Transcriptome analysis of two lines of Brassica oleracea in response to early infection with Xanthomonas campestris pv. campestris

Abstract

Black rot, caused by Xanthomonas campestris pv. campestris (Pammel) Dowson (Xcc), is a devastating disease of cruciferous crops worldwide and a threat to the economic production of Brassica oleracea L. in China. Strategies to control this disease are limited. In addition, knowledge of the molecular mechanisms of pathogenicity and host resistance are still only poorly understood. In this study, we performed transcriptomic analyses of the leaves of two contrasting B. oleracea lines (Xcc-resistant line QP07 and Xcc-susceptible line DBP71) to study the early defence response induced by infection with Xcc. We identified 3357 upregulated and 4091 downregulated genes between QP07 and DBP71 at 0, 12, 24, 48 and 96 h post-inoculation (hpi). A functional annotation pathways analysis of the differentially expressed genes indicated that the glucosinolate biosynthetic and catabolic pathways were enhanced, and ROS scavenging, hormonal, receptor-kinase-related genes and nucleotide binding site (NBS)-encoding resistance genes were expressed during the early infection period. In addition, we found that the host plants may respond to Xcc infection by actively regulating photosynthetic energy metabolism. The results provide new insights to help understand the different responses to Xcc in resistant and susceptible host genotypes, and to aid in the development of effective strategies for the prevention of black rot.

Résumé

La pourriture noire, causée par Xanthomonas campestris pv. campestris (Pammel) Dowson (Xcc), est une maladie dévastatrice qui s’attaque aux crucifères partout dans le monde et qui constitue une menace pour la production économique de Brassica oleracea L. en Chine. Les stratégies de lutte contre cette maladie sont limitées. De plus, les connaissances relatives aux mécanismes moléculaires de la pathogénicité et à la résistance de l’hôte sont à ce jour fragmentaires. Dans cette étude, nous avons procédé à des analyses trancriptomiques des feuilles de deux lignées opposées de B. oleracea (lignée résistante QP07 à Xcc et lignée réceptive DBP71 à l’égard de Xcc) afin d’étudier la réaction de défense première induite par l’infection à Xcc. Nous avons identifié 3 357 gènes régulés à la hausse et 4 091 gènes régulés à la baisse chez QP07 et DBP71, respectivement, à 0, 12, 24, 48 et 96 heures après inoculation. Une analyse fonctionnelle des voies d’annotation des gènes différentiellement exprimés a indiqué que les voies biosynthétiques et cataboliques du glucosinolate étaient améliorées et que le pillage des FRO, les gènes d’hormones et de récepteurs kinases ainsi que le site de fixation des nucléotides codant les gènes de résistance étaient exprimés au début de la phase infectieuse. En outre, nous avons noté que les plants hôtes peuvent réagir à l’infection à Xcc en régulant vivement le métabolisme de l’énergie photosynthétique. Les résultats fournissent un nouvel éclairage quant à la compréhension des diverses réactions à Xcc chez les différents génotypes hôtes, tant résistants que réceptifs, et contribuent à l’élaboration de stratégies efficaces pour prévenir la pourriture noire.

Keywords:

• black rot
• Brassica oleracea
• disease resistance
• transcriptome
• Xanthomonas campestris pv. campestris

Mots clés:

• Brassica oleracea
• pourriture noire
• résistance à la maladie
• transcriptome
• Xanthomonas campestris pv. campestris

Introduction

Black rot disease affects cruciferous crops worldwide and is caused by Xanthomonas campestris pv. campestris (Xcc), a vascular pathogen spread via infected seeds (Cook et al. 1952; Vicente & Holub 2013). In addition, the pathogen can be disseminated by insects, diseased plants, farm implements, irrigation water and the seed trade (Sharma et al. 2017). Under natural conditions, the Xcc enters the margins of the leaves across the hydathodes, or through wounds on the plant, forming a triangular chlorotic yellow spot along the leaf edge spreading with the vein, gradually expanding into a V-shaped chlorosis around the leaf. The vein then becomes necrotic. When the pathogen spreads along the veins and petioles to the stems and roots of the plants, the vascular tissue in the roots and stems turns black, and the whole plant dies, sometimes accompanied by a secondary soft rot (Vicente & Holub 2013). Traditional measures used to manage black rot include the control of weeds and diseased plants that harbour the pathogen, the use of disease-free seeds and bactericidal seed treatments. However, these methods are all preventative, and there is no effective method to control black rot once the plant is infected (Vicente & Holub 2013; Singh et al. 2018). Therefore, the development and planting of disease-resistant varieties is a more effective solution for black rot prevention and reduction of yield losses than on other management measures.

Currently, Xcc is divided into 11 races, and races 1 and 4 are the most pathogenic and important, with over 90% of the black rot in the world caused by these two physiological races (Kamoun 1992; Vicente et al. 2001; Fargier & Manceau 2007; Vicente & Holub 2013; Cruz et al. 2017). A number of sources of resistance have been identified in the A and B genome of the brassicas, including Brassica rapa (AA), Brassica nigra (BB), Brassica juncea (AABB), Brassica napus (AACC), and Brassica carinata (BBCC) (Taylor et al. 2002; Tonguç & Griffiths 2004; Griffiths et al. 2009; Sharma et al. 2016, 2017). However, there are only a few reports of resistance genes in the C genome; moreover, Brassica oleracea (CC) is more sensitive to Xcc than other brassica species and reports of resistance in B. oleracea are mostly race-non-specific (Sharma et al. 1972; Jamwal & Sharma 1986; Griesbach et al. 2003; Tonu et al. 2013). Recently, Dey et al. (2015) reported the creation of black rot-resistant lines in cauliflower through a combination of interspecific hybridization and embryo rescue. Saha et al. (2016) identified new sources of resistance to race 1 by screening B. oleracea germplasm sources. Interestingly, the majority of Arabidopsis thaliana genotypes showed broad-spectrum resistance to the main pathogenic races of Xcc (Holub 2007). Although some resistant lines have been identified in brassica crops, and some resistance genes have been mapped, none have been cloned so far. Thus, an understanding of host resistance mechanisms and the interaction between B. oleracea and Xcc is necessary to improve black rot management.

Plants have evolved a range of effective defence systems over the course of long-term interactions and co-evolution with pathogens (Keshavarzi et al. 2004; Jones & Dangl 2006;). The first layer of the plant immune system, known as PTI (PAMP trigger immunity), is triggered by pattern recognition receptor (PRR) proteins that identify pathogen-related molecular patterns (PAMPs) and provide a type of underlying defence. However, pathogens can inhibit PTI by secreting effectors into host cells. Therefore, the second layer of the plant immune system can play a significant role in the cell. Plant disease resistance (R) genes identify cognate pathogen effectors (avirulence proteins) and trigger the activation of defence (Flor 1971; Dangl & Jones 2001). This type of resistance is commonly referred to as ETI (effector-triggered immunity). Resistance (R) proteins are considered to be important regulators of plant defence signalling. It has been reported that many R genes are involved in the resistance of brassica crops to different pathogens. Two clubroot resistance genes CRa and Crr1a of B. rapa have been cloned and found to be classic R genes (Ueno et al. 2012; Hatakeyama et al. 2013). Lv et al. (2014) mapped Fusarium wilt-resistance gene FOC1 in B. oleracea, which encodes a TIR-NBS-LRR type R protein. Lee et al. (2015) identified four quantitative trait loci (QTL) acting against Xcc by whole-genome resequencing and identified different NBS-LRR genes. Afrin et al. (2018) identified R genes related to black rot resistance in B. oleracea, reporting that seven NBS-encoding genes were more highly expressed in Xcc-resistant vs. Xcc-susceptible lines. However, the response mechanisms related to these R genes have yet to be fully elucidated in B. oleracea when challenged with Xcc.

‘Omics’ technologies have been important tools to explore and understand the molecular mechanisms of the B. oleracea-Xcc interaction. Izzah et al. (2014) developed many EST-based markers by RNA sequencing of two cabbage parental lines (Xcc-resistant and Xcc-susceptible). Santos et al. (2019b) focused on the changes in microRNA expression in B. oleracea and identified four miRNAs that were up-regulated in a resistant genotype and down-regulated in a susceptible genotype. Various groups have explored changes in protein abundance after inoculation of Xcc-resistant and Xcc-susceptible B. oleracea lines with Xcc, finding that most of the differentially abundant proteins were involved in photosynthesis, energetic metabolism, hormone metabolics and disease/defence responses (Villeth et al. 2016; Ribeiro et al. 2018; Santos et al. 2019a).

In previous studies, growth of the Xcc bacterial population in a resistant line was observed only in the first 24 h following inoculation, and then decreased significantly, thus suggesting that resistant hosts block bacterial growth at 24 h (Villeth et al. 2009; Andrade et al. 2010). To clarify the response mechanisms of B. oleracea in response to Xcc, we used transcriptome sequencing to analyse the changes in gene expression in B. oleracea 12, 24, 48 and 96 h following infection by the pathogen.

Materials and methods

Isolation and identification of bacterial strains

The bacterial strains were isolated from brassica crop leaves, with typical ‘V’-shaped chlorosis of black rot disease, collected from diverse regions of Shaanxi Province, China. Surface contamination of the leaves was removed by rinsing the surface with distilled water and disinfecting it with 75% ethanol for 1 min. Leaf tissue segments were excised from the lesion margins by the addition of sterile 0.85% saline; the saline suspension was streaked on nutrient agar (NA) medium with an inoculation loop and the petri dishes were cultured at 28°C for 2 days (Burlakoti et al. 2018). Pale yellow colonies were picked using an inoculation loop and streaked in the new NA medium for further purification. We isolated a total of 20 strains (Supplementary Table 1). Presumptive Xcc isolates were identified to species using a species-specific PCR assay suggested by the International Seed Testing Association (ISTA; Roberts and Koenraadt 2014), and the races were identified using race 1 and 4 specific molecular markers for Xcc (Rubel et al. 2017).

Pathogen and plant material preparation

Two contrasting B. oleracea DH lines were used in our study, including line QP07 (Xcc-resistant) and line DBP71 (Xcc-susceptible). These two lines were developed by microspore culture after three rounds of inbreeding using a variety moderately resistant to Xcc. The plants were grown in pots under greenhouse conditions at the Northwest Agriculture and Forestry University, Yangling, China. Inoculations were done on 45-day-old plants. This stage was well suited for diseases resistance screening as the plants were showing vigorous growth. The Xcc strain YL-1 that we had isolated and purified was grown in agar-free nutrient medium at 28°C and 200 rpm in a shaking incubator. The bacterial suspension was prepared at the rate of 1 × 10⁸ cfu mL⁻¹ for inoculation.

Pathogen inoculation and sample collection

Forty-five-day-old plants of QP07 and DBP71 were inoculated with Xcc by spraying the leaves with the bacterial suspension. The inoculated seedlings were cultured in a greenhouse with a 14 h light photoperiod at 28°C and kept humid during the treatment stage. Sample leaves of QP07 and DBP71 seedlings were collected at five time points (before inoculation, 12, 24, 48, and 96 h after inoculation) for RNA-Seq analysis. Three biological replicates were used with five plants per replicate at each infected stage. Following collection, the leaves were immediately frozen in liquid nitrogen and stored at −80°C for RNA isolation.

RNA isolation

Total RNA was extracted using a plant RNA Extraction kit (Takara, Dalian, China). The quality of the extracted RNA was evaluated with a Nanodrop spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA) and by running aliquots on a 1.0% agarose gel. The total RNA after treatment with gDNase was used as the template to synthesize first strand cDNA with a FastKing RT Kit (TIANGEN BIOTECH, Beijing, China).

RNA Sequencing

cDNA library construction and sequencing were performed by the Biomarker Technology Company (Beijing, China). The mRNA was enriched from the total RNA using beads with Oligo(dT), and mRNA was then disrupted to form short fragments under an elevated temperature. The first-strand cDNA was synthesized using the mRNA fragments as templates with random primers, and then the second-strand cDNA was synthesized by adding DNA polymerase I, dNTPs, RNase H and buffer. The cDNA was subjected to end reparation and single nucleotide adenine addition. The short fragments were then connected with adapters, and the suitable fragments were selected from an agarose gel and enriched by PCR amplification. Finally, RNA-Seq was performed using the Illumina Hi-Seq X Ten sequencing platform. The raw RNA-Seq data have been deposited at SRA (http://www.ncbi.nlm.nih.gov/sra/) under the accession number SUB6182237.

Digital gene expression analysis

After sequencing, the sub-sequences and low-quality sequence reads were removed from the original data. The clean reads were then mapped to the B. oleracea genome (http://plants.ensembl.org/Brassica_oleracea/Info/Index) and reference gene sequences using TopHat2 (Parkin et al. 2014). The gene expression levels were estimated by the amount of reads mapped to the reference sequences using the RPKM method (Mortazavi et al. 2008). The resulting P values were corrected using the Benjamini and Hochberg approach (Benjamini 1995), and the corrected P-values were used to determine the false discovery rate (FDR). Sequences were regarded to be significantly differentially expressed if the |log2 (fold change)| was ≥1 and the FDR was <0.01.

DEGs analysis and annotation

For functional annotation, gene functions were subjected to the following databases: NCBI nonredundant protein sequences (Nr, ftp://ftp.ncbi.nlm.nih.gov/blast/db), Swiss-Prot database (http://ftp.ebi.ac.uk/pub/databases/swissprot), the cluster of Orthologous Groups (KOG/COG) databases (http://www.ncbi.nlm.nih.gov/COG), and protein family (PFam, http://pfam.xfam.org/). Additionally, to perform GO and pathway annotation to the DEGs, the Blast2GO programme was used to acquire the GO annotation of the unigenes. The WEGO and Top GO software programmes were used to perform GO functional classifications and enrichment analyses to understand the distributions of the gene functions. The KEGG pathway annotations and enrichment analyses were performed using BLASTALL software against the KEGG databases (http://www.genome.jp/kegg/).

Validation of the RNA-Seq data by qRT-PCR

To confirm the RNA-Seq results obtained, we selected 20 DEGs for qRT-PCR analysis. The gene-specific primers were designed using Primer3 web (http://primer3.ut.ee/) and are listed in Supplementary Table 2. The qRT-PCR reactions were conducted on a CFX 96 real-time PCR system (Bio-Rad, Hercules, CA). The thermal cycling conditions were as follows: 95°C for 3 min, followed by 40 amplification cycles of 95°C for 5 s, and 60°C for 30 s, and melting curves were generated by increasing the temperature from 55°C to 95°C in increments of 0.5°C per 10 s to verify the specificity of the PCR amplification. An analysis of the gene expression was performed for all the samples at 0, 12, 24, 48, and 96 h post-inoculation of QP07 and DBP71 with Xcc. B. oleracea Actin2 (BoActin2) was used as the housekeeping gene and three independent biological and technical replicates were performed per primer. The relative expression levels of 20 DEGs were calculated using the 2^−ΔΔCt method and results were analysed in MS Excel (Livak and Schmittgen 2001).

Results

Identification of bacterial strains and symptoms of the inoculated B. oleracea leaves

The bacterial isolates were identified to the species level using the species-specific PCR assay suggested by the International Seed Testing Association (ISTA; Roberts and Koenraadt 2014). Among the 20 strains that we isolated, we found three amplicons (370, 441 and 619 bp) that identified them all as Xcc (Fig. 1a,b,c). The Xcc strains that we isolated were identified using race 1 and 4 specific molecular markers (Rubel et al. 2017). None of the 20 Xcc strains were amplified by the Xcc race 1 primer Xcc_47R1, and the 462 bp amplicon specific for the Xcc race 4 was observed in 19 of the Xcc isolates (Fig. 1d). Therefore, Xcc strain YL-1 (race 4) was selected for inoculation.

Fig. 1 PCR amplification products from 20 Xanthomonas campestris pv. campestris (Xcc) strains in a species-specific PCR assay and race 4 specific molecular markers. (a), (b) and (c): Three amplicons (370, 441 and 619 bp) from all 20 strains in the PCR assay; (d): Amplicons (462 bp) obtained with Xcc-Race 4 specific molecular markers from all 20 strains. Lanes 1–20: 20 Xcc strains (the corresponding strains, host and sampling location for each lanes are included in Supplementary Table 1)

At approximately 72 h post-inoculation, DBP71 began to show slight chlorosis along the leaf edge. At approximately 1 week post-inoculation, a typical V-shaped chlorosis appeared around the leaf. Up through 2 weeks after the inoculation, we found severe V-shaped chlorosis around the leaf of line DBP71, but no black rot symptoms were detected in line QP07 (Fig. 2). This result indicated that QP07 showed more resistance to Xcc than DBP71.

Fig. 2 Black rot symptoms on the Brassica oleracea lines QP07 and DBP71 infected by Xanthomonas campestris pv. campestris (Xcc) at 0, 2, 3, 7, and 14 days (left to right) following inoculation with Xcc. The black rot symptoms are marked by red arrows

RNA-Seq analysis

To evaluate the molecular mechanisms of B. oleracea against Xcc, RNA-Seq analysis was performed on three biological replicates of QP07 and DBP71 samples collected at each time point (0, 12, 24, 48 and 96 hpi) after Xcc inoculation. Raw reads were generated for RNA-Seq using 150-bp paired end Illumina sequencing. After sequencing quality control, a total of 237.32 Gb of data were obtained, and the filtered data of each of the 30 cDNA libraries reached 6.06 Gb. The GC percentage of the sequence data from the 30 libraries was approximately 48%, and 91% of bases were Q30 or better, indicating that the accuracy and quality of the sequencing data were adequate for further analysis. Almost 89% of the sequenced reads were mapped to the reference B. oleracea genome (Supplementary Table 3). The mapped reads are aligned with different regions of the reference genome, and approximately 85% of the reads mapped to exons (Supplementary Table 3).

Analysis of differentially expressed genes (DEGs)

A total of 7430 DEGs between the two lines (QP07-vs-DBP71) were included, including 3357 up-regulated and 4091 down-regulated genes (Fig. 3, Supplementary Table 6). Compared with DBP71, 1942 genes were up-regulated, and 2399 genes were down-regulated in the 0 hpi QP07 sample. Following inoculation with Xcc, the numbers of DEGs in QP07 (as shown in Fig. 1) included 729 up-regulated and 1292 down-regulated at 12 hpi; 1856 genes were up-regulated and 2320 genes were-down regulated at 24 hpi; 2015 genes were up-regulated and 2533 genes were down-regulated at 48 hpi; 2026 genes were up-regulated and 2660 genes were down-regulated 96 hpi. Moreover, only 1388 DEGs were detected as common DEGs at all five points, while 517 genes were up-regulated, and 891 genes were down-regulated (Fig. 3).

Fig. 3 Number of differentially expressed genes in the Brassica oleracea line QP07 (a) up-regulated or (b) down-regulated (|log2 (fold change)| ≥1 and the FDR < 0.01) compared with DBP71. Numerals inside the parentheses indicate the number of genes expressed at each time-point. The total number of DEGs is noted at the bottom of each Venn diagram. hpi, h post-inoculation

Functional annotation analysis of the DEGs

Based on the GO terms, DEGs between the two lines (QP07-vs-DBP71) were assigned to three GO ontologies: cellular component, molecular function and biological process. For the five time-points, there were 6014 DEGs enriched in the cellular component ontology, and more DEGs were involved in the ‘cell,’ ‘cell part’ and ‘organelle’ categories (Fig. 4a). In the molecular function ontology, 4873 DEGs were assigned to 15 GO categories (Fig. 4b). The main enriched categories were ‘binding’ and ‘catalytic activity.’ There were 5477 DEGs enriched in the biological process ontology, with most of the genes involved in ‘cellular process,’ ‘single-organism process,’ ‘metabolic process,’ ‘response to stimulus’ and ‘biological regulation’ (Fig. 4c). The GO enrichment analysis of the DEGs in the biological process ontology shows that the top GO term at each time-point is represented as a directed acyclic graph (Supplementary Figs. 1–5). We extracted the top 10 GO terms at each time-point, and a total of 25 GO terms appeared in five stages (Table 1). Among these, the GO terms ‘glucosinolate catabolic process’ (GO:0019762) and ‘defense response signaling pathway, resistance gene-independent’ (GO:0010204) were present at all five time-points. Furthermore, the GO term ‘immune response-regulating signaling pathway’ (GO:0002764) was present at 0, 12, 24, and 96 hpi. In addition, we also found that ‘defense response to fungus, incompatible interaction’ (GO:0009817) and ‘regulation of plant-type hypersensitive response’ (GO:0010363) were only present at 48 hpi (Table 1).

Table 1. Top 10 Gene Ontology (GO) terms (biological process) that were significantly enriched at five time-points after inoculation of Brassica oleracea with Xanthomonas campestris pv. campestris.

Time-point (h post-inoculation, hpi)	GO terms number	GO terms
0 hpi, 12 hpi, 24 hpi, 48 hpi, 96 hpi	2	glucosinolate catabolic process (GO:0019762); defence response signalling pathway, resistance gene-independent (GO:0010204)
0 hpi, 12 hpi, 24 hpi, 48 hpi	1	Response to ozone (GO:0010193)
0 hpi, 12 hpi, 24 hpi, 96 hpi	2	Response to alcohol (GO:0097305); immune response-regulating signalling pathway (GO:0002764)
0 hpi, 12 hpi, 24 hpi	1	nitrile biosynthetic process (GO:0080028)
0 hpi, 24 hpi, 96 hpi	1	Response to stress (GO:0006950)
0 hpi, 48 hpi	1	Response to karrikin (GO:0080167)
0 hpi, 96 hpi	1	Response to fructose (GO:0009750)
24 hpi, 48 hpi	1	maltose metabolic process (GO:0000023)
24 hpi, 96 hpi	1	Vacuolar protein processing (GO:0006624)
0 hpi	1	Cinnamic acid ester metabolic process (GO:0009801)
12 hpi	4	Response to light stimulus (GO:0009416) response to chitin (GO:0010200) iron ion transport (GO:0006826) cellular response to iron ion starvation (GO:0010106)
24 hpi	1	Post-embryonic development (GO:0009791)
48 hpi	5	Defence response to fungus, incompatible interaction (GO:0009817) nitrate transport (GO:0015706) regulation of plant-type hypersensitive response (GO:0010363) protein targeting to membrane (GO:0006612) response to nitrate (GO:0010167)
96 hpi	3	Stamen development (GO:0048443) indolebutyric acid metabolic process (GO:0080024) positive regulation of long-day photoperiodism, flowering (GO:0048578)

Fig. 4 GO assignment of the differentially expressed genes (DEGs) in Brassica oleracea lines QP07 and DBP71. The unigenes were assigned to three main categories: cellular component (a), molecular function (b), and biological process (c). The x-axis indicates the number of annotated DEGs. hpi, h post-inoculation

The differentially expressed gene sequences were mapped to the reference canonical pathways in KEGG. A total of 1468 DEGs were annotated in the KEGG database and assigned to 126 KEGG pathways, of which 17 significantly enriched pathways (P ≤ 0.01) were detected (Supplementary Table 4). The most common term was ‘Cysteine and methionine metabolism,’ followed by ‘Photosynthesis – antenna proteins.’ The ‘Glutathione metabolism’ and ‘Sulfur metabolism’ pathways were also enriched. Based on the KEGG annotation analyses, we also extracted the top 10 metabolic pathways with the highest number of DEGs, and the top three metabolic pathways were ‘Biosynthesis of amino acids,’ ‘Carbon metabolism’ and ‘Plant hormone signal transduction’. The ‘Plant-pathogen interaction’ pathway was also enriched (Supplementary Table 5).

DEGs involved in the resistance to XCC

To survive infection by pathogens, plants require an effective response to restrict the further propagation of the pathogen. We utilized our transcriptome data analysis and the reported data in the literature to better understand the defence mechanism that B. oleracea employs against infection by Xcc. We analysed the response of photosynthesis (45 DEGs), ROS scavenging (40 DEGs), plant hormones (JA 17 DEGs, ET 16 DEGs, SA 9 DEGs), glucosinolates metabolites (32 DEGs), receptor kinase (30 DEGs) and resistance (R) genes (18 DEGs) after inoculation with Xcc (Supplementary Table 7).

The existence of ROS is a double-edged sword for plants, as they can inhibit the damage caused by pathogens, but also directly damage plant cell components. Therefore, the balance of ROS in cells is very important. Forty DEGs were identified as related to ROS scavenging in QP07 and showed an irregular regulation.

Identification of the DEGs involved in photosynthesis

According to the results of KEGG pathway analyses, photosynthesis was differentially regulated in QP07 compared with DBP71 after Xcc infection. Forty-five DEGs were identified in the KEGG pathway as related to photosynthesis in QP07, including ‘Photosynthesis,’ ‘Photosynthesis – antenna proteins,’ and ‘Carbon fixation in photosynthetic organisms’ (Supplementary Table 7). Compared with DBP71, the most DEGs associated with these pathways were down-regulated in QP07 after Xcc infection. However, interestingly, the number of up-regulated DEGs was significantly increased at 48 hpi, which included chlorophyll a-b binding protein, malate dehydrogenase, fructose-bisphosphate aldolase and PsbQ-like protein.

Identification of the DEGs involved in the SA, JA and ET signalling pathways

SA, JA and ET play a significant role in the plant’s defence system as the most important plant defence hormones. This study indicates that the biosynthesis of a number of these hormones was affected by Xcc infection. The results of the plant hormone pathway analysis in GO and KEGG indicated that there were 17 genes (10 up-regulated and seven down-regulated) related to the JA pathway. Sixteen genes (11 up-regulated and five down-regulated) were involved in the ET pathway, and nine genes (six up-regulated and three down-regulated) were associated with the ET pathway (Supplementary Table 7).

Quantitative RT-PCR validation

To confirm the results of the RNA-Seq analysis, 20 DEGs were selected for qRT-PCR analysis, including some involved in photosynthesis, ROS scavenging, plant hormone metabolism, glucosinolate metabolism, as well as resistance (R) genes and receptor like proteins (Supplementary Table 2). The results showed that all 20 DEGs were expressed in a manner similar to that observed via RNA-Seq analysis (Fig. 5), confirming the high degree of reliability of the RNA-Seq method used in this study.

Fig. 5 Validation of the gene expression patterns of selected genes identified via RNA sequencing using qRT-PCR analysis. The y-axis shows the fold-change between the two lines

Discussion

Black rot leads to a significant decline in the yield and quality of B. oleracea, and is one of the most destructive and widespread seed-borne bacterial diseases of this crop (Williams 1980; Singh et al. 2011, 2018). Unfortunately, few highly resistant lines have been identified in B. oleracea. Therefore, to develop an effective strategy for Xcc-resistance breeding, it is important to understand the mechanisms of resistance that different host lines utilize in response to black rot. In this study, we conducted a transcriptomic analysis of the defence responses to Xcc in resistant and susceptible varieties of B. oleracea.

Research has indicated that a range of resistance proteins have evolved in plants, which activate immune responses and prevent pathogen infection (Jones & Dangl 2006). Among these, plants use cell surface specific receptor proteins to perceive pathogen-secreted effectors to regulate immunity, resulting in PTI (Jones & Dangl 2006; He et al. 2018). In this study, we found a multitude of PRR genes differentially expressed between QP07 and DBP71. However, we found no differentially expressed key PRR genes that trigger PTI, such as flagellin sensing 2 (FLS2), chitin elicitor-binding protein (CEBiP), brassinosteroid insensitive 1-associated kinase 1 (BAK1) and chitin elicitor receptor kinase (CERK). These phenomena were also observed in the studies of Chen et al. (2016) and Wang et al. (2019). In addition, there are many examples of the differential expression of cysteine-rich receptor-like kinase genes (CRKs), some of which are related to plant defence and are up-regulated after bacterial infection in Arabidopsis (Czernic et al. 1999; Chen et al. 2003). We also noticed that the differentially expressed genes were more apparent on 0, 24, and 48 hpi. This suggests that the recognition between B. oleracea and Xcc occurs at an early stage of infection and within a relatively short period.

Disease resistance in plants can result from the interaction between pathogen avirulence (avr) genes and the corresponding host resistance (R) genes (Dangl & Jones 2001; Bent & Mackey 2007). Eighteen genes for R proteins were found to be differentially expressed between the two lines, including five transcripts homologous to Arabidopsis ‘ADR1-like 1,’ ‘Resistance to P. syringae pv maculicola 1,’ ‘Resistant to P. syringae 5,’ ‘Recognition of P. parasitica 8’ and ‘RPP8L4’, which were up-regulated at different time-points in QP07. Previous studies indicated that ADR1 contributed to resistance to the biotrophic pathogens Peronospora parasitica and Erysiphe cichoracearum (Grant et al. 2003). RPM1 can perceive the Pseudomonas syringae TTSS effectors AvrRpm1 and AvrB, leading to the hypersensitive reaction (HR) and restriction of Pseudomonas growth; overexpression of RPM1 conferred broad-spectrum resistance to the fungal pathogen Magnaporthe oryzae and the bacterial pathogen Xanthomonas oryzae pv. oryzae (Xoo) in rice (Gao et al. 2011; El Kasmi et al. 2017; Li et al. 2019). Ade et al. (2007) showed that RPS5 confers resistance to the bacterium P. syringae. It was also reported that RPP8 confers resistance to downy mildew in Arabidopsis (McDowell et al. 1998; Mohr et al. 2010). Therefore, these R genes may be vital for mediating resistance to Xcc in B. oleracea, and the gene functions may be conserved across species.

According to the lifestyles and infection strategies of pathogens, plant hormones participate in plant-pathogen interactions by regulating defence response signal networks, enabling plants to produce the most suitable defence response to pathogens (O’Donnell et al. 2001; Anderson et al. 2004; Prerostova et al. 2018). Salicylic acid (SA) and jasmonic acid (JA) are involved in the reaction to biotrophic pathogens and necrotrophic pathogens, respectively (Glazebrook 2005; Pieterse et al. 2009). To study whether SA- and JA/ET-mediated pathways were also involved in B. oleracea resistance to Xcc, we examined the pattern of expression of hormone-related DEGs. We found that the JA biosynthetic pathway was significantly induced and that LOX1, LOX2, LOX3, OPR1 and OPR2 were up-regulated. Lipoxygenase (LOX) is a master regulator of the JA biosynthetic pathways (Wasternack 2007). In addition, some studies have reported that LOX1 was involved in the hypersensitive response to X. campestris in plants (Jalloul et al. 2002; Patil et al. 2005). Some key genes involved in ET biosynthesis (e.g., ACO1, ACO4, SAM2, and SAM3) and the signalling pathway (e.g., EIN3, ERF4, and ERF15) were up-regulated in QP07. However, we found that NIMIN1 and NIMIN2 were only up-regulated at 0 hpi for the SA-mediated pathway. Combined with previous data, our results suggest that the JA signalling pathway and not the SA signalling pathway may play a critical role in host resistance to Xcc.

Photosynthetic metabolism plays a significant role in the interactions between pathogens and host plants, and the host can act against pathogens by reducing photosynthesis (Berger et al. 2007; Garavaglia et al. 2010). Previous research showed that most photosynthesis-related proteins were increased after Xcc infection in susceptible plants (Santos et al. 2019a). Compared with DBP71, most of the differentially expressed genes were down-regulated in QPO7. However, the differentially expressed genes increased significantly at 48 hpi. This may indicate that the susceptible B. oleracea requires more energy to cope with infection by Xcc. Among these DEGs, two fructose-bisphosphate aldolase (FBA) genes were up-regulated at 48 hpi in QP07 compared with DBP71. Cai et al. (2016) reported that overexpression of the FBA gene in tomato significantly increased the expression level and activity of some enzyme genes in the Calvin cycle, and the rate of regeneration of RuBP increased significantly. In addition, the accumulation of malondialdehyde (MDA) and ROS decreased significantly. A population dynamics study of the interaction between Xcc and host plants has shown that population growth was observed only in the first 24 h after inoculation, followed by a significant reduction in the bacterial population up to 48 h (Villeth et al. 2009). Taken together, our analysis suggests that resistant plants may reduce the energy supply of Xcc by reducing photosynthetic metabolism, thereby inhibiting its growth. In addition, the resilience of the resistant plants seems to be greater than that of the susceptible plants, which can quickly restore their normal photosynthetic energy metabolism. This result was consistent with previous observations in proteomic analyses of the interaction between Xcc and B. oleracea (Ribeiro et al. 2018; Santos et al. 2019a).

ROS are vital signalling molecules that regulate the onset of the hypersensitive response (Levine et al. 1994). Moreover, the accumulation of ROS can be toxic to the pathogen by inhibiting and/or reducing its survival (Jones & Dangl 2006; Zhang & Zhou 2010). However, ROS accumulation can lead to biofilm system damage and the oxidation of important cell components (Lushchak 2011). To adapt to ROS toxicity, plants usually protect themselves against ROS by enzymatic and non-enzymatic antioxidant systems, such as catalase activity, glutathione metabolism and ascorbate peroxidase. The B. oleracea antioxidants include catalase (Bo3g167210, Bo7g117190), superoxide dismutase (Bo8g112320), glutathione peroxidase 5 (Bo4g098230) and several glutathione S-transferases, which were up-regulated at different time-points in QP07 compared with DBP71 (Supplementary Table 7). Similar results were observed in a proteome analysis of the interaction between B. oleracea and Xcc (Santos et al. 2019a). These results may indicate a balance of oxidative stress response in B. oleracea in response to Xcc infection, maintaining some genes by up-regulation and others by down-regulation, which may be important to efficiently control the pathogen without causing extensive damage to the plant.

Plants defend themselves against pathogens by synthesizing and accumulating metabolites with antibiotic properties (Pastorczyk & Bednarek 2016). Glucosinolates, secondary metabolites of cruciferous plants, are implicated in the defence against plant pathogens (Grubb & Abel 2006). Previous research showed that glucosinolates could enhance resistance against Xcc in B. oleracea. Aires et al. (2009) reported the effect of glucosinolate hydrolytic products (GHP) on Xcc and showed that GHP could inhibit the growth of this bacterium. Similar results were obtained in a study by Velasco et al. (2013). Recently, Madloo et al. (2019) evaluated the role of three glucosinolates (the aliphatics sinigrin, glucoiberin and the indolic glucobrassicin) in the defence of kale against Xcc, showing that only the indolic glucobrassicin can inhibit infection by Xcc. In this study, the expression of MYB28, FMO GS-OX2, BASS5, MAM2, GSL-OH and BCAT4, which encode proteins involved in the glucosinolate biosynthetic pathway, was up-regulated after Xcc infection. The double mutant MYB28/29, which does not accumulate aliphatic GSLs, has been shown to negatively affect the plant’s ability to defend against Botrytis cinerea and Alternaria brassicicola (Buxdorf et al. 2013). In addition, many key genes involved in the glucosinolate catabolic process (e.g., TGG2, UGT74 C1, ESP, NSP4, and NSP5) were also up-regulated in QP07 compared with DBP71. It is interesting that the up-regulation of the genes involved in glucosinolate biosynthesis and catabolic processes is concentrated at 48 h. This is similar to previous findings reported for the interaction between glucosinolates and Xcc (Velasco et al. 2013). Collectively, our results suggest that glucosinolates are synthesized and catabolized to enhance resistance to Xcc at the early stages of infection.

In this study, we comparatively analysed gene expression profiles in the leaves of Xcc-resistant and Xcc-susceptible lines of B. oleracea during the early stages of Xcc infection. A total of 7430 DEGs were identified. During this early phase, the glucosinolate biosynthetic and catabolic pathways were enhanced, and ROS scavenging, hormonal, nucleotide binding site (NBS)-encoding resistance (R) and receptor-kinase related genes were differentially expressed in response to Xcc. Taken together, the results suggest that upon infection by Xcc, B. oleracea activates its defences to restrict pathogen development and spread by (i) synthesizing antibacterial substances and reducing the energy supply of Xcc to inhibit its growth (e.g., enhancing GSL biosynthesis and catabolism, reducing photosynthetic metabolism), and (ii) maintaining a balance of in the host oxidative stress response to efficiently control Xcc.

Acknowledgements

We thank Professor Baohua Li and Professor Xiaofeng Wang from our University for experimental advice and critical reading of this manuscript. We also wish to thank the reviewers and editor for their valuable comments.

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

We are grateful for the funding of this work supported by the National Key Research and Development Programme of China [2017YFD0101804].