Sclerotinia sclerotiorum utilizes host-derived copper for ROS detoxification and infection

Necrotrophic plant pathogen induces host reactive oxygen species (ROS) production, which leads to necrosis in the host, allowing the pathogen to absorb nutrients from the dead tissues. Sclerotinia sclerotiorum is a typical necrotrophic pathogen that causes Sclerotinia stem rot in more than 400 species, resulting in serious economic losses. Here, we found that three S. sclerotiorum genes involved in copper ion import/transport, SsCTR1, SsCCS and SsATX1, were significantly up-regulated during infection of Brassica oleracea. Function analysis revealed that these genes involved in fungal ROS detoxification and virulence. On the host side, four genes putatively involved in copper ion homeostasis, BolCCS, BolCCH, BolMT2A and BolDRT112, were significantly down-regulated in susceptible B. oleracea, but stably expressed in resistant B. oleracea during infection. Their homologs were found to promote resistance to S. sclerotiorum and increase antioxidant activity in Arabidopsis thaliana. Furthermore, copper concentration analysis indicated that copper flow from healthy area into the necrotic area during infection. A model was proposed that S. sclerotiorum utilizes host copper to detoxify ROS in its cells, whereas the resistant hosts may restrict the supply of essential copper nutrients to S. sclerotiorum by maintaining copper ion homeostasis during infection.

genes involved in copper ion import and transport in order to utilize the host-derived copper to activate ROS detoxification enzymes during infection, and that resistant hosts may limit the supply of copper to S. sclerotiorum by maintaining copper ion homeostasis. This research provides new insights into the interaction between S. sclerotiorum and the host, highlighting the subtle and complex role of copper in these interactions.

Copper is involved in the interaction between Brassica oleracea and S. sclerotiorum
S. sclerotiorum induces typical lesions, which are the main battlegrounds of gene interactions between S. sclerotiorum and the host. We previously detected differentially expressed genes (DEGs) by comparing gene expression in lesions of resistant and susceptible F 2 plants of B. oleracea [25]. Here, the set of transcriptome data was analyzed for dynamic changes of gene expression in S. sclerotiorum and hosts during infection. A total of 738 and 228 S. sclerotiorum DEGs (24 hours post inoculation [hpi] vs 12 hpi) were detected in lesions of resistant and susceptible B. oleracea, respectively (S1A Fig), which were significantly enriched for three overlapping Gene Ontology (GO) terms, 'oxidation-reduction process', 'copper ion transport' and 'copper ion import' (Fig 1A). Eight S. sclerotiorum DEGs involved in the 'copper ion transport' and 'copper ion import' processes were up-regulated during infection as revealed by both RNA-seq analysis and quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) analysis (S1B and S2A Figs).
A total of 5988 and 5441 DEGs (24 hpi vs 12 hpi) were detected and subjected to GO analysis in resistant and susceptible B. oleracea, respectively (S1C Fig). Interestingly, the biological process 'copper ion homeostasis' was significantly enriched in susceptible B. oleracea but not in resistant B. oleracea (Fig 1B). Among ten DEGs involved in 'copper ion homeostasis' (S1D

Copper distribution in and around lesions during early infection
Considering that the 'copper ion transport' and 'copper ion import' processes are promoted in S. sclerotiorum during infection, we speculate S. sclerotiorum absorbs copper from lesion for infection, producing copper flows from the healthy area into the necrotic area in host. We analyzed the copper distribution in the different sites of healthy leaves, and in and around lesions of infected leaves in several hosts, including parental B. oleracea (C01 and C41) and two cultivars in B. napus (cv. Zhongshuang 11 and Westar) at 48 hpi and A. thaliana (Col-0) at 12 hpi. Although the copper concentration exhibit significant differences in the healthy leaves among five hosts (P < 0.05, 0.043 ± 0.001 μg/cm 2 in B. oleracea C01, 0.030 ± 0.001μg/cm 2 in B. oleracea C41, 0.038 ± 0.001μg/cm 2 in B. napus cv. Zhongshuang11, 0.036 ± 0.001μg/cm 2 in B. napus cv. Westar and 0.036 ± 0.001μg/cm 2 in A. thaliana Col-0), no significant difference for

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum the copper concentration was detected in the different sites of healthy leaves of each host ( Fig  1D). However, the copper concentration exhibited significant difference among the necrotic, margin and uninfected areas of leaves in each host (P < 0.05) (Fig 1D). The copper concentration in necrotic areas was 27.86% and 62.08% higher than that of margin areas, and 54.61% and 86.32% higher than that of uninfected areas in parental B. oleracea C01 and C41, respectively ( Fig 1D). Similar results that highest copper concentration in necrotic areas were detected in B. napus cv. Zhongshuang11 and Westar and A. thaliana Col-0 ( Fig 1D). These data indicate that copper flows from healthy area into the necrotic area during early infection.

Copper ion homeostasis genes promote host resistance
To test whether copper ion homeostasis is associated with host resistance, the six genes in this process that were stably expressed or slightly up-regulated in resistant B. oleracea but significantly down-regulated in susceptible B. oleracea were aligned with four A. thaliana orthologs (AtCCS, AtMT2A, AtDRT112 and AtCCH) (S3A Fig). We tested the function of these A. thaliana homologs with respect to S. sclerotiorum resistance using T-DNA mutants (atccs, atmt2a, atdrt112 and atcch) and overexpression lines (OX-AtCCS, OX-AtMT2A, OX-AtDRT112 and OX-AtCCH) (S3B and S3C Fig). Notably, all of the reduced-expression T-DNA mutants were more susceptible to S. sclerotiorum compared to the wild-type line, while the overexpression lines displayed higher resistance (Fig 2A). At 24 hpi, the lesion size was 1.28-1.35 cm 2 in the T-DNA mutant lines, 1.06 cm 2 in the wild-type line and 0.59-0.71 cm 2 in the overexpression lines ( Fig 2B). More interesting, the necrotic areas of the T-DNA mutants showed the highest copper concentration, followed by that of wild-type line and overexpression lines of four A. thaliana genes (Fig 2C).

Copper ion homeostasis is associated with response to oxidative stress in the host
To further explore how copper ion homeostasis is associated with host resistance, the transcriptomes of leaves from A. thaliana wild-type line, T-DNA mutants and overexpression lines of AtCCS, AtMT2A, AtDRT112 and AtCCH at 0, 6, and 12 hpi were sequenced, producing an average of 22.9 million clean reads for each sample. On average, 17.71 and 0.97 million clean reads were mapped to the reference genome of A. thaliana and S. sclerotiorum per sample, respectively. As expected, the overexpression lines exhibited higher expression of the corresponding target genes than the T-DNA mutants and the wild-type line (S4A Fig). Five common GO terms, including 'response to oxidative stress', were detected by comparing the up-regulated DEGs between overexpression lines and T-DNA mutants and between overexpression lines and wild-type line at 12 hpi (S4B Fig). Considering that more DEGs were found between the overexpression lines and T-DNA mutants than between the overexpression lines and wild-type line, we conducted a Weighted Gene Co-expression Network Analysis (WGCNA) using a total of 7321 A. thaliana DEGs between the overexpression lines and T-DNA mutants (S1 Table). This analysis produced 17 modules (groups of genes with similar expression pattern), of which one (shown in red) showed a negative correlation with lesion size (r = -0.45, P = 0.03) (S5A Fig). It encompassed 394 genes, which were significantly enriched for the biological process of 'response to oxidative stress' (S5B Fig). In addition, total of 1273 S. sclerotiorum DEGs (overexpression lines vs. T-DNA mutants) were detected and used for WGCNA, which resulted in five significant modules (P < 0.05), two of which were highly correlated with the processes of 'peroxisome organization' and 'oxidation-reduction' (S2 Table and  The oxidative burst is a typical response of a host against pathogen attack [26]. To functionally test whether copper ion homeostasis genes are associated with response to oxidative stress, we analyzed the antioxidant activity of these overexpression lines and T-DNA mutants by

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum staining inoculated leaves at 0, 6 and 12 hpi with nitrotetrazolium blue chloride (NBT) for the accumulation of �O 2 -. We observed deeply stained areas around the inoculant column and more accumulation of �O 2 -(1.32 ± 0.07 μmol/g FW at 6 hpi, and 1.31 ± 0.13 μmol/g FW at 12 hpi) in the T-DNA mutants, but more lightly stained areas and less accumulation of �O 2 in the overexpression lines (0.63 ± 0.15 μmol/g FW at 6 hpi, and 0.35 ± 0.09 μmol/g FW at 12 hpi) (Fig 3A  and 3B), suggesting that antioxidant activity was enhanced in the overexpression lines but suppressed in the T-DNA mutants. This conclusion was further supported by a Cu/Zn SOD enzyme activity assay, in which the overexpression lines (43.30 ± 1.26 and 56.53 ± 7.02 U/mgprot at 6 and 12 hpi, respectively) exhibited higher Cu/Zn SOD enzyme activity than T-DNA mutants (19.26 ± 3.32 and 19.48 ± 8.00 U/mgprot at 6 and 12 hpi, respectively) ( Fig 3C). Wild-type line was in the middle (32.14 and 42.82 U/mgprot at 6 and 12 hpi, respectively). These results indicate that copper ion homeostasis is associated with the detoxification of ROS in the host.

S. sclerotiorum requires trace copper for pathogen infection
Copper is an essential nutrient for microbial pathogens and serves as an important cofactor of enzymes that scavenge ROS [8,10]. To test the hypothesis that S. sclerotiorum takes up and utilizes host-derived copper for the synthesis or activation of enzymes involved in ROS scavenging during infection, we explored the roles of three S. sclerotiorum DEGs annotated as involved in 'copper ion transport/import': SS1G_05578 (SsCTR1), SS1G_00102 (SsCCS) and To test the role of copper in the virulence of S. sclerotiorum, we sprayed low-concentration CuSO 4 solutions (0.01, 0.025 and 0.05 mg/L) onto the leaf surface and incubated for 30 min prior to inoculation with each of the silenced strains. The lesion size after inoculation with the silenced strains increased by 1.19-to 1.35-fold at 48 hpi in comparison with the non-CuSO 4 treatment (sprayed with ddH 2 O), but exhibited no significant difference from that inoculated with the wild-type strain and sprayed with ddH 2 O ( Fig 4C). However, treatment sprayed with highly concentrated CuSO 4 solutions (0.1 and 0.25 mg/L) significantly reduced disease symptoms of the silenced strains ( Fig 4C). Similar observations were detected in the wild-type strain and overexpression stains, which increased lesion size under low concentration of CuSO 4 , but decreased lesion size under high concentration of CuSO 4 ( Fig 4C). We further found that lowconcentration CuSO 4 promoted the growth of all the strains on potato dextrose agar (PDA) plates, and that high-concentration CuSO 4 suppressed the growth of S. sclerotiorum (S7 Fig). Especially, the growth inhibition was higher in the silenced strains than the overexpression strains under high-concentration CuSO 4 (0.25 mg/L) (S7 Fig). It suggests that low-concentration copper could promote the growth of strains and restores the virulence of the silenced strains. the standard deviation of six independent replicates with five leaves for every sample in one replicate. � : represents significant difference from WT at 0.05 level (Student's t-test). (C) Copper concentration in uninfected (U), margin (M), necrotic (N) and mock tissues of A. thaliana WT, T-DNA mutants and overexpression at 12 hpi. Error bars indicate the standard deviation of three independent replicates with twenty leaves for every sample in one replicate. Different letters indicate statistically significant differences (P < 0.05, Student's t-test). https://doi.org/10.1371/journal.ppat.1008919.g002

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum Furthermore, the copper concentration in the leaves infected by the silenced and overexpression strains of these three genes was determined (Fig 5). Similarly, the necrotic areas exhibited the higher copper concentration than the uninfected and margin areas of leaves. The copper concentration in the necrotic areas infected by the wild-type strain was 8.37%-13.09% higher than that of the silenced strains, and was 7.99%-11.27% lower than that of the overexpression strains (Fig 5). It suggests that SsCTR1, SsCCS and SsATX1 are in association with copper uptake in S. sclerotiorum during infection.

Copper ion transport/import involves in ROS scavenging in S. sclerotiorum
To further test the hypothesis that S. sclerotiorum utilizes copper to promote ROS scavenging, we observed superoxide (�O 2 -) accumulation among wild-type strain, the silenced and overexpression strains of SsCTR1, SsCCS and SsATX1 using NBT staining. The accumulation of �O 2 in the fungal hyphal tips was highest in the silenced strains, followed by the wild-type strain and the overexpression strains (Fig 6A and 6B). We further found that exogenous spraying of low-concentration CuSO 4 solutions (0.01, 0.025 and 0.05 mg/L) decreased the accumulation of   6C). These observations indicate that ROS can inhibit the growth of S. sclerotiorum and that these S. sclerotiorum genes are involved in the detoxification of ROS. Because Cu/Zn SOD is one of the primary superoxide scavengers [16], we analyzed the SsSOD1 expression level in the wild-type strain, silenced and overexpression strains. The SsSOD1 transcript accumulation of wild-type strain was 1.51, 1.86 and 1.37 times of silenced strains, and 0.61, 0.50 and 0.57 times of overexpression strains of SsCTR1, SsCCS and SsATX1 on average, respectively ( Fig 6D). In comparison with the wild-type strain (572.51 U/g FW), the Cu/Zn SOD enzyme activity was lower in silenced strains of SsCTR1, SsCCS and SsATX1 with average of 506.71, 353.51 and 415.23 U/g FW, and higher in their overexpression strains with average of 726.84, 881.30 and 759.57 U/g FW, respectively (Fig 6D). These results indicate that these three S. sclerotiorum copper-related genes promote the expression of SsSOD1 and increase enzyme activity of Cu/Zn SOD, resulting in the increase of ROS scavenging capacity in the fungal cells.

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum

Discussion
Copper is a component of numerous enzymes and plays a key role in the responses to oxidative stress [20,29]. By analyzing DEGs at lesion for dynamic changes in host and S. sclerotiorum during infection, we found that the genes in the process of 'copper ion import' and 'copper ion transport' with up-regulated expression involved in S. sclerotiorum copper uptake, virulence and ROS detoxification, and that the host genes in the process of 'copper ion homeostasis' with stable expression in the resistant line, but down-regulated expression in the susceptible line were associated with response to oxidative stress and resistance to S. sclerotiorum. Our data indicate a copper battlefield at the host-S. sclerotiorum interface where S. sclerotiorum utilizes host-derived copper to successfully detoxify ROS and colonize the host.
The generation of ROS at the infection site is one of the earliest responses of pathogen-associated molecular-pattern-triggered immunity (PTI) [26]. As secondary messengers, ROS are indispensable for signaling, stress responses and developmental processes [30]. However, excess ROS trigger programmed cell death (PCD) and cause host necrosis, which facilitates the growth of necrotrophic pathogens [23, 31,32]. Antioxidant components in the host, such as peroxidase, SODs, glutathione sulfhydryl transferase (GST) and glutathione (GSH) are associated with resistance against S. sclerotiorum [33][34][35][36][37][38], and resistance against S. sclerotiorum can be improved by decreasing the accumulation and/or production of ROS in the host [39,40]. The four DEGs in the biological process of 'copper ion homeostasis' we detected during infection are involved in various antioxidant activities. AtCCS is responsible for the activation of Cu/Zn SOD [41]. Metallothioneins (MTs), which act as heavy metal chelators and ROS scavengers, contribute to plant adaptation to abiotic stresses [42]. DRT112 is one of two A. thaliana plastocyanin genes (PETE2), which function to buffer excess copper [14]. AtCCH is a homolog of yeast ATX1, which functions to deliver copper into laccase or Fet3 [21]. In this study, we found that the overexpression of these genes in A. thaliana enhanced host ROS detoxification and resistance against S. sclerotiorum, and the genes in the biological process of 'copper ion homeostasis' were coordinately expressed with those for 'response to oxidative stress'.

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum Therefore, our research provides evidence that copper ion homeostasis is associated with ROS detoxification in host.
Free copper ions induce ROS production [43], which may be toxic to S. sclerotiorum. We found that highly concentrated CuSO 4 solutions suppressed the growth of S. sclerotiorum. In fact, copper is one of important active ingredients in many bactericidal and fungicidal agents, such as Bordeaux mixture [44]. However, S. sclerotiorum can grow in necrotic areas with a relatively high concentration of copper. Three aspects of our findings might help explain this phenomenon. (1) The copper uptake system of S. sclerotiorum was elevated during infection. SsCTR1, which functions to import extracellular copper into the fungal cells [45], exhibited up-regulated expression, indicating that host-derived copper may be imported into the S. sclerotiorum cells. (2) A few genes associated with copper detoxification, such as MTs (SsMT), Sur7 (SsSUR7) and P-type ATPase (SsATP7A) [46][47][48], were not significantly induced in S. sclerotiorum during infection (S9 Fig), indicating that the amount of host-derived copper was not high enough to produce toxic effects in S. sclerotiorum. (3) SsCCS and SsATX1 expression was up-regulated during infection. The homologs of SsCCS function to deliver copper to oxidative scavengers Cu/Zn SOD [9,27]. SsSOD1 is identified as an important virulence factor of S. sclerotiorum and plays critical roles in detoxification of ROS during host-pathogen interactions and is an important virulence factor of S. sclerotiorum [49,50]. The ATX1 gene could act as a multi-copy suppressor of oxidative damage in yeast cell lacking SOD1 [21]. It indicates that the copper might be utilized for the synthesis or activation of these ROS detoxification enzymes of S. sclerotiorum during infection. Thus, this study provides important insights into the question about survival of S. sclerotiorum at relatively high levels of ROS.
The idea of 'nutritional immunity' was first proposed to describe the resistance mechanism in human cells wherein they withhold transition metals, such as iron, zinc, manganese and copper to defend against microbial pathogen invaders [51][52][53][54]. To overcome this strategy, successful pathogenic species must evolve specialized mechanisms to adapt to the nutritionally restrictive environment of the host and cause disease. For example, the human kidney and brain can thwart Candida albicans growth by limiting the supply of copper nutrients, producing 'copper starvation' to C. albicans [51]. In response, C. albicans induces copper uptake machineries that enable it to survive in a copper-starved environment [55]. Here, the defense response in resistant lines that stabilized the expression of copper ion homeostasis genes and limited the availability of copper to S. sclerotiorum may be considered a form of nutritional immunity. To the best of our knowledge, this is the first description of nutritional immunity in plants. We therefore propose a possible host-S. sclerotiorum interaction model in which resistant plants induce nutritional immunity and restrict the supply of essential copper nutrients to S. sclerotiorum by maintaining copper ion homeostasis, while S. sclerotiorum enhances its copper uptake system and acquires host-derived copper, which activates its ROS scavenging system during infection and promotes its survival and virulence (Fig 7).

Experimental strains and plants
The wild-type strain of S. sclerotiorum 1980 [56] were used in this study. Fungal strains were grown on potato dextrose agar (PDA, 20% potato, 2% dextrose and 1.5% agar) at 22˚C. S. sclerotiorum transformants were cultured on PDA amended with 80 μg/ml hygromycin B (Calbiochem, quantity of S. sclerotiorum SsTubulin cDNA was used to normalize different samples. Error bars indicate the standard deviation of three independent samples. � : represents significant difference from wild-type strain 1980 at 0.05 level (Student's t-test). https://doi.org/10.1371/journal.ppat.1008919.g006

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum San Diego, CA) to stabilize the transformants. Seedlings from A. thaliana Col-0 (ecotype Columbia-0), T-DNA mutants and overexpression transgenic lines were grown in the autoclaved soil (Pindstrup) at 20 ± 2˚C under a 12 h light/dark cycle with 70% relative humidity.

Vector construction and transformation of S. sclerotiorum and A. thaliana
The fragments of three S. sclerotiorum genes (SsATX1: 221 bp, SsCCS: 221 bp and SsCTR1: 461 bp) were PCR-amplified from the cDNA library of S. sclerotiorum wild-type strain by using

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum the specific primers of RNAi vector construction in S3 Table. The sense and antisense fragments were ligated into the plasmid vector pCIT [57] at the corresponding sites, and a hygromycin resistance gene cassette from pSKH [58] was isolated and ligated into them, resulting the RNAi vectors, pSiatx, pSiccs and pSictr1. To generate S. sclerotiorum overexpression strains of these three genes, the full-length coding sequences were amplified by using the specific primers of overexpression vector construction in S3 Table and cloned into a hygromycin resistance containing vector, driven by the S. sclerotiorum constitutive expression promoter EF1-A (SS1G_06124, translation elongation factor 1 alpha), resulting the overexpression constructs, pOXSsATX1, pOXSsCCS and pOXSsCTR1. S. sclerotiorum transformations were conducted with a standard polyethylene glycol (PEG)-mediated transformation protocol [59].
The T-DNA mutants of A. thaliana (AtCCS: SALK_025986C, AtCCH: SALK_118605C, AtMT2A: SALK_021037C, and AtDRT112: SALK_135199C) were obtained from the Arabidopsis Biological Resource Center, Ohio State University, USA. The homozygous T-DNA insertion lines were confirmed with primers flanking the T-DNA insertions (S3 Table) and the left border primer LB1.3 (ATTTTGCCGATTTCGGAAC). To generate the overexpression lines, the coding sequences of these four genes were amplified from a cDNA library of A. thaliana Col-0 leaves. The amplicons were digested with Xba I and Xho I and ligated into the plant expression vector pBinGlyRed3, which contains a gene encoded red fluorescent protein (DsRed). The resulting vectors (pBinGlyRed-AtCCS, pBinGlyRed-AtCCH, pBinGlyRed-AtMT2A and pBinGlyRed-AtDRT112) were introduced into the Agrobacterium tumefaciens strain GV3101 by electroporation [60] and transformed into A. thaliana Col-0 by the floral dip method [61].

RNA-seq and network analysis
In our previous study [25], the stems of resistant and susceptible B. oleracea plants in a F 2 population which derived from the cross between a resistant B. oleracea genotype C01 (B. incana) and a susceptible B. oleracea genotype C41 (B. oleracea var. alboglabra) were inoculated with S. sclerotiorum, and lesions at 0, 12 and 24 hpi were collected for transcriptome sequencing. To analyze the dynamic changes during infection, here we analyzed the DEGs of S. sclerotiorum and B. oleracea by using the DESeq package [62].
To reveal the pathways associated with 'copper ion homeostasis', the RNA of infected leaves was sequenced from A. thaliana wild-type line, T-DNA mutants and overexpression lines during infection. Briefly, the detached leaves were inoculated with S. sclerotiorum wild-type strain, and RNA from the lesions was extracted with the RNAprep pure Plant Kit (DP 432, Tiangen Biotech (BEIJING) CO., LTD). The sequencing library was generated using the Illumina RNA Library Prep Kit (NEB, USA) following the manufacturer's recommendation, and sequenced on an Illumina HiSeq 4000 platform with three biological replicates. After removing low-quality reads and those with adapter sequences, poly-N sequence from the raw data, the clean reads were screened and aligned to the reference genomes of A. thaliana (https://www. arabidopsis.org/download/index.jsp) and S. sclerotiorum (http://fungidb.org/common/ downloads/Current_Release/Ssclerotiorum1980UF-70/) by using the TopHat program (http:// ccb.jhu.edu/software/tophat/index.shtml) [63] with default parameters except that the Q value was set to 100. Gene expression was quantified using htseq-count 0.6.1p2 (https://htseq. readthedocs.io/). The raw counts were normalized by TMM normalization using the edgeR package [64] and the differential expression analysis was carried out using the DESeq package [62].
The threshold determining the significance of DEGs among multiple tests was set at a false discovery rate (FDR) � 0.001 and |log 2 ratio| � 1. GO (Gene Ontology) and KEGG (Kyoto

Measurement of copper concentration
In order to test copper flow in the leaves infected with S. sclerotiorum, we measured the copper concentration in hosts, including B. napus cv. Zhongshuang 11 with the moderate resistance [68], B. napus cv. Westar with high susceptibility, parental B. oleracea (C01 and C41) and A. thaliana lines. The detached leaves were inoculated with S. sclerotiorum strains, and the uninfected (U), margin (M) and necrotic (N) area of leaves were collected to measure area size by ImageJ (https://imagej.nih.gov/ij/). The tissues were washed with distilled water, dried for one week at 80˚C, and then washed with 11 N HNO 3 . Copper content of tissues was measured using atomic absorption spectroscopy (SPECTR AA220) with flame at a wavelength of 324.8 nm. The copper concentration was determined with copper content per area (ug/cm 2 ).

Quantitative RT-PCR
Gene expression was analyzed by qRT-PCR using a Bio-Rad CFX96 Real Time System (Bio-Rad, USA) and QuantiTect SYBR Green PCR master mix (Bio-Rad, USA), according to the manufacturer's instructions. The SsTubulin and BoActin3 genes were used as the internal control for S. sclerotiorum and B. oleracea, respectively. All the qRT-PCR primers were listed in S3 Table. The PCR cycling conditions comprised 1 cycle of 95˚C for 30 s, then 39 cycles of 95˚C for 5 s and 55-70˚C for 1 min, followed by a melting curve ramping from 65˚C to 95˚C with temperature increasing by 0.5˚C every 5 s (1 cycle). Transcript levels were calculated from the threshold cycle using the 2 -ΔΔCT method [69]. Three replicates of same sample were performed for each gene and data were analyzed using CFX Manager v3.0.

Pathogenicity assays
Pathogenicity of S. sclerotiorum was evaluated by infecting B. napus and A. thaliana according to the procedure described previously [25,37]. The detached leaves at seedling stage of B. napus and A. thaliana were inoculated with mycelium-colonized agar plugs (0.6 cm for B. napus, 0.2 cm for A. thaliana) obtained from expanding margins of PDA-cultured colonies. The inoculation chamber was maintained at 85% relative humidity at 22˚C. The lesion size (S, cm 2 ) for the leaves was calculated with the formula S = π�a�b/4, where a and b represent the long and short diameter of an approximately elliptical lesion.

Antioxidant activity assays
The antioxidant activity of S. sclerotiorum hyphae and A. thaliana leaves was determined by NBT staining [70]. Hyphae on the PDA plate at 24 hpi were infiltrated under gentle vacuum with NBT staining solution for 5 hours and then washed 3 times with distilled water, prior to observation with a microscope. A. thaliana leaves were infiltrated under gentle vacuum with NBT staining solution for 5 hours and then the staining solution was replaced with bleaching solution (ethanol: acetic acid: glycerol = 3:1:1). After 15 ± 5 min in a boiling water bath (~90-95˚C), the bleaching solution was replaced with fresh bleaching solution and stained in 60% glycerin. The quantitative of �O 2 accumulation was tested using the superoxide anion assay kit

PLOS PATHOGENS
Association of copper with ROS detoxification in S. sclerotiorum (A052-1-1; Nanjing Jiancheng Bioengineering Institute) according to the manufacturer's protocol. The enzyme activity of Cu/Zn SOD of S. sclerotiorum hyphae and infected A. thaliana leaves was tested using a Cu/Zn SOD assay kit (A001-4-1; Nanjing Jiancheng Bioengineering Institute) according to the manufacturer's protocol.