GIV accumulation between the most resistant combination BN4303 × 00-100 s, which showed high levels of GIV, and the most susceptible combinations BN4059 × 00-100 s and BN4059 × 03-02 s, which showed low levels of GIV, highlight GIV as a potentially important contributor to glucosinolate-mediated resistance in adult cabbage plants (Figure 2). Moreover, a decrease in GIV levels upon infection in the two most susceptible combinations BN4059 × 03-02 s and BN4059 × 00-100 s further indicated that lower GIV levels are associated with reduced plant resistance against blackleg infection (Figure 2). By contrast, in the modestly resistant combinations BN4098 × 00-100 s and BN4303 × 03-02 s, GIV was completely absent and highly upregulated, respectively. This discrepancy indicates that GIV cannot provide full resistance alone, but rather is likely to be a contributor to resistance, possibly together with indolic GBS and NGBS (Figure 2). A higher accumulation of GIV coupled with GBS and NGBS was also found to show a positive association with lower feeding scores as reflected by less leaf area damage by the Diamondback moth in cabbage (Robin et al., 2017b).
MYB34 Likely Activates Glucosinolate Biosynthesis Genes, Leading to Increases in GBS and NGBS in Cabbage Lines Resistant to Blackleg Disease Experimental evidence suggests that MYB34 genes directly regulate the biosynthesis of indolic glucosinolates in Arabidopsis thaliana (Frerigmann and Gigolashvili, 2014) and Brassica oleracea Yi et al., 2016). MYB34 together with MYB51 and MYB122 provided resistance to the necrotrophic fungal pathogen Plectosphaerella cucumerina through the biosynthesis of indolic glucosinolates, where PEN2 plays a key role in activating indolic glucosinolates in response to pathogen attack (Frerigmann et al., 2016). MYB34 (Bol007760) is also induced in broccoli upon methyl jasmonate (MeJA) treatment which is a biotic elicitor that responds to jasmonic acid (JA) signaling . Likewise in this study, MYB34 genes Bol007760 and Bol017062 showed a 9.35-and 2.13-fold increase, respectively, upon inoculation with 00-100s in the resistant line BN4303 compared to mock-treated plants (Figure 4). Along with the corresponding increase in indolic glucosinolates, this indicates that MYB34 genes could play a role in the trans-activation of genes necessary for indolic glucosinolate biosynthesis in response to L. maculans infection (Figure 2). Consistent with our results, upregulation of MYB34 (Bol017062) in cabbage plants in a previous study was associated with the accumulation of indolic NGBS, which is a derivative of another indolic glucosinolate GBS .
Some notable MYB genes, MYB34 (Bol007760), MYB28 (Bol017019, Bol036743), and MYB29 (Bol008849), as well as other genes such as ST5b, ST5c and FMOGS-OX5 showed higher expression, in general, in BN4098 whereas CYP81F1 (Bol017375, BOl017376) genes showed comparatively higher expression in BN4072. However, these differences in expression did not lead to significant differences in glucosinolate accumulation of the respective indolic and aliphatic glucosinolates, i.e., glucosinolate accumulation was comparable in those lines to the other lines. Hence, this observation is a subject for further investigation. In a recent study, cabbage line BN3383 showed higher expression of MYB29 (Bol008849), ST5c and FMOGS-OX5 compared to other three cabbage inbred lines but that particular line exhibited comparatively lower levels of total glucosinolates  indicating that expressions of certain biosynthesis genes were not always consistent with higher accumulation of particular glucosinolates from the relevant biosynthesis pathways.

Accumulation of Indolic GBS in Cabbage Lines Resistant to Blackleg Disease Is Triggered by Increased Expression of Biosynthesis Genes
ST5a genes take part in secondary modifications of desulfoglucosinolates for the biosynthesis of GBS ( Figure S1, Liu et al., 2014;Yi et al., 2015). Concentrations of both GBS and NGBS, which have a proven role in antifungal responses in plants, increased in the resistant line BN4303 in response to 00-100s, and this was associated with a remarkable upregulation of ST5a (Bol026200), CYP81F4 (Bol032712), CYP81F4 (Bol032714), and CYP81F2 (Bol026044) (Figure 5). In a previous study, treatment with MeJA led to a 2,400-fold increase in CYP81F4 in broccoli and a 10-fold increase in cabbage , suggesting that resistance is mediated via signaling pathways involved in indolic glucosinolate metabolism. ST5a (Bol026200) and CYP81F2 (Bol026044) were also upregulated in the moderately resistant interaction between BN4303 and 03-02s and CYP81F4 (Bol032712) was upregulated in the moderately resistant BN4098 × 00-100 s interaction (Figure 5), suggesting that these genes are associated with the enhanced accumulation of indolic glucosinolates in these interactions. Sotelo et al. (2016) also reported a similar association between CYP81F2 expression and GBS levels in B. oleracea.
By contrast, accumulation of indolic GBS and NGBS and few other aliphatic glucosinolates were noted in susceptible combinations, coupled with upregulation of some notable genes, for example: CYP81F4 genes (Figures 6A,C). These observations raise the question of whether the abundance of intact glucosinolates at a particular time point always reflects their physiological functions. Abundance of glucosinolates in leaf tissues at a particular time-point can be a consequence of simultaneous activation of biosynthesis and catabolism by myrosinases, which can provide upregulation of the bioactive form of the compounds at a particular time point.
Increased Expression of CYP81F2 Leads to Accumulation of MGBS and Is Associated with Moderate Blackleg Resistance in Adult Cabbage Plants MGBS levels have previously been reported to increase by 30-47% in response to L. maculans infection in Brassica napus after 5-8 days of inoculation (Wretblad and Dixelius, 2000). In an in vitro study, Mithen et al. (1986) also reported anti-fungal activity for MGBS along with two other glucosinolates, SIN and GBS. Here, we observed an increase in MGBS in BN4098 after 00-100s inoculation (Figure 2). The facts that BN4098 showed moderate resistance to 00-100 s and that levels of none of the aliphatic glucosinolates increased suggest a role for MGBS in conferring resistance in this particular interaction. Since CYP81F2 (Bol026004) and CYP81F4 (Bol032712, Bol032714), which are involved in methoxylation and the conversion of GBS to 4-MGBS, are up-regulated in this genotype × isolate combination, it is likely that both the gene and the glucosinolate play roles in conferring resistance to BN4098 to 00-100 s. Our conclusion is consistent with findings of Bednarek et al. (2009) that CYP81F2 (Bol026004) and the myrosinase PEN2 induce antifungal defense (Figure 5).

CONCLUSIONS
Glucosinolate profiling and expression analysis of glucosinolaterelated genes in response to blackleg infection identified a direct association between the genes and their corresponding glucosinolates in 3-month-old cabbage plants as supported by both heat map and correlation analyses. This study revealed that the simultaneous, pathogen-induced accumulation of both aliphatic GIV and GER and indolic GBS and NGBS were associated with resistance to blackleg disease in a genotypespecific manner in cabbage. Different glucosinolate profiles were associated with different levels of resistance in a genotypeand isolate-specific manner. Although the presence of certain glucosinolates of either the aliphatic or indolic class was associated with moderate resistance, only the presence of both was associated with complete resistance. The glucosinolates and their corresponding genes identified in this study are candidate genetic and biochemical determinants of resistance and could be tested in efforts to improve blackleg resistance in cabbage.

AUTHOR CONTRIBUTIONS
IN, JP, and AR conceived of and designed the study. AR managed and inoculated the experimental plants, collected samples, prepared cDNA, performed the qPCR analysis, prepared samples for HPLC and wrote the manuscript. MH carefully commented on a draft of the manuscript. HK conducted the HPLC analysis. GY and RL assisted with the cDNA preparation and qPCR analysis.